Degradation kinetics of model hyperbranched chains with uniform subchains and controlled locations of cleavable disulfide linkages

Article abstract of DOI:10.1021/ma402438m

We developed a strategy to make model hyperbranched structure with uniform subchains and controlled locations of cleavable linkages. First, a novel seesaw-type tetrafunctional initiator with one alkyne, one disulfide linkage, and two bromine groups (≡-S-S-(Br)₂) was prepared. Using such an initiator, an AB₂-type macromonomer (azide～～alkyne～ ～azide) with one disulfide linkage at its center was prepared via successive atom transfer radical polymerization (ATRP) and azidation substitution reaction, where ～～ represents polystyrene chains. Further interchain "clicking" coupling between the azide and alkyne groups on the macromonomers led to model hyperbranched polystyrenes with uniform subchains and controllablly located cleavable disulfide linkages. The ¹H nuclear magnetic resonance spectra, Fourier transform infrared spectroscopy, and size exclusion chromatography with a multiangle laser light scattering detector confirmed the designed degradable hyperbranched structure. Armed with this novel sample, we studied its dithiothreitol (DTT)-induced degradation in various organic solvents by a combination of static and dynamic LLS. We found that the cleavage of disulfide bonds contains a fast and a slow process. The fast one reflects the degradation of disulfide bonds on the chain periphery; while the slow one involves those inside. Both the fast and slow degradation reaction rate constants (K_fast and K_slow) are a linear function of the initial DTT concentration ([DTT]₀), but the relative contribution of the two processes is mainly governed by the hyperbranched chain structure, nearly independent of [DTT]₀.

Full text of DOI:10.1021/ma402438m

Article
pubs.acs.org/Macromolecules
Degradation Kinetics of Model Hyperbranched Chains with Uniform
Subchains and Controlled Locations of Cleavable Disulfide Linkages
Lianwei Li,*,† Xu Wang,† Jinxian Yang,† Xiaodong Ye,*,† and Chi Wu†,‡
†Hefei National Laboratory for Physical Sciences at the Microscale, Department of Chemical Physics, University of Science and
Technology of China, Hefei, Anhui 230026, China
‡Department of Chemistry, The Chinese University of Hong Kong, Shatin N. T., Hong Kong
ABSTRACT: We developed a strategy to make model
hyperbranched structure with uniform subchains and con-
trolled locations of cleavable linkages. First, a novel seesaw-
type tetrafunctional initiator with one alkyne, one disulfide
linkage, and two bromine groups (≡−S−S−(Br)2) was
prepared. Using such an initiator, an AB2-type macromonomer
(azide∼∼alkyne∼∼azide) with one disulfide linkage at its
center was prepared via successive atom transfer radical
polymerization (ATRP) and azidation substitution reaction,
where ∼∼ represents polystyrene chains. Further interchain
“clicking” coupling between the azide and alkyne groups on the
macromonomers led to model hyperbranched polystyrenes
with uniform subchains and controllablly located cleavable
1
disulfide linkages. The H nuclear magnetic resonance spectra, Fourier transform infrared spectroscopy, and size exclusion
chromatography with a multiangle laser light scattering detector confirmed the designed degradable hyperbranched structure.
Armed with this novel sample, we studied its dithiothreitol (DTT)-induced degradation in various organic solvents by a
combination of static and dynamic LLS. We found that the cleavage of disulfide bonds contains a fast and a slow process. The fast
one reflects the degradation of disulfide bonds on the chain periphery; while the slow one involves those inside. Both the fast and
slow degradation reaction rate constants (Kfast and Kslow) are a linear function of the initial DTT concentration ([DTT]0), but the
relative contribution of the two processes is mainly governed by the hyperbranched chain structure, nearly independent of
[DTT]0.
On the other hand, it is well-known that disulfide bonds can
be cleaved into thiol groups in the presence of various reducing
agents,19 such as thiols,20 phosphines,21,22 and zinc dust;23 and
the resulting thiol groups can reversibly re-form disulfide bonds
upon oxidation. Therefore, the thiol−disulfide exchange
reaction has been widely used in various applications, such as
drug delivery,24 gene transfection,25 and construction of self-
healing materials.26,27 Meanwhile, a number of disulfide-
functionalized polymer structures have been prepared, namely,
linear (diblock24,28−30 and multiblock31,32), cyclic,31 comb-
like,33 hyperbranched,25,34−38 and cross-linking39−41 structures.
Generally, two main approaches have been used to introduce
disulfide linkage into the polymer backbone: (1) homopolyme-
rization of star precursors with thiol ends (Scheme 1A)31,37 or
(2) copolymerization of star precursors with disulfide-function-
alized linear small molecules (Scheme 1B).25,34−36 However,
the distributions of the subchain length and disulfide bond
location of hyperbranched polymers prepared by these two
approaches will be only partly controllable (star precur-
INTRODUCTION
■
In the past two decades, various nonideal hyperbranched
structures with uncontrollable branching subchains have been
extensively prepared and used for the structure−property
investigation.1−7 This is due to the fact that the preparation of
an ideal model hyperbranched structure with uniform and
adjustable subchains between any two neighboring branching
points is rather difficult because of synthetic methodology
limitation,8,9 which seriously hinders the relevant structure−
property study.10,11 Recently, using a novel seesaw-type
macromonomer B∼∼A∼∼B as the precursor in self-poly-
condensation reaction, where the reactive A and the two B
groups are located at the chain center and the two chain ends,
respectively,12−14 we have successfully prepared model hyper-
branched chains. Moreover, we have experimentally, for the first
time, established the scaling laws between their sizes (R),
intrinsic viscosities ([η]), and molar masses (M);13,15 found
how these chains pass through a small cylindrical nanopore
under an elongational flow field;16,17 and prepared amphiphilic
hyperbranched block copolymer chains and studied their
interchain and intrachain association in dilute and semidilute
solutions.18
Received: November 28, 2013
Revised: December 27, 2013
Published: January 6, 2014
© 2014 American Chemical Society
650
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
Scheme 1. Schematic of Topological Structures of Disulfide-Functionalized Hyperbranched Chains Prepared by Two Different
Approaches: (A) Homopolymerization of Telechelic Star Precursors with Thiol End Groups and (B) Copolymerization of
Disulfide-Functionalized Small Molecules and Star Chains
sors).25,33,34,36,42,43 To the best of our knowledge, it remains a
Scheme 3. Schematic of Synthesis and Degradation of
challenge to prepare model hyperbranched polymers with
uniform subchains and controlled locations of cleavable
disulfide linkages simultaneously.
Disulfide-Functional Hyperbranched Polystyrene Chains
HB−(S−S−PS)n
In the current study, applying our seesaw-type macro-
monomer stragtegy, we have successfully synthesized model
hyperbranched polystyrene chains with uniform subchains and
controllablly located cleavable disulfide linkages. As shown in
Scheme 2, seesaw-type macromonomer ≡−S−S−(PS−N3)2
Scheme 2. Schematic of Synthesis of Disulfide-
Functionalized Seesaw-Type Initiator and Linear
Polystyrene Macromonomer ≡−S−S−(PS−N3)2
EXPERIMENTAL SECTION
■
Materials. Styrene (Sinopharm, 97%) was purified by first passing
it through a basic alumina column and then distilling under a reduced
pressure over calcium hydride (CaH2). Dimethylformamide (DMF,
Sinopharm, AR) was first dried with anhydrous magnesium sulfate and
then distilled under a reduced pressure. N,N,N′,N′,N″-Pentamethyl-
diethylenetriamine (PMDETA, Aldrich, 99%), tetrahydrofuran (THF),
dichloromethane (DCM), and triethylamine (TEA) were distilled over
CaH2. DMF, THF, and DCM were bubbled with nitrogen for 5 h
before used. Dithiothreitol (DTT, Aladdin, 99%) was stored under
nitrogen atmosphere at 2−6 °C. All the other reagents from
Sinopharm or Aldrich were directly used as received unless otherwise
stated.
Characterization. 1H nuclear magnetic resonance (1H NMR)
spectra were recorded using a Bruker AV400 spectrometer with
tetramethylsilane (TMS) as an internal standard and deuterated
chloroform (CDCl3) as solvent. Fourier transform infrared spectra
(FTIR) were recorded using a Bruker VECTOR-22 IR spectrometer,
and each spectrum was collected over 64 scans with a spectral
resolution of 4 cm−1.
Size Exclusion Chromatography (SEC). The relative number-
and weight-average molar masses (Mn,RI and Mw,RI) and the absolute
number- and weight-average molar masses (Mn.MALLS and Mw,MALLS)
with one disulfide linkage at the chain center was first prepared
by atom transfer radical polymerization (ATRP) and azidation
substitution reaction. Further interchain “clicking” coupling of
macromonomers ≡−S−S−(PS−N3)2 led to disulfide-function-
alized hyperbranched polystyrene chains (Scheme 3). The well-
defined structures of macromonomer and hyperbranched
product were confirmed by various kinds of methods. Finally,
the degradation kinetics of hyperbranched polystyrene chains in
various organic solvents was monitored by a combination of
static and dynamic laser light scattering.
were determined at 35 °C by size exclusion chromatography (SEC,
Waters 1515) equipped with three Waters Styragel columns (HR2,
HR4, and HR6), a refractive index detector (RI, Wyatt WREX-02),
and a multiangle laser light scattering detector (MALLS, Wyatt
DAWN EOS). THF was used as the eluent at a flow rate of 1.0 mL/
min, and linear narrowly distributed polystyrenes were used as
standards.
Laser Light Scattering (LLS). A commercial LLS spectrometer
(ALV/DLS/SLS-5022F) equipped with a multi-τ digital time
correlator (ALV5000) and a cylindrical 22 mW UNIPHASE He−Ne
651
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
laser (λ0 = 632.8 nm) was used as the light source. In static LLS,44,45
the angular dependence of the absolute excess time-average scattering
intensity, known as the Rayleigh ratio RVV(q), can lead to the weight-
average molar mass (Mw), the root-mean-square gyration radius
(⟨Rg⟩), and the second virial coefficient A2 by using
⎛
⎜
⎞
KC
VV(q)
1
Mw
1
3
2
2
⎟
≅
1 + ⟨Rg ⟩Zq + 2A2C
⎝
⎠
R
(1)
4
where K = 4π2n2(dn/dC)2/(NAλ0 ) and q = (4πn/λ0) sin(θ/2) with n,
C, dn/dC, NA, and λ0 being the refractive index, the concentration of
the polymer solution, the specific refractive index increment,
Avogadro’s number, and the wavelength of light in a vacuum,
respectively. The dn/dC value of polystyrene in THF is 0.186 mL/
g.12,13
In dynamic LLS,46 the Laplace inversion of each measured
intensity−intensity time correlation function G(2)(q,t) in the self-
beating mode can result in a line-width distribution G(Γ). G(Γ) can be
converted to a translational diffusion coefficient distribution G(D) or
further to a hydrodynamic radius distribution f(Rh) via the Stokes−
Einstein equation, Rh = (kBT/6πη0)/D, where kB, T, and η0 are the
Boltzmann constant, the absolute temperature, and the solvent
viscosity, respectively, by using both the cumulants and CONTIN
analysis.
Synthesis of Intermediate Compound I. The general procedure
was outlined as follows. 30.0 g of 2,2-bis(hydroxymethyl)propionic
acid (223.7 mmol), 41.4 mL of 2,2-dimethoxypropane (DMP) (335.4
mmol), and 2.1 g of p-toluenesulfonic acid monohydrate (p-TSA, 11.1
mmol) were dissolved in 100 mL of acetone. The mixture was stirred
for 4 h at room temperature before a 3.0 mL mixture of an ammonia
aqueous solution (25%) and ethanol (50/50, v/v)) was added into the
reaction mixture to neutralize the catalyst. The solvent was removed
by evaporation under a reduced pressure at room temperature. The
residue was then dissolved in DCM (600 mL) and extracted with two
portions of water (80 mL). The organic phase was dried with
anhydrous Na2SO4 and then evaporated to give white powder (33.0 g,
47
1
84%). Its H NMR spectrum in CDCl3 can be found elsewhere.
Synthesis of Intermediate Compound II. The general procedure
was outlined as follows. 20.0 g of 2-hydroxyethyl disulfide (130 mmol)
and 3.6 g of propargyl bromide (65 mmol) were mixed in 200 mL of
THF. 3.0 g of NaH powder (80 wt %, 98 mmol) was added into the
reaction mixture in three successive batches under nitrogen within 3 h
at 0 °C. The mixture was further stirred for another 7 h at room
temperature before a few drops of water were added to stop the
reaction. The mixture was filtered, and the solvent was removed by
evaporation under a reduced pressure at 50 °C. The crude product was
purfied by column chromatography to give a pale-yellow clear oil: 6.5 g
Figure 1. 1H NMR spectra of the intermediate compounds (II−IV) in
CDCl3.
Synthesis of Intermediate Compound IV. The general procedure
was outlined as follows. 8.0 g of compound III (25.0 mmol) dissolved
in 100 mL of methanol was added with an excess of activated acidic
styrene cation exchange resin (Amberlite IR-120, aladdin). The
mixture was stirred for 4 h at room temperature. After the resin was
filtered, the solvent was removed by evaporation. The crude product
was purified by column chromatography to give a pale-yellow oil: 7.1 g
1
1
(52%). Figure 1A shows its H NMR spectrum in CDCl3. H NMR
(CDCl3, ppm): δ 4.20 (d, 2H, CHCCH2O−), 3.90 (t, 2H, CH
CCH2OCH2−), 3.81 (t, 2H, −CH2OH), 2.93 (t, 2H, CH
CCH2OCH2CH2−), 2.88 (t, 2H, −CH2CH2OH), 2.46 (t, 1H,
CHCCH2−), 1.93 (s, 1H, −CH2OH).
1
1
(89%). Figure 1C shows its H NMR spectrum in CDCl3. H NMR
(CDCl3, ppm): δ 4.45 (t, 2H, CHCCH2O−), 4.20 (t, 2H,
−CH2OOC−), 3.88 (d, 2H, CHCCH2OCH2−), 3.75 (m, 4H,
HOCH2−), 2.95 (m, 4H, −CH2SSCH2−), 2.46 (t, 1H, CH
CCH2−), 1.11 (s, 3H, CH3−).
Synthesis of Intermediate Compound III. The general procedure
was outlined as follows. 5.4 g of compound I (31.3 mmol), 6.0 g of
compound II (31.3 mmol), and 0.38 g of 4-(dimethylamino)pyridine
(DMAP) (3.1 mmol) were dissolved in 100 mL of DCM. 7.74 g of
N,N′-dicyclohexylcarbodiimide (DCC) (37.6 mmol) was dissolved in
50 mL of DCM. The two solutions were mixed under nitrogen flow at
0 °C within 0.5 h and further stirred overnight at room temperature.
The byproduct N,N′-dicyclohexylurea was filtered. The solution was
then diluted by 100 mL of DCM and extracted with two portions of
water (150 mL). The organic phase was dried with anhydrous sodium
sulfate. After removing the solvent by evaporation, the crude product
was purified by column chromatography to give a pale-yellow oil: 8.0 g
Synthesis of Compound V, Seesaw-Type Initiator. The general
procedure was outlined as follows. 6.5 g of compound IV (21.1 mmol)
and 2.6 g of TEA (25.3 mmol) were dissolved in 150 mL of DCM.
11.6 g of α-bromoisobutyryl bromide (50.6 mmol) dissolved in 50 mL
of DCM was added into the reaction mixture under nitrogen flow at 0
°C within 0.5 h. The mixture was stirred overnight. After filtration, the
mixture was diluted with 100 mL of DCM and extracted with 50 mL of
saturated sodium bicarbonate aqueous solution. The organic phase was
dried with anhydrous sodium sulfate. After removing the solvent by
evaporation, the crude product was purified by column chromatog-
1
1
(83%). Figure 1B shows its H NMR spectrum in CDCl3. H NMR
(CDCl3, ppm): δ 4.43 (t, 2H, CHCCH2O−), 4.20 (m, 4H,
(CH3)2C(OCH2−)), 3.79 (t, 2H, −CH2OOC−), 3.66 (m, 2H, CH
CCH2OCH2−), 2.96 (m, 4H, −CH2SSCH2−), 2.46 (t, 1H, CH
CCH2−), 1.40 (m, 6H, (CH3)2C(OCH2−)), 1.21 (s, 3H, CH3−).
1
raphy to give a dark-yellow oil: 9.0 g (71%). Figure 2 shows its H
1
NMR spectrum in CDCl3. H NMR (CDCl3, ppm): δ 4.43 (m, 4H,
(COOCH2)2CCH3−), 4.38 (t, 2H, CHCCH2O−), 4.20 (t, 2H,
−CH2OOCSS−), 3.80 (m, 2H, CHCCH2OCH2−), 2.95 (m, 4H,
652
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
removed by evaporation, and the residue was dissolved in THF and
precipitated into an excess of methanol. After drying in a vacuum oven
overnight at 40 °C, hyperbranched polymer HB-(S−S−PS)n was
obtained. The resultant polydispersed HB-(S−S−PS)n was further
purified by fractional precipitation in a mixed solution of toluene/
methanol to remove most of the low-molar-mass fractions. Finally,
hyperbranched HB-(S−S−PS)n fraction with a Mw of ∼3.7 × 105 g/
mol (n ≃ 35) was obtained and used in the degradation kinetics study.
DTT-Induced Degradation of HB-(S−S−PS)35 Chains. The DTT-
induced degradation of HB-(S−S−PS)35 was conducted in situ inside
the LLS cuvette. Stock DMF solutions of HB-(S−S−PS)35 (1.0 g/L)
and DTT (25.0 g/L) were purged by nitrogen to replace oxygen and
respectively clarified with 0.45 μm Millipore PTFE filters to remove
dust. In a typical experiment, a proper amount of dust-free DTT DMF
solution was rapidly added into 3.0 mL of dust-free HB-(S−S−PS)35
DMF solution to start the degradation. Both the scattered intensity
and G(2)(q,t) were recorded during the degradation for at least 1000
min.
1
Figure 2. H NMR spectrum of disulfide-functionalized initiator.
−CH2SSCH2−), 2.48 (t, 1H, CHCCH2−), 1.98 (d, 12H,
BrC(CH3)2−), 1.36 (s, 3H, CH3−).
RESULTS AND DISCUSSION
Preparation of Macromonomer ≡−S−S−(PS−Br)2 by ATRP. The
general procedure was outlined as follows. A three-necked flask
equipped with a magnetic stirring bar and three rubber septa was
charged with 0.4 g of initiator V (0.66 mmol), 12.1 mL of styrene
(105.9 mmol), and 124 μL of PMDETA (0.66 mmol). The flask was
degassed by three freeze−pump−thaw cycles and then placed in an oil
bath thermostated at 90 °C. After ∼2 min, 93.7 mg of CuBr (0.66
mmol) was added under nitrogen flow to start the polymerization.
Samples were withdrawn at different times during the polymerization
for SEC and monomer conversion measurements by the weighing
method. Namely, the conversion was calculated as conversion = m/m0,
where m0 and m represent the weights of sample solution before and
after removing styrene monomer by evaporation under a reduced
pressure at 50 °C for four cycles; in each cycle, a certain amount of
THF was added to enhance the extraction of styrene. After a few
hours, the flask was rapidly cooled in liquid nitrogen. The polymer
mixture was diluted with THF and passed through a short column of
neutral alumina to remove metal salt. The solvent was removed by
evaporation, and the residue was dissolved in THF and precipitated
into an excess of methanol and recovered by filtration. The above
purification cycle was repeated twice. After drying in a vacuum oven
overnight at 40 °C, macromonomer ≡−S−S−(PS−Br)2 was obtained.
Yield: 6.15 g (54%). Mn: 9.4 × 103 g/mol and Mw/Mn: 1.14 (by SEC);
■
As shown in Scheme 2, disulfide-functional initiator (V) was
prepared through a five-step synthesis. All the intermediate
compounds (II−IV) were carefully purified by column
chromatography, and their high purities were confirmed by
1H NMR characterization (Figure 1A−C).
Figure 2 shows the 1H NMR spectrum of the tetrafunctional
initiator, and the integral area ratio of peak (a), peak (i), and
peak (d + e) is 1.00/3.10/4.09, in consistency with its
theoretical ratio of 1.00/3.00/4.00. Using this functional
initiator, seesaw-type polystyrene macromonomer ≡−S−S−
(PS−Br)2 was successfully prepared by ATRP, and its typical
1H NMR spectrum is shown in Figure 3.
Mn: 9.0 × 103 g/mol (by H NMR in Figure 2).
1
Preparation of Macromonomer ≡−S−S−(PS−N3)2 by Azidation.
The general procedure was outlined as follows. A 100 mL round-
bottom flask was charged with 5.0 g of ≡−S−S−(PS−Br)2 (0.45
mmol), 40 mL of DMF, and 0.29 g of sodium azide (4.5 mmol). The
mixture was allowed to stir at room temperature for 24 h under
nitrogen flow. After removing DMF under a reduced pressure, the
remaining portion was diluted with DCM and passed through a
neutral alumina column to remove residual sodium salts. The solvent
was removed by evaporation, and the residue was dissolved in THF
and precipitated into an excess of methanol. After drying in a vacuum
oven overnight at 40 °C, azido-terminated polymer was obtained and
named as ≡−S−S−(PS−N3)2. Yield: 4.7 g (94%). Mn: 9.3 × 103 g/
mol and Mw/Mn: 1.12 (by SEC).
Preparation of Degradable Hyperbranched Polymer HB-(S−S−
PS)n by “Click” Chemistry. The general procedure was outlined as
follows. A 25 mL three-necked flask equipped with a magnetic stirring
bar and three rubber septa was charged with 4.5 g of macromonomer
≡−S−S−(PS−N3)2 (∼0.43 mmol, Mw,macromonomer = 1.04 × 104 g/
mol), 83 μL of PMDETA (0.40 mmol), and 15.0 mL of DMF. The
flask was degassed by three freeze−pump−thaw cycles and then placed
in a water bath thermostated at 35 °C. After ∼2 min, 57.5 mg of CuBr
(0.40 mmol) was added under nitrogen flow to start the polymer-
ization. Samples were withdrawn at different times and precipitated
into a mixture of methanol/water (90/10, v/v) for SEC measure-
ments. After 24 h, the flask was cooled in liquid nitrogen. The polymer
mixture was diluted with THF and passed through a short column of
neutral alumina for the removal of metal salt. The solvent was
Figure 3. 1H NMR spectrum of disulfide-functionalized ≡−S−S−
(PS−Br)2.
Figures 4−6 show how the monomer conversion, number-
average molar mass (Mn), polydispersity index (Mw/Mn), and
SEC curves of polystyrene evolved during the ATRP process.
As shown in Figure 4, the plot of semilogarithmic of ln(1/(1 −
conversion)) versus t presents a linear relation, indicating the
polymerization meets the required criteria of a living system.
Figure 5 shows that the experimental Mn increases linearly with
the conversion, and the Mw/Mn rapidly decreases with the
conversion and reaches a minimum of ∼1.14 at the end of
polymerization, indicating the whole ATRP process is well-
controlled, which is also reflected in the symmetric and
narrowly distributed elution curves (Figure 6).
In a typical ATRP process, the molar ratio of CuBr to
halogen atom on initiator is normally 1.0/1.0, while the ratio in
653
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
Figure 4. (A) Kinetic plot of semilogarithmic of ln(1/(1 −
conversion)) versus reaction time (t) during the ATRP process.
Reaction conditions: [styrene]/[initiator]/[CuBr]/[PMDETA] =
150/1/1/1; T = 90 °C; bulk.
Figure 7. IR spectra of (A) ≡−S−S−(PS−Br)2, (B) ≡−S−S−(PS−
N3)2, and (C) HB-(S−S−PS)n.
polystyrene chains with uniform subchains and controlled
locations of cleavable disulfide linkages. Scheme 3 schematically
shows the structures of the prepared hyperbranched chains and
its degradation products.
The alkyne−azide “click” cycloaddition was used here
because of its high efficiency and low susceptibility to side
reactions.50−52 DMF was used as solvent in the interchain
“clicking” of macromonomer ≡−S−S−(PS−N3)2 after the
removal of oxygen at 35 °C with a catalyst of CuBr−PMDETA.
Similar to our previous study,12,13 we prefer to use the weight-
average polycondensation degree (DPw), instead of the
number-average polycondensation degree (DPn), to describe
the whole interchain “clicking” process because it is the weight-
average value that could be measured accurately in SEC-
MALLS or stand-alone static LLS, where DP is defined as the
number of macromonomers, not monomer, chemically coupled
together inside each hyperbranched polystyrene chain; namely,
DPw = Mw,hyperbranched/Mw,macromonomer (Mw,macromonomer  1.04 ×
104 g/mol).
Figure 5. Monomer conversion dependence of number-average molar
mass (Mn) and polydispersity index (Mw/Mn) of macromonomer ≡−
S−S−(PS−Br)2 during the ATRP process. Reaction conditions:
[styrene]/[initiator]/[CuBr]/[PMDETA] = 150/1/1/1; T = 90 °C;
bulk.
Figure 8A shows that the two elution curves respectively
from the RI and MALLS detectors are very different. This is
Figure 6. Reaction time (t)-dependent SEC curves of macromonomer
≡−S−S−(PS−Br)2 during the ATRP process. Reaction conditions:
[styrene]/[initiator]/[CuBr]/[PMDETA] = 150/1/1/1; T = 90 °C;
bulk.
our study is actually 1.0/0.5. In addition, the monomer
conversion was intentionally controlled below 60%, which is
because both the CuBr concentration and monomer conversion
are the key factors to attain polymer chains with high end-
group functionalities. In principle, lowering both the CuBr
concentration and monomer conversion can effectively reduce
the chain termination and other side reactions.48,49
Further azidation substitution reaction of ≡−S−S−(PS−Br)2
by sodium azide in DMF led to macromonomer ≡−S−S−
(PS−N3)2. The successful installation of azide groups was
reflected in the appearance of a NNN antisymmetric
stretching absorption band near ∼2090 cm−1 in FT-IR spectra
(Figure 7). Using this seesaw-type macromonomer ≡−S−S−
(PS−N3)2, we are able to prepare degradable hyperbranched
Figure 8. (A) Comparison of two typical SEC curves of HB-(S−S−
PS)n monitored by two different detectors: RI (solid line) and MALLS
(circle symbol) and (B) reaction time (t)-dependent SEC-RI curves of
HB-(S−S−PS)n, where T = 35 °C and Cpolymer = 0.30 g/mL.
654
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
because the scattered light intensity in LLS is proportional to
the product of the weight-average molar mass and the weight
concentration, while the signal from RI is only proportional to
the weight concentration. Figure 8B shows the reaction time-
dependent SEC curves of HB-(S−S−PS)n during the self-
polycondensation. As the reaction proceeds, the macro-
monomer peak becomes smaller while the hyperbranched
polymer peaks emerged after a few hours. Gradually, DPw and
DPn increase to ∼28 and ∼6, respectively, and its polydispersity
index (Mw/Mn) changes from 1.14 to 4.90 (Figure 9) when
Figure 10. SEC-RI curves of HB-(S−S−PS)35 chains before and after
24 h DTT-induced reduction in DMF at 25 °C, where [DTT]/[−S−
S−] = 18.
Figure 9. Reaction time (t) dependence of polydispersity index (Mw/
Mn), number (DPn), and weight-average degree of polycondensation
(DPw) of HB-(S−S−PS)n, where T = 35 °C and Cpolymer = 0.30 g/mL.
Figure 11. Hydrodynamic radius distributions [f(Rh)] of HB-(S−S−
PS)35 chains before and after 24 h DTT-induced reduction in DMF at
25 °C, where [DTT]/[−S−S−] = 18.
most of the initial macromonomers are coupled together,
reflecting in a sharp decrease of the NNN stretching
absorption near ∼2100 cm−1 (Figure 7C). The self-
polycondensation nearly ceases after ∼14 h, similar to what
we observed in a previous study.13 The related polycondensa-
tion kinetics analysis can be found elsewhere.13
As discussed before, each branching point of HB-(S−S−PS)n
contains just one disulfide bond that is cleavable when a
reducing agent is added. The cleavage of each disulfide bond
results in two smaller fragemented hyperbranched chains with a
thiol end group. Using a combination of static and dynamic
LLS, we studied such a degradation kinetics of a fractionated
HB-(S−S−PS)n with a DPw of 35 and a Mw/Mn of 1.40. There
are different choices of reducing reagents, including thiols,20
phosphines,21,22 and zinc dust.23 DTT was selected in the
current study because (1) its induced disulfide reduction is very
efficient owing to its high conformational propensity to form a
six-member ring with an internal disulfide bond and (2) it is
soluble in a range of organic solvents that can also dissolve
polystyrene easily.28
Figures 10 and 11 show how HB-(S−S−PS)35 chains change
after the 24 h DTT induced reduction in DMF at 25 °C. As
shown in Figure 10, the elution peak located at ∼24 mL
completely disappears after the degradation, and only the
macromonomer peak (maybe contains few dimers) remains,
indicating that the DTT induced degradation of the initial HB-
(S−S−PS)35 chains is nearly 100%. The residual dimers are
attributed to the limited degradation time (24 h). Such chain
degradation is also directly reflected in the evolution of the
hydrodynamic radius distribution [f(Rh)], as shown in Figure
11; i.e., the peak position shifts from ∼19 to ∼3 nm after
degradation. However, it should be noted that the degradation
products are similar to the initial seesaw-type linear macro-
monomers except different end groups, as shown in Scheme 3.
Figure 12 shows how the DPw, ⟨Rh⟩, and average chain
density (⟨ρ⟩) of HB-(S−S−PS)35 change during the degrada-
tion at different molar ratios of [DTT]/[−S−S−]; for the
convenience of discussion, DPw,t is used to represent the value
of DPw at time t. As expected, DPw,t and ⟨Rh⟩ gradually decrease
as the degradation (fragmentation) proceeds for a given
[DTT]/[−S−S−], and the degradation rate becomes higher
when more DTT is added. We tested whether all the disulfide
bonds can be completely degraded in the range of [DTT]/
[−S−S−] = 3−180 and found that most of the initial disulfide
bonds were cleaved within 5 h even at [DTT]/[−S−S−] = 3.
Figure 12C shows that the average chain density (⟨ρ⟩), defined
as Mw/[NA(4/3)π⟨Rh⟩3], increases with the degradation time,
where Mw, ⟨Rh⟩, and NA are the weight-average molar mass, the
average hydrodynamic radius of hyperbranched chains, and
Avogadro’s number, respectively, revealing that larger branched
chains have a lower chain density than its degraded products
and the initial linear macromonomer has the highest ⟨ρ⟩.
To analyze the results in Figure 12, and considering the
nature of thiol−disulfide exchanging reaction, we first assumed
that the degradation follows the second-order kinetics, i.e.
d[−S−S−]t
−
= k[−S−S−]t [DTT]
t
(2)
dt
where [−S−S−]t and [DTT]t are the molar concentrations of
disulfide bond and DTT at time t, respectively, and k is a
reaction rate constant. At t = 0, [−S−S−]t = [−S−S−]0, and at
t = ∞, [−S−S−]t = 0. Since DTT is excessive, we can treat
[DTT]t = [DTT]0 as a constant in eq 2 so that we can rewrite
it as a pseudo-first-order equation
[−S−S−]t
[−S−S−]0
t
= e−k[DTT]
0
(3)
655
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
Figure 13. DTT-induced degradation time (t) dependence of
normalized weight-average degree of polycondensation (DPw,t
/
DPw,0) of HB-(S−S−PS)35 in DMF at 25 °C, where dashed and
solid lines represent a single-exponential fitting (eq 4) and a double-
exponential fitting (eq 5), respectively.
degradation reactions, respectively; Afast + Aslow = 1.0. The
reasonable double-exponential fitting actually reveals that the
degradation of disulfide bonds on the periphery of the
hyperbranched chain is much faster than those inside.
Moreover, Figure 14A shows that the fast degradation
contributes twice more than the slow one; namely, ∼2/3 of the
disulfide bonds are on the periphery. It is interesting to find
that the contributions of the fast and slow degradations are
independent of the DTT concentration, presumably because
Figure 12. Degradation time (t) dependence of (A) weight-average
degree of polycondensation (DPw,t), (B) average hydrodynamic radius
(⟨Rh⟩), and (C) average chain density (⟨ρ⟩) of HB-(S−S−PS)35 in
DMF at 25 °C.
On the other hand, it is well-known that polystyrene is
insoluble in diols so that most of DTT are excluded from the
interior of HB-(S−S−PS)n. Therefore, those external disulfide
bonds would be cleaved first. As shown in Scheme 3, the
cleavage of each external disulfide bond leads to a decrease of
DPw by one. In this way, we are able to rewrite [−S−S−]t/
[−S−S−]0 as DPw,t/DPw,0 so that eq 3 can be rewritten as
DP = DP e−Kt
(4)
w,t
w,0
where DPw,t and DPw,0 are the degrees of polycondensation of
hyperbranched chains at t = t and t = 0, respectively, and K =
k[DTT]0.
However, Figure 13A shows that the single-exponential
fitting using eq 3 is not satisfactory, which forces us to include
the degradation of those disulfide bonds inside the hyper-
branched polystyrene chain. It is expected that the degradation
of those internal disulfide bonds would be much slower due to
(1) the lower concentration of DTT inside the polystyrene core
and (2) the steric effect of the polystyrene segments around
each internal disulfide bond. Empirically, we found that adding
a slow degradation process into eq 4 enables us to fit all the
curves in Figure 13B by using a double-exponential fitting as
follows:
DP
w,t
t
t
−Kfast
−Kslow
= A
e
+ A
e
slow
fast
Figure 14. DTT initial concentration ([DTT]0) dependence of (A)
relative contributions (A) of fast and slow degradation processes and
(B) reaction rate constants kfast and kslow calculated from fitting of data
in Figure 13 on the basis of eq 5.
DP
(5)
w,0
where A and K are two constants, representing the relative
contributions and rate constants of the slow and fast
656
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
the distribution of the disulfide bonds inside a hyperbranched
chain is not influenced by the presence of DTT. In addition,
Figure 14B reveals that both Kfast and Kslow increase linearly
encapsulate active chemicals and release them in a controllable
fashion because hyperbranched chains have a smaller size and a
less tendency to aggregate with each other than their linear
counterparts for a given molar mass.
with the DTT concentration, i.e., Kfast = kfast[DTT]0 and Kslow
=
kslow[DTT]0, indicating that the cleavage of the disulfide bond
on a polymer chain by DTT is very similar to those small
molecule thiol−disulfide exchange reactions.53−55 It is worth
noting that both kfast and kslow are much smaller than the
apparent rate constants (105−106 min−1 M−1) of small
molecule thiol−disulfide exchange reactions in DMF.56
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail llw@mail.ustc.edu.cn (L.L.).
*E-mail xdye@ustc.edu.cn (X.Y.).
Finally, we studied the effect of solvent on the degradation
kinetics of HB-(S−S−PS)35 because it is practically important
to choose a proper solvent to degrade a given polymer chain.
Figure 15 shows that the degradation rate increases with the
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The financial support of the Ministry of Science and
Technology of China Key Project (2012CB933800), the
National Natural Scientific Foundation of China Projects
(20934005, 51173177 and 21274140), and the Hong Kong
Special Administration Region Earmarked Projects
(CUHK4036/11P, 2130281/2060431; CUHK4035/12P,
2130306/4053005; and CUHK7/CRF/12G, 2390062) is
gratefully acknowledged.
REFERENCES
■
(1) Kong, L.; Sun, M.; Qiao, H.; Pan, C. J. Polym. Sci., Part A: Polym.
Chem. 2010, 48, 454.
(2) Liu, D.; Zeng, S.; Hu, Q.; Yi, C.; Xu, Z. Polym. Bull. 2010, 64, 877.
(3) Xu, J.; Tao, L.; Boyer, C.; Lowe, A. B.; Davis, T. P.
Macromolecules 2009, 43, 20.
(4) Konkolewicz, D.; Gray-Weale, A.; Perrier, S. B. J. Am. Chem. Soc.
2009, 131, 18075.
Figure 15. DTT-induced degradation time (t) dependence of
normalized weight-average degree of polycondensation (DPw,t
/
DPw,0) of HB-(S−S−PS)35 in three different solvents at 25 °C,
where lines represent double-exponential fittings of eq 5.
(5) Hutchings, L. R. Soft Matter 2008, 4, 2150.
(6) Hutchings, L. R.; Dodds, J. M.; Roberts-Bleming, S. J. Macromol.
Symp. 2006, 240, 56.
(7) Hutchings, L. R.; Dodds, J. M.; Roberts-Bleming, S. J.
Macromolecules 2005, 38, 5970.
(8) Zimm, B. H.; Stockmayer, W. H. J. Chem. Phys. 1949, 17, 1301.
(9) de Gennes, P. G. Biopolymers 1968, 6, 715.
(10) Hutchings, L. R.; Dodds, J. M.; Rees, D.; Kimani, S. M.; Wu, J.
J.; Smith, E. Macromolecules 2009, 42, 8675.
(11) Konkolewicz, D.; Poon, C. K.; Gray-Weale, A.; Perrier, S. Chem.
Commun. 2010, 47, 239.
(12) He, C.; Li, L.; He, W.; Jiang, W.; Wu, C. Macromolecules 2011,
44, 6233.
solvent polarity. To our knowledge, this is the first study of how
the solvent polarity affects the degradation kinetics of disulfide-
functionalized hyperbranched polymers. Quatitatively, the fast
and slow degradation rate constants in DMF are ∼2 and ∼5
times higher than those in THF and DCM, respectively. It has
been known that the reducing power of DTT depends on how
easily the thiol groups of DTT are ionized into the thiolate
form −S− because only a free thiolate can attack the disulfide
linkage on a chain.28,52 Generally, organic solvents with a higher
polarity are able to polarize the thiol group to produce more
free thiolate.
(13) Li, L.; He, C.; He, W.; Wu, C. Macromolecules 2011, 44, 8195.
(14) He, C.; He, W.; Li, L.; Jiang, W.; Tao, J.; Yang, J.; Chen, L.; Ge,
X.; Chen, S. J. Polym. Sci., Part A: Polym. Chem. 2012, 50, 3214.
(15) Li, L.; Lu, Y.; An, L.; Wu, C. J. Chem. Phys. 2013, 138, 114908.
(16) Li, L.; He, C.; He, W.; Wu, C. Macromolecules 2012, 45, 7583.
(17) Wu, C.; Li, L. Polymer 2013, 54, 1463.
(18) Li, L.; Zhou, J.; Wu, C. Macromolecules 2012, 45, 9391.
(19) Jocelyn, P. C. Chemical reduction of disulfides. In Methods in
Enzymology; William, B., Jakoby, O. W. G., Eds.; Academic Press: New
York, 1987; Vol. 143, p 246.
(20) Cleland, W. W. Biochemistry 1964, 3, 480.
(21) Humphrey, R. E.; Hawkins, J. M. Anal. Chem. 1964, 36, 1812.
(22) Humphrey, R. E.; Potter, J. L. Anal. Chem. 1965, 37, 164.
(23) Overman, L. E.; Smoot, J.; Overman, J. D. Synthesis 1974, 1974,
59.
CONCLUSION
■
Using the interchain “clicking” of disulfide-functional macro-
monomer ≡−S−S−(PS−N3)2, we have successfully prepared
model hyperbranched polystyrene chains with both uniform
subchains and controlled locations of cleavable disulfide
linkage. The DTT-induced degradation of such model
hyperbranched chains in DMF contains a fast and a slow
process because the degradation of the disulfide bonds on the
chain periphery is much faster than those inside, attributing to a
higher DTT concentration outside and a weaker steric effect of
the polystyrene segments. Not only the category of monomer
and cleavable linkage but also the coupling method developed
in the current study can be alternated to make various
structure-defined degradable hyperbranched polymer chains,
which makes further studies of the correlation between
microscopic structures and macroscopic properties of degrad-
able hyperbranched polymers possible. We can envision that
such degradable hyperbranched chains can be used to
(24) Wang, W.; Sun, H.; Meng, F.; Ma, S.; Liu, H.; Zhong, Z. Soft
Matter 2012, 8, 3949.
(25) Deng, R.; Yue, Y.; Jin, F.; Chen, Y.; Kung, H.-F.; Lin, M. C. M.;
Wu, C. J. Controlled Release 2009, 140, 40.
(26) Canadell, J.; Goossens, H.; Klumperman, B. Macromolecules
2011, 44, 2536.
657
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658

Macromolecules
Article
(27) Yoon, J. A.; Kamada, J.; Koynov, K.; Mohin, J.; Nicolay, R.;
̈
Zhang, Y.; Balazs, A. C.; Kowalewski, T.; Matyjaszewski, K.
Macromolecules 2011, 45, 142.
(28) Tsarevsky, N. V.; Matyjaszewski, K. Macromolecules 2002, 35,
9009.
(29) Wang, Y.; Wang, F.; Sun, T.; Wang, J. Bioconjugate Chem. 2011,
22, 1939.
(30) Ko, N. R.; Yao, K.; Tang, C.; Oh, J. K. J. Polym. Sci., Part A:
Polym. Chem. 2012, 51, 3071.
(31) Whittaker, M. R.; Goh, Y.-K.; Gemici, H.; Legge, T. M.; Perrier,
S.; Monteiro, M. J. Macromolecules 2006, 39, 9028.
(32) Zhang, Q.; Noh, S. M.; Nam, J. H.; Jung, H. W.; Park, J. M.; Oh,
J. K. Macromol. Rapid Commun. 2012, 33, 1528.
(33) Zhang, M.; Liu, H.; Shao, W.; Miao, K.; Zhao, Y. Macromolecules
2013, 46, 1325.
(34) Tsarevsky, N. V.; Huang, J.; Matyjaszewski, K. J. Polym. Sci., Part
A: Polym. Chem. 2009, 47, 6839.
(35) Xu, J.; Tao, L.; Liu, J.; Bulmus, V.; Davis, T. P. Macromolecules
2009, 42, 6893.
(36) Yang, W.; Pan, C. Macromol. Rapid Commun. 2009, 30, 2096.
(37) Yan, J.; Hong, C.; You, Y. Macromolecules 2011, 44, 1247.
(38) Jiang, X.; Zhang, M.; Li, S.; Shao, W.; Zhao, Y. Chem. Commun.
2012, 48, 9906.
(39) Kamada, J.; Koynov, K.; Corten, C.; Juhari, A.; Yoon, J. A.;
Urban, M. W.; Balazs, A. C.; Matyjaszewski, K. Macromolecules 2010,
43, 4133.
(40) Steinhauer, W.; Keul, H.; Moller, M. Polym. Chem. 2011, 2,
1803.
(41) Zhang, Q.; Aleksanian, S.; Noh, S. M.; Oh, J. K. Polym. Chem.
2013, 4, 351.
(42) Tao, L.; Liu, J.; Tan, B. H.; Davis, T. P. Macromolecules 2009,
42, 4960.
(43) You, Y.; Hong, C.; Pan, C. Macromolecules 2009, 42, 573.
(44) Zimm, B. H. J. Chem. Phys. 1948, 16, 1099.
(45) Chu, B. Laser Scattering, 2nd ed.; Academic Press: New York,
1991.
(46) Berne, B.; Pecroa, R. Dynamic Light Scattering; Plenum Press:
New York, 1976.
(47) Li, L.; Yan, M.; Zhang, G.; Wu, C. Macromolecules 2013, 46,
8152.
(48) Lutz, J. F.; Matyjaszewski, K. J. Polym. Sci., Part A: Polym. Chem.
2005, 43, 897.
(49) Jakubowski, W.; Kirci-Denizli, B.; Gil, R. R.; Matyjaszewski, K.
Macromol. Chem. Phys. 2008, 209, 32.
(50) Binder, W. H.; Sachsenhofer, R. Macromol. Rapid Commun.
2007, 28, 15.
(51) Binder, W. H.; Sachsenhofer, R. Macromol. Rapid Commun.
2008, 29, 952.
(52) Bock, V. D.; Hiemstra, H.; van Maarseveen, J. H. Eur. J. Org.
Chem. 2006, 51.
(53) Gallogly, M. M.; Starke, D. W.; Mieyal, J. J. Antioxid. Redox
Signaling 2009, 11, 1059.
(54) Lukesh, J. C.; Palte, M. J.; Raines, R. T. J. Am. Chem. Soc. 2012,
134, 4057.
(55) Wiita, A. P.; Ainavarapu, R. K.; Huang, H. H.; Fernandez, J. M.
Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 7222.
(56) Singh, R.; Whitesides, G. M. J. Am. Chem. Soc. 1990, 112, 1190.
658
dx.doi.org/10.1021/ma402438m | Macromolecules 2014, 47, 650−658