Calcium orthophosphates are subject to intensive investigations owing to their biological importance. The ion-substituted non-stoichiometric nano-sized poorly crystalline calcium orthophosphates, mainly with apatite structure, build the inorganic component of hard tissues in the organisms. The main ion substitutes are the ions Na+, K+, Mg2+, Fe2+, Zn2+, Si2+, CO3 2-, Cl-, and F- (Dorozhkin, 2009; Daculsi et al., 1997) and they differ in variety and amount depending on the type of the hard tissue, its age as well as on individual peculiarities. The so called “biological apatite” is formed in the living organisms as a result of biomineralization processes, the mechanism of which is not yet clarified. These processes include precipitation, dissolution and growth of poorly-crystalline calcium orthophosphates taking place in the organic matrix, e.g., collagen in the case of bones (Dorozhkin, 2009; Palmer et al., 2008) or amelogenin in the case of enamel (Palmer et al., 2008), in the presence of body fluids. One of the ways to elucidate the elementary processes occurring during bone hard tissue mineralization is the biomimetic approach designed to study these processes. The knowledge of the elementary processes is crucial for the development of new bioactive calcium phosphate materials (close to the natural ones) that may be applied for bone repairing, reconstruction and remodeling.
The aim of this chapter is to throw light on the biomimetic precipitation and modification of calcium orthophosphates, XRD-amorphous calcium phosphate (ACP) and dicalcium phosphate dihydrate (DCPD) on the basis of authors’ kinetic, spectral (XRD and IR) and thermodynamic studies and literature data.
2. Calcium orthophosphates – short review
Eleven calcium orthophosphates are known in the literature. According to the methods of their preparation they are divided into two groups - calcium phosphates precipitates and calcium phosphates calcinates (Table 1). The preparation of calcium phosphates precipitates strongly depends on pH of the medium; that of calcium phosphates calcinates is a function of the calcination temperature.
Calcium phosphates are divided into three groups according to their structure (Chow & Eanes, 2001): (i) Ca-PO4 sheet-containing compounds (MCPA, MCPM, DCPA DCPD). DCPD has a monoclinic structure, space group Ia, where HPO4 2- ions are linked to Ca2+ ions forming linear chains, that are stacked and form corrugated sheets parallel to the (010) face. The water molecules are situated between the sheets, bonded to the Ca2+ ion. The packing of the Ca-HPO4 ions in chains or sheets determine several possible pseudohexagonal arrangements, similar to the glaserite type structure (Curry & Jones, 1971; Dickens et al., 1972; Dickens & Bowen, 1971); (ii) glaserite type compounds (α-TCP and β-TCP). Two types of columns along the c-axis in a pseudohexagonal arrangement, one containing only Ca2+ and other both Ca2+ and PO4 3- ions in a ratio 1:2 build the glaserite type structure of monoclinic α-TCP (Mathew et al., 1977). In rhombohedral β-TCP structure two types of columns contain both Ca2+ and PO4 3- ions (Dickens et al., 1974). One of the columns has vacancies at both cationic and anionic position; and (iii) apatite type compounds (OCP, TTCP, PCA and HA) (Chow & Eanes 2001; Mathai&Takagi, 2001). Commonly, HA has a hexagonal structure (space group P63/m) (Kay et al., 1964), where Ca2+ ions occupy two different crystallographic symmetry sites. Ca1 are located in columns along the c-axis, where is coordinated to nine O atoms. The Ca-O9 polyhedra are connected in chains parallel to c-axis. Ca2 are arranged in two triangular units. The Ca2 ions are 7-coordinated, with six O atoms and one OH- ion. Ca1 and Ca2 polyhedra are linked through oxygen atoms of the PO4 3- tetrahedra. Each OH- ion occupies statistically disordered positions.
OCP has a triclinic structure, which can be described as alternating along (100) “hydrated” and apatitic layers (Mathew et al., 1988). The atomic positions of the structure of OCP are very close to HA structure, which is the precondition for possible epitaxial growth and formation of interlayered structures, important for explanation of the process of biomineralization.
A special position holds the amorphous calcium phosphate (ACP) which structure is built of Ca9(PO4)6, so called Posner’s clusters, where Ca2+ and PO4 3- ions are arranged in a hexagonal dense packing (Betts et al., 1975; Blumenthal et al., 1977).
The existing symmetry relations between these structures ensure the easier phase transformations.
Calcium orthophosphates are sparingly soluble in water (Table 2). HA has the lowest solubility among them, which is its natural priority. The solubility of calcium phosphates strongly depends on pH of the medium and this feature is of significance for their preparation and biological behavior. Thus, the practically insoluble mono-phase bio-ceramics of dense HA do not actively participate in the process of bone remodeling (Tas, 2004). However, upon contact with body fluids they participate in the formation of a surface layer of bone-like apatite. Mono-phase α-TCP and β-TCP display higher solubilities and rapidly degrade in vitro and in vivo (Radin & Ducheyne, 1993, 1994). Mg- and Zn-doped TCP ceramics display lower solubility than pure TCP ceramics and thus reduce the resorption rate (Xue, 2008). Bi-phase mixtures of HA and β-TCP ceramics were developed in order to improve the biological behaviour of the mono-phase materials (Petrov et al., 2001; Teixeira et al., 2006).
The knowledge on the Ca2+, H+/ OH-, PO4 3-//H2O system and its sub-systems may be used as a theoretical base for predetermination or optimization of the conditions for the preparation of different calcium orthophosphates. Unfortunately, owing to the low solubility and narrow crystallization fields of the different stable and metastable salts, there are no systematic experimental studies of this system. Only single solubility data are available for the binary sub-system Ca2+/PO4 3-//H2O at 25ºC (Kirgintzev et al., 1972). More detailed studies were performed on the three-component Ca2+, H+/ PO4 3-//H2O sub-system and experimental data are available for the temperature range 0 – 100ºC (Flatt et al., 1961; Bassett, 1958; Flatt et al., 1956; Chepelevskii et al., 1955; Belopol’skii, 1940; Bassett, 1917). Two hydrous and two anhydrous salts, namely Ca(H2PO4)2, Ca(H2PO4)2.H2O, CaHPO4 and CaHPO4.2H2O are established at 25ºC and 40ºC respectively; there are contradictions about the existence and stability of the salt of lowest solubility CaHPO4.2H2O (Bassett, 1917; Belopolskii et al., 1940; Chepelevskii et al., 1955). The solubility of Ca(H2PO4)2 and Ca(H2PO4)2.H2O slightly increases at temperatures above 50ºC but CaHPO4.2H2O was not detected (Bassett, 1917; Chepelevskii et al., 1955).
The most appropriate method for evaluation of the solubility of sparingly soluble calcium phosphate salts is the thermodynamic modeling. The ion association model based on the extended Debye-Huckel theory was applied to the Ca2+, H+/ OH-, PO4 3-//H2O system (Chow & Eanes, 2001; Johnnson & Nancollas, 1992). Thermodynamic data for the solubility products (lgKsp 0) of all calcium orthophosphates and the complex formation constants (lgK0) of all complex species which may exist in aqueous calcium phosphate solutions are necessary for its application (Table 2). The calculations of Chow and Eanes (2001) have shown that DCPA is the least soluble salt in the Ca2+, H+/ OH-, PO4 3-//H2O system at pH < 4.2 and 25ºC while HA becomes the least soluble salt at pH > 4.2; TTCP is the most soluble salt at pH <8.2 while DCPD is the most soluble salt at pH > 8.2. In the pH region 7.3 - 7.4 typical for body fluids, the solubility of the salts at 25ºC (Chow & Eanes, 2001) and 37ºC (Johnnson & Nancollas, 1992) follows the order:
TTCP > α-TCP > DCPD > DCPA >OCP ~ β-TCP > HA
3. Electrolyte systems for biomimetic studies
Electrolyte solutions of different composition, designed to mimic the аcellular human body plasma, have become a modern way to test bone-bonding abilities of bioactive materials or to produce thin calcium-phosphate layers on materials (metals, alloys or glasses) for bone graft substitutes (Yang & Ong, 2005; Raghuvir et al., 2006; Jalota et al., 2006; Kontonasaki et al., 2002). The composition of the most popular ones is presented in Table 3.
Earle's balanced salt solution (EBSS, Ca/P = 1.8, HCO3 - - 26.2mmol.dm-3) (Earle et al., 1943) and Hank’s balanced salt solution (HBSS, Ca/P = 1.6, HCO3 - - 4.2mmol.dm-3) (Hanks & Wallace, 1949) were among the first simulated body solutions. Kokubo (1990) was the first to popularize a multicomponent inorganic solution, called conventional simulated body fluid (SBFc) which contains definite amounts of Na+, K+, Mg2+, Ca2+, Cl-, HCO3 2-, HPO4 2- and SO4 2- ions, has a Ca/P ratio of 2.5 (equal to that in the blood plasma), HCO3 - concentration of 4.2 mmol.dm-3 and physiologic pH of 7.3-7.4. To mimic the blood plasma in terms of the most important HCO3 - ions, Bayractar and Tas (1999) revised the SBFc by increasing HCO3 - concentration up to 27 mmol.dm-3 at the account of Cl- ions (revised simulated body fluid, SBFr). The concentrations of Ca2+ and Mg2+ ions in the ionic SBF (SBFi) correspond to those of free Ca2+ and Mg2+ ions (not bound to proteins), in the blood plasma (Oyane, et al., 2003).
These solutions were buffered to the pH of blood plasma with TRIS, BITRIS or HEPES buffers.
SBF modified with glycine (SBFg), essential for the biological system amino acid, was prepared on the basis of conventional SBF. Concentration of glycine was thermodynamically calculated so that the contents of free Ca2+ and Mg2+ ions to be analogous to SBFi.
4. Biomimetic precipitation of ion modified precursors
The biomimetic approach which includes precipitation processes of bioactive calcium phosphates in electrolyte medium of simulated body fluids and uses the influence of the medium composition on their formation and phase transformation have attracted extensive research interest (Xiaobo et al., 2009; Hui et al., 2009; Shibli & Jayalekshmi, 2009; Martin et al., 2009), because of their analogy to the biological mineralization processes. In the following, the authors’ studies on the precipitation of ion-modified ACP and DCPD precursors are summarized.
Various crystal chemical and kinetic factors affect the crystallization process. The ion-modified calcium phosphates are mixed crystals (non-stoichiometric compounds), where part of the ions building the crystal unit cell are substituted by other ions. The ability of the admixture ion to adopt the coordination of the substituted ion determines the substitution degree.
To enable ion modification of calcium phosphate precursors with Na+, K+, Mg2+ and Cl- ions we have performed all our studies using conventional SBFc that was modified for each concrete case. Modified calcium-free simulated body fluid (SBFc-Cam) was used as a solvent for K2HPO4 (Solution 1) and phosphorus-free simulated body fluid (SBFc-Pm) was used as a solvent for CaCl2 (Solutions 2 and 5), for CaCl2 and MgCl2 (Solution 3) and for ZnCl2 (Solution 4) (Table 4). In this way preliminary precipitation was avoided. pH of the mixed solutions was adjusted to 7.2-7.4 using 0.1M HCl or 0.05M 2-amino-2-hydroxymethil-1,3-propandiol.
The electrolyte medium provided by SBF plays a crucial role in the precipitation processes and influences the composition of the precipitated product. Precipitation, co-precipitation, ion substitution and ion incorporation reactions simultaneously take place. The cationic and anionic substitutions are mainly responsible for the calcium deficiency of the precipitated ACP precursors. Two methods – fast mixing or continuous co-precipitation of the reagents were applied in these studies. The method of precipitation affected the size, morphology and chemical homogeneity of the precipitate.
SBF-modified XRD-amorphous calcium-deficient phosphate (ACP) (Fig. 1) with a Ca/P ratio of 1.3 or 1.51 (Table 5) due to ion substitution and incorporation of Na+, K+, Mg2+ and Cl- ions from the SBFs at levels close to those of natural enamel, dentin and bone (Dorozhkin, 2009), was precipitated.
The fast precipitation was carried out by mixing Solution 1 and Solution 2 (Table 4) at a Ca/P ratio of 1.67 and pH of 11.5 (maintained by 1M KOH) under intense stirring at room temperature. It is known that the fast mixing, the high supersaturation and the presence of Mg2+ and CO3 2- ions provoke the precipitation of an amorphous calcium-deficient product (Sinyaev et al., 2001; Combes & Rey, 2010). The continuous co-precipitation was carried out by mixing Solution 1 and Solution 2 (Table 4) at a rate of 3 ml/min to precipitate in glycine buffer (Sykora, 1976) at room temperature and pH 8 (maintained by 1M KOH).
Zn- or Mg-modified amorphous calcium phosphate precursors with varying Me2+/(Ca2++Me2+) ratio from 0.01 to 0.16 (Table 6) due to Ca2+ ion substitution by Me2+ ions as well as Me2+ incorporation were precipitated by the method of continuous co-precipitation in electrolyte system only. All reagents (Solutions 1, 2 and 4 for Zn-modified precursors and Solutions 1 and 3 for Mg-modified precursors, Table 4) with a (Ca2++Me2+)/P ratio of 1.67 (Me2+ = Mg, Zn) were mixed to precipitate in glycine buffer with a rate of 3 ml/min at room temperature and pH 8 (maintained by 1M KOH). The modified conventional simulated body fluids provided ion modification of all Mg- and Zn-modified calcium phosphate precursors with Na+ (0.02 - 0.08 mmol/g), K+ (0.01 – 0.02 mmol/g, Mg 2+ (0.04 mmol/g) and Cl- (below 0.05 mmol/g) ions (Table 6).
By analogy with Bigi et al. (1995), we have established that the presence of Zn2+ or Mg2+ ions in the reaction mixture inhibits the crystallization of HA, so that XRD amorphous Mg- or Zn-modified calcium phosphate precursors are obtained. The Posner’s clusters (Betts et al., 1975; Blumenthal et al., 1977) of the complex formula CawMgxZnyNazKu(PO4)v(CO3)6-v (w+x+y+z+u ≤ 9) are the first particles formed in the studied complex electrolyte SBF – CaCl2 – MgCl2/ZnCl2 – KOH – H2O system. A modifying ion, whose ionic radius and electrical charge are closer to those of the Ca2+ ions, will be more readily incorporated into the Pozner’s clusters. The Zn2+ ionic radius (0.74 Å) is closer to that of the Ca2+ ion (1.0 Å) than the radius of the Mg2+ ion (0.65 Å). The substitution with Na+ (0.95 Å) and K+ (1.33 Å) ions is partial not only for geometrical reasons but also for electrostability. The results (Table 6) showed that all Zn2+ ions and only about half of the Mg2+ ions from the reaction solutions were included in the precipitated ACP. The different chemical behavior of Zn2+ and Mg2+ ions can be explained by the “softness-hardness” factor and by the Crystal Field Stabilization Energy (CFSE). According to Pearson’s concept of “hard” and “soft” Lewis acids and bases (Pearson, 1963), as well as the Klopman scale of hardness and softness (Klopman, 1968), “soft acids” predominantly coordinate “soft bases” and “hard acids” - predominantly “hard bases”. Mg is a “hard acid”, while Zn is a “soft acid”. The simulated body fluids contain high concentrations of Cl- ions which are “softer bases” than H2O, OH-, PO4 3-, SO4 2-, HCO3 - and HPO4 2-. Although Zn2+ ions are a “soft acid”, they form a negligible amount of chloride complexes due to the zero value of their CFSE and mainly exist as free Zn2+ ions in the studied solutions. In contrary, Mg2+ as a “hard acid” is preferentially coordinated by the H2O molecules (“hard base”) and are mainly present as [Mg(H2O)6]2+ complexes. The last ones are too large to be incorporated into the crystal structure of the calcium phosphate without its distortion. The necessity of overcoming the energy barrier for even partial dehydration of the [Mg(H2O)6]2+ complexes is another reason for the low substitution rate of these ions.
DCPD biomimetic precipitated precursors - Well crystallized dicalcium phosphate dihydrate (DCPD) (Fig.2) was precipitated by the method of fast mixing (room temperature and intense stirring) of Solution 1 and Solution 5 (Table 4) at a Ca/P ratio of 1 and pH 6 (maintained by 1M HCl). Differently from all modified ACP precursors, only negligible amounts of Mg2+ (0.001 mmol/g), Na+ (0.025 mmol/g), K+ (0.001 mmol/g) and Cl- (0.003 mmol/g) ions were found in biomimetic precipitated DCPD.
Thermodynamic modeling of biomimetic precipitation - The precipitation processes of SBF-modified ACP and DCPD as well as of Zn- and Mg-modified ACP were simulated by an ion-association model using the computer program PHREEQCI v.2.14.3 (Parkhurst, 1995). All possible association/dissociation and dissolution/ crystallization processes in the SBFs were taken into account. The formation of complexes and the precipitation of salts were considered by means of a mass-action expression using the appropriate formation constants or solubility products. The activity coefficients of all possible simple and complex species were calculated by the extended Debye-Huckel theory using an updated database (Todorov, et al., 2006).
where IAP is an ion activity product and K is a solubility product.
When the solution is supersaturated with respect to a certain salt (SI > 0), it will precipitate; when the solution is undersaturated (SI < 0), the salt will not precipitate; the solution and the salt will be in equilibrium when SI = 0.
Different calcium, magnesium, sodium and potassium salts can simultaneously co-precipitate in electrolyte SBF systems. Their number depends on the precipitation conditions (Table 7).
In the SBF with pH value of 11.5, nine salts display positive SI, namely Mg(OH)2, CaHPO4, Mg3(PO4)2.8H2O, MgCO3.Mg(OH)2.3H2O, CaCO3, Ca3(PO4)2(am), Ca8H2(PO4)6.5H2O, Ca9Mg(HPO4)(PO4)6 and Ca10(PO4)6(OH)2 (Table 7) and can co-precipitate. At pH 8 the same salts including CaHPO4.2H2O but except Mg(OH)2 can co-precipitate. The increase of the Mg2+ ion concentration in the system leads to co-precipitation of extra four metastable magnesium salts and favors the precipitation of Ca9Mg(HPO4)(PO4)6 (SI increases). The increase of the Zn2+ concentration in the system does not influence the co-precipitated salts. The only zinc phosphate salt Zn3(PO4)2.4H2O is not expected to precipitate (SI<0). In SBF of pH 6 where DCPD precipitates, only calcium phosphate salts can co-precipitate. In all cases, the highest SI and the highest thermodynamic stability are displayed by Ca10(PO4)6(OH)2 followed by Ca9Mg(HPO4)(PO4)6. Despite the thermodynamic stability of HA, the kinetic factors favor the formation of metastable phases – ACP at pH 8 and 11.5 and initial (Ca+Me)/P = 1.67 and DCPD at pH 6 and Ca/P = 1. These results are in compliance with Ostwald’s step rule, according to which the crystal phase that nucleates is not the phase that is most thermodynamically stable under these conditions, but rather is a metastable phase closest in free energy to the parent phase (Chung, et al., 2009). The highest crystallization rate and the lowest supersaturation necessary for nucleation should be exhibited by those salts in the saturated solution, for which there is a sufficient concentration of structural entities able to be incorporated unchanged or with small changes into the crystal structure.
5. Biomimetic modifications and phase transformations of ACP and DCPD
With the aim to elucidate the influence of micro-environmental surroundings on the phase transformation process of SBF-modified ACP, DCPD, and Zn-modified ACP we have investigated their biomimetic maturation in SBFs by means of kinetic, spectral and thermodynamic studies. The experiments were performed with three different SBFs –
conventional, SBFc, revised, SBFr, and conventional modified with glycine, SBFg (Table 3). Before maturation the precipitated precursors were filtered, washed with water and with acetone (solid-to-liquid ratio of 1:1) and lyophilized at -56ºC. Then the freeze-dried samples were matured for varying time periods (from 1 h to 6 months) at a solid-to-liquid ratio of 1:250, physiological temperature of 37ºC, in a static regime.
The biomimetic modification of SBF-modified ACP in SBFc, SBFr and SBFg gave rise to changes in the compositions of both solid and liquid phases during the maturation process (Fig. 3) and revealed that dissolution/crystallization processes are strongly influenced by the content of SBFs.
A similar behavior of PO4 3- (Fig. 3a), Ca2+ (Fig. 3b), and Mg2+ (Fig. 3c) ions was found in the three SBFs. The biggest changes were registered during the first 6 hours. The liquid phases were enriched in Ca2+, Mg2+, and PO4 3- ions during the first 2 - 4 h; afterwards, until the 6th hour, the enrichment rate gradually decreased. The highest increase was registered for Ca2+ (32%) and Mg2+ (46%) ions in SBFg. Their highest decrease was observed in SBFr. For Ca2+ ions this decrease was about 90% for 6 hours, while for Mg2+ ions the decrease continued after the 6th hour and reached 83% at the 72nd hour. The presence of glycine in SBFg and the higher content of HCO3 - ions in SBFr leads to formation of metal-glycine and metal-carbonate complexes that enhance the solubility of the salts. During maturation in SBFr which is richer in HCO3 -, crystallization of CaCO3 occurs, also confirmed by the increased Ca/P ratios in the solid phase (Fig.3d), whereas in SBFc and SBFg the formation of calcium phosphate dominates. The increase of the Mg/Ca ratio in the solid phases (Fig.3e) gives an evidence for the incorporation of Mg2+ in the amorphous phase.
The spectral studies (XRD and IR) confirmed the biomimetic phase transformation of amorphous calcium phosphate into the more stable poorly-crystalline apatite in the three SBFs, differing only in the phase transformation rate (Fig 4). Crystal phase was detected at the 4th hour of the maturation process in SBFc; at the 2nd hour in SBFr and at the 1st hour in SBFg (Fig 4a). The increase in the degree of crystallinity during the maturation process was confirmed by the observed splitting of the phosphate bands at 960, 1100, 562, and 603 cm-1, which are characteristic for the IR spectra of crystalline calcium phosphate (Fig 4b).
IR data also revealed a change in the carbonate content of the samples treated in SBFc and SBFr with different carbonate content (Fig. 5). As a measure of the carbonate content we used the ratio between the areas underneath the peaks corresponding to CO3 2- (1549-1336 cm-1) and PO4 3- (1280-914 cm-1) stretching bands. As can be seen, the amount of carbonate ions faster increases in samples matured in SBFr than in those matured in SBFc. Thermodynamic modeling of the maturation process of two different solid calcium phosphate products - a metastable amorphous product (*ACP), and a stable equilibrium product (**ACP) were done in the three solutions (SBFc, SBFr and SBFg) differing in their ionic content (Table 8).
The metastable amorphous product (*ACP) simulated the system behavior during the first 1-2 hours of maturation, when no Ca8H2(PO4)6.5H2O, Ca9Mg(HPO4)(PO4)6 and Ca10(PO4)6(OH)2 phase was yet formed, whereas the stable product (**ACP) simulated the equilibrium system. The maturation of the metastable amorphous product (*ACP) leads to a phase transformation that depends on the content of HCO3 - ions in SBF at the beginning of the process (Table 8). In a solution with a low HCO3 - content (SBFc and SBFg), dissolution phenomena of all magnesium salts occur (SI<0) during maturation and the system will be in equilibrium with the calcium salts (SI = 0), including the amorphous calcium phosphate. The increase in the HCO3 - content (SBFr) leads to dissolution and phase transformation of the amorphous calcium phosphate into thermodynamically more stable salts. The calculated equilibrium amounts of CaCO3 and Ca3(PO4)2(am) in the three investigated body fluids (Fig. 6) point to the significantly favored crystallization of CaCO3 (especially in SBFr) and the decreased amount of Ca3(PO4)2(am) due to dissolution processes. The calculations revealed that there was no influence of SBFs composition on the equilibrium product (**ACP), the system tending to thermodynamic equilibrium by dissolution of all co-precipitated solid phases and re-crystallization of the thermodynamically unstable amorphous calcium phosphate (with SI<0) into pure HA (with SI = 0) (Table 8).
The model cannot predict the formation of Mg-substituted carbonated hydroxyapatite due to the lack of thermodynamic data.
The calculated species distribution in the initial SBFs, in the SBFs at metastable (maturation of *ACP) and at stable (maturation of **ACP) equilibrium (Fig. 7) gives an evidence for the domination of Me2+ (Me = Ca, Mg) free ions in all studied cases. In the initial SBFs (Fg. 7a and 7b) and at a stable thermodynamic equilibrium (Fig. 7e and 7f) Me2+ free ions are dominating, followed by MeCl+ in SBFc, from MeHCO3 - and MeCl+ in SBFr and from CaH(Gly)2+ and CaCl+ and from MgCl+ and Mg(Gly)+ in SBFg. Significant changes in species distribution are observed at a metastable equilibrium (Fig. 7c and 7d), revealing the essential role of SBF ionic composition on the maturation process. The increased amount of MeHPO4 2-, CaPO4 - in SBFc and SBFr, as well as the increased amount of Me(Gly)+ species in SBFg is due to the dissolution of metastable salts while the decreased amount of MeHCO3 - species especially in SBFr is due to the crystallization of CaCO3.
These thermodynamic data explain the results from the maturation kinetics.
The biomimetic modifications of Zn-modified ACP were studied on three exemplary samples with different Zn2+/(Zn2+ + Ca2+) molar ratios (0.03, 0.05 and 0.10), treated in a conventional simulated body fluid (SBFc). It was found that the Zn content decreases by a factor of 2 during the first 2 hours (Fig. 8) when the samples are still amorphous (Fig. 9). Subsequently, the amorphous phase progressively converted into poorly-crystalline apatite. The Zn content influenced the transformation rate. At a higher Zn content the stability of the amorphous phase increased and the rate of the process slowed down (Fig. 9).
The kinetic studies of the biomimetic modifications of DCPD revealed that the compositions of the liquid and solid phases, similar to those of ACP, changed during the maturation process (Fig. 10). In SBFg the highest increase in PO4 3- and Ca2+ concentrations was registered as a result of the effect of glycine which promotes the dissolution of DCPD. In the solid phase the Ca/PO4 ratio is kept ≈1 during the first 6 hours, then gradually increases. The latter is an indication for the beginning of the transformation process and the formation of basic calcium phosphates with Ca/PO4> 1.
IR and XRD data (Fig.11) show that the initial DCPD transforms into poorly-crystalline B-type carbonate apatite via an intermediate phase of OCP. The rate of conversion of DCPD to carbonated apatite differs, depending on the type of SBF. The formation of carbonated apatite is faster in SBFr, as it is indicated by the disappearance of O-H and P-OH peaks typical of DCPD phase and the appearance of absorption bands near 1420 and 1480 cm-1 associated with B-type CO3 2- groups after 240 h of treatment (Rabadjieva et al., 2010). For samples treated in SBFc DCPD phase is still preserved after 72 h of treatment, while the carbonate peaks appeared after 720h (one month) maturation and increased in intensity after 6 months, indicating slow rate of PO4 3− substitution by CO3 2- groups. Intense absorption bands in the 1073-1122 cm-1 spectral range are due to the asymmetric stretching vibration mode of P-O, whereas the absorption band that appears near 963 cm-1 is associated with the symmetric stretching mode of P-O. Peaks at 602, 562 and at 470 cm-1 originate from O-P-O bending modes. O-H stretching bands of crystalline water appeared in the range of 3500-3100 cm-1, while H2O bending is at 1640-1650 cm-1. Peaks at 1297 and 1192 cm-1 are due to the P-OH bending mode of HPO4 2- groups, while those near 914 and 870 cm-1 arise from the stretching mode of HPO4 2- groups and partially overlapped with C-O vibrations. The absorption bands at around 878 cm-1 could be due to bending vibration of CO3 2- groups, because are related with intensive bands in the 1400-1500 cm-1 spectral interval, typical of C-O stretching vibrations. Peak at 526 cm-1 is associated with HO–PO3 bending mode in HPO4 2- (Mandel & Tas, 2010; LeGeros et al., 1989).
6. High-temperature phase transformations of ion modified ACP
The matured Zn- and Mg-modified ACP precursors were treated by a procedure including gelation with Xanthan gum, lyophilization at -56ºC, washing (solid-to-water ratio of 1:100) and secondary lyophilization. Then they were sintered at 600, 800 and 1000ºС in order to study the high-temperature phase transformations and to follow the effect of Mg2+ and Zn2+ on the Ca2+ ion substitution and on the phase composition and morphology of the sintered products. The working regime consisted of heating at a rate of 3ºC/min up to the desired temperature and keeping it constant for 1 hour. Both, concentration of Mg2+ and Zn2+ ions and temperature affect the spectral characteristics of the studied samples (Fig. 12). The hydroxyl stretching peaks near 3570 cm-1 and the hydroxyl librations at 633 cm-1 revealed the presence of HA phase, while the peaks near 1120, 1075, 1040, 975 and 950 cm-1 indicated the formation of β-TCP phase.
The XRD data (Fig. 13) revealed that all amorphous precursors were transformed into ion-modified HA or β-TCP crystalline phases depending on the additive and its amount as well as on the temperature. Both Zn and Mg substitutions promoted the ACP transformation to ion-modified HA and β-TCP but the effect was more pronounced in the case of Zn substitution. Zn-β-TCP and Mg-β-TCP were registered at 600ºC for samples with Me2+/(Me2+ + Ca2+) ratios higher than 0.05 and mixture of the ion-modified HA and β-TCP calcium phosphates at ratios lower than 0.05. The SEM analyses revealed that Zn and Mg substitutions influenced the morphology of the β-TCP grains (Fig 14). Sintering at 800-1000ºC leads to zinc-modified-β-TCP powders of idiomorphic crystals with sizes ranging from 500 to 5000 nm in all studied cases. Magnesium-modified-β-TCP fine powders with spherical grains of smaller size (250 - 1000 nm) were obtained at a Mg2+/(Mg2+ + Ca2+) ratio higher than 0.02 (Figure 13 and 14). The surface of the spherical grains of magnesium modified samples was covered by blanket, while clean grain boundary was observed in the zinc-modified samples.
The calculated unit cell parameters showed that low ion substitution (Me2+/(Me2++Ca2+) ratio up to 0.05) leads to a slight decrease in the parameter a and the volume V and to an increase in the parameter c of the Me-β-TCP unit cell (Table 9). This tendency became more pronounced at higher temperatures.
The inclusion of Mg2+ and Zn2+ ions into the crystal unit cell of the thermodynamically stable HA proceeds through Ca2+ ion substitution. As the ionic radii of Mg2+ (0.65 Å) and
Zn2+ (0.74 Å) are too small in comparison with that of Ca2+ (1.00 Å) the increase in their amount leads to unit cell distortion and volume decrease, established also by Ito et al. (2002) and Miyaji et al. (2005). Thus, the structure of Me2+ ion modified HA is destabilized and
its transformation into the more stable β-TCP structure could be expected. The latter structure includes different CaOn coordination polyhedra (n=3,6,7,8) (Yashima et al., 2003). The vacant sites of the smallest CaO3 polyhedron are the most suitable holes for the inclusion of the small Mg2+ and Zn2+ ions, thus the unit cell distortion and structure destabilization will be negligible. Ion substitution at a Me2+/(Ca2++Me2+) ratio higher than 0.05-0.15 leads to an increase in the Me2+ ion inclusion into the larger CaOn polyhedra (n=6-8), which destabilizes the structure. The appearance of a more stable high-temperature modification, α-TCP, could be expected in this case, but no α-TCP XRD peaks were detected in our experiments.
In a summary, original authors’ studies and literature data are presented on the biomimetic synthesis of XRD-amorphous calcium phosphate and dicalcium phosphate dihydrate and their biomimetic modifications and phase transformations into poorly-crystalline apatite in three types of simulated body fluids - conventional (SBFc), revised (SBFr) and modified with glycine (SBFg). The compositions of the different types of artificial body fluids that are known in the literature are compared in terms of their similarity to blood plasma; their advantages and disadvantages are highlighted. The authors’ studies and original results on chemical and phase compositions, kinetics and thermodynamic simulations are discussed. A new approach based on thermodynamic modeling (using the PHREEQCI v.2.14.3 computer program based on an ion-association model) was applied for simulation and explanation of the biomimetic precipitation of metastable XRD-amorphous calcium phosphate and dicalcium phosphate dihydrate instead of the thermodynamically stable hydroxyapatite and of their biomimetic phase transformations during the maturation processes. The crucial role of the SBF as an electrolyte system is emphasized.