The effect of hydroxyapatite reinforcement and preparation methods on the structure and mechanical properties of Mg-HA composites

2.1 Materials and processing

Magnesium powder with particle size between 50 and 250 μm and HA powder with particle size between 0.2 and 0.5 μm were used for preparation of composite material with 0, 2, 5, and 10 wt.% of HA. Materials were prepared by three different ways (Figure 1).

Figure 1:

Preparation of composite materials.

In the first way, magnesium powder and HA powder were mixed manually in hexane for 5 min. Hexane was selected to create doughy mass and prepare more a homogenous mixture. Mixed powders were pressed at 80 kN at room temperature to prepare green compacts 20 mm in diameter and 15 mm in height and then extruded at 400°C.

The second group of materials was prepared by milling the Mg and HA powders together in a planetary mill for 20 min with rotation speed of 400 rpm under the argon atmosphere. The weight ratio between powders and steel balls was set up to 1:70. Milled powders were pressed at 80 kN to green compacts and then extruded at 400°C. To characterize the effect of extrusion temperature on properties of prepared composites, pure magnesium and samples with 5 wt.% of HA were prepared using this method also at extrusion temperatures 300°C and 500°C.

As an alternative to the compaction by extrusion process, milled powder mixture with 5 wt.% of HA was processed by the SPS method. During this procedure, powder was pressed at 80 MPa and then heated up to 400°C for 5 min with a heating rate of 100°C/min. The preparation at 500°C and at same conditions was performed to estimate the effect on structure and mechanical properties of prepared composite.

2.2 Structure characterization

For structure observation, samples were grinded on emery papers (Buehler, Lake Bluff, IL, USA) and polished on diamond paste (Urdiamant s.r.o., Šumperk, Czech Republic) and colloidal silica (Urdiamant s.r.o., Šumperk, Czech Republic). Samples were etched in solution containing 10 ml of acetic acid, 4.2 g of picric acid, 10 ml of distilled water, and 70 ml of ethanol. The structure of powders and prepared composites was studied by light microscopy and scanning electron microscopy (Tescan Vega 3, TESCAN Brno, s.r.o., Brno, Czech Republic) equipped with energy dispersion spectrometry (AZtec, Oxford Instruments, Prague, Czech Republic). Phase composition was studied using X-ray diffraction (XRD) (X-Pert Pro, PANalytical GmbH, Kassel, Germany). The Rietveld analysis was performed to confirm the formation of fine-grained material during the preparation procedure. This analysis was implemented in HighScore Plus 3.0e (PANalytical, Almelo, The Netherlands). The split pseudo-Voigt analytical function was used to fit peak profiles. The structure models were taken from the Inorganic Crystal Structure Database (ICSD), and the atomic coordinates and B_iso parameters of all atoms were fixed for all phases.

2.3 Mechanical properties

Compressive tests on composite materials were carried out on LabTest 5.250SP1-VM machine (LABORTECH s.r.o., Opava, Czech Republic) at room temperature with strain rate of 0.001 s^-1 on cylindrical samples 6 mm in diameter and 8 mm high. Compressive yield strength (CYS), ultimate compressive strength (UCS), and total plastic deformation were determined from measurements. Cylindrical samples 5.5 mm in diameter and 35 mm in length were prepared for three-point bending tests. These measurements were performed with the same strain rate and with distance between two supporting points of 20 mm. Compressive and flexural mechanical properties were supplemented by Vickers hardness measurement at loading 5 kg (HV5).

3.1 Microstructure

The sizes of input powders were confirmed on the basis of the image analyses (Figure 2). Observation on transmission electron microscope (TEM) (Figure 2B) demonstrates that majority of the HA particles ranged from 0.05 to 0.2 μm. However, few particles reach almost 0.5 μm in diameter. Larger particles were predominantly spherical; however, few amounts of small particles of HA were characterized by blocks with sharp edges. Magnesium particles characterized by sharp edges were significantly coarser and ranged from 50 to 250 μm.

Figure 2:

Input powders: (A) Mg particles (50–250 μm), (B) HA particles (TEM).

The study of the structure of prepared composite materials revealed that specimens prepared by extrusion of HA and Mg powders mixed in hexane led to the creation of agglomerates of HA particles together with magnesium oxides. These agglomerates are oriented in rows parallel to the extrusion direction (Figure 3A). High amount of oxides was caused by oxidation during stirring of Mg with HA in hexane on air. In this case, greater amount of HA fraction causes the increase in amount and size of agglomerates in the structure.

Figure 3:

Mg-5HA extruded at 400°C with previous (A) mixing in hexane and (B) mixing in planetary mill. (C) Homogenously distributed HA particles in Mg after milling. (D) Structure of Mg-5HA after extrusion at 400°C (TEM).

Using the milling process, the mixed powders of nano-HA and Mg were characterized by homogenously distributed HA particles in the Mg matrix already after milling (Figure 3C). During the milling process, temperature rises and particles of Mg and HA are sintered and form homogeneous composite. The homogenous state was maintained also after the extrusion process (Figure 3B), and no specific rows of HA particles parallel to the extrusion direction were observed in the structure.

It is known that specimens are highly plastically deformed during the extrusion process. In combination with increased temperature, recrystallization process often takes place. In such case, small particles of HA may obstructs the growth of new grains [17]. As a consequence, prepared composites are fine-grained materials, and observed grain size does not exceed 1 μm (Figure 3D). This was also confirmed by XRD analyses. On Figure 4A and B, you can see continuous diffraction lines for phases observed in Mg-5HA composite prepared by extrusion at 400°C. Such effect is observable in the case that these phases are very fine and the grain size is commonly under 1 μm. On Figure 4C, there is a for-comparison image, which shows diffraction line of Mg powder used for composite preparation. It is clearly evident that diffraction lines are in this case partially discontinuously composed of points with different intensities, which points to the coarse-grained material. These results clearly confirm recrystallization processes during preparation of composites with a fine-grained final structure.

Figure 4:

XRD measurements: (A) Diffraction lines for Mg phase in Mg-5HA composite in cross-section, (B) diffraction lines for Mg phase in Mg-5HA composite in longitudinal section, (C) diffraction lines for Mg powder, and (D) XRD plots of studied materials.

Extruded materials are commonly characterized by specific texture. Therefore, Rietveld analysis was performed in both the cross- and longitudinal sections to indicate possible crystallographic textures. The split pseudo-Voigt analytical function was used to fit peak profiles. The structure models were taken from the ICSD, and the atomic coordinates and B_iso parameters of all atoms were fixed for all phases. At the first stage, seven parameters (zero shift, scale, unit cell parameters and U, W profile parameters for Mg phase) were refined to the final agreement factors R_p=0.148 and R_wp=0.206. The corresponding Rietveld plot showing the difference curve between measured and calculated data showed remarkable difference between measured and calculated intensity of reflection (010). Therefore, the March/Dollase parameter for preferred orientation (direction <010>) for the Mg phase was added into the Rietveld refinement. At the final stage, eight parameters were refined to the final agreement factors R_p=0.082 and R_wp=0.113, which shows significant improvement, so as the final Rietveld plot. The March/Dollase parameter was refined to the value 0.75(2), which is quite far from value 1 corresponding to crystallites randomly distributed and supports the existence of preferred orientation of crystallites oriented by <010> direction parallel to the sample normal base on the measurements from the cross-sectional cut. The existence of preferred orientation of crystallites oriented by <002> direction parallel to the sample normal was observed on the basis of the measurement from the longitudinal section of the sample. Although these results confirmed the presence of texture in material, it is worth mentioning that its strength is low. This is also evident from Figure 4A, where individual diffraction circles are characterized by only a slight difference in observed intensity, which corresponds only to the slight texture development during preparation of composites. Such behavior of prepared materials is favored to prevent possible differences in mechanical properties in various directions of loading.

One of the advantages of extrusion is friction of particles during process, which causes disruption of oxide shells and helps the particles to stick together. The same can be said about milling, which is another reason why particles of HA and Mg are properly sintered. In this case, combination of milling and subsequent extrusion leads to the homogeneous structure with uniform distribution of HA particles with no clusters or agglomerates (Figure 3B). Even though milling was performed under the argon atmosphere, there were oxides present in all samples. These oxides were confirmed by XRD (Figure 4D). Presence of oxide particles comes partially from oxides present in rough Mg powder before milling and also from oxygen containing argon gas (0.996%). Thankfully, a small amount of oxides in the structure also works as reinforcement and is considered not to have a negative effect on mechanical properties.

Properties of prepared composites can be affected by extrusion temperature. Therefore, different temperatures 300°C, 400°C, and 500°C were selected for material preparation. During extrusion at 300°C, the center of rods was significantly cracked (Figure 5A). On the contrary, application of higher temperatures 400°C and 500°C prevents cracking (Figure 5B and C). Such effect is caused by the lack of a slipping system in magnesium hexagonal close-packed (HCP) structure at low temperature and accumulation of stress at some areas during severe plastic deformation. This can lead finally to the formation of cracks in the structure. It is known that more slipping systems are activated during processing at higher temperatures, and as a consequence, plastic deformation occurs more easily. As a consequence, almost no cracks are observed in the samples extruded at 400°C and 500°C. On the contrary, there exists the danger of easier oxidation with increasing temperature. Unfortunately, a higher fraction of oxide particles was observed in the case of extrusion at 500°C (Figures 4D and 5C). This means that during extrusion, magnesium matrix is partially oxidized, and this effect is stronger at higher temperatures.

Figure 5:

Mg-5HA prepared by mixing in planetary mill and extruded at (A) 300°C, (B) 400°C, and (C) 500°C.

In Figure 6 you can see the samples prepared by the SPS method at 400°C and 500°C. Samples compacted at 400°C (Figure 6A) are characterized by significantly higher porosity, which is caused by a slower diffusion process during processing at this temperature. Temperature of 500°C is sufficient enough to gain an almost nonporous structure (Figure 6B).

Figure 6:

Mg-5HA prepared by mixing in planetary mill and (A) SPS at 400°C and (B) SPS at 500°C.

3.2 Mechanical properties

Figure 7A and B shows the compressive curves of composites prepared by two different procedures. In the first case (Figure 7A), magnesium and HA powders were homogenized in proprietary amount in a beaker with hexane. After sufficient homogenization of mixture, powders were pressed to green compacts. It is worth mentioning that pure Mg powder was pressed without preliminary mixing in hexane. It is evident that the addition of HA causes a decrease in CYS, but an improvement in UCS. The improvement of CYS and UCS is more significant with higher HA addition; however, the effect is not strong. All Mg-xHA (x=2, 5, 10 wt.%) are characterized by almost similar ultimate deformation under press and also decreased deformation compared to pure Mg. The main reason for such observations is connected with the presence of agglomerates of HA particles in the structure, which cause the decrease of CYS, UCS, and also deformation.

Figure 7:

Compressive curves of Mg-HA prepared by (A) mixing of powders in hexane, (B) mixing of powders in planetary mill, and subsequent extrusion at 400°C.

From Figure 7B, it is obvious that samples prepared after homogenization of powder mixture in planetary mill are characterized by improved mechanical properties, especially CYS compared to the mentioned procedure with hexane. Moreover, CYS and UCS are more significantly improved with higher volume fraction of HA in composite. Surprisingly, also compressive deformation is improved up to 5 wt.% of HA in composite. It seems that at higher amounts of HA, plasticity of materials is decreased as is obvious in the case of Mg-10HA. Compared to the previous procedure including the homogenization in hexane, these composites contained almost no agglomerates and HA particles were regularly distributed in the Mg matrix. This has main impact on the observed mechanical properties. It is worth mentioning that measured mechanical properties were characterized by good repeatability for different samples. This cannot be said about the previous procedure, where strong deviations were observed between identically prepared samples due to the inhomogeneity in the structure.

Figures 7B and 8A and B present the measured compressive curves of composites prepared from the same green compact but using a different extrusion temperature. Magnesium powder extruded at the same temperature is also added for comparison. Due to the fact that the addition of 10 wt.% of HA has a very small effect on CYS and UCS compared to Mg-5HA and, moreover, plasticity was significantly decreased, only Mg-2HA and Mg-5HA materials were considered for comparison of the effect of extrusion temperature on mechanical properties. Samples extruded at 300°C were characterized by a decreased CYS and especially UCS, compared to extrusion at 400°C. The reason for such behavior is connected with the presence of cracks in the structure, which has already been mentioned in the part deals with structure observations. As a consequence, these specimens cracked during tests in the center of rods. The presence of these cracks in the structure obviously affects also CYS, although higher values of CYS at higher extrusion temperatures could by potentially connected with the presence of oxide MgO particles. These particles were observed occasionally in the structure of samples extruded at 400°C and more regularly in samples extruded at 500°C; however, its amount was very small and only slightly measurable by analytical method, such as XRD (Figure 4D). Therefore, authors assume that the main effect on yield stress is connected with the volume fraction and distribution of HA and also the grain size of magnesium matrix. It is evident on Figure 8B that CYS is slightly decreased and UCS increased for materials extruded at 500°C compared to the same materials prepared by extrusion at 400°C. At 500°C, there is a higher tendency to the coarsening of the material [16], [29]. This can cause the observed decrease in yield points. Simultaneously, increase in UCS is observed because of the work hardening during plastic deformation, which is much easier in larger grains. It has been shown recently that increase in the extrusion temperature can cause the increase of mechanical properties such as yield strength, ultimate strength and also plasticity [29]. This is predominantly connected with improved particle distribution in matrix and also decreased danger of the formation of voids in materials especially on places of cracking of reinforcement particles. In the present work, HA particles were regularly distributed in magnesium matrix after the milling procedure, and therefore, no significant differences in particle distribution were observed after the extrusion process at different temperatures. In addition, because of the small character of HA particles, no significant cracking of them during extrusion was observed. The effect of extrusion temperature on mechanical properties of Mg and Mg-5HA is summarized in Figure 9A.

Figure 8:

Compressive curves of Mg-HA prepared by mixing of powders in planetary mill and subsequent extrusion at (A) 300°C and (B) 500°C.

Figure 9:

Compressive tests of studied materials: (A) The effect of extrusion temperature on the compressive behavior of pure Mg and Mg-5Ha prepared by mixing of powders in planetary mill and subsequent extrusion at 300°C, 400°C, and 500°C. (B) Compressive behavior of Mg-5HA prepared by self-sustained high-temperature synthesis (SHS) in comparison with extruded specimens.

SPS was selected as an alternative method to extrusion process. In this case, two temperatures, 400°C and 500°C, were used. Lower mechanical properties of the sample prepared by SPS at 400°C (Figure 9B) are connected with higher porosity (Figure 6A). During the procedure, compaction was insufficient and many pores remain in the structure. Higher temperature (500°C) eliminates the porosity and causes significant improvement of the mechanical properties. Obtained values of CYS for materials prepared by SPS at 500°C are nearly similar in comparison with composite prepared by extrusion at the same temperature. The slight difference is observable in the plastic behavior of both prepared composites. In the case of extruded Mg-5HA, deformation strengthening is more significant. This relates to the specific deformation texture of extruded Mg materials, which causes easier deformation during compressive testing parallel to the extrusion direction [30], [31], [32], [33]. On the contrary, there is no significant plastic deformation during SPS, and therefore, specimens after the process are not characterized by specific texture. Consequently, there is not any preferential testing direction for compressive tests in which plastic deformation is more favored. The measured values of mechanical properties of all studied materials are summarized in Table 2.

Different methods have been recently used for the preparation of magnesium based composites, particularly Mg-HA [2], [14], [16], [21], [23], [24], [34]. In [16], authors studied the effect of conditions on the final properties of Mg-1HA composite prepared from nanopowders using high-frequency induction. At proper conditions, Mg-1HA was characterized by CYS about 120 MPa and UCS nearly 200 MPa. It is worth mentioning that in present work, Mg-2HA prepared by combination of milling and extrusion was characterized by significantly better compressive mechanical properties with CYS and UCS about 260 and 350 MPa, respectively. Although, authors in [16] prepared the materials with very fine-grained structure with the grain size about 40 nm, grain size in the present study did not exceed 0.5 μm. Therefore, authors assume that there is no significant reason for the preparation of such composite materials from nanopowders of Mg to reach satisfactory mechanical properties, but it is sufficient to homogenously incorporate nanoparticles of reinforcement into the Mg matrix, which can slow down grain growth during recrystallization and also cause significant particle reinforcement.

In [17], authors prepared Mg-HA composites using the classical casting method after sufficient homogenization of melt and subsequent extrusion of ingots. However, obtained compressive properties of Mg-5HA were significantly lower (about 100 and 300 MPa for CYS and UCS, respectively) compared to results presented in this paper. The main advantage can be found in the application of very fine regularly shaped particles of HA in the magnesium matrix and also mentioned uniform distribution of particles.

It is known that compressive tests are less susceptible to structural disorders compared to the tensile or flexural testing. Therefore, flexural curves were measured on Mg-xHA (x=0, 2, 5, 10 wt.%) (Figure 10) to confirm the differences in mechanical properties gained on the basis of compressive tests. Also in this case, flexural strength is improved with higher HA volume fraction; however, ultimate flexural strain is improved only up to 5 wt.% of HA addition. The addition of 10 wt.% of HA to Mg leads to the deterioration of plasticity of Mg-HA composites. From the mechanical point of view, authors consider 10 wt.% and higher fraction of HA as excessive amounts, which cause undesirable deterioration of deformation behavior accompanied by insignificant increase of strength, which is not required because the strength of suitably prepared Mg-5HA already reached sufficient values (Figures 7B and 8B).

Figure 10:

The effect of HA volume fraction on flexural properties of Mg-HA prepared by mixing of powders in planetary mill and subsequent extrusion at 400°C.