Sintering of the Ultrahigh Pressure Densified Hydroxyapatite Monolithic

Journal of MATERIALS RESEARCH Welcome Comments Help

Sintering of the ultrahigh pressure densified hydroxyapatite monolithic xerogels Ján Majling, Peter Znáik, Angela Palová, and Stefan Svet´ık Department of Ceramics, Slovak Technical University, 812 37 Bratislava, Slovak Republic Stefan Koval´ık Department of Materials Science and Technologies, Slovak Technical University, 812 37 Bratislava, Slovak Republic Dinesh K. Agrawal and Rustum Roy Intercollege Materials Research Laboratory, The Pennsylvania State University, University Park, Pennsylvania 16802 (Received 16 February 1996; accepted 28 June 1996) Dense and translucent ceramics were prepared by sintering of cylindrical preforms of hydroxyapatite extruded from xerogels. Extruded specimens were dried as monoliths and then consolidated by applying cold isostatic pressure, ranging from 500 to 1500 MPa. Upon heating the samples began to densify at 610 ±C, and the densification/sintering was completed at 870 ±C as was evidenced by the dilatometry plot indicating no further shrinkage. The sintered specimens thus formed were translucent in appearance. Further heating of the samples up to 1200 ±C resulted in their “bloating” or creation of pores in the originally dense matrix. Pore creation within the structure is reproducible, it proceeds from the surface to the interior of the sample, and its spreading can be thermally controlled. Pore evolution within the single phase dense polycrystalline material is not related to the frequently occurring phenomenon of microcracking in ceramics during cooling.

I. INTRODUCTION powder). The consolidation of powder was achieved Hydroxyapatite (HAp) ceramics draw special inter- by the cold isostatic pressing (350 MPa), and sintering est for their biocompatibility. Dense apatite ceramics are of resultant pellets was accomplished in a microwave in commercial use for bone replacement in the human furnace. body. Fabrication procedures for dense hydroxyapatite Cold isostatic pressing of HAp powders by medium, ceramics are quite diversified. Most of the researchers high, and very high pressures is a most frequently have used methods involving the formation of suitable used technique to consolidate powders, in order to gel-like precipitates1,2 and then drying the obtained pow- smear the aggregates of particles. In this contribution ders. The resulting xerogels are then calcined,3,4 milled, we have examined the influence of isostatic pressures compacted,5,6 and finally subjected to sintering.7,8 (500–1500 MPa) on the densification of hydroxyapatite Aggregation of powders during drying and calcina- preforms which were subsequently sintered at a constant tion stages (dry route) causes nonuniformity in packing rate. The preforms were fabricated as monoliths by of particles in green compacts, and thus prevents full extruding the wet gel. During drying the consolidation of densification during sintering. Colloidal or sol/gel proc- particles was expected to contribute to the uniformness esses are therefore considered in order to prevent the of the green structure. aggregation of powders. A defect-free fine microstructure in a ceramic is II. EXPERIMENTAL synonymous with the optimum mechanical properties The hydroxyapatite gel (with Ca͞P ratio of 1.8) of polycrystalline materials and is usually manifested was precipitated from 0.2 M solutions of Ca(NO3)2 and by the translucency or transparency of the respective NH4H2PO4, with pH adjusted to 12. The two solutions materials. The preparation of a translucent1 HAp ce- were mixed together and heated in a closed plastic ves- ramics was reported in which solution route assisted sel at 50 ±C for three days. The resultant suspension in the partial consolidation of colloidal particles. The was homogenized by shaking several times a day. Am- transparent HAp was obtained after hot pressing.9,10 In monia was boiled out, and the precipitate was decanted a recent report, however, it was also prepared by the several times, filtered, and thoroughly washed with dis- pressureless sintering,11 using a noncalcination route and tilled water. The wet cake was extruded using an adapted hydrothermally synthesized powder (an aggregate free syringe into small cylinders of 1.5 cm in diameter. The

198 J. Mater. Res., Vol. 12, No. 1, Jan 1997  1997 Materials Research Society J. Majling et al.: Sintering of the ultrahigh pressure densified hydroxyapatite monolithic xerogels cylinders were dried while suspended vertically at room TABLE I. Relative densities of green and sintered (at 1000 ±C) temperature, and subsequently calcined at 300 to 500 ±C samples of hydroxyapatite. for 5 h. The calcined preforms were wrapped in a thin Pressure Green density Sinter density aluminum foil and encapsulated with a flexible rubber MPa A B A B coating by dipping them into a natural latex emulsion. The encapsulated samples were isostatically pressed in a 500 0.51 0.55 0.99 0.96 universal hydraulic press, equipped with a high pressure 1000 0.59 0.60 0.98 0.97 1500 0.66 0.62 0.96 0.90 steel vessel capable of withstanding pressures up to 2000 MPa. A ෇ preforms calcined at 300 ±C. ± The phase composition of hydroxyapatite gels was B ෇ preforms calcined at 500 C. analyzed by x-ray diffractometry (STADI, Stoe). Its specific surface area was measured by BET (Sorptomatic (1100 ±C) XRD pattern (Fig. 1). By heating the gel to 1900, Fisons). Sintering experiments were performed 1180 ±C (HT), the average crystallite size is increased, on ϳ 1 cm long cylinders using a Netzch dilatometer as is evident by the sharper XRD peaks in the HT pat- (Model 402 E) at a heating rate of 10 ±C͞min in a static tern (Fig. 1). air atmosphere. The stress induced in the sample by the push rod of the dilatometer was 0.5 N. Bulk density of Dilatometric measurements samples was measured by Archimedes’ technique. The After further heating of samples above 1000 ±Cat microstructures (fracture surfaces) of sintered samples which shrinkage of samples was completed, an expan- were observed by SEM (Tesla BS 300). sion of the samples was observed at about 1200 ±C as is evident from their dilatometric curves (Fig. 2). III. RESULTS AND DISCUSSION The expansion occurs at a temperature that is well above the temperature at which the shrinkage of samples The precipitation procedure of the hydroxyapatite completely ceased. Figure 2 shows a typical dilatometric gel was essentially the same as the one given in the 1,12 curve (for a sample pressed at 500 MPa). The dried literature with a slight modification that in our prepa- heated preforms of extruded cylinders, which were not ration the solutions of Ca(NO3)2 and NH4H2PO4 were subjected to cold isostatic pressing, did not exhibit any simply poured together. Instantaneous massive nucle- such expansion at all. ation occurred, and the primary particles were expected A fracture surface (perpendicular to the axis of to be smaller than those obtained in the method in which the sample cylinder) of the sample, cooled to room one solution was added dropwise to the second stirred temperature from the temperature at which the expansion solution. just commenced, showed an inner dense (D) and a The precipitate was filtered, thoroughly washed, ± circumferential porous (P) zone (Fig. 3). There exists dried, and when heated at 1200 C for 0.5 h, exhibited a transitional zone between these two constant density single phase material of well-crystallized hydroxyapatite zones which is also shown in detail in Figs. 4(b) and with no trace of tri calcium phosphate. This was consid- 4(c). Figures 4(a) and 4(d) are enlargements of the D and ered as a proof that the precipitated powder was not P zones from Fig. 3. A sample that has just completed a calcium deficient hydroxyapatite as reported in the literature.2 Dried filter cake showed a specific surface area of 106.2 m2͞g. The extruded wet cylinders shrank in their diameters from 1.5 cm in the wet state to approx. 0.52 cm in the dry state. Pressing to 500 MPa reduced the cylinder’s diameters further to approx. 0.39 cm. The green densities of pressed samples are listed in Table I. In the first run of experiments, the samples at sintering (dilatometric) experiments were heated only to 1000 ±C, a temperature well above the temperature at which the shrinkage of samples had completed. Samples after heating to 1000 ±C were translucent with a slight ruby color. Their relative densities are given in Table I. The theoretical density of hydroxyapatite was taken as 3.156 g͞cc.13 Crystallite size estimated from the surface area data FIG. 1. X-ray powder diffraction patterns for samples before and after is around 15–20 nm, so the particles are very fine an expansion, LT: sample heated to 1000 ±C; HT: sample heated to which is also reflected in the line broadening in LT 1180 ±C.

J. Mater. Res., Vol. 12, No. 1, Jan 1997 199 J. Majling et al.: Sintering of the ultrahigh pressure densiﬁed hydroxyapatite monolithic xerogels

FIG. 2. Dilatometric curve of the hydroxyapatite sample compacted at 500 MPa. LT, HT, and MS refer to thermal curing of samples taken for x-ray analysis (Fig. 1) and for the microstructure analysis (Figs. 3 and 4). expanding exhibits larger crystals and single phase HAp. In the x-ray diffraction pattern (HT), the diffraction peaks are much sharper than the corresponding diffrac- tions of the unexpanded sample, LT (Fig. 1). The reason for the poring of the originally dense matrix and a corresponding expansion of sample is believed to be due to the presence and the phase transition of tricalcium phosphate (CTP).14–16 As shown in Ref. 16 FIG. 3. Microstructure of a sample in which the expansion has just commenced, exhibiting inner dense (D) and a circumferential porous such a phase transformation causes the intragranular (P) zones. fracture without creating pores. The dilatometric curves in such a case show an immediate transition from a shrinkage to the expansion.15,16 Foaming (or bloating) is encountered in a number of considered, however, as the macro defect-free bodies, systems involving gas evolution. These systems invari- without microcracks, and the entrapped hydroxyls and ably necessitate the condition for the viscous flow to oc- water molecules can therefore escape the sample only cur. The resulting pores are invariably closed pores.17–19 by the long distance diffusion. It means that no defects Foaming of the pure, amorphous, single phase particulate (pores) larger than the primary particles are present. It systems such as silica gels20 and aerogels belong to is believed that high pressure densification may have this category. The foaming is caused by the residual eliminated macropores. hydroxyls and creation of pores. Prolonged tempering The pores are not completely eliminated at lower and removing of hydroxyls at lower temperatures elimi- temperatures (as indicated by the translucency of sam- nates foaming at higher temperatures. Sintering of silica ples at 1000 ±C), and the water vapor inside them might gels by microwaves prevents foaming of samples too,20 inhibit their complete elimination. The hydroxyls are probably by the inverse temperature profile. expected to be concentrated also at grain boundaries. The foaming of the HAp ceramics in the present At high enough temperature, which in the present case case is also associated with a presence of the residual corresponds nearly to 1200 ±C, water pressure in pores hydroxyls and their subsequent release. After the closure exerts a localized tension in the matrix which exceeds its of pores during sintering of samples, which occurs at strength. The poring of the structure starts therefore at the a relatively low temperature, the pores are filled with surface of the sample and advances in the interior. The the water vapor. The x-ray diffraction patterns taken individual newly formed pores “nucleate” statistically before and after foaming indicate the presence of HAp in the dense portion of the matrix well in advance to only, and no evidence of HAp dissociation into C3P etc. the already decomposed matrix (on residual pores) and was observed. The partially sintered compacts can be decrease the free energy necessary for the “nucleation”

200 J. Mater. Res., Vol. 12, No. 1, Jan 1997 J. Majling et al.: Sintering of the ultrahigh pressure densiﬁed hydroxyapatite monolithic xerogels

FIG. 4. Enlarged microstructures of (a) the dense zone, (b) transitional zone, (c) porous zone, and (d) transitional zone. of additional pores (which may be generated on account A poring of the dense hydroxyapatite ceramics indicates of hydroxyl ions present on grain boundaries). its thermal instability. Authors in Ref. 23 declared hy- It is interesting to note that these pores are homo- droxyapatite ceramics to be thermally unstable due to its geneously distributed and are small in size similar to ionic conductivity characteristics. the size of grains formed (ϳ15–20 nm). Their formation necessitates, therefore, a transient plastic deformation of the sample, without leading to the microcracking. In literature reported superplasticity of the hydroxyapatite IV. CONCLUSIONS ceramics suggests22 that such a mechanism of the de- The basic facts are reiterated pertaining to the transformation might be available. Equally interesting is to formation of the nearly fully dense and crystalline hy- notice that for the sufficiently well-organized particles of droxyapatite ceramics into the porous one. The poring hydroxyapatite in the green structures the temperatures of the structure is believed not to be due to the elastic of 1100 and 1150 ±C have been recommended as optimal deformation which is associated with the microcracking, ones for a preparation of the dense ceramics.1,8,10 These but to the presence of the residual hydroxyls and their temperatures are well below the temperature at which the association in pores at grain boundaries, which at a poring of the structure is observed in the present study. sufficiently high temperature exert the stress in the In this report we have identified some basic facts sample causing it to plastically deform, and hence the relevant to the transformation of dense hydroxyapatite resultant expansion. ceramics to the porous one and suggested a plausible The preparation of samples, especially the very high explanation. This transformation is believed to be dis- pressure consolidation of preforms leads to the micro similar to the conventional foaming or microcracking. In defect-free bodies, a prerequisite for the occurrence of order to understand this process in detail, an additional such a transition. Further experiments are necessary to experimental investigation of this system is necessary. fully explain this transformation.

J. Mater. Res., Vol. 12, No. 1, Jan 1997 201 J. Majling et al.: Sintering of the ultrahigh pressure densiﬁed hydroxyapatite monolithic xerogels

FIG. 4. (continued from previous page)

ACKNOWLEDGMENTS 11. D. K. Agrawal, Y. Fang, D. M. Roy, and R. Roy, in Microwave Processing of Materials III, edited by R. L. Beatty, W. H. Partial supports from the Copernicus project Sutton, and M. F. Oskander (Mater. Res. Soc. Symp. Proc. 269, Contract No. CIPA-CT94-0233 and NSF Grant No. Pittsburgh, PA, 1992). GER-9355460 are greatly appreciated. 12. Z. Bak´o and I. Kotsis, Ceramics Int. 18, 373 (1992). 13. G. De With, H. J. A. Van Dijk, N. Hattu, and K. Prijs, J. Mater. Sci. 16, 1592 (1981). REFERENCES 14. K. Yamashita, H. Owada, H. Nakagawa, T. Umegaki, and 1. M. Jarcho, C. Bolen, M. Thomas, J. Bobick, J. Kay, and T. Kanazawa, J. Am. Ceram. Soc. 69, 690 (1986). R. Doremus, J. Mater. Sci. 11, 2027 (1976). 15. Satoshi Nakamura, Ryohei Otsuka, Hideki Aoki, Masaru Akao, 2. A. Osaka, Y. Miura, K. Takeuchi, and K. Takahashi, J. Mater. Naoki Miura, and Takeyuki Yamamoto, Thermochim. Acta 165, Sci.: Materials in Medicine 2, 51 (1991). 57 (1990). 3. A. Ababou, D. Bernache-Assollant, and M. Heughebaert, Ceram. 16. Kiyoshi Itatani, Toshio Nishioka, Satoru Seike, F. S. Howell, Processing Sci. Technol., p. 111. Akira Kishioka, and Makio Kinoshita, J. Am. Ceram. Soc. 77, 4. Kazuo Kondo, Masahiko Okuyama, Hidetoshi Ogawa, and 801 (1994). Yoshimasa Shibata, J. Am. Ceram. Soc., C-222 (1984). 17. T. Fujiu and G. L. Messing, J. Mater. Synth. Process. 1, 33 (1993). 5. Huaxia Ji and P. Marquis, J. Mater. Sci. Lett. 10, 132 (1991). 18. J. Bridgwater, Mater. Design 14, 15–18 (1993). 6. G. J. Akimov, V. M. Timtschenko, and P. A. Arseneev, Ogneu- 19. D-J. Kim and P. Hrma, J. Am. Ceram. Soc. 75, 2959 (1992). pory. Nos. 1–5, 19 (1994). 20. J. P. Zhong, Z. Fathi, G. P. Latorre, D. C. Folz, and D. E. Clark, 7. E. G. Nordstrom and K. H. Karlsson, J. Mater. Sci.: Materials in Ceram. Eng. Sci. Proc. 15, 1003 (1994). Medicine 1, 182 (1990). 21. T. Woignier, J. Phalippou, and M. Prassas, J. Mater. Sci. 25, 3118 8. P. E. Wang and T. K. Chaki, J. Mater. Sci.: Materials in (1990). Medicine 4, 150 (1993). 22. F. Wakai, Y. Kodama, and S. Sakaguchi, J. Am. Ceram. Soc. 73, 9. K. Ioku, S. Soymia, and M. Yoshimura, J. Mater. Sci. Lett. 8, 457 (1990). 1203 (1989). 23. K. Yamashita, K. Kitagaki, and T. Umegaki, J. Am. Ceram. 10. M. Takagi, M. Mochida, N. Uchida, K. Saito, and K. Uematsu, Soc. 78, 1191 (1995). J. Mater. Sci.: Materials in Medicine 3, 199 (1992).

202 J. Mater. Res., Vol. 12, No. 1, Jan 1997