Materials Science and Engineering C 40 (2014) 121–126

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Importance of nucleation in transformation of octacalcium phosphate to hydroxyapatite Natsuko Ito a, Masanobu Kamitakahara a,⁎, Masahiro Yoshimura b, Koji Ioku c a b c

Graduate School of Environmental Studies, Tohoku University, 6-6-20, Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan Department of Materials Science and Engineering, National Cheng Kung University, No. 1, University Road, Tainan 70101, Taiwan Faculty of Economics, Keio University, Building #2-101B, 4-1-1, Hiyoshi, Kohoku-ku, Yokohama 223-8521, Japan

a r t i c l e

i n f o

Article history: Received 25 July 2013 Received in revised form 3 March 2014 Accepted 18 March 2014 Available online 25 March 2014 Keywords: Octacalcium phosphate Hydroxyapatite Transformation

a b s t r a c t Octacalcium phosphate (OCP) is regarded as an in vivo precursor of hydroxyapatite (HA). It is important to understand the mechanism of transformation of OCP to HA in order to reveal the mechanism of mineralization and help in the development of artificial bone-repairing materials. Herein, we have examined the behavior of OCP in a simulated body fluid (SBF) and pure water. The OCP particles immersed in the SBF at 37 °C did not transform to HA even after 720 h of immersion, though the particles showed crystal growth. In distilled water at 60 °C, the OCP particles transformed to HA but the unreactive period was observed. Although the immersed solution became supersaturated with HA within 12 h of immersion, the OCP was not transformed in the first 36 h of immersion. These results indicate that the nucleation of HA is the rate-determining step in the transformation of OCP to HA. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Hydroxyapatite (Ca10(PO4)6(OH)2, HA) is the main inorganic constituent of human bones and teeth. Considering the treatment of diseases and injuries in bone, it is important to understand the mechanism of calcification. Additionally, nowadays calcium phosphates are widely used to repair the bone defects [1–3] and it is also important to understand the behavior of calcium phosphates. Among the calcium phosphates, octacalcium phosphate (Ca8(HPO4)2(PO4)4·5H2O, OCP) is paid attention in the present study because it is regarded as an in vivo precursor of HA [4,5] and has become an important candidate for use as a biomaterial for bone repair [6–9]. OCP was found to be converted to HA and support bone regeneration in vivo [6,10]. Therefore, elucidation of the mechanism by which OCP is converted to HA would provide extremely useful information for the design and development of materials for bone repair. The transformation of OCP to HA can be also utilized for the material process [11]. By using this transformation, HA nanorods [12] and plate-like HA particles [13] can be synthesized. There are several papers about the transformation of OCP to HA [4, 14–24]. OCP has a layered structure composed of an apatitic layer, with a structure similar to that of HA, in addition to a hydrated layer. The atomic arrangements of the apatitic layers of OCP and HA are ex-

⁎ Corresponding author. Tel./fax: +81 22 795 7375. E-mail address: [email protected] (M. Kamitakahara).

http://dx.doi.org/10.1016/j.msec.2014.03.034 0928-4931/© 2014 Elsevier B.V. All rights reserved.

tremely alike [14]. It is suggested that the HA crystals can grow epitaxially on the OCP crystals because of this structural similarity [4,15,16]. LeGeros et al. suggested that the transformation of OCP to HA occurs by the dissolution of OCP and the subsequent precipitation of HA [17]. Iijima et al. suggested that the transformation of OCP to HA proceeds via in situ reorganization of lattice ions and/or apatitic clusters [18]. Recently, transmission electron microscope (TEM) observation of in situ transformation by electron beam of TEM [19] and nuclear magnetic resonance (NMR) measurement of the transformation [20] were also reported. However, they mainly focused on the crystal growth and discussed the relationship between the OCP structures and the formed HA, and did not focus on the initial transformation reaction, i.e. nucleation of HA. We previously examined the transformation of OCP in distilled water and revealed the existence of the unreactive period in the first stage [21]. However, the initial transformation reaction is still unclear. Moreover, the behavior of OCP under physiological conditions is still unclear. Ban et al. [22] and Yokoi et al. [25] reported that OCP did not transform to HA in a simulated body fluid (SBF), while several researchers [10,23,24] reported the in vitro transformation OCP into HA under physiological conditions. These different behaviors may also be explained by revealing the transformation behaviors from the view point of the nucleation. In the present study, we reveal the importance of the nucleation in the transformation of OCP to HA by immersing OCP in SBF and in pure water. The changes in both the samples and the solutions during the immersion were examined in detail, and the transformation mechanism was discussed from the view point of nucleation.

122

N. Ito et al. / Materials Science and Engineering C 40 (2014) 121–126

Intensity / a. u.

OCP

After immersion

Before immersion 10

20

30

40

50

60

2θ / ° (CuKα) Fig. 1. XRD patterns of sample before and after immersion in SBF at 37 °C for 720 h.

2. Materials and methods The OCP powder was synthesized from calcium carbonate and phosphoric acid according to the previously reported method [26,27]. A 20 mmol of phosphoric acid (85%, Wako Pure Chemical Industries, Ltd., Japan) was added to 200 cm3 of distilled water with stirring. The resultant homogeneous solution was heated to 60 °C, and then 26.6 mmol of calcium carbonate (Wako Pure Chemical Industries, Ltd., Japan) was added to the solution with stirring. The resulting suspension was stirred at 60 °C for 6 h. The product was collected using suction filtration, washed with distilled water, and dried at 37 °C for 24 h. The obtained OCP was immersed in an SBF prepared according to reports by Kokubo et al. [28,29], with ion concentrations similar to those of the human blood plasma. The SBF had a pH value of 7.40, and ion concentrations as follows (in mmol⋅dm−3): Na+ = 142.0; K+ = 2− = 5.0; Ca2 + = 2.5; Mg2 + = 1.5; Cl− = 147.8; HCO− 3 = 4.2; HPO4 1.0; and SO24 − = 0.5. A 15 mg sample of the OCP powder was placed in a polypropylene tube with 1.5 cm3 of the SBF. The tubes were kept in an incubator at 37 °C for 720 h. The SBF was replaced every day in order to simulate the fluid circulation within the body. From the preliminary experiment, the decrease in calcium and phosphorus concentrations in the SBF was observed when OCP was immersed in the SBF. As the changes in calcium and phosphorus concentrations in the SBF should affect the reaction of OCP due to the changes in degree of supersaturation, the renewal of the SBF was conducted during the SBF

Before

immersion test. After the immersion, the powders and solutions were separated by centrifugation. The powders were collected and washed with distilled water, and then dried at 37 °C for 24 h. In the next experiment, a 0.10 g sample of fresh OCP powder was placed in a polypropylene tube with 10 cm3 of distilled water. The tubes were then kept in an incubator at 60 °C for 240 h without exchanging the solution. As the phase change from OCP to HA was not observed in distilled water at 37 °C within 240 h in the preliminary experiment, the temperature was increased to 60 °C to accelerate the transformation reaction. After the immersion, the powders and solutions were separated by centrifugation. The powders were collected and washed with distilled water, and then dried at 37 °C for 24 h. The supernatant solutions were moved to fresh tubes for the subsequent analysis. The samples were characterized using X-ray diffractometry (XRD; RINT-2200VL, Rigaku, Japan) with CuKα radiation. Diffraction lines derived from OCP and/or HA were detected in all samples, with no other lines evident. The OCP content in the immersed powders was quantitatively determined using an internal standard method. As the standard, Si powder (NIST, USA) was mixed with the immersed samples, and the OCP content was quantified using a calibration curve prepared by plotting the ratio of OCP and Si integrated intensities against the OCP content in the known mixtures of OCP, HA, and Si. The 100 diffraction line (2θ = 4.7°) of OCP and the 111 diffraction line (2θ = 28.4°) of Si were used for the preparation of the calibration curve. The morphology of the immersed samples was observed using scanning electron microscopy (SEM; SU8000, Hitachi, Japan) and transmission electron microscopy (TEM; HF-2000, Hitachi, Japan). The particle sizes of the samples before and after immersion in the SBF for 720 h were measured from the obtained SEM images. The pH value and calcium and phosphorus concentrations of the solutions after the immersion were examined using a standard pH electrode and inductively coupled plasma atomic emission spectroscopy (ICP-AES; ICAP6000, Thermo Fisher Scientific, USA), respectively. Using the values for the pH and ion concentrations, the degree of supersaturation of the solution was calculated according to a previously reported method [18,30,31]. The degrees of supersaturation of OCP (S(OCP)) and HA (S(HA)) were calculated as shown in Eqs. (1) and (2), respectively.

SðOCPÞ ¼

SðHAÞ ¼

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi u 4  2þ   þ  3  3−  u a ca a H a PO 4 t 8

ð1Þ

Ksp ðOCPÞ

ffi vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u 5  2þ  3  3−   −  u a ca PO a a OH 4 t 9

ð2Þ

Ksp ðHAÞ

The activity coefficient was obtained using the Debye–Hückel equa+ − tion, and the activities (a(Ca2+), a(PO3− 4 ), a(H ) and a(OH )) were

After

4 µm Fig. 2. SEM images of sample before and after immersion in SBF at 37 °C for 720 h.

4 µm

N. Ito et al. / Materials Science and Engineering C 40 (2014) 121–126

123

0.01

Concentration / mol dm-3

216 h

Intensity / a. u.

72h

P

0.008

Ca 0.006 0.004 0.002 0

0

30

60

90

OCP HA

120

150

180

210

240

Time / h Fig. 5. Temporal change in calcium and phosphorus concentrations in solution in which sample was immersed at 60 °C.

36 h 10

20

30

40

50

60

2θ / ° (CuKα) Fig. 3. XRD patterns of sample after immersion in distilled water at 60 °C for 36 h, 72 h and 216 h.

obtained from the product of the activity coefficient and the concentrations of each ion. The ion concentrations were determined as follows: − + [Ca2+] = CCa − [CaH2PO+ 4 ] − [CaHPO4] − [CaPO4 ] − [CaOH ]

+ − [PO3− 4 ] = CP − [CaH2PO4 ] − [CaHPO4] − [CaPO4 ] − [H3PO4] − 2− − [H2PO4 ] − [HPO4 ]

Elemental concentrations of calcium and phosphorous (CCa and CP) were obtained using ICP, and the ion concentrations were determined using the association/dissociation constants [30,31]. 3. Results and discussion Fig. 1 shows the XRD patterns of the sample powder before and after immersion in the SBF at 37 °C for 720 h. No change was detected in the

XRD pattern after the immersion, with the crystal phase of the sample corresponding to that of pure OCP. Fig. 2 shows the SEM images of the sample before and after the immersion in SBF. Although the particles retained their plate-like morphology, their size increased in the direction of the long axis. This indicates that the crystal growth of OCP occurred in the long-axis direction in the SBF. The OCP particles immersed in the SBF were not transformed to HA; instead, they showed crystal growth. Our results are consistent with those reported by Ban et al. [22] who reported that OCP did not transform to HA in the SBF and Yokoi et al. [25] who reported that OCP did not transform to HA and OCP formation occurred on OCP crystals in the SBF. Although HA is the more stable crystal phase compared to that of OCP in the SBF, the crystal growth of OCP occurred preferentially. These phenomena can be theoretically explained by Lu et al. [32]. They conducted a theoretical analysis of calcium phosphate precipitation in the SBF, revealing that the nucleation rate of OCP is substantially higher than that of HA, while HA is thermodynamically more stable than OCP in the SBF. Because of the kinetically slow HA nucleation, the crystal growth of OCP was speculated to occur preferentially. Moreover, Iijima et al. [33] reported that OCP can grow without transforming into HA as long as the driving force to precipitate OCP is larger than that to form HA, and that a continuous Ca2+ supply and a static condition are favorable to OCP precipitation. As the SBF was replaced everyday and a static condition was used in the present system, the driving force for crystal growth of OCP was speculated to be remained. Under these conditions, the crystal growth of OCP is speculated to occur preferentially in the SBF.

100

8

80

7

pH

OCP content / %

9

60

6

40

5

20

4

0

0

30

60

90

120

150

180

210

240

Time / h Fig. 4. Temporal change in OCP content of sample immersed in distilled water at 60 °C.

3

0

30

60

90

120

150

180

210

240

Time / h Fig. 6. Temporal change in pH of solution in which sample was immersed at 60 °C.

124

N. Ito et al. / Materials Science and Engineering C 40 (2014) 121–126

Degree of supersaturation

15

S(HA) S(OCP)

10

5

1 0 0

30

60

90

120

150

180

210

240

Time / h Fig. 7. Temporal change in the degree of supersaturation of solution in which sample was immersed at 60 °C.

On the other hand, our results are not consistent with those reported by the researchers [10,23,24], who reported the in vitro transformation OCP into HA under physiological conditions. Moreover, the result that the transformation from OCP to HA was not observed at 37 °C in distilled water in the preliminary experiment cannot be explained by the explanation described above. Ban et al. [22] reported that OCP transformed

OCP

HA completely in the NaCl solution at 37 °C and speculated that magnesium ions in the SBF inhibit the transformation of OCP to HA. A likely explanation for the transformation of OCP to HA is that during the preparation of OCP, some formation of HA also occurred owing to its thermodynamic stability. Detection of a small amount of HA in OCP is difficult because XRD patterns of the HA and OCP are similar. In the presence of even a small amount of HA, crystal growth of HA would preferentially occur because the thermodynamic stability of HA is higher than that of OCP. Although OCP has been reported to transform to HA in vivo [6,10,22], this transformation did not occur in the present study when the powder was immersed in the SBF. When OCP is implanted on a bony site, it comes in contact with osteoblasts and osteoclasts, which make bones and resorb them, respectively. Although the roles of these cells on the transformation of OCP to HA are not clear, it is likely that they are involved. Based on the results obtained by the immersion of OCP in the SBF, we consider that it is necessary to understand the importance of HA nucleation for investigating the mechanism by which OCP transforms to HA. Therefore, OCP was then immersed in distilled water and its transformation was examined in detail. Fig. 3 shows the XRD patterns of the samples after immersion in distilled water at 60 ºC for different periods. No changes in the XRD patterns were evident during the early stage, with the powder immersed for 36 h showing only the characteristic peaks of pure OCP. However, it was observed that OCP was partially transformed to HA after

36 h

2 µm

96 h

2 µm Fig. 8. SEM images of sample before and after immersion in distilled water at 60 °C.

2 µm

N. Ito et al. / Materials Science and Engineering C 40 (2014) 121–126

immersion for 72 h, and that the sample immersed for 216 h comprised pure HA. Fig. 4 shows the temporal changes in the OCP content in the immersed sample, as determined from the XRD data. During the first 36 h of immersion, the sample phase comprised pure OCP. The OCP content decreased first significantly on increasing the period of immersion from 48 to 120 h. After 216 h of immersion, the sample phase comprised pure HA. Fig. 5 shows the temporal changes in the calcium and phosphorus concentrations in the solution in which the sample had been immersed, and Fig. 6 shows the temporal change in the solution pH. Fig. 7 shows the temporal change in the degree of supersaturation of the solution in which the sample had been immersed at 60 °C, as calculated from the data in Figs. 5 and 6. The calcium and phosphorus concentrations and the pH remained fairly constant for 12–36 h of immersion duration. During this period, the degree of supersaturation with respect to OCP was close to 1, while that with respect to HA was nearly 10. This indicates that the solution became saturated with OCP and highly supersaturated with HA within 12 h. On the other hand, the concentrations of calcium and phosphorus increased and the pH decreased rapidly after 36 h of immersion. It is speculated that the dissolved ions from OCP were used in the formation of HA, and the rest of the phosphoric ions remained in the solution because of the difference in the composition between OCP and HA. When OCP transforms to HA, phosphoric ions

125

are released because stoichiometric HA (Ca/P = 1.67) has a higher Ca/P molar ratio than OCP (Ca/P = 1.33). Moreover, the formation of HA seems to decrease the solution pH because HA incorporates hydroxide ions. The degree of supersaturation with respect to HA decreased dramatically after 48 h of immersion. As there was an unreactive period in the first stage of the transformation, which was followed by a rapid reaction, the HA crystals seemed to grow easily once the HA nuclei were formed. These phenomena are consistent with the results of the XRD analysis. Fig. 8 shows the SEM images of the initial OCP and the samples after immersion in distilled water at 60 °C for different periods. The OCP particles were plate shaped, and this morphology did not change even after 36 h of immersion. After 96 h, although the particles were observed to have slits at their edges, with some appearing to have dissolved slightly, the plate-like morphology derived from OCP was retained. Fig. 9 shows the TEM images and electron diffraction patterns of the sample after 96 h of immersion in distilled water at 60 °C. Based on the electron diffraction patterns, the edges and middle areas of the particles were assigned to HA and OCP, respectively. The c-axis of HA at the edge and that of OCP at the middle was parallel. This result implies that the HA crystals grew epitaxially on the OCP crystals, which is in agreement with previous reports [15,16].

Edge

Inside 1 μm Edge

Inside 002

002

110 220

Fig. 9. TEM image of sample after immersion in distilled water at 60 °C for 96 h, and electron diffraction patterns of the edge and inside of particle.

126

N. Ito et al. / Materials Science and Engineering C 40 (2014) 121–126

Based on the results obtained in the present study, the transformation of OCP to HA in distilled water is speculated to proceed via the initial dissolution of the OCP particles, with the solution becoming saturated with respect to OCP and supersaturated with respect to HA. Because OCP has an atomic arrangement similar to that of HA, it seems to provide sites that are favorable for HA nucleation. HA nucleation would occur when the dissolved ions are adsorbed on the surfaces of OCP with HA-like structure. However, the atomic arrangements of OCP and HA are not exactly the same, and it is speculated that this slight difference may have delayed the HA nucleation, resulting in the observed interval when no reaction occurred before HA transformation was observed. Once the HA nuclei were formed, HA crystals could easily grow, forming crystals by using calcium and phosphoric ions from the dissolved OCP. The HA-like structure would cause epitaxial growth of HA on the OCP. These results demonstrate the importance of HA nucleation for its transformation from OCP because the nucleation of HA is the rate-determining step in the transformation of OCP to HA. Although we have revealed here the importance of HA nucleation in the transformation of OCP to HA in vitro, the in vivo transformation of OCP to HA cannot be fully explained. Further researches are needed to understand the in vivo transformation behavior of OCP to HA. 4. Conclusion The OCP particles immersed in the SBF at 37 °C did not transform to HA even after 720 h of immersion. In distilled water at 60 °C, OCP started to transform to HA after 36 h of immersion, although the immersed solution became supersaturated with HA within 12 h of immersion. We have revealed here the importance of HA nucleation in the transformation of OCP to HA. Acknowledgments This work was partly supported by KAKENHI (11J04659) and the Nippon Sheet Glass Foundation for Materials Science and Engineering. References [1] L.L. Hench, J. Am. Ceram. Soc. 74 (1991) 1487–1510. [2] R.Z. LeGeros, Clin. Orthop. Relat. Res. 375 (2002) 81–98.

[3] S.V. Dorozhkin, Biomaterials 31 (2010) 1465–1485. [4] W.E. Brown, J.P. Smith, J.R. Lehr, A.W. Frazier, Nature 196 (1962) 1050–1055. [5] H. Tohda, M. Yamada, Y. Yamaguchi, T. Yanagisawa, J. Electron Microsc. 46 (1997) 97–101. [6] O. Suzuki, M. Nakamura, Y. Miyasaka, M. Kagayama, M. Sakurai, Tohoku J. Exp. Med. 164 (1991) 37–50. [7] O. Suzuki, Acta Biomater. 6 (2010) 3379–3387. [8] T. Handa, T. Anada, Y. Honda, H. Yamazaki, K. Kobayashi, N. Kanda, S. Kamakura, S. Echigo, O. Suzuki, Acta Biomater. 8 (2012) 1190–1200. [9] K. Miura, T. Anada, Y. Honda, Y. Shiwaku, T. Kawai, S. Echigo, T. Takahashi, O. Suzuki, Appl. Surf. Sci. 282 (2013) 138–145. [10] O. Suzuki, S. Kamakura, T. Katagiri, M. Nakamura, B. Zhao, Y. Honda, R. Kamijo, Biomaterials 27 (2006) 2671–2681. [11] G.S. Han, S. Lee, D.W. Kim, D.H. Kim, J.H. Noh, J.H. Park, S. Roy, T.K. Ahn, H.S. Jung, Cryst. Growth Des. 13 (2013) 3414–3418. [12] J. Zhan, Y.H. Tseng, J.C.C. Chan, C.Y. Mou, Adv. Funct. Mater. 15 (2005) 2005–2010. [13] M. Kamitakahara, N. Ito, S. Murakami, N. Watanabe, K. Ioku, J. Ceram. Soc. Jpn. 117 (2009) 385–387. [14] M. Mathew, W.E. Brown, L.W. Schroeder, B. Dickens, J. Crystallogr. Spectrosc. Res. 18 (1988) 235–250. [15] D.G.A. Nelson, J.D. McLean, Calcif. Tissue Int. 36 (1984) 219–232. [16] M. Iijima, H. Tohda, Y. Moriwaki, J. Cryst. Growth 116 (1992) 319–326. [17] R.Z. LeGeros, G. Daculsi, I. Orly, T. Abergas, W. Torres, Scanning Microsc. 3 (1989) 129–137. [18] M. Iijima, H. Kamemizu, N. Wakamatsu, T. Goto, Y. Doi, Y. Moriwaki, J. Cryst. Growth 181 (1997) 70–78. [19] R. Xin, Y. Leng, N. Wang, J. Cryst. Growth 181 (2006) 70–78. [20] Y.H. Tseng, C.Y. Mou, J.C.C. Chan, J. Am. Chem. Soc. 128 (2006) 6909–6918. [21] N. Ito, M. Kamitakahara, K. Ioku, Key Eng. Mater. 529–530 (2013) 11–14. [22] S. Ban, T. Jinde, J. Hasegawa, Dent. Mater. J. 11 (1992) 130–140. [23] R. Horváthová, L. Müller, A. Helebrant, P. Greil, F.A. Müller, Mater. Sci. Eng. C 28 (2008) 1414–1419. [24] J.V. Rau, M. Fosca, V.S. Komlev, I.V. Fadeeva, V.R. Albertini, S.M. Barinov, Cryst. Growth Des. 10 (2010) 3824–3834. [25] T. Yokoi, I.Y. Kim, C. Ohtsuki, Phosphorus Res. Bull. 26 (2012) 71–76. [26] M. Kamitakahara, C. Ohtsuki, A. Takahashi, M. Tanihara, J. Soc. Mater. Sci. Jpn. 55 (2006) 881–884. [27] M. Kamitakahara, H. Okano, M. Tanihara, C. Ohtsuki, J. Ceram. Soc. Jpn. 116 (2008) 481–485. [28] S.B. Cho, K. Nakanishi, T. Kokubo, N. Soga, C. Ohtsuki, T. Nakamura, T. Kitsugi, T. Yamamuro, J. Am. Ceram. Soc. 78 (1995) 1769–1774. [29] T. Kokubo, H. Takadama, Biomaterials 27 (2006) 2907–2915. [30] M.S. Tung, N. Eidelman, B. Sieck, W.E. Brown, J. Res. Natl. Bur. Stand. 93 (1988) 613–624. [31] H. McDowell, T.M. Gregory, W.E. Brown, J. Res. Natl. Bur. Stand. A Phys. Chem. 81A (1977) 273–281. [32] X. Liu, Y. Leng, Biomaterials 26 (2005) 1097–1108. [33] M. Iijima, H. Kamemizu, N. Wakamatsu, T. Goto, Y. Doi, Y. Moriwaki, J. Cryst. Growth 112 (1991) 467–473.