Effect of specimen gripping device, geometry and fixation method on microtensile bond strength, failure mode and stress distribution: Laboratory and finite element analyses

Abstract

Objectives

Innumerous modifications have been proposed for the microtensile test since its introduction; however, testing parameters are not often well described and wide variations in bond strength are commonly reported. The aim of this study was to evaluate the effect of the test specimen’s gripping device, specimen geometry and fixation method on microtensile bond strength, failure mode, and stress distribution when using an etch-and-rinse 2-step adhesive system bonded to human dentin.

Methods

Resin-based composite bonded to occlusal dentin from 21 human molars was used to fabricate dumbbell- and stick-shaped test specimens which were divided into three groups: Di – dumbbell-specimens placed in a Dircks device; GeS – stick-specimens gripped in a Geraldeli’s device with Superglue; GeZ – stick-specimens gripped in a Geraldeli’s device with Zapit. Specimens were tested to failure in tensile mode and the failure mode was examined under stereomicroscopy and fracture initiation sites were verified by scanning electron microscopy and energy dispersive X-ray spectroscopy. Three-dimensional models of each device/specimen were created and finite element calculations were performed.

Results

The effect of the gripping devices on the bond strength was not significant, unless the bond test areas were normalized. The failure mode was influenced by the type of device. Dircks device was less sensitive to human error than Geraldeli’s, and produced a more uniform stress distribution at the dumbbell specimen adhesive layer than did the Geraldeli’s device at the stick layer.

Significance

Microtensile testing parameters can directly influence the results and consequently inter-study comparisons.

Introduction

Conventional tensile and shear tests have been used for a long time , although it has been shown that the results assessed by these methodologies are not representative of the actual bond strength. In attempts to overcome these limitations, Sano et al. , proposed to evaluate interfacial bond strength from smaller bond test areas. With this methodology, it became possible to obtain several specimens from a single tooth, measuring bond strength in different regions while reducing scatter and achieving adhesive failures in the majority of specimens. Authors have described numerous advantages of this ‘microtensile’ methodology to assess bond strength of different materials . However, bond strength alone does not offer an integral description of adhesion and the adhesive interface should be also investigated thoroughly with spectro- and microscopic investigative tools, such as scanning electron microscopy (SEM), transmission electron microcopy (TEM) , and energy dispersive X-ray spectroscopy (EDS) .

A considerable number of studies presenting microtensile bond strength tests were published in the last decade . However, the description of materials and methods of these studies generally fails to report important test parameters , which can directly influence the bond strength and stress distribution of the tested specimens. The popularization of this methodology allowed innumerous modifications to be introduced by different researchers through the original approach , resulting in conflicting results between studies. These discrepant values are often credited to the different settings applied on the microtensile test, as differing approaches could directly influence the results. Several alterations had been proposed in the geometry of the original specimen, mainly in the trimming of the adhesive region, which resulted in differing bond strength results due to the geometrical characteristics of each specimen . Other changes, such as modification of the original squared cross-sectional area to a cylindrical shape by trimming the specimens in a lathe were proposed . Moreover, a non-trimming technique was also proposed, resulting in stick-shaped specimens and a greater number of samples per tooth, reducing the risk of pre-test failures due to elimination of the free-hand trimming of the adhesive region .

An important parameter, almost always neglected, is the correct alignment of the specimens in the jig used for loading. The misalignment of specimens can be extremely harmful, fatally leading to non-uniform stress distribution in the adhesive layer and consequently scattering of results . As a way to minimize the influence of erroneous alignment of specimens in the microtensile jigs, a metallic device formed by two identical aluminum parts and a 90° central notch named Geraldeli’s device was proposed, allowing self-alignment of specimens previously to test execution . Other parameter hardly standardized is the specimen fixation on the jig, which can be performed involving one or more faces, using active gripping with cyanoacrylate glues, passive gripping without adhesives, or even an association of both methods . As an alternative to active gripping jigs, a microtensile device consisting of two identical aluminum parts and a central self-aligning notch, with lateral shoulders that allow passive gripping of specimens was developed (Dircks device) . At last, a microtensile device that grips the specimen by its ends was introduced, intending to produce better alignment between the applied load and the center of the adhesive layer (“top-bottom” design) .

The type of quick-setting cyanoacrylate glues or fixation adhesives used to grip specimens in active gripping devices is another factor that requires attention, due to the residual stresses that could arise during the setting of these materials . Other parameters such as storage conditions , substrate preparation methods , specimen integrity prior to testing , specimen notch geometry , adhesive layer thickness , bond angle variations , load application speed , are also named as responsible for the great variation observed in bond strength values among microtensile studies. Thus, a great effort have been done to improve the microtensile methodology with the utilization of laboratorial investigations associated to finite element analyses, since the destructive tests alone are not capable to fully access the whole behavior of the tested specimens . On this way, several studies evaluated the effects of the microtensile variables on bond strength and stress distribution, showing an urgency for standardization of parameters used on this test .

Therefore, it is relevant to evaluate the real influence of the different parameters involving microtensile test execution. The aim of this study was to evaluate the effect of the gripping device, specimen geometry and fixation method on microtensile bond strength, failure mode, and stress distribution of an etch-and-rinse 2-step adhesive system to human dentin. The hypothesis that these parameters could influence the results of laboratorial tests and finite element simulations was investigated.

Materials and methods

Teeth and surface preparation

Twenty-one sound human molars extracted for clinical purposes were selected as substrate (gathered following informed consent approved by the Committee for Ethics in Research of the Federal University of Uberlândia: #366/08). Teeth were stored in distilled water at 4 °C and used within 3 months following extraction. The teeth were cleansed of soft tissues, roots notched, and mounted in dental stone blocks to facilitate manipulation. The occlusal surface was ground with carbide bur under copious water spray using the CNC Specimen Former (University of Iowa, Iowa City, IA, USA). Tooth surface was deepened in 0.3 mm increments until all remnants of enamel were removed. Complete removal of enamel was confirmed by 3–5 s application of 37% phosphoric acid, then subsequently, an additional 0.1 mm layer of dentin was removed to present an unaltered substrate for dentin bonding.

After surface preparation, the exposed dentin surfaces were bonded with an etch-and-rinse 2-step adhesive, applied according to the manufacturer’s instructions (Adper Single Bond II; 3M-ESPE, St. Paul, MN, USA). A thin layer of resin-based composite, approximately 0.5 mm in thickness, was bonded to the cured adhesive with margins entirely located in dentin. Then, a block of the same composite was built-up incrementally and light-cured to a height of 4–5 mm (Filtek Z250 A2 shade; 3M-ESPE) ( Table 1 ). Polymerizing steps were performed using a quartz-tungsten halogen curing unit (Optilux 501; Kerr, Orange, CA, USA) with the light guide held perpendicularly in a custom holding device and within 1 mm of the surface. The light output of the curing unit was verified at 850 mW/cm 2 throughout the experiment as monitored with a radiometer in the wavelength range of 400–500 nm (Curing Radiometer model 100, Demetron Research Corp., Danbury, CT, USA). The bonded teeth were stored for 24 h in artificial saliva (pH 7.1) at room temperature.

Table 1
Study materials.
Material Type Composition a Lot
Adper Single Bond 2 (3M-ESPE) Etch-and-rinse 2-step adhesive system Ethyl alcohol 25–35%, bisphenol A diglycidyl ether dimethacrylate 10–20%, silane treated silica (5 nm nanofiller) 10–20%, 2-hydroxyethyl methacrylate 5–15%, glycerol 1,3-dimethacrylate 5–10%, copolymer of acrylic and itaconic acids 5–10%, diurethane dimethacrylate 1–5%, water <5% BRW
Filtek Z250 A2 shade (3M-ESPE) Light-cure resin-based composite Silane treated ceramic 75–85%, bisphenol A polyethylene glycol diether dimethacrylate 5–10%, diurethane dimethacrylate 5–10%, bisphenol A diglycidyl ether dimethacrylate <5%, triethylene glycol dimethacrylate <5%, water <2%. 8HE
Cond AC 37% (FGM) Etching agent Water 63–65%; phosphoric acid 35–37%, synthetic amorphous silica 2–3%. 040309
Super Glue Gel (Loctite) Cyanoacrylate based glue Ethyl cyanoacrylate 60–100%. 86604
Zapit (DVA Inc.) Cyanoacrylate based glue Ethyl-2 cyanoacrylate 60–100%; poly methyl methacrylate 10–30%; hydroquinone 0–1% A39A

a Supplied by manufacturers.

Specimen forming and storage

Each restored tooth was sectioned perpendicular to the bonded surface into two halves, using a low speed diamond saw (Isomet 1000; Buehler, Lake Bluff, IL, USA). One of the halves was sectioned into four stick-shaped specimens with square cross-section of approximately 1 mm 2 . The other half was sectioned into two bar-shaped sticks of approximately 2 mm 2 , which were trimmed with an eight micron ultra-fine-grit diamond bur (ISO 806 314 012; Brasseler GmbH & Co., Lemgo, Germany) in the CNC Specimen Former to obtain dumbbell-shaped specimens. The cylindrical dumbbell-shaped specimens had a 0.51 ± 0.02 mm 2 round cross-sectional area within the 1 mm gauge length and a 0.6 mm radius of curvature. This dumbbell geometry and dimensions are required to match the passive gripping surface of the Dircks device for tensile testing. A digital caliper (Digimatic caliper, Mitutoyo Corporation, Tokyo, Japan) was used to verify the diameter of each specimen. The two dumbbell-shaped specimens were designated to one experimental group and the four stick-shaped bars from the same tooth were divided equally into other two groups. Each specimen was verified under a stereomicroscope (Stemi 2000; Carl Zeiss, Thornwood, NY, USA) for any interfacial defects that would exclude specimen from testing. The prepared specimens were stored in artificial saliva at room temperature for 24 h prior to testing.

Microtensile testing

The specimens were assigned to experimental groups as follows: Di – dumbbell-shaped specimens passively gripped in a Dircks device (University of Iowa) ( Fig. 1 A ) ; GeS – stick-shaped specimens actively gripped onto a Geraldeli’s device , with Superglue Gel cyanoacrylate (Loctite, Henkel Corp., Avon, OH, USA) ( Fig. 2 A ); GeZ – stick-shaped specimens actively gripped onto a Geraldeli’s device with Zapit cyanoacrylate (Dental Ventures of America Inc., Corona, CA, USA) ( Fig. 2 B). Each testing assembly was connected to a universal testing machine (Zwick Z2.5; Zwick GmbH, Ulm, Germany) and the specimens were stressed to failure under tension at 1 mm/min. At no time during specimen fabrication, storage, or testing were the specimens allowed to dehydrate.

Fig. 1
Dircks microtensile device scheme: (A) Dircks device with dumbbell specimen prepared for experimental testing; (B) dumbbell specimen 3D model; (C) meshing of the Dircks device 3D model; (D) surface contact simulating the passive gripping of the Dircks device; (E) FEA boundary conditions applied to the Dircks device model. von Mises stress distribution for: (F) full Dircks microtensile device; (G) passive gripping of the specimen by the device; (H) dumbbell-shaped specimen; (I) longitudinal cut of the model at the gage area.

Fig. 2
Geraldeli’s microtensile device scheme: (A) Geraldeli’s device with stick specimen and Super Glue; (B) Geraldeli’s device with stick specimen and Zapit glue; (C) stick specimen 3D model; (D) meshing of the Geraldeli’s device 3D model; (E) FEA boundary conditions applied to Geraldeli’s device model. von Mises stress distribution for: (F) full Geraldeli’s microtensile device; (G) active gripping of the specimen with cyanoacrylate glue; (H) stick-shaped specimen.

Failure mode determination

All fractured specimens were observed under a stereomicroscope (Stemi 2000, Carl Zeis, Oberkochen, Germany) and the failure mode was determined. The fractured surfaces were classified as: apparently interfacial (within the adhesive joint), cohesive in dentin (dentin in gage, dentin/glue or dentin neck), cohesive in the resin-based composite (resin-based composite in gage, resin-based composite/glue or resin-based composite neck), or mixed failures, in which the failures were recorded as the surfaces comprising the dominance of failure of each substrate (mixed dentin in gage, mixed resin-based composite in gage, mixed dentin/resin-based composite in gage).

Sequentially, specimens were dried, mounted on aluminum stubs with a flowable resin-based composite (Filtek Supreme Flowable; 3M-ESPE), and coated with graphite (carbon) . Fractured surfaces were examined using SEM (LEO 435VP; Carl Zeiss SMT, Oberkochen, Germany) to confirm the mode of failure based on the fracture origin and involvement . The fracture initiation sites and their surface composition were also verified using EDS.

Finite element analysis

Tri-dimensional models of both microtensile devices and respective specimens were created from measurements taken from the geometric pieces of each gripping jig and specimens, using CAD software (Rhinoceros 3D 4.0; McNeel North America, Seattle, WA, USA). Surfaces were created from the external lines obtained by the measurements, and sequentially, external and internal volumes of the jigs and specimens were generated.

The models of the specimens were developed simulating a bi-material body including resin-based composite, dentin and an adhesive layer region. Models were assumed to be 10.02 mm long, with 5 mm of dentin and 5 mm of resin-based composite, bonded with a 20 μm thickness adhesive layer . The dumbbell specimen model was defined with a cross-sectional cylindrical area of 0.5 mm 2 at the bonded interface, with 1 mm of gage length, 0.6 mm radius of curvature, and 2 mm of thickness ( Fig. 1 B). For the stick specimen model, the thickness was 1 mm and the bonded surface was 1 mm 2 consequently ( Fig. 2 C). The models of the Dircks and Geraldeli’s microtensile devices were defined as a simplification of the actual jigs. After finishing the jig models, the respective specimen models were added to the testing sites, such as in the experimental design. Dircks device was defined with a dumbbell specimen and Geraldeli’s device with a stick specimen attached with cyanoacrylate glue on its ends.

The geometrical models were then exported to the pre-processing CAE software (Computer Assisted Engineering – Femap 10.1; Velocity Series, Siemens PLM Software, USA), and meshing of each structure was performed using solid quadratic tetrahedral elements of 10 nodes. Meshing process was carefully controlled using specific tools, resulting in homogeneity and connectivity of the mesh ( Figs. 1C and 2D ). Materials were assumed as being isotropic, linear and elastic, and the mechanical properties used are described in Table 2 . For the Dircks device model, a uniform surface contact was established between the specimen neck and the shoulders of the device simulating a passive gripping ( Fig. 1 D). In the Geraldeli’s device model, a perfect adhesion was considered between the device’s metallic surface, the cyanoacrylate glue, and the specimen’s ends, simulating an active gripping . Additional ideal models of both dumbbell and stick specimens were also evaluated without the jigs, as controls.

Table 2
Mechanical properties used to develop finite element models.
Material Elastic modulus (GPa) Reference Poisson’s ratio ( v ) Reference
Human dentin 18.0 0.31
Resin-based composite (Filtek Z250) 18.5 0.308
Bonding adhesive (Adper Single Bond 2) 8.43 0.3
Cyanoacrylate glue 6.0 0.3
Aluminum 71.0 0.33

Boundary conditions were defined with a 30 N tensile load applied parallel to the model on its top surface, and a nodal displacement constraint applied on the bottom and lateral surfaces of the model on axes X , Y and Z ( Figs. 1E and 2E ). Models were then exported to FE processing software (NEi Nastran 9.2, Noran Engineering, Westminster, CA, USA) and the solution of each model was run. The solved files were exported to the post-processing software (Femap 10.1), and the results were analyzed qualitatively using von Mises criteria.

Statistical analysis

Bond strength and mode of failure

The Shapiro–Wilk test was conducted to test the data for normality. One-way ANOVA with post hoc Tukey’s HSD (Honestly Significant Difference) was used to determine whether there was a significant difference in the bond strength between gripping devices. When taking into account the correlated specimens obtained from the same tooth, the random effects in Mixed Model ANOVA (i.e. to allow correlation between three specimens obtained from the same tooth) was conducted to evaluate the effect of types of gripping devices on the bond strength. Furthermore, parametric Weibull regression models were performed to evaluate whether there was a significant association between the bond strength and the types of gripping devices by Wald Chi-square test. The random effect Weibull regression models with repeated events (multiple specimens from the same tooth) were used to examine the results of parametric Weibull regression models while considering the tooth random effect and correlations between specimens. Additionally, a difference in failure mode or in human error between gripping devices was tested using Fisher’s exact test and Chi-square test. A two-sample t -test was performed to detect a difference between Ge groups when regluing was required to complete a successful tensile test. All tests employed a 0.05 level of statistical significance and all statistical analyses were carried out with the statistical package SAS ® System version 9.2 (SAS Institute Inc., Cary, NC, USA).

Bond strength with equal specimen resin–dentin bond test area

The area or volume tested affects resultant bond strengths by the following expression: σ o / σ = ( A / A o ) 1/ m , in which an increase in specimen size increases the probability of finding a larger defect present and therefore a reduction in bond strength . Bond strength results for the larger specimens tested in the Dircks device were transformed using this formula: σ o / σ = ( A / A o ) 1/ m , where σ o is the Dircks dumbbell specimen microtensile strength and A o is its cross-sectional area (fixed at 0.5 mm 2 ), σ is the Geraldeli stick specimen microtensile strength and A is its cross-sectional area (matched to σ ), and m is the Dircks device testing reliability value (from our study = 3) . The same statistics procedures as described in previous section were used to analyze the transformed bond strengths.

Materials and methods

Teeth and surface preparation

Twenty-one sound human molars extracted for clinical purposes were selected as substrate (gathered following informed consent approved by the Committee for Ethics in Research of the Federal University of Uberlândia: #366/08). Teeth were stored in distilled water at 4 °C and used within 3 months following extraction. The teeth were cleansed of soft tissues, roots notched, and mounted in dental stone blocks to facilitate manipulation. The occlusal surface was ground with carbide bur under copious water spray using the CNC Specimen Former (University of Iowa, Iowa City, IA, USA). Tooth surface was deepened in 0.3 mm increments until all remnants of enamel were removed. Complete removal of enamel was confirmed by 3–5 s application of 37% phosphoric acid, then subsequently, an additional 0.1 mm layer of dentin was removed to present an unaltered substrate for dentin bonding.

After surface preparation, the exposed dentin surfaces were bonded with an etch-and-rinse 2-step adhesive, applied according to the manufacturer’s instructions (Adper Single Bond II; 3M-ESPE, St. Paul, MN, USA). A thin layer of resin-based composite, approximately 0.5 mm in thickness, was bonded to the cured adhesive with margins entirely located in dentin. Then, a block of the same composite was built-up incrementally and light-cured to a height of 4–5 mm (Filtek Z250 A2 shade; 3M-ESPE) ( Table 1 ). Polymerizing steps were performed using a quartz-tungsten halogen curing unit (Optilux 501; Kerr, Orange, CA, USA) with the light guide held perpendicularly in a custom holding device and within 1 mm of the surface. The light output of the curing unit was verified at 850 mW/cm 2 throughout the experiment as monitored with a radiometer in the wavelength range of 400–500 nm (Curing Radiometer model 100, Demetron Research Corp., Danbury, CT, USA). The bonded teeth were stored for 24 h in artificial saliva (pH 7.1) at room temperature.

Table 1
Study materials.
Material Type Composition a Lot
Adper Single Bond 2 (3M-ESPE) Etch-and-rinse 2-step adhesive system Ethyl alcohol 25–35%, bisphenol A diglycidyl ether dimethacrylate 10–20%, silane treated silica (5 nm nanofiller) 10–20%, 2-hydroxyethyl methacrylate 5–15%, glycerol 1,3-dimethacrylate 5–10%, copolymer of acrylic and itaconic acids 5–10%, diurethane dimethacrylate 1–5%, water <5% BRW
Filtek Z250 A2 shade (3M-ESPE) Light-cure resin-based composite Silane treated ceramic 75–85%, bisphenol A polyethylene glycol diether dimethacrylate 5–10%, diurethane dimethacrylate 5–10%, bisphenol A diglycidyl ether dimethacrylate <5%, triethylene glycol dimethacrylate <5%, water <2%. 8HE
Cond AC 37% (FGM) Etching agent Water 63–65%; phosphoric acid 35–37%, synthetic amorphous silica 2–3%. 040309
Super Glue Gel (Loctite) Cyanoacrylate based glue Ethyl cyanoacrylate 60–100%. 86604
Zapit (DVA Inc.) Cyanoacrylate based glue Ethyl-2 cyanoacrylate 60–100%; poly methyl methacrylate 10–30%; hydroquinone 0–1% A39A

a Supplied by manufacturers.

Specimen forming and storage

Each restored tooth was sectioned perpendicular to the bonded surface into two halves, using a low speed diamond saw (Isomet 1000; Buehler, Lake Bluff, IL, USA). One of the halves was sectioned into four stick-shaped specimens with square cross-section of approximately 1 mm 2 . The other half was sectioned into two bar-shaped sticks of approximately 2 mm 2 , which were trimmed with an eight micron ultra-fine-grit diamond bur (ISO 806 314 012; Brasseler GmbH & Co., Lemgo, Germany) in the CNC Specimen Former to obtain dumbbell-shaped specimens. The cylindrical dumbbell-shaped specimens had a 0.51 ± 0.02 mm 2 round cross-sectional area within the 1 mm gauge length and a 0.6 mm radius of curvature. This dumbbell geometry and dimensions are required to match the passive gripping surface of the Dircks device for tensile testing. A digital caliper (Digimatic caliper, Mitutoyo Corporation, Tokyo, Japan) was used to verify the diameter of each specimen. The two dumbbell-shaped specimens were designated to one experimental group and the four stick-shaped bars from the same tooth were divided equally into other two groups. Each specimen was verified under a stereomicroscope (Stemi 2000; Carl Zeiss, Thornwood, NY, USA) for any interfacial defects that would exclude specimen from testing. The prepared specimens were stored in artificial saliva at room temperature for 24 h prior to testing.

Microtensile testing

The specimens were assigned to experimental groups as follows: Di – dumbbell-shaped specimens passively gripped in a Dircks device (University of Iowa) ( Fig. 1 A ) ; GeS – stick-shaped specimens actively gripped onto a Geraldeli’s device , with Superglue Gel cyanoacrylate (Loctite, Henkel Corp., Avon, OH, USA) ( Fig. 2 A ); GeZ – stick-shaped specimens actively gripped onto a Geraldeli’s device with Zapit cyanoacrylate (Dental Ventures of America Inc., Corona, CA, USA) ( Fig. 2 B). Each testing assembly was connected to a universal testing machine (Zwick Z2.5; Zwick GmbH, Ulm, Germany) and the specimens were stressed to failure under tension at 1 mm/min. At no time during specimen fabrication, storage, or testing were the specimens allowed to dehydrate.

Fig. 1
Dircks microtensile device scheme: (A) Dircks device with dumbbell specimen prepared for experimental testing; (B) dumbbell specimen 3D model; (C) meshing of the Dircks device 3D model; (D) surface contact simulating the passive gripping of the Dircks device; (E) FEA boundary conditions applied to the Dircks device model. von Mises stress distribution for: (F) full Dircks microtensile device; (G) passive gripping of the specimen by the device; (H) dumbbell-shaped specimen; (I) longitudinal cut of the model at the gage area.

Nov 28, 2017 | Posted by in Dental Materials | Comments Off on Effect of specimen gripping device, geometry and fixation method on microtensile bond strength, failure mode and stress distribution: Laboratory and finite element analyses
Premium Wordpress Themes by UFO Themes