Home
Search
Collections
Journals
About
Contact us
My IOPscience
Crystallized nano-sized alpha-tricalcium phosphate from amorphous calcium phosphate: microstructure, cementation and cell response
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Biomed. Mater. 10 025009 (http://iopscience.iop.org/1748-605X/10/2/025009) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 147.188.128.74 This content was downloaded on 21/04/2015 at 15:02
Please note that terms and conditions apply.
Biomed. Mater. 10 (2015) 025009
doi:10.1088/1748-6041/10/2/025009
Paper
received
5 November 2014 re vised
17 March 2015
Crystallized nano-sized alpha-tricalcium phosphate from amorphous calcium phosphate: microstructure, cementation and cell response
accep ted for publication
20 March 2015 published
17 April 2015
Linda Vecbiskena1, Karlis Agris Gross1, Una Riekstina2 and Thomas Chung-Kuang Yang3 1
Biomaterials Research Laboratory, Riga Technical University, 3/7 Paula Valdena St., LV-1048 Riga, Latvia Department of Pharmacology, University of Latvia, 4 Kronvalda Blvd., LV-1010 Riga, Latvia 3 Department of Chemical Engineering and Biotechnology, National Taipei University of Technology, Chung-Hsiao E Rd., Taipei 106, Taiwan 2
E-mail: [email protected] and [email protected] Keywords: amorphous calcium phosphate, alpha tricalcium phosphate, Rietveld refinement, cell response, bone cements, regenerative medicine
Abstract New insight on the conversion of amorphous calcium phosphate (ACP) to nano-sized alpha tricalcium phosphate (α-TCP) provides a faster pathway to calcium phosphate bone cements. In this work, synthesized ACP powders were treated with either water or ethanol, dried, crystallized between 700 and 800 °C, and then cooled at different cooling rates. Particle size was measured in a scanning electron microscope, but crystallite size calculated by Rietveld analysis. Phase composition and bonding in the crystallized powder was assessed by x-ray diffraction and Fourier-transform infrared spectroscopy. Results showed that 50 nm sized α-TCP formed after crystallization of lyophilized powders. Water treated ACP retained an unstable state that may allow ordering to nanoapatite, and further transition to β-TCP after crystallization and subsequent decomposition. Powders treated with ethanol, favoured the formation of pure α-TCP. Faster cooling limited the growth of β-TCP. Both the initial contact with water and the cooling rate after crystallization dictated β-TCP formation. Nano-sized α-TCP reacted faster with water to an apatite bone cement than conventionally prepared α-TCP. Water treated and freeze-dried powders showed faster apatite cement formation compared to ethanol treated powders. Good biocompatibility was found in pure α-TCP nanoparticles made from ethanol treatment and with a larger crystallite size. This is the first report of pure α-TCP nanoparticles with a reactivity that has not required additional milling to cause cementation.
1. Introduction Preparation of alpha tricalcium phosphate (α-TCP) for producing resorbable bone cements has required high temperature processing to achieve the alpha phase [1], followed by milling to impart solubility [2], but the chain of events is very time consuming. Pyrolysis [3, 4] provides another avenue, but has a low efficiency. This paves the way for revisiting the transition of amorphous calcium phosphate to alpha tricalcium phosphate [5, 6]. The difference lies in the processing speed, processing temperature, the resulting particle size and a greater reactivity; passage through the metastable amorphous phase state needs special attention to prevent accompanying phases. α-TCP is primarily produced in a pure form, but may also appear as a secondary phase (a β-TCP © 2015 IOP Publishing Ltd
polymorph). Comparing β-TCP (obtained at lower temperature) to α-TCP (formed at higher temperature) points to a larger range in properties, which extends the use of pure nano-sized α-TCP beyond the application in biomaterials. For instance, the hygroscopic nature aids adsorption of moisture to prevent powder clumping in foodstuffs such as salt [7]. It readily dissolves in aqueous solutions and precipitates a calcium deficient apatite (CDHA) to offer cementing action [8–10]. The extensive use of α-TCP would benefit from an alternative synthesis route to provide nano-sized, spherical particles with a narrow size distribution for more precisely tuning the properties. A pure nano-sized α-TCP has a preferred form with great bio-reactivity. To obtain a pure α-TCP, β-TCP will undergo phase transformation above 1100 °C after 2 h [11]. The alternative route for α-TCP involves heating
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
of blended powders: CaCO3 and CaHPO4 [3, 12–14], CaCO3 and NH4H2PO4 [15], or CaCO3 and Ca2P2O7 [15, 16]. However, the resulting grain size will be larger in these high temperature processes, which significantly reduces the bio-reactivity. Commercially available α-TCP is not always 100% pure; ≥75% from Sigma-Aldrich Co. (St.Louis, USA); >95% pure for powder containing 1 mm particles from InnoTERE (Dresden, Germany); ≥99% from Ensail Beijing Co. Ltd (Beijing, China) [8]. These powders are typically made by solid-state synthesis above 1250 °C to produce >95% pure α-TCP [14]. Heattreatment of a flame synthesized ACP produces ≥94% α-TCP [8]. Solution synthesis in this work will be investigated to achieve an α-TCP pure phase. The established methods for obtaining pure αTCP with the desired properties involves heating to very high temperatures followed by time-consuming ball-milling to increase the reactivity [17, 18]. Yet, there is less control over the particle size with this method, since milling influences the particle size. Milling also reduces the degree of order in the particles, either introducing defects or an amorphous phase. This drives scientists to develop alternative synthesis pathways to extend the functionality of biomaterials for regenerative medicine applications. Here we will demonstrate a simple synthesis approach—the preparation of an amorphous tricalcium phosphate from aqueous solutions—that makes a homogeneous α-TCP and provides scopes for more elaborative design of calcium phosphate bone cements. According to the literature, the amorphous calcium phosphate precursor may be synthesized by solid [19], sol–gel process [20, 21] or a rapidly quenched molten liquid [22, 23]. The sol–gel process is also initially amorphous, but it included an organic fraction that needs to be removed, unlike amorphous phase phosphate that is produced solely from inorganic salts, allowing easier purification during washing and immediate formation of a crystalline phase upon heating. The sol–gel process is versatile, but this method includes some disadvantages such as expensive raw materials and long processing time [20]. The feasibility for conversion of amorphous tricalcium phosphate to α-TCP was first reported by Eanes in 70s [5] and then addressed by Kanazawa et al 10 years later [19]. Recent work has produced a 95% pure α-TCP [23–25], showing a slight increase in purity from 94% purity after heating at 600 °C to 96% after heating at 800 °C [25]. Heating freeze-dried amorphous calcium phosphate (ACP) to higher temperatures rapidly decreases the α-TCP purity. Pure α-TCP nanoparticles have still not been produced. The challenge in the process is to overcome the metastability of the amorphous tricalcium phosphate. In contact with water, hydrolysis of phosphate converts the amorphous phase to a calcium deficient hydroxyapatite [26]. At room temperature the amorphous phase when formed is stable up to about 5 h, but lower temperatures can be chosen to extend the processing time if required. 2
Table 1. Synthesis conditions of ACP. Drying schedule (hours)
Synthesis
Solution volume (ml)
I
400
II
800
With ethanol
400
Freeze drying
Oven drying
48 4
Instant drying then stops the process of hydrolysis, allowing further phase transformations to be explored at higher temperatures. This study takes on the challenge, with attention to microstructural characteristics, to generate α-TCP by heating ACP above the crystallization temperature to 800 °C. The effect of cooling rate on the phase purity of α-TCP will be investigated. The production of pure nano-sized α-TCP will be shown for the first time, a proof-of-principle for cement formation illustrated, and the variation in stem cell response for powders heated at different temperatures shown.
2. Materials and methods 2.1. Synthesis of α-TCP Tricalcium phosphate was obtained from an ACP precursor by double decomposition [25] of calcium nitrate tetrahydrate (Ca(NO3)2 · 4H2O, Sigma-Aldrich, analytical grade) and ammonium phosphate dibasic ((NH4)2HPO4, Sigma-Aldrich, analytical grade) in a 26% ammonium hydroxide solution (NH 4OH, Sigma-Aldrich, analytical grade). A 0.30 M calcium nitrate solution made by dissolving Ca(NO3)2 · 4H2O in deionised water together with 30 mL ammonia solution was combined with a 0.24 M ammonium phosphate solution (synthesis pH was 10). The resulting precipitate was immediately filtered and rinsed with deionised water and one batch treated with ethanol (Enola, 96% v/v). Three different powders were produced, as shown in table 1. The precipitate was rinsed with water and either lyophilised, or treated with ethanol and oven-dried. The powder was then heated between 675 and 800 °C for 10 min on platinum foil or in an alumina boat (20 min at 775 °C) in a cylindrical tube furnace (Ceramic Engineering, Sydney, Australia). A longer heating time in the alumina boat (70 mm long, 10 mm wide, 10 mm high) compensated for the slower heating after direct placement in the furnace; due to the lower thermal conductivity and larger thermal mass of the alumina boat. Powders were immediately removed from the furnace and either emptied onto a metal block or allowed to cool in the ceramic boat to give a range of cooling regimes. 2.2. Characterization X-ray powder diffraction (XRD) has been shown to differentiate the α- and β-polymorphs of TCP [8]. An
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
x-ray diffractometer (D8 Advance, Bruker, Germany) with a Cu Kα x-ray source generated at 40 kV and 40 mA produced an x-ray diffraction pattern in the range of 20°–40° at a scan speed of 0.20° min−1. The diffractometer was equipped with a variable divergence slit: 0.05–2 mm. Phases were identified with reference to JCPDS 9–348 for α-TCP and JCPDS 9–169 for β-TCP. A quantitative phase analysis and the crystallite size were calculated with the use of the BGMN Rietveld program [27]. Fourier-transform infrared (FTIR) spectroscopy identified the chemical bonding. An FTIR spectrophotometer (Frontier, Perkin Elmer, USA) collected the spectrum at a 4 cm−1 resolution in the range of 4000– 450 cm−1. Powder was prepared for analysis in a KBr tablet–300 mg KBr and 5 mg powder. Thermal analysis (TG-DTA) was performed to investigate the thermal stability of amorphous precursors. A thermo gravimetric analysis (TG) equipped with differential thermal analysis (DTA) was operated by thermal analyser (EXTAR 6000 TG/DTA 6300, Seiko, Japan) in air atmosphere at a heating rate of 10 °C min−1 up to 1000 °C (10 mg of powdered sample in a alumina cylindrical crucible, diameter: 4 mm). Powders were investigated using high-magnification scanning electron microscopy (JSAM-6500F, JOEL, Japan) at an accelerating voltage of 15 kV. Samples were sputtered with gold for 20 s for increasing the surface conductivity in the scanning electron microscope. 2.3. Cementation The hydrolysis of α-TCP was assessed by the following test: (a) powder was homogenized with a mortar and pestle, (b) paste was formed by mixing 0.1 g powder with 0.5 ml of deionized water, (c) cementation time in a closed vessel at 37 °C was assessed from x-ray diffraction of powders immersed in liquid for 1, 4, 6, 8, 12 and 24 h, (d) mixing with 1 ml acetone [28] stopped the reaction followed by drying at 105 °C/24 h and (e) examination with x-ray diffraction to detect the αTCP transformation into calcium deficient apatite. X-ray diffraction was chosen to measure the progress of reaction as a clearer indicator of the reaction with reference to the apatite cement phase. Previous work by Ten Huisen et al with calorimetry showed a peak maximum at 5–30 min (7–8% conversation of α-TCP), but XRD detected calcium deficient apatite at 10 h and completion at 18 and 21 h [29]. 2.4. Cytotoxicity testing Powders were investigated in vitro with a lactate dehydrogenase (LDH) cytotoxicity assay (Promega, USA) to determine the mesenchymal stem cell response (table 2). Briefly, mesenchymal stem cells from the human dermis [30] were seeded on 96-well plates 3
Table 2. Purity of α-TCP after heating at 700 and 775 °C. Sample name
a
Phase content (wt%)a
Temperature (°C)
α-TCP
β-TCP
HA
II-775
775
91.3
8.7
—
Ethanol-700
700
98.0
—
2.0
Ethanol-775
775
99.5
—
0.5
Data calculated by Rietveld analysis.
(Sarstedt, Germany) at a concentration of 5 × 103 cells per well in 100 µl of DMEM/F12 medium (Invitrogen, USA) supplemented with penicillin and streptomycin (100 μ ml−1, 100 µg ml−1; Invitrogen, USA) and 10% of fetal calf-serum (Invitrogen, USA). After 24 h, α-TCP powder (concentration of 1 mg ml−1) was added to the cells and 2-fold serial dilutions were made. Cell culture without TCP powder served as an experimental control for spontaneous lactate dehydrogenase release. The cell culture medium was harvested after 72 h and cytotoxicity determined by LDH assay (Promega, USA). The enzymatic activity of released LDH was measured by formation of red formazan product, which correlates to the amount of lysed cells. Cytotoxicity was gauged by experimental LDH release (OD490)/maximum LDH release (OD490) according to the manufacturer’s instructions (Promega, USA). To confirm the formation of a calcium deficient apatite, powders were also immersed in the same solution, dried and then analysed by x-ray diffraction.
3. Results 3.1. Effect of synthesis conditions on α-TCP purity Amorphous calcium phosphate (ACP), with a Ca/P molar ratio close to 1.5, is a necessary precursor for αTCP or β-TCP [5, 8]. Figure 1(a) shows the typical x-ray powder diffraction (XRD) pattern of lyophilized ACP– a broad amorphous peak centred at 30° characteristic of phosphate groups in solution formed ACP [31], rapidly quenched melted calcium phosphate nanoparticles [4] or microparticles [6, 32] and calcium phosphate glasses [33]. Differentiation between the amorphous phase and nano-sized tricalcium phosphate is difficult [34], and requires other techniques such as thermal analysis to show the exothermic peak of crystallization. According to Somrani et al, amorphous calcium phosphate is stable up to 500 °C, but α-TCP is thermodynamically more favourable above 600 °C, as revealed by an exothermic peak at 630 °C [25]. The thermal evolution of lyophilized and ovendried ACP occurred in several steps; explained on the basis of thermochemical transformations (figure 2). A small weight loss from 100 to 150 °C, reflected a water loss from ACP (synthesis I and II—a 15% weight loss, synthesis with ethanol—a 17% weight loss). This effect is seen as a broad and asymmetrical endothermic peak in the DTA curve; synthesis I has a larger endothermic response due to a larger adsorbed water content. Up to
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
Intensity (counts)
Synthesis I
(a)
Synthesis II
Synthesis with ethanol
(c)
(b)
10
20
30
40
50
60
30
31
32
30
31
32
Figure 1. XRD pattern of an amorphous calcium phosphate (the same pattern was seen in ACP from all synthesis conditions); ACP heated at (b) 725 °C and (c) 775 °C (α-TCP peak intensities were 3663 and 4142 counts for synthesis with ethanol, respectively) showing less β-TCP at the higher temperature. Legend: α = α-TCP, β = β-TCP.
0 0
100
200
300
400
500
600
700
800
900
1000
-1 -2
Synthesis with ethanol
DTA (uV)
-3 Synthesis II -4 -5 -6 -7
Synthesis I
-8 -9
Figure 2. DTA curves of an amorphous calcium phosphate heated up to 1000 °C.
600 °C, the samples remained amorphous; the crystallization from amorphous phase to α-TCP occurred between 600 and 650 °C. The amorphous TCP crystallization temperature was not affected by the treatment of ethanol; the difference between weight losses was detected. Amorphous calcium phosphate after crystallization showed α-TCP and β-TCP. α-TCP exhibited an intense peak at approximately 30.7°, but β-TCP displayed weaker peaks at 31.0°, 25.8° and 27.7° (figures 1(b) and (c)). Heating at a higher temperature of 775 °C produced more α-TCP, as shown in a previous study [25], however the beta phase was still present. The FTIR spectra of heated powders confirmed α-TCP; β-TCP could not be readily identified with the more easily resolved PO34− peaks at 1048 cm−1, 970 cm−1 and 945 cm−1 [35]. The ACP spectrum demonstrated two phosphate bands at 1050 cm−1 and 570 cm−1, and an adsorbed water band at 3450 cm−1 (figure 3). The representative PO34− bands for α-TCP were present at 1050, 950 and 550 cm−1 [8, 35–37]. 4
3.2. Influence of cooling conditions on α-TCP purity Cooling is the second most important factor in the controlling of α-TCP purity. Figure 4 shows the data from faster cooling conditions of a heated lyophilized powder. A 2 mm powder layer on platinum foil removed from the furnace and cooled on a metal block provided the fastest cooling rate. Slower cooling was investigated by two different cooling regimes. An overall slower cooling rate occurred from a 9 mm high powder load in an alumina boat. The ceramic boat was transferred from the central hot region to the cooler outer zone and then either emptied onto the metal block for a final fast quench, or kept in the boat for slower cooling. The slower cooling lowered the α-TCP purity during withdrawal from the cylindrical furnace. Final cooling on a metal block—for a quick quench—produced 81 ± 2 wt% α-TCP (processing synthesis II), but cooling in the ceramic boat—for the slowest cooling rate— produced 75 ± 2 wt% α-TCP (processing synthesis II, figure 5). Where β-TCP was already present, a slower cooling rate significantly increased the β-TCP content.
Absorbance
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
-TCP Synthesis with ethanol
Synthesis II Synthesis I Lyophilised ACP 3500
2500 Wavelength (cm-1)
1500
500
Figure 3. FTIR spectra of an amorphous calcium phosphate heated at 775 °C showing the resulting α-TCP.
Synthesis I
Synthesis II
Synthesis with ethanol
750 Temperature (
)
-TCP content (wt%)*
100 80 60 40 20 0 675
700
725
775
800
825
Figure 4. α-TCP content (*calculated by Rietveld analysis) after heating amorphous calcium phosphate on Pt foil at different temperatures followed by rapid cooling on a metal block.
Carrodeguas also found less α-TCP in large batches cooled slowly from 1300 °C and so recommended treating smaller powder volumes for a higher cooling rate [8]. The cooling rate is equally important for retaining a high alpha phase content when cooling from 1300 °C or from lower temperature when starting with amorphous calcium phosphate. 3.3. Nanoparticle formation The crystallite size, calculated from Rietveld analysis, increased with processing temperature, in our case from 50 to 100 nm, and was comparable to the particle size viewed in SEM micrographs (figures 6 and 7). The particle size was also controlled by changing the reactant mixing conditions. A faster stirring rate for the smaller 5
batch (synthesis I) created smaller ACP clusters, column 1 in figure 7. It is noteworthy that smaller particles (from drying clusters) induced more β-TCP, supporting the notion of greater susceptibility for smaller spheres to absorb moisture, create apatite-like nanoregions that upon heat treatment formed β-TCP [14]. Producing a pure α-TCP—from ethanol treated ACP heated at 775 °C—allowed setting the crystallinity closer to the crystallization temperature. Such a capability was not possible for lyophilized powder since pure α-TCP could not be formed. 3.4. Cementation and cellular biocompatibility Hydrolysis speed depended on the processing history of the powders. In this work, hydrolysis of α-TCP into
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
Intensity (a.u.)
Powder cooled in an alumina boat
Synthesis with ethanol
Synthesis II 20
22
24
26
28
30
32
34
36
38
40
Figure 5. XRD patterns of powder heated at 775 °C and cooled in a ceramic boat. The drying process (freeze drying, or ethanol treatment) influences α-TCP formation with a higher purity from ethanol treatment.
Synthesis I
Synthesis II
Synthesis with ethanol
120 R2 = 0.97
Crystallite size (nm)*
110 100
R2 = 0.68
90 R2 = 0.76
80 70 60 50 40 675
700
725
750 Temperature (
775
800
825
)
Figure 6. Variation of crystallite size (*calculated by Rietveld analysis) of nano-sized α-TCP with increasing temperature.
CDHA in deionized water at 37 °C was completed in 6–8 h. Lyophilised α-TCP (containing 7–15% β-TCP) completed the reaction within 6 h (figure 8(a)), whereas pure α-TCP treated with ethanol and dried in the oven after 8 h (figure 8(b)). Faster cementation occurred for our processed powder than for the conventional powder (made at 1200 °C, ball-milled and then heated from 500 °C and 800 °C) that completed 90% of the reaction after 10 h (shown by heat released in isothermal calorimetry) [38]. It was important to understand if the cementation reaction and the potential release of small particles affect the cell response. Therefore, the cytotoxicity was determined after 72 h, reflecting a surface modified by the dissolution–precipitation process. The 6
cell adherence was not affected by the presence of the nanoparticles. Pure α-TCP, after hydrolysis in liquid medium (DMEM/F12), demonstrated the highest biocompatibility (75% surviving cells at concentration 0.5 mg ml−1) whereas the samples containing 7–15% β-TCP showed a lower biocompatibility (
Search
Collections
Journals
About
Contact us
My IOPscience
Crystallized nano-sized alpha-tricalcium phosphate from amorphous calcium phosphate: microstructure, cementation and cell response
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Biomed. Mater. 10 025009 (http://iopscience.iop.org/1748-605X/10/2/025009) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 147.188.128.74 This content was downloaded on 21/04/2015 at 15:02
Please note that terms and conditions apply.
Biomed. Mater. 10 (2015) 025009
doi:10.1088/1748-6041/10/2/025009
Paper
received
5 November 2014 re vised
17 March 2015
Crystallized nano-sized alpha-tricalcium phosphate from amorphous calcium phosphate: microstructure, cementation and cell response
accep ted for publication
20 March 2015 published
17 April 2015
Linda Vecbiskena1, Karlis Agris Gross1, Una Riekstina2 and Thomas Chung-Kuang Yang3 1
Biomaterials Research Laboratory, Riga Technical University, 3/7 Paula Valdena St., LV-1048 Riga, Latvia Department of Pharmacology, University of Latvia, 4 Kronvalda Blvd., LV-1010 Riga, Latvia 3 Department of Chemical Engineering and Biotechnology, National Taipei University of Technology, Chung-Hsiao E Rd., Taipei 106, Taiwan 2
E-mail: [email protected] and [email protected] Keywords: amorphous calcium phosphate, alpha tricalcium phosphate, Rietveld refinement, cell response, bone cements, regenerative medicine
Abstract New insight on the conversion of amorphous calcium phosphate (ACP) to nano-sized alpha tricalcium phosphate (α-TCP) provides a faster pathway to calcium phosphate bone cements. In this work, synthesized ACP powders were treated with either water or ethanol, dried, crystallized between 700 and 800 °C, and then cooled at different cooling rates. Particle size was measured in a scanning electron microscope, but crystallite size calculated by Rietveld analysis. Phase composition and bonding in the crystallized powder was assessed by x-ray diffraction and Fourier-transform infrared spectroscopy. Results showed that 50 nm sized α-TCP formed after crystallization of lyophilized powders. Water treated ACP retained an unstable state that may allow ordering to nanoapatite, and further transition to β-TCP after crystallization and subsequent decomposition. Powders treated with ethanol, favoured the formation of pure α-TCP. Faster cooling limited the growth of β-TCP. Both the initial contact with water and the cooling rate after crystallization dictated β-TCP formation. Nano-sized α-TCP reacted faster with water to an apatite bone cement than conventionally prepared α-TCP. Water treated and freeze-dried powders showed faster apatite cement formation compared to ethanol treated powders. Good biocompatibility was found in pure α-TCP nanoparticles made from ethanol treatment and with a larger crystallite size. This is the first report of pure α-TCP nanoparticles with a reactivity that has not required additional milling to cause cementation.
1. Introduction Preparation of alpha tricalcium phosphate (α-TCP) for producing resorbable bone cements has required high temperature processing to achieve the alpha phase [1], followed by milling to impart solubility [2], but the chain of events is very time consuming. Pyrolysis [3, 4] provides another avenue, but has a low efficiency. This paves the way for revisiting the transition of amorphous calcium phosphate to alpha tricalcium phosphate [5, 6]. The difference lies in the processing speed, processing temperature, the resulting particle size and a greater reactivity; passage through the metastable amorphous phase state needs special attention to prevent accompanying phases. α-TCP is primarily produced in a pure form, but may also appear as a secondary phase (a β-TCP © 2015 IOP Publishing Ltd
polymorph). Comparing β-TCP (obtained at lower temperature) to α-TCP (formed at higher temperature) points to a larger range in properties, which extends the use of pure nano-sized α-TCP beyond the application in biomaterials. For instance, the hygroscopic nature aids adsorption of moisture to prevent powder clumping in foodstuffs such as salt [7]. It readily dissolves in aqueous solutions and precipitates a calcium deficient apatite (CDHA) to offer cementing action [8–10]. The extensive use of α-TCP would benefit from an alternative synthesis route to provide nano-sized, spherical particles with a narrow size distribution for more precisely tuning the properties. A pure nano-sized α-TCP has a preferred form with great bio-reactivity. To obtain a pure α-TCP, β-TCP will undergo phase transformation above 1100 °C after 2 h [11]. The alternative route for α-TCP involves heating
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
of blended powders: CaCO3 and CaHPO4 [3, 12–14], CaCO3 and NH4H2PO4 [15], or CaCO3 and Ca2P2O7 [15, 16]. However, the resulting grain size will be larger in these high temperature processes, which significantly reduces the bio-reactivity. Commercially available α-TCP is not always 100% pure; ≥75% from Sigma-Aldrich Co. (St.Louis, USA); >95% pure for powder containing 1 mm particles from InnoTERE (Dresden, Germany); ≥99% from Ensail Beijing Co. Ltd (Beijing, China) [8]. These powders are typically made by solid-state synthesis above 1250 °C to produce >95% pure α-TCP [14]. Heattreatment of a flame synthesized ACP produces ≥94% α-TCP [8]. Solution synthesis in this work will be investigated to achieve an α-TCP pure phase. The established methods for obtaining pure αTCP with the desired properties involves heating to very high temperatures followed by time-consuming ball-milling to increase the reactivity [17, 18]. Yet, there is less control over the particle size with this method, since milling influences the particle size. Milling also reduces the degree of order in the particles, either introducing defects or an amorphous phase. This drives scientists to develop alternative synthesis pathways to extend the functionality of biomaterials for regenerative medicine applications. Here we will demonstrate a simple synthesis approach—the preparation of an amorphous tricalcium phosphate from aqueous solutions—that makes a homogeneous α-TCP and provides scopes for more elaborative design of calcium phosphate bone cements. According to the literature, the amorphous calcium phosphate precursor may be synthesized by solid [19], sol–gel process [20, 21] or a rapidly quenched molten liquid [22, 23]. The sol–gel process is also initially amorphous, but it included an organic fraction that needs to be removed, unlike amorphous phase phosphate that is produced solely from inorganic salts, allowing easier purification during washing and immediate formation of a crystalline phase upon heating. The sol–gel process is versatile, but this method includes some disadvantages such as expensive raw materials and long processing time [20]. The feasibility for conversion of amorphous tricalcium phosphate to α-TCP was first reported by Eanes in 70s [5] and then addressed by Kanazawa et al 10 years later [19]. Recent work has produced a 95% pure α-TCP [23–25], showing a slight increase in purity from 94% purity after heating at 600 °C to 96% after heating at 800 °C [25]. Heating freeze-dried amorphous calcium phosphate (ACP) to higher temperatures rapidly decreases the α-TCP purity. Pure α-TCP nanoparticles have still not been produced. The challenge in the process is to overcome the metastability of the amorphous tricalcium phosphate. In contact with water, hydrolysis of phosphate converts the amorphous phase to a calcium deficient hydroxyapatite [26]. At room temperature the amorphous phase when formed is stable up to about 5 h, but lower temperatures can be chosen to extend the processing time if required. 2
Table 1. Synthesis conditions of ACP. Drying schedule (hours)
Synthesis
Solution volume (ml)
I
400
II
800
With ethanol
400
Freeze drying
Oven drying
48 4
Instant drying then stops the process of hydrolysis, allowing further phase transformations to be explored at higher temperatures. This study takes on the challenge, with attention to microstructural characteristics, to generate α-TCP by heating ACP above the crystallization temperature to 800 °C. The effect of cooling rate on the phase purity of α-TCP will be investigated. The production of pure nano-sized α-TCP will be shown for the first time, a proof-of-principle for cement formation illustrated, and the variation in stem cell response for powders heated at different temperatures shown.
2. Materials and methods 2.1. Synthesis of α-TCP Tricalcium phosphate was obtained from an ACP precursor by double decomposition [25] of calcium nitrate tetrahydrate (Ca(NO3)2 · 4H2O, Sigma-Aldrich, analytical grade) and ammonium phosphate dibasic ((NH4)2HPO4, Sigma-Aldrich, analytical grade) in a 26% ammonium hydroxide solution (NH 4OH, Sigma-Aldrich, analytical grade). A 0.30 M calcium nitrate solution made by dissolving Ca(NO3)2 · 4H2O in deionised water together with 30 mL ammonia solution was combined with a 0.24 M ammonium phosphate solution (synthesis pH was 10). The resulting precipitate was immediately filtered and rinsed with deionised water and one batch treated with ethanol (Enola, 96% v/v). Three different powders were produced, as shown in table 1. The precipitate was rinsed with water and either lyophilised, or treated with ethanol and oven-dried. The powder was then heated between 675 and 800 °C for 10 min on platinum foil or in an alumina boat (20 min at 775 °C) in a cylindrical tube furnace (Ceramic Engineering, Sydney, Australia). A longer heating time in the alumina boat (70 mm long, 10 mm wide, 10 mm high) compensated for the slower heating after direct placement in the furnace; due to the lower thermal conductivity and larger thermal mass of the alumina boat. Powders were immediately removed from the furnace and either emptied onto a metal block or allowed to cool in the ceramic boat to give a range of cooling regimes. 2.2. Characterization X-ray powder diffraction (XRD) has been shown to differentiate the α- and β-polymorphs of TCP [8]. An
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
x-ray diffractometer (D8 Advance, Bruker, Germany) with a Cu Kα x-ray source generated at 40 kV and 40 mA produced an x-ray diffraction pattern in the range of 20°–40° at a scan speed of 0.20° min−1. The diffractometer was equipped with a variable divergence slit: 0.05–2 mm. Phases were identified with reference to JCPDS 9–348 for α-TCP and JCPDS 9–169 for β-TCP. A quantitative phase analysis and the crystallite size were calculated with the use of the BGMN Rietveld program [27]. Fourier-transform infrared (FTIR) spectroscopy identified the chemical bonding. An FTIR spectrophotometer (Frontier, Perkin Elmer, USA) collected the spectrum at a 4 cm−1 resolution in the range of 4000– 450 cm−1. Powder was prepared for analysis in a KBr tablet–300 mg KBr and 5 mg powder. Thermal analysis (TG-DTA) was performed to investigate the thermal stability of amorphous precursors. A thermo gravimetric analysis (TG) equipped with differential thermal analysis (DTA) was operated by thermal analyser (EXTAR 6000 TG/DTA 6300, Seiko, Japan) in air atmosphere at a heating rate of 10 °C min−1 up to 1000 °C (10 mg of powdered sample in a alumina cylindrical crucible, diameter: 4 mm). Powders were investigated using high-magnification scanning electron microscopy (JSAM-6500F, JOEL, Japan) at an accelerating voltage of 15 kV. Samples were sputtered with gold for 20 s for increasing the surface conductivity in the scanning electron microscope. 2.3. Cementation The hydrolysis of α-TCP was assessed by the following test: (a) powder was homogenized with a mortar and pestle, (b) paste was formed by mixing 0.1 g powder with 0.5 ml of deionized water, (c) cementation time in a closed vessel at 37 °C was assessed from x-ray diffraction of powders immersed in liquid for 1, 4, 6, 8, 12 and 24 h, (d) mixing with 1 ml acetone [28] stopped the reaction followed by drying at 105 °C/24 h and (e) examination with x-ray diffraction to detect the αTCP transformation into calcium deficient apatite. X-ray diffraction was chosen to measure the progress of reaction as a clearer indicator of the reaction with reference to the apatite cement phase. Previous work by Ten Huisen et al with calorimetry showed a peak maximum at 5–30 min (7–8% conversation of α-TCP), but XRD detected calcium deficient apatite at 10 h and completion at 18 and 21 h [29]. 2.4. Cytotoxicity testing Powders were investigated in vitro with a lactate dehydrogenase (LDH) cytotoxicity assay (Promega, USA) to determine the mesenchymal stem cell response (table 2). Briefly, mesenchymal stem cells from the human dermis [30] were seeded on 96-well plates 3
Table 2. Purity of α-TCP after heating at 700 and 775 °C. Sample name
a
Phase content (wt%)a
Temperature (°C)
α-TCP
β-TCP
HA
II-775
775
91.3
8.7
—
Ethanol-700
700
98.0
—
2.0
Ethanol-775
775
99.5
—
0.5
Data calculated by Rietveld analysis.
(Sarstedt, Germany) at a concentration of 5 × 103 cells per well in 100 µl of DMEM/F12 medium (Invitrogen, USA) supplemented with penicillin and streptomycin (100 μ ml−1, 100 µg ml−1; Invitrogen, USA) and 10% of fetal calf-serum (Invitrogen, USA). After 24 h, α-TCP powder (concentration of 1 mg ml−1) was added to the cells and 2-fold serial dilutions were made. Cell culture without TCP powder served as an experimental control for spontaneous lactate dehydrogenase release. The cell culture medium was harvested after 72 h and cytotoxicity determined by LDH assay (Promega, USA). The enzymatic activity of released LDH was measured by formation of red formazan product, which correlates to the amount of lysed cells. Cytotoxicity was gauged by experimental LDH release (OD490)/maximum LDH release (OD490) according to the manufacturer’s instructions (Promega, USA). To confirm the formation of a calcium deficient apatite, powders were also immersed in the same solution, dried and then analysed by x-ray diffraction.
3. Results 3.1. Effect of synthesis conditions on α-TCP purity Amorphous calcium phosphate (ACP), with a Ca/P molar ratio close to 1.5, is a necessary precursor for αTCP or β-TCP [5, 8]. Figure 1(a) shows the typical x-ray powder diffraction (XRD) pattern of lyophilized ACP– a broad amorphous peak centred at 30° characteristic of phosphate groups in solution formed ACP [31], rapidly quenched melted calcium phosphate nanoparticles [4] or microparticles [6, 32] and calcium phosphate glasses [33]. Differentiation between the amorphous phase and nano-sized tricalcium phosphate is difficult [34], and requires other techniques such as thermal analysis to show the exothermic peak of crystallization. According to Somrani et al, amorphous calcium phosphate is stable up to 500 °C, but α-TCP is thermodynamically more favourable above 600 °C, as revealed by an exothermic peak at 630 °C [25]. The thermal evolution of lyophilized and ovendried ACP occurred in several steps; explained on the basis of thermochemical transformations (figure 2). A small weight loss from 100 to 150 °C, reflected a water loss from ACP (synthesis I and II—a 15% weight loss, synthesis with ethanol—a 17% weight loss). This effect is seen as a broad and asymmetrical endothermic peak in the DTA curve; synthesis I has a larger endothermic response due to a larger adsorbed water content. Up to
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
Intensity (counts)
Synthesis I
(a)
Synthesis II
Synthesis with ethanol
(c)
(b)
10
20
30
40
50
60
30
31
32
30
31
32
Figure 1. XRD pattern of an amorphous calcium phosphate (the same pattern was seen in ACP from all synthesis conditions); ACP heated at (b) 725 °C and (c) 775 °C (α-TCP peak intensities were 3663 and 4142 counts for synthesis with ethanol, respectively) showing less β-TCP at the higher temperature. Legend: α = α-TCP, β = β-TCP.
0 0
100
200
300
400
500
600
700
800
900
1000
-1 -2
Synthesis with ethanol
DTA (uV)
-3 Synthesis II -4 -5 -6 -7
Synthesis I
-8 -9
Figure 2. DTA curves of an amorphous calcium phosphate heated up to 1000 °C.
600 °C, the samples remained amorphous; the crystallization from amorphous phase to α-TCP occurred between 600 and 650 °C. The amorphous TCP crystallization temperature was not affected by the treatment of ethanol; the difference between weight losses was detected. Amorphous calcium phosphate after crystallization showed α-TCP and β-TCP. α-TCP exhibited an intense peak at approximately 30.7°, but β-TCP displayed weaker peaks at 31.0°, 25.8° and 27.7° (figures 1(b) and (c)). Heating at a higher temperature of 775 °C produced more α-TCP, as shown in a previous study [25], however the beta phase was still present. The FTIR spectra of heated powders confirmed α-TCP; β-TCP could not be readily identified with the more easily resolved PO34− peaks at 1048 cm−1, 970 cm−1 and 945 cm−1 [35]. The ACP spectrum demonstrated two phosphate bands at 1050 cm−1 and 570 cm−1, and an adsorbed water band at 3450 cm−1 (figure 3). The representative PO34− bands for α-TCP were present at 1050, 950 and 550 cm−1 [8, 35–37]. 4
3.2. Influence of cooling conditions on α-TCP purity Cooling is the second most important factor in the controlling of α-TCP purity. Figure 4 shows the data from faster cooling conditions of a heated lyophilized powder. A 2 mm powder layer on platinum foil removed from the furnace and cooled on a metal block provided the fastest cooling rate. Slower cooling was investigated by two different cooling regimes. An overall slower cooling rate occurred from a 9 mm high powder load in an alumina boat. The ceramic boat was transferred from the central hot region to the cooler outer zone and then either emptied onto the metal block for a final fast quench, or kept in the boat for slower cooling. The slower cooling lowered the α-TCP purity during withdrawal from the cylindrical furnace. Final cooling on a metal block—for a quick quench—produced 81 ± 2 wt% α-TCP (processing synthesis II), but cooling in the ceramic boat—for the slowest cooling rate— produced 75 ± 2 wt% α-TCP (processing synthesis II, figure 5). Where β-TCP was already present, a slower cooling rate significantly increased the β-TCP content.
Absorbance
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
-TCP Synthesis with ethanol
Synthesis II Synthesis I Lyophilised ACP 3500
2500 Wavelength (cm-1)
1500
500
Figure 3. FTIR spectra of an amorphous calcium phosphate heated at 775 °C showing the resulting α-TCP.
Synthesis I
Synthesis II
Synthesis with ethanol
750 Temperature (
)
-TCP content (wt%)*
100 80 60 40 20 0 675
700
725
775
800
825
Figure 4. α-TCP content (*calculated by Rietveld analysis) after heating amorphous calcium phosphate on Pt foil at different temperatures followed by rapid cooling on a metal block.
Carrodeguas also found less α-TCP in large batches cooled slowly from 1300 °C and so recommended treating smaller powder volumes for a higher cooling rate [8]. The cooling rate is equally important for retaining a high alpha phase content when cooling from 1300 °C or from lower temperature when starting with amorphous calcium phosphate. 3.3. Nanoparticle formation The crystallite size, calculated from Rietveld analysis, increased with processing temperature, in our case from 50 to 100 nm, and was comparable to the particle size viewed in SEM micrographs (figures 6 and 7). The particle size was also controlled by changing the reactant mixing conditions. A faster stirring rate for the smaller 5
batch (synthesis I) created smaller ACP clusters, column 1 in figure 7. It is noteworthy that smaller particles (from drying clusters) induced more β-TCP, supporting the notion of greater susceptibility for smaller spheres to absorb moisture, create apatite-like nanoregions that upon heat treatment formed β-TCP [14]. Producing a pure α-TCP—from ethanol treated ACP heated at 775 °C—allowed setting the crystallinity closer to the crystallization temperature. Such a capability was not possible for lyophilized powder since pure α-TCP could not be formed. 3.4. Cementation and cellular biocompatibility Hydrolysis speed depended on the processing history of the powders. In this work, hydrolysis of α-TCP into
Biomed. Mater. 10 (2015) 025009
L Vecbiskena et al
Intensity (a.u.)
Powder cooled in an alumina boat
Synthesis with ethanol
Synthesis II 20
22
24
26
28
30
32
34
36
38
40
Figure 5. XRD patterns of powder heated at 775 °C and cooled in a ceramic boat. The drying process (freeze drying, or ethanol treatment) influences α-TCP formation with a higher purity from ethanol treatment.
Synthesis I
Synthesis II
Synthesis with ethanol
120 R2 = 0.97
Crystallite size (nm)*
110 100
R2 = 0.68
90 R2 = 0.76
80 70 60 50 40 675
700
725
750 Temperature (
775
800
825
)
Figure 6. Variation of crystallite size (*calculated by Rietveld analysis) of nano-sized α-TCP with increasing temperature.
CDHA in deionized water at 37 °C was completed in 6–8 h. Lyophilised α-TCP (containing 7–15% β-TCP) completed the reaction within 6 h (figure 8(a)), whereas pure α-TCP treated with ethanol and dried in the oven after 8 h (figure 8(b)). Faster cementation occurred for our processed powder than for the conventional powder (made at 1200 °C, ball-milled and then heated from 500 °C and 800 °C) that completed 90% of the reaction after 10 h (shown by heat released in isothermal calorimetry) [38]. It was important to understand if the cementation reaction and the potential release of small particles affect the cell response. Therefore, the cytotoxicity was determined after 72 h, reflecting a surface modified by the dissolution–precipitation process. The 6
cell adherence was not affected by the presence of the nanoparticles. Pure α-TCP, after hydrolysis in liquid medium (DMEM/F12), demonstrated the highest biocompatibility (75% surviving cells at concentration 0.5 mg ml−1) whereas the samples containing 7–15% β-TCP showed a lower biocompatibility (
