The degree and rate of photopolymerization in resin-based dental composites will significantly affect polymer network formation and resultant material properties that may determine their clinical success. This study investigates the mechanical properties, the generation of stress from polymerization, tooth cusp deflection and marginal integrity of experimental resin composites that contain different photoinitiators.
Experimental light-activated resin composites (60 vol% particulate filled in 50/50 mass% bis-GMA/TEGDMA) were formulated using a monoacylphosphine oxide (MAPO) photoinitiator and compared with a conventional camphoroquinone (CQ)-based system. Similar radiant exposure was used (18 J cm −2 ) for polymerization of each material although the curing protocol was varied (400 mW cm −2 for 45 s, 1500 mW cm −2 for 12 s and 3000 mW cm −2 for 6 s). Degree and rate of polymerization was calculated in real-time by near infrared spectroscopy and the generation of stress throughout polymerization measured using a cantilever beam method. Flexural strength and modulus were acquired by three-point bend tests. Standardized cavities in extract pre-molar teeth were restored with each material, the total cuspal deflection measured and post-placement marginal integrity between the tooth and restoration recorded.
Generally, MAPO- exhibited a significantly higher degree of conversion (72 ± 0.8 to 82 ± 0.5%) compared with CQ-based materials (39 ± 0.7 to 65 ± 1.6%) regardless of curing protocol ( p < 0.05) and MAPO-based materials exhibited less difference in conversion between curing protocols. CQ-based materials exhibited between ∼85 and 95% of the maximum rate of polymerization at <15% conversion, whereas MAPO-based RBCs did not approach the maximum rate until >50% conversion. Higher irradiance polymerization had a significant deleterious effect on the mechanical properties of CQ-based materials ( p < 0.05) whereas MAPO-based materials exhibited increased strength and modulus and were less affected by the curing method. Total cuspal deflection in restored extracted teeth was higher for CQ- compared with MAPO-based materials cured at the lowest irradiance curing protocol (12.9 ± 4.0 and 8.3 ± 1.5 μm) and similar at 3000 mW cm −1 for 6 s (10.1 ± 3.5 and 9.0 ± 1.5 μm). A significant decrease in marginal integrity was observed for CQ-based RBCs cured at high irradiance for short exposure time compared with that of the MAPO-based RBC cured using a similar protocol ( p = 0.037).
Polymer network formation dictates the final properties of the set composite and the use MAPO photoinitiators may provide an effective restorative material that exhibits higher curing speeds, increased degree of conversion, strength and modulus without compromise in terms of polymerization stress and marginal integrity between tooth and restoration.
The development of dental materials and associated technologies is often driven to satisfy the demands of the clinical practitioner to reduce chair-side operation time, rather than by sound materials science principles. One such example is the manufacture of light-curing units (LCUs) with high power outputs that are designed to significantly decrease the necessary exposure time required to effectively polymerize light-activated resin-based composite (RBC) materials.
Since the inception of visible-light cured RBCs in the 1970s , the power output of quartz-tungsten halogen light sources has increased from 35 to over 340 W producing LCUs that exhibit irradiances from ∼300 to 3000 mW cm −2 . Although the manufacture of halogen LCUs has mostly stopped due to imminent governmental legislation banning the use of incandescent light sources, the development of light-emitting diode (LED) technology has resulted in a similar increasing trend of LCU irradiance. During the 1990s, the first LED LCU types were produced using an array of multiple diodes to deliver sufficient power since each chip only provided ∼0.03–0.06 W . Thereafter, advancements in LED technology introduced single diodes with significantly increased output power (1–15 W) and currently, LCU manufacturers build devices that contain single or multiple diodes with measured irradiances of ∼1500–5000 mW cm −2 .
Photopolymerization of RBCs using extremely high irradiance is based on the premise that adequate cure can be achieved using substantially shorter exposure times (from 20 to 40 s to less than 10, but even as brief as 1 s in the case of some modern units), thereby reducing chair-side procedure time, which has been suggested to result in significant financial benefit . However, for the same radiant exposure (irradiance × time), the assumption that an increase in irradiance and reciprocal decrease in cure time will lead to equivalent material properties is flawed. Material properties following photopolymerization not only rely on curing protocols, but also on intrinsic characteristics such as monomer viscosity and radical mobility and therefore, such reasoning cannot be considered as a general rule. More specifically, there exists a large body of evidence suggesting that irradiance and time independently affect the mechanical and physical properties of cured resins and RBCs . Recent studies have also reported the significant effect of resin viscosity on polymerization kinetics of resins and RBCs cured using high irradiance . For low viscosity (unfilled) resins containing greater than or equal to 40% diluent (TEGDMA) and specific (commercial) “flowable” RBCs with low viscosity parent resin , “fast” high irradiance curing regimes severely limit polymer conversion. It is known that higher system mobility in less viscous media increases the probability of radical loss by bimolecular termination, and at higher irradiance that loss is greater than for similar radiant exposure achieved at lower light intensity for longer time . Consequently, if “ultra-fast” (<10 s) cured composites are desirable, alternative materials chemistry should be explored that may avoid the potential for under-cured and physically inferior resin-composite restorations.
Previous work by our group has identified the potential of alternative photoinitiators, such as monoacylphosphine oxide (MAPO) that may replace camphoroquinone (CQ) in contemporary RBCs . MAPO is a Type I photoinitiator that does not require a co-initiator and exhibits much greater molar absorptivity and polymerization efficiency compared with CQ, which, if considered with their specific absorption range and used with a curing light that exhibits an appropriate spectral output, may allow for a reduction in curing time without significant reduction of polymer conversion at high curing light irradiance . Although the degree and rate of polymer conversion determine important material properties, irradiance and exposure time are well known to independently affect polymer chain length and crosslinking, which may ultimately determine mechanical strength and stiffness and degradation of the cured RBC. Furthermore, significantly increased curing rates achieved with photoinitiators such as MAPO may decrease the ability of flow within the curing composite and increase the magnitude of stress resulting from polymerization, which may ultimately compromise the integrity of the tooth-restoration margin. Consequently, this work aims to determine whether the photoinitiator type and rate of polymerization affect mechanical properties, the developed stress magnitude as well as the tooth-material interface evaluated through ex vivo cuspal deflection and microleakage.
50/50 mass% of bisphenol-A glycidyl methacrylate (BisGMA) and triethyleneglycol dimethacrylate (TEGDMA) resins with equimolar concentrations (0.0134 mol dm −3 ) of the photoinitiator, camphoroquinone (CQ) and co-initiator, dimethylaminoethyl methacrylate (DMAEMA) (0.20/0.80 mass%) or monoacylphosphine oxide (MAPO) (0.42 mass%) were mixed in a lightproof container at 60 °C for 2 h. Silanised barium silicate glass fillers (55 vol% of 1 μm and 5 vol% of 0.05 μm average diameter particles) were mixed for one minute at 2300 rpm, followed by one minute at 3500 rpm using a centrifugal mixer (Speedmixer, DAC 150, Hauschild and Co. KG, Hamm, Germany).
Curing protocol/spectral irradiance and absorption
Specimens were cured with an 11 mm diameter tip halogen Swiss Master Light (EMS, Switzerland), chosen for its broad spectrum, overlapping the absorption spectrum of both photoinitiators (albeit only partially with MAPO). Emission spectra were measured by calibrating a UV-Vis spectrometer (USB4000, Ocean Optics) using a deuterium tungsten light source (calibrated to NIST standards) with known spectral output in the UV-VIS and NIR range (Mikropack DH2000-CAL, Ocean Optics, Dunedin, USA), which allowed absolute measurements of irradiance for the light curing unit (LCU). The absorption spectra of the photoinitiators were determined in methyl methacrylate (Sigma–Aldrich, UK) using the UV–vis spectrometer coupled to a standard cuvette holder in absorbance mode. Three different curing protocols were applied to compare the applicability of exposure reciprocity law at similar radiant exposure of 18 J cm −2 : 400 mW cm −2 for 45 s, 1500 mW cm −2 for 12 s and 3000 mW cm −2 for 6 s. For cuspal deflection measurements, only the extreme curing protocols were used.
Degree of conversion and rate of polymerization
Fourier Transform near infrared spectroscopy (FT-NIRS) was used to calculate polymer conversion and curing rates (Nicolet 6700, Thermo Scientific, Hemel Hemstead, UK) by monitoring the height of the peak at 6164 cm −1 , which corresponded with the vinyl CH 2 absorbance . Curing kinetics of the resin composites were measured for 70 s through a 12 mm × 1 mm cylindrical white Teflon mold placed on a glass slide and covered by a thin glass cover slip ( n = 3).
Flexural strength properties
In order to avoid the overlapping curing regime stipulated in ISO4049: 2008 for three-point flexural strength testing, which would result in increased radiant exposure received by the irradiation overlap along 25 mm bar-shaped specimens, 12.5 mm × 1 mm × 1 mm specimens were light cured in a single-shot from a distance of 1 mm. The cured specimens were tested 24 h post-irradiation using a universal testing machine (Model 5544, Instron Ltd., High Wycombe, Bucks, England) operating at a crosshead speed of 0.75 mm/min. Each specimen was centrally loaded using a small knife-edge indenter across a 10 mm support span and the flexural strength and modulus calculated by Eqs. (1) and (2) , respectively:
σ f = 3 F L / ( 2 B H 2 )
E f = F 1 L 3 / ( 4 B H 3 D )
where F is the maximum load (N), F 1 is the maximum load measured at a convenient point on the straight line portion of the load-deflection trace and D is the deflection of the specimen at that point (mm); L is the distance, in millimeters, between the supports (10 mm); B is the breadth (mm) and H is the height (mm) of the specimen.
The ‘Bioman’ device, designed at the University of Manchester , was used to evaluate the generation of stress throughout and following cure. Briefly, the apparatus consisted of a cantilever beam that embodies a 500 kg capacity load-cell and holds a steel rod (10 mm diameter, 22 mm length) perpendicular to the load-cell axis. The device was calibrated periodically using a series of calibration weights to provide a linear calibration plot of load-cell signal (mV) versus applied load (N). The resin composite specimen was introduced to the testing device on a 3 mm thick glass plate and sandwiched between this plate, via a rigid stainless steel platform (with an aperture to permit transillumination through the glass plate), and the rod, forming a 10 mm diameter by 0.8 mm thickness uncured material disk. Both the rod and glass plate surface were grit-blasted with 50 μm alumina particles to aid bonding between the rod–material–glass interfaces. Any excess material was removed and the specimen allowed to stabilize in dark conditions for 2 min at 23 °C prior to photopolymerization using the same curing protocols described in Section 2.2 . Throughout polymerization, the developing stress in the constrained and shrinking specimen caused micro-displacement at the free end of the load cell, which was continually measured via the load-cell signal (N) and converted into stress (MPa) using the specimen face area. The compliance of the device remained constant throughout testing (circa 6 μm/MPa). To consider polymerization stress data that corresponded to a higher stiffness load cell and lower compliance (that might be expected in tooth cavities that have less compliance and generate more stress) a correction factor of ×4 was used .
Cuspal deflection and microleakage
32 maxillary premolars free from caries, hypoplastic defects or cracks were subjected to calculus deposit removal using a hand-scaler and distributed into 4 groups ( n = 8); CQ-45 s@400, CQ-6 s@3000, MAPO-45 s@400 and MAPO-6 s@3000 mW cm −2 . The maxillary premolars were fixed into a cubic stainless steel mold and stored in 0.5% chloramine solution at 23 ± 1 °C until they were required for the extensive cavity preparation. Large standardized mesial-occlusal-distal (MOD) cavities were prepared under copious water irrigation. The width of the approximal box was two-thirds the bucco-palatal width (BPW) of the maxillary premolar, the occlusal isthmus was prepared to half the BPW and the cavity at the occlusal isthmus was standardized to a depth of 3.5 mm from the tip of the palatal cusp and 1 mm above the amelocemental junction at the cervical aspect of the approximal boxes. The cavosurface margins were all prepared at 90° and all internal line angles were rounded.
Following cavity preparation, the tooth surfaces were prepared for bonding with a three-step adhesive (All-Bond 2 ® Universal Adhesive System, Bisco Inc., Schaumburg, IL, USA). The MOD cavity preparation was air-dried for 30 s, prior to the application of a 32% phosphoric acid etching gel (Uni-Etch ® ) for 15 s without agitation before rinsing with water. Following a light drying with an air-syringe for 1 s, five consecutive coats of the primer (a mixture of All-Bond 2 ® Universal Dental Adhesive System Primer A (Ref B-2511, Lot 1000007217) and Primer B (Ref B-2512, Lot 1000007218)) was applied with a saturated brush tip until the surface appeared glossy. The primer mixture was lightly dried with an air-syringe for 5 s. A thin layer of bonding resin (D/E Resin, Ref B-2502A, Lot 1000007219) was applied to the primed enamel and dentin and light irradiated for 20 s with a QTH LCU (Optilux 501, Kerr Mfg. Co., Orange, CA, USA) operating in standard mode at a light intensity of 660 ± 32 mW cm −2 . Each MOD cavity was restored using an oblique incremental technique with the placement of three triangular-shaped increments (∼2 mm thickness) in the mesial approximal box, three triangular-shaped increments in the distal approximal box and two occlusal increments.
A Tofflemire matrix band was shaped and placed around the maxillary teeth prior to resin composite placement and care was taken to ensure the buccal and lingual cusps of the maxillary premolar teeth were free to contact the receptors of the twin channel deflection measuring gauge (Twin Channel Analog Gauge Unit, Thomas Mercer Ltd., St. Alban’s, UK). To ensure consistency in the measuring technique, the palatal measuring gauge was placed 2.5 mm from the palatal cusp tip prior to recording the baseline cuspal deflection measurement. Following each curing protocol, deflection measurements were recorded at 180 s post-irradiation for each increment and the combined total cuspal deflection measurement (the sum of the buccal and palatal cusp deflections) was calculated for each maxillary premolar tooth.
The restored maxillary premolar teeth were polished using a slow hand-piece under water with Sof-Lex Finishing discs (3M ESPE, St. Paul, MN, USA) and 15 μm grit Composhape finishing diamond burs (Intensiv, Viganello-Lugano, Switzerland). Following polishing, the root apices of the maxillary premolar teeth were sealed with sticky wax and all tooth surfaces were sealed with nail varnish (Rimmel 60 Seconds, London, UK) with the exception of a 1 mm band around the margins of each restoration surface. The teeth were thermocycled for 500 cycles between two water-baths maintained at 4 ± 1 and 65 ± 1 °C and involved submerging the teeth for 10 s in each water-bath with a 25 s transfer between water-baths. The thermocycled teeth were immediately immersed in 0.2% basic fuchsin dye for 24 h. The maxillary premolar teeth were then sectioned mid-sagitally in the mesio-distal plane using a ceramic cutting disk (Struers, Glasgow, Scotland) operating at 125 rpm under an applied load of 100 g. The sectioned teeth were examined under a stereo-microscope (Wild M3C, Heerburg, Switzerland) at 25× magnification. The extent of the cervical microleakage was recorded by assigning a score where ‘0’ was no evidence of dye penetration; a score of ‘1’ was superficial dye penetration not beyond the ADJ; a score of ‘2’ was dye penetration along the gingival floor and up to the axial wall; a score of ‘3’ was dye penetration along the axial wall and across the pulpal floor and a score of ‘4’ was dye penetration into the pulp chamber from the pulpal floor.
Group mean comparisons were made by two-way analysis of variance (ANOVA) for photoinitiator type and curing protocol and one-way ANOVA at a significance level of p = 0.05. Supplementary Post hoc Tukey tests were also conducted. Where necessary, a non-parametric Kruskal–Wallis test followed by paired group comparisons using Mann–Whitney U tests were used ( p = 0.05).