| Sign In to gain access to subscriptions and/or personal tools. |
Joining Veneers to Ceramic Cores and Dentition with Adhesive Interlayers
Department of Materials Science and Engineering, University of Maryland, College Park, MD 20742-2115, USA; and Correspondence: * corresponding author, brian.lawn{at}nist.gov
Adhesive joining of veneers to cores offers potential simplicity and economy in the fabrication of all-ceramic crowns. We tested the hypothesis that resin-based adhesives can be used for such fabrication without compromising mechanical integrity of the crown structure. A simple test procedure for quantifying this hypothesis was proposed. A model glass veneer layer 1 mm thick (representative of porcelain), adhesively bonded onto a glass-like core substrate (ceramic or dental enamel), was loaded at its top surface with a hard sphere (occlusal force) until a radial crack initiated at the veneer undersurface. The critical loads for fracture, visually observable in the transparent glass, afforded a measure of the predisposition for the adhesive to cause veneer failure in an occlusal overload. Two adhesives were tested, one a commercial epoxy resin and the other a relatively stiff in-house-developed composite. The results confirmed that stiffer adhesives provide higher resistance to failure.
Key Words: adhesive joining • glass • contact loading • veneer failure
All-ceramic dental crowns generally consist of an aesthetic but weak porcelain veneer layer fused at elevated temperature to a functional and strong core ceramic, with the porcelain applied layer by layer by a dental technician (McLean, 1979, 1983). The process of joining is time- and labor-intensive, and is subject to potentially deleterious residual stresses from mismatch in coefficients of thermal expansion (CTE). An attractive alternative is to fabricate veneer and core individually by an alternative manufacturing route, and then to bond the layers together chemically at low temperatures by means of a polymeric resin-based adhesive. This route avoids CTE stresses (although perhaps not shrinkage stresses), and establishes a soft interfacial barrier for arresting cracks formed in any one layer (Clegg et al., 1990). Onlays and inlays are joined onto prepared dentition by an analogous cementation procedure. The use of resin-based adhesives has its own possible drawbacks, most obviously the introduction of a weak interface, rendering the system liable to delamination and spalling. Perhaps more important, because the interface is relatively compliant, the prospect arises for flexure during occlusal loading and consequent failure of the veneer by crack initiation from a flaw at the bottom surface (Chai and Lawn, 2000). Once initiated, such cracks spread radially outward from the source into an elongate elliptical configuration—hence the term "radial" crack—and eventually traverse the veneer to the top surface.
The challenge, then, is to find and test adhesives that are both stiff enough to minimize veneer flexure and strong enough to resist delamination. The present study pursues this challenge by the use of a model layer system (Fig. 1
Glass plates of thickness d = 1.0 mm (veneer) and 12.5 mm (base) were used as adjoining brittle layers. The thin plates were abraded at their bottom surfaces with 600 SiC grit (Chai et al., 1999). This treatment provided starting flaws of controlled size for the ensuing radial fracture initiation, equivalent to the sandblasting treatments used by dental clinicians, thereby producing a worst-case scenario for veneer strength properties (Zhang et al., 2006). The same plates were etched (10% hydrofluoric acid, 30 sec) at their top surfaces for removal of surface handling flaws, thus precluding spurious cone fractures during testing (Bhowmick et al., 2005). Side walls were polished to allow for direct observation of crack evolution. The veneer and base glass plates were joined by 2 different adhesives. An epoxy resin (Harcos Chemicals, Bellesville, NJ, USA) was used as a simple adhesive with good bonding, well-documented in studies of this kind (Chai et al., 1999). We produced a composite with uncommonly high modulus by loading 72 mass% spherical alumina particles (NanoTek, Nanophase Technologies Corp., Romeoville, IL, USA) of mean diameter 45 nm into a monomer blend of 50 mass% bisphenol-A-glycidyldimethacrylate (bis-GMA, Esstech, Essington, PA, USA) and 50 mass% triethylene glycol dimethacrylate (TEDGMA, Esstech, Essington, PA, USA). The mixtures were stirred for 4 hrs to break up any agglomeration of particles and to produce a uniform microstructure, as confirmed by scanning electron microscopy of sections (Wang et al., 2007). Opposing surfaces of the glass plates were first silanized (3M ESPE RelyX Ceramic Primer, St. Paul, MN, USA) and then joined with the 2 adhesives in their as-mixed forms. Epoxy joins were formed and cured at room temperature for one day. Composite joins were formed at room temperature and heated in an oven at 120°C for at least 6 hrs, to enable the organic matrix to crosslink. The resulting composite was relatively uniform, with alumina filler dispersed throughout the matrix. We obtained adhesive interlayers of prescribed thicknesses h by squeezing the opposing glass plates tight under light pressure (small h) or by inserting spacers between the layers (large h). Adhesive interlayer thicknesses were measured at specimen sections by a micrometer, producing a working range h = 65 µ 650 µm. This covers the range of dental cement thicknesses used in adhering crowns to tooth structure (McLean, 1979). Elastic moduli of the adhesives were measured at the same sections for different values of h by means of a nanoindenter (Oliver and Pharr, 1992), with contact diameter of ca. 50 µm (i.e., much greater than the 45-nm particle scale) (Wang et al., 2007). Means and standard deviations were Ei = 2.3 ± 0.1 GPa for epoxy and Ei = 20.4 ± 0.6 GPa for the filled composite.
Contact with a WC spherical indenter of radius r = 3.18 mm, mounted into a mechanical testing machine, was applied at the specimen top surface (Chai and Lawn, 2000; Bhowmick et al., 2005; Zhang et al., 2005; Hermann et al., 2006). Loading was increased monotonically until radial cracks initiated at the bottom surface of the veneer plate (Fig. 1 Variations in the measurement of experimental variables—load P, layer thickness d, and sphere radius r—were less than 1%, so that uncertainties in the ensuing data could be largely ascribed to material variation.
A side view of a glass/glass specimen bonded with composite adhesive of thickness h = 100 µm and contact-loaded at PR = 600 N revealed a typical radial crack pattern in the veneer layer (Fig. 2  ). Upon further loading, the crack arms expanded sideways and upward, ultimately penetrating through to the top surface. Views from below the base confirmed the incidence of several such radial cracks, at more or less equal angles, forming a characteristic "star" pattern (Chai et al., 1999). The side view in Fig. 2 highlights one such radial crack, approximately normal to the observation direction. In our experiments, no subsidiary fracture modes, e.g., interfacial delamination or top-surface cone cracking, were observed in any specimens up to loads P = 1200 N.
Load PR to produce veneer radial cracks was plotted as a function of adhesive thickness h, for epoxy and filled-composite adhesives (Fig. 3 ). A logarithmic scale was used simply to show the curve shifts for the different adhesives with greater clarity. The data points with error bars are means and standard deviations of 6 tests for each adhesive at one clinically representative thickness, h = 150 µm; the other data points are individual test results in specimens of various adhesive thicknesses. Of primary interest here is a distinct upward shift in the data, by more than a factor of 3, from epoxy to filled-composite adhesive. A standard t test analysis of the datasets at h = 150 µm yields p < 0.000002, so that this shift may be considered significant. Note also that each dataset indicates monotonically decreasing PR with increasing h, amounting to about a factor of 2 decline over the order-of-magnitude range covered. Thus, any decrease in fracture resistance from increase in adhesive thickness can easily be outweighed by a change to a stiffer adhesive.
The data trends (Fig. 3  ) can be quantified by an existing set of equations for the initiation of radial cracks in a veneer plate of modulus Ev and thickness d adhesively bonded to a thick substrate of modulus Es (Kim et al., 2003):
where B = 2 is a constant, S is the strength of the glass, and where we define an effective modulus E* of the combined adhesive/substrate underlayer
and where L is an exponent dependent on adhesive thickness
with
The formulation in Eqs. 1–3 establishes a basis for considering the role of geometric and material variables in veneer failure. As an illustrative example, retaining the same material parameters for glass, plots are shown for PR as a function of adhesive modulus Ei, using logarithmic coordinates, for fixed adhesive thickness h = 100 µm and selected values of veneer thickness d (Fig. 4
We have used a simple contact loading test to quantify the suitability of resin-based adhesives as a means for joining veneers to a base core layer. The test is simple, making use of transparent glass plates in a model layer system for in situ detection of veneer radial cracking. The critical loads PR to initiate such cracks provide a quantitative measure of resistance to veneer failure. In the clinical context, it is necessary to maintain PR >> 100 N, i.e., above the range of typical occlusal biting forces (McLean, 1979; Kelly, 1997, 1999). This condition is accomplished by both test adhesives, over a broad range of join thicknesses h, with an especially comfortable safety margin for the stiff composite. Increasing Ei above 20 GPa overcame any degradation from excessive adhesive thickness. The test also verified the bonding capacity of the adhesive. Recall that, in our tests, no debonding was observed up to PR = 1200 N. Contrast this to results from an earlier study on some commercial dental cements, where the layers delaminated catastrophically at low loads, indicating a total unsuitability of those cements for veneer/core joining (Kim et al., 2003). The present results therefore suggest that resin-based adhesives may provide a convenient means of joining brittle veneers to underlying ceramic cores (crowns) or tooth enamel (onlays and inlays). Potential advantages include: avoidance of CTE stresses (although care may be needed to avoid shrinkage stresses during polymeric curing); and arrest of veneer (or core) cracks at the soft/tough interlayer, thereby inhibiting spread of cracks from one layer to another. The adhesive needs to meet certain requirements:
We have demonstrated how Eqs. 1–3 can be used to elucidate the dependence of critical load PR for veneer fracture on adhesive modulus Ei and veneer thickness d. We may further note the appearance of strength S in Eq. 1, which suggests that control of the surface flaw state of the veneer is an important factor in veneer preparation (Zhang et al., 2006). Dummy tests on as-polished glass veneers show critical loads considerably higher than those with surface abrasion, it is reported here (Chai et al., 1999). Last, sphere radius r is a relatively unimportant parameter in our tests, because radial crack initiation occurs at the veneer undersurface remote from the contact zone—necessary only to ensure that r does not become too small, to avoid spurious cone cracking at the intensified Hertzian contact. While it is experimentally expedient to construct the test layer system entirely from glass, for the reasons outlined in the INTRODUCTION, Eq. 2 may, in principle, be expanded to cover the case of dissimilar veneer and substrate materials (Kim et al., 2003). However, the formulation in Eq. 2 tends to overestimate E* in the region of high Es, so that veneer failures remain a threat with even the stiffest core ceramics. An all-glass system usefully provides a lower bound to PR values. Finally, data such as those presented here may serve one more useful purpose, in cases where it becomes difficult to measure Ei of ultra-thin interlayers, and where it is suspected that Ei might differ from the modulus of the bulk material (e.g., by redistribution of the particulate density during joining). Given knowledge of the glass properties, along with the adhesive thickness, one may, in principle, ’deconvolute’ the modulus Ei numerically from the data by using Eqs. 1–3.
Discussions with Herzl Chai at Tel Aviv University on the formulations in Eqs. 1 to 3, and with Elaine Romberg at the University of Maryland at Baltimore on statistical analysis of the data, are gratefully acknowledged. This work was supported by a grant from the US National Institute of Dental and Craniofacial Research (PO1 DE10976). Certain equipment, instruments, or materials are identified in this paper to specify experimental details adequately. Such identification does not imply recommendation by the National Institute of Standards and Technology, nor does it imply that the materials are necessarily the best available for the purpose. Received for publication August 28, 2006. Revision received March 8, 2007. Accepted for publication April 15, 2007.
Journal of Dental Research, Vol. 86, No. 8,
745-748 (2007)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
= 1.18, β = 0.33, and 
= 3.13. These equations are based on an analysis of flexure of a contact-loaded brittle plate (veneer) on a compliant interlayer (adhesive), and assume that radial fracture occurs when the maximum tensile stress at the bottom surface of the plate reaches the strength S of the material (