Flexural behavior of glass fiber reinforced composite wires with two monomer compositions compared to steel wire used as an orthodontic retainer

Abstract

The aim of this study was to investigate the mechanical properties of fiber reinforced composite (FRC) wires with different polymer matrices and compare them with steel wires commonly used in orthodontic retention. Eight groups of the FRC wires (continuous unidirectional E-glass) and a control group of steel metal Penta One 0.0215′′ were tested with a 3-point bending test. The FRC wire groups consisted of two thicknesses of fiber rovings (300-tex and 600-tex) which were impregnated with a light-curing monomer resin system of bis-GMA/PMMA or bis-Mepp/dimetacrylate/prosphoric ester monomer. The bending was continued until breakage of the specimen or to the strain of 3 mm using a span length of 10 mm and cross-head speed 1.00 mm/minute. The data were analysed using analysis of variance (ANOVA). The maximum load values of the FRC wire groups varied between 1.3 and 20.0 N, and the control group was 2.4 N. Specimens of 600-tex groups had considerably higher load values than 300-tex groups. The load value of the control steel group was close to the load value of in the 300-tex groups. Bis-GMA/PMMA impregnated FRC demonstrated higher values of maximum load than bis-Mepp/dimethacrylate/prosphoric ester monomere resin impregnated FRC.

Keywords

FRC-wires orthodontic retainer

1. Introduction

Bonded retainers are frequently used after orthodontic treatment with fixed appliances [1]. Common indications for bonded retainers are to prevent relapse, crowding of the incisors, space reopening and rotational relapse of the teeth [2]. Different designs of bonded lingual retainers exist, having different wires, sizes and composition [1,3,4]. Most used bonded retainer types are made of a thick (0.030 or 0.032′′, corresponding to 0.752 or 0.0813 mm) plain metal bar bonded only to the canines or a thin (0.0215 or 0.0175′′, corresponding to 0.0546 or 0.445 mm) flexible, spiral wire bonded to each tooth [5–7].

Bonded retainers are independent of patient co-operation, and they are also quite invisible. Flexible spiral wire retainers allow little movement of all bonded teeth but if they are not bonded passively onto the enamel surface, they may also cause unwanted tooth movement [2,8–10]. At the moment, multistranded retainers are the gold standard for orthodontic retention, though periodontal complications have been reported [1,2,11].

However, metal retainers have some limitations: they are often esthetically compromised, particularly if used labially, and they can cause allergic reactions. Alternative retainers using a polyethylene fiber ribbon-reinforcement or a glass fiber reinforcement have therefore been developed. Of these, everStick^® Ortho contains unidirectional E-glass fibers which are pre-impregnated with a resin system of polymethylmethacrylate (PMMA) and bisphenol-A-glycidyl dimethacrylate bis-GMA. Angelus Interlig^® consists of E-glass and bis-GMA. DentaPre Splint^® contains E-glass, dimetachacrylates, initiators and stabilizers. Ribbond^® contains ultra-high molecular weight polyethylene (UHMWP) [12]. However, some investigators have reported lower success rates for glass fiber reinforced retainers [11] and polyethylene ribbon-reinforced retainers [7] compared to multistranded metal retainers, while others have not found significant differences between glass fiber reinforced retainers and multistranded metal retainers [13,14]. Glass fiber bonded retainers have a certain level of flexibility, but interdental areas are normally covered and splinted by particulate filler resin composite, which lower the flexibility of the glass fiber retainer system.

Shear stress typically predominates at the approximal regions, and glass fibers are not optimal fibers in resisting the fractures caused by shear stress. Difficulties in placing fiber reinforced retainers can be a further cause of higher failure rates compared to multistranded metal retainers [11].

The purpose of this study was to investigate the mechanical properties of commercial conventionally impregnated everStick Ortho wire and a new experimental impregnated FRC wire with another resin system and compare them with Penta One 0.0215′′ wire, which is commonly used in orthodontic retention. The null hypothesis was that the type of polymer matrix composition has an effect on the force level required to bend the FRC wires.

2. Materials and methods

Eight groups of FRC-wires and the control group of metal Penta One five stranded 0.0215′′ wire (Massel, Carlsbad, United States, LOT F1207164) were tested with a 3-point bending test (Fig. 1) to determine the mechanical properties of materials in bending (Table 1). The FRC wire groups consisted of two thicknesses of fiber roving materials (tex-300 or tex-600 E-glass fibers). Fibers were impregnated with two types of light curing resin systems: bisphenol-A-glycidyl dimethacrylate (bisGMA)-polymethylmethacrylate (PMMA) monomer system (equivalent with everStick Ortho resin system), and 2.2-bis(4-methacryloxypolyethoxyphenol)propane(bis-MEPP)-dimethacrylate-prosphoric ester monomer system. In both resins, initiator-activator system was based on camphorquinone and amine. There were also additional test specimens (Group “Coated”), which were coated with a layer of PMMA by a solvent process. There were seven specimens (length 15 mm) in each group ( $n = 7$ ). The diameter of the FRC specimens was measured with a digital sliding gauge. The diameters were as follows: tex 300 groups between 0.49–0.87 (SD 0.060–0.122) mm, tex 600 groups between 0.63–0.95 (SD 0.037–0.10) mm, and Penta One group 0.53 mm (Table 1).

Fig. 1.

The loading conditions. Schematic representation of the loading conditions of test specimens. Dimensions are in millimetres. For the diameter (d) of specimen of groups with fiber rovings of different text numbers, see Table 1.

Table 1

The content of FRC groups and the control group

Groups	Impregnation resin system	Fiber	Diameter, mm (SD)
1. F300coat-Hand	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-300 E-glass	0.49 (0.060)
2. F300no coat-Hand	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-300 E-glass	0.48 (0.016)
3. F300coat-Oven	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-300 E-glass	0.48 (0.026)
4. F300no coat-Oven	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-300 E-glass	0.46 (0.025)
5. E300coat-Hand	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-300 E-glass	0.77 (0.122)
6. E300no coat-Hand	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-300 E-glass	0.87 (0.105)
7. E300coat-Oven	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-300 E-glass	0.57 (0.066)
8. E300no coat-Oven	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-300 E-glass	0.65 (0.027)
9. F600coat-Hand	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-600 E-glass	0.67 (0.048)
10. F600no coat-Hand	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-600 E-glass	0.63 (0.064)
11. F600coat-Oven	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-600 E-glass	0.64 (0.037)
12. F600 no coat-Oven	bis-Mepp 40–50%, Dimetacrylate 45–55%, Phosphoric ester monomer 5–10%, Photo initiator and inhibitor	tex-600 E-glass	0.63 (0.051)
13. E600coat-Hand	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-600 E-glass	0.85 (0.095)
14. E600no coat-Hand	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-600 E-glass	0.74 (0.063)
15. E600coat-Oven	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-600 E-glass	0.95 (0.57)
16. E600no coat-Oven	bis-GMA 85–95%, PMMA 5–15%, Photo initiator and inhibitor	tex-600 E-glass	0.88 (0.10)
17. Control: Penta One	no impregnation	five stranded stainless steel	0.53

The FRC groups and the control group: the glass contents, the diameter: (mm) and standard deviation (SD), the impregnation resin systems, the coating with PMMA or no coating, and the curing method (with hand or additionally with post-curing oven). F-group refers to bisMepp/dimethacrylate/prosphoric ester monomer group, and group refers to bis-GMA/PMMA group. ‘Coat’ refers to coating with PMMA, ‘no coat’ refers to no-coating. ‘Hand’ refers to polymerization by hand light-curing unit, and ‘oven’ refers to post-curing by light and heat. Monomer compositions: 2.2-bis(4-ethacryloxypolyethoxyphenol)propane(bis-MEPP), bisphenol-A-glycidyl dimethacrylate (bis-GMA), polymethylmethacrylate (PMMA).

Resin matrix of specimens were light-cured with 3M Espe Elipar S10 (light intensity 1200 mV, Serial nr. 939123002122, 210 V/50/60 Hz) for 2 times 20 s (Group “Handunit”). In addition, there was a group of specimens which had undergone post-curing by heat with TargisPower (IVOCLAR, Shaan, FL) light-curing oven for 25 minutes, where the temperature rose gradually to +90°C (Group “Oven”). After curing, the specimens were placed into Falcon tubes (vol. 50 ml), which were filled with 25 ml distilled water (Mill-Q) purity 18.2 mΩ cm⁻¹. The tubes were then placed in to an incubator +37°C for 30 days.

The 3-point bending test was carried out with material testing machine Lloyd (Lloyd LRX30, Lloyd Instruments, Fareman, UK, load cell 250 N). The specimens were tested in air immediately after 30 days in water storage. Metal wires of the control group were stored and tested dry in air. The bending test was continued until the initial breakage of the specimens, that is where no visual fracture of the specimen could have been seen. The span length between the supports was 10 mm, and the test cross-head speed was 1.0 mm/min, preloading the specimen for the deflection of 0.05 mm, preload speed 5 mm/min. The test stopped at extension 3 mm. Load required to bend the specimen for 3.0 mm after preloading was registered as maximum load value.

The statistical analysis was completed by using analysis of variance (ANOVA) and SPSS 16.0 (SPSS Inc, USA) factor analysis. The independent variables were: thickness of FRC wire, composition of polymer matrix, presence of coating, and curing type (hand or hand and additional oven curing). The dependent variable was bending required to fracture the wire or case 3 mm deflection (strain).

3. Results

ANOVA revealed that the composition of polymer matrix, curing method, wire thickness, and coating had a significant (p < 0.05) effect on the force level and deflection values of FRC wires.

The descriptive results for the maximum load values of the 17 groups are shown in Figs 2 and 3. The load values in the FRC-wire groups varied between 1.3 and 20.0 N, and in the control metal wire group the load value was 2.4 N. 600-tex group FRC wires had much higher load values than 300-tex group wires, and the load value of control metal group was close to that of 300-tex FRC wires. In all post-curing groups, load values were higher than in the hand-curing groups. In all 600-tex FRC wires, the additional PMMA coating increased the load values, whereas in 300-tex FRC-wires the coating did not have an influence on load values with one exception (experimental 300-tex FRC wire group). In all bis-GMA/PMMA FRC wire groups, load values were higher, but they were also thicker than the comparable bis-Mepp/dimethacrylate/prosphoric ester monomer groups. The deflection at maximum load varied between 0.82 and 1.14 mm in the additionally PMMA coated FRC wires groups and between 0.79 and 1.23 mm in the non-coated FRC wires groups. With metal wires, the deflection was 1.29 mm (Figs 4 and 5). The deflection was higher in all post-cured FRC wires than those light-curing only, with the exception that in the non-coated 600-tex FRC wire group of bis-GMA/PMMA polymer matrix, the light-curing group had somewhat higher deflection values. Examples of the load–deflection curves of tex-600 FRC wires with additional coating PMMA are presented in Fig. 6, and the load-deflection curve of tex-300 FRC-wires (non-coated) and metal wires are in Fig. 7.

Fig. 2.

The maximum load values of the coated FRC wires and the metal Penta One 0.215′′ wire. The load values are before the initial breakage of the specimen with standard deviation (vertical bars). ‘Handunit’ refers to polymerization by hand light-curing unit and ‘oven’ refers to post-curing by light and heat. F-group refers to bisMepp/dimethacrylate/prosphoric ester monomer group, and E-group refers to bis-GMA/PMMA group.

Fig. 3.

The maximum load values of the no-coated FRC-wires and the metal Penta One 0.215′′. The load values are before the initial breakage of the specimen or to the extension of 3 mm with standard deviation (vertical bars). ‘Handunit’ refers to polymerization by hand light-curing unit, and ‘oven’ refers to post-curing by light. F-group refers to bisMepp/dimethacrylate/prosphoric ester monomer group, E-group, and refers to bis-GMA/PMMA group.

Fig. 4.

The deflection at maximum load of coated fibre reinforced composite (FRC) wires and the Penta One metal 0.215′′ wire. The deflection (millimeter) at maximum load (Newton) before the initial breakage or to the extension of 3 mm of the coated fibre reinforced composite (FRC) wires and the Penta One metal 0.215′′ wire with standard deviations (vertical bars). ‘Handunit’ refers to polymerization by hand, and ‘oven’ refers to post-curing by light and heat. F-group refers to bisMepp/dimethacrylate/prosphoric ester monomer group, an E-group refers to bis-GMA/PMMA group.

Fig. 5.

The deflection at maximum load of the no-coated fiber reinforced composite (FRC) wires and the Penta One metal 0.215′′ wire. The deflection (millimeter) at maximum load (Newton) before the initial breakage or to the extension of 3 mm of the no-coated fiber reinforced composite (FRC) wires and the Penta One metal 0.215′′ wire with standard deviations (vertical bars). ‘Handunit’ refers to polymerization by hand, and ‘oven’ refers to post-curing by light and heat. F-group refers to bisMepp/dimethacrylate/prosphoric ester monomer group, and E-group refers to bis-GMA/PMMA group.

Fig. 6.

The load-deflection curves for tex-600 coated groups. The examples of the load-deflection curves for tex-600 coated groups from below to top: (a) a dotted line/hand-cured bis-Mepp/dimethacrylate/prosphoric ester monomer group; (b) a cutted line/oven cured bis-Mepp/dimethacrylate/prosphoric ester monomer group; (c) a two dotted segment line/hand-cured bis-GMA/PMMA group; (d) segment dotted line/oven cured bis-GMA/PMMA.

Fig. 7.

The load-deflection curves for tex-300 no coated groups. The examples of the load-deflection curves for tex-300 no coated groups and the Penta One 0.0215′′ group from below to top: (e) a straight line/Penta One 0.0215′′ group, (a) a dotted line/hand-cured bis-Mepp/dimethacrylate/prosphoric ester monomer group; (b) a cutted line/oven cured bis-Mepp/dimethacrylate/prosphoric ester monomer group; (c) a two dotted segment line/hand-cured bis-GMA/PMMA group; (d) segment dotted line/oven cured bis-GMA/PMMA group.

4. Discussion

One major problem of FRC retainers is said to be their low flexibility [11]. Interdental areas of splinted teeth are filled with resin composite, which increases the diameter of the connecting composite system between the teeth. The multistranded bonded metal retainers are bonded to teeth, but interdental areas are not coated with composite which allow normal physiologic movement of teeth due to the bending of metal wire. In this investigation, the force levels of two different monomer compositions were compared with an attempt to lower the stiffness of the FRC material be closer to the of multistranded metal wire. There were two thicknesses of FRC wires, and they were coated or not coated with PMMA. PMMA as linear polymer has lower stiffness than cross-linked dimethactryate systems. Our control group was the commonly used Penta One 0.0215′′ metal wire. Because a multistranded metal wire is a gold standard for retention, the purpose of this study was to find whether FRC wire of various kinds could provide similar bending behavior.

Some investigations have indicated that the type of composite affects bond values of multistranded metal wire retainers and FRC retainers to tooth surface, possibly influencing the survival of the retainer [15,16]. Some investigations have suggested that one reason for lower success rates may be the type of composite used with FRC-wires. In Scribante’s research [15] the FRCs (evertStickORTHO, Stick Teck, Ltd, Turku, Finland) were bonded to bovine incisors with three different adhesive systems. Transbond XT^® exhibited the best bond strength 23.6 (SD 3.1) MPa compared to Tetric flow^® 13.4 (SD 3.1) MPa and Fuji Ortho LC^® 12.5 (SD 2.9) MPa [15]. However Reynolds [17] assumed that bonded orthodontic appliances should be able to resist a 5-8 MPa strength. and according to the Scribante’s study all groups were above that level. However, variations in testing methods of bond strength make direct comparison of the results difficult.

Most investigations have tried to achieve high strength values for FRC wires but perhaps flexibility of the wire is a more important property. In this study, flexibility of the FRC wire was intentionally influenced by the selection monomer systems for polymer matrix of FRC wire. Polymer matrix binds the glass fibers and forms the composite, and thus, the flexibility of the polymer matrix can considerably influence the physical properties of FRC wire. FRC wires with bis-GMA/PMMA monomer impregnation showed higher stiffness for the FRC wire than that with bis-Mepp monomer impregnation. This could be explained by differences in the molecular structure of the monomers. However, it should be noted that Bis-GMA is a more viscous monomer than bis-Mepp, which caused FRC wire to become larger in diameter than with bis-Mepp monomer system, which could have influenced the results.

In this investigation it was shown that all 600-tex FRC wire groups had considerably higher load values than 300-tex FRC wire groups and that the oven post-curing increased the strength even more. The effect of thermal post-curing on the stiffness and strength of resin composites has been documented previously [18–20]. Long-term retention of mechanical properties of the FRC wires is a fundamental requirement for orthodontic applications. In general, FRCs retain their mechanical properties for a long period of time if there is no degradation of the polymer matrix, no significant dissolution of glass fibers, and no deterioration of the silane promoted adhesion between glass fibers and polymer matrix. There are studies, which demonstrate influence of these aspects on the mechanical behavior in provocated aging conditions and even in ten years of in vitro conditions [21–23]. However, with regards to the stability of the presently used monomer compositions and corresponding polymers, bis-Mepp should be investigated in more detail as there is currently no information available on the hydrolytic stability of that specific polymer.

If FRC wires of the present study are compared with 0.0215′′ steel Penta One wires, it can be concluded that 300-tex FRC wires impregnated with monomer systems of bis-Mepp provided similar load values and deflection. On the other hand, commercial everStick Ortho wires proved to be considerably stiffer than Penta One 0.0215′′. Consequently, it is possible that the FRC wire stiffness, which may have contributed to the failures of everStick Ortho retention wires. One possible method of reducing this stiffness might lie in changing the monomer system to bis-Mepp. This however requires further preclinical and clinical testing.

5. Conclusion

Composition of the polymer and diameter of the FRC wire had an effect on the level of force required to bend the wire, and thus on the stiffness of the wire. Bis-Mepp/dimethacrylate/prosphoric ester monomer impregnated 300-tex FRC wires provided similar mechanical behavior compared to the control wire of metal (Penta One’s 0.0215′′).

Conflict of interest

Author ES is RD Director at Stick Tech Ltd – Member of GC Group and author PV consults Stick Tech Ltd – Member of GC Group in RD.

References

Bearn

D.R.

, Bonded orthodontic retainers: A review, Am. J. Orthod. Dentofacial Orthop. 108(2) (1995), 207–213.

Årtun

, Caries and periodontal reactions associated with long-term use of different type of bonded lingual retainers, Am. J. Orthod. 86 (1984), 112–118.

Zachrisson

B.U.

, Clinical experience with different bonded orthodontic retainers, Am. J. Orthod. 71 (1977), 440–448.

Årtun

Spadafora

A.T.

Shapiro

P.A.

McNeill

R.W.

and Chapko

M.K.

, Hygiene status associated with different types of bonded orthodontic canine to canine retainers, J. Clin. Periodontol. 14 (1987), 89–94.

Littlewood

S.J.

Millett

D.T.

Douleday

Bearn

D.R.

and Worthinton

H.V.

, Retention procedures for stabilising tooth position after treatment with orthodontic braces, Cochrane Database Syst. Rev. 1 (2004), CD002283.

Zachrisson

B.U.

and Büyükyilmaz

, Bonding in orthodontics, in: Orthodontics: Current Principles and Techniques Graber

L.W.

, ed., 4th edn, Mosby, St. Louis, 2005, pp. 621–659.

Rose

Frucht

and Jonas

I.E.

, Clinical comparison of a multistranded wire and a direct-bonded polyethene ribbon-reinforced resin composite used for lingual retention, Quintessence Int. 33 (2002), 579–783.

Årtun

and Thylstrup

, Clinical and scanning electron microscopic study of surface changes of incipient caries lesion after debonding, Scan. J. Dent. Res. 94(3) (1986), 193–201.

Dahl

E.H.

and Zachrisson

B.U.

, Long term experience with direct bonded lingual retainers, J. Clin. Orthod. 25(10) (1991), 619–630.

10.

Årtun

Spadafora

A.T.

and Shapiro

P.A.

, A 3-year follow-up study of various types of orthodontic canine-to-canine retainers, Eur. J. Orthod. 19(5) (1997), 501–509.

11.

Tacken

P.E.

Cosyn

De Wilde

Aerts

Govaerts

and Vannet

B.V.

, Glass fibre reinforced versus multistranded bonded orthodontic retainers: A 2 years prospective multi-centre study, Eur. J. Orthod. 32(2) (2010), 117–123. doi:10.1093/ejo/cjp100.

12.

Foek

D.L.

Özcan

Krebs

and Sandham

, Adhesive properties of bonded orthodontic retainers to enamel: Stainless steel wire vs fibre-reinforced composites, J. Adhes. Dent. 11(5) (2009), 381–390.

13.

Sfondrini

M.F.

Fraticelli

Castellazzi

Scribante

and Gandini

, Clinical evaluation of bond failures and survival between mandibular canine-to-canine retainers made of flexible wire and fiber-reinforced composite, J. Clin. Exp. Dent. 6(2) (2014), 145–149. doi:10.4317/jced.51379.

14.

Bolla

Cozzani

Doldo

and Fontana

, Failure evaluation after a 6-year retention period: A comparison between glass fiber-reinforced (GFR) and multistranded bonded retainers, Int. Orthod. 10(1) (2012), 16–28. doi:10.1016/j.ortho.2011.12.005.

15.

Scribante

Cacciafesta

and Sfondrini

, Effect of various adhesive systems on the shear bond strength of fiber reinforced composite, Am. J. Orthod. Dentofacial Orthop. 130(2) (2006), 224–227.

16.

Raju

P.S.

Gupta

Grag

Bhattacharya

and Agarwal

D.K.

, Evaluation of shear bond strength of fiber-reinforced composite using different adhesive systems, J. Dr. NTR Univ. Health Sci. 1(4) (2012), 249–252. doi:10.4103/2277-8632.105117.

17.

Reynolds

I.R.

, A review of direct orthodontic bonding, Br. J. Orthodont. 2 (1979), 171–178.

18.

Lassila

L.V.J.

Nohrström

and Vallittu

P.K.

, The influence of short-term water storage on the flexural properties of unidirectional glass fiber-reinforced composite, Biomaterials 23(10) (2002), 2221–2229.

19.

Lassila

L.V.J.

and Vallittu

P.K.

, The effect of fiber position and polymerization condition on the flexural properties of fiber reinforced composite, J. Contempt. Dent. Pract. 5(2) (2004), 14–26.

20.

Dyer

S.R.

Lassila

L.V.J.

Jokinen

and Vallittu

P.K.

, Effect of fiber position and orientation on fracture load of fiber-reinforced composite, Dent. Mater. 20(10) (2004), 947–955.

21.

Bouillaguet

Schutt

Alander

Vallittu

P.K.

Schwaller

Buerki

Michler

Cattani-Lorente

and Krejci

, Hydrothermal and mechanical stresses degrade fiber-matrix interfacial bond strength in dental fiber-reinforced composites, J. Biomed. Mater. Res. B Appl. Biomater. 76(1) (2006), 98–105.

22.

Vallittu

P.K.

, Effect of ten years of in vitro aging on the flexural properties of fiber-reinforced composites, Int. J. Prosthodont. 20(1) (2007), 43–45.

23.

Matinlinna

J.P.

Dahl

J.E.

Karlsson

Lassila

L.V.J.

and Vallittu

P.K.

, The effect of the novel silane system to the flexural properties of E-glass fiber reinforced composites for dental use, in: Silanes and Other Coupling Agents, Vol. 5, Koninklijke Brill NV, The Netherlands, 2009, pp. 107–121.