Abstract
Objectives
To provide background information and guidance as to how to use fractography accurately, a powerful tool for failure analysis of dental ceramic structures.
Methods
An extended palette of qualitative and quantitative fractography is provided, both for in vivo and in vitro fracture surface analyses. As visual support, this guidance document will provide micrographs of typical critical ceramic processing flaws, differentiating between pre- versus post sintering cracks, grinding damage related failures and occlusal contact wear origins and of failures due to surface degradation.
Results
The documentation emphasizes good labeling of crack features, precise indication of the direction of crack propagation (dcp), identification of the fracture origin, the use of fractographic photomontage of critical flaws or flaw labeling on strength data graphics. A compilation of recommendations for specific applications of fractography in Dentistry is also provided.
Significance
This guidance document will contribute to a more accurate use of fractography and help researchers to better identify, describe and understand the causes of failure, for both clinical and laboratory-scale situations. If adequately performed at a large scale, fractography will assist in optimizing the methods of processing and designing of restorative materials and components. Clinical failures may be better understood and consequently reduced by sending out the correct message regarding the fracture origin in clinical trials.
1
Foreword
First, the authors would like to make it clear that this paper by no means seeks to repeat or replace any of the current standards or published books that have marked the field of fractography and failure analysis . The National Institute for Standards and Technology (NIST) recommended practice guide by George Quinn, which has been recently updated, contains all the information needed for a fractographer to perform good descriptive (qualitative) and quantitative fractography and is available online at no cost .
This paper intends to provide selected guidance to assist any researcher who is starting to use dental fractography for in vitro or clinical failures in documenting and reporting the relevant findings appropriately. The focus of this paper as presented is on fractographic principles observed on brittle facture planes, as found in ceramics. Fractography of resin composites lies beyond the scope of this guidance project. Currently, the quality of the information delivered in the literature varies significantly, depending on the level of expertise of both the researcher and the reviewer. These guidelines therefore should help researchers to adopt a more standardized methodology for reporting fractographic data, especially within the dental literature, so that the discussions and conclusions drawn are fully supported by a thorough fractographic analysis. Specific documented examples of fracture origins identified on both in vitro specimens and in vivo restorations or replicas will reinforce the key role played by fractography in the dental field in relation to quality control, processing, prosthesis design, surface grinding damage, and in situ surface wear degradation. The examples presented herein are meant to open a new analytical dimension to strength-testing papers and studies dealing with clinical complications (survival and success rates) that usually list the existing fractures or chippings but rarely perform any type of fractographic failure analysis. With the currently available knowledge delivered through standards, books, scientific articles, and hands-on training courses, the dental community still needs to become more involved in applying this powerful analytical tool accurately.
2
Introduction
The first use of clinical fractography in the dental literature involving ceramic clinical failures goes back to 1989 and 1990, when Kelly et al. started to analyze fracture surfaces of failed all-ceramic dental restorations with the objective of finding the fracture origin . Since then, the use of fractography has increasingly grown particularly in the past 10 years for both in vitro lab-scale specimens or in vivo ceramic restorations. Fractography is a powerful tool that allows for an accurate failure analysis based on the interpretation of microscopic fracture surface features that reveal the direction of crack propagation, pointing back to the origin or cause of failure. Fractography can be performed in a qualitative or descriptive way by means of recognition of surface crack features that indicate the direction of crack propagation. This method is mostly used in clinical failure analysis . Quantitative fractography is applied in materials engineering studies and provides quantitative measurements of fracture surface features, particularly the critical flaw (origin), which has a specific shape and size and is used to determine the fracture toughness or the fracture stress based on fracture mechanics relations . Such quantitative evaluation has been applied to fractured clinical ceramic restorations, allowing determination of flaw sizes and estimation of failure stresses . Fractures may occur from critical stress concentrations, cyclic fatigue assisted by stress corrosion (slow crack growth), or from a combination of mechanisms involving processing methods and restoration design. Interpretation of clinical failures is not always straightforward as it is highly dependent on the available recovered parts and on the ceramic microstructure, glasses being easier to analyze than polycrystalline ceramics. Nevertheless, if the fractographic analysis is accurately performed, the information obtained may disclose relevant processing or design problems and then, measures can be taken to avoid similar failures in the future. The findings obtained from fractography have a key role to play in the mechanical performance of the product, the development or improvements of materials, in the manufacturing and design, in the handling, laboratory grinding adjustments and finishing/polishing procedures. The dental materials researcher, while testing existing products or new materials under development, can help to raise awareness of potential causes of failure and provide the necessary feed-back to all the actors involved.
Both descriptive and quantitative fractography should be part of research projects in which ceramic fracture is involved, whether during testing or clinical trials. The selected examples in this paper will show the various applications of fractography as a valuable and necessary tool to perform failure analysis on fractured surfaces of dental ceramics. Along this text the reader will find documentation regarding critical flaw identified on the fractured parts of dental ceramic in different testing conditions, such as flexural strength, fracture toughness, multilayer interfaces, thermal stress, fatigue, clinical use and wear.
2
Introduction
The first use of clinical fractography in the dental literature involving ceramic clinical failures goes back to 1989 and 1990, when Kelly et al. started to analyze fracture surfaces of failed all-ceramic dental restorations with the objective of finding the fracture origin . Since then, the use of fractography has increasingly grown particularly in the past 10 years for both in vitro lab-scale specimens or in vivo ceramic restorations. Fractography is a powerful tool that allows for an accurate failure analysis based on the interpretation of microscopic fracture surface features that reveal the direction of crack propagation, pointing back to the origin or cause of failure. Fractography can be performed in a qualitative or descriptive way by means of recognition of surface crack features that indicate the direction of crack propagation. This method is mostly used in clinical failure analysis . Quantitative fractography is applied in materials engineering studies and provides quantitative measurements of fracture surface features, particularly the critical flaw (origin), which has a specific shape and size and is used to determine the fracture toughness or the fracture stress based on fracture mechanics relations . Such quantitative evaluation has been applied to fractured clinical ceramic restorations, allowing determination of flaw sizes and estimation of failure stresses . Fractures may occur from critical stress concentrations, cyclic fatigue assisted by stress corrosion (slow crack growth), or from a combination of mechanisms involving processing methods and restoration design. Interpretation of clinical failures is not always straightforward as it is highly dependent on the available recovered parts and on the ceramic microstructure, glasses being easier to analyze than polycrystalline ceramics. Nevertheless, if the fractographic analysis is accurately performed, the information obtained may disclose relevant processing or design problems and then, measures can be taken to avoid similar failures in the future. The findings obtained from fractography have a key role to play in the mechanical performance of the product, the development or improvements of materials, in the manufacturing and design, in the handling, laboratory grinding adjustments and finishing/polishing procedures. The dental materials researcher, while testing existing products or new materials under development, can help to raise awareness of potential causes of failure and provide the necessary feed-back to all the actors involved.
Both descriptive and quantitative fractography should be part of research projects in which ceramic fracture is involved, whether during testing or clinical trials. The selected examples in this paper will show the various applications of fractography as a valuable and necessary tool to perform failure analysis on fractured surfaces of dental ceramics. Along this text the reader will find documentation regarding critical flaw identified on the fractured parts of dental ceramic in different testing conditions, such as flexural strength, fracture toughness, multilayer interfaces, thermal stress, fatigue, clinical use and wear.
3
Qualitative fractography (in vitro)
Qualitative fractography is performed by means of optical tools such as a stereo microscope, SEM, Field emission—SEM usually combined with Energy Dispersive X-ray spectroscopy (EDX) in case of a localized need for elemental identification. Detailed reading on tools and equipment can be found in George Quinn’s practice guide of fractography .
3.1
Origins in strength testing
Ceramics are known to have a brittle behavior, with only approximately 0.01% of elastic elongation and no detectable plastic deformation. Therefore, linear elastic fracture mechanics (LEFM) is applicable to ceramics and describes stresses at crack tips, slow and fast crack growth and catastrophic failure when the stress states have exceeded the materials’ fracture toughness. Identification of crack origins from which the fracture started to propagate can provide valuable information every time new material developments or designs are created. Determination of relevant mechanical properties such as (i) fracture toughness (K Ic ), relating defect size and stress at failure, (ii) strength, for which a probabilistic approach using the weakest link distribution is involved, or (iii) fatigue testing establishing fatigue limits, will regularly include the use of fractographic analysis.
When carrying out a strength test with ceramic materials one needs to remember that all materials have an inherent population of defects (also called flaws) due to processing, which include powder composition (purity, homogeneity), powder compaction (pressing parameters, agglomerate size, powder friction), sintering (cracks and pores), presintered or solid state sintered ceramic machining and handling (scratches, edge chipping, cracks) . These defects vary in size, shape, distribution and orientation. They are either volume distributed from the fabrication process (pores, large grains, inclusions) or surface distributed from surface treatment (precracks, edge chips, machining scratches or grooves) . The material will fail when the nominal strength is overcome by the stress peak concentrated around a defect. An example of brittle fracture is shown in a zirconia bend bar ( Fig. 1 ). The tensile side is at the bottom where the origin is found. A compression curl is at the top where the compression side is located. Flexure testing usually will show a compression curl on the opposite side to the failure origin. The very first approach using fractography of strength tested specimens is to orient the broken surface such as to have the tensile side at the bottom and compression side at the top. The compression curl is usually very easy to identify on bend bars and a quick stereomicroscope image with the right lateral illumination will highlight relevant fracture features including the area of the fracture origin ( Fig. 1 ). Further SEM analysis ( Fig. 2 ) at higher power will allow identification and characterization of the type of flaw responsible for failure.
Similar porous regions, as shown in Figs. 2 c and 3 bc, are the result of interstices or pores remaining between agglomerates that underwent poor deformation during powder compaction. Zirconia grains that sintered freely in air (i.e. unconstrained sinter conditions) will have round shaped grains versus a more angulated grain shape seen on the bulk fractured surface as a result of inter or transgranular fracture. The shape of the grains (round versus angled) may serve as a diagnostic tool during failure analysis and thus understand if the defect was introduced before or post sintering. An example is given in Fig. 4 , which shows grinding cracks introduced in a zirconia bend bar in the presintered state. Later in this paper, grinding cracks in the presintered state resulting from reshaping zirconia CAD–CAM framework will be shown ( Figs. 14 and 15 ).
3.2
Origins labeled in Weibull or S–N graphs
Specimens of the same ceramic material will not fail at one reproducible strength value but will have a strength distribution value based on their flaw population. The Weibull distribution is based on the theory that the largest structural defect in a loaded body controls it’s strength . The Weibull two parameter distribution contains a scale (σ 0 ) and a shape parameter (Weibull modulus, m). The scale parameter is called the characteristic strength σ 0 and corresponds to the 63.2% failure probability (see Fig. 5 ) whereas the Weibull modulus m indicates the homogeneity of the strength data, thus expressing reliability. A Weibull modulus in the range between 10 and 40 is related to ceramic materials with less variation during strength testing, i.e., a narrow flaw population and overall improved quality. A low m (between 1 and 10) means a wide distribution, large spread and low reliability, which requires improvement and thus a closer look at the reasons for failure. Generally speaking, increases in strength and reliability will come from a significant reduction in the critical flaw size by means of improvements made to the processing methods. A thorough and practical review of the Weibull analysis illustrating the importance of flaw size distributions when using different test configurations has been published by Quinn and Quinn . The authors analyzed every single specimen and identified different failure mechanisms leading to clustered flaw populations. An important prerequisite for a thorough Weibull analysis is a sufficient number of specimens. A sound analysis is based on thirty or more samples in a group (EN 843-5). What is very helpful in the data analysis is to present a fractography montage in conjunction with a Weibull strength graph , or to follow-up with SEM fractographic analysis to show the type of flaws encountered during strength testing . Overlooking the step of identification of the critical flaw population may lead to wrong conclusions when comparing reliability data between materials or when claiming that one ceramic is significantly stronger or weaker than another without understanding the causes of fracture. An example of a fractographic montage is given in Fig. 5 .
Similarly to strength specimens, fatigue tested specimens should also be fractographically analyzed for critical flaw assessment. The type of flaw (ceramic processing or surface induced) can be labeled directly on the S–N plot ( Fig. 6 ) and selected fractographic images of the critical flaws discussed to explain the findings. The early and high stress cycles may activate a different population of flaws compared to that activated in specimens cycled for longer times. In a S–N (Wöhler) fatigue study on 3Y-TZP , S–N data showed that processing defects (pressing of powder granules) were primarily activated under conditions of high stress/low cycle (<2000) fatigue, whereas the grinding surface flaws were the dominant ones when subjected to low stress/high cycle (from 2000 to 10 6 ) fatigue. Of course the location, size and shape of the critical flaw, whether intrinsic or from grinding, have to be considered, as well as the stress level and loading direction.
4
Quantitative fractography (in vitro)
Fracture toughness (K Ic ) expresses the material’s resistance to unstable crack extension and is a design relevant material property. A separate ADM guidance document introduces the principles of fracture mechanics and fracture toughness . Depending on the fracture mechanics literature the flaw size is denoted either “a” or “c”. The following relationship K Ic = Y σ f √a (Griffith Criterion, Eq. (1)) describes the critical stress intensity factor, K Ic , which is linked to a stress at failure (σ f ), a crack depth (a) and a stress intensity shape factor coefficient Y which will range between a value of 1.1 and 1.99 depending on the overall crack shape (width 2 c and depth a ). The crack depth can be precisely calculated or approximated from crack depth over width (a/c) ratios . One has however to be familiar with finding the origin using fractography and measuring the correct critical crack size for fracture toughness determination . The extent of the critical flaw size a is defined by the area of the fracture mirror which is typically a smooth and semielliptical region around the fracture initiating defect . Several fracture toughness tests exist and among them is the Surface-Crack in Flexure (SCF) , which requires specific crack size measurements using fractography. A few research papers in the dental literature have applied this test successfully and have provided images to show the delineated critical crack size for fracture toughness calculation.
One interesting aspect of failure analysis is to work with the parameters in Eq. (1). Hence, toughness can be estimated from crack size and stress measurements ( Fig. 7 ), or stress can be estimated from crack size and known toughness values , or estimates of crack sizes can be given from known toughness and stress information when one has to find the fracture origin and wants some magnification scale information to search for it. However, it is important to keep in mind that the material needs to meet the requirements of LEFM, and that failure must have occurred from a remote stress field and in mode I (tensile crack opening). An example of chip damage induced from grinding and viewed on a mirror polished bonded interface is shown in Fig. 7 . The chip depth is assimilated as a critical flaw of a crack size (a) (Eq. (1)). Fracture stress estimates were calculated from Eq. (1) based manufacturer’s K Ic and then extrapolated to potential losses of strength when comparing to the manufacturer’s reported strength .
The well delineated fracture origin in Fig. 7 has a peculiar microstructure that differs from the bulk microstructure and corresponds to excessive grain growth. This is due to a localized concentration of 2.8 wt% of CaO used as a sintering catalyst and was detected by energy dispersive X-ray spectroscopy (EDX). This flaw, made of large grains, was the fracture origin and was responsible for the low fracture stress (680 MPa) in this 4-point bending fractured bend bar specimen. The critical flaw (a) measured 13 μm deep and 37 μm wide (2c). The calculated critical stress intensity shape factor Y depth is 1.4 using solutions from Scherrer et al. . Hence, the resulting fracture toughness using Eq. (1) would be in this case only 3.4 MPa√m which is rather low compared to the usual value of approximately 4.5–5 MPa√m reported for 3Y-TZP.
By using the backscattering mode in the SEM, it becomes often possible to highlight structure features and distinguish different phases, which can complement a first SEM image. As an example, Fig. 8 shows another zirconia specimen of the same research with a failure origin from excessive grain growth similar to that seen in Fig. 7 . The backscattering mode at 3500× reveals many large grains forming one large critical flaw activated during 4 point-bending. For this particular specimen, the fracture stress σ f was 586 MPa, the measured critical flaw depth a = 20 μm and width 2c = 60 μm, the calculated Y depth = 1.42 using Scherrer et al. . Based on Eq. (1), the calculated K Ic was 3.7 MPa√m. One explanation for such low toughness may be the additional processing step of post sintering Hot Isostatic Pressing (post-HIP) performed at 1400 °C for 2 h which, together with a 2.8 wt% of CaO, may have contributed to this localized grain growth.
The problem of contamination or detrimental chemical additives involved with the failure origin is illustrated in Fig. 9 to which EDX was added. Two failure origins from strength testing are documented in 3Y-TZP bend bars. The first one is a large pore at the failure site containing alumina particles, possibly from powder contamination. The second is a flaw containing a concentration of 2.8 wt% CaO, which contributed to localized grain growth, weakening the zirconia and from which the crack started (data from Ref. ).
5
Qualitative fractography (in vivo)
5.1
Origin identification using a systematic approach
As with lab-scale in vitro specimens, fractographic failure analysis has to be systematic in its approach. The very first objective is to be familiarized with all the possible fracture surface features encountered in ceramics. The best available reference for that is the NIST recommended practice guide for fractography of glasses and ceramics by George Quinn . Several clinical papers have since then used a systematic approach for clinically failed specimens and it is gaining momentum within the dental research community but never or rarely used in clinical trials when reporting ceramic survival or failure rates. The purpose in this guidance document however, is not to repeat the content of reference papers but to provide a relevant hints on the major help of using fractography when attempting to understand the reason for premature failure of a clinical component. The cause of failure usually goes back to design issues, processing induced damage (manufacturer, lab and clinician) or surface contact damage and material degradation from the mechanically harsh environment of the oral cavity (i.e., peak loading, cyclic loading, friction or abrasive wear, excessive contact pressure, bruxism, multidirectional chewing, off-center loading of implant supported ceramic restorations, water and temperature exposure, slow crack growth, aging). In most cases one would find a combination of the above listed factors involved in the fracture process.
For each clinical failure, maximum of information should be collected before starting the fracture analysis. These include: (1) intra-oral pictures to secure the anatomical orientation of the fracture, (2) crown number in the FDI system, (3) knowledge of the ceramic material (core, veneer), (4) time to failure, (5) circumstances of the fracture event as provided by the patient, (6) retrieval information by the clinician (including cementation/adhesive luting procedures). Further documentation by the fractographer will include stereo microscopy and SEM. The photographic documentation should follow a systematic approach with (1) correct orientation of the entire broken part (occlusal surface on top), (2) labeling of the specimen regarding orientation (mesial, distal, buccal, palatal or lingual, (3) targeting zones of interest for detailed analysis and mapping the direction of crack propagation over the entire fractured part based on the correct identification of characteristic fracture surface features (wake hackle, twist hackle, arrest lines, compression curl and more) within these zones. It is important to give photographic evidence for such mapping. The origin should be thoroughly searched and at least a location identified from which the crack has started. If identified, than at least one higher magnification image of the origin is needed.
Clinical failures are often more complex than lab-scale failures, the latter being rather straightforward with regards to fractographic analysis. Unfortunately, due to an incomplete fractographic approach, some fractures of broken parts may be wrongly identified as an origin because the whole fracture surface was not properly analyzed in terms of direction of crack propagation (dcp). Hence, a secondary fracture event may be confused with a crack origin if only one picture is shown. Fig. 10 a illustrates such a misleading documentation. It corresponds to a partial view of fractured zirconia abutment which was directly veneered for a premolar crown reconstruction and screwed onto a titanium implant. The fracture occurred after 2 years of intra-oral use. The shoulder part of the abutment shows a chip fracture ( Fig. 10 a) which looks like the origin (white arrow) but in fact this is only a secondary fracture event. This can be proven by the very fine hackle lines indicating the main crack direction, as pointed by the black arrows. Looking at a different image further up ( Fig. 10 b), the fracture origin is clearly situated at the internal angle (large black arrow), below the screw head. A large twist hackle is visible emanating out of that corner followed by many fine hackles indicating the dcp ( Fig. 10 c, 800×). Chip damage on that inner side of the corner can be observed in connection with the crack origin. Reasonable speculation can be made that these damages were created during the screw tightening or screw contact while the crown is cyclic loaded during function. This case is further documented in Fig. 11 with the purpose of having an overall robust stereomicroscope documentation in which the various origins, dcp, and secondary chipping fractures are clearly indicated. These secondary events are visible on both mesial and distal shoulders as well as at the area of direct contact with the screw head. The chipping fracture at the screw head contact (slight metal traces visible on the inside) was arrested in the zirconia abutment. The black arrows show the overall dcp. The red arrows indicate multiple fracture origins at the corners. Fig. 11 c,d shows both fractured parts joined together in which the crack pattern is visible as well as the missing chipped ceramic at the shoulder level (secondary event). When available, assembling matching parts is essential in fractography to better understand the fracture as it provides information about the overall crack pattern, crack bifurcation, crack branching, presence of cone cracking and missing parts. In clinical dentistry, this has been very efficiently applied to recovered all-ceramic crowns using only stereo microscopy before adding SEM for more detailed views of the origins located at the crown margins . A reference paper by Quinn et al. has introduced this method to the dental community reassembling broken parts on three whole all-ceramic restorations and successfully describing the failure event, mapping the crack propagation along the fractured surfaces, and identifying the fracture origin. Usually the half showing a better view of the origin is published, but the other half should be also kept as a back-up to confirm the initial findings .
