3.1 Structural characterization of glasses using FTIR spectroscopy
In contrast to the XRD technique, FTIR spectroscopy is able to record the functional groups of glasses in the mid-IR region in a precise manner [44]. Based on this concept, FTIR spectroscopy was carried out to examine the molecular structure of the prepared glasses, as shown in Fig. 1. Before assigning the spectra obtained from the glass samples, it is preferable to briefly mention some facts about the chemistry of phosphate bioactive glass (PBG) as follows:
As discussed in Refs. [29,45], in the absence of modifier oxides, PBG is mainly composed of PO43− tetrahedral units in chain/ring structures, having much non-bridging oxygen (NBO) atoms. Therefore, these glasses are characterized by low chemical durability along with low melting temperatures. However, in the presence of modifier oxides (MO), the phosphate network is depolymerized as a result of the replacement of some of the P−O−P links with −P−O−M− ones leading to a considerable change in the glass structure from Q3 units to tetrahedral chains in the metaphosphate (Q2 units), pyrophosphate (Q1 units) and orthophosphate (Q0 anions) bearing in mind that this conversion depends on the type and contents of modifying oxides. On the other hand, borate consists of three boron-oxygen triangles forming boroxol rings/groups (B3O6). However, this coordination is changed from triangular to tetrahedral after adding alkali oxide (CaO) or alkaline earth oxide (Na2O). The FTIR absorption bands of phosphate groups can be classified into three fundamental regions; namely region of 1400-1150 cm-1 that represents the vibrations of non-bridging PO2 groups, region of 1150-900 cm-1 that symbolizes the vibrations of P−O and PO3 groups and region of 900-700 cm-1 that expresses the vibrations of bridging P−O−P groups. However, borate groups exhibit three spectral regions at 800-600, 1500-1200, and 1200-850 cm-1, which emerge from the B−O stretching mode of the borate part, B−O stretching mode of BO3 and BO4 units, respectively [23]. Importantly, the most characteristic region for bioactive silicate glasses falls in the spectral region 1050-850 cm-1. Based upon the abovementioned regions, the present spectra can be assigned according to the literature [21,23,46-49] as follows:
- B−O stretching vibrations of the triangular (BO3, BO2O−) units can be confirmed by the weak band located at 1250 cm-1, considering that the increase in B2O3 contents leads to a considerable increase in its intensity until it reaches medium one.
- A single sharp, strong band can indicate the symmetric stretching mode of the P=O bond at 1200 cm-1. As expected, an increase in B2O3 contents causes its intensity to decrease until it reaches a medium intensity.
- Intense broadband located near 1045 cm-1 can be attributed to the superposition of symmetric stretching mode of Q2 units, symmetric stretching mode of O−P−O bonds, asymmetric stretching mode of Si−O−Si vibrations, and asymmetric stretching vibration of tetrahedral borate units.
- A weak band at 970 cm-1 can be ascribed to symmetric stretching vibrations of Si−O−Si and B−O bond of tetrahedral BO4 groups. Of note, by increasing B2O3 content in the glass composition, the intensity of this band increases.
- The asymmetric stretching mode of the P−O−P bonds accompanying the formation of Q2 chains can be represented by having two shoulders at 925 and 820 cm-1. On the other hand, a medium band at 765 cm-1 can be assigned to the symmetric stretching vibration of P−O−P and the bending/deformation modes of different borate units. Moreover, the shoulder at 925 cm-1 represents an asymmetric stretching mode of the Si−O bond.
- Broadband at 470 cm-1 can be ascribed to the overlapping of bending vibrations of the B−O−B bonds, Q1 units of phosphate groups, and Si−O bonds. It is reasonable to expect that the intensity of this band considerably increases due to the increased contents of B2O3, which also results in the conversion of some Q2 phosphate units to Q1 ones.
- Bending vibrations of P−O−P and Si−O bonds can be indicated by the presence of a medium band found at 465 cm-1.
3.2 Evaluation of bioactivity behavior (bone-bonding) of prepared glass samples
3.2.1 XRD analysis
In this work, glass samples were subjected to XRD technique before and after dipping in SBF solution for different time periods, i.e., 3, 7, and 10 days to ensure their amorphous structure as well as the formation of an apatite-like layer on their surfaces as shown in Fig. 2 (a-d). It is suggested from Fig. 2a that all samples have an amorphous structure as indicated by the presence of broadband at 2θ=20-30°. This observation is extremely important as it refers to the correct preparation procedures for glass samples [50]. On the other hand, the soaked samples in SBF solution exhibit crystalline HA layer, which is easily identified according to (JCPDS No. 19-0272), bearing in mind that the bioactivity of samples are positively affected by increasing soaking time as well as B2O3 contents where the best bioactivity is observed for S4 sample after its soaking in SBF solution for 10 days. The most likely explanation for the enhancement of the bioactivity of the examined samples with increasing incubation time is that increasing the soaking time leads to an increase of Ca2+ and PO43− ions on their surfaces [46]. It is generally considered that the existence of Na+ and Ca2+ in the prepared glasses act as modifiers for the network where they trigger the formation of non-bridging oxygen (NBO) atoms which in turn increase the degradation of glasses. In this regard, it is very meaningful to discuss the famous mechanism proposed by Hench et al. [8], which described the formation of an HA-layer on silicate-based glass surfaces, which usually occurs through specific steps. These steps begin by exchanging Na+ ions, from the soaked glass sample, with H+ ones, from SBF solution, causing breakage in the glass network and forming Si-OH groups at the glass/SBF interface. Then, these groups are polycondensed, leading to the formation of a silica gel layer on the glass sample surface. On the opposite side, OH− groups work as effective nucleation sites for the precipitation as well as crystallization of HA layer. These stages are finalized by the formation of sufficient HA nuclei and immigration of Ca2+ and PO43− groups from SBF solution and glass matrix producing amorphous calcium phosphate (Ca−P) film. Consequently, significant growth for HA layer is achieved. Since the current glass system has a low SiO2 content while being rich in P2O5 and B2O3, we should expect other mechanisms for the bioactivity of the prepared glass samples.
As discussed in Ref. [51], the affinity of Na+ and Ca2+ ions to the phosphate network is much higher than the silicate network. Therefore, the latter network suffers from polymerization, causing the formation of Si−O−Si and Si−O−P linkages. The latter linkages are responsible for reducing the bioactivity of glass. Fortunately, the present glass system possesses a lower contribution of SiO2 than that of P2O5; the bioactivity of the examined S1 sample is not greatly affected by Si−O−P groups. Increasing the content of B2O3 contributes to the improvement of bioactivity considerably. One possible reason for this positive result is that when B2O3 increases, at the expense of P2O5, it helps to reach a non-stoichiometric ratio of Ca:P. Another reason is the high solubility of B2O3, which contributes to breaking the glass network and the formation of B−OH bonds that act as a catalyst for bioactivity [52]. Noteworthy, the aforementioned stages greatly contribute to precipitation of calcium phosphate (Ca-P) layer on the glasses surfaces and subsequently turn into a crystalline layer, i.e., HA. Importantly, the formation of HA layer can be clearly understood with the help of the following Eqn. [47,48]:
10Ca2+ + 6PO43─ + 2OH─ → Ca10(PO4)6(OH)2 ………..…… (8)
It cannot be denied that HA nuclei's heterogeneous nucleation can be achieved by the super-saturation limit elevation above the critical level. Notably, the increase of released Na+ ions is responsible for such an increase in the super-saturation level of the soaking solution in connection to HA. Then, enlargement of HA nuclei occurs due to the consumption of Ca2+, PO43─ and OH─ ions from the SBF [53].
3.2.2. SEM analysis
In order to follow the formation of HA layer on the surface of S4 sample after soaking in SBF solution for 3, 7, and 10 days, SEM was used in this work, as shown in Fig. 3 (a-c), respectively. It should be noted that the selection of this sample for SEM examination is based on the XRD results obtained, which show that the best bioactivity characterized S4 sample among all tested samples as a result of containing the highest percentage of B2O3. It is possible to see from Fig. 3a the formation of a large number of HA crystals seen individually without the formation of the characteristic shape of the layer reflecting that the poor crystallization of these crystals. However, as can be seen from Fig. 3b, the prolonged soaking time of up to 7 days causes the number of HA crystals to increase with somewhat formation of the characteristic shape of the HA layer. Lastly, it is clearly evident from Fig. 3c that soaking the same sample in SBF solution for 10 days leads to the formation of a dense layer with few HA crystals. This indicates that the sample needs additional time to crystallize all the crystals formed or that another layer is starting to form. As discussed above, the soaking time improves the formation of the apatite layer on the surface of the same sample.
Undoubtedly, the perfect match between the bioactive ceramics and the mineral part of human bone is an important factor in forming a strong bond between these interfaces. Given that the formation of the HA layer is highly dependent on the surface texture of the examined specimen, glass grains are preferred for many biomedical applications, including antibiotics, gene delivery, and growth factors [54,55].
3.3 The properties of the prepared glass samples
3.3.1 Density
It cannot be denied that density is considered a fundamental property and can be calculated based on glass composition [56]. The density values for the as-prepared glass specimens were measured and illustrated in Fig. 4. It is interesting to see from this figure that the density decreases with increasing boron contents in the as-prepared samples despite the slight increase in the density value of B2O3 (2.46 g/cm3) compared to that of P2O5 (2.39 g/cm3). The reason for this result is that the successive increase in the content of B2O3, at the expense of P2O5, increases the number of NBOs as a result of the conversion of some Q2 units to Q1 ones, which was confirmed by the obtained FTIR spectra, that leads to a slight decrease in the density. This explanation is supported by the results discussed by Yamusa et al. [57], who reported that the increase in the number of NBOs in the glass structure is responsible for decreasing the density of the glass samples. It should be noted that because B2O3 is a network former, the density values are not always a good indicator of the packing density within the glass structure [58].
3.3.2 Mechanical properties
Generally, the clinical applications of bone substitute materials are greatly affected by their mechanical properties [44]. In this sense, microhardness, fracture toughness, compressive strength, longitudinal and shear velocities, Young's modulus, longitudinal modulus, bulk modulus, shear modulus, and Poisson's ratio of all glasses were measured and illustrated in Figs. 5-8. The figures show the role of increasing B2O3 contents in improving the mechanical properties of all samples. On the contrary, fracture toughness exhibits an opposite trend to other mechanical properties as it decreases by successive increases in B2O3 contents. These results can be explained on the basis that the current glass system is composed of three formers, i.e., SiO2, P2O5, and B2O3, along with glass modifiers. When B2O3 is added to P2O5, the phosphate glass structure is changed from a link structure to a framework structure due to the formation of B−O−P bonds, which in turn improves the mechanical properties of the investigated samples [59]. Notably, the obtained values of compressive strength and fracture toughness closely match those of cortical bone, i.e., 100-180 MPa and 2-12 MPa.m0.5, respectively [60]. Noteworthy, these results are in good agreement with those reported in Ref. [61], which mentioned that the increase of boron content, at the expense of phosphate ones, is responsible for giving the studied glass samples this recorded increase in their mechanical properties. Importantly, the good mechanical properties of these glasses give them a substantial advantage to be used in load-bearing sites [36,37].
3.4 Finite element model analysis
Model and boundary condition
A finite element model (FEM) was created using the ANSYS software to predict the compositions' mechanical characteristics and dynamic behavior. Four different concentrations of boric acid have varied the designs of the investigated work; hence this variation reflected on the experimental mechanical properties discussed before. The studied compositions samples were modeled according to the real data acquired from the experimental tests. The isotropic elasticity parameters such as Young’s modulus, density, and the maximum compressive stress were fed to the model for each composition. The samples were designed to match the purpose of its application; for example, the biomedical plate used to fix the bone was modeled, as shown in Fig. 9. With a Young's modulus of 20 GPa and a Poisson ratio of 0.3, the bone material is considered to be isotropic and homogeneous throughout the bone. Hence the calculations will be based on the load response on the investigated plates; we have used isotropy herein for convenience sake because our goal is to demonstrate the effect of different loads on the plates that are fixed to the bone.
Meshing techniques and size greatly influence the accuracy of the solution result for both mechanical and dynamic behavior. Many attempts were made to select the greatest possible resolution that could be attained and matched with the experimental test. According to the FEA solutions and meshing types, three meshing approaches were examined to choose the closest value to the experimental test: multi-zone, triangles, and quadrilateral dominant. The investigated model uses the Quad/Tri meshing method with 8108 nodes and 4166 elements. Boundary conditions were assigned using fixed support from one side of the investigated model. Static structural module and eigenvalue buckling were utilized to predict and model the mechanical properties of the studied compositions. At the same time, Modal analysis was conducted for the modeled sample; the first six mode shapes and their natural frequencies were computed. 10N force was applied at the end of the plate to calculate and plot the frequency response function. Furthermore, the harmonic response module was used in its superposition mode as part of the solution method, and the solution interval was set to 100 increment points to increase the resolution of the results in the frequency response function (FRF) plot.
3.4.2. Deflection effect
Fig. 10 shows the effect of the applied load on the deflection of the investigated compositions. It is evident from the figure that there is a linear relationship between increased loads and their effect on the deflection of the structure. As the B2O3 contents increased, the structure deflection decreased. This can be attributed to the increase in the composition strength. The improvement in the resistance to deflect (from 0 to 25 mol% B2O3) is achieved by 39%. While from (0 to 10 mol% B2O3) was enhanced by noticed 8.11 %. A significant improvement is achieved when increasing the B2O3 contents are more than 20 (mol.%).
3.4.3 Buckling analysis
Recent research shows that these structural glasses can sustain substantial dead loads caused by body weight. Because of their great slenderness, these glass samples are prone to buckling failure and must thus be examined under this probable limit condition. Analytical buckling analysis of glass columns may typically be conducted using a conventional formula for columns subjected to axial stresses. Fig. 11-a shows the effect of applied load on load multiplier; hence, buckling occurs when a component loses stability, regardless of material strength. This loss of stability happens within the material's elastic range. The load multiplier for all investigated samples was greater than 1; hence the loads applied are fewer than the critical load that can cause buckling. The exponential relationship between the applied load and the load multiplier reflects that increasing loads are getting too close to the critical load, causing the plates to collapse. Using a 1kN load, the S4 compositions enhance the buckling behavior by 73% more than the S1. The buckling deflection takes various behavior during failure mode; Fig. 11-b shows that increasing the boric acid contents in the investigated composition enhances and alerts the failure results from the extra loads. Fig. 12 illustrates the simulation of the primary first sixth buckling failure modes.
3.4.4 Stress analysis
The Von-Mises stress determines whether a material will yield or fracture. A material will yield if its value located under load is equal to or greater than its simple tension yield limit. Fig. 13 shows the effect of the different percentages of boric acid on the maximum equivalent stresses. There is no significant impact on the stress value throughout all the investigated compositions, as illustrated in Table 2. By increasing the load, the equivalent Von-Mises stress is increased, and this result logically due to the loads are approaching the yield point.
3.4.5 Dynamic behavior
The structures and materials have inherent vibration frequencies at which they will deform, a phenomenon known as resonance. Each natural frequency is connected with a mode, which is a shape that describes the deformation of the structure. The first mode is the lowest frequency at which deformation occurs. When a structure vibrates, the initial mode generally specifies the highest loads in the structure or how that structure will interact with the rest of the system surrounding it. The FEA showed the effect of different compositions properties on the dynamics behavior; hence, the resonant frequency on the most important parameters used to identify the dynamic behavior of materials subjected to dynamic or excitation sources near the structure's natural frequency. The modal analysis was used to predict the resonant frequency throughout the various mode shapes. In the current investigation, the primary sixth modes are simulated as shown in Fig. 14, and the corresponding resonant frequencies are represented in Fig. 15. As the excitation source is increased, the resonant frequency moves from resonant mode to another at a specific value. The higher resonant frequencies were achieved by (25 mol.% B2O3) content. The FRF was plotted in Fig. 16; at the first mode shape, the maximum resonant frequency was 360.67 Hz for the composition (25 mol.% B2O3), while the minimum was recorded 232.66 Hz for the (0 mol.% B2O3). According to these results, increasing the mechanical properties of isotropic parameters leads to an increase in the natural frequency value.