A Quantum-Chemical Insight into the Role of Charge-Transfer States in Organic Emitters for Electroluminescence

Abstract

Open AccessCCS ChemistryMINI REVIEW1 Aug 2020A Quantum-Chemical Insight into the Role of Charge-Transfer States in Organic Emitters for Electroluminescence Xian-Kai Chen Xian-Kai Chen *Corresponding author: E-mail Address: [email protected] Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ 85721 https://doi.org/10.31635/ccschem.020.202000281 SectionsAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail The rapid progress of the design and development of efficient organic electroluminescent emitters without heavy transition metals has been achieved in recent experimental investigations. In this minireview, based on his recent theoretical works on purely organic emitters, the author provides a quantum-chemical insight into the role of charge-transfer (CT) electronic states in electroluminescence exploiting organic thermally activated delayed fluorescence (TADF) and radical emitters. The author discusses the electronic structure factors impacting the efficiencies of organic TADF and radical emitters, such as the characteristics of the singlet, doublet, and triplet excited states, the singlet–triplet energy gap, the spin–orbit coupling, the solid-state polarization effect, and the second-order spin–vibronic coupling mechanism in spin–flip processes. The author hopes this minireview will be useful for further understanding of the photophysical mechanism and designing more novel and efficient organic electroluminescent materials (not limited to TADF and radical emitters). Download figure Download PowerPoint Introduction Since the pioneering work of Tang and VanSlyke1 on organic light-emitting diodes (OLEDs), much attention has been paid to OLEDs by the academic and industrial communities. In an OLED device, singlet excitons and triplet excitons are generated with a 1∶3 ratio via electric injection, due to quantum spin statistics.2 In early devices exploiting conventional organic fluorescence emitters, the radiative decay of singlet excitons leads to prompt fluorescence emission, while triplet excitons are usually deactivated through nonradiative decay since the phosphorescence emission of triplet excitons is spin forbidden due to negligible spin–orbit couplings (SOCs) in purely organic molecules.3 Assuming a light outcoupling factor of 20%, this means that the maximum external quantum efficiency (EQE) of devices is limited to approximately 5%. To overcome the 5% efficiency limit of early fluorescent OLED devices, many efforts have been devoted to efficiently utilizing triplet excitons. In 1998, organometallic coordination complexes with heavy transition metals (e.g., iridium or platinum) were exploited as emitters in the active layers.4,5 In these OLEDs, strong SOCs (up to several hundreds of cm−1) induced by heavy metal atoms enable efficient phosphorescence emission, thus harvesting the triplet excitons generated via electric injection.3,6 Their internal quantum efficiencies (IQE) can reach up to ca. 100%,7 and EQEs over 30% have now been demonstrated.8 While these phosphorescent emitters are exploited in the OLEDs currently on the market, there is strong impetus to exploit organic emitters without heavy metals, in order to: (1) lower the cost associated with heavy metals and (2) increase the ability to tune the electronic and optical properties via the synthetic flexibility associated with organic compounds.9–11 In 2012, Uoyama et al.12 proposed a promising strategy to harvest the triplet excitons in purely organic molecular materials, based on the thermally activated delayed fluorescence (TADF) mechanism. In the TADF mechanism, low-lying (or "cold") triplet excitons are upconverted as singlet excitons via thermal activation (see Figure 1a). A large number of TADF molecular emitters or exciplexes have been designed by combining electron donor (D) units with electron acceptor (A) units through either chemical bonds or intermolecular interactions.9,13 Impressive photophysical properties and device performances have been reported, with IQEs also reaching nearly 100%.12,14 Recently, Li et al.15 and Yao et al.16 introduced another promising pathway, that is, the so-called hot-exciton mechanism, toward harvesting triplet excitons generated under electric injection. Different from the TADF mechanism, in the hot-exciton mechanism upper-lying (or "hot") triplet excitons are converted as singlet excitons by reverse intersystem crossing (RISC; Figure 1b). In these OLED devices, the D–A-type molecular emitters exploiting the hot-exciton mechanism have also achieved high ratios (up to approximately 100%) of singlet exciton formation. Figure 1 | Schematic diagram of the electroluminescence processes of (a) organic TADF, (b) hot-exciton, and (c) radical emitters. Here, S, singlet state; T, triplet state; D, doublet state; PF, prompt fluorescence; DF, delayed fluorescence; HEIF, hot-exciton induced fluorescence; IC, internal conversion; RISC: reverse intersystem crossing; TADF, thermally activated delayed fluorescence. Note that for radical emitters, the electroluminescent mechanism remains unclear, and the simplest case is thus displayed here. Download figure Download PowerPoint In addition to organic electroluminescence arising from radiative singlet and triplet excitons, in 2015, Peng et al.17 exploited a neutral radical molecule with D–A-type chemical structure as an emitter in the active layer. For organic radical emitters with a single unpaired electron, doublet excitons generated via electric injection radiatively emit light and then decay to the doublet ground state (GS; Figure 1c). Recently, they further boosted the EQE efficiencies of radical-based devices with deep red/near infrared emissions to a record high value of 27%.18 For either D–A-type molecular emitters or D–A exciplexes employed in purely organic electroluminescence devices, the (intramolecular or intermolecular) charge-transfer (CT) electronic state plays a critical role in their photophysical properties and device performances. In this minireview, the author has primarily summarized recent quantum-chemical (QC) calculation results on organic TADF and radical emitters to provide a theoretical insight, at the molecular level, into the role of CT states in the electroluminescent mechanism and molecular design for purely OLED materials. Organic TADF Materials TADF is based on a thermally activated RISC from the triplet to singlet excited state.12,19,20 In the context of electroluminescence, the singlet excitons generated via RISC processes can decay radiatively to the electronic GS, leading to delayed fluorescence. Thus, for TADF materials, a fast RISC rate (kRISC) is the key to harvest triplet excitons as efficiently as possible. Nature and energies of excited states and SOCs If RISC processes involve only the lowest triplet excited state (T1) and the lowest singlet excited state (S1), the kRISC of RISC process from T1 to S1 state is expressed, via the Marcus–Levich–Jortner rate equation derived from first-order time-dependent perturbation theory,3,21 as: k RISC = 2 π ℏ ( | H SO S 1 T 1 | ) 2 1 4 π λ M k B T Σ n = 0 ∞ S n e − s n ! exp ⁡ ( − ( ▵ E S 1 − T 1 + λ M + n ℏ ω eff ) 2 4 λ M k B T ) (1)where | H SO S 1 T 1 | is the SOC between S1 and T1 states, ΔES1–T1 is the energy gap between adiabatic excitation energies of S1 and T1 states, λM denotes the Marcus reorganization energy related to the classical low-frequency vibrations; S is the Huang–Rhys factor corresponding to a nonclassical high-frequency vibrational mode with an effective energy of ℏωeff, kB denotes the Boltzmann constant, and T denotes temperature. Equation (1) suggests that small ΔES1–T1 and large | H SO S 1 T 1 | values are required to facilitate the T1 → S1 RISC process for TADF materials. Due to the purely organic feature of TADF emitters, the SOCs are small, commonly on the order of ≤1 cm−1.21,22 However, as shown in Equation (1), an increase in SOC by 10 times leads to an enhancement in kRISC by 100 times. Besides the SOC mechanism, RISC processes induced by hyperfine coupling23,24 or the Δg mechanism25,26 were also reported in other relevant investigations. Herein, the author only discusses the RISC processes induced by SOC. The minimization of ΔES1–T1 is a strategy commonly adopted to accelerate RISC processes. To achieve small ΔES1–T1, TADF materials were initially designed as either D–A-type molecules generally showing a large twist angle between D and A units,27 or D–A complexes.28 In these systems, the wave functions of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) are spatially separated, meaning that HOMO and LUMO are localized on the D and A units, respectively. When the main electronic configurations of both S1 and T1 states correspond to an electronic transition from the HOMO to the LUMO, the electron exchange energy (KHL) vanishes, and ΔES1–T1 is very small since ΔES1–T1=2KHL.20 As a consequence, both the S1 and T1 states are dominated by the characteristics of CT excitation from HOMO to LUMO. However, the negative outcome of this initial molecular design strategy is a significantly reduced SOC between two CT states with singlet or triplet spin20,29 (since excited-state characteristics of the S1 and T1 states must be maintained differently in order to conserve the total angular momentum). For example, in 2′,7′-bis(di-p-tolylamino)-9,9′-spirobi-[fluorene]-2,7-dicarbonitrile (Spiro-CN), the electron and hole natural transition orbitals (NTOs) describing the S1 and T1 states are separated in space due to a twist angle of nearly 90° between the diphenylamine-fluorene and fluorene-dicarbonitrile segments, see Figure 2a. This means that both the S1 and T1 states have remarkable CT excitation characteristics, thus leading to small ΔES1–T1 (0.01 eV) and vanishing | H SO S 1 T 1 | (0.01 cm−1) values simultaneously.21 In the recent Perspective paper of Olivier et al.29, they also quantitatively discussed the impact of CT degrees of S1 and T1 states on | H SO S 1 T 1 | and ΔES1–T1. In addition, the CT excitation characteristic also has a significant impact on radiative and nonradiative decay rates of the emissive S1 state, as discussed in detail in our earlier paper.20 Figure 2 | Chemical structures of (a) Spiro-CN, (b) ACRFLCN, and (c) 4CzIPN; calculated NTO analysis for the S1, T1, and T2 excited states [here the predominant hole (h) and electron (e) wave functions with the largest weight v are shown]. Figures 2a and 2b were adapted with permission from ref 21. Copyright 2017 American Chemical Society. Figure 2c was adapted with permission from ref 30. Copyright 2019 Springer Nature Limited. Download figure Download PowerPoint To reliably describe the characteristics and energies of excited states, it is critical to correctly select QC approaches such as density functional theory (DFT) or wave function-based electron correlation methods. For TADF molecules composed of several tens or even over a hundred atoms, DFT approaches are most widely exploited. As discussed in earlier works,31 the application of popular hybrid DFT functionals, such as B3LYP or PBE0, incorrectly output the characteristics and energies of excited states, because of the many-electron self-interaction errors.32 Long-range corrected DFT functionals (such as ωB97XD or LC-ωPBE) are thus strongly recommended.33 In D–A-type molecules, the T1 states usually have a local excitation (LE) or hybrid LE–CT excitation characteristic, while the S1 states show a CT-dominant excitation characteristic.21,34 For instance, in 10-phenyl-10H-spiro[acridine-9,9-fluorene]-2,7-dicarbonitrile (ACRFLCN), the calculation results obtained at the optimally tuned long-range corrected LC-ωPBE level indicate that the hole and electron NTOs describing the T1 state are both localized on the cyano-substituted fluorene unit, which is supported by the clear vibronic feature of its phosphorescence spectroscopy,35 while the S1 and T2 states similarly maintain a substantial CT excitation characteristic, see Figure 2b. As a result, the ΔES1–T1 for ACRFLCN is large, approximately 0.24 eV.35 However, the application of the B3LYP functional underestimates the CT state energy and incorrectly outputs the CT excitation characteristic of the T1 state; as a consequence, the ΔES1–T1 calculated by B3LYP is significantly underestimated to 0.02 eV.31 Although long-range corrected DFT functionals provide reliable understanding of excited-state properties for most TADF molecules with a twisted D–A structure, these standard DFT functionals usually fail in the description of molecular electronic structures showing strong electronic correlation.36 It is worth noting that planar TADF emitters with multiresonant effects first reported by Hatakeyama et al.37 show impressive electronic structure and photophysical properties, such as small ΔES1–T1 (<0.2 eV), high photoluminescence efficiency (up to 90%), and narrow emission spectra (full width at half maximum of approximately 30 nm). For these TADF molecules, time-dependent DFT (TDDFT) approaches incorporating only one-electron excitation significantly overestimate ΔES1–T1 values (to 0.4–0.6 eV). The QC calculations of Pershin et al.38 by a high-level coupled-cluster approach successfully reproduced experimental data well, and their calculation results demonstrate that these impressive properties result from short-range reorganization of the electron density taking place upon electronic excitation of these multiresonant structures. In addition, the calculation results of de Silva39 by high-level electron-correlated approaches [e.g., EOM-CCSD (equation of motion coupled cluster with single and double excitations) or ADC(2) (second-order algebraic-diagrammatic construction)] also demonstrate that electron correlation due to a substantial contribution of double-electron excitations leads to a negative value of ΔES1–T1, as shown in the case of the cycl[3.3.3]azine molecule. Actually, OLED devices exploiting TADF emitters with multiresonant effects usually show significant efficiency roll-off, due to slow kRISC. In future investigations, a thorough understanding of their ultrafast photophysical properties is required either experimentally or in theory, to further improve kRISC. Besides the lowest-lying S1 and T1 states, the upper-lying excited states (e.g., T2) also play an important role in RISC processes,22 which will be discussed in the next section. For isolated TADF molecules, the electronic structures of molecular fragments impact excitation characteristics and energies of T2 states in the whole TADF molecules. For example, for the ACRFLCN and Spiro-CN with the same fluorene-dicarbonitrile acceptor fragment, the different donor fragments lead to completely different excitation characteristics of their T1 and T2 states. In Spiro-CN, the T1 is a CT-dominant state and the T2 state is localized at the donor fragment, while in ACRFLCN, the situation is reversed (Figures 2a and 2b). In fact, this results from the competition between the triplet exciton energies of the individual molecular fragments and the CT triplet-state energies of the D–A molecules. Apparently, since the energies of CT states with a permanent dipole are sensitive to the polarity of surrounding media, the excitation characteristics and energies of the T1 and T2 states are significantly impacted by the choice of solvent or host matrix. Such a big difference in the excitation characteristics of the triplet states impacts not only the ΔES1–T1 and ΔET2–T1 but also the SOC values. For Spiro-CN, since the S1 and T1 states have a CT-dominant nature, the S1–T1 SOC is small. On the other hand, the difference in the nature of the S1 and T2 excited states gives rise to a much bigger SOC (0.46 cm−1). Different from the cases of simple D–A(–D) molecules, our recent investigation on complex TADF molecules, for example, 4CzIPN (Figure 2c) with multiple donors, has demonstrated that the T2 state of the whole molecule corresponds to the T1 state of a partial molecular structure, with the hybrid LE–CT excitation characteristic rather than LE or CT dominance.30 In such TADF emitters, multiple donor or acceptor moieties lead to the appearance of a dense manifold of triplet states with different excitation characters, which facilitates the spin–flip processes. In addition to the isolated molecules, the excited-states properties of TADF D–A complexes strongly depend on the electronic structures of single D and A components and their molecular packing, which is similar to the case of organic solar cells.40 In fact, the energy of D–A intermolecular CT state (ECT) is related to the energy offset between the ionization potential (IPD, or more crudely HOMO) of the donor, the electron affinity (EAA or LUMO) of the acceptor, as well as the electron–hole Coulombic interaction (ECoul) impacted by the D–A intermolecular distance, with the relationship ECT = IPD − EAA + ECoul. Such an intermolecular CT state is hybridized with the LE state on the D or A component, via the electronic coupling between D and A molecules that is significantly impacted by intermolecular packing. Thus, the hybridization extent between the intermolecular CT state and LE state governs the excitation characteristics and energies of the adiabatic excited states (e.g., S1, S2, T1, and T2), thus ΔES1–T1 and ΔET2–T1, as well as the SOC values (see Figures 3a and 3b). As shown in earlier investigations on solid-state D–A exciplexes,41,42 the energies and oscillator strength of fluorescent emission, singlet–triplet energy gaps, and kRISC constants are all impacted by molecular structures and energy-level alignments of D and A components and D–A intermolecular packing (e.g., intermolecular distance and orientation). Recently, Kabe and Adachi43 achieved organic long persistent luminescence (OLPL) by using a D–A complex that lasts for more than 1 h at room temperature, and they attributed OLPL to radiative recombination of long-lived charge-separated states, rather than TADF. In 1952, Mulliken44 proposed a quantum-mechanical theory to rationalize the formation of intermolecular CT states in D–A complexes, a number of theoretical and experimental investigations were dedicated to understanding D–A complexes.45–49 However, so far the mechanism of RISC from triplet to singlet and the appearance of long-lived charge-separated states in D–A complexes remain unclear. Thus, a systematic investigation on the above scientific problems is critical to establish the relationship among the electronic structure properties of single D and A components, D–A intermolecular packings, and photophysical properties of D–A complexes. Figure 3 | (a) Energy-level diagrams in a model including the four diabatic electronic states 1CT, 3CT, 1LE, and 3LE. 1/3CT denotes the pure charge-transfer singlet/triplet state corresponding to a full electron transfer from D to A, and 1/3LE denotes the LE state fully localized on an individual component. (b) Blend of TCTA and B4PyMPM molecules, simulated by molecular dynamics (MD) simulation (left); one random dimer was extracted from the blend (right). Figure 3b was adapted with permission from ref 41. Copyright 2018 American Chemical Society. Download figure Download PowerPoint For both D–A isolated molecules and D–A complexes, environmental polarization has a significant effect on the characteristics and excitation energies of excited states as well as SOC values. In an earlier investigation, our results showed that environmental polarization implicitly described by dielectric constant causes the S1 and T1 states to be dominated by CT excitations, thus giving rise to a substantial reduction in ΔES1–T1.50 Recently, Tu el al.51 further employed the combination of quantum-mechanics and embedded-charge approaches to explicitly account for solid-state polarization effects. Their results demonstrated that the CT contribution to the S1 state is relatively larger compared with the T1 state, and, thus, the S1 energy is more stabilized by electronic polarization, leading to smaller ΔES1–T1 in solids. Moreover, the SOC can simultaneously be large due to the different characteristics of the S1 and T1 states in solids. These theoretical investigations provide one of the reasons why the light-emitting performances of TADF emitters in OLED devices are sensitive to the choice of host In addition, the of excited-state properties (such as fluorescent and phosphorescent for TADF emitters upon solvent polarity is also a useful experimental to their excitation characteristics and photophysical spin–vibronic coupling mechanism in RISC When the RISC processes involve only the T1 and S1 states, then the Marcus–Levich–Jortner rate Equation (1) can be employed to the As by et their results that for ΔES1–T1 the are fast on the order of (Figure as shown in most of the efficient TADF molecules. Figure 4 | (a) kRISC constant as a of ΔES1–T1 for different values of the SOC and Marcus reorganization energy (b) Chemical structures of of the for the low-frequency vibrational for and potential energy of the excited T1, S1, and states in with permission from and respectively. Copyright 2017 and American Chemical Society. Download figure Download PowerPoint In addition, Figure also that in the case of ΔES1–T1 the will be significantly reduced or even small, that the triplet harvesting via RISC is However, recent have reported that a number of small molecules (such as in Figure that have large ΔES1–T1 values of show clear earlier work introduced the second-order time-dependent perturbation theory to rationalize the experimental results and that the second-order spin–vibronic coupling via molecular plays an important role in the of RISC processes. In the vibrational related to between the and are these have energies of that their activation at room is via thermal energy see Figure the between the and units approach the electronic coupling between the two units is significantly As a consequence, the HOMO and orbitals localized at the two lead to of the S1 and states, as well as the T1 and T2 states, see Figure This is important because it that in addition to the S1 and T1 states, the upper-lying and T2 play in the RISC However, the first-order perturbation theory the of the and T2 states to the RISC Thus, the second-order time-dependent perturbation theory is required to simultaneously SOC and coupling the so-called spin–vibronic As shown in the kRISC derived from the second-order perturbation theory, the second-order in addition to the first-order Equation have to the its is as: k RISC | ( H SO S 1 T 2 ▵ E T 1 − T 2 T 2 | | T 1 ) | 2 T 2 | | T 1 denotes the between T1 and T2 states related to the mode with the and | H SO S 1 T 2 | denotes the SOC between the S1 and T2 states. Equation (2) that the T2 state plays an role in the process from T1 to it that an increase in the | H SO S 1 T 2 | and T 2 | | T 1 and a reduction in the energy gap between T1 and T2 states can the Thus, the between the and units in reach a strong between the two T1 and T2 states (since this coupling is to the energy gap between the a in addition to a The multiple spin–flip due to the spin–vibronic coupling mechanism rationalize the efficient RISC process and TADF in organic molecules with large ΔES1–T1 values. As in the second-order spin–vibronic contribution to kRISC is up to of the total the second-order spin–vibronic coupling mechanism the initial molecular design to which fast RISC processes only from small ΔES1–T1. The investigations of et and and on D–A(–D) emitters that between the lowest LE triplet state on the D or A and the CT triplet state can significantly increase the due to large SOC between the LE triplet and CT singlet recent investigation on TADF emitters with multiple (e.g., 4CzIPN and also demonstrated that the RISC process from the T1 to S1 state is by the T2 Different from the case of simple D–A(–D) emitters, the T1 and T2 states in 4CzIPN show a hybrid LE–CT excitation The low-frequency between D and A units excitation characteristics and energies of S1, T1, and T2 states as well as the SOC which facilitates the RISC processes by T2 states. the energy of the T2 states significantly impact the activation energies of the RISC processes. In the case of 4CzIPN T2 is above S1, the RISC activation energy is by the energy offset ΔET2–T1 between the T2 and T1 states. In to in the case of the T2 energy is much lower than the S1 energy and to the T1 such an T2 state coupled with the T1 state leads to a smaller RISC activation energy than ΔES1–T1. Moreover, in these TADF emitters with multiple donor units, the characteristics and energy of T2 states which correspond to the T1 state of a partial molecular structure are tuned by and number of the design of organic molecules with multiple donor or acceptor moieties is promising to achieve TADF emitters with fast Organic Emitters Recently, et exploited a neutral radical molecule with an electronic structure as an OLED and achieved devices with EQE of In organic or emitters with a electronic structure, singlet triplet exciton is Different from emitters, radical emitters the doublet state with for and the emissive excited state Due to such a feature of electronic structure, the electronic transition from the lowest excited double state to the ground doublet state is a and the limit of IQEs of the OLEDs exploiting organic radical emitters is It is worth noting in the context of electroluminescence, both the doublet and excited states can be generated via electric since the excited states are their is for OLEDs exploiting radical emitters the of efficiently harvesting triplet

Topics & Concepts