Alfiya F.
Suleymanova
a,
Marsel Z.
Shafikov
b,
Xinrui
Chen
c,
Yafei
Wang
*c,
Rafal
Czerwieniec
*b and
Duncan W.
Bruce
*a
aDepartment of Chemistry, University of York, Heslington, YORK YO10 5DD, UK. E-mail: duncan.bruce@york.ac.uk; Tel: (+44) 1904 324085
bInstitut für Physikalische Chemie, Universität Regensburg, Universitätsstrasse 31, 93053 Regensburg, Germany. E-mail: rafal.czerwieniec@chemie.uni-regensburg.de; Tel: (+49) 941 943 4463
cSchool of Materials Science & Engineering, Changzhou University, Changzhou 213164, P. R. China. E-mail: qiji830404@hotmail.com
First published on 29th August 2022
The device performance is reported for three compounds which show both thermally activated delayed fluorescence and liquid crystallinity, and use the donor 3,6-bis(3,4-didodecyloxyphenyl)carbazole. Two of the compounds, whose photophysics were reported previously, are based on a terephthalonitrile acceptor. A third and new compound is based on an isophthalonitrile acceptor and shows a more temperature-accessible mesophase and enhanced solution emission quantum yield. Two of the compounds show device external quantum efficiencies of between 2–3% and exhibit very small efficiency roll off. The responses are evaluated in terms of the specific nature of the materials.
Recently, there has been increasing interest in thermally activated delayed fluorescent (TADF) materials showing up to 100% of internal quantum efficiency without the need for a heavy transition metal such as iridium.12–14 This is due to the design of molecules with a minimised energy gap between the lowest excited singlet (S1) and triplet (T1) states, thereby promoting efficient reverse intersystem crossing (rISC) from the non-radiative triplet state to the radiative singlet state.15,16 Although TADF as a phenomenon has been known for many decades,17–19 it became very popular in application when Adachi and co-workers first demonstrated use of TADF molecules in organic light-emitting diodes (OLEDs).20–22 Today, these materials are used in many different areas such as time-resolved fluorescence imaging,23–27 organic photocatalysis,28–31 photodynamic therapy32–35 and in improving the efficiency of OLED devices.36,37
Constructing molecules exhibiting both LC and TADF properties is an interesting approach for potential application in OLED devices and, for example, there is a very recent report of LC properties in some disc-like, multi-resonant TADF materials.38 Some of us have previously reported LCs based on dihydroacridine donor and diphenylsulfone acceptor, which show external quantum efficiencies in devices of up to 15%,39 whereas in another, earlier study, we reported multifunctional materials based on carbazole-substituted terephthalonitriles (Fig. 1, 1, 2).40 In this latter work, modification of well-known donor–acceptor systems14 led to the combination of TADF and LC properties in one material, but the contribution of TADF to the overall emission was only 5–10% and the emission quantum yields of the materials were low. Nonetheless, we were keen to investigate the behaviour of these materials in device constructs, which we now report along with studies on a new LC TADF emitter (3) based on isophthalonitrile and a carbazole and showing increased quantum yield and TADF contribution.
g • 35 • Colh • 76 • Iso |
![]() | ||
Fig. 2 (a) Polarising optical microscopy image of 3 at 76 °C on cooling (magnification 20×) and (b) small-angle diffraction pattern recorded at 63.7 °C. |
Time-correlated single photon counting (TCSPC) measurements in degassed toluene solution at room temperature reveal that the emission decay kinetics consist of a fast process with a decay time of 7.5 ns and a slower process with a decay time of 920 ns (Fig. 4). When re-measured following air-equilibration, the fast process does not show a notable change in kinetics whereas the decay time of the slower process shortens to 360 ns. Sensitivity of the decay rate to oxygen indicates the involvement of the triplet state, which can decay via Dexter-type energy transfer to molecular oxygen. On this basis, the shorter-lived component of emission is assigned to the prompt fluorescence (S1 → S0) and the long-lived part to delayed fluorescent emission (also S1 → S0) following S1 → T1 intersystem crossing and the thermally activated reverse intersystem crossing (rISC) T1 → S1, i.e. via thermalisation of the S1 and T1 states. It is noted that the contribution of T1 → S0 phosphorescence to the longer component is improbable as in purely organic materials it is typically observed only at low temperatures and characterised with decay times in the range of milliseconds and above. The overall emission quantum yield of compound 3 in degassed toluene at room temperature is 11% which is a notable increase when compared to the previously reported terephthalonitrile-based analogues 1 and 2 (Table 1). The integrated contribution of the TADF to the total emission intensity is about 10% which is also an increase compared to compounds 1 and 2, revealing the relative superiority of the molecular design of 3 as a TADF material.
λ em/nm | φ PL | τ PF/ns | τ DF/s | k PF /s−1 | k nr(S1) | k ISC | |
---|---|---|---|---|---|---|---|
a Data for 1 and 2 from ref. 39. b Values in parentheses represent percentile share of delayed emission intensity in the total quantum yield. c Calculated as kPF = φPF(S1)/τPF. d Calculated as knr(S1) = (1 − (φPF + φDF)/τPF, e.g., assuming that all non-radiative relaxation to the ground state occurs via the S1 → S0 pathway, thus representing the upper limit. e Calculated as kISC = φDF/τPF, e.g., assuming that all non-radiative relaxation to the ground state occurs via the S1 → S0 pathway, thus representing the lower limit. | |||||||
1 | 560 | 5 (5) | 6.9 | 3.5 | 0.6 × 107 | 1.5 × 108 | ≥3.6 × 105 |
2 | 620 | 1.5 | 1.2 | ≤1 | 1.3 × 107 | 8.2 × 108 | |
3 | 585 | 11 (10) | 7.5 | 0.92 | 1.3 × 107 | ≤1.2 × 108 | ≥1.5 × 106 |
In the ground state (S0), the planes of the carbazole groups stand oriented with respect to the plane of the isophthalonitrile acceptor at dihedral angles of 78.3° (Cz1), 78.3° (Cz2), 79.0° (Cz3) and 63.2° (Cz4). This separates the π systems of the carbazole derivatives (donors) and isophthalonitrile moiety (acceptor), which is a prerequisite for a small value for ΔE(S1–T1) and hence TADF behaviour. In the T1 state, upon relaxation of the molecular geometry, the dihedral angle between the planes of the carbazole derivatives and of the isophthalonitrile moiety change slightly to 77.9° (Cz1), 76.5° (Cz2), 78.0° (Cz3), and increase by a few degrees to 68.9° (Cz4).
According to TD-DFT calculations, conducted with a geometry optimised in the T1 state, the S1 and T1 states originate from HOMO → LUMO transitions. The HOMO (−5.175 eV) is localised on the π–π-stacked carbazole units with the major contribution of Cz2 (84%) and minor contribution of Cz1 (7%) and Cz3 (8%). In contrast, the LUMO (−2.920 eV) is localised primarily on the iso-phthalonitrile moiety (93%) and has just 7% of the electron density on the two nitrogen-phenyl bonds of Cz1 and Cz3 (Fig. 6 and Table S1 in the ESI†). Thus, states S1 and T1 have predominantly charge-transfer character allowing for a weak exchange interaction and consequently a small value of the energy gap between them (S1 = 1.632 eV and T1 = 1.618 eV). Indeed, the calculated value of the gap is just ΔE(S1–T1) = 110 cm−1 (13.7 meV), indicating that the S1 state can easily be populated at room temperature by thermal activation from T1 through the rISC process. It is noted that although the computed excited state energies are underestimated in comparison to the experimental data, it probably has only a weak effect on the energy gap between them. The small computed ΔE(S1–T1) gap value agrees well with our assignment of the slow component of the TCSPC decay profile to TADF. It is also worth noting that, according to the TD-DFT calculations, the higher-lying excited states S2/T2 (HOMO−1 → LUMO) and S3/T3 (HOMO−2 → LUMO) are also charge transfer in character, originating from different carbazole groups in the molecule (Fig. 6). This is consistent with the absorption spectrum, which features >1 charge-transfer bands at longer wavelengths.
![]() | ||
Fig. 6 Iso-surface contour plots (iso-value = 0.03) of selected molecular orbitals of compound 3 calculated at T1 state geometry. Hydrogens are omitted for clarity. |
Comparing now compound 3 to the previously reported, isoelectronic compound 2 (Fig. 1), calculated at the same theory level, the latter has lower energies of both HOMO (−5.430 eV) and LUMO (−3.343 eV). The ΔE(HOMO–LUMO) energy gap, however, decreases from 3 (2.255 eV) to 2 (2.087 eV) as in 2 the relative stabilisation of the LUMO is greater than that of the HOMO. The larger ΔE(HOMO–LUMO) gap of 3 results in relatively higher-lying S1 and T1 states as manifest by the blue shift of the emission from 2 (λmax = 620 nm) to 3 (λmax = 585 nm) accompanied by an increase in quantum yield according to the energy gap law.48
Notably, the π–π-stacking keeps the three carbazole units in 3 at much larger dihedral angles with respect to the isophthalonitrile, as compared to the analogous phthalonitrile based compound 1 (Fig. 1) where the dihedral angle is about 55° (computed at the same level of theory – see ESI† for computational data). Consequently, compound 1 has a much larger value of ΔE(S1–T1) = 1270 cm−1 (0.158 meV), resulting in less efficient thermal population of emissive S1 states via rISC and more efficient, non-radiative T1 → S0 relaxation. Thus, a higher emission quantum yield for 3 is observed, despite the lower energy of the emitting state (Table 1) compared to 1.
The significantly higher ISC rate of 3, compared to 1 (Table 1), also agrees with the importance of molecular vibrations within the π–π-stacked carbazole units. Such vibrations, displacing the atoms out of the conjugation plane of the carbazole groups, admix the σπ* and σπ* character to the primary ππ * character of S1 and T1 states. This enhances the spin–orbit coupling of the two states which will enhance the rates of S1 ↔ T1 ISC/rISC.40,49
The discussions presented above reveal the overall advantageous molecular design of 3 as a TADF material when compared to 1 and 2.
Compound | Dopant (wt%) | V on (V) | L max (cd m−2) | CEmax (cd A−1) | EQEmax (%) | CIE (x, y) | λ (nm) |
---|---|---|---|---|---|---|---|
1 | 1 | 8.0 | 994.5 | 6.82 | 2.29 | (0.40, 0.54) | 552 |
1 | 3 | 6.8 | 2065 | 6.58 | 2.19 | (0.42, 0.53) | 558 |
1 | 6 | 11.6 | 1148 | 2.00 | 0.76 | (0.44, 0.52) | 564 |
2 | 1 | 8.4 | 274.5 | 1.13 | 0.66 | (0.53, 0.43) | 598 |
2 | 3 | 6.8 | 407.2 | 1.20 | 0.59 | (0.54, 0.41) | 606 |
2 | 6 | 8.8 | 222.9 | 0.40 | 0.26 | (0.56, 0.42) | 610 |
3 | 1 | 8.0 | 1387 | 7.69 | 2.97 | (0.48, 0.49) | 578 |
3 | 3 | 8.4 | 1704 | 6.52 | 2.60 | (0.49, 0.49) | 578 |
3 | 6 | 10.4 | 837.1 | 3.27 | 1.50 | (0.51, 0.48) | 588 |
As shown in the inset to Fig. 8a, all devices show broad EL spectra, similar to the PL profiles and, compared to compound 1, compounds 2 and 3 exhibit red-shifted EL spectra owing to the greater number of donor moieties leading to a stabilisation of the HOMO. This is most pronounced in compound 2 where, in addition, the cyano acceptor groups are conjugated leading to emission at about 604 ± 6 nm. However, there is an additional emission band at about 420 nm for the device fabricated using compound 2. This is assigned to the host matrix and implies incomplete energy transfer between host and emitter, suggesting inferior device performance.
![]() | ||
Fig. 8 Device performance at 1 wt% dopant concentration: (a) EQE-luminance curves (inset: the EL spectra); (b) J–V–L curves. |
As seen from Table 2, the EL spectra have a clear red shift with the increasing dopant concentration, moving between 10 and 12 nm from 1 to 6% doping. This is interpreted as reflecting the LC nature of the materials and the driving force to self-organise leading to relatively strong intermolecular interactions. In addition, the devices exhibit a relatively high turn-on voltage of around 8 V (@ 1 cd m−2), which may be attributable to the rather high coverage of the periphery of the chromophore by flexible chains adversely affecting the carrier mobility (Fig. 8b). Thus, a passable device performance with an external quantum efficiency (EQE) of 2.97% is achieved for a device using compound 3 doped into the host at a concentration of 1 wt%, with a concomitant current efficiency of 7.69 cd A−1 and luminance of 1387 cd m−2. Compound 1 has a not dissimilar efficiency of 2.29% at 1 wt% dopant, whereas compound 2 shows a much poorer EQE of 0.66%, also doped at 1 wt%. As expected, the device performance decreases with the increased dopant concentration due to the concentration quenching. Impressively however, the devices based on compounds 1 and 3 display a very small efficiency roll-off (Fig. 8a) with EQEs of 2.85 and 2.24% at 100 cd m−2 and 2.1 and 1.8% at 500 cd m−2, respectively. These results demonstrate that both compounds 1 and 3 show very stable in the device.
Impressively, compound 3 showed enhanced emission quantum yield compared to compound 1 and 2, leading to it showing the best device performance with an acceptable EQEmax value in a liquid crystalline-based TADF OLED and, in particular, the very small efficiency roll-off is a very pleasing observation. Nonetheless, in order to avoid concentration quenching, the loading of the emitter in the host is kept low (1–2%), which precludes the expression of the liquid crystal properties. However, in the only other study of LC-TADF materials in devices, 20 wt% loading of the luminophore were achieved,39 which is a very encouraging observation and provides strong motivation to continue the development of LC-TADF emitters in order for the potentially advantageous properties, which include enhanced charge-transport, to be realised.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cp02684h |
This journal is © the Owner Societies 2022 |