Materials Science and Engineering C 35 (2014) 21–28

Contents lists available at ScienceDirect

Materials Science and Engineering C

j ourna l homepage: www.e lsev ie r .com/ locate /msec

Effect of sintering conditions on the microstructural and mechanicalcharacteristics of porous magnesium materials prepared bypowder metallurgy

Jaroslav Čapek ⁎, Dalibor VojtěchDepartment of Metals and Corrosion Engineering, Institute of Chemical Technology, Prague, Technická 5, 166 28 Prague 6, Czech Republic

⁎ Corresponding author. Tel.: +420 220444055.E-mail address: Jaroslav.Capek@vscht.cz (J. Čapek).

0928-4931/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.msec.2013.10.014

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 May 2013Received in revised form 14 October 2013Accepted 21 October 2013Available online 1 November 2013

Keywords:Porous magnesiumSintering conditionsPowder metallurgyMechanical properties

There has recently been an increased demand for porous magnesium materials in many applications, especiallyin themedicalfield. Powdermetallurgy appears to be a promising approach for the preparation of suchmaterials.Manyworks have dealtwith the preparation of porousmagnesium;however, the effect of sintering conditions onmaterial properties has rarely been investigated. In this work, we investigated porous magnesium samples thatwere prepared by powder metallurgy using ammonium bicarbonate spacer particles. The effects of the purity ofthe argon atmosphere and sintering time on themicrostructure (SEM, EDX and XRD) andmechanical behaviour(universal loadingmachine and Vickers hardness tester) of porousmagnesiumwere studied. The porosities of theprepared samples ranged from 24 to 29 vol.% depending on the sintering conditions. The purity of atmosphereplayed a significant role when the sintering time exceeded 6 h. Under a gettered argon atmosphere, a prolongedsintering time enhanced diffusion connections between magnesium particles and improved the mechanicalproperties of the samples, whereas under a technical argon atmosphere, oxidation at the particle surfaces causeddeterioration in the mechanical properties of the samples. These results suggest that a refined atmosphere isrequired to improve the mechanical properties of porous magnesium.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Magnesiumandmagnesiumalloys have recently been studied for useinmany applications, such as in the automotive and aerospace industriesbecause of their low densities and good mechanical properties [1].Many successful studies have been performed on biocompatible and bio-degradable magnesium-based materials, which are considered suitablematerials for orthopaedic applications, such as for nails, screws, splints,etc. [2–6]. For some applications, porous implants, called scaffolds, arerequired because they possessmechanical properties, such as themodu-lus of elasticity, that are relatively similar to those found in natural bonetissue [7–14]. The mechanical biocompatibility is important to aid in theremodelling of new tissue [7–14]. Moreover, an interconnected porousstructure allows the transport of body fluids to damaged or woundedtissue and supports the incorporation of new tissue in the implant[7,15,16]. Therefore, porous magnesium materials and their preparationmethods have been extensively studied in recent years [7–14].

Many methods have been developed for fabricating porous metallicmaterials [17]. However, because biomaterials should contain intercon-nected pores [12] and should not be contaminated with harmful impu-rities [4], only a few of these methods are used for the production ofbiomaterials. In the available literature, there are five “non-machining”

ghts reserved.

approaches that have been reported for the fabrication of porousmagnesiummaterials: (1) injection of an inert gas into amelt [4], (2) di-rectional solidification of the metal–gas eutectic (the GASAR process)[9], (3) plaster casting [4], (4) negative salt pattern moulding [12] and(5) powder metallurgical techniques [11,13,14]. However, the firsttwo methods mentioned above do not necessarily produce open-cellstructures, and the following twomethodsmay contaminate or corrodethe final product during pattern removal [18]. Suitable modifications topowder metallurgy (PM), for example, using space-holder particles,allow the fabrication of materials with interconnected pores. Thismodification consists of preparing a green compact that contains apowdered mixture of magnesium and a space-holder material, whichis later removed by leaching or thermal decomposition. Subsequently,the porous green body is sintered at high temperatures [8,10,13,14].In principle, any solidmattermay be used as the space-holdermaterial;however, in practice, this selection is limited because the spacermaterial has to be removed without contaminating the product. In theliterature, urea and ammonium bicarbonate have been successfullyused as spacer materials in the preparation of PM porous magnesium[8,11,13,14,19]. Hao et al. [8] removed the urea by leaching thematerialin a NaOH-water solution, whereas other authors removed the space-holder particles by thermal decomposition [10,11,13,14]. The majorityof these authors used urea as the spacer material, even though itscomplete decomposition occurs at temperatures above the meltingpoint of magnesium. Because urea only partially decomposes at lower

Fig. 1. SEM micrographs (SE detector) of sample cross sections prepared under the following sintering conditions: (a) technical argon for 6 h, (b) technical argon for 24 h, (c) getteredargon for 6 h and (d) gettered argon for 24 h.

22 J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

temperatures, contaminationmay occur [20]. For example, Zhuang et al.[13] found traces of carbon and oxygen on porewalls, which they attrib-uted to contamination by urea residues. Ammonium bicarbonate, whichdecomposes at significantly lower temperatures (36 °C–60 °C), appearsto be a more suitable spacer material for the preparation of PM porousmagnesium. In addition, leaching the spacer material has the disadvan-tage of possibly hermetically enclosing some of the spacer particles inthe magnesium matrix, which prevents their dissolution. Moreover,water-based dissolving agents are corrosive to magnesium [21].

The mechanical properties of PM porous magnesium are influencednot only by the total porosity but also by the pore size, distribution andshape and the connection between magnesium particles [8,10,13].These structural characteristics can be adjusted by selecting the optimalsintering time and temperature, compacting pressure and size, shapeand volume ratio of the starting material powders [8,11,13,19,22].Grain coarsening within each magnesium particle, which is expectedduring the sintering process, may also affect the mechanical propertiesof the final material [23,24].

In our previous work [19], we showed that porous magnesium can,in principle, be successfully prepared by PMusing ammoniumbicarbon-ate as the space-holder material. Although several works on the prepa-ration of porous magnesium have been reported, to the best of ourknowledge, no systematic study on the influence of the processing

Table 1The influence of sintering conditions on sample porosity (in vol.%).

Atmosphere/sintering time 0 h 3 h 6 h 12 h 24 h

Technical argon 29 ± 2 28 ± 1 28 ± 2 29 ± 3 31 ± 4Gettered argon 29 ± 2 28 ± 1 27 ± 2 25 ± 3 24 ± 2

parameters, namely, the sintering time and atmosphere purity, on themicrostructure and mechanical properties of porous magnesium isavailable. Therefore, our study focused on sintering kinetics and theeffect of the sintering atmosphere on the important characteristics ofPM porous magnesium.

Fig. 2. XRD patterns of samples sintered for 6 h.

Fig. 3. XRD patterns of samples sintered for 24 h.

23J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

2. Experimental

A purchased Mg powder (purity of 99.6 wt.%, mesh −100 + 200,Alfa Aesar) and NH4HCO3 powder (p.a. purity, 250–500 μm) wereused as starting materials. An Mg/NH4HCO3 volume ratio of 90:10 wasused because, according to our previous work [19], this ratio imparts agood combination of strength and corrosion resistance to the resultingmaterial. Materials prepared using this ratio have been shown to pos-sess better mechanical properties that are more comparable to naturalbone tissue than porous biomaterials, which have recently been usedfor medical applications [19]. To avoid segregation, the powders wereintensively blended with 30 vol.% hexane for 30 min. Subsequently,pre-weighed mixtures were pressed into cylindrical green compacts(10 mm in diameter and 30 mm in length) at a pressure of 265 MPausing a LabTest 5.250SP1-VM universal loading machine. The greencompacts were then subjected to a two-step procedure. First, the

Fig. 4. SEMmicrographs (BSE detector) and elementalmaps determined by EDX of samples sintbar 50 μm.

compactswere annealed for 4 h at 130 °C in amuffle furnace in air. Dur-ing this step, the decomposition of ammonium bicarbonate and theevaporation of hexane occurred. Afterwards, sintering was performedat 550 °C in a tube furnace. The green compacts were sintered for 0, 3,6, 12 and 24 h. The sintering process was performed under two typesof flowing atmosphere at a flow rate of 0.1 l/min: (1) argon with tech-nical purity (99.996 vol.%) and (2) argon purified by a 55-mm-thicklayer of Mg chips (250–500 μm in size), which was placed around thesintered material and acted as a getter. After sintering, the average ma-terial porosity was determined according to Eq. (1) [8,13]:

P ¼ 1–ρ=ρMg

� �� 100%; ð1Þ

where P is the porosity, ρ is the density of the porous sample (calculatedfrom the dimensions andweight) and ρMg is the density of puremagne-sium (ρMg = 1738 kg/m3). This approach for determining porositywasused because it describes the volumetric 3D porosity more preciselythan analysing 2D micrographs.

The material mechanical properties were characterised by flexural,compression and Vickers hardness tests that were performed at roomtemperature. Samples 26 mm and 15 mm in length were used for flex-ural and compression testing, respectively. Five samples of each serieswere used for the flexural tests, and three samples of each series wereused for the compression tests. The average hardness was calculatedfrom ten values. Standard deviations for all tests were calculated andare shown as error bars in the figure plots. The deformation rates duringthe flexural and compression tests were 0.5 mm/min and 1 mm/min,respectively. A LabTest 5.250SP1-VM universal loading machine wasused for these tests. After flexural testing, the fracture surfaces wereobserved by a TESCAN VEGA-3 LMU scanning electron microscope(SEM). Metallographic cross sections were also prepared and examinedby SEM. Afterwards, themetallographic cross sectionswere etched (2 gof picric acid, 10 ml of 99% acetic acid, 10 ml of water and 70 ml of eth-anol), and the microstructures of the powder particles were observedusing an Olympus PME 3 light metallographic microscope. The phasecomposition and elemental distributions were examined by X-ray dif-fraction using a PANalytical X'Pert PRO X-ray diffractometer equippedwith a Cu anode (XRD) and a TESCAN VEGA-3 LMU SEM equippedwith an Oxford Instruments INCA 350 EDX analyser (SEM-EDX). TheVickers hardnesswasmeasured for each serieswith a load of 3 kg (HV 3).

ered 24 h under (a) technical argon atmosphere and (b) gettered argon atmosphere. Scale

Fig. 5.Microstructure of powder particles during the material fabrication process: (a) the initial powder, (b) compacted powder after annealing at 130 °C for 4 h, (c) compacted powderafter annealing at 130 °C for 4 h and sintering at 550 °C for 3 h under technical argon and (d) compacted powder after annealing at 130 °C for 4 h and sintering at 550 °C for 24 h undertechnical argon.

24 J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

3. Results and discussion

3.1. Microstructure and porosity

The microstructures of samples sintered under two types of atmo-spheres for different periods of time are shown in Fig. 1.

In the magnesium matrix, two types of pores were observed:(1) pores approximately 250–500 μm in size that were formed fromthe decomposition of spacer particles (“Type I” in Fig. 1) and (2) smallerpores that most likely originated from imperfect compaction duringpressing and the expansion of trapped gas during sintering (“Type II”in Fig. 1). No significant difference in microstructure was observed for

Fig. 6. Flexural stress–strain curves of selected samples.

samples sintered for 6 h under the different atmospheres. A prolongedsintering time caused annihilation of “Type II” pores under the getteredatmosphere, whereas under the technical argon atmosphere, theamount of these pores did not significantly change. The sample porositywas measured, and the effects of the sintering conditions on porosityare shown in Table 1. This table indicates that samples sintered underthe gettered argon atmosphere became more compact with longersintering times, i.e., contained less pores. In contrast, the technicalargon atmosphere produced materials whose porosities appear almostthe same.

Fig. 7. Ultimate flexural strength of prepared samples versus sintering time.

Fig. 8. Fracture surfaces of samples after flexural testing under the following sintering conditions: (a) technical argon for 6 h - overview, (b) technical argon for 6 h - detail, (c) technicalargon for 24 h and (d) gettered argon for 24 h. The white arrow in Fig. 8c denotes oxide particles. In Fig. 8d, the arrow indicates traces of plastic deformation.

25J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

This phenomenon suggests that sintering under the gettered argonatmosphere enhanced the diffusion connection between themagnesiumparticles, which led to annihilation of some “Type II” pores (Fig. 1). Incontrast, the use of technical argon caused a slight increase in total poros-ity. As shown below, porosity is directly related to the presence of oxidesin the material.

Fig. 9. Compressive stress–strain curves of selected samples.

The XRD patterns of the samples sintered under various conditionsare shown in Figs. 2 and 3.

The XRD patterns indicate that there is no significant difference inthe amount ofmagnesium oxide in the samples sintered for 6 h, regard-less of the atmosphere type (Fig. 2). In this figure, the peaks that areassigned toMgO are similar for both samples sintered for 6 h. However,

Fig. 10. The compressive yield strength versus sintering time of prepared samples.

Fig. 11. The ultimate compressive strength versus sintering time of prepared samples.

26 J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

samples thatwere sintered for 24 h in technical argon (Fig. 3) containedsignificantly higher amounts of MgO than those sintered in getteredargon. No traces of ammonium bicarbonate or the reaction products ofmagnesium and ammonium bicarbonate were found by the XRDanalysis.

To determine the difference between the samples sintered for 24 h,EDX analysis was performed. The results are shown in Fig. 4. The X-rayelementalmaps show the presence ofMg particles surrounded by oxidelayers. Comparing Fig. 4a and b reveals that more extensive oxidationoccurs at the grain boundaries during sintering under the technicalargon atmosphere than under the gettered atmosphere. This difference,which is observed after sintering for more than 6 h, is in good agree-ment with the results acquired from XRD analysis (Fig. 3). Traces ofoxygen and water are present in technical argon and cause surfaceoxidation, which weakens the diffusion connection between themagnesium particles. Thus, annihilation of the smaller “Type II” poresis slowed during longer sintering periods (Table 1). Moreover, oxideson the sample and particle surface may spall off and increase the mea-sured porosity. In contrast, a lower extent of oxidation was observedunder the gettered atmosphere, which supports the annihilation ofsome “Type II” pores because of the diffusive processes of the magne-sium particles, as shown in Table 1.

The presence of nitrogen and carbon was also investigated ascontamination by either ammonium bicarbonate or its decompositionproducts. No traces of these elements were found; therefore, theirdistribution is not shown in the XRD elemental maps.

Grain coarsening within magnesium particles during sinteringwas also investigated because it may have an impact on sampleproperties. Fig. 5 shows detailed views of themicrostructure of individ-ual powder particles.

The initial powder (Fig. 5a) possesses a relativelyfinemicrostructure(grain size up to 13 μm). After compaction and annealing at 130 °C for4 h (Fig. 5b), the grain size slightly enlarged; however, the microstruc-ture remained fine (approximate grain size of 15 μm). The grain sizesignificantly increased after sintering at 550 °C for 3 h in technicalargon (approximate grain size of 45 μm) (Fig. 5c). Sintering for longerperiods did not cause any further significant grain growth. The averagegrain size after sintering for 24 h at 550 °C was approximately 48 μm(Fig. 5d). The type of atmosphere did not influence grain growth duringsintering.

Fig. 12. The HV 3 versus sintering time of prepared samples.

3.2. Mechanical properties

The flexural stress–strain curves of selected samples are shown inFig. 6.

Samples sintered in technical argon had a significantly lower modu-lus of elasticity than samples sintered under the gettered atmosphere.This result can be explained by the increased amount of oxygen at theinterface between powder particles, which results in decreased cohe-siveness between the particles. Oxides also impact flexural strength,which will be shown and discussed below.

The flexural stress–strain curves of all the materials contained onlyelastic regions; no regions of macroscopic plastic deformation wereobserved (Fig. 6). The ultimate flexural strength (UFS) as a function ofsintering time is plotted in Fig. 7.

The ultimate flexural strength increasedwith longer sintering times,up to 6 h, independent of the atmosphere. The samples attainedapproximately the same UFS, 12 MPa, after 6 h of sintering underboth atmospheres. For longer sintering times under the gettered argonatmosphere, theUFS increased to approximately 15 MPa. The differencebetween samples sintered for 12 and 24 h under the gettered argonatmosphere is negligible. In contrast, prolonging the sintering periodto more than 6 h under technical argon had a completely oppositeeffect. The UFS rapidly decreased to 5 MPa after 12 h of sintering andcontinued to decrease to approximately 3 MPa after 24 h of sintering.

A study of the fracture surfaces helped to explain this behaviour. Thefracture surfaces of samples after flexural testing are illustrated in Fig. 8.For all the samples, the fracture surfaces were similar (Fig. 8a); howev-er, slight differences were observed at highermagnifications. Generally,the fracture surfaces were brittle, but traces of plastic deformationwereobserved, especially in samples sintered under gettered argon for longertime periods (12 and 24 h) (Fig. 8d). The fracture surfaces were nearlyidentical and oxide free after sintering for up to 6 h, independent ofthe atmosphere (Fig. 8b). After longer sintering times, some differenceswere observed between the fracture surfaces of samples preparedunder the different atmospheres. Samples sintered in gettered argonhad nearly oxide-free fracture surfaces with features of plastic deforma-tion (Fig. 8d), whereas a large amount of oxide was present on thefracture surfaces of samples sintered in technical argon (Fig. 8c). Thisobservation confirms the XRD and EDX results that were mentionedabove (Figs. 2–4).

Table 2Mechanical properties of porous biomaterials.

Material Porosity (vol.%) Pore size (μm) UFS (MPa) CYS (MPa) UCS (MPa) Reference

Natural bone — — 2–150 — 2–180 [13]Porous Mg 29–31 250–500 3–15 13–53 20–70 This studyPorous Mg 23–38 250–500 4–17 — — [19]Porous Mg 14–44 250–500 2–5 — — [22]Porous Mg 36–55 200–400 14–27 — 15–31 [13]Porous Mg 52–70 ~1250 — — 4–14 [8]Porous Mg 50 200–500 — — 2 [14]Porous Mg 28 170 — — 24 [7]Porous Mg 35–55 100–400 — — 12–17 [11]Porous Ti 78 200–500 — — 35 [14]Porous hydroxyapatite 50–77 200–400 2–7 — 1–17 [13]Porous hydroxyapatite — 366–444 — — 30 [25]Porous composite (poly-L-lactide + 20–50 wt.% of bioglass) 77–88 ~100 1–4 — ~0.4 [13]Porous polycaprolactone 48–77 — — 2–3 — [25]Porous polycaprolactone 37–55 — — — 2–3 [25]Porous polycaprolactone 55–56 — 2–3 [26]Porous polylactide-co-glycolide 31 116 — — 0.5 [27]Porous composite of polylactide-co-glycolide and 45S5 bioglass 43 89 — — 0.4 [27]

27J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

Fig. 9 shows the compressive stress–strain curves of selected samples.From the figure, the trends of the compressive and flexural behaviour ofthe material are similar, with the exception of the effect of atmosphericcondition on the compressivemodulus of elasticity,which is not as strong.

The compressive yield strength (CYS) and ultimate compressivestrength (UCS) as functions of the sintering time are plotted in Figs. 10and 11.

Under the gettered argon atmosphere, both the yield strength (CYS)and the ultimate strength (UCS) increased with longer sintering times.The maximum CYS and UCS values after sintering for 24 h are 53 and69 MPa, respectively. These high strength values result from a lowamount of oxide, decreased porosity and good contact between theparticles in the samples (Figs. 2–4, Table 1). In contrast, for samplessintered under technical argon, the maximum CYS and UCS wereattained after 6 h of sintering. Prolonging the sintering period underthis atmosphere led to a significant decrease in compressive properties.Thisfinding can be attributed to the extended oxidation that occurs dur-ing sintering, which weakens the diffusion bonds between particles. Aninteresting observation in Figs. 10 and 11 is that the CYS of samplessintered in gettered argon increased almost linearly with sinteringtime, whereas the UCS increased for up to 6 h of sintering and then in-creased significantly more slowly. These results indicate that the “TypeII” pores, which decrease in quantity during long sintering periods(Table 1), influence the yield strength rather than the ultimate strength.The ultimate compressive strength is probably more influenced by thelarger “Type I” pores, whose amount is almost constant for all samples.The reason may be that the large pores concentrate the stress in theirvicinity and initiate the crack nucleation and subsequent collapse ofthe porous structure. On the other hand, the small “Type II” pores con-centrate the stress to a smaller extent and plastic deformation takesplace in their surroundings. Therefore, their partial vanishing duringlonger sintering times seems to increase the stress needed for the plasticstrain to begin.

Fig. 12 reveals the effect of sintering time on sample hardness.Samples sintered in gettered argon possessed a higher hardness

than those sintered in technical argon. The effect of sintering time onhardness was small for samples sintered in technical argon. However,prolonging the sintering period in the gettered argon led to an increasein hardness, with a peak hardness occurring in samples sintered for12 h. Sintering up to 24 h slightly decreased sample hardness. Thisresult may be caused by oxidation beginning at the grain boundariesand decreases the cohesiveness of the magnesium particles.

A comparison of our results with previously reported results is diffi-cult because of differences in the porosities and pore sizes of the studiedmaterials. Nevertheless, we conducted a small review that summarisesthe mechanical properties of natural bone tissue and porous materials

that are considered suitable biomaterials. The results of this researchare summarised in Table 2.

From this table, it is clear that porous magnesium-based metallicmaterials exhibit mechanical properties similar to those of naturalbone tissue. However, non-metallic porous materials (especiallypolymer-based materials) have weaker mechanical properties com-pared to natural bone tissue. It is also important to note that the mate-rials prepared in our study have higher UFS and UCS values than porousmagnesium materials that have been prepared in the majority of otherworks and non-metallic porous materials. Moreover, the materials pre-pared in this work possess mechanical properties that are comparableto natural bone tissue.

4. Conclusions

Magnesium samples with porosities of 29–31 vol.% were preparedby powder metallurgy under different sintering conditions. When com-pared to non-metallic biomaterials, the samples exhibited enhancedmechanical properties that were similar to human bone tissue. Aftersintering for up to 6 h, no significant effect of atmosphere purity onsample microstructure or mechanical behaviour was observed. Underthe gettered argon atmosphere, longer sintering times decreased poros-ity but enhanced the mechanical properties of the sample. Under thetechnical argon atmosphere, the opposite trend was observed, whichcan be attributed to oxidation of the powder surface. These resultssuggest that when sintering is longer than 6 h, the purity of the argonatmosphere plays an important role in determining the mechanicalproperties of PM magnesium.

Acknowledgements

The authors would like to thank the Czech Science Foundation(project no. P108/12/G043) for supporting this research.

References

[1] F.H. Froes, Mater. Sci. Eng. A 184 (1994) 119–133.[2] N. Li, Y.F. Zheng, J. Mater. Sci. Technol. 29 (2013) 489–502.[3] H. Waizy, J.M. Seitz, J. Reifenrath, A. Weizbauer, F.W. Bach, A. Meyer-Lindenberg, B.

Denkena, H. Windhagen, J. Mater. Sci. 48 (2013) 39–50.[4] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006) 1728–1734.[5] S. Virtanen, Mater. Sci. Eng. B 176 (2011) 1600–1608.[6] M. Hradilova, J. Kubasek, P. Lejcek, L. Tanger, Metal 2011: 20th Anniversary Interna-

tional Conference on Metallurgy and Materials, 2011, pp. 920–924, (full text avail-able on http://www.metal2013.com/files/proceedings/metal_11/lists/papers/978.pdf, 31.7.2013).

[7] X.N. Gu, W.R. Zhou, Y.F. Zheng, Y. Liu, Y.X. Li, Mater. Lett. 64 (2010) 1871–1874.[8] G. Hao, F. Han, W. Li, J. Porous. Mater. 16 (2009) 251–256.

28 J. Čapek, D. Vojtěch / Materials Science and Engineering C 35 (2014) 21–28

[9] Y. Liu, Y.X. Li, J. Wan, H.W. Zhang, Mater. Sci. Eng. A 402 (2005) 47–54.[10] C.E. Wen, Y. Yamada, K. Shimojima, Y. Chino, H. Hosokawa, M. Mabuchi, Mater. Sci.

Forum 419–4 (2003) 1001–1006.[11] C.E. Wen, Y. Yamada, K. Shimojima, Y. Chino, H. Hosokawa, M. Mabuchi, Mater. Lett.

58 (2004) 357–360.[12] F. Witte, H. Ulrich, M. Rudert, E. Willbold, J. Biomed. Mater. Res. Part A 81A (2007)

748–756.[13] H.Y. Zhuang, Y. Han, A.L. Feng, Mater. Sci. Eng. C 28 (2008) 1462–1466.[14] C.E. Wen, M. Mabuchi, Y. Yamada, K. Shimojima, Y. Chino, T. Asahina, Scripta Mater.

45 (2001) 1147–1153.[15] A.H. Yusop, A.A. Bakir, N.A. Shaharom, M.R. Abdul Kadir, H. Hermawan, Int. J.

Biomater. 2012 (2012) 10.[16] A. Biswas, Acta Mater. 53 (2005) 1415–1425.[17] G.J. Davies, S. Zhen, J. Mater. Sci. 18 (1983) 1899–1911.[18] G. Song, A. Atrens, Adv. Eng. Mater. 5 (2003) 837–858.

[19] J. Čapek, D. Vojtěch, Mater. Sci. Eng. C 33 (2013) 564–569.[20] P.A. Schaber, J. Colson, S. Higgins, D. Thielen, B. Anspach, J. Brauer, Thermochim. Acta

424 (2004) 131–142.[21] F. Witte, N. Hort, C. Vogt, S. Cohen, K.U. Kainer, R. Willumeit, F. Feyerabend, Curr.

Opin. Solid State Mater. Sci. 12 (2008) 63–72.[22] J. Capek, D. Vojtech, Proceedings of the 20th Anniversary International Conference

on Metallurgy and Materials, 2011, pp. 903–908, (full text available on http://www.metal2013.com/files/proceedings/metal_11/lists/papers/728.pdf, 31.7.2013).

[23] A. Yamashita, Z. Horita, T.G. Langdon, Mater. Sci. Eng. A 300 (2001) 142–147.[24] P. Donelan, Mater. Sci. Technol. 16 (2000) 261–269.[25] S.J. Hollister, Nat. Mater. 5 (2006) 518–524.[26] D.W. Hutmacher, T. Schantz, I. Zein, K.W. Ng, S.H. Teoh, K.C. Tan, J. Biomed. Mater.

Res. 55 (2001) 203–216.[27] H.H. Lu, S.F. El-Amin, K.D. Scott, C.T. Laurencin, J. Biomed. Mater. Res. Part A 64A

(2003) 465–474.