Homologous series of dendronized polymethacrylates with a methyleneoxycarbonyl spacer between the backbone and dendritic side chain: Synthesis, characterization, and some bulk properties

Article abstract of DOI:10.1021/ja0494205

First through fourth generation (G1-G4) dendronized macromonomers, 3, 5, 7, and 9, with a methyleneoxycarbonyl spacer between the polymerizable group and dendritic side chain (dendron) were synthesized, and their polymerization behavior to the corresponding dendronized polymers PG1s, PG2s, PG3s, and PG4s, respectively, was investigated by heating the monomers to 55 °C without intentional addition of initiator. This self-induced polymerization is referred to as thermally induced radical polymerization (TRP). The molar masses of PG1s-PG4s were determined by gel permeation chromatography in DMF calibrated to a recently developed G1 dendronized polymer standard (PG1). A comparison of this homologous series' polymerization results with those of an already existing one, which differed only by the lack of this spacer (referred to as PG1-PG4), was made to contribute to the issue of whether short spacers have an effect on polymerization. Several representatives of both series were also used in the first systematic and generation-dependent investigation of these unusual comb polymers' bulk properties. Both structure and dynamics were investigated by DSC, X-ray diffraction, and dynamic mechanical measurements.

Full text of DOI:10.1021/ja0494205

Published on Web 05/07/2004
Homologous Series of Dendronized Polymethacrylates with a
Methyleneoxycarbonyl Spacer between the Backbone and
Dendritic Side Chain: Synthesis, Characterization, and Some
Bulk Properties
Afang Zhang,† Lidia Okrasa,§ Tadeusz Pakula,*,‡ and A. Dieter Schlu¨ter*,†,|
Contribution from the Institut fu¨r Chemie, Freie UniVersita¨t Berlin, Takustrasse 3,
14195 Berlin, Germany, Max-Planck Institut fu¨r Polymerforschung, Postfach 3148,
D-55021 Mainz, Germany, and Department of Molecular Physics, Technical UniVersity of Lodz,
Lodz, Poland
Received February 2, 2004; E-mail: adschlue@chemie.fu-berlin.de; pakula@mpip-mainz.mpg.de
Abstract: First through fourth generation (G1-G4) dendronized macromonomers, 3, 5, 7, and 9, with a
methyleneoxycarbonyl spacer between the polymerizable group and dendritic side chain (dendron) were
synthesized, and their polymerization behavior to the corresponding dendronized polymers PG1s, PG2s,
PG3s, and PG4s, respectively, was investigated by heating the monomers to 55 °C without intentional
addition of initiator. This self-induced polymerization is referred to as thermally induced radical polymerization
(TRP). The molar masses of PG1s-PG4s were determined by gel permeation chromatography in DMF
calibrated to a recently developed G1 dendronized polymer standard (PG1). A comparison of this
homologous series’ polymerization results with those of an already existing one, which differed only by the
lack of this spacer (referred to as PG1-PG4), was made to contribute to the issue of whether short spacers
have an effect on polymerization. Several representatives of both series were also used in the first systematic
and generation-dependent investigation of these unusual comb polymers’ bulk properties. Both structure
and dynamics were investigated by DSC, X-ray diffraction, and dynamic mechanical measurements.
1. Introduction
render it difficult to decide whether there is a spacer effect. For
example, such simple things as the monomers’ achievable degree
of purity as well as their melting points, mobilities in highly
concentrated solutions or in the melt, and solubilities in an
unpredictable way depend on the presence or absence of a spacer
and may well delude one to draw wrong conclusions from
experimental results. Especially, the latter two aspects may
become critical because polymerizations of macromonomers are
usually done in highly concentrated solution or even in bulk.4
If the monomer with spacer happens to give more highly
concentrated solutions, it may give higher molar mass product
just for that reason. We therefore set out to look more deeply
and systematically into this matter. Recently, a series of
polymethacrylate-based first through fourth generation (G1-
G4) dendronized polymers (PG1-PG4) were prepared.5 High
molar mass material was obtained up to the third generation
(PG3), and even G4 monomers gave high oligomers (PG4) in
high yields. We here report the synthesis and polymerization
of the exact same series of monomers with the only difference
being a methyleneoxycarbonyl spacer between the polymerizable
group (methacrylate) and the dendritic substituent. The three
factors of yield, approximate polymerization time, and molar
Up to now, most macromonomers applied in the synthesis
of dendronized polymers1 carry a spacer between the polymer-
izable group and the appendent dendron. Those without are
sometimes reported to only give oligomeric product.2 It was
reasoned that having a distance between these units may be
essential for the obtainment of high molar mass materials
because a spacer could help decrease steric congestion around
the polymerizable unit. On the other hand, Percec reported on
a considerable rate enhancement of the radical polymerization
of styrenic and acrylic dendronized macromonomers which did
not carry a spacer but were preorganized prior to polymerization
so that the polymerizable units were kept at a short distance
from one another by the packing behavior of their adjacent
mesogenic dendrons.3 Besides the question of whether pre-
organization plays a role, there are many other factors which
† Freie Universita¨t Berlin.
‡ Max-Planck Institut fu¨r Polymerforschung.
§ Technical University of Lodz.
| New address: Department of Materials, Institute of Polymers ETH-
Ho¨nggerberg, D-MATL, CH-8093 Zu¨rich, Switzerland.
(1) (a) Zhang, A.; Shu, L.; Bo, Z.; Schlu¨ter, A. D. Macromol. Chem. Phys.
2003, 204, 328-339. (b) Schlu¨ter, A. D.; Rabe, J. P. Angew. Chem., Int.
Ed. 2000, 39, 864-883.
(2) For example, see: (a) Scrivanti, A.; Fasan, S.; Matteoli, U.; Seraglia, R.;
Chessa, G. Macromol. Chem. Phys. 2000, 201, 326-329. (b) Chen, Y.-
M.; Chen, C.-F.; Li, X.-F.; Xi. F. Macromol. Rapid Commun. 1996, 17,
401-407. (c) Neubert, I.; Klopsch, R.; Claussen, W.; Schlu¨ter, A. D. Acta
Polym. 1996, 47, 455-459. (d) Draheim, G.; Ritter, H. Macromol. Chem.
Phys. 1995, 196, 2211-2222.
(3) Percec, V.; Ahn, C.-H.; Barboin, B. J. Am. Chem. Soc. 1997, 119, 12978-
12979.
(4) For example, see: Shu, L.; Schlu¨ter, A. D. Macromol. Chem. Phys. 2000,
201, 239-245.
(5) Zhang, A.; Zhang, B.; Wa¨chtersbach, E.; Schmidt, M.; Schlu¨ter, A. D.
Chem.-Eur. J. 2003, 9, 6083-6092.
9
6658
J. AM. CHEM. SOC. 2004, 126, 6658-6666
10.1021/ja0494205 CCC: $27.50 © 2004 American Chemical Society

Homologous Series of Dendronized Polymethacrylates
A R T I C L E S
Scheme 1 a
a Reagents and conditions: (a) 1a, TFA, 0 °C, 12 h (95%); (b) 1a, KOH, 50 °C, 6 h (91%); (c) 2a, 25% HCl, THF, 0 °C, 4 h (98%); (d) 2a, HEMA,
DMAP, DCC, DCM, room temperature, 12 h (98%); (e) 4a, DCC, HOSu, DCM, 12 h, -20 °C (98%); (f) 4a, 25% HCl, THF, 0 °C, 4 h (98%); (g) 4a, DCC,
DMAP, HEMA, DCM, room temperature, 12 h (92%); (h) 1b, 4b, TEA, DCM, -30 °C, 12 h (87%); (i) 6a, KOH, MeOH, THF, 12 h, 50 °C (86%); (j) 6b,
DCC, HOSu, DCM, -20 °C (92%); (k) 6b, HEMA, DMAP, DCC, DCM, room temperature, 12 h (96%).
mass of the obtained polymers PG1s-PG4s were used for
comparison with the above series and to address the spacer issue
(the small s indicates the existence of the spacer). Having both
of these homologous series in hand, we had another important
goal of substantially furthering the scarce knowledge on the
bulk properties of these unusual comb polymers by performing
9
J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004 6659

A R T I C L E S
Zhang et al.
Scheme 2 a
a Reagents and conditions: (a) 2b, 6c, TEA, DCM, MeOH, -20 °C, 12 h (75%); (b) 4b, 4c, TEA, MeOH, DCM, -30 °C, 12 h (78%); (c) 8, HEMA,
DCC, DMAP, DCM, room temperature, 12 h (92%).
a systematic, generation-dependent investigation by DSC, X-ray
diffraction, and dynamic mechanical analysis. As a side effect,
this should also provide insight into whether the short spacer
has an influence on bulk properties.
synthetic effort involved in 6c was considered too high. Finally,
the synthesis of 8 was done through the coupling of the two
G2 building blocks 4b and 4c. Dendron 8 was obtained in a
yield of 80%. The corresponding monomer 9 was obtained by
esterification with HEMA as described above. All intermediate
compounds and the four monomers were characterized by their
2. Results and Discussion
1
fully assigned H, 13C NMR spectra, mass spectra (FAB or
Macromonomer Synthesis. The syntheses of macromono-
mers 3, 5, 7, and 9 are shown in Schemes 1 and 2. For the first
two monomers 3 and 5, they are closely related to known work5,6
and will therefore not be further discussed in detail here. For
the couplings of the dendrons 2a and 4a to the polymerizable
unit, hydroxyethyl methacrylate (HEMA), the DCC method was
applied.7 The G3 dendron 6a was obtained from G2 active ester
4b and the G1 branching unit 1b in a yield of 78% despite the
relatively low excess of 4b (1.05 equiv per amine). This way
only two coupling steps were required and the achievable yield
was expected to be higher than in the opposite case (not shown)
in which a G1 fragment is attached to a G2 dendron with its
four amine coupling sites. The saponification of 6a required a
longer amount of time than in the related lower generation cases.
Coupling of G3 acid 6b to HEMA afforded the G3 macromono-
mer 7. The synthesis of the G4 dendron 8 was attempted from
different sides. First, the corresponding succinidyl active ester
6c was coupled to branching unit 1b. This gave a G4 dendron
with an ester group at its focal point (not shown). Its saponifica-
tion could not be achieved with standard methods because of
the relatively narrow range of conditions applicable. Basically,
starting material was recovered. Next, the reaction between the
active ester dendron 6c and the branching unit 2b was tried.
Branching unit 2b, in contrast to 1b, has a carboxylic acid at
its focal point. Although the coupling to 8 worked nicely, the
MALDI-TOF), and correct data from combustion analysis.
Thermally Induced Polymerization. Highly concentrated
solutions of the monomers 3, 5, 7, and 9 were heated to 55 °C
without any initiator added, and the resulting polymers PG1s-
PG4s, respectively, were isolated after predetermined times by
simple precipitation. Details on these polymerizations are given
in Table 1 and can also be found in the Supporting Information.
1
The polymers were characterized by their H and 13C NMR
spectra and the correct data from combustion analysis. Upon
increases in both generation and molar mass, the NMR signals
became broader and broader, and, thus, the assignment was less
reliable. The molar masses of all polymers were determined by
GPC in DMF using a recently developed PG1 standard.5
Representative results are shown in Table 1. Each entry has
been reproduced at least three times for PG1 and PG2 and two
times for PG3. The polymerization to PG4 has been carried
out only once.
A comparison of the dendronized polymers’ molar masses
achieved in both homologous series, PG1-PG4 and PG1s-PG4s,
can only be done on a qualitative level for reasons which include
the possible presence of unknown trace impurities in some of
the samples and the fact that for each monomer pair of the same
generation, but with and without oxyethylene spacer, respec-
tively, the concentrations of the saturated polymerization media
varied up to a few percent. For experimental simplicity, all
polymerizations were carried out in the same amount of solvent
using the same quantity of macromonomer. This concentration
(6) Zhang, A.; Vetter, S.; Schlu¨ter, A. D. Macromol. Chem. Phys. 2001, 202,
3301-3315.
(7) Sheenan, J. C.; Ledis, S. L. J. Am. Chem. Soc. 1973, 95, 875-879.
9
6660 J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004

Homologous Series of Dendronized Polymethacrylates
A R T I C L E S
Table 1. Conditions and Results for the Radical Polymerization of Monomers 3, 5, 7, and 9 To Give Polymers PG1s-PG4s, Respectively
(Entries 1-7), and for Comparison Purposes the Corresponding Information for PG1-PG4 (Entries 8-11)a
polymerization
conditionsb
GPC (PG1)c
M /M
entry
monomers
[M], mol/L
time, h
yield %
M × 10-4
M × 10-4
DP
n
DPw
n
w
w
n
1
2
3
4
5
3
3
5
5
7
7
9
G1
G2
G3
G4
1.22
1.22
0.51
0.51
0.25
0.25
0.14
1.36
0.60
0.30
0.14
16
72
14
20
16
38
48
5.5
48
30
48
89
88
52
57
48
50
12
78
86
50
88
162.5
149.2
84.8
77.0
40.8
43.1
11.7
150.8
73.0
81.9
15.0
560.6
608.7
351.9
365.0
119.5
123.3
16.3
642.4
247.5
325.1
24.2
3.45
4.08
4.15
4.74
2.93
2.86
1.39
4.26
3.39
3.97
1.61
2962
2719
715
650
166
175
23
3074
647
341
30
10 219
11 094
2967
3081
486
500
32
12 980
2193
1354
48
6
7d
8e
9e
10e
11e
a PG1s-PG4s and PG1-PG4 are referred to as dendronized polymethacrylate with and without a methyleneoxycarbonyl spacer between the polymerizable
group and dendritic side chain, respectively. b Carried out at 55 °C in benzene. c GPC in DMF at 30 °C calibrated with light scattering data obtained for
PG1. d DMF as polymerization solvent. e Data from Table 1 of ref 5.
difference may have had some effect on the achieved molar
mass. The comparison is based on entries 1-7 and 8-11 of
Table 1, respectively, the latter of which had already been
published.5 It can be seen that both of the molar mass ranges
achieved for the different generations vary only by ap-
proximately (10% except for the G3 case. Here, PG3s has by
a factor of 2 lower molar mass than PG3 which is actually in
contrast to what one intuitively might have expected. Also, the
trend to shorter chains with increasing dendron generation is
the same. Finally, as judged from the total polymerization times,
Figure 1. More adequate than commonly defined spacers in the “spacerless”
series of dendronized polymers PG1-PG4 and in the series with a so-
called methyleneoxycarbonyl spacer PG1s-PG4s.
there also does not seem to be much difference in polymerization
rate for comparable generations, although a thorough kinetic
study would be required to substantiate this point. Thus,
considering the large amount of samples of both series involved
in this study, it can be concluded with some confidence that
there is no significant spacer effect for all four macromonomer
pairs. Even for the G4 monomers, where steric hindrance
supposedly plays the most important role, basically the same
results were obtained (except for the yields).
connected to it in the para-position) represents a length of 3.7
Å. The “spacerless” PG1-PG4 series would then, in fact, have
a spacer of approximately 3.7 Å, and the PG1s-PG4s series
would have a spacer of 7.5 Å (Figure 1).
Polymer Characterization. Differential Scanning Calo-
rimetry (DSC). All polymers were at room temperature in an
amorphous glassy state. When heated, they underwent softening
related to the glass transition (Tg). Figure 2 shows the second
heating DSC traces for both PG1-PG4 and PG1s-PG4s series
with clearly visible enthalpy steps in all samples. The Tg values
were determined from the middle point in these steps and are
listed in Table 2. In both series, they increase with dendron
generation up to the third generation but become lower again
for the fourth generation.
This increase by approximately 10 °C is remarkable and
means that the segmental mobility for a temperature slightly
above the highest Tg could for different generations differ even
by 4 orders of magnitude (see below for a comparison with the
mechanical results). It can also be noticed that with increasing
generation the transitions become broader, which may be better
seen in the differentiated DSC traces shown in Figure 2c with
the PG1s-PG4s series as an example. Variation of both the
glass transition temperature and the width of the transition with
increasing dendron generation should be explained by the
corresponding change of the molecular architecture. Increasing
generation is, on one hand, related to an increase of the number
of branching points, but, on the other hand, it involves an
increase of the number of end groups. From the dynamic point
of view, the branching will locally reduce the segmental mobility
responsible for the glass transition, whereas the end groups will
Comment on Spacers. There seems to be no common
understanding as to what a spacer is in polymer synthesis. Most
vinyl polymerizable units require additional substituents to attain
sufficient reactivity to polymerize. For acrylates and styrenes,
for example, this would be the ester and phenyl group,
respectively. Because these substituents are a prerequisite for
the polymerization, they are not normally referred to as spacers,
although, besides their activating nature, they certainly contribute
to keep the substituents at the ester or phenyl groups apart from
the polymerizing vinyl. As a result, an acrylate monomer with
a methyleneoxy “spacer” keeps a substituent bound to the spacer
not much further apart from the polymerizable unit as a styrene
to which this very same substituent is bound directly in the para-
position, even though the use of the term spacer in the former
implies this. Although several factors influence the propensity
of a given polymerizable unit to polymerize in differently
substituted monomers of the same kind (e.g., G1-G4 den-
dronized methacrylates), it may be helpful to more rigorously
define the spacer as the entire fragment in its stretched
conformation which connects the vinyl carbon atom of the
polymerizable unit with an arbitrarily chosen part of the
substituent in question (e.g., a dendron). This way, the ester
group of MMA represents a spacer of approximately 2.5 Å
length, and the phenyl group in styrenes (if the substituent is
9
J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004 6661

A R T I C L E S
Zhang et al.
Figure 2. Second heating DSC thermograms for the dendronized polymers
without (a) and with spacer (b and c). In (c), the differentiated traces are
shown to better visualize the shift and broadening of the glass transition.
Figure 3. The X-ray diffractograms of (a) PG(n) and (b) PG(n)s series of
samples.
Table 2. Glass Transition Temperatures (Tg), Bulk
Backbone-Backbone Spacing (d), and the Effective Backbone
Length Per Repeat Unit (Lru) for Both Series of Dendronized
Polymers PG1-PG4 and PG1s-PG4s
increasing generation can be considered analogous to what has
been reported by others for flexible dendrimers in bulk, for
which the 1/M (M ) molar mass) dependence was detected,
indicating an increase to a limiting value at high molecular
mass.8 The trend of increasing Tg also does not continue beyond
G3 in the systems studied here. Both G4 homologues PG4 and
PG4s have an unexpectedly low Tg. This, however, may reflect
their comparatively low degrees of polymerization (DPs of
approximately 40) resulting in an effective reduction of the less
mobile core fraction because of the naturally higher mobility
at the backbone ends. This indicates that the problem of
segmental mobility in the dendronized polymers becomes
additionally related to the chain length. A separation of the
effects resulting from variation of possible parameters (genera-
tion, spacer length, backbone length) would, however, require
samples with better controlled molecular weights.
polymers
T [°C]
d [nm]
L [nm]
g
ru
PG1sa
PG2sb
PG3sc
PG4sd
PG1e
PG2f
55.8
71.4
79.6
68.2
55.9
64.2
76.5
65.6
2.78
4.03
5.17
0.151
0.155
0.195
2.69
4.06
5.05
5.59
0.145
0.145
0.199
0.335
PG3g
PG4h
a Entry 2 in Table 1. b Entry 3 in Table 1. c Entry 6 in Table 1. d Entry
7 in Table 1. e Entry 8 in Table 1. f Entry 9 in Table 1. g Entry 10 in Table
1. h Entry 11 in Table 1.
become more mobile. This can lead to a broader distribution of
segmental relaxation times and consequently a broadening of
the transition. This dynamic heterogeneity is probably also
spatially nonuniformly distributed because the highly branched
fragments of the repeat units are at least topologically closer to
the polymer backbone than the free ends. Consequently, the
probably cylindrically shaped macromolecules consist of the
less mobile core around the backbone and more mobile
periphery where elements of adjacent macromolecules inter-
penetrate each other to an extent which might be dependent on
dendrimer generation. An increase of the glass transition
temperature is generally related to a reduction of the segmental
mobility. Consequently, the observed considerable increase of
Tg for the first three generations should be related to a globally
slower segmental dynamics. It can be conjectured that the
increasing branching and additionally probably stronger seg-
mental zipping at the more and more crowded dendron
peripheries, which can be expected for higher generations,
contribute to this effect. The observed increase of Tg with
X-ray Diffraction. The polymers have an amorphous char-
acter; nevertheless, there are ordering effects, as detected by
X-ray diffraction. Examples of X-ray diffractograms recorded
at room temperature for samples of both series PG1-PG4 and
PG1s-PG4s are shown in Figure 3. Independent of the
polymer, the diffractograms indicate an amorphous halo at wide
angles and low angle peaks with positions, intensities, and
widths depending remarkably on the size of the side groups.
The wide angle halo seems to be asymmetric or consisting of
two components. The more intensive maximum corresponds to
real space distances of 0.4 nm, which can be attributed to spatial
correlations between some repetitive elements such as the
dendrons’ phenyl rings. The peaks at low angles correspond to
(8) (a) Farrington, P. J.; Hawker, C. J.; Fre´chet, J. M. J.; Mackay, M. E.
Macromolecules 1998, 31, 5043-5050. (b) Wooley, K. L.; Hawker, C. J.;
Pochan, J. M.; Fre´chet, J. M. J. Macromolecules 1993, 26, 1514-1517.
9
6662 J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004

Homologous Series of Dendronized Polymethacrylates
A R T I C L E S
dendronized polymers seem to follow this relation for the two
first generations but obviously deviate (Figure 4) for higher
generations. Here, the d values increase slower with increasing
Mru than would take place at constant backbone extension.
Qualitatively, this can be interpreted as an increasing backbone
extension with increasing dendron size and means that more
extended chains just thicken not as much with increasing side
group volume. This suggests that higher extension and conse-
quently higher effective stiffness of the backbone for higher
generations must be considered.
Taking into account the molar masses of the repeat units
(Mru), the backbone-backbone distances (d) can be estimated
as corresponding to a diameter of the cylindrical macromolecule
which can be described as follows:10
Figure 4. Dependencies of the backbone-backbone correlation distance
on the molecular weight of the repeat units in the dendronized polymers.
The dashed line represents a d ∼ M1/2 dependence for the maximally
extended backbone. Different symbols correspond to dendrimer units with
and without spacer.
d ) 2 M /πFN l
(1)
x
ru
A ru
distances exceeding one nanometer and in such amorphous
systems reflect the thickness of the densely branched polymer.9,10
High spatial requirements of the bulky dendritic side groups
probably make the backbone surroundings predominantly filled
by elements of the same chain. This means the polymers can
only weakly penetrate each other and consequently highly
exclude each other in space.
where F is the polymer bulk density (here assumed F ) 1,
because of lack of an experimental value), NA is the Avogadro
constant, and lru is the length of the backbone per monomer.
This dependence is shown in Figure 4 by a dashed line for the
special case of fully extended backbone (lru ) 0.25 nm) which
represents the smallest possible diameter for a given Mru. On
the other hand, having determined d, it is possible to estimate
the mean effective length of the backbone per repeat unit by
means of the same relation. The corresponding values (lru) are
listed in Table 2. They indicate that the chain extension increases
very fast starting from G3 for which values are reached close
to those expected for a completely extended backbone (0.195
and 0.199). For G4, in one case, no maximum is observed in
the expected s-range (Figure 3b), and, in the other case, the
peak’s scattering intensity is relatively weak (Figure 3a). The
corresponding s value translates into a backbone extension
exceeding the accessible limit (0.34). This might mean that the
fourth generation dendronized polymers of the given here
architecture have already too large mass to be polymerized into
long linear chains. Increasing packing problems result when the
more and more bulky and compact macromonomers are
anchored to the linear backbone one after the other. The
presented structural effect seems to explain the observed
limitations in achieving higher polymerization degrees for the
macromonomers of the fourth generation. Short chains could
still be possible because the intramolecular crowding can still
to some extent be relaxed at chain ends. Theoretical12 and
experimental studies13 on the curvature of dendronized polymers
are available.
This exclusion results in backbone-backbone distance cor-
relations as well as correlations between end groups surrounding
the cylindrical, but at higher temperatures still flexible, mac-
romolecules. Both of these correlations can contribute to the
low angle maxima in the recorded diffractograms and allow
extraction of information related to the polymer thicknesses.
Increasing intensities and decreasing width of the low angle
peaks for generations increasing up to 3 confirm the above
conjectures. This tendency seems, however, to break down at
the fourth generation. In the system without spacer, only a
considerable weaker low angle intensity peak is observed, and
no intensity maximum at all in this range is detected for the
samples with additional spacers. The values d ) K/smax are listed
in Table 2 and can be considered as nearly reflecting the
macromolecular diameters or the backbone-backbone distances.
The constant K is not exactly known if the correlation does not
extend over longer distances. Following the arguments of
Guinier,11 we assume here K ) 1.23, which is usually considered
for nearest neighbor correlations (for a hexagonal arrangement
of locally oriented cylindrical chains, K ) 1.15 should be taken,
for example). The systems studied seem from any point of view
to behave as amorphous. Therefore, larger scale order is not
expected.
To understand the above results, let us consider that the
backbone-backbone distance correlations d in a dense system
are controlled by both the spatial requirements of the side groups
(i.e., their sizes) and the backbone extension. For a backbone
contour length not influenced by the side group’s size, the
dependence between d and the molar mass of the repeat unit
(Mru) should be d ∼ Mru1/2, as a result of simple volume balance
consideration at a constant bulk density. The macromolecule is
regarded as a dense but flexible cylinder. This dependence is
represented in Figure 4 by a dashed line. Both series of the
Dynamic Mechanical Measurements. Despite their struc-
tured packing in the bulk, there is a considerable mobility in
dendronized polymers as detected by means of dynamic
mechanical measurements. They were characterized in a way
typical for amorphous melts, that is, by means of master curves
representing frequency dependencies of the real (G′) and
imaginary (G′′) components of the complex shear modulus in
a broad frequency range. This representation of the mechanical
behavior reflects usually a spectrum of properties of materials
at a chosen reference temperature in relation to various
relaxation processes of molecular or supramolecular elements.
(9) Pakula, T.; Minkin, P.; Beers, K. L.; Matyjaszewski, K. Abstr. Pap. Am.
Chem. Soc. 2001, 221, 559-562.
(10) Zhang, B.; Zhang, S.; Okrasa, L.; Pakula, T.; Stephan, T.; Schmidt, M., to
be submitted.
(12) Christopoulos, D. K.; Photinos, D. J.; Stimson, L. M.; Terzis, A. F.;
Vanakaras, A. G. J. Mater. Chem. 2003, 13, 2756-2764.
(13) Ecker, C.; Severin, N.; Shu, L.; Schlu¨ter, A. D.; Rabe, J. P. Macromolecules
2004, 37, 2484-2489.
(11) Guinier, A. X-ray Diffraction in Crystals, Imperfect Crystals and Amorphous
Bodies; W. H. Freeman & Co.: San Francisco, CA, 1963.
9
J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004 6663

A R T I C L E S
Zhang et al.
Figure 6. Temperature dependencies of the segmental relaxation times
for the dendronized polymers of various generations. The arrows point
toward Tg’s as detected by DSC. The vertical dashed line shows the reference
temperature for the results of Figure 5.
geneity of the systems with increasing dendron generations is,
as discussed earlier, also reflected in the mechanical spectra by
a remarkable broadening of the segmental relaxation process.
An attempt to characterize this broadening by means of the
parameters of the Fourier transformed stretched exponential
function, exp(-(t/τ)â), using just a graphical approximation,
indicated that the exponent â decreases from the value of nearly
0.4 to a very low value of 0.2 between G1 and G4, respectively.
The mechanical spectra shown in Figure 5, when considered
at deformation frequencies below the frequency at which the
segmental relaxation is observed, can be correlated further with
the parameters of the specific architecture of the polymers. These
parts of the spectra can also be considered as reflecting behavior
above the glass transition temperature when tested isochroni-
cally. In this range of testing conditions, the samples become
viscoelastic with a broad spectrum of relaxations characteristic
for each of them. Two parameters, the backbone length and
the sizes of the anchored dendrons, determine this behavior.
Unfortunately, these parameters in the studied samples do not
change independently, and the G′ and G′′ dependencies are
determined in a limited frequency range, because measurements
at temperatures higher than 130 °C led to irreversible changes
of the structure and properties probably resulting from cross-
linking. Above this temperature, the systems cannot be consid-
ered as rheologically simple. An example of such an effect is
illustrated in Figure 7, in which the low-frequency part of the
viscoelastic spectrum changes irreversibly with increasing
temperature, indicating a formation of the rubbery state with
modulus increasing with annealing temperature and time.
A common observation in the results presented in Figure 5
is that when the segmental relaxation starts, the modulus drops,
however, in different ways for different samples. Taking into
account the limited accessibility of the mechanical characteristics
of these polymers, the low-frequency part of the observed
properties can be understood on the basis of analogies of these
polymers with other relevant systems such as brushlike poly-
mers10,15,16 or stars.14,17 There are indications in the results
presented in Figure 5 that for all polymers, there is a slower
relaxation process which can be attributed to relaxations of the
Figure 5. Frequency dependencies of the real (G′) and imaginary (G′′)
parts of the complex shear modulus (master curves at a reference temperature
Tref ) 100 °C) for the dendronized polymers (a) generations 1-3 and (b)
generations 3 and 4.
The properties may usually vary between those characteristic
for the glassy state and those for a liquidlike state as indicated
by the dynamic mechanical behavior at high and at low
frequencies, respectively.14 Details of the behavior between these
extremes depend on the molecular structure. Examples of the
results of the dynamic mechanical characterization of the series
of dendronized polymers at the common reference temperature
of 100 °C are shown in Figure 5.
As already observed in the DSC traces, the polymers have a
glassy state. According to the mechanical tests, they are
characterized in this state by modulus values exceeding 108 Pa
(the plateaus at high frequencies). This range of the spectra
corresponds to deformation rates (frequencies) which are faster
than segmental mobilities of the polymers, and the cross points
between G′ and G′′ observed at the low-frequency edges of these
plateaus indicate the deformation rates which become compa-
rable with the segmental relaxations. The results presented in
Figure 5 indicate considerable differences in the segmental
mobility between different samples which reflect the differences
in the glass transition temperatures observed in the DSC. In
Figure 6, the segmental relaxation times determined from the
mechanical behavior of polymers are shown as a function of
temperature and compared with temperatures of the glass
transitions determined by DSC (full triangles). The results show
that the segmental relaxation times in various polymers reach
values of 10 s nearly at temperatures corresponding to the Tg
values. This shows a good agreement between the mechanical
and calorimetric observations. The increasing dynamic hetero-
(15) Tsukahara, Y.; Namba, S.; Iwasa, J.; Nakano, Y.; Kaeriyama, K.; Takahashi,
M. Macromolecules 2001, 34, 2624-2629.
(16) Pakula, T.; Minkin, P.; Beers, K. L.; Matyjaszewski, K., in preparation.
(17) Pakula, T.; Vlassopoulos, D.; Fytas, G.; Roovers, J. Macromolecules 1998,
31, 8931-8940.
(14) Pakula, T.; Geyler, S.; Edling, T.; Boese, D. Rheol. Acta 1996, 35, 631-
644.
9
6664 J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004

Homologous Series of Dendronized Polymethacrylates
A R T I C L E S
dendronized polymers PG1s-PG4s as analytically pure materi-
als. The first three members of this homologous series exhibit
high molar masses. The G4 monomer still gave high oligomers
in high yield. This series of polymers was compared with an
already existing one (PG1-PG4), which lacks this spacer, to
find out whether the spacer has an effect on polymerization
behavior. As judged by the factors of total polymerization time,
yield, and molar mass, no significant difference was observed
between these two series. This indicates that in cases where a
spacer effect was proposed other reasons should also be taken
into consideration.
Analysis of the structure of the dendronized polymers PG1-
PG4 and PG1s-PG4s in bulk indicates backbone-backbone
distance correlations giving rise to a distinct small angle peak
in the X-ray scattering intensities, however, without any
signatures of a longer range order as that suggested for systems
studied by others.20,21 The correlation improved with increasing
generation up to G3 but was almost lost for G4. Considering
the detected correlation distances under the condition of dense
packing of polymers in bulk leads to the conclusion that
anchoring of bulky dendrons to the backbone with a high
grafting density results in backbone extensions. These extensions
increase drastically for the higher generations (G3 and G4). In
the case of the highest studied generation (G4), the packing
problems seem to impose limits on the growth process so that
high numbers of repeat units cannot be reached. Probably
because of the small length of the methyleneoxycarbonyl spacer,
there are no significant differences between the samples with
and without spacer.
Figure 7. Frequency dependencies of the real (G′) and imaginary (G′′)
parts of the complex shear modulus (master curves at a reference temperature
Tref ) 100 °C) for the dendronized polymer. At higher measurement
temperatures, the system irreversibly changes properties (see the low-
frequency plateau of G′).
side groups. This relaxation can be characterized by a plateau
of G′ at a typical polymeric level (between 105 and 106 Pa) and
by relaxation times becoming slower with increasing side group
sizes. For the polymer PG3s, the G′ stays above the level of
105 Pa up to the lowest seen frequencies, which can mean that
this relaxation dominates the behavior becoming too slow to
be observed in the accessible range.
Relaxation of the side groups leads to a further drop of the
G′ values, which is seen in polymers PG1s and PG2s, but only
to the level of the order of 104 Pa which can be considered as
controlled by the relaxation of the entangled backbones if
sufficiently long. A flow is not observed in any of the samples,
but the low-frequency G′ plateau is more distinct for the PG1
with a higher backbone length. The new G′ plateau at the lowest
frequencies means that the global polymer relaxation takes
longer times. This terminal relaxation is not really observed in
any of the systems studied here, but this seems to be limited by
the instability of these systems for thermally induced cross-
linking as indicated in Figure 7. The cross-linking seems to start
at temperatures lower than those at which the chain relaxation
would have an observable rate. The plateaus of G′ observed
for samples PG1s and PG2s at the lowest frequencies (Figure
5) are in the range that is characteristic for polymeric or colloidal
systems with a large number of pendant groups.9,15-19 The
behavior of PG4s cannot be precisely interpreted, but probably
the small backbone length and high polydispersity are mainly
responsible for the fast drop of G′ despite the largest side groups.
An exact analysis of the systems in the discussed frequency
range would require polymers with comparable lengths of the
backbones and possibly with a lower polydispersity, but first
with polymers of higher thermal stability.
Analysis of the dynamics in the studied systems indicates
remarkable effects of the polymer architecture on the segmental
relaxation controlling the transition to the glassy state. Increasing
the dendron generations led to a considerable increase of the
temperature of the materials’ main softening. This temperature
increase is equivalent to a slowing down of the segmental
relaxation, when the different polymers are compared at a given
temperature. This observation was in good agreement with DSC
results. Here, the increase of the dendron generations involved
also a considerable broadening of the segmental relaxation
process, indicating an increasing dynamic heterogeneity of the
systems. Above the glass transition, the viscoelastic properties
are dominated by either side group or backbone relaxation
depending on the sizes of these structural elements. In the case
of the longer chains with the smaller G1 and G2 dendrons, the
side group relaxations are relatively fast, and the low-frequency
G′ plateaus at the level of 104 Pa are indicative of nonrelaxed
backbones. For the more internally crowded and extended
polymers with G3 dendrons, the behavior was controlled by
the nonrelaxed side groups. This was concluded because the
low-frequency plateau remained on the level exceeding 105 Pa.
Unfortunately, the thermal instability of the systems at temper-
atures higher than 130 °C made impossible a comprehensive
analysis of the polymers’ dynamic behavior up to the flow range.
In the performed characterization of structure and dynamics,
no remarkable differences have been detected for samples with
and without spacers.
(18) Gohr, K.; Pakula, T.; Tsutsumi, K.; Scha¨rt, W. Macromolecules 1999, 32,
7156-7165.
3. Conclusions
(19) Lindenblatt, G.; Scha¨rtl, W.; Pakula, T.; Schmidt, M. Macromolecules 2001,
34, 1730-1736.
The thermally induced radical polymerization of the G1-
G4 dendronized methacrylate macromonomers 3, 5, 7, and 9
with a methyleneoxycarbonyl spacer gave the corresponding
(20) Kwon, Y. K.; Chvalun, S. N.; Blackwell, J.; Percec, V.; Heck, J. A.
Macromolecules 1995, 28, 1552-1558.
(21) Rapp, A.; Schnell, I.; Sebastiani, D.; Brown, S. P.; Percec, V.; Spiess, H.
W. J. Am. Chem. Soc. 2003, 125, 13284-13297.
9
J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004 6665

A R T I C L E S
Zhang et al.
Also, in the dynamics and resulting properties, no effect of the
short spacer has been observed.
(Mainz) for doing the molar mass determinations and for help
with the interpretation.
Acknowledgment. This work was financially supported by
the Deutsche Forschungsgemeinschaft (Sfb 448, TP A1), the
Fonds der Chemischen Industrie, and the Bundesministerium
fu¨r Bildung, Forschung, und Technologie (WTZ-Project POL
01/048). We cordially thank Prof. Manfred Schmidt and his crew
Supporting Information Available: All synthetic procedures
and analytical data. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0494205
9
6666 J. AM. CHEM. SOC. VOL. 126, NO. 21, 2004