Poly(L-lactic acid) (PLLA), a typical biodegradable polymer made
from bioresources, has extensively been used in many industrial fields
such as automotive, electrical, and medical industries. PLLA has some
potential advantages such as environmental degradability and
biocompatibility and also strength and modulus comparable to those of
commercially available engineering polymers. In the medical industry,
PLLA also has great potential of applications as bioabsorbable medical
devices such as bone fixation plates, screws, and rods used in
orthopedic and oral surgeries [1-3]. However, PLLA exhibits brittle
fracture behavior, especially, under impact loading conditions and,
therefore, the toughening of PLLA becomes one of the most important
issues in the field of biopolymer engineering . Several approaches
have been used to improve the mechanical properties of PLLA, for
example, blending with a ductile biodegradable polymer is known to be an
effective way [5-20]. The common ductile biodegradable polymers used in
such PLLA blending are poly([epsilon]-caprolactone) (PCL) [8, 17, 18],
poly(butylene succinate) (PBS) [13, 15, 20], poly(butylene
succinate-co-L-lactate) (PBSL) [13, 14, 20], and poly(butylene
succinate-co-[epsilon]-caprolactone) (PBSC) [19, 20]. It is known that
in general these PLLA polymer blends exhibit phase-separation
morphology, e.g., secondary polymers are dispersed in PLLA matrix when
the polymers are directly blended without any additives . Just
recently, it was found that addition of lysine triisocyanate (LTI) can
effectively improve the immiscibility in PLLA/PCL, PLLA/PBS, and
PLLA/PBSL blends, and, as a result, their mechanical properties become
higher [15-19]. However, effects of LTI addition on PLLA/PBSC blend have
not been clarified yet.
The objective of this study was therefore to characterize the
effects of LTI addition on the microstructures and the mechanical
properties of PLLA/PBSC blend. Differential scanning calorimetry (DSC),
Fourier transform infrared (FTIR) analyses, and scanning electron
microscopy (SEM) of cryofractured surfaces were performed to
characterize the microstructural modification generated in the blend due
to LTI addition. Mode I fracture and three-point bending tests were also
carried out to assess the change appeared in the mechanical properties.
These macroscopic properties were then correlated with the microscopic
Materials and Specimens
PLLA pellets ([M.sub.w] = 1.45 X [10.sup.5]) and PBSC pellets
([M.sub.w] = 1.7 X [10.sup.5]) were supplied by Toyota Motor Co. and
Daicel Chemical Industries, respectively. These pellets were held into a
desiccator to keep them dry and to prevent them from degradation because
of hydrolysis by moisture. High viscous solution of LTI was supplied by
Kyowa Co. The chemical structures of PLLA, PBSC, and LTI are shown in
Fig. 1. Blends of PLLA and PBSC, with and without addition of LTI, were
prepared by melt-mixing in a conventional melt-mixer at 190[degrees]C
and at a rotor speed of 50 rpm for 20 rain. Blend ratios of PLLA and
PBSC were chosen to be 90/10, 80/20, and 70/30 in weight fraction, and
the amount of LTI mixed with the blends was fixed to be 2 phr. The
mixtures were then compression molded at 30 MPa for 190[degrees]C, and
then followed by cooling process using a water cooling system to produce
sheets of 140 X 140 X 2 [mm.sup.3]. Sheets of neat PLLA were also
fabricated through the same molding process. Bend specimens and
single-edge-notch-bend (SENB) specimens for mode I fracture tests were
then prepared from these sheets. The dimensions of both testing samples
were approximately 50 X 10 X 2 [mm.sup.3]. Table 1 exhibits the sample
labeling with corresponding compositions.
[FIGURE 1 OMITTED]
FTIR and DSC Measurements
FTIR measurements were carried out using Bio-Rad FTS-6000
spectrometer with resolution of 8 [cm.sup.-1] for 32 scans over wave
number range of 400-4000 [cm.sup.-1]. The samples were obtained by
casting films of polymer resolution on the KBr disc. The solvent
(chloroform) was evaporated at room temperature for at least 5 min
before recording the IR spectra.
DSC analysis was conducted using DSC-60 installed with TA-60WS
thermal analysis system (Shimadzu Co.). Approximately 8-10 mg samples
were weighted and put into an aluminum pan with a cover. The samples
were scanned from-100 to 230[degrees]C at a scan rate of
10[degrees]C/min under nitrogen atmosphere to determine the thermal
property of the samples. The melting temperature, [T.sub.m], and the
glass transition temperature, [T.sub.g], of the neat polymers and the
blends were then evaluated from the DSC curves, and the intrinsic degree
of crystallinity, [X.sub.c] PLLA, was also determined based on the
[X.sub.c PLLA](%) = [[([DELTA][H.sub.c] - [DELTA][H.sub.m]) x
100]/[93 x [X.PLLA]]] (1)
where [delta][H.sub.c] and [delta][H.sub.m] are the enthalpies of
crystallization and melting of PLLA, respectively, and the constant of
93 J/g is the fusion enthalpy of PLLA . [X.sub.PLLA] is the weight
fraction of PLLA.
Three-point bending tests of the beam specimens were performed at a
load-rate of 1 mm/min using a servo-hydraulic testing machine.
Load-displacement relations were recorded using a digital recorder. For
all testing measurements, at least five specimens were tested to obtain
the average and standard deviation (SD). The ultimate bending strength
and the flexural modulus were then determined using the following
[[sigma].sub.f] = [[3PL]/[2b[h.sup.2]]] and E =
where P is the maximum load. S is the slope of the initial linear
portion of the load-displacement curve. L, b, and h are the span length,
width, and the thickness, respectively.
Mode I fracture tests of the SENB specimens were performed at a
loading-rate of 1 mm/min by using the servo-hydraulic testing machine.
Load-displacement relations were recorded using the digital recorder.
Mode I fracture properties such as the initial fracture energy,
[J.sub.in], which is defined as the critical J-integral value at crack
initiation, and the averaged fracture energy, [J.sub.f], were then
evaluated using the following formulae:
[J.sub.in] = [[[eta][U.sub.in]]/[B(W - a)]] (3)
[J.sub.f] = [[U.sub.f]/[B(W - a)]] (4)
where [U.sub.in] is the critical energy at crack initiation that is
defined as the point where the stiffness of the specimen starts to
rapidly decrease. [U.sub.f] is the total fracture energy that is
dissipated by the complete fracture of the specimen. B and W are the
specimen thickness and width, respectively, [alpha] is the initial crack
length and [eta] the geometrical correction factor, where [eta] = 2 for
the standard SENB specimen.
Cryo-fracture surfaces were obtained by immersing the specimens
into liquid nitrogen for 30 min. The surfaces were then observed using a
field emission SEM (FE-SEM; HITACHI S-4100) to characterize
microstructural morphology. Fractured surfaces of the SENB specimens
were also observed by FE-SEM to characterize microscopic deformation
mechanisms and effect of LTI addition on the fracture behavior.
RESULTS AND DISCUSSION
Figure 2 shows the FTIR spectra of PLLA, PBSC, LTI, [AC.sub.20],
and [ACI.sub.20]. It is seen that PLLA, PBSC, and the blends exhibited
very similar spectra patterns with the C=O peak at about 1700
[cm.sup.-1]. However, the C=O peak of PBSC (d) was much higher than that
of PLLA (a). It is, therefore, understood that the higher C=O peak of
[AC.sub.20] (b) than PLLA was a result of PBSC blending. It is clearly
seen that the C=O peak of [ACI.sub.20] (c) became lower than that of
[AC.sub.20], suggesting that the mobility of PLLA and PBSC molecules
reduced due to LTI addition. On the other hand, the spectra of LTI (e)
were obviously characterized by the existence of the large NCO peak
around 2200 [cm.sub.-1]. It is noted that there was no apparent peak of
NCO in the spectra of [ACI.sub.20], indicating that NCO groups acted as
compatibilizer by attributing secondary process between the two polymers
PLLA and PBSC. This secondary process is polar interaction and
[FIGURE 2 OMITTED]
The best way to analyze the FTIR data based on these propositions
would be to compare the spectra of [AC.sub.20] and [ACI.sub.20] (refer
to Fig. 3). Based on FTIR analysis, the peak responding to C = O group
appeared higher peak intensity for the sample without the addition of
LTI. This refers to difference in bonding motion between that of PLLA
and PBSC. However, after addition of LTI, these peaks shift lower, which
indicate an identical bond motion in carbonyl group between the two
copolymers. This can be attributed to the effect of the LTI, which
formed good polar interaction between both PLLA and PBSC. This
interaction induces the mobilities of the respective carbonyl bonds as
unit entity. This is in agreement with the previous works that showed
changes in peak shift and intensity as the results interchange of
interaction because of the secondary process [21, 22].
[FIGURE 3 OMITTED]
Figure 4 shows the DSC curves for heating thermograms of neat PLLA,
PBSC, [AC.sub.20], and [ACI.sub.20]. Points corresponding to the
crystallization temperature ([T.sub.c]), the melting temperature
([T.sub.m]), and the glass transition temperature ([T.sub.g]) are also
shown in Fig. 4. The values of [T.sub.c], [T.sub.m], [T.sub.g], and the
degree of crystallinity of PLLA are summarized in Table 2. In the neat
PLLA thermogram (a), the heat jump at about 64[degrees]C corresponded to
[T.sub.g], followed by the melting peak at about 177[degrees]C. On the
other hand, the thermogram of the neat PBSC reveals that its [T.sub.m]
occurred at about 105[degrees]C and [T.sub.g] was found at
-42.8[degrees]C, although the heat jump was not clearly seen in this
thermogram. According to previous work , the syntheses of PBSC with
different CL/BL ratios show melting and glass transition temperatures
ranging between 39 to 107[degrees]C and -38 to -58[degrees]C,
respectively, to be illustrating PBSC in this research.
[FIGURE 4 OMITTED]
For the thermogram of [AC.sub.20], it is obvious that the [T.sub.m]
peaks of PLLA and PBSC and the heat jump at [T.sub.g] of PLLA still
exist and, furthermore, [T.sub.g] of PBSC was also defined. Two
exothermic peaks existed in the [AC.sub.20] thermogram, the former
appearing at 90[degrees]C was attributed to the typical cold
crystallization of PLLA, and the second appearing just before the
melting point of PLLA could be explained by the possibility of
recrystallization of lower perfection crystals of polylactide into
[alpha] crystal of the higher perfection . For the thermogram of
[ACI.sub.20], it is noted that the isothermic peaks corresponding to
[T.sub.m] and [T.sub.g] of PBSC were totally disappeared, indicating
that the immissibility between PLLA and PBSC was effectively improved
because of LTI addition, so that PBSC molecules could not freely be
The levels of crystallinity, [X.sub.c], of the blends are shown in
Table 2. It is clearly seen that the blends of PLLA/PBSC with and
without LTI possessed the crystallinity values, whereas the neat PLLA
exhibited almost amorphous state corresponding to no apparent
crystallization peak. For the blends, it is obvious that ACI showed
lower crystallinity than AC because of the improvement of immiscibility.
FE-SEM micrographs of the cryo-fractured surfaces are shown in Fig.
5. Evolution of spherulitic morphology appeared on all of PLLA/PBSC
blends. The spherical structures constructed the secondary phase and,
therefore, corresponded to PBSC spherulites. It is apparent that the
spherical structures became large with increasing PBSC content. This
kind of morphological result agrees with immiscible polymer blends that
generally create macrophase separation of the two components due to
difference of solubility parameter . Such phase separation usually
affects the physical and mechanical properties of the blend . On the
contrary, for PLLA/PBSC blends with LTI, apparent circular structures of
PBSC could not be observed on the surfaces. This is due to improvement
of miscibility between PLLA and PBSC. Such microstructural improvement
is understood to be caused by the chemical reaction between NCO and OH
groups as discussed based on the FTIR results given in the previous
[FIGURE 5 OMITTED]
Bending mechanical properties are listed in Table 3. The bending
strength decreased due to blending with PBSC having much lower strength
than PLLA. It is clearly shown that LTI addition improved the strength
of the blends. Effects of blending and LTI addition on the flexural
modulus were very similar to those on the strength. The total fracture
energy values dissipated during bending deformation are also shown in
Table 3. Blending and LTI addition tended to improve the bending energy;
however, the energy of [ACI.sub.30] was lower than that of PLLA and the
other blends, because effect of the ductile secondary PBSC phase on the
bending deformation became significant.
Typical load-displacement curves obtained from the mode I fracture
tests are shown in Fig. 6. In general, the maximum load almost coincides
with the onset of crack growth in the tip of notch introduced. It is,
therefore, understood that the much higher maximum load of [ACI.sub.20]
and [ACI.sub.30] indicated dramatic improvement of resistibility to
crack initiation. The slope of the load-displacement curve after the
maximum load usually corresponds to the rate of crack growth. Therefore,
the much gentler slopes of the blends indicated slower crack growth than
the rapid growth in PLLA characterized by a steep slope.
[FIGURE 6 OMITTED]
Effects of LTI addition on the initial fracture energy, [J.sub.in],
and the averaged fractured energy, [J.sub.f], are shown in Fig. 7. It is
seen that [J.sub.in] of PLLA/PBSC blends were slightly higher than that
of neat PLLA; on the contrary, [J.sub.in] of PLLA/PBSC/LTI increased
with increase of PBSC content, and [J.sub.in] of [ACI.sub.30] was about
three times larger than that of neat PLLA. [J.sub.f] values of AC blends
were higher than that of PLLA, and the improvement of [J.sub.f] due to
blending reached its peak at 20 wt% of PBSC. On the contrary, [J.sub.f]
of ACI effectively increased with increase of PBSC content, and
[J.sub.f] of [ACI.sub.30] was about seven times larger than that of
PLLA. These experimental results clearly exhibited that LTI addition
effectively improved the mode I fracture properties such as [J.sub.in]
and [J.sub.f] of PLLA/PBSC polymer blends.
[FIGURE 7 OMITTED]
Fracture Surface Morphology
FE-SEM micrographs of fracture surfaces in the note-tip regions are
shown in Fig. 8. It is apparent that PLLA created a flat and smooth
surface compared to the rough fracture surfaces of PLLA/PBSC blends. The
blends showed plastic deformation behavior characterized by elongated
fibril structures. It is also seen that the surfaces of PLLA/PBSC,
characterized by the existence of many voids, are much rougher than
those of PLLA/PBSC/LTI. It is naturally understood that these voids were
created by removal of the PBSC spherulites shown in Fig. 5. The size of
the spherulites tended to increase with increasing PBSC content as shown
in Fig. 6 and, therefore, the size of the voids also increased as PBSC
content increased. It is noted that the voids were created due to
debonding at the interfaces between the spherulites and the matrix under
a low stress level, because of the lower interfacial strength than the
strength of the base polymer. The crystallization, i.e., the spherulite
formation process of PBSC takes place at a lower temperature than the
crystallization process of PLLA and, therefore, the PBSC spherulites
cannot firmly connect with the surrounding PLLA matrix at the
interfaces. It is easily understood that the voids cause localized
stress concentration in the surrounding matrix and, therefore,
accelerates fracture initiation in the notch-tip region. Larger
spherulites correspond to wider interfacial regions and also stronger
interactions between the spherulites, and, therefore, faster the
interfacial failure and lower the fracture energy as shown in Fig. 6.
When LTI was added to the mixture of PLLA and PBSC, the isocyanate
groups of LTI tried to create chemical bonding with the hydroxyl groups
of PLLA and PBSC, known as the urethane bonding, and, therefore, the
PLLA molecules were firmly entangled with the PBSC molecules. As a
result, the immiscibility of the polymers and, therefore, the fracture
energy values were dramatically improved.
[FIGURE 8 OMITTED]
Effects of LTI addition on polymer blends of PLLA and PBSC results
significant improvement in mode I fracture and bending properties of the
blends. On the fracture surfaces of PLLA/PBSC blends, many voids were
observed, on the contrary, for PLLA/PBSC/LTI blends, void formation was
suppressed due to increase of entanglement of PLLA and PBSC molecules
caused by chemical reaction. This kind of morphology change due to LTI
addition resulted in the dramatic improvement of the mode I fracture
properties. FE-SEM observation also showed that for PLLA/PBSC blends
with LTI, apparent circular structures of PBSC that were spherulites,
could not be observed on the cryo-fractured surfaces. This corresponded
to the improvement of immiscibility between PLLA and PBSC due to LTI
addition. DSC curves exhibited that the isothermic peaks corresponding
to [T.sub.m] and [T.sub.g] of PBSC were totally disappeared in
PLLA/PBSCLTI blends, indicating that the immiscibility between PLLA and
PBSC was effectively improved due to LTI addition, so that the PBSC
molecules could not freely be moved. FTIR results suggested that the NCO
groups of LTI were acted as compatibilizer by attributing secondary
process between the two polymers PLLA and PBSC. It is thus suggested
that LTI worked as a compatibilizer connecting between PLLA and PBSC
(1.) S.R. Bhattarai, N. Bhattarai, H.K. Yi, P.H. Hwang. D.I. Cha,
and H.Y. Kim, Biomaterials, 25, 2595 (2004).
(2.) O.M. Bostman, J. Bone Joint Surg., 73-A, 148 (1991).
(3.) J.W. Leenslag, A.J. Pennings, R.R.M. Bos. F.R. Rozema, and J.
Boering, Biomaterials, 8, 70 (1987).
(4.) M. Todo, S.D. Park, T. Takayama, and K. Arakawa, Eng. Fract.
Mech., 74, 1872 (2007).
(5.) A.J. Domb, J. Polym. Sci. Part A: Polym. Chem., 31, 1973
(6.) M. Hiljanen-Vainio, P. Varpomaa, J. Seppala, and P. Tormala,
Macromol Chem. Phys., 197, 1503 (1996).
(7.) L. Wang. W. Ma, R.A. Gross, and S.P. McCarthy, Polym. Degrad.
Stab. 59, 161 (1998).
(8.) H. Tsuji, A. Mizuno, and Y. Ikada, J. Appl. Polym. Sci., 70.
(9.) Y. Yuan and E. Ruckenstein, Polym. Bull., 40, 485 (1998).
(10.) R. Dell'Erba, G. Groeninckx, G. Maglio, M. Malinconico,
and A. Migliozzi, Polymer, 42, 7831 (2001).
(11.) X. Cao, A. Mohamed, S.H. Gordon, J.L. Willett, and D.J.
Sessa, Thermochim. Acta, 406, 115 (2003).
(12.) K.S. Anderson, S.H. Lim, and M.A. Hillmyer, J. Appl. Polym.
Sci., 89, 3757 (2003).
(13.) M. Shibata, Y. Inoue, and M. Miyoshi, Polymer, 47, 3557
(14.) M. Shibata, N. Teramoto, and Y. Inoue, Polymer, 48, 2768
(15.) M. Harada, T. Ohya, K. Iida, H. Hayashi, K. Hirano, and H.
Fukuda, J. Appl. Polym. Sci., 106, 1813 (2007).
(16.) M. Harada, K. Iida, K. Okamoto, H. Hayashi, and K. Hirano,
Polym. Eng. Sci., 48, 1359 (2008).
(17.) T. Takayama and M. Todo, J. Mater. Sci., 41, 4989 (2006).
(18.) T. Takayama, M. Todo, H. Tsuji, and K. Arakawa, J. Mater.
Sci., 41, 6501 (2006).
(19.) V. Vannaladsaysy, M. Todo, T. Takayama, M. Jaafar, Z. Ahmad,
and K. Pasomsouk, J. Mater. Sci., 11, 3006 (2009).
(20.) M. Todo, A. Harada, and H. Tsuji, CD-ROM, Proc. 16th Int.
Conf. Compos. Mater., July 8-13. TuGM1-07, Kyoto, Japan (2007).
(21.) J. Dong and Y. Ozaki, Macromolecules, 30, 286 (1997).
(22.) O.R. Pardini and J.I. Amalvy, J. Appl. Polym. Sci., 107, 1207
(23.) A. Cao, T. Okamura, C. Ishiguro, K. Nakayama, Y. Inoue, and
T. Masuda, Polymer. 43, 671 (2002).
(24.) J. Zhang, Y. Duan, H. Sato, H. Tsuji. I. Noda, S. Yan, and Y.
Ozaki, Macromolecules, 38, 8012 (2005).
(25.) H. Tsuji, Biomaterials, 24, 537 (2003).
Vilay Vannaladsaysy,(1),(2) Mitsugu Todo,(1) Mariatti Jaafar, (2)
Zulkifli Ahmad, (2) Korakanh Pasomsouk (3)
(1) Research Institute for Applied Mechanics, Division of
Fundamental Mechanics, Kyushu University, Kasuga, Fukuoka, Japan
(2) School of Materials and Mineral Resources Engineering,
Universiti Sains Malaysia, Nibong Tebal 14300, Pulau Pinang, Malaysia
(3) Department of Mechanical Engineering, Faculty of Engineering,
National University of Laos, Vientiane, Laos
Correspondence to: Mitsugu Todo; e-mail: firstname.lastname@example.org
Published online in Wiley InterScience
[C] 2010 Society of Plastics Engineers
TABLE 1. Sample designation and composition of materials.
Sample code Composition Weight percent (wt%)
PLLA PLLA 100
PBSC PBSC 100
[AC.sub.10] PLLA/PBSC 90/10
[ACI.sub.10] PLLA/PBSL/LTI 90/10/2
[AC.sub.20] PLLA/PBSC 80/20
[ACI.sub.20] PLLA/PBSC/LTI 80/20/2
[AC.sub.30] PLLA/PBSC 70/30
[ACI.sub.30] PLLA/PBSC/LTI 70/30/2
TABLE 2. Thermal properties of neat PLLA, PBSC, and the blends.
Materials [T.sub.g] [T.sub.m] [T.sub.c] ([degrees]C)
PLLA 64.44 177.09 --
PBSC -- -- --
[AC.sub.10] 63.24 174.44 88.66
[ACI.sub.10] 62.95 173.86 95.18
[AC.sub.20] 63.65 177.25 90.78
[ACI.sub.20] 62.88 177.1 97.36
[AC.sub.30] 63.68 176.59 89.69
[ACI.sub.30] 62.08 176.48 97.79
Materials [T.sub.g] [T.sub.m] ([degrees]C) [X.sub.cPLLA] (%)
PLLA -- -- --
PBSC -44.14 105.13 --
[AC.sub.10] -42.9 98.25 68.98
[ACI.sub.10] -40.24 -- 73.67
[AC.sub.20] -40.1 100.47 59.54
[ACI.sub.20] -35.71 -- 40.07
[AC.sub.30] -42.18 99.89 58.99
[ACI.sub.30] -32.49 -- 43.79
TABLE 3. Bending properties of neat PLLA, PBSC, and the blends.
Materials Strength (MPa) Modulus (GPa)
PLLA 99.62 [+ or -] 7.1 3.99 [+ or -] 0.35
PBSC 32.1 [+ or -] 1.1 0.71 [+ or -] 0.18
[AC.sub.10] 91.98 [+ or -] 2.62 3.56 [+ or -] 0.1
[ACI.sub.10] 92.3 [+ or -] 6.6 3.91 [+ or -] 0.26
[AC.sub.20] 90.7 [+ or -] 6.3 3.29 [+ or -] 0.2
[ACI.sub.20] 94.21 [+ or -] 4.7 3.72 [+ or -] 0.12
[AC.sub.30] 73.94 [+ or -] 4.9 2.56 [+ or -] 0.3
[ACI.sub.30] 84.8 [+ or -] 1.7 3.73 [+ or -] 0.17
Materials Energy (kJ/[m.sup.2])
PLLA 21.5 [+ or -] 0.2
PBSC 6.6 [+ or -] 0.7
[AC.sub.10] 22.72 [+ or -] 1.1
[ACI.sub.10] 24.06 [+ or -] 2.2
[AC.sub.20] 21.31 [+ or -] 2.2
[ACI.sub.20] 23.88 [+ or -] 1.23
[AC.sub.30] 12.8 [+ or -] 3.7
[ACI.sub.30] 20.7 [+ or -] 0.8