UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Strength testing variables in dental ceramics

Wang, H.

Link to publication

Citation for published version (APA):Wang, H. (2008). Strength testing variables in dental ceramics.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 29 Dec 2019

Strength testing variables in dental ceramics

Hang Wang

Strength testing variables in dental ceramics

H

. Wang

2008

I N V I T A T I O N to attend the public

defense of the thesis of

Hang Wang

Monday, April 28, 2008 at

14:00

in the aula of the

Universiteit van Amsterdam

Agnietenkapel

Oudezijds Voorburgwal 231

Amsterdam

Strength testing variables

in dental ceramics

Hang Wang

Strength testing variables in dental ceramics

Hang Wang

Printed by: Ponsen & Looijen BV, Wageningen Copyright: © H. Wang All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, mechanically, by photocopy, by recording or otherwise, without permission by the author.

Strength testing variables in dental ceramics

ACADEMISCH PROEFSCHRIFT Ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam

op gezag van de Rector Magnificus

prof. dr. D.C. van den Boom

ten overstaan van een door het college voor promoties ingestelde

commissie, in het openbaar te verdedigen in de Agnietenkapel

op maandag 28 april 2008, te 14:00 uur

door

Hang Wang

geboren te Baoji, Shaanxi, China

Promotiecommissie Promotor : Prof. dr. A.J. Feilzer Co-promotor : Dr. M.J. Frijn

Overige leden : Dr. S.S. Scherrer, PD

Prof. dr. H.F. Kappert

Prof. dr. ir. J.M. van der Zel

Faculteit der Tandheelkunde

This thesis was supported by a grant of the Chinese government and IOT/ACTA and prepared at the Department of Dental Materials Science at the Academic Center for Dentistry Amsterdam (ACTA), Universiteit van Amsterdam and Vrije Universiteit, the Netherlands.

CONTENTS

Chapter 1 Introduction 9

Chapter 2 Influence of Test Methods on Fracture Toughness of a

Dental Porcelain and a Soda Lime Glass

35

Chapter 3 Fracture toughness comparison of three test methods with four dental

porcelains

53

Chapter 4 Fracture toughness determination of two dental porcelains with the

indentation strength in bending method

67

Chapter 5 Indentation-strength fracture toughness: the role of the indentation load at

smaller flaw sizes

81

Chapter 6 Strength influencing variables on

CAD/CAM zirconia frameworks

97

Chapter 7 Staircase evaluation of the fatigue strength of sandblasted zirconia 113

Chapter 8 Summary and conclusions

Samenvatting en conclusies

125

Acknowledgment 131

9

CHAPTER 1

Introduction

Chapter 1

10

1.1.1 Strength testing variables in dental ceramics Understanding the mechanical properties of ceramics and porcelains is essential

for the explanation of the effects of applied forces on these materials and their responses to the forces. For that reason an overview of the structural as well as mechanical properties of dental ceramics is presented in this chapter. 1.1.2 Ceramics and porcelains

The first successful applications, in 1774, of ceramic materials in dentistry are all porcelain dentures. Until that moment tooth-colored substitutes of lost teeth had mainly been made from ivory, bone, animal teeth, and even extracted human teeth. Aided by the achievements of science and technology, the translucency and color of dental porcelains were improved and production techniques for dental restorative applications developed. Fused porcelain inlays and crowns were introduced around 1886, however, without any worry over the structural weakness [1, 2].

A milestone in the widespread application of dental porcelains was reached in 1950’s [3] with the invention of porcelains with a high coefficient of thermal expansion due to the addition of leucite, which enabled them to be fused to supporting structures made from gold alloys. This new and layered structure, with the patent title of ‘porcelain fused to metal’ (PFM) crowns [4, 5] and the practical name ‘metal-ceramic’, combined the favorable properties of metals and porcelain. This new technology and combination of materials enabled the fabrication of complete crowns and fixed partial dentures with a reliable durability. The rigid metal framework resists the firing contraction of porcelains and guarantees a precise fit on the prepared tooth. In the next decades metal, porcelain, their bonding, and their fabricating technology were developed further to reach the level of a clinical standard [6-8]. However, metal alloys also have disadvantages with regard to biocompatibility and esthetics. As the main reason to combine ceramics with metals is the lack of strength of the dental porcelains, much effort has been taken to develop strong ceramic materials to replace the metals.

Generally, porcelains belong to ceramics. Ceramics are classified into crystalline and non-crystalline ceramics. Non-crystalline ceramics are composed of glass that has been enriched with alumina or other powders, and are called ‘dispersion-reinforced porcelains’ or ‘glass-ceramics’. This second phase in glass results from devitrification or addition of particles or whiskers. Most crystalline ceramics are only partly crystalline as they are mainly ceramic composites; a glass matrix phase and a reinforcing filler fraction. A special group of dental ceramics is fully crystalline as

Introduction

11

they lack a matrix phase. In dentistry the term ‘polycrystalline’ typically refers to ceramics, which are fully crystalline.

In the middle of the 1960’s the introduction of alumina as a reinforcement phase started the era of development of all-ceramic crown systems. First for the all-ceramic Jacket crown made of conventional dental porcelain, which was reinforced with alumina [9-11]. These restorations were only reliable in the anterior, low stress bearing region. Although products, like Hi-ceram, Vitadur N systems, were introduced in the 1970’s [12], it was only in the 1980’s before all-ceramic crown systems came up to clinical acceptability [13]. Among these early representatives were the “shrink-free” Cerestore [14] and a castable glass-ceramic Dicor [15]. Clinical experience with both systems showed that they were still not strong enough for application in the posterior region of the mouth [16-20]. When Empress 1, with its hot-press technique enabling a precise fit of restorations, came on the market the indication area for all-ceramic restoration was extended to the premolar region [21-23].

Prefabricated, i.e. already fired porcelains, glass-ceramics, and CAD/CAM were developed in the 1980’s. Milling the dental restoration from prefabricated ceramic blocks that do not require further sintering, effectively prevents high contraction during firing, and the amount of intrinsic flaws is reduced in this production method. The CAD/CAM-based production method, which is required for this method provides a clean and fast fabrication, and enables chair-side production of small all-ceramic restorations [24]. The pronounced development of full range dental all-ceramic systems began in the 1990’s when materials with enhanced mechanical properties like Empress 2, In-Ceram series, and Procera Allceram with strength’s of 300 to 700 MPa became available.

Empress 2 utilizes the hot-press production technology, equal to Empress 1 and shows an improved strength because it contains more than 60 vol% lithium disilicate interlocking grains. This material can be used in the anterior and premolar region for both crown and three-unit bridge restorations [25, 26].

In-Ceram series is a novel ceramic named ‘interpenetrating phase composite’, which combines glass and crystalline ceramic in three-dimensional and structural continuity. The size characteristics of the powder and the sintering procedure result in a partially sintered porous framework with more than 80 vol% crystals, which is subsequently infiltrated with molten glass by capillary action. Glass offers transmissivity and color as well as eliminates the pores. The main disadvantage of this system is the need to create enough space for the restoration especially at the margin to

Chapter 1

12

obtain enough structural resistance to fracture. The three products in the In-Ceram series are Spinell, Alumina, and Zirconia, in order of increasing strength or decreasing translucency. The recommended indications are anterior crowns, anterior crowns and three-unit bridges combined with posterior crowns, posterior crowns, and three-unit bridges, respectively, according to the trade-off of the mechanical and esthetic qualities. In-Ceram Zirconia, as radio-opaque as metal was the first all-ceramic system recommended for short bridges in the posterior region [27-34].

The first polycrystalline ceramic used in dentistry was the alumina-based Procera AllCeram composed of micrometric pure alumina grains. Applying CAD/CAM technology to produce enlarged green bodies solved the problem of the large sintering shrinkage of polycrystalline ceramics. This kind of alumina is ivory white but to some extent also translucent. Despite the remarkable mechanical properties and clinical results, this technique was not suitable for bridge fabrication due to the dry pressing forming technique on enlarged die models [35-42].

A milestone of this development was reached at the end of the last century when poly-crystalline yttria stabilized zirconia, e.g products like Cercon, LAVA, and Everest, was introduced in dental industry. Zirconia ceramics are usually rather strong, above 900 MPa [43-48]. The main difference from other ceramics is its toughening mechanism, based on grain phase transformation. Depending on the temperature and external pressure zirconia has three crystallographic phases. With pure zirconia only the monoclinic phase (m) is stable at room temperature and pressure. The tetragonal phase (t) exists between 1170°C and 2370°C while the structure becomes cubic above 2370°C. So after pure zirconia is sintered at least at 100°C higher than 1170°C, it undergoes a phase transformation during cooling from the tetragonal to the monoclinic phase, which is accompanied by a volume increase of approximately 4.5%. Dopant additions such as CaO, MgO, Y2O3, and CeO2 dramatically reduce the transformation temperature and stabilize the tetragonal phase at clinical temperatures. Stress may transform the stabilized tetragonal phase to monoclinic. This unique transformation in response to stress is very favorable, as it generates compressive stress, which contributes to crack arrest and superior mechanical properties. Due to the strength and toughness they are suitable as a substrate for veneering porcelain, as recommended for example as copings of crowns and frameworks of bridges in both the anterior and posterior regions. These materials became a most promising candidate for substitution of metal in esthetic restorations. The zirconia ceramics are often white and opaque, but they can be stained to be dentin-colored.

Introduction

13

1.2.1 Mechanical aspects of dental ceramics Essential for the mechanical properties is to understand that most materials

applied in dentistry are composite materials in which a matrix material is ‘reinforced’ with filler particles. In glass ceramics the matrix is composed of an amorphous glass phase while the fillers are composed of crystalline glass or ceramic materials like leucite, alumina, or zirconia. The production of dental structures with these materials is achieved by sintering the milled ceramic powders, which are heated until the matrix melts and the filler particles fuse. The sintering causes shrinkage in excess of the thermal shrinkage from cooling to room temperature. The ability of the glass matrix to flow allows this type of materials to be condensed to metal structures without fracturing during cool-down to room temperature. This is not the case for the more homogenous poly-crystalline alumina or zirconia ceramics, in which no matrix phase is present. For this reason, these materials have to be shaped into an enlarged model in the green state, which shrinks down to a precise fitting shape after sintering. As in general for composites, the tensile strength is much lower than the compression strength. In tensile stress the weakest link in the materials, i.e. the glass phase, is loaded, while in compression the stronger filler is loaded. For poly-crystalline materials the inter-grain interface is probably the weakest link in tensile loading. The structural difference between these types of ceramics is of major importance for understanding their fracture behaviour [49]. 1.2.2 Stress-strain relations

The deformation of a material with an applied force is described as a stress-strain curve, which is one of the most important material parameters. From a stress-strain relationship the elastic modulus, fracture strength, and yield strength can be determined. Brittleness can be identified by the lack of deformation before fracture. Ceramics, porcelains, and glasses are brittle materials, which fracture elastically already at little elastic strain [49, 50]. 1.2.3 Strength

The strength (σf) of a material refers to the material's ability to resist an applied force. It is defined as the maximum or critical stress that a material can endure before fracture. The incapability of ceramics, as brittle materials, to relieve tensile stresses by plastic deformation at the tip of arising cracks explains their low resistance to tensile forces compared to compressive forces. From this point of view, the tensile strength is more meaningful than the compressive strength for dental ceramics. Routinely, the

Chapter 1

14

tensile fracture strength of brittle materials is measured with 3-point, 4-point or biaxial flexural tests [51-53]. In clinical dentistry, it was proved that stronger restorative ceramics including structurally uni-layered or substrate ceramics, had a higher survival rate [54-58].

For brittle materials as ceramics or porcelains, the outcome of fracture strength tests is a range of widely scattered values. According to ‘Weakest Link’ theory, fracture of a brittle material always starts at the biggest flaw in the loaded area. Due to the uncertainty of the distribution of those flaws in size, type and location in the material, the measured strength depends on the individual composition of a specimen and its flaw characteristics. To assess the strength distribution, failure probability was introduced as a statistic assessment, which is based on Weibull’s theory stating that the strength depends on the weakest element [59]. For ceramic materials, the so-called ‘strength-control’ or ‘strength-limiting flaw’ is the weakest element, capable of initiating ultimate fracture at a given stress, depending on size, shape, direction, and the location where the stress is applied. Using the relationship of probability with strength analysed with the Weibull approach, the ‘failure probability’ or ‘risk of rupture’ of a structural specimen at any applied stress may be predicted. The strength evaluation in which the impact of the tensile stress status (magnitude, distribution) is taken into account on fracture probability also considers the influence of specimen volume and surface area on the chance of a large flaw [60, 61]. For example, for specimens with same size and shape, the measured strength in a 4-point bending test might be lower than in a 3-point test. The Weibull analysis therefore offers a good risk and reliability evaluation for a material as well as a structure. The Weibull modulus characterises the distribution or variability of strength. A lower value of m indicates a greater range of strength values, while a higher m suggests a smaller range, thus a better reliability and safety. 1.2.4 Fracture toughness

With brittle materials fatal fracture is caused by a crack, which propagates through the material until the integrity of structure is lost. As a consequence the strength of a material may be considered as the stress required to initiate and propagate a crack to the fracture point. The driving “force” for a crack to grow is the stress intensity factor K near the crack tip [60-67]. The stress intensity factor K at a given stress σ is related firmly as the following equation:

cYK σ= or cY

K=σ

Introduction

15

where Y is a constant depending on the geometry and location of the crack, crack shape and loading configuration, and c is the radius of the crack tip. The stress intensity factor is positively proportional to stress. When stress intensity factor rises sufficiently, a crack is propagated leading to structural failure. In tensile or mode I (mode I: tensile opening; mode II: parallel sliding; mode III: lateral tearing) loading the critical level of stress intensity factor is denoted as KIc, which indicates the ability of a material to resist crack propagation and its consequent catastrophic failure.

KIc or the fracture toughness is a quantitative way of expressing the material resistance to brittle fracture and one of the most important material properties in fracture mechanics of brittle materials. For ceramics fracture toughness is therefore more elucidating than the strength. Strength of a brittle material depends on flaw size and its values are statistically scattered as a consequence of the flaw size distribution. Meanwhile, the fracture toughness is theoretically or generally a stable parameter independent of the crack size. If a material has a high fracture toughness it will probably undergo ductile fracture. Brittle fracture is characteristic of materials with low fracture toughness. High fracture toughness materials have improved clinical performance and reliability over low fracture toughness ones [68-70]. In dental ceramics, fracture toughness assessment may help to evaluate the suitability in long-term clinical performance of brittle materials. 1.2.5 Fatigue

Fatigue is another mode of failure, which occurs after repeated or continuous loading at stress levels smaller than the strength measured in one-load-failure tests. For dental ceramics, the term is related to cyclic loading in oral cavity [71, 72]. In applications of structural materials, this delayed failure property is most important for the durability of the structure. Fatal failure, especially for brittle ceramics may occur without a clear symptom [49, 50]. Early all-ceramic restorations failed at a high rate, but at a stress lower than material strength.

Progressive stages of fatigue damage, derived from a detailed classification are given [50] as:

a. Substructural and micro-structural changes, which cause nucleation of permanent

damage. b. The creation of microscopic cracks. c. The growth and coalescence/fusion of microscopic crack to form a dominant crack

that potentially or eventually lead to fatal failure. From a practical standpoint, this

Chapter 1

16

stage is the demarcation between crack initiation and propagation. d. Stable propagation of the dominant macroscopic cracks. e. Structural instability or complete fracture. With knowledge of fracture toughness and strength, it is clear that structural failure due to fatigue loading results from crack growth from initial size till critical in stress concentration areas. During this event smaller cracks may fuse or individual cracks become larger. The growing crack decreases the strength of the structure until the final loading cycle causes complete failure. 1.2.6 Crack initiation

Based on their origin, three categories of crack initiation flaws may be recognized. Intrinsic flaws are pores, agglomerates, inclusions, irregularity of grain distribution, and micro-cracks due to release of thermal residual stresses at grain boundaries and interfaces [49, 50, 71-77]. Scratches, grooves, pits, notches, fissures, usually present on or near surface, are extrinsic sources and often are damage from machining, grinding, and sandblasting. Crack initiation during service is a dynamic process, which arises from defects, which may already be micro-cracks or will lead to micro-cracks. Any defect may grow to complete fracture. Dental ceramics and porcelains are subjected to a mechanically, chemically, thermally, and tribologically rigorous environment [49, 50, 71-85]. 1.2.7 Cyclic loadings and cyclic contact damage

Serious loss of ceramic strength may be caused by cracks that arise from accumulated damage in cyclic loading in aqueous environments. This mechanism will be mentioned in the part of crack growth. Cyclic loading may lead to irreversible microscopic deformations, which engenders different failure mode. Under these loading conditions, the accumulated permanent strain inevitably increases pre-present damage. Micro-cracks may nucleate at grain boundaries in single-phase materials, interphase regions in multiphase, and along interfaces between matrix and filler in brittle composites [50].

Cyclic uni-axial contact damage is well modelled in the Hertzian contact test, in which spherical indenters apply a load to a flat surface. A major advantage of controlled contact testing is that simple analytic relations can be derived for critical loads in terms of layer thickness, indenter radius, and basic material properties such as modulus, hardness, toughness, or strength [86].

Introduction

17

Under cyclic uni-axial contact loading two basic damage modes are identified. The tensile-driven brittle mode produces single “cone” cracks that initiate from the surface; the other shear-driven quasi-plastic mode leads to distributed micro-cracks from the subsurface Hertzian “yield” zone. The former occurs mainly in brittle glasses, porcelains, and fine-grained ceramics. The quasi-plastic mode is predominant in coarse-grained tougher ceramics such as alumina- and zirconia-based ceramics. Both damage modes increase with cyclic loading; the brittle mode by slow extension of the cone crack, and the quasi-plastic mode by relatively rapid micro-crack coalescence, ultimately leading to the formation of subsurface radial cracks.

In layered all-ceramic structures, a more dangerous mode is interior radial cracking at a layer boundary, which is reported to derive from pre-existing flaws. Bi-axial cyclic contact loading modes like sliding or fretting in wear processes may cause severe surface ageing on the relatively weak veneering porcelains [87]. Uni-axial cyclic contact fatigue studies are well reported in literature [86]. 1.2.8 Wear

Wear is another source of unwanted cracks and surface fracture, which may grow under occlusal force with time, while their propagation enhances mechanisms like chemical erosion. Subsequent degradation of the surface microstructure, accumulation of unwanted residual stress, irreversible strain, initiation and growth of cracks are in the process [78-84, 88-90].

Chemical or corrosive wear is caused by aggressive chemical agents like acids, bases, and enzymes, in the food, drinks, and saliva [85]. For example, acid etching and leaching can lead to dissolution of the silica network through ion exchange and Si-O bond network breakdown. This causes both surface loss and roughness with defects leading to a higher vulnerability to wear and stress. Although ceramics and porcelains are very stable materials, the well-known breakage of SiO2 bonds explains the sensitivity of glass-containing ceramics and porcelains to water. 1.2.9 Thermal loading

The defects may be transformed into cracks at an additional stressing mode: repeated driven forces induced by thermal cycles in oral cavity, Changes of temperature especially sudden thermal shocks due to eating, drinking and breathing generate thermal transient stresses [28, 91-99]. Thermal shocking is often performed by moving subjects between two water containers at different temperatures. In laboratory experiments thermal shocking results in clear crack formation and/or

Chapter 1

18

fracture, which is assumed to be driven by repeated thermal transient stress. The temperature range is often defined from 0ºC to 60ºC, very hot water for oral tissues. This may not be as severe as in other fields of engineering, yet thermal shocking is a potential weakening factor. Several investigations showed that porcelains intended for dental use are less resistant to thermo-cycling fatigue than core materials, especially polycrystalline ones like alumina and zirconia. 1.2.10 Crack growth and its involvement in lifetime predication

The first part of a fatigue process is characterized by subcritical crack growth (SCG) until a critical size is reached and unstable and fast fracture occurs. When the magnitude of SCG is known, it is possible to estimate the service life of a mechanical structure.

The crack growth of brittle materials is difficult to characterize, especially when they are subjected to cyclic loading because several processes are involved [50]:

A) Frictional sliding of the mating surfaces of microscopic cracks. B) Progressive wear and breakage of the bridging ligaments connecting the faces of

cracks and long flaws. C) Wedging of the crack surfaces by debris such as entrapped particles. D) Inelastic strain from shear or dilatational transformations such as mechanical twins

or martensitic lamellae. Progressive wear of ligaments, and wedging are more important under cyclic loading [50,100-102].

Although cyclic loading leads to complex processes of crack enlargement, simplified models may help to understand these. Sub-critical crack growth can be explained with a typical (KI, ν) curve, shown in Figure 1.1.

Region 1. in which crack motion is assumed to be controlled by stress-enhanced

chemical processes at the crack tip. Region 2. where the crack velocity does not depend strongly on the stress intensity

factor (n = 0, below). Crack motion is controlled by the diffusion rate of corrosive agents at the crack tip.

Region 3. The crack velocity increases rapidly when KI reaches the critical stress intensity factor, KIc. The crack motion is governed by brittle fracture mechanisms, associated with a high exponent of crack propagation.

Introduction

19

Figure 1.1 Typical plot of crack growth versus stress intensity factor. All three regions may be described by the form [61, 103]

nIAK

dtdc

==ν

where v is the crack velocity, c is the crack length, and A is a constant. The propagation exponent n, is a characteristic of the resistance to mechanical fatigue. A and n are recognized as SCG parameters. A greater n indicates a better resistance to stress-driven environment-assisted SCG.

Experimental tests may be divided into direct and indirect techniques. Direct techniques directly measure the stable subcritical crack growth speed through a specimen having a big pre-crack such as Double torsion (DT), Compact Tension (CT), and Double Cantilever Beams (DCB) [72, 104-107] and the help of fracture mechanic analysis. Indirect techniques are based on fracture strength measurements [108-110]. However, resistance to crack growth may be crack-size dependent for example with rising R-curve materials. The resistance to growth of large cracks in direct techniques may not be identical to that in small cracks.

Meanwhile indirect techniques may take advantage of the initial surface condition, which is expected to simulate the natural crack system, including flaws. Fracture strength tests in indirect techniques use a controlled applied stress σf0 (static fatigue test) or a controlled stress rateσ& (dynamic fatigue test). The combination of these two indirect techniques is often employed for the determination of SCG parameters and for the estimation of clinical lifetime of all-ceramic prosthetic devices using

Chapter 1

20

( ) σσσ &20

1 1 −+ += nf

n nB and ( )22

2

2

−=

−

nAYKB

nIc ,

The measurement of the fracture strength at different stressing rates allows the determination of the SCG parameters. After mathematic processing, the time to failure at a constant stress σ becomes

nnff Bt −−= σσ 2

0

The development of Strength–Probability–Time (SPT) diagrams looks

theoretically and practically promising in perspective of structural design, material selection, and as a tool for the improvement and development of materials and their processing.

Integrated with failure probability due to the random distribution of crack size and the consequent dispersion of strength values, SPT diagrams provide a combination of failure probabilities as a function of strength and time [108].

0lnln2

ln21

1lnln σσ mBn

mtn

mP

n

ff

−−

+−

=−

Pf is the failure probability. SPT diagrams provide a combination of failure probabilities as a function of strength and time [111-113].

Static fatigue lifetime may be depicted in another manner, focused on the driving force of the stress intensity using the initial value of the stress intensity factor, KIi.

( ) ( )nIc

nIif KK

nAYt −−

−

−−

= 222

2

22σ

Note the strong effect of KIc on tf. Taking ( )2/ σYKc I= , the equation illustrates the effect of the initial crack size Ci on SCG and the lifetime of mechanical parts. The closer it is to the critical size Cc, the more rapid the crack grows to a fatal stage.

( ) ⎟⎟⎠

⎞⎜⎜⎝

⎛−

−=

−−−2

22

2

22 n

c

n

in

n

f CCnAY

t σ

Unfortunately, the SCG knowledge of cyclic fatigue is far away from the level

of static and dynamic fatigue due to its complex failure mechanism. Crack

Introduction

21

susceptibility to KI or ΔKI depends on size, as small cracks are more sensitive and advance below Kth (threshold of KI for SCG) found for large-crack specimens. Assuming equality of the subcritical crack enhance rate or process under different loading modes, [114] gave a formulation of the number of cycles to failure Nf with substitution of the applied stress σ of the static test with an alternating stress Δσ as the subtraction of the minimum stress σmin from the maximum stress σmax.

( )( ) ⎟⎟

⎠

⎞⎜⎜⎝

⎛−

−Δ

=−−−2

22

2

22 n

c

n

in

n

f CCnAY

N σ

Another modified numeric estimation of cyclic fatigue lifetime and SCG was

developed by Munz and Fret [115],

( ) ( ) nI

nI

nnI KAKRAKA max,

*max,1 =−=Δ=ν

where ΔKI is the amplitude of the applied stress intensity factor (KI,max–KI,min), R is the applied stress ratio KI,min/KI,max, and n and A (and A*) are the subcritical crack growth parameters for cyclic fatigue.

2*0 +=

mm

n f ,

( )( )nf

nf

nIc

nYNK

Aσ

σΔ−

=−−

22

20,

20,0

2

and ( )( )nf

nf

nIc

nYNK

Amax

20,

20,0

2*

22

σσ−

=−−

where m* and Nf,0 are the Weibull modulus and the characteristic number of cycles to failure from the lifetime distribution, mf0 and σf0,0 are the Weibull modulus and characteristic strength of the inert strength σf0, which are based on

0,00 lnln1

1lnln fff

mmP

σσ −=⎟⎟⎠

⎞⎜⎜⎝

⎛

−

and 0** lnln

11lnln ff

f

NmNmP

−=⎟⎟⎠

⎞⎜⎜⎝

⎛

−

The numbers of cycles leading to failure at a σmax of about half of σf0, are adopted for the Weibull analysis. The prediction for the number of cycles then becomes

Chapter 1

22

( )( )nnf

nIc

f nYAK

Nmax

2*

20

2

22

σσ

−=

−−

or ( ) nn

ff BN −−= max2

0* σσ , with B* based on A*, which looks similar to the static fatigue

lifetime formulation nnff Bt −−= σσ 2

0 , with Nf,0 merged into B*.

This method employs the SCG data obtained from cyclic loading rather than static and dynamic fatigue tests, which might be a better representation of cyclic SCG, and has been reported in the evaluation of dental ceramics [116-119].

Stress-assisted and corrosion-driven crack growth can be characterized with the knowledge of SCG as described above. By applying engineering tools as finite element analysis and with the help of those formulations, the SPT relation could be established provided that all the necessary basic and SCG properties are available. Fatigue behaviour may also be calculated with the relevant software like CARES/LIFE (Ceramic Analysis and Reliability Evaluation of Structure Life Prediction) (NASA Lewis Research Center, Cleveland, OH, USA) [109, 120, 121], which also accounts for the phenomenon of subcritical crack growth by utilizing the power law [115]. Nevertheless, in dental ceramics, SCG parameters are mainly assessed with dynamic fatigue test and lifetime predictions based on static fatigue theories. Although the parameters are based on experimental data, these results rely on theoretical calculations and the assumption of similarity of SCG behaviour under distinctly different loading conditions. In practice such fatigue estimations seem inadequate for complex conditions like the oral cavity [49, 50, 104-106, 122], despite that several scientific publications [109, 123] confirmed the consistence of SCG parameters from one method or experimental test with other methods and clinical failure trends. 1.2.11 S-N curve for strength degadation test In other experimental fatigue approaches the applied stress, or the residual strength is plotted to the loading time in-service time or number of cycles to failure, resulting in S-N (Stress-Number of cycles) diagrams, which expose initially “natural” surfaces to controlled stresses [50, 72, 124-126]. There are two definitions of “time” which could be introduced in stress fatigue in prosthodontics: total life and half life or endurance limit. Total life is the average number of cycles until failure, while half-life is defined as the number of cycles where 50% of the specimens have failed. The technical problem is that the total-life S-N curve approach may be rather time-consuming. To mitigate this, specimens are often polished or roughened to control the surface profile [108-111, 113, 116-119], in which relatively large surface flaws are

Introduction

23

preferred or introduced to get a narrower range of degradation strength or time/cyclic numbers.

Half-life can be assessed with the staircase test design, which is a straightforward approach for endurance limit in which the load for each successive specimen is slightly increased or decreased, depending on success or failure of the previous specimen [72].

Normally in the biomaterial industry, the processing of products like artificial heart valves and hip joints is standardized, the microstructure of the used materials is well controlled and proof testing is routine practice [50]. Characteristic of the custom made processing in dentistry is the individual shape of the structures and that proof testing is not feasible because it is hardly possible to make more than one identical restoration. 1.2.13 Complications in lifetime prediction

Service time prediction (STP) is an achievement in materials engineering, which may be applied in prosthetic dentistry too for the prediction of longevity of prosthetic restorations. Obviously fatigue is a process of time-dependent evolution of damage or damage accumulation. Besides functional loading, restorations in the oral cavity have to function in a wet environment with wear, contact and impact damage, thermal residual stresses due to ceramic and porcelain sintering, and repeated thermal transient stresses from temperature changes. It is hard to couple all factors and failure modes in the acquirement of data of strength degradation as well as lifetime prediction and experimental design. At present, this is an impossible mission.

It can be concluded that masticatory complex loading modes, damaging modes, flaw size and distribution, SCG parameters of small cracks and clinical failure modes should be well investigated and known before a useful prediction of the durability of a dental prosthetic restoration is possible. The study presented in this thesis aims to contribute to this knowledge.

Chapter 2 is a study into the influence of test methods on fracture toughness of a dental porcelain and a soda lime glass. While in chapter 3 three test methods to determine fracture toughness are compared. In chapter 4 the indentation strength in bending method is further evaluated with dental porcelains. Chapter 5 deals with the role of the indentation load at smaller flaw sizes in the indentation-strength fracture toughness. In chapter 6 the strength influencing variables on CAD/CAM zirconia frameworks and in chapter 7 the staircase strength degradation of sandblasted zirconia due to cyclic fatigue are investigated further.

Chapter 1

24

1.3 References [1] J.R Kelly, I. Nishimura I and S.D. Campbell, Ceramics in dentistry: historical

roots and current perspectives. J Prosthet Dent 75 (1996), pp. 18-32. [2] J.W. McLean, Evolution of dental ceramics in the twentieth century. J Prosthet

Dent 85 (2001), pp. 61-6. [3] A.S. Freese, Porcelain fused to iridio-platinum crowns. J Prosthet Dent 9

(1959), pp. 847-50. [4] M. Weinstein, S. Katz and A.B. Weinstein, Fused porcelain-to-metal teeth. US

Patent 3,052,982, 1962. [5] M. Weinstein and A.B. Weinstein, Porcelain-covered metal-reinforced teeth.

US Patent 3,052,983, 1962. [6] N.H. Creugers, A.F. Käyser and M.A. van 't Hof, A meta-analysis of durability

data on conventional fixed bridges. Community Dent Oral Epidemiol. 22 (1994), pp. 448-52.

[7] P.J. Leempoel, A.F. Käyser, G.M. Van Rossum and A.F. De Haan, The survival rate of bridges. A study of 1674 bridges in 40 Dutch general practices. J Oral Rehabil. 22 (1995), pp. 327-30.

[8] A. Eliasson, C.F. Arnelund and A. Johansson, A clinical evaluation of cobalt-chromium metal-ceramic fixed partial dentures and crowns: A three- to seven-year retrospective study. J Prosthet Dent. 98 (2007), pp. 6-16.

[9] J.W. McLean, The alumina reinforced porcelain jacket crown. J Am Dent Assoc. 75 (1967), pp. 621-8.

[10] J.W. McLean, A higher strength porcelain for crown and bridge work. Br Dent J. 21 (1965), pp. 268-72.

[11] J.W. McLean and T.H. Hughes, The reinforcement of dental porcelain with ceramic oxides. Br Dent J. 21 (1965), pp. 251-67.

[12] H. Hennies, Vita Hi-Ceram--an advance in development of ceramics. Quintessenz. 39 (1988), pp. 235-46.

[13] W. Riedling and H.F. Kappert, Longevity of Hi-Ceram full ceramic bridges. Quintessenz. 38 (1987), pp. 1493-8.

[14] R.B. Sozio and E.J. Riley, The shrink-free ceramic crown. J Prosthet Dent. 49 (1983), pp. 182-7.

[15] K.A. Malament and D.G. Grossman, The cast glass-ceramic restoration. J Prosthet Dent. 57 (1987), pp. 674-83.

[16] G. Sjögren, R. Lantto and A. Tillberg, Clinical evaluation of all-ceramic crowns

Introduction

25

(Dicor) in general practice. J Prosthet Dent. 81 (1999), pp. 277-84. [17] K.A. Malament and S.S. Socransky, Survival of Dicor glass-ceramic dental

restorations over 14 years: Part I. Survival of Dicor complete coverage restorations and effect of internal surface acid etching, tooth position, gender, and age. J Prosthet Dent. 81 (1999), pp. 23-32.

[18] K.A. Malament and S..S. Socransky, Survival of Dicor glass-ceramic dental restorations over 16 years. Part III: effect of luting agent and tooth or tooth-substitute core structure. J Prosthet Dent. 86 (2001), pp. 511-9.

[19] K.A. Malament and S.S. Socransky, Survival of Dicor glass-ceramic dental restorations over 14 years. Part II: effect of thickness of Dicor material and design of tooth preparation. J Prosthet Dent. 81 (1999), pp. 662-7.

[20] W.P. Kelsey 3rd, T. Cavel, R.J. Blankenau and W.W. Barkmeier, T.M. Wilwerding and M.A. Latta, 4-year clinical study of castable ceramic crowns. Am J Dent. 8 (1995), pp. 259-62.

[21] J.K. Dong, H. Luthy, A. Wohlwend and P. Schärer, Heat-pressed ceramics: technology and strength. Int J Prosthodont. 5 (1992), pp. 9-16.

[22] O. El-Mowafy and J.F. Brochu, Longevity and clinical performance of IPS-Empress ceramic restorations - a literature review. J Can Dent Assoc. 68 (2002), pp. 233-7.

[23] M. Fradeani and M. Redemagni, An 11-year clinical evaluation of leucite-reinforced glass-ceramic crowns: a retrospective study. Quintessence Int. 33 (2002), pp. 503-10.

[24] G.J. Christensen, Computerized restorative dentistry. State of the art. J Am Dent Assoc. 132 (2001), pp. 1301-3.

[25] W. Höland, M. Schweiger, M. Frank and V. Rheinberger, A comparison of the microstructure and properties of the IPS Empress 2 and the IPS Empress glass-ceramics. J Biomed Mater Res. 53 (2000), pp. 297-303.

[26] P. Marquardt and J.R Strub, Survival rates of IPS empress 2 all-ceramic crowns and fixed partial dentures: results of a 5-year prospective clinical study. Quintessence Int. 37 (2006), pp. 253-9.

[27] A. Wassermann, M. Kaiser and J.R. Strub, Clinical long-term results of VITA In-Ceram Classic crowns and fixed partial dentures: A systematic literature review. Int J Prosthodont. 2006 Jul-Aug;19(4):355-63.

[28] H.F. Kappert, H. Knode and R. Schultheiss, Strength of the In-Ceram system under mechanical loads and thermocycling in artificial saliva. Dtsch Zahnarztl Z. 46 (1991), pp. 129-31.

Chapter 1

26

[29] M.J. Heffernan, S.A. Aquilino, A.M. Diaz-Arnold, D.R. Haselton, C.M. Stanford and M.A. Vargas, Relative translucency of six all-ceramic systems. Part II: core and veneer materials. J Prosthet Dent. 88 (2002), pp. 10-5.

[30] M.J. Heffernan, S.A. Aquilino, A.M. Diaz-Arnold, D.R. Haselton, C.M. Stanford and M.A. Vargas, Relative translucency of six all-ceramic systems. Part I: core materials. J Prosthet Dent. 88 (2002), pp. 4-9.

[31] P. Magne and U. Belser, Esthetic improvements and in vitro testing of In-Ceram Alumina and Spinell ceramic. Int J Prosthodont. 10 (1997), pp. 459-66.

[32] S.D. Campbell, L.B. Pelletier, R.L. Pober and R.A. Giordano, Dimensional and formation analysis of a restorative ceramic and how it works. J Prosthet Dent. 74 (1995), pp. 332-40.

[33] J. Fischer, M. Schmid, H.F. Kappert and J.R. Strub, Texture formation of the dental ceramic core material In-Ceram and thermal expansion of its individual components. Dtsch Zahnarztl Z. 46 (1991), pp. 461-3.

[34] M.J. Suárez, J.F. Lozano, M. Paz Salido and F. Martínez, Three-year clinical evaluation of In-Ceram Zirconia posterior FPDs. Int J Prosthodont. 17 (2004), pp. 35-8.

[35] M. Andersson and A. Odén, A new all-ceramic crown. A dense-sintered, high-purity alumina coping with porcelain. Acta Odontol Scand. 51 (1993), pp. 59-64.

[36] N.U. Zitzmann, M.L. Galindo, E. Hagmann and C.P. Marinello, Clinical evaluation of Procera AllCeram crowns in the anterior and posterior regions. Int J Prosthodont. 20 (2007), pp. 239-41.

[37] M.L. Galindo, E. Hagmann, C.P. Marinello and N.U. Zitzmann, Long-term clinical results with Procera AllCeram full-ceramic crowns. Schweiz Monatsschr Zahnmed. 116 (2006), pp. 804-9.

[38] Y. Kokubo, C. Ohkubo, M. Tsumita, A. Miyashita, P. Vult von Steyern and S. Fukushima, Clinical marginal and internal gaps of Procera AllCeram crowns. J Oral Rehabil. 32 (2005), pp. 526-30.

[39] F.H. Rasetto, C.F. Driscoll, V. Prestipino, R. Masri and J.A. von Fraunhofer, Light transmission through all-ceramic dental materials: a pilot study. J Prosthet Dent. 91 (2004), pp. 441-6.

[40] P. Odman and B. Andersson, Procera AllCeram crowns followed for 5 to 10.5 years: a prospective clinical study. Int J Prosthodont. 14 (2001), pp. 504-9.

[41] A. Odén, M. Andersson, I. Krystek-Ondracek, D. Magnusson, Five-year clinical evaluation of Procera AllCeram crowns. J Prosthet Dent. 80 (1998), pp.

Introduction

27

450-6. [42] M. Andersson, M.E. Razzoog, A. Odén, E.A. Hegenbarth and B.R. Lang,

Procera: a new way to achieve an all-ceramic crown. Quintessence Int. 29 (1998), pp. 285-96.

[43] J.R. Kelly and I. Denry, Stabilized zirconia as a structural ceramic: An overview. Dent Mater (2007) [Epub ahead of print]

[44] I. Sailer, A. Fehér, F. Filser, L.J. Gauckler, H. Lüthy and C.H. Hämmerle, Five-year clinical results of zirconia frameworks for posterior fixed partial dentures. Int J Prosthodont. 20 (2007), pp. 383-8.

[45] H. Yilmaz, C. Aydin and B.E. Gul, Flexural strength and fracture toughness of dental core ceramics. J Prosthet Dent. 98 (2007), pp. 120-8.

[46] I. Denry and J.R. Kelly, State of the art of zirconia for dental applications. Dent Mater. (2007) [Epub ahead of print]

[47] C. Piconi and G. Maccauro, Zirconia as a ceramic biomaterial. Biomaterials. 20 (1999), pp. 1-25.

[48] R.H.J. Jannink, P.M. Kelly and B.C. Muddle, Transformation toughening in zirconia-containing ceramics. J Am Ceram Soc. 83 (2000), pp. 461-87.

[49] T.L. Anderson, Fracture Mechanics: Fundamentals and Applications, Vol. 1, Edited by T.L. Anderson. CRC Press, Boca Raton, 1991.

[50] S. Suresh, Fatigue of Materials, 2nd Edition, Edited by S. Suresh. Cambridge University Press, Cambridge, 1998.

[51] S. Ban and K.J. Anusavice, Influence of test method on failure stress of brittle dental materials. J Dent Res. 69 (1990), pp. 1791-9.

[52] W.C. Wagner and T.M. Chu, Biaxial flexural strength and indentation fracture toughness of three new dental core ceramics. J Prosthet Dent. 76 (1996), pp. 140-4.

[53] K. Zeng, A. Odén and D. Rowcliffe, Flexure tests on dental ceramics. Int J Prosthodont. 9 (1996), pp. 434-9.

[54] A.J. Raigrodski, Contemporary materials and technologies for all-ceramic fixed partial dentures: a review of the literature. J Prosthet Dent 92 (2004), pp. 557–62.

[55] H.J. Conrad, W.J. Seong and I.J. Pesun, Current ceramic materials and systems with clinical recommendations: a systematic review. J Prosthet Dent. 98 (2007), pp. 389-404.

[56] J.R. Kelly, Dental ceramics: current thinking and trends. Dent Clin North Am. 48 (2004), pp. 513-30.

Chapter 1

28

[57] J.R. Kelly, Ceramics in restorative and prosthetic dentistry. Annu Rev Mater Sci. 27 (1997), pp. 443-68.

[58] J. Tinschert, D. Zwez, R. Marx and K.J. Anusavice, Structural reliability of alumina-, feldspar-, leucite-, mica- and zirconia-based ceramics. J Dent. 28 (2000), pp. 529-35.

[59] W. Weibull, A statistical distribution function of wide application. Journal of Applied Mechanics. 9 (1951), pp. 293-7.

[60] J.R. Kelly, Perspectives on strength. Dent Mater. 11 (1995), pp. 103-10. [61] J.E. Ritter, Predicting lifetimes of materials and material structures. Dent

Mater. 11 (1995), pp. 142-146. [62] J.J. Mecholsky, Fracture mechanics principles. Dent Mater. 11 (1995), pp. 111–

112. [63] G.D. Quinn, R.J. Gettings and K. Xu, Standard reference material 2100:

ceramic fracture toughness, Ceram Eng Sci Proc. 20 (1999), pp. 513–523. [64] G.D. Quinn, The fracture toughness round robins in VAMAS: what we have

learned. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), pp. 1409.

[65] G.D. Quinn, K. Xu, R.G. Gettings, J.A. Salem and J.J. Swab, Does anyone know the real fracture toughness? SRM 2100: the world's first ceramic fracture toughness reference material. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), pp. 1409.

[66] M.G. Jenkins, J.A. Salem, G.D. Quinn and I. Bar-On, Development, verification, and implementation of a national consensus fracture toughness test method standard for advanced ceramics. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), pp. 1409.

[67] American Society for Testing Materials (ASTM 399-74), Standard test method for plan-strain fracture toughness of metallic materials. E399-74, ASTM, Philadelphia, PA (1983) pp. 923–36.

[68] H. Fischer and R. Marx, Fracture toughness of dental ceramics: comparison of bending and indentation method. Dent Mater. 18 (2002), pp. 12–19.

Introduction

29

[69] J.B Quinn, V. Sundar and I.K. Lloyd, Influence of microstructure and chemistry on the fracture toughness of dental ceramics. Dent Mater. 19 (2003), pp. 603–611.

[70] M. Guazzato, M. Albakry, M.V. Swain and J. Ironside, Mechanical properties of In-ceram alumina and In-ceram zirconia. Int J Prosthodont. 15 (2002), pp. 339–346.

[71] J.R. Kelly, Clinically relevant approach to failure testing of all-ceramic restorations. J Prosthet Dent. 81 (1999), pp. 652-61.

[72] H.W. Wiskott, J.I. Nicholls and U.C. Belser, Stress fatigue: basic principles and prosthodontic implications. Int J Prosthodont. 8 (1995), pp. 105-16.

[73] J.Y. Thompson, B.R. Stoner and J.R. Piascik, Ceramics for restorative dentistry: critical aspects for fracture and fatigue resistance. Mater Sci Eng C. 27 (2007), pp. 565-9.

[74] D.R. Rekow and V.P. Thompson, Engineering long term clinical success of advanced ceramic prostheses. J Mater Sci: Mater Med. 18 (2007), pp. 47-56.

[75] J.W. McLean, The failed restoration: causes of failure and how to prevent them. Int Dent J. 40 (1990), pp. 354-8.

[76] K.J. Anusavice, Degradability of dental ceramics, Adv Dent Res 6 (1992), pp. 82-9.

[77] S.S. Scherrer, J.B. Quinn, G.D. Quinn and J.R. Kelly, Failure analysis of ceramic clinical cases using qualitative fractography. Int J Prosthodont. 19 (2006), pp. 185-92.

[78] J.B. Quinn, G.D. Quinn, J.R. Kelly and S.S. Scherrer, Fractographic analyses of three ceramic whole crown restoration failures. Dent Mater. 21 (2005), pp. 920-9.

[79] R.G. Luthardt, M.S. Holzhüter, H. Rudolph, V. Herold and M.H. Walter, CAD/CAM-machining effects on Y-TZP zirconia. Dent Mater. 20 (2004), pp. 655-62.

[80] H. Wang, M.N. Aboushelib, and A.J. Feilzer, Strength influencing variables on CAD/CAM zirconia frameworks. Dental Mater. (2007), [Epub ahead of print].

[81] R.G. Luthardt, M. Holzhuter, O. Sandkuhl, V. Herold, J.D. Schnapp, E. Kuhlisch and M. Walter, Reliability and properties of ground Y-TZP-zirconia ceramics. J Dent Res. 81 (2002), pp. 487-91.

[82] D. Rekow and V.P. Thompson, Near-surface damage--a persistent problem in crowns obtained by computer-aided design and manufacturing. Inst Mech Eng [H]. 219 (2005), pp. 233-43.

Chapter 1

30

[83] M. Braem, P. Lambrechts and G. Vanherle, Clinical relevance of laboratory fatigue studies. J Dent. 22 (1994), pp. 97-102.

[84] P. Lambrechts, K. Goovaerts, D. Bharadwaj, J.D. Munck, L. Bergmans, M. Peumans and B. van Meerbeek, Degradation of tooth structure and restorative materials: A review. Wear 261 (2006), pp. 980-986.

[85] W.S. Oh, R. Delong and K.J. Anusavice, Factors affecting enamel and ceramic wear: a literature review. J Prosthet Dent. 87 (2002), pp. 451-9.

[86] B.R. Lawn, Y. Deng and V.P. Thompson, Use of contact testing in the characterization and design of all-ceramic crownlike layer structures: a review. J Prosthet Dent. 86 (2001), pp. 495-510.

[87] J.W. Kim, J.H. Kim, V.P. Thompson and Y. Zhang, Sliding contact fatigue damage in layered ceramic structures. J Dent Res. 86 (2007), pp. 1046-50.

[88] Y.A.H, Mashal, The role of fracture mechanics parameters in glass/ceramic wear. Eng Fract Mech. 52 (1995), pp. 43-47.

[89] P. Pallav, Occlusal wear in dentistry, fundamental mehanisms, clinical implications, and laboratory assessment. Ph.D. Thesis, Academisch Proefschrift (ACTA) (1996).

[90] L.H. Mair, T.A. Stolarski, R.W. Vowles and C.H. Lloyd, Wear: mechanisms, manifestations and measurement. Report of a workshop. J Dent. 24 (1996), pp. 141-8.

[91] P.J. Steiner, J.R. Kelly and A.A. Giuseppetti, Compatibility of ceramic-ceramic systems for fixed prosthodontics. Int J Prosthodont. 10 (1997), pp. 375-80.

[92] M. Rosentritt, M. Behr, R. Gebhard and G. Handel, Influence of stress simulation parameters on the fracture strength of all-ceramic fixed-partial dentures. Dent Mater. 22 (2006), pp. 176-82.

[93] P. Vult von Steyern, S. Ebbesson, J. Holmgren, P. Haag and K. Nilner, Fracture strength of two oxide ceramic crown systems after cyclic pre-loading and thermocycling. J Oral Rehabil. 33 (2006), pp. 682-9.

[94] O. Addison, G.J. Fleming and P.M. Marquis, The effect of thermocycling on the strength of porcelain laminate veneer (PLV) materials. Dent Mater. 19 (2003), pp. 291-7.

[95] J.R. Strub and S.M. Beschnidt, Fracture strength of 5 different all-ceramic crown systems. Int J Prosthodont. 11 (1998), pp. 602-9.

[96] R.K. Kellerhoff and J. Fischer, In vitro fracture strength and thermal shock resistance of metal-ceramic crowns with cast and machined AuTi frameworks. J Prosthet Dent. 97 (2007), pp. 209-15.

Introduction

31

[97] K.J. Anusavice, P.H. Dehoff, S.W. Twiggs and P.C. Lockwood, Thermal shock resistance of porcelain discs. J Dent Res. 62 (1983), pp. 1082-5.

[98] K.J. Anusavice, S.W. Twiggs, P.H. DeHoff and C.W. Fairhurst, Correlation of thermal shock resistance with thermal compatibility dat for porcelain-metal systems. J Dent Res. 61 (1982), pp. 419-22.

[99] M.G. Tróia MG Jr, G.E. Henriques, M.A. Nóbilo and M.F. Mesquita, The effect of thermal cycling on the bond strength of low-fusing porcelain to commercially pure titanium and titanium-aluminium-vanadium alloy. Dent Mater. 19 (2003), pp. 790-6.

[100] C.J. Gilbert, J.J. Cao, W.J. Moberlychan, L.C. Dejonghe and R.O. Ritchie, Cyclic fatigue and resistance-curve behavior of an in situ toughened silicon carbide with Al-B-C additions. Acta Mater. 44 (1996), pp. 3199-214.

[101] H. El Attaoui, M. Saadaoui, J. Chevalier and G. Fantozzi, Static and cyclic crack propagation in Ce-TZP ceramics with different amounts of transformation toughening. J Eur Ceram Soc. 27 (2007), pp. 483-86.

[102] F. Guiu, M.J. Reece and D.A.J. Vaughan, Cyclic fatigue of ceramics. J Mater Sci. 26 (1991), pp. 3275-86.

[103] J.E. Ritter, Critique of test methods for lifetime predictions. Dent Mater. 11 (1995), pp. 147-51.

[104] F. Guiu, M.J. Reece and D.A.J. Vaughan, Cyclic fatigue of ceramics. J Mater Sci. 26 (1991), pp. 3275-86.

[105] J. Chevalier, C. Olagnon and G. Fantozzi, Subcritical Crack Propagation in 3Y-TZP Ceramics: Static and Cyclic Fatigue. J. Am. Ceram. Soc. 82 (1999), pp. 3129–38.

[106] C.J. Gilbert, J.J. Cao, W.J. Moberlychan, L.C. Dejonghe and R.O. Ritchie, Cyclic fatigue and resistance-curve behavior of an in situ toughened silicon carbide with Al-B-C additions. Acta Mater. 44 (1996), pp. 3199-214.

[107] H. El Attaoui, M. Saadaoui, J. Chevalier and G. Fantozzi, Static and cyclic crack propagation in Ce-TZP ceramics with different amounts of transformation toughening. J Eur Ceram Soc. 27 (2007), pp. 483-86.

[108] J. Tinschert, G. Natt, N. Mohrbotter, H. Spiekermann and K.A. Schulze, Lifetime of alumina- and zirconia ceramics used for crown and bridge restorations. J Biomed Mater Res B Appl Biomater. 80 (2007), pp. 317-21.

[109] H. Fischer, M. Weber and R. Marx, Lifetime prediction of all-ceramic bridges by computational methods. J Dent Res. 82 (2003), pp. 238-42.

[110] E.C. Teixeira, J.R. Piascik, B.R. Stoner and J.Y. Thompson, Dynamic fatigue

Chapter 1

32

and strength characterization of three ceramic materials. J Mater Sci Mater Med. 18 (2007), pp. 1219-24.

[111] S. Raynaud, E. Champion, D. Bernache-Assolant and D. Tetard, Dynamic fatigue and degradation in solution of hydroxyapatite ceramics. J Mater Sci: Mater in Med. 9 (1998), pp. 221-227.

[112] R.G. Chadwick RG. Strength-probability-time (SPT) diagram - an adjunct to the assessment of dental materials? J Dent. 22 (1994), pp. 364-9.

[113] U. Lohbauer, A. Petschelt and P. Greil, Lifetime prediction of CAD/CAM dental ceramics. J Biomed Mater Res (Appl Biomater) 63 (2002), pp. 780–5.

[114] R.O. Ritchie and R.H. Dauskardt. Cyclic fatigue ceramics: a fracture mechanics approach to subcritical crack growth and life prediction. J Ceram soc Jpn. 99 (1991), pp. 1047-62.

[115] D. Munz and T. Fett, Ceramics: mechanical properties, failure behaviour, materials selection. Berlin: Springer-Verlag; 1999.

[116] A.R. Studart, F. Filser, P. Kocher and L.J. Gauckler, In vitro lifetime of dental ceramics under cyclic loading in water. Biomaterials. 28 (2007), pp. 2695-705.

[117] A.R. Studart, F. Filser, P. Kocher, H. Lüthy and L.J. Gauckler, Cyclic fatigue in water of veneer-framework composites for all-ceramic dental bridges. Dent Mater. 23 (2007), pp. 177-85.

[118] A.R. Studart, F. Filser, P. Kocher, H. Lüthy and L.J. Gauckler, Mechanical and fracture behavior of veneer-framework composites for all-ceramic dental bridges. Dent Mater. 23 2007 Jan;23(1):115-23.

[119] A.R. Studart, F. Filser, P. Kocher and L.J. Gauckler, Fatigue of zirconia under cyclic loading in water and its implications for the design of dental bridges. Dent Mater. 23 (2007), pp. 106-14.

[120] N.N. Nemeth, J.M. Manderscheid and J.P. Gyekenyesi, Ceramics analysis and reliability evaluations of structures (CARES). NASA Technical Paper 2916; Cleveland, OH: Glenn Research Center, 1989.

[121] N.N Nemeth, L.M. Powers, L.A. Janosik and J.P. Gyekenyesi, CARES/LIFE. Users and programmers manual. Cleveland, OH: NASA, Glenn Research Center, 1993.

[122] F. Sudreau, C. Olagnon C and G. Fantozzi, Lifetime Prediction of Ceramics: Importance of the Test Method. Ceram Inter. 20 (1994), pp. 125-35.

[123] U. Lohbauer, N. Krämer, A. Petschelt, R. Frankenberger, Correlation of in vitro fatigue data and in vivo clinical performance of a glassceramic material. Dent Mater. 24 (2008), pp. 39-44.

Introduction

33

[124] A.G. Evans, Overview No. 125 Design and life prediction issues for high-temperature engineering ceramics and their composites. Acta Mater. 45 (1997), pp. 23-40.

[125] H.Y. Chen, R. Hickel, J.C. Setcos, K.H. Kunzelmann, Effects of surface finish and fatigue testing on the fracture strength of CAD-CAM and pressed-ceramic crowns. J Prosthet Dent. 82 (1999), pp. 468-75.

[126] T. Ohyama, M. Yoshinari, Y. Oda, Effects of cyclic loading on the strength of all-ceramic materials. Int J Prosthodont. 12 (1999), pp. 28-37.

Chapter 1

34

35

CHAPTER 2

Influence of Test Methods on Fracture Toughness of a Dental Porcelain and a Soda Lime Glass

Keywords: Dental porcelains, Fracture toughness, Chevron notched, Indentation strength, Single-edge-notched beam

Chapter 2

36

2.1 Abstract Objective: The aim of this study was to investigate the influence of the test method on fracture toughness of a dental porcelain and a soda lime glass. Materials and methods: Three methods were used to determine fracture toughness: the indentation strength (IS) by bending, chevron-notched beam (CNB), and the single-edge-notched beam (SENB). In the IS method, the ratio of elastic modulus to hardness (E/H) in the formula was determined by two methods: individual measurement for E and H (ISM) as well as direct estimation from Knoop's indentation method (ISK). The tested materials were a dentin porcelain, a traditional feldspar-based leucite-reinforced glass ceramic (Carrara Vincent), and a soda lime glass. Result: Carrara Vincent showed a higher toughness (P<0.01) than glass with all three test methods. The toughness values manifested significant differences between the methods used (P<0.01). The two-way analysis of variance suggested that the materials tested and the test methods used had interaction effects, which statistically means that differences in materials and methods influenced the comparability of the toughness result. In this study, a first step was made to compare different toughness test methods by testing the toughness of a traditional feldspar-based leucite-reinforced glass ceramic and a soda lime glass that has a homogeneous microstructure. Conclusion: An interaction effect of the method and the material used was shown. As a consequence, none of the methods tested is suitable as a universal fracture toughness test method. Further research is needed to investigate more extensively the influence of material composition on the fracture toughness test methods' comparability.

Influence of Test Methods on Fracture Toughness

37

2.2 Introduction Strength is commonly seen as an important parameter for understanding the

clinical performance of dental ceramic restorations as it reflects an important mechanical property. Since the tensile strength of ceramics is much lower than the compressive strength, ceramic restorations often fail in areas of tensile stress. The traditional method for restoration design and material selection is based on attempts to reduce the tensile stresses generated in the structure under load and to select materials with greater strengths than the expected applied stresses [1-4].

Unfortunately, with extremely brittle materials such as ceramics, high strength does not imply a satisfactory fracture resistance [2-4]. Fracture is caused by a propagating crack, which often originates from flaws and extends when the applied stress exceeds a certain threshold. In very brittle materials, this threshold largely depends on the crack tip radius, flaw size, flaw distribution, and fracture toughness.

Fracture toughness is one of the most important material properties in fracture mechanics for brittle materials and is assumed to be independent of flaw size, specimen shape, and the stress concentration acting on the surface. Fracture toughness (KIc) of a brittle material is characterized by a critical level of the stress intensity factor near the crack tip at which a crack will start to propagate. For ceramics that have a primary disadvantage of brittleness and contain many flaws, fracture toughness is, therefore, more elucidating than strength. Obviously, the availability of an accurate method for fracture toughness determination is very important. In the dental literature, greater attention is being drawn to fracture toughness measurements, resulting in an increasing number of publications [5-12].

Various test techniques have been developed for the determination of fracture toughness: the single-edge-notched beam (SENB, by three- or four-point bending) [3, 4, 12] and its derivatives single-edge-precrack-notched beam (SEPB/SEPNB) [4, 7, 13] and single-edge-V-notched beam (SEVNB), [5, 6] chevron-notched (CN) specimen (short rod/bar, three/four-point bending long beam/rod), [8,13-15] double-torsion, double-cantilever beam (DCB), indentation strength (IS) by bending with Vickers or Knoop indentations, [12, 16] indentation fracture (IF) method, [6, 12, 17] and surface-crack-by-flexure (SCF, also called controlled surface microcrack).[11] In dental ceramics research, IF, SENB, and IS methods have frequently been used for the ranking of fracture toughness. The CN method, which rarely appears in the dental literature for ceramic evaluation, was developed in the late 1970s. It is different from SENB in that the designed crack for extension has the shape of a triangle instead of a rectangle.

Chapter 2

38

In spite of the fact that fracture toughness is a material property and therefore the test method to determine this property should not affect the value, it might be expected that just like strength variation attributed to the difference of test techniques, [18] fracture toughness values might be sensitive to and might be affected by the testing and processing methods used.

The aim of this study was to determine the influence of test methods on fracture toughness data of dental ceramics. In this study, fracture toughness values of a dental porcelain and a soda lime glass were determined using three different methods. 3.3 Materials and Methods Specimen Preparation

Three fracture toughness tests were selected for the study: the IS method, the CN beam (CNB) method, and the SENB method. For each method, at least 10 beam-shaped specimens were prepared (Table 2.1). Carrara Vincent porcelain (Elephant Dental B.V., Hoorn, the Netherlands) is a dental dentin porcelain for metal and all-ceramic applications; a soda lime glass was used as reference. Green bodies of porcelain powder were formed in a metal fixture and were then fired in a dental porcelain oven (STRATOS, Elephant Dental B.V.) according to the manufacturer's instructions, except that the cooling time was extended to 1 h to prevent breakage of the baked bodies caused by fast cooling, because their size and thickness were much greater than that of dental ceramic restorations. After firing, the porcelain bodies were ground to a thickness of 3.00±0.05 mm in a grinding device (VEM Metallurgy, Vos & Van Eijk Metallurgie B.V., Houten, the Netherlands) with 30 μm diamond pastes. Subsequently, the bodies were cut into bars with a 0.5 mm diamond saw (ISOMET 1000, Buehler Ltd, Lake Bluff, IL).

All IS beams followed a series of wet grinding and polishing procedures (ECOMET Grinder/Polisher, Buehler Ltd, Evanston, IL) with silicon carbide paper (400, 600, and 1200 grit), during which the sharp edges of rectangular specimens were blunted.

Annealing treatments of specimens were carried out to remove residual stresses. Annealing temperatures were held at 450°C for 1.5 h. Then, specimens were cooled to 100°C in the furnace for another 1.5 h. The annealing temperature was close to the glass transition temperature of the Carrara Vincent porcelain.

Influence of Test Methods on Fracture Toughness

39

Soda–lime–silica glass beams were cut from glass plates and were prepared according to the same processing method to obtain comparable sizes and surface roughness. Table 2.1 The test configuration of three measurement techniques

Test method CNB SENB IS Precrack technique Notch Notch Micro-crack Precrack type Large crack Large crack Small crack Precrack processing method

sawing sawing Vickers indentation

Fracture plane Chevron triangle Rectangle through half

section Rectangle through

whole section Load mode 3-point bending 3-point bending 3-point bending Load speed 0.05 mm/min 0.05 mm/min 0.05 mm/min Specimen dimension proportion(d:b:S)

2 : 3 : 12 1 : 2 :8 2 : 3 : 22

for porcelain Dimension (d:b:S) for porcelain

3 × 4.5 × 26 mm

3 × 6 × 26 mm

2 × 3 × 25 mm

Dimension (d:b:S) for glass

3.8 ×5.7 × 30 mm 3.8 × 7.6 × 50 mm 3.8 ×4.5 × 50mm

with span of 40mm IS Based on the Bending Method

Vickers hardness indentations were made (HM-124 Hardness Testing Machine, Mitutoyo Corp., Kanagawa, Japan) on the tensile surface in the middle of the beams at a load of 9.8 N for glass and 19.6 N for dental porcelain. Variation in the indentation load was introduced to make the ratio of crack length to indentation diagonal equal to about 2.0. Within 0.5 h following the indentation, the beam was loaded in a tensilometer (Instron Universal Mechanical test machine, Instron Corp., Hy Wycombe, U.K.) on the side opposite the indentation in three-point bending until fracture occurred. Specimens in which the fracture did not originate from the indentation notch were excluded; therefore, testing was continued until at least 10 acceptable specimens were acquired.

The fracture strength of indented specimens (IS) was calculated according to the following formula:

Chapter 2

40

223bdWL

f =σ (1)

where σf is the bending strength, W is the fracture load, L is the span length, b is the specimen width, and d is the specimen thickness. The fracture toughness values were then calculated using the following formula [16]:

( ) ( ) 4/33/18/1/ PHEK fIc ση= (2)

where η is a geometrical constant (0.59), E is the elastic modulus, H is the Vickers hardness, σf is the IS, and P is the indentation load.

For each material, the elastic modulus was determined with a three-point bending test on beams without indentation (n=10). The bending deflection (q) of the loaded specimens was recorded after failure. The modulus was calculated using the equation:

qbdWLE 3

3

4= (3)

The Vickers hardness was measured on broken specimens (n=10) using a load of 1.96 N for 15 s. The hardness was calculated using H=1.854P/(2l)2, where P is the indentation load (1.96 N), and l is half of the average length of two diagonals of the indentation measured with a precision of 0.1 μm. The ratio of elastic modulus to hardness (E/H) was determined in two different ways: first, E and H were measured individually as described above and their ratios (E/H)M were calculated; second, the ratio was estimated from the length/width ratio of a Knoop indentation, with y'/x'=y/x−α/(E/H)K, where x and y are defined by the Knoop indenter geometry, y/x=7.11. x' and y' are the measured length of long and short diagonals of Knoop indentation; α is the fitting gradient and amounts to 0.45 [19].

The (E/H)M and (E/H)K ratios were used in the formula for IS fracture toughness and led to two toughness results: ISM and ISK, respectively. CNB Method

Three-point bending was performed on the CN specimens according to previous research on the CNB test technique for plane-strain fracture toughness of brittle

Influence of Test Methods on Fracture Toughness

41

materials [20] (see Figure 2.1). A chevron notch was made with a 0.1 mm diamond sawing disk (ISOMET 1000, Buehler Ltd), which yielded to a chevron angle θ of 60°±1.5° and a pre-crack ratio (r0/b) of 0.1–0.35. As with the traditional three-point bending test, the load was recorded during the experimental procedure, and the maximum load W was obtained from this record. The Chevron angle θ and Chevron precrack length r0 were measured on the two fractured sections of each specimen by optical microscopy (× 10, measuring precision 1 μm). The fracture toughness was calculated according to the following equation:

)/()2/(tan)1(2 02121223 daf

dWK Ic ⋅

−=

θν (4)

2

000 )/(53.179)/(708.20959.17)/( dadadaf ++=

where the Poisson's ratio ν was assumed to be 0.25, as recommended by ISO 6872 for the biaxial flexural strength calculation, and f(r0/b) is the stress intensity shape factor. The actual values of r0 and b of each specimen were used.

Figure 2.1 Chevron-notched beam test. SENB Method

For the SENB test, the notch of a specimen was machined with the 0.1 mm thick diamond saw disk. The saw depth, a, was nearly half that of the specimen's width, b (see Figure 2.2). The specimens were fractured in a three-point bending fixture. The two halves of the broken specimens were used for notch depth measurement under an optical microscope (magnification × 10, reading precision 1 μm). The depth, a, was the average of six values at three locations of the notch: middle point and two sides of each section. The fracture toughness value was calculated according to the following formula [21]:

Chapter 2

42

)/(23 dafbdWLK Ic ⋅= (5)

2927252321 )/(7.38)/(6.37)/(8.21)/(6.4)/(9.2)/( dadadadadadaf +−+−=

where σf is the fracture stress, W is the critical load, d is the specimen thickness, b is the specimen width, L is the span length between supports, a is the notch depth, and f(a/b) is the stress intensity shape factor. The actual values of a and b of each specimen were used. For brittle materials, Fc is equivalent to the maximum load.

d

a

d

b

L

W

Figure 2.2 Single-edged-notched-beam test. Statistical Analysis

Two-way analysis of variance was utilized for determining the significance of the material effect, the test method effect, and the interaction between these effects. The difference within each material group was analyzed by pairwise comparisons of tests for a simple method factor within the other factor–material group. The test methods were assigned to three groups comprising combination 1 of ISM, CNB, and SENB, combination 2 of ISK, CNB, and SENB, and a paired combination of ISM and ISK. All statistical analyses were assessed at an α level of 0.05. 2.4 Results

The fracture toughness results for the tested materials are listed in Table 2.2. The statistical analysis results are presented in Tables 2.3, 2.4, and 2.5. In all cases, data scatter was low, as the standard deviations ranged from 2.5% to 7.4% of the mean values.

The main material effect and the main method effect as well as the material by method interaction were significant (P ≤ 0.05) for all method combinations. The traditional leucite-reinforced feldspar-based glass-ceramic Carrara Vincent was slightly tougher than glass, which was demonstrated with each of the three test methods. At the same time, for each material, the toughness results differed slightly by

Influence of Test Methods on Fracture Toughness

43

the test method. Moreover, the differences between the materials were also influenced by the methods used.

Table 2.2 Fracture toughness (in MPa m-1/2) of five materials with four test methods

KIc (average, standard deviation, and scatter in percent) Materials

ISM ISK CNB SENB

Glass 0.77±0.02(2.5%) 0.77±0.02(2.5%) 0.70±0.03(4.0%) 0.77±0.05(6.0%)

Carrara Vincent 0.82±0.05(6.5%) 0.83±0.05(6.5%) 0.84±0.06(7.4%) 0.88±0.03(3.6%)

KIC=Fracture toughness of mode I crack, ISK=Indentation strength method using (E/H)K, ISM = Indentation strength method using (E/H)M, CNB=Chevron notched beam method, SENB=Single edge notched beam method. Table 2.3 Tests of between-Subjects Effects

Subjects combination Effect source F P

material 9.827 0.005method 2508.939 <0.001

ISM and ISK

material-method interaction 876.556 <0.001material 91.868 <0.001method 8.069 0.001

ISM, SENB and CNB

material-method interaction 7.326 0.001material 95.649 <0.001method 7.917 0.001

ISK, SENB and CNB

material-method interaction 6.139 0.004

Although mean toughness values for ISM and for ISK were very close together for both materials, the differences between means was statistically significant. The difference for glass (0.002) was somewhat smaller when compared with that for Carrara Vincent (0.009). Meanwhile, for both materials, SENB indicated a greater fracture toughness than CNB. However, statistically, for glass, both ISM and ISK gave toughness values comparable to that for SENB, and higher than that for CNB, whereas for Carrara Vincent, the mean values were comparable to that for CNB but lower than that for SENB.

Chapter 2

44

Table 2.4 Tests for simple method within material effects Subjects combination Material F P

glass 219.304 <0.001ISM and ISK Carrara Vincent 3043.405 <0.001glass 8.959 <0.001ISM, SENB and CNB Carrara Vincent 6.394 0.003glass 9.289 <0.001ISK, SENB and CNB Carrara Vincent 4.779 0.012

Table 2.5 Pairwise comparisons with regard to toughness as dependent variable Subjects combination Material Method comparisons Difference P

ISM vs SENB 0.002 0.905 CNB 0.069 <0.001

glass

SENB vs CNB 0.067 0.001ISM vs SENB -0.066 0.001 CNB -0.026 0.163

ISM, SENB and CNB

Carrara Vincent

SENB vs CNB 0.040 0.040ISK vs SENB 0.005 0.798 CNB 0.072 <0.001

glass

SENB vs CNB 0.067 0.001ISK vs SENB -0.056 0.004 CNB -0.016 0.379

ISK, SENB and CNB

Carrara Vincent

SENB vs CNB 0.040 0.041 2.5 Discussion

The pronounced effect of the materials on the fracture toughness can be explained by their microstructure. Soda–lime–silica glass is microstructurally homogeneous and devoid of crystals. Glass was chosen as a control since it is believed to be a good counterpart for ceramic materials because of its brittleness and toughness, which is usually 0.77 Mpa m−1/2 [22]. The leucite in the traditional leucite-reinforced feldspar-based glass ceramic Carrara Vincent may not only have a reinforcing effect but may also have a toughening effect, while the chemical composition of the matrix may obviously contribute to its toughness increase in any toughness test by deflecting crack extension [5, 7].

Influence of Test Methods on Fracture Toughness

45

It is well-known that for proper material evaluation and development, the critical stress intensity factor, KIc, better known as fracture toughness, should be properly obtained. At present and historically, the IF method, the IS by bending method, the SENB and its derivates SEVNB and SEPB, and SCF methods have been applied in the fracture toughness testing of dental ceramics.

SENB is a traditional method, established to measure straightforward fracture toughness, and has often worked as a reference or standard test technique [13, 23-28]. Values obtained from this test proved to be valid only for ceramics containing coarse crystals (20–40 μm). For ceramics containing fine crystals, less than 2–4 μm in size, the values obtained are expected to be overestimated. In most instances, the notch tip width is usually and practically 100 μm (as in this study) or wider, which tends to increase the fracture toughness measured on finer crystalline materials. There are two modifications based on this straight-through notch technique. Single-edge-precrack (-notched) beam (SEPNB, or SEPB) specimens have a fatigue-induced crack from the tip of the notch and often serve as a reliable reference method in comparison of toughness measurement tests [7, 13, 25, 27]. The SEVNB was developed to overcome the shortcoming of a sharp crack front in SEPB. The saw-cut notch is sharpened with a razor blade and oil-lubricated diamond paste in order to reduce the notch width to twice the grain size [5-7, 23-25, 27]. Both derivatives are time and technique consuming and the pre-cracking/V-notching process easily confers pre-crack front aberrations or sawing-generated penny-shaped flaws, which can lead to a discrepancy in the measured toughness [5]. The SEPB and SEVNB techniques consume more material than the CN specimens, which is a disadvantage as newly developed materials are usually provided in very small quantities. Another problem is that the beam configuration tends to be unstable because most test machines have difficulties in achieving stable crack growth in brittle SENB specimens [13].

In dental journals, only one report was based on the use of the CN method for dental ceramics [8]. This technique is particularly suitable for very brittle materials such as ceramics. The CN toughness is based on plane strain rather than the straight-through-crack theory [13-15, 24, 25, 27]. Because of the high stress concentration at the tip of the chevron notch, a crack initiates at a low applied load. When the crack propagates, the stress intensity factor decreases to a minimum, because the crack front becomes wider because of the triangular shape of the notch. At the minimum value, the crack grows stably with increasing load. Above this point, the crack propagation becomes unstable and the specimen snaps. At the point of minimum stress intensity factor, the fracture toughness can be evaluated with the maximum test load.

Chapter 2

46

An important advantage is that pre-cracking of the specimen is not necessary as this occurs during the first and stable part of the test before the maximum load is reached. Moreover, this is a way of creating a pre-crack with a tip, which resembles that of a propagating crack as much as possible, at least theoretically.

As pointed out in the literature, [14, 15] the calculation of fracture toughness should take the stress intensity factor (instead of using only the maximum fracture value) and the load–line displacement as a function of crack length (generally a0/W) into account in order to obtain compliance calibrations to verify that the R curve has become horizontal and to some extent compensate when it has not. In addition to introduction of Poisson's ratio into the formula, this may engender the comparability of results from a CN specimen with other results [3, 14, 15]. With regard to this consideration, the assumption of 0.25 for the Poisson's ratio of the materials tested in this study might cause a small variance from reality. However, when ν in the formulas is changed to 0.3 or to 0.2, KIc changes to +1.50% or −1.18%, respectively. This discrepancy is very low in consideration of data scatter.

SCF is a good method to assess the fracture toughness, with the advantage of simulating a real surface flaw, the size of which is required in the calculation of the fracture toughness [3, 10, 11, 29, 30]. The flaws are often induced by a Knoop indentation rather than a Vickers indentation. With the help of fractographic analysis, this indentation-induced flaw is demarcated after a three- or four-point flexural specimen is fractured through this artificial flaw on the tensile side. A disadvantage is that the flaw identification requires experience and knowledge, and might be affected by subjective judgment [11, 29]. The residual stress and lateral cracks caused by the indentation interfere with the true value of SCF fracture toughness tests by influencing the maximum load at fracture [30]. Unfortunately, the exact amount of material removed from the residual stress zone and lateral cracks is not well-known at present, and requires detailed and individual investigation. Such an interference is assumed to exist in most dental ceramics since this phenomenon appears in brittle materials with toughnesses less than 3 MPa·m−1/2 [30]. For materials with evident porosity, coarse grain, and ready generation of lateral cracks, SCF is not preferred because of the difficulty in delineating the artificial flaw and its size, and the necessity that fracture of all valid samples tested must originate from that artificial flaw/indentation.

Introduced by Chantiful et al. (1981), the IS method is a convenient method that does not require preparation of a macrocrack (i.e., notch); instead, a small crack (microcrack) is induced in the tensile surface in the form of a Vickers indentation, similar to the technique used for the SCF method [3, 5, 9, 11, 12, 16, 28]. Moreover, it

Influence of Test Methods on Fracture Toughness

47

is unnecessary to gauge the produced crack size for calculation of the fracture toughness, which only requires the indentation load, and as described, the post-indentation crack propagation does not affect the test result very strongly [16]. Nevertheless, increasing the time interval between indentation and the bending test may contribute to data scatter, and lateral cracks generated by the Vickers indentation, especially at a higher load, may cause inaccuracy [5]. Subcritical crack growth of ceramics is believed to be associated with residual stress and moisture. Thus, use of the same criteria for the timeframe from indentation to fracture and oil coverage or an inert gas atmosphere at the indentation site should be useful. Annealing of the specimens before indentation may also be helpful in reducing the error [5, 28] but may be inconsistent with clinical practice. As a small crack, the indentation-induced crack in IS specimens has to be sufficiently distinct from the intrinsic flaws within the specimens. Similar to the SCF test, when the fracture initiates at a location other than the artificial flaw, the test results become invalid. Conceivably, changing the test methods such as three-/four-point, biaxial bending, etc., might cause a greater discrepancy in the calculated toughness value [9, 18] This simple method needs more study for general usage and recommendations for standard organizations.

All fracture toughness methods inevitably face the influence of a material's R-curve behavior [2, 3, 5, 11, 13-15]. Although a rising R curve reflects a generally favorable material property, it also renders any toughness test invalid. During the first part of crack propagation with such materials, the toughness increases, while for the tests it should have reached its steady-state value, which is believed to be size independent.

Also, in this respect, the methods are divided into two types: long, large-crack, or macro-crack toughness methods, and short, small-crack, or microcrack toughness methods. The former includes SENB, SEPB, SEVNB, and CN tests, and the latter type is comprised of SCF and IS tests. Although the fracture toughness of a rising R-curve ceramic is no longer a material constant, results from large- and small-crack test methods represent different performances of the materials' resistance to crack propagation under specific service conditions [2, 4]. In other words, methods could be chosen according to service conditions of the materials being tested.

By reviewing the weaknesses of each test method and individual test method difficulties, one can better understand the results of this study, especially the material–method interaction and method effect on the fracture toughness, which can be influenced by many factors. As reported in the literature, the results of individual fracture toughness tests may or may not be comparable to the results of other

Chapter 2

48

investigators [5-7, 23-29]. The fracture toughness can be reported as apparent fracture toughness. Although the outcome of this study revealed differences by material and test technique used, the goal in refining KIc test methods is to achieve consistent, repeatable results. A few investigations have been performed in an effort to demonstrate that CNB, SEPB, and SCF techniques are in good agreement with each other when using an updated revision of the stress-intensity shape factor and elimination of bias factors [30-33]. With the V-notch technique, other researchers have also attempted to improve the comparability of toughness data [27]. The statistical results of this study imply that in practice, fracture toughness tests always require extremely careful processing of the notch or indentation techniques. 2.6 Conclusion

In this study, a first step was made to compare different toughness test methods by testing the toughness of a traditional feldspar-based leucite-reinforced glass ceramic and a soda lime glass that has a homogeneous microstructure. An interaction effect of the method and the material used was shown. As a consequence, none of the methods tested is suitable as a universal fracture toughness test method. Further research is needed to investigate more extensively the influence of material composition on the fracture toughness test methods' comparability. 2.7 References [1] J.R. Kelly, Clinically Relevant Approach to Failure Testing of All-Ceramic

Restorations, J Prosthet Dent 81 (1999), pp. 652–61. [2] J.R. Kelly, Perspective on Strength, Dent Mater 11 (1995), pp. 103–10. [3] T.L. Anderson, "History and Overview"; pp. 3–34 in Fracture Mechanics:

Fundamentals and Applications, Vol. 1, Edited by T.L. Anderson. CRC Press, Boca Raton, 1991.

[4] J.J. Mecholsky, Fracture Mechanics Principles, Dent Mater 11 (1995), pp. 111–2.

[5] S.S. Scherrer, I.L. Denry and H.W. Wiskott, Comparison of Three Fracture Toughness Testing Techniques Using a Dental Glass and a Dental Ceramic, Dent Mater 14, 246–55 (1998).

[6] H. Fischer and R. Marx, Fracture Toughness of Dental Ceramics: Comparison of Bending and Indentation Method, Dent Mater 18 (2002), pp. 12–9.

Influence of Test Methods on Fracture Toughness

49

[7] J.B. Quinn, V. Sundar and I.K. Lloyd, Influence of Microstructure and Chemistry on the Fracture Toughness of Dental Ceramics, Dent Mater 19 (2003), pp. 603–11.

[8] M.Y. Wen, H.J. Mueller, J. Chai and W.T. Wozniak, Comparative Mechanical Property Characterization of 3 All-Ceramic Core Materials, Int J Prosthodont 12 (1999), pp. 534–41.

[9] M. Albakry, M. Guazzato and M.V. Swain, Fracture Toughness and Hardness Evaluation of Three Pressable All-Ceramic Dental Materials, J Dent 31 (2003), pp. 181–8.

[10] J. Thompson, K.J. Anusavice, A. Naman and H.F. Morris, Fracture Surface Characterization of Clinically Failed All-Ceramic Crowns, J Dent Res 73 (1994), pp. 1824–32.

[11] S.S. Scherrer, J.R. Kelly, G.D. Quinn and K. Xu, Fracture Toughness (KIc) of A Dental Porcelain Determined by Fractographic Analysis, Dent Mater 15 (1999), pp. 342–8.

[12] K. Kvam, H. Herö and G. Øilo, Fracture Toughness Measurements of Some Dental Core Ceramics: A Methodologic Study, Scand J Dent Res 99 (1991), pp. 527–32.

[13] T.L. Anderson, "Fracture Testing in Nonmetals"; pp. 485–530 in Fracture Mechanics: Fundamentals and Applications, Vol. 1, Edited by T.L. Anderson. CRC Press, Boca Raton, 1991.

[14] J.C. Newman, "A Review of Chevron-Notched Fracture Specimens. 'Chevron-Notched Specimens: Testing and Stress Analysis'"; pp. 5–31 in ASTM STP 855, Edited by J.H. Underwood, W.E. Freiman, and F.I. Baratta. American Society for Testing and Materials, Philadelphia, PA, 1984.

[15] J.H. Hanson and A.R. Ingraffea, "Standards for Fracture Toughness Testing of Rock and Manufactured Ceramics: What Can We Learn for Concrete?,"Cement Concrete Aggregates, CCAGDP 19 (1997), pp. 103–11.

[16] P. Chantikul, G.R. Anstis, B.R. Lawn and D.B. Marshall, A Critical Evaluation of Indentation Techniques for Measuring Toughness: II. Strength Method, J Am Ceram Soc 64 (1981), pp. 539–54.

[17] G.R. Anstis, P. Chantikul, B.R. Lawn and D.B. Marshall, A Critical Evaluation of Indentation Techniques for Measuring Fracture Toughness: I. Direct Crack Measurements, J Am Ceram Soc 64 (1981), pp. 533–8.

[18] G.D. Quinn and R. Morrell, Design Data for Engineering Ceramics: A Review of The Flexure Test, J Am Ceram Soc 74 (1991), pp. 2037–66.

Chapter 2

50

[19] D.B. Marshall, T. Noma, and A.G. Evans, A Simple Method for Determining Elastic-Modulus-to-Hardness Ratios Using Knoop Indentation Measurements, Comm Am Ceram Soc (1982) C-175–6.

[20] Z.D. Guan and G.S. Sun, "Study on Method to Determine KIc Values of Ceramics by Chevron-Notched 3-pt Bending Specimens"; pp. 586–91 in 3rd International Symposium on Ceramic Materials and Components for Engines. Las Vegas, Nov. 1988, Edited by V. J.Tennery. The American Ceramic Society Inc., Westerville, OH, 1989.

[21] ASTM 399-74. American Society for Testing Materials. Standard Test Method for Plane-Strain Fracture Toughness of Metallic Materials. ASTM 1983 E399-74, pp. 923–36. ASTM, Philadelphia, PA, 1983.

[22] V.M. Sglavo and D.J. Green, Influence of Indentation Crack Configuration on Strength and Fatigue Behaviour of Soda-Lime Silicate Glass, Acta Metall Mater 43 (1995), pp. 965–72.

[23] T. Nishida, Y. Hanaki and G. Pezzotti, Effect of Notch-Root Radius on the Fracture Toughness of a Fine-Grained Alumina, J Am Ceram Soc, 77 (1994), pp. 606–8.

[24] A.K. Mukhopadhyay, S.K. Datta and D. Chakraborty, Fracture Toughness of Structural Ceramics, Ceram Int 25 (1999), pp. 447–54.

[25] G.A. Gogotsi, Fracture Toughness of Ceramics and Ceramic Composites, Ceram Int 29 (2003), pp. 777–84.

[26] Orange, H. Tanaka and G. Fantozzi, Fracture Toughness of Pressureless Sintered Silicon Carbide: A Comparison of KIc Measurement Methods, Ceram Int 13 (1987), pp. 159–65.

[27] H. Awaji, T. Watanabe and Y. Sakaida, Fracture Toughness Measurements of Ceramics by V Notch Technique, Ceram Int 18 (1992), pp. 11–7.

[28] S. Mashio and O. Sbaizero, Fracture Toughness of (Ce) Stabilized ZrO2/Al2O3 Composites, Ceram Int 15, (1989), pp. 363–8.

[29] G.D. Quinn, J.J. Swab and M.D. Hill, Fracture Toughness by the Surface Crack in Flexure (SCF) Method: New Test Results, Ceram Eng Sci Proc 18 (1997), pp. 163–72.

[30] G.D. Quinn and J.A. Salem, Effect of Lateral Cracks on Fracture Toughness Determined by the Surface-Crack-in-Flexure Method, J Am Ceram Soc 85 (2002), pp. 873–80.

Influence of Test Methods on Fracture Toughness

51

[31] S.S. Scherrer, I.L. Denry, H.W. Wiskott and U.C. Belser, Effect of Water Exposure on the Fracture Toughness and Flexure Strength of a Dental Glass, Dent Mater 17 (2001), pp. 367–71.

[32] J. Salem, L. Ghosn, M. Jenkins and G.D. Quinn, Stress Intensity Factor Coefficients for Chevron-Notched Flexure Specimens and a Comparison of Fracture Toughness Methods, Ceram Eng Sci Proc 20 (1999), pp. 503–12.

[33] G.D. Quinn, R.J. Gettings and K. Xu, Standard Reference Material 2100: Ceramic Fracture Toughness, Ceram Eng Sci Proc 20 (1999), pp. 513–23.

Chapter 2

52

53

CHAPTER 3

Fracture toughness comparison of three test methods with four dental porcelains

Keywords: Dental porcelains, Fracture toughness, Chevron notched, Indentation strength, Single-edge-notched beam

Chapter 3

54

3.1 Abstract Objective: The aim of this study was to compare three fracture toughness test methods, using four commercial dental porcelains. Materials and methods: The fracture toughness test techniques involved were: the single-edge-notched beam (SENB), the indentation strength method (IS), and a rather convenient ASTM standard for advanced ceramics, which is still rarely used in dental ceramic research, the Chevron-notched beam method (CN). Duceram, Duceram LFC, Sintagon Zx and Carrara Vincent were chosen for study. Data was analyzed by two-way and paired ANOVA. Result: No statistical difference was found between the CN and SENB methods with four dental porcelains, but IS was not always in statistical agreement with SENB or CN. Statistical agreement among all three methods occurred only with Duceram LFC. Conclusion: The different test methods did not always lead to the same ranking or values of fracture toughness. Yet the toughness results of the SENB method were comparable to those of the CN method for all the four dental porcelains tested in this study.

Fracture toughness of three test methods

55

3.2 Introduction Ceramics have been increasingly employed in dental restorations because of

their esthetics and biocompatibility. Because these materials are very brittle they are also vulnerable to tensile stress, which opens small pre-existing defects, which work as stress raisers. Smoothly finished ceramics are often very strong, but scratches and other small defects, which cumulate with time, seriously degrade the strength of these materials. With little simplifications, a crack initiates or grows when the stress intensity (KI) around a defect or crack tip exceeds the fracture toughness (KIc), which is a material property [1-4]. The magnitude of the stress intensity depends on the size and shape (tip radius) of the pre-existing defect or crack, and the applied stress.

In dentistry, increasing attention has been paid to KIc. The research has mainly been focused on two issues: studies into the fracture toughness test methods, and to use the methods to rank dental materials. As many test methods of fracture toughness have limitations, it is essential to assess the consistency of the various techniques by comparing the values obtained with various dental ceramics [3, 4].

A rather convenient ASTM standard for advanced ceramics, which is still rarely used in dental research, is the Chevron-notched beam method (CN). It was hypothesized that this method might be a better method for determining the KIc of dental ceramics than the more popular single-edge-notched beam (SENB) or the indentation strength method (IS). The aim of this study is to assess the comparability of these three test techniques with four dental veneering porcelains. 3.3 Materials and Methods

The dentin powders of four veneer ceramic systems were used. These were silicates or glass ceramics, reinforced with leucite crystals. Dentin porcelains are intended to add color and should provide opacity to mask the core material. Duceram and Duceram LFC (Degussa Dental GmbH, Hanau, Germany) are used with metal–ceramic systems. Sintagon Zx (Elephant Dental B.V., Hoorn, The Netherlands) is bonded to zirconia all-ceramic substrates. Carrara Vincent (Elephant Dental B.V., Hoorn, The Netherlands) has been designed for metal–ceramic and for Carrara Pressable Core (CPC) all-ceramic systems.

Three fracture toughness test techniques were compared, the indentation strength method (IS), the Chevron-notched beam method (CN) and the single-edge-notched beam method (SENB). Beam-shaped specimens were prepared for at least 10 measurements for each test method. The test configuration is listed in Table 3.1.

Chapter 3

56

Table 3.1 The test configuration of three measurement techniques. KIc test method CN SENB IS Precrack processing sawed notch sawed notch Vickers’ indentation

Precrack type growing crack large crack (notch)radial cracks from

indentation

Fracture plane Chevron triangle

half section rectangle

whole section rectangle

Proportions (B:W:S) 2:3:12 1:2:8 2:3:22 Dimensions (BxWxL) 3×4.5×26 mm 3×6×26 mm 2×3×25 mm Specimen preparation

The porcelain powders were condensed in a brass mold (4 mm × 30 mm × 40 mm). The blocks were removed from the mold, and fired in a dental porcelain furnace (STRATOS, Elephant Dental B.V., Hoorn, The Netherlands) in compliance with the manufacturer's recommendations. The cooling time was prolonged to 1 h to accommodate the relatively large size of the blocks as compared to normal applications. The upper and lower surfaces of the fired porcelain blocks were ground parallel and the blocks were sawn into rectangular beams (Table 3.1). The specimens were polished with wet silicon carbide papers (grits 400, 600, and 1200, successively) on all surfaces except the end planes, and the sharp edges were chamfered. The specimens were annealed to release residual stresses by keeping these close to the glass transition temperature for 90 min and cooling to 100 °C for another 90 min. This was 450 °C for Duceram LFC, Carrara Vincent, and Sintagon Zx, and 575 °C for Duceram. Indentation strength method (IS)

Vickers indentations were made (HM-124 Hardness Testing Machine, Mitutoyo Corp., Kanagawa, Japan) in the middle of the tensile surfaces of the beams at a load of 19.6 N. The radial cracks, which arise from this load, serve as the pre-cracks in this test. Because these cracks continue to grow during the first few minutes following indentation, the beams were loaded after 20–30 min in a three-point bending set up at 0.05 mm/min until fracture occurred in a tensilometer (ACTA Intense, ACTA, Amsterdam, The Netherlands). Specimens, where the fracture did not originate from the Vickers indentation, were excluded from the study, and testing was continued until

Fracture toughness of three test methods

57

at least 10 acceptable test results were obtained. The fracture strength (σf) of indented specimens was calculated according to following formula:

223WBFS

f =σ

where W is the fracture load; L the span length; b the specimen width; d is the specimen thickness.

For each material, the elastic modulus (E) was determined with a three-point bending test on beams without indentation (n = 10). The bending deflection (q) of the specimens loaded until failure was recorded. The modulus was calculated with:

lWBFSE 3

3

4=

The Vickers hardness (H) was measured on broken specimens (n = 10) using a load of 1.96 N for 15 s, which magnitude prevented the introduction of radial cracks. The hardness was calculated with H = 1.854P/(2a)2, where P is the indentation load (1.96 N), and 2a is the average of the two diagonals of the indentation.

The facture toughness (KIc) was obtained by calculation of this equation [5],

( ) ( ) 4/33/18/1/ PHEK fIc ση=

where η is the geometrical constant (0.596), and P is the indentation load on the IS beams. The geometrical constant is slightly greater than the 0.59 used by Chantikul et al. [5], because they used 2 instead of 1.854 in the Vickers equation. Chevron-notched beam method (CN)

Conforming to a previous study [6], a 0.1 mm diamond-sawing disc (ISOMET 1000, Buehler, Lake Bluff, USA) was used to create a notch (Figure 3.1) with a Chevron angle θ of 60 ± 1.5° and a0/W ratio of 0.1–0.35. The beams were loaded in a three-point bending test. The simple variant of the CN was used, where just the maximum force Fm is used for the calculations and no beams are rejected, regardless of the (omissible) load displacement plots, which were made at two samples per second. The Chevron angle θ and notch length a0 were measured at the two fractured sections of each specimen using optical microscopy (10×, measuring precision 1 μm). The toughness was calculated with the following equation:

Chapter 3

58

)/()2/(tan)1(2 0212123 Waf

WFK m

Ic ⋅−

=θν

( ) ( ) ( )2000 /53.179/708.20959.17/ WaWaWaf ++=

where S is the span; B the specimen width; W the specimen height; f(a0/W) the stress intensity shape factor; ν is the Poisson's ratio. The Poisson's ratio was rounded to 0.25 as recommended in ISO 6872 for biaxial flexural strength calculation.

Figure 3.1 Chevron-notched beam test. Single-edge-notched beam method (SENB)

The notches of the specimens were cut with a 0.1 mm diamond saw disc. The saw depth c was nearly half of the specimen's height W (Figure 3.2). The specimens were fractured in a three-point bending test. The two halves of the broken samples were used for the measurement of the notch depth c under an optical microscope. The length c was the average of the six values at three locations of the notch: in the middle and at two lateral sides of each section. The toughness value was calculated according to the following formula [7]:

( )WcfW

SBF

K cIc /2/3 ⋅⋅=

( ) ( ) ( ) ( ) ( ) ( ) 2/92/72/52/32/1 /7.38/6.37/8.21/6.4/9.2/ WcWcWcWcWcWcf +−+−=

where Fc is the critical load; B the specimen width; S the supporting span; f(c/W) is the stress intensity shape factor.

Fracture toughness of three test methods

59

Figure 3.2 SENB test. Statistical analysis

The significances of the material effect (difference between porcelains) and the test method effect (difference between methods) were evaluated by two-way ANOVA. The difference within each factor group was assessed by pairwise comparisons. The significance level of all statistical analysis was p = 0.05. 3.4 Results

The results are listed in Table 3.2. The data variation coefficients (data scatter), calculated as the standard deviations divided by the means (in percent), were low and ranged from 3.6 to 10.2%. Table 3.2 Fracture toughness (MPa m−1/2) of the tested dental porcelains Material Mean fracture toughness±SD and the data scatter in parentheses IS CN SENB Duceram Carrara Vincent Sintagon ZX Ducera LFC

0.884±0.088 (10.0%) 0.809±0.053 (6.5%) 0.779±0.063 (8.1%) 0.692±0.056 (8.0%)

0.949±0.059 (6.2%) 0.843±0.062 (7.4%) 0.698±0.045 (6.4%) 0.677±0.050 (7.4%)

0.944±0.043 (4.6%) 0.882±0.032 (3.6%) 0.665±0.047 (7.1%) 0.708±0.072 (10.2%)

A total of three Duceram IS specimens were rejected because the fracture did

not originate from the indentation, and extra beams were tested. As indicated in Table 3.3, SENB displayed statistical agreement with CN for

all four dental porcelains tested in this study. IS was in agreement with CN for two materials, Duceram LFC and Carrara Vincent, and with SENB only with Duceram LFC.

Chapter 3

60

Table 3.3 Pairwise comparisons of the method effect within material group Material Method comparisons Difference(ns) P

IS vs SENB -0.060 0.020

CN -0.066 0.011 Duceram

SENB vs CN (-0.006) 0.823

IS vs SENB -0.073 0.005

CN (-0.034) 0.187 Carrara Vincent

SENB vs CN (0.039) 0.128

IS vs SENB 0.114 0.000

CN 0.081 0.002 Sintagon Zx

SENB vs CN (-0.033) 0.194

IS vs SENB (-0.016) 0.525

CN (0.015) 0.550 Ducera LFC

SENB vs CN (0.031) 0.225

Table 3.4 Pairwise comparisons of the material effect within method group Material Method comparisons Difference(ns) P

Duceram vs Carrara Vincent 0.075 0.003 Sintagon Zx 0.105 0.000 Ducera LFC 0.191 0.000

Carrara Vincent vs Sintagon Zx (0.030) 0.238 Ducera LFC 0.117 0.000

IS

Sintagon Zx vs Ducera LFC 0.087 0.001 Duceram vs Carrara Vincent 0.107 0.000

Sintagon Zx 0.251 0.000 Ducera LFC 0.272 0.000

Carrara Vincent vs Sintagon Zx 0.144 0.000 Ducera LFC 0.165 0.000

CN

Sintagon Zx vs Ducera LFC (0.021) 0.403 Duceram vs Carrara Vincent 0.611 0.021

Sintagon Zx 0.278 0.000 Ducera LFC 0.235 0.000

Carrara Vincent vs Sintagon Zx 0.217 0.000 Ducera LFC 0.174 0.000

SENB

Sintagon Zx vs Ducera LFC (-0.043) 0.096

Fracture toughness of three test methods

61

In Table 3.4, all methods found differences between all materials except with Sintagon Zx; IS found no difference with Carrara Vincent, and CN and SENB found no difference between Duceram LFC and that material.

The ranking of the porcelains by the average KIc of all three methods from high to low is Duceram, Carrara Vincent, Sintagon Zx and Duceram LFC. 3.5 Discussion

Many fracture toughness test methods have been developed [3] and [8]: the single-edge-notched beam (SENB), and its two derivatives, the single-edge pre-cracked beam (SEPB) and the single-edge-V-notched beam (SEVNB), the Chevron-notch method (CN), surface-crack-in/by-flexure (SCF, previously also called controlled surface micro-crack), double torsion (DT), the double cantilever beam (DCB), indentation strength in/by bending (IS) with Vickers’ or Knoop's indentation, the indentation fracture method (IF), etc.

The IF, SENB and IS methods have been frequently involved in dental ceramics research [9-15]. The IF and IS, which use hardness indentations have been investigated in most reports. Other studies compared IF to SEVNB, IS to SEVNB, or IS to SCF [9-15].

As many methods often do not produce the same result for the same material, or even not give the same ranking for a set of materials [9-10, 12-15], it is clear that many methods are not always accurate. Other literature suggests that fracture toughness assessment is rather sensitive to the type of method, configuration, and processing procedures [17-24], similar to strength determination [16]. Obtained values may be inaccurate, not always consistent with each other, and different methods may offer confusing values and rankings for comparison.

For reliable and reproducible measurements of fracture toughness, recommendations of standardization organizations could be followed. At present, the single-edge pre-cracked beam (SEPB), V-notched beam (SEVNB), the surface crack in flexure (SCF) and the Chevron-notched beam (CN) have been accepted as standard methods by the ASTM or the CEN. These have been proved to give very close and reproducible KIc values for several advanced ceramics. The IS, IF, and SENB are not considered any further in these rather thorough standardization processes [25-29].

Many methods have shortcomings [3, 8-15, 17-29]. With most specimen configurations crack growth is unstable and KIc is assessed at the onset of crack extension, because at that moment the specimen breaks. Therefore, certain materials can be rather sensitive to the sharpness of the pre-crack tip, which ideally should have

Chapter 3

62

the same radius as the propagating crack, which causes failure. In the SENB, the crack initiates at a not very predictable radius, related to the roughness of the notch tip, but small enough with the present materials, as the results were similar to the CN. With the SEVNB it is relatively easy to produce the V-shaped tip at the end of the notch with a razor blade. Although the microstructure of the ceramic influences the radius at the tip of the sharpened V, it is much smaller, more easily predictable, and suited to much smaller particle size materials. With the SEPB method, the specimens are pre-cracked in a specially designed bridge-anvil structure with a carefully applied high compressive load, until the slight sound of the pre-crack is heard. This should yield a quite realistic tip radius over the entire width, and some toughening (rising R-curve) may stabilize, but it is a technically sensitive method. For the SCF, it is important to locate the border of the pre-crack, which requires fractographic experience and is difficult for ceramics with porosities or a coarse crystalline microstructure. Furthermore, it is critical to remove the residual stress around the Knoop's indentation. The amount of removal, however, varies across materials.

The CN method has been designed to overcome the problems with elaborate pre-cracking. The triangularly widening crack front leads to insensitivity to crack furcation and origin deviation [3, 8, 25-31]. With the CN notch, pre-cracking occurs automatically and inevitably when the load is applied. Already at a small load, a crack initiates in the tip of the triangle, due to stress concentration. At first the crack should grow stably with increasing load, because the crack front widens. This was visible as bending of the load displacement plots. The formula calculates KIc at the point where the maximum load occurs, which is at the crack length where further crack propagation may become unstable and the specimen breaks. Most specimens fractured beyond the maxima of the load displacement plots, where the load had been constant or decreasing a few percent for several seconds, which indicates stable crack growth. With each material two or three beams fractured before, or shortly before, the plots had fully bent horizontally. The KIc of one Duceram CN beam was 3% overestimated as it showed a small pop-in, i.e. the load increased linearly to a sharp peak, where it dropped about 4% and increased more slowly to a normally rounded peak of about 3% less. This beam's KIc was 7% greater than the Duceram average (Table 3.2). One LCF CN beam showed a single pop-in, i.e. it fractured with a single linear peak, suggesting unstable cracking. This beam's KIc was 17% greater than its average. With 10 specimens tested, these pop-ins have increased the averages by 0.3 and 1.7%, respectively. It seems safe to use just the maximum force with the CN method.

Fracture toughness of three test methods

63

Although only a large notch is sawn, the pre-crack at which KIc is assessed, is generally a growing crack. This method could rival the SEPB, but the crack front is less wide. Because the CN is reportedly elaborate, it should perhaps be stated that it is not when the beams are prepared one by one as in the present study. Compared to the simple SENB each specimen needs to be rotated and sawn again, which is perhaps less effort than the IS specimens, as these require extra (E and H) experiments and specimens. The two broken halves of a long rod CN experiment may be used for two short-rod experiments, when small amounts of experimental material are available. In dental literature, one article [32] used the short-rod CN method to measure the fracture toughness of three dental ceramics.

Because dental porcelains and ceramics have a broad range in microstructure, chemistry, and fracture toughness [33], it could be useful to systematically assess the comparability of the various test methods with the standardized (ASTM, CEN) methods, since these have been proved to work on advanced ceramics.

The IS method seems an economically affordable alternative to KIc estimation only when a diamond saw is not available, as its limitations are much more complex than the straightforward notch problem of the SENB.

Considering its history, the Chevron-notched beam values of the present study could perhaps be seen as more accurate than the others. With the present four dental porcelains only the SENB method was statistically comparable to the CN. 3.6 Conclusion 1. The CN method, and recording just the highest load seems accurate with dental

ceramics. This ATSM standard requires less effort than many other tests. 2. With the present materials the SENB was statistically close to CN. 3. IS was not always in statistical agreement with SENB or CN and further

investigation is needed, as the limitations are unclear. 4. The fracture toughness sequence from high to low is Duceram, Carrara Vincent,

Sintagon Zx, and Duceram LFC, according to the three test methods. 3.7 References [1] J.R. Kelly, Clinically relevant approach to failure testing of all-ceramic

restorations, J Prosthet Dent 81 (1999), pp. 652–661. [2] J.R. Kelly, Perspective on strength, Dent Mater 11 (1995), pp. 103–110.

Chapter 3

64

[3] T.L. Anderson, History and overview. In: T.L. Anderson, Editor, Fracture mechanics: fundamentals and applications vol. 1, CRC Press, Boca Raton (1991), pp. 3–34.

[4] J.J. Mecholsky, Fracture mechanics principles, Dent Mater 11 (1995), pp. 111–112.

[5] P. Chantikul, G.R. Anstis, B.R. Lawn and D.B. Marshall, A critical evaluation of indentation techniques for measuring toughness: II. Strength method, J Am Ceram Soc 64 (1981), pp. 539–554.

[6] Z.D. Guan and G.S. Sun, Study on method to determine KIc values of ceramics by Chevron-notched 3-pt bending specimens. In: V.J. Tennery, Editor, Proceedings of the Third International Symposium on Ceramic Materials and Components for Engines (1989), pp. 586–591.

[7] American Society for Testing Materials (ASTM 399-74), Standard test method for plan-strain fracture toughness of metallic materials. E399-74, ASTM, Philadelphia, PA (1983) p. 923–36.

[8] T.L. Anderson, Fracture testing in nonmetals. In: T.L. Anderson, Editor, Fracture mechanics: fundamentals and applications vol. 1, CRC Press, Boca Raton (1991), pp. 485–530.

[9] K. Kvam, H. Herö and G. Øilo, Fracture toughness measurements of some dental core ceramics: a methodologic study, Scand J Dent Res 99 (1991), pp. 527–532.

[10] S.S. Scherrer, I.L. Denry and H.W. Wiskott, Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic, Dent Mater 14 (1998), pp. 246–255.

[11] S.S. Scherrer, I.L. Denry, H.W. Wiskott and U.C. Belser, Effect of water exposure on the fracture toughness and flexure strength of a dental glass, Dent Mater 17 (2001), pp. 367–371.

[12] S.S. Scherrer, J.R. Kelly, G.D. Quinn and K. Xu, Fracture toughness (KIc) of a dental porcelain determined by fractographic analysis, Dent Mater 15 (1999) (5), pp. 342–348.

[13] M. Guazzato, M. Albakry, M.V. Swain and J. Ironside, Mechanical properties of In-ceram alumina and In-ceram zirconia, Int J Prosthodont 15 (2002) (4), pp. 339–346.

[14] H. Fischer and R. Marx, Fracture toughness of dental ceramics: comparison of bending and indentation method, Dent Mater 18 (2002), pp. 12–19.

Fracture toughness of three test methods

65

[15] M. Albakry, M. Guazzato and M.V. Swain, Fracture toughness and hardness evaluation of three pressable all-ceramic dental materials, J Dent 31 (2003), pp. 181–188.

[16] G.D. Quinn and R. Morrell, Design data for engineering ceramics: a review of the flexure test, J Am Ceram Soc 74 (1991), pp. 2037–2066.

[17] T. Nishida, Y. Hanaki and G. Pezzotti, Effect of notch-root radius on the fracture toughness of a fine-grained alumina, J Am Ceram Soc 77 (1994), pp. 606–608.

[18] A.K. Mukhopadhyay, S.K. Datta and D. Chakraborty, Fracture toughness of structural ceramics, Ceram Int 25 (1999), pp. 447–454.

[19] G.A. Gogotsi, Fracture toughness of ceramics and ceramic composites, Ceram Int 29 (2003), pp. 777–784.

[20] G. Orange, H. Tanaka and G. Fantozzi, Fracture toughness of pressureless sintered silicon carbide: a comparison of KIc measurement methods, Ceram Int 13 (1987), pp. 159–165.

[21] H. Awaji, T. Watanabe and Y. Sakaida, Fracture toughness measurements of ceramics by V notch technique, Ceram Int 18 (1992), pp. 11–17.

[22] S. Mashio and O. Sbaizero, Fracture toughness of (Ce) stabilized ZrO2/Al2O3 composites, Ceram Int 15 (1989), pp. 363–368.

[23] G.D. Quinn, J.J. Swab and M.D. Hill, Fracture toughness by the surface crack in flexure (SCF) method: new test results, Ceram Eng Sci Proc 18 (1997), pp. 163–172.

[24] G.D. Quinn and J.A. Salem, Effect of lateral cracks on fracture toughness determined by the surface-crack-in-flexure method, J Am Ceram Soc 85 (2002), pp. 873–880.

[25] J. Salem, L. Ghosn, M. Jenkins and G.D. Quinn, Stress intensity factor coefficients for Chevron-notched flexure specimens and a comparison of fracture toughness methods, Ceram Eng Sci Proc 20 (1999), pp. 503–512.

[26] G.D. Quinn, R.J. Gettings and K. Xu, Standard reference material 2100: ceramic fracture toughness, Ceram Eng Sci Proc 20 (1999), pp. 513–523.

[27] G.D. Quinn, The fracture toughness round robins in VAMAS: what we have learned. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), p. 1409.

[28] G.D. Quinn, K. Xu, R.G. Gettings, J.A. Salem and J.J. Swab, Does anyone know the real fracture toughness? SRM 2100: the world's first ceramic fracture

Chapter 3

66

toughness reference material. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), p. 1409.

[29] M.G. Jenkins, J.A. Salem, G.D. Quinn and I. Bar-On, Development, verification, and implementation of a national consensus fracture toughness test method standard for advanced ceramics. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, American Society for Testing and Materials (ASTM STP), West Conshohocken, PA (2002), p. 1409.

[30] J.C. Newman, A review of Chevron-notched fracture specimens. In: J.H. Underwood, W.E. Freiman and F.I. Baratta, Editors, Chevron-notched specimens: testing and stress analysis, American Society for Testing and Materials (ASTM STP 855), Philadelphia, PA (1984), pp. 5–31.

[31] J.H. Hanson and A.R. Ingraffea, Standards for fracture toughness testing of rock and manufactured ceramics: what can we learn for concrete?, Cement Concrete Aggreg (CCAGDP) 19 (1997), pp. 103–111.

[32] M.Y. Wen, H.J. Mueller, J. Chai and W.T. Wozniak, Comparative mechanical property characterization of 3 all-ceramic core materials, Int J Prosthodont 12 (1999), pp. 534–541.

[33] J.B. Quinn, V. Sundar and I.K. Lloyd, Influence of microstructure and chemistry on the fracture toughness of dental ceramics, Dent Mater 19 (2003), pp. 603–611.

67

CHAPTER 4

Fracture toughness determination of two dental porcelains with the indentation strength in bending method

Keywords: Dental porcelain, Fracture toughness, Indentation strength

Chapter 4

68

4.1 Abstract Objective: This study was to investigate the influence of the bending test configurations and the crosshead displacement speeds on the fracture toughness (KIc) of dental porcelains obtained with the indentation strength in bending (ISB) method. Methods: The strength of the dental veneering porcelains Duceram and Sintagon Zx, Vickers’ indented at a load of 2 kg was measured at a crosshead speed of 0.5 mm/min with three test configurations, which were 3-point, 4-point, and biaxial bending. Two more groups of Sintagon Zx were tested the same way, but at speeds of 0.1, and 0.05 mm/min, respectively. Both porcelains, the three crosshead speeds, and the three test configurations were compared statistically. Results: Duceram had a higher toughness than Sintagon Zx with all three test configurations and there was no significant difference between three test configurations with either porcelain. Within the crosshead speed groups of Sintagon Zx, a significant difference was found only in the 0.5 mm/min group between the 3-point, and 4-point configurations.

Within the configuration groups, significant differences were found between all speeds with the 3-point configuration and only between the highest and lowest speed with the 4-point and the biaxial tests. Conclusion: The crosshead displacement speed can cause statistically different results of fracture toughness obtained with the ISB method.

Fracture toughness determination of two dental porcelains

69

4.2 Introduction Application of porcelain and ceramic in prosthodontics resulted in a naturally

appearing restoration, no matter if it belongs to metal- or all-ceramic restorations, compared with metal restorations [1, 2]. However, dental porcelains and ceramics are brittle materials, which generally fail in tension due to their limited ductility, which restricts the ability to absorb a great deal of elastic strain energy before fracture [3]. A major weakness of these materials is the sensitivity to flaws, which may have developed as a result of thermal, chemical or mechanical processes, and act as local stress raisers. At a certain critical applied stress, a crack can originate from a flaw and propagate, engendering final catastrophic fracture [3-6]. Fracture toughness, KIc is defined as the critical stress intensity level at which a given flaw starts extending and provides insight into the potential resistance to crack growth of a material [7-9]. So in the last decade in the field of dental porcelains and ceramics research, much attention has been paid to the fracture toughness [2].

Fracture toughness is reported to be an intrinsic property of ceramics, which indicates the ability of dental porcelains and ceramics to resist crack extension. So the accurate measurement of KIc is essential. However, determination of KIc is technically rather sensitive, and the obtained values and rankings may be different depending on the techniques and procedures used [8-10].

Among the many methods for fracture toughness determination, the indentation strength in bending (ISB) method as introduced by Chantikul et al. [11], is relatively simple, and reportedly accurate and reproducible, compared to other methods [8, 9, 12, 13]. It uses the bending strength of for example beams, which have an indentation at the center of the tensile surface. There is no need to determine the initial size of the flaw, because an entry of the indentation load in the equation is used instead. Since the 1990s there has been a dramatic increase in the use of this method [14-26]. However, uncertainty may be introduced with the test configuration [27, 28], which influences the result of KIc. For example, first, 3-point, 4-point, and biaxial bending tests, lead to different values due to different effective surface areas or volumes subjected to stress. Second, obtained strength values could depend on the crosshead speed.

In this light, the aim of the present study was to compare fracture toughness values obtained for two dental porcelains with the ISB method with different test configurations and crosshead speeds. The assumption subject to investigation was that there was no statistical difference between KIc values obtained with the ISB method, when different test configurations and crosshead speeds were used.

Chapter 4

70

4.3 Materials and Methods Two veneering dentine porcelains were involved in the comparison: Duceram

(Degussa Dental GmbH, Hanau, Germany) and Sintagon Zx (Elephant Dental B.V., Hoorn, The Netherlands) (Table 4.1). Table 4.1 Elastic moduli and Vickers’ hardness in GPa and standard deviation in

parentheses Materials Elastic modulus Vicker's hardness Duceram Sintagon Zx

60.8 (2.4) 64.5 (1.6)

4.97 (0.11) 4.43 (0.08)

Specimen preparation

For the beam specimens in the 3-point and 4-point setups, porcelain blocks were made by condensing the powder a in brass mould (4 mm × 30 mm × 40 mm). The discs for the biaxial tests were formed by condensing the powders in a stainless steel ring with an inner diameter of 21 mm and a height of 2.5 mm.

The blocks and the discs were removed from the mould, and fired in a dental porcelain furnace (STRATOS, Elephant Dental B.V., Hoorn, The Netherlands) in compliance with the manufacturer's recommendations. The cooling time was prolonged to 1 h to accommodate for the relatively large size of the bodies as compared to normal applications. The upper and lower surfaces of the square blocks were ground parallel to a thicknesses of 3 mm and sawed into rectangular beams of 2 mm × 3 mm × 26 mm. The round discs were ground parallel to a thickness of 2 mm and the diameter of approximately 18 mm was left unchanged. All specimens were polished on all surfaces with wet silicon carbide paper of the grits 400, 600, and 1200 successively and the sharp edges were chamfered, except the end planes of the beams and the circumference of the discs, which were left as fired.

The specimens were annealed to release residual stresses by keeping these close to the glass transition temperature for 90 min and cooling to 100 °C for another 90 min. This was 450 °C for Sintagon Zx, and 575 °C for Duceram. Fracture toughness with the ISB method

Vickers’ indentations were made (HM-124 Hardness Testing Machine, Mitutoyo Corp., Kanagawa, Japan) at the center of the tensile surface of the specimens at a load of 19.6 N during 15 s. The radial cracks, which arise with this load, serve as

Fracture toughness determination of two dental porcelains

71

the pre-crack in this test. Within 30 min following the indentation, the specimens were loaded until fracture occurred in a tensilometer (ACTA Intense, ACTA, Amsterdam, The Netherlands) [27-30]. Specimens where the fracture did not originate from the Vickers’ indentation were excluded from the study and testing was continued until at least 10 acceptable test results were acquired.

The facture toughness was obtained with [11],

( ) ( ) 4/33/18/1/ PHEK fIc ση=

where η is the geometrical constant (0.59), E the elastic modulus, H the Vickers’ hardness, and P is the indentation load. Calculation of the strength, σf, depends on the test setup and is discussed later. For each porcelain the elastic modulus was determined with a 3-point bending test on 20 beams without indentation at a crosshead speed of 0.5 mm/min. The bending deflection of the specimens was recorded. The modulus was calculated with:

qbd

WLE 3

3

4=

where E is the elastic modulus, W the fracture load, L the span length (20 mm), B the specimen width, D the specimen thickness, and q is the bending deflection.

The Vickers’ hardness (H) was measured on broken specimens (n = 10) using a load of 1.96 N during 15 s, which magnitude prevented the introduction of radial cracks. The hardness was calculated with H = 1.854P/(2a)2, where P is indentation load (1.96 N), and 2a is the average of the two diagonals of the indentation. The test configurations and crosshead speeds The three test configurations used, were 3-point, 4-point, and biaxial bending. The last was carried out in a ball-on-ring setup. The strength with the 3-point, and 4-point bending test was calculated according to the following formula [27]:

221

2)(3

BDLLW

f−

=σ

Chapter 4

72

where σf is the strength, W the fracture load, L1 the supporting span length of 20 mm, L2 the loading span length, which is 10 mm in the 4-point test and 0 mm in the 3-point test, B the specimen width, and D is the specimen thickness. The strength with the biaxial bending test (ball-on-ring) was calculated as the following equation:

2

222

4)]1/())1()/)}((2/(1{)/ln(21)[1(3

BDDDbbDW sss

fπ

νυνσ +−−+++=

where σf is the strength, W the fracture load, Ds the diameter of the support circle (16 mm), b the diameter of the area with uniform load (=2B/3), B the thickness, and D is the diameter of disc. For ν, the Poisson's ratio, a value of 0.25 was used, as recommended with ISO 6872 [28-30].

Sintagon Zx was tested in all three configuration with crosshead speeds of 0.05, 0.1, and 0.5 mm/min. Duceram was tested only at 0.5 mm/min. Statistic analysis The fracture toughness values among different groups were subject to two-way ANOVA and pair wise comparison, where a p-value of less than 0.05 was considered statistically significant. The comparisons were carried out between all groups of Sintagon Zx, between the configurations with Duceram, and between Duceram and Sintagon Zx at 0.5 mm/min. 4.4 Results

Elastic moduli and Vickers’ hardness values are listed in Table 4.1 and the fracture toughness values and standard deviations are listed in Table 4.2. Table 4.2 ISB fracture toughness in MPa m1/2 with standard deviation in parentheses

Crosshead speed (mm/min) Material

Bending test methods

0.5 Duceram

0.5 Sintagon Zx

0.1 Sintagon Zx

0.05 Sintagon Zx

3-Point 4-Point Biaxial

1.100 (0.068) 1.099 (0.065) 1.095 (0.083)

0.885 (0.064) 0.829 (0.041) 0.850 (0.047)

0.826 (0.081) 0.784 (0.050) 0.815 (0.040)

0.760 (0.069) 0.761 (0.047) 0.787 (0.042)

Fracture toughness determination of two dental porcelains

73

With all three test configurations the ISB KIc values for Duceram were

significantly higher than for Sintagon Zx and no significant difference existed between the test configurations with either porcelain in Table 4.3 and Table 4.4. Table 4.3 Pairwise comparison of the configuration effect within the material effect with

Duceram and Sintagon Zx Material Configurations p Duceram 3-Point vs. 4-point

3-Point vs. biaxial 4-Point vs. biaxial

0.975 0.848 0.872

Sintagon Zx 3-Point vs. 4-point

3-Point vs. biaxial 4-Point vs. biaxial

0.051 0.215 0.462

The differences are not statistically significant at the 0.05 level. Table 4.4 Pairwise comparison of the material effect within the configuration effect

between Duceram and Sintagon Zx Configuration Materials p 3-Point 4-Point Biaxial

Duceram vs. Sintagon Zx Duceram vs. Sintagon Zx Duceram vs. Sintagon Zx

<0.001* <0.001* <0.001*

* The difference is statistically significant at the 0.05 level.

No significant difference was found between the test configurations within the speed groups in Table 4.5, except at 0.5 mm/min between the 3-point and 4-point tests. The 3-point test showed significant differences between all crosshead speeds in Table 4.6. Both the 4-point and the biaxial configuration produced a significant difference only between the 0.05 and the 0.5 mm/min group.

Chapter 4

74

Table 4.5 Pairwise comparison of the configuration effect within the crosshead speed effect of Sintagon Zx

Crosshead speed Configurations p 0.05 mm/min 3-Point vs. 4-point

3-Point vs. biaxial 4-Point vs. biaxial

0.958 0.263 0.297

0.1 mm/min 3-Point vs. 4-point

3-Point vs. biaxial 4-Point vs. biaxial

0.094 0.656 0.227

0.5 mm/min 3-Point vs. 4-point

3-Point vs. biaxial 4-Point vs. biaxial

0.026* 0.158 0.403

* The difference is statistically significant at the 0.05 level. Table 4.6 Pairwise comparison of the crosshead speed effect within the configuration

effect of Sintagon Zx Configuration Crosshead speeds p 3-Point 0.05 vs. 0.1 mm/min

0.05 vs. 0.5 mm/min 0.1 vs. 0.5 mm/min

0.007* 0.000* 0.017*

4-Point 0.05 vs. 0.1 mm/min

0.05 vs. 0.5 mm/min 0.1 vs. 0.5 mm/min

0.335 0.008* 0.079

Biaxial 0.05 vs. 0.1 mm/min

0.05 vs. 0.5 mm/min 0.1 vs. 0.5 mm/min

0.276 0.014* 0.165

* The difference is statistically significant at the 0.05 level.

Fracture toughness determination of two dental porcelains

75

4.5 Discussion Although in some literature the ISB method has shown good agreement with

conventional fracture mechanics tests [8, 9, 11-13], the consistency of the results with standard methods were mixed [8-10, 14, 20]. The fracture toughness, determined with the ISB method, showed dependence on the indentation load, in which a higher indentation load leads to an apparently higher KIc value [8, 16, 17]. This dependence on indentation load provides information on the R-curve behavior of dental ceramics [31]. Test configuration

In dental literature, few investigations evaluated the influence of the configuration of the test, as 3-point, 4-point and biaxial bending tests could be used in the ISB test. These test configurations may demonstrate different strength values in general because different areas and volumes are exposed to stress [27, 28]. Although the indentation introduced flaw, which fixates the point of cracking, should theoretically prevent this influence, the role of the tests configuration in the ISB results is uncertain. Furthermore, alignment errors of the load point with the indentation contribute to scatter in the data with the 3-point and the biaxial test.

Albakry et al. [20] have found a significant difference in the ISB fracture toughness between the 3-point and the biaxial tests with Empress 2 and their experimental glass–ceramic, two all-ceramic substrate materials with much higher KIc values. Nevertheless, Empress with a low toughness of about 1.3 MPa m1/2 had very close ISB KIc values in those two test configurations. In the present study, with a crosshead speed of 0.5 mm/min, all test configurations, that is 3-point, 4-point, and biaxial bending displayed consistent ISB fracture toughness values for both Duceram and Sintagon Zx, which materials had low toughness values too. This seems to suggest that with low fracture toughness porcelains and ceramics the ISB results with different test configurations are comparable, while significant differences between the configurations may occur with ceramics of higher toughness. If this is true, further investigation should be given to the role of the test configurations in ISB fracture toughness determination, since dental porcelains and ceramics have a large range of fracture toughness and a distinct diversity in chemistry and microstructure [2, 34]. More different types and varieties of dental porcelains and ceramics should be investigated in ISB KIc measurement with different strength tests and their comparability to standard methods.

Chapter 4

76

Crosshead speed In bending procedures the strength is known to increase with crosshead speed,

which is the basis for the determination of slow crack growth parameters with a dynamic test [32, 33].

In the present study It is understandable that at a lower loading speed, more time is available for the stored strain energy to produce crack growth, leading to slower crack propagation with lower stress. On the contrary, at a higher load speed, a material needs more energy to drive the growing crack [27, 32].

In dental literature, mostly crosshead speeds of 0.5 and 0.1 mm/min have been used. In the current investigation, crosshead speeds ranging from 0.5 to 0.05 mm/min were used, which covered the range of speeds often used in ISB test and it was found that a high load speed elevated the ISB fracture toughness. In other words, a high loading speed causes overestimation of KIc while it may be underestimated with a low crosshead speed. Despite the absence of many significant differences, all test configurations, especially 3-point bending test manifested obvious sensitivity to crosshead speed. Based on the present study, such a crosshead speed range may cause confusion in comparison even of values within the ISB method itself. 4.6 Conclusion

More different types of dental porcelains and ceramics with a diversity in fracture toughness, chemistry and microstructure should be involved in ISB fracture toughness determination and comparison in order to investigate the comparability of the ISB method to recognized standards when different test configurations are used.

Fracture toughness results, obtained with the ISB method are sensitive to the crosshead speed, which may lead to statistically different and perhaps confusing results. 4.7 References [1] J.R. Kelly, I. Nishimura and S.D. Campbell, Ceramic in dentistry: historical

roots and current perspectives, J Prosthet Dent 75 (1996), pp. 18–32. [2] J.R. Kelly, Dental ceramics: current thinking and trends, Dent Clin N Am 48

(2004) (viii), pp. 513–530. [3] J.J. Mecholsky, Fracture mechanics principles, Dent Mater 11 (1995), pp. 111–

112. [4] J.R. Kelly, Perspective on strength, Dent Mater 11 (1995), pp. 103–110.

Fracture toughness determination of two dental porcelains

77

[5] J.Y. Thompson, K.J. Anusavice, A. Naman and H.F. Morris, Fracture surface characterization of clinically failed all-ceramic crowns, J Dent Res 73 (1994), pp. 1824–1832.

[6] J.R. Kelly, S.D. Campbell and H.K. Bowen, Fracture-surface analysis of dental ceramics, J Prosthet Dent 62 (1989), pp. 536–541.

[7] T.L. Anderson, History and overview. In: T.L. Anderson, Editor, Fracture mechanics: fundamentals and applications, vol. 1, CRC Press, Boca Raton (1991), pp. 3–34.

[8] G.D. Quinn, The fracture toughness round robins in VAMAS: what we have learned. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, ASTM STP 1409, American Society for Testing and Materials, West Conshohocken, PA (2002).

[9] G.D. Quinn, K. Xu, R.G. Gettings, J.A. Salem and J.J. Swab, Does anyone know the real fracture toughness? SRM 2100: the world's first ceramic fracture toughness reference material. In: J.A. Salem, G.D. Quinn and M.G. Jenkins, Editors, Fracture resistance testing of monolithic and composite brittle materials, ASTM STP 1409, American Society for Testing and Materials, West Conshohocken, PA (2002).

[10] H. Wang, G. Isgrò, P. Pallav and A.J. Feilzer, Influence of test methods on fracture toughness of a dental porcelain and a soda lime glass, J Am Ceram Soc 88 (2005), pp. 2868–2873.

[11] P. Chantikul, G.R. Anstis, B.R. Lawn and D.B. Marshall, A critical evaluation of indentation techniques for measuring toughness: II, strength method, J Am Ceram Soc 64 (1981), pp. 539–543.

[12] G.D. Quinn, J. Salem, I. Bar-On, K. Cho, M. Foley and H. Fang, Fracture toughness of advanced ceramics at room temperature, J Res Natl Inst Stand Technol 97 (1992), pp. 579–607.

[13] Primas RJ, Gstrein R. ESIS TC 6 round robin on fracture toughness. EMPA report No. 155,088, 3-41. Dübendorf, Switzerland: EMPA; 1995.

[14] K. Kvam, H. Herö and G. Øilo, Fracture toughness measurements of some dental core ceramics: a methodologic study, Scand J Dent Res 99 (1991), pp. 527–532.

[15] K.J. Anusavice, N.Z. Zhang and J.E. Moorhead, Influence of colorants on crystallization and mechanical properties of lithia-based glass–ceramics, Dent Mater 10 (1994), pp. 141–146.

Chapter 4

78

[16] S.S. Scherrer, I.L. Denry and H.W. Wiskott, Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic, Dent Mater 14 (1998), pp. 246–255.

[17] S.S. Scherrer, J.R. Kelly, G.D. Quinn and K. Xu, Fracture toughness (KIc) of a dental porcelain determined by fractographic analysis, Dent Mater 15 (1999), pp. 342–348.

[18] M. Guazzato, M. Albakry, M.V. Swain and J. Ironside, Mechanical properties of in-ceram alumina and in-ceram zirconia, Int J Prosthodont 15 (2002), pp. 339–346.

[19] W.S. Oh, J.M. Park and K. Anusavice, Fracture toughness (KIC) of a hot-pressed core ceramic based on fractographic analysis of fractured ceramic FPDs, Int J Prosthodont 16 (2003), pp. 135–140.

[20] M. Albakry, M. Guazzato and M.V. Swain, Fracture toughness and hardness evaluation of three pressable all-ceramic dental materials, J Dent 31 (2003), pp. 181–188.

[21] W.S. Oh, N.Z. Zhang and K.J. Anusavice, Effect of nucleation temperature on fracture toughness (KIC) of fluorcanasite-based glass–ceramic, Int J Prosthodont 16 (2003), pp. 505–509.

[22] M. Guazzato, M. Albakry, S.P. Ringer and M.V. Swain, Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part I. Pressable and alumina glass-infiltrated ceramics, Dent Mater 20 (2004), pp. 441–448.

[23] M. Guazzato, M. Albakry, S.P. Ringer and M.V. Swain, Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part II. Zirconia-based dental ceramics, Dent Mater 20 (2004), pp. 449–456.

[24] I.L. Denry and J.A. Holloway, Effect of post-processing heat treatment on the fracture strength of a heat-pressed dental ceramic, J Biomed Mater Res B Appl Biomater 68 (2004), pp. 174–179.

[25] I.L. Denry and J.A. Holloway, Effect of heat pressing on the mechanical properties of a mica-based glass–ceramic, J Biomed Mater Res B Appl Biomater 70 (2004), pp. 37–42.

[26] M. Albakry, M. Guazzato and M.V. Swain, Influence of hot pressing on the microstructure and fracture toughness of two pressable dental glass–ceramics, J Biomed Mater Res B Appl Biomater 71 (2004), pp. 99–107.

[27] G.D. Quinn and R. Morrell, Design data for engineering ceramics: a review of the flexure test, J Am Ceram Soc 74 (1991), pp. 2037–2066.

Fracture toughness determination of two dental porcelains

79

[28] J. Jin, H. Takahashi and N. Iwasaki, Effect of test method on flexural strength of recent dental ceramics, Dent Mater J 23 (2004), pp. 490–496.

[29] G. With de and H.H.M. Wagemans, Ball-on-ring test revisited, J Am Ceram Soc 72 (1989), pp. 1538–1541.

[30] ASTM Designation: F394-78, American standard test method for biaxial flexure strength (modulus of rupture) of ceramic substrates, Annual book of ASTM standards, American Society for Testing and Materials, Philadelphia, PA (1978) p. 434–40 [re-approved 1984].

[31] H. Fischer, W. Rentzsch and R. Marx, R-curve behavior of dental ceramic materials, J Dent Res 81 (2002), pp. 547–551.

[32] C.W. Fairhurst, P.E. Lockwood, R.D. Ringle and S.W. Twiggs, Dynamic fatigue of feldspathic porcelain, Dent Mater 9 (1993), pp. 269–273.

[33] G.A. Thompson, Determining the slow crack growth parameter and weibull two-parameter estimates of bilaminate disks by constant displacement-rate flexural testing, Dent Mater 20 (2004), pp. 51–62.

[34] J.B. Quinn, V. Sundar and I.K. Lloyd, Influence of microstructure and chemistry on the fracture toughness of dental ceramics, Dent Mater 19 (2003), pp. 603–611.

Chapter 4

80

81

CHAPTER 5

Indentation-strength fracture toughness: the role of the indentation load at smaller flaw sizes

Keywords: Dental porcelains, Fracture toughness, Indentation strength, flaw size

Chapter 5

82

5.1 Abstract Objective: The aim of the study was to investigate the influence of small indentation loads on ISB fracture toughness determination of dental veer ceramics. Materials and Methods: Duceram Plus and Sintagon Zx were tested in 3-point, 4-point, and biaxial bending while a soda lime glass was tested in 4-point bending as a reference. Duceram Plus was indented at 9.8, and 19.6 N, Sintagon Zx at 4.9, 9.8, and 19.6 N, and glass at 1.96, 2.94, 4.9, 9.8, and 19.6 N. Unindented specimens were tested, too. The specimens which fractured at a flaw other than the precrack were rejected. Results: All toughness values increased significantly with indentation load. Duceram Plus was significantly tougher than Sintagon Zx and glass. No statistical differences were found between the setups, except with Duceram Plus in the 4-point setup, which produced lower toughness values. At a fixed indentation load, different materials are compared at different flaw sizes, where tougher materials have smaller indentation flaws and more rejections. At small flaw sizes, rejected ISB specimens are often stronger than average. Conclusions: Testing the unindented strength of identical specimens to verify that the indented strength is sufficiently lower is a safer requirement for a valid experiment than (only) setting a maximum to the number of rejected specimens.

Indentation-strength fracture toughness

83

5.2 Introduction Fracture phenomena such as chipping, delamination, or complete failure, are a

major reason for failure of all-ceramic dental restorations [1-5]. Cores of strong ceramics like zirconia increasingly replace metal cores underneath veneer porcelains and new phenomena occur. For example, metal alloy core materials are perhaps half as strong as good quality zirconia but depending on restoration design, type of loading, and other circumstances, cracks may initiate in the much weaker veneer, propagate into the ceramic core and cause complete fracture [4, 5]. Perhaps not surprisingly, a good quality, well-fused bond between the core and the veneer increases this problem. The strength is a rather variable property of these brittle materials, which is determined by two other properties; the presence of irregularities, flaws, roughness, scratches, etc. that lead to stress concentrations and two, the fracture toughness or KIc, which is a material constant. Flaws may be related to a composite structure and as such be a material constant, but surface roughness and scratches generally are related to the finishing procedure or service life [6, 7]. The toughness, i.e. the resistance to crack extension is not easily assessable and many techniques have been developed to measure the fracture toughness of brittle materials. A few methods, such as the Single-Edge-Pre-cracked-Beam (SEPB), the Chevron-Notched beam or rod (CN), or the Surface-crack-by-Flexure (SCF) produce consistent and reproducible KIc values and have been accepted by standardization organizations such as the ASTM or the ISO [8-11]. Many other methods, the Indentation-strength in Bending (ISB), Indentation Fracture (IF), etc, are more convenient or easier to perform, but lead to a variety of toughness values depending on many factors [10-24]. The toughness of dental porcelains and ceramics has been evaluated with various test methods [14-25]. The Indentation-Strength-in-Bending (ISB) method, which is often used because of its convenience, seems to be an appropriate test to evaluate the fracture toughness of dental ceramics. However, differences in micro-structural properties of the tested materials [17, 18, 23], the crosshead speed [24], as well as the indentation load [15, 16] have been shown to clutter the resulting ISB-toughness in a complex way. With this method the tensile surfaces are polished and precracked with a Vickers’ indenter. The radial cracks at the indentation are considered a reasonable reproduction of typical surface damage [12]. After the tests, the fracture surfaces are examined to verify that failure occurred from the indentation and the toughness is obtained with a formula, which uses the failure stress, the precrack force, the Vickers’ hardness, and the elastic modulus. The hardness may be measured on broken specimens the elastic

Chapter 5

84

modulus requires extra experiments with unindented specimens. A minimum precrack load generally is required to prevent a large amount of rejected specimens because a smaller precrack increases the chance that another, already present flaw is effectively larger and that fracture occurs there and not at the precrack. This aim of this study was to investigate the influence of small indentation loads on ISB experiments with relatively weak dental veneer ceramics in the 3-point, 4-point, and biaxial bending tests, respectively. 5.3 Materials and Methods

Apart from a soda lime glass sheet material, which was used as a reference, the dentin powders of two veneer ceramic systems were tested. These silicates or glass ceramics are intended to increase the strength of the final structure and provide opacity to mask the core. Leucite particles with a high thermal contraction coefficient are added to match the thermal contraction of the core material and to increase the toughness. Duceram Plus (Degussa Dental GmbH, Hanau, Germany, LOT: 41598) is used with porcelain-fused-to-metal crowns, Sintagon Zx (Elephant Dental B.V., Hoorn, The Netherlands, LOT: A-1811) is intended for zirconia all-ceramic systems. Specimen preparation

The powders with their mixing liquids were condensed in brass moulds of 30 x 50 x 4 mm3 for the 3- and 4-point tests and of 24 mm diameter x 2.5 mm thickness for the ball-on-ring tests. The blocks were removed from the moulds and fired in a dental oven (STRATOS, Elephant Dental B.V., Hoorn, The Netherlands) according to the manufacturers’ instructions. The cooling time was extended to 30 min to avoid thermal stresses. The upper and lower surfaces of the fired blocks were ground parallel with 30 μm diamond paste in a grinding device (VEM Metallurgy, Vos & Van Eijk Metallurgie B.V., Houten, the Netherlands). The rectangular blocks were reduced to a thickness of 3.0 mm and the round discs to 1.9 mm. The blocks were cut into bars (2.0 height x 3.0 width x 25.0 mm3) with a 0.5 mm diamond saw (ISOMET Q2 1000, Buehler Ltd, Lake Bluff, IL). The glass beams (2.0 height x 3.9 width x 25mm3) were cut from 3.9 mm sheet material. The tensile surface of all bars and discs were polished with wet silicon carbide papers of 400, 600, and 1200 grit successively, and the sharp edges of the rectangular bars were chamfered. The specimens of glass and Sintagon Zx were annealed at 450°C and the Duceram Plus specimens at 600°C during 90 min and cooled to 100°C in another 90 min. The roughness, i.e. the Ra value was measured in three places on six random

Indentation-strength fracture toughness

85

beams of two series of each material (SZ 700 profilometer, Mitutoyo Corp., Kanagawa, Japan). Strength testing

Vickers’ indentations were made (HM-124 Hardness Testing Machine, Mitutoyo Corp., Kanagawa, Japan) in the middle of the tensile surfaces. Unindented specimens of each material were tested as well. Duceram Plus specimens were indented at 9.8 and 19.6 N, Sintagon Zx specimens at 4.9, 9.8, and 19.6 N, and the glass beams at loads of 1.96, 2.94, 4.9, 9.8, and 19.6 N. The radial cracks underneath the Vickers’ indentation serve as the pre-crack in this test. Because of the residual stresses around the indentation the pre-cracks continue to grow during the first few minutes following the indentation. The specimens were loaded after 20 to 30 min at 0.05mm/min until fracture occurred in a tensilometer (ACTA Intense, ACTA, Amsterdam, NL). Specimens, where the fracture did not originate at the Vickers’ indentation were excluded from the toughness but not from the strength assessments and testing was continued until at least 10 acceptable toughness results were obtained. Setup configurations

The two dental veneers were tested in three configurations: 3- and 4-point bending of bars, and biaxial bending of discs in a ball-on-ring setup. The glass beams were tested in 4-point bending. In the 3- and 4-point bending tests the strength was calculated according to [26]:

221

2)(3

BDLLW

f−

=σ

Where σf is the strength, W is the fracture load, L1 the support span (20.0 mm), L2 the loading span (10.0 mm for 4-point bending, 0 mm for 3-point bending), B the specimen width, and D the specimen thickness. Strength with the ball-on-ring test was calculated with the following equation:

2

22

2

4

]11))(

21()(21)[1(3

TDD

Db

bD

LnW s

S

s

f πννν

σ+−

−+++=

Where ν is the Poisson’s ratio, Ds the diameter of the support ring (16.0 mm), D the diameter of disc (21 mm), T the thickness, and b = 2T/3 is the diameter of the area with

Chapter 5

86

a uniform load. A value of 0.25 was used for Poisson’s ratio, as recommended by ISO 6872 [27, 28]. The diameter of the loading ball is 5 mm. The support ring is a circle of 16 cylindrical 2.5 mm pins with rounded tops. The pins are supported as pistons entering a shared and sealed compartment filled with soft (Shore 10) silicone rubber, which works as a non-leaking hydraulic fluid. Discs, which are not absolutely flat, for example when surfaces are left as fired, may also be tested this way. The elastic modulus was determined with the 3-point bending setup on ten bars without indentation but at a crosshead speed of 0.5 mm/min.

qBDWLE 3

3

4=

Where E is the elastic modulus, L the support span (20.0 mm), and q the bending deflection. The Vickers’ hardness, HV, was measured on broken specimens (n = 10) using an indentation load, P, of 1.96 N, with H = 1.854P/l2, where l is the average of the two diagonals of the indentation. The ISB facture toughness was obtained with the ISB equation [12]:

( ) ( ) 4/33/18/1/ PHEK fIc ση=

Where η is the geometrical constant (0.596), E the elastic modulus, H the Vickers’ hardness, and P the precrack load. The geometrical constant is slightly greater than the 0.59 used by Chantikul et al., because they used 2 instead of 1.854 in the Vickers’ equation.

Three-way analysis of variance (ANOVA) and SNK pair wise comparison was carried out to assess the individual and interaction effect of materials, strength test configurations, and indentation loads on the fracture toughness, at a significance level of 0.05. The software used was SPSS 10.0 (SPSS inc., Chicago, USA). 5.4 Results

Figure 5.1 shows the toughness and Figure 5.2 the strength values, which are listed numerically in Table 5.1 The table also lists the numbers of bars and strength of the bars, which were rejected for the calculation of the ISB toughness. Note that the average values in the stress column do include the values of the rejected bars. In all cases the reason for rejection was fracture at another flaw than the indentation.

Indentation-strength fracture toughness

87

Table 5.1 Average ISB toughness and strength values, numbers of specimens tested and rejected and strength of each rejected specimen. The standard deviations are in parentheses. The elastic modules, E, and Vickers’ hardness’s, HV, are in GPa. Not significant differences as connected with index letters with 0N indentation load series are based on strength values, all others on toughness values.

Indentation load ISB values Specimens tested and

Configuration rejected

Material N Toughness MPa·m½

Strength MPa

Strength of rejected bars – MPa

0 - a,b 46.9(2.0) 8

9.8 0.825(0.032)a 48.6(2.6) 17/6 52.1(4.4) Duceram Plus

4-p

19.6 0.923(0.067)b 44.9(4.3) 14/3 44.2/48.6/52.7

0 - 60.3(1.7) 8

9.8 0.891(0.037) 53.8(3.0) 12/2 52.5/57.4 E =61.6(1.4)

3-p

19.6 1.004(0.066) 50.2(4.4) 11/0

0 - c 56.5(3.6) 8 HV =5.33(0.16) 9.8 0.909(0.061)c 55.4(5.0) 15/4 59.2/61.5/64.0/71.4

biax

19.6 1.010(0.047) 50.5(3.1) 12/0

0 - 52.6(7.4) 8

4.9 0.667(0.073) 44.4(6.6) 13/1 45.8

9.8 0.717(0.060) 38.9(4.4) 11/1 47.8 Sintagon Zx

4-p

19.6 0.761(0.047) 33.4(2.8) 11/1 50.3

0 - 62.6(8.8) 8

4.9 0.675(0.044) 45.1(4.0) 12/0 E =64.5(1.6)

9.8 0.725(0.039)d 39.4(2.7) 12/2 25.2/63.0 3-p

19.6 0.760(0.069)d 33.3(4.0) 11/0 HV =4.43(0.08) 0 - 46.7(3.1) 8

4.9 0.652(0.060) 43.1(5.2) 11/1 60.7

9.8 0.716(0.025) 38.7(1.8) 10/0

biax

19.6 0.787(0.042) 34.9(2.5) 10/0

Glass 0 - 96.4(11.6) 8

1.96 0.645(0.044)e 61.6(6.5) 12/3 59.3/60.2/75.5 E=58.6(3.5)

2.94 0.652(0.024)e 57.2(10.2) 12/2 64.0/87.3

4.9 0.674(0.019)e,f 47.2(1.7) 12/0 HV =5.37(0.05) 9.8 0.708(0.027)f 40.0(2.0) 12/0

4-p

19.6 0.778(0.042) 36.0(2.6) 12/0

Chapter 5

88

0

0.2

0.4

0.6

0.8

1

1.2

0 5 10 15 20

indentation load (N)

toughness (MPa.m½)

Duceram Plus 4-pointSintagon Zx 3-pointGlass biaxial

Figure 5.1 ISB toughness values of all series

0

10

20

30

40

50

60

70

80

90

100

0 5 10 15 20

indentation load (N)

strength (MPa)

Duceram Plus 4-pointSintagon Zx 3-pointGlass biaxial

Figure 5.2 ISB indented strength of all series

Indentation-strength fracture toughness

89

The average Ra values of all materials are similar, 0.05 to 0.06 µm. The toughness values of Duceram Plus at 9.8 N precrack load in the 4-point and in the biaxial configuration are not acceptable, because of the large number of rejected specimens.

All ISB-KIc values increase with indentation load. In all cases the differences between the lowest (≠ 0) and highest precrack loads are significant, but the differences between the successive steps of indentation load often are not. In all test configurations and indentation loads the ISB toughness values of Duceram Plus are statistically higher than those of Sintagon Zx and the glass. The differences between the toughness values of glass and Sintagon Zx are not significant. The differences between the setup configurations are not significant, except the 4-point setup with Duceram Plus, which produced lower KIc values with both indentation loads. The failure stresses of all indented specimens were significantly lower than the unindented ones with the exceptions of Duceram Plus in the 4-point setup and at 9.8 N in the biaxial setup, where the differences were not significant. A total of 26 rejected specimens are listed. 20 of these, or about three out of four, are stronger than the average of the series in the stress column. On average of all 26, the failure stresses of the rejected specimens are about 14% greater than the average of each one’s series.

The results of the unindented specimens are rather different. Glass is considerably stronger than the dentin veneers, between which the difference varies with the setup, although Sintagon Zx is significantly stronger than Duceram Plus in the 4-point setup. All differences between the setup configurations are significant, except between the 3-point and the biaxial tests with Duceram Plus and between the three bending tests with Sintagon Zx. 5.5 Discussion The results confirm [15, 16] a structural positive influence of the indentation load on the ISB toughness values of the tested materials, which is not caused by a rising R-curve [29]. The increases as shown in Figure 5.1, must be attributed to the ISB method, because especially glass has a flat R-curve. The influence of the precrack load on the apparent toughness seems reasonably linear. In previous, Chevron Notch (CN) experiments [14, 25] the toughness values of glass and Sintagon Zx were the same as well. The value of 0.70MPa·m½ in those CN experiments suggests that these materials should be tested at a precrack load just less than 9.8 N in order to obtain correct ISB values (Figure 5.1). The Sintagon Zx was of

Chapter 5

90

the same batch number as the material in the present experiments, the glass came with the same delivery and should be identical too. Materials The unindented strengths of the veneers are so much less than that of the glass, because their composite structure gives rise to stress concentrations similar to those at other flaws. The leucite particles, which are added for this purpose, have a much greater thermal contraction coefficient than the glassy matrix. This generates residual thermal stresses after annealing and sometimes cracks in the matrix around the particles [30], which work as a minimum flaw size and are known to decrease the strength at smaller flaw sizes including the unindented strength but also to increase the toughness [31]. Moreover some porosity is inevitable with the powder-liquid technique [32,33]. The pores found in Duceram Plus were larger and more numerous than in Sintagon Zx. Sintagon Zx seems to contain just enough leucite to reduce the unindented strength to a level similar to that of Duceram Plus, but not enough to make it any tougher than glass. Although under different conditions the veneers may be more difficult to polish than glass also because of their structure, grinding the specimens with wet silicon carbide paper up to grit 1200 seems to produce a fairly constant roughness with an average Ra value of 0.05 to 0.06 µm for all materials tested. The difference in unindented strength between glass and the veneers therefore cannot be attributed to the roughness, but largely is related to irregularities, such as pores or leucite particles, present near or cut at the polished tensile surfaces. Precrack size The strength plot (Figure 5.2) repeats that strength is not a physical constant of these materials. This figure also gives some idea of the damage tolerance of the materials, although at the same precrack load the resulting flaw size is actually smaller with the tougher material. At a fixed indentation load different toughness materials are ranked at different flaw sizes, which could mean that extra test are necessary to establish an appropriate precrack load for each material when the flaw sizes should be similar. Single materials on the other hand may be tested at a range of precrack loads or flaw sizes. The results of Duceram Plus and glass confirm that more rejections occur at smaller flaw sizes. A fixed precrack load amplifies variations in the toughness of the specimens, as a tougher specimen is stronger, not only because it is tougher but also because the flaw

Indentation-strength fracture toughness

91

is smaller. This increases the differences between materials as well as the data scatter within the series, but these phenomena are compensated when the strength values are converted to toughness values with the ISB equation. At smaller sizes of the precrack, this mechanism increases the chance that rejected specimens are stronger than the average of the series, because the smaller precracks in tougher specimens increase the chance of failure at another flaw. Moreover the stress required for fracture at the precrack should be greater than the failure stress of these specimens. Rejecting these specimens of the toughness assessment therefore should decrease the resulting average more than leaving these in. Unindented strength The results of Duceram Plus at 9.8N, with a large amount of rejected specimens in all setups, illustrate that before or without the indentation the specimens should be sufficiently stronger (Figure 5.2) than is required to indentedly fracture at the strength for the appropriate toughness value, as the numbers of rejected specimens, 2, 4, and 6, seems to agree with the ratios of the indented strength to the unindented strength, 89%, 98%, and 104%, in the 3-point, biaxial, and 4-point setup respectively in Table 1. This looks consistent with the much smaller number of rejections with Sintagon Zx at 4.9N where the strength of the indented specimens is 72% (3-point), 84% (4-point), and 92% (biaxial) of the strength of the unindented specimens. In Figure 5.2 the strength values of Duceram Plus in the 4-point setup show that indentations simply do not increase the strength of the specimens, which prevents the 4-point setup to produce toughness values in Figure 5.1 similar to those of the 3-point and biaxial tests. The Duceram Plus 19.6 N 4-point toughness value would just be acceptable, considering the three rejected specimens, but the comparison with the unindented strength in Figure 5.2 shows that it clearly isn’t acceptable, because the precrack load is much to low. It is not clear however why a majority of the specimens still fractured acceptably at the indentation. Apparently, the precrack load should be high enough to decrease the strength to a value, which is less than 80 to 90% of the unindented strength, for an acceptable test and not only to prevent rejections due to fracture at another flaw [12]. Setup configurations The number of bars that have to be rejected because fracture occurs at another flaw than the precrack is greater in the 4-point setup because a greater part of the bars is loaded at the maximum stress than in the 3-point setup.

Chapter 5

92

For the same reason, the strength of the unindented specimens is greater in the 3-point setup. It is not very clear how much area is effectively loaded to the maximum stress when the 3-point and the biaxial setup are compared [34, 35]. In the biaxial tests a 1.3 mm circle with a diameter of two thirds of the 1.9 mm thickness of the disc is loaded at the maximum stress. In the 3-point setup this is only a very thin 3mm line opposing the middle roller, but for example 1 mm away from this line the stress is still 90%. Moreover, materials, flaws, and indentations do not necessarily respond similarly to uni- or biaxial stresses. At biaxial stress the crack tip is exposed to perpendicular stress too and with the Vickers’ indentation both radial cracks are loaded, which yields an extra chance on fracture. Fracture from internal flaws, which may occur with unindented specimens should also be sensitive to the rate at which the stress decays towards the neutral line, which then becomes equivalent to the volume of the subsurface part loaded at the maximum stress, which is proportional to the thickness. With both veneers the unindented strength is greater in the 3-point tests. With Duceram Plus the unindented strength in the biaxial setup is not much less than in the 3-point setup, but with Sintagon Zx it is even less than in the 4-point test 5.6 Conclusion The present study supports the following conclusions. • ISB toughness values of glass ceramics increase with the used indentation load. • At a fixed indentation load, different materials are compared at different flaw

sizes. • At small indentation loads, rejected ISB specimens, which fracture at another

flaw than the precrack, often are stronger than specimens, which fracture acceptably.

• Accepting rather than rejecting one or two such specimens might be more accurate.

• Testing the unindented strength of identical specimens to verify that the indented strength is sufficiently lower is a safer requirement for a valid experiment than setting a maximum to the number of rejected specimens.

5.7 References [1] J.R. Kelly, Dental ceramics: current thinking and trends. Dental clinics of North

America 48:viii (2004), pp. 513-530.

Indentation-strength fracture toughness

93

[2] J.J. Mecholsky, Fracture mechanics principles. Dental Materials 11 (1995), pp. 111–112.

[3] J.R. Kelly, Perspective on strength. Dental Materials 11 (1995), pp. 103–110. [4] J.Y. Thompson, K.J. Anusavice, A. Naman and H.F. Morris, Fracture surface

characterization of clinically failed all-ceramic crowns. Journal of Dental Research 73 (1994), pp. 1824–1832.

[5] J.R. Kelly, S.D. Campbell and H.K. Bowen, Fracture-surface analysis of dental ceramics. Journal of Prosthetic Dentistry 62 (1989), pp. 536–541.

[6] W.S. Oh, R. Delong and K.J. Anusavice, Factors affecting enamel and ceramic wear: a literature review. Journal of Prosthetic Dentistry 87 (2002), pp. 451-9.

[7] T.L. Anderson, Fracture testing in nonmetals. In: Anderson TL, editor. Fracture mechanics: fundamentals and applications, Vol 1. Boca Raton: CRC Press; 1991. p. 485-530.

[8] G.D. Quinn, J. Salem, I. Bar-On, K. Cho, M. Foley and H. Fang, Fracture toughness of advanced ceramics at room temperature. Journal of Research of the National Institute of Standards and Technology 97 (1992), pp. 579–607.

[9] R.J. Primas and R. Gstrein, ESIS TC 6 Round Robin on fracture toughness. EMPA report No. 155’088, 3–41, EMPA, Dubendorf, Switzerland, 1995.

[10] Quinn GD. “The fracture toughness round robins in VAMAS: what we have learned”. Fracture Resistance Testing of Monolithic and Composite Brittle Materials. ASTM STP 1409, JA Salem, GD Quinn, and MG Jenkins, Eds., American Society for Testing and Materials, West Conshohocken, PA, 2002.

[11] G.D. Quinn, K. Xu, R.G. Gettings, J.A. Salem and J.J. Swab, “Does anyone know the real fracture toughness? SRM 2100: the world’s first ceramic fracture toughness reference material”, Fracture Resistance Testing of Monolithic and Composite Brittle Materials. ASTM STP 1409, JA Salem, GD Quinn, and MG Jenkins, Eds., American Society for Testing and Materials, West Conshohocken, PA, 2002.

[12] P. Chantikul, G.R. Anstis, B.R. Lawn and D.B. Marshall, A critical evaluation of indentation techniques for measuring toughness: II, strength method. Journal of American Ceramic Society 64 (1981), pp. 539–543.

[13] H. Fischer and R. Marx, Fracture toughness of dental ceramics: comparison of bending and indentation method. Dental Materials 18 (2002), pp. 12-9.

[14] H. Wang, G. Isgro, P. Pallav and A.J. Feilzer, Influence of Test Methods on Fracture Toughness of a Dental Porcelain and a Soda Lime Glass. Journal of American Ceramic Society 88 (2005), pp. 2868–2873.

Chapter 5

94

[15] S.S. Scherrer, I.L. Denry and H.W. Wiskott, Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic. Dental Materials 14 (1998), pp. 246-255.

[16] S.S. Scherrer, J.R. Kelly, G.D. Quinn and K. Xu, Fracture toughness (KIc) of a dental porcelain determined by fractographic analysis. Dental Materials 15 (1999), pp. 342–348.

[17] M. Guazzato, M. Albakry, M.V. Swain and J. Ironside, Mechanical properties of In-Ceram Alumina and In-Ceram Zirconia. International Journal of Prosthodontics 15 (2002), pp. 339-346.

[18] M. Albakry, M. Guazzato and M.V. Swain, Fracture toughness and hardness evaluation of three pressable all-ceramic dental materials. Journal of Dentistry 31 (2003), pp. 181–188.

[19] M. Guazzato, M. Albakry, S.P. Ringer and M.V. Swain, Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part I. Pressable and alumina glass-infiltrated ceramics. Dental Materials 20 (2004), pp. 441-448.

[20] M. Guazzato, M. Albakry, S.P. Ringer and M.V. Swain, Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part II. Zirconia-based dental ceramics. Dental Materials 20 (2004), pp. 449-456.

[21] I.L. Denry and J.A. Holloway, Effect of post-processing heat treatment on the fracture strength of a heat-pressed dental ceramic. Journal of Biomedical Materials Research Part B: Applied Biomaterials 68 (2004), pp. 174-179.

[22] I.L. Denry and J.A. Holloway, Effect of heat pressing on the mechanical properties of a mica-based glass-ceramic. Journal of Biomedical Materials Research Part B: Applied Biomaterials 70 (2004), pp. 37-42.

[23] M. Albakry, M. Guazzato and M.V. Swain, Influence of hot pressing on the microstructure and fracture toughness of two pressable dental glass-ceramics. Journal of Biomedical Materials Research Part B: Applied Biomaterials 71 (2004), pp. 99-107.

[24] H. Wang, G. Isgro, P. Pallav and A.J. Feilzer, Fracture toughness determination of two dental porcelains with the indentation strength in bending method. Dental Materials 23 2007, pp. 755-759.

[25] H. Wang, P. Pallav, G. Isgro and A.J. Feilzer, Fracture toughness comparison of three test methods with four dental porcelains. Dental Materials 23 2007, pp. 905-910.

[26] G.D. Quinn and R. Morrell, Design Data for Engineering Ceramics: A Review of

Indentation-strength fracture toughness

95

The Flexure Test. Journal of American Ceramic Society 74 1991, pp. 2037–2066.

[27] G. de With and H.H.M Wagemans, Ball-on-ring test revisited. Journal of American Ceramic Society 72 1989, pp. 1538–1541.

[28] ASTM Designation: F394-78. American Standard Test Method for Biaxial Flexure Strength (Modulus of Rupture) of Ceramic Substrates. In: Annual Book of ASTM Standards. Philadelphia, PA: American Society for Testing and Materials (re-approved 1984), 1978. p. 434–440.

[29] D. Munz, What Can We Learn from R-Curve Measurements? Journal of American Ceramic Society 90 (2007), pp. 1–15.

[30] M.J. Cattell, T.C. Chadwick, J.C. Knowles, R.L. Clarke and E. Lynch, Flexural strength optimisation of a leucite reinforced glass ceramic. Dental Materials 17 (2001), pp. 21-33.

[31] P.F. Cesar, H.N. Yoshimura, W.G. Miranda Junior and C.Y. Okada, Correlation between fracture toughness and leucite content in dental porcelains. Journal of Dentistry 33 (2005), pp. 721-9.

[32] Y. Zhang, J.A. Griggs and A.W. Benham, Influence of powder/liquid mixing ratio on porosity and translucency of dental porcelains. Journal of Prosthetic Dentistry 91 (2004), pp. 128-35.

[33] K.C. Cheung and B.W. Darvell, Sintering of dental porcelain: effect of time and temperature on appearance and porosity. Dental Materials 18 (2002), pp. 163-73.

[34] K. Zeng, A. Oden and D. Rowcliffe, Flexure tests on dental ceramics. International Journal of Prosthodontics 9 (1996), pp. 434-9.

[35] H. Takahashi and N. Iwasaki, Effect of test method on flexural strength of recent dental ceramics. Dental Materials Journal 23 (2004), pp. 490-496.

Chapter 5

96

97

CHAPTER 6

Strength influencing variables on CAD/CAM zirconia frameworks

Keywords: Zirconia; Particle abrasion; Surface damage; CAD/CAM

Chapter 6

98

6.1 Abstract Objective: any studies in the dental literature look at the effect of different surface treatment methods on the flexure strength of zirconia where polished zirconia has been used as control. However, zirconia is subjected to different types of surface damage as a result of the CAD/CAM milling procedure and also to damage produced by other laboratory procedures in use daily. The aim of this work was to evaluate the effect of different surface treatment methods and in particular the effect of the CAD/CAM milling procedure on the flexure strength of zirconia frameworks. Materials and methods: At least 20 zirconia bars (17 mm × 2 mm × 1 mm) for each group were prepared by either cutting and polishing zirconia milling blocks or by using a CAD/CAM device (Cercon) which left behind characteristic surface features related to the milling process. The fully sintered bars received either of the following surface treatments: air-borne particle abrasion (with 50 and 120 μm aluminum oxide particles, or both). Some bars received a heat treatment commonly used in baking veneer ceramics before or after particle abrasion. The surface roughness was measured for all bars, which were finally loaded in a three-point device. The fractured bars were examined using scanning electron microscopy. Data were analyzed using one-way analysis of variance and survivability was estimated using Weibull analysis (α < 0.05). Results: There were significant differences in the flexure strength (in MPa) between the tested groups subjected to different surface treatments which can be categorized into four strength levels: (1074–1166 MPa) for polished zirconia and the CAD/CAM bars that were particle abraded (50 μm Al2O3) whether with or without heat treatment (936 MPa) for the ground bars that were particle abraded (50 μm Al2O3), (708–794 MPa) for CAD/CAM bars and for the polished bars that were particle abraded (120 μm Al2O3), and (546 MPa) for the ground bars that were particle abraded (120 μm Al2O3) being the weakest. There was a strong correlation between flexure strength and the severity of surface damage as indicated by surface roughness (R2 = 0.912). Scanning electron microscopy revealed different types of surface and subsurface damage produced by the different surface treatments. Conclusions: The surface damage produced by the CAD/CAM milling procedure significantly reduced the strength of zirconia which could be further weakened by different surface treatment methods resulting in unexpected failures at stresses much lower than the ideal strength of the material. It is advised to consider the effect of the CAD/CAM procedure on the characteristic strength when designing zirconia-based fixed partial dentures.

Strength influencing on CAD/CAM zirconia

99

6.2 Introduction The introduction of zirconia to the dental field opened up the design and

application limits of all-ceramic restorations. The superior mechanical properties of zirconia combined with the state-of-the-art CAD/CAM fabrication procedure allowed for the production of large and complex restorations with high accuracy and success rate [1].

The strength of zirconia can be directly influenced by different surface treatment methods which exert different degrees and types of surface damage. These areas of surface flaws act as stress concentration sites and even though they are microscopic in nature, they act as potential sites for crack initiation and propagation [2]. Dental literature has focused on studying the effect of different surface treatments on the strength of zirconia-based materials and reduction in strength was generally associated with the degree of surface damage [3-5]. On the other hand, some studies reported an increase in strength after air-borne particle abrasion and related such finding to the creation of compressive fields as a result of the induced tetragonal–monoclinic transformation of the surface crystals [6-9].

A point worth noting is that in some of the previous studies polished zirconia was used as a reference point, while grinding with different grits of silicon carbide paper and air-borne particle abrasion at a high pressure were commonly used as examples of different levels of surface damage. On the other hand, under daily circumstances, zirconia is subjected to a different type of surface damage as a result of the milling procedure, which leaves behind characteristic trace lines and different patterns of surface damage and flaws [5]. Additionally, the combined effect of the CAM milling procedure and common laboratory procedures such as air-borne particle abrasion and multiple firing cycles used in baking the ceramic veneer, could be different from that expected for polished or disc-ground zirconia [10]. The mechanical properties and the long-term stability of CAD/CAM zirconia will be a function of the exerted surface damage, the degree of transformation, and the loading environment in terms of peak stresses and number of cycles [11].

The aim of this study was to evaluate the damage induced by the CAD/CAM milling procedure, combined with different surface finishing procedures, on the mechanical properties of zirconia and to analyze the interaction between these variables.

Chapter 6

100

6.3 Materials and Methods At least 20 zirconia bars (17 mm × 2 mm × 1 mm) for each group where

prepared by either of the following methods: cutting zirconia milling blocks in a sawing machine (ISOMET Q2 1000, Buehler Ltd., Lake Bluff, IL) using a diamond coated disc saw (ground bars) or by using CAD/CAM technology (Cercon, Degudent GmbH, Hanau-Wolfgang, Germany) where individual bars were milled by the machine using wax replicas (CAM bars). The bars were sintered in the relevant manufacturer equipment (Cercon Heat uses a 6.5 h firing cycle at a maximum temperature of 1350 °C). Some of the ground bars were polished using ascending grit silicon carbide paper (ECOMET Grinder/Polisher, Buehler Ltd., Evanston, IL) up to 1200 grit. The ground or CAM bars received either of the following surface treatments: 1. Air-borne particle abrasion using either 50 or 120 μm aluminum oxide particles at

0.35 MPa pressure for 25 s/cm2 at a distance of 2.0 cm. These bars were used to investigate the direct effect of air-borne particle abrasion on the mechanical properties of zirconia.

2. Some of the bars treated as above were subjected to one firing cycle used to bake the preformed ceramic veneer (1 min hold time in vacuum at a maximum temperature of 910 °C), either before or after air-borne particle abrasion (Austromat 3001, Dekema Dental-Keramiköfen GmbH & Co, Germany). These bars were used to investigate the effect of temperature on relieving surface pre-stresses caused by particle abrasion.

The surface roughness of the prepared bars was measured using a contact sensor (SJ-400, Mitutoyo Corporation, Japan) after which the bars were loaded in a three-point flexure device (15 mm span) at a cross-head speed of 0.5 mm/min in a universal testing machine (Instron 6022, Instron Limited, High Wycombe, UK). The load to failure was extracted from a computer-generated file and the flexure strength (MPa) was calculated using the relevant formula [12]. At least twenty bars were measured for each test group (n ≥ 20). Scanning electron microscopy (SEM) was conducted before and after flexure strength testing to examine the effect of the treatment applied to the surface structure and to analyze the fractured surface (XL20, Philips, Eindhoven, The Netherlands).

One-way analysis of variance and SNK post hoc tests were used to analyze the data. The failure probability under one-cycle loading was investigated using Weibull modulus. Pearson's correlation test was used to investigate the relation between flexure strength and surface roughness. According to the significance level (α = 0.5) and the

Strength influencing on CAD/CAM zirconia

101

sample size (n = 20), the test of choice had a power (1 − β = 0.9) to detect large effect size differences (F = 0.4), which in terms of flexure strength could be of clinical significance. 6.4 Results

Statistical analysis revealed significant differences in the three-point flexure strength values between the groups tested as a result of the different surface treatment methods applied (F = 35.5, p < 0.000). According to strength values, the groups tested could be divided into the following categories: (1074–1166 MPa) for polished zirconia and CAM and polished bars which were particle abraded (50 μm Al2O3), (936 MPa) for ground bars that were particle abraded (50 μm Al2O3), (708–794 MPa) for CAM bars, ground bars, and polished bars that were particle abraded (120 μm Al2O3), and (546 MPa) for the ground bars that were particle abraded (120 μm Al2O3) being the weakest. Thermal firing either before or after particle abrasion, had no significant influence on the flexure strength (Table 6.1).

Pearson's correlation test indicated that there was a significant relation between the reduction in flexure strength and the severity of the surface damage induced by different surface treatments (R2 = 0.912). As indicated by the surface roughness measured, polished zirconia bars had the lowest surface roughness values, followed by particle abrasion with 50 μm particles, CAD/CAM defects, and particle abrasion with 120 μm (Figure 6.1). The survival probability as calculated by Weibull modulus was also directly related to the degree of surface damage introduced by different surface treatments (Figure 6.2). The characteristic strength of the groups tested was calculated at 63% failure probability. A point worth noting is that CAD/CAM bars had the highest value of surface roughness, but on the other hand they were not the weakest compared with the other groups, which is explained in the following text. Scanning electron microscopy of the fractured bars revealed that the critical crack always started from the surface flaws created and propagated to split the loaded bars in half. Compression curls were observed under the loading point and demarcated the shift from tensile to compressive fields. Air-borne particle abrasion resulted in severe surface damage as sharp scratches, cracks, grain pull-out, and material loss. The severity of this surface damage was more severe for the 120 μm compared to 50 μm particles. Additionally, the damage not only related to the surface, but extended to the subsurface region to a depth of few microns (Figure 6.3).

Chapter 6

102

Cha

ract

-er

istic

st

reng

th

1244

.17

790.

23

809.

92

809.

22

1130

.70

1128

.15

1127

.94

855.

79

588.

48

1018

.56

758.

93

820.

65

1180

.92

Wei

bull

mod

ulus

7.26

5.

28

7.50

5.

38

9.91

11

.80

10.5

8 6.

13

6.19

5.

08

6.74

3.

99

9.99

Min

imum

st

reng

th

851.

38

470.

43

524.

00

485.

15

705.

11

783.

48

825.

44

562.

72

379.

49

545.

75

497.

19

381.

17

874.

64

Max

imum

st

reng

th

1528

.78

1010

.27

1026

.16

1041

.98

1228

.06

1213

.40

1233

.23

1173

.47

753.

00

1321

.06

951.

51

1084

.46

1399

.09

S.D

. of

stre

ngth

192.

28

164.

44

122.

51

160.

36

111.

21

103.

95

118.

73

156.

47

103.

58

211.

56

125.

55

213.

74

135.

10

Ave

rage

st

reng

th a

116

6.39

A

727

.54B

7

60.7

2B

45

.99B

10

74.5

7A

1080

.67A

10

76.0

9A

94.

83B

546.

56

936.

13

708.

72B

743.

35B

1123

.93A

Rv

0.35

5.

01

5.00

4.

66

2.76

2.

60

2.60

4.

38

6.27

2.

87

3.50

7.

20

6.33

Rp

0.17

5.

22

3.96

4.

29

2.77

2.

73

2.86

6.

51

4.18

2.

63

3.10

8.

99

5.81

Rou

ghne

ss

Ra

0.04

1.

18

1.10

1.

11

0.65

0.

62

0.59

1.

18

1.22

0.

66

0.87

1.

91

1.61

Tab

le 6

.1

F

lexu

re st

reng

th (M

Pa) a

nd su

rfac

e ro

ughn

ess (μm

) of t

he te

sted

gro

ups

Surf

ace

cond

ition

Polis

hed

Polis

hed

+ SB

120

Polis

hed

+ SB

120

+ fir

ed

Polis

hed

+ SB

120

+ fir

ed +

SB

120

Polis

hed

+ SB

50

Polis

hed

+ SB

50 +

fire

d

Polis

hed

+ SB

50 +

fire

d +

SB50

Gro

und

Gro

und

+ SB

120

Gro

und

+ SB

50

Polis

hed

+ SB

120

+ SB

50

CA

M

CA

M +

SB

50

SB12

0: sa

ndbl

aste

d w

ith p

owde

r of a

mea

n si

ze o

f 120

μm

; SB

50: s

andb

last

ed w

ith p

owde

r of a

mea

n si

ze o

f 50 μm

; CA

M: a

s mac

hine

d by

C

AM

/DA

M d

evic

e. a A

vera

ge fl

exur

e st

reng

th o

f the

gro

ups w

ith sa

me

supe

rscr

ipt l

ette

r was

not

stat

istic

ally

sign

ifica

ntly

diff

eren

t (F

= 3

5.5,

p <

0.0

01).

Strength influencing on CAD/CAM zirconia

103

400500600700800900

1000110012001300

Polish

ed

Polish

ed+S

B120

Polish

ed+S

B120+

Fired

Polish

ed+F

ired+

SB120

Polish

ed+S

B50

Polish

ed+S

B50+F

ired

Polish

ed+F

ired+

SB50

Polish

ed+S

B120+

SB50

Ground

Ground

+SB12

0

Ground

+SB50

CAM

CAM+SB50

Surface Treatment

Flex

ure

stre

ngth

(MPa

) -113579111315

Roughness(µm

)Strength Ra Rp Rv

Figure 6.1 Correlation between flexure strength and surface roughness.

Figure 6.2 One-cycle load failure probability of zirconia bars with different surface treatments.

0%

20%

40%

60%

80%

100%

400 600 800 1000 1200 1400 1600

Flexure stress (MPa)

PolishedPolished+SB120Polished+SB120+Fired Polished+SB120+Fired+SB120 Polished+SB50Polished+SB50+Fired+SB50 Polished+SB50+Fired GroundGround+SB120Ground+SB50Polished+SB120+SB50 CAMCAM+SB50

Frac

ture

P

roba

bilit

y

Chapter 6

104

Figure 6.3 (A top) Grain pull-out, sharp scratches, and pitting after particle abrasion with

120 μm aluminum oxide. (B bottom) Deep sharp crack (30 μm) as a result of particle abrasion with 120 μm aluminum oxide on ground zirconia.

Strength influencing on CAD/CAM zirconia

105

Figure 6.3 (C top) Critical crack originated from the damaged surface as result of particle

abrasion with 120 μm aluminum oxide. (D bottom) Cross-section view demonstrating the subsurface damage induced by particle abrasion with 50 μm aluminum oxide.

Chapter 6

106

Figure 6.4 (A top) Trace lines left on the surface as a result of CAM milling procedure.

The observed macroscopic features were not effective as crack initiation sites. (B bottom) Oblique image demonstrating that the critical crack (white arrow) was not related to CAM milling trace lines (black arrow).

Strength influencing on CAD/CAM zirconia

107

Figure 6.5 Severe damage introduced by grinding zirconia. On the other hand, the CAM milling procedure resulted in the creation of milling trace lines, which in the present research were parallel to the long axis of the bars. These trace lines did not act as crack initiation sites (Figure 6.4). Additionally, other milling defects, such as premature contact with the milling burs and detached islands of zirconia grains, were also observed all of which resulted in the large variations in the measured surface roughness and naturally resulted in a large standard deviation for the strength measured for this group. Grinding zirconia frameworks before sintering (ground bars) resulted in severe damage to the surface in the form of large islands of detached zirconia grains and deep defects (Figure 6.5). 6.5 Discussion

As the main function of the underlying framework is to support the ceramic veneer and to carry the loading forces, different laboratory tests were used to evaluate the internal strength of zirconia frameworks. The design of fixed partial dentures can be considered as a simple beam and different flexure strength tests are frequently used for strength evaluation [4]. On the other hand, standard flexure strength tests do not take into account important factors such as the effect of design, variation in thickness of the framework, and the nature of human occlusion and loading environment [13]. Nevertheless they offer a controlled environment for evaluation of the interacting variables of interest.

The strength of all-ceramic materials is directly related to the size, population, and distribution of structural defects and flaws. Additionally, the location of these

Chapter 6

108

defects also plays an important role. While bulk defects are more shielded and protected by the surrounding material, surface defects directly act as stress concentration sites, which magnify the applied stresses according to the severity of the surface flaw [14]. This severity is a function of the size and shape of the surface defect. Sharper and deeper defects increase the concentration of stress at their crack tips and thus are more likely to act as crack initiation sites [15].

Achieving a flaw-free state of a material seems only possible in theoretical applications as each material has some characteristic flaw population related to the fabrication and handling procedures selected. On the other hand, fine polishing tends to reduce the severity and the population of surface defects and flaws to a degree where the internal strength of the material becomes the dominant factor determining its mechanical performance [4]. According to the results of this study, the flexure strength values associated with the different surface treatment methods used could be categorized into four groups with polished zirconia being the strongest (Table 6.1).

It is clearly depicted in Figure 6.1 that there is a direct relationship between variations in the flexure strength and the associated surface roughness for each test group (Pearson correlation factor = 0.912). Three parameters were selected to analyze surface roughness, the Ra value which describes the average surface roughness as a mean of the elevations and depressions measured from an estimated surface, the RP value which represents the average vertical elevations measured from the estimated average surface, and the RV value which is the surface depressions measured from the estimated surface and in this case was the most representative of the resultant surface flaws and defects acting as effective stress concentration sites. The presented data are in agreement with Luthardt et al. who also reported similar surface roughness values but relatively lower flexure strength for ground zirconia [4].

The only apparent misleading deviation from this linear relation is that CAM bars were associated with the highest average surface roughness (Rp = 8.9 μm) and at the same time were not the weakest. This could be explained by the fact that the milling procedure was executed before the final sintering, while the other surface procedures tested were carried out after the final sintering. The sintering procedure might have a healing effect on the surface damage caused by the CAM grinding procedure. Moreover, the CAM milling burs have a circular cross-section and they contact the surface of the framework in calculated plans and contact points leaving behind milling trace lines (Figure 6.4). These macroscopic surface roughness or elevations do not directly act as stress concentration sites, but in fact it is the

Strength influencing on CAD/CAM zirconia

109

microscopic surface roughness represented by sharp cracks and scratches that act most effectively as crack initiation sites [16].

While even the smallest damage introduced either by the CAM milling procedure or particle abrasion may at the beginning seem trivial, it will have a detrimental effect on the fatigue life of zirconia restorations. With repetitive cyclic loading during mastication, small defects tend to grow in size until reaching a critical size where catastrophic failure will eventually result. Thus any temporary beneficial effects induced by different surface treatment methods will be counter weighted by crack growth phenomena [17, 18]. With consideration that crack growth starts at a threshold intensity factor, which for zirconia (KI0 = 3.1 MPa m1/2) [19] is much lower compared to its critical crack intensity factor (KIc = 7.4 MPa m1/2) [2], the smallest surface defect could be large enough to act as an effective stress concentration site, finally increasing the fracture probability [20].

In partial agreement with other studies which reported an increase in flexure strength after particle abrasion, relating such findings to the creation of a compressive field as a result of tetragonal–monoclinic transformation [6-9], the data of this study indicate that particle abrasion with 50 μm aluminum oxide resulted in an increase in the strength of CAM and ground bars, possibly by removing weakly attached surface grains and by the elimination of milling and grinding trace lines [6]. Supporting such explanation is the reduction of the surface roughness observed after particle abrasion (Table 6.1). On the other hand, if the increase in flexure strength was due to phase transformation, such effect would be lost after thermal firing which was not the case in this study. On the contrary, particle abrasion with 120 μm aluminum oxide resulted in significantly weakening all the bars tested and in increasing the surface roughness, which also explains such a deteriorating effect. Even though the strength of the CAM or ground bars that were air-borne particle abraded with 50 μm aluminum oxide was higher compared with the non-abraded bars, the long-term effect of the induced damage should be considered.

In addition to surface roughness, thermal firing of the framework before or after air-borne particle abrasion was investigated. While other studies found that thermal firing after particle abrasion lowers the strength of zirconia and related such finding to the reverse of the monoclinic phase back to tetragonal, which relieves the created compressive fields [6], in the present work thermal firing had no influence on the flexure strength whether performed before or after particle abrasion but on the other hand it resulted in an increase in the Weibull modulus of the air-borne particle abraded polished zirconia either with 50 or 120 μm particles from 9.91 to 11.8 and from 5.28 to

Chapter 6

110

7.5, respectively. As zirconia is a glass-free material, whether this is related to reverse transformation or due to relieving any present pre-stresses, it remains a point for further investigation [6, 21, 22].

The probability of the fracture of zirconia under a single-load cycle (Figure 6.2) indicates that zirconia is a highly reliable material for construction of all-ceramic frameworks and it can handle surface damage parameters, which if introduced to other ceramics would result in catastrophic failure at much lower loads. Nevertheless, one should not overestimate the strength of zirconia, as catastrophic failure could result at extremely unexpected low loads, 379 MPa minimum strength for ground zirconia with 120 μm particle abrasion.

Bearing in mind that CAD/CAM fabricated zirconia frameworks are already weakened by the surface damage induced by the milling procedure, all strength calculations should be redesigned using the characteristic strength of CAD/CAM fabricated zirconia, 820 MPa instead of 1244 MPa as calculated for polished zirconia. Careful selection of the surface treatment method of choice, which does not induce further excessive surface damage, is a prerequisite for the success of zirconia restorations. 6.6 References [1] M.N. Aboushelib, N. de Jager, C.J. Kleverlaan and A.J. Feilzer, Microtensile

bond strength of different components of core veneered all-ceramic restorations, Dent Mater 21 (2005), pp. 984–991.

[2] M.N. Aboushelib, N. de Jager, C.J. Kleverlaan and A.J. Feilzer, Effect of loading method on the fracture mechanics of two layered all-ceramic restorative systems, Dent Mater 23 (2007), pp. 952–959.

[3] Y. Zhang, B.R. Lawn, E.D. Rekow and V.P. Thompson, Effect of sandblasting on the long-term performance of dental ceramics, J Biomed Mater Res 71B (2004), pp. 381–386.

[4] R.G. Luthardt, M. Holzhuter, O. Sandkuhl, V. Herold, J.D. Schnapp and E. Kuhlisch et al., Reliability and properties of ground Y-TZP-zirconia ceramics, J Dent Res 81 (2002), pp. 487–491.

[5] R.G. Luthardt, M.S. Holzhuter, H. Rudolph, V. Herold and M.H. Walter, CAD/CAM-machining effects on Y-TZP zirconia, Dent Mater 20 (2004), pp. 655–662.

Strength influencing on CAD/CAM zirconia

111

[6] M. Guazzato, L. Quach, M. Albakry and M.V. Swain, Influence of surface and heat treatments on the flexural strength of Y-TZP dental ceramic, J Dent 33 (2005), pp. 9–18.

[7] A.R. Curtis, A.J. Wright and G.J. Fleming, The influence of surface modification techniques on the performance of a Y-TZP dental ceramic, J Dent 34 (2006), pp. 195–206.

[8] T. Kosmac, C. Oblak, P. Jevnikar, N. Funduk and L. Marion, The effect of surface grinding and sandblasting on flexural strength and reliability of Y-TZP zirconia ceramic, Dent Mater 15 (1999), p. 426.

[9] T. Kosmac, C. Oblak, P. Jevnikar, N. Funduk and L. Marion, Strength and reliability of surface treated Y-TZP dental ceramics, J Biomed Mater Res 53 (2000), pp. 304–313.

[10] J. Tinschert, D. Zwez, R. Marx and K.J. Anusavice, Structural reliability of alumina-, feldspar-, leucite-, mica- and zirconia-based ceramics, J Dent 28 (2000), pp. 529–535.

[11] I.L. Denry and J.A. Holloway, Microstructural and crystallographic surface changes after grinding zirconia-based dental ceramics, J Biomed Mater Res B: Appl Biomater 76 (2006), pp. 440–448.

[12] L. Xiao-Ping, T. Jie-Mo, Z. Yun-Long and W. Ling, Strength and fracture toughness of MgO-modified glass infiltrated alumina for CAD/CAM, Dent Mater 18 (2002), pp. 216–220.

[13] J. Kelly, Clinically relevant approach to failure testing of all-ceramic restorations, J Prosthet Dent 81 (1999), pp. 652–661

[14] A.R. Curtis, A.J. Wright and G.J. Fleming, The influence of simulated masticatory loading regimes on the bi-axial flexure strength and reliability of a Y-TZP dental ceramic, J Dent 34 (2006), pp. 317–322.

[15] S.S. Scherrer, I.L. Denry and H.W. Wiskott, Comparison of three fracture toughness testing techniques using a dental glass and a dental ceramic, Dent Mater 14 (1998), pp. 246–255.

[16] M. Albakry, M. Guazzato and M. Vincent Swain, Effect of sandblasting, grinding, polishing and glazing on the flexural strength of two pressable all-ceramic dental materials, J Dent 32 (2004), pp. 91–99.

[17] Y. Zhang, B. Lawn, A. Malament and P. Thompson, Damage accomulation and fatigue life of particel-abraded ceramics, Int J Prosthodont 19 (2006), pp. 442–448.

Chapter 6

112

[18] Y. Zhang and B.R. Lawn, Fatigue sensitivity of Y-TZP to microscale sharp-contact flaws, J Biomed Mater Res B: Appl Biomater 72 (2005), pp. 388–392.

[19] R. Marx, F. Jungwirth and P. Walter, Threshold intensity factors as lower boundaries for crack propagation in ceramics, BioMed Eng OnLine 3/1/41 (2004), pp. 1–9.

[20] Y. Deng, B.R. Lawn and I.K. Lioyd, characterization of damage modes in dental ceramic bilayer structures, J Biomed Mater Res 63 (2001), pp. 137–145.

[21] M. Guazzato, M. Albakry, L. Quach and M.V. Swain, Influence of surface and heat treatments on the flexural strength of a glass-infiltrated alumina/zirconia-reinforced dental ceramic, Dent Mater 21 (2005), pp. 454–463.

[22] S. Deville, J. Chevalier and H. El Attaoui, Atomic force microscopy study and qualitative analysis of martensite relief in zirconia, J Am Ceram Soc 88 (2005), pp. 1261–1267.

113

CHAPTER 7

Staircase evaluation of the fatigue strength of sandblasted zirconia

Keywords: dental ceramics, zirconia, strength, fatigue, sandblasting

Chapter 7

114

7.1 Abstract Objective: The aim of this study was to assess the fatigue behavior of sandblasted zirconia using the staircase method in search of better performance. Methods: 20 specimens for each group of a dental Y-TZP ceramic were polished or sandblasted with 50 µm or 120 µm alumina powders and subjected to cyclic or static fatigue loading. The specimen were loaded in a 3-point bending setup with 10,000 cycles at 1 s per cycle and a load time of 0.5 s, or with a constant load for 5000 s. The staircase method with 7.5% stress increments or decrements was used. Results: Compared to the single load strength, the fatigue strength in cyclic tests was 86.3% for polished zirconia, 73.4% for sandblasted with 50 µm particles, and 42.3% with 120 µm particles. The fatigue strength in static tests was 85.9%, 78.5%, and 51.5%, respectively. Statistically significant differences were found between the surface treatments, but not between cyclic and static test. Conclusion: Sandblasting with 50 µm particles produces significantly less degradation of the strength of this kind of zirconia than sandblasting with 120 µm particles.

Fatigue strength of sandblasted zirconia

115

7.2 Introduction Dental restorations are exposed to the aggressive oral environment. Because of

design and production related causes, these may fracture completely due to extreme incidental bite forces. Also the regular use causes fatigue-induced degradation and wear [1-3]. For all-ceramic restorations with a relatively brittle character it is a true challenge to withstand such odyssey. With the advantage of esthetics and biocompatibility many advanced ceramics have emerged in dentistry, which show improved strength [4, 5]. Currently, zirconia ceramics have been proved being the strongest and toughest commercially available all-ceramic core materials.

Zirconia has three crystallographic phases depending on the temperature and external pressure. The monoclinic phase is stable for pure zirconia at room temperature and pressure. The tetragonal phase exists at the temperature ranged between 1170°C and 2370°C while the structure becomes cubic at a temperature above 2370°C. So after pure zirconia is sintered at a temperature that is often at least 100°C higher than 1170°C, it undergoes a phase transformation during cooling from tetragonal to monoclinic phase starting at about 1170°C, which is accompanied by a volume increase of approximately 4.5%. However, dopant additions such as CaO, MgO, Y2O3, and CeO2, etc. may stabilize the tetragonal phase at room temperature and can be involved in the favorable stress-induced transformation from tetragonal to monoclinic phase, which generate compressive stress, contributing to crack arrest and superior mechanical properties. With the introduction of zirconia in dentistry the era of the design and structural buildup of extensive all-ceramic bridges started. Nevertheless, zirconia ceramics are still brittle materials and have serious drawbacks; they have little tendency to deform, are rendered fragile by fairly small flaws, which are randomly distributed, and fracture quickly upon critical crack growth [1-5]. In fact, surface treatments such as milling, grinding, sandblasting, etc, which are routine procedures in the production of zirconia structures, may significantly compromise the strength [6-8]. Surface or near-surface damage from finishing procedures [9] would be great enough to initiate the crack from the flaws and accelerate the subcritical crack growth from the initial size to a critical size [10], which then jeopardizes of the structural integrity of the material, i.e. may cause fracture.

Sandblasting can introduce surface/near-surface defects and, hence, it could influence zirconia strength as present defects play a crucial role in initiating and driving crack growth under stress or stress intensity factor. It is hypothesized that sandblasting not only decreases the strength immediately but also compromises the fatigue behavior of zirconia. The purpose of the study was to assess the fatigue

Chapter 7

116

strength of zirconia with different sandblasting surface treatment, using a staircase method. The zirconia tested in the present study was an yttrium-stabilized tetragonal zirconia polycrystals, Y-TZP, for dental use [5]. 7.3 Materials and Methods

Zirconia (Cercon, Degudent GmbH, Hanau-Wolfgang, Germany) specimens were prepared in a same way in a previous study [7], from which also the single load values were used (Table 7.1). Milling zirconia blocks were cut in a sawing machine (ISOMET Q2 1000, Buehler Ltd, Lake Bluff, IL) using a diamond coated sawing disc into bars, which were sintered as recommended in a special oven (Cercon Heat, Degudent GmbH, Hanau-Wolfgang, Germany) for 6.5 hours at a maximum temperature of 1350°C. The sintered bars (17 mm x 2 mm x 1 mm) were polished with sandpapers in a sequence from 600 to 1200 grit (ECOMET Grinder/Polisher, Buehler Ltd, Evanston, Q3 IL). The surfaces that were to be exposed to tensile forces in the subsequent strength tests were sandblasted using either 50 µm or 120 µm alumina powder at 0.35 MPa pressure for 5 s at a distance of 20 mm (P-G 400/3, Harnisch+Rieth, Winterbach, Germany).

The specimens were mounted in a 3-point strength test holder immersed in distilled water. The 3-point setups were loaded cyclically in an air-driven fatigue machine (ACTA Cyclic Fatigue Tester, ACTA, the Netherlands). The valley-loading force was set as continuous at 5 N to keep the specimen from moving away from the original position. The peak-loading force was set according to ratio of the initial strength, i.e. single-load-to-failure strength, which was obtained from a previous study [7] (see Table 7.1). The cycle time was set at 1 s and the on-time at 0.5 s. The on-time load was kept at a same level during each loading. The specimens were subjected to 104 load cycles. Based on a pilot study and the requirements for the staircase method, the peak loading force started at 80% of the initial strength for polished bars, 60% for sandblasted with 50 μm alumina, and 42.5% for sandblasted with 120 μm alumina. If a bar did not fail within 104 cycles, the peak-loading force for the next one would be increased by 5%, or otherwise, decreased by 5%. 20 bars were involved in each group.

The static fatigue tests were carried out in a tensilometer (ACTA intense, ACTA, Amsterdam, NL) for three more groups with the same surface treatments and staircase procedure as in the cyclic fatigue tests. The static load was set up like the peak force for the cyclic load but kept without change for 5000 s. The calculation of the peak-loading force for cyclic test or the hold force for static test of an individual specimen was based on the normal 3-point flexure strength formula [11]:

Fatigue strength of sandblasted zirconia

117

223BDWL

f =σ

Where σf is the strength, W is the fracture load, L the support span, B the specimen width, and D the specimen thickness. The residual strength of the specimens, which had survived the cyclic and static staircase tests, was determined in the same 3-point setup.

The staircase result was analyzed with multiple logistic regressions [12]. The means and standard deviation of the fatigue strength were computed and compared with the ANOVA and Turkey’s pair wise multiple comparisons at a significant level of 0.05. 7.4 Results

In Figures 7.1 and 7.2 the staircase result of the cyclic and static fatigue are presented. The results are summarized in Tables 7.1 and 7.2. Statistic analysis of fatigue strength showed that there are no significant differences between the cyclic and static fatigue values with the same surface treatment, but all differences between the surface treatments are significant. Among these, the groups sandblasted with 120 μm alumina not only had the lowest fatigue strength (Table 7.1), but also presented the greatest decrease of strength compared to the single load strength and a much higher decrease than those sandblasted with 50 μm alumina or polished (Table 7.2). Table 7.1 Fatigue, single load strength* (in MPa) and roughness* (Ra, Rp and Rv, in μm)

with the standard deviations in parentheses.

Surface Loading

Polished SB50 SB120

Cyclic Static Single load*

1006.5 (60.1) 1001.5 (121.9) 1166.4 (192.3)

789.0 (103.9) 843.7 (121.4) 1074.6 (111.2)

307.5 (6.0) 374.8 (80.4) 727.5 (164.4)

Ra* 0.04 (0.00) 0.65 (0.06) 1.18 (0.11) Rp* 0.17 (0.05) 2.77 (0.45) 5.22 (1.12) Rv* 0.35 (0.16) 2.76 (0.29) 5.01 (0.71) * data from reference [7]. SB50 / 120: sandblasted with 50 / 120μm alumina

Chapter 7

118

Table 7.2 Ratio of the fatigue strength values to the single load strength

Surface Loading

Polished SB50 SB120 Cyclic Static

86.3% 85.9%

73.4% 78.5%

42.3% 51.5%

SB50 / 120: sandblasted with 50 / 120μm alumina

0%10%20%30%40%50%60%70%80%90%

100%110%

1 3 5 7 9 11 13 15 17 19

Specimen number

load

ratio

to m

ean

stre

ngth

Polished SB50 SB120

Figure 7.1 Staircase results of the 104 cycles fatigue tests (solid marks are surviving

specimens).

0%10%20%30%40%50%60%70%80%90%

100%110%

1 3 5 7 9 11 13 15 17 19

Specimen number

load

ratio

to m

ean

stre

ngth

Polished SB50 SB120

Figure 7.2 Staircase results of the 5000 s static fatigue tests (solid marks are surviving

specimens).

The mean residual strength values of the survived specimens are greater than the fatigue values of the same groups and similar for the cycling and the static tests. The residual strength of the polished and SB50 groups seems somewhat greater than

Fatigue strength of sandblasted zirconia

119

the initial (single-load) strength, whereas in the SB120 groups the residual strength is less (Table 7.3). Table 7.3 Residual strength (in MPa) of survived zirconia specimens from staircase test

with the standard deviations in parentheses

7.5 Discussion An advantage of the staircase method is to give a good assessment of the mean

fatigue strength and the feasibility of the computation of the standard deviation and statistical analysis [12]. In order to conduct a more severe simulating test, the specimens were submerged in an aqueous environment, which an important factor in mechanical aging of zirconia [28]. However, the loading cycles and the stress values in the present investigation were not representative for the long-term loading level of dental restorations in vivo. A relatively high cyclic peak stress for an accordingly small number of cycles was employed here, which is similar to maximum stress or low-cycle fatigue approach [13] and the outcome should not be used for a calculative prediction of the service life at a stress level in vivo. Yet, the current study demonstrates the sensitivity of zirconia to the particles size of sandblasting powders, which is consistent with the literature [14, 15]. The extrapolation can be made that at more dental stress levels, polished zirconia or sandblasted with 50 μm powder is much more resistant to fatigue degradation than zirconia sandblasted with 120 μm powder. Sandblasting with 120 μm powder not only reduces the (single load) strength, but decreases the fatigue strength to an even greater extent, which may yield very short service times of zirconia-based all-ceramic restorations in the oral cavity, presumably due to their severe surface/near-surface damage introduced by impact of large blasting particles [7, 14, 15].

Loading Surface Residual Strength Ratio to single load Ratio to fatigue Cyclic Polished

SB50 SB120

1351.6 (120.9) 1253.4 (131.5) 604.8 (98.7)

115.9% 116.7% 83.2%

134.3% 158.9% 196.7%

Static Polished

SB50 SB120

1227.6 (221.8) 1201.2 (139.2) 604.2 (105.5)

105.3% 111.8% 83.1%

122.6% 142.4% 161.2%

Chapter 7

120

According to Griffith’s theory and Weibull distribution [29] the residual strength of the survived specimens should perhaps be greater than the initial strength, as weaker specimens had been removed with the staircase tests. This is true for SB50 and polished zirconia, but the residual strength of the SB120 specimens is much less than the initial strength, which could be caused by stress corrosion in water. Fatigue studies often find much greater differences between static and cycling tests, because of the breakdown of crack bridging / shielding ligaments behind the tip. These mechanisms progressively enhance the crack-tip driving force, and the wedging effect of fracture particles [21-23], which occur with cycling and are promoted at lower values of the stress intensity factor [24]. Such studies, however, use numbers of cycles, which are at least a hundred times greater and it could be that the small number of cycles in the current study and the consequently high loading force is largely responsible for the absence of difference between static and cyclic fatigue in the current study. A purpose of this study was to illustrate the effect of surface treatments on fatigue life [16] rather than only subcritical crack growth (SCG) parameters and inert strength [17-20]. Predictions with computations based on data from SCG parameters and inert strength may be inaccurate where they fail to take the deleterious influence of certain surface treatments on (fatigue) strength into consideration. Inert strength has often been tested with artificial large surface defects and a high loading speed. Mechanical machining operations such as milling, grinding, polishing, sandblasting, etc. can result in different levels of surface damages and resistance to fatigue crack nucleation. Particle impact during sandblasting may induce potential surface/near-surface damage, like surface valleys acting as stress concentrator, nucleation of micro-defects, and the creation of microscopic, even dominant cracks, may be induced by [7, 9, 13-16], which is essential for the rate of the later advance of crack growth because the pre-existing defects are closer to the critical crack size [10]. Therefore, it becomes understandable that the condition of the surface has a decisive role in the initiation of fatigue cracks, which determine the service life, i.e. cause complete fracture. This susceptibility is revealed clearly in a very straightforward procedure with the staircase method [27] other than SCG methods.

Y-TZP does show some phase transformation, but it could be much more brittle compared to other stabilized zirconia such as Ce-TZP or Mg-PSZ. Although it shows the highest initial strength, it also has the lowest resistance against aging, shows a sudden drop of strength above a critical crack size and behaves rather like ordinary ceramics with a flat R-curve [25, 26]. Furthermore, the transformation in Y-TZP

Fatigue strength of sandblasted zirconia

121

ceramics behaves in a different manner. The same zirconia, tested in a different study [30] showed much less transformation on the fracture surfaces than another Y-TZP, which suggests that the zirconia in the current study has less resistance to crack initiation and growth.

Being flaw-sensitive rather than tolerant, the yttrium stabilized zirconia tested in the present study is still a brittle material [25], which implies that during production procedures, attention and effort should be paid to control the condition of surfaces in a attempt to minimize the effect of fabrication-induced flaws. In previous studies sandblasting with 50 μm alumina improved the strength of milled and ground zirconia, possibly by removal of a damaged layer and creating a smoother surface, while sandblasting with 120 μm alumina was detrimental [7]. Also in the present study sandblasting with 50 μm alumina deteriorated the fatigue strength significantly less than 120 μm. Sandblasting with 50 μm alumina could be performed after sintering and before cementation. Diminishing flaws is helpful to delay the degradation of strength and prolong the functioning duration of zirconia-based restorations. Flaws should be eliminated and restorations treated gently to favor long-term clinical service. 7.6 Conclusion

Sandblasting with coarse particles, such as 120 μm alumina seriously degrades the fatigue strength of zirconia. Sandblasting with 50 μm particles is considerably less damaging and if sandblasting is necessary, is recommended. 7.7 References [1] J.R. Kelly, Ceramics in restorative and prosthetic dentistry, Annu Rev Mater Sci

27 (1997), pp. 443-68. [2] J.R. Kelly, Clinically relevant approach to failure testing of all-ceramic

restorations, J Prosthet Dent 81 (1999), pp. 652– 61. [3] K.J. Anusavice, Degradability of dental ceramics, Adv Dent Res 6 (1992), pp.

82-9. [4] J.Y. Thompson, B.R. Stoner and J.R. Piascik, Ceramics for restorative dentistry:

Critical aspects for fracture and fatigue resistance, Mater Sci Eng–C 27 (2007), pp. 565-9.

[5] I. Denry and J.R. Kelly, State of the art of zirconia for dental applications. Dental Mater (2007), e-pub.

Chapter 7

122

[6] R.G. Luthardt, M.S. Holzhüter, H. Rudolph, V. Herold and M.H. Walter, CAD/CAM-machining effects on Y-TZP zirconia. Dent Mater 20 (2004), pp. 655-62.

[7] H. Wang H, M.N. Aboushelib and A.J. Feilzer, Strength influencing variables on CAD/CAM zirconia frameworks, Dental Mater (2007), e-pub.

[8] R.G. Luthardt, M. Holzhuter, O. Sandkuhl, V. Herold, J.D. Schnapp, E. Kuhlisch and M. Walter, Reliability and properties of ground Y-TZP-zirconia ceramics. J Dent Res 81 (2002), pp. 487-91.

[9] D. Rekow and V.P. Thompson, Proc. Near-surface damage--a persistent problem in crowns obtained by computer-aided design and manufacturing. Inst Mech Eng [H] 219 (2005), pp. 233-43.

[10] R.O. Ritchie and R.H. Dauskardt, Cyclic fatigue ceramics: a fracture mechanics approach to subcritical crack growth and life prediction. J Ceram soc Jpn 99 (1991), pp. 1047-62.

[11] H. Wang, Y. Liao, Y. Chao and X. Liang, Shrinkage and strength characterization of an alumina–glass interpenetrating phase composite for dental use. Dent Mater 23 (2007), pp. 1108-13.

[12] W.J. Dixon and A.M. Modd, A method for obtaining and analyzing sensitivity data. J Am Stat Assn 43 (1948), pp. 109-26.

[13] S. Suresh, Introduction and Overview. In: Suresh S., editor. Fatigue of Materials. 2nd edition. Cambridge: Cambridge University Press, (1998), pp. 1-35.

[14] Y. Zhang, B.R. Lawn, E.D. Rekow and V.P. Thompson, Effect of sandblasting on the long-term performance of dental ceramics. J Biomed Mater Res B Appl Biomater 71 (2004), pp. 381-6.

[15] Y. Zhang, B.R. Lawn, K.A. Malament, P. Van Thompson and E.D. Rekow, Damage accumulation and fatigue life of particle-abraded ceramics. Int J Prosthodont 19 (2006), pp. 442-48.

[16] A.G. Evans, Overview No. 125 Design and life prediction issues for high-temperature engineering ceramics and their composites. Acta Mater 45 (1997), pp. 23-40.

[17] A.R. Studart, F. Filser, P. Kocher and L.J. Gauckler, Fatigue of zirconia under cyclic loading in water and its implications for the design of dental bridges. Dent Mater 23 (2007), pp. 106-14.

Fatigue strength of sandblasted zirconia

123

[18] J. Tinschert, G. Natt, N. Mohrbotter, H. Spiekermann and K.A. Schulze, Lifetime of alumina- and zirconia ceramics used for crown and bridge restorations. J Biomed Mater Res B Appl Biomater 80 (2007), pp. 317-21.

[19] H. Fischer, M. Weber and R. Marx, Lifetime prediction of all-ceramic bridges by computational methods, J Dent Res 82 (2003), pp. 238-42.

[20] E.C. Teixeira, J.R. Piascik, B.R. Stoner and J.Y. Thompson. Dynamic fatigue and strength characterization of three ceramic materials. J Mater Sci Mater Med 18 (2007), pp. 1219-24.

[21] C.J. Gilbert, J.J. Cao, W.J. Moberlychan, L.C. Dejonghe and R.O. Ritchie. Cyclic fatigue and resistance-curve behavior of an in situ toughened silicon carbide with Al-B-C additions. Acta Mater 44 (1996), pp. 3199-214.

[22] H. El Attaoui, M. Saadaoui, J. Chevalier and G. Fantozzi G. Static and cyclic crack propagation in Ce-TZP ceramics with different amounts of transformation toughening. J Eur Ceram Soc 27 (2007), pp. 483-86.

[23] F. Guiu, M.J. Reece and D.A.J. Vaughan, Cyclic fatigue of ceramics. J Mater Sci 26 (1991), pp. 3275-86.

[24] S. Suresh, Fatigue crack growth in brittle solids. In: Suresh S, Editor. Fatigue of Materials. 2nd edition. Cambridge: Cambridge University Press; (1998), pp. 383-407.

[25] M.J. Readey, C.L. McCallen and P.D. McNamara, Correlations between flaw tolerance and reliability in zirconia. J Mater Sci 28 (1993), pp. 6748-52.

[26] S.K. Lee, R. Tandon, M.J. Readey and B.R. Lawn, Scratch damage in zirconia ceramics. J Am Ceram Soc 83 (2000), pp. 1428-32.

[27] H.W. Wiskott, J.I. Nicholls and U.C. Belser, Stress fatigue: basic principles and prosthodontic implications. Int J Prosthodont 8 (1995), pp. 105-16.

[28] J. Chevalier, What future for zirconia as a biomaterial? Biomaterials 27 (2006), pp. 535–43.

[29] J.R. Kelly, Perspectives on strength. Dent Mater 11 (1995), pp. 103-10. [30] J. Chai, F.C. Chu, T.W. Chow, B.M. Liang, Chemical solubility and flexural

strength of zirconia-based ceramics, Int J Prosthodont 20 (2007), pp. 587-95.

Chapter 7

124

125

CHAPTER 8

Summary and conclusions

The present series of studies investigated the mechanical behavior of dental porcelains and ceramics under various conditions. Strength demonstrates the “initial” resistance of a material against fracture. Besides the strength, fracture toughness and fatigue are important factors in the survival of porcelain and ceramic based restorations. The fracture toughness and the fatigue resistance describe the ability of a material to resist crack growth. This thesis focused on different methods to evaluate strength, fracture toughness, and fatigue with the aim to obtain clinically relevant values.

In Chapter 2, three test methods were used to compare the fracture toughness of the dental porcelain Carrara Vincent, and a soda lime glass. The soda lime glass has a homogeneous microstructure, while the traditional dental porcelain is feldspar-based and leucite-reinforced, which improved the fracture toughness. The fracture toughness was determined with the following methods; Indentation-Strength-in-Bending method (ISB), Single-Edge Notched Beam (SENB), and Chevron-Notched Beam (CNB). The results of the different fracture toughness tests were dependent on test methods and material. ISB, SENB, and CNB were not consistent with each other. Because only two materials were investigated it was hard to generalize the conclusions of this study that microstructure and chemistry of materials may influence the fracture toughness results of individual test methods. It seemed that technical reasons, like preparation and way

126

of testing, were also important factors. Anyhow, it showed that one should realize that fracture toughness values from different test methods might not be comparable.

In Chapter 3 the influence of the same three fracture toughness test methods was further evaluated with four commercial dental porcelain materials with different KIc values. For these materials the CNB and SENB test were statistically consistent. According to the results mentioned in Chapter 2, Carrara Vincent had statistically different toughness when evaluated with CNB and SENB, while in the current study, the same different values were not significantly different. Furthermore, it was shown that the ISB method was not always in agreement to CNB or SENB. Also in literature, fracture toughness methods may present different values for same materials and different ranking of varied materials, in a same manner as strength tests. For dental porcelains and ceramics, which have a broad variation in microstructure, chemistry, and fracture toughness, it is useful to systematically assess the comparability of the different test methods with standardized method.

The effect of the strength test configuration and the crosshead speed with the indentation-strength-in-bending method (ISB), which is easy and fast compared to SENB and CNB, was evaluated in Chapter 4. The fracture toughness and the 3-point, 4-point, and biaxial strength were determined for two dental porcelains. It was found that in an ISB test the 3-point, 4-point, and biaxial strength setups did not have a significant influence. However, the crosshead displacement speed was of influence for some test conditions.

In Chapter 5, the influence of the indentation load of the ISB method was evaluated. A higher load significantly increased the found toughness values. Specimens were rejected if they did not fail from the indentation, but from a flaw on the tensile surface. The tougher materials showed to have smaller indentation flaws and more rejected specimens. At small flaw sizes, rejected ISB specimens were often stronger than average. Testing the un-indented strength of identical specimens to verify that the indented strength is sufficiently lower can be a safe requirement for a valid experiment. When using ISB methods for comparing materials with different toughnesses, the indentation load applied to create predictable surface flaws should be adjusted for each material.

Chapter 2, 3, 4 and 5 were helpful to understand fracture toughness assessment and its sensitivity to the test configurations. Meanwhile, one has to keep in mind that small-crack and large-crack methods for fracture toughness measurement may lead to different resistance to crack growth. The mentioned chapters were focused on dental porcelains and ceramics, which are used in PFM restorations and as layering porcelain

Summary and conclusions

127

in all-ceramics. The zirconia in all-ceramic restorations is used as core material instead of a metal. During the fabrication processes, like shaping, trimming, sandblasting, machining and grinding, the zirconia surface is damaged. The induced flaws may compromise the strength, toughness and fatigue resistance.

In Chapter 6, after CAD/CAM machining and/or sandblasting with 120 μm, the tested zirconia showed a significantly reduced strength compared to polished specimens. However, sandblasting with 50 μm alumina did not decrease strength of zirconia so much and was statistically close to polishing. On the contrary, it increased not only the strength of ‘as-sintered’ and machined zirconia but also their structural reliability, which may be explained by the removal of large surface damages.

In Chapter 7, the fatigue resistance of the same zirconia was investigated. It was found that sandblasting with 50 µm particles produces significantly less degradation than sandblasting with 120 µm particles. For this zirconia, sandblasting with 50 μm was recommended for long-term clinical performance. Based on the obtained fatigue resistance of these zirconia ceramics, it was shown that different fabrication processes might lead to different strength degradation. The insensitivity of the studied zirconia to sandblasting might be explained by the grain interface and phase transformation-triggering stress levels.

Strength, toughness, and fatigue resistance tests of dental porcelains and ceramics were investigated. For toughness measurements, the test setup design should be carefully considered. The method should be chosen based on the materials’ service condition. The processing of the materials and specimens should be considered, as during the fabrication process flaws are introduced, which may lead to a reduction of the strength, fracture toughness, fatigue resistance and ultimately the clinical survival rate. In conclusion, strength and toughness tests must be well designed and performed for the estimation of the clinical survival rate of dental ceramic systems.

128

Samenvatting en conclusies

In dit proefschrift zijn de mechanische eigenschappen van verschillende tandheelkundige porseleinen en keramische materialen onder verschillende test omstandigheden onderzocht. Sterkte is gerelateerd aan de ‘initiële’ weerstand van het materiaal tegen breuk. Naast sterkte zijn de breuktaaiheid en vermoeiingsweerstand belangrijke factoren bij de levensduur van porseleinen en keramische restauraties. De taaiheid en de vermoeiingsweerstand beschrijven de weerstand van een materiaal tegen propagerende scheuren. De nadruk van dit proefschrift ligt op de evaluatie van de sterkte, breuktaaiheid en vermoeiingsweerstand in relatie tot het klinische gebruik.

In Hoofdstuk 2 zijn de resultaten van drie test methoden voor bepaling van de breuktaaiheid van een tandheelkundig porselein Carrara Vincent en een standaard glas met elkaar vergeleken. Het gewone glas heeft een homogene microstructuur en het traditionele tandheelkundige porselein is een samengesteld materiaal, met leuciet versterkt veldspaat, wat de breuktaaiheid verhoogd. De breuktaaiheid is bepaald met de volgende methoden: Indentation-Strength-in-Bending method (ISB), Single-Edge Notched Beam (SENB) en Chevron-Notched Beam (CNB). Uit de resultaten van de verschillende taaiheids testen blijkt, dat de breuktaaiheid afhankelijk is van de test methode en het materiaal. ISB, SENB en CNB waren onderling niet consistent. Maar omdat er maar twee materialen onderzocht zijn, is het niet mogelijk om algemene conclusies te trekken ten aanzien van invloed van de microstructuur en de chemie van de materialen op de breuktaaiheid, die volgens een bepaalde methode is bepaald. Het lijkt er op dat meer technische redenen, zoals de preparatie van de monsters en de manier van testen belangrijke factoren zijn. Deze studie laat zien dat de breuktaaiheid van de verschillende test methoden verschillende resultaten kunnen geven.

In Hoofdstuk 3 zijn dezelfde drie breuktaaiheidstesten methoden verder geëvalueerd met vier commercieel verkrijgbare tandheelkundige porseleinen met verschillende breuktaaiheidswaarden (KIc). Bij deze materialen waren de uitkomsten van de CNB- en de SENB-testen niet significant verschillend. Uit de resultaten van hoofdstuk 2 bleek dat met de CNB en SENB test Carrara Vincent significant

Summary and conclusions

129

verschillende breuktaaiheden had, terwijl dat met de testen beschreven in dit hoofdstuk niet het geval was. Verder blijkt, dat de ISB methode niet altijd overeenkomt met de CNB of de SENB methode. Ook in de literatuur worden verschillende breuktaaiheden en sterkten gevonden voor de verschillende methoden. Tandheelkundige porseleinen en keramische materialen hebben verschillende microstructuren, chemische samenstelling en een andere breuktaaiheid. Het systematisch onderzoeken van de methoden maakt deze bruikbaar en vergelijkbaar.

Het effect van de configuratie van de sterktetest en de snelheid van drukbank bij de indentation-strength-in-bending (ISB) methode wordt onderzocht in Hoofdstuk 4. Dit is gedaan, omdat deze methode gemakkelijk en snel is vergeleken met de SENB en CNB methoden. De breuktaaiheid en de 3-punts, 4-punts en de bi-axiale sterkte zijn bepaald voor twee tandheelkundige porseleinen. Uit de resultaten blijkt, dat de configuratie van de 3-punts, 4-punts en de bi-axiale sterktetest geen significante invloed heeft, maar dat de snelheid van de trekbank wel een effect op de testresultaten kan hebben.

In Hoofdstuk 5 is de invloed van de indentatie kracht van de ISB methode bestudeerd. Hogere indentatie krachten resulteren in een significant hogere breuktaaiheid. Monsters, die niet gebroken zijn op de plaats van de indentatie maar ter plaatse van een ander oppervlaktedefect zijn buiten beschouwing gelaten. Taaiere materialen hadden minder diepe indentaties en meer van deze afgekeurde testmonsters. Bij kleinere oppervlaktedefecten blijken deze afgekeurde testmonsters vaak sterker. Dezelfde testmonsters, maar dan zonder indentatie, zijn gebruikt als referentie om te controleren of de sterkte inderdaad is afgenomen vergeleken met een testmonster met een indentatie. Als de ISB methode wordt gebruikt voor het vergelijken van de breuktaaiheid, blijkt dat de indentatie-kracht, die gebruikt wordt voor het maken van het oppervlakte defect, moet worden aangepast aan de eigenschappen van het te testen materiaal.

De Hoofdstukken 2, 3, 4 en 5 zijn behulpzaam bij het begrijpen van het bepalen van de breuktaaiheid en de gevoeligheid voor de test-configuratie. Ondertussen kan er ook nog verschil zijn in de grootte van de initiële scheur waar de verschillende methodes gebruik van maken. Ook dit kan tot verschillende resultaten leiden. In de bovengenoemde hoofdstukken is de aandacht gericht op de in de tandheelkunde toegepaste porseleinen en keramische materialen. Deze materialen worden meestal toegepast in metaal-keramische restauraties en als opbakporselein bij volledige keramische restauraties. Zirconia wordt in volledige keramische restauraties gebruikt voor de onderstructuur, in plaats van het metaal zoals bij metaal-keramische

130

restauraties. Tijdens de fabricage worden processen zoals boren, schuren, slijpen en zandstralen gebruikt. Al deze processen kunnen het zirconia-oppervlak beschadigen en al deze oppervlakte-defecten kunnen invloed hebben op de sterkte, breuktaaiheid en vermoeiingsweerstand.

In Hoofdstuk 6 blijkt, dat het bewerken van zirconia met CAD/CAM apparatuur c.q. zandstralen met 120 μm alumina deeltjes het materiaal significant verzwakt vergeleken met gepolijste testmonsters. Zandstralen met 50 μm alumina deeltjes geeft deze verzwakking echter veel minder en laat sterkten zien, die statistisch vergelijkbaar zijn met die van gepolijste testmonsters. Dit verbeterderde niet alleen de sterkte vergeleken met de ruwe of de bewerkte monsters, maar ook de variatie in sterktewaarden (betrouwbaarheid). Dit laatste kan worden verklaard, doordat polijsten leidt tot het verwijderen van eventuele oppervlakte defecten.

In Hoofdstuk 7 is de vermoeiingsweerstand van hetzelfde zirconia onderzocht. Uit deze studie blijkt, dat het zandstralen met 50 μm aluminiumoxide deeltjes in vergelijking met 120 μm deeltjes tot minder degradatie van de vermoeiingsweerstand leidt. Voor dit type zirconia valt het aan te raden om met 50 μm aluminiumoxide deeltjes te zandstralen om een optimale langdurige klinische levensduur te verkrijgen. Op basis van de vermoeiingsweerstands-waarden van dit type zirconia blijkt, dat het fabricatie-proces kan leiden tot verschillende degradatie-processen. De ongevoeligheid van dit type zirconia voor zandstralen met 50 μm deeltjes kan worden verklaard door de microstructuur van het materiaal, wat de spanningswaarde beinvloedt waarbij fase-transformatie kan optreden.

De sterkte, breuktaaiheid en vermoeiingsweerstand testen van verschillende tandheelkundige porseleinen en keramische materialen zijn onderzocht. Bij de breuktaaiheid-metingen blijkt, dat de meetopstelling een belangrijke factor kan zijn. De gekozen methode moet vergelijkbaar zijn met de situatie zoals die is tijdens het normale (klinisch) functioneren. Ook spelen de manier waarop de materialen worden verwerkt tijdens het fabricageproces een belangrijke rol. Tijdens dit proces kunnen er defecten ontstaan, die een negatieve invloed hebben op de sterkte, breuktaaiheid en de vermoeiingsweerstand. Uiteindelijk zal dit dus ook een negatieve invloed hebben op het klinisch functioneren. Alles bij elkaar moet men optimaal ontworpen en uitgevoerde sterkte- en taaiheidtesten toepassen om een schatting te kunnen geven over het klinisch functioneren van tandheelkundige keramische materialen.

131

ACKNOWLEDGMENT My gratitude goes to Prof. dr. Albert J. Feilzer, who accepted me as a visiting scholar first and then started the research program. He did more than a promoter to take care of my things both academic and personal for an easier and happier stay and work in the Netherlands. I wish to thank Pallav who always supported my behavior in thought, performance and writing of research. The doctorate committee is gratefully acknowledged for evaluating my thesis. Cees, Toon and Arie who warm-heartedly helped the project in arrangement and technical issues. Jacqueline is appreciated for her kind catering and showing around. The following persons are acknowledged for their friendly care: Moustafa, Hesam, Joris, Filip, Alma, Niek, Jef, Giuseppe, Ahmed, Tony, Anne, Leontine, and the staff in the ACTA dental lab. I would like to thank my other colleagues in ACTA, friends in the Netherlands and colleagues in the West China College of Stomatology for their support, help and care. My Dean, Prof. dr. Xuedong ZHOU and Chairman, Prof. dr. Zhimin ZHU made the collaborative research feasible. The financial support by the CSC of the Chinese government and the IOT/ACTA of Universiteit van Amsterdam and Vrije Universiteit, the Netherlands are gratefully acknowledged, and surely, my wife, Bo Wu, for her encouragement.