X-ray microdiffraction, TEM characterization and texture analysis of human dentin and enamel

Authors


Correspondence to: Dr. Wei Li, State Key Laboratory of Oral Diseases, West China Hospital of Stomatology, Sichuan University, No. 14 South Renmin Road, Chengdu 610041, P.R. China. Tel: +86-28-85501317; fax: +86-28-85582167; e-mail: leewei2000@sina.com

Summary

Human tooth is a complex bioceramic composite, which consists of enamel, dentin and the interface, the dentin–enamel junction (DEJ). The crystal properties and ultrastructure of the inorganic phase through the thickness of healthy human molar teeth were investigated using X-ray microdiffraction (μXRD), electron diffraction and transmission electron microscopy (TEM) techniques. The XRD data were analysed using the Le Bail profile fitting approach. The size and the texture of the crystallites forming enamel and dentin in the crown part of teeth were measured using both techniques and then compared. Results showed that the thickness of dentin crystallites was found to decrease towards the DEJ, whereas the thickness of the enamel crystallites increased from the DEJ towards the outer layers. It was demonstrated that enamel exhibited an increase of texture in 002 lattice planes from the DEJ towards the outer layers. Texture was also detected in 102 lattice planes. The texture effect in 002 planes at the scale of less than 1 μm was also demonstrated in dentin. The variation of lattice parameters as a function of the position within the thickness of dentin and enamel was also observed. The values of the nonuniform microstrain in the dentin and enamel crystallites were from 1.40 × 10−6% to 4.44 × 10−5%. The good correlation between XRD and TEM indicated that μXRD is a useful technique to study crystallography and microstructure of heterogeneous enamel and dentin. The observed gradient characteristics of texture and crystallite size in enamel and dentin maybe an evolutionary outcome to resist wear and fracture, thereby contributing to the excellent mechanical properties of teeth.

Introduction

Human teeth consist of two mineralized layers: highly mineralized enamel coating and the underlying softer dentin interior. Both of these layers protect the pulp. The principal mechanism of protection in teeth is by stress shielding, whereby the comparatively stiff enamel coat supports the bulk of occlusal loading in dental function. Dentin, by contrast, is tougher and resists the entry of cracks propagating through the enamel layer (Imbeni et al., 2005). Considerable research has been carried out to model bi-layer systems that capture the essential macroscale characteristics of hard coating/soft substrate structures (Lawn et al., 2007) as well as attempts to incorporate the role of specimen surface curvature (dome-shaped specimens) to take one step closer to real tooth geometry (Qasim et al., 2005), which is of high interest for their relevance to design of dental crowns (Jung et al., 1999) and other biomechanical implant structures (Lawn, 2002).

Enamel is a complex biomaterial, a composite of elongated mineral crystallites in the form of biological apatite (90% vol.) bonded by polymeric proteins and peptide chains (2% vol.) saturated with water (8% vol.) (Fincham et al., 1999). It has a hierarchical structure with crystals of hexagonal cross-section tightly packed into rods (prisms) (width ∼5 μm) (Reyes-Gasga et al., 2012; Macho et al., 2003) and enclosed by polymeric sheaths (width 0.1 μm) (Fincham et al., 1999). The rods are highly aligned, approximately normal to the outer enamel surface but with some variation in emergent angle around the tooth surface (Lawn et al., 2009). Enamel is stiffer along the long axis of the rods and weaker between rods, where sliding and separation may occur. Also the decussation (crossing) of the rods is observed in human teeth in the region immediately adjacent to the dentin–enamel junction (DEJ), which is thought to confer toughness by inhibiting propagation of cracks (Lucas et al., 2008).

Dentin is a mineralized connective tissue that constitutes the bulk of the tooth. It is intimately related to the dental pulp, with which it shares the same embryological origin from dental papilla (Arana-Chavez & Massa, 2004). Dentin forms the hard tissue portion of the dentin–pulp complex, whereas the dental pulp is the living, soft connective tissue that retains the vitality of dentin. Dentin is a hydrated structure, composed of a mineral phase in the form of needle- and plate-shaped biological calcium-deficient, carbon-rich apatite crystallites (48% vol.). The organic component is in the form of predominantly type I collagen and a minor presence of other proteins (29% vol.). Also dentin is hydrated with fluid (which is similar to plasma) to about 23% by volume (Nanci, 2008). Dentin contains multiple closely packed dentinal tubules of about 2 μm in diameter oriented between the pulp and dentin–enamel junction. Each tubule is surrounded by a cuff of highly mineralized peritubular dentin and intertubular dentin between them (Gotliv & Veis, 2007). The main structural components of dentin are type I collagen fibrils impregnated and surrounded by mineral crystallites, which have c-axes aligned with the collagen fibril axis (Linde, 1989). The interactions between collagen and nano-size mineral crystallites give rise to the stiffness of the dentin structure.

Both enamel and dentin demonstrate hierarchical structure at the nano- and micro-scales with organic and inorganic components organized in substructures optimized by evolution to withstand environmental stresses associated with their biological function (Xie et al., 2008). Although dental restorative composite materials technology has succeeded in creating highly specialized, high performance materials, it is still far from being able to replicate the elegance of biological composites, such as enamel and dentin. Therefore dental restorative procedures are generally temporary with a lifespan of only few years before the restorative material or supporting tooth structure fails. From this perspective new generations of synthetic restorative materials are needed better resembling the ultrastructure and mechanical properties of normal enamel and dentin. The prerequisite for producing such biomimetic materials with properties close to biological composites is to understand the influence of the geometry of the ultrastructural elements, such as size and preferred orientation (texture), of enamel and dentin nanocrystallites on the overall mechanical properties. Therefore a systematic investigation of tooth structures is needed to understand this correlation.

X-ray diffraction (XRD) is one of the most powerful techniques to study the hierarchical structure of biological mineral composites such as teeth and bone. The traditional technique applied to study dental structures was powder XRD, which required a considerable amount of sample powder to generate diffraction patterns. This constraint makes the analysis of small specific areas of a specimen unachievable. Another important constraint was that it was a destructive technique because a sample must be ground into a fine powder, which resulted in the loss of structural information such as heterogeneous distribution in composite, phases, stresses and texture (He, 2004). Furthermore, powder XRD was not effective in the identification of small quantities of secondary minerals (<5%) (Flemming et al., 2005). Both enamel and dentin are heterogeneous structures and it has long been recognized that such structures should be investigated in situ by a high-resolution position-sensitive method in order to relate measurements to the exact position. With that goal in mind, micro-beam techniques have been applied to study dental structures but the resolution of such instruments proved to be insufficient for reliable structural analysis, particularly when dealing with samples with relatively large grain sizes, inhomogeneous phase distribution and preferred orientation (Omnell et al., 1960; Bergman & Lind, 1966; Angmar-Månsson, 1971; Endo et al., 1989). The development of micro-XRD (μXRD) instruments has overcome many of these constraints and it is now possible to apply this nondestructive technique to study selected areas of specimens down to 50 μm in diameter in situ (He, 2004; Friedel et al., 2005; Zioupos & Rogers, 2006). In our past research, we used high-resolution μXRD to analyze tooth enamel and dentin. We tested two types of μXRD equipments, Panalytical X'pert Pro and Bruker D8 DISCOVER, the results indicated that both equipments could be used for tooth in situ μXRD analysis and got meaningful results. However, Bruker D8 equipment is better than X'pert Pro equipment for in situ tooth analysis (Xue et al., 2008a,b). In this study, we will use Bruker D8 equipment for μXRD test.

The purpose of this study was to investigate the crystal structure and texture of enamel and dentin through the thickness of the crown of human molar teeth with μXRD, transmission electron microscopy (TEM) and electron diffraction techniques. Such ultrastructural characterization may in turn provide insights into further improvements of dental restorative biomimetic materials.

Materials and methods

Human molar noncarious teeth (four specimens from 12 to 17-year-old subjects) were extracted for orthodontic reasons and collected for this study according to the protocols approved by the Ethics Review Committee of Sydney South West Area Health Service, reference №X07–0217 & 07/RPAH/47 and by the Human Research Ethics Committee of the University of Sydney, reference №10541.

The specimens were fixed in Modified Karnovsky's fixative in 0.2 M Sorenson's buffer titrated to pH 7.2 for 24 hours at 2°C and then rinsed several times with 0.2 M Sorenson's buffer. The teeth were sectioned at the cement–dentin junction to separate crowns and then 1-mm-thick sections were prepared from the crowns along mesio-distal axis using a low-speed diamond saw (Isomet, Buehler Ltd., Lake Bluff, USA) under constant water irrigation. A total of six points, forming a straight line, were selected for analysis: areas 1–3 were located within dentin (point 3 was at the dentin–enamel junction) and areas 4–6 were located within enamel (Fig. 1).

Figure 1.

Dark field optical image of enamel, DEJ (a), mantle (b) and circumpulpal dentin. The light contrast lines in dentin perpendicular DEJ are tubules (c). The light contrast thick lines protruding from DEJ into the thickness of enamel are tufts (d). The areas selected within dentin and enamel for μXRD and TEM experiments are schematically shown.

X-ray microdiffraction

The measurements were performed by XRD (Bruker AXS D8 Discover) equipped with the Eulerian 1/4 cradle HI-STAR (2D-Detector), fine collimator with 50 μm diameter exposed spot and GADDS (general area detector diffraction solution) software. The X-ray beam (Cu Kα line,λ = 1.5418 Ǻ) was generated by 2.2 kW X-ray tube and Göbel mirror optics. The 2D position sensitive detector had 1024 × 1024 pixels with a 5 cm × 5 cm beryllium window. The tooth section was fixed to the sample holder. The sample to detector distance was 15.1 cm, 2θ scanning angle range was from 8° to 102°, the step-size was 0.02° and the counting time was 1800 s. All measurements were performed at room temperature (290 ± 0.1 K). The analysis spots were determined using an optical microscope equipped with a laser marker. The instrumental peak broadening of the XRD instrument was determined by scanning a Lanthanum hexaboride (LaB6) reference sample. Rietveld refinement and analysis was completed with Rietica software package (Hunter, 1998, http://www.rietica.org), the XRD lines were identified by comparing the measured patterns to the JCPDS (Joint Committee on Powder Diffraction Standards) data cards (International Centre for Diffraction Data, 2005). The obtained XRD patterns were refined using Rietveld method with Le Bail algorithm (Le Bail, 2005).

The crystallite size was calculated from physical broadening of Bragg reflection peaks using Scherrer equation (peak shapes represented by a Gauss function) (Zhang et al., 2003):

display math(1)

where thkl was the averaged crystalline dimension perpendicular to the reflecting planes, λ was the wave-length of the X-ray radiation, KG ≈ 0.94, broadening Bhkl was peak full-width at half-max (FWHM) and θhkl was the centroid peak position. Also Scherrer-Wilson approximation was employed to calculate grain size (represented by a Lorentzian function) and microstrain (Gaussian function) (Klug & Alexander, 1974):

display math(2)

where math formula was the root-mean-square of the lattice strain in the hkl direction and math formula, where Ip was the peak intensity, I(2θ) was the intensity at 2θhkl and KL ≈ 0.9. Plotting math formula versus math formula, the root-mean-square of the lattice strain was extracted from the intercept with the ordinate and the crystallites size thkl was determined from the slope.

It should be noted that the size of the crystallites measured using XRD techniques are the volume-averaged size of the crystalline domains and not actual geometrical characteristics of the particles forming mineral phase of dentin and enamel.

The crystallites preferred orientation (texture) indices (Rhkl) in dentin and enamel were calculated using the following intensity ratios:

display math(3)

where the coefficient math formula was calculated using intensities according to the JCPDS card 09-0432 (International Centre for Diffraction Data, 2005) for a random hydroxyapatite powder mixture (k002 = 0.42 and k300 = 0.55), I211 intensity was corresponding to the observed 211 reflections and Ihkl were the intensities corresponding 002 and 300 reflections. For R≈1, the grains were considered randomly oriented. R values are greater or lower than 1 indicated the presence of preferred orientation of the crystallites in that crystallographic direction (Low, 2004).

Transmission electron microscopy

Using a low-speed diamond saw the dentin specimens were cut into smaller blocks of approximately 1 × 1 × 1 mm representing areas 1 and 2 (Fig. 1) and then rinsed several times with 0.2 M Sorenson's buffer. The specimens were dehydrated in graded alcohols, infiltrated with the epoxy resin for 24 hours and then embedded into epoxy resin. Thin, 80-nm-thick, sections were cut with a diamond knife (Diatome, Bienne, Switzerland) and mounted onto 200 mesh copper grids (Ted Pella, Inc.).

To prepare ultrathin sections at the DEJ and from normal undecalcified enamel, ion-beam thinning technique was used to reduce the thickness to electron transparency in TEM (Fig. 1, areas 3–6). Pre-thinning was completed with tripod wedge polishing technique using diamond-impregnated plastic film grits of 30 μm, 9 μm, 6 μm, 3 μm and final polish was with 1 μm diamond film. Further thinning of the specimens was completed with ion-beam thinning using a cold stage (liquid nitrogen) to minimize damage to the specimens. During milling of the cross-sectional specimens (Fig. 1, area 3), shields were used to protect the interface and to reduce the effect of different milling rates of enamel and dentin.

The prepared specimens were then carbon coated and observed in a CM12 (Philips, Eindhoven, The Netherlands) TEM operated at 120 kV equipped with a nitrogen-cooled anticontamination device.

The width of at least 40 crystallites was measured for each area from bright-field TEM images following the procedure described in (Zavgorodniy et al., 2008b). The means and standard deviations of these data were calculated. Statistical differences were examined using a two-tail unpaired t-test and the corresponding p values reported (p < 0.05 were considered statistically significant).

Results

Lattice parameters of the crystallites in dentin and enamel determined by Le Bail method (Le Bail, 2005) are presented in Table 1. It can be noticed that the lattice parameters vary depending on the position within the tooth structure.

Table 1. Refined structural parameters of dentin and enamel crystallites. Space group P63/m, α = β = 90°, γ = 120° remained unchanged. Profile residues (Rp) and weighted profile residues (Rwp) are presented
Test pointa (Å)c (Å)V (Å)Rp (%)Rwp (%)
19.43776.8830530.945.637.34
29.41396.8593526.445.236.78
39.42696.8642528.277.5210.02
49.43136.8671528.996.008.46
59.42856.8697528.885.457.54
69.43596.8699529.725.417.12

The analysis of XRD profiles of dentin (Fig. 2A) showed that the physical broadening of peaks 211 and 300 were decreasing from the circumpulpal dentin towards DEJ, which in turn suggested the increased crystallinity of dentin towards DEJ. Peak intensity corresponding to 002 reflections was decreasing towards DEJ, whereas the intensity of the 211 peaks remained the same. Depth profile 2-D XRD GADDS frames of dentin (Fig. 2, inset A1) showed Debye-Scherrer rings at lower 2θ angles (8–42°), which were continuous. In area 1 (Fig. 1) the diffraction rings corresponding to 112, 211 and 300 planes were overlapped and formed a broad intense ring, whereas in areas 2 and 3 (Fig. 1) the diffraction rings 112, 211 and 300 were not overlapped. The 210 rings became more intense (about two times intenser) towards DEJ. At higher 2θ angles (42–102°) the contrast of the Debye-Scherrer rings was weak.

Figure 2.

Refined XRD profiles of dentin (A) and enamel (B). The observed data points are represented by the crosses, the calculated plot is represented by the solid line. Below the profile is the difference plot and the vertical lines represent the peak position markers for hydroxyapatite phase. (A) XRD profile of circumpulpal dentin (area 2, Fig. 1). The corresponding Debye-Scherrer rings were shown in inset A1. (B) XRD profile of enamel (area 4, Fig. 1). The corresponding Debye-Scherrer rings were shown in inset B1.

XRD peaks observed in enamel (Fig. 2B) were sharper than the corresponding peaks in dentin (Fig. 2A). The line broadening of the 112, 211 and 300 peaks remained the same across the thickness of enamel, whereas the peak intensity of 112, 211 and 300 were decreasing from inner enamel towards outer enamel. In enamel almost all of the Debye-Scherrer rings (Fig. 2B, inset B1) had better contrast than the corresponding rings observed in dentin. The discontinuous rings corresponding to 002 and 102 were observed, which suggested the preferred orientation in these planes. The most intense discontinuous rings were 002, whereas 102 rings appeared weak in areas 4 and 5 (Fig. 1) and was not detected in the outer layer of enamel (area 6, Fig. 1). The X-ray spectra and Debye-Scherrer ring spectra of all 6 test points were shown in Figure S1 and Figure S2 in online supporting materials.

The mean sizes of the crystallites were calculated according to Eqs. (1) and (2) and measured through the direct observations in TEM with statistical analysis (Table 2).

Table 2. Mineral crystallite widths (200 reflections) and length (002 reflections) in dentin and enamel calculated from XRD data and TEM. Microstrain values in dentin and enamel mineral crystallites calculated from XRD data
 200 reflections002 reflectionsMicrostrain (%)
   TEM (nm)    
 ScherrerScherrer-Wilson   pScherrerScherrer-WilsonCross-Long
Test pointequation (nm)equation (nm)MeanSDComparisonvalueequation (nm)equation (nm)sectiondimension
111.6010.807.001.28Point 1 vs. 20.07112.0417.409.12 × 10−65.40 × 10−6
219.0516.206.521.34Point 1 vs. 3<0.000119.0350.204.44 × 10−57.76 × 10−6
316.7413.706.001.32Point 2 vs. 30.03621.2350.403.92 × 10−51.95 × 10−6
421.5118.2027.015.86Point 4 vs. 50.00721.8622.205.60 × 10−61.40 × 10−6
524.4120.9037.8412.77Point 4 vs. 60.000124.9724.907.83 × 10−62.95 × 10−6
622.8320.9041.429.20Point 5 vs. 60.4823.1121.303.55 × 10−61.69 × 10−6

Table 2 demonstrated variability of the crystallites sizes through the thickness of dentin and enamel. Generally the crystallites were becoming thinner and longer from the middle circumpulpal dentin (areas 1 and 2, Fig. 1) to the outer layer of circumpulpal dentin and further towards DEJ (area 3, Fig. 1). The lengths of the dentin crystallites near DEJ calculated with Scherrer method (Eq. (1)) appeared to have similar size to the crystals forming the outer layer of the circumpulpul dentin, whereas calculations based on Scherrer-Wilson equation (Eq. (2)) showed a substantial increase in length of grain particles embedded into dentin crystallites at the DEJ in comparison to the crystallites in the middle circumpulpal dentin.

The average width of enamel crystals was substantially increasing through the thickness of enamel towards the superficial layer (200 reflections, Table 2), whereas their length was increasing towards the outer layer of enamel (002 reflections, Table 2).

TEM was also employed to determine and compare the width of dentin and enamel crystallites. The width of the crystallites forming the inorganic phase of dentin (Fig. 1) was decreasing towards DEJ (Table 2). It could also be noticed that the crystallites in circumpulpul dentin (Fig. 3A) demonstrated a decrease of length towards DEJ, whereas mantle dentin (Fig. 3B) showed a significant increase in the crystallites length in comparison with the circumpulpul dentin crystallites (Fig. 3A).

Figure 3.

TEM images of circumpulpal (A) and mantle (B) dentin. Mineral crystallites were shown with arrows.

The mineral phase of mantle dentin located 4 μm underneath the DEJ observed in TEM appeared in a form of a porous reticulate texture (Fig. 4A) with numerous needle-like crystallites (Fig. 3B), which were significantly longer than those observed in circumpulpal dentin. Indexing of the corresponding selected area electron diffraction (SAED) pattern confirmed that the crystallites were apatite phase (Fig. 4B). The ring pattern of the SAED confirmed that dentin near DEJ was a polycrystalline material. The preferred orientation of the crystallites in the 002 planes was observed at the characteristic scale of 1 μm. The preferred orientation at the scales greater than 1 μm appeared to be lost and the crystallites were oriented randomly in all directions. This texture was in contrast to the circumpulpal intertubular dentin found deeper in the crown (areas 1 and 2, Fig. 1) where the preferred orientation of the crystallites was also random at the scales greater than 1 μm but not in all directions and only in the incremental growth planes orthogonal to the tubules. No peritubular dentin was observed in mantle dentin. Needle-like crystallites were tightly packed together forming dark contrast structures with lighter contrast material between them. Although the observed crystallites were predominantly straight, the bending of the crystallites around the circumference of the voids was observed (Figs 3B and 4, arrows). Arrows indicated bending of crystallites around the voids. Indexing of the corresponding SAED pattern (Fig. 4B) confirmed the presence of apatite and a highly polycrystalline structure.

Figure 4.

(inset A) Transitional dentin at 4 μm below DEJ. Arrows indicate bending of crystallites around the voids. Indexing of the corresponding SAED pattern (inset B) confirms the presence of apatite and a highly polycrystalline structure at the characteristic scale of 1 μm.

The underlying ultrastructure of enamel was revealed in the TEM observations (Fig. 5A). The enamel mineral phase consisted of needle-like crystals, which were tightly packed with gaps between them of less than 5 nm. Rings formed the SAED pattern of enamel (Fig. 5B), which was consistent with its polycrystallinity. Indexing confirmed the apatitic nature of the observed crystallites. The enamel crystals preferred orientation was along c-direction, long axis of a needle-like apatite crystal (Fig. 5A), which was confirmed by the slightly spread spot structure of the enamel SAED pattern formed by 002 reflections (Fig. 5B).

Figure 5.

(inset A) Enamel crystallites forming prisms (1) and (2) at the interprismatic junction and the corresponding SAED pattern (inset B) with indexing confirming the presence of apatite phase with preferred orientation in c-direction of the crystallites. Arrows point to (002) reflections of the prisms (1) and (2).

The enamel crystals were oriented in nearly the same direction within the prism but demonstrated variation in the directions, when compared between the prisms. Figure 5 showed two enamel prisms and the interprismatic junction. The crystals in the enamel prism (1) were oriented at approximately 37° to the prism (2) (Fig. 5A). The slight variation of the crystallites orientation within an enamel prism of less than 1.5° was due to a slight orientation variation of the crystallites near the boundaries of the enamel prisms. This orientation variation was clearly observed at the enamel SAED patterns, e.g. the slight spread of 002 reflections at (1) and (2) (Fig. 5B). The preferred orientation in 102 planes was also visible (Fig. 5B). The width of the enamel crystallites was increasing towards the superficial layer (Table 2).

Microstrain calculated from 002 and 200 reflections of X-ray profiles was tensile for all crystallites in dentin and enamel. The strain within the cross section of the crystallites was significantly increasing from the middle circumpulpal dentin towards DEJ (Microstrain, Table 2). The strain within the long axes of the crystallites was changing relatively insignificantly from the middle circumpulpal dentin towards the outer layer of circumpulpal dentin but notably decreased in the mantle area of dentin at DEJ.

Microstrain within the cross-section of the enamel crystals remained similar through the thickness of the enamel layer. The microstrain within the long axes of the enamel crystals also remained unchanged through the thickness of the enamel layer and were comparable to the values of dentin crystallites in the mantle layer of dentin (area 3, Fig. 1).

Texture indices of dentin and enamel polycrystalline structures were calculated according to the Eq. (3) (Fig. 6). Texture indices in enamel corresponding to the reflections 002 were significantly greater than 1, which indicated the preferred orientation of crystallites in these planes. Texture indices corresponding to the reflections 300 calculated for both dentin and enamel were close to 1, which indicated random orientation of crystallites in these planes.

Figure 6.

Texture index variation in dentin and enamel. Texture indices were calculated using 002 reflections, which correspond to the crystallites orientation along their long axis. Texture indices calculated using 300 reflections correspond to the crystallites orientation along their short axis. The solid curves are the spline of the experimental data.

Discussion

The structure of enamel is designed to masticate nutrients placed in the oral cavity. To fulfill this function it must exhibit wear and fracture resistance (Xie et al., 2008). The increased preferred orientation and thickness of the crystal grains towards the outer surface of enamel (Fig. 6), which are approximately perpendicular to the enamel surface, is designed to reduce wear. Also the observed increasing degree of the thickness of the enamel crystallites fits within the previous observations of the increased mineralization of enamel towards its superficial layer. On the other hand, interprismatic cleavages (Fig. 5A) may not only limit crack propagation but may also allow limited deformation. The varying degrees of enamel anisotropy (Fig. 6) may help to direct stresses from the geometrically complex enamel occlusal surfaces to the resilient underlying dentin.

It has been demonstrated that enamel apatite crystals are at least 100 μm long, which is inline with the hypothesis that the crystallites are continuous from DEJ to the outer surface of the tooth (Daculsi et al., 1984). Therefore the increasing degree of the enamel hardness (Cuy et al., 2002) towards the enamel occlusal surface is unlikely to be attributed to an increase in the density of crystallites but to the observed increase in the width of the crystallites.

The ultrastructural observations that dentin crystals were becoming smaller towards the outer layers of circumpulpal dentin and then further appeared thinner but longer in mantle dentin may possibly be attributed to the special role of the DEJ and outer layers of dentin in transferring mastication and other types of stresses through the buccal and lingual sides of enamel into coronal dentin through mid-way between the cusp tip and the cervical margin onto root dentin (Yettram et al., 1976; Goel et al., 1991). It was demonstrated that the coronal dentin zone of 200–300 μm adjacent to the DEJ has a compression elastic modulus in the range of 1–12 GPa measured on fully hydrated teeth by means of noncontact laser-speckle interferometer (ESPI) (Zaslansky et al., 2006). This value is significantly lower than the value of stiffness measured with resonant ultrasound spectroscopy (RUS) for a hydrated bulk crown dentin of 24.4 GPa (Kinney et al., 2004), by an atomic force microscope (AFM) of fully hydrated dentin specimens, 22.8–24.5 GPa (Habelitz et al., 2002), or measured with nanoindentation technique, 24.8 GPa (Fong et al., 2000). Mechanical deformation and the in-plain strain distribution mapping of tooth slices by the Moiré fringe interferometry technique demonstrated that most strain was concentrated in this softer 200 μm dentin layer adjacent to the DEJ and it was suggested that this zone functions as a cushion, which allows enamel and dentin work together (Wang & Weiner, 1997). This zone in coronal outer layer of dentin was described as a broad transition zone, or ‘interphase’, which also contributes to the bonding between enamel and dentin (Zaslansky et al., 2006).

Nano-size crystallites generally demonstrate relatively high surface energy, which increases with the reduction of the size of the crystallites and therefore results in the presence of uniform and nonuniform microstrains in the particles. Uniform microstrain is characterised by shifts in the crystal interplanar lattice spacings to the lower or higher side, which depends upon whether the strain is compressive or tensile. When the strain changes from one part of the grain to another, it is called nonuniform strain. This variation of the microstrain within one crystallite leads to the further increase in the broadening of the X-ray lines (Klug & Alexander, 1974). The variation of microstrain in the cross-section of the crystallites resulted in the cross-sectional microstrain, whereas the variation of microstrain in c-direction resulted in the microstrain along the long axes of the crystallites. The variation of microstrain of crystal grains forming dentin and enamel is expected to depend on the surface to volume ratio of the crystallites (∼1/width), i.e. with the increase of the ratio the microstrain increases. This variation of microstrain was particularly significant in the dentin crystallites due to the substantial change of the ratio through the thickness of dentin towards the DEJ (Table 2). Enamel crystals, however, indicated only a trend of the increase of microstrain with the increase of the ratio probably due to the fact that the ratio change was significantly smaller than that observed in dentin. The values of the microstrain in all instances were relatively low and were corresponding to stresses of less than few megapascal.

It was interesting to notice the dependence of a- and c-lattice parameters of the position of measurement through the thickness of dentin and enamel (Table 1). Such variation has been observed in dentin (Zioupos & Rogers, 2006) and enamel (Al-Jawad et al., 2007). These observations maybe interpreted in terms of the variation of the chemical composition of crystals through the thickness of dentin and enamel. It has been reported that biological apatite crystals in dental structures are calcium-deficient and carbon-rich hydroxyapatite (Marshall Jr. et al., 1997) with substitution of (CO3)2− for (PO4)3− group leading to a lower Ca/P ratio than that of stoichiometric hydroxyapatite (Zavgorodniy et al., 2008a). This kind of substitution (B-substitution) leads to the change of both a- and c-lattice parameters. However, other compositional changes, including the ionic substitutions to maintain the charge balance for the carbonate-phosphate exchange (Zioupos & Rogers, 2006; Zavgorodniy et al., 2008a), may affect a- and c-lattice parameters.

Conclusions

X-ray microdiffraction is a useful technique to study cryst-allography and microstructure of heterogeneous structures such as teeth. Using μXRD and TEM techniques it was demonstrated that the crystallites in mantle dentin appeared longer than in circumpulpal dentin; nonuniform microstrain was insignificant in both dentin and enamel crystal grains; μXRD and TEM techniques demonstrated texture in (002) and (102) planes in the enamel crystallites, and texture effect was increasing in 002 planes from the DEJ towards outer layers of enamel. Scherrer method applied to enamel demonstrated results, which were partly in agreement with TEM, whereas Scherrer-Wilson method was in better agreement with TEM results in dentin, which was discussed in terms of effect of microstrain in crystallites.

Acknowledgements

This work was supported by Research Fund for the Doctoral Program of Higher Education of China (No. 20100181120056). The first author also acknowledges China Scholarship Council (CSC) for financial support for this cooperation research. The second author acknowledges a scholarship from the South West Sydney Area Health Service as well as partial support from the Australian Dental Research Foundation (ADRF) and Australian Dental Industry Association (ADIA). The authors also acknowledge the facilities and technical assistance from staff in Electron Microscope Unit, the University of Sydney, Sydney Dental Hospital and Bruker AXS Company.

Ancillary