Introduction

Since the seminal publication by Tang and Van Slyke in 1987, Organic Light-Emitting Diodes (OLEDs) have emerged as a dominant technology within the display and lighting industries1. Significant progress in the OLED field has been driven by the development of new energy-efficient materials and OLED fabrication architectures2. The first generation of OLED devices utilized organic fluorescent emitters. When charge is injected into an OLED device, it results in the simultaneous generation of one singlet and three triplet excitons, as determined by spin statistics. However, fluorescent emitters limit the internal quantum efficiency (IQE) of first-generation OLEDs to 25% due to nonradiative loss of the triplet excitons3. In contrast, the second and third generations of OLED devices achieve up to 100% IQE by harvesting all singlet and triplet excitons. This improvement is achieved through phosphorescence and thermally activated delayed fluorescence (TADF) mechanisms4,5,6,7,8,9,10.

A significant challenge in the OLED field lies in developing energy-efficient blue OLEDs with high operating stability11. Phosphorescent and TADF emitters often suffer from self-quenching due to triplet–triplet or triplet-polaron annihilation processes12,13. To mitigate this, the “solid-state dilution” approach is employed, necessitating the use of a host matrix12,13. Similar to emitter molecules, host materials must meet several criteria to enable energy-efficient and stable blue OLEDs. These include (1) a higher triplet (T1) energy level compared with the emitter molecules to prevent Dexter energy transfer to the host; (2) the highest occupied (HOMO) and lowest unoccupied (LUMO) molecular orbital energy levels for the host should confine those of the emitter molecule; (3) balanced charge-carrier (electron and hole) injection and mobility within the host; (4) high thermal and redox stability; (5) high volatility during the deposition process; (6) ability to enable slow roll-off of efficiency at high luminance, and many more14. All these requirements pose a unique challenge for blue OLED hosts, limiting the number of appropriate materials. Scheme 1 presents some of the most frequently used high triplet energy materials, illustrating various molecular design strategies used to achieve the necessary energetic properties for the blue OLED host materials15,16,17. Synthetic protocols to obtain such hosts (Scheme 1) are optimized but involve expensive reagents and lengthy multistep syntheses which frequently employ transition metal catalysts that contribute to the final material cost. Additionally, the environmental impact of producing advanced materials (Scheme 1) is increased by necessary purification steps, such as a double sublimation18. In this work, we introduce a novel molecular design for a blue host material containing a 1,3,5-oxadiazine core which is unprecedented in the literature. Such NON-type materials are synthesised in a simple ‘one-pot’ reaction using inexpensive reagents in a good yield. We demonstrate that this new NON-type (Scheme 1) host material is compatible with deep-blue emitters. An external quantum efficiency (EQE) up to 21% was achieved in OLEDs with a device performance on par with hosts of complex molecular design, such as DPEPO19 or PPBi20.

Scheme 1
scheme 1

Molecular structures for organic host materials. Molecular structure for popular DPEPO19, PPBi20, UGH342, oCBP29,43 and NON (R = H, CF3 or Br) OLED host materials with triplet energy level more than 3.0 eV.

Full size image

Results and discussion

Synthesis and characterization of NON-materials

The novel 1,3,5-oxadiazine compounds NON, NON-CF3 and NON-Br were synthesized in one step by heating the substituted arylhydrazine hydrochlorides and cyclohexylcarbaldehyde in glacial acetic acid at 60 °C for 3 h (Scheme 2). The reaction forms three heterocyclic rings and two stereogenic centres (C2 and C3 carbon atoms, Fig. 1). The reaction yields varied from moderate for NON and NON-Br (22–26%) to high for NON-CF3 (82%). Garg et al.21 previously reported that lactols and hemiaminals behave as masked carbaldehydes under the conditions of a Fisher indole synthesis, resulting in the formation of various indoline derivatives with two heterocyclic rings and two stereogenic centres (Scheme 2, top). Therefore, we propose that the formation of the NON-R compounds occurs via an interrupted Fischer indolization22,23. It’s worth noting that the 1,3,5-oxadiazine moiety is a powerful heterocyclic framework with applications in pharmaceuticals, agriculture, and industry. This makes the facile synthesis of NON-R (R = H, CF3 and Br) materials interesting in its own right24,25,26,27,28.

Scheme 2
scheme 2

Synthetic protocols. One-step synthesis of the compounds NON, NON-CF3 and NON-Br.

Full size image
Fig. 1: X-ray crystal structure of NON-compound.
figure 1

X-ray crystal structure of NON-compound front view (left, hydrogen atoms omitted for clarity) and side view (right, with selected hydrogen atoms). Ellipsoids are shown at the 50% level.

Full size image

All NON-compounds are white solids, stable in air and soluble in common polar organic solvents dichloromethane, acetone, THF and poorly soluble in hexane. The purity of the products was confirmed by 1H, 13C and 19F NMR, High-Resolution Mass Spectrometry and X-ray diffraction analysis. The thermal stability of NON-materials increases from NON-Br (235 °C) to NON (256 °C) to NON-CF3 (262 °C, Fig. S12). The central oxadiazine ring possesses two groups of distinctive proton resonances: the first as a doublet at 4.9 ppm for C1–H1, and the second is a singlet at 5.2 ppm for C2–H2 and C3–H3 (see Supplementary File). Crystals of NON-materials were grown by layering dichloromethane solutions with hexanes (Figs. 1, S8–S10). All three compounds possess meso-stereochemistry (R for C2 and S for C3). Similar stereochemical observations for the fused 1,3,5-oxadiazine rings have been reported by Ghinet et al.24. The X-ray crystal structure revealed that the central oxadiazine six-membered ring adopts a chair-conformation where the H2 and H3 atoms occupy an axial position, whereas the H1 atom occupies an equatorial position (Fig. 1). Both indoline and spiro-cyclohexyl moieties of the NON-materials point in opposite directions with respect to the central oxazine core, thus forming voids above and below the central ring, which are capable of accommodating solvent molecules (for instance, hexane, see Fig. S8). Molecules of the NON-materials form a three-dimensional network through weak C–H∙∙∙π, C–H∙∙∙F, C–H∙∙∙Br or C–Hδ+(phenyl)···δ+H–C(cyclohexyl) interactions between neighbouring molecules.

Cyclic voltammetry experiments were performed to investigate the redox properties of NON, NON-Br, and NON-CF3 in 1,2-difluorobenzene solution (Fig. S11). The data is summarised in Table 2. The NON-compound exhibits non-reversible oxidation at a peak potential of Ep = +0.5 V. The NON-CF3 material shows two oxidation processes: the first one is reversible (E1/2= +0.90 V), while the second is irreversible with Ep at +1.38 V. The NON-Br compound shows two reversible oxidations at E1/2 = +0.62 and +1.12 V. The onset of the first oxidation potential (Eonset) values was used to calculate the energy level of the highest occupied molecular orbital (EHOMO, see Table 1) for the NON-R materials. The EHOMO energy correlates with the electron-withdrawing strength of the substituent R, i.e., it increases in line from –5.76 (R = H) to –5.90 (R = Br) and –6.14 eV (R = CF3). The oxidation process of the NON-compounds is likely centred on the indoline moiety, which is supported by the theoretical calculations which predict the localization of the HOMO to this region (see Fig. 3c, Table S4). The reduction potential for all three compounds was beyond the solvent discharge potential, therefore, a direct calculation of the LUMO energy level is impossible. Instead, the ELUMO energy was estimated indirectly using the equation ELUMO = Eopt-gap + EHOMO, where Eopt-gap is taken as the red-onset of the lowest energy absorption band in the UV-vis absorption spectra in a dichloromethane solution (vide infra, Table 1).

Table 1 Cyclic voltammetry and additional photophysical parameters
Full size table

Photophysical properties

To investigate photophysical properties, UV-visible absorption (Fig. 2) and photoluminescence (Fig. 3) spectra were recorded for all NON-compounds. The data is summarized in Tables 1 and 2. All compounds show a strong absorption band in the range 260–280 nm (extinction coefficient ε up to 2.8 × 104 M−1 cm−1, Table S3) and a low-energy absorption band at ca. 304 nm with ε coefficient up to 4 × 103 M−1 cm−1 (Table S3). Theoretical calculations revealed that oscillator strength coefficients (f) for S0→S1 (HOMO→LUMO) are nearly zero for all NON-materials. Further analysis of the TD-DFT results (Fig. 2), for instance for the NON-compound, revealed the intense vertical excitation: S0→S2 (f = 0.08; HOMO→LUMO + 1) and S0→S3 (f = 0.17; HOMO→LUMO + 2) which are all ascribed to π–π* transitions within the indolines of the NON-materials (Tables S4, S6 and S7). The optical gap (Eopt-gap) values for all NON-materials were calculated from the red-onset of the lowest energy absorption band: 3.72 eV for NON, 3.86 eV for NON-Br and 3.90 eV for NON-CF3. All absorption bands demonstrate very little solvatochromic effect, as shown in Fig. 2b for the compound NON-CF3 (see Supporting Files, Figs. S13–S15 for all compounds). This indicates only a minor change in the dipole moment upon vertical excitation from the ground S0 to excited S1 states and is reflected in the TD-DFT theoretical calculations (Table S5).

Fig. 2: UV-vis absorption and frontier molecular orbitals distribution for NON-materials.
figure 2

UV-vis absorption for NON-materials in dichloromethane solution (a) and solvent of different polarity for NON-CF3 (b); molecular orbitals distribution involved in vertical excitations for NON-materials (c). There is ca 40 nm offset between the gas-phase calculated energy of the transition and observed UV-vis absorption in solution.

Full size image
Fig. 3: Photoluminescence spectra for NON-materials.
figure 3

Photoluminescence (PL) spectra for compounds NON (a), NON-CF3 (b) and NON-Br (c) in MeTHF solution (at 298 K and frozen glass at 77 K); PL for NON-compounds in 1% by weight Zeonex matrix at 298 K (d).

Full size image
Table 2 Photophysical properties of the NON-materials in MeTHF solution and 1% Zeonex matrix at 298 and 77 K
Full size table

All NON-compounds exhibit featureless and narrow near-UV-fluorescence with a full width at half maximum (FWHM) of 3400 cm–1 (40 nm) in MeTHF solution or Zeonex films at 295 K (Figs. 3 red line, S16–S21). The electron-withdrawing substituents on the indoline moieties lead to a blue shift in emission from 338 nm for NON to 310 and 315 nm for NON-CF3 and NON-Br, respectively. This observation is corroborated by theoretical calculations that show the HOMO is localised over the indoline moieties. Therefore, increasing the electron-withdrawing strength of the substituents attached to the indoline moiety results in the stabilization of the HOMO energy level and an increase in the emission energy. The photoluminescence quantum yields (PLQY) for UV-luminescence increase from NON-Br (3%) to NON-CF3 (12%) to NON (23%) in the MeTHF solution at 295 K.

All NON-materials, in all media, exhibit an excited state lifetime on the nanosecond time scale in the range of 0.9–1.8 ns, which is typical for the fluorescence emission mechanism. Figure 3 depicts the emission profiles of the NON-compounds in frozen MeTHF glasses at 77 K (black line, Table 2). The emission spectra of NON at 77 K is dominated by a deep-blue and vibronically resolved phosphorescence, peaking at 400 nm. The excited state lifetime is too long to be measured by a gated Xenon flashlamp. However, the emission was observed with the naked eye to persist for more than 3 s after the excitation ceased. Unlike NON, the emission spectra of NON-CF3 and NON-Br compounds are dominated by UV-fluorescence (306–337 nm), with a minor contribution from the deep-blue phosphorescence component at 77 K (Fig. 3b, c). Similar to NON, the excited state lifetime of the phosphorescence is longer than 3 s, while the fluorescence excited state lifetime only slightly increases up to 5.9 ns upon cooling.

The energy levels of the singlet (S1) and triplet (T1) excited states for all NON-materials were deduced from the blue-onsets of the fluorescent and phosphorescent profiles (Figs. 3, S23, Table 1) at 77 K. The NON and NON-CF3 materials exhibit a high triplet energy level at 3.3 eV, which is higher than the most popular host material DPEPO (3.0 eV) and on par with the most stable deep-blue OLED host material – PPBi (3.3 eV). The computed values align well with the experimental results (Table S6). The theoretical calculations show that both frontier orbitals, HOMO and LUMO, are primarily localized over the indoline rings of the NON-compounds (Table S4), resulting in a large frontier orbital overlap integral (0.62–0.66) for all NON-materials (Table S4). This is largely due to the absence of spatially separated donor and acceptor moieties, which results in a significant stabilization of the T1 energy level and an energy difference with the singlet excited state, ΔEST, up to 1.15 eV. Note that the central oxadiazine core with spiro-sp3 carbon atoms acts as an insulator, preventing extended conjugation between two indoline moieties for the NON-materials, which prevents further lowering of the T1 energy level. It has previously been demonstrated that small molecule hosts29,30 show triplet energy variance depending on the degree of conjugation and magnitude of the dihedral angle between the cycles involved, for instance, carbazole moieties in CBP (T1 = 2.64 eV), mCBP (T1 = 2.84 eV), oCBP (T1 = 3.0 eV, Scheme 1)29 and others29,30. Therefore, the novelty of NON-host molecular design strongly benefits from the bulky cyclohexyl moieties, absence of any inter-cycle bonds and spiro-sp3 carbon atom breaking the conjugation thus explaining the high triplet energy level and making NON-hosts more robust towards the local environment.

Device fabrication, characterisation, and performance

We assessed all NON-materials against the criteria outlined in the introduction to identify a promising host material for the development of deep-blue OLED devices. The NON-CF3 material exhibits good thermal stability, a wide energy gap (3.9 eV) thanks to a well-stabilised HOMO (–6.1 eV) and a destabilised LUMO (–2.2 eV), a high triplet energy level of 3.3 eV, and excellent volatility during the sublimation process due to three cyclohexyl groups. These attributes motivated us to test the NON-CF3 material as a host with our recently developed Carbene-Metal-Amide (CMA) TADF material P170 (Fig. 4a)31. First two triplet states of complex P170 are T1 (3CT 2.77 eV) and T2 (3LE 3.29 eV)31 are lower in energy compared to the first triplet state of the NON-CF3 host (T1 is 3.3 eV) thus preventing the Dexter energy transfer back to the host material. OLED devices were fabricated via thermal vapour deposition. The device architecture is shown in Fig. 4, along with the chemical structures of the materials used. The 30 nm layer of 1,1,-bis{4-[N,N-di(4-toyl)amino]phenyl}cyclohexane (TAPC) and a 10 nm layer of 9,9′-biphenyl-2,2′-diylbis-9H-carbazole (oCBP) function as hole transport layers. In each device, the 20 nm thick emissive layer (EML) consists of an emitting material doped at 10% by weight into the NON-CF3 host. A 45 nm layer of diphenyl-4-triphenylsilyl-phenylphosphine oxide (TSPO1) is used for electron transport. Four EML compositions were prepared using four different host materials, namely PPBi, TCP, mCP and NON-CF3 (Fig. 4a), to evaluate and compare the performance of NON-CF3.

Fig. 4: Vapour-deposited OLED device architectures.
figure 4

a OLED architecture for P170 doped into different hosts at 10% by weight and chemical formulae of materials used in the devices; b Normalised electroluminescence spectra from devices incorporating P170 10% by weight in host-guest structures; c current density-voltage plot; d luminance-voltage plot; e EQE versus luminance plot.

Full size image

The electroluminescence spectra for devices at 6 V are shown in Fig. 4, and a summary of the performance data for each device is given in Table 3. The use of NON-CF3 as a host material results in a blue shift of up to 31 nm for the electroluminescence from P170 compared to the other host materials tested. This leads to a highly desirable 448 nm EL from P170 with a narrow FWHM of ca. 60 nm in a NON-CF3 host. This result is favourable when comparing NON-CF3 to the DPEPO host, where the deep-blue OLED (EML: P170 10% in DPEPO) demonstrated a 30 nm wider FWHM value (ca. 90 nm)31. Therefore, a commonly cited disadvantage of organometallic TADF materials, such as their broad EL emission profiles associated with the emission occurring from a charge transfer (CT) state, could be circumvented by using the NON-CF3 host material. This enables narrow EL profiles and better CIE-colour coordinates. The phenomena of the PL31,32,33,34 and EL31,35 blue shift of the CT emission for CMA materials were studied in detail and explained by thermally activated diffusion and electrostatic interactions between CMA emitter with host molecules.

Table 3 Performance data of evaporated OLEDs
Full size table

Devices utilizing NON-CF3 exhibit an increased turn-on voltage, which is reported as the applied bias for luminescence to equal 1 cd m–2. This suggests a higher barrier for charge injection into the emitting layer. Figure 4 shows the current density–voltage–luminescence (J–V–L) curves for each OLED device. Devices experience luminescence roll-off when operating above 11 V, indicating a need for further improvement in the molecular design of the NON-host materials. However, the OLED devices hosted by NON-CF3 achieved an External Quantum Efficiency EQEmax value of 21.0%, the highest value among all tested host materials. Impressively, OLED devices containing P170 doped into NON-CF3 were operational at 100 cd m−2 with an EQE of 9.2%, which is comparable to commercially available host materials such as TCP and mCP tested in this work. The EQE performance is likely associated with higher PLQY values for P170 films in the NON-CF3 host (82%) compared to TCP films (38%). These findings suggest a high potential for NON-host materials to be used as host materials for deep-blue OLEDs.

Conclusion

We discovered an unexpected product of the interrupted Fischer indolization reaction which has enabled a straightforward synthesis of optionally substituted 1,3,5-oxadiazines (NON) in good yields. Our protocol involves a single-step condensation reaction between cost-effective substituted hydrazines and carbaldehydes to prepare NON-compounds with good thermal stability up to 262.5 °C. We demonstrated that all NON-materials emit only fast UV-fluorescence (310–338 nm) at room temperature, while a long-lived (over 3 s) deep-blue phosphorescence emerges upon cooling to 77 K. Our theoretical calculations revealed large frontier orbitals overlap integral values up to 0.66 for the NON-materials which is rationalised by the significant energy difference between singlet (S1) and triplet (T1) excited states up to 1.15 eV.

Our photophysical and electrochemical investigation into the new 1,3,5-oxadiazine materials revealed their suitability, from an energy level standpoint, for use as host materials in OLED devices. NON-CF3 was selected for use in proof-of-concept devices based on its good thermal stability, wide energy gap (3.9 eV), well-stabilised HOMO (–6.1 eV) and destabilised LUMO (–2.2 eV), high triplet energy level (3.3 eV), and excellent volatility, thanks to the bulky cyclohexyl groups that prevent strong intermolecular contacts. Given the novel molecular design of NONs and their remarkably simple synthesis, we are impressed that OLED devices using a CMA TADF emitter and NON-CF3 host achieved EQE values up to 21%. Furthermore, we demonstrated how the use of a NON-host material can facilitate excellent deep-blue colour emission by producing a blue-shifted and narrow (60 nm) electroluminescence profile in comparison to conventional host materials.

This work introduces a new class of OLED host materials. Further modifications to the host material structure can be made by decorating with electron donating or withdrawing groups to enhance hole and electron transport mobility and pave the way to high-performance and stable deep-blue OLEDs.

Methods

Material preparation

The Au1 complex P170 was obtained according to the literature. All NON-materials were purified on a silica gel column, recrystallised from hexane, while NON-CF3 was sublimed to have high purity for OLED fabrication (See Supplementary Methods).

Photophysical properties

Solution UV-visible absorption spectra were recorded on a Cary 500 UV-vis-NIR spectrometer for a wavelength range 250–700 nm. Photoluminescence measurements were recorded using an Edinburgh Instruments FLS920 spectrometer. All excited state lifetimes were measured on FLS920 spectrometer with mono- and biexponential fitting provided by Edinburgh Instruments Fluoracle software v2.6.1. Absolute photoluminescent quantum yields were measured directly with a Quantaurus-QY Absolute PL quantum yield spectrometer.

Computational details

The ground and excited states of all complexes were studied by DFT and TD-DFT36 respectively, using Tamm-Dancoff approximation37. All calculations were carried out by Gaussian 1638, employing the global hybrid MN15 functional39 and the def2-TZVP basis set40. The overlap integrals between HOMO and LUMO were calculated using the Multiwfn program41.

OLED fabrication and characterisation

OLED devices were fabricated by high-vacuum (10−7 Torr) thermal evaporation on ITO-coated glass substrates with sheet resistance of 15 Ω/□. Substrates were cleaned by sonication in a non-ionic detergent, deionised water, acetone, and isopropyl alcohol and subject to an oxygen plasma treatment for 10 min. Layers were deposited at rates of 0.1–2 Ås−1. TCP, PPBi and mCP were purchased from Luminescence Technology Corp. TPBi, DPEPO and TSPO1 were purchased from Shine Materials. All purchased materials were used as received. OLED current density-voltage measurements were made using a Keithley 2400 source-meter unit. The luminance was measured on-axis using a 1 cm2 calibrated silicon photodiode at a distance of 15 cm from the front face of the OLED. Electroluminescence spectra were measured using a calibrated OceanOptics Flame spectrometer. Lifetime measurements were measured with a Keithley 2400 source-meter unit and a 0.75 cm2 silicon photodiode. The devices were held under rough vacuum (~10−3 Torr).