A comprehensive study of substituent effects on poly(dibenzofulvene)s

Article abstract of DOI:10.1039/C6NJ02861F

We herein report the first cross comparison of 14 poly(dibenzofulvene) derivatives (13 novel examples and the parent poly(dibenzofulvene)) in order to understand how the choice of substituent affects the physical properties of this interesting class of semiconducting polymers. Electron withdrawing substituents decreased the polymerization reactivity and resulted in very low molecular-weight products. Di-substituted poly(dibenzofulvene)s were found to be much less soluble than the mono-analogues which can be explained by the Hansen solubility parameter system. Analysis based on absorption, emission and electrochemistry profiles suggests that polymer solubility is a very important factor that controls the degree of stacking present in the polymer due to synthetic issues. For the first time, the thermal analysis of the parent poly(dibenzofulvene) and its derivatives is reported and it was believed that depolymerization occurred much earlier than the melting transition. We have also demonstrated the orthogonal synthesis of dibenzofulvene monomers using three distinct routes (lithiation, oxidation and Wittig) to cope with functional group compatibility.

Full text of DOI:10.1039/C6NJ02861F

NJC
View Article Online
View Journal
PAPER
A comprehensive study of substituent effects on
poly(dibenzofulvene)s†
Cite this:DOI: 10.1039/c6nj02861f
Michael Y. Wong* and Louis M. Leung
We herein report the first cross comparison of 14 poly(dibenzofulvene) derivatives (13 novel examples
and the parent poly(dibenzofulvene)) in order to understand how the choice of substituent affects the
physical properties of this interesting class of semiconducting polymers. Electron withdrawing
substituents decreased the polymerization reactivity and resulted in very low molecular-weight products.
Di-substituted poly(dibenzofulvene)s were found to be much less soluble than the mono-analogues
which can be explained by the Hansen solubility parameter system. Analysis based on absorption,
emission and electrochemistry profiles suggests that polymer solubility is a very important factor that
controls the degree of stacking present in the polymer due to synthetic issues. For the first time, the
thermal analysis of the parent poly(dibenzofulvene) and its derivatives is reported and it was believed
that depolymerization occurred much earlier than the melting transition. We have also demonstrated the
orthogonal synthesis of dibenzofulvene monomers using three distinct routes (lithiation, oxidation and
Wittig) to cope with functional group compatibility.
Received (in Montpellier, France)
11th September 2016,
Accepted 30th November 2016
DOI: 10.1039/c6nj02861f
www.rsc.org/njc
coworkers prepared a novel p-stacked polymer (PVMSiF) by
Introduction
replacing the nitrogen atom in poly(9-vinylcarbazole) with silicon.8
Poly(dibenzofulvene) (poly(DBF)) is an interesting polymer in PVMSiF possessed excellent nanofuse property with an ON–OFF
which the pendant fluorene moieties have been shown by ratio as high as 4 ꢀ 106, which was attributed to the reversible
Nakano et al. to adopt a p-stacking conformation (Fig. 1), which p-stacking conformation of the polymer under hole/electron
gives rise to a number of characteristic observations in the transport. Thanks to the absence of extended main-chain
photophysics and electrochemistry profiles of the polymer.1–4 In conjugation, poly(DBF) is also more advantageous than traditional
terms of practical application, poly(DBF) has a great potential to conducting polymers in that it is colorless and stable towards
serve as an excellent organic transistor which has stacked fluorenes photo-oxidation.
as a delocalization tunnel for charge transport, contrary to typical
While poly(DBF) is an attractive transistor or nanofuse material,
conducting polymers such as polythiophene and polyaniline it is surprising that the number of reported poly(DBF) derivatives is
that delocalize charges through the conjugated main chain.5,6 still very limited so far. These include alkyl,9–11 thiophene,12 ether13
For instance, poly(DBF) exhibits a high hole mobility of and N-alkylated amino14 groups and most of them were reported
B3 ꢀ 10ꢁ4 cm2 Vꢁ1 sꢁ1 which is of the same order of magnitude independently without cross comparison with each other. Sub-
as the inorganic selenium semiconductor.7 Recently, Li and stituent effects can play a significant role in the physical properties
of p-stacked polymers. For example, the molecular weight of poly-
(2,7-di-tert-butyldibenzofulevene) was limited to the trimer due to
the steric bulk of the tert-butyl groups.9 Moreover, the molecular
weights, photophysical behaviors and propensity to thermal depoly-
merization of poly(benzofulvene) (poly(BF))have been found to be
substituent-dependent.15 It must be pointed out that poly(BF) and
poly(DBF) are two very different classes of polymers because, while
the former can have a very high number-average molecular weight
(Mn) up to B1 700 000 g molꢁ1, the latter can only achieve an Mn of
Fig. 1 Chemical structure of poly(DBF) (left) and a schematic showing
p-stacking of fluorene moieties in the polymer chain (right).
several thousand g molꢁ1 due to high steric hindrance at the
polymerization site. Therefore, there is a definite need to study how
the substituents affect the physical properties of poly(DBF).
Department of Chemistry, Hong Kong Baptist University, Kowloon Tong,
In this contribution, we prepared 13 novel16 mono- and di-
Hong Kong SAR, China. E-mail: myw1011@gmail.com
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6nj02861f
substituted poly(DBF)s containing a wide range of substituents
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
Results and discussions
Monomer synthesis
All substituted dibenzofulvene monomers were prepared by the
Wittig route (Scheme 2, top) reported by us previously,18 with
the exception of 2-dimethylaminobenzofulvene and 2-fluorodibenzo-
fulvene. 2-Dimethylaminobenzofulvene could not be prepared
via the Wittig route because the C9-bromination of the 2-dimethyl-
aminofluorene intermediate was seriously interrupted by the ring
bromination due to the strong activation effect of the electron-
donating dimethylamino group, and thus the monomer was
prepared by the lithiation route (Scheme 2, middle). On the other
hand, 2-fluorodibenzofulvene was prepared by the oxidation
route (Scheme 2, bottom) simply because 2-fluorofluorenone
was much more commercially available than 2-fluorofluorene.
Therefore, we have demonstrated that the three synthetic routes to
dibenzofulvene monomers can be used in an orthogonal manner.
Scheme 1 A list of mono- and di-substituted poly(DBF)s generated by
free-radical polymerization in this study.
Polymer synthesis
such as electron-donating groups (e.g. MeO and NMe2), Table 1 summarizes the polymerization conditions, yields and
electron-withdrawing (e.g. CN and NO2) groups and halogens molecular weight distributions of the substituted poly(DBF)s.
(e.g. F, Br and I) which are shown in Scheme 1. We chose the All polymerizations were affected at 80 1C with AIBN as an
free-radical method for polymerization because it enjoys the initiator in degassed toluene under nitrogen atmosphere for
greatest functional group tolerance. We shall present how 20 h. Despite higher stereochemical defect formation using a
substituents affect the photophysical and electrochemical more diluted monomer concentration,2 poly(NO2DBF) and
profiles of the polymers. We found that polymer solubility is poly(Br2DBF) employed lower monomer concentrations due
an important parameter in that it controls the degree of to limited solubilities of their monomers in toluene. It has been
stacking present in the poly(DBF)s. We shall also present the reported that high molecular-weight poly(DBF) was intractable,2
thermal analysis of the parent poly(DBF) and its derivatives, therefore an usually heavy initiator loading of 5.0 mol% was
which, to the best of our knowledge, has never been reported employed as the molecular weight control. For polymers with
before.
solubilizing groups (alkyl or alkoxy), a lighter loading of 2.5 mol%
Scheme 2 Preparation of dibenzofulvene monomers via the Wittig route (top). Preparation of 2-dimethylaminobenzofulvene by the lithiation route
(middle). Preparation of 2-fluorodibenzofulvene by the oxidation route (bottom).
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Table 1 Summary of polymerization conditions,a yields and molecular weight distributions of the substituted poly(DBF)s
[M]
Initiator
(mol%)
Yields in solubleb/
soluble/total (%/%/%)
c
(g mLꢁ1
)
Mn (Xn)
PDI
Poly(DBF)
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.05
0.10
0.10
0.10
0.10
0.10
0.05
5.0
2.5
2.5
5.0
5.0
5.0
5.0
5.0
2.5
2.5
2.5
2.5
2.5
5.0
15.0/16.8/31.8
0/42.9/42.9
0/25.0/25.0
1409 (7.9)
2313 (11.1)
3752 (17.0)
1610 (8.2)
1141 (4.4)
1335 (4.4)
506, 325, 116 (—)
515, 306 (—)
348 (—)
2194 (6.8)
—
1212 (3.8)
3304 (9.3)
—
1.35
2.03
2.13
1.87
1.78
1.69
—
—
—
1.66
—
1.78
1.67
—
Poly(MeODBF)
Poly(NMe2DBF)
Poly(FDBF)
Poly(BrDBF)
Poly(IDBF)
0/36.9/36.9
38.4/29.7/68.1
53.1/31.3/84.4
45.6/21.1/66.7
26.1/29.1/55.2
70.9/23.9/94.8
0/60.1/60.1
69.1/6.3/75.4
45.8/17.2/63.0
0/39.8/39.8
Poly(CNDBF)
Poly(NO2DBF)
Poly(NO2PrODBF)
Poly(NO2HexODBF)
Poly(BrMeODBF)
Poly(BrPrODBF)
Poly(BrHexODBF)
Poly(Br2DBF)
51.6/0/51.6
a
b
All polymerizations were affected at 80 1C with AIBN as an initiator in degassed toluene under nitrogen atmosphere for 20 h. Insoluble in THF.
With respect to the polystyrene standards.
c
was employed. Many substituted poly(DBF)s in this study have significantly smaller molecular weights than the fluorine analogue,
both soluble and insoluble fractions (THF used as the extracting poly(FDBF), which can be attributed to the steric hindrance of
solvent). According to the infrared absorption analysis (Fig. S10 heavy halogens.9 In addition, poly(CNDBF) and poly(NO2DBF) gave
and S11, ESI†), the insoluble fraction has the same chemical only oligomers (Mn o 500). Given that the cyano group should not
structure as the soluble counterpart and thus it can be deduced have caused significant side reactions during the polymerization
that the former fraction contains polymer aggregates with a (because AIBN contains the cyano functionality as well) and that it
higher molecular weight than the latter fraction.2 In the pioneering imposes little steric hindrance due to its linearity, it seems to be
work on the parent poly(DBF) by Nanako et al., they obtained total appropriate to attribute the lack of polymerization reactivity to the
yields and Mns from 22 to 73% and from 250 to 2190, respectively, electronic effect exerted by the electron-withdrawing effect of cyano
which agree well with our results.2 Monomers with heavy halogens and nitro groups.
1
and electron withdrawing substituents were found not to poly-
The H NMR spectra of poly(MeODBF) and poly(NMe2DBF)
merize smoothly. For example, poly(BrDBF) and poly(IDBF) gave and their corresponding fluorenes are shown in Fig. 2. In these
Fig. 2 1H NMR spectra of poly(MeODBF) and poly(NMe2DBF) and their corresponding fluorenes to demonstrate anisotropic shielding as the
consequence of p-stacking conformation in the polymers.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
polymers, both the aromatic and methyl protons exhibit significant
upfield shifts compared with the monomeric fluorenes which is
attributed to anisotropic shielding in the p–p stacking system.4,5
Similar phenomena are observed in poly(benzofulvene),19,20
and oligophenylurea.21 On the other hand, poly(9,90-dimethyl-
2-vinylfluorene), a close analogue to poly(DBF) but without p–p
stacking, does not show such an upfield shift for the aromatic
protons.4
Polymer solubilities
The solubility of the substituted poly(DBF)s can be defined as
the relative ratio of the soluble fraction to the insoluble counter-
part. We believe that solubility is a very important parameter of
the poly(DBF)s because we noticed that more soluble poly(DBF)s
(i.e. high ratio of soluble fraction yield to insoluble counterpart)
demonstrate higher molecular weights which greatly affect their
photophysical and electrochemical properties (vide infra). For
example, by lengthening the alkoxy group, molecular weights
were enhanced from poly(NO2PrODBF) to poly(NO2HexODBF)
and from poly(BrPrODBF) to poly(BrHexODBF), because the less
soluble propoxy-substituted polymers suffered more seriously
from premature precipitation of the insoluble growing chain
during polymerizations.22 In addition, poly(NMe2DBF), with the
highest molecular weight among all poly(DBF)s, is totally solu-
ble. We noted that di-substituted poly(DBF)s have much lower
solubilities than the mono-substituted analogues. For example,
even with additional solubilizing alkoxy groups, di-substituted
poly(BrMeODBF) and poly(NO2PrODBF) are significantly less
soluble than mono-substituted poly(BrDBF) and poly(NO2DBF),
respectively. Furthermore, poly(Br2DBF) is completely insoluble.
These observations can be explained by the well-noted Hansen
solubility parameter (d) system23 using the following expression:
Fig. 3 Alternate stacking of electron-rich (blue) and electron-deficient (red)
phenyl rings in the di-substituted poly(BrRODBF)s and poly(NO2RODBF)s due
to favorable dipole interactions, where R denotes general alkyl groups. dpp is
hence much larger than dps
.
solvent (dds) due to the additional large bromine atom with a
significant amount of polarizable electron density. Based on the
same reason, a trend of decreasing solubilities is also observed
from mono-substituted poly(FDBF) to poly(IDBF) when the size
of halogens increases (Table 1).
Photophysical and electrochemical properties
The absorption, photophysical and electrochemical properties
of the substituted poly(DBF)s are summarized in Table 2. The
lowest-energy absorption peak wavelengths of the substituted
poly(DBF)s and their corresponding fluorenes were recorded
and the difference between them (Dl) were compiled. A positive
Dl (i.e. bathochromic shift after polymerization) is the result of
the stacking conformation of the adjacent fluorene moieties in
the polymer chain.4 The bathochromic shift has been shown to
level off when the number of stacking fluorene units approaches 5.4
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ÁÃ
Â
Ã
Â
Ã
2
2
2
Dd ¼
2 ddp ꢁ dds þ dpp ꢁ dps þ dhp ꢁ dhs
(1)
where the first subscripts ‘‘d’’, ‘‘p’’ and ‘‘h’’ refer to the disper- Therefore, it is reasonable for low molecular-weight polymers such
sion force, polarity and hydrogen-bonding, respectively; and the as poly(CNDBF), poly(NO2DBF) and poly(NO2PrODBF) to have
second subscripts ‘‘p’’ or ‘‘s’’ refer to the polymer and solvent, negligible Dl (ꢁ1 to 2 nm). As the Dl depends on the molecular
respectively. Simply put, a small Dd results in a good solubility of weight, it follows naturally that poly(DBF)s with solubilizing
the polymer in the particular solvent in question. dh should be substituents exhibit a larger Dl, which is confirmed by the compar-
irrelevant because p–p stacking is not a hydrogen-bonding ison between poly(NO2PrODBF) and poly(NO2HexODBF) as well as
interaction and there is no –OH or –NH substituent used in this between poly(BrPrODBF) and poly(BrHexODBF). In addition,
study which can form hydrogen bonds. For the di-substituted the magnitude of Dl was found to depend on the nature of the
poly(BrMeODBF) and poly(NO2PrODBF), the polar parameters substituent. For example, while poly(DBF), poly(MeODBF) and
(dp) should be the main factors that result in their low solubilities. poly(NMe2DBF) have considerably higher molecular weights
As demonstrated in Fig. 3, bromine and nitro groups are among the polymers studied, their Dls (6, 13 and 10 nm
electron-withdrawing and make the phenyl ring attached to be respectively) are considerably different. This suggests that the
electron-deficient (shown in red), whereas the alkoxy group is effectiveness of conjugation between adjacent stacked fluorene
electron-donating and makes the other phenyl ring electron-rich moieties might be affected by substituents. Therefore, caution
(shown in blue). As a result, it is likely that in the di-substituted must be taken when using Dl to estimate the degree of p–p
poly(DBF)s the electron-deficient and electron-rich phenyl rings stacking present in the polymers. In general, a very small Dl
will be stacked with each other in an alternate fashion due to (e.g. o5 nm) may translate to little or non-existent stacking,
electrostatic attraction. Therefore, dpp is much larger than dps whereas a large Dl (e.g. 410 nm) may suggest the presence of a
followed by a large Dd and insolubility. In the case of considerable amount of stacking (Fig. 4).
poly(Br2DBF) which is totally insoluble, the dispersion force
The cyclic voltammograms of the substituted poly(DBF)s showed
parameter of the polymer (ddp) is much larger than that of the that all of them underwent irreversible oxidations except for
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
Table 2 Photophysical and electrochemical data of the substituted poly(DBF)s in this study
Absorptiona
Emissionb
Electrochemistryc
d
f
g
l
polymer/lfluorene
Dle
(nm)
lpolymer/ldefect
Eox,polymer/Eox,fluorene
(V/V vs. Fc/Fc+)
DEox
(nm nmꢁ1
)
(nm nmꢁ1
)
(mV)
Poly(DBF)
309/303
323/310
339/329
311/305
317/307
317/308
314/312
329/330
355/355
360/355
—
6
13
10
6
10
9
403/322
409/333
437/371
397/321
402/338
404/334
428/336
432/375
535/402
537/401
—
0.859/0.999
0.807/0.854
0.213/0.246
0.955/1.032
1.057/1.053
1.025/1.019
1.132/1.125
1.273/1.281
0.951/0.940
0.864/0.935
—
ꢁ140
ꢁ47
ꢁ33
ꢁ77
4
Poly(MeODBF)
Poly(NMe2DBF)
Poly(FDBF)
Poly(BrDBF)
Poly(IDBF)
6
7
Poly(CNDBF)
2
Poly(NO2DBF)
Poly(NO2PrODBF)
Poly(NO2HexODBF)
Poly(BrMeODBF)h
Poly(BrPrODBF)
Poly(BrHexODBF)
Poly(Br2DBF)h
ꢁ1
ꢁ8
0
11
5
—
6
13
—
ꢁ71
—
ꢁ23
ꢁ67
—
319/313
326/313
—
401/339
403/340
—
0.843/0.866
0.805/0.872
—
a
b
c
Measured in THF at a concentration of 100 ppm. Measured in THF at a concentration of 10 ppm. Measured in THF under nitrogen atmosphere
d
e
with 0.1 M n-Bu4NPF6 as the supporting electrolyte. Absorption maxima of the corresponding substituted fluorenes. Dl = lpolymer ꢁ lfluorene
.
f
g
h
Oxidation potential of the corresponding substituted fluorenes. DEox = Eox,polymer ꢁ Eox,fluorene
.
Not measured due to the lack of the soluble part.
internal standard and the LUMO levels are estimated from the
optical band gaps of the polymers from absorption onset wave-
lengths, and the results are summarized in Table S1 (ESI†). The
strongly electron-donating dimethylamino group (poly(NMe2DBF))
raises the HOMO level to ꢁ5.01 eV, whereas the potent electron-
withdrawing cyano (poly(CNDBF)) and nitro groups (poly(NO2DBF),
poly(NO2PrODBF) and poly(NO2HexODBF)) deepen the LUMO
levels to ꢁ2.58 eV and from ꢁ3.07 eV to ꢁ3.14 eV respectively.
For the rest of the poly(DBF)s, the HOMO and LUMO levels are
in general ꢁ5.6 eV and ꢁ2.0 eV, respectively. The oxidation
potentials of the polymers are compared with those of the
corresponding fluorenes and the differences (DEoxs) are listed
in Table 2. A negative DEox results from the additional stabilization
of the resulting radical cation by delocalization through p-stacking
present in the polymers.4 In most cases, DEox agrees well with
Dl reported in the aforementioned absorption analysis. For
example, DEox becomes progressively more negative (i.e. more
pronounced p-stacking in the polymer) from poly(NO2PrODBF)
to poly(NO2HexODBF) and from poly(BrPrODBF) to poly-
(BrHexODBF) due to the increase in molecular weights. Yet,
the case of halogen-substituted poly(DBF)s is less trivial. While
poly(FDBF) shows a clear agreement between DEox and Dl,
poly(BrDBF) and poly(IDBF) do not. One plausible explanation
is that while Dl measures the ground state properties of the
polymers, DEox involves the stability of the radical cation (i.e.
excited species).4 We deduce that Dl probably gives a more
accurate picture than DEox about the degree of stacking present
in the poly(DBF)s because the polymer chain in the excited state
has such a high energy that a weak secondary interaction like
p–p stacking may be broken, or that the excited polymer chain
interacts with the solvent or another chain which is not relevant
to the p stacking in the polymers.
Fig. 4 Normalized absorption spectra of poly(MeODBF) (top) and poly-
(NMe2DBF) (bottom) together with their corresponding fluorenes, showing
red-shifts in the lowest-energy absorption peaks after polymerization due
to stacking conformation.
All substituted poly(DBF)s showed both characteristic excimer3,4
and defect (i.e. monomeric fluorene) emissions (see Fig. 5 for
poly(NMe2DBF) (Fig. S17 and S18, ESI†). The HOMO levels of examples). The presence of defects is inevitable under high
the substituted poly(DBF)s are determined from the ferrocene temperature conditions used in free-radical polymerization.2
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
Table 3 DSC and TGA data of the substituted poly(DBF)s, both soluble
and insoluble parts
DSCa
TGAb
Tpeakc/DHd (1C/kJ molꢁ1
)
Tonset (1C)
e
Soluble
part
Insoluble Soluble
Insoluble
part
part
part
Poly(DBF)
171/2.05
236/5.18
227/8.20
Not observed
164/4.24
158/4.48
188/9.91
220/13.69
230/19.83
178/2.95
—
—
290
327, 596
338
285
277
331
—
—
—
331
369
Poly(MeODBF)
Poly(NMe2DBF)
Poly(FDBF)
Poly(BrDBF)
Poly(IDBF)
Poly(CNDBF)
Poly(NO2DBF)
Poly(NO2PrODBF)
—
166/5.32
156/6.48
254
201/11.14 273, 592 f 346, 662
230/14.50 300
236/21.55 315
—
206/12.07
209/12.59 281
—
209/9.08
338
362
—
346
354
—
Poly(NO2HexODBF) 228/22.55
342
—
Poly(BrMeODBF)
Poly(BrPrODBF)
Poly(BrHexODBF)
Poly(Br2DBF)
—
204/11.15
196/12.32
—
300
—
338, 585
a
b
Under nitrogen atmosphere at a heating rate of 20 1C minꢁ1
.
Under
nitrogen atmosphere at a heating rate of 40 1C minꢁ1
.
Temperature of
c
d
endotherm peak. Enthalpy change associated with the endotherm.
e
f
Temperature at 5% weight loss. Second phase of weight loss.
Fig. 5 Normalized emission spectra of poly(MeODBF) (top) and poly-
(NMe2DBF) (bottom) showing both excimer and defect emissions. The arrows
indicate the emissions from defects in the polymer chains which agree very
well with the emission of the corresponding fluorenes.
We found that the relative emission intensity of the excimer to
the defect is not a reliable indicator of the degree of stacking
present in the polymer chain. For example, poly(CNDBF)
showed exclusively excimer emission at 428 nm with very small
defect emission at 324 nm (Fig. S16, ESI†). However, both GPC
and absorption analyses suggest that poly(CNDBF) has a small
molecular weight followed by limited degree of stacking. Indeed,
unlike absorption and electrochemical properties which depend
Fig. 6 DSC traces of some substituted poly(DBF)s in this study.
on the stacking of several adjacent fluorene units, excimer 1H NMR and GPC techniques and it was found that the dibenzo-
emission could be observed in much smaller oligomers, even in fulvene monomer was regenerated (i.e. characteristic peak at
dimers.4
d = 6.0 ppm) and that the polymer was decomposed into species
with lower molecular weights (Fig. S19, ESI†). Thus, the
endotherm can be assigned to either depolymerization or melting
Thermal properties
The thermal properties of the substituted poly(DBF)s were processes. Given the large steric bulk at the vinyl functional group
studied by differential scanning calorimetry (DSC) and thermo- in DBF monomers at which polymerization occurs, it can be
gravimetric analysis (TGA) and the data are summarized in inferred that the ceiling temperature (Tc) for this type of polymers
Table 3. Previous studies of poly(DBF) focused heavily on the should be low. For example, a sterically hindered polymer like
conformation4,24 of the pendant fluorene units and charge poly(a-methyl styrene) has a Tc as low as 61 1C (under the
transport5–7 and this study, to the best of our knowledge, conditions of a neat a-methyl styrene monomer).25 In addition,
reveals for the first time the thermal behaviors of poly(DBF) free-radical polymerizations of the parent poly(DBF) at 80 1C
and its derivatives. The DSC trace (Fig. 6) of the parent soluble reported by Nakano’s group2 and us consistently produced poly-
poly(DBF) shows an endotherm with a temperature peak (Tpeak
)
mers with low molecular weights (e.g. Xn o 10), further suggesting
at 171 1C with an enthalpy change (DH) of 2.05 kJ molꢁ1. The low Tcs for this class of polymers. Therefore, it is very unlikely that
annealed sample after the DSC experiment was analyzed by the depolymerizations of poly(DBF)s were initiated at high
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
temperatures like above 170 1C (DSC endotherms shown in
1H and 13C NMR spectra were recorded on a Bruker-AF301
Fig. 6). On the other hand, the DH of the endotherms (Fig. 6) AT 400 MHz spectrometer using CDCl3 as the solvent and
has a magnitude close to typical melting process for polymers.26 tetramethylsilane as the internal standard. High resolution mass
Furthermore, DH depends largely on the nature of the substituent. spectrometry (HRMS) was carried out using a Bruker autoflex
For example, the DHs of soluble poly(NO2DBF), poly(NO2PrODBF) MALDI-TOF mass spectrometer. Melting points were measured
and poly(NO2HexODBF) are 13.69 kJ molꢁ1, 19.83 kJ molꢁ1 and on a MEL-TEMP capillary tube apparatus and uncorrected.
22.5 kJ molꢁ1, respectively which suggests that additional Infrared spectra were recorded using a Nicolet Magna 550 Series
heat was used to melt the alkoxy chains. The length of the II FTIR spectrometer using KBr pellets. UV-vis absorption spectra
alkoxy chains in the three polymers should have little effect were measured in THF at a 100 ppm concentration using a
on the depolymerization process because of their remote dis- Varian Cary 200 spectrophotometer. Fluorescence spectra were
tance from the polymerization site. Depolymerization processes also measured in THF at a 10 ppm concentration using a Perkin
for vinyl polymers have been reported to show either a sharp15 Elmer LS55 luminescence spectrometer. Differential scanning
or a broad19,27 peak in DSC analysis. DBF monomers, due calorimetry (DSC) analyses were performed on a Perkin-Elmer
to their large steric bulk at the polymerization site, should Pyris Diamond DSC under nitrogen purge at a heating rate of
have a low heat of polymerization (DHp).25 Therefore, it is safe 20 1C minꢁ1. Thermogravimetric analyses (TGA) were achieved
to assume that the depolymerizations of the substituted using a Perkin-Elmer TGA-6 thermal analyzer under nitrogen at a
poly(DBF)s occurred at a much lower temperature before the heat rate of 40 1C minꢁ1. Cyclic voltammetry (CV) measurements
melting endotherms and were hardly discernible in DSC were performed on a BAS CV-50W electrochemical analyzer
analysis because the depolymerization endotherms were too adapted with a conventional three-electrode configuration
broad and low. While the insoluble part of the substituted consisting of platinum working and auxiliary electrodes and a
poly(DBF)s have practically the same Tpeak as their soluble Ag/AgCl reference electrode. All measurements were affected in
counterparts, their DHs are invariably higher which can be 0.1 M THF solution with tetrabutylammonium hexafluorophos-
explained by the presence of stronger inter-chain interactions in phate as the electrolyte and ferrocene as the internal standard.
the insoluble part. The substituted poly(DBF)s show moderate Molecular weights were determined by gel permeation chromato-
thermal stability with their onset decomposition temperature graphy (GPC) using a HP 1050 series HPLC with THF as the eluent
(Tonset, defined as 5 wt% loss) ranging from 254 1C to 369 1C. and a UV absorption detector (at 254 nm) calibrated against
The insoluble part always has a higher Tonset compared with polystyrene MW standards.
their soluble counterpart.
Synthesis
2-N,N-Dimethylaminofluorene was prepared according to the
literature.28 2-Fluorofluorene was prepared by the Wolff–Kishner
Conclusions
reduction of 2-fluorofluorenone.29 2-Methoxyfluorene, 2-cyano-
This study is the first comprehensive cross comparison of sub-
stituent effects on poly(DBF)s and it contributes to the under-
standing of the structure–property relationship of p-stacked
polymers, which is relatively underexplored. Di-substituted
poly(DBF)s are much less soluble than the mono-analogues
which can be explained by the Hansen solubility parameters.
Solubility should not be overlooked because it controls the
degree of stacking present in the poly(DBF)s. Lastly, it is
believed that depolymerization occurred much earlier than
the melting transition in DSC analysis which means that
poly(DBF)s have limited thermal stability.
fluorene, 2-nitrofluorene, 2-nitro-7-alkoxyfluorenes and 2-bromo-7-
alkoxyfluorenes were prepared according to the literature.18 Except
for 2-N,N-dimethylaminodibenzofulvene and 2-fluorodibenzo-
fulvene, the synthesis of other DBF monomers have been
reported previously.18
2-N,N-Dimethylaminodibenzofulvene
To an ice-water cooled solution of 2-N,N-dimethylamino-
fluorene (500 mg, 2.39 mmol) in dry THF, 1.6 M n-butyl-
lithium (1.49 mL, 2.39 mmol) was dropwise added. The reaction
mixture was stirred for 10 min. Paraformaldehyde (71.7 mg,
2.39 mmol) was then added in a single portion. After 30 minutes,
Experimental
General information
Fluorene, 2-bromofluorene, 2-iodofluorene, 2,7-dibromofluorene, the reaction was quenched by pouring into water. The mixture
paraformaldehyde, 2-fluorofluorenone, p-toluenesulfonic acid, was extracted with DCM several times. The combined organic
n-butyllithium (1.6 M in hexane solution) and methylmagnesium was dried with anhydrous magnesium sulfate. After concen-
bromide (3.0 M in ether solution) were purchased from Sigma tration, the crude was purified by flash column chromatography
Aldrich. Toluene used in polymerization was distilled freshly using an eluent (EtOAc: hexanes = 1 : 2, v : v, with 3% Et3N
once from calcium hydride and AIBN were recrystallized from as an additive) to give 9-hydroxymethyl-2-N,N-dimethylamino-
methanol before use. Other solvents were of analytical grade and fluorene (275 mg, 1.15 mmol). The alcohol was then mixed
used as received.
with potassium hydroxide (1 g) in methanol (10 mL) and
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.

View Article Online
Paper
NJC
allowed to stir overnight. DCM was then added and the reaction The filtrates thus collected were concentrated. The soluble
mixture was filtered through a short plug of silica gel. fractions were recovered by precipitation again in hexane.
After concentration, the crude was purified by flash column
chromatography using an eluent (DCM : hexanes = 1 : 5, v : v,
with 3% Et3N as an additive) to offer 2-N,N-dimethylamino-
Acknowledgements
dibenzofulvene (238 mg, 45.1%) as a yellow solid. Mp 118 1C; dH
We are grateful to receive financial support for this work from
(CDCl3) 7.64 (1H, d, J = 7.6 Hz, ArH), 7.52 (2H, d, J = 8.0 Hz,
the Hong Kong Research Grants Council (HKBU 2105/06E).
ArH), 7.29 (1H, dt, J = 7.6 and 0.8 Hz, ArH), 7.15 (1H, dt, J = 7.6
and 0.8 Hz, ArH), 7.11 (1H, d, J = 2.4 Hz, ArH), 6.76 (1H, dd,
J = 8.0 and 2.4 Hz, ArH), 6.15 (1H, s, vinyl), 6.12 (1H, s, vinyl),
3.03 (6H, s, NMe2); dC (CDCl3) 150.53, 144.03, 141.03, 139.50,
Notes and references
137.68, 129.69, 128.73, 125.14, 120.71, 120.40, 118.43, 113.44,
106.71, 105.08, 41.08; HRMS (MALDI) [M+] calcd for C16H15N:
221.1204; found: 221.1195.
1 T. Nakano, Polym. J., 2010, 42, 103–123.
2 T. Nakano, T. Yade, Y. Fukuda, T. Yamaguchi and S. Okumura,
Macromolecules, 2005, 38, 8140–8148.
2-Fluorodibenzofulvene
3 T. Nakano, K. Takewaki, T. Yade and Y. Okamoto, J. Am. Chem.
Soc., 2001, 123, 9182–9183.
4 T. Nakano and T. Yade, J. Am. Chem. Soc., 2003, 125,
15474–15484.
5 W. S. Kim, J. Kim, J. K. Park, S. Mukamel, S. K. Rhee, Y. K.
Choi and J. Y. Lee, J. Phys. Chem. B, 2005, 109, 2686–2692.
6 V. Coropceanu, T. Nakano, N. E. Gruhn, O. Kwon, T. Yade,
To an ice-water cooled solution of 2-fluoro-9-fluorenone
(500 mg, 2.72 mmol) in dry THF, 3.0 M methylmagnesium
bromide (1.1 mL, 3.26 mmol) was added dropwise. After stirring
for 30 minutes, the reaction was quenched by pouring into
water. The mixture was extracted with chloroform several times.
The combined organic fractions were dried with anhydrous
magnesium sulfate. After concentration in vacuo, the crude alcohol
was mixed with p-toluenesulfonic acid (2.34 g, 13.6 mmol) and
refluxed in 10 mL benzene/chloroform (1 : 1) for 30 minutes.
The solvent was then removed and the crude was purified
´
K.-I. Katsukawa and J.-L. Bredas, J. Phys. Chem. B, 2006, 110,
9482–9487.
7 T. Nakano, T. Yade, M. Yokoyama and N. Nagayama, Chem.
Lett., 2004, 33, 296–297.
8 H. Li, Z. Wang, C. Song, Y. Wang, Z. Lin, J. Xiao, R. Chen,
C. Zheng and W. Huang, J. Mater. Chem. C, 2014, 2, 6946–6953.
9 T. Yade and T. Nakano, J. Polym. Sci., Part A: Polym. Chem.,
2006, 44, 561–572.
by column chromatography using hexane as the eluent to 10 T. Nakano, PCT Int. Appl., WO 2002088202 A1 20021107,
offer 2-fluorodibenzofulvene as colorless oil (331 mg,
2002.
68.7%). dH (CDCl3) 7.70 (1H, d, J = 7.6 Hz, ArH), 7.63–7.58 11 T. Nakano, O. Nakagawa, M. Tsuji, M. Tanikawa, T. Yade
(2H, m, ArH), 7.39–7.33 (2H, m ArH), 7.27 (1H, dt, J = 7.6 and 0.8
and Y. Okamoto, Chem. Commun., 2004, 144–145.
Hz, ArH), 7.08–7.03 (1H, m, ArH), 6.09 (1H, s, vinyl), 6.03 12 T. Nakano, A. Kotani and M.-T. Gao, Jpn Kokai Tokkyo Koho,
(1H, s, vinyl); dC (CDCl3) 163.96, 161.53, 142.78, 142.75,
JP 2007302729 A 20071122, 2007.
140.13, 140.04, 139.45, 138.02, 136.16, 136.14, 128.95, 126.75, 13 T. Nakano, K. Tanaka and H. Higashimura, Jpn Kokai Tokkyo
121.06, 120.82, 120.73, 119.47, 115.80, 115.56, 114.09, 108.93,
Koho, JP 2013209606 A 20131010, 2013.
108.46, 108.23; HRMS (MALDI) [M+] calcd for C14H9F: 196.0688; 14 T. Nakano, A. Kotani and M. Gao, Jpn Kokai Tokkyo Koho,
a
found: 196.0677.
JP 2008280377 A 20081120, 2008.
15 A. Cappelli, S. Galeazzi, G. Giuliani, M. Anzini, A. Donati,
L. Zetta, R. Mendichi, M. Aggravi, G. Giorgi, E. Paccagnini
and S. Vomero, Macromolecules, 2007, 40, 3005–3014.
General polymerization procedures
Freshly prepared monomers (250 mg), a catalytic amount
of AIBN (2.5 mol% or 5.0 mol%) and distilled toluene (2.5 or 16 We noticed that Li et al. reported poly(Br2DBF),17 but it
5.0 mL) were added into a Schlenk tube. The system was then
degassed by five freeze–pump–thaw cycles. The polymerization
was then affected by stirring at 80 1C for 20 h under nitrogen
atmosphere. For monomers that produced only soluble pro-
ducts, the polymerization mixture was a homogeneous solution
must be pointed out the chemical structure of their polymer
was incorrectly drawn (Polymer, 2006, 47, 7889–7899, page
7984). Under Suzuki polymerization conditions, it should be
the bromine atoms which underwent the polymerization
but not the vinyl group.
at the end of the reaction. The polymers were collected by 17 Z. Li, C. Di, Z. Zhu, G. Yu, Z. Li, Q. Zeng, Q. Li, Y. Liu and
precipitation in hexane twice. On the other hand, for mono- J. Qin, Polymer, 2006, 47, 7889–7899.
mers that yielded both soluble and insoluble products, the 18 M. Y. Wong and L. M. Leung, Tetrahedron, 2010, 66,
polymerization mixture was a suspension. The suspension was 3973–3977.
first precipitated in hexane. The polymer thus obtained was 19 A. Cappelli, S. Galeazzi, G. Giuliani, M. Anzini, M. Aggravi,
soaked in three successive batches of 10 mL THF to resolve the
soluble and insoluble fractions. The solids that remained
insoluble in THF were identified as the insoluble fraction.
A. Donati, L. Zetta, A. C. Boccia, R. Mendichi, G. Giorgi,
E. Paccagnini and S. Vomero, Macromolecules, 2008, 41,
2324–2334.
New J. Chem.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016

View Article Online
NJC
Paper
20 A. Cappelli, G. P. Mohr, M. Anzini, S. Vomero, A. Donati, 24 T. Nakano and T. Yade, Chem. Lett., 2008, 37, 258–259.
M. Casolaro, R. Mendichi, G. Giorgi and F. Makovec, J. Org. 25 A. Rudin and P. Choi, The Elements of Polymer Science and
Chem., 2003, 68, 9473–9476.
21 F. D. Lewis, T. L. Kurth and G. B. Delos Santos, J. Phys. 26 J. E. Mark, Physical Properties of Polymers Handbook, Springer,
Chem. B, 2005, 109, 4893–4899. New York, 2007.
22 We observed that most of the polymerization solutions 27 M. Elladiou and C. S. Patrickios, Macromolecules, 2015, 48,
Engineering, Academic Press, Cambridge, 2012.
turned to a suspension soon after the polymerization was
initiated (B0.5–1 h).
23 J. W. Van Dyk, H. L. Frlsch and D. T. Wu, Ind. Eng. Chem.
Prod. Res. Dev., 1985, 24, 473–478.
7503–7512.
28 C.-W. Lee, M.-P. Kung, C. Hou and H.-F. Kung, Nucl. Med.
Biol., 2003, 30, 573–580.
29 H. Minlon, J. Am. Chem. Soc., 1946, 68, 2487–2488.
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2016
New J. Chem.