- Email: [email protected]
Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds
Journal of Alloys and Compounds xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds Yeon-wook Kim* Department of Advanced Materials Engineering, Keimyung University, Daegu, 42403, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 9 October 2019 Received in revised form 28 November 2019 Accepted 29 November 2019 Available online xxx
A simple and effective way to fabricate highly porous scaffolds with controlled porosity and pore size is demonstrated in this study. Ti50Ni49.7Mo0.3 shape memory alloy fibers were prepared by a melt overflow process. The scaffolds with high porosity of 60~80% and large pores of 100~600 mm in size were fabricated by sintering the as-solidified fibers. The martensitic transformation behavior of the porous scaffold was determined and it was found that superelasticity could be achieved at human body temperature. An effect of porosity on mechanical properties of porous scaffolds was investigated using compressive tests. As the porosity increases from 60% to 80%, elastic modulus and compressive strength decrease from 1.0 to 0.2 GPa and from 19 to 1.2 MPa, respectively, which mimics the mechanical behaviors of cancellous bone. © 2019 Elsevier B.V. All rights reserved.
Keywords: Intermetallics Rapid-solidification Shape memory Elasticity Thermal analysis
1. Introduction TiNi alloys, which are known as the shape memory alloys, have already been extensively applied as metallic biomaterials in making biomedical apparatus and implant devices such as orthodontic arch wires, bone plates and stents due to their good biocompatibility, excellent corrosion resistance and low elastic modulus [1e4]. Furthermore, on the basis of the thermoelastic martensitic transformation derived from the austenite phase to the martensite phase, TiNi-based alloys exhibit the novel shape memory effect and superelasticity [5,6]. TiNi-based alloys in porous form have attracted an additional interest as biomaterials for implantation since the introduction of pores into the bulk material provides ingrowth of living tissues and firm fixation in addition to producing light weight ductile shape memory alloys [7e10]. Especially, it is possible to decrease the elastic modulus of these alloy implants by increasing their porosity. The increased porosity reduces the mismatch in stiffness between bone and implant and then eliminates stress shielding effects which shorten the lifetime of the implant through bone resorption and loosening [11]. It is well known that the proper powder metallurgy processes are the promising methods to fabricate porous near-net-shape
* 1095 Dalgubeol-daero, Dalseo-gu, Daegu, 42601, Republic of Korea E-mail address: [email protected].
components. Li et al. [12] successfully synthesized porous TiNi shape memory intermetallics with a maximum porosity of 40% from pure elemental Ti and Ni powders by the combination of mechanical alloying and powder sintering. Self-propagating high temperature synthesis (SHS) processes were used to prepare porous TiNi alloys with larger porosity of 60% from pure elemental Ti and Ni powders [13]. In these processes, the alloys are formed in solid state via inter-diffusion reactions between Ti and Ni powders. The porous TiNi alloys synthesized by hot isostatic pressing and a conventional sintering method exhibited martensitic transformations, but often with reduced latent heat, suggesting incomplete formation of the TiNi intermetallic phase. In fact, all TiNi alloys synthesized via solid-state reaction processes were found to contain considerable amounts of undesired secondary intermetallics such as Ti2Ni and TiNi3 [14]. The presence of Ti2Ni and TiNi3 in the TiNi matrix may cause a degradation of the shape memory effects and superelasticity. Zhao et al. [15] fabricated porous TiNi alloys from the pre-alloyed TiNi raw powders by spark plasma sintering (SPS). Their results showed that porous TiNi alloys with 13% porosity had excellent superelasticity similar to dense TiNi alloys. Unfortunately, the porosity of the TiNi prepared by this method could not exceed 20%. Wu et al. [16] synthesized porous TiNi alloys with 48% porosity from pre-alloyed TiNi powders by capsule-free hot isostatic pressing (CF-HIP) with ammonium acid carbonate as a space-holder. Although the higher porosity caused by the space-holder resulted in very low elastic modulus of
https://doi.org/10.1016/j.jallcom.2019.153220 0925-8388/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
2
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
6~11 GPa, many circular pores were enclosed in the bulk and then the porosity of open-pores was only 40~70%. A number of conventional powder metallurgy techniques were used to produce porous TiNi shape memory alloys. However, there were a lot of limitations in terms of appropriate pore size, homogeneous pore distribution, high density of open-pore and high level of porosity. It is necessary to develop a better processing method which provides porous scaffolds with functionally upgraded pore structure. In this study, very fine fibers are fabricated by a melt overflow process and the rapidly solidified fibers are consolidated in a graphite mold. In order to utilize the superelastic property of TiNi-based shape memory alloys for the application of implants in human body, the crystal structure of the scaffold alloy should be austenite phase at human body temperature (36 C). However, martensitic transformation start temperature (Ms) and austenite transformation finish temperature (Af) of Ti50Ni50 alloy are 47 and 80 C, respectively [17]. It was reported that the substitution of Mo for Ni changed the martensitic transformation behavior from a B2B190 one-step transformation to a B2-R-B9’ two-step martensitic transformation and decreased the transformation temperatures [4]. Therefore, a Ti50Ni49.7Mo0.3 alloy system was investigated for this study. And the effect of high porosity on the mechanical performances of Ti50Ni49.7Mo0.3 shape memory alloys was investigated systemically. 2. Material and methods Ti50Ni49.7Mo0.3 ingots were prepared from high purity elements of titanium, nickel and molybdenum by non-consumable arcmelting system under argon atmosphere. The mother alloy was remelted more than five times under the high purity of argon atmosphere to achieve the homogeneous chemical composition. In the present study, the experimental works to make fine alloy fibers were performed using a laboratory scale arc melt overflow. About 30 g ingot was placed in a water-cooled hearth, and skull melted under argon atmosphere by plasma beam. Then the hearth was tilted about the rotating quenching wheel, which was made of molybdenum. The liquid metal overflowed over a relatively horizontal edge or pour spot, to contact the cooling wheel surface. The quenching wheel substrate served as a continuous permanent mold, against which the casting solidified. The thickness of fibers can be controlled by the rotating speed of the wheel. The dimension of Mo cooling wheels is 122 mm diameter and 10 mm width. Its tip was acutely machined, in order to produce the shape of fibers or filaments. The speed of the cooling wheel was kept at 1200 rpm to produce relatively thin fibers. These rapidly solidified fibers were cut into segments with the length range from 5 to 10 mm. The fiber segments were uniformly put into the predetermined packing chamber of the graphite mold and then pressure was applied by screwing the bolts to conform the strong contact of fibers during sintering. Sintering was carried out in vacuum induction furnace under a high vacuum atmosphere. The sintering temperature and time were 1200 C and 30 min, respectively. When the sintering was completed, the sample was cooled to room temperature under the high vacuum (103 Pa). Fig. 1 shows the disassembled (upper three piece parts and four screws) and assembled (lower-left) mold pressing equipment and two sintered porous cubes. The porous cubes of 15 mm 15 mm x 15 mm were fabricated. The scaffolds with three different porosities were fabricated to investigate the effect of porosity on the mechanical properties. Fractional porosity is defined as the ratio of pore volume to total volume [16]. Then, the porosity is expressed as
Fig. 1. Photograph of graphite mold pressing equipment and sintered porous cubes.
M 100 Porosityð%Þ ¼ 1 r$V where r and M are the density (6.4 g/cm3) and mass of the alloy fiber, respectively. V is the packing cavity volume (15 x 15 15 mm3) of the mold-pressing equipment. The scaffolds of 60, 70 and 80% porosity were fabricated by filling the cavity of the designed mold with 8.37, 6.28 and 4.18 g of the Ti50Ni49.7Mo0.3 alloy fibers, respectively. In order to analyse the transformation behaviors, differential scanning calorimetry (DSC; TA Instrument Q-20) with a cooling and heating rate of 10 C/min was used. Microstructural investigations were performed by FE-SEM (ZEISS GeminiSEM 300). The mechanical properties of the porous shape memory alloy scaffolds were investigated by uniaxial compression experiments (an Instron 6820 model) with a strain rate of 0.24 mm/min. 3. Results and discussion The rapidly solidified Ti50Ni49.7Mo0.3 alloy fibers prepared by the melt overflow processing are shown in Fig. 2(a). The continuous fibers of about 70 mm diameter were fabricated at the linear speed of 7.7 m/s under high vacuum atmosphere. The insert image in Fig. 2(a) shows the typical longitudinal section of as-solidified fibers. The columnar microstructure was observed in the rapidly solidified fibers. Because the melt came into contact with the quenching wheel during the high speed casting of the melt overflow system, the heat would be extracted to the direction of the wheel during solidification, so that the fibers consisted of long grains, crystallographically oriented with their columnar directions normal to the surface, which was the reverse direction of the heatextraction. As-solidified fibers were sintered under very low press to make the firm bond that linked one fiber to another and then large pores were formed. Fig. 2(b) shows an optical image of the highly porous Ti50Ni49.7Mo0.3 scaffold. The scaffold was 5 mm 5 mm x 12 mm in size and fabricated by cutting the sintered porous cubes of 15 mm 15 mm x 15 mm shown in Fig. 1. This scaffold showed good quality to go further with mechanical testing. Martensitic transformation behaviors of as-solidified fibers and sintered scaffolds were investigated by means of DSC. The scaffolds were solution-treated at 900 C for 1 h in vacuum, and followed by rapid quenching in ice water. All DSC curves show two exothermic
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
3
Fig. 2. Photographs of (a) as-solidified fibers fabricated by a melt overflow process and (b) a porous scaffold processed from the sintered cube for compressive tests. Insert image in Fig. 2(a) corresponds to the typical longitudinal section of as-solidified fibers.
peaks on cooling and one endothermic peak on heating as shown in Fig. 3. According to XRD analysis of Ti-Ni-Mo alloys [4], only the diffraction peaks corresponding to B2 (cubic) parent phase were found at high temperature (above 50 C). However, the diffraction peaks of R (trigonal) martensite started to appear on cooling. On further cooling, the diffraction peaks of the B2 and R martensite disappeared and the diffraction peaks of B19’ (monoclinic) martensite were found. Therefore, the DSC peaks designated Rs and Ms on cooling in the curves of Fig. 3 are attributed to B2/R and R/B190 martensitic transformation, respectively. On heating of DSC curves of Fig. 3, only one large endothermic peak is shown because the reverse transformation temperatures of B19’/R and R/B2 are approaching each other. The Ms and Af of the rapidly solidified fibers are 17 and 42 C, respectively. However, it is found that Af of the porous scaffolds drops to 34 C and the DSC peak is relatively broad and asymmetric. These results might be related to the release of internal stresses such as high dislocation density and small grain size, which were introduced by rapid solidification process, and the Ti-depleted zones which were formed by precipitating Ti2Ni phase from the rapidly solidified fibers during the sintering process [18]. When these scaffolds are used as the implants, they may be cooled to a temperature below Ms and then they are really soft and easy to be deformed before being inserted into the body. The shape of scaffolds must be recovered by shape memory effect during reverse martensitic transformation in human body because the Af of the scaffolds is well below the body temperature. Furthermore, the scaffolds achieve the superelasticity under the external impact or load by the stress-induced
Fig. 3. DSC curves of (a) as-solidified fibers and (b) sintered scaffold of Ti50Ni49.7Mo0.3 alloy.
transformation from austenite phase to martensitic phase. Fig. 4 (a)~(c) show SEM micrographs of the scaffolds with 60, 70 and 80% porosity, respectively. In Fig. 4 (d), a magnified microstructure of the sintered joint area exhibits the achievement of good bonding of fibers. The pores, which are formed by threedimensionally networking structures of the thin fibers, are fully interconnected as open-cell geometries and large enough to provide space for tissue cells and extensive body fluid transport through the scaffold matrix. The pore size of the scaffolds varies widely from 100 to 600 mm. However, the average pore size can be controlled by a porosity variation and then the increase of porosity results in a substantial increase of the pore size as shown in Fig. 4(a) ~(c). Itin et al. [19] reported that the optimal size of pores for implant stability ranged from 100 to 500 mm for the fixation establishing with surround tissues. The size and shape of pores in the present study can meet the requirement of the biomaterials for new bone tissue ingrowth and body fluid supply. Shape memory alloys have the ability to recover strain after deformation. This unique property is directly related to phase transformation between martensitic and austenite phase, and such transformations can be induced by heating (or cooling) and by loading (or unloading) in some temperature regimes. Loadingunloading compressive tests were carried out to investigate the
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
4
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
Fig. 4. SEM images of sintered scaffolds of (a) 60, (b) 70, and (c) 80% porosity and (d) high-magnification SEM image showing the joint-section of fibers formed by sintering at 1250 C for 30 min.
mechanical behaviors of porous Ti-Ni-Mo shape memory alloys. The compressive test samples (5x5x7 mm3) were cut by diamond cutting machining, as shown in Fig. 2(b). The samples with porosity of 60, 70 and 80% were used in the compressive tests. The compressive stress-strain results are shown in Fig. 5. All samples were compressed to the strain of 5% at 40 C, which was a litter higher than the austenite transformation finish temperature (34 C) of the scaffolds. The stress plateau regions in all stress-strain curves are observed, which are introduced by stress-induced martensitic transformation. The stress for martensitic transformation of the scaffolds decreases rapidly from 17 MPa to 1 MPa, when the porosity increases from 60 to 80%. After unloading, the stress-strain curves exhibit incomplete superelastic behaviors. The elastic strain, which is recovered by superelasticity after unloading, decreases as the porosity of scaffolds increases. These incomplete superelastic behaviors were also observed in Ti-Ni alloys [20]. However, most of this residual strain was recovered by heating due to the shape memory effect which occurred during the reverse martensitic transformation at the temperature higher than Af as shown Fig. 5. In the stress-strain curves of porous scaffolds, it is also found that the compressive strength is strongly dependent on the porosity. The increase of the porosity from 60 to 80% results in a rapid decrease of the compressive strength from 19 to 1.2 MPa at 5% strain. The apparent elastic modulus was evaluated in the linear unloading portion of each stress-strain curve, in order to avoid a possible contribution of plastic deformation of the thinnest segments of the porous specimen during the initial elastic loading portion of the cycle [21]. In Fig. 5, variation of the elastic moduli obtained from the stress-strain curves is shown as the function of the porosity. Upon increasing the porosity from 60 to 80%, the elastic modulus decreases from 1.0 to 0.2 GPa. Their low elastic
moduli are valuable mechanical property of porous scaffolds. Elastic modulus of cancellous bone is between 0.1 and 2 GPa and its compressive strength is between 2 and 20 MPa [22]. Even though it is well known that TiNi alloys have similar deformation behavior with that of bone compared with other metallic materials [23], the elastic moduli of fully dense TiNi alloys range from 40 to 90 GPa for austenite phase and from 20 to 50 GPa for martensite phase [24], which are still higher than that of cancellous bone. Depending on the property, the elastic moduli of porous Ti50Ni49.7Mo0.3 scaffolds can be tailored to the elastic modulus of human bone closer than dense metals can, thus reducing the problems associated with stress shielding. Furthermore, in this study, compressive strength and pore size could be controlled as a function of porosity by this simple method to make highly porous scaffolds. 4. Conclusions Ti50Ni49.7Mo0.3 fibers were prepared by a melt overflow process and porous scaffolds were fabricated by solid state sintering of assolidified fibers. The porous scaffolds with a range of porosity from 60 to 80% were successfully processed by filling the different volume fraction of fibers in the restricted container of a sintering mold. The pores exhibited three-dimensionally networking structures of the thin fibers and were fully interconnected as open-cell. The pore size varied widely from 100 to 600 mm. These open-cell pores were large enough to transport extensive body fluid and facilitate bone ingrowth. Two-stage B2/R/B19’ martensitic transformation occurred in the sintered scaffold and its austenite transformation finish temperature (34 C) was enough low to show superelasticity at human body temperature. Upon increasing the porosity of scaffolds from 60 to 80%, the elastic modulus and the compressive
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
5
Reviewing and Editing. Declaration of competing interest None. Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2017R1A2B4005693). References
Fig. 5. Compressive stress-strain curves of sintered scaffolds of (a) 60, (b) 70 and (c) 80% porosity.
strength decreased from 1.0 to 0.2 GPa and from 19 to 1.2 MPa, respectively. These important mechanical properties for implant materials matched well with those of cancellous bone, which could reduce the stress shield effect. Author contribution Yeon-wook Kim: Conceptualization, Methodology, Investigation, Resources, Writing- Original draft preparation, Writing-
[1] J.I. Helsen, H.J. Breme, Metals as Biomaterials, first ed., John Wiley & Sons, Chichester, United Kingdom, 1998. [2] L.L. Hench, E.C. Ethridge, Biomaterials, an Interfacial Approach, Academic Press, New York, 1992. [3] L.S. Castleman, S.M. Motzkin, F.P. Alicandri, V.L. Bonawit, A.A. Johnson, Biocompatibility of nitinol alloy as an implant material, J. Biomed. Mater. Res. 10 (1976) 695e731. [4] Y.W. Kim, B.G. Jo, S. Young, T.H. Nam, Shape memory characteristics of porous TieNieMo alloys prepared by solid state sintering, Mater. Res. Bull. 82 (2016) 45e49. [5] K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni-based shape memory alloys, Prog. Mater. Sci. 50 (2005) 511e678. [6] K. Otsuka, X.B. Ren, Recent developments in the research of shape memory alloys, Intermetallics 7 (1999) 511e528. [7] J.Y. Xiong, Y.C. Li, X.J. Wang, P.D. Hodgson, C.E. Wen, Titaniumenickel shape memory alloy foams for bone tissue engineering, J. Mech. Behav. Biomed. Mater. 1 (2008) 269e273. [8] P. Filip, J. Lausmaa, J. Musialek, K. Mazanec, Structure and surface of TiNi human implants, Biomaterials 22 (2001) 2131e2138. [9] S.L. Wu, P.K. Chu, X.M. Liu, C.Y. Chung, J.P.Y. Ho, C.L. Chu, S.C. Tjong, K.W.K. Yeung, W.W. Lu, K.M.C. Cheung, K.D.K. Luk, Surface characteristics, mechanical properties, and cytocompatibility of oxygen plasma-implanted porous nickel titanium shape memory alloy, J. Biomed. Mater. Res. A 79 (2006) 139e146. [10] C. Greiner, S.M. Oppenheimer, D.C. Dunand, High strength, low stiffness, porous NiTi with superelastic properties, Acta Biomater. 1 (2005) 705e716. [11] G. Ryan, A. Pandit, B. Apatsidis, Fabrication methods of porous metals for use in orthopaedic applications, Biomaterials 27 (2006) 2651e2670. [12] B.Y. Li, L.J. Rong, Y.Y. Li, Stressestrain behavior of porous NieTi shape memory intermetallics synthesized from powder sintering, Intermetallics 8 (2000) 643e646. [13] B.Y. Li, L.J. Rong, Y.Y. Li, V.E. Gjunter, Synthesis of porous NieTi shape-memory alloys by self-propagating high-temperature synthesis: reaction mechanism and anisotropy in pore structure, Acta Mater. 48 (2000) 3895e3904. [14] S.L. Wu, X.M. Liu, P.K. Chu, C.Y. Chung, C.L. Chu, K.W. Yeung, Phase transformation behavior of porous NiTi alloys fabricated by capsule-free hot isostatic pressing, J. Alloy. Comp. 449 (2008) 139e143. [15] Y. Zhao, M. Taya, Y. Kang, A. Kawasaki, Compression behavior of porous NiTi shape memory alloy, Acta Mater. 53 (2005) 337e343. [16] S. Wu, C.Y. Chung, X. Liu, P.K. Chu, K.J.P. Y Ho, C.L. Chu, Y.L. Chan, K.W.K. Yeung, W.W. Lu, K.M.C. Cheung, K.D.K. Luk, Pore formation mechanism and characterization of porous NiTi shape memory alloys synthesized by capsule-free hot isostatic pressing, Acta Mater. 55 (2007) 3437e3451. €m, C. Quiu, New modelling of the B2 phase [17] W. Tang, B. Sundmann, R. Sandstro and its associated martensitic transformation in the TieNi system, Acta Mater. 47 (1999) 3457e3468. [18] J. Mentz, J. Frenzel, M. Wagner, K. Newking, G. Eggeler, H.P. Buchkremer, €ver, Powder metallurgical processing of NiTi shape memory alloys with D. Sto elevated transformation temperatures, Mater. Sci. Eng. A 491 (2008) 270e278. [19] V.I. Itin, V.E. Gyunter, S.A. Shabalovskaya, R.L.C. Sachdeva, Mechanical properties and shape memory of porous nitinol, Mater. Character 32 (1994) 179e187. [20] Y.W. Kim, Mechanical properties of highly porous Ti49.5Ni50.5 biomaterials, Intermetallics 62 (2015) 56e59. [21] O. Scalzo, S. Turenne, M. Gauthier, V. Brailovski, Mechanical and microstructural characterization of porous NiTi shape memory alloys, Metall. Mater. Trans. A 40A (2009) 2061e2070. [22] S. Bose, M. Roy, A. Bandyopadhyay, Recent advances in bone tissue engineering scaffolds, Trends Biotechnol. 30 (2012) 546e554. [23] S.A. Shabalovskaya, Surface, corrosion and biocompatibility aspects of Nitinol as an implant material, Mater. Eng. 12 (2002) 69e109. [24] Y.F. Zheng, B.B. Zhang, B.L. Wang, L. Li, Q.B. Yang, L.S. Cui, Introduction of antibacterial function into biomedical TiNi shape memory alloy by the addition of element Ag, Acta Biomater. 7 (2011) 2758e2767.
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds Yeon-wook Kim* Department of Advanced Materials Engineering, Keimyung University, Daegu, 42403, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 9 October 2019 Received in revised form 28 November 2019 Accepted 29 November 2019 Available online xxx
A simple and effective way to fabricate highly porous scaffolds with controlled porosity and pore size is demonstrated in this study. Ti50Ni49.7Mo0.3 shape memory alloy fibers were prepared by a melt overflow process. The scaffolds with high porosity of 60~80% and large pores of 100~600 mm in size were fabricated by sintering the as-solidified fibers. The martensitic transformation behavior of the porous scaffold was determined and it was found that superelasticity could be achieved at human body temperature. An effect of porosity on mechanical properties of porous scaffolds was investigated using compressive tests. As the porosity increases from 60% to 80%, elastic modulus and compressive strength decrease from 1.0 to 0.2 GPa and from 19 to 1.2 MPa, respectively, which mimics the mechanical behaviors of cancellous bone. © 2019 Elsevier B.V. All rights reserved.
Keywords: Intermetallics Rapid-solidification Shape memory Elasticity Thermal analysis
1. Introduction TiNi alloys, which are known as the shape memory alloys, have already been extensively applied as metallic biomaterials in making biomedical apparatus and implant devices such as orthodontic arch wires, bone plates and stents due to their good biocompatibility, excellent corrosion resistance and low elastic modulus [1e4]. Furthermore, on the basis of the thermoelastic martensitic transformation derived from the austenite phase to the martensite phase, TiNi-based alloys exhibit the novel shape memory effect and superelasticity [5,6]. TiNi-based alloys in porous form have attracted an additional interest as biomaterials for implantation since the introduction of pores into the bulk material provides ingrowth of living tissues and firm fixation in addition to producing light weight ductile shape memory alloys [7e10]. Especially, it is possible to decrease the elastic modulus of these alloy implants by increasing their porosity. The increased porosity reduces the mismatch in stiffness between bone and implant and then eliminates stress shielding effects which shorten the lifetime of the implant through bone resorption and loosening [11]. It is well known that the proper powder metallurgy processes are the promising methods to fabricate porous near-net-shape
* 1095 Dalgubeol-daero, Dalseo-gu, Daegu, 42601, Republic of Korea E-mail address: [email protected].
components. Li et al. [12] successfully synthesized porous TiNi shape memory intermetallics with a maximum porosity of 40% from pure elemental Ti and Ni powders by the combination of mechanical alloying and powder sintering. Self-propagating high temperature synthesis (SHS) processes were used to prepare porous TiNi alloys with larger porosity of 60% from pure elemental Ti and Ni powders [13]. In these processes, the alloys are formed in solid state via inter-diffusion reactions between Ti and Ni powders. The porous TiNi alloys synthesized by hot isostatic pressing and a conventional sintering method exhibited martensitic transformations, but often with reduced latent heat, suggesting incomplete formation of the TiNi intermetallic phase. In fact, all TiNi alloys synthesized via solid-state reaction processes were found to contain considerable amounts of undesired secondary intermetallics such as Ti2Ni and TiNi3 [14]. The presence of Ti2Ni and TiNi3 in the TiNi matrix may cause a degradation of the shape memory effects and superelasticity. Zhao et al. [15] fabricated porous TiNi alloys from the pre-alloyed TiNi raw powders by spark plasma sintering (SPS). Their results showed that porous TiNi alloys with 13% porosity had excellent superelasticity similar to dense TiNi alloys. Unfortunately, the porosity of the TiNi prepared by this method could not exceed 20%. Wu et al. [16] synthesized porous TiNi alloys with 48% porosity from pre-alloyed TiNi powders by capsule-free hot isostatic pressing (CF-HIP) with ammonium acid carbonate as a space-holder. Although the higher porosity caused by the space-holder resulted in very low elastic modulus of
https://doi.org/10.1016/j.jallcom.2019.153220 0925-8388/© 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
2
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
6~11 GPa, many circular pores were enclosed in the bulk and then the porosity of open-pores was only 40~70%. A number of conventional powder metallurgy techniques were used to produce porous TiNi shape memory alloys. However, there were a lot of limitations in terms of appropriate pore size, homogeneous pore distribution, high density of open-pore and high level of porosity. It is necessary to develop a better processing method which provides porous scaffolds with functionally upgraded pore structure. In this study, very fine fibers are fabricated by a melt overflow process and the rapidly solidified fibers are consolidated in a graphite mold. In order to utilize the superelastic property of TiNi-based shape memory alloys for the application of implants in human body, the crystal structure of the scaffold alloy should be austenite phase at human body temperature (36 C). However, martensitic transformation start temperature (Ms) and austenite transformation finish temperature (Af) of Ti50Ni50 alloy are 47 and 80 C, respectively [17]. It was reported that the substitution of Mo for Ni changed the martensitic transformation behavior from a B2B190 one-step transformation to a B2-R-B9’ two-step martensitic transformation and decreased the transformation temperatures [4]. Therefore, a Ti50Ni49.7Mo0.3 alloy system was investigated for this study. And the effect of high porosity on the mechanical performances of Ti50Ni49.7Mo0.3 shape memory alloys was investigated systemically. 2. Material and methods Ti50Ni49.7Mo0.3 ingots were prepared from high purity elements of titanium, nickel and molybdenum by non-consumable arcmelting system under argon atmosphere. The mother alloy was remelted more than five times under the high purity of argon atmosphere to achieve the homogeneous chemical composition. In the present study, the experimental works to make fine alloy fibers were performed using a laboratory scale arc melt overflow. About 30 g ingot was placed in a water-cooled hearth, and skull melted under argon atmosphere by plasma beam. Then the hearth was tilted about the rotating quenching wheel, which was made of molybdenum. The liquid metal overflowed over a relatively horizontal edge or pour spot, to contact the cooling wheel surface. The quenching wheel substrate served as a continuous permanent mold, against which the casting solidified. The thickness of fibers can be controlled by the rotating speed of the wheel. The dimension of Mo cooling wheels is 122 mm diameter and 10 mm width. Its tip was acutely machined, in order to produce the shape of fibers or filaments. The speed of the cooling wheel was kept at 1200 rpm to produce relatively thin fibers. These rapidly solidified fibers were cut into segments with the length range from 5 to 10 mm. The fiber segments were uniformly put into the predetermined packing chamber of the graphite mold and then pressure was applied by screwing the bolts to conform the strong contact of fibers during sintering. Sintering was carried out in vacuum induction furnace under a high vacuum atmosphere. The sintering temperature and time were 1200 C and 30 min, respectively. When the sintering was completed, the sample was cooled to room temperature under the high vacuum (103 Pa). Fig. 1 shows the disassembled (upper three piece parts and four screws) and assembled (lower-left) mold pressing equipment and two sintered porous cubes. The porous cubes of 15 mm 15 mm x 15 mm were fabricated. The scaffolds with three different porosities were fabricated to investigate the effect of porosity on the mechanical properties. Fractional porosity is defined as the ratio of pore volume to total volume [16]. Then, the porosity is expressed as
Fig. 1. Photograph of graphite mold pressing equipment and sintered porous cubes.
M 100 Porosityð%Þ ¼ 1 r$V where r and M are the density (6.4 g/cm3) and mass of the alloy fiber, respectively. V is the packing cavity volume (15 x 15 15 mm3) of the mold-pressing equipment. The scaffolds of 60, 70 and 80% porosity were fabricated by filling the cavity of the designed mold with 8.37, 6.28 and 4.18 g of the Ti50Ni49.7Mo0.3 alloy fibers, respectively. In order to analyse the transformation behaviors, differential scanning calorimetry (DSC; TA Instrument Q-20) with a cooling and heating rate of 10 C/min was used. Microstructural investigations were performed by FE-SEM (ZEISS GeminiSEM 300). The mechanical properties of the porous shape memory alloy scaffolds were investigated by uniaxial compression experiments (an Instron 6820 model) with a strain rate of 0.24 mm/min. 3. Results and discussion The rapidly solidified Ti50Ni49.7Mo0.3 alloy fibers prepared by the melt overflow processing are shown in Fig. 2(a). The continuous fibers of about 70 mm diameter were fabricated at the linear speed of 7.7 m/s under high vacuum atmosphere. The insert image in Fig. 2(a) shows the typical longitudinal section of as-solidified fibers. The columnar microstructure was observed in the rapidly solidified fibers. Because the melt came into contact with the quenching wheel during the high speed casting of the melt overflow system, the heat would be extracted to the direction of the wheel during solidification, so that the fibers consisted of long grains, crystallographically oriented with their columnar directions normal to the surface, which was the reverse direction of the heatextraction. As-solidified fibers were sintered under very low press to make the firm bond that linked one fiber to another and then large pores were formed. Fig. 2(b) shows an optical image of the highly porous Ti50Ni49.7Mo0.3 scaffold. The scaffold was 5 mm 5 mm x 12 mm in size and fabricated by cutting the sintered porous cubes of 15 mm 15 mm x 15 mm shown in Fig. 1. This scaffold showed good quality to go further with mechanical testing. Martensitic transformation behaviors of as-solidified fibers and sintered scaffolds were investigated by means of DSC. The scaffolds were solution-treated at 900 C for 1 h in vacuum, and followed by rapid quenching in ice water. All DSC curves show two exothermic
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
3
Fig. 2. Photographs of (a) as-solidified fibers fabricated by a melt overflow process and (b) a porous scaffold processed from the sintered cube for compressive tests. Insert image in Fig. 2(a) corresponds to the typical longitudinal section of as-solidified fibers.
peaks on cooling and one endothermic peak on heating as shown in Fig. 3. According to XRD analysis of Ti-Ni-Mo alloys [4], only the diffraction peaks corresponding to B2 (cubic) parent phase were found at high temperature (above 50 C). However, the diffraction peaks of R (trigonal) martensite started to appear on cooling. On further cooling, the diffraction peaks of the B2 and R martensite disappeared and the diffraction peaks of B19’ (monoclinic) martensite were found. Therefore, the DSC peaks designated Rs and Ms on cooling in the curves of Fig. 3 are attributed to B2/R and R/B190 martensitic transformation, respectively. On heating of DSC curves of Fig. 3, only one large endothermic peak is shown because the reverse transformation temperatures of B19’/R and R/B2 are approaching each other. The Ms and Af of the rapidly solidified fibers are 17 and 42 C, respectively. However, it is found that Af of the porous scaffolds drops to 34 C and the DSC peak is relatively broad and asymmetric. These results might be related to the release of internal stresses such as high dislocation density and small grain size, which were introduced by rapid solidification process, and the Ti-depleted zones which were formed by precipitating Ti2Ni phase from the rapidly solidified fibers during the sintering process [18]. When these scaffolds are used as the implants, they may be cooled to a temperature below Ms and then they are really soft and easy to be deformed before being inserted into the body. The shape of scaffolds must be recovered by shape memory effect during reverse martensitic transformation in human body because the Af of the scaffolds is well below the body temperature. Furthermore, the scaffolds achieve the superelasticity under the external impact or load by the stress-induced
Fig. 3. DSC curves of (a) as-solidified fibers and (b) sintered scaffold of Ti50Ni49.7Mo0.3 alloy.
transformation from austenite phase to martensitic phase. Fig. 4 (a)~(c) show SEM micrographs of the scaffolds with 60, 70 and 80% porosity, respectively. In Fig. 4 (d), a magnified microstructure of the sintered joint area exhibits the achievement of good bonding of fibers. The pores, which are formed by threedimensionally networking structures of the thin fibers, are fully interconnected as open-cell geometries and large enough to provide space for tissue cells and extensive body fluid transport through the scaffold matrix. The pore size of the scaffolds varies widely from 100 to 600 mm. However, the average pore size can be controlled by a porosity variation and then the increase of porosity results in a substantial increase of the pore size as shown in Fig. 4(a) ~(c). Itin et al. [19] reported that the optimal size of pores for implant stability ranged from 100 to 500 mm for the fixation establishing with surround tissues. The size and shape of pores in the present study can meet the requirement of the biomaterials for new bone tissue ingrowth and body fluid supply. Shape memory alloys have the ability to recover strain after deformation. This unique property is directly related to phase transformation between martensitic and austenite phase, and such transformations can be induced by heating (or cooling) and by loading (or unloading) in some temperature regimes. Loadingunloading compressive tests were carried out to investigate the
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
4
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
Fig. 4. SEM images of sintered scaffolds of (a) 60, (b) 70, and (c) 80% porosity and (d) high-magnification SEM image showing the joint-section of fibers formed by sintering at 1250 C for 30 min.
mechanical behaviors of porous Ti-Ni-Mo shape memory alloys. The compressive test samples (5x5x7 mm3) were cut by diamond cutting machining, as shown in Fig. 2(b). The samples with porosity of 60, 70 and 80% were used in the compressive tests. The compressive stress-strain results are shown in Fig. 5. All samples were compressed to the strain of 5% at 40 C, which was a litter higher than the austenite transformation finish temperature (34 C) of the scaffolds. The stress plateau regions in all stress-strain curves are observed, which are introduced by stress-induced martensitic transformation. The stress for martensitic transformation of the scaffolds decreases rapidly from 17 MPa to 1 MPa, when the porosity increases from 60 to 80%. After unloading, the stress-strain curves exhibit incomplete superelastic behaviors. The elastic strain, which is recovered by superelasticity after unloading, decreases as the porosity of scaffolds increases. These incomplete superelastic behaviors were also observed in Ti-Ni alloys [20]. However, most of this residual strain was recovered by heating due to the shape memory effect which occurred during the reverse martensitic transformation at the temperature higher than Af as shown Fig. 5. In the stress-strain curves of porous scaffolds, it is also found that the compressive strength is strongly dependent on the porosity. The increase of the porosity from 60 to 80% results in a rapid decrease of the compressive strength from 19 to 1.2 MPa at 5% strain. The apparent elastic modulus was evaluated in the linear unloading portion of each stress-strain curve, in order to avoid a possible contribution of plastic deformation of the thinnest segments of the porous specimen during the initial elastic loading portion of the cycle [21]. In Fig. 5, variation of the elastic moduli obtained from the stress-strain curves is shown as the function of the porosity. Upon increasing the porosity from 60 to 80%, the elastic modulus decreases from 1.0 to 0.2 GPa. Their low elastic
moduli are valuable mechanical property of porous scaffolds. Elastic modulus of cancellous bone is between 0.1 and 2 GPa and its compressive strength is between 2 and 20 MPa [22]. Even though it is well known that TiNi alloys have similar deformation behavior with that of bone compared with other metallic materials [23], the elastic moduli of fully dense TiNi alloys range from 40 to 90 GPa for austenite phase and from 20 to 50 GPa for martensite phase [24], which are still higher than that of cancellous bone. Depending on the property, the elastic moduli of porous Ti50Ni49.7Mo0.3 scaffolds can be tailored to the elastic modulus of human bone closer than dense metals can, thus reducing the problems associated with stress shielding. Furthermore, in this study, compressive strength and pore size could be controlled as a function of porosity by this simple method to make highly porous scaffolds. 4. Conclusions Ti50Ni49.7Mo0.3 fibers were prepared by a melt overflow process and porous scaffolds were fabricated by solid state sintering of assolidified fibers. The porous scaffolds with a range of porosity from 60 to 80% were successfully processed by filling the different volume fraction of fibers in the restricted container of a sintering mold. The pores exhibited three-dimensionally networking structures of the thin fibers and were fully interconnected as open-cell. The pore size varied widely from 100 to 600 mm. These open-cell pores were large enough to transport extensive body fluid and facilitate bone ingrowth. Two-stage B2/R/B19’ martensitic transformation occurred in the sintered scaffold and its austenite transformation finish temperature (34 C) was enough low to show superelasticity at human body temperature. Upon increasing the porosity of scaffolds from 60 to 80%, the elastic modulus and the compressive
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220
Y.-w. Kim / Journal of Alloys and Compounds xxx (xxxx) xxx
5
Reviewing and Editing. Declaration of competing interest None. Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2017R1A2B4005693). References
Fig. 5. Compressive stress-strain curves of sintered scaffolds of (a) 60, (b) 70 and (c) 80% porosity.
strength decreased from 1.0 to 0.2 GPa and from 19 to 1.2 MPa, respectively. These important mechanical properties for implant materials matched well with those of cancellous bone, which could reduce the stress shield effect. Author contribution Yeon-wook Kim: Conceptualization, Methodology, Investigation, Resources, Writing- Original draft preparation, Writing-
[1] J.I. Helsen, H.J. Breme, Metals as Biomaterials, first ed., John Wiley & Sons, Chichester, United Kingdom, 1998. [2] L.L. Hench, E.C. Ethridge, Biomaterials, an Interfacial Approach, Academic Press, New York, 1992. [3] L.S. Castleman, S.M. Motzkin, F.P. Alicandri, V.L. Bonawit, A.A. Johnson, Biocompatibility of nitinol alloy as an implant material, J. Biomed. Mater. Res. 10 (1976) 695e731. [4] Y.W. Kim, B.G. Jo, S. Young, T.H. Nam, Shape memory characteristics of porous TieNieMo alloys prepared by solid state sintering, Mater. Res. Bull. 82 (2016) 45e49. [5] K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni-based shape memory alloys, Prog. Mater. Sci. 50 (2005) 511e678. [6] K. Otsuka, X.B. Ren, Recent developments in the research of shape memory alloys, Intermetallics 7 (1999) 511e528. [7] J.Y. Xiong, Y.C. Li, X.J. Wang, P.D. Hodgson, C.E. Wen, Titaniumenickel shape memory alloy foams for bone tissue engineering, J. Mech. Behav. Biomed. Mater. 1 (2008) 269e273. [8] P. Filip, J. Lausmaa, J. Musialek, K. Mazanec, Structure and surface of TiNi human implants, Biomaterials 22 (2001) 2131e2138. [9] S.L. Wu, P.K. Chu, X.M. Liu, C.Y. Chung, J.P.Y. Ho, C.L. Chu, S.C. Tjong, K.W.K. Yeung, W.W. Lu, K.M.C. Cheung, K.D.K. Luk, Surface characteristics, mechanical properties, and cytocompatibility of oxygen plasma-implanted porous nickel titanium shape memory alloy, J. Biomed. Mater. Res. A 79 (2006) 139e146. [10] C. Greiner, S.M. Oppenheimer, D.C. Dunand, High strength, low stiffness, porous NiTi with superelastic properties, Acta Biomater. 1 (2005) 705e716. [11] G. Ryan, A. Pandit, B. Apatsidis, Fabrication methods of porous metals for use in orthopaedic applications, Biomaterials 27 (2006) 2651e2670. [12] B.Y. Li, L.J. Rong, Y.Y. Li, Stressestrain behavior of porous NieTi shape memory intermetallics synthesized from powder sintering, Intermetallics 8 (2000) 643e646. [13] B.Y. Li, L.J. Rong, Y.Y. Li, V.E. Gjunter, Synthesis of porous NieTi shape-memory alloys by self-propagating high-temperature synthesis: reaction mechanism and anisotropy in pore structure, Acta Mater. 48 (2000) 3895e3904. [14] S.L. Wu, X.M. Liu, P.K. Chu, C.Y. Chung, C.L. Chu, K.W. Yeung, Phase transformation behavior of porous NiTi alloys fabricated by capsule-free hot isostatic pressing, J. Alloy. Comp. 449 (2008) 139e143. [15] Y. Zhao, M. Taya, Y. Kang, A. Kawasaki, Compression behavior of porous NiTi shape memory alloy, Acta Mater. 53 (2005) 337e343. [16] S. Wu, C.Y. Chung, X. Liu, P.K. Chu, K.J.P. Y Ho, C.L. Chu, Y.L. Chan, K.W.K. Yeung, W.W. Lu, K.M.C. Cheung, K.D.K. Luk, Pore formation mechanism and characterization of porous NiTi shape memory alloys synthesized by capsule-free hot isostatic pressing, Acta Mater. 55 (2007) 3437e3451. €m, C. Quiu, New modelling of the B2 phase [17] W. Tang, B. Sundmann, R. Sandstro and its associated martensitic transformation in the TieNi system, Acta Mater. 47 (1999) 3457e3468. [18] J. Mentz, J. Frenzel, M. Wagner, K. Newking, G. Eggeler, H.P. Buchkremer, €ver, Powder metallurgical processing of NiTi shape memory alloys with D. Sto elevated transformation temperatures, Mater. Sci. Eng. A 491 (2008) 270e278. [19] V.I. Itin, V.E. Gyunter, S.A. Shabalovskaya, R.L.C. Sachdeva, Mechanical properties and shape memory of porous nitinol, Mater. Character 32 (1994) 179e187. [20] Y.W. Kim, Mechanical properties of highly porous Ti49.5Ni50.5 biomaterials, Intermetallics 62 (2015) 56e59. [21] O. Scalzo, S. Turenne, M. Gauthier, V. Brailovski, Mechanical and microstructural characterization of porous NiTi shape memory alloys, Metall. Mater. Trans. A 40A (2009) 2061e2070. [22] S. Bose, M. Roy, A. Bandyopadhyay, Recent advances in bone tissue engineering scaffolds, Trends Biotechnol. 30 (2012) 546e554. [23] S.A. Shabalovskaya, Surface, corrosion and biocompatibility aspects of Nitinol as an implant material, Mater. Eng. 12 (2002) 69e109. [24] Y.F. Zheng, B.B. Zhang, B.L. Wang, L. Li, Q.B. Yang, L.S. Cui, Introduction of antibacterial function into biomedical TiNi shape memory alloy by the addition of element Ag, Acta Biomater. 7 (2011) 2758e2767.
Please cite this article as: Y.-w. Kim, Martensitic transformation behavior and mechanical properties of highly porous Ti-Ni-Mo scaffolds, Journal of Alloys and Compounds, https://doi.org/10.1016/j.jallcom.2019.153220