B ← N Lewis Pair Fusion of N,N‐Diaryldihydrophenazines: Effect on Structural, Electronic, and Emissive Properties

Abstract

Doping of polycyclic aromatic hydrocarbons (PAHs) with boron and/or nitrogen is emerging as a powerful tool to tailor the electronic structure and photophysical properties. As N‐doped analogues of anthracene, N,N‐dihydrophenazines play important roles as redox mediators, battery materials, luminophores, and photoredox catalysts. Although benzannulation has been used successfully as a structural constraint to control the excited state properties, fusion of the N‐aryl groups to the phenazine backbone has rarely been explored. Herein, we report the first examples of dihydrophenazines, in which the N‐aryl groups are fused to the phenazine backbone via B←N Lewis pair formation. This results in structural rigidification, locking the molecules in a bent conformation, while also modulating the electronic structure through molecular polarization. B─N fusion in BNPz1−BNPz3 induces a quinoid resonance structure with significant C─N(py) double bond character and reduces the antiaromatic character of the central pyrazine ring. Borylation also lowers the HOMO/LUMO (highest occupied/lowest unoccupied molecular orbital) energies and engenders bathochromic shifts in the emission. Further rigidification in the solid state gives rise to enhanced emission quantum yields, consistent with aggregation‐induced emission enhancement (AIEE) observed upon water addition to solutions in tetrahydrofuran (THF). The demonstrated structural control and fine‐tuning of optoelectronic properties are of great significance to potential applications as emissive materials and in photocatalysis.

B─N Lewis pair fusion of N,N‐dipyridyldihydrophenazine and its benzofused homologs results in steric constraints that enforce a bent structure in both the ground and excited states, reduces the antiaromatic character of the central dihydropyrazine heterocycle and promotes a quinoidal resonance structure, bathochromically shifts the absorption and emission bands, and induces aggregation induced emission enhancement.

Introduction

Polycyclic aromatic hydrocarbons (PAHs) show desirable photophysical and electronic properties, acting as chromophores and functional materials in organic electronics, (bio)imaging, and as chemical sensors and catalysts.^{[ 1} , ^{2 ]} The optical and electronic characteristics can be dramatically altered by doping with heteroatoms.^{[ 1} , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ^{12 ]} For instance, 9,10‐dihydro‐9,10‐diboraanthracene may be viewed as an electron‐deficient analog of anthracene as the boron atoms feature empty p‐orbitals that overlap with the adjacent benzene π‐systems, resulting in local antiaromatic character of the central C₄B₂ heterocycle (formally a 4π‐electron system).^{[ 13} , ^{14 ]} Typically referred to simply as “diboraanthracenes,” aryl‐substituted derivatives such as A (Figure 1) have been studied extensively as luminescent materials, building blocks for supramolecular constructs, Z‐type ligand frameworks, and Lewis acid (co)catalysts.^{[ 15} , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ^{21 ]} Recent efforts have culminated in further elongated acenes that are doped with multiple boron centers.^{[ 22} , ²³ , ²⁴ , ²⁵ , ^{26 ]}

Figure 1.

a) Examples of B‐ and N‐doped analogs of anthracene (A–D), their core geometry in the ground state (GS) and excited state (ES), and emission characteristics (AIE: aggregation‐induced emission, VIE: vibration‐induced emission, CPL: circularly polarized emission). b) B ← N Lewis pair fused anthracenes (E) and N,N‐diaryldihydrophenazines (F) reported in here.

Contrary to the electron‐deficient character of diboraanthracenes, incorporation of nitrogen in the 9,10‐positions results in electron‐rich 9,10‐dihydrophenazines (B, Figure 1).^{[ 27 ]} The nitrogen atoms feature filled p‐orbitals (lone pairs) that overlap with the adjacent benzene π‐systems, again bringing about local antiaromatic properties of the central C₄N₂ heterocycle (formally an 8π‐electron system).^{[ 28} , ²⁹ , ^{30 ]} N,N‐diaryldihydrophenazines have been widely utilized as redox mediators, in battery applications, and as emissive materials.^{[ 28} , ³¹ , ^{32 ]} Recently, they have been reported to exhibit aggregation‐induced emission (AIE) properties.^{[ 33 ]} They have also attracted great interest as photoredox catalysts in organic synthesis and metal‐free polymerization catalysis.^{[ 34} , ^{35 ]} N,N‐diphenyldihydrophenazine adopts a planar geometry, indicative of extensive π‐delocalization, and similar planar structures have been reported for the extended naphtho‐fused dihydrodiazatetracene and dihydrodiazapentacene derivatives.^{[ 29} , ^{36 ]} Sterically constrained dihydrophenazines in which planarization is unfavorable have attracted much recent interest for their unusual excited state properties.^{[ 37} , ³⁸ , ³⁹ , ^{40 ]} For instance, dihydrodibenzo[a,c]‐phenazines (C, Figure 1) adopt a saddle‐shaped bent structure in the ground state (GS) with a bending angle between two planes of opposite aryl rings of Θ _b = 137°, caused by steric interference between the fused benzene rings and the N‐phenyl substituents.^{[ 37 ]} However, in the excited state (ES) the molecular structure relaxes toward a planarized conformation, from which emission occurs with large Stokes shifts. When more sterically demanding substituents are placed on nitrogen, normal solvent polarity‐dependent emission with moderate Stokes shifts is also seen because planarization becomes less favorable. Characteristic differences between the higher energy emission from locally excited states and lower energy emission from vibrationally relaxed excited states (“vibrationally induced emission,” VIE) have been exploited for the development of white light emitters, phosphorescent materials, molecular thermometers, (micro)viscosity probes, and organic vapor sensors.^{[ 41} , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ^{47 ]} Efforts have also been undertaken to connect the pendent N‐aryl groups through a linker, thereby controlling the conformational flexibility of the molecules and enabling catenation.^{[ 48} , ^{49 ]} On the other hand, fusion of the N‐aryl groups to the phenazine backbone has sparingly been explored, possibly because of a lack of convenient synthetic methods. A rare example is the introduction of bridging sulfur or sulfone moieties (e.g., D, Figure 1), which is accomplished by fusion of two preformed phenothiazine molecules rather than post‐modification.^{[ 50} , ⁵¹ , ⁵² , ^{53 ]} Rigidification in D has been shown to result in interesting emissive properties, including circularly polarized thermally activated delayed fluorescence (CP‐TADF) in the case of twisted molecules with X = CN, and to enable use as receptors for fullerenes.^{[ 53 ]}

B ← N Lewis pair fusion at the periphery is emerging as a very convenient approach for the structural rigidification and electronic modulation of PAHs. N‐directed electrophilic borylation is synthetically straightforward and typically leads to extension of π‐conjugation into the N‐heterocyclic substituents, lowers the LUMO level, and improves the rigidity, thus promoting radiative over nonradiative decay.^{[ 54} , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ^{62 ]} B─N functionalized π‐conjugated materials have been used as emitters in organic light‐emitting devices (OLEDs), electron transporting materials in transistors, acceptor materials for organic solar cells (OSCs), and in bioimaging applications.^{[ 63} , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ^{69 ]} They also show unique photochemical reactivity patterns,^{[ 70} , ^{71 ]} stimuli‐responsive properties,^{[ 72} , ^{73 ]} and circularly polarized luminescence (CPL) behavior.^{[ 74} , ^{75 ]} Furthermore, empowered by the reversibility of the B ← N Lewis pair formation, they can act as molecular switches.^{[ 73} , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ^{80 ]}

We have a long‐standing interest in exploring the effect of B ← N Lewis pair functionalization of PAHs on the structural and optoelectronic properties.^{[ 81} , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ , ⁸⁷ , ⁸⁸ , ⁸⁹ , ^{90 ]} Previously, we demonstrated that B ← N Lewis pair fusion of 9,10‐dipyridylanthracene (E, Figure 1) induces strong geometric distortions with bend angles between the peripheral benzene rings ranging from Θ_b = 156° to 163°.^{[ 81} , ⁸² , ⁸³ , ^{90 ]} B ← N Lewis pair fusion also leads to polarization of the molecules, resulting in largely red‐shifted absorption and emission from cross‐conjugated excited states with significant intramolecular charge transfer (ICT) character and unusual self‐sensitized reactivity^{[ 91} , ⁹² , ⁹³ , ^{94 ]} toward singlet oxygen. As such, without any external sensitizer reversible formation of the respective endoperoxides is achieved, from which singlet oxygen is released at elevated temperature.^{[ 81} , ⁸² , ⁸³ , ^{90 ]}

We surmised that B ← N Lewis pair formation could also serve as a powerful approach for structural rigidification and electronic tuning of dihydrophenazines, locking the molecules in a semi‐rigid non‐planar conformation while also modulating the electronic structure through molecular polarization. We introduce here a versatile synthetic approach to the first examples of dihydrophenazines in which the N‐aryl groups are fused through B ← N Lewis pair formation to the phenazine backbone. N‐doped analogues of anthracene (BNPz1), tetracene (BNPz2), and pentacene (BNPz3) are accessed, and their solid‐state structures are examined by single crystal X‐ray analysis (Figure 1). The structural studies show that B ← N Lewis pair formation promotes a polarized resonance structure, in which the pyridyl groups are dearomatized and linked to anionic borate and cationic iminium groups. Computational studies on both the ligand precursors and the borylated products offer detailed insights into the effect of B ← N Lewis pair fusion on the electronic structure and local antiaromaticity of the central ring. We find that the more flexible ligand precursors are moderately emissive in the solid state but more strongly emissive in a poly(methyl methacrylate) (PMMA) matrix, presumably due to competing effects of aggregation‐caused quenching (ACQ) and restriction of molecular motions. The opposite is true for the B─N fused systems, which are more highly emissive in the bulk than in PMMA, indicative of aggregation‐induced emission enhancement (AIEE) effects.

Results and Discussion

Synthesis and Structural Features

The synthetic routes to the ligands and borylated complexes are illustrated in Scheme 1. Following a literature procedure,^{[ 35 ]} the precursor 5,10‐dihydrophenazine (1) was synthesized by reduction of phenazine with sodium dithionite in refluxing ethanol. Melt reaction of naphthalene‐2,3‐diol with benzene‐1,2‐diamine at 180 °C gave 5,12‐dihydrobenzo[b]phenazine (2) in 80% yield;^{[ 95 ]} similarly, melt reaction of naphthalene‐2,3‐diol with naphthalene‐2,3‐diamine at 180 °C gave 6,13‐dihydrodibenzo[b,i]phenazine (3) in 78% yield.^{[ 36 ]} The desired ligands were obtained as yellow solids in good to excellent yields (Pz1,^{[ 33 ]} 77%; Pz2, 90%; Pz3, 97%) by Buchwald−Hartwig coupling of the respective dihydrodiazaacenes with 2‐bromopyridine in toluene at 130 °C using Pd₂(dba)₃ (4 mol%) / ^tBu₃P (3 mol%) as the catalyst system and NaO ^t Bu as the base.

Scheme 1.

Synthesis of ligands Pz1−Pz3 and B ← N Lewis pair‐functionalized BNPz1−BNPz3 (R = Et) and plots of the X‐ray crystal structures of (top) BNPz1 (CCDC 387675), (middle) BNPz2 (CCDC 387676), and (bottom) BNPz3 (CCDC 387677); thermal ellipsoids at 50% probability, C grey, N blue, B green, and H atoms omitted for clarity. For BNPz2, only one of two independent molecules in the unit cell is shown, and a cocrystallized EtOAc molecule is omitted for clarity; for BNPz3, disordered solvent molecules (EtOAc/hexanes) were removed using the Platon squeeze routine.

In general, B ← N Lewis pair formation can be achieved by lithiation−borylation reaction sequences, hydroboration of vinyl functionalized derivatives,^{[ 96 ]} main group and transition metal‐catalyzed C─H borylation,^{[ 97} , ⁹⁸ , ^{99 ]} or Lewis base‐directed electrophilic^{[ 56} , ^{100 ]} borylation.^{[ 101 ]} We pursued the B─N heterocycle formation by N‐directed electrophilic borylation with BBr₃ in the presence of N ⁱ Pr₂Et (DIPEA) as a sterically hindered base. Murakami and coworkers^{[ 102 ]} previously reported that the borylation of 2‐(pyrid‐2‐yl)naphthalene with BBr₃/DIPEA produces a mixture of regioisomeric 5‐membered B─N heterocycles, and Wendt and coworkers^{[ 103 ]} showed that borylation of 1‐(pyrid‐2‐yl)naphthalene under similar conditions preferentially occurs at the more nucleophilic peri‐position, resulting in a 6‐membered B─N heterocycle. When using BCl₃/AlCl₃ for the borylation of 1,4‐dipyridylnaphthalene and 3,10‐dipyridylperylene, Ji and coworkers obtained isomer mixtures of the diborylated products from which B ← N Lewis pair‐fused products with two 6‐membered rings were isolated in low yields.^{[ 59} , ^{104 ]} In earlier work, we found that the double borylation of 9,10‐di(pyrid‐2‐yl)anthracene with a mixture of BCl₃ and AlCl₃ (4 equivs each) in the presence of 2,6‐di‐t‐butylpyridine as the base selectively generates the trans‐diborylated isomer with boryl groups attached to different benzene rings (trans‐isomer of E, Figure 1).^{[ 81} , ^{82 ]} In recent work, we show that the selective generation of the trans‐isomer of E under these conditions is related to the formation of borenium ion intermediates in the presence of excess AlCl₃.^{[ 90 ]} Further, we discovered that with only two equivalents of AlCl₃ a mixture of cis‐ and trans‐isomers is obtained, and the relative amount of the cis‐isomer of E (Figure 1) can be optimized by modification of the reaction conditions. Efforts to diborylate Pz1−Pz3 with BCl₃ / AlCl₃ in the presence of 2,6‐di‐t‐butylpyridine were not successful. However, the borylation of ligands Pz1−Pz3 with 2 equivs BBr₃ in the presence of DIPEA, followed by treatment with Et₂Zn, selectively generated the cis‐diborylated products BNPz1−BNPz3 as confirmed by ¹H, ¹³C, ¹¹B NMR, ESI‐MS, and XRD analysis. Only small amounts of a compound that is tentatively attributed to the trans‐product (ca. 1:5 ratio of trans:cis) could be identified in the crude reaction mixture of BNPz1, and efforts are currently under way to generate the trans‐isomer more selectively and to fully examine its properties. The cis‐disubstituted B ← N Lewis pair‐functionalized dihydrodiazaacenes BNPz1 (59%), BNPz2 (36%), and BNPz3 (61%) were isolated in moderate to good yields. Signals at −0.2 (BNPz1), 1.7 (BNPz2), and 0.9 ppm (BNPz3) in the ¹¹B NMR spectra confirmed the presence of tetracoordinate boron atoms. Furthermore, the ¹H NMR spectral patterns with a characteristic singlet for the acene backbone in BNPz1 at 7.68 ppm, four doublets of doublets for BNPz2, and a more complex signal pattern indicative of an unsymmetric structure for BNPz3 strongly suggested the formation of cis‐ as opposed to trans‐isomers, with preferential borylation of the larger naphthalene rather than the smaller benzene π‐system in the case of BNPz2.

The formation of two fused 6‐membered B─N heterocycles by regioselective double electrophilic C─H borylation of the same benzene ring was further confirmed by X‐ray structure analyses.^{[ 105 ]} Single crystals of BNPz1–BNPz3 suitable for X‐ray diffraction analysis were grown by slow evaporation of solutions in a mixture of ethyl acetate and hexanes at room temperature. In all boron complexes, the dihydrodiazaacene core is strongly bent with interplanar angles between the benzene rings Ph_B and Ph_H ranging from Θ_b = 131.2° to 143.2° (Scheme 1, Table 1). The deviation from planarity increases with benzannulation and is most pronounced for BNPz2. The observed bending stands in contrast to the planar dihydrophenazine backbone of ligand precursor Pz1 (Θ_b = 180°) and the almost planar geometry of phenyl‐substituted dihydrophenazine (B, Θ_b = 176.2°).^{[ 33} , ³⁵ , ^{106 ]} As noted earlier, bent structures are seen for benzannulated dihydrophenazines C for which a V‐shaped conformation is favored in the GS due to steric constraints (Θ_b = 137°).^{[ 37} , ^{39 ]} Bent and twisted structures have also been reported for ortho‐methylated derivatives.^{[ 107 ]} The structures of BNPz1–BNPz3 are reminiscent of those of B ← N Lewis pair‐functionalized anthracenes (E, Θ_b ca. 156–163°)^{[ 81} , ⁸² , ^{90 ]} but the bending of the dihydrophenazines is much more pronounced. For BNPz1, the boron atoms lie essentially in the same plane as the benzene ring they are bound to, whereas for BNPz2 and BNPz3, both boron atoms are positioned slightly above the arene plane. The pyridyl groups approach the boron atoms from the same side, above the dihydropyrazine units. Hence, the molecules show an approximate mirror symmetry, although the B1‐C1 distances are slightly different from the B2‐C2 distances, and the same is true for the B1‐N1 and B2‐N2 distances. The B─N bond distances of 1.643(2)–1.671(2) Å are on the longer side of typical tetracoordinate boron complexes and slightly longer than those of the anthracene derivatives E (R = Et), which show B─N bond distances of 1.632(2) Å for the trans‐isomer and 1.632(3)/1.635(3) Å for the cis‐isomer.^{[ 90 ]} A possible explanation is that the electron‐donating effect of the amino groups in the dihydrophenazine core reduces the Lewis acid strength of the borane moieties. Importantly, the C─N bonds to the pyridyl groups (N─C(Py) = 1.375(2)–1.384(2) Å) are significantly shorter than those within the dihydrophenazine framework (N─C(Ph) = 1.417(2)–1.440(2) Å). This is opposite to the metric parameters for the non‐borylated N,N‐dipyridyldihydrophenazine Pz1, for which the C─N bonds to the pyridyl groups (N─C(Py) = 1.434(2) Å) were reported to be longer than those within the phenazine units (N‐C(Ph) = 1.409(2), 1.410(2) Å).^{[ 33 ]} We conclude that electronic delocalization from the dihydrophenazine nitrogen lone pairs to the pendent pyridyl groups becomes more favorable upon B ← N Lewis pair formation (which withdraws electron density from the pyridyl groups). Conversely, the nitrogen lone pairs delocalize less into the dihydrophenazine benzene rings upon B ← N Lewis pair formation. Within the dihydrophenazine moiety, the C─N bonds to the borylated benzene rings (N─C(Ph_B) = 1.430(2)–1.440(2) Å) are consistently longer than those to the non‐borylated benzene rings (N─C(Ph_H) = 1.417(2)–1.423(2) Å). This suggests that delocalization into the boron‐substituted benzene rings is relatively less favorable because of their more electron‐rich character endowed by the σ‐donating effect of the tetracoordinate boron moieties. Collectively, the structural features indicate that the planarization and molecular polarization upon B ← N Lewis pair formation promote N─C π‐bonding between lone pairs on the dihydrophenazine nitrogens and the electron‐deficient pyridyl groups and reduce N─C π‐bonding within the dihydrophenazine framework.

Table 1.

Selected bond distances (Å) and angles (°) for BNPz1, BNPz2 (two independent molecules, A and B), and BNPz3.

Compound	BNPz1	BNPz2 (A)	BNPz2 (B)	BNPz3
B1‐C1	1.611(2)	1.626(2)	1.634(2)	1.631(2)
B2‐C2	1.609(2)	1.627(2)	1.624(2)	1.630(2)
B1‐N1	1.647(2)	1.654(2)	1.671(2)	1.643(2)
B2‐N2	1.658(2)	1.667(3)	1.653(2)	1.656(2)
N3‐C3 (PhB)	1.435(2)	1.439(2)	1.433(2)	1.432(2)
N4‐C4 (PhB)	1.430(2)	1.434(2)	1.438(2)	1.440(2)
N3‐C5 (PhH)	1.423(2)	1.417(2)	1.417(2)	1.423(2)
N4‐C6 (PhH)	1.418(2)	1.423(2)	1.422(2)	1.420(2)
N3‐C11 (Py)	1.378(2)	1.381(2)	1.375(2)	1.381(2)
N4‐C21 (Py)	1.378(2)	1.377(2)	1.375(2)	1.384(2)
N1‐B1‐C1	103.92(9)	105.35(10)	104.73(10)	105.67(10)
N2‐B2‐C4	104.59(9)	105.15(14)	106.19(11)	105.84(12)
Θb a)	143.2	131.2	132.9	136.9

^a) Interplanar angle Θ_b between benzene rings Ph_B and Ph_H attached to the central dihydropyrazine heterocycle (Pz).

Electronic Structure Calculations

DFT calculations were performed at the B3LYP/6–311G** level of theory to assess in more detail the electronic structures of the free ligands Pz1–Pz3 and the changes encountered when forming the B ← N Lewis pair complexes BNPz1'–BNPz3' (boron‐bound ethyl were replaced with methyl groups, Figures S27 and S28, Tables S3 and S4). Even further extended dihydrodiazaheptacene derivatives Pz4 and BNPz4' were also included in the computational studies. The computed B─N distances of 1.680–1.694 Å are slightly longer, and the interplanar angles between the benzene rings of Θ_b = 139.1–147.7° slightly larger than those derived from the X‐ray structure analyses (1.643(2)–1.671(2) Å; 131.2–143.2°). Thus, in the computed structures, relatively longer B─N bonds result in a slight decrease of the steric strain. Nevertheless, the structural parameters of the geometry‐optimized complexes are overall in good agreement with those derived from the single crystal X‐ray analyses (Table S4). Furthermore, the trends seen in the X‐ray data upon extension of conjugation from BNPz1' to BNPz3' are well reproduced. For instance, BNPz1' (Θ_b = 147.7°) adopts a more planar conformation than BNPz2' (139.1°) and BNPz3' (139.7°), consistent with what is seen in the X‐ray crystal structure data. A more detailed comparison of the metric parameters of BNPz1'–BNPz4' shows that the B─N and B─C bond lengths and the angles Θ_b decrease (structure becomes more bent) upon going from derivatives with benzene (BNPz1') to naphthalene (BNPz2') and anthracene (BNPz3') units. The computational data also suggest that the larger angle Θ_b derived from single crystal structures for BNPz3 relative to BNPz2 is an outlier and a result of crystal packing effects (not reproduced by computations). The computed N─C(Py) bond distances for the B─N Lewis pair complexes (BNPz1' 1.379/1.378 Å, BNPz2' 1.374/1.372 Å, BNPz3' 1.376/1.374 Å, BNPz4 1.377/1.376 Å) are much shorter than those for the ligand precursors (Pz1 1.430/1.430 Å, Pz2 1.432/1.433 Å, Pz3 1.434/1.434 Å, Pz4 1.436/1.436 Å), independent of the number of additional benzene rings, which is also in excellent agreement with experimental data (Figures S27 and S28).

The frontier orbitals and corresponding energy levels for BNPz1'–BNPz4' are illustrated in Figure 2, and those of the ligand precursors Pz1–Pz4 are provided in Figure S29. The HOMOs of Pz1–Pz4 are localized on the dihydrodiazaacene core with significant contributions from the nitrogen lone pairs, and the HOMO energy gradually decreases from −4.34 eV for Pz1 to −4.84 eV for Pz4 as the acene π‐system is enlarged. A set of two degenerate LUMOs is seen for Pz1–Pz3 consisting of linear combinations of orbitals that are localized on the pyridyl rings with minor contributions of the central diazine core in the case of Pz3 (Figure S32). Again, the energy of these orbitals slightly decreases from Pz1 (−1.21/−1.20 eV) to Pz3 (−1.35/−1.32 eV). The dihydrodiazaacene‐centered π* orbitals (LUMO+4) are relatively higher in energy at +0.11 to −0.97 eV (Figure S32). In contrast, for Pz4 with its further extended dihydrodiazaheptacene framework, the π* orbital becomes the LUMO at −1.77 eV, lower in energy than the corresponding pyridyl‐centered orbitals. Formation of the B ← N Lewis pair complexes greatly influences the nature of the frontier orbitals. Although little to no direct contributions are seen from the boron atoms and their exocyclic substituents, the Lewis acidic bridging borons withdraw electron density from the pyridyl groups and promote a more planarized and extended π‐system, thus favoring the orbital delocalization (Figure 2). This is seen most prominently for BNPz1', for which the HOMO is delocalized over the entire molecule except for the BMe₂ groups. For BNPz2'–BNPz4', the corresponding orbital becomes the HOMO‐1, because the expansion of the electron‐rich boron‐substituted acene moieties shifts the respective acene‐centered orbital higher in energy. Hence, the gradual increase in the HOMO energy levels in the order BNPz1' < BNPz2'  ≈ BNPz3' < BNPz4' is attributed to a combination of the σ‐donating character of the tetracoordinate borons and the expansion of the acene π‐system that they are attached to. The ordering of the unoccupied orbitals also changes with expansion of the acene π‐systems. For BNPz1'–BNPz2', the LUMOs are predominantly localized on the pyridyl substituents, not unlike for the free ligands. In contrast, for BNPz3' and BNPz4', the LUMO is predicted to be localized on the non‐borylated acene unit, whereas the pyridyl‐centered orbitals become the LUMO+1/LUMO+2. For BNPz3', the orbital energy difference between the LUMO and LUMO + 1 orbitals is very small. Because of these trends, the LUMO energies decrease with increasing acene π‐conjugation length, in the order of BNPz1' > BNPz2' > BNPz3' >> BNPz4'.

Figure 2.

Schematic illustration of computed orbital energy levels (not to scale) and frontier orbital plots of B ← N Lewis pair fused BNPz1 '–BNPz4 ' (RB3LYP/6–311G**; energies in eV).

Overall, B ← N Lewis pair‐functionalization induces a very pronounced lowering of the LUMOs and to a lesser degree the HOMOs (that of diazadihydroheptacene is slightly raised). As a result, relatively smaller HOMO‐LUMO energy gaps (Figure S32) are seen for BNPz2'–BNPz4' (3.20, 3.16, 2.36 eV) in comparison to Pz2–Pz4 (3.28, 3.40, 3.07 eV). An outlier is BNPz1' for which a dramatic decrease in the HOMO energy from −4.34 to −5.28 eV upon complexation outcompetes the lowering of the LUMO. Because of that, an increase of the HOMO‐LUMO energy gap from 3.13 to 3.47 eV is predicted. To shine further light onto the effect of B─N substitution, we have also performed computational studies on CCPz1', the all‐carbon analog of BNPz1'. The corresponding orbital energy levels and orbitals are illustrated in Figure 3 in direct comparison to the free ligand Pz1 and the isosteric B─N species BNPz1'. It is apparent that the HOMO‐LUMO gap is dramatically lower for the B─N species in comparison to the all‐carbon species. This is due to a far lower LUMO energy level and a more moderate decrease in the HOMO energy level for BNPz1' in comparison to CCPz1'. The much lower LUMO energy can be traced back to a predominant localization of the LUMO on the electron‐deficient pyridyl groups for BNPz1', whereas it shows strong contributions from the unfused benzene ring for CCPz1'. This has direct implications on the computed absorption characteristics. As illustrated in Table S8e, the lowest energy absorption for BNPz1' corresponds to a HOMO‐to‐LUMO transition, whereas that of CCPz1' is HOMO‐to‐LUMO+1 in nature. Furthermore, the computed transition for BNPz1' is much more intense and significantly lower in energy at 3.58 eV (f = 0.337) in comparison to that of CCPz1' at 3.83 eV (f = 0.036). Similar trends are seen for the emission energies (Table S9c). Overall, we conclude that the B ← N Lewis pair‐functionalization of dihydrodiazaphenazines presents a very effective way to lower the HOMO and LUMO energies, while an increase in the acene conjugation length decreases the HOMO‐LUMO gap.

Figure 3.

Schematic illustration of computed orbital energy levels (not to scale) and frontier orbital plots of free ligand Pz1, B ← N Lewis pair fused BNPz1', and the all‐carbon analog CCPz1' (RB3LYP/6–311G**; energies in eV).

To assess the effect of B ← N Lewis pair formation on the aromaticity of N,N‐diaryldihydrophenazines, nucleus‐independent chemical shift (NICS) calculations were performed (Table S5). Given their planar structure, for the free ligands the NICS(1)_zz and NICS(−1)_zz values are identical, but for the bent B ← N Lewis pair complexes slight differences are seen at the central dihydropyrazine ring; hence, averaged values are given. As previously reported for other dihydrophenazine derivatives,^{[ 95 ]} the central dihydropyrazine heterocycle in the free ligands exhibits local antiaromatic character. Enhanced delocalization upon extension of π‐conjugation from Pz1 to Pz4 impacts the electronic structure of the central ring, significantly reducing its antiaromatic character (Figure 4). More importantly, the antiaromaticity of the dihydropyrazine heterocycle is greatly reduced by B─N fusion, resulting in essentially a locally non‐aromatic central ring for BNPz1'–BNPz4', independent of the acene chain length. Conversely, the aromaticity of the adjacent benzenoid rings slightly increases upon B ← N Lewis pair formation, especially for the borylated benzene ring (Figure S33).

Figure 4.

Graphical illustration of trends in NICS(0) and NICS(1)_zz values for Pz1−Pz4 and B ← N Lewis pair fused BNPz1'−BNPz4'; averaged values of NICS(1) and NICS(−1) are given for the bent boron complexes.

Comparison of the structural parameters and NICS(1)_zz values with those of corresponding reference systems offers valuable insights into the origin of the reduced antiaromatic character of the central dihydropyrazine heterocycle upon B ← N Lewis pair formation. As illustrated in Figure 5, the computed exocyclic N─C(py) bond lengths of BNPz1' (1.378 Å) are much shorter than those of Pz1 (1.430 Å), indicative of extensive N lone pair delocalization into the electron‐deficient pyridyl rings. Conversely, the N─C bonds to the fused benzene rings, especially the borylated benzene rings (1.439 Å), are elongated relative to those in Pz1 (1.409 Å), suggesting that π‐delocalization is less efficient. These observations are consistent with the reduced antiaromatic character of the central dihydropyrazine ring, as well as the relatively more aromatic character of the borylated than the non‐borylated benzene ring derived from the NICS calculations. Of note is also the enhanced bond length alternation and the reduced aromatic delocalization seen in the pyridyl rings of BNPz1'. Together with the reduced N─C(py) bond lengths, these data are indicative of a significant contribution of the quinoid resonance structure of BNPz1' illustrated in Figure 5.

Figure 5.

Comparison of computed metric parameters and NICS(1)_zz values for BNPz1', ligand Pz1, and isostere CCPz1' with B─N units replaced by C─C units; averaged values of NICS(1) and NICS(−1) are given for the bent compounds.

Another question that arises is whether i) the steric effects of a more coplanar orientation of the pyridyl groups with the dihydrophenazine or ii) the electronic effect of the Lewis acidic boron centers (withdrawal of electron density from the pyridyl groups) is primarily responsible for the observed electronic structure modulation. A comparison of the structural features of BNPz1' to those of the carbonaceous analog CCPz1' (Figure 5) indicates that fusion of the pendent group in CCPz1' does impact the N─C bond lengths and aromatic character of the central ring (NICS(1)_zz = 2.4), but to a lesser degree than in BNPz1'. Similarly, Lewis acid–base complex formation in Pz1(BH₃)₂ shortens the C─N(py) (1.410/1.419 Å), elongates the other C─N bonds (1.417–1.421 Å), and reduces the antiaromatic character of the diphydrophenazine central ring (NICS(1)_zz = 4.1), but not to the same extent as what is seen for BNPz1' (Table S6). Collectively, these results suggest that the dramatic changes in the electronic structure upon borylative fusion of Pz1 to give BNPz1 originate from both coplanarization of the pyridyl pendent groups and electronic effects of Lewis acid–base adduct formation.

Electrochemical Properties

Next, we investigated the electrochemical properties by cyclic voltammetry in THF containing 0.1 m Bu₄N[PF₆] (Figure 6, Table 2). For ligand Pz1, a single irreversible reduction was detected within the accessible electrochemical window, whereas Pz2 and Pz3 showed two closely spaced consecutive reductions. The onset potential for the first reduction falls into a close range for Pz1 (−3.08 V), Pz2 (−2.91 V), and Pz3 (−3.00 V), consistent with pyridyl‐centered LUMOs that are similar in energy as predicted by the computational studies. The boron complexes BNPz1–BNPz3 also show two closely spaced quasi‐reversible reductions. The first reduction occurs between −2.62 and −2.63 V, which is much less negative than for Pz1–Pz3, consistent with the predicted decrease of the LUMO levels upon B ← N Lewis pair functionalization. This further confirms the strong effect of the Lewis acidic boron centers in removing electron density from the pyridyl rings and the non‐borylated acene framework.

Figure 6.

Cyclic voltammetry data for a) ligands Pz1–Pz3 and b) B ← N Lewis pair complexes BNPz1–BNPz3 (0.1 m Bu₄N[PF₆] in anhydrous THF, versus Fc^+/0, ν = 250 mV s⁻¹).

Table 2.

Electrochemical data and comparison to computational results.

Compound	E ox, CV a) (V)	E red, CV a) (V)	HOMO b) (eV)	LUMO b) (eV)	E g,CV b) (eV)	E g,DFT c) (eV)	E g,UV d) (eV)
Pz1	−0.17	−3.08	−4.63	−1.72	2.91	3.13	2.88 e)
Pz2	0.01	−2.91	−4.81	−1.89	2.92	3.28	2.86
Pz3	0.07	−3.00	−4.87	−1.80	3.07	3.40	2.92
BNPz1	0.29	−2.62	−5.09	−2.18	2.91	3.47	2.74
BNPz2	0.30	−2.63	−5.10	−2.17	2.93	3.20	2.79
BNPz3	0.30	−2.62	−5.10	−2.18	2.92	3.16	2.83

^a) Onset oxidation and reduction potentials, derived from CV scans in anhydrous THF versus Fc^+/0

^b) E _LUMO = –(4.8 + E _red), E _HOMO = –(4.8 + E _ox) derived from E _onset of CV scans.

^c) From DFT calculations at B3LYP/6‐311G** level of theory for compounds with methyl in place of ethyl groups.

^d) Estimated from absorption onset in THF.

^e) Estimated based on low energy tailing of absorption band.

For the oxidation processes, two well‐separated quasi‐reversible anodic waves are seen with onset potentials for the first wave at −0.17 V (Pz1), 0.01 V (Pz2), and 0.07 V (Pz3). Upon benzannulation in Pz2 and Pz3, the oxidation potential gradually increases, consistent with the decreasing HOMO energy because of the electron‐withdrawing nature of the enlarged π‐conjugated system. The boron complexes BNPz1–BNPz3 also show two distinct oxidation processes, but in contrast to those of the ligand precursors, they are more closely spaced and irreversible (Table S11a). The onset oxidation potentials for BNPz1 (0.29 V), BNPz2 (0.30 V), and BNPz3 (0.30 V) fall into a narrow range. They are much more positive than those of the ligands Pz1–Pz3, as the B ← N Lewis pair functionalization makes the system less electron‐rich (lowers the HOMO levels). The trends in the HOMO‐LUMO gaps determined by CV are by and large consistent with the HOMO‐LUMO gaps determined by DFT methods (and closely match those derived from the onset of UV–vis absorptions, vide infra) (Table 3). An exception is that the DFT‐predicted dramatic lowering of the HOMO of BNPz1 and ensuing larger HOMO‐LUMO gap is not well reproduced experimentally, suggesting that the computations may overestimate the effect of the borylation on the HOMO delocalization of this compound.

Table 3.

Comparison of photophysical properties of free ligands and B ← N Lewis pair complexes in solution, 2 wt% doped in PMMA film, and in the solid state.

Compound	λabs,TDDFT / nm	λabs,THF a) / nm	εTHF a) / 104 M−1cm−1	λFl,THF a) / nm	Stokes a) / cm−1	τFl, tol c) / ns	ϕF, tol b) / %	k r / k nr d) / 107 s−1	λFl,ss e) / nm	λFl,PMMA e) / nm	ϕFl,ss f) / %	ϕFl,PMMA f) / %
Pz1	287 g)	370 i) , 319	1.07	542	8580	4.5 (100%)	1.0	0.12 / 22	481	515	0.14	2.1
Pz2	335 h)	391, 289	1.06, 6.86	487	5040	5.0 (70%)	3.0	0.52 / 20	473	458	0.25	3.1
Pz3	349	416, 394, 374, 299	2.06, 1.55, 0.88, 8.43	421, 446, 474	2940	4.4 (71%)	17.1	3.5 / 19	463	420	0.08	1.6
BNPz1	347	399, 264	1.41, 1.36	565	7360	1.2 (97%)	2.7	1.4 / 82	527	508	3.6	0.53
BNPz2	353	383, 272	0.78, 2.45	549	7890	1.9 (85%)	2.7	0.82 / 52	481	509	4.1	0.12
BNPz3	355	385, 288	1.37, 4.25	546	7660	2.1 (83%)	8.6	0.89 / 47	511	513	6.5	j)

^a) Data in THF solution; excited at longest wavelength absorption maximum (c = 4 × 10⁻⁵ M, except for Pz3 at c = 2 × 10⁻⁵ M).

^b) Relative quantum yield in toluene solution.

^c) Fluorescence lifetime in toluene solution, excited with a nanoLED at 300 nm (major component of double exponential fit given, see Supporting Information for detailed contributions to multi‐exponential fits).

^d) Radiative (k _r) and non‐radiative (k _nr) decay rate constants in toluene solution estimated using the equations k _r = Φ / τ, k _nr = (1 − Φ) / τ) based on major decay component.

^e) Emission in the solid state or 2 wt% doped in PMMA film; excited at longest wavelength maximum of excitation spectra.

^f) Absolute quantum yield in the solid state or 2 wt% doped in PMMA film.

^g) S₀←S₆ transition, the oscillator strength for the S₁←S₀ transition at 373 nm is 0.

^h) S₀←S₃ transition, the oscillator strength for the S₁←S₀ transition at 354 nm is 0.

ⁱ⁾ Shoulder band.

^j) Quantum yield too low to determine accurately.

Photophysical Properties

The photophysical data of the free ligands and B ← N Lewis pair‐functionalized diazadihydroacenes are summarized in Table 3, and their absorption and emission spectra in THF solution are depicted in Figure 7. For the ligand precursors, a dominant high‐energy band is observed at 319 nm (Pz1), 289 nm (Pz2), and 299 nm (Pz3) in THF solution. This band is accompanied by weaker structured absorptions in the region from 350 to 430 nm for Pz2 and Pz3. Only a faint long wavelength tail is seen for Pz1 at ∼370 nm. TDDFT calculations (rcam‐B3LYP/6‐311G**) suggest that the weak low energy absorptions are primarily HOMO‐LUMO in nature and involve significant ICT from the acene backbone to pyridyl‐centered orbitals, especially for Pz1 and Pz2 (Table S7, Figure S34).

Figure 7.

UV–vis absorption spectra of a) ligands and b) B ← N Lewis fused dihydrodiazaacenes in THF solution, spectra are normalized to longest wavelength absorption maxima. Fluorescence spectra of c) ligands and d) B ← N Lewis fused dihydrodiazaacenes in THF solution (c = 4 × 10⁻⁵ m, except for Pz3 at c = 2 × 10⁻⁵ m; excited at 319 nm (Pz1), 391 nm (Pz2), 416 nm (Pz3), 399 nm (BNPz1), 383 nm (BNPz2), and 385 nm (BNPz3)). Photographs of solutions of e) ligands and f) B ← N Lewis pair fused dihydrodiazaacenes, irradiated with a handheld UV lamp at 365 nm.

The emission spectra of the ligands show maxima at wavelengths between 421–542 nm in THF solution. Pz1 and Pz2 give rise to broad emission bands at 542 and 487 nm, respectively, whereas a structured higher energy emission is detected for Pz3 with maxima at 421, 446, and 474 nm. The very large Stokes shifts and broad featureless emissions for Pz1 and Pz2 support pronounced structural reorganization in the excited state with emission from relaxed states with ICT character for the smaller dihydrophenazines.^{[ 33 ]} In good agreement, the emission maximum of Pz1, and to some degree Pz2, experiences a redshift with increasing solvent polarity (Figure S37, Table S14a). In sharp contrast, the structured nature of the emission of Pz3, and the fact that the emission shows almost no solvent dependence suggests that radiative decay occurs from a locally excited state with π–π* character. The shift of the emission to shorter wavelengths when going from Pz1 to Pz3 is also well reproduced by TDDFT calculations performed on the compounds at the S₁‐optimized geometries (Table S9).

The lowest energy absorptions for the B ← N Lewis pair‐functionalized complexes appear in a very narrow wavelength range (λ _abs = 399, 383, and 385 nm for BNPz1–BNPz3) and are accompanied by higher energy bands as also seen for the free ligands (Figure 7). When compared to the ligand precursors, upon boron complexation, the lowest energy absorption redshifts in the case of BNPz1 but slightly blueshifts for BNPz2 and BNPz3. Importantly, the absorption maxima of all complexes are red‐shifted relative to that of the parent non‐fused N,N‐diphenyldihydrophenazine (λ _{abs, THF} = 357 nm).^{[ 33 ]} The absorption maxima of BNPz1–BNPz3 experience a modest blueshift with increasing polarity (Figure S38, Table S14b), which indicates a certain degree of ground state stabilization by more polar solvents. TDDFT calculations suggest that the lowest energy transitions arise from complex contributions of multiple orbitals, but the HOMO‐LUMO excitation dominates. The transitions involve significant ICT from acene‐ to pyridyl‐centered orbitals, similar to the ligand precursors (Table S8, Figure S35). Only small changes in excitation energy are predicted with additional benzo‐fusion from BNPz1 to BNPz3, not unlike what is seen in the experimental spectra. All three complexes show a single unstructured emission band with maximum intensity in the range of ca. 546–565 nm. Although the emission of BNPz1 closely mirrors that of the free ligand Pz1, the emission bands of BNPz2 and BNPz3 are bathochromically shifted relative to those of the ligand precursors Pz2 and Pz3, signifying a strong effect of B ← N Lewis pair functionalization on the electronic structure. Similar to the absorptions, the emissions are largely red‐shifted relative to the non‐fused parent molecule, N,N‐diphenyldihydrophenazine (λ _{em THF} = 462 nm).^{[ 33 ]} The emission maxima experience a modest redshift with increasing polarity (Figure S38, Table S14b), which is consistent with ICT character in the excited state. The solvatochromic effect is most pronounced for BNPz1 and least for BNPz3.

The large Stokes shifts observed for compounds BNPz1–BNPz3 raise the possibility that, despite the B─N Lewis pair fusion, the compounds experience a certain degree of planarization in the excited state, emitting from a structurally partially relaxed state. This possibility is further supported by computational studies where the geometry of the complexes was optimized in the S₁ excited state. The optimized structures show consistently a slight increase in the bend angle Θ_b (Ph_B//Ph_H) toward a more planar conformation, from 139.1–146° in the ground state to 143.6–152.6° in the first singlet excited state (Table S4). At the same time, one of the B─N bonds is significantly shortened, resulting in an unsymmetric, more twisted conformation. The fluorescence decay curves were fitted to double exponentials with lifetimes in the low nanosecond range. indicating that two (or more) different decay processes occur simultaneously (Table S12). The fact that the fitted decay lifetimes τ ₁ and τ ₂ are in a similar range may suggest that emission occurs from different excited state conformers, consistent with the partial relaxation (planarization) of the excited state geometry as detailed in Table S4. When acquiring emission data on PMMA thin films doped with the complexes at 2 wt%, we consistently observed a blueshift in the emission maximum and a slight increase in the fluorescence lifetime (Figure 8a–c, Table S14b). Despite being incorporated into the rigid polymer matrix, the fluorescence quantum yield for BNPz1–BNPz3 in PMMA was low (<1%, Table 3), indicating that structural relaxation promotes non‐radiative decay.

Figure 8.

Normalized emission spectra of a) BNPz1, b) BNPz2, and c) BNPz3 as solutions in 2‐methyltetrahydrofuran (MeTHF), 2 wt% doped into a PMMA film, and in the solid state (PMMA films and powders excited at λ _max of excitation spectra); insets show photographs of powder samples under irradiation with UV light (365 nm). Emission spectra of d) BNPz1, e) BNPz2, and f) BNPz3 in THF containing varying amounts of water; insets show photographs of solutions containing 5% THF/95% H₂O under irradiation with UV light (365 nm).

Solid State Emission and AIEE Effect

We next explored the potential for AIEE^{[ 108 ]} through molecular rigidification in the solid state. AIEE has been previously reported for N,N‐diaryldihydrophenazine derivatives, including compound Pz1;^{[ 33 ]} hence, the possible impact of B─N Lewis pair fusion on the AIEE properties was of interest to investigate. When gradually adding water to solutions of ligands Pz1–Pz3 in tetrahydrofuran, the emission was largely quenched, and no emission enhancement could be observed up to 95% water content (Figure S39). In stark contrast, for the B ← N Lewis pair fused complexes BNPz1–BNPz3, a clear AIEE effect was detected (Figure 8d–f). Initially, when water was added up to 70%, the emission intensity decreased gradually to  ∼50% relative to that in pure THF, possibly because of an increase in solvent polarity. However, when further increasing the water content to 95%, the trend was reversed, and the emission intensity was greatly boosted for all compounds, reaching about four times the intensity relative to pure THF for BNPz1. Upon aggregate formation, a slight blueshift of the emission was seen for all compounds. This blueshift of the emission is consistent with solid‐state emission data, which show shifts to higher energy for all complexes (Figure 8a–c). For BNPz2, the shift in the solid state is relatively larger, resulting in a blue–green emission, possibly due to crystal packing effects (vide infra). The measured solid‐state quantum yields for complexes BNPz1–BNPz3 (3.2%–5.8%) are about an order of magnitude higher than for the doped PMMA films (<0.6%). They are also much higher than for the non‐borylated ligand precursors, Pz1–Pz3, which give rise to very low emission quantum yields in the solid state (<0.3%), indicative of aggregation‐caused quenching (ACQ) effects.

Closer inspection of the extended structures of BNPz1–BNPz3 derived from the single crystal X‐ray diffraction data offers possible clues as to the origin of the enhanced solid‐state emission and AIEE effects seen for the B ← N Lewis pair complexes. For BNPz1, two molecules come together in the unit cell through π–π stacking interactions between pyridyl groups (Figure S23a) and for BNPz2, two crystallographically independent molecules in the asymmetric unit assemble into an extended chain‐like 1D polymer structure through weak C─H π‐interactions (Figure S23b). Finally, for BNPz3, intermolecular interactions promote the formation of cylindrical pores that extend throughout the crystal, which accommodate disordered EtOAc/hexanes solvent molecules (Figure S26). We conclude that the borylative fusion in BNPz1–BNPz3 not only sterically disfavors ACQ effects seen in Pz1–Pz3 but also promotes AIEE, as is evidenced by the water‐titration experiments and increased solid‐state quantum yields.

Conclusion

Although structural extension through benzannulation of N,N‐diaryldihydrophenazines has been established as a powerful approach to control the ground and excited state structures and to enable unusual emissive characteristics, the effects of direct fusion of the N‐aryl groups to the dihydrohenazine backbone have so far sparsely been explored. Herein, we introduce a facile approach for the fusion of N‐aryl groups to the phenazine backbone via B ← N Lewis pair formation. B ← N Lewis pair fusion in BNPz1–BNPz3 is readily achieved by N‐directed electrophilic borylation of N,N‐dipyridyldihydrophenazine and its benzannulated congeners. X‐ray structure analyses reveal a strongly bent structure in the solid state. The B─N bond lengths are in the typical range, and they are retained in solution according to multinuclear NMR studies. DFT calculations reveal that the B─N fusion enforces a bent conformation also in the singlet excited state, albeit slightly more planarized (from 139.1–147.7° in the ground state to 143.6–152.6° in the excited state) and less symmetric (twisting occurs because one of the B─N bonds becomes shorter). Importantly, B ← N Lewis pair formation induces polarization of the molecules and favors a quinoid resonance structure in which the exocyclic C─N(py) bonds show significant double bond character. This is also reflected in changes in the aromaticity of the central dihydropyrazine ring that turns from antiaromatic to non‐aromatic and the pendent pyridyl groups that show reduced aromatic character after borylation according to our NICS calculations. The B─N fused dihydrophenazines are weakly emissive in solution, giving rise to yellow emission that is bathochromically shifted from that of the ligand precursors. The emission intensity is enhanced upon addition of water to solutions in THF, as well as in the bulk state, indicative of AIEE effects. Thus, we demonstrate that B─N fusion can serve as a new approach to induce AIEE effects in molecules that do not otherwise exhibit such behavior. Collectively, our results show that electrophilic B ← N Lewis pair functionalization not only offers facile access to novel fused dihydrodiazaacenes but also opens new avenues for fine‐tuning their electronic structures and optoelectronic properties with implications for potential application as emissive materials and in photocatalysis. Our results will also guide future work on additional rigidification of dihydrodiazaacenes through B─N Lewis pair formation that is in progress, including attempts to isolate the trans‐isomers that are fused on both sides of the dihydrophenazine core. Preliminary studies on the photostability of the B─N Lewis pair complexes BNPz1 suggest good stability in the solid state but photodegradation occurs upon prolonged exposure to UV irradiation in solution, most likely through the formation of ethyl radicals. This can be prevented by the use of phenyl in place of ethyl pendent groups on boron. Current work is in progress on the exploration of derivatives with aryl pendent groups on boron as luminophores and the exploration of the ethyl derivates presented here as initiators in photo‐induced free radical polymerization of vinyl monomers.^{[ 109} , ^{110 ]}

Supporting Information

The authors have cited additional references within the Supporting Information.^{[ 111} , ¹¹² , ¹¹³ , ¹¹⁴ , ¹¹⁵ , ¹¹⁶ , ¹¹⁷ , ¹¹⁸ , ¹¹⁹ , ^{120 ]}

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Supporting Information

Sahoo A., Patel A., Lalancette R. A., Jäkle F., Angew. Chem. Int. Ed. 2025, 64, e202503658. 10.1002/anie.202503658

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.