Bonding of resin cements to an aluminous ceramic: A new surface treatment

Bonding of resin cements to an aluminous ceramic: A new surface treatment

Dent Mater 10:185-189, May, 1994 Bonding of resin cements to an aluminous ceramic: A new surface treatment Michael S a d o u n ~, E r i k A s m u s ...

856KB Sizes 0 Downloads 3 Views

Dent Mater 10:185-189, May, 1994

Bonding of resin cements to an aluminous ceramic: A new surface treatment Michael

S a d o u n ~, E r i k A s m u s s e n


1Groupe de Recherche Biomatdriaux, Facultd de Chirurgie Dentaire, Universitd Paris V, Paris, FRANCE 2Department of Dental Materials, School of Dentistry, Faculty of Health Sciences, University of Copenhagen, DENMARK

ABSTRACT Objectives. The purpose of the work reported here was to develop a surface treatment of an alumina-based ceramic (In Ceram) that would make reliable bonding to a resin-based luting agent possible. Methods.The surface treatment studied was the application of a suspension of a fine-grained, refractory powder, which after drying was sintered to the surface at 960°C. The adherence potential of the surface was determined by measurement of bond energy. Results. It was found that the surface treatment, in conjunction with a heat-treated, silane coupling agent, resulted in mean bond energies of 47 (_+19), 56 (_+22), and 525 (+116) J/m 2 for the three resin cements studied. Significance. It was concluded that the new surface treatment makes reliable bonding possible, which may allow new indications for this material. INTRODUCTION An increasing demand for esthetic restorations during the last decade has led to the development of several new ceramic systems. One ofthese is In Ceram (Vita Zahnfabrik, Sackingen, Germany), a high-strength ceramic based on a skeleton of sintered aluminum oxide infused with glass (Tyszblat, 1988). This aluminous ceramJc is normally employed as core material in conjunction with a more translucent ceramic, and thus replaces less esthetic metal cores. A core of aluminous ceramic may be luted with phosphate cement or with glass ionomer cement. However, when the retentive area is small, retention may be inadequate (Kaufmann et al., 1961) and an adhesive luting agent desirable. Ceramic restorations such as inlays, onlays, and crowns are commonly adhesively bonded with resin-based luting agents. Some types of ceramic material have been found to be fracture prone (Erpenstein and Kerschbaum, 1991; Stenberg and Matsson, 1993). Increased resistance to fracture may be obtained with high-strength aluminous ceramic (Seghi et al., 1990) or by means of a "sand-

wich" using the aluminous material as a reinforcing base. Adherence of the resin cement to the tooth side of the luted joint is ensured by acid etching of enamel and/or dentin, and by use of a dentin adhesive. Adherence to the ceramic side of the joint is normally obtained by etching of the ceramic with hydroiluoric acid. By dissolution of the glassy phase, the etching of the ceramic creates a rough surface, favoring adherence relying on mechanical interlocking (Edris et al., 1990). The adherence may be improved by application of a silane coating to the etched surface (Barley and Bennett, 1988; Sheth et al., 1988). With the aluminous ceramic (In Ceram, Vita Zahnfabrik), the main part of the surface is an acid-resistant aluminum oxide. Thus, hydrofluoric acid does not appreciably roughen the surface so bending to a resin cement is minimal (Pape et al., 1991). This lack of adherence may limit the use of an otherwise satisfactory ceramic system. It was the aim of this study to develop a surface treatment for an aluminous ceramic that could be used to obtain a reliable bond between the ceramic and a resin cement. A further aim was to test ffthe bonds obtained were affected by the type of resin cement.

MATERIALS AND METHODS Micro-mechanical retention requires a rough surface but as mentioned above, acid etching was not effective on the investigated aluminous ceramic. Therefore, instead of a differential removal of substance from the surface by acid etching, an alternative route for enhancing roughness was to add a substance to the surface. The new surface treatment was based on the idea that it should be possible to sinter a fine-grained, refractory powder to the surface of an aluminous ceramic, and thus create considerable surface roughness. It was essential that the powder be fine-grained so that it would not interfere with the seating of the restoration. The powder was to be applied as an ethanolic suspension. The adherence potential of the surface treatment was assessed by the wedge test (Cognard, 1986). The wedge test gives information on Dental Materials/May 1994 185

"rhBLE 1: MATERIALSUSED Manufacturer

Batch No.

Silane Porcelain Primer

Bisco, Downers Grove, IL, USA 099200


Kulzer, Wehrheim, Germany

A: 28 B: 904026

Hydrofluoric acid Porcelain Etchant Bisco


Resin Cement All-Bond C & B



Sun Medical, Japan

Uni: 119280

Cat.: 119080 Monomer: 20603

Cat.: 210021 Polymer: 20101 Twinlook


Base: 94.06.30 025

Cat.: 93.04 027

bond energy, and was chosen instead of measurements of bond strength because of recent criticism of the latter type of measurements (Van Noort et al., 1991). Also, evidence has been put forward that bend energy may be more relevant for the survival of bonded restorations than bend strength (Degrange et al., 1994). The wedge test makes use of two identical beams that are joined together with an adhesive. A box-shaped mold, or rather sleeve, of a vinyl polysiloxane impression material (President, Coltene, Altst~itten, Switzerland) was produced by taking an impression of a beam of acrylic resin having dimensions of 80 mm x 12 mm x 20 mm. After the impression material had set, the sleeve was poured with gypsum. After the setting and drying of the gypsum, the gypsum block was pushed partly out of the sleeve of impression material to form the floor of a box-shaped mold of the following dimensions: length = 80 mm, width = 12 mm, and height = 6 mm. The specimens of aluminous ceramic were then produced by slip casting as described by the manufacturer. Powder, liquid, and antiflocculant were mixed, poured into the mold, and left to harden by water evaporation. The specimen was carefully separated from the mold and heated to 1120°C in accordance with the recommended heating procedure. After cooling, the sintered skeleton of aluminum oxide was ground on 320 grit carborundum paper to dimensions of ~ 80 mm x 5 mm x 4 mm. The specimen was then impregnated with the glass of the system at 1080°C and after cooling, excess glass was removed by grinding on a 40~m diamond plate. In this manner, 24 beams were produced. The beams were then divided into 12 pairs, and ground pairwise on the diamond plate to a width of about 5 mm and a thickness of about 3.8 ram. The deviation in width and thickness of the two beams of a pair never exceeded 0.06 mm, and mostly was less than 0.03 mm. The adherence potential of several different surface states was investigated. The surface state was the result of either no further treatment other than grinding on the diamond plate, or resulted from a combination of the following surface 186 Sadoun & Asmussen / Bonding to aluminous ceramic

treatments: 1) grinding on a 40 pm diamond wheel; 2) application of silicone resin (Dow Coming, Brussels, Belgium, 62230 resin, 0.5 g in 50 g ethanol). A~er evaporation of the ethanol, the treated specimen was heated to 960°C for 0.5 h; 3) application of a silane coupling agent (Table 1). The silane was either left to dry at room temperature, or heated for 2 min with a hair dryer. The temperature attained with the hair dryer was estimated as 90-95°C using a thermocouple; 4) application of hydrofluoric acid (Table 1) for 2 min, followed by rinsing with water and drying with the hair dryer; 5) application of a suspension of silica. The suspension was composed of 2 g Aerosil 380 (Degussa, Frankfurt, Germany), 0.09 g Beigostat phosphate diester antiflocculant, 0.1 g of the above silicone resin, and 40 g of ethanol. According to the manufacturer, the powder of Aerosil 380 has a mean particle size of about 7 nm. After evaporation of the ethanol, the treated specimens were heated to 960°C for 0.5 h. The concentration of Aerosil in ethanol was selected ai~r a number of pre-experiments in which different suspensions were applied to glass slabs and the ethanol evaporated. The concentration of the suspension selected was found by visual inspection to leave a smooth, homogeneous layer on the surface of the glass. The application of treatments 2 - 5 was performed with a small brush. The combination of treatments investigated appears in Table 2. After surface treatment, a pair of beams was joined with one of the resin cements in Table 1. For this purpose, a specimen holder was constructed, which enabled a parallel alignment of the beams. The mixed resin cement was applied to both of the bonding surfaces, and using a micrometer screw, the beams were pressed against each other until a film thickness of 100 pan was attained. The assembly was then left for 10 min at ambient temperature, after which time the joined beams were disengaged from the specimen holder and placed at 36 _+ 2°C for at least 24 h. Excess resin cement was removed by grinding with 500 grit carborundum paper, exposing the joint in its full length at both sides of the specimen. To produce a fissure, a steel wedge of thickness between 375 and 420 ]am was introduced in the joint by means of another specimen holder, as described previously (Asmussen et al., 1993). The separated assembly was then placed in water for at least 24 h, at which time the fissure had reached its full, equilibrium length (Cognard, 1986; Asmussen et al., 1993). The specimen was then removed from the water, dried, and the length (l) of the fissure determined in a stereo microscope at magnification 40x as earlier described (Asmussenet al., 1993). The separation (d) of the beams caused by the wedge was determined in a microscope fitted with a measuring occular at a magnification of 50x. The thickness (t) of the beams had been measured with a micrometer screw before surface treatment. The energy of adherence Wwas calculated (Cognard, 1986) as: 3E t 3 W •d 2 (1) 16 " 14 where/, t, and d are defined above, and E is the modulus of elasticity of the beam material, assumed to be 300 GPa. To assist in the interpretation of the results, the beam surfaces were studied in the scanning electron microscope. The number of measurements carried out for each combination of surface treatment and resin cement appears in Table 2.












Mean + S.D. J/m 2 (n)


<1o (4)*

<10 (4)*

+ +


<10 (4)*



<10 (4)*

154 _+121 (8)

Calcined silicone

<10 (4)*

resin +

<1 o {4)*



<51 _+ 40 (8) +

47 __.19 (7)

573 + 194 (3)

56 + -22 (8)

525 + 116 (8) (n) = number of specimens. * In all specimens tested, the fissure propagated the entire length of the specimen. ÷ In 3 specimens, the fissure propagated the entire length of the specimen.

After the measurements, resin cement and organic material were removed by burnout at 500°C. The surfaces were then reground on the diamond disc, the thickness of the specimens measured again, and a new series of measurements commenced. In a previous study (Asmussen et al., 1993), it was found that the standard deviations of measured energies of adherence could not be assumed to be the same for all groups. However, by using In W as a transformed variable (Hald, 1952), standard deviations were obtained that might be considered identical. In the present study, the same transformation was performed, enabling the use of analysis of variance and the student t-test as statistical tools (Hal& 1952). RESULTS The results of the measurements are presented in Table 2. In a number of cases, the fissure propagated the entire length of the specimen. The energy of adherence for such a case is less than about 10 J/m 2, which corresponds to a fissure length of 80 ram. It can be seen from the table that for surfaces that were only ground or silane treated (Porcelain Primer) without heat, adherence was not present (Experiments 1 and 2). Ground specimens silanated with heat gave an energy of adherence of 154 _+ 121 J/m 2 with Superbond, but energies lower than 10 J/m 2 with All-Bond and Twinlook (Experiment 3). A student t-test showed that the adherence of the Superbond was significantly different from that of All-Bond or Twinlook (p < 0.01). The relatively large standard deviation (coefficient of variation = 79%) may reflect an inherent unreliability of

the pretreatment. Etching with hydrofluoric acid (Porcelain Etchant) did not improve bond energies in a measurable way (Experiment 4). Surface treatment with calcined silicone resin resulted in bends of low energy (Experiments 5 and 6). Treatment with the suspension followed by silanization (Porcelain Primer) without heat (Experiment 7) resulted in a bend that was not reliable in 3 specimens. The fissure propagated the entire length of the specimen, and the standard deviation was relatively large (coefficient of variation = 79%). In contrast, treatment with suspension followed by a heat-treated silane (Experiments 8 and 9) gave rise to reliable adherence. The standard deviations were relatively small (coefficients ofvariation = 22 - 40%), and the statistical analysis showed that the adherences were significantly different from zero. No difference was measured in the efficacy of the silanes (Porcelain Primer and Silicoup) (Experiments 8 and 9, p > 0.40). Superbend gave rise to an energy of adherence that was superior to those of All-Bond a n d Twinlook (Experiment 8, p < 0.001). DISCUSSION Only a few of all the possible combinations of Table 2 were investigated. The reason for this was that it did not seem reasonable to use time and effort on combinations which according to the first series of experiments could be expected to give poor bending. When more effective pretreatments were discovered in later series ofexperiments, the decision was made to concentrate on these combinations. In the measurements, the 12 pairs of beams were re-used several times. It may be argued that the results are not stoDental Materials~May 1994 1 8 7

Fig. 1. Surface of a beam of aluminous ceramic after drying and sintering of the adhesion-promoting suspension. Numerous cracks in the coating may be observed.

Fig. 3. Surface of a beam of aluminous ceramic after drying and sintering of the adhesion-promoting suspension. The contribution to the roughness of individual particules of the suspension may be observed.

chastically independent, and that some of the statistical analyses (comparisons between groups) are less reliable. However, inspection of the results showed no particular pattern in the behavior of any of the 12 pairs of beams. On this basis, it was assumed that a possible beam effect was negb'gible. The present study has shown that it is possible to impart the investigated aluminous ceramic with a surface that is amenable to bonding. Fig. 1 shows a surface after treatment with the suspension of silica, and sintering. A surface of considerable roughness has been produced, and bending based on mechanical interlocking is easily imaginable. It may be noted that at the magnification of Fig. 1, it is not the individual grains of the suspension that contribute to the roughness, but rather the cracking of the drying surface of the applied suspension. At higher magnification (Fig. 2), the roughness present in the walls of the cracks is discernible, and at still higher magnification (Fig. 3), the contribution to the roughness of individual, sintered particles can be observed. The suspension of silica was sintered to the surface of the aluminous ceramic at a temperature of 960°C. This is close to 188 Sadoun & Asmussen / Bonding to aluminous ceramic

Fig. 2. Surface of a beam of aluminous ceramic after drying and sintering ofthe adhesion-promoting suspension. The roughness in the walls of the cracks may be observed.

the firing temperature of feldspathic porcelains commonly employed as veneering material. In such cases, it may seem advisable to apply and sinter the suspension before a veneering of the aluminous core. However, at the present time, it is not known how later multiple firings for feldspathic porcelain may affect surface morphology and bond energies. This may be a topic for future research. The positive effect of the silane treatment is in agreement with earlier studies (Barley and Bennett, 1988; Sheth et al., 1988). The effect of heat treatment of the silanated ceramic surface is probably twofold: first, it ensures a complete evaporation of solvent from the minute surface pores seen in Fig. 3; and second, it enhances the efficacy of the silane, as suggested by manufacturers oforgano-functional silanes (HtilsTmisdorf: 1992). The unmeasurable influence on bonding of hydrofluoric acid corroborates earlier studies (e.g., Pape et al., 1991). In the present study, bond energy has been determined, while more commonly, ceramic-polymer bonds are described in terms of bend strength. Bond strength and bond energy have been found to be poorly correlated (Degrange et al., 1991). In a study on bonding to sandblasted metal, energies of adherence were found in the range 24-121 J/m 2, depending on the resin cement (Asmussen et al., 1993). In the case of sandblasted metal, bond strengths in the range of 13-38 MPa have been found, depending on the method and on the resin cement (Atta et al., 1988; Garcia-Godoy et al., 1991). However, although bond energies were found to be 2-5 times higher to the treated aluminous ceramic than to sandblasted metal, it seems unlikely that the bond strength to the aluminous ceramic should also be 2-5 times higher. The lack of correlation between energy of adherence and bond strength may, in part, be due to difficulties in obtaining true values of bond strength, as discussed by van Noort (1991). In many cases, an unknown blend of adhesion, mechanical properties ofthe resin composite, and shape factors enter into reported values ofbond strength. As mentioned in the introduction, data have been put forward that bond energy may provide a better measure of clinical predictablility than does bond strength (Degrange et al., 1994). The difference in bond energies between Superbond and Twinlook has been demonstrated in an earlier study (Asmussen


Fig. 4. Surface of a beam of aluminous ceramic along which a fissure has propagated. Remnants of the cement may be observed in the cracks of the sintered suspension.

et al., 1993) and is most probably linked to the viscoelastic properties of Superbond. That mechanical properties of the resin cement affect the measured bond energies is supported by Fig. 4, showing a beam surface along which a fissure has propagated. The cement has fractured, and remnants of the cement may be seen in the cracks of the sintered suspension. The high bond energies of Superbond, even in equilibrium with an aqueous environment, contrast with several reports on bond strength in which the water uptake has been found to give rise to reduced bond strengths (Marx, 1987; Pfeiffer, 1989). Marginal integrity is an important parameter of cemented restorations. The adhesive surface developed in the present study was formed by addition and sintering of a refractory powder to the bonding side of the aluminous ceramic. It is well known that even small dimensional changes to a prosthesis may cause large effects in marginal seating. Inspection in the SEM showed that the thickness of the sintered silica did not exceed 10 pm. Marginal openings of cemented ceramic restorations have been the subject of several studies (e.g., Pfeiffer et al., 1993; Sorensen et al., 1990). With the aluminous ceramic investigated in the present study, marginal openings 24 - 67 ~m were reported. At present, it is not known how a 10 ~ n layer of sintered silica will affect marginal fit. It is conceivable that the dimensional change of less than 10 pzn can be accommodated for by proper die spacing technique (Sorensen et al., 1990). It may be concluded that the adhesive surface treatment developed with the present work may lead to a significant increase in the applicability of the material. Received August 2, 1993/Accepted April 22,1994 Address correspondence and reprints to: E. Asmussen Department of Dental Materials School of Dentistry Faculty of Health Sciences Universityof Copenhagen NSrre All~20 DK-2200 Copenhagen N DENMARK

Asmussen E, Attal J-P, Degrange M (1993). Adherence of resin-based luting agents assessed by the energy of fracture. Acta Odontol Scand 51:235-240. Atta MO, Brown D, Smith BGN (1988). Bond strength of contemporary bridge cements to a sandblasted or electrolyrically etched nickel-chromiumalloy Dent Mater4:201-207. Bailey LF, Bennett RJ (1988). Dicor surface treatments for enhanced bonding. J Dent Res 67:925-931. Cognard J (1986). Theemechanics of the wedge test. JAdhesion 20:1-13. Degrange M, Edard V, Toumelin JP (1991). Comparison of two testing methods for resin bonded systems. JDent Res 70:741, Abstr. No. 868. Degrange M, Chattier J-L, Attal J-P, Asmussen E (1994). Adherence of luting cements for resin-bonded bridges. J Dent Suppl 1:22:$28-$32. Edris AA, Jabr A ~ Cooley RL, Barghi N (1990). SEM evaluation of etch patterns by three etchants on three porcelains. JProsthetDent 64:734-739. Erpenstein H, Kerschbaum T t1991). Frakturrate von Dicor% Kronen unter klinischen Bedingungen. Dtsch Zahn~rztl Z 46:124-128. Garcia-Godoy F, Kaiser DA, Malone WFP, Hubbard G (1991). Shear bond strength of two resin adhesives for acid-etched metal prostheses. JProsthetDent 65:787-789. Hald A (1952). Statistical theory with engineering applications. New York: John Wiley & Sons. Hills Troisdorf(1992). Dynasilan. Adhesion promoters. Organo-functional silanes. Technical information. Hills Troisdorf &G. D-5210 Troisdo~ Germany. Kaufmann EG, Coelho DH, Colin L (1961). Factors influencing the retention of cemented gold castings. J Prosthet Dent 11:487-502. Marx R (1987).Ursachen ffir das Versagen von Klebeverbindungen beiAdh~isivbrficken. Schweiz Monat Zahnmed 97:1081-1086. Pape F-W, Pfeiffer P, Marx R (1991). Haftfestigkeit von ge~itztem In-Ceram® an Zahnschmelz. Zahn Welt und Rundschau 100:450-453. PfeifferP (1989). Spezifi~he adh~rente Klebung abgestrahlter Legierungsoberfl~ichen. Dtsch Zahnaerztl Z 44:936-937. Pfeiffer P, Sommer MC, KnofR (1993). Zur Eignung des InCeram®-Systems ffir Inlays.Dtsch Zahn~rztl Z 48:419-422. Seghi RR, Sorensen JA, Engelman MJ, Ronmanas E, Torres TJ (1990). Flexural strength of new ceramic materials. J Dent Res 69:299, Abstr. No. 1521. Sheth J, Jensen M, Tolliver D (1988). Effect of surface treatment on etched porcelain bond strength to enamel. Dent Mater 4:328-337. Sorensen JA, Torres TJ, Kang SK, Avera SP (1990). Marginal fidelity of ceramic crowns with different margin designs. J Dent Res 69:279, Abstr. No. 1365. Stenberg R, Matsson L (1993). Clinical evaluation of glass ceramic inlays (Dicor). Acta Odontol Scand 51:91-97. Tyszblat M (1988). Process for the preparation of a dental prosthesis by slight solid phase fritting of a metal oxide based infrastructure. U.S. Patent 4,772,436. Van Noort R, Cardew GE, Howard IC, Noroozi S (1991). The effect of local interracial geometry on the measurement of the tensile bond strength to dentin. JDent Res 70:889-893. Dental Materials~May 1994 189