Cationic polymer grafted starch from nonsymmetrically substituted macroinitiators

Article abstract of DOI:10.1021/ma048063f

The aim of this work was the synthesis of starch macroinitiators for cationic polymer grafted starches that: (i) are free of cationic homopolymer, and (ii) display a high degree of conversion of the cationic monomer. We show that this can be achieved by a free-radical polymerization reaction using the cationic monomer N-methacryloyloxyethyl-N,N-dimethyl-N-benzylammonium chloride (MADAM-BQ) and a new starch-based macroazoinitiator. For this purpose, the acid chloride of 4-tert-butylazo-4-cyanovaleric acid was synthesized and bound covalently to starch (predominantly in the C6 position) to form a nonsymmetrically substituted macroinitiator that was used to polymerize MADAM-BQ in aqueous media. Essentially no MADAM-BQ homopolymer was formed. The initiator decomposes thermally to starch radicals of high reactivity and low-molar mass radicals that do not initiate polymerization. The reason for the different reactivities of the radicals is presumably due to the nonsymmetric constitution of the starch-bound azo groups. The graft polymerization of MADAM-BQ in aqueous solution performs according to an ideal overall kinetic. The structure of the synthesized starch-graft-poly(MADAM-BQ) products is similar to that of block copolymers because of the low radical efficiency of the starch initiators in aqueous solution. Especially, starch substrates with a higher content of azo groups did not lead to graft products with shorter graft distances because the state of solution of these macroinitiators becomes worse and aggregation occurs with an increasing degree of substitution.

Full text of DOI:10.1021/ma048063f

Macromolecules 2005, 38, 7251-7261
7251
Cationic Polymer Grafted Starch from Nonsymmetrically Substituted
Macroinitiators
Stefano Bruzzano,* Nathalie Sieverling, Christoph Wieland, and Werner Jaeger
Fraunhofer Institute for Applied Polymer Research, Geiselbergstrasse 69,
1
4476 Potsdam-Golm, Germany
Andreas F. Th u1 nemann*
Federal Institute for Materials Research and Testing, Richard-Willst a¨ tter-Strasse 11,
1
2489 Berlin, Germany
J u1 rgen Springer
Technical University of Berlin, Strasse des 17 Juni, Berlin, Germany
Received September 20, 2004; Revised Manuscript Received May 8, 2005
ABSTRACT: The aim of this work was the synthesis of starch macroinitiators for cationic polymer grafted
starches that: (i) are free of cationic homopolymer, and (ii) display a high degree of conversion of the
cationic monomer. We show that this can be achieved by a free-radical polymerization reaction using the
cationic monomer N-methacryloyloxyethyl-N,N-dimethyl-N-benzylammonium chloride (MADAM-BQ) and
a new starch-based macroazoinitiator. For this purpose, the acid chloride of 4-tert-butylazo-4-cyanovaleric
acid was synthesized and bound covalently to starch (predominantly in the C6 position) to form a
nonsymmetrically substituted macroinitiator that was used to polymerize MADAM-BQ in aqueous media.
Essentially no MADAM-BQ homopolymer was formed. The initiator decomposes thermally to starch
radicals of high reactivity and low-molar mass radicals that do not initiate polymerization. The reason
for the different reactivities of the radicals is presumably due to the nonsymmetric constitution of the
starch-bound azo groups. The graft polymerization of MADAM-BQ in aqueous solution performs according
to an ideal overall kinetic. The structure of the synthesized starch-graft-poly(MADAM-BQ) products is
similar to that of block copolymers because of the low radical efficiency of the starch initiators in aqueous
solution. Especially, starch substrates with a higher content of azo groups did not lead to graft products
with shorter graft distances because the state of solution of these macroinitiators becomes worse and
aggregation occurs with an increasing degree of substitution.
1
. Introduction
an uncharged comonomer like acrylamide. This is also
evident for grafting cellulose.30
The physical and chemical modification of starch
A further possibility of initiating the graft polymer-
ization is to bind initiating groups directly to the starch
molecule, which can be activated thermally or chemi-
offers many possibilities to adjust its properties to those
required in tailor-made products. One attractive method
of chemical modification of starch is the grafting of
synthetic polymer chains from starch (e. g. starch-graft-
polystyrene).
initiated reactions.
cally. For example, a thiocarbamate derivative of starch
31,32
can be grafted by the addition of H2O2
or iron
and azo
1
,2
Most of these graftings are radical-
33
34-36
37,38
peroxodisulfate. Peroxy,
azonium,
3
-6
Here, the graft polymers are
groups are the common thermally cleavable groups;
starch derivatives of these groups are typical starch
macroinitiators. All these starch macroinitiators show
the same disadvantages as the above-named initiating
systems, such as the formation of low-molar mass
radicals inducing homopolymers.
generated by the formation of radicals on the starch
backbone from which the polymerization of vinylic
monomers is initiated. The formation of the initiating
radicals by γ, â, or UV radiation is nonspecific and
results in the formation of a significant amount of
7
-10
homopolymer.
Classical initiating systems for the
Azo compounds of starch have been described in the
chemical formation of starch radicals are permangan-
patent literature in which two anhydroglucose units
11-14
15-17
ates in the presence of acids,
the Fenton agent,
were linked by symmetrically substituted azobis
and peroxodisulfates.18-20 Large amounts of homopoly-
mers are formed in all of these systems because the
primary radicals, which are not bound to starch, also
initiate homopolymerization. The unwanted side reac-
tions can be suppressed by generating radicals with the
39-42
groups.
An advantage of these starch initiators is
that their thermal decomposition exclusively produces
macroradicals. A strong disadvantage, however, is that
the polymer analogue reaction of such bifunctionalized
azo compounds leads generally to cross-linked struc-
tures whose accessibility to the graft monomer is
hindered. This limits a homogeneous graft polymeriza-
tion in solution.
2
1-26
27-29
help of Ce(IV)
or Mn(III),
but it is difficult to
remove these metal ions from the colored products.
However, even in this case, the formation of homopoly-
mer cannot be excluded. In particular, the grafting of
cationic monomers leads exclusively to homopolymers
unless the graft process is not promoted by the use of
The two aims of this work are to avoid the formation
of a cationic homopolymer and to graft a cationic vinyl
monomer without the need to add noncharged mono-
mers such as acrylamide, which gives grafted copolymer
chains.43 Our idea was that these aims can be achieved
with a starch macroinitiator that contains nonsym-
*
Corresponding authors. E-mail: stefano.bruzzano@iap.
fraunhofer.de (S.B.); andreas.thuenemann@bam.de (A.F.T.).
1
0.1021/ma048063f CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/19/2005

7252 Bruzzano et al.
Macromolecules, Vol. 38, No. 17, 2005
. Materials and Methods
2
Materials. All chemicals were obtained from Fluka unless
otherwise stated. Laevulinic acid (g97%), tert-butylhydrazo-
nium chloride (g97%), potassium cyanide (g98%), and 0.01
-
1
mol L hydrochloric acid prepared by 1:10 dilution of 0.1 N
HCl (Merck, standard solution), phosphor pentoxide (Sicapent),
sodium hydroxide (98%), nitrogen (Linde, 5.0), chlorine (Linde,
2
.8), pentane (g95%), toluene (g99.5%), thionyl chloride
(
g99.5%), N,N-dimethyl-acetamide (Merck, for synthesis
g99%), triethylamine (g99.5%), N-methacryloyloxyethyl-N,N-
dimethyl-N-benzylammonium chloride (Ato-Chem, 75% aq
solution, determined by potentiometry of Cl), 2,3-dimethyl-1-
vinylimidazolium chloride (R o¨ hm), hydroquinone (g99.5%),
4
,4′-azobis-4-cyanovaleric acid (WAKO, V-501), 4-tert-butylazo-
Figure 1. Concept for the synthesis of cationic grafted starch.
4-cyanovaleric acid, and 4-tert-butylazo-4-cyanovaleric starch
A starch-based nonsymmetrically substituted macroazoinitia-
ester.
1 2
tor is formed in a first step whose groups R and R have
Starch. We used a commercially available enzymatic waxy
maize starch hydrolysate from Roquette (France). This was
chosen for the starch modification in homogeneous aqueous
solution because its aqueous solutions are stable for long
periods of time, showing a low tendency of retrogradation due
different symmetries. Its thermal decomposition (second step)
produces highly reactive radicals R * that are bound to the
anhydroglucose units (AGU) of the starch and the low-molar
mass radicals R * that do not initiate polymerization. The
cationic monomer M polymerizes exclusively initiated by R
third step), while the amount of polymerization initiated by
* can be neglected (no cationic homopolymer is formed).
1
2
1
*
4
4
to its high content of amylopectin. The broad molar mass
distribution of this starch (cf. Figure 3, solid line) with molar
(
R
2
3
7
-1
masses in the range of 10 -10 g mol and a polydispersity
of 14.1 was fractionated by ultrafiltration (Minisette Omega,
metrically substituted azo groups. It should decompose
thermally into highly reactive radicals R1* being bound
to the starch and into low-molar mass radicals R2*,
which do not initiate polymerization. There were indica-
tions in the literature that R2* can be a tert-butyl radical
that unexpectedly does not initiate polymerization of
5
-1
FA. Pall Filtron, cutoff was 10 g mol ) to a starch with a
5
7
-1
3
molar mass in the range of 10 -10 g mol , Mw ) 660 × 10
-
1
g mol and a polydispersity of 1.5 (cf. Figure 3, dotted line).
Synthesis of Nonsymmetrically Substituted Starch
Azo Initiators. 4-tert-Butylhydrazono-4-cyanovaleric Acid
(cf. structure 3 in Figure 2). An amount of 11.6 g (0.1 mol) of
laevulinic acid was dissolved in 50 mL of water, and 17.5 g
37
homopolymer. Presuming that no initiating species are
formed by transfer reactions, this should give a graft
polymer that is essentially free of cationic homopolymer.
Our concept is shown schematically in Figure 1. To
verify this concept, we synthesized a suitable nonsym-
metrically substituted azo compound, 5, and bound it
to starch, which resulted in the starch macroazoinitiator
(0.14 mol) of tert-butylhydrazonium chloride and 6.5 g (0.1 mol)
of potassium cyanide were added. A white precipitate was
formed while stirring the solution at room temperature
overnight. The precipitation was separated, washed with
diluted hydrochloric acid, and dried over phosphor pentoxide.
The yield was over 90%. CHN analysis found: C 56.7%, H
8.9%, N 19.6%; calcd: C 56.3%, H 9.0%, N 19.7%. Melting
point: 104.5 ( 0.5 °C.
7
(cf. Figure 2). The polymerization products and the
polymerization kinetics were investigated systemati-
cally.
4-tert-Butylazo-4-cyanovaleric Acid (cf. structure 4 in
Figure 2). An amount of 4.0 g (0.1 mol) of sodium hydroxide
Figure 2. Three-step synthesis of a low-molar mass azo compound 5 suitable for covalent binding to starch (left-hand figure).
Synthesis of the nonsymmetrically substituted starch-based macroazoinitiator 7 using the acid chloride 5 and amylopectin 6
(
right-hand figure).

Macromolecules, Vol. 38, No. 17, 2005
Cationic Polymer Grafted Starch 7253
Figure 3. Molar mass distribution of the waxy maize starch
hydrolysate in its state as received (solid line) with a polydis-
persity of 14.1 and after ultrafiltration (dotted line) with a
3
-1
polydispersity of 1.5 (M
w
) 660 × 10 g mol ).
Figure 4. 1H NMR spectrum (400 MHz, DMSO-d
) of the
6
was dissolved in 50 mL of water, and 21.3 g (0.1 mol) of 3 was
added while stirring. The solution was cooled to 5 °C under
nitrogen atmosphere. Then chlorine was introduced until a
thick-yellow flocculation was formed while the reaction tem-
perature was kept to between 5 and 10 °C. The product was
separated and crystallized in a pentane/toluene solvent mix-
starch macroazoinitiator I10 for the determination of the degree
of substitution (DS) which is here 0.10. The DS was deter-
1
mined from the H NMR signal intensity ratios of the starch
(
1
4.2-6.0 ppm) to the protons of the tertiary butyl group (1.0-
.8 ppm).
1
3
ture (50:10). Analysis of 4: C NMR (CDCl
3
, δ in ppm): 23.81
An amount of 3.78 g (0.01 mol) of 75% aq solution of MADAM-
BQ, 0.1057 g (0.5 mmol) of 4-tert-butylazo-4-cyanovaleric acid
(
q, 1C); 26.54 (q, 3C); 28.87 (t, 1C); 32.82 (t, 1C); 68.61 (s, 1C);
7
0.72 (s, 1C); 118.89 (s, 1C); 177.78 (s, 1C). The yield was in
(
or 0.1402 g (0.5 mmol) of 4,4′-azo-bis-4-cyanovaleric acid) were
the range of 50-70%. CHN analysis found: C 55.8%. H 8.0%,
weighed in a 100 mL measuring flask, solved in deionized
water, and filled up to the calibration mark. The solution (20
mL) was filled into a reaction ampule (Schlenk tube). The
ampule was frozen with an ethanol/dry ice mixture, degassed
with a vacuum pump, and defrosted inside a water bath of
room temperature. After repeating this process four times, the
clear solution was overlaid with argon. The thermal initiation
of the polymerization solution was performed by heating the
reaction ampule in a water bath at 70 °C. After stirring for
N 18.5%; calcd: C 56.9%, H 8.1%, N 19.9%. FT-IR (KBr,
-1
-1
selected bands): 2974 cm (νas CH
3
2
), 2937 cm (νas CH ), 2237
-
1
-1
-1
cm (ν CN, weak), 1712 cm (ν CdO), 1580 cm (ν NdN,
weak). Melting point: 82.5 ( 0.5 °C.
4-tert-Butylazo-4-cyanovaleric Acid Chloride (cf. struc-
ture 5 in Figure 2). An amount of 21.1 g (0.1 mol) of 4 was
dissolved in 100 mL of water-free toluene and adjusted to 0-5
°
C. Then 9 mL (0.12 mol) of thionyl chloride was added in
droplets, and the reaction mixture was stirred at room
temperature for 4 h. The obtained product was a yellow-
brownish liquid after the distillation of the excess thionyl
chloride and solvent. The yield was 80%. It had to be cooled
for storage at 4 °C to avoid decomposition. FT-IR (KBr, selected
1
80 min, 0.010 g of hydroquinone (500 ppm) was added and
the solution was diluted with cold deionized water, filled into
3
a dialysis tube (regenerated cellulose, cutoff was 6 × 10 g
-
1
mol ), and dialyzed against water for 5 days to remove
unreacted cationic monomer (control of conductivity). After
-
1
-1
-1
bands): 2977 cm (νas CH
3
2
), 2937 cm (νas CH ), 2237 cm
-
1
-1
freeze-drying a white, cottonlike solid was obtained.
(
ν CN, weak), 1798 cm (ν CdO), 1580 cm (ν NdN, weak).
1
H NMR (400 MHz, D
), 3.5-4.2 (N-CH2-), 7.2-8.0 (-C
The proton signal intensities do not correspond completely
2
O): δ (in ppm) ) 0.5-1.4 (C-CH
3
),
4
-tert-Butylazo-4-cyanovaleric Acid Starch Ester (cf.
structure 7 in Figure 2, Table 1). An amount of 5.0 g of starch
corresponding to 0.03 mol anhydroglucose units) was dissolved
in 100 mL of N,N-dimethyl-acetamide (DMA) and heated to
70 °C. Then 30 mL of DMA was distilled in a nitrogen stream
2
.7-3.4 (N-CH
3
6 5
H
).
(
to the structure of poly(MADAM-BQ). This may due to the
different state of solution/solvation and mobility of the polymer
protons, especially the signals of protons being attached to
hydrophobic backbone are partially broadened and suppressed.
This phenomenon is known and described for other ionic
1
to remove the water from the starch. The solution was cooled
to room temperature and transferred to a 150 mL double-wall
reactor. A mixture of 3.6 g (0.036 mol) of triethylamine and
4
5
10 mL of DMA was added. The solution was cooled to 8 °C,
methacrylate structures by Laschewsky and Zerbe.
and 1.2 g (0.0052 mol) of 4-tert-butylazo-4-cyanovaleric acid
chloride, which was dissolved in 10 mL DMA, was added
slowly, and the reaction mixture was stirred for 24 h at 8 °C.
The starch derivative was precipitated in 1 L of methanol,
dissolved in water, and dialyzed for 4 days at 4 °C (mem-
Graft Polymerization of (MADAM-BQ) N-Methacryl-
oxyethyl-N,N-dimethyl-N-benzylammonium Chloride
(MADAM-BQ) (cf. Figure 7, Example no. 2 of Table 3). A
4-tert-butylazo-4-cyanovaleric acid starch ester of DS ) 0.05
5
(I ) (0.648 g, 4 mmol) and 0.756 g of a 75% aq solution of
3
-1
brane: regenerated cellulose, cutoff was 6 × 10 g mol ) and
then freeze-dried. The degree of substitution of the resulting
starch macroinitiator (compound 7 in Figure 2) was 0.05, as
MADAM-BQ (2 mmol) were solved in 20 mL of deionized
water. The solution was filled into a reaction ampule (Schlenk
tube), frozen with an ethanol/dry ice mixture, degassed via a
vacuum pump, and defrosted inside a water bath to room
temperature. After repeating this process four times, the clear
solution was overlaid with argon. The thermal initiation of the
polymerization solution was performed by heating the reaction
ampule in a water bath at 70 °C. After stirring for 180 min,
0.010 g of hydroquinone (500 ppm) was added and the solution
was diluted with cold deionized water, filled into a dialysis
1
revealed by H NMR signal intensity ratios of starch (4.2-6.0
ppm) to methyl group protons (1.0-1.8 ppm) (cf. Figure 4).
The degree of substitution varied in the range of 0.04-0.72
by adjusting the reaction conditions (cf. Table 1). The starch
initiator with the low degree of substitution (DS ) 0.05, I
5
of
3
Table 1) was characterized by SEC-MALLS (M
w
) 637 × 10
-
1
3
-1
g mol and M
n
) 504 × 10 g mol ), while the higher-
3
-1
substituted initiators showed adsorption effects on the SEC
tube (regenerated cellulose, cutoff was 6 × 10 g mol ), and
dialyzed against water for several days to remove unreacted
cationic monomer (control of conductivity). After freeze-drying
a white, cottonlike solid was obtained, showing an intrinsic
columns. CHN analysis of I70 (DS ) 0.72) calcd: C 52.5%, H
6
(
.9%, H 9.5%; found C 51.5%, H 7.5%,N 8,7%. Water content
Karl Fischer titration): 2.87% (w/w).
3
-1
-1
Homopolymerization of N-Methacryloyloxyethyl-N,N-
viscosity of 35 cm g (0.2 mol L aq solution of Na
2 4
SO +
dimethyl-N-benzylammonium Chloride (MADAM-BQ).
1% (w/w) acetic acid). The synthetic mass ratio of the obtained

7254 Bruzzano et al.
Macromolecules, Vol. 38, No. 17, 2005
product was quantified by densiometry (37%, see section on
graft polymerization kinetics) and also by UV-vis measure-
ments at a wavelength of 268 nm (43%), resulting in an
average amount of 40% (w/w). The synthetic molar ratio of
decoupling, without NOE decoupling) was carried out with a
UNITY 400 (Varian, Germany). The UV-vis spectroscopy was
performed with an UVIKON 939 photometer (Kontron, UK)
using a temperature-controlled quartz cuvette (d ) 10 mm).
Solvents were N,N-dimethyl-acetamide and water, respec-
tively. Analytical ultracentrifugation was performed with a
XL-I (Beckman, Germany) at 20 ( 0.1 °C and a rotation rate
1
the graft product was additionally determined with H NMR.
It was calculated to 26.5% mol/mol by means of comparing the
signal intensities of the acetal starch protons and the aromatic
protons of MADAM-BQ. This value is equal to a synthetic mass
4
-1
of 4 × 10 min . The sedimentation of the macromolecules
-
1
-1
fraction (w
U
) of 39% (w
U
) (0.265 g mol × 283.4 g mol )/
was detected by interference measurements using water con-
-
1
-1
-1
(
0.265 g mol × 283.4 g mol ) + (0.735 × 162.1)).
2 4
taining 0.2 mol L Na SO and 1% (w/w) acetic acid as solvent
1
-1
H NMR (400 MHz, D
2
O): δ (in ppm) ) 0.5-1.4 (C-CH3),
and different sample concentration of 1-5 g L . The calcula-
2
.7-3.3 (N-CH3), 3.4-4.2 (protons of the starch ring at C2-
tion of the relative molar mass M
r
followed the method of
4
7
C6), 5.2-5.6 (acetal proton at C1), 7.2-8.0 (-C6H5).
To get different synthetic products, the initiator and mono-
mer concentrations were varied as shown in the Tables 3 and
Linow and Phillip using
3/2
s η
M ) 2.406 × 10 mol-1
2
5
0
0
([η] kSB)
1/2
4
. In addition, we also used a corresponding starch initiator
r
(1 - vj F0)
(
I
10) of DS ) 0.10 to graft MADAM-BQ.
St
Graft Process Characterization. Because of the absence
of homopolymer, the graft efficiency of the polymerization
experiments was in each case 100%. The conversion U was
calculated by the use of equation U ) (1/w100% - 1)/(1/w
1
-13
where s
of the solvent in g cm , F
η] is the intrinsic viscosity in cm /g, kSB is the Schulz-
0 0
is the Svedberg constant in 10 s, η is the viscosity
-1
3
0
is the density of the solvent in g/cm ,
U
-
[
3
), where w represents the synthetic mass ratio of the
U
Blaschke constant, and vj St is the partial specific volume of the
dialyzed reaction product determined by UV-vis measure-
ments (268 nm), and w100% is the corresponding value for 100%
conversion. It must be taken into account that nongrafted
starch remains after polymerization, and the purified product
is a mixture of the starch graft copolymer and free starch. Its
3
starch substrate in cm /g.
w n
The weight- and number-average molar masses (M and M ,
respectively) of the polymers were determined by size exclusion
chromatography with multiangle laser light scattering detec-
tion (SEC-MALLS: degaser SCM 400, isocratic pump P1000
and auto sampler AS 1000 from Spectra Physics, Egelsbach,
Germany; TSK columns (polyglycidyl(meth)acrylate gel) PWH-
U
mass fraction is wstarch ) (1 - w /wP-MADAM-BQ).
Molecular Characteristics of Starch-graft-poly(MAD-
AM-BQ). wP-MADAM-BQ represents the synthetic add-on of the
pure graft copolymer, which is the ratio of the mass of grafted
poly(MADAM-BQ) and the total mass of the starch backbone
and its grafts. SEC was used to determine wP-MADAM-BQ by the
coupling of RI and UV detection. The ratio of the simulta-
neously measured UV and RI signals (c(λ ) 268 nm) and c(RI),
respectively) was calculated. It was found that the ratio c(268
nm)/c(RI) is approximately constant for small elution volumes,
where graft copolymer fractions elutes exclusively and de-
creases at higher volumes where graft copolymer fractions and
free starch elutes at the same time (not shown).
5
7
Guard + 6000 (7.5 × 300 mm, 17 µm) 5 × 10 to 5 × 10 g/mol,
4
6
+
3
5000 (7.5 × 300 mm, 17 µm) 5 × 10 to 7 × 10 g/mol, +
2 4
000 (7.5 × 300 mm, 10 µm) 10 to 6 × 10 g/mol by Tosohaas,
Stuttgart, Germany; Hema Bio column (hydroxyethyl meth-
2
4
acrylate gel) (8.0 × 300 mm, 10 µ) 10 to 2 × 10 g/mol Polymer
Standards Service GmbH, Mainz, Germany; laser light scat-
tering detector Dawn DSP, 632.8 nm, cell type K5; 18 detectors
from 22.5° up to 147° from Wyatt, Woldert, Germany; OPTI-
LAB DSP interferometric refractometer, 632.8 nm, cell type
P10 from Wyatt, Woldert, Germany; UV-vis detector UV 2000
by Spectra Physics, Egelsbach, Germany; software Astra 4.7
from Wyatt, Woldert, Germany). The accuracy of the deter-
mined molar masses is about (10%. MADAM-BQ was detected
The wP-MADAM-BQ is equal to the c(268 nm)/c(RI) ratio in the
constant region, presuming that the correct refractive index
increment is used. Refractive index increments were calculated
from the polymer composition according to Bushuk and
-
1
at a wavelength of 268 nm. A mixture of 0.2 mol L Na
2 4
SO
and 1% (w/w) acetic acid in deionized water was used as eluent.
4
6
Benoit, which is dn/dc ) (wP-MADAM-BQ × (dn/dc)P-MADAM-BQ
)
-1
The flow rate was 0.65 mL min at 20 °C using a sample loop
+
((1 - wP-MADAM-BQ) × (dn/dc)starch) for the grafted starch. An
-
1
of 100 µL and sample concentrations of 1-4 g L
.
interferometric refractometer was used for the (dn/dc) deter-
mination of poly(MADAM-BQ) and starch. The increment (dn/
Viscosity measurements were performed using an Ub-
belohde viscometer with automatic dilution (Viscoboy 2 from
-
1
dc)P-MADAM-BQ was 0.182 mL g , and (dn/dc)starch is 0.147 mL
-
1
Lauda) at 30 ( 0.1 °C. An aqueous mixture of Na SO (0.2
2 4
g
at a wavelength of 633 nm in an aq solution of 0.2 mol/L
SO with 1% (w/w) acetic acid. The determination of
P-MADAM-BQ or dn/dc was carried out iteratively.
The graft length P represents the average number of
monomer units per grafted poly(MADAM-BQ). P was deter-
mined by SEC-MALLS after the complete hydrolysis of the
-
1
mol L ) and 1% (w/w) acetic acid with a pH of 3-4 was used
as solvent. The thickness of the capillary used was 0.58 mm.
Na
w
2
4
The kinetic measurements were performed with a density
meter (DMA 60/602, Paar, Austria). It is based on a vibrating
mass whose resonance frequency is influenced by the density
l
l
-
1
48
starch (5% (w/w) of the dialyzed graft product in 0.1 mol L
HCl at 80 °C for 4 h) using P ) number-average molar mass
of the polymer solution (oscillating U-tube principle). The
correctness as given for the density of degassed water was
l
-
6
3
of the cationic graft chains Mn,P-MADAM-BQ/molar mass of
the monomer unit MMADAM-BQ. The ester structure of the
poly(MADAM-BQ) is not affected by the applied conditions of
hydrolysis. The graft frequency νgraft reflects the average
distance between two graft chains quantified by the average
number of anhydroglucose units (AGU) per grafted poly-
(0.5‚10 g/cm and the temperature was held constant at
(0.01 °C.
3. Results and Discussion
3.1. Starch Macroinitiator. The synthesis of 5 was
carried out in three steps, as shown in Figure 2. First,
the reaction of tert-butylhydrazonium chloride (2) with
laevulinic acid (1) in the presence of cyanide results in
(
MADAM-BQ) chain. Mn,P-MADAM-BQ and wP-MADAM-BQ were
used to calculate the graft frequency via νgraft ) ((1 -
w
P-MADAM-BQ) × Mn,P-MADAM-BQ)/(wMADAM-BQ × MAGU). It was
assumed that no recombination of growing side chains occur
during the graft polymerization. This is well-known for the
4-tert-butylhydrazono-4-cyanovaleric acid (3). Second, 3
was oxidized by chlorine to the azo compound 4. Third,
compound 4 was reacted with thionyl chloride to the
low-molar mass azo initiator 5, which is a brownish
liquid. The coupling of 5 to starch (6) to produce the
starch macroinitiator (7) was carried out analogous to
(
graft) polymerization of methacrylics with bulky substituents
such as MADAM-BQ. The average number of grafted chains
per starch molecule, Ngraft, was calculated from νgraft and
M
n, starch , the number-average molar mass of the starch
backbone Ngraft ) Mn,starch/(vgraft × MAGU).
Measurements. H NMR (400 MHz) and 13C NMR (400
1
a tosylating procedure for cellulose in N,N-dimethyl-
acetamide in the presence of triethylamine.4
9,50
A low
MHz) spectroscopy (solvent DMSO-d
6 3
, CDCl , inverse gated

Macromolecules, Vol. 38, No. 17, 2005
Cationic Polymer Grafted Starch 7255
Table 1. Reaction Parameters for the Synthesis of the Starch Macroinitiators (I5 to I70), Their Degree of Substitution
DS), Intrinsic Viscosities ([η]), and Huggins Constants (kH)
(
b
4
-tert-butylazo-4-cyanovaleric
triethylamine
[mol]
starch
[mol]
[η]
3
number
acid chloride [mol]
DS
kH
[cm g-1]
starch
I5
0
0.6
1.4
4.1
6.0
11.9
a
26
23
18
17
14
a
0.010
0.015
0.020
0.025
0.030
0.040
0.100
0.090
0.080
0.070
0.075
0.070
0.030
0.035
0.040
0.045
0.035
0.030
0.05
0.07
0.10
0.13
0.60
0.72
I7
I10
I13
I60
I70
a
a
a
Turbid aqueous solutions (viscosities were not measured). b Determined at 25 °C in 0.2 M Na2SO4 with 1% (w/w) acetic acid.
dominantly to carbon C6 from the high intensity of the
substituted C6 signal and its comparison to the whole
1
DS as measured by H NMR spectroscopy. This finding
is similar to the result of Katsura et al.,51 who revealed
a preferred functionalization of C6 for the etherification
of amylopectin with diethylaminoethyl chloride and
3
-chloro-2-hydroxypropyltrimethylammonium chloride
using aqueous alkaline conditions.
Properties in Solution. The functionalization of the
starch with the azo compound 5 corresponds to a
hydrophobization of the starch, and we expected that
the solution properties of the starch and the starch
macroinitiator differ significantly. Therefore, we mea-
sured the viscosity of their aqueous salt solutions (0.2
Figure 5. 13C NMR spectrum (400 MHz, DMSO-d
-1
6
) of the
mol L Na2SO4 and 1% (w/w) acetic acid) and deter-
starch macroinitiator I10 for the determination of the preferred
position of the substitution. The signals of the substituted
carbon C6 (6s, 61.5 ppm) and the nonsubstituted C6 (6, 60.5
ppm) are clearly separated.
mined their intrinsic viscosities, [η], as well as their
Huggins constants, kH, (see Table 1). The kH, which is
3
-1
0
.6 for the pristine starch (Mw ) 660 × 10 g mol ),
increases strongly with an increasing degree of substi-
tution (1.4 for I5 to 11.9 for I13). The initiators with the
highest degree of substitution (I60 and I70) were not
measured because they formed turbid aqueous solu-
tions. Typical coiled polymers with linear chains in a
good solvent have a kH in the range of 0.3-0.6, while
higher values of 0.6-0.8 are known for branched
structures such as amylopectin.52 We assume that the
reason for the increasing kH values of the macroinitia-
tors result from less water solubility with an increase
of DS. Probably, an increasing DS leads to more compact
structures of the initiators compared to those of the
pristine starch. This assumption is consistent with the
values of the intrinsic viscosities that decrease from 26
reaction temperature of 8 °C should give predominantly
a selective C6 substitution of the anhydroglucose units
and should also avoid the thermal decomposition of the
azo groups. No gelation was observed during the reac-
tion, indicating the absence of intermolecular reactions
of starch molecules. The questions arose how the degree
of substitution depends on the composition of the
reaction mixture and whether the substitution of 5 at
the starch in C6 position is the most likely. The
properties of the synthesized macroinitiators are sum-
marized in Table 1.
Degree of Substitution (DS). The degree of sub-
stitution of the starch macroinitiator was determined
3
-1
3
-1
1
cm g (starch, DS ) 0) to 14 cm g (I13, DS ) 0.13).
Indications of the formation of aggregates were observed
by analytical ultracentrifugation, e.g., for the macro-
initiator I10; its molar mass Mr was calculated as 470 ×
using H NMR spectroscopy (400 MHz, DMSO-d6). An
example is given in Figure 4. The signals between 4.2
and 6.0 ppm can be assigned to the three protons of the
hydroxyl groups and the proton of the semiacetal of the
starch. The characteristic signal of the tertiary butyl
group occurs at 1.2 ppm and the signal of the methyl
group at 1.6 ppm. We determined the degree of substi-
tution from the ratio of the signal area of the anhydro-
glucose unit to that of the substituent, which is 0.10 in
the example shown in Figure 4. The degree of substitu-
tion of the starch macroinitiators was in the range of
3
-1
1
1
0 g mol for the molecular soluble part and 1500 ×
3 -1
0 g mol for the aggregates.
Decomposition of the Initiator. The thermal de-
composition of the initiators in aqueous solution was
measured at 85 °C by UV-vis detection at 356 nm,
where 4 has its absorption maximum. The decomposi-
tion constants, kd, were then determined from the initial
slope of the logarithmic absorption curves (i.e., ln
Absorbance(t) ) -kdt, not shown). The kd values are
0
.05-0.72 (cf. Table 1).
Position of the Azo Groups. 13C NMR measure-
-
6
-1
ments of the starch macroinitiators were carried out to
determine the position of the substitution. An example
is shown in Figure 5. The signal of the carbons can be
distinguished clearly whether the carbon in the C6
position is substituted (6s, 61.5 ppm) or not (6, 60.5
ppm). The substitution at the C2 and C3 positions
cannot be quantified because of the strong superposition
of other signals in the region of 70-76 ppm. Neverthe-
less, we can conclude that the binding of 5 was pre-
(9.5 ( 0.2) 10
(31.7 ( 14.4) 10
and (180.5 ( 37.5) 10
s
(4-t-butylazo-4-cyanovaleric acid),
-
6
-1
-6 -1
s
(I5), (93.3 ( 18.9) 10
s
(I10),
-6
-1
s
(I13) (for thermal decomposi-
tion curves see also Figure S1 in Supporting Informa-
tion). It is surprising that the macroinitiators decompose
faster than their low-molar mass analogue and faster
with increasing DS. The reason for this is not clear.
Possibly, the decomposition of the azo groups along the
starch backbone is a cooperative process. We observed

7256 Bruzzano et al.
Macromolecules, Vol. 38, No. 17, 2005
Figure 6. SEC traces of products of poly(2,3-dimethyl-1-
Figure 7. SEC traces of products of poly(MADAM-BQ)
vinylimidazolium chloride) (MV) grafted from the starch
5
grafted from the starch macroinitiator I at different reaction
5
macroinitiator I at different reaction times. The straight lines
times. The straight line shows the peak maxima positions.
indicate the peak maxima positions at different reaction times.
-
1
-1
5
Reaction conditions were [M] ) 0.1 mol L , [I ] ) 0.1 mol L ,
Reaction conditions were monomer concentration [MV] ) 0.2
-
1
-1
T ) 90 °C, termination with 500 ppm hydroquinone, UV
5
mol L , [I ] ) 0.2 mol L , T ) 90 °C, termination with 500
detection at 268 nm.
ppm hydroquinone, UV detection at 222 nm.
zation and remained as nongrafted starch. With that,
our working hypothesis is proven. Cationic starch
grafted polymers are available without the formation
of homopolymers, provided that the macroinitiated
polymerization will be carried out with vinyl monomers
of very low transfer activity.
the formation of gels for longer decomposition times,
which is promoted for higher DS.
3
.2. Graft Polymerization of Different Cationic
Vinyl Monomers. To estimate the initiating properties
of the starch initiators as well as the structure of the
resulting polymers, we investigated the polymerization
of the two cationic monomers 2,3-dimethyl-1-vinyl-
imidazolium chloride (MV) and 2-N-methacryloyloxy-
ethyl-N,N-dimethyl-N-benzylammonium chloride (MAD-
AM-BQ) when initiated with I5.
Further information about the specific properties of
starch in solution, poly(MADAM-BQ), and the corre-
sponding graft copolymer are also available from SEC-
MALLS measurements. Figure 11 describes the increase
of the radius of gyration with increasing molecular
mass. The slope increases from starch (0.26) to the graft
copolymer (0.41) to poly(MADAM-BQ) (0.63 for the
grafted chains), which was obtained after hydrolysis (5%
Figure 6 shows the SEC traces of the products from
a polymerization mixture that contained equimolar
amounts of I5 and MV (equimolar with respect to
monomer units). The reaction was carried out at 90 °C
and stopped at different times with hydroquinone so
that the products could be investigated by SEC. The
cationic polymer structures of MV were detected specif-
ically at a wavelength of 222 nm. The presence of
bimodal distributions can be seen clearly in the SEC
traces (see Figure 6). This indicates the formation of the
aspired graft polymers as well as cationic homopoly-
mers. The graft polymers are expected to be found at
lower elution volumes than that of homopolymers due
to their higher molar masses. This was verified by the
nearly constant position of the peak maxima of the
homopolymer, which is especially expected for low
conversions of a free-radical polymerization (neglecting
the consumption of monomer). By contrast, one has to
expect a shift of the peak maxima to lower elution
volumes for the graft polymer as a result of its increas-
ing molecular mass with reaction time. The high
amount of homopolymer formed is probably a result of
a radical transfer to the monomer, which has a reactive
carbon at position 5.
A comparable experiment was carried out again with
MADAM-BQ using analogue procedures of polymeriza-
tion and SEC analysis. The molar monomer concentra-
tion was reduced to 50% to produce similar molar
masses, only, because of the significantly higher reactiv-
ity of MADAM-BQ compared to MV. It can be seen in
Figure 7 that, in this case, only a single peak, which
was detected at 268 nm, is present in the SEC traces.
The peak maximum shifts to lower elution volumes with
increasing reaction times, as expected for increasing
molecular masses with time due to successful grafting.
Consequently, we can exclude the formation of signifi-
cant amounts of MADAM-BQ homopolymer on the basis
of the SEC traces. But it must be mentioned that some
amount of initiator may not have started a polymeri-
-
1
(
w/w) of the dialyzed graft product in 0.1 mol L HCl
at 80 °C for 4 h) of the graft copolymers. This indicates
a more compact structure of the starch backbone and a
1
remarkable swelling after the graft reaction. The H
NMR-spectra in D2O of poly(MADAM-BQ) and starch-
graft-poly(MADAM-BQ) are shown, for comparison, in
Figure S2 of Supporting Information.
3.3. Different Reactivities of the Radicals. It
contradicts the a priori expectation that R1* (iso-
butyronitrile radical bound to the starch, Figure 2)
should be less reactive in initiating the polymerization
than R2* (tert-butyl radical). The latter displays a strong
nucleophilic character and is known to react fast with
substituted alkenes.53 However, it was shown experi-
mentally by Simionescu et al.54 that this expected
behavior is, for example, not observed for a poly(dim-
ethylsiloxane) macroazoinitiator containing terminal
4-tert-butylazo-4-methylcyanobutyryl groups. In their
system, the macroinitiated polymerization of vinyl
monomers yields exclusively block copolymers. As dem-
onstrated by SEC and extraction experiments, no ho-
mopolymer is observed, which has to be produced if the
tert-butyl radicals were able to initiate the polymeriza-
tion. Simionescu suggested that the reason for the
absence of homopolymer formation is the rapid recom-
bination of tert-butyl radicals due to their high reactiv-
ity. This explanation is not satisfactory, though their
experimental results are convincing. Therefore, we
hoped also to observe a similar behavior for the starch
macroazoinitiator 7, i.e., the tert-butyl radical does not
initiate polymerization of cationic homopolymers.
To examine whether the tert-butyl radical initiates
polymerization when it is a fragment of a nonsymmetri-
cally substituted initiator, we compare the polymeriza-
tion of MADAM-BQ when: (i) initiated with 4-tert-

Macromolecules, Vol. 38, No. 17, 2005
Cationic Polymer Grafted Starch 7257
Figure 8. Chemical structures of low-molar mass initiators 4,4′-azobis-4-cyanovaleric acid (ABCVA) and 4-tert-butylazo-4-
cyanovaleric acid (t-BACVA) and their corresponding starch macroinitiators (I
5
, I10, and I13) for the investigation of the radical
efficiency of the tert-butyl radical and the cyanovaleric acid radical.
Figure 9. Rate of polymerization, R
starch macroinitiator I13 (dotted lines). It shows an ideal kinetic in aqueous solution, i.e., R
increases linearly with the initial monomer concentration [M] and the square root of initiator concentration [I]
p
, for MADAM-BQ when initiated with ABCVA (solid lines), t-BACVA (dashed lines), and
p
(determined by density measurements)
0
.5
0
0
. Reaction
-1
-
1
-3
-1
conditions were [M]
0
) 0.05-0.60 mol L , [I]
0
) 5.00 × 10 mol L for the variation of [M]
0 0 0
and [M] ) 0.30 mol L , [I] )
-
3
-1
1
.25-7.50 × 10 mol L for the variation of [I] , T ) 70 °C.
0
butylazo-4-cyanovaleric acid (t-BACVA), the low-molar
mass analogue to 7, and (ii) 4,4′-azobis-4-cyanovaleric
acid (ABCVA), the corresponding symmetric azo com-
pound (cf. Figure 8).
fragments, i.e., of the tert-butyl radical (ft-butyl) and the
cyanovaleric acid radical (fCVA). Then we get the ratio
ft-BACVA/fABCVA ) (fCVA + ft-butyl)/(2fCVA), which can only
be close to the calculated value of 0.53 if ft-butyl is close
to zero. Accordingly, we proved that the tert-butyl
radical does not initiate the polymerization of MADAM-
BQ in aqueous solution. This is in agreement with the
finding of Simionescu et al.37 and our SEC experiments
(cf. Figure 7). Nevertheless, the reason why the tert-
butyl radical does not initiate the polymerization of
MADAM-BQ is not clear yet. Probably, the highly
reactive tert-butyl radicals recombine. However, this
hypothesis seems impossible if the local concentration
is low. Alternatively, the highly reactive tert-butyl
radical could abstract hydrogen atoms from starch. This
would lead also to a grafting from the starch that cannot
be distinguished analytically from the expected graft
positions.
We chose MADAM-BQ because it is known as a
monomer showing an ideal kinetic polymerization in
aqueous solution when initiated, for example, with a
commercial initiator 2,2′-azobis(2-amidinepropane)
5
5
hydrochloride. The rate of the free-radical polymeri-
zation of MADAM-BQ in that study follows Rp )
0
.5
-1 0.5
K[I] [M]t)0 with K ) kp(f × kd × kt
) . The K is the
t)0
-
1 -2 0.5
overall rate constant in (L mol s ) , kp is the rate
-
1
-1
constant of the chain propagation in L mol s , kt is
-
1
the rate constant of the chain termination in L mol
-1
s , kd is the rate constant of the initiator decomposition
-
1
in s , and f is the radical efficiency. The [I]t)0 and
[
M]t)0, respectively, are the concentrations of initiator
and monomer when the polymerization starts (t ) 0).
2
We rearrange the expression for K to K /(f × kd) )
3.4. Graft Polymerization Kinetics of MADAM-
BQ. Determination of the Rate of Polymerization.
Conversion and rate of the graft polymerization were
estimated by measuring the change of the density of the
reaction mixture in situ using a very precise density
meter based on the oscillating U-tube principle to
monitor the polymerization of MADAM-BQ (partial
specific volumes of monomer and polymer differ). Details
of the use of the oscillating U-tube capillary for the
investigation of polymerization kinetics are given else-
where.5
2
-1
kp × kt , which is constant. Then we determined K
and kd experimentally. Figure 9 shows the rates of
polymerization of ABCVA, t-BACVA, and I13 (i.e., I with
DS ) 0.13) as a function of the initial monomer
concentration and initiator concentration, respectively.
The temperature was always 70 °C. It can be seen that
the rate of polymerization increases linearly with [M]t)0
0
.5
and [I] . The results show clearly that the Rp of
t)0
ABCVA is much larger than those of t-BACVA and I13.
Further, the Rp of the low-molar mass initiator t-BACVA
is close to that of the macromolecular one, I13. We
determined the overall rate constants by linear inter-
polation (slopes in Figure 9) to be K ) (6.67 ( 0.44) ×
6,57
In principle, this method is applicable by a foregoing
calibration of density versus conversion of a given
polymerizing system. We will show the possibility of
making precise measurements without such calibration.
The density of the solution, F, is given by
-
3
0.5
-0.5
-1
1
0
0
(L mol ) s for t-BACVA, and K ) (16.58 (
.41) × 10 (L0.5 mol ) s for ABCVA.
-
3
-0.5
-1
-
6
-1
The kd values are (9.5 ( 0.2) × 10
s
for t-BACVA
-
6
-1
and (31.1 ( 0.6) × 10
s
for ABCVA. The ratio of the
F ) F + (1 - v F )c
0
∑i
i
0
i
radical efficiencies was calculated as ft-BACVS/fABCVA )
2
2
(
Kt-BACVA /kd,t-BACVA)/(KABCVA /kd,ABCVA) ) 0.53. We can
-
3
presume that the radical efficiency of the initiator is
given as sum of the radical efficiencies of its two
where F0 is the density of the solvent (water) in g cm ,
-
3
ci is the mass concentration of the analyte i in g cm ,

7258 Bruzzano et al.
Macromolecules, Vol. 38, No. 17, 2005
and vi is the partial specific volume of the analyte i in
Table 2. Partial Specific Volumes of MADAM-BQ Polymer
3
-1
(v ) and Monomer (v ), the Density of Water, and the
cm g . Here, we have the three analytes MADAM-BQ,
poly(MADAM-BQ), and starch. It must be mentioned
that whether polymeric MADAM-BQ is bound to starch
or not cannot be distinguished by the density measure-
ment. This has to be proved by SEC. The sum of the
concentrations of MADAM-BQ and poly(MADAM-BQ)
p
m
Constant B at Temperatures of 20 and 70 °C
temperature
vp
vm
F0
B )
(vm - vp) F0
3
3
-3 a
[
°C ]
[cm g-1] [cm g-1] [g cm
]
20
70
0.7850
0.8071
0.8547
0.8954
0.99820
0.97784
0.06957
0.08634
(
in monomer units) is always equal to the starting
a
-3
The values given in the literature are 0.99820 g cm for 20
monomer concentration, cM,0. The density of the polymer
-3
61
°
C and 0.97776 g cm for 70 °C
reaction mixture as a function of time, F(cp,t), is given
by
F(c ) ) F + (1 - v F )c + (1 - v F )c +
M,0
p,t
0
st
0
st
M
0
(
v - v )F c
M P 0 P,t
where the v represents the specific volumes of starch
vst), MADAM-BQ (vM), and poly(MADAM-BQ) (vP).
(
After summarizing the terms that stay constant during
the polymerization and after introducing the conversion,
U ) (cM,0 - cM,t)/cM,0 ) cP,t/cM,0, we get
F(U) ) A + BcM,0
U
with
A ) F + (1 - v F )c + (1 - v F )c
0
st
0
st
M
0
M,0
Figure 10. (Inset) Density of the reaction mixture as a
function of the conversion at a temperature of 70 °C. The
monomer concentration of MADAM-BQ was [M]t)0 ) 0.1 mol
and B ) (v - v )F
0
M
P
-
1
L
5
, and the macroazoinitiator concentration was [I10] )
t)0
The density can be rewritten as a function of the
molar monomer concentration, [M], of the time t:
-3 -1
× 10 mol L . From the linear fit (straight line), we
calculated B ) (v - v ) F ) 0.09655. The density of the
m
p
0
reaction mixture as a function of reaction time during the
polymerization of MADAM-BQ at 70 °C (large figure). The
-3
-3
F([M]) ) A′ - BM [M](10 L cm ) with
M
-1
monomer concentration at reaction start was 0.1 mol L , the
initiator concentration (tert-butylazocyanovaleric acid) was
A′ ) F + (1 - v F )c + (1 - v F )c
0
st
0
st
p
0
M,0
-2
-1
10
mol L .
and B ) (v - v )F
0
M
P
2
,2′-azobis-(2-amidinopropane) dihydrochloride as an
-
1
MM is the molar mass of MADAM-BQ (283.4 g mol ).
58
initiator displays this ideal behavior. Therefore, it was
reasonable to assume that the polymerization kinetics
of MADAM-BQ would show a similar behavior when
started with the starch macroinitiators.
Differentiation with respect to time results in
dF([M])
dt
d[M]
dt
-3
-3
)
-BM (10 L cm )
M
The applicability of the linear relation between den-
sity and conversion for understanding the graft polym-
erization of MADAM-BQ from starch was proved. We
measured first the partial specific volumes of monomeric
and polymeric MADAM-BQ at 20 and 70 °C in water
which is equivalent to the overall reaction rate of the
polymerization, Rp. Assuming ideal kinetics for a free-
radical polymerization, the initial reaction rate, Rp0, is
given by
(Table 2). The high value of B at 70 °C calculated with
these data explains why a precise density measurement
is also sensitive for the measurement in the ratio of
monomer to polymer. Furthermore, B is positive, which
means that the density of the polymer is higher than
that of the monomer, and therefore, as expected, the
density of the reaction mixture increases during the
polymerization also in this special case.
d[M]t)0
Rp0
) -(
)
dt
t)0
fkd 0.5
0
.5
Rp0 ) K[I] [M]t)0 with K ) kp
t)0
(
)
kt
-
1
-2 0.5
where K is the overall rate constant in (L mol s )
,
Confirmation of the validity of the upper equation B
was determined by both conversion and density mea-
surements of a graft polymerization (70 °C, [M]t)0 ) 0.1
kp is the rate constant of the chain propagation in L
-
1
-1
mol s , kt is the rate constant of the chain termina-
-
1 -1
tion in L mol s , kd is the rate constant of the initiator
-1
-3
-1
mol L , [I10]t)0 ) 5 × 10 mol L ). After polymeri-
zation, we measured the conversion by SEC separation
of monomer and graft product (ratio of the peak areas
of the UV signal at a wavelength of 268 nm). It was
found that the conversion, as well as the density,
increases linearly with time as expected (see Figure 10,
insert). The conversation data and density data were
used to determine how the density of the solution
depends on the conversion of the monomer after the
elimination of the variable time. The result is shown in
Figure 10. From the slope, we calculated B ) 0.09655,
which is about 10% higher than given in Table 2. Both
values are within the expected experimental error.
-
1
decomposition in s , and f is the radical efficiency of
the initiator. The [I]t)0 and [M]t)0, respectively, are the
concentrations of initiator and monomer when the
polymerization starts. Combining both equations re-
sults in
dF([M]t)0
)
-3
-3
0.5
)
t)0
BM (10 L cm )K[I] [M]t)0
M t)0
(
)
dt
which describes ideal radical polymerization kinetics.
Further, Rp ) K[I] [M]t)0. It was shown earlier that
the polymerization of MADAM-BQ in solution using
0.5
t)0

Macromolecules, Vol. 38, No. 17, 2005
Cationic Polymer Grafted Starch 7259
of MADAM-BQ proves that the transfer activity of
MADAM-BQ can be neglected.
Second, we investigated the homopolymerization of
MADAM-BQ, initiated with 4-tert-butylazo-4-cyanova-
leric acid in the presence of different concentrations of
-1
starch (0.0-0.4 mol L ). We found a constant value for
-
6
-3 -1
(
dF[M]/dt)t)0 of (0.35 ( 0.02) × 10 g cm
s
and a
6
-1
constant molar mass of Mw ) (2.7 ( 0.2) × 10 g mol
-
1
when starting with [M]t)0 ) 0.30 mol L and [I]t)0 )
-
3
-1
5
× 10 mol L at T ) 70 °C. This proves that the
transfer activity of starch can also be neglected (CM ≈
) and that starch has no significant degrading influ-
0
ence, i.e., reduction of the polymer mass of poly-
Figure 11. lgR
z
-lgM-plot, as determined by SEC-MALLS,
/d lgM) increases with
(
MADAM-BQ).
shows that the slope (a ) d lgR
z
Next, we determined the overall rate constants for the
increasing M of starch-graft-poly(MADAM-BQ), starch and
poly(MADAM-BQ). The low value of a ) 0.26 found for starch
is indicative for a compact structure of the graft substrate in
aqueous solution and an increased swelling after graft polym-
erization.
polymerization of MADAM-BQ using starch macroini-
tiators with different DS. The values for K are (6.0 (
-3
-1 -2 0.5
-3
0
.6) × 10 (L mol s ) for I5, (6.3 ( 0.6) × 10 (L
-1 -2 0.5 -3 -1 -2 0.5
mol s ) for I10, and (5.7 ( 0.7) × 10 (L mol s )
for I13. Obviously, all the Ks are identical within the
error range. For comparison, the K of the low-molar
mass initiator 4-tert-butylazo-4-cyanovaleric acid is
Overall Rate and Transfer Constants. One aim
of this section is to prove whether transfer reactions
occur. Therefore, homopolymerizations of MADAM-BQ
were carried out by using different concentrations of a
low-mass initiator and also with the addition of starch.
The molar masses were determined by SEC-MALLS in
order to compare the polymerization when using differ-
ent initial concentrations of the monomer, [M]t)0, initia-
tor, [I]t)0, and the presence of chain-transfer agents. The
evaluation was carried out assuming kinetics as de-
-
3
-1 -2 0.5
slightly higher (6.7 ( 0.4) × 10 (L mol s ) . In
contrast to K, we found that the decomposition constant
of the initiators kd increases with higher DS. The kd
-
6
-1
values are (9.5 ( 0.1) × 10
s
for 4-tert-butylazo-4-
-
6
-1
cyanovaleric acid, (31.7 ( 4.4) × 10
s
for I5, (93.3 (
-
6
-1
-6 -1
1
8.9) × 10
s
for I10, and (180.5 ( 37.4) × 10
s
for I13. When kd increases and K is constant, it can be
concluded that the radical efficiency f decreases signifi-
cantly with an increase of DS (kd × f ∼ constant). This
seems to be evident when we assume that the higher
the DS of starch substrate, the more compact is its
structure in aqueous media such as monomer solutions
(cf. Section 3.1 on viscosity and SEC-MALLS measure-
ments). The thermally induced starch radicals are
probably not accessible enough for the monomers to
react and initiate graft polymerization efficiently. We
expected, therefore, that a higher DS does not neces-
sarily result in a shorter distance between two graft
chains (decrease of graft frequency). To investigate the
influence of the initiator and MADAM-BQ concentration
on the polymer characteristics, we used I5 with DS 0.05,
which is the macroinitiator with the lowest DS showing
best solubility in water.
5
9
scribed by the Mayo equation, which was applied to
the graft polymerization of MADAM-BQ from starch by
the expression:
[
I]0
[St]0
1
Pn
x
)
K′
+ C + C
M
[M]
st[M]
0
0
1/2
2
(fk k )
2fkd
d
t
with K′ )
)
akp
aK
Pn is the degree of polymerization of the grafted
poly(MADAM-BQ) chains (i.e., graft length), CM and Cst
are the chain-transfer constants of MADAM-BQ and
starch, respectively. [St]0 is the starch concentration at
the beginning of the polymerization, and K′ is a com-
bined constant. f is the initiator efficiency, kp is the rate
3.5. Polymer Structure. Variation of the Initia-
tor Concentration. The initial concentration of I5 was
varied for the graft polymerization at 70 °C in the range
-
1
-1
constant of chain propagation in L mol s , K is the
-
1
overall rate constant of the polymerization in (L mol
-
2 0.5
-1
-3
-1
s ) , and a is an integer (a ) 1 for termination by
disproportionation and a ) 2 for recombination).
For simplicity, we assume that neither water nor the
initiating azo group have chain-transfer properties. The
same assumption cannot be made a priori for the
MADAM-BQ monomer and starch because both are
capable of hydrogen abstraction. Therefore, we per-
formed the homopolymerization of MADAM-BQ with
and without the presence of starch using 4-tert-butylazo-
of 0.10-0.40 mol L (5-20 × 10 mol L concentra-
tion of azo groups), while the concentration of MADAM-
-
1
BQ was constant at 0.1 mol L . The results of the
molecular characteristics are summarized in Table 3.
It can be seen that the molar mass Mn,P-MADAM-BQ of
the grafted poly(MADAM-BQ) chains decreases from
3
-1
3
-1
289 × 10 g mol to 202 × 10 g mol with increasing
initiator concentration. These correspond to graft lengths,
Pl, of 1020-714 monomer units per chain. For initiator
-
1
4
-cyanovaleric acid as low-molar mass initiator.
concentration above 0.2 mol L , the monomer conver-
sions are similar (90-94%) and do not depend signifi-
cantly on the initiator concentration. The mass fraction
of poly(MADAM-BQ), wP-MADAM-BQ, decreases from 58
to 32%, and the fraction of starch that was not grafted
with poly(MADAM-BQ), wstarch, increases from 10 to
25% with increasing initiator concentration. The graft
frequency, νgraft, representing the average distance
between two graft chains, exceeds from 1290 to 2650
with increasing initiator concentration. This means
First, the homopolymerization was carried out with
different initial concentrations of the monomer [M]t)0
without starch and the same concentration of 4-tert-
butylazo-4-cyanovaleric acid. The transfer constant CM
was then determined from a Mayo plot of Pn-1 vs
1
/[M]t)0 (not shown). It was found that CM is in the
-
5
-5
range of 4.8 × 10 to 6.5 × 10 , which is comparable
-
5
to the values of methyl methacrylate (1 × 10 at 70
-
5
60
°C) and styrene (6 × 10 at 60 °C). The low CM value

7260 Bruzzano et al.
Macromolecules, Vol. 38, No. 17, 2005
Table 3. Graft Polymerization of MADAM-BQ with Different Concentrations of the Starch Azo Initiator I5 (DS ) 0.05)a
polymer
no.
[I5]
Mn,P-MADAM-BQ
wP-MADAM-BQ
wstarch
[%]
[mol L-1]
U
[10 g mol
3
-1
]
Pl
[%]
νgraft
Ngraft
1
2
3
4
0.10
0.20
0.30
0.40
74
91
94
90
289
260
243
202
1020
916
851
714
58
48
45
32
10
17
31
25
1291
1735
1819
2654
2.41
1.79
1.71
1.17
a
Monomer concentration of MADAM-BQ was 0.1 mol L-1, the polymerization was carried out at a temperature of 70 °C and stopped
after 180 min. U is the conversion of MADAM-BQ, Mn,P-MADAM-BQ is the number-average molar mass of the graft chains, Pl is the graft
length (number of MADAM-BQ monomers units per grafted chain), wP-MADAM-BQ is the mass fraction of poly(MADAM-BQ) in starch-
graft-poly(MADAM-BQ), wstarch is the mass fraction of nongrafted starch, νgraft is the graft frequency, and Ngraft is the number of grafted
chains per starch molecule.
Table 4. Graft Polymerization of MADAM-BQ with Different Concentrations of the Monomera
polymer
no.
[M]
Mn,P-MADAM-BQ
wP-MADAM-BQ
wstarch
[%]
[mol L-1]
U
[10 g mol
3
-1
]
Pl
[%]
νgraft
Ngraft
5
6
7
8
0.05
0.10
0.15
0.20
58
82
86
89
107
260
396
445
377
918
1399
1569
16
47
61
66
34
26
23
32
3463
1810
1564
1413
0.90
1.72
1.99
2.20
a
Initiator concentration (I5, DS ) 0.05) was 0.25 mol L-1 (12.5 × 10 mol L-1 concentration of azo groups), the polymerization was
carried out at a temperature of 70 °C and stopped after 180 min. U is the conversion of MADAM-BQ, Mn,P-MADAM-BQ is the number-
average molar mass of the graft chains, Pl is the graft length (number of MADAM-BQ monomers units per grafted chain), wP-MADAM-BQ
is the mass fraction of poly(MADAM-BQ) in starch-graft-poly(MADAM-BQ), wstarch is the mass fraction of nongrafted starch, vgraft is the
graft frequency, Ngraft is the number of grafted chains per starch molecule.
-3
that, in any case on the starch backbone, only less than
one of a thousand anhydroglucose units is grafted. By
combining the graft frequency with the molar mass
numbers, we calculated the number of poly(MADAM-
BQ) chains per starch molecule, Ngraft, which was found
to decrease from 2.41 to 1.17 with increasing concentra-
tion of starch initiator. This means that, on average,
we have only one to two grafted poly(MADAM-BQ)
chains per starch molecule. Consequently, the structure
of the starch-graft-poly(MADAM-BQ) can be assumed
to be similar to block copolymers of the AB and ABA
type, but with nondefined positions of the A blocks
with varying the initiator and monomer concentrations
are consistent.
Variation of DS. One may assume that a higher DS
of the initiator gives products with more grafted poly-
(
MADAM-BQ) chains per starch molecules. Therefore,
we used I10 to start the polymerization using the same
conditions as for I5 in polymerization number 2 (see
Table 3). Because of the higher DS of I10, we used for
-1
comparison a concentration of 0.10 mol L instead of
0
-1
.20 mol L . We found that the conversion is slightly
higher (93% instead of 91%) and, therefore, also the
mass fraction of poly(MADAM-BQ) (57% instead of
(poly(MADAM-BQ)) on the B blocks (starch). Structures
4
8%). Surprisingly, the length of the grafted poly-
similar to polymer brushes, which may be assumed a
priori for the grafting from starch, can be excluded on
the basis of the present data. It was of interest to see
whether the variation of the monomer concentration
gives similar results.
(
MADAM-BQ) chains increases (1220 instead of 916),
but the graft frequency is almost the same (1611 instead
of 1735); also, the number of grafted chains per starch
molecule remaining almost the same (1.93 instead of
1
.79). Obviously, the use of a starch macroinitiator with
Variation of the Monomer Concentration. For
this purpose, the concentration of MADAM-BQ was
a higher DS does not result in graft products with a
significantly lower graft frequency. We assume that the
reason for this unexpected result is a reduction in the
efficiency of the macroradicals with increasing DS due
to the decrease in solubility and tendency to aggregate
of higher-substituted starch initiators. Regarding the
amount of decomposed azo groups of I10 as a function
-
1
varied in the range of 0.05-0.20 mol L , while the
molar concentration of azo groups of I5 was held
-
3
-1
constant at 12.5 × 10 mol L . The results of the
molecular characteristics are summarized in Table 4.
It can be seen there that the Mn,P-MADAM-BQ of the
grafted poly(MADAM-BQ) chains increases from 107 ×
of the polymerization time t with ∆DS(t) ) DS ×
3
-1
3
-1
1
0 g mol to 445 × 10 g mol with increasing mono-
-kd×t
(1 - e
), the specific value of ∆DS (180 min) is equal
mer concentration. These correspond to graft lengths,
Pl, of 377-1569 monomer units per graft chain. The
conversions of MADAM-BQ increase with increasing
MADAM-BQ concentration from 58 to 89%. The per-
centage of poly(MADAM-BQ) by mass, wP-MADAM-BQ,
increases from 16 to 66%, and the percentage of starch
that was not grafted with poly(MADAM-BQ) varies
nonsystematically between 23 and 34%. The graft
frequency decreases from 3463 to 1413 with increasing
MADAM-BQ concentration. Again, less than one of a
thousand anhydroglucose units is grafted. The number
of poly(MADAM-BQ) chains per starch molecule in-
creases from 0.90 to 2.20 with increasing monomer
concentration. Again, on average, we have only one to
two grafted poly(MADAM-BQ) chains per starch mol-
ecule. In summary, the results of the polymerization
-6 -1
to 0.0635 for DS ) 0.10 and kd ) 93.3 × 10 s .
Thus, the reciprocal of 0.0635 (∼16) represents the
minimal graft distance, which can be obtained at this
point of the graft process, i.e., 100% radical efficiency.
Because we observed a hundred times larger graft
distance of 1611, the radical efficiency of I5 is expected
-2
to be in the region of 10 . The loss of radical efficiency,
especially for higher-substituted starch initiators, was
also verified by the previous kinetic measurements.
4
. Conclusions
The object of this work has been the design of a new
starch-based macroinitiator for the synthesis of cationic
polymer grafted starch. It was shown that the binding
of a nonsymmetric azo group is suitable by reaction of

Macromolecules, Vol. 38, No. 17, 2005
Cationic Polymer Grafted Starch 7261
starch with 4-tert-butylazo-4-cyanovaleric acid chloride.
The azo compound is linked to the starch predominantly
at carbon in the C6 position. This functionalization leads
to a hydrophobization of the starch, resulting in de-
creasing water solubility with increasing degree of
substitution as revealed by viscometry in solution.
The azo macroinitiator starts the polymerization of
MADAM-BQ, resulting in graft products with a high
monomer conversion and which are essentially free of
cationic homopolymers. In contrast to MADAM-BQ, it
was shown that the use of 2,3-dimethyl-1-vinylimid-
azolium chloride results in significant amounts of cat-
ionic homopolymer due to high transfer activity of this
monomer. It was found that, independent of the DS of
the macroinitiator, the number of grafted poly(MADAM-
BQ) chains from starch is nearly constant in the case
of the used starch substrate between one and two.
Therefore, the structures of the formed graft products
were always similar to that of block copolymers. Poly-
mer brush structures can be excluded. Obviously only
few azo groups of the starch initiators are able to start
the polymerization of MADAM-BQ. Their radical ef-
ficiency is significantly lower than those of low-mass
initiators and decreases with increasing DS. This is
mainly due to the compactness of the starch substrate
and the corresponding initiator, which in addition, leads
to some aggregation. We expect that the synthesized
starch-initiators could be useful as stabilizer and also
as initiator in emulsion polymerization. Investigations
of these applications are in progress.
(14) Khalil, M. I.; Mostafa, Kh. M.; Hebeish, A. Starch 1990, 42,
107.
(
(
(
(
15) Brockway, C. E.; Moser, K. B. J. Polym. Sci., Part. A: Polym.
Chem. 1963, 1, 1025.
16) Fanta, G. F.; Burr, R. C.; Russell, C. R.; Rist C. E. J. Appl.
Polym. Sci. 1970, 14, 2601.
17) McCormick, C. L.; SoonPark, L. J. Appl. Polym. Sci. 1981,
26, 1705.
18) Brockway, C. E.; Estes, R. R.; Smith, D. R. U.S. Patent
3,095,391, 1963.
(19) Jones, E. I.; Morgan, L. B.; Roberts, J. F. L.; Todd, S. M. Br.
Patent 715,194, 1954.
(
20) Khalil, M. I.; Mostafa, Kh. M.; Hebeish, A. Angew. Makromol.
Chem. 1993, 213, 43.
(
21) Mino, G.; Kaizerman, S. J. Polym. Sci. 1958, 31, 242.
(22) Vazquez, M. B.; Goni, I.; Gurruchaga, M.; Valero, M.; Guz-
man, G. M. J. Polym. Sci., Part A: Polym. Chem. 1992, 30,
1541.
(23) Vazquez, M. B.; Goni, I.; Gurruchaga, M.; Areizaga, J.; Valero,
M.; Guzman, G. M. Makromol. Chem. 1992, 193, 2189.
(24) Okieimen, F. E.; Ebhodaghe, J. E. Macromol. Rep. 1996, A33,
Suppl. 7/8, 401.
(
25) Patil, D. R.; Fanta, G. F J. Appl. Polym. Sci. 1993, 47, 1765.
26) Liu, M.; Cheng, R.; Wu, J.; Ma, C. J. Polym. Sci., Part A:
Polym. Chem. 1993, 31, 3181.
(
(27) Singh, H.; Thampy, R. T.; Chipalkatti, V. B. J. Polym. Sci.,
Part. A: Polym. Chem. 1965, 3, 4289.
(
28) Mehrotra, R.; R a˚ nby, B. J. Appl. Polym. Sci. 1977, 21, 1647.
29) Gao, J.-P.; Tian, R.-C.; Yu, J.-G.; Duan, M.-L. J. Appl. Polym.
Sci. 1994, 53, 1091.
(
(
30) Gruber, E.; Granzow C. Das Papier 1996, 6, 293.
(31) Faessinger, R. W.; Conte, J. S. U.S. Patent 3,095,391, 1963.
(
(
(
32) Maeher, G. G. Polymer 1979, 11, 85.
33) Bayazeed, A.; Farag, S.; Hebeish, A. Starch 1986, 38, 268.
34) Katai, A. A.; Schuerch, C. J. Polym. Sci., Part A-1 1966, 4,
2683.
(35) Kargin, V. A.; Kozlov, P. V.; Plate, N. A.; Konoreva, I. I.
Vyskomol. Soedin. 1959, 1, 114.
(36) Smets, G. J. Polym. Sci. 1961, 52, 1.
(37) Simionescu, C. I.; Dumitriu, S. J. Polym. Sci., Part C 1972,
Acknowledgment. This work was supported by
Fachagentur f u¨ r Nachwachsende Rohstoffe, the Fraun-
hofer Society, the Graduate School “Polymeric Materi-
als” of Technical University of Berlin, the Federal
Institute for Materials Research and Testing, and the
S u¨ dzucker AG, Ochsenfurt. We thank E. G o¨ rnitz (Fraun-
hofer IAP) for help with analytical ultracentrifugation
measurements, A. Haji Begli (S u¨ dzucker AG, Ochsen-
furt) for advice in starch modification, and T. Linker
and A. Laschewsky for helpful discussions.
37, 187.
(38) Speakman E. L., U.S. Patent 3,511,830, 1970.
(39) Kniewske, R.; Krause, F. European Patent 0173517A2, 1986.
(
40) Sugiyama, K.; Hosokawa, Y. Japanese Patent 10/087.702,
1998.
(
(
41) Tanaka, K.; Miura, H. Japanese Patent 06/287.243, 1994.
42) Tanaka, K.; Miura, H. Japanese Patent 06/412.503, 1994.
(43) Bhuniya, S. P.; Rahman, S.; Satyanand, A. J.; Gharia, M.
M.; Dave, A. M. J. Polym. Sci., Part A: Polym. Chem. 2003,
41, 1650.
(
44) Elfstrand, L.; Frigard, T.; Andersson, R.; Eliasson, A. C.;
Jonsson, M.; Reslow, M.; Wahlgren, M. Carbohydr. Polym.
2004, 57, 389.
Supporting Information Available: Thermal decompo-
1
sition of ABCVA, t-BACVA and I13 in aqueous solution; H
(45) Laschewsky, A.; Zerbe, I. Polymer 1991, 32, 11.
NMR-spectra in D
2
O of poly(MADAM-BQ) and starch-graft-
(46) Bushuk, W.; Benoit, H. Can. J. Chem. 1958, 36, 1616.
(
47) Linow, K.; Phillipp, B. Faserforsch. Textiltech. 1968, 19, 509.
48) Kratky, O.; Leopold, H.; Stabinger, H. Z. Angew. Phys. 1969,
poly(MADAM-BQ). This material is available free of charge
via the Internet at http://pubs.acs.org.
(
27, 273.
(
(
(
(
(
(
(
(
(
(
49) Heinze, T.; Rahn, K.; Jaspers, M.; Berghmans, H. J. Appl.
Polym. Sci. 1996, 60, 1891.
References and Notes
50) Rahn, K.; Diamantoglou, M.; Klemm, D.; Berghmans, H.;
Heinze T. Makromol. Chem. 1996, 238, 143.
(
1) Cho, C. G.; Lee, K. Carbohydr. Polym. 2002, 48, 125.
2) Janarthanan, P.; Yunus, W. M. Z. W.; Bin Ahmad, M. J Appl.
Polym. Sci. 2003, 90, 2053.
(
51) Katsura, S.; Isogai, A.; Onabe, F.; Usuda, M. Carbohydr.
Polym. 1992, 18, 283.
(3) Block and Graft Copolymers of Starch; Ceresa, R. J., Ed; John
52) Bohdanecky, M.; Kovar, J. Viscosity of Polymer Solutions;
Polymer Science Libary 2; Elsevier: New York, 1982.
53) Wong, M. W.; Pross, A.; Radom, L. J. Am. Chem. Soc. 1994,
Wiley & Sons: London, 1973; Vol. 1.
(4) Masuda, F.; Nishida, K.; Nakamura, A. U.S. Patent 4,076,663,
1978.
116, 11938.
(
(
5) Mooth, R. A. U.S. Patent 4,155,880, 1979.
6) Denzinger, W.; Hartmann, H. European Patent 441197A2,
54) Harabagiu, V.; Hamciuc, V.; Giurgiu, D.; Simionescu, B. C.;
Simionescu, C. I. Macrom. Rapid Commun. 1990, 11, 433.
55) Wandrey, C.; Zimmermann, A.; Jaeger, W.; Reichert, K.-H.
Polym. News 1997, 22, 390.
1991.
(7) The Carbohydrates. Chemistry and Biochemistry; Pigman, W.,
Horton, D., Wander, J. D., Eds.; Academic Press: New York,
56) Jaeger, W.; Wandrey, C.; Reinisch, G.; Linow, K.-J. Acta
Polym. 1979, 30, 113.
1980.
(
(
8) Mishina, A.; Nikumi, Z. Nippon Nogei Kagaku Kaishi 1964,
57) Wandrey, C.; Jaeger, W.; Reinisch, G. Acta Polym. 1981, 32,
38, 372.
197-202.
9) Fanta, G. F.; Burr, R.; Doane, W. M.; Russel, C. R. J. Appl.
Polym. Sci. 1977, 21, 425.
58) Zimmermann, A.; Jaeger, W.; Reichert, K.-H. Polym. News
998, 22, 390.
1
(
(
(
10) Herold, R.; Fouassier, J.-P. Starch 1981, 33, 90.
11) Carlsohn, H.; Hartmann, M. Acta Polym. 1982, 33, 640.
12) Hebeish, A.; Abd El-Thalouth, I.; El-Kashouti, M. A.; Abdel-
Fattah, S. H. Angew. Makromol. Chem. 1979, 78, 101.
13) El-Sisy, F.; Abadel-Hafiz, S. A.; El-Rafie, M. H.; Hebeish, A.
Acta Polym. 1990, 41, 324.
(
59) Mayo, F. R.; Lewis, F. M. J. Am. Chem. Soc. 1944, 66, 1594.
60) Makromolekulare Chemie, 2nd ed.; Lechner, M. D., Gehrke,
K., Nordmeier, E. H., Eds.; Birkh a¨ user Verlag: Basel, 1996.
61) Bettin, H.; Spieweck, F. PTB-Mitt. 1990, 100, 195.
(
(
(
MA048063F