3.1.1.

Exciplex emission

The radiative recombination of CT excitons is known as exciplex emission, as shown in Fig. 3(a). Exciplex emission always shows a long wavelength and a broad spectrum because a small energy difference between the HOMO of HTL and the LUMO of ETL. In addition, exciplex emission is radiative recombination of singlet CT excitons and therefore generally exhibits low efficiency.^24,25

Due to a large separation distance between the electrons and holes, CT excitons have been demonstrated recently to have a very small singlet–triplet energy splitting, which enables an efficient transition from nonradiative triplet states to radiative singlet states via a reverse-intersystem-crossing (RISC) process.^26,27 Therefore, fluorescent OLEDs based on exciplex emission can harvest both singlet and triplet excitons through prompt and delayed fluorescent decay channels. Indeed, a high external quantum efficiency (EQE) ($\sim 15.4\%$) fluorescent OLED based on exciplex emission has been reported recently by mixing a hole transporting material with a bipolar host material to form CT excitons.²⁸ Recently, a simple heterojunction device with a stack of $p$-type HTL and an $n$-type ETL also has been shown to yield an efficient ($\mathrm{EQE}\sim 12.02\%$) exciplex emission.²⁹

It is known that the energy of exciton is lower than the energy difference in the unbound electron and hole pair. That is, the energy of exciplex emission, which originates from direct radiative recombination of CT excitons, should be lower than the energy difference between the HOMO of HTL and the LUMO of ETL. As shown in Fig. 3(b), however, the actual energies of exciplex emission are higher than the energy differences between the HOMOs of HTLs and the LUMOs of ETLs. Empirically, the exciplex emission energy can be expressed as follows:²¹

where $L_{\mathrm{A}}$ is the LUMO of acceptor (ETL), $H_{\mathrm{D}}$ is the HOMO of donor (HTL), and the constant $\alpha$ is $0.20\pm 0.15\mathrm{eV}$, which is comparable to a reported value $0.15\pm 0.10\mathrm{eV}$ measured from exciplex emission in solution.³⁰

3.1.2.

Energy transfer

For a guest–host device, CT excitons can transfer their energies to chromophores when the excitonic energies of chromophores on either side of heterojunction are lower than that of the interfacial CT excitons.^31,32 The excitons on the chromophores will recombine radiatively. As shown in Fig. 4, there are two mechanisms through which CT excitons can transfer energy to dopants: Förster energy transfer and Dexter energy transfer.

Förster energy transfer is also called resonance energy transfer, which occurs through the Coulombic dipole–dipole interaction. In this case, an exciton generated on a donor molecule recombines to the ground state and transfers its energy to excite an acceptor. Förster energy transfer is a long-range process and mainly emerges between the two singlet excitons. The rate of Förster energy transfer is given by³³

where $R$ is the distance between donor and acceptor, $\tau$ is the natural lifetime of the exciton, $R_0$ is the Förster radius, which describes the critical transfer distance that excitation transfer and spontaneous deactivation of the donor are of equal probability, and is given by the equation³³ where $N_{\mathrm{A}}$ is the Avogadro’s number, $\kappa$ is the dipole orientation factor, ${\mathrm{\Phi }}_{\mathrm{D}}$ is the fluorescent quantum yield of the donor, $n$ is the index of refraction of the medium containing the donor and acceptor, and $J_{\mathrm{DA}}$ is the overlap integral of the normalized donor emission spectrum and acceptor absorption spectrum.

Dexter energy transfer is through exciton hopping, which involves electron hopping from a donor’s LUMO to an acceptor’s LUMO and electron hopping from the acceptor’s HOMO to the donor’s HOMO. Dexter energy transfer occurs within a short distance due to its overlap requirement of the wavefunction between the donor and acceptor. The rate constant for Dexter energy transfer is given by³⁴

where $K$ describes specific orbital interactions and $L$ is the van der Waals radii of the donor and acceptor.

These energy transfer mechanisms have been demonstrated to be a successful strategy for producing high-efficiency OLEDs. For a phosphorescent dopant, both singlet and triplet of CT excitons transfer energy to the phosphorescent dopant via Förster and Dexter processes, and then the singlets of phosphorescent dopant convert quickly into triplets via an efficient intersystem crossing (ISC). Therefore, both singlet and triplet of CT excitons can excite phosphorescent dopant yielding possibly 100% internal quantum efficiency. Indeed, high EQEs of 29.5%, 32.3%, and 35.6% for blue, green, red phosphorescent OLEDs, respectively, have been reported through effective use of the Förster and Dexter energy transfer from CT excitons that are formed by mixing HTL and ETL as a cohost.^35–37 In the case of fluorescent dopants, both Förster and Dexter energy transfer also occur as well, also producing singlet and triplet excitons on the fluorescent dopant. But only singlet excitons can be used for light emission, triplet excitons decay by nonradiative transitions. Therefore, fluorescent-doped devices in general show a low efficiency. It is worth noting that shown in Fig. 4, however, triplet of CT excitons may upconvert into singlet of CT excitons via RISC process and then transfer to singlets of dopant via Förster energy transfer channel before it transfer to triplet of dopants through Dexter energy transfer process. Indeed, a high-efficiency ($\mathrm{EQE}\sim 14.5\%$) fluorescent-dopant-based OLED has been achieved recently by using an extremely low dopant concentration to minimize Dexter energy transfer from the triplet of CT excitons to triplet of fluorescent dopant.³⁸

3.2.1.

Phonon emission

In addition to the radiative recombination pathways, CT excitons may decay nonradiatively via phonon emission. In fact, nonradiative recombination is a common process in semiconductor devices. CT excitons may recombine nonradiatively following an energy-gap law, which describes the rate of nonradiative decay increases exponentially with decreasing difference in energy between the ground and excited states. The rate constant for nonradiative decay is given by³⁹

where $E^{*}$ is the energy of excited state, $k_0$ and $A$ are the constants related to phonon.

Indeed, an extremely high nonradiative recombination of CT excitons has been demonstrated for $\mathrm{NPB}/{\mathrm{C}}_{60}$ heterojunction recently.⁴⁰ As shown in Fig. 5(a), a unique energy level alignment and a large energy barrier lead to electrons and holes accumulation at $\mathrm{NPB}/{\mathrm{C}}_{60}$ interface and form CT excitons, which recombine nonradiatively by phonon emission at a very fast rate. This extremely fast CT excitons recombination makes $\mathrm{NPB}/{\mathrm{C}}_{60}$ heterojunction behave like an ideal Ohmic contact and thus leads to a very high rectifying characteristics from the electrode/organic contacts [see Fig. 5(b)].

3.2.2.

Auger recombination

The CT exciton recombination may take another nonradiative pathway to release its energy by exciting an adjacent electron to higher energy states.^17–21 This process is referred to as Auger recombination. Auger recombination is a common electronic process in inorganic semiconductors and quantum dots, in particular when the carrier concentration is high.^41,42 In organic semiconductors, Auger recombination may involve bimolecular recombination, which CT excitons recombine at the heterojunction interface by transferring their energy to excite adjacent electrons to higher energy levels.²²

As shown in Fig. 6, there are two possible types of Auger recombination at organic heterojunctions: (i) biexciton Auger recombination, where a ground-state biexciton decay into an excited state mono-exciton; (ii) negative trion Auger recombination, where a three-bound-particle (one hole and two electrons, i.e., charged exciton) state decay into one excited electron. The trion states were predicted theoretically and have been observed experimentally in other types of semiconductors such as quantum dots and quantum wells.^43,44 Moreover, the trion was observed in carbon nanotubes with a large energy separation (0.1 to 0.2 eV) from the bright excitons.⁴⁵ This value is consistent with the 0.22 eV energy difference between Auger recombination and exciplex emission at NPB/B4PyPPM heterojunction device.²¹ The Auger recombination rate at organic heterojunctions is given by²¹

where $C_a$ is the coefficient of Auger recombination, $n$ and $p$ are the concentration of electron and hole, respectively.

Figure 7 shows a schematic energy diagram of negative trion Auger recombination at an organic heterojunction. The CT exciton recombination at a heterojunction interface releases its energy to excite a LUMO electron to higher energy states. This electron on a high energy state may either relax down a lower energy state via inelastic collisions or ballistically injects into the LUMO of donor. The energy offset condition for Auger-electron injection is given as²⁰

where $E_g$ is the HOMO–LUMO energy gap of donor, $L_{\mathrm{A}}$ and $L_{\mathrm{D}}$ are the LUMO of acceptor and donor, respectively. Under this condition, the turn-on voltage of OLED for Auger-electron injections is no longer limited by the energy gap of the emitter molecule. The lowest possible turn-on voltage is²⁰ where $L_{\mathrm{A}}$ is the LUMO of acceptor and $H_{\mathrm{D}}$ is the HOMO of donor.