Chinese Chemical Letters  2019, Vol. 30 Issue (10): 1717-1730   PDF    
Thermally activated delayed fluorescence molecules and their new applications aside from OLEDs
Wenlong Chena, Fengling Songa,b,*     
a State Key Laboratory of Fine Chemicals, School of Chemical Engineering, Dalian University of Technology, Dalian 116024, China;
b Institute of Molecular Sciences and Engineering, Shandong University, Qingdao 266237, China
Abstract: Thermally activated delayed fluorescence (TADF) organic molecules feature with long-lived delayed fluorescence, because they can undergo not only efficient intersystem crossing (ISC), but also efficient reverse intersystem crossing (RISC) at room temperature. As a new type of luminescent molecules, they have exhibited successful applications in organic light emitting diodes (OLEDs). Aside from OLEDs, they are also found to have potential applications in time-resolved luminescence imaging based on long-lived fluorescence property. Meanwhile, due to their excited triplet characteristic originated from efficient ISC, they were found to be applied in triplet-triplet annihilation upconversion (TTA-UC), photodynamic therapy (PDT) and organic photocatalytic synthesis. This review briefly summarizes the characteristics and excellent photophysical properties of TADF organic compounds, then covers their applications to date aside from OLEDs based on their highly efficient ISC ability and RISC ability at room temperature.
Keywords: Thermally activated delayed fluorescence     Intersystem crossing     Time-resolved luminescence     Triplet-triplet annihilation upconversion     Photodynamic therapy     Organic photocatalytic synthesis    
1. Introduction

Organoluminescent molecules have witnessed successful applications in various fields such as electroluminescent material [1-3], dye lasers [4, 5], photocatalysts [6-9], solar cells [10-13], and molecular sensor and fluorescent probes bioimaging [14-16] and photodynamic therapy (PDT) [17-21]. The development of organic luminescent materials benefits from the easy derivatization of molecular structures and have made much progress recently [22-24]. The luminescent organic molecules can be classed into three types: fluorescent molecules, phosphorescent molecules and thermally activated delayed fluorescence molecules according to their different luminescence mechanisms, which is depicted specifically in Fig. 1.

Fig. 1. The simple energy level diagram of three types of luminescent organic molecules: (a) fluorescence, (b) phosphorescence and (c) thermally activated delayed fluorescence.

Fluorescence is a universal photoluminescent phenomenon in most organic molecules with aromatic conjugated structure. Fluorescent molecules can be photoexcited to the singlet excited state, then radiatively deactivated by emission within nanoseconds. This quick emission is called fluorescence, also named prompt fluorescence (PF). Under normal conditions, pure organic luminescent molecules exhibit fluorescence but not phosphorescence owning to forbidden intersystem crossing transition (ISC) between the lowest singlet excited state (S1) and the lowest triplet excited state (T1). The energy gaps ΔEST between S1 and T1 are generally large (above 0.5 eV) for fluorescence molecules (Fig. 1a) [25, 26].

For phosphorescence molecules, the process of ISC can be dramatically enhanced by introducing heavy atoms or by some other complicated design strategies [17, 27-31], resulting in high population of lowest triplet excited state and subsequent radiative phosphorescence, although the ΔEST is still large (Fig. 1b). Comparing with fluorescent molecules, phosphorescent molecules with high-efficiency triplet quantum yield characteristics can used as photosensitizers in triplet-triplet annihilation upconversion (TTA-UC), photodynamic therapy (PDT), organic photocatalytic synthesis and as luminescence materials in time-resolved imaging and oganic light emitting diodes (OLEDs). However, the low molar absorption efficiency, possible toxicity and potential high cost in phosphorescent molecules are obstacles to their utilization in these field, considering that room-temperature phosphorescence cannot be easily obtained. These challenges cannot be resolved until the successful development of pure organic thermally activated delayed fluorescence (TADF) molecules.

TADF molecules with relatively small ΔEST can not only enhance the ISC ability to fulfill efficient triplet quantum yield without the use of heavy atoms, but also achieve efficient reverse intersystem crossing (RISC) by molecule thermal motion at room temperature (Fig. 1c). TADF molecules possess advantages of both fluorescent molecules and phosphorescent molecules and exhibit a great potential in many fields. In this review, we will briefly introduce the progress of photophysical properties of the TADF organic molecules and focus on their applications in time-resolved imaging, TTA-UC, PDT, organic photocatalytic synthesis. Because TADF molecules are the hot spot in OLEDs field, the recent related literatures has been covered in previous review paper [22, 32, 33].

2. Molecular structures and photophysical properties

TADF, as a special photoluminescence phenomenon, was firstly found in eosin in 1961 [34]. It has received much attention again until it was successfully applied into OLEDs in 2012 [3]. It is necessary to go through the photophysical properties of the TADF molecules afresh before understanding their applications.

The ISC rate constant kISC can be expressed by the following equation (Eq. (1)) according to the theory of nonradiative transitions [35, 36]:


where HSO is the Hamiltonian for spin–orbit coupling (VSO). It is clear that kISC is positively correlated with HSO, but negatively correlated with ΔEST. A small ΔEST should induce an enhanced ISC rate by bringing a tight resonance between S1 and T1 [37, 38]. Based on this principle, Adachi and colleagues developed [3] a series of pure organic molecules with high ISC efficiencies and long-lived T1. Meanwhile, a small ΔEST is also critical to maximize the RISC rate constant kRISC at room temperature given by Eq. (2) [39, 40]:


where kB is the Boltzmann constant and T is the temperature. According to the Eq. (2), this endothermic upconversion of the RISC process from the T1 to S1 state can be thermally activated and produce TADF considering a small ΔEST under room temperatures or lower temperature [39]. In addition to the small ΔEST, a high radiative decay rate kr of the S1 excitons is also important for TADF. For example, aromatic ketones such as benzophenone derivatives are well known to have fairly small ΔEST of 0.1–0.2 eV [41]. However, these molecules exhibit intense phosphorescence at low temperature but not fluorescence and TADF.

Many reported works focused on meeting these two prerequisites of a small ΔEST and a fairly high kr. Adachi group designed and synthesized a molecular PIC-TRZ with a twisted electron donor–acceptor (D-A) structure (Fig. 2). The molecular structure of PIC-TRZ was significantly distorted by steric hindrance introduced through bulky substituents, which resulted in efficient TADF with a small ΔEST 0.11 eV and a reasonably high kr [42]. This molecular design strategy for TADF compounds is to achieve such D–A molecules with twisted intramolecular charge transfer (TICT), which has proven to be an efficient pathway for designing pure organic molecules with highly efficient TADF [3]. Comparing to PIC-TRZ, 4CzIPN was developed as a star molecule with a very high external quantum efficiency (EQE) (19.3%) in OLEDs by reducing ΔEST from 110 meV to 83 meV. Furthermore, DACT-Ⅱ was found to have nearly zero ΔEST, which can realize approximately 100% upconversion of the triplet to singlet excited state. To the best of our knowledge, DACT-Ⅱ exhibits the highest EQE of 29.6% [43].

Fig. 2. The structures of the TADF compounds PIC-TRZ, 4CzIPN and DACT-Ⅱ. Yellow region represents electron donor, and red region represents electron acceptor. Such description styles of molecule structures are used in the following figures.

In addition to D–A molecules (Fig. 3a) with TICT, Zheng's group have summarized two other strategies for realizing high efficient TADF [44]. One is to design TADF molecules with through-space charge transfer (TSCT). Three kinds of TSCT molecules have been reported, including homoconjugation structure molecules such as TPA-QNX(CN)2 [45], spiro-conjugation structure molecules such as ACRFLCN, and D–A type TSCT molecules XPT [46] (Fig. 3b). The other one is to design molecules with multi-resonance induced TADF (MR-TADF), such as molecular DABNA-1 (Fig. 3c) [44].

Fig. 3. Three types of TADF molecule structures. Reproduced with permission [44]. Copyright 2019, Wiley-VCH Verlag GmbH & Co. KGaA.

Still, designing D–A molecules with TICT has become the most popular strategy to develop new TADF molecules. Molecular structure modification can efficiently improve photophysical properties of these TADF D–A molecules. Song's group developed a new TADF molecule DCF-MPYM with TICT based on a 2, 7-dichlorofluorescein, a water-soluble fluorophore (Fig. 4) [47]. This molecule can exhibit a long-lived red emission of 22.11 μs. Furthermore, when introducing a strong electron-withdrawing 2-(3-cyano-4, 5, 5-trimethylfuran-2(5H)-ylidene)malononitrile (TFM), the maximum emission wavelengths of these two new TADF molecules STFM-DCF and DTFM-DCF were extended from 630 nm to 681 nm and 755 nm [48]. At the same time, STFM-DCF and DTFM-DCF still kept the property of thermally activated delayed fluorescence and exhibit long-lived luminescence with 9.02 μs and 0.43 μs respectively in deaerated dichloromethane at room temperature. Song's group also proposed another idea to adjust the wavelength of TADF emission based on a FRET system FL-CyN by adopting the TADF molecule DCF-MPYM as FRET donor and a NIR fluorophore CyN as FRET acceptor [49]. FL-CyN performed long-life fluorescence emission phenomenon at long wavelengths 754 nm. The FRET-based strategy is also used to improve the performance of OLEDs [50, 51]. By introducing aromatic carbonyl groups to increase the rate of intersystem crossing, Song group found that the luminescence lifetime of the new TADF molecules DPK-DCF and DTK-DCF can be enhanced to 31.29 μs and 52.05 μs. Long lifetime emission is highly desired and more beneficial to time-resolved luminescence imaging by greatly improving signal-to-noise ratio [52].

Fig. 4. Fluorescein derivatives of TADF molecule structures.

Molecular structure modification has also been done to resolve the aggregation-caused fluorescence quenching (ACQ) or lifetime falloff, which is often found in the TADF molecules. Designing TADF molecules with aggregation-induced emission (AIE) is a smart approach to overcome these problems.Wang'sgroupdesignedand synthesizedtwo molecules TXO-PhCz and TXO-TPAwith TADFand AIE properties. These two AIE molecules employed 9H-thioxanthen-9-one-10, 10-dioxide (TXO) core as the acceptor unit, and N-phenylcarbazole (PhCz) or triphenylamine (TPA) as the donator unit. The single crystal analysis shows that TPA and PhCz units adopt a highly twisted conformation and can prevent both the TADF molecules from forming excimers and exciplexes in solid state. Their AIE behavior of TXO-TPA and TXO-PhCz provides their high photoluminescence quantum yield (ΦPL) in solid film (Fig. 5a) [53]. Recently, Chi's group developed a new approach to achieve an AIE-enhanced TADF by introducing nonplanar donors [54]. In their work, TADF molecule DCZ-SF showed a typical ACQ effect [55]. But compounds PTCZ-SF and DPT-SF were designed to modulate the molecular π-π interactions to induce AIE by replacing carbazole rings with nonplanar phenothiazine (Fig. 5b). And these two molecules provide high emission efficiency in their aggregated states (a high TADF quantum yield of 93.3% in the solid state for PTCZ-SF). Similarly, TADF molecule O2C has ACQ effect due to its planar donors. When nonplanar donors were introduced into O2P and OPC, these two new TADF molecules were proved to have AIE property (Fig. 5c). In fact, OPC achieved white emission in non-doped solid state [56]. Afterwards, numerous AIEgens with TADF began to be explored in the application of OLEDs [57-60].

Fig. 5. Three types of molecule structures with TADF and AIE characteristics.

What is more, the same strategy of integrating AIE and TADF characteristics into a single molecule has also been employed in bioimaging applications. TADF molecules with AIE property were found to avoid their emission be quenched by oxygen, which made them more suitable for time resolved luminescence bioimaging. Zhao et al. reported a series of TADF molecules by chosing benzophenone (BP) as the acceptor (A) and nonplanar phenoxazine (PXZ) or phenothiazine (PTZ) as donors (D) (Fig. 6). Single crystals of BP-2PXZ and BP-PXZ showed that these molecules adopt a highly twisted conformation, which facilitates the occurrence of TADF. To improve water solubility of these TADF molecules, water-soluble nanoparticles (NPs) was generated by encapsulating them within a BSA matrix. The NPs still kept the intriguing AIE and TADF characteristics and were used in fluorescence lifetime bioimaging applications [61].

Fig. 6. Molecular structures of BP-2PXZ and BP-PXZ.

3. Applications aside from OLEDs

As mentioned above, TADF molecules have unique triplet excited state properties and singlet excited state properties based on the their small ΔEST. There are mainly five potential applications well studied in recent literatures as the Fig. 7 describes. As a singlet emitter, TADF compounds can be applied in OLEDs and fluorescence bioimaging. And as a triplet photosensitizer, TADF molecule can be applied in triplet-triplet annihilation up conversion (TTA-UC), photodynamic therapy (PDT) and organic photocatalytic synthesis.

Fig. 7. The main application based on TADF compounds.

TADF molecules was firstly used in OLEDs by Adachi group in 2009 [62]. It is well known that exciton formation under electrical excitation typically results in 25% singlet and 75% triplet excitons according to spin statistics [2, 63]. However, only 25% singlet excitons can be harvested by fluorescence emitter. Currently OLEDs use rare metallic complexes as luminescence materials which can achieve nearly 100% IQE. However, these phosphoresce materials are subject to the growing cost and potential environmental contamination of these heavy rare metals [64, 65]. The most promising and efficient approach to harvest both singlet and triplet excitons has been agreed to be the utilization of TADF emitters. OLEDs fabricated from pure organic TADF materials have already displayed colors covering the whole visible region with comparable EQE with phosphoresce-based OLEDs [22]. Since the reviews of OLEDs application of TADF molecules are sufficient enough [22, 32, 33], we focus on applications of TADF molecules in time resolved luminescence imaging, TTA-UC, PDT and organic photocatalytic synthesis in this review.

3.1. Time resolved luminescence imaging

TADF has almost as same emission spectrum as PF since both delayed and prompt fluorescence correspond to the S1 to S0 transition. But TADF has a much longer decay time for involving in the long lifetime triplet state, which makes TADF suitable for time-resolved fluorescence imaging. Compared with fluorescence imaging, time-resolved fluorescence microscopy exploits advantages in eliminating the short-lived background fluorescence in the order of nanoseconds, which can provide high signal-to-noise ratios in monitoring target fluorescence [66]. Lanthanides and transition-metal complexes are famous for their long luminescence lifetimes and used in time-resolved fluorescence imaging. However, the toxicity of these heavy metal-based complexes must be considered in practical applications [67]. In order to meet the requirements for time-resolved imaging technology, TADF molecules have become the perfect alternative for lanthanides and transition-metal complexes. Some long-lived luminescent probes based TADF molecules have been reported recently to be used in time-resolved luminescence bioimaging to improve the detection sensitivity and accuracy.

Pure organic TADF molecules have been successfully applied in OLEDs since 2011 [42]. But their poor water solubility limits their application in bioimaging. Song's group reported a water-soluble fluorescein derivative DCF-MPYM (Fig. 8) which exhibits long-wavelength emission (>600 nm), large Stokes shift characteristics and fluorescence-enhanced response to bovine serum albumin (BSA) [68]. Further studies have found that DCF-MPYM exhibits long-lived triplet excited states of 23.89 μs and long-lived fluorescence emission with a fluorescence lifetime of 22.11 μs. And the fluorescence lifetime and time-resolved fluorescence intensity of the molecule were found to be sensitive to oxygen Figs. 9a and b). Based on these results, it is proposed that the molecule DCF-MPYM is a TADF compound [47]. A strong red luminescence signal for the stained cells by DCF-MPYM was clearly observed in time-resolved fluorescence imaging Figs. 9c and d) compared to steady-state fluorescence imaging Figs. 9e and f). These results prove that the time-resolved imaging method can effectively remove interference from the autofluorescence and scattered light and improve the accuracy and clarity of cell imaging. It is noteworthy that these results support that DCF-MPYM has good enough water solubility to be directly used in bioimaging.

Fig. 8. Molecular structures of DCF-MPYM and fluorescence-enhanced response to bull serum albumin.

Fig. 9. (a) Lifetime decay of DCF-MPYM at room temperature, after nitrogen bubbling and deoxygenation (red curve) and deoxidation sample after 6 h under air conditions (black curve). (b) Time delayed fluorescence spectrum of air condition DCF-MPYM (10 μmol/L in CH3CN) (■ black line) and delay of DCF-MPYM under Ar atmosphere spectrum (● red line). DCF-MPYM for MCF-7 time-resolved fluorescence imaging image, where (c) is a bright field, (d) is a time-resolved fluorescence field; DCF-MPYM (20 μmol/L) for MCF-7 steady-state fluorescence images, where (e) is a bright field, (f) is a steady-state fluorescence field. Reproduced with permission [47]. Copyright 2014, American Chemical Society.

Based on this TADF molecule DCF-MPYM, Song's group reported a fluorescent probe molecule DCF-MPYM-thiol that can specifically respond to cysteine [69]. As shown in Fig. 10, the fluorescence of probe molecule is designed in the "OFF" state. When the probe molecule is treated with cysteine, the fluorescence can be rapidly restored by forming the DCF-MPYM. Time-resolved imaging of cysteine in cells was achieved by using this probe molecule.

Fig. 10. Reaction principle of probe molecule DCF-MPYM-thiol when treated with cysteine, the detecting target.

Most other TADF molecules have poor water solubility due to their aromatic structures, which limits their application in the biological field. A variety of strategies have been employed to resolve the problem of water solubility. Huang's groups reported an organic dots CPy-Odots which formed by coating the TADF molecule CPY (Fig. 11a) with the ethylene glycol analog DSPE-PEG2000 [70]. CPy-Odots not only shows high brightness and strong stability, but also possesses AIE and TADF properties with a long fluorescence lifetime of 9.3 μs. Time-resolved imaging with cells and zebrafish angiography were achieved using this TADF material. Using the similar approach, Tang's group [61] reported a nanoparticle which is consisted of a series of TADF molecule molecules (BP-2POZ, BP-2PSZ, BP-POZ and BP-PSZ) with AIE properties (Fig. 11b) and coated in a cavity of bovine serum albumin for effectively isolating oxygen. It not only maintains the AIE properties and TADF properties, but also is successfully applied to time-resolved fluorescence imaging in cells.

Fig. 11. Molecule structures of CPY, DSPE-PEG2000 (a) and BP-2POZ, BP-2PSZ, BP-POZ and BP-PSZ (b).

Some classical TADF compounds such as 4CzIPN, NAI-DPAC, and BTZ-DMAC have excellent photophysical properties (Fig. 12a). But they are limited to be used in the biological field due to their poor water solubility. Zhao's group adopted a biocompatible amphiphilic polypeptide chain [F6G6(rR)3R2] to form self-assemble nanoparticles with TADF compounds 4CzIPN, NAI-DPAC and BTZ-DMAC [16]. The outer layer polypeptide of nanoparticles can effectively transport these TADF molecules into the cells, which solves the problem that these TADF molecules are difficult to enter the cells. What is more important, the TADF compounds encapsulated in the nanoparticle was effectively isolated from external oxygen, so that the long-lived fluorescence emission cannot be affected by oxygen. The forming TADF nanoparticles can be used in time-resolved imaging.

Fig. 12. (a) Structures of the amphiphilic peptide [F6G6(rR)3R2] and three TADF compounds: 4CzIPN, NAI-DPAC and BTZ-DMAC. (b) Structures of TXO and the amphiphilic triblock copolymer (PEG-b-PPG-b-PEG).

Basedona similar strategy, Fan's group collaborated with Huang's group successfully synthesized a kind of semiconductor nanoparticles TXO-NPs assembled by amphiphilic triblock copolymer polymer PEG-b-PPG-b-PEG and the TADF compound TXO with AIE property by nanoprecipitation [71] (Fig. 12b). The nanoparticles havegoodwater solubilityand two-photon absorption properties.It is noteworthy that a microsecond level of fluorescence lifetime was kept even in an oxygen environment. Therefore, TXO-NPs achieved good performance in time-resolved fluorescence imaging in cells.

With the continuous development of TADF molecules, some TADF compounds with good biocompatibility can be directly applied to biological imaging. Yang group reported a TADF compounds PXZT [72] (Fig. 13a) with characteristic of both TADF, AIE and crystallization-induced room temperature phosphorescence. The molecule ZnPXZT1 was synthesized by complexing PXZT with Zn2+, which caused the quenching of the fluorescence of PXZT due to the enhanced intramolecular charge transfer. And Zn2+ can be dissociated under certain conditions to get free PXZT accompanying with fluorescence restore. The dissociation process was confirmed in solution using EDTA. What is more important, the free hydrophobic PXZT can aggregate in situ to give TADF due to its AIE property. And the restored TADF is free from the influence of oxygen in biological environment. Time-resolved fluorescence lifetime imaging was achieved by applying ZnPXZT1 to HEIA and 3T3 cells. The authors speculate that some biomolecules in the cells dissociated ZnPXZT1, and the free PXZT can undergoes aggregation-induced luminescence. Later, Yang group reported a new case of the TADF molecule NID-TPP with mitochondria-targeted AIE effect (Fig. 13b) [73]. NID-TPP behaves as a typical AIE property but exhibits fluorescence "OFF" state under dilute solution. NID-TPP can aggregates around the mitochondria due to negative potential in the mitochondrial membrane, and then induces AIE. It is important that the TADF produced by aggregated NID-TPP is free from the influence of oxygen in biological environment. NID-TPP was applied to intracellular mitochondrial time-resolved fluorescence imaging.

Fig. 13. (a) TADF compounds PXZT and ZnPXZT1 conversion mechanism. (b) Schematic diagram of mitochondrion-induced NID-TPP aggregation. Reproduced with permission [72]. Copyright 2018, The Royal Society of Chemistry.

3.2. TTA-UC

A small S1-T1 energy gap ΔEST < 100 meV) can enhance the ISC efficiency and provide TADF molecules a long-lived triplet excited state, which make TADF molecules meet the requirements of photosensitizers (PS) of TTA-UC (Fig. 14). From the point of TTA-UC system, TADF molecules are appealing owing to their two advantages: (ⅰ) The small ΔEST can induce very efficient ISC process without the use of heavy atoms. (ⅱ) A small ΔEST brings about less energy loss during ISC process in TADF molecules than in common phosphorescent photosensitizers, which can match well with an acceptor with a much higher triple state energy level. The later advantage of TADF molecules can eventually realize a relatively larger anti-stokes shift for TTA-UC system.

Fig. 14. Diagram for the common mechanism of TTA-UC that consists of the main photophysical process: intersystem crossing (ISC), donor-to-acceptor triplet–triplet energy transfer (TTET) and triplet–triplet annihilation (TTA).

Some TTA-UC systems sensitized by TADF molecules have been reported in recent years [74]. In 2015, Baldo's group reported the first example of pure organic TADF molecule 4CzTPN-Ph as a photosensitizer in TTA-UC system with the blue emitter 9, 10-diphenylanthracene (DPA) as an energy acceptor (Fig. 15). A two-layer solid film consisted of photosensitizer and acceptor was built to achieve the upconversion emission [75]. And a complete study the TTA-UC system was done in detail involving triplet to triplet energy transfer (TTET) between the TADF photosensitizers and acceptor, and triplet-triplet annihilation (TTA) of the acceptor which is also called emitter or annihilator. The sensitization efficiency was calculated to be 9.1%, and the upconversion quantum yield ηUP = 0.28%. Although the upconversion quantum yield needs to be further improved, this work confirmed that TADF compounds without heavy atoms can be used as a new generation of photosensitizers even in solid-state TTA-UC system. It has opened a new direction for the application of TADF compounds as photosensitizers in TTA-UC.

Fig. 15. Chemical structures of photosensitizer 4CzTPN-Ph and acceptor DPA.

In 2016, based on another TADF molecule 4CzIPN as photosensitizer, an upconversion system was established by Yanai and Kimizuka's group [76] in solution system (Fig. 16). Thanks to the extremely small ΔEST of TADF photosensitizer, the energy loss during ISC is less, which widens the choice of acceptors. It is means that TADF photosensitizer can increases the potential to sensitize emitters with high T1/S1 energy levels and maximize the anti-Stokes shift. An upconversion emission from green visible light to ultraviolet was achieved by matching the acceptor p-quarterphenyl (QP) and p-terphenyl (TP) respectively. The authors asserted that it is the highest T1 (2.5 eV) and S1 (3.6 eV) energy levels of emitters (TP) ever employed in TTA-UC. The sensitization efficiency and upconversion emission quantum efficiency for 4CzIPN/QP were as large as 20% and 3.9%, respectively. It proves again that the use of TADF molecules as photosensitizers can reduce the energy loss of ISC process to achieve large anti-Stokes shift TTA-UC even in ultraviolet region.

Fig. 16. Chemical structures of photosensitizer 4CzIPN and acceptor TP and QP.

Although, the above studies demonstrated that TADF molecules can be used as photosensitizers to achieve TTA-UC, the reported upconversion quantum efficiency is still relatively low. There are many factors affecting the upconversion quantum efficiency. Among them, the sensitizing efficiency from the photosensitizer to the emitter is the key one. Increasing the sensitizing ability of the TADF photosensitizer to the acceptor can make the triplet energy transfer from the photosensitizer to the emitter more efficient. In 2016, Ma group developed a novel TTA-UC system by attaching the acceptor unit 2, 7-di-tert-butylpyrene (DBP) to the TADF molecule 4CzPN by covalent attachment (Fig. 17) [77]. This TTA-UC system can be effectively realized even in polyurethane film. It was confirmed that covalent attachment of DBP groups to 4CzPN was very critical to achieve efficient sensitization and upconversion performance in solid matrices. It was proposed that the RISC process can be suppressed, which effectively increase the TTEF efficiency from triplet photosensitizer to the emitters. Therefore, the TTA-UC luminescence efficiency can be eventually improved in the solid state.

Fig. 17. Molecular structures and simple schematic energy diagrams of the important processes in TTA UC sensitized by 4CzPN (left) and 4CzPN–1DBP (right). FL: fluorescence; GS: ground state; RISC: reverse intersystem crossing; iTTET: intramolecular TTET. Reproduced with permission [77]. Copyright 2016, The Royal Society of Chemistry.

Recently, some new types of TADF photosensitizers have been used in TTA-UC. In 2017, Yang's group reported that a TADF molecule BTZ-DMAC with red emission (Fig. 18) as a photosensitizer to achieve TTA-UC from green to blue light [78]. The red TADF molecule BTZ-DMAC has a low triplet energy level, which can be matched with the acceptor DPA. Compared with TADF molecule 4CzIPN with green emission, the energy loss in the TTET process can be effectively reduced. So, a large anti-Stokes shift of 97 nm was obtained. The upconversion emission phenomenon in the solution system can be clearly observed even under air conditions.

Fig. 18. Chemical structures of photosensitizer BTZ-DMAC and acceptor DPA.

Compared with traditional phosphoresce photosensitizers, TADF photosensitizers need to improve the upconversion emission quantum efficiency. Recently, Song's group has successfully developed a new TTA-UC system by using the TADFmolecule DCF-MPYM as photosensitizer with perylene and DPA as acceptor (Fig. 19) [79]. The very small ΔEST of the DCF-MPYM can effectively avoid the energy loss during ISC process to reach a high triplet state energy level and obtain a long triplet state lifetime. These features finally contributed to the large anti-stokes shifts and high upconversion emission quantum yields for this TTA-UC system. Compared with perylene, DPA was found to be the better acceptor for DCF-MPYM photosensitizer.

Fig. 19. (a) Chemical structures of the new photosensitizer DCF-MPYM and blue light-emitting acceptors perylene and DPA. (b) Simple schematic diagram of TTA photophysical process. Reproduced with permission [79]. Copyright 2019, The Royal Society of Chemistry.

To our best knowledge, the anti-Stokes shift (207 nm) is the largest one in the reported TTA-UC systems using pure organic TADF molecules as photosensitizers. And the upconversion emission quantum yield of 11.2% is also comparable to those of phosphoresce photosensitizers. We expect this TTA-UC system has a bright prospect in solar energy conversion and biomedical applications in the future.

3.3. PDT

Another valuable application based on TADF photosensitizers is photodynamic therapy [20]. Owing to its spatiotemporal selectivity and noninvasive nature, photodynamic therapy (PDT) has become a clinically promising approach for the treatment of a wide range of cancers and other diseases. However, the full potential of PDT has not been achieved. The lack of optimal photosensitizers is one of the problems to be solved [80]. Thus far, photosensitizers employed in clinical applications are phosphoresce photosensitizers, such as porphyrin, chlorin, phthalocyanine derivatives, hypericin, methylene blue, rose Bengal and BODIPY. However, the intrinsic cytotoxicity and non-degradation issues of heavy-metal-containing photosensitizers have to consider before used in PDT. It turns out that pure organic compounds such as TADF molecules have become the better alternative because TADF molecules have excellent triplet nature and can sensitizeoxygen into reactive oxygen species to kill the cancer cells. It is believed that TADF molecules are promising to become the excellent replenishments for traditional triplet photosensitizers.

However, mostof the existingTADF compoundshave ahydrophobic structure, which makes it difficult to apply them to biological environment due to their poor water solubility. Lee's group resolved problem of the water solubility by coating serval classical TADF molecules with nanoparticles (Fig. 20) [81]. The nanoparticles were indicated to have the ability to sensitize oxygen to produce singlet oxygen to be used in PDT.

Fig. 20. Simple schematic diagram of PDT (a) and the structures of 2CzPN, 4CzIPN, 4CzTPN-Ph, D1:TAPC, D2:NPB, and A:DPTPCz (b).

Although most of the existing TADF molecules have the hydrophobicity problem, there are still some water soluble TADF molecules been developed which can be directly applied in biological environment. Recently, Song's group made full use of the good hydrophilicity of fluorescein derivatives and designed two molecules compound 1 and compound 2 with enzymeactivated TADF characteristics (Fig. 21a) [82]. As PDT photosensitizers, water soluble TADF molecules have another valuable advantage. Besides the ability of sensitization of oxygen to reactive oxygen species, their long-lifetime fluorescence emission can be used in time-resolved fluorescence imaging to achieve the goal of theranostics. The two probes compound 1 and compound 2 employed p-nitrobenzyl group as a nitroreductase specific reaction site which can not only quench the fluorescence of the probes through the photoinduced electron transfer (PET) mechanism, but also weaken the ISC process of the probes to reduce the efficiency of the excitation triplet state. Before nintroreductase activation, the fluorescence and PDT properties of the probes compound 1 and compound 2 are in the "OFF" state. When the probes react with the nitroreductase to form the molecule DCF-MPYM, both TADF and PDT effect are effectively restored. Compound 2 was chosen to be used in vivo PDT, and excellent PDT efficiency was achieved Figs. 21b and c). The experimental mice kept good weights and healthy organ tissues because the PDT effect of compound 2 can be activated only in tumor regions Figs. 21d and e).

Fig. 21. (a) Two probe molecular structures, compound 1 and compound 2, and the product DCF-MPYM after reduction of nitroreductase with compound 1 and compound 2. (b) The pictures of four groups of tumor-bearing mice with different experimental treatments. (c) Changes of tumor volumes of the four groups with different experimental treatments. (d) Changes of the weight of the mice after the phototoxicity test. (e) H&E staining of tumor tissue sections from four groups of tumor-bearing mice with different experimental treatments. Reproduced with permission [82]. Copyright 2019, American Chemical Society.

3.4. Organic photocatalytic synthesis

The last but not the least application of the TADF molecules is organic photocatalytic synthesis. Visible-light photoredox catalysis has undoubtedly proven to be a powerful tool for the activation of small molecule under mild conditions with high tolerance to functional groups. Organic photocatalytic synthesis has found widespread applications in a myriad of syntheses and functionalization considering the green chemistry nature of organic catalysts which are also called photosensitizers [83].

A good photosensitizer of photocatalytic synthesis calls for a sufficiently long-lived electronically excited state (ES), which is beneficial to single-electron-transfer (SET) process or energy transfer (EnT) activation pathway [84]. Pure organic TADF molecules have been becoming the most attractive alternative to metalorganic complexes which have the risk and concerns of potential metal contamination.

The TADF molecules usually have a relatively high triplet excited quantum yield (ΦT), and a long-lifetime triplet excited state. Both the two features provide an efficient TTET activation in photocatalysis process. The redox potential distribution of D-A type photosensitizers mainly depends on the LUMO of the donor part and the HOMO of the acceptor part. Therefore, the redox property of the photosensitizer can be controlled by simply adjusting the D-A structure to satisfy catalyst demand of the organic photocatalytic synthesis.

The application potential of pure organic TADF photosensitizers has been demonstrated in the field of organic photocatalysis. Zhang's group reported a series of low-cost photosensitizers with different reducing properties based on the classical TADF compounds carbazole benzonitrile [85] (Fig. 22a). The authors have achieved the regulation of the redox potential of photosensitizers by simply changing the number and substitution sites of the donor carbazolyl group and acceptor benzonitrile group. Furthermore, the C(sp3)-C(sp2) cross-coupling photocatalytic reaction was achieved with the photoredox/Ni dual catalysis mode (Fig. 22b). Subsequently, Zhang's group further studied the triplet-triplet energy transfer catalysis mechanism of TADF photocatalysts [84]. The authors proposed that the energy order between 1CT (chargetransfer singlet state) and 3LE (locally excited triplet state) directly determine and affect the accessibility of the 3LE and 3CT (chargetransfer triplet state) and effectively reducing the 1CT can improve the organic photocatalytic performance involved of the chemically more tunable 3CT state via a combined classical diphenylethylene visible-light isomerization reaction and transient absorption spectroscopic study (Fig. 23a). Based on the fact that the 1CT energy level can get lowered by effectively by increasing D-A CT character, the authors designed a new photocatalyst 4DPAPN by increase of the donor strength of the electron donor, which realized the cross-coupling reaction of more kinds of carboxylic acids and halogenated aromatics with Ni(Ⅱ) as a cocatalyst Figs. 23b and c). Subsequently, photocatalysts based on TADF compounds 4CzIPN modification have also been reported [83]. It has been fully confirmed that the pure organic TADF photosensitizers with the D-A structure have the advantage of easily structural modification to achieve catalytic performance.

Fig. 22. Molecular structures of pure organic TADF compounds (a) and the C(sp3)-C(sp2) cross-coupling photocatalytic reaction achieved with the photoredox/Ni dual catalysis mode (b). Reproduced with permission [85]. Copyright 2016, American Chemical Society.

Fig. 23. Scheme of triplet-triplet energy transfer catalysis mechanism (a). Molecular structure of pure organic TADF compound 4DPAPN (b) and the Ni(Ⅱ)-catalyzed crosscoupling of carboxylic acids and aryl halides (c). Reproduced with permission [84]. Copyright 2018, American Chemical Society.

The classic TADF molecule 4CzIPN was also found in the development of other new photocatalytic organic synthesis. Mariano's group found that the aryl acylation of aryl olefins can be done by selecting 4CzIPN as the photocatalyst and diethoxyacetic acid as the formylating reagent (Fig. 24) [86]. By using this new protocol, aldehydes can be generated regioselectively up to 90% yield. Similarly, Molander's group recently reported the utility of 4CzIPN as a photocatalyst for the synthesis of DNA-encoded compound libraries [87]. The 4CzIPN/Ni dual catalytic system was applied to the modification of short-chain DNA by C(sp3)-C(sp2) cross-coupling reactions and radical/polar cross-alkylation reactions (Fig. 25). The authors demonstrated 140 reactions to expand the DNA-encoded compound libraries.

Fig. 24. Organo-photoredox catalyzed hydroformylation reactions by TADF 4CzIPN. Reproduced with permission [86]. Copyright 2017, American Chemical Society.

Fig. 25. Ni/photoredox dual catalytic C(sp3)-C(sp2) cross-coupling on DNA. Reproduced with permission [87]. Copyright 2019, American Chemical Society.

4. Conclusion

Pure organic TADF molecules have attracted much attention due to their successful application in OLEDs, which have stimulated photophysical and theorical research works and the development of new TADF molecules. In turn, these research works have further expanded the application fields aside from OLEDs. This review mainly summarizes the two major characteristics of TADF compounds: triplet characteristic due to high ISC efficiency and strong delayed fluorescence due to high RISC efficiency. And then we focus on covering the expanded applications of these TADF molecules in time resolved luminescence imaging, TTA-UC, PDT, and photocatalytic organic synthesis. Of course, there might be some new application areas of TADF molecules that are not mentioned here. And we also believe that there will be more new application areas based on TADF molecules in the future due to their rich photophysical processes.


This work was financially supported by the National Natural Science Foundation of China (Nos. 21877011, 21576038, 21421005) and the Talent Fund of Shandong Collaborative Innovation Center of Eco-Chemical Engineering (No. XTCXYX03).

J.H. Burroughes, D.D.C. Bradley, A.R. Brown, et al., Nature 347 (1990) 539-541. DOI:10.1038/347539a0
M.A. Baldo, D.F. O'Brien, Y. You, et al., Nature 395 (1998) 151-154. DOI:10.1038/25954
H. Uoyama, K. Goushi, K. Shizu, H. Nomura, C. Adachi, Nature 492 (2012) 234-238. DOI:10.1038/nature11687
J.H. Boyer, A.M. Haag, G. Sathyamoorthi, et al., Heteroatom Chem. 4 (1993) 39-49. DOI:10.1002/hc.520040107
M. Álvarez, F. Amat-Guerri, A. Costela, et al., Opt. Commun. 267 (2006) 469-479. DOI:10.1016/j.optcom.2006.06.059
B.D. Ravetz, A.B. Pun, E.M. Churchill, et al., Nature 565 (2019) 343-346. DOI:10.1038/s41586-018-0835-2
C. Kerzig, O.S. Wenger, Chem. Sci. 9 (2018) 6670-6678. DOI:10.1039/C8SC01829D
N.A. Romero, D.A. Nicewicz, Chem. Rev. 116 (2016) 10075-10166. DOI:10.1021/acs.chemrev.6b00057
A. Hossain, A. Bhattacharyya, O. Reiser, Science 364 (2019) eaav9713. DOI:10.1126/science.aav9713
A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, J. Am. Chem. Soc. 131 (2009) 6050-6051. DOI:10.1021/ja809598r
A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253-278. DOI:10.1039/B800489G
T.F. Schulze, T.W. Schmidt, Energy Environ. Sci. 8 (2015) 103-125. DOI:10.1039/C4EE02481H
M.A. Green, A. Ho-Baillie, H.J. Snaith, Nat. Photonics 8 (2014) 506-514. DOI:10.1038/nphoton.2014.134
L. He, B. Dong, Y. Liu, W. Lin, Chem. Soc. Rev. 45 (2016) 6449-6461. DOI:10.1039/C6CS00413J
D. Wu, A.C. Sedgwick, T. Gunnlaugsson, et al., Chem. Soc. Rev. 46 (2017) 7105-7123. DOI:10.1039/C7CS00240H
Z. Zhu, D. Tian, P. Gao, et al., J. Am. Chem. Soc. 140 (2018) 17484-17491. DOI:10.1021/jacs.8b08438
Z. Wang, Y. Gao, M. Hussain, et al., Chemistry 24 (2018) 18663-18675. DOI:10.1002/chem.201804212
X. Zhu, Q. Su, W. Feng, F. Li, Chem. Soc. Rev. 46 (2017) 1025-1039. DOI:10.1039/C6CS00415F
Y. Liu, X. Meng, W. Bu, Coord. Chem. Rev. 379 (2019) 82-98. DOI:10.1016/j.ccr.2017.09.006
S. Monro, K.L. Colon, H. Yin, et al., Chem. Rev. 119 (2019) 797-828. DOI:10.1021/acs.chemrev.8b00211
G.X. Lan, K.Y. Ni, W.B. Lin, Coord. Chem. Rev. 379 (2019) 65-81. DOI:10.1016/j.ccr.2017.09.007
Z. Yang, Z. Mao, Z. Xie, et al., Chem. Soc. Rev. 46 (2017) 915-1016. DOI:10.1039/C6CS00368K
J. Mei, N.L. Leung, R.T. Kwok, J.W. Lam, B.Z. Tang, Chem. Rev. 115 (2015) 11718-11940. DOI:10.1021/acs.chemrev.5b00263
Z. He, W. Li, G. Chen, Y. Zhang, W.Z. Yuan, Chin. Chem. Lett. 30 (2019) 933-936. DOI:10.1016/j.cclet.2019.03.015
N.J. Turro, V. Ramamurthy, J.C. Scaiano, Photochem. Photobiol. 88 (2012) 1033-1033. DOI:10.1111/j.1751-1097.2012.01178.x
A. Köhler, H. Bässler, Electronic Processes in Organic Semiconductors: An Introduction, 1st ed., Wiley-VCH, Weinheim, 2015.
A.N. Amin, M.E. El-Khouly, N.K. Subbaiyan, et al., Chem. Commun. 48 (2012) 206-208. DOI:10.1039/C1CC16071K
M.B. Smith, J. Michl, Chem. Rev. 110 (2010) 6891-6936. DOI:10.1021/cr1002613
R.J. Dillon, G.B. Piland, C.J. Bardeen, J. Am. Chem. Soc. 135 (2013) 17278-17281. DOI:10.1021/ja409266s
Z. Wang, J. Zhao, A. Barbon, et al., J. Am. Chem. Soc. 139 (2017) 7831-7842. DOI:10.1021/jacs.7b02063
L. Jiao, F. Song, J. Cui, X. Peng, Chem. Commun. 54 (2018) 9198-9201. DOI:10.1039/C8CC04582H
Y. Tao, K. Yuan, T. Chen, et al., Adv. Mater. 26 (2014) 7931-7958. DOI:10.1002/adma.201402532
M.Y. Wong, E. Zysman-Colman, Adv. Mater. 29 (2017). DOI:10.1002/adma.201605444
C.G. Hatchard, Trans. Faraday Soc. 57 (1961) 1894-1904. DOI:10.1039/tf9615701894
P.T. Chou, Y. Chi, M.W. Chung, C.C. Lin, Coord. Chem. Rev. 255 (2011) 2653-2665. DOI:10.1016/j.ccr.2010.12.013
Y.L. Chen, S.W. Li, Y. Chi, et al., ChemPhysChem 6 (2005) 2012-2017. DOI:10.1002/cphc.200500252
Y. Olivier, B. Yurash, L. Muccioli, et al., Phys. Rev. Mater. 1 (2017) 075602. DOI:10.1103/PhysRevMaterials.1.075602
S. Xu, Y. Yuan, X. Cai, et al., Chem. Sci. 6 (2015) 5824-5830. DOI:10.1039/C5SC01733E
M.N. Berberan-Santos, J.M.M. Garcia, J. Am. Chem. Soc. 118 (1996) 9391-9394. DOI:10.1021/ja961782s
F.B. Dias, T.J. Penfold, A.P. Monkman, Methods Appl. Fluoresc. 5 (2017) 012001. DOI:10.1088/2050-6120/aa537e
J. Saltiel, H.C. Curtis, L. Metts, et al., J. Am. Chem. Soc. 92 (1970) 410-411. DOI:10.1021/ja00705a617
A. Endo, K. Sato, K. Yoshimura, et al., Appl. Phys. Lett. 98 (2011) 083302. DOI:10.1063/1.3558906
H. Kaji, H. Suzuki, T. Fukushima, et al., Nat. Commun. 6 (2015) 8476. DOI:10.1038/ncomms9476
X. Liang, Z.L. Tu, Y.X. Zheng, Chem.-Eur. J. 25 (2019) 5623-5642. DOI:10.1002/chem.201805952
K. Kawasumi, T. Wu, T. Zhu, et al., J. Am. Chem. Soc. 137 (2015) 11908-11911. DOI:10.1021/jacs.5b07932
H. Tsujimoto, D.G. Ha, G. Markopoulos, et al., J. Am. Chem. Soc. 139 (2017) 4894-4900. DOI:10.1021/jacs.7b00873
X. Xiong, F. Song, J. Wang, et al., J. Am. Chem. Soc. 136 (2014) 9590-9597. DOI:10.1021/ja502292p
X. Peng, F. Song, K. Han, et al., J. Photonics Energy 8 (2018) 032103.
Z. Liu, F. Song, B. Song, et al., Sens. Actuators B-Chem. 262 (2018) 958-965. DOI:10.1016/j.snb.2018.01.240
T. Yamanaka, H. Nakanotani, S. Hara, T. Hirohata, C. Adachi, Appl. Phys. Express 10 (2017) 074101. DOI:10.7567/APEX.10.074101
H. Nakanotani, T. Higuchi, T. Furukawa, et al., Nat. Commun. 5 (2014) 4016. DOI:10.1038/ncomms5016
Y. Wu, F. Song, W. Luo, et al., ChemPhotoChem 1 (2017) 79-83. DOI:10.1002/cptc.201600027
H. Wang, L. Xie, Q. Peng, et al., Adv. Mater. 26 (2014) 5198-5204. DOI:10.1002/adma.201401393
S. Xu, T. Liu, Y. Mu, et al., Angew. Chem. Int. Ed. 54 (2015) 874-878. DOI:10.1002/anie.201409767
Q. Zhang, J. Li, K. Shizu, et al., J. Am. Chem. Soc. 134 (2012) 14706-14709. DOI:10.1021/ja306538w
Z. Xie, C. Chen, S. Xu, et al., Angew. Chem. Int. Ed. 54 (2015) 7181-7184. DOI:10.1002/anie.201502180
F. Rizzo, F. Cucinotta, Isr. J. Chem. 58 (2018) 874-888. DOI:10.1002/ijch.201800049
B. Huang, W.C. Chen, Z. Li, et al., Angew. Chem. Int. Ed. 57 (2018) 12473-12477. DOI:10.1002/anie.201806800
S. Gan, W. Luo, B. He, et al., J. Mater. Chem. C 4 (2016) 3705-3708. DOI:10.1039/C5TC03588K
J. Guo, X.L. Li, H. Nie, et al., Chem. Mater. 29 (2017) 3623-3631. DOI:10.1021/acs.chemmater.7b00450
S. Gan, J. Zhou, T.A. Smith, et al., Mater. Chem. Front. 1 (2017) 2554-2558. DOI:10.1039/C7QM00286F
A. Endo, M. Ogasawara, A. Takahashi, et al., Adv. Mater. 21 (2009) 4802-4806. DOI:10.1002/adma.200900983
H. Wang, X. Lv, L. Meng, et al., Chin. Chem. Lett. 29 (2018) 471-474. DOI:10.1016/j.cclet.2017.07.025
C. Adachi, M.A. Baldo, M.E. Thompson, S.R. Forrest, J. Appl. Phys. 90 (2001) 5048-5051. DOI:10.1063/1.1409582
H. Xu, R. Chen, Q. Sun, et al., Chem. Soc. Rev. 43 (2014) 3259-3302. DOI:10.1039/C3CS60449G
K. Hanaoka, K. Kikuchi, S. Kobayashi, T. Nagano, J. Am. Chem. Soc. 129 (2007) 13502-13509. DOI:10.1021/ja073392j
K.Y. Zhang, Q. Yu, H. Wei, et al., Chem. Rev. 118 (2018) 1770-1839. DOI:10.1021/acs.chemrev.7b00425
X. Xiong, F. Song, S. Sun, J. Fan, X. Peng, J. Org. Chem. 2 (2013) 145-149. DOI:10.1002/ajoc.201200109
X. Xiong, L. Zheng, J. Yan, et al., RSC Adv. 5 (2015) 53660-53664. DOI:10.1039/C5RA08539J
T. Li, D. Yang, L. Zhai, et al., Adv. Sci. 4 (2017) 1600166. DOI:10.1002/advs.201600166
W. Hu, L. Guo, L. Bai, et al., Adv. Healthc. Mater. 7 (2018) 1800299. DOI:10.1002/adhm.201800299
F. Ni, Z. Zhu, X. Tong, et al., Chem. Sci. 9 (2018) 6150-6155. DOI:10.1039/C8SC01485J
F. Ni, Z. Zhu, X. Tong, et al., Adv. Sci. 6 (2019) 1801729. DOI:10.1002/advs.201801729
N. Yanai, N. Kimizuka, Acc. Chem. Res. 50 (2017) 2487-2495. DOI:10.1021/acs.accounts.7b00235
T.C. Wu, D.N. Congreve, M.A. Baldo, Appl. Phys. Lett. 107 (2015) 031103. DOI:10.1063/1.4926914
N. Yanai, M. Kozue, S. Amemori, et al., J. Mater. Chem. C 4 (2016) 6447-6451. DOI:10.1039/C6TC01816E
J. Peng, X. Guo, X. Jiang, D. Zhao, Y. Ma, Chem. Sci. 7 (2016) 1233-1237. DOI:10.1039/C5SC03245H
D. Wei, F. Ni, Z. Zhu, Y. Zou, C. Yang, J. Mater. Chem. C 5 (2017) 12674-12677. DOI:10.1039/C7TC04096B
W.L. Chen, F.L. Song, S.L. Tang, et al., Chem. Commun. 55 (2019) 4375-4378. DOI:10.1039/C9CC01868A
X. Li, S. Lee, J. Yoon, Chem. Soc. Rev. 47 (2018) 1174-1188. DOI:10.1039/C7CS00594F
J. Zhang, W. Chen, R. Chen, et al., Chem. Commun. 52 (2016) 11744-11747. DOI:10.1039/C6CC05130H
Z.W. Liu, F.L. Song, W.B. Shi, et al., ACS Appl. Mater. Interfaces 11 (2019) 15426-15435. DOI:10.1021/acsami.9b04488
E. Speckmeier, T.G. Fischer, K. Zeitler, J. Am. Chem. Soc. 140 (2018) 15353-15365. DOI:10.1021/jacs.8b08933
J. Lu, B. Pattengale, Q. Liu, et al., J. Am. Chem. Soc. 140 (2018) 13719-13725. DOI:10.1021/jacs.8b07271
J. Luo, J. Zhang, ACS Catal. 6 (2016) 873-877. DOI:10.1021/acscatal.5b02204
H. Huang, C. Yu, Y. Zhang, Y. Zhang, et al., J. Am. Chem. Soc. 139 (2017) 9799-9802. DOI:10.1021/jacs.7b05082
J.P. Phelan, S.B. Lang, J. Sim, et al., J. Am. Chem. Soc. 141 (2019) 3723-3732. DOI:10.1021/jacs.9b00669