Evaluation of Tensile Retention of Y-TZP Crowns Cemented on Resin Composite Cores: Effect of the Cement and Y-TZP Surface Conditioning

Zirconia crowns have superior flexural strength (> 1000 MPa)¹  and fracture toughness (> 9-10 MPam^1/2)²  than those of other ceramics due to the presence of tetragonal crystals stabilized by yttrium oxide in the material microstructure. By means of thermal, mechanical, or chemical stimulation, these tetragonal grains transform into a monoclinic arrangement, leading to a significant increase in grain volume, generating localized compressive stresses that hinder both slow and fast crack propagation.³  The outstanding mechanical behavior of yttria-stabilized polycrystalline tetragonal zirconia (Y-TZP) makes it an excellent candidate material to be used as the infrastructure of both all-ceramic crowns veneered with porcelain and monolithic prostheses.⁴ 

However, the physical and chemical characteristics of Y-TZP do not favor bonding to current resin cements.^5,6  Although it has been reported by Casucci and others⁷  and Blatz and others⁸  that resin-based luting agents are the most appropriate materials for the purposes of marginal seal, retention, and fracture resistance, the use of surface pretreatments to improve bonding to the luting agent is mandatory in this case. Different surface treatments have been proposed to modify the zirconia surface and promote mechanical and/or chemical adhesion.^9-13  According to Inokoshi and others,¹⁴  the combination of mechanical and chemical pretreatments can be recommended to promote a stable bond to zirconia.

Sandblasting with aluminum oxide particles coated by silica, followed by the application of silane (silanization),^15,16  has been widely used to increase surface roughness (micromechanical retention) and to establish chemical bonds with resin cements.^17,18  An alternative method is the application of a thin layer of glassy porcelain (rich in silica), followed by etching with hydrofluoric acid and silanization.^12,19-23  According to Ntala and others,²¹  that surface treatment increases the capacity of the Y-TZP surface to establish both chemical and micromechanical interactions with resin composites.

Another important factor to be considered when using all-ceramic restorations is the substrate on which the crown will be cemented. When the dentin remaining is insufficient, it is recommended to use an intraradicular post and a composite core buildup to support the crown. In this clinical situation, the use of prefabricated fiber posts associated with a composite core will have a positive impact on the final esthetic result of a cemented zirconia crown.²⁴  However, it is important to keep in mind that the adhesion process of resin cements to dentin is significantly different than that observed for composite resins. A resin composite core has very few unreacted methacrylate groups at the surface when the cementation procedure takes place. This fact reduces the potential for bonding to a new resinous material, such as the resin cement.²⁵  In order to improve the adhesion forces, it is necessary to modify the composite core surface to improve chemical bonding to resin cements. An example of a surface treatment for the resin core is the use of self-etching adhesives, which have been proven to promote high bond strength values between aged and new resin due to an efficient wettability provided by the self-etching systems.²⁶ 

Therefore, it can be concluded that it is necessary to assess not only the bond strength of the resin cements to dentin but also the interface between the resin cement and the composite resin core. Thus, the objective of this study was to compare different treatments of the inner surface of frameworks made of Y-TZP and different types of cements in terms of crown retention (tensile forces) in preparations made mostly of resin composite. The hypotheses were that 1) the different adhesive cements would lead to similar crown retention values; 2) adhesive resin cements would promote higher retention values than the self-adhesive, glass ionomer, and zinc phosphate cements; and 3) the Y-TZP intaglio surface treatment would improve retention regardless of the cement used.

In order to determine the sample size, a sample calculation was made based on two other articles.^27,28  Considering a statistical power of 80%, mean standard deviation of 1.7, and a detectable difference of 2.3 MPa, the sample size was established as n = 12, for a total of 108 teeth.

The teeth were donated by the Human Teeth Bank of the Federal University of Santa Maria and stored in distilled water (4°C) until needed. The 108 teeth were numbered and assigned randomly into nine testing groups (Table 1), using a computer program (www.randomizer.org).

The coronal part of the tooth was glued to an adapted surveyor to keep the long axis of the tooth perpendicular to the ground (horizontal plane) when embedding each root into the acrylic resin. Self-cured acrylic resin (Dencrilay, Dencril, Caieiras, SP, Brazil) was prepared and poured into the matrix. Then the tooth was inserted into the resin, up to 3 mm below the cemento-enamel junction.

The occlusal surfaces of all teeth were cut with a diamond blade mounted on a cutting machine (Isomet 1000, Buehler, Lake Bluff, IL, USA), 4 mm above the cemento-enamel junction. Conical trunk diamond burs (KG 3139 and KG 3139FF, KG Sorensen, Cotia, Brazil) were mounted in a high-speed hand piece and fixed to a modified optic microscope to obtain reductions as parallel as possible to the long axis of the tooth. Thus, the axial walls were reduced at depths of 1.5 mm (similar to the bur diameter), and a standard angle of convergence was created. The height of the preparation was 4 mm.

1. A vinyl polysiloxane impression (Elite Light Body-normal set, Zhermack, Badia Polesine, Italy) of each dental preparation was performed. Then master dies were produced, and a silicone matrix was fabricated on each master die. Thus, the future composite core reconstruction had the same characteristics as the full crown preparation for each tooth.

2. All of the dental preparations were removed by cutting with a diamond blade (Isomet 1000).

3. For the post cementation, the greatest root canal of each molar received intracanal preparation with a custom #2 drill of the glass fiber post system (White Post DC, FGM, Joinville, Brazil). Afterward, the coronal and root dentin received an etch-and-rinse single-bottle adhesive system (Ambar, FGM). The dentin was etched with 37% phosphoric acid, rinsed with water for 20 seconds, and dried with absorbent papers. The adhesive agent was then applied according to the manufacturer's instructions and light-cured for 20 seconds (Radii-Cal, SDI, Australia). The posts received silane application (Prosil, FGM) and were cemented with a dual-cure resin cement (Allcem, FGM). The core was built up with a composite resin (Opallis, Joinville, Brazil), using the silicone matrix previously made. All preparations were finished with a fine conical trunk diamond bur (3139FF, KG Sorensen) under low rotation. The specimens were stored in water (37°C) for 24 hours.

The preparations of the specimens were molded with polyvinylsiloxane (Elite Light Body-normal set, Zhermack). Afterward, master dies (CAM-BASE type 4, Dentona, Dortmund, Germany) were obtained and taken to the CEREC MC XL IN LAB for the crown manufacture using Software Inlab 3.60. The copings were designed with retentive features on the occlusal surface for subsequent tensile testing, and milling was performed using VITA In-Ceram 2000 YZ CUBES (VITA Zahnfabrik, Bad Säckingen, Germany) and sintered in a Zircomat furnace (VITA Zahnfabrik) as recommended by the manufacturer.

Before Y-TZP framework cementation, the inner surface of Y-TZP copings of the groups Multilink and RelyX ARC were treated using three different methods, according to Table 1.

After the Y-TZP surfaces were conditioned, the surfaces were silanized with a silane agent (ESPE-Sil, 3M ESPE, St. Paul, MN, USA) that was applied for 60 seconds and then thoroughly air-dried. All of the infrastructures were cemented using a device that exercised a force of 750g over the assembled tooth/Y-TZP infrastructure.

After cementation, all specimens were stored in distilled water at 37°C for 24 hours and then submitted to thermocycling (6000 cycles) between 5°C and 55°C, according to Ernst and others²⁷  and Palacios and others.²⁸ 

Before tensile testing, part of the cemented infrastructure was embedded in acrylic resin (Dencrilay, Dencril, SP, Brazil) until the resin covered the retention form on the crown. This procedure was performed following the same axis as used with root embedding with the aid of an adapted surveyor.

For testing, the lower base of the assembly was fixed on a universal testing machine (DL-2000, Emic, Pinhais, PR, Brazil), and the upper base was connected to a mobile device that contained a load cell of 1000 N, and the tensile strength test was performed until fracture with a speed of 0.5 mm/min (Figure 1).

The fractured interfacial surfaces of the tested specimens were analyzed under a stereomicroscope (SteREO Discovery V12, Carl Zeiss, Gottingen, Germany). The failure mode was classified as: over 50% of cement in crown; over 50% of cement in substrate; and catastrophic failure (post debonding, root fracture and removal of the root from the acrylic resin).

Three Y-TZP discs (10 mm in diameter, 3 mm in thickness) were treated with tribochemical silica coating, vitrification, and no treatment, exactly as if conditioning of the inner surfaces of the framework. The micromorphological changes were inspected using scanning electron microscopy (SEM).

The tensile retention data were submitted to Kruskal-Wallis and Dunn multiple comparison tests (α=0.05). A statistical analysis was performed among the groups that were cemented with Multilink and RelyX ARC in order to compare only adhesive resin cements.

The data of the groups MultC and RelC were subject to the Dunn multiple comparison test to verify the effect of the cement on the crown without surface treatment. The same was performed for the other surface treatments. The data of the groups MultC, MultS, and MultV were subject to the Dunn multiple comparison test to verify the effect of the surface treatment on the crown cemented with Multilink cement. The same was performed for the RelyX ARC cement.

Another statistical analysis was performed among all groups without the zirconia surface treatment. The data of the groups MultC, RelC, Ion, Self, and Zinc were subject to the Dunn multiple comparison test to compare the cement types without the zirconia surface treatment.

The median, maximum, minimum, and first and third quartiles (Q1 and Q3) of all groups are shown in the box plot (Figure 2). The data are shown as mean and standard deviation for better comparison with others studies. The comparisons of the groups cemented with the adhesive resin cements are shown in Table 2. The vitrification method showed significantly higher retention values when compared to the control group (no treatment) only for the groups cemented with Multilink. However, the vitrification and silicatization surface treatments were statistically similar for the Multilink cement. The factor “cement type” was not statistically significant.

Table 3 shows the comparisons between cement groups without conditioning of the framework inner surfaces. Bis-GMA- and HEMA-based resin cements showed the highest retention values. The other cements showed similar retention values.

The failure analysis indicated that most of the groups had high percentages of cement remaining over the composite core preparation after the tensile testing, except for crowns cemented with Multilink (Figure 3), since most failures for those crowns were catastrophic. For the crowns cemented with a resin-modified glass ionomer (RelyX Luting), all specimens showed a higher percentage of cement on the crown side. Representative images for the failure analysis can be seen in Figure 4.

Representative SEM images of the conditioned Y-TZP surfaces are shown in Figure 5. Relevant changes were observed after the different surface treatments (Figure 5B,C) when compared to the untreated surface (Figure 5A). Pits and microretentions caused by selective etching with hydrofluoric acid can be seen on the glazed Y-TZP surface (Figure 5C).

The retention values obtained for both the chemically activated dual-cure resin cement (Multlink) and the BIS-GMA-based cement were not statistically different from each other. Thus, the first hypothesis of this study was accepted. Multilink Automix contains HEMA molecules, which possess a molecule that contains one hydroxyl radical, increasing the stability of the monomer under moist and acidic conditions.²⁹  Additionally, Multilink is chemically activated, which minimizes polymerization shrinkage due to the slower polymerization reaction.

It is important to note, though, that all three groups cemented with Multilink showed higher percentages of catastrophic failures, especially for group MultS. It may be concluded that the posts debonded from the intracanal dentin before the crown debonded from the preparation, indicating that the bond strength between the post and the dentin substrate was not as high as the one obtained between the cement and dentin. If the post had not debonded from the root canal, the retention values of these groups might have been higher. This observation is in line with the results obtained from Palacios and others,²⁸  who stated that the tensile retention values of some of the experimental groups were influenced by the cohesive failure of the zirconia coping and the teeth, thereby underestimating the real retention values. In the present investigation, specimens cemented with Multilink showed significantly more catastrophic failures than those cemented with RelyX ARC. Thus, it might be inferred that the Multilink cement could promote greater retention values if the posts had not debonded. According to Luthy and others,⁵  RelyX ARC is a conventional dual-cure resin cement (based on BIS-GMA) that does not have any functional monomer in its composition, keeping it from developing any kind of chemical bonding to the crown or the core.

When considering the factor “surface treatment,” it was significant only for the Multilink cement, and therefore the second hypothesis was rejected. Silica coating did not increase the retention values of the crowns for the Multilink and RelyX ARC cements when compared to the control group. This fact may be related to the difficulty in applying the silica layer on the zirconia surface since Y-TZP has very high hardness and fracture toughness. Ntala and others²¹  found that this surface treatment did not create the micromechanical retention required for efficient adhesive bonding, confirming the findings of the present study. Additionally, according to Borges and others³⁰  and Oyague and others,³¹  airborne particle abrasion using 50 μm of alumina oxide has little effect on the morphological surface features of zirconium dioxide ceramics and does not result in a deep surface modification. In the present study, smaller particles were used (30 μm), which might have prevented surface modification. Perhaps this is the reason why the group cemented with silica coating and Relyx ARC showed the highest amount of remaining cement on the core surface and not on the crown cementation surface. For the Multilink cement, most of the failures were catastrophic, indicating that this group (MultS) probably would show greater retention values if post debonding were avoided.

The vitrification process was efficient only for the Multilink cement, as it resulted in improved retention values when compared to the control groups. However, for the MultV group, most failures were either catastrophic or between the crown and cement with higher amounts of cement remaining on the core surface. As previously mentioned, if the specimens that fractured catastrophically could have been tested, the failures would probably have occurred between the crown and cement because no failures occurred between the core and the cement in the MultV group.

It is probable that the 10-μm-thick¹²  glass layer on the inner surface might have caused increased friction between the crown and the preparation walls, leading to higher retention values. In addition, Vanderlei and others¹²  showed that one of the limitations of the vitrification technique is the difficulty in standardizing the glaze application inside the Y-TZP infrastructure since the glass layer applied on the intaglio surface of the Y-TZP infrastructure creates a layer thick enough to interfere with seating the infrastructure. Thus, technical improvements need to be tested for reducing the effect on the marginal adaptation.

When all cements without zirconia surface treatment were compared (Table 3), the groups that had the highest retention values were the BIS-GMA- and HEMA-based cements. Thus, the third hypothesis of this study was accepted. Regarding RelyX U100, although its chemical composition containing methacrylated phosphoric esters has not been fully disclosed by the manufacturer, it has been shown that these monomers can bond to ceramic surfaces by means of the same mechanisms previously described for the monomer 10-MDP (also a methacrylated phosphoric acid ester).³²  Yap and others³³  reported that the bonding mechanism of RelyX Unicem is reminiscent of the self-adhesiveness of glass ionomer cements and that a possible improvement in bond strength may occur after cement maturation, over time. However, according to the failure analysis, which shows that the largest amount of cement remained attached in the core, it can be concluded that the bond failure occurred between the cement and the zirconia.

Manufacturers of the glass ionomer based–cement reported that, although this material has a chemical affinity for dentin hydroxyapatite, it has little affinity to the resin composites, which explains their low performance in terms of retention values and failure mode (between cement and composite resin core for all specimens).

In the present study, the manufacturers of the zinc phosphate cement (control group) still claim that this material can be used for cementation of zirconia crowns. It is interesting to note that this cement showed statistically similar retention values when compared to those obtained by the ionomer and self-adhesive resin cements. This retention behavior of the zinc-phosphate cements may be attributed only to the high friction coefficient between the crown and composite resin core walls, as this material does not bond to either the ceramic or the preparation substrate.³⁴ 

One of the limitations of this study was the fact that it was not possible to measure the adhesive area of the preparation in order to calculate the nominal tensile retention stress for all of the ceramic crowns. However, it is believed that the retention force values (in kgf) used by this current investigation were reliable enough once samples were randomized and preparations were standardized. Palacios and others²⁸  compared the tensile strength (in MPa) and the load for crown pull-out (tensile retention in kg) and showed similar outcomes for both measurement units due to the homogeneity of preparation total areas. Those authors also calculated the average total preparation area for each experimental group and showed that the adhesive areas were similar for the different teeth used.

The clinical relevance of this study is that it simulated different cementation protocols for Y-TZP crowns in an in vitro design, as a clinical crown cemented on a tooth reconstructed with resin composite core and a fiber post. Further studies should be conducted to investigate other factors involved in the retention of Y-TZP crowns, such as longitudinal fatigue testing, the evaluation of different cementation strategies, and other surface treatments for the inner surface of Y-TZP frameworks.

1. The type of resin cement (BIS-GMA- or HEMA-based) did not affect the crown retention values.

2. The conditioning of the zirconia intaglio surface by the application of a thin low-fusing glass porcelain layer plus silanization was capable of improving the retention force for the HEMA-based cement.

3. For the untreated intaglio surfaces of Y-TZP crowns, resin cements showed significantly higher retention values when compared to those obtained for the self-adhesive resin cement, glass ionomer, and zinc phosphate cements.

This study received grant support from the Sao Paulo Research Foundation (FAPESP; São Paulo, Brazil) (# 2010/52672-4, # 2012/03436-1). This study is based on a Doctorate Thesis submitted to the São José dos Campos Dental School, São Paulo State University [UNESP] (Brazil) as part of the requirements for the doctorate degree (Dr Marília Pivetta Rippe).

The authors of this manuscript certify that they have no proprietary, financial, or other personal interest of any nature or kind in any product, service, and/or company that is presented in this article.