Polymer Vol. 39 No. 22, pp.

5515–5521, 1998
q 1998 Elsevier Science Ltd.
All rights reserved
PII: S0032-3861(97)10203-8 0032-3861/98/$ - see front matter

Crystallization behaviour of poly(L-lactide)

Tadakazu Miyata a ,* and Toru Masuko b
a
Oji Paper Co., Ltd, Advanced Technology Research Laboratory, 1-10-6, Shinonome,
Koto-ku, Tokyo 135, Japan
b
Faculty of Engineering, Yamagata University, 4-3-16, Jonan, Yonezawa, Yamagata
992, Japan
(Received 18 September 1997; accepted 6 October 1997)

The crystallization characteristics of poly(l-lactide) (PLLA) films of different molecular weights have been
studied by differential scanning calorimetry, optical microscopy and a depolarized-light intensity (d.l.i.) method.
In non-isothermal crystallization modes from the isotropic melt, the crystallinity of PLLA films increased
remarkably with a decrease in the cooling rate, accompanied by a kind of annealing effect. Overall isothermal
crystallization rates of the polymer determined by d.l.i. in the range from 908C to 1408C indicated the maximum
value at 1058C. This temperature is lower than that (1208C) at which the maximum spherullite growth rate is
observed. The PLLA polymer exhibited an Avrami exponent n ¼ 4 in the isothermal d.l.i. experiments over the
whole temperature range examined. Molecular parameters governing the isothermal crystallization behaviour
have been estimated by the application of secondary nucleation theory. q 1998 Elsevier Science Ltd. All rights reserved.

(Keywords: poly(l-lactide); morphology; crystallization kinetics)

INTRODUCTION non-isothermal process is required from a fundamental

viewpoint.
In recent years, much attention has been focused on
In our previous paper18, the morphology and structural
biodegradable and biocompatible polymers, particularly
features of solution-grown crystals of PLLA were examined
from an ecological viewpoint. They are expected to be
by transmission electron microscopy with the diffraction
utilized for surgical implant materials and drug delivery
mode, wide-angle X-ray diffractometry and atomic force
systems. The mechanical properties and degradation rates of
microscopy. In this study, we examine the crystallization
biodegradable polymers, in general, depend on their
morphology or crystallinity as well as on their chemical characteristics of PLLA bulk films under various thermal
conditions with respect to the change in crystallinity and
structures.
spherulite morphology.
Poly(l-lactide) (PLLA) is one biodegradable, synthetic
polymer of great interest. Many publications concerning the
degradation mechanism of PLLA both in vitro and in vivo EXPERIMENTAL
have been published1–11. In parallel with degradative
investigations, several studies on the solid state of PLLA Polymer synthesis
have been carried out by X-ray diffraction (XRD) and Well-purified l,l-lactide (Tokyo Kasei Kogyo Co., Ltd,
conformational analysis of the molecular chain12–15. In Japan) was placed in evacuated tubes containing 0.03% (w/
recent X-ray works, it is proposed that the a-form of the w) stannous octanoate (Sigma Chem. Corp., USA) and
PLLA crystal is composed of 10/3 helical chains13–15. 0.01% (w/w) lauryl alcohol (Tokyo Kasei Kogyo Co., Ltd,
Furthermore, Eling et al.7 were the first to find another Japan) as a catalyst activator and chain control agent4.
crystal modification (the b-form) which contains 3/1 helical Polymerization was carried out in the tube dipped into a
chains of PLLA. silicone oil bath at 2208C for 5 min to 4 h. The solid
Examinations concerned with the crystallization beha- polymer in the tube was dissolved in chloroform and then
viour of PLLA have also been published by several precipitated repeatedly from a large quantity of methanol at
investigators. For example, Vasanthakumari and Pennings16 room temperature. The resultant polymer was white and
reported the growth characteristics of PLLA spherulites as fibrous in appearance. Molecular characterization results of
functions of undercooling and molecular weight, mainly by the PLLA used in this work are listed in Table 1.
means of optical microscopy. Along with that report, Kishor
Film specimens
and Vasanthakumari17 examined the crystallization para-
meters for PLLA by means of a differential scanning The PLLA polymer was sandwiched between two glass
calorimetry (d.s.c.) technique. However, the isothermal plates, heated to above its respective melting point (T m) and
crystallization mechanism for PLLA seems not to have then quenched into ice water. The quenched films were
been examined sufficiently in those publications. The crystallized isothermally at various temperatures or sub-
relationship between structural features and crystalliza- jected to non-isothermal crystallization at different cooling
tion (or annealing) conditions under an isothermal or rates in a nitrogen atmosphere. For optical microscope
observations, solution-cast films 10 mm thick were cast from
a 1 wt% chloroform solution onto a glass plate in an air oven
* To whom correspondence should be addressed at 268C, and then dried in a vacuum oven at 458C.

POLYMER Volume 39 Number 22 1998 5515

Crystallization of PLLA: T. Miyata and T. Masuko

Table 1 Characteristics of the PLLA used in this work

Infra-red absorption CyO ester 1760
(cm ¹1)
–O– ester 1190
1
H nuclear magnetic a 5.16
resonance (ppm)
b 1.58
Gel permeation A M w ¼ 2.0 3 10 5
chromatography
B M w ¼ 1.0 3 10 5
C M w ¼ 5.0 3 10 4

D.s.c. measurements
To examine the thermal properties of the PLLA films, a
SEIKO DSC-200 calorimeter was utilized to obtain
endothermic and exothermic changes in the range of 0–
2008C at various heating (or cooling) rates in a nitrogen purge.
Optical microscopy
Spherulitic textures in the thin cast films (10 mm) were
examined with an Olympus BH-2 polarizing microscope
equipped with a video camera system. A Linkam THM-600
hot-stage device was also used in the range from 258C to the
temperature above the respective T m of the PLLA samples.
Density measurements
Figure 1 D.s.c. traces of the PLLA film in heating/cooling cycles
The densities of the polymer films were determined at
258C by using a density-gradient column filled with an
aqueous CaBr 2 solution.
XRD measurements
Graphite-monochromatized Cu K a radiation was
employed for XRD experiments on a MAC Science
M18XHF-SRA generator operated at 50 kV and 350 mA.
XRD photographs and integral diffraction intensities for the
specimens were taken at 238C by means of an imaging plate
system (DIP-220). Spacings of the reflection planes were
measured, and correlated with those of standard cholesterol
powder.
Depolarized-light intensity measurement
Isothermal crystallization measurements were carried out
by use of a Kotaki MK-801 depolarized-light intensity
(d.l.i.) apparatus19 in the region from 908C to 1408C. The Figure 2 The logarithm of the exothermic enthalpy change as a function
electronic signals transformed from the measured d.l.i. were of the cooling rate from the isotropic melt
amplified and then recorded on a personal computer which
accommodated an Avrami analysis software. This implies that film specimens cooled from the isotropic
melt a rate greater than 108C min ¹1 remain amorphous.
RESULTS AND DISCUSSION Thus, we examined first the crystallization behaviour of
PLLA on the cooling process from the isotropic melt.
Non-isothermal crystallization behaviour of the PLLA films The quenched film was kept at 2008C for 10 min in the
Figure 1 illustrates the heating/cooling d.s.c. traces at a rate d.s.c. apparatus, and then cooled to room temperature at
of 108C min ¹1 for the melt-cast PLLA film of specimen A (M w various cooling rates. Figure 2 shows the logarithmic
¼ 2.0 3 10 5) mainly employed in this section. In the first exothermic enthalpy change observed in the cooling process
heating run (curve 1) can be observed the glass transition (T g) as a function of the cooling rate. The crystallinity of the
at ,558C, a single exothermic peak at 1158C, and endothermic PLLA film increases with decreasing cooling rate. In
peaks (melting point, T m) at 1718C. Upon cooling from 2008C, Figure 2, a fairly good linear relationship between both
only a baseline change with T g appears at 508C (curve 2). quantities is recognized. The exothermic enthalpy change
When the specimen is subsequently heated, a similar obtained by extrapolating the cooling rate to 08C min ¹1 in
thermogram to that in the first heating run is observed (curve Figure 2 is calculated to be almost 60 J g ¹1. When these
3). After two heating/cooling cycles, no change in the d.s.c. specimens were reheated at 108C min ¹1 to 2008C, the d.s.c.
traces is observed at the same heating/cooling rate. traces obtained indicate that the baseline change at T g,
Interestingly, no exothermic peak can be observed in the DC p(T g), decreases with reducing cooling rate. The DC p(T g)
first cooling run at 108C min ¹1 (curve 2), and the change finally disappears for the specimen cooled at
exothermic enthalpy change at 1158C is close to the heat 0.58C min ¹1. Also, the exothermic enthalpy change clearly
of fusion at 1718C (T m) on the second heating run (curve 3). increases with increasing cooling rates in the reheating process.

5516 POLYMER Volume 39 Number 22 1998

 Crystallization of PLLA: T. Miyata and T. Masuko

10 5) at 1408C, similar to that obtained by Pennings and co-

workers6,16 and Tsuji et al.22. The G values for the
specimens of different molecular weight are shown in
Figure 7. Below 1008C, it was difficult to measure the
growth rate because of low segmental mobilities of the
PLLA.
Figure 7 shows that the value of G increases with
increasing crystallization temperature, T c; after reaching the
maximum, it then decreases with increasing T c. The
crystallization rate increases with decreasing molecular
weight of PLLA. The maximum in the rate of spherulite
growth is located at about 1208C on the temperature axis.
These types of behaviour are very common among ordinary
synthetic polymers23, and are also essentially consistent
with those reported by Pennings and co-workers16.
To characterize the mode of crystallization, the following
Avrami equation
ÿ

Xc ¼ 1 ¹ exp ¹ ktn (1)
Figure 3 Changes in film density as a function of cooling rate from the
isotropic melt was applied to the d.l.i. data obtained, where k is the tem-
perature-dependent rate constant and n an integer between 1
and 4 indicating the mode of dimension of crystal growth.
Figure 3 shows that the density of the annealed speci- X c(t) represents the fraction of transformed or crystallized
mens increases with a decrease in the cooling rate. This material after time t. In the depolarized-light intensity24,25
tendency is similar to the relationship between the change in technique, it can be expressed as
cooling rates and exothermic enthalpy changes in the I` ¹ It
cooling process. These results indicate that the crystallinity Xc ¼ (2)
in the specimens depends on their annealing conditions, I` ¹ I0
being especially influenced by slow cooling rates. where I `, I 0 and I t represent the transmitted-light intensity at
Figure 4 presents optical micrographs of PLLA spher- very long time, the intensity at initial time and the intensity
ulites grown at various cooling rates from 2008C ( . T m) to at an intermediate time, t, respectively.
room temperature. Obviously, the size of the PLLA Avrami theory is useful for determining the mode of
spherulites increases with a decrease in the cooling rate; crystal growth (spherical, disc-like, etc.) and the overall rate
particularly at slower cooling rates ( , 28C min ¹1) the of nucleation-controlled crystallization in polymeric mate-
PLLA film crystallizes effectively from the isotropic melt rials. The values of overall crystallization rate, k, from the
with an accompanying annealing treatment. Whenever the reciprocal induction period (1/t ¼ k t) are plotted against T c
PLLA films were cooled at rapid rates ( . 208C min ¹1) the in Figure 8a. Also, the other k values (k t1/2) shown in
resultant films became completely amorphous, as confirmed Figure 8b are obtained by use of a fractional conversion of
by the X-ray photograph shown in Figure 5. The density of 0.5. Both k values from the different methods increase with
these films is almost constant and determined experimen- increasing T c; then, after reaching the maximum, decrease
tally as 1.256 g cm ¹3, presumably equal to the ‘amorphous with further increase in T c. These tendencies agree well with
density’ of PLLA. that of the spherulite growth rate in Figure 7. The maximum
As shown in Figure 6, a linear relationship seems to hold in the rate of overall crystallization occurs at a temperature
between exothermic enthalpy change and density for PLLA of about 1058C, lower than that (1208C) corresponding to
films. The density change is generally governed by crystal- the maximum spherulite growth rate.
linity of the polymer. Since the crystal density of PLLA (a- Avrami plots of these data for the sample with M w ¼ 2 3
form) was estimated13,18 to be 1.285 g cm ¹3, the enthalpy of 10 5 are shown in Figure 9. The average value of n was
fusion for a large crystal of infinite size corresponds to estimated to be 4 by least-squares fits. Although not shown
135 J g ¹1 as deduced from this relationship. This estimated in the figures, similar values for the Avrami indices were
value is far larger than the 93 J g ¹1 calculated by Fischer et obtained for all PLLA samples of different molecular
al.20, but more close to value (140 J g ¹1) of Loomis et al.21. weight. Thus, the spherulite growth mode of PLLA is three-
As described in the preceding section, however, the dimensional with a homogeneous nucleation mode. As
exothermic enthalpy change obtained by extrapolating the shown in Figure 9, the n value gradually becomes 3 at
cooling rate to 08C min ¹1 (see Figure 2) is about 60 J g ¹1. In longer time scales, since larger spherulites sandwiched
this case, the maximum crystallinity of the annealed PLLA between two glass plates have grown into a disc-like shape.
film is calculated to be almost 0.44. The enthalpy of fusion
for the film crystallized at 1408C for 24 h was determined to Growth rate parameters in crystallization of PLLA
be 70 J g ¹1. Therefore, it might be difficult to achieve a The kinetic data obtained from Figure 7 have been
degree of crystallinity for PLLA films exceeding the value examined in terms of secondary nucleation theory23. The
of about 0.5. general expression for the growth rate of a linear polymer
crystal with folded chains is given by
Isothermal crystallization behaviour of PLLA !  
The spherulite radius of PLLA films measured from ¹ Up ¹ Kg
G ¼ G0 exp ÿ
exp (3)
optical micrographs is confirmed to change linearly with R Tc ¹ T` Tc DTf
crystallization time. The spherulite growth rate, G, thus
evaluated was 2.0 mm min ¹1 for specimen A (M w ¼ 2.0 3 where K g is the nucleation constant; DT is the undercooling

POLYMER Volume 39 Number 22 1998 5517

Crystallization of PLLA: T. Miyata and T. Masuko

Figure 4 Spherulites of PLLA grown at various cooling rates from the isotropic melt

defined by Tm8 ¹ Tc ; Tm8 is the equilibrium melting point; f is from their slopes. Figure 11 shows plots of log G þ U*/
a factor given as 2Tc =(Tm8 þ Tc ); U* is the activation energy 2.303R(T ¹ T `) versus 1/T cDTf, which give the values of K
for segment diffusion to the site of crystallization; R is the g and log G 0 from their slopes and intercepts, respectively.
gas constant; T ` is the hypothetical temperature where all The results obtained from the secondary nucleation analysis
motion associated with viscous flow ceases, T g ¹ 50; and G 0 are listed in Table 2. There is still insufficient direct
is the front factor. We apply the equation to our data of information on the growth front of PLLA spherulites, since
PLLA shown in Figure 7. the lamellae structure of PLLA has not yet been examined
As illustrated in Figure 10, log G changes linearly with 1/ adequately. When the growth front is assumed to be the
(T c ¹ T `). These linear relationships give the U* values (1 1 0) plane of PLLA as in the case of high-density

5518 POLYMER Volume 39 Number 22 1998

 Crystallization of PLLA: T. Miyata and T. Masuko

Figure 5 X-ray photograph of the glassy PLLA film obtained with a rapid
cooling rate of 208C min ¹1 from the isotropic melt

Figure 8 Isothermal overall crystallization rate as a function of crystal-

lization temperature (T c) for the PLLA films. (a) k t is evaluated from
the reciprocal induction period; (b) indicates k t1/2 obtained at a
fractional conversion of 0.5. A, M w ¼ 2.0 3 10 5; W, M w ¼ 1.0 3 10 5;
K, M w ¼ 5.0 3 10 4

Figure 6 Relationship between exothermic enthalpy change and density

for PLLA film specimens

Figure 9 Avrami plots for isothermal crystallization of PLLA films of M w

¼ 2.0 3 10 5 at various T cs from the isotropic melt at 2008C kept for 5 min.
The lines indicate the gradients of 3 and 4

polyethylene, the value of the layer thickness should be

taken as 0.53 nm for PLLA crystals18. Accordingly, the
values of K g and surface free energies, jj e, can be
calculated by considering Regime II growth23.
The nucleation parameter K g is given by:
2bjje Tm8
Kg ¼ (4)
Dhf kB
where b is the layer thickness; j is the lateral surface free
energy; j e is the fold surface free energy; Dh f the heat of
fusion per unit volume; and k B Boltzmann’s constant.
The values of jj e estimated are in the range 560–990 3
10 ¹6 J 2 m ¹4, similar to those found for other polymers23.
Figure 7 Growth rate G (mm min ¹1) as a function of isothermal
However, these values are slightly larger than Vasanthaku-
crystallization temperature (T c) for PLLA films: A, M w ¼ 2.0 3 10 5; W, mari and Pennings’ results16 for PLLA. It is not clear at
M w ¼ 1.0 3 10 5; K, M w ¼ 5.0 3 10 4 present whether or not this difference is caused by the

POLYMER Volume 39 Number 22 1998 5519

Crystallization of PLLA: T. Miyata and T. Masuko

The j value was then calculated as 9.2 3 10 ¹3 J m ¹2 for all

PLLA samples with different molecular weight. When this
value of j was used to obtain j e from their jj e values
estimated above, the j e values of specimen A, specimen B
and specimen C were 107 3 10 ¹3 J m ¹2, 100 3 10 ¹3 J m ¹2
and 63 3 10 ¹3 J m ¹2, respectively. Clearly there seems to
be a molecular weight dependence; namely, the value of j e
increases with increasing molecular weight. This implies
that the fold surface of a high-molecular-weight specimen
would be predominantly composed of loose-loop folds.
Vasanthakumari and Pennings have reported that Regime
I could be applicable to the crystallization data of PLLA
since axialite-like morphology is observed16. However, we
can hardly distinguish the axialite reported by Pennings and
co-workers. More detailed examination of the surface
morphology of the negatively birefringent PLLA spherulites
by electron microscopy and diffraction techniques should be
informative concerning the mechanism and mode of
Figure 10 Changes in ln G as a function of 1/(T c ¹ T `). A, M w ¼ 2.0 3 spherulite growth.
10 5; W, M w ¼ 1.0 3 10 5; K, M w ¼ 5.0 3 10 4

CONCLUSIONS
The crystallization characteristics of PLLA have been
studied by optical microscopy and a depolarized-light
intensity method. The results obtained are as follows.
(1) The crystallinity of PLLA depends closely on the cool-
ing rate from its isotropic melt during non-isothermal
crystallization.
(2) The overall crystallization rate increases with decreas-
ing molecular weight of PLLA. This behaviour is very
similar to that of ordinary synthetic polymers.
(3) The PLLA polymer shows an Avrami exponent n ¼ 4 in
isothermal crystallization experiments.
(4) Molecular parameters of the crystallization behaviour
have been estimated by the application of secondary
nucleation theory.

ACKNOWLEDGEMENTS
Figure 11 Changes in ln G þ U*/RðT ¹ T` ) as a function of 1/T cDTf. A, The authors appreciate the contributions of Mr Kazuo
M w ¼ 2.0 3 10 5; W, M w ¼ 1.0 3 10 5; K, M w ¼ 5.0 3 10 4
Sasaki of Yamagata University, who carried out some parts
of the XRD work and materials arrangements, and Mr Taku
Yoshizawa and Miss Rumiko Shimada for their help with
Table 2 Results of the secondary nucleation analysis the d.l.i. experiments.
Mw T g (8C) U* (kJ mol ¹1) jj e (J 2 m ¹4)
2.0 3 10 5
56 5730 991 3 10 ¹6
1.0 3 10 5 54 4740 925 3 10 ¹6 REFERENCES
5.0 3 10 4 52 3427 575 3 10 ¹6
1. Kulkarni, R. K., Moore, E. G., Hegyeli, A. F. and Lenard, F., J.
Parameters used in this analysis Biomed. Mater. Res., 1971, 5, 169.
2. Brady, J. M., Cutright, D. E., Miller, R. A., Bottistone, G. C. and
Quantity Value Source Hunsuck, E. E., J. Biomed. Mater. Res., 1973, 7, 155.
3. Miller, R. A., Brady, J. M. and Cutright, D. E., J. Biomed. Mater.
Dh f (J m ¹3) 174 3 10 6 this work
Res., 1977, 11, 711.
Tm8 (8C) 203 16
4. Gilding, D. K. and Reed, A. M., Polymer, 1979, 20, 1459.
d 110 (m) 5.3 3 10 ¹10 18
5. Reed, A. M. and Gilding, D. K., Polymer, 1981, 22, 494.
6. Kalb, B. and Pennings, A. J., Polymer, 1980, 21, 607.
7. Eling, B., Gogolewski, S. and Pennings, A. J., Polymer, 1982, 23,
1587.
theoretical equation applied. The values of U* are estimated 8. Zwiers, R. J. M., Gogolewski, S. and Pennings, A. J., Polymer,
to range from 3500 to 5700 J mol ¹1, slightly smaller than 1983, 24, 167.
those reported for other polymers. Also, these values have a 9. Eenink, M. J. D., Feijin, J., Olijslager, J. M., Albers, J. H., Riek, J. C.
molecular weight dependence. and Geidanus, P. J., J. Controlled Release, 1987, 6, 225.
10. Kimura, Y., Matsuzaki, Y., Yamane, H. and Kitao, T., Polymer,
We then estimated the lateral surface energy, j, by using 1989, 30, 1342.
the Thomas–Stavely equation23 11. Reeve, M. S., McCarthy, S. P., Downey, M. J. and Gross, R. A.,
Macromolecules, 1994, 27, 825.
j ¼ 0:1·Dhf ·b (5) 12. De Santis, P. and Kovacs, A. J., Biopolymers, 1968, 6, 299.

5520 POLYMER Volume 39 Number 22 1998

 Crystallization of PLLA: T. Miyata and T. Masuko

13. Hoogsteen, W., Postema, A. R., Pennings, A. J., tenBrinke, G. and 20. Fischer, E. W., Sterzel, H. J. and Wegner, G., Kolloid-Z. u.Z. Poly-
Zugenmaier, P., Macromolecules, 1990, 23, 634. mere, 1973, 251, 980.
14. Kobayashi, J., Asahi, T., Ichiki, M., Okikawa, A., Suzuki, H., Wata- 21. Loomis, G. L., Mudoch, J. R. and Gardner, K. H., Polym. Prepr.
nabe, T., Fukada, E. and Shikinami, Y., J. Appl. Phys., 1995, 77, (Am. Chem. Soc., Div. Polym. Chem.), 1990, 31, 55.
2957. 22. Tsuji, H. and Ikada, Y., Polymer, 1995, 36, 2709.
15. Brizzolara, D., Cantow, H. J., Diederichs, K., Keller, E. and Domb, 23. Hoffman, J. D., Davis, G. T. and Lauritzen, J. I., Treatise on
A. J., Macromolecules, 1996, 29, 191. Solid State Chemistry, vol. 3, ed. N. B. Hannay. Plenum, New
16. Vasanthakumari, R. and Pennings, A. J., Polymer, 1983, 24, 175. York, 1976.
17. Kishore, K. and Vasanthakumari, R., Colloid Polym. Sci., 1988, 24. Magill, J. H., Nature, 1960, 187, 771.
266, 999. 25. Magill, J. H., Nature, 1961, 191, 1092.
18. Miyata, T. and Masuko, T., Polymer, 1997, 38, 4003.
19. Masuko, T., Okuizumi, R., Yonetake, K. and Magill, J. H., Macro-
molecules, 1989, 22, 4636.

POLYMER Volume 39 Number 22 1998 5521