Random catalyst walking along polymerized poly(3-hexylthiophene) chains in kumada catalyst-transfer polycondensation

Article abstract of DOI:10.1021/ja102210r

A walking process of Ni catalysts during Kumada catalyst-transfer polycondensation along polymerizing poly(3-hexylthiophene), P3HT, chains was investigated. To simplify polymer end group identifications, a compound Br-C₆H₄-Ni(dppe)-Br was prepared and used as an externally addable initiator. Normally, aryl moieties present in initiators incorporate into the structure of the resulting P3HT as the starting groups. We demonstrate that due to the presence of the C-Br group located in the para-position to the Ni substituent of the initiator, two different polymeric products are formed. One of them is the normal product, that is, P3HT with a para-bromophenyl end group, whereas another one has the phenyl ring inside the P3HT chain. The content of the product with the internal phenyl ring increases with the increase of the polymerization degree. Control experiments demonstrated that no intermolecular catalyst transfer takes place in the conditions used. Such results suggest that catalytic Ni(0) species are able to walk along the polymerizing chain containing many tens of thienyl rings up to the opposite end and can initiate polymerization there. Numerical analysis of a random hopping model was undertaken, which revealed that a combination of a random catalyst walking along the chain and a sticking effect at the end groups is operative in Kumada catalyst-transfer polycondensation.

Full text of DOI:10.1021/ja102210r

Published on Web 05/13/2010
Random Catalyst Walking along Polymerized
Poly(3-hexylthiophene) Chains in Kumada Catalyst-Transfer
Polycondensation
Roman Tkachov, Volodymyr Senkovskyy,* Hartmut Komber, Jens-Uwe Sommer,*
and Anton Kiriy*
Leibniz-Institut fu¨r Polymerforschung Dresden e.V., Hohe Strasse 6, 01069 Dresden, Germany
Received March 29, 2010; E-mail: senkovskyy@ipfdd.de; sommer@ipfdd.de; kiriy@ipfdd.de
Abstract: A “walking” process of Ni catalysts during Kumada catalyst-transfer polycondensation along
polymerizing poly(3-hexylthiophene), P3HT, chains was investigated. To simplify polymer end group
identifications, a compound Br-C6H4-Ni(dppe)-Br was prepared and used as an externally addable initiator.
Normally, aryl moieties present in initiators incorporate into the structure of the resulting P3HT as the starting
groups. We demonstrate that due to the presence of the C-Br group located in the para-position to the Ni
substituent of the initiator, two different polymeric products are formed. One of them is the “normal” product,
that is, P3HT with a para-bromophenyl end group, whereas another one has the phenyl ring inside the
P3HT chain. The content of the product with the internal phenyl ring increases with the increase of the
polymerization degree. Control experiments demonstrated that no intermolecular catalyst transfer takes
place in the conditions used. Such results suggest that catalytic Ni(0) species are able to walk along the
polymerizing chain containing many tens of thienyl rings up to the opposite end and can initiate
polymerization there. Numerical analysis of a random hopping model was undertaken, which revealed that
a combination of a random catalyst walking along the chain and a “sticking effect” at the end groups is
operative in Kumada catalyst-transfer polycondensation.
Introduction
chain-growth catalyst-transfer polymerization of 2-bromo-5-
chloromagnesio-3-hexylthiophene (1) into regioregular poly(3-
Ni-catalyzed chain-growth Kumada polycondensation1-18
nowadays emerges into a powerful synthetic route to well-
defined conjugated polymers,1,2 block copolymers,9-13 and other
polymer architectures.3-8 Interest in Ni-catalyzed Kumada
polycondensation was greatly renewed after the discovery of a
alkylthiophene)s (P3ATs).1,2 This process is of great practical
importance because it paves the way to earlier inaccessible
materials useful for various optoelectronic applications.15 This
reaction is also interesting from a theoretical point of view16
since it exemplifies a new way for the transformation of
polycondensations that normally have step-growth character into
more attractive chain-growth polymerizations.17-19
(1) Sheina, E. E.; Liu, J.; Iovu, M. C.; Laird, D. W.; McCullough, R. D.
Macromolecules 2004, 37, 3526. Iovu, M. C.; Sheina, E. E.; Gil, R. R.;
McCullough, R. D. Macromolecules 2005, 38, 8649–8656. Osaka, I.;
McCullough, R. D. Acc. Chem. Res. 2008, 41, 1202–1214.
(2) Yokoyama, A.; Miyakoshi, R.; Yokozawa, T. Macromolecules 2004,
37, 1169. Miyakoshi, R.; Yokoyama, A.; Yokozawa, T. J. Am. Chem.
Soc. 2005, 127, 17542–17547.
It is now generally accepted that under “classical” Mc-
Cullough/Yokozawa conditions, Grignard methathesis (GRIM)
polymerization or Kumada catalyst-transfer polycondensation
(12) Miyakoshi, R.; Yokoyama, A.; Yokozawa, T. Chem. Lett. 2008, 37,
1022–1023. Yokoyama, A.; Kato, A.; Miyakoshi, R.; Yokozawa, T.
Macromolecules 2008, 41, 7271–7273.
(3) Senkovskyy, V.; Khanduyeva, N.; Komber, H.; Oertel, U.; Stamm,
M.; Kuckling, D.; Kiriy, A. J. Am. Chem. Soc. 2007, 129, 6626–6632.
(4) Khanduyeva, N.; Senkovskyy, V.; Beryozkina, T.; Bocharova, V.;
Simon, F.; Nitschke, M.; Stamm, M.; Gro¨tzschel, R.; Kiriy, A.
Macromolecules 2008, 41, 7383–7389.
(13) Dai, C.-A.; Yen, W.-C.; Lee, Y.-H.; Ho, C.-C.; Su, W.-F. J. Am. Chem.
Soc. 2007, 129, 11036. Sommer, M.; Lang, A. S.; Thelakkat, M.
Angew. Chem., Int. Ed. 2008, 47, 7901–7904. Zhang, Q.; Cirpan, A.;
Russell, T. P.; Emrick, T. Macromolecules 2009, 42, 1079–1082.
Rajaram, S.; Armstrong, P. B.; Kim, B. J.; Frechet, J. M. J. Chem.
Mater. 2009, 21, 1775–1777.
(5) Beryozkina, T.; Senkovskyy, V.; Kaul, E.; Kiriy, A. Macromolecules
2008, 41, 7817–7823.
(6) Khanduyeva, N.; Senkovskyy, V.; Beryozkina, T.; Horecha, M.;
Stamm, M.; Uhrich, C.; Riede, M.; Leo, K.; Kiriy, A. J. Am. Chem.
Soc. 2009, 131, 153–161.
(14) Wu, S.; Bua, L.; Huang, L.; Xinhong, Y.; Han, Y.; Geng, Y.; Wang,
F. Polymer 2009, 50, 6245–6251.
(7) Senkovskyy, V.; Tkachov, R.; Beryozkina, T.; Komber, H.; Oertel,
U.; Horecha, M.; Bocharova, V.; Stamm, M.; Gevorgyan, S. A.; Krebs,
F. C.; Kiriy, A. J. Am. Chem. Soc. 2009, 131, 16445–16453.
(8) Tkachov, R.; Senkovskyy, V.; Horecha, M.; Oertel, U.; Stamm, M.;
Kiriy, A. Chem. Commun. 2010, 46, 1425–1427.
(15) Snaith, H. J.; Whiting, G. L.; Sun, B.; Greenham, N. C.; Huck,
W. T. S.; Friend, R. H. Nano Lett. 2005, 5, 1653.
(16) Lanni, E. L.; McNeil, A. J. J. Am. Chem. Soc. 2009, 131, 16573–
16579.
(17) Yokoyama, A.; Yokozawa, T. Macromolecules 2007, 40, 4093–4101.
Yokoyama, A.; Yokozawa, T. Chem. ReV. 2009, 109, 5595–5619.
(18) Yokoyama, A.; Suzuki, H.; Kubota, Y.; Ohuchi, K.; Higashimura, H.;
Yokozawa, T. J. Am. Chem. Soc. 2007, 129, 7236–7237.
(19) Beryozkina, T.; Boyko, K.; Khanduyeva, N.; Senkovskyy, V.; Horecha,
M.; Oertel, U.; Simon, F.; Komber, H.; Stamm, M.; Kiriy, A. Angew.
Chem., Int. Ed. 2009, 48, 2695–2698.
(9) Kaul, E.; Senkovskyy, V.; Tkachov, R.; Bocharova, V.; Komber, H.;
Stamm, M.; Kiriy, A. Macromolecules 2010, 43, 77–81.
(10) Miyakoshi, R.; Shimono, K.; Yokoyama, A.; Yokozawa, T. J. Am.
Chem. Soc. 2006, 128, 16012–16013.
(11) Adachi, I.; Miyakoshi, R.; Yokoyama, A.; Yokozawa, T. Macromol-
ecules 2006, 39, 7793–7795.
9
10.1021/ja102210r  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 7803–7810 7803

A R T I C L E S
Tkachov et al.
Scheme 1. State-of-the-Art Vision of the Mechanism of
favored for lengthy monomers since in this case the distance
for the RW is enlarged, which, in principle, would favor other
reaction pathways.
Chain-Growth KCTP16
Recently, we demonstrated a “robustness” of the chain-growth
mechanism in examples of KCTP of thiophene-based oligomers
comprising two and three thiophene rings.5 In all cases, the
chain-growth mechanism was a dominating polymerization
pathway with a selectivity of the intramolecular catalyst-transfer
elementary step (versus the intermolecular diffusion step) of
∼99% for polymerization of the dithiophene and of ∼97% for
polymerization of terthiophenes. This means that Ni species are
able to undergo selective RW and intramolecular transfer over
distances, larger than 1 nm. The observed selectivity of the
intramolecular transfer over the intermolecular diffusion pathway
is likely due to a strong propensity of the Ni(0) species to form
intramolecular complexes with conjugated systems. The latter
is caused by unsaturated coordination of the Ni(0) center formed
at the RE step in the absence of appropriate stabilizing ligands.25
However, important details of the RW remain unknown. It
is, for example, not clear whether (1) the movement of Ni is
unidirectional toward the nearest C-Br bond (“unidirectional
RW”) or, instead, (2) it freely “walks” back and forth along
the growing chain via a set of intermediate η2 (η4 or η6)
complexes until it finds a C-Br bond to undergo the OA
(“random RW”).25 If the variant 1 is realized and if, to provide
the chain-growth propagation, the Ni catalyst should necessarily
always move in a correct direction toward the nearest C-Br
bond, then the extension of the KCTP scope onto lengthy or
strongly polarized monomers may be problematic. In principle,
the directionality of the catalyst walking toward the terminal
thiophene ring would be provided by the orientation effect of
the electron-deficient C-Br group; however, it should decay
quickly with the increase of the monomer unit length. On the
other hand, a complex substitution of monomer units with
electron-donating and -withdrawing groups would also mask
the orientation effect of the terminal C-Br bond. From these
regards, the random “RW” (the variant 2) seems to be more
promising mechanism. Indeed, if during the polymerization of
1 into P3HT Ni catalysts walk randomly along the polymerizing
chains and, nevertheless, they do not dissociate from the chain
and provide a near living chain propagation (that is defacto
observed1-18), then the binding propensity of Ni(0) species
toward π-conjugated systems is indeed huge. If so, the chain-
growth polymerization of lengthy and complex monomers
should be possible since anyway upon polymerization of even
short monomer 1 the Ni catalysts overcome distances much
larger than the monomer length.
A detailed revealing of the structure of polymeric products
of GRIM polymerization of 1 catalyzed by NidpppCl2 would,
in principle, shed light on the RW process. Preparation of the
initiator in GRIM polymerization involves two consecutive TM
steps between NidpppCl2 and two monomer molecules leading
to symmetrical adduct 2 (Scheme 2). In the next step, the Ni
catalyst undergoes RE and then oxidatively adds either to the
left-hand or to the right-hand thienyl rings with an equal
probability leading to 3. If the polymerization of 1 initiated by
3 involves the unidirectional RW, then the chain should extend
only in a one direction predetermined by the initial catalyst
choice (made at the very first OA step). The polymerization in
this case should lead to P3HT in which the tail-to-tail (TT) defect
is the starting group. If, instead, GRIM polymerization involves
the random RW, then both ends of the initiator 3 should be
reactive and the chain extends in both directions, leading to
(KCTP) can be initiated by catalytic species formed from 1
equiv of NidppeCl2 [dppe ) ethane-1,3-diylbis(diphenylphos-
phane)], NidpppCl2 [dppp ) propane-1,3-diylbis(diphenylphos-
phane)], or similar catalyst precursors and 2 equiv of Grignard
type monomer molecules. Alternatively, KCTP can be initiated
by externally added Ar-Ni(L2)-X initiators, synthesized in an
independent step.3-9,20-22 Preparation of the externally addable
initiators involves a two-step reaction of aryl halides (Ar-X)
with diethylbipyridylnickel [Et2Ni(bipy)] followed by addition
of dppp or dppe to replace bipy.7 An advantage of this approach
is that it allows initiation of the polymerization from predesigned
functions positioned at desired locations, such as from surfaces
leading to P3AT brushes7,8 or from bifunctional initiators
allowing preparation of rod-coil block copolymers.9
A chain-propagation catalytic cycle of KCTP (or GRIM
polymerization) involves transmetalation (TM), reductive elimi-
nation (RE), and oxidative addition (OA) elementary steps
(Scheme 1).1,2,16 The chain-growth mechanism is provided by
the fact that within the chain-propagation cycles, the Ni(0)
catalytic species eliminated upon the RE step do not dissociate
from the chain and do not react with another monomer
molecules or other halides but, instead, undergo the intramo-
lecular OA into a C-Br bond present in the same chain. It is
noteworthy that Ni(0) species do not exchange their host chains
and that the intramolecular transfer proceeds at the level of each
individual growing chain,23 as it was proved previously.3 If so,
during polymerization, Ni(0) species should “walk”24,25 over a
distance at least equal to the length of one monomer unit. The
ring-walking (RW) process is one of the most intriguing and
not yet understood aspects of the KCTP mechanism. Under-
standing of this process is desirable for successful involvement
in the chain-growth polymerization of other monomer units,
including those that comprise several aromatic rings. It could
be a priori expected that the chain-growth mechanism is less
(20) Doubina, N.; Ho, A.; Jen, A. K.-Y.; Luscombe, C. K. Macromolecules
2009, 42, 7670–7677.
(21) Bronstein, H. A.; Luscombe, C. K. J. Am. Chem. Soc. 2009, 131,
12894–12895.
(22) Smeets, A.; Bergh, K.; Winter, J.; Gerbaux, P.; Verbiest, T.; Koeck-
elberghs, G. Macromolecules 2009, 42, 7638–7641.
(23) Lamps, J.-P.; Catala, J.-M. Macromolecules 2009, 42, 7282–7284.
(24) Strawser, D.; Karton, A.; Zenkina, O. V.; Iron, M. A.; Shimon,
L. J. W.; Martin, J. M. L.; van der Boom, M. E. J. Am. Chem. Soc.
2005, 127, 9322–9322.
(25) Zenkina, O. V.; Karton, A.; Freeman, D.; Shimon, L. J. W.; Martin,
J. M. L.; van der Boom, M. E. Inorg. Chem. 2008, 47, 5114–5121.
9
7804 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

Catalyst Walking along Polymerized P3HT Chains
A R T I C L E S
Scheme 2. Possible Mechanisms of Unidirectional and Random
Catalyst Walking Taking Place upon GRIM Polymerization
Resulting in Different Positions of the TT Defect1,2
tion). In this work, the dppe-based Ni initiator was used because
of its higher solubility (than solubility of dppp-based initiators),
which simplified handling of the complex and NMR charac-
terizations. We demonstrate here that the use of dppe-supported
initiators does not necessarily cause worsening of the polym-
erization control if they are introduced in a dissolved state.16
NidppeCl2 and NidpppCl2 are poorly soluble in THF; therefore,
their addition into the reaction mixture in the solid state does
not provide the chain initiation simultaneously by all Ni species
added. Even though NidppeCl2 has a better solubility than
NidpppCl2, this problem is more pronounced for polymerizations
induced by the former initiator since the polymerization rate is
much higher with NidppeCl2.26 However, as the initiator 4 is
soluble in THF, efficient initiation is provided, which leads to
narrowly distributed polythiophenes with a near complete
incorporation of para-bromophenyl group into the P3HT
structure (see MALDI-TOF data, Figure S4 of the Supporting
Information).
Scheme 4 depicts possible transformations that may occur
upon mixing of 1 and 4. In the first step, 1 reacts with 4 (TM
step) giving nonsymmetrical adduct 6 that undergoes RE,
leading to 7 in which the eliminated Ni(0) species is associated
with either benzene or thienyl rings. Afterward, 7 may undergo
OA, reacting with the C-Br bonds of either phenyl or thienyl
rings, forming 8 or 9, respectively, in the ratio corresponding
to the relative reactivity of the C-Br bonds. Let us further
consider two chain-propagation scenarios, one of them corre-
sponding to the unidirectional catalyst walking and another one
corresponding to the random RW mechanism. Upon the
unidirectional RW, the chain should be polymerized either from
the thienyl or from benzene rings, resulting in P11 or P12, which
have 2-bromo-3-hexyl-thien-5-yl or para-bromophenyl starting
groups, respectively. The opposite terminal groups for both P11
and P12 should be H atoms attached to the 2-position of the
thiophene ring due to termination by HCl. An alternative random
catalyst walking assumes the extension of the initiator 4 in both
directions via intermediates 6-10 and 13-15 with the assistance
of the same Ni catalytic species that freely move (walk) along
the polymerized chain from one end to the opposite leading to
P16 in which the 1,4-connected phenyl ring is located inside
the chain. One can also assume that these two RW pathways
(unidirectional and random) may both contribute to a real
polymerization process in a certain degree.
It is noteworthy that the position of the phenyl ring in the
structure of the polymerization product would be a clear
indication of a certain RW mechanism only if the polymerization
cannot be restarted intermolecularly from the para-bromophenyl
group of the polymer P12 to form P16. We demonstrated
previously that no intermolecular catalyst transfer occurs for
the polymerizations catalyzed by Ni supported by monodentate
phosphorus ligands.5 In the present work, we performed similar
control experiments using NidpppCl2 and NidppeCl2 catalysts.
In particular, polymerizations of 1 initiated by NidpppCl2 and
NidppeCl2 were conducted in the presence of a large excess of
either 1,4-dibromobenzene 5, bromobenzene, or 2,5-dibro-
mothiophene. We found that these additives did not affect the
normal polymerization course and that P3HT with low poly-
dispersities in a 1.1-1.2 range were obtained. No polymerization
products with phenyl (or unsubstituted thienyl) end functions
Scheme 3. Synthesis of the Initiator
P3HT with the TT defect located inside the chain. Hence, one
would determine the character of the RW, if the position of the
initiator residue in the structure of the final polymeric product
would be detectable. Unfortunately, identification of the exact
position of the above-mentioned TT defect in the P3HT structure
is highly challenging.
To circumvent the identification problem, in the present work,
we prepared Br-C6H4-Ni(dppe)-Br (4), which has an easily
1
recognizable (by H NMR) para-substituted phenyl ring, and
used it as initiator for the polymerization of 1. The structure of
the products formed and the course of polymerization were
investigated experimentally and theoretically, which allowed us
to deduce that during the KCTP, catalytic Ni(0) species are able
to walk along the polymerizing chain containing many tens of
thienyl rings up to the opposite end and can initiate polymer-
ization there.
Results and Discussion
The initiator 4 was cleanly obtained using a recently
developed procedure in our lab7-9 by reacting 5 with Et2Nibipy
followed by ligand exchange with an excess of dppe (Scheme
3). Only one bromine atom in 5 was reactive toward Et2Nibipy,
(26) Alternatively, narrowing of polydispersities can be achieved via a
preinitiation approach, as described in ref 16, according to which some
portion of the monomer is mixed with NidppeCl2 to give soluble
“living” short oligomers that are used as initiators.
1
which was confirmed by H and 31P NMR spectroscopy (for
experimental details and NMR spectra, see Supporting Informa-
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7805

A R T I C L E S
Tkachov et al.
Scheme 4. Unidirectional versus Random Catalyst Walking upon KCTP
were formed in these experiments; hence, the intermolecular
catalyst-transfer pathway can be ruled out in the particular
conditions. On the other hand, experiments discussed below
strongly suggest that the polymerization involves the intramo-
lecular catalyst-transfer mechanism, in full agreement with
previous observations.1-22
calculated from the intensities of signals (H2 + H3) and H2′,
and the intensity of the R-CH2 group signal of the hexyl residue
was related to the intensity of two phenyl protons [(H2 + H3 +
H2′)/2] to calculate an averaged degree of polymerization (DP).
Integration of the spectra shows a near quantitative incorporation
of the phenyl moiety into the structure of the resulting polymer.
In the experiments conducted, the P12 and P16 products were
formed in substantial quantities but at a variable ratio, whereas
P11 was not found among the polymeric reaction products.27
For the thiophene proton in the single bromo-3HT unit in P11,
a chemical shift of 7.03 ppm was calculated based on the
chemical shift of HBr, taking into account that replacement of a
P3HT chain by a P3HT-C6H4- moiety results in a low-field
shift of 0.20 ppm for that position (compare H2′ and HP3HT).
However, no signal that could prove this group is observed.
Only in very low-molecular products (DP ∼ 3) was such a
structure found (see Table S1 and Figure S1 of the Supporting
Information).
As a typical example, we further discuss results of the
experiment conducted at RT and a 1/150 feed ratio. Three
samples were withdrawn in this polymerization run, at 90 s,
210 s, and 30 min, and P3HT samples with DPs of 40, 70, and
168, respectively, were obtained, as determined by NMR (these
DPs agree rather well with DPs determined from GPC data,
which gave DPs of 46, 90, and 186, respectively; Figure S5 of
the Supporting Information). The reaction shows a clean chain-
growth performance, allowing preparation of P3HT with number
average molecular weight (Mn) up to 30000 g/mol with a
satisfactory integration of starting and end groups. This result
In the next step, polymerizations of 1 initiated by 4 were
conducted at room temperature (RT) at different monomer-to-
initiator (feed) ratios of 3, 7, 20, 40, and 150. To follow the
polymerization course, several samples at different polymeri-
zation times (corresponding to different conversions of 1) were
withdrawn from the polymerization mixture. P3HT with a
polydispersity index in a 1.1-1.2 range (determined for
nonfractionated crude products washed with water and methanol
only) and controllable (by the feed ratio and monomer conver-
sion) molecular weight (MW) were obtained, reflecting a near
“living” polymerization performance. Somewhat increased poly-
dispersities (∼1.4) were observed when the MW of the resulting
polymer exceeded 30000 g/mol value, presumably because of
a limited solubility of a high MW P3HT-promoted chain
termination and reinitiation (Figure S5 of the Supporting
Information). A typical 1H NMR spectrum of the reaction
mixture quenched with HCl and washed with methanol is shown
in Figure 1. Whereas the characteristic signals of terminal
thiophene units (HBr and HH) were well-known from the
literature,1-7 the additional AB spin system (signals H2 and H3)
is due to the para-bromo phenyl starting group in P12, and the
singlet (H2′) represents the phenyl units on both sides substituted
by P3HT in P16. The rotating frame Overhauser effect spec-
troscopy (ROESY) spectrum allows us to assign the neighboring
proton of the P3HT chain (H6 and H6′, respectively) and thus
to prove the terminal and internal positions, respectively, of these
phenyl moieties. The fraction of both substitution pattern was
(27) Trace amounts of P11 were found only in a fraction corresponding to
DP ∼ 3, demonstrating a negligible probability of the formation of
P11.
9
7806 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

Catalyst Walking along Polymerized P3HT Chains
A R T I C L E S
Figure 2. Dependence of the P12 fraction on the polymerization degree
for the polymerization of 1 initiated by 4 at RT.
polymerization course to variations in the composition of the
reaction mixture (and hence, on viscosity, molecular diffusion,
etc.) corroborates with the assumption that intermolecular
migrations of Ni catalysts are not responsible for the formation
of P16.
In the next experiment, a “livingness” of the polymerization
was tested. In this experiment, 3 equiv of the monomer 1 was
polymerized in the presence of 4 for a few minutes, and
afterward, the first sample was withdrawn; the rest of the
reaction mixture was kept overnight under argon at RT, and
then, the second sample was withdrawn; finally, 20 equiv of 1
was added to the reaction mixture, which was allowed to
polymerize for a further 10 min, giving the third sample. The
1
first and the second samples had the same DP (∼3, H NMR
data) and the same composition (P16/P12 ratio of 12/88).
Importantly, the addition of a new portion of 1 resulted in a
clean extension of the 3-mer into the 28-mer with a negligible
amount of reinitiated H/Br-terminated P3HT. Furthermore, the
ratio between P12 and P16 in this product was exactly the same
as in other experiments, in which polymerization was conducted
in a single run without a storage of the “living” prepolymer.
This result further demonstrates the livingness of the polym-
erization initiated by externally added initiator 4.
Figure 1. (a) 1H NMR spectrum (region) of a P3HT obtained upon Kumada
catalyst-transfer polymerization of 1 initiated by 4 at RT and (b) the
corresponding region from the ROESY spectrum. Br and H, thiophene
signals from Br and H chain termination; *, 13C satellites of the P3HT
backbone signal.
further proves that each Ni species polymerizes only a single
polymer chain, and no Ni exchange between the growing chains
takes place in these conditions.28
It is important for interpretation of the results to understand
that the formation of para-bromophenyl terminated product P12
requires that all chain-propagation steps occur in the same
direction (corresponding to the right-hand side direction in
Scheme 4) from intermediate 6 and via 9 without activation of
the para-bromophenyl group. On the other hand, even a single
activation of the para-bromophenyl group followed by at least
one chain-propagation step from this end will result in P16. In
other words, P16 should be formed with a much higher
probability if the reactivity of the two C-Br bonds located in
phenyl and thienyl rings would be equal. However, P12 is a
major product in most polymerization runs (with DP up to 40),
suggesting that the activation of the para-bromophenyl group
is a much less efficient process than the activation of the C-Br
bond of the thienyl ring. To verify this hypothesis, a competition
Kumada coupling experiment was conducted in which 1.1 equiv
of para-phenyl-magnesium chloride (17) was reacted with 1
equiv of phenylbromide (18) and 1 equiv of 2-bromo-3-
hexylthiophene (19) in the presence of 0.05 equiv of NidppeCl2
for 10 min (Scheme 5). Analysis of the reaction mixture by
gas chromatography with a mass spectrometer detector revealed
a largely preferential consumption of 19 (the ratio between
unreacted 18 and 19 of ∼1/6 was detected). The obtained results
clearly show a higher reactivity of the thiophenic bromide than
The fraction of para-bromophenyl terminated product P12
decays, whereas the fraction of P16 with internal phenyl groups
grows with the increase of an overall DP of P3HT (Figure 2).
For instance, only ∼12% of P16 was found in products with
DP ∼ 3, whereas P16 and P12 products are present in near
equal quantities at DP ∼ 40. It is noteworthy that at a fixed
polymerization temperature, relative yields of P12 and P16
products depend only on DP and are independent of the
concentration of monomer 1 and the feed ratio. For example,
P3HT with a DP ∼ 40 was obtained as the final product in the
polymerization with the set feed ratio of 40, P3HT with
the same DP was isolated at incomplete conversion of 1 in the
experiment with the “ordered” DP of 150, and the same content
of P16 of ∼50% was obtained in both cases. In another pair of
experiments, the same content of P16 was observed in products
when polymerizations were conducted at two different monomer
concentrations of 0.067 and 0.016 M. Such insensitivity of the
(28) This is in a full agreement with previously reported Yokozawa et al.
and McCullough et al. data (refs 1, 2, and 23) as well as with our
previous papers (refs 3-9) but contradicts with the paper of Rawlins
et al., who did not succeed in obtaining P3HT with Mn higher than
8000 g/mol and to control Mn by the feed ratio but, instead, observed
a dependence of Mn on the monomer concentration. Achord, B. C.;
Rawlins, J. W. Macromolecules 2009, 42, 8634–8639.
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7807

A R T I C L E S
Tkachov et al.
Scheme 5. Competitive Kumada Cross-Coupling Experiment
Demonstrating a Higher Reactivity of Thienyl Bromide 19 than of
Phenylbromide 18
Scheme 6. Schematic Picture of the Various Steps Involved in the
Chain-Hopping and Reaction Process
(Scheme 2), then both ends should propagate equally, leading
to P3HT with the TT defect located inside the P3HT chain but
not being only the terminal group, as represented in some
articles.1
of the bromobenzene in Kumada cross-coupling reaction that
qualitatively corroborates with the postulated higher reactivity
of the thienyl than the para-bromobenzene ends in KCTP. The
higher reactivity of the thiophene derivatives in Kumada
coupling could be either due to a higher reactivity at the OA
step of the C-Br bond as a result of activation of this bond by
electron-withdrawing sulfur atom, due to a higher complexation
ability of the thienyl ring than of the phenyl one with Ni
species,29 or due to a combination of these factors.
Numerical Analysis. The obtained data reveal that the Ni
catalyst walks randomly during KCTP; however, these data do
not exclude a possible contribution from a unidirectional catalyst
walking. To get deeper inside the walking process, the experi-
mental data were used to test the random walker model (random
hopping of the catalyst) quantitatively. The model is defined as
follows: At the beginning of the process, the Ni catalyst is
initially placed between the two different end groups, which
we call A and B for simplicity (which correspond to the
bromothienyl and para-bromophenyl ends, respectively; Scheme
6).
Within a given unit of time, it takes a random step with equal
probability to the neighboring monomers. Whenever the catalyst
reaches the end groups A or B, we assume that a monomer is
added with probabilities pA or pB, corresponding to relative
reactivities of the two different ends (equates to the OA step),
which, after following TM and RE steps, results in prolongation
of the chain (N f N + 1). Such different reactivities corroborate
with the experimental finding that 8 and 9 are not formed with
the same probability from 7. In the following time step, the
catalyst is set to the position next to the reactive side (equates
to TM and RE steps). As discussed above, the new chain end
is always of type A. In the course of time, the catalyst walks
unbiased and randomly along the chain. When a given number
of monomers are consumed, a given DP N is realized. The
fraction of chains of type AB and length N shall be denoted by
pAB(N).
In a view of this, the obtained results can be best explained
in terms of walking of the Ni catalyst along the polymerizing
chain with a preferential OA to C-Br of the thienyl end. In the
first OA step, the intermediate 7 is mainly transformed into 9,
whereas the amount of 8 (product of OA into the C-Br of the
phenyl ring) is low after the first monomer addition step.
Extrapolation of the curve in Figure 2 to DP ) 1 shows that
only about 4-5% of P16 is formed, corresponding to only
4-5% of Ni species undergoing OA into C-Br of the phenyl
ring at the first polymerization step. The major pathway implies
the consecutive addition of monomer 1 to the intermediate 9,
resulting in 10, “a living” precursor of P12 (Scheme 4).
However, the experimental observation is that the product P16
with an internal phenyl ring progressively accumulates with the
DP rise and a substantial increase of P16 fraction is still
observed at relatively high DPs. Such results together with the
fact that P16 is not forming via intermolecular reinitiation from
the para-bromophenyl end group (see control experiments)
unambiguously suggest that at least some number of catalytic
Ni(0) species walks along the polymerizing chain up to the
opposite end over distances of several tens of nanometers. At
each polymerization step, some (relatively minor) portion of
such Ni(0) species that reached the opposite end oxidatively
adds to the para-bromophenyl end group, forming intermediates
13 and 14, which initiate polymerization into P16 in which the
phenyl ring is located inside the chain. At the same time,
the extension of the π-conjugation system gradually decreases
the probability for Ni species to reach the opposite terminus
that reduces the rate of OA into the para-bromophenyl end
group that is qualitatively consistent with the proposed random
walk polymerization model.
The above-described process can be easily simulated. A
comparison of the simulation with the experimental results
(Figure 2) for the fraction of AB type chains is given in Figure
3. To obtain a best fit to the experimental results for small
oligomers (N ) 2, 3), a ratio of pB/pA ) 0.07 has to be chosen
(for simplicity, we assume pA ) 1). The result for this model is
displayed as the dashed blue curve in Figure 3. This curve
decays too fast, actually with an asymptotic power law of 1/N.
The asymptotic power law decay of pAB(N) ∼ N-1 does not
depend on the values of pA and pB and is in marked contrast to
the experimental data. The best fit to the data predicts an power
If, however, both ends of the initiator are identical, such as
in the classical GRIM polymerization initiated by NidppeCl2
law decay according to N-0.23
.
(29) Arene π-complexation to Ni(0) is the first irreversible step in cross-
coupling reactions of some aromatic halides with Grignards. Yoshikai,
N.; Matsuda, H.; Nakamura, E. J. Am. Chem. Soc. 2008, 130, 15258–
15259.
Now, let us reconsider the case that the Ni(0) catalyst is sitting
next to the reactive end group A that corresponds to the
intermediate formed after each TM followed by RE steps.
9
7808 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010

Catalyst Walking along Polymerized P3HT Chains
A R T I C L E S
Scheme 7. Representation of KCTP-Based Block Copolymerization
Performed via Sequential Addition of Two Different Monomers: It
Is Favorable if Ni(L2) Species Form a Stronger Complex with the
First Added Residue of the Second Monomer Rather than with
Already Polymerized Polymer 1
of the catalyst to the end group is necessary to explain the decay
of the fraction pAB(N). In other words, the obtained experimental
data and calculations suggest that both random and unidirec-
tional catalyst walking mechanisms are operative in KCTP in
such a way that the catalyst walks randomly when it is far away
from the end groups, but it experiences a “sticking” effect when
it approaches the end groups (Figure S6 of the Supporting
Information).31
Revealing the catalyst walking mechanism may shed light
onto some not yet fully understood experimental observations.
For example, detailed kinetic studies recently performed by
Lamps and Catala23 revealed that GRIM polymerization has two
kinetic rate constants, one of them related to early, whereas the
second one to later, stages of the GRIM polymerization, and
the former is four times higher than the latter. This phenomenon
may be explained from the positions of the catalyst walking
mechanism. It is expected that chain elongation during the
polymerization derives an “unproductive” random RW over
large distances that should inevitably decrease the probability
for the Ni catalyst to be present near the propagation-chain end,
thus decreasing the rate constant. The same consideration may
also explain difficulties that one meets aiming at the preparation
of very high MW polymers using GRIM or KCTP.
Figure 3. Number fraction of AB type chains (pAB) as a function of the
chain length N. The experimental data are compared with the result of a
random walk simulation using various parameters as discussed in the text.
The inset shows the data in linear coordinates.
Because the chemical environment left and right to the catalyst
is not equivalent, it is not evident that the hopping rate is
symmetrical in this position. Indeed, it can be assumed that an
electronegative bromine atom present at the end group induces
polarization of the aromatic system so that complexation of the
Ni(0) species with a more electronegative bromothienyl end
group becomes more favored than complexation with any other
(internal) thiophene rings [we suppose that Ni(0) has a higher
affinity to more electronegative groups since in the next step it
undergoes an energetically favored oxidation by the C-Br
bond]. In our model, a preferential hopping of Ni(0) to the
bromothienyl end group is described by a “stickiness” δ of the
catalysts to A, which gives rise to a probability to jump on A
of 1/2 + δ/2 and a probability to jump in the opposite direction
(i.e., into the nonterminal thiophene ring when DP > 2) of 1/2
- δ/2; see Scheme 6.30 We hasten to note that the stickiness
does only apply to the site next to A and not to any other site
in the chain; thus, the walk is still globally unbiased and random.
Using a finite value of δ implies a higher value of pB/pA, since
the stickiness enlarges the fraction of AB type chains at a given
ratio of pB/pA. Interestingly, the value of δ is directly related
with the asymptotic decay of pAB(N) (Supporting Information),
which gives access to δ without considering other parameters
of the model. The best fit to experimental data was obtained
for δ ) 0.6 and pB/pA ∼ 0.3 as displayed by the solid red curve
in Figure 3. The fit using pB/pA ) 1/6 (0.166), that is, the value
obtained from the competition Kumada cross-coupling, also gave
quite satisfactory agreement with the experimental data (green
curve in Figure 3).
On the other hand, a recent observation of Yokozawa et al.12
related to the synthesis of all-conjugated block copolymers about
the importance of the monomers addition order could also be
explained in terms of the random catalyst walking mechanism.
In particular, it was reported that polythiophene-polyphenylene
block copolymers formed smoothly only when the polyphe-
nylene block was polymerized first, whereas polymerization
failed to give a good quality block copolymer at the reverse
monomer addition order.12,14 In block copolymerization, prepa-
ration of the second block is initiated by poly1-Ni(L2)-Br,
which upon TM with a monomer 2 should give poly1-
Ni(L2)-mono2 (Scheme 7). A subsequent RE leads to liberation
of Ni(0) and its complexation with the polymerized conjugated
system. We propose that polymerization of the second block
proceeds smoothly if the monomer 2 and the polymer chain
derived from it form stronger complexes with Ni(0) than
monomer units in already prepolymerized poly1. In this case,
Ni(0) preferentially complexes with the terminal monomer unit
so that its unproductive RW is prohibited. This situation
corresponds to a relatively high “stickiness” (δ) value and,
consequently, to a larger contribution of unidirectional catalyst
Thus, we have obtained a quantitative analysis of the
experimental data in terms of a random walker model of the
catalyst, which reveals some fine details about the nature of
the hopping process. In particular, we can show that a stickiness
(30) To simplify the model, we assume that the probability for Ni(0) to
hop into the bromophenyl ring is also given by 1/2-δ/2, although
strictly to say, the stickiness coefficient δ may be different here;
therefore, the application of a correct value would further improve
the model. We, however, intuitively suggest that a negative value of
the stickiness (i.e.,-δ/2) correctly describes the trend since the phenyl
ring should have a lower complexation ability toward Ni(0) than the
thiophene ring due to favored interactions between lone electron pairs
of sulfur atoms and unoccupied d orbitals in the structures of thiophene
and Ni(0), respectively. Quantum mechanic calculations may elucidate
quantitatively this issue, and such calculations are planned.
(31) The obtained results further suggest that Ni(0) complexes associated
with the π-conjugated systems are really existing intermediates and
they play an important role in the catalyst-transfer polymerization.
Works aiming at a direct detection of such Ni(0) complexes by 31P
NMR are currently underway in our lab.
9
J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010 7809

A R T I C L E S
Tkachov et al.
walking. It can be postulated that complexation of Ni(0) with
phenyl rings of polyphenylene is less favored than with thienyl
units; therefore, in the synthesis of polythiophene-polyphenylene
block copolymer, it is essential to start polymerization from the
polyphenylene block.
opposite end and initiate polymerization there.32 Numerical
analysis of a random hopping model was undertaken, which
revealed that a combination of a random catalyst walking along
the chain and a “sticking effect” at the end groups is operative
in KCTP. The ability of Ni catalysts to walk along polymerizing
chains must be taken into account when one uses KCTP for
the preparation of conjugated polymers.
In conclusion, the “walking” process of the Ni catalyst during
KCTP along the polymerizing poly(3-hexylthiophene), P3HT,
chains was investigated. To simplify polymer end-group iden-
tifications, a compound Br-C6H4-Ni(dppe)-Br was prepared
and used as an externally added initiator. Normally, an aryl
moiety present in an initiator incorporates into the structure of
the resulting P3HT as the end group. We demonstrated in this
work that due to the presence of C-Br group located in the
para-position to the Ni substituent of the initiator used, two
types of differently terminated polymeric products were formed.
One of them is a “normal” product, P3HT with a para-
bromophenyl end group, whereas another one has the phenyl
ring inside the P3HT chain. The content of the product with
the internal phenyl ring increases with the increase of the
polymerization degree. Control experiments demonstrated that
no intermolecular catalyst transfer takes place in the conditions
used. Such results unambiguously show that catalytic Ni(0)
species are able to walk along the polymerizing chain up to the
Acknowledgment. We thank Mrs. Harnisch for gas chromato-
graphic analyses and Dr. Luxenhofer for MALDI-TOF measure-
ments. We thank DFG for financial support (KI-1094/3-1 and KI-
1094/4-1).
Supporting Information Available: Experimental details, 1H
NMR characterizations, and details of numerical analysis. This
material is available free of charge via the Internet at http://
pubs.acs.org.
JA102210R
(32) An ability of metal catalysts to walk over σ-bound chains via upon
ethylene polymerization was also demonstrated. Johnson, L. K.;
Killian, C. M.; Brookhart, M. J. Am. Chem. Soc. 1995, 117, 6414–
6415. Mo¨hring, M.; Fink, G. Angew. Chem., Int. Ed. 1985, 24, 1001–
1002.
9
7810 J. AM. CHEM. SOC. VOL. 132, NO. 22, 2010