1,8-naphthalimides in phosphorescent organic LEDs: The interplay between dopant, exciplex, and host emission

Four different 1,8-naphthalimide derivatives were examined in phosphorescent organic light emitting diodes (OLEDs), i.e., 1,8-naphthalimide, N-phenyl-1,8-naphthalimide, N-2,6-dibromophenyl-1,8-naphthalimide (niBr), and bis-NN-1,8-naphthalimide. Photoluminescence from all four naphthalimides have violet-blue fluorescence and phosphorescent bands between 550 and 650 nm (visible at 77 K). While all four compounds gave good glassy films when doped with a phosphorescent dopant, only the niBr films remained glassy for extended periods. OLED studies focused on niBr, with two different architectures. One OLED structure (type 1) had the niBr layer as a doped luminescent layer and an undoped niBr layer to act as a hole-blocking layer. The alternate structure (type 2) utilizes a doped CBP layer as the luminescent layer and the niBr layer is used as a hole-blocking layer only (CBP = 4,4'-N,N-dicarbazolylbiphenyl). Type 1 and 2 OLEDs were prepared with green, yellow, and red emissive phosphorescent dopants (lrppy, btlr, and btplr, respectively). The dopants were organometallic lr complexes, previously shown to give highly efficient OLEDs. Of the three dopants, the btplr-based OLEDs showed the best device performance in both structures (peak efficiencies for type 2: 3.2% and 2.3 lum/W at 6.3 V; type 1: 1.7% and 1.3 lm/W at 6.1 V). The green and yellow dopants gave very similar performance in both type 1 and 2 devices (peak efficiencies are 0.2-0.3%), which were significantly poorer than the btplr-based OLEDs. The emission spectrum of the btlr- and btplr-based devices (type 1 and 2) are the same as the solution photoluminescence spectrum of the dopant alone, while the lrppy device gives a broad red emission line (m,, = 640 nm). The red lrppy.niBr emission line is assigned to an lrppy.niBr exciplex. The type 2 lrppy-based device gave a voltage-dependent spectrum, with the red emission observed at low bias (4-8 V), switching over to strong green emission as the bias was raised. All other devices showed bias-independent spectra. Estimates of HOMO, LUMO, and excited-state energies (dopant, niBr, and exciplex) were used to explain the observed spectral properties of these devices. btplr-based devices emit efficiently from isolated dopant states (external efficiencies = 3.2 2.3 lum/W). Irppy-based devices emit only from exciplex states, with low efficiency (external efficiency = 0.3%). btlr.niBr films have very similar energies for the dopant, exciplex, and niBr triplet states, such that relaxation can go through any of these states, leading to low device efficiency (external efficiency = 0.4%). High device efficiency is achieved only when dopant emission is the dominant pathway for relaxation, since exciplex and niBr triplet states give either weak or no electroluminescence.