Polymer library comprising fluorene and carbazole homo- and copolymers for selective single-walled carbon nanotubes extraction

Article abstract of DOI:10.1021/ma201890g

To date, (n, m) single-walled carbon nanotubes (SWNTs) cannot be selectively synthesized. Therefore, postprocessing of SWNTs including solubilization and sorting is necessary for further applications. Toward this goal, we have synthesized a polymer library consisting of fluorene- and carbazole-based homo- and copolymers. Variations of the connection of these aromatics together with the incorporation of further conjugated monomers give access to a broad diversity of polymers. Their ability to selectively wrap specific (n, m) species is investigated toward HiPco SWNTs raw material which contains more than 40 (n, m) species. Absorption and fluorescence spectroscopies were used to analyze SWNTs/polymer suspensions. These results provide evidence for selective SWNTs/polymer interactions and allow a more detailed assessment of polymer structure-property relationships, thus paving the way toward custom synthesis of polymers for single (n, m) SWNTs extraction.

Full text of DOI:10.1021/ma201890g

Article
pubs.acs.org/Macromolecules
Polymer Library Comprising Fluorene and Carbazole Homo- and
Copolymers for Selective Single-Walled Carbon Nanotubes
Extraction
Fabien Lemasson,† Nicolas Berton,† Jana Tittmann,† Frank Hennrich,† Manfred M. Kappes,*,†,‡,§
and Marcel Mayor*,†,§,⊥
†Institute for Nanotechnology (INT), Karlsruhe Institute of Technology (KIT), 76021 Karlsruhe, Germany
‡Institute of Physical Chemistry, Karlsruhe Institute of Technology, 76021 Karlsruhe, Germany
§DFG Center for Functional Nanostructures, 76028 Karlsruhe, Germany
⊥Department of Chemistry, University of Basel, 4003 Basel, Switzerland
S
* Supporting Information
ABSTRACT: To date, (n, m) single-walled carbon nanotubes (SWNTs) cannot be selectively synthesized. Therefore,
postprocessing of SWNTs including solubilization and sorting is necessary for further applications. Toward this goal, we have
synthesized a polymer library consisting of fluorene- and carbazole-based homo- and copolymers. Variations of the connection of
these aromatics together with the incorporation of further conjugated monomers give access to a broad diversity of polymers.
Their ability to selectively wrap specific (n, m) species is investigated toward HiPco SWNTs raw material which contains more
than 40 (n, m) species. Absorption and fluorescence spectroscopies were used to analyze SWNTs/polymer suspensions. These
results provide evidence for selective SWNTs/polymer interactions and allow a more detailed assessment of polymer structure−
property relationships, thus paving the way toward custom synthesis of polymers for single (n, m) SWNTs extraction.
functionalization.4 The former approach induces defects on
the SWNT sidewalls which degrade their physical properties. In
INTRODUCTION
Single-walled carbon nanotubes (SWNTs) exhibit unique
■
the latter approach, the SWNT intrinsic properties remain
physical and mechanical properties, making them excellent
mainly unchanged due to the preservation of the sp2-carbon
candidates for applications in the fields of optics, electronics,
high-strength fibers, and electrolytes.1 The corresponding raw
hybridization of their backbone. This approach was chosen to
solubilize SWNTs for enrichment and sorting and several
materials are produced in bulk scale using several different
methods were developed for this purpose, which include DNA
wrapping followed by ion-exchange chromatography,5 solubili-
zation using detergents followed by either dielectrophoresis,6
density gradient centrifugation7 or nonlinear density gradient
centrifugation,8 dispersion with small molecules,9 or polymer
wrapping.10 Recently, several types of conjugated polymers have
been shown to be selective toward a few (n, m) species like
poly(9,9-dialkyl-2,7-fluorene) for large chiral angles (close to
processes including chemical vapor deposition (CVD), laser
vaporization, or arc discharge. However, all these methods are
not selective in terms of electronic structure type (metallic or
semiconducting) or chiral angle (n, m).2 Despite many efforts in
this direction, selective synthetic methods yielding monodis-
perse (n, m) SWNT samples are still not available. As their
physical properties are mainly dictated by their chiral angle,
postprocessing including sorting of SWNTs is needed to tackle
fundamental issues as well as for applications in electronics,
optoelectronics, and photonics.3
Received: August 19, 2011
Revised: December 16, 2011
Published: December 29, 2011
Any postprocessing approach requires the preparation of
SWNT suspensions, either using covalent or noncovalent
© 2011 American Chemical Society
713
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
armchair: θ ≥ 25°)10a,c,h,i or poly(N-decyl-2,7-carbazole) for
lower chiral angles (typically 10° ≤ θ ≤ 20°).10b At the same
time they also exhibit strong selectivity toward semiconducting
SWNTs. This approach is of great interest due to the relatively
cheap availability of these materials and due to the perspectives
for further improvement thanks to the great variety of structural
modifications possible with these types of structures.
For convenience, the yields and molecular weights of the
polymers described in this article are summarized in Table 2.
Diboronic ester 2, which displays a 2,7-fluorenyl moiety, was
used to synthesize a large part of the library and was reacted in a
Suzuki coupling with 11 comonomers to access a variety of
polymers. Thus, reaction with its parents 2,7-dibromo-9,9-
didodecylfluorene 1 yielded poly(9,9-didodecylfluorene-2,7-diyl)
P1, in very good yield (99%). Copolymer P2, featuring a 2,7 and
a 3,6 fluorenyl moiety, was obtained by reaction of diboronic
ester 2 with dibromid 10 in very good yield (99%). Strictly
alternating copolymers P6 and P10, in which a 2,7 and 3,6
carbazolyl moiety is installed, were synthesized from dibromide 3
and 11 in 98% and 94% yield, respectively. Electron-withdrawing
groups such as BTD (benzo-2,1,3-thiadiazole) and anthraqui-
none were also introduced via their corresponding 4,7-dibromo-
2,1,3-benzothiadiazole and 1,5-dichloroanthraquinone to yield
polymers P13 and P15 in 96% and 80% yields, respectively.
Finally, a series of five polyaromatics comonomers were reacted
with diboronic ester 2, namely 1,4-dibromonaphthalene, 1,5-
bis(trifluoromethylsulfonyloxy)naphthalene, 1,5-dibromoanthra-
cene, 2,6-dibromoanthracene, and 9,10-dibromoanthracene to
access polymers P23 (90%), P17 (95%), P19 (88%), P21
(94%), and P22 (88%), respectively.
Diboronic ester 4, displaying a 2,7-carbazolyl moiety, was also
reacted in a Suzuki coupling with various (pseudo)haloaromatics
to access a large diversity of polymers. Reaction with its parent
dibromide 3 yielded poly(N-decyl-2,7-carbazole) in P7 in 76%
yield. Polymers P3 and P11, which display a 3,6-fluorenyl and
carbazolyl moieties, were obtained from dibromide 10 and 11 in
94 and 89% yield, respectively. The reddish polymer P14 was
synthesized from dibromide 5 in 29% yield. The low yield might
be attributed to the low solubility of this polymer. Finally,
reaction of diboronic ester 4 with 1,5-dichloroanthraquinone,
1,5-bis(trifluoromethylsulfonyloxy)naphthalene, and 1,5-dibro-
moanthracene allowed the synthesis of polymers P16 (39%),
P18 (97%), and P20 (79%), respectively.
Despite numerous studies comprising experimental probes as
well as molecular modeling of polymer/SWNT interactions, the
reasons for selective solubilization of SWNTs are not fully under-
stood.10−12 In order to investigate the correlation between the
polymer structures and its SWNT dispersing features, a
systematic variation of the polymers structure is desirable.
Here we have therefore investigated a small polymer library
comprising homo- and copolymers based on 9,9-didodecyl-
fluorene and N-alkylcarbazole units connected either in the
2,7 or 3,6 position. The selectivity of these polymers was
investigated toward HiPco SWNTs as this material contains
more than 40 different (n, m) tube types ranging from 0.7 to
1.3 nm diameter.13 Despite a wide range of parameters able to
influence the polymer/SWNT interactions, significant informa-
tion based on polymeric structural aspect was identified.
Although many variables are still not identified, the structural
diversity of the library allows to get more insight into specific
interactions of polymer/SWNTs and hopefully paves the way
toward custom sorting of single (n, m) species.
RESULTS AND DISCUSSION
■
Synthesis of Monomers and Polymers. The library
comprises four families which include 23 homo- and copoly-
mers stemming from (i) 9,9-didodecyl-2,7-fluorenyl, (ii) 9,9-
didodecyl-3,6-fluorenyl, (iii) N-alkyl-2,7-carbazolyl, and (iv)
N-alkyl-3,6-carbazolyl. Additionally, combinations of the
aforementioned subunits either with available dihaloaromatics
or with themselves completed the collection of copolymers. An
overview of the library is given in Table 1. All these conjugated
polymers were synthesized by Suzuki coupling, except for
polymer P5 which was synthesized by Yamamoto coupling and
polymer P9 which was assembled by Kumada coupling
(Scheme 3). Details of synthetic procedures for monomers
and polymers are given in the Supporting Information. The
synthesis of several polymers presented in this study has already
been reported, namely polymers P1,14 P7,10b and P15−P20.10f
Yamamoto and Kumada couplings required dihaloaromatics
as buiding blocks, whereas Suzuki cross-coupling required
boronic esters and a building block comprising two halides or
pseudohalides. Diboronic esters 2,15 4,10b and 616 were used in
this study to access the diversity of the library. They were
synthesized by palladium-catalyzed Miyaura borylation of their
corresponding dibromides 1, 3, and 5, respectively (Scheme 1).
Monoboronic ester 12 (Scheme 2) was synthesized by
monolithiation of dibromide 1117 followed by quenching with
isopropoxypinacolborane, yielding monomer 12 in 37% yield.
The synthesis of 3,6-dibromofluorene 9 was shortened from
five to three steps18 by starting from the commercially available
phenanthrenequinone to access 3,6-dibromo-9,9-didodecyl-
fluorene 10 after alkylation (Scheme 2). Direct reduction of
ketone 8 using either hydrazine hydrate19 or polysiloxane in
combination with tris(pentafluorophenyl)borane afforded
3,6-dibromofluorene in good yields. Reduction of ketone 8
using magnesium in dry methanol yields the alcohol.20 Other
haloaromatics were commercially available or synthesized
according to the literature (see Supporting Information).
Diboronic ester 6 was reacted first with dibromide 10 to
access polymer P4 (87%) which allowed the introduction of a
BTD moiety to a 3,6-fluorenyl moiety. Moreover, the low
solubility of polymer P14 led us to increase the number of
carbon atoms in the N-alkyl side chain. Thus, diboronic ester 6
was reacted with carbazole 7, which displays a branched
2-hexyldecyl chain at the carbazole nitrogen atom, to give the
reddish polymer P8 in 76% yield.
Finally, three homopolymers were obtained as follow. A
Suzuki coupling of the bifunctional carbazole 12 yielded poly(N-
decylcarbazole-3,6-diyl) P12, and homopolymer P5 was
obtained via Yamamoto homocoupling of its corresponding
dibromide 10. As previously observed for similar 3,6-linked
1
structures, it should be noticed that H NMR spectrum and
SEC data of P5 and P12 point at the presence of cyclic
oligomers within the polymer sample.18,21 The first attempt to
synthesize polymer P9 was performed via a Yamamoto coupling
of dibromide 7 which yielded high molecular weight polymer P9
barely soluble in toluene (see Supporting Information). As
further investigation of the polymers takes place in toluene, the
polymers should be soluble in this solvent. Therefore, dibromide
7 was engaged in a Kumada coupling to yield a lower molecular
weight polymer P9 (Mn = 1189 g mol−1) which exhibits much
higher solubility in common organic solvents including toluene.
Preparation of Polymer/SWNT Dispersions. Dispersions
of SWNTs were obtained by sonicating pristine HiPco SWNTs
in toluene with an excess of the polymer under investigation.
714
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
a
Table 1. Overview of the Conjugated Polymers Library
a
Polymers dispersing SWNTs are shaded in green (good), light green (medium), and yellow (poor dispersing properties). Pale red shading indicates
that the corresponding suspension was not stable.
Typically, ∼1 mg of HiPco SWNTs was added to a solution of
50 mg of the investigated polymer in toluene (15 mL), and the
suspension was sonicated for 1 h. Following an established
procedure, centrifugation with a mild centripetal acceleration
(10 000 rpm, 10 min) was applied to remove large bundles of
nanotubes and remaining catalyst particles. A density gradient
centrifugation (DGC) was applied to remove the excess polymer
and to investigate the possibility of further enrichment.10f The
dispersions were analyzed by absorption and photolumines-
cence−excitation spectroscopy (PLE), allowing the identification
and quantification of semiconducting SWNT species (see also
Supporting Information).13,22
procedure) chiral angle (θ) vs diameter (Ø) maps in order to
visualize the composition of the SWNTs dispersion (Figure 2).
The SWNT distribution in HiPco material is represented by
sodium cholate dispersions in D2O, assuming that all SWNT are
equally well dispersed in this surfactant. Relative intensity maps
of each polymer suspension are also presented in order to allow
facile comparison to the original nonuniform distribution of
SWNTs in the HiPco material (Figure 3). Relative intensity
maps represent the ratio between normalized PL maps and the
normalized HiPco SWNTs suspended in D2O using sodium
cholate (i.e., if a polymer disperses all SWNTs equally well, then
the relative intensity map would contain only circles with the
same diameter). We note that the accuracy of such relative
intensity maps is likely to be lower in regions of higher diameters
due to the correspondingly low amount of SWNTs present.
Dispersing Properties. All recorded PL maps (Figure 1 and
Figures S3, S4, S6, and S7) were translated into normalized (see
Supporting Information for details concerning the normalizing
715
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
now. Thus, the polymers were classified into three categories
according to the amount of SWNT (in mg/L) which was
dispersed (cf. Table S1 in Supporting Information and related
color code in Table 1). Note that this qualitative estimation
depends on the parent composition of HiPco tubes. Having a
polymer which exclusively selects a (n, m) species which has
a low concentration in the starting raw material would lead to a
low concentration in the polymer/SWNT suspension. More-
over, this estimation fits fairly well with the absorption spectra of
the suspensions displayed in Figure S2. Because of a low signal-
to-noise ratio, the absorption spectra of low concentrated
suspensions were not displayed.
Scheme 1. Synthesis of Boronic Esters 2, 4, and 6 Used for
the Synthesis of the Polymer Library
a
Finally, some of the reported polymers formed unstable
suspensions with SWNTs and precipitation occurred within a
few minutes or hours. These polymers, namely P2−P5, P10,
P14, and P20, were not investigated further.
Polymers Containing a 2,7-Fluorene Moiety. As pre-
viously mentioned, poly(2,7-dialkylfluorene)s are known to
be extremely selective toward close-to-armchair SWNTs
(θ > 25°) and poly(9,9-didodecylfluorene-2,7-diyl) P1 was
found to behave similarly (Figure 2). It should be noticed that
the length of the alkyl chains at the C-bridging atom has an
influence on the diameter distribution which shifts toward
higher diameter tubes. Indeed, the main species in the polymer
P1/SWNT dispersion is the (7, 6) species, while in comparable
conditions (dispersions of HiPco SWNTs in toluene) poly(9,9-
dioctylfluorene-2,7-diyl) was reported to mainly disperse the
(8, 6) species having a slightly larger diameter.10b,c,h Also,
poly(9,9-dihexylfluorene-2,7-diyl) mainly disperses (8, 7) and
(9, 7) SWNTs species under similar conditions.10c,e Thus, our
results confirm that the diameter of the dispersed nanotubes is
inversely proportional to the alkyl chain length, whereas the
θ-selectivity is not affected by alkyl chain length variations. This
pronounced effect of the length of the alkyl chains of poly(9,9-
dialkylfluorene-2,7-diyl) on the diameter selectivity, though not
fully understood, is consistent with the predominance of CH/π
interactions between the alkyl chains and the zigzag bonds of
the fused hexagons on the nanotubes walls as has recently been
suggested by molecular dynamics simulations.10b,12 However,
a
Reagents and conditions: (a) PdCl2(dppf), bis(pinacolato)diboron,
dioxane, 80 °C, overnight.
For most of the polymers, the DGC did not allow a signifi-
cant enrichment of the dispersions and similar PL maps were
obtained before and after DGC. Figure 2 presents the chiral
angle (θ) vs diameter (Ø) maps of these suspensions, and Figure 3
contains their corresponding relative intensities (versus sodium
cholate as described in the previous paragraph). However, we
observed a clear enhancement of the selectivity caused by DGC
for three copolymers comprising a 9,9-didodecyl-2,7-fluorenyl
subunit (P6, P13, and P23). The corresponding chiral angle
(θ) vs diameter (Ø) maps obtained from PL maps before and
after DGC treatment are presented in Figure 4.
In order to get a qualitative estimate of the SWNT dispersion
efficiency of HiPco tubes for each polymer (i.e., how many
SWNTs from the parent HiPco mixture are suspended by a
polymer regardless to the (n, m) indice), the sum of the
intensities of all PL maps peaks attributed to SWNTs (before
DGC) was considered. This estimation does not consider any
difference in quantum yields of each SWNT as the data required
to provide an analysis considering individual quantum yields
for each (polymer-wrapped) SWNT type are not available until
a
Scheme 2. Synthesis of Monomers 7, 10, and 11
a
Reagents and conditions: (a) hydrazine monohydrate, KOH, triethylene glycol, 80 °C to reflux, 68%;19 (b) B(C6F5)3, poly(triethoxysilane),
CH2Cl2, RT, 67%; (c) n-BuLi, isopropoxypinacolborane, THF, −78 °C to RT, 37%.
716
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
Scheme 3. Overview of the Polymerization Procedures Used in This Study
Table 2. Summary of the Synthesis of the Polymers
Presented in Table 1 and Used in This Study, Including
Yield, Number-Average Molecular Weight, Weight-Average
Molecular Weight, and Polydispersity Index
polymer
yield (%)
Mn (Da)
Mw (Da)
PDI
ref
P1
99
99
94
87
83
98
12878
8608
53510
14136
8404
4.2
1.6
1.5
2.0
5.6
3.1
14b
P2
P3
5471
P4
2350
4760
Figure 1. Representative PLE maps of HiPco SWNTs suspensions
prepared from polymers P9 and P12 in toluene which allow the
identification of each semiconducting (n, m) species. These maps are
translated into θ/Ø maps allowing an overview of diameter and chiral
index selectivity (Figure 2).
P5
11920
18840
66990
58620
P6
a
b
b
b
P7
20 (76)
3402
5333
1.6
10b
P8
76
17510
50490
2.9
c
c
c
c
P9
87
1189
1987
1.7
properties, polymer P1 is also estimated to have good disper-
sion efficiency (cf. Table S1).
P10
P11
P12
P13
P14
P15
P16
P17
P18
P19
P20
P21
P22
P23
94
89
98
96
10722
4461
19125
6912
1.6
1.6
1.5
2.0
Changing the selectivity of poly(2,7-fluorene) can be achieved
by introduction of aromatic moieties in the polymer main chain.
Indeed, polymer P19 exhibits strong diameter selectivity toward
SWNTs having Ø > 0.95 nm, which corresponds to a minor
fraction in the starting HiPco material (Figures 2 and 3).10f In
contrast, there is no significant chiral angle selectivity: the
dispersed SWNTs chiral angles θ range from close to zigzag
(12, 1) to armchair (8, 7). Despite the dispersion of only a
minor fraction of HiPco nanotubes, polymer P19 yields high
concentrations of SWNTs in suspension with regards to the
other polymers; i.e., it disperses very efficiently large diameter
SWNTs (cf. Table S1). Similar but not so pronounced selecting
properties were observed for polymer P15 having an electron-
poor 1,5-anthraquinone moiety in the polymer backbone
(Figure 2). Its relative intensity map indicates affinity for high
diameter SWNTs (Figure 3). However, the dispersion ability of
polymer P15 is much lower than that of polymer P19.
Noteworthy is that SWNTs with θ < 10° were not selected.
Introduction of a 1,5-naphthyl moiety as in polymer P17 did not
lead to pronounced diameter selectivity. Structural selectivity is
directed toward close-to-armchair SWNTs motifs (Figures 2
and 3). The dispersion ability of P17 is also reduced compared
with polymer P19. The anthracene connectivity of polymer P19
was varied from 1,5 to 2,6 and 9,10 positions leading to
3035
4552
29000
58000
a
e
e
e
16 (29)
2250
2450
1.1
80
39
95
97
88
11861
2390
25764
3245
2.2
1.4
3.7
5.7
2.4
10f
10f
10f
10f
10f
10f
59735
23278
21165
220990
133480
51175
a
e
e
e
24 (79)
6900
16800
2.4
94
12364
17100
16800
29632
24000
29200
2.4
1.4
1.7
d
38 (88)
90
a
Fraction stemming from Soxhlet extraction with toluene. The
percentage in parentheses corresponds to the total yield which
comprises a further extraction with chloroform. Value given for the
toluene fraction. Synthesized by Kumada coupling. Value given for
the chloroform fraction. The percentage in parentheses corresponds to
the total yield which comprises a further extraction with toluene.
Value given for the toluene fraction which was not fully soluble in
THFthe solvent in which the SEC measurement was performed.
b
c
d
e
it can be noticed that the relative intensities of P1/SWNT
dispersion presented in Figure 2 suggests a less pronounced
selectivity toward θ > 25° for Ø > 0.95 nm. Besides its sorting
717
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
Figure 2. Normalized θ/Ø maps of HiPco SWNTs dispersed in aqueous solution using sodium cholate compared to HiPco SWNTs dispersed in
toluene using the polymers of this study (after DGC) (see Table 1). Within an individual θ/Ø map, the circles areas are proportional to the
concentration of the corresponding SWNT species in this dispersion. In the sodium cholate map (top left), the circles are assigned to the (n, m)
indices of the corresponding SWNT.
polymers P21 and P22, respectively. Neither of these two
polymers is efficient at dispersing SWNTs having Ø < 0.85 nm
(Figures 2 and 3). The suspension stemming from polymer P22
is relatively more concentrated in low chiral index SWNTs com-
pared to polymer P21/SWNTs dispersions. However, polymer
P22 exhibits poor overall solubilizing properties compared with
polymer P21. Neither of them exhibits a diameter selectivity as
pronounced as observed with the parent polymer P19.
Polymers Containing a 3,6-Fluorene Moiety. One major
observation of these results stems from the family of polymers
containing 3,6-fluorenyl units. Whereas polymers derived from
9,9-dialkyl-2,7-fluorene units show good to outstanding dispersing
properties and high wrapping selectivity, all homologous polymers
containing 9,9-dialkyl-3,6-fluorenes, namely P2−P5, yield un-
stable suspensions with SWNTs precipitating within minutes.
Such behavior demonstrates that the 3,6-connectivity prevents
SWNTs dispersion, which can in turn be attributed to an
unfavorable polymer conformation that does not enable efficient
wrapping of the nanotubes. An analogous unfavorable con-
formation has been proposed in the case of foldamers which are
able to reversibly suspend and release SWNTs by adopting a
different conformation by tuning solvent polarity.23
Polymers Containing a 2,7-Carbazole Moiety. While
poly(9,9-didodecylfluorene-2,7-diyl) P1 exhibits selectivity
718
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
Figure 3. θ/Ø maps presenting normalized relative intensities of HiPco SWNTs dispersed in toluene using the polymers of this study (after DGC).
Intensities are relative to those of the same (n, m) SWNT species in sodium cholate/water dispersion.
toward high chiral angle SWNTs, its structural proxy poly(N-
decylcarbazole-2,7-diyl) P7 exhibits selectivity for lower chiral
angles, namely between 10° < θ < 20°, making both polymers
nicely complementary.10b Another noticeable difference is the
dispersing efficiency which was estimated to be 5 times lower in
the case of polymer P7 compared to polymer P1. This can be
attributed to the single linear alkyl chain attached at the N
atom of the carbazole moiety. A way to increase its dispersing
efficiency is to introduce a branched alkyl chain at the N atom,
yielding poly(N-2-hexyldecylcarbazole-2,7-diyl) P9. Indeed, the
dispersing efficiency is then estimated to increase by a factor of 2.
Polymer P9 exhibits selectivity features relatively similar to those
of its parent poly(2,7-carbazole) P7 despite the increased
bulkiness of the side chain (Figures 1−3). Note that polymer P7
disperses mainly the (11, 3) and (10, 5) species, while polymer
P9 selects preferentially (9, 4) and (9, 5) SWNTs, which
represents a slight shift toward smaller diameters and higher
θ values (around 20°). Thus, as observed for poly(9,9-
dialkylfluorene)s, the nature of the alkyl chain only slightly
influences the selectivity. As previously postulated and simulated,
the selectivity for lower chiral indices may arise from the sp2
hybridization of the N-bridging atom which reduces the steric
hindrance at this position, thus allowing a tighter complex with
the SWNT than the poly(9,9-dialkyl-2,7-fluorene)s which have a
sterically more demanding sp3-hybridized C-bridging atom.10b
The low solubility of copolymer P14 did not allow us to test
its dispersing properties toward SWNTs (vide supra). The
introduction of a branched alkyl chain at the N atom of the
carbazole moiety afforded the more soluble polymer P8, and
its dispersing properties toward HiPco SWNTs were tested.
719
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
Figure 4. Normalized θ/Ø maps of HiPco SWNTs dispersed in toluene using polymers P6, P13, and P23 before DGC (second row). First row:
normalized relative intensity θ/Ø maps corresponding to the second row. Third and fourth row: normalized θ/Ø maps after DGC. In the case of
polymer P6, DGC fractions 25 and 35 are presented due to an inhomogeneous enrichment along the DGC tube, in contrast to suspensions arising
from polymers P13 and P23 which exhibit homogeneous enrichment.
P8 was found to disperse most SWNTs species, regardless of the
diameter and chiral angle, with the (7, 6) and (8, 6) nanotubes
dominating the distribution (Figure 2). The θ/Ø map rep-
resenting the relative intensities of SWNTs dispersed by
polymer P8 shows a tendency to suspend SWNTs of higher
diameter regardless of the chiral angle (Figure 3). Furthermore,
polymer P8 exhibits high dispersion efficiency.
Copolymers P16 and P18 which contain a 2,7-carbazole
moiety do not exhibit the same trend and their selectivity is
mainly driven by the 1,5-linked anthraquinone and naphthalene
groups respectively, as previously reported.10f Indeed, polymer
P16 is selective toward small diameter SWNTs having Ø < 0.9 nm
(Figures 2 and 3) but shows extremely low dispersing effi-
ciency. Polymer P18 dominantly disperses high chiral indices
θ > 25° with only a minor fraction of θ < 25°. Part of this effect
may also be attributable to the HiPco SWNT distribution as
shown by the related relative intensity θ/Ø map (Figure 3).
Even though polymer P18 yields higher concentration of SWNTs
than polymer P16, it is not an efficient dispersing agent.
Polymers Containing a 3,6-Carbazole. As shown in
Figures 1 and 2, poly(N-decylcarbazole-3,6-diyl) P12 is able to
disperse SWNTs, in contrast to its poly(3,6-fluorene) homo-
logue P5. However, no strong selectivity was observed as P12
disperses most of the nanotubes species initially present and
exhibits a θ/Ø map similar to the one obtained for sodium
cholate aqueous dispersions. The corresponding relative
intensity θ/Ø map confirms this behavior while also underlining
that large diameter SWNTs are well dispersed. Moreover,
polymer P12 shows excellent dispersing efficiency (cf. Table
S1). The other mixed polymeric structures P10 and P11, both
having a 3,6-linked carbazole subunit in their backbone, yielded
unstable suspensions of SWNTs, thus showing behavior similar
to their 3,6-fluorene homologues P2 and P3. It should be
pointed out that P11/SWNTs suspensions are slightly more
stable and precipitated within a few hours instead of minutes,
thus allowing fluorescence measurements to be recorded.
Although the intensity was relatively low, the PL map indicated
selectivity toward θ ≥ 15° and diameters ≥ 0.9 nm (Figure S7).
Enhancement Induced by DGC: Polymers P6, P13,
and P23. In most cases the polymer wrapping is not that
uniform that further fractionation by DGC is possible. However,
in the case of the polymers P6, P13, and P23, besides their
720
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
intrinsic wrapping selectivity (vide infra), a clear enhancement
caused by DGC treatment was observed (Figure 4). Copolymer
P6, which possesses strictly alternating 2,7-linked fluorene and
carbazole moieties, is an interesting example of a combination of
wrapping selectivities. Given the complementary selectivity of
poly(9,9-dialkylfluorene-2,7-diyl) and poly(N-alkylcarbazole-2,7-
diyl) toward high- and low-chiral angle SWNTs, respectively, it
was expected that P6 would combine both properties. Indeed, as
evidenced from the θ/Ø map of P6/SWNT dispersion in
toluene before DGC (Figure 4), the dispersion of a wide range
of SWNTs was observed. This tendency is slightly attenuated in
the relative intensity θ/Ø map which shows the preferential
solubilization of higher diameter SWNTs together with the
absence of θ sorting (Figure 4). Moreover, the dispersion
efficiency of polymer P6 is high. Interestingly, a post-DGC
inhomogeneous enrichment effect was found in this system by
harvesting/analyzing discrete fractions within the density
gradient. In particular, DGC fraction 25 was enriched with
SWNTs having θ ≥ 25°, mainly the (7, 6) and (8, 6) species,
whereas DGC fraction 35 showed diameter-selective features
with Ø ≥ 1.0 nm. The corresponding evolution of the SWNTs
distributionreflecting characteristic buoyant density differ-
enceswas further confirmed by absorbance spectrometry
(Figure S5). Thus, the nanotubes having the biggest diameter
move to higher (= lower density) regions within the density
gradient than do smaller nanotubes. The opposite trend has
been observed for aqueous dispersions of SWNTs using sodium
cholate, where DGC-induced diameter sorting results in smaller-
diameter SWNTs settling out at higher points in the centrifuge
tube.4a This indicates that in the case of P6 the polymer coating
strongly influences the density of the polymer/SWNT
complexes. Note that in most other cases of polymer/SWNTs
dispersions studied here we did not find analogous enrichment
effects via DGC, possibly because of nonuniform coating of the
polymer on the SWNT.10f In contrast, the coating density of
polymer P6 on SWNTs might be relatively uniform for each
single (n, m) species and vary significantly from one species to
another, leading to the differentiation of the buoyant coefficients
for polymer/SWNTs complexes, depending upon diameter/
(n, m) characteristics.
BTD-containing copolymer P13 showed reduced dispersion
efficiency compared to copolymer P8. Copolymer P13 leads to
a suspension in which (8, 6) nanotubes are predominant, with a
noticeable amount of (9, 4) and (10, 5) as well as other tubes
(Figure 4). In this case, applying DGC to the dispersion resulted
in a highly enriched suspension of (8, 6) and (10, 5) tubes. All
other species, in particular the (9, 4) tube, were efficiently
removed during the DGC step. An opposite trend was observed
by replacing the electron-deficient BTD unit in copolymer P13
by a naphthalene unit in polymer P23. Polymer 23 did not show
significant sorting properties before DGC. However, the relative
intensity map indicates the preferential dispersion of smaller
diameter SWNTs (Figure 4). Moreover, after DGC treatment
the relative amount of the (9, 4) tubes dramatically increased
compared to the other tubes, which might indicate specific
interactions between polymer P23 and the (9, 4) tube. It is
worth noting that energy transfer is observed in the case of the
low-band-gap polymer P13 (Figure S6). Indeed, we attributed
the PL signals observed at excitation wavelength around
500 nm (corresponding to the absorption onset of polymer
P13, Figure S8) with emission wavelength at around 1200 and
1280 nm, corresponding to the (8, 6) and (10, 5) tubes to
energy transfers from the polymer to the corresponding tubes.24
Further features stemming from similar energy transfer were not
observed, probably due to the PL analysis window which is out
of the absorption region of the described polymers.
CONCLUSION
■
We have synthesized a polymer library comprising conjugated
homopolymers and strictly alternating copolymers based on
fluorene and carbazole groups in which their connectivity was
varied in either 2,7- or 3,6-positions. The selectivity of these
polymers toward specific (n, m) SWNTs was investigated with
HiPco material containing more than 40 different nanotube
species. The dispersing properties were estimated from optical
absorption spectra and from photoluminescence intensities,
allowing to classify the polymers into three categories depending
upon their dispersing efficiency. An indirect correlation between
dispersing efficiency and selectivity is observed due to somewhat
inhomogeneous distribution of (n, m) species in the raw
material. Exceptions are polymer P19, which shows high con-
centrated suspensions despite the low “availability” of SWNTs
having Ø > 0.95 nm in raw HiPco material, and polymers P16,
P18, and P22, which show low dispersion efficiency despite
selecting tubes of high availability.
The trends observed in the SWNTs dispersing features with
this polymer library can be summarized as follows: (a) In
contrast to the stable suspensions formed with 2,7-connected
fluorenyl and/or carbazolyl, SWNTs dispersions derived from
polymers containing 3,6-connected building blocks exhibited
rapid precipitation of the SWNTs, with the noticeable
exception of poly(N-decyl-3,6-carbazole) P12 which efficiently
disperses all (n, m) SWNTs species. (b) While poly(9,9-
didodecylfluorene-2,7-diyl) P1 is selective toward high chiral
angle-SWNTs (θ > 20°), poly(N-alkyl-2,7-carbazole) P7 and
P9 show a strong preference toward SWNTs with 10° < θ <
20°. Interestingly, this trend was not affected by the increased
bulkiness of the N-alkyl side chain of poly(2,7-carbazole), which
confirms that the sp2 hybridization at the N atom may be the
origin of the difference in selectivity with poly(9,9-dialkyl-2,7-
fluorene). However, the dispersing efficiency was strongly
improved, which can be attributed to the enhanced solubility of
the polymer. (c) The behavior of alternating fluorene- and
carbazole-based copolymers toward SWNTs can be tuned via
the chemical nature and connectivity of the subunits. In
particular, efficient diameter sorting is achieved by introducing
1,5-linked anthracene units (polymer P19). (d) Although in
most cases DGC has no impact on the selectivity, DGC-induced
separation of SWNTs was observed for some of the copolymers
comprising 9,9-didodecylfluorene-2,7-diyl subunits. In particular,
DGC fraction-dependent SWNT compositions were observed
for copolymers P6, P13, and P23.
This study provides valuable information on structure−
property relationships of conjugated homo- and copolymers
featuring 2,7- and 3,6-fluorenyl and carbazolyl building blocks
in particular, as pertaining to their interactions with semi-
conducting SWNTs. Fine tuning of comonomers allowed a high
degree of enrichment in terms of (n, m) species as well as in
terms of diameters. Ultimately, we expect such studies to pave
the way toward custom polymer design for comprehensive direct
extraction of single species of SWNTs corresponding to the
requirements for specific applications in electronics, optoelec-
tronics, or photonics.
In spite of the systematic structural variation within the
(co)polymer libraries of this and other recent studies, clear
design rules for a polymer to be able to enrich a particular (n, m)
721
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722

Macromolecules
Article
Sadokierska, I.; Lebedkin, S.; Kappes, M. M.; Michel, T.; Alvarez, L.;
Sauvajol, J.-L.; Meunier, S.; Mioskowski, C. Nano Lett. 2008, 8, 1830−
1835.
SWNT species still remain to be established. On the experimental
side, we are currently investigating structural aspects like the
peripheral alkyl chain or the backbone composition of the polymer
in further detail. At the same time we hope that our present
study will stimulate further computational modeling of polymer
wrapping of semiconducting (and metallic) SWNTs.
(10) (a) Sturzl, N.; Hennrich, F.; Lebedkin, S.; Kappes, M. M. J. Phys.
̈
Chem. C 2009, 113, 14628−14632. (b) Lemasson, F. A.; Strunk, T.;
Gerstel, P.; Hennrich, F.; Lebedkin, S.; Barner-Kowollik, C.; Wenzel,
W.; Kappes, M. M.; Mayor, M. J. Am. Chem. Soc. 2011, 133, 652−655.
(c) Hwang, J.-Y.; Nish, A.; Doig, J.; Douven, S.; Chen, C. W.; Chen,
L.-C.; Nicholas, R. J. J. Am. Chem. Soc. 2008, 130, 3543−3553.
(d) Cheng, F.; Imin, P.; Maunders, C.; Botton, G.; Adronov, A.
Macromolecules 2008, 41, 2304−2308. (e) Nish, A.; Hwang, J.-Y.;
Doig, J.; Nicholas, R. J. Nature Nanotechnol. 2007, 2, 640−646.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental section containing synthetic procedures and
characterization of monomers and polymers, description of
the preparation and characterization of polymer/SWNTs
dispersions, and additional PL maps. This material is available
free of charge via the Internet at http://pubs.acs.org.
(f) Berton, N.; Lemasson, F.; Tittmann, J.; Sturzl, N.; Hennrich, F.;
̈
Kappes, M. M.; Mayor, M. Chem. Mater. 2011, 23, 2237−2249.
(g) Ozawa, H.; Fujigaya, T.; Song, S.; Suh, H.; Nakashima, N. Chem.
Lett. 2011, 40, 470−472. (h) Chen, F.; Wang, B.; Chen, Y.; Li, L.-J.
Nano Lett. 2007, 7, 3013−3017. (i) Wang, W. Z.; Li, W. F.; Pan, X. Y.;
Li, C. M.; Li, L.-J.; Mu, Y. G.; Rogers, J. A.; Chan-Park, M. B. Adv.
Funct. Mater. 2011, 21, 1643−1651.
AUTHOR INFORMATION
Corresponding Author
*E-mail: marcel.mayor@unibas.ch (M.M.); manfred.kappes@
kit.edu (M.M.K.).
■
(11) Star, A.; Stoddart, J. F.; Steuerman, D.; Diehl, M.; Boukai, A.;
Wong, E. W.; Yang, X.; Chung, S. W.; Choi, H.; Heath, J. R. Angew.
Chem., Int. Ed. 2001, 40, 1721−1725. Kang, Y. K.; Lee, O. S.; Deria, P.;
Kim, S. H.; Park, T. H.; Bonnell, D. A.; Saven, J. G.; Therien, M. J.
Nano Lett. 2009, 9, 1414−1418. Caddeo, C.; Melis, C.; Colombo, L.;
Mattoni, A. J. Phys. Chem. C 2010, 114, 21109−21113.
ACKNOWLEDGMENTS
■
We thank Peter Gerstel and Christopher Barner-Kowollik for
size exclusion chromatography measurements. We also
acknowledge ongoing generous support by the DFG Center
for Functional Nanostructures (CFN), by the Helmholtz
Programme POF NanoMikro, by the Karlsruhe Institute of
Technology (KIT), and by the University of Basel.
(12) (a) Ozawa, H.; Fujigaya, T.; Niidome, Y.; Hotta, N.; Fujiki, M.;
Nakashima, N. J. Am. Chem. Soc. 2011, 133, 2651−2657. (b) Gao, J.;
Loi, M. A.; de Carvalho, E. J. F.; dos Santos, M. C. ACS Nano 2011, 5,
3993−3999.
(13) Bachilo, S. M.; Strano, S. M.; Kittrell, C.; Hauge, R. H.; Smalley,
R. E.; Weisman, R. B. Science 2002, 298, 2361−2366.
(14) (a) Teetsov, J.; Fox, M. A. J. Mater. Chem. 1999, 9, 2117−2122.
REFERENCES
■
(b) Lemasson, F.; Tittmann, J.; Hennrich, F.; Sturzl, N.; Malik, S.;
̈
(1) (a) Reich, S.; Thomsen, C.; Maultzsch, J. Carbon Nanotubes:
Basic Concepts and Physical Properties; Wiley-VCH: Weinheim, 2004.
(b) O’Connell, M. J. Carbon Nanotubes: Properties and Applications;
CRC Press Taylor & Francis Group: Boca Raton, FL, 2006. (c) Ajayan,
P. M. Chem. Rev. 1999, 99, 1787−1799. (d) Moniruzzaman, M.;
Winey, K. I. Macromolecules 2006, 39, 5194−5205. (e) Ryu, S.; Lee, Y.;
Hwang, J.-W.; Hong, S.; Kim, C.; Park, T. G.; Lee, H.; Hong, S. H.
Adv. Mater. 2011, 23, 1971−1975.
Kappes, M. M.; Mayor, M. Chem. Commun. 2011, 47, 7428−7430.
(15) Utsa, H.; Facchetti, A.; Marks, T. J. Org. Lett. 2008, 10, 1385−
1388.
(16) Zhang, M.; Tsao, H. N.; Pisula, W.; Yang, C.; Mishra, A. K.;
Mullen, K. J. Am. Chem. Soc. 2007, 129, 3472−3473.
̈
(17) Zhang, Z.-B.; Fujiki, M.; Tang, H.-Z.; Motonaga, M.; Torimitsu,
K. Macromolecules 2002, 35, 1988−1990.
(18) Fomina, N.; Hogen-Esch, T. E. Macromolecules 2008, 41, 3765−
(2) Joselevich, E.; Dai, H.; Lui, J.; Hata, K.; Windle, A. H. In Carbon
Nanotubes; Jorio, A., Dresselhaus, G., Dresselhaus, M. S., Eds.; Topics
in Applied Physics, Vol. 111; Springer: Berlin, 2008; pp 101−104.
(3) (a) Martel, R. ACS Nano 2008, 2, 2195−2199. (b) Adam, E.;
Aguirre, C. M.; Marty, L.; St-Antoine, B. C.; Meunier, F.; Desjardins,
3768.
(19) During this study, the following publication appeared where
ketone 8 was reduced using hydrazine monohydrate: Song, Y.; Xu,
W.; Zhu, D. Tetrahedron Lett. 2010, 51, 4894−4897.
(20) Khurana, J. M.; Bansal, K.; Pandey, R. R. Monatsh. Chem. 2003,
134, 1365−1371.
(21) Ostrauskaite, J.; Strohriegl, P. Macromol. Chem. Phys. 2003, 204,
1713−1718.
(22) Weisman, R. B.; Bachilo, S. M. Nano Lett. 2003, 3, 1235−1238.
Lebedkin, S.; Hennrich, F.; Kiowski, O.; Kappes, M. M. Phys. Rev. B
2008, 77, 165429.
́
P.; Menard, D.; Martel, R. Nano Lett. 2008, 8, 2351−2355. (c) Cao,
Q.; Kim, H. S.; Pimparkar, N.; Kulkarni, J. P.; Wang, C.; Shim, M.;
Roy, K.; Alam, M. A.; Rogers, J. A. Nature 2008, 454, 495−500.
(d) Xia, F.; Steiner, M.; Lin, Y. M.; Avouris, P. Nature Nanotechnol.
2008, 3, 609−613. (e) Holt, J. M.; Ferguson, A. J.; Kopidakis, N.;
Larsen, B. A.; Bult, J.; Rumbles, G.; Blackburn, J. L. Nano Lett. 2010,
10, 4627−4633.
(23) Zhang, Z.; Che, Y.; Smaldone, R. A.; Xu, M.; Bunes, B. R.;
Moore, J. S.; Zang, L. J. Am. Chem. Soc. 2010, 132, 14113−14117.
(24) Berton, N.; Lemasson, F.; Hennrich, F.; Kappes, M. M.; Mayor,
M. Manuscript submitted.
(4) (a) Hersam, M. C. Nature Nanotechnol. 2008, 3, 387−394.
(b) Komatsu, N.; Wang, F. Materials 2010, 3, 3818−3844.
(5) (a) Tu, X.; Manohar, S.; Jagota, A.; Zheng, M. Nature 2009, 460,
250−253. (b) Zheng, M.; Semke, E. D. J. Am. Chem. Soc. 2007, 129,
6084−6085.
(6) Krupke, R.; Hennrich, F.; von Loehneysen, H.; Kappes, M. M.
Science 2003, 301, 344−347.
(7) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam, M.
C. Nature Nanotechnol. 2006, 1, 60−65.
(8) Gosh, S.; Bachilo, S. M.; Weisman, R. B. Nature Nanotechnol.
2010, 5, 443−450.
(9) (a) Peng, X.; Komatsu, N.; Bhattacharya, S.; Shimawaki, T.;
Aonuma, S.; Kimura, T.; Asuka, A. Nature Nanotechnol. 2007, 2, 361−
365. (b) Wang, F.; Matsuda, K.; Mustafizur Rahman, A. F. M.; Peng,
X.; Kimura, T.; Komatsu, N. J. Am. Chem. Soc. 2010, 132, 10876−
10881. (c) Rahman, A. S. F. M.; Wang, F.; Matsuda, K.; Kimura, T.;
Komatsu, N. Chem. Sci. 2011, 2, 862−867. (d) Marquis, R.; Greco, C.;
722
dx.doi.org/10.1021/ma201890g | Macromolecules 2012, 45, 713−722