Melt Chain Dimensions of Polylactide

Article abstract of DOI:10.1021/ma0357523

Melt chain dimensions of two polylactide samples were measured using small-angle neutron scattering, Three polylactides were synthesized: a deuterated polylactide containing 26% R-stereocenters (d-PLA-26), a hydrogenous polylactide with an R-content similar to the deuterated polylactide (PLA-28), and a hydrogenous polylactide that contained no R-stereocenters (PLA-0). The hydrogenous polylactides were each solution blended with the d-PLA-26 at a volume fraction of 0.2 for the deuterated polymer, and the melt chain dimensions of these polymers were determined. Small-angle neutron scattering experiments were performed at 30°C for the d-PLA-26/PLA-28 blend and at 200°C for both the d-PLA-26/PLA-28 and d-PLA-26/PLA-0 blends. Using three analysis methods, the average values for the statistical segment lengths, based on a C₆ repeat unit, were found to be 10.0 ± 0.2 A for the PLA-28 at 30°C, 8.9 ± 0.2 A for the PLA-28 at 200°C, and 9.9 ± 0.4 A for the PLA-0 at 200°C.

Full text of DOI:10.1021/ma0357523

Macromolecules 2004, 37, 1857-1862
1857
Melt Chain Dimensions of Polylactide
Kelly S. An d er son a n d Ma r c A. Hillm yer *
Department of Chemistry, 207 Pleasant St. SE, University of Minnesota,
Minneapolis, Minnesota 55455-0431
Received November 21, 2003; Revised Manuscript Received J anuary 6, 2004
ABSTRACT: Melt chain dimensions of two polylactide samples were measured using small-angle neutron
scattering. Three polylactides were synthesized: a deuterated polylactide containing 26% R-stereocenters
(d-PLA-26), a hydrogenous polylactide with an R-content similar to the deuterated polylactide (PLA-28),
and a hydrogenous polylactide that contained no R-stereocenters (PLA-0). The hydrogenous polylactides
were each solution blended with the d-PLA-26 at a volume fraction of 0.2 for the deuterated polymer,
and the melt chain dimensions of these polymers were determined. Small-angle neutron scattering
experiments were performed at 30 °C for the d-PLA-26/PLA-28 blend and at 200 °C for both the d-PLA-
2
6/PLA-28 and d-PLA-26/PLA-0 blends. Using three analysis methods, the average values for the statistical
segment lengths, based on a C repeat unit, were found to be 10.0 ( 0.2 Å for the PLA-28 at 30 °C, 8.9
0.2 Å for the PLA-28 at 200 °C, and 9.9 ( 0.4 Å for the PLA-0 at 200 °C.
6
(
In tr od u ction
temperature for 48 h. Diethylaluminum ethoxide (Et
2
AlOEt)
(Aldrich, 1.6 M solution in toluene) and tin(II) 2-ethylhexa-
Polylactide is arguably the most important plastic
derived from renewable resources given its omnipresent
status in the biomedical field and recent commercial
noate (Aldrich) were used as received. Toluene used for the
ring-opening polymerization of lactide was purified by passage
through an activated alumina column and a supported copper
1
2
14
production as a commodity material. The foundation
catalyst on a home-built system. All other solvents were used
of current and future applications of polylactide will de-
pend strongly on the use of reliable physical and chem-
ical property data. As such, the determination of inher-
ent material properties of polylactide has been and will
continue to be an imperative endeavor. Of these char-
acteristics, the unperturbed dimensions of a polymer
chain in the melt are certainly one of the most funda-
without further purification.
Syn th esis of d
,3,3-d acid (6.22 g, 0.0668 mol) was heated to 175 °C under
nitrogen for 24 h, giving low molecular weight polylactic acid
3.20 g). Tin(II) 2-ethylhexanoate (0.098 g, 0.242 mmol) was
6
-La ctid e (Deu ter a ted La ctid e). L-Lactic-
3
3
(
then added to the flask, and the low molecular weight
polylactic acid was heated to 230 °C under reduced pressure
(50 Torr). The pressure was lowered to 12 Torr over a period
3
mental properties of any macromolecule. In addition
of 1 h, during which time d -lactide was distilled into a
6
to rheological properties relevant for polymer processing
and product manufacturing, the chain dimensions are
intimately tied to mechanical properties such as tough-
collection flask (2.65 g, 83% yield). The crude d -lactide was
purified by recrystallization in ethyl acetate. After three
6
successive recrystallizations, the d
lected and dried in the vacuum oven at room temperature for
8 h (0.985 g, 31% purified yield). The synthesized d -lactide
6
-lactide crystals were col-
4
ness, impact resistance, and tensile strength. Further-
4
6
more, differences in chain dimensions have been impli-
cated in polymer-polymer miscibility,5,6 are important
was found to contain 26% of the R-stereocenters from both
D-lactide and meso-lactide based on polarimetry (see the
equipment/general procedures section for further details on
for the self-assembly of conformationally asymmetric
block copolymers,7 and play an important role in inter-
,8
the polarimetry measurements). The fact that pure d
6
-L-lactide
9
,10
facial adhesion and polymer welding.
Many studies
(S,S) was not obtained from the deuterated L-lactic acid was
aimed at the careful determination of the unperturbed
surprising, since pure hydrogenous L-lactide (S,S) was obtained
when the above procedure was performed with hydrogenous
L-lactic acid. The reason for the significant amount of R-
stereocenters in the d -lactide is not known, but epimerization
6
of the S-stereocenter during ring closure (at high temperature)
is the most likely cause.
polylactide chain dimensions have appeared.1
1-13
How-
ever, these literature values are not in agreement. In
examination of this literature, we realized that the
determination of the melt chain dimensions for poly-
lactide by small-angle neutron scattering analysis of
H/D isotopic blends, debatably the most direct measure-
ment of melt chain dimensions, had not been carried
out. In this study we report the unperturbed chain di-
mensions of polylactide in the melt using small-angle
neutron scattering.
Syn th esis of P olyla ctid e. The following is a general pro-
cedure for the synthesis of polylactide with a targeted molec-
-1
ular weight of 60 kg mol (assuming 60% conversion of lac-
tide). In a glovebox, a Teflon-coated stir bar, 2.0 g of lactide
2
(0.013 88 mol), 14 mL of toluene, and 12.5 µL of Et AlOEt (1.6
M in toluene, 0.02 mmol) were placed into a 38 mL Chemglass
high-pressure vessel. The mixture was stirred in an oil bath
at 80 °C for ∼ 24 h. Acetic anhydride (3 mL) was then added
to the reaction vessel to end-cap the polylactide. The high-
pressure vessel was placed back in the 80 °C oil bath for 24 h.
The polylactide was then precipitated into methanol and dried
in a vacuum oven for 48 h at 65 °C (1.4 g, 68% conversion of
lactide). The deuterated polymer (d-PLA) and hydrogenous
polymers with varying amounts of R-stereocenters were syn-
thesized using this method. The end-capping of the polymers
Exp er im en ta l Section
3
Ma ter ia ls. L-Lactic-3,3,3-d acid (methyl deuterated) was
obtained from Isotec as an 85% aqueous solution with a
minimum D atom content of 98% according to the manufac-
turer. L-Lactide (Aldrich), DL-lactide (Aldrich), and deuterated
lactide (prepared as described below) were recrystallized from
ethyl acetate and subsequently dried in a vacuum oven at room
1
was verified using H NMR spectroscopy, and the percentage
*
To whom correspondence should be addressed. E-mail:
of the R-stereocenters in the polymers was determined using
polarimetry. The molecular characteristics of the polylactides
hillmyer@chem.umn.edu.
1
0.1021/ma0357523 CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/13/2004

1
858 Anderson and Hillmyer
Macromolecules, Vol. 37, No. 5, 2004
Ta ble 1. Ch a r a cter iza tion of P olyla ctid e Hom op olym er s
where [R] and [R]
0
are the optical rotations of the sample and
an optically pure sample, respectively. The optically pure
sample was taken to be either the L-lactide obtained from
Aldrich or the PLLA that was synthesized from L-lactide,
Mna
Mnb
(kg mol-1)
PDIa
(kg mol-1)
Mw/Mnb
% Rc
polymer
PLA-0
PLA-28
d-PLA-26
57.8
56.2
70.4
1.20
1.24
1.30
34.8
44.0
48.6
1.08
1.15
1.31
0
28
26
which was found to contain no stereodefects based on methine
decoupled 1H NMR spectroscopy. The percentage of the R-
enantiomer in the lactide or polylactide was then determined
from
a
Determined by SEC vs polystyrene standards (RI detector).
b
Determined by SEC with multiangle light scattering detection.
See Supporting Information for SEC chromatographs. c Deter-
mined by polarimetry.
1
00 - OP
%
{R} )
(2)
2
are given in Table 1. (The nomenclature that will be used in
this study for referring to the polylactides is PLA-x, where x
is the percentage of R-stereocenters in the polymer.) The
polydispersity of d-PLA-26 is slightly higher than either of the
hydrogenous polymers. The reason for the broadening of the
polydispersity is unknown but is likely due to the slightly
higher conversion of monomer for the d-PLA-26 polymer. The
well-controlled nature of this type of polymerization has been
demonstrated with narrow polydispersities being obtained at
low conversions while broadening occurs as the conversion
Neu tr on Sca tter in g Mea su r em en ts. To prepare the
small-angle neutron scattering samples, each of the hydrog-
enous polymers was codissolved with the d-PLA-26 in chloro-
form (5 wt % solution). The blends were then precipitated
into methanol and dried in a vacuum oven at 65 °C for
2
4 h. All blends were prepared with a volume fraction of
1
d-PLA-26 of 0.2 (verified by H NMR spectroscopy). The poly-
mer blends were then pressed into aluminum disks (1 mm ×
1
6 mm i.d.) using a hot press at 200 °C. The samples were
kept under pressure at 200 °C for 5 min after which they were
quenched to ca. 30 °C using circulating water over a period of
1
5
1
increases. Decoupled H NMR spectroscopy was also per-
formed in order to examine the stereosequence distribution
3
min. (Some of the polymer blends needed to be pressed sev-
1
6
in the hydrogenous polymers. Using this technique, the
stereosequence distribution of the hydrogenous PLA-28 was
found to be similar to that of the d-PLA-26.
eral times before a bubble-free sample was obtained; however,
the molecular weights of all of the polymer blends were
determined by SEC with refractive index detector after press-
ing, and no obvious degradation was evident.) The pressed
samples contained in the aluminum disks were then sand-
wiched between quartz disks and sealed using a high-temper-
ature silicone sealant. Blank samples containing only hydrog-
enous PLA-28 and d-PLA-26 were also prepared for incoherent
scattering measurements (see Supporting Information).
The neutron scattering experiments were performed on the
NIST/Exxon/ University of Minnesota 30 m SANS instrument
at NIST in Gaithersburg, MD. A wavelength of 6 Å was used
with a sample-to-detector distance of 7 m, which gave a q range
Equ ip m en t a n d Gen er a l P r oced u r es. The molecular
weights and the molecular weight distributions of the polymers
were determined on a Hewlett-Packard 1100 series liquid
chromatograph equipped with a Hewlett-Packard 1047A re-
fractive index detector and three J ordi poly(divinylbenzene)
4
3
columns with 10 , 10 , and 500 Å pore sizes. The mobile phase
-
1
was tetrahydrofuran (40 °C and 1 mL min ). The columns
were calibrated using polystyrene standards from Polymer
Laboratories. In addition, SEC analysis was done using a
liquid chromatograph equipped with three Phenomenex Phe-
5
4
3
nogel columns with pore sizes of 10 , 10 , and 10 Å and a
Wyatt DAWN multiangle light scattering detector. The eluent
was tetrahydrofuran, and the dn/dc for polylactide in THF was
-1
-1
of 0.006 < q ) 4πλ sin(Θ/2) < 0.10 Å where Θ is the
scattering angle. Data were collected at 30 °C for the as-
prepared (pressed) amorphous blend and at 200 °C for both
polymer blends. Sample transmissions were measured for both
blends at 30 °C. In addition, data were corrected for back-
ground radiation (blocked beam and empty cell scattering),
sample thickness, detector sensitivity, and incoherent scat-
tering. The data were then converted to an absolute scale
based on the open beam intensity. Excess scattering was
1
7
found to be 0.0576.
The 1H NMR spectra were recorded using CDCl
INOVA 500 MHz NMR spectrometer. The following H NMR
resonances are representative of the polylactide synthe-
on an
3
1
sized in this study: 5.1 (m, -C(O)-C(H )(CH
m, -C(O)-C(H)(CH )-O-), 4.2 (q, CH -CH -O-), 2.1 (s,
O-C(O)-CH ). For the d-PLA-26 there was a small reso-
3
)-O-), 1.6
(
3
3
2
-
1
-
3
observed at q < 0.01 Å for all of the polymer blends as well
as the incoherent blank samples. Although we have no direct
evidence, we believe this is due to foreign matter or voids in
nance at 1.6 ppm due to a small amount (7 mol %) of
nondeuterated methyl groups. The amount of nondeuterated
material found in the d-PLA-26 is inconsistent with the
manufacturers’ claim of 98% D atom content in the deuterated
lactic acid. The reason for this is unknown. The homonuclear
-
1
the samples. Therefore, all data below q ) 0.01 Å were
discarded when performing analysis of the scattering data.
This type of excess scattering has been reported in the
literature previously, resulting in similar treatment to the low
q data.1
1
decoupled H NMR spectra of the hydrogenous polylactides
8,19
were obtained by decoupling using the methyl resonance at
1
.57 ppm. The molecular weights of the polylactides were also
1
determined from the H NMR spectra (PLA-0 ) 55.3 kg/mol;
PLA-28 ) 69.3 kg/mol; d-PLA-26 ) 79.7 kg/mol); however, this
analysis gave anomalously high values for the molecular
weights when compared to the absolute molecular weights
determined by SEC with the multiangle light scattering
detector given in Table 1. These large molecular weight values
are believed to be due to the significant error associated with
the integration of the end-group peak owing to its low
concentration in the sample.
Resu lts
We prepared deuterated polylactide by first synthe-
sizing deuterated lactide from S-d3-lactic acid (deuter-
ated in the methyl position). Formation of the d6-lac-
tide led to some epimerization of the chiral centers;
thus, the deuterated lactide obtained contained ap-
preciable amounts of R stereocenters (presumably in
R,S-meso-lactide, R,R-D-lactide). As a result, the poly-
lactide prepared from this stereochemically impure
lactide gave an amorphous, atactic microstructure with
a slight isotactic tendency, confirmed using H NMR
spectroscopy. This is the only deuterated polylactide
we prepared, so to make a matched hydrogenous mate-
rial we polymerized a D,L mixture of hydrogenous
lactide with approximately the same level of R stereo-
centers and overall molecular weight as the deuterated
Specific optical rotation measurements, [R], were deter-
mined for the lactide and polylactide samples using a J asco
DIP-370 digital polarimeter with a Na lamp at a wavelength
of 589 nm. Solutions of 10 mg/mL in chloroform were analyzed
at room temperature in a 100 mm cell. The optical purity (OP)
was determined using eq 1:
1
16
[
R]
OP )
× 100
(1)
[
R]0
1
sample. By H NMR spectroscopy, this hydrogeneous

Macromolecules, Vol. 37, No. 5, 2004
Polylactides 1859
F igu r e 1. Chemical structures of the polymers used in this
study: (a) deuterated polylactide (d-PLA-26). (b) hydrogenous
polylactides (PLA-28 and PLA-0).
polymer had a similar stereosequence distribution as
the deuterated polymer (see Supporting Information).
In addition, we prepared a purely isotactic polylactide
sample from L-lactide. To avoid transesterification reac-
tions of the deuterated and hydrogenous polymers in
the blends used for SANS, the polylactide chains were
end-capped with an acetate group using acetic anhy-
F igu r e 2. Nonlinear least-squares best fits of coherent
scattering using eq 3 for d-PLA-26/PLA-0 at 200 °C (squares)
(data shifted up by 20× for clarity), d-PLA-26/PLA-28 at 200
°C (circles) (data shifted up by 5× for clarity), and d-PLA-26/
PLA-28 at 30 °C (triangles).
2
0
dride.
Figure 1 shows the structures of both the hydrogenous
and deuterated polymers used in this study, and Table
where ∑ib is the sum of the coherent scattering
lengths per C6 unit for the hydrogenous or deuterated
polymers, vi are the molar volumes per C6 unit at the
temperature of the experiment, φi is the volume frac-
tion of hydrogenous or deuterated polymer, gpi is the
polydisperse Debye function, øij is the Flory-Huggins
interaction parameter between the hydrogenous and
deuterated polymers, Nni and Nwi are the number- and
weight-average degrees of polymerization, respectively,
Rg,ni is the number-average radius of gyration, and
a is the statistical segment length. For this anlaysis,
the repeating unit molar volumes for the hydroge-
nous and deuterated polylactides were assumed to be
1
gives the molecular characteristics of these polymers.
For the SANS experiment, the deuterated polymer (d-
PLA-26) was separately blended with the hydrogenous
polymer of similar stereosequence distribution (PLA-
2
8) and the hydrogenous polymer containing no R-ster-
eocenters (PLA-0). Both blends contained d-PLA-26 at
a volume fraction of 0.2. Although the SANS scattering
intensity is attenuated using such an asymmetric blend,
this volume fraction was chosen in an effort to prevent
macrophase separation of the H/D blends (deuterium
isotope effect).21 The blends were pressed at 200 °C
and subsequently quenched. The scattering experiments
on the d-PLA-26/PLA-28 amorphous polymer blend
were performed at 30 and 200 °C. The d-PLA-26/PLA-0
blend was examined at 200 °C only. The molecular
characteristics of the polymer blends were determined
after the scattering experiment, and some degradation
in the molecular weight of the polymers was noted;
however, the level of degradation in these samples did
not significantly impact the SANS results for these
blends (see Supporting Information).
To determine the statistical segment lengths of the
hydrogenous polylactides in the polymer blends, three
different analysis methods were used. In the first
method, the coherent small-angle scattering from the
single phase polymer blend was modeled on the basis
of de Gennes’ random phase approximation for polydis-
equal (v ) 115.5 cm mol at 30 °C; v ) 130.6 cm3
3
-1
-
1
24
mol at 200 °C). The molecular weights determined
from SEC with the multiangle light scattering detector
were used when analyzing the SANS results (Table 1).
The statistical segment length of the hydrogenous
polylactide in the blends was found by performing a
simultaneous fit of the scattering data using the func-
tions given in eq 3 and floating both the statistical
segment length and ø (see Supporting Information for
details on the fitting procedures and relevant calcula-
tions). The nonlinear least-squares regressions are
shown in Figure 2. When performing the fit for the
d-PLA-26/PLA-28 blend, the statistical segment lengths
of the deuterated and hydrogenous polylactides were
assumed to be equal given that they contain essentially
the same number of R-stereocenters and consequently
should have the same chain dimensions. This value of
the statistical segment length was then fixed for d-PLA-
26 when fitting the d-PLA-26/PLA-0 blend data, and the
statistical segment length of PLA-0 and ø were floated.
The statistical segment lengths determined from the fit
of the data are given in Table 2.
1
9,22,23
perse components given by
2
ib
jb
∑
∑
Icoh(q) )
-
S(q)
(
)
v
v
j
i
1
1
1
)
+
- 2øij
S(q) N φ g (x ) N φ g (x )
i
i
pi ni
j j pj nj
2
xni -ki
The second analysis method makes use of the inter-
2
g (x ) )
x - 1 + 1 +
mediate q range approximation to eq 3. In this case, the
pi ni
ni
(x
)
[
(
)
]
ni
ki
-1
2
slope of S(q) vs q yields the statistical segment length
(see Supporting Information for details). In addition, the
Flory-Huggins interaction parameter can also be de-
termined from the y-intercept. These plots are shown
in Figure 3. To calculate the statistical segment length
of the hydrogenous polymers from the slope, the follow-
-
1
N
2
2
wi
x ) q R
ki )
- 1
ni
g,ni
(
)
Nni
2
N a
2
ni
Rg,ni
)
(3)
25
6
ing relationships were used:

1
860 Anderson and Hillmyer
Macromolecules, Vol. 37, No. 5, 2004
2
F igu r e 3. Plots of v/S(q) vs q from scattering data for d-PLA-
2
F igu r e 4. Kratky plots for scattering data of d-PLA-26/PLA-0
blend at 200 °C (squares) (data shifted up by +0.003 for clarity,
plateau value ) 0.0030), d-PLA-26/PLA-28 blend at 200 °C
(circles) (data shifted up by +0.001 for clarity, plateau value
) 0.0033), and d-PLA-26/PLA-28 blend at 30 °C (triangles)
6/PLA-0 blend at 200 °C (squares) (data shifted up by +1.0
for clarity), d-PLA-26/PLA-28 blend at 200 °C (circles) (data
shifted up by +0.5 for clarity), and d-PLA-26/PLA-28 blend at
3
0 °C (triangles).
(plateau value ) 0.0030).
2
2
S(q)-
1
)
1
+
1
- 2ø +
q a
1
(4)
(5)
ij
using eq 6, since we assumed that the statistical
segment lengths of the hydrogenous and deuterated
polymer were equal. The statistical segment length of
the hydrogenous polylactide in the d-PLA-26/PLA-0
blend was then determined using eq 5 and the statistical
segment length of d-PLA-26 from the d-PLA-26/PLA-
N φd N φh
12 φ φ
nd
nh
d h
2
2
ah
ad
2
a )
+
vhφh vdφd
2
8 blend. The statistical segment length values from
In the case of the d-PLA-26/PLA-28 blend the statisti-
this analysis are given in Table 2.
cal segment lengths of the deuterated and hydrogenous
polymers were assumed to be equal when performing
the analysis, so the statistical segment length was
calculated directly from the slope. For the d-PLA-26/
PLA-0 blend, to determine the statistical segment
length of the hydrogenous polymer, PLA-0, eq 5 was
used. The value for the statistical segment length of the
deuterated polymer, ad, was taken as the statistical
segment length determined for the d-PLA-26/PLA-28
blend. The statistical segment lengths of the hydrog-
enous polymers from this analysis method are given in
Table 2.
Discu ssion
All three analysis methods gave values for the sta-
tistical segment lengths of the hydrogenous polymers
that were in agreement with each other (Table 2). For
the matched H/D blend, d-PLA-26/PLA-28, the statisti-
cal segment length at 30 °C is 10.0 ( 0.2 Å and at 200
°C it is 8.9 ( 0.2 Å. For the d-PLA-26/PLA-0 blend,
where the hydrogenous polymer contains no stereode-
fects, the statistical segment length of the PLA-0 is
9.9 ( 0.4 Å at 200 °C. In addition, the average values
and standard deviations for the Flory-Huggins interac-
tion parameter between the hydrogenous and deuter-
ated polymer in the two different blends are given in
Table 2.
Finally, the statistical segment lengths were deter-
2
mined from the Kratky plateau. In the limit that the q
2
term in eq 4 dominates, a plot of q I(q) vs q comes to a
constant value. The statistical segment length is then
determined from the following relationship:25
The ratio of the unperturbed chain dimensions to the
molecular weight was calculated for both PLA-28 and
PLA-0 using the following relationship:
2
6
2
ib
jb
∑ ∑
-
1
2φ φd
h
(
)
v
v
2
2
2
i
j
〈R 〉
6〈R 〉
2
2
0
g
0
a N
a
a )
(6)
2
)
)
)
(7)
(
q I(q))
M
M
m N m0
plateau
0
Kratky plots for the polylactide blends at 30 and at
00 °C are given in Figure 4 (see Supporting Informa-
tion for details on Kratky plateau determination). The
value for the statistical segment length of the hydrog-
enous polymer in the d-PLA-26/PLA-28 blend was found
where m0 is the molecular weight of the C6 repeat unit
2
(144.217 g/mol). These values are also given in Table 2.
2
By directly comparing the value of 〈R 〉0/ M from this
study to other values reported in the literature for poly-
2
lactide, we see some interesting differences. A 〈R 〉0/ M
Ta ble 2. P LA Sta tistica l Segm en t Len gth s a n d F lor y-Hu ggin s In ter a ction P a r a m eter sa
a (Å)
R 〉0/M (Å mol g-1)
2
2
ø (10-3
b
〈
)
polymer
RPA (eq 3)
intermediate q (eq 4)
Kratky (eq 6)
averageb
averageb
average
PLA-0 (200 °C)
PLA-28 (30 °C)
PLA-28 (200 °C)
10.3
10.2
9.0
9.6
9.8
8.7
9.8
10.1
9.1
9.9 ( 0.4
10.0 ( 0.2
8.9 ( 0.2
0.681 ( 0.050
0.699 ( 0.029
0.554 ( 0.026
7.1 ( 2.0
5.7 ( 2.0
3.7 ( 1.9
a
Flory-Huggins interaction parameters between hydrogenous polymer and d-PLA-26. b (1 standard deviation.

Macromolecules, Vol. 37, No. 5, 2004
Polylactides 1861
2
of 0.380 Å mol g-1 was reported for a stereoregular
2
a
2
7
C )
(9)
poly(L-lactide) by Fetters et al. based on experiments
∞
2
performed by Flory et al.28 This value for 〈R 〉0/ M is
2
n l
v
significantly lower than the value determined for the
2
-1
In this analysis, we have chosen nv ) 6 and l ) 1.41 Å
11
PLA-0 at 200 °C (0.685 Å mol g ). The light scattering
experiments by Flory et al. were performed at 85 °C
rather than 200 °C; however, on the basis of the trend
for polylactide.
For PLA-28 the average value of the characteristic
ratio at 30 °C is 8.4 ( 0.3 and at 200 °C it is 6.7 ( 0.3.
For PLA-0 the average value is 8.2 ( 0.6 at 200 °C. In
comparison, Pennings et al. found the characteristic
ratio of a polylactide containing 29 mol % D-lactide to
2
for PLA-28, the 〈R 〉0/ M for PLA-0 should be larger at
lower temperatures. The most likely reason for the
large discrepancy between the values is that the experi-
ments of Flory were done in bromobenzene, and the
quality of this solvent for light scattering measure-
1
1
be 9.4 at 25 °C. In addition, Hsu et al. found the
characteristic ratio at 25 °C to be 8.5 for a polylactide
1
2
ments of polylactide has been brought into question.
1
3
containing 20% D-stereocenters. These values are in
general agreement with the characteristic ratio of 8.4
determined in this study for the PLA-28 at 30 °C. To
be able to compare the characteristic ratio of PLA-0 to
other values found in the literature, we also estimated
the temperature dependence of the unperturbed chain
dimensions for PLA-28 by calculating the temperature
2
-1
More recently, a value of 0.77 Å mol g was reported
by Adachi et al. for polylactide with a D/L ratio of 1/3
2
9
(
ca. PLA-25). This value agrees quite well with the
2
-1
value determined for PLA-28 at 30 °C (0.699 Å mol g ).
In the case of Adachi et al., light scattering measure-
ments were taken at room temperature, and benzene
was used as the solvent, which is a good solvent for
polylactide. This could account for their slightly larger
chain dimensions than those that we determined in the
melt.
2
coefficient, κ ) d ln Rg / dT. Using the statistical
segment lengths at 30 and 200 °C, κ was found to be
-1
-
0.00141 K . If we assume that the PLA-0 has the
same temperature dependence as PLA-28, the radius
of gyration and hence the statistical segment length of
PLA-0 at 30 °C can be determined. Using this assump-
tion, the statistical segment length of PLA-0 at 30 °C
was calculated to be 11.2 Å. From eq 9, the character-
istic ratio at room temperature for PLA-0 is 10.5. In
comparison, a characteristic ratio of 12.4 was reported
for a PLA-0 at 25 °C by Hsu et al., and a value of 11.8
was reported by Pennings et al. Our value is signifi-
cantly lower, indicating that the assumption of similar
temperature dependence for PLA-28 and PLA-0 may not
be valid. However, in the case of Pennings et al., rather
than making a direct measurement they extrapolated
to zero D content, and Hsu et al. used a solvent mixture
in their light scattering experiment that may have
caused a larger coil dimension than would be deter-
mined in the melt. Nevertheless, both Pennings et al.
and Hsu et al. found that the characteristic ratio of
polylactide decreased as the R-content increased, which
is consistent with our results at 200 °C.
To determine whether the value we found for the
statistical segment length of polylactide was reasonable
compared to other common polymers, we calculated the
statistical segment lengths based on a commonly used
-
22
polyolefin reference volume (vref ) (1.084 × 10 ) exp-
-
4
3
[
6.85 × 10 (T - 296)], with vref in cm and T in absolute
19,30
temperature).
The statistical segment length of
polylactide was determined by first calculating the effec-
tive repeat unit molecular weight for polylactide using
the polyolefin reference volume and the density of
polylactide. With this value the statistical segment
length based on the polyolefin reference volume was
2
found by means of eq 7 and the 〈R 〉0/ M value in Table
2
(
. At 30 °C the statistical segment length for PLA-28
7.6 Å)31 is in the range of the statistical segment
lengths for other common polymers (based on the
previously defined polyolefin reference volume). For
instance, the statistical segment length of polyethylene
2
7
3
at 25 °C is 8.9 Å, for 1,4-polybutadiene it is 7.2 Å, for
3
,32
Finally, we used our value of the statistical segment
length to reevaluate the conformational asymmetry
parameter, ꢀ, which we previously reported for the poly-
(ethylene-alt-propylene)-b-polylactide (PEP-PLA) sys-
polystyrene it is 5.5 Å,
and for poly(cyclohexylethyl-
ene) it is 4.6 Å.33 Furthermore, the ratio of the unper-
turbed chain dimensions to the molecular weight,
2
2
-1
〈
R 〉0/ M ) 0.699 Å mol g , is also intermediate when
compared to the values reported for these polymers (PE
3
5
tem. The conformational asymmetry parameter is
3
6
calculated from the following equation:
2
7
3
3,32
)
1.42, PBD ) 0.876, PS ) 0.434,
PCHE ) 0.337
2
-1 33
Å mol g
).
-
1
2
2
ꢀ
) (νA aA /νB-1a
)
(10)
B
Using the experimentally determined statistical seg-
ment length, the characteristic ratio, C∞, of polylactide
was also calculated. This parameter is of interest
because it gives an indication as to the intrinsic flex-
ibility of the polylactide chain. The characteristic ratio
This parameter accounts for the differences in the molar
volumes and statistical segment lengths between the
two polymers. Using the characteristic ratio reported
11
by Pennings et al., a statistical segment length was
3
4
was determined using the following equation:
determined, and the ꢀ value between PLA and PEP was
calculated to be 1.39. However, on the basis of the
experimentally determined morphology diagram for
these block copolymers, we expected the ꢀ value to be
closer to 1.0, since the phase diagram appears sym-
metric around fPLA ) 0.5. We can now recalculate this
parameter on the basis of our experimental data for the
polylactide statistical segment length. Using the poly-
olefin reference volume previously defined and making
use of κ to determine the chain dimensions of PLA-28
at 140 °C, the statistical segment length of PLA-28 is
7.0 Å and for PEP the statistical segment length is 6.8
2
〈
R 〉
0
C )
(8)
∞
2
nl
2
where 〈R 〉0 is the mean-square end-to-end distance of
a polymer coil and n is the number of statistical skeletal
units of length l in the polymer chain. Since n ) Nnv,
where nv is the number of real or virtual bonds per
repeat unit, and making use of relationships given in
eq 7, this equation can then be reduced to

1
862 Anderson and Hillmyer
Macromolecules, Vol. 37, No. 5, 2004
3
Å. From these values the conformational asymmetry
parameter between PLA and PEP is now 1.06. This
value is in good agreement with the symmetric phase
diagram that was observed.
(6) Bates, F. S.; Schulz, M. F.; Rosedale, J . H.; Almdal, K.
Macromolecules 1992, 25, 5547-5550.
7) Bates, F. S.; Schulz, M. F.; Khandpur, A. K.; Forster, S.;
(
Rosedale, J . H.; Almdal, K.; Mortensen, K. Faraday Discuss.
1
994, 98, 7-18.
We also experimentally obtained the morphology
(8) Almdal, K.; Koppi, K. A.; Bates, F. S.; Mortensen, K.
diagram for PS-PLA block copolymers.37 Polystyrene
Macromolecules 1992, 25, 1743-1751.
9) Schnell, R.; Stamm, M.; Creton, C. Macromolecules 1998, 31,
(
and polylactide apparently have a higher degree of
asymmetry, resulting in a shift of the experimental
morphology diagram from the idealized symmetric case.
The statistical segment length for PLA-28 at 110 °C
2
284-2292.
(
(
(
10) Cole, P. J .; Cook, R. F.; Macosko, C. W. Macromolecules 2003,
36, 2808-2815.
11) J oziasse, C. A. P.; Veenstra, H.; Grijpma, D. W.; Pennings,
A. J . Macromol. Chem. Phys. 1996, 197, 2219-2229.
12) Grijpma, D. W.; Penning, J . P.; Pennings, A. J . Colloid Polym.
Sci. 1994, 272, 1068-1081.
(
using κ) is 7.1 Å, and for PS it is 5.5 Å based on the
polyolefin reference volume at this temperature. From
these statistical segment lengths, the asymmetry pa-
rameter is 1.67. Fortuitously, Whitmore et al. calculated
what the position of the cylinder to lamellar phase
boundary would be for a block copolymer with a øN
value of 80 and an asymmetry parameter of 1.67. They
predict that the boundary will shift from 0.33 to 0.40.
Experimentally, we determined that the cylinder to
lamellar phase boundary in the intermediate to strong
segregation limit is between 0.43 and 0.44, which is in
qualitative agreement with their prediction.
(13) Kang, S.; Zhang, G.; Aou, K.; Hsu, S. L.; Stidham, H. D.;
Yang, X. J . Chem. Phys. 2003, 118, 3430-3436.
14) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R.
(
K.; Timmers, F. J . Organometallics 1996, 15, 1518-1520.
(
15) Wang,Y.;Hillmyer,M.A.Macromolecules 2000,33,7395-7403.
3
6
(16) Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Kolstad, J . J .;
Lindgren, T. A.; Doscotch, M. A.; Siepmann, J . I.; Munson,
E. J . Macromolecules 1997, 30, 2422-2428.
(
17) Williams, C. K.; Breyfogle, L. E.; Choi, S. K.; Nam, W.; Young,
V. G., J r.; Hillmyer, M. A.; Tolman, W. B. J . Am. Chem. Soc.
2
003, 125, 11350-11359.
(18) Bates, F. S.; Fetters, L. J .; Wignall, G. D. Macromolecules
1
988, 21, 1086-1094.
Con clu sion s
(19) Weimann, P. A.; J ones, T. D.; Hillmyer, M. A.; Bates, F. S.;
Londono, J . D.; Melnichenko, Y.; Wignall, G. D.; Almadal,
K. Macromolecules 1997, 30, 3650-3657.
The melt chain dimensions of two polylactide samples
were investigated using small-angle neutron scattering.
The statistical segment lengths were determined to be
0.0 ( 0.2 Å for PLA-28 at 30 °C, 8.9 ( 0.2 Å for PLA-
8 at 200 °C, and 9.9 ( 0.4 Å for PLA-0 at 200 °C based
on a C6 repeating unit. The chain dimensions were
found to decrease with both increasing temperature and
increasing D content. On the basis of its statistical
segment length, the characteristic ratio of PLA-28 at
(20) J amshidi, K.; Hyon, S.-H.; Ikada, Y. Polymer 1988, 29, 2229-
2
234. The procedure for end-capping is not the same as that
used in this study; however, the effect that end-capping has
on the transesterification of polylactide is discussed. In
addition, an experiment was performed where end-capped
polylactide was placed in a vacuum oven at 200 °C for 2 h
and no broadening of the polydispersity was observed,
indicating insignificant transesterification.
1
2
(
21) Bates, F. S.; Wignall, G. D.; Koehler, W. C. Phys. Rev. Lett.
1
985, 55, 2425-2428.
3
0 °C was calculated to be 8.4, which is in general
(22) de Gennes, P. G. Scaling Concepts in Polymer Physics, Cornell
agreement with other values reported in the literature
for polylactides with similar D contents. However, the
characteristic ratio of PLA-0 at 30 °C was found to be
somewhat smaller than other reported values for iso-
tactic PLLA. Finally, we calculated the conformational
asymmetry parameters for both PEP-PLA and PS-
PLA block copolymer systems based on the experimen-
tally determined chain dimensions for polylactide and
found them to be in agreement with the experimentally
determined morphology diagrams.
University Press: Ithaca, NY, 1979.
23) Higgins, J .; Benoit, H. Polymers and Neutron Scattering;
Oxford University Press: Oxford, UK, 1994.
(
(
24) Calculated from the density of polylactide and the repeat unit
molecular weight. The density at room temperature was
taken to be 1.248 g/cm3 (see: Grijpma, D. W.; Penning, J . P.;
Pennings, A. J . Colloid Polym. Sci. 1994, 272, 1068-1081).
3
The density at 200 °C was taken to be 1.104 g/cm (see:
Witzke, D. R.; Narayan, R.; Kolstad, J . J . Macromolecules
1
997, 30, 7075-7085).
(25) J ones, T. D.; Chaffin, K. A.; Bates, F. S.; Annis, B. K.;
Hagaman, M. K.; Kim, M. H.; Wignall, G. D.; Fan, W.;
Waymouth, R. Macromolecules 2002, 35, 5061-5068.
(
26) Flory, P. J . Statistical Mechanics of Chain Molecules; Inter-
science: New York, 1969.
Ack n ow led gm en t. We acknowledge the support of
the National Institute of Standards and Technology,
U.S. Department of Commerce, in providing the neutron
research facilities used in this work. The authors also
thank Professors Timothy Lodge and Frank Bates for
helpful input during the preparation of this manuscript.
The David and Lucille Packard Foundation is acknowl-
edged for financial support of this work.
(
27) Fetters, L. J .; Lohse, D. J .; Graessley, W. W. J . Polym. Sci.,
Part B: Polym. Phys. 1999, 37, 1023-1033.
(28) Tonelli, A. E.; Flory, P. J . Macromolecules 1969, 2, 225-227.
(
29) Ren, J .; Urakawa, O.; Adachi, K. Macromolecules 2003, 36,
2
10-219.
(
30) Owens, J . N.; Gancarz, I. S.; Koberstein, J . T.; Russell, T. P.
Macromolecules 1989, 22, 3380-3387.
(31) The value for the statistical segment length of PLA-28 at 200
°
C based on the polyolefin reference volume is 6.7 Å and for
Su p p or tin g In for m a tion Ava ila ble: Size exclusion chro-
matographs of all polymers used, coherent scattering data,
intermediate q-range analysis, Kratky plateau determination,
effect of sample degradation on data analysis, and comparison
of stereosequence distribution in polylactide. This material is
available free of charge via the Internet at http://pubs.acs.org.
PLA-0 at 200 °C it is 7.5 Å.
(32) Gehlsen, M. D.; Weimann, P. A.; Bates, F. S.; Harville, S.;
Mays, J . W.; Wignall, G. D. J . Polym. Sci., Part B: Polym.
Phys. 1995, 33, 1527-1536.
(
33) Krishnamoorti, R.; Graessley, W. W.; Zirkel, A.; Richter, D.;
Hadjichristidis, N.; Fetters, L. J .; Lohse, D. J . J . Polym. Sci.,
Part B: Polym. Phys. 2002, 40, 1768-1776.
(
34) Mattice, W. L.; Suter, U. W. Conformational Theory of Large
Molecules. The Rotational Isomeric State Model in Macro-
molecular Systems; Wiley: New York, 1994.
Refer en ces a n d Notes
(
1) Vert, M.; Li, S. M.; Spenlehauer, G.; Guerin, P. J . Mater.
Sci.: Mater. Med. 1992, 3, 432-446.
(
(
(
35) Schimdt, S. C.; Hillmyer, M. A. J . Polym. Sci., Part B: Polym.
Phys. 2002, 40, 2364-2376.
36) Vavasour, J . D.; Whitmore, M. D. Macromolecules 1993, 26,
(2) Tullo, A. Chem. Eng. News 2000, 78, 13.
(
3) Fetters, L. J .; Lohse, D. J .; Richter, D.; Witten, T. A.; Zirkel,
A. Macromolecules 1994, 27, 4639-4647.
7
070-7075.
37) Zalusky, A. S.; Olayo-Valles, R.; Wolf, J . H.; Hillmyer, M. A.
(
4) Wu, S. Polym. Int. 1992, 29, 229-247.
J . Am. Chem. Soc. 2002, 124, 12761-12773.
(
5) Bates, F. S.; Fredrickson, G. H. Macromolecules 1994, 27,
1
065-1067.
MA0357523