Properties of A Model Self-Healing Microcapsule-Based Dental Composite Reinforced with Silica Nanoparticles

Aim: The purpose of this study was to evaluate the mechanical properties of an experimental self-healing dental composite model (SHDC) composed of SiO2 nanoparticles with varying percentages of triethylene glycol dimethacrylate (TEGDMA) monomer and N,N-dihydroxyethyl-p-toluidine (DHEPT) amine microcapsules. Materials and methods: Microcapsules were prepared by in-situ polymerisation of PUF shells, as explained in our previous work. The model SHDC included bisphenol A glycidyl dimethacrylate (Bis-GMA:TEGDMA) (1:1), 1 wt% phenyl bis(2,4,6-trimethylbenzoyl) phosphine oxide (BAPO), 0.5 wt% benzoyl peroxide (BPO) catalyst, 20 wt% silanised silica dioxide (SiO2) (15 nm) and (0, 2.5, 5, 7.5, 10 wt%) of microcapsules (120 ± 45 μm). Light transmission, hardness, degree of conversion (DC), flexural strength and elastic modulus of the SHDC model were measured. Results: The degree of conversion of the SHDC ranged from 73 to 76% 24 h after polymerisation. Hardness measurements ranged from 22 to 26 VHN (p > 0.05); however, the flexural strength was adversely affected from 80 to 55 MPa with increasing microcapsules of up to 10 wt% in the composites (p < 0.05). Conclusion: Only flexural strength decreased drastically ~30% with increasing microcapsules (>10 wt%) in the composites. All other measured properties were not significantly affected. Accordingly, we recommend a stronger composite material that could be created by increasing the filler content distribution in order to achieve a hybrid self-healing composite with enhanced mechanical properties.


Introduction
Dental composites are commonly used as the restorative material of choice in contemporary dentistry [1,2]. Composites can bond to the tooth structure via bonding agents to repair destructed or decayed teeth with acceptable aesthetics, satisfactory mechanical properties and ease of use [3]. Although substantial efforts have been made to improve the clinical performance and properties of resin composites [4][5][6][7], composite restorations have been shown to encounter two main downsides: secondary caries and bulk fracture [8]. Survival rate for composites has been reported to be 91.7% at 5 years and 82.2% at 10 years [9][10][11]. In many cases, composite failure can be attributed to the accumulation of microcracks, which are generated from masticatory forces and thermal stresses [12].
The introduction of self-healing composites and polymers involving microcapsule systems has shown that providing long-life structural materials is possible [13]. Selfhealing composites can be understood as materials that have the capability to recover when mechanical damage occurs [13,14]. The potential for crack repair and recovery of

Preparation of Self-Healing Dental Composite (SHDC)
The organic phase consisted of Bis-GMA:TEGDMA mixed at a (1:1, w/w) ratio. BAPO photo initiator was added at 1 wt% and self-healing initiator BPO was also added at 0.5 wt% to the mixture. The inorganic phase consisted of SiO 2 nanoparticles that were coated with silane coupling agent and dispersed at 20 wt%, with an average particle size of 15 nm. Microcapsules were dispersed gently into the experimental photo-activated composites with low-speed mechanical mixing at the following percentages: 0, 2.5, 5, 7.5 and 10 wt% (Table 1). The microcapsules were prepared by in-situ polymerisation of poly(urea-formaldehyde) shells encapsulating TEGDMA monomer and 1 wt% DHEPT amine as healing agents. The average microcapsule diameter was between 150 and 300 µm.

Light Transmittance Measurements
Five cylindrical specimens were prepared in each experimental composite group using a PTFE disc mould (4 mm internal diameter, 2 mm height). The mould was placed on a glass slab (2 mm thick) with Mylar strips (Moyco Union Broach, York, ME, USA) on the bottom and top surfaces of the mould and pressed with another glass slab on the top to remove excess resin. One increment was placed for each specimen during preparation.
The prepared SHDC composites were first placed and cured for 20 s in a MARC TM spectrophotometer resin calibrator (BlueLight Analytics Inc., Halifax, NS, Canada). The light curing unit used was a 455 (±10) nm LED light (S10 Elipar, 3M ESPE, St Paul, MN, USA). Prior to experimentation, the maximum light irradiance of the top sensor of the resin calibrator was recorded, and the mean irradiance over 20 s duration was reported. After curing, specimens were demoulded and incubated for 24 h at 37 • C prior to hardness evaluation.

Hardness Measurements
A hardness instrument (FM-700, Future-Tech Corp., Kawasaki, Kanagawa, Japan) was used to measure the top and bottom surfaces for the formulated composites. A Vickers indenter was used with an applied load of 300 gf (2.94 N) for 15 s. Five indentations were measured for top and bottom surfaces, with a distance of four times the diagonal length of the indenter between each indentation (4D, around 200 nm). In order to have standardised readings, indentations were made randomly in the matrix or microcapsules surfaces. The mean of five readings for each specimen was recorded. In addition, the means and the standard deviations of the five specimens in each group were calculated and recorded.

Degree of Conversion Measurements
A Fourier transfer-infrared spectrometer (FT-IR) was used to measure the unpolymerised carbon-carbon double bond (Avatar 360, Nicolete Analytical Instrument, Thermo Electron Corp., Cambridge, UK). The spectra of five specimens in each group were recorded. A mould (PTFE disc, 4 mm internal diameter, 0.5 mm height) was placed on top of the diamond ATR crystal reader in the FT-IR, filled with resin composite paste, covered on the top surface with a Mylar strip and pressed to remove the excess resin. It was then photo cured for 40 s using a 455 (±10) nm LED light.
The DC was reported using the spectra of the composite immediately post-cure and 24 h after polymerisation. Specimens were stored in distilled water and incubated at 37 • C for 24 h. The means and standard deviations of the post-polymerisation spectra were recorded.
The spectra of the composites were recorded over the range of 4000 to 400 cm −1 , with 32 scans at a resolution of 4 cm −1 . Software was used to record the peak intensity ratio of aliphatic C=C at 1637 cm −1 against the internal reference immediately post-cure and 24 h after polymerisation. Meanwhile, the aromatic C=C peak at 1608 cm −1 was selected as the internal reference.
The following equation was used to calculate the degree of conversion (Equation (1)): peak height a f ter curing Monomer(C = C/C = O)peak height be f ore curing × 100 (1) Equation (1): degree of conversion from peaks on the FT-IR spectrum (C = C aliphatic 1637 cm −1 , C = C aromatic 1608 cm −1 ).

Flexural Strength Measurements
The prepared SHDC composite paste was placed in a PTFE mould (2 mm × 2 mm × 25 mm) on a glass slab (2 mm thick) separated with Mylar strips on the bottom and top surfaces of the mould, then pressed with another glass slab on the top to remove the excess resin. Each resin specimen was photo cured in five overlapping points for 20 s on each open side of the mould using a 455 (±10) nm LED light. Upon light curing from one point to another on the specimen, the curing tip was positioned on half the diameter of the cured area by the earlier exposure to ensure optimal polymerisation of the entire specimen. After curing, specimens were demoulded and sharp edges were smoothed with P800 grit SiC abrasive paper. All specimens were stored in distilled water at 37 • C for 24 h prior to testing.
A computer-controlled universal testing machine (Zwick/Roell Z020, Ulm-Einsingen, Germany) was used to measure the flexural strength of the specimens in a three-point flexure bending test that used a span of 20 mm, a crosshead speed of 1 mm/min and a load cell of 500 N. Prior to testing, five specimens in each group were removed from the storage medium (distilled water). The thickness of each composite specimen was measured at the centre and the two ends of the specimen using a digital caliper (Mitutoyo, Tokyo, Japan), and the mean thickness was used for cross sectional area estimations.
The following equation was used to calculate the flexural strength (MPa) (Equation (2)): Equation (2): S-flexural strength; P max -the load at failure; L-span; bh-specimen width and thickness.
The elastic modulus was determined from the slope of the linear region of the load deflection curve, and the following formula was used to calculate the elastic modulus (GPa) (Equation (3)): Equation (3): E-elastic modulus; P-load; d-displacement of the slope in the linear elastic region of the load displacement curve.
The fractured surfaces of selected composite specimens, which were prepared by sputter coating with gold (7 nm), were examined by SEM (Zeiss EVO60, Oberkochen, Germany), to investigate the fracture mechanism of the microcapsules in the resin composite at the fracture plane.

Statistical Analysis
Collected data were statistically analysed using SPSS software (IBM SPSS Statistics 23.0, SPSS Inc., Chicago, IL, USA). The Shapiro-Wilk test was implemented to confirm the normality of the data. Levene's test also confirmed the equality of variance.
One-way analysis of variance (ANOVA) and independent sample t-tests were conducted to detect any significant effects of variables for the study groups at a significance level of (p ≤ 0.05). All data were plotted using Sigma Plot (SigmaPlot 13.0, Systat Software Inc., San Jose, CA, USA).

Light Transmittance Measurements
The maximum light irradiance measurement of the light curing unit at the top sensor of the resin calibrator was 2050 (±24) mW/cm 2 over 20 s prior to experimentation. The light transmittance through the SHDC did not show any statistical differences between any of the groups (p > 0.05). The light transmittance through the 2 mm-thick specimens was found in the control group's MC 0 composite at 698 (±26) mW/cm 2 , while the MC 10 composite showed a lower value of 641 (±42) mW/cm 2 ( Figure 1).
One-way analysis of variance (ANOVA) and independent sample t-tests were con ducted to detect any significant effects of variables for the study groups at a significance level of (p ≤ 0.05). All data were plotted using Sigma Plot (SigmaPlot 13.0, Systat Software Inc., San Jose, CA, USA).

Light Transmittance Measurements
The maximum light irradiance measurement of the light curing unit at the top sensor of the resin calibrator was 2050 (±24) mW/cm 2 over 20 s prior to experimentation. The ligh transmittance through the SHDC did not show any statistical differences between any o the groups (p > 0.05). The light transmittance through the 2 mm-thick specimens was found in the control group's MC0 composite at 698 (±26) mW/cm 2 , while the MC10 compo site showed a lower value of 641 (±42) mW/cm 2 ( Figure 1).

Degree of Conversion Measurements
The mean degree of conversion values immediately after curing the (DC0) of MC2.5 to MC10 microcapsules in the composite was around 67%, with no statistical differences be tween test groups (p > 0.05). The DC0 value of 68% was shown in the MC5 group. However the control group, MC0, showed a value of 65% at DC0, which was statistically differen compared to MC2.5 and MC5 (p < 0.05). The mean degree of conversion 24 h after polymer isation (DC24) was between 73 and 76%, with the highest score being for the control group MC0. The DC24 values showed no statistical differences between any of the groups (p > 0.05).
Intragroup comparisons were also performed by independent sample t-tests between DC0 and DC24. There were statistically significant differences in all composite groups be tween the DC0 and DC24 measurements (p < 0.05) ( Figure 2).

Degree of Conversion Measurements
The mean degree of conversion values immediately after curing the (DC 0 ) of MC 2.5 to MC 10 microcapsules in the composite was around 67%, with no statistical differences between test groups (p > 0.05). The DC 0 value of 68% was shown in the MC 5 group. However, the control group, MC 0 , showed a value of 65% at DC 0 , which was statistically different compared to MC 2.5 and MC 5 (p < 0.05). The mean degree of conversion 24 h after polymerisation (DC 24 ) was between 73 and 76%, with the highest score being for the control group, MC 0 . The DC 24 values showed no statistical differences between any of the groups (p > 0.05).
Intragroup comparisons were also performed by independent sample t-tests between DC 0 and DC 24 . There were statistically significant differences in all composite groups between the DC 0 and DC 24 measurements (p < 0.05) (Figure 2).

Hardness Measurements
Vickers hardness (VHN) measurements for the top and bottom surfaces were recorded at 24 h after photo curing for the SHDC groups. The mean value for the top surface hardness was in the control group MC 0 at 26 (±2) VHN, while lower values were recorded for the MC 7.5 and MC 10 groups-around 22 (±2) VHN. There were statistical differences between MC 10 and both MC 0 and MC 5 (p < 0.05); however, no significant differences were reported for the rest of the groups (p > 0.05) (Figure 3). None of the DC values showed any statistical differences between the groups with different microcapsule percentages (p > 0.05). Some differences could be noticed in DC0 between the control group, MC0, and both MC2.5 and MC5 groups (p < 0.05). Differences in values between DC0 and DC24 were shown in MC0, MC7.5 and MC10 (p < 0.05). Diagram is reused from [37].

Hardness Measurements
Vickers hardness (VHN) measurements for the top and bottom surfaces were recorded at 24 h after photo curing for the SHDC groups. The mean value for the top surface hardness was in the control group MC0 at 26 (±2) VHN, while lower values were recorded for the MC7.5 and MC10 groups-around 22 (±2) VHN. There were statistical differences between MC10 and both MC0 and MC5 (p < 0.05); however, no significant differences were reported for the rest of the groups (p > 0.05) (Figure 3).  . None of the DC values showed any statistical differences between the groups with different microcapsule percentages (p > 0.05). Some differences could be noticed in DC 0 between the control group, MC 0 , and both MC 2.5 and MC 5 groups (p < 0.05). Differences in values between DC 0 and DC 24 were shown in MC 0 , MC 7.5 and MC 10 (p < 0.05). Diagram is reused from [37]. None of the DC values showed any statistical differences between the groups with different microcapsule percentages (p > 0.05). Some differences could be noticed in DC0 between the control group, MC0, and both MC2.5 and MC5 groups (p < 0.05). Differences in values between DC0 and DC24 were shown in MC0, MC7.5 and MC10 (p < 0.05). Diagram is reused from [37].

Hardness Measurements
Vickers hardness (VHN) measurements for the top and bottom surfaces were recorded at 24 h after photo curing for the SHDC groups. The mean value for the top surface hardness was in the control group MC0 at 26 (±2) VHN, while lower values were recorded for the MC7.5 and MC10 groups-around 22 (±2) VHN. There were statistical differences between MC10 and both MC0 and MC5 (p < 0.05); however, no significant differences were reported for the rest of the groups (p > 0.05) (Figure 3).  Intragroup comparisons between the top and bottom surface hardness values indicated statistical differences in the MC 5 and MC 7.5 groups (p < 0.05); however, the rest of the groups showed no statistically significant differences between the top and bottom surface values (p > 0.05) (Figure 3).

Flexural Strength Measurements
Flexural strength and elastic modulus measurements of the SHDC groups showed values of 80 (±12) MPa and 1.95 GPa in the control group, MC 0 . Reduced values of 55 (±4) MPa and 1.81 GPa were found in the MC 10 group. There were statistically significant differences between the control group, MC 0 , and the MC 5 , MC 7.5 and MC 10 groups; similar differences were also noticed in MC 2.5 compared to MC 7.5 and MC 10 (p < 0.05). The elastic modulus measurements of the SHDC groups showed no statistically significant differences between the groups (p > 0.05) (Figure 4).
lower hardness values of 21 (±1) VHN were recorded on the bottom surfaces for composites MC7.5 and MC10. Statistical differences in the hardness values of the bottom surfaces were found in the MC0 and MC2.5 groups when compared to the MC7.5 and MC10 groups (p < 0.05); no difference was recorded for the rest of the groups (p > 0.05).
Intragroup comparisons between the top and bottom surface hardness values indicated statistical differences in the MC5 and MC7.5 groups (p < 0.05); however, the rest of the groups showed no statistically significant differences between the top and bottom surface values (p > 0.05) (Figure 3).

Flexural Strength Measurements
Flexural strength and elastic modulus measurements of the SHDC groups showed values of 80 (±12) MPa and 1.95 GPa in the control group, MC0. Reduced values of 55 (±4) MPa and 1.81 GPa were found in the MC10 group. There were statistically significant differences between the control group, MC0, and the MC5, MC7.5 and MC10 groups; similar differences were also noticed in MC2.5 compared to MC7.5 and MC10 (p < 0.05). The elastic modulus measurements of the SHDC groups showed no statistically significant differences between the groups (p > 0.05) (Figure 4). Representative fractured surfaces of self-healing composites are illustrated in Figure  5; the examined specimens contained 5 wt% microcapsules. The fractured surfaces showed typical step-nail patterns in the same direction as the crack propagation ( Figure  5A). Polymer film formation at the microcracks in the composite can be seen in Figure 5B The cracked microcapsules released TEGDMA-DHEPT healing agent then polymerised Size variation of microcapsules can be shown in Supplementary Materials Figure S1. Representative fractured surfaces of self-healing composites are illustrated in Figure 5; the examined specimens contained 5 wt% microcapsules. The fractured surfaces showed typical step-nail patterns in the same direction as the crack propagation ( Figure 5A). Polymer film formation at the microcracks in the composite can be seen in Figure 5B

Discussion
In our previous work, TEGDMA-DHEPT microcapsules were successfully synthesised via emulsion polymerisation to achieve poly(urea-formaldehyde) capsular shells [38,39]. Microcapsules were prepared by in-situ polymerization where EMA acted as a surfactant, which helped to form an oil-in-water (O/W) emulsion (the oil being TEGDMA-DHEPT liquid). Our findings from the SEM imaging showed that PUF nanoparticles existed on the external surface of the capsular shells. PUF nanoparticles in the outer surface

Discussion
In our previous work, TEGDMA-DHEPT microcapsules were successfully synthesised via emulsion polymerisation to achieve poly(urea-formaldehyde) capsular shells [38,39]. Microcapsules were prepared by in-situ polymerization where EMA acted as a surfactant, which helped to form an oil-in-water (O/W) emulsion (the oil being TEGDMA-DHEPT liquid). Our findings from the SEM imaging showed that PUF nanoparticles existed on the external surface of the capsular shells. PUF nanoparticles in the outer surface of the microcapsules can facilitate the mechanical interlocking interface (micromechanical retention) to the host resin polymer matrix upon photo activation (light curing) [26]. This retention interface allows the microcapsules to break when subjected to cracking. However, if the outer surface of the microcapsules had a smooth, non-porous morphology, the interlocking interface would be missing and the crack could bypass the microcapsules without breaking them, resulting in no healing liquid to fill the crack [26].
Regarding the biocompatibility of the self-healing system, which includes TEGDMA monomer and DHEPT amine in the microcapsules and BPO catalyst within the resin composite matrix, the materials were all approved by the Food and Drug Administration, and are available in commercial resin-based dental composites. In addition, human gingival fibroblast cytotoxicity tests in vitro have proven that TEGDMA-DHEPT microcapsules exhibit a good biocompatibility, hence the incorporation of microcapsules in resin did not drastically compromise the cell viability [26]. However, the possibility of leakage of unreacted free formaldehyde from PUF shells should be investigated.
Self-healing dental composites (SHDCs) were formulated by incorporation of Bis-GMA: TEGDMA (1:1), 20 wt% SiO 2 nanoparticles (15 nm) and different percentages of TEGDMA-DHEPT microcapsules (0, 2.5, 5, 7.5, 10 wt%) with an average diameter of 120 ± 45 µm (n: 100). The photo-activated resin composite had 1 wt% of type I photo initiator phenyl bis(2,4,6-trimethylbenzoyl) phosphine oxide (BAPO), which generated free radicals by fragmentation of the photo-initiator molecule. This is unlike type II photoinitiator systems, which need a co-initiator to donate electrons and protons in their excited state to generate the radicals for polymerisation [40]. A 0.5 wt% BPO initiator was also added to promote chemical or auto-polymerisation by free radical formation, which in turn targeted the carbon double bond (C=C) leading to polymerisation [34]. The most commonly used activator is tertiary aromatic amine N,N-bis(2-hydroxyethyl)-p-toluidine, which reacts with the initiator BPO [34].
During photo polymerisation, light transmitted through a resin composite is absorbed and scattered; this causes the light intensity to be attenuated and its efficiency drops as the depth increases [41,42]. Many factors affect the depth of cure, such as light irradiance [43], exposure time [44,45], materials composition [46], resin composite shades [47] and translucency [48,49]. Light scattering is the most important limiting factor for the depth of cure and it is maximised when the filler particle size is near to half of the wavelength emitted by the light source [50]. Studies have shown that other factors can also influence the light transmittance of resin-based composites, such as fillers and the polymeric matrix refractive index, filler type and content and monomer type [51,52]. The light transmittance through the SHDC in the present study showed no statistically significant differences between the groups (p > 0.05). The differences in the maximum light transmittance measurements between the experimental groups were mainly due to microcapsules' particle sizes of ≥150 µm.
The preferred mode of cure in resin-based dental composites is photo polymerisation. Ideally, the photo-polymerisation of the restorative dental composite will convert all monomers to polymers [53]; yet, most of dimethacrylate monomers showed a degree of conversion (DC) ranging from 55 to 85% under conventional curing conditions [54][55][56][57] and a further reduction in DC rate could be associated with depth [53]. The mechanical properties of composites are directly influenced by the degree of conversion [58].
The degree of conversion of the SHDC in this study ranged from 65 to 68% immediately after polymerisation (DC 0 ). After 24 h at 37 • C in distilled water (DC 24 ), scores higher than DC 0 were noted, ranging from 73 to 76%. The degree of conversion increased after 24 h in conditions matching the intraoral temperatures and hydrophilic environment of the oral cavity. A recent study reported that a higher degree of conversion could be seen after 24 h [59]. The type and concentration of photo initiator can influence the degree of conversion in the resin matrix. The high reactivity of the BAPO photo initiator seems to improve the mechanism of free radical formation. As absorption of light energy (BAPO*) takes place, the molecule undergoes α-cleavage in the excited triplet state of the C-P bond; this may occur twice, which results in four free radicals being generated per molecule [60]. In other words, the first free radical is less reactive than the second radical by 2 to 6 times [61].
BAPO photo initiators, according to UV-VIS spectra, have a λ max value ranging from 365 to 440 nm. The λ max of BAPO was greater than 400 nm [62]. BAPO has proven advantages, such as a white colour that helps light penetrate and increases the polymerisation depth in the resin [63,64]. Although the most commonly used photo-initiator system in dental resin composites is type II (e.g., camphorquinone, CQ), this requires an amine co-initiator (e.g., DMAEMA) to generate free radical polymerisation [65]. This may compromise the aesthetics of the dental restoration as the CQ molecule is yellow and the amine undergoes yellowing over time [65][66][67]. The model self-healing composite requires the addition of a BPO initiator in the resin matrix, which requires the use of an amine co-initiator with a CQ photo initiator to prevent resin matrix polymerisation with BPO upon mixing [68].
Surface hardness of self-healing composites has not been widely explored in the literature. The presented study reported Vickers hardness (VHN) for the SHDC, however, the VHN scores were considered low (22)(23)(24)(25)(26) when compared to commercial dental composites. The surface hardness of dental composites is affected by the volume fraction of fillers rather than the filler particle hardness [69]. Composite hardness is also affected by hydrolytic degradation and water sorption [70,71]. Material hardness is strongly associated with compressive strength, degree of conversion and wear resistance [72,73]; low hardness values may indicate a poor chemical/physical bonding in the matrix/filler interface [73].
The flexural strength and elastic modulus of SHDCs can provide insight into the effect of microcapsules in dental composites on the mechanical properties of the material. According to ISO standard 4049:2009 of polymer based restorative materials, flexural strength values shall be equal to or greater than 80-100 MPa. However, the flexural strength values of the experimental SHDCs ranged from 55-80 MPa (MC 10 to MC 0 groups, respectively). Flexural strength values significantly decreased with the addition of 7.5 to 10 wt% microcapsules into the composite. The elastic moduli of the SHDCs did not show significant differences between the groups (1.95-1.81 GPa from MC 0 to MC 10 ). These findings were in agreement with another study involving TEGDMA-DHEPT microcapsules in dental composites, which concluded that at 15 wt% of microcapsules in the composite the flexural strength dropped drastically, and at 20 wt% of microcapsules the value reached 30 MPa, indicating deterioration of the mechanical properties [26]. The inclusion of microcapsules into resin composites may have an adverse effect in the mechanical properties of the composite. This is due to the size and strength of microcapsules' shells, as they hold the healing liquid. Thus, this property of the microcapsule is reduced and deteriorates the mechanical properties of the composites.
The mechanical properties of fibre-reinforced composites (FRCs) differ considerably from conventional and bulk particulate-filled composites (PFC), as they display superior fracture toughness [74,75]. One approach aimed at improving the mechanical properties of SHDCs is to use millimetre-scale, discontinuous fibre-reinforced composites (millimetrescale, discontinuous-FRCs). Fibre toughening mechanisms rely on their ability to deflect crack propagation, to stretch, bridge and resist the opening and further spreading of the crack. These properties induce a closure force onto the crack itself [76]. This action subsequently lowers the stress concentration at the tip of the crack, thus slowing or preventing further progression. Further study is warranted to investigate the suitability of using fibres in SHDCs.
SEM of the fractured surfaces of the self-healing composite specimen indicated possible polymer film formation into the microcracks at the fracture plane. This indicated that in the presence of a BPO initiator in resin composites, and upon microcapsule rupture, the healing agent is released. The amine accelerator, DHEPT, in the healing agent had the ability to polymerise once it contacted the BPO, then monomer polymerisation sealed the microcracks. The polymerisation capability of the healing agent (TEGDMA-DHEPT) and BPO initiator was also consistent with FT-IR spectra results; a degree of conversion of 60.3% was showed in our previous report.
The incorporation of microcapsules in resin composites can change the properties of the host polymeric material; the goal was to obtain a self-healing capability without diminishing the original mechanical properties of the material. The durability and reliability of the restorative material could be compromised by weaker mechanical properties, including flexural strength and fracture toughness [77][78][79]. Previous reports on the incorporation of microcapsules in resin-based materials showed significant improvements in tensile strength [80] and flexural strength in the polymer matrix [81]. Yet, other reports showed a continuous reduction in the strength of the resin polymer with high percentages of microcapsule involvement [82].

Conclusions
TEGDMA-DHEPT microcapsules were synthesised by in-situ polymerisation of an O/W emulsion. The microcapsules had the ability to polymerise when they were ruptured and triggered by a BPO catalyst in the host composite. The self-healing dental composite (SHDC) incorporated Bis-GMA:TEGDMA (1:1), 1 wt% BAPO photo-initiator, 0.5 wt% BPO catalyst, 20 wt% silica dioxide nanoparticles and different percentages of microcapsules. The composite properties were not drastically affected in terms of light transmittance, degrees of conversion, hardness and elastic moduli. Microcapsule inclusion did not improve the mechanical properties; rather, a reduction of~65% in flexural strength measurements showed at 10 wt% of microcapsules in the composite. A way to improve the properties of SHDCs is to increase the filler content percentage with different filler size distributions in order to have a hybrid composite incorporating nano-, micro-and macro-sized particles. The current study suggested that improvements in mechanical properties of SHDCs are required in order to meet the clinical requirements of a restorative material.