Citation
The Effects of Residual Stress, Viscoelastic and Thermodynamic Parameters on Apparent Fracture Toughness of Dental Bilayer Ceramic Composites

Material Information

Title:
The Effects of Residual Stress, Viscoelastic and Thermodynamic Parameters on Apparent Fracture Toughness of Dental Bilayer Ceramic Composites
Creator:
TASKONAK, BURAK ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Ceramic materials ( jstor )
Compressive stress ( jstor )
Empresses ( jstor )
Flexural strength ( jstor )
Fracture mechanics ( jstor )
Fracture strength ( jstor )
Heat treatment ( jstor )
Lawns ( jstor )
Residual stress ( jstor )
Veneers ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Burak Taskonak. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/7/2004
Resource Identifier:
56799445 ( OCLC )

Downloads

This item has the following downloads:

taskonak_b ( .pdf )

taskonak_b_Page_025.txt

taskonak_b_Page_084.txt

taskonak_b_Page_072.txt

taskonak_b_Page_102.txt

taskonak_b_Page_015.txt

taskonak_b_Page_106.txt

taskonak_b_Page_007.txt

taskonak_b_Page_053.txt

taskonak_b_Page_091.txt

taskonak_b_pdf.txt

taskonak_b_Page_122.txt

taskonak_b_Page_009.txt

taskonak_b_Page_100.txt

taskonak_b_Page_108.txt

taskonak_b_Page_020.txt

taskonak_b_Page_065.txt

taskonak_b_Page_077.txt

taskonak_b_Page_129.txt

taskonak_b_Page_092.txt

taskonak_b_Page_142.txt

taskonak_b_Page_026.txt

taskonak_b_Page_032.txt

taskonak_b_Page_022.txt

taskonak_b_Page_048.txt

taskonak_b_Page_018.txt

taskonak_b_Page_119.txt

taskonak_b_Page_134.txt

taskonak_b_Page_111.txt

taskonak_b_Page_045.txt

taskonak_b_Page_130.txt

taskonak_b_Page_120.txt

taskonak_b_Page_063.txt

taskonak_b_Page_139.txt

taskonak_b_Page_044.txt

taskonak_b_Page_114.txt

taskonak_b_Page_060.txt

taskonak_b_Page_033.txt

taskonak_b_Page_061.txt

taskonak_b_Page_005.txt

taskonak_b_Page_121.txt

taskonak_b_Page_041.txt

taskonak_b_Page_021.txt

taskonak_b_Page_138.txt

taskonak_b_Page_109.txt

taskonak_b_Page_067.txt

taskonak_b_Page_090.txt

taskonak_b_Page_118.txt

taskonak_b_Page_013.txt

taskonak_b_Page_003.txt

taskonak_b_Page_075.txt

taskonak_b_Page_082.txt

taskonak_b_Page_140.txt

taskonak_b_Page_054.txt

taskonak_b_Page_057.txt

taskonak_b_Page_094.txt

taskonak_b_Page_116.txt

taskonak_b_Page_097.txt

taskonak_b_Page_043.txt

taskonak_b_Page_027.txt

taskonak_b_Page_078.txt

taskonak_b_Page_036.txt

taskonak_b_Page_030.txt

taskonak_b_Page_019.txt

taskonak_b_Page_034.txt

taskonak_b_Page_011.txt

taskonak_b_Page_049.txt

taskonak_b_Page_113.txt

taskonak_b_Page_031.txt

taskonak_b_Page_112.txt

taskonak_b_Page_131.txt

taskonak_b_Page_058.txt

taskonak_b_Page_083.txt

taskonak_b_Page_128.txt

taskonak_b_Page_080.txt

taskonak_b_Page_016.txt

taskonak_b_Page_127.txt

taskonak_b_Page_039.txt

taskonak_b_Page_099.txt

taskonak_b_Page_103.txt

taskonak_b_Page_066.txt

taskonak_b_Page_073.txt

taskonak_b_Page_069.txt

taskonak_b_Page_004.txt

taskonak_b_Page_029.txt

taskonak_b_Page_125.txt

taskonak_b_Page_117.txt

taskonak_b_Page_135.txt

taskonak_b_Page_079.txt

taskonak_b_Page_132.txt

taskonak_b_Page_056.txt

taskonak_b_Page_064.txt

taskonak_b_Page_081.txt

taskonak_b_Page_085.txt

taskonak_b_Page_088.txt

taskonak_b_Page_074.txt

taskonak_b_Page_040.txt

taskonak_b_Page_014.txt

taskonak_b_Page_137.txt

taskonak_b_Page_055.txt

taskonak_b_Page_086.txt

taskonak_b_Page_059.txt

taskonak_b_Page_002.txt

taskonak_b_Page_096.txt

taskonak_b_Page_141.txt

taskonak_b_Page_107.txt

taskonak_b_Page_006.txt

taskonak_b_Page_017.txt

taskonak_b_Page_126.txt

taskonak_b_Page_104.txt

taskonak_b_Page_115.txt

taskonak_b_Page_071.txt

taskonak_b_Page_042.txt

taskonak_b_Page_098.txt

taskonak_b_Page_046.txt

taskonak_b_Page_133.txt

taskonak_b_Page_136.txt

taskonak_b_Page_008.txt

taskonak_b_Page_076.txt

taskonak_b_Page_093.txt

taskonak_b_Page_038.txt

taskonak_b_Page_101.txt

taskonak_b_Page_010.txt

taskonak_b_Page_001.txt

taskonak_b_Page_062.txt

taskonak_b_Page_037.txt

taskonak_b_Page_051.txt

taskonak_b_Page_047.txt

taskonak_b_Page_024.txt

taskonak_b_Page_110.txt

taskonak_b_Page_105.txt

taskonak_b_Page_070.txt

taskonak_b_Page_028.txt

taskonak_b_Page_089.txt

taskonak_b_Page_087.txt

taskonak_b_Page_023.txt

taskonak_b_Page_052.txt

taskonak_b_Page_050.txt

taskonak_b_Page_095.txt

taskonak_b_Page_143.txt

taskonak_b_Page_124.txt

taskonak_b_Page_012.txt

taskonak_b_Page_068.txt

taskonak_b_Page_123.txt

taskonak_b_Page_035.txt


Full Text












THE EFFECTS OF RESIDUAL STRESS, VISCOELASTIC AND
THERMODYNAMIC PARAMETERS ON APPARENT FRACTURE TOUGHNESS
OF DENTAL BILAYER CERAMIC COMPOSITES















By

BURAK TASKONAK


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Burak Taskonak
































This dissertation is dedicated to my loving family, Leyla, Erim and Bahadir Taskonak.















ACKNOWLEDGMENTS

There have been many persons whose empowerment gave me the strength,

confidence and motivation while working on this project. Without their presence and

support such a challenging process would not have been such a rewarding experience.

I would like to first and foremost thank Dr. Kenneth John Anusavice and Dr. John

Joseph Mecholsky, Jr., whose work ethic, compassion, support, understanding, and

guidance inspired me through this project. Second, I would like to thank my dearest

friends, Ibrahim Keklik and Deniz Rende, whose love and support were reflective of a

family.

I owe special thanks to my mother, Leyla Taskonak, who by her brilliance, vision

and courage set an astonishing example for my personal and professional growth. I thank

my father, Erim Taskonak, whose hard work, generosity, and principles taught me

truthfulness. Also, I thank my brother Bahadir Taskonak, whose presence in my life alone

is an enjoyment.

I am forever indebted to my colleagues and friends Dr. Karl-Johan Soderholm,

Dr. Chiayi Shen, Mr. Ben Lee, Ms. Allyson Barrett, Mr. Nai Zheng Zhang and Ms. Amy

Corrbitt who exemplified sensitivity to my personal needs, and whose conversations

sustained the spirit of my professional development. I thank my lovely friends and

colleagues at the University of Florida, Department of Materials Science and

Engineering, whose support and encouragement showed me the universality of

friendship.









Finally, I am thankful to my other committee members Dr. Wolfgang Sigmund,

Dr. Darryl Butt, Dr. Mark Yang, and Dr. Arthur (Buddy) Clark for their great support,

guidance and scholarship.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................ iv

LIST OF TABLES .............. .................................. .. .... ......... viii

LIST OF FIGURES ......... ............................ .. ........ ...... .......... ix

A B S T R A C T ......................................................................................... x i

CHAPTER

1 IN TR O D U CTIO N ...... .............................. ....... .. 1

2 TWO-YEAR CLINICAL EVALUATION OF LITHIA-DISILICATE-BASED
CERAM IC FIXED PARTIAL DENTURES ............................................................... 6

2.1. M materials and M ethods.................................................... .............................. 7
2 .2 R e su lts................. .................................. ....... ......... ...... 13
2 .3 D iscu ssio n ..................................................................... .............. 16

3 QUANTITATIVE FRACTURE SURFACE ANALYSIS OF CLINICALLY
FAILED CERAMIC FIXED PARTIAL DENTURES........................ ........... 19

3.1 M materials and M ethods........................................................... .......................... 20
3 .2 R e su lts................. .................................. ....... ......... ...... 2 3
3.3. D discussion .................................... ............................... ......... 27

4 ROLE OF INVESTMENT INTERACTION LAYER ON STRENGTH AND
TOUGHNESS OF CERAMIC LAMINATES ...................................................... 31

4.1. M materials and M ethods........................................................... ......................... 33
4 .2 R e su lts................. .................................. ....... ......... ...... 3 9
4 .3. D iscu ssion .............................................................. ................ .. 42

5 RESIDUAL STRESSES IN BILAYER DENTAL CERAMICS............................ 48

5.1. M materials and M ethods.................................................................. ................... 49
5.1.1. Specim en Preparation ............................ .... ................................... 49
5.1.2 Mechanical Testing Methods.......................... ...................... 51









5 .2 R e su lts............................................... 54
5 .3 D iscu ssion ..................................................... 5 8

6 OPTIMIZATION OF RESIDUAL STRESSES IN BILAYER.............. ............... 64
DENTAL CERAMIC COMPOSITES ......................................................... 64

6.1. M materials and M ethods........................................................ .......................... 65
6.1.1. Sample Preparation ................. ....... .. ... ..... ........... ....... 66
6.1.2. Compositional, Physical and Thermodynamic Property Characterization of
S p ecim en s ....................................................................... 7 1
6.1.3 Mechanical Testing Methods....................... ................ 72
6.2 R esults............................. ............... ..... 74
6.3. Discussion ......... ................ ............................... 84

7 C O N C L U SIO N S........................................................ .. .. ............ .. .............. 92

APPENDIX

A COMPOSITIONAL CHARACTERIZATION OF CERAMIC COMPOSITE
COMPONENTS ............................ ............. .. 95

B STRESS CALCULATIONS USING COMPOSITE BEAM THEORY..................... 97

C VISCOELASTIC THEORY AND FRACTURE MECHANICS METHODS FOR
DETERMINATION OF RESIDUAL STRESSES............................................. 101

C.1 Viscoelastic Theory. ............................. .................... ......... ......... ..... 101
C.2 Determination of Residual Stresses Using Fracture Mechanics ...................... 104

D APPARENT FRACTURE TOUGHNESS AND RESIDUAL STRESS DATA ...... 107

L IST O F R E FE R E N C E S ................................................... ...................................... 127

BIOGRAPHICAL SKETCH ............................................................. .............. 131
















LIST OF TABLES


Table Page

2-1 Location, survival time (in months) and failure type of Empress 2 FPDs.................. 8

2-2 Number of Empress 2 crowns according to their evaluation time............. .............. 9

2-3 Citeria for the direct evaluation of the restorations ........................................... 12

2-4 FPD restorations: Scores of the clinical evaluation (%) at baseline, year one and
year two ...... ..................................... .............. 14

2-5 Crown restorations: Scores of the clinical evaluation (%) at baseline, year one and
year two ...... ..................................... .............. 15

2-6 Scores of the clinical evaluation for plaque and gingival index (%) at baseline,
year one, and year tw o ............... ...................................... ......................... 15

3-1 Failure stress (of), residual stress (GR), critical flaw size (c), semiminor axis(a),
and semimajor axis (2b) of the FPD specimens .................................................. 27

4-1 Mean flexural strength (o),characteristic strength (oo), critical Flaw size (c) and
Weibull modulus (m) values ofbilayer ceramic laminates. .................................... 41

5-1 Mean flexural strength (o), residual stress (GR), indentation-induced crack sizes
and apparent fracture toughness (K,) of bilayer and monolithic specimens............. 55

6-1 Glass veneers and firing schedules used for sintering ............................................ 68

6-2 Heat treatment and cooling schemes for bilayer ceramic groups............................. 70

6-3 Physical, mechanical and thermal properties of core and veneer ceramics ........... 75

6-4 Mean flexural strength (o), residual stress (oR) indent crack sizes, and apparent
fracture toughness (K,) of bilayer and monolithic veneer specimens. ................... 79

A-1 Calculation of centroids for each type of core/veneer combination ..................... 99

C-l Elastic Viscoelastic Analogy ...... .............................................. .............. 103
















LIST OF FIGURES


Figure Page

1-1 Boussinesq stress field, for principal normal stresses 511, C22 and C33.. ..................... 3

2-1 Facial view of prepared maxillary left central incisor, maxillary left canine,
mandibular right central incisor and mandibular left lateral incisor......................... 10

2-2 Finished prostheses that are placed on the prepared teeth. ...................................... 10

2-3 Fractured connector in anterior Empress 2 fixed partial denture after 11 months
of clinical service. .......................................... 14

2-4 Kaplan-Meier survival statistics of Empress 2 fixed partial dentures (n=20). ......... 16

3-1 Arrest lines in the form of ridges are present on the fracture surfaces One of the
arrest lines is a result of the intersection of two fracture paths............................. 24

3-2.Wake hackle markings were used to establish the reference points for determining
the fracture origin in the glass veneer..... ............. ..... ................................ 24

3-3 Fracture propagated quickly in the glass veneer without any increase in stress from
the point where the semi-minor axis of the fracture origin ends. ........................... 29

4-1 Cross section of Empress 2 core ceramic with residual investment interaction
layer on its outer surface ............... ...... .. .... ......................... ........ .. .. ........... 39

4-2 Fracture surface of ceramic laminate (SEM view) ....................................... 42

4-3 Representative graph of flexural strength (of) versus c-1/2...................................... 44

4-4 Representative graph of flexural strength (of) versus c-1/2 for Empress 2 veneer
m onolithic specim en s......................................... .......................... ................ .. 4 5

5-1 Schematic illustration of an indented beam specimen placed in four-point flexure
Indent cracks were induced in the tension surface........................................ 52









5-2 Fracture surface and fracture origin of a monolithic veneer specimen that reveal
an indent crack as a fracture origin. ........................................ ....................... 56

5-3 SEM image of a fracture surface of an E2V specimen that exhibits a large pore as
the fracture origin.......................................................................................... 57

5-4 Theoretical calculations and experimental measurements of residual stress in
borosilicate glass resulting from viscoelastic relaxation behavior ..................... 61

5-5 Stress distribution in layers of a bilayer specimen when the outer layer is in
compression. ............................................ 62

5-6 Schematic illustration of radial and lateral cracks at the surface with or without
residual stress. .......................................................................... 62

6-1 Profiles of surface temperature versus time for the two cooling conditions........... 69

6-2 Design of tempering apparatus (Anusavice et al., 1989).................................. 69

6-3 X-ray diffraction patterns of ceramic components. ............................................... 76

6-4 Natural logarithm of shear viscosity (qs) versus inverse absolute temperature for
experimental ((Li20*2Si02 glass ceramic) core. .............................................. 78

6-5 Interfacial delamination between zirconia core and a glass veneer layer ............... 80

6-6 A fracture surface of a specimen where the fracture origin is lost........................... 82

6-7 Residual stress as a function of heat treatment temperature in bilayer
lithia-disilicate core/glass veneer ceramic composites. .......................................... 87

6-8 Fracture initiation from a median (radial) indent crack...................... ............. 91

6-3 X-ray diffraction patterns of ceramic components. ............................................... 96

A-1 Distances for the calculation of centroid of the composite parts............................. 99

A-2 Distances from the centroidal-axis and the centroid of the composite parts .......... 100















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE EFFECTS OF RESIDUAL STRESS, VISCOELASTIC AND
THERMODYNAMIC PARAMETERS ON APPARENT FRACTURE TOUGHNESS
OF DENTAL BILAYER CERAMIC COMPOSITES


By

Burak Taskonak

August 2004

Chair: Kenneth J. Anusavice
Cochair: John J. Mecholsky, Jr.
Major Department: Materials Science and Engineering

Bilayer dental ceramic composites used for fixed partial dentures are becoming

more widely used in dental practices because of their biocompatibility, aesthetic

properties, and chemical durability. However, large statistical variations in the strength of

ceramics are associated with the structural flaws as a result of processing and complex

stress states within the surfaces of the materials because of thermal properties of each

layer. In addition, partial delaminations of the veneer layer and connector fractures of

bilayer ceramic fixed partial dentures (FPDs) have been observed in a clinical study

which is a part of this dissertation. Analysis of fracture surfaces of failed FPDs reveals

that such fractures of the veneering ceramic are most likely caused by lateral crack

growth. Global residual stresses associated with the coefficient of thermal expansion

differences between core and veneering ceramics can cause lateral crack initiation. Also,









rapid cooling of bilayer ceramics from the sintering temperature of the glass veneer may

not allow the interfacial stresses in the viscoelastic glass to relax to equilibrium values.

This can further contribute to the propagation of lateral cracks. Furthermore, local

residual stresses that develop in the plastic deformation zone below sharp contact areas

on the occlusal surface are another contributor to lateral crack growth. Superposition of

global residual stresses and a Boussinesq stress field can incrementally increase the

possibility of lateral crack growth. The long-range goals of this study are to critically

analyze the lateral crack growth mechanisms associated with residual stresses, to modify

residual tensile stress distributions by controlled heat treatment, and to minimize the

probability of veneering ceramic fractures.

Four approaches were used to accomplish these goals: (1) clinical evaluation of a

bilayer ceramic fixed partial denture system; (2) fracture surface analysis of clinically

failed FPDs; (3) determination of residual stresses using fracture mechanics techniques;

and (4) optimizing residual stresses using heat treatment methods.

This study suggests that the compressive global residual stresses within the

ceramic surface can strengthen the material; however, excessive compressive residual

stresses can cause lateral cracks to grow and propagate to the surface, which will

eventually cause failure of the material. When a glass layer is used in a bilayer ceramic

composite, heat treatment above and below the glass transition temperature (Tg) of this

glass will induce different magnitudes of stresses within the surface of the material. This

phenomenon can be used to modify the residual stresses and reduce the risk for fracture.














CHAPTER 1
INTRODUCTION



The use of all-ceramic systems by dental laboratory technicians and dentists has

increased because of their enhanced aesthetic potential compared with metal-ceramics

(Piddock and Qualtrough, 1990). However, partial delamination of the veneering ceramic

in bilayer fixed partial dentures (FPD) has been reported in clinical studies by Sorensen

and Cruz (1998), Pospiech et al. (2000), and Edelhoff et al. (2002).

Residual stresses play an important role in the reliability and failure of dental

materials, especially in bilayer and trilayer ceramic and glass composites. Two classes of

residual stresses are important, global and local residual stresses. Global residual stresses

are usually introduced during fabrication by cooling the external surface of a ceramic

structure faster than the interior of the structure. If designed properly, the thermal

expansion and contraction mismatch between adjacent layers of a bilayer composite

should produce global residual compressive stress within the surface of the desired region

in a ceramic prosthesis. These global residual compressive stresses generally increase the

effective strength of the component. Local residual stresses are present at indentation or

impact sites. These stresses can be described using a Boussinesq stress field surrounding

the contact event. This stress distribution is considered local because the magnitude of the

stress decreases rapidly in magnitude and is concentrated only around the contact area.

The residual stresses remain because plastic (inelastic) deformation occurs as a result of









the contact load. Local residual stresses are both tensile and compressive in nature and

the tensile stresses control the fracture initiation and propagation processes.

Lateral cracks can initiate near the base of the plastic deformation zone below the

point of contact loading and they spread laterally on a plane approximately parallel to the

specimen surface (Marshall et al., 1982, Lawn and Swain, 1975; Lawn and Wilshaw,

1975). In a severe contact event these cracks can propagate to the surface to cause

localized ceramic fragmentation. Lawn and Swain (1975) and Lawn et al. (1980) reported

that lateral crack extension occurs during indenter unloading in much the same way as

radial cracks are produced by sharp-contact fracture. Residual stress is the primary

driving force for lateral cracking (Marshall et al., 1982) that can result in spallation.

If the two materials in a bilayer composite are crystalline (or polycrystalline), then

the primary behavior is elastic, the global residual stress is well defined, and the analysis

is straightforward. However, if one or both of the materials is amorphous, such as glass

veneer, the glass will exhibit viscoelastic behavior and the determination of the global

residual stress becomes much more complex and dependent on cooling history relative to

the glass transition temperature, Tg. In addition, the local stress field surrounding the

contact site is affected by residual stresses and complex crack characteristics (locations

and geometry). Associated with the indentation or impact site are two types of crack

geometries, radial and lateral. The radial cracks generally control strength and the lateral

cracks control the potential for spallation.

An axially symmetric Boussinesq stress field can be superimposed on residual

stresses during contact at the point of loading (Fig. 1-1). Superposition of global residual

stresses associated with viscoelastic relaxation or rapid cooling of the core/veneer









ceramic composites on the Boussinesq contact stress field can lead to early lateral

fracture of the material. These conditions will allow a better understanding of these

failure mechanisms and the reasons for spallation from the surface of all ceramic

prostheses. In view of the key role played by lateral fracture in the spallation and wear

processes (Lawn and Swain, 1975) for brittle substrates, it is imperative that the

mechanisms) of lateral cracking be analyzed to accurately estimate the survival time of

ceramic dental prostheses.


a) (b)






7-"
0.0005 .002



0.002












contours, side view (Johnson, 1985).
-0.008



Figure 1-1. Boussinesq stress field, for principal normal stresses o51, 522 and 533, (a)
Stress contours, half surface view (top) and side view (bottom). (b) Stress
contours, side view (Johnson, 1985).

The specific aims of this study are as follows:

Specific Aim 1: To determine the two year clinical survivability of a bilayer ceramic FPD

system.









This study was designed to evaluate the clinical performance of Empress 2 system

over a two-year clinical period

Specific Aim 2: To characterize the fracture features and failure stresses of clinically

failed FPDs.

Specific Aim 3: To test the hypothesis that the interaction of a core ceramic with

investment material can significantly reduce the flexural strength and the fracture

toughness of core / veneer ceramic laminates.

Specific Aim 4: To investigate the magnitude of residual stresses in bilayer ceramic

composites and determine their effects on apparent fracture toughness. The hypothesis of

this aim is that the global residual stresses within the surface of veneered all ceramic

fixed partial dentures (FPD) are responsible for both the observed strength increase in

bilayer dental ceramics and for partial spallation of the veneering ceramic.

Specific Aim 5: To determine the threshold levels of global compressive residual stress

that will strengthen the core/veneer dental ceramic composite and minimize the risk for

lateral crack propagation. The objective of this aim is to demonstrate that a compressive

residual stress can be selectively distributed in bilayer dental ceramics to strengthen the

material without causing lateral crack propagation within the surface.

Two fracture mechanics approaches were designed to analyze residual stresses in

this study: determination of residual stresses using fracture surface analysis, and

determination of residual stresses using an indentation technique.

Chapter two will demonstrate the clinical failure source of all ceramic FPDs.

Chapter three will identify the major cause of these clinical failures using fractographic

techniques. Chapter four will show that investment material that is used in bilayer all






5


ceramic FPD fabrication is not the cause of failures, nor does it contribute significantly to

the failures. Chapter five will prove that there is a significant level of residual stress that

can increase strength and at can also cause spallation failures. Chapter six will

demonstrate how residual stress can be controlled to potentially improve clinical

reliability.














CHAPTER 2
TWO-YEAR CLINICAL EVALUATION OF LITHIA-DISILICATE-BASED
CERAMIC FIXED PARTIAL DENTURES



Dental ceramics are known for their natural appearance and long-term color

stability. Most recently, the popularity of all-ceramic systems has increased because of

their enhanced aesthetic potential compared with metal-ceramics (Piddock and

Qualtrough, 1990). However, dentists recognize their limited fracture resistance, potential

abrasivity, and variations in marginal integrity (Anusavice, 1989). These concerns have

led to the development of new dental ceramic restorative materials and techniques.

Despite the excellent translucency compared with traditional metal-ceramic systems,

ceramic systems still have a limited long-term fracture resistance, especially when they

are used in posterior areas or for fixed partial dentures (Sjogren et al., 1999). Compared

with traditional metal-ceramic restorations, ceramic dental prostheses, including crowns

and FPDs, are attractive to the dental community because of their superior esthetics and

biocompatibility.

Fabrication of ceramic dental crowns is challenging because exceptional skills of

a technician are required to provide minimal stress concentration areas and accurate

marginal fit (Kelly et al., 1996). In addition, ceramic crowns must be translucent and

resistant to fracture even in clinical situations where inadequate thickness precludes

optimal design. Natural translucency is needed to achieve an appearance similar to that of

human teeth. The core of ceramic prostheses can be fabricated from feldspathic porcelain,









aluminous porcelain, lithia-disilicate-based ceramic, glass infiltrated magnesia aluminate

spinel, glass-infiltrated alumina, and glass-infiltrated zirconia, (Campbell and Sozio,

1988). However, poor resistance to fracture has been a limiting factor in their use,

especially for long-span or multiunit ceramic FPD prostheses (Suarez et al., 2004).

Lithia-disilicate (Li20-2Si02) based IPS Empress 2 (Ivoclar-Vivadent, Schaan,

Liechtenstein) is one of the all-ceramic systems that was developed in response for the

high demand for all-ceramic materials for FPDs. However, long term clinical studies with

the Empress 2 system are required to determine whether they can serve as a feasible

replacement for metal-ceramic systems. Sorensen et al., (1998) reported 3% fracture rate

in a clinical study based on observation of 60 three-unit IPS Empress 2 FPDs over a 10

month period. Pospiech et al., (2000) reported one failure due to fracture in a clinical

study with 51 FPDs after one year of service. The aim of this study was to determine the

clinical survivability of IPS Empress 2 crowns and FPDs in a two-year period using

modified U.S. Public Health Service evaluation criteria (Ryge et al., 1973, 1980).

2.1. Materials and Methods

A total of 20 crowns and 20 FPDs placed in an experimental population of 15

patients (3 men and 12 women, ages 21-59) were analyzed in this study. Crown

placements in 20 of the cases were indicated because of existing crowns associated with

secondary caries, apical lesions, fracture or lack of esthetics. Indications for FPDs

included replacement of an incisor or a first premolar tooth, or an inadequate existing

anterior FPD. Three unit FPDs were fabricated following a design that requires a

minimum thickness of 3.5 mm buccolingually, occlusogingivally and mesiodistally in the

connector areas. The locations, survival time, and evaluation time (in months) of the

FPDs as well as failure types are shown in Table 2-1. In addition Table 2-2 shows the









distribution of the Empress 2 crowns according to their evaluation time (in months). Six

of the prepared teeth had received endodontic treatment. For those that were

endodontically treated, three were abutments for FPDs and three were restored with cast,

post and core (cosmopost, Ivoclar AG, Liechtenstein).


Table 2-1. Location, survival time (in months) and failure type of Empress 2 FPDs.


Location Survival Time (mo) Failure Type
31-33 9 Veneer fracture (chipping)
23-25 10 Connector fracture
21-23 11 Connector fracture
33-35 11 Connector fracture
21-23 13 Veneer fracture (chipping)
42-31 19 Connector fracture
11-22 20 Intact
23-25 20 Intact
21-23 20 Intact
21-23 21 Intact
12-11 22 Connector fracture
11-22 22 Connector fracture
13-15 22 Intact
41-43 22 Intact
23-25 23 Connector fracture
13-15 23 Connector fracture
12-21 24 Intact
13-15 24 Intact
33-35 25 Intact
41-32 27 Intact









Table 2-2. Number of Empress 2 crowns according to their evaluation time


il tie ( Number of Crowns
Survival time (mo) .
Incisors Premolars Molars Total
19 2 2
20 1 1 1 3
21 2 1
22 1 2 3
23 2 1 3
24 6 2 8


The tooth preparations consisted of a shoulder finish line with rounded, smooth

contours, avoiding sharp angles (Fig. 2-1) to obtain maximum fit of the finished

protheses. To optimize the load carrying capacity of the ceramic prostheses and to

maximize esthetics, a shoulder width of -1.5 mm was obtained. Occlusal reduction of the

prepared tooth was -2 mm. Fine diamond burs were used for final tooth contouring and

finish line enhancement. The smoothness of the finish line and the ability to transfer the

details to the refractory die are essential for the precision and the fit of the coping. To

ensure high definition on impressions, retraction cord (Stay-put; Roeko, Langenau,

Germany) was used. Polyvinylsiloxane impression material (Extrude, Kerr, Romulus, MI,

USA) was used for complete arch recordings. Temporary crowns and FPDs were

prepared to maintain gingival health and to maintain tooth position. All prostheses were

prepared by the same certified dental technician using a layering technique. Margin

integrity and periodontal health were recorded after cementation. Both gingival margins

and the prostheses finish line were clinically excellent at the cementation appointment

(Fig. 2-2).





























Figure 2-1. Facial view of prepared maxillary left central incisor, maxillary left canine,
mandibular right central incisor and mandibular left lateral incisor for a
Empress 2 fixed partial denture.


Figure 2-2. Finished prostheses that are placed on the prepared teeth as shown in Fig. 2.1.
Labial view of two Empress 2 fixed partial dentures for a maxillary left
central incisor, maxillary left lateral incisor, maxillary left canine, mandibular
right central incisor, mandibular left central incisor, and mandibular left lateral
incisor region at baseline.









After try-in, the internal surface of the ceramic prosthesis was etched (5% HF, IPS

ceramic etching gel, Vivadent, Schaan, Liechtenstein) for 60 s, rinsed, dried and silanated

for 60 s (Monobond-S, Vivadent, Schaan, Liechtenstein). Prepared tooth surfaces were

conditioned with 37% H3PO4 (Email Preparator GS; Vivadent) for 30 s. Syntac dentin

primer (Vivadent) and Syntac dentin adhesives were applied to rinsed and partially dried

dentin surfaces. Subsequently, Heliobond (Vivadent) bonding medium was brushed on

the dentin surface and the internal surface of the prosthesis. Variolink II low viscosity

(Vivadent) luting composite catalyst and base were selected and mixed according to the

color of dentin. Cementation was performed immediately after coating the internal

surface of the prosthesis with luting agent. Excess cement was removed using a thin

brush, explorer and dental floss respectively. Luting agent was polymerized from each

surface using UV light for 60 s. Occlusion and articulation were controlled after

cementation. Clinical procedures were performed by the same clinician for all the

restorations.

Each restoration was evaluated two days after cementation (baseline), and

thereafter one year and two years. Evaluations were performed by two clinicians using a

mirror, explorer and intraoral photographs. Agreement between the two clinicians was

95%. United Public Dental Health (USPHS) criteria were used to evaluate the quality of

the restorations (Table 2-3). Disagreement was resolved by consent.









Table 2-3. Citeria for the direct evaluation of the restorations


Category Score Criteria
Anatomic
omc Alpha Restoration is continuous with tooth anatomy
form
Slightly under- or overcontoured restoration; marginal ridges
Bravo slightly under-contoured; contact slightly open (may be self-
correcting); occlusal height reduced locally
Restoration is undercontoured, dentin or base exposed;
Charlie contact is faulty, not self-correcting; occlusal height reduced;
occlusion affected
Marginal Alpha Restoration is continuous with existing anatomic form,
adaptation explorer does not catch
Bravo Explorer catches, no crevice is visible into which explorer
Bravo *
will penetrate
Charlie Crevice at margin, enamel exposed
Color
match and
ach ad Alpha Excellent color match and smooth surface
Surface
texture
Bravo Good color match and slightly rough or pitted surface
Chae Slight mismatch in color, shade or translucency and rough
surface, cannot be refinished
No evidence of caries contiguous with the margin of the
Caries Alpha restoration
restoration
Bravo Caries is evident contiguous with the margin of the prosthesis
Post
operative Alpha No sensitivity
sensitivity
Bravo Slight sensitivity


Kaplan-Meier statistics were used to analyze the survival rates of the restorations

(Kaplan and Meier, 1958). Alpha, Bravo and Charlie rankings were recorded and percent


distributions were analyzed for each year.









2.2. Results

Patients were evaluated at the recall appointments. Of the 20 Empress 2 FPDs

evaluated, 50% were rated satisfactory and 100% of the 20 crowns were satisfactory.

Distributions of the scores of the evaluated variables, color and surface, anatomic form,

marginal integrity, and postoperative sensitivity are presented in Tables 2-4, 2-5, and 2-6.

Fractures in the connector area of the five FPDs (25%) were recorded at the one year

recall exam (Fig. 2-3). In addition, five more fractures (25%) were observed in the

remaining FPDs in the second year. Eight (40%) fractures occurred in the connector

areas. Additionally, two local chipping failures (10%) were observed in the FPDs. Even

though the score for surface texture was 85% Alpha at the baseline exam, it decreased to

80% Alpha at the end of the first year and to 60% at the end of the second year (Table 2-

4). A significant difference was not observed in the scores of anatomic form, caries, and

sensitivity for the FPDs. All were given an Alpha rating. However, the score for marginal

integrity significantly decreased from an 85% Alpha score to a 54% Alpha score (Table

2-4).

Fractures were not observed in any of the Empress 2 crowns. Crown restorations

were rated an 80% Alpha score for color and surface parameter at the baseline exam. This

value decreased to 65% at the second year recalls (Table 2-5). Marginal adaptation

received a 70% Alpha score at baseline and decreased to a 40% Alpha score by the end of

first year and to a 25 % Alpha score by the end of second year (Table 2-5). The crowns

were not associated with secondary caries during the two-year evaluation period. Even

though tooth sensitivity was rated a 95% Alpha score at the baseline exam, it increased to

a 100% Alpha score by the end of second year (Table 2-5). There was no significant









difference between the baseline, first year and second year values of plaque index and

gingival index of FPD and crown restorations (Table 2-6).


Figure 2-3. Fractured connector in anterior Empress 2 fixed partial denture after 11
months of clinical service.


Table 2-4. FPD restorations: Scores of the clinical evaluation (%) at baseline, year one
and year two (Surface texture and color, anatomic form values were calculated
for n=15 for the first year and for n=10 for the second year).

Baseline Year 1 Recall Year 2 Recall
Empress
2 FPDs Alpha Bravo Alpha Bravo Charlie Alpha Bravo Charlie
(n=20)
Surface
texture 85 15 80 20 60 40
and color
Anatomic
Anatomic 75 25 67 33 70 30
form
Marginal
Marginal 85 15 60 15 25 54 13 33
adaptation
Caries 100 100 100
Post
operative 90 10 90 10 100
sensitivity ________









Table 2-5. Crown restorations: Scores of the clinical evaluation (%) at baseline, year one
and year two.

Baseline Year 1 Recall Year 2 Recall
Empress 2
crowns Alpha Bravo Charlie Alpha Bravo Charlie Alpha Bravo Charlie
(n=20)
Surface
texture 80 20 80 20 65 35
and color
Anatomic
Anatomic 90 10 90 10 80 20
form
Marginal
adaptation 70 20 10 40 40 20 25 55 20
adaptation
Caries 100 100 100
Post
operative 95 5 95 5 100
sensitivity________


Table 2-6. Scores of the clinical evaluation for plaque and gingival index (%) at baseline,
year one, and year two.

Empress 2 FPDs Empress 2 Crowns
Ba e Year 1 Year 2 Year 1 Year 2
Baseline Baseline
Recall Recall Recall Recall
Plaque index
0 100 67 80 100 70 60
1 33 20 30 40
Gingival
index
0 90 85 85 95 80 75
1 10 15 15 5 10 25


Kaplan-Meier statistics revealed that the survival rate for Empress 2 FPDs at 2


years was 50% (Fig. 2-4).










1,0

,9

,8

7


S 15

,4
3

-2 Survival Function


0,0
0 6 12 18 24 30 36 42

Time (months)

Figure 2-4. Kaplan-Meier survival statistics of Empress 2 fixed partial dentures (n=20).



2.3 Discussion

Metal-ceramic prostheses are commonly used for FPDs (Suarez et al., 2004). Previous

studies revealed that they have a high survival rate of 98% (Creugers et al., 1994), 90%

(Scurria et al., 1998), and 85% (Walton, 2002) at 5, 10, 15 years respectively. Also, all-

ceramic FPDs are becoming more popular and more desired due to their esthetic capacity.

However, there are very few clinical studies with all-ceramic prostheses that evaluate

their long term success (Suarez et al., 2004). Recent clinical studies reveal high failure

rates of all-ceramic restorations compared with metal-ceramic restorations, especially

when they are used in the posterior region (Sorensen et al., 1998, Olsson et al., 2000).

Empress 2 core ceramic is composed of crystalline and glass phases. The

crystalline Empress 2 core consists of elongated lithia disilicate crystals (Li2Si205). The









lithia disilicate crystalline content in the hot-pressed core ceramic is 70 + 5 vol % (Della

Bona, 2001). Empress 2 veneer also consists of a glass and one crystalline phase. The

latter is reported to be fluorapatite and the crystal volume fraction is less than that of the

Empress 2 core (Holand, 2000) (Appendix A). X-ray diffraction analysis showed that the

Empress 2 veneer ceramic is an amorphous glass-ceramic (Appendix A). Empress 2

core ceramic has a flexural strength of 215 MPa and fracture toughness of 3.4 (MPaml/2)

(Della Bona, 2001). Fabrication of three-unit anterior FPDs using the Empress 2 system

is recommended by the manufacturer because of high mechanical properties of Empress

2 core ceramic (Della Bona, 2001). However, when the core layer is coated with the

low-strength glass veneer, the resulting ceramic composite exhibits a significantly lower

strength compared to the core ceramic. Distribution of stresses is affected by the elastic

modulus differences between the glass veneer and ceramic core. These stresses in bilayer

ceramic composite are different than those in the monolithic core material. In our specific

case, the outer layer is a glass and will result in lower failure stress in core/veneer

composite than in the monolithic core ceramic.

In the present study, a high fracture rate (50%) was observed for veneered

Empress 2 FPDs during the examination period. This outcome is related to the presence

of a low toughness Empress 2 glass veneer (66 MPa) bonded to the core ceramic.

In a previous study, a 3% fracture rate was reported in 41 Empress 2 FPDs after

a 10-month clinical evaluation period (Sorensen et al., 1998). Also, Postpiech et al.

(2000) reported a 10% failure rate at the one year clinical examination. In addition, they

reported no clinical failures in 76 crown restorations. The results of our study confirmed









these results for crown restorations; however, failure rates for FPD restorations were

significantly higher in our study.

Even though special attention was paid during occlusal adjustment sessions to

minimize the occlusal loads on FPD connectors, most of the fractures occurred at the

connector areas. We conclude that connector design and thickness has a significant effect

on the long term clinical survival of Empress 2 FPDs. This finding is in agreement with

the results reported in a previous in vitro study (Oh et al., 2002).

In our study we did not observe a significant difference between the base line,

first year, and second year values of surface and color, plaque index and gingival index

parameters.

In a two-year clinical evaluation period, 50 % of the 20 Empress 2 FPDs had

fractured. For these failed prostheses, 80% fractured at the connector regions and 20%

chipped within the veneer layer. None of the crown restorations had fractured. The

marginal integrity criteria of the restorations in the second year recalls were poorer than

their baseline. Fracture surface analysis of the clinically fractured restorations is needed

to understand the cause of the failures. The outcome of this study combined with fracture

surface analysis of the failed FPDs can provide critically important information that can

lead to improvement in ceramic properties and increased survival probabilities for full-

ceramic prostheses. Thus, the next chapter will present the fracture surface analysis of

clinically failed fixed partial dentures.














CHAPTER 3
QUANTITATIVE FRACTURE SURFACE ANALYSIS OF CLINICALLY FAILED
CERAMIC FIXED PARTIAL DENTURES



Bilayer ceramics used for fixed partial dentures are becoming more widely used

because of their aesthetic properties in the oral environment. However, the large

statistical variation in the strength of these ceramics is associated with their low

toughness and structural flaws. In addition, localized fracture of the veneering ceramic

and connector fractures within the veneer layer of a bilayer full ceramic FPD system have

been reported in previous clinical studies (Sorensen and Cruz, 1998; Pospiech et al., 2000

and Edelhoff et al., 2002).

An increasing interest in ceramic fixed prostheses has followed improvements in

strength, aesthetics and ease of processing. Such advances include introduction of lithia

disilicate (Li20-2SiO2) reinforced glass-ceramics for dental use. The moderately high

strength and improved esthetics of these systems are well documented in the dental

literature (Holand et al., 2000). However, the mechanisms for the failure of dental

ceramics have not been studied extensively, although fractographic analysis is a key

element in the design and development of dental structural materials (Kelly et al., 1989).

Fractographic analysis of retrieved clinical specimens has been crucial in efforts

to investigate failure mechanisms, to identify fracture initiation sites, and to determine the

probable cause of failure. Fractography can provide information about the cause and

source of failure. One can determine whether the failure is caused by a processing defect









or by an overload condition. Quantitative fractography or fracture surface analysis, is the

application of fracture mechanics to characteristic features on the fracture surface

including the size of the fracture-initiating flaw. However, although fractography is based

on the science of fracture, proper interpretation depends on the skill and knowledge of the

examiner, particularly the ability to recognize fracture markings. This paper provides

guidelines for observing the patterns of cracks and features on fracture surfaces of failed

ceramic prostheses. The overall objective of this study will be to identify the principal

cause of failure in ceramic fixed partial dentures (FPDs) using fractographic techniques.

3.1 Materials and Methods

FPDs that were produced from veneered and nonveneered lithia-disilicate-based

glass-ceramic core ceramics were analyzed in this study. Veneered glass-ceramic FPDs

were made from the Empress 2 ceramic system (Ivoclar Vivadent AG, Schaan,

Liechtenstein). X-ray diffraction analysis revealed that the veneering ceramic consists

primarily of amorphous glass (Appendix A). All core FPDs were made using an

experimental lithia-disilicate glass-ceramic that has a greater crystal volume fraction than

Empress 2 core ceramic (Appendix A). Clinically failed FPDs were retrieved from a

clinical study. All of the FPDs were made by the same technician using the same

fabrication procedures.

Four-point bending specimens were prepared from Empress 2 veneer, Empress

2 core ceramic, lithia-disilicate-based experimental core ceramic, and bilayer Empress

2 core-veneer ceramic. The final dimensions of beam specimens were 1.6 mm (height) x

4 mm (width) x 25 mm (length). The veneer thickness was 0.6 mm for each bilayer

specimen. The core/veneer thickness ratio was 10/6. The span length/specimen thickness

ratio was 15/1.6 to avoid large deflections and high shear stresses within the beam









specimens. After polishing, dimensions of the specimens were measured with a

micrometer. A total of 10 specimens were used for each group for fracture toughness

calculations. Thus, 30 beam specimens were used in the study. Specimens were indented

on the veneer surface with a Vickers indenter at a load of 4.9 N to produce controlled

cracks. Specimens were stored in air for 24 h to ensure complete crack growth. They

were then loaded to fracture at a crosshead speed of 0.5 mm/min.

Beam specimens were loaded using a four-point bending fixture and an Instron

universal testing machine. All flexure experiments were performed using the same four-

point flexure fixture with an 18 mm outer span and a 6 mm inner span. The veneer layer

was placed in tension for bilayer flexure test specimens. The mean strength (of) of the

laminated composites was calculated using composite beam theory (Beer and Johnston,

1981) (Appendix B).

Fracture toughness of the beam specimens was measured using quantitative

fractography. Fracture in brittle materials generally occurs by the unstable propagation of

a defect as a result of the combination of high stress and large flaws (Mecholsky, 2001).

Flaws occur in dental ceramics as a result of processing or preparation. Fracture origins

may be volume flaws such as internal cracks, pores, agglomerates, regions of

inhomogeneous density or composition, or surface origins such as cracks caused by

machining, surface pits or voids, and impact damage.

Almost all of mechanically induced cracks can be idealized as semi-elliptical,

sharp cracks of depth, a, and half-width, 2b (Fig. 3-1) (Mecholsky, 1991). The crack sizes

are approximated by an equivalent semicircular crack size, c [c = (ab)12]. Fracture









toughness, Kc, is calculated using the stress at fracture, or strength, cf, and the crack size,

c:

Kc= Y f(c1/2) (3.1)

where Kc is the critical stress intensity factor (fracture toughness), and Y is a geometric

factor, which accounts for the shape of the fracture-initiation crack and loading condition.

The quantity Y depends on the ratio a/b. The approximation [c = (ab)1/2] allows many

irregular crack shapes to be analyzed and avoids the complications of calculating a

geometric factor for each crack (Mecholsky, 1991). For surface cracks that are small

relative to the thickness of the sample, Y-1.24. For sharp cracks that are induced by a

Vickers or Knoop indentation, Y-1.65, and for internal cracks, Y-1.4 (Mecholsky, 1991).

The fracture origins were determined by examining the fracture surface and

tracing the fracture surface markings back to the initiation site. These markings include

twist hackle (river marks), wake hackle (fracture tails), cleavage steps, Wallner lines, and

branching locations.

When characterizing fracture origins, photographs were made of the overall

sample and of an enlargement of the fracture origin region. The general fractographic

procedure is outlined in ASTM standard C1322 (ASTM, 1999).

Each specimen was studied using a stereoscopic microscope (Bauch & Lomb Inc.,

Rochester, NY, USA) at 160 X magnification. Crack initiating flaws (a, b) were

measured to determine the fracture toughness of each specimen. Scanning electron

microscopic (JSM-6400, Jeol, Tokyo, Japan) examination was performed on selected

specimens.









Residual stress caused by the thermal expansion/contraction mismatch of the

veneer ceramic and core ceramic was estimated using the following equation (Lawn,

1993):

OR=AcAT/[( 1 +vc)/2Ec+( 1-2vv)Ev] (3-2)

where aR is the residual stress, Au is the difference between thermal expansion

coefficients of the veneer and core ceramics, and AT is the difference between the glass

transition temperature of the veneer ceramic and room temperature. Subscripts C and V

refer to core and veneer ceramics, respectively. The terms v and E are Poisson's ratio and

Young's modulus, respectively.

3.2. Results

We estimated from preliminary examination of the fracture surfaces, that eight

(89%) of the nine connector failures and two (66%) of the three veneer ceramic failures

were acceptable for fracture surface analysis. Primary fracture origins of the two

discarded specimens were missing as a result of a chip fracture near the origin. All

fracture surfaces exhibited multiple crack initiation sites as a result of multidirectional

and repeated loading. Arrest lines in the form of ridges were present on the fracture

surfaces (Fig. 3-1). In addition, some curved markings resulted from the intersection of

two crack fronts (Fig. 3-1). Twist hackle markings were helpful in determining the

primary fracture origin in the core ceramic (Fig. 3-2). Wake hackle markings from pores

were selected as reference points for determining the location of fracture origins in the

glass veneer (Fig 3-2). The wake hackle markings were observed next to pores on the

opposite side from the fracture origin in the veneer layer. Porosity exists within the glass

veneer because of slurry preparation and sintering of veneer powders.




























Figure 3-1. All fracture surfaces exhibit multiple crack initiation sites as a result of
multidirectional loading. Arrest lines in the form of ridges are present on the
fracture surfaces. One of the arrest lines is a result of the intersection of two
fracture paths.


Figure 3-2.Wake hackle markings were used to establish the reference points for
determining the fracture origin in the glass veneer. Wake hackle markings
were also observed next to porosities as an outcome of crack propagation
through pores. The markings indicate the direction of the fracture origin in the
veneer layer. Black arrows indicate the direction of the fracture path.









Fracture, in seven of the eight (88%) connector failures examined, initiated within

the surfaces of the prostheses. Six (75%) of the eight connector failures originated in the

occlusal surface of the FPD and one initiated from the gingival surface.

Fracture origins of the bilayer FPDs occurred within the veneer layer (Table 3-1).

Two FPDs failed in their posterior abutment crowns and one of them had an identifiable

fracture origin in the margin area. The fracture initiation site in the latter prosthesis was

located on the outer surface of the margin and fracture occurred along the mesiodistal axis

of the abutment. Two failures within the veneer layer failed from chipping and both

originated at internal flaws.

The fracture toughness of the lithium-disilicate-based glass-ceramic core and the

glass veneer were determined from beam specimens using quantitative fractography, i.e.,

equation (3-1). The mean fracture toughness of the core ceramic was 3.1 + 0.1 MPa-ml/2

and the mean fracture toughness of the glass veneer was 0.7 + 0.1 MPa-ml/2

Fractographic analysis can only determine the stress that caused the fracture to occur

from a crack or flaw of a particular size if the toughness or apparent toughness is known.

In the case of all core FPDs, the failure stress can be estimated directly from the crack

size and the fracture toughness determined from the core ceramic bar specimens. The

fracture toughness of the glass veneer was used to calculate the stress at failure (Eq. 3-1)

for the FPD specimens in which fracture initiation occurred within the veneer surface.

However, for the veneered FPDs, we have shown that residual compressive stress is

caused by an elastic thermal mismatch and viscoelastic (Appendix C) relaxation.

Therefore, the calculated stress should include an additional term, which includes the

compressive residual stress that has to be overcome before tensile stress can develop.









Dilatometric analysis showed that there was a slight difference between the thermal

expansion coefficients of the core and veneer. The values were 10.0 (ppm-K-1)and 10.4

(ppm-K) for the core and glass veneer, respectively. The glass transition temperature

(Tg) of the glass veneer is 5400C and the difference between the Tg and room temperature

is 5150C. Poisson's ratio of the glass veneer, vv is 0.23 and for the core ceramic, vc, it is

0.24. In addition, the measured Young's modulus of the glass veneer, Ev is 64 GPa and

for the core ceramic, Ec is 96 GPa. The estimated residual stress for veneered prostheses

using the above parameters and equation (3.2) was 13 MPa compressivee) as shown in

Table 3-1. This calculation assumes that veneer/core materials are identical in their

thermal history and geometry.

The residual stress was estimated from laminated beams fabricated using firing

schedules recommended by the manufacturer. This is the value listed in the Table 3-1 as

CR. The critical flaw sizes, estimated failure stresses, and fracture initiation sites for each

ceramic FPD are listed in Table 3-1. The critical flaw size, c, of the specimens for which

an origin could be found, ranged from 240 to 939 im. Estimated failure stresses of the

FPDs that had fracture initiation sites within the veneer layers, ranged from 19 to 68

MPa. Failure stresses of all-core FPDs that fractured in the connector area ranged from

107 to 161 MPa.









Table 3-1. Failure stress (of), residual stress (oR), critical flaw size (c), semiminor axis(a),
and semimajor axis (2b) of the FPD specimens.

Sample ID a 2b c Failure Crack initiation Mti
Material
(pm) (pm) (pm) Stress (MPa) site
1 376 1129 460 116 Connector Experimental
1 376 1129 460 116
(gingival surface) core ceramic

2 94 1223 240 161 Posterior crown Experimental
margin core ceramic
3 Connector Experimental
3 (occlusal surface) core ceramic
4 361 918 407 124xperimental
(occlusal surface) core ceramic

5 188 706 258 156 onnector Experimental
4 282 941 364 131

(occlusal surface) core ceramic
7 Connector Experimental
5 188 706 258 156
(occlusal surface) core ceramic

Connector Experimental
7 376 1543 539 1073(R) core ceramic
(occlusal surface) core ceramic
Connector Experimental
6 2 8 3 1 367 440 27+13(oR) occ re core ceramic/
(occlusal surface)veneer


Experimental
9 791 2232 939 19+13(CR) Veneer chip-out core ceramic /
glass veneer
Experimental
8 229 829 308 32+13(oR) core ceramic /
glass veneer
Experimental



10 283 1367 440 68+13(CR) Veneer chip-out core ceramic /
glass veneer



3.3. Discussion

Fracture surface analyses of failed ceramic FPDs showed that failure origins occurred

mostly at surface flaws except for the cases of chipping failures where fracture occurred

within the veneer layer of the prostheses. Previous investigators reported that fracture

initiated typically along the veneer/core interface of In-ceram alumina based ceramic

crowns (Kelly et al, 1989; Thompson et al, 1994). However, this result was not observed









in our study. For the most part the fractures initiated within the occlusal surface of the

FPDs, most likely because of contact damage caused by the opposing teeth or premature

occlusal contacts. Also, mechanical damage resulting from the occlusal adjustment by the

dentist or the dental technician can introduce flaws in the FPDs. We suspect that fracture

initiation at the margin of an abutment of a specimen occurred because of a flaw that was

introduced by the dentist in the margin area during the try-in procedure.

Since most of the connector failures were associated with fractures that initiated

from occlusal surfaces, we suspect that these flaws were introduced as a result of contact

damage. Evidence that supports this conclusion is that fractures were multidirectional

(Fig. 3-1). During the mastication process the mandible makes lateral, centric and

protrusive movements that allow the opposing cusp tip in the maxilla to exert

multidirectional forces on the prosthesis. As a result, fracture can occur in the most

vulnerable part of the ceramic FPD, i.e., the connector (Oh et al., 2002). Fracture

markings in specimen 3 (Table 3-1) indicated that there were two fracture origins on the

occlusal surface of the FPD (Fig. 3-1). To determine the primary fracture origin that led

to failure, we analyzed fracture markings that were farther away from both fracture

origins. These markings indicate that the primary fracture origin is the one with the

longest path (left side of the fracture surface shown in Fig. 3-1). The ridge between the

two fracture origins represents the intersection of two of the propagating cracks.

In two of the three chipping failures, fracture origins were visible. In the third

case, fracture initiated from the internal surface of the veneer layer and propagated in two

directions. The fracture origin was not present on the fracture surface.









Fracture toughness of the glass veneer was used to calculate the stress at failure of

specimens in which the fracture origin occurred within the veneer layer (Fig 3-3). The

fracture initiating crack propagates immediately at the failure stress. Even though the

core layer is tougher than the veneer, once crack propagation begins in the veneer, the

crack does not stop. Crack progression is not impeded by the core ceramic at the interface

between the core and veneer. Thus, the toughness of the veneer is used to calculate the

stress from the crack size (Eq. 3-1). The bilayer materials also include a term for the

compressive residual stress generated by the thermal expansion anisotropy ( 13 MPa)

and viscoelastic (Appendix C) processes that occur.

Imm



















Figure 3-3. Fracture propagated quickly in the glass veneer without any increase in stress
from the point where the semi-minor axis of the fracture origin ends.


The calculated failure stresses in the veneered lithia-disilicate-based glass ceramic

specimens are relatively low compared with those reported by Holand et al. (2000). We

reported that the increase in strength of bilayer core/veneer ceramics occurred because of









a global compressive residual stress (Taskonak et al., 2002) (Table 3-1). However, it is

not only the global residual stress that plays a role in the failure mechanism of bilayer

dental ceramics. Local residual tensile and compressive stresses adjacent to points of

contact damage from previous loading and tensile stresses from flexural and/or

subsequent contact loading also can lead to additional stress or cause failure. The

superposition of these stresses can cause lateral cracks to develop and/or propagate to the

surface (Lawn, 1993). Even if the stresses are not sufficient to propagate median cracks,

they might be sufficient to propagate lateral cracks and cause chipping of the glass

veneer. This was most probably the case for the chipping failures. We conclude that

fracture initiation sites of these glass-ceramic FPDs occurred primarily within the

occlusal surfaces of the veneered units and the crack propagation patterns appear to be

controlled by the loading orientation.














CHAPTER 4
ROLE OF INVESTMENT INTERACTION LAYER ON STRENGTH AND
TOUGHNESS OF CERAMIC LAMINATES



The demand for ceramic dental prostheses has increased with the introduction of

pressable ceramics. These materials provide excellent esthetics in the oral environment

and simplify the match in color and translucency of anterior fixed partial restorations to

that of adjacent tooth structure. However, fractures related to surface flaws and low

flexural strength appear to be common problems in veneered pressable ceramics

(Anusavice et al., 1989). Failures of dental ceramic structures are multi-factorial and may

be associated with improper crown and bridge design, thermal incompatibility, stresses in

layered structures, the presence of critical structural flaws, and improper processing

techniques (Pospiech et al., 2000).

Clinical failures reported for Empress 2 core/Empress 2 veneer ceramic

specimens occurred by chipping of the veneer layer near the restoration surface (Pospiech

et al., 2000; Edelhoff et al., 2002). Spalling, in which a crack grows beneath the surface

before propagating to the surface, thereby forming a chip, occurs in situations such as

local contact or indentations (Marshall et al., 1982), delamination of the layered materials

(Hutchinson et al., 1992), and during machining (Thoules, 1989).

The Empress 2 core ceramic is composed of crystalline and glass phases. The

crystalline Empress 2 glass-ceramic core consists of elongated lithia disilicate crystals

(Li2Si205). The lithia disilicate crystalline content in the hot-pressed core ceramic is 70 +









5 vol % (Della Bona, 2002). Empress 2 veneer also consists of a glass and a crystalline

phase. The latter is reported to be fluorapatite where crystal volume fraction is less than

that of the Empress 2 core (Holand et al., 2000). X-ray diffraction analysis showed that

the Empress 2 veneer and Eris veneer are amorphous glass (Appendix A). The primary

difference between Empress 2 core and the experimental core is the crystal size. The

experimental core has a smaller particle size than the Empress 2 core ceramic. The mean

surface area of the particles for Empress 2 core is 0.42 tnm2 compared with 0.20 tnm2 for

crystals in the experimental core ceramic (Della Bona, 2002). Although the experimental

core ceramic has smaller crystals and a lower volume fraction than the Empress 2 core

ceramic, their physical properties are almost identical (Della Bona, 2002) (Appendix A).

In any bilayer composite, such as the core/veneer system, the interface between

the two layers can control fracture behavior. If the toughness of the interface is greater

than that of either of the components, the propagating crack would be expected to

progress across the interface. However, if the interface is less tough than that of either of

the phases, then the propagating crack would be expected to propagate along or close to

the interface (Thoules, 1989, Thompson, 2000). The loading and relative toughness of the

interface and each phase control the behavior and the direction of any crack that forms or

propagates in the core veneer system (Thoules, 1989).

Hot pressing is the process used for the fabrication of the core portion of ceramic

crowns and fixed partial dentures. The ceramic core structure is formed by the application

of pressure to cause flow of viscous ceramic in an investment mold of the desired shape

(Tooley, 1985). The shape in the investment mold is formed using the lost-wax process

(Tooley, 1985). One disadvantage of this technique is the difficulty in removal of the









investment material from the final ceramic shape. Grit blasting and acid etching are

common methods used to remove the divesting material from the ceramic fixed partial

denture. However, if divesting is not completed properly, a residual investment

interaction layer can remain on the ceramic surface. Potentially, the remaining layer

could cause bonding problems between the veneer and ceramic core.

The primary goal of this study was to analyze the effect of the investment

interaction layer on the flexural strength, fracture toughness, and fracture path of four

bilayer ceramic laminates. Two hypotheses were proposed:

(1) The fracture path of bilayer specimens will be controlled by the presence of

the interaction layer at the interface.

(2) The flexure strength of bilayer ceramic laminates will decrease in the presence

of an investment interaction layer at the interface.

4.1. Materials and methods

Bilayer ceramic composites were fabricated following accepted dental laboratory

procedures. Self-cured acrylic resin (Pattern Resin, GC Corp., Tokyo, Japan) was used to

prepare master models for rectangular Empress 2 core and experimental core layers.

Impressions were made from the master model using a polyvinyl siloxane impression

material (Extrude, Kerr Corp., Romulus, MI, USA). Acrylic resin material was poured

into the molds and cured to make rectangular bar patterns (1.7 mm x 4.0 mm x 25 mm).

The dimensions of the rectangular bars were made uniform using a milling machine (PGF

100, Cendres & Metaux Sa., Biel-Bienne, Switzerland).









Following the preparation of the resin bars, four specimens of each bar was

sprued and invested in each investing ring. Thus, either an Empress 2 core or

experimental core ingot (Ivoclar AG, Schaan, Liechtenstein) was used for every four

specimens. A preheating furnace was used for the burnout procedure (Radiance, Jelrus

Int., Hicksville, NY, USA). The following two-stage burnout sequence was used: (1) heat

at 5C/min to 2500C and hold for 30 min hold; and (2) heat at 5C/min to 8500C and hold

for 1 h. After the preheating stage, the investment cylinders were immediately transferred

to the pressing furnace (EP500, Ivoclar AG, Schaan, Liechtenstein). The pressing

temperatures for Empress 2 core and experimental core ceramics were 9200C and

9100C, respectively.

Following the pressing procedure, the investment cylinders were removed from

the pressing furnace and cooled for 2 h in a ventilated room. The cooled specimens were

divested by grit blasting with 80 |tm glass beads (Williams glass beads, Ivoclar North

America, Amherst, NY, USA) at an air pressure of 0.28 MPa. Before etching, the sprues

were cut away and excess sprue segments were removed by grinding from the specimen

surfaces using water as a coolant.

Four core specimens were placed in one plastic bottle containing 20 mL of 1% HF

solution (Invex Liquid, Ivoclar AG, Schaan, Liechtenstein) and these bottles were placed

in an ultrasonic bath. After etching, the specimens were cleaned under running tap water

for 10 s and then dried thoroughly.

A metal rod, 3 cm in length, was attached to the tip of the grit blasting hose (1.6

mm in diameter) to assist in standardizing the distance from the tip to the specimen. Grit

blasting was performed with 100 |tm A1203 particles (Blasting Compound, Williams-









Ivoclar North America Inc) at an air pressure of 0.1 MPa. The timing of etching and grit

blasting was controlled according to each procedure described below. Divested

specimens were cleaned after grit blasting with a steam spray under pressure before

veneering.

The following conditions were designed to change the thickness of the interaction

layer for each hot-pressed core ceramic group:

Group 1: 1 % HF acid etch for 30 min and grit blast the surface to be veneered for 30 s

Group 2: 1 % HF acid etch for 30 min and grit blast the surface to be veneered for 15 s

Group 3: 1 % HF acid etch for 15 min and grit blast the surface to be veneered for 30 s

Group 4: 1 % HF acid etch for 15 min and grit blast the surface to be veneered for 15 s

Prepared core bars were veneered with Empress 2 veneer and eris veneer

ceramics (Ivoclar AG, Schaan, Liechtenstein) according to the combinations described

below. Veneer powders were incorporated with a mixing liquid (Ivoclar AG) to obtain a

slurry solution, which was brushed onto the core ceramics that were prepared in slightly

oversized silicone molds.

The following core / veneer specimens were prepared:

Combination A: Empress 2 pressed core ceramic, Empress 2 veneer

Combination B: Empress 2 pressed core ceramic, Eris veneer

Combination C: Experimental pressed core ceramic, Empress 2 veneer

Combination D: Experimental pressed core ceramic, Eris veneer









Sixteen specimens were prepared for each of these 16 groups (four ceramic groups x four

divesting conditions).

The Empress 2 veneer-ceramic specimens were sintered in a furnace (P80,

Ivoclar AG) with the firing cycle set for a 6-min climb to 8000C at 600C/min, a 2-min

hold, and cooling to 1800C. A vacuum was applied at 4500C and 7590C. Three layers of

veneering ceramic, including the wash layer, were sintered on each core ceramic

specimen. Furnaces were calibrated each day before the firing procedures. The Eris

veneer-ceramic specimens were sintered in the same manner as described above except

that the maximum firing temperature was 765 C.

Following sintering, any excess veneer was ground away using a 75-grit diamond

embedded disk, the veneer surface was sequentially polished to a 2000-grit finish on a

metallographic polisher (Model 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while

exposed to a continuous flow of tap water. Edges of each specimen were beveled to

minimize edge failure during flexure testing.

The final dimensions of the bimaterial bar specimens were 1.7 mm (height) x 4.0

mm (width) x 25.0 mm (length). The veneer thickness was 0.6 mm for each specimen.

The veneer/core thickness ratio was 6/11. This ratio was chosen after composite beam

analysis revealed that the tensile stress at the interface would be maximized (Beer and

Johnston, 1981) (Appendix B). If the interface was at the neutral axis, no stress would

develop. Thus, the thickness of the veneer-core specimens were adjusted to shift the

neutral axis away from the core and to ensure as large a tensile stress as possible at the

interface. The ratio of test span length/thickness was chosen as 15mm/1.7mm to avoid

large deflections and large shear stresses during loading. After final polishing, the









dimensions of the specimens were measured with a digital micrometer. All specimens

were then stored in distilled water for 72 h before testing.

The flexural strength of the ceramic laminates was determined using a four-point

flexure test fixture. The flexure fixture was enclosed in a testing chamber and specimens

were tested in circulating water at 370C to approximate an oral environment. The flexure

tests were performed using a universal mechanical testing machine (Model 4465, Instron

Corp., Canton, MA) with four-point flexure fixture (15 mm outer span, 5 mm inner span)

at a crosshead speed of 0.5 mm/min. For all flexure tests, the top fixture was held in place

using a three-post alignment apparatus and the bottom fixture was placed on a stainless

steel ball to provide a fully articulating configuration. The veneer side was placed in

tension for all flexural test specimens, because most of the observed failures in previous

clinical studies and in our fractographic analysis of clinically failed FPDs in chapter 3

occurred in the veneer layer (Pospiech et al., 2000 and Edelhoff et al., 2002 ). The

maximum tensile stress occurs in the outer veneer surface within the inner loading span.


The strength, of, of bilayer composites was calculated using composite beam

theory (Beer and Johnston, 1981) (Appendix B). In the past, these types of analyses have

been used to determine the mechanical behavior of several dental laminates (DeHoff et

al., 1989; Thompson, 2000). Weibull analysis was performed to evaluate the structural

integrity of the bilayer ceramic specimens. Characteristic strength values were also

calculated.

To determine the fracture origins of the specimens, the fracture surfaces of the

specimens were coated with gold-palladium using a sputter-coating machine (Technics

Inc., Alexandria, VA, USA). Each specimen was studied with a stereomicroscope (Bauch









& Lomb Inc., Rochester, NY, USA) at 160 X magnification. The length of crack-

initiating flaws and the calculated strengths were used to determine the apparent fracture

toughness (K,) of each specimen (Eq. 4-1). Scanning electron microscopic (JSM-6400,

Jeol, Tokyo, Japan) examination was performed on selected specimens and SEM images

were recorded from representative fracture surfaces. Apparent fracture toughness, Kc,

was calculated using fractographic analysis and the fracture mechanics equation:

Kc = Y () of(c)1/2 (4-1)


where Y(O) is a geometric factor that has a value of 1.24 for an equivalent semicircular

flaw, c, of a semi-elliptical flaw of depth, a, and half width, b, (Mecholsky, 1994), c =

(ab)/2, (assuming no local residual stress), and of is the calculated flexural strength. The

crack size is equivalent to that obtained using the length of the semiminor axis of the flaw

with the appropriate Y(O) factor for the appropriate semi-elliptical crack geometry

(Mecholsky, 1991).

Two specimens were randomly chosen from each divesting group to analyze the

amount of residual investment interaction layer that was left on the core surface after

divesting (Fig. 4-1). Quantitative EDS analysis of the surfaces was made at 80 X

magnification using a scanning electron microscope. The percentage of each element that

was present on the scanned area was recorded after analyzing the surface. No difference

was found in elemental composition between the investment interaction layer and core

layers for each group.

Two-way analysis of variance was used to determine whether the differences

between group means for flexural strength, crack size, and fracture toughness of









specimens were statistically significant. Linear regression analysis of a versus c1/2 was

performed to determine the relationship between critical flaw size and flexural strength

(Mecholsky, 1994; Mecholsky, 1991), i.e.


Of = (Ca+ OR) = [K, / Y(O)] c-1/2 (4-2)

where Ca is the applied stress to failure and oR is any global residual stress that may be

present. If there is no residual stress, then R = 0 and yf= oa.




















Figure 4-1. Cross section of Empress 2 core ceramic with residual investment
interaction layer on its outer surface. Representative sample was divested
using 15 min etching and 15 s grit blasting surface treatment. (A) Residual
investment interaction layer (B) Empress 2, glass-ceramic core layer.

4.2. Results

The mean flexure strength and standard deviation of each group are summarized

in Table 4-1. Based on two-way ANOVA and Duncan's multiple range test, the

differences in mean strength values of groups 1 and 2 were statistically different (p <

0.05) and both group means are significantly different from the values for groups 3 and 4.









No significant difference was found between the means for the flexure strengths of

subgroups 3 and 4 (p > 0.05). Combinations A, B, C and D represent different ceramic

laminates. Also, there were no significant differences between mean flexural strengths of

groups 1, 2, 3, and 4 within combinations A, B, C and D, which represent different

surface (divesting) treatments (p > 0.05).

Weibull modulus and characteristic strength values for each group are

summarized in Table 4-1. The Weibull modulus gives an indication of the variability of

the flexural strength, with greater values indicating a narrower distribution of flexural

strength.

Without exception, failure origins occurred within the veneer surface of each

specimen. The critical cracks, within the tensile region of the laminate beams, were

approximately semi-elliptical in shape (Fig. 4-2). Independent of the crack origin, all of

the specimens showed a crack path approximately perpendicular to, and through the

laminate interface without delamination. Surface porosity in the veneer was a common

flaw that initiated crack propagation. Wake hackle markings, informally known as

"fracture tails", were observed at pores on the fracture surfaces (Fig. 4-2A). These

markings were helpful in determining the crack direction and critical flaw site. Two-way

ANOVA showed that the mean flaw size dimensions (c) of groups C and B were

significantly different from those of combinations A and D (p < 0.05) (Table 4-1). There

was no significant difference between the mean flaw size dimensions (c) of groups 1, 2, 3

and 4 for ceramic combinations A, B, C and D, respectively (Table 4-1). Ordinarily the

largest flaws are expected to correspond to the lowest strength values. This clearly was

not the case among the groups (Table 4-1).









Table 4-1. Mean flexural strength (o), standard deviation (SD), number of specimens per
group (n), characteristic strength (oo), critical Flaw size (c) and Weibull
modulus (m) values of bilayer ceramic laminates.

a (S.D.) Go (c) S.D.
Divesting Combination n m
(MPa) (MPa) (10 m)

Combination 94(17) 16 6 92 12+3
Group 1 A 84 (20) 16 4 98 15+6
Etch 30 min, Empress 2
84(18) 17 5 96 14+5
grit blast 30 s core / Empress
2 veneer 83 (16) 17 5 94 15+5

Combination 123 (18) 17 7 125 9+3
Combination
Group 2 B 126(15) 16 8 104 8+2
Empress 2
Etch 30 min, Empress 2 133 (15) 17 9 134 7+2
core / Ens
grit blast 15 s veneer 122 (20) 17 6 119 9+3

Combination 118(27) 17 4 104 107
Group 3 C
Gro 3 Experimental 112(21) 17 4 105 12+7
Etch m core/Empress 106(24) 16 5 103 12+4
grit blast 30 s 2 veneer
106(23) 17 9 102 10+2

94(26) 16 3 110 17+12
Combination
Group D 112(22) 17 4 115 12+6
Etch 15 min, Experimental
gritblcore/ Eris 111(20) 17 5 113 12+5
grit blast 15 s17 4 110 14
105 (11) 17 4 110 14+8







































Figure 4-2. Fracture surface of ceramic laminate (SEM view). (A) The fracture origin
(arrow), and wake hackle markings [12] in the veneer ceramic confirm that
failure initiated from the tensile surface in the lower middle portion of the
micrograph as marked by the larger arrow. (B) Fracture surface of the
specimen in Figure 2A, at higher magnification. The crack initiation site is a
pore defect. (C) Fracture surface of a laminate specimen. The core ceramic is
shown by the lighter area (top), and the veneer is the darker area (bottom).

4.3. Discussion

For each core-veneer combination and surface treatment group, there is an

expected correlation (Figure 4-3) between the flaw size and strength, i.e., high strength,

corresponds to a small flaw size. However, there is a lack of correlation between critical

flaw sizes and strength, i.e., some of the combinations that have large values of strength

also have large values of flaw sizes. Generally, it is expected that high strength









corresponds to small flaw sizes. This lack of correlation for the bilayer groups in Table 4-

1 suggests that a more complex mechanism controls the fracture process. To explain

these results, the graphs of flexural strength versus c-12 values were made for each group.

A representative graph, shown in Fig. 4-3, demonstrates the expected relationship

between cy and c determined from Equation 4-2. Notice that the graph in Figure 4-3 does

not intersect at y =0 but rather at cy 27.5 MPa. This implies that there is a global

compressive residual stress in the bilayer specimens. The slope of this graph is equal to

K,/Y. Using Y=1.24 for equivalent semicircular cracks (Mecholsky, 1994), the calculated

Kc value was 0.7 MPa-m1/2 (Figure 4-3). This calculation was made for all bilayer

specimens and results lead to the same conclusion. The similarity in toughness values for

different core-veneer combinations occurs because all failures originated in the veneer

and the veneer toughness controlled these failures. The toughness of monolithic veneer

specimens using fractographic analysis was 0.7 + 0.1 MPa-m1/2 (Figure 4-4). This value is

in agreement with the above calculations.

Since there is no difference among the mean values for groups 1 to 4 for each

divesting treatment, it is concluded that the residual interaction layer (Fig. 4-1) did not

significantly affect the flexural strength, Weibull moduli, characteristic strength or

fracture toughness of the bilayer ceramic laminates. Thus, both hypotheses are rejected.

However, different bilayer ceramic combinations had significantly different values of

mean flexural strength, characteristic strength, and apparent fracture toughness. Thus, the

material properties appear to be more important factor in the mechanical behavior of

bilayer laminates than the residual investment interaction layer as long as the interaction

layer results in a well-bonded interface. There was also no significant difference in failure










modes and crack sizes within these groups and combinations. Thus, the differences

observed in strength for the core/veneer combination groups are likely associated with

global residual stress caused by the thermoelastic and viscoelastic (Appendix C) behavior

of the bilayer specimens. Test methods that involve application of loads directly to the

interface should be used to understand precisely the interfacial strength of bilayer

specimens with the investment interaction layer.




Group A-4


140

120

100
80

"* 60

3 40

20

0
0 20 40 60 80 100 120 140 160

C-1/2 (m-1/2


Figure 4-3. Representative graph (Group A-4) of flexural strength (of) versus c-12.
Graphs for each group showed that the data lie on a best-fit linear line (R2
0.89). Extrapolation of the best-fit linear line indicates a residual stress of
27.5 MPa, i.e., the intercept with the ordinate (slope = 0.6 MPa-m/2).




For any bilayer composite, the combination of loading and toughness of interfaces

will determine if failure occurs because of delamination of the interface or flexural/tensile









failure of the composite (Thoules, 1989 and Thompson, 2000). In the absence of interface

fracture, toughened multiphase and laminated ceramics are susceptible to brittle failure in

a manner similar to fine grained, homogenous ceramics (Kerans et al., 1989 and Prakash

et al., 1995).





Empress 2 veneer monolithic specimens


0 20 40 60 80 100

C-1/2 (m-1/2)


Figure 4-4. Representative graph of flexural strength (of) versus c-12 for Empress 2
veneer monolithic specimens. Graphs for each group showed that the data lie
on a best-fit linear line (R2 = 0.89). Extrapolation of the best-fit linear line
indicates a residual stress of 3.6 MPa, i.e., the intercept on the ordinate (slope
= 0.59 MPa-m/2).



The Weibull moduli values varied from group to group. These values reflected the

scatter in flexural strength for the specimens in each group. As expected, the flexural









strength generally varied inversely with the square root of the flaw dimensions for each

group (Fig. 4-3 and Fig 4-4). A majority of the critical flaws in the specimens initiated

from surface porosity in the veneer ceramic. Porosity may result from air entrapment

during mixing or condensation, the use of a narrow particle size distribution, and gases

produced during firing (McLean and Hughes, 1965).

Specific emphasis should be placed on crack-initiating flaws and crack origins of

the specimens retrieved from clinical studies. Since a majority of such failures initiated

from surface porosity at the tensile surface, chipping failures that occur in clinical studies

may be a result of volume flaws at the surface of the veneer layer. The driving force for

subsequent crack growth is most likely the residual stresses generated from thermal

expansion mismatches and viscoelastic relaxation processes (Appendix C) in the

core/veneer laminates (Scherer, 1986).

Elastic-viscoelastic relaxation behavior of a glass veneer layer affects this residual

stress. When a glass veneer/ceramic core bilayer composite is heat treated at temperatures

near the glass transition temperature, the density of the glass veneer will increase because

of structural relaxation whereas little or no change takes place in the ceramic core. As the

glass reaches its equilibrium structure, there is a rapid increase in the residual stress

initially; then there is a gradual increase until the residual stress equilibrates at a greater

stress level (Scherer, 1986; Anokye, 1989).

All fracture origins in the four-point flexural test specimens occurred at the tensile

surface and there was no interface delamination in any of the specimens. The Empress 2

core and Eris veneer combination showed a global residual stress of -28 MPa and the

greatest flexural strength. The Empress 2 core / Empress 2 veneer specimen group









exhibited the lowest mean fracture toughness and flexural strength. The mean flexure

strength of specimens between material combinations varied from group to group (Table

4-1). Linear regression analysis of flexural strength versus negative square root of flaw

dimension data is consistent with a linear relationship expected from linear elastic

fracture mechanics (Mecholsky, 1994 and Mecholsky, 1991).

The hypothesis that different divesting methods influence the flexural strength of

the specimens was rejected. The hypothesis that the interaction layer controlled the

fracture path was also rejected. Since the investment interaction layer results in a

coherent, well-bonded interface, then there is no effect on strength or toughness of the

bilayer laminates. Different combinations of core and veneer ceramics had significantly

different apparent fracture toughness values, flexural strength, and Weibull moduli. The

latter result implies that thermoelastic and viscoelastic (Appendix C) properties of the

ceramic-veneer combinations may control the strength.

The flexural strength and fracture toughness of layered ceramics loaded in flexure

(with strong bonding at the interface) are governed by the flaw characteristics of the

material in tension and residual stress.














CHAPTER 5
RESIDUAL STRESSES IN BILAYER DENTAL CERAMICS



An increasing interest by dentists and patients in ceramic fixed prostheses has led

to improvements in strength, toughness, aesthetics and ease of processing. However, the

large statistical variation in the strength of these ceramics is associated with their low

toughness and structural flaws (cracks and porosity). In addition, chipping fracture, i.e.

spallation from the veneer layer of bilayer ceramic prostheses has been reported in

clinical studies (Pospiech et al, 2000; Edelhoff et al., 2002). Empress 2 (E2C),

experimental (EXC) core glass-ceramics, Empress 2 veneer (E2V) and Eris veneer

(ERV) are the lithia-disilicate-based glass ceramic core and silicate amorphous glass

veneer ceramics (Appendix A), respectively, selected for this study. The increased

strength and improved esthetics of these systems have received much attention in the

dental literature (Holand et al, 2000).

Bilayer ceramics that are widely used for dental crown and bridge prostheses can

sustain relatively high global compressive residual stresses within the veneer surface

(DeHoff and Anusavice, 1989). Compressive stresses can result from thermal expansion

coefficient differences between the core and veneer ceramics or from differences in the

elastic-viscoelastic behavior between the two layers (Scherer, 1986). Veneered ceramics

with such residual stresses are susceptible to chipping, i.e. spallation (Marshall and Lawn,

1980). During mastication, local stresses are applied to the occlusal surfaces of









prostheses. With the superposition of global compressive residual stresses on these local

stresses, a tensile stress region can develop below the surface of the veneer layer. Under

these stress states lateral cracks can form and/or propagate toward the surface, potentially

resulting in chipping of the veneering ceramic (Hutchinson and Suo, 1992).

This paper presents a four-step fracture mechanics approach to determine the

existence and cause of any global residual stress in bilayer dental ceramics.

5.1. Materials and methods

5.1.1. Specimen Preparation

Bilayer ceramic composites were fabricated following accepted dental laboratory

procedures. Self-cured acrylic resin (Pattern Resin GC Corp., Tokyo, Japan) was used to

prepare E2C and EXC bar specimens.

Four bar specimens per mold were sprued and invested with IPS Empress 2

special investment material (Ivoclar Vivadent AG, Schaan, Liechtenstein). The invested

bar patterns were eliminated using a two-stage burnout for 60 min at 8500C in a

preheating furnace (Radiance, Jelrus Int., Hicksville, NY, USA) and the mold cavity was

filled with core ceramic by hot pressing either E2C or EXC ingots using a pressing

furnace (EP500, Ivoclar Vivadent AG, Schaan, Liechtenstein) (Appendix A). The

pressing temperatures for E2C and EXC ceramics were 9200C and 9100C, respectively.

The cooled specimens were divested by particle abrasion with 80 |tm glass beads

(Williams glass beads, Ivoclar Vivadent North America, Amherst, NY) at a pressure of

0.28 MPa. The remaining investment was cleaned by placing each bar in 1% HF (Invex

Liquid, Ivoclar Vivadent AG, Schaan, Liechtenstein) for 30 min.









Each specimen was abraded using 100 |tm A1203 particles (Blasting Compound, Williams

Ivoclar North America Inc., Amherst, NY, USA) at an air pressure of 0.1 MPa for 30 s.

Divested specimens were cleaned with a steam cleaner before veneering. Prepared core

ceramic bars were veneered with Empress 2 veneer (E2V) and Eris veneer (ERV)

ceramics (Ivoclar Vivadent AG, Schaan, Liechtenstein) (Appendix A). Veneer ceramic

powders were incorporated with mixing liquid (Ivoclar Vivadent AG, Schaan,

Liechtenstein) to obtain a slurry solution, which was brushed onto the core ceramics that

had been placed in slightly oversized silicone molds.

The E2V ceramic was sintered in a furnace (P 80, Ivoclar Vivadent AG, Schaan,

Liechtenstein) according to a firing cycle consisting of a 6 min rise to 8000C at

600C/min, a 2 min hold, and cooling to 1800C in 45 s. A vacuum was applied at 4500C

and released at 7990C. Three layers of veneering ceramic, including the wash layer, were

sintered on each core ceramic specimen. The furnaces were calibrated each day before

the firing procedures.

Following sintering, the excess veneering ceramic was ground with a 75 grit

diamond embedded disk and sequentially polished to a 2000 grit finish on a

metallographic polisher (Model 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while

exposed to a continuous flow of tap water.

The following core/veneer specimens were prepared:

Combination A: E2C / E2V

Combination B: E2C / ERV


Combination C: EXC / E2V









Combination D: EXC / ERV

Six specimens were prepared for each of these four groups. Monolithic veneer specimens

from E2V and ERV powders were prepared in addition to the bilayer specimens.

The final dimensions of the bimaterial bar specimens were 1.7 mm (height) x 4.0

mm (width) x 25.0 mm (length). The veneer thickness was 0.6 mm for each specimen.

The veneer/core thickness ratio was 6/11. The selected test span length/specimen

thickness ratio was 15/1.7 to avoid large deflections of the beam. After final polishing,

dimensions of the specimens were measured with a micrometer.

X-ray diffraction analysis was performed to determine the crystal phases in the

sintered veneering ceramics. Thermal expansion coefficients of the glass veneers and

core ceramics were measured using a dilatometer. The maximum residual stress caused

by the thermal expansion coefficient mismatch between the veneer ceramic and core

ceramic was estimated using the following equation (Lawn, 1993):

CR=AcXAT/[(1 +vc)/2Ec+(1-2vv)Ev] (5-1)

where GR is the residual stress, Au is the difference between linear thermal expansion

coefficients of the veneer and core ceramics (av-ac), AT is the difference between the

glass transition temperature of the veneer ceramic and room temperature, and v and E are

Poissons's ratio and Young's modulus, respectively. Subscripts C and V refer to core and

veneer ceramics, respectively.

5.1.2 Mechanical testing methods

A four-step fracture mechanics approach was used to determine residual stress in

the bilayer dental ceramics. This technique included:









(1) Indentation of the specimens to measure and compare indentation induced crack sizes.

(2) Flexural strength determination for monolithic and bilayer specimens.

(3) Calculation of apparent fracture toughness for bilayer specimens and fracture

toughness determination for monolithic specimens.

(4) Calculation of residual stress in bilayer specimens using a fracture mechanics

equation.

Indentation cracks were induced within the veneer surface of all specimens using

a Vickers indenter at a load of 4.9 N. Indentation induced longitudinal and transverse

cracks were measured optically using a reticulated eyepiece (Fig. 5-1).



Longitudinal crack Transverse crack















Figure 5-1. Schematic illustration of an indented beam specimen placed in four-point
flexure. Indent cracks were induced in the tension surface.


The mean flexural strength of each of the four types of ceramic composites was

determined using a four-point flexure test fixture with a 15 mm outer span and a 5 mm

inner span at a crosshead speed of 0.5 mm/min using a universal testing machine (Model









4465, Instron Corp., Canton, MA). The veneer surface was placed in tension for all

flexure test specimens.

The strength (of) of the laminated composites was calculated using mechanics and

composite beam theory (Beer and Johnston, 1981) (Appendix B). The strength (of) of the

monolithic specimens was calculated using simple beam theory. In previous studies,

composite beam theory (DeHoff and Anusavice, 1989; Thompson, 2000) and finite

element stress analysis have been applied to analyze failure stresses in ceramic/ceramic

and metal/ceramic composites (Scherrer, 1986).

Each fractured specimen was analyzed with a stereomicroscope (Bauch & Lomb

Inc., Rochester, NY, USA) at 160 X magnification. Crack initiating flaws were measured

to determine the fracture toughness of each specimen. Scanning electron microscopic

(JSM-6400, Jeol, Tokyo, Japan) examination was performed on selected specimens.

Fracture toughness, Kc, was then calculated using the fracture mechanics

equation:

Kc= Y(O) Gf(c)1/2 (5-2)


where Y(O) is a geometric factor that has a value of 1.65 for an indentation induced flaw

with local residual stress from indentation (Mecholsky, 1991) and 1.24 for flaws without

local residual stress and of is the calculated flexural strength. For equivalent semicircular

flaws of depth "a" and half width "b" (Mecholsky, 1994), c is the crack size [c= (ab)2]

(Mecholsky, 1991). Residual stress is calculated from the following equation (Conway

and Mecholsky, 1989):









or = [Y(2/71/2) GaC1/2 Kc] / [Y(2/1/2)C1/2] (5-3)

where or is the residual stress, Ca is the applied stress at failure, and Kc is the fracture

toughness of the glass veneers and the other terms are as listed above (Conway and

Mecholsky, 1989).

One-way analysis of variance was performed to determine whether the differences

between group means for flexural strength and fracture toughness of specimens were

statistically significant. The same statistical analysis was used to determine whether the

differences between the mean longitudinal and transverse indentation induced crack sizes

were statistically significant.

5.2. Results

The mean longitudinal and transverse indentation-induced crack sizes and standard

deviation of each group are summarized in Table 5-1. Based on one-way ANOVA and

Duncan's multiple range tests, the differences between the mean indentation induced

longitudinal and transverse crack sizes of bilayer specimens were significantly different

(p < 0.05). However, no significant difference was found between the means for the

longitudinal and transverse crack sizes of monolithic glass veneer specimens (p > 0.05).

One-way ANOVA showed that there was a statistically significant difference

between the mean flexure strengths of monolithic and bilayer specimens (p < 0.05), but

there was no significant difference (Table 5-1) between the mean flexural strength of

bilayer specimen groups (p > 0.05).

Without exception, failure origins were located within the tension surface of each

specimen and not at the interface between the veneer and core. For most of the composite

specimens, the indentation cracks were the critical flaws that controlled crack









propagation (Fig 5-2). However, some specimens failed from porosity at the surface.

These critical flaws, within the tensile region of the composite bars were approximately

semi-elliptical in shape (Fig. 5-3).


Table 5-1. Mean flexural strength (o), standard deviation (SD), residual stress (GR),
longitudinal and transverse indentation-induced crack size and apparent
fracture toughness (K,) of bilayer and monolithic specimens. Negative values
of YR indicate compressive stress in the veneer.

Longitudinal Transverse
Groups aS.D. indent crack indent crack K._ S.D.
(MPa) () (pm) (MPa) (MPaml/2)
(Pm) (Pm)

Eris Veneer 48 + 7 62 +11 59+ 11 0 0.7 0.1



Empress 2 Veneer 43 8 61 + 15 58+14 0 0.7 0.1


Experimental Core
66 11 55 14 44+14 -23 0.9 0.1
/Eris Veneer


Experimental Core

/Empress 2 73 + 15 50+ 18 34+12 -25 1.0 0.2

Veneer


Empress 2 Core /
61 +12 63+ 11 46+12 -22 1.0 0.2
Eris Veneer


Empress 2 Core /
72 +13 58+ 12 9 -22 1.0 0.2
Empress46 + 9
Empress 2 Veneer









Even though all specimens failed from the veneer surfaces, there was a

statistically significant difference between the calculated apparent fracture toughness

values (Eq. 5-2) of bilayer specimens and monolithic veneer specimens (p < 0.05). The

fracture toughness of monolithic ERV and E2V (veneers) was identical (0.7 MPaml/2)

(Table 5-1).

-_,% _AP-Mr_ .+'~L -~~~ ~lg~PX'-Z~ -4 2r-


Figure 5-2. Fracture surface and fracture origin (arrows) of a monolithic veneer specimen
that reveal an indent crack as a fracture origin.




X- ray diffraction analysis was performed to identify crystal phases in the veneer

layers that may increase fracture toughness. However, the veneer layers were found to be

essentially amorphous because there were no distinct peaks observed in the X-ray

diffraction graph between the angles of 5 and 110 degrees.









Residual stress values of bilayer specimens (Eq. 5-3) ranged between -40 MPa

and -60 MPa compressivee) (Table 5-1). The highest compressive residual stresses were

observed in bilayer specimens with Empress 2 veneer layers. No residual stress was

detected in the monolithic specimens (Table 5-1).


Figure 5-3. SEM image of a fracture surface of an E2V specimen that exhibits a large
pore (arrows) as the fracture origin. The bottom figure is an enlarged view of
the top figure.


Dilatometric analysis showed that there was a slight difference between the

thermal expansion coefficients of the core and veneer. The values were 9.8 (ppm-K-)and

10.2 (ppm-K1) for the core and veneer ceramic, respectively. The glass transition

temperature (Tg) of the veneer ceramic was determined to be 5400C and the difference


r









between Tg and room temperature is 5150C. The Poisson's ratios of the veneer ceramic

(vv) and the core ceramic (vc) are 0.23 and 0.24, respectively. The Young's moduli of the

veneer ceramic (Ev) and core ceramic (Ec) are 64 GPa and 96 GPa, respectively. The

calculated residual stress for the bilayer specimens using the parameters above and

equation (5-1) is 13 MPa (compression).

5.3. Discussion

The increase in strength and toughness for the bilayer composites compared with the

glass veneer can be explained by several possible phenomena. These include

crystallization of the veneer layer, the increased toughness of the ceramic core,

compressive residual stress associated with thermal expansion anisotropy, and

compressive residual stress caused by viscoelastic structural relaxation (Appendix C).

X-ray diffraction analysis revealed no evidence of a crystal phase, indicating that

the veneering ceramics consist predominantly of an amorphous glass phase. Although

there are reports that fluorapatite crystals exist in the veneering ceramics (Holand, 2000),

the volume fraction is evidently lower than detection limits. Thus, the strengthening

mechanism for the bilayer composites is not caused by crystallization of the veneer

layers.

Observation of the fracture surfaces (Figs. 5-2 and 5-3) showed that all fractures

occurred within the veneer layer at the surface. Crack propagation continued through the

veneer/core interface. Thus, the increased toughness of the core ceramic compared with

the veneer ceramic did not affect crack initiation or propagation.

Bilayer specimens showed longer longitudinal cracks than transverse cracks

(Table 5-1 and Fig. 5-1). However, there was no statistically significant difference









between the mean indentation-induced transverse and longitudinal cracks of monolithic

glass veneer layers (p > 0.05). This difference in mean crack lengths was the first

indication of residual stress in the bilayer specimens. There are two possibilities for the

difference in the length of indent-induced cracks in the veneer layer. The longer

longitudinal cracks could be a result of residual tensile stress perpendicular to their

direction or the shorter cracks may be a result of the compressive residual stress

perpendicular to their short cracks. In Table 5-1, we see that the difference in mean

length of longitudinal cracks in the monolithic veneer and bilayer specimens is not

statistically significant (p > 0.05), whereas, the mean transverse cracks in the bilayer

specimens are significantly shorter than the ones in the monolithic veneer specimens (p <

0.05). Thus, we conclude that compressive residual stress exists in the longitudinal

direction within the veneer layer of the bilayer specimens.

Although, fractographic analyses of the fracture surfaces showed that the fracture

origins of all specimens occurred within the veneer surface, the flexural strength and

apparent fracture toughness of the bilayer specimens were significantly greater than those

of the monolithic specimens (p < 0.05). These effects represent further evidence of

compressive residual stress in the veneer layers of bilayer ceramics.

The residual stress calculated using equation (5-2) further validates the presence

of compressive residual stress. The distinction between global and local residual stress

should be made. Local residual stress refers to the remaining stresses caused by

indentation. The stresses in this case are near the indentation site and drop off rapidly.

Global residual stresses are those associated with the entire (externally) unloaded

specimen and can be introduced by thermal processing, e.g., from intentional or









unintentional rapid cooling. Each of these residual stress systems is not uniform. Thus,

for detailed calculations, an assumption must be made as to the nature of the stress

profiles. In the present case, we assume that the residual stresses in the vicinity of the

crack are constant and can be superimposed on the applied stress.

Considerable research has been conducted to estimate and measure residual stress

in bilayer glass caused by a mismatch between the coefficients of linear thermal

expansion (Varsheneya and Petti, 1978). Dilatometric analysis showed that the maximum

difference in expansion coefficient between the glass veneer and core materials was 0.4

(ppm-K1). Thus, the calculated residual compressive stress caused by thermal expansion

differences between the veneer and core ceramics was 13 MPa and does not fully explain

the observed strength differences between the bilayer and monolithic specimens.

Previous studies showed that tensile or compressive residual stresses can develop

because of different viscoelastic relaxation mechanisms (Appendix C) in elastic-

viscoelastic composites (Anokye, 1989; Jessen and Mecholsky, 1989). Tensile or

compressive stress can develop and increase in the glass as shown in Figure 5-4 (Anokye,

1989). Even though heat treatments above and below the glass transition temperature of

the glass veneer may create tensile or compressive global residual stresses, these stresses

vary linearly through each layer of the beam, as shown in Figure 5-5 (Scherrer, 1986).














20




10
S-1 Experimental

G 0 Theoretical






-10
-10- I I I I I
380 400 420 440 460 480 500 520

Heat Treatment Temperature (C) (4 h hold)




Figure 5-4. Theoretical calculations and experimental measurements of residual stress in
borosilicate glass resulting from viscoelastic relaxation behavior, as a function
of different heat treatment temperatures above and below the glass transition
temperature for a bilayer system where borosilicate glass (C 7052) was
bonded to Kovar (Iron-Nickel-Cobalt alloy) metal (Anokye, 1989).




There are several potential sources for stresses that result in spallation: global

residual compressive stresses within the surface from elastic and viscoelastic processes,

local residual tensile and compressive stresses adjacent to points of contact damage from

previous loading, and tensile stresses from flexural and/or subsequent contact loading.

We suggest that the superposition of these stresses cause lateral cracks to develop and/or

propagate to the surface (Lawn, 1993) resulting in spallation of segments of the veneering

ceramic (Fig 5-6).



















Compression


Tension


Figure 5-5. Stress distribution in layers of a bilayer specimen when the outer layer is in
compression. GR1, GR2 are the stresses in ceramic 1 and ceramic 2,
respectively (Scherrer, 1986).


Figure 5-6. Schematic illustration of radial and lateral cracks at the surface with or
without residual stress.



Lateral cracks are initiated near the base of the plastic deformation zone below the

contact area and spread out laterally on a plane closely parallel to the specimen surface

(Lawn and Wilshaw, 1975). In a severe contact event these cracks can propagate to the


GR1 -_ 1

-G2R2 2
________________________A__


Median (Radial) Crack









surface to cause material chipping. Observations of sharp contact fracture in glass

showed that lateral extension of a crack occurs during unloading (Lawn and Swain,

1975). Residual stresses were identified as the primary driving force for lateral cracking

in another study (Evans and Wilshaw, 1976).

Flexural strength and the apparent fracture toughness of bilayer ceramic bars are

mainly determined by the veneer layer when the critical crack initiates from the veneer

surface and the critical crack is confined totally within the veneer layer. Global

compressive residual stress in the veneer layer significantly increases the flexural

strength of bilayer ceramic composite bars. However, global compressive residual

stresses that lead to greater tensile stresses under the contact point are also the main cause

for the observed chipping, i.e. spallation fracture of bilayer all-ceramic prostheses.














CHAPTER 6
OPTIMIZATION OF RESIDUAL STRESSES IN BILAYER DENTAL CERAMIC
COMPOSITES


In addition to material selection and geometric design, residual stresses in brittle

materials can be a major factor in the improvement of the apparent fracture toughness of

bilayer ceramic composites. Therefore, it is important to determine the magnitude and

distribution of residual stresses. Residual stresses can be caused because of rapid cooling

or mismatch in thermal coefficients of expansion in the components of a ceramic

composite. Also, existing cracks can propagate and cause failure due to residual stresses

during the use of the material.

Residual stresses can be tailored for increasing strength and apparent toughness of

the brittle materials. Previously, evaluation of residual stresses and enhancing its effects

through heat treatment techniques in metal-ceramic composites have been studied

(Dehoff and Anusavice, 1989; Hsueh and Evans, 1985; Scherrer, 1986; Anusavice et al.,

1989; Rekhson, 1979). However, the effect of residual stress on apparent fracture

toughness on bilayer dental ceramic composites is still relatively unknown.

In addition to the factors described above, the viscoelasticity (Appendix C) of the

glass components in ceramic composites is another determining cause of residual stresses

(Scherrer, 1986). When glasses are cooled from a temperature above their glass transition

temperature to one which is below it, they remain in a nonequlibrium thermodynamic

state, from which their properties thermodynamicc, mechanical, etc.) evolve slowly









towards equilibrium (Scherrer, 1989). This process must be taken into account to predict

accurately the long-term performance of glasses, especially when they are used in bilayer

dental ceramic composites.

Bilayer dental ceramics can exhibit a significant amount of residual stress

(Taskonak et al., 2002). These stresses influence the flexural strength and apparent

fracture toughness of bilayer ceramic restorations, mainly when the critical crack initiates

within the veneer surface. Global compressive residual stress in the veneer layer

significantly increases the flexural strength of bilayer materials. However, global residual

stresses may also be the main cause for the observed chipping or spelling fracture of

bilayer all-ceramic prostheses (Chapter 3).

The objective of this study is to demonstrate that a compressive residual stress can

be selectively distributed in bilayer dental ceramics to strengthen the material without

causing lateral crack propagation within the surface.

6.1. Materials and Methods

An analytical procedure based on fracture mechanics was used to obtain the

magnitude of residual stress in the ceramic veneer and determine its effect on lateral

crack growth based on measurements on the fracture surface. Marshall and Lawn (1979)

described an indentation technique to measure near-surface residual stresses. We also

used another technique in this study similar to the indentation technique based on the

superposition of stress intensity factors. The use of these two techniques provides an

opportunity to validate the correctness of our data. Conway and Mecholsky (1991)

proposed the analysis of residual stresses using fracture surface analysis. Our technique

includes five steps to determine the effect of residual stress on the lateral crack

propagation in bilayer ceramic core/veneer structures. These steps include:









1. Heat treatment of the bilayer ceramic specimens above, at, or below the glass

transition temperature of the applied glass veneer to introduce variable

magnitudes and distributions of compressive and tensile residual stresses at the

veneer surface.

2. Introduction of sharp indentation cracks using a Vickers indenter at load of 4.9 N

to produce controlled median-radial cracks and lateral cracks. These indentations

also introduce local residual stresses. Preliminary data suggest that lateral cracks

grow at an indentation load of 4.9 N (Taskonak et al., 2002).

3. Calculation of the flexural strength, apparent fracture toughness of heat-treated

bilayer ceramic specimens, and the "true" fracture toughness of monolithic glass

veneer specimens using fractographic analysis.

4. Calculation of the maximum residual stress in bilayer ceramic specimens using a

fracture mechanics equation.

5. Measurement of lateral cracks using fractography and determination of the change

in lateral crack length as a result of local and global residual stress differences.

6.1.1. Sample Preparation

For these procedures bilayer ceramic combinations were fabricated following

accepted dental laboratory procedures. Bar specimens were fabricated from a lithia-

disilicate-based core ceramic (Li20*2SiO2) (Empress 2, Ivoclar Vivadent AG, Schaan,

Liechtenstein) and a zirconia core ceramic (3Y-TZP) (Lava, 3M ESPE, Seefeld,

Germany). A hot pressing technique was used for the preparation of lithia-disilicate-

based core ceramic. However, zirconia core specimens were provided by the

manufacturer since CAD-CAM instruments were needed for their preparation.









Prepared core bars were veneered with a silicate glass (Empress 2 veneer, Ivoclar

Vivadent AG, Schaan, Liechtenstein) and Lava Ceram veneer (Lava, 3M ESPE, Seefeld,

Germany). Empress 2 veneer was applied to Empress 2 core ceramic and Lava zirconia

core was veneered with Lava Ceram veneer. Glass veneers have been selected with

coefficients of thermal expansion (CTE) compatible with those of their core ceramics to

ensure that the residual stress contribution caused by a CTE difference is negligible.

Ceramic veneer powders were incorporated with mixing liquid (Ivoclar Vivadent AG,

Schaan, Liechtenstein) to obtain slurry solutions, which were brushed onto the core

ceramics that were placed in slightly oversized silicone molds. Additionally monolithic

veneer specimens were prepared to serve as control groups.

The following core/veneer specimens were prepared:

Combination A: Lithia disilicate glass ceramic core (Experimental) / glass

veneer (Empress 2 veneer) (Appendix A)

Combination B: Zirconia core (Lava core) / glass veneer (Lava veneer)

(Appendix A)

Following sintering of the veneering ceramics using firing schedules listed in Table 6-

1 (See Fig 6-1 for fast cooling), the excess veneering ceramic was ground with a 75 grit

diamond disk and sequentially polished to a 2000 grit finish on a metallographic polisher

(Model 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while exposed to a continuous flow

of tap water. Fifteen specimens were prepared for each group. Monolithic veneer

specimens from Empress 2 veneer and Lava Ceram powders were prepared in addition

to the bilayer ceramic specimens.









Table 6-1. Glass veneers and firing schedules used for sintering
Heating Vacuum
Glass Starting Heating Firing T Holding t Cooling
rate T on-off
Veneer T (C) (C/m (0C) (min) o Rate
("C/min) (00
Lava Fast
450 40 810 1 451-810 F
veneer Cooling
Empress 2 403 60 802 1 450-801 Fast
Veneer Cooling



The final dimensions of the ceramic bilayer and monolithic bar specimens were 1.7

mm (height) x 4.0 mm (width) x 25.0 mm (length). The veneer thickness of bilayer

specimens was 0.6 mm for each specimen. The core/veneer thickness ratio was 6/11. The

testing span length/specimen thickness ratio was selected to be 15/1.7 to avoid large

beam deflections. After final polishing, dimensions of the specimens were measured with

a micrometer.

To optimize residual stresses, each core/veneer material group was heat-treated at

40C below and 20C, and 40C above, the glass transition temperature, and at the glass

transition temperature of the glass veneer (Table 6-2). In addition to the heat treated and

tempered specimens a control group for each bilayer material was included. Specimens in

the control group were neither heat-treated nor tempered. However their cooling rate

when veneered was considered as fast cooling (Fig 6-1). Profiles of surface temperature

of bilayer bars versus time for the two cooling conditions are shown in Figure 6-1. This

plot was described previously by Anusavice et al. (1989). Since the same cooling

techniques were performed, we used the data in the graph from that publication as a

reference. Tempering was performed by blasting compressed air directly on them as they

were removed from the furnace. A nozzle with a 4 mm diameter (Fig. 6-2) was placed 20

mm above the disc at a pressure of 0.34 MPa for 90 s.







69



Specimens with no heat treatment specimens were fast cooled as recommended by the


manufacturer using the cooling rate shown in Fig 6-1. Subsequently, specimens were


indented using a Vickers indenter at load of 4.9 N.


1000-

900-

800-

700 -

600 -

500 -

400-

300 -

200-

100 -


S I I I I I I I -
0 60 120 180 240 300 360 420
TIME (sec)


Figure 6-1. Profiles of surface temperature versus time for the two cooling conditions
(Anusavice et al., 1989).


SUPPORT BASE


INSULATING
BLANKET








SAGGER TRAY


MULTIPLE ORIFICE
NOZZLE (4 mm)

S0.34 MPa (50 psi)






20 mm






Sample


Figure 6-2. Design of tempering apparatus (Anusavice et al., 1989)


S\

\.
\
\


Fast Cooling



Tempered
-










Table 6-2. Heat treatment and cooling schemes for bilayer ceramic groups.
Grous Starting Heating rate Holding T Holding t C g
Groups Cooling Rate
T (oC) (oC/min) (0C) (min)
Lava Core/ Lava
veneer 400 60 525 60 Tempered
(Group 1)
Lava Core/ Lava
veneer 400 60 565 60 Tempered
(Group 2)
Lava Core/ Lava
veneer 400 60 585 60 Tempered
(Group 3)
Lava Core/ Lava
veneer 400 60 605 60 Tempered
(Group 4)
Lava Core/ Lava
veneer No heat No heat No heat No heat Fast cooling
veneer Fast cooling
treatment treatment treatment treatment
(Control Group)
Experimental Core
/ Empress 2 400 60 500 60 Tempered
Veneer (Group 1)
Experimental Core
/ Empress 2 400 60 540 60 Tempered
Veneer (Group 2)
Experimental Core
/ Empress 2 400 60 560 60 Tempered
Veneer (Group 3)
Experimental Core
/ Empress 2 400 60 580 60 Tempered
Veneer (Group 4)
Experimental Core
/ Empress 2 No heat No heat No heat No heat Ft
Fast cooling
Veneer (Control treatment treatment treatment treatment
Group)









6.1.2. Compositional, Physical and Thermodynamic Property Characterization of
Specimens

Glass transition temperatures of the glass veneers were determined using differential

thermal analysis (DTA) vs. temperature plots. The thermal expansion coefficients of each

ceramic composite component were measured using a single pushrod dilatometer and a

quartz standard (Orton Ceramic Foundation, Westerville, OH). Beam-bending viscometer

tests were performed on a lithia-disilicate-glass ceramic core (experimental core) to

determine its viscoelastic behavior (Appendix C) at high temperature since it has a glass

matrix.

Residual stress caused by the thermal mismatch of the veneer ceramic and core

ceramic was estimated using the equation below (Lawn, 1993):

CR=AcXAT/[(1 +vc)/2Ec+(1-2vv)Ev] (6-1)

where GR is the residual stress, Ac is the difference between linear thermal expansion

coefficients of the veneer and core ceramics, AT is the difference between the glass

transition temperature of the veneer ceramic and room temperature, and v and E are

Poissons's ratio and Young's modulus, respectively. Subscripts C and V refer to core and

veneer ceramics, respectively. In addition, the elemental composition of each ceramic

composite component was determined using X-ray diffraction. Youngs' moduli and

Poissons' ratios were measured using an ultrasonic technique. Also, the density of the

each ceramic component was measured using a pycnometer.









6.1.3 Mechanical Testing Methods

6.1.3.1 Determination of residual stresses using an indentation technique

A Vickers diamond pyramid indenter was used to produce well-defined radial and

lateral crack patterns at the center of each sample (Fig 5-1). Specimens were stored in air

for 48 h at room temperature to allow crack growth that is caused by residual stresses.

Microscopic examination of the contact sites was made to record post-indentation crack

development (Marshall and Lawn, 1980). Indentation induced radial cracks were

measured and recorded for each sample to observe crack behavior associated with

residual stresses.

The 2c dimensions of the radial half-penny cracks (Fig. 5-1) were measured for each

of ceramic bilayer composite and heat treatment condition. The actual fracture toughness

values of the monolithic glass specimens were determined from the following relation,

with use of values of load (P), crack size (c), and hardness (H) obtained from micro

indentation data:

Ke=1.6 x 10-(P/c2)(E/H)1/2 (6-2)

Where Kc = fracture toughness in MPa-m1/2 (actual fracture toughness of glass used for

cR calculations).

P = indentation load = 4.9 N

c = crack size (m)

E = elastic modulus (GPa)

H = hardness (GPa) = 0.47 P/a2

a = projected length of the indenter half-diagonal dimension (m)









The surface residual stress values in MPa for the heat treated (T) specimens were

calculated from the following relation, which was adopted from the theory developed by

Marshall and Lawn (1978):

GR = (KC/2)(XT/Cb)1/2[ 1-(Cm/Cb)3/2] (6-3)

where Cb and Cm are indent crack sizes (Fig 5-1) for the bilayer composite and

monolithic glass specimens, respectively.

6.1.3.2 Determination of residual stresses using fracture mechanics

The flexural strength of the specimens was determined using a four-point flexure test

fixture with a 15 mm outer span and a 5 mm inner span crossheadd speed of 0.5 mm/min)

using a universal testing machine (Model 4465, Instron Corp., Canton, MA). The veneer

surface was placed in tension for all flexure test specimens since the crack initiation sites

of the clinically failed FPDs were on the veneer surface (Chapter 3).

The flexure strength (of) of the bilayer ceramic composites was calculated using

mechanics, classical laminated-plate theory and the strength (of) of the monolithic

specimens was calculated using simple beam theory (Beer and Johnston, 1981)

(Appendix B).

Each specimen was studied with a stereoscopic microscope (Bauch & Lomb Inc.,

Rochester, NY, USA) at 160 X magnification. Crack initiating flaws and failure stresses

were measured to determine the fracture toughness of each specimen. Scanning electron

microscopic (JSM-6400, Jeol, Tokyo, Japan) examinations were performed on

representative specimens.









Apparent fracture toughness, Kc, was calculated using the fracture mechanics

equation:

Kc= Y(O) of(c)1/2 (6-4)



where Y(O) is a geometric factor equal to 1.65 for a surface indentation induced flaw with

local residual stress and 1.24 for surface flaws without local residual stress, for equivalent

semicircular flaws of depth "a" and half width "b" (Mecholsky, 1994), of is the

calculated flexural strength, and c is the crack size [c= (ab) ] (Mecholsky, 1991).

Residual stress was calculated from the following equation:



or = [Y(2/71/2) ac1/2 Kc] / [Y(2/T1/2)c1/2] (6-5)



where yr is the residual stress, Ca is the applied stress at failure, and Kc is the fracture

toughness of the glass veneers (Conway and Mecholsky, 1989).

Two-way analysis of variance was performed to determine whether the

differences between group means for flexural strength and residual stresses of specimens

are statistically significant. Statistical analysis was also performed to determine whether

the differences between the mean lateral crack sizes were statistically significant

depending on the local and global residual stresses of the groups.

6.2. Results

Measured thermal and physical properties of ceramic composite components are

given in Table 6-3.









Table 6-3. Physical, mechanical and thermal properties of core and veneer ceramics
Lava Veneer Experimental Empress 2
S Lava Core Core Veneer
Material Properties (zirconia) (Amorphous (Li20-2SiO2 (Amorphous
glass) glass-ceramic) glass)

Density (p) 6.22 g/cm3 2.53 g/cm3 2.56 g/cm3 2.53 g/cm3

Elastic modulus (E) 155 GPa 58 GPa 96 GPa 64 GPa


Poisson's ratio (u) 0.34 0.27 0.24 0.23

Tg 5650C 6870C 540C
CTE (ct)
10.7 (ppm.K-1) 10.2 (ppm.K-1) 10.4 (ppm.K-1) 9.8 (ppm.K-1)
(250C-6000C)
Kc 5.5 MPaml/2 0.7 MPa-m1/2 3.1 MPa-ml2 0.7 MPa-ml2

Vickers Hardness (H) 11.2 GPa 4.4 GPa 6.1 GPa 4.9 GPa

Density 6.22 g/cm3 2.53 g/cm3 2.56 g/cm3 2.53 g/cm3
Flexure Strength (o)
634 MPa 53 MPa 228 MPa 52 MPa
(4-point flexure)

X- ray diffraction analysis was performed to identify any crystal phases in the

veneer layers. However, the veneer layers were found to be essentially amorphous by X-

ray diffraction analysis because there were no distinct peaks observed in the graph

between the angles of 5 and 110 degrees. X-ray diffraction analyses of Empress 2

veneering ceramic revealed a large amorphous background signal that represents the

concentration of the glass phase. However, we observed the main peaks of lithia disilicate

in the experimental core ceramic (Li20-2SiO2 glass-ceramic) at diffraction angles (20) of

23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the dominant peak (highest

intensity) at 25.0 degrees, which corresponds to the (111) crystallographic plane of the

monoclinic phase. Also, the other veneering ceramic, Lava veneer, exhibits an









amorphous background with three minor peaks. These peaks were observed at 20

diffraction angles of 26.7, 34.0, and 51.7 degrees that revealed a slight presence of alpha

quartz. The results are shown in Fig. 6-3. Lava core ceramic is 3Y-TZP (Suarez et al.,

2004) (Appendix A).


Figure 6-3. (A) X-ray diffraction pattern of Empress 2 veneer shows an amorphous
structure. (B) X-ray diffraction pattern of Lava veneer shows slight alpha
quartz peaks at 20 values of 26.7, 34.0, and 51.7 degrees. (C) X-ray diffraction
pattern of an experimental core ceramic reveals lithia disilicate peaks at 20
values of 23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the
dominant peak (highest intensity) at 25.0 degrees.




Dilatometric analysis showed that there was a slight difference between the

thermal expansion coefficients of Empress 2 veneer and Experimental core ceramics

(Table 6-3). The values were 9.8 (ppm-K-1) and 10.4 (ppm-K-1) for the Empress 2 veneer

and experimental core ceramic, respectively. The glass transition temperature (Tg) of the

Empress 2 veneer ceramic was determined to be 5400C and the difference between Tg


I .' '. _









and room temperature is 5150C. The Poisson's ratios of the veneer ceramic (vv) and the

core ceramic (vc) are 0.23 and 0.24, respectively. The Young's moduli of the veneer

ceramic (Ev) and core ceramic (Ec) are 64 GPa and 96 GPa, respectively. The calculated

residual stress for the specimens using the parameters above and equation (6-1) is 13

MPa (compression).

In addition, dilatometric analysis also showed that there was a slight difference

between the thermal expansion coefficients of Lava veneer and Lava core ceramics

(Table 6-3). The values were 10.2 (ppm-K1) and 10.7 (ppm-K-)for the Lava veneer and

Lava core ceramics, respectively. The glass transition temperature (Tg) of the Lava

veneer ceramic is 5650C and the difference between Tg and room temperature is 5400C.

The Poisson's ratios of the veneer ceramic (vv) and the core ceramic (vc) are 0.27 and

0.34, respectively. The Young's moduli of the veneer ceramic (Ev) and core ceramic (Ec)

are 58 GPa and 155 GPa, respectively. The calculated residual stress for the specimens

using the parameters above and equation (6-1) is 22 MPa (compression).

Beam-bending viscometer data revealed creep occurring above 7000C in

experimental core ceramic (Li20*2SiO2 glass-ceramic) (DeHoff, personal

communications). As a result we calculated the relaxation function for the experimental

core ceramic. This plot revealed a glass transition temperature for experimental core

ceramic which is 6870C (Fig 6-4). Shown in Fig 6-4 is the plot for the experimental core

for which the intersection of the two straight lines occurs at a temperature of

approximately 6870C (Tg). Although the transition from liquid to glass occurs over a

range of temperatures, we have selected the intersection of the two lines as the glass

transition temperature (Tg) for the experimental core ceramic.














40

35

30

25

20

15

10

5

0


0.00094 0.00099 0.00104 0.00109 0.00114 0.00119 0.00124


1/T(K')


Figure 6-4. Natural logarithm of shear viscosity (rs) versus inverse absolute temperature
for experimental ((Li20*2SiO2 glass ceramic) core (DeHoff, personal
communications).





For all specimens, the mean longitudinal and transverse indentation-induced crack


sizes with standard deviation of each group are summarized in Table 6-4. Based on one-


way ANOVA analysis, the differences between the mean indentation-induced


longitudinal and transverse crack sizes of bilayer specimens are not significantly different


(p > 0.05). However, the difference between mean indent crack sizes of monolithic


specimens and bilayer specimens are statistically significant (p < 0.05).


For groups with a lithia-disilicate-based glass-ceramic core, a one-way ANOVA


analysis showed that there was a statistically significant difference between the mean


flexure strengths of monolithic and bilayer specimens (p < 0.05), but there was no


y = -395 63x + 34.52
R2 0.0016
y= 71 547x 40.504
R = 0.8557


Series A
Series B
A Series C
low temp
x high temp









statistically significant difference (Table 6-4a) between the mean flexural strength of

bilayer specimen groups (p > 0.05).


Table 6-4a. Mean flexural strength (o), standard deviation (SD), residual stress (CR) from
fracture surface analysis (FSA) and indent crack sizes, longitudinal and
transverse indentation induced crack sizes, residual stresses of groups
calculated using flexural strength differences between monolithic veneer
(of(m)) and bilayer specimens (of(b)) and apparent fracture toughness (K,) of
bilayer (Experimental core/Empress 2 veneer) and monolithic veneer
specimens.


Transverse
Groups Longitudinal Transverse
indent CR S.D. GR S.D. TR
T S.D. indent crack K
(Experimental entcrack (Indent) (FSA) C)
core/Empress 2 (MPa) length ln t (MPa-ml2)
veneer) (pm) ((MPa) (Ma) (MPa)

Tg-40 64 +11 76 14 74 12 -60 48 -18 8 12 1
(5000C)



Tg 67 +6 76 +16 73 + 15 -70 +61 -19 +11 15 1
(5400C)



Tg+20 64 +10 69 +10 76 12 -50 + 37 -21 + 14 12 1
(560C)



Tg+40 67 +9 74 +9 80 + 6 -39 + 34 -16 +7 15 1
(5800C)



No heat 66 6 72 +14 84 14 -20 + 16 -14 + 12 14 1
treatment

No heat 52 +10 93 +22 92 +15 0 0 0 0.7*
treatment
(Empress 2
Veneer)

Actual fracture toughness









Specimens in groups with zirconia core delaminated at the veneer-core interface

without exception (Fig 6-5). In order to determine the flexural strength of the glass

veneer in these specimens, composite beam theory (Beer and Johnston, 1981) (Appendix

B) was used; however, simple beam theory was used for calculation of failure stress for

the remaining zirconia layer that failed after interface delamination. Failure stresses of

both glass layers and zirconia layers are shown in Table 6-4b. There was a statistically

significant difference between the mean veneer layer failure stresses of the group heat-

treated at 5250C and the group that was not heat treated (control) (p < 0.05). There was

no statistical difference between the glass layer failure stresses of other heat treatment

groups and the control group (p > 0.05). Also, there was no statistical difference between

the mean failure stresses of the zirconia core ceramics (p > 0.05).


Figure 6-5. Interfacial delamination between zirconia core and a glass veneer layer.










Table 6-4b. Mean flexural strength (o), standard deviation (SD), residual stresses (GR) of
groups calculated using flexural strength differences between monolithic
veneer (of(m)) and bilayer specimens (Gf(b)), longitudinal and transverse
indentation induced crack sizes, and apparent fracture toughness (K,) of
bilayer (Lava core/Lava veneer) and monolithic veneer specimens.

Transverse
Longitudinal
Groups (core) nitd indent 1R S.D.
Y(veneer)> S.D. indent crack rack
(Lava core/Lava S.D. length crack Of(b) Of(m)
(Lava core/Lava (N) length length
veneer) (Ma) (MPa) (n) (MPa)


Tg-40 (5250C) 146 13 659 166 78 + 9 80 10 -93 + 40




Tg (5650C) 123 11 553 +116 84 +15 84 10 -70 11




Tg+20 (5850C) 118 +15 615 +102 85 10 90 +12 -65 15




Tg+40 (6050C) 119 16 631 117 68 +10 73 +9 -66 +16




No heat 112 7 629 151 66 +6 72 +14 -59 7
treatment



No heat 53 + 6 101 +39 103 +30 0
treatment
(Lava Veneer)



For groups with a lithia disilicate crystal phase, failure origins were located within

the tension surface of each specimen and not at the interface between the veneer and core.

For most of the composite specimens, the indentation cracks were the critical flaws that

controlled crack propagation (Fig 5-2a). However, some specimens failed from porosity









at the surface. These critical flaws, within the tensile region of the composite bars were

approximately semi-elliptical in shape (Fig. 5-2b). Also, for heat-treated groups

spallation occurred within the veneer surface during fracture and it was not possible to

observe the fracture origin (Fig. 6-6). Since delamination occurred in specimens with a

zirconia core following the fracture of the glass veneer, actual fracture origins were lost

and it was not possible to perform fracture surface analysis (Fig. 6-5).


I 100um _


4F


Figure 6-6. The fracture origin of the specimen was not detectable. Wake hackle
markings indicate the direction of the fracture origin in the veneer layer.
However, the part of the specimen containing the fracture origin chipped away
from the veneer layer.


Even though all specimens failed from the veneer surfaces, there was a

statistically significant difference between the calculated apparent fracture toughness

values (Eq. 6-4) of bilayer specimens with lithia disilicate glass-ceramic based core









(experimental core) and monolithic veneer (Empress 2 veneer) specimens (p < 0.05)

(Table 6-4a). The fracture toughnesses of monolithic Empress 2 veneer and Lava veneers

were identical [0.7 MPa-m1/2] (Table 6-3.). However, the fracture toughness of the

zirconia core (Lava core) was 5.5 MPa-m1/2 whereas the fracture toughness of lithia

disilicate-based glass-ceramic core (experimental core) was 3.1 MPa-m/2 (Table 6-3.).

Residual stress values of experimental core (lithia-disilicate-based glass-ceramic)

/Empress2 veneer specimens (Eq. 6-5) ranged between -24 MPa and -14 MPa

compressivee) (Table 6-4). The highest compressive residual stress was observed in a

group that was heat treated 200C above the glass transition temperature, whereas the

lowest residual stress was observed in the group with no heat-treatment. No residual

stress was detected in the monolithic specimens (Table 6-4a). There was no statistically

significant difference between the residual stresses of the heat-treated groups (p > 0.05).

Residual stress values of Lava zirconia core/Lava veneer ranged between -92

MPa and -58 MPa (compression) (Table 6-4b). The highest residual compressive stresses

were present in the group that was heat treated 200C below the glass transition

temperature whereas the lowest residual stress was observed in the group with no heat

treatment. No residual stress was detected in the monolithic specimens (Table 6-4b).

There was a statistically significant difference between the residual stresses of the heat

treated groups (p < 0.05). Since the glass veneer layers of these specimens were

delaminated and fractured in many pieces (Fig.6-5), fracture surface analysis was not

possible to perform. As a result equation 6-5 could not be used to determine residual

stresses. Instead differences between the mean failure stresses of glass veneer layers of









heat treated groups and the mean strength of monolithic glass sample were used to

calculate residual stresses in the Lava zirconia core/Lava glass veneer specimens

Residual compressive stresses obtained using the indentation technique were

greater than the failure stresses of experimental core/Empress 2 bilayer ceramic

composites. As a result, the data obtained using this technique were ignored and the data

from fracture surface analysis were used for statistical analysis.

6.3. Discussion

It is important to understand how the stress distribution in dental ceramics can be

used to predict failure mechanisms. Stress development mechanisms become more

complex especially when trilayer or bilayer composite ceramics are used for dental

restorations. Clinical success is mostly predicted by the manufacturers using thermal

expansion mismatch data of the components in multilayer ceramic composites (DeHoff

and Anusavice, 2004).

Bilayer dental ceramics can exhibit a significant amount of residual stress. These

stresses influence the flexural strength and apparent fracture toughness of bilayer all-

ceramic restorations and occur mainly when the critical crack initiates within the veneer

surface. Global residual compressive stress in the veneer layer significantly increases the

flexural strength of bilayer materials. However, global residual stresses may also be the

main cause of chipping or spelling fracture in bilayer all-ceramic prostheses.

Calculations of residual stress in bilayer dental ceramics using a fracture mechanics

equation (Conway and Mecholsky, 1989) suggest that there is a significant amount of

global residual compressive stress within the veneer layers of some bilayer materials

(Table 6-4). The distinction between global and local residual stress should be









emphasized. Local residual stress refers to the residual stresses induced by indentation.

The stresses are near the indentation site and drop off rapidly. The distribution of these

stresses can be represented by the three-dimensional Boussinesq stress field (Lawn, 1993)

(Fig. 1-1). Each of these residual stress systems is not uniform. Thus, in detailed

calculations an assumption must be made as to the nature of the stress profiles. In the

present case, we assume that the residual stresses in the vicinity of the crack are locally

constant and can be superimposed. Global residual stress refers to overall stresses

distributed from the surface toward all over the sample.

Consequently, tensile or compressive stresses will develop and increase in the

glass. At temperatures lower than the glass transition temperature of the glass veneer,

viscosity is high and molecular motions are very slow. Thus, it will take a long time for

the glass to stabilize. At higher temperatures, but lower than the glass transition

temperature of the glass veneer, the viscosity is low and the glass stabilizes rapidly upon

cooling, and residual compressive stress will develop within the glass layer (Anokye,

1989).

In our study we observed a significant difference between the residual stress in

experimental core/Empress 2 bilayer ceramics developed by contraction mismatch (13

MPa), and the total residual stress (21 MPa) calculated using Equation 6-5. Difference is

mostly caused by the viscoleastic relaxation of the glass layer near its glass transition

temperature. In addition there were slight changes in residual stress values between heat

treatment groups. These changes show that residual stresses can be increased or

decreased using different heat treatments in the ceramic-ceramic composites. However,

the slight changes are most likely associated with the viscoelastic behavior (Appendix C)









of the experimental core ceramic above 687C. Viscoelastic behavior of the glass layer

controls the magnitude of residual stress during cooling from 6870C to 5400C (Tg).

The possibility that crystallization of the veneer layer can result in an increase in

strength and toughness of the bilayer composites compared with the glass veneer is

rejected because X-ray diffraction analysis revealed no evidence of a crystal phase,

indicating that the veneering ceramics consist predominantly of an x-ray amorphous glass

phase. Thus, the strengthening mechanism for the bilayer composites is not caused by

crystallization of the veneer layers.

The results of the evaluation of residual stress as a function of heat treatment around

the glass transition temperatures are shown in Figures 6-7a and 6-7b. During the

isothermal heat treatment of the bilayer ceramic composites in the glass transition range,

the density of the glass veneer will increase because of structural relaxation, while there

is no change in the ceramic core (Scherrer, 1986). Depending on the viscosity of the glass

veneer at the heat treatment temperature and cooling rate, either tensile or compressive

global residual stresses will develop (Scherrer, 1986, Anokye, 1989). For example at

temperatures below Tg, e.g., 5000C, for the Empress 2 glass veneer, the viscosity is high

and molecular motions are very slow. As a result it will take long time for the glass to

stabilize (Fig. 6-7a). At temperatures above Tg, e.g., 5800C, the viscosity is low and the

glass stabilizes rapidly, immediately inducing stress relaxation. Near the glass transition

temperature, i.e., 540C, the structure does not stabilize as quickly as it does at higher

temperatures (Fig. 6-7a). Also, we have to account for the fact that the glass transition

temperature range is a function of the cooling rate. Additionally, the cooling rate will






87


determine the amount of stress relaxation. These factors can play an important role in

evaluation of the global residual stresses in bilayer systems.


0 -
480


500 520 540 560 580
Heat Treatment Temperature C (1 h hold)


600


Figure 6-7a. Residual stress as a function of heat treatment temperature in bilayer lithia-
disilicate core/glass veneer ceramic composites [Residual stresses are given in
Table 6.4a: OR (FSA)].




In the zirconia-glass bilayer ceramic system (Lava) all specimens delaminated, which

is considered a failure for dental ceramic composites (Fig 6-5).

Residual stresses in the zirconia core-glass veneer system (Lava) were greater than

these in the lithia disilicate core-glass veneer system (Fig. 6-7b). This may be associated

with the viscoelastic behavior of the lithia disilicate glass ceramic core above 687C.

However, zirconia is elastic through the entire heating-cooling temperature range. Also










this factor provides a more significant residual stress difference between heat treatment

groups of the zirconia core-glass veneer system (Lava) (Fig. 6-7b).


100

90

80

70

60

50

40

30

20

10

0
520


540 560 580 600
Heat Treatment Temperature oC (1 h hold)


Figure 6-7b. Residual stress as a function of heat treatment temperature in bilayer
zirconia core/glass veneer ceramic composites [Residual stresses are
summarized in Table 6.4b: GR = Of(b)- Of(m)].



The reason that the zirconia core-glass veneer system (Lava) exhibited the greatest

residual stress may be because the glass transition temperature (Tg) shifts as a function of

cooling rate. The glass transition temperature was determined to be 5650C from DTA

experiments. However, during processing the cooling rate is faster and thus the Tg most

likely shifts to a lower temperature. This shift increases the viscosity of the glass at lower




Full Text

PAGE 1

THE EFFECTS OF RESIDUAL STRESS, VISCOELASTIC AND THERMODYNAMIC PARAMETERS ON APPARENT FRACTURE TOUGHNESS OF DENTAL BILAYER CERAMIC COMPOSITES By BURAK TASKONAK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Burak Taskonak

PAGE 3

This dissertation is dedicated to my loving family, Leyla, Erim and Bahad r Taskonak.

PAGE 4

iv ACKNOWLEDGMENTS There have been many persons whose empowerment gave me the strength, confidence and motivation while working on th is project. Without their presence and support such a challenging process would not have been such a rewarding experience. I would like to first and foremost tha nk Dr. Kenneth John Anusavice and Dr. John Joseph Mecholsky, Jr., whose work ethic, compassion, support, understanding, and guidance inspired me through this project. Se cond, I would like to thank my dearest friends, Ibrahim Keklik and Deniz Rende, w hose love and support were reflective of a family. I owe special thanks to my mother, Leyl a Taskonak, who by her brilliance, vision and courage set an astonishing example for my personal and professional growth. I thank my father, Erim Taskonak, whose hard wor k, generosity, and principles taught me truthfulness. Also, I th ank my brother Bahad r Taskonak, whose presence in my life alone is an enjoyment. I am forever indebted to my colleague s and friends Dr. Karl-Johan Sderholm, Dr. Chiayi Shen, Mr. Ben Lee, Ms. Allyson Barrett, Mr. Nai Zheng Zhang and Ms. Amy Corrbitt who exemplified sensitivity to my personal needs, and whose conversations sustained the spirit of my professional de velopment. I thank my lovely friends and colleagues at the University of Florida, Department of Materials Science and Engineering, whose support and encourag ement showed me th e universality of friendship.

PAGE 5

v Finally, I am thankful to my othe r committee members Dr. Wolfgang Sigmund, Dr. Darryl Butt, Dr. Mark Yang, and Dr. Art hur (Buddy) Clark for their great support, guidance and scholarship.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS................................................................................................iv LIST OF TABLES..........................................................................................................viii LIST OF FIGURES..........................................................................................................ix ABSTRACT......................................................................................................................x i CHAPTER 1 INTRODUCTION.........................................................................................................1 2 TWO-YEAR CLINICAL EVALUATION OF LITH IA-DISILICATE-BASED CERAMIC FIXED PARTIAL DENTURES.................................................................6 2.1. Materials and Methods............................................................................................7 2.2. Results.................................................................................................................. .13 2.3 Discussion..............................................................................................................16 3 QUANTITATIVE FRACTURE SURFACE ANALYSIS OF CLINICALLY FAILED CERAMIC FIXED PARTIAL DENTURES.................................................19 3.1 Materials and Methods...........................................................................................20 3.2. Results.................................................................................................................. .23 3.3. Discussion.............................................................................................................27 4 ROLE OF INVESTMENT INTERAC TION LAYER ON STRENGTH AND TOUGHNESS OF CERAMIC LAMINATES............................................................31 4.1. Materials and Methods..........................................................................................33 4.2. Results.................................................................................................................. .39 4.3. Discussion.............................................................................................................42 5 RESIDUAL STRESSES IN BI LAYER DENTAL CERAMICS................................48 5.1. Materials and Methods..........................................................................................49 5.1.1. Specimen Preparation.....................................................................................49 5.1.2 Mechanical Testing Methods...........................................................................51

PAGE 7

vii 5.2. Results.................................................................................................................. .54 5.3. Discussion.............................................................................................................58 6 OPTIMIZATION OF RESIDUAL STRESSES IN BILAYER...................................64 DENTAL CERAMIC COMPOSITES........................................................................64 6.1. Materials and Methods..........................................................................................65 6.1.1. Sample Preparation.........................................................................................66 6.1.2. Compositional, Physical and Therm odynamic Property Characterization of Specimens.................................................................................................................71 6.1.3 Mechanical Testing Methods...........................................................................72 6.2. Results.................................................................................................................. .74 6.3. Discussion.............................................................................................................84 7 CONCLUSIONS..........................................................................................................92 APPENDIX A COMPOSITIONAL CHARACTERIZATION OF CERAMIC COMPOSITE COMPONENTS.........................................................................................................95 B STRESS CALCULATIONS USING COMPOSITE BEAM THEORY.....................97 C VISCOELASTIC THEORY AND FRACTURE MECHANICS METHODS FOR DETERMINATION OF RESIDUAL STRESSES...................................................101 C.1 Viscoelastic Theory............................................................................................101 C.2 Determination of Residual St resses Using Fracture Mechanics........................104 D APPARENT FRACTURE TOUGHNESS AND RESIDUAL STRESS DATA......107 LIST OF REFERENCES...............................................................................................127 BIOGRAPHICAL SKETCH.........................................................................................131

PAGE 8

viii LIST OF TABLES Table Page 2-1 Location, survival time (in months) a nd failure type of Empress 2 FPDs..................8 2-2 Number of Empress 2 crowns acco rding to their evaluation time..............................9 2-3 Citeria for the direct evaluation of the restorations..................................................12 2-4 FPD restorations: Scores of the clinical evaluation (%) at baseline, year one and year two.....................................................................................................................14 2-5 Crown restorations: Scores of the clini cal evaluation (%) at ba seline, year one and year two.....................................................................................................................15 2-6 Scores of the clinical evaluation for plaque and gingival index (%) at baseline, year one, and year two..............................................................................................15 3-1 Failure stress ( f), residual stress ( R), critical flaw size (c), semiminor axis(a), and semimajor axis (2b) of the FPD specimens........................................................27 4-1 Mean flexural strength ( ),characteristic strength ( 0), critical Flaw size (c) and Weibull modulus (m) values of bilayer ceramic laminates......................................41 5-1 Mean flexural strength ( ), residual stress ( R), indentation-induced crack sizes and apparent fracture toughness (Kc) of bilayer and monolithic specimens.............55 6-1 Glass veneers and firing sc hedules used for sintering..............................................68 6-2 Heat treatment and cooling sche mes for bilayer ceramic groups.............................70 6-3 Physical, mechanical and thermal prop erties of core and veneer ceramics..............75 6-4 Mean flexural strength ( ), residual stress ( R) indent crack sizes, and apparent fracture toughness (Kc) of bilayer and monolithic veneer specimens......................79 A-1 Calculation of centroids for each type of core/veneer combination.........................99 C-1 Elastic – Viscoelastic Analogy...............................................................................103

PAGE 9

ix LIST OF FIGURES Figure Page 1-1 Boussinesq stress field, for principal normal stresses 11, 22 and 33.......................3 2-1 Facial view of prepared maxillary left central incisor, maxillary left canine, mandibular right central incisor and mandibular left lateral incisor.........................10 2-2 Finished prostheses that are placed on the prepared teeth........................................10 2-3 Fractured connector in anterior Empress 2 fixed partial denture after 11 months of clinical service......................................................................................................14 2-4 Kaplan-Meier survival statistics of Em press 2 fixed partial dentures (n=20)..........16 3-1 Arrest lines in the form of ridges ar e present on the fracture surfaces One of the arrest lines is a result of the in tersection of two fracture paths.................................24 3-2.Wake hackle markings were used to establish the reference points for determining the fracture origin in the glass veneer.......................................................................24 3-3 Fracture propagated quickl y in the glass veneer without any increase in stress from the point where the semi-minor axis of the fracture origin ends..............................29 4-1 Cross section of Empress 2 core ceramic with residua l investment interaction layer on its outer surface...........................................................................................39 4-2 Fracture surface of ceramic laminate (SEM view)...................................................42 4-3 Representative graph of flexural strength ( f) versus c-1/2........................................44 4-4 Representative graph of flexural strength ( f) versus c-1/2 for Empress 2 veneer monolithic specimens................................................................................................45 5-1 Schematic illustration of an indented beam specimen placed in four-point flexure Indent cracks were induced in the tension surface....................................................52

PAGE 10

x 5-2 Fracture surface and fracture origin of a monolithic veneer specimen that reveal an indent crack as a fracture origin...........................................................................56 5-3 SEM image of a fracture surface of an E2 V specimen that exhibits a large pore as the fracture origin......................................................................................................57 5-4 Theoretical calculations a nd experimental measurements of residual stress in borosilicate glass resulting from viscoelastic relaxation behavior.........................61 5-5 Stress distribution in layers of a bilaye r specimen when the outer layer is in compression..............................................................................................................62 5-6 Schematic illustration of radial and late ral cracks at the surf ace with or without residual stress............................................................................................................62 6-1 Profiles of surface temperature versus time for the two cooling conditions.............69 6-2 Design of tempering apparatus (Anusavice et al ., 1989)..........................................69 6-3 X-ray diffraction patterns of ceramic components...................................................76 6-4 Natural logarithm of shear viscosity ( s) versus inverse absolute temperature for experimental ((Li2O•2SiO2 glass ceramic) core.......................................................78 6-5 Interfacial delamination between zirc onia core and a glass veneer layer................80 6-6 A fracture surface of a specimen wher e the fracture origin is lost...........................82 6-7 Residual stress as a function of heat treatment temperat ure in bilayer lithia-disilicate core/glass ve neer ceramic composites.............................................87 6-8 Fracture initiation from a me dian (radial) indent crack............................................91 6-3 X-ray diffraction patterns of ceramic components...................................................96 A-1 Distances for the calculation of centroid of the composite parts.............................99 A-2 Distances from the centroidal-axis and the centroid of the composite parts..........100

PAGE 11

xi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF RESIDUAL STRESS, VISCOELASTIC AND THERMODYNAMIC PARAMETERS ON APPARENT FRACTURE TOUGHNESS OF DENTAL BILAYER CERAMIC COMPOSITES By Burak Taskonak August 2004 Chair: Kenneth J. Anusavice Cochair: John J. Mecholsky, Jr. Major Department: Materials Science and Engineering Bilayer dental ceramic composites used for fixed partial dentures are becoming more widely used in dental practices b ecause of their biocompatibility, aesthetic properties, and chemical durability. However, la rge statistical variations in the strength of ceramics are associated with the structural flaws as a result of processing and complex stress states within the surfaces of the ma terials because of thermal properties of each layer. In addition, partial dela minations of the veneer layer and connector fractures of bilayer ceramic fixed partial dentures (FPD s) have been observed in a clinical study which is a part of this dissertation. Analys is of fracture surfaces of failed FPDs reveals that such fractures of the veneering cera mic are most likely cau sed by lateral crack growth. Global residual stresses associated with the coefficient of thermal expansion differences between core and veneering ceram ics can cause lateral crack initiation. Also,

PAGE 12

xii rapid cooling of bilayer ceramic s from the sintering temperat ure of the glass veneer may not allow the interfacial stresses in the viscoe lastic glass to relax to equilibrium values. This can further contribute to the propaga tion of lateral cracks. Furthermore, local residual stresses that develop in the plastic deformation z one below sharp contact areas on the occlusal surface are another contributo r to lateral crack grow th. Superposition of global residual stresses and a Boussinesq stress field can incrementally increase the possibility of lateral crack gr owth. The long-range goals of this study are to critically analyze the lateral crack growth mechanisms a ssociated with residual stresses, to modify residual tensile stress distribu tions by controlled heat trea tment, and to minimize the probability of veneering ceramic fractures. Four approaches were used to accomplish th ese goals: (1) clinic al evaluation of a bilayer ceramic fixed partial denture system; (2) fracture surface analysis of clinically failed FPDs; (3) determination of residual st resses using fracture mechanics techniques; and (4) optimizing residual stresse s using heat treatment methods. This study suggests that the compressive global residual stre sses within the ceramic surface can strengthen the material; however, excessive compressive residual stresses can cause lateral cracks to grow and propagate to the surface, which will eventually cause failure of the material. When a glass layer is used in a bilayer ceramic composite, heat treatment above and be low the glass transition temperature (Tg) of this glass will induce different magnitudes of stress es within the surface of the material. This phenomenon can be used to modify the residual stresses and reduce the risk for fracture.

PAGE 13

1 CHAPTER 1 INTRODUCTION The use of all-ceramic systems by dental laboratory technicians and dentists has increased because of their enhanced aesthetic potential compared with metal-ceramics (Piddock and Qualtrough, 1990). However, partia l delamination of the veneering ceramic in bilayer fixed partial dentures (FPD) has b een reported in clinical studies by Sorensen and Cruz (1998), Pospiech et al. (2000), and Edelhoff et al. (2002). Residual stresses play an important role in the reliability and failure of dental materials, especially in bilayer and trilayer ceramic and glass com posites. Two classes of residual stresses are important, global and local residual stre sses. Global residual stresses are usually introduced during fabrication by cooling the external surface of a ceramic structure faster than the in terior of the structure. If designed properly, the thermal expansion and contraction mismatch between adjacent layers of a bilayer composite should produce global residual compressive stre ss within the surface of the desired region in a ceramic prosthesis. These global residual compressive stresses generally increase the effective strength of the component. Local resi dual stresses are presen t at indentation or impact sites. These stresses can be descri bed using a Boussinesq stress field surrounding the contact event. This stress distribution is considered local because the magnitude of the stress decreases rapidly in magnitude and is concentrated only around the contact area. The residual stresses remain because plastic (inelastic) deformation occurs as a result of

PAGE 14

2 the contact load. Local residual stresses are both tensile and compre ssive in nature and the tensile stresses control the fracture initiation and propa gation processes. Lateral cracks can initiate near the base of the plastic deformation zone below the point of contact loading and they spread late rally on a plane approxima tely parallel to the specimen surface (Marshall et al., 1982, La wn and Swain, 1975; Lawn and Wilshaw, 1975). In a severe contact event these crack s can propagate to the surface to cause localized ceramic fragmentation. Lawn and Swain (1975) and Lawn et al (1980) reported that lateral crack extension occurs during indenter unloadi ng in much the same way as radial cracks are produced by sharp-contact fracture. Residual stress is the primary driving force for lateral cracking (Marshall et al., 1982) that can result in spallation. If the two materials in a bilayer composite are crystalline (or pol ycrystalline), then the primary behavior is elastic, the global re sidual stress is well defined, and the analysis is straightforward. However, if one or both of the materials is am orphous, such as glass veneer, the glass will exhibit viscoelastic behavior and the determination of the global residual stress becomes much more complex a nd dependent on cooling history relative to the glass transition temperature, Tg. In addition, the local stress field surrounding the contact site is affected by residual stresses and complex crack characteristics (locations and geometry). Associated with the indentat ion or impact site are two types of crack geometries, radial and lateral. The radial cracks generally co ntrol strength and the lateral cracks control the pote ntial for spallation. An axially symmetric Boussinesq stress field can be superimposed on residual stresses during contact at the point of loadi ng (Fig. 1-1). Superposition of global residual stresses associated with viscoelastic rela xation or rapid coolin g of the core/veneer

PAGE 15

3 ceramic composites on the Bou ssinesq contact stress field can lead to early lateral fracture of the material. These conditions will allow a better understanding of these failure mechanisms and the reasons for spallation from the surface of all ceramic prostheses. In view of the key role played by lateral fract ure in the spallation and wear processes (Lawn and Swain, 1975) for brittle substrates, it is im perative that the mechanism(s) of lateral cracking be analyzed to accurately estimate the survival time of ceramic dental prostheses. Figure 1-1. Boussinesq stress fi eld, for principal normal stresses 11, 22 and 33, (a) Stress contours, half surface view (top) and side view (bottom). (b) Stress contours, side view (Johnson, 1985). The specific aims of this study are as follows: Specific Aim 1: To determine the two year clinical survivability of a bilayer ceramic FPD system.

PAGE 16

4 This study was designed to evaluate the cl inical performance of Empress 2 system over a two-year clinical period Specific Aim 2: To characterize the fracture features and failure stresses of clinically failed FPDs. Specific Aim 3: To test the hypothesis that the in teraction of a core ceramic with investment material can significantly redu ce the flexural stre ngth and the fracture toughness of core / veneer ceramic laminates. Specific Aim 4: To investigate the magnitude of residual stresses in bilayer ceramic composites and determine their effects on a pparent fracture toughness. The hypothesis of this aim is that the global residual stresses within the surface of veneered all ceramic fixed partial dentures (FPD) ar e responsible for both the obs erved strength increase in bilayer dental ceramics and for partia l spallation of the veneering ceramic. Specific Aim 5: To determine the threshold levels of global compressive residual stress that will strengthen the core/veneer dental ceramic composite and minimize the risk for lateral crack propagation. The objective of this aim is to demonstrate that a compressive residual stress can be se lectively distributed in bilayer de ntal ceramics to strengthen the material without causing lateral cr ack propagation within the surface. Two fracture mechanics approaches were de signed to analyze residual stresses in this study: determination of residual stre sses using fracture surface analysis, and determination of residual stresses using an indentation technique. Chapter two will demonstrate the clinical failure source of all ceramic FPDs. Chapter three will identify the major cause of these clinical failures using fractographic techniques. Chapter four will show that invest ment material that is used in bilayer all

PAGE 17

5 ceramic FPD fabrication is not the cause of fa ilures, nor does it contribute significantly to the failures. Chapter five will prove that there is a significant level of residual stress that can increase strength and at can also cause spallation failures. Chapter six will demonstrate how residual stress can be c ontrolled to potentially improve clinical reliability.

PAGE 18

6 CHAPTER 2 TWO-YEAR CLINICAL EVALUATION OF LITHIA-DISILICATE-BASED CERAMIC FIXED PARTIAL DENTURES Dental ceramics are known for their na tural appearance and long-term color stability. Most recently, the popularity of a ll-ceramic systems has increased because of their enhanced aesthetic pot ential compared with me tal-ceramics (Piddock and Qualtrough, 1990). However, dentists recognize th eir limited fracture re sistance, potential abrasivity, and variations in marginal inte grity (Anusavice, 1989). These concerns have led to the development of new dental ceram ic restorative materials and techniques. Despite the excellent translucency compared with traditional metal-ceramic systems, ceramic systems still have a limited long-term fracture resistance, especially when they are used in posterior areas or fo r fixed partial dentures (Sjgren et al., 1999). Compared with traditional metal-ceramic restorations, ceramic dental prostheses, including crowns and FPDs, are attractive to the dental commun ity because of their superior esthetics and biocompatibility. Fabrication of ceramic dental crowns is challenging because exceptional skills of a technician are required to provide mini mal stress concentration areas and accurate marginal fit (Kelly et al., 1996). In addition, ceramic crowns must be translucent and resistant to fracture even in clinical situ ations where inadequa te thickness precludes optimal design. Natural translucency is needed to achieve an appearance similar to that of human teeth. The core of ceramic prostheses can be fabricated from feldspathic porcelain,

PAGE 19

7 aluminous porcelain, lithia-disi licate-based ceramic, glass infiltrated magnesia aluminate spinel, glass-infiltrated alumina, and gla ss-infiltrated zirconia, (Campbell and Sozio, 1988). However, poor resistance to fracture ha s been a limiting factor in their use, especially for long-span or multiunit ceram ic FPD prostheses (Surez et al., 2004). Lithia-disilicate (Li2O2SiO2) based IPS Empress 2 (Ivoclar-Vivadent, Schaan, Liechtenstein) is one of the all-ceramic syst ems that was developed in response for the high demand for all-ceramic materials for FPDs. However, long term clinical studies with the Empress 2 system are required to determine whether they can serve as a feasible replacement for metal-ceramic systems. Sorensen et al. (1998) reported 3% fracture rate in a clinical study based on observa tion of 60 three-unit IPS Empress 2 FPDs over a 10 month period. Pospiech et al. (2000) reported one failure du e to fracture in a clinical study with 51 FPDs after one year of service. The aim of this study was to determine the clinical survivability of IPS Empress 2 crowns and FPDs in a two-year period using modified U.S. Public Health Service ev aluation criteria (Ryge et al., 1973, 1980). 2.1. Materials and Methods A total of 20 crowns and 20 FPDs placed in an experimental population of 15 patients (3 men and 12 women, ages 21-59) were analyzed in this study. Crown placements in 20 of the cases were indicated because of existing crowns associated with secondary caries, apical lesi ons, fracture or lack of esthetics. Indications for FPDs included replacement of an incisor or a first premolar tooth, or an inadequate existing anterior FPD. Three unit FPDs were fabr icated following a design that requires a minimum thickness of 3.5 mm buccolingually, oc clusogingivally and mesiodistally in the connector areas. The locations, survival ti me, and evaluation time (in months) of the FPDs as well as failure types are shown in Table 2-1. In addition Table 2-2 shows the

PAGE 20

8 distribution of the Empress 2 crowns accordi ng to their evaluation time (in months). Six of the prepared teeth had received end odontic treatment. For those that were endodontically treated, three were abutments fo r FPDs and three were restored with cast, post and core (cosmopost, Ivoclar AG, Liechtenstein). Table 2-1. Location, survival time (in months ) and failure type of Empress 2 FPDs. Location Survival Time (mo) Failure Type 31-33 9 Veneer fracture (chipping) 23-25 10 Connector fracture 21-23 11 Connector fracture 33-35 11 Connector fracture 21-23 13 Veneer fracture (chipping) 42-31 19 Connector fracture 11-22 20 Intact 23-25 20 Intact 21-23 20 Intact 21-23 21 Intact 12-11 22 Connector fracture 11-22 22 Connector fracture 13-15 22 Intact 41-43 22 Intact 23-25 23 Connector fracture 13-15 23 Connector fracture 12-21 24 Intact 13-15 24 Intact 33-35 25 Intact 41-32 27 Intact

PAGE 21

9 Table 2-2. Number of Empress 2 crowns according to their evaluation time The tooth preparations consisted of a shoulder finish line with rounded, smooth contours, avoiding sharp angles (Fig. 2-1) to obtain maximum fit of the finished protheses. To optimize the load carrying cap acity of the ceramic prostheses and to maximize esthetics, a shoulder width of ~1.5 mm was obtained. Occlus al reduction of the prepared tooth was ~2 mm. Fine diamond burs were used for final tooth contouring and finish line enhancement. The smoothness of the finish line and the ab ility to transfer the details to the refractory die ar e essential for the precision a nd the fit of the coping. To ensure high definition on impressions, re traction cord (Stay-put; Roeko, Langenau, Germany) was used. Polyvinylsiloxane impressi on material (Extrude, Kerr, Romulus, MI, USA) was used for complete arch record ings. Temporary crowns and FPDs were prepared to maintain gingival health and to maintain tooth position. All prostheses were prepared by the same certified dental techni cian using a layering technique. Margin integrity and periodontal health were record ed after cementation. Both gingival margins and the prostheses finish line were clinica lly excellent at the cementation appointment (Fig. 2-2). Number of Crowns Survival time (mo) Incisors Premolars Molars Total 19 2 2 20 1 1 1 3 21 2 1 22 1 2 3 23 2 1 3 24 6 2 8

PAGE 22

10 Figure 2-1. Facial view of prepared maxillary left central incisor, maxillary left canine, mandibular right central in cisor and mandibular left lateral incisor for a Empress 2 fixed partial denture. Figure 2-2. Finished prostheses that are placed on the prepared teeth as shown in Fig. 2.1. Labial view of two Empress 2 fixed partial dentures for a maxillary left central incisor, maxillary left lateral incisor, maxillary left canine, mandibular right central incisor, mandi bular left central incisor, and mandibular left lateral incisor region at baseline.

PAGE 23

11 After try-in, the internal su rface of the ceramic prosthes is was etched (5% HF, IPS ceramic etching gel, Vivadent, Schaan, Liechte nstein) for 60 s, rinsed, dried and silanated for 60 s (Monobond-S, Vivadent, Schaan, Liechtenstein). Prepared tooth surfaces were conditioned with 37% H3PO4 (Email Preparator GS; Vivade nt) for 30 s. Syntac dentin primer (Vivadent) and Syntac de ntin adhesives were applied to rinsed and partially dried dentin surfaces. Subsequently, Heliobond (Viv adent) bonding medium was brushed on the dentin surface and the inte rnal surface of the prosthesis Variolink II low viscosity (Vivadent) luting composite catalyst and base were selected and mi xed according to the color of dentin. Cementation was performe d immediately after coating the internal surface of the prosthesis with luting agen t. Excess cement was removed using a thin brush, explorer and dental floss respectivel y. Luting agent was polymerized from each surface using UV light for 60 s. Occlusion and articulation were controlled after cementation. Clinical procedures were pe rformed by the same clinician for all the restorations. Each restoration was evaluated two da ys after cementation (baseline), and thereafter one year and two years. Evaluati ons were performed by tw o clinicians using a mirror, explorer and intraoral photographs. Agreement between the two clinicians was 95%. United Public Dental Health (USPHS) crite ria were used to evaluate the quality of the restorations (Table 2-3). Disagr eement was resolved by consent.

PAGE 24

12 Table 2-3. Citeria for the direct evaluation of the restorations Kaplan-Meier statistics were used to anal yze the survival rates of the restorations (Kaplan and Meier, 1958). Alpha, Bravo and Ch arlie rankings were recorded and percent distributions were analyzed for each year. Category Score Criteria Anatomic form Alpha Restoration is cont inuous with tooth anatomy Bravo Slightly underor overcontoured restoration; marginal ridges slightly under-contoured; contac t slightly open (may be selfcorrecting); occlusal he ight reduced locally Charlie Restoration is undercontoure d, dentin or base exposed; contact is faulty, not self-corr ecting; occlusal height reduced; occlusion affected Marginal adaptation Alpha Restoration is continuous w ith existing anatomic form, explorer does not catch Bravo Explorer catches, no crevice is visible into which explorer will penetrate Charlie Crevice at margin, enamel exposed Color match and Surface texture Alpha Excellent color match and smooth surface Bravo Good color match and sligh tly rough or pitted surface Charlie Slight mismatch in color, sh ade or translucency and rough surface, cannot be refinished Caries Alpha No evidence of caries contiguous with the margin of the restoration Bravo Caries is evident contiguous with the margin of the prosthesis Post operative sensitivity Alpha No sensitivity Bravo Slight sensitivity

PAGE 25

13 2.2. Results Patients were evaluated at the recall a ppointments. Of the 20 Empress 2 FPDs evaluated, 50% were rated satisfactory and 100% of the 20 crowns were satisfactory. Distributions of the scores of the evaluated variables, color and surface, anatomic form, marginal integrity, and postoperative sensitivit y are presented in Tables 2-4, 2-5, and 2-6. Fractures in the connector area of the five FPDs (25%) were recorded at the one year recall exam (Fig. 2-3). In addition, five mo re fractures (25%) were observed in the remaining FPDs in the second year. Eight (4 0%) fractures occurre d in the connector areas. Additionally, two local ch ipping failures (10%) were observed in the FPDs. Even though the score for surface texture was 85% Alpha at the baseline exam, it decreased to 80% Alpha at the end of the firs t year and to 60% at the end of the second year (Table 24). A significant difference was not observed in the scores of anatomic form, caries, and sensitivity for the FPDs. All were given an Alpha rating. However, the score for marginal integrity significantly decreased from an 85% Alpha score to a 54% Alpha score (Table 2-4). Fractures were not observed in any of the Empress 2 crowns. Crown restorations were rated an 80% Alpha score for color and surface parameter at the baseline exam. This value decreased to 65% at the second year recalls (Table 2-5). Marginal adaptation received a 70% Alpha score at baseline and d ecreased to a 40% Alpha score by the end of first year and to a 25 % Alpha score by the e nd of second year (Table 2-5). The crowns were not associated with secondary caries during the two-year evaluation period. Even though tooth sensitivity was rate d a 95% Alpha score at the ba seline exam, it increased to a 100% Alpha score by the end of second year (Table 2-5). There was no significant

PAGE 26

14 difference between the baseline, first year a nd second year values of plaque index and gingival index of FPD and crow n restorations (Table 2-6). Figure 2-3. Fractured connector in anterior Empress 2 fixed partial denture after 11 months of clinical service. Table 2-4. FPD restorations: Sc ores of the clinical evaluation (%) at baseline, year one and year two (Surface textur e and color, anatomic form values were calculated for n=15 for the first year and for n=10 for the second year). Baseline Year 1 Recall Year 2 Recall Empress 2 FPDs (n=20) Alpha Bravo Alpha Bravo CharlieAlpha Bravo Charlie Surface texture and color 85 15 80 20 60 40 Anatomic form 75 25 67 33 70 30 Marginal adaptation 85 15 60 15 25 54 13 33 Caries 100 100 100 Post operative sensitivity 90 10 90 10 100

PAGE 27

15 Table 2-5. Crown restorations: Scores of the clinical evaluation (%) at baseline, year one and year two. Baseline Year 1 Recall Year 2 Recall Empress 2 crowns (n=20) Alpha Bravo Charlie Alpha Bravo CharlieAlpha Bravo Charlie Surface texture and color 80 20 80 20 65 35 Anatomic form 90 10 90 10 80 20 Marginal adaptation 70 20 10 40 40 20 25 55 20 Caries 100 100 100 Post operative sensitivity 95 5 95 5 100 Table 2-6. Scores of the clini cal evaluation for plaque and gi ngival index (%) at baseline, year one, and year two. Empress 2 FPDs Empress 2 Crowns Baseline Year 1 Recall Year 2 Recall Baseline Year 1 Recall Year 2 Recall Plaque index 0 100 67 80 100 70 60 1 33 20 30 40 Gingival index 0 90 85 85 95 80 75 1 10 15 15 5 10 25 Kaplan-Meier statistics revealed th at the survival rate for Empress 2 FPDs at 2 years was 50% (Fig. 2-4).

PAGE 28

16 zaman (ay)42 36 30 24 18 12 6 0Kmlatif Basari Yzdesi1,0 ,9 ,8 ,7 ,6 ,5 ,4 ,3 ,2 ,1 0,0 Survival Function Censored Figure 2-4. Kaplan-Meier survival statistics of Em press 2 fixed partial dentures (n=20). 2.3 Discussion Metal-ceramic prostheses are commonly used for FPDs (Surez et al., 2004). Previous studies revealed that they have a hi gh survival rate of 98% (Creugers et al., 1994), 90% (Scurria et al., 1998), and 85% (Walton, 2002) at 5, 10, 15 years respectively. Also, allceramic FPDs are becoming more popular and more desired due to their esthetic capacity. However, there are very few clinical studies with all-ceramic prostheses that evaluate their long term success (Surez et al., 2004). Recent clinic al studies reveal high failure rates of all-ceramic restorations compared with metal-ceramic restorations, especially when they are used in th e posterior region (Sorensen et al., 1998, Olsson et al., 2000). Empress 2 core ceramic is composed of crystalline and glass phases. The crystalline Empress 2 core consists of elongated lithia disilicate crystals (Li2Si2O5). The Time (months) Cumulative Survival SurvivalFunction

PAGE 29

17 lithia disilicate crystalline content in the hot-pressed core ceramic is 70 5 vol % (Della Bona, 2001). Empress 2 veneer also consists of a glass and one crystalline phase. The latter is reported to be fluor apatite and the crystal volume fraction is less than that of the Empress 2 core (Hland, 2000) (Appendix A). Xray diffraction analysis showed that the Empress 2 veneer ceramic is an amorphous glass-ceramic (Appendix A). Empress 2 core ceramic has a flexural strength of 215 MPa and fracture toughness of 3.4 (MPam1/2) (Della Bona, 2001). Fabrication of threeunit anterior FPDs us ing the Empress 2 system is recommended by the manufacturer because of high mechanical properties of Empress 2 core ceramic (Della Bona, 2001). However, when th e core layer is coated with the low-strength glass veneer, the resulting ceramic composite exhibits a significantly lower strength compared to the core ceramic. Dist ribution of stresses is affected by the elastic modulus differences between the glass veneer and ceramic core. These stresses in bilayer ceramic composite are different than those in the monolithic core material. In our specific case, the outer layer is a glass and will resu lt in lower failure stress in core/veneer composite than in the monolithic core ceramic. In the present study, a high fracture ra te (50%) was observed for veneered Empress 2 FPDs during the examination period. This outcome is related to the presence of a low toughness Empress 2 glass veneer (66 MPa) bonde d to the core ceramic. In a previous study, a 3% fracture ra te was reported in 41 Empress 2 FPDs after a 10-month clinical evaluation period (Sorensen et al., 1998). Also, Postpiech et al (2000) reported a 10% failure rate at the one year clinical examina tion. In addition, they reported no clinical failures in 76 crown restorations. The results of our study confirmed

PAGE 30

18 these results for crown restorations; however failure rates for FPD restorations were significantly highe r in our study. Even though special attention was paid during occlusal adjustment sessions to minimize the occlusal loads on FPD connectors, most of the fractures occurred at the connector areas. We conclude that connector design and thickness has a significant effect on the long term clinical survival of Empress 2 FPDs. This finding is in agreement with the results reported in a pr evious in vitro study (Oh et al., 2002). In our study we did not observe a signif icant difference between the base line, first year, and second year values of surface and color, plaque index and gingival index parameters. In a two-year clinical evaluation period, 50 % of the 20 Empress 2 FPDs had fractured. For these failed prostheses, 80% fractured at the conn ector regions and 20% chipped within the veneer layer. None of the crown restorations had fractured. The marginal integrity criteria of the restorations in the second year recalls were poorer than their baseline. Fracture surface analysis of the clinically fractured restorations is needed to understand the cause of the failures. The out come of this study combined with fracture surface analysis of the failed FPDs can provide critically important information that can lead to improvement in ceramic properties and increased survival probabilities for fullceramic prostheses. Thus, the next chapter wi ll present the fracture surface analysis of clinically failed fixed partial dentures.

PAGE 31

19 CHAPTER 3 QUANTITATIVE FRACTURE SURFACE ANALYSIS OF CLINICALLY FAILED CERAMIC FIXED PARTIAL DENTURES Bilayer ceramics used for fixed partial de ntures are becoming more widely used because of their aesthetic pr operties in the oral envir onment. However, the large statistical variation in the strength of th ese ceramics is associated with their low toughness and structural flaws. In addition, lo calized fracture of the veneering ceramic and connector fractures within the veneer laye r of a bilayer full ceramic FPD system have been reported in previous clinical st udies (Sorensen and Cruz, 1998; Pospiech et al. 2000 and Edelhoff et al., 2002). An increasing interest in ceramic fixed prostheses has followed improvements in strength, aesthetics and ease of processing. Su ch advances include in troduction of lithia disilicate (Li2O2SiO2) reinforced glass-ceramics for dental use. The moderately high strength and improved esthetics of these sy stems are well documented in the dental literature (Hland et al., 2000). However, the mechanisms for the failure of dental ceramics have not been studied extensivel y, although fractographic analysis is a key element in the design and development of dental structural materials (Kelly et al., 1989). Fractographic analysis of retr ieved clinical specimens ha s been crucial in efforts to investigate failure mechanisms, to identify fracture initiation sites, and to determine the probable cause of failure. Fr actography can provide information about the cause and source of failure. One can determine whether the failure is caused by a processing defect

PAGE 32

20 or by an overload condition. Quantitative fract ography or fracture surface analysis, is the application of fracture mechanics to charac teristic features on the fracture surface including the size of the fract ure-initiating flaw. However, although fractography is based on the science of fracture, pr oper interpretation depends on th e skill and knowledge of the examiner, particularly the ability to rec ognize fracture markings. This paper provides guidelines for observing the patterns of cracks and features on fracture surfaces of failed ceramic prostheses. The overall objective of th is study will be to identify the principal cause of failure in ceramic fixed partial de ntures (FPDs) using fr actographic techniques. 3.1 Materials and Methods FPDs that were produced from veneered and nonveneered lithia-disilicate-based glass-ceramic core ceramics were analyzed in this study. Veneered glass-ceramic FPDs were made from the Empress 2 ceramic system (Ivoclar Vivadent AG, Schaan, Liechtenstein). X-ray diffraction analysis re vealed that the veneering ceramic consists primarily of amorphous glass (Appendix A) All core FPDs were made using an experimental lithia-disilicate glass-ceramic th at has a greater crystal volume fraction than Empress 2 core ceramic (Appendix A). Clinically failed FPDs were retrieved from a clinical study. All of the FPDs were made by the same technician using the same fabrication procedures. Four-point bending specimens we re prepared from Empress 2 veneer, Empress 2 core ceramic, lithia-disilicate-based experimental core ceramic, and bilayer Empress 2 core-veneer ceramic. The final dimensions of beam specimens were 1.6 mm (height) x 4 mm (width) x 25 mm (length). The veneer thickness was 0.6 mm for each bilayer specimen. The core/veneer thickness ratio wa s 10/6. The span length/specimen thickness ratio was 15/1.6 to avoid large deflections and high shear stresses within the beam

PAGE 33

21 specimens. After polishing, dimensions of the specimens were measured with a micrometer. A total of 10 specimens were used for each group for fracture toughness calculations. Thus, 30 beam specimens were used in the study. Specimens were indented on the veneer surface with a Vickers indenter at a load of 4.9 N to produce controlled cracks. Specimens were stored in air for 24 h to ensure complete crack growth. They were then loaded to fracture at a crosshead speed of 0.5 mm/min. Beam specimens were loaded using a f our-point bending fixt ure and an Instron universal testing machine. All flexure experi ments were performed using the same fourpoint flexure fixture with an 18 mm outer span and a 6 mm i nner span. The veneer layer was placed in tension for bilayer flexur e test specimens. The mean strength ( f) of the laminated composites was calculated using co mposite beam theory (Beer and Johnston, 1981) (Appendix B). Fracture toughness of the beam specime ns was measured using quantitative fractography. Fracture in brittle materials generally occurs by the unstable propagation of a defect as a result of the combination of high stress and large flaws (Mecholsky, 2001). Flaws occur in dental ceramics as a result of processing or preparat ion. Fracture origins may be volume flaws such as internal cracks, pores, agglomerates, regions of inhomogeneous density or composition, or surf ace origins such as cracks caused by machining, surface pits or voids, and impact damage. Almost all of mechanically induced cr acks can be idealized as semi-elliptical, sharp cracks of depth, a, and half-width, 2b (Fig. 3-1) (Mecholsky, 1991). The crack sizes are approximated by an equivalent semicircular crack size, c [ c = (ab)1/2]. Fracture

PAGE 34

22 toughness, KC, is calculated using the stre ss at fracture, or strength, f, and the crack size, c : KC = Y f ( c1/2) (3.1) where KC is the critical stress intensity factor (fracture toughness), and Y is a geometric factor, which accounts for the sh ape of the fracture-initiati on crack and loading condition. The quantity Y depends on the ratio a/b The approximation [ c = (ab)1/2] allows many irregular crack shapes to be analyzed a nd avoids the complications of calculating a geometric factor for each cr ack (Mecholsky, 1991). For surf ace cracks that are small relative to the thickness of the sample, Y ~1.24. For sharp cracks that are induced by a Vickers or Knoop indentation, Y ~1.65, and for internal cracks, Y ~1.4 (Mecholsky, 1991). The fracture origins were determined by examining the fracture surface and tracing the fracture surface markings back to the initiation site. These markings include twist hackle (river marks), wake hackle (fracture tails), cleavage steps, Wallner lines, and branching locations. When characterizing fracture origins, photographs were made of the overall sample and of an enlargement of the fract ure origin region. The general fractographic procedure is outlined in ASTM standard C1322 (ASTM, 1999). Each specimen was studied using a stereo scopic microscope (Bauch & Lomb Inc., Rochester, NY, USA) at 160 X magnificati on. Crack initiating flaws (a, b) were measured to determine the fracture to ughness of each specimen. Scanning electron microscopic (JSM-6400, Jeol, Tokyo, Japan) examination was performed on selected specimens.

PAGE 35

23 Residual stress caused by the thermal e xpansion/contraction mismatch of the veneer ceramic and core ceramic was es timated using the following equation (Lawn, 1993): R= T/[(1+ C)/2 EC+(1-2 V) EV] (3-2) where R is the residual stress, is the difference between thermal expansion coefficients of the veneer and core ceramics, and T is the difference between the glass transition temperature of the veneer ceramic and room temperature. Subscripts C and V refer to core and veneer cer amics, respectively. The terms and E are PoissonÂ’s ratio and YoungÂ’s modulus, respectively. 3.2. Results We estimated from preliminary examination of the fracture surfaces, that eight (89%) of the nine connector failures and two (66%) of the three veneer ceramic failures were acceptable for fracture surface analys is. Primary fracture origins of the two discarded specimens were missing as a result of a chip fracture near the origin. All fracture surfaces exhibited multiple crack initia tion sites as a result of multidirectional and repeated loading. Arrest lines in the form of ridges were present on the fracture surfaces (Fig. 3-1). In addition, some curved markings resulted from the intersection of two crack fronts (Fig. 3-1). Twist hackle ma rkings were helpful in determining the primary fracture origin in the core ceramic (F ig. 3-2). Wake hackle markings from pores were selected as reference points for determ ining the location of fracture origins in the glass veneer (Fig 3-2). The wake hackle ma rkings were observed next to pores on the opposite side from the fracture origin in the vene er layer. Porosity ex ists within the glass veneer because of slurry preparation and sinterin g of veneer powders.

PAGE 36

24 Figure 3-1. All fracture surfaces exhibit multiple cr ack initiation sites as a result of multidirectional loading. Arrest lines in the form of ridges are present on the fracture surfaces. One of the arrest lines is a result of the intersection of two fracture paths. Figure 3-2.Wake hackle markings were us ed to establish the reference points for determining the fracture origin in the glass veneer. Wake hackle markings were also observed next to porositie s as an outcome of crack propagation through pores. The markings indicate the di rection of the fracture origin in the veneer layer. Black arrows indicate the direction of the fracture path. origin Occlusal Surface Twist hackle markings Wake hackle marking Origin

PAGE 37

25 Fracture, in seven of the eight (88%) conn ector failures examined, initiated within the surfaces of the pros theses. Six (75%) of the eight conne ctor failures originated in the occlusal surface of the FPD and one initiated from the gingival surface. Fracture origins of the bilaye r FPDs occurred within the veneer layer (Table 3-1). Two FPDs failed in their posterior abutment cr owns and one of them had an identifiable fracture origin in the margin area. The fracture initiation site in the latter prosthesis was located on the outer surface of the margin and fracture occured along the mesiodistal axis of the abutment. Two failures within the veneer layer failed from chipping and both originated at internal flaws. The fracture toughness of the lithium-disili cate-based glass-ceramic core and the glass veneer were determined from beam specimens using quantit ative fractography, i.e., equation (3-1). The mean fracture toughness of the core ceramic was 3.1 0.1 MPam1/2 and the mean fracture toughness of th e glass veneer was 0.7 0.1 MPam1/2. Fractographic analysis can only determine the stress that caused the fracture to occur from a crack or flaw of a particular size if the toughness or appa rent toughness is known. In the case of all core FPDs, the failure stress can be estimated directly from the crack size and the fracture toughness determined fr om the core ceramic bar specimens. The fracture toughness of the glass veneer was used to calculate the stress at failure (Eq. 3-1) for the FPD specimens in which fracture initia tion occurred within the veneer surface. However, for the veneered FPDs, we have shown that residual compressive stress is caused by an elastic thermal mismatch a nd viscoelastic (Appendix C) relaxation. Therefore, the calculated stress should in clude an additional term, which includes the compressive residual stress that has to be overcome before tensile stress can develop.

PAGE 38

26 Dilatometric analysis showed that there was a slight difference between the thermal expansion coefficients of the core and veneer. The values were 10.0 (ppm K-1) and 10.4 (ppm K-1) for the core and glass veneer, respect ively. The glass transition temperature (Tg) of the glass veneer is 540 C and the difference between the Tg and room temperature is 515 C. PoissonÂ’s ratio of the glass veneer, V is 0.23 and for the core ceramic, C, it is 0.24. In addition, the measured YoungÂ’ s modulus of the glass veneer, EV is 64 GPa and for the core ceramic, EC is 96 GPa. The estimated residual stress for veneered prostheses using the above parameters and equation ( 3.2) was 13 MPa (compressive) as shown in Table 3-1. This calculation assumes that ve neer/core materials are identical in their thermal history and geometry. The residual stress was estimated from laminated beams fabricated using firing schedules recommended by the manufacturer. This is the value listed in the Table 3-1 as R. The critical flaw sizes, estimated failure st resses, and fracture initiation sites for each ceramic FPD are listed in Table 3-1. The critical flaw size, c of the specimens for which an origin could be found, ranged from 240 to 939 m. Estimated failure stresses of the FPDs that had fracture initiation sites within the veneer layers, ranged from 19 to 68 MPa. Failure stresses of all-core FPDs that fractured in the connect or area ranged from 107 to 161 MPa.

PAGE 39

27 Table 3-1. Failure stress ( f), residual stress ( R), critical flaw size (c), semiminor axis(a), and semimajor axis (2b) of the FPD specimens. 3.3. Discussion Fracture surface analyses of failed ceramic FPDs showed that failure origins occured mostly at surface flaws except for the cases of chipping failures where fracture occurred within the veneer layer of the prostheses. Pr evious investigators re ported that fracture initiated typically along the veneer/core in terface of In-ceram alumina based ceramic crowns (Kelly et al 1989; Thompson et al 1994). However, this result was not observed Sample ID a (m) 2b (m) c (m) Failure Stress (MPa) Crack initiation site Material 1 376 1129460 116 Connector (gingival surface) Experimental core ceramic 2 94 1223240 161 Posterior crown margin Experimental core ceramic 3 361 918 407 124 Connector (occlusal surface) Experimental core ceramic 4 282 941 364 131 Connector (occlusal surface) Experimental core ceramic 5 188 706 258 156 Connector (occlusal surface) Experimental core ceramic 7 376 1543539 107 Connector (occlusal surface) Experimental core ceramic 6 283 1367440 27+13( R) Connector (occlusal surface) Experimental core ceramic / glass veneer 8 229 829 308 32+13( R) Connector (occlusal surface) Experimental core ceramic / glass veneer 9 791 2232939 19+13( R) Veneer chip-out Experimental core ceramic / glass veneer 10 283 1367440 68+13( R) Veneer chip-out Experimental core ceramic / glass veneer

PAGE 40

28 in our study. For the most part the fractures initiated within the o cclusal surface of the FPDs, most likely because of contact damage caused by the opposing teeth or premature occlusal contacts. Also, mechanical damage resulting from the occlusal adjustment by the dentist or the dental technici an can introduce flaws in the FPDs. We suspect that fracture initiation at the margin of an abutment of a specimen occurred because of a flaw that was introduced by the dentist in the margin area during the try-in procedure. Since most of the connector failures were associated with frac tures that initiated from occlusal surfaces, we suspect that these flaws were introduced as a result of contact damage. Evidence that supports this conclusion is that fractures were multidirectional (Fig. 3-1). During the mastication process the mandible makes lateral, centric and protrusive movements that allow the oppos ing cusp tip in the maxilla to exert multidirectional forces on the prosthesis. As a result, fracture can occur in the most vulnerable part of the ceramic FPD, i.e. the connector (Oh et al ., 2002). Fracture markings in specimen 3 (Table 3-1) indicated that there were two fracture origins on the occlusal surface of the FPD (Fig. 3-1). To de termine the primary fract ure origin that led to failure, we analyzed fracture markings that were farther away from both fracture origins. These markings indicate that the primary fracture origin is the one with the longest path (left side of the fracture surf ace shown in Fig. 3-1). The ridge between the two fracture origins represents the inters ection of two of the propagating cracks. In two of the three chipping failures, fr acture origins were visible. In the third case, fracture initiated from the internal surf ace of the veneer layer and propagated in two directions. The fracture origin was not present on the fracture surface.

PAGE 41

29 Fracture toughness of the glass veneer was us ed to calculate the stress at failure of specimens in which the fracture origin occurr ed within the veneer layer (Fig 3-3). The fracture initiating crack propa gates immediately at the fail ure stress. Even though the core layer is tougher than the veneer, once crack propagation begins in the veneer, the crack does not stop. Crack progression is not im peded by the core ceramic at the interface between the core and veneer. Thus, the toughness of the veneer is used to calculate the stress from the crack size (E q. 3-1). The bilayer materials also include a term for the compressive residual stress generated by the thermal expansion anisotropy (~ 13 MPa) and viscoelastic (Appendix C) processes that occur. Figure 3-3. Fracture propagated qu ickly in the glass veneer w ithout any increase in stress from the point where the semi-minor axis of the fracture origin ends. The calculated failure stresses in the veneer ed lithia-disilicate-based glass ceramic specimens are relatively low compar ed with those reported by Hland et al (2000). We reported that the increase in strength of bila yer core/veneer ceramics occurred because of Interface Ori g in

PAGE 42

30 a global compressive residual stress (Taskonak et al ., 2002) (Table 3-1). However, it is not only the global residual stre ss that plays a role in the failure mechanism of bilayer dental ceramics. Local residual tensile and compressive stresses adjacent to points of contact damage from previous loading and tensile stresses from flexural and/or subsequent contact loading al so can lead to additional stress or cause failure. The superposition of these stresses can cause lateral cracks to deve lop and/or propagate to the surface (Lawn, 1993). Even if the stresses are not sufficient to propagate median cracks, they might be sufficient to propagate late ral cracks and cause chipping of the glass veneer. This was most probably the case fo r the chipping failures. We conclude that fracture initiation sites of these glass-cer amic FPDs occurred primarily within the occlusal surfaces of the veneered units and the crack propagation patterns appear to be controlled by the loading orientation.

PAGE 43

31 CHAPTER 4 ROLE OF INVESTMENT INTERAC TION LAYER ON STRENGTH AND TOUGHNESS OF CERAMIC LAMINATES The demand for ceramic dental prostheses has increased with the introduction of pressable ceramics. These materials provide ex cellent esthetics in the oral environment and simplify the match in color and translucency of anterior fixed pa rtial restorations to that of adjacent tooth structure. However, fractures related to surface flaws and low flexural strength appear to be common problems in veneered pressable ceramics (Anusavice et al., 1989). Failures of dental ceramic st ructures are multi-factorial and may be associated with improper crown and bridge design, thermal incompatibility, stresses in layered structures, the presence of critical structural flaws, and improper processing techniques (Pospiech et al., 2000). Clinical failures reported for Empress 2 core/Empress 2 veneer ceramic specimens occurred by chipping of the veneer layer near the restoration surface (Pospiech et al., 2000; Edelhoff et al., 2002). Spalling, in which a crack grows beneath the surface before propagating to the surface, thereby form ing a chip, occurs in situations such as local contact or indentations (Marshall et al., 1982), delamination of the layered materials (Hutchinson et al., 1992), and during machining (Thoules, 1989). The Empress 2 core ceramic is composed of crystalline and glass phases. The crystalline Empress 2 glass-ceramic core consists of elongated lithia disilicate crystals (Li2Si2O5). The lithia disilicate crys talline content in the hot-pressed core ceramic is 70

PAGE 44

32 5 vol % (Della Bona, 2002). Empress 2 veneer also consists of a glass and a crystalline phase. The latter is reported to be fluorapatite where crystal volume fraction is less than that of the Empress 2 core (Hland et al., 2000). X-ray diffraction analysis showed that the Empress 2 veneer and Eris veneer are amor phous glass (Appendix A). The primary difference between Empress 2 core and the experimental co re is the crystal size. The experimental core has a smaller particle size than the Empress 2 core ceramic. The mean surface area of the part icles for Empress 2 core is 0.42 m2 compared with 0.20 m2 for crystals in the experimental core ceramic (Della Bona, 2002). Although the experimental core ceramic has smaller crystals and a lower volume fraction than the Empress 2 core ceramic, their physical properties are almost identical (Della Bona, 2002) (Appendix A). In any bilayer composite, such as the co re/veneer system, the interface between the two layers can control fracture behavior. If the toughness of the interface is greater than that of either of the components, the propagating crack would be expected to progress across the interface. Howe ver, if the interface is less tough than that of either of the phases, then the propagating crack would be expected to propaga te along or close to the interface (Thoules, 1989, Thompson, 2000). The loading a nd relative toughness of the interface and each phase control th e behavior and the direction of any crack that forms or propagates in the core ve neer system (Thoules, 1989). Hot pressing is the process used for the fabrication of the core portion of ceramic crowns and fixed partial dentures. The ceramic core structure is formed by the application of pressure to cause flow of viscous ceramic in an investment mold of the desired shape (Tooley, 1985). The shape in the investment mo ld is formed using the lost-wax process (Tooley, 1985). One disadvantage of this tech nique is the difficulty in removal of the

PAGE 45

33 investment material from the final cerami c shape. Grit blasting and acid etching are common methods used to remove the divesti ng material from the ceramic fixed partial denture. However, if dive sting is not completed prope rly, a residual investment interaction layer can remain on the ceramic surface. Potentially, the remaining layer could cause bonding problems between the veneer and ceramic core. The primary goal of this study was to an alyze the effect of the investment interaction layer on the flexural strength, fracture toughness, and fr acture path of four bilayer ceramic laminates. Two hypotheses were proposed: (1) The fracture path of bilayer specimens will be controlled by the presence of the interaction laye r at the interface. (2) The flexure strength of bilayer ceramic laminates will decrease in the presence of an investment interac tion layer at the interface. 4.1. Materials and methods Bilayer ceramic composites were fabric ated following accepted dental laboratory procedures. Self-cured acrylic resin (Pattern Resin, GC Corp., Tokyo, Japan) was used to prepare master models for rectangular Empress 2 core and experimental core layers. Impressions were made from the master m odel using a polyvinyl siloxane impression material (Extrude, Kerr Cor p., Romulus, MI, USA). Acrylic resin material was poured into the molds and cured to make rectangul ar bar patterns (1.7 mm x 4.0 mm x 25 mm). The dimensions of the rectangular bars were made uniform using a milling machine (PGF 100, Cendres & Metaux Sa., Bi el-Bienne, Switzerland).

PAGE 46

34 Following the preparation of the resin bars, four specimens of each bar was sprued and invested in each inves ting ring. Thus, either an Empress 2 core or experimental core ingot (Ivoc lar AG, Schaan, Liechtenstein) was used for every four specimens. A preheating furnace was used for the burnout procedure (Radiance, Jelrus Int., Hicksville, NY, USA). The following tw o-stage burnout sequence was used: (1) heat at 5 C/min to 250 C and hold for 30 min hold; and (2) heat at 5 C/min to 850 C and hold for 1 h. After the preheating stage, the invest ment cylinders were immediately transferred to the pressing furnace (EP500, Ivoclar AG, Schaan, Liechtenstein). The pressing temperatures for Empress 2 core and experimental core ceramics were 920 C and 910 C, respectively. Following the pressing procedure, the inve stment cylinders were removed from the pressing furnace and cooled for 2 h in a ve ntilated room. The cooled specimens were divested by grit blasting with 80 m glass beads (Williams glass beads, Ivoclar North America, Amherst, NY, USA) at an air pre ssure of 0.28 MPa. Before etching, the sprues were cut away and excess sprue segments were removed by grinding from the specimen surfaces using water as a coolant. Four core specimens were placed in one plastic bottle containing 20 mL of 1% HF solution (Invex Liquid, Ivoclar AG, Schaan, Liechtenstein) and these bottles were placed in an ultrasonic bath. After etching, the specimens were cleaned under running tap water for 10 s and then dried thoroughly. A metal rod, 3 cm in length, was attached to the tip of the grit blasting hose (1.6 mm in diameter) to assist in standardizing the distance from the tip to the specimen. Grit blasting was performed with 100 m Al2O3 particles (Blasting Compound, Williams-

PAGE 47

35 Ivoclar North America Inc) at an air pressure of 0.1 MPa. The timing of etching and grit blasting was controlled according to each procedure described below. Divested specimens were cleaned after grit blasting with a steam spray under pressure before veneering. The following conditions were designed to change the thickness of the interaction layer for each hot-pressed core ceramic group: Group 1: 1 % HF acid etch for 30 min and grit blast the surface to be veneered for 30 s Group 2: 1 % HF acid etch for 30 min and grit blast the surface to be veneered for 15 s Group 3: 1 % HF acid etch for 15 min and grit blast the surface to be veneered for 30 s Group 4: 1 % HF acid etch for 15 min and grit blast the surface to be veneered for 15 s Prepared core bars were veneered with Empress 2 veneer and eris veneer ceramics (Ivoclar AG, Schaan, Liechtenstein) according to the combinations described below. Veneer powders were incorporated with a mixing liquid (I voclar AG) to obtain a slurry solution, which was brushed onto the core ceramics that were prepared in slightly oversized silicone molds. The following core / veneer specimens were prepared: Combination A: Empress 2 pressed core ceramic, Empress 2 veneer Combination B: Empress 2 pressed core ceramic, Eris veneer Combination C: Experimental pressed core ceramic, Empress 2 veneer Combination D: Experimental pres sed core ceramic, Eris veneer

PAGE 48

36 Sixteen specimens were prepared for each of these 16 groups (four ceramic groups x four divesting conditions). The Empress 2 veneer-ceramic specimens were sintered in a furnace (P80, Ivoclar AG) with the firing cycle set for a 6-min climb to 800 C at 60 C/min, a 2-min hold, and cooling to 180 C. A vacuum was applied at 450 C and 759 C. Three layers of veneering ceramic, including the wash layer, were sintered on each core ceramic specimen. Furnaces were calibrated each day be fore the firing procedures. The Eris veneer-ceramic specimens were sintered in the same manner as described above except that the maximum firing temperature was 765 C. Following sintering, any excess veneer wa s ground away using a 75-grit diamond embedded disk, the veneer surface was seque ntially polished to a 2000-grit finish on a metallographic polisher (Mode l 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while exposed to a continuous flow of tap water. Edges of each specimen were beveled to minimize edge failure during flexure testing. The final dimensions of the bimaterial bar specimens were 1.7 mm (height) x 4.0 mm (width) x 25.0 mm (length). The veneer thickness was 0.6 mm for each specimen. The veneer/core thickness ratio was 6/11. This ratio was chosen after composite beam analysis revealed that the tensile stress at the interface would be maximized (Beer and Johnston, 1981) (Appendix B). If the interf ace was at the neutral axis, no stress would develop. Thus, the thickness of the veneer-c ore specimens were ad justed to shift the neutral axis away from the core and to ensure as large a tensile st ress as possible at the interface. The ratio of test span length /thickness was chosen as 15mm/1.7mm to avoid large deflections and large shear stresse s during loading. After final polishing, the

PAGE 49

37 dimensions of the specimens were measured with a digital micrometer. All specimens were then stored in distilled water for 72 h before testing. The flexural strength of the ceramic lami nates was determined using a four-point flexure test fixture. The fle xure fixture was enclosed in a testing chamber and specimens were tested in circulating water at 37 C to approximate an oral environment. The flexure tests were performed using a universal mech anical testing machine (Model 4465, Instron Corp., Canton, MA) with four-poi nt flexure fixture (15 mm outer span, 5 mm inner span) at a crosshead speed of 0.5 mm/min. For all flex ure tests, the top fixture was held in place using a three-post alignment a pparatus and the bottom fixt ure was placed on a stainless steel ball to provide a fully articulating c onfiguration. The veneer side was placed in tension for all flexural test specimens, because most of the observed failures in previous clinical studies and in our fr actographic analysis of clinically failed FPDs in chapter 3 occurred in the veneer layer (Pospiech et al., 2000 and Edelhoff et al., 2002 ). The maximum tensile stress occurs in the outer ve neer surface within th e inner loading span. The strength, f, of bilayer composites was calculated using composite beam theory (Beer and Johnston, 1981) (Appendix B). In the past, these types of analyses have been used to determine the mechanical behavi or of several dental laminates (DeHoff et al., 1989; Thompson, 2000). Weibull analysis was performed to evaluate the structural integrity of the bilayer ceramic specimens. Characteristic strength values were also calculated. To determine the fracture origins of the specimens, the fracture surfaces of the specimens were coated with gold-palladium using a sputter-coating machine (Technics Inc., Alexandria, VA, USA). Each specimen wa s studied with a stereomicroscope (Bauch

PAGE 50

38 & Lomb Inc., Rochester, NY, USA) at 160 X magnification. The length of crackinitiating flaws and the calculated strengths we re used to determine the apparent fracture toughness (Kc) of each specimen (Eq. 4-1). Scanning electron microscopic (JSM-6400, Jeol, Tokyo, Japan) examination was perfor med on selected specimens and SEM images were recorded from repres entative fracture surfaces. A pparent fracture toughness, KC, was calculated using fractographic analys is and the fracture mechanics equation: KC = Y ( ) f ( c )1/2 (4-1) where Y( ) is a geometric factor that has a valu e of 1.24 for an equivalent semicircular flaw, c of a semi-elliptical flaw of depth, a, and half width, b, (Mecholsky, 1994), c = (ab)1/2, (assuming no local residual stress), and f is the calculated flexural strength. The crack size is equivalent to that obtained using the length of the semiminor axis of the flaw with the appropriate Y( ) factor for the appropriate se mi-elliptical crack geometry (Mecholsky, 1991). Two specimens were randomly chosen from each divesting group to analyze the amount of residual investment interaction laye r that was left on the core surface after divesting (Fig. 4-1). Quantitative EDS anal ysis of the surfaces was made at 80 X magnification using a scanning electron microsc ope. The percentage of each element that was present on the scanned area was recorded after analyzing the surface. No difference was found in elemental composition between th e investment interaction layer and core layers for each group. Two-way analysis of variance was used to determine whether the differences between group means for flexural strengt h, crack size, and fracture toughness of

PAGE 51

39 specimens were statistically significan t. Linear regression analysis of versus c-1/2 was performed to determine the relationship between critical flaw size a nd flexural strength (Mecholsky, 1994; Mecholsky, 1991), i.e. f = ( a+ R) = [Kc / Y( )] c-1/2 (4-2) where a is the applied stress to failure and R is any global residual stress that may be present. If there is no residual stress, then R = 0 and f = a. Figure 4-1. Cross section of Empress 2 core ceramic with residual investment interaction layer on its outer surface. Representative sample was divested using 15 min etching and 15 s grit blas ting surface treatment. (A) Residual investment interaction layer (B) Empress 2, glass-ceramic core layer. 4.2. Results The mean flexure strength and standard deviation of each group are summarized in Table 4-1. Based on two-way ANOVA a nd DuncanÂ’s multiple range test, the differences in mean strength values of gr oups 1 and 2 were statistically different (p 0.05) and both group means are significantly diff erent from the values for groups 3 and 4. A B

PAGE 52

40 No significant difference wa s found between the means for the flexure strengths of subgroups 3 and 4 (p > 0.05). Combinations A, B, C and D represent different ceramic laminates. Also, there were no significant diffe rences between mean fl exural strengths of groups 1, 2, 3, and 4 within combinations A, B, C and D, which represent different surface (divesting) treatments (p > 0.05). Weibull modulus and characteristic strength values for each group are summarized in Table 4-1. The Weibull modulus gives an indication of the variability of the flexural strength, with gr eater values indicating a narro wer distribution of flexural strength. Without exception, failure origins occurre d within the veneer surface of each specimen. The critical cracks, within the tensile region of the laminate beams, were approximately semi-elliptical in shape (Fig. 42). Independent of the crack origin, all of the specimens showed a crack path approxi mately perpendicular to, and through the laminate interface without delamination. Surf ace porosity in the veneer was a common flaw that initiated crack propagation. Wake hackle markings, informally known as “fracture tails”, were observe d at pores on the fracture surfaces (Fig. 4-2A). These markings were helpful in determining the crack direction and critical flaw site. Two-way ANOVA showed that the mean flaw size dimensions ( c ) of groups C and B were significantly different from thos e of combinations A and D (p 0.05) (Table 4-1). There was no significant difference between the mean flaw size dimensions ( c ) of groups 1, 2, 3 and 4 for ceramic combinations A, B, C and D, respectively (Table 4-1). Ordinarily the largest flaws are expected to correspond to the lowest strength valu es. This clearly was not the case among the groups (Table 4-1).

PAGE 53

41 Table 4-1. Mean flexural strength ( ), standard deviation (SD) number of specimens per group (n), characteristic strength ( 0), critical Flaw size (c) and Weibull modulus (m) values of b ilayer ceramic laminates. Divesting Combination (S.D.) (MPa) n m 0 (MPa) (c) S.D. ( 10-5 m) 94 (17) 16 6 92 12 3 84 (20) 16 4 98 15 6 84 (18) 17 5 96 14 5 Group 1 Etch 30 min, grit blast 30 s Combination A Empress 2 core / Empress 2 veneer 83 (16) 17 5 94 15 5 123 (18) 17 7 125 9 3 126 (15) 16 8 104 8 2 133 (15) 17 9 134 7 2 Group 2 Etch 30 min, grit blast 15 s Combination B Empress 2 core / Eris veneer 122 (20) 17 6 119 9 3 118 (27) 17 4 104 10 7 112 (21) 17 4 105 12 7 106(24) 16 5 103 12 4 Group 3 Etch 15 min, grit blast 30 s Combination C Experimental core / Empress 2 veneer 106 (23) 17 9 102 10 2 94 (26) 16 3 110 17 12 112 (22) 17 4 115 12 6 111(20) 17 5 113 12 5 Group 4 Etch 15 min, grit blast 15 s Combination D Experimental core / Eris veneer 105 (11) 17 4 110 14 8

PAGE 54

42 Figure 4-2. Fracture surface of cer amic laminate (SEM view). (A) The fracture origin (arrow), and wake hackle markings [12] in the veneer ceramic confirm that failure initiated from the tensile su rface in the lower middle portion of the micrograph as marked by the larger arrow. (B) Fracture surface of the specimen in Figure 2A, at higher magnifi cation. The crack initiation site is a pore defect. (C) Fracture surface of a la minate specimen. The core ceramic is shown by the lighter area (top), and the veneer is the darker area (bottom). 4.3. Discussion For each core-veneer combination and surface treatment group, there is an expected correlation (Figure 4-3) between the flaw size and strength, i.e., high strength, corresponds to a small flaw size. However, ther e is a lack of correlation between critical flaw sizes and strength, i.e., some of the combinations that have large values of strength also have large values of flaw sizes. Generally, it is expected that high strength Wake hackle marking

PAGE 55

43 corresponds to small flaw sizes. This lack of correlation for the bilayer groups in Table 41 suggests that a more complex mechanism c ontrols the fracture process. To explain these results, the graphs of flexural strength versus c-1/2 values were made for each group. A representative graph, shown in Fig. 4-3, demonstrates the expected relationship between and c determined from Equation 4-2. Notice that the graph in Figure 4-3 does not intersect at =0 but rather at 27.5 MPa. This implies that there is a global compressive residual stress in the bilayer specim ens. The slope of this graph is equal to Kc/Y. Using Y=1.24 for equivalent semicircular cracks (Mecholsky, 1994), the calculated Kc value was 0.7 MPa m1/2 (Figure 4-3). This calculat ion was made for all bilayer specimens and results lead to the same conc lusion. The similarity in toughness values for different core-veneer combinations occurs be cause all failures originated in the veneer and the veneer toughness controlled these failures. The toughness of monolithic veneer specimens using fractographic analysis was 0.7 0.1 MPa m1/2 (Figure 4-4). This value is in agreement with the above calculations. Since there is no difference among the mean values for groups 1 to 4 for each divesting treatment, it is concluded that the residual interaction la yer (Fig. 4-1) did not significantly affect the flexural strength, Weibull moduli, characteristic strength or fracture toughness of the bilayer ceramic lami nates. Thus, both hypotheses are rejected. However, different bilayer ceramic combinati ons had significantly different values of mean flexural strength, characte ristic strength, and apparent fracture toughness. Thus, the material properties appear to be more impor tant factor in the m echanical behavior of bilayer laminates than the resi dual investment interaction la yer as long as the interaction layer results in a well-bonded interface. There was also no significant difference in failure

PAGE 56

44 modes and crack sizes within these groups and combinations. Thus, the differences observed in strength for the core/veneer comb ination groups are like ly associated with global residual stress caused by the thermoelastic and viscoelastic (Appendix C) behavior of the bilayer specimens. Test methods that i nvolve application of lo ads directly to the interface should be used to understand precise ly the interfacial strength of bilayer specimens with the investment interaction layer. Group A-40 20 40 60 80 100 120 140 020406080100120140160c-1/2 (m-1/2)Flexural Strength (MPa) Figure 4-3. Representative graph (Group A4) of flexural strength ( f) versus c-1/2. Graphs for each group showed that the da ta lie on a best-fit linear line (R2 = 0.89). Extrapolation of the best-fit lin ear line indicates a residual stress of 27.5 MPa, i.e., the intercept with the ordinate (slope = 0.6 MPa m1/2). For any bilayer composite, the combinati on of loading and toughness of interfaces will determine if failure occurs because of delamination of the interface or flexural/tensile

PAGE 57

45 failure of the composite (Thoules, 1989 and Thompson, 2000). In the ab sence of interface fracture, toughened multiphase and laminated cer amics are susceptible to brittle failure in a manner similar to fine grained, homogenous ceramics (Kerans et al., 1989 and Prakash et al., 1995). 0 10 20 30 40 50 60 020406080100120Flexural Strength (MPa) Empress 2 veneer monolithic specimensc-1/2 (m-1/2) Figure 4-4 Representative graph of flexural strength ( f) versus c-1/2 for Empress 2 veneer monolithic specimens. Graphs for each group showed that the data lie on a best-fit linear line (R2 = 0.89). Extrapolation of the best-fit linear line indicates a residual stress of 3.6 MPa, i.e., the inter cept on the ordinate (slope = 0.59 MPa m1/2). The Weibull moduli values varied from group to group. These values reflected the scatter in flexural strength for the specim ens in each group. As expected, the flexural

PAGE 58

46 strength generally varied inversely with the s quare root of the flaw dimensions for each group (Fig. 4-3 and Fig 4-4). A majority of the critical fl aws in the specimens initiated from surface porosity in the veneer ceramic. Porosity may result from air entrapment during mixing or condensation, the use of a na rrow particle size distribution, and gases produced during firing (McLean and Hughes, 1965). Specific emphasis should be placed on crack -initiating flaws and crack origins of the specimens retrieved from clinical studies. Since a majority of such failures initiated from surface porosity at the tensile surface, chip ping failures that occur in clinical studies may be a result of volume flaws at the surface of the veneer layer. The driving force for subsequent crack growth is most likely th e residual stresses generated from thermal expansion mismatches and viscoelastic re laxation processes (Appendix C) in the core/veneer laminates (Scherer, 1986). Elastic-viscoelastic relaxati on behavior of a glass veneer layer affects this residual stress. When a glass veneer/ceramic core bilaye r composite is heat treated at temperatures near the glass transition temperature, the density of the glass veneer will increase because of structural relaxation whereas little or no change takes place in the ceramic core. As the glass reaches its equilibrium structure, ther e is a rapid increase in the residual stress initially; then there is a gradual increase unt il the residual stress equilibrates at a greater stress level (Scherer, 1986; Anokye, 1989). All fracture origins in the fou r-point flexural test specime ns occurred at the tensile surface and there was no interface delamination in any of the specimens. The Empress 2 core and Eris veneer combination showed a global residual stress of 28 MPa and the greatest flexural strength. The Empress 2 core / Empress 2 veneer specimen group

PAGE 59

47 exhibited the lowest mean fracture toughness and flexural strength. The mean flexure strength of specimens between material comb inations varied from group to group (Table 4-1). Linear regression analysis of flexural strength versus negative square root of flaw dimension data is consistent with a linear relationship expected from linear elastic fracture mechanics (Mecholsky, 1994 and Mecholsky, 1991). The hypothesis that different divesting met hods influence the flexural strength of the specimens was rejected. The hypothesis that the interaction layer controlled the fracture path was also rejected. Since the investment interactio n layer results in a coherent, well-bonded interface, then there is no effect on strength or toughness of the bilayer laminates. Different combinations of core and veneer ceramics had significantly different apparent fracture toughness values, flexural streng th, and Weibull moduli. The latter result implies that thermoelastic and viscoelastic (Appendix C) properties of the ceramic-veneer combinations may control the strength. The flexural strength and fracture toughness of layered ceramics loaded in flexure (with strong bonding at the inte rface) are governed by the flaw characteristics of the material in tension and residual stress.

PAGE 60

48 CHAPTER 5 RESIDUAL STRESSES IN BILAYER DENTAL CERAMICS An increasing interest by dentists and pa tients in ceramic fixed prostheses has led to improvements in strength, toughness, aesthe tics and ease of processing. However, the large statistical variation in the strength of these ceramics is associated with their low toughness and structural flaws (cracks and poros ity). In addition, chipping fracture, i.e. spallation from the veneer layer of bilayer ceramic prostheses has been reported in clinical studies (Pospiech et al, 2000; Edelhoff et al., 2002). Empress 2 (E2C), experimental (EXC) core glass-ceramics, Empress 2 veneer (E2V) and Eris veneer (ERV) are the lithia-d isilicate-based glass ceramic co re and silicate amorphous glass veneer ceramics (Appendix A), respectivel y, selected for this study. The increased strength and improved esthetics of these syst ems have received much attention in the dental literature (Hland et al, 2000). Bilayer ceramics that are widely used fo r dental crown and bridge prostheses can sustain relatively high global compressive resi dual stresses within the veneer surface (DeHoff and Anusavice, 1989). Compressive st resses can result from thermal expansion coefficient differences between the core and ve neer ceramics or from differences in the elastic-viscoelastic be havior between the two layers (S cherer, 1986). Veneered ceramics with such residual stresses are susceptible to chipping, i.e. spallation (Marshall and Lawn, 1980). During mastication, local stresses are applied to the occlusal surfaces of

PAGE 61

49 prostheses. With the superposition of globa l compressive residual stresses on these local stresses, a tensile stress region can develop be low the surface of the veneer layer. Under these stress states lateral cracks can form a nd/or propagate toward the surface, potentially resulting in chipping of the veneer ing ceramic (Hutchinson and Suo, 1992). This paper presents a four-step fracture mechanics approach to determine the existence and cause of any global residual stress in bi layer dental ceramics. 5.1. Materials and methods 5.1.1. Specimen Preparation Bilayer ceramic composites were fabricated following accepted dental laboratory procedures. Self-cured acrylic resin (Pattern Resin GC Corp., Tokyo, Japan) was used to prepare E2C and EXC bar specimens. Four bar specimens per mold were sprued and invested with IPS Empress 2 special investment material (Ivoclar Vivadent AG, Schaan, Liechtenstein). The invested bar patterns were eliminated using a two-stage burnout for 60 min at 850C in a preheating furnace (Radiance, Jelrus Int., Hick sville, NY, USA) and the mold cavity was filled with core ceramic by hot pressing e ither E2C or EXC ingots using a pressing furnace (EP500, Ivoclar Vivadent AG, Sc haan, Liechtenstein) (Appendix A). The pressing temperatures for E2 C and EXC ceramics were 920 C and 910 C, respectively. The cooled specimens were divest ed by particle abrasion with 80 m glass beads (Williams glass beads, Ivoclar Vivadent North America, Amherst, NY) at a pressure of 0.28 MPa. The remaining investment was clea ned by placing each bar in 1% HF (Invex Liquid, Ivoclar Vivadent AG, Schaan, Liechtenstein) for 30 min.

PAGE 62

50 Each specimen was abraded using 100 m Al2O3 particles (Blasting Compound, Williams Ivoclar North America Inc., Amherst, NY, USA) at an air pressure of 0.1 MPa for 30 s. Divested specimens were cleaned with a steam cleaner before veneering. Prepared core ceramic bars were veneered with Empress 2 veneer (E2V) and Eris veneer (ERV) ceramics (Ivoclar Vivadent AG, Schaan, Liech tenstein) (Appendix A). Veneer ceramic powders were incorporated with mixi ng liquid (Ivoclar Vivadent AG, Schaan, Liechtenstein) to obtain a slur ry solution, which was brushed onto the core ceramics that had been placed in slightly oversized silicone molds. The E2V ceramic was sintered in a furn ace (P 80, Ivoclar Vivadent AG, Schaan, Liechtenstein) according to a firing cycl e consisting of a 6 min rise to 800 C at 60 C/min, a 2 min hold, and cooling to 180 C in 45 s. A vacuum was applied at 450 C and released at 799 C. Three layers of veneering ceram ic, including the wash layer, were sintered on each core ceramic specimen. The furnaces were calibrated each day before the firing procedures. Following sintering, the excess veneeri ng ceramic was ground with a 75 grit diamond embedded disk and sequentially polished to a 2000 grit finish on a metallographic polisher (Mode l 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while exposed to a continuous flow of tap water. The following core/veneer specimens were prepared: Combination A: E2C / E2V Combination B: E2C / ERV Combination C: EXC / E2V

PAGE 63

51 Combination D: EXC / ERV Six specimens were prepared for each of th ese four groups. Monolithic veneer specimens from E2V and ERV powders were prepared in addition to the bilayer specimens. The final dimensions of the bimaterial bar specimens were 1.7 mm (height) x 4.0 mm (width) x 25.0 mm (length). The veneer thickness was 0.6 mm for each specimen. The veneer/core thickness ratio was 6/11. Th e selected test span length/specimen thickness ratio was 15/1.7 to a void large deflections of the beam. After final polishing, dimensions of the specimens were measured with a micrometer. X-ray diffraction analysis was performed to determine the crystal phases in the sintered veneering ceramics. Thermal expans ion coefficients of the glass veneers and core ceramics were measured using a dila tometer. The maximum residual stress caused by the thermal expansion coefficient mismat ch between the veneer ceramic and core ceramic was estimated using the following equation (Lawn, 1993): R= T/[(1+ C)/2 EC+(1-2 V) Ev] (5-1) where R is the residual stress, is the difference between linear thermal expansion coefficients of the veneer and core ceramics ( VC), T is the difference between the glass transition temperature of the ven eer ceramic and room temperature, and and E are PoissonsÂ’s ratio and YoungÂ’s modulus, respectively. S ubscripts C and V refer to core and veneer ceramics, respectively. 5.1.2 Mechanical testing methods A four-step fracture mechanics approach wa s used to determine residual stress in the bilayer dental ceramics. This technique included:

PAGE 64

52 (1) Indentation of the specimens to measure and compare indentation induced crack sizes. (2) Flexural strength determination for monolithic and bilayer specimens. (3) Calculation of apparent fracture t oughness for bilayer specimens and fracture toughness determination for monolithic specimens. (4) Calculation of residual stress in b ilayer specimens using a fracture mechanics equation. Indentation cracks were induced within th e veneer surface of all specimens using a Vickers indenter at a load of 4.9 N. Indentation induced longitudinal and transverse cracks were measured optically using a reticulated eyepiece (Fig. 5-1). Figure 5-1. Schematic illustration of an indented beam specimen placed in four-point flexure. Indent cracks were induced in the tension surface. The mean flexural strength of each of the four types of ceramic composites was determined using a four-point flexure test fixture with a 15 mm outer span and a 5 mm inner span at a crosshead speed of 0.5 mm/min using a universal testing machine (Model

PAGE 65

53 4465, Instron Corp., Canton, MA). The veneer surface was placed in tension for all flexure test specimens. The strength ( f) of the laminated composites wa s calculated using mechanics and composite beam theory (Beer and Johns ton, 1981) (Appendix B). The strength ( f) of the monolithic specimens was calculated using simple beam theory. In previous studies, composite beam theory (DeHoff and A nusavice, 1989; Thompson, 2000) and finite element stress analysis have been applied to analyze failure stresses in ceramic/ceramic and metal/ceramic composites (Scherrer, 1986). Each fractured specimen was analyzed w ith a stereomicroscope (Bauch & Lomb Inc., Rochester, NY, USA) at 160 X magnifica tion. Crack initiating flaws were measured to determine the fracture toughness of each specimen. Scanning electron microscopic (JSM-6400, Jeol, Tokyo, Japan) examinati on was performed on selected specimens. Fracture toughness, KC, was then calculated using the fracture mechanics equation: KC= Y( ) f ( c )1/2 (5-2) where Y( ) is a geometric factor that has a valu e of 1.65 for an indent ation induced flaw with local residual stress from indentati on (Mecholsky, 1991) and 1.24 for flaws without local residual stress and f is the calculated flexural stre ngth. For equivalent semicircular flaws of depth “a” and half width “b” (Mecholsky, 1994), c is the crack size [c= (ab)] (Mecholsky, 1991). Residual st ress is calculated from th e following equation (Conway and Mecholsky, 1989):

PAGE 66

54 r = [ Y (2/ 1/2) ac1/2 KC] / [ Y (2/ 1/2)c1/2] (5-3) where r is the residual stress, a is the applied stress at failure, and KC is the fracture toughness of the glass veneers and the other terms are as listed above (Conway and Mecholsky, 1989). One-way analysis of variance was performe d to determine whether the differences between group means for fle xural strength and fracture t oughness of specimens were statistically significant. The same statistical analysis was used to determine whether the differences between the mean longitudinal and transverse indentation induced crack sizes were statistically significant. 5.2. Results The mean longitudinal and transverse inde ntation-induced crack sizes and standard deviation of each group are summarized in Table 5-1. Based on one-way ANOVA and DuncanÂ’s multiple range tests, the differen ces between the mean indentation induced longitudinal and transverse crack sizes of b ilayer specimens were significantly different (p 0.05). However, no significant differen ce was found between the means for the longitudinal and transverse crack sizes of monolithic glass veneer specimens (p > 0.05). One-way ANOVA showed that there was a statistically sign ificant difference between the mean flexure strengths of monolithic and bilayer specimens (p 0.05), but there was no significant difference (Table 51) between the mean flexural strength of bilayer specimen groups (p > 0.05). Without exception, failure origins were locat ed within the tensi on surface of each specimen and not at the interface between the ve neer and core. For most of the composite specimens, the indentation cracks were th e critical flaws that controlled crack

PAGE 67

55 propagation (Fig 5-2). However, some speci mens failed from porosity at the surface. These critical flaws, within the tensile regi on of the composite bars were approximately semi-elliptical in shape (Fig. 5-3). Table 5-1. Mean flexural strength ( ), standard deviation (SD), residual stress ( R), longitudinal and transverse indentatio n-induced crack size and apparent fracture toughness (Kc) of bilayer and monolithic specimens. Negative values of R indicate compressive stress in the veneer. Groups S.D. (MPa) Longitudinal indent crack (m) Transverse indent crack (m) R (MPa) Kc S.D. (MPam1/2) Eris Veneer 48 7 62 11 59 11 0 0.7 0.1 Empress 2 Veneer 43 8 61 15 58 14 0 0.7 0.1 Experimental Core /Eris Veneer 66 11 55 14 44 14 -23 0.9 0.1 Experimental Core / Empress 2 Veneer 73 15 50 18 34 12 -25 1.0 0.2 Empress 2 Core / Eris Veneer 61 12 63 11 46 12 -22 1.0 0.2 Empress 2 Core / Empress 2 Veneer 72 13 58 12 46 9 -22 1.0 0.2

PAGE 68

56 Even though all specimens failed from the veneer surfaces, there was a statistically significant difference between the calculated apparent fracture toughness values (Eq. 5-2) of bilayer specimen s and monolithic veneer specimens (p 0.05). The fracture toughness of monolithic ERV and E2V (veneers) was identical (0.7 MPam1/2) (Table 5-1). Figure 5-2. Fracture surface and fracture origin (arro ws) of a monolithic veneer specimen that reveal an indent cr ack as a fracture origin. Xray diffraction analysis was performed to identify crystal phases in the veneer layers that may increase fracture toughness. Howe ver, the veneer layers were found to be essentially amorphous because there were no distinct peaks observed in the X-ray diffraction graph between the a ngles of 5 and 110 degrees.

PAGE 69

57 Residual stress values of bilayer specimens (Eq. 5-3) ranged between -40 MPa and -60 MPa (compressive) (Table 5-1). The highest compressive residual stresses were observed in bilayer specimens with Empress 2 veneer layers. No residual stress was detected in the monolithic specimens (Table 5-1). Figure 5-3. SEM image of a fracture surface of an E2V specimen that exhibits a large pore (arrows) as the fracture origin. The bottom figure is an enlarged view of the top figure. Dilatometric analysis showed that there was a slight difference between the thermal expansion coefficients of the co re and veneer. The values were 9.8 (ppm K-1)and 10.2 (ppm K-1) for the core and veneer ceramic, respectively. The glass transition temperature (Tg) of the veneer ceramic was determined to be 540 C and the difference

PAGE 70

58 between Tg and room temperature is 515 C. The PoissonÂ’s ratios of the veneer ceramic ( V) and the core ceramic ( C) are 0.23 and 0.24, respectively. The YoungÂ’s moduli of the veneer ceramic ( EV) and core ceramic ( EC) are 64 GPa and 96 GPa, respectively. The calculated residual stress for the bilayer sp ecimens using the parameters above and equation (5-1) is 13 MPa (compression). 5.3. Discussion The increase in strength and toughness for th e bilayer composites compared with the glass veneer can be explained by seve ral possible phenomena. These include crystallization of the veneer layer, th e increased toughness of the ceramic core, compressive residual stress associated with thermal expansion anisotropy, and compressive residual stress caused by viscoelast ic structural relaxa tion (Appendix C). X-ray diffraction analysis revealed no evidence of a crystal phase, indicating that the veneering ceramics consist predominantly of an amorphous glass phase. Although there are reports that fluorapa tite crystals exist in the veneering ceramics (Hland, 2000), the volume fraction is evidently lower than detection limits. Thus, the strengthening mechanism for the bilayer composites is not caused by crystallizat ion of the veneer layers. Observation of the fracture surfaces (Figs. 5-2 and 5-3) showed that all fractures occurred within the veneer layer at the su rface. Crack propagation continued through the veneer/core interface. Thus, the increased toughness of the core ceramic compared with the veneer ceramic did not affect crack initiation or propagation. Bilayer specimens showed longer longitudinal cracks th an transverse cracks (Table 5-1 and Fig. 5-1). However, there was no statistically significant difference

PAGE 71

59 between the mean indentation-induced transv erse and longitudinal cracks of monolithic glass veneer layers (p > 0.05). This differe nce in mean crack lengths was the first indication of residual stress in the bilayer specimens. There are two possibilities for the difference in the length of indent-induced cracks in the veneer layer. The longer longitudinal cracks could be a result of resi dual tensile stress perpendicular to their direction or the shorter cracks may be a result of the compressive residual stress perpendicular to their short cracks. In Tabl e 5-1, we see that the difference in mean length of longitudinal cracks in the monolithic veneer and bilayer specimens is not statistically significant (p > 0.05), whereas, the mean transverse cracks in the bilayer specimens are significantly shorter than the ones in the monolithic veneer specimens (p 0.05). Thus, we conclude that compressive re sidual stress exists in the longitudinal direction within the veneer layer of the bilayer specimens. Although, fractographic analyses of the fract ure surfaces showed that the fracture origins of all specimens occurred within th e veneer surface, the flexural strength and apparent fracture toughness of the bilayer specimens were sign ificantly greater than those of the monolithic specimens (p 0.05). These effects represent further evidence of compressive residual stress in the ve neer layers of bilayer ceramics. The residual stress calculated using e quation (5-2) further validates the presence of compressive residual stress. The distinct ion between global and local residual stress should be made. Local residual stress refe rs to the remaining stresses caused by indentation. The stresses in th is case are near the indentati on site and drop off rapidly. Global residual stresses are those associated with the entire (externally) unloaded specimen and can be introduced by thermal processing, e.g ., from intentional or

PAGE 72

60 unintentional rapid cooling. Each of these re sidual stress systems is not uniform. Thus, for detailed calculations, an assumption must be made as to the nature of the stress profiles. In the present case, we assume that the residual stresses in the vicinity of the crack are constant and can be supe rimposed on the applied stress. Considerable research has been conducte d to estimate and me asure residual stress in bilayer glass caused by a mismatch between the coefficients of linear thermal expansion (Varsheneya and Pett i, 1978). Dilatometric analys is showed that the maximum difference in expansion coefficient between th e glass veneer and core materials was 0.4 (ppm K-1). Thus, the calculated residual compre ssive stress caused by thermal expansion differences between the veneer and core ce ramics was 13 MPa and does not fully explain the observed strength differences between the bilayer and monolithic specimens. Previous studies showed th at tensile or compressive re sidual stresses can develop because of different viscoelastic relaxa tion mechanisms (Appendix C) in elastic– viscoelastic composites (Anokye, 1989; Je ssen and Mecholsky, 1989). Tensile or compressive stress can develop and increase in the glass as shown in Figure 5-4 (Anokye, 1989). Even though heat treatments above and below the glass transition temperature of the glass veneer may create tensile or compre ssive global residual st resses, these stresses vary linearly through each layer of the beam as shown in Figure 5-5 (Scherrer, 1986).

PAGE 73

61 Figure 5-4. Theoretical calculati ons and experimental measurem ents of residual stress in borosilicate glass resulting from viscoela stic relaxation beha vior, as a function of different heat treatment temperatur es above and below the glass transition temperature for a bilayer system wh ere borosilicate glass (C 7052) was bonded to Kovar (Iron-Nickel-Cobalt alloy) metal (Anokye, 1989). There are several potential sources for stresses that result in spallation: global residual compressive stresses within the surface from elastic and viscoelastic processes, local residual tensile and compressive stresses adjacent to points of contact damage from previous loading, and tensile stresses from fl exural and/or subse quent contact loading. We suggest that the superposition of these st resses cause lateral cracks to develop and/or propagate to the surface (Lawn, 1993) resulting in spallation of segments of the veneering ceramic (Fig 5-6). Experimental Theoretical T g = 465 C Heat Treatment Temperat ure (C) (4 h hold) Residual Stress (MPa)

PAGE 74

62 Figure 5-5. Stress distribution in layers of a bilaye r specimen when the outer layer is in compression. R1, R2 are the stresses in ceramic 1 and ceramic 2, respectively (Scherrer, 1986). Figure 5-6. Schematic illustrati on of radial and lateral crac ks at the surface with or without residual stress. Lateral cracks are initiated near the base of the plasti c deformation zone below the contact area and spread out laterally on a pl ane closely parallel to the specimen surface (Lawn and Wilshaw, 1975). In a severe cont act event these cracks can propagate to the Median (Radial) Crack Lateral Cracks No global residual stress Global residual compressive stress Compression + Tension R2 R1 1 2

PAGE 75

63 surface to cause material chipping. Observati ons of sharp contact fracture in glass showed that lateral extension of a crac k occurs during unloading (Lawn and Swain, 1975). Residual stresses were identified as th e primary driving force for lateral cracking in another study (Evans and Wilshaw, 1976). Flexural strength and the apparent fractu re toughness of bilayer ceramic bars are mainly determined by the veneer layer when the critical crack initiates from the veneer surface and the critical crack is confined totally within the veneer layer. Global compressive residual stress in the veneer layer significantly increases the flexural strength of bilayer cerami c composite bars. However, global compressive residual stresses that lead to greater tensile stresses under the contact point are also the main cause for the observed chipping, i.e. spallation fracture of bila yer all-ceramic prostheses.

PAGE 76

64 CHAPTER 6 OPTIMIZATION OF RESIDUAL STRESSE S IN BILAYER DENTAL CERAMIC COMPOSITES In addition to material selection and ge ometric design, residual stresses in brittle materials can be a major factor in the impr ovement of the apparent fracture toughness of bilayer ceramic composites. Therefore, it is important to determine the magnitude and distribution of residual stresses. Residual stre sses can be caused because of rapid cooling or mismatch in thermal coefficients of expansion in the components of a ceramic composite. Also, existing cracks can propagate and cause failure due to residual stresses during the use of the material. Residual stresses can be tailored for in creasing strength and a pparent toughness of the brittle materials. Previously, evaluation of residual stresses and enhancing its effects through heat treatment techniques in meta l-ceramic composites have been studied (Dehoff and Anusavice, 1989; Hsueh and Evans, 1985; Scherrer, 1986; Anusavice et al. 1989; Rekhson, 1979). However, the effect of residual stress on apparent fracture toughness on bilayer dental ceramic co mposites is still relatively unknown. In addition to the factors described above the viscoelasticity (Appendix C) of the glass components in ceramic composites is another determin ing cause of residual stresses (Scherrer, 1986). When glasses are cooled from a temperatur e above their glass transition temperature to one which is below it, they remain in a nonequlibrium thermodynamic state, from which their properties (therm odynamic, mechanical, etc.) evolve slowly

PAGE 77

65 towards equilibrium (Scherrer, 1989). This proc ess must be taken into account to predict accurately the long-term performance of glasses, especially when they are used in bilayer dental ceramic composites. Bilayer dental ceramics can exhib it a significant amount of residual stress (Taskonak et al., 2002). These stresses influe nce the flexural stre ngth and apparent fracture toughness of bilayer cera mic restorations, mainly when the critical crack initiates within the veneer surface. Global compre ssive residual stress in the veneer layer significantly increases the flexur al strength of bilayer material s. However, global residual stresses may also be the main cause for th e observed chipping or spalling fracture of bilayer all-ceramic prostheses (Chapter 3). The objective of this study is to demonstrate that a co mpressive residual stress can be selectively distribu ted in bilayer dental ceramics to strengthen the material without causing lateral crack propag ation within the surface. 6.1. Materials and Methods An analytical procedure based on fractu re mechanics was used to obtain the magnitude of residual stress in the ceramic veneer and determine its effect on lateral crack growth based on measurements on the fr acture surface. Mars hall and Lawn (1979) described an indentation tec hnique to measure near-surface residual stresses. We also used another technique in this study simila r to the indentation technique based on the superposition of stress intensity factors. The use of these two techniques provides an opportunity to validate the correctness of our data. Conway and Mecholsky (1991) proposed the analysis of resi dual stresses using fracture surf ace analysis. Our technique includes five steps to determ ine the effect of residual stress on the lateral crack propagation in bilayer ceramic core/ven eer structures. These steps include:

PAGE 78

66 1. Heat treatment of the bilayer ceramic specimens above, at, or below the glass transition temperature of the applied glass veneer to introduce variable magnitudes and distributions of compressive and tensile residual stresses at the veneer surface. 2. Introduction of sharp indentation cracks us ing a Vickers indenter at load of 4.9 N to produce controlled median-radial cracks and lateral cracks. These indentations also introduce local re sidual stresses. Preliminary data suggest that lateral cracks grow at an indentation load of 4.9 N (Taskonak et al., 2002). 3. Calculation of the flexural strength, apparent fracture toughness of heat-treated bilayer ceramic specimens, and the “tru e” fracture toughness of monolithic glass veneer specimens using fractographic analysis. 4. Calculation of the maximum residual stre ss in bilayer ceramic specimens using a fracture mechanics equation. 5. Measurement of lateral cracks using fractogr aphy and determination of the change in lateral crack length as a result of local and global residual stress differences. 6.1.1. Sample Preparation For these procedures bilayer ceramic co mbinations were fabricated following accepted dental laboratory procedures. Bar specimens were fabricated from a lithiadisilicate-based core ceramic (Li2O•2SiO2) (Empress 2, Ivoclar Vivadent AG, Schaan, Liechtenstein) and a zirconia core cera mic (3Y-TZP) (Lava, 3M ESPE, Seefeld, Germany). A hot pressing technique was used for the preparation of lithia-disilicatebased core ceramic. However, zirconia core specimens were provided by the manufacturer since CAD-CAM instrument s were needed for their preparation.

PAGE 79

67 Prepared core bars were veneered with a silicate glass (Empress 2 veneer, Ivoclar Vivadent AG, Schaan, Liechtenstein) and Lava Ceram veneer (Lava, 3M ESPE, Seefeld, Germany). Empress 2 veneer was applied to Empress 2 core ceramic and Lava zirconia core was veneered with Lava Ceram veneer Glass veneers have been selected with coefficients of thermal expansion (CTE) compa tible with those of their core ceramics to ensure that the residual stress contributi on caused by a CTE difference is negligible. Ceramic veneer powders were incorporated with mixing liquid (Ivoclar Vivadent AG, Schaan, Liechtenstein) to obtain slurry so lutions, which were brushed onto the core ceramics that were placed in slightly oversiz ed silicone molds. Additionally monolithic veneer specimens were prepared to serve as control groups. The following core/veneer specimens were prepared: Combination A: Lithia disilicate gl ass ceramic core (Experimental) / glass veneer (Empress 2 veneer) (Appendix A) Combination B: Zirconia core (Lava core) / glass veneer (Lava veneer) (Appendix A) Following sintering of the veneering ceramic s using firing schedules listed in Table 61 (See Fig 6-1 for fast cooling), the excess veneering ceramic was ground with a 75 grit diamond disk and sequentially polished to a 200 0 grit finish on a metallographic polisher (Model 41-1512, Buehler Ltd., Lake Bluff, IL, USA) while exposed to a continuous flow of tap water. Fifteen specimens were prepared for each group. Monolithic veneer specimens from Empress 2 veneer and Lava Ceram powders were prepared in addition to the bilayer ceramic specimens.

PAGE 80

68 Table 6-1. Glass veneers and firi ng schedules used for sintering Glass Veneer Starting T (C) Heating rate (C/min) Firing T (C) Holding t (min) Vacuum T on-off (C) Cooling Rate Lava veneer 450 40 810 1 451-810 Fast Cooling Empress 2 Veneer 403 60 802 1 450-801 Fast Cooling The final dimensions of the ceramic bilayer and monolithic bar specimens were 1.7 mm (height) x 4.0 mm (width) x 25.0 mm (lengt h). The veneer thickness of bilayer specimens was 0.6 mm for each specimen. The core/veneer thickness ratio was 6/11. The testing span length/specimen thickness ratio was selected to be 15/1.7 to avoid large beam deflections. After final polishing, dimensi ons of the specimens were measured with a micrometer. To optimize residual stresses, each core/ven eer material group was heat-treated at 40C below and 20C, an d 40C above, the glass transition temperature, and at the glass transition temperature of the glass veneer (Table 6-2). In addition to the heat treated and tempered specimens a control group for each bilayer material was included. Specimens in the control group were neither heat-treated nor tempered. However their cooling rate when veneered was considered as fast cooli ng (Fig 6-1). Profiles of surface temperature of bilayer bars versus time for the two coo ling conditions are shown in Figure 6-1. This plot was described previously by Anusavice et al. (1989). Since the same cooling techniques were performed, we used the da ta in the graph from that publication as a reference. Tempering was performed by blasting compressed air directly on them as they were removed from the furnace. A nozzle with a 4 mm diameter (Fig. 6-2) was placed 20 mm above the disc at a pressure of 0.34 MPa for 90 s.

PAGE 81

69 Specimens with no heat treatment specimens were fast cooled as recommended by the manufacturer using the cooli ng rate shown in Fig 6-1. S ubsequently, specimens were indented using a Vickers inde nter at load of 4.9 N. Figure 6-1. Profiles of surface temperature ve rsus time for the two cooling conditions (Anusavice et al., 1989). Figure 6-2. Design of temper ing apparatus (Anusavice et al ., 1989) Sample

PAGE 82

70 Table 6-2. Heat treatment and cooling schemes for bilayer ceramic groups. Groups Starting T (C) Heating rate (C/min) Holding T (C) Holding t (min) Cooling Rate Lava Core/ Lava veneer (Group 1) 400 60 525 60 Tempered Lava Core/ Lava veneer (Group 2) 400 60 565 60 Tempered Lava Core/ Lava veneer (Group 3) 400 60 585 60 Tempered Lava Core/ Lava veneer (Group 4) 400 60 605 60 Tempered Lava Core/ Lava veneer (Control Group) No heat treatment No heat treatment No heat treatment No heat treatment Fast cooling Experimental Core / Empress 2 Veneer (Group 1) 400 60 500 60 Tempered Experimental Core / Empress 2 Veneer (Group 2) 400 60 540 60 Tempered Experimental Core / Empress 2 Veneer (Group 3) 400 60 560 60 Tempered Experimental Core / Empress 2 Veneer (Group 4) 400 60 580 60 Tempered Experimental Core / Empress 2 Veneer (Control Group) No heat treatment No heat treatment No heat treatment No heat treatment Fast cooling

PAGE 83

71 6.1.2. Compositional, Physical and Thermodyn amic Property Characterization of Specimens Glass transition temperatures of the glass veneers were determined using differential thermal analysis (DTA) vs. temperature plots. The thermal expansion coefficients of each ceramic composite component were measured using a single pushrod dilatometer and a quartz standard (Orton Ceramic Foundation, We sterville, OH). Beam-bending viscometer tests were performed on a lithia-disilicateglass ceramic core (experimental core) to determine its viscoelastic behavior (Appendix C) at high temperature since it has a glass matrix. Residual stress caused by the thermal mismatch of the veneer ceramic and core ceramic was estimated using the equation below (Lawn, 1993): R= T/[(1+ C)/2 EC+(1-2 V) Ev] (6-1) where R is the residual stress, is the difference between linear thermal expansion coefficients of the veneer and core ceramics, T is the difference between the glass transition temperature of the veneer ceramic and room temperature, and and E are PoissonsÂ’s ratio and YoungÂ’s modulus, respectively. S ubscripts C and V refer to core and veneer ceramics, respectively. In addition, the elemental composition of each ceramic composite component was determined using X-ray diffraction. YoungsÂ’ moduli and PoissonsÂ’ ratios were measured using an ul trasonic technique. Also, the density of the each ceramic component was measured using a pycnometer.

PAGE 84

72 6.1.3 Mechanical Testing Methods 6.1.3.1 Determination of residual stresse s using an indentation technique A Vickers diamond pyramid indenter was us ed to produce well-defined radial and lateral crack patterns at the cen ter of each sample (Fig 5-1). Specimens were stored in air for 48 h at room temperature to allow crack growth that is caused by residual stresses. Microscopic examination of the contact sites was made to record post-indentation crack development (Marshall and Lawn, 1980). Inde ntation induced radial cracks were measured and recorded for each sample to observe crack behavior associated with residual stresses. The 2c dimensions of the radial half-pe nny cracks (Fig. 5-1) were measured for each of ceramic bilayer composite and heat treat ment condition. The actua l fracture toughness values of the monolithic glass specimens we re determined from the following relation, with use of values of load (P), crack size (c), and hardness (H) obtained from micro indentation data: Kc=1.6 x 10-8(P/c2)(E/H)1/2 (6-2) Where Kc = fracture toughness in MPam1/2 (actual fracture toughness of glass used for R calculations). P = indentation load = 4.9 N c = crack size (m) E = elastic modulus (GPa) H = hardness (GPa) = 0.47 P/a2 a = projected length of the i ndenter half-diagonal dimension (m)

PAGE 85

73 The surface residual stress valu es in MPa for the heat tr eated (T) specimens were calculated from the following relation, which was adopted from the theory developed by Marshall and Lawn (1978): R = (Kc/2)( /Cb)1/2[1-(Cm/Cb)3/2] (6-3) where Cb and Cm are indent crack sizes (Fig 51) for the bilayer composite and monolithic glass specimens, respectively. 6.1.3.2 Determination of residual stre sses using fracture mechanics The flexural strength of the specimens was determined using a four-point flexure test fixture with a 15 mm outer span and a 5 mm i nner span (crosshead speed of 0.5 mm/min) using a universal testing machine (Model 4465, Instron Corp., Canton, MA). The veneer surface was placed in tension for all flexure te st specimens since the crack initiation sites of the clinically failed FPDs were on the veneer surface (Chapter 3). The flexure strength ( f) of the bilayer ceramic co mposites was calculated using mechanics, classical laminatedplate theory and the strength ( f) of the monolithic specimens was calculated using simple beam theory (Beer and Johnston, 1981) (Appendix B). Each specimen was studied with a stereoscopic microscope (Bauch & Lomb Inc., Rochester, NY, USA) at 160 X magnification. Cr ack initiating flaws and failure stresses were measured to determine the fracture toughness of each specimen. Scanning electron microscopic (JSM-6400, Jeol, Tokyo, Japa n) examinations were performed on representative specimens.

PAGE 86

74 Apparent fracture toughness, KC, was calculated using the fracture mechanics equation: KC= Y( ) f ( c )1/2 (6-4) where Y( ) is a geometric factor equal to 1.65 for a surface indentation induced flaw with local residual stress and 1.24 for surface flaws wi thout local residual stress, for equivalent semicircular flaws of depth “a” a nd half width “b” (Mecholsky, 1994), f is the calculated flexural strength, and c is the crack size [c= (ab)] (Mecholsky, 1991). Residual stress was calculated from the following equation: r = [ Y (2/ 1/2) ac1/2 KC] / [ Y (2/ 1/2)c1/2] (6-5) where r is the residual stress, a is the applied stress at failure, and KC is the fracture toughness of the glass veneers (C onway and Mecholsky, 1989). Two-way analysis of variance was performed to determine whether the differences between group mean s for flexural strength and re sidual stresses of specimens are statistically significant. Statistical analysis was also performed to determine whether the differences between the mean lateral crack sizes were statistically significant depending on the local and global re sidual stresses of the groups. 6.2. Results Measured thermal and physical properties of ceramic composite components are given in Table 6-3.

PAGE 87

75 Table 6-3. Physical, mechanical and thermal properties of core and veneer ceramics Material Properties Lava Core (zirconia) Lava Veneer (Amorphous glass) Experimental Core (Li2O2SiO2 glass-ceramic) Empress 2 Veneer (Amorphous glass) Density ( ) 6.22 g/cm3 2.53 g/cm3 2.56 g/cm3 2.53 g/cm3 Elastic modulus (E) 155 GPa 58 GPa 96 GPa 64 GPa PoissonÂ’s ratio ( ) 0.34 0.27 0.24 0.23 Tg 565C 687 C 540C CTE ( ) (25C-600C) 10.7 (ppm K-1) 10.2 (ppm K-1) 10.4 (ppm K-1) 9.8 (ppm K-1) Kc 5.5 MPam1/2 0.7 MPam1/2 3.1 MPam1/2 0.7 MPam1/2 Vickers Hardness (H) 11.2 GPa 4.4 GPa 6.1 GPa 4.9 GPa Density 6.22 g/cm3 2.53 g/cm3 2.56 g/cm3 2.53 g/cm3 Flexure Strength ( ) (4-point flexure) 634 MPa 53 MPa 228 MPa 52 MPa Xray diffraction analysis was performe d to identify any crystal phases in the veneer layers. However, the veneer layers were found to be essentially amorphous by Xray diffraction analysis because there were no distinct peaks observed in the graph between the angles of 5 and 110 degrees. X-ray diffraction analyses of Empress 2 veneering ceramic revealed a large amor phous background signal that represents the concentration of the glass phase. However, we observed the main peaks of lithia disilicate in the experimental core ceramic (Li2O2SiO2 glass-ceramic) at diffraction angles (2 ) of 23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the dominant peak (highest intensity) at 25.0 degrees, whic h corresponds to the (111) crystallographic plane of the monoclinic phase. Also, the other veneeri ng ceramic, Lava veneer, exhibits an

PAGE 88

76 amorphous background with three minor peak s. These peaks were observed at 2 diffraction angles of 26.7, 34.0, and 51.7 degrees th at revealed a slight presence of alpha quartz. The results are shown in Fig. 6-3. Lava core ceramic is 3Y-TZP (Surez et al., 2004) (Appendix A). Figure 6-3. (A) X-ray diffraction pattern of Empress 2 veneer shows an amorphous structure. (B) X-ray diffraction pattern of Lava veneer shows slight alpha quartz peaks at 2 values of 26.7, 34.0, and 51.7 de grees. (C) X-ray diffraction pattern of an experimental core ceramic reveals lithia disilicate peaks at 2 values of 23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the dominant peak (highest intensity) at 25.0 degrees. Dilatometric analysis showed that there was a slight difference between the thermal expansion coefficients of Empress 2 veneer and Experimental core ceramics (Table 6-3). The values were 9.8 (ppm K-1) and 10.4 (ppm K-1) for the Empress 2 veneer and experimental core ceramic, respectiv ely. The glass transition temperature (Tg) of the Empress 2 veneer ceramic was determined to be 540 C and the difference between Tg A B C

PAGE 89

77 and room temperature is 515 C. The Poisson’s ratios of the veneer ceramic ( V) and the core ceramic ( C) are 0.23 and 0.24, respectively. The Young’s moduli of the veneer ceramic ( EV) and core ceramic ( EC) are 64 GPa and 96 GPa, respectively. The calculated residual stress for the specimens using the pa rameters above and equation (6-1) is 13 MPa (compression). In addition, dilatometric anal ysis also showed that th ere was a slight difference between the thermal expansion coefficients of Lava veneer and Lava core ceramics (Table 6-3). The values were 10.2 (ppm K-1) and 10.7 (ppm K-1)for the Lava veneer and Lava core ceramics, respectively. The glass transition temperature (Tg) of the Lava veneer ceramic is 565 C and the difference between Tg and room temperature is 540 C. The Poisson’s ratios of the veneer ceramic ( V) and the core ceramic ( C) are 0.27 and 0.34, respectively. The Young’s moduli of the veneer ceramic ( EV) and core ceramic ( EC) are 58 GPa and 155 GPa, respectively. The calcu lated residual stress for the specimens using the parameters above and equa tion (6-1) is 22 MPa (compression). Beam-bending viscometer data re vealed creep occurring above 700 C in experimental core ceramic (Li2O•2SiO2 glass-ceramic) (DeHoff, personal communications). As a result we calculated the relaxation function for the experimental core ceramic. This plot revealed a glass transition temperature for experimental core ceramic which is 687 C (Fig 6-4). Shown in Fig 6-4 is the plot for the experimental core for which the intersection of the two straight lines occurs at a temperature of approximately 687 C (Tg). Although the transition from liquid to glass occurs over a range of temperatures, we have selected th e intersection of the two lines as the glass transition temperature (Tg) for the experimental core ceramic.

PAGE 90

78 y = -395.63x + 34.52 R2 = 0.0016 y = 71547x 40.504 R2 = 0.8557 0 5 10 15 20 25 30 35 40 0.000940.000990.001040.001090.001140.001190.00124 Series A Series B Series C low temp high temp Figure 6-4. Natural logarith m of shear viscosity ( s) versus inverse absolute temperature for experimental ((Li2O•2SiO2 glass ceramic) core (DeHoff, personal communications). For all specimens, the mean longitudinal and transverse indentation-induced crack sizes with standard deviation of each gr oup are summarized in Table 6-4. Based on oneway ANOVA analysis, the differences be tween the mean indentation-induced longitudinal and transverse crack sizes of bila yer specimens are not significantly different (p > 0.05). However, the difference between mean indent crack sizes of monolithic specimens and bilayer specimens are statistically significant (p 0.05). For groups with a lithia-disilicate-ba sed glass-ceramic core, a one-way ANOVA analysis showed that there was a statistical ly significant difference between the mean flexure strengths of monolithic and bilayer specimens (p 0.05), but there was no 1/T(K-1) Ln( s)

PAGE 91

79 statistically significant difference (Table 64a) between the mean flexural strength of bilayer specimen groups (p > 0.05). Table 6-4a. Mean flexural strength ( ), standard deviation (SD), residual stress ( R) from fracture surface analysis (FSA) and i ndent crack sizes, longitudinal and transverse indentation induced crac k sizes, residual stresses of groups calculated using flexural strength differences between monolithic veneer ( f(m)) and bilayer specimens ( f(b)) and apparent fracture toughness (Kc) of bilayer (Experimental core/Empress 2 veneer) and monolithic veneer specimens. Actual fracture toughness Groups (Experimental core/Empress 2 veneer) S.D. (MPa) Longitudinal indent crack length (m) Transverse indent crack length (m) R S.D. (Indent) (MPa) R S.D. (FSA) (MPa) R f(b) f(m) (MPa) Kc (MPam1/2) Tg-40 (500 C) 64 11 76 14 74 12 -60 48 -18 8 12 1 Tg (540 C) 67 6 76 16 73 15 -70 61 -19 11 15 1 Tg+20 (560 C) 64 10 69 10 76 12 -50 37 -21 14 12 1 Tg+40 (580 C) 67 9 74 9 80 6 -39 34 -16 7 15 1 No heat treatment 66 6 72 14 84 14 -20 16 -14 12 14 1 No heat treatment (Empress 2 Veneer) 52 10 93 22 92 15 0 0 0 0.7

PAGE 92

80 Specimens in groups with zirconia core delaminated at the veneer-core interface without exception (Fig 6-5). In order to de termine the flexural strength of the glass veneer in these specimens, composite beam theory (Beer and Johnston, 1981) (Appendix B) was used; however, simple beam theory wa s used for calculation of failure stress for the remaining zirconia layer that failed af ter interface delamination. Failure stresses of both glass layers and zirconia layers are s hown in Table 6-4b. There was a statistically significant difference between the mean veneer layer failure stresse s of the group heattreated at 525 C and the group that was not heat treated (control) (p 0.05). There was no statistical difference between the glass laye r failure stresses of other heat treatment groups and the control group (p > 0.05). Also, there was no statistical difference between the mean failure stresses of the zirconia core ceramics (p > 0.05). Figure 6-5. Interfacial delamination between zirconia core and a glass veneer layer. Zirconia core layer Glass veneer layer Delamination

PAGE 93

81 Table 6-4b. Mean flexural strength ( ), standard deviation (SD), residual stresses ( R) of groups calculated using flexural stre ngth differences between monolithic veneer ( f(m)) and bilayer specimens ( f(b)), longitudinal and transverse indentation induced crack sizes, a nd apparent fracture toughness (Kc) of bilayer (Lava core/Lava veneer) and monolithic veneer specimens. For groups with a lithia disilicate crystal phase, failure origins were located within the tension surface of each specimen and not at the interface between the veneer and core. For most of the composite specimens, the in dentation cracks were th e critical flaws that controlled crack propagation (Fig 5-2a). Howe ver, some specimens failed from porosity Groups (Lava core/Lava veneer) (veneer) S.D. (MPa) (core) S.D. (MPa) Longitudinal indent crack length (m) Transverse indent crack length (m) R S.D. f(b) f(m) (MPa) Tg-40 (525 C) 146 13 659 166 78 9 80 10 -93 40 Tg (565 C) 123 11 553 116 84 15 84 10 -70 11 Tg+20 (585 C) 118 15 615 102 85 10 90 12 -65 15 Tg+40 (605 C) 119 16 631 117 68 10 73 9 -66 16 No heat treatment 112 7 629 151 66 6 72 14 -59 7 No heat treatment (Lava Veneer) 53 6 ______ 101 39 103 30 0

PAGE 94

82 at the surface. These cr itical flaws, within the tensile region of the composite bars were approximately semi-elliptical in shape (Fig. 5-2b). Also, for heat-treated groups spallation occurred within the veneer surf ace during fracture and it was not possible to observe the fracture origin (F ig. 6-6). Since delamination occurred in specimens with a zirconia core following the fracture of the gla ss veneer, actual fracture origins were lost and it was not possible to perform fract ure surface analysis (Fig. 6-5). Figure 6-6. The fracture origin of the sp ecimen was not detectable. Wake hackle markings indicate the direction of the fracture origin in the veneer layer. However, the part of the specimen cont aining the fracture origin chipped away from the veneer layer. Even though all specimens failed from the veneer surfaces, there was a statistically significant difference between the calculated apparent fracture toughness values (Eq. 6-4) of bilayer specimens with lithia disilicate gla ss-ceramic based core Wake hackle marking Spallation site

PAGE 95

83 (experimental core) and monolithic veneer (Empress 2 veneer) specimens (p 0.05) (Table 6-4a). The fracture toughnesses of m onolithic Empress 2 veneer and Lava veneers were identical [0.7 MPam1/2] (Table 6-3.). However, the fracture toughness of the zirconia core (Lav a core) was 5.5 MPam1/2 whereas the fracture toughness of lithia disilicate-based glass-ceramic core (experimental core) was 3.1 MPam1/2 (Table 6-3.). Residual stress values of experimental co re (lithia-disilicate-based glass-ceramic) /Empress2 veneer specimens (Eq. 6-5) ranged between -24 MPa and -14 MPa (compressive) (Table 6-4). The highest comp ressive residual stress was observed in a group that was heat treated 20 C above the glass transition temperature, whereas the lowest residual stress was observed in th e group with no heat-treatment. No residual stress was detected in the monolithic specimens (Table 6-4a). There was no statistically significant difference between the residual stresse s of the heat-treated groups (p > 0.05). Residual stress values of Lava zirconia core/Lava veneer ranged between -92 MPa and -58 MPa (compression) (Table 6-4b). The highest residual compressive stresses were present in the group that was heat treated 20 C below the glass transition temperature whereas the lowest residual st ress was observed in the group with no heat treatment. No residual stress was detected in the monolithic specimens (Table 6-4b). There was a statistically significant differen ce between the residual stresses of the heat treated groups (p 0.05). Since the glass veneer la yers of these specimens were delaminated and fractured in many pieces (Fig.6-5), fracture surface analysis was not possible to perform. As a result equation 65 could not be used to determine residual stresses. Instead differences between the mean failure stresses of glass veneer layers of

PAGE 96

84 heat treated groups and the mean strength of monolithic glass sample were used to calculate residual stresses in the Lava zirconia core/Lava glass veneer specimens Residual compressive stresses obtained using the indentatio n technique were greater than the failure stresses of expe rimental core/Empress 2 bilayer ceramic composites. As a result, the data obtained us ing this technique were ignored and the data from fracture surface analysis were used for statistical analysis. 6.3. Discussion It is important to understand how the stress distribution in dental ceramics can be used to predict failure mechanisms. Stress development mechanisms become more complex especially when trilayer or bilayer composite ceramics are used for dental restorations. Clinical success is mostly predicted by the manuf acturers using thermal expansion mismatch data of the component s in multilayer ceramic composites (DeHoff and Anusavice, 2004). Bilayer dental ceramics can exhibit a significant amount of residual stress. These stresses influence the flexural strength and apparent fracture toughness of bilayer allceramic restorations and occur mainly when th e critical crack initiates within the veneer surface. Global residual compressive stress in the veneer layer significantly increases the flexural strength of bilayer materials. Howeve r, global residual stre sses may also be the main cause of chipping or spalling fract ure in bilayer all-ceramic prostheses. Calculations of residual stre ss in bilayer dental cerami cs using a fracture mechanics equation (Conway and Mecholsky, 1989) suggest that there is a significant amount of global residual compressive stress within the veneer layers of some bilayer materials (Table 6-4). The distinction between gl obal and local residual stress should be

PAGE 97

85 emphasized. Local residual stress refers to th e residual stresses induced by indentation. The stresses are near the inde ntation site and drop off rapi dly. The distribution of these stresses can be represented by the three-di mensional Boussinesq stress field (Lawn, 1993) (Fig. 1-1). Each of these residual stress systems is not uniform. Thus, in detailed calculations an assumption must be made as to the nature of the stress profiles. In the present case, we assume that the residual stresses in the vicinity of the crack are locally constant and can be superimposed. Global re sidual stress refers to overall stresses distributed from the surface toward all over the sample. Consequently, tensile or compressive st resses will develop an d increase in the glass. At temperatures lower than the gla ss transition temperature of the glass veneer, viscosity is high and molecular motions are ve ry slow. Thus, it will take a long time for the glass to stabilize. At higher temperatures, but lowe r than the glass transition temperature of the glass veneer, the viscosity is low and the glass stabilizes rapidly upon cooling, and residual compressive stress will develop within the glass layer (Anokye, 1989). In our study we observed a significant difference betw een the residual stress in experimental core/Empress 2 bilayer ceramic s developed by contraction mismatch (13 MPa), and the total residual st ress (21 MPa) calculated using Equation 6-5. Difference is mostly caused by the viscoleastic relaxation of the glass layer near its glass transition temperature. In addition there were slight changes in residua l stress values between heat treatment groups. These changes show that residual stresses can be increased or decreased using different heat treatments in the ceramic-ceramic composites. However, the slight changes are most likely associated with the viscoelastic behavior (Appendix C)

PAGE 98

86 of the experimental core ceramic above 687 C. Viscoelastic behavi or of the glass layer controls the magnitude of resi dual stress during cooling from 687 C to 540 C (Tg). The possibility that crystallization of the ve neer layer can result in an increase in strength and toughness of the bilayer compos ites compared with the glass veneer is rejected because X-ray diffraction analysis revealed no evidence of a crystal phase, indicating that the veneering ceramics consist predominantly of an x-ray amorphous glass phase. Thus, the strengthening mechanism fo r the bilayer compos ites is not caused by crystallization of the veneer layers. The results of the evaluation of residual st ress as a function of heat treatment around the glass transition temperatures are s hown in Figures 6-7a and 6-7b. During the isothermal heat treatment of the bilayer cera mic composites in the gl ass transition range, the density of the glass veneer will increase because of structural relaxation, while there is no change in the ceramic core (Scherrer, 1986). Depending on the viscosity of the glass veneer at the heat treatment temperature and cooling rate, either te nsile or compressive global residual stresses w ill develop (Scherrer, 1986, Anokye, 1989). For example at temperatures below Tg e.g., 500 C, for the Empress 2 glass veneer, the viscosity is high and molecular motions are very slow. As a re sult it will take long time for the glass to stabilize (Fig. 6-7a). At temperatures above Tg, e.g. 580 C, the viscosity is low and the glass stabilizes rapidly, imme diately inducing stress relaxatio n. Near the glass transition temperature, i.e ., 540 C, the structure does not stabilize as quickly as it does at higher temperatures (Fig. 6-7a). Also, we have to account for the fact that the glass transition temperature range is a function of the coo ling rate. Additionally, the cooling rate will

PAGE 99

87 determine the amount of stress relaxation. These factors can play an important role in evaluation of the global residual stresses in bilayer systems. 560 580 500 540 0 5 10 15 20 25 480500520540560580600 Heat Treatment Temperature C (1 h hold)Residual Stress (MPa) Figure 6-7a. Residual stress as a function of h eat treatment temperature in bilayer lithiadisilicate core/glass veneer ceramic composites [Residual stresses are given in Table 6.4a: R (FSA)]. In the zirconia-glass bilayer ceramic system (Lava) all specimens delaminated, which is considered a failure for dental ceramic composites (Fig 6-5). Residual stresses in the zirconia core-glass veneer system (Lava) were greater than these in the lithia disilicate core-glass ven eer system (Fig. 6-7b). This may be associated with the viscoelastic behavior of the lith ia disilicate glass ceramic core above 687 C. However, zirconia is elastic through the enti re heating-cooling temperature range. Also

PAGE 100

88 this factor provides a more significant residual stress diffe rence between heat treatment groups of the zirconia core-glass ven eer system (Lava) (Fig. 6-7b). 525 605 585 565 0 10 20 30 40 50 60 70 80 90 100 520540560580600620 Heat Treatment Temperature C (1 h hold)Residual Stress (MPa) Figure 6-7b. Residual stress as a function of heat treatment temp erature in bilayer zirconia core/glass veneer ceramic composites [Residual stresses are summarized in Table 6.4b: R = f(b) f(m)]. The reason that the zirconia core-glass ven eer system (Lava) exhibited the greatest residual stress may be because the glass transition temperature (Tg) shifts as a function of cooling rate. The glass transition te mperature was determined to be 565 C from DTA experiments. However, during processing the cooling rate is fast er and thus the Tg most likely shifts to a lower temperature. This shif t increases the viscosity of the glass at lower

PAGE 101

89 heat treatment temperatures, which results in greater residual stre ss values at those temperatures. The reason that the indentation techni que (Marshall and Lawn, 1977) provided unusually large residual stress values, i.e, greater compressive stress than the failure stress of the composite (Table 6-4a), can be because of an assumption that was used to derive Equation 6-3. Equation 6-3 was developed for tempered glass specimens. In determining the appropriate indentation fracture func tion for surface residual stresses ( R), it is assumed that the thermal tempering process does not affect the fracture parameters of the test specimen (Marshall and Lawn, 1977). The indent crack size in the absence of the residual stress field is assumed to be governed by a simple relation (Marshall and Lawn, 1977): P/Co 3/2 = Kc/ = constant (6-6) where Kc is the fracture toughness for the residual stress-free ceramic, is a dimensionless indenter constant, P is the indenter load, and Co is the median indent crack size without global residual stress. Using the assumption above, Equation 6-3 is derived from: P/CT 3/2 = (Kc/ )[1+ R( CT)1/2 Kc] (6-7) where is a dimensionless crack geometry term ( 4/ 2) and CT is the crack size with global residual stress. Howeve r, it is expected that the factor depends on the residual stress value. Therefore, the constant in Equation 6-7 should ch ange, since the crack size is affected by the presence of the residual st ress from tempering. Future studies need to be conducted to determine an accurate wa y of modifying and using equation 6-3.

PAGE 102

90 Residual stresses in bilayer dental compos ite systems can be tailored through heat treatment around the glass transition temperature of glass veneers. Even though the heat treatment either above or below the glass transition temperature of the glass veneer can modify the magnitude and distribution of residual tensile or compressive stresses, these stresses vary linearly through each layer of the beam (Scherer, 1986). The stress distri bution in bilayer specimens can be estimated using finite element analysis and it can be confirmed us ing a viscoelastic analysis method (Appendix C) (Scherer, 1986). An axially symmetric Boussinesq stress fiel d can be used to describe the residual stress distribution during contact under the point of loading (Lawn, 1993). Superposition of global residual stresses asso ciated with viscoelastic relaxation (Appendix C) on the Boussinesq contact stress field may lead to la teral crack growth within the veneer layer. The combination of these stresses causes latera l cracks to develop and/or propagate to the surface (Lawn, 1993) resulting in sp allation of segments of the veneering ceramic (Fig. 68). Previous studies have shown that subsurface cracks can grow to the surface with repeated cyclic loading that present the appearance of spallati on (Chen and Mecholsky, 1996). The results from the present study a llow a better understanding of these failure mechanisms and the reasons for spallation from the surface of ceramic prostheses.

PAGE 103

91 Figure 6-8. (A) Fracture initia tion from a median (radial ) indent crack. (B) Higher magnification of (A). Lateral crack growth was caused by residual surface stresses. B A Glass veneer Lithia-disilicate core Fracture origin Lateral Cracks

PAGE 104

92 CHAPTER 7 CONCLUSIONS The main objectives of this work were to determine the failure mechanisms in bilayer dental ceramic composites and to in crease their reliability using residual stress optimization through thermal processing parame ters. It is essen tial to understand all aspects of residual stress development to optim ize stress distributions. We used fracture mechanics techniques to analyze residual stresse s in bilayer ceramic composites. Fracture toughness was measured using fracture surface an alysis and indentation techniques. The main conclusions are summarized as follows. 1In a two-year clinical evaluation period preceding the start of this dissertation, 50 % of the 20 Empress 2 FPDs had fractured. Relative to these failed prostheses, 80% fractured at the connector regions a nd 20% chipped within the veneer layer. Because of this high failure rate, fractu re surface analysis of the clinically fractured restorations is indicated to determine th e cause of the failures. 2Fracture surface analysis of clinically failed Empress 2 FPDs based on lithiadisilicate-glass ceramic reveal that crack s initiate from the veneer surface and propagate through the core layer without stopping at the inte rface of the bilayer ceramic prostheses. We calculated relatively low failure stresses in core/veneer specimens that are compared with those reported by Hland et al (2000). We reported that the increase in strength of th e bilayer ceramics occu rred because of a global compressive residual stress (Taskonak et al ., 2002) (Table 3-1). However,

PAGE 105

93 it is not only the global resi dual stress that plays a role in the failure mechanism of bilayer dental ceramics. Local resi dual tensile and compressive stresses adjacent to points of contact damage fro m previous loading, and tensile stresses from flexural and/or subsequent cont act loading also can cause failure. The superposition of these stresses can ca use lateral cracks to develop and/or propagate to the surface (Lawn, 1993). Even if the stresses are not sufficient enough to propagate median cracks, they mi ght be sufficient to propagate lateral cracks and cause chipping of the glass veneer. This was most probably the case for the chipping failures. We conclude that fracture initiation sites of these glassceramic FPDs occurred primarily within oc clusal surfaces in the veneered units and the crack propagation patterns app ear to be controlled by the loading orientation. 3The flexural strength and fracture toughness of layered ceramics loaded in flexure (with strong bonding at the in terface) are governed by the flaw characteristics of the material in tension, and residual stress. Plots of failure stress vs. c-1/2 show that there is a compressive re sidual stress contribution ( 28 MPa) that leads an increase in the failure stress of lithia disilicate-glass-ceramic-based bilayer ceramic bars. 4Flexural strength and the apparent fracture toughne ss of bilayer all-ceramic restorations are mainly determined by th e veneer layer when the critical crack initiates from the veneer surface. Global co mpressive residual stress in the veneer layer significantly increases the flexural st rength of bilayer materials. However,

PAGE 106

94 global compressive residual stresses are also the main cause for the observed chipping, i.e. spallation, fracture of bila yer all-ceramic prostheses. 5Viscoelasticity of the glass ve neer, cooling rate, and thermal expansion/contraction coefficient differe nces between core and veneer layers govern the residual stresses in bilayer dental composites. 6An axially symmetric Boussinesq stress fi eld can be used to represent residual stresses during contact near the point of loading (Lawn, 1993). Superposition of global residual stresses associated with visc oelastic relaxation or rapid cooling of the core/veneer ceramic composites on th e Boussinesq contact stress field may lead to early lateral fractu re of the material. The results from this study will allow a better understanding of these failure mechanisms and the reasons for spallation from the surface of all ceramic prostheses. 7Residual stresses in bilayer dental compos ite systems can be tailored through heat treatment near the glass tr ansition temperature of gla ss veneers by modifying the viscosity of the glass veneer layer and selectively controlling the stress relaxation and structural relaxation times.

PAGE 107

95 APPENDIX A COMPOSITIONAL CHARACTERIZATION OF CERAMIC COMPOSITE COMPONENTS Xray diffraction analysis was performe d to identify any crystal phases in the veneer layers. However, the veneer layers were found to be essentially X-ray diffraction amorphous because there were no distinct peaks observed in the graph between the angles of 5 and 110 degrees. X-ray diffraction analyses of Empress 2 veneering ceramic and Eris veneering ceramic revealed a larg e amorphous background signal that represents the concentration of the glass phase. Howeve r, we observed the main peaks of lithium disilicate in the experimental core ceramic and Empress 2 core ceramic (Li2O2SiO2 glass-ceramic) at diffraction angles (2 ) of 23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the dominant peak (highest inte nsity) at 25.0 degrees, which corresponds to the (111) crystallographic plane of the monoclin ic phase. Previously, it was reported that the size of the elongated Li2O2SiO2 crystals ranged from 0.5 m to 4 m for Empress 2 core ceramic and from 0.5 to 0.2 m for e xperimental core ceramic (Della Bona, 2000). This is the only difference between e xperimental core ceramic and Empress 2 core ceramic. Also, the other veneering cerami c, Lava veneer, reveals an amorphous background with 3 minor peaks. Th ese peaks were observed at 2 diffraction angles of 26.7, 34.0, and 51.7 degrees that revealed a slight presence of alpha quartz. The results are shown in Fig. A-1. Lava core ceramic consis ts of yttria-stabilized tetragonal zirconia (3Y-TZP) (Surez et al., 2004).

PAGE 108

96 Figure 6-3. (A) X-ray diffrac tion pattern of Empress 2 veneer shows an amorphous structure. (B) X-ray diffraction pattern of Lava veneer shows slight alpha quartz peaks at 2 values of 26.7, 34.0, and 51.7 de grees. (C) X-ray diffraction pattern of an experimental core ceramic reveals lithia disilicate peaks at 2 values of 23.8, 24.4, 25.0, 30.6, 37.9, 38.5, and 39.4 degrees, with the dominant peak (highest intensity) at 25.0 degrees. A B C

PAGE 109

97 APPENDIX B FLEXURAL STRESS CALCULATIONS USING COMPOSITE BEAM THEORY Flexural stresses within the tensile surface of the specimens were determined using a transformation technique for bilayer beam specimens (Beer and Johnston, 1981; Thompson, 2000). The simple beam formula applies to monolithic materials. Laminated composites need to be analyzed in a different manner. There are several ways to accomplish the task of analyzing laminated composite beams. One way, selected here, is the transf ormed beam method. The transf ormation of a bilayer beam of core ceramic and veneer into a uniform beam of monolithic core ceramic is shown in Fig A-1. Empress 2 core / Empress 2 veneer bilayer composite used in this description to demonstrate flexural stress calculatio n using composite beam theory. The transformation factor is determined as follo ws: Empress 2 core has an elastic modulus of EC 96 GPa, and Empress 2 veneer has an elastic modulus of EE2V 64 GPa; therefore n=Ec/EE2V=1.5. The transformed width of the veneer layer was calculated using the aforementioned values [i.e., C = (1/n)(actual width)]. The transformed width of the veneer layer with E2V is 2.66 mm (C1=2.66 mm). Calculation of the centroid is shown in Table A-1. K, L, M, N, O, and P as shown in Fig. A-1 are the distances for the centroid of the composite parts. Centroid. Transformed and T-shaped cross se ction of a specimen was divided into two rectangles and the centroid of each type of specimen was calculated as shown in Table A-1. Centroidal Moment of Inertia. From the parallel axis theorem, the crosssectional moment of inertia (I) for the transformed beam in Fig. A-2 can be determined.

PAGE 110

98 I =[1/12 (k)(l)3+(k)(l)(m-n)2] +[1/12(o)(p)3+(o)(p)(r-s)2] (A-1) I =1.34 mm4 for composites with E2V Maximum Tensile Stress. Flexural stress ( ) determination in a transformed beam loaded in a 4-point can be determined by = Mc/I (A-2) where M is the maximum moment, c is the pe rpendicular distance to the centroid from the bottom of the transformed beam, and I is the moment of inertia. The maximum moment (M) for a simply supported beam loaded in the 4-point fixture configuration is defined by: M= P (L0-Li)/4 (A-3) where P is the failure load, Lo is the outer support span length and Li is the inner support span length.

PAGE 111

99 Table A-1. Calculation of centroids for each type of core/veneer combination. K, L, M, N, O, and P in Figure A.1 are the distances needed for the calculation of the centroid for the composite parts. A = volume, (6.0mm2) = 5.54mm3; = 0.92mm; = Centroid Figure A-1. Distances for the calculation of centroid of the composite parts Area (mm2) A, D (mm) (A, D) x Area (mm3) 1 (E2V) 2(Core Ceramic) (L) (M) (0.6)(2.66) = 1.6 (O) (P) (1.1)(4) = 4.4 Area 6.0 0.3 = K 1.15 = N 0.48 5.06 volume = 5.54 Core Ceramic P O L M K N

PAGE 112

100 Figure A-2. Distances from the centroidal-axis and the centroid of the composite parts o p l k m n r centroid y-axis z-axis centroidal-axis s

PAGE 113

101 APPENDIX C THE VISCOELASTIC THEORY AND FRACTURE MECHANICS METHODS FOR DETERMINATION OF RESIDUAL STRESSES C.1 Viscoelastic Theory When a strain is imposed on a glass bar in th e glass transition temper ature range, there is an immediate (elastic) response in addition to a time-dependent (viscoelastic) response. This immediate elastic response occurs because of the vibrational activity of atoms in the glass, while the time-dependent viscoelastic response occurs because of translational atomic activity or viscous flow of the glass (Anokye, 1989). During this flow, atoms rearrange themselves and the volume of the glass will recover to its level before the strain application. Thus, the change in thickness is compensated for by the change in the length of the glass bar. At this point the stress in the bar is zero. In the glass transition temperature range a sudden temperature change in a glass structure is accompanied by structural cha nges, which, in turn, produce changes in its properties. Thus, when a bilayer ceramic-gla ss composite is heat-treated in the glass transition temperature range, the density of the glass component will increase during structural relaxation while no changes take place in the ceramic components that do not contain a glass phase. Even th e initial stresses are zero, compressive or tensile stresses can develop in the glass component depending on the heat treatment temperature (above or below the glass transition temperature). The stresses will increase rapidly at first, and then they will continue to increase slowly until they reach a plat eau. The stabilization

PAGE 114

102 time will depend on the viscosity of the glass at the heat treatment temperature. If the viscosities are less than 1014-1016 Pa, the structure of the glass will stabilize rapidly. At these viscosity levels, the stresses in the glass should relax rapi dly at first then continue to relax more slowly until they disappear (Scherrer, 1986). Stress and structural relaxation times ar e controlled by viscosity. Stress relaxation is determined by the extent of the structural relaxation. As a result, it is expected that, at the upper heat treatment temperatures, structur al and stress relaxati on will proceed faster than at the lower annealing temperatures, be cause the greater the viscosity decreases the slower the molecular motion, and the less stru ctural relaxation there is (Anokye, 1989). In addition, at the lower heat-treat ment temperatures there is very little stress relaxation because the viscosity is high a nd the relaxation time is long. Scherrer (1986) introduced a viscoelast ic analogy through th e application of Laplace transform equations. Use of the La place transform demonstrates the analogy between the elastic constitutive equations a nd the transformed viscoe lastic constitutive equations. The Laplace transform of th e function f(t, x) with re spect to t is defined by L[f(t, x)] = f*(p, x) = e-pt f(t, x) dt (C-1) For example, if a uniaxial stress 1 is applied to an elastic material, the resulting strains are 1 = 1/E (C-2) and 2 = 3 = 1 (C-3) The corresponding equations for a vi scoelastic materials are: 0

PAGE 115

103 1 = 1 */M (C-4) and 2 = 3 = -N 1 (C-5) where M and N are apparent Young’s modulus a nd apparent Poisson’s ratio, respectively. They are given in terms of the transformed moduli as M = (3pG1 *G2 *) / (2G2 *+G1 *) (C-6) and N= (G2 *-G1*) / (2G2 *+G1 *) (C-7) To evaluate M and N we need the transforms of the relaxation moduli or compliances, which are represented by the b-function. For example, G1(t) = 2Goexp[-(t/ s)b] (C-8) where s is the shear viscosity. The equivalences of elastic equations that were derived using viscoelastic theory are shown in Table C-1 (Scherrer, 1986). Table C-1. Elastic – Viscoelastic Analogy Elastic Transformed Viscoelastic Equivalance Sk = 2Gek Sk = pG1 *ek 2G pG1 = 3K( -3 f) = pG2 ( *-3 f) 3K pG2 E=[3(3K)(2G)]/[2(3K)+2G] E=[3(pG2 *)( pG1 *)]/[2(pG2 *)+ pG1 *] E M = [3K-2G]/ [2(3K)+2G] = [pG2 *pG1 *]/ [2(pG2 *)+ pG1 *] N x = f + (1/E)[ x( y+ x)] x = f + (1/M)[ x N( y+ x)] E, M,N In the above table (Table C-1), G = shear modulus, K = bulk modulus, E = Young’s modulus, = Poisson’s ratio, M and N are the transformed viscoelastic

PAGE 116

104 counterparts of E and respectively, G1* and G2* are the transformed shear and bulk relaxation modulus respectively, p is the transform parameter, and Sk is the shear stress, = hydrostatic stress. Using viscoelastic theory and by knowi ng the elastic solutio n to a problem one can determine the viscoelas tic solution by using the transforms described above. C.2 Determination of Residual St resses Using Fracture Mechanics A four-step fracture mechanics approach wa s used to determine residual stress in the bilayer dental ceramics. This technique included: (1) Indentation of the specimens to co mpare indentation induced crack sizes. (2) Flexural strength measurement and compar ison of monolithic and bilayer specimens. (3) Apparent fracture toughness determina tion of bilayer specimens and fracture toughness determination of monolithic specimens. (4) Calculation of residual stress in b ilayer specimens using fracture mechanics equations. A fracture-mechanics analysis establishe s a quantitative basi s for the technique (Marshall and Lawn, 1977). The stress inte nsity factor for th e half-penny crack configuration can be written as: K = P/c3/2-2m R(c/ )1/2 (C-9) Where is the dimensionless contact constant which incorporates details of the indenter/specimen contact (Lawn and Wilsha w, 1975), m is a dimensionless modification factor which is unity when free surface effects and the stress gradients over the prospective crack depth are neglected (Mar shall and Lawn, 1977). The above equation (eq. C-9) includes two terms the first of which represents th e indentation driving force on

PAGE 117

105 the crack and the second of which repres ents the residual resistance force. The indentation crack remains stable if it sa tisfies the condition (M arshall and Lawn, 1977): K = Kc = constant (C-10) where Kc is a material property. Since, there will be no residual stress effect on nontempered (stress free) monolithic ceramic s, one can determine the surface stresses using the comparative indent crack measur ements between stressed and unstressed (control) specimens. Equation C-9 is modified to determine surface stresses using indent crack sizes (Marshall and Lawn, 1978): R = (Kc/2)( /Cb)1/2[1-(Cm/Cb)3/2] (C-11) where Cb and Cm are indent crack sizes (Fig 5-1) for the residually stressed and nonstressed specimens, respectively. In addition to the fracture mechanics technique described above we used another method for determination of residual stresses in bilayer dental ceramics, which is based on fracture surface analysis to validate the da ta. This method was described previously by Conway and Mecholsky (1989). This techniqu e employs an apparent fracture toughness factor which is described as: Kapp = Kc + KR (C-12) Apparent fracture toughness, KC, was calculated using the fracture mechanics equation: KC= Y( ) f (c)1/2 (C-13) where Y( ) is a geometric factor equal to 1.65 fo r an surface indenta tion induced flaw with local residual stress a nd 1.24 for surface flaws without local residual stress, for

PAGE 118

106 equivalent semicircular flaws of dept h “a” and half width “b” (Mecholsky, 1994), f is the calculated flexural strength, and c is the crack size [c= (ab)] (Mecholsky, 1991). Residual stress was calculated from the following equation: r = [Y(2/ 1/2) ac1/2 KC] / [Y(2/ 1/2)c1/2] (C-14) where r is the residual stress, a is the applied stress at failure, and KC is the actual fracture toughness of the glass vene ers (Conway and Mecholsky, 1989).

PAGE 119

107 APPENDIX D APPARENT FRACTURE TOUGHNESS AND RESIDUAL STRESS DATA The data contained here refers to information in Chapter 6. Specimen number (Experimental core/Empress 2 veneer) Tg-40 (500 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 55 62 1.6 -105 2 92 60 1.6 -118 3 64 79 1.1 -34 4 83 53 1.9 -171 5 85 64 1.5 -92 6 Lost data Lost data Lost data Lost data 7 85 75 1.2 -47 8 47 70 1.3 -62 9 75 85 1.0 -19 10 85 89 0.9 -10 11 96 89 0.9 -10 12 85 75 1.2 -47 13 77 89 0.9 -10 14 66 77 1.1 -40 15 66 68 1.3 -71

PAGE 120

108 Specimen number (Experimental core/Empress 2 veneer) Tg-40 (500 C) Fracture Load (N) F (MPa) [Composite beam theory] Flaw type a (m) 2b (m) c (m) Kc (MPam1/2) R (FSA) (MPa) 1 33 48 indent0.000100.000340.00013 0.9 16 2 35 51 indent0.000090.000310.00012 0.9 18 3 45 65 indent0.000110.000330.00013 1.2 33 4 33 48 indent0.000090.000320.00012 0.9 15 5 indent0.000110.000300.00013 6 48 69 pore 0.000110.000320.00013 1.0 21 7 44 64 indent0.000030.000100.00004 0.7 6 8 46 66 lost 9 56 81 pore 0.000090.000090.00006 0.8 16 10 43 62 indent0.000030.000120.00005 0.7 8 11 41 59 indent0.000040.000140.00005 0.7 9 12 57 83 indent0.000030.000120.00005 0.9 29 13 43 62 indent0.000060.000140.00006 0.8 16 14 55 80 indent0.000020.000150.00004 0.8 22 15 43 62 indent0.000110.000140.00009 0.9 23

PAGE 121

109 Specimen number (Experimental core/Empress 2 veneer) T g ( 540 C ) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 64 85 1.0 -19 2 60 58 1.7 -134 3 66 75 1.2 -47 4 89 62 1.6 -105 5 66 62 1.6 -105 6 62 75 1.2 -47 7 75 53 1.9 -171 8 62 75 1.2 -47 9 104 85 1.0 -19 10 66 77 1.1 -40 11 81 81 1.0 -28 12 70 51 2.1 -193 13 79 62 1.6 -105 14 104 96 0.8 0 15 66 68 1.3 -71

PAGE 122

110 Specimen number (Experimental core/Empress 2 veneer) Tg (540 C) Fracture Load (N) F (MPa) [Composite beam theory] Flaw type a (m) 2b (m) c (m) Kc (MPam1/2) R (FSA) (MPa) 1 42 60 lost 2 51 73 indent0.000110.000340.00013 1.41 42 3 37 53 lost 4 44 64 indent0.000050.000250.00008 0.95 24 5 45 65 lost 6 43 62 pore 0.000100.000400.00014 0.91 16 7 48 70 pore 0.000090.000170.00009 0.81 14 8 44 64 lost 9 52 76 pore 0.000240.000250.00017 1.23 29 10 48 70 pore 0.000140.000150.00010 0.87 16 11 50 73 buble 0.000070.000120.00007 0.74 11 12 48 70 pore 0.000030.000120.00005 0.78 16 13 44 63 pore 0.000090.000110.00007 0.66 4 14 49 71 lost 15 44 64 indent0.000060.000140.00006 0.83 18

PAGE 123

111 Specimen number (Experimental core/Empress 2 veneer) Tg+20 (560 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 77 102 0.7 8 2 66 75 1.2 -47 3 70 75 1.2 -47 4 60 58 1.7 -134 5 66 81 1.0 -28 6 72 66 1.4 -81 7 62 64 1.5 -92 8 77 79 1.1 -34 9 51 70 1.3 -62 10 64 66 1.4 -81 11 89 85 1.0 -19 12 64 83 1.0 -23 13 83 96 0.8 0 14 72 75 1.2 -47 15 58 70 1.3 -62

PAGE 124

112 Specimen number (Experimental core/Empress 2 veneer) Tg+20 (560 C) Fracture Load (N) F (MPa) [Composite beam theory] Flaw type a (m) 2b (m) c (m) Kc (MPam1/2) R (FSA) (MPa) 1 41 60 indent0.000050.000120.00005 0.71 10 2 53 77 indent0.000100.000270.00012 1.36 43 3 46 67 indent0.000030.000090.00004 0.70 9 4 39 56 lost 5 47 67 pore 0.000110.000220.00011 0.88 16 6 56 81 lost 7 42 60 lost 8 41 60 lost 9 46 66 lost 10 52 75 pore 0.000020.000020.00001 0.34 11 31 46 lost 12 38 55 lost 13 47 68 indent0.000060.000230.00009 1.04 29 14 49 70 pore 0.000010.000020.00001 0.30 15 35 51 lost

PAGE 125

113 Specimen number (Experimental core/Empress 2 veneer) Tg+40 (580 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 68 94 0.8 -3 2 70 75 1.2 -47 3 70 98 0.8 3 4 72 72 1.2 -54 5 62 64 1.5 -92 6 7 70 64 1.5 -92 8 89 85 1.0 -19 9 89 92 0.9 -7 10 64 66 1.4 -81 11 66 87 0.9 -14 12 92 89 0.9 -10 13 70 79 1.1 -34 14 81 81 1.0 -28 15 75 68 1.3 -71

PAGE 126

114 Specimen number (Experimental core/Empress 2 veneer) Tg+40 (580 C) Fracture Load (N) F (MPa) [Composite beam theory] Flaw type a (m) 2b (m) c (m) Kc (MPam1/2) R (FSA) (MPa) 1 45 65 pore 0.000110.000090.00007 0.68 5 2 45 66 lost 3 53 77 corner0.000090.000140.00008 0.97 25 4 51 73 lost 5 56 81 pore 0.000080.000150.00008 0.88 19 6 51 74 lost 7 45 65 indent0.000050.000190.00007 0.90 21 8 41 59 lost 9 35 51 lost 10 50 73 lost 11 47 68 indent0.000050.000090.00005 0.78 16 12 43 62 pore 0.000090.000260.00011 0.79 11 13 50 72 lost 14 31 46 lost 15 48 70 lost

PAGE 127

115 Specimen number (Experimental core/Empress 2 veneer) No heat treatment Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 89 72 1.2 -54 2 53 81 1.0 -28 3 60 87 0.9 -14 4 102 107 0.7 12 5 64 89 0.9 -10 6 7 64 89 0.9 -10 8 77 72 1.2 -54 9 64 87 0.9 -14 10 66 70 1.3 -62 11 77 87 0.9 -14 12 64 87 0.9 -14 13 81 117 0.6 21 14 15

PAGE 128

116 Specimen number (Experimental core/Empress 2 veneer) No heat treatment Fracture Load (N) F (MPa) [Composite beam theory] Flaw type a (m) 2b (m) c (m) Kc (MPam1/2) R (FSA) (MPa) 1 40 58pore 0.000080.000190.00009 0.685 2 44 63indent0.000100.000300.00012 1.1631 3 47 67pore 0.000050.000190.00007 0.696 4 54 79indent0.000070.000180.00008 1.1538 5 46 67pore 0.000060.000290.00010 0.8214 6 43 63corner0.000060.000210.00008 0.7913 7 47 68indent0.000030.000090.00004 0.688 8 43 62indent0.000030.000090.00004 0.644 9 48 69pore 0.000070.000210.00009 0.7913 10 52 75pore 0.000110.000240.00011 1.0023 11 46 67pore 0.000030.000120.00004 0.56-4 12 45 65indent0.000090.000110.00007 0.9021 13 42 61lost 14 41 59lost 15 40 58pore 0.000080.000190.00009 0.685

PAGE 129

117 Specimen number (Lava core/Lava veneer) Tg-40 (525 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 70 81 1.1 -9 2 94 96 0.9 13 3 89 92 0.9 8 4 75 85 1.0 -1 5 79 81 1.1 -9 6 72 68 1.4 -44 7 68 66 1.5 -53 8 79 72 1.3 -30 9 62 96 0.9 13 10 83 70 1.4 -37 11 75 75 1.3 -24 12 77 87 1.0 2 13 81 81 1.1 -9 14 92 77 1.2 -19 15 77 72 1.3 -30

PAGE 130

118 Specimen number (Lava core/Lava veneer) Tg-40 (525 C) Fracture Load (Veneer) (N) F (veneer) (MPa) [Composite beam theory] Fracture Load (core) (N) F (core) (MPa) [Simple beam theory] R F(b) F(m) (MPa)) 1 81 138 296 770 85 2 93 158 260 676 105 3 92 156 146 380 103 4 78 133 360 937 80 5 84 143 285 741 90 6 84 143 148 386 90 7 79 134 255 662 81 8 81 138 333 865 85 9 84 143 293 762 90 10 207 539 11 81 138 209 543 85 12 104 177 187 485 124 13 97 165 275 714 112 14 82 139 238 620 86 15 85 145 310 806 92

PAGE 131

119 Specimen number (Lava core/Lava veneer) Tg (565 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 77 77 1.2 -19 2 94 96 0.9 13 3 87 68 1.4 -44 4 81 75 1.3 -24 5 77 87 1.0 2 6 94 83 1.1 -5 7 79 83 1.1 -5 8 66 81 1.1 -9 9 89 75 1.3 -24 10 94 77 1.2 -19 11 83 96 0.9 13 12 128 102 0.8 20 13 81 94 0.9 11 14 75 83 1.1 -5 15 64 81 1.0 -9

PAGE 132

120 Specimen number (Lava core/Lava veneer) Tg (565 C) Fracture Load (Veneer) (N) F (veneer) (MPa) [Composite beam theory] Fracture Load (core) (N) F (core) (MPa) [Simple beam theory] R F(b) F(m) (MPa)) 1 171 444 2 186 484 3 268 697 4 79 134 305 792 81 5 77 131 144 375 78 6 216 562 7 79 134 205 534 81 8 77 131 201 522 78 9 64 109 264 686 56 10 66 112 247 643 59 11 63 107 188 489 54 12 184 478 13 76 129 170 441 76 14 74 126 189 490 73 15 66 112 251 652 59

PAGE 133

121 Specimen number (Lava core/Lava veneer) Tg +20 (585 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 81 89 1.0 5 2 75 77 1.2 -19 3 92 85 1.0 -1 4 87 89 1.0 5 5 77 81 1.1 -9 6 81 89 1.0 5 7 89 98 0.8 16 8 75 89 1.0 5 9 96 119 0.6 31 10 83 94 0.9 11 11 70 85 1.0 -1 12 94 81 1.1 -9 13 109 111 0.7 26 14 83 85 1.0 -1 15 81 79 1.2 -14

PAGE 134

122 Specimen number (Lava core/Lava veneer) Tg+20 (585 C) Fracture Load (Veneer) (N) F (veneer) (MPa) [Composite beam theory] Fracture Load (core) (N) F (core) (MPa) [Simple beam theory] R F(b) F(m) (MPa)) 1 78 133 181 469 80 2 82 139 213 553 86 3 62 105 267 694 52 4 84 143 267 693 90 5 67 114 270 702 61 6 65 111 243 632 58 7 71 121 161 418 68 8 79 134 276 717 81 9 56 95 248 646 42 10 182 474 11 61 104 234 609 51 12 70 119 236 613 66 13 262 682 14 69 117 295 767 64 15 60 102 215 558 49

PAGE 135

123 Specimen number (Lava core/Lava veneer) Tg +40 (605 C) Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 79 102 0.8 20 2 66 66 1.5 -53 3 87 96 0.9 13 4 53 66 1.5 -53 5 92 89 1.0 5 6 100 81 1.1 -9 7 77 62 1.7 -72 8 89 83 1.1 -5 9 75 72 1.3 -30 10 66 64 1.6 -62 11 79 85 1.0 -1 12 81 45 2.7 -218 13 66 62 1.7 -72 14 75 72 1.3 -30 15 89 79 1.2 -14

PAGE 136

124 Specimen number (Lava core/Lava veneer) Tg+40 (605 C) Fracture Load (Veneer) (N) F (veneer) (MPa) [Composite beam theory] Fracture Load (core) (N) F (core) (MPa) [Simple beam theory] R F(b) F(m) (MPa)) 1 59 100 238 619 47 2 307 798 3 56 95 245 637 42 4 81 138 348 904 85 5 62 105 260 676 52 6 71 121 278 723 68 7 228 594 8 257 669 9 82 139 249 647 86 10 73 124 229 594 71 11 237 617 12 187 485 13 70 119 178 462 66 14 76 129 204 530 76 15 198 514

PAGE 137

125 Specimen number (Lava core/Lava veneer) No heat treatment Longitudinal indent crack length (m) Transverse indent crack length (m) Kc (Indent) (MPam1/2) R (Indent) (MPa) 1 66 83 1.1 -5 2 72 75 1.3 -24 3 53 66 1.5 -53 4 75 81 1.1 -9 5 81 85 1.0 -1 6 70 53 2.1 -126 7 47 66 1.5 -53 8 55 68 1.4 -44 9 79 81 1.1 -9 10 64 64 1.6 -62 11 75 70 1.4 -37 12 70 68 1.4 -44 13 75 81 1.1 -9 14 79 81 1.1 -9 15 60 66 1.5 -53

PAGE 138

126 Specimen number (Lava core/Lava veneer) No heat treatment Fracture Load (Veneer) (N) F (veneer) (MPa) [Composite beam theory] Fracture Load (core) (N) F (core) (MPa) [Simple beam theory] R F(b) F(m) (MPa)) 1 64 109 241 626 56 2 65 111 207 539 58 3 66 112 301 783 59 4 70 119 262 682 66 5 65 111 339 881 58 6 59 100 175 456 47 7 63 107 282 734 54 8 64 109 174 452 56 9 65 111 258 672 58 10 70 119 283 735 66 11 64 109 312 812 56 12 62 105 164 427 52 13 63 107 154 400 54 14 74 126 263 684 73 15 70 119 215 558 66

PAGE 139

127 LIST OF REFERENCES Anokye WF, 1989. The effects of residual stress, processing and thermodynamic parameters on toughness of glass-metal s eals. Dissertation, Penn State University: State College, PA. Anusavice KJ, 1989. Citeria for selection of restorative materials: properties versus technique sensitivity. In: Anusavice KJ, editor. Quality evaluation of dental restorations: criteria for placement and replacement. Chicago: Quintessence; pp. 15-59. Anusavice KJ, DeHoff PF, Hojjatie B, Gr ay A, 1989. Influence of tempering and contraction mismatch on crack developm ent in ceramic surfaces. J Dent Res, 68:1182-1187. American Society for Testing and Mate rials (ASTM), 1999. Standard Practice for Fractography and Characterization of Fr acture Origins in Advanced Ceramics, ASTM Designation C1322-96a. Annual B ook of ASTM Standards, Vol. 15.01. American Society for Testing and Ma terials (ASTM), West Conshohocken, PA. Beer FP, Johnston ER, 1981. Mechanics of Materials. 2nd rev. ed. New York: McGrawHill, pp. 150-233. Campbell SD, Sozio RB, 1988. Evaluation of the fit and strength of all ceramic fixed partial denture. J Prosthet Dent; 59:301-306. Chen Z, Cuneo JC, Mecholsky, JJ, Jr ., Hu S,1996. Damage processes in Si3N4 bearing material under contact loading. Wear 198:19 Conway JC Jr, Mecholsky JJ Jr ., 1989 Use of crack branchi ng data for measuring nearsurface residual stresses in tempered glass. J Am Ceram Soc, 72(9): 1584-1587. Creugers NHJ, Kyser AF, vanÂ’t Hof MA, 1994. A meta-analysis of durability data on conventional fixed bridges. Community Dent Oral Epidemiol, 22:448-452. Dehoff PH, Anusavice KJ, 1989. Ef fect of visco-elastic beha vior on stress development in a metal-ceramic system. J Dent Res, 68(8): 1223-1230. Dehoff PH, Anusavice KJ, 2004. Creep functions of dental ceramics measured in a beambending viscometer. Dent Mater, 20:297-304.

PAGE 140

128 128 Della Bona A, 2001. Interfacial adhesion of de ntal, ceramic-resin systems. Dissertation, University of Florida, Gainesville, FL. Edelhoff D, Brauner J, Spiekermann H, Yildirim M, 2002. Two-year clinical evaluation of crowns and bridges made of IPS Em press 2. J Dent Res, 81, Abstr. No. 2584. Evans AG, Wilshaw TR, 1976. Quasi-plastic solid particle damage in brittle materials. Acta Metall, 24:939-956. Hland W, Schweiger M, Frank M, Rh einberger V, 2000. A comparison of the microstructure and properties of the IPS Empress 2 and the IPS Empress glassceramics. Applied Biomaterials, 53:297-303. Hsueh CW, Evans AG, 1985. Residul stress in metal-ceramic bonded strips. J Am Ceram Soc, 68(5):241-248. Hutchinson JW, Suo Z, 1992. Mixed mode cracki ng in layered materials. Advances in Applied Mechanics, 29:63-191. Jessen TL, Mecholsky JJ Jr., 1989. Viscoelastic effect of h eat treatment on the fracture toughness of metal-particulate / glass-matrix composites. J Am Ceram Soc, 72(11): 2094-2097. Kaplan EL, Meier P, 1958. Nonparametric es timation from incomplete observations. J Am Stat Assoc; 53:457-65. Kelly JR, Campbell SD, Bowen HK, 1989. Fracture surface analysis of dental ceramics. J Prosth Dent, 62:536-541. Kelly JR, Nishimura I, Campbell SD, 1996. Ceramics in dentistry: Historical roots and current perspectives. J Prosthet Dent; 75:18-32. Kerans RJ, Hay R S, Pagano NJ, Partasar athy TA, 1989. The role of the fiber-matrix interface in ceramic composites. American Ceramic Society Bulletin, 68:429-441. Khaund AK, Krstic VD, Nicholson PS, 1977. Infl uence of elastic and thermal mismatch on the local crack-driving force in britt le composites. J Mater Sci, 12(11):22692273. Lange FF Lawn BR, Evans AG, Marshall DB 1980. Elastic/Plastic indentation damage in ceramics: The medi an/radial crack system J Am Ceram Soc, 63(9-10):571-584. Lawn BR, 1993. Fracture of brittle solids. 2nd rev. ed. Cambridge: Cambridge University Press, pp. 194-306. Lawn BR, Swain MV, 1975. Microfracture beneath point indentations in brittle solids. J Mater Sci, 10(1):113-122.

PAGE 141

129 129 Lawn BR, Wilshaw TR, 1975. Indentation fractur e: Principles and applications. J Mater Sci 10(6): 10491081. Marshall DB, Lawn BR, 1977. An indentation technique for measuring stresses in tempered glass surfaces, J Am Ceram Soc, 60:86-87. Marshall DB and Lawn BR, 1978.Measurement of non uniform distribution of residual stresses in tempered glass di scs. Glass Technology, 19:57-58. Marshall DB and Lawn BR, 1979. Residual stress effects in sharp contact cracking, Part 1. Indentation fracture mechanic s, J Mater Sci, 14:2001-2012. Marshall DB, Lawn BR, 1980. Flaw characterist ics in dynamic fatigue: The influence of residual contact stresses. J Am Ceram Soc, 63(9-10): 532-536. Marshall DB, Lawn BR, Evans AG, 1982. Elas tic/Plastic indentation damage in ceramics: The lateral crack system J Am Ceram Soc, 65(11):561-566. McLean JW, Hughes TH, 1965. The reinforcem ent of dental porcelain with ceramic oxides. British Dental Journal, 119:251-267 Mecholsky JJ Jr., 1991. Quantitative fractography : An assessment. In: Fractography of Glasses and Ceramics. Proceedings of th e conference on the fractography of glasses and ceramics, August 3-6, 1986, Alfred, NY. Varner JR, Frchette VD, Quinn GD, editors. Westerville, OH: American Ceramic Society, pp. 413-451 Mecholsky JJ Jr., 1994. Quantitative fractographic analysis of fracture origins in glass. In: Fractography of glass. Bradt RC, Tre ssler RE, editors. New York, NY: Plenum Press, pp. 39-73. Mecholsky JJ Jr., 2001. Fractography of brittle materials. In: Encyclopedia of Materials: Science and Technology. St. Loui s, MO: Elsevier Science Ltd. Oh W, Gtzen N, Anusavice KJ, 2002. In fluence of connector design on fracture probability of ceramic fixed partial dentures. J Dent Res, 81(9):623-627. Olsson K, Furst B, Andersson B, Carlsson GE, 2000. A long term retrospective study and clinical follow-up study of In-Ceram Alumina FPDs. Int J Prosthodont, 13: 131135. Piddock V, Qualtrough AJ, 1990. Dental Cera mics – An update. J Dent, 18: 227-235. Pospiech P, Kistler S, Frasch P, Rammelsb erg P, 2000. Clinical evaluation of posterior crowns and bridges of Empress 2: Prelim inary results after one year. J Dent Res, 78:445. Prakash O, Sarkar P and Nicholson PS, 1995. Crack deflection with ceramic/ceramic laminates with strong interfaces. J Am Ceram Soc, 78:1125-1127.

PAGE 142

130 130 Rekhson SM, 1979. Annealing of glass-metal an d glass-ceramic seals. Part 1, Theory, Glass Thechnology, 20(1):27-35. Ryge G, Snyder M, 1973. Evaluating the clin ical quality of restorations. J Am Dent Assoc, 87:369-77. Ryge G, 1980. Clinical criteria. Int Dent J, 30:347-58. Scherer GW, 1986. Relaxation in Glass & Composites. 1st ed. New York: WileyInterscience publication pp. 75-174. Scurria MS, Bader JD, Shugars DA, 1998. Me ta-analysis of fixed partial denture survival: Prostheses and abutment s. J Prosthet Dent, 79:459-464. Sjgren G, Lantto R, Tillberg L, 1999. Clinical evaluation of all-ce ramic crowns (Dicor) in general practice. J Prosthet Dent, 81:277-84. Sorensen JA, Cruz M, 1998. A clinical inves tigation on three-unit fixed partial dentures fabricated with a lithium disilicate gla ss-ceramic. Pract Periodont Aesthet Dent, 11:95-106. Sorensen JA, Kang S-K, Torres TJ, Knode H, 1998. In-Ceram fixed partial dentures: Three year clinical trial results J Calif Dent Assoc, 26:207-214. Surez MJ, Lozano JFL, Salido MP, Martnez F, 2004. Three-year clin ical evaluation of In-Ceram Zirconia posterior FPD s. Int J Prosthodont, 17:35-38. Taskonak B, Mecholsky JJ Jr, Anusavice KJ, 2002. Determination of residual stresses in bilayer ceramics using fracture mechanics. J Dent Res, 78:445. Thompson GA, 2000. Influence of relative laye r height and testing method on the failure mode and origin in a bilayer dental ceramic composite. Dent Mater, 16:235-243. Thompson JY, Anusavice KJ, Naman A, Morris HF, 1994. Fracture surface characterization of clinically failed al l-ceramic crowns. J Dent Res, 73:1824-1832. Thoules MD, 1989. Some mechanics for the adhe sion of thin films. Thin Solid Films, 181:397-406. Tooley FV, 1985. Handbook of glass manufacture, 2nd ed. New York: Ashlee Publishing Company, Inc. Varshneya AK, Petti RJ, 1978. Finite element analys is of stress in glass-metal foil seals. J Am Ceram Soc, 61(11-12): 498-503. Walton TR, 2002. An up to 15-year longitidunal study of 515 metal-ceramic FPDs: Part 1 Outcome. Int J Prosthodont, 15; 439-445.

PAGE 143

131 BIOGRAPHICAL SKETCH Burak Taskonak was born in Edirne, Turkey. He was admitted to dental school, at Marmara University, Istanbul, Turkey in 1992 and he received his Doctor of Dental Surgery (D.D.S.) degree in February, 1998. Burak completed coursework require ments for the specialty degree in prosthodontics while working as a private pract itioner in his dental office. During his specialty training between th e years of 1998 and 2000, he also instructed clinical and preclinical dental students at Marmar a University, School of Dentistry. He was awarded a fellowship in 1999 fr om New York University, College of Dentistry for conducting resear ch in dental biomaterials. In August 2000, Burak was admitted to the Doctor of Philosophy program at the Department of Materials Engineering and Scienc e, University of Florida. He specialized in dental biomaterials. He tutored incoming freshman dental students at the college of Dentisty for over two years. He worked in the Department of Dental Biomaterials, College of Dentistry during his doctorate education. Upon completion of the doctoral degree program, Burak plans to work as a facu lty member in a dental school and conduct scientific research. He aspires to take activ e roles in international research organizations.