- Email: [email protected]
Fatigue testing of enamel bonds with self-etch and total-etch adhesive systems
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
available at www.sciencedirect.com
journal homepage: www.intl.elsevierhealth.com/journals/dema
Fatigue testing of enamel bonds with self-etch and total-etch adhesive systems Robert L. Erickson a,∗ , Anton J. De Gee b , Albert J. Feilzer b a b
Creighton University, School of Dentistry, Omaha, NE, USA Academic Center for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands
a r t i c l e
i n f o
a b s t r a c t
Article history:
Objective. Etching of enamel by self-etching adhesive systems is not as pronounced as with
Received 3 May 2005
phosphoric acid employed with most total-etch adhesive systems. This may result in differ-
Received in revised form
ences in the effectiveness of the bonds for the two types of systems. The aim of this study
22 September 2005
was to compare two such systems by bond strength and fatigue testing.
Accepted 2 November 2005
Materials and methods. Shear bond strengths for Single Bond (SB) and Adper Prompt-L-Pop (PLP) were obtained on bovine enamel surfaces. Fatigue measurements were made with the same test fixtures using cyclical loading at 2 Hz for a maximum of 105 cycles or until
Keywords:
the bond failed. Four selected loads between 40 and 70% of the shear bond strength were
Dental materials
tested for each adhesive and the number of cycles to failure was recorded. S–N curves were
Enamel
constructed from the data and fatigue stress limits were determined. The Mann–Whitney
Adhesion
U-test and t-tests were used for statistical comparisons of the results. SEM analysis of resin
Fatigue
tag formation into enamel surfaces was carried out. Results. A significantly greater bond strength was found for SB (25.3 MPa) than for PLP (19.2). Sustainable stresses after 105 load cycles were substantially lower than the corresponding shear bond strengths and fatigue limits were found to be 8.4 MPa for PLP and 14.6 MPa for SB. The ratio of fatigue limit to bond strength was less for PLP (44%) than for SB (57%). Resin tag penetration into enamel was substantially less for PLP than for SB. Significance. Fatigue testing in conjunction with bond strength testing can provide a better means for assessing the performance of adhesive systems used for bonding to enamel. © 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
A typical, in vitro, evaluation of the effectiveness of an adhesive system involves measurement of bond strengths to enamel or dentin by a shear or tensile test. Such tests are useful for evaluating differences in performance between materials but the results are difficult to relate to in vivo effectiveness. Bond strength tests are dynamic tests involving the application of a monotonically increasing force to the bond until it fails, giving an indication of its strength. This, however, is not
∗
a likely mode of failure for bonds in the mouth, where failure is considered to result from repeated loading over many months or years and at stresses well below the ultimate bond strength. This suggests that fatigue studies, where cyclic loading of bonded specimens is evaluated, may provide better insight to in vivo performance and give more realistic values of sustainable stresses. Testing by cyclic loading is frequently referred to as fatigue testing and a review of such testing for dental materials has recently been published [1]. There are only a small number
Corresponding author at: Creighton University, School of Dentistry, 6101-Lynn Way, Woodbury, MN 55129, USA. Tel.: +1 651 459 7962. E-mail address: [email protected] (R.L. Erickson).
0109-5641/$ – see front matter © 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2005.11.021
982
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
of studies where cyclic loading has been used to examine the effect on adhesive bonds to enamel or dentin directly. Specifically for bonds to enamel, one study [2] developed S–N curves for a three-step total-etch adhesive system and projected a fatigue limit of about 41% of the shear bond strength. S–N curves provide a relationship between applied stress and the logarithm of the number of cycles at which failure of the bond occurs. A more popular method of fatigue testing, referred to as the staircase method, involves selecting a starting stress of about 50–60% of the ultimate strength and an upper limit for the number of cycles, and tests one specimen until it fails or survives. If it fails, the stress is decreased by a set amount for the next specimen but if it survives, the stress is raised by that same set amount. Continuing in this manner for a number of specimens the test focuses on a stress that is likely to produce 50% failures. This value is termed the fatigue limit for the number of cycles selected. A statistical analysis applicable to staircase testing has been described [3]. Two studies, using a staircase approach, have examined fatigue of adhesive bonds to enamel [4,5] and found fatigue limits of 37–46% of the respective bond strengths. These studies point out a common finding of fatigue studies, i.e. fatigue limits are much lower than the typically measured bond strengths and may be in the order of 40–60% of that strength. Also, materials can show different fatigue behavior and a material with a greater bond strength may have a lower fatigue limit than a comparative material [5]. Such differences in fatigue performance may be important for understanding failures in vivo. In recent years, self-etching adhesives have become popular due to their purported ease of use. However, they generally do not etch enamel as strongly as phosphoric acid, which is typically used with total-etch adhesive systems. Some recent studies on bonding to enamel have found that self-etching adhesive systems, whether two-step or one-step, have inferior bond strengths compared with total-etch systems used as control materials [6,7]. In addition, a pre-etch of enamel with phosphoric acid was shown to improve the bond strengths of two-step adhesives compared to their normal application method [8]. These studies suggest that inadequate etching may be at least part of the reason that self-etching adhesive systems do not achieve the same level of performance on enamel as total-etch systems. The purpose of this study was to test the hypothesis that the two-step, total-etch adhesive system, Single Bond (SB; 3M
ESPE, St. Paul, MN, USA), can provide better bond strength and fatigue strength when bonded to enamel, than the one-step, self-etch adhesive, Adper Prompt-L-Pop (PLP; 3M ESPE, St. Paul, MN, USA). A secondary goal was to construct S–N curves to see if the fatigue behavior differs for the two adhesive systems.
2.
Materials and methods
2.1.
Enamel specimen preparation
Cylindrical cores of enamel and attached dentin, 6.0 mm diameter were cut from bovine incisors, normal to the labial surface, using a hollow core diamond drill with copious water cooling (Diamant Boart Inc., Vianen, The Netherlands). The fabrication procedure for each test specimen started with etching the enamel along the cut cylindrical wall of a core with 35% phosphoric acid, followed by rinsing and drying. The etched enamel was coated with a bonding agent (Single Bond; 3M ESPE, St. Paul, MN, USA), air dried and light-cured. The core was inserted into a 6.0 mm hole in a 2.0 mm thick, stainless steel plate (Fig. 1a), with the enamel surface above the plate by about 0.5 mm. A light-curable, flowable composite (Tetric Flow, Ivoclar Vivadent, Schaan, Liechtenstein) was used to cement the core into place. The exposed enamel surface was ground, by hand, on 240 grit abrasive paper with water cooling, until it was flush with the stainless steel plate. This ground surface was treated with one of the two adhesive systems being investigated as described in Table 1. Manufacturer’s instructions were followed except for doubling the light curing time to 20 s to ensure good cure of the adhesives. Next, the stainless steel plate, with the enamel specimen, was placed on a pin-indexing jig and a second matching plate was placed over it (Fig. 1b). The pins helped position a 4.0 mm hole, in this second plate, over the center of the enamel surface. A resin-composite (Z 250; 3M ESPE, St. Paul, MN, USA) was condensed into this cylindrical mold in two 1.0 mm increments, each light-cured for 20 s. The pins were then pushed out to allow transfer of the completed test specimen to a test fixture. This fixture consisted of a cylindrical cavity in a brass block, with shallow, diametrically opposed grooves that were machined to precisely fit the two-plate specimen assembly (Fig. 1c). This allowed free vertical movement of the assembly with negligible lateral movement. In this manner specimens were made, with each adhesive, for both shear bond strength testing and fatigue testing. Bonded specimens were stored for
Table 1 – Investigated adhesive systems and procedures Adhesive system
Code
Etchant
Procedures
Single Bond (lot# 3HK)
SB
35 wt.% phosphoric acid
1. Etch enamel 15 s; rinse; dry thoroughly with air 2. Apply two consecutive coats of adhesive and dry thoroughly with air 3. Light-cure 20 s
Adper Prompt-L-Pop (lot# 135864)
PLP
Self-etch
1. Mix adhesive in package according to instructions 2. Apply adhesive with a rubbing action for 15 s 3. Dry thoroughly with gentle air stream 4. Apply second coat without rubbing and dry thoroughly 5. Light-cure 20 s
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
983
Fig. 1 – Specimen test fixture shown with pin-indexing jig for facilitating assembly. (a) Exploded view, with lower plate showing enamel specimen in 6.0 mm hole and upper plate having a 4.0 mm hole for composite. (b) Assembled view. After applying adhesive to enamel the lower plate is placed on pin-indexing jig and then the upper plate is positioned on top with hole centered on enamel specimen. Composite is condensed into this hole and light-cured to complete the bond. Pins are pushed out to allow transfer of the test specimen. (c) Test fixture consisting of a cylindrical cavity in a brass block, with shallow, diametrically opposed grooves to precisely fit the two-plate specimen assembly.
24 h in distilled water at 37 ◦ C after the bonding procedures were completed and were maintained under those conditions throughout the bond strength and fatigue testing procedures described below.
2.2.
Shear bond strength testing
The specimen assembly when placed in the brass test fixture was seated to the bottom of the cavity with the end of the second plate protruding out of the cavity. By applying a slowly increasing force onto the end of this plate, a shear force was exerted on the bonded interface until total bond failure occurred. Ten specimens were tested for shear bond strength (SBS) in an Instron testing apparatus (model no. 6022; Instron, High Wycombe, Bucks, UK) at a cross-head speed of 1.0 mm/min, for each adhesive system.
2.3.
Cyclic load testing
Load cycling of bonded specimens was carried out with the same test configuration as SBS testing. A twelve-station fatigue testing apparatus utilizing air pressure to regulate and deliver the force was employed for load cycling and has been described in another publication [9]; a photograph and schematic diagram are shown in Fig. 2. For each station, a test fixture was positioned so that a flat-ended rod, driven by the piston delivering the load, impinged on the end of the protruding plate of the specimen assembly. The brass text fixture sat on the metal base of a custom water bath. Each station had a
built-in load cell and separate controls for adjusting the maximum and minimum loads for each cycle. A counter on each station recorded the number of cycles elapsed and the station shut off automatically when bond failure occurred. The load was applied at a rate of 2 Hz with a 50% duty cycle and took the form of a damped square wave, there being a 40 ms time constant in switching between load settings. The load was cycled between a fixed lower setting of 12 N, an arbitrary value used to maintain some load at all times, and preset higher loads ranging from 40 to 70% of the SBS determined for each of the adhesive systems. The range of loads was estimated to be adequate for determining S–N curves from previous unpublished fatigue studies and from published studies [2,4,5], the goal being to cover the range where specimens mostly fail to mostly survive. Cyclic loading was carried out to 105 cycles or until the specimens failed by total loss of the bond and the number of cycles at failure were recorded for each specimen. There were 15–17 specimens tested at each load value for each adhesive. Specimens that survived 105 cycles were tested for SBS by the methods described earlier.
2.4.
SEM analysis
Cylindrical cores of bovine enamel, polished as already described, were used to evaluate the degree of resin tag penetration into the enamel surface during the bonding process. The procedures used to bond the surface were identical to those used for other test procedures already described, except that the enamel specimens were not cemented into stainless
984
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
Fig. 2 – Left: The ACTA fatigue tester. Right: Schematic diagram of a single station with specimen depicted in place.
steel plates. Four bonded specimens for each adhesive system were immersed in 50 vol.% nitric acid until the enamel was fully dissolved and the resin interfaces exposed. These specimens were rinsed with distilled water and dried gently with an air stream. They were then mounted on SEM stubs and sputter-coated with gold. SEM observations were made on these specimens and representative micrographs were recorded.
3.
in Fig. 3 and Tables 3 and 4, where the load values in Newtons have been converted to Megapascals. Fig. 3 shows S–N data points associated with the minimum, median and maximum number of cycles before bond failure for each load level tested. The connecting lines through the data points provide approximate S–N curves for those same parameters. The minimum and maximum curves form an envelope that encompasses all of the data and the median curve gives an indication of how
Results
Results for shear SBS testing are shown in Table 2, expressed in Megapascal units, using the failure load in Newtons and the area of the hole (mm2 ) in which the composite was placed as the nominal bond area. A statistically significant difference was found between the mean bond strengths of the two adhesive systems (SB: 25.3 MPa; PLP: 19.2 MPa), using a t-test (p = 0.01). Load levels chosen for cyclic testing were determined from the above SBS values. For SB they were 165, 187.5, 210 and 232.5 N and for PLP the load levels were 100, 120, 140 and 165 N. It should be noted that 165 N was the only common load for the two adhesive systems. Results for cyclic loading are shown
Table 3 – Summary of the cyclic loading results for SB showing minimum, median and maximum number of cycles survived by specimens at four stress levels # Cycles Min Med Max % Survivors
13.1 (MPa)
14.9 (MPa)
16.7 (MPa)
1.5 × 104 105 105 69
2.2 × 103 2.6 × 104 105 25
1 2.7 × 103 105 13
18.5 (MPa) 1 80 6 × 103 0
The percentage of specimens surviving 105 cycles is shown at the bottom for each stress level.
Table 4 – Summary of the cyclic loading results for PLP showing minimum, median and maximum number of cycles survived by specimens at four stress levels Table 2 – Mean shear bond strength of adhesive systems evaluated Shear bond strength (MPa) (S.D.) SB PLP
25.3 (5.6)a 19.2 (3.2)b
Different superscript letter indicates a statistically significant difference (p < 0.01; t-test).
# Cycles
8.0 (MPa)
Min Med Max % Survivors
4.8 × 10 105 105 80
3
9.6 (MPa) 2 2 × 103 105 20
11.2 (MPa)
13.1 (MPa)
1 2.9 × 102 2.9 × 103 0
1 10 1.9 × 103 0
The percentage of specimens surviving 105 cycles is shown at the bottom for each stress level.
985
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
Fig. 3 – S–N curves for the minimum, median and maximum number of cycles until failure, for composite bonded to enamel, using two test adhesives. The dashed arrows signify that these data points are for specimens that survived and could have gone to larger numbers of cycles. A maximum of 105 cycles were used and four stress levels for each adhesive material. Fifteen to seventeen specimens were used at each stress level. (a) Single Bond, (b) Adper Prompt-L-Pop.
Fig. 4 – Representative SEM micrographs of resin tags, exposed by dissolving enamel from bonded specimens; 2000× magnification. (a) Single Bond, (b) Adper Prompt-L-Pop.
the data is distributed within that envelope. From the above data, it can be observed that there is no overlap of data at the 13.3 MPa load level and this immediately indicates that the two adhesives have a statistically significant difference in their fatigue performance at that level, as confirmed by a Mann–Whitney U-test (p = 0.01). Further, using the proportions of failures after 105 cycles (Tables 3 and 4) in a calculation of the mean shear fatigue stress (SFS) [3], the values shown in Table 5 were determined. A statistically significant difference was found between the SFS values for the two adhesives using a t-test (p = 0.01). The ratio of SFS to SBS was found to be smaller for PLP (44%) than for SB (57%).
Table 5 – Comparison of shear bond strength (SBS) and shear fatigue strength (SFS) and the ratio of these two strengths for the two adhesive systems evaluated
SB PLP
SBS (MPa)
SFS (MPa)
SFS/SBS (%)
25.3 (5.6)a 19.2 (3.2)b
14.6 (1.6)a 8.4 (0.5)b
57 44
Different superscripts in columns indicate statistically significant differences (p < 0.01; t-test).
Table 6 – Comparison of initial shear bond strengths with shear bond strengths of specimens surviving cyclic loading
SB PLP
Initial strength (MPa)
Strength of survivors (MPa)
25.3 (5.6)a 19.2 (3.2)b
27.1 (5.0)a 20.6 (4.2)b
Different superscript letters indicate a statistically significant difference (p = 0.01).
Table 6 compares the initial SBS values to the SBS for all specimens that survived 105 cycles, regardless of load. No statistically significant differences were found using a t-test (p = 0.05). Fig. 4 shows representative images of resin tags from SEM micrographs, taken at 2000× magnification, for the two adhesives. It is clear that the tags are longer and more extensive for the SB adhesive than for the PLP adhesive.
4.
Discussion
Bond strength tests are the most common way of evaluating the effectiveness of an adhesive system. In this study
986
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
SBS were determined for the two adhesive systems, both as a direct comparison of their effectiveness in this type of test, and also to provide baseline values with which the SFS values could be compared. A statistically significant difference between the SBS values of SB and PLP was found, with the value for PLP being 76% of that for SB, which supports that part of the original hypothesis for this study. Similar results were found in another study of the same materials [7] where the ratio was 73%. In that study, five self-etch systems were compared to SB and four of the five, including PLP, were found to have lower SBS compared to SB. In a study where microtensile bond strengths were determined [8], three out of four selfetch systems tested, one being PLP, were lower in SBS than for a three-step, self-etch adhesive control material, Optibond FL (Kerr). In this case the SBS ratio was 45%. The possibility that insufficient etching may reduce the SBS of self-etch adhesive systems has been tested by including a pre-etch of enamel with phosphoric acid before applying the self-etching adhesive system [8]. This resulted in improved SBS and one two-step system, Clearfil SE (Kuraray), nearly doubled in SBS compared to using the system in its recommended way and was comparable to SB, which was the control material. These studies, taken together, suggest that insufficient etching of the enamel may partly account for the lower SBS found for selfetch systems. From the cyclic loading data presented in Fig. 3 and Tables 3 and 4 it can be seen that there is a large spread in the number of cycles at which failures occur. Fatigue failure in cyclic loading is a result of nucleation of micro-cracks, propagation of these cracks and eventual coalescence of cracks leading to catastrophic failure [1]. Nucleation is thought to occur at imperfections in the material or at interfaces, such as scratches, voids, inclusions, etc., which might be focal points where high stress intensities can develop under load application. Existing defects or large numbers of imperfections may hasten the process of crack development, while a defect free interface could require many cycles to nucleate micro-cracks. Therefore, the wide range in data might reflect variation in the degree of perfection of the adhesive and interfaces involved in the bond. It is likely that similar variation would exist in clinical situations. The S–N data is presented in a way that tries to clarify this widespread data [2] by showing the maximum, median and minimum numbers of cycles survived for each load utilized. An interesting difference can be seen in the fatigue behavior between PLP and SB. There is a rapid onset of specimens surviving 105 cycles as the stress is lowered, changing from 20% at 9.6 MPa to 80% at 8.0 MPa (Table 4). This is in contrast to the more gradual trend toward increased surviving specimens displayed for SB. It is possible that this behavior is associated with reduced etching of the enamel surface but could also be due to properties of the adhesive. Further work is planned to examine this issue. It would be desirable if the S–N relationships showed clear evidence of a plateau developing, suggesting an endurance limit, i.e. a stress below which no failures would occur for any number of cycles. In a study of bonds to enamel using Scotch Bond Multipurpose adhesive [2], an endurance limit of about 10 MPa was projected, but this seems unlikely as only two of 15 specimens survived 105 cycles at a stress level of 12 MPa
(about 50% of the shear bond strength). In the present study, both SB and PLP had greater numbers of specimens surviving 105 cycles but it would be difficult to try and project endurance limits from the data. Furthermore, it is believed that some materials, including polymers, may not exhibit endurance limits at all [1,3]. It may be more useful to define a practical endurance limit as the stress at which only 5% of specimens fail for a given number of cycles. Using a two-point methodology [5] for describing the trend of the data and extrapolating to a 5% failure gives a practical limit of about 13 MPa for SB and about 7 MPa for PLP at 105 cycles. However, such a procedure is rather speculative since it assumes a normal distribution of the data and a somewhat large extrapolation for the number of specimens tested. The ultimate usefulness of fatigue data would be in correlating the number of cycles and load levels in vitro to similar parameters in vivo. However, there is a wide range of stresses that have been reported for functioning of the dentition (see Ref. [2]). Also, it would be a complex problem to determine what magnitude and type of stress might occur at a given bonded area from occlusal stresses on teeth or restorations. The situation regarding the number of occlusal contacts that occur in vivo, for some period of time, is just as ill-defined. One source [2] cited a number for chewing and swallowing contacts of 1800 per day, which translates into 105 in 2 months. Another [10] suggests that there are 106 , active cycles in 20 years of service and still another estimate [11] considers that 106 cycles represents 5 years of functional life. For the present study, where 105 cycles was the maximum number, this gives a range of 2 months to 2 years for the equivalent period in vivo. One could perhaps settle for 1 year as a compromise. However, this argues for fatigue studies that extend to at least 106 cycles or more, to assure that testing is comparable to expected lifetimes of restorations. For this study the choice of 105 cycles as the upper limit was a concession to the time available for collecting data, while using a relatively large number of cycles and a realistic repetition rate of 2 Hz. Other studies have used lower numbers of cycles from 1000 [4] to 5000 [3], which seems to be too conservative for observing the full effect of fatigue stresses on materials. Using 105 cycles in this study, the SFS for SB was found to be 14.6 MPa and the ratio of SFS to SBS was 57%, while the corresponding values for PLP were 8.4 MPa and 44%. This supports the hypothesis that the fatigue performance of SB would be better than that of PLP. An underlying premise for the hypothesis of this study was that reduced etching of the enamel for PLP would lead to a more tenuous attachment to the enamel and would be observed in SBS and fatigue results. Resin interfaces of bonded specimens, exposed by dissolving the enamel in acid, were examined in the SEM and representative images are provided in Fig. 4 for the two adhesives tested. It can be observed that resin penetrated substantially further into the enamel surface for SB than was the case for PLP. This is, presumably, due to a greater degree of etching of the enamel by the phosphoric acid used with the SB system. While this result correlates with SBS and fatigue results it is not conclusive evidence that etching differences alone are responsible for these results and further work will be necessary to clarify this issue. Table 6 shows a comparison of the initial SBS for the two adhesives and the SBS for pooled specimens that survived 105
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
cycles at the different load levels for the respective adhesive system. There was an expectation that a difference would be observed for the two groups, but no statistically significant differences were found. This suggests that specimens that survived the cycling regimen did not incur damage that resulted in a lower SBS than found initially. Neither did cyclic loading selectively remove specimens that would have been representative of the lower end of the distribution for the initial specimens and thereby increase the measured SBS for the surviving specimens [2]. In fact, the means and standard deviations are quite similar for the two groups, for each adhesive. Results of this study, as with other fatigue studies, found that the sustainable stresses with cyclic loading of bonded specimens were substantially lower than measured shear bond strengths. For the two materials examined, the hypothesis posed for this study was confirmed by finding lower SBS and SFS for the PLP adhesive system compared to the SB adhesive system. In addition the S–N behaviors of the two adhesives were different, with PLP exhibiting a rapid onset of surviving specimens at a lower relative stress. The penetration of resin into the enamel surface for the PLP system was substantially less than for SB, as anticipated, but it is inconclusive that this is the sole cause of the above described differences in bond strength and fatigue results.
references
[1] Baran G, Boberick K, McCool J. Fatigue of restorative materials. Crit Rev Oral Biol Med 2001;12:350–60.
987
[2] Ruse N, Shew R, Feduik D. In vitro fatigue testing of a dental bonding system on enamel. J Biomed Mater Res 1995;29:411–5. [3] Draughn R. Compressive fatigue limits of composite restorative materials. J Dent Res 1979;58:1093–6. [4] Dewji H, Drummond J, Fadavi S, Punwani I. Bond strength of Bis-GMA and glass ionomer pit and fissure sealants using cyclic fatigue. Eur J Oral Sci 1998;106:594–9. [5] Zardiackas L, Caldwell D, Caughman W, Lentz D, Comer R. Tensile fatigue of resin cements to etched metal and enamel. Dent Mater 1988;4:163–8. [6] Lopes G, Marson F, Vieira L, de Andrada M, Baratieri L. Composite bond strength to enamel with self-etching primers. Oper Dent 2004;29:424–9. [7] De Munck J, Vargas M, Iracki J, Van Landuyt K, Poitevin A, Lambrechts P, et al. One-day bonding effectiveness of new self-etch adhesives to bur-cut enamel and dentin. Oper Dent 2005;30:39–49. [8] Miguez P, Castro P, Nunes M, Walter R, Pereira P. Effect of acid-etching on the enamel bond of two self-etching systems. J Adhes Dent 2003;5:107–12. [9] Bolhuis H, De Gee A, Pallav P, Feilzer A. Influence of fatigue loading on the performance of adhesive and nonadhesive luting cements for cast post-and-core buildups in maxillary premolars. Int J Prosthodont 2004;17:571–6. [10] Wiskott A, Nicholls J, Belser U. Fatigue resistance of soldered joints: a methodological study. Dent Mater 1994;10: 215–20. [11] Outhwaite W, Twiggs S, Fairhurst C, King G. Slots vs pins: a comparison of retention under simulated chewing stresses. J Dent Res 1982;61:400–2.
available at www.sciencedirect.com
journal homepage: www.intl.elsevierhealth.com/journals/dema
Fatigue testing of enamel bonds with self-etch and total-etch adhesive systems Robert L. Erickson a,∗ , Anton J. De Gee b , Albert J. Feilzer b a b
Creighton University, School of Dentistry, Omaha, NE, USA Academic Center for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands
a r t i c l e
i n f o
a b s t r a c t
Article history:
Objective. Etching of enamel by self-etching adhesive systems is not as pronounced as with
Received 3 May 2005
phosphoric acid employed with most total-etch adhesive systems. This may result in differ-
Received in revised form
ences in the effectiveness of the bonds for the two types of systems. The aim of this study
22 September 2005
was to compare two such systems by bond strength and fatigue testing.
Accepted 2 November 2005
Materials and methods. Shear bond strengths for Single Bond (SB) and Adper Prompt-L-Pop (PLP) were obtained on bovine enamel surfaces. Fatigue measurements were made with the same test fixtures using cyclical loading at 2 Hz for a maximum of 105 cycles or until
Keywords:
the bond failed. Four selected loads between 40 and 70% of the shear bond strength were
Dental materials
tested for each adhesive and the number of cycles to failure was recorded. S–N curves were
Enamel
constructed from the data and fatigue stress limits were determined. The Mann–Whitney
Adhesion
U-test and t-tests were used for statistical comparisons of the results. SEM analysis of resin
Fatigue
tag formation into enamel surfaces was carried out. Results. A significantly greater bond strength was found for SB (25.3 MPa) than for PLP (19.2). Sustainable stresses after 105 load cycles were substantially lower than the corresponding shear bond strengths and fatigue limits were found to be 8.4 MPa for PLP and 14.6 MPa for SB. The ratio of fatigue limit to bond strength was less for PLP (44%) than for SB (57%). Resin tag penetration into enamel was substantially less for PLP than for SB. Significance. Fatigue testing in conjunction with bond strength testing can provide a better means for assessing the performance of adhesive systems used for bonding to enamel. © 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
A typical, in vitro, evaluation of the effectiveness of an adhesive system involves measurement of bond strengths to enamel or dentin by a shear or tensile test. Such tests are useful for evaluating differences in performance between materials but the results are difficult to relate to in vivo effectiveness. Bond strength tests are dynamic tests involving the application of a monotonically increasing force to the bond until it fails, giving an indication of its strength. This, however, is not
∗
a likely mode of failure for bonds in the mouth, where failure is considered to result from repeated loading over many months or years and at stresses well below the ultimate bond strength. This suggests that fatigue studies, where cyclic loading of bonded specimens is evaluated, may provide better insight to in vivo performance and give more realistic values of sustainable stresses. Testing by cyclic loading is frequently referred to as fatigue testing and a review of such testing for dental materials has recently been published [1]. There are only a small number
Corresponding author at: Creighton University, School of Dentistry, 6101-Lynn Way, Woodbury, MN 55129, USA. Tel.: +1 651 459 7962. E-mail address: [email protected] (R.L. Erickson).
0109-5641/$ – see front matter © 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2005.11.021
982
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
of studies where cyclic loading has been used to examine the effect on adhesive bonds to enamel or dentin directly. Specifically for bonds to enamel, one study [2] developed S–N curves for a three-step total-etch adhesive system and projected a fatigue limit of about 41% of the shear bond strength. S–N curves provide a relationship between applied stress and the logarithm of the number of cycles at which failure of the bond occurs. A more popular method of fatigue testing, referred to as the staircase method, involves selecting a starting stress of about 50–60% of the ultimate strength and an upper limit for the number of cycles, and tests one specimen until it fails or survives. If it fails, the stress is decreased by a set amount for the next specimen but if it survives, the stress is raised by that same set amount. Continuing in this manner for a number of specimens the test focuses on a stress that is likely to produce 50% failures. This value is termed the fatigue limit for the number of cycles selected. A statistical analysis applicable to staircase testing has been described [3]. Two studies, using a staircase approach, have examined fatigue of adhesive bonds to enamel [4,5] and found fatigue limits of 37–46% of the respective bond strengths. These studies point out a common finding of fatigue studies, i.e. fatigue limits are much lower than the typically measured bond strengths and may be in the order of 40–60% of that strength. Also, materials can show different fatigue behavior and a material with a greater bond strength may have a lower fatigue limit than a comparative material [5]. Such differences in fatigue performance may be important for understanding failures in vivo. In recent years, self-etching adhesives have become popular due to their purported ease of use. However, they generally do not etch enamel as strongly as phosphoric acid, which is typically used with total-etch adhesive systems. Some recent studies on bonding to enamel have found that self-etching adhesive systems, whether two-step or one-step, have inferior bond strengths compared with total-etch systems used as control materials [6,7]. In addition, a pre-etch of enamel with phosphoric acid was shown to improve the bond strengths of two-step adhesives compared to their normal application method [8]. These studies suggest that inadequate etching may be at least part of the reason that self-etching adhesive systems do not achieve the same level of performance on enamel as total-etch systems. The purpose of this study was to test the hypothesis that the two-step, total-etch adhesive system, Single Bond (SB; 3M
ESPE, St. Paul, MN, USA), can provide better bond strength and fatigue strength when bonded to enamel, than the one-step, self-etch adhesive, Adper Prompt-L-Pop (PLP; 3M ESPE, St. Paul, MN, USA). A secondary goal was to construct S–N curves to see if the fatigue behavior differs for the two adhesive systems.
2.
Materials and methods
2.1.
Enamel specimen preparation
Cylindrical cores of enamel and attached dentin, 6.0 mm diameter were cut from bovine incisors, normal to the labial surface, using a hollow core diamond drill with copious water cooling (Diamant Boart Inc., Vianen, The Netherlands). The fabrication procedure for each test specimen started with etching the enamel along the cut cylindrical wall of a core with 35% phosphoric acid, followed by rinsing and drying. The etched enamel was coated with a bonding agent (Single Bond; 3M ESPE, St. Paul, MN, USA), air dried and light-cured. The core was inserted into a 6.0 mm hole in a 2.0 mm thick, stainless steel plate (Fig. 1a), with the enamel surface above the plate by about 0.5 mm. A light-curable, flowable composite (Tetric Flow, Ivoclar Vivadent, Schaan, Liechtenstein) was used to cement the core into place. The exposed enamel surface was ground, by hand, on 240 grit abrasive paper with water cooling, until it was flush with the stainless steel plate. This ground surface was treated with one of the two adhesive systems being investigated as described in Table 1. Manufacturer’s instructions were followed except for doubling the light curing time to 20 s to ensure good cure of the adhesives. Next, the stainless steel plate, with the enamel specimen, was placed on a pin-indexing jig and a second matching plate was placed over it (Fig. 1b). The pins helped position a 4.0 mm hole, in this second plate, over the center of the enamel surface. A resin-composite (Z 250; 3M ESPE, St. Paul, MN, USA) was condensed into this cylindrical mold in two 1.0 mm increments, each light-cured for 20 s. The pins were then pushed out to allow transfer of the completed test specimen to a test fixture. This fixture consisted of a cylindrical cavity in a brass block, with shallow, diametrically opposed grooves that were machined to precisely fit the two-plate specimen assembly (Fig. 1c). This allowed free vertical movement of the assembly with negligible lateral movement. In this manner specimens were made, with each adhesive, for both shear bond strength testing and fatigue testing. Bonded specimens were stored for
Table 1 – Investigated adhesive systems and procedures Adhesive system
Code
Etchant
Procedures
Single Bond (lot# 3HK)
SB
35 wt.% phosphoric acid
1. Etch enamel 15 s; rinse; dry thoroughly with air 2. Apply two consecutive coats of adhesive and dry thoroughly with air 3. Light-cure 20 s
Adper Prompt-L-Pop (lot# 135864)
PLP
Self-etch
1. Mix adhesive in package according to instructions 2. Apply adhesive with a rubbing action for 15 s 3. Dry thoroughly with gentle air stream 4. Apply second coat without rubbing and dry thoroughly 5. Light-cure 20 s
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
983
Fig. 1 – Specimen test fixture shown with pin-indexing jig for facilitating assembly. (a) Exploded view, with lower plate showing enamel specimen in 6.0 mm hole and upper plate having a 4.0 mm hole for composite. (b) Assembled view. After applying adhesive to enamel the lower plate is placed on pin-indexing jig and then the upper plate is positioned on top with hole centered on enamel specimen. Composite is condensed into this hole and light-cured to complete the bond. Pins are pushed out to allow transfer of the test specimen. (c) Test fixture consisting of a cylindrical cavity in a brass block, with shallow, diametrically opposed grooves to precisely fit the two-plate specimen assembly.
24 h in distilled water at 37 ◦ C after the bonding procedures were completed and were maintained under those conditions throughout the bond strength and fatigue testing procedures described below.
2.2.
Shear bond strength testing
The specimen assembly when placed in the brass test fixture was seated to the bottom of the cavity with the end of the second plate protruding out of the cavity. By applying a slowly increasing force onto the end of this plate, a shear force was exerted on the bonded interface until total bond failure occurred. Ten specimens were tested for shear bond strength (SBS) in an Instron testing apparatus (model no. 6022; Instron, High Wycombe, Bucks, UK) at a cross-head speed of 1.0 mm/min, for each adhesive system.
2.3.
Cyclic load testing
Load cycling of bonded specimens was carried out with the same test configuration as SBS testing. A twelve-station fatigue testing apparatus utilizing air pressure to regulate and deliver the force was employed for load cycling and has been described in another publication [9]; a photograph and schematic diagram are shown in Fig. 2. For each station, a test fixture was positioned so that a flat-ended rod, driven by the piston delivering the load, impinged on the end of the protruding plate of the specimen assembly. The brass text fixture sat on the metal base of a custom water bath. Each station had a
built-in load cell and separate controls for adjusting the maximum and minimum loads for each cycle. A counter on each station recorded the number of cycles elapsed and the station shut off automatically when bond failure occurred. The load was applied at a rate of 2 Hz with a 50% duty cycle and took the form of a damped square wave, there being a 40 ms time constant in switching between load settings. The load was cycled between a fixed lower setting of 12 N, an arbitrary value used to maintain some load at all times, and preset higher loads ranging from 40 to 70% of the SBS determined for each of the adhesive systems. The range of loads was estimated to be adequate for determining S–N curves from previous unpublished fatigue studies and from published studies [2,4,5], the goal being to cover the range where specimens mostly fail to mostly survive. Cyclic loading was carried out to 105 cycles or until the specimens failed by total loss of the bond and the number of cycles at failure were recorded for each specimen. There were 15–17 specimens tested at each load value for each adhesive. Specimens that survived 105 cycles were tested for SBS by the methods described earlier.
2.4.
SEM analysis
Cylindrical cores of bovine enamel, polished as already described, were used to evaluate the degree of resin tag penetration into the enamel surface during the bonding process. The procedures used to bond the surface were identical to those used for other test procedures already described, except that the enamel specimens were not cemented into stainless
984
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
Fig. 2 – Left: The ACTA fatigue tester. Right: Schematic diagram of a single station with specimen depicted in place.
steel plates. Four bonded specimens for each adhesive system were immersed in 50 vol.% nitric acid until the enamel was fully dissolved and the resin interfaces exposed. These specimens were rinsed with distilled water and dried gently with an air stream. They were then mounted on SEM stubs and sputter-coated with gold. SEM observations were made on these specimens and representative micrographs were recorded.
3.
in Fig. 3 and Tables 3 and 4, where the load values in Newtons have been converted to Megapascals. Fig. 3 shows S–N data points associated with the minimum, median and maximum number of cycles before bond failure for each load level tested. The connecting lines through the data points provide approximate S–N curves for those same parameters. The minimum and maximum curves form an envelope that encompasses all of the data and the median curve gives an indication of how
Results
Results for shear SBS testing are shown in Table 2, expressed in Megapascal units, using the failure load in Newtons and the area of the hole (mm2 ) in which the composite was placed as the nominal bond area. A statistically significant difference was found between the mean bond strengths of the two adhesive systems (SB: 25.3 MPa; PLP: 19.2 MPa), using a t-test (p = 0.01). Load levels chosen for cyclic testing were determined from the above SBS values. For SB they were 165, 187.5, 210 and 232.5 N and for PLP the load levels were 100, 120, 140 and 165 N. It should be noted that 165 N was the only common load for the two adhesive systems. Results for cyclic loading are shown
Table 3 – Summary of the cyclic loading results for SB showing minimum, median and maximum number of cycles survived by specimens at four stress levels # Cycles Min Med Max % Survivors
13.1 (MPa)
14.9 (MPa)
16.7 (MPa)
1.5 × 104 105 105 69
2.2 × 103 2.6 × 104 105 25
1 2.7 × 103 105 13
18.5 (MPa) 1 80 6 × 103 0
The percentage of specimens surviving 105 cycles is shown at the bottom for each stress level.
Table 4 – Summary of the cyclic loading results for PLP showing minimum, median and maximum number of cycles survived by specimens at four stress levels Table 2 – Mean shear bond strength of adhesive systems evaluated Shear bond strength (MPa) (S.D.) SB PLP
25.3 (5.6)a 19.2 (3.2)b
Different superscript letter indicates a statistically significant difference (p < 0.01; t-test).
# Cycles
8.0 (MPa)
Min Med Max % Survivors
4.8 × 10 105 105 80
3
9.6 (MPa) 2 2 × 103 105 20
11.2 (MPa)
13.1 (MPa)
1 2.9 × 102 2.9 × 103 0
1 10 1.9 × 103 0
The percentage of specimens surviving 105 cycles is shown at the bottom for each stress level.
985
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
Fig. 3 – S–N curves for the minimum, median and maximum number of cycles until failure, for composite bonded to enamel, using two test adhesives. The dashed arrows signify that these data points are for specimens that survived and could have gone to larger numbers of cycles. A maximum of 105 cycles were used and four stress levels for each adhesive material. Fifteen to seventeen specimens were used at each stress level. (a) Single Bond, (b) Adper Prompt-L-Pop.
Fig. 4 – Representative SEM micrographs of resin tags, exposed by dissolving enamel from bonded specimens; 2000× magnification. (a) Single Bond, (b) Adper Prompt-L-Pop.
the data is distributed within that envelope. From the above data, it can be observed that there is no overlap of data at the 13.3 MPa load level and this immediately indicates that the two adhesives have a statistically significant difference in their fatigue performance at that level, as confirmed by a Mann–Whitney U-test (p = 0.01). Further, using the proportions of failures after 105 cycles (Tables 3 and 4) in a calculation of the mean shear fatigue stress (SFS) [3], the values shown in Table 5 were determined. A statistically significant difference was found between the SFS values for the two adhesives using a t-test (p = 0.01). The ratio of SFS to SBS was found to be smaller for PLP (44%) than for SB (57%).
Table 5 – Comparison of shear bond strength (SBS) and shear fatigue strength (SFS) and the ratio of these two strengths for the two adhesive systems evaluated
SB PLP
SBS (MPa)
SFS (MPa)
SFS/SBS (%)
25.3 (5.6)a 19.2 (3.2)b
14.6 (1.6)a 8.4 (0.5)b
57 44
Different superscripts in columns indicate statistically significant differences (p < 0.01; t-test).
Table 6 – Comparison of initial shear bond strengths with shear bond strengths of specimens surviving cyclic loading
SB PLP
Initial strength (MPa)
Strength of survivors (MPa)
25.3 (5.6)a 19.2 (3.2)b
27.1 (5.0)a 20.6 (4.2)b
Different superscript letters indicate a statistically significant difference (p = 0.01).
Table 6 compares the initial SBS values to the SBS for all specimens that survived 105 cycles, regardless of load. No statistically significant differences were found using a t-test (p = 0.05). Fig. 4 shows representative images of resin tags from SEM micrographs, taken at 2000× magnification, for the two adhesives. It is clear that the tags are longer and more extensive for the SB adhesive than for the PLP adhesive.
4.
Discussion
Bond strength tests are the most common way of evaluating the effectiveness of an adhesive system. In this study
986
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
SBS were determined for the two adhesive systems, both as a direct comparison of their effectiveness in this type of test, and also to provide baseline values with which the SFS values could be compared. A statistically significant difference between the SBS values of SB and PLP was found, with the value for PLP being 76% of that for SB, which supports that part of the original hypothesis for this study. Similar results were found in another study of the same materials [7] where the ratio was 73%. In that study, five self-etch systems were compared to SB and four of the five, including PLP, were found to have lower SBS compared to SB. In a study where microtensile bond strengths were determined [8], three out of four selfetch systems tested, one being PLP, were lower in SBS than for a three-step, self-etch adhesive control material, Optibond FL (Kerr). In this case the SBS ratio was 45%. The possibility that insufficient etching may reduce the SBS of self-etch adhesive systems has been tested by including a pre-etch of enamel with phosphoric acid before applying the self-etching adhesive system [8]. This resulted in improved SBS and one two-step system, Clearfil SE (Kuraray), nearly doubled in SBS compared to using the system in its recommended way and was comparable to SB, which was the control material. These studies, taken together, suggest that insufficient etching of the enamel may partly account for the lower SBS found for selfetch systems. From the cyclic loading data presented in Fig. 3 and Tables 3 and 4 it can be seen that there is a large spread in the number of cycles at which failures occur. Fatigue failure in cyclic loading is a result of nucleation of micro-cracks, propagation of these cracks and eventual coalescence of cracks leading to catastrophic failure [1]. Nucleation is thought to occur at imperfections in the material or at interfaces, such as scratches, voids, inclusions, etc., which might be focal points where high stress intensities can develop under load application. Existing defects or large numbers of imperfections may hasten the process of crack development, while a defect free interface could require many cycles to nucleate micro-cracks. Therefore, the wide range in data might reflect variation in the degree of perfection of the adhesive and interfaces involved in the bond. It is likely that similar variation would exist in clinical situations. The S–N data is presented in a way that tries to clarify this widespread data [2] by showing the maximum, median and minimum numbers of cycles survived for each load utilized. An interesting difference can be seen in the fatigue behavior between PLP and SB. There is a rapid onset of specimens surviving 105 cycles as the stress is lowered, changing from 20% at 9.6 MPa to 80% at 8.0 MPa (Table 4). This is in contrast to the more gradual trend toward increased surviving specimens displayed for SB. It is possible that this behavior is associated with reduced etching of the enamel surface but could also be due to properties of the adhesive. Further work is planned to examine this issue. It would be desirable if the S–N relationships showed clear evidence of a plateau developing, suggesting an endurance limit, i.e. a stress below which no failures would occur for any number of cycles. In a study of bonds to enamel using Scotch Bond Multipurpose adhesive [2], an endurance limit of about 10 MPa was projected, but this seems unlikely as only two of 15 specimens survived 105 cycles at a stress level of 12 MPa
(about 50% of the shear bond strength). In the present study, both SB and PLP had greater numbers of specimens surviving 105 cycles but it would be difficult to try and project endurance limits from the data. Furthermore, it is believed that some materials, including polymers, may not exhibit endurance limits at all [1,3]. It may be more useful to define a practical endurance limit as the stress at which only 5% of specimens fail for a given number of cycles. Using a two-point methodology [5] for describing the trend of the data and extrapolating to a 5% failure gives a practical limit of about 13 MPa for SB and about 7 MPa for PLP at 105 cycles. However, such a procedure is rather speculative since it assumes a normal distribution of the data and a somewhat large extrapolation for the number of specimens tested. The ultimate usefulness of fatigue data would be in correlating the number of cycles and load levels in vitro to similar parameters in vivo. However, there is a wide range of stresses that have been reported for functioning of the dentition (see Ref. [2]). Also, it would be a complex problem to determine what magnitude and type of stress might occur at a given bonded area from occlusal stresses on teeth or restorations. The situation regarding the number of occlusal contacts that occur in vivo, for some period of time, is just as ill-defined. One source [2] cited a number for chewing and swallowing contacts of 1800 per day, which translates into 105 in 2 months. Another [10] suggests that there are 106 , active cycles in 20 years of service and still another estimate [11] considers that 106 cycles represents 5 years of functional life. For the present study, where 105 cycles was the maximum number, this gives a range of 2 months to 2 years for the equivalent period in vivo. One could perhaps settle for 1 year as a compromise. However, this argues for fatigue studies that extend to at least 106 cycles or more, to assure that testing is comparable to expected lifetimes of restorations. For this study the choice of 105 cycles as the upper limit was a concession to the time available for collecting data, while using a relatively large number of cycles and a realistic repetition rate of 2 Hz. Other studies have used lower numbers of cycles from 1000 [4] to 5000 [3], which seems to be too conservative for observing the full effect of fatigue stresses on materials. Using 105 cycles in this study, the SFS for SB was found to be 14.6 MPa and the ratio of SFS to SBS was 57%, while the corresponding values for PLP were 8.4 MPa and 44%. This supports the hypothesis that the fatigue performance of SB would be better than that of PLP. An underlying premise for the hypothesis of this study was that reduced etching of the enamel for PLP would lead to a more tenuous attachment to the enamel and would be observed in SBS and fatigue results. Resin interfaces of bonded specimens, exposed by dissolving the enamel in acid, were examined in the SEM and representative images are provided in Fig. 4 for the two adhesives tested. It can be observed that resin penetrated substantially further into the enamel surface for SB than was the case for PLP. This is, presumably, due to a greater degree of etching of the enamel by the phosphoric acid used with the SB system. While this result correlates with SBS and fatigue results it is not conclusive evidence that etching differences alone are responsible for these results and further work will be necessary to clarify this issue. Table 6 shows a comparison of the initial SBS for the two adhesives and the SBS for pooled specimens that survived 105
d e n t a l m a t e r i a l s 2 2 ( 2 0 0 6 ) 981–987
cycles at the different load levels for the respective adhesive system. There was an expectation that a difference would be observed for the two groups, but no statistically significant differences were found. This suggests that specimens that survived the cycling regimen did not incur damage that resulted in a lower SBS than found initially. Neither did cyclic loading selectively remove specimens that would have been representative of the lower end of the distribution for the initial specimens and thereby increase the measured SBS for the surviving specimens [2]. In fact, the means and standard deviations are quite similar for the two groups, for each adhesive. Results of this study, as with other fatigue studies, found that the sustainable stresses with cyclic loading of bonded specimens were substantially lower than measured shear bond strengths. For the two materials examined, the hypothesis posed for this study was confirmed by finding lower SBS and SFS for the PLP adhesive system compared to the SB adhesive system. In addition the S–N behaviors of the two adhesives were different, with PLP exhibiting a rapid onset of surviving specimens at a lower relative stress. The penetration of resin into the enamel surface for the PLP system was substantially less than for SB, as anticipated, but it is inconclusive that this is the sole cause of the above described differences in bond strength and fatigue results.
references
[1] Baran G, Boberick K, McCool J. Fatigue of restorative materials. Crit Rev Oral Biol Med 2001;12:350–60.
987
[2] Ruse N, Shew R, Feduik D. In vitro fatigue testing of a dental bonding system on enamel. J Biomed Mater Res 1995;29:411–5. [3] Draughn R. Compressive fatigue limits of composite restorative materials. J Dent Res 1979;58:1093–6. [4] Dewji H, Drummond J, Fadavi S, Punwani I. Bond strength of Bis-GMA and glass ionomer pit and fissure sealants using cyclic fatigue. Eur J Oral Sci 1998;106:594–9. [5] Zardiackas L, Caldwell D, Caughman W, Lentz D, Comer R. Tensile fatigue of resin cements to etched metal and enamel. Dent Mater 1988;4:163–8. [6] Lopes G, Marson F, Vieira L, de Andrada M, Baratieri L. Composite bond strength to enamel with self-etching primers. Oper Dent 2004;29:424–9. [7] De Munck J, Vargas M, Iracki J, Van Landuyt K, Poitevin A, Lambrechts P, et al. One-day bonding effectiveness of new self-etch adhesives to bur-cut enamel and dentin. Oper Dent 2005;30:39–49. [8] Miguez P, Castro P, Nunes M, Walter R, Pereira P. Effect of acid-etching on the enamel bond of two self-etching systems. J Adhes Dent 2003;5:107–12. [9] Bolhuis H, De Gee A, Pallav P, Feilzer A. Influence of fatigue loading on the performance of adhesive and nonadhesive luting cements for cast post-and-core buildups in maxillary premolars. Int J Prosthodont 2004;17:571–6. [10] Wiskott A, Nicholls J, Belser U. Fatigue resistance of soldered joints: a methodological study. Dent Mater 1994;10: 215–20. [11] Outhwaite W, Twiggs S, Fairhurst C, King G. Slots vs pins: a comparison of retention under simulated chewing stresses. J Dent Res 1982;61:400–2.