Effect of Novel Bioactive Glass-Containing Dentin Adhesive on the Permeability of Demineralized Dentin

This study aimed to evaluate the effect of a novel bioactive glass (BAG)-containing dentin adhesive on the permeability of demineralized dentin. Bioactive glass (85% SiO2, 15% CaO) was fabricated using the sol-gel process, and two experimental dentin adhesives were prepared with 3 wt% silica (silica-containing dentin adhesive; SCA) or BAG (BAG-containing dentin adhesive; BCA). Micro-tensile bond strength (μTBS) test, fracture mode analysis, field-emission scanning electron microscopy (FE-SEM) analysis of adhesive and demineralized dentin, real-time dentinal fluid flow (DFF) rate measurement, and Raman confocal microscopy were performed to compare SCA and BCA. There was no difference in μTBS between the SCA and BCA (p > 0.05). Multiple precipitates were evident on the surface of the BCA, and partial occlusion of dentinal tubules was observed in FE-SEM of BCA-approximated dentin. The DFF rate was reduced by 50.10% after BCA approximation and increased by 6.54% after SCA approximation. Raman confocal spectroscopy revealed an increased intensity of the hydroxyapatite (HA) peak on the dentin surface after BCA application. The novel BAG-containing dentin adhesive showed the potential of both reducing dentin permeability and dentin remineralization.


Introduction
The demineralization of dentin is usually caused by organic acids from oral bacteria or acid etching during an adhesive procedure, which results in the destruction of the dentin BAG 85S (85 mol% SiO 2 , 15 mol% CaO) was prepared without amination as described by Lee et al. [32]. All the reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA). Briefly, it was prepared by sol-gel synthesis using a mixture of a precursor (calcium nitrate tetrahydrate) with deionized water (DW), co-solvents (ethanol and 2-ethoxyethanol), surfactant (hexadecyltrimethylammonium bromide; CTAB), and a catalyst (aqueous ammonia) at room temperature. After the mixture was stirred for 30 min, tetraethyl orthosilicate was added at a molar ratio of 85:15 (Si:Ca). Next, the solution was stirred for 4 h to form a gel-state solution. Then, a precipitate obtained from the gel-state solution was filtered, washed, and dried in air for 24 h. Next, the solution was heated to remove CTAB. After calcination at 600 • C in air for 5 h, the precipitate was washed with ethanol and DW, and the nanoparticles were dried under vacuum to obtain BAG. Table 1 shows the formulations of the prepared adhesives used in this study. All the reagents were also purchased from Sigma-Aldrich. The monomers were diurethane dimethacrylate (UDMA) and 2-hydroxyethyl methacrylate (HEMA). Camphorquinone and ethyl-4-(dimethylamino) benzoate (EDMAB) were added as the photoinitiator systems. As an antioxidant, 2,6-di-tert-butyl-4-methylphenol (BHT) was added. For the control dentin adhesive, 3 wt% of inert non-silanated silica was added (silica-containing adhesive, SCA). For the experimental group, BAG 85S was added at a concentration of 3 wt% to form the experimental dentin adhesive (BAG-containing adhesive, BCA).

Micro-Tensile Bond Strength (µTBS) Test
Ten teeth were assigned to undergo the µTBS test and were randomly divided into two groups according to the type of adhesive: SCA and BCA. The fresh, superficial dentin surface was exposed to a high-speed diamond saw (IsoMet 5000; Buehler, Lake Bluff, IL, USA). Then, each dentin surface was etched with 37% phosphoric acid gel (Etch-37; Bisco, Schaumburg, IL, USA) for 15 s, after which it was rinsed and blot dried. Each adhesive was applied to dentin for 15 s and gently air-dried. Then, the adhesive was light-cured for 20 s using an LED curing unit (Bluephase G2; Ivoclar Vivadent, Schaan, Liechtenstein) emitting 1200 mW/cm 2 . A resin composite (Any-Com; MEDICLUS, Cheongju, Korea) was incrementally built up to a thickness of 2 mm to a height of 4 mm. Each increment was light-cured for 20 s, with light placed directly over the composite. The specimens were stored in distilled water at 37 • C for 24 h.
After storage, each composite-tooth specimen was sectioned occluso-gingivally into 1.0 mm thick serial slabs using a high-speed diamond saw (IsoMet 5000). These slabs were further sectioned into composite-dentin beams (1.0 × 1.0 mm 2 ). Forty composite dentin beams were produced in each group. These beams were divided into two subgroups and stored in distilled water for 24 h (n = 20). No treatment was performed on the first 20 beams in the immediate group. The other 20 beams were accelerated-aged in 10 % sodium hypochlorite (NaOCl) for 1 h before testing [33]. The prepared beams were mounted on a testing jig with a cyanoacrylate adhesive (Zapit; Dental Ventures of America, Corona, CA, USA). The µTBS test was performed with a universal testing machine (AGS-X STD; Shimadzu, Kyoto, Japan) at a crosshead speed of 1.0 mm/min.

FE-SEM of the Adhesive Surface
FE-SEM analysis was performed to evaluate the cured-surface changes of the SCA and BCA. A total of 10 resin composite blocks (Any-Com) were prepared with a silicone mold (6.0 × 6.0 × 4.0 mm 3 ). The blocks were randomly divided into two groups (n = 5). A 37% phosphoric acid gel (Etch-37) was applied for 20 s on the surface of the block and rinsed with water. Then, either SCA or BCA was applied and light-cured for 20 s.
The specimens were immersed in simulated body fluid (SBF) for 2 weeks, and the solution was changed every 2 days to prevent autogenous precipitation. SBF was prepared according to the method of Tas et al. [34]. The composition of SBF solution with 27 mM HCO 3 − is presented in Table 2. After storage, the surface of the adhesive on the resin composite block was thoroughly rinsed with distilled water for 3 min. The blocks were then treated as described by Perdigao et al. [35]. They were subsequently examined using FE-SEM (S-4700; Hitachi, Tokyo, Japan) at 10 kV.

FE-SEM of Dentin Surface
Ten teeth were assigned and randomly divided into two groups (SCA and BCA, n = 5). To simulate demineralized dentin, each dentin surface was etched with 37% phosphoric acid gel (Etch-37) for 60 s. Ten composite blocks (Any-Com) were fabricated, and each dentin adhesive was applied and light-cured, as previously described in Section 2.4. The adhesiveapplied surface of the composite block was approximated to the demineralized dentin surface and then fixed with a flowable resin (G-Fix; GC, Tokyo, Japan). The composite-tooth specimens of the two groups were then stored in SBF solution at 37 • C for 2 weeks. The SBF was changed every 2 days to prevent autogenous precipitation. After storage, the composite block was removed from the specimen, and the dentin surface was thoroughly rinsed with distilled water for 3 min. The FE-SEM analysis was performed on the dentin surface, as described in Section 2.4.

Dentinal Fluid Flow (DFF) Rate Measurement
Five teeth were used for real-time DFF rate analysis for each experimental group (n = 5). A schematic diagram of the DFF-rate analysis device is shown in Figure 1. The root portion of each tooth was removed 5 mm below the cemento-enamel junction using a high-speed diamond saw (Isomet 5000). The pulp tissue was completely removed using endodontic files and tissue forceps. Then, the tooth specimen was mounted on an acrylic plate with a hole drilled in its center. A metal tube that was 0.9 mm in diameter was inserted into the hole, ending within the pulp chamber. The acrylic with the tooth specimen was fixed, and the remaining exposed root surface was completely sealed with dentin adhesive (All-Bond Universal; Bisco) and flowable resin composite (Gaenial-Flow Universal; GC). This was then stored for 24 h in distilled water. Subsequently, each specimen was connected to a water reservoir under a hydrostatic pressure of 30 cm H 2 O to simulate the physiological pressure of the pulp tissue [36]. Each specimen underwent a stabilization procedure for 10 min after being connected to a subnanoliter-scaled fluid flow measuring device (NanoFlow; IB Systems, Seoul, Korea) [37].

Dentinal Fluid Flow (DFF) Rate Measurement
Five teeth were used for real-time DFF rate analysis for each experimental group (n = 5). A schematic diagram of the DFF-rate analysis device is shown in Figure 1. The root portion of each tooth was removed 5 mm below the cemento-enamel junction using a high-speed diamond saw (Isomet 5000). The pulp tissue was completely removed using endodontic files and tissue forceps. Then, the tooth specimen was mounted on an acrylic plate with a hole drilled in its center. A metal tube that was 0.9 mm in diameter was inserted into the hole, ending within the pulp chamber. The acrylic with the tooth specimen was fixed, and the remaining exposed root surface was completely sealed with dentin adhesive (All-Bond Universal; Bisco) and flowable resin composite (Gaenial-Flow Universal; GC). This was then stored for 24 h in distilled water. Subsequently, each specimen was connected to a water reservoir under a hydrostatic pressure of 30 cm H2O to simulate the physiological pressure of the pulp tissue [36]. Each specimen underwent a stabilization procedure for 10 min after being connected to a subnanoliter-scaled fluid flow measuring device (NanoFlow; IB Systems, Seoul, Korea) [37]. The prepared specimen was connected to a glass capillary filled with water containing an air bubble. Abbreviations: SCA, silica-containing adhesive; BCA, BAG-containing adhesive; CEJ, cementoenamel junction; LED, light-emitting diode A real-time DFF rate analysis was performed continuously for 5 min before acid etching to measure the baseline flow rate ( ). Acid etching was then conducted for 60 s. Then, the specimen was rinsed with distilled water and blot-dried. The same analysis was performed for 5 min to measure the pre-application flow rate through dentin simulating demineralization ( ). After measuring the , each tooth was approximated to the composite block where one of the adhesives (SCA or BCA) was cured on its surface and fixed as described in Section 2.5. They were then stored in the SBF solution at 37 °C for 2 weeks, with the solution being changed every 2 days to prevent autogenous precipitation. After storage, the composite block was removed from the specimen and thoroughly rinsed with distilled water for 3 min, and DFF rate analysis was then performed for 5 min to measure the post-application flow rate ( ). The real-time changes in the DFF rate were calculated by: A real-time DFF rate analysis was performed continuously for 5 min before acid etching to measure the baseline flow rate (DFF Base ). Acid etching was then conducted for 60 s. Then, the specimen was rinsed with distilled water and blot-dried. The same analysis was performed for 5 min to measure the pre-application flow rate through dentin simulating demineralization (DFF Pre ). After measuring the DFF Pre , each tooth was approximated to the composite block where one of the adhesives (SCA or BCA) was cured on its surface and fixed as described in Section 2.5. They were then stored in the SBF solution at 37 • C for 2 weeks, with the solution being changed every 2 days to prevent autogenous precipitation. After storage, the composite block was removed from the specimen and thoroughly rinsed with distilled water for 3 min, and DFF rate analysis was then performed for 5 min to measure the post-application flow rate (DFF Post ). The real-time changes in the DFF rate were calculated by:

Confocal Raman Spectroscopy
Six teeth were selected and divided into two groups (SCA and BCA). They were prepared as described in Section 2.5. To identify the chemical changes, we used a confocal Raman spectroscopy system (UniDRON, Yongin, Korea) with a 785-nm diode laser with 100 mW power and a 10× objective lens with 0.25 NA. With this system, Raman spectra were obtained three time points (before demineralization, after demineralization, and after SCA or BCA application) at 25 random locations (n = 75 per each time point) within a spectral range of 400-1100 cm −1 , spectral resolution of 2 cm −1 , and an acquisition time of 10 s.

Statistical Analysis
To identify differences in µTBS according to dentin adhesive and accelerated aging, two-way analysis of variance (ANOVA) was performed. To analyze DFF rate according to measurement time and ∆DFF of dentin adhesives, one-way ANOVA was conducted. The Bonferroni test was used for post-hoc analysis. The p-value was estimated using an R-squared approximation for degrees of freedom (α = 0.05).

Micro-Tensile Bond Strength (µTBS) Test
The µTBS results are presented in Table 3. There was no difference in the µTBS of the SCA and BCA groups in either immediate or accelerated aging (p > 0.05). However, the µTBS of each adhesive significantly decreased after accelerated aging (p < 0.05).

FE-SEM of the Adhesive Surface
Representative FE-SEM images of the SCA and BCA groups are shown in Figure 2. There were no precipitates on the surface of the SCA group (Figure 2A,B), whereas a bunch of precipitates were observed on the surface of the BCA group ( Figure 2C). In the magnified image, small particles with a diameter of 200 nm were observed ( Figure 2D).

FE-SEM of the Dentin Surface
Representative FE-SEM images are shown in Figure 3. In the SCA group, a demineralized dentin surface with open dentinal tubules was apparent ( Figure 3A). The demineralized collagen network was clearly observed around the dentinal tubules, especially in the magnified image ( Figure 3B). Some precipitates appeared around the dentinal tubules. However, the dentin surface was predominantly covered with small precipitates in the BCA group ( Figure 3C), and many precipitates were observed in the dentinal tubules, which partially occluded dentinal tubules ( Figure 3D).

FE-SEM of the Dentin Surface
Representative FE-SEM images are shown in Figure 3. In the SCA group, a demineralized dentin surface with open dentinal tubules was apparent ( Figure 3A). The deminer-Materials 2021, 14, 5423 7 of 13 alized collagen network was clearly observed around the dentinal tubules, especially in the magnified image ( Figure 3B). Some precipitates appeared around the dentinal tubules. However, the dentin surface was predominantly covered with small precipitates in the BCA group ( Figure 3C), and many precipitates were observed in the dentinal tubules, which partially occluded dentinal tubules ( Figure 3D). tion view (×10,000) of SCA-approximated dentin surface. The exposed collagen network can be seen on the demineralized dentin surface. Some precipitates covering the dentin surface can be seen. (C).
Low magnification view (×5000). (D). High magnification view (×10,000) of BCA-approximated dentin surface. The demineralized dentin surface was predominantly covered with small precipitates, and precipitates were observed inside the dentinal tubules, which obstructed dentinal tubule orifices.

Dentinal Fluid Flow (DFF) Rate Measurement
The DFF values of both experimental groups are shown in Table 4, and the graph of real-time DFF rate analysis is shown in Figure 4.
values for the SCA and BCA groups were 1.38 and 1.79 nL/s, respectively. After demineralization, the values of the groups increased to 3.68 and 3.71 nL/s, respectively. Although both adhesives showed similar and values, their values were different. The DFF rate slightly increased by 6.54% in the SCA group but dramatically decreased by 50.10% in the BCA group. As shown in Figure 4, the DFF rates of both groups temporarily decreased at blot dry and increased gradually after etching. The DFF changes in both groups were similar before the indirect approximation of the two adhesives. While the of the SCA group did not change after adhesive approximation, that of the BCA group decreased considerably.  . High magnification view (×10,000) of SCA-approximated dentin surface. The exposed collagen network can be seen on the demineralized dentin surface. Some precipitates covering the dentin surface can be seen. (C). Low magnification view (×5000). (D). High magnification view (×10,000) of BCA-approximated dentin surface. The demineralized dentin surface was predominantly covered with small precipitates, and precipitates were observed inside the dentinal tubules, which obstructed dentinal tubule orifices.

Dentinal Fluid Flow (DFF) Rate Measurement
The DFF values of both experimental groups are shown in Table 4, and the graph of real-time DFF rate analysis is shown in Figure 4. DFF Base values for the SCA and BCA groups were 1.38 and 1.79 nL/s, respectively. After demineralization, the DFF Pre values of the groups increased to 3.68 and 3.71 nL/s, respectively. Although both adhesives showed similar DFF Base and DFF Pre values, their DFF Post values were different. The DFF rate slightly increased by 6.54% in the SCA group but dramatically decreased by 50.10% in the BCA group. As shown in Figure 4, the DFF rates of both groups temporarily decreased at blot dry and increased gradually after etching. The DFF changes in both groups were similar before the indirect approximation of the two adhesives. While the DFF Post of the SCA group did not change after adhesive approximation, that of the BCA group decreased considerably.   Figure 5 shows the representative changes in the Raman spectra of the SCA and BCA groups (n = 6). There were four different phosphate ion peaks (433, 579, 959, and 1043 cm −1 ) in both groups before demineralization ( Figure 5A,C). The most intense peak was at 959 cm −1 , which decreased dramatically after demineralization in both groups ( Figure 5B,E). In addition, the intensities of another three peaks slightly decreased. After the approximation of both adhesives, these peaks slightly increased in the SCA group ( Figure 5C). On the other hand, the intensity of peak at 959 cm −1 increased considerably in the BCA group ( Figure 5F).   Figure 5 shows the representative changes in the Raman spectra of the SCA and BCA groups (n = 6). There were four different phosphate ion peaks (433, 579, 959, and 1043 cm −1 ) in both groups before demineralization ( Figure 5A,C). The most intense peak was at 959 cm −1 , which decreased dramatically after demineralization in both groups ( Figure 5B,E). In addition, the intensities of another three peaks slightly decreased. After the approximation of both adhesives, these peaks slightly increased in the SCA group ( Figure 5C). On the other hand, the intensity of peak at 959 cm −1 increased considerably in the BCA group ( Figure 5F). groups (n = 6). There were four different phosphate ion peaks (433, 579, 959, and 1043 cm −1 ) in both groups before demineralization ( Figure 5A, C). The most intense peak was at 959 cm −1 , which decreased dramatically after demineralization in both groups ( Figure 5B, E). In addition, the intensities of another three peaks slightly decreased. After the approximation of both adhesives, these peaks slightly increased in the SCA group ( Figure 5C). On the other hand, the intensity of peak at 959 cm −1 increased considerably in the BCA group ( Figure 5F).

Discussion
Remineralization of dentin is an important issue because it can enhance tooth structure and further prevent collagen degradation. Although various methods have been suggested to induce tooth remineralization, most have only been possible in in vitro situations. For dentin remineralization agents to be clinically useful, the underlying demineralized dentin must be directly contacted. Thus, dentin adhesive is an ideal material for the incorporation of a remineralization agent. This study evaluated BAG-containing dentin adhesive using various methods and proved that it could partially remineralize adjacent dentin, thus reducing the permeability of demineralized dentin while maintaining bond strength.
BAG 85S was prepared using a sol-gel process. It has smaller particle size (160 nm) and larger surface area (423 m 2 /g) than conventional, melt-quenching derived BAG. Vollenweider et al. [23] compared the remineralization ability of sol-gel and melt-quenching derived BAG and showed that nano-sized BAG was far more efficacious than micrometer-sized BAG for remineralization. The surface area of BAG is important for HCA layer formation because ion release only occurs on its surface. Thus, the nanostructured BAG synthesized in the present study has sufficient potential for dentin remineralization. The formation of the HCA layer on the BAG surface plays an important role in the remineralization process. HCA is considered a precursor of hydroxyapatite (HA), which is thought to interact with the exposed collagen fibers to promote tissue mineralization [38]. When BAG is exposed to body fluids, there is an ion exchange from the fluids, forming silanol bonds on the surface of BAG and eventually a silica (SiO 2 )-rich layer. When exposed to increased pH, the Si-O-Si bonds of the silica-rich layer are disrupted, and the layer absorbs calcium ions (Ca 2+ ) from the glass. This allows the glass surface to react with phosphate ions [(PO 4 ) 3− ] to form an amorphous HCA layer. Over time, the amorphous HCA layer crystallizes into the HA layer.
Some dentin adhesives contain silanized inorganic fillers. The silanization of inorganic fillers plays an important role in chemical bonding to the organic matrix, thus reinforcing the mechanical properties of the adhesive [39]. However, it was thought that incorporation of BAG 85S might adversely affect dentin bond strength because it was not silanized. Although silanization of inorganic filler is necessary to chemically combine the organic matrix and inorganic filler, silanization of BAG may reduce its ion release ability by increasing its hydrophobicity. According to the results of the µTBS test, those of the SCA and BCA groups were not significantly different in both the immediate and acceleratedaged modes. These results indicate that the incorporation of BAG into the adhesive did not adversely affect the adhesion performance, which is consistent with the results of a previous study [27]. A common point of two studies was that the content of BAG was low (3% in this study, and 0.5-2% in Jun et al. [27], respectively).
For accelerated aging of the specimens, NaOCl storage was used in this study. Using this method, unprotected, exposed collagen fibrils after acid etching are deproteinized, which reduces bond strength [33]. It was selected to confirm whether the incorporated BAG hindered adhesive infiltration due to its particle size (160 nm). The interfibrillar spaces of demineralized dentin after etching have been reported to be within 20 nm [40] thus, there was a potential concern with adhesive infiltration. However, the µTBS values of both groups were not significantly different from each other after accelerated aging. This means that the BAG 85S used in this study did not hamper adhesive infiltration.
To evaluate the surface changes of SCA and BCA, the cured surfaces of both dentin adhesives were analyzed using FE-SEM. Although there were no precipitates on the surface of SCA, cluster-like precipitates were obvious on the dentin treated with BCA ( Figure 2). It was assumed that the precipitates were part of the HCA layer formed on the surface of the BAG. This formation of precipitates was also reported in our previous study [31].
To identify the effect of precipitates on demineralized dentin, the dentin surface was also analyzed. The reason for observing the top surface of dentin without direct application of the adhesive was that it was technically impossible to physically remove only the thin adhesive and hybrid layers without damaging the underlying demineralized dentin. This experimental method was also utilized in other studies that evaluated the effect of a BAG-containing glass ionomer cement and composite resin [31,41]. There were more precipitates on the BCA-approximated dentin surface than on the SCA-approximated one. Theoretically, it was expected that there would not be any precipitates on the dentin surface of the SCA group. However, the FE-SEM analysis revealed some precipitates on the SCA-approximated dentin surface. It was attributed to the effect of storage solution. In this study, a 27 mM HCO 3 − Tris SBF was used for specimen storage, as its chemical composition is more similar to human blood plasma than the original SBF prepared by Kokubo and Takadama [42], and is more proper in simulating dentinal tubule fluid [34]. However, it may also form autogenous calcium phosphate precipitates in the specimen due to supersaturation [43]. Thus, we changed the SBF every 2 days to prevent the formation of undesirable precipitates. Our previous study revealed that the difference between artificial saliva and 27 mM HCO 3 − Tris SBF did not affect the dentin remineralization with every 2 days change [41]. However, this study showed different result in the FE-SEM analysis. It should be investigated further.
The indirect approximation of SCA and BCA to demineralized dentin was used in this study. It was due to the exclusion of the effect of adhesive resin, which could seal the exposed dentinal tubule. This indirect method has been used by Kim et al. [31] and Jang et al. [41] successfully. Considering the results of the FE-SEM analysis, the DFF rates of both adhesives were expected to decrease. However, the DFF rate of the SCA group increased by 6.54%, indicating that the precipitates did not contribute to the reduction of DFF rate. In the FE-SEM analysis, dentinal tubules were not occluded by precipitates ( Figure 3A,B). However, the DFF rate was reduced by 50.10% in the BCA group, suggesting that the formed precipitates were HA crystals and might contribute to tubule occlusion ( Figure 3C,D). Considering the remarkable reduction in DFF rate in the BCA group, it is assumed that formed precipitates may exist in deeper areas of the dentinal tubule.
Reducing the dentin permeability is important for dentin adhesion. Increased permeability may cause rapid degradation of the adhesive layer via continuous fluid movement, thus eventually deteriorating dentin adhesion [44]. As mentioned in the Introduction, BAG has the potential to remineralize dentin. This study not only proved its remineralization ability to dentin, but also reduced dentin permeability. The reduction in permeability was attributed to the formation of HA crystals inside the dentinal tubules. Considering these abilities of BAG, a synergistic reduction effect is expected if BCA is directly applied to demineralized dentin.
The chemical properties of the precipitates were analyzed by confocal Raman spectroscopy. Four distinct peaks (433, 579, 959, and 1043 cm −1 ), which are representatives for HA crystal were investigated. They refer to (PO 4 ) 3− ( ν2 ), (PO 4 ) 3− ( ν4 ), (PO 4 ) 3− ( ν1 ), and (PO 4 ) 3− ( ν3 ) respectively. The peak of 959 cm −1 was most sensitive to mineral change of the dentin. This result is consistent to that of Sauro et al. [45] and Marin et al. [46]. The SCA group showed a slightly increased HA crystal peaks after the approximation of the adhesive in spite of the absence of BAG incorporation. It may be attributed to autogenous calcium phosphate precipitates on the dentin surface observed in the FE-SEM images of this group. The precipitates in the BCA group showed more intense peaks than those in the SCA group. This suggests that the BCA approximated dentin surface was partly remineralized by newly formed HA crystals. This conclusion is also supported by the FE-SEM results. Because this study only investigated phosphate peaks to identify the formation of HA crystals, it was not possible to evaluate the quality of dentin remineralization. Two analysis of Raman parameters have been suggested. The measurement of mineral-to-matrix ratio between 960 and 1650 cm −1 can represent the volumetric fraction of mineral with respect to collagen [47]. Full width at half maximum (FWHM) at the peak of 959 cm −1 can also reveal the degree of crystallinity and defects in the HA crystals [48,49].
In addition, further studies for the changes of mechanical properties such as microhardness, nanoindentation, or elastic modulus measurement, should be performed to clarify the remineralization effect of BAG. These methods can provide the proof of intra-fibrillar remineralization of the dentin. Micro-hardness measurement about elastic modulus measurement of BAG-containing dentin adhesive and BAG-containing resin composite also reported the possibility of dentin remineralization [29,41].

Conclusions
Within the limitations of this study, BAG-containing dentin adhesive reduced the permeability of demineralized dentin via its remineralization potential while maintaining bond strength to dentin. This study provides reliable evidence for incorporating BAG into dentin adhesive as a multifunctional agent.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the institutional review board (KHD IRB 1701-1).
Informed Consent Statement: Patient consent was waived because comprehensive consent was obtained when the patient's teeth were extracted in our institution. This indicates that the extracted teeth might be used for future research. Therefore, individual patient consent was not required in the studies using only previously extracted teeth.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.