Cyclometalated platinum polymers: Synthesis, photophysical properties, and photovoltaic performance

Article abstract of DOI:10.1021/cm9029038

The synthesis and characterization of platinum-containing conjugated polymers in which the platinum atom is attached to the conjugated backbone via a CN ligand is presented. The newly designed platinum-containing monomer can be polymerized under both Stil

Full text of DOI:10.1021/cm9029038

Chem. Mater. 2010, 22, 1977–1987 1977
DOI:10.1021/cm9029038
Cyclometalated Platinum Polymers: Synthesis, Photophysical Properties,
and Photovoltaic Performance
,†,§
Tabitha A. Clem,† David F. J. Kavulak,†,§ Erik J. Westling,†,§ and Jean M. J. Frechet*
ꢀ
†College of Chemistry, University of California, Berkeley, California 94720-1460 and §Materials Sciences
Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720
Received September 17, 2009. Revised Manuscript Received December 1, 2009
The synthesis and characterization of platinum-containing conjugated polymers in which the
platinum atom is attached to the conjugated backbone via a C∧N ligand is presented. The newly
designed platinum-containing monomer can be polymerized under both Stille and Suzuki conditions.
The polymers exhibit optical bandgaps between 2.1 and 1.65 eV depending on the choice of
comonomer. Triplet exciton formation is detected indirectly by measuring photosensitized emission
of singlet oxygen in both solution and in film. The ability of the materials to sensitize formation of
singlet oxygen varies both with excitation wavelength and with the change from solution to solid
state. This study provides design principles for developing conjugated polymers with significant
triplet yields in the solid state. The photovoltaic performance of these polymers was also evaluated in
preliminary experiments with power conversion efficiencies as high as 1.3% obtained for a bulk
heterojunction cell with PCBM.
Introduction
Materials with large triplet yields may well provide
access to increased current in organic photovoltaic de-
vices.11 Triplet excited state lifetimes are typically in the
microsecond regime, which is 3 orders of magnitude
longer than the nanosecond decays typically observed
for singlet excited states in conjugated polymers. Given
that exciton diffusion length (LD) is determined by exci-
Conjugated polymers with their highly tunable opto-
electronic properties and ease of processing1,2 are interest-
ing materials in the challenging search for efficient low-
cost photovoltaic devices. Recently, there has been signi-
ficant interest in conjugated organometallic polymers as
donor polymers for use in bulk heterojunction (BHJ)
solar cells in combination with fullerene acceptors.3-10
In particular, polyplatinynes have drawn some attention
as a result of their absorption profiles extending as far as
the near-infrared,10 and their promising charge transport
properties with reported hole mobilities as high as 1.0 ꢀ
10-2 cm2 V-1 s-1 in field-effect transistors (FETs).9 In
addition, polyplatinynes have demonstrated triplet exci-
ton formation.3,7
ton lifetime (τ) and exciton diffusivity (D) according to
pffiffiffiffiffiffi
the equation LD
¼
τD, a significant increase in lifetime
should lead to a longer exciton diffusion length, which
would be advantageous for bilayer photovoltaic devices.
Triplet-forming polymers have been reported to inhibit
geminate pair recombination3 and to increase the exciton
diffusion length12 in bulk heterojunction devices. How-
ever, the conjugated organometallic polymers that have
been shown to form triplets in the solid state are also large
bandgap systems12 (>2.5 eV), and therefore power con-
version efficiencies are limited by the poor overlap of their
absorption spectrum with the solar irradiance. While
lower band gap systems have demonstrated formation
of triplet excitons in solution, this result does not neces-
sarily translate to the solid state. Furthermore, perfor-
mance of these lower band gap polymers are limited by
charge mobility.7 It is also important to consider that
formation of triplet excitons in a bulk heterojunction
device is anticipated to improve the photocurrent pri-
marily by decreasing geminate recombination, as charge
separation in a well-mixed bulk heterojunction occurs on
the femtosecond time scale13 while intersystem crossing
*Author to whom correspondence should be addressed: E-mail: frechet@
berkeley.edu.
(1) Thompson, B. C.; Frechet, J. M. J. Angew. Chem., Int. Ed. 2008, 47,
ꢀ
58–77.
(2) Blom, P. W. M.; M., V. D.; Koster, L. J. A.; Markov, D. E. Adv.
Mater. 2007, 19, 1551–1566.
(3) Guo, F.; Kim, Y.-G.; Reynolds, J. R.; Schanze, K. S. Chem.
Commun. 2006, 1887–1889.
(4) Wong, Wai-Yeung Macromol. Chem. Phys. 2008, 209, 14–24.
(5) Wong, W.-Y.; Wang, X.-Z.; He, Z.; Chan, K.-K.; Djurisic, A. B.;
Cheung, K.-Y.; Yip, C.-T.; Ng, A. M.-C.; Xi, Y. Y.; Mak, C. S. K.;
Chan, W.-K. J. Am. Chem. Soc. 2007, 129, 14372–14380.
(6) Wu, P.-T.; Bull, T.; Kim, F. S.; Luscombe, C. K.; Jenekhe, S. A.
Macromolecules 2009, 42, 671–681.
(7) Mei, J.; Ogawa, K.; Kim, Y.-G.; Heston, N. C.; Arenas, D. J.;
Nasrollahi, Z.; McCarley, T. D.; Tanner, D. B.; Reynolds, J. R.;
Schanze, K. S. ACS Appl. Mater. Interfaces 2009, 1, 150–161.
(8) Sudha Devi, L.; Al-Suti, M. K.; Zhang, N.; Teat, S. J.; Male, L.;
€
Sparkes, H. A.; Raithby, P. R.; Khan, M. S.; Kohler, A. Macro-
molecules 2009, 42, 1131–1141.
(9) Baek, N. S.; Hau, S. K.; Yip, H.-L.; Acton, O.; Chen, K.-S.; Jen,
A. K.-Y. Chem. Mater. 2008, 20, 5734–5736.
(10) Wang, X.-Z.; Wong, W.-Y.; Cheung, K.-Y.; Fung, M.-K.;
Djurisic, A. B.; Chan, W.-K. Dalton Trans. 2008, 5484–5494.
(11) Y. Shao, Y. Y. Adv. Mater. 2005, 17, 2841–2844.
(12) Schulz, G. L.; Holdcroft, S. Chem. Mater. 2008, 20, 5351–5355.
(13) Moses, D.; Dogariu, A.; Heeger, A. J. Chem. Phys. Lett. 2000, 316,
356–360.
r
2009 American Chemical Society
Published on Web 12/31/2009
pubs.acs.org/cm

1978 Chem. Mater., Vol. 22, No. 6, 2010
Clem et al.
occurs in the nanosecond regime.14 Therefore, in a bulk
heterojunction system which has a closely mixed morpho-
logy, charge separation will most likely occur before
triplet excitons can form. Thus, any improvement in
photovoltaic performance for triplet-forming materials
in a bulk heterojunction is expected to be primarily the
result of decreased geminate recombination or another
factor.
Figure 1. General structure of cyclometalated Pt polymers.
In polyplatinynes the poor size and energetic overlap
between the 5d Pt and 2p C orbitals leads to a significantly
smaller effective conjugation length when compared to
structurally analogous poly(phenylene ethynylenes).15
Given that exciton diffusion length is a function of both
lifetime and diffusivity, this exciton localization, or de-
creased effective conjugation length, may ultimately limit
the photovoltaic performance and/or the exciton diffu-
sion lengths observed for these polymers even if long-
lived triplets are realized. Organometallic conjugated
polymer architectures other than polyplatinynes are
therefore worthy of investigation as donor materials to
better understand the role of heavy atoms on triplet
formation in conjugated polymers for possible applica-
tion in photovoltaics.
We report a new class of platinum-containing conju-
gated polymers with optoelectronic properties that are
more suitable for photovoltaic applications. These poly-
mers contain a 2-(20-thienyl)thiazole C∧N ligand with an
O∧O diketonate ligand that creates a structure analogous
to a fused bithiophene system but in which platinum is
adjacent to the conjugated backbone and connects the
aryl groups into a coplanar conformation. The general
structure of these new organometallic polymers is shown
in Figure 1. Fused bithiophenes have recently received
attention as a building block in low bandgap conjugated
polymers, where the optical and electronic properties
of the polymer vary with the choice of connecting
atom.16-19 In contrast to the platinum acetylides, the
platinum atom is peripheral to the conjugated polymer
backbone, leading to a greater involvement from the
orbitals of Pt in both the ground and the excited states
of the material.20 This design attempts to minimize any
potential decrease in effective conjugation length from
the poor overlap of Pt and C orbitals, and creates a
platform for the study of the influence of a heavy atom
on more diffuse excitons. This design also provides flex-
ibility in tuning the absorption profile and energy levels of
the materials simply by varying the Pt ligand and co-
monomer.
Experimental Section
Synthesis. All chemicals were purchased from commercial
sources and used without further purification unless otherwise
stated. 2,5-bis(trimethylstannyl)thiophene21 and 2-(trimethyl-
stannyl)thiophene22 were prepared as described in the literature.
Tetrahydrofuran (THF), N,N-dimethylformamide (DMF), and
toluene were dried by passing through neutral alumina prior to
1
use. All small molecules were characterized by H NMR (400
MHz) and 13C NMR (100 MHz) on a Bruker AVB 400 or AVQ
400. Polymer 1H NMR (500 MHz) spectra were obtained on a
Bruker DRX 500. All NMR spectra were referenced to TMS.
High-resolution mass spectroscopy and elemental analysis
(CHN) were performed at the University of California, Berkeley
Department of Chemistry analytical services laboratory. IR
spectra were recorded on a Varian 3100 FT-IR as KBr pellets
or as solid or liquid films drop-cast on KBr discs. Melting points
are uncorrected. Column chromatography was performed with
silica (230 ꢀ 400 mesh). Polymer molecular weights were deter-
mined by size exclusion chromatography (SEC) using either
dichlorobenzene or THF as the mobile phase as specified. SEC
with THF mobile phase (1 mL/min, 45 °C) was carried out using
a Waters GPC 150-CV plus system (Milford, MA) with an
attached M486 tunable absorbance detector (254 nm), calibra-
ted with polystyrene standards. SEC in dichlorobenzene was
performed using HPLC grade dichlorobenzene at a flow rate of
0.8 μL/min on two 300 ꢀ 8 mm linear S SDV, 5 μm columns
(Polymer Standards Services, USA Inc.) at 70 °C using a Waters
(Milford, MA) 2690 separation module and a Waters 486
tunable absorption detector monitored at 350 nm. The instru-
ment was calibrated vs polystyrene standards (1,050 - 135,000 g/
mol) and data was analyzed using Millenium 3.2 software.
2-(20-Thienyl)thiazole (1). 2-(trimethylstannyl)thiophene (1
eq, 6.7 g, 27 mmol) and 2-bromothiazole (1.1 eq, 5.41 g, 33
mmol) were dissolved in anhydrous toluene (240 mL) and
anhydrous DMF (60 mL). The solution was degassed by pur-
ging with nitrogen for 20 min, then Pd(PPh3)4 was added under a
flow of nitrogen. The solution was kept at 100 °C overnight, then
cooled to room temperature and water was added. The aqueous
layer was then washed with ether (3 ꢀ 100 mL). The organic
layers were combined, washed with 1 M HCl three times then
dried with anhydrous NaSO4. The solution was filtered and
solvent removed to afford an off-white oil. Following purifica-
tion by column chromatography (SiO2, 15% EtOAc in hexanes)
4.49 g of the title compound (99% yield) was obtained as a clear
oil. 1H NMR (400 MHz, CDCl3) δ 7.76 (d, 1H), 7.50 (dd, 1H),
7.37 (dd, 1 H), 7.22 (d, 1 H), 7.06 (td, 1 H). 13C NMR (100 MHz,
CDCl3) δ 162.0, 143.3, 137.3, 127.9, 127.7, 126.6, 118.2. HRMS
(14) Turro, N. J. Modern Molecular Photochemistry; University Science
Book: Sausalito, CA, 1991; p 628.
(15) Silverman, E. E.; Cardolaccia, T.; Zhao, X.; Kim, K.-Y.;
Haskins-Glusac, K.; Schanze, K. S. Coord. Chem. Rev. 2005,
249, 1491–1500.
(16) Liao, L.; Dai, L.; Smith, A.; Durstock, M.; Lu, J.; Ding, J.; Tao, Y.
Macromolecules 2007, 40, 9406–9412.
(17) Zhou, E.; Nakamura, M.; Nishizawa, T.; Zhang, Y.; Wei, Q.;
Tajima, K.; Yang, C.; Hashimoto, K. Macromolecules 2008, 41,
8302–8305.
(18) Soci, C.; H., I.-W.; Moses, D.; Zhu, Z.; Waller, D.; Gaudiana, R.;
Brabec, C. J.; Heeger, A. J. Adv. Funct. Mater. 2007, 17, 632–636.
ꢀ
€
(19) Moule, A. J.; Tsami, A.; Bunnagel, T. W.; Forster, M.; Kronenberg,
N. M.; Scharber, M.; Koppe, M.; Morana, M.; Brabec, C. J.;
Meerholz, K.; Scherf, U. Chem. Mater. 2008, 20, 4045–4050.
(21) Van, P. C.; Macomber, R. S.; Mark, H. B.; Zimmer, H. J. Org.
Chem. 1984, 49, 5250–5253.
(20) Brooks, J.; Babayan, Y.; Lamansky, S.; Djurovich, P. I.; Tsyba, I.;
Bau, R.; Thompson, M. E. Inorg. Chem. 2002, 41, 3055–3066.
(22) Rossi, R.; Carpita, A.; Ciofalo, M.; Lippolis, V. Tetrahedron 1991,
47, 8443–8460.

Article
Chem. Mater., Vol. 22, No. 6, 2010 1979
(EI) calcd. for C7H5NS2 167.9897 found, 167.9891. CHN Calcd.
50.27% C, 3.01% H, 8.37% N, 38.34% S. Found 50.16% C,
3.06% H, 8.24% N, 38.04% S.
dichloromethane to yield an orange solid with mp at 131-
133 °C (0.76 g, 51% yield). 1H NMR (400 MHz, CDCl3) δ
8.00 (dd, 4H), 7.31 (d, 2H), 7.57 (d, 1H), 7.30 (d, 1H), 7.16 (d,
1H), 6.94 (dd, 4H), 6.71 (s, 1H), 4.02 (t, 4H), 1.82 (q, 4H), 1.49
(m, 4H), 1.34 (m, 16H), 0.91 (t, 6H) 13C NMR (100 MHz,
CDCl3) δ 177.74, 176.79, 173.39, 161.48, 146.17, 137.57, 131.97
(d, J = 103 Hz), 129.77 (d, J = 85 Hz), 128.76, 114.24, 113.5,
95.31, 68.17, 31.81, 29.35, 29.23, 29.17, 26.02, 22.65, 14.11.
CHN Calcd. 54.27% C, 5.63% H, 1.67% N. Anal. 54.31% C,
5.32% H, 1.38% N. HRMS (FAB) for C38H47NO4PtS2 calcd.
840.2638, found 840.2637 IR (KBr film) 3097, 3061, 2924, 2850,
2362, 2311, 1604, 1586, 1524, 1492, 1469, 1444, 1380, 1306, 1263,
1229, 1175, 1129, 1007, 940, 902, 882, 841, 784, 740, 703, 609,
5-Bromo-2-(5-bromothiophen-2-yl)-1,3-thiazole (2). 2-(20-
Thienyl)thiazole (1 eq, 7.2 mmol, 1.4 g) was dissolved in 100
mL DMF in an ice bath and treated dropwise with a solution of
N-bromosuccinimide (2.5 eq, 18 mmol, 3.3 g) in 100 mL DMF.
The solution was allowed to warm to room temperature and
was stirred overnight. The reaction mixture was poured over
ice water, and a white solid was formed. The solution was
then extracted with diethyl ether (3 ꢀ 100 mL). The combined
organic layers were then washed with water (3 ꢀ 100 mL). The
ether layer was dried with NaSO4, filtered, and the solvent
removed to obtain a pale orange solid. This solid was applied
to a plug of silica and eluted with 20% ethyl acetate in hexanes to
obtain the product as a white solid with mp 118-121 °C in 77%
yield (2.1 g). 1H NMR (400 MHz, CDCl3) δ 7.63 (s, 1H), 7.18 (d,
1H), 7.04 (d, 1H). 13C NMR (100 MHz, CDCl3) δ 162.2, 144.7,
138.3, 131.0, 126.9, 116.2, 108.2. HRMS (EI) calcd. for C7H3-
Br2NS2 326.8033, found 326.8031. CHN Calcd. 25.87% C,
0.93% H, 4.31% N. Found 25.97% C, 0.83% H, 4.22% N.
(2Z)-3-hydroxy-1,3-bis[4-(octyloxy)phenyl]prop-2-en-1-one (3).
To a flame-dried, three-neck flask containing 8 (1 eq, 30 mmol,
7.45 g) and 9 (1 eq, 30 mmol, 7.93 g) was added 1,2-dimethoxy-
ethane (150 mL) via cannula. NaH (5 eq, 150 mmol, 3.6 g) was
then added in one portion to the stirred solution kept under a
strong flow of nitrogen. Once the spontaneous evolution of
bubbles from the reaction ceased, the solution was heated to
reflux for 24 h, then cooled to room temperature and slowly
added to ice water under rapid stirring. This solution was then
extracted with ether three times. The combined organic extracts
were dried over MgSO4 and solvent was removed under reduced
pressure to yield a pale orange solid. This material was recrys-
tallized from a mixture of THF and methanol to yield an off-
white solid 3 melting at 76 -77 °C in 48% yield. 1H NMR (400
MHz,CDCl3) δ 17.25 (s, 1H), 7.99 (d, 4H), 7.0 (d, 4H), 6.77 (1,
1H), 4.07 (t, 4H), 1.89 (m, 4H), 1.51 (m, 4H), 1.36 (m, 16H), 0.94
(t, 6H). 13C (100 MHz, CDCl3) δ 184.57, 162.61, 129.02, 127.88,
114.35, 91.35, 68.23, 31.79, 29.32, 29.21, 29.11, 25.98, 22.64,
14.09. HRMS (FAB) for C31H44O4 calcd. 481.3318, found
481.3323. CHN Calcd. 77.46% C, 9.23% H. Anal. 77.42% C,
9.48% H. IR (KBr Pellet) 2957, 2923, 2855, 2362, 2343, 1545,
1507, 1491, 1472, 1429, 1385, 1308, 1263, 1227, 1174, 1125, 1105,
1064, 1030, 999, 972, 945, 922, 895, 859, 847, 785, 721, 699, 668,
614, 549 cm-1
.
Platinu0m(II) (5-bromo-2-(5-bromothiophen-2-yl)-1,3-thiazo-
lato-N,C2 ) ((2Z)-3-hydroxy-1,3-bis[4-(octyloxy)phenyl]prop-2-
en-1-onato-O,O) (5). Potassium tetrachloroplatinate (0.45 g,
1.1 mmol, 1 equiv) was dissolved in 20 mL H2O and 5-bromo-
2-(5-bromothiophen-2-yl)-1,3-thiazole (0.7 g, 2.2 mmol, 2
equiv) was dissolved separately in 70 mL ethoxyethanol. The
solutions were combined and heated to 90 °C overnight under
nitrogen. The reaction mixture was then cooled to room tem-
perature and poured over 90 mL H2O. A dark green precipitate
was collected by filtration, rinsed with water and methanol, then
dried under vacuum overnight (0.6 g). It was used in subsequent
steps without additional purification. The dried precipitate (1
eq, 1.35 mmol, 1.52 g), 3 (3 eq, 4.05 mmol, 1.97 g), and Ag2O
(1.85 eq, 2.5 mmol, 0.59 g) were dissolved in dry THF (15 mL)
and heated to reflux for 24 h. The solution was then cooled to
room temperature and solvent removed under reduced pressure.
The product was purified using column chromatography (silica,
CH2Cl2, Rf = 1.0). The solvent was removed and further
purification achieved by recrystallizing the obtained solid from
a mixture of CH2Cl2 and methanol. The product 5 was obtained
1
as a yellow solid with mp 162-165 °C. H NMR (400 MHz,
CDCl3) δ 7.93 (t, 4H), 7.68 (s, 1H), 7.10 (s, 1H), 6.93 (t, 4H), 6.58
(s, 1H), 4.04 (t, 4H), 1.82 (m, 4H), 1.52 (m, 4H), 1.35 (m, 16H),
0.93 (t, 6H). 13C (100 MHz, CDCl3) δ 177.71, 176.67, 173.37,
161.78 (d, J = 12 Hz), 148.86, 138.16, 134.04 (d, J = 81 Hz),
131.82 (J = 82 Hz), 129.03, 118.50, 114.51, 101.96, 95.54, 68.44,
53.64, 32.06, 29.61, 29.48, 29.45, 26.28, 22.90, 14.34. HRMS
(FAB) for C38H45Br2NO4PtS2 calcd. 998.0841, found 998.0844.
CHN Calcd. 45.7% C, 4.54% H, 1.40% N. Anal. 45.47% C,
4.45% H, 1.38% N. IR (KBr Pellet) 2920, 2853, 2361, 2343,
1605, 1585, 1559, 1524, 1524, 1472, 1426, 1383, 1305, 1264, 1229,
653, 643, 597 cm-1
.
20
Platinum(II) (2-(20-thienyl)thiazolato-N,C ) ((2Z)-3-hydroxy-
1,3-bis[4-(octyloxy)phenyl]prop-2-en-1-onato-O,O) (4). Pota-
ssium tetrachloroplatinate (0.62 g, 3 mmol, 1 equiv) was dis-
solved in 60 mL H2O, and 2-(20-thienyl)thiazole (0.5 g, 6 mmol,
2 equiv) was dissolved separately in 180 mL ethoxyethanol. The
solutions were combined and heated to 90 °C overnight under
nitrogen. The solution was then cooled to room temperature
and poured over 100 mL of ice-cold water. The dark green preci-
pitate which formed was collected by filtration and washed with
water and methanol, then dried under vacuum overnight with-
out further purification (0.456 g). The dried precipitate (1.0 g,
1.8 mmol, 1 equiv) was treated with 3 (2.59 g, 5.4 mmol, 3 equiv)
and Ag2O (0.76 g, 3.33 mmol, 1.85 equiv) in dry, degassed THF
(18 mL), and heated to reflux overnight. The reaction mixture
was filtered, and the remaining solid was washed extensively
with dichloromethane. After removing the solvent, the resulting
brown solid was passed through a silica column (CH2Cl2). The
product collected as the first orange spot was further puri-
fied by recrystallization from a mixture of methanol and
1178, 1132, 1107, 1037, 959, 839, 781, 667, 635, 601, 578 cm-1
.
4-Octyloxyacetophenone (8). 4-hydroxyacetophenone (1 eq,
60 mmol. 8.17 g) was dissolved in anhydrous DMF (120 mL)
with anhydrous K2CO3 (3 eq, 180 mmol, 24.88 g) and 1-bromo-
octane (1.2 eq, 72 mmol, 13.90 g). The solution was stirred at
60 °C overnight. The reaction was cooled to room temperature,
then poured over 100 mL water and extracted into ether (3 ꢀ
100 mL). The organic fractions were combined and washed with
water three times to remove any remaining DMF. The organic
fractions were then dried with MgSO4 and the solvent removed
under reduced pressure to yield a brown oil. This oil was flushed
through a plug of silica in 10% EtOAc in hexanes to give a pale
brown oil. The oil was stirred under vacuum at 60 °C overnight
to remove any residual bromooctane. The oil was then placed in
the freezer at -20 °C, whereupon it became a solid and remained
a solid with mp at 32-35 °C. Yield: 12.89 g, 85%. 1H NMR (400
MHz,CDCl3): δ 7.99 (d, 2H), 6.90 (d, 2H), 4.01 (t, 2H), 2.55 (s,
3H), 1.82 (quintet, 2H), 1.2-1.5 (m, 10H), 0.87 (t, 3H). 13C
NMR (100 MHz, CDCl3) δ 196.76, 163.08, 130.54, 130.02,

1980 Chem. Mater., Vol. 22, No. 6, 2010
Clem et al.
114.07, 68.21, 31.77, 29.29, 29.19, 29.06, 26.30, 25.95, 22.62,
14.08. HRMS (EI) for C16H24O2 calcd. 248.1776, found
248.1771 CHN Calcd. 77.38% C, 9.74% H. Anal. 77.48% C,
9.96% H. mp 28-31 °C IR (KBr film) 2925, 2856, 2364, 1677,
1601, 1578, 1541, 1509, 1468, 1420, 1357, 1306, 1255, 1171, 1114,
2956, 2925, 2854, 2365, 1674, 1604, 1586, 1527, 1492, 1467,
1425, 1383, 1305, 1259, 1229, 1175, 1134, 1093, 1024, 841, 801,
725 cm-1
.
Optical and Electrochemical Measurements. Cyclic voltam-
metry (CV) was performed using a Solartron 1285 potentiostat.
For the polymers, CV was performed on thin films dip-coated
onto a Pt wire working electrode and submerged in CH3CN
freshly distilled from CaH2. For small molecules, the cyclic
voltammograms were measured in dichloromethane solution
with a Pt wire working electrode. All measurements were
performed using a silver wire pseudoreference electrode, a
platinum auxiliary electrode, and were referenced to the ferro-
cene/ferrocenium couple, which was taken to be -5.1 eV relative
to vacuum.23 Tetrabutylammonium tetrafluoroborate (NBu4-
BF4) was the supporting electrolyte for all measurements.
UV-visible absorption spectra were obtained using a Cary
50 UV-visible spectrophotometer. For thin film measurements
polymers were spin coated onto untreated glass slides from
chlorobenzene solution (10 mg mL-1). A model P6700 Spin-
coater was used to spin coat the films at 1500 rpm for 60 s.
Photoluminescence spectra were obtained using a Horiba Jobin
Yvon Nanolog fluorimeter. For deoxygenated samples, the
solvent was degassed by three freeze-pump-thaw cycles, and
then the solutions were freshly prepared inside of a glovebox
under argon atmosphere. For aerated samples, the solution was
purged with air for 5 min prior to measurement.
1077, 1019, 955, 834, 814, 725, 590 cm-1
.
4-Octyloxy Benzoic Acid Methyl Ester (9). 4-Hydroxybenzoic
acid methyl ester (1 eq, 60 mmol, 9.13 g) was dissolved in
anhydrous DMF (120 mL) with anhydrous K2CO3 (3 eq, 180
mmol, 24.88 g) and 1-bromooctane (1.2 eq, 72 mmol, 13.90 g)
and stirred at 60 °C overnight. The reaction was cooled to room
temperature, then water and ether were added. The organic
layer was washed three times with water to remove any remain-
ing DMF. The organic fractions were then dried with MgSO4
and the solvent removed to give an off-white solid. This solid
was applied to a silica column and eluted with 10% EtOAc in
hexanes. Recrystallization from ethanol afforded a white solid
1
with mp 35-36 °C. Yield: 12.0 g, 76%. H NMR (400 MHz,
CDCl3): δ 7.99 (d, 2H), 6.91 (d, 2H), 3.98 (t, 2H), 3.88 (s, 3H),
1.77 (m, 2H), 1.49 (m, 2H), 1.32 (m, 8H), 0.89 (t, 3H). 13C (100
MHz, CDCl3) δ 167.11, 163.16, 131.75, 122.49, 114.24, 68.39,
52.00, 32.00, 29.52, 29.42, 29.31, 26.18, 22.85, 14.29. HRMS (EI)
for C16H24O3 calcd. 265.1759, found 265.1761 CHN Calcd.
72.69% C, 9.15% H. Anal. 72.85% C, 9.38% H. IR (KBr disk)
2929, 2857, 1721, 1610, 1607, 1579, 1512, 1469, 1435, 1394, 1315,
1280, 1255, 1169, 1105, 1023, 972, 847, 771, 697, 648 cm-1
.
Device Fabrication and Measurement. Photovoltaic devices
consisted of a standard ITO/PEDOT:PSS/Pt-polymer:PCBM/
Al architecture. Indium-doped tin oxide (ITO) coated glass
substrates were purchased from Thin Film Devices, Inc. The
substrates (150 nm sputtered pattern, 10 Ω 0-1) were cleaned by
20 min of sonication in acetone, 2% Helmanex soap in water,
and finally isopropanol. The substrates were then dried under a
stream of air before being coated immediately with a filtered
(0.45 μm GHP) dispersion of PEDOT:PSS in water (Baytron-
PH) via spin coating for 30 s at 4000 rpm. The resulting polymer
layer was ∼30 nm thick after baking at 140 °C for 20 min. All
subsequent device fabrication was performed inside a glovebox
under inert Ar atmosphere with water and oxygen levels below
1 ppm. Each Pt-polymer was dissolved at a concentration of
16 mg mL-1 in chlorobenzene. PCBM (purchased from Nano-
C) was dissolved separately at 40 mg mL-1 in chlorobenzene and
all solutions were allowed to stir overnight at 120 °C. The
solutions were then combined in various ratios from 1:1 to 1:6
polymer:PCBM along with additional chlorobenzene as needed
to a final polymer concentration of 8 mg mL-1, before spin-
casting onto the PEDOT:PSS-treated ITO at 1200 rpm for 30 s.
100 nm aluminum electrodes were deposited by thermal resis-
tance evaporation at pressures of approximately 10-6 torr to
complete the device structure. The shadow mask used during
thermal deposition yielded eight independent devices per sub-
strate each with a surface area of 0.03 cm2. Completed devices
were then tested under Ar(g) using a 300 W Thermo-Oriel
Xenon arc-lamp with flux control spectrally corrected to AM
1.5 G with one filter (Thermo-Oriel No. 81088). The AM 1.5 G
light was further attenuated using a 0.5 O.D. neutral density
filter, and the intensity of the AM 1.5 G light was calibrated to be
100 mW cm-2 by a spectrally matched Hamamatsu S1787-04
photodiode (calibrated by NREL and obtained through Na-
nosys Inc.). I-V behavior was measured using a computer-
controlled Keithley 236 SMU.
Pt-T1. To a flame-dried three neck flask was added 5 (1 eq.,
0.100 g, 0.100 mmol) and 5 mL anhydrous THF and the solution
was treated with 2,5-bis(trimethylstannyl)-thiophene (1 eq,
0.041 g, 0.100 mmol). The solution was stirred, purging with
nitrogen for 20 min. Following degassing, Pd(PtBu3)2 (0.02 eq,
1 mg, 0.002 mmol) and CsF (4 eq, 0.060 g, 0.402 mmol) were
added quickly under strong flow of nitrogen. The solution was
then stirred and heated to 40 °C overnight. After 24 h, the
solution was cooled to room temperature and precipitated into
methanol. The crude polymer was recovered by filtration and
further purified by Soxhlet extraction with methanol, chloro-
form, and chlorobenzene. The chlorobenzene fraction was
concentrated and reprecipitated into methanol to obtain
29 mg (32% yield) of polymer free of low molecular weight
material.
1H NMR (500 MHz, C2D2Cl2, 90 °C) δ 6.5-8.0 (br, 13H),
3.96 (br, 4H), 0.5-2.0 (br, 30H). CHN Calcd. 54.77% C, 5.14%
H, 1.52% N, 10.44% S. Anal. 51.66% C, 5.49% H, 1.51 N,
9.29% S. SEC (dichlorobenzene) Mn 45 kDa, Mw 84 kDa, PDI
1.9 IR (KBr Pellet) 2923, 2849, 2366, 2344, 1602, 1585, 1528,
1490, 1472, 1418, 1384, 1303, 1256, 1227, 1173, 1132, 1108, 1022,
969, 954, 837, 779, 667, 653, 630, 561, 519 cm-1
.
F8TZPt. To a three-neck flask was added 5 (1 eq, 0.060 g,
0.060 mmol) and 9,9-dioctylfluorene-2,7-bis(trimethylborate) (1
eq, 0.032 g, 0.06 mmol), and potassium fluoride (3.3 eq, 0.012 g,
0.198 mmol). THF (3 mL) and H2O (1 mL) were added and the
solution purged with nitrogen for 20 min. Pd(PtBu3)2 (0.08 eq,
0.02 g, 0.004 mmol) and Pd2dba3 (0.04 eq, 0.02 g, 0.002 mmol)
were added under strong flow of nitrogen and the solution was
stirred at 40 °C for 24 h. After 24 h the solution was precipitated
into methanol, and then purified by Soxhlet extraction with
methanol, then chloroform. The chloroform fraction was con-
centrated and reprecipitated into methanol to obtain 54 mg of
1
polymer (73% yield). H NMR (500 MHz, CDCl3) δ 8.0-8.5
(br, 5H), 7.3-7.9 (m, 6H), 6.6-7.2 (m, 6H), 4.1 (s, 4H), 1.7-2.4
(br, 8H), 1.0-1.5 (m, 38H), 0.5-1.0 (m, 18H). CHN Calcd.
65.44% C, 7.13% H, 1.14% N. Anal. 61.32% C, 6.93% H, 1.12
N. SEC (THF) Mn 24 kDa, Mw 47 kDa, PDI 1.9. IR (KBr film)
(23) Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta. 2000, 298,
97–102.

Article
Chem. Mater., Vol. 22, No. 6, 2010 1981
Scheme 1. Synthesis of Monomer and Model Complex
Polymer mobility was measured using a diode configuration
thiophene. The thiazole unit should also contribute fa-
vorably to charge mobility29,30 while providing a nitrogen
atom for Pt coordination and minimizing donor-accep-
tor interactions in the polymer. The diphenyl ketonate
ligand provides solubility via the peripheral dialkoxy
groups. Since platinum complexes containing the 2-(20-
thienyl)thiazole ligand have not been reported previously,
model complex 4 without terminal bromine moieties was
synthesized to enable a comparison of optical and elec-
tronic properties of the repeat unit to those of the corres-
ponding polymers. Stille and Suzuki polymerizations
were carried out using Pd(PtBu3)2, and conditions were
based on those developed by the Fu group,31,32 as shown
in Scheme 2. This palladium system was used in place of
the Pd(PPh3)4 analog, because compound 5 was found to
degrade under polymerization conditions using the more
classical catalyst. This degradation was confirmed by
of ITO/PEDOT:PSS/Polymer/Al in the space charge limited
current regime. At sufficient potential the conduction of charges
in the device can be described by
9
V2
L3
JSCLC ¼ εrε0μ
ð1Þ
8
where ε0 is the permittivity of space, εR is the dielectric of the
polymer (assumed to be 3), μ is the mobility of the majority
charge carriers, V is the potential across the device (V = Vapplied
- Vbi - Vr), and L is the polymer layer thickness. The series and
contact resistance of the device (13-21 Ω) was measured using a
blank (ITO/PEDOT/Al) and the voltage drop due to this
resistance (Vr) was subtracted from the applied voltage. The
built-in voltage (Vbi), which is based on the relative work
function difference of the two electrodes, was also subtracted
from the applied voltage. The built-in voltage can be estimated
from the difference in work function between the cathode and
anode and is found to be about 1 V.
1
significant changes to the H NMR spectrum of 5 after
adding a solution of PPh3. The exact nature of this
degradation is unknown, but Pt(acac)2 complexes have
been shown to react with triaryl phosphine ligands by
displacing one of the O atoms of the diketonate ligand33
or by forming a pentacoordinate complex.34 The tri(tert-
butyl)phosphine ligand is significantly more bulky than
triphenylphosphine, as evidenced by its ability to form a
stable divalent Pd(0) complex at room temperature,32 and
by its larger cone angle of 182° versus the 145° determined
Results and Discussion
Synthesis. The new Pt-containing dibrominated mono-
mer 5 was synthesized as shown in Scheme 1, using a
modification of the procedure of Venkatesan et al.24 In
order to take advantage of the high photovoltaic efficien-
cies,25,26 high charge carrier mobilities27 and strong light
absorption28 associated with polythiophenes, we chose
the 2-(20-thienyl)thiazole C∧N ligand which also contains a
(24) Venkatesan, K.; Kouwer, P. H. J.; Yagi, S.; Muller, P.; Swager,
T. M. J. Mater. Chem. 2008, 18, 400–407.
(25) Hou, J.; Chen, H.-Y.; Zhang, S.; Li, G.; Yang, Y. J. Am. Chem. Soc.
2008, 130, 16144–16145.
(29) Hong, X. M.; Katz, H. E.; Lovinger, A. J.; Wang, B.-C.;
Raghavachari, K. Chem. Mater. 2001, 13, 4686–4691.
(30) Li, W.; Katz, H. E.; Lovinger, A. J.; Laquindanum, J. G. Chem.
Mater. 1999, 11, 458–465.
(26) Lee, J. K.; Ma, W. L.; Brabec, C. J.; Yuen, J.; Moon, J. S.; Kim,
J. Y.; Lee, K.; Bazan, G. C.; Heeger, A. J. J. Am. Chem. Soc. 2008,
130, 3619–3623.
(27) Chua, L.-L.; Zaumseil, J.; Chang, J.-F.; Ou, E. C. W.; Ho, P. K. H.;
Sirringhaus, H.; Friend, R. H. Nature 2005, 434, 194–199.
(28) Kim, Y.; Cook, S.; Tuladhar, S. M.; Choulis, S. A.; Nelson, J.;
Durrant, J. R.; Bradley, D. D. C.; Giles, M.; McCulloch, I.; Ha,
C.-S.; Ree, M. Nat. Mater. 2006, 5, 197–203.
(31) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122, 4020–
4028.
(32) Littke, A. F.; Schwarz, L.; Fu, G. C. J. Am. Chem. Soc. 2002, 124,
6343–6348.
(33) Okeya, S.; Miyamoto, T.; Ooi, S. I.; Nakamura, Y.; Kawaguchi, S.
Inorg. Chim. Acta. 1980, 45, L135–L137.
(34) Ooi, S. I.; Matsushita, T.; Nishimoto, K.; Okeya, S.; Nakamura,
Y.; Kawaguchi, S. Bull. Chem. Soc. Jpn. 1983, 56, 3297–3301.

1982 Chem. Mater., Vol. 22, No. 6, 2010
Clem et al.
Scheme 2. Synthesis of Polymers
Table 1. Molecular Weights of Polymers Studied
polymer
Mn
Mw
PDI
Pt-T1a
45 kDa
24 kDa
84 kDa
47 kDa
1.9
1.9
F8TZPtb
a SEC in Dichlorobenzene. b SEC in THF (both calibrated with
polystyrene standards).
for triphenylphosphine.35 This bulkiness likely hinders
coordination of the ligand on the Pt center of 5, allowing
the polymerization to proceed without disturbance to the
metal complex. In addition, the high activity of the
Pd(PtBu3)2 catalyst system makes it possible to achieve
high molecular weight polymers via both Stille and
Suzuki coupling routes at 40 °C, a lower reaction tem-
perature than the 80-100 °C frequently used for analo-
gous Pd-catalyzed polymerizations. The reduced reaction
temperature should also contribute to the improved
stability of the platinum complex during polymerization.
The crude polymers were isolated by precipitation into
methanol, followed by Soxhlet extraction with methanol
and hexanes to remove low molecular weight impurities,
and finally chloroform or chlorobenzene to collect the
desired polymer fraction. The chlorobenzene fraction of
Pt-T1 and the chloroform fraction of F8TZPt were
isolated and reprecipitated into methanol for further
study. As shown in Table 1, monomer 5 is an excellent
monomer for the preparation of a variety of Pt-contain-
ing polymers via Stille or Suzuki protocols affording high
molecular weight soluble polymers. 1H NMR and FT-IR
analysis of the polymers indicated that the platinum
complex remains intact after polymerization.
Optical and Electronic Properties. The optical and
electronic properties of 4, F8TZPt, and Pt-T1 were
studied to determine the suitability of these materials
for photovoltaic applications, and to further understand
the effects of extending conjugation through a cyclome-
talated platinum complex. Figure 2a shows the UV-vis
absorbance spectra for 4, Pt-T1, and F8TZPt in dilute
chloroform solution and Figure 2b shows the UV-vis
Figure 2. Overlaid absorption spectra of polymers and small molecule in
chloroform solution (a) and thin film (b): 4 (black), Pt-T1 (blue), and
F8TZPt (red).
absorbance spectra for 4, Pt-T1 and F8TZPt in thin films.
Both polymers exhibit a strong transition at approxi-
mately 350 nm, which, given the similarity between this
peak and the absorption profile of the small molecule 4, is
attributed to the direct excitation of the metal complex.
(35) Tolman, C. A. Chem. Rev. 1977, 77, 313–348.

Article
Chem. Mater., Vol. 22, No. 6, 2010 1983
were used to determine their oxidation and reduction
potentials. The HOMO level was determined by the onset
of oxidation, the LUMO was determined by the onset of
reduction when it was observed and determined by the
difference between the HOMO and the optical bandgap.
The oxidations and reductions observed were all irrever-
sible, which is consistent with previously reported Pt(II)
complexes.37 For the small molecule 4 no reduction peak
was observed, but an irreversible oxidation was observed
at 0.3 V corresponding to a HOMO of -5.4 eV. This
HOMO level is comparable to that of the C∧N ligand 1,
which has a HOMO of -5.3 eV, suggesting that the
HOMO of the platinum complex is primarily based on
the C∧N ligand. The polymer Pt-T1 undergoes an irrever-
sible oxidation at -5.4 eV, and an irreversible reduction
at -3.5 eV, for an electrochemical bandgap of 1.9 eV. The
optical bandgap for Pt-T1 is slightly smaller (1.65 eV).
However for F8TZPt the HOMO is lower at -5.6 eV. The
HOMO of poly(9,9-dioctyfluorene) (PFO) is -5.7 eV,27
suggesting that the HOMO of F8TZPt is at least partly
influenced by the fluorene moiety. An electrochemical
reduction was not observed for F8TZPt, but based on the
difference between the optical bandgap and the HOMO,
the LUMO is estimated to be -3.5 eV. The optical and
electronic properties of all the materials are summarized
in Table 2. Taken together the data indicate that both the
optical bandgap and the HOMO levels of the materials
may be tuned by varying the comonomer, whereas the
LUMO is primarily determined by the platinum complex.
Photoluminescence and Singlet Oxygen Generation.
Organometallic conjugated polymers with large intersys-
tem crossing yields have demonstrated increased photo-
current in photovoltaic devices. This increase in perfor-
mance has been attributed to decreased geminate recom-
bination3 and increased exciton diffusion length due to
the forbidden nature of recombination from the triplet
state.12 However, these previously reported polymers
have bandgaps larger than 2.5 eV, resulting in poor
overlap with the solar spectrum. The polymers presented
here absorb strongly in the visible spectrum; they are also
expected to possess more delocalized excitons than the
polyplatinynes, providing a new platform to study the
effect of a heavy atom on intersystem crossing in con-
jugated polymers.
Figure 3. Previously reported platinum-containing polyfluorenes.36
Both polymers also show strong transitions at longer
wavelengths, with peaks at 610 and 520 nm for solutions
of Pt-T1 and F8TZPt, respectively. These peaks are
attributed to excitations delocalized along the conjugated
polymer chain. The absorption for F8TZPt is signifi-
cantly red-shifted when compared to those of the pre-
viously reported platinated polyfluorenes P-1 and P-2
shown in Figure 3.36
Changing the C∧N ligand in the fluorene copolymers
from 2-phenylpyridine in P1 or P2 to 2-(20-thienyl)thiazole
in F8TZPt causes the lower energy absorption maximum
to shift from 450 nm36 to 520 nm. The small molecule com-
plexes containing 2-phenylpyridine20 and 2-(20-thienyl)-
thiazole have similar absorption spectra, so this larger
bathochromic shift suggests that F8TZPt may have sig-
nificant donor-acceptor character. Further, these results
show that the optical properties of these platinum-contain-
ing polymers can be tuned by modifying either or both the
platinum complex and the comonomer. The UV-vis
spectrum of F8TZPt also exhibits a redshift from 520 to
555 nm when changing from solution to film. This shift in
absorption to longer wavelengths together with the en-
hanced vibronic structure is indicative of increased π-π
stacking in the solid state. In contrast, the spectrum for
Pt-T1 broadens and redshifts slightly going from solution
to the solid state, but does not exhibit vibronic structure.
Based on the onset of absorption in the solid state, the
optical bandgap of F8TZPt is 2.1 eV. The optical band-
gap of Pt-T1 is 1.65 eV, close to the theoretically ideal
bandgap of 1.5 eV for a donor material in a polymer:
PCBM solar cell.18
To investigate the influence of the Pt center on the
F8TZPt and Pt-T1 polymers the emission spectra of all
materials (including monomer precursors) were mea-
sured at both long and short excitation wavelengths. All
experiments were conducted in degassed benzene solu-
tion. At room temperature, 4 exhibits a phosphorescence
peak with an onset at 575 nm (2.16 eV) and a maximum at
660 nm when excited at 370 nm. This peak can be assigned
to a T1 f S0 transition as evidenced by the large Stokes
shift and long lifetime (3.8 μs). The triplet energy of the
small molecule is at a lower energy relative than reported
for several other C∧NPt(O∧O) complexes, which is most
In addition to the absorption properties, the HOMO
and LUMO of the donor material are key parameters that
influence the overall performance of a photovoltaic de-
vice by affecting the efficiency of charge separation and
the maximum attainable open-circuit voltage (VOC).
Cyclic voltammograms (CV) of Pt-T1, F8TZPt, and 4
(36) Thomas, S. W. Y., S.; Swager, T. M. J. Mater. Chem. 2005, 15,
2829–2835.
(37) Kvam, P.-I. P.; Michael, V.; Balashev, Konstantin P.; Songstad,
Jon. Acta Chem. Scand. 1995, 49, 335–343.

1984 Chem. Mater., Vol. 22, No. 6, 2010
Clem et al.
Table 2. Optical and Electronic Properties
sample
λmax, CHCl3 solution
λonset, film
Eg optical (eV)
HOMO (eV)
LUMOa (eV)
LUMOb (eV)
4
Pt-T1
F8TZPt
330 nm
335 nm, 610 nm
375 nm, 520 nm
475 nm
750 nm
600 nm
2.6
1.65
2.1
-5.4
-5.4
-5.6
N/A
-3.5
N/A
-2.8
-3.75
-3.5
a Determined by onset of reduction. b Estimated according to LUMOdHOMO-Eg optical.
Table 3. Singlet Oxygen Generation by Polymers and Model Complex
singlet oxygen
generation
sample
λexcitation
4a
370 nm
370 nm
665 nm
370 nm
665 nm
370 nm
530 nm
yes
Pt-T1a
no
no
yes
no
yes
yes
Pt-T1a
Pt-T1b
Pt-T1b
F8TZPta
F8TZPta
a C6H6 solution. b 2:1 C6H5Cl:C6H6 solution.
decay pathway, possibly via intersystem crossing, may be
more favorable.
In order to probe the formation of triplets before
nonradiative decay, the ability of both solutions and films
of these materials to generate singlet oxygen was explored
as a function of excitation wavelength. The efficiency of
singlet oxygen photosensitization is strongly correlated to
triplet quantum yield in conjugated polymers.38 Molecu-
lar oxygen is able to quench triplet excited states by a
variety of energy transfer pathways, leading to lumine-
scence of singlet oxygen that is readily detected at 1270
nm.40 The results of these experiments are summarized in
Table 3.
At room temperature, an aerated benzene solution of 4
shows strong singlet oxygen emission when excited at
370 nm. In contrast, a benzene solution of Pt-T1 does not
sensitize singlet oxygen formation at either short or long
excitation wavelengths. Because the sample of Pt-T1 used
in these studies was isolated from chlorobenzene, aggre-
gation of the polymer in a weaker solvent such as benzene
was thought to be responsible for quenching. Aggrega-
tion of conjugated polymers has been shown to quench
excitons via an energy transfer process occurring on a
picosecond time scale,41 faster than the nanosecond time
scale typical for intersystem crossing. Indeed, singlet oxy-
gen formation was observed upon excitation at 370 nm
when a more strongly solvating 2:1 mixture of chloro-
benzene and benzene was used to dissolve Pt-T1. The
solvent dependence for singlet oxygen generation using
Pt-T1 as a sensitizer indicates that Pt-T1 is most likely
aggregated in pure benzene solution. However, when
excited at 665 nm, no singlet oxygen generation is obser-
ved from Pt-T1 even in the better solvent system. Excita-
tion of Pt-T1 at 370 nm is analogous to excitation of 4,
which quickly undergoes intersystem crossing due to the
proximity of the heavy Pt atom and the strong orbital
overlap with Pt orbitals in the excited state. Therefore
Figure 4. Absorbance and emission of 4: absorbance (black) and emis-
sion (red).
likely due to the electron rich nature of the C∧N and O∧O
ligands.20 Further, the triplet energy of the complex is
above the triplet energy of poly(3-octylthiophene) (1.65
eV) and below that of PFO (2.3 eV),38 suggesting that in
Pt-T1 energy transfer from the triplet of the complex to
the triplet of the polymer may readily occur. Conver-
sely, energy transfer from the metal complex may be
disfavored for the fluorene copolymer F8TZPt. The
absorbance and photoluminescence of 4 are shown in
Figure 4.
The observed phosphorescence of the complex at room
temperature is in sharp contrast to the reported beha-
vior36 of other diphenyl ketonate-substituted platinum
complexes, which show no emission from the triplet state
as a result of the thermal equilibrium between the C∧N
ligand-centered (3LC) emissive state and the nonemissive
Pt/diphenyl ketonate charge transfer (3CT) state.36 The
significant phosphorescence from 4 at room temperature
indicates that the triplet state of 1 lies sufficiently below
the triplet state of the diphenyl ketonate to inhibit this
nonradiative decay pathway. At room temperature, de-
gassed solutions of both polymers F8TZPt and Pt-T1
show weak emission with small Stokes shifts suggesting
emission from the singlet state. Although phosphores-
cence is not observed for Pt-T1 or F8TZPt, they may still
undergo intersystem crossing to the triplet state followed
by nonradiative decay back to the ground state.39 The
weak fluorescence that is observed suggests that another
(38) Burrows, H. D.; Melo, J. S. d.; Serpa, C.; Arnaut, L. G.; Monkman,
A. P.; Hamblett, I.; Navaratnam, S. J. Chem. Phys. 2001, 115,
9601–9606.
(39) Rachford, A. A.; Goeb, S.; Castellano, F. N. J. Am. Chem. Soc.
2008, 130, 2766–2767.
(40) Schweitzer, C.; Schmidt, R. Chem. Rev. 2003, 103, 1685–1758.
(41) Fakis, M.; Anestopoulos, D.; Giannetas, V.; Persephonis, P.
J. Phys. Chem. B 2006, 110, 24897–24902.

Article
Chem. Mater., Vol. 22, No. 6, 2010 1985
Figure 5. The different means of promotion and inhibition of triplet formation in F8TZPt and Pt-T1 for long wavelength excitations.
triplet formation is facile when exciting at 370 nm. Excit-
ing at 665 nm corresponds to more delocalized excitations
that are spread along the polymer backbones. This more
delocalized excitation clearly does not exhibit enhanced
intersystem crossing. As the exciton becomes more delo-
calized, the Pt atomic orbitals are expected to make a
proportionally smaller contribution to the molecular
orbitals associated with the exciton. Given that the singlet
energy of Pt-T1 is lower in energy than the triplet energy
of 4, it is expected that the singlet of Pt-T1 will not
undergo significant energy transfer to the triplet of 4. Fur-
ther, it is anticipated that nonradiative decay rates in-
crease relative to the intersystem crossing rate as excitons
become more delocalized in a conjugated material.42
Benzene solutions of F8TZPt exhibit singlet oxygen
emission at 1270 nm when excited at either 370 nm or at
530 nm. As was the case for Pt-T1, excitation at 370 nm is
attributed to excitation of the small molecule chromo-
phore, which rapidly undergoes intersystem crossing. In
contrast to Pt-T1, F8TZPt also generates triplet excitons
at longer excitation wavelength. For F8TZPt, the initially
delocalized exciton may ultimately localize to the plati-
num moiety because the platinum repeat unit is lower in
energy than poly(9,9-dioctylfluorene)43 as evidenced by
their relative energy levels. The singlet energy of F8TZPt
is close in energy to the triplet energy of 4, suggesting that
energy transfer from the singlet of F8TZPt to the triplet of
4 can occur readily. This more localized exciton exhibits
strong spin-orbit coupling through the platinum atom,
leading to formation of triplet excitons as evidenced by
sensitized formation of singlet oxygen. The difference in
triplet formation observed between F8TZPt and Pt-T1 at
long excitation wavelength is illustrated schematically in
Figure 5. Red dashed lines represent the excitons initially
formed at long excitation wavelengths, with energy transfer
occurring to the monomer unit of 4 in the case of F8TZPt.
While both polymers are able to sensitize formation of
singlet oxygen in solution, neither polymer was observed
to sensitize singlet oxygen formation in a thin film. The
difference in triplet formation from solution to solid
phase, as evidenced by singlet oxygen generation, most
likely arises as the result of intermolecular interactions
providing a faster decay pathway, analogous to aggre-
gates in solution. For example, regioregular poly(3-
alkyl)thiophenes (P3ATs) have a large triplet yield in
solution (ΦT = 0.77),38 but do not readily form triplets
in a film.44 Formation of triplet states in the solid state is
(42) Wilson, J. S.; Chawdhury, N.; Al-Mandahary, M. R. A.; Younus,
€
M.; Khan, M. S.; Raithby, P. R.; Kohler, A.; Friend, R. H. J. Am.
Chem. Soc. 2001, 123, 9412–9417.
(43) Liao, L. S.; Fung, M. K.; Lee, C. S.; Lee, S. T.; Inbasekaran, M.;
Woo, E. P.; Wu, W. W. Appl. Phys. Lett. 2000, 76, 3582–3584.
€
(44) Jiang, X. M. O., R.; Korovyanko, O.; An, C. P.; Horovitz, B.;
Janssen, R. A. J.; Vardeny, Z. V. Adv. Funct. Mater. 2002, 12, 587–
597.

1986 Chem. Mater., Vol. 22, No. 6, 2010
Table 4. Device Parameters: F8TZPt:PCBM and Pt-T1:PCBM Blend Photovoltaic Devices and SCLC Hole Mobility
Clem et al.
active layer
Voc
Jsc
FF
η
μh(SCLC)
F8TZPt:PCBM
Pt-T1:PCBM
0.38 V
0.65 V
3.5 mA cm-2
5.3 mA cm-2
0.30
0.37
0.40%
1.29%
2.5 ꢀ 10-9 cm2 V-1 -1
s
s
1.0 ꢀ 10-5 cm2 V-1 -1
disfavored for P3ATs in part because the formation of
more delocalized, lower energy excimers is more favor-
able.45 Formation of triplet excitons in P3AT films is also
partly disfavored as a result of decreased twisting between
adjacent thiophene units; an increased twist angle bet-
ween adjacent thiophenes leads to increased spin-orbit
coupling.46 Given that spin-orbit coupling is dominated
by the platinum atom in both F8TZPt and Pt-T1, the
change in triplet formation from solution to solid state in
F8TZPt and Pt-T1 (as evidenced by singlet oxygen gen-
eration) should be primarily the result of lower-energy
excimers forming in the polymer films and not a decrease
in twisting between adjacent thiophene units. Together,
these results suggest that polymers with excitons centered
on molecular orbitals with significant contribution from a
heavy atom and with lowered aggregation in the solid
state would be desirable to generate triplet excitons in a
film for photovoltaic devices.
of Pt-T1 is significantly better than F8TZPt, which is partly
the result of its superior overlap with the solar spectrum.
Based on the electrochemistry described previously, it is also
possible that holes localize at the platinum unit of F8TZPt
and this localization inhibits charge transport in a device.
The HOMO of 4 is higher than the HOMO of F8TZPt,
suggesting that the platinum-containing monomer may act
as a local energy minimum for holes in the device.
In order to better understand potential limiting factors
on photovoltaic performance of polymers containing the
cyclometalated platinum complex, the space charge lim-
ited current (SCLC) mobilities of Pt-T1 and F8TZPt were
measured. Unlike FETs, SCLC mobility determines
charge mobility in the vertical direction under no gate
bias, making it potentially a more relevant measurement
of charge mobility in the context of photovoltaic devices.
The zero-field hole mobility of F8TZPt is 2.5 ꢀ 10-9 cm2
V s-1. The inefficient charge transport measured for this
polymer clearly contributes to its poor photovoltaic per-
formance, and further suggests that the platinum complex
may act as a charge and energy trap in this polymer in
which the platinum complex is copolymerized with the
wider bandgap fluorene unit. In contrast, the zero-field
hole mobility of Pt-T1 was measured as 1 ꢀ 10-5 cm-2
V-1 s-1. The SCLC mobilities for the pristine polymers
and the bulk heterojunction photovoltaic device para-
meters are summarized in Table 4. The striking difference
in SCLC hole mobility between Pt-T1 and F8TZPt
demonstrates that comonomer selection may affect both
charge transport and photophysics of the resulting poly-
mer. The SCLC hole mobility of Pt-T1 is lower than that
of polythiophenes47,48 (10-4 to 10-3 cm2 V-1 s-1), but
higher than the reported SCLC hole mobilities of poly-
Photovoltaic Devices and Charge Mobility. The poly-
mers Pt-T1 and F8TZPt were studied in bulk heterojunction
photovoltaic devices with PCBM to better understand the
influence of the cyclometalated platinum moiety on perfor-
mance. The optimal blending ratio for both F8TZPt and Pt-
T1 was found to be 1:4 (polymer:PCBM). Results optimized
for polymer:PCBM ratio, thickness, and annealing give
power conversion efficiencies of 0.40% for F8TZPt, and
1.29% for Pt-T1 under AM 1.5 illumination at 100 mW
cm-1. Figure 6 shows the performance of the devices without
annealing. Thermal annealing was found to significantly
decrease the performance of the devices, presumably as a
result of excessive phase segregation between polymer and
PCBM. For F8TZPt the Jsc is 3.5 mA cm-2, the Voc is 0.38
V, and the fill factor is 0.30. For Pt-T1 the Jsc is 5.3 mA cm-2
,
the Voc is 0.65 V, and the fill factor is 0.37. The performance
platinynes7 (10-8 to 10-7 cm2 V-1 -1
s ). The improved
SCLC hole mobility observed for Pt-T1 relative to the
polyplatinynes shows that platinum-containing conju-
gated polymers with connectivity via a C∧N ligand are
an attractive alternative route to organometallic poly-
mers for photovoltaics.
Conclusions
A new approach to low bandgap platinum-containing
conjugated polymers based on a platinum monomer with
C∧Nand O∧O diketonate ligands has been investigated. In
these materials the platinum atom is attached adjacent to
(45) Ohkita, H.; Cook, S.; Astuti, Y.; Duffy, W.; Tierney, S.; Zhang, W.;
Heeney, M.; McCulloch, I.; Nelson, J.; Bradley, D. D. C.; Durrant,
J. R. J. Am. Chem. Soc. 2008, 130, 3030–3042.
(46) Beljonne, D.; Shuai, Z.; Pourtois, G.; Bredas, J. L. J. Phys. Chem. A
2001, 105, 3899–3907.
(47) Thompson, B. C.; Kim, B. J.; Kavulak, D. F.; Sivula, K.; Mauldin,
ꢀ
C.; Frechet, J. M. J. Macromolecules 2007, 40, 7425–7428.
(48) Huang, Y.; Wang, Y.; Sang, G.; Zhou, E.; Huo, L.; Liu, Y.; Li, Y.
Figure 6. Photovoltaic Performance of F8TZPt (red) and Pt-T1 (black)
under AM 1.5 illumination.
J. Phys. Chem. B 2008, 112, 13476–13482.

Article
Chem. Mater., Vol. 22, No. 6, 2010 1987
the conjugated backbone, and therefore does not inhibit
exciton delocalization along the polymer chain, providing
a means to study the effect of a heavy atom on diffuse
excitons. Photovoltaic devices fabricated from these
materials yield efficiencies approaching 1.3%, demon-
strating that cyclometalated platinum polymers are an
attractive new class of materials for organic photovol-
taics. Photophysical studies indicate that for some mate-
rials long wavelength excitations experience rapid non-
radiative decay leading to no observable triplet exciton
formation, while other materials exhibit localization of an
initially delocalized exciton. This localization promotes
triplet formation but is disadvantageous to charge trans-
port in a photovoltaic device. The results presented here
suggest that the development of conjugated materials
having both significant triplet yields and overlap with
the visible spectrum in the solid state will require new
materials designed to minimize nonradiative decay path-
ways at longer excitation wavelengths.
Acknowledgment. This work was supported by the Direc-
tor, Office of Science, Office of Basic Energy Sciences,
Materials Sciences and Engineering Division, of the U.S.
Department of Energy under Contract No. DE-AC02-
05CH11231. We also thank Thomas W. Holcombe and Jill
Millstone for helpful discussions.