Thermally activated delayed fluorescence in a deep red dinuclear iridium(iii) complex: a hidden mechanism for short luminescence lifetimes

The high luminescence efficiency of cyclometallated iridium(iii) complexes, including those widely used in OLEDs, is typically attributed solely to the formally spin-forbidden phosphorescence process being facilitated by spin–orbit coupling with the Ir(iii) centre. In this work, we provide unequivocal evidence that an additional mechanism can also participate, namely a thermally activated delayed fluorescence (TADF) pathway. TADF is well-established in other materials, including in purely organic compounds, but has never been observed in iridium complexes. Our findings may transform the design of iridium(iii) complexes by including an additional, faster fluorescent radiative decay pathway. We discover it here in a new dinuclear complex, 1, of the form [Ir(N^C)2]2(μ-L), where N^C represents a conventional N^C-cyclometallating ligand, and L is a bis-N^O-chelating bridging ligand derived from 4,6-bis(2-hydroxyphenyl)-pyrimidine. Complex 1 forms selectively as the rac diastereoisomer upon reaction of [Ir(N^C)2(μ-Cl)]2 with H2L under mild conditions, with none of the alternative meso isomer being separated. Its structure is confirmed by X-ray diffraction. Complex 1 displays deep-red luminescence in solution or in polystyrene film at room temperature (λem = 643 nm). Variable-temperature emission spectroscopy uncovers the TADF pathway, involving the thermally activated re-population of S1 from T1. At room temperature, TADF reduces the photoluminescence lifetime in film by a factor of around 2, to 1 μs. The TADF pathway is associated with a small S1–T1 energy gap ΔEST of approximately 50 meV. Calculations that take into account the splitting of the T1 sublevels through spin–orbit coupling perfectly reproduce the experimentally observed temperature-dependence of the lifetime over the range 20–300K. A solution-processed OLED comprising 1 doped into the emitting layer at 5 wt% displays red electroluminescence, λEL = 625 nm, with an EQE of 5.5% and maximum luminance of 6300 cd m−2.


Introduction
2][3][4][5] The archetypal example is the green-emitting, tris-cyclometallated Ir(III) complex fac-Ir(ppy) 3 (ppyH = 2-phenylpyridine).The triplet radiative rate constants (k T r ) of such metal complexes are generally several orders of magnitude larger than those of metal-free room-temperature phosphors. 6The k T r value depends on the spin-orbit coupling (SOC) between the singlet and triplet states, which is in turn determined by the extent to which metal centred d orbitals contribute to the lowest excited states. 7In order to red-shi the emission to obtain red-and near-infrared (NIR)-emitting metal complexes, organic ligands featuring more extended conjugation are required, but this is typically

EDGE ARTICLE
accompanied by a reduction in metal admixtures, with predominantly ligand-centred character to the lowest excited states. 8,9The k T r value suffers as a result, 8,10 limiting the luminescence efficiency that is achievable.Recently, it has been discovered that certain dinuclear complexes of platinum(II) 11,12 and iridium(III) 13 show faster triplet radiative decay, thanks to more efficient mixing of singlet character into the emissive triplet state.5][16] The further acceleration of radiative decay rates through TADFregardless of the amplitude of spin-orbit coupling (SOC) induced by the metal centres 17,18 also has potentially major signicance to the design of blue phosphors, as faster radiative decay helps to attenuate the pathways of emitter degradation in an OLED.
The relative values of phosphorescence k T r and uorescence k S r decay rates, as well as the magnitude of DE ST , determine the extent of involvement of TADF at a given temperature. 1,14While increasing k S r and reducing DE ST generally promotes the thermally activated mechanism, 1,6,19 delayed uorescence will only be apparent if it leads to a faster decay pathway than via the normal phosphorescence (T 1 / S 0 ) route.For this reason, TADF is most commonly found in luminophores displaying long phosphorescence lifetimes, such as in metal-free molecules 20,21 or metal complexes with relatively weak SOC.Contribution of the TADF mechanism to the luminescence is apparent in some complexes of Cu(I), 19,22,23 Ag(I), 24 Au(I), 25 Au(III), [26][27][28] Pd(II), 17,29,30 Pt(II), [14][15][16] and also Zn(II), 31 W(VI) 32 or Sn(IV) 33 complexes where the triplet radiative decay lifetimes span from several microseconds to milliseconds.In Ir(III) complexes, on the other hand, the level of S-T mixing is generally larger than in the former examples, leading to faster triplet decay rates, such that any TADF contribution may be easily overlooked.One study has reported some Ir(III) complexes 34 that demonstrate behaviour consistent with that observed in Pt(II) emitters featuring a TADF contribution, 16 although TADF per se was not identied as the underlying mechanism.
In this work, we demonstrate unequivocal evidence for the role of TADF in actively accelerating radiative decay in the new, red-emitting, dinuclear iridium(III) complex 1 (Scheme 1).The TADF leads to a reduction in the room-temperature radiative decay lifetime of 1 from ∼2 to ∼1 ms.Our study demonstrates that TADF may indeed lead to the shortening of radiative decay lifetimes of Ir(III) complexes, an effect which would otherwise be incorrectly attributed solely to phosphorescence.The application of such an Ir(III) complex in an OLED is presented, where TADF is used for the rst time to accelerate the radiative rate.
2 Molecular design, synthesis, and structure of 1 A range of dinuclear complexes incorporating two cyclometallated iridium(III) centres have been reported. 357][38] The excited-state properties are then typically similar to those of the corresponding individual constituent metal complexes, albeit with energy-transfer potentially occurring between them when the ligand sets differ such that the excited-state energies associated with the two units are different.0][41][42] Our previous work has employed bis-N^C-coordinating ligands based on 4,6-diarylpyrimidines and bis-aryl-substituted thiazolo [5,4-d]thiazoles, to bridge two cyclometallated Ir(III) centres. 13,43In the present work, we explored an alternative bridge, namely the bis-N^Ocoordinating ligand L derived from 4,6-bis(2-hydroxyphenyl) pyrimidine H 2 L, which bridges two Ir(N^C-Meppy) 2 units in the new complex 1 {MeppyH = 2-(p-tolyl)pyridine} (Scheme 1).
The requisite ditopic, bridging proligand H 2 L was synthesized in 67% yield by the palladium-catalysed cross-coupling of 4,6-dichloropyrimidine with 2-hydroxybenzene boronic acid (Scheme 1).Subsequent treatment of the dichloro-bridged complex [Ir(Me 2 ppy) 2 (m-Cl)] 2 with H 2 L in the presence of base (K 2 CO 3 ), in a mixture of chloroform and methanol at reux, gave the desired dinuclear complex 1 in 68% yield aer purication by column chromatography.These conditions are notably milder than those used to introduce bis-N^C bridges, where the activation energy associated with cyclometallation necessitates higher temperatures and the use of Ag + ions to scavenge liberated chloride.The identity and the purity of the complex were conrmed by 1 H and 13 C NMR spectroscopy, elemental analysis and, subsequently, by X-ray crystallography (vide infra).
Due to the intrinsic C 2 (or D 3 ) chirality of bis-and trisbidentate Ir(III) complexes, dinuclear compounds based on Ir(N^C) 2 {or Ir(N^C) 3 } units may comprise of a mixture of meso (DL) and rac (LL, DD) diastereomers.That can be problematic, as the diastereomers may display different properties from one another.The use of enantiomerically pure mononuclear building blocks can circumvent the problem, but such chiral separation is oen difficult and not feasible on larger scales, and the starting congurations must also be retained under the conditions required to introduce the bridge.Another strategy that has been used to avoid the formation of diastereomeric mixtures is to employ non-stereogenic metal centres wherein the metal ions are coordinated by symmetric tridentate ligands. 34evertheless, in some binuclear systems comprising two bisbidentate unitsespecially when a short linker is used such that steric interactions between the units inuences the relative stabilities of the productsthe formation of one diastereomer may occur diastereoselectively, or even diastereospecically.Such is the case, for example, in the synthesis of the well-known chloro-bridged dimers of the form [Ir(N^C) 2 (m-Cl)] 2 . 44The product is uniquely the rac pair (LL and DD).The meso (DL) product is apparently disfavoured through steric interactions of the N^C ligands on neighbouring metal centres, owing to their proximity.
In the synthesis of 1, our tentative expectation was that the short distance between the metal centres, dictated by the compact pyrimidine bridging unit, would similarly lead selectively to the rac diastereoisomer.And, indeed, only one diastereomer was isolated from the chromatography column, based on 1 H and 13 C spectroscopy.X-ray diffraction analysis of a crystal of 1 conrms the hypothesis.The molecular structure (Fig. 1; CCDC no.2288521) shows the expected structure, with the two Ir(N^C) 2 centres both N^O-coordinated by the bridging ligand.The molecule in the crystal is located on a 2-fold rotation axis along C1/C3 (see also Scheme 1), such that the two metal centres have the same conguration; i.e., LL (the enantiomer shown in Fig. 1) or DD.The Ir/Ir distance is 6.086(2) Å. Apparently, then, the choice of compact bridging ligand ensures that the reaction proceeds with a level of diastereoselectivity such that the meso isomer is not formed in signicant amounts.Note that the three heterocyclic nitrogen atoms around each Ir(III) centre are coordinated in a meridional arrangement, with the two pyridine rings occupying positions trans to each other.This is the same as the conguration observed in the chlorobridged dimers [Ir(N^C) 2 (m-Cl)] 2 .Thus, the mer / fac rearrangement that typically occurs during the formation of Ir(N^C) 3 complexes thermallywhich necessitates higher temperatures than those used hereis not observed.In those cases, the mer isomer is destabilised relative to the fac owing to two of the strongly s-donating cyclometallating rings being positioned trans to one another.In 1, the cyclometallated rings are not trans to one another.The outcome is essentially the same as what is usually observed for [Ir(N^C) 2 (N^N)] + complexes when prepared under similarly mild conditions from the chloro-bridged dimers, but with the N^O unit in place of the N^N.

Absorption and steady-state emission spectra
The absorption and photoluminescence spectra of 1 are shown in Fig. 2, with corresponding numerical data compiled in Table 1.The absorption spectrum is quite typical of triscyclometallated iridium(III) complexes in that it displays a set of quite intense bands of 3 of the order of 10 4 M −1 cm −1 in the visible region, attributed to charge-transfer transitions, as well as bands around 4× more intense in the far UV, associated with ligand-centred transitions.The visible-region bands are, however, substantially red-shied compared to the related mononuclear complex Ir(Meppy) 3 , and the long-wavelength tail of 1 extends to around 600 nm compared to scarcely beyond 500 nm for Ir(Meppy) 3 .The red-shi is consistent with our previous work on pyrimidine-bridged dinuclear Ir(III) and Pt(II) complexes, attributed primarily to the lower-energy p* orbitals associated with the bis-coordinated pyrimidine that lead to correspondingly lower-energy 1,3 MLCT transitions.
The complex displays deep red photoluminescence (PL), giving a broad, featureless spectrum in solution at room temperature, typical of Ir(III)-based 3 MLCT emitters.As in absorption, the emission is strongly red-shied relative to the mononuclear analogue: l em = 655 nm for 1 versus 510 nm for Ir(Meppy) 3 .There is no signicant solvatochromism, neither in absorption nor emission, as is clear from Fig. 2 which shows the spectra in toluene and chlorobenzene superimposed on those in 2-MeTHF.The PL quantum yield of 0.30 in toluene is respectable for a deep-red emitter; values are slightly lower in the other two solvents (Table 1).
It is important to notice the clear overlap of the tail of the absorption band with the onset of the PL spectrum, around 550-580 nm.Such behaviour is not normally expected for a triplet emitter, if the lowest-energy absorption band of signicant intensity is indeed a singlet, unless the S-T gap is very small (as would be required for a TADF contribution).][16] However, in the case of Ir(III) complexes, the strength of singlettriplet mixing through SOC is sufficient to enhance the oscillator strength of the normally strongly forbidden S 0 / T 1 transitions in the absorption spectrum, to the point that the bands have quite high molar absorptivities.The observation of pronounced overlap of absorption and emission is thus not, on its own, sufficient to infer a TADF contribution to the PL.
The temperature-dependence of the PL was recorded in toluene over the range 160-300 K.The evolution of the steadystate spectra with temperature is suggestive of two luminescent components: a higher-energy component, rather broad and featureless, which dominates at higher temperatures, and a lower-energy component with a discernible vibronic shoulder at lower temperatures (Fig. 3).A clear transition between the two spectral proles is observed with a distinct iso-emissive point, indicating that the two components arise from two separate emissive states in equilibrium.This experimental picture is consistent with the behaviour of diplatinum(II) complexes which display TADF, 14,16 namely the lower-energy component being due to phosphorescence from T 1 and the higher to uorescence from S 1 , thermally populated from T 1 .Our attempts to deconvolve the luminescence spectra into separate phosphorescence and uorescence spectra (which can be useful in understanding the thermodynamics 45 ) were hampered by thermal broadening effects inuencing the proles at different temperature.

Time-resolved photoluminescence
The PL intensity of 1 displays mono-exponential decay, with lifetimes s obs in the range 0.47-0.85ms at room temperature according to the solvent (Table 1).The variation of the experimentally observed lifetime with temperature s obs (T) (Fig. 4) ts well to the model described by eqn (1), 1,47 where DE ST is the S 1 -T 1 energy gap in J mol −1 ; s PH is the phosphorescence lifetime (s); k S r is the radiative rate constant of S 1 (s −1 ); R is the universal gas constant, 8.314 J mol −1 K −1 ; and T is the temperature in K.The best t gives a value of DE ST = 47 ± 7 meV and k S r of (1.2 ± 0.2) × 10 7 s −1 (corresponding to a natural radiative lifetime for  S 1 of about 83 ns).The DE ST is a similar magnitude to that reported previously in a diplatinum(II) TADF complex, 16 while k S r is consistent with values observed for charge-transfer singlet states in metal-free TADF emitters. 48obs ðTÞ ¼ 4 Solid-state photophysics The temperature dependence of the PL spectra and lifetimes of 1 in polystyrene lm were investigated over the range 300-20 K (Fig. 5 and S5.4-5.6 †).Access to the lower temperature range <80 K is insightful as it allows the splitting of the sublevels of T 1 to be studied. 1,49The energy gap DE 1,3 between sublevels 1 and 3 (DE 1,3 )also known as the zero-eld splitting (ZFS)is oen of the order of tens of cm −1 in 3rd row complexes, 1 such that the thermal population of sublevel 3 inuences the observed lifetime over the temperature range 20-80 K (DE 1,2 is usually much smaller and so even lower temperatures <20 K are required to probe it).The magnitude of ZFS is associated with the strength of the SOC between the singlet and triplet manifolds, and a correlation has been observed between it and the phosphorescence radiative rate constant of metal complexes through the work of Yersin and co-workers in particular. 1 The lifetime data in Fig. 5 are tted to eqn (2), which takes into account the thermal equilibrium between the 1 and 3 sublevels of the lowest triplet state while assuming that DE 1,2 is sufficiently small for the 1 and 2 sublevels to be considered equally populated over the temperature range used.
The PL lifetime of 1 increases only modestly from 1.1 ms at 300 K to 2.3 ms at 100 K, but then increases steeply to 6.5 ms at 21 K.Of course, it is the experimentally observed lifetime s obs (T) that is measured, rather than the natural radiative lifetime.However, since the steady-state PL intensity remains essentially invariant over a wide range of temperatures including <100 K (Fig. S5.6 †), it may be concluded that the variation in s obs (T) is largely reecting changes in the radiative rate as opposed to simple suppression of non-radiative decay.The initial increase in s obs can then probably be associated with the lowering inuence of the TADF mechanism of re-population of S 1 as the temperature is reduced, and then the subsequent larger increase is attributed to sublevel 3 being less thermally populated from sublevels 1 and 2 at the lowest temperatures.The experimental data give an excellent t to eqn (2) (blue dashed line in Fig. 5b).On the other hand, if the data are tted to a form  of eqn (2) in which no TADF mechanism is included (i.e., where DE ST = N and/or k S r = 0), the t is poor (red dotted line).The modelled difference between the two scenarios (i.e., with or without TADF) suggests that the effect of delayed uorescence is to halve the radiative decay lifetime of 1 at room temperature.The t that incorporates TADF gives values of DE ST = 53 ± 17 meV, DE 1,3 = 5 ± 1 meV (=40 cm −1 ), and k S r = (8 ± 5) × 10 6 s −1 (or s S approx.100 ns), values which are consistent with those obtained in toluene (vide supra).On the other hand, if the TADF mechanism is excluded, the ZFS would need to be 430 cm −1 to account for the behaviour.As that value is more than double the largest-ever recorded ZFS in Ir(III) complexes with the highest contributions of MLCT character to the emitting states, it is improbable that efficient SOC alone could be responsible for the fast decay of 1.
Finally, we note that the time-resolved photoluminescence spectra reveal an invariance of the PL spectrum with time delay at any temperature (Fig. S5.5 †), which conrms that the delayed uorescence and phosphorescence have the same lifetime and thus that these two emissive states are in equilibrium.It also rules out intermolecular processes being responsible for delayed uorescence.Collectively, the observations provide very strong evidence for TADF as the sole up-conversion mechanism. 47,50Calculations Density functional theory (DFT) and time-dependent DFT (TD-DFT) as well as the quasi-degenerate perturbation theory (QDPT) 51,52 with zeroth-order regular approximation (ZORA) 53,54 implemented in Orca 55,56 were used to gain additional insight into the luminescence mechanisms operating in 1.The ground state (S 0 ) and triplet excited state (T 1 ) geometries were optimised at the BP86 (ref.57)/def2-TZVP 58 /CPCM(toluene) level of theory.Singlet and triplet radiative rate constants were calculated using the ZORA-corrected def2-TZVP basis sets 58 for light atoms and a segmented all-electron, relativisticallycontracted (SARC) def2-TZVP basis set for Ir.
Spin-orbit coupled excited states (SOC states) are represented as mixed states with singlet and triplet admixtures.They are summarised in Table S4.1 and Fig. S4.2 in the ESI.† The lowest four SOC states are considered, which represent the triple-degenerate sublevels of the T 1 (states 1 to 3), and the state 4, interpreted as S 1 .States 1 to 3 are dominated by T 1 character with ∼10% admixtures from other states (mainly upper triplet states T n ).State 4 is 81% S 1 with triplet state admixtures.The orbital topology of the S 1 and T 1 states is therefore considered.S 1 has dominant HOMO / LUMO character (>98%), while T 1 is also dominated by the HOMO/LUMO transition (92%), but with small contributions from HOMO-1 / LUMO (4%) and HOMO-3 / LUMO (1%) (Fig. 6).The HOMO and HOMO-1 involve d orbitals on a different metal centre and ligand orbitals coordinating the respective metal ion: the oxygen p orbital of the phenolate ligand, as well as the phenyl p orbitals of the C,N ligands (Fig. 6).HOMO-3 is similar to HOMO and HOMO-1 but it involves both metal centres.The LUMO is localised on the diphenylpyrimidine bridge, as in Pt(II) complexes containing related structural motifs. 14,16Such orbital topology of S 1 and T 1 states gives them a clear MLCT + IL (interligand) character.Since the two lowest excited states differ in orbital topology, a relatively large < T 1 jH SO jS 1 > = 91 cm −1 arises (cf.values of the order of ∼1 cm −1 expected in metal-free TADF emitters 59 ), suggesting that direct S 1 4 T 1 ISC and reverse-ISC (RISC) should be fast.This does not exclude other states being involved in the RISC/ISC process, but rather indicates that the S 1 4 T 1 spin-ip may be sufficiently fast on its own to explain the experimental behaviour of the complex.
A relatively small splitting between the lowest SOC states is observed, with DE 1,2 = 1 meV (or 8 cm −1 ) and DE 1,3 = 4 meV The temperature dependence of the luminescence lifetime of 1 has been modelled using eqn (3), which takes into account the radiative decay of all four of the SOC states, using the calculated energy gaps between them (here, k T1À3 r represent the radiative rates of states 1 to 3).The plot of the calculated data (Fig. 7b) perfectly reproduces the observed changes in the photoluminescence lifetime, reecting rst the thermally activated occupation of the higher triplet sublevels at the lowest temperatures and then the population of the S 1 (i.e., the TADF contribution).The excellent agreement between the calculated and experimental data thus demonstrates the importance of the TADF mechanism to the fast luminescence decay of 1.Its behaviour at room temperature (indeed, any temperature above about 150 K) cannot be adequately explained without inclusion of the TADF contribution.

OLED devices
Fast radiative decay in the red region of the spectrum evidently renders complexes like 1 of great interest as emitters for deepred OLEDs.Proof-of-concept, solution-processed OLEDs have therefore been fabricated in this work, using 1 as the luminescent dopant.The electrical and electroluminescent characteristics of the OLEDs are presented in Fig. 8, with pertinent numerical data in Table 2.The OLED structure used in this Fig. 6 Relevant molecular orbital iso surfaces (iso = 0.05) for 1 calculated at the T 1 geometry.work has been adapted from our previous studies, 14,60 but we used a different hole-transport material to better match the HOMO of the luminescent dopant: ITOjAI4083 (30 nm)jTCTA : PO-T2T (80 : 20) co. 1 (x%) (65 nm)jPO-T2T (50 nm)jLiF (0.8 nm)j Al (100 nm).It comprises of PEDOT AI4083 as the hole-injection layer; an emissive layer consisting of a blend host for 1 of TCTA {4,4 ′ ,4-tris(carbazol-9-yl)triphenylamine} as a hole-transport component and PO-T2T {2,4,6-tris [3-(diphenylphosphinyl)  phenyl]-1,3,5-triazine} as an electron-transport component; PO-T2T as the electron-transport layer; LiF as the electroninjection layer; an Al cathode.
The of doping concentration on OLED properties has been studied, showing that low loads lead to higher efficiency.Interestingly, a distinct red-shi of the intensity-normalised electroluminescence (EL) spectrum is observed as the concentration of 1 increases; e.g.l EL = 625 and 635 nm at 5 and 12% w/ w respectively.Given the overlap between the absorption and PL spectra observed in solution, it is likely that the self-absorption of EL within the emissive layer is the most likely reason for the apparent red-shi.The most efficient OLED Device 1 displays a maximum external quantum efficiency (EQE) of 5.5% and a maximum luminance of 6300 cd m −2 .

Conclusions
We present here the rst example of an iridium(III) complex where a TADF mechanism has been demonstratedboth computationally and experimentallyto accelerate the luminescence decay signicantly, and the rst report of the use of such an emitter in an OLED.This work proves the concept that TADF can lead to an important, additional route to accelerating the luminescence decay of iridium(III) complexes.It complements the spin-orbit coupling model based solely on partially allowed phosphorescence, which has hitherto always been assumed to be the only pathway for radiative decay of Ir(III) materials.The mechanism leads to a k r of 3.5 × 10 5 s −1 at room temperature, a remarkably high value for such a deep red emitter new, and it arises from the small ∼50 meV S 1 -T 1 gap.An OLED incorporating 1 displays an EQE of 5.5% with l EL = 625 nm.
In this work we present a vital, new luminescent mechanism in iridium(III) complexes that may profoundly change the way one would design metal complexes with a short photoluminescence decay.Evidently, TADF can signicantly shorten PL lifetimes even if not dominating the PL spectrum, and this feature can readily be exploited in multiple areas of research.Shortening the luminescence lifetimes of iridium(III) complexes using our new strategy may be relevant, for example, to the development of blue and NIR OLEDs, where short decay times are crucial to reduced device degradation and increased efficiency of long-wavelength EL, respectively.
Our work extends the understanding of the possible emissive pathways available for iridium(III) complexes and beyond the spin-orbit coupling-based model.TADF is likely to prove to be  an overlooked pathway in metal complexes.It does not undermine the importance of the heavy atom effect and phosphorescence, but rather serves as a potentially powerful alternative strategy for designing more efficient luminescent complexes.

Fig. 1
Fig.1The molecular structure of dinuclear complex 1 in the crystal.

Fig. 2
Fig. 2 Absorption and photoluminescence (l ex = 365 nm) spectra of 1 in the three solvents indicated at a concentration of approximately 10 −5 M.

Fig. 3
Fig. 3 (a) PL spectra of 1 in toluene (c = 10 −5 M) at various temperatures from 300 to 160 K. (b) The same set of spectra shown normalised at the l max values.l ex = 365 nm.

Fig. 4
Fig.4The experimentally observed PL decay lifetime of 1 (monitored at 640 nm) as a function of temperature (black circles) together with the best fit of eqn (1) to the data (red line).

Fig. 5
Fig. 5 (a) The PL spectra of 1 in polystyrene matrix at (0.1% w/w) at the temperatures indicated.(b) The corresponding variation in s obs .The experimental data points are represented by black circles and the best fit to eqn (2) by the dashed blue line.The corresponding "fit" in the absence of the TADF component is shown by the dotted red line.l ex = 355 nm.

Fig. 7
Fig. 7 (a) Computational model of the photophysics of 1; (b) simulated natural lifetime, s 0 , of 1 as a function of temperature obtained using eqn (3).The grey line represents the lifetime simulated without inclusion of a TADF component (i.e., phosphorescence only, k S r = 0 or DE ST = N), and the red line the corresponding simulation in which the TADF component to the decay is included.

Table 1
Summary of the spectroscopic properties of 1 in the solvents indicated at 295 K a Absorbance maxima and corresponding extinction coefficients.bEmissionmaxima.cPhotoluminescencequantum yield recorded against rhodamine 6G in ethanol (F PL = 0.91 (ref.46)).d Experimentally determined photoluminescence lifetime.e Radiative k r and non-radiative k nr rate constants, estimated assuming that the emitting state is formed with unit efficiency such that k r = F PL /s and k nr = (1 − F PL )/s.