Divalent osmium complexes: Synthesis, characterization, strong red phosphorescence, and electrophosphorescence

Article abstract of DOI:10.1021/ja0176705

We report new divalent osmium complexes that feature strong red metal-to-ligand-charge-transfer (MLCT) phosphorescence and electrophosphorescence. The general formula of the complexes is Os(II)(N-N)₂L-L, where N-N is either a bipyridine or a phenanthroline and L-L is either a phosphine or an arsine. New polypyridyl ligands synthesized are 4,4′-di(biphenyl)-2,2′-bipyridine (15) and 4,4′-di(diphenyl ether)-2,2′-bipyridine (16), and the 1,10-phenanthroline derivatives synthesized are 4,7-bis(p-methoxyphen-yl)-1,10-phenanthroline (17), 4,7-bis(p-bromophenyl)-1,10-phenanthroline (18), 4,7-bis(4′-phenoxybiphen-4-yl)-1,10-phenanthroline (19), and 4,7-bis(4-naphth-2-ylphenyl)-1,10-phenanthroline (20). 4,4′-Diphenyl-2,2′-bipyridine (21) and 4,7-diphenyl-1,10-phenanthroline (22) were also used in these studies. Strong π-acid ligands used were 1,2-bis(diphenylarseno) ethane (23), cis- 1,2-bis(diphenylphosphino)ethylene (24), and cis-1,2-vinylenebis(diphenylarsine) (25). Ligand 25 is used for the first time in these types of luminescent osmium complexes. These compounds feature strong MLCT absorption bands in the visible region and strong red phosphorescent emission ranging from 611 to 651 nm, with quantum efficiency up to 45% in ethanol solution at room temperature. Red organic light-emitting diodes (OLEDs) were successfully fabricated by doping the Os(II) complexes into blend of poly(N-vinylcarbazole) (PVK) and 2-tert-butylphenyl-5-biphenyl-1,3,4-oxadiazole (PBD). Brightness over 1400 cd/m² for a double-layer device has been reached, with a turn-on voltage of 8 V. The maximum external quantum efficiency was 0.64%. Commission Internationale de I'Eclairage (CIE) chromaticity coordinates (x, y) of the red electrophosphorescence from the complexes are (0.65, 0.34), which indicates pure red emission.

Full text of DOI:10.1021/ja0176705

Published on Web 11/01/2002
Divalent Osmium Complexes: Synthesis, Characterization,
Strong Red Phosphorescence, and Electrophosphorescence
Brenden Carlson,† Gregory D. Phelan,† Werner Kaminsky,† Larry Dalton,*,†
Xuezhong Jiang,‡ Sen Liu,‡ and Alex K.-Y. Jen‡
Contribution from the Departments of Chemistry and Materials Science and Engineering,
UniVersity of Washington, Seattle, Washington 98195
Received December 3, 2001. Revised Manuscript Received June 3, 2002
Abstract: We report new divalent osmium complexes that feature strong red metal-to-ligand-charge-transfer
(MLCT) phosphorescence and electrophosphorescence. The general formula of the complexes is Os(II)-
(N-N)2L-L, where N-N is either a bipyridine or a phenanthroline and L-L is either a phosphine or an
arsine. New polypyridyl ligands synthesized are 4,4′-di(biphenyl)-2,2′-bipyridine (15) and 4,4′-di(diphenyl
ether)-2,2′-bipyridine (16), and the 1,10-phenanthroline derivatives synthesized are 4,7-bis(p-methoxyphen-
yl)-1,10-phenanthroline (17), 4,7-bis(p-bromophenyl)-1,10-phenanthroline (18), 4,7-bis(4′-phenoxybiphen-
4-yl)-1,10-phenanthroline (19), and 4,7-bis(4-naphth-2-ylphenyl)-1,10-phenanthroline (20). 4,4′-Diphenyl-
2,2′-bipyridine (21) and 4,7-diphenyl-1,10-phenanthroline (22) were also used in these studies. Strong π-acid
ligands used were 1,2-bis(diphenylarseno)ethane (23), cis-1,2-bis(diphenylphosphino)ethylene (24), and
cis-1,2-vinylenebis(diphenylarsine) (25). Ligand 25 is used for the first time in these types of luminescent
osmium complexes. These compounds feature strong MLCT absorption bands in the visible region and
strong red phosphorescent emission ranging from 611 to 651 nm, with quantum efficiency up to 45% in
ethanol solution at room temperature. Red organic light-emitting diodes (OLEDs) were successfully fabricated
by doping the Os(II) complexes into blend of poly(N-vinylcarbazole) (PVK) and 2-tert-butylphenyl-5-biphenyl-
1,3,4-oxadiazole (PBD). Brightness over 1400 cd/m2 for a double-layer device has been reached, with a
turn-on voltage of 8 V. The maximum external quantum efficiency was 0.64%. Commission Internationale
de l’Eclairage (CIE) chromaticity coordinates (x, y) of the red electrophosphorescence from the complexes
are (0.65, 0.34), which indicates pure red emission.
various platinum porphyrin derivatives,7,8 and divalent ruthenium
Introduction
(Ru) complexes such as Ru(II)tris(2,2′-bipyridine) [Ru(II)-
Rapid growth in the use of organic light-emitting devices
(OLED) is expected in the coming years1 due to their potential
application in large-screen flat-panel displays.2-4 For full-color
displays, efficient OLEDs emitting three primary colors, i.e.,
blue, green, and red, are required. Pure red color has been
intrinsically difficult to obtain from conjugated polymers and
small molecules. This is due to difficulty in tuning the energy
levels of the compounds and the energy gap law, which
describes the decrease in luminescence intensity with red-
shifting emission. Because of the expected growth in the
application of OLEDs, there has been much effort put into the
development of novel emitters for use in OLED applications.
Recently there has been interest in the use of triplet-emitting
compounds primarily focusing on trivalent iridium complexes,5,6
tris(26)].9-11 Until recently,12-15 osmium (Os) complexes have
largely been ignored for applications in OLEDs. One reason
for this is that Os tris(polypyridyl) complexes in general have
low quantum yields when compared to their Ru(II) counterparts.
It has been reported that tris(polypyridyl)Os(II) complexes
have very low quantum yields (on the order of 0.1%) and very
short emission lifetimes (on the order of 50 ns).16 The emission
(5) Chen, F.; Yang, Y.; Thompson, M. E.; Kido, J. Appl. Phys. Lett. 2002, 80
(13), 2308.
(6) Lamansky, S.; Djurovich, P.; Murphy, D.; Abdel-Razzaq, F.; Lee, H.;
Adachi, C.; Burrows, P. E.; Forrest, S. R.; Thompson, M. E. J. Am. Chem.
Soc. 2001, 123 (18), 4304.
(7) Kwong, R. C.; Sibley, S.; Dubovoy, T.; Baldo, M.; Forrest, S. R.;
Thompson, M. E. Chem. Mater. 1999, 11 (12), 3709.
(8) Chang, S. E.; He, G.; Chen, F. C.; Guo, T. F.; Yang, Y. Appl. Phys. Lett.
2001, 79, 2088.
(9) Handy, E. S.; Pal, A. J.; Rubner, M. F. J. Am. Chem. Soc. 1999, 121 (14),
3525.
* Corresponding author.
(10) Liu, C.-Y.; Bard, A. J. J. Am. Chem. Soc. 2002, 124 (16), 4190-4191.
(11) Rudmann, H.; Shimada, S.; Rubner, M. F. J. Am. Chem. Soc. 2002, 124
(17), 4918.
† Department of Chemistry.
‡ Department of Materials Science and Engineering.
(1) Allen, K. Solid State Technol. 2001, August, 42.
(2) Burroughes, J. H.; Bradley, D. D. C.; Brown, A. R.; Marks, R. N.; Mackay,
K.; Friend, R. H.; Burn, P. L.; Holmes, A. B. Nature 1990, 347, 539.
(3) Friend, R. H.; Gymer, R. W.; Holmes, A. B.; Burroughes, J. H.; Marks, R.
N.; Taliani, C.; Bradley, D. D. C.; Dos Santos, D. A.; Bredas, J. L.;
Logdlund, M.; Salaneck, W. R. Nature 1999, 397, 121.
(4) Bernius, M. T.; Inbasekaran, M.; O’Brien, J.; Wu, W. AdV. Mater. 2000,
12 (23), 1737.
(12) Carlson, B.; Jiang, X. Z.; Liu, M.; Liu, S.; Jen, A. K.-Y.; Dalton, L. Polym.
Mater. Sci. Eng. 2002, 86, 210.
(13) Jiang, X. Z.; Jen, A. K.-Y.; Carlson, B.; Dalton, L. R. Appl. Phys. Lett.
2002, 80 (5), 713.
(14) Bernhard, S.; Gao, X.; Malliaras, G. G.; Abruna, H. D. AdV. Mater. 2002,
14 (6), 433.
(15) Ma, Y. G.; Zhang, H. Y.; Shen, J. C.; Che, C. M. Synth. Met. 1998, 94,
245.
9
14162
J. AM. CHEM. SOC. 2002, 124, 14162-14172
10.1021/ja0176705 CCC: $22.00 © 2002 American Chemical Society

Divalent Osmium Complexes
A R T I C L E S
Figure 2. Structures of strong π acid ligands.
Figure 3. Synthetic scheme for cis-1,2-vinylenebis(diphenylarsine).
(such as arsines, phosphines, DMSO, CN, and CO; see Figure
2) in combination with polypyridyl ligands may yield Os(II)
complexes with quantum yields approaching 24%. These strong
π acid ligands (arsines and phosphines) strongly back-bond with
the Os, but their σ* or d accepting orbitals are very high in
energy21 compared to the polypyridyl π*; thus, the charge
transfer to ligand bands are dominantly Os (t2g) f bpy (π*).
By reducing the number of polypyridyl ligands, there is less
ability to accept an electron; thus, there is a shift to higher energy
in the absorption and emission bands from the tris(polypyridyl)
counterparts. This makes the mixed-ligand system suitable for
OLED applications by blue-shifting the emission from the
infrared (750-850 nm) for the Os(II)tris(polypyridyl) complexes
to orange-red (600-650 nm) for the mixed ligand system. These
changes in structure cause an increase in radiative decay and
leads to emission yields up to 45%.
Using triplet-based emitting centers in organic and polymer
OLEDs eliminates the 25% limit for maximum internal quantum
efficiency, which is the expected singlet exciton fraction
generated by electrical injection, and potentially allows for
displays with 100% internal quantum efficiency.22 Strong back-
bonding with a metal center, which exhibits a large spin-orbit
coupling constant, facilitates intersystem crossing by breaking
down the spin selection rules, thus leading to stronger triplet-
state emission. Improved phosphorescence provides a possibility
to design a high-efficiency OLED device. Triplet-harvesting red
and green OLEDs based on Pt and Ir complexes have demon-
strated very high external quantum efficiency.23-27 Europium
complexes also show triplet emission and have been used in
red OLEDs.28-30 The characteristic of the lowest excited states
(triplet states) of these heavy-metal complexes can be systemati-
cally varied from largely ligand-centered (LC) to metal-to-
Figure 1. Polypyridyl ligands and numbering position on the ring: 2,2′-
bipyridine is on the left and 1,10-phenanthroline is on the right.
bands of tris(polypyridyl) Os(II) complexes are greatly red-
shifted from their Ru(II) counterparts and occur in the near-
infrared, which makes them unsuitable for use as OLEDs. This
is due to the fact that Os(II) is more easily oxidized than
Ru(II). According to the gap law, decreasing the energy
difference between the excited and ground states enhances
nonradiative decay of the excited state. Thus, red-shifting the
emission band could decrease emission quantum yields; how-
ever, this does not have to be the case.
In designing a synthetic strategy, the above-discussed prob-
lems must be taken into account. It has been repeatedly shown
that the use of phenyl groups on the polypyridyl ligand (Figure
1) greatly increases quantum yields. For example, the quantum
yield for emission of Ru(II)tris(phenanthroline) is 1.9%, while
the yield for Ru(II)tris(22) is 36.6%,17 despite a red shift in
emission from 595 nm for the phenanthroline complex to 610
nm for the 22 complex. Similar results can be observed for
Ru(II) tris complexes of both 21 and 26. The reason for this is
that the phenyl groups undergo a conformational change18 in
the excited state, contributing to and extending to the π system.
In this conformation there is a barrier to C-C bond rotation
due to the extended π system. This reduces C-C bond rotation
and vibration of the substituent, both of which may lead to an
increase in nonradiative decay. This barrier would not exist for
other substituents such as a methyl or other aliphatic groups.
The substitution also replaces a C-H bond vibration para to
the nitrogen, which may quench luminescence. Therefore,
phenyl or other aromatic derivatives of polypyridyls were used
in this study to increase quantum yields.
(19) Kober, E.; Caspar, J.; Sullivan, P.; Meyer, T. Inorg. Chem. 1988, 27, 4587.
(20) Kober, E. M.; Sullivan, B. P.; Dressick, W. J.; Caspar, J. V.; Meyer, T. J.
J. Am. Chem. Soc. 1980, 102 (24), 7383.
(21) Periodic Correlations; Rich, R. L., Ed.; Benjamin: New York, 1965; p 6.
(22) Adachi, C.; Baldo, M.; Thompson, M. E.; Forrest, S. R. J. Appl. Phys.
2001, 90 (10), 5048.
(23) Baldo, M. A.; O’Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.;
Thompson, M. E.; Forrest, S. R. Nature 1998, 395, 151.
(24) Kwong, R. C.; Sibley, S.; Dubovoy, T.; Baldo, M. A.; Forrest, S. R.;
Thompson, M. E. Chem. Mater. 1999, 11, 3709.
(25) Baldo, M. A.; Lamansky, S.; Burrows, P. E.; Thompson, M. E.; Forrest,
S. R. Appl. Phys. Lett. 1999, 75, 4.
(26) Tsutsui, T.; Yang, M.; Yahiro, M.; Nakamura, K.; Watanabe, T.; Tsuji,
T.; Fukuda, Y.; Wakimoto, T.; Miyaguchi, S. Jpn. J. Appl. Phys. 1999,
38, L1502.
(27) Adachi, C.; Baldo, M. A.; Forrest, S. R.; Lamansky, S.; Thompson, M. E.;
Kwong, R. C. Appl. Phys. Lett. 2001, 78, 1622.
(28) Kido, J.; Hayase, H.; Hongawa, K.; Nagai, K.; Okuyama, K. Appl. Phys.
Lett. 1994, 65, 2124.
While Os(II)tris(22) and similar derivatives exhibit some
increase in quantum yield, the emission is still in the infrared,
making them unsuitable for use as OLEDs. Meyer and co-
workers19,20 showed that that the use of strong π acid ligands
(16) Fetterolf, M. L.; Offen, H. W. J. Phys. Chem. 1985, 89 (15), 3320.
(17) Alford, P.; Cook, M.; Lewis, A.; McAuliffe, G.; Skarda, V.; Thomson, A.;
Glasper, J.; Robbins, D. J. Chem. Soc., Perkin Trans. 2 1985, 705.
(18) Damrauer, N.; McCusker, J. K. Inorg. Chem. 1999, 38, 4268.
(29) McGehee, M. D.; Bergstedt, T.; Zhang, C.; Saab, A. P.; O’Regan, M. B.;
Bazan, G. C.; Srdanov, V. I.; Heeger, A. J. AdV. Mater. 1999, 11, 1349.
(30) Jiang, X.; Jen, A. K.-Y.; Huang, D.; Phelan, G.; Londergan, T.; Dalton, L.
Synth. Met. 2002, 125, 331-336.
9
J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002 14163

A R T I C L E S
Carlson et al.
Figure 4. Synthetic scheme for 4,4′-dibromo-2,2′-bipyridine and structures of Os(II) complexes based on resulting ligands.
ligand-charge-transfer (MLCT) character. The triplet emission
character depends on the strength of the back-bonding between
the metal center and the ligand and the relative energies of the
π* (LC) transition vs the dπ* (MLCT) transition. The emission
of Eu complexes (sharp bands at around 615 nm) is completely
inner-shell electronic f to d transitions and is determined by
the energetics of the central Eu3+ ion.31 The emission from
Pt(II) porphyrins is ligand-based, and Ir(III) complexes are
largely ligand-based, though MLCT complexes have been
reported for some Ir complexes as well.25 Luminescence of the
Os(II) complexes being reported is from the MLCT state.
Furthermore, these third-row heavy-metal complexes tend to
be thermally, chemically, and photochemically robust, which
is favorable for device stability. Extremely long device lifetime
has been reported for a triplet OLED device with platinum
octaethylporphyrin (PtOEP) as LC emitting center with a 298
K triplet lifetime of ∼50 µs. The long device lifetime is
speculated to be an intrinsic property of electrophosphorescent
OLEDs, where radiative phosphors significantly shorten the
lifetime of potentially reactive triplet states in the conductive
host material.32 In Os complexes, due to strong back-bonding
from Os to the ligands, the triplet MLCT emission has a very
short lifetime (0.4-2 µs). Although Eu, Pt, and Ir complexes
have been studied to some extent, the application of Os
complexes in OLED has barely been explored.12-15 In this paper
we report the use of various novel Os(II) complexes featuring
strong red emission and high quantum yields in OLED devices.
(32) Burrows, P. E.; Forrest, S. R.; Zhou, T. X.; Michalski, L. Appl. Phys. Lett.
2000, 76, 2493.
(31) Richardson, F. S. Chem. ReV. 1982, 82, 541.
9
14164 J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002

Divalent Osmium Complexes
A R T I C L E S
Experimental Section
Syntheses of Materials. 1,2-Bis(diphenylarseno)ethylene (23) and
cis-1,2-bis(diphenylphosphino)ethane (24) were purchased from either
Aldrich or Alfa and recrystallized 3 times from butanol before use.
4,4′-Diphenyl-2,2′-bipyridine (21) and bathophenanthroline (22) were
purchased from GFS Chemicals. Compound 21 was ∼50% pure and
was purified by repeated washings and recrystallization from dimeth-
ylformamide (DMF) to yield 2.3 g out of 15 g purchased. Potassium
hexachloroosmiate was purchased from Alfa.
cis-1,2-Vinylenebis(diphenylarsine) (25) (Figure 3) was prepared
by a modification of a previous method, resulting in an improved yield
of product.33 Diphenylarsine (Organometallics, 25.00 g, 108.6 mmol)
was used as received and was added to 400 mL of freshly dried (sodium/
benzophenone) tetrahydrofuran (THF). The solution was stirred under
nitrogen and cooled to -78 °C in an acetone/dry ice bath. To this
solution was added n-butyllithium (1.6 M in hexane, 1.05 equiv, 114.1
mmol). The solution was allowed to stir for 1 h. The acetone bath was
then removed and cis-dichloroethylene (TCI-America, 10.66 g, 110.0
mmol) was added. The solution was allowed to slowly warm to 18.5
°C and react overnight. Water was added and the THF was removed
by rotary evaporation under vacuum at 40 °C. The water was removed
by filtration and the solid material was washed with large amounts of
deionized water. The sample was dried under vacuum and then
recrystallized three times from butanol. Yield: 23.97 g (91%). 1H NMR
(DMSO): 7.63 (s, 2H), 7.35 (20H). Anal. Calcd: C, 64.48%; H, 4.58%.
Found: C, 64.10%; H, 4.28%.
4,4′-Dibromo-2,2′-bipyridine (Figure 4) was prepared by an adapta-
tion and combination of the methods reported by Haginiwa34 and
Maerker and Chase.35
4,4′-Di(biphenyl)-2,2′-bipyridine (15) was synthesized from a
palladium-catalyzed Suzuki cross-coupling reaction of 4,4′-dibromo-
2,2′-bipyridine and biphenyl-4-boronic acid neopentyl glycol ester. 1H
NMR (DCCl3) as a Ru(II)(heptafluorobutyrate)2 [Ru(II)(HFB)2] com-
plex: 9.08 (d, 6H), 7.97 (m, 18H), 7.80 (m, 18H), 7.61 (m, 12H), 7.45
(m, 18 H).
Figure 5. Synthetic scheme for aromatic 4,7-substituted 1,10-phenanthro-
lines and structures of Os(II) complexes based on resulting ligands.
4,4′-Bis(diphenyl ether)-2,2′-bipyridine (16) was made following
the same procedure as 15: Yield 0.95 g (30.4%). H NMR (DMSO):
Table 1. Elemental Analysis Data
1
calculated
H
found
H
8.750 (d, 2H), 8.631 (m, 4H), 7.904 (m, 4H), 7.800 (m, 4H), 7.459 (t,
2H), 7.147 (m, 8H).
compd
C
N
C
N
mass
yield (%)
4,7-Bis(p-methoxyphenyl)-1,10-phenanthroline (17) and 4,7-bis-
(p-bromophenyl)-1,10-phenanthroline (18). Syntheses of these com-
pounds have been reported previously.17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
25
54.49 3.28 3.26 54.65 3.36 3.30
57.49 3.34 3.44 58.00 3.41 3.33
54.41 3.42 3.52 54.51 3.48 3.36
58.89 3.60 2.86 59.01 3.59 2.88
60.54 3.59 2.77 59.92 3.34 2.70
60.88 3.73 2.96 61.01 4.02 2.85
58.91 3.51 3.62 58.78 3.52 3.66
55.75 3.32 3.42 56.05 3.34 3.44
61.26 4.36 3.11 61.36 4.41 3.05
64.47 4.47 3.27 64.57 4.65 3.25
61.33 4.25 3.11 61.64 4.33 3.12
51.56 3.31 2.87 52.01 3.38 2.88
647.1a
602.7a
646.1a
830.2a
646.1a
798.2a
626.2a
670.1a
731.2a
686.2a
730.1a
827.9a
90
85
94
90
90
89
92
90
89
95
92
89
93
88
51
30
32
50
64
68
91
Phenanthroline derivatives 19 and 20 were synthesized from 18.
4,7-Bis(4′-phenoxybiphenyl-4-yl)-1,10-phenanthroline (19) was
made following the same procedure as 15 and was recrystallized from
benzene to give a colorless crystalline solid. 1H NMR (DMSO): 9.20
(d, 2H), 8.88-7.65 (m, 18H), 7.45 (t, 4H), 7.25-7.06 (m, 8H).
4,7-Bis(4-naphth-2-ylphenyl)-1,10-phenanthroline (20) was made
following the same procedure as 15 and was recrystallized from DMF
1
to give a colorless crystalline solid. H NMR (DCCl3): 9.29 (d, 2H),
69.38 4.28 2.38 69.07 4.36 2.44 1006.1a
8.14 (s, 2H), 8.03-7.75 (m, 14 H), 7.72-7.63 (m, 6H), 7.57-7.46
(m, 4H).
70.25 4.33 2.56 70.56 4.32 2.48
88.67 5.25 6.08 89.06 5.32 6.21
82.91 4.91 5.69 83.09 4.99 5.77
79.57 5.14 7.14 79.45 5.26 7.20
58.81 2.88 5.71 58.88 2.79 5.75
86.20 4.82 4.19 86.32 4.98 4.28
90.38 4.83 4.79 90.45 4.86 4.81
64.48 4.58 0.00 64.10 4.28 0.08
922.7a
461.6b
493.2b
393.3b
490.9b
669.3b
561.5b
c
General Procedure19 for Synthesis of Osmium Complexes. The
osmium complexes were synthesized by reacting 1.000 g (2.08 mmol)
of K2OsCl6 with 2.05 equiv of polypyridyl (N-N) ligand in 25 mL of
refluxing DMF (Aldrich) under an inert atmosphere for 3 h. The
resulting solution was filtered, washed with DMF, cooled to 0 °C, and
then added dropwise to a water solution of sodium dithionite (2.00 g
in 400 mL) at 0 °C. The resulting purple precipitate of Os(N-N)2Cl2
was filtered and washed with deionized water. Os(N-N)2Cl2 was
reacted with 1.05 equiv of arsine or phosphine ligand (23, 24, or 25)
in a refluxing mixture of 2,2′-ethoxyethoxyethanol (Aldrich) and
a m/2z. b m/z. c Highly fragmented, no m/z.
glycerol (75:25 by volume) for 2 h under an inert atmosphere. The
complexes were precipitated by dropwise addition to a water solution
of the appropriate counterion: tosylate (Ts), triflate (Tf), heptafluo-
robutyrate (HFB), and PF6. The complex structures are shown in Figures
4 and 5. Elemental analysis data, mass spectrometry results, and yields
(33) Mitchener, J. P.; Aguiar, A. M. Org. Prep. Proced. 1969, 1 (4), 259-261.
(34) Haginiwa, J. Yakugaku Zasshi 1955, 6, 731.
(35) Maerker, G.; Chase, F. J. Am. Chem. Soc. 1958, 80, 5.
9
J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002 14165

A R T I C L E S
Carlson et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
[Os(II)bis(4,4′-diphenyl-2,2′-bipyridine) 1,2-bis(diphenylarseno)ethane]2+
[Os(II)(4,4′-diphenyl-2,2′-bipyridine)2 cis-1,2-bis(diphenylphosphino)ethylene]2+(HFB)2
[Os(II)(4,4′-diphenyl-2,2′-bipyridine)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Tf)2
[Os(II)(4,4′-bis(p-diphenyl ether)-2,2′-bipyridine)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Tf)2
[Os(II)(4,4′-bis(p-biphenyl)-2,2′-bipyridine)2 1,2-bis(diphenylarseno)ethane]2+(HFB)2
[Os(II)(4,4′-bis(p-biphenyl)-2,2′-bipyridine)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Tf)2
[Os(II)(bathophenanthroline)2 cis-1,2-bis(diphenylphosphino)ethylene]2+(Tf)2
[Os(II)(bathophenanthroline)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Ts)2
[Os(II)(4,7-bis(p-methoxyphenyl)-1,10-phenanthroline)21,2-bis(diphenylarseno)ethane]2+(Ts)2
[Os(II)(4,7-bis(p-methoxyphenyl)-1,10-phenanthroline)2 cis-1,2-bis(diphenylphosphino)ethylene]2+(Ts)2
[Os(II)(4,7-bis(p-methoxyphenyl)-1,10-phenanthroline)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Ts)2
[Os(II) (4,7-bis(p-bromophenyl)-1,10-phenanthroline)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Ts)2
[Os(II)(4,7-bis(4′-phenoxybiphenyl-4-yl)-1,10-phenanthroline)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Ts)2
[Os(II)(4,7-bis(4-naphth-2-ylphenyl)-1,10-phenanthroline)2 cis-1,2-vinylenebis(diphenylarsine)]2+(Ts)2
are presented in Table 1. More detailed synthesis and characterization
data can be found in the Supporting Information.
Characterization. Elemental analyses were carried out by Oneida
Research Services, Inc, Whitesboro, NY. UV-vis absorption spectra
were measured on a Shimadzu UV-1601 spectrophotometer. Quantita-
tive measurements were obtained by using 1.000 cm path length quartz
cells with absolute ethanol as the solvent. Electrospray mass spectros-
copy was measured on either an Esquire-LC ion trap mass spectrometer
(Bruker and Hewlett-Packard) or an Applied Biosystems Mariner ESI-
TOF mass spectrometer. 1H NMR was carried out on a 200 MHz Bruker
FT-NMR spectrometer.
Emission spectra of ethanol solutions were collected on a Perkin-
Elmer LS50B fluorescence spectrophotometer. The wavelength sensi-
tivity of the instrument was calibrated prior to measurements using a
standard 20 W tungsten lamp of known output. All emission spectra
were corrected to the calibration curve calculated from the known lamp
output. The solutions were degassed using argon for 30 min before the
measurement. Photoluminescence (PL) quantum yields to (10% of the
Os complexes (ΦOs) in ethanol solutions were obtained with Ru(II)
tris-22 dichloride as the standard, which has a known quantum yield
of 0.366, from the following equation:36
abs Ru area Os
area Ru abs Os
ΦOs
)
× 36.6%
(1)
Figure 6. Structure of tetraphenyldiamine containing perfluorocyclobutane
polymer (BTPD-PFCB) and FIB7 polymer.
where abs is the absorbance of the sample and area is the integration
of the emission curve. Samples were excited through the LC state at
280 nm with absorption of 0.050. Temperature for the measurements
was 25 ( 2 °C.
use. A layer of ∼40 nm thick hole-transport material (HTL), a
tetraphenyldiamine-containing perfluorocyclobutane polymer (BTPD-
PFCB) (Figure 6), was first fabricated by spin-coating the monomer
from its 1,2-dichloroethane (DCE) solution and annealing at 225 °C
under nitrogen atmosphere.40 Then a layer of 3.0 wt % Os complex
doped blend of poly(N-vinylcarbazole) and 2-(t-butyl)phenyl-5-biphen-
yl-1,3,4-oxadiazole (PVK:PBD, 70:30 by weight) or poly(2-vinylnaph-
thalene) and PBD (PVN:PBD, 70:30 by weight) was spin-coated from
the corresponding DCE solution (∼12 mg/mL) at 2000 rpm. A layer
of 30-nm-thick Ca was vacuum-deposited at below 1 × 10-6 Torr
through a mask in an argon-protected evaporator, and another layer of
120-nm-thick Ag was deposited as a protective layer. All testing was
carried out in air at room temperature. Current-voltage characteristics
were measured on a Hewlett-Packard 4155B semiconductor parameter
analyzer. EL spectra were measured with an Oriel InstaSpec IV CCD
camera or a Photo Research PR650 colorimeter. The EL emission power
was measured by use of a Newport 2835-C multifunction optical meter
in combination with a calibrated photodiode. Brightness was calculated
from the emission power and EL spectra of the devices, assuming
Lambertian distribution of the EL emission.41 Thickness of the films
was measured on a Sloan Dektak 3030 profilometer.
For lifetime measurement, the Os complexes were dissolved into a
paint solution of FIB737,38 polymer (Figure 6) and trifluorotoluene. FIB7
is a very photostable polymer and does not absorb light above 235
nm. The solution was spray-painted onto a polished aluminum plate
and dried at 50 °C. The samples were put into a sample holder,38 placed
under vacuum, and excited at 338 nm with a nitrogen pulse laser. The
luminescence decay was monitored and the lifetime was calculated.39
Samples for photodegradation and temperature dependence were
prepared following the same procedures as for the lifetime measure-
ments detailed above. A tungsten-halogen lamp filtered by a FIV-
026 band-pass filter was used as the excitation source (400 nm, FWHM
) 20 nm). The Power density was 925 µW cm-2. The emission intensity
was monitored with a photomultiplier tube.37,38 The illumination for
degradation was continuous and the temperature was set to 25 °C, while
for temperature dependence the shutter time was 1 s and the temperature
ranged from 5 to 50 °C.
Device Fabrication and Testing. OLED devices were fabricated
on ITO substrates that were cleaned and treated with O2 plasma before
(36) Wallace, L.; Rillema, D. P. Inorg. Chem. 1993, 32, 3836.
(37) Puklin, E.; Carlson, B.; Gouin, S.; Costin, C.; Green, E.; Ponomarev, S.;
Tanji, H.; Gouterman, M. J. Appl. Polym. Sci. 2000, 77 (13), 2795.
(38) Carlson, B.; Gouterman, M. U.S. Patent 5,965,642, 1999.
(39) Coyle, L. Dissertation, University of Washington, 1999.
(40) Jiang, X. Z.; Liu, S.; Liu, M. S.; Ma, H.; Jen, A. K.-Y. Appl. Phys. Lett.
2000, 76, 2985.
(41) Greenham, N. C.; Friend, R. H.; Bradley, D. D. C. AdV. Mater. 1994, 6,
491.
9
14166 J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002

Divalent Osmium Complexes
A R T I C L E S
Table 2. Crystallographic Data for the Structures Provided
cis-1,2-vinylenebis-
property
(diphenylarsine)
complex 3
complex 12
empirical formula
formula weight
temperature, K
wavelength, Å
crystal/color
crystal system
space group
unit cell dimensions
a, Å
C26H22As2
484.28
130(2)
0.710 73
prism/clear
monoclinic
Pc (no. 7)
C84H70As2N4O7OsS2
1651.60
130(2)
C90H66As2Br4Cl6N4O6OsS2
2235.92
130(2)
0.710 73
0.710 73
plate/dark red
monoclinic
P21/c (no. 14)
prism/dark red
orthorhombic
P21212 (no. 18)
12.5920(4)
5.6160(2)
17.0320(6)
90
117.1651(13)
90
1071.59(6)
1.501
14 847/4081
18.329(5)
17.240(10)
22.048(9)
90
99.055(11)
90
6880(5)
1.594
14 1517/7042
14.5150(2)
26.0710(3)
11.1480(7)
90
90
90
4218.6(3)
1.760
35 455/8380
b, Å
c, Å
R, deg
â, deg
γ, deg
volume, Å3
density, Mg/m3
reflections collected/unique
final R indices [I > 2σ(I)]
R1
0.0376
0.0851
0.0667
0.0661
0.0554
0.1171
wR2
R indices (all data)
R1
wR2
0.0431
0.0872
0.1780
0.0840
0.0853
0.1276
Table 3. Bond Lengths for Free Ligand and Complexes 3 and 12a
Table 4. C-As-C Bond Angles for Complexes for Free Ligand
and Complexes 3 and 12a
complex
N1
N2
N3
N4
As1
As2
arsenic
C(Ph)−As−C(br)
C(Ph)−As−C(br)
C(Ph)−As−C(Ph)
3
12
2.054(9) 2.094(8) 2.011(11) 2.057(9) 2.3997(15) 2.420(15)
2.096(7) 2.091(7) 2.096(7) 2.091(7) 2.4165(9) 2.4165(9)
As1
As1a
As2
As2a
3
98.9(2)
88.35(14)
96.9(2)
86.21(2)
103.7(5)
105.9(5)
96.1(2)
82.15(15)
97.6(2)
81.23(15)
103.6(5)
101.0(5)
96.88(2)
81.87(2)
95.81(2)
89.82(2)
101.3(4)
101.8(6)
arsenic
C(br)
C(Ph)
C(Ph)
CdC(br)
As1
As1a
As2
As2a
3
1.956(5)
1.8285
1.955(5)
1.8159
1.932(11)
1.895(10)
1.971(4)
2.171(5)
1.972(5)
2.149(5)
1.930(10)
1.946(9)
1.973(5)
2.330(6)
1.976(4)
2.145(5)
1.945(11)
1.790(3)
1.308(7)
1.4581
1.308(7)
1.4581
1.296(11)
1.374(19)
12
complex
C(Ph)−As−Os
C(Ph)−As−Os
C(br)−As−Os
3
12
122.7(3)
120.3(4)
115.7(3)
116.3(3)
107.8(3)
109.3(3)
12
a Bond lengths are given in angstroms; br ) bridge; Ph ) phenyl.
a Bond angles are given in degrees; br ) bridge, Ph ) phenyl.
X-ray Diffraction. Crystals of cis-1,2-vinylenebis(diphenylarsine)
(25) were grown by dissolving the compound in hot 1-butanol and
slowly cooling to room temperature. Crystals of complexes 3 and 12
were grown by dissolving the complex as the Ts salt in chloroform
and then adding an equal volume of ethyl acetate. The solutions were
then allowed to slowly evaporate. The crystals were mounted in random
orientation on a glass fiber on a Kappa CCD diffractometer, Mo KR
(λ ) 0.710 73 Å). Measurements were performed at 130 ( 2 K. Cell
constants and an orientation matrix for data collections were obtained
by least-squares refinements of the diffraction data from up to 141 517
full and partial reflections. The structures were solved by direct methods
with SIR97 and DIRDIFF, provided by the refinement package
MaXus.42 Missing atoms were found by difference Fourier synthesis.
The non-hydrogen atoms were refined with anisotropic temperature
factors. Scattering factors are from Waasmaier and Kirfel.43 The
structures were refined with SHELXL-97, and Ortep plots were
generated with ORTEP32.44 Tables 2-4 summarize the crystal data,
collection information, and refinement data for these structures.
Figure 7. Crystal structure with 50% probability spheres for cis-1,2-
vinylenebis(diphenylarsine).
Results and Discussion
X-ray Diffraction. Crystal structures (Figures 7-9) for 25
and complexes 3 and 12 are given. Compound 25 exhibits
disorder of 88.9(2)% for As1C1C2As2 and 11.1(2)% for
As1aC1aC2aAs2a. In addition, this noncentrosymmetric struc-
ture is complete by racemerized Flack-enantiopole parameter
equal to 0.503(14). The bond angles around the arsenic vary
greatly between the two conformations. As main-group elements
beyond the second row of the periodic table are reluctant to
hybridize, it is expected that the bonds that arsenic forms would
(42) Mackay, S.; Edwards, C.; Henderson, A.; Gilmore, C.; Stewart, N.;
Shankland, K.; Donald, A. MaXus, University of Glasgow, Scotland, 1997.
(43) Waasmaier, D.; Kirfel, A. Acta Crystallogr. A 1995, 51, 416.
(44) Farrugia, L. J. Appl. Cryst. 1997, 30, 565.
9
J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002 14167

A R T I C L E S
Carlson et al.
angles for 12 were as follows: N(2)#1-Os(1)-N(2),
169.8(3)°; N(1)-Os(1)-As(1)#1, 178.1(2)°; N(1)#1-Os(1)-
As(1), 178.1(2)°. A perfect octahedron would require the bite
angles to be 90° and the trans bond angles to be 180°. Complex
12 is somewhat closer to ideal than 3 in that more of its trans
bonds are closer to 180°. There was very little trans effect of
the arsenic on the N-Os bond as the Os-N bond length is
largely unaffected by the Os-N bond being trans to N or to
As. This shows that the electron density being donated to
osmium by the arsenic is in balance with removal through back-
bonding. The bond angles around the arsenic deviate when
complexed to the osmium from that of the uncoordinated ligand.
The carbon-arsenic-carbon bond angles in both complexes 3
and 12 are broader than that of the free ligand and range from
101.0(5)° to 105.9(5)°. The osmium-arsenic-carbon bond
angles range from 107.8(3)° to 122.7(3)°. The structure seems
to be migrating to a distorted tetrahedron around the arsenic. A
question arises if the arsenic is hybridizing when complexed to
osmium or if steric forces are “pushing” the phenyl groups into
a new geometry around the arsenic. Hybridization would be
due to localization of the arsenic 4s “lone pair” in forming a
bond with osmium. Examination of reported arsenic compounds
gives the following bond angles: AsH3, 91.8°; AsF3, 95.97°;
AsCl3, 97.7°; AsBr3, 97.7°; AsI3, 99.7°; and As(Ph)3, 100.1°.45
Previous reports on the arsenic bond angles in this series and
other similar arsenic compounds have concluded that hybridiza-
tion of the arsenic is indeed taking place.46-48 The arsenic bond
angles in complexes 3 and 12 are larger than the series listed
above and larger than that of the free ligand. However, the
smallest of the arsenic bond angles in complexes 3 and 12,
101.0(5)°, is close to what is observed for As(Ph)3. The bond
angle around the ethylene bridge (examined for signs of strain)
on the ligand range were observed at 124.3(4)° and 123.2(4)°
for structure 1 of the crystal and 120.1° and 123.7° for the
second structure in the crystal lattice. These bond angles also
remain nearly unchanged in the complexes 3 and 12 and were
122.7(3)° and 122.6(4)° for complex 3 and 118.9(3)° for
complex 12. It would seem that very little strain was placed
upon the ethylene bridge by coordinating to osmium. If the
arsine ligand is under strain, it is being observed only in the
phenyl groups. However, there is space between the phenyl
groups on the arsenic and the polypyridyl structures. It would
be this type of interaction that would lead to steric effects that
would change the geometry around arsenic. Evidence for this
is librative disorder of the phenyl groups on the two arsenic
atoms in structure 12. From discussions in previous reports and
the crystallographic evidence provided, it is proposed that
p-character and possibly d character of the 4s on the arsenic is
increasing due to the localization of electron pair contained
therein in forming a bond with osmium. A structure study of a
model compound that minimizes steric strain such as (Cl)4-
Os(IV)[25] could be used to determine if the arsenic is
hybridizing.
Figure 8. Crystal structure with 50% probability spheres for complex 3.
Counterions and hydrogen have been removed for clarity.
Figure 9. Crystal structure with 50% probability spheres for complex 12.
Counterions and hydrogen have been removed for clarity.
be made with unhybridized p orbitals. The resulting bond angles
would be 90° around arsenic. The conformer with probability
88.9(2)% had arsenic bond angles that ranged from 95.81(2)°
to 98.9(2)°, or greater than expected with an average at
97.02(6)°. The conformer comprising 11.1(2)% of the structure
of ligand 25 had bond angles of 81.23(15)-89.82(2)° with an
average at 84.94(6)°, or smaller than expected. The average bond
angle was 90.98(6)°, close to the 90° theoretically expected for
organoarsenic compounds.
The structure for complexes 3 and 12 compares to a distorted
octahedron. The bite angles for the ligands (21) in complex 3
were observed at 75.0(4)° and 77.0(4)° while the bite angle for
the arsine ligand (25) was observed at 83.08(5)°. For complex
12, the bite angles for both 18 were observed at 77.7(3)° while
the arsine ligand was 83.28(5)°. The trans bond angles for
complex 3 were as follows: N2-Os-As1, 172.0(3)°; N4-Os-
As2, 177.7(3)°; and N1-Os-N3, 166.8(4)°. The trans bond
UV-Visible and Photoluminescence Spectroscopy. Table
5 and Figures 10-13 summarize absorption and emission
(45) Norman, N. C. Chemistry of Arsenic, Antimony, and Bismuth; Blackie
Academic: London, 1998.
(46) Svergun, V. I.; Babushkina, T. A.; Shvedova, G. N.; Kudryavtseva, L. V.;
Semin, G. K. IzV. Akad. Nauk SSSR, Ser. Khim. 1970, 2, 482.
(47) Grechishishkin, V. S.; Yusupov, M. Z. Zh. Strukt. Khim. 1973, 14 (6), 1028.
(48) Siddiqui, R. A.; Raj, P.; Saxena, A. K.; Dixit, S. K. Synth. React. Inorg.
Met.-Org. Chem. 1996, 26 (7), 1189.
9
14168 J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002

Divalent Osmium Complexes
A R T I C L E S
Table 5. Absorption Maxima for the Various Ligand and Charge Transfer Bands, Extinction Coefficients (ꢀ), and Emission Properties of the
Os(II) Complexes Being Reported
complex
LC (nm) (ꢀ)
1MLCT (nm) (ꢀ)
3MLCT (nm) (ꢀ)
emissiona
τ,b ns
Φc
1
2
3
4
5
6
7
8
262 (52 000), 306 (60 000)
301 (68 000)
304 (70 000)
301 (62 000), 330 (63 000)
308 (89 000), 331 (87 000)
304 (91 000), 334 (88 000)
282 (66 000)
405 (16 000)
382 (17 000)
397 (18 000)
402 (22 000)
408 (26 000)
402 (28 000)
378 (21 000)
391 (22 000)
397 (36 000)
364 (50 000)
391 (39 000)
391 (28 000)
393 (44 000)
374 (52 000)
520 (3700)
484 (4500)
512 (4700)
518 (4300)
650
623
640
645
651
643
613
623
635
611
629
635
637
637
410
520
460
470
430
0.19
0.23
0.25
0.27
0.22
0.28
0.33
0.38
0.27
0.36
0.45
0.39
0.40
0.41
508 (10 000)
522 (11 000)
507 (sh, 4100)
524 (sh, 4000)
505 (sh, 9000)
487 (sh, 9000)
500 (sh, 8000)
500 (sh, 6300)
500 (sh, 7500)
500 (sh, 8400)
450
1810
1530
1200
1970
1550
1400
1310
1260
279 (69 000)
9
274 (66 000), 331 (45 000)
269 (69 000), 327 (52 000)
273 (71 000), 329 (43 000)
290 (75 000)
273 (101 000), 344 (49 000)
283 (108 000), 331 (51 000)
10
11
12
13
14
a Emission peak. b Luminescence lifetime. c Luminescence quantum yield.
Figure 10. Absorption of complex 3 and emission (640 nm) at three
different excitations, 300, 400, and 500 nm. The emission curves are offset
for clarity.
Figure 12. Absorption (ꢀ) and emission of complexes 9 (O), 10 (b), and
11 (s).
ligand π-π* transition bands exhibit the strongest ꢀ that is
>60 000 L‚cm-1‚mol-1. Absorption bands that occur at roughly
1
3
390 and 500 nm are the MLCT and spin-forbidden MLCT
bands. These are weaker bands with ꢀ of 16 000 and 3700
L‚cm-1‚mol-1 for complex 1. Extending the conjugation length
of the polypyridyl ligand increases the strength of all absorption
bands. Complexes 1, 2, and 3 are based upon ligand 21. These
have ꢀ of 17 000 L‚cm-1‚mol-1 for the 1MLCT, 5000
3
L‚cm-1‚mol-1 for MLCT, and 70 000 L‚cm-1‚mol-1 for the
LC state. Additional substitution or extending the conjugation
length as in complexes 4, 5, and 6 affords ꢀ of 28 000
1
L‚cm-1‚mol-1 for the MLCT, 11 000 L‚cm-1‚mol-1 for the
3MLCT, and 91 000 L‚cm-1‚mol-1 for the LC state. The same
trends were observed for the phenanthroline-containing com-
plexes (7, 8 vs 13, 14). With the extended conjugation in these
systems, additional LC bands were observed between the main
Figure 11. Absorption (ꢀ) and emission of complexes 6 (O), 7 (b), and 8
(s).
1
polypyridyl LC peak and the MLCT transition.
properties of the complexes. Absorbance and emission spectra
for complexes 1, 2, and 3 are reported in our previous report.13
Complexes with bipyridyl ligands exhibit an adsorption band
at >300 nm, while complexes with phenanthroline ligands
exhibit an adsorption band at <290 nm. These bands are
attributed to the π-π* transition centered on the ligand. These
At room temperature the complexes feature smooth, unstruc-
tured exponential Gaussian emission typical of MLCT emission.
The emission of the arsine complexes is to the red of the
emission of the phosphine complexes. This offers the ability to
tune the emission of Os(II) complexes by the use of different
ligands to the specific application. Emission lifetime for the
9
J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002 14169

A R T I C L E S
Carlson et al.
Figure 14. Photodegradation of Ru(22)3 and 7 in FIB7 polymer. Samples
were illuminated for 90 min at 925 µW cm-2
.
complexes have a reported emission at 600 nm. Here we are
reporting significantly red-shifted complexes (630-640 nm)
with significantly greater quantum yields in seeming defiance
of the energy gap law. The use of phenyl derivatives of 2,2′-
bipyridyls and 1,10-phenanthrolines has been shown to increase
quantum yields.17 The radiative and nonradiative rate terms in
the quantum efficiencies of Ru(II) complexes has also been
discussed before.17 An explanation of this phenomenon may
be the reduction of bond vibrations and rotations that quench
luminescence. C-H, N-H, and O-H bond vibrations and C-C
bond rotations are well-known to quench luminescence. The
increase in quantum yield by reduction in C-H bond vibrations
is observable in this research report where complexes based
upon ligand 23 (complexes 1, 5, and 9) have significantly
weaker quantum yield than similar complexes based upon ligand
25 (complexes 3, 6, and 11). In the crystal structures given for
complexes 3 and 12, the phenyl groups are rotated out of the
plane of the main polypyridyl structure. It has been shown18
that phenyl groups on polypyridyl complexes become coplanar
with the main polypyridyl structure in the excited state. This
extends the π-system of the ligand, which hinders C-C bond
rotation in the excited state. Thus, the use of phenyl groups
reduces pathways (such as C-H bond vibration and C-C bond
rotation) of nonradiative deactivation of the excited state.
Photodegradation. Illustrated in Figure 14 is a degradation
comparison of Os complex 7 and Ru complex [Ru(22)3PF6].
After an exposure period of 90 min at 925 mW cm-2 and 400
nm, the Os complex photo degraded 1%, while the Ru complex
photodegraded 20% over the same time frame. This result can
be explained by the energy levels of the metal-centered state
for Ru(II) and Os(II) d6 octahedral complexes36,49 as illustrated
in Figure 15. Going down the periodic table affords an increase
in the energy of the MC state of tris(polypyridyl) complexes of
Fe2+ to Os2+. This is due to an increase in 10Dq value. For
Fe2+, the MC state is lowest in energy; thus, the polypyridyl
complexes of Fe2+ are nonluminescent and are easily formed
and degrade. Ru2+ has the MLCT state as the lowest in energy
and the MC state is intermediate in energy between the LC and
MLCT states; thus Ru2+ complexes are luminescent. However,
the low-lying MC state may be thermally populated, making it
more susceptible to degradation and leading to loss in quantum
yield with increasing temperature. Like Ru2+, Os2+ has the
MLCT state as the lowest energy excited state. As Os2+ is more
easily oxidized than Ru2+, the MLCT state is generally lower
Figure 13. Absorption (ꢀ) and emission of complexes 12 (b), 13 (O), and
14 (s).
bipyridyl complexes (1-6) were observed at roughly 450 ns,
while the lifetimes for the phenanthroline complexes (7-14)
were 1.2-2.0 µs. The difference in lifetime between the
bipyridine- and phenanthroline-containing complexes may be
due to the extended ring system of the phenanthroline. The
complexes with arsine ligands (23, 25) had shorter emission
lifetimes than those with phosphine ligands (24). This may be
due to the fact that arsenic is a heavier atom than phosphorus,
thus increasing the rate of intersystem crossing and rate of
phosphorescence. There was some effect of extending the
conjugation length on the outer portion of the polypyridyl
ligands and emission lifetime, as complexes with extended π
systems (5, 6, 13, and 14) had slightly shorter emission lifetimes.
The large spin-orbit coupling constant of osmium (∼3500
cm-1), and strong back-bonding between ligand and metal is
resulting is short emission lifetime of the complexes with minor
contributions from the arsenic, phosphorus, and the extended
π system. Complex 11 was the most efficient photoluminescence
emitter with 45% quantum yield (Φ), which is given by the
following expression:
kp
Φ )
(2)
kp + knr + kq[Q]
where kp is the rate of radiative decay, knr is the nonradiative
decay rate, and kq is the quenching rate. A common quencher
of luminescence is oxygen.
The 1,10-phenanthroline complexes had quantum yields in
excess of 30%. The complexes with ligand 25 in general have
greater quantum yields than the complexes with ligand 24. This
may be due to the heavy atom effect. From the lifetime data,
complexes with arsine ligands have shorter lifetimes than
complexes with phosphine ligands. The heavier arsenic increases
spin-orbit coupling, which increases the rate of intersystem
crossing. This may have the effect of making kp more competi-
tive with knr (eq 2) in the arsine complexes, thus increasing
quantum yields of the complexes with ligand 25. Osmium(II)
complexes have been reported with quantum yields up to 24%
in the literature.14 The reported complexes are based upon ligand
24 and other non-phenyl-substituted bipyridyl ligands. The
(49) Demas, J. N.; DeGraff, B. A. Anal. Chem. 1991, 63 (17), 829.
9
14170 J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002

Divalent Osmium Complexes
A R T I C L E S
Table 6. Temperature Dependencea of Osmium Complexes vs
Ruthenium Complexes at 5-50 °C
complex
X-
%/°C (760 Torr)
%/°C (vac)
1
2
3
7
8
HFB
Tf
Tf
Ts
Ts
-0.09
-0.31
-0.10
-0.33
-0.23
-1.2
-0.08
-0.28
-0.08
-0.28
-0.20
-0.95
-0.44
Ru(22)3
PF6
PF6
Ru(22)2DCBuBPYb
-0.48
a Loss of quantum yield with increasing temperature. b [Ruthenium-
bis(22)-4,4′-bis(carboxybutyl ester)-2,2′-bipyridine]2+. This compound has
emission at 673 nm. We thank John Bullock (Central Washington State
University) and Tim Hance for their generous supply of this compound.
faster rate than their Os(II) counterparts. Another consequence
of this is that Ru2+ compounds exhibit greater loss in emission
quantum yield with increasing temperature as shown in Table
6, where complexes 1, 2, 3, 7, and 8 are compared to two
Ru(II) complexes, one that emits at 610 nm and one that emits
at 673 nm. As temperature is increased, the MC state becomes
more thermally populated, which leads to loss in quantum yield
since the MC state relaxes by nonradiative means. As the MLCT
is separated energetically from the MC state by red-shifting
emission (or by increasing the energy of the MC state), the MC
state becomes less populated, which leads to lower emission
temperature dependence as observed in Table 6. The MC state
of the Os(II) complexes is so high that the most blue-shifted
(613 nm) Os(II) emitter (in Table 6) is less temperature-
dependent than the most red-shifted (673 nm) Ru(II) emitter.
Figure 15. Lowest triplet state orderings for group VIII divalent cations
in octahedral symmetry. LC or MLCT as the lowest excited state leads to
a luminescent compound; MC state as the lowest excited state leads to a
nonluminescent compound.
Organic LEDs. The Os complexes demonstrate good phos-
phorescence efficiency and short excited-state lifetime (see Table
5), which are very desirable properties for light-emitting diode
applications. To study the device performances of these
complexes, double-layer devices were fabricated by doping the
Os complexes at a weight ratio of 3 wt % into PVK:PBD or
PVN:PBD blends. BTPD-PFCB was used as the hole-transport-
ing layer. At the doping level of 3 wt %, the EL spectra of the
devices are almost identical to the PL spectra of the Os
complexes. No emission from the host materials was observed.
Table 7 summarizes the performance of the devices. As can be
seen, even with a simple double-layer structure and PVK:PBD
as the host (type I devices), relatively good performances can
be achieved. Among type I devices, the best external quantum
efficiency of 0.78% was obtained from complex 11 with Ts as
the counterion, while the highest brightness of 1430 cd/m2 was
obtained from complex 10 with Ts as the counterion. In general,
the complexes utilizing arsine ligand 25 have better quantum
yields than those with ligand 24. Interestingly, the device
efficiency follows this trend. It has been found that the Os
complexes trap both electrons and holes, which facilitates the
direct recombination of holes and electrons on the complex sites
and benefits the device efficiency.13 It should be noted that the
counterion used in the complexes also affects the device
performance, presumably through affecting the charge trap/
transport property of the complex, thereby providing an ad-
ditional way of tuning the device properties. Better external
quantum effciencies were achieved from type II devices, where
PVN:PBD was used as the host and excitation was transferred
from the host more efficiently to the Os complex dopants,
Figure 16. Schematic representation of two limiting cases for the relative
positions of the MC and LC (or MLCT) excited states.
3
3
3
in energy than the MLCT state for Ru2+ compounds. The MC
for Os2+ compounds state is very high in energy, much higher
than even the LC. Much more thermal energy is required to
populate the MC state in polypyridyl Os2+ compounds. There-
fore the state is more difficult to populate, slowing the rate of
photodegradation. Samples of Os(II) complexes 1, 2, 3, 7, and
8 in solution have been kept exposed to sunlight for over a year
without change, while their Ru(II) counterparts degrade in just
a few days.
Figure 1650 illustrates the consequence of the positioning of
the various excited states of the Os(II) or Ru(II) complexes. If
the MLCT or LC state is lowest in energy, the resulting
compound may be luminescent. However, if the MC state is
lowest in energy, the resulting compound is not luminescent.
Figure 16 illustrates that for the MLCT or LC states there is
very little distortion along the metal-ligand bond axis, while
for the MC state there is severe lengthening of the metal-ligand
bonds. This is the result of the antibonding nature of the MC
state. Once the MC state is populated, the complex may relax
back to the ground state by nonradiative means, or the complex
may lose a ligand and photoproducts result. The MC state is
lower in energy for Ru(II) than for Os(II); hence, it is more
easily populated. As a result, Ru(II) compounds degrade at a
(50) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Zelewsky,
A. Coord. Chem. ReV. 1988, 84, 85.
9
J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002 14171

A R T I C L E S
Carlson et al.
Table 7. Performance of Os Complex Doped LEDs
efficiency of 1.9 cd/A, while that of a complex 12 doped PVN:
PBD device is 0.79%. Substitution of Br in complex 12 with a
naphthyl group in complex 14 almost doubles the extinction
coefficient of 1MLCT absorption of the Os complex (see Table
5). A larger extinction coefficient contributes to a larger spectral
overlap integral between the emission spectrum of PVN:PBD
host and the absorption of complex 14. This also contributes to
a more efficient energy transfer from PVN:PBD to complex
14. However, we note that compared to the devices based on
PVN:PBD, the devices based on PVK:PBD have a lower turn-
on voltage and higher brightness, mainly due to the better hole
transport property of PVK. Nevertheless, the device data clearly
demonstrate that Os complexes, when carefully designed, can
be good candidates for light-emitting device applications.
OLEDs have been reported for ruthenium complexes with
similar emission profiles as the Os(II) complexes being reported
here. Various polymer host devices with emissions between 611
and 665 nm gave brighntness in the range of 200-650 cd/m2
and efficiency in the range of 0.08-2.5%.51-53 In comparison,
we report Os(II) devices that give efficiency up to 2.2% with a
brightness of 870 cd/m2 at 637 nm emission, or brightness of
1210 cd/m2 with efficiency of 0.78% at 629 nm emission. The
reported ruthenium complexes have emission quantum yields
of 3.6-6.2%, significantly less than the reported osmium
complexes here. At 611 nm emission we report a device with
brighntness of over 1400 cd/m2 and efficiency of 0.48%.
2
Os complex
X-
devicea
V b (V)
Bmaxc (cd/m )
ηmaxd (%)
1
1
HFB
HFB
Ts
I
9.3
7.5
8.7
7.6
7.5
8.4
6.7
6.9
6.4
9.4
7.6
8.0
7.5
6.0
12.2
8.2
7.8
7.0
14.2
310
260
970
410
725
600
799
750
1030
460
1430
1210
760
710
470
780
960
1090
870
0.64
0.27
0.27
0.60
0.42
0.32
0.28
0.20
0.31
0.39
0.48
0.78
0.29
0.19
0.79
0.31
0.45
0.48
2.2
2
I
2
I
3
HFB
Ts
I
3
I
4
Tf
I
6
Tf
I
7
Tf
I
8
Tf
I
9
Ts
I
10
11
12
12
12
13
14
14
14
Ts
I
Ts
I
Ts
I
PF6
PF6
Ts
I
II
I
Ts
PF6
PF6
I
I
II
a ITO/BTPD-PFCB/Os complex, PVK:PBD (∼45 nm)/Ca (type I), or
ITO/BTPD-PFCB/Os complex, PVN:PBD(∼45 nm)/Ca (type II). b Voltage
needed for brightness of 1 cd/m2. c Maximum brightness. d Maximum
external quantum efficiency.
Conclusion
Red-emitting Os(II) complexes were synthesized for light-
emitting applications from different ligand systems and coun-
terions. These compounds feature strong MLCT absorption
bands in the visible region and strong red phosphorescent
emission ranging from 611 to 651 nm, with quantum yields up
to 45%. The electronic structure and emission properties of the
Os(II) complexes can be modified by changing the ligand
structures. Electrophosphorescent devices were demonstrated by
use of the Os complexes with doped PVK:PBD or PVN:PBD
as the emitting layer. When PVK:PBD was used as the host
matrix, brightness of over 1400 cd/m2 was achieved. The best
external quantum efficiency of 2.2%, which corresponds to a
photometric efficiency of 1.9 cd/A, was achieved when PVN:
PBD was used as the host matrix. It was found that the
counterion also affects the performance of the complexes and
devices, providing an additional way of tuning the material and
device properties.
Figure 17. Electroluminescent spectra of ITO/PVK:PBD/Ca and ITO/PVN:
PBD/Ca devices.
presumably through a Fo¨rster mechanism. Figure 17 shows the
EL spectra of ITO/PVK:PBD/Ca and ITO/PVN:PBD/Ca de-
vices. As can be seen, compared with the EL emission of PVK:
PBD host, the EL emission of PVN:PBD host peaks at shorter
wavelength and provides a much better spectral overlap with
the absorption spectra of the Os complexes. In Fo¨rster energy
transfer, the energy transfer rate is proportional to the integral
of the spectral overlap between the emission of the energy donor
and the absorption of the energy acceptor. Therefore, PVN:PBD
can transfer energy more efficiently to the Os complexes.
Consequently, devices with a PVN:PBD host are more efficient
than devices with a PVK:PBD host. It is very interesting to
further compare the performance of complexes 12 and 14 doped
type II devices. The maximum efficiency of a complex 14 doped
PVN:PBD device is 2.2%, corresponding to a photometric
Acknowledgment. We thank the Air Force Office of Scien-
tific Research (AFOSR) for support through the MURI and the
DURIP program. A.K.-Y.J. thanks the Boeing-Johnson Founda-
tion for financial support. B.C. thanks Professor Martin Gou-
terman for his help and guidance for several years. We thank
Martin Sadilek and Ross Lawrence for their help with mass
spectrometry.
Supporting Information Available: Synthesis and charac-
terization data (PDF); X-ray crystallographic data (CIF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(51) Chan, W. K.; Ng, P. K.; Gong, X.; Hou, S. Appl. Phys. Lett. 1999, 75,
3920.
(52) Lee, J.; Yoo, D.; Rubner, M. F. Chem. Mater. 1997, 9, 1710.
(53) Rudmann, H.; Rubner, M. F. J. Appl. Phys. 2001, 90 (9), 4338.
JA0176705
9
14172 J. AM. CHEM. SOC. VOL. 124, NO. 47, 2002