Manufacturing mechanical characterization and In vitro Performance of bioactive glass 13-93 Fibers E Pirhonen, H. Niranen, T. Niemela, 1 M. Brink, 2 P. Tormala' Tampere University of Technology, Institute of Biomaterials, Tampere, Finland 2 Sydvast Polytechnic, Malmgatan 5, Abo, Finland Received 11 March 2005: revised 21 June 2005: accepted 22 June 2005 shedonline28October2005inWileyInterscience(www.interscience.wileycom).Dol:10.1002/jbn.b Abstract: Fibers were manufactured from the bioactive glass 13-93 by melt spinning. The fibers were further characterized by measuring their tensile and flexural strength, and their in vitro performance was characterized by immersing them in simulated body fluid, which analyzed changes in their mass, their flexural strength, and surface reactions. The strength of glass fibers is highly dependent on fiber diameter, test method, and possible surface flaws, for example, cracks due to abrasion. In this study, the thinnest fibers(diameter between 24 and 33 um) possessed the highest average tensile strength of 861 MPa. The flexural strength nitially 1353.5 MPa and it remained at that level for 2 weeks. The Weibull modulus for both ensile and flexural strength values was initially about 2. 1. The flexural strength started to decrease and was only 20%o of the initial strength after 5 weeks. During the weeks 5-40, only a slight decrease was detected. The fiexural modulus decreased steadily from 68 to 40 GPa during this period. The weight of the samples initially decreased due to leaching of ions and further started to increase due to precipitation of calcium phosphate on the fiber surfaces. The mass change of the bioactive glass fibers was dependent on the surface area rather than initial weight of the sample The compositional analysis of the fiber surface after 24 h and 5 weeks immersion did confirm the initial leaching of ions and later the precipitation of a calcium phosphate layer on the bioactive glass 13-93 fiber surface in vitro. 2005 Wiley Periodicals, Inc J Biomed Mater Res Part B: Appl Biomater 77B: 227-233, 2006 Keywords bioactive glass; mechanica al properties; degradation; fiber NTRODUCTION phosphorus oxides, and silica. Brink et al. studied different glasses in a system containing boron, sodium, potassium, Tissue engineering has brought new challenges to the area of magnesium, calcium and phosphorus oxides, and silica. They biomaterials research. Especially in bone and cartilage tissue found that some of the glasses had a wide working range, engineering applications, there is a need for even more suit- which enables the production of fibers. One of the most able scaffold materials to be developed. Bioactive glasses promising glasses was the glass coded 13-93. Histological have been clinically used over three decades and have been studies by Brink et al. on glass 13-93 rods have shown that this found to possess an osteoconductive nature. Also an osteo- glass is bioactive in bone tissue and slowly resorbable in soft promotive performance of bioactive glass, namely Bio- tissue . In the early tests on bioactive glass 13-93, thin fibers glass@, has been reported. Therefore, bioactive glasses are were also produced by drawing from the melt. The in vivo considered advantageous compared with many other materi- performance of these fibers was tested by implanting fit als used as bone fillers today. For clinical purpose, the use of subcutaneously in rats. The resorbtion of fibers was initiated almost immediately, but no inflammation reaction was detect the use of bioactive glass as fibers has provoked lots of ed. 0 The use of bioactive glass as fibers does extend the variety interest amongst a number of research groups. Vita Finzi of possible implants. For example, there is Zalman et al. developed bioactive glass fibers with glass bioactive glass fibers as a reinforcing phase in biopolymers and system containing sodium, potassium, calcium, aluminum, in the manufacturing of porous scaffolds. 1.12 The aim of this study was to manufacture bioactive glass 13-93 fibers and study the mechanical properties of these orrespondence to: E Pirhonen, Inion Oy, Laakirinkatu 2, F1-33520 Tampere. fibers when immersed in simulated body fluid(SBF). The Contract grant sponsor: Academy of Finland mass loss and surface reactions of in vitro conditions were o 2005 Wiley Periodicals, Inc also studied
Manufacturing, Mechanical Characterization, and In Vitro Performance of Bioactive Glass 13–93 Fibers E. Pirhonen,1 H. Niiranen,1 T. Niemela¨ ,1 M. Brink,2 P. To¨ rma¨ la¨ 1 1 Tampere University of Technology, Institute of Biomaterials, Tampere, Finland 2 Sydva¨ st Polytechnic, Malmgatan 5, Åbo, Finland Received 11 March 2005; revised 21 June 2005; accepted 22 June 2005 Published online 28 October 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.30429 Abstract: Fibers were manufactured from the bioactive glass 13–93 by melt spinning. The fibers were further characterized by measuring their tensile and flexural strength, and their in vitro performance was characterized by immersing them in simulated body fluid, which analyzed changes in their mass, their flexural strength, and surface reactions. The strength of glass fibers is highly dependent on fiber diameter, test method, and possible surface flaws, for example, cracks due to abrasion. In this study, the thinnest fibers (diameter between 24 and 33 m) possessed the highest average tensile strength of 861 MPa. The flexural strength was initially 1353.5 MPa and it remained at that level for 2 weeks. The Weibull modulus for both tensile and flexural strength values was initially about 2.1. The flexural strength started to decrease and was only 20% of the initial strength after 5 weeks. During the weeks 5– 40, only a slight decrease was detected. The flexural modulus decreased steadily from 68 to 40 GPa during this period. The weight of the samples initially decreased due to leaching of ions and further started to increase due to precipitation of calcium phosphate on the fiber surfaces. The mass change of the bioactive glass fibers was dependent on the surface area rather than initial weight of the sample. The compositional analysis of the fiber surface after 24 h and 5 weeks immersion did confirm the initial leaching of ions and later the precipitation of a calcium phosphate layer on the bioactive glass 13–93 fiber surface in vitro. © 2005 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater 77B: 227–233, 2006 Keywords: bioactive glass; mechanical properties; degradation; fiber INTRODUCTION Tissue engineering has brought new challenges to the area of biomaterials research. Especially in bone and cartilage tissue engineering applications, there is a need for even more suitable scaffold materials to be developed.1,2 Bioactive glasses have been clinically used over three decades and have been found to possess an osteoconductive nature. Also an osteopromotive performance of bioactive glass, namely Bioglass®, has been reported.3–5 Therefore, bioactive glasses are considered advantageous compared with many other materials used as bone fillers today. For clinical purpose, the use of bioactive glasses has mainly been limited to particulates, but the use of bioactive glass as fibers has provoked lots of interest amongst a number of research groups. Vita Finzi Zalman et al. developed bioactive glass fibers with glass system containing sodium, potassium, calcium, aluminum, phosphorus oxides, and silica.6 Brink et al. studied different glasses in a system containing boron, sodium, potassium, magnesium, calcium and phosphorus oxides, and silica. They found that some of the glasses had a wide working range, which enables the production of fibers.7 One of the most promising glasses was the glass coded 13–93. Histological studies by Brink et al. on glass 13–93 rods have shown that this glass is bioactive in bone tissue and slowly resorbable in soft tissue.8,9 In the early tests on bioactive glass 13–93, thin fibers were also produced by drawing from the melt. The in vivo performance of these fibers was tested by implanting fibers subcutaneously in rats. The resorbtion of fibers was initiated almost immediately, but no inflammation reaction was detected.10 The use of bioactive glass as fibers does extend the variety of possible implants. For example, there is an interest to use bioactive glass fibers as a reinforcing phase in biopolymers and in the manufacturing of porous scaffolds.11,12 The aim of this study was to manufacture bioactive glass 13–93 fibers and study the mechanical properties of these fibers when immersed in simulated body fluid (SBF). The mass loss and surface reactions of in vitro conditions were also studied. Correspondence to: E. Pirhonen, Inion Oy, La¨a¨ka¨rinkatu 2, FI-33520 Tampere, Finland (e-mail: eija.pirhonen@inion.fi) Contract grant sponsor: Academy of Finland © 2005 Wiley Periodicals, Inc. 227
228 PIRHONEN ET AL. MATERIALS AND METHODS elemental analyses were performed, one on manufactured bioactive glass fibers and the other on the remnants of the ing of the Glas glass in the platinum crucible The bioactive glass 13-93 contains 6 wt Na,o, 12 wt %o Tensile Test. Tensile tests were performed on the newly K2O, 5 wt MgO, 20 wt Cao, 4 wt P, Os, and 53 wt drawn bioactive glass fibers, normally wit n davs after the 96 SiO 2. The raw materials used, analytical grade Na- CO3, manufacturing. The tensile test results reported in this study weighed and mixed in a plastic container. The mixed raw are results from various manufacturing batches. The ASTM materials were melted in a platinum crucible for 3 h at D3379-75 instructions were followed in the tensile testing, in hieve a homogeneous glass the main. A gauge length of 50 mm and a testing speed of 1 was crushed into mm/min were used. The Instron materials testing machine m' pieces and reheated for 3 h at(Series IX Automated Materials Testing System 1.38)with either a 10N or a 500N load cell was used, depending on the Processing of Glass Fibers thickness of the fiber. The diameter of the fiber was measured from both free ends of the sample with a micrometer with an The glass fiber manufacturing machinery was built to perform accuracy of 0.001 mm. The average of the two measured melt spinning tests. The apparatus contained a furnace with values was used in calculations of the strength. The diameter an opening at the bottom, as well as a fiber spinning unit, of the fibers varied from approximately 25 to 180 um. Prior which was placed -1. 5 m below the furnace to enable the to the testing, care was taken not to touch the fibers to avoid fibers to fully cool down before spinning into a roll. The any contamination of fiber surfaces. Since bioactive glass speed of the spinning roll was controlled by adjusting the fibers are brittle, they do not have a well-defined tensile rolling speed of the motor and by changing the roll diameter. strength. The stress at which they fail depends mostly on the To perform the melt spinning of the fibers, the formed glass presence of flaws, which may occur randomly along the block was heated in a platinum crucible with seven orifices at length of the fiber. The Weibull distribution is a statistical e bottom. After a while, the glass melt started to drain model used to describe the scatter in strength data. slowly, forming continuous fibers from each of the nozzles. strength values are ranked in descending order; the highest The drained glass drops/fibers were attached to a spinning roll strength value having the rank j=l, the next j=2, and so on creating continuous fibers. The achieved fiber diameter could until the smallest value j=N. a likelihood to survive, be adjusted from approximately 20 up to 300 um. The Sj=j/(N+1), is then assigned to each value of strength. A optimum fibrillation parameters were previously studied in minimum of 30 values of fracture strength are required for the project by variation of the melt temperature and the statistical validity. A Weibull's diagram is obtained by plot diameter of the orifices. In this study, a furnace set temper- ting InIn(1/S)) against In(strength). The gradient of the line of ature of 960C and an orifice nozzle diameter of -3.7 mm regression on In In(1/s)) upon In(strength) is the Weibul were found to be optimal for the melt spinning of glass modulus m, which enables the probability of a material to 3-93 survive under a given stress to be estimated. The Weibull modulus of glass 13-93 fibers were evaluated as outlined SBF Study earlier To study fiexural strength retention, changes of mass, and Strength Retention. The flexural strength was measured surface reactions in vitro, fiber samples were immersed in a using a three-point bending test fixture on fibers with diam- SBF by Kokubo. Fiber samples were kept in closed plastic containers stored in thermo closet at +37C. The sample eters from 229 to 338 um. A span length of 3. 8 mm and the surface area to SBF volume(SA/V) ratio of 0.1 cm was crosshead speed of 0.5 mm/min were used in this testing Prior to the testing, care was taken to avoid any contamina- used for all the test samples. The pH of the SBF solution was tion of the fiber surfaces. After immersion in SBF, the sam- sets of analysis were performed for samples with in vitro ples were flushed with distilled water and kept moist until the conditioning, namely (1)flexural strength retention of fiber testing was performed. The Weibull modulus of bioactive rods,(2)change in mass, and (3)scanning electron micro- glass 13-93 fibers was evaluated, as outlined earlier for flexural strength series where the number of samples was ic(SEM)analysis and compositional analysis of the above 30. Also the average and median strength and modulus glass surface after immersion in SBF. values were calculated for fiber series immersed in Sbf for Tests and Analysis Performed various time periods Elemental Analyses. Elemental analyses were performed Changes in Mass. Three series of fiber samples were find out whether there was a change in the glass compo- prepared, namely fibers with a nominal fiber diameter of 38, sition during the fiber manufacturing process. The elemental 100, and 210 um. The surface area/volume(SA/) ratio of analyses were performed by Rautaruukki Steel(Raahe, Fin- 0. 1 cm was used for all tested samples. The surface area of land) with X-ray spectroscopy(Philips PW 2404 RGT). Two samples was kept constant with all the samples, and so the
MATERIALS AND METHODS Manufacturing of the Glass The bioactive glass 13–93 contains 6 wt % Na2O, 12 wt % K2O, 5 wt % MgO, 20 wt % CaO, 4 wt % P2O5, and 53 wt % SiO2. The raw materials used, analytical grade Na2CO3, K2CO3, MgO, CaCO3, CaHPO42H2O as well as SiO2, were weighed and mixed in a plastic container. The mixed raw materials were melted in a platinum crucible for 3 h at 1360°C. To achieve a homogeneous glass, the formed glass was crushed into 1 cm3 pieces and reheated for 3 h at 1360°C. Processing of Glass Fibers The glass fiber manufacturing machinery was built to perform melt spinning tests. The apparatus contained a furnace with an opening at the bottom, as well as a fiber spinning unit, which was placed 1.5 m below the furnace to enable the fibers to fully cool down before spinning into a roll. The speed of the spinning roll was controlled by adjusting the rolling speed of the motor and by changing the roll diameter. To perform the melt spinning of the fibers, the formed glass block was heated in a platinum crucible with seven orifices at the bottom. After a while, the glass melt started to drain slowly, forming continuous fibers from each of the nozzles. The drained glass drops/fibers were attached to a spinning roll creating continuous fibers. The achieved fiber diameter could be adjusted from approximately 20 up to 300 m. The optimum fibrillation parameters were previously studied in the project by variation of the melt temperature and the diameter of the orifices. In this study, a furnace set temperature of 960°C and an orifice nozzle diameter of 3.7 mm were found to be optimal for the melt spinning of glass 13–93. SBF Study To study flexural strength retention, changes of mass, and surface reactions in vitro, fiber samples were immersed in a SBF by Kokubo.13 Fiber samples were kept in closed plastic containers stored in thermo closet at 37°C. The sample surface area to SBF volume (SA/V) ratio of 0.1 cm1 was used for all the test samples. The pH of the SBF solution was monitored and solution changed once in every 2 weeks. Three sets of analysis were performed for samples with in vitro conditioning, namely (1) flexural strength retention of fiber rods, (2) change in mass, and (3) scanning electron microscopic (SEM) analysis and compositional analysis of the glass surface after immersion in SBF. Tests and Analysis Performed Elemental Analyses. Elemental analyses were performed to find out whether there was a change in the glass composition during the fiber manufacturing process. The elemental analyses were performed by Rautaruukki Steel (Raahe, Finland) with X-ray spectroscopy (Philips PW 2404 RGT). Two elemental analyses were performed, one on manufactured bioactive glass fibers and the other on the remnants of the glass in the platinum crucible. Tensile Test. Tensile tests were performed on the newly drawn bioactive glass fibers, normally within days after the manufacturing. The tensile test results reported in this study are results from various manufacturing batches. The ASTM D3379 –75 instructions were followed in the tensile testing, in the main. A gauge length of 50 mm and a testing speed of 1 mm/min were used. The Instron materials testing machine (Series IX Automated Materials Testing System 1.38) with either a 10N or a 500N load cell was used, depending on the thickness of the fiber. The diameter of the fiber was measured from both free ends of the sample with a micrometer with an accuracy of 0.001 mm. The average of the two measured values was used in calculations of the strength. The diameter of the fibers varied from approximately 25 to 180 m. Prior to the testing, care was taken not to touch the fibers to avoid any contamination of fiber surfaces. Since bioactive glass fibers are brittle, they do not have a well-defined tensile strength. The stress at which they fail depends mostly on the presence of flaws, which may occur randomly along the length of the fiber. The Weibull distribution is a statistical model used to describe the scatter in strength data.14 N strength values are ranked in descending order; the highest strength value having the rank j1, the next j2, and so on until the smallest value j N. A likelihood to survive, Sjj/(N1), is then assigned to each value of strength. A minimum of 30 values of fracture strength are required for statistical validity. A Weibull’s diagram is obtained by plotting ln ln(1/Sj) against ln(strength). The gradient of the line of regression on ln ln(1/Sj) upon ln(strength) is the Weibull modulus m, which enables the probability of a material to survive under a given stress to be estimated. The Weibull modulus of glass 13–93 fibers were evaluated as outlined earlier. Strength Retention. The flexural strength was measured using a three-point bending test fixture on fibers with diameters from 229 to 338 m. A span length of 3.8 mm and the crosshead speed of 0.5 mm/min were used in this testing. Prior to the testing, care was taken to avoid any contamination of the fiber surfaces. After immersion in SBF, the samples were flushed with distilled water and kept moist until the testing was performed. The Weibull modulus of bioactive glass 13–93 fibers was evaluated, as outlined earlier for flexural strength series where the number of samples was above 30. Also the average and median strength and modulus values were calculated for fiber series immersed in SBF for various time periods. Changes in Mass. Three series of fiber samples were prepared, namely fibers with a nominal fiber diameter of 38, 100, and 210 m. The surface area/volume (SA/V) ratio of 0.1 cm1 was used for all tested samples. The surface area of samples was kept constant with all the samples, and so the 228 PIRHONEN ET AL.
BIOACTIVE GLASS 13-93 FIBERS TABLE I. The Results of the Elemental Analyses solid glass block was successfully drawn to fibers. Glass Glass Glass Theoretical particles were used in early tests, but it was soon noticed that Sample Fibers Remnants their surfaces started to crystallize at the melt spinning tem- Na.o 5.93 6 perature. Some crystallites were transferred to the fibers as 11 well. When a solid block of glass was used, the homogeneous Moo 4.83 fibers were successfully manufactured and only the upper surface of the glass crystallized. Using this manufacturin P,O 4 method, -50% of the weight of the initial glass block wa 53.0 53.0 successfully drawn to fibers in each case All values are given in weight percentage. Elemental Analyses The results of the elemental analyses are shown in Table L. average sample sizes were as follows: 27 mg for 38 um The compositions are given in weight percentage. As seen in fibers, 67 mg for 100 um fibers, and 137 mg for 210 um Table I, there are no significant compositional changes in the fibers. The initial weight of the sample as well as the weight glass during the fiber manufacturing process. The manufac after immersion was measured with accuracy of 0.01 mg. tured glass fibers as well as the remnants of the glass(partly There were two parallel samples tested for each series at each crystallized) possessed an identical compositional structure time point, and the average of these two measurements is It can also be noticed that with this manufacturing method, reported in this study homogeneous glass batches can be achieved with only minor variation in compositions compared with the theoretically SEM and Compositional Analysis. For the SEM analy- calculated values techniques. The fibers were placed in\s a specimen hols a sis, test specimens were first prepared using hot mount Tensile Testing perpendicular to the surface prior to mounting so as to hieve a cross-sectional view of the fiber. After casting, the All the fibers exhibited a brittle failure and a linear force- specimens were ground and finally polished using No. 4000 deflection response to the point of failure Sic-paper. A SEM Model XL30(Philips, The Netherlands) As seen in tensile test results in Table I, there is a large equipped with an energy dispersive spectrometer(EDS) variation in tensile strength results of bioactive glass 13-93 fibers. The thinner fibers possess higher mechanical strength positional analysis. The analysis was performed by scanning values compared with the thicker fibers, but also the Weibull over the formed surface layers to achieve a compositional modulus (1.8)is the lowest for the thin fibers. The Weibull nalysis for each layer For each sample, two or three parallel modulus is higher(3. 1)for fibers with the largest diameter runs were performed. The analysis data were transferred to When all the tested fibers are considered as one lot, the Excel where, using the initial images, the distances could be Weibull modulus value is 2.1 calibrated. The compositional analysis data presented in this study are taken from samples that have been immersed in Flexural Strength Retention In Vitro SBF for 24 h and 5 weeks The results of the three-point bending tests after variou immersion times are shown in Table ll. The flexural strength RESULTS of fibers first started to decrease when immersed in Sbf. but the strength was recovered after immersion in Sbf between Processing of the Fibers 2 days and I week. After I week immersion, the flexural strength started to decrease again, and after 7 weeks, the During melt spinning, the upper surface(the one in contact flexural strength was less than 15%o of the initial strength. The with the air)of the glass block crystallized slightly. Only a decrease in strength was reasonably low from 7 to 40 weeks TABLE IL. The Effect of the Fiber Diameter on the values of Median Tensile Strength, Average Tensile Strength, and Weibull Modulus of the bioactive glass 13-93 Fibers Fiber Number Median Lowest-Highest Standard Strength (MPa) Modulus 25-33 910.10 65.3-1693.9 861.80 35742 176 34-47 809.85 147.8-1918.6 795.69 31945 2.20 93-160 424.75 150.2-725.7 15122
average sample sizes were as follows; 27 mg for 38 m fibers, 67 mg for 100 m fibers, and 137 mg for 210 m fibers. The initial weight of the sample as well as the weight after immersion was measured with accuracy of 0.01 mg. There were two parallel samples tested for each series at each time point, and the average of these two measurements is reported in this study. SEM and Compositional Analysis. For the SEM analysis, test specimens were first prepared using hot mounting techniques. The fibers were placed into a specimen holder perpendicular to the surface prior to mounting so as to achieve a cross-sectional view of the fiber. After casting, the specimens were ground and finally polished using No. 4000 SiC-paper. A SEM Model XL30 (Philips, The Netherlands) equipped with an energy dispersive spectrometer (EDS) (Model DX-4, EDAX International, USA) was used for compositional analysis. The analysis was performed by scanning over the formed surface layers to achieve a compositional analysis for each layer. For each sample, two or three parallel runs were performed. The analysis data were transferred to Excel where, using the initial images, the distances could be calibrated. The compositional analysis data presented in this study are taken from samples that have been immersed in SBF for 24 h and 5 weeks. RESULTS Processing of the Fibers During melt spinning, the upper surface (the one in contact with the air) of the glass block crystallized slightly. Only a solid glass block was successfully drawn to fibers. Glass particles were used in early tests, but it was soon noticed that their surfaces started to crystallize at the melt spinning temperature. Some crystallites were transferred to the fibers as well. When a solid block of glass was used, the homogeneous fibers were successfully manufactured and only the upper surface of the glass crystallized. Using this manufacturing method, 50% of the weight of the initial glass block was successfully drawn to fibers in each case. Elemental Analyses The results of the elemental analyses are shown in Table I. The compositions are given in weight percentage. As seen in Table I, there are no significant compositional changes in the glass during the fiber manufacturing process. The manufactured glass fibers as well as the remnants of the glass (partly crystallized) possessed an identical compositional structure. It can also be noticed that with this manufacturing method, homogeneous glass batches can be achieved with only minor variation in compositions compared with the theoretically calculated values. Tensile Testing All the fibers exhibited a brittle failure and a linear forcedeflection response to the point of failure. As seen in tensile test results in Table II, there is a large variation in tensile strength results of bioactive glass 13–93 fibers. The thinner fibers possess higher mechanical strength values compared with the thicker fibers, but also the Weibull modulus (1.8) is the lowest for the thin fibers. The Weibull modulus is higher (3.1) for fibers with the largest diameter. When all the tested fibers are considered as one lot, the Weibull modulus value is 2.1. Flexural Strength Retention In Vitro The results of the three-point bending tests after various immersion times are shown in Table III. The flexural strength of fibers first started to decrease when immersed in SBF, but the strength was recovered after immersion in SBF between 2 days and 1 week. After 1 week immersion, the flexural strength started to decrease again, and after 7 weeks, the flexural strength was less than 15% of the initial strength. The decrease in strength was reasonably low from 7 to 40 weeks. TABLE I. The Results of the Elemental Analyses Sample Glass Fibers Glass Remnants Theoretical Value Na2O 5.93 5.90 6 K2O 11.7 11.8 12 MgO 4.82 4.83 5 CaO 20.4 20.2 20 P2O5 4.05 4.05 4 SiO2 53.0 53.0 53 All values are given in weight percentage. TABLE II. The Effect of the Fiber Diameter on the Values of Median Tensile Strength, Average Tensile Strength, and Weibull Modulus of the Bioactive Glass 13–93 Fibers Fiber Diameter (m) Number of Samples Median Strength (MPa) Lowest-Highest Strength (MPa) Average Strength (MPa) Standard Deviation (MPa) Weibull Modulus 25–33 31 910.10 165.3–1693.9 861.80 357.42 1.76 34–47 30 809.85 147.8–1918.6 795.69 319.45 2.20 48–91 31 601.50 210.6–1887.5 651.48 389.09 1.99 93–160 32 424.75 150.2–725.7 440.57 151.22 3.06 BIOACTIVE GLASS 13–93 FIBERS 229
PIRHONEN ET AL. TABLE IIL Flexural Test Results for bioactive glass 13-93 Fibers Average Modulus Standard Immersion Number Median Lowest-Highest Average Deviation Deviation Strength Values Weibull strength (Strength) Elasticity (Modulus) (GPa) (GPa) 1353.5 3334-3073.0 1443.0 696.7 68.12 6.77 1 day 937.1 500.4-2433.0 1097.8 5009 63.46 7.61 2 days 40 969.9 525.7-19700 3.59 1024.7 326.8 66.85 5.13 3 day 86.2-2727.0 1368 428.7 62.55 1583.5 663.9-2633.0 3.29 522.6 477.4 63.36 5.24 1317.5 523.2-1948.0 421.3 61.31 960.7 90.2-26880 51 732.1 34579 575.81259-1548.01.66 618.8 57.39 798 238.4 116.39593 2.47 284.8 161.5 324 41 .71 2054 71.65-637.6 210.1 39.61 11 weeks 40.1-3523 2.18 76.6 13 weeks 4908009 206.3 1089-337.1 214.6 67.3 43.58 15 weeks 148.8 24.23-282.3 153.8 14 20 weeks 168.3 59.32-242.6 163.6 30.98-354.6 157.2 43.8 30 weeks 42.3-244.1 1598 47.1 49.0 9.72 120.6 63.87-363.6 140.2 41.61 40 weeks 50.28-1939 123.8 40.1 44.13 The flexural modulus of bioactive glass 13-93 fibers started same behavior was also observed with fibers with a nominal to decrease when the glass fibers were immersed in SBF. The diameter of 100 and 210 um variation in modulus values was initially minor, but increased For all the three sets of fibers with variable diameters fter 5 weeks of immersion. The Weibull modulus was ini- namely 38, 100, and 210 um, there was almost identical tially -2 and it reached its maximum 3.6 after 2 days of change in actual mass during the time period, even though the immersion. After 3 weeks of immersion, the Weibull modu- initial mass of samples did differ significantly (27 mg for 38 lus decreased and a large variation in strength as well as um fibers, 67 mg for 100 um fibers, and 137 mg for 210 um lower strength values can be detected. fibers). The mass increase with the thinnest(38 um) fibers is slightly lower after 52 weeks of immersion compared with Changes in Mass the thicker fibers The changes in mass of bioactive glass 13-93 fibers are Compositional Analysis presented in Figures I and 2 In Figure l, the changes in mass of fibers with a diameter of -38 um are shown for the first The compositional analysis performed on samples immersed 72 h in SBF. within the first 12 h, there was a rapid decrease for 24 h is presented in Figure 3. From the compositional in mass but after 24 h the mass had again increased. The Change in mass of bioactive glass 13-93 fibers Change in mass of bioactive glass 13-93 fibers after immersion into SBF with diameter of 38 microns 38 microns 100 microns 三 -+38 microns -210 microns mersion time I weeks Immersion timeS Figure 2. The change in mass of bioactive glass 13-93 fibers after Figure 1. The change in mass of bioactive glass 13-93 fibers with immersion in SBF. Initial masses of samples were as follows: 27 mg diameter of 38 um during the first three days of immersion in SBF. for 38 um fibers, 67 mg for 100 um fibers, and 137 mg for 210 um Initial mass of sample was 27 mg
The flexural modulus of bioactive glass 13–93 fibers started to decrease when the glass fibers were immersed in SBF. The variation in modulus values was initially minor, but increased after 5 weeks of immersion. The Weibull modulus was initially 2 and it reached its maximum 3.6 after 2 days of immersion. After 3 weeks of immersion, the Weibull modulus decreased and a large variation in strength as well as lower strength values can be detected. Changes in Mass The changes in mass of bioactive glass 13–93 fibers are presented in Figures 1 and 2. In Figure 1, the changes in mass of fibers with a diameter of 38 m are shown for the first 72 h in SBF. Within the first 12 h, there was a rapid decrease in mass, but after 24 h the mass had again increased. The same behavior was also observed with fibers with a nominal diameter of 100 and 210 m. For all the three sets of fibers with variable diameters, namely 38, 100, and 210 m, there was almost identical change in actual mass during the time period, even though the initial mass of samples did differ significantly (27 mg for 38 m fibers, 67 mg for 100 m fibers, and 137 mg for 210 m fibers). The mass increase with the thinnest (38 m) fibers is slightly lower after 52 weeks of immersion compared with the thicker fibers. Compositional Analysis The compositional analysis performed on samples immersed for 24 h is presented in Figure 3. From the compositional TABLE III. Flexural Test Results for Bioactive Glass 13–93 Fibers Immersion Time in SBF Number of Samples Median Strength (MPa) Lowest-Highest Strength Values (MPa) Weibull Modulus Average strength (MPa) Standard Deviation (Strength) (MPa) Average Modulus of Elasticity (GPa) Standard Deviation (Modulus) (GPa) 0 40 1353.5 333.4–3073.0 2.09 1443.0 696.7 68.12 6.77 1 day 40 937.1 500.4–2433.0 2.60 1097.8 500.9 63.46 7.61 2 days 40 969.9 525.7–1970.0 3.59 1024.7 326.8 66.85 5.13 3 days 42 1335.5 486.2–2727.0 3.20 1368.1 428.7 62.55 5.54 1 week 40 1583.5 663.9–2633.0 3.29 1522.6 477.4 63.36 5.24 2 weeks 40 1317.5 523.2–1948.0 3.08 1265.4 421.3 61.31 5.55 3 weeks 40 960.7 190.2–2688.0 1.51 1177.1 732.1 54.43 6.09 4 weeks 40 575.8 125.9–1548.0 1.66 618.8 369.7 57.39 7.98 5 weeks 41 238.4 116.3–959.3 2.47 284.8 161.5 47.73 10.38 7 weeks 40 276.9 99.7–1164.0 1.86 324.1 226.7 41.93 10.71 9 weeks 35 205.4 71.65–637.6 2.48 210.1 99.5 39.61 7.18 11 weeks 34 175.9 40.1–352.3 2.18 177.6 76.6 42.87 11.04 13 weeks 19 206.3 108.9–337.1 214.6 67.3 43.58 6.71 15 weeks 20 148.8 24.23–282.3 153.8 64.8 49.69 14.63 20 weeks 18 168.3 59.32–242.6 163.6 50.2 53.85 12.43 25 weeks 20 151.7 30.98–354.6 157.2 69.0 43.83 18.04 30 weeks 20 158.1 42.3–244.1 159.8 47.1 49.02 9.72 35 weeks 19 120.6 63.87–363.6 140.2 70.0 41.61 12.93 40 weeks 20 121.9 50.28–193.9 123.8 40.1 44.13 12.88 Figure 1. The change in mass of bioactive glass 13–93 fibers with diameter of 38 m during the first three days of immersion in SBF. Initial mass of sample was 27 mg. Figure 2. The change in mass of bioactive glass 13–93 fibers after immersion in SBF. Initial masses of samples were as follows: 27 mg for 38 m fibers, 67 mg for 100 m fibers, and 137 mg for 210 m fibers. 230 PIRHONEN ET AL
BIOACTIVE GLASS 13-93 FIBERS Elemental analysis of bioactive glass 13-93 fiber The flexural strength of glass fibers decreased slightly after 24 hours of immersion in SBF after immersion in sbf. but it further increased from the nitial strength value after 1 week of immersion. As brittle materials,glass fibers are very sensitive to the presence of 0600 microscopic imperfections, or flaws, which behave as stress concentrators. As the si-gel forms on the bioactive glass surface, it overcomes the microscopic imperfections or flaw in the surface and the strength of the fibers is increased. this can also be detected from the Weibull modulus value that is 2 for nonimmersed fibers, but increases up to 3.6 for im- Distance from fiber surface/um mersed fibers, onto which a Si-rich layer has been formed. Figure 3. The EDXA analysis result of bioactive glass 13-93 fiber After I week immersion time, the flexural strength starts to surface after 24 h of immersion in SBF. Distance from the fiber surface decrease and this is probably due to increasing thickness of measured toward the centre of the fiber the CaP-layer. This layer forms a brittle core and decrease the overall strength of the fibers. from 5 to 40 weeks. the further decrease in strength is rather low, and this may be analysis data, it can be noticed that after 24 h the Na+, K+. explained by the fact that the CaP-layer reaches sufficient and Ca2+have started to diffuse out from the glass surface, thickness and slows down the further degradation and leach- and there is an increase in the amount of silicon in the glass ing of ions from the glass. The flexural modulus starts to surface. The transformed layer has a thickness varying from decrease steadily after 1 week of immersion. This is most 2 to 3 um after 24 h of immersion probably due to the transfer of the stiff glass phase to the low In the compositional analysis for fibers immersed for 5 modulus Si-gel layer. The high standard deviation in modulus weeks into SBE, there is a clear Ca-and P-rich layer on the values after 2 weeks immersion is probably due to the inho fiber surface and a Si-rich layer beneath the CaP-rich layer en in Figures 4 and 5. The Cap-rich layer has uniform The change in mass in vitro correlates closely to the thickness of-2 um all over the fiber surface. The thickness surface area of the glass samples, as the samples with uniform of the Si-rich layer is variable over the fiber surface, and as surface areas, but variable initial masses, had a uniform shown in Figure 6, the thickness of the Si-rich layer is -5-30 change in overall mass, and so the change in mass was um after 5 weeks of immersion in SBF. From Figure 6, it can proportional to the surface area of the samples, rather than the be noticed that the remnants of an additional CaP-rich layer total volume of the samples. The initial leaching of ions can can be detected on top of the continuous CaP-rich layer. be detected as a loss of mass. Later in the experiment, the During sample preparation, it was noticed that as the samples mass of the samples increased, which is caused by the con- were dried, the CaP-rich surface became very brittle and tinuous increase in thickness of the CaP-layer. The diffusion some of the outer layer detached from the fiber surface of the elements from the surface rules the degradation of structures manufactured from bioactive glass. The total deg radation of the bioactive glass structures can be adjusted by controlling the surface/volume ratio. Thin fibers are expected DISCUSSION to resorb and transform to calcium phosphate faster than thicker fibers The formation of a hydroxycarbonate apatite layer on top of From the compositional analysis, on the reacted surfaces, the bioactive glass is a result of the following steps: (1) the dissolution of ions was clearly observed. A Si-rich layer leaching and formation of silanols (SiOH); (2) loss of soluble silica and formation of silanols; (3) polycondensation of Elemental analysis of bioactive glass 13-93 fiber silanols to form a hydrated silica gel; (4)formation of an after 5 weeks immersion in SBF amorphous calcium phosphate layer; and(5)crystallization of a hydroxycarbonate apatite layer. The formation of the Si nd CaP-rich layers in glass surfaces are usually analyzed be 25 by techniques such as X-ray diffraction(XRD), Fourier trans form infrared(FTIR)spectroscopy, or SEM. 6 The effect of formation of Si-and CaP-rich layers to the mechanical prop- $150 erties of the glasses has not been studied earlier. When 8 100 bioactive glass is drawn to fibers, very uniform test speci mens are achieved from which, for example, mechanical properties can be measured. In this study, the formation of 4 6 8 reactive layers on top of the surface of bioactive glass 13-93 Distance from fiber surface / um fibers was investigated, and its effect to mechanical proper- Figure 4. The result of EDXA analysis of bioactive glass 13-93 fiber ties and mass have been analyzed surface after 5 weeks of immersion in SBf
analysis data, it can be noticed that after 24 h the Na, K, and Ca2 have started to diffuse out from the glass surface, and there is an increase in the amount of silicon in the glass surface. The transformed layer has a thickness varying from 2 to 3 m after 24 h of immersion. In the compositional analysis for fibers immersed for 5 weeks into SBF, there is a clear Ca- and P-rich layer on the fiber surface and a Si-rich layer beneath the CaP-rich layer as seen in Figures 4 and 5. The CaP-rich layer has uniform thickness of 2 m all over the fiber surface. The thickness of the Si-rich layer is variable over the fiber surface, and as shown in Figure 6, the thickness of the Si-rich layer is 5–30 m after 5 weeks of immersion in SBF. From Figure 6, it can be noticed that the remnants of an additional CaP-rich layer can be detected on top of the continuous CaP-rich layer. During sample preparation, it was noticed that as the samples were dried, the CaP-rich surface became very brittle and some of the outer layer detached from the fiber surface. DISCUSSION The formation of a hydroxycarbonate apatite layer on top of the bioactive glass is a result of the following steps: (1) leaching and formation of silanols (SiOH); (2) loss of soluble silica and formation of silanols; (3) polycondensation of silanols to form a hydrated silica gel; (4) formation of an amorphous calcium phosphate layer; and (5) crystallization of a hydroxycarbonate apatite layer.15 The formation of the Siand CaP-rich layers in glass surfaces are usually analyzed be by techniques such as X-ray diffraction (XRD), Fourier transform infrared (FTIR) spectroscopy, or SEM.16 The effect of formation of Si- and CaP-rich layers to the mechanical properties of the glasses has not been studied earlier. When bioactive glass is drawn to fibers, very uniform test specimens are achieved from which, for example, mechanical properties can be measured. In this study, the formation of reactive layers on top of the surface of bioactive glass 13–93 fibers was investigated, and its effect to mechanical properties and mass have been analyzed. The flexural strength of glass fibers decreased slightly after immersion in SBF, but it further increased from the initial strength value after 1 week of immersion. As brittle materials, glass fibers are very sensitive to the presence of microscopic imperfections, or flaws, which behave as stress concentrators. As the Si-gel forms on the bioactive glass surface, it overcomes the microscopic imperfections or flaws in the surface and the strength of the fibers is increased. This can also be detected from the Weibull modulus value that is 2 for nonimmersed fibers, but increases up to 3.6 for immersed fibers, onto which a Si-rich layer has been formed. After 1 week immersion time, the flexural strength starts to decrease and this is probably due to increasing thickness of the CaP-layer. This layer forms a brittle core and decreases the overall strength of the fibers. From 5 to 40 weeks, the further decrease in strength is rather low, and this may be explained by the fact that the CaP-layer reaches sufficient thickness and slows down the further degradation and leaching of ions from the glass. The flexural modulus starts to decrease steadily after 1 week of immersion. This is most probably due to the transfer of the stiff glass phase to the low modulus Si-gel layer. The high standard deviation in modulus values after 2 weeks immersion is probably due to the inhomogenity of the Si-gel and CaP layers. The change in mass in vitro correlates closely to the surface area of the glass samples, as the samples with uniform surface areas, but variable initial masses, had a uniform change in overall mass, and so the change in mass was proportional to the surface area of the samples, rather than the total volume of the samples. The initial leaching of ions can be detected as a loss of mass. Later in the experiment, the mass of the samples increased, which is caused by the continuous increase in thickness of the CaP-layer. The diffusion of the elements from the surface rules the degradation of structures manufactured from bioactive glass. The total degradation of the bioactive glass structures can be adjusted by controlling the surface/volume ratio. Thin fibers are expected to resorb and transform to calcium phosphate faster than thicker fibers. From the compositional analysis, on the reacted surfaces, the dissolution of ions was clearly observed. A Si-rich layer Figure 4. The result of EDXA analysis of bioactive glass 13–93 fiber surface after 5 weeks of immersion in SBF. Figure 3. The EDXA analysis result of bioactive glass 13–93 fiber surface after 24 h of immersion in SBF. Distance from the fiber surface measured toward the centre of the fiber. BIOACTIVE GLASS 13–93 FIBERS 231
PIRHONEN ET AL. 200kV5.04000XBSE99 10vkd2 Figure 5. The SEM image of the observed area. The fiber has been immersed in SBF for 5 weeks. The layers in figure from top to bottom are glass, Si-rich layer, CaP-rich layer, and had formed already after 24 h of immersion and a slight Cap had disintegrated from the surface during sample prep- increase of Ca and P ions could already be detected at this aration time point. After 5 weeks of immersion, the thickness of the There is interest to develop new materials for biomedical ilica layer varied from 5 to 30 um and the inner Cap layer purposes and bioactive glass fibers bring interesting possibi showed a constant thickness of 2 um. Although the Cap ities for example as a reinforcement in polymeric composites content of this sample was initially higher, the outer layers of Thus, the idea of using bioactive glass fibers to reinforce biopolymers is not new, and several research groups have studied this issue. For example, the processing and mechan- ical properties of bioactive or resorbable glass fiber rein- forced composite materials have been reported by Dunn et al (1985), Lin(1986), and Krebs et al. (1993).Marcolongo et al. have also published in vitro and in vivo studies on bioactive glass fiber reinforced polysulfones. They found that bone tissue exhibited direct contact with the glass fibers and strengths significantly higher than with polymer controls ic djacent polymer matrix, resulting in interfacial bo As reinforcement in composites, the fiber strength is the most infuential factor on the strength of the composite. The strength of the glass fiber is dependent on the defect structure, for example, cracks, voids, and impurities of the fiber, and possibly also on the stress state of the fiber. Thinner fibers have a smaller surface area, that is, a smaller probability for the existence of flaws, and this will lead to higher strength values. This may also explain the variation in tensile and 100 um flexural stress values in this study. In the three-point testing. Ovc the area under maximum stress is significantly smaller com- pared with the situation in tensile testing. Figure 6. A SEM image using back scattered electron (BCE)of a fiber The cooling of the glass fiber in melt spinning happens immersed in SBF for 5 weeks. The glass is seen as a light gray, silica rather quickly. This causes stress distribution that affects the as a dark gray, and CaP as a white zone in the image. The black arrow glass fiber, resulting in compression stress in the surface parts shows remnant of a second Cap layer and tensile stress in the inner parts. This toughening phen
had formed already after 24 h of immersion and a slight increase of Ca and P ions could already be detected at this time point. After 5 weeks of immersion, the thickness of the silica layer varied from 5 to 30 m and the inner CaP layer showed a constant thickness of 2 m. Although the CaP content of this sample was initially higher, the outer layers of CaP had disintegrated from the surface during sample preparation. There is interest to develop new materials for biomedical purposes and bioactive glass fibers bring interesting possibilities for example as a reinforcement in polymeric composites. Thus, the idea of using bioactive glass fibers to reinforce biopolymers is not new, and several research groups have studied this issue. For example, the processing and mechanical properties of bioactive or resorbable glass fiber reinforced composite materials have been reported by Dunn et al. (1985),17 Lin (1986),18 and Krebs et al. (1993).19 Marcolongo et al. have also published in vitro and in vivo studies on bioactive glass fiber reinforced polysulfones. They found that bone tissue exhibited direct contact with the glass fibers and adjacent polymer matrix, resulting in interfacial bond strengths significantly higher than with polymer controls.20 As reinforcement in composites, the fiber strength is the most influential factor on the strength of the composite. The strength of the glass fiber is dependent on the defect structure, for example, cracks, voids, and impurities of the fiber, and possibly also on the stress state of the fiber.21 Thinner fibers have a smaller surface area, that is, a smaller probability for the existence of flaws, and this will lead to higher strength values. This may also explain the variation in tensile and flexural stress values in this study. In the three-point testing, the area under maximum stress is significantly smaller compared with the situation in tensile testing. The cooling of the glass fiber in melt spinning happens rather quickly. This causes stress distribution that affects the glass fiber, resulting in compression stress in the surface parts and tensile stress in the inner parts. This toughening phenomFigure 5. The SEM image of the observed area. The fiber has been immersed in SBF for 5 weeks. The layers in figure from top to bottom are glass, Si-rich layer, CaP-rich layer, and epoxy. Figure 6. A SEM image using back scattered electron (BCE) of a fiber immersed in SBF for 5 weeks. The glass is seen as a light gray, silica as a dark gray, and CaP as a white zone in the image. The black arrow shows remnant of a second CaP layer. 232 PIRHONEN ET AL
BIOACTIVE GLASS 13-93 FIBERS enon is caused by the stress distribution, which is a result of 4. Xynos ID, Hukkanen MJJ, Batten JJ, Buttery LD, Hench LL, density differences in the different layers of the fiber caused Polak JM. Bioglass 45S5 stimul turnover and by cooling. For thinner fibers, the cooling is more rapid than enhances bone formation in vitro d applications this may also cause the difference in 5. Hench LL Biomaterials: a forecast for the rength properties. In technical glass fiber manufacturing, the 998:19:1419-1423 ooling of the fibers is normally done by spraying water 6. Vita Finzi Zalman E, Locardi B, Gabbi C, Tranquilli Leali P. raight to the fiber when draining out from the nozzle, and Bioactive vitreous composition for bone implants, filaments nis further decreases the cooling time of thin fibers Another made therefrom and method. PCT Wo 91/12032. 1991 7. Brink M. Bioactive Glasses with a Large Working range. benefit of manufacturing technical grade glass fibers is the Doctoral Thesis. Turku, Finland: Abo Aka use of sizing agents to cover the formed fiber surface already within the fiber spinning process. The reported tensile 8. Brink M, Turunen T, Happonen R-P, Yli-Urpo A Com ength for technical sized E-glass fiber is-2 GPa or above tional dependence of bioactivity of glasses in the system Na,O for glass fibers with a diameter of -15 um. The reported K2O-Mgo-CaO-B2O3-P2Os-Sio med Mater Res 1997 37:114-121 Weibull modulus for technical E-glass fibers is -5.522.23 De 9. Brink M, Yli-Urpo S, Yli-Urpo A. The resorption of a bioactive Diego et al. have studied the tensile properties of 45S5 lass implanted into rat soft tissue. In the 5th World Biomate- MPa, with the Weibull modulus being between 3.0 and 3.5. 10. Brink M, Laine p otg K, Yli-Urpo A. Implantation of bio- Bioglass. The obtained tensile strength was at the level of 90 The tested fibers were thick, from 165 to 310 um.4 The ctive and inert glass fibres in rats-Soft tissue response and Weibull modulus for the manufactured bioactive glass 13-93 short term reactions of the glass. In: Sedel L, Rey C, editors fibers tested in this study is rather low, m= 2, which Bioceramics. New York: Elsevier Science: 1997. Vol 10, p 61-64 represents considerable uncertainty about the stress level at 11. Pirhonen E, Grandi G, Tormala P. Bioactive glass fibre/poly- which a fiber is likely to fail. As glass fibers are highly sensitive to abrasion and flaws drastically affect the strength 12. Pirhonen E, Moimas L, Haapanen J Porous bioactive 3-D glass properties, it would be beneficial to use sizing or coupling agents for bioactive glass fibers by sintering technique. Key Eng Mater 2003: 240-242: 237-240. 13. Kokubo T. Kushitani H. Sakka S. Kitsugi. Yamamuro T Solu- tions able to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-w. J Biomed Mater Res 1990: 24 CONCLUSIONS 721-734 14. Hull D, Clyne Tw. An Introduction to Composite Materials, In this work, fiber manufacturing process for bioactive glass 15. Hench LL, Andersson 0. Bioactive glasses. In: Hench L,Wil- 13-93 was developed. Strong glass fibers with a diameter son j. editors. An Introduction to Bioceramics. S adjustable from 20 to 300 um were successfully manufac World Scientific Publishing: 1993 tured Bioactive glass 13-93 fibers with a diameter of -250 16. Izquierda-Barba I, Salinas A, Vallet-Regi M In vitro calcium um retained their flexural strength up to 3 weeks in SBF and phosphate layer formation on sol-gel glasses of the Cao-Sio system J Biomed Mater Res 1999: 47: 243-24 started to decrease in strength after that. However, there Is 17. Dunn RL, Cassper RA, Kelley BS. Biodegradable composites only minor strength loss within 7-40 weeks immersion in In the llth Annu Mtg Soc Biomater, San Diego, CA, Apri SBF. The flexural modulus started to decrease when the fibers 25-28,1985.p213 were immersed in SBE The change in mass when immersed 18. Lin TC. Totally absorbable fibre reinforced composite for in- in SBF was proportional to the surface area of the samples, termal fracture fixation devices. In the 12th Annu Mtg Soc rather than the total volume of the samples. As the degrada- Biomater, Minneapolis-St Paul, MN, May 29-June 1, 1986. p tion rate of the samples is determined by the diffusion of ions 19. Krebs S, Lin S, King R Characterization of HA and resorbable from the surface, the degradation of the material can be glass fibers reinforced PLLA screws In the 19th Annu Mtg Soc djusted by controlling the surface area/volume ratio and wall Biomater, April 28-May 2, Birmingham, AL, 1993. P 317 hickness of bioactive glass structures. Marcolongo M, Ducheyne P, Garino J, Schepers E. Bioactive glass fibre/composites bond to bone tissue. J Biomed Mater Res The authors thank mrs. tiina aaltonen for the technical assis- 1997:37:440-448 tance and Dr. Kari Kolppo for help with the compositional analysis. 21. Jarvela P. Properties of Glass Fibres and Their Applications in Porous Composites, Doctoral Thesis. Tampere, Finland: Tam- REFERENCES 22. Pardini LC, Manhani LGB. Infiuence of the testing gage length on the strength, Youngs modulus and Weibull modulus of 1. Hutmacher Dw. Scaffolds in tissue engineering bone and car- carbon fibres and glass fibres. Mater Res 2002: 5: 411-420 lage. Biomaterials 2000: 21: 2529-2543 23. Gurvich MR. Dibenedetto at. Pegoretti A. Evaluation of the 2. Bonassar LJ. Vacanti CA. Tissue engineering: The first decade statistical parameters of a Weibull distribution. J Mater Sci and beyond. J Cell Biochem Suppl 1998: 30/31: 297-303. 1997:32:3711-3716 3. Hench LL. Bioactive materials: The potential for tissue regen- 24. De Diego MA, Coleman NJ, Hench LL. Tensile eration. In the 24th Annu Mtg Soc biomater. San Diego, CA, bioactive fibres for tissue engineering applications. J Biomed 1998.p5l1-518 Mater Res B Appl Biomater 2000: 53: 199-203
enon is caused by the stress distribution, which is a result of density differences in the different layers of the fiber caused by cooling. For thinner fibers, the cooling is more rapid than for thicker fibers, and this may also cause the difference in strength properties. In technical glass fiber manufacturing, the cooling of the fibers is normally done by spraying water straight to the fiber when draining out from the nozzle, and this further decreases the cooling time of thin fibers. Another benefit of manufacturing technical grade glass fibers is the use of sizing agents to cover the formed fiber surface already within the fiber spinning process. The reported tensile strength for technical sized E-glass fiber is 2 GPa or above for glass fibers with a diameter of 15 m. The reported Weibull modulus for technical E-glass fibers is 5.5.22,23 De Diego et al. have studied the tensile properties of 45S5 Bioglass®. The obtained tensile strength was at the level of 90 MPa, with the Weibull modulus being between 3.0 and 3.5. The tested fibers were thick, from 165 to 310 m.24 The Weibull modulus for the manufactured bioactive glass 13–93 fibers tested in this study is rather low, m 2, which represents considerable uncertainty about the stress level at which a fiber is likely to fail. As glass fibers are highly sensitive to abrasion and flaws drastically affect the strength properties, it would be beneficial to use sizing or coupling agents for bioactive glass fibers. CONCLUSIONS In this work, fiber manufacturing process for bioactive glass 13–93 was developed. Strong glass fibers with a diameter adjustable from 20 to 300 m were successfully manufactured. Bioactive glass 13–93 fibers with a diameter of 250 m retained their flexural strength up to 3 weeks in SBF and started to decrease in strength after that. However, there is only minor strength loss within 7– 40 weeks immersion in SBF. The flexural modulus started to decrease when the fibers were immersed in SBF. The change in mass when immersed in SBF was proportional to the surface area of the samples, rather than the total volume of the samples. As the degradation rate of the samples is determined by the diffusion of ions from the surface, the degradation of the material can be adjusted by controlling the surface area/volume ratio and wall thickness of bioactive glass structures. The authors thank Mrs. Tiina Aaltonen for the technical assistance and Dr. Kari Kolppo for help with the compositional analysis. REFERENCES 1. Hutmacher DW. Scaffolds in tissue engineering bone and cartilage. Biomaterials 2000;21:2529 –2543. 2. Bonassar LJ, Vacanti CA. Tissue engineering: The first decade and beyond. J Cell Biochem Suppl 1998;30/31:297–303. 3. Hench LL. Bioactive materials: The potential for tissue regeneration. In the 24th Annu Mtg Soc Biomater, San Diego, CA, 1998. p 511–518. 4. Xynos ID, Hukkanen MJJ, Batten JJ, Buttery LD, Hench LL, Polak JM. Bioglass 45S5 stimulates osteoblast turnover and enhances bone formation in vitro. Implications and applications for bone tissue engineering. Calcif Tissue Int 2000;67:321–329. 5. Hench LL. Biomaterials: A forecast for the future. Biomaterials 1998;19:1419 –1423. 6. Vita Finzi Zalman E, Locardi B, Gabbi C, Tranquilli Leali P. Bioactive vitreous composition for bone implants, filaments made therefrom and method. PCT WO 91/12032, 1991. 7. Brink M. Bioactive Glasses with a Large Working Range, Doctoral Thesis. Turku, Finland: Åbo Akademi University; 1997. 8. Brink M, Turunen T, Happonen R-P, Yli-Urpo A. Compositional dependence of bioactivity of glasses in the system Na2OK2O-MgO-CaO-B2O3-P2O5-SiO2. J Biomed Mater Res 1997; 37:114 –121. 9. Brink M, Yli-Urpo S, Yli-Urpo A. The resorption of a bioactive glass implanted into rat soft tissue. In the 5th World Biomaterials Congress, Toronto, 1996. p 48. 10. Brink M, Laine P, Narva K, Yli-Urpo A. Implantation of bioactive and inert glass fibres in rats—Soft tissue response and short term reactions of the glass.In: Sedel L, Rey C, editors. Bioceramics. New York: Elsevier Science; 1997. Vol 10, p 61– 64. 11. Pirhonen E, Grandi G, To¨ rma¨la¨ P. Bioactive glass fibre/polylactide composite. Key Eng Mater 2001;192–195:725–728. 12. Pirhonen E, Moimas L, Haapanen J. Porous bioactive 3-D glass fibre scaffolds for tissue engineering applications manufactured by sintering technique. Key Eng Mater 2003;240 –242:237–240. 13. Kokubo T, Kushitani H, Sakka S, Kitsugi, Yamamuro T. Solutions able to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-W3 . J Biomed Mater Res 1990;24: 721–734. 14. Hull D, Clyne TW. An Introduction to Composite Materials, 2nd ed. Cambridge: Cambridge University Press; 1996. 15. Hench LL, Andersson O¨ . Bioactive glasses. In: Hench L, Wilson J, editors. An Introduction to Bioceramics. Singapore: World Scientific Publishing; 1993. 16. Izquierda-Barba I, Salinas AJ, Vallet-Regi M. In vitro calcium phosphate layer formation on sol-gel glasses of the CaO-SiO2 system. J Biomed Mater Res 1999;47:243–244. 17. Dunn RL, Cassper RA, Kelley BS. Biodegradable composites. In the 11th Annu Mtg Soc Biomater, San Diego, CA, April 25–28, 1985. p 213. 18. Lin TC. Totally absorbable fibre reinforced composite for internal fracture fixation devices. In the 12th Annu Mtg Soc Biomater, Minneapolis-St. Paul, MN, May 29 –June 1, 1986. p 166. 19. Krebs S, Lin S, King R. Characterization of HA and resorbable glass fibers reinforced PLLA screws. In the 19th Annu Mtg Soc Biomater, April 28 –May 2, Birmingham, AL, 1993. p 317. 20. Marcolongo M, Ducheyne P, Garino J, Schepers E. Bioactive glass fibre/composites bond to bone tissue. J Biomed Mater Res 1997;37:440 – 448. 21. Ja¨rvela¨ P. Properties of Glass Fibres and Their Applications in Porous Composites, Doctoral Thesis. Tampere, Finland: Tampere University of Technology; 1983. 22. Pardini LC, Manhani LGB. Influence of the testing gage length on the strength, Young’s modulus and Weibull modulus of carbon fibres and glass fibres. Mater Res 2002;5:411– 420. 23. Gurvich MR, Dibenedetto AT, Pegoretti A. Evaluation of the statistical parameters of a Weibull distribution. J Mater Sci 1997;32:3711–3716. 24. De Diego MA, Coleman NJ, Hench LL. Tensile properties of bioactive fibres for tissue engineering applications. J Biomed Mater Res B Appl Biomater 2000;53:199 –203. BIOACTIVE GLASS 13–93 FIBERS 233