Journal of Sol-Gel Science and Technology

, 47:335

Acid character control of bioactive glass/polyvinyl alcohol hybrid foams produced by sol–gel

Authors

    • Department of Metallurgical and Materials EngineeringFederal University of Minas Gerais
  • V. Ciminelli
    • Department of Metallurgical and Materials EngineeringFederal University of Minas Gerais
  • M. S. S. Dantas
    • Department of Metallurgical and Materials EngineeringFederal University of Minas Gerais
  • H. S. Mansur
    • Department of Metallurgical and Materials EngineeringFederal University of Minas Gerais
  • M. M. Pereira
    • Department of Metallurgical and Materials EngineeringFederal University of Minas Gerais
Original Paper

DOI: 10.1007/s10971-008-1777-1

Cite this article as:
de Oliveira, A.A.R., Ciminelli, V., Dantas, M.S.S. et al. J Sol-Gel Sci Technol (2008) 47: 335. doi:10.1007/s10971-008-1777-1

Abstract

Bioactive glass/polymer hybrids are promising materials for biomedical applications because they combine the bioactivity of bioceramics with the flexibility of polymers. In previous work hybrid foams with 80% bioactive glass and 20% polyvinyl alcohol were prepared by the sol–gel method. The produced hybrids presented a high acidic character due to the catalysts added. In this work different methods to control the acidity and toxicity of the hybrids were also evaluated, through changes in the synthesis pH and use of different neutralization solutions. The hybrids were prepared with inorganic phase composition of 70%SiO2–30%CaO and PVA fractions of 20–60% by the sol–gel method. The characterization of the obtained foams was done by FTIR, SEM, Raman Spectroscopy, Helium Picnometry and TGA. The immersion of hybrids in a calcium acetate solution was the most adequate neutralization method. The foams presented porosity of 60–85% and pore diameters of 100–500 μm with interconnected structure.

Keywords

Hybrid foamsBioactive glassPolyvinyl alcoholNeutralization methodsChemical structure

1 Introduction

When a significant loss of tissue occurs as a result of trauma or disease, the total healing occurs only with the support of grafts implanted. Human therapies or other species living or non-living have been restricted due to the limited availability of material, and the surgical complications of multiple stages in the way of the site of collection, and the risk of disease transmission. These factors create a big demand for synthetic substitutes specially designed and manufactured to act as matrices for tissue engineering [1]. Tissue engineering is a very important area of research which uses biomaterials as matrices to support the cell cultures in order to develop living tissues. A biomaterial gives the basis for the tissue in growth, promoting the repair, proliferation and the natural regeneration of tissue, rebuilding and replacing damaged tissue. Thus, tissues can be repaired from cultures made of the patient’s own cells, without waste material by the use of artificial biodegradable materials. This way, cellular responses that normally only occur naturally, can be induced by accelerating the process of healing and rehabilitation [2].

Some requirements for the materials to be applied as matrices in tissue engineering include: (i) three-dimensional structure with high porosity of interconnected macropores (diameters of the order of 100 μm), to have cell migration and nutrition throughout the material; (ii) biodegradable compositions, with controlled degradation rates, compatible with the cell/tissue growth in vitro or in vivo, (iii) bioactive surface, to promote cell proliferation and differentiation; (iv) mechanical properties appropriate to the physiological conditions and the tissue replaced, as well as the neighboring tissues; (iv) uniformity of structure, to promote cell homogenous bond; (v) to be easily processed into large variety of shapes and sizes [3]. In particular, for the replacement of bone tissue, moreover, to all the features already mentioned, it is desirable that the matrix have mechanical behavior compatible to bone, with Young’s Modulus of 300–500 MPa and elastic deformation of 5–10%. These properties must be maintained associated with the degradation of the matrix and the growth of new tissue [4].

Bioactive glasses have good biological characteristics that indicate it as a promising matrix for bone tissue engineering. When implanted in the body, they induce an interfacial bioactive answer [5]. Furthermore, the dissolution products of bioactive glass exert control over genetic factors of bone growth [6]. In previous works we developed a process that allows the production of bioactive porous foams using the sol–gel method [7]. These structures have appropriate porosity and interconnectivity, but have low mechanical strength and toughness, which limits its application in tissue engineering.

An approach to improve the mechanical properties of bioactive glasses is the production of organic–inorganic hybrids, where an inorganic phase, with nanodimensions, is inserted in a polymer matrix. The addition of ceramics to polymer matrices allows the production of new materials with superior properties [8]. Thus, one material can combine the bioactivity of ceramics with the flexibility of polymers [9]. The sol–gel process is potentially useful in enabling such combination in molecular and nano scales. It allows the preparation of ceramics at temperatures compatible with the polymers’ processing [10].

Hybrid foams of the system bioactive glass/polyvinyl alcohol (PVA) was studied by Pereira [11, 12], that produced porous foams by the sol–gel method for application in tissue engineering. The foams were obtained with the polymer content of 20% and different compositions of the inorganic phase. Hybrids were obtained by hydrolysis of tetraethyl orthosilicate (TEOS) in the presence of hydrochloric acid (HCl) solution and subsequent addition of calcium chloride. A solution of PVA, surfactant and hydrofluoric acid (HF) solution were added to the sol, and the mixture vigorously agitated for foam formation. The foam was put into containers where the gelation occurred. The samples were dried at low temperature, because the treatment at high temperatures causes the degradation of the polymer. The foams obtained had porosity of 60–90% and pore diameters between 10 μm and 600 μm. The presence of the polymer increased the toughness compared to the bioactive glass foams. The hybrid foams also showed greater resistance and greater deformation.

Hybrids produced by this route had a high acid character because of the catalysts added during the process. Consequently, an additional step of neutralization was necessary to produce biocompatible foams. The neutralizing method used was the immersion of foams in solution of the ammonium hydroxide. It changed the structure of pores of the material and decreased the amount of calcium in hybrids. After neutralizing the hybrids showed strength and deformation slightly lower than before the procedure. The change of behavior of the material is related to the change of composition and pore structure. The neutralization method should be adjusted.

In this work it was evaluated the effect of increasing the PVA content on structural characteristics of hybrid foams produced by method sol–gel. Different methods to control the acidity of hybrids produced were evaluated by variations in the pH of synthesis and the use of different neutralizing solutions.

2 Materials and methods

The reagents used to synthesize the hybrids were tetraethyl orthosilicate (TEOS) 98% by Across Organics, hydrochloric acid (HCl) 1 mol/L by Merck, calcium chloride (CaCl2 · 2H2O) by Vetec, poly(vinyl alcohol) (PVA) 80% hydrolyzed by Sigma-Aldrich, the surfactant sodium lauryl ether sulfate 27% by Sulfal, hydrofluoric acid (HF) 48% by Merck, ammonium hydroxide 28% by Sigma-Aldrich and calcium acetate (Ca(CH3COO)2 · H2O) by Vetec. NH4F solution was made with ammonium hydroxide and hydrofluoric acid. The hybrids were prepared with inorganic phase composition of 70%SiO2–30%CaO and PVA fractions of 20–60% by two different routes of the sol–gel method.

Route 1: the sol was prepared by acid hydrolysis of TEOS followed by the addition of CaCl2 · 2H2O and 20 wt% PVA solution. The surfactant and HF solution were added. The mixture was vigorously stirred to form the foam, which was then cast in a container were it gelled and was dried at 40 °C in an air circulation oven for one week. Hybrid samples were immersed three times, 30 min each, in different neutralization solutions: aqueous and alcoholic ammonium hydroxide solutions and aqueous and alcoholic calcium acetate solutions. The most effective neutralization method was chosen by the evaluation of calcium and PVA losses and pH measurements. After neutralization, the hybrids were dried again at 40 °C in an air circulation oven for one week and, after, under vacuum for 48 h.

Route 2 was proposed to control pH of synthesis and, consequently, to control the acidity of hybrids. It consisted of the hydrolysis of TEOS for 3 days, without addition of acid catalyst, followed by the addition of CaCl2, PVA solution, surfactant. NH4F was chosen as catalyst instead of HF. The mixture was vigorously stirred and cast in a container were it gelled and was dried at the same way as route 1.

The characterization of the foams obtained by the two routes used was done by several techniques. A Horiba Jobin Yvon LABRAM-HR 800 spectrograph with a monochromator slit an Olympus BHX microscope, were used to obtain the Raman spectra in the wavenumber between 4,000 cm−1 to 400 cm−1. The excitation source was a linearly polarized 633 nm helium-neon laser, with 20 mW. The Raman spectra of all samples were obtained by placing the fluid directly under the objective of the microscope. A large numbers of short scans were accumulated to yield the 2-min integration time. The transmittance infrared analyses were obtained in Centauros FTIR Nicolet Nexus 470 in the wavenumber range of 4,000–400 cm−1. The porosity of the foams was obtained by Archimedes Method, Helium Picnometry (AccuPyc 1330, Micromeritics) and qualitative analysis of the images obtained by Jeol Scanning Electron Microscope (SEM). Thermogravimetry Analysis (TGA) was done in equipment Perkin-Elmer with heat rate of 10 °C/min with temperature from 25 °C to 1,100 °C, under nitrogen atmosphere.

3 Results and discussion

3.1 Synthesis routes

Raman spectroscopy was used to evaluate, in the molecular scale, the sol structure until the gel formation by the two routes proposed in this work. The results obtained were compared to literature [1349]. The Raman spectra obtained at various times during the sol–gel reaction, in which TEOS hydrolysis was acid catalyzed, with water/alkoxide ratio 12 at room temperature (route 1), is shown in Fig. 1. The spectra were compared with those of TEOS reagent and ethanol, a reaction product. The most significant reactions occurred for reaction times between 10 min and 20 min.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig1_HTML.gif
Fig. 1

Raman spectra obtained at various times during the sol–gel reaction by the route 1. The traces line was added to mark the 1,001 cm−1 peak that refers to the container where the reaction was conducted

Table 1 lists the major assignments of Raman bands observed in the 4,000–450 cm−1 region of the spectra and the structure relation with the mode vibration of these bands. TEOS produced bands at 311, 404, 655, 810, 934, 962 1,033 and 1,092 cm−1. The bands in the range from 2,700 cm−1 to 3,000 cm−1 are related to vibration contributions from CH2 and CH3 groups, in addition to the vibrations of –OH groups in ethanol, silanol and the water adsorbed on dimmers [17]. Due to the large number of species that have vibration at this wave number, it is very difficult to define the characteristic bands of each specie, so it was not possible to make a more detailed analysis of these bands.
Table 1

Structural assignments of Raman bands observed during TEOS acid hydrolysis

Vibration modes

Bond

Structure

Raman bands (cm−1)

References

δ(C–C–O)

C–C–OSi

Monomers

311

[15, 17, 21, 26, 36]

Dimmers and trimmers

306

[17, 21, 26]

C–C–OH

Ethanol

432

[14, 17, 21, 25, 28, 36]

δ(O–Si–O)

CO–Si–OC

Monomers

404

[1517, 19, 21, 26, 29, 33, 36, 37]

HO–Si–OC

Hydrolyzed monomer

406–436

[16, 17, 19, 21, 26, 29]

HO–Si–OH

Fully hydrolyzed monomer

436

[1719, 21, 26, 27, 29, 31, 33, 37, 3942, 45]

νs(Si–O)

Si–OC

Monomers

655

[17, 20, 26, 29, 30, 33, 34, 43, 44]

Si–OC e Si–OSi

Oligomers

621

[17, 2226, 29, 33, 34, 44]

Rings with 2–4 Si

537

[17, 23, 26, 29, 33, 34, 42, 44]

Rings with 4–6 Si

511

[17, 26, 29, 33, 42, 44]

Rings with 6–8 Si

506

[17, 23, 26, 29, 3436, 42, 44]

Network

487

[15, 17, 2127, 29, 31, 3340, 4348]

νas(Si–O)

Si–OC

Monomers

1092

[15, 17, 2226, 29, 35, 44]

Si–OC e Si–OH

Partially hydrolyzed monomers

1090–1089

[17, 18, 2125, 29, 3342, 4448]

νs(C–C)

C–COSi

Monomers

1033

[17, 24, 32, 36]

C–COSi e C–COH

Hydrolyzed monomers and ethanol

1047–1049

[14, 17, 20, 32]

C–COH

Ethanol

1052

[17, 28, 32]

νs(C–O)

C–OSi

Monomers

810

[17, 2225, 32]

C–OSi e C–OH

Hydrolyzed monomers and ethanol

882

[17, 2225, 30, 32, 35]

C–OH

Ethanol

884

[14, 17, 2225, 28]

ν(–OH)

Si–OH

Hydrolyzed oligomers

2700–3000

[16, 17, 19, 21, 29, 37, 40, 44]

C–OH

Ethanol

[13, 14, 17, 19]

H–OH

Water

[13, 17, 19, 4449]

νs(O–H)

SiO–H

Hydrolysed oligomers

3100–3700

[13, 16, 17, 19, 31, 39, 40, 44]

CO–H

Ethanol

[13, 14, 16, 17, 28]

HO–H

Water

[13, 16, 17, 4449]

The overlapping of the bands in the region between 3,100 cm−1 and 3,700 cm−1 was attributed to the contributions of O–H in water, silanol, and ethanol; in addition to the vibration of the water molecules and hydrogen bonds between different species. It can be noted that these bands were beginning to emerge in a reaction time of 10 min, showing that there were already reaction products, water and ethanol.

The intermediate species, produced by the hydrolysis and condensation reactions of the sol–gel method, are assigned to bands with vibration wave numbers very close or overlapping, which also changed their position with the reaction progress. When these transitions occur only in bands characteristics of these changes in the structure, it can be considered that there is a displacement of wave numbers if these variations are above 2 cm−1. These displacements are confirmed by the position of a band of reference, which should be the same in all spectra [38]. In this case the band at 1,001 cm−1 was used as a reference because it is the band of the container in which the material was tested, so they didn’t change over the reaction. The convolution method was used to help to elucidate these modes of vibration. The graphs shown in Fig. 2 are results of convolutions obtained by the method of Lorenzian Magnitude (with 100 interactions, at confidence intervals of 95% and a R2 less than 0.9996) adjusting the best baseline for each curve.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig2_HTML.gif
Fig. 2

Convolution results by Lorentzian Method, at 750–650 and 600–200 cm−1, for (a) TEOS, and for various times during the sol–gel reaction: (b) 10, (c) 20 and (d) 120 min

The reaction can be monitored by the evolution of TEOS in different species. TEOS molecules were hydrolyzed in several species involving the changes of groups etoxy (–OCH2CH3) for (–OH), that caused the displacement of Raman bands to higher wave numbers, due to the reduction in weight caused by these substitutions. The displacement of the band at 404 cm−1, for TEOS, to values between 406 cm−1 (0–10 min) and 436 cm−1 (20–120 min), as seen in Fig. 2, were related to the O–Si–O bending mode in monomers partial and completely hydrolyzed, respectively.

Intermediate species partially hydrolyzed are involved in the condensation reactions. Because of the mass increase that occurs when monomers reacts forming dimmers, trimmers, tetramers and, finally, the network, the corresponding bands to these species move to smaller wave numbers. This can be seen in the displacement of the band at 311 (TEOS) to 306 cm−1 (dimmers and trimmers). These reactions produce a low intensity band at 620 cm−1, which amplitude increases as the intensity of the band at 655 cm−1 (TEOS) decreases. Thus, the band at 621 cm−1 is assigned to Si–O symmetric stretching in species containing at least one Si–O–Si bond, which indicates that the condensation reaction started in the time of 0 min. The bands at 934 cm−1 and 962 cm−1 are assigned to species not condensed until the time of 10 min. Moreover, a small displacement of the band at 1,092 cm−1 (TEOS) to values between 1,090 cm−1 and 1,088 cm−1 are assigned to Si–O–C stretching vibrations of different species partially hydrolyzed.

The bands at 655 cm−1 and 306 cm−1 almost disappeared at the time of 20 min, indicating a marked reduction of TEOS and simple condensed structures. The band at 621 cm−1 remained there until the time of 120 min, but its intensity becomes smaller because of the evolution of this specie to cyclical condensed species. These new species are confirmed by the presence of bands at 560–480 cm−1, which started appearing at the time of 10 min, associated with rings with 2–4 and 4–6 atoms of silicon. These structures tend to evolve into a three-dimensional network, which can be seen by the decrease in the number of bands at the time of 120 min, leaving only the bands at 506 and 487 cm−1, referring to vibrations in rings with more than 6 members and the network, respectively.

The ethanol and water formation during the process can also be used to make a reaction analysis. When two hydrolyzed molecules, complete or partial, have a condensation reaction involving silanol groups, they produced Si–O–Si bonds and water or alcohol as product. At 20 min the ethanol related bands were formed (432, 882 and 1,052 cm−1). The 432 cm−1 band was overlapped, at 20–120 min, by the 436 cm−1 band, assigned to completely hydrolyzed monomers. The band at 882 cm−1, associated with the C–O–H species, is a result of the displacement of the 810 cm−1 band, caused by the change of –OCH2CH3 for –OH groups.

During the reaction, the characteristic band of C–C bond moved from 1,033 cm−1 (TEOS) to 1,047 cm−1 (20 and 60 min) and 1,049 cm−1 (120 min), highlighting the transition of C–COSi to C–COH species. There was still the presence of not hydrolyzed groups at the time of 120 min, because of the C–C assigned band was at 1,049 and not at 1,052 cm−1, as was expected in completely hydrolyzed species.

The Raman spectra for various times of sol–gel reaction without catalyst added (route of synthesis 2) are shown in Fig. 3, and compared with the TEOS and ethanol spectra. The spectra for route 2 had the same bands that the spectra of route 1, a difference being observed only in time and intensity of these bands. It can be noted that 48 h reaction time was not enough to form the three-dimensional structure of the network. From these results, the time of hydrolysis was increased to 72 h, time very close to gelation, which occurred in approximately 75 h. The spectra generated at that time presented no major changes in relation to the 48 h spectra, indicating that the structure formed by the neutral route does not characterize a the network until the moment of gelation.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig3_HTML.gif
Fig. 3

Raman spectra obtained at various times during the sol–gel reaction by route 2. The traces line was added to mark the reference band

The different structures formed during the reactions by the sol–gel routes 1 and 2 can be explained by the different reaction pH’s. In the method sol–gel, when the hydrolysis is acid catalyzed, below pH 2, the speed of condensation reaction is greater than the speed of hydrolysis, forming long chains early in the process. These structures are sensitive to the water concentration. High water concentration favors the hydrolysis and decreases the condensation rate. Thus, an excess of water, even under acid conditions, can be used to ensure the complete hydrolysis at an early stage of the reaction, resulting in a high density of crosslinks [17]. Route 1 occurred at pH 1.2 approximately. Below this pH, hydrolyzed particles collide with each other, giving rise to chains and, subsequently, to a three-dimensional solid, confirmed by the presence of Raman bands associated with rings having more than 6 Si atoms and the network. The growth of crosslinks occurs by the incorporation of small particles of gel primary to the network. A representative model of this sequence of events is presented in Fig. 4a.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig4_HTML.gif
Fig. 4

Models of structures formed during hydrolysis and condensation sol–gel reactions by routes 1 (a) and 2 (b)

Route 2 occurred at pH value close to 7.0 where the processes of growth and aggregation of particles occur together and are indistinguishable. Since the silica solubility is low at this pH, the growth of particles stops when they have sizes between 2 nm to 4 nm, where the solubility is greatly reduced. Thus, the gelation occurs without the network structured, with only an aggregation of particles, as illustrated by Fig. 4b. When the glass is sintered, these molecules bind, forming a three-dimensional network [17]. It should be remembered that in the production of hybrid foams it is not possible to perform a heat treatment at high temperatures, because it would cause the PVA decomposition, which is part of the hybrid. This may be the reason why the foams produced by route 2 resulted in much weaker samples. The absence of Raman bands associated with three-dimensional network reinforces this analysis. Moreover, at the time of 48 h, there was still a large amount of TEOS and species not condensed, showing that the reaction had lower income.

Route 2 was not a route of synthesis better than route 1, because it resulted in very fragile samples and with bigger synthesis times. Consequently, route 2 has been discarded as an alternative route of synthesis for the production of neutral bioactive glass/PVA hybrid foams. Thus, the neutralization tests had to be studied to allow the production of biocompatible samples.

3.2 Chemical structure

The analysis of silicate structures by infrared spectroscopy (FTIR) has been reported by several research groups in literature, which were used as reference [12, 17, 24, 26, 32, 37, 4244, 4968]. Figure 5 shows the IR spectra of the hybrid foams produced by acid catalysis, route 1, with 20–60% of PVA, in addition to the spectra of bioactive glass and PVA. The spectra for hybrids containing 30% and 50% of PVA were omitted to improve the view of the bands in other spectra.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig5_HTML.gif
Fig. 5

FTIR spectra obtained for hybrid foams with PVA contents of 20–60%. The spectra of bioactive glass and PVA are presented for comparison

The band between 3,800 cm−1 and 3,000 cm−1 was observed in all the spectra and it was assigned to stretching vibration of several hydroxyl groups. This band was composed by the overlap of the SiO–H stretching vibration at: 3,750 cm−1, isolated SiO–H vicinal stretching; 3,660 cm−1, stretching of H connected and/or internal SiO–H stretching; 3,450 cm−1, SiO–H stretching in the surface silanols groups that forms Hydrogen Bonds with water molecules; 3,430–3,420 cm−1, CO–H stretching in PVA chains and ethanol; 3,500–3,400 cm−1, O–H stretching in water molecules. For O–H stretching vibrations, there was also the band at 1,630 cm−1, assigned to deformation mode of the adsorbed water molecules.

The bands at 1,200 and 1,090 cm−1 were associated with LO and TO vibration modes, respectively, of asymmetric stretching in several cyclical species, species with 4–8 Si atoms and the network. The 964 cm−1 band was assigned to Si–OH vibration mode, attributed to the presence of Ca2+ ions. The band in the range from 830 cm−1 to 760 cm−1 was the result of overlapping of the bands at 806 cm−1, related to the Si–O–Si symmetric stretching and vibration modes of rings.

The band at 830 cm−1, was assigned to the Si–O–C symmetric stretching in not hydrolyzed etoxy groups that remained in hybrids. Based on the increase in the intensity of this band compared to the bioactive glass, it may indicate the formation of the Si–O–PVA–O–Si bond. It should be pointed that the increase in the intensity of the band at 830–760 cm−1 was only an indicative of the incorporation of PVA in the network, considering that the increase in the 830 cm−1 band occurred because of the increase in PVA content. Moreover, it was considered that the intensity associated at Si–O–Si in the 806 cm−1 band, should not increase because of the glass fraction in the hybrid composition decreases.

The band at 2,940 cm−1, observed in hybrids and in PVA, was assigned to C–H asymmetric stretching in –CH2– groups. The presence of this band in hybrids indicates came from the PVA chains. The 1,430 cm−1 band in the PVA was associated to C–H scissor deformations in –CH2– groups. The band related to this vibration in the hybrid moved to 1,410 cm−1 because of the overlap with the 1,400 cm−1 band assigned to Si–O–Si bending mode.

From the results obtained by FTIR and Raman spectra analysis, it was possible to propose a model for the structure of the three-dimensional network formed, Fig. 6, based on models proposed in the literature [17, 24, 41, 42, 44, 69]. The TEOS hydrolysis under acidic conditions produced several species involving the replacement of etoxy (–OCH2CH3) by hidroxyl (–OH) groups. These species may have degrees of hydrolysis between 0 (TEOS) and 4 (fully hydrolyzed). The intermediate species, partially hydrolyzed, were involved in condensation reactions, forming dimmers, trimmers, tetramers. These species evolved to cyclical condensed structures, with 2–6 atoms of silicon in each ring. The structures tend to evolve into rings with more than 6 members, and, finally, for the three-dimensional network.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig6_HTML.gif
Fig. 6

Models proposed for the structures formed during the production of bioactive glass/PVA hybrids

After the CaCl2 addition to the sol, we suggest that there was an increase in the number of Si–OH bonds, formed by the collapse of the Si–O–Si bonds and/or ion exchange by the cations presence, resulting in Si–O–Ca–O–Si structures [7, 48].

The microstructure of the bioactive glass/PVA hybrid may result of two possible forms of incorporation of PVA particles to the gel: (i) the chains of PVA occupy the gel micropores; (ii) the condensation reaction of Si–OH with C–OH groups of PVA could occur by acid catalysis, forming various crosslinks in the hybrid, Si–O–PVA–O–Si [67].

3.3 Chemical composition

The thermal degradation kinetics of the hybrids, PVA and bioactive glass are presented in Fig. 7. In TGA curves, the mass loss was small for bioactive glass. In hybrids, the mass loss started due to water evaporation and organic waste decomposition (50 and 150 °C), followed by a pronounced loss because of the PVA decomposition (200 and 800 °C). Considering that the remaining mass after the temperature of 800 °C corresponds to the mass of bioactive glass in the hybrids, the composition of glass and PVA could be estimated based on these values (Table 2). From these estimations, it can be seen that the nominal composition of hybrids with up to 40% of PVA was above the estimated PVA fraction, and under valued in relation to the glass fraction. This difference probably was due to the fact that the sol–gel reaction was not complete, leaving groups not bonded to the hybrids (byproducts of hydrolysis and condensation, besides not reacted TEOS).
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig7_HTML.gif
Fig. 7

TGA curves of hybrid foams. The curves of the bioactive glass and PVA were presented as references

Table 2

Estimated composition of the hybrids based on TGA analysis

Nominal composition (PVA %)

20%

30%

40%

50%

60%

Bioactive glass (%)

56

52

48

42

39

PVA (%)

27

32

38

43

49

%Water/Waste

17

16

14

15

12

Bioactive glass/PVA composition

67/33

62/38

56/44

49/51

44/56

3.4 Acid character control

Figure 8 shows some hybrid foams produced. The foams presented porosity of 60–85% (measured by Archimedes Method) and pore diameters of 100–500 μm with interconnected structure. After drying these foams were highly acidic and the neutralization methods described previously were used to obtain samples with neutral pH. The best neutralization method was determined after evaluation of the loss of calcium, the sample final pH, PVA loss and time of drying. To determine these parameters, the techniques of EDS, TGA, and FTIR were used, in addition to pH measurements.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig8_HTML.jpg
Fig. 8

MEV images, increased 30x, of hybrid foams with PVA contend of (a) 20 and (b) 60%

During the neutralization tests with sodium hydroxide and deionized water, it was observed, by EDS analysis (not shown) that there was loss of calcium. The analysis of samples submitted to neutralization with calcium acetate solutions indicated that the hybrids incorporated calcium in their structures.

The pH of the foams before and after the neutralization tests was measured and the results are shown in Table 3. All samples had neutralized pH values close to 7, except the samples neutralized only with deionized water.
Table 3

pH of neutralizing test solutions, before and after foam immersion and indirect measurement of sample final pH in deionized water

Neutralizing tests

Before

After

Water

NH4OH (aqueous)

10.1

8.9

7.6

NH4OH (alcoholic)

10.1

8.7

7.1

Ca(CH3COO)2 (aqueous)

7.2

7.0

7.2

Ca(CH3COO)2 (alcoholic)

7.7

7.6

7.1

Deionized water

7.0

2.2

2.6

FTIR analysis was used to monitor the changes in the hybrid structures in each neutralization solution, Fig. 9. The intensities were normalized dividing the respective bands by the intensity of a reference band. The spectra of foams neutralized in aqueous and alcohol solutions of ammonium hydroxide and deionized water showed that the band between 3,800 cm−1 and 3,000 cm−1 had its intensity decreased, which could be related to water and/or hydrolyzed species reduction. It can be noted that the neutralizing tests caused a reduction of the band intensity at 2,940 cm−1, that may an indication of PVA loss. There was also a decrease in the band assigned to Ca2+ ions, 964 cm−1, in the spectra of the sample neutralized in ammonium hydroxide and deionized water solutions. This band had its intensity increased in the spectra of samples neutralized in calcium acetate solutions.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig9_HTML.gif
Fig. 9

FTIR Spectra obtained from foams with 20% of PVA, without neutralization and after the neutralization tests

PVA loss due to the neutralization procedures used was estimated by TGA analysis for each hybrid composition. Figure 10 shows the curves obtained for 20% PVA hybrid foams. A larger water loss was observed compared to not neutralized samples due to the small drying time used for these samples. The mass loss in the range 50–150 °C is attributed to evaporation of water and decomposition of organic residues (TEOS and ethanol). In the range 200–800 °C the mass loss is attributed to PVA decomposition.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig10_HTML.gif
Fig. 10

TGA curves of hybrid foams subjected to the neutralization tests in different solutions: deionized water; calcium acetate aqueous and alcoholic solutions; ammonium nitrate aqueous and alcoholic solutions

Table 4 shows the relative values of PVA loss in samples submitted to neutralization tests. PVA loss was calculated by the difference between the samples before and after each neutralizing test, in the temperature range of 200–800 °C. The method that had larger PVA loss percentage was neutralization in deionized water. The method that presented lower PVA loss was neutralization in calcium acetate aqueous solution.
Table 4

PVA relative mass loss after neutralization of hybrid foams in different solutions

Hybrid (% PVA)

PVA mass loss (%)

Neutralization solutions

H2O

NH4OH (aq.)

NH4OH (alc.)

Ca(CH3COO)2 (aq.)

Ca(CH3COO)2 (alc.)

20

−7.1

−5.9

−4.3

−3.9

−5.5

30

−8.4

−7.1

−5.1

−4.6

−6.5

40

−10.0

−8.3

−6.1

−5.5

−7.7

50

−11.3

−9.4

−6.9

−6.2

−8.8

60

−12.9

−10.7

−7.8

−7.1

−10.0

The time necessary to complete drying of the samples after neutralization was also a parameter measured. Samples neutralized in aqueous solution and deionized water had the drying time of 1 week in oven at 40 °C with air circulation, then 48 h under vacuum. The samples submitted to alcoholic neutralization solutions had their drying times reduced to 2 days, in oven, then 12 h, under vacuum. Table 5 shows the results of the apparent porosity of the foams after neutralization and drying. It was observed that after neutralization the porosity decreases, but the reduction is smaller, between 2% and 5%, in foams that were neutralized in alcoholic solutions. It can be observed by MEV images of foams obtained before and after the neutralizing with calcium acetate solution, Fig. 11. After these neutralizing, the foams presented crystals of, probably, calcium chloride, analyzed by EDS.
https://static-content.springer.com/image/art%3A10.1007%2Fs10971-008-1777-1/MediaObjects/10971_2008_1777_Fig11_HTML.jpg
Fig. 11

MEV images, increased 100×, of hybrid foams with 20% PVA contend, (a) before and (b) after calcium acetate neutralization method

Table 5

Apparent porosity of dried foams before and after neutralization, measured by Archimedes method

Foams %PVA

Apparent porosity (%)

Aquous solutions

Alcoholic soutions

Before

After

Variation

Before

After

Variation

Glass

87.9

79.0

10.1

87.9

83.2

5.3

20

81.8

74.1

9.5

81.8

77.1

5.8

30

77.0

69.8

9.3

77.0

74.3

3.4

40

73.0

66.7

8.6

73.0

70.5

3.5

50

68.0

62.7

7.8

68.0

65.8

3.2

60

66.4

61.4

7.5

66.4

65.0

2.1

Taking into account pH, composition change and sample contraction we conclude that immersion in calcium acetate alcoholic solution is the most interesting neutralization method to avoid calcium loss (actually, occurred calcium incorporation) and to facilitate the drying without major contractions, resulting in foams with final pH near 7.0, and PVA loss of 5–10%.

4 Conclusions

The obtained foams presented porosity of 60–85% and pore diameters of 100–500 μm with interconnected structure. The hybrids obtained by a neutral pH sol–gel route were weaker and had higher synthesis times. This route was discarded as an alternative route of synthesis for the production of neutral hybrid foams. From the FTIR and Raman spectra analysis, it was possible to propose a model for the structure of the three-dimensional network of the bioactive glass/PVA hybrids. Calcium acetate alcoholic solution was the best neutralizing method, resulting in foams with final pH of about 7, 0, and loss of PVA from 5% to 10%. The apparent porosity of foams decreased between 2% and 5% after neutralization.

Acknowledgments

The authors acknowledge National Counsel of Technological and Scientific Development, The State of Minas Gerais Research Foundation and CAPES for financial support on this project.

Copyright information

© Springer Science+Business Media, LLC 2008