1

Introduction

Since being hailed as “ceramic steel” by Garvie et al. zirconia has been investigated for myriad applications . As an engineering material, zirconia provides a unique set of material properties allowing it to be tailored for specific technologies. It is this combination of mechanical properties (similar to those of metals) and optical properties that makes zirconia attractive for dental restorations . Zirconia provides better fracture resistance and, potentially, enhanced long-term viability than porcelain and other inorganic, non-metallic alternatives.

There is a wealth of information in the scientific literature regarding the use of zirconia in dental applications; such as frameworks for all-ceramic posterior crowns and fixed partial dentures, root canal posts, implants, and implant abutments . Although superior in terms of mechanical performance (strength, toughness, fatigue resistance) when compared to alternative materials, a consistent problem associated with zirconia is poor adhesion to the variety of substrates (synthetic or tissues) that can be encountered in dental or other biomedical applications. Conventional cementation/attachment techniques used with zirconia components do not provide sufficient bond strength for many of these clinical applications due to the inert nature of zirconia . This also explains why zirconia frameworks are resistant to aggressive chemical agents such as most acids, alkalis, and/or organic/inorganic solvents . It is thus important for high retention, prevention of microleakage, and increased fracture and fatigue resistance, that reliable bonding techniques be established. This report presents a novel pretreatment technique, whereby the zirconia surface is converted to a more reactive zirconium oxyfluoride, enabling improved chemical bonding to other dental substrates via conventional silanation approaches.

Strong resin bonding relies on both micromechanical interlocking and chemical adhesion to the ceramic surface . Thus, both surface roughening for the mechanical interlocking and surface activation for chemical adhesion are often required. In some instances, high strength ceramic restorations do not require chemical adhesive bonding to tooth structure and can be placed using conventional cements which rely only on micromechanical retention. However, strong resin bonding is desirable in many clinical situations. In addition, it is likely that the air-particle abrasion would lead to enhanced long-term fracture and fatigue resistance in the oral environment. Therefore, a simple, non-destructive method for pre-treating zirconia-based ceramic surfaces would be clinically beneficial. Such a pre-treatment would facilitate the use of resin-based cements to bond zirconia to various substrates; including tooth structure and zirconia implant abutments, and potentially eliminate necessary surface damage. Also, it would be highly desirable if the surface treatment is compatible with more traditional silica-based based dental ceramics.

There are numerous reports of novel bonding approaches; the work has been summarized in multiple review manuscripts . As previously stated, traditional adhesive chemistry is ineffective on zirconia surfaces due to the material’s stable, non-reactive properties. Some of the popular approaches to enhance zirconia adhesion include:

• •

  Surface abrasion or roughening (airborne-particle abrasion, diamond bur) establishes adhesion through micro-mechanical retention with no chemical bonding benefits. It has been reported by some that particle abrasion results in the creation of sharp crack tips and structural defects that render zirconia susceptible to radial cracking during service . However, Magne et al. noted that using particles up to 50 μm in size for surface abrasion does not reduce the strength of zirconia (per manufacturer’s data) .

• •

  Application of a tribochemical silica coating allows for chemical bonds to a silane coupling agent and resin cement. This is a commercial product, but could be affected by user variability and does not produce bond strengths as high as those reported for silane-bonded porcelain .

• •

  Application of phosphate ester primers and phosphate-modified resin cements creates a relatively stable bond, but has been shown to be insufficient during exposure to simulated intraoral environments. This method coupled with particle abrasion has been shown to display high bond strength . However, bond strength values are generally lower than those reported for tribochemical silica coating coupled with silane and resin cement.

• •

  Silica – coating techniques (fusion of glass bead and plasma spray) to facilitate a siloxane bond between a glass-like zirconia surface and resin cement display functional bond strengths ; however, it introduces complicated techniques to achieve adequate coverage (e.g. thermal spray).

• •

  Chloro – silane treatment allows for a thin Si _x O _y film to coat the zirconia surfaces enabling traditional silane chemistries and resin cements to be used . While significantly more thin than the silica-coating techniques above, this Si _x O _y thin-film process, reported bond strengths higher than tribochemical functionalization and statistically similar to porcelain.

• •

  Selective infiltration etching (SIE) process creates an inter-grain nano-porosity at the surface where the primer and adhesive resin can infiltrate and interlock, facilitating higher bond strength . While still under development, this material-removal technique appears mechanically favorable according to the above reports, yet it was noted by the authors that long-term resistance to microleakage and bond viability may still be a concern.

The currently available approaches for adhesive bonding of ZrO ₂ bioceramics are not adequate for all clinical applications, and it should be noted that long-term durability has yet to be investigated for most. More recently, limited clinical studies of ZrO ₂ prosthetics have reported loss of retention as a reason for clinical failure, further emphasizing the need for an improved, robust bonding solution.

Interestingly, studies have yet to investigate a chemical bonding structure modification of zirconia surfaces to create a “reactive” surface that would facilitate chemical bonding. Various fluorination processes have been used in the past to modify the physical and chemical properties of other inert surfaces. For example, both the electronic and wear properties of diamond can be altered through either gas- or liquid-phase fluorination. Fullerenes and carbon nanotubes may also be modified electrically and optically through surface fluorination, a process that converts the predominantly sp ² – hybridized carbon into a mixture of sp ² /sp ³ . Related to the present study, Pantono and Brow reported on the reactive properties of fluorozirconate glasses and the relative classification of sub-oxides of fluorinated zirconium dioxide (or zirconium oxyfluoride, ZrO _x F _y ); however, little work has been reported on the chemical reactivity of these oxyfluoride materials . Hess and Kemnitz did evaluate the surface acidity and catalytic behavior of modified Zr- and Ti-dioxides, demonstrating that fluorinated surfaces increased reactivity for certain acid–base reactions . The general properties of the fluorozirconate glasses along with the catalytic behavior of refractory metal oxyfluorides led these authors to study the materials in more detail, and support the premise that an oxyfluoride-modified zirconia surface could be more chemically active than pure zirconia.

Presented in this study, a simple fluorination vapor technique (shown in Fig. 1 ) is used to create an oxyfluoride conversion layer on zirconia surfaces to enhance reactivity. The overall goals were to analyze stoichiometric differences of as-received, compared to fluorinated zirconia specimens and then evaluate their adhesive strength when bonded to a dental composite. Simple shear bond tests were employed to measure adhesion on polished and roughened zirconia specimens and compared to alternative pretreatment techniques.

Schematic representation of the plasma surface modification process. Fluorinated plasma is applied to the zirconia surface, converting the top 1–3 nm into a zirconium oxyfluoride (ZrO _x F _y ). The oxyfluoride surface reacts with organo-silanes, enabling silicon attachment to the surface.

2

Materials and methods

Blocks of pre-sintered zirconia (ZirCAD, Ivoclar-Vivadent, Schaan, Liechtenstein) measuring 14 mm × 12 mm × 20 mm were obtained from the manufacturer and sectioned into 2 mm plates. Composite cylinders (Filtek Supreme, 3M-ESPE, St. Paul, MN) were fabricated by condensing the material into a Teflon mold (2 mm diameter × 3 mm height) and UV light-activated for 40 s at 500 mW/cm ² . Surfaces of each material were highly polished through 50 μm diamond grit polishing paper to ensure starting surface roughness. After polishing, select surfaces were air-abraded (50 μm alumina abrasive, 0.29 MPa, 20 s) prior to chemical surface treatments and/or bonding procedures. Abraded specimens were rinsed with iso-propanol and submersed in DI ultra-sonic bath for 5 min.

Zirconia specimens were fluorinated in a planar, inductively coupled 13.56 MHz plasma reactor at 800 W with a dc bias of – 300 V. A continuous flow source gas of SF ₆ at 25 sccm was used to maintain a pressure of 35 mT for 2 min . X-ray photoelectron spectroscopy (XPS) was used to evaluate surface chemistry and stoichiometry of the conversion layer. A Kratos Analytical Axis Ultra XPS system with a monochromatic Al kα source operated at 15 kV and pass energy of 20 eV was used to obtain Zr 3 d core level spectra. The spectra was then deconvoluted using CasaXPS™ software employing a Shirley background subtraction and mixed Gaussian–Lorentzian (G–L) peaks associated with the oxide and oxyfluoride components. The spectra were referenced to the Zr 3 d _5/2 peak at 182.2 eV for ZrO ₂ .

Below are the seven groups ( n = 10) from which shear bond specimens were fabricated, with variations for each surface treatment. All shear bond specimens were prepared using the same bonding procedure. Zirconia surfaces were modified (see below for modification techniques) and treated with an organosilane (Monobond-S, Ivoclar-Vivadent, Schaan, Liechtenstein) prior to resin cement bonding. Composite cylinders were coated with resin cement (Rely-X Unicem, 3M-ESPE, St. Paul, MN), placed on the zirconia surface and UV-light cured under a defined load (5 N):

• Groups 1 and 2 ( control ): (1) Polished, untreated surface and (2) roughened, untreated surface.

• Groups 3 and 4: Surfaces were polished (3) and roughened (4) were modified with a 3 nm Si _x O _y layer, this procedure is described in detail in Piascik et al. .

• Group 5: Zirconia surfaces were silica-coated using 30 μm alumina particles modified with salicylic acid (CoJet, 3M-ESPE, St. Paul, MN – 0.28 MPa, 5–10 mm working distance, 15 s).

• Groups 6 and 7: Zirconia surfaces were polished (6) or roughened (7) were exposed to the fluorination process described above.

Shear bond test specimens were stored in DI water at 37 °C for a period of 24 h prior to testing. Specimens were then fixed to a custom vise fixture to ensure vertical compliance. All specimens were subjected to a uniform force (semi-circular, notched crosshead fitting the diameter of the composite cylinder) at a crosshead speed of 0.5 mm/min in an electro-mechanical testing device (Instron Corp., Norwood, MA). Shear bond strengths were calculated by dividing peak load by the cross-sectional area of the composite cylinder. Single-factor analysis of variance (ANOVA) at a 5% confidence level was performed for the bonding strength data. Optical microscopy and scanning electron microscopy (SEM) were used to evaluate and quantify failure surfaces.

2

Materials and methods

Blocks of pre-sintered zirconia (ZirCAD, Ivoclar-Vivadent, Schaan, Liechtenstein) measuring 14 mm × 12 mm × 20 mm were obtained from the manufacturer and sectioned into 2 mm plates. Composite cylinders (Filtek Supreme, 3M-ESPE, St. Paul, MN) were fabricated by condensing the material into a Teflon mold (2 mm diameter × 3 mm height) and UV light-activated for 40 s at 500 mW/cm ² . Surfaces of each material were highly polished through 50 μm diamond grit polishing paper to ensure starting surface roughness. After polishing, select surfaces were air-abraded (50 μm alumina abrasive, 0.29 MPa, 20 s) prior to chemical surface treatments and/or bonding procedures. Abraded specimens were rinsed with iso-propanol and submersed in DI ultra-sonic bath for 5 min.

Zirconia specimens were fluorinated in a planar, inductively coupled 13.56 MHz plasma reactor at 800 W with a dc bias of – 300 V. A continuous flow source gas of SF ₆ at 25 sccm was used to maintain a pressure of 35 mT for 2 min . X-ray photoelectron spectroscopy (XPS) was used to evaluate surface chemistry and stoichiometry of the conversion layer. A Kratos Analytical Axis Ultra XPS system with a monochromatic Al kα source operated at 15 kV and pass energy of 20 eV was used to obtain Zr 3 d core level spectra. The spectra was then deconvoluted using CasaXPS™ software employing a Shirley background subtraction and mixed Gaussian–Lorentzian (G–L) peaks associated with the oxide and oxyfluoride components. The spectra were referenced to the Zr 3 d _5/2 peak at 182.2 eV for ZrO ₂ .

Below are the seven groups ( n = 10) from which shear bond specimens were fabricated, with variations for each surface treatment. All shear bond specimens were prepared using the same bonding procedure. Zirconia surfaces were modified (see below for modification techniques) and treated with an organosilane (Monobond-S, Ivoclar-Vivadent, Schaan, Liechtenstein) prior to resin cement bonding. Composite cylinders were coated with resin cement (Rely-X Unicem, 3M-ESPE, St. Paul, MN), placed on the zirconia surface and UV-light cured under a defined load (5 N):

• Groups 1 and 2 ( control ): (1) Polished, untreated surface and (2) roughened, untreated surface.

• Groups 3 and 4: Surfaces were polished (3) and roughened (4) were modified with a 3 nm Si _x O _y layer, this procedure is described in detail in Piascik et al. .

• Group 5: Zirconia surfaces were silica-coated using 30 μm alumina particles modified with salicylic acid (CoJet, 3M-ESPE, St. Paul, MN – 0.28 MPa, 5–10 mm working distance, 15 s).

• Groups 6 and 7: Zirconia surfaces were polished (6) or roughened (7) were exposed to the fluorination process described above.

Shear bond test specimens were stored in DI water at 37 °C for a period of 24 h prior to testing. Specimens were then fixed to a custom vise fixture to ensure vertical compliance. All specimens were subjected to a uniform force (semi-circular, notched crosshead fitting the diameter of the composite cylinder) at a crosshead speed of 0.5 mm/min in an electro-mechanical testing device (Instron Corp., Norwood, MA). Shear bond strengths were calculated by dividing peak load by the cross-sectional area of the composite cylinder. Single-factor analysis of variance (ANOVA) at a 5% confidence level was performed for the bonding strength data. Optical microscopy and scanning electron microscopy (SEM) were used to evaluate and quantify failure surfaces.