Transition-metal-free synthesis of conjugated polymers from bis-grignard reagents by using TEMPO as Oxidant

Article abstract of DOI:10.1002/chem.201000236

(Figure Presented) Increasing the TEMPO! Transition metals are not necessary for oxidative polymerization of 2,7-di-magnesated fluorenes. Oxidation readily occurs by using the commercially available 2,2,6,6-tetramethyl-piperine 1 -oxyl radical (TEMPO) as

Full text of DOI:10.1002/chem.201000236

DOI: 10.1002/chem.201000236
Transition-Metal-Free Synthesis of Conjugated Polymers from Bis-Grignard
Reagents by Using TEMPO as Oxidant**
Modhu Sudan Maji,[a] Thorben Pfeifer,[b] and Armido Studer*[a]
½10ꢀ
2R-MgX þ 2TEMPO
R-R þ 2TEMPO-MgX
ƒ!
Conjugated polymers have become important materials as
the emissive medium in light-emitting diodes (LED).[1] In
that regard poly(2,7-fluorenes) (PF) are often used due to
their valuable optical and electrical properties.[2] Different
methods for the preparation of this important substance
class have been developed. Initially, polyfluorene was pre-
pared by oxidative coupling of fluorene.[3] However, regiose-
lectivity problems occurred. Regioselective polymerization
has been achieved by the reductive coupling of 2,7-dibromo-
9,9-dialkylfluorenes with Zn as a reductant and NiCl2 as the
ð1Þ
ð2Þ
Ar ¼ aryl, alkenyl, alkynyl
nXMg-Ar-MgX þ 2mTEMPO !
ꢁðArÞyꢁ þ 2mTEMPO-MgX
Ar ¼ fluorene, p-diethynylbenzene, diethynylfluorene,
diethenylfluorene
We recently reported on transition-metal-free homocou-
pling of aryl, alkenyl, and alkynyl Grignard reagents by
using the 2,2,6,6-tetramethyl-piperidine 1-oxyl radical
(TEMPO) as oxidant (Eq. [1])[10,11] Since excellent yields
were obtained for these homocoupling reactions we as-
sumed that these processes might be applicable for polymer
synthesis in which excellent yields for the individual cou-
pling steps are mandatory. Herein, we will show that oxida-
tive homocoupling of Grignard compounds can be used for
the synthesis of polyfluorenes, poly(fluorenylenebutadiyny-
lenes), and poly(fluorenylenebutadienylenes) ([Eq. (2)]).[12]
For successful polymerization to occur it is absolutely
mandatory that the bis-Grignard reagents are generated in
high quality. If 2 contains some of the corresponding partial-
ly reacted mono-Grignard compound as an impurity, oxida-
tive coupling of the growing polymer with the mono-
Grignard derivative will lead to termination of the polymeri-
zation at one end. We therefore intensively studied clean
formation of bis-Grignard reagent 2 first starting with 2,7-di-
bromo-9,9-dioctylfluorene (1) by using different protocols.
We were pleased to find that by applying Mg turnings in the
presence of catalytic amounts of I2 in refluxing THF with
subsequent addition of TEMPO polyfluorene 3 was indeed
formed (Scheme 1). This experiment clearly showed that
our novel approach toward polyfluorenes was working.
However, polymerization was not efficient as 3 was formed
with an unsatisfactory low mean molecular weight (Mn =
2400 gmolꢁ1, polydispersity index (PDI)=1.53, Table 1,
entry 1, gel-permeation chromatography (GPC) was mea-
sured against a polystyrene standard with THF as the sol-
[5]
catalyst.[4] Suzuki- and Stille-type reactions[6] and NiII-cat-
alyzed Grignard metathesis (GRIM)[7] have also been suc-
cessfully used for the synthesis of polyfluorenes. However,
for Suzuki- and Stille-type coupling processes the corre-
sponding boronic acid or trialkyl tin derivatives have to be
synthesized mainly from the corresponding bromides or io-
dides prior to polymerization. This of course requires addi-
tional synthetic efforts, contributes to costs and leads to
waste generation. The currently applied methods for the
preparation of polyfluorenes all require transition metals as
catalysts. For polymers with complex side chains removal of
transition metals might be a problem.[4] Moreover, it is
known that impurities, including metals, act as quenching
sites for excitons.[8] Therefore, it is desirable to develop tran-
sition-metal-free processes for the preparation of polyfluor-
enes and other conjugated polymers.[9]
[a] Dr. M. S. Maji, Prof. Dr. A. Studer
NRW Graduate School of Chemistry
Organisch-Chemisches Institut
Westfꢀlische Wilhelms-Universitꢀt
Correnstrasse 40, 48149 Mꢁnster (Germany)
Fax : (+49)281-83-36523
E-mail: studer@uni-muenster.de
[b] T. Pfeifer
Institut fꢁr Anorganische und Analytische Chemie
Westfꢀlische Wilhelms-Universitꢀt
48149 Mꢁnster (Germany)
[**] TEMPO=2,2,6,6-tetramethyl-piperidine 1-oxyl radical.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000236.
5872
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5872 – 5875

COMMUNICATION
ment with air[10] rendering our new method economically at-
tractive. GPC measurements showed that the polymer was
formed with a Mn of 4500 gmolꢁ1 and a PDI of 1.96
(Table 1, entry 2). Increasing the concentration of the bis-
Grignard reagent to 0.25m did not have a large effect either
on Mn or on PDI (Table 1, entry 3). By using 7,7’-diiodo-
9,9,9’,9’-tetraoctylACTHNUTRGNEUGN[2,2’]bifluorenyl (4) as the starting mono-
mer under similar conditions polyfluorene 3 was obtained
with 90% yield and larger molecular weight (Mn =
9100 gmolꢁ1; PDI=1.97, Table 1, entry 4).
We also tested generation of the bis-Grignard reagent by
a Li–Br exchange reaction followed by a Li–Mg transmetal-
lation starting with monomer 1 (Scheme 1). Br–Li exchange
was conducted at ꢁ788C by addition of tBuLi for 0.5 h. Sub-
sequent Li–Mg transmetallation was achieved by adding
MgBr2 at ꢁ788C for 3 h at 0.07m monomer concentration.
Oxidative polymerization of the resulting bis-Grignard re-
agent provided 3 with a Mn of 6200 gmolꢁ1 and a PDI of
1.97 (Table 1, entry 5). Increasing the monomer concentra-
tion to 0.14m did not influence reaction outcome (Table 1,
entry 6). The best result was obtained when the Br–Li ex-
change reaction was performed at 08C over 0.5 h (tBuLi was
added at ꢁ788C), followed by a Li–Mg exchange reaction at
room temperature for 2 h at a monomer concentration of
0.20m. Oxidative polymerization allowed formation of 3
with a Mn value of 9000 gmolꢁ1 (Table 1, entry 7). Under
similar conditions with 7,7’-dibromo-9,9,9’,9’-tetraoctyl-
Scheme 1. Optimized protocol: a) tBuLi, ꢁ78 to 08C, THF, 0.5 h.
b) MgBr2·OEt2 in Et2O, 0 8C to RT, 2 h. c) TEMPO, reflux, 5 h, 94%.
Table 1. Oxidative polymerization of bis-Grignard reagents with
TEMPO under different conditions to give 3.
Entry
Monomer
Yield [%]
Mn [gmolꢁ1
]
PDI
1[a]
2[b]
3[b]
4[b]
5[c]
6[c]
7[c]
8[c]
1 [0.17m]
7 [0.17m]
7 [0.25m]
4 [0.17m]
1 [0.07m]
1 [0.14m]
1 [0.20m]
5 [0.07m]
59
92
91
90
88
81
94
94
2400
4500
4600
9100
6200
6900
9000
9700
1.53
1.96
2.00
1.97
1.97
2.01
2.12
2.11
[a] Grignard prepared with Mg turnings and catalytic I2. [b] Grignard
prepared by I–Mg exchange reaction. [c] Grignard prepared by Br–Li ex-
change reaction followed by Li–Mg transmetallation.
AHCTUNGTRENG[UNN 2,2’]bifluorenyl (5) as the starting dihalide, Mn was further
increased to 9700 gmolꢁ1 (Table 1, entry 8). To conclude
these studies, we can state that variation of the concentra-
tion did not influence the polymerization outcome to a large
extent and that Br–Li halogen exchange followed by trans-
metallation to Mg provided the best results.
vent). We assumed that the bis-Grignard reagent was not
quantitatively formed under the applied conditions.
After intensive unrewarding experimentation with the bis-
bromide 1 we decided to switch to the more reactive 2,7-
diiodo-9,9-dioctylfluorene (7) as substrate. Direct Grignard
formation by using Mg turnings was not very successful. We
therefore planned to approach formation of the requested
bis-Grignard compound of type 2 through I–Mg exchange
by using iPrMgCl·LiCl following a procedure developed by
Knochel.[13] The halogen Mg exchange reaction was per-
formed at 08C and was complete within 1.7 h as confirmed
by GC analysis of a reaction aliquot obtained after hydroly-
sis. It is important to mention that on prolonged exposure of
the bis-Grignard reagent 6 to iPrI, which was formed during
I–Mg exchange reaction as a side product, we started to ob-
serve formation of the 2-isopropyl-9,9-dioctylfluorene
mono-Grignard. Therefore, Grignard reagent 6 (0.17m) was,
after its generation (1.7 h), immediately subjected to poly-
merization by addition of TEMPO. Oxidative polymerizing
coupling was performed under refluxing conditions for 5 h.
After aqueous work-up, the crude polymer was dissolved in
dichloromethane and precipitated upon adding MeOH. The
purification protocol allowed removal of TEMPO and the
corresponding reduced hydroxylamine. It is important to
note that the hydroxylamine obtained as the TEMPO reduc-
tion product can readily be reoxidized to TEMPO by treat-
We next tested whether our oxidative polymerization
method would also allow the preparation of polymers con-
taining the 1,3-diyne moiety in the back bone. This should
be possible by homocoupling of Mg-bis-alkynyl compounds.
It is well known that polymers with bis-acetylene back
bones are not very soluble.[14] In fact, we faced serious solu-
bility problems during attempted polymerization of 10
(Scheme 2).[15] Pleasingly, solubility problems were not ob-
served during transition metal free polymerization of 9. Mo-
nomer 8 was readily synthesized from 7 (see the Supporting
Information). The bis-alkynylmagnesium chloride 9 was gen-
Scheme 2. a) iPrMgCl, THF, RT, 2.5 h. b) TEMPO, reflux, 5.3 h.
Chem. Eur. J. 2010, 16, 5872 – 5875
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5873

A. Studer et al.
erated by double deprotonation of 8 with iPrMgCl at room
temperature. After addition of TEMPO, the resulting reac-
tion mixture was refluxed for 5.3 h. Polymer 11 was obtained
in 88% yield with a Mn of 17000 gmolꢁ1 and a PDI of 2.39.
Note that upon using the Hey procedure with Pd/Cu cata-
lysts and O2 as the oxidant, the reaction is far slower and
takes about three days.[14]
We then decided to synthesize conjugated polymers con-
taining the 1,3-diene moiety in the polymer back bone using
our method. To this end monomer 12 was readily prepared
(see the Supporting Information). The bis-alkenyl Grignard
reagent was generated by an I–Mg exchange reaction and
oxidative polymerization with TEMPO was performed in re-
fluxing THF (Scheme 3). Purification by precipitation from
*
Figure 2. Emission spectra of polymer 3 and 11 (CH2Cl2); =3 (Table 1,
&
*
entry 7), =3 by GRIM method, =11.
ing Information).[5] Moreover, absorption and emission
maxima of 3 are congruent with those measured for PF in-
dependently synthesized, by us, using the GRIM method[7b]
(Mn =14000 gmolꢁ1, absorption maximum 385 nm and emis-
1
sion maximum 416 nm).[17] In addition, the H and 13C NMR
spectra of 3 further supported the regular bond connection
during polymerization (see the Supporting Information).[18]
We also characterized 11 by NMR spectroscopy (see the
Supporting Information) and by measuring absorption and
emission spectra (Figure 1 and Figure 2).
Finally, we measured the amount of transition metals
present in polymers 3 and 11. Trace analysis was conducted
for Ni, Mn, Cu, Pd, and Co (Table 2 and the Supporting In-
Scheme 3. a) iPrMgCl·LiCl, THF, 08C, 0.5 h. b) TEMPO, reflux, 5 h.
MeOH provided 14 in 84% yield (Mn =4800, PDI=1.89).[16]
To show that formation of PF took place regioregularly
(2,7-linked), we measured absorption and emission spectra
in solution and in film. Polymer 3 shows an absorption maxi-
mum at around 383 nm in CH2Cl2, which is in agreement
with the literature value (384 nm for PF with Mn =
9700 gmolꢁ1, see Figure 1 and the Supporting Informa-
tion).[7b] Polymer 3 emits at around 415 nm, which is very
characteristic for regioregular PF (Figure 2 and the Support-
Table 2. Trace analysis (all data in ppm= mggꢁ1).
Metal
3[a]
3[b]
3[c]
11
Ni
1.64ꢂ0.49
0.58ꢂ0.17
9.77ꢂ2.93
0.12ꢂ0.04
0.05ꢂ0.02
15.7ꢂ4.7
0.62ꢂ0.18
24.4ꢂ7.3
0.02ꢂ0.01
0.09ꢂ0.03
1200ꢂ360
9.19ꢂ2.76
47.6ꢂ14.3
0.03ꢂ0.01
0.09ꢂ0.03
7.27ꢂ2.18
2.47ꢂ0.74
64.0ꢂ19.2
0.19ꢂ0.06
0.14ꢂ0.04
Mn
Cu
Pd
Co
[a] Table 1, entry 4. [b] Table 1, entry 7. [c] Polymer
3
(Mn =
14000 gmolꢁ1, PDI=2.24) was prepared by the GRIM method according
to reference [7a] by using [Ni
nylphosphino)propane.
ACHTUNGTRNEN(UNG dppp)Cl2] (2 mol%); dppp=1,3-bis(diphe-
formation). PF 3 prepared by our novel method contained
transition metals as impurities in the range of 0.02 to
16 ppm. However, for polymer 11 we found a slightly larger
Cu-content (64 ppm). As reported earlier,[10a] the source of
these transition-metal impurities in our polymers lies in the
Mg (turnings or isopropyl Grignard) used to prepare the Mg
reagents. For comparison, we also prepared 3 following the
reported GRIM procedure[7b] by using iPrMgCl·LiCl from
the same batch. Polymer 3 prepared by GRIM contained
Mn (9 ppm) in slightly larger amounts than the Mn content
of 3 synthesized by the TEMPO method. However, more
*
Figure 1. Absorption spectra of polymers 3 and 11 in CH2Cl2; =3
&
*
(Table 1, entry 7), =3 by GRIM method, =11.
5874
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 5872 – 5875

Transition-Metal-Free Synthesis of Conjugated Polymers
COMMUNICATION
[9] a) S. J. Park, S.-H. Jung, J. M. Kim, H.-N. Cho, Mater. Sci. Eng. C
2004, 24, 99–102; b) H. Meng, D. F. Perepichka, M. Bendikov, F.
Wudl, G. Z. Pan, W. Yu, W. Dong, S. Brown, J. Am. Chem. Soc.
2003, 125, 15151–15162; c) A. Patra, Y. H. Wijsboom, S. S. Zade, M.
Li, Y. Sheynin, G. Leitus, M. Bendikov, J. Am. Chem. Soc. 2008,
130, 6734–6736.
[10] a) M. S. Maji, T. Pfeifer, A. Studer, Angew. Chem. 2008, 120, 9690–
9692; Angew. Chem. Int. Ed. 2008, 47, 9547–9550; b) M. S. Maji, A.
Studer, Synthesis 2009, 2467–2470.
significantly, the Ni content (1200 ppm) of the GRIM
sample is very high as compared with the Ni content of the
other samples.
In conclusion, we have presented a novel method for the
synthesis of conjugated polymers by transition-metal free
homocoupling of various bis-organomagnesium compounds
using commercially available TEMPO as an organic oxidant.
Polyfluorenes with Mn of up to 10000 gmolꢁ1 were obtained.
The method also allowed the synthesis of conjugated co-
polymers containing the bis-alkyne or the bis-ethene moiety
alternating with the fluorene unit in the polymer back bone.
Experiments were easy to run by using standard laboratory
equipment under the exclusion of air and moisture.
[11] T. Vogler, A. Studer, Synthesis 2008, 1979–1993.
[12] a) M. Kijima, I. Kinoshita, T. Hattori, H. Shirakawa, Synth. Met.
1999, 100, 61–69; b) I. Kinoshita, M. Kijima, K. Yoshikawa, Y. Mis-
hima, N. Sasaki, Synth. Met. 2003, 137, 1059–1060; c) H. N. Cho,
J. M. Hong, D. K. Moon, C. Y. Kim, Synth. Met. 2000, 111–112, 429–
431; d) K. West, C. Wang, A. S. Batsanov, M. R. Bryce, Org. Biomol.
Chem. 2008, 6, 1934–1937.
[13] a) P. Knochel, W. Dohle, N. Gommermann, F. F. Kneisel, F. Kopp, T.
Korn, I. Sapountzis, V. A. Vu, Angew. Chem. 2003, 115, 4438–4456;
Angew. Chem. Int. Ed. 2003, 42, 4302–4320; b) H. Ren, A. Krasov-
skiy, P. Knochel, Org. Lett. 2004, 6, 4215–4217.
Acknowledgements
[14] I. Kinoshita, M. Kijima, H. Shirakawa, Synth. Met. 2001, 119, 187–
188; see also, ref. [12a].
[15] During oxidative polymerization of 10 a precipitate was formed that
could not be dissolved for GPC measurements.
We thank the NRW Graduate School of Chemistry (stipend to M.S.M.)
and Novartis Pharma AG (Young Investigator Award to A.S.) for sup-
porting our work.
[16] The polymer was formed as mixture of cis/trans double bond iso-
mers. We previously showed that oxidative homocoupling of alkenyl
Grignard reagents with TEMPO occurs highly stereospecifically, see
reference [10]. Unfortunately, we were not able to synthesize iso-
merically pure starting iodide (several methods were tested).
[17] As suggested by a referee, the small absorption peak at around
440 nm for 3 in Figure 1 might derive from TEMPO impurities re-
maining in the polymer after precipitation. We proved by analyzing
a polymer sample with EPR spectroscopy, which is very sensitive for
TEMPO detection, that TEMPO traces indeed remained in that
polymer after precipitation. Importantly, we found that the absorp-
tion spectrum of PF 3 prepared from monomer 5 (see Table 1,
entry 8) showed only a minor impurity at 440 nm (see Figure S5 in
the Supporting Information) and the PF 3 prepared from monomer
7 (Table 1, entry 3) did not show any absorption at 440 nm (see Fig-
ure S4 in the Supporting Information) indicating that the peak at
440 nm derived from an impurity (likely to be TEMPO).
ꢁ
Keywords: C C coupling
oxidation · polymerization
·
fluorenes
·
nitroxides
·
[1] A. C. Grimsdale, K. L. Chan, R. E. Martin, P. G. Jokisz, A. B.
Holmes, Chem. Rev. 2009, 109, 897–1091.
[2] a) U. Scherf, E. J. W. List, Adv. Mater. 2002, 14, 477–487; b) R.
Abbel, A. P. H. Schenning, E. W. Meijer, J. Polym. Sci. Part A
Polym. Chem. 2009, 47, 4215–4233.
[3] M. Fukuda, K. Sawada, K. Yoshino, J. Polym. Sci. Part A Polym.
Chem. 1993, 31, 2465–2471.
[4] Q. Pei, Y. Yang, J. Am. Chem. Soc. 1996, 118, 7416–7417.
[5] M. Ranger, D. Rondeau, M. Leclerc, Macromolecules 1997, 30,
7686–7691.
[6] B. Tsuie, J. L. Reddinger, G. A. Sotzing, J. Soloducho, A. R. Katritz-
ky, J. R. Reynolds, J. Mater. Chem. 1999, 9, 2189–2200.
[7] a) L. Huang, S. Wu, Y. Qu, Y. Geng, F. Wang, Macromolecules 2008,
41, 8944–8947; b) M. C. Stefan, A. E. Javier, I. Osaka, R. D. McCul-
lough, Macromolecules 2009, 42, 30–32.
[18] See the Supporting Information for 1H and 13C NMR spectra and
also for absorption and emission spectra in film. As proved by mass
spectroscopic analysis (MALDI), all polymers prepared were H-ter-
minated.
[8] a) M. Reufer, F. Schindler, S. Patil, U. Scherf, J. M. Lupton, Chem.
Phys. Lett. 2003, 381, 60–66; b) J. M. Lupton, A. Pogantsch, T. Piok,
E. J. W. List, S. Patil, U. Scherf, Phys. Rev. Lett. 2002, 89, 167401–1–
167401–4.
Received: January 27, 2010
Published online: April 21, 2010
Chem. Eur. J. 2010, 16, 5872 – 5875
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5875