J O U R N A L O F C H E M I S T R Y Materials Growth of calcium phosphate onto coagulated silica prepared by using modified simulated body fluids Juan Coren�o,a Rogelio Rodrý�guez,*a,b Miguel A. Araizac and Victor M. Castan�ob aDepartamento de Fý�sica, Universidad Auto�noma Metropolitana-Iztapalapa, Apdo. Postal 55-534, Me�xico, D.F. 09340 bInstituto de Fý�sica, Universidad Nacional Auto�noma de Me�xico, Apdo.Postal 1-1010, Quere�taro, Qro. Me�xico 76001 cFacultad de Odontologý�a, Universidad Nacional Auto�noma de Me�xico, Ciudad Universitaria, Me�xico, D.F. 04510 Received 11th May 1998, Accepted 25th September 1998 Silica sols prepared by the alkaline hydrolysis of tetraethylorthosilicate were coagulated by adding an excess of a CaCl2 aqueous solution. The aggregates were immersed into three diVerent modified simulated body fluids at 90 °C to allow the growth of a calcium phosphate phase onto the silica aggregates.The apatite phase grew faster compared to previous studies. Also, the amount of the crystalline apatite yield was higher when the simulated body fluid employed had the largest Ca/P ratio, as measured by X-ray diVraction. The relevance of these findings is discussed in terms of the current and future trends in biomaterials research and development.A hydrated silica gel layer is formed on the glass surface Introduction and provides favorable sites for apatite nucleation. It has The preparation of synthetic biomaterials, aimed to be been suggested in the literature11 that, in contrast to dense employed for prosthestic applications in living organisms, has silica, this silica layer produced on the bioglasses, is flexible attracted growing interest in the last few decades.For instance, enough to provide the oxide–oxide spatial requirements to apatite coatings have been applied onto diVerent substrates to match the bone lattice, thus providing epitaxial sites for bone produce a number of interesting materials with potential growth.biomedical applications.1 The development of novel nano- The potential role of hydrated silica species on biological composites opens a broad new field of research since, in mineralization9 also points out the relevance of producing principle, it is possible to control, to a great extent, the materials with high surface areas, controlled porosity and corresponding morphology, which in turn is related to the appropriate chemical groups available for the desired interbiomedical and physical properties.2 actions.One possible approach is the coating of diVerent A number of important requirements, in addition to the substrates with biocompatible apatite layers, by using various obvious biocompatibility, ought to be fulfilled if synthetic techniques, which range from standard vacuum evaporamaterials are to be used for human applications. In fact, one tion to chemical vapor deposition (CVD) technologies. of the most important characteristics is the corresponding Unfortunately, long-term animal studies suggest that the HAp porous structure, since the extra-cellular fluids must be allowed coatings may degrade or come oV.9 to flow through the inner structure of the biomedical device One interesting method to produce calcium phosphate to allow an adequate osteoconduction.3,4 apatites at low temperature is through the immersion of silica Among the materials that are known to have an appropriate gels into simulated body fluid (SBF).12–17 Besides the promisbiological activity when implanted in living organisms calcium ing potential applications of these experiments, from a practical phosphates deserve special attention.5 Historically speaking, standpoint, the available reports in the literature reveal a very low reaction kinetics and a limitation on the number of the modern use of these materials for bone repair was probably available surface chemical groups for the reaction with the pioneered by Albee in the 1920s6 which are, nowadays, widely surrounding SBF, since the silica gels reportedly employed employed for low-loading hard tissue repair and augmentation only react on the areas exposed to the SBF.Accordingly, an of living bone.4 The excellent histological behaviour of these alternative method, by starting from silica sols rather than materials, either synthetic or natural, is attributed to their gels, which are coagulated in situ by adding controlled amounts chemical similarity to the mineral phase of natural bone, of calcium ions, and by using three diVerent modified SBFs, hydroxyapatite [Ca10(PO4)6(OH)2] or HAp, which is the is reported here.The higher temperature of reaction, namely classical example.In fact, the crystallographic and chemical 90 °C, makes the available reacting chemical groups more properties of HAp closely resemble those of bone and tooth reactive; additionally, due to the use of nanometer-sized sols minerals.7 and to the coagulation process, one can obtain faster reaction Bioactive glasses have also been used for implants because times, higher yields and more controllable conditions these materials bond to bone through an intervening bonecompared with previous studies.like apatite layer formed on their surfaces in a living bodylike environment.5,8 Interestingly this apatite layer is calciumdeficient compared with hydroxyapatite, being formed by small Experimental crystallites and containing small amounts of carbonate.9 In CaO–SiO2-based bioactive glasses,10 it has been found that Coagulated silica was obtained by preparing pure silica sol by silica forms a low solubility matrix in which the network of the sol–gel method under alkaline conditions of the hydrolysis silicate chains acts as a framework for ionic species (Ca2+, reaction, followed by a coagulation procedure of the silica PO43-, Na+, etc.), whose role is to stimulate the biochemical particles through the addition of an excess of calcium chloride.It is important to point out that, unlike previous reports in environment surrounding the bioactive glass. J. Mater. Chem., 1998, 8, 2807–2812 2807the literature, in the present studies, coagulated silica sols were 9 days. After this period the solid was washed and dried at 100 °C for 2 h.employed instead of the standard silica gels. SBF and mSBFs solutions were prepared by dissolving, at diVerent concentrations, reagent grade NaCl, NaHCO3, KCl, Sol preparation Na2HPO4·12H2O, MgCl2·H2O, CaCl2 or Na2SO4, buVered A solution of 4 moles of distilled water in 6 moles of ethanol at pH=7.4 by using tris(hydroxymethyl )aminomethane (reactive grade)(Baker Co.) was added to another solution of [(CH2OH)3CNH2] and hydrochloric acid. 1 mole of tetraethylorthosilicate (TEOS) (Aldrich Chem. Co.) in 6 moles of ethanol, under vigorous stirring. The starting Characterization techniques pH of water was adjusted to 12 by using NH4OH. The resulting mixture was poured into a round double-necked The dynamic light scattering (DLS) apparatus used to measure the particle size was a Brookhaven Instrument with a digital flask, and heated to reach reflux conditions.A profile of the silica particle size as a function of the reaction time was correlator model 9000; in all cases the scattering angle was set to 90°, the measurements were done at room temperature and obtained by sampling at regular time intervals by using dynamic light scattering techniques.Once a constant particle the light source was an argon ion laser operating at 488 nm. The phosphate concentration was measured by the molyb- diameter was obtained, the sol–gel reaction was stopped by diluting with ethanol and cooling the system to prevent denum blue method18 using a UV–VIS absorption spectrophotometer (Perkin Elmer Lambda 5) at 690 nm.The X-ray gelation. Silica sols prepared under these (basic) conditions are electrically stabilized; consequently, they can be coagulated diVractograms (XRD) of the samples were obtained in a Phillips diVractometer in the range 20–50° with a scanning at will by the addition of a suitable salt. rate of 2° min-1. Scanning electron microscopy (SEM) was carried out in a Zeiss model DSMSol coagulation ground and carbon-coated.m-Raman characterization was An aqueous 0.1 M CaCl2 solution was added dropwise to a carried out in a DILOR apparatus model Labram equipped fixed volume of colloidal silica suspension under stirring; the with a He–Ne laser as a light source and a confocal optical final concentration of calcium ions was 0.01 M.After stirring microscope; the wavenumber range of the scattered light was for one hour, the flocculate was allowed to settle under the varied from 100 to 1200 cm-1. 31P NMR spectra were taken influence of gravity; the liquid was decanted and the solid on a Bruker ASX300 NMR spectrometer using a 4 mm CP centrifuged. The solid part was washed twice by resuspending MAS probe at 5 kHz.H3PO4 was used as reference. it in distilled water and stirring gently for 20 min. The coagulates were dried at 400 °C for 2 days and ground by using an Results and discussion agate mortar. Fig. 1 shows the X-ray diVraction pattern corresponding to Calcium–phosphates growth pure HAp where the plane indices of the main reflections are shown. XRD for pure silica, prior to the immersion into the The growth of calcium phosphate crystals was achieved by mSBF and with the same thermal treatment as all samples, resuspending the silica coagulates in an aqueous solution was also obtained (not shown); for this sample no crystalline containing calcium and phosphate ions.Four diVerent ionic reflections were observed but only a broad peak around 23° solutions were used: due to the amorphous phase. 1) SBF at 37 °C XRD for samples containing calcium phosphate grown on their surfaces at diVerent immersion times in diVerent mSBFs 2) mSBF-1.0: modified SBF with Ca/P=1.00 at 90 °C are shown in Fig. 2–4. Fig. 2 corresponds to the diVractograms 3) mSBF-1.7: modified SBF with Ca/P=1.67 at 90 °C 4) mSBF-2.5: modified SBF with Ca/P=2.50 at 90 °C Table 1 summarizes the ionic composition for the four SBF solutions; human plasma is also reported for comparison purposes.Each experiment using mSBF was carried out in three round flasks. The first flask contains only 20 ml of mSBF, to rule out a possible spontaneous crystallization. The second flask contains 0.35 g of 0.125–0.25 mm ground silica in 30 ml of mSBF and the third contains 0.20 g of coagulated silica ground to sizes smaller than 0.125 mm in 20 ml of mSBF. The third flask was sampled on a regular basis to measure phosphate concentration in the solution as a function of time and to obtain X-ray diVractograms of the corresponding products.The experiment at 37 °C was taken as reference. In this case, the silica aggregates were used without previous drying.They Fig. 1 X-Ray diVractogram of pure crystalline hydroxyapatite. were added to the SBF and kept at 37 °C under agitation for Table 1 Ion concentration of standard blood plasma, SBF, and the modified SBF used for calcium phosphates growth onto silica aggregates Ion concentration/mM Ca/P Na+ K+ Ca2+ Mg2+ Cl- HCO3- HPO42- SO42- Blood plasma 2.5 142.0 5.0 2.5 1.5 103.0 27.0 1.0 0.5 SBF 2.5 142.0 5.0 2.5 1.5 147.0 4.2 1.0 0.5 Modified SBF, No. 1 1.67 142.4 — 2.0 — 144.0 — 1.2 — Modified SBF, No. 2 1.0 143.0 — 1.5 — 143.0 — 1.5 — Modified SBF, No. 3 2.5 142.0 — 2.5 — 145.0 — 1.0 — 2808 J. Mater. Chem., 1998, 8, 2807–2812Fig. 2 X-Ray diVractograms of samples immersed in mSBF with Ca/P=1.0 at 90 °C and diVerent immersion times: (a) 10 h, (b) 23 h, Fig. 5 X-Ray diVractogram of a sample prepared at 37 °C immersed (c) 33 h and (d) 48 h. in SBF for 7 days. equivalent to Fig. 2 but using mSBF-1.7 and mSBF-2.5, respectively. For comparison purposes, the diVractogram for the experiment at 37 °C is shown in Fig. 5. The diVractograms of Fig. 2–4 show, in addition to the contribution of the amorphous silica, the crystalline reflection corresponding to the apatite phase. As the Ca/P ratio in the mSBFs was increased, the intensity of the reflection was also increased.This is clearly observed in Fig. 6 where the intensity of the characteristic apatite peak is plotted as a function of the immersion time for the three diVerent mSBFs. This plot shows that, at the beginning of the crystal growth on the silica surface, the amount of the crystalline phase is small, corresponding to an induction period where the silica surface oVers multi-nucleating sites where the apatite can grow.Up to 33 h the amount of crystals remains practically constant as revealed Fig. 3 As for Fig. 2 but for samples immersed in mSBF with by the intensity of the reflection, which is nearly the same as Ca/P=1.67. after induction.However, after this time, for Ca/P ratios of 1.67 and 2.5, their heights have increased producing sharper and stronger X-ray reflections. Since the height of the diVraction peak is known to be proportional to the amount of crystals which contribute to the peak, there is a 2.2 fold increase in this crystal content for mSBF-2.5 compared with mSBF-1.0. It is important to mention that these growth times are considerably shorter than typical HAp formation periods reported previously.19 m-Raman spectra were obtained for all samples and pure HAp.The Raman spectrum for pure HAp is shown in Fig. 7(a), while the sample with Ca/P 2.5 is shown in Fig. 7(b) where it is possible to observe a small signal corresponding to the apatite superimposed on a strong fluorescence signal produced by the silica substrate.As before, only the strongest apatite band at 962 cm-1 due to the phosphate symmetric stretching vibration20 is observed because this phase represents only a small percentage of the resulting material. Similar Fig. 4 As for Fig. 2 but for samples immersed in mSBF with Ca/P=2.5. of the samples immersed in mSBF-1.0 at 90 °C for 10, 23, 33 and 48 h, respectively.The whole set of reflections for the crystalline phase is diYcult to obtain since it is the minor component of the resulting material (ca. 10%). It is worth mentioning that, since the apatite growth takes place on the surface of small coagulated silica particles, the silanol groups on their surfaces and the open structure produced by the coagulating process allow the coating to be both internal and external, unlike the methods which use silica gels; this means that the reflections may also contain contributions from the internal coating.Since the reaction conditions favor apatite formation,10 the characteristic reflection for the crystalline phase is located around 32°; this peak is assigned to an overlap of the diVraction bands of three crystalline spacings: (211), Fig. 6 Plot of the 32° reflection intensity as a function of the immersion time for samples immersed in the three mSBFs. (112) and (300) according to the literature.19 Fig. 3 and 4 are J. Mater. Chem., 1998, 8, 2807–2812 2809Fig. 10 Consumption of phosphorus as a function of time for all Fig. 7 m-Raman spectra of pure HAp (a) and for the sample Ca/P 2.5 experiments.immersed in mSBF for 48 h. spectra (not shown) were obtained for the remainder of apatites with Ca/P ranging from 1.14 to 1.66 (d 2.8±0.2) the samples. which is clearly diVerent from those corresponding to other Fig. 8 and 9 show the 31P MAS NMR spectra for pure HAp calcium phosphates.21 This result confirms the identification and for the sample immersed in mSBF-2.5 for 48 h, respect- of the crystalline phase formed as apatite.ively. In Fig. 9 only one 31P resonance with weak sidebands is Fig. 10 shows the change in P concentration for the four observed. The isotropic chemical shifts obtained are d 2.710 experiments. Both the rate and the amount of P consumption for pure HAp and d 2.924 for the sample. These values are in are similar for all the three experiments at 90 °C.The decrease accord with the chemical shift reported for calcium phosphate in P concentration for these experiments ends approximately 23 h after the beginning of the reaction, whereas for the experiment at 37 °C, it starts approximately at this point. It is interesting that, even for the experiment at 37 °C, which corresponds to the slowest apatite growth, an apatite growth rate faster than those reported for gels dried at 400 °C is observed.19 There are a number of factors which explain this behavior: (a) the present experiments were carried out using coagulated silica sols instead of the standard gels, which increases the active surface area available for the chemical reaction, (b) the higher temperature of reaction makes the available chemical groups more reactive so reducing the reaction time.Additionally, the Ca ions adsorbed on the silica sol may favour the growth of the apatite phase by increasing the ionic activity of the surrounding solution near the flocs. It is important to mention that, for the series of experiments without silica sols (i.e., containing only mSBF), there was no P consumption during the same time intervals reported here.SEM micrographs for the whole set of samples are shown in Fig. 11–15. Fig. 11 reveals a smooth surface corresponding to pure silica. The micrographs shown in Fig. 12–14 correspond to experiments where the apatite grew onto Fig. 8 31P MAS NMR spectrum of pure hydroxyapatite. Fig. 9 31P MAS NMR spectrum of the sample Ca/P 2.5 immersed in Fig. 11 SEM micrograph of pure silica. mSBF for 48 h. 2810 J. Mater. Chem., 1998, 8, 2807–2812Fig. 14 SEM micrograph from a sample with Ca/P=1.0 for an Fig. 12 SEM micrograph from a sample with Ca/P=1.67 for an immersion time of 48 h. immersion time of 48 h. Fig. 13 SEM micrograph from a sample with Ca/P=2.5 for an immersion time of 48 h. Fig. 15 SEM micrograph from a sample at 37 °C and 7 days immersion in SBF. 0.125–0.25 mm ground silica particles at 90 °C during 48 h. These do not show the typical spherical morphology that has been observed in similar work at 37 °C.14,19 modified by the presence of some ions. In this work it is observed that the Ca/P ratio aVects both the crystal size and Fig. 12 shows the micrograph corresponding to samples immersed in mSBF-1.7; this shows smaller and more uniform the way they cluster together, but not significantly the crystal morphology.It seems that the eVect of Ca/P ratio on the crystallite sizes (ca. 12 mm) with respect to the other two experiments at diVerent Ca/P ratios. Fig. 13 shows a micro- crystal morphology is diminished at the higher temperature employed for the apatite growth.graph corresponding to Ca/P =2.5 where the apatite crystals form a loose cluster with a wider particle size distribution It is interesting to note the dependence of the apatite morphology on the nature of the substrate. In this work, the compared with Fig. 12. These characteristics are not observed in Fig. 14 (Ca/P=1.0) where the crystals in the aggregate are sample prepared in SBF at 37 °C does not show the flake-like crystals reported when fired silica was used,22 but rather a tightly distributed. For comparison, the micrograph for the experiment at 37 °C uniformly distributed crystal layer without a defined morphology.Also, this layer was grown more rapidly than that is shown in Fig. 15. Here, it can be seen that this rough surface is quite diVerent from the other SEM micrographs.The apatite previously reported.19 These observations can be explained by the greater silanol density present in a wet coagulated silica phase shows smaller and poorly defined crystallites, compared with samples at 90 °C. This sample was kept in the SBF until sol, compared with fired silica, since it has been proposed that silanol groups are responsible for apatite nucleation.11 XRD characteristic peaks were observed (after 7 days).These results demonstrate the strong influence of temperature on the All the above results demonstrate that the silica is covered with a layer of apatite whose morphology depends on tempera- growth of the crystalline phase. For apatite growth onto silica fired at 400 °C22 it has been ture, immersion time and the type of substrate.The morphology of this nano-composite can also be modified, either reported that the morphology is Ca/P dependent, and can be J. Mater. Chem., 1998, 8, 2807–2812 28113 L. L. Hench and J. Wilson, Science, 1984, 226, 630. by changing the Si5Ca/P ratio, or by modifying the route of 4 F. H. Albee, Ann. Surg., 1920, 71, 32. adding the constituents. 5 R.Z. LeGeros, Calcium Phosphates in Oral Biology and Medicine, One important advantage of the present approach over the Karger, Basel, Switzerland, 1991. previous reports is that, since the silica sol was coagulated, 6 K. de Groot, in Contemporary Biomaterials, ed. J. W. Boretos and the active sites suitable for apatite growth are evenly distributed M. Eden, Noyes Publications, Park Ridge, NJ, 1984, pp. 477–492. over the whole sample surface, consequently it is expected that 7 R. H. Doremus, J. Mater. Sci., 1992, 27, 285. the crystalline phase will grow both inwardly and outwardly 8 M. M. Pereira, A. E. Clark and L. L. Hench, J. Am. Chem. Soc., 1995, 18, 2463. on the silica flocs. This shows the clear advantage of using 9 L. L. Hench, J. Am. Ceram. Soc., 1991, 74, 1487.sols rather than gels to allow the presence of more active 10 L. L. Hench and A. E. Clark, Biocompatibiity of Orthopedic groups available for the corresponding reaction. Additionally, Implants, ed. E. F. Williams, CRC Press, Boca Raton, FL, 1982, by choosing the appropriate starting silica particle size it is vol. 2, pp. 129–170. possible to control the porosity and the interstitial volume of 11 L.L. Hench and E. C. Ethridge, Biomaterials: An Interfacial the material. The immersion time, and, in turn, the thickness Approach, Biophysics and Bioengineering Series, ed. of the HAp layers, also slightly modifies the average pore size. A. Noordergraaf, Academic Press, New York, 1982, vol. 4, p. 139. 12 A. Ravaglioli and A. Krajewski, Bioceramics: Materials, Properties and Applications, Chapman & Hall, London, 1992, Conclusions pp. 140 and 175. 13 T. J. Kokubo, J. Non-Cryst. Solids, 1990, 120, 138. A novel composite material with potential biomedical 14 P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, applications prepared by coagulating silica sols with calcium T. Nakamura and T. Yamamuro, J. Am. Ceram. Soc., 1992, 75, ions, is reported. With this procedure, a silica substrate covered 2094.inwardly and outwardly with a crystalline apatite phase can 15 T. Kokubo, Biomaterials, 1991, 12, 1155. 16 R. Fresa, A, Constantini, A. Buri and F. Branda, J. Non-Cryst. be produced. The role of the liquid phase (SBF) was analyzed Solids, 1995, 16, 1249. by using several modified simulated body fluids. The kinetics 17 M. Tanahashi, T. Kokubo, T. Nakamura, Y. Katsura and of each process is certainly an area worth further study, not M. Nagano, J. Non-Cryst. Solids, 1996, 17, 47. only because of its relevance for the production of novel 18 C. Sung-Baek, N. Kazuki, T. Kokubo, N. Soga, C. Ohtsuki, biomaterials, but also because it opens the possibility of T. Nakamura, T. Kitsugi and T. Yamamuro, J. Am. Ceram. Soc., preparing other ceramic-like systems at low temperature. 1995, 78, 1769. 19 P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, T. Nakamura and T. Yamamuro, J. Mater. Sci. Mater. Med., 1993, Acknowledgments 4, 127. 20 M. A.Walters, Y. C. Leung, N. C. Blumenthal, R. Z. Legeros and The authors are indebted to Ing. Francisco Rodrý�guez K. A. Konsker, J. Inorg. Biochem., 1990, 39, 193. Melgarejo from CINVESTAV, Quere�taro and to Dra. 21 J. W. P. Rothwell, J. S.Waugh and J. P. Yesinowski, J. Am. Chem. Antonieta Mondrago�n from IFUNAM, for their valuable Soc., 1980, 102, 2637. assistance in the Raman measurements. 22 P. Li, K. Nakanishi, T. Kokubo and K. de Groot, Biomaterials, 1993, 14, 963. References 1 K. A. Khor, P. Cheang and Y. Wang, JOM, 1997, 49, 51. Paper 8/03503B 2 R. Rodrý�guez, J. Coren�o and V. Castan� o, Adv. Comp. Lett., 1996, 5, 25. 2812 J. Mater. Chem., 1998, 8, 280