A covalent-chemistry approach to giant macromolecules and their wetting behavior on solid substrates

Article abstract of DOI:10.1002/anie.200460390

Hair raising! Poly(methyl methacrylate) chains grown on a dendronized polymer (denpol) having repeat units with terminal 2-bromoisobutyramide groups produce a hairy surface. When adsorbed on mica, MoS₂, and highly oriented pyrolytic graphite, the hairy denpol molecules have a quasi two dimensional backbone conformation. The 50-nm-long hairs bend back on the MoS₂ (see image) and graphite substrates.

Full text of DOI:10.1002/anie.200460390

Angewandte
Chemie
Radical Polymerization
A Covalent-Chemistry Approach to Giant
Macromolecules and Their Wetting Behavior on
Solid Substrates**
Afang Zhang, Jörg Barner, Illdiko Göessl,
Jürgen P. Rabe,* and A. Dieter Schlüter*
Synthetic polymers commonly lack any sizable extension into
the second dimension (thickness) and are best described as
more or less flexible threads. Recently there have been two
synthetic approaches to address the problem of how to make
macromolecules “thick”: one is referred to as molecular
brushes,[1,2] the other as dendromized polymers.[3] Dendrom-
ized polymers (denpols) differ from molecular brushes in that
their side chains are not linear but highly and regularly
branched (dendrons),which results in a considerable density
of molecular segments in the direct neighborhood of the
central backbone. Denpols,especially those of higher gen-
eration,have the built-in character of a cylindrical object,a
feature which found application in basic experiments in
nanoscience in which the bottom-up approach was inves-
tigated.[4] These objects can reach lengths of several hundred
nanometers and widths of 5–7 nm depending on their
chemical structure and generation number and,in this
regard,are amongst the largest molecules ever prepared. It
seems difficult,for both synthetic and space reasons,to extend
the denpol concept to a degree such that the dimensions of
biologically relevant cylindrical objects such as the tobacco
mosaic virus (TMV envelope: diameter 18 nm,length
ca. 300 nm) can be mimicked. Therefore,we set out to
merge the brush and denpol concepts. Herein,we report a
joint synthetic and scanning force microscopy (SFM) inves-
[*] Dipl.-Phys. J. Barner, Dr. I. Göessl, Prof. Dr. J. P. Rabe
Institut für Physik, Humboldt Universität zu Berlin
Newtonstrasse 15, 12489 Berlin (Germany)
Fax: (+49)30-2093-7632
E-mail: rabe@physik.hu-berlin.de
Prof. Dr. A. Zhang,+ Prof. Dr. A. D. Schlüter++
Institut für Chemie, Freie Universität Berlin
Takustrasse 3, 14195 Berlin (Germany)
Fax: (+49)30-838-53357
E-mail: adschlue@chemie.fu-berlin.de
dieter.schluter@mat.ethz.ch
[+] New address:
College of Materials Engineering, Zhengzhou University
Daxue 75, Zhengzhou 450052 (China)
[
++] New address:
Department of Materials, Swiss Federal Institute of Technology
Institute of Polymers ETHZürich-Hönggerberg
Wolfgang-Pauli-Strasse 10, HCI J 541
8093 Zürich (Switzerland)
[**] We thank the German Science Foundation (Sfb 448, TPs A1 and
A11) for financial support. We cordially thank C. Ecker, HU Berlin,
for helpful discussions on aspects of persistence lengths.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 5185 –5188
DOI: 10.1002/anie.200460390
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5185

Communications
tigation of our attempts to further increase the diameter of
denpols. The synthetic issues comprise: a) synthesis of a
denpol macroinitiator whose number of initiator sites per
repeat unit (four) is inherently fixed; b) growth of poly-
(methyl methacrylate) (PMMA) chains off this denpolꢀs
“surface”[5] by using atom-transfer radical polymerization
(ATRP);[2a,6] c) determination of polymer hair length; and
d) estimation of the initiation efficiency. Since the starting
material already has a molar mass easily in the range of 1 ”
106 Daltons,it was clear that a complete characterization of
the product according to organic chemistry standards would
not be feasible. We therefore carried out a detailed character-
ization of the starting material,restricted ourselves at this
initial stage to second-generation (G2) denpols,and per-
formed model reactions for the growing process. SFM was
used to investigate the interaction of the hairy denpols with
solid substrate surfaces and served not only to further
characterize these huge molecules but also,and more
importantly,to study the response of hairy denpols to the
nature of the surface on which they are adsorbed. These
investigations specifically included: a) the imaging of spin-
coated samples on three different substrates,namely mica,
MoS2,and highly oriented pyrolytic graphite (HOPG);
b) characterization of the denpol backbone contour as well
as the structure of the hairy shell; and c) observation of
whether there is a systematic tendency of the hairs to fold
back as their interaction with the substrate becomes increas-
ingly less favorable. This latter aspect is considered important
for a distant goal of this research,which is to force the hairs to
completely fold back and produce cylindrical molecular
entities with unprecedented dimensions,for example,by
subsequent photochemical cross-linking.
The present study was carried out with the G2 denpol 2c,
which has four terminal 2-bromoisobutyramide groups in
each repeat unit. They serve as initiator sites for the ATRP of
methyl methacrylate. The synthesis starts from the known
G2 dendron 1,whose amine functions are completely con-
verted by reaction with the succinidyl active ester of 2-
bromoisobutyrate,and then the resultant 2a is converted into
the methacrylate G2 macromonomer 2b (Scheme 1). Com-
pounds 2a and 2b were fully characterized (see Supporting
Information). As has already been discovered for similar
dendronized macromonomers,simple heating of 2b to 558C
for a few hours in an extremely highly concentrated solution
(0.6 g of 2b in 0.4 mL of benzene) resulted in its polymer-
ization to the high-molecular weight polymer 2c. Notably,
polymer 2c automatically has four initiator sites per repeat
unit in this mode of synthesis.[7] Standard workup and
lyophilization from dioxane[8] afforded a colorless foam
(yields: 82–86%) which,despite its hygroscopic character,
gave correct data from combustion analysis (see Supporting
Information). Additionally,the 1H (500 MHz) and 13C NMR
spectra could be fully assigned,although the signals were
relatively broad as a result of the high molecular weight. Gel
permeation chromatography (GPC) in DMF calibrated with
G1 denpol standards[9] furnished number-average molar
masses between 900000 and 1500000 gmolꢀ1 and polydisper-
sity indices (PDI) of 3.4–4.0. The GPC elution curves were
basically monomodal (see Supporting Information). For the
Scheme 1. Reagents and conditions: a) 1, 2,5-dioxopyrrolidin-1-yl 2-bro-
moisobutyrate, TEA, CH2Cl2, MeOH, ꢀ308C, 15 h (86%); b) 2a, DCC,
HEMA, DMAP, CH2Cl2, RT, 15 h (68%); c) 2b, DCM, 558C, 8 h (85%).
TEA=triethylamine, DCC=N,N’-dicyclohexylcarbodiimide, HEMA=
hydroxyethylmethacrylic acid, DMAP=4-dimethylaminopyridine.
following experiments the highest molecular weight sample
with PDI = 3.4 was used.
Molecular brushes were prepared by grafting polymeric
chains from the backbone,[1e,f,2a] in which the most critical
aspect is cross-linking. ATRP is a so-called controlled radical
polymerization (CRP) procedure,which means that the
lifetime of a free radical chain end is reduced by the action
of a transition metal (CuI in the case of ATRP). Thus,chain
termination by radical recombination (and disproportiona-
tion),which would lead to intra- and intermolecular cross-
linking,is reduced. A rough estimation of the number of
initiator sites shows that polymer 2c can initiate 6000 PMMA
chains per individual 2c chain in the optimum case.[10]
Although not many of the chain ends will be active at the
same time,from this simple estimation it suggests that the
experimental conditions should be better designed such that
5186
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5185 –5188

Angewandte
Chemie
cross-linking is reduced as far as possible. Approximately 30
independent experiments to grow hairs off the denpols were
performed,each using 58 mg of 2c and 5 g of methyl
methacrylate (MMA). CuIBr and pentamethyldiethylenetri-
amine (PMDETA) were always employed in a constant molar
ratio to 2c; the reaction time,however,was varied from 0.5 to
2 hours to change the lengths of the hairs (for a detailed
synthetic procedure,see the Supporting Information). The
volume of the solvent DMF was critical.[11] In a number of
cases where 58 mg of polyinitiator 2c and 5 g of monomer
were dissolved in 5 mL of DMF cross-linking could instanta-
neously be visually recognized and no soluble product could
be isolated. Cross-linking still occurred if 10 mL of DMF was
used but some hairy denpols,specifically those with “short
hair” at low conversion and short reaction times (10 min),
could be isolated. The optimum results were finally obtained
with 20 mL of DMF,where only small amounts of insoluble
product were obtained (ca. 5 mass%). For both reaction
times all other material was fully soluble and was purified by
passing it through a short column of silica gel (100–200 mesh,
eluent: dichloromethane),which left the Cu salts behind as a
blue layer on the top of the column. This process was
associated with a considerable loss of material,which stayed
on the column and could not be eluted despite the fact that it
was fully soluble before.[12] Removal of solvent furnished 1 g
(0.5 h) and 1.2 g (2 h) of a colorless,transparent film (20 and
25% conversion,respectively,based on MMA). The apparent
number-average molar masses of 3 obtained by GPC were
2000000 (0.5 h) and 3500000 (2 h) (apparent PDIs: 2.5 and
1.4,respectively). These values need to be considered with
care because there is no appropriate standard available for
such complex molecular structures. The actual molar masses
will most probably be higher.[9] The hair lengths were roughly
estimated by hydrolyzation experiments with denpol 3 and a
model compound to be on the order of 20000–25000 gmolꢀ1
(PDIs ca. 1.3; see Supporting Information). These figures
were used to derive the number of initiator sites per repeat
unit of 2c that had actually been active. Assuming the above
hair molecular weights are correct and doing a simple mass
consideration,an initiator efficiency of roughly 50% (two out
of four) was obtained (see Supporting Information).
Denpol 3,which had been purified from nonbound
PMMA chains,[13] was spin-coated from diluted chloroform
solution onto mica,MoS 2,and highly oriented pyrolytic
graphite (HOPG) and investigated by SFM in the tapping
mode (Figure 1a–c). In all cases the adsorbed denpol
molecules exhibit a backbone conformation in quasi two
dimensions.[14] Single denpol molecules are found on mica and
MoS2 on atomically flat terraces,whereas on HOPG they are
predominantly in islands and at step edges because of a high
mobility on the HOPG terraces. The image on mica
(Figure 1a) reveals single denpol molecules with a central
backbone and a frayed-out corona,with a total width of about
122 nm (average value from 64 cross-sectional profiles) and a
thickness between 0.5 and 1 nm,which decreases gradually
from the backbone toward the edge. In addition,there are
small islands with the same apparent thickness as the corona
and no backbone. We attribute the corona to extended flat-
lying PMMA chains covalently linked to the denpol backbone
and the islands to free PMMA chains,which obviously had
not been removed during purification. Given a number-
average molar mass of Mn = 20000 gmolꢀ1,a fully extended
chain (hair) would on average be 50 nm long,which is in good
agreement with half of the corona width (61 nm). The
backbone conformation was analyzed by plotting hcosVi
versus ‘,where V is the change in orientation between two
segments at a distance ‘ apart along the backbone contour. A
fit to the wormlike chain model in 2D[15] would result in a
persistence length of 65 nm (see Supporting Information).
The corona exhibits holes and a sharp edge on MoS2.
Moreover,it is narrower (90 nm,average value from 114
cross-sectional profiles) and thicker (between 1 and 2 nm),
and the apparent persistence length of the backbone amounts
to 164 nm. The corona width is not so well defined on HOPG
since,because of their high mobility,free PMMA molecules
diffuse across the surface and into the corona of the denpol,
Figure 1. SFM images and cross-sectional profiles of 3 with PMMA hairs of Mn =20000–25000 gmolꢀ1 on a) mica, b) MoS2, and c) HOPG.
Angew. Chem. Int. Ed. 2004, 43, 5185 –5188 ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5187

Communications
[2] For recent reviews,see: a) K. Matyjaszewski,J. Xia, Chem. Rev.
2001, 101,2921 – 2990; b) N. Hadjichristidis,M. Pitsikalis,H.
Iatrou,S. Pispas, Macromol. Rapid Commun. 2003, 24,979 –
1013.
[3] A. D. Schlüter,J. P. Rabe, Angew. Chem. 2000, 112,860 – 880;
Angew. Chem. Int. Ed. 2000, 39,864 – 883; A. Zhang,L. Shu,Z.
Bo,A. D. Schlüter, Macromol. Chem. Phys. 2003 204,328 – 339;
A. D. Schlüter, Top. Curr. Chem. 2004,in press.
thereby increasing its thickness. A typical distance between
the backbones on the order of 50 nm is observed in the
islands,which is attributed to a corona width. The height
profile indicates a homogeneous layer of roughly constant
height of about 2 nm,which decreases sharply at the edge.
The determined hair lengths of approximately 50 nm
correspond well with half of the corona width on mica,which
indicates that the PMMA chains are more or less stretched
out perpendicular to the backbone direction. The fraying
reflects the polydispersity of the hairs. The narrower corona
on MoS2 shows that the hairs now deviate from an all-
stretched conformation by bending back. They dewet into
bundles that form a sharp corona edge,thus indicating a less
favorable interaction of the PMMA chains with MoS2 as
compared to mica. The relatively small distance of 50 nm
between the backbones on HOPG indicates that the PMMA
chains also bend back on this substrate. There is no indication
of the above kind of dewetting. The continuous increase in
apparent corona height on going from mica to MoS2 and
HOPG (Figure 1,bottom) suggests that the hair/substrate
interaction on HOPG is the least favorable.
The results show that the hair/substrate interaction has a
considerable effect on hair conformation. The hairs behave
quite differently on MoS2 and HOPG,but in either case the
all-stretched conformation perpendicular to the backbone
direction is avoided,which results in a much narrower corona
than on mica. The hairy denpols,with their sensitive response
to solid substrates,open the possibility of controlling the
persistence of a single polymer at a surface. This could
potentially be exploited for polymers having hairs of different
natures and lengths on the same substrate or,alternatively,for
the same polymer on different substrates. SFM studies of
hairy denpols clearly also allows the independent investiga-
tion of hairs and backbone. This can lead to a better
understanding of the mutual hair/backbone interaction.
The next steps include increasing the interaction of the
hairs with the denpol backbone and decreasing their inter-
action with the substrate. Also,it seems likely that the use of a
“thicker” backbone (e.g.,G4),where the hairs near to the
backbone “see” only other hairs,may also have a positive
effect.
[4] L. Shu,A. D. Schlüter,C. Ecker,N. Severin,J. P. Rabe,
Angew.
Chem. 2001, 113,4802 – 4805; Angew. Chem. Int. Ed. 2001, 40,
4666 – 4669; J. Barner,F. Mallwitz,L. Shu,A. D. Schlüter,J. P.
Rabe, Angew. Chem. 2003, 115,1976 – 1979; Angew. Chem. Int.
Ed. 2003, 42,1932 – 1935.
[5] A. Zhang,A. D. Schlüter, Polym. Prepr. Am. Chem. Soc. Div.
Polym. Chem. 2003, 44,524 – 525.
[6] M. Kamigaito,T. Ando,M. Sawamoto, Chem. Rev. 2001, 101,
3689 – 3746.
[7] In brush synthesis the polymeric initiators are normally prepared
by attaching initiator sites to a precursor polymer in an
independent synthesis step,which may lead to complications
and incompleteness. Structural proof of the resulting macro-
initiators does not always meet accepted standards.
[8] If the lyophilization was done in DMF or benzene,the polymer
was obtained in a morphology which did not dissolve or
molecularly disperse in DMF,the medium for the following
ATRP step,but instead formed a gel.
[9] A. Zhang,B. Zhang,E. Wꢁchtersbach,M. Schmidt,A. D.
Schlüter, Chem. Eur. J. 2003, 9,6083 – 6092; A Zhang,L. Okrasa,
T. Pakula,A. D. Schlüter, J. Am. Chem. Soc. 2004,126,6658 –
6666.
[10] Given an average molecular weight of 1500000 and 1000 for
polymer 2c and its repeat unit,respectively,an individual chain
has 1500 ” 4 = 6000 initiator sites. Of course,one has to consider
that not all the initiator sites will actually act as such for steric
reasons.
[11] Several experiments were also tried in toluene but cross-linking
could not be prevented. Presumably toluene does not dissolve or
disperse 2c. In a recent report on cross-linking reactions in
ATRP,Cu II salts were proposed to deactivate the reactivity of
the chain end: K. Matyjaszewski,S. Qin,J. R. Boyce,D.
Shiravanyants,S. S. Sheiko, Macromolecules 2003, 36,1843 –
1849.
[12] In three independent experiments these losses amounted to
59% (100 mesh),50% (100 mesh),and 76% (300–400 mesh). It
is assumed that it is the high-molecular weight hairy denpols
which strongly adhere to the silica surface. Tests confirmed that
nonconnected PMMA easily passed through the column.
[13] The denpols were purified by preparative GPC such that no
PMMA was detectable. Nevertheless,all samples still contained
some nonbound PMMA hairs,as observed in the SFM images.
[14] There must have been a process which converts the 3D struc-
tures into the 2D structures in solution.
Received: April 21,2004
Keywords: chain structures · dendrimers · polymers ·
.
scanning probe microscopy
[15] C. Rivetti,M. Guthold,C. Bustamante, J. Mol. Biol. 1996, 264,
919 – 932.
[1] a) Y. Tsukahara,K. Mizuno,A. Segawa,Y. Yamashita, Macro-
molecules 1989, 22,1546 – 1550; b) M. Wintermantel,M.
Schmidt,Y. Tsukahara,K. Kajiwara,S. Kohjiya,
Macromol.
Rapid Commun. 1994, 15,279 – 284; c) P. Dziezok,S. S. Sheiko,
K. Fischer,M. Schmidt,M. Möller, Angew. Chem. 1997, 109,
2894 – 2897; Angew. Chem. Int. Ed. Engl. 1997, 36,2812 – 2815;
d) S. Qin,K. Matyjaszewski,H. Xu,S. S. Sheiko,
Macromole-
cules 2003, 36,605 – 612; e) M. Zhang,T. Breiner,H. Mori,
A. H. E. Müller, Polymer 2003, 44,1449 – 1458; f) C. Li,N.
Gunari,K. Fischer,A. Janshoff,M. Schmidt, Angew. Chem. 2004,
116,1121 – 1124; Angew. Chem. Int. Ed. 2004, 43,1101 – 1104.
5188
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5185 –5188