- Email: [email protected]
Micromechanical properties of single crystal hydroxyapatite by nanoindentation
Available online at www.sciencedirect.com
Acta Biomaterialia 5 (2009) 2206–2212 www.elsevier.com/locate/actabiomat
Micromechanical properties of single crystal hydroxyapatite by nanoindentation Saeed Saber-Samandari *, Ka¯rlis A. Gross Department of Mechanical Engineering, The University of Melbourne, Parkville, Vic. 3010, Australia Swinburne University of Technology, Faculty of Engineering and Industrial Sciences, IRIS, Hawthorn, Vic. 3122, Australia Received 13 October 2008; received in revised form 4 February 2009; accepted 5 February 2009 Available online 11 February 2009
Abstract Knowledge of the intrinsic properties of hydroxyapatite (HAp) single crystals is important for the design of natural systems and will assist further improvements of manufactured biomaterials. Nanoindentation provides a useful tool for determining mechanical properties such as the hardness, elastic modulus and fracture toughness of small samples. A Berkovich indenter was placed on the side and basal planes of a natural single crystal of Durango HAp. The hardness and elastic modulus values obtained revealed higher values for the base (7.1 and 150.4 GPa) compared to the side (6.4 and 143.6 GPa). The cracking threshold, i.e., the load at which cracking initiates, revealed earlier crack formation on the base (at 8 mN) compared to the side (at 11 mN). Fracture toughness was measured as 0.45 ± 0.09 and 0.35 ± 0.06 MPa m1/2 for the side and basal plane, respectively. These results suggest that crystals are less prone to cracking and resist microcrack events better on the side, which is useful in bone, while exposing the base, the hardest face, to minimize mass loss from abrasion with teeth. Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Hydroxyapatite; Single crystal; Mechanical properties; Nanoindentation
1. Introduction Details about the bond strength and elasticity within crystals can be best obtained from single crystals, which can provide important information on the hardness, strength and elastic modulus. Since apatite is more commonly found as nanocrystals, special techniques need to be employed to produce a single crystal of apatite. An alternative is to source naturally occurring single crystals of apatite. Apatite used in biomedical applications is always available as micron- or nanometre-sized crystals of hydroxyapatite (HAp). Various techniques have been employed to prepare the nano-sized HAp, including precipitation, hydrothermal [1–3], sol–gel [4], crystal conversion [5] and microwave techniques [6]. Recently, new techniques have *
Corresponding author. Tel.: +61 3 8344 0024. E-mail address: [email protected] (S. Saber-Samandari).
evolved to prepare micro-sized HAp [7,8]. The crystal size, shape, chemical composition, defect concentration and assemblage all depend on the manufacturing technology. As a result, it is difficult to distinguish the contributions of the microstructure, chemistry, morphology, charge and residual stress to properties such as hardness. Studies on single crystals should be able to determine the base properties, and serve as an important reference for other studies that intend to improve on these properties of apatites. Natural apatite in bone has elongated crystals incorporated into collagen. Bone cements, used as bone fillers, also have long crystals. Recent advances in thermal spraying have revealed the ability to create oriented single crystals [9]. The properties of these materials depend on the orientation of the crystals. Testing on such fine crystals is challenging; an easier approach would be to conduct testing on large single crystals. This would not only provide values for mechanical properties of the side (parallel to the c-axis), but also provide enough material for testing of the base
1742-7061/$ - see front matter Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2009.02.009
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
(parallel to the a-axis). Such information could form the basis for understanding the effect of crystal orientation in the different HAp products used in the medical industry. Large-sized single crystals of HAp are required for mechanical property testing. These can be found in millimeter sizes in nature at Cerro de Mercado, close to Durango in Mexico. No experimental study on the micromechanical properties of a natural HAp single crystal has been reported. The goal of this work was therefore to employ nanoindentation techniques to probe such mechanical properties as hardness, elastic modulus and fracture toughness. Nanoindentation would be performed on the side and the basal plane of the sample to confirm the effect of orientation on its mechanical properties. To this end, a natural single crystal of HAp from Durango was selected. It is believed that this is the first comprehensive study of single crystal HAp. 2. Experimental procedure The single crystal was embedded in epoxy resin to provide support for cutting the cross-section and metallographically polishing it to produce the reasonably flat surface required for nanoindentation testing. Both a cross-section and longitudinal section were produced to assess the mechanical properties relative to the crystal faces. Fine grinding was conducted with silicon carbide abrasive paper of 2400 grit and polishing involved a 1 lm diamond suspension on a Dur surface (Struers, Denmark) for 5 min. Micro-Raman spectroscopy was used to confirm the purity of the sample (Fig. 1). Analyses were conducted with an excitation wavelength of 514 nm at a spectrum resolution of 1 cm1. Spectra were recorded within the range 500– 1500 cm1. A very narrow peak was found at 960 cm1 for the side and basal plane, confirming that the only calcium phosphate present was HAp [10]. The residual
impression from nanoindentation was observed using an XL30 Philips scanning electron microscope (SEM). The samples were coated with gold using an Edwards S150B sputter coater for examination with the SEM at 20 keV. The SEM images were also used for measuring the length of microcracks to calculate the fracture toughness. A Berkovich three-sided pyramidal indenter was used throughout this study. The area function for the diamond was calibrated by indentation in fused silica (FS) to determine the contact area for a given depth. The contact area as a function of depth was found to be Aðhc Þ ¼ 22h2c þ 2500hc by the polynomial-fitting procedure. The measured depth was then adjusted for the effect of instrument compliance (0.499 nm mN1) in the instrument software for analysis (NanoTest, MML, Wrexham, UK). The hardness and reduced elastic modulus of FS, assuming a Poisson’s ratio of 0.165, were measured to be 8.8 ± 0.1 and 69.6 ± 0.2 GPa, respectively. After calibration, the instrument was set up for two series of 10 indentations in both the side and the basal plane of the sample: one using low loads, ranging from 1 to 20 mN, and the other using high loads, ranging from 20 to 200 mN. The hardness and elastic modulus were measured using the standard method proposed by Oliver and Pharr [11]. The separation distance between the indents was selected based on the size of the residual impression (Berkovich shape). The spacing between adjacent indentations was about 15 times the maximum width of the indentations at low load and more than four times bigger than those at high load to ensure that the deformed region surrounding one indentation did not interfere with the formation of the next indentation. Since the maximum depth at the highest load (200 mN) was about 1.3 lm, indentations were spaced approximately 40 lm apart in a given row and the rows (low load and high load) were spaced at 100 lm. Indentations started approximately at the centre of the sample to avoid any testing near the edges. The loading/unloading rate was set to 10% of minimum load. The dwell periods at maximum load and for drift correction were 5 and 30 s, respectively. All experiments were performed at room temperature. Fracture toughness is another mechanical property that can be measured using nanoindentation techniques. For the calculation of the fracture toughness, the relationship between the fracture toughness and the length of the radial cracks using pyramidal sharp indenters was considered using the following equation [12] K IC
Fig. 1. Raman shifts of the natural single crystal HAp.
2207
1=2 E P ¼W H c3=2
where KIC is the mode I critical stress intensity factor, W is a material-independent constant that must be calibrated by indentation of well-characterized materials, H is the hardness, E is the elastic modulus, P is the peak load and c is the radius of the surface crack. The empirical constant was found to be 0.016 for Berkovich-type indenters [13].
2208
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
3. Results and discussion Loading–unloading curves of 10 indentations made on the base of the HAp single crystal are depicted in Fig. 2, where the load varied from 20 to 200 mN. This clearly reveals that the sample is homogeneous. Shown in Fig. 3 is the hardness and reduced elastic modulus of both the side and the basal plane of the single crystal HAp as a function of load. Plots illustrate that applying a 30 mN load or greater allows determination of the true hardness (a hardness that remains unchanged from the size of the indentation [14]). The indentation size effect was observed at low load, where hardness decreases with increasing load up to 30 mN. Hardness and reduced elastic modulus as functions of indentation depth are depicted in Fig. 4. It can be inferred that true hardness and reduced elastic modulus could be determined with a penetration depth of more than 500 nm (Fig. 4). The indentation size effect was also noticed at penetration depths shallower than 500 nm. The measured moduli are expected to be constant with indentation depth (or load); however, results reveal an approximately 5% overestimation with loads less than 30 mN or depths shallower than 500 nm. The discrepancy between the expected and obtained results could be due to the pile-up phenomenon. Oliver and Pharr’s proposed method is best applied when no pile-up but some sink-in occurs when the sample is indented. In the case of pileup, contact depth is always more than the total depth of indentation, thus the projected area is always underestimated. This underestimated area could overestimate both hardness and reduced elastic modulus, with a more significant effect on hardness. The reduced elastic modulus and hardness are calculated with the following expressions: 2S pffiffiffi Er ¼ pffiffiffi A p P max H¼ A
Fig. 3. The hardness and reduced elastic modulus vs applied load results from nanoindentation tests on both the side and basal planes of single crystal HAp.
Fig. 4. The hardness and reduced elastic modulus vs indentation depth results from nanoindentation tests on both the side and basal planes of single crystal HAp.
Fig. 2. Load–displacement curves for 10 indentations on the base of the single crystal of HAp. The applied load varies from 20 to 200 mN.
Recent nanoindentation studies on single crystals of different materials (i.e., metals, semiconductors, ceramics and oxides) show pop-in phenomena (plateau behaviour in loading) from applying low loads [15–17]. Interestingly, other nanoindentation studies on single crystals that have a hexagonal close-packed (hcp) lattice structure (i.e., sapphire, GaN and ZnO) show multiple load plateau behaviour [18–20]. HAp has a hexagonal lattice structure and therefore multiple load plateau behaviour is expected. Load plateau behaviour could be a result of a microcrack [21], a pore underneath the indenter [22] or the pile-up phenomenon. In the case of single crystals, existence of pores has not been reported. Therefore, surface topography
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
tools, such as atomic force microscopy and SEM at high magnification, are needed to reveal the existence of pileup and/or microcracks. Fig. 5A shows a loading–unloading curve obtained from nanoindentation on the side of the sample under a 7.5 mN load. Discontinuities on the loading curve are observed as expected, since HAp has an hcp crystal structure. In addition, an SEM image of the residual indentation at very high magnification (120,000) reveals no crack, but deformation at the edges of the impression (Fig. 5B). Clearly, pile-up only occurs during loading. However, multiple pop-in events are visible in Fig. 5C when the applied load is increased to about 11 mN. A SEM image of its corresponding loading–unloading curve reveals both microcracks and deformation at the edges (Fig. 5D). It was therefore concluded that the cracking threshold for single crystal HAp is about 11 mN on the side face using a Berkovich indenter. It is noted that, with a Vickers indenter, cracking thresholds in most ceramic materials are about 250 mN or more [23]. Significantly lower thresholds can be achieved using the cube-corner indenter [24], but no report on the cracking threshold of ceramic materials using Berkovich (the most commonly used) indenter are known.
2209
It is interesting to compare the cracking threshold of the side with that of the base. Fig. 6A and B show the loading– unloading curve and its corresponding SEM image of the indentation for the base of the single crystal HAp. Pop-in events are observed with a 5 mN load (Fig. 6A). Fig. 6B reveals no microcrack but a few occurrences of pile-up. In Fig. 6C, multiple pop-in events are evident under a maximum load of 8 mN, while its corresponding SEM image reveals both pile-up and three microcracks (Fig. 6D). As a result, the cracking threshold of the single crystal HAp on its basal plane is approximately 8 mN. Therefore, cracks propagate sooner on the base of the HAp single crystal than on its side, indicating that the base is more brittle. Overloading the pyramidal Berkovich indenter in this study causes crack propagation around the indent. Observation and length measurement of such cracks allow the fracture toughness to be determined. In this study, the average crack lengths were taken since the crack lengths were found to be anisotropic (Fig. 7A). The fracture toughness values (average of 12 measurements) obtained for the side and basal planes were 0.45 ± 0.09 and 0.35 ± 0.06 MPa m1/2, respectively. The side is tougher than the base but has lower hardness and elastic modulus values, (Table 1).
Fig. 5. Nanoindentation on the side of single crystal Hap. (A) Load–unload curve showing pop-in events; (B) an SEM image (120,000) showing no microcrack, with a maximum load of 7.5 mN; (C) multiple pop-in events from an 11 mN applied load; and (D) the corresponding SEM image showing pile-up and three microcracks.
2210
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
Fig. 6. Nanoindentation made on the base of single crystal Hap. (A) Load–unload curve showing pop-in events using a maximum load of 5 mN; (B) an SEM image (120,000) showing pile-up on the top two sides; (C) multiple pop-in events from an 8 mN applied load; and (D) the corresponding SEM image with pile-up and three microcracks.
Shown in Fig. 7 are residual indentation SEM images of the HAp single crystal with cracked corners from application of an 80 mN load on the side and basal plane. Fig. 7A shows a crack length of 7.73 lm on the side, while a larger crack length of 8.09 lm is observed on the base, confirming the lower fracture toughness of the base. Anisotropy found in crack lengths could be due to anisotropy in the surface energy, as the base of the HAp crystal has a higher surface energy than the side [25]. Moreover, a polycrystalline sample shows a higher fracture toughness than a single crystal due to grain boundaries, which reduce the crack growth. The fracture toughness for polycrystalline HAp was reported to be between 0.6 and 1.5 MPa m1/2 [26]. The results obtained in this study are in good agreement with the values of fracture toughness for the single crystal HAp reported by Viswanath et al. [27]. They reported the mechanical properties of manufactured single crystal HAp below pop-in events and found fracture toughness values of 0.48 and 0.28 MPa m1/2 for the side and basal plane, respectively. These results outline the anisotropic mechanical properties of single crystal HAp. Since about 70% of bone is made up of the inorganic mineral HAp, anisotropy in the mechanical properties of bone is expected. This behaviour has been observed in fibrolamellar bone [28], single osteon
lamellae in bone tissue [29], human lumbar vertebral cortical and trabecular bone [30], and human tibial cortical bone [31]. Most studies on bone show that differences in mechanical properties may be due to collagen fibre orientation [32]. This study reveals the contribution of HAp orientation to mechanical properties. The assembly of HAp crystals will influence the properties of bone. The arrangement of the crystals aligned at an angle to the long bone in the femur will provide 5% greater stiffness. This is not a large difference, but it does contribute to the anisotropy. Crystal geometry is also an important factor. The small surface area of the rod end lowers the potential for crack formation along the c-axis and positions the crystals where they may be more effective in bending. More energy will be required to form a crack in the side of a rod. The greater resistance to crack formation on the side of the crystal will influence the microcrack path that is formed in bone. Teeth contain the highest concentration of apatite crystals within the body, and have the side of the crystal towards the outer surface of the tooth. This presents the crystal side for easier remodeling, an event that is continually taking place in the aqueous environment of the oral cavity. The higher hardness of the crystal side indicates that nature has optimized the crystal orientation to minimize
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
2211
were determined to be 11 and 8 mN, respectively. The results of these experiments indicated higher hardness and elastic modulus values for the base (by about 11% and 5%, respectively) than for the side. Unlike the hardness and elastic modulus, fracture toughness was found to be higher (approximately 25%) for the side, as longer crack lengths were observed on the base compared to the side under the same applied load. Acknowledgements The authors acknowledge support from an ARC Project # DP0774251. The nanoindenter was purchased with funds provided by the Rowden White Fund and supplemented by the ARC LIEF Grant No. LE0775715. The single crystal was kindly provided by Prof. Andrew Gleadow and Assoc. Prof. Barry Kohn from the School of Earth Sciences, University of Melbourne. This work has partially evolved from a past Ramaciotti Grant. References
Fig. 7. Residual indentation image of single crystal HAp with cracked corners from application of an 80 mN load on (A) the side, with a 7.73 lm crack length; and (B) the base, with 8.09 lm crack length.
Table 1 Mechanical properties of single crystal HAp. Orientation
H (GPa)
Er (GPa)
E (GPa)
KIC (MPa m1/2)
Side Base
6.41 7.06
137.98 142.92
143.56 150.38
0.45 0.35
abrasion. The abrasion from opposing teeth and third body wear from mineral inclusions in foods will thus be minimized [33]. The additive effect of strengthening from fluoride [34] and from crystal orientation shows the benefits of separate microstructural elements in optimizing materials for their function. Similarly, the crystal shape, size and arrangement will be important for HAp materials used in biomedical applications. The properties of the single crystal will provide important reference information for the assessment of presently used HAp materials and support for further improvements. Further work will determine the mechanical properties of HAp used in biomedical implants. 4. Conclusions Thresholds for crack-initiation on the side and basal planes of single crystal HAp using a Berkovich indenter
[1] Ito A, Nakamura S, Aoki H, et al. Hydrothermal growth of carbonate-containing hydroxyapatite single crystals. J Crys Growth 1996;163:311–7. [2] Yoshimura M, Suda H. Hydrothermal synthesise of biocompatible whiskers. Mater Sci 1994;29:3399–402. [3] Liu J, Ye X, Wang H, et al. The influence of PH and temperature on the morphology of hydroxyapatite synthesized by hydrothermal method. Ceram Inter 2003;29:629–33. [4] Jillavenkatesa A, Hoelzer DT, Condrate RA. An electron microscopy study of the formation of hydroxyapatite through sol–gel processing. Mater Sci 1999;34:4821–30. [5] Zhan J, Tseng Y. Biomimetic formation of hydroxyapatite nanorods by a single-crystal-to-single-crystal transformation. Adv Funct Mater 2005;15:2005–10. [6] Anabel L, Jaime G, Rafael R. Nanosized hydroxyapatite precipitation from homogenous calcium/citrate/phosphate solutions using microwave and conventional heating. Adv Mater 1998;10:49–53. [7] Madsen H, Christensson F, Polyak LE, et al. Calcium phosphate crystallization under terrestrial and microgravity conditions. J Cryst Growth 1995;152:191–202. [8] Teng S, Shi J, Chen L. A novel method to synthesize large-sized hydroxyapatite rods. J Cryst Growth 2006;290:683–8. [9] Saber-Samandari S, Gross KA. Nanoindentation reveals mechanical properties within thermally sprayed hydroxyapatite coatings. Surf Coat Technol 2009;203:1660–4. [10] Gross KA, Phillips MR. Identification and mapping of the amorphous phase in plasma-sprayed hydroxyapatite coatings using scanning cathodoluminescence microscopy. J Mater Sci: Mater Med 1998;9:797–802. [11] Oliver WC, Pharr GM. Measurement of hardness and elastic modulus by instrumented indentation: advances in understanding and refinements to methodology. J Mater Res 2004;19:3–20. [12] Lawn BR, Evans AG, Marshall DB. Elastic/plastic indentation damage in ceramics: the median/radial crack systems. J Am Ceram Soc 1980;63:574–81. [13] Pharr GM, Harding DS, Oliver WC. Measurement of fracture toughness in thin films and small volumes using nanoindentation methods. In: Nastasi M, Parkin DM, Gleiter H, editors. Mechanical properties and deformation behavior of materials having ultra-fine microstructures. Dordrecht: Kluwer Academic Press; 1993. p. 449–61. [14] Manika I, Maniks J. Size effects in micro- and nanoscale indentation. Acta Mater 2006;54:2049–56.
2212
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
[15] Gerberich WW, Nelson JC, Lilleodden ET, Anderson P, Wyrobek JT. Indentation induced dislocation nucleation: the initial yield point. Acta Mater 1996;44:3585–98. [16] Van Vliet KJ, Li J, Zhu T, Yip S, Suresh S. Quantifying the early stages of plasticity through nanoscale experiments and simulations. Phys Rev B 2003;67:104105. [17] Kearney C, Zhao Z, Bruet BJF, et al. Nanoscale anisotropic plastic deformation in single crystal aragonite. Phys Rev Lett 2006;96: 2555051–54. [18] Page TF, Oliver WC, McHargue CJ. The deformation behavior of ceramic crystals subjected to very low load (nano)indentations. J Mater Res 1992;7:450473. [19] Bradby JE, Kucheyev SO, Williams JS, et al. Indentation-induced damage in GaN epilayers. Appl Phys Lett 2002;80:383–5. [20] Kucheyev SO, Bradby JE, Williams JS, Jagadish C, Swain MV. Mechanical deformation of single-crystal ZnO. Appl Phys Lett 2002;80:956–8. [21] Fischer-Cripps AC. Nanoindentation. New York: Springer-Verlag; 2002. [22] Lu X, Wang X, Xiao P. Nanoindentation and residual stress measurements of yttria-stabilized zirconia composite coatings produced by electrophoretic deposition. Thin Solid Films 2006;494: 223–7. [23] Lankford J, Davidson DL. The crack-initiation threshold in ceramic materials subject to elastic/plastic indentation. Mater Sci 1979;14: 1662–8. [24] Pharr GM. Measurement of mechanical properties by ultra-low load indentation. Mater Sci Eng A 1998;253:151–9.
[25] Busscher HJ, de Jong HP, Arends J. Surface free energies of hydroxyapatite, fluorapatite and calcium fluoride. Mater Chem Phys 1987;17:553–8. [26] Ruys AJ, Wei M, Sorrell CC, et al. Sintering effects on the strength of hydroxyapatite. Biomaterials 1995;16:409–15. [27] Viswanath B, Raghavan R, Ramamurty U, Ravishankar N. Mechanical properties and anisotropy in hydroxyapatite single crystals. Scripta Mater 2007;57:361–4. [28] Seto J, Gupta HS, Zaslansky P, Wagner HD, Fratzl P. Tough lessons from bone: extreme mechanical anisotropy at the mesoscale. Adv Funct Mater 2008;18:1905–11. [29] Hofmann T, Heyroth F, Meinhard H, Franzel W, Raum K. Assessment of composition an anisotropic elastic properties of secondary osteon lamellae. J Biomech 2006;39:2282–94. [30] Roy ME, Rho JY, Tsui TY, Evans ND, Pharr GM. Mechanical and morphological variation of the human lumbar vertebral cortical and trabecular bone. J Biomed Mater Res 1999;44:191–7. [31] Swadener JG, Rho JY, Pharr GM. Effects of anisotropy on elastic moduli measured by nanoindentation in human tibial cortical bone. J Biomed Mater Res 2001;57:108–12. [32] Carando S, Barbos MP, Ascenzi A, Boyde A. Orientation of collagen in human tibial and fibular shaft and possible correlation with mechanical properties. Bone 1989;10:139–42. [33] Sanson GD, Kerr SA, Gross KA. Do silica phytoliths really wear mammalian teeth? J Archaeol Sci 2007;34:526–31. [34] Gross KA, Rodrı´guez-Lorenzo LM. Sintered hydroxyfluorapatites. Part II: Mechanical properties of solid solutions determined by microindentation. Biomaterials 2004;25:1385–94.
Acta Biomaterialia 5 (2009) 2206–2212 www.elsevier.com/locate/actabiomat
Micromechanical properties of single crystal hydroxyapatite by nanoindentation Saeed Saber-Samandari *, Ka¯rlis A. Gross Department of Mechanical Engineering, The University of Melbourne, Parkville, Vic. 3010, Australia Swinburne University of Technology, Faculty of Engineering and Industrial Sciences, IRIS, Hawthorn, Vic. 3122, Australia Received 13 October 2008; received in revised form 4 February 2009; accepted 5 February 2009 Available online 11 February 2009
Abstract Knowledge of the intrinsic properties of hydroxyapatite (HAp) single crystals is important for the design of natural systems and will assist further improvements of manufactured biomaterials. Nanoindentation provides a useful tool for determining mechanical properties such as the hardness, elastic modulus and fracture toughness of small samples. A Berkovich indenter was placed on the side and basal planes of a natural single crystal of Durango HAp. The hardness and elastic modulus values obtained revealed higher values for the base (7.1 and 150.4 GPa) compared to the side (6.4 and 143.6 GPa). The cracking threshold, i.e., the load at which cracking initiates, revealed earlier crack formation on the base (at 8 mN) compared to the side (at 11 mN). Fracture toughness was measured as 0.45 ± 0.09 and 0.35 ± 0.06 MPa m1/2 for the side and basal plane, respectively. These results suggest that crystals are less prone to cracking and resist microcrack events better on the side, which is useful in bone, while exposing the base, the hardest face, to minimize mass loss from abrasion with teeth. Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Hydroxyapatite; Single crystal; Mechanical properties; Nanoindentation
1. Introduction Details about the bond strength and elasticity within crystals can be best obtained from single crystals, which can provide important information on the hardness, strength and elastic modulus. Since apatite is more commonly found as nanocrystals, special techniques need to be employed to produce a single crystal of apatite. An alternative is to source naturally occurring single crystals of apatite. Apatite used in biomedical applications is always available as micron- or nanometre-sized crystals of hydroxyapatite (HAp). Various techniques have been employed to prepare the nano-sized HAp, including precipitation, hydrothermal [1–3], sol–gel [4], crystal conversion [5] and microwave techniques [6]. Recently, new techniques have *
Corresponding author. Tel.: +61 3 8344 0024. E-mail address: [email protected] (S. Saber-Samandari).
evolved to prepare micro-sized HAp [7,8]. The crystal size, shape, chemical composition, defect concentration and assemblage all depend on the manufacturing technology. As a result, it is difficult to distinguish the contributions of the microstructure, chemistry, morphology, charge and residual stress to properties such as hardness. Studies on single crystals should be able to determine the base properties, and serve as an important reference for other studies that intend to improve on these properties of apatites. Natural apatite in bone has elongated crystals incorporated into collagen. Bone cements, used as bone fillers, also have long crystals. Recent advances in thermal spraying have revealed the ability to create oriented single crystals [9]. The properties of these materials depend on the orientation of the crystals. Testing on such fine crystals is challenging; an easier approach would be to conduct testing on large single crystals. This would not only provide values for mechanical properties of the side (parallel to the c-axis), but also provide enough material for testing of the base
1742-7061/$ - see front matter Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2009.02.009
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
(parallel to the a-axis). Such information could form the basis for understanding the effect of crystal orientation in the different HAp products used in the medical industry. Large-sized single crystals of HAp are required for mechanical property testing. These can be found in millimeter sizes in nature at Cerro de Mercado, close to Durango in Mexico. No experimental study on the micromechanical properties of a natural HAp single crystal has been reported. The goal of this work was therefore to employ nanoindentation techniques to probe such mechanical properties as hardness, elastic modulus and fracture toughness. Nanoindentation would be performed on the side and the basal plane of the sample to confirm the effect of orientation on its mechanical properties. To this end, a natural single crystal of HAp from Durango was selected. It is believed that this is the first comprehensive study of single crystal HAp. 2. Experimental procedure The single crystal was embedded in epoxy resin to provide support for cutting the cross-section and metallographically polishing it to produce the reasonably flat surface required for nanoindentation testing. Both a cross-section and longitudinal section were produced to assess the mechanical properties relative to the crystal faces. Fine grinding was conducted with silicon carbide abrasive paper of 2400 grit and polishing involved a 1 lm diamond suspension on a Dur surface (Struers, Denmark) for 5 min. Micro-Raman spectroscopy was used to confirm the purity of the sample (Fig. 1). Analyses were conducted with an excitation wavelength of 514 nm at a spectrum resolution of 1 cm1. Spectra were recorded within the range 500– 1500 cm1. A very narrow peak was found at 960 cm1 for the side and basal plane, confirming that the only calcium phosphate present was HAp [10]. The residual
impression from nanoindentation was observed using an XL30 Philips scanning electron microscope (SEM). The samples were coated with gold using an Edwards S150B sputter coater for examination with the SEM at 20 keV. The SEM images were also used for measuring the length of microcracks to calculate the fracture toughness. A Berkovich three-sided pyramidal indenter was used throughout this study. The area function for the diamond was calibrated by indentation in fused silica (FS) to determine the contact area for a given depth. The contact area as a function of depth was found to be Aðhc Þ ¼ 22h2c þ 2500hc by the polynomial-fitting procedure. The measured depth was then adjusted for the effect of instrument compliance (0.499 nm mN1) in the instrument software for analysis (NanoTest, MML, Wrexham, UK). The hardness and reduced elastic modulus of FS, assuming a Poisson’s ratio of 0.165, were measured to be 8.8 ± 0.1 and 69.6 ± 0.2 GPa, respectively. After calibration, the instrument was set up for two series of 10 indentations in both the side and the basal plane of the sample: one using low loads, ranging from 1 to 20 mN, and the other using high loads, ranging from 20 to 200 mN. The hardness and elastic modulus were measured using the standard method proposed by Oliver and Pharr [11]. The separation distance between the indents was selected based on the size of the residual impression (Berkovich shape). The spacing between adjacent indentations was about 15 times the maximum width of the indentations at low load and more than four times bigger than those at high load to ensure that the deformed region surrounding one indentation did not interfere with the formation of the next indentation. Since the maximum depth at the highest load (200 mN) was about 1.3 lm, indentations were spaced approximately 40 lm apart in a given row and the rows (low load and high load) were spaced at 100 lm. Indentations started approximately at the centre of the sample to avoid any testing near the edges. The loading/unloading rate was set to 10% of minimum load. The dwell periods at maximum load and for drift correction were 5 and 30 s, respectively. All experiments were performed at room temperature. Fracture toughness is another mechanical property that can be measured using nanoindentation techniques. For the calculation of the fracture toughness, the relationship between the fracture toughness and the length of the radial cracks using pyramidal sharp indenters was considered using the following equation [12] K IC
Fig. 1. Raman shifts of the natural single crystal HAp.
2207
1=2 E P ¼W H c3=2
where KIC is the mode I critical stress intensity factor, W is a material-independent constant that must be calibrated by indentation of well-characterized materials, H is the hardness, E is the elastic modulus, P is the peak load and c is the radius of the surface crack. The empirical constant was found to be 0.016 for Berkovich-type indenters [13].
2208
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
3. Results and discussion Loading–unloading curves of 10 indentations made on the base of the HAp single crystal are depicted in Fig. 2, where the load varied from 20 to 200 mN. This clearly reveals that the sample is homogeneous. Shown in Fig. 3 is the hardness and reduced elastic modulus of both the side and the basal plane of the single crystal HAp as a function of load. Plots illustrate that applying a 30 mN load or greater allows determination of the true hardness (a hardness that remains unchanged from the size of the indentation [14]). The indentation size effect was observed at low load, where hardness decreases with increasing load up to 30 mN. Hardness and reduced elastic modulus as functions of indentation depth are depicted in Fig. 4. It can be inferred that true hardness and reduced elastic modulus could be determined with a penetration depth of more than 500 nm (Fig. 4). The indentation size effect was also noticed at penetration depths shallower than 500 nm. The measured moduli are expected to be constant with indentation depth (or load); however, results reveal an approximately 5% overestimation with loads less than 30 mN or depths shallower than 500 nm. The discrepancy between the expected and obtained results could be due to the pile-up phenomenon. Oliver and Pharr’s proposed method is best applied when no pile-up but some sink-in occurs when the sample is indented. In the case of pileup, contact depth is always more than the total depth of indentation, thus the projected area is always underestimated. This underestimated area could overestimate both hardness and reduced elastic modulus, with a more significant effect on hardness. The reduced elastic modulus and hardness are calculated with the following expressions: 2S pffiffiffi Er ¼ pffiffiffi A p P max H¼ A
Fig. 3. The hardness and reduced elastic modulus vs applied load results from nanoindentation tests on both the side and basal planes of single crystal HAp.
Fig. 4. The hardness and reduced elastic modulus vs indentation depth results from nanoindentation tests on both the side and basal planes of single crystal HAp.
Fig. 2. Load–displacement curves for 10 indentations on the base of the single crystal of HAp. The applied load varies from 20 to 200 mN.
Recent nanoindentation studies on single crystals of different materials (i.e., metals, semiconductors, ceramics and oxides) show pop-in phenomena (plateau behaviour in loading) from applying low loads [15–17]. Interestingly, other nanoindentation studies on single crystals that have a hexagonal close-packed (hcp) lattice structure (i.e., sapphire, GaN and ZnO) show multiple load plateau behaviour [18–20]. HAp has a hexagonal lattice structure and therefore multiple load plateau behaviour is expected. Load plateau behaviour could be a result of a microcrack [21], a pore underneath the indenter [22] or the pile-up phenomenon. In the case of single crystals, existence of pores has not been reported. Therefore, surface topography
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
tools, such as atomic force microscopy and SEM at high magnification, are needed to reveal the existence of pileup and/or microcracks. Fig. 5A shows a loading–unloading curve obtained from nanoindentation on the side of the sample under a 7.5 mN load. Discontinuities on the loading curve are observed as expected, since HAp has an hcp crystal structure. In addition, an SEM image of the residual indentation at very high magnification (120,000) reveals no crack, but deformation at the edges of the impression (Fig. 5B). Clearly, pile-up only occurs during loading. However, multiple pop-in events are visible in Fig. 5C when the applied load is increased to about 11 mN. A SEM image of its corresponding loading–unloading curve reveals both microcracks and deformation at the edges (Fig. 5D). It was therefore concluded that the cracking threshold for single crystal HAp is about 11 mN on the side face using a Berkovich indenter. It is noted that, with a Vickers indenter, cracking thresholds in most ceramic materials are about 250 mN or more [23]. Significantly lower thresholds can be achieved using the cube-corner indenter [24], but no report on the cracking threshold of ceramic materials using Berkovich (the most commonly used) indenter are known.
2209
It is interesting to compare the cracking threshold of the side with that of the base. Fig. 6A and B show the loading– unloading curve and its corresponding SEM image of the indentation for the base of the single crystal HAp. Pop-in events are observed with a 5 mN load (Fig. 6A). Fig. 6B reveals no microcrack but a few occurrences of pile-up. In Fig. 6C, multiple pop-in events are evident under a maximum load of 8 mN, while its corresponding SEM image reveals both pile-up and three microcracks (Fig. 6D). As a result, the cracking threshold of the single crystal HAp on its basal plane is approximately 8 mN. Therefore, cracks propagate sooner on the base of the HAp single crystal than on its side, indicating that the base is more brittle. Overloading the pyramidal Berkovich indenter in this study causes crack propagation around the indent. Observation and length measurement of such cracks allow the fracture toughness to be determined. In this study, the average crack lengths were taken since the crack lengths were found to be anisotropic (Fig. 7A). The fracture toughness values (average of 12 measurements) obtained for the side and basal planes were 0.45 ± 0.09 and 0.35 ± 0.06 MPa m1/2, respectively. The side is tougher than the base but has lower hardness and elastic modulus values, (Table 1).
Fig. 5. Nanoindentation on the side of single crystal Hap. (A) Load–unload curve showing pop-in events; (B) an SEM image (120,000) showing no microcrack, with a maximum load of 7.5 mN; (C) multiple pop-in events from an 11 mN applied load; and (D) the corresponding SEM image showing pile-up and three microcracks.
2210
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
Fig. 6. Nanoindentation made on the base of single crystal Hap. (A) Load–unload curve showing pop-in events using a maximum load of 5 mN; (B) an SEM image (120,000) showing pile-up on the top two sides; (C) multiple pop-in events from an 8 mN applied load; and (D) the corresponding SEM image with pile-up and three microcracks.
Shown in Fig. 7 are residual indentation SEM images of the HAp single crystal with cracked corners from application of an 80 mN load on the side and basal plane. Fig. 7A shows a crack length of 7.73 lm on the side, while a larger crack length of 8.09 lm is observed on the base, confirming the lower fracture toughness of the base. Anisotropy found in crack lengths could be due to anisotropy in the surface energy, as the base of the HAp crystal has a higher surface energy than the side [25]. Moreover, a polycrystalline sample shows a higher fracture toughness than a single crystal due to grain boundaries, which reduce the crack growth. The fracture toughness for polycrystalline HAp was reported to be between 0.6 and 1.5 MPa m1/2 [26]. The results obtained in this study are in good agreement with the values of fracture toughness for the single crystal HAp reported by Viswanath et al. [27]. They reported the mechanical properties of manufactured single crystal HAp below pop-in events and found fracture toughness values of 0.48 and 0.28 MPa m1/2 for the side and basal plane, respectively. These results outline the anisotropic mechanical properties of single crystal HAp. Since about 70% of bone is made up of the inorganic mineral HAp, anisotropy in the mechanical properties of bone is expected. This behaviour has been observed in fibrolamellar bone [28], single osteon
lamellae in bone tissue [29], human lumbar vertebral cortical and trabecular bone [30], and human tibial cortical bone [31]. Most studies on bone show that differences in mechanical properties may be due to collagen fibre orientation [32]. This study reveals the contribution of HAp orientation to mechanical properties. The assembly of HAp crystals will influence the properties of bone. The arrangement of the crystals aligned at an angle to the long bone in the femur will provide 5% greater stiffness. This is not a large difference, but it does contribute to the anisotropy. Crystal geometry is also an important factor. The small surface area of the rod end lowers the potential for crack formation along the c-axis and positions the crystals where they may be more effective in bending. More energy will be required to form a crack in the side of a rod. The greater resistance to crack formation on the side of the crystal will influence the microcrack path that is formed in bone. Teeth contain the highest concentration of apatite crystals within the body, and have the side of the crystal towards the outer surface of the tooth. This presents the crystal side for easier remodeling, an event that is continually taking place in the aqueous environment of the oral cavity. The higher hardness of the crystal side indicates that nature has optimized the crystal orientation to minimize
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
2211
were determined to be 11 and 8 mN, respectively. The results of these experiments indicated higher hardness and elastic modulus values for the base (by about 11% and 5%, respectively) than for the side. Unlike the hardness and elastic modulus, fracture toughness was found to be higher (approximately 25%) for the side, as longer crack lengths were observed on the base compared to the side under the same applied load. Acknowledgements The authors acknowledge support from an ARC Project # DP0774251. The nanoindenter was purchased with funds provided by the Rowden White Fund and supplemented by the ARC LIEF Grant No. LE0775715. The single crystal was kindly provided by Prof. Andrew Gleadow and Assoc. Prof. Barry Kohn from the School of Earth Sciences, University of Melbourne. This work has partially evolved from a past Ramaciotti Grant. References
Fig. 7. Residual indentation image of single crystal HAp with cracked corners from application of an 80 mN load on (A) the side, with a 7.73 lm crack length; and (B) the base, with 8.09 lm crack length.
Table 1 Mechanical properties of single crystal HAp. Orientation
H (GPa)
Er (GPa)
E (GPa)
KIC (MPa m1/2)
Side Base
6.41 7.06
137.98 142.92
143.56 150.38
0.45 0.35
abrasion. The abrasion from opposing teeth and third body wear from mineral inclusions in foods will thus be minimized [33]. The additive effect of strengthening from fluoride [34] and from crystal orientation shows the benefits of separate microstructural elements in optimizing materials for their function. Similarly, the crystal shape, size and arrangement will be important for HAp materials used in biomedical applications. The properties of the single crystal will provide important reference information for the assessment of presently used HAp materials and support for further improvements. Further work will determine the mechanical properties of HAp used in biomedical implants. 4. Conclusions Thresholds for crack-initiation on the side and basal planes of single crystal HAp using a Berkovich indenter
[1] Ito A, Nakamura S, Aoki H, et al. Hydrothermal growth of carbonate-containing hydroxyapatite single crystals. J Crys Growth 1996;163:311–7. [2] Yoshimura M, Suda H. Hydrothermal synthesise of biocompatible whiskers. Mater Sci 1994;29:3399–402. [3] Liu J, Ye X, Wang H, et al. The influence of PH and temperature on the morphology of hydroxyapatite synthesized by hydrothermal method. Ceram Inter 2003;29:629–33. [4] Jillavenkatesa A, Hoelzer DT, Condrate RA. An electron microscopy study of the formation of hydroxyapatite through sol–gel processing. Mater Sci 1999;34:4821–30. [5] Zhan J, Tseng Y. Biomimetic formation of hydroxyapatite nanorods by a single-crystal-to-single-crystal transformation. Adv Funct Mater 2005;15:2005–10. [6] Anabel L, Jaime G, Rafael R. Nanosized hydroxyapatite precipitation from homogenous calcium/citrate/phosphate solutions using microwave and conventional heating. Adv Mater 1998;10:49–53. [7] Madsen H, Christensson F, Polyak LE, et al. Calcium phosphate crystallization under terrestrial and microgravity conditions. J Cryst Growth 1995;152:191–202. [8] Teng S, Shi J, Chen L. A novel method to synthesize large-sized hydroxyapatite rods. J Cryst Growth 2006;290:683–8. [9] Saber-Samandari S, Gross KA. Nanoindentation reveals mechanical properties within thermally sprayed hydroxyapatite coatings. Surf Coat Technol 2009;203:1660–4. [10] Gross KA, Phillips MR. Identification and mapping of the amorphous phase in plasma-sprayed hydroxyapatite coatings using scanning cathodoluminescence microscopy. J Mater Sci: Mater Med 1998;9:797–802. [11] Oliver WC, Pharr GM. Measurement of hardness and elastic modulus by instrumented indentation: advances in understanding and refinements to methodology. J Mater Res 2004;19:3–20. [12] Lawn BR, Evans AG, Marshall DB. Elastic/plastic indentation damage in ceramics: the median/radial crack systems. J Am Ceram Soc 1980;63:574–81. [13] Pharr GM, Harding DS, Oliver WC. Measurement of fracture toughness in thin films and small volumes using nanoindentation methods. In: Nastasi M, Parkin DM, Gleiter H, editors. Mechanical properties and deformation behavior of materials having ultra-fine microstructures. Dordrecht: Kluwer Academic Press; 1993. p. 449–61. [14] Manika I, Maniks J. Size effects in micro- and nanoscale indentation. Acta Mater 2006;54:2049–56.
2212
S. Saber-Samandari, K.A. Gross / Acta Biomaterialia 5 (2009) 2206–2212
[15] Gerberich WW, Nelson JC, Lilleodden ET, Anderson P, Wyrobek JT. Indentation induced dislocation nucleation: the initial yield point. Acta Mater 1996;44:3585–98. [16] Van Vliet KJ, Li J, Zhu T, Yip S, Suresh S. Quantifying the early stages of plasticity through nanoscale experiments and simulations. Phys Rev B 2003;67:104105. [17] Kearney C, Zhao Z, Bruet BJF, et al. Nanoscale anisotropic plastic deformation in single crystal aragonite. Phys Rev Lett 2006;96: 2555051–54. [18] Page TF, Oliver WC, McHargue CJ. The deformation behavior of ceramic crystals subjected to very low load (nano)indentations. J Mater Res 1992;7:450473. [19] Bradby JE, Kucheyev SO, Williams JS, et al. Indentation-induced damage in GaN epilayers. Appl Phys Lett 2002;80:383–5. [20] Kucheyev SO, Bradby JE, Williams JS, Jagadish C, Swain MV. Mechanical deformation of single-crystal ZnO. Appl Phys Lett 2002;80:956–8. [21] Fischer-Cripps AC. Nanoindentation. New York: Springer-Verlag; 2002. [22] Lu X, Wang X, Xiao P. Nanoindentation and residual stress measurements of yttria-stabilized zirconia composite coatings produced by electrophoretic deposition. Thin Solid Films 2006;494: 223–7. [23] Lankford J, Davidson DL. The crack-initiation threshold in ceramic materials subject to elastic/plastic indentation. Mater Sci 1979;14: 1662–8. [24] Pharr GM. Measurement of mechanical properties by ultra-low load indentation. Mater Sci Eng A 1998;253:151–9.
[25] Busscher HJ, de Jong HP, Arends J. Surface free energies of hydroxyapatite, fluorapatite and calcium fluoride. Mater Chem Phys 1987;17:553–8. [26] Ruys AJ, Wei M, Sorrell CC, et al. Sintering effects on the strength of hydroxyapatite. Biomaterials 1995;16:409–15. [27] Viswanath B, Raghavan R, Ramamurty U, Ravishankar N. Mechanical properties and anisotropy in hydroxyapatite single crystals. Scripta Mater 2007;57:361–4. [28] Seto J, Gupta HS, Zaslansky P, Wagner HD, Fratzl P. Tough lessons from bone: extreme mechanical anisotropy at the mesoscale. Adv Funct Mater 2008;18:1905–11. [29] Hofmann T, Heyroth F, Meinhard H, Franzel W, Raum K. Assessment of composition an anisotropic elastic properties of secondary osteon lamellae. J Biomech 2006;39:2282–94. [30] Roy ME, Rho JY, Tsui TY, Evans ND, Pharr GM. Mechanical and morphological variation of the human lumbar vertebral cortical and trabecular bone. J Biomed Mater Res 1999;44:191–7. [31] Swadener JG, Rho JY, Pharr GM. Effects of anisotropy on elastic moduli measured by nanoindentation in human tibial cortical bone. J Biomed Mater Res 2001;57:108–12. [32] Carando S, Barbos MP, Ascenzi A, Boyde A. Orientation of collagen in human tibial and fibular shaft and possible correlation with mechanical properties. Bone 1989;10:139–42. [33] Sanson GD, Kerr SA, Gross KA. Do silica phytoliths really wear mammalian teeth? J Archaeol Sci 2007;34:526–31. [34] Gross KA, Rodrı´guez-Lorenzo LM. Sintered hydroxyfluorapatites. Part II: Mechanical properties of solid solutions determined by microindentation. Biomaterials 2004;25:1385–94.