Introduction

Bone tissue regeneration is attracting increasing attention due to the large number of patients suffering from bone damage. Calcium phosphate (CaP)-based biomaterials such as hydroxyapatite (HA) and β-tricalcium phosphate (β-TCP)1,2 are commonly used as orthopedic and dental implants.3,4,–5 HA possesses tremendous biocompatibility6 with close resemblance to the mineralized phase of natural bone, which affects the bioresorption rate in the body.7 Several attempts have been made to form HA composites with bioactive glasses (BaGs). The main objective is to enhance the bioactivity, control the resorption rate, and enhance the osseointegration.8,9 However, high-temperature sintering of such HA-based systems results in crystallization of BaGs and/or excessive reaction with HA, which has negative effects on the bioactivity, resorption, and osseointegration.10,11,–12 In this study, Bioglass®45S5 (BG) was incorporated into the HA matrix. Among all bioceramics, BG possesses the highest bioactivity (IB > 8) and forms s bonds with the surrounding tissues.8 Several efforts have been made to improve the biological performance of HA by reinforcing it with BG.13 The primary aim of incorporating BG into the HA matrix is to enhance the bioactive response. The bonding ability of BG to natural bones within a few minutes of implantation is associated with the crystallization of hydroxyl–carbonate apatite (HCA) due to the presence of Si.14 The freshly developed HCA provides a platform for bone cell adhesion, which proliferates and participates in the formation of collagen. The early stages of bone calcification are facilitated by Si, by favoring the adhesion, proliferation, and differentiation of osteoblast-like cells.15 Si plays significant roles in enhancing the skeletal development, connective tissue metabolism, and stimulation of gene activity.16 The replacement of SiO4− and Na+ in the HA unit cell results in enhanced bioactivity.17,18 The bioactivity enhancement of Si-substituted hydroxyapatite (SHA) compared with pure HA is an established phenomenon.19,20 However, the mechanisms responsible for the increased bioactivity of HA in the presence of SiO44− ions remain unclear. BG promotes the decomposition of HA into β-TCP when sintered at low concentrations (< 5 wt.%); however, the formation of other phases occurs at higher concentrations. A volume expansion occurs upon the precipitation of β-TCP.14

In this study, a lower sintering temperature (below the melting point of BG, i.e., 1200°C21,22) was used to avoid excessive reactions between HA and BG which occur when melting BG at higher temperatures. This lower temperature sintering prevents the undesirable chemical reactions between the HA and BG from taking place even at high BG content of 10 wt.%. The poor biological performance of the composites is largely due to the formation of undesirable phases such as Ca5(PO4)2SiO4, Na3Ca6(PO4)5, Na2Ca2Si3O9, etc.23 due to the chemical reactions occurring at higher temperatures. Other researchers have reported severe chemical reactions between the constituents even at a lower BG content of 2.5 wt.%.24 Researchers have reported delayed/slowed apatite layer development and loss of bone resembling CaP phases as a result of these reactions.7,25

Experimental Procedures

BG 45S5 microspheres (200 µm) were procured from XL Sci-Tech Inc. (USA). The BG microspheres were milled for 360 min into powder using a ball mill in a zirconia (ZrO2) jar with ZrO2 balls. The particle size of the BG powder after milling was between 3 µm and 10 µm. Commercial hydroxyapatite (code 21223; Sigma Aldrich) with submicron particle size was procured from Sigma Aldrich (Malaysia). BG and HA were mixed together with different BG contents from 0 wt.% to 30 wt.% to prepare several compositions (Table I) and ball-milled in a polyethylene jar for 360 min using ZrO2 balls. Low-pressure consolidation was carried out at 106.6 MPa in a square die with area of 1 inch2 with filleted edges, followed by conventional sintering (at 1130°C) to prepare the series of HBC composites.

Table I Composition and designations of HA–BG composites
Full size table

Materials Characterization

A diamond cutter was used to cut the sintered samples for further characterization. A PANalytical EMPYREA x-ray diffraction (XRD) instrument with Cu Kα radiation source was used in the phase analysis. The XRD analysis was performed at 30 mA and 40 kV in the 2θ scanning range from 10° to 80° at a rate of 2° min−1 with step size of 0.02°. The microstructural analyses were carried out on a Phenom Pro X table-top scanning electron microscope (SEM), while a high-resolution field emission scanning electron microscope (FESEM, FEI Quanta 200F) was used to observe the morphology of the apatite layer. A thin layer of Au/Pt was deposited on the prepared samples to avoid charging effects during SEM and FESEM analyses. The chemical composition of the samples was determined by energy-dispersive x-ray analyzer (EDAX) attached to the table-top SEM and FESEM. The hardness of the samples was measured using a Shimadzu HMV-G21 DT (Japan) microhardness tester, taking the average of ten hardness readings for each sample to improve the accuracy.

The theoretical density was calculated using the rule of mixture as follows:

$$ \rho_{\text{c}} = \rho_{\text{HA}} \times f_{\text{HA}} + \rho_{\text{BG}} \times f_{\text{BG}} , $$
(1)

where ρc, ρHA, and ρBG are the theoretical densities of the composite, HA, and BG, respectively, while fHA and fBG are the volume fractions of HA and BG, respectively. For these calculations, the theoretical density of fully dense HA was taken as 3.156 g/cm3 and that for BG as 2.45 g/cm3.26

The bulk density of the composite samples was measured using the Archimedes principle as follows:

$$ {\text{BD }} = \frac{{W_{\text{d}} }}{{W_{\text{w}} - W_{\text{s}} }} $$
(2)

where BD is the bulk density of the composite, while Wd, Ww, and Ws are the weights of the composite in the dry, wet, and suspended state, respectively.

In Vitro Bioactivity Analysis

In vitro bioactivity analysis was carried out by examining the development of bone-like apatite film on the samples in simulated body fluid (SBF) as described by Kokubo and Takadama.27 The pH (7.4) and ion concentrations (Na+ 142.0, K+ 5.0, Mg2+ 1.5, Ca2+ 2.5, Cl− 147.8, HCO3− 4.2, HPO24− 1.0, and SO24− 0.5 mM) of SBF are similar to those of human blood plasma. Briefly, SBF was prepared by dissolving reagent-grade K2HPO4·3H2O, CaCl2, NaHCO3, KCl, MgCl2·H2O, NaCl, and Na2SO4 in distilled water at pH 7.4 with hydrochloric acid (HCl) and tris(hydroxymethyl)aminomethane (Tris). The as-sintered samples with average surface area of 650 mm2 were soaked in 65 ml SBF in polyethylene beakers and incubated at 37°C for 7 days. The volume of SBF and the total surface area of the samples were determined as follows:

$$ V_{\text{s}} = \frac{\text{SA}}{10} $$
(3)

where Vs and SA are the volume of SBF and surface area of the sample, respectively.

An appropriate method for determination of the volume of SBF from Eq. 3 was described by Kokubo and Takadama.27 The SBF was refreshed after every 48 h to maintain the ion concentration. FESEM (Hitachi SU 8220) and EDAX analysis were used to analyze the microstructure of the soaked samples and the elemental composition of the precipitated HA. The characteristic infrared spectra of the precipitated apatite were analyzed by Fourier-transform infrared spectroscopy (FTIR, Spectrum 400; PerkinElmer).

In Vitro Resorption Analysis

In vitro biodegradation of the HA–BG composites was carried out to study their resorption behavior. The biodegradation behavior of the CaP-based materials for specific applications must be estimated and optimized to match the bone regeneration rate. Recent studies have suggested that the use of CaP-based materials with an optimized resorption rate could boost bone tissue regeneration through an immunomodulatory activity.28 The in vitro biodegradation test was performed in a Ca-free Tris-based electrolyte, as per the protocol explained by Ito et al..28 Briefly, 0.05 M solution was prepared by dissolving 6.057 g Tris at pH 7.3 ± 0.05 by buffering with dilute (0.1 M) HCl. The composite samples were then soaked in this solution (200 ml) for 5 h under noncontact stirring at 200 rpm at 37 ± 1°C. The electrolyte samples were collected and analyzed for the presence of Ca using atomic absorption spectroscopy (AAS). A standard sample was prepared for calibration of AAS by dissolving 1.250 g CaCO3 (dried for 60 min at 180°C before weighing) in deionized distilled water and dissolved carefully with diluted (0.1 M) HCI. This solution was then diluted to 1000 ml using deionized distilled water (1 ml = 0.5 mg Ca). In the next step, La2O3 (29 g) was dissolved in 250 ml concentrated (12 M) HCI and diluted to 500 ml using deionized distilled water. The La2O3 solution was added to the CaCO3 solution and electrolyte sample (1 ml:10 ml, respectively) after soaking. The Ca concentration in the electrolyte sample with La2O3 solution was then analyzed by AAS (AAnalyst 400 AA spectrometer; PerkinElmer).

Results and Discussion

Conventional Processing of Premixed Compositions

Mechanically stable composite disks were obtained from low-pressure consolidation. As a general finding, the volume shrinkage (densification) during sintering was hampered with increasing BG content.

Phase Analysis of Conventionally Processed Samples

BG reinforcement in HA (< 5 wt.%) is generally considered to be a sintering aid that enhances the densification. The presence of other phases was observed when the BG content was more than 10 wt.%.14 Conventional sintering of HA is usually performed at a temperature ≥ 1200°C. The procured HA (Sigma Aldrich, cod. 21223) had a very high tendency to transform into the β-TCP phase even at a lower sintering temperature of 900°C, as observed by Cuccu et al.29,30 Likewise, the HB 0 C sample (Table I) was completely transformed into β-TCP phase after conventional sintering. The presence of BG at up to 10 wt.% did not promote any undesired reactions between the precursors, and all the samples were composed of a single crystalline phase, i.e., β-TCP (and glassy phase). Demirkiran et al. 31,32 observed undesirable chemical reactions between the precursors at 10 wt.% BG during sintering at temperature of 1200°C, but in the present study, additional phase formation was absent up to 10 wt.% BG addition. Severe chemical reactions between the precursors occur at this temperature as the BG melts at 1200°C.21,22 In this study, the melting of BG was avoided by sintering at 1130°C, which could avoid the undesired reactions between the precursors up to 10 wt.% BG addition. Sintering of the HA–BG system beyond the melting point of BG has been studied and show to yield rhenanite and other undesirable phases resulting in improved fracture toughness and compressive strength.2,3 However, in the absence of these undesirable phases (upon addition of < 5 wt.% BG), decreased compressive strength in comparison with pristine HA has been reported.31,32 Precipitation of β-TCP (which is a less dense and mechanically weaker phase compared with HA) is primarily responsible for this decreased compressive strength.23,26 In the present study, a similar trend was observed in the microhardness when sintering was performed at a temperature below the melting point of BG. In the absence of undesirable phases, the microhardness decreased, whereas upon precipitation of rhenanite, the mechanical properties (microhardness in this case) improved. Figure 1a, b, c, and d shows the XRD patterns of the HB 0 C, HB 2.5 C, HB 5 C, and HB 10 C samples, respectively.

Fig. 1
figure 1

XRD patterns of conventionally processed HB 0 C (a), HB 2.5 C (b), HB 5 C (c), HB 10 C (d), HB 15 C (e), HB 20 C (f), HB 25 C (g), and HB 30 C (h)

Full size image

XRD peaks corresponding to CaPs were completely absent when sintering was performed at 1200°C with BG content > 10 wt.% (i.e., up to 25 wt.%).31,32 Similarly, Bellucci et al.25 reported the absence of CaPs phase but the presence of sodium calcium phosphate (rhenanite) and calcium silicate phases after sintering at 1150°C. In this study, conventional sintering at 1130°C resulted in the combination of the β-TCP and rhenanite phases. The XRD peak intensity of rhenanite increased with the BG content, as shown in Fig. 1e, and g. A single-crystalline rhenanite phase was achieved with the absence of CaPs at a higher BG content in HB 30 C (Fig. 1h). Gong et al.33 proposed the following chemical reaction between CaP and glass to produce the rhenanite phase:

$$ [2{\text{Ca}}_{5} \left( {{\text{PO}}_{4} } \right)_{3} \left( {\text{OH}} \right)]_{\text{cryst}} + [{\text{xNa}}_{2} {\text{O}}^{ \cdot } {\text{ySiO}}_{2} ]_{\text{glass}}\rightarrow [6{\text{NaCaPO}}_{4} ]_{\text{cryst}} [\left( {{\text{x}}{-}3} \right){\text{Na}}_{2} {\text{O}}^{ \cdot } 4{\text{CaO}}^{ \cdot } {\text{ySiO}}_{2} ]_{\text{glass}} + {\text{H}}_{2} {\text{O}}\;[33] $$

Physical and Mechanical Characterization

Figure 1 presents the XRD results, which suggest that HA transforms into the β-TCP phase when sintered conventionally even in the absence of BG, thus the addition of BG into HA could not enhance the densification as reported in literature.14 The presence of BG only decreases the bulk density and the average hardness of the homogenous matrix phase. The porosity increases with increasing BG content, as presented in Table II. On the other hand, the porosity of the bioceramics plays a vital role in driving their biological behavior when these materials are exposed to physiological conditions. Porous CaP bioceramics exhibit enhanced bone-forming bioactive character due to their increased surface area and reactivity.34 The osseointegration of CaPs is primarily influenced by the chemical composition, bioactivity, degradation behavior, and macro/microporosity.35 Several in vivo analyses have confirmed that bone formation takes place preferentially on concave surfaces (formed due to porosity) and internal pores.36 An increase of porosity resulting in poor mechanical properties is associated with higher BG content, as reported by Demerkiran et al.31,32 The glassy region (for HB 10 C to HB 30 C) exhibits very high hardness index compared with the CaP and rhenanite phases. The average Vickers hardness index and relative density % (RD %) are plotted in Fig. 2a. The average hardness of the CaP phase decreases significantly with the presence of BG, due to the lower density. The rhenanite phase exhibited intermediate hardness values, lying between those of the CaP and glassy phases for all the samples, as presented in Table II. The microhardness indentation of the HB 10 C sample is shown in Fig. 2b, indicating the presence of a glassy region and high hardness index compared with the CaP and rhenanite phases, as observed in the HB 15 C to HB 30 C samples. The secondary cracks appearing in the micrograph are usually related to the lower hardness values observed at low indentation loads (as used in this study) due to the reverse indentation size effect (RISE). However, the higher hardness values presented herein are intended only to support the idea regarding the presence of glassy phase in the composite samples. The very high hardness, approaching the theoretical hardness of BG (i.e., 458 HV), supports this hypothesis.13

Table II Comparison of physical, mechanical, and resorption characteristics of scaffolds
Full size table
Fig. 2
figure 2

RD and average hardness of conventionally processed samples (a) and micro Vickers hardness indent on glassy phase of the HB 10 C sample (b)

Full size image

In Vitro Resorption Analysis

The dissolution/resorption rate is a critical parameter in the design of bone tissue engineering scaffolds.28 The scaffold resorption behavior must be optimized according to the specific requirements.14 The factors which affect the resorption behavior of CaP-based scaffolds are the porosity, presence of phases, processing technique, and chemistry of the biological fluid. Furthermore, the latest research findings correlate the optimized resorption behavior of the CaP-based materials with the enhanced regeneration of bony tissues.37 The in vitro Ca dissolution exhibited an increasing trend with increasing BG content. Increasing the BG content enhanced the Ca dissolution due to the higher Ca dissolution in BG compared with CaP. The poor densification due to the increase in porosity at higher BG content is another reason for the enhanced Ca dissolution. Table II presents the Ca dissolution values (mg/l) observed in the HB 0 C, HB 10 C, HB 20 C, and HB 30 C samples and their densities.

Microstructural and Elemental Analyses

SEM and elemental analysis were performed on HB 0 C, HB 10 C, HB 20 C, and HB 30 C to investigate the microstructural characteristics and chemical composition of the phases. Figure 3 shows the microstructure and elemental composition of the HB 0 C sample with a homogenously distributed porosity in the darker regions. The brighter region represents the completely sintered HA, which transformed into the β-TCP phase as explained in the previous section. The EDAX analysis confirmed the elemental homogeneity of the single-phase β-TCP. The red plus sign indicates the point of the EDAX investigation in Figs. 3, 4 and 5. Ca and P are the main elements apart from O, which is present in the CaPs and represents the porosity, as shown in Fig. 3a.

Fig. 3
figure 3

Microstructure and EDAX analysis of HB 0 C (a) and HB 10 C (b)

Full size image
Fig. 4
figure 4

Microstructure and EDAX analysis of HB 20 C

Full size image
Fig. 5
figure 5

Microstructure and EDAX analysis of HB 30 C

Full size image

Figure 3b presents the microstructure and EDAX pattern of the HB 10 C sample. Two distinct phases and increased dark regions with higher porosity are clearly visible due to the higher BG content. Figure 3b shows the lower-density regions with higher BG content. The lower content of Na in the glassy phase is due to the outward diffusion of Na from the BG phase. The microstructure of the central glassy region assumes a rounded morphology with lower Na content due to the outward diffusion. The crystallization ability of the glass decreases due to the outwards diffusion of Na from the glassy region.38, 39 The higher Si content, circular morphology, and absence of any Si-containing crystalline phase (Fig. 1) confirms the development of the glassy phase. A glassy phase with similar composition was also reported in earlier studies.

Three distinct phases can be observed in Fig. 4. The region on the right of the image with uneven surface due to the homogeneous open porosity is rich in Na, which confirms the presence of the rhenanite phase. The glassy region is also visible, with higher Si content but lower Na content. The top-left region appears similar in morphology to the CaP phase, as shown by HB 10 C in Fig. 3b. CaP is the dominant phase according to the elemental mapping of the HB 20 C sample presented in Fig. 6.

Fig. 6
figure 6

FESEM image and EDAX pattern of apatite layer formed on HB 0 C (a–d), apatite layer formed on HB 15 C (e, f), apatite layer with EDAX pattern of HB 30 C (g–i), and FTIR patterns of the apatite layer developed on samples (j)

Full size image

The microstructure of the HB 30 C sample is presented in Fig. 5 and shows very high porosity and a Si-rich glassy phase. The nonglassy region shows very high Na content and relatively lower Si content, which is expected to consist of the rhenanite phase as presented in Fig. 1g, and h. Large air pockets (porosity) are particularly visible at the interface of the rhenanite and glassy region. The elemental distribution behavior is confirmed for the HB 20 C sample through elemental mapping as shown in Supplementary Fig. S1.

In Vitro Bioactivity Analysis

Bioactive materials form bonds with living tissues via the formation of a HA layer on the surface. This biologically active apatite (HCA) is a fundamental prerequisite for artificial bone materials to bridge with the scaffolds and surrounding host tissues.40 The HB 0 C, HB 15 C, and HB 30 C samples (650 mm2) were immersed in SBF for 7 days at 37°C to investigate the bioactivity behavior of the conventionally processed composite scaffolds. The surface of the samples was completely covered with a porous and cracked apatite layer. Such cracking and porosity of the HCA layer is commonly observed due to the drying of the prepared samples after the immersion process.41

The deposited layer was investigated by EDAX and FTIR analyses (Fig. 6). The EDAX analysis (Fig. 6) confirmed the presence of Ca and P, with minor content of Na and Mg. Na and Mg are commonly observed with Cl due to the precipitation of chloride compounds in SBF.42 The Ca/P ratio of the apatite layers was more than 1.4, which is close to the ideal Ca/P ratio of HCA.

The FESEM microstructures and EDAX spectra of the developed apatite layers are presented in Fig. 6. Figure 6a, b, c, and d presents the apatite layer developed on the HB 0C sample and the EDAX spectra. The surface was completely covered with apatite, and thick growth of apatite can be seen in Fig. 6a. The characteristic cauliflower appearance of the apatite is visible in Fig. 6d. Figure 6e, and f presents the apatite layer on the HB 15 C sample at different magnifications, while Fig. 6g, and h presents the apatite layer on HB 30 C at similar magnifications. Comparison of Fig. 6a and g (at the same magnification) reveals that the developed apatite layers exhibited some morphological differences. The developed apatite layer assumed a more homogenous and flatter morphology with increased BG content (in the case of HB 15 C and HB 30 C), while an increased presence of blisters could be observed on the firm layer in the absence of BG (for HB 0 C). These differences in the morphology of the apatite indicate a delay in the apatite formation in samples with higher BG content compared with the HB 0 C sample due to the formation of rhenanite phase. Such a delay in the deposition of apatite due to the formation of phases resulting from excessive reactions has already been reported in literature.7,25

Furthermore, the FTIR spectra of the formed apatite layer significantly resemble the characteristic FTIR pattern of HCA.43 The FTIR spectra of the samples confirmed the characteristic V4 PO43− bending mode at around 550 cm−1 and 660 cm−1 and the V3 PO43− bending mode at around 1050 cm−1. The C–O bending vibration occurs at around 870 cm−1, while the V2 O–H vibration occurs at around 1630 cm−1 for all the samples, as shown in Fig. 6j.

Conclusion

  1. 1.

    Conventional sintering of HA–BG composites resulted in severe chemical reactions between the precursors, due to the higher heat exposure. The formation of rhenanite (NaCaPO4) phase occurred due to the reaction between the HA and BG, yielding a combination of β-TCP, rhenanite, and glassy phases (for BG contents between 10 wt.% and 25 wt.%). However, CaP phase was absent from HB 30 C, which was thus composed of crystalline rhenanite and glass phase only.

  2. 2.

    The physical/mechanical properties of the conventionally processed HA-BG composites were largely degraded with increasing BG content. Highly porous regions were particularly found close to the CaP and glassy interface. The dissolution (resorption) of Ca increased with the increased porosity and BG content. Very high content of Na could be observed in the nonglassy regions of HB 20 C and HB 30 C, due to the reaction between the HA and BG to form the rhenanite phase.

  3. 3.

    HCA formation on the conventionally processed composites was visible for all samples after 1 week of immersion in SBF. Apart from the β-TCP and BG phases, literature also confirms the bioactivity of the rhenanite phase. However, the morphology of the apatite layer suggests that the apatite layer growth on the HB 0 C was at an advanced stage compared with the HB 15 C and HB 30 samples. This difference in the morphology of HCA could be associated with a delay in the apatite formation tendency (though still bioactive) of the rhenanite phase, as reported in literature.