SEARCH

SEARCH BY CITATION

Keywords:

  • foam;
  • bioactive glass;
  • scaffold;
  • sol-gel macroporous

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

Resorbable 3D macroporous bioactive scaffolds have been produced for tissue-engineering applications by foaming sol-gel–derived bioactive glasses of the 58S (60 mol% SiO2, 36 mol% CaO, 4 mol% P2O5) composition with the aid of a surfactant. Bioactive glasses are known to have the ability to regenerate bone, and to release ionic biological stimuli that promote bone-cell proliferation by gene activation. The foams exhibit a hierarchical structure, with interconnected macropores (10–500 μm), which provide the potential for tissue ingrowth and mesopores (2–50 nm), which enhance bioactivity and release of ionic products. Many factors in the sol-gel and foaming processes can be used to control these pore sizes and distributions. This work concentrates on the effect of the processing temperature, gelling agent concentration, and the amount of water used for the foam generation on the structure, pore morphology, and the properties of the foam scaffold. The simplest and most reproducible method for controlling the modal pore diameter was by the amount of water added during the foaming process. The in vitro dissolution and bioactivity of the bioactive foams were compared to that of unfoamed monoliths and powders (< 20 μm in diameter) of the same composition. © 2003 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater 68B: 36–44, 2004


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

Due to the shortage of organ or tissue donors and difficulties with immunoresponse, there is great demand for tissue-engineered replacements.1 Tissue-engineering techniques may employ the use of scaffolds that can be seeded with cells in vitro to develop the basis of a replacement tissue, which is then implanted.2

Certain compositions of bioactive glasses containing SiO2-CaO-P2O5 bond to both soft and hard tissue without forming scar tissue.3 Results of in vivo implantation show that these compositions are nontoxic.4 Melt-derived bioactive glasses such as Bioglass® have been used clinically as treatment for periodontal disease and bone defects and are available commercially in powder form.5 The bioactivity has been associated with the formation of a crystalline hydroxy carbonate apatite (HCA) surface layer, similar in structure to the inorganic region of bone, on contact with body fluid.5 Recently, Xynos et al.6 showed that bioactive glass dissolution products cause rapid expression of genes that regulate osteogenesis and the production of growth factors. These discoveries have stimulated extensive investigations for using bioactive glass as scaffolds for tissue engineering.

An ideal scaffold should combine the beneficial properties of bioactive glasses with a structure consisting of an interconnected network with macropores (greater than 100 μm) to enable tissue ingrowth7 and nutrient delivery to the center of the regenerated tissue and mesopores (2 nm < pore size < 50 nm), which may promote cell adhesion. The scaffold should also resorb at controlled rates to match that of tissue repair and be made from a processing technique that can produce irregular shapes to match that of the defect in the bone of the patient. The foaming of sol-gel–derived bioactive glasses provides the potential to make such a scaffold.7

Sol-gel derived glasses have been found to exhibit enhanced resorbability and bioactivity in vitro9–11 and improved bone bonding in vivo compared to melt-derived bioactive glasses.12 This has been attributed to the gel glasses exhibiting a mesoporous texture (pores in the range 2–50-nm diameter), which is inherent to the sol-gel process and which causes an increase in the specific surface area.13–15 The porous texture, and therefore the bioactivity and resorbability, of the gel glasses can be tailored by controlling the composition and processing temperature of the sol-gel process.16

Each stage in the sol-gel foaming process (Figure 1) affects the structure and properties of the foam scaffolds and can be used to obtain specific architectures, to produce specific pore-size ranges, and to control rates of glass dissolution. Such process variables include the temperature at which the foaming process is carried out, surfactant type and concentration, gelling agent type and concentration, added water concentration, and glass composition. This work investigates how the process variables affect the foam volume produced and hence the pore structure, textural characteristics and glass dissolution rates of the ternary 58S (60 mol% SiO2, 36 mol% CaO, 4 mol% P2O5) composition, which is thought to be the most bioactive of the sol-gel–derived bioactive glasses.17

thumbnail image

Figure 1. Flow diagram of sol-gel foaming process.

Download figure to PowerPoint

If the macropores exhibit interconnected diameters of less than 100 μm, tissue ingrowth and vascularization will not occur efficiently in vivo and the damaged tissue cannot be fully regenerated. Changes in the macroporosity and textural porosity of the foams, due to changes to the foaming process, may lead to changes in the dissolution (resorption) rate and bioactivity (rate of HCA layer formation) of the scaffold. A change in dissolution rate of the glass would also affect the number of ionic products released by the glass that can act as genetic stimuli. It is important to be able to control the dissolution rate of a scaffold so that it can be matched to the rate of regeneration of the tissue and supply the required dosage of ions for genetic stimulation.2

The aim of the work was to optimize the foaming process and develop a protocol to produce scaffolds with interconnected pores with diameters in excess of 100 μm, maximum handling strength, and controlled resorbability in vitro.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

Foams of 58S (60 mol% SiO2, 36mol% CaO, 4 mol% P2O5) composition were produced with the use of the process shown in Figure 1.

58S Sol Preparation

Sol-gel precursors tetraethoxyl orthosilicate [TEOS, Si(OC2H5)4], triethoxyl orthophosphate [TEP, OP(OC2H5)3], and calcium nitrate Ca(NO3)2 · 4H2O were mixed in D.I. (deionized) water in the presence of 2N nitric acid (HNO3), a catalyst for hydrolysis.16 Simultaneous hydrolysis and polycondensation reactions occur to begin formation of a silica network.

Foaming

Viscosity of the sol increased as the condensation reaction continued. On completion of hydrolysis, aliquots of 50 ml sol were foamed by vigorous agitation with the addition of 1.5-ml surfactant (Teepol, a detergent containing a low-concentration mixture of anionic and nonionic surfactants), D.I. water (improves foamability of surfactant), and 5 vol% hydrofluoric acid (HF, acatalyst for polycondensation). The foaming process was performed in a transparent polypropylene graduated beaker (HF resistant).

The surfactant stabilized the bubbles that were formed by air entrapment during the early stages of foaming by lowering the surface tension of the solution.18, 19 As viscosity rapidly increased and the gelling point was approached the solution was cast into airtight molds. The gelling point is the point at which the meniscus of the foamed sol does not move, even if the mold is tilted.20 Casting must take place immediately prior to the gelation point. The gelation process provided permanent stabilization for the bubbles.

The foams were then aged at 60 °C for 72 h, dried at 130 °C for 48 h and thermally stabilized at 600 °C for 22 h, according to established procedures.16

From the results of a pilot study (not shown), addition of 1.5 ml of Teepol, 1.5 ml water, and 1.5 ml HF and foaming at 25 °C produced foams with pores connected by windows of up to 100 μm diameter. This was the control foam recipe used for the 58S composition in this study. In the following work the recipe from the pilot study will be referred to as the control recipe. The temperature, concentration of added water, and the concentration of HF in the solution were varied independently to obtain foams of different porosities. At least three separate batches were produced for each recipe to ensure reproducibility.

Temperature

Aliquots of 50 ml 58S sol were foamed with the use of 1.5 ml Teepol, 1.5 ml added water, and 1.5 ml HF at temperatures of 20, 25, 29, and 35 °C. The foaming temperature was thermostatically controlled with the use of a water bath. The volume of foam produced from 50 ml sol (foam volume) was measured from the graduated beaker 20 s prior to the gelling point and the gelling time was recorded.

Gelling Agent

Aliquots of 50 ml 58S sol were foamed at 25 °C, using 1.5 ml Teepol, 1.5 ml added water and 1, 1.5, 2.5, 3.5, and 4 ml HF. The foam volume and gelling time were recorded.

In order to investigate the effect of HF on the textural pore (mesopore) size, two sets of 58S monoliths were prepared with 1.5 ml Teepol per 50 ml sol so that their composition was the same as the foams, but no agitation was carried out. One set of monoliths was gelled with the use of 1.5 ml HF per 50 ml sol, and the other set was allowed to gel over 3 days with no HF. The textural properties of each set of monoliths were compared using the nitrogen sorption techniques.

Added Water

Aliquots of 50 ml 58S sol were foamed at 25 °C; with the use of 1.5 ml Teepol; 1.5 ml HF; and 0, 1, 1.5, 2.5, 3.5, and 4.5 ml of deionized water, added as foaming commenced. The foam volume and gelling time were recorded.

Characterization

The geometrical bulk densities of the foam scaffolds were calculated from dimensional and mass measurements. Archimedes' principle could not be used, as the foams have a lower density than water and begin to undergo dissolution reactions in solution.

The foams were characterized with the use of field-emission scanning electron microscopy (Leo 1525) with 2-kV accelerating voltage and mercury intrusion porosimetry (PoreMaster 33, Quantachrome) to measure interconnected macropore21 size distributions. Mercury porosimetry only measures pores that are interconnected; therefore, the mode of the interconnected macropore distributions was used as a guide to the interconnectivity of the foams and their potential for use in tissue- engineering applications.

Monoliths were analyzed with the use of nitrogen adsorption (Autosorb AS6, QuantaChrome) to establish specific surface area (BET) and mesopore size distributions (B.J.H.).22

Compositional Analysis

In order to investigate whether the composition of the control recipe foam was similar to that of the 58S composition, lithium metaborate fusion was carried out at 1100 °C. After cooling, 80 ml 10 vol% HNO3 and the solution was mixed with the use of a magnetic stirrer, until all the fused solid had dissolved. Induced couple plasma (ICP, Thermo FI ARL 3580 B) analysis was then run on the solution for Ca, P, and Si. Three measurements per element were recorded in ppm and a mean value taken. The weight percentage of each component could then be calculated, assuming that no other components were present in the foam.

Three control recipe foams were scanned for the presence of fluoride ions with the use of magic angle spinning nuclear magnetic resonance (MAS-NMR).

Dissolution and Bioactivity Tests

The dissolution and bioactivity of foams was compared to that of unfoamed monoliths and powders (mean particle size of 10 μm23) with the use of simulated body fluid (SBF) electrolyte solution at a starting pH of 7.25 at 37 °C. The 58S gel-glass monoliths were prepared by conventional methods16 without the use of catalyst, surfactant, or agitation. The powders were supplied by USBiomaterials (Alachua, FL).

A previous study found that bioactive glass powders exhibited maximum bioactivity at a glass concentration of 1.5 mg l−1.24 Therefore 0.075 g of foam, monolith, or powder were immersed in 50 ml of SBF and placed in an orbital shaker at 37 °C, for 30 min, 1, 2, 5, 22, and 72 h, at an agitation rate of 175 Hz. Three samples were run per test, with the average values reported.

Extracts obtained by filtration (1 μm paper) were analyzed by ICP (inductive coupled plasma spectroscopy, Thermo FI ARL 3580 B) for Si, Ca, Na, and P concentration in solution. The instrument detection limits for Si, Ca, Na, and P were 0.05, 0.10, 0.10, and 0.20 ppm, respectively. The filtrated powder was rinsed with acetone to terminate any ongoing reactions, dried, and then evaluated by X-ray diffraction (XRD) and Fourier-transform infrared (FTIR) analyses for the presence of calcium–phosphate phases, and formation of a hydroxy-carbonate apatite (HCA) layer. Absorbance FTIR spectra were collected with the use of a Mattson Genesis II spectrometer, with a Pike Technologies EasiDiff diffuse reflectance accessory in the range 400–1500 cm−1.

Error Bars

The error bars on all graphs represent ± 1 standard deviation from the mean value of tests performed in at least triplicate.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

Compositional Analysis

The composition of the 58S foam produced with 1.5ml of Teepol, added water, and HF was 61.1 wt% SiO2, 29.2 wt% CaO, and 9.7 wt% P2O5, compared to the theoretical composition of 58 wt% SiO2, 33 wt% CaO, and 9 wt% P2O5. No fluoride ions were detected in the foams from the NMR scan, which implies that all of the HF was removed during thermal processing.

Effect of Temperature

58S foams of the control recipe were produced at foaming temperatures of 20, 25, 29, and 35 °C, where foaming temperature is defined as the temperature of the water bath in which the sol was foamed. Figure 2(a) shows a graph of gelling time as a function of foaming temperature (Tf). Figure 2(a) shows that as foaming temperature was increased from 20 to 35 °C, the gelling time decreased from 11 min 10 s to 6 min 20 s. All gelling times were reproducible with a ± 5% range. The gelling time is thought to have decreased because the condensation rate increased as a result of the higher temperature. Figure 2(b) shows that foam volume as a function of foaming temperature followed a similar relationship, decreasing from approximately 180 ml at 20 °C to 70 ml at 35 °C. This corresponded to bulk density (ρb) values of 0.30 g cm−3, when foamed at 20 and 25 °C, which increased to 0.35 g cm−3 at 29 °C and 0.40 g cm−3 at 35 °C. All foams exhibited a shrinkage of 65–75% during thermal processing.

thumbnail image

Figure 2. Graphs of (a) gelling time as a function of foaming temperature, and (b) foam volume as a function of gelling time for 58S gel-glass.

Download figure to PowerPoint

Figure 3 shows macropore distributions, attained by mercury porosimetry from foams produced at each foaming temperature. The vertical axis (− dV/d log D) is a differential of the volume of mercury intruded (V) at each interconnected pore diameter (D). A full derivation can be found in the literature.25 All pore distributions were wide, implying that foams produced at each of the four temperatures contained some pores greater than 200 μm (limit of the mercury porosimeter). Scaffolds foamed at temperatures of 20 and 25 °C, exhibited approximately normal pore distributions, with modal pore diameters of 95 μm (ρb = 0.30 g cm−3). Above 25 °C pore distributions followed a positive skew and the amount of skew increased as temperature increased, with a modal pore diameter of approximately 35 μm at 29 °C (ρb = 0.35 g cm−3) and 22 μm at 35 °C (ρb ° 0.40 g cm−3). The mode of the pore distribution is of most interest for tissue engineering applications as it represents the pore size that was most frequent in the scaffold.

thumbnail image

Figure 3. Macropore distributions (mercury intrusion porosimetry) as a function of foaming temperature for 58S gel-glass.

Download figure to PowerPoint

Figure 4(a) shows an SEM micrograph of a scaffold, foamed at 25 °C to obtain a foam volume of 110 ml [Figure 2(b)]. The foam exhibited a crack-free pore network with spherical pores with diameters up to 600 μm and interconnected pores of up to 100 μm in diameter. It is these interconnections that are vital for vascularization and tissue ingrowth. Figure 4(b) shows an SEM micrograph of a scaffold that has been foamed at 35 °C, which exhibited small isolated pores, was highly cracked and was fragile to the touch.

thumbnail image

Figure 4. SEM micrographs of 58S scaffolds foamed at (a) 25 and (b) 35 °C.

Download figure to PowerPoint

Effect of Gelling Agent Concentration

Figure 5 shows a graph of gelling time as a function of gelling agent (HF) concentration. Unsurprisingly, the gelling time decreased as gelling agent concentration increased; however, the foam volume was approximately constant as gelling time decreased and therefore the bulk density and macropore size of the scaffolds were unaffected.

thumbnail image

Figure 5. Gelling time as a function of HF concentration for the foaming of 58S gel glass.

Download figure to PowerPoint

58S monoliths containing no HF exhibited a specific surface area of 132 m2 g−1, a pore volume of 0.81 ccg−1) and a modal pore diameter (BJH) of 17.5 nm. 58S monoliths containing 1.5 ml HF per 50-ml sol exhibited a specific surface area of 120 m2 g−1, a pore volume of 0.80 ccg−1, and a modal pore diameter (BJH) of 12.5 nm. The specific error for specific surface area was ± 12 m2 g−1, for pore volume was ± 0.5 ccg−1, and for modal mesopore diameter was ± 0.6 nm. The only parameter that changed on HF addition was that the modal pore diameter decreased from 17.5 to 12.5 nm.

Effect of Added Water Content

Figure 6 shows a graph of foam volume as a function of the quantity of added water. As added water content increased the foam volume increased approximately linearly to 170 ml at a water content of 3.5 ml, and therefore modal pore size increased (130 μm at 3.5 ml water). As water content increased from 3.5 to 5 ml the foam volume increased to a greater extent, reaching 400 ml at 5 ml water; that is, eight times the original sol volume. Such a volume increase produced very large air bubbles in the sol, and therefore high interconnectivity. However, the struts between the pores (cell walls) were so thin that the foam could not support its own weight and it collapsed on gelation. A maximum of 170 ml of foam could be produced from 50 ml of 58S sol, as is shown by the foam survival limit line in Figure 6.

thumbnail image

Figure 6. Foam volume as a function of added water content for the foaming of 58 gel glasses.

Download figure to PowerPoint

Dissolution and Bioactivity Tests

Figure 7 shows an SEM micrograph of a 58S foam produced with the use of the control recipe. Figure 7 shows that such a foam exhibited a highly interconnected crack-free macroporous network. The macropores were approximately spherical and were as large as 600 μm in diameter. The windows connecting the pores were spherical and were in the range 10–200 μm, most being approximately 80–120 μm in diameter, which agrees with the pore distributions observed in Figure 3. Images of the 58S powders can be seen in previous work.23

thumbnail image

Figure 7. SEM micrograph of a 58S gel-glass scaffold foamed with the use of 1.5-ml Teepol, 1.5 ml HF, and 1.5 ml H2O.

Download figure to PowerPoint

Figure 8 shows the dissolution profiles (from ICP analysis) of the 58S foam (control recipe), 58S powders and 58S monoliths. Figure 8 shows that 58S powders exhibited the most rapid silicon dissolution rates, increasing from 0 to 36 ppm in 30 min and then to 70 ppm after 8 h, where silicon content of the SBF reached saturation. The foams released just 10 ppm in the first 30 min of immersion and saturated the silicon content of the SBF after 22 h. From 22 to 72 hours immersion, the silicon concentration was identical for the foams and powders; that is, the SBF was saturated for both cases.

thumbnail image

Figure 8. Dissolution profiles of 0.075 g 58S foam, powder, and monolith in 50-ml SBF at 37 °C under 175-rpm agitation.

Download figure to PowerPoint

The calcium ion concentration for immersed powders increased from 84 ppm (amount in unreacted SBF) to 200 ppm in the first 30 min of immersion, compared to 143 ppm for the foam. The calcium ion content then increased very slowly to 230 ppm and 281 ppm at 3 days immersion for the powders and foams, respectively.

The amount of phosphate in solution decreased with time, indicating formation of the HCA layer.15 Due to the more rapid dissolution of the powders, the amount of phosphate removed from solution was also greater for the powders than the foams at short times, but after 22 h, most of the phosphate was removed from the SBF by both powders and foams.

The 58S monoliths underwent the slowest dissolution. During 72 h of monolith immersion the concentration of silicon increased from 0.3 ppm in SBF to just 6 ppm, and calcium-ion concentration increased from 107 to 170 ppm. The phosphate concentration in solution decreased from 31 to 28 ppm, a decrease of just 3 ppm.

Figure 9 shows the FTIR spectra for a 58S powder, foam and monolith after 2 h immersion in SBF. The characteristic P-O bend vibrational bands at 570 and 600 cm−1 that indicate formation of the HCA surface layer11 were of higher intensity for the powders and of lowest intensity for the monolith.

thumbnail image

Figure 9. FTIR spectra for 58S powder, foam, and monolith after 2-h immersion in SBF at 37 °C and 175-rpm agitation.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

Effect of Temperature

The relationship between gelling time and foaming temperature [Figure 2(a)] can be fitted to a first-order exponential decay (r2 = 0.9998). It is difficult to completely separate hydrolysis and condensation during gelation, as they tend to occur simultaneously. Colby et al.26 assumed that hydrolysis is completed after mixing of TEOS with the water as a first approximation, leaving the condensation reaction as the rate-determining process. Hydrolysis is an exothermic reaction. By measuring the temperature of the sol in this work in situ the temperature of solution was found to increase to a peak after 7 min of mixing TEOS and water. This implies that hydrolysis was completed soon after mixing and the condensation reaction is the rate-determining step for gelation. The condensation reaction is complex, but if the gelling time (tgel) is the time over which viscosity has increased from approximately 10−1 to 104 P, then tgel can be considered as an averaged rate of gelation. Therefore the exponential decay shown in Figure 2(a) can be fitted to an Arrhenius equation:

  • equation image

where A is an Arrhenius constant, R is the gas constant and E* is an apparent activation energy.

Therefore, there is an apparent activation energy barrier that has to be overcome for the condensation reaction to complete and for gelling to occur. The activation energy for gelation was calculated from the slope of a plot of ln[tgel] as a function of 1/Tf and was found to be 10.3 kcal/mol. As temperature is increased the kinetics of the condensation reaction is increased (more monomers come into contact with each other), reducing the gelling time. Activation energy for gelling is affected by the pH and composition of the sol.

As foaming temperature increased, gelling time decreased, which meant that agitation time decreased, and therefore foam volume and macropore diameter decreased. This agrees well with the work of Colby et al. work on an HF catalyzed condensation of TEOS in an ethanol mixture.26 They found that the gelation time decreased from 9.2 h to 30 min as temperature increased from 25 to 70 °C. An apparent activation energy value of 13.2 kcal/mol was attained for HF catalysis. Silva and Vasconcelos27 also found an activation energy of 11.8 kcal/mol for silica gels catalyzed by HF at different temperatures.

Gelling Agent

The decrease in gelling time as gelling agent concentration was increased is in agreement with work by Silva and Vasconcelos27 on the gelation of sol-gel-derived silica with the use of HF catalysis. However, they found that the mean diameter of mesopores increased from 7 to 21 nm as HF concentration increased from 0.02 to 0.12 M, but the surface area decreased and the permeability increased. In contrast, the modal mesopore diameter of the monoliths in this study decreased from 17.5 to 12.5 nm as HF concentration increased from 0 to 1.5 ml per 50 ml sol. The surface area decreased from 130 to 120 ccg−1, which is a very small change and within the accuracy limits of nitrogen gas sorption techniques. The increase in pore diameter with HF concentration in the work of Silva and Vasconcelos, which was not observed in this work, could be that they used a single-component system (silica), whereas a tertiary system was used in this work. Therefore, with regard to tissue-engineering applications, the HF concentration is not critical to the 58S composition, as long as it is low enough to allow time for a foam to be created (approximately 3 min, when 1.5 ml water and 1.5 ml Teepol are used).

Added Water

The marked effect of water on the foam volume produced is because water is the critical component necessary for film formation. Unreacted water is necessary for foam production because the hydrophilic portions of surfactant molecules attach to free water, allowing film formation when air is introduced and stabilizes the foam.18, 19 The R ratio of the sol (TEOS: water ratio) defines the amount of free water available for foaming. In this work, excess water was used in the hydrolysis of TEOS (sol preparation), which means there was always some free water available for foaming. Adding water directly to the sol on addition of surfactant increased the local R value and improved efficiency of the surfactant.

Dissolution and Bioactivity Tests

The dissolution rate of bioactive glasses depends on glass composition, specific surface area, type of dissolution medium, pH, and geometry (e.g., particle size).15 The foams and the powders each exhibited specific surface areas of ∼155 ccg−1.23 Therefore the geometry is the only parameter that was different on immersion. The foam exhibited a lower dissolution rate compared with the powder because it had less surface area in contact with the solution at one time compared to the powders, as the powders moved at random throughout the solution under the agitation. The foam acted as a cluster of powders. The dissolution rate of the monolith was due to an even lower surface area in contact with the SBF; that is, it did not exhibit the many surfaces of the particles, or the macropores of the foam.

The dissolution rate of bioactive glasses affects the rate of formation of the HCA layer15 as confirmed by Figure 8.

Handling Strength of Foams

The processing variables affect the porosity of the foam, which has a direct effect on the mechanical properties of the scaffolds. Foam scaffolds produced at 35 °C were highly cracked [figure 4(b)] and were fragile on handling compared to foams produced at 25 °C, which were crack free and durable to handling [Figure 4(a)]. The cracked foams (Tf = 35 °C) exhibited small isolated pores with low interconnectivity (ρb = 0.40 g cm−3 and modal pore diameter was 22 μm) because a foam volume of just 65 ml was achieved [figure 2(b)]. Cracks are formed in the foam during the drying stage of the thermal processing (Figure 1). Pore liquor (by-products of the polycondensation reaction during gelation) must evaporate from within the gel network, resulting in capillary stresses. Capillary stresses increase as the interconnected pore diameter decreases. When capillary stresses exceed a critical value cracks form, nucleating at points of highest stress concentration, that is, pore edges. Foam volumes of at least 110 ml are therefore desired from a 50-ml sol in order to attain stable scaffolds with pores suitable for tissue engineering. A balance is required between obtaining pores large enough for tissue engineering applications and foams that do not crack during drying, for example, not producing pores so large that the foam is weakened by thin pore walls.

Recommended Protocol

The foaming temperature chosen for further investigations was therefore 25 °C, as this temperature produced foams of suitable modal pore size for tissue-engineering applications (i.e., pores with interconnected diameters in excess of 100 μm). The reason 20 °C was not chosen was that it is below ambient temperature for much of the year and therefore difficult to maintain between experiments. One can add 2.5 ml of water with 1.5 ml Teepol to produce stable foams with ρb = 0.25 g cm−3 and modal interconnected pore diameters of 120 μm. HF (2.5 ml) can be used to reduce processing time without affecting the macroporous structure of the scaffold.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

All the variables (foaming temperature, water content, and HF content) investigated in this study influence the porosity and structure of CaO-P2O5-SiO2 foam scaffolds. However, the one variable that produced the simplest control over pore size of the foams is the amount of water added to aid the surfactant. Small changes in added water concentration can be used at constant temperature and constant concentrations of surfactant and gelling agent to produce different pore networks at reproducible gelling times. Although the foaming process changes the sol-gel process slightly, compared to the production of monoliths and powders, the foams still exhibit a mesoporous texture characteristic of gel-derived glasses and are still highly bioactive.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES

The authors would like to thank USBiomaterials (Alachua, FL) for supplying the bioactive glass powders.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. REFERENCES