Polymer 45 (2004) 8809–8823

www.elsevier.com/locate/polymer

The influence of block copolymer microstructure on the toughness of

compatibilized polylactide/polyethylene blends
Kelly S. Anderson, Marc A. Hillmyer*
Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, MN 55455-0431, USA
Received 20 July 2004; received in revised form 4 October 2004; accepted 7 October 2004

Abstract
Poly(L-lactide) (PLLA) was melt blended with a set of polyethylenes (PE) in an effort to increase the impact strength of the PLLA. As
compatibilizers, we prepared a series of molecularly distinct polylactide–polyethylene block copolymers. The influence of the copolymer
structure on the matrix/dispersed phase interfacial adhesion was correlated with the mechanical properties of the PLLA composites. For the
lowest modulus PE dispersed phase, the block copolymer that gave the strongest interfacial adhesion was necessary to achieve the most
significant improvement in toughening. Whereas with the stiffest PE, the block copolymer that gave the weakest interfacial adhesion resulted
in the greatest improvement in impact strength. For the intermediate stiffness PE, an intermediate degree of adhesion was necessary to obtain
the largest increase in the impact properties. The impact properties of the composites were also found to be highly dependent on the dispersed
phase properties.
q 2004 Elsevier Ltd. All rights reserved.

Keywords: Polylactide; Toughening; Interfacial adhesion

1. Introduction two homopolymers we found that toughened binary blends

could be obtained; however, upon addition of the PLLA-PE
The toughening of polylactide, an environmentally block copolymer a significant reduction in impact strength
friendly thermoplastic [1], has been investigated using a variability and an increase in the overall toughness was
wide variety of toughening agents, mostly emphasizing observed [9]. While the addition of a compatibilizer was
biocompatible materials. Successful methodologies include useful for preparing toughened PLLA composites, we aim to
copolymerization strategies [2–4], plasticization with a better understand the role of the compatibilizer. Specifi-
miscible component [5–8] and blending with an immiscible cally, we are interested in the influence of crystallinity in the
homopolymer [9–19] or a block copolymer [13,14,20–22]. block copolymer both on the interfacial adhesion between
However, in the case of immiscible poly(L-lactide) (PLLA) the matrix and the dispersed phase, and the distribution of
blends, unifying principles that govern toughening are the dispersed phase particles. To uncover the role of these
lacking [23]. parameters on the composite toughness we will use three
We previously reported the toughening of PLLA by polyethylene dispersed phases with differing levels of
blending with linear low-density polyethylene (LLDPE) [9, crystallinity and thus modulus.
10], concentrating on the compatibilization with a poly(L- Fig. 1 gives a schematic view of the variables that
lactide)-polyethylene (PLLA-PE) block copolymer. With a will be explored in this study. Both the PLLA and
solution blending technique, toughening was only achieved polyethylenes are commercially available. As compatibili-
after the PLLA-PE block copolymer was introduced as a zers, we prepared four polylactide–polyethylene block
compatibilizer [10]. On the contrary, by melt blending the copolymers that are essentially identical except for their
abilities to co-crystallize with the homopolymers (altered
* Corresponding author. Tel.: C1 612 625 7834; fax: C1 612 624 7801. through variations in tacticity or defect site density). Of the
E-mail address: [email protected] (M.A. Hillmyer). four block copolymers used, one material contains two
0032-3861/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2004.10.047
8810 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

necessary to promote rubber particle cavitation [34].

However, in all of the above instances the dispersed phase
properties were not investigated as to potential reasons for
the observed effects. While the influence of the dispersed
phase properties on the resultant toughening has been
explored [35–37], in these cases the role of interfacial
adhesion was essentially ignored. By examining the
combined effects of interfacial adhesion and dispersed
phase properties on the toughening of PLLA, we aim to
uncover generic requirements for the toughening of PLLA
with other semi-crystalline polymers. We hope the results of
this study can be applied more generally to compatibilized
Fig. 1. Overview of parameters explored in this work.
semi-crystalline/semi-crystalline blends [38].
semi-crystalline blocks, one is completely amorphous and
the other two contain one semi-crystalline component and
one amorphous component. In principle, all blocks will be 2. Experimental
able to mix and entangle with the homopolymers, but the
interlocking of the chains into crystallites will only be 2.1. Materials
possible for a subset of the block copolymers. Therefore, the
varying abilities of the block copolymers to co-crystallize Commercial grade PLLA (98.7 mol% L-isomer) was
with the homopolymers can be correlated with the adhesion/ supplied by Cargill-Dow Polymer, LLC (Natureworkse).
mechanical properties [24]. While some improvement in the The two commercial grade LLDPEs were obtained from
Dow Chemical Company (Engage EG 8100, ethylene/oc-
adhesion due to entanglements is likely [25], a significant
tene copolymer containing 13.2 mol% octene, MFIZ
improvement is expected when the blocks are able to co-
1.0 g/10 min, rZ0.870 g/cc and Engage EG 8540, ethyle-
crystallize [26,27].
ne/octene copolymer containing 4.8 mol% octene, MFIZ
In addition to the adhesion between the dispersed and
1.0 g/10 min, rZ0.908 g/cc). These homopolymers will be
matrix phases, the particle size and distribution of the
referred to as LLDPE1 and LLDPE2. A commercial grade
dispersed phase can significantly affect the resultant
HDPE was also obtained from Dow Chemical Company
toughening. As first proposed by Wu [28,29], there is a
(Dowlex IP40, MFIZ40 g/10 min, rZ0.952 g/cc) and will
critical distance between neighboring dispersed phase
be referred to as HDPE. The PLLA-PE and PLA-PE block
particles known as the critical matrix ligament thickness.
copolymers were prepared using the procedure previously
Wu suggests that in order to observe toughening one must described [9]. Briefly, 1,3-butadiene was anionically poly-
be able to attain dispersion of the rubber such that the matrix merized in cyclohexane using sec-butyl-lithium as the
ligament thickness is below the critical value. This critical initiator and subsequently end-capped with ethylene oxide
distance is believed to be related to the overlap of the stress to give hydroxyl-terminated 1,4-polybutadiene containing
fields which surround the dispersed phase particles. Once ca. 93% of the 1,4 regioisomer. This polybutadiene was then
you are below the critical matrix ligament thickness and the hydrogenated to give hydroxyl-terminated polyethylene,
stress fields are able to interact, the matrix phase will which was utilized in combination with AlEt3 as a
undergo enhanced shear yielding resulting in tough macroinitiator in the ring-opening polymerization of L-
behavior. By preparing polylactide–polyethylene block lactide (PLLA) or D,L-lactide (PLA). Two block copolymers
copolymers with varying crystallinities we hope to modu- were prepared, both with a PE block of ca. 30 kg/mol. One
late the interfacial adhesion while maintaining an adequate block copolymer had a PLLA block of ca. 30 kg/mol
dispersion, and thus decouple matrix ligament thickness (PLLA-PE) and the other had a PLA block of ca. 30 kg/mol
effects from interfacial adhesion effects. (PLA-PE), as determined by 1H-NMR spectroscopy. In
While the role of interfacial adhesion on toughening has addition, PLLA-PEP and PLA-PEP block copolymers were
been thoroughly explained, there are system to system synthesized using the same general procedure as described
variations that have not been reconciled. For example, Wu above with the exception that isoprene was anionically
suggested that only van der Waals adhesion is necessary to polymerized to give hydroxyl-terminated 1,4 polyisoprene
achieve toughening since, according to his analysis, the containing ca. 94% of the 1,4 regioisomer. The polyisoprene
main parameter controlling the toughening is the matrix was then hydrogenated to give hydroxyl-terminated poly
ligament thickness [28]. In contrast, others have found that (ethylene-alt-propylene) (PEP), which was used as a
increased interfacial adhesion was either detrimental to the macroinitiator in the ring-opening polymerization of lactide
impact properties [30,31] or had no significant effect on the as described above. Again, two block copolymers were
toughening [32,33], while still others report that good synthesized, both with a PEP block of ca. 30 kg/mol. One
adhesion between the matrix and the rubbery particles are block copolymer had a PLLA block of ca. 30 kg/mol
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8811

(PLLA-PEP) and the other had a PLA block of ca. 30 kg/

mol (PLA-PEP). The detailed molecular characteristics of

TT (MJ/m3)
the homopolymers and block copolymers are given in Table

1.5G0.1

19.4G6.8
96G32
56G3
1, and the block copolymer structures are shown in Fig. 2.

2.2. Equipment/general procedures

The molecular weights of the PLLA homopolymer,

62.1G1.0
10.8G0.4
21.7G2.7
10.4G3.7
sb (MPa)
PLLA-PEP block copolymer and PLA-PEP block copoly-
mer were determined on a Hewlett–Packard 1100 series
liquid chromatograph equipped with a Hewlett–Packard
1047A refractive index detector and three Jordi poly(divi-
nylbenzene) columns with 104, 103, and 500 Å pore sizes.
The mobile phase was tetrahydrofuran (40 8C and

740G140
960G40

150G40
4G1
1 mL/min). The molecular weights of the two LLDPEs

3b (%)
and the HDPE were determined on a Waters 150-C at
135 8C in 1,2,4-trichlorobenzene with a flow rate 1 mL/min.
The molecular weights of the PLLA-PE and the PLA-PE

Modulus (MPa)
block copolymers were determined on a Polymer Labora-

8.0G0.5
2400G60

440G60
tories PL-GPC 220 at 140 8C in 1,2,4-trichlorobenzene with

120G5
a flow rate of 1 mL/min. In all cases, the columns were
calibrated using polystyrene standards from Polymer
Laboratories.
Images of the cryo-fractured tensile bars, impact bar
fracture surfaces and the adhesion sample surfaces were
obtained using a JEOL JSM-6500F scanning electron

Ethylene/octene copolymer (Engage 8100) with 13.2 mol% octene, supplied by Dow Chemical Company.
PDIb

1.67
1.94
2.59
2.72
1.20
1.15
1.13
1.11

Ethylene/octene copolymer (Engage 8540) with 4.8 mol% octene, supplied by Dow Chemical Company.
microscope with an accelerating voltage of 5 kV. The
sample surfaces were coated with 50 Å of platinum.
Particle size analysis was performed on the LLDPE and
HDPE particles. The analyzed samples were obtained by
Mnb (kg/mol)

dissolving the PLLA matrix of a tested impact bar using

The Tg of the PLLA could not be determined due to the broad melting peak of the PE.
chloroform. This gave a LLDPE or HDPE particle
85.6
60.5
49.4
25.4
82.8
92.0
70.8
70.5

suspension. The analysis was done using a Coulter LS 230

High density polyethylene (Dowlex IP40), supplied by Dow Chemical Company.

Nature Workse, 98.7 mol% L-isomer, supplied by Cargill-Dow Polymer, LLC.

The Tg of the PE could not be determined accurately due to the small signal.
Tma (8C)

170/102
167

105
128

105
173
60
Characteristics of homopolymers and block copolymers

SEC versus polystyrene standards (see Section 2).

59/K61
51/K61
Tga (8C)

K58

50/f
58

f,h
f
f
Crystallinitya

48/33
0/29
49/0

DSC (see Section 2).

(%)

0/0
11

36
70
6

PLLA-PEP
PLLA-PE
LLDPE1d
LLDPE2e

PLA-PEP
PLA-PE
Table 1

Sample

HDPEg
PLLAc

Fig. 2. Block copolymers used to investigate the influence of

b

g
h
a

e
f

interfacial adhesion on the toughening of PLLA with polyethylene.

8812 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

particle analyzer (Coulter Beckman Co.). To prevent mixing chamber, the polymers were manually pre-mixed in
particle coalescence, a sonicator was used during the the desired compositions. All blends were prepared with
experiment. For each sample, the average particle diameter, 80 wt% of PLLA and 20 wt% of polyethylene (either
d, and the particle size distribution parameter, s, were LLDPE1, LLDPE2 or HDPE) [43]. When the block
determined from the following equations [39]: copolymer was used, 5 wt% of the block copolymer was
PN added based on the total weight of the 80/20 blend. Once the
iZ1 ni ln di material was pre-mixed it was added to the chamber of the
ln d Z P N (1)
iZ1 ni mixer, which was heated to 190 8C and had the blades
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rotating at 30 rpm. The plunger was then lowered to enclose
PN 2
iZ1 ni ðln di K ln dÞ the chamber and the rotation was increased to 50 rpm. After
ln s Z PN (2) 15 min, the blades were stopped and the blends were
iZ1 ni
removed from the chamber using a spatula and cut into
Where ni is the number of particles having the diameter di. pieces using scissors. The blend was allowed to cool to
The particle size distribution parameter is a measure of the room temperature on the bench top.
polydispersity of the particle size with sZ1 being
monodisperse. The particle size and particle size distri-
bution parameter were then used to calculate the matrix 2.4. Mechanical testing
ligament thickness, given by [39]:
 1=3  Mechanical testing samples were prepared by com-
p pression molding the melt blends and the PLLA
T Zd expð1:5 ln2 sÞ K expð0:5 ln2 sÞ (3)
6f
In Eq. (3), f is the volume fraction of rubber, d is the
average particle diameter and s is the particle size
distribution parameter, given in Eqs. (1) and (2),
respectively.
Thermal analysis was performed on all materials used in
this study. A 5–10 mg sample was heated at 5 8C/min on a
PerkinElmer DSC-7. The samples were allowed to anneal
for 5 min at their highest test temperature to remove any
thermal history and were cooled to 5 8C at 2 8C/min. The
samples were then held for 5 min at 5 8C before heating
again for the thermal analysis. For the PLLA, HDPE and the
PLLA-PE, PLA-PE, PLLA-PEP and PLA-PEP block
copolymers, the samples were heated from 5 to 200 8C.
The two LLDPE samples were heated from 5 to 150 8C. The
impact bar samples were heated from 0 to 200 8C at
10 8C/min. One sample was tested for each data point
reported. The percent crystallinity was calculated using the
following equation:
DHm
% crystallinity Z 100 ! (4)
f DHmideal
Where DHm is the measured heat of fusion, f is the weight
fraction of the component in question and DHmideal is the
enthalpy of fusion for a crystal having infinite crystal
thickness (94 J/g for PLLA [40,41], 277 J/g for PE [42]). To
determine the Tg of the LLDPEs and the PEP in the PLA-
PEP and PLLA-PEP block copolymers, DSC analysis for
these samples were also performed on the PerkinElmer
Pyris-1 instrument.

2.3. Blend preparation Fig. 3. SEM images of 80/20 wt% PLLA/LLDPE2 tensile bars
after fracture. Images were taken directly below fracture surface.
The melt blending was performed in a Haake Rheomix (a) Without block copolymer and (b) with 5 wt% PLLA-PE block
600 batch mixer at 190 8C. Before being placed in the copolymer. The stress direction is vertical.
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8813

homopolymer at 200 8C for 10 min. The polyethylene when testing was performed on the PLLA/LLDPE1 sample
homopolymers were compression molded at 150 8C for with the PLLA-PEP block copolymer. The peak in the
10 min. All samples were prepared with an ultimate adhesion indicates that the optimum layer thickness is
pressure of 0.20 MPa. Once molded, the samples were attained with this concentration [24]. For the PLA-PE and
cooled to 100 8C with circulating water and held for 5 min. PLLA-PE block copolymers, films were spin coated onto
They were then immediately cooled to ca. 30 8C and the PLLA plaque at 3000 rpm from a 7 mg/mL solution in
removed from the mold. For the polyethylene homopoly- chlorobenzene. These conditions gave the thinnest layer
mers, the 100 8C equilibration step was not performed. (w26 nm) while still being able to achieve uniform
Impact test bars of dimensions 12.7!63.5!3.2 mm3 thickness as determined from profilometer measurements.
with notch depth of 2.54 mm were prepared for all melt The PLLA-PE and PLA-PE block copolymer layer
blends and the PLLA homopolymer. The notched Izod thicknesses were not as thin as that determined to be the
impact testing was performed on a CEAST USA, Resil 25 optimum thickness with the PLLA-PEP block copolymer
impact tester at room temperature according to ASTM (w12 nm). Therefore, the values reported for the PLLA-PE
D256. The impact resistance was calculated by dividing the and PLA-PE block copolymers are actually lower than the
total energy required to break the sample by the thickness of theoretically achievable adhesion [24].
the impact bar. A minimum of six samples were tested for In all cases the block copolymer solutions were heated to
each material. In our previous study [9], the impact energies 90 8C prior to spin coating in order to dissolve the block
were determined based on the area under a force-displace- copolymer and obtain a homogenous solution. The plaques
ment curve; however, in this study the energies were with the spin coated block copolymer layer were kept at
determined based on the hammer height after impact, due to ambient conditions for 48 h and placed in a 100 8C oven for
the load cell being unavailable during testing. Therefore, the 1 h prior to bilayer sample preparation in order to allow for
values reported for the PLLA/LLDPE1 blends are slightly residual solvent evaporation. The bilayer adhesion samples
higher than previously reported. were then prepared by placing the PLLA and PE plaques
Tensile bars were also prepared with dimensions as into contact with an ultimate pressure of 0.20 MPa at 200 8C
specified in ASTM D1708 for all melt blends and all for 5 min. The samples were then cooled to 100 8C and held
hompolymers. Tensile testing was performed at room for 5 min before being quenched to 30 8C. Spacers were
temperature on an Instron 5500 at a crosshead speed of used to control the final thickness of the bilayer adhesion
10 mm/min. A minimum of four samples were tested for samples. Any overflow that was formed during pressing was
each material. Stress–strain curves were obtained and the trimmed away using a hot wire cutter. In addition, to avoid
ultimate elongation, tensile modulus and tensile toughness artificially high values for the adhesive strength, the layer
(area under the stress–strain curve) were determined. The thicknesses were varied in order to find the minimum
modulus was calculated from the slope of the initial linear critical strain energy release rate [44].
portion of the stress–strain curve. While ASTM D1708 For the HDPE samples, a mold with dimensions of 20!
suggests that the modulus of elasticity be determined using 35!4.2 mm3 was used to prepare the adhesion samples.
other standards, the use of D1708 was necessary due to the This was necessary to avoid uneven sample surfaces due to
limited amount of sample available. the lower molecular weight of the HDPE which caused it to
flow when it was not confined on all sides. Due to the
2.5. Adhesion testing sample preparation technique the layer thickness could not
be varied greatly; therefore, the minimum critical strain
The adhesion between the PLLA and the LLDPE2 and energy release rate may not have been attained with these
HDPE was determined via the dual cantilever beam crack samples. However, the layer thickness was similar within
propagation test [44]. The plaques for the adhesion test the HDPE samples so a fair comparison can be made among
samples were prepared by compression molding under an these samples.
ultimate pressure of 0.20 MPa for 10 min. The PLLA To perform the adhesion testing on the PLLA/LLDPE2
plaques were pressed at 200 8C, while the LLDPE2 and and PLLA/HDPE bilayer samples, a razor blade was
HDPE plaques were pressed at 150 8C. In order to prepare inserted at the interface of the adhesion sample. One hour
adhesion samples with block copolymer at the interface, the after insertion, the crack length ahead of the razor blade tip
block copolymers were spin coated onto the PLLA plaque was measured. In addition, the thickness of each layer was
surface. Prior to spin coating, the PLLA plaques were measured after testing. The strain energy release rate was
placed in a 100 8C oven for 24 h to remove residual water in then calculated using the following equation [44]:
order to prevent bubble formation during bilayer sample
preparation. 3D2 E1 E2 h31 h32 ðC12 E2 h32 C C22 E1 h31 Þ
Gc Z (5)
For the PLA-PEP and PLLA-PEP block copolymers, 8a4 ðC13 E2 h32 C C23 E1 h31 Þ2
spin coating was done at 2000 rpm from a 3 mg/mL solution
in chlorobenzene. This concentration was found to give the Where: CiZ1C0.64(hi/a), D is the razor blade thickness, Ei
greatest adhesion out of seven other concentrations tested is the elastic modulus, a is the crack length ahead of the
8814 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

razor blade and hi is the layer thickness. A minimum of six (Fig. 2), which influences their crystallinity. The semi-
measurements were made for each block copolymer. crystalline PLLA-PE block copolymer used in our previous
Due to the very low modulus of the LLDPE1, preparation studies was also used in this study [9,10]. In addition, a
of the dual cantilever crack propagation test samples with block copolymer with a semi-crystalline PE block but an
the proper geometry for obtaining adhesion results proved to amorphous polylactide block (PLA-PE), a block copolymer
be difficult. Therefore, for the PLLA/LLDPE1 system, a with a semi-crystalline PLLA block and an amorphous
1808 peel test was performed to determine the adhesion poly(ethylene-alt-propylene) block (PLLA-PEP), and a
between these materials [45]. The plaques and adhesion completely amorphous PLA-PEP block copolymer were
samples were prepared as described above for the dual prepared.
cantilever crack propagation test samples; however, the
LLDPE1 plaque was over twice as long as the PLLA plaque 3.2. Melt miscibility of homopolymers and block copolymers
to allow for the material to bend back over the adhesion
sample and be gripped in the tensile machine. The testing While the miscibility of the PLLA homopolymer and the
was done on an Instron 5500 at a crosshead speed of 24 mm/ polylactide segments in the block copolymers is inferred
min. Three samples were tested for each condition. The from established studies [49], the miscibility of the
adhesion value was calculated by dividing the average load polyethylene homopolymers used in this study with the
required to peel the LLDPE1 by the width of the adhesion PE and PEP blocks in the block copolymers will be
sample. When determining the average load, the first and important in determining their capacity to compatibilize the
last 20 mm of data was discarded to account for startup and binary blends. Based on literature reports, we assume that
edge effects. the PE in the block copolymers is miscible with the three
polyethylene homopolymers. For instance, HDPE has been
reported to be miscible with butene-based LLDPEs with less
3. Results than 40 branches per 1000 backbone carbons [50–52]. The
PE segments in both PLLA-PE and PLA-PE are essentially
3.1. Blend components identical to butene-based LLDPEs with low branch
contents. In addition, a hexene based LLDPE was shown
Two linear low-density polyethylenes (LLDPE1 and to be miscible with a LDPE at similar molecular weights
LLDPE2) and a high-density polyethylene (HDPE) were and compositions as those in this study [53]. The
chosen as the dispersed phase polymers. The two LLDPEs insensitivity of the miscibility to the LLDPE branch length
are both ethylene/octene copolymers that differ only in the (butene versus hexene) was also discussed [53], indicating
amount of octene incorporation, which significantly influ- that the octene based LLDPEs in our study should behave in
ences the degree of crystallinity (HDPEOLLDPE2O a similar fashion to the hexene based LLDPE and thus be
LLDPE1) (Table 1). miscible with the PE blocks.
The tensile properties of the PLLA and PE homopoly- Due to the amorphous nature of the PEP block and the
mers are also given in Table 1, which includes the ultimate low level of crystallinity in LLDPE1, the miscibility of
elongation (3b), tensile modulus, ultimate tensile stress (sb), these two polymers was verified through thermal analysis
and tensile toughness (TT). The LLDPE1 is a rubbery and the presence of a single Tg intermediate between the Tgs
material with a tensile modulus of only 8.0 MPa and an of the PEP and LLDPE1 (in a 50/50 wt% blend of LLDPE1
ultimate elongation of 960%, while LLDPE2 is a stiffer and PLA-PEP block copolymer) [54]. Based on this result,
material with a tensile modulus of 120 MPa and an ultimate we can infer that LLDPE2 should also be miscible with
elongation of 740%. The HDPE used in this study has a PEP, since the LLDPE2 has a lower branch content and the
tensile modulus of 440 MPa and an ultimate elongation of degree of miscibility is reported to increase with a decrease
150%. However, even the HDPE has a tensile modulus that in branch content for the LLDPEs [51]. However, the
is significantly lower than the homopolymer PLLA miscibility of the HDPE homopolymer with the PEP block
(z2.4 GPa). All of the PE homopolymers have a relatively was not definitively determined [55].
low ultimate tensile stress compared to the PLLA, while the
tensile toughness of these polymers is significantly higher 3.3. Mechanical properties of PLLA composites
than the PLLA.
The molecular characteristics of the block copolymers 3.3.1. Tensile properties
are also given in Table 1. The block copolymers were The tensile testing results for the PLLA/PE blends are
synthesized based on established synthetic procedures given in Table 2. In general, when any of the three
developed in our laboratory [10,46]. All of the block polyethylenes are introduced into the PLLA the tensile
copolymers contain blocks of approximately 30 kg/mol modulus decreases. The addition of block copolymer to the
each (above the entanglement molecular weights of both blend causes a further reduction in the modulus; however,
polylactide and polyethylene) [47,48]. The four block the modulus of the ternary blends is fairly insensitive to the
copolymers differ only in their repeat unit microstructures block copolymer structure. The morphologies of the PLLA/
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8815

Table 2
Properties of PLLA and PLLA/PE blends

Dispersed Block Modulus 3b (%) sb (MPa) TT (MJ m3) IS (J/m) MLT (mm) PLLA crystal-
phase PE copolymer (MPa) linity (%)
Nonea None 2400G60 4G1 62.1G1.0 1.5G0.1 20G2 NA
LLDPE1 None 1680G140 23G6 21.9G6.4 6.7G1.6 490G200 1.1G0.5 3.5
LLDPE1 PLLA-PE 1320G110 31G18 24.3G1.1 7.7G4.8 760G50 0.4G0.1 7.6G6.7
LLDPE1 PLA-PE 1200G40 26G16 23.0G1.3 6.3G4.0 730G40 0.5G0.1 1.1G0.5
LLDPE1 PLLA-PEP 1190G20 19G7 23.0G1.8 4.3G1.5 710G40 0.6G0.1 0.2G0.5
LLDPE1 PLA-PEP 1200G40 27G17 22.6G0.7 5.9G3.9 660G40 0.6G0.1 0
LLDPE2 None 1570G70 16G7 29.5G1.9 4.7G2.2 50G8 1.5G0.1
LLDPE2 PLLA-PE 1330G60 33G27 26.7G3.4 8.5G7.0 230G30 0.3G0.2
LLDPE2 PLA-PE 1270G90 62G46 23.8G1.7 16G12 420G30 0.5G0.1
LLDPE2 PLLA-PEP 1310G60 15G6 25.2G0.9 3.8G1.4 220G30 0.5G0.1
LLDPE2 PLA-PEP 1290G70 26G9 23.4G1.0 6.1G2.1 150G30 0.5G0.1
HDPE None 1710G60 2.9G0.2 41.8G2.1 0.7G0.1 12G6 1.2G0.1
HDPE PLLA-PE 1590G60 3.7G0.5 44.5G1.9 1.0G0.2 24G9 0.8G0.1
HDPE PLA-PE 1540G90 4.2G0.8 42.3G2.5 1.2G0.3 32G7 0.8G0.1
HDPE PLLA-PEP 1440G90 14G8 27.7G1.1 3.7G2.4 42G7 0.8G0.1
HDPE PLA-PEP 1390G60 13G4 25.0G1.1 3.3G1.0 64G20 0.8G0.1
a
PLLA homopolymer was melt blended as a control following the blend preparation procedure described in Section 2.

LLDPE2 blends with and without the PLLA-PE block in an increased impact strength for the resulting ternary
copolymer were examined by SEM (Fig. 3) (tensile bars blends.
cryo-fractured parallel to the test direction). Both of these To fairly compare the toughening ability of the different
tensile bars broke at similar elongations (w28%) and had systems, the matrix ligament thickness for each composite
similar values of yield stress (w37 MPa); however, with the was calculated (see Section 2) and is given in Table 2. As
PLLA-PE compatibilizer there is a significant increase in Wu has shown [28,29], the matrix ligament thickness has a
the deformation of both the matrix PLLA and the LLDPE2 significant influence on the impact properties. From our
as compared to the binary blend. The reduction in modulus previous study, we determined a critical matrix ligament
for the ternary blend can be associated with both increased thickness for PLLA of approximately 1.0 mm [9]. For all of
plastic deformation of the PLLA and an increased the binary blends, the matrix ligament thicknesses are above
mechanical coupling of the PLLA and the LLDPE2. the critical value which partially accounts for their inferior
In addition to a significant decrease in the tensile impact properties compared to the compatibilized blends.
modulus, the ultimate tensile stress in the composites is Due to the immiscibility of the PLLA and PE homopoly-
reduced by a factor of 2–3 in all cases relative to the PLLA mers, a block copolymer is necessary in order to
homopolymer. Conversely, an increase in the ultimate compatibilize the homopolymers and achieve a uniform
elongation over that of the homopolymer PLLA is seen dispersion of the polyethylenes in the matrix PLLA. When
when the polyethylene is added with the exception of three the block copolymers are used, the matrix ligament
of the HDPE blends. For the LLDPE1 and LLPDE2 blends, thicknesses of all of the blends are well below the critical
the tensile toughness increases relative to pure PLLA by a value resulting in an increased impact resistance over that of
factor of 2–4. The smallest reduction in the tensile modulus the binary blends. However, the level of improvement in the
and the ultimate tensile stress was obtained when HDPE was impact properties is dependent on which block copolymer is
used for the dispersed phase; however, these blends also used. Furthermore, the block copolymer that gives the
gave the smallest increase in the ultimate elongation and largest improvement in toughening depends on which PE is
tensile toughness. Thus, at 20 wt% of polyethylene used as the dispersed phase.
incorporation, a reduction in the tensile modulus must be In the case of the ternary LLDPE1 blends, to a first
tolerated to achieve a significant increase in the tensile approximation, the matrix ligament thicknesses were
toughness. identical within error (Table 2). The highest impact
resistance was obtained when using the PLLA-PE block
3.3.2. Impact properties copolymer. SEM images of the impact bar fracture surfaces
The impact resistance of the composites generally showed distinctly different fracture behavior between the
increases relative to the PLLA homopolymer when poly- PLLA-PE (highest impact strength) and the PLA-PEP
ethylene is added (Table 2). The ultimate impact resistance (lowest impact strength) compatibilized blends (Fig. 4).
that can be obtained in the blends increases as the modulus When the PLLA-PE block copolymer was used significant
of the dispersed phase polyethylene decreases. In all cases, matrix deformation is evident which is characteristic of
addition of a block copolymer to the binary blends resulted shear yielding, and the LLDPE1 particles are not clearly
8816 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

reduced compared to the PLA-PE block copolymer;

however, around the LLDPE2 particles that did ‘partially
debond’ from the matrix we observed significant damage to
the LLDPE2 particles (Fig. 5(d)) in contrast to the PLA-PE
blends where only slight deformation of the LLDPE2
particles was seen.
When the HDPE was used as the dispersed phase, the
most significant increase in the impact resistance occurs
with the PLA-PEP block copolymer. With the PLLA-PE
block copolymer, the impact properties were nearly
identical to the homopolymer PLLA. As with the other
blends, the matrix ligament thicknesses of the block
copolymer blends are similar to one another and below
the critical value of 1.0 mm. With the PLA-PEP block
copolymer very distinct interfaces are seen around the
particles indicating debonding at the interface (Fig. 6(a));
whereas with the PLLA-PE block copolymer damage to the
HDPE particles, with an extensive number of fibrils, was
observed (Fig. 6(b)).
The crystallinity of the PLLA matrix phase can influence
the mechanical properties of the composites, so the level of
crystallinity for the matrix PLLA in the impact bars was
determined using differential scanning calorimetry (DSC)
(Table 2). The crystallinity of the PLLA in the blends with
LLDPE2 and HDPE could not be reliably determined using
thermal analysis due to overlap of the PLLA crystallization
exotherm with the polyethylene melting endotherm. How-
ever, based on the DSC data for the PLLA/LLDPE1
composites and estimates obtained from the DSC data for
the PLLA/LLDPE2 and PLLA/HDPE samples, the level of
crystallinity in the PLLA matrix for these blends is also
expected to be relatively low (!10%).
In addition, an annealing study was performed in order to
Fig. 4. SEM images of 80/20 wt% PLLA/LLDPE1 impact bar
fracture surfaces. Images were taken behind notch area. (a) With determine the influence of crystallinity on the impact
5 wt% PLLA-PE block copolymer and (b) with 5 wt% PLA-PEP properties. Impact bars of PLLA/LLDPE/PLLA-PE
block copolymer. quenched directly to room temperature from 200 8C showed
no (0%) matrix phase crystallinity, and an impact strength of
750G30 J/m was obtained. This impact strength is
defined from the matrix PLLA (Fig. 4(a)). However, when comparable to the impact strength given in Table 2 for the
the PLA-PEP block copolymer was used both debonding of samples with 7.6% crystallinity. However, when the PLLA/
the particle/matrix interface and deformation of the LLDPE1/PLLA-PE samples were annealed at 100 8C for
LLDPE1 particles is evident (Fig. 4(b)). 30 min the level of crystallinity in the PLLA matrix
For the LLDPE2 dispersed phase, all of the ternary increased to 46.7% and the impact strength decreased to
blends had similar matrix ligament thicknesses and the 470G30 J/m. Based on this data, the impact strengths of
highest impact strength was obtained when using the PLA- these composites are not that sensitive to the level of
PE block copolymer. The SEM images of these blends show crystallinity at relatively low levels (0 to ca. 8%); however,
that although some deformation of the LLDPE2 particles is at higher levels of crystallinity, the impact strength
evident when using the PLA-PE block copolymer, the decreases by nearly 40%. This phenomenon is currently
LLDPE2 does appear to be able to debond from the matrix under investigation.
PLLA (Fig. 5(a)). By changing to the PLA-PEP block
copolymer, an increase in the density of debonded particles 3.4. Adhesion testing
was observed (Fig. 5(b)). With the PLLA-PEP block
copolymer, we observed near complete debonding at the Adhesion testing was used to directly measure the
interface with minimal deformation of the LLDPE2 influence of the block copolymer on the interfacial adhesion
particles (Fig. 5(c)). Using the PLLA-PE block copolymer, between the PLLA and the three polyethylene homopoly-
the density of debonded particles from the matrix was mers. For the LLDPE2 and HDPE, a dual cantilever crack
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8817

Fig. 5. SEM images of 80/20 wt% PLLA/LLDPE2 impact bar fracture surfaces. Images were taken behind notch area. (a) With 5 wt% PLA-
PE block copolymer, (b) with 5 wt% PLA-PEP block copolymer, (c) with 5 wt% PLLA-PEP block copolymer and (d) with 5 wt% PLLA-PE
block copolymer (higher magnification).

propagation test was used [44]. For the LLDPE1, we increase in the adhesion when the PLLA-PEP or PLA-PEP
employed a 1808 peel test due to difficulties associated with block copolymer was used compared to the bare interface.
preparing specimens for the dual cantilever crack propa- However, using the PLLA-PE block copolymer resulted in
gation test [56] (see Section 2). The adhesion strengths for an increase of the adhesion strength by a factor of 160 over
each combination are given in Table 3. Because of the the no block copolymer case, and using the PLA-PE block
difference in testing geometry and testing rates, the adhesion copolymer caused the adhesion strength to increase by a
strengths using the two tests cannot be directly compared to factor of approximately 170. After considering the large
each other. standard deviation associated with the PLLA-PE measure-
For the PLLA/LLDPE1 system, there is only a slight ment, the adhesion values for these two block copolymers

Table 3
Adhesion testing results

Adhesion value (J/m2)

LLDPE1 (1808 Peel test) LLDPE2 (Dual cantilever test) HDPE (Dual cantilever test)
No block copolymer 9.1G1.4 Delaminated 0.3G0.0
PLLA-PE 1476G250 20.2G6.3 6.1G2.7
PLA-PE 1520G110 28.5G5.5 7.0G0.8
PLLA-PEP 10.7G1.7 Delaminated 0.5G0.1
PLA-PEP 9.6G1.0 Delaminated 0.6G0.1
8818 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

7. No significant features are seen for the interface

containing no block copolymer (Fig. 7(a)). When the
PLA-PEP and PLLA-PEP block copolymers were used
some surface roughness on both the LLDPE2 and PLLA
sides was noted but significant damage was not evident
(data not shown). Extensive deformation of the LLDPE2
was observed when either the PLA-PE or the PLLA-PE
block copolymer was used (Fig. 7(b) and (c)). In addition,
the PLLA surface showed fibril like material for both the
PLLA-PE and PLA-PE samples and a higher magnification
image of the PLLA surface from the adhesion sample
containing PLA-PE is shown in Fig. 8.
Finally, adhesion testing was performed between the
PLLA and the HDPE. The adhesion with no block
copolymer and with either the PLLA-PEP or PLA-PEP
block copolymers was poor. The SEM image of the HDPE
surface of the adhesion sample when no block copolymer
was present is given in Fig. 9(a). The polyethylene side of
the interface shows significant surface roughness and this is
presumably responsible for the measurable adhesion
between the PLLA and HDPE [57,58]. The surfaces of the
PLLA-PEP and PLA-PEP block copolymer-containing
samples appeared similar to the adhesion sample without
block copolymer. As in the previous cases, the PLLA-PE
and PLA-PE block copolymers gave the largest improve-
ment in adhesion for the PLLA/HDPE interface. The
surfaces of the PLA-PE and the PLLA-PE block copolymer
adhesion samples were similar and only the HDPE surface
from the PLLA-PE block copolymer-containing sample is
shown in Fig. 9(b). The surface of the HDPE was markedly
different when comparing the samples without (Fig. 9(a))
Fig. 6. SEM images of 80/20 wt% PLLA/HDPE impact bar and with (Fig. 9(b)) block copolymer. However, unlike the
fracture surfaces. Images were taken behind notch area. (a) With LLDPE2 samples (e.g. Fig. 7(b)), significant damage to the
5 wt% PLA-PEP block copolymer and (b) with 5 wt% PLLA-PE HDPE is not evident with the PLLA-PE or PLA-PE block
block copolymer (higher magnification). copolymer.
are comparable. SEM images of the adhesion sample
surfaces with PEP-containing block copolymers appeared
4. Discussion
essentially the same as the surfaces when no block
copolymer was used. The LLDPE1 surfaces after the
Central to this study is the influence of the block
adhesion testing of the samples with the PE-containing copolymer structure on the interfacial adhesion between the
block copolymers showed some surface roughness, but the dispersed polyethylenes and the matrix PLLA. The adhesion
PLLA surfaces were very similar to the samples that testing results suggest that the ability of the polyolefin in the
contained no block copolymer. block copolymer to co-crystallize with the dispersal phase
The adhesion strength for three of the PLLA and polyethylene results in a significant improvement in the
LLDPE2 laminates could not be measured due to spon- interfacial adhesion (Table 3). The crystallinity of the
taneous delamination (no block copolymer, PLA-PEP and polylactide block in the block copolymers made little
PLLA-PEP), which was likely due to thermal stresses difference in the interfacial adhesion. This is likely due to
induced by the difference in thermal expansion coefficients the mechanical interlocking of the entangled polylactide
between the two polymers [25]. When the PLLA-PE and blocks by virtue of polylactide being in a glassy state at the
PLA-PE block copolymers were used, the adhesion testing temperature. The low strain rate experienced in the
increased to a measurable value. Again there was only a adhesion test cannot detect the difference in the adhesion
slight difference in the adhesion values between the PLLA- due to the ability of the PLLA block to co-crystallize
PE and PLA-PE block copolymers, with the PLA-PE block compared to the PLA block being entangled in the glassy
copolymer giving slightly higher adhesion. The SEM matrix.
images of the adhesion sample surfaces are shown in Fig. To test this hypothesis, an adhesion test was done on the
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8819

Fig. 7. SEM images of PLLA/LLDPE2 adhesion samples after testing. The LLDPE2 surfaces are shown on the left side and the PLLA
surfaces are shown on the right side of the figure. (a) With no block copolymer, (b) with PLA-PE block copolymer and (c) with PLLA-PE
block copolymer.
8820 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

PLLA/LLDPE2/PLLA-PE and PLLA/LLDPE2/PLA-PE

samples at 80 8C, a temperature above the Tg of the
PLLA. The crack was initiated at room temperature with a
razor blade and the crack length ahead of the blade was
measured after equilibration. The samples were then placed
in an 80 8C oven and allowed to equilibrate for 2 h, after
which the crack length was measured again. With the
PLLA-PE block copolymer at the interface, the crack grew
an average of 0.2G0.1 mm over that of the initial crack
length; however, with the PLA-PE block copolymer at the
interface the crack grew on average 0.6G0.2 mm. This
result shows that in fact the PLA-PE block copolymer does
give a lower level of adhesion if the samples are evaluated
above the Tg of polylactide where the amorphous PLA is no
longer locked into the glassy matrix.
At impact testing speeds the difference in the interfacial
Fig. 8. Higher magnification SEM images of PLLA surface of adhesion between block copolymers is clear. For instance,
PLLA/LLDPE2/PLA-PE adhesion samples after testing. in Fig. 5(d) there are a significant number of fibrils and
damaged particles when the PLLA-PE block copolymer was
used with the LLDPE2 but when the PLA-PE block
copolymer was used (Fig. 5(a)) only slight deformation of
the LLDPE2 particles was seen. From this we conclude that
the LLDPE2/PLLA interfacial adhesion with the PLLA-PE
block copolymer is stronger compared with the composite
using the PLA-PE block copolymer, contrary to the room
temperature adhesion testing where these two block
copolymers behaved in a similar manner. One possibility
for the differences between the adhesion testing and the
impact testing is that at high testing rates the samples
undergo adiabatic heating due to poor thermal diffusivity
[59–61]. This could increase the temperature of the sample
above the Tg of the PLLA and allow slipping of the PLA
block and thus a decrease in adhesion during impact testing
for the PLA-PE block copolymer. Taking this into
consideration, we propose that the interfacial adhesion
strengths between the matrix PLLA and the dispersed phase
polyethylenes at impact testing speeds decrease in the
following order PLLA-PEOPLA-PEOPLLA-PEPOPLA-
PEP. The higher adhesion for the PE containing blocks
relative to the PEP containing blocks is consistent with the
adhesion testing data.
The interfacial adhesion requirements for optimum
toughening with the different dispersed phases can now be
determined from the impact testing data. For the rubbery
LLDPE1 dispersed phase, the strongest interfacial adhesion
is necessary to obtain the highest impact strength, since the
most significant improvement was seen with the PLLA-PE
block copolymer. The fracture surface of the PLA-PEP
block copolymer sample shown in Fig. 4(b) reveals that the
particles debonding during fracture (because of the low
interfacial adhesion) is most likely the cause of the lower
impact properties due to poor transfer of stress between the
matrix and dispersed phases. According to Dompas et al.
Fig. 9. SEM image of HDPE surfaces of PLLA/HDPE adhesion [62] in an Izod impact test, relief of the hydrostatic tension
sample after testing. (a) With no block copolymer and (b) with by microvoiding (either cavitation or debonding) is essential
PLLA-PE block copolymer. for the initiation of matrix shear yielding. In the case of the
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8821

rubbery LLDPE1, in order to achieve the optimum impact Comparing the overall values of the impact properties of
properties debonding is not sufficient. Rather, good the composites with the different PE dispersed phases we
interfacial adhesion is necessary in order to allow for observed significant differences. The stiff HDPE gives the
putative rubber particle cavitation. lowest values, while the rubbery LLDPE1 gives the highest
Conversely, for the stiff HDPE dispersed phase, the block overall impact properties. To directly compare all of the
copolymer that leads to the lowest interfacial adhesion PLLA composites with the three different polyethylenes, a
(PLA-PEP) gives the greatest improvement in impact plot of impact resistance versus matrix ligament thickness is
properties. With the PLA-PEP block copolymer very given in Fig. 10. The block copolymer blends for the PLLA/
distinct interfaces are seen around the particles in Fig. HDPE system clearly have larger matrix ligament thick-
6(a) indicating debonding at the interface; whereas with the nesses than the other blends; however, they are still below
PLLA-PE block copolymer significant damage to the HDPE the critical value so a larger improvement in impact
particles, with an extensive number of fibrils, can be seen properties would be expected for the PLA-PEP blends that
(Fig. 6(c)). This damage indicates that stresses were show debonding of the particle/matrix interface. A possible
effectively transferred across the interface to the stiff reason for the lower impact properties of the PLLA/HDPE/
HDPE particles. The poor impact properties for the PLA-PEP blend could be that even the adhesion from the
PLLA-PE blends are most likely due to the inability of the entanglements of PEP with HDPE is strong enough to
interface to debond or the HDPE to cavitate. The need for hinder debonding when required for significant property
complete debonding at the interface when stiff dispersed improvement.
phase materials are used has been noted in the literature and In addition to the HDPE blends, the impact properties of
is consistent with this result [59]. the PLLA/LLDPE2 blends are significantly lower than the
For the intermediate modulus LLDPE2 dispersed phase, PLLA/LLDPE1 blends. This trend shows that the properties
the PLA-PE block copolymer composites showed the best of the dispersed phase play an important role in the ultimate
impact properties. In fact when the interface is able to toughness of the system. The matrix ligament thicknesses
completely debond (PLA-PEP block copolymer), the were essentially the same between the PLLA/LLDPE2 and
impact properties were diminished relative to the PLA-PE the PLLA/LLDPE1 blends so the matrix ligament thickness
composite. In addition, when stress transfer is at its can be ruled out as a possible difference for these blends.
maximum (PLLA-PE block copolymer) the impact proper- Although, several researchers have suggested that the
ties are still not able to reach their highest levels. With the critical matrix ligament thickness does shift as the dispersed
PLLA-PE block copolymer blend a significant number of phase properties change, getting smaller as the modulus of
fibrils are seen at the particle interface and damaged the dispersed phase increases [31,37]. However, the same
LLDPE2 particles can be found (Fig. 5(d)), thus indicating critical matrix ligament thickness was found when toughen-
that stresses were transferred to the LLDPE2 but complete ing of HDPE was performed with both rubbery materials
debonding of the interface was not possible due to the and CaCO3, a stiff material, with only a slight decrease in
increased interfacial adhesion. Similar to the case of HDPE the ultimate impact properties [59].
blends, the inability of the LLDPE2 particles to fully
cavitate is most likely responsible for the lower impact
properties of the blends with the PLLA-PE block copolymer 5. Conclusions
where complete debonding could not occur. These results
indicate that an intermediate level of adhesion is necessary We investigated the influence of interfacial adhesion and
for optimal toughening with the LLDPE2 dispersed phase. dispersed phase properties on the toughening of PLLA. The
Interestingly, when the PLLA-PEP block copolymer was interfacial adhesion was altered by introducing different
used in the PLLA/LLDPE2 blends, where the adhesion on block copolymers with varying abilities to crystallize.
the PLLA side would be good, the impact properties are still Direct measurements of interfacial adhesion were obtained
significantly lower than when the PLA-PE block copolymer using a 180 8C peel test and dual cantilever crack
was used. The reason for this can be seen in the SEM images propagation test. Using these tests, we found increased
shown in Fig. 5. When the PLA-PE block copolymer is adhesion strength with the PLLA-PE and PLA-PE block
present, the interface between the PLLA and LLDPE2 is copolymers, an effect seemingly due to the co-crystal-
able to debond (Fig. 5(a)); however, some deformation and lization of the PE block with the polyethylene homopoly-
stretching of the LLDPE2 particles is evident indicating the mers. The effect of the interfacial adhesion on the impact
debonding occurs after some transfer of stress. In the lower properties of PLLA was determined using three different
impact strength PLLA-PEP block copolymer blend, we polyethylenes as the dispersed phase. With a rubbery
observed complete debonding at the interface without any polyethylene, LLDPE1, we found that the strongest
deformation of the LLDPE2 particles (Fig. 5(c)). Therefore, interfacial adhesion, PLLA-PE block copolymer, was
in the case of the PLLA/LLDPE2 blends it appears that necessary in order to achieve the greatest improvement in
debonding must occur after some stress transfer to the toughening. Using the stiff polyethylene, HDPE, we found
dispersed phase particles. that the weakest interfacial adhesion, PLA-PEP block
8822 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823

Fig. 10. Variation of impact resistance with matrix ligament thickness for all PLLA/dispersed phase blends. (B) LLDPE1 blends, (6)
LLDPE2 blends, (,) HDPE blends.

copolymer, gave the best impact properties. Finally, with a [9] Anderson KS, Lim SH, Hillmyer MA. J Appl Polym Sci 2003;89:
polyethylene that had an intermediate modulus, LLDPE2, 3757–68.
[10] Wang Y, Hillmyer MA. J Polym Sci, Part A: Polym Chem 2001;39:
an intermediate level of adhesion, PLA-PE block copoly- 2755–66.
mer, was necessary to give the greatest toughening. This [11] Tams J, Joziasse CAP, Bos RRM, Rozema FR, Grijpma DW,
result shows that subtle differences in the degree of adhesion Pennings AJ. Biomaterials 1995;16:1409–15.
between the matrix and dispersed phase can have a [12] McCarthy SP, Ranganthan A, Ma W. Macromol Symp 1999;144:
significant influence on the resultant toughening in these 63–72.
[13] Grijpma DW, Van Hofslot RDA, Super H, Nijenhuis AJ, Pennings AJ.
blends. In addition, the properties of the dispersed phase
Polym Eng Sci 1994;34:1674–84.
play an important role in the attainable toughening. The [14] Hiljanen-Vainio M, Varpomaa P, Seppala J, Tormala P. Macromol
lowest degree of toughening was achieved with the stiff Chem Phys 1996;197:1503–23.
HDPE while the most significant increase in toughening was [15] Nijenhuis AJ, Colstee E, Grijpma DW, Pennings AJ. Polymer 1996;
found with the rubbery LLDPE1. 37:5849–57.
[16] Kim K, Chin I, Yoon JS, Choi HJ, Lee DC, Lee KH. J Appl Polym Sci
2001;82:3618–26.
[17] Focarete ML, Scandola M, Dobrzynski P, Kowalczuk M. Macromol-
ecules 2002;35:8472–7.
Acknowledgements [18] Tsuji H, Ikada Y. J Appl Polym Sci 1996;60:2367–75.
[19] Wang L, Ma W, Gross RA, McCarthy SP. Polym Degrad Stab 1998;
The authors would like to thank Laura Crawford for 59:161–8.
performing the tensile testing of the Engage 8540 and [20] Joziasse CAP, Topp MDC, Veenstra H, Grijpma DW, Pennings AJ.
Polym Bull 1994;33:599–605.
Engage 8100 polyethylenes. The authors also thank
[21] Hiljanen-Vainio M, Kylma J, Hiltunen K, Seppala JV. J Appl Polym
Professor Christopher W. Macosko and Dr Yonathan Thio Sci 1997;63:1335–43.
for helpful input during the preparation of this manuscript. [22] Ruckenstein E, Yuan Y. Polym Bull 1998;40:485–90.
The David and Lucile Packard Foundation and the Toyota [23] Our focus has been on the toughening of the semi-crystalline, isotactic
Motor Company are acknowledged for financial support of PLLA, since it is the more commercially relevant material (as
compared with amorphous, atactic polylactide) due to its higher end-
this work.
use temperature and superior mechanical properties See: Perego G,
Cella GD, Bastioli C. J Appl Polym Sci 1996;59:37.
[24] Benkoski JJ, Flores P, Kramer EJ. Macromolecules 2003;36:
3289–302.
References [25] Cole PJ, Cook RF, Macosko CW. Macromolecules 2003;36:2808–15.
[26] Chaffin KA, Knutsen JS, Brant P, Bates FS. Science 2000;288:
[1] Tullo A. Chem Eng News 2000;78:13. 2187–90.
[2] Grijpma DW, Pennings AJ. Macromol Chem Phys 1994;195: [27] Chaffin KA. Semi-crystalline polymer blends: thermodynamics,
1649–63. morphology, and toughness enhancement. PhD Thesis, University
[3] Hiljanen-Vainio M, Karjalainen T, Seppala J. J Appl Polym Sci 1996; of Minnesota; 1999.
59:1281–8. [28] Wu S. Polymer 1985;26:1855–63.
[4] Ruckenstein E, Yuan Y. J Appl Polym Sci 1998;69:1429–34. [29] Wu S. J Appl Polym Sci 1988;35:549–61.
[5] Baiardo M, Frisoni G, Scandola M, Rimelen M, Lips D, Ruffieux K, [30] Liu Z, Zhu X, Wu L, Li Y, Qi Z, Choy C, Wang F. Polymer 2001;42:
Wintermantel E. J Appl Polym Sci 2003;90:1731–8. 737–46.
[6] Jacobsen S, Fritz HG. Polym Eng Sci 1999;39:1303–10. [31] Van der Sanden MCM, de Kok JMM, Meijer HEH. Polymer 1994;35:
[7] Labrecque LV, Kumar RA, Dave V, Gross RA, McCarthy SP. J Appl 2995–3304.
Polym Sci 1997;66:1507–13. [32] Bartczak Z, Argon AS, Cohen RE, Weinberg M. Polymer 1999;40:
[8] Martin O, Averous L. Polymer 2001;42:6209–19. 2331–46.
 K.S. Anderson, M.A. Hillmyer / Polymer 45 (2004) 8809–8823 8823

[33] Borggreve RJM, Gaymans RJ. Polymer 1989;30:63–70. [48] Fetters LJ, Lohse DJ, Richter D, Witten TA, Zirkel A. Macromol-
[34] Muratoglu OK, Argon AS, Cohen RE, Weinburg M. Polymer 1995; ecules 1994;27:4639–47.
36:921–30. [49] Tsuji H, Ikada YJ. Appl Polym Sci 1995;58:1793–802.
[35] Sue HJ, Huang J, Yee AF. Polymer 1992;33:4868–71. [50] Alamo RG, Graessley WW, Krishnamoorti R, Lohse DJ, Londono JD,
[36] Borggreve RJM, Gaymans RJ, Eichenwald HM. Polymer 1989;30: Mandelkern L, Stehling FC, Wignall GD. Macromolecules 1997;30:
78–83. 561–6.
[37] Borggreve RJM, Gaymans RJ, Schuijer J. Polymer 1989;30:71–7. [51] Hussein IA. Macromolecules 2003;36:2024–31.
[38] Chaffin KA, Bates FS, Brant P, Brown GM. J Polym Sci, Part B: [52] Choi P. Polymer 2000;41:8741–7.
Polym Phys 2000;38:108–21. [53] Hameed T, Hussien IA. Polymer 2002;43:6911–29.
[39] Liu ZH, Zhang XD, Zhu XG, Li RKY, Qi ZN, Wang FS, Choy CL. [54] Unfortunately, due to the relatively high level of crystallinity in
Polymer 1998;39:5019–25. LLDPE2, HDPE and the PE block of PLLA-PE and PLA-PE, a
[40] Witzke DR, Narayan R, Kolstad JJ. Macromolecules 1997;30: miscibility study based on the glass transition temperatures, Tg, of
7075–85. these blended materials was not feasible.
[41] Sodergard A, Stolt M. Prog Polym Sci 2002;27:1123–63. [55] Maurer WW, Bates FS, Lodge TP, Almdal K, Mortensen K,
[42] Brandrup J, Immergut EH. Polymer handbook, 3rd ed. New York: Fredrickson GHJ. Chem Phys 1998;108:2989–3000.
Wiley; 1989. [56] Due to the very low elastic modulus of the LLDPE1, the thickness of
[43] After processing, there is some degradation of the PLLA. For the LLDPE1 plaque would need to be extremely thin in order to have
example, we found from SEC analysis that the molecular weight of the crack propagate at the interface using the dual cantilever test.
the matrix PLLA decreased from MnZ85.6 kg/mol (PDIZ1.67) for [57] Yuan BL, Wool RP. Polym Eng Sci 1990;30:1454–64.
an as received pellet, to MnZ67.3 kg/mol (PDIZ2.00) for material [58] Janarthanan V, Garrett PD, Stein RS, Srinivasarao M. Polymer 1997;
that was taken from an impact bar prepared as described in Section 2. 38:105–11.
[44] Creton C, Kramer EJ, Hui CY, Brown HR. Macromolecules 1992;25: [59] Bartczak Z, Argon AS, Cohen RE, Weinberg M. Polymer 1999;40:
3075–88. 2347–65.
[45] Cole PJ. Polymer–polymer adhesion in melt-processed multilayered [60] Muratoglu OK, Argon AS, Cohen RE, Weinberg M. Polymer 1995;
structures. PhD Thesis, University of Minnesota; 2002. 36:4771–86.
[46] Schmidt SC, Hillmyer MA. Macromolecules 1999;32:4794–801. [61] Dijkstra K, ter Laak J, Gaymans RJ. Polymer 1994;35:315–22.
[47] Grijpma DW, Penning JP, Pennings AJ. Colloid Polym Sci 1994;272: [62] Dompas D, Groeninckx G, Isogawa M, Hasegawa T, Kadokura M.
1068–81. Polymer 1995;36:437–41.