Indian Journal of Chemistry Vol. 47A, November 2008, pp. 1626-1631 Synthesis of nano sized hydroxyapatite powder using sol-gel technique and its conversion to dense and porous bodies Iis Sopyan a, *, Ramesh Singh b & Mohammed Hamdi c a Department of Manufacturing & Materials Engineering, Faculty of Engineering, International Islamic University Malaysia, P.O. Box 10, 50728 Kuala Lumpur, Malaysia Email: sopyan@iiu.edu.my b Department of Mechanical Engineering, College of Engineering, University Tenaga Nasional, Malaysia c Department of Engineering Design and Manufacture, Faculty of Engineering, University Malaya, Malaysia Received 13 May 2008; revised 20 October 2008 Hydroxyapatite powder has been prepared via sol-gel procedure using calcium nitrate tetrahydrate and diammonium hydrogen phosphate as the precursors for calcium and phosphorus, respectively. XRD measurement shows that the powder contains hydroxyapatite crystals with β-TCP and calcium oxide as secondary phases. Hydroxyapatite powder of higher purity, i. e., the correct Ca/P ratio, has been obtained by adding an appropriate amount of diammonium hydrogen phosphate and heating with stirring. Morphological evaluation by SEM measurement shows that the particles of the HA are tightly agglomerated and globular in shape with an average size of 1-2 µm. The primary particulates have average diameters of 50-200 nm, as detected by SEM and nanoparticle sizer. Purity (almost 100%) of the obtained hydroxyapatite has been confirmed by XRD analysis. Its performance has been tested by making dense and porous samples. IPC Code: Int Cl. 8 B82B3/00; C01B25/45; C01F11/00 Hydroxyapatites (HA) are particularly attractive materials for bone and tooth implants since they closely resembles human tooth and bone mineral and have proved to be biologically compatible with these tissues. 1 Since the use of hydroxyapatite for the first time in 1981 for periodontal lesion filling, its use in the medical field has extended to solid blocks, solid components, and films for dental implants. Many studies have shown that HA ceramics show no toxicity, inflammatory response, pyrogenetic response, or fibrous tissue formation between implant and bone. Also, these materials have the ability to bond directly to the host bone. 2 The main limitation of HA ceramics as well as all other bioactive materials is that they have poor mechanical properties. Basically, all bioceramics which have good mechanical properties and suitable for load bearing applications should be bioinert. Hydroxyapatite, on the other hand, has high bioactivity, with many medical applications in the form of porous, dense, granules, and, as coatings. 2,3 Several research groups have developed prepara- tive procedures for hydroxyapatite. Traditionally, two main methods are employed for preparation of HA powders: wet (chemical) method (including precipitation method, 4 hydrothermal technique, 5 and hydrolysis 6 ) and dry (solid state reaction) method (refs cited in Ref. 7). Differences in the preparative routes leads to variations in morphology, stoichiometry, and level of crystallinity. Other methods, such as sol-gel, 8,9 spray pyrolysis, 10,11 mechano-chemical method, 12 have also been developed recently and are well documented. 7 Sol-gel procedure was first employed for the preparation of HA by Sakka and co-workers. 8 They used calcium diethoxide and phosphorus triethoxide as starting materials. Hydrolysis – polycondensation of the monomers in neutral and acidic solutions gave HA powders of high purity. Extraordinarily fine amorphous particulates with less than 10 nm in diameter were obtained from precipitation of the solutions, which increased to only ca. 100 nm after calcination at 900°C. In this present study, we have developed sol-gel procedure for preparing HA powder. It is well known that sol-gel techniques have several advantages for producing ceramic particulates of high purity, high crystallinity, nano sizes, and high reactivity. Sol-gel process, however, has some drawbacks such as expensive raw materials and low homogenity of the final product. We report herein a novel sol-gel method for preparing extraordinarily fine hydroxyapatite SOPYAN et al.: SYNTHESIS OF NANO SIZED HYDROXYAPATITE POWDER USING SOL-GEL TECHNIQUE 1627 powder, utilizing easily obtainable raw materials of relatively low cost. Simplicity of experimental execution, in respect of methods employing wet chemical reaction, is one of the most important advantages offered by this method. Physico-chemical characterization of the hydroxyapatite powder obtained from the sol-gel procedure has been carried out. Morphology as well as the mechanical properties of dense and porous bodies prepared using the powder have also been investigated. Materials and Methods For preparation of hydroxyapatite powder, calcium nitrate tetrahydrate and diammonium hydrogen phosphate (reagent grade) were used as calcium and phosphorus precursors, respectively. Both reagents were purchased from Merck KGaA, Germany. Urea (R&M Chemicals, UK) was used as gelling and ammonium donor agent. EDTA (Merck KGaA) was used as chelating agent to prevent an immediate precipitate formation calcium ions in the course of gel formation. The reaction was conducted in basic solution using ammonium solution (R&M Chemicals, UK) as solvent. Preparation of the stoichiometric hydroxyapatite powder Ammoniun solution was heated at 60°C, and EDTA (181 g) was added while stirring until it dissolved. Into this, 200 mL aqueous solution of 129 g calcium nitrate tetrahydrate was poured. Diammonium hydrogen phosphate (39.83 g) and urea (45.20 g) were subsequently added. The mixture was heated at 100°C for 3-4 h. The obtained gel was dried at 350°C under ambient static air and subsequently subjected to an 820°C calcination under flowing air. The powder was examined by X-ray diffraction techniques to determine the phases formed. It was observed that Ca/P molar ratio is ca. 1.8. Accordingly, to compensate the upward deviation from the stoichiometric ratio (1.667), the powder was mixed with an appropriate amount of diammonium hydrogen phosphate, followed by suspending in water and heating at 90°C with rigorous stirring. This procedure restored the Ca/P ratio of hydroxyapatite powder to 1.67. After drying, pure hydroxyapatite powder was obtained. Preparation of dense and porous samples For preparation of dense bodies, the as-prepared hydroxyapatite powder (50 g), was mixed with poly(vinyl alcohol) of 15.000 MW(2.5 g) and 100 ml distilled water. The suspension was homogenized using a magnetic stirrer followed by spray-drying on a Buchi mini spray dryer (B-290 type). Well dispersed powders obtained were compressed uniaxially using cold pressing technique at 800 kg/cm 2 . The pellets thus obtained were sintered in air at 1250°C for 1h. The same procedure was applied for a commercial hydroxyapatite powder (Sigma Aldrich). For preparation of porous bodies, slurries were prepared with the as-prepared HA powder, and adjusted loading of HA (ca. 25 wt%), using Duramax of D-3021 or D-3005 type (Rohm and Haas, USA) as the dispersant. Commercial cellulosic sponges were used and an initial impregnation procedure with a considerable fluidity slurry was employed. After soaking into the slurry, the sponges were dried in ambient air for 72 h and then subjected to heat treatment at 600°C for 1 h to eliminate organic matrix. Sintering was carried out at 1250°C for 1h. Characterization Scanning electron microscopic measurement for morphology evaluation of the powder was performed on a Jeol FESEM instrument (model JSM 6700F). The particle size distribution and mean particle size were measured using a back scattering method on a Malvern Nanosizer (model Nano-S). Differential and thermogravimetric analyses were performed on the as-prepared of HA powders and dried gel in ambient air using Perkin Elmer instrument (model PYRIS Diamond) with a 10°C/min heating rate. The crystalline phase compositions of the powders and of the dense samples were evaluated in a Rigaku diffractometer with copper K α radiation and a scan rate of 2° in 2θ min -1 . XRD patterns obtained were utilized for quantitative phase analysis according to literature. 13 Density of porous bodies was measured as apparent density (geometrical weight/volume measurement). The compressive strength was measured on cylindrical specimens (10 mm ht × 10 mm dia.) using an Instron 1195 apparatus. The compressive strength was calculated from the maximum load registered during the test divided by the original area. Several specimens were used for the testing. Density of cylindrical dense bodies was measured as apparent density (geometrical weight/volume measurement). Bars of the polycrystalline hydroxyapatite of dimension 25 mm × 2.0 mm × 2.5 mm were cut from cylindrical plates with a diamond saw. No further chemical treatment was INDIAN J CHEM, SEC A, NOVEMBER 2008 1628 performed on bars before testing. The flexural tests were conducted in four-point bending, using a Lloyd LR10K plus mechanical tester. Results and Discussion After refluxing the reaction mixture at 100°C for ca. 4 h, concentrated sol was formed. Subsequently it was converted to white gel through in situ solvent evaporation. Reaction of hydroxyapatite formation can be expressed as follows: 5Ca(NO 3 ) 2 + 3(NH 4 ) 2 HPO 4 + 4NH 4 OH Ca 5 (PO 4 ) 3 OH + 10NH 4 NO 3 + 3H 2 O …(1) The black dried gel obtained was then subjected to TG-DTA for thermal characterization. DTA-TG curves of the gel dried at 350°C shows the first weight drop of 10% at 100°C due to water evaporation. A subsequent decrease in weight of ca. 50% occurs until 700°C which is attributed to decomposition and elimination of ammonia, nitrate, urea, organic compounds, and carbon dioxide. There are two exothermic peaks in the curve; the first ranging from 350 – 420°C, is attributed to decomposition of ammonia and organic compounds. The second one, from 420 – 500°C is a large exothermic peak which may be due to decomposition of urea and carbon dioxide. Subsequently, calcination of dried gel at 820°C for 2h under flowing air converted it into hydroxyapatite powder. The yields of the HA powders were 90-95%. XRD pattern of crystalline phase of the powder after the heat treatment at 820°C (Fig. 1) shows that hydroxyapatite is the main component in the powder (ca. 75%). Calcium oxide (5%) and β-tricalcium phosphate (20%) were present as secondary phases. Generally, the powder mixture obtained at this stage contained 75-85%, 15-20%, and 4-6% of HA, β-TCP, and CaO, respectively. At this composition, the Ca/P molar ratio is about 1.8. Since optimum stoichiometric Ca/P molar ratio is 1.667, to compensate for the upward deviation of Ca/P, diammonium hydrogen phosphate was added to the suspension of the mixture HA, and heated at 90°C with rigorous stirring until the solvent was completely removed. In the XRD pattern of the pure powder obtained after this treatment the peaks of β-TCP and CaO disappear, proving 100% purity of the hydroxyapatite powder. We also checked the possiblility of the presence of calcium hydroxide in the powder. A phenolphtalein test shows that no the hydroxide is present in the powder. SEM photograph of the as-prepared HA powder (Fig. 2) shows that individual hydroxyapatite particles formed in globular shape with an average size of ~ 50-200 nm in diameter. The nanometric primary particles agglomerated tightly into micrometric aggregates of various shape and size. On the other hand, the particle size distribution of the hydroxyapatite powder as measured by nanoparticle sizer shows two separate size distributions; the lower distribution ranging from ca. 50 – 500 nm may be attributed to individual particles and the higher distribution from 2000 – 7000 nm may be attributed to tightly bonded particle agglomerates. Quite likely it Fig. 1 — XRD pattern of HA powder mixed with impurities CaO and β-TCP. Fig. 2 — SEM photograph of the sol-gel derived HA powder. SOPYAN et al.: SYNTHESIS OF NANO SIZED HYDROXYAPATITE POWDER USING SOL-GEL TECHNIQUE 1629 is difficult to disperse all agglomerates even after rigorous stirring for hours. The specific surface area measured by BET method gave a low value of 7 m 2 /g. This value is unusual for particles as fine as in the present case at nano levels, hence it is considered that the surface area measured by BET is for agglomerates and not of the particles. The particle size of the HA powder obtained in this study is considerably fine, as confirmed by SEM measurement, in respect of the HA powder prepared by sol-gel technique. Earlier studies have reported that sol-gel derived HA powders have particle size of about 100 nm in diameter 8 . Figure 3 shows the IR spectra of the dried gel (a) and hydroxyapatite powder (b). The dried gel’s spectrum clearly shows broad peaks, a characteristic of amorphous products. Bands of carboxylic and carbonate (1400-1600 cm -1 ), phosphate (900- 1100 cm -1 ), ammino (1400, 1600 and 3200 cm -1 ) and acetate (2800, 2300 cm -1 ) groups were detected. Obviously, CO 3 2- and HPO 4 2- groups are present, partially substituting groups of PO 4 3- and/or OH - in the HA structure. In Fig. 3(b), FTIR spectrum of the final powder shows the characteristic peaks corresponding to OH - (630, 3560 cm -1 ) and PO 4 3- (960, 1050, 1090 cm -1 ) vibrations, together with weak bands of CO 3 2- group at 870 and 1540 cm -1 , which indicates that the sol-gel HA powder is partially carbonated hydroxyapatite, as commonly observed in synthesis involving organic reagents. Dense and porous samples Dense and porous bodies were prepared to evaluate the performance of the obtained sol-gel HA powder. Dense samples were also prepared using commercial HA powder. Table 1 lists the physical properties of dense samples prepared using the sol-gel HA and commercial HA powders after sintering at 1250°C. The sintered dense bodies for the sol-gel derived HA powder showed flexural strength of 57.7 MPa and an apparent density of 2.855 g/cm 3 at 90% density (i. e., 10% porosity). Although the flexural strength is not so high, it is still in the range of the flexural strength of human cortical bones. The value is in fact much better than that obtained using commercial powder (36.0 MPa), at 89% relative density. Figure 4 shows XRD pattern of the cylindrical dense bodies. It is clear that α-TCP appeared at 2θ = 30.7° as the secondary phase after the sintering, Fig. 3 — FTIR spectra of (a) dried gel HA and (b) sol-gel derived HA powder. Table 1 — Physical properties of dense samples prepared using the sol-gel HA and commercial HA powders. Type of HA powder Sintering temp. (ºC) Flexural strength (MPa) Apparent density (g/cm 3 ) Rel. density (%) Sol gel 1250 57.7 2.855 90 Commercial 1250 36.0 2.810 89 Fig. 4 — XRD pattern of the HA dense body after 1250°C sintering. INDIAN J CHEM, SEC A, NOVEMBER 2008 1630 although with an acceptable quantity (less than 3%). The formation of α-TCP itself is unusual since normally β-TCP will form first at temperatures <1250°C, then α-TCP and tetracalcium phosphate appear at higher temperature. 14 This phenomenon was observed in HA powders prepared by precipitation method. It is deduced that the phenomena was attributed to the high reactivity of the obtained HA powder. However, this is an advantage since sintering process at lower temperatures suffice to obtain comparable particle densification level as indeed proven by the much higher mechanical strengths of the present HA dense bodies. Figure 5 shows SEM pictures of the fractured surface of dense samples. Micropores of 1-2 µm in diameter are present on the surface, which are responsible for high porosity (10%). In the dense samples. Besides micropores, a macropore of ca. 20 µm in size was found, representing a less- compacted region which was probably the origin of the fracture. Porous bodies were also prepared by the polymeric sponge method as reported in literature. 15,16 Figure 6 illustrates the resulting macrostructure of the porous hydroxyapatite. It is clear that the sample has circular open-cell pores of 500 microns - 2 mm in diameter. Presumably, the pore structures are similar to that of the original matrix. A 44% sintering shrinkage was obtained after 1250°C sintering, with an apparent density of 1.290 g/cm 3 and a relative porosity of 59%. On the other hand, measurement of mechanical properties of porous bodies provided a compressive strength of 1.96 MPa. It is important to note that conditions for slurry preparation have not been optimised asyet. In fact, large cavities were found in several samples as the result of incomplete slurry penetration into the pores. The slurry itself had to be ultrasonically treated to prevent coagulation. So, the compressive strength obtained, 1.96 MPa, was considered low for such value of apparent density (1.290 g/cm 3 ). It is also reasonable to presume that such low compressive strength is partly attributed to high macroporosity of the sample. Figure 7 shows the microporos structure of the wall of the open pores in which micropores of 1-2 µm in diameter were observed. Hing et al. reported that they succceded in preparing porous HA with compressive stress of 1 – 11 MPa for the increased apparent density from 0.38 – 1.25 g/cm 3 . 17 Even though it is compared in respect of porosity, the value of 1.96 MPa is still low if we refer to the value reported Fig. 5 — SEM picture of the fractured surface of dense samples. Fig. 6 — Macrostructure of the porous hydroxyapatite. Fig. 7 — Microporosity structure of the wall of the open pores. SOPYAN et al.: SYNTHESIS OF NANO SIZED HYDROXYAPATITE POWDER USING SOL-GEL TECHNIQUE 1631 by Guicciardi et al 18 . (16.1 MPa) for the speciment of 56% total porosity. The preliminary test on powder performance shows that the sol-gel powders have excellent physical properties which may be used to produce dense and porous bioactive bone implants with desired properties. Conclusions Hydroxyapatite (HA) powders with particles having nano- to submicro-metric size were succesfully prepared via sol-gel procedure using calcium nitrate tetrahydrate and diammonium hydrogen phosphate as the precursors. The primary particulates have globular shape with an average diameter of 50-200 nm, as detected by SEM. The results of the the nanoparticle sizer measurements are in good agreement with those obtained by morphological investigation. All the powders obtained have good purity (nearly 100%). Dense samples were prepared successfully from the HA powder using cold pressing technique. Sintering at 1250°C led to a total porosity of 10%, with apparent density of 2.855 g/cm 3 . The mechanical test of the dense samples provided a higher flexural strength (57.7 MPa) than that using commercial HA powder (36.0 MPa), thus showing its suitability for load bearing bone implant applications. Porous bodies produced via sponge impregnation technique using the sol-gel derived HA powder have apparent density of 1.290 g/cm 3 , with 59% total porosity and 44% sintering shrinkage. The porous bodies contained open macropores of 0.5-2 mm diameter as the result of removal of the polymeric matrix as well as micropores of 1-2 µm diameter. The compressive strength measurement yielded a value of 1.96 MPa, proving that porous HA samples can be used for human spongy bone substitutes. Thus, physical characterization of the sol-gel HA powder followed by the preliminary test on powder performance in making porous and dense samples shows that the powder has excellent physical properties and may be to produce dense and porous bioactive bone implants with desired properties. References 1 Willmann G, Interceram, 42 (1993) 206. 2 Willmann G, British Ceram Trans, 95 (1996) 212. 3 Levernia C & Schoenung, J M, Ceram Bull, 70 (1991) 95. 4 Tas A C, Biomaterials, 21 (2000) 1429. 5 Fujishiro Y, Sato T & Okuwaki, A., J Mater Sci Mater Med, 6 (1995) 172. 6 Monma H & Kamiya T, J Mater Sci, 22 (1987) 4247. 7 Suchanek W & Yohsimura M, J Mater Res, 13 (1998) 94. 8 Masuda Y, Matsubara K & Sakka S, Nippon Seramikkusu Kyokai Gakujitsu Ronbunshi, 98 (1990) 1255. 9 Liu D M, Troczynski T & Tseng W J, Biomaterials, 22 (2001) 1721. 10 Luo P & Nieh T G, Mater Sci Eng, C3 (1995) 75. 11 Aizawa M, Itatani K, Howell F S & Kishioka A, J Ceram Soc Japan, 103 (1995) 1214. 12 Toriyama M, Ravglioli A, Krajewki A, Galassi C, Roncari E & Piancastelli A, J Mater Sci, 30 (1995) 3216. 13 Tampieri A, Celotti G, Sprio S & Mingazzini C, Mater Chem Phys, 64 (2000) 54. 14 Bertoluzza A, Cacciari S, Tinti A, Vasina M & Morelli M A, J Mater Sci Mater Med, 6 (1995) 76. 15 Galassi C, Landi E, Tampieri A & Celotti G, Am Ceram Soc Ceram Trans, 112 (2001) 215. 16 Saggio Woyansky J, Scott C E & Minnear W P, Am Ceram Soc Bull, 71 (1992) 1674. 17 Hing K A, Best S M & Bonfield W, J Mater Sci Mater Med, 10 (1999) 135. 18 Guicciardi S, Galassi C, Landi E, Tampieri A, Satou K & Pezzoti G, J Mater Res, 16 (2001) 163.