Synthesis and characterization of hyperbranched poly(ε-caprolactone)s having different lengths of homologous backbone segments

Article abstract of DOI:10.1021/ma034170i

Hyperbranched poly(ε-caprolactone)s (HPCLs) were synthesized by moisture-sensitive catalyst-free polycondensation of AB₂ macromonomers, 2,2-bis[ω-hydroxy oligo(ε-caprolactone)methyl]-propionic acids. The HPCLs were designed to incorporate diffe

Full text of DOI:10.1021/ma034170i

8
630
Macromolecules 2003, 36, 8630-8637
Synthesis and Characterization of Hyperbranched Poly(ꢀ-caprolactone)s
Having Different Lengths of Homologous Backbone Segments
†
,‡
J eon gsoo Ch oi a n d Seu n g-Yeop Kw a k *
Research Institute of Advanced Materials (RIAM), and School of Materials Science and Engineering,
Seoul National University, San 56-1, Shinlim-dong, Kwanak-ku, Seoul 151-744, Korea, and
Hyperstructured Organic Materials Research Center (HOMRC), and School of Materials Science and
Engineering, Seoul National University, San 56-1, Shinlim-dong, Kwanak-ku, Seoul 151-744, Korea
Received February 10, 2003; Revised Manuscript Received J uly 10, 2003
ABSTRACT: Hyperbranched poly(ꢀ-caprolactone)s (HPCLs) were synthesized by moisture-sensitive
catalyst-free polycondensation of AB macromonomers, 2,2-bis[ω-hydroxy oligo(ꢀ-caprolactone)methyl]-
2
propionic acids. The HPCLs were designed to incorporate different lengths of linear oligomeric segments
consisting of 5, 10, and 20 ꢀ-caprolactone monomer units on the branched backbone chains, accordingly
1
named as HPCL-5, -10, and -20, respectively. End-group analyses were performed on H NMR spectra of
the three HPCLs, which provided information about the average number of ꢀ-caprolactone units
incorporated in the AB macromonomers and thus the average number of AB macromonomer units
2 2
incorporated in the resulting hyperbranched polymers. Consequently, the absolute values of molecular
weights for the HPCLs were calculated. Size exclusion chromatography equipped with multiangle laser
light scattering detector (SEC-MALLS) was employed to measure absolute molecular weights and
1
molecular weight distributions of HPCLs. Then, the results were compared with those obtained from H
NMR end-group analyses, indicating that there was a good agreement between them. From small-angle
X-ray scattering (SAXS), the radii of gyration of the HPCLs and their linear counterpart, poly(ꢀ-
caprolactone) (LPCL), were determined from the initial slope of the reciprocal of scattered intensity, 1/I(q),
2
vs the square of scattering vector, q , curves, fit by the Zimm scattering function. The ratio of mean-
square radius of gyration of each HPCL to that of LPCL, termed the branching ratio, resulted in the
relative degree of branching for individual HPCLs. As the length of linear oligo(ꢀ-caprolactone) segments
decreased, the more highly branched polymers were obtained, i.e., in the order of HPCL-5, -10, and -20
from higher to lower degree of branching. The melting temperatures of HPCLs were lower than that of
LPCL and decreased with increase in the degree of branching. In contrast, the thermal decomposition
temperatures of HPCLs were shown to be higher than that of LPCL, becoming higher as the degree of
branching increased.
In tr od u ction
polymers with a range of structural variations such as
different lengths of incorporated oligo(ꢀ-caprolactone)
segments. Recently, Skaria et al. have reported a new
Recently, hyperbranched polymers have received
considerable attention due to the expectation that their
8
synthetic concept, the enzyme-catalyzed synthesis of
highly branched structure will impart unique physical
9
and chemical properties.1-5 Hyperbranched polymers
HPCLs. This synthetic route enables to skip multistep
reactions of preparing AB2 macromonomers in the
synthesis of HPCLs of Trollsås et al. However, these
two HPCLs have the distinct difference in the chain
structure. The HPCLs of Skaria et al. have randomly
incorporated oligo(ꢀ-caprolactone) segments, while rather
precisely controlled and readily characterizable oligo-
(ꢀ-caprolactone) segments are incorporated in the HP-
CLs of Trollsås et al., which is more preferable in
combining the properties of conventional linear poly-
mers with the new properties of hyperbranched poly-
mers and in investigating the effects of structural
variables such as branching and the length of incorpo-
rated oligo(ꢀ-caprolactone) segments on the resulting
HPCLs. Nevertheless, it is more desirable to modify the
original synthetic route into an alternative one such as
a moisture-sensitive catalyst-free polyesterification in
order to be applicable to the industrial scale production
with easier process control for its further practical
application such as ecologically friendly polymeric ad-
ditives.
may be considered as irregular analogues of the den-
drimers that have well-defined and perfectly branched
structure. Although dendrimers are prepared through
divergent or convergent approaches that are composed
of step-by-step synthesis, hyperbranched polymers are
prepared by direct one-pot polymerization of ABx-type
6
monomers. This synthetic simplicity along with the fact
that many of the physical/mechanical properties of both
hyperbranched polymers and dendrimers have been
reported to be similar when possessing the same repeat-
7
ing units has encouraged many researchers to develop
tailor-made materials for a specific application where
high performance and/or novel functionality are needed.
Trollsås et al. have reported the synthesis of hyper-
branched poly(ꢀ-caprolactone)s (HPCLs) from polyes-
terification of homologous AB2 macromonomers with
different lengths of linear oligo(ꢀ-caprolactone)s using
4
-(dimethylamino)pyridinium 4-toluenesulfonate (DPTS)
8
and 1,3-dicyclohexylcarbodiimide (DCC) as catalysts.
This novel synthetic approach was shown to allow
production of a homologous series of hyperbranched
Determination of the degree of branching in hyper-
branched polymers is one of prerequisites to describe
*
To whom correspondence should be addressed.
Research Institute of Advanced Materials.
Hyperstructured Organic Materials Research Center.
†
their structural features as well as to correlate the
‡
10,11
structure with their properties.
The average degree
1
0.1021/ma034170i CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/22/2003

Macromolecules, Vol. 36, No. 23, 2003
Hyperbranched Poly(ꢀ-caprolactone)s 8631
of branching (DB), the term first proposed by Fr e´ chet
Sch em e 1
1
2
et al., is commonly calculated as the sum of the
terminal and dendritic units divided by the total sum
of all repeating units as the following equation:
∑D +
∑
T
DB )
(1)
D + L +
T
∑
∑ ∑
where ∑D, ∑T, and ∑L are the numbers of dendritically,
terminally, and linearly incorporated monomer units in
the hyperbranched polymers. Determination of DB is
usually limited to some cases in which the fractions of
each component (D, L, T) are distinguishable, mostly
observable with the spectroscopic methods.13 Otherwise,
an alternative approach to obtain information about DB
is possible from measurements of the dimensions of the
hyperbranched polymer molecules with reference to
their linear analogue, recognizing that the DB is a
matter of relativeness of the partially branched to the
completely linear structure. The hyperbranched polymer
chains are, in general, less densely packed than the
linear of the same molecular weight in the solid state
due to their amorphous nature. However, in the com-
pletely dissolved solution state, the dimensions of the
linear polymers become larger than those of the branched
molecules of the same molecular weight because the
entangled chains and the crystalline regions get dis-
solved to be free of any restricting force and to make
the molecular size larger, while the branched molecules
are less solvated due to the chemically bonded branched
structure. Moreover, the higher the DB, the smaller
the dimensional size among the hyperbranched poly-
mers of the same chemical structure and the analogous
molecular weights. Thus, the ratio of the mean-square
radius of gyration of the given hyperbranched polymer
to that of its linear analogue, termed as branching ratio,
g, provides an alternative means to characterize the
branching structure and hence obtain information about
relative DBs among the homologous series of hyper-
branched polymers:15
this study is to investigate the thermal characteristics
of HPCLs in conjunction with their branching features.
Exp er im en ta l Section
Ma ter ia ls. 2,2-Bis(hydroxymethyl)propionic acid (bis-MPA,
98%), potassium hydroxide (KOH, 99%), benzyl bromide (98%),
ꢀ
5
-caprolactone (99%), palladium on activated carbon (Pd/C,
wt % palladium), tin(II) 2-ethylhexanoate (Sn(Oct) , 99%),
2
and p-toluenesulfonic acid (TSA, 98.5%) were purchased
from Aldrich Chemical, being used as received. N,N-Dimethyl-
formamide (DMF, 99%), diethyl ether (99%), toluene (99%),
methanol (MeOH, 99%), tetrahydrofuran (THF, 99%), and
ethyl acetate (EtOAc, 99%) were purchased from Daejung
Chemicals & Metals (Korea) and used without further puri-
fication. Linear poly(ꢀ-caprolactone) (LPCL), as a linear coun-
terpart of the hyperbranched poly(ꢀ-caprolactone) (HPCLs)
prepared in this study, was reported by the supplier (Aldrich
1
4
n
Chemical) to have number-average molecular weight, M , of
1
0 000.
Syn th esis. 2,2-Bis(h yd r oxym eth yl)p r op yl Ben zoa te, 1.
2
7 g (200 mmol) of bis-MPA and 12.9 g (230 mmol) of KOH
were dissolved in 150 mL of DMF, and the potassium salt was
allowed to form by stirring reaction mixture at 100 °C for 1 h.
Benzyl bromide (42.4 g) was then added dropwise to a solution,
and the mixture was stirred and allowed to react with
maintaining temperature for 15 h. After completion of the
reaction, the solvent was evaporated; the residue was dissolved
in 600 mL of diethyl ether and extracted with distilled water
(3 × 150 mL). The crude product was then purified by
recrystallization from toluene to give 2,2-bis(hydroxymethyl)-
propyl benzoate, 1 (Scheme 1), as white crystals. Yield: 65%.
2
〈
R 〉
g
branched
g )
(2)
2
〈
R 〉
g linear
In the present study, hyperbranched poly(ꢀ-capro-
lactone)s (HPCLs) having different lengths of linear
oligo(ꢀ-caprolactone) segments incorporated in their
branched backbone chains are synthesized by a vacuum
melt polycondensation reaction free from using moisture-
sensitive catalysts. Then, the average number of ꢀ-
caprolactone monomer units in AB2 macromonomers
1
6,17
The structural verification has been described elsewhere.
Rin g-Op en in g P olym er iza tion of E-Ca p r ola cton e. 2,2-
Bis[ω-hydroxy oligo(ꢀ-caprolactone)methyl]propyl benzoate (ben-
zyl protected AB
opening polymerization which was initiated with 1 using a
catalytic amount of Sn(Oct) . The target number of ꢀ-capro-
2
macromonomer), 2, was synthesized by ring-
〈
Nꢀ-CL〉, the average number of AB2 macromonomers in
2
HPCLs 〈NAB 〉, and the resulting absolute value of
2
lactone monomer units incorporated in 2 was determined by
the monomer-to-initiating hydroxyl group molar ratio. The
reaction was carried out in the melt at 110 °C for 20 h. The
crude product was then dissolved in THF and precipitated into
cold MeOH to give 2 as white solids. Yield: 92-96%. The
complete polymerization procedure and the structural verifica-
tion of 2 have been reported elsewhere.8
Hyd r ogen a tion . 2,2-Bis[ω-hydroxy oligo(ꢀ-caprolactone)-
2
methyl]propionic acid (R-carboxylic-ω-dihydroxy AB macro-
monomer), 3, was obtained by hydrogenation reaction of 2
using a Pd/C catalyst. Hydrogenation reaction was carried out
in a solution of 2 in THF/EtOAc (20/80, v/v) mixed solvent.
After the Pd/C (10 wt % to 2) was added to a solution, the
number-average molecular weights of the HPCLs are
1
determined by end-group analyses on H nuclear mag-
netic resonance (NMR) spectra. These results are then
compared and confirmed with the absolute molecular
weights and molecular weight distribution of the HPCLs
determined by means of size exclusion chromatography
(
(
SEC) equipped with multiangle laser light scattering
MALLS) detector. To analyze the degree of branching,
the branching ratio among the HPCLs is compared, in
which the absolute values of radius of gyration for both
HPCLs and LPCL are obtained from small-angle X-ray
scattering (SAXS) experiments. Another focal point of
2
reaction vessel was evacuated and put under a hydrogen (H )

8
632 Choi and Kwak
Macromolecules, Vol. 36, No. 23, 2003
F igu r e 1. 1H NMR spectrum (300 MHz) of hyperbranched poly(ꢀ-caprolactone)s (HPCLs).
micro Styragel, 105 × 2, 10 , 10 , 500 Å) was used. A
MiniDAWN (Wyatt Technology) multiangle laser light scat-
tering (MALLS) detector equipped with 690 nm semiconductor
laser was connected between the size exclusion columns and
a Wyatt Optilab DSP differential refractometer. The mixed
4
3
atmosphere. The reaction was stopped after 24 h, and the Pd/C
was filtered off. The reaction mixture was concentrated by
evaporation of solvent and then poured into cold MeOH to give
3
2
as white solids. The AB macromonomers in the above were
designed to possess different lengths of linear oligomeric
segments consisting of 5, 10, and 20 ꢀ-caprolactone units; they
3
solvent of THF, DMF, and lithium nitrate (LiNO ) was used
were designated as AB
2
-5, -10, and -20, respectively. Yield:
as a mobile phase at a flow rate of 1.0 mL/min. A 100 µL
sample of a 0.1 mg/mL solution, which was filtered through a
0.2 µm Whattman filter prior to use, was injected for all
measurements. The software Astra (ver. 4.90.04, Wyatt Tech-
nology) was used to compute the calibration curves and the
molecular weights distribution. The calibration of detectors
was achieved using polystyrene standards of narrow molecular
weight distribution with known molecular weights.
8
8-97%. The structural verification has been described else-
8
where.
P olycon d en sa tion w ith Con tin u ou s Wa ter Rem ova l.
Hyperbranched poly(ꢀ-caprolactone)s (HPCLs), 4, were syn-
thesized by polycondensation with continuous water removal
reaction of 3. Then, 3 and TSA were carefully mixed in a three-
necked flask equipped with an argon inlet, a drying tube, and
a stirrer. The mixture was allowed to react at 110 °C under a
stream of argon, removing the water formed during the
reaction. After 2 h, the argon stream was stopped purging,
and the flask was sealed and connected to a vacuum line (10
Torr, cooling trap) for 3 h. After polycondensation reaction was
completed, the mixture was dissolved in THF and precipitated
into MeOH to give 4 as white solids. On the basis of the AB
macromonomers of different lengths of oligomeric ꢀ-caprolac-
Sm a ll-An gle X-r a y Sca tter in g (SAXS). The radii of
gyration, R , of HPCLs and LPCL, which were too small to be
g
measured by SEC-MALLS, were determined from SAXS. The
SAXS intensity distribution, I(q), was measured with a rotat-
ing-anode X-ray generator (Bruker AXS Nanostar) operated
at 40 kV and 35 mA. The X-ray source was monochromatized
Cu KR (λ ) 1.54 Å) radiation, and the scattering angle ranges
between 0.5° and 2.5°. For the measurement, sample solutions
(1 mg/mL) in THF were injected into glass capillary in liquid
sample holder frame. The background correction was carried
-2
2
tone segments used for syntheses, the resulting HPCLs were
accordingly named as HPCL-5, -10, and -20, respectively.
1
Yield: 87-94%. H NMR (CDCl
3
): Figure 1.
out using pure THF solvent. The values of R were estimated
g
by curve fitting of reciprocal plots of I(q) as a function of q
with the Zimm scattering function, which is described in detail
later.
1
Ch a r a cter iza tion . H NMR An a lysis. The chemical struc-
tures of synthesized HPCLs were elucidated by 1H NMR
spectroscopy, employing a Bruker Avance DPX-300 spectrom-
eter with the tetramethylsilane (TMS) proton signal as an
internal standard in CDCl
Differ en tia l Sca n n in g Ca lor im etr y (DSC) a n d Th er -
m a l Gr a vim etr ic An a lysis (TGA). The melting tempera-
1
3
. H NMR spectra was also analyzed
to determine the average number of ꢀ-caprolactone monomer
units incorporated in AB macromonomers, 〈Nꢀ-CL〉, the average
number of AB macromonomer units incorporated in HPCLs,
AB2〉, and consequently the number-average molecular weights
of HPCLs, Mn,NMR. The methods for determining the values of
ꢀ-CL〉, 〈NAB2〉, and Mn,NMR were explained in detail in the
m
tures, T ’s, of HPCLs and LPCL were determined by DSC with
2
a TA Instruments DSC 2920 at heating rates of 10 °C/min
under a nitrogen atmosphere. The temperatures with 10%
weight loss of the polymers, Td10’s, were determined by TGA
with a TA Instruments TGA 2050 at the heating rate of 10
°C/min under a nitrogen atmosphere.
2
〈
N
〈
N
Results and Discussion section.
Mea su r em en t of Differ en tia l Refr a ctive In d ex In cr e-
m en t, d n /d c. The dn/dc was measured with a Wyatt Optilab
DSP interferometric differential refractometer at 690 nm at
room temperature for the five concentrations (1 mg/mL < c <
Resu lts a n d Discu ssion
The AB2-type macromonomers, 2,2-bis[ω-hydroxy oligo-
(ꢀ-caprolactone) methyl]propionic acids, 3, used for the
preparation of hyperbranched poly(ꢀ-caprolactone)s (HP-
CLs), 4, were synthesized according to a reaction
9
1
mg/mL) of poly(ꢀ-caprolactone) in a mixed solvent (THF/DMF
0/1 v/v; LiNO 2 g/L). The dn/dc value was determined with
3
1
8,19
the following relationship:
8
developed by Trollsås et al. The recipe and the method
dn/dc ) k d(∆V)/dc
(3)
for the synthesis of HPCLs were, however, modified
from those of Trollsås et al., considering the applicability
to the industrial scale production wherein the syntheses
must be much less sensitive to the moisture. Table 1
summarizes the conditions for the preparation of AB2
macromonomers and HPCLs. Ring-opening polymeri-
zation with variable ꢀ-caprolactone-to-initiating hy-
droxyl group molar ratios ([ꢀ-CL]0/[-OH]0 ) 5, 10, and
20), followed by hydrogenation reaction, produced three
different AB2 macromonomers (i.e., AB2-5, -10, and -20),
where k is an instrument calibration constant and ∆V is the
refractive index detector voltage. The k value was determined
-
4
to be 1.56 × 10 with standard sodium chloride (NaCl)
samples at the wavelength of 690 nm, 25 °C.
Size Exclu sion Ch r om a togr a p h y w ith Mu ltia n gle La -
ser Ligh t Sca tter in g (SEC-MALLS). A size exclusion chro-
matography (SEC) modular system consisted of a Dionex DG-
1
210 degasser, a Dionex P680 HPLC pump, a Dionex automatic
sample injector, and sets of five size exclusion columns (Ultra-

Macromolecules, Vol. 36, No. 23, 2003
Hyperbranched Poly(ꢀ-caprolactone)s 8633
Ta ble 1. P r ep a r a tion of AB2 Ma cr om on om er s a n d
targeted values solely predictable from [ꢀ-CL]0/[-OH]0
ratios (Table 1). In a similar manner, the values of
Hyp er br a n ch ed P oly(E-ca p r ola cton e)s
macro-
ꢀ-CL]0/ monomer
〈NAB 〉 of HPCLs were calculated from the following
2
[
yieldd
(%)
equation based on the branching theories developed by
a
〈Nꢀ-CL〉b 〈NAB2〉c Mn,NMR
sample [-OH]0
entry
21
Flory:
AB2-5
AB2-10
AB2-20
HPCL-5
HPCL-10
HPCL-20
5
10
20
5.7
10.3
20.1
1435
2485
4720
11510
12700
15100
88
94
97
87
94
92
m
〈NAB2〉 )
(4)
2
〈Nꢀ-CL〉 - m
AB2-5
AB2-10
AB2-20
8.1
5.1
3.3
where m represents the ratios of the integrated area of
peak R′ to the integrated area of peak E′ of HPCLs
a
ꢀ
-Caprolactone monomer-to-initiating hydroxyl group molar
ratio. b Average number of ꢀ-caprolactone units incorporated in
(Figure 2, bottom). Consequently, the number-average
1
c
AB2 macromonomers determined by H NMR. Average number
of AB2 macromonomer units incorporated in HPCLs determined
molecular weights of AB2 macromonomers and HPCLs
were calculated from the following equation and also
listed in Table 1 along with the results of 〈Nꢀ-CL〉 and
1
d
by H NMR. Water/methanol (50/50, v/v)-insoluble part for AB2-5
and cold methanol-insoluble part for the other samples.
〈
NAB 〉:
2
Mn,NMR(AB -n) ) MW
+ 2MWꢀ-CL〈Nꢀ-CL
bis-MPA
〉
(5)
2
Mn,NMR(HPCL-n) ) Mn,NMR(AB -n)〈NAB2〉 +
2
MW [〈NAB2〉 - 1] (6)
H O
2
where MWbis-MPA, MWꢀ-CL, and MWH O are the molecular
2
weights of bis-MPA, ꢀ-caprolactone, and water, respec-
tively. From the molecular weight data of HPCLs
prepared in the same condition, it was found that larger
macromonomers produced slightly lower molecular
weight HPCLs (Table 1). The decrease in the reactivity
of larger macromonomers can be explained by increase
in the steric hindrance as well as decrease in the
concentration of reacting sites in a unit volume.
1
To prove and confirm the results from H NMR end-
group analyses, a size exclusion chromatography (SEC)
with multiangle laser light scattering (MALLS) detector
was employed as an another means to obtain absolute
information about molecular weights. From an SEC, the
smooth and monomodal chromatograms were observed
for all of AB2 macromonomers and HPCLs as illustrated
in Figure 3a for AB2-10 and HPCL-10. Compared in
Figure 3b are the concentration chromatograms of three
HPCLs and LPCL; although the molecular weight of
HPCL-20 was higher than that of LPCL on the basis of
F igu r e 2. 1H NMR spectra focusing the peaks of the chain
ends (E and E′) and of the repeating methylene units (R and
R′) in AB
2
macromonomers (top) and hyperbranched poly(ꢀ-
caprolactone)s (bottom).
1
determination from H NMR end-group analyses, their
chromatograms were shown in the almost same reten-
tion volume area. Such unexpectedness that usually
occurs in characterizing molecular weights of hyper-
branched polymers with gel permeation chromatogra-
phy is generally ascribed to the inevitable decrease in
hydrodynamic volume imposed by branching,22 which
in turn exceeds the relative discrepancy with the
polystyrene standard. The same trend was also found
in the cases of HPCL-5 and -10; even though the
molecular weights of the two HPCLs determined by end-
group analyses were quite similar to that of LPCL, their
chromatograms lay on the higher retention volume area
relative to the LPCL. This also indicates the smaller
hydrodynamic volumes of HPCL-5 and -10 than that of
LPCL, despite the almost similar molecular weights. An
SEC equipped with MALLS detector overcomes such
limitation typically brought about by the conventional
SEC for hyperbranched polymers and dendrimers.
3
, having different lengths of linear oligo(ꢀ-caprolac-
tone)s. The HPCLs were successfully produced from 3
by polycondensation with continuous water removal
with p-toluenesulfonic acid (TSA), which was known to
be effective for the syntheses of hyperbranched polyes-
ters with the lower sensitivity to the moisture com-
pared to 4-(dimethylamino)pyridinium 4-toluenesulfonate
2
0
2
3
(DPTS)/1,3-dicyclohexylcarbodiimide (DCC) catalyst sys-
8
1
tem reported previously. The H NMR spectra of both
AB2 macromonomers and HPCLs were analyzed to
determine the average number of ꢀ-caprolactone mono-
mer units incorporated in AB2 macromonomers, 〈Nꢀ-CL〉,
and the average number of AB2 macromonomer units
incorporated in HPCLs, 〈NAB 〉. As shown in Figure 2,
the peak assigned to the chain ends (E, -CH2OH, δ
3
methylene units (R, -COCH2, δ 2.22) are quite distin-
guishable. Thus, the 〈Nꢀ-CL〉 values of AB2 macromono-
mers could be readily calculated from the ratios of the
integrated area of peak R to that of peak E (Figure 2,
top), which were in close agreement with those of
2
.59) and the peak corresponding to the repeating
In MALLS detecting, the angular dependence of the
excess absolute time-average scattered light intensity,
known as the excess Rayleigh ratio R(θ), of a dilute
polymer solution at concentration C (g/mL) and the

8
634 Choi and Kwak
Macromolecules, Vol. 36, No. 23, 2003
Ta ble 2. Molecu la r Weigh t Da ta for Hyp er br a n ch ed
P oly(E-ca p r ola cton e)s (HP CLs) a n d Lin ea r
P oly(E-ca p r ola cton e) (LP CL) Deter m in ed by 1H NMR
En d -Gr ou p An a lysis, Con ven tion a l SEC, a n d SEC w ith
MALLS Detector
1H NMR SEC
SEC-MALLS
sample
Mn,NMR Mn,SEC Mw/Mn Mn,MALLS Mw,MALLS Mw/Mn
HPCL-5
HPCL-10 12 700
HPCL-20 15 100 10 500
LPCL 10 100 10 200
11 510
8 300
9 500
1.7
1.6
1.6
1.4
11 800
12 600
15 700
10 700
20 900
20 400
24 200
15 200
1.8
1.6
1.5
1.4
and the wavelength of light, respectively. The optical
constant K is given by
2
2
4
π n
dn 2
dC
0
K )
(9)
4
(
)
λ N
A
with n0, NA and dn/dC being the refractive index of the
solvent at the given wavelength, Avogadro’s number,
and the specific refractive index increment of the
polymer in the given solvent at the given wavelength.
The value of dn/dC was determined beforehand in a
separate experiment with the five concentrations of
poly(ꢀ-caprolactone) solutions (1 < C < 9 mg/mL) in the
solvent (THF/DMF 10/1 v/v; LiNO3 2 g/L) at the
wavelength (690 nm) used for SEC-MALLS measure-
ments. From the plot of ∆V vs concentration, the dn/dC
value was obtained to be 0.0652. In SEC-MALLS
measurements, the sample concentrations for each data
slice (Ci) are determined from concentration detector
output, and the molecular weights for each data slice
2
(
(
Mi) are determined from the KC/R(θ)) vs sin (θ/2) plot
Debye plot). With the values of Ci and Mi for each data
slice along the entire chromatogram, the number-
average molecular weights and weight-average molec-
ular weights were calculated from the following rela-
tionship:
F igu r e 3. Size exclusion chromatograms of typical AB
macromonomer and hyperbranched poly(ꢀ-caprolactone) (a)
and of all three hyperbranched poly(ꢀ-caprolactone)s (HPCL-
2
5
, -10, and -20) and linear poly(ꢀ-caprolactone) (b).
Ci
∑
Mn,MALLS
)
)
(10)
(11)
scattering angle θ was measured. The Rayleigh ratio,
R(θ), is defined as22
∑
i
C /Mi
C Mi
∑
i
2
Mw,MALLS
(
Iθ - Iθ,solvent)r
Iθ - Iθ,solvent
∑C
i
R(θ) )
) f
(7)
I0V
I0
Thus, the absolute molecular weights and the polydis-
persity index determined by the SEC-MALLS are sum-
marized in Table 2, the results from H NMR end-group
where Iθ is the intensity of light scattered in the
direction at an angle θ to the incident light. I0, Iθ,solvent,
V, r, and f are the intensity of the incident light, the
scattered intensity of the solvent, the scattering volume,
the distance between the scattering volume and the
detector, and an instrumental calibration constant,
respectively. The value of R(θ) is related to the molar
1
analysis and SEC being also listed for comparison. For
the HPCLs, the number-average molecular weights
determined by SEC-MALLS, Mn,MALLS, were found to be
1
in good agreement with those of H NMR end-group
analysis, Mn,NMR. Whereas, the results from conven-
tional SEC, Mn,SEC, showed deviation from the values
of Mn,MALLS and Mn,MNMR as was expected.
Another important material characteristic to be char-
acterized to describe the structural features of the
hyperbranched polymers is the degree of branching
2
1/2
mass M, root-mean-square radius 〈Rg 〉 (equivalently
called the radius of gyration, Rg), and concentration C
as following the Zimm formalism of the Rayleigh-
Debye-Gans light scattering model:1
9,22
(DB). The DB is commonly determined from the relative
2
proportions among the NMR resonance peak areas
pertinent to the respective dendritic (D), linear (L), and
terminal (T) units of the given hyperbranched polymer
whenever there exist the distinguishable corresponding
chemical shifts. In some previous articles, the DB of the
HPCLs were determined from the chemical shifts of the
characteristic peaks of the methyl group of the bis-MPA
KC
1
1
1
M
16π
2
θ
=
+ 2A C )
1 +
sin
+ 2A C
2
2
(
( )
2
)
2
R(θ) M P(θ)
3
λ
(8)
where θ, P(θ), A2, and λ are the scattering angle, the
particle scattering function, the second virial coefficient,

Macromolecules, Vol. 36, No. 23, 2003
Hyperbranched Poly(ꢀ-caprolactone)s 8635
corresponding to each D, L, and T unit.8,9 However, the
chemical environments of the methyl group among D,
L, and T units in the HPCLs of the present study were
not significantly different due to intrinsically attached
long oligo(ꢀ-caprolactone) segments, which made the
HPCLs of the present study show neither distinguish-
1
able H resonance peaks useful for the calculation of
DB nor quaternary carbons unique for the determina-
tion of DB. In such a case, information about DB can
be alternatively achieved from measurements of the
dimensions of the hyperbranched polymer molecules
with reference to their linear analogue, recognizing that
the DB is a matter of relativeness of the partially
branched to the completely linear structure. One pos-
sible method is based on the plots of a radius of gyration
against molecular weight directly measurable from
2
4,25
SEC-MALLS,
in which information about DB can
F igu r e 4. Plot of ∆V vs concentration (g/mL) for poly(ꢀ-
caprolactone) in THF/DMF/LiNO mixed solvent at 690 nm.
3
be inferred from the values (less than 0.5) of the slope
in the plots. However, the radii of gyration of the three
HPCLs were found to be too small to be exactly
measured from the SEC-MALLS. Therefore, it seeks
another approach to measure the radius of gyration by
means of a small-angle X-ray scattering (SAXS) and
then determine the branching ratio, g, which denotes
the ratio of mean-square radii of gyration between a
given HPCL and the reference linear poly(ꢀ-caprolac-
tone). It is noteworthy that the parameter g is always
less than 1 and decreases with increase in the degree
of branching because the branched molecules become
more compactly packed rather than the more loosely
branched molecules or the corresponding linear coun-
terpart of the similar molecular weight in the completely
dissolved solution state.22 Scattering curves obtained
from polymer solutions, after corrections for instrumen-
tal effects, generally consist of scattering contributions
from the single particle scattering, molecular size and
shape of a single particle, and interparticle interference.
Thus, the scattering from a group of particles can be
expressed in terms of
I(q) ) kNP(q) S(q)
(12)
where I(q) is the experimental scattered intensity, q is
the scattering vector, k is an electron density contrast
factor, N is the number of particles, P(q) is the single
particle scattering function, and S(q) is the interparticle
scattering function.26 Assuming that the sample solu-
tions of HPCLs and LPCL (Cpolymer ) 1 mg/mL) were
very dilute and in a noninteracting system, the inter-
particle scattering effect can be ignored (i.e., S(q) f 1),
providing scattering information solely from the particle
scattering functions. The most common method for
determining the radius of gyration, Rg, from scattering
data is classified into the graphical techniques known
as the Zimm27 and Guinier methods, which involve the
28
F igu r e 5. Small-angle X-ray scattering (SAXS) profiles of
hyperbranched poly(ꢀ-caprolactone) (HPCL-10): Zimm plot (a)
and Guinier plot (b). The dashed lines indicate the fits to eqs
2
plot of either 1/I(q) or ln I(q) vs q . Figure 5 is the typical
scattering curves for HPCLs observed by SAXS, al-
though those only for HPCL-10 was shown here; the
1
4 and 15, respectively.
2
Zimm plot (a) is represented as 1/I(q) vs q and the
2
Guinier plot (b) as ln I(q) vs q . For the Zimm plot in
are only available for higher q values, the determination
2
5,26
Figure 5a, the HPCL-10 appeared linear over a rela-
tively large q range, while the HPCL-10 for the Guinier
plot had extremely strong curvature near the origin as
shown in Figure 5b. When data for qRg , 1 are
available, Rg can be easily extracted from the initial
slope of the curve at q ) 0 in Zimm and/or Guinier
plot, with all of these methods giving equivalent Rg
values in this limit. However, when experimental data
of Rg becomes rather model dependent.
With regard
to the SAXS experiments in this work, the entire
accessible q range varied from qmin ) ∼0.036 to ∼0.177,
and the region of the particle scattering term varied
from qminRg ) ∼1.6 for the smallest particles to ∼1.9
for the largest, which means a simple estimation of Rg
based on expansions about q ) 0 may be inappropriate.
An alternative approach to extract Rg values from

8
636 Choi and Kwak
Macromolecules, Vol. 36, No. 23, 2003
Ta ble 3. Ra d iu s of Gyr a tion (Rg), Br a n ch in g Ra tio (g),
Meltin g (Tm ) a n d Degr a d a tion (Td 10) Tem p er a tu r e,
En th a lp y of Meltin g (∆Hm ), a n d Degr ee of Cr ysta llin ity
(
Xc) of Hyp er br a n ch ed P oly(E-ca p r ola cton e)s (HP CLs)
a n d Lin ea r P oly(E-ca p r ola cton e) (LP CL)
Rg
(nm)
Tm
(°C)
∆Hm
Xc
(%)
Td10
a
b
b
(°C)c
sample
g
(J /g)
HPCL-5
HPCL-10
HPCL-20
LPCL
4.57
4.77
5.18
5.23
0.76
0.83
0.98
1
44
49
52
57
67.7
75.4
78.8
80.2
49.6
55.3
57.8
58.8
354
348
342
331
a
Determined from SAXS curves fit by the Zimm scattering
b
function. DSC melting transition observed on second heating.
Temperature of 10% weight loss.
c
scattering curves obtained over the relatively larger
qRg region is performing weighted nonlinear, least-
squares fits to the scattering curves over the q range,
F igu r e 6. DSC thermograms of hyperbranched poly(ꢀ-capro-
lactone)s and linear poly(ꢀ-caprolactone).
[qmin, qmax]:
I(q) ) AP(R ,q) + B, [qmin, qmax
]
(13)
g
Previous results by Prosa et al.26 revealed that the
scattering curves of hyperbranched polymers are fit
much better by particle scattering function suggested
by Zimm, PZimm(q), while those of dendrimers are fit
better by Guinier function, PGuinier(q). Particle scatter-
ing functions suggested by Zimm and Guinier are as
follows:
2
2
q R
g
PZimm(R ,q) ) 1/ 1 +
(14)
(15)
g
(
)
3
2
2
q R
g
PGuinier(R ,q) ) exp -
g
(
)
3
As shown in Figure 4, the Zimm fit was more accurate
appearing linear over the larger q range than the
Guinier fit for the HPCLs, which was well agreeable
with the previous results by Prosa et al. Summarized
in Table 3 are the Rg values of all three HPCLs and
LPCL determined by curve fitting of the Zimm plots
with PZimm as well as branching ratios, g’s, subsequently
determined from the mean-square values of Rg. The g
values were shown to decrease in the order of HPCL-
F igu r e 7. TGA thermograms of hyperbranched poly(ꢀ-capro-
lactone)s and linear poly(ꢀ-caprolactone).
LPCL were determined from the enthalpy of melting
using the following equation:
∆
Hm
X (%) )
× 100
(16)
c
0
∆
Hm
2
0, -10, and -5 from the value of 1.0 for LPCL, indicating
that the relative degree of branching increased as the
length of linear oligo(ꢀ-caprolactone) segments incorpo-
rated in the backbone chains became shortened. These
are in accordance with the predictions from the value
where ∆Hm is the apparent enthalpy of melting of each
0
samples and ∆Hm is the extrapolated value of enthalpy
corresponding to the melting of a 100% crystalline poly-
of 〈NAB 〉 in Table 1 because the HPCLs obtained from
2
(
1
ꢀ-caprolactone), which was previously reported to be
36.4 J /g. As listed in Table 3, Xc’s decreased, and thus
the smaller macromonomers have more numbers of
intrinsic branching points when all the HPCLs have
similar molecular weights.
3
0
the amorphous fraction in the samples increased, with
increase in the degree of branching. However, the
temperatures of 10% weight loss, Td10’s, of HPCLs and
LPCL, determined from TGA thermograms in Figure 7
and summarized in Table 3, indicated a reverse trend
compared to the Tm’s. This observation in increase of
thermal stability for the HPCLs relative to the LPCL
has motivated our further study in that determination
of the polymer chain mobility may correlate with and
explain the increase of thermal stability endowed by
branching as well as how the degree of branching may
affect the crystallization behavior of the HPCLs in
conjunction with the thermal degradation, which will
be of the near future.
As the length of linear oligo(ꢀ-caprolactone) segments
affected considerably the degree of branching in the
resulting HPCLs, their thermal transition and degrada-
tion characterized with DSC and TGA were found to be
also dependent on this structural variation. Figure 6
represents the DSC thermograms of HPCLs and LPCL
with appearance in sharp melting transition peaks. The
melting temperatures, Tm’s, of HPCLs are all lower than
that of LPCL and gradually decreased as the degree of
branching became increased, which is expected because
the presence of branches in a polymer may render
2
9
crystallization more difficult than in a linear polymer.
The degrees of crystallinity (Xc’s) of HPCLs as well as

Macromolecules, Vol. 36, No. 23, 2003
Hyperbranched Poly(ꢀ-caprolactone)s 8637
Con clu sion
valuable discussion about this work during his sabbati-
cal years.
1
. A series of hyperbranched poly(ꢀ-caprolactone)s
(
HPCLs) with homologous structural variation in the
Refer en ces a n d Notes
backbone chains were synthesized through moisture-
sensitive catalyst-free polycondensation with continuous
water removal of AB2 macromonomers, 2,2-bis[ω-hy-
droxy oligo(ꢀ-caprolactone)methyl] propionic acids, hav-
ing different lengths of oligomeric ꢀ-caprolactone seg-
ments. The targeted structural variation, which was the
oligomeric segments consisted of 5, 10, and 20 ꢀ-capro-
lactone monomer units, were confirmed by calculating
the average number of ꢀ-caprolactone units present AB2
(1) Newkome, G. R.; Moorefield, C. N.; V o¨ gtle, F. Dendritic
Molecules: Concepts, Syntheses, Perspectives; VCH Publish-
ers: New York, 1996; Chapter 6.
(2) Tomalia, D. A.; Naylor, A. M.; Goddard, W. A. G., III. Angew.
Chem., Int. Ed. Engl. 1990, 29, 138.
(3) Voit, B. I. Acta Polym. 1995, 46, 87.
(4) Fr e´ chet, J . M. J .; Hawker, C. J .; Gitsov, I.; Leon, J . W. J .
Macromol. Sci., Pure Appl. Chem. 1996, A33, 1399.
(5) Malmstr o¨ m, E.; Hult, A. J . Macromol. Sci., Rev. Macromol.
Chem. Phys. 1997, C37, 555.
1
macromonomer, 〈Nꢀ-CL〉, from H NMR spectra.
(6) Newkome, G. R. Advances in Dendritic Molecules; J AI
Press: Greenwich, CT, 1994; Vol. 1.
2
. The molecular weights of the resulting HPCLs with
respect to their linear counterpart, LPCL, were deter-
mined by (1) end-group analysis on H NMR spectra to
calculate the average number of macromonomer units
incorporated in HPCL, 〈NAB 〉, and (2) size exclusion
(7) Wooley, K. L.; Fr e´ chet, J . M. J .; Hawker, C. J . Polymer 1994,
35, 4489.
1
(
(
8) Trollsås, M.; Hedrick, J . L. Macromolecules 1998, 31, 4390.
9) Skaria, S.; Smet, M.; Frey, H. Macromol. Rapid Commun.
2002, 23, 292.
2
chromatography capable of a multiangle laser light
scattering (SEC-MALLS); the latter was also compared
with a conventional SEC results. The absolute number-
average molecular weights, Mn’s of HPCLs both from
(10) Kwak, S.-Y.; Lee, H. Y. Macromolecules 2000, 33, 5536.
(
(
11) Kwak, S.-Y.; Ahn, D. U. Macromolecules 2000, 33, 7557.
12) Hawker, C. J .; Lee, R.; Fr e´ chet, J . M. J . J . Am. Chem. Soc.
1
991, 113, 4583.
(
13) Bolton, D. H.; Wooley, K. L. Macromolecules 1997, 30, 1890.
(
1) and (2) were reasonably coincident in the range
between ca. 11 000 and 15 000.
. Small-angle X-ray scattering (SAXS) enabled to
(14) Roovers, J . In Encyclopedia of Polymer Science and Engineer-
ing; Mark, H. F., Bikales, N. M., Overberger, C. G., Menges,
G., Eds.; Wiley: New York, 1985; Vol. 2, p 478.
3
(
(
(
(
15) Zimm, B. H.; Stockmayer, W. H. J . Chem. Phys. 1949, 17,
estimate the radii of gyration, Rg, of HPCLs and LPCL,
1
301.
particularly by curve fits of the Zimm plot represented
16) Ihre, J .; Hult, A.; S o¨ derlind, E. J . Am. Chem. Soc. 1996, 118,
6388.
17) Trollsås, M.; Hedrick, J . L.; Mecerreyes, D.; Dubois, Ph.;
J e´ r oˆ me, R.; Ihre, H.; Hult, A. Macromolecules 1997, 30, 8508.
18) B u¨ y u¨ ktanir, E.; K u¨ c¸ u¨ kyavuz, Z. J . Polym. Sci., Part B:
Polym. Phys. 2000, 38, 2678.
2
as 1/I(q) vs q with the Zimm scattering function. Then,
the branching ratio, g, of the mean-square value of Rg
for HPCL to that for LPCL was used to characterize the
relative degree of branching (DB) of HPCLs relative to
LPCL. The DBs of HPCLs were found to be considerably
influenced by the length of the linear olig(ꢀ-caprolac-
tone) segments repeated in the backbone chains; the
shorter the olig(ꢀ-caprolactone) segments, the higher the
DB in the order HPCL-5 > HPCL-10 > HPCL-20. Also
found were the presence of any correlative effect of DB
on the thermal transition and degradation. The melting
transitions of HPCLs were elevated in accordance with
decrease in the number of oligo(ꢀ-caprolactone) seg-
ments and hence increase in the degree of branching,
while vice versa for the thermal degradation.
(19) Huglin, M. B. Light Scattering from Polymer Solutions;
Academic: London, 1972.
20) Malmstr o¨ m, E.; J ohansson, M.; Hult, A. Macromolecules
(
1
995, 28, 1698.
(
(
21) Flory, P. J . J . Am. Chem. Soc. 1952, 75, 2718.
22) Podzimek, S. J . Appl. Polym. Sci. 1994, 54, 91.
(23) Claesson, H.; Malmstr o¨ m, E.; J ohansson, M.; Hult, A. Polymer
002, 43, 3511.
2
(
24) Yoshikawa, K.; Ofuji, N.; Imaizumi, M.; Moteki, Y.; Fujimaki,
T. Polymer 1996, 37, 1281.
(25) Nakamura, Y.; Wan, Y.; Mays, J . W. Iatrou, H.; Hadjichris-
tidis, N. Macromolecules 2000, 33, 8323.
(26) Prosa, T. J .; Bauer, B. J .; Amis, E. J .; Tomalia, D. A.;
Scherrenberg, R. J . Polym. Sci., Part B: Polym. Phys. 1997,
3
5, 2913.
Ack n ow led gm en t. The authors are grateful to the
Ministry of Environment, Republic of Korea, for their
support of this study through the Eco-Technopia 21
project. They also acknowledge Hyperstructured Or-
ganic Materials Research Center (HOMRC) at Seoul
National University, Republic of Korea, for the use of
SEC-MALLS and SAXS equipments. The corresponding
author especially appreciates Professor Colin A. Fyfe at
The University of British Columbia, Canada, for the
(
27) Higgins, J . S.; Beno ´ı t, H. C. Polymers and Neutron Scattering;
Clarendon Press: Oxford, 1994.
(28) Guinier, A.; Fournet, G. Small Angle Scattering of X-Rays;
Wiley: New York, 1955.
29) Magnusson, H.; Malmstr o¨ m, E.; Hult, A.; J ohansson, M.
Polymer 2002, 43, 301.
30) Crescenzi, V.; Manzini, G.; Calzolari, G.; Borri, C. Eur. Polym.
J . 1972, 8, 449.
(
(
MA034170I