journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

Dissolution behavior of high strength bioresorbable glass

fibers manufactured by continuous fiber drawing

Timo J. Lehtonenn, Jukka U. Tuominen, Elina Hiekkanen

Vivoxid Ltd, FI-20750 Turku, Finland

ar t ic l e in f o abs tra ct

Article history: This article describes the dissolution behavior of three silica-based resorbable glasses
Received 12 October 2012 manufactured by an industrial-type continuous fiber drawing process yielding fibers with
Received in revised form tensile strength of 1800–2300 MPa. The results of a long-term in vitro degradation testing of
24 January 2013 the manufactured high strength bioresorbable glass fibers are presented. The degradation
Accepted 28 January 2013 was performed by exposing the glass fibers to SBF and TRIS for 26 weeks at physiological
Available online 28 February 2013 conditions at 37 1C. All fibers showed continuous resorption throughout the study and two
of the fibers revealed bioactivity by forming a calcium phosphate (CaP) layer in SBF.
& 2013 Elsevier Ltd. All rights reserved.

1. Introduction Over the past few decades, resorbable polymers have verified
their potential in the fracture fixation field (Nair and Laurencin,
In most cases, bone fractures need an internal fixation to give 2007; Middleton and Tipton, 2000; Hofmann, 1995; Kellomäki
stability and alignment for the fractured bone during the et al., 2000; Rokkanen et al., 2000). Nevertheless, contrary to
healing process. Currently titanium alloys are the standard metals these materials have lower modulus and strength than
materials for the majority of the fracture fixation treatments bone and are therefore suitable only for non-load-bearing
due to their outstanding mechanical properties and versatile applications. Additionally they are lacking osteoconductive
product forms. Even though they dominate the fracture properties. Composites consisting of bioactive filler particles
fixation field, metal and in general non-resorbable implant and resorbable polymers have been investigated as alternatives
materials are not without problems. They can require to metallic implants for fracture fixation (Wagoner Johnson and
removal via a second surgery following the fracture healing, Herschler, 2011; Zhou et al., 2007, 2008, 2012; Navarro et al., 2005;
give rise to a reaction to a foreign body, and interfere with Zhang et al., 2004; Niemelä et al., 2005, 2008; Navarro et al.,
X-ray and magnetic resonance imaging (MRI) examinations. 2008). These types of composites gave good results with respect
In addition to these disadvantages, metal materials such as to bone ingrowth, but the material strength properties even
titanium and surgical stainless steel which are used in decreased with the increasing amount of filler from the unfilled
orthopedic prostheses, possess very high mechanical proper- materials.
ties, especially a high elastic modulus compared to cortical There is still a major need for a resorbable fixation material
bone. This leads to a stress shielding phenomenon when the that would mimic the mechanical properties of bone and
stresses directed towards the implant are not transferred moreover promote bone growth. A potential solution to these
from the implant to the surrounding bone, eventually leading problems is to reinforce the bioresorbable polymer with rein-
to bone loss in the implant site (Gerhardt and Boccaccini, forcing fibers (Andriano et al., 1992; Felfel et al., 2011; Ahmed
2010). et al., 2009; Ahmed et al., 2008). Glass fibers and carbon fibers

n
Corresponding author. Present address: Arkelsedijk 46, 4200 AA Gorinchem, Netherlands. Tel.: þ31 629 582 932; fax: þ31 183 695 606.
E-mail address: [email protected] (T.J. Lehtonen).

1751-6161/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jmbbm.2013.01.032
 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386 377

are reinforcements for technical advanced composites and the After formation of the silicic acid gel, calcium phosphate
fiber strength and stress transfer from the polymer to the fibers layers precipitate on the surface of the silicic acid gel. The
are the most influential factor on the strength of the compo- formation of calcium-phosphate layers on silicic acid gel
sites (Agarwal, 1990; Paradini and Manhani, 2002). Composites surface occur in two stages. First, silanol groups of the gel
made with these fibers are characterized by a high strength/ react with phosphate from solution according to the neutra-
density ratio and their properties are intrinsically related to lization reaction:
their microstructure, i.e., amount and orientation of the fibers,
surface treatment, among other factors. Processing parameters
have an important role in the fiber mechanical behavior
(strength and modulus) (Paradini and Manhani, 2002). In the Then the formed free chains Si–O–PO2–O will be linked
past decades, many different bioactive glass compositions have together by calcium cations.
been manufactured and studied. Nevertheless, most of these
glasses have not been suitable for fiber drawing due to the
thermo-physical behavior (Arstila, 2008), or at least not in an
industrial-type continuous process and have yielded low glass
fiber strengths compared to e.g. E-glass (Ahmed et al., 2008;
Pirhonen et al., 2006; Rinehart et al., 1999), which is typically
used as the reinforcement in dental (Ballo et al., 2007) and These reactions give rise to a three-dimensional structure
technical (Agarwal, 1990) glass fiber reinforced composites. and stabilize the silicate gel. According to the reported in vivo
Besides the good mechanical properties, the degradation rate studies during the initial stages of the formation of newly
of the fibers has to be slow enough to meet the requirements of formed bone the values of the ratio Ca/P are low. Upon
mechanical properties for medical devices in load-bearing crystallization of low-cross-linked carbonate hydroxyapatite
applications (Ahmed et al., 2008; Cozien-Cazuc et al., 2005). (CHA) on the surface of the calcium-phosphate layer, the Ca/P
Given the fact that glass dissolution rate depends on the ratio gradually increases. Eventually, a strong bonding occurs
surface to volume ratio, the dissolution rate and behavior of between the bone tissue and the surface of the implant, when
fibers cannot be directly derived from published values for bulk carbonate hydroxyapatite, and tricalcium phosphate crystal-
bioglasses (Wagoner Johnson and Herschler, 2011; Zhou et al., lize in the matrix. If the implanted material is resorbed, the
2007, 2008, 2012; Navarro et al., 2005, 2008; Zhang et al., 2004; newly formed bone gradually replaces the implant.
Niemelä et al., 2005, 2008) containing similar oxide combina- The aim of this work was to study the dissolution and
tions but made from particles with much smaller surface to degradation behavior of three bioresorbable glass composi-
volume ratios. tions that were drawn to fibers in an industrial-type contin-
Beletskii and Sventskaya (2009) reviewed the studies of the uous spinning process.
resorption mechanism of silica-based bioactive glasses both
in vitro and in vivo. According to their review the bonding
with the bone and soft tissues is determined by the chemical 2. Materials and methods
composition of the glass and the contact medium. The
dissolution of biodegradable glass is caused first by sodium, 2.1. Materials
found in sodium silicate (yNa2O–zCaO–xSiO2) compounds in
the glass. Glass may be chemically stable, bioresorbable and/ Three bioresorbable melt-derived glasses (grades BG1, BG2 and
or bioactive depending on the type of glass dissolution. The BG3, Table 1) of a system Na2O–CaO–MgO–Al2O3–B2O3–P2O5–SiO2
glass type is derived from ‘‘thickness’’ of the silicic acid gel were manufactured and drawn to fibers (targeted average
and calcium phosphate layers during dissolution and the diameter 13 mm) (Lehtonen et al., 2010). The bioresorbable bulk
subsequent bonding reactions. Silicic acid gel forms on the glasses (BG1, BG2 and BG3) were prepared by the traditional
surface of the implant when exposed to physiological media method from mixed raw materials by melting, casting and
(Beletskii and Sventskaya, 2009; Hench, 1998; Banchet et al., annealing. To ensure the homogeneity of the bulk glass, the
2006) according to the reaction: glass was melted twice for 3 h in 1420 1C and between melting’s
crushed to cullet. Annealing was carried out at 600 1C after
which the temperature was slowly reduced to ambient. The
casted bulk glass cubes (15  15  15 mm3) were drawn into

Table 1 – The nominal oxide compositions of the fibers [wt%].

Glass SiO2 Na2O CaO MgO P2 O 5 B2O3 Al2O3

BG1 69.8 12.0 12.0 3.1 1.5 1.1 0.5

BG2 67.8 14.0 9.0 5.4 1.5 2.3 –
BG3 70.0 17.5 10.0 2.0 0.5 – –
378 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

fibers (204 glass filaments) at the Leibniz Institute of Polymer E-glass. The measurements were carried out after 0, 4, 8, 16
Research, Dresden (IPF) in a platinum-covered melting furnace and 26 weeks, but for the ion release and composition also at
with platinum bushing (Lehtonen et al., 2010; Massera et al., 2 weeks’ time point. BG1 and BG2 were studied in both
2011). The fiber manufacturing processing parameters are pre- solutions, BG3 only in TRIS and E-glass only in SBF. Over this
sented in Table 2. A sizing was used to coat the fibers for easier period the pH of the buffer solutions was monitored, and the
handling and to increase adhesion between filaments to form buffer solutions were changed to maintain a sufficient buffer-
bundles and moreover increase adhesion to the polymer matrix ing capacity and sink conditions. The TRIS buffer solution
in the composite manufacturing. The sizing was added on-line was changed once a week and SBF once a month. Concomi-
as a water-based dispersion consisting of the coupling agent tant with each buffer change, the TRIS buffer was filtered and
3-glycidoxypropyltriethoxysilane (Dynasylan GLYEO, Evonik collected for ICP-OES analysis.
Degussa GmbH) and compatibilizer polycaprolactone PCL
(Lehtonen and Tuominen, 2010). Sizing content i.e. organic 2.3. Characterization of the bioresorbable glass fibers
content in the surface of the glass was determined by pyrolysis
at 650 1C for 30 min (Nabertherm L15/11). Glacial acetic acid 2.3.1. Mechanical properties by single fiber tensile test
(Sigma Aldrich) was used as the catalyst for hydrolysis of GLYEO. The tensile strength of the single glass fiber filaments were
The drawing was successfully accomplished with this contin- determined using the Lloyd LRX universal mechanical tester
uous process without considerable single filament breakage equipped with a 5 N load cell, according to ASTM C1557-03
yielding several kilograms of fibers. BG-glass fiber bobbins were (Standard Test Method for Tensile Strength and Young’s
dried in an air-circulating oven (Memmert) at 75 1C for 12 h, and Modulus of Fibers). The test uses a gauge length of 25.4 mm
strands were multiplied to achieve the final glass fiber yarn TEX and a cross head velocity of 2.0 mm/min. The diameters of
280. The fibers were stored in a vacuum oven at RT prior to fibers were measured with a Struers Duramin-5 micro hard-
sample preparation. Epoxysilane sized E-glass was direct drawn ness tester’s microscope, with an accuracy of 0.1 mm. The
roving grade R338-300 and was used as received from Ahlström, mechanical properties were measured for all fibers in both
and used as a chemically stable control material. buffer solutions after every time period. A test was performed
for dried fibers (drying 1 day/r5 mbar/RT), and a total of
2.2. Dissolution approximately 30 parallel fibers were tested. Tensile strength
was calculated per the formula:
Dissolution of the fibers was conducted in two buffer solu- a
X¼ ð1Þ
tions, namely Tris(hydroxymethyl) aminomethane (TRIS) and pðd=1000=2Þ2
Simulated Body Fluid (SBF), to determine the degradation rate
where X is the tensile strength [MPa], a is the maximum
and the change in material properties after different time
load [N] and d is the diameter of the fiber [mm] and results are
periods. TRIS solution was prepared by diluting TRIZMA pre-
presented as mean value with standard deviation.
set crystals (Sigma-Aldrich) into de-ionized water. SBF was
Due to the behavior of thin single fiber filaments and the
prepared following the instructions made by Kokubo et al.
high number of parallel measurements, a statistical evalua-
(1990). 1.5 g of fibers was immersed in 45 ml of buffered
tion is required to understand mechanical properties of glass
solution in a 50 ml falcon tube. TRIS was selected for ion
fibers. The most used statistical tool to describe the varia-
release and composition studies to follow the resorption of
bility in strength for the brittle materials is the Weibull
the glass fibers BG1, BG2 and BG3, because TRIS does not
modulus (Spencer and Stevens, 1990). The Weibull modulus
contain the same inorganic components as the glasses and
is not a material constant, but gives a good indication of how
therefore does not interfere with the ion release measure-
homogeneous the material is. The Weibull modulus can be
ment. Correspondingly, SBF was used to analyze the calcium
regarded as a flaw frequency distribution factor. High values
phosphate precipitation of BG1, BG2 and the reference
(44) indicates that flaws are evenly distributed throughout
Table 2 – Fiber drawing parameters. the material, whether they are plentiful or not, and hence
strength is nearly independent of the length. Low values are
Glass Drawing speed (m/min) Average titer Yield (kg) indicating that flaws are fewer and less evenly distributed,
causing greater scatter in strength, thus Weibull modulus is a
BG1 424 69.2 TEX 3.9
BG2 381 69.1 TEX 4.0
tool to evaluate the fiber’s homogeneity in terms of proper-
BG3 410 69.5 TEX 2.8 ties. The Weibull modulus is determined by plotting (Eq. (2))
ln [ln(1/(1P))] as function of s and then the Weibull

Table 3 – Initial properties of the studied glass fibers.

Fiber Average diameter (lm) Tensile strength (Mpa) Young’s modulus (Gpa) Sizing content (wt%)

BG1 13.073.0 23117258 68.677.3 1.0370.02

BG2 13.972.7 19837244 65.375.7 0.6370.07
BG3 14.572.2 18127306 67.378.7 0.8970.01
E-glass 12.771.1 29557440 75.775.4 0.5070.03
 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386 379

Fig. 1 – Change of tensile strength (MPa), Young’s modulus (GPa) and Weibull modulus during degradation in TRIS and SBF.
(Loss of integrity of the sample due to degradation.)
380 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

parameter (m) is obtained by linear regression (y¼axþb) SEM-EDXA was used to monitor changes in the glass surface
(Paradini and Manhani, 2002; Spencer and Stevens, 1990). (compositional analysis of the formed surface layers).
 
1
ln ln ¼ mln smlns0 ð2Þ
1P
3. Results
where P is the probability of failure, m is the Weibull
modulus, s is the tensile strength and s0 is the scale 3.1. Characterization of the glass fibers
parameter.
All three resorbable glass compositions were successfully
drawn to amorphous fibers and sized on-line. In Table 3,
2.3.2. Composition by XRF and ion release by ICP-OES
the average fiber diameters, sizing contents and tensile
The changes in the glass compositions were analyzed with
properties with their standard deviations after spinning are
X-ray fluorescence (XRF) at Labtium Oy (Espoo, Finland). The
presented. The targeted theoretical value for sizing amount
fibers immersed in TRIS were removed from the containers
for BG-glass fibers was 1.0 wt%. The tensile properties of the
using tweezers and dried in a vacuum oven before analysis.
spun resorbable single glass fibers were determined from
All of the glass components were analyzed except boron,
app. 30 parallel measurements.
which is not reliably measurable by XRF and B2O3 was
determined as the non-measured rest from the difference
of the sum of concentrations with 100%. For the ion release 3.2. Tensile properties
analysis, the dissolution media (TRIS and the released ions
from the glasses) was filtered at each buffer change. Analyses The tensile strength properties of the bioresorbable glasses
were performed with inductively coupled plasma optical during the dissolution were studied both in TRIS and in SBF
emission spectrometry (ICP-OES, PerkinElmer Optima 5300 up to 26 weeks. The results of the single fiber tensile strength
DV) at Åbo Akademi University (Turku, Finland). The results tests are presented in Fig. 1. From the single fiber tensile tests
of XRF were compared to the ICP-OES results as the ion conducted it was seen that the bioresorbable glass composi-
release to the surrounding dissolution media decreases in tion had a significant effect on the degradation behavior. The
proportion to the concentration of elements in the degrading decrease in BG2 fiber mechanical properties was drastic,
glass fibers. whereas BG1 and BG3 lost their tensile strength more slowly.
BG2 lost 55% of its initial strength within 4 weeks of dissolu-
tion in TRIS, after which time the mechanical properties
2.3.3. SEM analysis remained fairly stable. On the contrary, the tensile strength of
A system combining a scanning electron microscope (SEM, BG1 and BG3 decreased more gradually. The buffer composi-
Leo 1530 Gemini from Zeiss) and Energy-Dispersive X-Ray tion also had an effect on the degradation of the glasses. BG1
analysis (EDXA, Vantage by Thermo Electron Corporation) lost its tensile strength faster in SBF than in TRIS and for BG2
was used for the determination of the fibers physical and the effect seemed to be reversed. The BG2 fiber strength after
chemical changes in SBF at Åbo Akademi University (Turku, 2 weeks in TRIS was the same as that after 8 weeks in SBF
Finland). For the imaging, the fiber bundles were molded into and after 26 weeks in SBF BG2 fibers had lost its integrity and
Epoxy resin and cut perpendicularly to the fibers to achieve a the measurement was impossible to conduct. Over the 26
cross-sectional view of the fiber. The cut surface was polished weeks of dissolution, the tensile strength of BG1 decreased
using no. 320, 1200, 2400 and lastly 4000 silicon carbide from the initial 2311 MPa to 1568 MPa in TRIS and 928 MPa in
papers (Struers) and carbon sputtered for the SEM analysis. SBF, of BG2 from 1983 MPa to 682 MPa in TRIS and to zero in

Fig. 2 – Linear regression to obtain Weibull modulus of BG1 in TRIS.

 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386 381

SBF, of BG3 from 1812 MPa to 1147 MPa in TRIS and 963 MPa in The plot of Eq. (2) and linear regression to obtain Weibull
SBF. Although E-glass is generally considered to be rather modulus of BG1 in TRIS is presented in Fig. 2 as an example of
stable, it also lost 44% of its initial tensile strength in 26 Weibull modulus determination. The Weibull modulus of all
weeks of SBF immersion. fiber types (Fig. 1) indicted narrow scatter in strength (m44)

Fig. 3 – Release of ions to the degradation medium (TRIS) during dissolution.

382 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

data before dissolution (time point 0 weeks) and bioresorb- silicon was detected from the specimens, especially from
able glass fibers had Weibull modulus at same level as BG3, even though from compositional XRF analysis (Table 4)
commercial E-glass fiber. At every time point after 8 weeks, the silicon amount increases, which indicates formation of Si
the BG2 fiber Weibull modulus was less than 3, indicating rich layer in the surface of fibers. The composition of the
very large scatter in strength data. On the other hand, BG1 glasses has a significant effect on the ion release and thereby
and BG3 fiber Weibull modulus were significantly higher the degradation of the glasses. The released amount of
during the whole study, and these fibers also retained their sodium is directly related to the sodium content in the
mechanical properties longer. studied glasses. Similar phenomena were detected for the
other alkaline metals (Ca2þ and Mg2þ) as well.
3.3. Ion release and change in composition As can be observed from the ICP-OES data the degradation
and ion release of fibers starts quite rapidly and then
Ion release from glasses BG1, BG2 and BG3 in TRIS was decreases. The ion release profile in BG1 and BG2 was similar
analyzed with ICP-OES after 2, 4, 8, 16, and 26 weeks. From for all the elements, where the highest dissolution occurs
the data obtained (Fig. 3), a rapid release of sodium and with silicon. With BG3 glass the highest dissolution occurs
with sodium and also the total ion release in higher level
Table 4 – Change in glass composition during dissolution than with BG1 and BG2 (Fig. 4). There were only very slight
in TRIS [wt%]. changes in the phosphorus concentrations leaching from BG3
glass, since BG3 glass contains only one-third the phosphor-
SiO2 Na2O CaO MgO P2 O 5 B2O3a Al2O3
ous of BG1 and BG2 (0.5% vs. 1.5%), this low concentration
BG1 in TRIS might indicate a non-bioactive nature for BG3. Only a mini-
0 69.5 12.2 11.9 3.4 1.6 1.1 0.4 mal amount of aluminum was detected from BG1 fibers
2 69.7 12.0 11.8 3.3 1.5 1.1 0.4
soaking in TRIS during the 26-week study. Aluminum can
4 69.8 12.0 11.8 3.3 1.5 1.1 0.4
migrate to the surface of the glass, but it does not leach into
8 70.0 11.8 11.7 3.3 1.5 1.1 0.4
16 70.5 11.6 11.6 3.2 1.5 1.1 0.5 the soaking solution (Anderson, 1990). Therefore the amount
26 70.8 11.4 11.4 3.3 1.5 1.1 0.5 of aluminum in the glass should be small enough to reduce
BG2 in TRIS the dissolution rate but not to entirely prohibit the other ions
0 67.8 14.2 9.2 5.1 1.5 2.3 – from leaching as is the case with E-glass fibers. As the
2 68.0 13.9 9.2 5.0 1.5 2.3 – degradation proceeded, the relative amount of silicon in all
4 68.1 13.8 9.2 5.0 1.5 2.3 –
the studied glasses increased (Table 4). This was expected as
8 68.3 13.7 9.2 5.0 1.5 2.3 –
16 68.4 13.6 9.1 5.0 1.5 2.3 –
silica gel is usually the final degradation product of silica
26 68.6 13.5 9.0 5.0 1.5 2.3 – based glasses (Cozien-Cazuc et al., 2005; Beletskii and
BG3 in TRIS Sventskaya, 2009; Hench, 1998; Anderson, 1990).
0 70.4 17.3 9.3 2.5 0.5 – –
2 70.8 17.0 9.3 2.4 0.5 – –
3.3.1. SEM analysis
4 70.9 16.8 9.4 2.4 0.5 – –
8 71.2 16.6 9.3 2.4 0.5 – –
SEM analyses were conducted on the BG1, BG2 and E-glass
16 71.7 16.3 9.2 2.4 0.5 – – fibers before soaking and for post-degradation periods at 4, 8,
26 72.3 15.9 9.0 2.3 0.5 – – 16 and 26 weeks. Both BG1 and BG2 revealed CaP layer on the
fiber surface at 8 weeks’ time point, thus the CaP formation
a
B2O3 is determined as the non-measured rest from the difference
occurs during 4–8 weeks’ time period soaking in SBF, which
of the sum of concentrations with 100%.
was also verified by EDX sweep analysis (Fig. 6). After initial

Fig. 4 – Total ion release to the degradation medium (TRIS) during dissolution.
 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386 383

precipitation of CaP to the glass surface in both BG1 and BG2 calcium phosphate formation on the fiber surface at any
(Fig. 5 8 weeks), the CaP layer continues to grow thick in time point of the study (Figs. 5 and 6). However, the surface of E-
(Fig. 5 16 weeks, BG1 and BG2) until the fiber diameter starts glass did not show absolute physical stability as SEM pictures
to decrease faster than CaP-layer is growing. The fiber revealed cracking of the surface after immersion in SBF,
diameter reduction is caused by fiber resorption through which was also observed as decrease in tensile strength.
surface erosion due to the high surface to volume ratio of
thin (app. 13 mm) fibers. The loose Si-rich gel layer has been
measured as fragments in the vicinity of fibers and under the 4. Discussion
CaP-layer as shown in graphical abstract (SEM figure and EDX
graph). The EDX sweep analysis shows a decline in Si- Fiber strength and stress transfer from fibers to matrix are
concentration, but its relative amount to the other glass the most critical aspects when designing a composite for
elements increases in the range of 3.5–5.0 mm before Ca, P load-bearing applications. Therefore, the properties of struc-
and O concentrations becomes predominant in the range of tural composites are a function of both fiber distribution
5.0–8.0 mm. Moreover the BG2 fibers showed thicker and more strength at short gauge lengths and fiber elastic modulus. On
uniform precipitation of calcium phosphate than BG1 fibers the other hand, composite properties are highly influenced by
and the BG2 fibers lost their integrity in 26 weeks’ time point. fiber/polymer matrix bonding (Paradini and Manhani, 2002).
SEM analysis also revealed the surface erosion type resorp- One of the drawbacks when considering the use of bioresorb-
tion mechanism for the BG glass fibers, which is in accor- able fibers as reinforcement in a biodegradable polymer
dance with mechanical and dissolution results. The SEM composite has been the insufficient fiber strength. In this
images of the cross-sections of the glass fibers after immer- study, three novel bioresorbable glasses were developed and
sion in SBF are presented in Fig. 5. The control used for this successfully drawn to fibers with an industrial type of con-
investigation (E-glass) demonstrated chemical stability by not tinuous spinning process. The defects generated during
exhibiting any chemical alterations i.e. silica-rich layer or processing and from handling, such as ultra micropores, fibril

Fig. 5 – SEM images of cross-sections of fibers during immersion in SBF: (a–d) BG1 at 4, 8, 16 and 26 weeks respectively; (e–h)
BG2 at 4, 8, 16 and 26 weeks respectively; (i–l) E-glass at 4, 8, 16 and 26 weeks respectively. CaP formation on the BG1 and BG2
fiber surfaces at 8, 16 and 26 weeks. (Si gel layer dark gray region between BG1 fiber and CaP layer.)
384 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

Sweep Epoxy
CaP layer direction (void) BG Fiber Si gel CaP layer

—O
Si gel
—Si

BG Fiber —Ca
—Na
—Mg —Al —P

Epoxy
(void) E-glass Fiber
—O

—Si
E-glass
Fiber —Ca

Sweep —Al
direction
—Na —Mg

Fig. 6 – EDX elemental sweep analysis from core to surface (a) SEM-EDX from BG1 after 26 weeks dissolution in SBF and
(b) SEM-EDX from E-glass after 26 weeks dissolution in SBF.

misalignments and impurities are the main source for However, the tensile strengths of the bioresorbable fibers
decreasing strength and failure of fibers by crack initiation. were after 26 weeks of immersion still much higher than the
Therefore industrial type of continuous fiber spinning and average initial tensile strength of e.g. phosphate glass fibers
handling process is extremely important to use in order to which are largely studied as a potential reinforcement for
yield high strength bioresorbable fibers comparable to E-glass bone fracture fixation devices (Ahmed et al., 2008; Paradini
fibers, which is typically used as a reinforcing phase in dental and Manhani, 2002). In contrast BG2 fiber lost its integrity
and technical applications (Ballo et al., 2007). The goodness of after 26 weeks in SBF, showing the importance of under-
fiber processing and handling was evaluated in this work standing the effects of different dissolution media properties.
with the Weibull modulus describing flaw frequency distribu- However, at the same time the ion release should occur
tion. The Weibull modulus before dissolution for E-glass was sufficiently fast to support the osteogenesis of the tissue. All
8 and for bioresorbable glass fibers 6.2–9.9. So, the in scatter the studied bioresorbable glasses showed signs of resorption
in strength and the quality of the bioresorbable fibers were in in TRIS in terms of ion release. In addition, with BG1 and BG2
the same range as E-glass fiber. The tensile strength of the glass fibers the ion release and composition studies showed
fibers after spinning was 1800–2300 MPa, which is the highest CaP precipitation in SBF, an indication of bioactivity in vitro.
tensile strength reported (Ahmed et al., 2008; Paradini and The calcium phosphate precipitation was verified with SEM
Manhani, 2002; Arstila, 2008; Pirhonen et al., 2006; Rinehart images and EDX analysis where the calcium phosphate layer
et al., 1999) for bioresorbable glasses and reasonably close to was detected between 4 and 8 weeks of immersion in SBF.
that of E-glass which is a commonly used high strength fiber Nevertheless, the CaP layer was not always detected as a
grade for applications where high fiber strength is required. uniform layer on the fiber surface but as fragments in the
In bioresorbable load-bearing composites, it is essential vicinity of the fibers. In addition SEM-EDX analysis showed a
that the fibers also retain an adequate level of strength by silica rich layer for BG1. This phenomenon, where soluble
dissolving slowly enough to ensure sufficient reinforcement silica is first lost to the solution and then polymerized back
for the composite. The tensile strength of the manufactured onto the fiber surface, is common to silica-based bioactive
high strength bioresorbable glass fibers was studied in TRIS glasses, thus indicating bioactivity in BG1 and BG2 glasses.
and SBF. The loss of strength occurred gradually for BG1 and Although BG3 retains its mechanical properties longer
BG3 whereas the degradation profile for BG2 was more rapid. than BG2, there is a concern about its rapid release of alkali
 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386 385

metals, especially sodium. When used as reinforcement in a Ballo, A.M., Lassila, L.V., Vallittu, P.K., Närhi, T.O., 2007. Load
bioresorbable polymer, the ion release of the glass fibers is of bearing capacity of bone anchored fiber-reinforced composite
great importance. The leaching of alkalis should occur rela- device. Journal of Materials Science: Materials in Medicine 18,
2025–2031.
tively slowly so as not to give rise to local pH, which occurs in
Banchet, V., Michel, J., Jallot, E., Wortham, L., Bouthors, S.,
the interfacial (fiber–polymer) region. The rapid raise of local
Laurent-Maquin, D., Balossier, G., 2006. Interfacial reactions
pH is known to occur with standard bioactive glasses such as of glasses for biomedical application by scanning
45S and S53P4 (Anderson, 1990; Zhang et al., 2008). When transmission electron microscopy and microanalysis. Acta
leached from the glass, the alkalis react with the hydroxyl Biomaterialia 2, 349–359.
ions of the surrounding fluids forming alkali hydroxides, Beletskii, B.I., Sventskaya, N.V., 2009. Silicon in living organism
strong bases, which in turn are very corrosive for both the and new-generation biocomposite materials (review). Glass
glass fibers and polymer in the composite material Ceramic 66, 104–108.
Cozien-Cazuc, S., Parsons, A.P., Burling, L.K., Walker, G.S., Jones,
(Ehrenstein and Schmiemann, 1990). The degradation of the
I.A., Robitaille F., Rudd, C.D., 2005. Suitability of phosphate
interfacial region causes pores to form in the structure, and glass fibres for resorbable bio-medical composites. In:
further facilitates the hydrolytic breakdown of the polymer Proceedings of the International Conference on Composite
leading to premature rupture of the composite. The effect of materials, ICCM15, Durban, South Africa.
the ion release rate on the degradation of the composite has Ehrenstein, G.W., Schmiemann, A., 1990. Corrosion phenomena
been studied separately by authors by impregnating the glass in glass-fiber reinforced thermosetting resins. In:
fibers with biodegradable polymer and studying the changes Cheremisinoff, N.P. (Ed.), Handbook of Ceramics and
Composites, vol 1: Synthesis and Properties. Marcel Dekker
in the composite properties in vitro (Lehtonen et al., 2013).
Inc., New York and Basel, pp. 231–268.
Felfel, R.M., Ahmed, I., Parsons, A.J., Walker, G.S., Rudd, C.D., 2011.
In vitro degradation, flexural, compressive and shear
properties of fully bioresorbable composite rods. Journal of the
5. Conclusions Mechanical Behavior of Biomedical Materials 4, 1462–1472.
Gerhardt, L.-C., Boccaccini, A.R., 2010. Bioactive glass and glass-
We have designed, manufactured and characterized three ceramic scaffolds for bone tissue engineering. Materials 3,
novel bioresorbable glass fibers with initial tensile properties 3867–3910.
close to E-glass fibers. Continuous dissolution of the biore- Hench, L.L., 1998. Bioceramics. Journal of the American Ceramic
sorbable glasses showed degradation for all three of the novel Society 81, 1705–1728.
Hofmann, G.O., 1995. Biodegradable implants in traumatology: a
glasses. Two of these glasses additionally showed bioactive
review on the state-of-the-art. Archives of Orthopaedic and
characteristics when immersed in SBF by calcium phosphate
Trauma Surgery 114, 123–132.
formation on the fiber surface. The studied fibers show great Kellomäki, M., Pohjonen, T., Törmälä, P., 2000. Self-reinforced
potential as reinforcement in load-bearing medical devices polylactides. Optimization of degradation and mechanical
for bone fracture fixation in orthopedics. Additionally, by properties. In: Arshady, R. (Ed.), The PBM Series, vol. 2. Citus
tailoring the composition of the glasses, it is possible to Books, pp. 211–235.
manufacture reinforcing glass fibers for a range of biomedical Kokubo, T., Kushitani, H., Sakka, S., Kitsugi, T., Yamamuro, T.,
1990. Solutions able to reproduce in vivo surface-structure
applications with tailored mechanical properties, degree of
changes in bioactive glass-ceramic A-W3. Journal of
bioactivity and different resorption profiles.
Biomedical Materials Research A 24, 721–734.
Lehtonen, T., Tuominen, J., Ollila F., 2010. Resorbable and
r e f e r e n c e s biocompatible fiber glass compositions and their uses. World
Pat. Inf. WO2010122019A1.
Lehtonen, T., Tuominen, J., 2010. Biocompatible composite and its
Agarwal, B.D., 1990. Fracture toughness of fiber-reinforced use. Granted Eur. Pat. EP 2243500B1.
composites. In: Cheremisinoff, N.P. (Ed.), Handbook of Lehtonen, T.J., Tuominen, J.U., Hiekkanen, E., 2013. Resorbable
Ceramics and Composites, vol 1: Synthesis and Properties. composites with bioresorbable glass fibers for load-bearing
Marcel Dekker Inc., New York and Basel, pp. 269–305. applications. In vitro degradation and degradation
Ahmed, I., Cronin, P.S., Abou Neel, E.A., Parsons, A.J., Knowles, mechanism. Acta Biomaterialia 9, 4868–4877.
J.C., Rudd, C.D., 2009. Retention of mechanical properties and Massera, J., Claireaux, C., Lehtonen, T., Tuominen, J., Hupa, L.,
cytocompatibility of a phosphate-based glass fiber/polylactic Hupa, M., 2011. Control of the thermal properties of slow
acid composite. Journal of Biomedical Materials Research B bioresorbable glasses by boron addition. Journal of Non-
89B, 18–27. Crystalline Solids 357, 3623–3630.
Ahmed, I., Parsons, A.J., Palmer, G., Knowles, J.C., Walker, G.S., Middleton, J.C., Tipton, A.J., 2000. Synthetic biodegradable
Rudd, C.D., 2008. Weight loss, ion release and initial polymers as orthopedic devices. Biomaterials 21, 2335–2346.
mechanical properties of a binary calcium phosphate glass Nair, L.S., Laurencin, C.T., 2007. Biodegradable polymers as
fibre/PCL composite. Acta Biomaterialia 4, 1307–1314. biomaterials. Progress in Polymer Science 32, 762–798.
Anderson, Ö., 1990. The bioactivity of silicate glass. Doctoral Navarro, M., Ginebra, M.P., Planell, J.A., Barrias, C.C., Barbosa,
Dissertation. Åbo Akademi University, Turku, Finland. M.A., 2005. In vitro degradation behavior of a novel
Andriano, K.P., Daniels, A.U., Heller, J., 1992. Biocompatibility and bioresorbable composite material based on PLA and a soluble
mechanical properties of a totally absorbable composite CaP glass. Acta Biomaterialia 1, 411–419.
material for orthopaedic fixation devices. Journal of Applied Navarro, M., Engel, E., Planell, J.A., Amaral, I., Barbosa, M.,
Biomaterials 3, 197–206. Ginebra, M.P., 2008. Surface characterization and cell
Arstila, H., 2008. Crystallization characteristics of bioactive response of a PLA/CaP glass biodegradable composite
glasses. Doctoral dissertation. Åbo Akademi University, material. Journal of Biomedical Materials Research A 85A,
Turku, Finland. 477–486.
386 journal of the mechanical behavior of biomedical materials 20 (2013) 376 –386

Niemelä, T., Niiranen, H., Kellomäki, M., Törmälä, P., 2005. Self- Spencer, A.M., Stevens, G.T., 1990. Techniques for strain
reinforced composites of bioabsorbable polymer and bioactive characterization of materials. In: Cheremisinoff, N.P. (Ed.),
glass with different bioactive glass contents. Part I: Initial Handbook of Ceramics and Composites, vol 1: Synthesis and
mechanical properties and bioactivity. Acta Biomaterialia 1, Properties. Marcel Dekker Inc., New York and Basel, pp. 1–21.
235–242. Wagoner Johnson, A.J., Herschler, B.A., 2011. A review of the
Niemelä, T., Niiranen, H., Kellomäki, M., 2008. Self-reinforced mechanical behavior of CaP and CaP/polymer composites for
composites of bioabsorbable polymer and bioactive glass with applications in bone replacement and repair. Acta
different bioactive glass contents. Part II: In vitro degradation. Biomaterialia 7, 16–30.
Acta Biomaterialia 4, 156–164. Zhang, D., Hupa, M., Hupa, L., 2008. In situ pH within particle
Paradini, L.C., Manhani, L.G.B., 2002. Influence of the testing gage beds of bioactive glasses. Acta Biomaterialia 4, 1498–1505.
Zhang, K., Wang, Y., Hillmyer, M.A., Francis, L.F., 2004. Processing
length on the strength, Young’s modulus and Weibull
and properties of porous poly(L-lactide)/bioactive glass
modulus of carbon fibres and glass fibres. Materials Research
composites. Biomaterials 25, 2489–2500.
5, 411–420.
Zhou, Zhi-hua, Ruan, Jian-ming, Zou, Jian-peng, Zhou, Zhong-
Pirhonen, E., Niiranen, H., Niemelä, T., Brink, M., Törmälä, P.,
cheng, Shen, Xiong-jun, 2007. Bioactivity of bioresorbable
2006. Manufacturing, Mechanical Characterization, and In
composite based on bioactive glass and poly-L-lactide.
Vitro Performance of Bioactive Glass 13–93 Fibers. Journal of
Transactions of Nonferrous Metals Society of China 17,
Biomedical Materials Research B: Applied Biomaterials 77B, 394–399.
227–233. Zhou, H., Lawrence, J.G., Bhaduri, S.B., 2012. Fabrication aspects of
Rinehart, J.D., Taylor, T.D., Tian, Y., Latour Jr., R.A., 1999. Real-Time PLA-CaP/PLGA-CaP composites for orthopedic applications: a
Dissolution Measurement of Sized and Unsized Calcium review. Acta Biomaterialiahttp://dxdoi.org/10.1016/
Phosphate Glass Fibers. Journal of Biomedical Materials j.actbio.2012.01.031.
Research B: Applied Biomaterials 48, 833–840. Zhou, Zhi-hua, Ruan, Jian-ming, Zou, Jian-peng, Zhou, Zhong-
Rokkanen, P.U., Böstman, O., Hirvensalo, E., Mäkelä, E.A., Partio, cheng, Chen, Liang-long, 2008. Preparation of poly-L-lactide/
E.K., Pätiälä, H., Vainionpää, S., Vihtonen, K., Törmälä, P., 2000. bioactive glass composite and evaluation of cytotoxicity
Bioabsorbable fixation in orthopaedic surgery and in vitro. Transactions of Nonferrous Metals Society of China
traumatology. Biomaterials 21, 2607–2613. 18, 1151–1156.