SynthesisofceramicsinMO –SiO systems through sol–gel

5656

Journal of the Ceramic Society of Japan 116 [1] 56-62 2008 Paper

Synthesis of ceramics in MOn/2–SiO2 systems through sol–gelprocessing under coexistence of polyethylene glycol andin vitro evaluation of their bioactivity

IllYong KIM, Masanobu KAMITAKAHARA,Giichiro KAWACHI, Koichi KIKUTA,SungBaek CHOand Chikara OHTSUKI

Graduate School of Engineering, Nagoya University, Furocho, Chikusaku, Nagoya 464–8603 JapanGraduate School of Environmental Studies, Tohoku University, 6–6–20, Aoba, Aramaki, Aobaku, Sendai 980–8579 JapanKorea Institute of Geosciences and Mineral Resources (KIGAM), Kajungdong 30, Yusongku, Taejon 305–350 Korea

Sol–gel processing from tetraethoxysilane (TEOS) under coexistence of polyethylene glycol (PEG) allows controlled morphol-ogy of the formed silica because of the phase separation during the polycondensation of the hydrated sol. The presence of PEGcontrolled the morphology effective in preparation of the ceramics, even in the CaO–SiO2 binary system, starting from TEOSand Ca(NO3)2. In the present study, some ceramics were prepared using the MOn/2–SiO2 binary systems, where M is given asNa＋ (n＝1), Mg2＋ (n＝2) and Al3＋ (n＝3), through sol–gel processing in the presence of PEG, to compare the compositionalchanges in the morphology of the ceramics. In the case of NaO1/2–SiO2, the framework size increased remarkably with increas-ing amounts of Na＋ in the starting compositions up to 30 mol, while continuous pores of approximately 10 mm were observedat 5 mol of Na＋. In the case of MgO-SiO2, the framework size increased with increasing amounts of Mg2＋ up to 20 mol inthe starting composition. Aggregation of particles with approximately 3 mm in size was observed at 25 mol and more. Thetrend in morphological change with addition of Mg2＋ was quite similar to that with Ca2＋. With AlO3/2-SiO2, particles wereobtained when the starting composition was more than 10 mol Al3＋. The size of the particles appeared to decrease withincreasing amounts of Al3＋. The morphology and phase separation behavior are strongly affected by the charge value of themetal ions, because the number of non-bridging oxygens coordinated to the metal ion increased with increasing charge value inthe sol–gel processing. A specimen prepared with starting composition 20NaO1/2･80SiO2 and heated at 600°C formed hydroxy-apatite on the particle surfaces within 7 d after soaking in a simulated body fluid (SBF), suggesting a potential for bioactivityon application of bone repair.

Key-words : Sol–gel processing, Silicate, Polyethylene glycol, Morphology

[Received October 10, 2007; Accepted November 15, 2007] ＠2008 The Ceramic Society of Japan

1. IntroductionSol–gel processing has been widely studied for the produc-

tion of silica particles and porous ceramics from tetraethox-ysilane (TEOS). Nakanishi and his colleagues1),2) reportedthat silica gel with controlled size and morphology can besynthesized using sol–gel processing in the presence of poly-ethylene glycol (PEG). The morphology of the synthesizedsilica gel was governed by phase separation in solutions con-taining water-soluble polymers during the polycondensationthat forms the siloxane network.

We expected that sol–gel processing in the presence ofPEG could be applied to fabrication of bioactive ceramicparticles, because calcium silicate is reported to act as amaterial that could induce bone-like apatite deposition on itssurface after exposure to body fluids, and because of resultsin osteoconduction through the surface bone-like apatitelayer. Osteoconduction, the so-called bioactivity of ceramicsfor bone repair, is the property whereby newly formed bonetissue grows directly on the surface of materials implanted inbony defects, and hence may be useful in assessing novelbiomaterials for bone reconstruction. We have previouslyreported the preparation of calcium silicate with a form ofmicrometer- and nanometer-sized particles, through a modi-fied process starting from TEOS and Ca(NO3)2.3) The sol–gel behavior in the TEOS–Ca(NO3)2–PEG system shows a

decrease in gelation time with increasing Ca(NO3)2 in thestarting composition. The compositional range in whichinterconnected pore structures could be fabricated in theTEOS–Ca(NO3)2–PEG was smaller than that for TEOS-PEG. This suggests that additives for producing metal oxidesto modify the siloxane network strongly affect the morphol-ogy of the resultant gels. However, the effects of such addi-tives have not been investigated in sol–gel processing in thepresence of PEG.

This study focuses on the compositional dependence of thestructural changes of silica gels after incorporation of sometypes of metal oxides, through sol–gel processing in thepresence of PEG. Ceramics were prepared from MOn/2–SiO2

binary systems, where M may be Na＋ (n＝1), Mg2＋ (n＝2)and Al3＋ (n＝3), using sol–gel processing in the presence ofPEG, to compare the composition-based changes in themorphology of the resulting ceramics. The potential forbone-like apatite formation was also preliminarily evaluatedin vitro using the simulated body fluid (SBF) proposed byKokubo et al.4),5) The microstructure and characteristics ofthe materials were investigated with comparisons to those ofthe CaO–SiO2 system.

57

Table 1. Starting Compositions of Solutions Containing Metal Nitrate, TEOS and PEG

57

Journal of the Ceramic Society of Japan 116 [1] 56-62 2008 JCSJapan

2. Experimental2.1 Synthesis of the specimens using sol–gel process-

ingTetraethoxysilane (TEOS, Si(C2H5O)4, Nacalai Tesque

Inc., Japan) was used as a silica source. Some metal nitratessuch as sodium nitrate (NaNO3), magnesium nitrate tetra-hydrate (Mg(NO3)2(4H2O) and aluminum nitrate ennahy-drate (Al(NO3)3(9H2O) were purchased as metal sourcesfrom Nacalai Tesque Inc., Japan. Nitric acid (HNO3,Nacalai Tesque Inc., Japan) was used a catalyst. Polyethy-lene glycol (PEG, Aldrich, USA) with a molecular weight of10,000 was used. All the chemicals were used for the prepa-ration of the gels without further purification. The startingcompositions for preparation of the gels are summarized inTable 1.

All the gels were prepared using methods describedpreviously.3) PEG and metal nitrates were dissolved inappropriate amounts of distilled water (7.71–4.76 cm3 forNaO1/2–SiO2 gels, 7.54–5.27 cm3 for MgO-SiO2 gels and7.46–4.76 cm3 for AlO3/2–SiO2 gels), and 60 mass nitricacid solution (0.56–0.41 cm3) was added. Then TEOS wasadded to the aqueous solution with vigorous stirring. Afterstirring for 20 min, the resultant solution was transferredinto a polystyrene square case (10×6.5×2.8 cm3), with itstop tightly sealed, and then kept at 40°C for 1 d for gelationand aging. The gelation time was determined by simplytilting the container to discover the time at which the bulkfluidity of the solution was lost. The wet gel obtained wasimmersed in distilled water for washing for 3 h to removePEG. The distilled water was renewed every 1 h. The wet gelwas then dried at 40°C for 7 d and then heated at 600°C for2 h.

Crystalline phases of the specimens were analyzed by X-ray diffraction (XRD, RINT PC 2100, Rigaku Co., Japan).The binding energy of O 1s of the specimens was measuredusing an X-ray photoelectron spectrometer (XPS, JPS–9000MC, JEOL, Japan). In the XPS measurements,AlKaX-rays were used as an excitation source. The measuredbinding energies were corrected with reference to the bindingenergy of C 1s (284.6 eV) of the hydrocarbon adsorbed onthe sample surfaces. The morphology of the prepared speci-

mens was observed using a scanning electron microscope(SEM, JSM5600, JEOL Ltd., Japan), after coating thespecimens with platinum. To analyze the metal content ofthe gels after heating at 600°C for 2 h, elementary characteri-zation was carried out using energy-dispersive spectroscopy(EDS, EX–54140, JEOL Ltd., Japan).

2.2 In vitro evaluation of bioactivityFormation of a bone-like apatite layer in the simulated

body fluid proposed by Kokubo and his colleagues4),5) wasexamined for some specimens to evaluate the osteoconduc-tion potential of the materials. Rectangular specimens withdimensions of 10×10×2 mm3 were cut from the bulk afterheat treatment at 600°C. The specimens were then soaked in30 cm3 of a simulated body fluid (SBF, Na＋ 142.0, K＋ 5.0,Mg2＋ 1.5, Ca2＋ 2.5, Cl－ 147.8, HCO －

3 4.2, HPO 2－4 1.0,

SO 2－4 0.5 mol･m－3) with ion concentrations nearly equal to

those of human blood plasma with pH 7.25 at 36.5°C.4),5)

After the specimens were soaked in SBF for predeterminedperiods of 1 and 7 d in the incubator at 36.5°C, they wereremoved from the SBF, and gently washed with distilledwater and then dried at room temperature. The surfacestructures of the specimens were analyzed by XRD. The sur-face morphology of the specimens was observed under SEM.

3. ResultsAll the prepared solutions remained transparent during

stirring for 20 min, and there were no appearances ofprecipitates. Figure 1 shows the gelation times of the solu-tions, when they were prepared without or with the presenceof PEG. When the solution was prepared without PEG, thegelation time of the solution decreased with increasing theadded amounts of magnesium (Mg2＋) and aluminum(Al3＋) increased, but increased with increasing sodium(Na＋), The decreasing degree of gelation time of the Al3＋

added solution is larger than that of the Mg2＋ added solu-tion. When the solutions were prepared with PEG present,the gelation time of the Na＋ added solution also decreasedwith increasing amounts of Na＋ in the starting composition.The gelation times of Mg2＋ added solutions and Al3＋ addedsolutions decreased with increasing amounts of Mg2＋ and

58

Fig. 1. Gelation time of the solutions. (a) without PEG and (b) with PEG.

Fig. 2. SEM photographs of the NaO1/2–SiO2 (xNa(100－x)Si) specimens heated at 600°C for 2 h.

58

JCSJapanKim et al.: Synthesis of ceramics in MOn/2–SiO2 systems through sol–gel processing under coexistence of

polyethylene glycol and in vitro evaluation of their bioactivity

Al3＋ with the similar degree of decreases, while the decreasein gelation time for the Na＋ added solution was smaller thanthose of the Mg2＋ and Al3＋ added solutions.

After gelation and aging, all the gels were white opaqueand crack-free, irrespective of their starting compositions.The specimens prepared with the addition of sodium nitratelooked partially black after heating at 600°C. Figure 2 showsSEM photographs of the NaO1/2–SiO2 specimens after heat-ing at 600°C for 2 h. It can be seen from Fig. 2 that the sizeof the frameworks and particles of the specimens preparedwith sodium nitrate increased with increasing sodium con-tent up to 30 mol. It is evident that the size of the frame-works increased remarkably with increasing amounts of Na＋

in the starting compositions from 10 mol, while continu-ous pores of approximately 10 mm were observed at 5 molof Na＋. In a previous study on the effect of calcium nitrates(Ca(NO3)2･2H2O),3) the aggregation of spherical particles

could be obtained with the addition of 25 calcium contentin the CaO–SiO2 system. In contrast, spherical particlescould not be obtained with the addition of sodium nitrate inthe range of starting compositions from 0 to 30 molarratio in the NaO1/2–SiO2 system.

Figure 3 shows SEM photographs of the MgO–SiO2 speci-mens after heating at 600°C for 2 h. The framework and par-ticle sizes increased with increasing magnesium content.Aggregation of particles to sizes of approximately 3 mm wasobserved at 25 mol and more in the MgO–SiO2 system,while an interconnected porous structure was observed forcompositions of less than 20. The trend in morphologicalchange with the addition of Mg2＋ was quite similar to thatwith Ca2＋. Figure 4 shows SEM photographs of the AlO3/2–SiO2 specimens after heating at 600°C for 2 h. The intercon-nected porous structure changed to the aggregation of parti-cles by addition of aluminum nitrate at 10 and more in the

59

Fig. 3. SEM photographs of the MgO–SiO2 (xMg(100－x)Si) specimens heated at 600°C for 2 h.

Fig. 4. SEM photographs of the AlO3/2–SiO2 (xAl(100－x)Si) specimens heated at 600°C for 2 h.

59


AlO3/2–SiO2 system. Particles were obtained with startingcompositions of more than 10 mol. The size of the parti-cles appears to decrease slightly with increasing amounts ofAl3＋.

From these morphological observations, the structures ofthe formed gels varied from porous products to the aggrega-tion of spherical particles, depending on the kinds andamounts of added metal (M) ions. With magnesium, theaggregation could be obtained at the starting composition,Mg/Si＝25/75 molar ratio, while with aluminum, it was

obtained at the starting composition, Al/Si＝10/90. Incontrast, there are no suitable compositions to obtain sphereparticles in the NaO1/2–SiO2 system up to Na/Si＝30/70molar ratio at starting compositions.

Figure 5 shows XRD patterns of NaO1/2–SiO2 gels afterheating at 600°C for 2 h. It is apparent that the specimenscrystallized to form cristobalite with this heat treatment.From the intensity of the peaks assigned to cristobalite, itseems that the crystallization temperature decreased withincreasing amounts of added sodium nitrate. In the SEM

60

Fig. 5. XRD patterns of NaO1/2–SiO2 (xNa(100－x)Si) specimensheated at 600°C for 2 h.

Fig. 6. Relationships between starting compositions and resultantcom positions of the obtained specimens heated at 600°C for 2 h,determined by EDS analysis.

Fig. 7. O 1s peak areas of the specimens prepared from 20M80Sicompositions, determined by XPS analysis. No metal was preparedfrom composition. Na＋ (n＝1), Mg2＋ (n＝2) and Al3＋ (n＝3),where n is ion charge of metal.

60



photographs of the NaO1/2–SiO2 gels in Fig. 2, small chipson the frameworks or particles were observed at higher Na＋

contents. These chips seem to be the cristobalite formedfrom the gels during the heat treatment. In the specimensprepared with magnesium and aluminum, we could notdetect any peaks assigned to crystalline phases after heattreatment. No previous studies on the preparation of speci-mens in the system CaO–SiO2 reported any crystalline phasesfrom the XRD measurement.

Figure 6 shows the results of compositional analysis ofthe prepared specimens, determined by EDS for the gelsafter heating at 600°C for 2 h. The analyzed proportions ofadded metals to silicon (M/Si) in molar ratio increased withincreasing metal contents. There is no significant differencein the incorporated metal ions among the resultant speci-mens, but all the compositions of the resultant specimensshow a lower proportion of metal ions than was present inthe starting mixture. Some amount of metal salt may exist inthe salinated solution after extrusion during shrinking.Moreover, in this study, the prepared wet gels were washed

with distilled water for 3 hours. Metal salts were dissolved tothe distilled water. Thus, to obtain a predetermined compo-sition in the resultant specimens, excess metal nitrate isrequired in the starting composition.

Figure 7 shows O 1s peak areas of the gels prepared withthe addition of 20 mol metal nitrate by XPS analysis afterheat treatment at 600°C for 2 h. The assignment of each peakwas conducted according to the literature reported by Miyajiet al.6) The specimens prepared without the addition of metalnitrate did not show any presence of non-bridging oxygen,implying the assignment of Si–O–M. The 20Al80Si specimenhas the largest amount of non-bridging oxygen among thethree specimens 20Na80Si, 20Mg80Si, and 20Al80Si. Thismeans that the amount of non–bridging oxygen (Si–O–M)increases with increasing charge value of the metal cations,even when added in the same composition. The amount ofsiloxane (Si–O–Si) decreased with increasing non-bridgingoxygen in the specimen. The number of silanol groups(Si–OH) was almost the same among the prepared gels.20Na80Si seemed to have a smaller number of silanol groupsthan the others; this may be attributed to the crystallizationof cristobalite.

Figure 8 shows XRD patterns of the specimens after heat-ing at 600°C, before and after soaking in SBF for 7 d. A veryweak peak at about 32 in 2u was detected with the specimenprepared with starting composition of 20Na80Si after soak-ing in SBF for 7 d. The other specimens did not form the par-ticles that indicate deposition of hydroxyapatite on their sur-faces after soaking in SBF. Figure 9 shows an SEM photo-graph of the specimens after heating at 600°C, before andafter soaking in SBF for 1 and 7 d. These results indicate thatceramic prepared using the composition 20Na80Si formedapatite on its surface after soaking in SBF for 7 d. Theapatite formation resulted from increased pH because of therelease of sodium from the materials.7) This means thatapatite formation can be controlled by the amount of addedsodium.

4. DiscussionIn this study, we obtained crack-free monoliths of MOn/2–

61

Fig. 8. XRD patterns of the specimens of 20M80Si heated at 600°C, followed by soaking in SBF for 1 and 3 d.

Fig. 9. SEM photographs of the 20M80Si specimens heated at 600°C followed by soak ing in SBF for 1 and 7 d.

61


SiO2 binary systems through sol–gel processing. In thepresence of PEG, the gelation time of all the compositionsdecreased with increasing metal content. The effects of theaddition of metal ions on gelation time were confirmed.From the results of SEM observations (see Figs. 2, 3 and 4),the effects of the metal ions on the structural change of thegel increased with increasing ionic charge of the cations.Higher charge values of metal additives may allow formationof particles of the resultant gels with lower amounts ofadditives. For example, the required addition of sodium ions

is estimated to be 50 in molar ratio to obtain sphericalparticles, because calcium is a divalent cation having twicethe charge of the sodium ion. Bansal et al.8) reported theinfluence of the addition of several metal ions on the prepa-ration of silica gel through sol–gel processing. They classi-fied the added metals (Cu2＋, Al3＋, La3＋, Y3＋, Li＋, Na＋,Mg2＋, Ca2＋, Sr2＋) ions into two groups. One group resultsin a decrease of gelation time and has ions that are generallyclassified as glass network modifiers. The other group resultsin an increase of gelation time and has ions that are classified

6262



as glass intermediates. Our previous study reported thatthe gelation time of the CaO–SiO2 system decreased withincreasing amounts of added calcium compound. Thepresent study also showed that the gelation times of thegels in the systems NaO1/2–SiO2, MgO–SiO2 and AlO3/2–SiO2

decreased with increasing metal ion concentrations when thesol–gel processing is conducted in the presence of PEG.

Comparison of results using sodium, magnesium and alu-minum ions reveals that aluminum ions bond to oxygenmore rapidly than sodium or magnesium ions because of therapid gelation with aluminum. This can be related to differ-ences in ionic field strength among these metal ions. Theionic field strength, F, is defined as:

F＝Zr 2, (1)

where Z is the valence of the ion and r is its ionic radius.9)

This equation means that metal ions with high positivevalences have high ionic field strengths. The ionic fieldstrength of aluminum ions is thus greater than those of sodi-um and magnesium. The higher ionic field strength allows ahigher rate of gelation in sol–gel processing for productionof silicate compounds from TEOS.10) Therefore, the solu-tions prepared with the addition of aluminum nitrate gelledfaster than the solutions prepared with the addition of mag-nesium or sodium.

The size and morphology of gels is affected by their ratesof gelation.2) According to the previous research on thepreparation of silica gel through sol–gel processing of siliconalkoxides such as TEOS, the gelation time is affected by theratio of water and the alkoxide, as well as by the catalyst andits type. These parameters were the same among the exa-mined compositions. However, the gel formation was con-ducted with various kinds and amounts of the added metalions. Thus, the metal ions strongly affected the gelation reac-tion of the binary MOn/2–SiO2 system. In our previous study,the morphology of the silica gels prepared with the additionof calcium nitrate changed from the porous interconnectedbody to an aggregation of spherical particles. In this study,the morphology of the gel prepared with the addition ofsodium nitrate did not change; however, the frameworkincreased with increasing sodium content (see Fig. 2). Withthe addition of magnesium nitrate, the change in the mor-phology of the gel was similar to that of the gel preparedwith the addition of calcium nitrate.3) With the addition ofaluminum nitrate, the morphology of the gel was changedto spherical particles at lower concentrations than wererequired with calcium.

These changes in morphology were related to the chargevalence of the metal ions, and are similar to the formation ofnon-bridging oxygen in glass structures. From the viewpointof glass formation, each divalent cation such as Mg2＋, Ca2＋

and Ba2＋ form two non-bridging oxygen ions, while eachmonovalent cation such as Li＋, Na＋, K＋ forms one of non-bridging oxygen.11) With aluminum ions, the XPS results ofthe gels prepared with 20 mol of metal in the startingcomposition show that the gel containing aluminum ionhas a high proportion of non-bridging oxygen comparedwith specimens prepared with sodium and magnesium (seeFig. 7). It therefore seems that aluminum ions act as a net-work modifier, making non-bridging oxygen, and that analuminum ion bonds to three SiO4 tetrahedras.

The present study also showed the potential of using sodi-

um ions for producing bioactive materials. Cho et al.reported11) that silica gel porous bodies prepared in thepresence of water-soluble polymers formed hydroxyapatiteon their surfaces when they were immersed in SBF within 2weeks. In contrast, silica gel monoliths prepared in theabsence of water-soluble polymers did not form it. This sug-gests that the ability of apatite formation in SBF depends onthe porous structure of the silica gel determined by process-ing. Thus, the bioactivity of pure silica gel is not enough forapplication as bioactive material, because of its low ability ofapatite formation. On the other hand, it was reported7),12),13)

that CaO–SiO2 and Na2O–SiO2 glasses have high ability ofapatite formation in SBF because the glasses allow increasesin degree of supersaturation of the surrounding fluid withrespect to hydroxyapatite, attributed to the dissolution ofcalcium and sodium, as well as the formation of silanolgroups on their surfaces. We can conclude that modificationof silica gel with calcium or sodium can enhance the abilityof apatite formation in the body environment, i.e., bioactivi-ty, even when prepared in the presence of water-soluble poly-mer.

5. ConclusionSol–gel processing in the presence of PEG was conducted

to fabricate ceramics in the binary systems MOn/2–SiO2 (M＝Na, Mg, Al). After sol–gel processing, gels could beobtained in a form with interconnected pores with low metalconcentration, and with aggregation of spherical particleswith high metal concentrations. The morphological changesof the forms from porous products to spherical particlesdepended on the charge valences of the metal cations. Theproducts of the system NaO1/2–SiO2 showed the ability toform hydroxyapatite in SBF, indicating potential bioactivityfor bone repair.

References1) K. Nakanishi and N. Soga, J. Am. Ceram. Soc., 74,

2518–2530 (1991).2) K. Nakanishi, H. Komura, R. Takahashi and N. Soga, Bull.

Chem. Soc. Jpn., 67, 1327–1335 (1994).3) I. Y. Kim, C. Ohtsuki, M. Kamitakahara and S. B. Cho, Key

Engineering Materials, 330–332, 177–180 (2007).4) S. B. Cho, K. Nakanishi, T. Kokubo, N. Soga, C. Ohtsuki, T.

Nakamura, T. Kitsugi and T. Yamamuro, J. Am. Ceram.Soc., 78, 1769–1774 (1995).

5) T. Kokubo and H. Takadama, Biomaterials, 27, 2907–2915(2006).

6) F. Miyaji, M. Iwai, T. Kokubo and T. Nakamura, J. Mater.Sci. Mater. Med., 9, 61–65 (1998).

7) L. L. Hench and J. Wilson,“An Introduction to bio-ceramics,”World Scientific, Singapore (1993) pp. 41–52.

8) N. P. Bansal, J. Am. Ceram. Soc., 73, 2647–2652 (1990).9) D. D. Petrova,“Phase Separation in Glass,”Ed. by O. V.

Mazurin and E. A. Porai-Koshits, Amsterdam (1984) pp.35–44.

10) P. Saravanapavan and L. L. Hench, J. Non-Cryst. Solids,318, 1–13 (2003).

11) S.B. Cho, K. Nakanishi, T. Kokubo, N. Soga, C. Ohtsuki andT. Nakamura, J. Biomed. Mater. Res., 33, 145–151 (1996).

12) C. Ohtsuki, T. Kokubo, T. Takatsuka and T. Yamamuro, J.Ceram. Soc. Japan, 99, 1–6 (1991).

13) T. Kokubo, S. Ito, Z. T. Huang, T. Hayashi, S. Sakka, T.Kitsugi and T. Yamamuro, J. Biomed. Mater. Res., 24,331–343 (1990).

Documents

SynthesisofceramicsinMO –SiO systems through sol–gel