- Email: [email protected]
Characterization of immersed hydroxyapatite-bioactive glass coatings in Hank’s solution
Materials Chemistry and Physics 64 (2000) 229–240
Characterization of immersed hydroxyapatite-bioactive glass coatings in Hank’s solution J.H. Chern Lin1,∗ , H.J. Lin, S.J. Ding, C.P. Ju Department of Materials Science and Engineering, National Cheng Kung University, Tainan 70101 Taiwan Received 10 November 1999; accepted 15 December 1999
Abstract Monolithic HA and hydroxyapatite-bioactive glass (HA–BG) coatings have been plasma-sprayed on Ti–6A1–4V substrate using sinter-granulated HA and HA–BG powders. The immersion behavior of these coatings in Hank’s physiological solution was investigated. The results showed that sinter-granulated monolithic HA powder was irregularly-shaped and appeared quite dense. When 5 wt.% or more BG was added in HA, the powder became rough and porous. X-ray diffraction patterns showed that the presence of BG enhanced the decomposition of HA structure during fabrication of the powder. The plasma spray process itself also enhanced the decomposition of apatite. Surface morphology of the present HA–BG coatings was not much different from that of monolithic HA coating, that comprised patches of smooth and shiny glassy film and irregularly-shaped particles on its surface. Reasonably high bond strengths were obtained from all as-plasma-sprayed coatings. When immersed in Hank’s solution, these bond strengths largely declined. © 2000 Published by Elsevier Science S.A. All rights reserved. Keywords: Hydroxyapatite; Bioactive glass; Plasma spray; Bond strength; Immersion
1. Introduction Due to its favorable biocompatibility and mechanical properties, hydroxyapatite (HA)-coated Ti–6A1–4V alloy has been accepted as one of the most promising implant materials for orthopaedic and dental applications [1–7]. Various deposition techniques, such as dip coating-sintering, immersion coating, hot isostatic pressing (HIP), ion plating, sputtering, pulsed laser deposition and plasma spray, have been explored [3–7]. Among the many techniques, plasma spray is perhaps most popularly used due to its process feasibility as well as reasonably high coating bond strength and mechanical properties. Several attempts have been made to improve the performance (e.g. bioactivity [8] and mechanical properties) of bulk HA by incorporating bioactive glass (BG) or other second phase particles into HA. For example, Mochida et al. [9] developed an apatite-bioactive glass composite (ABC) material by sintering the mixture of HA and 45SF 1/4 bioactive glass. The composition of 45SF 1/4 was similar to that of 45S5 Bioglass except that 6.7 mol% CaO in 45S5 Bioglass was substituted with CaF2 . Due to the high tempera∗ Corresponding author. Tel.: +886-6-2748086; fax: +886-6-2748086. E-mail address: [email protected] (J.H. Chern Lin) 1 Professor
ture (1000◦ C) interaction between the BG and HA phases, the incorporation of HA into 45SF1/4 caused the glass phase to crystallize. Knowles and co-workers [10,11] found that phase transformations of HA into Ca5 (PO4 )2 SiO4 , ␣-Ca3 (PO4 )2 (␣-TCP) and -Ca3 (PO4 )2 (-TCP) took place when small amounts (up to 5 wt.%) of phosphate-based glasses were added. Such transformations were more profound in the HA containing soda type glass (Na2 O–P2 O5 ) than that containing lime type glass (CaO–P2 O5 ). A strengthening effect was observed when 2.5 wt.% CaO–P2 O5 glass was added in the HA resulting in an HA/␣-TCP sintered body. In their study of thermal stability of sintered HA-doped glass (SiO2 –CaO–Na2 O–P2 O5 –A12 O3 –B2 O3 ), Kangasniemi et al. [12] showed that, when sintered at temperatures higher than 700◦ C, sodium ions diffused into doped HA particles inducing the phase transformation of HA into NaCaPO4 . These phase transformations led to changes in properties of the glass. The biodegradation phenomenon of HA in both in vitro and animal studies has been reported by Ramselaar et al. [13], Blitterswijk et al. [14], and others [6,15,16]. Kummer and Jaffe [6] evaluated the stability of a cyclically-loaded HA coating in in-vitro solutions and found increased instability of the HA coating. Reis et al. [15] suggested that degradation of their plasma-sprayed monolithic HA coating in Hank’s or
0254-0584/00/$ – see front matter © 2000 Published by Elsevier Science S.A. All rights reserved. PII: S 0 2 5 4 - 0 5 8 4 ( 0 0 ) 0 0 2 0 9 - 1
230
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Table 1 Composition of raw HA powder used in this studya Chloride Fluoride Sulfate Heavy metals Arsenic Iron Water Assay a
Table 2 Composition of Hank’s physiological solution used in this study <0.05 wt.% <0.005 wt.% <0.3 wt.% <0.003 wt.% <0.0002 wt.% <0.02 wt.% 1.7 wt.% 96.5%
Merck #2196; E. Merck, Germany.
NaCl solution under loading was not only due to loosening of solid particles, but also to coating dissolution. When the coating was subjected to a higher load in the solution, the reduction in coating thickness became larger. Yang et al. [16] found that bond degradation of ≈25–33% of the original strength was measured when the specimens were immersed in a pH-buffered, serum-added simulated body fluid. Chern Lin et al. [7,17] studied a series of HA–BG composite coatings plasma-sprayed on Ti6Al4V substrate. The powders used for plasma spray were prepared by directly mixing in-house fabricated BG powder and a commercial HA powder. Their results showed that directly mixed type
Fig. 1. Schematic diagram of bond strength test for this study (ASTM C633-79).
Constituent
g/l
NaCl NaHCO3 KCl MgCl2 ·6H2 O CaCl2 Na2 HPO4 ·2H2 O MgSO4 ·7H2 0 KH2 PO4 Glucose
8.00 0.35 0.40 0.10 0.14 0.06 0.06 0.06 1.00
HA–BG composite coatings were generally more resistant to moisture attack than monolithic HA coating in terms of bond strength degradation. In a recent study, Ding et al. [18] found that the dissolution depth of plasma-sprayed coatings was largely dependent on the type and composi-
Fig. 2. LVSEM micrographs of various powders fabricated in-house for plasma spray.
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
231
Fig. 3. XRD patterns of various powders for plasma spray (䊊: apatite; A: ␣-TCP; B: -TCP).
tion of the coating. The granulated type HA–BG coatings were less dissolvable in simulated body fluid than monolithic HA or mixed type HA–BG coatings. In the present study, sinter-granulated HA and HA–BG coatings have been plasma-sprayed on Ti6Al4V substrate. The structure, bond strength, and immersion behavior in Hank’s physiological solution of these coatings were investigated.
Fig. 5. LYSEM micrographs of various as-plasma-sprayed coatings.
2. Materials and methods
Fig. 4. XRD patterns of various plasma-sprayed coatings (䊊: apatite; T: TTCP; C: CaO; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
A commercial HA powder (Merck A.G., Darmstadt, Germany) with particle size <1 m was used in this study. The composition of this powder is listed in Table 1. The in-house fabricated BG used in this study contained frits of SiO2 , Na2 O, CaO and P2 O5 (SiO2 :Na2 O:CaO:P2 O5 =10:3:5:2 in weight). In preparing the BG, appropriate amounts of the frits were mixed and melted at 1350◦ C for 2 h. The melt was then quenched in water and fractured fragments of the glass were collected. Detailed description of the fabrication of the BG powder has been given in an earlier paper [7]. The HA–BG powders were prepared using sintergranulation method [18]. The sinter-granulation method involved mixing and ball-milling appropriate amounts of HA and BG powders (HA:BG=95:5 and 92.5:7.5 in weight ratio) in ethyl alcohol, followed by drying and sintering at 1300◦ C for 24 h. The sintered granules were then crushed,
232
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 6. Cross-sectional LVSEM micrographs of various as-plasma-sprayed coatings.
milled and sieved to obtain particles with sizes 44–149 m, a particle size range recommended by the manufacturer of the plasma spray system. Throughout this study the sinter-granulated powders and the plasma-sprayed coatings produced from such powders were designated as the same code. Commercially available cylindrical rods and plates of Ti6Al4V alloy were used as the substrate material. Prior to plasma spray, the substrate surface was mechanically polished to #1200 grit level, cleaned and sandblasted with 450 m SiC particles. To obtain a uniform coating, the substrate was mounted on a disk which could rotate during spraying in air using a Plasma-Technik A-3000 system (Plasma Technik A.G., Wohien, Switzerland). The coatings had a thickness in the range of 100–140 m. The ASTM C633-79 method suggested for testing flame-sprayed coating bond strength was used to measure the bond strength of the present coatings, as shown in Fig. 1. In performing this test, two identical cylindrical Ti6Al4V rods were prepared, one with coating on a flat surface and the other without. The flat surface of the
uncoated rod, which was to be bonded to the coated rod, was sandblast-roughened to enhance resin adherence. A thin layer of bonding glue (Plasma Technik, Klebbi glue, Switzerland) was applied onto both coated and uncoated surfaces. After the two rods were carefully aligned using a special device [17], a compressive stress was applied to both rods during curing to assure an intimate contact between resin and the two surfaces. After curing at 180◦ C in an oven, the bonded rods were bench cooled to room temperature. The pressure was then released and the resin-bonded rods were removed from the device. The coatings were pulled out from substrate using an Instron 8562 tester at a cross head speed of 1.0 mm/min. In the present study, a Hank’s physiological solution (Table 2), recommended by Pourbaix [19] for testing corrosion behavior of implant materials, was used for the immersion test. The coated specimens, each with a surface area of 1 cm2 , were immersed up to 30 days in vials containing 50-ml Hank’s solution, which were maintained at 37◦ C throughout tests. The solution was agitated daily to help maintain uniform ion concentrations. After predetermined
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 7. XRD patterns of monolithic HA coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; C: CaO; T: TTCP).
time periods of immersion, the specimens were removed from the vials and stored in a desiccator for characterization. Phases of the various powders and coatings before and after immersion were analyzed by an X-ray diffractometer (Rigaku D-max IIIV, Japan) with Ni-filtered CuKα radiation operated at 30 kV and 20 mA with a scanning speed
233
Fig. 9. XRD patterns of 92.5HA/7.5BG coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
of 1◦ /min. Microstructure of the specimens were examined using a low vacuum scanning electron microscope (Topcon SM-300, Tokyo, Japan). Using this low vacuum SEM (LVSEM), specimens could be observed under a ‘wet’ condition and the conventional deposition of a conducting film was not needed, thus largely eliminating a variety of difficulties such as high vacuum and charging-induced structural damage. Microchemistry was characterized using an energy dispersive spectroscopy (EDS) system. Specimens for cross-sectional microscopy were prepared by mounting the coated specimens in epoxy resin, followed by polished through 0.3 m alumina. More important, the bond strength variation as a function of immersion time in Hank’s solution was evaluated by means of ASTM C633-79 method. One-way ANOVA statistical analysis was used to evaluate the statistical significance of bond strength data. Scheff multiple comparison test was used to determine the significance of the deviations in bond strength of each coating for different immersion times. In all cases, the results were considered statistically different with p<0.05.
3. Results and discussion 3.1. Characterization of HA and HA–BG powders
Fig. 8. XRD patterns of 95HA/5BG coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
As shown in Fig. 2, the sinter-granulated monolithic HA powder was irregularly-shaped and quite. When 5 wt.% BG was added in HA (95HA/5BG), the powder became porous. The majority of the pores had sizes smaller than 10 m.
234
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 10. LVSEM micrographs of monolithic HA coating after immersion in Hank’s solution.
Fig. 11. LVSEM micrographs of 95HA/5BG coating after immersion in Hank’s solution.
Kangasmiemi et al. [12] attributed the BG-induced formation of porosity to the release of hydroxyl group from HA. Wang et al. [20] confirmed that the presence of BG not only enhanced the breakdown process of hydroxyl group, but also enhanced the decomposition of HA structure. It seems
that at least two factors, i.e., BG-enhanced release of hydroxyl group and BG-induced liquid phase sintering effect [21], were operating competitively to determine the eventual porosity level of the sinter-granulated HA–BG powders. When 7.5 wt.% BG was added (92.5HA/7.5BG), the powder
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
235
Fig. 12. LVSEM micrographs of 92.5HA/7.5BG coating after immersion in Hank’s solution.
surface revealed many microcracks on the surface. The formation of the microcracks might be explained by the differences in density and coefficient of thermal expansion (CTE) among the various phases that involved phase transformations at high temperatures [18]. X-ray diffraction (XRD) patterns of the various powders are shown in Fig. 3. As shown in the figure, HA was a highly crystalline phase in accordance with JCPDS No. 9-432 file. When sinter-granulated powder contained 5 wt.% BG, the decomposition of apatite into -TCP and ␣-TCP phases was readily detected. As BG content increased (92.5HA/7.5BG), the apatite intensity decreased, while the intensities of ␣-TCP and -TCP phases increased. It was suggested that, when an HA–BG mixture was sintered at elevated temperatures, the sodium ions in BG might diffuse into the HA structure, causing release of a portion of hydroxyl group from HA. The HA structure with deficient hydroxyl group became less stable at high temperatures and had greater tendency to decompose [11,12,18]. The present XRD results seem to support this theory.
3.2. Structural characterization of as-plasma-sprayed HA and HA–BG coatings As shown in the XRD patterns in Fig. 4, all as-plasmasprayed coatings had lower apatite crystallinity than their respective initial powders, in agreement with previous studies [17,18]. In addition to the broadening of apatite peaks, Ca4 P2 O9 (TTCP), CaO, ␣-TCP and -TCP phases were
observed in monolithic HA coating. The high temperature involved in plasma spray process had obviously enhanced the decomposition of apatite as well as chemical reactions among different phases. The intensities of ␣-TCP and -TCP in 95HA/5BG coating were larger than those in monolithic HA coating. A large amount of sodium calcium silicate (Na2 CaSiO4 ) was observed. As BG content increased to 7.5 wt.%, ␣-TCP, -TCP and Na2 CaSiO4 continued to increase at the expense of the apatite phase [18]. The formation of sodium calcium silicate might result from the reactions between apatite and such ions as silicon and sodium in glass. Broad face and cross-sectional LVSEM micrographs of as-plasma-sprayed HA and HA–BG coatings are shown in Figs. 5 and 6, respective. Randomly distributed pores of different sizes and numerous cracks could be observed on all coated surfaces. Patches of smooth and shiny glassy film and irregularly-shaped particles were observed on the surface of HA coating, as shown in Fig. 5. In their transmission electron microscopic (TEM) study, Ji et al. [22] showed that the plasma spray-induced glassy structure was an amorphous form of calcium phosphate. During plasma spraying, small size powder was completely melted, while larger size powder, such as the one used in the present study, was only partially melted resulting in a coating morphology comprising the observed irregularly-shaped particles [23,24]. The overall broad face surface morphology of both HA–BG coatings was not much different from that of monolithic HA coating. The thickness of monolithic HA, 95HA/5BG, and 92.5HA/7.5BG coatings were roughly 115–145, 100–130,
236
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 13. Cross-sectional LVSEM micrographs of monolithic HA coating after immersion in Hank’s solution.
and 100–130 m, respectively. The rather large deviations in thickness are quite common for plasma-sprayed due to their surface roughness. 3.3. Structural characterization of immersed HA and HA–BG coatings The XRD patterns of HA, 95HA/5BG, and 92.5HA/ 7.5BG-coated Ti–6A1–4V after a series of immersion times are shown in Figs. 7–9. When immersed in Hank’s solution, the intensities of such phases as CaO, ␣-TCP, -TCP, and TTCP in monolithic HA and 95HA/5BG coatings (Figs. 7 and 8) gradually decreased, indicating either the disso-
lution of these phases or the deposition of an amorphous apatite phase. The major apatite peaks at 2θ roughly between 32 and 33◦ were slightly broadened as a result of the deposition of amounts of amorphous apatite phase after immersion [7,18,25,26]. The XRD patterns of immersed 92.5HA/7.5BG coating (Fig. 9) were somewhat different from that of monolithic HA or 95HA/5BG coating. After immersion for 15 days or longer, the intensities of ␣-TCP and apatite increased significantly. The intensity of Na2 CaSiO4 intensities was still significant due to its large quantity in the coating. The broad face LVSEM micrographs of immersed HA, 95HA/5BG, 92.5HA/7.5BG coatings are shown respectively
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
237
Fig. 14. Cross-sectional LVSEM micrographs of 95HA/5BG coating after immersion in Hank’s solution.
in Figs. 10–12, while the cross-sectional micrographs of the three immersed coatings are shown respectively in Figs. 13–15. As shown in Fig. 10, the surface morphology of the HA coating immersed for 1 day was similar to that of as-sprayed coating. After 7 days of immersion, the glassy phase and numerous irregularly-shaped small particles on the surface of as-sprayed monolithic HA coating were no longer present. Instead, the surface started to be covered with clusters of precipitated apatite spherulites (indicated by the arrows as shown in Fig. 13). Higher magnification examination revealed that most of the spherulites had sizes of several microns. It appears that, during the immersion test, dissolu-
tion of the coating surface and deposition of the Ca/P-rich layer had both taken place, in agreement with the XRD results. Similar apatite spherulites ca. 2–3 m in diameter had also been observed by Andersson and Kangasniemi [27] in their immersed 45S5 glass in SBF for 8 h and by Chern Lin et al. [7]. According to Hench and Clark [25] and Kokubo et al. [26], a Ca/P-rich amorphous layer could be formed in vivo or in vitro on the surface of a bioactive glass, ceramic, or glass-ceramic. This Ca/P-rich layer was later crystallized to form small carbonate-containing HA crystals [25,26]. Those cracks observed on the coating surface were a result of drying process.
238
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 15. Cross-sectional LVSEM micrographs of 92.5HA/7.5BG coating after immersion in Hank’s solution.
As 5 wt.% BG was added in HA, the apatite spherulites were not observed until after 15 days of immersion (Fig. 11). When the BG content increased to 7.5 wt.%, these spherulites were not observed even after 30-day immersion (Fig. 12). The precipitated apatite layer could be distinguished more clearly from the original coating layer in cross-sectional LVSEM micrographs, as those indicated by arrows in Figs. 13 and 14. Such a precipitated layer with a distinct contrast was not observed in the 92.5HA/7.5BG coating even after immersion for 30 days, confirming the broad face examination.
3.4. Bond strengths of HA and HA–BG coatings Although the ASTM C633 method has often been used for measuring the bond strength of HA coatings, whether the measured bond strength values using this method could represent ‘true’ bond strengths is still in debate [23,28,29]. For example, it was argued [28] that, during press-curing, the bonding resin might infiltrate through open porosity into coating interior, introducing uncertainties to the measured values. Another difficulty arises from the determination of the ‘true’ fracture area. Post-test observations indicated that
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
239
Fig. 16. Bond strengths of various plasma-sprayed coatings after immersion in Hank’s solution.
all major fractures of the present coated specimens occurred within coating interior. In many cases only small portions of coatings were pulled out right from substrate surface. The major portions of the coatings were still bonded to the substrate when the coatings were pulled out. For simplicity, in this study, the original cross-sectional area of Ti–6A1–4V substrate was used at all times in the calculation of bond strengths. Compared to reported values [1,7,18,23], reasonably high bond strengths were obtained from all present as-plasma-sprayed coatings (in the range 50–60 MPa), as shown in Fig. 16. One-way ANOVA statistical analysis showed that there were no significant differences in bond strength among these as-sprayed coatings (p=0.05). When immersed in Hank’s solution, however, all three coatings largely lost their bond strengths. The statistical analysis using Scheffé multiple comparison test showed that the bond strength of monolithic HA coating significantly declined from 53.6 MPa, the as-coated bond strength, down to 21.8 MPa with a reduction of ca. 59% after immersion for Table 3 Ca/P ratios of as-coated and tensile-fractured surfaces of coated specimens immersed in Hank’s solution for 30 days Coating
As-coated surface
Fractured surface
HA 95HA/5BG 92.5HA/7.5BG
2.40±0.01 2.51±0.04 2.46±0.12
1.60±0.08 1.76±0.12 1.84±0.13
7 days (p<0.001). 95HA/5BG and 92.5HA/7.5BG coatings lost 62 and 55% in bond strength, respectively, after 7-day immersion. This deterioration in bond strength seems unavoidable for plasma-sprayed HA-based coatings immersed in simulated body fluid and also observed in other studies [6,15,16]. A question was raised regarding whether this immersioninduced degradation in bond strength could arise from a poor bonding between the plasma-sprayed coating and the precipitated apatite layer. To answer this question, the post-test fracture surfaces of three coated materials immersed for 30 days were examined and Ca/P ratios were measured. The observation indicated that all major failures of the immersed specimens occurred within coating interior). As shown in Table 3, the Ca/P ratios of all fractured surfaces were much lower than those of as-coated surfaces. For example, the Ca/P ratios of monolithic HA coating before and after 30 days of immersion were 2.40 and 1.60, respectively; the Ca/P ratios of 95HA/5BG before and after immersion were 2.51 and 1.76, respectively; while the Ca/P ratios of 92.5HA/7.5BG before and after immersion were 2.46 and 1.84, respectively. The much higher Ca/P ratios (than the stoichiometric Ca/P ratio, 1.67, of HA) on as-sprayed coating surfaces were not surprising and also observed in other plasma-sprayed HA coatings [30] due to the fact that a large quantity of phosphorus was vaporized during the high temperature plasma spray process. The lower Ca/P ratios indicated that those failures probably occurred within the precipitated apatite layer with a Ca/P ratio closer to 1.67. The numerous drying-induced
240
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
cracks observed on immersed surfaces (Figs. 10–12) also suggest that it was very likely for the failures to occur within this poorly-bonded precipitated layer.
4. Conclusions 1. The sinter-granulated monolithic HA powder was irregularly-shaped and appeared quite dense. When 5 wt.% or more BG was added in HA, the powder became rough and porous. XRD patterns showed that the presence of BG enhanced the decomposition of HA structure during fabrication of the powders. As BG content increased, the apatite intensity decreased, while the intensities of ␣-TCP and -TCP phases increased. 2. The as-sprayed coatings had lower apatite crystallinity than their respective initial powders. The plasma spray process itself enhanced the decomposition of apatite. 3. Surface morphology of the present HA–BG coatings was not much different from that of the monolithic HA coating, that was composed of patches of smooth and shiny glassy film and irregularly-shaped particles on its surface. 4. Reasonably high bond strengths were obtained from all as-sprayed coatings. When immersed in Hank’s solution, the bond strengths of all coatings largely declined.
Acknowledgements The authors acknowledge with appreciation the support for this research by the National Science Council of the Republic of China under the contract No. NSC 86-2213-E-033 & 034 & 035. References [1] R.G.T. Geesink, K. de Groot, C.P.A.T. Klein, Clin. Orthop. 225 (1987) 147. [2] J.E. Lemons, Clin. Orthop. 235 (1988) 220. [3] W.R. Lacefield, Bioceramics: material characteristics versus in vivo behavior, Ann. N.Y. Acad. Sci. 523 (1982) 72. [4] C.K. Wang, J.H. Chern Lin, C.P. Ju, H.C. Ong, R.P.H. Chang, Biomaterials 18 (1997) 1331. [5] C.S. Kim, P. Ducheyne, Biomaterials 12 (1991) 461.
[6] F.J. Kummer, W.L. Jaffe, J. Appl. Biomater. 3 (1992) 211. [7] J.H. Chern Lin, M.L. Liu, C.P. Ju, J. Biomed. Mater. Res. 28 (1994) 723. [8] L.L. Hench, in: T. Yamamuro, L.L. Hench, J. Wilson (Eds.), Handbook of Bioactive Ceramics, Vol. I, Bioactive Glasses and Glass-ceramics, CRC Press, Boca Raton, FL, 1990, 7 pp. [9] M. Mochida, I. Fujiu, M. Ogino, Yogyo-Kyokai-Shi 95 (1987) 798. [10] J.C. Knowles, W. Bonfield, J. Biomed. Mater. Res. 27 (1993) 1591. [11] J.D. Santos, J.C. Knowles, R.L. Reis, F.J. Monteiro, G.W. Hastings, Biomaterials 15 (1994) 5. [12] I.M.O. Kangasniemi, K. de Groot, J. Wolke, Ö. Andersson, Z. Luklinska, J.G.M. Becht, M. Lakkisto, A. Yli-urpo, J. Mater. Sci. Mater. Med. 2 (1991) 133. [13] M.M.A. Ramselaar, F.C.M. Driessens, W. Kalk, J.D. de Wijn, P.J. van Mullen, J. Mater. Sci. Mater. Med. 2 (1991) 63. [14] C.A. van Blitterswijk, J.J. Grote, W. Kuijpers, W.Th. Daems, K. de Groot, Biomaterials 7 (1986) 137. [15] R.L. Reis, F.J. Monteiro, G.W. Hastings, J. Mater. Sci. Mater. Med. 5 (1994) 457. [16] C.Y. Yang, B.C. Wang, E. Chang, B.C. Wu, J. Mater. Sci. Mater. Med. 6 (1995) 258. [17] J.H. Chern Lin, M.L. Liu, C.P. Ju, J. Mater. Sci. Mater. Med. 5 (1994) 279. [18] S.J. Ding, C.P. Ju, J.H. Chern Lin, J. Mater. Sci. Mater. Med. (in press). [19] M. Pourbaix, Biomaterials 5 (1984) 122. [20] C.K. Wang, J.H. Chern Lin, C.P. Ju, Mater. Chem. Phys. 53 (1997) 138. [21] J.D. Santos, R.L. Reis, F.J. Monteiro, J.C. Knowles, G.W. Hastings, J. Mater. Sci. Mater. Med. 6 (1995) 348. [22] H. Ji, C.B. Ponton, P.M. Marquis, J. Mater. Sci. Mater. Med. 3 (1992) 283. [23] J.G.C. Wolke, C.P.A.T. Klein, K. de Groot, in: A. Ravaglioli, A. Krajewski (Eds.), Bioceramics and the Human Body, Elsevier, London, England, 1992, 166 pp. [24] W. Tong, J. Chen, X. Li, Y. Cao, Z. Yang, J. Feng, X. Zhang, Biomaterials 17 (1996) 1507. [25] L.L. Hench, A.E. Clark, in: D.F. Williams, (Ed.), Biocompatibility of Orthopaedic Implants, Vol. 2, CRC Press, Boston, 1982, 129 pp. [26] T. Kokubo, S. Ito, Z.T. Huang, T. Hayashi, S. Sakka, T. Kitsugi, T. Yamamuro, J. Biomed. Mater. Res. 24 (1990) 331. [27] O.H. Andersson, I. Kangasniemi, J. Biomed. Mater. Res. 25 (1991) 1019. [28] K. de Groot, C.P.A.T. Klein, J.G.C. Wolke, J.M.A. de Blieck-Hogervorst, in: T. Yamamuro, L.L. Hench, J. Wilson (Eds.), Handbook of Bioactive Ceramics, Vol. II, Calcium Phosphate and Hydroxylapatite Ceramics, CRC press, Boca Raton, FL, 1990, 133 pp. [29] M.J. Filiaggi, N.A. Coombs, R.M. Pilliar, J. Biomed. Mater. Res. 22 (1991) 1211. [30] R. McPhyerson, N. Gane, T.J. Hastow, J. Mater. Sci. Mater. Med. 6 (1995) 327.
Characterization of immersed hydroxyapatite-bioactive glass coatings in Hank’s solution J.H. Chern Lin1,∗ , H.J. Lin, S.J. Ding, C.P. Ju Department of Materials Science and Engineering, National Cheng Kung University, Tainan 70101 Taiwan Received 10 November 1999; accepted 15 December 1999
Abstract Monolithic HA and hydroxyapatite-bioactive glass (HA–BG) coatings have been plasma-sprayed on Ti–6A1–4V substrate using sinter-granulated HA and HA–BG powders. The immersion behavior of these coatings in Hank’s physiological solution was investigated. The results showed that sinter-granulated monolithic HA powder was irregularly-shaped and appeared quite dense. When 5 wt.% or more BG was added in HA, the powder became rough and porous. X-ray diffraction patterns showed that the presence of BG enhanced the decomposition of HA structure during fabrication of the powder. The plasma spray process itself also enhanced the decomposition of apatite. Surface morphology of the present HA–BG coatings was not much different from that of monolithic HA coating, that comprised patches of smooth and shiny glassy film and irregularly-shaped particles on its surface. Reasonably high bond strengths were obtained from all as-plasma-sprayed coatings. When immersed in Hank’s solution, these bond strengths largely declined. © 2000 Published by Elsevier Science S.A. All rights reserved. Keywords: Hydroxyapatite; Bioactive glass; Plasma spray; Bond strength; Immersion
1. Introduction Due to its favorable biocompatibility and mechanical properties, hydroxyapatite (HA)-coated Ti–6A1–4V alloy has been accepted as one of the most promising implant materials for orthopaedic and dental applications [1–7]. Various deposition techniques, such as dip coating-sintering, immersion coating, hot isostatic pressing (HIP), ion plating, sputtering, pulsed laser deposition and plasma spray, have been explored [3–7]. Among the many techniques, plasma spray is perhaps most popularly used due to its process feasibility as well as reasonably high coating bond strength and mechanical properties. Several attempts have been made to improve the performance (e.g. bioactivity [8] and mechanical properties) of bulk HA by incorporating bioactive glass (BG) or other second phase particles into HA. For example, Mochida et al. [9] developed an apatite-bioactive glass composite (ABC) material by sintering the mixture of HA and 45SF 1/4 bioactive glass. The composition of 45SF 1/4 was similar to that of 45S5 Bioglass except that 6.7 mol% CaO in 45S5 Bioglass was substituted with CaF2 . Due to the high tempera∗ Corresponding author. Tel.: +886-6-2748086; fax: +886-6-2748086. E-mail address: [email protected] (J.H. Chern Lin) 1 Professor
ture (1000◦ C) interaction between the BG and HA phases, the incorporation of HA into 45SF1/4 caused the glass phase to crystallize. Knowles and co-workers [10,11] found that phase transformations of HA into Ca5 (PO4 )2 SiO4 , ␣-Ca3 (PO4 )2 (␣-TCP) and -Ca3 (PO4 )2 (-TCP) took place when small amounts (up to 5 wt.%) of phosphate-based glasses were added. Such transformations were more profound in the HA containing soda type glass (Na2 O–P2 O5 ) than that containing lime type glass (CaO–P2 O5 ). A strengthening effect was observed when 2.5 wt.% CaO–P2 O5 glass was added in the HA resulting in an HA/␣-TCP sintered body. In their study of thermal stability of sintered HA-doped glass (SiO2 –CaO–Na2 O–P2 O5 –A12 O3 –B2 O3 ), Kangasniemi et al. [12] showed that, when sintered at temperatures higher than 700◦ C, sodium ions diffused into doped HA particles inducing the phase transformation of HA into NaCaPO4 . These phase transformations led to changes in properties of the glass. The biodegradation phenomenon of HA in both in vitro and animal studies has been reported by Ramselaar et al. [13], Blitterswijk et al. [14], and others [6,15,16]. Kummer and Jaffe [6] evaluated the stability of a cyclically-loaded HA coating in in-vitro solutions and found increased instability of the HA coating. Reis et al. [15] suggested that degradation of their plasma-sprayed monolithic HA coating in Hank’s or
0254-0584/00/$ – see front matter © 2000 Published by Elsevier Science S.A. All rights reserved. PII: S 0 2 5 4 - 0 5 8 4 ( 0 0 ) 0 0 2 0 9 - 1
230
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Table 1 Composition of raw HA powder used in this studya Chloride Fluoride Sulfate Heavy metals Arsenic Iron Water Assay a
Table 2 Composition of Hank’s physiological solution used in this study <0.05 wt.% <0.005 wt.% <0.3 wt.% <0.003 wt.% <0.0002 wt.% <0.02 wt.% 1.7 wt.% 96.5%
Merck #2196; E. Merck, Germany.
NaCl solution under loading was not only due to loosening of solid particles, but also to coating dissolution. When the coating was subjected to a higher load in the solution, the reduction in coating thickness became larger. Yang et al. [16] found that bond degradation of ≈25–33% of the original strength was measured when the specimens were immersed in a pH-buffered, serum-added simulated body fluid. Chern Lin et al. [7,17] studied a series of HA–BG composite coatings plasma-sprayed on Ti6Al4V substrate. The powders used for plasma spray were prepared by directly mixing in-house fabricated BG powder and a commercial HA powder. Their results showed that directly mixed type
Fig. 1. Schematic diagram of bond strength test for this study (ASTM C633-79).
Constituent
g/l
NaCl NaHCO3 KCl MgCl2 ·6H2 O CaCl2 Na2 HPO4 ·2H2 O MgSO4 ·7H2 0 KH2 PO4 Glucose
8.00 0.35 0.40 0.10 0.14 0.06 0.06 0.06 1.00
HA–BG composite coatings were generally more resistant to moisture attack than monolithic HA coating in terms of bond strength degradation. In a recent study, Ding et al. [18] found that the dissolution depth of plasma-sprayed coatings was largely dependent on the type and composi-
Fig. 2. LVSEM micrographs of various powders fabricated in-house for plasma spray.
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
231
Fig. 3. XRD patterns of various powders for plasma spray (䊊: apatite; A: ␣-TCP; B: -TCP).
tion of the coating. The granulated type HA–BG coatings were less dissolvable in simulated body fluid than monolithic HA or mixed type HA–BG coatings. In the present study, sinter-granulated HA and HA–BG coatings have been plasma-sprayed on Ti6Al4V substrate. The structure, bond strength, and immersion behavior in Hank’s physiological solution of these coatings were investigated.
Fig. 5. LYSEM micrographs of various as-plasma-sprayed coatings.
2. Materials and methods
Fig. 4. XRD patterns of various plasma-sprayed coatings (䊊: apatite; T: TTCP; C: CaO; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
A commercial HA powder (Merck A.G., Darmstadt, Germany) with particle size <1 m was used in this study. The composition of this powder is listed in Table 1. The in-house fabricated BG used in this study contained frits of SiO2 , Na2 O, CaO and P2 O5 (SiO2 :Na2 O:CaO:P2 O5 =10:3:5:2 in weight). In preparing the BG, appropriate amounts of the frits were mixed and melted at 1350◦ C for 2 h. The melt was then quenched in water and fractured fragments of the glass were collected. Detailed description of the fabrication of the BG powder has been given in an earlier paper [7]. The HA–BG powders were prepared using sintergranulation method [18]. The sinter-granulation method involved mixing and ball-milling appropriate amounts of HA and BG powders (HA:BG=95:5 and 92.5:7.5 in weight ratio) in ethyl alcohol, followed by drying and sintering at 1300◦ C for 24 h. The sintered granules were then crushed,
232
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 6. Cross-sectional LVSEM micrographs of various as-plasma-sprayed coatings.
milled and sieved to obtain particles with sizes 44–149 m, a particle size range recommended by the manufacturer of the plasma spray system. Throughout this study the sinter-granulated powders and the plasma-sprayed coatings produced from such powders were designated as the same code. Commercially available cylindrical rods and plates of Ti6Al4V alloy were used as the substrate material. Prior to plasma spray, the substrate surface was mechanically polished to #1200 grit level, cleaned and sandblasted with 450 m SiC particles. To obtain a uniform coating, the substrate was mounted on a disk which could rotate during spraying in air using a Plasma-Technik A-3000 system (Plasma Technik A.G., Wohien, Switzerland). The coatings had a thickness in the range of 100–140 m. The ASTM C633-79 method suggested for testing flame-sprayed coating bond strength was used to measure the bond strength of the present coatings, as shown in Fig. 1. In performing this test, two identical cylindrical Ti6Al4V rods were prepared, one with coating on a flat surface and the other without. The flat surface of the
uncoated rod, which was to be bonded to the coated rod, was sandblast-roughened to enhance resin adherence. A thin layer of bonding glue (Plasma Technik, Klebbi glue, Switzerland) was applied onto both coated and uncoated surfaces. After the two rods were carefully aligned using a special device [17], a compressive stress was applied to both rods during curing to assure an intimate contact between resin and the two surfaces. After curing at 180◦ C in an oven, the bonded rods were bench cooled to room temperature. The pressure was then released and the resin-bonded rods were removed from the device. The coatings were pulled out from substrate using an Instron 8562 tester at a cross head speed of 1.0 mm/min. In the present study, a Hank’s physiological solution (Table 2), recommended by Pourbaix [19] for testing corrosion behavior of implant materials, was used for the immersion test. The coated specimens, each with a surface area of 1 cm2 , were immersed up to 30 days in vials containing 50-ml Hank’s solution, which were maintained at 37◦ C throughout tests. The solution was agitated daily to help maintain uniform ion concentrations. After predetermined
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 7. XRD patterns of monolithic HA coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; C: CaO; T: TTCP).
time periods of immersion, the specimens were removed from the vials and stored in a desiccator for characterization. Phases of the various powders and coatings before and after immersion were analyzed by an X-ray diffractometer (Rigaku D-max IIIV, Japan) with Ni-filtered CuKα radiation operated at 30 kV and 20 mA with a scanning speed
233
Fig. 9. XRD patterns of 92.5HA/7.5BG coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
of 1◦ /min. Microstructure of the specimens were examined using a low vacuum scanning electron microscope (Topcon SM-300, Tokyo, Japan). Using this low vacuum SEM (LVSEM), specimens could be observed under a ‘wet’ condition and the conventional deposition of a conducting film was not needed, thus largely eliminating a variety of difficulties such as high vacuum and charging-induced structural damage. Microchemistry was characterized using an energy dispersive spectroscopy (EDS) system. Specimens for cross-sectional microscopy were prepared by mounting the coated specimens in epoxy resin, followed by polished through 0.3 m alumina. More important, the bond strength variation as a function of immersion time in Hank’s solution was evaluated by means of ASTM C633-79 method. One-way ANOVA statistical analysis was used to evaluate the statistical significance of bond strength data. Scheff multiple comparison test was used to determine the significance of the deviations in bond strength of each coating for different immersion times. In all cases, the results were considered statistically different with p<0.05.
3. Results and discussion 3.1. Characterization of HA and HA–BG powders
Fig. 8. XRD patterns of 95HA/5BG coating after immersion in Hank’s solution (䊊: apatite; A: ␣-TCP; B: -TCP; S: Na2 CaSiO4 ).
As shown in Fig. 2, the sinter-granulated monolithic HA powder was irregularly-shaped and quite. When 5 wt.% BG was added in HA (95HA/5BG), the powder became porous. The majority of the pores had sizes smaller than 10 m.
234
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 10. LVSEM micrographs of monolithic HA coating after immersion in Hank’s solution.
Fig. 11. LVSEM micrographs of 95HA/5BG coating after immersion in Hank’s solution.
Kangasmiemi et al. [12] attributed the BG-induced formation of porosity to the release of hydroxyl group from HA. Wang et al. [20] confirmed that the presence of BG not only enhanced the breakdown process of hydroxyl group, but also enhanced the decomposition of HA structure. It seems
that at least two factors, i.e., BG-enhanced release of hydroxyl group and BG-induced liquid phase sintering effect [21], were operating competitively to determine the eventual porosity level of the sinter-granulated HA–BG powders. When 7.5 wt.% BG was added (92.5HA/7.5BG), the powder
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
235
Fig. 12. LVSEM micrographs of 92.5HA/7.5BG coating after immersion in Hank’s solution.
surface revealed many microcracks on the surface. The formation of the microcracks might be explained by the differences in density and coefficient of thermal expansion (CTE) among the various phases that involved phase transformations at high temperatures [18]. X-ray diffraction (XRD) patterns of the various powders are shown in Fig. 3. As shown in the figure, HA was a highly crystalline phase in accordance with JCPDS No. 9-432 file. When sinter-granulated powder contained 5 wt.% BG, the decomposition of apatite into -TCP and ␣-TCP phases was readily detected. As BG content increased (92.5HA/7.5BG), the apatite intensity decreased, while the intensities of ␣-TCP and -TCP phases increased. It was suggested that, when an HA–BG mixture was sintered at elevated temperatures, the sodium ions in BG might diffuse into the HA structure, causing release of a portion of hydroxyl group from HA. The HA structure with deficient hydroxyl group became less stable at high temperatures and had greater tendency to decompose [11,12,18]. The present XRD results seem to support this theory.
3.2. Structural characterization of as-plasma-sprayed HA and HA–BG coatings As shown in the XRD patterns in Fig. 4, all as-plasmasprayed coatings had lower apatite crystallinity than their respective initial powders, in agreement with previous studies [17,18]. In addition to the broadening of apatite peaks, Ca4 P2 O9 (TTCP), CaO, ␣-TCP and -TCP phases were
observed in monolithic HA coating. The high temperature involved in plasma spray process had obviously enhanced the decomposition of apatite as well as chemical reactions among different phases. The intensities of ␣-TCP and -TCP in 95HA/5BG coating were larger than those in monolithic HA coating. A large amount of sodium calcium silicate (Na2 CaSiO4 ) was observed. As BG content increased to 7.5 wt.%, ␣-TCP, -TCP and Na2 CaSiO4 continued to increase at the expense of the apatite phase [18]. The formation of sodium calcium silicate might result from the reactions between apatite and such ions as silicon and sodium in glass. Broad face and cross-sectional LVSEM micrographs of as-plasma-sprayed HA and HA–BG coatings are shown in Figs. 5 and 6, respective. Randomly distributed pores of different sizes and numerous cracks could be observed on all coated surfaces. Patches of smooth and shiny glassy film and irregularly-shaped particles were observed on the surface of HA coating, as shown in Fig. 5. In their transmission electron microscopic (TEM) study, Ji et al. [22] showed that the plasma spray-induced glassy structure was an amorphous form of calcium phosphate. During plasma spraying, small size powder was completely melted, while larger size powder, such as the one used in the present study, was only partially melted resulting in a coating morphology comprising the observed irregularly-shaped particles [23,24]. The overall broad face surface morphology of both HA–BG coatings was not much different from that of monolithic HA coating. The thickness of monolithic HA, 95HA/5BG, and 92.5HA/7.5BG coatings were roughly 115–145, 100–130,
236
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 13. Cross-sectional LVSEM micrographs of monolithic HA coating after immersion in Hank’s solution.
and 100–130 m, respectively. The rather large deviations in thickness are quite common for plasma-sprayed due to their surface roughness. 3.3. Structural characterization of immersed HA and HA–BG coatings The XRD patterns of HA, 95HA/5BG, and 92.5HA/ 7.5BG-coated Ti–6A1–4V after a series of immersion times are shown in Figs. 7–9. When immersed in Hank’s solution, the intensities of such phases as CaO, ␣-TCP, -TCP, and TTCP in monolithic HA and 95HA/5BG coatings (Figs. 7 and 8) gradually decreased, indicating either the disso-
lution of these phases or the deposition of an amorphous apatite phase. The major apatite peaks at 2θ roughly between 32 and 33◦ were slightly broadened as a result of the deposition of amounts of amorphous apatite phase after immersion [7,18,25,26]. The XRD patterns of immersed 92.5HA/7.5BG coating (Fig. 9) were somewhat different from that of monolithic HA or 95HA/5BG coating. After immersion for 15 days or longer, the intensities of ␣-TCP and apatite increased significantly. The intensity of Na2 CaSiO4 intensities was still significant due to its large quantity in the coating. The broad face LVSEM micrographs of immersed HA, 95HA/5BG, 92.5HA/7.5BG coatings are shown respectively
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
237
Fig. 14. Cross-sectional LVSEM micrographs of 95HA/5BG coating after immersion in Hank’s solution.
in Figs. 10–12, while the cross-sectional micrographs of the three immersed coatings are shown respectively in Figs. 13–15. As shown in Fig. 10, the surface morphology of the HA coating immersed for 1 day was similar to that of as-sprayed coating. After 7 days of immersion, the glassy phase and numerous irregularly-shaped small particles on the surface of as-sprayed monolithic HA coating were no longer present. Instead, the surface started to be covered with clusters of precipitated apatite spherulites (indicated by the arrows as shown in Fig. 13). Higher magnification examination revealed that most of the spherulites had sizes of several microns. It appears that, during the immersion test, dissolu-
tion of the coating surface and deposition of the Ca/P-rich layer had both taken place, in agreement with the XRD results. Similar apatite spherulites ca. 2–3 m in diameter had also been observed by Andersson and Kangasniemi [27] in their immersed 45S5 glass in SBF for 8 h and by Chern Lin et al. [7]. According to Hench and Clark [25] and Kokubo et al. [26], a Ca/P-rich amorphous layer could be formed in vivo or in vitro on the surface of a bioactive glass, ceramic, or glass-ceramic. This Ca/P-rich layer was later crystallized to form small carbonate-containing HA crystals [25,26]. Those cracks observed on the coating surface were a result of drying process.
238
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
Fig. 15. Cross-sectional LVSEM micrographs of 92.5HA/7.5BG coating after immersion in Hank’s solution.
As 5 wt.% BG was added in HA, the apatite spherulites were not observed until after 15 days of immersion (Fig. 11). When the BG content increased to 7.5 wt.%, these spherulites were not observed even after 30-day immersion (Fig. 12). The precipitated apatite layer could be distinguished more clearly from the original coating layer in cross-sectional LVSEM micrographs, as those indicated by arrows in Figs. 13 and 14. Such a precipitated layer with a distinct contrast was not observed in the 92.5HA/7.5BG coating even after immersion for 30 days, confirming the broad face examination.
3.4. Bond strengths of HA and HA–BG coatings Although the ASTM C633 method has often been used for measuring the bond strength of HA coatings, whether the measured bond strength values using this method could represent ‘true’ bond strengths is still in debate [23,28,29]. For example, it was argued [28] that, during press-curing, the bonding resin might infiltrate through open porosity into coating interior, introducing uncertainties to the measured values. Another difficulty arises from the determination of the ‘true’ fracture area. Post-test observations indicated that
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
239
Fig. 16. Bond strengths of various plasma-sprayed coatings after immersion in Hank’s solution.
all major fractures of the present coated specimens occurred within coating interior. In many cases only small portions of coatings were pulled out right from substrate surface. The major portions of the coatings were still bonded to the substrate when the coatings were pulled out. For simplicity, in this study, the original cross-sectional area of Ti–6A1–4V substrate was used at all times in the calculation of bond strengths. Compared to reported values [1,7,18,23], reasonably high bond strengths were obtained from all present as-plasma-sprayed coatings (in the range 50–60 MPa), as shown in Fig. 16. One-way ANOVA statistical analysis showed that there were no significant differences in bond strength among these as-sprayed coatings (p=0.05). When immersed in Hank’s solution, however, all three coatings largely lost their bond strengths. The statistical analysis using Scheffé multiple comparison test showed that the bond strength of monolithic HA coating significantly declined from 53.6 MPa, the as-coated bond strength, down to 21.8 MPa with a reduction of ca. 59% after immersion for Table 3 Ca/P ratios of as-coated and tensile-fractured surfaces of coated specimens immersed in Hank’s solution for 30 days Coating
As-coated surface
Fractured surface
HA 95HA/5BG 92.5HA/7.5BG
2.40±0.01 2.51±0.04 2.46±0.12
1.60±0.08 1.76±0.12 1.84±0.13
7 days (p<0.001). 95HA/5BG and 92.5HA/7.5BG coatings lost 62 and 55% in bond strength, respectively, after 7-day immersion. This deterioration in bond strength seems unavoidable for plasma-sprayed HA-based coatings immersed in simulated body fluid and also observed in other studies [6,15,16]. A question was raised regarding whether this immersioninduced degradation in bond strength could arise from a poor bonding between the plasma-sprayed coating and the precipitated apatite layer. To answer this question, the post-test fracture surfaces of three coated materials immersed for 30 days were examined and Ca/P ratios were measured. The observation indicated that all major failures of the immersed specimens occurred within coating interior). As shown in Table 3, the Ca/P ratios of all fractured surfaces were much lower than those of as-coated surfaces. For example, the Ca/P ratios of monolithic HA coating before and after 30 days of immersion were 2.40 and 1.60, respectively; the Ca/P ratios of 95HA/5BG before and after immersion were 2.51 and 1.76, respectively; while the Ca/P ratios of 92.5HA/7.5BG before and after immersion were 2.46 and 1.84, respectively. The much higher Ca/P ratios (than the stoichiometric Ca/P ratio, 1.67, of HA) on as-sprayed coating surfaces were not surprising and also observed in other plasma-sprayed HA coatings [30] due to the fact that a large quantity of phosphorus was vaporized during the high temperature plasma spray process. The lower Ca/P ratios indicated that those failures probably occurred within the precipitated apatite layer with a Ca/P ratio closer to 1.67. The numerous drying-induced
240
J.H. Chern Lin et al. / Materials Chemistry and Physics 64 (2000) 229–240
cracks observed on immersed surfaces (Figs. 10–12) also suggest that it was very likely for the failures to occur within this poorly-bonded precipitated layer.
4. Conclusions 1. The sinter-granulated monolithic HA powder was irregularly-shaped and appeared quite dense. When 5 wt.% or more BG was added in HA, the powder became rough and porous. XRD patterns showed that the presence of BG enhanced the decomposition of HA structure during fabrication of the powders. As BG content increased, the apatite intensity decreased, while the intensities of ␣-TCP and -TCP phases increased. 2. The as-sprayed coatings had lower apatite crystallinity than their respective initial powders. The plasma spray process itself enhanced the decomposition of apatite. 3. Surface morphology of the present HA–BG coatings was not much different from that of the monolithic HA coating, that was composed of patches of smooth and shiny glassy film and irregularly-shaped particles on its surface. 4. Reasonably high bond strengths were obtained from all as-sprayed coatings. When immersed in Hank’s solution, the bond strengths of all coatings largely declined.
Acknowledgements The authors acknowledge with appreciation the support for this research by the National Science Council of the Republic of China under the contract No. NSC 86-2213-E-033 & 034 & 035. References [1] R.G.T. Geesink, K. de Groot, C.P.A.T. Klein, Clin. Orthop. 225 (1987) 147. [2] J.E. Lemons, Clin. Orthop. 235 (1988) 220. [3] W.R. Lacefield, Bioceramics: material characteristics versus in vivo behavior, Ann. N.Y. Acad. Sci. 523 (1982) 72. [4] C.K. Wang, J.H. Chern Lin, C.P. Ju, H.C. Ong, R.P.H. Chang, Biomaterials 18 (1997) 1331. [5] C.S. Kim, P. Ducheyne, Biomaterials 12 (1991) 461.
[6] F.J. Kummer, W.L. Jaffe, J. Appl. Biomater. 3 (1992) 211. [7] J.H. Chern Lin, M.L. Liu, C.P. Ju, J. Biomed. Mater. Res. 28 (1994) 723. [8] L.L. Hench, in: T. Yamamuro, L.L. Hench, J. Wilson (Eds.), Handbook of Bioactive Ceramics, Vol. I, Bioactive Glasses and Glass-ceramics, CRC Press, Boca Raton, FL, 1990, 7 pp. [9] M. Mochida, I. Fujiu, M. Ogino, Yogyo-Kyokai-Shi 95 (1987) 798. [10] J.C. Knowles, W. Bonfield, J. Biomed. Mater. Res. 27 (1993) 1591. [11] J.D. Santos, J.C. Knowles, R.L. Reis, F.J. Monteiro, G.W. Hastings, Biomaterials 15 (1994) 5. [12] I.M.O. Kangasniemi, K. de Groot, J. Wolke, Ö. Andersson, Z. Luklinska, J.G.M. Becht, M. Lakkisto, A. Yli-urpo, J. Mater. Sci. Mater. Med. 2 (1991) 133. [13] M.M.A. Ramselaar, F.C.M. Driessens, W. Kalk, J.D. de Wijn, P.J. van Mullen, J. Mater. Sci. Mater. Med. 2 (1991) 63. [14] C.A. van Blitterswijk, J.J. Grote, W. Kuijpers, W.Th. Daems, K. de Groot, Biomaterials 7 (1986) 137. [15] R.L. Reis, F.J. Monteiro, G.W. Hastings, J. Mater. Sci. Mater. Med. 5 (1994) 457. [16] C.Y. Yang, B.C. Wang, E. Chang, B.C. Wu, J. Mater. Sci. Mater. Med. 6 (1995) 258. [17] J.H. Chern Lin, M.L. Liu, C.P. Ju, J. Mater. Sci. Mater. Med. 5 (1994) 279. [18] S.J. Ding, C.P. Ju, J.H. Chern Lin, J. Mater. Sci. Mater. Med. (in press). [19] M. Pourbaix, Biomaterials 5 (1984) 122. [20] C.K. Wang, J.H. Chern Lin, C.P. Ju, Mater. Chem. Phys. 53 (1997) 138. [21] J.D. Santos, R.L. Reis, F.J. Monteiro, J.C. Knowles, G.W. Hastings, J. Mater. Sci. Mater. Med. 6 (1995) 348. [22] H. Ji, C.B. Ponton, P.M. Marquis, J. Mater. Sci. Mater. Med. 3 (1992) 283. [23] J.G.C. Wolke, C.P.A.T. Klein, K. de Groot, in: A. Ravaglioli, A. Krajewski (Eds.), Bioceramics and the Human Body, Elsevier, London, England, 1992, 166 pp. [24] W. Tong, J. Chen, X. Li, Y. Cao, Z. Yang, J. Feng, X. Zhang, Biomaterials 17 (1996) 1507. [25] L.L. Hench, A.E. Clark, in: D.F. Williams, (Ed.), Biocompatibility of Orthopaedic Implants, Vol. 2, CRC Press, Boston, 1982, 129 pp. [26] T. Kokubo, S. Ito, Z.T. Huang, T. Hayashi, S. Sakka, T. Kitsugi, T. Yamamuro, J. Biomed. Mater. Res. 24 (1990) 331. [27] O.H. Andersson, I. Kangasniemi, J. Biomed. Mater. Res. 25 (1991) 1019. [28] K. de Groot, C.P.A.T. Klein, J.G.C. Wolke, J.M.A. de Blieck-Hogervorst, in: T. Yamamuro, L.L. Hench, J. Wilson (Eds.), Handbook of Bioactive Ceramics, Vol. II, Calcium Phosphate and Hydroxylapatite Ceramics, CRC press, Boca Raton, FL, 1990, 133 pp. [29] M.J. Filiaggi, N.A. Coombs, R.M. Pilliar, J. Biomed. Mater. Res. 22 (1991) 1211. [30] R. McPhyerson, N. Gane, T.J. Hastow, J. Mater. Sci. Mater. Med. 6 (1995) 327.