Synthesis and characterization of hydroxyapatite whiskers by hydrothermal homogeneous precipitation using acetamide

Acta Biomaterialia 6 (2010) 3216–3222

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Synthesis and characterization of hydroxyapatite whiskers by hydrothermal homogeneous precipitation using acetamide Hongquan Zhang a, Brian W. Darvell b,* a b

Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong SAR, China Department of Bioclinical Sciences, Faculty of Dentistry, Kuwait University, PO Box 24923, Safat 13110, Kuwait

a r t i c l e

i n f o

Article history: Received 29 October 2009 Received in revised form 11 January 2010 Accepted 4 February 2010 Available online 10 February 2010 Keywords: Hydroxyapatite whiskers Acetamide Hydrothermal homogeneous precipitation Morphology Crystallinity

a b s t r a c t Long and uniform HA whiskers with high crystallinity, controlled morphology and high aspect ratio were successfully synthesized by hydrothermal homogeneous precipitation using acetamide. Compared with the additive urea, which is commonly used to raise the pH to drive nucleation and growth of HA crystals, acetamide has a low hydrolysis rate under the required hydrothermal conditions. This allows better and easier control, giving rise to rapid growth of whiskers at a low supersaturation. Whisker length and width were in turn determined by solution conditions, including the concentration of Ca and PO4. Whiskers had a mean length of 60–116 lm and an aspect ratio of 68–103 for starting solutions containing 42– 84 mmol L1 Ca and 25–50 mmol L1 PO4 with a fixed Ca/P ratio of 1.67. Such whiskers are favourable for their improved bone bonding and bioactivity, as well as their mechanical properties. Whiskers were slightly Ca-deficient with Ca/P = 1.60–1.65, with the preferred direction of growth along the c-axis. Variation of acetamide concentration did not affect the constitution, the crystallinity or the crystal growth habit. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction The use of hydroxyapatite (Ca10(PO4)6(OH)2) (HA) in restorative dentistry and biomedical substitution has proved to offer several advantages, including intrinsic radio-opacity, improved wear behaviour and good bonding ability to bone directly through a Ca–P-rich layer under both load-bearing and non-load-bearing conditions [1–3]. HA-filled resin composites have shown low polymerization exotherm, low shrinkage and better mechanical properties than the parent resin [4], as would be expected from composite theory. However, general clinical applications of synthetic HA are restricted to areas free of dynamic loading due to its relative weakness and brittleness, and most HA particle-reinforced composites have fallen short in terms of durability and mechanical strength [5–7]. Whisker-reinforced composites have superior application potential as biomaterials because of their structural similarity with hard tissues, and the use of both whiskers and bioactive filling materials has drawn much attention [8,9]. Unfortunately, most available supposedly bioinert ceramic whiskers, such as Al2O3, ZrO2, TiO2, SiC and Si3N4, show decreased biocompatibility in composites; worse, some have been found to be cytotoxic in vitro [10– 12]. Thus, reinforcements such as absorbable calcium phosphate and bioactive glass materials have been developed to improve * Corresponding author. Tel.: +965 2498 6698; fax: +965 2532 6049. E-mail address: [email protected] (B.W. Darvell).

the mechanical properties and biocompatibility of biomaterials [13–15]. Both biodegradable and non-degradable composites reinforced with such materials have shown better mechanical properties and good biocompatibility in vitro. The use of HA whiskers combines both reinforcement and bioactivity. Since no indication of cytotoxicity has been found, occupational health issues associated with airborne whiskers can be avoided in manufacture and handling. Consequently, such whiskers may be useful not only as a promising reinforcement in dental restorative and bone-substitute materials, but also as materials for the preparation of bioactive scaffolds with large porosity but good mechanical properties for tissue engineering. HA whiskers and fibrous HA have been synthesized by various methods, such as dissolution–reprecipitation of acidic calcium phosphate under hydrothermal conditions, homogeneous precipitation, solid synthesis at high temperature and gel systems [16–19]. However, both long and uniform HA whiskers, i.e. with controlled morphology and composition, are difficult to obtain by most commonly used methods because structure and properties are very sensitive to the preparation conditions, and both crystallinity and thermal stability are commonly inferior. Although uniform fibrous HA can be obtained by homogeneous precipitation below 100 °C using urea as the pH control agent, it is then generally calcium-deficient, with relatively low crystallinity and a low aspect ratio [20,21]. Furthermore, the poor dispersibility of these whiskers due to entanglement or agglomeration becomes a problem when mixing with matrix materials, and the reinforcing

1742-7061/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2010.02.011

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222

efficacy is low due to the low aspect ratio and short length [22,23]. While hydrothermal homogeneous precipitation appears to be a useful method, the exact conditions that would enable the preparation of long and non-aggregated whiskers with high crystallinity, controlled aspect ratio, high purity and low dislocation density remain to be identified [24,25]. Acetamide (AA) shows a low hydrolysis rate in both acidic and basic conditions, releasing acetate and ammonia (which do not substitute in the HA lattice), e.g.:

CH3 CONH2 þ H3 Oþ ) CH3 CO2 H þ NHþ4 It has been used as an agent to drive homogeneous precipitation at <100 °C, yielding large rod-like, well-crystallized HA particles [26,27], rather than whiskers. However, no such work seems to have been conducted at or above 100 °C. Previously, it was shown that the hydrolysis rate for AA is substantially lower than that of urea in an initially modestly acidic solution from 80 to 200 °C [28], where it was also found that 180 °C as a plateau temperature for 10–15 h were the most favourable processing conditions. Since the role of such amide additives is to drive the pH higher on hydrolysis, and so drive the nucleation and growth of HA whiskers as the degree of supersaturation then rises, control of this assumes considerable importance. That is, they are to maintain a modest degree of supersaturation rather than a rapidly increasing value, which would tend to generate large numbers of small, poorly-crystalline particles. A further consideration is that urea generates carbonate, the avoidance of which is necessary to keep the purity of the resulting material high and its solubility low, as substitution for phosphate or hydroxide ions occurs [29,30]. The aim of the present investigation, therefore, was to explore the possibility of using AA under hydrothermal conditions (i.e. at 180 °C) to maximize HA crystal uniformity, length and aspect ratio, and to identify the optimum concentrations of calcium, phosphate and AA. 2. Experimental procedure 2.1. Preparation of HA whiskers Aqueous solutions containing 10–334 mmol L1 calcium and 6– 200 mmol L1 phosphate, with the Ca/P ratio fixed at 1.67 throughout, were prepared by dissolving analytical grade reagents Ca(NO3)24H2O and (NH4)2HPO4 in 0.05 mol L1 HNO3 (all AnalaR, BDH, Poole, Dorset, England) solution, with 0.25–1.5 mol L1 of acetamide (99%, Alfa Aesar, Heysham, Lancashire, England). The initial pH was adjusted to 3.0 with 0.1 mol L1 HNO3 or 1:1 ammonium hydroxide. The hydrothermal processing was in aliquots of 100 ml of a solution, as above, in an autoclave. The solution was contained in a covered 195 ml borosilicate glass test tube, supported on a cooling coil, the outer space being filled with deionized water (MilliRO, Milli-Q; Millipore, Bedford MA, USA) to give a total filling ratio of the pressure vessel volume of 70%. After processing at 180 °C for 10 or 15 h [28], the mixture was cooled naturally over 12 h to ambient temperature (25 °C). The product was then filtered off and washed quickly with deionized water four times before finally being dried in air at 80 °C, to give a yield of 85% by mass. 2.2. Characterization of HA whiskers The product was characterized using X-ray powder diffraction (XRD) (X’Pert Pro, PANalytical BV, Almelo, The Netherlands) in the range 2h = 5–80° with Cu Ka radiation (k = 0.154 nm) at 40 kV and 40 mA, and Fourier transform infrared spectroscopy (FTIR) (FTS-165, Bio-Rad, Hercules, CA, USA) in the range 4000–

3217

400 cm1 with a resolution of 4 cm1 using KBr pellets. XRD pattern-processing software (MDI Jade 5, Materials Data, Livermore, CA, USA) was used for phase identification and lattice parameter calculation. Morphology and microstructure were observed using scanning electron microscopy (SEM) (XL30CP, Philips Electron Optics, Eindhoven, The Netherlands) with an accelerating voltage of 5–10 kV at a working distance of 10–15 mm. Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) observations and selected-area electron diffraction (SAED) patterns were also made (Tecnai G2 20 S-TEM, FEI, Philips, Hillsboro, OR, USA) using a LaB6-type filament at an acceleration voltage of 200 kV with resolution 0.2 nm. Image-processing software (QWin, Leica Microsystems Imaging Solutions, Cambridge, UK) was used to measure whisker dimensions and the d-spacings of lattice planes using SEM micrographs and HRTEM images, respectively. Four or five SEM micrographs of non-overlapping fields were used to randomly select more than 80 whiskers (showing the whole length) and to determine their length and width. The aspect ratio (length/width) was then calculated for each. In addition, the chemical elements and Ca/P ratio were analysed by energy-dispersive X-ray spectroscopy (EDX) using a field-emission scanning electron microscope (LEO 1530, Oxford Instruments, Abingdon, UK). The EDX results were believed to be accurate to about 5%. The pH of the solution before the synthesis was measured with a combination electrode (GK2401C, Radiometer Analytical, Copenhagen, Denmark) at room temperature (25 ± 2 °C). After processing, the pH of the filtrate was measured as representing the final pH of the synthesis solution.

3. Results 3.1. Effect of calcium, phosphate concentration Figs. 1 and 2 show the XRD patterns and FTIR spectra of the products prepared using various concentrations of Ca and PO4 ([Ca, PO4]) with 1 mol L1 AA. All were identified as HA, with all peaks matching those of the reference data (JCPDS PDF 9-432) well, despite some variation in peak intensity ratios. The sharp peaks indicated high crystallinity, and variation of [Ca, PO4] did not affect this. The strongest peak intensity appeared for the (3 0 0) lattice plane rather than the usual (2 1 1), consistent with previous reports [19,31]. Furthermore, the peak intensity for the (3 0 0), (2 0 0) and (1 0 0) planes was rather stronger than in the reference pattern. The peak intensity ratio I(300)/I(002) showed an increase from 10 to 83 with decreasing [Ca, PO4], indicating that growth along the c-axis direction was favoured. However, the ratio I(300)/ I(210) for specimens prepared using 10–21 mmol L1 Ca was greater than at higher [Ca, PO4], indicative of rapid growth in the ab-plane, and tending therefore to give plate- or lath-like crystals. The lattice parameters calculated from the XRD patterns for the various [Ca, PO4] varied slightly, for a from 9.4285 to 9.4391 Å and for c from 6.8750 to 6.8965 Å; the unit cell had a volume of 529.28–532.73 Å3. These values coincided well with the reference values (JCPDS PDF 9-432), i.e. a = 9.418 Å, c = 6.884 Å and volume = 528.80 Å3. The a value of these whiskers was slightly larger than that of stoichiometric HA, in agreement with Trautz [32], who reported that precipitated Ca-deficient apatite had a similar c-axis parameter but an a-axis parameter larger by 0.01–0.02 Å. Elemental analysis by EDX showed that the Ca/P ratio ranged from 1.60 to 1.65 for [AA] = 1 mol L1 (Table 1), but the starting [Ca, PO4] had no apparent systematic effect. The Ca-deficiency was presumably balanced mainly by the substitution of HPO4 for PO4 [32]. This was supported by the FTIR analysis: the characteristic phosphate group bands at 1093, 1034, 602 and 564 cm1 and the bands for hydroxyl at 3571 and 633 cm1 were clearly visible [33]. Decreasing [Ca, PO4] did not affect the intensity of those characteristic bands,

3218

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222

40.0

PO43-

10

[ Ca] / mmol/L I(300)/I(210)

(200)

(100)

I(300)/I(002)

[Ca] / mmol/L

5.6

10

8.5

83.3

21

4.8

52.6

HPO42-

21

Relative Transmittance

(211) (300)

Relative Intensity (Counts)

42

OH84

42 167

4.4

11.9

H2O

OH-

84

PO43-

PO434.6

(210)

(002)

10.3

3.3

1.5

10

20

4000

167

30

3000

2000

50

60

1000

750 600 500 400

Wavenumber / cm -1

PDF 9-432, Hydroxyapatite, syn.

40

1500

70

Fig. 2. FTIR spectra of HA prepared using various [Ca, PO4] and 1 mol L1 AA for 10 h.

2θ Fig. 1. XRD patterns of HA prepared using various [Ca, PO4] and 1 mol L1 AA for 10 h.

1

HPO2 4 ,

and the bands at 871 cm , assigned to indicating calcium deficiency [34], were also of near-constant intensity. Fig. 3 shows the systematic variation in the mean length and width of the whiskers for 10 and 15 h with variation in [Ca, PO4]. Lower concentrations favoured large crystals, but with relatively low aspect ratios, the largest value of which was found at 42 mmol L1 Ca, above which concentration the length and aspect ratio declined. No obvious effect of time is present. The variation of the morphology of the whiskers with concentration is shown in Fig. 4. For 10–84 mmol L1 Ca, the precipitated material was of uniform morphology, but for [Ca] < 21 mmol L1 crystals were lathlike, while for [Ca] > 84 mmol L1 agglomerates of small fibres- or plate-like particles and stubby rods were found to accompany the whiskers. Too great a dilution led away from whisker formation, while too great a concentration gave a mixed morphology, the latter presumably due to too great a rate of nucleation initially, until the concentration had fallen to an appropriate value. 3.2. Acetamide concentration The length and aspect ratio of the resulting whiskers varied with [AA] (Fig. 5). While both the length and width increased with lower [Ca, PO4], a greater length was obtained at high [AA] with increasing [Ca, PO4]. However, the whisker width was hardly af-

Table 1 Summary of effect of [Ca, PO4] and [AA] on Ca/P ratio of products, all at 180 °C, for 10 h, starting pH 3, Ca/P = 1.67. [AA] (mol L1)

[Ca] (mmol L1)

Final pH

Ca/P ratio of product

1.00

167 84 42 21 10

5.09 5.47 5.67 5.90 6.10

1.65 1.62 1.62 1.60 1.62

0.50 0.75 1.00 1.25 1.50

42

5.32 5.55 5.67 5.82 5.86

1.58 1.59 1.62 1.60 1.57

fected. [Ca] = 33–63 mmol L1 favoured uniform long whiskers (68–172 lm) with a high aspect ratio (75–103) at 0.75–1 mol L1 AA, but at fixed [Ca, PO4], variation in [AA] had no significant effect on morphology. As in Fig. 4, for [Ca] < 33 mmol L1, a long plate- or lath-like morphology was obtained, which was unaffected by [AA]. However, irregular particles or tiny crystals often accompanied the whisker or lath-like HA for [AA] > 1.25 mol L1 (Fig. 6), and the length of the whiskers decreased. Overall, [AA] = 0.75–1.25 mol L1 was effective for preparing long and uniform HA whiskers at [Ca] 42 mmol L1. Again, it would appear that driving the nucleation rate too high (i.e. by having the pH rise too rapidly with high [AA]) led to a mixed morphology. Fig. 7 shows the XRD patterns of samples prepared using various [AA] at 42 mmol L1 Ca. All peaks were identified as HA only.

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222

3219

the ratio I(300)/I(002) was low, reflecting the morphology seen in Fig. 6. Variation in [AA] had no detectable effect on band positions or intensities in the FTIR spectra, nor on the resulting Ca/P ratio (Table 1). Again, to determine whether the slight variations in values are significant would require a more detailed analysis. Variation in neither [AA] nor [Ca, PO4] affected either the constitution or the crystallinity of the product. 3.3. Growth direction of HA whiskers Spots in the SAED pattern near the (0 0 0) plane for material prepared under near-optimal conditions (Fig. 8) were identified  (1 1 0) and (1 1 2), taken from the [2 2  0] zone axis; they as (0 0 2),

Fig. 3. Relationship of mean length and width of whiskers (log scales) with [Ca] at 1 mol L1 AA.

While there was a slight variation with the ratio I(300)/I(002) from 45 to 53 and I(300)/I(210) from 4 to 5, these values indicate the preferred c-axis growth orientation. There was no obvious change in growth habit for [AA] = 0.5–1.25 mol L1. For [AA] > 1.25 mol L1,

matched the vector relationship of crystal planes. Reflections were  (0 0 2) and (0 0 4) spots were aligned pervery strong, and (0 0 2), pendicular to the long axis of the HA whiskers (Fig. 8a), indicating good crystallinity and c-axis growth [35]. An HRTEM image recorded from a small area near the edge of a crystal showed the atomic arrangement clearly, and was used to calculate the two lattice periodicities from the Fourier transform pattern (inset in Fig. 8b): the d-spacings for the (0 0 1) and (1 1 0) lattice planes were 0.682 and 0.472 nm, respectively, in agreement with the XRD results. The longitudinal axis of the HA whiskers was parallel to the [0 0 1] direction. All these analyses were consistent. The whiskers were all single-crystal, with the

Fig. 4. SEM images of HA prepared at 1 mol L1 AA and various concentrations of Ca (mmol L1): (a) 334, (b) 167, (c) 84, (d) 42, (e) 21, and (f) 10.

3220

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222 500

[Ca] / mmol/L Length / μm

400

21 31 33 42 63

300 200 100 0 18 16

Width / μm

14 12 10 8 6 4 2 0 160

Aspect ratio

140 120 100 80 60 40 20 0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

[AA] / mol/L Fig. 5. Variation of dimensions and aspect ratio of HA whiskers with [AA].

same structure and composition. In addition, no dislocations or defect structure is visible in the TEM and HRTEM images (Fig. 8). Similar results were obtained for material prepared using other [Ca] values. 4. Discussion 4.1. Concentration of calcium and phosphate Precipitation of calcium phosphate from aqueous solutions is complicated by the possibility of several solid phases, depending on the solution composition and pH. Preparation of HA from solution always encounters two major difficulties: the incorporation of impurities and the formation of precursors prior to crystalline HA. At pH 7.4 and low supersaturation, HA can be precipitated without the formation of other calcium phosphate phases; control of the concentration of each component was therefore critical [36,37]. Here, the degree of supersaturation was controlled by the hydrolysis rate of AA. More HA nuclei are generated at high [Ca, PO4], and so there is a greater bulk precipitation rate, but this limits the crystal size and favours the formation of agglomerates. Hence, many irregular particles and their agglomerates were seen here for high [Ca, PO4]. Boskey and Posner [36] proposed that the degree of supersaturation with respect to HA should be within 105–109 to precipitate HA without ACP. When the total [Ca] and total [PO4] were each less than 2 mmol L1, the precipitate formed at pH 7.4 showed an XRD pattern distinct from that of ACP prepared at high [Ca] and [PO4] (each >10 mmol L1). In order to obtain uniform HA whiskers with

Fig. 6. SEM images of HA at 10 h using (a): 1.25 mol L1 AA and 31.3 mmol L1 Ca; (b): 1. 5 mol L1 AA and 42 mmol L1 Ca; and (c): 1.5 mol L1 AA and 62.6 mmol L1 Ca.

a high aspect ratio, it has been suggested that low supersaturation should be maintained, with the ideal concentration product of Ca and PO4 ranging from 4 to 50 mM2, at least when urea is used as the precipitation agent [18]. Here, for low [Ca, PO4], rapid growth of HA benefited from fewer nuclei being formed during heating, especially along the c-axis direction, coincident with the change in the XRD peak intensity ratio I(300)/I(002). As shown in Fig. 3, the length of the whiskers increased with decreasing [Ca, PO4], while a rapid growth in width occurred for [Ca] < 42 mmol L1. The appearance of lath-like HA depended on the final pH, which rose slightly but steadily as the initial [Ca] was reduced. That higher pH changes the crystal growth habit has been shown for other systems [38]. The HA showed preferred growth along both the c- and a-axis directions. Therefore, the peak intensity ratio I(300)/I(002) was low and I(300)/I(210) was high for the material prepared with 10 mmol L1 Ca. It follows that the concentration of Ca and PO4

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222

3221

(211)

(300)

Relative Intensity (Counts)

10

(210)

(002)

14.5

(200)

(100)

I(300)/I(002)

[AA] / mol/L

I(300)/I(210)

1.50

4.1

1.25

45.5

4.8

52.6

4.8

1.00

50.0

4.9

0.75

45.5

5.0

0.50

1.5

4.1

20

30

PDF 9-432, Hydroxyapatite, syn.

40

2θ

50

60

70

Fig. 7. XRD patterns for HA prepared at 42 mmol L1 Ca, 0.50–1.50 mmol L1 AA and 180 °C for 10 h.

should be controlled within 42–84 mmol L1 to prepare long and uniform HA whiskers, i.e. the concentration product of Ca and PO4 should be controlled in the range of 1.04–4.16 mM2 when acetamide is used under the present conditions.

4.2. Concentration of acetamide Generally, mineral precipitation occurs from supersaturated solution, and the rate of both nucleation and subsequent growth depend on the degree of supersaturation, itself controlled by the activity product of ions (IP) in the solution [39,40]. Because the combination of [Ca] and [PO4] was chosen here to give a solution supersaturated with respect to HA, IP greatly depends on the pH of the solution. Since the hydrolysis of amines is accelerated at higher temperature, releasing ammonia, the supersaturation of the synthesis solution was mainly controlled by the hydrolysis rate and concentration of precipitation agent. Because AA has a low hydrolysis rate, the pH rose gradually with increasing temperature, and the hydrolysis reaction occurred continuously with time at the designed (plateau) temperature, maintaining low supersaturation, giving smaller numbers of nuclei and facilitating whisker growth without interference from precursors. Long, uniform whiskers of high aspect ratio were thereby obtained. Although AA has seldom been used in this context [27], it now clearly appears to be the agent of choice in a hydrothermal environment, and especially if

Fig. 8. TEM images of HA prepared at 180 °C for 10 h using a solution containing 42 mmol L1 Ca and 1 mol L1 AA. (a) Low magnification, SAED pattern inset; (b) HRTEM image, Fourier transform pattern inset.

carbonate contamination is to be avoided. Urea hydrolyses rapidly from 80 °C, accelerating with temperature, and the plateau temperature had no obvious effect on pH change when this was above 160 °C, though high supersaturation was rapidly attained, resulting in more nuclei. While the crystals were needle-like, their aspect ratio was low. Thus, despite urea being widely used [18,21], it does not appear appropriate for use under hydrothermal conditions – or, indeed, in any case, if carbonate is to be avoided. On the other hand, compared with the concentration of AA used for preparation of HA by wet precipitation methods under normal conditions at <100 °C [27], the [AA] used here was low, easily keeping the solution at a low supersaturation. The final pH increased both with [AA] at constant [Ca, PO4] and with the decrease in [Ca, PO4] at fixed [AA]. At low [Ca, PO4], a small number of nuclei were formed, and the growth of HA was determined by the

3222

H. Zhang, B.W. Darvell / Acta Biomaterialia 6 (2010) 3216–3222

concentration of OH caused by the hydrolysis of AA. In the case of high [Ca, PO4], more OH ions were needed, so that the optimal value for [AA] showed a dependence on [Ca, PO4]. With increasing [AA], growth in both length and width was obtained; however, with excess [AA], more nuclei would be generated due to the increase in the degree of supersaturation. If the component concentrations were not harmonized to match the Ca/P ratio of the product, the solution Ca/P ratio would change continuously during the process. The Ca/P ratio of the solid would then be affected, and presumably the morphology and other characteristics as well. Here, no morphological differences were apparent other than the length and aspect ratio. Variation of [AA] did not affect the constitution, the crystallinity or the crystal growth habit of the whiskers. The low value in the XRD intensity ratio I(300)/I(002) for whiskers prepared with 1.5 mol L1 AA reflects the decrease in the aspect ratio. 5. Summary Well-crystallized, long HA whiskers with controlled morphology were successfully synthesized using hydrothermal homogeneous precipitation over a wide range of Ca and PO4 concentrations. Nucleation and growth of the crystals were controlled well through the slow hydrolysis of AA; however, at too high a value of [AA] small crystals with other morphologies were found. For long, uniform, single-crystal HA whiskers of high aspect ratio, [Ca] = 21–84 mmol L1 at Ca/P = 1.67 was required; this gave a length and an aspect ratio of 60–116 lm and 68–103 respectively, and Ca/P = 1.60–1.65, for preparation at 180 °C with [AA] = 1 mol L1. The growth orientation along the c-axis direction was increasingly preferred with decreasing initial [Ca, PO4]. Processing time had little effect between 10 and 15 h. Acknowledgements This work was done in partial fulfilment of the requirements of the degree of Ph.D. for Hongquan Zhang at and supported by the Faculty of Dentistry, The University of Hong Kong. We are grateful to Dental Materials Science technicians Tony D.B. Yuen and Paul K.D. Lee for technical support; Oral Bioscience technician Simon Lee for assistance with the SEM observations; and Electron Microscope Unit technicians for TEM, SEM and EDX analysis. Appendix A. Figure with essential colour discrimination Certain figures in this article, particularly Figures 3 and 5, are difficult to interpret in black and white. The full-colour images can be found in the on-line version, at doi:10.1016/ j.actbio.2010.02.011. References [1] Arcis RW, Lopez-Macipe A, Toledano M, Osorio E, Rodriguez-Clemente R, Murtra J, et al. Mechanical properties of visible light-cured resins reinforced with hydroxyapatite for dental restoration. Dent Mater 2002;18:49–57. [2] Lucas ME, Arita K, Nishino M. Toughness, bonding and fluoride-release properties of hydroxyapatite-added glass ionomer cement. Biomaterials 2003;24:3787–94. [3] Kobayashi M, Nakamura T, Okada Y, Fukumoto A, Furukawa T, Kato H, et al. Bioactive bone cement: comparison of apatite and wollastonite containing glass-ceramic, hydroxyapatite, and b-tricalcium phosphate fillers on bonebonding strength. J Biomed Mater Res 1998;42:223–37. [4] Deb S, Aiyathurai L, Roether JA, Luklinska ZB. Development of high-viscosity, two-paste bioactive bone cements. Biomaterials 2005;26:3713–8. [5] Palin WM, Fleming GJ. Low-shrink monomers for dental restorations. Dent Update 2003;30:118–22. [6] Nicholson JW. Adhesive dental materials and their durability. Int J Adhes Adhes 2000;20:11–6.

[7] Kasuga T, Ota Y, Nogami M, Abe Y. Preparation and mechanical properties of polylactic acid composites containing hydroxyapatite fibers. Biomaterials 2001;22:19–23. [8] Xu HHK, Smith DT, Simon CGCG. Strong and bioactive composites containing nano-silica-fused whiskers for bone repair. Biomaterials 2004;25:4615–26. [9] Xu HHK, Sun L, Weir MD, Antonucci JM, Takagi S, Chow LC, et al. Nano DCPAwhisker composites with high strength and Ca and PO4 release. J Dent Res 2006;85:722–7. [10] Yamamoto A, Honma R, Sumita M, Hanawa T. Cytotoxicity evaluation of ceramic particles of different sizes and shapes. J Biomed Mater Res A 2004;68A:244–56. [11] Fisher GL, McNeill KL, Smith JT. In vitro effects of fibrous and nonfibrous silicon nitride on bovine pulmonary macrophages. Environ Res 1989;50:279–88. [12] Ishihara Y, Kohyama N, Nagai A, Kagawa J. Cellular biological effects and a single transtracheal injection test in three types of whisker fibers. Inhal Toxicol 1998;10:275–91. [13] Brauer DS, Russel C, Vogt S, Weisser J, Schnabelrauch M. Degradable phosphate glass fiber reinforced polymer matrices: mechanical properties and cell response. J Mater Sci: Mater Med 2008;19:121–7. [14] De Diego MA, Coleman NJ, Hench LL. Tensile properties of bioactive fibers for tissue engineering applications. J Biomed Mater Res 2000;53:199–203. [15] Ibnabddjalil M, Loh IH, Chu CC, Blumenthal N, Alexander H, Turner D. Effect of surface plasma treatment on the chemical, physical, morphological, and mechanical-properties of totally absorbable bone internal-fixation devices. J Biomed Mater Res 1994;28:289–301. [16] Mizutani Y, Hattori M, Okuyama M, Kasuga T, Nogami M. Large-sized hydroxyapatite whiskers derived from calcium tripolyphosphate gel. J Eur Ceram Soc 2005;25:3181–5. [17] Ota Y, Iwashita T, Kasuga T, Abe Y. Novel preparation method of hydroxyapatite fibers. J Am Ceram Soc 1998;81:1665–8. [18] Zhang HQ, Wang YF, Yan YH, Li SP. Precipitation of biocompatible hydroxyapatite whiskers from moderately acid solution. Ceram Int 2003;29:413–8. [19] Tas AC. Molten salt synthesis of calcium hydroxyapatite whiskers. J Am Ceram Soc 2001;84:295–300. [20] Kieswetter K, Bauer TW, Brown SA, Vanlente F, Merritt K. Characterization of calcium-phosphate powders by ESCA and EDXA. Biomaterials 1994;15:183–8. [21] Aizawa M, Ueno H, Itatani K, Okada I. Syntheses of calcium-deficient apatite fibres by a homogeneous precipitation method and their characterizations. J Eur Ceram Soc 2006;26:501–7. [22] Becher PF, Tiegs TN, Angelini P. Whisker toughened ceramic composites. In: Mazdiyasn KS, editor. Fiber Reinforced Ceramic Composites: Materials, Processing and Technology. Park Ridge, NJ: Noyes Publications; 1990. p. 311–27. [23] Baek YK, Kim CH. The effect of whisker length on the mechanical properties of alumina–SiC whisker composites. J Mater Sci 1989;24:1589–93. [24] Yoshimura M, Suda H. Hydrothermal processing of hydroxyapatite: past, present, and future. In: Brown PW, Constantz B, editors. Hydroxyapatite and Related Materials. Boca Raton, FL: CRC Press; 1994. p. 45–72. [25] Laudise RA. The Growth of Single Crystals. Englewood Cliffs, NJ: Prentice–Hall; 1970. [26] Yasukawa A, Kamiuchi K, Yokoyama T, Ishikawa T. Preparation of lead–calcium hydroxyapatite solid solutions by a wet method using acetamide. J Solid State Chem 2002;163:27–32. [27] Yasukawa A, Takase H, Kandori K, Ishikawa T. Preparation of calcium hydroxyapatite using amides. Polyhedron 1994;13:3071–8. [28] Zhang H, Darvell BW. Effect of amine additives on the morphology of hydroxyapatite. Key Eng Mater 2008;361–363:115–8. [29] LeGeros RZ, Tung MS. Chemical stability of carbonate and fluoride-containing apatites. Caries Res 1983;17:419–29. [30] Ito A, Maekawa K, Tsutsumi S, Ikazaki F, Tateishi T. Solubility product of OHcarbonated hydroxyapatite. J Biomed Mater Res 1997;36:522–8. [31] Suchanek W, Suda H, Yashima M, Kakihana M, Yoshimura M. Biocompatible whiskers with controlled morphology and stoichiometry. J Mater Res 1995;10:521–9. [32] Trautz OR. X-ray diffraction of biological and synthetic apatites. Ann NY Acad Sci 1955;60:696–712. [33] Koutsopoulos S. Synthesis and characterization of hydroxyapatite crystals: a review study on the analytical methods. J Biomed Mater Res 2002;62:600–12. [34] Elliott JC. Structure and Chemistry of the Apatites and Other Calcium Phosphates. Amsterdam: Elsevier Science B.V.; 1994. [35] Aizawa M, Porter AE, Best SM, Bonfield W. Ultrastructural observation of single-crystal apatite fibres. Biomaterials 2005;26:3427–33. [36] Boskey AL, Posner AS. Formation of hydroxyapatite at low supersaturation. J Phys Chem 1976;80:40–5. [37] Nancollas GH, Tomazic B. Growth of calcium-phosphate on hydroxyapatite crystals–effect of supersaturation and ionic medium. J Phys Chem 1974;78:2218–25. [38] Pamplin BR. Crystal Growth. Oxford: Pergamon Press; 1980. [39] Pach L, Duncan S, Roy R, Komarneni S. Morphological control of precipitated calcium carbonates and phosphates by colloidal additives. J Mater Sci 1996;31:6565–9. [40] Koutsopoulos S, Dalas E. Hydroxyapatite crystallization on sodium cholate. J Cryst Growth 2001;222:279–86.