Structural analysis of a series of strontium-substituted apatites

Structural analysis of a series of strontium-substituted apatites

Available online at www.sciencedirect.com Acta Biomaterialia 4 (2008) 1455–1464 www.elsevier.com/locate/actabiomat Structural analysis of a series o...

395KB Sizes 1 Downloads 41 Views

Available online at www.sciencedirect.com

Acta Biomaterialia 4 (2008) 1455–1464 www.elsevier.com/locate/actabiomat

Structural analysis of a series of strontium-substituted apatites M.D. O’Donnell *, Y. Fredholm, A. de Rouffignac, R.G. Hill Imperial College London, Department of Materials, South Kensington Campus, Exhibition Road, London SW7 2AZ, UK Received 6 November 2007; received in revised form 26 March 2008; accepted 16 April 2008 Available online 3 May 2008

Abstract A series of Sr-substituted hydroxyapatites, (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00, were made by a standard wet chemical route and investigated using X-ray diffraction (XRD), Rietveld refinement and Raman spectroscopy. We report apatites manufactured by two synthesis routes under 90 °C, and only the fully Sr-substituted sample had a small amount of an impurity phase, which is believed to be strontium pyrophosphate. Lattice parameters (a and c), unit cell volume and density were shown to increase linearly with strontium addition and were consistent with the addition of a slightly larger and heavier ion (Sr) in place of Ca. XRD Lorentzian peak widths increased to a maximum at x = 0.50, then decreased with increasing Sr content. This indicated an increase in crystallite size when moving away from the x = 0.50 composition (d  9.4 nm). There was a slight preference for strontium to enter the Ca(II) site 1 in the mixed apatites (6 to 12% depending on composition). The position of the Raman band attributed to v1 PO3 in 4 at around 963 cm 1 hydroxyapatite decreased linearly to 949 cm at full Sr-substitution. The full width at half maximum of this peak also correlated well and increased linearly with increasing crystallite size calculated from XRD. Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Raman; Apatite; Strontium; XRD; Rietveld

1. Introduction It is known that strontium and other divalent cations with a similar charge-to-size ratio to calcium can readily substitute in the lattice of hydroxyapatite. Also it has been reported that there is immiscibility in this system and complete solid-solution cannot occur, although this may be due to poor sample preparation [1]. Apatites have many applications, from optoelectronics [2] to waste immobilization [3] and biomaterials [4]. Strontium is also beneficial for biological applications for bone regeneration due to the recent success of treatments such as strontium ranelate [5], which stimulates bone formation, decreases bone resorption and reduces the risk of vertebral fractures in postmenopausal osteoporosis. Both calcium- and strontium-hydroxyapatite are hexagonal (space group P63/m) [6,7]. No complete systematic studies have been performed on this system manufactured by a wet chemical route to our knowledge, and *

Corresponding author. Tel.: +44 20 7594 6814; fax: +44 20 7594 6757. E-mail address: [email protected] (M.D. O’Donnell).

samples are typically only analysed by X-ray diffraction (XRD). Heijligers et al. [1] performed a structural study of Sr–Ca apatites made by solid state synthesis using XRD but had significant tristrontium phosphate present for compositions above 80% Sr-substitution. Bigi et al. looked at the structure of 0, 20 and 60% substitution of Sr for Ca in hydroxyapatite [8]. Raman microspectrometry is powerful investigative tool, as the probe is 10–100 times smaller than similar micro-infrared spectroscopic techniques. It is useful for biological samples due to its non-invasive and non-destructive nature, and fluorescence is minimized. No special sample preparation is required as the technique is performed in reflection and hence artefacts from the preparation process are eliminated. Complex studies can be performed, such as at interfaces and line-scans of cross-sections. The functional groups present in biological and synthetic apatites,  such as phosphate (PO3 4 ), hydroxyl (OH ), acidic phos2 2 phate (HPO4 ), carbonate (CO3 ) and fluoride (F), can be probed, and peak width, position and multiplicity correlated with composition and structure.

1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2008.04.018

1456

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

The purpose of this study was to investigate the crystalline structure of a series of apatites, where Ca is progressively substituted for Sr, by XRD, Rietveld refinement and Raman spectroscopy. 2. Experimental 2.1. Synthesis Strontium-substituted hydroxyapatites were synthesized by a wet chemical route. Two reaction routes were used [9]. 2.1.1. Route 1 Solution A preparation. Calcium nitrate 4-hydrate (Ca(NO3)24H2O) and/or strontium nitrate (Sr(NO3)2) were dissolved in 200 ml of distilled water. Table 1 gives the weights used for each apatite composition. The pH of the solution was adjusted to 11 using 30% ammonia solution, then 400 ml of distilled water was added to the solution. The solution was constantly stirred. Solution B preparation. Diammonium hydrogen orthophosphate ((NH4)2HPO4) was dissolved in 120 ml of distilled water. The amount of (NH4)2HPO4 used is given in Table 1. The pH of the solution was adjusted to 11 using ammonia solution; however, no hydroxides precipitated, which have lower solubility than amide salts, tricalcium phosphate and apatite. Distilled water (160 ml) was then added to the constantly stirred solution. The solution was filtered using a Buchner funnel, as described [9]. Solution A–solution B mix. Solution B was added dropwise to the constantly stirred solution A over a period of approximately 2 h. The pH was kept at around 11. A precipitate formed and was stirred for 1 h, before being left overnight. The precipitate was centrifuged and washed twice with distilled water. Finally, the precipitate was dried for 20 h at 85 °C. The dried material was crushed using a Gyro-mill and sieved with a 38 lm sieve into coarse and fine fractions. 2.1.2. Route 2 Solution A preparation. Calcium hydroxide (Ca(OH)2) and strontium hydroxide (Sr(OH)2) were dissolved in 500 ml of distilled water. Table 2 gives the weights used for each apatite composition. Solution B preparation. Orthophosphoric acid solution (85% H3PO4) was diluted in 500 ml of distilled water. The amount of H3PO4 used is given in Table 2.

Solution A–solution B mix. Solution B was added dropwise to the constantly stirred solution A over a period of 2 h. The pH of the mixture was adjusted to about 11 with a 30% ammonia solution, if needed. The precipitate formed was stirred for 1 h, before being left overnight. The precipitate was centrifuged and washed twice with distilled water. Finally, the precipitate was dried for 20 h at 85 °C. The dried material was powdered using a Gyro-mill and sieved with a 38 lm sieve. 2.2. XRD A Phillips powder diffractometer with a copper (Cu Ka) X-ray source (Philips PW 1700 series diffractometer, Philips, Eindhoven, the Netherlands) was used to characterize the glass samples. The powdered samples (<38 lm particle size) were scanned between 2h values of 10 and 80° with a step size of 2h = 0.04° in order to determine the crystal structure of each apatite. This step size results in a d-spac˚ at 20° 2h, consistent with ing deviation of around 0.009 A the errors reported in the refinement below. These parameters gave sufficient resolution for the refinement, with the most intense peaks occurring in the 2h = 30–35° region. XRD was then carried out and the results analysed using software containing a database of standard diffraction files. Rietveld refinement was performed with GSAS [10] and EXPGUI [11] software using a 10-term shifter-Chebyschev background function. Initial atomic coordinates and unit cell dimensions were taken from previously published data [6,7]. Parameters varied in the refinement were atomic coordinates (x, y and z), unit cell parameters (a and c), displacement parameters (Uiso), peak intensity scaling and peak profile parameters related to particle size and stress. For the mixed apatites, the fractional occupancies of both divalent cation sites were initially set at the stoichiometric values, then refined to total unity as there is some evidence that strontium preferentially substitutes onto the Ca(II) site [8]. To determine the Bragg peak widths, the Bragg reflection at around 2h = 25.9°, corresponding to the 0 0 2 hkl reflection, was Lorentzian deconvoluted with Microcal Origin software. 2.3. Raman The Raman system used consisted of a Renishaw RM 2000 spectrometer (Renishaw LC, UK) connected to a Leica microscope with an objective (50). A 785 nm 300 mW line focus laser with 100 mW power at the sample

Table 1 Experimental weights (in g) for route 1 synthesis for the series (SrxCa1x)5(PO4)3OH, where x = 0.00 and 1.00 Sr (x)

Ca(NO3)2  4H2O

wt.%

mol.%

Sr(NO3)2

wt.%

mol.%

(NH4)2HPO4

wt.%

mol.%

1.00 0.50 0.25 0.00

– 23.62 35.42 47.23

– 38.95 57.27 74.87

– 31.25 46.87 62.50

42.34 21.17 10.58 –

72.76 34.91 17.11 –

62.50 31.25 15.62 –

15.85 15.85 15.85 15.85

27.24 26.14 25.63 25.13

37.50 37.50 37.51 37.50

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

1457

Table 2 Experimental weights (in g) for route 2 synthesis for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00 Sr (x)

Ca(OH)2

wt.%

mol.%

Sr(OH)2

wt.%

mol.%

H3PO4

wt.%

mol.%

1.00 0.75 0.00

– 4.58 18.34

– 10.45 52.39

– 14.80 59.27

30.10 22.58 –

64.36 51.52 –

59.26 44.46 –

16.67 16.67 16.67

35.64 38.03 47.61

40.74 40.74 40.73

Overlapping peak positions were obtained by Lorentzian deconvolution using Microcal Origin software.

was used. This wavelength and 100% power was sufficient to produce a high signal to noise ratio Raman spectrum. The laser spot size was calculated to be 10 lm wide by 20 lm high and 40 lm deep. The spectrometer was set up with a spectrometer slit of 50 lm and 8 CCD (charge-coupled device) pixels. The powdered samples were placed on a glass slide to collect the spectra. Ten 10 s spectra were taken between Raman wave number shifts of 100 and 1500 cm1. Quartz was used as a calibration material with the main Raman resonance peak at around 521 cm1.

a

3. Results and discussion 3.1. XRD and Rietveld refinement Fig. 1a shows the XRD traces of the Sr-substituted apatites and Rietveld refinement. Traces were shifted on the y-axis for display purposes. The fitting was good for all

3500

Intensity (a.u.)

3000 2500 2000 1500 1000 500 0 10

20

30

40

50

2-theta / x=0.00

x=0.25

60

70

80

o

x=0.50

x=0.75

x=1.00

600 500

Intensity

400 300 200 100 0 -100 10

20

30

40

50

2-theta / Data

Fit

60

70

80

o

Background

Residual

Fig. 1. (a) XRD traces and Rietveld refinement of the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00: data points (XRD) and solid lines (Rietveld); (b) Rietveld refinement for x = 0.50, with residual and positions of Bragg reflections.

1458

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

samples, as demonstrated in the high R2 values (all >0.9, not shown) indicating excellent correlation. Fig. 1b shows a typical fit for the series for the x = 0.5 sample, also showing the residual and background. The diffraction peaks shift to lower 2h values with Sr-addition, indicating an increase in d-spacings and hence lattice parameters. This would be expected as Sr is slightly larger than Ca (118 and 100 pm for octahedral coordination [12]). Peak intensity also increases with Sr-addition. This is to be expected as Sr is heavier and contains more electrons than Ca, and will therefore more effectively scatter X-rays, so there is a general increase in crystallinity with Sr-addition (see below). Lattice parameters (a and c) increase linearly with Sraddition, as shown in Figs. 2 and 3, consistent with a larger ion entering the apatite lattice. As a result of this, the unit cell volume also increases linearly, as seen in Fig. 4. Density calculated from the Rietveld refinement in Fig. 5 also increases linearly with Sr-addition due to the replacement of a heavier ion for Ca in the crystal structure (87.62 and 40.078 g mol1, respectively [12]), and the Ca-rich end of the series agree well with previously published experimental data on pure Ca-hydroxyapatites made by aqueous gel casting [13] and sintering [14] and a low Sr-substitution Ca-hydroxyapatite [15]. Due to a lack of published data

on pure Sr-hydroxyapatite and mixed Ca–Sr-hydroxyapatite experimental densities, Fig. 5 shows the density of a Yb-doped Sr-fluorapatite [2]. As F and OH have approximately the same mass and ionic radii, the density of these materials should not differ greatly from Srhydroxyapatite. As Fig. 5 shows, there is good agreement between the published and Rietveld data. Crystal sizes, d, were calculated from the full width at half maximum values (FWHMs) of the most intense Bragg reflection at around 2h = 25.9°, corresponding to the 0 0 2 hkl reflection and using the formula equation shown below in Eq. (1) [16] d¼

0:9k w cos hx

ð1Þ

˚ ), w is where k is the wavelength of the X-rays (1.5406 A the FWHM of the Bragg peak and hx the angle of the Bragg reflection. The crystallite size decreased to a minimum at around x = 0.5 then increased with Sr content (Fig. 6). This is consistent with the observations of Li et al. [17], who showed, with a series of three Sr–Ca apatites (0.3, 1.5 and 15% Sr-substitution), that crystallite size decreases as Sr-substitution approaches 15%. This study did not report higher Sr-substitutions. It has been indicated that strontium addition to hydroxyapatite

Table 3 Parameters from Rietveld refinement of the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00, with standard deviations ˚ ˚ ˚ ˚ ˚3 ˚3 Sr Ca a/A 2ra / A c/A 2rc / A V/A 2rV / A q/g.cm3

d / nm

0.00 0.25 0.50 0.75 1.00

17.55 11.37 9.38 12.94 20.50

1.00 0.75 0.50 0.25 0.00

9.411 9.505 9.596 9.659 9.777

0.004 0.007 0.004 0.005 0.003

6.877 6.950 7.054 7.182 7.288

0.003 0.006 0.004 0.004 0.003

527.5 543.8 562.6 580.3 603.3

0.378 0.648 0.452 0.740 0.472

3.163 3.384 3.692 3.897 4.068

9.80 9.75

y = 0.35x + 9.41 2 R = 0.99

9.70

a /Å

9.65 9.60 9.55 9.50 9.45 9.40 0.0

0.2

0.4

0.6

0.8

1. 0

Fraction Sr Fig. 2. Variation in lattice parameter, a, for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00; error bars: +2r from Table 3.

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

1459

7.30

y = 0.42x + 6.86 2 R = 0.99

7.25 7.20

c /Å

7.15 7.10 7.05 7.00 6.95 6.90 6.85 0.0

0.2

0.4

0.6

0.8

1.0

Fraction Sr Fig. 3. Variation in lattice parameter, c, for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00; error bars: ±2r from Table 3.

610 y = 75.25x + 525.88 2 R = 1.00

600 590

V / Å3

580 570 560 550 540 530 520 0.0

0.2

0.4

0.6

0.8

1.0

Fraction Sr Fig. 4. Variation in unit cell volume for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00; error bars: ±2r from Table 3.

4.2 y = 0.93x + 3.18 2 R = 0.99

4.0

-3 ρ / g.cm

3.8 3.6 3.4 3.2 3.0 0.0

0.2

0.4

0.6

0.8

1.0

Fraction Sr Rietveld

Literature

Fig. 5. Variation in density for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00; error bars ±1%.

1460

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

22 2

y = 37.73x - 34.74x + 17.57 2 R = 1.00

20

d / nm

18 16 14 12 10 8 0.0

0.2

0.4

0.6

0.8

1.0

x Fig. 6. Variation in crystallite size from XRD Lorentzian peak widths for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00.

Table 4 Occupancies of M(I) and M(II) sites from Rietveld refinement

0.00

0.25

0.50

0.75

1.00

1.00 0.00 1.00 0.00 0.00

0.81 0.19 0.75 0.25 0.22

0.52 0.48 0.45 0.55 0.52

0.31 0.69 0.20 0.80 0.74

0.00 1.00 0.00 1.00 1.00

broadens the crystal size distribution, which contributes to the improved mechanical strength of bone [17]. When x = 0.5, the Sr–Ca apatite will be at its most disordered; therefore it would be expected that the crystallite size approaches a minimum. The Rietveld refinement gave a good fit with a slight preference for strontium occupancy on the Ca(II) site, as observed by Bigi et al. [8]. There is a 6 to 12% preference of Sr for the M(II) site, with increasing preference as the Sr content increases. The occupancies of the sites can be seen in Table 4 and is plotted graphically in Fig. 7a, and the correlation to the stoichiometric compositions agrees well, as seen in Fig. 7b. 3.2. Raman spectroscopy Figs. 8–10 show the Raman spectra of the apatites synthesized in this study, normalized to the v1 PO3 4 band at around 955 cm1 and shifted on the y-axis for presentation purposes. It can be seen from the Raman spectra in Fig. 8 that there is a general shift to a lower wave number of Raman bands with Sr-addition (see Table 5). The main Raman band at around 955 cm1 due to the v1 PO3 4 vibration decreases linearly with Sr-addition (see Figs. 8, 10 and

Occupancy - Rietveld

Ca(I) Sr(I) Ca(II) Sr(II) Mean Sr

x (Sr)

0.75

0.50

0.25

0.00 0.00

0.25

0.50

0.75

1.00

Sr content (x ) - stoichiometry Ca(I)

Sr(I)

Ca(II)

Sr(II)

1.00 y = 1.01x - 0.01 2

Sr content (x ) - Rietveld

Site

1.00

R = 1.00 0.75

0.50

0.25

0.00 0.00

0.25

0.50

0.75

1.00

Sr content (x ) - stoichiometry Fig. 7. (a) Occupancies of M(I) and M(II) sites and (b) mean Sr compositions from occupancies.

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

11). As the frequency of a Raman band is dependent on lattice vibrations, the masses of the atoms/ions present

1461

and the strength of the forces between the atoms/ions define the position of the vibration. The frequency of the

Wavelength / nm 810

820

830

840

850

860

Normalised intensity (a.u.)

3

x=1.00 2 x=0.75

x=0.50 1 x=0.25

x=0.00 0 400

600

800

1000

1200

-1

Raman shift / cm

Fig. 8. Raman spectra for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00, with 782 nm excitation.

Wavelength / nm 812

814

816

818

820

822

824

826

0.40

Normalised intensity (a.u.)

0.35 x=1.00

0.30

x=0.75

0.25

x=0.50 0.20

0.15 x=0.25 0.10 x=0.00 0.05 400

450

500

550

Raman shift / cm

600

650

-1

Fig. 9. Raman spectra forth series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00, with 782 nm excitation (400–650 cm1 region).

1462

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

Wavelength / nm 845

850

855

860

2.5

Normalised intensity (a.u.)

2.0 x=1.00 1.5 x=0.75 x=0.50

1.0

x=0.25 0.5 x=0.00 0.0 900

950

1000

1050

1100

1150

-1

Raman shift / cm

Fig. 10. Raman spectra for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00, with 782 nm excitation (900–1150 cm1 region).

Table 5 Raman band positions (in cm1) for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00 x

v2 PO3 4

v4 PO3 4

0.00 0.25 0.50 0.75 1.00

430, 426, 424, 425, 422,

580, 579, 582, 576, 573,

446 447 443 442 442

592, 589, 601 585, 581,

609 606 599 595

P–O–P

v1 PO3 4

v3 PO3 4

– – – – 714, 738

963 958 956 952 949

1031, 1022, 1034, 1027, 1023,

O–P–O

1048, 1048, 1051, 1044, 1051,

1070 1069 1064 1065 1058

– – – – 1051, 1058

Position of PO4 v1 band / cm-1

965

y = -13.16x + 961.98 2 R = 0.99 960

955

950

945 0.0

0.2

0.4

0.6

0.8

1.0

x Fig. 11. Change in the

v1 PO3 4

Raman band position for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00.

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

1463

FWHM of PO4 v1 band / cm-1

18

y = -0.94x + 26.34 2 R = 0.90

16 14 12 10 8 6 8

10

12

14

16

18

20

22

d / nm Fig. 12. Change in the FWHM of v1 PO3 4 Raman band plotted against crystallite size for the series (SrxCa1x)5(PO4)3OH, where x = 0.00, 0.25, 0.50, 0.75 and 1.00.

vibration, f, can be described classically by a harmonically oscillating ball and spring model, shown in Eq. (2), known as the Szigeti relationship [18] sffiffiffi 1 k ð2Þ f ¼ c l where c is the velocity of light in a vacuum (2.998108 m s1), k is the bond force constant (typically in N m1) and l is the reduced mass of the two bonding atoms (equal to m1m2/(m1 + m2)). As the bond strengths of Ca–O and Sr–O are similar (bond enthalpies of 402.1 and 425.5 kJ mol1, respectively [19]), k would be expected to be similar for both bonds. Therefore, as Sr is substituted for Ca, l will increase, shifting f, and hence the Raman shift to lower values as PO3 4 is associated with the heavier Sr cation. In practice, this shift could be used to determine the change in Sr in apatite compositions in future in vivo and in vitro studies, in implant materials for example. The following bands can be seen in all samples (Figs. 8– 10) and are tabulated in Table 5 [20]: two bands at around 430 cm1, attributed to v2 PO3 4 ; three bands around 590 cm1, attributed to v4 PO3 (except the x = 0.50 sam4 ple, where only two bands could be resolved); one band at around 955 cm1, attributed to v1 PO3 4 ; and three bands around 1050 cm1, attributed to v3 PO3 4 . In addition, two bands were seen at around 725 cm1 for the pure Srhydroxyapatite (x = 1.00). Unusually, for this sample the intensity of the bands in the region of the normally weak v3 PO3 4 vibrations were approximately half the strongest v1 PO3 4 band. This spectral region contains the carbonate vibration in apatite: 1100 cm1 for A-type carbonate (OH) and 1070 cm1 for the more common B-type carbonate (PO3 4 ) [20]. However, these vibrations are weak in biological apatites, which typically show carbonate contents of between 2 (dental enamel) and 8% (dentine and

bone) relative to the v1 phosphate peak, and any carbonate contamination from the reactants would be small. These features could be due to a small amount strontium pyrophosphate impurity (Sr2P2O7), as Raman bands have been previously reported in a series of phosphate glasses at around 775 and 1070 cm1, identified as P–O–P and O– P–O vibrations in P2 O4 7 dimmers, respectively [21], with the latter band higher in intensity, as seen here. This impurity is likely to be minor only, as one peak can be seen that may correspond to this phase at 2h = 33.04° in Fig. 1 for x = 1.00. This may be due to the 3 2 0 hkl reflection and is the third most intense peak in Sr2P2O7 (space group Pmna) [22]. The highest and second highest intensity peaks are likely to be hidden in the strong apatite peaks. Finally, the FWHM of the v1 PO3 band was plotted 4 against crystallite size from Table 3 and Fig. 6, and is shown in Fig. 12. There is a good correlation between crystal size and peak width, with linear broadening with decreasing size, regardless of the composition. Again, this could be used to estimate disorder in apatite samples. 4. Conclusions  A series of Sr-substituted hydroxyapatites were successfully synthesized by two routes under 90 °C with only the fully substituted material containing a small impurity, likely to be Sr2P2O7.  Lattice parameters, unit cell volume and density decreased linearly with strontium content, consistent with addition of a larger, heavier ion.  Crystallite size decreased with up to 25% substitution of Sr then increased with higher content.  There was a slight preference for Sr-occupancy of the M(II) site (6 to 12%) in agreement with previous studies.

1464

M.D. O’Donnell et al. / Acta Biomaterialia 4 (2008) 1455–1464

 The v1 PO3 Raman vibration at around 963 cm1 4 decreased linearly with Sr-addition, indicating that this could be used to accurately determine compositions of solid-solution compounds.  The FWHM of the v1 PO3 4 Raman band decreased linearly with crystallite size, which could also be used to determine disorder in chemically dissimilar apatite systems. Acknowledgements MDO would like to thank Dr. Molly Stevens, Dr. Gavin Jell and Mr. Robin Swain at IC for use of the Raman equipment and Dr. Stephen Skinner at IC for proofreading the manuscript. References [1] Heijligers HJM, Driessens FCM, Verbeeck RMH. Lattice-parameters and cation distribution of solid-solutions of calcium and strontium hydroxyapatite. Calcified Tissue Int 1979;29(2):127–31. [2] Deloach LD, Payne SA, Smith LK, Kway WL, Krupke WF. Laser and spectroscopic properties of Sr5(PO4)3F–Yb. J Opt Soc Am B: Opt Phys 1994;11(2):269–76. [3] Ma QY, Traina SJ, Logan TJ, Ryan JA. In situ lead immobilization by apatite. Environ Sci Technol 1993;27(9):1803–10. [4] Ohtsuki C, Kokubo T, Yamamuro T. Mechanism of apatite formation on CaO–SiO2P2O5 glasses in a simulated body-fluid. J Non-Cryst Solids 1992;143(1):84–92. [5] Meunier PJ, Roux C, Seeman E, Ortolani S, Badurski JE, Spector TD, et al. The effects of strontium ranelate on the risk of vertebral fracture in women with postmenopausal osteoporosis. N Engl J Med 2004;350(5):459–68. [6] Sudarsanan K, Young RA. Significant precision in crystal structural details: holly springs hydroxyapatite. Acta Crystallogr B 1969;25:1534–43.

[7] Sudarsanan K, Young RA. Structure of strontium hydroxide phosphate, Sr5(PO4)3OH. Acta Crystallogr B 1972;28:3668–70. [8] Bigi A, Falini G, Gazzano M, Roveri N, Tedesco E. Structural refinements of strontium substituted hydroxylapatites. Mater Sci Forum 1998;278(2):814–9. [9] Best SM, Bonfield W, Gibson IR, Jha LJ, Santos JDDS, inventors, Abonetics Limited, assignee. Silicon-substituted apatites and process for the preparation thereof. United States Patent No. 6312468; 2001. [10] Larson AC, Von Dreele RB. General structure analysis system (GSAS). Los Alamos National Laboratory; 1994. [11] Toby BH. EXPGUI, a graphical user interface for GSAS. J Appl Crystallogr 2001;34:210–3. [12] Greenwood NN, Earnshaw A. Chemistry of the elements. New York: Pergamon Press; 1984. [13] Chen B, Zhang Z, Zhang J, Dong M, Jiang D. Aqueous gel-casting of hydroxyapatite. Mater Sci Eng A 2006;435–436(5):198–203. [14] Gross KA, Rodrı´guez-Lorenzo LM. Sintered hydroxyfluorapatites. Part I. Sintering ability of precipitated solid solution powders. Biomaterials 2004;25(7–8):1375–84. [15] Kikuchi M, Yamazaki A, Otsuka R, Akao M, Aoki H. Crystal structure of Sr-substituted hydroxyapatite synthesized by hydrothermal method. J Solid State Chem 1994;113(2):373–8. [16] Cullity BD. Elements of X-ray diffraction. Reading, MA: AddisonWesley; 1959. [17] Li ZY, Lam WM, Yang C, Xu B, Ni GX, Abbah SA, et al. Chemical composition, crystal size and lattice structural changes after incorporation of strontium into biomimetic apatite. Biomaterials 2007;28(7):1452–60. [18] Parker JM, Seddon AB. Infrared-transmitting optical fibres. In: Cable M, Parker JM, editors. High-performance glasses. London: Blackie; 1992. p. 252–86. [19] Kerr JA. CRC handbook of chemistry and physics. 81st ed. Florida (FL): CRC Press; 2000. [20] Penel G, Leroy G, Rey C, Bres E. MicroRaman spectral study of the PO4 and CO3 vibrational modes in synthetic and biological apatites. Calcified Tissue Int 1998;63(6):475–81. [21] Santos LF, Almeida RM, Tikhomirov VK, Jha A. Raman spectra and structure of fluoro aluminophosphate glasses. J Non-Cryst Solids 2001;284(1–3):43–8. [22] Barbier J, Echard JP. A new refinement of alpha-Sr2P2O7. Acta Crystallogr C 1998;54:2.