Toward a rational design of poly(2,7-carbazole) derivatives for solar cells

Article abstract of DOI:10.1021/ja0771989

On the basis of theoretical models and calculations, several alternating polymeric structures have been investigated to develop optimized poly(2,7-carbazole) derivatives for solar cell applications. Selected low band gap alternating copolymers have been obtained via a Suzuki coupling reaction. A good correlation between DFT theoretical calculations performed on model compounds and the experimental HOMO, LUMO, and band gap energies of the corresponding polymers has been obtained. This study reveals that the alternating copolymer HOMO energy level is mainly fixed by the carbazole moiety, whereas the LUMO energy level is mainly related to the nature of the electron-withdrawing comonomer. However, solar cell performances are not solely driven by the energy levels of the materials. Clearly, the molecular weight and the overall organization of the polymers are other important key parameters to consider when developing new polymers for solar cells. Preliminary measurements have revealed hole mobilities of about 1 × 10^-3 cm2·V^-1·s^-1 and a power conversion efficiency (PCE) up to 3.6%. Further improvements are anticipated through a rational design of new symmetric low band gap poly(2,7-carbazole) derivatives.

Full text of DOI:10.1021/ja0771989

Published on Web 12/21/2007
Toward a Rational Design of Poly(2,7-Carbazole) Derivatives
for Solar Cells
Nicolas Blouin,† Alexandre Michaud,† David Gendron,† Salem Wakim,†,‡
Emily Blair,† Rodica Neagu-Plesu,† Michel Belleteˆte,§ Gilles Durocher,§
Ye Tao,‡ and Mario Leclerc*,†
Canada Research Chair on ElectroactiVe and PhotoactiVe Polymers, De´partement de Chimie,
UniVersite´ LaVal, Quebec City, Quebec, G1K 7P4, Canada, Institute of Microstructural
Sciences, National Research Council of Canada, Ottawa, Ontario, K1A 0R6, Canada, and
Laboratoire de Photophysique Mole´culaire, De´partement de Chimie, UniVersite´ de Montreal,
C.P. 6128, Succ. Centre-Ville, Montreal, Quebec, H3C 3J7, Canada
Received September 19, 2007; E-mail: mario.leclerc@chm.ulaval.ca
Abstract: On the basis of theoretical models and calculations, several alternating polymeric structures
have been investigated to develop optimized poly(2,7-carbazole) derivatives for solar cell applications.
Selected low band gap alternating copolymers have been obtained via a Suzuki coupling reaction. A good
correlation between DFT theoretical calculations performed on model compounds and the experimental
HOMO, LUMO, and band gap energies of the corresponding polymers has been obtained. This study
reveals that the alternating copolymer HOMO energy level is mainly fixed by the carbazole moiety, whereas
the LUMO energy level is mainly related to the nature of the electron-withdrawing comonomer. However,
solar cell performances are not solely driven by the energy levels of the materials. Clearly, the molecular
weight and the overall organization of the polymers are other important key parameters to consider when
developing new polymers for solar cells. Preliminary measurements have revealed hole mobilities of about
1 × 10-3 cm2‚V-1‚s-1 and a power conversion efficiency (PCE) up to 3.6%. Further improvements are
anticipated through a rational design of new symmetric low band gap poly(2,7-carbazole) derivatives.
1. Introduction
butyric acid methyl ester (PCBM), as the electron acceptor. To
further improve the device performances, one can develop new
device architectures,10,11 synthesize new polymers12,13 and new
electron acceptors,5,14-16 or work on both ends.
In the past few years, several groups of chemists proposed
new polymer structures as alternatives to P3HT and MDMO-
PPV since the performances of these two polymers are somehow
limited by their relatively large band gap.17,18 For many years,
internal charge transfer (ICT) from an electron-rich unit to an
electron deficient moiety has been extensively used to obtain
low band gap conjugated polymers.19-23 Using an ICT strategy,
Solar cells are one key technology for solving world energy
needs. The development of new materials such as the semicon-
ducting conjugated polymers as active components in bulk
heterojunction1-4 (BHJ) photovoltaic devices could help to
significantly reduce the fabrication cost of such devices. For
these purposes, two polymeric materials have been extensively
studied in the past decade: poly[2-methoxy-5-(3′,7′-dimethy-
loctyloxy)-p-phenylenevinylene] (MDMO-PPV)5,6 and regio-
regular poly(3-hexylthiophene) (P3HT).7-9 The utilization of
these polymeric materials has led to power conversion efficien-
cies between 3.0 and 5.0% when mixed with [6,6]-Phenyl C61
(10) Hadipour, A.; de Boer, B.; Wildeman, J.; Kooistra, F. B.; Hummelen, J.
C.; Turbiez, M. G. R.; Wienk, M. M.; Janssen, R. A. J.; Blom, P. W. M.
AdV. Funct. Mater. 2006, 16, 1897-1903.
(11) Kim, J. Y.; Lee, K.; Coates, N. E.; Moses, D.; Nguyen, T.-Q.; Dante, M.;
Heeger, A. J. Science 2007, 317, 222-225.
† Universite´ Laval.
‡ National Research Council of Canada.
§ Universite´ de Montreal.
(1) Brabec, C. J.; Sariciftci, N. S.; Hummelen, J. C. AdV. Funct. Mater. 2001,
(12) Scharber, M. C.; Mu¨hlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.;
Heeger, A. J.; Brabec, C. J. AdV. Mater. 2006, 18, 789-794.
(13) Bundgaard, E.; Krebs, F. C. Sol. Energy Mater. 2007, 91, 954-985.
(14) Brabec, C. J.; Cravino, A.; Meissner, D.; Sariciftci, N. S.; Fromherz, T.;
Rispens, M. T.; Sanchez, L.; Hummelen, J. C. AdV. Funct. Mater. 2001,
11, 374-380.
11, 15-26.
(2) Hoppe, H.; Sariciftci, N. S. J. Mater. Res. 2004, 19, 1924-1945.
(3) Coakley, K. M.; McGehee, M. D. Chem. Mater. 2004, 16, 4533-4542.
(4) Gunes, S.; Neugebauer, H.; Sariciftci, N. S. Chem. ReV. 2007, 107, 1324-
1338.
(5) Wienk, M. M.; Wiljan, J. M. K.; Verhees, J. H.; Knol, J.; Hummelen, J.
C.; Hal, P. A. v.; Janssen, R. A. J. Angew. Chem., Int. Ed. 2003, 42, 3371-
3375.
(15) Backer, S. A.; Sivula, K.; Kavulak, D. F.; Frechet, J. M. J. Chem. Mater.
2007, 19, 2927-2929.
(16) Kooistra, F. B.; Knol, J.; Kastenberg, F.; Popescu, L. M.; Verhees, W. J.
H.; Kroon, J. M.; Hummelen, J. C. Org. Lett. 2007, 9, 551-554.
(17) Alstrup, J.; Norrman, K.; Jorgensen, M.; Krebs, F. C. Sol. Energy Mater.
2006, 90, 2777-2792.
(6) Shaheen, S. E.; Brabec, C. J.; Sariciftci, N. S.; Padinger, F.; Fromherz, T.;
Hummelen, J. C. Appl. Phys. Lett. 2001, 78, 841-843.
(7) Ma, W.; Yang, C.; Gong, X.; Lee, K.; Heeger, A. J. AdV. Funct. Mater.
2005, 15, 1617-1622.
(18) Chabinyc, M. L.; Street, R. A.; Northrup, J. E. Appl. Phys. Lett. 2007, 90,
123508-3.
(8) Reyes-Reyes, M.; Kim, K.; Carroll, D. L. Appl. Phys. Lett. 2005, 87,
083506-3.
(19) Havinga, E. E.; ten Hoeve, W.; Wynberg, H. Synth. Met. 1993, 55, 299-
(9) Li, G.; Shrotriya, V.; Huang, J.; Yao, Y.; Moriarty, T.; Emery, K.; Yang,
Y. Nat. Mater. 2005, 4, 864-868.
306.
(20) Zhang, Q. T.; Tour, J. M. J. Am. Chem. Soc. 1998, 120, 5355-5362.
9
732
J. AM. CHEM. SOC. 2008, 130, 732-742
10.1021/ja0771989 CCC: $40.75 © 2008 American Chemical Society

Design of Poly(2,7-Carbazole) Derivatives
A R T I C L E S
new polymers have been developed to better harvest the solar
spectrum, especially in the 1.4-1.9 eV region. Several low band
gap polythiophene derivatives have been reported by Krebs24-26
and Reynolds,27,28 but until now, relatively low performances
in solar cells have been obtained. Polyfluorene derivatives29-32
show promising features with power conversion efficiencies
(PCE) between 2.0 and 4.2%. However, these polymers show
relatively low carrier mobility, limiting the device performances.
Recently, benzothiadiazole and cyclopentadithiophene copoly-
mers showed very interesting PCE values (3.2-5.5%)33,34 and
be below the air oxidation threshold (ca. -5.27 eV or 0.57 V
vs SCE).45 Furthermore, this relatively low value assures a
relatively high open circuit potential (VOC) in the final device.
Second, the LUMO energy level of the polymer must be
positioned above the LUMO energy level of the acceptor (i.e.,
[6,6]-Phenyl C61 butyric acid methyl ester (PCBM)) by at least
0.2-0.3 eV to ensure efficient electron transfer from the polymer
to the acceptor.46,47 Therefore, the ideal donor LUMO energy
level should be between -3.7 and -4.0 eV. Finally, the optimal
band gap, considering the solar emission spectrum and the open
circuit potential of the resulting solar cell, should range between
1.2 and 1.9 eV. Therefore, the ideal polymer HOMO energy
level should range between -5.2 and -5.8 eV. Taking into
account the LUMO energy level of PCDTBT (-3.60 eV), one
needs to reduce this parameter to further increase the perfor-
mance of a polymeric cell while keeping the HOMO energy
level within the same energy value (i.e., -5.45 eV).
On the basis of these initial data and theoretical models, we
investigated several alternating polymeric structures to develop
the most suitable poly(2,7-carbazole) derivatives for BHJ solar
cell devices.12,27,44 To estimate the polycarbazole performances
using those models, we performed quantum calculations on the
repeat unit of the planned polymers to determine both the
HOMO and LUMO energy levels. These values were then
compared to the experimental values of the polymers. These
new polymers were tested in solar cell devices, and their
performances were analyzed in terms of polymer organization,
molecular weight, charge carrier mobility, and LUMO energy
level.
high carrier mobility (10-2-10-1 cm2‚V-1‚s-1 33,35
demonstrat-
)
ing that one can synthesize ICT polymers having low band gap
and high carrier mobility.
Poly(2,7-carbazole) derivatives are other excellent potential
candidates for BHJ solar cells. Indeed, as observed with poly-
(2,7-fluorene)s,36,37 the physical properties of poly(2,7-carba-
zole)s38 can be easily modulated. As an electron-rich mole-
cule,38-40 the carbazole unit is perfect for the development of
ICT polymers. Moreover, poly(N-vinylcarbazole) (PVK) is
among the best photoconductive polymeric materials.39 Initial
studies by K. Mu¨llen41 and our group42 have produced poly-
carbazole materials that exhibit relatively low efficiency in solar
cells (0.6-0.8%). In general, those polymers were poorly soluble
and showed a lack of organization. However, we recently
reported a new polycarbazole derivative (PCDTBT)43 bearing
a secondary alkyl side chain on the nitrogen atom of the
carbazole unit that shows high solubility and some organization,
resulting in a very good PCE (3.6%). It was concluded that, by
improving the electronic properties of the polymeric materials,
much higher efficiencies could be reached.
2. Results and Discussion
Recently, several models have been proposed to estimate the
polymer performance in BHJ solar cells.12,27,44 First, the polymer
must be air-stable; therefore, the HOMO energy level needs to
2.1. Theoretical Calculations. Predicting the behavior of both
the HOMO and LUMO energy levels for new polymers is
crucial to make a rational design of optimized BHJ solar cells.
Recently, semiempirical calculations (MNDO) have been suc-
cessfully applied on several similar low band gap fluorene
derivatives.48 In the present work, we have estimated the HOMO
and LUMO energy levels of the repetitive units of the corre-
sponding alternating copolymers by using the density functional
theory (DFT) as approximated by the B3LYP functional and
employing the 6-31G* basis set. The computational methodol-
ogy is described in the Supporting Information. DFT/B3LYP/
6-31G* has been found to be an accurate formalism for
calculating the structural and optical properties of many
molecular systems.49-55
(21) Mullekom, H. A. M. v.; Vekemans, J. A. J. M.; Havinga, E. E.; Meijer, E.
W. Mater. Sci. Eng., R. 2001, 32, 1-40.
(22) Ajayaghosh, A. Chem. Soc. ReV. 2003, 32, 181-191.
(23) Roncali, J. Macromol. Rapid Commun. 2007, 28, 1761-1775.
(24) Petersen, M. H.; Hagemann, O.; Nielsen, K. T.; Jorgensen, M.; Krebs, F.
C. Sol. Energy Mater. 2007, 91, 996-1009.
(25) Bundgaard, E.; Krebs, F. C. Sol. Energy Mater. 2007, 91, 1019-1025.
(26) Bundgaard, E.; Krebs, F. C. Macromolecules 2006, 39, 2823-2831.
(27) Thompson, B. C.; Kim, Y. G.; Reynolds, J. R. Macromolecules 2005, 38,
5359-5362.
(28) Thompson, B. C.; Kim, Y. G.; McCarley, T. D.; Reynolds, J. R. J. Am.
Chem. Soc. 2006, 128, 12714-12725.
(29) Wang, F.; Luo, J.; Yang, K.; Chen, J.; Huang, F.; Cao, Y. Macromolecules
2005, 38, 2253-2260.
(30) Zhang, F.; Mammo, W.; Andersson, L. M.; Admassie, S.; Andersson, M.
R.; Ingana¨s, O. AdV. Mater. 2006, 18, 2169-2173.
(31) Andersson, L. M.; Fengling, Z.; Olle, I. Appl. Phys. Lett. 2007, 91,
071108-3.
(32) Slooff, L. H.; Veenstra, S. C.; Kroon, J. M.; Moet, D. J. D.; Sweelssen, J.;
Koetse, M. M. Appl. Phys. Lett. 2007, 90, 143506-3.
(33) Mu¨hlbacher, D.; Scharber, M.; Morana, M.; Zhu, Z.; Waller, D.; Gaudiana,
R.; Brabec, C. AdV. Mater. 2006, 18, 2884-2889.
(45) de Leeuw, D. M.; Simenon, M. M. J.; Brown, A. R.; Einhard, R. E. F.
Synth. Met. 1997, 87, 53-59.
(46) Choulis, S. A.; Nelson, J.; Kim, Y.; Poplavskyy, D.; Kreouzis, T.; Durrant,
J. R.; Bradley, D. D. C. Appl. Phys. Lett. 2003, 83, 3812-3814.
(47) Mihailetchi, V. D.; Duren, J. K. J. v.; Blom, P. W. M.; Hummelen, J. C.;
Janssen, R. A. J.; Kroon, J. M.; Rispens, M. T.; Verhees, W. J. H.; Wienk,
M. M. AdV. Funct. Mater. 2003, 13, 43-46.
(34) Peet, J.; Kim, J. Y.; Coates, N. E.; Ma, W. L.; Moses, D.; Heeger, A. J.;
Bazan, G. C. Nat. Mater. 2007, 6, 497-500.
(35) Zhang, M.; Tsao, H. N.; Pisula, W.; Yang, C.; Mishra, A. K.; Mullen, K.
J. Am. Chem. Soc. 2007, 129, 3472-3473.
(48) Admassie, S.; Inganas, O.; Mammo, W.; Perzon, E.; Andersson, M. R.
Synth. Met. 2006, 156, 614-623.
(36) Leclerc, M. J. Polym. Sci., Part A: Polym. Chem. 2001, 39, 2867-2873.
(37) Akcelrud, L. Prog. Polym. Sci. 2003, 28, 875-962.
(38) Morin, J.-F.; Leclerc, M.; Ade`s, D.; Siove, A. Macromol. Rapid Commun.
2005, 26, 761-778.
(49) Tretiak, S.; Mukamel, S. Chem. ReV. 2002, 102, 3171-3212.
(50) Belleteˆte, M.; Durocher, G.; Hamel, S.; Coˆte´, M.; Wakim, S.; Leclerc, M.
J. Chem. Phys. 2005, 122, 104303-9.
(39) Grazulevicius, J. V.; Strohriegl, P.; Pielichowski, J.; Pielichowski, K. Prog.
Polym. Sci. 2003, 28, 1297-1353.
(51) Belletete, M.; Blouin, N.; Boudreault, P. L. T.; Leclerc, M.; Durocher, G.
J. Phys. Chem. A 2006, 110, 13696-13704.
(40) Drolet, N.; Morin, J. F.; Leclerc, N.; Wakim, S.; Tao, Y.; Leclerc, M. AdV.
Funct. Mater. 2005, 15, 1671-1682.
(52) Perrier, A.; Maurel, F.; Aubard, J. J. Photochem. Photobiol. A 2007, 189,
167-176.
(41) Li, J.; Dierschke, F.; Wu, J.; Grimsdale, A. C.; Mu¨llen, K. J. Mater. Chem.
2006, 16, 96-100.
(53) Kurashige, Y.; Nakajima, T.; Kurashige, S.; Hirao, K.; Nishikitani, Y. J.
Phys. Chem. A 2007, 111, 5544-5548.
(42) Leclerc, N.; Michaud, A.; Sirois, K.; Morin, J. F.; Leclerc, M. AdV. Funct.
Mater. 2006, 16, 1694-1704.
(54) Osuna, R. M.; Ortiz, R. P.; Okamoto, T.; Suzuki, Y.; Yamaguchi, S.;
Hernandez, V.; Lopez, Navarrete, J. T. J. Phys. Chem. B 2007, 111, 7488-
7496.
(43) Blouin, N.; Michaud, A.; Leclerc, M. AdV. Mater. 2007, 19, 2295-2300.
(44) Koster, L. J. A.; Mihailetchi, V. D.; Blom, P. W. M. Appl. Phys. Lett.
2006, 88, 093511-3.
(55) Liu, T.; Gao, J.-S.; Xia, B.-H.; Zhou, X.; Zhang, H.-X. Polymer 2007, 48,
502-511.
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 733

A R T I C L E S
Blouin et al.
Figure 1. Theoretical energy levels for some carbazole derivatives. Insert: repeating unit of the correlated alternating copolymer.
As shown in Figure 1, we first investigated several potential
polycarbazole analogues to the low band gap polyfluorenes
reported by Andersson/Ingana¨s.30,56-58 The TQ comonomer
should lead to a relatively low band gap polymer, but as found
with polyfluorene derivatives,57,58 the resulting polycarbazole
should have a too low LUMO energy level, assuming that
PCBM is used as an electron acceptor. Other candidates (TP
and TP-Ph) raise also some concerns. For these two copoly-
mers, the calculations revealed an important increase of the
HOMO energy value without any major decrease of the LUMO
energy value compared to the other candidates. This phenom-
enon could result in lower Voc and, more importantly, to air-
unstable polymers. In contrast, pyridine-based candidates (PP,
PT, and PX) revealed an interesting trend; the pyridine moiety
decreases both the HOMO and LUMO energy levels. The
resulting polymers should therefore exhibit higher open circuit
potentials while being air-stable. Replacing the sulfur atom (BT,
PT) by an oxygen atom (BX, PX) has a similar effect. Based
on these calculations, it was decided to synthesize the BX, PP,
PT, PX copolymers. For comparison purposes, we also synthe-
sized the Qx polycarcarbazole derivative and the previously
reported BT derivative.
2.2. Synthesis. 2.2.1. Monomer Synthesis. The synthetic
route to the chosen five new comonomers is depicted in Scheme
1. First, this synthetic strategy required an electron-deficient
dibromoheteroarene as the elementary block, as already pro-
posed by Yamashita and co-workers for similar compounds.59
Starting from the 2,3-diaminopyridine (1), compound 2 was
obtained from bromine in hydrobromic acid according to the
work of Lee and Yamamoto.60 An alternative purification
process was used to eliminate any remaining mineral residues.
Without this additional purification, the ring closing reaction
leading to compounds 3 and 4 is significantly affected.
Compound 3 was prepared from a double imide formation using
aqueous glyoxal in butanol. The conversion of compound 2 to
compound 4 was only possible with thionyl chloride (75 equiv),
following a strategy similar to that reported by Thomas and
Sliwa.61 Surprisingly, the bromine atom at the 4-position is
converted into a chlorine atom, as demonstrated from both mass
spectra and monocrystal X-ray analyses (Figure 1S). Reaction
concentration, addition of base (pyridine or triethylamine), or
substitution of thionyl chloride for thionylaniline or thionyl
bromide did not improve the reaction yield and generally
resulted in undesired products. The only viable way found to
obtain the 4,7-dibromo-2,1,3-benzoxadiazole (6) was from
melted 2,1,3-benzoxadiazole (5) mixed with iron as catalyst and
bromine. The product was subsequently treated with a K2CO3
saturated aqueous solution according to already reported pro-
cedures.62,63 Under milder conditions, nearly pure starting
compound is recovered. To ensure that the bromine atoms insert
at the 4,7-positions, monocrystals were grown from ethanol and
subsequently analyzed by X-ray diffraction. Figure 2S (Sup-
porting Information) confirms the structure of the desired
compound, demonstrating a perfect selectivity for the 4,7-
positions.
From these electron-deficient dibromoheteroarene building
blocks, the comonomers 12 to 15 were synthesized. Starting
from the dibromoheteroarene compounds 3, 4, 6, or 764 and
2-tributylstannylthiophene, a Stille coupling reaction65 in dim-
ethylformamide (DMF) afforded precursors 8 to 11. These
precursors were treated with n-bromosuccinimide (NBS) in 1,2-
dichlorobenzene (ODCB) at 55 °C, according to a methodology
previously reported for 4,7-di(2′-bromothien-5′-yl)-2,1,3-ben-
(56) Zhang, F.; Perzon, E.; Wang, X.; Mammo, W.; Andersson, M. R.; Ingana¨s,
O. AdV. Funct. Mater. 2005, 15, 745-750.
(57) Wang, X.; Perzon, E.; Oswald, F.; Langa, F.; Admassie, S.; Andersson,
M. R.; Ingana¨s, O. AdV. Funct. Mater. 2005, 15, 1665-1670.
(58) Wang, X.; Perzon, E.; Mammo, W.; Oswald, F.; Admassie, S.; Persson,
N.-K.; Langa, F.; Andersson, M. R.; Inganas, O. Thin Solid Films 2006,
511-512, 576-580.
(61) Thomas, A.; Sliwa, W. Heterocycles 1983, 20, 1043-1048.
(62) Pilgram, K.; Zupan, M. J. Heterocycl. Chem. 1974, 11, 813-814.
(63) Hammick, D. L.; Edwakdes, W. A. M.; Steiner, E. R. J. Chem. Soc. 1931,
3308-3313.
(64) Edelmann, M. J.; Raimundo, J.-M.; Utesch, N. F.; Diederich, F.; Boudon,
C.; Gisselbrecht, J.-P.; Gross, M. HelV. Chim. Acta 2002, 85, 2195-2213.
(65) Stille, J. K. Angew. Chem., Int. Ed. 1986, 25, 508-524.
(59) Kitamura, C.; Tanaka, S.; Yamashita, Y. Chem. Mater. 1996, 8, 570-578.
(60) Lee, B. L.; Yamamoto, T. Macromolecules 1999, 32, 1375-1382.
9
734 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008

Design of Poly(2,7-Carbazole) Derivatives
A R T I C L E S
Scheme 1. Synthesis of the Comonomers
zothiadiazole.66 The resulting comonomers 12 to 15 are then
easily recovered. In contrast, when treated under more classical
conditions (i.e., chloroform/acetic acid, 1:1), the resulting
products are not suitable for polymerization reactions. The last
comonomer (18) is prepared from an alternative synthetic
pathway originally proposed by Gorohmaru et al.67 The ethyl
4,7-bis(5-bromothien-2-yl)[1,2,5]oxadiazolo[3,4-c]pyridine-6-
carboxylate67 (16) is hydrolyzed with aqueous NaOH and
neutralized with aqueous hydrochloride to obtain the carboxylic
acid 17. This compound was decarboxylated at high temperature
in 1,2,4-trichlorobenzene to afford comonomer 18.
reaction as that already reported for PCDTBT43 was chosen to
obtain these new copolymers. Indeed, 2,7-bis(4′,4′,5′,5′-tetram-
ethyl-1′,3′,2′-dioxaborolan-2′-yl)-N-9′′-heptadecanylcarba-
zole43 (20) and the desired comonomers were copolymerized
through a Suzuki coupling reaction68 and produced three
symmetric copolymers (PCDTQx, PCDTBT, and PCDTBX) and
three asymmetric copolymers (PCDTPP, PCDTPT, and PCDT-
PX). PCDTBT and PCDTBX have higher molecular weights
than the other polymers even if each monomer showed similar
purities and was polymerized using the same conditions (Table
1). Other Suzuki coupling polymerization conditions were tried,
especially for PCDTPT and PCDTPX, to improve the molecular
weight but, up to now, without any success. We attribute these
results to the possible complexation69-71 of the palladium
catalyst by monomers 14 and 18 or the resulting polymers,
slowing down or stopping the polymerization reaction.
2.2.2. Polymer Synthesis. The five new polymers were
synthesized according to Scheme 2. The same polymerization
The parent symmetric PCDTQx and asymmetric PCDTPP
show similar number-average molecular weights (9.0 KDa and
11 KDa, respectively). Therefore, comparisons of the polymer
properties and device performances are possible without taking
into account the influence of the molecular weight, as generally
observed for P3HT.72,73 In general, the asymmetric polymers
are more soluble in common solvents (i.e., chloroform and
(66) Zhang, C. U.S. Patent 20040229925, 2004.
(67) Gorohmaru, H.; Thiemann, T.; Sawada, T.; Takahashi, K.; Nishi-i, K.; Ochi,
N.; Kosugi, Y.; Mataka, S. Heterocycles 2002, 56, 421-431.
(68) Towns, C.; Wallace, P.; Allen, I.; Pounds, T.; Murtagh, L. U.S. Patent
20050014926, 2005.
(69) Tsuboyama, A.; Iwawaki, H.; Furugori, M.; Mukaide, T.; Kamatani, J.;
Igawa, S.; Moriyama, T.; Miura, S.; Takiguchi, T.; Okada, S.; Hoshino,
M.; Ueno, K. J. Am. Chem. Soc. 2003, 125, 12971-12979.
(70) Liu, P.; Huang, M.; Pan, W.; Zhang, Y.; Hu, J.; Deng, W. J. Lumin. 2006,
121, 109-112.
(71) Kokil, A.; Yao, P.; Weder, C. Macromolecules 2005, 38, 3800-3807.
(72) Schilinsky, P.; Asawapirom, U.; Scherf, U.; Biele, M.; Brabec, C. J. Chem.
Mater. 2005, 17, 2175-2180.
Figure 2. UV-vis spectra at 135 °C in 1,2,4-trichlorobenzene for the six
polycarbazoles derivatives.
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 735

A R T I C L E S
Blouin et al.
Scheme 2. Structure and Synthesis of the 2,7-Carbazole-Based Alternating Copolymers
1
Table 1. Number-Average Molecular Weight (Mn), Polydispersity
2.3.2. H NMR Analyses of the Polymers. The symmetric
(PCDTQx, PCDTBT, PCDTBX) and asymmetric (PCDTPP,
PCDTPT, PCDTPX) natures of the copolymers were clearly
revealed by 1H NMR analyses. For such measurements, a high
temperature (in ODCB at 130 °C) was necessary to suppress
both the polymer aggregation and alkyl side chain atropisom-
erism configuration observed at room temperature.43 Differences
between the two structures are particularly evident in the
aromatic region (see Figure 3). For instance, the signal of
proton-2 on the PCDTQx is split into two proton signals (2,
12) for PCDTPP. Moreover, the two other proton signals (1
and 3) for PCDTQx result in four different signals (1, 3, 11,
and 13) for PCDTPP although the splitting is less evident.
Overall, there are three proton signals associated with the
carbazole for the PCDTQx, whereas there are six peaks for the
PCDTPP. The asymmetry in the polymer backbone is therefore
attributable to the additional nitrogen atom present on the
pyridine core and its relative position to both sides of the
carbazole unit. Similar but less evident observations have been
made for the other copolymers.
2.3.3. Thermal Properties. The thermal stability of these
polymers is also a very important parameter for the performance
of the BHJ devices. All present copolymers are thermally stable
up to 340 °C and show a glass transition temperature (Tg) over
100 °C (Table 1), except for PCDTPX whose Tg was not
detected by DSC. Clearly, the oxygen-containing polymers
(PCDTBX and PCDTPX) show lower thermal stability than the
sulfur-containing polymers (PCDTBT and PCDTPT). Interest-
ingly, PCDTBT melts into a liquid crystalline state around
268 °C and recrystallizes around 259 °C (Figure 4). Similar
transitions have been reported for poly(fluorene-alt-dithienyl-
benzothiadiazole)s (175-180 °C).74 The low supercooling and
the very low energy involved (between 0.34 and 0.39 J/g in
both heating and cooling) support a mesophase nature. This
assumption is also supported by observations made with
polarized optical microscopy (POM) since the melting of
PCDTBT and PCDTQx leads to birefringent liquids. The third
symmetric polymer, PCDTBX, shows some birefringence but
degrades before melting. No birefringence was noticed for the
asymmetric polymers, and this fact confirms a higher ordering
capability for the symmetric structures. The same alkyl side
chain has a similar impact over the polymer solubility and
organization of other N-substituted heterocycles,75 demonstrating
the importance to better understand the role also played by the
alkyl side chain.
Index (PDI), Glass Transition Temperature (Tg), Degradation
Temperature (Td), and Optical and Electrochemical Properties of
Polycarbazole Derivatives
a
b
elec
Mn
Tg,
Td,
EHOMO
eV
,
ELUMO
eV
,
Eg
eV
,
Egopt thin
filmc, eV
polymer
kDa
PDI
°
C
°C
PCDTQx
PCDTPP 11
PCDTBT 36
PCDTPT
PCDTBX 26
PCDTPX 4.5 1.33
9.0 1.53 116 440 -5.46 -3.42 2.04
2.02
1.89
1.88
1.75
1.87
1.67
1.70 133 420 -5.52 -3.67 1.85
1.54 130 470 -5.45 -3.60 1.85
4.0 1.50 111 410 -5.53 -3.80 1.73
1.97 145 380 -5.47 -3.65 1.82
340 -5.55 -3.93 1.62
-
a Ionization potential (EHOMO; HOMO: highest occupied molecular
orbital) measured from the cathodic onset. The SCE level used for the
HOMO calculation was -4.7 eV.27 b Electron affinity (ELUMO; LUMO:
lowest unoccupied molecular orbital) measured from the anodic onset. The
SCE level used for the LUMO calculation was -4.7 eV.27 c Measurements
performed on spin-coated films from the onset of the absorption band.
tetrahydrofuran) than their symmetric counterparts. Surprisingly,
PCDTBT is soluble in chloroform, chlorobenzene (CB), or
orthodichlorobenzene (OBCB), whereas PCDTBX is far less
soluble in these solvents. This explains the lower soluble content
recovered after Soxhlet extractions. It was also noticed that 16
h of polymerization for the PCDTBT were more efficient than
the 72 h previously reported,43 leading to an improved polym-
erization yield (soluble fraction in chloroform) and a less
polydisperse polymer. Similar attempts to improve PCDTBX
and PCDTQx polymerization reactions did not result in the same
improvements.
2.3. Polymer Characterization. 2.3.1. UV-vis Absorption.
The characterization of the UV-visible absorption features of
our polymers is essential for photovoltaic applications and has
been performed both in solution (in 1,2,4-trichlorobenzene
(TCB) at 135 °C, see Figure 2) and in the solid state. For
solution measurements, such a high temperature was required
since a significant bathochromic shift (ca. 15-25 nm) is
observed at 25 °C (Figures 3S to 8S, Supporting Information),
suggesting, in agreement with SEC data, the presence of
aggregation.
The solid-state UV-vis spectra of the polymers show a
bathochromic shift (ca. 30-40 nm) compared to the solution
spectra obtained in TCB at 135 °C (Figures 3S to 8S, Supporting
Information). The red shift observed from solution to the solid
state suggests a higher structural organization in the solid state.
As expected, the optical band gap (Table 1) varies as a function
of the electron-withdrawing properties of the comonomers.
Therefore, the optical band gap (Table 1) ranges from 2.02 eV
(614 nm) for PCDTQx to 1.67 eV (743 nm) for PCDTPX.
(74) Ingana¨s, O.; Svensson, M.; Zhang, F.; Gadisa, A.; Persson, N. K.; Wang,
X.; Andersson, M. R. Appl. Phys. A 2004, 79, 31-35.
(75) Koeckelberghs, G.; DeCremer, L.; Persoons, A.; Verbiest, T. Macromol-
ecules 2007, 40, 4173-4181.
(73) Hiorns, R. C.; de Bettignies, R.; Leroy, J.; Bailly, S.; Firon, M.; Sentein,
C.; Khoukh, A.; Preud’homme, H.; Dagron-Lartigau, C. AdV. Funct. Mater.
2006, 16, 2263-2273.
9
736 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008

Design of Poly(2,7-Carbazole) Derivatives
A R T I C L E S
1
Figure 3. Aromatic region of the H NMR spectra of PCDTQX and PCDTPP in 1,2-dichlorobenzene-d4 at 130 °C.
Figure 5. X-ray diffraction patterns of powdery polycarbazole derivatives.
Figure 4. Differential scanning calorimetry of PCDTBT at 5, 10, and
20 °C/min. Insert: micrograph under polarized light.
mers have no effect on the d1 distance. In contrast, the nature
of the comonomers has an important effect on the d2 distance.
For instance, PCDTBT exhibits the shortest π stacking distance
(d2 ≈ 4.4 Å). Such a phenomenon could be related to the highly
favorable dipole-dipole and π-π interactions between ben-
zothiadiazole units and conjugated polymer chains.79-81 These
results suggest a higher structural organization in the solid state
for the symmetric polymers (PCDTQx, PCDTBT, and PCDT-
BX) compared to the asymmetric polymers (PCDTPP and
PCDTPT) with an exception for the PCDTPX.
2.3.5. HOMO-LUMO Energy Levels. Copolymer energy
levels depend on several factors and have been characterized
from electrochemical analyses. The anodic peak potential (Epa)
for the three symmetric polymers (PCDTQx, PCDTBT, PCDT-
BX) is, within the experimental error, the same (ca. 1.01 V), as
2.3.4. X-ray Analyses. To confirm the structural order
observed by POM and DSC measurements, X-ray diffraction
analyses on powdery polymers were performed (Figure 5). All
six polymers show a peak at d1 ≈ 21-22 Å, but only a very
weak diffraction is observed for PCDTPP and PCDTPT.
Contrastingly, the three symmetric polymers show d2 ≈ 4.4-
4.9 Å, whereas only PCDTPX shows a peak at the same distance
spacing. These d values (d1 and d2) are similar to those reported
for poly(9,9-dioctylfluorene) in a nematic form.76-78 It is
reasonable to assume that d1 values correspond to the distance
between π-conjugated main chains separated by the long alkyl
side chain, whereas d2 values correspond to the distance between
the coplanar π-conjugated main chains. Logically, the comono-
(76) Grell, M.; Bradley, D. D. C.; Ungar, G.; Hill, J.; Whitehead, K. S.
Macromolecules 1999, 32, 5810-5817.
(79) Chen, C. T. Chem. Mater. 2004, 16, 4389-4400.
(80) Yamamoto, T.; Arai, M.; Kokubo, H.; Sasaki, S. Macromolecules 2003,
36, 7986-7993.
(81) Donley, C. L.; Zaumseil, J.; Andreasen, J. W.; Nielsen, M. M.; Sirringhaus,
H.; Friend, R. H.; Kim, J. S. J. Am. Chem. Soc. 2005, 127, 12890-12899.
(77) Chen, S. H.; Su, A. C.; Su, C. H.; Chen, S. A. Macromolecules 2005, 38,
379-385.
(78) Chen, S. H.; Su, A. C.; Chen, S. A. J. Phys. Chem. B 2005, 109, 10067-
10072.
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 737

A R T I C L E S
Blouin et al.
Figure 6. Cyclic voltammogram (second scan) of (a) symmetric and (b) asymmetric polymer films cast on a platinum wire in Bu4NBF4/acetonitrile at 50
mV/s.
Figure 7. Experimental and theoretical energy levels for the six polycarbazole derivatives.
shown in Figure 6a. The same observation is also true for the
three asymmetric polymers (PCDTPP, PCDTPT, PCDTPX) with
an Epa value of ca. 1.12 V (Figure 6b). Therefore, the substitution
of the electron-deficient center unit has almost no effect on the
polymer oxidation potential. However, replacing the benzene
core on the electron-deficient center unit by a pyridine core
induces an increase of the Epa by ca. 0.10 V. From these CV
curves, it was calculated that the three symmetric polymers
(PCDTQx, PCDTBT, PCDTBX) show HOMO energy levels
at ca. -5.46 eV and the three asymmetric polymers (PCDTPP,
PCDTPT, PCDTPX) at ca. -5.53 eV (Table 1 and Figure 7).
More importantly, based on the oxidation potential data, the
neutral polymers should show good air stability, the HOMO
energy level being largely below the air oxidation threshold (ca.
-5.27 eV or 0.57 V vs SCE).45 On the other hand, the polymer
LUMO energy level behaves quite differently. The LUMO
energy level is clearly affected by the electron-deficient center
on the bithiophene comonomer, a stronger electron-deficient unit
resulting in a lower LUMO energy level. The pyridine core also
affects the LUMO energy level, lowering the LUMO energy
level by ca. 0.25 eV when compared to the benzene core
polymers. It is worth also noting that the electrochemical and
optical band gap values are relatively in good agreement (Table
1).
9
738 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008

Design of Poly(2,7-Carbazole) Derivatives
A R T I C L E S
hand, the correlation for the small energy variations of the
HOMO energy levels is not as good, resulting in a lower
correlation factor (Figure 8a). Nevertheless, DFT calculations
on model compounds provide a good estimation of the HOMO
energies considering the small energy difference between the
lowest and highest values. As expected, theoretical energy band
gaps of the repeating units of the polymers are overestimated
compared to the experimental values of the polymers (see Fig-
ure 7), but a useful correlation can be found between the
theoretical and experimental band gaps as shown in Figure 8c.
From these data, one can conclude that DFT calculations
performed on the repetitive units can indeed provide good
estimations of the HOMO, LUMO, and band gap energy trends,
thus allowing a rapid screening of the most promising polymeric
structures.
2.3.7. Performance Predictions. All six copolymers have a
HOMO energy level within the desired range (-5.2 to -5.8
eV), assuring both an air-stable photovoltaic device and a
relatively high VOC. One can also expect similar VOC values
from all devices made with these polymers since they have very
similar HOMO energy level values. In contrast, only two
polymers (PCDTPT and PCDTPX) have an ideal LUMO energy
level (-3.7 to -4.0 eV). The four other polymers have higher
LUMO energy levels and, consequently, should present lower
efficiencies. Using the design rules proposed by Scharber et
al.12 which assume a charge carrier mobility of 10-3 cm2‚V-1‚s-1,
a fill factor of 0.65, and an optimal morphology, one can
estimate the overall power conversion efficiency from the optical
band gap and the LUMO values of the copolymers (Figure 9).
This model predicts that the asymmetric polymers (PCDTPP,
PCDTPT, and PCDTPX) should be more efficient than the
symmetric polymers (PCDTQx, PCDTBT, PCDTBX). Further-
more, based on this model, predicted device performances made
from PCDTPP, PCDTBT, and PCDTBX should be similar.
Finally, one can expect a power conversion efficiency near 10%
for PCDTPX (Figure 9).
2.4. Device Performances and Characterization. 2.4.1.
Field-Effect Transistors. As indicated above, transport proper-
ties are important parameters for efficient photovoltaic devices.
The electron transport properties of pure PCBM have been
already reported in details and are known to be sufficient for
high photovoltaic performances (∼10-3 cm2‚V-1‚s-1).46,47 In
order to avoid important photocurrent loss by recombination, a
hole mobility of g10-3 cm2‚V-1‚s-1 is required for the donor
conjugated polymer.12
The hole mobilities of our pure copolymers or in blends with
PCBM were studied using organic field-effect transistors
(OFETs) as the testing vehicle. Top contact OFETs of these
poly(2,7-carbazole) derivatives were fabricated on OTS-treated
SiO2/Si substrates as described in the Supporting Information.
All these poly(2,7-carbazole) derivatives were found to exhibit
typical p-type organic semiconductors characteristics. The hole
mobilities calculated in the saturation regime82 at VDS ) -100
V, the on/off current ratios, and the threshold potentials of these
polymers are summarized in Table 2.
Figure 8. Theoretical and experimental HOMO (A), LUMO (B), and band
gap (C).
Pure polymers spin-coated on OTS-treated substrates and
heated at 50 °C for 10 min show hole mobilities between 2 ×
10-5 and 1 × 10-3 cm2 V-1‚s-1. PCDTBT exhibits the highest
2.3.6. Validation of the Theoretical Calculations. An
excellent correlation can be established between the theoretical
LUMO energy levels of the repetitive units and the experimental
LUMO energy levels of the polymers (Figure 8b). On the other
(82) Dimitrakopoulos, C. D.; Malenfant, P. R. L. AdV. Mater. 2002, 14, 99-
117.
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 739

A R T I C L E S
Blouin et al.
Figure 9. Estimated power conversion efficiency for the six polycarbazole derivatives. (Adapted with permission from Wiley-VCH from the work of
Scharber and collaborators.12)
Table 2. Open Circuit Potential (Voc), Short Circuit Current (ISC), Fill Factor (FF), Power Convertion Efficiency (ηe), Hole Carrier Mobility (µ),
On/Off Current Ratios (Ion/Ioff), and Threshold Voltage (VT) for Polycarbazole Derivatives
PCBM/polymer (4:1)
Pristinea
PCBM/polymer (4:1)a
Voc
V
,
ISC
,
ηe
%
,
µ
,
VT,
(V)
µ
,
VT,
(V)
2
1
1
1
1
s-
polymer
mA
‚
cm-
FF
cm2
‚
V-
‚
s-
I
on/Ioff
cm2
‚
V-
‚
I
on/Ioff
PCDTQx
PCDTPP
PCDTBT
PCDTPT
PCDTBX
PCDTPX
0.95
0.90
0.86
0.71
0.96
0.85
-3.0
-2.6
-6.8
-2.9
-3.7
-1.4
0.56
0.44
0.56
0.32
0.60
0.60
1.8
1.1
3.6
0.7
2.4
0.8
3 × 10-4
2 × 10-5
1 × 10-3
4 × 10-5
1 × 10-4
5 × 10-4
1 × 104
1 × 102
3 × 104
1 × 102
5 × 103
1 × 103
-21
-40
-16
-48
-21
-46
7 × 10-6
no FEb
4 × 10-4
6 × 10-5
c
30
-
-29
-
-1
-32
c
2 × 102
2 × 102
c
3 × 10-4
4 × 102
-9
a Polymers and polymer/PCBM blends were deposited from a chloroform solution on OTS-treated SiO2/Si substrates and heated at 55 °C before Au
contact deposition. b No field effect (FE). c No thin films were obtained on OTS-treated SiO2/Si substrates.
hole mobility and on/off current ratio. As one can see in Table
2, the hole mobilities and on/off current ratios of symmetric
polymers are higher (by more than 1 order of magnitude) than
those of their asymmetric counterparts, except for PCDTBX
and PCDTPX which exhibit similar performances. This differ-
ence between the symmetric and asymmetric polymers could
be attributed to different solid-state packings and/or molecular
weights. The structural organization present in both PCDTBX
and PCDTPX, as observed in the X-ray diffraction pattern
(Figure 5), may explain the similar mobilities observed for both
polymers. Comparing PCDTQx and PCDTPP, which have
similar molecular weights, one can suggest that the higher
mobility of PCDTQx is due to its symmetric structure which
could induce some structural ordering in the thin film, as
revealed by polarized optical microscopy and X-ray measure-
ments. First attempts for OFET optimization by film annealing
at 150 °C for 20 min demonstrated interesting enhancement
(about 3 times) in performances for PCDTPP and PCDTBT.
For example, in the case of PCDTBT, a hole mobility and an
on/off current ratio up to 3 × 10-3 cm2‚V-1‚s-1 and 1 × 105
were obtained, respectively (Figure 10).
which should introduce more defects and consequently should
reduce the hole mobility, we note that the hole mobilities of
these polymers in the blends are very close to those obtained
with the pure polymers, except for PCDTQx and PCDTPP. In
fact, different research groups have found from studies on
OFETs and diodes that the blends of different polymers give
higher hole mobility values with increasing fractions of
PCBM.83-86
2.4.2. Photovoltaic Solar Cells. Devices using each of the
polymers were then made according to procedures described
in the Supporting Information. AFM images (see Supporting
Information, Figure 9S) reveal no important variation in the
morphology of PCBM/polymer (4:1) thin films. However, larger
structures are observed for the PCBM/PCDTBX and PCBM/
PCDTPP blends. We attribute those structures to polymer
aggregates present in chloroform solution. Indeed, these two
polymers have a lower solubility in chloroform compared to
the other polymers. Photovoltaic results are summarized in Table
2 and Figure 11. At first glance, and contrary to what was
proposed from the theoretical calculations, the most efficient
polymers are those with a symmetric structure. One can
conclude that the structural organization observed for the
All these polymers, except PCDTPP, when blended with
PCBM in a 1:4 stoichiometry show hole mobility values
between 5 × 10-6 and 4 × 10-4 cm2‚V-1‚s-1 in OFETs (Table
2). The PCDTBT/PCBM blend exhibits the highest hole
mobility. The higher structural organization present in PCDTBT,
as observed in DSC, POM, and X-ray analyses, is consequent
with such results. Despite the high ratio of PCBM in the blends
(83) Tuladhar, S. M.; Poplavskyy, D.; Choulis, S. A.; Durrant, J. R.; Bradley,
D. D. C.; Nelson, J. AdV. Funct. Mater. 2005, 15, 1171-1182.
(84) Mihailetchi, V. D.; Koster, L. J. A.; Blom, P. W. M.; Melzer, C.; de Boer,
B.; van Duren, J. K. J.; Janssen, R. A. J. AdV. Funct. Mater. 2005, 15,
795-801.
(85) Gadisa, A.; Wang, X.; Admassie, S.; Perzon, E.; Oswald, F.; Langa, F.;
Andersson, M. R.; Inganas, O. Org. Electron. 2006, 7, 195-204.
(86) Andersson, L. M.; Inganas, O. Appl. Phys. Lett. 2006, 88, 082103-3.
9
740 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008

Design of Poly(2,7-Carbazole) Derivatives
A R T I C L E S
Figure 10. Output at different gate voltages (VG) and transfer characteristics in the saturation regime at constant source-drain voltage (VDS ) -100 V) for
OFETs using spin-coated PCDTBT on OTS-treated SiO2/Si substrates.
Figure 11. Current density-potential characteristics of a diode made of PCBM/Polymer (4:1) under illumination with AM1.5G solar simulated light ([)
and in the dark (9).
symmetric polymers has also a major impact on the device
performances. Comparisons between PCDTPP and PCDTQx,
which have similar molecular weights, clearly illustrate this
behavior. In BHJ solar cells, their PCE is respectively of 1.1%
and 1.8%. These results are also in good agreement with the
OFET measurements. The hole mobility in pure PCDTQx is
about 1 order of magnitude higher than that obtained with
PCDTPP. Therefore, this higher hole mobility in PCDTQx,
presumably related to a better organization, could explain the
higher value of the short-circuit current (Isc) with PCDTQx,
despite the lower band gap for PCDTPP which is supposed to
increase its Isc.
weight polymers (ca. 4-5 KDa) generally exhibit low mobili-
ties,72,73 resulting in low Isc. However, in the case of PCDTPX,
the low Isc could not be explained solely by a low mobility of
the material as the hole mobility of PCDTPX (pure and in the
blend with PCBM) is about 4 × 10-4 cm2‚V-1‚s-1. The cause
for the low efficiencies for PCDTPX is poorly understood at
this time, but it is likely to be related to geminate charge
recombination at the interface due to stable charge-transfer
states, which limits the photocurrent.87 Moreover, low molecular
weight also has an impact on the film mechanical properties,
the interface resistance, and the nanoscale morphology, resulting
in a low open-circuit potential (Voc) and a low fill factor (FF).
PCDTBT gave the best results with a PCE of 3.6%. We note
here that this high efficiency is obtained due to the excellent Isc
According to the predicted performances, one can expect
relatively high PCEs for PCDTPT and PCDTPX. In contrast,
these polymers showed very low PCEs (under 1%). These results
can be explained by different factors. First, the low molecular
(87) Morteani, A. C.; Sreearunothai, P.; Herz, L. M.; Friend, R. H.; Silva, C.
Phys. ReV. Lett. 2004, 92, 247402-4.
9
J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008 741

A R T I C L E S
Blouin et al.
and FF values both attributed to the good hole mobility (see
Table 2) resulting from the higher structural organization present
in PCDTBT, as previously noted. However, one cannot exclude
the possibility of better contact electrode interfaces, better charge
transfers at the blend interface, or specific interactions induced
by the benzothiadiazole unit, since other benzothiadiazole-based
copolymers32-34,88-90 also demonstrated high PCE values.
the asymmetric polymers (except for PCDTPX). For all these
reasons, the polymer backbone (pyridine or benzene core) has
an important impact on the polymer hole mobility and power
conversion efficiency.
Taking all these data into account, one can conclude that, up
to now, PCDTBT has the best potential for photovoltaic devices.
It is anticipated that this value could be improved through
modifications in the device fabrication (film thickness, anneal-
ing, solvent, etc.). Optimization of the PCDTBX and PCDTPX
polymerization reactions should also provide better materials
with higher solubility and higher molecular weight and, hope-
fully, better performances in BHJ solar cell devices. Finally,
symmetric and lower band gap polycarbazole derivatives with
a LUMO energy level value similar to that of PCDTPX will
certainly be good candidates for BHJ solar cells.
3. Conclusions
We have demonstrated a versatile synthetic strategy to obtain
various low band gap polycarbazole derivatives via a Suzuki
coupling reaction. DFT theoretical calculations performed on
the repeat unit of the polymers provided good estimations of
the trends to be observed for the HOMO, LUMO, and band
gap energies of the corresponding polymers. Overall, the HOMO
energy level seems to be fixed by the carbazole moiety, whereas
the LUMO energy level seems to be directly related to the nature
of the electron-withdrawing unit. From these data, one can
expect to easily fine-tune the LUMO energy level for optimal
solar cell performances while synthesizing air-stable polymers
since the HOMO is more or less constant. The polymer
organization is also influenced by the polymer backbone
(benzene or pyridine core). As revealed by DSC, polarized
optical microscopy, and X-ray diffraction measurements, the
symmetric polymers show better structural organization than
Acknowledgment. This work was supported by discovery
and strategic grants from the Natural Sciences and Engineering
Research Council (NSERC) of Canada. N.B. thanks NSERC
for a Ph.D. Canada Graduate Scholarships (CGS D) and the
Fonds Que´be´cois de la recherche sur la nature et les technologies
(FQRNT) for a Ph.D. Energy research scholarship. The authors
would like to thank Pr. F.-G. Fontaine for helpful discussions
and Christian Tessier for monocrystal X-ray analyses.
Supporting Information Available: Details of syntheses,
characterization, computation method, device fabrication and
testing (pdf), and X-ray crystallographic information files of
compounds 4 and 6 (CIF). This material is available free of
charge via the Internet at http://pubs.acs.org.
(88) Brabec, C. J.; Winder, C.; Sariciftci, N. S.; Hummelen, J. C.; Dhanabalan,
A.; Hal, P. A. v.; Janssen, R. A. J. AdV. Funct. Mater. 2002, 12, 709-712.
(89) Boudreault, P.-L. T.; Michaud, A.; Leclerc, M. Macromol. Rapid Commun.
2007, 28, 2176-2179.
(90) Wong, W.-Y.; Wang, X.-Z.; He, Z.; Djurisic, A. B.; Yip, C.-T.; Cheung,
K.-Y.; Wang, H.; Mak, C. S. K.; Chan, W.-K. Nat. Mater. 2007, 6, 521-
527.
JA0771989
9
742 J. AM. CHEM. SOC. VOL. 130, NO. 2, 2008