Influence of Cement on Survival of All-Ceramic Restorations

Home , Glass ionomer cement, Luting agent

INFLUENCE OF CEMENT ON SURVIVAL OF ALL-CERAMIC RESTORATIONS

THESIS

Presented in Partial Fulfillment of the Requirements for The Degree of Master of Science in the Graduate School of The Ohio State University

Enas Elbahie Alakhras, B.D.S Graduate Program in Dentistry The Ohio State University 2011

Thesis Committee: Dr. Robert Seghi, DDS, MS (Advisor) Dr. William A. Brantley, PhD Dr. Noriko Katsube, PhD

ABSTRACT

Previous research has shown that new technology in adhesive dentistry improves the performance of all-ceramic restorations. However, the major reason for failure of these restorations remains the occurrence of fractures. The overall objective of this research project was to investigate the influence of cement on the survival of all-ceramic restorations.

A preliminary study was performed to evaluate the influence of the cement as a supporting structure on the survival of a simulated all-ceramic restoration. A trilayer simulation of a model restoration subjected to a clinically relevant condition of functional mastication was used. The results from the preliminary study showed that adhesively bonded specimens had higher survival rates than those conventionally cemented and that one of the adhesive cements had a significant higher survival rate than the other.

Based on results from the preliminary study, three other studies were performed to investigate why adhesive cementation improves the performance of all-ceramic restorations. Results from these studies showed: (1) Resin cements had fewer defects or were void-free at the ceramic-cement interface of our ceramic model, while conventional cements showed areas of voids at the this interface. (2) The resin cement had no influence on ceramic sensitivity to slow crack growth (SCG). (3) While the actual mechanism for

resin strengthening could not be determined, it may involve the formation of a more durable bond at the ceramic/cement interface.

iii

Dedicated to My Mother, Father and Husband

ACKNOWLEDGEMENTS

I would like to thank my thesis advisor; Dr. Robert Seghi for taking me under his wing and teaching me everything I know about research. You have an enormously generous heart. I am always inspired by you and I thank you for standing by me and helping me constructs and executes this study. Thank you so much for giving me an opportunity to work with you and for you always support, encouraging and helping me see the true value in research and inspiring me to be a better researcher.

I would like to also thank Dr. William Brantley whose knowledge and guidance contributed tremendously to this project. It is an honor to me to be Dr. Brantley‟s student and learning from him does not only contribute to my academic performance but also influenced my personal life.

To my parents, Sana and Mohammed, I would like to first and foremost thank you both for the amazing love and support you have given me throughout the years. Thank you for encouraging me endlessly to pursue higher education and explaining to me how through education and knowledge comes enlightenment and humility. Both of you have inspired me to always do my best and to constantly challenge and better myself. As well, in my difficult moments you were both endlessly comforting and even more supporting.

You have done your best to give me the opportunities and the means to pursue my dreams that were not available to either one of you growing up, and for that I am

eternally grateful. I love you both very much and I dedicate everything to my loving parents. I am nothing without you and I am everything with you.

To my wonderful beloved husband Mohamed, whose love, patient and support brighten my life and enabled me to finish my masters program. Thanks you for being always by my side, inspite of being busy, but you were always there when it most needed.

To my mother in law Safiya, for her love, encouragement and patient.

To my wonderful sisters, Amira, Eman, Sara and Dalia, even though we have been far apart for a long time, I still feel the love and care every time we meet. I wish you the very best in everything you pursue in life.

I would like to also thank all my faculty members at The Ohio State University for educating me and being great mentors.

VITA

December 09, 1983…………………………………………Sharkiya, EGYPT

2005…………………………………………………………B.D.S ………………………………………………………………Suez Canal University ………………………………………………………………Ismailia, EGYPT

2007……………………………………………………...….Instructor ………………………………………………………………Dental Materials Department ………………………………………………………………School of Dentistry ………………………………………………………………Suez Canal University ………………………………………………………………Ismailia, EGYPT

2008-09……………………………………………………...Research Assistant ………………………………………………………………School of Dentistry ………………………………………………………………The Ohio State University ………………………………………………………………Columbus, Ohio

2009 to present………………………………………………MS, Dental Material Program ………………………………………………………………The Ohio State University ………………………………………………………………Columbus, Ohio

2010 to present……………………………………………...Bench Instructor ………………………………………………………………School of Dentistry ………………………………………………………………The Ohio State University ………………………………………………………………Columbus, Ohio

vii

FIELDS OF STUDY

Major Field: Dentistry

Specialty: Dental Materials

viii

TABLE OF CONTENTS

Page

Abstract...... ii

Dedication...... iv

Acknowledgements ...... v

Vita...... vii

Chapter 1 Introduction ...... 1

Chapter2 Influence of cement on the interface initiated fracture rate of all ceramic restorations under cyclic loading...... 14

Chapter 3 Nondestructive acoustic microscopic Evaluation of the cement / glass-ceramic interface...... 36

Chapter 4 Influence of resin cement on slow crack growth parameter of two glass-ceramic…...... 51

Chapter 5 Effect of short and long terms aging conditions on the micro-tensile bond strength of resin cement to glass-ceramic…………………………….…………………………...... 78

Chapter 6 summary and conclusion………………………………………………………………...93

References………………………….………………………………….……………...…94

Chapter 1

Introduction

Over the past decade the high demand for esthetically pleasing restorations has driven the development of all-ceramic systems for use in dental restorations (inlays, onlays, crowns, fixed partial dentures (FPD‟s), and implant-supported restorations). The absence of a metallic substructure in all-ceramic restorations allows them to have improved esthetics, as the underlying tooth structure can potentially influence the final shade of the restoration and more closely imitate the optical effects of the natural teeth and result in a more “life-like” or natural appearance. These restorations are nonmetallic and biocompatible, which provide an advantage to soft tissue health since lesser amounts of plaque and adherence molecules are recovered from ceramic surfaces 1. It is often acceptable in all-ceramic restorations to leave the margin of the prosthesis supragingival or at the gingival margin, which adds the benefit of more predictable and less traumatic impression making. Improved clinical performance, especially with the use of higher strength ceramics and adhesives for bonding the ceramic restoration to tooth structure, have led to a resurgence of interest in all-ceramic restorations and make these restorations a more favorable choice by patients and dentists.

Dental ceramics can be considered as “composites” because they have two or more distinct entities in their compositions. Dental ceramic materials can be classified according to their crystalline content into three main divisions: (1) predominantly glassy materials, (2) particle-filled glasses, and (3) polycrystalline ceramics. Defining characteristics are provided for each of these ceramic types 2, 3. High-esthetic dental ceramic restorations consist predominantly of glass, while high-strength ceramics are generally polycrystalline.

Predominantly glassy ceramics are the best at mimicking the optical properties of enamel and dentin. Glasses in dental ceramics are derived from a group of mined minerals called feldspars, and are based on silicon and aluminum oxides. Hence, the dental feldspathic porcelains belong to a family called aluminosilicate glasses 3. Feldspar glass-ceramics are resistant to crystallization (devitrification) during firing and have long firing temperature ranges. They resist slumping if firing temperatures rise above optimal.

There is a wide range of materials used for all-ceramic restorations, and they contain a variety of crystalline phases. Theses crystalline phases act primarily as reinforcements to the ceramic matrix and improve the material properties. The physical and mechanical properties of ceramics are very sensitive to the nature, amount, and particle size distribution of the crystalline phases 4. Filler particles are usually added to the base glass composition to improve mechanical properties and to control optical properties such as color, opalescence, and opacity. These fillers are usually crystalline but can also be particles of a higher melting glass. Particle-filled glass-ceramics can be

classified by the type and amount of filler particles they contain, the reason why the particles were added, and the manner in which they were incorporated into the glass.

Leucite crystals were the first and most common particles added to glass-ceramics 2, 3.

Leucite was chosen primarily because its index of refraction is close to that of feldspathic glasses. The match between the refractive indices of the crystalline phase and glassy matrix is a key factor for controlling the translucency of the porcelain.

Leucite also has a high thermal expansion/contraction coefficient (≈ 20 x 10-6/°C) compared with feldspathic glasses (≈ 8 x 10-6/°C). Leucite particles help improve the thermal compatibility of glass-ceramics with dental alloys during firing when used in amounts of 17 ‒ 25 mass %. Therefore, the thermal expansion behavior of dental porcelains is quite sensitive to changes in leucite concentration. Incorporation of leucite crystals may cause a moderate increase in strength through a process termed „„dispersion strengthening”, when used at concentrations of ≈ 40 ‒ 55 mass % 5.

Different approaches have been used to incorporate the crystalline filler particles into the glass: (1) They can be added mechanically to the glass, by mixing together crystalline and glass powders before firing. (2) In a more recent approach the filler particles are grown inside the glass object (prosthesis or pellet for pressing into a mold) after the object has been formed, by subjecting the object to a special heat treatment causing precipitation and growth of crystallites within the glass 5. The crystallization is achieved when the glass is submitted to a heat treatment during which crystal growth occurs. Proper control of the crystallization heat treatment is necessary to ensure the

growth of crystals to an effective size 2. Because these fillers are derived chemically from atoms of the original glass, the composition of the remaining glass is altered as well during this process, which is termed „„ceraming.‟‟ Such particle-filled composites are called glass-ceramics and contain a glassy matrix phase and a crystalline phase (filler particles). 5

High-strength ceramics were developed by gradually increasing the volume percentage of crystalline material while decreasing the glass content and finally excluding all glass content from the mixture, resulting polycrystalline solids5.

Fully polycrystalline ceramics have no glassy components, and all atoms are densely packed into regular (crystalline) arrays that make crack propagation much more difficult than in the less dense and irregular network of atoms found in the noncrystalline glasses. Hence, polycrystalline ceramics are generally much tougher and stronger than glassy ceramics (entirely or predominantly glass phase). However, polycrystalline ceramics are more difficult to process into complex shapes than are glassy ceramics 5.

All-ceramic restorations combine esthetic veneering porcelains with strong ceramic cores. Veneering porcelains consist of a glass and a crystalline phase of fluoroapatite, aluminum oxide, or leucite. Ceramic cores are usually lithium disilicate, aluminum oxide, or zirconium oxide. Veneering a ceramic core with glass allows dental technicians to customize these restorations in terms of form and esthetics.

From the glass-ceramic family, lithium disilicate (SiO2-Li2O), leucite (SiO2-

Al2O3-K2O) and feldspathic porcelain (SiO2-Al2O3-Na2O·K2O) are currently used in all- ceramic dental restorations. Lithium disilicate glass-ceramics have been used in crowns, onlays and anterior FPD‟s because of better mechanical properties than leucite-reinforced glass-ceramics which have been used primarily for crowns and onlays. Feldspathic porcelain is used mainly for laminate veneers due to its excellent esthetics, as it primarily consists of glass.

Ceramics having similar compositions may be fabricated by different laboratory techniques, and each method results in a different distribution of flaws, opportunity for depth of translucency, and accuracy of fit. These techniques are sintering, heat-pressing, slip-casting and milling. Sintering is a process of fabricating a coherent mass from closely packed particles by heating to a specified temperature without melting; the dense strong structure is achieved as a result of bonding, diffusion, and flow phenomena 6. Two main types of dental ceramic materials are used for this technique: (1) Alumina-based ceramics, which have high strength due to dispersion of aluminum oxide (Al2O3) crystals within the glassy matrix and excellent bonding between alumina and the glass phase 4. (2)

Leucite-reinforced feldspathic porcelain, containing up to 45 volume % of tetragonal leucite crystals, which act to increase the flexural and compressive strength.

Sintered all-ceramic restorations are being gradually replaced by heat-pressed all- ceramic restorations with simplified processing steps. This technique relies on application of high pressure on a heated ceramic ingot, which is slowly pressed into a wax mold that

disappears at elevated temperatures (lost wax technique). This process can be expected to produce a well-controlled and homogeneous material without large pores, having improved dispersion of the crystalline phase within the glassy matrix, hence resulting in maximized mechanical properties 4. Pressable ceramics usually have application only as core and framework materials. Pressable veneering materials are available, but the depth of layered esthetics may be limited when using pressable ceramics for veneering materials 7. Pressable ceramic materials vary according to the type of crystalline phase and pressing temperature. Leucite-based, lithium disilicate-based and lithium phosphate- based ceramics are currently being used. The advantages of heat-pressed ceramics include good esthetics for the leucite-reinforced ceramics, high strength but higher opacity for the lithium disilicate, and the ability to use the well-known lost wax technique4.

Another technique that is also used for fabricating all-ceramic restorations is slip casting, which involves application of a porcelain slurry slip on a refractory die, where capillary action helps to remove water and densify the deposited particles. The formed piece is then fired, and a molten glass is later infiltrated into the pores by capillary action.

The resulting material will exhibit less porosity and higher toughness 4, 6. Use of this method in dentistry has been limited to one series of three products for glass infiltration

(In-Ceram, Vita Zahnfabrik, Bad Säckingen, Germany) that includes In-Ceram Alumina,

In-Ceram Spinell, and In-Ceram Zirconia which are primary used as core or framework materials. The limited application of slip casting in dentistry is probably because the

method requires a complicated series of steps, which provide a challenge to achieving accurate fit.

The evolution of computer-aided design/computer-aided manufacturing (or computer-assisted machining) (CAD/CAM) systems for the production of machined dental restorations (inlays, onlays, veneers, and crowns) has led to the development of a new generation of ceramics that are machinable. CAD-CAM ceramics are available as prefabricated ingots. These ingots are milled or cut by computer-controlled tools. The ceramic ingot may be supplied in the pre-sintered or sintered condition. In the case of pre-sintered ceramics, the ingots are porous, which enables rapid milling without bulk fracture of the ceramic. The disadvantage of pre-sintered ingots is the need for a subsequent sintering treatment to eliminate the porosity. Densely sintered ceramics are available in non-porous ingots, which are more difficult to mill, but they do not require any further sintering 7.

Successful application of all-ceramic materials depends on the ability of the clinician to select the appropriate material, manufacturing technique, and cementation or bonding procedure, to match intraoral conditions and esthetic requirements 8. No cement is considered ideal for all situations; some require multiple or technique-sensitive steps.

To achieve a good clinical outcome, the clinician needs to be familiar with the strengths, shortcomings, and handling requirements of each cement material. Malament and

Socransky [26] stated that “A luting agent might not only ʻattachʼ a restoration to a tooth

but also act as a physical shield to mask the weak qualities of dentin and other low modulus materials used as cores”.

To attach the reconstructive work to the prepared remaining tooth structure, luting cements are required. The main role of such agents is to retain the restoration on the tooth, to seal the exposed dentin, and to fill the unavoidable gap between them. The attachment between restoration/cement/tooth could be (a) mechanical, through the micro- irregularities on both restoration and tooth surfaces, (b) chemical, using a bonding technique or (3) a combination of both mechanisms. An ideal dental adhesive should provide a durable bond between the tooth and restoration, possess good compressive and tensile strengths, have sufficient fracture toughness to prevent dislodgment as a result of interfacial or cohesive failures, possess the ability to wet the tooth and the restoration, have adequate film thickness and viscosity to ensure complete seating, be resistant to degradation in the oral cavity, be biocompatible, and have adequate working and setting times 6.

There are 5 types of commercially available luting agents for the long-term cementation of fixed prostheses: zinc phosphate, zinc polycarboxylate, glass ionomer, resin composite, and resin-modified or “hybrid” glass ionomer cements. Each type is physically and chemically unique, and no luting agent is ideal for all situations 9. (1) Zinc phosphate cement does not chemically bond to any substrate and provides a retentive seal

(sealing) by mechanical means only. Thus, the taper, length, and surface area of the tooth preparation are critical to its success. (2) Zinc polycarboxylate cements exhibit chemical

adhesion to tooth structure through the interaction of free carboxylic acid groups with calcium. After hardening, these cements exhibit significantly greater plastic deformation than zinc phosphate cement and thus are not well-suited for use in regions of high masticatory forces or in the cementation of long-span prostheses. (3) Glass ionomer cements are thought to adhere to tooth structure by formation of ionic bonds at the tooth/cement interface as a result of chelation of the carboxyl groups in the acid with the calcium and/or phosphate ions in the apatite of enamel and dentin 10. The main disadvantage of glass ionomer cement is its sensitivity to water during the initial period of setting, which results in some solubility. (5) Resin modified glass ionomer (RMGI) cement or hybrid ionomer has a mainly fluroalumonisilicate powder while the liquid contains an aqueous solution of polyalkenoic acids modified with pendant methacrylate groups. The setting reaction of RMGI cement occurs in two steps. The first step is an acid-base reaction, which is the same as for the conventional glass ionomer, and the second step is either a light-cured or self-cured polymerization reaction of the pendant methacrylate groups. Therefore, two types of chemical bonding to tooth structure will occur: an ionic bond resulting from the acid-base reaction and a hybrid layer bond 4.

The RMGI cements have compressive and diametral tensile strengths greater than zinc phosphate, polycarboxylate, and some glass ionomers but less than resin composite11. These cements may have some cariostatic potential and resistance to marginal leakage. A significant disadvantage of the RMGI cements is the hydrophilic nature of polyHEMA, which results in increased water sorption and subsequent plasticity and hygroscopic expansion. This behavior is analogous to a synthetic hydrogel. The

potential for substantial dimensional change contraindicates their use with all-ceramic feldspathic-type restorations 9.

Resin-based composite cements usually consist of a bis-GMA/TEGDMA (2,2- bis[4-(2-hydroxy-3-methacryloyloxypropoxy) phenylpropane / triethyleneglycol dimethacrylate) or polyurethane matrix in which micro-filler particles (0.04-0.2µm) of quartz are embedded. Heavy metals such as zinc, barium, strontium, or yttrium are incorporated into the glass to obtain radio-opacity. For a composite to have successful properties, a good bond should be formed between the inorganic filler and the organic matrix. This bond can be achieved by coating the filler particles with a coupling agent compound during manufacturing. This compound is an organic silicon compound, termed silane, which allows formation of a chemical bond between the resin matrix and the fillers 4, 12. Resin composite cement can be polymerized through a chemically-initiated mechanism, photo-polymerization, or a combination of both.

These resin composite cements are available in various shades and opacities, and their chemistry allows adherence to many dental substrates. Adhesion to enamel occurs primarily through surface irregularities created after acid etching (usually phosphoric acid). The etchant is used to remove the smear layer from tooth preparation, preferentially dissolving the hydroxyapatite crystals. Micromechanical retention will be obtained when the fluid adhesive subsequently penetrates the surface irregularities and becomes locked after polymerization 6.

Dentin bonding is more challenging than enamel bonding as dentin contains more water and is very hydrophilic. The dentin-bonding technique includes three different processes: etching, priming and bonding. Etching is usually performed with a phosphoric acid solution or gel (37%), which causes removal of the smear layer; dentin plugs open and dentinal tubules are widened. The acid dissolves and extracts the apatite mineral phase that covers the collagen fibers of the dentin matrix and opens 20 ‒ 30 nm channels around the collagen fibers. These channels provide an opportunity to achieve mechanical retention of the subsequently placed hydrophilic adhesive monomer. A primer (such as

HEMA) is a bifunctional agent, i.e., it is both hydrophilic, which enables a bond to dentin, and hydrophobic, which enables a bond to the adhesive. When the primer is applied, it wets and penetrates the dentin surface to create “resin tags” which give micromechanical retention. The later process during priming is termed “hybrid layer formation”, and the resulting layer is termed the “resin-interpenetration zone” 4, 9. The adhesive resin is then applied to the “primed” surface to stabilize the primer-infiltrated demineralized dentin and to penetrate into the dentinal tubules. Most resin adhesives are filled with 50 ‒ 70 % by weight of glass or silica 13. However, high filler content increases viscosity, which in turn reduces flow and increases film thickness of the adhesive.

Resin composite cements exhibit high compressive strength, resistance to tensile fatigue, and are virtually insoluble in the oral environment 6. They have the ability to bond chemically to resin composite restorative materials and to silanated porcelain. Their ability to adhere to multiple substrates, high strength, insolubility in the oral environment,

and shade-matching potential have made resin composite cements the adhesive of choice for esthetic type restorations, including resin composite inlays and onlays, all-ceramic inlays and onlays, veneers, crowns, FPDs, and the newly developed fiber-reinforced composite restorations.

Due to the lack of a strong metal substructure in an all-ceramic restoration, the cement has the task of supporting this brittle material during loading. As mentioned previously, many dental luting agents are available in the market, and some are used for cementing all-ceramic restorations, including glass ionomer, resin-modified glass ionomer and adhesive resin cements. Finite element analysis and quantitative fractography of failed restorations have demonstrated that all-ceramic crowns fracture due to the extension of pre-existing surface defects that occupy the inner “fit” surface of the restoration under tensile loading. Clinically, the environment of the inner surface defects is influenced by the dental cement used to retain the restoration on the prepared tooth structure 14.

Resin-based luting materials are considered the best choice for cementation of all- ceramic inlays, onlays and full crowns, as they provide the optimum strength and good micro-mechanical bonding capability which is required as the foundation for these otherwise brittle restorative materials. Resin-based composite bonding and luting technology is considered an inherent part of the state-of-the-art of all-ceramic or all- porcelain restorations 12. Many in vitro studies 15-18 have reported the significant strengthening of dental ceramics when bonded to resin cement. Two theories have been

proposed for the apparent strengthening. Marquis [16] suggested that the resin cement modified the surface flaw population by a process of crack healing, which increased the resistance to fracture. In contrast, Nathanson [28] proposed that the polymerization shrinkage of resin cements “stresses” the molecules together, which strengthens the porcelain. However, the apparent strengthening mechanism behind the proposed theories has not been proven.

Chapter 2

Influence of Cement on the Interface Initiated Fracture Rate of All-Ceramic Restorations Under Cyclic Loading

2.1 Introduction

The cement is believed to play an important role in supporting all-ceramics restorations. Complete all-ceramic crowns and fixed partial dentures (FPD) do not have the underlying metal support of the porcelain-fused-to-metal (PFM) constructions; therefore the underlying cement/tooth combination should provide the support for these brittle materials during loading. The dynamic and repetitive forces associated with oral function provide a great challenge to any luting cement, and dislodged or inadequately retentive crowns and bridges are a common cause of failure 12, which may be disastrous if micro-movement occurs, particularly in the case of FPD abutments.

The long-term clinical survival of all-ceramic restorations continues to be a concern, especially for posterior teeth. Conrad et al. 8 performed a comprehensive review of the literature covering all-ceramic materials and systems, with respect to survival, material properties, cementation and bonding, color and esthetics, and provided clinical

recommendations for their use. From the reviewed literature, they reported survival rates for all-ceramic restorations (in percent) that ranged from 88 to 100% after 2-5 years of service and 84 to 97 % after 5-14 years of service. They also reported that most common complications associated with the clinical survival rates were fracture and debonding.

Many factors can affect the mechanical properties and fracture strength of all-ceramic restorations, such as microstructure of the ceramic material 19, the fabrication technique, the surface finish, the luting methods, and the storage conditions before loading until fracture 20, 21. Different cementation and bonding techniques have been applied to modern all-ceramic restorations. Until the last three decades, zinc phosphate cement was the most commonly used cementing agent, but due to its lack of adhesion to tooth structure, high dissolution in the mouth, potential for chemical and thermal injury to the pulp, and minimal anti-cariogenic effect 6, many efforts were placed to develop a more efficient cementing material. Different classes of cements include glass ionomer and resin-based cements, which have adhesive potential to tooth structure and most prosthetic materials.

Glass ionomers are susceptible to early water degradation, resulting in microcracks, which may initiate cracks and promote crack propagation in the cement 22. Resin- modified glass ionomer cement sets through a combination of an acid-base reaction and photo- or chemically-initiated polymerization. Combining chemical adhesion advantages of traditional glass-ionomer cements with advantages of composite resins results in improved strength, fracture toughness, and wear resistance. Zinc phosphate, zinc polycarboxylate, and conventional glass-ionomer cements set through an acid-base reaction that could have a tendency to exacerbate pre-existing flaws in ceramic

restorations due to the corrosive effect of the acid-base cement 23. Zinc phosphate cement has a high elastic modulus which may promote the effective strength of the glass–ceramic by reducing the stresses transferred to glass ceramic, but this effect can only be realized if there is intimate contact with the glass ceramic 24.

Currently, retentive or adhesive cements are very desirable. The use of the resin- based luting agent on hydrofluoric acid-etched and silanated glass-ceramic surfaces may be an important method for improving the fracture resistance of these types of ceramics

15. Many different factors may be responsible for the strengthening of resin-bonded crowns such as : 1) alteration of surface energy by an acid-etching process with improved wetting by resin , 2) improved retention to etched surfaces by creating deep involuted spaces into which resin can flow and interlock, creating micromechanical retention which improves the transfer of stress to the supporting structure, and 3) minimal concentration of free water at the ceramic/bonding-agent interface 24.

Significant differences in failure rates and probability of survival have been obtained between bonded and unbonded glass-ceramic restorations, where bonded restorations showed higher survival rates and lower risks of failure 25. Acid-etched Dicor restorations luted with resin composite exhibited more favorable survival than restorations luted with glass ionomer or zinc phosphate cement 26. Malament and Grossman 27 reported a clinical failure rate for resin-bonded Dicor crowns which was significantly lower than that previously reported for Dicor crowns luted with zinc phosphate cement. In vitro studies by Grossman and Nelson and Eden and Kacicz [both cited in Reference 24] showed an

increase in the failure load of Dicor® crowns of up to 97% when the crowns were etched with hydrofluoric acid and luted with light-activated resin-based cement compared with the use of unetched crowns luted with zinc phosphate cement.

In conventional glass-ceramic restorations, the adhesive technique is critical for successful bonding. Surface treatment of the porcelain by etching with 5% to 9.5% hydrofluoric acid and etching of the tooth structure with 37% phosphoric acid and application of a silane coupling agent provided the highest bond strength of adhesive- resin cement to feldspathic material. Two theories have been proposed for the apparent improvement in the strength of ceramic when luted with resin cement. Marquis 16 suggested that the resin cement modified the surface flaw population by a process of crack healing, which increased the resistance to fracture. In contrast, Nathanson 28 proposed that within certain limitations, resin shrinkage during polymerization exerts compressive stresses on the inner porcelain surface which causes porcelain molecules to move closer together rather than away from each other. Conversely, it may be conjectured that resin-modified glass ionomer cements, which have been found to expand, thereby increase the risk of crack propagation. Another strengthening mechanism described by Ritter and Lin 29 is the reduced stress-corrosion of the glass-ceramic attributable to the resin-coating reducing the ability of water to be transported to the crack tip, thereby lessening the slow crack growth.

The available research strongly suggests that using resin cement for bonding ceramic restorations to tooth structure is indicated. This supporting research comes from three

different sources: laboratory fracture studies comparing restorations luted with resin vs other materials, clinical studies, and laboratory studies examining the surface sealing/strengthening effect of resin on ceramic. Laboratory studies also confirm the enhanced resistance to fracture of crowns cemented with an adhesive procedure 30. Al-

Makramani et al. reported a higher fracture resistance of resin-bonded Procera All Ceram copings when compared to other conventional cements (zinc phosphate and glass ionomer) 31. Goodacre et al. 25 reported that the use of dentin-bonding agents and resin cements may improve longevity of all-ceramic crowns.

In-vitro studies for evaluating the survival rates of any dental restorations should be clinically relevant, provide meaningful information to clinicians and aid in elucidating clinically important variables. For test data to be relevant, laboratory tests should cause the same type(s) of damage observed in cases of clinical failure. Laboratory tests that create damage uncharacteristic of clinical situations provide misleading guidance to clinicians 32. Stress conditions and the subsequent failure response of brittle materials during contact loading are governed by some variables, including load, contact area, and the elastic moduli of the contacting materials. In addition, intraoral failure involves the active participation of water, a phenomenon termed “chemically-assisted crack growth” or “static fatigue”33, 34.

Finite element analysis (FEA) performed by Anusavice et al. 35 showed that for a ceramic layer uniformly supported by and bonded to a less stiff material, high tensile stresses developed in the ceramic at its interface with the cement, directly below the

loaded area. These interfacial stresses arise from strain differences in the ceramic, cement, and dentin because of the ceramic having higher elastic modulus (being much stiffer) than either the cement or the dentin. These tensile stresses are extremely sensitive to the ratios of elastic moduli between the ceramic and the cement and dentin, and to a much lesser extent the thickness of the ceramic and the cement 36.

All-ceramic restorations have a finite service life, and their failure in the oral cavity usually occurred by ceramic fracture 25, 37. The most common failure mode reported for ceramic restorations of all types is bulk fracture in the ceramic material.

Fracture of ceramics initiates at surfaces or interfaces due to stress concentration acting on microdefects that are either preexisting or developed during service 32, 37. As reported by Thompson et al. 38, different factors can be associated with crack initiation and propagation in ceramic restorations, including: (1) restoration shape; (2) material microstructure; (3) size and distribution of flaws; (4) residual stresses and stresses induced by surface treatment like polishing or thermal processing; (5) environment in contact with restoration; (6) ceramic/cement interfacial features; (7) thickness and thickness variation of restoration; (8) elastic moduli of restoration components; and

(9) magnitude and orientation of applied loads. The possible interactions among these variables complicate the interpretation of failure analysis observations.

Three different stress states can develop at least 3 different crack systems. Stress states include (1) sharp indentation stresses, (2) blunt indentation stresses, and (3) interfacial stresses (cementation surface)36. Sharp indentation (small contact area) will

results in what is called “Median-radial” crack systems which open onto the surface on unloading and the surface crack traces are clearly visible. Sharp indentation may also result in “Lateral” cracks which develop during unloading and result in surface material loss by chipping. Indenter with a bigger contact area or blunt indentation, develops a”Hertizan cone” cracks which pop-in from initial ring cracks located just outside of the contact area on the loaded surface. When surface contact stresses around a blunt indenter are kept low enough to avoid Hertzian ring and cone formation, a single crack can be developed from the cementation surface of the ceramic and propagated toward the loaded surface Initially, while still under the original load, this crack could be arrested if it encountered a large compressive stress field below the indenter. However, once this large crack exists in the ceramic part, complete failure occurs under loading from a slightly different angle or under cyclic loading conditions 32.

Evaluation of clinically failed glass-ceramic restorations demonstrated that a majority of these fractures (> 90%) were initiated from flaws and stresses originating from the adhesive resin cement interface rather than from the contact surface of the restoration itself 39-41.

Statement of Problem

Inherent mechanical properties, fabrication techniques, luting agents, and intraoral conditions are primary factors attributing to survival and longevity of all-ceramic restorations. Before doing time-consuming and costly clinical studies, preclinical in vitro studies should be conducted to evaluate the durability of these restorations. This study investigated the influence of different luting agents and cyclic loading under wet conditions on the survival rates of all-ceramic restorations.

2.2 Material and Methods

Chewing consists of high numbers of low cyclic loads. Cyclic loading and wet environment are the conditions encountered in the mouth during mastication. Therefore, these conditions must be considered during in vitro testing of dental restorations. In this study a fatigue test was developed that applies clinically relevant conditions of force and controlled cyclic loads to a simple tri-layer simulation of a model “restoration” .The model was fatigued under clinically relevant conditions of function to mimic the dental restorations under function. This dynamic model was used to develop survival curves reported in the form of percent survival as a function of number of cycles. The resulting data can be statistically compared in the same manner as actual clinical data. This model specifically addresses the interface initiated fracture mode of failure.

2.2.1 Specimen Preparation

Restoration Model

A simple tri-layer model consisting of a dental glass-ceramic plate cemented to the top of a flattened extracted human tooth (carious-free molar or premolar) was used to simulate a ceramic restoration Figure 2.1. The roots of extracted teeth stored in a 1% chloramine solution (Sigma-Aldrich Co, St. Louis, Mo) were embedded in an autopolymerizing acrylic resin (Jet Acrylic; Lang Dental Mfg Inc, Wheeling, Ill) to 2 mm below the cemento-enamel junction using 1/2 inch diameter PVC pipe about 10 mm in length. After mounting, the occlusal surfaces of the teeth were cut flat with a slow- speed diamond wheel saw (Series 15LC Diamond; Buehler Ltd, Lake Bluff, Ill) under water coolant to expose the coronal dentin. The flattened embedded teeth were placed in a small machine lathe (Unimat-SL; American Edelstaal Inc, Tenafly, NJ), and a high- speed handpiece (430 SWL; StarDental, Lancaster, Pa) with a medium-grit diamond rotary cutting instrument (5837KR; Brasseler USA, Savannah, Ga) mounted to the lathe feed table was used to machine the occlusal surface of the tooth perpendicular to the long axis. The.occlusal surface of each tooth was machined flat under water spray perpendicular to the long axis of the tooth to simulate the clinical prepared tooth surface with intact smear layer, and then stored in water until the cementation procedure.

Ceramic Materials and Fabrication

Two different glass-ceramic materials were used: a Leucite-reinforced glass- ceramic (Pro-CAD; Ivoclar Vivadent, Schaan, Liechtenstein) and a Lithium disilicate- based glass-ceramic (eMax-CAD; Ivoclar Vivadent). The ceramic materials were

supplied in premanufactured CAD/CAM blocks with a 12.5 x 14.5-mm rectangular cross- section. The bocks were sectioned into plates, with a uniform thickness of approximately

0.60 mm, using a slow-speed diamond wheel saw (Series 15LC Diamond; Buehler Ltd) and water coolant. The eMax-CAD plates were further heat-treated according to the manufacturers‟ instructions to reach the fully crystallized state. The ceramic plates were hand-ground on one side using a circular motion with 600 grit SiC/water slurry on a glass plate to remove any remaining visible saw marks and provide an approximately uniform surface finish. The specimens were ultrasonically cleaned in distilled water for 5 min following the grinding procedure This study had two main groups depending on glass- ceramic type used Figure 2.2.

Cement Groups

The cements used for each group were based on clinical recommendations.

Leucite-reinforced glass-ceramics have moderate strength but are not recommended for use with conventional cements by the manufacturers. Lithium disilicate glass-ceramics exhibit improved strength and are recommended by the manufacturers to be applied with either conventional or resin cements.

(1) eMax-CAD Restoration groups

The eMax-CAD plates were cemented to the flattened teeth using four different types of cements. According to the manufacturer instructions, eMax-CAD restorations can be cemented with either adhesive or non adhesive cement. The four cements used were: a conventional glass Ionomer cement (Ketac™ Cem Maxicap™, 3M ESPE,

St.Paul, MN), a resin modified glass ionomer (RelyX, 3M ESPE), a total-etch resin cement (Nexus®, Kerr Corp, Orange, Calif), and a self-adhesive resin cement (RelyX™

Unicem, 3M ESPE) Table 2.1.

(2) ProCAD Restoration groups

For this glass-ceramic the manufacturer recommends using only adhesive cements. Two different adhesive resin cements were selected: a total etch adhesive resin cement (Nexus® Kerr Corp, Orange, Calif), and a self-adhesive resin cement (RelyX™

Unicem, 3M ESPE). Table 2.2 shows the cements used with ProCAD ceramic and the required tooth and ceramic pretreatment. Cementation procedures were performed according to the manufacturers‟ instructions for each luting agent.

Tooth and Ceramic Pretreatment Conditions for Each Cement Group

(1) Total Etch Cement System

The teeth were etched for 15 sec with 35% phosphoric acid gel (Gel Etchant;

Kerr Corp), rinsed for 10 sec, and lightly air dried with gentle puffs of air to ensure that the dentin remained moist. The dentin adhesive (Optibond Solo Plus; Kerr Corp) was applied to teeth surface for 15 sec using a gentle rubbing motion. After air thinning, the adhesive was polymerized for 20 sec using a handheld halogen light (Optilux VCL401;

Demetron Research Corp, Danbury, Conn) with an approximate output of 500 mW/cm2.

For the ceramic plates the cemented surface was etched with 5% HF acid gel (IPS

Ceramic Etching Gel Ivoclar Vivadent) for 60 sec, then thoroughly rinsed with

air/water syringe and air dried with air stream. A silane coupling agent (Silane Primer;

Kerr Corp) was applied to the etched surface, rubbed for 15 sec and air thinned, and then protected from light exposure until cementation.

Prepared ceramic disks were then bonded to the prepared teeth. The resin cement was applied to the tooth and ceramic surface before cementation, and a light load applied to the restoration through a loading device. Resin-based cements were light-polymerized for 40 seconds using a handheld halogen light (Optilux VCL401; Demetron Research

Corp) with an approximate output of 500 mW/cm2. Specimens were then placed in water and stored for 24 hr prior to testing.

(2) Self-Etch Cement System

No tooth pretreatment was required with this system, and ceramic treatment was the same as in total etch system.

(3) Glass Ionomer Cement

No tooth or ceramic pretreatment was required. The cement was applied on the dry tooth surface, and the ceramic plate was seated with application of light load.

(4) Resin-Modified Glass Ionomer Cement

Tooth and ceramic treatments were the same as for the total etch group.

Table 2.1: Cement materials used with eMax-CAD ceramic plates

Group Cement Ceramic Cement type Tooth treatment name treatment

Ketac™ No treatment No treatment GI Glass ionomer Cem required required

Resin modified RMGI RelyX Etch/rinse/adhesive Etch/rinse/silane glass ionomer

Dual-cure, total-etch N Nexus® Etch/rinse/adhesive Etch/rinse/silane resin cement

U RelyX™ Dual-cure, self-etch No treatment Etch/rinse/silane Unicem resin cement required

Table 2.2: Cement materials used with Pro-CAD ceramic plates

Cement Name Cement type Tooth treatment Ceramic treatment

Dual-cure, total- Nexus® Etch/rinse/adhesive Etch/rinse/silane etch resin cement

RelyX™ Dual-cure, self-etch No treatment required Etch/rinse/silane Unicem resin cement

A total amount of 126 specimens were fabricated and stored in 100% humidity at

37°C in an incubator. The specimens were then placed in the loading pan, to be fatigued under cyclic loading.

Figure 2.1: Steps for sample preparation and mounting

Study groups

Pro-CAD eMax-CAD

Non Adhesive Adhesive cement Adhesive cement cement

RelyX™Unicem RelyX™Unicem RMGI Three trials Three trials One trial n= 27 n= 27 n= 9

Nexus® Nexus® Ketac™ Cem Three trials Three trials One trial n= 27 n= 27 n= 9 Figure 2.2: Diagram showing the main study groups and subgroups

2.2.2 Fatigue Device

To allow for controlled loading conditions and to assure interface-initiated radial fractures only, the ceramic surface was loaded with a 12.5 mm diameter brass ball that was coated with a vacuum-formed (Sta-Vac; Buffalo Dental Mfg Co, Syosset, NY) 0.5 mm thick layer of polyethylene (Product 31725; Buffalo Dental Mfg Co) to avoid surface contact damage. Samples were cyclically loaded at a clinically relevant magnitude (200-300 N), contact area (3-4 mm2) and frequency (1.6 Hz) in water at room temperature. The loading device Figure 2.3 was designed to fatigue 18 specimens at one time with precision spring/displacement-controlled loading forces. To simulate the clinical examination, periodically during the simulated mastication process (cyclic loading), the “restorations” were evaluated after each 0.5 million cycle interval under x10 magnification using a stereomicroscope (Nikon SMZ-1; Nikon, Tokyo, Japan), and transillumination with a fiber optic 120 W halogen light (Philips Electronics, New

York, NY), for evidence of crack initiation. The presence of interface-initiated “radial” fractures determines failure. The specimens were subjected to a total of 5 million loading cycles while in water at room temperature. The cyclic loading tests took approximately 8 wk. The number of cycles between the current and previous observation period was recorded as the restoration lifetime. Generated data were then reported in the form of

Kaplan-Meier survival curves to simulate actual clinical data, and the “lifetime” was considered in terms of cycles rather than years. Samples that did not fail during the 5 million cycles were handled as censored data.

Figure 2.3: Fatigue device used to apply cyclic loading on simulated “Model” restoration

2.3 Statistical Analysis

As noted already, the generated data were reported in the form of Kaplan-Meier survival curves to simulate actual clinical data, and the “lifetime” was considered in terms of cycles rather than years. The samples that survived the 5 million cycle test were handled as censored data in the analysis. Log-Rank was used to determine significant differences between survival curves, and a Bonferroni correction method used when more than two groups were compared.

2.4 Results

The simulated restorations were periodically examined after each half million cycles. The presence of interface-initiated “radial” fracture was considered to determine failure Figure 2.4. The radial fracture was differentiated between “Hertizan” or surface- initiated fractures in that the Hertzian fractures usually appear curved and just peripheral to the contact surface center.

eMaxCAD results

Log-Rank analysis showed a significant difference (p <0.001) between survival curves Figure 2.5. Using a Bonferroni correction method (p <.008), no significant different (p <.037) was found in the survival between both conventional cements (RMGI and GI), while a statistically significant difference was found between groups N and

RMGI (p <.008), N and GI (p <.008), U and GI p <.008), and U and RMGI (p <.008).

Between the two adhesive cements there is no significant different in the survival curves

(p <.01), however group U showed to have a high survival rate than N.

ProCAD Results

Group U was significantly higher than N (p <.001) as shown in Figure 2.6.

Figure 2.4: Interface-initiated “radial” fracture determination

Figure 2.5: Kaplan-Meier survival curves, showing influence of cement type on eMaxCAD glass ceramic survival

Figure 2.6: Kaplan-Meier Survival curves, showing influence of cement type on ProCAD

glass ceramic survival

2.5 Discussion

From the obtained results, it was shown that adhesive cements play a significant role in the survival of all-ceramic restorations. Both conventional cements (GI and

RMGI) resulted in dramatically earlier failures. The mode of failure between groups (GI) and (RMGI) was different, where specimens cemented with GI were totally debonded from the restoration without fracture and the bottom surface of the plate (cemented surface) was clean from any cement remnants. This means that the bond between glass ionomer and ceramic was very poor or did not exist. Plates bonded with RMGI were fractured. The survival of conventionally cemented specimens was much lower than for specimens bonded using adhesive cements. In a study by Rosenstiel et al.18 the investigators reported that ceramics plated cemented with glass ionomer cements were delaminated during flexural strength testing, which also showed that no bond existed between glass ionomer and ceramic. The results from the current study disagree with the

manufacturer‟s recommendation for cementing eMaxCAD crowns with conventional cements like glass ionomer. But our results totally agreed with many in vitro and in vivo studies that suggest using adhesive cementation to improve ceramic restoration survival rates 12, 14, 18, 26, 30.

Unexpectedly, the adhesive cements used in the current study did not show similar results. The reasons for the significantly higher survival rates of group (U) than group (N) were not clear, even after the experiment was repeated three times (increase of sample size to 27 in groups U and N with both ceramic types). This led to a fundamental question: Do all resin adhesive systems behave the same? The differences in survival rates could not be explained by any difference in elastic moduli, as both cements have the same value for the elastic modulus. The actual mechanism is unclear, but it might be related to the bond strength between cement and ceramic. Another suggestion might be the chemical nature and degradation behavior of each cement type, which might allow one cement to hold water or degrade more than the other and thereby further accelerate the process of slow crack growth. The change in cement support over time certainly influences the stress state of the ceramic interface and thus the subcritical crack growth

(SCG). Because many mechanisms are possible, further research is needed to obtain definitive information.

2.6 Conclusions

Within the limitations of this study, it has been shown that non-adhesive cements

(GI and RMGI) have significantly lower survival rates. We can conclude that adhesive cementation significantly improves the survival of our restoration model. The following other conclusions can also be drawn:

1. The mode of failure was different for both conventional cements, where GI-

cemented ceramic plates were totally debonded without fracture while RMGI-

cemented specimens were fractured. The need for a nondestructive test method is

evident that will provide insight about the ceramic/cement interface and capture

any difference before, during and after cyclic loading

2. Adhesively bonded specimens had higher survival rates than specimens that were

not adhesively bonded

3. One adhesive cement (Unicem) had higher survival rate than the other (Nexus).

While both cements have the same elastic modulus, the difference in survival

rates could not be explained by the elastic moduli. However, the ceramic/cement

bond strength might have an influence on this difference.

Based on the observations from this preliminary study, the specific aims for the following studies are:

1. Evaluation of the ceramic/cement interface using a nondestructive method, and

assessing the interface status before, during, and after fatiguing.

2. Evaluate the influence of resin cement on the slow crack growth resistance of

glass-ceramic

3. Measure and compare the effect of aging on ceramic/resin microtensile bond

strength using two different resin cements

Chapter 3

Nondestructive Acoustic Microscopic Evaluation of the Cement / Glass-Ceramic Interface

3.1 Introduction

Although ultrasound has been used for diagnostic applications in medicine since the mid 1940s 42, it has not developed into a mature diagnostic technique in dentistry. In the late 1960s it was shown that dental tissues can be observed using ultrasound 43.

Acoustic microscopy methods can be used for tooth structure study without the need for processing, fixing, dehydration, cleaning, staining, preparing replicas or contrasting, when compared to light electron microscopy and other conventional techniques which require sample preparation that is destructive in nature. Acoustic microscopy can also be used for evaluation of physicomechanical properties that does not require any destruction of objects, as in cases of hardness, compression, bending and breakage tests. Ultrasonic waves are practically non-harmful for the living body, which means that acoustic microscopy methods can be used as a basis in design of clinical diagnostic instruments that are safe both for patient and personnel 44.

Application of acoustic microscopy methods for dental diagnosis has been one of the most attractive clinical tools in the last few years. One of the important aspects of acoustic microscopy application in dentistry is quantitative nondestructive evaluation of local physicomechanical properties of tooth microstructure elements 44. Strong penetration, good directionality and high sensitivity of ultrasound waves allow them to be used in material inspection, including internal defects, dimensional measurements, material characterization, and interfacial assessment 45.

Ultrasonic non-destructive testing is one of the most widely used techniques by engineers to evaluate adhesive joints for the presence of defects like weak bonding between the adhesive and adherent, weak adhesive layer, cracks, voids or disbonding 46.

Valuable information is thereby obtained when evaluating important structures that involve safety risk (such as gas pipes) 45.

In dentistry, a non-invasive method to precisely evaluate and locate interface damage between dental tissue and the restorative ceramic material is highly desirable, and could provide valuable information about interface evaluation and allow understanding the relationship between interface degradation and material fracture.

Relatively few studies have reported the use of ultrasonic scanning as a non-invasive method of evaluating material interfaces in dentistry. The first attempts for using high- frequency focused ultrasound was to visualize the cement/dentin interface 47, 48, and it has shown that scanning with a focused ultrasonic probe (acoustic lens) enables one to get valuable information about the bonding conditions at the restoration–tooth interface, as

well as the condition and spatial distribution of the restorative material itself at the tooth- material interface. The early study by Ghoraeb et al. 49 showed that focused ultrasound transducer can be used to determine the thickness of amalgam restoration in an extracted tooth. Recently, Liudmila et al. 50 used the scanning acoustic microscopic (SAM) technique to evaluate the cement-dentin interface, and concluded that “this experimental approach may be effectively implemented for pre-clinical dental materials characterization, particularly for the real time monitoring of a restoration/dentine interface during long-term in vitro studies”. The authors also suggested that the basic principles of failed adhesion detection that were described in their study can be used in the development of in vivo ultrasound diagnostic techniques for the application within dental clinics.

Aim of the Study

Data from our previous study (Chapter 2) showed that glass-ceramic specimens luted with glass ionomer (GI) and resin-modified glass ionomer (RMGI) cements debonded rapidly in less than 500,000 loading cycles. The current study was an attempt to evaluate and capture any change or degradation in the interface between ceramic and cement before debonding occurs, using a nondestructive acoustic microscopy technique.

The evaluation was performed initially after specimen preparation and successively after short fatiguing cycles to capture any change in the Interface.

3.2 Material and Methods

3.2.1 Specimen preparation

“Simulated” bonded ceramic restorations were prepared as discussed in Chapter

2. In this study three different cements were used to bond the eMax-CAD glass-ceramic plates to tooth dentin, as shown in Table 3.1

Table 3.1: Three groups of cements used with eMax-CAD ceramic plates

Group Cement used

GI Glass Ionomer :Ketac™ Cem Maxicap™ (3M ESPE)

RMGI Resin modified glass ionomer: RelyX (3M ESPE)

N Adhesive Resin cement Nexus® (Kerr Corp, Orange, Calif)

3.2.2 Equipment for acoustic microscopy scanning

Acoustic images were obtained with a high-frequency (50 MHz) focused ultrasonic transducer (Panametrics, Inc. Waltham, MA) with focal length of 0.5 inch.

The acoustic microscope includes an ultrasonic pulser/receiver, precision scanner and personal computer with special software “Multiscan” (Panametrics, Inc.) to control the data-acquisition process.

3.2.3 Scanning procedure

The high-frequency focused transducer producing high-resolution imaging was utilized for scanning all samples using the non-contact immersion technique in pulse- echo mode. In this procedure water was used as a coupling medium between the transducer and the simulated ceramic specimen. This required the use of an immersion tank in which the specimen was placed, and the end of the transducer was immersed directly over the specimen surface. Prior to scanning, proper sample alignment was obtained. The transducer was focused on the flattened top ceramic surface using a Z-axis motor-driven stage with 0.01 mm precision. The incident ultrasonic beam was always normally aligned to the top sample surface by maximizing the first ultrasonic reflection signal, using an adjustable tilt table on which the specimen was placed. The transducer holder also allowed alignment in the two orthogonal planes. After alignment, the transducer was moved toward the specimen (in Z direction) to focus below the ceramic surface onto the cement interface.

To obtain information on the distribution of acoustical properties across the specimen, the lens was focused on its surface, and the ultrasonic pulses reflected from the specimen were received by the transducer (piezoelectric element) and recorded as a voltage vs. time delay. The resulting oscillogram that shows the succession of reflections is termed the A-scan, as shown in Figure 3.1. When the lens scans horizontally in one direction, the A-scan is produced at each position by the scanning line, and is then combined into a two-dimensional image termed the B-scan, where the horizontal axis represents the scanning direction and the vertical axis represents the time delay. By

scanning the lens in two perpendicular horizontal directions (X/Y scanning), which is termed C-scans, a three-dimensional (3D) ultrasonic image of an acoustical cross-section of the specimen is produced. By selecting a particular time interval (termed the “C-scan gate”), corresponding to a reflection of interest, a 3D image of the acoustical cross- section under the specimen surface can be obtained Figure 3.1.

Each specimen was automatically scanned over the entire surface at a continuous linear scanning speed of 2 mm/s and an interline scan increment of 0.01 inch. Ultrasonic scanning of the specimens was performed initially after specimen fabrication and at various intervals of short loading cycles (fatiguing), as shown in Table 3.2. The scans for each specimen were performed at the same acoustic parameters each time. The scanned data were stored in the computer for the subsequent analysis.

Table 3.2: Intervals of loading cycles at which acoustic scans performed

Group Number of cycles

GI Initial (0) 15,000 30,000 60,000 100,000 250,000

RMGI Initial (0) 15,000 30,000 60,000 100,000 250,000

N Initial (0) 15,000 30,000 60,000 100,000 250,000

3.2.4 Data analysis

ScanView software (Panametrics, Inc, ) was used for data analysis. Data were analyzed by interpretation of the A-scan, B-scan and C-scan plots, which were in the form of high resolution images for each specimen.

3.3 Results

Acoustic images obtained from various scan outputs are shown in Figure 3.1. A single point scan is represented as the A-scan or time domain format (amplitude versus time), which displays the amplitude of received ultrasonic signals as a function of time elapsed by the ultrasonic waves (TOF: time of flight). The amplitude is presented along the vertical axis, and the elapsed time is plotted along the horizontal axis. The B-scan is a plot of the signal time domain (µs) as a function of location (mm) along a single scan line. In the B-scan mode, the time of flight of the sound energy is displayed along the vertical axis, and the linear position of the scan head is displayed along the horizontal axis. The signal amplitudes are represented by a color scale as shown on the left of

Figure3.1. The C-scan represents a point by point plot along the specimen surface of a portion of the A-scan data. Each pixel represents an area of about 100 µm diameter, and the signal amplitude is represented by the color scale shown at the left.

To obtain C-scans of the interface, a signal-limiting acoustic “gate” can be specified around the signal of interest, and only that data which resides in the “gate” is

represented in the C-scan output. Placing the gate around the second (2nd) signal (solid lines) limits the C-scan to those signals that are reflected from the ceramic/cement interface (solid box) while placing the gate around the third (3rd) signal (dashed lines) displays the C-scan of the cement/tooth interface (dotted box).

Figure 3.1: Images obtained from various scan outputs

Figure 3.2: Ultrasonic beam reflections through specimen layers: 1st reflection is from nd rd surface of ceramic, 2 reflection is from ceramic/cement interface, 3 reflection is from cement/tooth interface.

As shown in Figure 3.2, the reflection of the signal from the top surface of the ceramic is represented by the waveform labeled 1st which is the first reflection arriving at the detector due to the shorter distance traveled. The 2nd waveform represents the amplitude of the signal that is reflected from the ceramic/cement interface, which arrives at the detector after the first one. The reflection from cement/tooth interface which is labeled “3rd” is the one that arrive last. The characterized ceramic/cement interfaces were obtained in the form of C-scan images in this study. Figure 3.3 shows the C-scan images of this interface (ceramic/cement) for the three different groups. The difference in signal

strength resulting from the presence of a defect (dis-bonding) in the interface is indicated as a shift in color pattern from green to purple.

Figure 3.3: C-scan images for specimens from 3 different groups: (a) GI, (b) RMGI, (c) N

The acoustic images of the initial (as-prepared) specimens were distinctly different for each cement used, as shown in Figure 3.3. Initial scanning for specimens cemented with GI showed large areas of high-reflection signals at the ceramic/cement interface, which meant areas of delamination at this interface. Specimens cemented with

RMGI showed less areas of delamination during the initial scanning, while those bonded with N resin cement did not show any delamination in the initial scanning. For each group, C-scan images were obtained for samples initially and after specific numbers of fatigue cycles, as indicated in Table 3.2.

Figure 4 (A and B) show the full range of C-scan images for two representative

samples in group (GI). Figure 3.4 (C and D) show 3D C-scan images for these samples

initially and after 250,000 cycles.

(A)

(B)

(C) (D) Figure 3.4: (A and B) C-scan images for two samples in GI group. Scanning was performed after initial preparation and successively after each fatigue cycle. (C and D) 3D C-scan images for samples A and B respectively, initially and after last fatiguing cycle (at 250,000 cycles)

Figure 3.5 shows representative C-scan images for samples in group RMGI after

different fatiguing cycles. Representative images for group N are shown in Figure 3.6.

(A) (B) (C) Figure 3.5: (A, B and C) Representative C-scan images with their corresponding 3D images for three samples in RMGI group, after preparation (initial) and after fatiguing

cycles

(A)

(B)

Figure 3.6: (A and B) Representative C-scan images for two samples in group N, after

preparation(initial), 100,000 and 250,000 cycles. Right side shows 3D C-scan image for samples (A and B), initially and after the last fatiguing cycle (at 250,000 cycles).

3.4 Discussion

An ultrasonic pulse traveling through the three specimen layers

(ceramic/cement/tooth) is modified, and the energy is reflected due to passing through layers of different acoustic impedance. When air or delamination (gap) is present between the ceramic and cement, the ultrasonic pulse will be totally reflected due to the large difference in the acoustic impedance between the ceramic and air, and this will produce an echo of high amplitude 50. Information about the area of the specimen illuminated by the ultrasonic beam is obtained through a display of the variation of the magnitude of the reflected echo with time (A-scan), which can be used to detect the area of the defect. Changes in the echo amplitude indicate the presence of a defect, and a record of defect location can be obtained by plotting amplitude against position. Such a plot is termed a C-scan.

In the current study, acoustic microcopy was used successfully to characterize the ceramic/cement interface. Employing an immersion pulse-echo technique, a wave traveling from the ceramic to the adhesive layer then to dentin was partially reflected and partially transmitted through these layers, as shown in Figure 3.2. The larger the impedance mismatch between any two layers (materials), the higher will be the reflected amplitude, which is proportional to the reflection coefficient 45. The presence of a defect containing air or another low-density substance that has a very low acoustic impedance relative to the adhesive or adherent causes the ultrasonic pulse to be almost totally reflected. After initial study of the signals yielded familiarity with the technique, the data

acquisition channels (gates) were set at the reflections corresponding to the adhesive layer (interface) between the ceramic and cement.

The C-scan consists of several thousand pixels; the color (scale from red to purple) of each pixel is linearly proportional to the average of the absolute value of the signal within the thickness of the C-scan gate. Analysis of the initial C-scan images for specimens in each group Figure 3.3 showed a variation in the ceramic/cement interface, where specimens luted with glass Ionomer cement had large defect areas (delamination) at the ceramic/cement interface. This behavior can be explained from the difference in the amplitude of the echo signal, which is shown on the color scale in this figure. When the signal amplitude is high, the color changes from green to purple in a scale as shown in

Figure 3.3. Specimens cemented to the resin-modified glass ionomer showed a smaller defect area at the interface, while those bonded to the Nexus resin cement were free from any interface defect.

Cyclic loading in water caused a change in the interface in specimens luted with

GI and RMGI, where the size of defect was found to increase gradually with an increase in the number of fatigue cycles as shown in Figures 3.4 and 3.5. The change in the delaminated areas occurred very rapidly and at small intervals of cyclic loading in specimens cemented with glass ionomer, while less change was found in specimens bonded with resin-modified glass ionomer. This may explain why the ceramic plates luted with glass ionomer were totally debonded from the cement surface without fracture while those bonded with the resin-modified glass ionomer were fractured. The present

results suggest that the bond between glass-ceramic and glass ionomer cement does not exist initially, which leads to early debonding of the ceramic plates. While the bond between ceramic plates and resin-modified glass ionomer cement may exist due to surface pretreatment for both ceramic and tooth, this bond appears to be very poor.

Fracture of the ceramic plates occurred very early under cyclic loading due to concentration of stresses at the delaminated area, and these regions were observed to grow slowly when compared to the delaminated regions in specimens cemented with glass ionomer. Specimens bonded with adhesive cement did not show any areas of delamination in the initial scanning, and only minimal changes were observed after cyclic loading.

In conclusion, we hypothesize that the difference in survival (or fracture mode) associated with the different cements is a direct result of the different character of initial delamination between the ceramic and cement. This delamination leads to increased local stress and subsequent early fracture or debonding.

The limitation of this study was the inability to characterize the cement/tooth interface, because the strength of the signal reflected was not strong enough to obtain meaningful and reliable data, which is related to the attenuation of the ultrasonic waves.

However, this technique nonetheless provided useful information about the behavior of the cement/tooth interface during cyclic loading, and in subsequent research it may enable a comparison of the difference in tooth/restoration interfaces when different luting materials are used

Chapter 4

Influence of Resin Cement on Slow Crack Growth Parameter of Two Glass- Ceramics

4.1 Introduction

All-ceramic dental restorations exhibit excellent esthetics and biocompatibility when compared to metal-based prostheses 6. However, long-term or repetitive low-level loading under water can decrease the strength of the ceramic in the service environment and lead to a failure at load levels insufficient to cause failure of the initial prosthesis.

The strength of a ceramic restoration is determined by the size of flaws present on the surface, which can be influenced over time by the amount of stress corrosion that has occurred during service in the moist oral environment and under functional loading. The practical manifestations of this fatigue effect for the ceramic are almost ubiquitous 51.

Radial cracks have been identified as a primary source of premature failure in glass- ceramic crowns 32, 37. The effect of these cracks is enhanced by the long time period of fatigue in the oral environment.

Fracture of ceramics in service occurs with little or no prior plastic deformation.

The major limitations for using ceramics widely in dentistry are their low strength and fracture toughness over long periods of time, due to their tendency for subcritical crack growth (SCG) or stress corrosion 52. SCG is a problematic fatigue behavior because subcritical growth of inherent pre-existing flaws or defects can take place until a critical size for catastrophic failure is reached. Charles and Hillig first reported that SCG causes a reduction in strength or delayed failure in ceramics due to the stress enhanced chemical reaction between the surface of the glass and the water vapor in the air 53.

Since the local stress is greatest at the ends of small cracks within the surface of the glass, the chemical reaction proceeds from the tips of these cracks, causing crack growth. The growth continues until the stress at the crack tip exceeds the ultimate strength of the glass, at which point catastrophic failure occurs.

Oxide ceramics (for example, SiO2, Al2O3 and ZrO2) are more sensitive to slow crack growth than non-oxide ceramics (for example, SiC and SiN), which is associated with the nature of the ionic bond that exist between silicon and oxygen 54. The slow crack growth process is affected by many factors such as environment, temperature, pH and glass composition 51. The classical theory that accounts for this phenomenon involves the chemical reaction between water molecules and silica, which takes place at the crack tip

53.

Slow crack growth in ceramics can be determined directly or indirectly. In the first case, crack velocity is measured directly as a function of stress intensity factors using complex specimens and sophisticated equipment. Indirect methods use simpler specimens and tests. Two indirect test techniques have been commonly used for glass and other ceramics to characterize the slow crack growth behavior 54. For the first technique, the static stress rupture test, the glass under investigation is loaded in constant stress conditions, which allows the flaw to grow. The times to failure obtained at various applied stress levels are used in evaluating the SCG parameters. This test is generally very time-consuming, as it is conducted over long periods of time (several hundred hours) 55. The second technique, the dynamic fatigue test, on the other hand is quicker and can enable the testing of a reasonable-size statistical population. In the dynamic test, the susceptibility of a material to fatigue is estimated from flexural strength tests that are performed at different, but constant, displacement rates. At high rates of loading there is little time for SCG to occur, and the calculated strength of the material approaches its inherent strength. At lower rates of loading, there is more time for stress corrosion to contribute to SCG, and the test specimens reach their critical stress intensity levels at lower applied stresses. The SGC parameters are calculated from an analysis of the effect of loading rates on strength 54, 56. The resulting variable strength values for different specimens of the same material reflect the differing population of flaws that exist in these specimens and decreasing the rate of loading effectively increases the size of the flaws, allowing some SCG to occur before failure. For accurate lifetime predictions and following ASTM Standard C1368 57, it is preferable to test the specimens under conditions of loading (dynamic) that more closely reflect the service life conditions.

Baran et al. suggested that dynamic fatigue tests can determine crack velocity parameters in brittle materials without directly measuring crack-growth rates 58. Also, these tests are preferred over fracture-mechanics-based crack propagation tests because more conservative lifetime estimates are obtained and the failure flaws more closely simulate those encountered under normal service conditions. The fatigue behavior is affected by the parameters of the subcritical crack growth (n and D) as well as by the initial strength level. Therefore, for reliable life-time predictions, it is important to evaluate both the characteristic “strength” value of a ceramic material and the SCG parameters as well 59.

Hard-coating layers provide mechanical, thermal and chemical protection to soft underlayers in biomechanical structures (such as teeth, dental crowns and hip prostheses).

Bilayer ceramic structures are subject to fatigue damage at the top (from surface contact loads) and bottom (from continuous concentrated loads) surfaces 60. In vivo and in vitro studies have demonstrated that the longevity of all ceramic restorations was markedly increased when luted with resin-based cements, compared to conventional cementation 9,

14, 26, 30. Bernal et al. reported that significantly increased loads were required to initiate the fracture of castable glass crowns cemented with resin-based cements when compared with conventional luting agents 61. Similarly, Yoshinari and Derand reported that a significant increase in loads to failure was obtained for feldspathic, heat-pressed, castable glass and glass-infiltrated crowns cemented with resin-based materials when compared with zinc phosphate-cemented crowns 62.

The strengthening mechanism when resin cement is used with ceramic restorations is not fully understood. Marquis suggested that the resin cement modified the surface flaw population by a process of crack healing, which increased the resistance to fracture 16. Fleming and Addison suggested that the infiltration of surface cracks by a surface coating partially healed and therefore reduced the crack length 14. In contrast,

Nathanson proposed that, within certain limitations, resin shrinkage during polymerization exerts a force or compressive stresses on the inner porcelain surface which causes the porcelain molecules to move more closely together 28.

Aim of Study

The reduction in the strength of the restoration during clinical service is of much concern, and both the adhesive interface and ceramic material may play a role in this process. However, the exact nature of the relationship between the cement and ceramic remains unclear. Oxide ceramic materials are known to undergo static fatigue or stress corrosion during service. The purpose of this study was to investigate the influence of the resin cement on the slow crack growth (SCG) characteristics of two glass-ceramic materials by assessing the constant-stress SCG parameters of a bilayer ceramic/resin structure. The overall objective of this project was to determine the nature of the relationship between the resin cement and SCG rates in a glass-ceramic.

4.2 Material and Methods

4.2.1 Specimen preparation

Two different types of glass-ceramics designed for the CEREC CAD/CAM system were used: leucite-reinforced (ProCAD, Ivoclar Vivadent, Schaan, Liechtenstein) and lithium disilicate (IPS e-max CAD, Ivoclar Vivadent). Ceramic plates (0.65 mm thick for e-MAX and 1.2 mm thick for ProCAD) were sectioned from the supplied blocks, using a slow-speed diamond saw (Series 15LC Diamond; Buehler Ltd, Lake Bluff, Ill) under water coolant. The e-max CAD plates were tempered according to the manufacturer‟s instructions to achieve the fully crystallized state. One surface of each specimen was finished by hand under light pressure with 600 grit SiC slurry, using a circular motion for fifty strokes. The specimens were examined at x10 magnification under a stereomicroscope (Nikon SMZ-1; Nikon, Tokyo, Japan) until the surface appeared uniform and had no evidence of visible diamond saw cut marks. If saw marks were visible, the specimens were hand-finished again for 20-50 strokes and re-examined in order to create a controlled surface flaw population. The final thickness of each specimen was measured in the center to the nearest 0.001 mm with a digital dial indicator

(Digimatic Indicator; Mitutoyo America, Aurora, Ill) having an accuracy of 0.0025 mm and a flat anvil face to avoid damaging the specimen then recorded Specimens were then ultrasonically cleaned in distilled water for 5 min.

The finished surfaces of the ceramic specimens were etched for 1 min with 5% hydrofluoric acid gel and rinsed with water for 10 sec. Prior to the cementation, plastic

spacers of 150 µm thickness were placed on the four corners of the rectangular specimens to obtain the desired resin cement thicknesses Figure 4.1. The etched surfaces were silanized (Silane Primer; Kerr Corp, Orange, Calif) and bonded with either Nexus®

(Kerr Corp) or Rely X Unicem, (3M-ESPE, St. Paul, MN) following the manufacturer‟s directions. Therefore, the bilayer specimen structure comprised the monolithic ceramic and the luting cement. Cement thickness was obtained by measuring the thickness at the center of each bilayer and subtracting the previously determined thickness of the glass-ceramic. Table 4.2 shows the mean values of ceramic and resin cement thicknesses for each group.

Figure 4.1: Diagram showing how specimens were prepared to obtain bilayer structure prior to fatigue testing. Glass-ceramic plates were bonded with resin cement using 150 µm spacers placed on corners of specimens in order to standardized film thickness of cement.

4.2.2 Experimental groups:

Specimens from each ceramic group were divided into four experimental groups as shown in Figure 4.2. The group names and the treatment applied to each group are listed in Table 4.1. Each group was divided into 3 sub-groups depending on the crosshead speed (stress rate) that was used during testing.

Each sub-group had 10-20 specimens, following ASTM specification C1368 57 which indicates that a minimum of 10 specimens per applied stress is required. All specimens from each group were tested in water at room temperature at three crosshead speeds (CHS) of (0.01, 0.1 and 1 mm/min) in displacement-controlled mode, with 10 specimens at each CHS.

Study Groups

eMax-CAD Pro-CAD

Group Group Group (O) Group (N) Group (N) Group (O) Group (N) Group (U) (E/S) (E/S)

Figure 4.2: Diagram showing test groups for each glass-ceramic type

Table 4.1: Group names with corresponding applied treatments

Group Name Treatment Final Structure

(O) No treatment Monolayer

(E/S) Samples etched and silanated only (Control group) Monolayer

(N) Etch/silane/resin-coated (Nexus resin cement film) Bilayer

(U) Etch/silane/resin-coated (Unicem resin cement film) Bilayer

Table 4.2: Mean(±SD) of ceramic and resin cement thicknesses for each group Group Ceramic thickness (mm) Cement thickness (mm) Group (O) 0.63(±.01) - Group (E/S) 0.61(±.01) - eMaxCAD Group (U) 0.61(±.01) 0.26(±.06) Group (N) 0.61(±.01) 0.24(±.04) Group (O) 1.13(±.05) - Group (E/S) 1.04(±.03) - ProCAD Group (U) 1.01(±.03) 0.20(±.03) Group (N) 1.00(±.03) 0.22(±.09)

4.2.3 Constant stress-rate testing (Dynamic fatigue)

Ceramic plate specimens were tested in the biaxial flexural mode following procedures similar to those described in the ASTM C1499 63 using a universal testing machine (Instron Model 4400). The biaxial flexural strength tests have several advantages over uniaxial tests: (1) multiaxial stress states, which provide more conservative estimates of strength, (2) no edge failure effect, and (3) biaxial loading configuration that closely simulates the loading conditions of our clinical laboratory simulation model (cyclic loading in Chapter 1)55. A “balls-on-ring” testing fixture is shown in Figure 4.3. The mechanical testing apparatus was equipped with the displacement-controlled loading mode, and slow crack growth parameters were estimated from log (flexural strength) – log (displacement rate) curves.

Figure 4.3: “Balls-on-ring” testing orientation

4.2.4 Biaxial flexural strength determination

The plates were placed on the ring of balls and a vertical load was applied using a

12.5 mm diameter spherical indenter attached to a universal testing machine (Instron

Model 4400). A piece of polyethylene tape was placed on the ceramic surface prior to indentation to reduce friction and provide a uniform distribution of load. The load was applied at the specified crosshead speed, and the load and displacement data were recorded until specimen fracture. The flexural strength for monolayer plate specimens was determined using the maximum load at fracture from the following equation, which shows the relationship between applied load and resulting stress for this geometry 63:

where: p = Failure load t = Thickness of specimen plate,

D = 0.54  (11+12)/2. where l1 and l2 are the length of the edges

 = Poisson‟s ratio a = Radius of support circle b = Radius of uniform loading at center

The flexural strength for bilayer specimens was determined from the stress in the top and bottom layers according to the following equation Rosenstiel el at. 18:

where: t = Thickness of specimen (ta= top layer; tb = bottom layer), a = Radius of supporting circle,

E = Young's modulus (Ea = top layer; Eb = bottom layer).

4.2.5 SCG parameter calculation

The specimens within each experimental group were divided into three equal subgroups (n=20). Each subgroup was tested at a different crosshead speed of 0.01, 0.1 or

1 mm/min, using a displacement-controlled mode with the universal testing machine. The load and displacement data were recorded until specimen fracture. The displacement data

(μm) was converted to time (sec) based on the assigned crosshead speed. The stress rate

(σr = dσ/dt) for each specimen was determined from the linear portion of the stress(Mpa)-time(Sec) data using linear regression analysis. By calculating the slope of the line we can get the stress rate (σr = dσ/dt). For each specimen within an experimental group, log σr was plotted against log σf (on a flexural strength-stress rate diagram). The

SCG parameters n and D were determined by a linear regression analysis using all log

[strength] values over the complete range of individual log [stress rates], based on the following equation57:

The slope of linear regression line was calculated as follows 57:

where α was the slope and K was the total number of specimens tested in a valid manner for the whole series of tests. Then the SCG parameter n was determined as follows 57:

The SCG parameter D was determined from the expression 57:

where β is the intercept on the log σr ‒ log σf plot.

4.3 Statistical Analysis

Strength distributions of quasi-brittle materials are properly described by Weibull statistics 64. The biaxial flexure strength data were arranged in ascending order, and a

Weibull analysis was performed on the resultant data 64. The basic form of the Weibull distribution is shown as

where:

m (constant) is the Weibull modulus characterizing the “brittleness” of a material. A higher value of m indicates a closer grouping of the flexure strength (σ) data.

σ0 is the normalizing constant or the characteristic Weibull stress (MPa) which is calculated at 63.21% failure probability.

The number of nominally identical brittle plate specimens used in the experiment to determine the Weibull fatigue constants (m and σ0) determined the confidence in the accuracy of these predictions 65. The confidence limits for the groups were calculated, and differences were considered to be significant when the confidence intervals did not overlap.

Regression lines were fitted to data for each group on logarithmic scales of log

(flexural strength) versus log (stress rate). The SCG parameters (n and D) were calculated according to ASTM-C1368 57.

4.4 Results

The mean biaxial flexural strengths and the results from the Weibull analysis of e-

MaxCAD plates in different groups and crosshead speeds are presented in Table 4.3.

Table 4.3: Mean biaxial flexural strength (MPa) and Weibull analysis results for different groups of e-MaxCAD plates and different CHS Group Mean Characteristic Confidence Weibull Confidence Strength strength interval (95%) modulus interval ±S.E. (MPa) for (m) (95%) for (MPa) characteristic Weibull strength modulus (O) n=30 0.01mm/min 244.89±14.3 261.64 238.01-286.17 8.06 4.38-13.24 0.1mm/min 321.56±6.63 330.53 318.05-342.67 19.55 11.22-30.50 1mm/min 316.37±14.5 333.77 307.30-360.64 9.80 5.32-15.94 (E/S) n=30 0.01mm/min 287.39±10.65 300.54 282.69-318.43 12.34 6.88-19.65 0.1mm/min 350.18±4.32 355.95 348.12-363.49 33.62 19.39-51.81 1mm/min 383.22±7.78 394.06 378.43-409.30 18.49 5.32-15.94 (U) n=60 0.01mm/min 283.19±10.14 301.38 281.58-321.13 7.89 5.26-11.19 0.1mm/min 329.11±12.59 351.96 328.24-375.84 7.20 4.90-10.01 1mm/min 351.35±15.67 379.54 347.78-412.18 5.73 3.96-7.81 (N) n=60 0.01mm/min 235.38±12.37 256.94 231.68-283.22 4.85 3.34-6.65 0.1mm/min 275.91±12.63 297.93 274.22-322.21 6.04 4.11-8.37 1mm/min 315.14±14.42 339.10 314.85-363.90 6.73 4.49-9.50

For all specimen groups, the characteristic strength (σ0) had the lowest value for all specimen groups at the slowest CHS of 0.01 mm/min and the highest value at the most rapid crosshead speed of 1 mm/min. There was no evident trend in Weibull modulus (m) with variation in CHS for the flexural strength testing. The m values observed were mostly in the range expected for ceramic materials (5-15) 66.The resulted data were

plotted on a log (flexural strength) – log (displacement rate curve) for each of the four e-

MaxCAD experimental groups, as shown in shown in Figures 4.4 – 4.7. The parameter n was determined by plotting a regression line through the unaveraged data points. The inverse slope of the regression line minus one gives the slow crack growth exponent, as shown in the equation on page 51. A positive slope corresponds to a loss in strength with decreasing test rate and is indicative of a material susceptible to slow crack growth.

Figure 4.8 shows regression plots for all four e-Max CAD experimental groups so that the differences between these groups can be seen.

eMaxCAD- no treatment [O] 2.7 2.6 2.5 y = 0.0563x + 2.4136 2.4 n= 16

2.3 Log Stress Log 2.2 2.1 2 -0.5 0 0.5 1 1.5 2 2.5 Log Stress rate

Figure 4.4: Regression line of log (flexural strength) versus log (stress rate) for

group (O) of e-Max CAD ceramic.

eMaxCAD-Etch /Silane [E/S] 2.7 2.6

2.5 y = 0.0596x + 2.4725 2.4 n= 15

2.3 Log Stress Log 2.2 2.1 2 -0.5 0 0.5 1 1.5 2 2.5 Log Stress rate

Figure 4.5: Regression line of log (flexural strength) versus log (stress rate) for group (E/S) of e-Max CAD ceramic.

eMaxCAD-Etch/Silane/Unicem [U] 2.7 2.6 2.5 y = 0.0425x + 2.4692 2.4 n= 22

2.3 Log Stress Log 2.2 2.1 2 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.6: Regression line of log (flexural strength) versus log (stress rate) for group (U) of e-Max CAD ceramic.

eMaxCAD-Etch/Silane/Nexus [N] 2.7 2.6 2.5 2.4 y = 0.0568x + 2.3928 n= 16

2.3 Log Stress Log 2.2 2.1 2 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.7: Regression line of log (flexural strength) versus log (stress rate) for

group (N) of e-Max CAD ceramic.

Figure 4.8 Analysis of covariance for eMaxCAD groups, showing that no significant difference exists between slopes (p = .80)

Table 4.4 summarizes the results of the flexural strength ‒ stress rate diagrams.

The R2 values slope α and y-intercept β for the regression lines, as well as the SCG parameters n and D, for each of the four experimental e-Max CAD groups are given.

Table 4.4: Values of R2, slope (α), y-intercept (β), and SCG parameters n and D from regression lines for four experimental e-Max CAD groups Group R2 α β n D

(O) 0.3229 0.0563 2.4136 16.85 258.82 (E/S) 0.6722 0.0596 2.4725 15.77 296.48 (U) 0.1816 0.0425 2.4692 22.80 294.44 (N) 0.2164 0.0568 2.3928 16.60 246.60

Biaxial Flexural strength (MPa) data and Weibull results were also obtained for all ProCAD experimental groups, as listed in Table 4.5. Flexural strength – stress rate diagrams using log scales were also obtained, showing the regression lines for all

ProCAD groups (Figures 4.9 - 4.13), while a summary of results from the flexural strength – stress rate diagrams are listed in Table 4.6.

Table 4.5: The mean biaxial flexural strength (MPa) and Weibull analytical results of ProCAD plates in different groups and XHS

Group Mean Characteristic Confidence Weibull Confidence Strength strength (MPa) interval (95%) for modulus interval (95%) ±S.E. (MPa) characteristic (m) for Weibull strength modulus (O) n=30 0.01mm/min 135.31±2.08 138.28 133.97-142.45 23.59 13.80-35.80 0.1mm/min 140.28±3.85 145.31 138.29-152.24 15.16 8.67-23.58 1mm/min 141.89±6.51 150.79 138.74-163.11 8.06 4.86-12.11 (E/S) n=30 0.01mm/min 136.28±3.63 140.90 134.53-147.20 16.16 9.24-24.94 0.1mm/min 162.59±3.49 167.17 160.70-173.53 18.86 10.93-28.73 1mm/min 165.55±4.44 172.50 162.18-182.72 12.21 7.07-18.98 (U) n=60 0.01mm/min 173.15±5.57 181.16 167.59-195.01 9.43 5.72-13.79 0.1mm/min 206.84±6.58 216.37 200.79-232.10 9.93 5.94-14.83 1mm/min 206.89±7.99 217.38 202.34-232.40 10.60 5.99-16.85 (N) n=60 0.01mm/min 161.78±3.63 166.25 160.49-171.81 21.70 12.02-35.35 0.1mm/min 170.29±4.44 176.05 167.34-184.47 16.13 8.88-26.02 1mm/min 193.16±5.98 200.74 186.82-214.85 11.75 6.56-18.15

ProCAD-no treatment [O] 2.45

2.35

2.25

Log Stress Log 2.15 y = 0.0071x + 2.1367 2.05 n= NA

1.95 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.9: Regression line of log (flexural strength) versus log (stress rate) for group (O) of ProCAD ceramic

ProCAD-Etch/Silane [E/S] 2.45

2.35

2.25

2.15

Log Stress Log y = 0.0404x + 2.1609 n= 24 2.05

1.95 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.10: Regression line of log (flexural strength) versus log (stress rate) for group (E/S) of ProCAD ceramic

ProCAD- Etch/Silane/Unicem [U] 2.45

2.35

2.25 y = 0.0351x + 2.2698 n= 27

2.15 Log Stress Log

2.05

1.95 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.11: Regression line of log (flexural strength) versus log (stress rate) for group (U) of ProCAD ceramic

ProCAD-Etch/Silane/Nexus [N] 2.45

2.35

2.25 y = 0.0357x + 2.2242

2.15 n= 27 Log Stress Log

2.05

1.95 -1 -0.5 0 0.5 1 1.5 2 Log Stress rate

Figure 4.12: Regression line of log (flexural strength) versus log (stress rate) for group (N) of ProCAD ceramic

Figure 4.13: Analysis of covariance for ProCAD groups, showing that no significant difference exists between slopes (p = .30)

Table 4.6: Values of R2, slope (α), y-intercept (β), and SCG parameters n and D from regression lines for four experimental Pro CAD groups Group R2 α Β n D

(O) 0.016 0.0071 (NA) 2.1367 N/A 136.77 (E/S) 0.4407 0.0404 2.1609 23.75 144.84 (U) 0.3175 0.0351 2.2698 27.49 186.12 (N) 0.4507 0.0357 2.2242 27.01 173.78

4.5 Discussion

This study was concerned about the influence of resin cement on the slow crack growth parameters of two glass-ceramic systems. Equibiaxial flexural strength for e-Max

CAD and ProCAD glass-ceramics plates was determined in accordance with ASTM standard C 1499 63. The experiment for calculating SCG parameters was conducted following ASTM standard C 1368 57, which is an indirect method for assessing slow crack growth using dynamic, controlled displacement rates (0.01, 0.1 and 1 mm/min) until failure (fracture) was observed. The amount of SCG is affected by several factors.

The environment can have a strong effect on crack growth, with aqueous environments leading to more SCG, and hence lower strength and decreased survival times.

Flexural strength was obtained as a function of stress rate in a given environment.

(All specimens from each group were tested in water at room temperature) From the results obtained, flexural strength values in each ceramic subgroup tended to decrease with decreasing displacement rate, although this decrease was not statistically significant

(Tables 4.3 and 4.4). This decrease in strength is consistent with slow crack growth behavior in ceramic materials, which is a time-dependent degradation of flexural strength

55. Flexural strength data were higher in all e-maxCAD groups than in ProCAD groups

(Tables 4.3 and 4.4), and this result was expected due to the difference in the microstructure and the amount of glassy matrix in each ceramic type 15. The e-Max CAD ceramic has approximately 70% fine-grain lithium disilicate crystals (Li2Si2O5). The fiber-like elongated lithium disilicate crystals embedded homogenously in a glass matrix

would act as an interlocking structure. This could prevent crack propagation and therefore enhance the flexural strength and fracture toughness 67.

Characteristic strength values were significantly different between the monolayer groups (O and E/S) of e-MaxCAD, where etched and silanated ceramic plates had higher flexural strength at all crosshead speeds (0.01, 0.1, and 1 mm/min) than untreated monolayer ceramic plates (group O).

The decrease in fracture stress values with the decrease in stress rate observed for both glass-ceramic materials was expected since at lower stress rates the flaw, which initiated fracture, had more time to grow, reaching larger sizes. Biaxial flexural strength values were higher in bilayer ceramic/resin plates than in monolayer ceramic plates. This result was found for both glass-ceramic types (ProCAD and e-Max CAD) at all displacement rates. At the highest displacement rate (1 mm/min), flexural strength values were significantly higher (both glass-ceramic types) for bilayer resin coated plates (U and

N) compared to monolayer plates (O and E/S). This result agrees with several earlier studies 14, 15, 18, 26 which showed that cementing glass-ceramic restorations with resin cement reduced the incidence of fracture and improved their survival rates. Rosenstiel et al. reported that the strengthening observed by resin-coating is most likely the result of bridging crack faces caused by either silane molecules entering the crack or because of strong bonding of the resin to the glass-ceramic surface, which does not allow the crack mouth to open freely 18.

Slow crack growth, static fatigue and stress corrosion are different expressions that have been used to describe the same general process of extension (growth) of subcritical cracks in ceramic materials, which may result from, but is not restricted to, such mechanisms as environmentally-assisted stress corrosion or diffusive crack growth, and this process causes a decrease in material strength. The strengths of many ceramics change with time in the presence of moisture because of slow crack growth; this is attributed to the chemical reaction occurs between water and the atomic bonds in the ceramic. In the current study the slow crack growth parameters n and D were determined by the linear regression analysis when log (ﬂexural strength) was plotted as a function of log (stress rate). The parameters (n) and (D) are estimated as constants in the flexural strength (MPa) ‒ stress rate (MPa/sec) equation, which represent a measure of susceptibility to slow crack growth of a material 57.

The SCG parameters for both ceramic materials are listed in Tables 4.4 and 4.6.

The results showed that for the e-Max CAD groups, the highest value for parameter (n) was in group (U) , but it was not significantly different from the value of (n) for the other groups (O,E/S and N). The same results was obtained in ProCAD ceramic group, but the values of the (n) parameter for groups (U) and (N) were very close and not significantly significant from other groups. The present data showed that resin cement bonded to ceramic plates does not influence the resistance of the ceramic to slow crack growth.

Results for SCG parameter n values can be compared to the results by Gonzaga, et al.68. They evaluated the n parameter for two glass ceramic systems (Empress I and II)

by dynamic fatigue method, their results showed that leucite reinforced glass ceramic has a higher n values (stress corrosion susceptibility coefficient) than lithium disilicate(30 and 17, respectively). They explained it due to the difference in the microstructure resulting from processing 69. But in our study we did not perform any microstructure analysis.

The limitations of the current study may include: small sample size and multiple steps in sample preparation (especially in e-MAX CAD group) which might induce flaws that affect on the fracture strength.

4.6 Conclusion

Within the limitations of this study and the limited sample size, it was found that biaxial flexural strength values increased after resin coating for both glass-ceramic groups. However, this thin film of resin cement does not appear to have an influence on ceramic resistance to stress corrosion. Future research should utilize the procedure of storing the resin-coated ceramic plates in water before testing, which might give more reliable information about the effect of resin cement on ceramic resistance to slow crack growth.

Chapter 5

Effect of Short and Long Term Aging Conditions on Micro-tensile Bond Strength of Resin Cement to Glass-Ceramic

5.1 Introduction

Currently, improvements in the properties of dental ceramics and bonding systems are of major clinical concern in esthetic dentistry. The influence of adhesive resin cement is believed to play a major role in strengthening these restorations 15, 18, 70.

Adhesive cementation resulted in higher survival rates for bonded ceramic crowns 26, 70.

Many mechanisms were suggested. For example, it was proposed by Nathanson, that polymerization shrinkage of resin cement may increase the apparent resistance of cemented discs by causing the molecules to move closer together, rather than away from each other28.

In order to establish a strong and durable bond, which is necessary for adequate biomechanical performance of the tooth-restoration system, appropriate treatment of the respective surfaces is crucial. A strong ceramic-resin bond relies on micromechanical interlocking and chemical bonding to the ceramic surface, which requires roughening and cleaning to obtain adequate bond strength 71. Hydrofluoric acid (HF) etching is

commonly used to condition the surface prior to bonding, and can achieve a proper surface texture and roughness 72. The mechanism for etching includes selective removal of the glassy matrix and exposure of the crystalline structures, which creates microporosities to help in mechanical retention of the resin 71. Silane coupling agents are bifunctional molecules that bond with the silicon dioxide on the ceramic surface. They also have a degradable functional group that copolymerizes with the organic matrix of the resin. Application of silane to etched ceramic surface creates a chemical covalent and hydrogen bond 73, which is considered a major factor for a sufficient resin bond to silica- based ceramics 74.

The ceramic-resin bond is susceptible to chemical, thermal, and mechanical influences during service in the oral environment. The simulation of such influences in the laboratory is important to draw conclusions on the long-term durability of a specific bonding procedure and to identify superior materials and techniques. Long-term water storage and thermocycling of bonded specimens are accepted methods to simulate aging and to stress the bonding interface 71. Most studies that apply these methods reveal significant differences between early and late bond strength values 75-77.

According to the International Organization for Standardization (ISO) 78, “Bond strength is the force per unit area required to break a bonded assembly with failure occurring in or near the adhesive/adherend interface” 79. Bond strength testing is only one of several parameters used to evaluate efficacy of adhesives; it allows gathering data and prediction of the eventual clinical outcome. In order to measure the bonding effectiveness

of adhesives to ceramics, diverse methodologies can be used currently 80. The bond strength can be measured using a macro- or micro-test set-up, basically depending upon the size of the bond area. Different methods or small modifications of the same test method produce a significance difference in bond strength values for the same product.

The main factors having influence on the results, apart from the product itself are: details of testing method, storage condition before testing, quality of substrates (especially dentin), and quality of the material and how it is handled 79. Finger 81 concluded that

“Bond strength figures should be reported with the mode of failure and the test method in order for the data to be comparable”.

The macro-bond strength, when the bond area larger than 3 mm2, can be measured in shear, tensile, or using a push-out protocol 80. The microtensile bond strength test (µTBS) is the most common micro-bond strength test. The bond area tested is much smaller compared to that of the macro-bond strength tests, being about 1 mm2 or less. In the tensile test, the bond is broken by a force acting perpendicular to the bond interface

(adhesive/adherend interface). The µTBS test is currently the most often used method to evaluate bond strength 71, 80, 82. Moreover, testing the bond strength by tensile loading produces more adhesive failures, which is relevant for the true adhesive bond strength under clinical conditions 83.

Specific aims of study

From the results obtained in some previous studies (Chapters 2 and 4), the survival of simulated bonded ceramic restorations under cyclic loading using resin cement could not be explained by dynamic fatigue testing. The results from Chapter 4 showed that resin cement has no influence on the slow crack growth parameters and does not increase the resistance of the ceramic to stress corrosion regardless of the cement type. Another experimental approach was needed to explain why one resin cement

(RelyX™ Unicem) behaved differently or had a higher survival rate than the other

(Nexus®) under the same aging condition (cyclic loading).

The first aim of this new study was to evaluate the effect of different aging conditions, thermal cycling (TC) and water storage (W), on the µTBS of leucite- reinforced glass-ceramic microbars [ceramic blocks of ProCAD (Ivoclar) bonded to two commercial composite resin cements: Nexus® (Kerr Corp, Orange, Calif) or RelyX™

Unicem, (3M ESPE, St. Paul, MN)].

The second aim was to compare the µTBS values of these two cements before and after aging.

Hypotheses to be tested:

The first hypothesis was that aging, including both thermal cycling (TC) and water storage (W), will cause a significant decrease in µTBS values regardless of the cement type used.

The second hypothesis was that there would be no significant difference in the measured µTBS between the two resin cements used before and after aging.

5.2 Materials and Methods

5.2.1 Sample preparation

ProCAD (Ivoclar Vivadent, Schaan, Liechtenstein) ceramic blocks, designed for the CEREC CAD/CAM system, were used as substrate material. Ceramic plates 5 mm thick were sectioned from these blocks as supplied, using a slow-speed diamond wheel saw (Series 15LC Diamond; Buehler Ltd, Lake Bluff, Ill) under water coolant.

The surfaces for bonding to were wet ground with 600-grit SiC slurry. Plates were then examined under a stereomicroscope (x10 magnification, Model SMZ-1; Nikon, Tokyo,

Japan) until the surface appeared uniform and had no evidence of visible diamond saw cut marks. Plates were ultrasonically cleaned for ten min in distilled water and air dried.

Ground surfaces were then etched for 1 min using 5% HF acid gel (IPS Ceramic

Etching Gel, Ivoclar Vivadent, Schaan, Lichtenstein), rinsed with water for 10 sec and air dried. Prior to cement application, a silane coupling agent (Silane Primer; Kerr

Corp) was applied to etched surfaces, which were then air dried. Two plates were cemented together with either Nexus® (Kerr Corp) or RelyX™ Unicem, (3M ESPE) following the directions of each manufacturer. Bonded samples were light-cured for 40

sec using a hand-held halogen light (Optilux VCL401; Demetron Research Corp,

Danbury, Conn) with an approximate output of 500 mW/cm2.

Figure 5.1: Schematic preparation of micro-tensile test specimen bars from prefabricated ceramic blocks and microbar fixed to modified microtensile testing device. Upper

stationary part is connected to lower articulating part through 0.35 mm thick brass sheet in back of device. Tensile force is applied to upper part via rod.

Using a low-speed diamond wire saw (Well Diamond Wire Saws, Norcross,

Ga; Model 3241), each bonded ceramic block was cut into slabs of 0.7 mm thickness, starting at the top surface of the block, through the ceramic perpendicular to the bonded interface. The cutting advanced until 1 mm remained in order to keep the slabs fixed in position. The block of slabs was then rotated 90⁰ and again cut perpendicular to the bonded interface to yield 0.7 mm2 rectangular microbars. Before the second cutting procedure, a light-body impression material (AquasilTM, DENTSPLY Caulk Milford

DE) was injected in order to hold the microbars after the second cutting procedure. After

cutting was completed, the microbars from each cement group were randomly divided into subgroups as listed in Table 5.1. Microbars were subjected to different aging conditions (water storage for 1 day, 1 week, 1 month or 1 year) and different amounts of thermal cycling (12,000, 17,000, 22,000 or 50,000 cycles) in a distilled water bath between 5°C and 55°C with 30 seconds dwell time. Before testing, the cross-sectional area (length and width) was measured for each microbar using a digital caliper.

Table 5.1: Thermal Cycling and Water Storage Conditions for Aging Number of thermal cycles Period of water storage 12,000 1 day 17,000 1 week 22,000 1 month 50,000 1 year

5.2.2 Testing device

The device for µTBS testing was specially designed by El Zohairy et al. 84 to facilitate accurate alignment of the microbar with the applied force during testing. This device consisted of two stainless steel articulating members, which were attached to each other at one end by a 0.35 mm thick brass sheet Figure 5.2. This attachment allowed hinge movement of the two parts and ensured application of a pure tensile force to the microbar specimens during testing. Values of µTBS were determined using a universal testing machine (Instron Model 4400) at a crosshead speed of .05 mm/min. Prior to

testing, each microbar specimen was glued to the testing device by means of a flowable light-curing adhesive (Helioseal, Ivoclar Vivadent).

Figure 5.2: Microtensile bond strength testing device

To determine the mode of failure, all specimens were observed immediately after fracturing under x10 magnification using a stereomicroscope (Model SMZ-1; Nikon).

The fracture surfaces were classified as follows Figure 5.3:

A = cohesive failure in ceramic B = adhesive failure at ceramic–cement interface

C = cohesive failure in cement D = mixed A and B modes E = mixed B and C modes

Figure 5.3: Failure modes for microtensile bond strength specimens

5.2.3 Statistical analysis

Two way analysis of variance (ANOVA) was performed with the bond strength as the dependent variable. The type of cement (Nexus or RelyX Unicem) and aging condition (water storage or thermal cycling) were treated as between-subject factors.

Whenever interaction or main effects were significant, they were further analyzed by the

Tukey multiple comparisons test. The level of significance was established at α = 0.05.

5.3 Results

The results are summarized in Figures 5.4 and 5.5. The means and standard deviations of the microtensile bond strength (μTBS) values and the differences between subgroups are listed in Tables 5.2 and 5.3. Results of the two-way ANOVA are presented in Tables 5.4 and 5.5, and show that cement type (p<.001), thermal cycling

(p<.001) and water storage (p<.003) significantly influenced the μTBS values. The interaction between cement type and thermal cycling conditions was significant (p<.001), and the interaction between cement type and water storage conditions was also significant

(p<.01).

For similar aging conditions (thermal cycling), significantly higher bond strengths were obtained with the RelyX™ Unicem (U) resin cement than with Nexus (N) through all thermal cycle periods except for 50,000 cycles. When comparing the effect of water storage on the bond strength within the two types of cements, it was found that both were affected similarly, when compared to the control (dry condition), but in the 1-day group microbars bonded with Nexus showed a large decrease in bond strength compared to their initial value and to similarly aged RelyX™ Unicem microbars.

Stereomicroscopic examination of the failure mode for each specimen revealed that most microbars had predominantly adhesive failure (mode B in Figure 5.3) while a few specimens showed a mixture (mode C) of adhesive failure (mode B) and cohesive failure in the cement (mode C).

Figure 5.4: Effect of water storage on µTBS values for Nexus (N) and RelyX™ Unicem (U). Mean values and standard deviations (SD) are shown in bars (both in MPa).

Table 5.2: Tukey test results for difference within groups Cement and Group µTBS(MPa) [Mean ± SD] Initial (Dry) N A 53.96 ± 14.65 Initial (Dry) U A B 52.54 ± 14.47 1 Day U B C 35.12 ± 8.00 1 Week U C 33.80 ± 20.45

1Month U B C D 33.87 ± 14.13 1 Year U C D 25.52 ± 13.92

1 Day N D 15.95 ± 9.29 1 Week N B C D 31.77 ± 6.03 1 Month N C D 21.82 ± 12.77 1 Year N D 16.81 ± 12.05)

Note: Specimen groups with different Tukey letter codes are significantly different (p < 0.05). The cement letter codes are the same as in Figure 5.4, and are also used in Figure 5.5 and Table 5.3

Figure 5.5: Effect of thermal cycling on µTBS values Nexus (N) and RelyX™ Unicem (U). Mean values and standard deviations (SD) are shown in bars (both in MPa).

Table 5.3: Tukey test results showing differences within groups Cement and Group µTBS(MPa) [Mean ±(SD)] Initial (Dry) N A 53.96 ± 14.65 Initial (Dry) U A 52.54 ± 14.47 TC(12,000 cycles) U B 31.18 ± 18.20 TC(17,000 cycles) U B 23.11 ± 10.93 TC(22,000 cycles) U B 34.29 ± 11.72 TC(50,000 cycles) U B C 30.30 ± 11.46 TC(12,000 cycles) N C 8.99 ± 8.14 TC(17,000 cycles) N C 12.62 ± 7.11 TC(22,000 cycles) N C 15.11 ± 9.07 TC(50,000 cycles) N C 13.56 ± 7.89

Table 5.4: Two-way ANOVA results for water storage Source of variation F value p- value Cement 20.36 <.001 Water storage 8.89 <.003 Cement  water storage 3.42 <.01

Table 5.5: Two-way ANOVA results for thermal cycling Source of variation F value p-value Cement 45.93 <.001 Thermal cycling 44.41 <.001 Cement x Thermal cycling 7.885 <.001

5.4 Discussion

Various methods have been reported in the literature for measuring bond strength79, 84, 85. The shear bond test is very common but it is very sensitive to the method of application of the adhesive and design of the testing arrangement. It often produces cohesive bulk fracture of the substrate away from the bonding interface, which gives only limited information about the true bond strength 83. The unpredictable modes of failure can be caused by several factors: surface flaws; internal material flaws in the substrate material, the adhesive layer, or the bonded composite; and flaws in the interfacial region.

Using the µTBS test with small specimen dimensions and small interfacial bonding zone for the samples reduces the number of these defects and results in a more uniform distribution of the applied stresses 85.

The first hypothesis was accepted as the result of the study showed a significant influence of water storage and thermal cycling on the microtensile bond strength, regardless of cement type, in which initial values for dry microbars were significantly higher than values after aging. The initial bond strength values for the two resin cements used were very close, but microbars bonded with RelyX™ Unicem cement was less influenced by the thermal cycling periods than the Nexus group. The 1 day of water storage also had a dramatic effect on bond strength for microbars bonded with the

Nexus cement, as a considerable decrease occurred after 24 hr water storage. This is attributed to a rapid degradation of Nexus compared to RelyX™ Unicem.

The result of this study could provide some explanation for the previous results obtained in Chapter 2. The higher survival rate of RelyX™ Unicem-bonded ceramic model “restorations” than Nexus-bonded model “restorations” may be attributed to the difference in the chemical structure of each material and the tendency for degradation in water. The chemical composition of RelyX™ Unicem allows an automatic change in properties from an initial hydrophilic character (providing better adaptation on the tooth surface during application) to a hydrophobic (end state) cross-linked matrix after setting, which provides a good barrier against moisture and long-term hydrolytic stability of the cement. This would ensure better bond strength over time.

Observation of the modes of failure after µTBS testing of the two resin cements bonded to ceramics revealed that the majority of the fractures were through the adhesive interface. This is in agreement with Della Bona et al. who found that most of the failures obtained from µTBS testing of composite resin bonded to hot-pressed ceramic materials occurred within the adhesion zone 86.

5.5 Conclusion

Within the limitation of this study we can conclude that both thermal cycling and water storage significantly decreased the μTBS, regardless of resin cement type.

However, RelyX™ Unicem was significantly less affected by both aging conditions than

Nexus, which might be related to the chemical composition and hydrophobic nature of the cured polymer matrix.

Chapter 6

Summary and Conclusion

Research has shown that the lack of metal substructure in all-ceramic restorations results in a high percentage of restoration failure due to ceramic fracture. New technologies in adhesive cementation help to improve the performance of these restorations and to reinforce the remaining tooth structure, although the actual mechanism of resin strengthening is currently unclear. Several studies were performed to evaluate the influence of the cement on the survival of these restorations. Under the conditions of these studies, the following conclusions could be drawn:

1. Use of resin cement increases the survival of all-ceramic restoration compared

to use of conventional cement.

2. The resin cement does not influence ceramic degradation through the slow

crack growth (SCG) process.

3. Although the actual mechanism for resin strengthening is unclear, it may

involve the formation of a more durable bond between the ceramic and

cement.

4. Acoustic microscopy is a powerful non-destructive technique that can help to

visualize tooth/ceramic interfaces in terms of sound signals and to provide

information about the interfacial degradation process.

References

1. Kawai K, Urano M. Adherence of plaque components to different restorative

materials. Oper Dent 2001;26(4):396-400.

2. Denry IL. Recent advances in ceramics for dentistry. Crit Rev Oral Biol Med

1996;7(2):134-43.

3. Giordano R. A comparison of all-ceramic restorative systems: Part 2. Gen Dent

2000;48(1):38-40, 43-5.

4. Powers JM, Sakaguchi RL, Craig RG. Craig's restorative dental materials. 12th ed.

St. Louis, Mo.: Mosby Elsevier; 2006.

5. Kelly JR. Dental ceramics: current thinking and trends. Dent Clin North Am

2004;48(2):viii, 513-30.

6. Anusavice K. Phillips‟ science of dental materials. 10th ed. Philadelphia: WB

Saunder; 1996.

7. Griggs JA. Recent advances in materials for all-ceramic restorations. Dent Clin

North Am 2007;51(3):713-27.

8. Conrad HJ, Seong WJ, Pesun IJ. Current ceramic materials and systems with

clinical recommendations: a systematic review. J Prosthet Dent 2007;98(5):389-

404.

9. Diaz-Arnold AM, Vargas MA, Haselton DR. Current status of luting agents for

fixed prosthodontics. J Prosthet Dent 1999;81(2):135-41.

10. Wilson AD, Prosser HJ, Powis DM. Mechanism of adhesion of polyelectrolyte

cements to hydroxyapatite. J Dent Res 1983;62(5):590-2.

11. White SN, Yu Z. Compressive and diametral tensile strengths of current adhesive

luting agents. J Prosthet Dent 1993;69(6):568-72.

12. Davidson CL. Luting cement, the stronghold or the weak link in ceramic

restorations? Adv Eng Mater 2001;3(10):763-67.

13. Jacobsen PH, Rees JS. Luting agents for ceramic and polymeric inlays and

onlays. Int Dent J 1992;42(3):145-9.

14. Fleming GJP, Addison O. Adhesive Cementation and the Strengthening of All-

Ceramic Dental Restorations. J Adhes Sci Technol 2009;23(7-8):945-59.

15. Pagniano RP, Seghi RR, Rosenstiel SF, Wang R, Katsube N. The effect of a layer

of resin luting agent on the biaxial flexure strength of two all-ceramic systems. J

Prosthet Dent 2005;93(5):459-66.

16. Marquis P. The influence of cements on the mechanical performance of dental

ceramics. Bioceramics 1992; 5:317-24.

17. Fleming GJ, Maguire FR, Bhamra G, Burke FM, Marquis PM. The strengthening

mechanism of resin cements on porcelain surfaces. J Dent Res 2006;85(3):272-6.

18. Rosenstiel SF, Gupta PK, Van der Sluys RA, Zimmerman MH. Strength of a

dental glass-ceramic after surface coating. Dent Mater 1993;9(4):274-9.

19. Oh SC, Dong JK, Luthy H, Scharer P. Strength and microstructure of IPS

Empress 2 glass-ceramic after different treatments. Int J Prosthodont

2000;13(6):468-72.

20. Burke FJ, Watts DC. Effect of differing resin luting systems on fracture resistance

of teeth restored with dentin-bonded crowns. Quintessence Int 1998;29(1):21-7.

21. Tinschert J, Natt G, Mautsch W, Augthun M, Spiekermann H. Fracture resistance

of lithium disilicate-, alumina-, and zirconia-based three-unit fixed partial

dentures: a laboratory study. Int J Prosthodont 2001;14(3):231-8.

22. Knobloch LA, Kerby RE, Seghi R, Berlin JS, Lee JS. Fracture toughness of resin-

based luting cements. J Prosthet Dent 2000;83(2):204-9.

23. Fleming GJ, Narayan O. The effect of cement type and mixing on the bi-axial

fracture strength of cemented aluminous core porcelain discs. Dent Mater

2003;19(1):69-76.

24. Thompson JY, Anusavice KJ. Effect of surface etching on the flexure strength

and fracture toughness of Dicor disks containing controlled flaws. J Dent Res

1994;73(2):505-10.

25. Goodacre CJ, Bernal G, Rungcharassaeng K, Kan JY. Clinical complications in

fixed prosthodontics. J Prosthet Dent 2003;90(1):31-41.

26. Malament KA, Socransky SS. 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 2001;86(5):511-9.

27. Malament KA , Grossman GD. Clinical application of Dicor® crowns: A two-

year report (Abstract). J Dent Res 1990; 69 299.

28. Nathanson D. Principles of porcelain use as an inlay/onlay material. In: Porcelain

and composite inlays and onlays: esthetic posterior restorations. Garber DA,

Goldstein RE, editors. Chicago, IL: Quintessence 1993:23-32.

29. Ritter JE, Lin MR. Effect of polymer coatings on the strength and fatigue

behaviour of indented soda-lime glass. Glass Technology 1991; 32:51-54.

30. Burke FJ, Fleming GJ, Nathanson D, Marquis PM. Are adhesive technologies

needed to support ceramics? An assessment of the current evidence. J Adhes Dent

2002;4(1):7-22.

31. Al-Makramani BMA, Razak AAA, Abu-Hassan MI. Evaluation of load at

fracture of Procera AllCeram copings using different luting cements. J

prosthodont 2008;17(2):120-4.

32. Kelly JR. Clinically relevant approach to failure testing of all-ceramic

restorations. J Prosthet Dent 1999;81(6):652-61.

33. Charles RJ. Static fatigue of glass. I. J Appl Phys 1958;29:1549-53.

34. Charles RJ. Static fatigue of glass. II. J Appl Phys 1958;29:1554-60.

35. Anusavice KJ., Hojjatie B. Tensile stresses in glass-ceramic crowns: effect of

flaws and cement voids. Int J Prosthodont 1992;5:351-8.

36. Morris DR., Kelly JR. Failure loads of bonded ceramics influenced by hydrostatic

compressive stresses. J Dent Res 1995;74:220 (abstract 1666).

37. Kelly JR. Ceramics in restorative and prosthetic dentistry. Annu. Rev. Mater. Sci.

1997;27:443-68.

38. Thompson JY, Anusavice KJ, Naman A, Morris HF. Fracture surface

characterization of clinically failed all-ceramic crowns. J Dent Res

1994;73(12):1824-32.

39. Kramer N, Frankenberger R. Leucite-reinforced glass ceramic inlays after six

years: wear of luting composites. Oper Dent 2000;25(6):466-72.

40. Fradeani M, D'Amelio M, Redemagni M, Corrado M. Five-year follow-up with

Procera all-ceramic crowns. Quintessence Int 2005;36(2):105-13.

41. Felden A, Schmalz G, Hiller KA. Retrospective clinical study and survival

analysis on partial ceramic crowns: results up to 7 years. Clin Oral Investig

2000;4(4):199-205.

42. Donald I, Macvicar J, Brown TG. Investigation of abdominal masses by pulsed

ultrasound. Lancet 1958;1(7032):1188-95.

43. S. Lees FEB. Looking into teeth with ultrasound. Science 161 (3840) 1968: 477–

78 http://www.sciencemag.org/cgi/content/abstract/161/3840/477.

44. Roman Gr. Maev LAD, Elena Yu. Maeva And Alexey A. Denissov. New data on

histology and physico-mechanical properties of human tooth tissue obtained with

acoustic microscopy. Ultrasound in Med. & Biol 2002; Vol. 28(No. 1):131–36.

45. Peter J. Shull. Ultrasound Non Destructive Evaluation: Theory, Techniques, and

Application. New york: Marcel Dekker; 2002. p. 63-7.

46. R. D. Adams BWD. Nondestructive testing of adhesively-bonded joints. NDT&E

International 1997;30(2):93-98.

47. Zheng JP ME, Denisov AA, Maev RGr. Ultrasound imaging of human teeth using

desktop scanning acoustic microscope. In: Hua L, editor. Acoustical imaging.

New York–London–Moscow: Kluwer Academic/Plenum; 2000. p. 165–71.

48. Denisova LA NN, Maev RGr, Grigorian AS, Grigoriants LA, Denisov AA. Using

an acoustic microscope for the investigation of dental cements designed for the

retrograde restoration of teeth roots. Stomatologia 2002;81(1):26-31 [in Russian].

49. Ghoraeb SR VT. Experimental evaluation of human teeth using noninvasive

ultrasound: echodentograpgy IEEE Trans UFFC 2002;49(10):1437-43.

50. Liudmila A. Denisova RGM, Valery K. Leontjev, Alexandre F. Denisov, George

G. Grayson, Fedor S. Rusanov, Evgeny Yu. Bakulin, Dmitry Yu. Gavrilov,

Tatiana V. Grineva. A study of the adhesion between dental cement and dentin

using a nondestructive acoustic microscopy approach. Dent Mater 2009;25:557–

65.

51. Gy R. Stress corrosion of silicate glass: a review. J Non Cryst.Solids

2003;316(1):1-11.

52. Zhu QS, de With G, Dortmans LJMG, Feenstra F. Subcritical crack growth

behavior of Al2O3-glass dental composites. J Biomed Mater Res B Appl

Biomater 2003;65B(2):233-38.

53. R.J. Charles, W.B. Hillig. The kinetics of glass failure by stress corrosion.

Symposium on Mechanical Strength of Glass and Ways of Improving it. USCV,

Charleroi, Belgium 1962: .

54. Munz D, Fett T. Subcritical Crack Growth. Ceramics : mechanical properties,

failure behaviour, materials selection. Berlin , New York: Springer; 1999. p. 77-

103.

55. Thompson GA. Determining the slow crack growth parameter and Weibull two-

parameter estimates of bilaminate disks by constant displacement-rate flexural

testing. Dent Mater 2004;20(1):51-62.

56. Seshadri SG, Srinivasan M, Weber GW. Evaluation of Slow Crack Growth-

Parameters for Silicon-Carbide Ceramics. J Mater Sci 1982;17(5):1297-302.

57. ASTM. C 1368 Standard Test Method for Determination of Slow Crack Growth

Parameters of Advanced Ceramics by Constant Stress-Rate Flexural Testing at

Ambient Temperature.

58. Baran G, Boberick K, McCool J. Fatigue of restorative materials. Crit.Rev.Oral

Biol.Med 2001;12(4):350-60.

59. Fischer H, Dautzenberg G, Marx R. Nondestructive estimation of the strength of

dental ceramic materials. Dent Mater 2001;17(4):289-95.

60. Zhang Y, Lawn B. Long-term strength of ceramics for biomedical applications. J

Biomed Mater Res B Appl Biomater 2004;69(2):166-72.

61. Bernal G, Jones RM, Brown DT, Munoz CA, Goodacre CJ. The effect of finish

line form and luting agent on the breaking strength of Dicor crowns. Int J

Prosthodont 1993;6(3):286-90.

62. Yoshinari M, Derand T. Fracture strength of all-ceramic crowns. Int J Prosthodont

1994;7(4):329-38.

63. ASTM. C1499 Test Method for Monotonic Equibiaxial Flexural Strength of

Advanced Ceramics at Ambient Temperature.

64. Jung YG, Peterson IM, Kim DK, Lawn BR. Lifetime-limiting strength

degradation from contact fatigue in dental ceramics. J Dent Res 2000;79(2):722-

31.

65. White SN, Li ZC, Yu Z, Kipnis V. Relationship between static chemical and

cyclic mechanical fatigue in a feldspathic porcelain. Dent Mater 1997,13(2):103-

10.

100

66. Tinschert J, Zwez D, Marx R, Anusavice KJ. Structural reliability of alumina-,

feldspar-, leucite-, mica- and zirconia-based ceramics. J Dent 2000;28(7):529-35.

67. Holand W, Schweiger M, Frank M, Rheinberger V. A comparison of the

microstructure and properties of the IPS Empress 2 and the IPS Empress glass-

ceramics. J Biomed Mater Res 2000;53(4):297-303.

68. Gonzaga CC, Cesar PF, Miranda WG, Jr., Yoshimura HN. Slow crack growth and

reliability of dental ceramics. Dent Mater 2011;27(4):394-406.

69. Gonzaga CC, Okada CY, Cesar PF, Miranda WG, Jr., Yoshimura HN. Effect of

processing induced particle alignment on the fracture toughness and fracture

behavior of multiphase dental ceramics. Dent Mater 2009;25(11):1293-301.

70. Malament KA, Socransky SS. 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

1999;81(1):23-32.

71. Blatz MB, Sadan A, Kern M. Resin-ceramic bonding: a review of the literature. J

Prosthet Dent 2003;89(3):268-74.

72. Chen JH, Matsumura H, Atsuta M. Effect of etchant, etching period, and silane

priming on bond strength to porcelain of composite resin. Oper Dent

1998;23(5):250-7.

73. Bailey JH. Porcelain-to-composite bond strengths using four organosilane

materials. J Prosthet Dent 1989;61(2):174-7.

101

74. Suliman AH, Swift EJ, Jr., Perdigao J. Effects of surface treatment and bonding

agents on bond strength of composite resin to porcelain. J Prosthet Dent

1993;70(2):118-20.

75. Roulet JF, Soderholm KJ, Longmate J. Effects of treatment and storage conditions

on ceramic/composite bond strength. J Dent Res 1995;74(1):381-7.

76. Kato H, Matsumura H, Tanaka T, Atsuta M. Bond strength and durability of

porcelain bonding systems. J Prosthet Dent 1996;75(2):163-8.

77. Hernandez AI, Roongruangphol T, Katsube N, Seghi RR. Residual interface

tensile strength of ceramic bonded to dentin after cyclic loading and aging. J

Prosthet Dent 2008;99(3):209-17.

78. ISO. 6345:1982 Adhesion-Vocabulary. Geneva, Switzerland.

79. G.Oilo. Bond strength testing-What does it mean? Int.Dent.J. 1993;43:492-98.

80. Van Meerbeek B, Peumans M, Poitevin A, Mine A, Van Ende A, Neves A.

Relationship between bond-strength tests and clinical outcomes. Dent Mater

2010;26(2):e100-21.

81. Finger WJ. Dentin bonding agents. Relevance of in vitro investigations. Am J

Dent 1988;1 Spec No:184-8.

82. J. De Munck AM, A. Poitevin, A. Van Ende, B. Van Meerbeek. Testing bond

strength: A review of the literature. Dent Mater 2010;26: e125–e57.

83. Della Bona A, van Noort R. Shear vs. tensile bond strength of resin composite

bonded to ceramic. J Dent Res 1995;74(9):1591-6.

102

84. El Zohairy AA, De Gee AJ, Mohsen MM, Feilzer AJ. Microtensile bond strength

testing of luting cements to prefabricated CAD/CAM ceramic and composite

blocks. Dent Mater 2003;19(7):575-83.

85. Pashley DH, Carvalho RM, Sano H, Nakajima M, Yoshiyama M, Shono Y. The

microtensile bond test: a review. J Adhes Dent 1999;1(4):299-309.

86. Della Bona A, Anusavice KJ, Shen C. Microtensile strength of composite bonded

to hot-pressed ceramics. J Adhes Dent 2000;2(4):305-13.

103