This is an author version of the contribution published on:
Questa è la versione dell’autore dell’opera:
[Crystal Growth & Design, Volume 16, Issue 2, 2016,
DOI: 10.1021/acs.cgd.5b01431]
The definitive version is available at:
La versione definitiva è disponibile alla URL:
[http://pubs.acs.org/doi/abs/10.1021/acs.cgd.5b01431]
Monetite-Assisted Growth of Micrometric Ca-hydroxyapatite Crystals
from Mild Hydrothermal Conditions
Linda Pastero1,2,3*, Dino Aquilano1
1Department of Earth Sciences, University of Torino, Via Valperga Caluso 35, 10125 Torino, Italy
2NIS Interdepartmental Center for Nanostructured Interfaces and Surfaces, Via Pietro Giuria 7, Torino, Italy
3CrisDi Interdepartmental Center for Crystallography, Via Pietro Giuria 7, Torino, Italy
Keywords: Hydroxyapatite; bio-ceramic; crystal growth; twinning; epitaxy; templating effect
* Corresponding Author e-mail: linda.pastero@unito.it
Dipartimento di Scienze della Terra, Università degli Studi di Torino Via Valperga Caluso 35 – 10125 Torino (Italy)
Abstract
Calcium-hydroxyapatite, HAp hereinafter, Ca5(OH)(PO4)3, the main component of vertebrate bones and teeth, plays a strategic role mainly in biomedical applications because of its
bioactivity, biocompatibility, and slow-degradation rate. It is a critical bio-ceramic material due to its properties of osteo-conduction, -integration and -induction. Moreover, HAp has a role in catalysis, agricultural and pharmaceutical products, protein chromatography and water and soil treatment as well. The bulk of investigations about HAp concerns nanosized individuals, owing to the difficulties encountered when growing large laboratory crystals. Then, deep information about surface and even bulk properties are unavoidably lacking. In this paper, we do face the relationship between the HAp polymorphism and its growth morphology both from the experimental and theoretical point of view. The micron-sized and well grown crystals we obtained are exploitable for morphological investigations, in order to better understand detailed surface properties determining the crystal reactivity. Further, a clean and effective HAp method of chemical synthesis is proposed and the involved crystal-growth mechanisms are extensively investigated as well. Finally, the unexpected synergic effect between the low supersaturation of the HAp solution and the templating effect of the monetite (CaHPO4) crystal, used as precursors, is recognized.
1. Introduction
As a strategic material for its multiple applications, the apatite group of minerals has been the subject of a wide variety of studies. The issue of the growth, stability and crystal quality of apatites has been faced from many points of view: biomedical (bioceramic applications in implantology, dentistry, reconstructive bone applications), catalysis and pharmaceutics, but also environmental and geological (ore deposits, waste remediation, soil science).
In order to prepare large calcium hydroxyapatites single crystals, many routines have been proposed both by wet and dry methods. Most of these routines involve the mass crystallization of nanoscaled HAp crystals. A wide range of techniques has been applied in order to answer to the considerable amount of questions rising from the natural variability of
the composition and structure of these crystals, from the electron microscopy to many spectroscopic methods.
A comprehensive report of the methods for the synthesis of apatites was published in 1951 by the U.S. Geological Survey 1. It represents a complete collection of the experimental work done on the crystal growth of apatites until the ‘50s.
Only a few attempts brought to the growth of clear and large HAp crystals useful for both single crystal diffraction and morphology (bulk-surface) investigations. Among all these studies, for example, in 1932, Schleede, Smith and Kindt 2 obtained the HAp from brushite (CaHPO4·2H2O) hydrolysis working under hydrothermal conditions for 500 hours and, in 1955, Hayek and coworkers 3 obtained non-pure HAp (with a few percent of sodium) from nanoscaled HAp, hydrothermally recrystallized under basic conditions.
In 1956, Perloff and Posner 4 proposed a wet method to obtain pure and large HAp crystals working at low hydrothermal conditions (300°C and evaluated water vapor pressure of 87 bar for 10 days). They started from a monetite (CaHPO4) suspension in water and obtained HAp crystals about 300 µm long. The monetite/water ratio was constrained to 1/100 in order to obtain a complete conversion of the precursor into HAp. The method brought to a decrease in pH value during HAp growth. The final pH ranged between 2 and 2.5 due to the phosphoric acid generated during the reaction.
In 1968, Kirn and Leidheiser 5 suggested a modification to the method proposed by Perloff, leaving an open door to further work. They obtained large HAp crystals from monetite hydrolysis at 350°C and about 600 bar, starting from a suspension of monetite in a diluted phosphoric acid solution. As Perloff stated, the final pH ranged around 2.5: it was also found that, at this point, the growth reaches its end-point.
In 1983, Chiranjeevirao 6 and coworkers proposed a very interesting method, ideally able to modulate the carbonate content of the apatites grown from solution at very low temperature. Unfortunately, the crystals they obtained were nanosized.
Recently, Ma and Liu 7 obtained HAp from the hydrolysis of brushite, as already proposed by Schleede in 1932. 2 Their crystals were suitable for TEM analysis and it was finally shown that monoclinic HAp could grow at low temperature.
It must be stressed that one of the main problems to be faced when growing HAp crystals, is the lack of reliable solubility data from literature. Due to the importance of the HAp stability, mainly in oral biology, a mess of studies about its solubility and its dependence from the pH have been carried out, but the values of the solubility product vary over a range of ten orders of magnitude, as observed by Wazer, 8 Larsen 9 and Chen. 10 Moreover, the solubility of HAp is incongruent and strongly dependent on the presence of calcium counter-ions like carbonate, acetate and so on. 11–16 This is a really hard point to deal with, in order to control the supersaturation of the system, and should be analyzed in depth.
In our work, we obtained large and clear crystals useful for bulk morphological studies and surface analysis. Our early experiments were performed modifying the method proposed by Chiranjeevirao, avoiding CO2 into the crystallization environment by bubbling N2 into the system and imposing a positive nitrogen pressure inside the reactor. All hydroxides were substituted by ammonia so as to limit the presence of carbonates from the hydroxide carbonation: pure nanosized HAp crystals were obtained. Then, we followed the suggestion by Kirn and Leidheiser, improving the performance of the growth method firstly proposed by Perloff and Posner. The effects of the chemistry of the system on the morphology of the HAp crystals were analyzed. From the hydrolysis reaction of the precursor in mild hydrothermal conditions, we obtained large and pure needle-like HAp crystals with well-developed morphology and some interesting surface structures.
2. Experimental
Monetite (CaHPO4), as HAp precursor, was synthesized following the procedure indicated by Perloff 4 and already described in a previous paper 17. Analytical grade H3PO4 (85%) from
Sigma-Aldrich was diluted 1:5 in ultrapure water 18.2MΩ. This solution was saturated at room temperature in calcium phosphate using analytical grade Ca3(PO4)2. The solution was heated slightly under the boiling temperature and continuously stirred. Then, the precipitate was filtered, washed with ultrapure water and dried at 40°C overnight and then measured by XRPD and Raman spectroscopy to ensure purity.
HAp were grown from a monetite/water suspension using 45mL PFTE lined stainless steel autoclaves (PARR Instrument Company). A small amount (typically 0.1g) of the precursor was sealed into the PFTE liner with ultrapure 18.2 MΩ water. The precursor/water weight ratio was kept constant at 1/100. In order to avoid CO2 contamination, all the samples were prepared and sealed into the steel autoclaves in N2 atmosphere, the precursor was kept free from gas before use and the CO2 water content was controlled by bubbling N2. The autoclaves were put in an oven at constant temperature of 220°C (the higher limit for the PFTE liners), and autogenic pressure (estimated of 20 bar by water vapor pressure at that temperature). Reaction products were left into the reactors at constant temperature for a time ranging between five days and three weeks; then, the autoclaves were quenched and the final product filtered and air dried all over night.
A variation of the routine was performed using microwave heating instead of the traditional convection oven. In this case a 45 mL, chemically inert vessel for microwave applications (PARR Instrument Company) was used. Growth runs were performed in household microwave oven at 600W for five minutes. Due to the operating hazards related to the use of such autoclaves, the samples were cooled slowly to room temperature and then the crystallization product was filtered and dried all over night as well.
2.1. XRPD
X-ray powder diffraction was carried out on all the precursors and all the obtained HAp samples to check the sample purity. A Siemens D5000 diffractometer whit Bragg-Brentano
geometry, 2.5° < 2θ < 100°, step size 0.01, 1 s/step scan time was used for the collection of diffraction data useful for phase determination. When committed to peak shape analysis, the XRPD diagrams were collected in the interval 30° < 2θ < 60°, step size 0.005 and 20 s/step. 2.2. SEM characterization
SEM imaging of the samples was performed using both a Cambridge S-360 Scanning Electron Microscope equipped with secondary electron (SE) and backscattered electron (BSE) detectors. The typical experimental conditions were: W filament, EHT 25 kV, probe current 100 pA, working distance 5 mm and a ZEISS SUPRA™ 40 Field Emission Scanning Electron Microscope (FESEM) (WD = 3mm, Aperture size = 30.00µm, EHT = 5kV)
2.3. TEM characterization
Transmission electron microscopy was performed on HAp samples showing the smallest size. A JEOL 3010 UHR-HR TEM (300 kV, Beam current = 114µA LaB6 source and 2k x 2k pixels Gatan CCD camera attached).
2.4. AFM characterization
AFM measurements were performed using a DME SPM Microscope (DME Igloo, Denmark) equipped with a DS95-50E scanner (scan volume 50x50x5µm). Data were acquired using MikroMasch Ultrasharp NSC16/Si3N4 Cr-Au back-coated cantilevers with typical resonance frequency 190 kHz, force constant 45 N/m, tip radius lower than 35nm and full tip cone angle 40°. All measurements were performed in alternated contact mode.
2.5. Raman spectroscopy
A high-resolution confocal µ-Raman system by Horiba Jobin Yvon HR800 was used, equipped with two gratings (1800 and 600 grooves/mm), air cooled CCD detector and a green polarized lasers (solid state Nd, 532 nm, 250 mW). The system is completed by Edge filters (532 nm) and interferential filters. The maximum resolution with the grating 1800 grooves/mm and green laser is 2 cm-1.
pH and conductivity measurements were obtained using a Thermo Scientific Orion 4-star benchtop pH/ISE meter equipped with an Orion™ 9107BN Triode™ 3-in-1 pH electrode for temperature compensated pH measurements and a Orion™ DuraProbe™ 4-Electrode conductivity ATC cell with nominal cell constant of 0.475cm-1. Ion chromatography on solutions was performed using a Metrohm 883 Basic IC plus ion chromatograph.
2.7. Equilibria calculations
Aqueous solution speciation and equilibria were calculated using PhreeqC software using the Minteq v.4 database for thermodynamic data because of the availability of HAp, brushite and monetite as crystalline phases.
2.8. XRPD Peak profile analysis
The profile analysis on powder diffraction peaks was performed by using the Fityk software 18. Due to the instrumental broadening, the PearsonVII profile was chosen. The built-in function was modified in order to calculate both the Kα1 and Kα2 contributions to the peak shape.
3. Results and Discussion
3.1. HAp from monetite hydrolysis following the Perloff’s routine
The routine proposed by Perloff and Posner in 1956 has proven to be a highly effective approach to the growth of large (more than 300µm) and optically clear HAp crystals with smooth and well developed surfaces. For the sake of clarity, the reference cell used hereinafter for the HAp is the same reported in previous papers dealing with the HAp equilibrium shape and twinning 17,19 (monoclinic, P21/c, a0 = 9.4214 Å, b0 2a0, c0 = 6.881 Å; α = β = 90°, γ 120°)
Monetite (triclinic, P ´1 , a0=6.91 Å, b0=6.627 Å, c0=6.998 Å, α=96.34°, β=103.82°, γ=88.33°) as synthesized by the method described in the following, shows a usual {010} flattened habit, with an approximately rhombic outline.
XRPD and µ-Raman measurements confirmed the purity of the precursor. Synthesis products, obtained by the hydrothermal way and described in Section 2.2, were confirmed by XRPD data.
At the end of each run, it is possible to recognize at least three generations of HAp crystals: i) the first generation of large crystals optically clear, showing an undeniable
pseudo-hexagonal morphology and frequently twinned, with the Original Composition Plane of the twin (OCP) parallel to the elongation axis of the crystal 20, as can be seen in Figure 1a;
ii) the second generation of smaller crystals with a pseudo-hexagonal morphology and evident twinning both parallel and perpendicular to the crystal elongation; iii) the third generation of small crystals, with an average size of a few micrometers
and a tape-like (or ribbon-like) morphology similar to that described by Elliott 20 for the biogenic HAp, and by Ma 7 for the HAp obtained from brushite hydrolysis. Also the third generation is not twin-free, showing the OCP parallel to the crystal elongation (Figure 1b and Figure S1, see Supporting Information).
The multiple nucleation of HAp lead to a broad dispersion of crystal sizes, ranging from a few nanometers up to 300 µm.
Moreover, partially dissolved crystals of monetite are frequently detected, even for long reaction times, as shown in Figure 1c, d, pointing out the incompleteness of the reaction.
Figure 1. a) Many generations of crystals can be detected in a sample grown from the hydrolysis of monetite at low hydrothermal conditions; b) ribbon-like crystal belonging to the last generation are commonly found. Usual twinning is parallel to the elongation of the crystal. c), d) Under these conditions, not completely dissolved crystals of monetite are regularly found.
In order to obtain a sequence of snapshots of the chemical behavior of the system during HAp precipitation, growth experiments are repeated and the samples are quenched at 0 °C (at the water triple point) after 4, 15, 18, 21.5, 32, 56, 72, 96 and 168 hours. pH, conductivity, calcium and phosphate concentrations are measured. The chemical trends of the crystal/solution system during the crystallization process are reported in Figure 2. The values for the 96 hours long run are not reported because of the unsatisfactory quality of the sample.
Figure 2. pH, conductivity and concentration of free calcium (black dots) and free phosphate species (white dots) of the main species in solution are recorded during a growth experiment. The system reaches the chemical stability between 18 and 56 hours. After 56 hours, the production of phosphoric acid rises, lowering the pH. The bottom right graph shows the trend of the gap between the concentration of the phosphate and the calcium species in solution.
The HAp stability field calculated with the PHREEQ software 21 indicates that HAp is stable, starting from pH=4.5 up to basic values for temperatures higher than 25°C, while the pH of the experimental system moves toward low pH values. In these conditions, the presence of HAp crystals with well-developed surfaces and sharp edges cannot be explained in terms of homogeneous nucleation; as a matter of fact, a foreign crystal phase (monetite) can promote the HAp nucleation, remembering that heterogeneous nucleation can be favored as well, in systems unsaturated with respect to the nucleating phase, when good epitaxial conditions between the two phases are fulfilled.
The reaction of monetite in water and the crystallization of HAp can be described as a “three-step process”:
i) during the early stages and immediately after the thermal equilibrium is reached, monetite dissolves and HAp heterogeneous nucleation occurs. The pH suddenly decreases from 7.4 to 3.3. At the same time, the conductivity of the solution rises, due to the increasing amount of free ions in solution (mainly hydrogen ions). After 18 hours of reaction, large and clear HAp crystals are already present in the sample (first generation of crystals), but monetite is still abundant, as shown in Figure S2a (see Supporting Information). Even if other transient phases, for example octacalcium phosphate, known as a HAp precursor in enamel and bones formation 22,23 could be taken in consideration during this first step, XRPD measurements confirmed that only monetite and HAp are present.
ii) Both pH and concentration of the free ions in solution reach a plateau after 18 hours. HAp nucleates and grows, indicating that the system is substantially far from equilibrium. Three HAp crystal generations can be described, at least, as detailed at the beginning of the section. The presence of multiple generations of crystals evidences an oscillating supersaturation with a maximum (corresponding to the dissolution of the monetite) and a minimum (corresponding to the HAp nucleation and growth. Moreover, between 18 and 72 hours of reaction, the concentration of calcium and phosphate ions follows the same trend, with a gap that lingers constant until 56 hour of reaction. Inasmuch, as the Ca++/PO43- ratio in monetite is 1 and in HAp is 5/3, the constant gap between calcium and phosphate during the process of dissolution and recrystallization can be obtained only when an oscillatory behavior of the supersaturation is considered.
iii) After 72 hours, the concentration of the free phosphate rises, the pH value decreases to 2.2, while the conductivity of the solution suddenly rises, so pointing out that the hydrolysis of the monetite becomes the driving process. The lack of information in this interval is due to the low quality of the data obtained from the sample extracted after 96 hours of reaction. After a week of growth (Figure S2 b and c), a few crystals of monetite are still present into the solution and HAp crystals exhibit their typical pseudo-hexagonal aspect showing evident composition planes of the twinning both along and perpendicular to the crystal elongation.
After a week of reaction, the quality of the HAp surfaces becomes worthy of careful attention. All crystal surfaces show a rough behavior during the first stages of growth, as shown in Figure S2 d, mainly if the closure of the crystal tips is considered.
Figure 3. a) General behavior of the crystallization of HAp after three weeks of growth; b) pseudo-hexagonal morphology of the crystals of the first generation and c) apparent lower symmetry of the second generation of crystals; d) crystal twinning
After this period, the HAp single crystals in general exhibit a pseudo-hexagonal morphology, as shown in Figure 3. Two kinds of OCPs twinning can be found: the most frequent parallel to the elongation axis of the twin (as we just described) and the second one coincident with the 010 plane, and hence perpendicular to the previous one. The detailed description of the twinning law can be found in a recent paper 17.
A modification of the Perloff’s method was introduced by growing HAp crystals from a monetite/water suspension in a microwave oven. This allowed to exclude the role of heat transport by convection on the presence of several crystal generations. After 5 minutes of microwave radiation (600W), HAp crystals growth is comparable with that obtained after a week in a conventional oven.
At the SEM scale, some evidences of growth relationships between the dissolving precursor and the growing crystals are found, as shown in Figure 4.
Figure 4. a) Evidence of some growth relationships between monetite and HAp. b) The white dashed line corresponds to the [010] elongation of the HAp crystal. Solid white lines correspond to the [001] direction of the monetite. c) The geometrical and morphological relationship between the two crystals is shown in the sketch.
As illustrated in Figure 4, the angle of 14° approximates the angle (13.82°) arising between the [010] direction of the monoclinic HAp and the [001] direction of the monetite when the two crystals are epitaxially related. The corresponding geometrical relationship between the two phases is described in Table 1. The crystal forms we considered are confined to the 010
of monetite and to those HAp forms enclosing the pseudo-hexagonal prism, due to their morphological importance in natural and synthetic crystals
3.1.1. Geometric epitaxial relationships between host (monetite) and guest (HAp) crystal.
There are two extreme kinds of coincidence cells. In the first case, (see Table 1a along with Figure 5) their multiplicity are decidedly low (especially for the coupling 010monetite/ 001HAp). Moreover, the angular misfit (obliquity) for the just mentioned coupling does not exceed 1.68°, so proving that the density of screw dislocations at this epitaxial interface should be very low. Thus, the epitaxy between the two phases has to be considered fairly good, due to the high short-range interaction across the common interface.
Table 1a. Short-range geometric epitaxial relationships between host (monetite) and guest (HAp) crystal.
Crystal form (host) Monetite
2D-lattice (Å) of the host form
Crystal form (guest) HAp
2D-lattice (Å) of the guest form
Misfit % 2D-Coincidence cell multiplicity 010 4 [100] = 27.64 001 3 [100] = 27.9759 +1.21 4 (010)monetite [10-1 ´1 ] = 10.946 [1-1 ´1 0] = 11.6305 +6.25 Area of the 2D coincidence cell (Å)2 187.825 194.44 +3.52 5 d020 = 16.466 2 d002 = 16.156 1.92 010 8 [100] = 55.28 100 3 [001] = 55.9308 +1.18 8 (010)monetite [301] = 20.238 [01-1 ´1 ] = 19.897 1.71 Area of the 2D coincidence cell (Å)2 375.65 388.736 +3.48 5 d020 = 16.466 4 d100 = 16.156 1.92 010 8 [100] = 55.28 10-2 3 [201] =55.965 +1.24 8 (010)monetite [301] = 20.238 [211] = 19.9076 1.66 Area of the 2D coincidence cell (Å)2 375.65 388.973 +3.54
In the second case, the multiplicities are fairly high, so proving that long-range interactions are needed at the epitaxial interface, even if the obliquity is null, in order the epitaxy could occur.
Table 1b. Long-range geometric epitaxial relationship between host (monetite) and guest (HAp) crystal.
Crystal form (host) Monetite
2D-lattice (Å) of the host form
Crystal form (guest) HAp
2D-lattice (Å) of the guest form
Misfit % 2D-Coincidence cell multiplicity 010 4 x [100] = 27.64 001 3 x [100] = 27.9759 1.2 16 (010)monetite [104] = 27.1816 4 x [010] = 27.8012 2.3 2D area 773.25 777.76 0.6 010 8 x [100] = 55.28 100 3 x [001] = 55.93 1.2 32 (010)monetite [104] = 27.1816 4 x [010] = 27.8012 2.3 2D area 1502.6 1555.5 3.5 010 4 x [100] = 55.28 10-2 3 x [201] =55.94 1.2 32 (010)monetite [104]= 27.1816 4 x [010] = 27.8012 2.3 2D area 1502.6 1555.2 3.5
Figure 5. The two extreme coincidence cells of monetite (dots) and HAp (crosses) are shown: the short-range coincidence across the common interface (red) and the long-range one (blue).
In between the two extreme cases just illustrated, two other families of lattice coincidences can be found at the interface between the {010} form of monetite and the HAp “prismatic” forms, as documented in the Supporting Information (Figure S3). It is worth outlining that in all the cases we explored the angle formed between the elongation axis of HAp with the z axis of monetite is always 13.82°, so confirming our preliminary conclusion at the end of Section 3.1.
Summing up, the experimental observation shown in Figure 4 along with the just evidenced lattice coincidences, unambiguously prove that HAp and monetite can be intimately related during growth, from low to high supersaturation values (with respect to HAp) in the mother phase.
3.2. HAp from monetite hydrolysis following a modified Perloff’s routine
As realized by Perloff and Posner in 1956, the result of the reaction is strictly dependent on the final pH of the solution: “…Apparently the controlling factor for the hydrolysis is the final pH of the liquid. As long as this pH stays above 2.0 to 2.5, the reaction will proceed in the desired direction...”
In our batches, the measured chemical quantities follow the trends illustrated in Figure 2 and the final pH reaches a value of 2.2. As stated in the previous paragraph, this value is related to the excess of phosphoric acid during the hydrolysis reaction of monetite in water.
Starting from this observation, a modification of the Perloff’s method was introduced by adding small amounts of phosphoric acid to the initial HAp/water suspension, in order to impose a low pH value since the beginning of the experiment. In this way three goals were reached:
i) the dissolution of monetite is almost complete since the early stages of the reaction;
ii) the supersaturation does not fluctuate during the simultaneous processes of monetite dissolution and HAp nucleation. In this way, the multiple nucleation effect is avoided and only one generation of crystals is obtained.
iii) the supersaturation of the solution, with respect to the HAp, is lowered, so moving the system out of the HAp stability field, and hence to the decreasing of the nucleation rate.
At low pH values, monetite does totally turn to HAp. Moreover, moving out the system to the stability field of the HAp, lowers the nucleation frequency, so leading to less HAp crystals with narrower size dispersion and controlled growth morphology. The presence of many generations of HAp as the result of a multiple nucleation process and the presence of monetite as a leftover of the incomplete dissolution process are avoided. The monotonic decrease of the supersaturation allows a crystal size selection and a very good crystal quality, as can be seen in Figure 6 a and b.
In order to confirm the effect of the pH on crystal size and morphology, a countercheck was performed growing HAp crystals from a monetite suspension (obtained using a phosphate buffer solution: pH 7.4, Cold Spring Harbor Protocol 24, CSP). The grown crystals are smaller and their quality is noticeably lower, as shown in Figure 6 c and d.
Figure 6. a) and b) HAp grown from a low pH (2.5) suspension of monetite in diluted phosphoric acid. c) and d) HAp crystals from a buffered suspension of monetite (CSP buffer, pH 7.4)
A simulation of the system was carried out changing the pH by adding small amount of H3PO4: the PHREEQ software was used along with a Minteq v.4 database with thermodynamic data for HAp, brushite and monetite. Being acquainted with the low reliability of thermodynamic data and solubilities of phosphates (the problem of the solubility data discussed previously), we are interested in the relative trend of the saturation indices, not in their absolute values. At high temperature, the HAp becomes supersaturated when pH>4. This trend confirms the experimental results: lowering the pH suspension allows to control both the 3D homogeneous nucleation rate and the final size of HAp crystals.
When pH < 2, monetite recrystallizes in well-formed and large crystals.
3.3. Crystal characterization
A discussion about equilibrium and growth morphology of single and twinned HAp crystals (both experimental and theoretical) has already been presented in our previous works. 17,25,26 At first glance, the growth morphology of HAp crystals seems to be hexagonal, with a well-developed {100} six-faces prism and {101} pyramids, but a careful morphological analysis excludes the hexagonal symmetry. Monoclinic single crystals show three pinacoids: {100}, {10 ´2 } and {001} belonging to the zone [010] and the related prisms {110}, {11 ´2 } and {012}. The pseudo-hexagonal growth morphology of the single crystals comes out from the sharp surface resemblance of the three pinacoids and then to the quasi- equivalence of their kinetic behavior.
The twinning is generated through a three-fold rotation around the [010] direction (A3 21 axis), as already discussed in the recent literature. The original composition face (OCF) can either belong to one of the three {h0l} pinacoids or to the basal {010} pinacoid. Twinning occurs since the early stages of growth, as proven by the presence of large stitches resulting from the concave dihedral angle effect 27.
In Figure 7a, at the tip of the crystal, the illusory hexagonal symmetry of the crystal decreases to a trigonal pseudo-symmetry. Moreover, a deep groove runs along the crystal, pointing out the existence of the twinning. In Figure 7b, the three fold twinning is once more demonstrated by the three-tips funnel rising at the crystal termination, whereas a single crystal has no reason to be concave.
Figure 7c shows a pseudo-hexagonal double terminated HAp crystal. Deep grooves can be seen in correspondence of the twinning sutures, both parallel and perpendicular to the crystal elongation. Both twins follow the same law, but in this case the original composition face belongs to the {010} pinacoid.
Two twins, viewed along the [010] twin axis, are shown in Figure 7d, e, the mismatch between the two individuals (building each twin) being underlined by an irregular suture. In Figure 7c, d, the pseudo-bipyramid shows an overlapping structure that will seal the OCF suture.
A manifest evidence of the misleading hexagonal pseudo-symmetry is shown in Figure 7e and f, where the lack of symmetry is stressed by the surface patterns that are present only on one side of the crystal edge and consequently, cannot be ascribable to the hexagonal symmetry.
Figure 7. a) The apparent morphological hexagonal symmetry of the crystal decreases to a trigonal pseudo-symmetry (white solid arrows). A large stitch runs along the crystal elongation (white dashed arrow). b) The threefold twinning of the crystal is evidenced by the presence of the triangular hole at the crystal tip (white dotted arrow, see also Figure S5) and of the deep stitches running along the crystal (black dotted lines and arrow). c) A pseudo-hexagonal double terminated crystal of HAp showing both the original composition faces, parallel and perpendicular to the crystal elongation. d) and e) twinned crystals with deep stitches. In e) and f) a surface structure with a symmetry not compatible with an hexagonal one is detected.
The AFM surface analysis has been carried on the pseudo-hexagonal prisms of some large and clear crystals. Their surfaces are atomically flat. The stitches between twinned crystals, both along and perpendicular to the crystal elongation are well detected and their behavior is shown in Figure S5.
A careful analysis of the XRPD pattern of a representative sample can be useful to fully investigate the symmetry of the HAp crystals obtained by the hydrolysis of the precursor. As already shown 17,28 the peak shape analysis is a reliable method to reveal the overlapping of many reflections lying under a seemingly individual peak.
The decomposition of the diffraction pattern is reported in detail in Figure S6. Neither smoothing nor filters were applied to the experimental data. The obviousness of the multiplicity of some diffraction peaks is impressive. For example, in the range 31.5° < 2θ <33.5°, the hexagonal polymorph should show only three diffraction peaks whereas the monoclinic one should allow 17 diffraction peaks. The decomposition of the XRPD pattern allow us to definitely assign nine diffraction peaks. The loss of multiplicity is related to the twin operation (120° rotation around the [010] axis) of the monoclinic crystals, that illusory increases the symmetry of the phase, from monoclinic to hexagonal.
4. Conclusions
In this work, we modified the Perloff’s routine, and showed that large and optically clear HAp crystals can be produced with controlled size. As a matter of fact, the size dispersion which is usually related to more than one nucleation events, is avoided. Our objectives have been achieved according to the following path:
i) The pH of the reaction has proven to be the most important variable to control, in order to select the crystal size and improve the quality of their surfaces. Most of the methods proposed up to now require high pH values (usually ranging between
10 and 12) to obtain the HAp mass precipitation; under these conditions the supersaturation of the solution is definitely too high to obtain large crystals. Indeed, in most cases, only nanosized HAp crystals are obtained, and it’s not without reason that, up to now, nothing about surface properties and quality has been mentioned. Lowering the pH, shifts the crystallizing system at the limit of the HAp stability field, so reducing the nucleation frequency and allowing to grow larger crystals.
ii) Moreover, we demonstrated as well that the epitaxial interaction between HAp and monetite used as a precursor phase, should favor the 2D heterogeneous HAp nucleation, so allowing the HAp to appear beyond its usual growth conditions. The effect of the choice of the precursor on the HAp growth under mild hydrothermal conditions will be described in a forthcoming paper.
It is mandatory to bring up a great issue concerning the solubility of the HAp: the values found in literature show an unbelievable variability, due mostly to the incongruent crystal solubility and to its dependence from the presence of many omnipresent calcium counter-ions, like carbonate. This is an unavoidable “impasse” when growing HAp crystals because it prevents the control of the operational supersaturation of the system.
From the solubility viewpoint, the low pH conditions adopted in this work contribute to simplify the system, avoiding the inadvertent presence of CO2, even working in nitrogen atmosphere. This implies a lower effect of the incongruent solubility of the HAp and a further simplification of the system with respect to the most diffused routine syntheses that require the use of a basis in order to trigger the HAp precipitation.
As concerns the HAp growth morphology, new acquisitions have been obtained:
- During the early stages of growth, HAp crystals exhibit a ribbon-like morphology frankly monoclinic and are unambiguously twinned. The ubiquitous three fold twin axis leads to the typical pseudo-hexagonal morphology of the crystal and, during
growth, the stitches that characterize the twin shape disappear, leaving a pseudo-hexagonal crystal morphology, usually misunderstood. The twinning evidences are either aligned with the crystal elongation (with the original composition plane pertaining to the forms {001}, {10 ´2 } and {100} that limit the pseudo-hexagonal prism) or across the crystal elongation (with the original composition plane pertaining to the form {010}). Both the twin growth morphology have been described in a previous paper. 17
- The pseudo-hexagonality of the crystals is demonstrated also by some surface patterns detected at the SEM scale. The surface quality of the pseudo-prisms has been investigated by atomic force microscopy. At the AFM scale, the forms {001}, {10 ´2 } and {100} of the largest crystals are atomically flat and frequently show scars of stitches, that have been generated by the twinning and, in turn, restored through the dihedral angle effect.
Finally, we can say that this is the first report of a reproducible, clean and easy to use method to obtain HAp crystals with well-developed morphologies that can be used for both bulk and surface investigations.
Supporting Information Available: Figure S1. TEM images of some twinned HAp crystals. Figure S2. CaHAp twinned crystals from monetite hydrolysis after 18 and 24 hours of reaction. Figure S3. The intermediate coincidence cells of monetite and HAp. Figure S4. A schematic representation of a monoclinic HAp crystal twinned around a three-fold axis parallel to the [010] of the crystal Figure S5. AFM topography of the pseudo-prism surfaces showing some evidences of the closure of the stitches originated by twinning. Figure S6. XRD peak decomposition performed using a Pearson VII peak profile.
Acknowledgements
This research was supported by MIUR (GEO-TECH Project, PRIN 2011-2012). The hydrothermal growth facilities and the AFM laboratory at the Department of Earth Sciences, University of Torino were funded by Fondazione CRT (grant. N. 2014.2187 and grant N. 2014.1042)
The Authors are deeply grateful to Dr. Enrico Destefanis for chromatographic measurements, Dr. Ruggero Vigliaturo and Dr. Mauro Giorcelli for FESEM imaging, and would like to thank the anonymous Referee for both criticisms and fruitful suggestions.
X(1) Jaffe, E. B. Abstracts of the literature on synthesis of apatites and some related
phosphates, GEOLOGICAL SURVEY CIRCULAR 135, 1951.
(2) Schleede, H. A.; Schmidt, W.; Kindt, H. Ztschr. Elektrochem. 1932, 38, 633–641. (3) Hayek, E.; Lechleitner, J.; Böhler, W. Angew. Chemie 1955, 67, 326.
(4) Perloff, A.; Posner, A. S. Science 1956, 124, 583–584.
(5) Kirn, J. F.; Leidheiser, H. J. Cryst. Growth 1968, 2, 111–112.
(6) Chiranjeevirao, S. V.; Voegel, J. C.; Frank, R. M. Inorganica Chim. Acta 1983, 78, 43– 46.
(7) Ma, G.; Liu, X. Y. Cryst. Growth Des. 2009, 9, 2991–2994. (8) Van Wazer, J. R. Angew. Chemie 1961, 73, 552–552. (9) Larsen, S. Nature 1966, 212, 605–605.
(10) Chen, Z.-F.; Darvell, B. W.; Leung, V. W.-H. Arch. Oral Biol. 2004, 49, 359–367. (11) Nancollas, G. H.; Mohan, M. S. Arch. Oral Biol. 1970, 15, 731–745.
(13) Wu, W.; Nancollas, G. H. Adv. Colloid Interface Sci. 1999, 79, 229–279.
(14) Xie, A.; Shen, Y.; Li, X.; Yuan, Z.; Qiu, L.; Zhang, C.; Yang, Y. Mater. Chem. Phys. 2007, 101, 87–92.
(15) Chuong, R. J. Dent. Res. 1973, 52, 911–914.
(16) Clark, J. S.; Peech, M. Soil Sci. Soc. Am. J. 1955, 19, 171.
(17) Aquilano, D.; Bruno, M.; Rubbo, M.; Pastero, L.; Massaro, F. R. Cryst. Growth Des. 2015, 15, 411–418.
(18) Wojdyr, M. J. Appl. Crystallogr. 2010, 43, 1126–1128.
(19) Aquilano, D.; Bruno, M.; Rubbo, M.; Massaro, F. R.; Pastero, L. Cryst. Growth Des. 2014, 14, 2846–2852.
(20) Elliott, J. C. Nature 1971, 230, 72.
(21) Parkhurst, D. L.; Appelo, C. A. J. In 6; USGS, 2013; p 497.
(22) Suzuki, O.; Imaizumi, H.; Kamakura, S.; Katagiri, T. Curr. Med. Chem. 2008, 15, 305– 313.
(23) Suzuki, O. Acta Biomater. 2010, 6, 3379–3387. (24) Cold Spring Harb. Protoc. 2006, 2006, pdb.rec8303.
(25) Aquilano, D.; Bruno, M.; Rubbo, M.; Massaro, F. R.; Pastero, L. Cryst. Growth Des. 2014, 14, 2846–2852.
(26) Pastero, L.; Cámara, F.; Bruno, M.; Rubbo, M.; Aquilano, D. Acta Crystallographica
Section A: Foundations and Advances. International Union of Crystallography August
5, 2014, p C1115.
(27) Boistelle, R.; Aquilano, D. Acta Crystallogr. Sect. A Found. Crystallogr. 1978, 34, 406–413.
For Table of Contents Use Only
Monetite-Assisted Growth of Micrometric Ca-hydroxyapatite Crystals
from Mild Hydrothermal Conditions
Linda Pastero, Dino Aquilano
Micron-sized and well-grown hydroxyapatite crystals useful for morphological investigations and detailed surface analysis are obtained from mild hydrothermal conditions. The clean and effective method of synthesis is described and the crystal-growth mechanisms involved are extensively investigated. The unexpected synergy between the low supersaturation of the solution and the templating effect of the precursor (monetite) on the growth of hydroxyapatite crystals is recognized and described.
