Detecting structural polydispersity in branched polybutadienes

Article abstract of DOI:10.1021/ma101803h

The structural details of a set of highly entangled H-shaped polybutadienes (PBDs) prepared by anionic polymerization were examined in detail by three reputable laboratories using size exclusion chromatography (SEC) and temperature gradient interaction chromatography (TGIC). While SEC data indicated that samples having the desired structures (i.e., nearly monodisperse H-shaped polymer) had been produced, additional SEC data from other laboratories showed that the samples were structurally more complex than originally thought. TGIC data revealed that while the samples did not contain high molecular weight byproducts, they did contain low molecular weight byproducts. To discern these structural details of the branched PBDs, small amounts of sample were fractionated by TGIC. By combining knowledge of the polymerization process with the TGIC data of fractionated samples, it was possible to work out the detailed compositions of the samples and the branching structures of each component.

Full text of DOI:10.1021/ma101803h

208 Macromolecules 2011, 44, 208–214
DOI: 10.1021/ma101803h
Detecting Structural Polydispersity in Branched Polybutadienes
Si Wan Li,† Heon E. Park, John M. Dealy,* and Milan Maric
Department of Chemical Engineering, McGill University, Montreal, QC, Canada H3A 2B2
Hyojoon Lee,† Kyuhyun Im,‡ Heungyeal Choi, and Taihyun Chang
Department of Chemistry and Division of Advanced Materials Science, Pohang University of Science and
Technology (POSTECH), Pohang 790-784, Korea
M. Shahinur Rahman and Jimmy Mays
Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37996, United States.
†Equal contribution. ‡Current address: Samsung Advanced Institute of Technology, Korea
Received September 27, 2010; Revised Manuscript Received November 9, 2010
ABSTRACT: The structural details of a set of highly entangled H-shaped polybutadienes (PBDs) prepared
by anionic polymerization were examined in detail by three reputable laboratories using size exclusion
chromatography (SEC) and temperature gradient interaction chromatography (TGIC). While SEC data
indicated that samples having the desired structures (i.e., nearly monodisperse H-shaped polymer) had been
produced, additional SEC data from other laboratories showed that the samples were structurally more
complex than originally thought. TGIC data revealed that while the samples did not contain high molecular
weight byproducts, they did contain low molecular weight byproducts. To discern these structural details of
the branched PBDs, small amounts of sample were fractionated by TGIC. By combining knowledge of the
polymerization process with the TGIC data of fractionated samples, it was possible to work out the detailed
compositions of the samples and the branching structures of each component.
Introduction
had been obviated, the samples were considerably more complex
than originally thought. This was discouraging from the point of
The polymers described here were synthesized for a study of
the effects of long-chain branching and polydispersity on the
rheological behavior of entangled polymers, the results of which
were to be used for the evaluation of molecular models of
viscoelasticity. The desired structural details (arm and crossbar
lengths) of these polymers were estimated prior to synthesis so
that the polymers would be highly entangled but with terminal
behavior accessible using available rheometric techniques. In
order to facilitate the evaluation of molecular models as pre-
dictors of rheological behavior, a series of molecularly uniform
H-shaped polybutadienes (H-PBDs) with a high concentration of
1,4-addition were required, and polydispersity was to be generated
by blending. In this way the detailed structures of the samples
would be known. H-PBD was chosen because H is the simplest
structure containing two branch points, and PBD has a relatively
low molecular weight between entanglements (Me = 2 kg/mol1)
among polymers that can be synthesized by anionic polymeriza-
tion. The preference for high 1,4-addition is due to the fact that
1,2-addition results in a vinyl side group (a short branch), and the
doublebond onthe vinyl side group isvulnerable to cross-linking.
A novel anionic polymerization method2 was used to prepare
samples having the desired characteristics.
view of the intended rheological study, but if the detailed com-
position of the samples could somehow be discerned, they could
still be of use. We explain below how this was accomplished. The
H-PBD samples are named according to the intended molecular
weights of the arms and crossbars; e.g., “HA12B40” denotes a
sample with four arms of equal molar mass of 12 kg/mol and a
crossbar with a molar mass of 40 kg/mol.
Experimental Section
Synthesis of H-PBDs. A novel strategy for preparing H-PBDs
by anionic polymerization was developed by Rahman et al.2 in
order to minimize the formation of high molecular weight by-
products like those reported by Perny et al.3 It is based on the
idea that the coupling of two side arms with trichloromethylsi-
lane prior to any reaction with the crossbar should significantly
reduce the production of high molecular weight byproducts.4
This method requires the use of high-vacuum and ampouliza-
tion techniques.5 Details of the synthesis procedures and dilute
solution properties have been described previously.2 In general,
the synthesis involves a difunctional diphenylethylene-based
coupling agent 4-(dichloromethylsilyl)diphenylethylene (DCM-
SDPE), sec-BuLi, butadiene, and benzene. As shown in Figure 1,
the synthesis consists of five main steps, and this reaction scheme
will be used to infer the molecular structures of the resulting
polymers in a later section.
While size exclusion chromatography (SEC) data indicated
that this synthesis method had produced the desired materials,
later analyses using SEC with different chromatographic detectors
and temperature gradient interaction chromatography (TGIC)
revealed that while the production of high molecular byproducts
Determination of Isomeric Composition. To elucidate micro-
structures, 1H nuclear magnetic resonance (NMR) was used. As
the sensitivity and resolution of NMR spectra increase with
magnetic field strength, the characteristic peaks for cis and trans
*Corresponding author. E-mail: john.dealy@mcgill.ca.
pubs.acs.org/Macromolecules
Published on Web 12/20/2010
r 2010 American Chemical Society

Article
Macromolecules, Vol. 44, No. 2, 2011 209
Alliance 2695 separations module and a Viscotek TDA model
302 consisting of a RI detector at a wavelength of 660 nm, a
4-capillary viscometer, and a 7ꢀ light scattering detector (LS) in
3 mW power at a wavelength of 670 nm were used for these
samples. Two Polymer Laboratories Polypore columns (300 ꢀ
7.5 mm, 5 μm) were used, and the detectors were temperature
controlled in the same oven as the Viscotek TDA. Measure-
ments were done at 40 ꢀC with THF as the mobile phase at a flow
rate of 1.0 mL/min; dn/dc was allowed to float between 0.133
and 0.138 mL/g. Each sample was injected three times, and the
data were mathematically averaged before processing. The third
SEC characterization (SEC-3) was performed on the final
product, using a Wyatt miniDAWN LS detector in 60 mW
power at a wavelength of 658 nm, and a Shodex RI-101 RI
detector uses white light as the light source with two PLgel
mixed-C columns (300ꢀ7.5 mm, 5 μm). This characterization
was performed at 40 ꢀC with THF as the mobile phase at a flow
rate of 0.8 mL/min; dn/dc was 0.128 mL/g.
Determination of Molecular Weights by TGIC. Unlike SEC,
TGIC6,7 is an interaction chromatographic technique in which
the separation is driven by enthalpic interactions between the
sample and the stationary phase. The interaction strength is
controlled by varying the column temperature7,8 and eluent com-
position,9,10 and the molecular weight resolution is little affected
by chain architecture. Therefore, TGIC is believed to be more
sensitive to the chemical nature of a polymer7-9 and to have a
higher resolution than SEC.3,11-13 The TGIC separations were
carried out using a standard high-performance liquid chroma-
tography (HPLC) system equipped with a C18 bonded silica
˚
column (Phenomenex, Kromasil, 300 A pore, 150ꢀ4.6 mm, 5 μm
particle size); the mobile phase was 1,4-dioxane at a flow rate of
0.5 mL/min, and dn/dc was 0.095 mL/g. The system was equip-
ped on line with a Wyatt miniDAWN LS detector and a Shodex
RI-101 RI detector. The column temperature profile is indicated
by solid lines in Figure 5 which were adjusted in a way to cover
the range of molar mass of each sample.
Results and Discussion
Isomeric Composition. Characteristic peaks determined by
1H NMR for the cis/trans 1,4 and 1,2 vinyl configurations
were identified,14,15 and the corresponding areas under the
peaks were calculated by the operating software. As proposed
by Tanaka et al.,15 the mole fractions of 1,2 and 1,4 units can
be calculated by eqs 1 and 2 where the bracket denotes the
mole fraction of the respective isomeric unit indicated and I(δ)
is the relative integrated intensity of the signal with chemical
shift δ. The ratio of the integrated intensity of the signal of
the trans type protons to that of the cis ones in the 2 ppm
(aliphatic) or 5 ppm (olefinic) region indicates the ratio of
trans to cis in the polymer.14,16 As noted by Santee et al., the
aliphatic region is not as distinct as the olefinic region due to
spin-spin coupling effects.14 Thus, the trans-to-cis ratio was
calculated by eq 3 using the olefinic region.
Figure 1. Synthesis steps for the anionic polymerization of H-PBD.
configurations can be resolved by use of 500 MHz field. Our
NMR studies were carried out on a 500 MHz Varian unity spec-
trometer at room temperature. A sample of 10 mg was dissolved
in 0.7 mL of deuterated chloroform (Cambridge Isotope Labo-
ratories Inc.) in 5 mm 508 Up NMR tubes. Samples were inject-
ed into the probe, shimmed, and scanned 32 times.
Determination of Molecular Weights by SEC. Molecular
weights of the H-PBDs were determined in three laboratories
under the supervision of qualified scientists. The first character-
ization (SEC-1) was performed using SEC equipped with two-
angle laser light scattering (TALLS) in 30 mW power at a
wavelength of 685 nm, a refractive index (RI) detector at a
wavelength of 680 nm, and a Viscotek differential viscometer to
determine the precursors and final products. The columns were
Waters Ultrastyragel HR series, HR-2, HR-4, HR-5E, and HR-
Ið1:3Þ
2½1, 2ꢁ
¼
ð1Þ
Ið2:0- 2:1Þ
½1, 2ꢁ þ 4½1, 4ꢁ
3
4
5
˚
6E, with pore sizes 10 , 10 , and 10 A. Measurements were done
at 40 ꢀC with tetrahydrofuran (THF) as the mobile phase at a
flow rate of 1.0 mL/min. An average value of the refractive index
increment (dn/dc) 0.130 mL/g was used in determining the molar
mass of the intermediate products. The characterization was
performed in parallel with the synthesis2 and thus provided
information on the product of each stage of the polymerization
(i.e., step 2 and step 4 in Figure 1) in addition to the final
product. “Final product” refers to the result of step 5 in Figure 1
after fractionation with toluene/methanol as the solvent/non-
solvent pair.2 The second SEC characterization (SEC-2) was
performed on the final product using a SEC system of Waters
½1, 2ꢁ þ ½1, 4ꢁ ¼ 1
ð2Þ
ð3Þ
Ið5:4Þ
Ið5:3Þ
:
Ið5:4Þ þ Ið5:3Þ Ið5:4Þ þ Ið5:3Þ
The mole fractions of cis 1,4 and trans 1,4 are then
calculated by multiplying the above ratio by the overall 1,4
content given by eq 2. Results of isomeric content are sum-
marized in Table 1, which shows that all samples had high
levels of 1,4-addition (g94%) as had been desired.

210 Macromolecules, Vol. 44, No. 2, 2011
Li et al.
Characterization by SEC. It is well-known that the molar
mass averages and molar mass distributions (MMDs) deter-
mined by SEC can vary significantly from one laboratory to
another, and a comparison of data from 15 sources for a
branched polyethylene revealed an interlaboratory variabil-
ity for Mw of about 18% regardless of the detectors used.17
Possible causes of these variations in molar mass average and
MMD among different laboratories have been examined in
detail by Trathnigg.18 And in our study, the molar mass
averages and MMDs determined in three laboratories were
quite different from each other. The average molar masses
determined by SEC are listed in Table 2. An “arm” is the
product of step 2 in Figure 1, “1/2-H” refers to that of step 4
in Figure 1, and “final product” refers to that of step 5, al-
though this may contain byproducts other than an H mole-
cule, even after fractionation using toluene/methanol as the
solvent/nonsolvent pair.2
by SEC-3 were broader than those found by SEC-1, likely
due to better separation in SEC-3. For HA30B40 and
HA40B40, the molecular weights found by SEC-3 were com-
parable to those found by SEC-1. However, lower molecular
weights were indicated for the other samples using SEC-3
Since the molar mass characterization by SEC-1 was
performed in parallel with the synthesis,2 it is the only one
that provides information on precursors at various stages of
polymerization. In other words, it contains molar mass
1
information on arms and /2-Hs. Complete elution curves
at various stages of polymerizations can be found in a pre-
vious publication.2 Generally, the SEC-1 elution profiles of
final products, shown in Figure 2, had single peaks and narrow
distributions. Moreover, a previous study on the branching
parameter gfor our samples showed very good to fair agreement
with literature values on H-shaped polystyrene.2
However, the RI elution profiles for the same final pro-
ducts characterized by SEC-2, as shown in Figure 3, had
double peaks except for HA12B40. The molecular weights at
the two peaks are in a ratio of about 1:2, indicating the
Figure 2. SEC-1 (LS response at 15ꢀ) elution profiles on final products;
samples were analyzed in THF at 40 ꢀC at a flow rate of 1.0 mL/min.
1
presence of /2-H molecules, which are equivalent to asym-
metric 3-arm stars, implying that most of the samples were
actually mixtures of H molecules and asymmetric 3-arm stars
(1/2-H). The weight fraction calculated from the area under
the peaks after deconvolution of the RI signal by a Gaussian
1
distribution indicates that the amount of /2-H material is
significant, ranging from 0 to 47 wt %, as shown in Table 2.
We note that the MMD obtained from SEC-2 was the
broadest of the distributions from the three laboratories.
The elution profiles obtained by SEC-3 shown in Figure 4
also indicate the presence of low molecular weight bypro-
ducts in most of the samples. In most cases, MMDs obtained
Table 1. Microstructural Characteristic of H-PBDs
microstructure
sample code
cis 1,4 (%)
trans 1,4 (%)
vinyl 1,2 (%)
HA12B40
HA30B40
HA40B40
HA12B100
51
52
52
53
43
42
42
41
6
6
6
6
Figure 3. SEC-2 (RI response) elution profiles on final products;
samples were analyzed in THF at 40 ꢀC at a flow rate of 1.0 mL/min.
Table 2. Summary of SEC Characterizations in Three Laboratories (Molar Mass in kg/mol)
SEC-1
1/2-H
SEC-2
final product
wt %
SEC-3
arm
final product
final product
a
Mp
sample code
Mn
PDI
Mn
PDI
Mn
PDI
Mn
PDI
PDI
HA12B40
HA30B40
10.6
29.6
1.01
1.01
41
80.2
1.11
1.05
82.3
161
1.03
1.06
70.8
93.3
100
8
1.25
1.32
70.6
164
N.A.b
N.A.
174
92
23
77
47
53
HA40B40
41.6
15.3
1.01
1.03
106
82.4
1.06
1.07
212
158
1.05
1.04
109.6
218.8
81.3
1.22
1.28
216
N.A.
N.A.
HA12B100
114
196
170
a Peak molar mass. b N.A.: not available.

Article
Macromolecules, Vol. 44, No. 2, 2011 211
samples were structurally more complex than originally
thought.
Characterization by TGIC. The SEC profiles shown in
Figures 2-4 suggested that some of the samples were binary
blends (H and 3-arm stars), but the TGIC chromatograms
shown in Figure 5 indicated that all the samples were mix-
tures of three or more components, even those fractionated
with toluene/methanol as the solvent/nonsolvent pair.2 All
samples contained multiple, low molecular weight bypro-
ducts as revealed by the molecular weight corresponding
to each peak, i.e., Mp shown in Table 3. Although TGIC
showed higher resolution than SEC by resolving more peaks,
further information is needed to identify the branching
structures of these extra components. It is noted that the
column temperature profile of HA12B40 is different from
the other samples. A lower starting temperature is used on
HA12B40 because the sample contains relatively low molec-
ular weight components (e.g., peaks 1, 2, and 3 with Mp <
45 kg/mol). If the starting temperature of HA12B40 is chosen
at 23 ꢀC, i.e., same starting temperature as the other samples,
these low molecular weight components cannot be resolved.
On the other hand, an increasing temperature is needed to
elute any high molecular weight components beyond peak 6
of HA12B40.
Figure 4. SEC-3 (RI and LS response at 90ꢀ) elution profiles on final
products; samples were analyzed in THF at 40 ꢀC at a flow rate of
0.8 mL/min.
Inference of Structures. Currently our ability to identify
branching structures is quite limited.19 A commonly used
characterizing parameter is the branching contraction
2
2
2
parameter g=ÆRg æb/ÆRg æl, where ÆRg æ is the mean-square
radius of gyration and the subscripts b and l denote
branched and linear polymer of identical molar mass,
respectively. This parameter reflects the effect of branching
2
on ÆRg æ at a given molar mass. Several quantitative
relationships have been proposed to relate g to branching
structure.20,21 Recently, van Ruymbeke et al.22 developed a
method to obtain structural information for systems con-
taining linear and star molecules. The MMD obtained
from SEC was split into contributions of various structures
using a statistical approach. They found that the calculated
structural compositions depended strongly on the MMD
used; i.e., the amounts of linear, 3-arm star, and 4-arm star
material found by SEC with a RI detector (SEC/RI) were
very different from those found by SEC with a multiangle
laser light scattering detector (SEC/MALLS). To avoid
such discrepancies, they used MMD data from the SEC/RI
for low and SEC/MALLS for high molar mass material.
However, the switching point from one MMD to the other
is arbitrary, and we concluded that for our samples more
reliable structural composition information could be ob-
tained by taking advantage of the higher resolution of
TGIC.
Figure 5. TGIC chromatograms for final products; samples were
processed in 1,4-dioxane at a flow rate of 0.5 mL/min. Corresponding
peak molar mass is shown in Table 3. Column temperature is also
shown in the plot.
Table 3. Peak Molar Mass Measured by TGIC
First, a small amount of material was fractionated using
TGIC separation columns, and eluates over given time
intervals were collected into separate vials. Average molar
masses (Mn and Mw) for each collected fraction of eluates
were then determined by SEC-3. We found that the Mn of the
last collected fraction Mn,highest molecular component (highest
molecular weight fraction in Table 4) of each sample was
peak molar mass Mp (kg/mol)
sample code peak 1 peak 2 peak 3 peak 4 peak 5 peak 6
HA12B40
HA30B40
HA40B40
HA12B100
20
65
110
103
28
117
121
170
43
131
205
198
57
164
70
83
1
close to 2 times that of /2-H molecules (Table 1) with an
compared to SEC-1, despite the fact that both techniques
are supposed to yield absolute molecular weights. Although
it is well-known that molar mass averages and MMDs vary
from one laboratory to another, especially since the light
scattered intensity is proportional to the square of dn/dc,
using different dn/dc values would cause deviation in molar
mass measurement by LS; it should not be confused with
our case in which one SEC result shows a single peak while
others show double peaks. Thus, we suspected that our
average difference of 4% between the two Mn values. Taking
into account that the uncertainty of molecular weights
measured by SEC/TALLS is typically (5%,23 we concluded
that the highest molecular weight fraction of each sample
consisted of H molecules. The lower molecular weight
materials were hypothesized to be the byproducts shown in
Figure 6 which was based on the synthesis mechanism shown
in Figure 1, and the molecular weights of these components
were estimated.

212 Macromolecules, Vol. 44, No. 2, 2011
Table 4. Estimated Structures and Compositions of H-PBDs
Li et al.
samples
fraction name
Mn (kg/mol)
PDI
wt %
inferred structurea
HA12B40
10
20
30
40
50
60
10
20
30
10
20
30
40
10
20
30
22
29
43
55
69
79
103
168
187
83
108
138
167
113
129
205
1.01
1.00
1.01
1.00
1.01
1.01
1.02
1.01
1.02
1.01
1.01
1.00
1.00
1.01
1.01
1.00
3
6
2
linear
linear
linear or 3-arm star
4-arm star
3-arm star
H
linear or 3-arm star
3-arm star
H
linear or 3-arm star
4-arm star
3-arm star
H
3
4 or 6
8
9
15
21
25
30
33
15
52
18
21
27
34
6
HA12B100
HA30B40
4 or 6
9
4 or 6
8
9
HA40B40
4 or 6
7
linear or 3-arm star
3-arm star
H
17
77
a Number refers to the structure identification number shown in Figure 7.
Figure 6. The most possible byproducts formed in the H-PBD synthesis. The number shown in brackets corresponds to the structure identification
number shown in Figure 7, while the length of each segment is shown in parentheses with a and b equal to the arm length shown in Table 2 and the
crossbar length calculated by eq 4, respectively.
We estimated Mn,crossbar (Mn of the crossbar) by use of
eq 4:
most likely to be produced during synthesis according to the
composition of arms and crossbars indicated in parentheses
in Figure 7. The formation of byproducts other than those
shown in Figure 7, with higher molar masses or more com-
plicated structures (e.g., branch-on-branch) is possible but very
unlikely due to steric hindrance of the bulky aromatic groups in
the DCMSDPE and the poor mobility of a heavy molecule.
Mn,crossbar ¼ Mn,highest molecular component - 4Mn,arm ð4Þ
where Mn,arm is Mn of the arm given in Table 2. Using Mn,crossbar
and Mn,arm, we calculated the Mn values of the byproducts

Article
Macromolecules, Vol. 44, No. 2, 2011 213
The TGIC data shown in Figure 8 also indicate negligible
amounts of higher molecular weights byproducts compared
to those shown in Figure 7. We then assigned structures to
the remaining fractions by matching their Mn with those of
the byproducts shown in Figure 7. Using this method, we
found that the measured Mn values of the fractions of eluate
shown in Table 4 were very close to the calculated Mn values
of the most probable byproducts for all samples, with an
average difference of 4%. Finally, to obtain the weight
fraction of each structure, the RI signals shown in Figure 8
were deconvoluted using Gaussian fitting, and the results are
shown in Table 4.
We use HA12B40 data to illustrate this procedure. The
vertical solid lines in the TGIC curve of Figure 8 indicate
the time intervals to collect various fractions of the eluate,
while the dashed lines are the Gaussian fits of each peak
obtained from deconvolution. The measured Mn of fraction
60 (79 kg/mol) is very close to 2 times that of a 1/2-H
molecule (82 kg/mol), implying that fraction 60 consists of
H molecules. Taking Mn,arm to be 10.6 kg/mol, as measured
on the product of step 2 in Figure 1, allows us to assign
fraction 50 to H molecules with a missing arm (i.e., structure 9).
Using eq 4, Mn,crossbar was found to be 36.7 kg/mol. With this
information, fraction 40 (Mn=55 kg/mol) is best described as
an asymmetric 4-arm star (i.e., structure 8). The same
analysis was applied to the remaining fractions, whereas
fractions 10 and 20 are best described as linear chains
(structures 2 and 3, respectively), and fraction 30 can be
identified as either a linear or an asymmetric 3-arm star
(structure 4 or 6).
Figure 7. Schematic representations of the most probable struc-
tures of low molecular weight byproducts; the length of each
segment is shown in parentheses with a and b equal to the arm
length shown in Table 2 and the crossbar length calculated by eq 4,
respectively.
Conclusions
Size exclusion chromatography (SEC) techniques were un-
able to resolve the structural compositions of our long chain
branched polymers containing various byproducts due to its
inherent low resolution. Furthermore, data depending on
sample treatment, quality of the separation columns, data
acquisition, and processing revealed significant variation in
SEC results among three reputable laboratories. Nonetheless,
the SEC results implied that samples originally thought to be
monodisperse H-shaped polymers were more complex, and
this was confirmed by temperature gradient interaction chro-
matography (TGIC) analysis. We combined knowledge of the
synthesis mechanism with TGIC data to infer the detailed
structural composition of the samples. The results showed that
while the synthesis method had successfully obviated the
production of high-molecular-weight byproducts, our poly-
mers were mixtures of the intended H molecules with low-
molecular-weight linear, 3-arm star, and 4-arm star bypro-
ducts. This made it possible to use these samples for further
rheological and molecular modeling studies. At the same time,
the results clearly demonstrated the limitation in the SEC-
based characterization of branched polymers and the necessity
of employing complementary tools such as TGIC for a rigorous
characterization.
Acknowledgment. The H-PBDs were produced under the
National Science Foundation (NSF) grant-DMR 0604965. The
Natural Sciences and Engineering Research Council of Canada
(NSERC)providedfinancial supportfor the structure study. T.C.
acknowledges the support from NRF via NRL (R0A-2007-000-
20125-0), SRC (R11-2008-052-03002), and WCU (R31-2008-
000-10059-0) programs. S.W.L. gratefully acknowledges the
assistance at Drs. Willem deGroot, Ray Brown, and Tianzi
Huang at the Dow Chemical Company for the SEC-2 data and
Dr. Jacques Roovers for his valuable advice regarding anionic
polymerization.
Figure 8. Fractionation of final products by TGIC and the SEC-3
elution profile (RI response) of each collected fraction. The rectangles
indicate the time intervals to collect various fractions of the eluate for
SEC characterization.

214 Macromolecules, Vol. 44, No. 2, 2011
Li et al.
References and Notes
(13) Lee, H. C.; Chang, T.; Harville, S.; Mays, J. Macromolecules 1998,
31, 690–694.
(14) Santee, E. R.; Chang, R.; Morton, M. J. Polym. Sci., Polym. Lett.
Ed. 1973, 11, 449–452.
(15) Tanaka, Y.; Takeuchi, Y. J. Polym. Sci., Part A-2 1971, 9, 43–57.
(16) Sadeghi, G. M.; Barikani, M.; Morshedian, J.; Taromi, F. A. Iran.
Polym. J. 2003, 12, 515–521.
(17) D’Agnillo, L.; Soares, J. B. P.; Penlidis, A. J. Polym. Sci., Part B:
Polym. Phys. 2002, 40, 902–921.
(18) Trathnigg, B. Size-exclusion chromatography of polymers. In
Encyclopedia of Analytical Chemistry; John Wiley & Sons Ltd.:
Chichester, 2000.
(19) Dealy, J. M.; Larson, R. G. Structure and Rheology of Molten
Polymers: From Structure to Flow Behavior and Back Again;
Hanser-Gardner Publications: Cincinnati, OH, 2006.
(20) Bonchev, D.; Markel, E. J.; Dekmezian, A. H. Polymer 2002, 43,
203–222.
(21) Zimm, B. H.; Stockmayer, W. H. J. Chem. Phys. 1949, 17, 1301–
1314.
(22) van Ruymbeke, E.; Coppola, S.; Balacca, L.; Righi, S.; Vlassopoulos,
D. J. Rheol. 2010, 54, 507–538.
(1) Fetters, L. J.; Lohse, D. J.; Colby, R. H. Chain dimensions and
entanglement spacings. In Physical Properties of Polymers Hand-
book; 2nd Ed., Springer: New York, 2006.
(2) Rahman, M. S.; Aggarwal, R.; Larson, R. G.; Dealy, J. M.; Mays,
J. Macromolecules 2008, 41, 8225–8230.
(3) Perny, S.; Allgaier, J.; Cho, D.; Lee, W.; Chang, T. Macromolecules
2001, 34, 5408–5415.
(4) Roovers, J.; Toporowski, P. M. Macromolecules 1981, 14, 1174–1178.
(5) Uhrig, D.; Mays, J. W. J. Polym. Sci., Part A: Polym. Chem. 2005,
43, 6179–6222.
(6) Lee, H. C.; Chang, T. Polymer 1996, 37, 5747–5749.
(7) Chang, T. J. Polym. Sci., Part B: Polym. Phys. 2005, 43, 1591–1607.
(8) Chang, T. Adv. Polym. Sci. 2003, 163, 1–60.
€
(9) Glockner, G. Gradient HPLC of Copolymers and Chromatographic
Cross-Fractionation; Springer Verlag: New York, 1992.
(10) Pasch, H.; Trathnigg, B. HPLC of Polymers; Springer Verlag: Berlin,
1998.
(11) Lee, W.; Lee, H. C.; Chang, T.; Kim, S. B. Macromolecules 1998,
31, 344–348.
(12) Lee, W.; Lee, H.; Cha, J.; Chang, T.; Hanley, K. J.; Lodge, T. P.
(23) Terao, K.; Mays, J. W. Eur. Polym. J. 2004, 40, 1623–1627.
Macromolecules 2000, 33, 5111–5115.