Radical polymerization tracked by single molecule spectroscopy

Article abstract of DOI:10.1002/anie.200704196

(Figure Presented) Polymerization - a single molecule view: Molecular motion of free and incorporated single perylene diimide derivatives during radical polymerization of styrene and styrene networks, studied by a combination of fluorescence correlation spectroscopy and wide-field microscopy, can be observed over the entire conversion range from free diffusion to immobilization of the dyes. In particular, the presence of evolving heterogeneities could be visualized.

Full text of DOI:10.1002/anie.200704196

Angewandte
Chemie
DOI: 10.1002/anie.200704196
Polymerization
Radical Polymerization Tracked by Single Molecule Spectroscopy**
Dominik Wöll, Hiroshi Uji-i, Tobias Schnitzler, Jun-ichi Hotta, Peter Dedecker,
Andreas Herrmann, Frans C. De Schryver, Klaus Müllen, and Johan Hofkens*
Products from polymerization have influenced our daily lives
tremendously over the past hundred years. It is therefore not
surprising that enormous efforts have been and are being
made to fully understand each detail of the polymerization
process. Prominent examples of analytical techniques used
are ESR spectroscopy,[1] pulsed-laser-initiated polymerization
in conjunction with size-exclusion chromatography[2] and
mass spectrometry,[3] as well as NMR[4] and fluorescence
spectroscopy.[5–11] With these techniques, very detailed knowl-
edge about polymerization kinetics could be gained. How-
ever, most of these techniques cannot probe over a large
extent of conversion and all average over an ensemble of
molecules. As a consequence, none of these techniques can
detect heterogeneities occurring at molecular level during
polymerization, a phenomenon which influences the final
polymer properties. In contrast to ensemble techniques, single
molecule spectroscopy (SMS) can elucidate such heteroge-
neities.
related to the freedom of molecules to move within the
monomer solution or matrix formed by the polymer. FCS
measures the time diffusing molecules remain within a
defined volume and allowed determination of D >
10ꢀ13 m2 sꢀ1 up to high conversions U before motion became
too slow and thus limited the applicability of this method.
Wide-field microscopy (WFM) directly visualizes the position
of fluorescent molecules and is a suitable method to track
slow moving molecules (D < 10ꢀ12 m2 sꢀ1) and even to detect
molecules which are immobilized. Both methods, therefore,
complement each other, and in combination, they permit
following translational motion of dyes for the entire poly-
merization process. The detection of heterogeneities is an
important advantage over determination of an average value
of D from viscosity measurements using the Stokes–Einstein
relation.
Perylenediimide derivatives were used as probing dyes.
Using dye molecules 1 or 2, the polymerization of styrene in
absence and presence of a cross-linker was studied. The dye 1
was of particular interest as it moves more slowly because of
its large size and, therefore, allows WFM detection already at
lower U. Additionally, perylenediimide derivative 3 bearing
two styrenyl groups (for the synthesis, see the Supporting
Information) allowed for the formation of polystyrene with
the dye acting as a potential cross-linker which is covalently
SMS has already been used to study the dynamics of single
molecules[12–14] or single polymer chains[15,16] in a polymer
matrix. In particular, fluorescence correlation spectroscopy
(FCS) allowed the study of diffusion in polymer solutions, gels
with different cross-linker concentrations,[17] poly(acrylic
acid) grafted on poly(ethylene terephthalate) films,[15] and
thrombin-induced fibrin aggregation.[18]
Herein we present SMS measurements performed for the
first time during bulk radical polymerization in the absence
and presence of a cross-linker. In particular, we follow
polymerization by detecting changes of the diffusion constant
D of dye molecules acting as probes. Changes in D can be
[*] Dr. D. Wöll, Dr. H. Uji-i, Dr. J.-I. Hotta, P. Dedecker,
Prof. Dr. F. C. De Schryver, Prof. Dr. J. Hofkens
Department of Chemistry
Katholieke Universiteit Leuven
Celestijnenlaan 200 F, 3001 Heverlee (Belgium)
Fax: (+32)1632-7990
E-mail: johan.hofkens@chem.kuleuven.be
T. Schnitzler, Dr. A. Herrmann, Prof. Dr. K. Müllen
Max-Planck-Institut für Polymerforschung
Ackermannweg 10, 55128 Mainz (Germany)
[**] We thank Prof. Dr. J. Enderlein (Universität Tübingen) and Wako
Chemicals (especially Dr. A. Kraetzschmar) for helpful advice and
Prof. Dr. C. Detrembleur (UniversitØ de Li›ge) for supplying V70.
Support from the Fonds voor Wetenschapplijk Onderzoek Vlaan-
deren (FWO grantG.0366.06), the KULeuven Research Fund (GOA
2006/2), the Flemish Ministry of Education (ZWAP 04/007), and the
Federal Science Policy of Belgium (IAP-VI/27) are gratefully
acknowledged. D.W. thanks the German Academic Exchange
Service (DAAD) for a postdoctoral scholarship.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2008, 47, 783 –787
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
783

Communications
incorporated into the polymer.[19] Dye 2 is of similar size as 3
and was used as reference.
All polymerizations were initiated with the thermal
radical initiator 4 and were performed at room temperature.
Three sets of experimental conditions were selected: the
polymerization of styrene probed with 1 or 2, the polymer-
ization of styrene with cross-linker 5 probed with 1, and the
polymerization of styrene with 3. The kinetics of the
polymerizations are presented in the Supporting Information.
For the polymerization probed with 1 and no cross-linker,
FCS autocorrelation curves were fitted with one diffusion
time t (Equation S2, Supporting Information) up to high
conversion (0.83). The resulting diffusion constants D are
shown as gray circles in Figure 1a and Figure 1b. The
reasonable fits with one D value indicate that translational
diffusion of the dyes is rather homogeneous. Even at higher
conversion U, a lot of freedom remains for the dye molecules
to move because the surrounding polymer chains are loosely
entangled and are able to move along each other. The motion
of dye molecules with D below about 3 ” 10ꢀ12 m2 sꢀ1 can be
directly observed in WFM (for movies see the Supporting
Information). However, at early stages of polymerization, the
dyes move too fast for the time resolution of the CCD camera
(an integration time of 38 ms per frame (26 Hz) was used for
all reported WFM experiments). Even though no dye
molecules could be localized at this stage, their traces can
be recognized in the WFM movies at 0.42U. The progress of
polymerization causes the translation of the dye molecules to
gradually slow down. The movie taken at 0.64U shows that
molecules become occasionally detectable for a few hundreds
of milliseconds before they move out of the focus (see
Figure 2b, left). It takes until about 0.70U before the motion
of most molecules slows down sufficiently for tracking (see
Figure S5 and movies in Supporting Information). At the final
stage of polymerization, the slow translational diffusion, still
detectable at 0.85U, finally stops at about 0.90U. At that
stage, the strongly hindered motion of molecules results in
very low polymerization rates.[20]
A second set of experiments focused on the diffusion in
polymer networks. For that purpose, the above mentioned
experimental conditions were repeated, but with addition of
1,4-divinylbenzene (5) as cross-linker in a concentration of
1% and 3%, respectively (Figure 1a and b). At low U,
autocorrelation curves, the quality of their fits, and diffusion
constants were similar to the experiment without 5. However,
when gelation started, the FCS curves could only be fitted
acceptably with two diffusion constants (Equation S3,
Supporting Information) as indicated by triangles (fast:
green, slow: red; black circles: weighted average given by
Equation S5, Supporting Information). The fast component
of D shows a decrease relative to that of D for 1 in the
experiment without cross-linker (gray circles). In contrast, at
approximately the reaction time when gelation was observed,
the slow component of D dropped by about one order of
magnitude for a cross-linker concentration of 1% and circa
two orders of magnitude for 3%. At 3%, gelation also started
Figure 1. Dependence of the diffusion constant D obtained by FCS on
U and on the polymerization time: a,b) for polymerization of styrene
probed with 1 and 0%, 1%, and 3% w/w of cross-linker 5. The
correlation between polymerization time and conversion is valid until
0.9U, 0.6U, and 0.4U, respectively (Figure S2, Supporting
Information) c) for the polymerization of styrene with 3 and as
comparison of styrene probed with 2. Each measurementwas repeated
three times at different positions. The dashed line indicates the point
atwhich all molecules were immobilized as determined by WFM.
earlier than at 1% (observed ratio of gelation times, 0.47/0.25
pffiffiffi
ꢁ 1.9, is close to the theoretical value of 3[21]). The WFM
movies (see Supporting Information) show that before
784
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 783 –787

Angewandte
Chemie
Figure 2. a) Movie extract from the indicated region in Figure 2b (middle, gray rectangle) with one immobilized and one moving dye (only every
second frame is shown). b) WFM pictures at circa 0.64U including tracks for up to 20 steps for the three types of experiments discussed in the
text. c) Schematic representation of dyes in their surrounding. Dyes are shown as circles for which the color indicates their current velocity (white:
too fast for detection by WFM, yellow: slow enough for WFM, red: very slow/immobilized).
gelation, more than 90% of the molecules move too fast to be
localized owing to our WFM time resolution. However, some
dye molecules move very slowly or are even immobilized. We
assume that these molecules are situated in regions in which a
polymer network has already formed and therefore are
hindered in their movement. With gelation the number of
slow and immobile molecules increases significantly. The
heterogeneity arising from the concomitant presence of fast,
slow and immobilized dyes is not only obvious in WFM but
also causes the FCS autocorrelation fits with one diffusion
constant to become unacceptable. In the course of the further
polymerization, the diffusion of all molecules becomes slower
and the concentration of immobilized molecules gradually
increases, as seen in the WFM movies. Finally, translational
motion stops entirely. This cessation happens earlier in time
for higher cross-linker concentrations.
Dye 3 with two styrenyl groups can be incorporated with
one or two ends into the growing polymer chain. Therefore,
the translational motion of free and incorporated dye can be
studied simultaneously. Dye 3 was added in nanomolar
concentrations to a solution of styrene and initiator. At this
low concentration an influence on viscosity or other proper-
ties of the reaction mixture can be neglected. The FCS
autocorrelation curves at all times had to be fitted with two
diffusion constants. At any time and with comparable
conditions, the fast component corresponds well with the
diffusion constant of free 2, a dye similar in size to 3
(Figure 1c), suggesting that this fraction represents the dye
which has not been incorporated into the polymer chains. The
slow component has a diffusion constant about one order of
magnitude smaller and can be associated with 3 incorporated
into polystyrene. The dye can be incorporated into chains of
different length, each of them having a characteristic D, but
only an average value could be determined by FCS. In
comparison to the experiment with the even larger dye 1, for a
similar dye concentration, the fraction of dyes detectable in
WFM is higher at any time. This can be explained by the
incorporation of 3 into the growing polymer chains which
slows down their motion already at an earlier stage. As in the
experiment with 1 or 2, motion (of both incorporated and free
dye 3) can no longer be detected at ca. 0.90U.
The WFM measurements were compared at ca. 0.64U. At
this conversion clear differences in lateral diffusion can be
observed (see Figure 2 and Movie 5 in the Supporting
Information). To quantify the WFM observations, the step-
length distributions[22,23] of the lateral motion of single dye
molecules between frames was measured and presented in a
histogram[24] (see Figure 3a, and Figure S6 in the Supporting
Information). In the polymerization of styrene without 5 and
probed with 1, the dye molecules move fast and therefore
tracking can be done only for a few molecules that occasion-
ally slow down. On average, however, even these “slow”
Angew. Chem. Int. Ed. 2008, 47, 783 –787
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
785

Communications
fast for a localization by WFM. However, the motion of dyes
incorporated into polymer chains is much slower and can be
tracked. D was found to be 0.261” 10 ꢀ12 m2 sꢀ1. Differences
between dyes incorporated into chains of different length or
even cross-linked dyes could not be evaluated quantitatively.
In conclusion, we present a novel way to follow radical
polymerization over an extensive conversion range. Our
method is based on the detection of single dye molecules and
the fact that their mobility changes during polymerization. In
absence of a cross-linker, 1 and 2 diffuse freely in the
surrounding medium, but in presence of 5 the influence of
heterogeneity that arises during the formation of a network
on the motion of reporter molecules could be verified.
Furthermore, using 3 in low concentration for example (ca.
10ꢀ9 m), the incorporation of the monomer units into the
growing polystyrene chains could be visualized.
Our investigations can be extended to other polymeri-
zation systems such as interpenetrating networks and nano-
composites during their formation process and in particular
will provide a deeper understanding of heterogeneities in
molecular motion. With this knowledge a better control over
polymerization and the properties of the resulting polymers
and polymer networks might be gained.
Received: September 11, 2007
Published online: December 11, 2007
Keywords: diffusion · fluorescence correlation spectroscopy ·
.
polymerization · styrene · wide-field microscopy
Figure 3. a) Step-length distributions for time lags of 38 and 114 ms
for the three types of experiments mentioned in the text; b) hr2i for
several single molecules of 1 at0.65 U in the experiment with 1% of 5.
[1] B. Yamada, D. G. Westmoreland, S. Kobatake, O. Konosu, Prog.
Polym. Sci. 1999, 24, 565 – 630.
moving molecules diffuse more freely with larger steps
compared to the other two experiments (D = 1.14 ”
10ꢀ12 m2 sꢀ1).
[2] S. Beuermann, M. Buback, Prog. Polym. Sci. 2002, 27, 191 – 254.
[3] J. Schweer, J. Sarnecki, F. Mayer-Posner, K. Müllen, H. J. Räder,
J. Spickermann, Macromolecules 1996, 29, 4536 – 4543.
[4] F. Heatley, P. A. Lovell, J. McDonald, Eur. Polym. J. 1993, 29,
255 – 268.
[5] P. Bosch, F. Catalina, T. Corrales, C. Peinado, Chem. Eur. J. 2005,
11, 4314 – 4325.
[6] M. S. Frahn, L. H. Luthjens, J. M. Warman, Polymer 2003, 44,
7933 – 7938.
[7] J. M. Warman, R. D. Abellon, L. H. Luthjens, J. W. A. Suyker-
buyk, H. J. Verhey, J. W. Verhoeven, Nucl. Instrum. Methods
Phys. Res. Sect. B 1999, 151, 361– 366.
[8] H. J. Verhey, B. Gebben, J. W. Hofstraat, J. W. Verhoeven, J.
Polym. Sci. Part A 1995, 33, 399 – 405.
The situation is different in case of the growing network
polymer. Significant heterogeneity of the motion of dyes was
observed. The step-length distributions reveal two fractions
(see Supporting Information, Figure S7). One fraction shows
free diffusion with D = 0.469 ” 10ꢀ12 m2 sꢀ1 and can be
assigned to molecules moving in areas where the polymer
network is less dense. For the other fraction the step length
does not significantly increase at longer time lags, and thus it
can be attributed to molecules immobilized in the network.
Further evidence that the observed heterogeneity is due to
different fractions of molecules with a different motion is
presented in Figure 3b. The mean square displacement hr2i is
plotted against time lag for several single molecules. The
immobile fraction of dyes shows values of hr2i which remain
close to zero for all time intervals. Another fraction diffuses
normally and can be related to the slow fraction detected in
FCS. However, in contrast to FCS, WFM reveals a distribu-
tion of hr2i for different molecules as can be seen from the
spread in Figure 3b. The fast fraction detected in FCS is too
quick for the WFM time resolution used.
[9] M. Levitus, J. L. Bourdelande, G. Marques, P. F. Aramendia, J.
Photochem. Photobiol. A 1999, 126, 77 – 82.
[10] R. O. Loutfy, Pure Appl. Chem. 1986, 58, 1239 – 1248.
[11] R. O. Loutfy, Macromolecules 1981, 14, 270 – 275.
[12] H. Uji-i, S. M. Melnikov, A. Deres, G. Bergamini, F. de Schryver,
A. Herrmann, K. Müllen, J. Enderlein, J. Hofkens, Polymer 2006,
47, 2511 – 2518.
[13] A. Schob, F. Cichos, J. Schuster, C. von Borczyskowski, Eur.
Polym. J. 2004, 40, 1019 – 1026.
[14] R. A. L. VallØe, P. Marsal, E. Braeken, S. Habuchi, F. C.
De Schryver, M. van der Auweraer, D. Beljonne, J. Hofkens, J.
Am. Chem. Soc. 2005, 127, 12011 – 12020.
In the experiment using 3, free dyes and incorporated dyes
are present simultaneously. The former are in most cases too
[15] X. Pan, C. Aw, Y. Du, H. Yu, T. Wohland, Biophys. Rev. Lett.
2006, 1, 1 – 9.
786
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 783 –787

Angewandte
Chemie
[16] G. T. Gavranovic, S. Csihony, N. B. Bowden, C. J. Hawker, R. M.
Waymouth, W. E. Moerner, G. G. Fuller, Macromolecules 2006,
39, 8121 – 8127.
[17] A. Michelman-Ribeiro, H. Boukari, R. Nossal, F. Horkay,
Macromolecules 2004, 37, 10212 – 10214.
[18] N. Bark, Z. Fꢀldes-Papp, R. Rigler, Biochem. Biophys. Res.
Commun. 1999, 260, 35 – 41.
[19] Dye 3 can act as a cross-linker. However, cross-linking was not
considered in the discussion because the concentration of these
molecules was ca. 9 orders of magnitudes smaller than the
styrene concentration.
[20] H. Yamazoe, P. B. Zetterlund, B. Yamada, D. J. T. Hill, P. J.
Pomery, Macromol. Chem. Phys. 2001, 202, 824 – 829.
[21] P. J. Flory, Principles of Polymer Chemistry, Cornell Univ. Press,
Ithaca, New York, 1953.
[22] M. C. Konopka, J. C. Weisshaar, J. Phys. Chem. A 2004, 108,
9814 – 9826.
[23] T. Kues, R. Peters, U. Kubitscheck, Biophys. J. 2001, 80, 2954 –
2967.
[24] R. Bausinger, K. von Gersdorff, K. Braeckmans, M. Ogris, E.
Wagner, C. Bräuchle, A. Zumbusch, Angew. Chem. 2006, 118,
1598 – 1602; Angew. Chem. Int. Ed. 2006, 45, 1568 – 1572.
Angew. Chem. Int. Ed. 2008, 47, 783 –787
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
787