J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 3 ( 2 0 1 0 ) 1 1 6 – 1 2 0

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/jmbbm

Short communication

Fracture of tooth enamel from incipientmicrostructural defects

Herzl Chaia, James J.-W. Leeb,c,∗, Brian R. Lawnc

a School of Mechanical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv, IsraelbDepartment of Anthropology, George Washington University, Washington, DC 20052, USAcCeramics Division, National Institute of Standards and Technology, Gaithersburg, MD 20899-8520, USA

A R T I C L E I N F O

Article history:

Received 15 May 2009

Received in revised form

21 July 2009

Accepted 3 August 2009

Published online 11 August 2009

A B S T R A C T

We present definitive evidence for crack growth from internal defects called ‘tufts’ in

human enamel. Transverse slices (normal to the tooth axis) sawn from extracted human

teeth are embedded in a polycarbonate sandwich configuration and tested in simple

flexural loading. The evolution of ensuing cracks across the enamel sections is viewed in

situ by a video camera. The observations unequivocally identify tufts as sources of internal

tooth fracture. In sufficiently thin slices the enamel becomes translucent, allowing for

through-thickness observations of the crack topography. Crack segments that appear to be

disjointed on a section surface link up into a contiguous primary crack below the surface,

suggesting some crack resistance by ‘bridging’ behind the advancing crack tip. The role of

these and other microstructural factors in determining the resilience of tooth structures is

considered.

Published by Elsevier Ltd

o

1. Introduction

The remarkable resilience of mammalian teeth has been welldocumented (Janis and Fortelius, 1988; Maas and Dumont,1999; Chai et al., 2009a). Teeth are built to withstand alifetime of heavy function, sustaining repeated biting loadsup to 1000 N. Yet the protective enamel coat is highlybrittle, with a toughness comparable to that of glass (He andSwain, 2008; Xu et al., 1998). Enamel also contains a highdensity of incipient defects or flaws within its microstructure.Chief among these defects are ‘tufts’, hypocalcified, protein-filled fissures extending outward from the dento-enameljunction (DEJ) (Sognnaes, 1949; Osborn, 1969; Ten Cate, 1989;Lucas, 2004). Tufts tend to lie along weak interfaces within

∗ Corresponding author at: Ceramics Division, National Institute of SE-mail addresses: herzl@eng.tau.ac.il (H. Chai), james.lee@nist.g

1751-6161/$ - see front matter. Published by Elsevier Ltddoi:10.1016/j.jmbbm.2009.08.002

tandards and Technology, Gaithersburg, MD 20899-8520, USA.v (J.J.-W. Lee), brian.lawn@nist.gov (B.R. Lawn).

the organic sheaths that delineate enamel prisms. Theyhave a ‘wavy’ appearance (like tufts of grass—hence theirname), associated with undulating and crossing (decussating)prism orientations. Recent studies of tooth sections havehypothesized tufts as elements of weakness within theenamel microstructure, acting as sources for initiation andpropagation of large-scale fractures (Lucas et al., 2008; Chaiet al., 2009a; Lee et al., 2009). Notwithstanding the ubiquitouspresence of these defects, tooth enamel remains highlyeffective in its capacity to protect the soft dentin–pulpinterior. The resilience of tooth enamel is attributable to thecapacity to contain cracks rather than to prevent them (Lawnand Lee, 2009).

However, evidence for the role of tufts as incipient so-urces of crack generation has thus far been indirect, relying

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 3 ( 2 0 1 0 ) 1 1 6 – 1 2 0 117

Specimen

Plate 1

Plate 2Epoxy

Fig. 1 – Schematic of mechanical testing of transversetooth slices. Loading of thin slice embedded in epoxybetween two polycarbonate plates, with specimen on thetensile side of the neutral axis (dashed line). A videocamera records the crack evolution from tufts viewed frombelow in combined reflected and transmitted light.

principally on post-mortem examinations of fractured teeth(Chai et al., 2009a; Lee et al., 2009). Part of the problem restswith the high degree of opacity of tooth enamel, precludingdirect observation of small-scale internal events duringevolution to failure. In this study we present unequivocalevidence for crack growth from tufts, by sawing slices fromextracted human teeth and observing the faces in situ duringloading. We confirm that cracks emanate from tufts at theDEJ and propagate steadily, albeit in a somewhat disjointedmanner, with increasing stress across the enamel section tothe outer surface.

2. Materials and methods

Tests were made on 8 human molar teeth extracted frommale and female patients 18 to 25 years of age. Only thoseteeth with minimal damage from the extraction procedurewere chosen for testing. These were stored in aqueous solu-tion soon after extraction and kept moist as far as possiblethrough all stages of preparation. Transverse slice specimenswere sawn from these specimens for subsequent mechanicaltesting, as shown in Fig. 1. The slices were cut perpendicularto the tooth axis, at prescribed depths below the uppermostcusp, to thickness 400 µm, and again polished to 1 µm sur-face finish. The slabs were placed between two polycarbon-ate plates of unequal thickness, 12.5 mm and 1.5 mm, andbonded to the plates with epoxy. The composite sandwichlayers were then loaded in plate flexure, load axis alignedwith the slice center (Fig. 1). With the encapsulated enamelsegments located below the neutral axis, the flexure resultedin a hoop tensile stress closely normal to the tuft directions. Avideo camera located below the thin polycarbonate plate wasused to view the response of the enamel to increasing stresslevels in mixed reflected–transmitted light. After testing, thespecimens were examined in transmitted light in an opticalmicroscope, to determine subsurface crack topography.

Values of the tensile stresses in each specimen wereestimated by routine finite element analysis using measuredapplied loads and appropriate elastic constants for theenamel/dentin tooth slice and polycarbonate/epoxy layers

(Chai et al., 1999, 2009b). The stresses showed gradients ofsome 20% within the enamel slice, so average values weredetermined.

3. Results

Fig. 2 is a sequence of a transverse slice specimen in thesandwich flexure configuration of Fig. 1. The direction oftensile stress is horizontal in the figure, i.e. closely normal tothe tufts—values of FEA-estimated stresses at each stage ofthe sequence are indicated in the caption. Note the presenceof 3 cracks between the DEJ and outer boundary of the enamelat the onset of loading (Fig. 2(A)), indicating some preexistingdamage from the past history of the tooth. On increasing theload, these preexisting cracks begin to open up and becomemore visible (Fig. 2(B)). At slightly higher load, new cracksextend into the enamel from those tufts at the DEJ marked byarrows (Fig. 2(C)). Note the abrupt jump of the more centralof these cracks at the same load (Fig. 2(D)), with ultimatecompletion of through-width propagation (Fig. 2(E)) shortlyafter. In no instance did we observe cracks propagatingfrom the outer surface inward. Note that the new cracks lieapproximately midway between their neighbors, indicatingthat it is the tufts at locations where stress has undergonegreatest relaxation that provide favored fracture sites.

All 8 transverse slices examined showed the same crackbehavior as in Fig. 2, albeit with some variations in stresslevels and crack densities depending on specimen history.In 5 specimens the newly formed slices were free of anycracks, whereas in the remaining 3 specimens preexistingcracks similar to those in Fig. 2(A) were observed. Thissuggests that cracks could be initiated either during oralfunction or specimen preparation. Regardless, initiationof subsequent cracks always occurred from tufts in ourspecimens. An estimated tensile stress 43 ± 16 MPa atnew crack formation (mean and standard deviation) fromfinite element analyses compares with the reported strength≈30 MPa of free-standing enamel (Lucas, 2004), the scatterpresumably reflecting specimen-to-specimen variation in tuftsize and density.

Enamel segments of the kind shown in Fig. 2 examined byadjusting the focal plane in transmission microscopy revealinteresting features of the crack development in relation tothe tufts. Crack C in Fig. 3 is an example of another through-width crack, with focal plane (A) at the specimen surface and(B) within the enamel slice at depth ≈100 µm. The image inFig. 3(A) shows a disjointed crack at location B, suggestive ofa ‘mother–daughter’ initiation process ahead of the tuft edge(Imbeni et al., 2005). However, the image in Fig. 3(B) showsthe tuft/crack to be fully connected without any discontinuity,indicating a smooth extension process at this level. It wouldtherefore seem likely that the crack in this specimen hasinitiated at some subsurface tuft segment, branching ontoadjacent surfaces as it extends upward and outward, butalways remaining connected on a common, contiguous path.

118 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 3 ( 2 0 1 0 ) 1 1 6 – 1 2 0

A

B

C

D

E

1 mm

Enamel

Dentin

Fig. 2 – Transverse slice embedded between polycarbonateplates of unequal thickness (Fig. 1). Plate flexure places theenamel segment in tension (oriented horizontally in thefigure), as evaluated by FEA: (A) 5 MPa, (B) 45 MPa, (C)48 MPa, (D) 48 MPa, (E) 53 MPa. Sequential loading causescracks to generate at tufts and propagate from DEJ to theouter surface. Note progressive crack propagation of tuftsmarked by arrows in sequence viewed in combinedreflected and transmitted light.

4. Discussion

The results presented in Figs. 2 and 3 provide evidence forcrack initiation and growth from internal defects or tuftson weak inter-prism interfaces at the DEJ. Analogous obser-vations of crack extension from tufts have previously beenreported in longitudinal sections parallel to the tooth axis

300 µm

Enamel

Dentin

C

C

B

B

A

B

Fig. 3 – Transverse slice of the kind shown in Fig. 2, butviewed in a transmission microscope after testing, (A)surface focus and (B) subsurface focus. Tuft C in (A) appearsto be physically separated from the ensuing crack at B.Defocus view in (B) shows tuft and crack to be contiguouson the common internal crack surface.

(Chai et al., 2009a), but the present results provide more com-pelling evidence for such behavior. Certain fracture modes inteeth, such as ‘radial’ cracks beneath the contact and ‘mar-gin’ cracks at the base of the enamel crown, are believedto occur at the enamel–dentin junction, and tufts presentthemselves as prime sources for these fractures (Lucas et al.,2008; Lee et al., 2009). As indicated, some of the preexistingcracks observed in Fig. 2 may be of these types, and may haveoccurred during life or even during specimen preparation.Human teeth, especially in older patients, tend to containmany such cracks or ‘lamellae’ running longitudinally alongthe teeth walls (Sognnaes, 1950; Bodecker, 1953).

Tufts appear to be a common feature of mammalianenamel, not only in humans but also in great apes andother animals with blunt, low-crowned (‘bunodont’) teeth,such as sea otters (Lucas, 2004). The tufts have the characterof closed cracks filled with adhesive interlayers, most likelyoriginating from intrusion of protein-rich fluids duringdevelopment (Sognnaes, 1949; Osborn, 1969; Palamara et al.,1989; Lucas, 2004). Such fluid intrusion might continue tooperate during subsequent function, so that newly formedcracks undergo a continual process of ‘self-repair’ duringaging (Myoung et al., 2009). While the adhesive ‘glue’ provides

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 3 ( 2 0 1 0 ) 1 1 6 – 1 2 0 119

some measure of strengthening to the enamel structure,the tufts nevertheless remain a source of weakness, as hasbeen clearly demonstrated by emplacing Vickers indentationsimmediately adjacent to tuft interfaces—corner cracks fromthe Vickers imprints can occasionally propagate through theinterfaces, but on occasion also can cause the interface todelaminate (Chai et al., 2009a). Tooth enamel is therefore aninherently weak structure, bound with proteinaceous ‘glue’but with an intrinsic population of strength-degrading defectsand cracks. Crack avoidance does not seem to be a strategy oftooth evolution (Lucas et al., 2008; Lawn and Lee, 2009).

Where the tooth gains resilience is in its capacity to re-distribute a largely compressive biting force in such a wayas to minimize tensile stresses within the enamel coat,and to contain the cracks once they start. It is an exam-ple of a highly damage-tolerant structure. Some features ofthe microstructural complexion appear to contribute to thisresilience. It has long been suggested in the literature thatdecussation may provide some resistance to fracture, by forc-ing cracks to deflect and arrest at locations where the prismschange orientation (Koenigswald et al., 1987). The appear-ance of disconnected segments of material along the pathof extending cracks might at first sight seem to suggesta ‘mother–daughter’ mechanism, whereby microcracks arereinitiated ahead of the advancing tip, along surfaces with fa-vorably oriented prisms. However, the observation in Fig. 3that such disjointed segments link up below the surface indi-cates an alternativemechanism, one in which a primary crackfragments along its front but maintains a contiguous sur-face, with bridging segments attached to opposite crack wallsin the wake. Such a mechanism has been well documentedin polycrystalline ceramics (Swanson et al., 1987; Swanson,1988). Bridging elements of this kind have also been reportedin crack propagation tests on dentin from human and ele-phant teeth (Kinney et al., 1999; Kruzic et al., 2003). Quantita-tive determinations of the way toughness may be augmentedin this way so as to generate increasing toughness character-istic with continued extension (R-curve behavior) in enamelare only now being made (Bajaj and Arola, 2009). However,even allowing for such processes, the toughness of enamel re-mains low, barely higher than that of silicate glass (Xu et al.,1998; He and Swain, 2008).

Another element that would appear to contribute toenamel resilience is ‘stress shielding’ (Chai et al., 2009a). Thetufts occur in dense populations, such that their lengths wellexceed the distance between neighbors. The presence of suchneighbors acts to screen any given tuft from the externallyapplies stress field, thereby reducing the stress intensity fac-tor driving crack extension, the more so with continued ex-tension (Chai et al., 2009a). A manifestation of such shieldingis observed in the crack patterns of Fig. 2, in which fully de-veloped through-thickness cracks continue to proliferate withmonotonic stressing, each time initiating approximately mid-way between preceding fractures. The same phenomenon isobserved in the generation of through-thickness ‘channel’ or‘tunnel’ cracks in brittle coatings (Gille and Wetzig, 1983) andlaminated composites (Xia et al., 1993) under tensile loading.

Acknowledgements

Molar teeth were obtained from Gary Schumacher, SabineDickens and Anthony Guiseppetti in the PaffenbargerResearch Center (PRC) laboratories at the National Instituteof Standards and Technology (NIST). Approval to test thesespecimens was granted by the NIST Internal Review Board.

R E F E R E N C E S

Bajaj, D., Arola, D.D., 2009. On the R-curve behavior of humantooth enamel. Biomaterials 30, 4037–4046.

Bodecker, C.F., 1953. Enamel lamellae and their origin. J. Dent. Res.32, 239–245.

Chai, H., Lawn, B.R., Wuttiphan, S., 1999. Fracture modes in brittlecoatings with large interlayermodulusmismatch. J. Mater. Res.14, 3805–3817.

Chai, H., Lee, J.J.-W., Constantino, P.J., Lucas, P.W., Lawn, B.R.,2009a. Remarkable resilience of teeth. Proc. Nat. Acad. Sci. 106,7289–7293.

Chai, H., Lee, J.J.-W., Kwon, J.-Y., Lucas, P., Lawn, B.R., 2009b. Asimple model for enamel fracture from margin cracks. ActaBiomater. 5, 1663–1667.

Gille, G., Wetzig, K., 1983. Investigations on mechanical behaviorof brittle wear-resistant coatings. 1. Experimental results. ThinSolid Films 110, 37–54.

He, L.H., Swain, M.V., 2008. Understanding the mechanical behav-ior of human enamel from its structural and compositionalcharacteristics. Journal of the Mechanical Behavior of Biomed-ical Materials 1, 18–29.

Imbeni, V., Kruzic, J.J., Marshall, G.W., Marshall, S.J., Ritchie, R.O.,2005. The dentin–enamel junction and the fracture of humanteeth. Nature Mat. 4, 229–232.

Janis, C.M., Fortelius, M., 1988. On the means whereby mammalsachieve increased functional durability of their dentitions,with special reference to limiting factors. Biol. Rev. CambridgePhilos. Soc. 63, 197–230.

Kinney, N.H., Balooch, M., Marshall, G.M., Marshall, S.J., 1999. Amicromechanics model of the elastic properties of humandentine. Arch. Oral Biol. 44, 813–822.

Koenigswald, W.v., Rensberger, J.M., Pretzschner, H.U., 1987.Changes in the tooth enamel of early paleocene mammalsallowing increased diet diversity. Nature 328, 159–162.

Kruzic, J.J., Nalla, R.K., Kinney, J.H., Ritchie, R.O., 2003. Crack blunt-ing, crack bridging and resistance-curve fracture mechanics indentin. Biomaterials 24, 5209–5221.

Lawn, B.R., Lee, J.J.-W., 2009. Analysis of fracture and deformationin teeth subjected to occlusal loading. Acta Biomater. 5,2213–2221.

Lee, J.J.-W., Kwon, J.-Y., Chai, H., Lucas, P.W., Thompson, V.P., Lawn,B.R., 2009. Fracture modes in human teeth. J. Dent. Res. 88.

Lucas, P.W., 2004. Dental Functional Morphology: How TeethWork. Cambridge University Press, Cambridge, UK.

Lucas, P.W., Constantino, P.J., Wood, B.A., Lawn, B.R., 2008. Dentalenamel as a dietary indicator in mammals. BioEssays 30,374–385.

Maas, M.C., Dumont, E.R., 1999. Built to last: The structure,function, and evolution of primate dental enamel. Evol.Anthrop. 8, 133–152.

120 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 3 ( 2 0 1 0 ) 1 1 6 – 1 2 0

Myoung, S., Lee, J.J.-W., Constantino, P.J., Lucas, P.W., Chai, H.,Lawn, B.R., 2009. Morphology and Fracture of Enamel. J.Biomech. 42, 1947–1951.

Osborn, J.W., 1969. The 3-dimensional morphology of the tufts inhuman enamel. Acta Anat. 73, 481–495.

Palamara, J., Phakey, P.P., Rachinger, W.A., Orams, H.J., 1989. Theultrastructure of spindles and tufts in human dental enamel.Adv. Dent. Res. 3, 249–257.

Sognnaes, R.F., 1949. The organic elements of enamel. II. Theorganic framework of the internal part of the enamel, withspecial regard to the organic basis for the so-called tufts andschreger bands. J. Dent. Res. 28, 549–557.

Sognnaes, R.F., 1950. The organic elements of enamel. IV. Thegross morphology and the histological relationship of thelamellae to the organic framework of the enamel. J. Dent. Res.29, 260–269.

Swanson, P.L., 1988. Crack-interface traction: A fracture-resistance mechanism in brittle polycrystals. In: Varner, J.,Frechette, V.D. (Eds.), Fractography of Glasses and Ceramics.American Ceramic Society, Columbus, OH, pp. 135–155.

Swanson, P.L., Fairbanks, C.J., Lawn, B.R., Mai, Y.-W., Hockey, B.J.,1987. Crack-interface grain bridging as a fracture resistancemechanism in ceramics: I. Experimental study on alumina.J. Am. Ceram. Soc. 70, 279–289.

Ten Cate, A.R., 1989. Oral Histology: Development, Structure, andFunction. St. Louis, Mosby.

Xia, Z.C., Carr, R.R., Hutchinson, J.W., 1993. Transverse cracking infiber-reinforced brittle matrix, cross-ply laminates. Acta Mater.41, 2365–2376.

Xu, H.H.K., Smith, D.T., Jahanmir, S., Romberg, E., Kelly, J.R.,Thompson, V.P., 1998. Indentation damage and mechanicalproperties of human enamel and dentin. J. Dent. Res. 77,472–480.