- Email: [email protected]
Formation of calcium-deficient hydroxyapatite from α-tricalcium phosphate
Biomaterials 19 (1998) 2209 — 2217
Formation of calcium-deficient hydroxyapatite from a-tricalcium phosphate Kevor S. TenHuisen1, Paul W. Brown* The Pennsylvania State University, Materials Research Laboratory, University Park, PA 16802, USA Received 10 February 1998; accepted 5 May 1998
Abstract This study investigated the factors influencing the kinetics of Ca HPO (PO ) OH (calcium deficient hydroxyapatite or CDHAp) 9 4 45 formation from a-Ca (PO ) (a-TCP). The kinetics of CDHAp formation were investigated by isothermal calorimetry at constant 3 42 temperatures ranging between 30 and 75°C and by changes in pH at 37.4 and 70°C. The calorimetric curves were characterized by two reaction peaks. Activation energies were calculated for the events resulting in these peaks. Values obtained were 48.4 and 67.7 kJ mol~1, respectively, indicating nucleation and growth mechanisms for both events. Temperature had a significant effect on the growth rate as indicated by a decrease in surface area (26.5—15.0 m2 g~1) of the CDHAp with increasing temperature (30—75°C). A linear relationship between hydrolysis temperature and CDHAp surface area was observed. The morphology of the CDHAp was plate-like and the crystallites became more regular as the reaction temperature was increased. A rapid elevation in pH upon mixing with water indicated the synthesis method initially used did not entirely eliminate slight compositional variations within the a-Ca (PO ) . Rapid elevation in pH retarded subsequent reaction. This effect was eliminated by increasing the duration of high3 42 temperature firing during a-TCP synthesizing. ( 1998 Elsevier Science Ltd. All rights reserved
1. Introduction The development of calcium phosphate cements (CPC) as a class of viable biomaterials has increased interest in their use as artificial hard tissues. One of the earliest CPCs investigated was based on the hydration of a-tricalcium phosphate (a-Ca (PO ) or a-TCP) [1]. a-TCP 3 42 was shown to hydrate to a calcium-deficient hydroxyapatite (CDHAp) in dilute aqueous solutions and at elevated temperatures (60—100°C) [2]. This calcium phosphate is unique in that it can hydrolyze by itself to form a cement. The rate of a-TCP hydration has been observed to increase with temperature and decrease with increasing pH. The composition of the apatite formed could be varied by adjusting the initial pH; as the pH was increased, the Ca/P of the apatite increased from 1.50 to 1.67 [3, 4]. Later it was shown that moderate a-TCP hydration rates could be obtained at near-physiologic * Corresponding author. 1 Johnson and Johnson Corporate Biomaterials Center, Somerville, NJ, USA.
temperatures. Hydration rates were significantly increased when monovalent cationic salts of sulfates, chlorides, nitrates, and a variety of organic species were used [5]. Conversely, analogous salts containing divalent cations acted as retarders. The extents of conversions varied between 0 and '90% when hydration was carried out for 3 h at 40°C. Recently, the setting behavior and property development of a variety of CPCs including those formulated with a-TCP and a number of other calcium phosphate salts and/or CaO or CaCO as the precursor(s) have 3 been investigated [6—10]. Depending on the proportions and reactivities of the precursors, a variety of products were formed. These included octacalcium phosphate (OCP, Ca (HPO ) (PO ) ) 5H O), dicalcium phosphate 8 42 44 2 dihydrate (DCPD, CaHPO ) 2H O), and hydroxyapa4 2 tites of variable composition (Ca (HPO ) (PO ) 10~x 4x 4 6~x (OH) ). Unfortunately, these studies did not report the 2~x extents of conversion for each formulation and only reported the setting times and/or pH. Although setting times provide information about the initial reaction rates, they do not indicate the extents or rates of reaction of these cements.
0142-9612/98/$—See front matter ( 1998 Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 1 3 1 - 8
2210
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
It has been shown in comparative studies [11—14] that the rate of implant resorption in vivo is a function of composition with HAp'b-TCP'a-TCP. Although this trend follows from the ‘relative’ solubilities of these phases [15], bulk density, pore size distribution, and surface area all play critical roles in the rates of resorption. Differences in these properties appear to explain the variability observed [16] in the resorption rates in vivo of preformed HAp (Ca/P"1.45) synthesized from a-TCP and DCPD. These implants were observed to bond to bone and resorb over time. New bone formed between the implant and host bone with no indication of an inflammatory reaction. Conversely, others [17] did not observe resorption of implants consisting of partially hydrated a-TCP blocks placed subperiosteally on rabbit mandibles. However, new bone did form between the native tissue and implant. In vitro studies have been performed on the dissolution behavior of calcium orthophosphates including that of a-TCP in a calcium- and phosphate-free Tris buffer solution [18]. Among the salts investigated, it was found that the dissolution rates of both tetracalcium phosphate (TetCP, Ca (PO ) O) and a-TCP were the highest. Pre4 42 cipitation studies [19] on these calcium orthophosphates in a simulated physiologic solution imitating plasma indicated the formation of amorphous calcium phosphate (ACP) on the surfaces of the a-TCP. Based on FTIR spectroscopy, the a-TCP disappeared within 1 week. In spite of these citations, an in-depth investigation of the kinetics of the hydration of a-TCP to CDHAp has not been performed. The present study investigates relationships between the rates of hydrolysis, the reaction temperature, and microstructural development. Establishing these relationships is required to determine the reaction mechanism(s) and rate limiting steps and is essential in tailoring the hydration of a-TCP for in vivo applications.
2. Materials and methods 2.1. Precursor synthesis a-TCP was synthesized by firing a mixture of calcium carbonate and ammonium hydrogen phosphate at 1150°C for 2 h (a-TCP1). In one instance, synthesis was performed by firing at 1300°C for 24 h (a-TCP2). The mechanistic path for this reaction has been discussed in an earlier publication [20]. The fired product was identified as a-TCP by X-ray diffraction. This solid was milled to a mean particle size of 4.7 lm with a standard deviation of 2.8 lm, as determined by light scattering analysis. The surface area of the a-TCP was determined to be 1.3 m2 g~1 as measured by single-point nitrogen adsorption after outgassing at 150°C for 45 min.
2.2. Calorimetry Isothermal calorimetric analyses of the rates of hydrolysis of a-TCP were performed at various temperatures. All experiments were performed at a liquid-to-solids ratio of 1 : 1 (v/w). In a typical experiment, 3000 g of precursor powder was weighed into a Teflon-coated copper calorimeter cup. The cup was sealed with ParafilmTM to minimize the endothermic effect of water evaporation and then placed into the body of the calorimeter. Deionized water (3.0 cm3) was drawn into a syringe and placed into a copper manifold. The body of the calorimeter and the copper manifold were connected to a constant temperature water bath with a temperature variance of $0.01°C. Upon thermal equilibrium, the water was injected into the calorimeter cup. Thermopiles surrounding the copper cup detected small temperature gradients between the hydrolysis occurring in the calorimeter cup and the body of the calorimeter. This temperature gradient resulted in a voltage output. The voltage output was used to calculate the rate of heat evolution (mW) using a thermoelectric calibration constant. Data were collected by a microprocessor interfaced to the calorimeter through an analog to digital converter. A datum point was collected every 5—15 s depending on the temperature of the reaction. Rates of reaction and cumulative heat were both normalized to moles of CDHAp formed. The activation energy of the hydrolysis of a-TCP to CDHAp was determined from calorimetry experiments performed at 30.0, 37.4, 45.0, 55.0, 65.0, and 75.0°C. 2.3. Solution chemistry analyses Calcium and phosphate concentrations were monitored along with pH during the hydration of a-TCP to CDHAp. A 400 ml jacketed reaction vessel connected to a constant temperature water bath adjusted to 37.4°C ($0.1°C) was filled with 375 ml of deionized water. Upon thermal equilibration, 37.5 g of a-TCP were added to the deionized water. The powder was stirred continuously at 200 rpm with a Teflon-coated stir bar. Nitrogen gas was bubbled through the suspension to minimize the formation of carbonates. The pH was measured every 20 s with a combination glass electrode calibrated using pH 4.00 and 10.00 buffers. The pH meter was connected to a laptop microprocessor for data collection. At selected times &10 ml aliquots of slurry were removed and centrifuged for &30 s to separate the solids. The supernatant was filtered through 0.22 lm PTFE syringe filters and analyzed for calcium and phosphate by DC plasma emission spectroscopy. Solids were frozen in liquid nitrogen to arrest further reaction and freeze-dried. The pH was also monitored during the hydration of a-TCP at 75.0°C. The experimental setup was the same as described above. A pH datum point was collected
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2211
every 5 s due to the more rapid reaction. Aliquots of the suspension were extracted at selected times, filtered, and rinsed in acetone. The solids were then dispersed in acetone, filtered, and dried under vacuum over P O . 2 5 2.4. Reaction product characterization All samples were characterized by X-ray powder diffraction (Scintag PAD V) utilizing copper k radiation. a Samples were scanned over a 2h range of 3 to 40° at a scan rate of 2° 2h per min. Single-point nitrogen adsorption (Quantachrome Monosorb MS-12) was performed on CDHAp formed from the hydration of pure a-TCP after drying overnight at 150°C and outgassing for 45 min at 150°C. Selected samples were lightly coated with gold and observed by scanning electron microscopy (ISI DS-130).
3. Results and discussion 3.1. Isothermal calorimetry The effects of temperature on the kinetics of a-TCP hydrolysis in deionized water are shown in Figs. 1 and 2. Times to complete reaction increase from about 2 h to about 2 days as the hydration temperature is decreased from 75.0 to 30.0°C. X-ray diffraction analysis indicated all reactions approached completion in that only CDHAp was observed. The hydration reaction proceeds in two steps. The first step is a period of rapid hydrolysis resulting in a
Fig. 2. Rate curves for the hydration of a-TCP to CDHAp at various temperatures. The scales of the ordinates (W mol~1) of the inset figure and the main figure are the same. The first reaction peak is apparent in the inset.
calorimetric peak which shows a maximum between 5 and 30 min depending on the temperature (see inset to Fig. 2). Sufficient heat is evolved during this event to induce the initial set. It has been shown [5] that 7—8% conversion of a-TCP will result in setting. This corresponds to 8—10 kJ mol~1 of heat evolution indicating hydrolysis at 37.4°C should result in set within the first hour of reaction. After this initial reaction peak, there is a period of slow heat evolution for hydrolysis temperatures below 55.0°C. This initial peak is followed by the occurrence of a second calorimetric peak. The majority of the heat of reaction is evolved during this second reaction step. The second thermal event resembles the behavior observed in the hydration of silicate cements [21]. Activation energies (E ) for the two reaction steps ! during a-TCP hydrolysis were determined. An Arrhenius relationship between the rate of hydrolysis and temperature was found (see Fig. 3). The slopes of these curves are equal to !E /R. The E for the first reaction ! ! peak was determined to be 48.4 kJ mol~1. A value of 67.7 kJ mol~1 was obtained for the second reaction peak. Values of this magnitude indicate that both thermal events are controlled by a nucleation and growth mechanism. 3.2. Surface area
Fig. 1. Heat evolution curves for the hydration of a-TCP to CDHAp at various temperatures. The reaction temperature in °C is shown for each curve.
Effects of temperature on the surface area of CDHAp formed by a-TCP hydration are shown in Fig. 4. A linear relationship is observed. The surface area of the CDHAp
2212
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
reacted at temperatures above 35°C to form CDHAp or stoichiometric HAp (Ca/P"1.67) [22] although apatites having much higher surface areas were produced. The differences in HAp surface area resulting from reaction between DCPD and TetCP and that of a-TCP having a similar particle size is related to the rates of HAp formation at equivalent temperatures. The DCPDTetCP reactions occur more rapidly, thereby, favoring a higher nucleation density and thus a smaller average HAp crystallite size. 3.3. Microstructure
Fig. 3. Arrhenius plots determining the apparent activation energy (E ) ! of the first and second reaction peaks of a-TCP hydration.
Figure 5 shows the microstructure of the a-TCP and those of CDHAp formed by its hydrolysis at various temperatures. As discussed, the a-TCP (Fig. 5a) had a mean particle size of 4.7 lm and a surface area 1.3 m2 g~1 with a microstructure typical of a milled solid-state product. Figure 5b shows the morphology of CDHAp formed at 37.4°C to be characterized by thin irregularly shaped plates. Little evidence of the original a-TCP morphology is evident other than an occasional spherical particle composed of many small crystallites. The microstructure is characterized by interpenetrating plates showing a high degree of interconnectivity. There appears to be a high degree of intergrowth between the initial reactant particles even though the water was added by inoculation, not by mechanically mixing. The effect of reaction temperature on CDHAp microstructure can be seen by comparing Fig. 5b—d. CDHAp formed at higher temperatures exhibits a more regular microstructure. Similarly, increases in microstructural regularity have previously been observed in CDHAp formed from DCPD and TetCP [22]. Three microstructural features are apparent after reaction at 75°C (Fig. 5c and d): elongated hexagonal plates, rectangular parallelogram shaped plates, and needles. There is again a high degree of interconnectivity between the interpenetrating plates and little evidence of the initial a-TCP particles. These plates exhibit a morphology reminiscent of OCP, but it will be demonstated that these reactions did not form OCP as an intermediate. 3.4. Solution chemistry
Fig. 4. CDHAp surface areas resulting from the hydrolysis of a-TCP as a function of the reaction temperature. The measurements were carried out on the products from isothermal calorimetry.
formed decreases from 26.5 to 15 m2 g~1 as the temperature is increased from 30.0 to 75.0°C. The decrease in surface area with increasing temperature indicates that the ratio of nucleation to growth rates decreases with increasing temperature, although the overall reaction rate increases significantly with increasing temperature. A similar trend was observed when DCPD and TetCP
Figure 6 shows the results of solution chemistry experiments performed during a-TCP hydrolysis. The curve labeled a-TCP1 corresponds to the variations in pH when a-TCP formed at 1150°C (2 h) was hydrated at 37.4°C. Companion variations in the calcium and phosphate concentrations are also shown. Initially the pH rises to &10.8 and then slowly decreases over time. The pH would not be expected to rise to this value as a consequence of the congruent dissolution of compositionally homogeneous TCP. During the first hour of reaction the calcium concentration increased to &1.1 mM and then
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2213
Fig. 5. Scanning electron micrographs of a-TCP and CDHAp obtained from isothermal calorimetry experiments: (a) a-TCP precursor powder; (b) CDHAp from the hydration of a-TCP at 37.4°C; (c) and (d) CDHAp from the hydration of a-TCP at 75.0°C. The magnification in (a—c) is 5 k]; (d) is 10 k].
decreased to 0.075 mM. This change in calcium concentration correlates with the first calorimetric reaction peak as discussed previously. The phosphate concentration remains nominally constant at values around 0.01 mM during this period. Two hypotheses can explain these observations. a-TCP dissolution could be occurring incongruently, or there could be slight compositional inhomogeneities in the a-TCP as a result of the synthesis conditions. Compositional variations would result in calcium-rich regions likely to hydrolyze more rapidly. To establish if such variations were likely, the variations in pH during the hydrolysis of a-TCP formed at 1300°C for 24 h (a-TCP2) were determined. The resulting pH curve, labeled a-TCP2, is shown in Fig. 6. The pH reached a maximum of &9.2 before decreasing to a value of &8.75 where it remained until &5 h. This comparison indicates slight compositional differences exist in the aTCP fired at 1150°C for 2 h. While these data do not
preclude incongruent a-TCP dissolution, dissolution studies are difficult to perform on compounds lacking true solubilities, although solubility products have been reported for both TetCP [23] and a-TCP [24]. The variations of the calcium and phosphate concentrations with pH observed in this study compare favorably to the calculated solubility isotherm of a-TCP shown by Chow [15]. In contrast to reactions between DCPD and TetCP to form HAp [22], where the calcium concentration is higher than the phosphate during the formation reaction, the Ca/P molar ratio in solution is very close to 1 during the hydration of a-TCP. This suggests the apatite initially forming has a Ca/P'1.50. HAp does not have a fixed composition and it has been shown that its composition varies with the solution pH [25]. A Ca/P ratio greater than 1.5 would be expected to form at pH values observed during the first several hours of hydration.
2214
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
Fig. 6. Calcium (Ca, open circles) and phosphate (P, filled circles) concentrations and pH as a function of a-TCP hydration time. a-TCP1 was synthesized by firing at 1150°C for 2 h. a-TCP2 was synthesized by firing at 1300°C for 24 h.
X-ray diffraction analyses of solids extracted as a function of hydration time at 37.4°C (see Fig. 7) confirm the bulk of the hydration occurs between &8 and 20 h as indicated by isothermal calorimetry. The pH decreases from &7.7 to &7.0 during this period. It has been previously reported that a significant increase in the hydration rate of a-TCP occurs at pH values below 7.0—7.5 [5]. As the pH decreases, the calcium and phosphate concentrations increase as expected from prior solubility studies and thermodynamic calculations [15]. Complete reaction, as indicated by X-ray diffraction analyses, occurs at &21 h although most of the a-TCP has been consumed by 18 h (see Fig. 7). Upon complete a-TCP consumption, the calcium and phosphate concentrations increased as the pH continued to decrease until the experiment was terminated (4 weeks). Similar trends were observed when CDHAp formed from DCPD and TetCP. The Ca/P ratio in solution continued to decrease until it reached a value of 0.42 at 4 weeks. This Ca/P value would be expected at a pH of 4.95 because the prominent ionic species in solution are H PO~ and 2 4 Ca2`. These solution conditions result from the incongruent solubility of CDHAp [22]. The effects of the initial pH value on kinetics can be seen in the curves labeled a-TCP1 and a-TCP2 in Fig. 6. Complete a-TCP consumption occurs shortly after the final decrease in pH in every experiment. A reduction in the initial pH rise by reducing compositional fluctuations in the a-TCP reduces the time to complete the reaction by &2.5 h. A pH curve for a-TCP1 hydration at 75.0°C is also shown in Fig. 6. This experiment was performed to determine if OCP forms as an intermediate. As previously discussed, the morphology of CDHAp formed
Fig. 7. X-ray diffraction patterns as a function of a-TCP hydration time at 37.4°C. The times of hydration are labeled.
at 75.0°C is reminiscent of OCP. As previously observed [22] increasing the temperature reduces the pH at any given stage of reaction and increases the rate of reaction involving the formation of HAp. X-ray diffraction analyses performed on samples extracted at 10 min intervals indicated the absence of OCP and complete reaction to CDHAp occurring at 1.5 h. 3.5. Microstructural evolution Figure 8 shows microstructural evolution as a function of a-TCP hydration time at 37.4°C. These solids were extracted in the determinations of calcium and phosphate concentrations and pH for the curves labeled aTCP1 in Fig. 6. Microstructural evidence of a product phase are evident as early as 1.5 h (see Fig. 8a). Small needle-like crystals are occasionally observed on some of the a-TCP particles. At 5 h, small needles and platelike crystallites were present on the surfaces of all of the a-TCP particles (see Fig. 8b). However, evidence of HAp could not be observed by X-ray diffraction until about 9—10 h (see Fig. 7). By this time the majority of the a-TCP
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2215
Fig. 8. Microstructural evolution during a-TCP hydrolysis to CDHAp at 37.4°C. Times are (a) 1.5 h (20 k]); (b) 5 h (15 k]); (c) 10 h (10 k]); (d) 13 h (10 k]); (e) 21 h (10 k]). All bars on the micrographs equal 1 lm.
surfaces are covered with submicron needles and plate (see Fig. 8c). The microstructure at 13 h (see Fig. 8d) suggests complete hydration has occurred. Although the original
a-TCP surface cannot be observed, X-ray diffraction indicates a significant amount of a-TCP is still present (see Fig. 7). Isothermal calorimetry indicates less than 65% of the total heat produced has been evolved. The
2216
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
predominant morphological feature is submicron plates growing perpendicular to the a-TCP surfaces. Even though the a-TCP particles appear to be isolated from the solution by apatite, the perpendicular orientation of these plate-like HAp crystallites are conducive to allow ion transport and do not appear to retard the rate of hydrolysis. Figure 8e shows the CDHAp microstructure at complete reaction. The CDHAp morphology is characterized by irregularly shaped hexagonal plates of larger size than those observed at 13 h. The formation of ‘amorphous calcium phosphate’ (ACP) intermediate was observed when CDHAp formed at 37.4°C by reaction of DCPD and TetCP [22]. Unlike reactions involving DCPD and TetCP, the hydration of a-TCP does not appear to form significant amounts of ACP as indicated by X-ray diffraction. In addition, CDHAp formed by a-TCP hydration exhibits a lower surface area and higher crystallinity than that formed from DCPD and TetCP. Although higher pH values favor the formation ACP [26], lower calcium and phosphate concentrations in solution reduce the likelihood of ACP formation. 4. Conclusions Hydrolysis of a-TCP to CDHAp occurs by a mechanism involving two steps. The first step occurred within the first hour and resulted in high initial pH values commensurate with a spike in the calcium concentration. This initial step was a result of slight compositional inhomogeneities attributable to synthesis conditions initially employed. The elevated pH also resulted in a reduction of the rate of hydrolysis. The second reaction step resulted in the bulk of hydrolysis. The activation energies for these steps were determined; they indicate a nucleation and growth mechanism. The microstructure of the resulting CDHAp differed greatly from that produced by the DCPD-TetCP reaction at equivalent temperatures. In particular, crystallites are much larger, resulting in lower surface areas, and the hydroxyapatite formed is more crystalline that that produced by reaction of DCPD with TetCP at equivalent temperatures. This difference is attributable to the slower rate of TCP hydrolysis combined with the absence of the formation of ACP as an intermediate. The CDHAp crystallites grew perpendicular to the a-TCP surface, thus minimizing the formation of a diffusion barrier. Understanding the factors influencing the hydrolysis of calcium phosphate salts to apatites by cement reactions is essential in controlling the setting times and times to complete reaction. The properties must be understood and controlled if these types of materials are to be used in vivo applications. The crystallite size, morphology, and the crystallinity of CDHAp from a-TCP hydration should change its resorption characteristics when compared with HAp formed by other techniques.
Acknowledgements The authors gratefully acknowledge the support of NIDR Grant DE09421.
References [1] Monma H, Ueno S, Kanazawa T. Properties of hydroxyapatite prepared by the hydrolysis of tricalcium phosphate. J Chem Tech Biotechnol 1981;31:15—24. [2] Monma H, Kanazawa T. The hydration of a-tricalcium phosphate. Yogyo-Kyokai-Shi 1976;84:209—13. [3] Monma H, Ueno S, Tsutsumi M, Kanazawa T. Changes in composition and structure of tricalcium phosphate under different pH conditions. Yogyo-Kyokai-Shi 1978;86:590—7. [4] Monma H, Kanazawa T. Wet-process formation of nonstoichiometric hydroxyapatite from tricalcium phosphate. YogyoKyokai-Shi 1978;86:73—6. [5] Monma H, Goto M, Kohmura T. Effect of additives on hydration and hardening of tricalcium phosphate. Gypsum and Lime 1984;188:11—6. [6] Driessens FCM, Boltong MG, Bermudez O, Planell JA. Formulation and setting times of some calcium orthophosphate cements: a pilot study. J Mater Sci: Mater Med 1993;4:503—8. [7] Driessens FCM, Boltong MG, Bermudez O, Planell JA, Ginebra MP, Fernandez E. Effective formulations for the preparation of calcium phosphate bone cements. J Mater Sci: Mater Med 1994;5: 164—70. [8] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Compressive strength and diametrial tensile strength of some calciumorthophosphate cements: a pilot study. J Mater Sci: Mater Med 1993;4:389—93. [9] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Development of some calcium phosphate cements from combinations of a-TCP, MCPM, and CaO. J Mater Sci: Mater Med 1994a;5: 160—3. [10] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Optimization of a calcium orthophosphate cement formulation occurring in the combination of monocalcium phosphate monohydrate with calcium oxide. J Mater Sci: Mater Med 1994b;5:67—71. [11] de Groot K. Clinical usefulness of calcium phosphate ceramics. Zahnarztl Mitteil 1985;75:1938—40. [12] Klein CP, Patka P, den Hollander W. Macroporous calcium phosphate bioceramics in dog femora: a histological study of interface and biodegradation. Biomaterials 1989;10:59—62. [13] Krause A, Wehrbein H, Jacobs HG, Morich S. An alpha-modified tricalcium phosphate-ceramic for filling large bone defects after cystectomy in the anterior maxillary region. A report of clinical experience. Dtsch Z Mund- Kiefer- Gesichts Chir 1986;10:108—12. [14] Renooij W, Hoogendoorn HA, Visser WJ, Lentferink RHF, Schmitz MG, Van Ieperen H, Oldenburg SJ, Janssen WM, Akkermans LMA, Wittebol P. Bioresorption of ceramic strontium85-labeled calcium phosphate implants in dog femora. Clin Orthoped Relat Res 1985;197:272—85. [15] Chow LC. Development of self-setting calcium phosphate cements. J Ceram Soc Japan 1991;99:954—64. [16] Kurashina K, Ogiso A, Kotani A, Takeuchi H, Hirano M. Histological and microradiographic evaluation of hydrated and hardened a-tricalcium phosphate/calcium phosphate dibasic mixtures. Biomaterials 1994;15:429—32. [17] Nagase M, Chen R-B, Asada, Y, Nakajima T. Radiographic and microscopic evaluation of subperiosteally implanted blocks of hydrated and hardened a-tricalcium phosphate in rabbits. J Oral Maxillofac Surg 1989;47:582—6.
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217 [18] Ducheyne P, Radin S, King L. The effect of calcium phosphate ceramic composition and structure on in vitro behavior. I. Dissolution. J Biomed Mater Res 1993;27:25—34. [19] Radin S, Ducheyne P. The effect of calcium phosphate ceramic composition and structure on in vitro behavior. II. Precipitation. J Biomed Mater Res 1993;27:35—45. [20] TenHuisen KS, Brown PW. Phase evolution during the formation of a-Ca (PO ) . J Am Ceram Soc (accepted). 3 42 [21] Taylor HFW. Cement chemistry. San Diego, CA: Academic Press, 1991. [22] TenHuisen KS, Brown PW. Variations in the solution chemistry during calcium-deficient and stoichiometric hydroxyapatite formation from CaHPO ) 2H O and Ca (PO ) O. J Biomed 4 2 4 42 Mat Res 1997;36:233—41.
2217
[23] Chow LC, Brown WE. Singular points in the chemistry of teeth. J Dent Res 1975;54(abstract number 120):74. [24] Fowler BO, Kuroda S. Changes in heated and in laser-irradiated human tooth enamel and their probable effects on solubility. Calcif Tissue Int 1986;38:197—208. [25] Zawacki SJ, Heughebaert JC, Nancollas GH. The growth of nonstoichiometric apatite from aqueous solution at 37°C. II. Effects of pH upon the precipitated phase. J Colloid Interface Sci 1990;135:33—44. [26] Holt C, Van Klemenade MJJM, Harries JE, Nelson LS, Biailey RT, Hulkins DWL, Hasnain SS, DeBruyn PL. Preparation of amorphous calcium-magnesium phosphates at pH 7 and characterization by X-ray absorption and Fourier transform infrared spectroscopy. J Crystal Growth 1988;92:239—52.
Formation of calcium-deficient hydroxyapatite from a-tricalcium phosphate Kevor S. TenHuisen1, Paul W. Brown* The Pennsylvania State University, Materials Research Laboratory, University Park, PA 16802, USA Received 10 February 1998; accepted 5 May 1998
Abstract This study investigated the factors influencing the kinetics of Ca HPO (PO ) OH (calcium deficient hydroxyapatite or CDHAp) 9 4 45 formation from a-Ca (PO ) (a-TCP). The kinetics of CDHAp formation were investigated by isothermal calorimetry at constant 3 42 temperatures ranging between 30 and 75°C and by changes in pH at 37.4 and 70°C. The calorimetric curves were characterized by two reaction peaks. Activation energies were calculated for the events resulting in these peaks. Values obtained were 48.4 and 67.7 kJ mol~1, respectively, indicating nucleation and growth mechanisms for both events. Temperature had a significant effect on the growth rate as indicated by a decrease in surface area (26.5—15.0 m2 g~1) of the CDHAp with increasing temperature (30—75°C). A linear relationship between hydrolysis temperature and CDHAp surface area was observed. The morphology of the CDHAp was plate-like and the crystallites became more regular as the reaction temperature was increased. A rapid elevation in pH upon mixing with water indicated the synthesis method initially used did not entirely eliminate slight compositional variations within the a-Ca (PO ) . Rapid elevation in pH retarded subsequent reaction. This effect was eliminated by increasing the duration of high3 42 temperature firing during a-TCP synthesizing. ( 1998 Elsevier Science Ltd. All rights reserved
1. Introduction The development of calcium phosphate cements (CPC) as a class of viable biomaterials has increased interest in their use as artificial hard tissues. One of the earliest CPCs investigated was based on the hydration of a-tricalcium phosphate (a-Ca (PO ) or a-TCP) [1]. a-TCP 3 42 was shown to hydrate to a calcium-deficient hydroxyapatite (CDHAp) in dilute aqueous solutions and at elevated temperatures (60—100°C) [2]. This calcium phosphate is unique in that it can hydrolyze by itself to form a cement. The rate of a-TCP hydration has been observed to increase with temperature and decrease with increasing pH. The composition of the apatite formed could be varied by adjusting the initial pH; as the pH was increased, the Ca/P of the apatite increased from 1.50 to 1.67 [3, 4]. Later it was shown that moderate a-TCP hydration rates could be obtained at near-physiologic * Corresponding author. 1 Johnson and Johnson Corporate Biomaterials Center, Somerville, NJ, USA.
temperatures. Hydration rates were significantly increased when monovalent cationic salts of sulfates, chlorides, nitrates, and a variety of organic species were used [5]. Conversely, analogous salts containing divalent cations acted as retarders. The extents of conversions varied between 0 and '90% when hydration was carried out for 3 h at 40°C. Recently, the setting behavior and property development of a variety of CPCs including those formulated with a-TCP and a number of other calcium phosphate salts and/or CaO or CaCO as the precursor(s) have 3 been investigated [6—10]. Depending on the proportions and reactivities of the precursors, a variety of products were formed. These included octacalcium phosphate (OCP, Ca (HPO ) (PO ) ) 5H O), dicalcium phosphate 8 42 44 2 dihydrate (DCPD, CaHPO ) 2H O), and hydroxyapa4 2 tites of variable composition (Ca (HPO ) (PO ) 10~x 4x 4 6~x (OH) ). Unfortunately, these studies did not report the 2~x extents of conversion for each formulation and only reported the setting times and/or pH. Although setting times provide information about the initial reaction rates, they do not indicate the extents or rates of reaction of these cements.
0142-9612/98/$—See front matter ( 1998 Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 1 3 1 - 8
2210
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
It has been shown in comparative studies [11—14] that the rate of implant resorption in vivo is a function of composition with HAp'b-TCP'a-TCP. Although this trend follows from the ‘relative’ solubilities of these phases [15], bulk density, pore size distribution, and surface area all play critical roles in the rates of resorption. Differences in these properties appear to explain the variability observed [16] in the resorption rates in vivo of preformed HAp (Ca/P"1.45) synthesized from a-TCP and DCPD. These implants were observed to bond to bone and resorb over time. New bone formed between the implant and host bone with no indication of an inflammatory reaction. Conversely, others [17] did not observe resorption of implants consisting of partially hydrated a-TCP blocks placed subperiosteally on rabbit mandibles. However, new bone did form between the native tissue and implant. In vitro studies have been performed on the dissolution behavior of calcium orthophosphates including that of a-TCP in a calcium- and phosphate-free Tris buffer solution [18]. Among the salts investigated, it was found that the dissolution rates of both tetracalcium phosphate (TetCP, Ca (PO ) O) and a-TCP were the highest. Pre4 42 cipitation studies [19] on these calcium orthophosphates in a simulated physiologic solution imitating plasma indicated the formation of amorphous calcium phosphate (ACP) on the surfaces of the a-TCP. Based on FTIR spectroscopy, the a-TCP disappeared within 1 week. In spite of these citations, an in-depth investigation of the kinetics of the hydration of a-TCP to CDHAp has not been performed. The present study investigates relationships between the rates of hydrolysis, the reaction temperature, and microstructural development. Establishing these relationships is required to determine the reaction mechanism(s) and rate limiting steps and is essential in tailoring the hydration of a-TCP for in vivo applications.
2. Materials and methods 2.1. Precursor synthesis a-TCP was synthesized by firing a mixture of calcium carbonate and ammonium hydrogen phosphate at 1150°C for 2 h (a-TCP1). In one instance, synthesis was performed by firing at 1300°C for 24 h (a-TCP2). The mechanistic path for this reaction has been discussed in an earlier publication [20]. The fired product was identified as a-TCP by X-ray diffraction. This solid was milled to a mean particle size of 4.7 lm with a standard deviation of 2.8 lm, as determined by light scattering analysis. The surface area of the a-TCP was determined to be 1.3 m2 g~1 as measured by single-point nitrogen adsorption after outgassing at 150°C for 45 min.
2.2. Calorimetry Isothermal calorimetric analyses of the rates of hydrolysis of a-TCP were performed at various temperatures. All experiments were performed at a liquid-to-solids ratio of 1 : 1 (v/w). In a typical experiment, 3000 g of precursor powder was weighed into a Teflon-coated copper calorimeter cup. The cup was sealed with ParafilmTM to minimize the endothermic effect of water evaporation and then placed into the body of the calorimeter. Deionized water (3.0 cm3) was drawn into a syringe and placed into a copper manifold. The body of the calorimeter and the copper manifold were connected to a constant temperature water bath with a temperature variance of $0.01°C. Upon thermal equilibrium, the water was injected into the calorimeter cup. Thermopiles surrounding the copper cup detected small temperature gradients between the hydrolysis occurring in the calorimeter cup and the body of the calorimeter. This temperature gradient resulted in a voltage output. The voltage output was used to calculate the rate of heat evolution (mW) using a thermoelectric calibration constant. Data were collected by a microprocessor interfaced to the calorimeter through an analog to digital converter. A datum point was collected every 5—15 s depending on the temperature of the reaction. Rates of reaction and cumulative heat were both normalized to moles of CDHAp formed. The activation energy of the hydrolysis of a-TCP to CDHAp was determined from calorimetry experiments performed at 30.0, 37.4, 45.0, 55.0, 65.0, and 75.0°C. 2.3. Solution chemistry analyses Calcium and phosphate concentrations were monitored along with pH during the hydration of a-TCP to CDHAp. A 400 ml jacketed reaction vessel connected to a constant temperature water bath adjusted to 37.4°C ($0.1°C) was filled with 375 ml of deionized water. Upon thermal equilibration, 37.5 g of a-TCP were added to the deionized water. The powder was stirred continuously at 200 rpm with a Teflon-coated stir bar. Nitrogen gas was bubbled through the suspension to minimize the formation of carbonates. The pH was measured every 20 s with a combination glass electrode calibrated using pH 4.00 and 10.00 buffers. The pH meter was connected to a laptop microprocessor for data collection. At selected times &10 ml aliquots of slurry were removed and centrifuged for &30 s to separate the solids. The supernatant was filtered through 0.22 lm PTFE syringe filters and analyzed for calcium and phosphate by DC plasma emission spectroscopy. Solids were frozen in liquid nitrogen to arrest further reaction and freeze-dried. The pH was also monitored during the hydration of a-TCP at 75.0°C. The experimental setup was the same as described above. A pH datum point was collected
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2211
every 5 s due to the more rapid reaction. Aliquots of the suspension were extracted at selected times, filtered, and rinsed in acetone. The solids were then dispersed in acetone, filtered, and dried under vacuum over P O . 2 5 2.4. Reaction product characterization All samples were characterized by X-ray powder diffraction (Scintag PAD V) utilizing copper k radiation. a Samples were scanned over a 2h range of 3 to 40° at a scan rate of 2° 2h per min. Single-point nitrogen adsorption (Quantachrome Monosorb MS-12) was performed on CDHAp formed from the hydration of pure a-TCP after drying overnight at 150°C and outgassing for 45 min at 150°C. Selected samples were lightly coated with gold and observed by scanning electron microscopy (ISI DS-130).
3. Results and discussion 3.1. Isothermal calorimetry The effects of temperature on the kinetics of a-TCP hydrolysis in deionized water are shown in Figs. 1 and 2. Times to complete reaction increase from about 2 h to about 2 days as the hydration temperature is decreased from 75.0 to 30.0°C. X-ray diffraction analysis indicated all reactions approached completion in that only CDHAp was observed. The hydration reaction proceeds in two steps. The first step is a period of rapid hydrolysis resulting in a
Fig. 2. Rate curves for the hydration of a-TCP to CDHAp at various temperatures. The scales of the ordinates (W mol~1) of the inset figure and the main figure are the same. The first reaction peak is apparent in the inset.
calorimetric peak which shows a maximum between 5 and 30 min depending on the temperature (see inset to Fig. 2). Sufficient heat is evolved during this event to induce the initial set. It has been shown [5] that 7—8% conversion of a-TCP will result in setting. This corresponds to 8—10 kJ mol~1 of heat evolution indicating hydrolysis at 37.4°C should result in set within the first hour of reaction. After this initial reaction peak, there is a period of slow heat evolution for hydrolysis temperatures below 55.0°C. This initial peak is followed by the occurrence of a second calorimetric peak. The majority of the heat of reaction is evolved during this second reaction step. The second thermal event resembles the behavior observed in the hydration of silicate cements [21]. Activation energies (E ) for the two reaction steps ! during a-TCP hydrolysis were determined. An Arrhenius relationship between the rate of hydrolysis and temperature was found (see Fig. 3). The slopes of these curves are equal to !E /R. The E for the first reaction ! ! peak was determined to be 48.4 kJ mol~1. A value of 67.7 kJ mol~1 was obtained for the second reaction peak. Values of this magnitude indicate that both thermal events are controlled by a nucleation and growth mechanism. 3.2. Surface area
Fig. 1. Heat evolution curves for the hydration of a-TCP to CDHAp at various temperatures. The reaction temperature in °C is shown for each curve.
Effects of temperature on the surface area of CDHAp formed by a-TCP hydration are shown in Fig. 4. A linear relationship is observed. The surface area of the CDHAp
2212
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
reacted at temperatures above 35°C to form CDHAp or stoichiometric HAp (Ca/P"1.67) [22] although apatites having much higher surface areas were produced. The differences in HAp surface area resulting from reaction between DCPD and TetCP and that of a-TCP having a similar particle size is related to the rates of HAp formation at equivalent temperatures. The DCPDTetCP reactions occur more rapidly, thereby, favoring a higher nucleation density and thus a smaller average HAp crystallite size. 3.3. Microstructure
Fig. 3. Arrhenius plots determining the apparent activation energy (E ) ! of the first and second reaction peaks of a-TCP hydration.
Figure 5 shows the microstructure of the a-TCP and those of CDHAp formed by its hydrolysis at various temperatures. As discussed, the a-TCP (Fig. 5a) had a mean particle size of 4.7 lm and a surface area 1.3 m2 g~1 with a microstructure typical of a milled solid-state product. Figure 5b shows the morphology of CDHAp formed at 37.4°C to be characterized by thin irregularly shaped plates. Little evidence of the original a-TCP morphology is evident other than an occasional spherical particle composed of many small crystallites. The microstructure is characterized by interpenetrating plates showing a high degree of interconnectivity. There appears to be a high degree of intergrowth between the initial reactant particles even though the water was added by inoculation, not by mechanically mixing. The effect of reaction temperature on CDHAp microstructure can be seen by comparing Fig. 5b—d. CDHAp formed at higher temperatures exhibits a more regular microstructure. Similarly, increases in microstructural regularity have previously been observed in CDHAp formed from DCPD and TetCP [22]. Three microstructural features are apparent after reaction at 75°C (Fig. 5c and d): elongated hexagonal plates, rectangular parallelogram shaped plates, and needles. There is again a high degree of interconnectivity between the interpenetrating plates and little evidence of the initial a-TCP particles. These plates exhibit a morphology reminiscent of OCP, but it will be demonstated that these reactions did not form OCP as an intermediate. 3.4. Solution chemistry
Fig. 4. CDHAp surface areas resulting from the hydrolysis of a-TCP as a function of the reaction temperature. The measurements were carried out on the products from isothermal calorimetry.
formed decreases from 26.5 to 15 m2 g~1 as the temperature is increased from 30.0 to 75.0°C. The decrease in surface area with increasing temperature indicates that the ratio of nucleation to growth rates decreases with increasing temperature, although the overall reaction rate increases significantly with increasing temperature. A similar trend was observed when DCPD and TetCP
Figure 6 shows the results of solution chemistry experiments performed during a-TCP hydrolysis. The curve labeled a-TCP1 corresponds to the variations in pH when a-TCP formed at 1150°C (2 h) was hydrated at 37.4°C. Companion variations in the calcium and phosphate concentrations are also shown. Initially the pH rises to &10.8 and then slowly decreases over time. The pH would not be expected to rise to this value as a consequence of the congruent dissolution of compositionally homogeneous TCP. During the first hour of reaction the calcium concentration increased to &1.1 mM and then
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2213
Fig. 5. Scanning electron micrographs of a-TCP and CDHAp obtained from isothermal calorimetry experiments: (a) a-TCP precursor powder; (b) CDHAp from the hydration of a-TCP at 37.4°C; (c) and (d) CDHAp from the hydration of a-TCP at 75.0°C. The magnification in (a—c) is 5 k]; (d) is 10 k].
decreased to 0.075 mM. This change in calcium concentration correlates with the first calorimetric reaction peak as discussed previously. The phosphate concentration remains nominally constant at values around 0.01 mM during this period. Two hypotheses can explain these observations. a-TCP dissolution could be occurring incongruently, or there could be slight compositional inhomogeneities in the a-TCP as a result of the synthesis conditions. Compositional variations would result in calcium-rich regions likely to hydrolyze more rapidly. To establish if such variations were likely, the variations in pH during the hydrolysis of a-TCP formed at 1300°C for 24 h (a-TCP2) were determined. The resulting pH curve, labeled a-TCP2, is shown in Fig. 6. The pH reached a maximum of &9.2 before decreasing to a value of &8.75 where it remained until &5 h. This comparison indicates slight compositional differences exist in the aTCP fired at 1150°C for 2 h. While these data do not
preclude incongruent a-TCP dissolution, dissolution studies are difficult to perform on compounds lacking true solubilities, although solubility products have been reported for both TetCP [23] and a-TCP [24]. The variations of the calcium and phosphate concentrations with pH observed in this study compare favorably to the calculated solubility isotherm of a-TCP shown by Chow [15]. In contrast to reactions between DCPD and TetCP to form HAp [22], where the calcium concentration is higher than the phosphate during the formation reaction, the Ca/P molar ratio in solution is very close to 1 during the hydration of a-TCP. This suggests the apatite initially forming has a Ca/P'1.50. HAp does not have a fixed composition and it has been shown that its composition varies with the solution pH [25]. A Ca/P ratio greater than 1.5 would be expected to form at pH values observed during the first several hours of hydration.
2214
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
Fig. 6. Calcium (Ca, open circles) and phosphate (P, filled circles) concentrations and pH as a function of a-TCP hydration time. a-TCP1 was synthesized by firing at 1150°C for 2 h. a-TCP2 was synthesized by firing at 1300°C for 24 h.
X-ray diffraction analyses of solids extracted as a function of hydration time at 37.4°C (see Fig. 7) confirm the bulk of the hydration occurs between &8 and 20 h as indicated by isothermal calorimetry. The pH decreases from &7.7 to &7.0 during this period. It has been previously reported that a significant increase in the hydration rate of a-TCP occurs at pH values below 7.0—7.5 [5]. As the pH decreases, the calcium and phosphate concentrations increase as expected from prior solubility studies and thermodynamic calculations [15]. Complete reaction, as indicated by X-ray diffraction analyses, occurs at &21 h although most of the a-TCP has been consumed by 18 h (see Fig. 7). Upon complete a-TCP consumption, the calcium and phosphate concentrations increased as the pH continued to decrease until the experiment was terminated (4 weeks). Similar trends were observed when CDHAp formed from DCPD and TetCP. The Ca/P ratio in solution continued to decrease until it reached a value of 0.42 at 4 weeks. This Ca/P value would be expected at a pH of 4.95 because the prominent ionic species in solution are H PO~ and 2 4 Ca2`. These solution conditions result from the incongruent solubility of CDHAp [22]. The effects of the initial pH value on kinetics can be seen in the curves labeled a-TCP1 and a-TCP2 in Fig. 6. Complete a-TCP consumption occurs shortly after the final decrease in pH in every experiment. A reduction in the initial pH rise by reducing compositional fluctuations in the a-TCP reduces the time to complete the reaction by &2.5 h. A pH curve for a-TCP1 hydration at 75.0°C is also shown in Fig. 6. This experiment was performed to determine if OCP forms as an intermediate. As previously discussed, the morphology of CDHAp formed
Fig. 7. X-ray diffraction patterns as a function of a-TCP hydration time at 37.4°C. The times of hydration are labeled.
at 75.0°C is reminiscent of OCP. As previously observed [22] increasing the temperature reduces the pH at any given stage of reaction and increases the rate of reaction involving the formation of HAp. X-ray diffraction analyses performed on samples extracted at 10 min intervals indicated the absence of OCP and complete reaction to CDHAp occurring at 1.5 h. 3.5. Microstructural evolution Figure 8 shows microstructural evolution as a function of a-TCP hydration time at 37.4°C. These solids were extracted in the determinations of calcium and phosphate concentrations and pH for the curves labeled aTCP1 in Fig. 6. Microstructural evidence of a product phase are evident as early as 1.5 h (see Fig. 8a). Small needle-like crystals are occasionally observed on some of the a-TCP particles. At 5 h, small needles and platelike crystallites were present on the surfaces of all of the a-TCP particles (see Fig. 8b). However, evidence of HAp could not be observed by X-ray diffraction until about 9—10 h (see Fig. 7). By this time the majority of the a-TCP
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
2215
Fig. 8. Microstructural evolution during a-TCP hydrolysis to CDHAp at 37.4°C. Times are (a) 1.5 h (20 k]); (b) 5 h (15 k]); (c) 10 h (10 k]); (d) 13 h (10 k]); (e) 21 h (10 k]). All bars on the micrographs equal 1 lm.
surfaces are covered with submicron needles and plate (see Fig. 8c). The microstructure at 13 h (see Fig. 8d) suggests complete hydration has occurred. Although the original
a-TCP surface cannot be observed, X-ray diffraction indicates a significant amount of a-TCP is still present (see Fig. 7). Isothermal calorimetry indicates less than 65% of the total heat produced has been evolved. The
2216
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217
predominant morphological feature is submicron plates growing perpendicular to the a-TCP surfaces. Even though the a-TCP particles appear to be isolated from the solution by apatite, the perpendicular orientation of these plate-like HAp crystallites are conducive to allow ion transport and do not appear to retard the rate of hydrolysis. Figure 8e shows the CDHAp microstructure at complete reaction. The CDHAp morphology is characterized by irregularly shaped hexagonal plates of larger size than those observed at 13 h. The formation of ‘amorphous calcium phosphate’ (ACP) intermediate was observed when CDHAp formed at 37.4°C by reaction of DCPD and TetCP [22]. Unlike reactions involving DCPD and TetCP, the hydration of a-TCP does not appear to form significant amounts of ACP as indicated by X-ray diffraction. In addition, CDHAp formed by a-TCP hydration exhibits a lower surface area and higher crystallinity than that formed from DCPD and TetCP. Although higher pH values favor the formation ACP [26], lower calcium and phosphate concentrations in solution reduce the likelihood of ACP formation. 4. Conclusions Hydrolysis of a-TCP to CDHAp occurs by a mechanism involving two steps. The first step occurred within the first hour and resulted in high initial pH values commensurate with a spike in the calcium concentration. This initial step was a result of slight compositional inhomogeneities attributable to synthesis conditions initially employed. The elevated pH also resulted in a reduction of the rate of hydrolysis. The second reaction step resulted in the bulk of hydrolysis. The activation energies for these steps were determined; they indicate a nucleation and growth mechanism. The microstructure of the resulting CDHAp differed greatly from that produced by the DCPD-TetCP reaction at equivalent temperatures. In particular, crystallites are much larger, resulting in lower surface areas, and the hydroxyapatite formed is more crystalline that that produced by reaction of DCPD with TetCP at equivalent temperatures. This difference is attributable to the slower rate of TCP hydrolysis combined with the absence of the formation of ACP as an intermediate. The CDHAp crystallites grew perpendicular to the a-TCP surface, thus minimizing the formation of a diffusion barrier. Understanding the factors influencing the hydrolysis of calcium phosphate salts to apatites by cement reactions is essential in controlling the setting times and times to complete reaction. The properties must be understood and controlled if these types of materials are to be used in vivo applications. The crystallite size, morphology, and the crystallinity of CDHAp from a-TCP hydration should change its resorption characteristics when compared with HAp formed by other techniques.
Acknowledgements The authors gratefully acknowledge the support of NIDR Grant DE09421.
References [1] Monma H, Ueno S, Kanazawa T. Properties of hydroxyapatite prepared by the hydrolysis of tricalcium phosphate. J Chem Tech Biotechnol 1981;31:15—24. [2] Monma H, Kanazawa T. The hydration of a-tricalcium phosphate. Yogyo-Kyokai-Shi 1976;84:209—13. [3] Monma H, Ueno S, Tsutsumi M, Kanazawa T. Changes in composition and structure of tricalcium phosphate under different pH conditions. Yogyo-Kyokai-Shi 1978;86:590—7. [4] Monma H, Kanazawa T. Wet-process formation of nonstoichiometric hydroxyapatite from tricalcium phosphate. YogyoKyokai-Shi 1978;86:73—6. [5] Monma H, Goto M, Kohmura T. Effect of additives on hydration and hardening of tricalcium phosphate. Gypsum and Lime 1984;188:11—6. [6] Driessens FCM, Boltong MG, Bermudez O, Planell JA. Formulation and setting times of some calcium orthophosphate cements: a pilot study. J Mater Sci: Mater Med 1993;4:503—8. [7] Driessens FCM, Boltong MG, Bermudez O, Planell JA, Ginebra MP, Fernandez E. Effective formulations for the preparation of calcium phosphate bone cements. J Mater Sci: Mater Med 1994;5: 164—70. [8] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Compressive strength and diametrial tensile strength of some calciumorthophosphate cements: a pilot study. J Mater Sci: Mater Med 1993;4:389—93. [9] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Development of some calcium phosphate cements from combinations of a-TCP, MCPM, and CaO. J Mater Sci: Mater Med 1994a;5: 160—3. [10] Bermudez O, Boltong MG, Driessens FCM, Planell JA. Optimization of a calcium orthophosphate cement formulation occurring in the combination of monocalcium phosphate monohydrate with calcium oxide. J Mater Sci: Mater Med 1994b;5:67—71. [11] de Groot K. Clinical usefulness of calcium phosphate ceramics. Zahnarztl Mitteil 1985;75:1938—40. [12] Klein CP, Patka P, den Hollander W. Macroporous calcium phosphate bioceramics in dog femora: a histological study of interface and biodegradation. Biomaterials 1989;10:59—62. [13] Krause A, Wehrbein H, Jacobs HG, Morich S. An alpha-modified tricalcium phosphate-ceramic for filling large bone defects after cystectomy in the anterior maxillary region. A report of clinical experience. Dtsch Z Mund- Kiefer- Gesichts Chir 1986;10:108—12. [14] Renooij W, Hoogendoorn HA, Visser WJ, Lentferink RHF, Schmitz MG, Van Ieperen H, Oldenburg SJ, Janssen WM, Akkermans LMA, Wittebol P. Bioresorption of ceramic strontium85-labeled calcium phosphate implants in dog femora. Clin Orthoped Relat Res 1985;197:272—85. [15] Chow LC. Development of self-setting calcium phosphate cements. J Ceram Soc Japan 1991;99:954—64. [16] Kurashina K, Ogiso A, Kotani A, Takeuchi H, Hirano M. Histological and microradiographic evaluation of hydrated and hardened a-tricalcium phosphate/calcium phosphate dibasic mixtures. Biomaterials 1994;15:429—32. [17] Nagase M, Chen R-B, Asada, Y, Nakajima T. Radiographic and microscopic evaluation of subperiosteally implanted blocks of hydrated and hardened a-tricalcium phosphate in rabbits. J Oral Maxillofac Surg 1989;47:582—6.
K.S. TenHuisen, P.W. Brown / Biomaterials 19 (1998) 2209—2217 [18] Ducheyne P, Radin S, King L. The effect of calcium phosphate ceramic composition and structure on in vitro behavior. I. Dissolution. J Biomed Mater Res 1993;27:25—34. [19] Radin S, Ducheyne P. The effect of calcium phosphate ceramic composition and structure on in vitro behavior. II. Precipitation. J Biomed Mater Res 1993;27:35—45. [20] TenHuisen KS, Brown PW. Phase evolution during the formation of a-Ca (PO ) . J Am Ceram Soc (accepted). 3 42 [21] Taylor HFW. Cement chemistry. San Diego, CA: Academic Press, 1991. [22] TenHuisen KS, Brown PW. Variations in the solution chemistry during calcium-deficient and stoichiometric hydroxyapatite formation from CaHPO ) 2H O and Ca (PO ) O. J Biomed 4 2 4 42 Mat Res 1997;36:233—41.
2217
[23] Chow LC, Brown WE. Singular points in the chemistry of teeth. J Dent Res 1975;54(abstract number 120):74. [24] Fowler BO, Kuroda S. Changes in heated and in laser-irradiated human tooth enamel and their probable effects on solubility. Calcif Tissue Int 1986;38:197—208. [25] Zawacki SJ, Heughebaert JC, Nancollas GH. The growth of nonstoichiometric apatite from aqueous solution at 37°C. II. Effects of pH upon the precipitated phase. J Colloid Interface Sci 1990;135:33—44. [26] Holt C, Van Klemenade MJJM, Harries JE, Nelson LS, Biailey RT, Hulkins DWL, Hasnain SS, DeBruyn PL. Preparation of amorphous calcium-magnesium phosphates at pH 7 and characterization by X-ray absorption and Fourier transform infrared spectroscopy. J Crystal Growth 1988;92:239—52.