Substantial enhancement of corrosion resistance and bioactivity of magnesium by incorporating calcium silicate particles

Abstract

Biodegradable metal matrix composites (MMCs) with pure magnesium as the matrix and bioceramic calcium silicate (CS) as the reinforcement phase were fabricated by means of spark plasma sintering (SPS). The microstructure, mechanical properties and degradation behavior of the composites as well as the cellular responses to these composites were investigated. The formation of CS networks in the Mg matrix was observed when the CS content reached 20%. Among the composites with CS weight percentages ranging from 10% to 40%, the composite containing 20% CS possessed the highest structural compactness and compressive strength. Immersion tests in simulated body fluid (SBF) revealed that the Mg–20% CS composite exhibited a substantially enhanced corrosion resistance as compared with pure Mg, which was attributed to the formation of a hydroxyapatite (HA) layer on the surface as a result of the presence of the CS networks throughout the Mg matrix. In addition, ionic products from the interaction between the Mg–20% CS composite and SBF brought about a significant stimulatory effect on the alkaline phosphate (ALP) expression of MC3T3-E1 osteoblast cells. Our results indicate that CS is an effective reinforcement phase to improve both the corrosion resistance and bioactivity of Mg and the Mg–CS composites developed in this research are able to overcome the inherent drawbacks of magnesium as a biodegradable implant material.

---

1. Introduction

Magnesium, and its alloys, as biodegradable orthopedic implant materials have, over the last decade, attracted a great deal of attention as these materials offer an intriguing solution to long-standing problems associated with conventional metallic and polymeric materials for orthopedic implants. This is mainly because Mg-based alloys are not only biodegradable, but also possess favorable mechanical properties, especially their elastic moduli being closer to those of bone than any other biometals.^1,2 However, a number of inherent drawbacks of biodegradable Mg alloys remain an obstacle to their clinical applications. The biggest drawback concerns the fast corrosion in chloride-containing solutions, including human body fluids, and blood plasma, which leads to the rapid release of hydrogen gas and a loss of mechanical integrity before the complete regeneration of defect bone.^3,4 In addition, Mg has insufficient bioactivity; although some recent studies have shown that Mg alloy may enhance bone regeneration by releasing Mg ions that may be able to stimulate bone marrow stromal cells,⁵ gaps could still be observed between the implant and surrounding bone tissue after implantation in earlier in vivo trials.^6,7

Obviously, to meet the requirements of biodegradable materials for orthopedic applications, further efforts are needed to slow down the degradation of Mg-based materials, in order to match the growth of new bone, and also to improve their bioactivity. In recent years, the development of Mg-based metal matrix composites (MMCs) have proven to be an effective way towards the desired improvements in the biodegradability and bioactivity of Mg-based biomaterials.^8–10 Driven by the automotive industry for lightweight structural parts, Mg-MMCs were first developed, aiming for high thermal stability along with high specific strength, elastic modulus and wear resistance.¹¹ Only in recent years have a limited number of studies been conducted on Mg-MMCs for biomedical applications with the aim of reducing the biodegradation rate of magnesium and enhancing its bioactivity. Almost all researchers used a type of calcium phosphate ceramic as the reinforcement phase in Mg-MMCs. Ye et al.,¹² for example, studied Mg-MMCs with a Mg-Zn-Zr alloy as the matrix and hydroxyapatite (HA) particles as the reinforcement phase, and found that the composites were cytocompatible biomaterials with enhanced corrosion resistances. Similar findings were reported on Mg-MMCs with nano-HA particles or β-tricalcium phosphate (β-TCP) particles as reinforcements.^13,14 However, from both biodegradability and bioactivity points of view, the chosen calcium phosphate ceramics with a high stability, e.g., sintered HA or β-TCP, may not be the most appropriate choices because of their low bioactivity.¹⁵ In the opinion of the present authors, for Mg-based MMCs intended for orthopedic applications, a more preferable reinforcement phase should be one that is more biodegradable and bioactive than calcium phosphate ceramics. Silicate-based ceramics are just such a class of biodegradable and bioactive materials that may better serve the purpose of developing biomedical Mg-MMCs. In recent years, silicate-based bioceramics have received more and more attention, mainly because they exhibit better bioactivity than calcium phosphate ceramics.^16–18 Among all the silicate-based bioceramics, calcium silicate (CS) has the simplest chemical composition. Recent studies have confirmed that it can promote bone formation by promoting human mesenchymal stem cells (hMSCs) and osteoblast-like cell adhesion, proliferation and differentiation, thanks to the release of Si ions.^17,19,20 In addition, in vivo evaluations have confirmed that CS can gradually degrades after implantation without inducing local or systemic toxicity, inflammation or foreign-body responses.²¹ Having these characteristic properties, CS appears to be an interesting reinforcement material for pure Mg and Mg alloys, especially with regard to the potential to enhance their biomedical properties.

The present study was intended to prove the hypothesis that CS would indeed be an adequate reinforcement material for Mg to form Mg–CS MMCs as a new class of biodegradable and bioactive composite materials for orthopedic applications. In this connection, Mg–CS MMCs with different weight percentages of CS particles were fabricated by means of spark plasma sintering (SPS), also known as pulsed electric current sintering (PECS). The rationale behind the selection of SPS as the fabrication method was its advantages over other powder metallurgy methods. The features associated with SPS, such as rapid heating and cooling capabilities, sintering under pressure and in a desired atmosphere, or in vacuum, were considered. The particular importance for this processing of MMCs²² is because all these factors would lead to minimum contaminations to the MMCs and facilitate the attainment of a high degree of structural compactness during the process. For the processing of Mg-based materials, and Mg-MMCs in particular, during SPS, an oxide film present on powder particle surfaces, due to the high reactivity of Mg, ruptures as a result of high pressure and high local temperatures on powder particle surfaces generated by the pulsed electrical current.^22,23 With the Mg–CS MMCs successfully prepared, their microstructures, mechanical strength and degradation behavior were characterized and their in vitro surface bioactivity was evaluated by performing immersion tests in simulated body fluid (SBF). The effect of ionic products formed during in vitro degradation on the proliferation and differentiation behavior of osteoblast-like cells were assessed and compared with those of Mg without CS.

2. Experimental

2.1 Material fabrication, structural characterization and mechanical tests

The starting materials were: Mg elemental powder with spherical particles (99.9% purity, supplied by Haotian Nano Technology Co., Ltd.) and CS powder. The Mg powder was produced by using the inert gas (argon) atomization technique, and had a mean particle size of 20 μm. The CS (CaSiO₃) powder was synthesized by making use of a chemical precipitation method developed earlier,²² in which Ca(NO₃)·4H₂O and Na₂SiO₃·9H₂O were used as the raw materials. The mean particle size of the CS powder was 10 μm.

Prior to SPS, the Mg and CS powders were mixed for 30 min to prepare homogeneous mixtures with 0, 10, 20, 30 and 40% CS by weight. The powder mixtures were then loaded into a cylindrical graphite die with a diameter of 10 mm, and sintered using an SPS system (Dr Sinter 2040, Sumitomo Coal Mining Co.) in vacuum (less than 10⁻⁴ Pa) under a pressure of 60 MPa. The heating rate was set at 100 °C min⁻¹ and the holding time at the sintering temperature of 550 °C was 5 min. In other studies, such an SPS condition could densify Mg- and Mg-alloy-based composites fully,^22,23 although the SPS temperature was lower than the melting point of pure Mg (650 °C). In the present materials preparation, an SPS temperature higher than 550 °C was not considered an option in order to avoid severe interactions between the constituents in the Mg–CS mixtures, as indicated by the appearance of numerous cracks in sintered composites in our preliminary SPS trials, likely due to exothermic reactions.

To identify the phases in the Mg–CS MMCs prepared by using SPS, an X-ray diffractometer (XRD, Geigerflex, Rigaku, Japan) was used with 2-theta angles ranging from 10 to 80°. For microstructure observation, specimens were ground down to a grid size of 1200 and polished with lubricants containing 3 μm diamond particles with the addition of dish liquid. Surface morphologies and elemental distributions of the Mg–CS MMCs were determined by using a scanning electron microscope (SEM; JSM-6700F, JEOL, Japan) equipped with an energy dispersive X-ray (EDX) spectrometer (INCA Energy, Oxford Instruments).

Compression tests at room temperature were carried out using a universal testing machine (Shimadza, AG-5KN, Japan). Cylindrical specimens with a diameter 10 mm and a height of 20 mm were tested and all the tests were conducted in triplicate. The fracture surfaces of the tested specimens were observed using SEM (JSM-6700F, JEOL, Japan).

2.2 Immersion tests

To determine the degradation behavior of the Mg–CS MMCs, samples in the form of a disk with a diameter of 10 mm and a thickness of 2 mm were prepared and immersed in simulated body fluid (SBF), prepared according to the method described by Kokubo,²⁴ and its composition is presented in Table 1. The immersion tests were performed at 37.0 °C in a water bath. For each sample, 250 mL SBF was used based on the consideration that the influence of the solution volume/surface area (SV/SA) ratio on the corrosion rate of the sample would become negligible when the SV/SA ratio was higher than 6.7.²⁵ After immersing in the SBF for 7 days, samples were taken out of the solution and cleaned with chromate acid (200 g L⁻¹ CrO₃ + 10 g L⁻¹ AgNO₃) for 5 min to remove surface corrosion products.²⁶ Then, they were rinsed with distilled water and ethanol followed by drying in air. The dried samples were weighed and the corrosion rate (CR) was calculated according to eqn (1):²⁶

CR = Δm/At (1)

where CR is the corrosion rate (mg cm⁻² h⁻¹), Δm the weight loss of the sample (mg), A the original surface area exposed to the corrosive medium (cm²), and t the immersion time (h). For reproducibility, three samples in each group were subjected to the immersion test and weight change measurements.

Table 1 Composition of SBF (1000 mL) used in the immersion tests

	NaCl	NaHCO3	KCl	K2HPO4·3H2O	MgCl2·6H2O
Amount	8.035 g	0.355 g	0.255 g	0.231 g	0.311 g
Purity (%)	99.5	99.5	99.5	99.0	98.0

---

	1.0 M HCl	CaCl2	Na2SO4	Tris	1.0 M HCl
Amount	39 mL	0.292 g	0.072 g	6.118 g	0–5 mL
Purity (%)	—	95.0	99.0	99.0	—

---

Samples in the form of a disk 10 mm in diameter and 2 mm in thickness were immersed in the SBF at 37.0 °C with a surface area to volume ratio of 0.1 cm⁻¹.²⁷ After 24 hours of immersion, the samples were taken out of the medium, gently rinsed with deionized water and dried in air. Their surfaces were then characterized by grazing incidence X-ray diffraction (GIXRD; Geigerflex, Rigaku, Japan) and SEM (JSM-6700F, JEOL) equipped with an EDX spectrometer (INCA Energy, Oxford Instruments), in order to investigate the surface layer formed at the early stage of immersion.

2.3 Cytotoxicity tests of the degradation products

MC3T3 (MC3T3-E1 Subclone 14) cells were supplied by the cell bank of the Chinese Academy of Sciences (Shanghai, China). Cytotoxicity tests were carried out using an extract assay, one of the testing methods that is most commonly used for evaluating the cytocompatibility of Mg-based biodegradable materials, as influenced by extraction medium concentration.^28,29 The extracts were prepared by immersing the samples in Dulbecco's modified Eagle's medium (DMEM) with an medium volume to surface area ratio of 1 mL cm⁻². The supernatant fluid was withdrawn, centrifuged and then stored at 4 °C before use. The ion concentrations of DMEM and the as-prepared extracts were measured by using an inductively coupled plasma optical emission spectrometer (ICP-OES, Varian 715-ES). MC3T3 cells were seeded in cell culture plates with 2 × 10³ cells per 100 μL medium in each well. After incubation for 24 hours, the medium was replaced with 100 μL of the as-prepared extracts. The control groups were treated by adding DMEM with 10% fetal bovine serum (FBS). MTT tests were performed to assess the number of viable cells.³⁰ After incubation for preselected periods, 100 μL of the 0.5 mg mL⁻¹ MTT solution was added to the well plate. After further incubation for 4 h at 37 °C, 100 μL of dimethyl sulfoxide was added to each well. Subsequently, optical density (OD) at 590 nm was measured with an enzyme-linked immunoadsorbent assay plate reader (ELX800, Bio-TEK). For each group, six samples were tested and the results were compared in OD units.

2.4 Alkaline phosphate activity assay

To investigate the alkaline phosphate (ALP) activity, MC3T3 cells were cultured for 10 days under the same culture conditions as described above. The method to quantitatively determining the ALP activity was adapted from Lowry et al.³¹ MC3T3 cells were extracted from the extracts and permeabilized by using 0.1% Triton X-100 solution (Sigma). Thereafter, cell lysates from each sample was used for ALP assays. A spectrophotometer (UV-VIS 8500, Shanghai, China) was used to measure the absorbance at 405 nm. The ALP activity was calculated from a standard curve after normalizing to the total protein content, for which the results were expressed as nanomoles of p-nitrophenol produced per min per mg of protein. The ALP activity of the cells cultured in the medium without the addition of extracts was taken as the control. Five samples were tested for each group and the test was performed in triplicate for reproducibility.

2.5 Statistical methods

All experimental data were analyzed using the Student's t-test and expressed by the mean values ± standard deviation (SD) in which a p-value < 0.05 was considered statistically significant.

3. Results and discussion

3.1 Phase compositions and microstructural characteristics of the Mg–CS MMCs

XRD patterns of pure Mg, Mg–20% CS, Mg–40% CS and pure CS are shown in Fig. 1. It can be seen that only the peaks corresponding to the Mg matrix and the CS phase were found in the XRD patterns of the pure Mg and CS samples, respectively. For the Mg–CS composites, the peaks related to Mg and CS were identified and no new crystalline compounds, e.g., MgO or intermetallic compounds, could be found. The intensities of the XRD peaks corresponding to the CS phase increased and those corresponding to the Mg phase decreased with increasing content of CS in the composites. The results of the XRD analysis suggested that no severe chemical reactions took place during the SPS process at 550 °C. Both Mg and CS could therefore maintain their bulk chemical properties under the current SPS conditions.

Fig. 1 XRD spectra of Mg (a), Mg–20% CS composite (b), Mg–40% CS composite (c) and CS (d).

The microstructures of pure Mg and the composites were observed using the back-scattered electron mode in SEM, and the results are presented in Fig. 2. As observable in Fig. 2a, the pure Mg sample was well sintered and contained almost no pores, indicating that the sintering temperature (550 °C) used in the present study was high enough to obtain a nearly compact structure, which was in agreement with the findings from earlier studies on the microstructures of other Mg-based materials after SPS.^22,32 In contrast, the microstructures of the Mg–CS composites were composed of dual phases, i.e., a dark background and continuously distributed white particles (Fig. 2b–e), being consistent with the results obtained from the XRD analysis (Fig. 1). It was obvious that the dark background and white particles belonged to the Mg and CS phases, respectively. It was also observed that, when the content of CS was 10%, the CS phase was scattered as individual particles and part of this phase also formed interconnected particle clusters distributed in other places of the matrix (inset of Fig. 2b). The Mg matrix and CS reinforcement particles appeared to be well bonded at the interface. When the content of CS increased from 10% to 40% (Fig. 2c–e), CS particles gradually formed continuous networks throughout the Mg matrix. These networks were composed of agglomerates of CS ceramic particles and grew with increasing content of CS in the composites. It is worth mentioning that the addition of CS up to 40% to Mg, in our study, did not induce the formation of pores or cracks in the Mg matrix, indicating that CS particles did not interfere with the inherent sintering behavior of Mg particles. It is, however, important to note that the entanglement of the CS networks with the Mg matrix became disturbed as the content of CS increased. As seen in the insets of Fig. 2b and c, the CS phase completely permeated into the Mg matrix when the weight percentages of CS were 10% and 20%. This could be attributed to the behavior of Mg during the SPS process. The Mg phase underwent plastic deformation and was pressed into the agglomerates of CS particles, thus serving as a binder between CS particles. The same observation was made in other studies on the microstructures of Mg-MMCs after SPS.^24,33 However, as the content of CS increased to 30%, Mg failed to infiltrate completely into the CS networks and, as a result, voids appeared between CS particles at the center part of the networks as pointed out by arrows in the inset of Fig. 2d. With a further increase in CS content up to 40%, the inner part of the CS networks presented itself as a porous structure, as shown in Fig. 2e and its inset. The failure to fill the voids during SPS could be attributed to the temperature being much lower than what was necessary for the sintering of CS powder (above 900 °C) and insufficient pressure to force Mg to flow into the gaps between CS particles.^19,34 From the results presented in Fig. 2, it is clear that a completely compact structure can be obtained if the weight percentage of CS is limited to 20%.

Fig. 2 SEM micrographs of pure Mg (a), Mg–10% CS (b), Mg–20% CS (c), Mg–30% CS (d) and Mg–40% CS (e) composites. The insets show the interfaces between the matrix and agglomerated CS particles. Red arrows point to the voids within the CS networks in the composites with 30% and 40% CS.

To investigate the interface between the Mg matrix and the CS phase further, an EDX line scan analysis of the Mg–20% CS sample across an interface was performed and the results are depicted in Fig. 3. Clearly, SPS yielded sound Mg–CS interfaces that were free from oxides and undesired porosity. Such well bonded interfaces must be favorable for the mechanical strength of the composites, as it is known that the interfaces between the metal matrix and ceramic reinforcement particles are preferential sites for the initiation of cracks, once the composite is under stress, and are often responsible for its premature failure,^35,36 especially when the product of interfacial reactions weakens the interfacial bonding between reinforcement particles and the matrix.³⁷ In addition, the results from the EDX analysis further confirmed our thought, based on the results of the XRD analysis (Fig. 1), that both the Mg and CS phases were well retained after SPS processing. This indicated that the individual chemical, physical and biomedical properties of the Mg matrix and CS particles were well retained.

Fig. 3 SEM morphology of the Mg–20% CS composite (a), on which a line scan EDS analysis across a boundary region from CS particles to the Mg matrix was performed to reveal the elemental distributions of Mg (b), Ca (c), Si (d) and O (e).

3.2 Mechanical properties of the Mg–CS MMCs

The compressive stress–strain curves of Mg and the Mg–CS MMCs are shown in Fig. 4a. The compressive strengths of the composites with different CS weight percentages are shown in Fig. 4b. It can be seen from Fig. 4a that all the samples exhibited extensive elastic deformation followed by strain hardening until a stress peak appeared. Fractures occurred shortly after brief inhomogeneous deformation. The compressive strength of Mg was indeed improved with the addition of 10%, 20% and 30% CS to different extents, and the composite with 10% CS showed the highest strength value (Fig. 4b). Comparison between the compressive strengths of the composites showed that there was no significant difference between the values of Mg–10% CS and Mg–20% CS (p < 0.05). However, the compressive strength of the Mg–30% CS composite was significantly lower than the strengths of the Mg–10% CS and Mg–20% CS composites, although this strength value was still higher than that of pure Mg (p < 0.05). With a further increase in CS content to 40% (Mg–40% CS), the compressive strength of the composite further decreased to a value comparable to that of pure Mg (p > 0.05).

Fig. 4 Compressive stress–strain curves (a) and strengths (b) of the Mg–CS composites with different weight percentages of CS.

The fracture surfaces of the samples after the compression tests are shown in Fig. 5. For pure Mg, the surface was composed of cleavage-like smooth fractures that were interspersed with some shallow ridges (Fig. 5a) typical of magnesium with intrinsic brittle characteristics. For the Mg–20% CS composite (Fig. 5b), the fracture surface of the Mg matrix part exhibited similar characteristics to that of pure Mg, and no severe voids and cracks were discernable at the interfaces between CS agglomerates and the Mg matrix. However, as the CS content increased to 40% (Fig. 5c), CS particles appeared to have been crushed and cracks initiated from agglomerated CS particles at an early stage of compressive testing quickly propagated, resulting in small plastic strains (Fig. 4a). The improved compressive strengths of the composites with 10–30% CS could be attributed primarily to the effective transfer of applied compressive load to the networks composed of CS particles, a strengthening mechanism that has been well established both experimentally and theoretically.^22,34,38,39 In addition, as the CS networks were entangled with the Mg matrix, extensive boundaries between CS particles and the Mg phase could act as effective obstacles to dislocation movement, thereby also contributing to the strengthening of the matrix material.^38–40 It was of interest to note that the presence of the networks of reinforcement particles seemed to be more effective in enhancing the compressive strength of the Mg matrix (by more than 30%) as compared with that of homogeneously dispersed particles. Feng et al.,⁴⁰ for example, reported that calcium polyphosphate (CP) particles added to ZK60A magnesium alloy that were homogeneously dispersed in the composites containing 10%, 20% and 30% CP, contributed by extrusion deformation after hot pressing, and the composites showed improvements in compressive strength up to 9%. Further studies on the strengthening mechanisms of composites with different reinforcing particles and their distributions will be necessary to ascertain the relative importance of different strengthening mechanisms.

Fig. 5 SEM micrographs of the compressive fracture surfaces of pure Mg (a), the Mg–20% CS composite (b) and the Mg–40% CS composite (c).

3.3 Degradation behavior of the Mg–CS MMCs

The corrosion behaviors of pure Mg and the Mg–CS composites were evaluated by performing immersion tests in SBF. It was observed that, as compared with the response of a pure Mg sample to SBF, fewer hydrogen bubbles emerged from the surfaces of Mg–10% CS, Mg–20% CS and Mg–30% CS composite samples during the first several minutes of immersion, whereas a large number of hydrogen bubbles rose from the surface of the Mg–40% CS composite sample. The average weight changes of pure Mg and Mg-MMC samples during the 7 days of immersion in SBF are shown in Fig. 6. It is clear that the composites with 10% and 20% CS had significantly less weight loss compared with pure Mg, and the Mg–20% CS composite had the lowest rate of weight loss (0.23 mg cm⁻² h⁻¹), being only about one sixth of the weight loss rate of pure magnesium (1.4 mg cm⁻² h⁻¹). However, as the content of CS in the composite increased to 30%, the weight loss rate increased to 1.66 mg cm⁻² h⁻¹, and some detached corrosion products in the form of particles were observed at the bottom of the beaker. When the CS content increased further to 40%, the composite sample had the highest weight loss rate among all the samples, and a large number of drop-off particles were observed on the surface of the sample. The significantly increased weight loss rates of the Mg–30% CS and Mg–40% CS composites were attributed to the fact that these composites reached relatively low levels of structural compactness achieved during SPS, which facilitated the penetration of the immersion electrolyte into the materials and thus led to dramatically accelerated corrosion.^41,42

Fig. 6 Average corrosion rates determined from mass loss testing of pure Mg and Mg-MMC with different weight percentages of CS.

It is clear from the results of the immersion tests that the addition of an appropriate weight percentage of CS could indeed improve the corrosion resistance of pure Mg under the condition that the as-fabricated composite had a compact structure without voids. To illuminate the role of CS particles in affecting the corrosion of the composite further, the composition and microstructure of the surface corrosion layer on the Mg–20% CS composite sample was investigated, as it showed the most enhanced corrosion resistance among all the samples, and compared with those of the pure Mg sample. Fig. 7 shows the GIXRD diffraction patterns of pure Mg (Fig. 7a) and the Mg–20% CS composite (Fig. 7b) samples after immersion in SBF for 24 h. For pure Mg, only the peaks from Mg and Mg(OH)₂ were identified. By contrast, in addition to those found in the GIXRD patterns of pure Mg, peaks of CS, MgO and hydroxyapatite (Ca₅(PO₄)₃(OH), HA) with low crystallinity on the surface of the Mg–20% CS composite sample were detected. SEM and EDX analysis of the surfaces of pure Mg and the Mg–20% CS composite samples after immersion in SBF for 24 h indicated interactions between the samples and SBF. SEM micrographs (Fig. 8a and d) clearly show that both the surfaces of the samples were covered by a reaction layer. However, the morphology and composition of the layer on the Mg sample differed from those on the Mg–20% CS sample. A large number of cracks were observed on the surface layer of the Mg sample (Fig. 8a) and the layer was composed of plate-like crystals (Fig. 8b). EDX analysis unveiled Mg and O as the main elements in the layer (Fig. 8c), indicating Mg(OH)₂ to be the main compound of the layer. By contrast, there were much fewer cracks in the surface layer of the Mg–20% CS composite sample (Fig. 8d) and the layer presented itself as a more compact structure. Higher magnification SEM images showed that the layer on the composite sample was composed of aggregates of nanocrystals (Fig. 8d) and EDX analysis revealed that the layer contained Mg, O, Ca, and P, stemming from Mg(OH)₂ and HA compounds, according to the results of GIXRD (Fig. 7).

Fig. 7 XRD spectra of Mg (a) and Mg–20% CS composite (b) after immersion in SBF for 24 hours.

Fig. 8 SEM morphologies and EDS results of pure Mg (a–c) and the Mg–20% CS composite (d–f) after immersion in SBF for 24 hours.

Based on the results presented above, the corrosion process and reasons for the improved corrosion resistance of the Mg–20% CS composite material can be fathomed by adopting and combining the degradation mechanisms of Mg and CS, proposed by Song⁴³ and Ni,⁴⁴ respectively. In the case of pure Mg, a Mg(OH)₂ corrosion layer is formed immediately after immersion in SBF and acts as a barrier against further corrosion.⁴⁵ However, the effectiveness of such a layer in protecting the Mg matrix from corrosion is rather limited, since the insoluble Mg(OH)₂ compound will be transformed into the soluble MgCl₂ compound due to the presence of corrosive chloride ions in SBF, and hence, cracks tend to form in the layer that facilitate penetration of the electrolyte into the Mg matrix leading to accelerated corrosion thereafter.^43,45 It has been proposed that, during immersion in SBF, the deposition of calcium phosphate above the Mg(OH)₂ layer can improve the corrosion resistance of the surface, but it usually takes more than 21 days before a sufficient quantity of calcium phosphate crystals are formed.⁴⁵ Although an addition of nano-size HA particles to Mg can accelerate the precipitation of calcium phosphate from SBF, it still needs more than 72 h for the formation of Mg-containing calcium phosphate on the surface.⁴⁶ From the observations made in this study, we believe that the beneficial effect of CS on the corrosion resistance of the Mg matrix is brought about because the presence of CS can accelerate the deposition of HA on the surface. It has been found in studies on bioceramics that CS possesses a superior ability to induce the deposition of HA from a SBF solution to CP bioceramics.⁴⁴ This has been attributed to the release of Si ions from CS during its degradation that effectively facilitates the fast nucleation of HA crystals within 24 h.^44,46,47 It is now confirmed that the presence of CS particles in the Mg-matrix composites can significantly improve the early-stage corrosion resistance of the magnesium matrix, which is of great importance for biomedical applications of Mg-based implant materials.⁴⁸ In addition, a high level of bone-like apatite formation found on the Mg–20% CS sample can improve the surface bioactivity of the material by enhancing osteoblastic activity, thus exerting a stimulatory effect on the growth of new bone tissue.⁴⁹

3.4 In vitro cytocompatibility and ALP activity

The differences in ion concentration between the original DMEM and the extracts are shown in Fig. 9. It can be seen that, as a result of the corrosion of Mg, the amounts of Mg in the extracts of both pure Mg and the Mg–20% CS composite samples were significantly increased compared to the amount in the original DMEM. Mg ion concentrations in the pure Mg extract were significantly higher than that in the Mg–20% CS extracts, which was in agreement with the weight loss rates found from the immersion tests reported earlier. The amounts of calcium and phosphate were not different between the original DMEM and the pure Mg group. However, contact of the Mg–20% CS composite with DMEM led to statistically significant decreases in the concentrations of Ca and P (p < 0.05) due to the deposition of HA and the consumption of these ions in the solution.⁵⁰ It was of special interest to note the presence of Si ions in the extract of the Mg–20% CS group that obviously resulted from the dissolution of CS particles in the composite. This confirmed our hypothesis that, unlike sintered CP ceramic particles, the CS phase could degrade along with the degradation of the Mg matrix.

Fig. 9 Concentrations of Mg, Ca, Si and P ions in pure Mg and Mg–20% CS composite extracts as compared with those in the original DMEM.

Fig. 10 shows the viability of MC3T3 osteoblasts cultured in the pure Mg and Mg–20% CS composite extraction media at dilution ratios of 1 (i.e., the original extract), 1/4, 1/16, 1/64 and 1/128 for 1, 3 and 7 days. Cell viability is expressed as a percentage of the viability of cells cultured in the control. It can be seen that the original pure Mg and the Mg–20% CS composite extracts showed significantly reduced cell viability values throughout the incubation period that could be attributed to the inappropriately high concentrations of Mg ions due to the corrosion of Mg and the Mg matrix.⁵¹ However, the viability values of MC3T3 osteoblasts cultured in the pure Mg and Mg–20% CS composite extraction media after dilution by ratios of 1/4 to 1/128 were close to the viability values of the control after 1, 3 and 7 days. The results indicated that the Mg–20% CS composite material had a cytocompatibility similar to pure Mg. As the biocompatibility of pure Mg has been well acknowledged,^52,53 we can expect that the composite will possess an equally good biocompatibility.

Fig. 10 MTT assay of osteoblast-like (MC3T3-E1) cells cultured in Mg and Mg–20% CS extracts at different dilution ratios.

The ALP activities of MC3T3 cells cultured in pure Mg and the Mg–20% CS composite extraction media at dilution ratios of 1/4 to 1/128 for 10 days are presented in Fig. 11. It was found that, regardless of the dilution ratio, pure Mg extracts did not induce significant changes in the ALP expression of MC3T3 cells compared with the control group (p > 0.05). This is in line with the previous finding that the presence of Mg ions alone might not be adequate to stimulate the ALP expression of osteoblast cells.⁵⁴ In contrast, the influence of the Mg–20% CS composite extracts on the ALP activity was found to be dependent on the dilution ratio. At a dilution ratio of 1/4 composite extracts did not induce significant changes in the ALP expression of osteoblast cells compared with the control group. However, when the dilution ratio increased to 1/16 and 1/64, composite extracts significantly enhanced the ALP expression of the cells compared with the control group and with pure Mg extracts at the same dilution ratios (p < 0.05). However, when the dilution ratio further increased to 1/128, the ALP expression of the cells in the composite extract diminished to the level that was comparable to the control group, indicative of weakened stimulating effects of ions in the composite extract as their concentrations became too low.

Fig. 11 ALP activity of MC3T3-E1 cells in the presence of pure Mg and the Mg–20% CS composite extracts at different dilution ratios. The symbols * and # represent a significant difference compared with the control and pure Mg extract, respectively.

It is well known that ions released from degrading metallic or inorganic minerals can influence the osteogenic differentiation of osteoblast-like cells.^55,56 Thus, the differences in the ALP expression of osteoblast cells in contact with Mg and Mg–20% CS composite extracts may be due to their differing ionic compositions and concentrations. The Mg ion concentrations of the extracts used for the ALP assay could be calculated to range from 2.7 to 86.6 μg mL⁻¹ for pure Mg extracts and from 1.6 to 50.4 μg mL⁻¹ for the Mg–20% CS extracts, which clearly overlapped each other and could thus be considered to fall into the same dimension.⁵ Therefore, the presence of Si ions in the Mg–20% CS composite extracts must have played a crucial role in enhancing the ALP expression of osteoblast cells. Numerous reports have confirmed that Si ions in a certain concentration range are indeed able to stimulate the ALP expression of osteoblast cells.^19,56,57 However, it should be noted that, in the Mg–20% CS extract, Mg ions might have assisted Si ions in enhancing the ALP expression of osteoblast cells, as earlier studies on Mg-containing silicate bioceramics have suggested.^58,59 Further investigations with sophisticated material designs are needed to clarify the individual and combined stimulatory effects of Mg and Si ions. Nevertheless, as ALP is one of the most widely recognized markers of osteogenic differentiation that is one of the key steps to determine the success in the bone regeneration,^60,61 it could be firmly believed that the Mg–20% CS composite material will possess an enhanced ability to stimulate bone formation compared with pure Mg.

4. Conclusions

Mg-MMCs reinforced with CS bioceramic particles as the reinforcement phase were fabricated by means of SPS. CS networks were formed in the Mg matrix when the content of the reinforcement phase was more than 10%. The composite with 20% CS possessed the most compact structure. With an addition of CS bioceramic particles to Mg, the compressive strength of the latter could be enhanced by more than 30%. Furthermore, an addition of CS particles by 20% improved the corrosion resistance of Mg, which was mostly due to accelerated hydroxyapatite precipitation on the surface exposed to SBF. In addition, the Mg–20% CS composite exhibited a superior ability to stimulate the ALP expression of osteoblast-like cells compared to pure Mg, which was attributed to the release of Si ions from the CS phase. In summary, our study showed that CS could be an effective reinforcement phase to be added to Mg or Mg-based alloys to enhance their mechanical properties, corrosion resistance and biological performance. In other words, by incorporating CS particles into Mg or Mg-based alloys, improvements in corrosion resistance and bioactivity could simultaneously be achieved, thereby overcoming the major drawbacks of Mg and Mg-based alloys for orthopedic implant applications.

Acknowledgements

Financial support from the National Natural Science Foundation of China (Grant No. 81401529), Shanghai Pujiang Program (Grant No. 14PJ1409400) and the One-Hundred Talent Program of SIC-CAS (Grant No. Y36ZB1110G) is greatly acknowledged. This work was also partly supported by the External Cooperation Program of the Chinese Academy of Sciences (Grant No. GJHZ1211) and the Netherlands Organization for Health Research and Development (ZonMw) under the project 1163500004.

Notes and references

1. M. P. Staiger, A. M. Pietak, J. Huadmai and G. Dias, Biomaterials, 2006, 27, 1728–1734.
2. H. Wang, Z. M. Shi and K. Yang, Adv. Mater. Res., 2008, 32, 207–210.
3. R. Zeng, W. Dietzel, F. Witte, N. Hort and C. Blawert, Adv. Eng. Mater., 2008, 10, B3–B14.
4. G. Song, Corros. Sci., 2007, 49, 1696–1701.
5. S. Yoshizawa, A. Brown, A. Barchowsky and C. Sfeir, Acta Biomater., 2014, 10, 2834–2842.
6. D. Vojtech, H. Ciova and K. Volenec, Kovove Mater., 2006, 44, 211–223.
7. E. Zhang, L. Xu and K. Yang, in Beijing International Materials Week, June 25–30 2006, Beijing, China, p. 317.
8. F. Witte, F. Feyerabend, P. Maier and J. Fisher, et al., Biomaterials, 2007, 28, 2163–2174.
9. S. Z. Khalajabadi, M. R. Abdul Kadir, S. Izman and R. Ebrahimi-Kahrizsangi, Mater. Des., 2015, 88, 1223–1233.
10. T. Lei, W. Tang, S. Cai, F. Feng and N. Li, Corros. Sci., 2012, 54, 270–277.
11. M. Rahman, A. Adebisi and M. Maleque, Int. J. Automot. Mech. Eng., 2011, 4, 471–480.
12. X. Ye, M. Chen, M. Yang, J. Wei and D. Liu, J. Mater. Sci.: Mater. Med., 2010, 21, 1321–1328.
13. Y. Huang, D. Liu, L. Anguilano, C. You and M. Chen, Mater. Sci. Eng., C, 2015, 54, 120–132.
14. S. Kalita and H. Bhatt, Mater. Sci. Eng., C, 2007, 27, 837–848.
15. S. Liu, F. Jin, K. Lin, J. Lu, J. Sun, J. Chang, K. Dai and C. Fan, Biomed. Mater., 2013, 8, 025008.
16. H. Li and J. Chang, Acta Biomater., 2013, 9, 5379–5389.
17. C. Kao, T. Huang, Y. Chen, C. Hung, C. Lin and M. Shie, Mater. Sci. Eng., C, 2014, 43, 126–134.
18. S. Ni, J. Chang, L. Chou and W. Zhai, J. Biomed. Mater. Res., Part B, 2007, 80, 174–183.
19. H. Li, W. Zhai and J. Chang, J. Biomater. Appl., 2009, 24, 231–246.
20. S. Xu, K. Lin, Z. Wang, J. Chang, L. Wang, J. Lu and C. Ning, Biomaterials, 2008, 29, 2588–2596.
21. W. Muhammada, S. Sajuri, Y. Mutohd and Y. Miyashita, J. Alloys Compd., 2011, 509, 6021–6029.
22. L. Long, L. Chen and J. Chang, Ceram. Int., 2006, 32, 457–460.
23. J. Liu, X. P. Song, P. Pei and G. L. Chen, J. Alloys Compd., 2009, 486, 338–342.
24. T. Kokubo and H. Takadama, Biomaterials, 2006, 27, 2907–2915.
25. L. Yang and E. Zhang, Mater. Sci. Eng., C, 2009, 29, 1691–1696.
26. M. Zhao, P. Schmutz, S. Brunner, M. Liu, G. L. Song and A. Atrens, Corros. Sci., 2009, 51, 1277–1292.
27. T. Kokubo, J. Non-Cryst. Solids, 1990, 120, 138–151.
28. X. N. Gu, X. H. Xie, N. Li, Y. F. Zheng and L. Qin, Acta Biomater., 2012, 8, 2360–2374.
29. K. Bobe, E. Willbold, I. Morgenthal, O. Andersen, T. Studnitzky, J. Nellesen, W. Tillmann, C. Vogt, K. Vano and F. Witte, Acta Biomater., 2013, 9, 8611–8623.
30. A. H. Cory, T. C. Owen and J. A. Barltrop, Cancer Commun., 1991, 3, 207–212.
31. O. H. Lowry, N. R. Roberts, M. Wu, W. S. Hixton and E. J. Crawford, J. Biol. Chem., 1954, 207, 19–37.
32. K. Kondoh, H. Fukuda, J. Umeda, H. Imai, B. Fugetsu and M. Endo, Mater. Sci. Eng., A, 2010, 527, 4103–4108.
33. T. Aydogmus, Mater. Sci. Eng., A, 2015, 624, 261–270.
34. K. Lin, J. Chang, Y. Zeng and W. Qian, Mater. Lett., 2004, 58, 2109–2113.
35. I. Ibrahim, F. Mohamed and E. Lavernia, J. Mater. Sci., 1991, 26, 1137–1156.
36. M. Nai, J. Wei and M. Gupta, Mater. Des., 2014, 60, 490–495.
37. B. Chua, L. Lu and M. Lai, Compos. Struct., 1999, 47, 595–601.
38. Y. Su, Q. Ouyang, W. Zhang, Z. Li, Q. Guo, G. Fan and D. Zhang, Mater. Sci. Eng., A, 2014, 597, 359–369.
39. G. Ghosh and P. Saha, Mater. Des., 2011, 32, 139–145.
40. A. Feng and Y. Han, J. Alloys Compd., 2010, 504, 585–593.
41. H. Zhuang, Y. Han and A. Feng, Mater. Sci. Eng., C, 2008, 28, 1462–1466.
42. E. Aghion and Y. Perez, Mater. Charact., 2014, 96, 78–83.
43. G. L. Song and A. Atrens, Adv. Eng. Mater., 2003, 5, 837–858.
44. S. Ni and J. Chang, J. Biomater. Appl., 2009, 24, 139–158.
45. Y. Wang, M. Wei, J. Gao, J. Hu and Y. Zhang, Mater. Lett., 2008, 62, 2181–2184.
46. M. Razavi, M. Fathi and M. Meratian, Mater. Charact., 2010, 61, 1363–1370.
47. H. Takadama, H. Kim, T. Kokubo and T. Nakamura, Chem. Mater., 2001, 13, 1108–1113.
48. X. Wan, C. Chang, D. Mao, L. Jiang and M. Li, Mater. Sci. Eng., C, 2005, 25, 455–461.
49. S. Shadanbaz and G. Dias, Acta Biomater., 2012, 8, 20–30.
50. L. Xu, E. Zheng, D. Yin, S. Zeng and K. Yang, J. Mater. Sci.: Mater. Med., 2008, 19, 1017–1025.
51. A. Zadpoor, Mater. Sci. Eng., C, 2014, 35, 134–143.
52. X. Gu, Y. Zheng, Y. Cheng, S. Zhong and T. Xi, Biomaterials, 2009, 30, 484–498.
53. M. Staiger, A. Pietak, J. Huadmai and G. Dias, Biomaterials, 2006, 27, 1728–1734.
54. L. Qiao, J. Gao, Y. Wang, S. Wang, S. Wu and Y. Xue, Mater. Sci. Forum, 2007, 546–549, 459–462.
55. Y. Yun, Z. Dong, D. Yang, M. Schulz, V. Shanov, S. Yarmolenko, Z. Xu, P. Kumta and C. Sfeir, Mater. Sci. Eng., C, 2009, 29, 1814–1821.
56. Z. Sun, J. Wataha and C. Hanks, J. Biomed. Mater. Res., 1997, 34, 29–37.
57. A. Hoppe, N. Güldal and A. Boccaccini, Biomaterials, 2011, 32, 2757–2774.
58. S. Zou, D. Ireland, R. Brooks, N. Rushton and S. Best, J. Biomed. Mater. Res., Part B, 2008, 90, 123–130.
59. W. Zhai, H. Lu, C. Wu, L. Chen, X. Lin, K. Naoki, G. Chen and J. Chang, Acta Biomater., 2013, 9, 8004–8014.
60. H. Sun, C. Wu, K. Dai, J. Chang and T. Tang, Biomaterials, 2006, 27, 5651–5657.
61. J. Farley and D. Baylink, Metabolism, 1986, 35, 563–571.

---