Bond Strength of Fiber-Reinforced Composite Resin Restorations Douglas Galan, D. M. D., M. Sc.* Edward Lynch, B. D. S , ED. S f M . Robin Heath, Ph.D., B. D. S., I*: D.S.
Severe tooth wear is common in older dentate individuals, with one treatment option being composite resin restorations reinforced with a suitable matrix. This study evaluated the use of high modulus polyethylene (Celanese) fibers as a reinforcing matrix for composite resin. Human mandibular incisor teeth were sectioned to simulate severe tooth wear. Sectioned surfaces were measured, the teeth paired and assigned to control or test sample groups, and further assigned to be tested with a labial or lingual shearing force. A phosphorylated dentin bonding adhesive was applied to the abraded dentin surface. Labial and lingual intra-enamel bevel preparations were cut on each specimen. A piece of Celanese fabric was bonded onto the acid-etched labial and lingual bevels of the test specimens using an enamel bonding agent. Class IV composite resin restorations were then placed onto the test and control specimens. Following water storage, the specimens were subjected to shearing forces. Bond strengths for test specimens were significantly greater (p c 0.03)than the controls. Specimens with a labially applied force also had significantly higher bond strengths (p < 0.001). SEM analysis revealed adhesive bond failures over dentin surfaces, with cohesivebond failures within the composite resin. Celanese fibers maintained the restorations on the teeth, although adhesive failures were seen between the fibers and enamel bonding agent.
evere tooth wear of anterior teeth is an increasingly frequent problem suffered by older dentate individuals. A s the populations of Canada and the United States continue to age, with a larger proportion maintaining more of their natural dentition, this problem of tooth wear will continue to increase. Traditionally, the restoration of these teeth to their original form and function usually involves extensive, complex, and expensive prosthodontic procedures. However, in some instances where the loss of tooth structure has been minimal, composite resin restorations may be advo-
cated. Composite resin restorations are easy to place, are considerably less expensive than prosthodontic procedures, and have a favorable resistance strength to withstand compressive masticatory forces. Unfortunately, the low shear bond strength of composite resins may limit the prognosis of large incisal restorations. Furthermore, in instances of severe tooth wear, the amount of enamel remaining for acid-etching and subsequent bonding is often limited, which further restricts the use of composite resins. Clinical cases of severe tooth wear are usually best restored using prosthodontic treatment modalities. However, in certain older individuals, especially those with a compromised physical health or those that are of advanced years, prosthodontic treatment may not be appropriate and is often not advocated. Instead composite resins, despite their limitations, may have to be utilized for restoring worn anterior teeth. To counteract the inherent weaknesses of composite resins, these materials may have to be strengthened by a reinforcing matrix or framework. One suitable matrix may be produced by using reinforcing fibers. Several different types of fibers are available including carbon-graphite fibers, aramid fibers, fiberglass fibers, and silk fibers.
Communlty Dentistry Programs. Faculty of Dentistry. University of Manitoba. Winnipeg. Manitoba. Canada. t Adult Dental Health and Restorative Dentistry, Dental School. The London Hospital Medical College, University of London. London. England. t Prosthetic Dentistry. Dental School. The London Hospital Medical College, University of London. London. England. Thls study was undertaken in partial fulRllment of the degree of Masters of Science in Gerodontics. University of London. England. Address reprint requests t o Douglas G a l a . D.M.D., M.Sc., Community Dentistry Programs. Faculty of Dentistry. University of Manitoba. 780 Bannatyne Avenue. Winnipeg. Manitoba. Canada. R3E OW2 01992 Decker Periodicals Inc.
Bond Strength of Composite Resin Restorations
The carbon-graphite and aramid fibers have been found to be poor reinforcers when directly incorporated into composite resins due to fiber placement difficulties and because the composite resin was ultimately weakened.‘-3 When carbon-graphite fibers are used as a reinforcingmatrix for periodontal ~ p l i n t sor~under .~ allcomposite resin bridges6 their performance has been very satisfactory. Esthetically, however, the black color of carbon-graphite fibers limits their use intraorally. Fiberglass and silk fibers have been used with composite resin to create orthodontic fixed retainers’ and splints,*-” in addition to some esthetic dentistry applications.’*.l 3 These fiber systems, in particular the silk fibers, have not been subjected to any extensive, long-term clinical trials, nor have any in uitro evaluations been conducted on them. As a result, the reports on the successful use of these fibers should be considered carefully. Ultra-high modulus polyethylene (Celanese)fibers could be another fiber suitable for reinforcing composite resin restorations. Celanese fibers have an enhanced modulus, are ductile, and can be placed esthetically because of their neutral color. To enhance bonding, the fibers are usually surface treated by an electric plasma method using oxygen or helium gas. This surface treatment causes the fibers to become etched, thereby permitting a mechanical bond to be created. These fibers have already demonstrated that they can successfully reinforce denture-base acrylics.l4 Not only do Celanese fibers increase the impact strength of polymethylmethacrylate. but they are easy to incorporate. The aim of this study was to determine if Celanese fibers could significantly increase the strength of class IV composite resin restorations placed on artificially worn teeth and, by distributing shearing forces more uniformly,improve the compositeresin to enamel bond. Because restored incisor teeth with an overbite are subjected to different shearing forces from labial and lingual directions, it was desirable to determine the results of bond testing by using realistically shaped restorations, and by applying the shearing force from two directions. Finally, since the tooth wear was artificially produced, the possibility that the direction of sectioning will affect the bond strength was investigated.
with 16- and 6- pm aluminum oxide slurries to produce a dentin surface resembling an abraded tooth. These abrasion surfaces were measured with a reflex microscope (ReflexMeasurement Ltd., London, England) and the abrasion surface areas used for pairing specimens. A t random, half of the matched pairs were assigned to the test sample and half to the control sample. The newly exposed dentin surface was covered with a phosphorylate dentin bonding adhesive (DBA) (Bondlite, Kerr/Sybron UK Ltd.) without prior smear layer removal, light cured with a fiberoptic light (Command 2 Fiber Optic Light, Kerr/Sybron UK Ltd.) for 30 s, covered with a thin film of composite resin (HerculiteX R , Kerr/Sybron UK Ltd.), and light cured for 60 s. This composite resin was applied to protect the DBAfrom the acid used for enamel etching. Standardized 0.5 mm intra-enamel bevels were prepared on the labial and lingual tooth surfaces, leaving the proximal surfaces intact. The bevels extended 2 mm cervically. beginning at the abrasion surface. These beveled and abraded enamel surfaces were etched with 37 percent phosphoric acid gel etchant for 15 s, washed for 30 s, and air dried for 60 s. Test specimens had 2 strips of Celanese fabric (University of Leeds, Physics Department, Leeds, England) approximately 3 mm x 8 mm placed prior to the composite resin restorations. A powder-liquid chemically cured enamel bonding agent (EBA) was used to facilitate Celanese fabric bonding. Each strip was initially soaked in the EBA liquid for 1.5min to fully wet the fabric, while a mixed EBA was brushed onto the etched lingual and incisal enamel. After soaking, a freshly mixed EBA was applied to one side of the strip, which was firmly compressed onto the lingual bevel, 0.5 mm short of the bevel margin, with 1mm of the strip adapted onto the abrasion surface (Fig. 1).The Celanese strip was then allowed to stiffen for 4 min. The same procedures were repeated to bond the second strip onto the labial and abrasion surfaces, and onto the already attached Celanese strip. To ensure bonding, a mixed EBAwas placed onto the labial surface of the previously attached strip prior to the application of the second strip. Following an additional 4-min setting period, the Celanese framework was reduced with a 50 pm diamond stone (DentsplyUK Ltd.) to a height of 2.0 mm to ensure that the frameworkwould be completely covered by the composite resin. The Celanese framework was then re-etched with acid for 15s,rinsed for 30 s, and air dried for 60 s to cleanse the fiber framework following rinsing and to prepare it for the composite resin application. Test and control samples had class IV composite resin restorations built up in a similar manner. Mixed EBA was brushed onto the etched enamel or Celanese framework, and allowed to set for 4 min. Each tooth was restored to a crown height of 9.5 mm using three applications of composite resin. Every composite application was subjected to four light exposures in a rotating
METHOD AND MATERIALS Following a pilot study on this restoration technique, it was determined that 16 pairs of recently extracted, caries and restoration-free, human mandibular incisors would be needed. Immediately following extraction, all teeth were stored with thymol crystals at 100 percent humidity. The incisal edges of all specimens were horizontally sectioned to simulate severe tooth wear by having 3.5 mm of crown height removed from the incisal portion. Following sectioning, the incisal edges were ground 25
both control and test specimens were: (1) the side adjacent to the applied force (SAF): (2)the side opposite the applied force (SOF); (3)the abrasion surface (AS); and (4)the restoration surface (RS) bonded onto the AS.
RESULTS Bond Strength CELANESE STRIP I
CELAKESE STRIP 2
Results from the bond strength tests show that the Celanese reinforced test specimens had a higher mean bond strength than the unreinforced controls (Table 1). In two of the test specimens, the tooth fractured before the restoration, therefore the bond strength was computed a t that fracturing force and expressed as a “greater than” value. These “greater than” values have been included in the calculation of the mean values for the test specimens. A numerical analysis revealed that the direction of tooth sectioning did not affect bond strength, therefore the two-way analysis of variance was conducted for evaluation of the reinforcement technique and direction of force (Table 1).The analyses for the reinforcement technique indicated that there was a significant difference between the test and control specimens (p < 0.03). The analyses also revealed that the forces needed to fracture the restorations from the labial direction were highly significantly greater than those from the lingual direction (p < 0.001). When the radiographs were assessed and the remaining enamel and applied composite resin were measured, the enamel thickness on average was found to be greater on the labial (La) surface thanonthelmgual (Li)(me% = 0.9 mm;meanu = 0.6 mm), a s was the thickness of the composite resin (mean, = 1.0 mm; mean,, = 0.8 mm). This directly related to the force application differences.
ABRASION SURFACE BEVEL MARGIN CELANESE FABRIC COMPOSITE RESIN
Figure 1. Celanese strip application to a prepared specimen. Strip 1 was attached to the lingual bevel of the specimen and partly onto the AS. Strip 2 was then attached onto the labial bevel, AS surface, and the already attached Strip 1 . Proximally the strips were placed 0.5 mm from the mesial and distal tooth margins. Composite resin was used to restore the crown, completely covering the Celanese fiber framework.
and overlapping fashion for 60 s per exposure. Excess composite resin material was removed with high-speed 12-and 30-bladed tungsten carbide burs (Dentsply UK Ltd.) using water coolant. Radiographs were taken of each restoration from the proximal aspect to determine the labiolingual thickness of the remaining enamel and the applied composite resin. The restorations were then stored for 14 days in distilled water at 37°Cbefore bond testing. Prior to bond strength testing, the roots of the specimens were mounted in polymethylmethacrylate. A random sample of half of the test specimens and their matched controls received a labially applied force (Fl), and the remaining specimens received a lingually applied force (F2).The shearing force was applied to the midpoint of the lingual or labial tooth surface 1 mm below and perpendicular to the incisal edge a t a crosshead speed of 1 mm/min until fracture. Means and standard deviations of the force at fracture for control and test specimens were computed. These data were then subjected to a two-way analysis of variance (ANOVA) using n- 1 degrees of freedom. Both sets of specimens were observed under light microscopy to ascertain possible sites and modes of failure. Selected fracture surfaces were then assessed with scanning electron microscopy (SEM) to examine the mode of failure. Sites chosen for examination on
Mode of Failure Control specimens exhibited a consistent mode of failure. In all instances, the composite resin remained attached to the bevel preparation areas, indicating a Table 1. Bond Strengths (MPa) and ANOVA Analyses for Test and Control Specimens
Shear bond strength
Standard deviation (n-1) Range ANOVA
Flt F2 Mean
b10.0 7.7 >8.9
8.6 5.0 6.8
< * 1.8
7.1 -12.0 Test vs control F1 vs F2 (all samples)
3.5-11.3 p > 0.03 p 0.01
c or > indicates where a value was included for teeth
that fractured prior to restoration failure t F1 = shearing force applied to the labial surface of the specimen t F2 = shearing force applied to the lingual surface of the specimen
Bond Strength of Composite Resin Restorations
cohesive resin failure occurring close to the restorationAS interface. Restorations over the AS showed clear
margins separating the enamel and dentin portions (Fig. 2). Replicas of sectioning ridges and troughs are seen on the dentin portions of these restoration surfaces, indicating an adhesive bonding failure. In contrast, the enamel portion of the AS revealed a cohesive failure because sectioning ridges and troughs were absent, and the tooth surface still contained the bonded EBA. With one exception, all of the test specimens remained attached to the teeth. On the SAF. multiple sites of fracture occurred suggesting a complicated failure pattern. Most fracture lines extended internally to the Celanese fiber framework. In each specimen, only one fracture line was seen through the restoration-AS interface. Fibers generally remained attached to the composite resin on both the bevel preparation areas and within the fracturing restoration (Fig. 3). Porosities were also evident in the fracture lines, primarily near the bevel preparation areas and within the fiber framework. In the specimen where the reinforced restoration separated from the tooth, the SAF and SOF differed in the fracture pattern produced. On the SAF, the fibers attached to the bevel preparation area were severed in two within the composite resin of the restoration, in a position incisal to the restoration-AS junction. SEM assessment revealed that these fibers pulled out of the composite resin and protruded above the restorationAS interface, remaining embedded in the resin still attached to the SAF bevel. On the corresponding RS of the fractured specimen, empty channels were observed where the fibers had pulled out. This SAF failure was adhesive between the fibers and the composite resin, and cohesive in the composite resin attached to the SAF
Figure 3. SEM of the SAF fracture separation gap of a test specimen. Some fibers appear to be tom (A). while others were unaffected. A composite resin fracture particle (B) can be seen between the fibers. A lack of EBA coating on the fibers. with the odd exception (C). indicates an adhesive bond failure between the fibers and EBA.
bevel. On the SOF bevel, the composite resin remained attached to the bevel area, but the fibers entirely pulled out of this resin and remained embedded in the fracturing restoration. As a result, the SEM assessment of the SOFbevel revealed empty channels within the composite resin that once contained the fibers (Fig. 4). An adhesive failure was present between the composite resin and the fibers, while a composite resin cohesive failure existed on the SOF bevel area. The mode of failure on the AS of this specimen had a similar appearance to the control specimens, except for an area where the smear layer was removed and the dentinal tubules exposed.
Figure 2. SEM of the abrasion surface from a fractured control specimen. Dentin portion (D)reveals ridges and troughs produced during tooth sectioning. The smear layer was evident, without any DBA. suggesting an adhesive bond failure. Enamel portion (E) reveals no sectioning ridges and troughs, with an elevation at the DEJ ( m w )indicating that EBAwas present on the enamel.
Figure 4. SEM of the SOF bevel of the test specimen that fractured offthe tooth. Composite resin remains attached to the bevel preparation area, where filler particles (A) and stress lines (B) suggest a cohesive bond failure. The fibers pulled out of the composite resin, leaving empty channels.
JOLRY.AI, OF ESTIiETIC DENI'ISI'IW VOLUME 4, NUMBER 1 Jmuary/Februarv 1992
could fonn2:'Thesecracks would then propagate when the restorations were stressed by shearing forces. Fiber fracture was seen in some cases, but more commonly the fibers lost their adhesion to the resin and pulled out leaving empty channels. Two different factors probably contributed to the loss of the fiber to composite resin adhesion: limited fiber wetting and a lack of plasma treatment of the fibers. The difficulty of ensuring adequate fiber wetting with the powderlliquid EBA used in this studyjustifies further investigation of a less viscous, unfilled EBA to serve as the fiber bonding agent. Plasma treatment of Celanese fibers is usually desired for enhanced bonding, unfortunately no plasma treated fibers were available for this study and untreated fibers were used. The use of plasma-treated fibers has already been shown to increase resistance to pull-out forces and improve the adhesion between the fibers and polymethylmethacrylate. l 4 Plasma-treated fibers should demonstrate a n improved EBA-fiber bond, and in recent work involving the application of BE-GMA to plasma-treated fibers, a material ofhigh strength was produced without fiber pull-out.24 Resistance to forces from different directions was tested because of its clinical importance. The results showed a highly significant difference between the two force directions employed. Some of the differences were almost inevitable because of the contrasts between the convex labial and concave lingual surfaces of the restorations. These contrasts were evident in the thickness difference shown radiographically for the enamel and composite resin. An inadequate thickness of composite resin overlying a lingual AS-SAF bevel junction would decrease the resistance to shearing forces in that area. This area in a restoration would therefore be a weak point that would fracture easily. The tendency to fracture easily was due in part to the inherent brittleness of composite resin, especially when there was an inadequate amount of material. In addition, the enamel remaining after bevel preparation would be closer to the dentoenamel junction and would therefore have a higher organic content and produce poorer etch patterns.25 This weaker enamel would not produce as strong a bond with composite resin and would ultimately reduce the bond strength between composite resin and enamel. The decreased enamel and composite resin thickness, and thus the strength of the SAFbevel, were more important to the net bond strength of the restoration than the SOFbevel. The most obvious clinical relevance is for restorations of maxillary anterior teeth where a n overbite exists. Here the tendency is to minimize the palatal contour (equivalent to the SAF surface), which is typically concave, as opposed to providing a greater thickness, which would be stronger. These points highlight the importance of the SAF/SOFdifferences in the fracture patterns of the fiber-reinforced restorations and confirms that the complex structure does indeed give potential for improving restorations. Finally, since the data failed to indicate any differences related to the sectioning direction, it appears
Celanese-reinforced test specimens were able to resist shearing forces more effectivelythan the composite-only control specimens. The improved bond strengths achieved by using these reinforcing fibers in large composite resin restorations justifies further development and investigation of this technique. Bond strengths achieved, however, were not as high as those for composite resin bonded only to acid-etched enamel,15,l6 but the differences may indicate some of the complexities that are real for restorations. One difference may be that approximately one-half of the restoration was placed onto dentin and not solely on enamel as was the case in other studies. This difference would dramatically affect the bond strength of the specimens used in this study because adhesive dentin bonding is weaker than the mechanical bond to acid-etched enamel. As a result, any large composite resin restoration that is not solely bonded to enamel would have a lower bond strength. Examination of the mode of failure further suggests that the most important factor for enhancing bond strength would be improved dentin bonding. Adhesive failures of the DBA on dentin in both test and control specimens was shown by SEM. Similar adhesive failure patterns have been described by others, with localized smear layer removal also beingdocumented. 17-19The bond strength achieved for Celanese reinforced specimens approached that which might be expected for shear bond strength values from a flat surface, half of which was dentin. In the case ofthe control specimens, which were appreciablyweaker, other factors must be involved that influenced fracturing. Realistic hydration by storage in water for a period of 14 days may also have influenced the DBA adhesion to dentin. Huang and SolderholmZ0found that the Bondlite-dentin interface was not stable after water storage and surmised that the dentin surface itself could be the weak link in dentinal bonding. Similar findingswere shown by Bassiouny et al,21where storage periods of 1 to 30 days at different pH levels (acidic, neutral, and basic) reduced the Bondlite to dentin bond strength. However, recent bond strength improvements demonstrated by the newer dentin bonding agents should reduce this problem significantly.22 The porosities that were identified with SEM undoubtedly affected the restoration bond strength and contributed to thevariations seen in the data. Porosities found in the restorations may have been caused by air being entrapped during the mixing and application of the EBA, despite care taken during these procedures. These porosities may also partially contribute to the complexfracturing patterns seen. Subsurface porosities adjacent to the Celanese framework a t the bond interface could localize stress concentration in those areas. Stress concentration may arise during resin polymerization, and because the stress concentration remains high around the porosities, external and internal cracks 28
Bond Strength of Composite Resln Restorations
unlikely that this aspect of tooth preparation was important to this in vitro investigation, and it would not likely be relevant to the clinical situation.
7. Diamond M. Resin fiberglass bonded retainer. J Clin Orthod 1987: 21:182-183. 8. Friskopp J, Blomlof L. Soder PO. Fiber glass splints. J Periodontol 1979: 50:193-197. 9. Anderson L. Friskopp J. Blomlof L. Fiberglass splinting of traumatized teeth. J Dent Child 1983: 50:21. 10. Friskopp J. Blomlof L. Intermediate fiberglass splints. J Prosthet Dent 1984; 51:334-337. 11. Levenson MF. The use of a clear, pliable film to form a fiberglass-reinforcedsplint. JAmDentAssoc 1986: 112:7980. 12. Golub JE. The Manhattan Bridge - A new silk wrap technique. NY J Dent 1985: 56:226-228. 13. Golub JE. Silk wrapping has bond strength. Dent Today 1987: 6:24-27. 14. Braden M. Davy KWM. Parker S, Ladizesky NH. Ward IM. Denture base poly(methylmethacry1ate) reinforced with ultra-high modulus polyethylene fibers. Br Dent J 1988: 164:109-113. 15. Laswell HR. Welk DA, Regenos JW.Attachment of resin restorations to acid pretreated enamel. J Am Dent Assoc 1971: 82:558-563. 16. Eakle SA. Fracture resistance of teeth restored with class I1 bonded composite resin. J Dent Res 1986: 65:149-153. 17. Odin A, Olio G. Tensile bond strength of dentin adhesives. Dent Mater 1986: 2:207-213. 18. Retief DH, Gross J D , Bradley EL, Denys FR. Tensile bond strengths of dentin bonding agents to dentin. Dent Mater 1986; 2172-77. 19. Chan DCN. Reinhardt JW,Schulein TM. Bond strengths of restorative materials to dentin. Gen Dent 1985; 33:236238. 20. Huang G. Solderholm KJM. An in vitro investigation of the shearbond strength ofaphosphate based dentinalbonding agent. Scand J Dent Res 1989; 97234-92. 21. Bassiouny MA, Kame1 IL, Ellison SR. Simulated oral environment effect on stability of dentin adhesion promotors. J Dent Res 1989: 68:344. Abstr. No. 1299. 22. Fasanaro TS. Chairside instructor: resin-dentin bonding third generation materials. J Esthet Dent 1990; 2:89-92. 23. Bates D, Retief DH, Jamison HC. Effects of acid etch parameters on enamel topography and composite resinenamel bond strength. Pediatr Dent 1982; 4: 106-1 10. 24. Braden M. Personal communication. 25. Lynch E, Galan D. Tay WM. Enamel bonding - clinical aspects. Tay WM.Rear S, eds. General dental treatment. London: Churchill Livingstone. 1990: 1-9.
CONCLUSIONS The use of ultra-high modulus polyethylene (Celanese) fibers a s a reinforcement material increases the bond strength of Class N composite resin restorations. Further improvement appears possible, and the clinical value of these simple restorations justifies further investigation. In addition, this study indicates the peculiar importance of the palatal contour of incisal adhesive restorations for anterior teeth. Acknowledgments We thank Professor M. Braden of the London Hospital Medical College Dental School for providing the Celanese fibers and for his advice: Dr. D. Auger of the London Hospital Medical College Dental School for his SEM expertise: Kerr/Sybron UK Ltd. for their kind donation of the composite resin materials: Ms. L. Mayer for assisting in manuscript and figure preparation: and finally to the Canadian Fund for Dental Education (CFDE)for their financial assistance.
REFERENCES 1. Kilfoil BM, Hesby RA, Pellen GB. The tensile strength of a composite resin reinforced with carbon fibers. J Prosthet Dent 1983: 50:40-43. 2. Diehl MC, Hetzer MT. Effect of compounding technique on diametral tensile strength of fiber-reinforced composite resin. J Dent Res 1987: 66:209. Abstr. No. 819. 3. Griffee NC. House RC. Pellen GB. Tupp NW. Flexural strength of a composite resin reinforced with aramid fiber. J Dent Res 1988; 67:220. Abstr. No. 858. 4. Amoric M. Contention collee am fibres de carbone. Rev Orthop Dento Faciale 1988: 2 2 6 3 1 4 3 2 . 5. Sfeir E. Contention post-traumatique a la fibre de verre: la souplesse esthetique. Chir Dent Fr 1989: 59:5942. 6. Malquarti G. Bermet RG, Bois D. Prosthetic use of carbon fiber-reinforced epoxy resin for esthetic crowns and fixed partial dentures. J Prosthet Dent 1990; 63:251-257.