Krzysztof
Bartkowski‡
a,
Abhishek Kumar
Gupta‡
b,
Tomas
Matulaitis
b,
Maja
Morawiak
a,
Eli
Zysman-Colman
*b and
Marcin
Lindner
*a
aInstitute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland
bOrganic Semiconductor Centre, EaStCHEM School of Chemistry, University of St Andrews, St Andrews, UK
First published on 4th December 2023
We report the synthesis and characterization of the first pyrrolizine-embedded PAH systems. The bipolar core, based on a naphthalimide and indole fusion, was successfully synthesized through a one-pot cascade of Suzuki/Boc deprotection/Buchwald–Hartwig transformations. A facile post-functionalization of the unoccupied β-position of the heterocycle with aromatic amines as donor subunits led to a set of donor–acceptor architecture-based dyes. Effective narrowing of the HOMO–LUMO gap, along with its tuning by changing donor strength together with stabilization of π conjugation within the pyrrolizine–NMI core, provided a route to near-infrared (NIR) emission not only in solution (∼740 nm) but also in the solid-state (∼700 nm).
Regardless of the type of derivatization performed, these moieties exhibited a planar geometry that fosters a strong packing arrangement (even though double N-annulated species adopted a bowl-like geometry as displayed in Fig. 1, Ib)15 which is detrimental from the viewpoint of emissive properties. Instead, N-annulated PDIs were successfully exploited as non-fullerene acceptors for organic solar cells. In a similar fashion, but creating a 7-membered, azepine-bridged PDI (see Fig. 1, Ic), Shinokubo and co-workers16 presented a twisted species which in spite of possessing the desired non-planar arrangement shows modest optical features. Facing this challenge, very recently the second approach in which an electron-rich system is thoroughly fused and conjugated with an electron-poor moiety has also been demonstrated by our group (Fig. 1, IIa); that is, we showed the formation of a PAH consisting of 1,8-NMI and 3,6-di-tert-butylcarbazole (dtBuCz) via a 6-membered ring closure utilizing one-pot Pd-catalysed cascade transformations encompassing the formation of C–N/C–C bonds by virtue of Buchwald–Hartwig cross-coupling/direct arylation reactions (Fig. 1, IIa).17 This compound is very bright (PLQY ∼ 95%). We demonstrated its utility as a terminal emitter in a so-called hyperfluorescent OLED that showed a maximum external quantum efficiency of 26%. Shortly after, Würthner and co-workers18 reported another approach for synthesizing similar types of annulated compounds that were accessed following a cascade of Suzuki cross-coupling/nucleophilic addition reactions that emit in the green region with moderate-to-high PLQY (Fig. 1, IIb). In a similar fashion, but with 8H-benzo[3,4][1,2]azaborinino[5,6,1-jk]carbazole-8-ol as a precursor, Luo and Liu19 reported a route to compounds containing a 7-membered ring consisting of a combination of Suzuki coupling and Buchwald–Hartwig amination reactions, as demonstrated in Fig. 1, IIc. Notably, all of these structures contain highly conjugated ambipolar architectures that are emissive as a result of the presence of a low-lying charge-transfer excited state. Thus, modification of their electron-rich segment is envisaged as a route to developing compounds that should emit in the red and NIR regions.
Here, we report a straightforward synthesis of the first example of fused NMI and pyrrolizine fragments, in which the β-position (see Fig. 1, IId) is functionalized with a set of electron-rich donor moieties, allowing us to establish an efficient approach toward a new family of highly emissive green-to-NIR dyes.
The synthesis of 3a–c was accomplished in a three-step sequence (Scheme 1). We first tried to react N-(2,6-diisopropylphenyl)-4,5-bromo-NMI with unsubstituted indole; however, this approach failed to provide compound 1. We next employed N-Boc-indole-2-boronic acid as a precursor. To enable the thermal decomposition of the Boc protecting group, the temperature of the reaction was increased, which proved to be the key to the sequential Buchwald–Hartwig amination and direct arylation, leading to fused intermediate 1 in 40% yield. With this in hand, its bromination led to 2 in quantitative yield, which was employed in a Buchwald–Hartwig amination with dtBuCz, 9,9-dimethyl-9,10-dihydroacridine (DMAc) and phenoxazine (PXZ) donor moieties. Dyes 3a–c were obtained in moderate yields ranging from 30–45%. The structures of all intermediates and target compounds 3a–c were unambiguously verified using 1H NMR and 13C NMR spectroscopy and HRMS (see the ESI† for more details). To gain more insights into the solid-state structure of these dyes, X-ray quality, single crystal structures were obtained for compounds 1 and 3c by slow evaporation from solutions of DCM and chloroform, respectively (Fig. 2A and B). Compound 1 crystalized in the triclinic P space group with two molecules per unit cell adopting antiparallel orientation of neighbouring molecules (Fig. 2A). Short intermolecular distances (3.52 Å) exist in the crystalline phase of unfunctionalized 1, which imply strong π–π interactions between head-to-tail arranged molecules (Fig. 2A). Compound 3c crystallized in monoclinic P21, also with two molecules per unit cell. The presence of the twisted quasi-axial PXZ donor (60°, Fig. 2B) influences the interactions in the crystal structure, which is reflected in the prevention of π-stack creation (Fig. S15†).
![]() | ||
Fig. 2 X-ray structures and their packing of (A) 1 and (B) 3c. The thermal ellipsoids are shown at 50% probability. The hydrogen atoms and solvent molecules have been removed for clarity. |
To gain insight into the photophysical behaviour of these compounds, we computationally modelled their structure and properties using density functional theory (DFT) at the PBE0/6-31G(d,p)20 level of theory in the gas phase. Compound 1 possesses a structure where the central “pyrrolizine-fused-NMI (PF-NMI)” core is planar, while the diisopropylphenyl moiety is attached orthogonally (dihedral angle > 89°) to the PF-NAI group (Fig. S10†). Compounds 3a–c possess a similar structure, but with the decorated donors adopting a twisted conformation with respect to the PF-NAI plane. The dihedral angle between the donor and PF-NAI plane is 50.8° in 3a, 67.4° in 3b and 60.4° in 3c (Fig. S10†); the computed structure broadly matches with that observed in the crystal structure of 3c.
The HOMO and LUMO of 1 are both delocalized across the PF-NMI core, with respective energies of −6.02/−2.77 eV. In 3a–c, the LUMOs are mainly located on the fused-NMI core and have comparable energies of −2.76, −2.81 and −2.83 eV, respectively, while the HOMOs are increasingly localized on the donor substituent.
The HOMO levels reflect the relative strength of the donor substituent, with the HOMO of 3c (−5.23 eV) being destabilized in comparison to those of both 3a (−5.56 eV) and 3b (−5.40 eV) because of the presence of stronger PXZ. The computed HOMO–LUMO gaps of 1, 3a, 3b and 3c are 3.25, 2.80, 2.59 and 2.40 eV, respectively (Fig. 3). This trend in HOMO–LUMO gaps is reflected in the S1 and T1 energies. The S1 and T1 energies of 1 are 2.88 eV and 1.84 eV, respectively, which are stabilized to 2.38 and 1.67 eV for 3a, 1.98 and 1.65 eV for 3b, and 1.87 and 1.50 eV for 3c. The calculated ΔEST values of 1 (1.04 eV) and 3a (0.71 eV) are much larger than those of 3b (0.33 eV) and 3c (0.37 eV). These large singlet–triplet gaps almost certainly preclude that these compounds would show TADF behaviour. The oscillator strength (f) of the S0 → S1 transition is higher at 0.4397 for 3a in comparison to 3b (0.1097), 3c (0.1950) and 1 (0.0655), reflecting the larger overlap of the frontier molecular orbitals (FMOs) in 3a.
![]() | ||
Fig. 3 Theoretical modelling of the energies of the HOMO/LUMO and the S1 and T1 states of 1 and 3a–c. |
The experimental HOMO and LUMO energies were deduced from electrochemistry, using cyclic and differential pulse voltammetry (CV and DPV) in DCM (Fig. 4 and Table 1). The electrochemical data were referenced to Fc/Fc+ and are reported relative to a saturated calomel electrode (SCE). All compounds show two highly reversible reduction waves at around −1.0 and −1.46 V, which were assigned to successive reductions of the PF-NMI acceptor. Compound 1 shows a quasi-reversible oxidation wave at 1.43 V, assigned to oxidation of the whole molecule. Compounds 3a–c show two reversible oxidation waves: the first wave is associated with oxidation of the donor moiety and the second one is associated with oxidation of the pyrrolizine-attached diisopropylphenyl moiety. The oxidation/reduction potentials of 1, 3a, 3b and 3c were determined from the first oxidation/reduction peaks from DPV and they are 1.43/−1.01, 1.16/−0.94 V, 1.07/−0.97 V, and 0.89/−0.93 V, respectively (Fig. 4). Based on these data, HOMO/LUMO values were determined to be −5.77/−3.33 eV, −5.49/−3.40 eV, −5.41/−3.37 and −5.23/−3.41 eV for 1, 3a, 3b, and 3c, respectively (Table 1) and the estimated electrochemical band gaps were 2.44, 2.09, 2.04, and 1.82 eV respectively, which align with the trend observed from the DFT calculations.
![]() | ||
Fig. 4 Cyclic and differential pulse voltammograms of 1, 3a, 3b, and 3c in degassed DCM (scan rate = 100 mV s−1). |
Compound | Toluene | 5% in CBP | |||||||
---|---|---|---|---|---|---|---|---|---|
HOMOa/eV | LUMOa/eV | λ UV /nm (ε/×103 M−1 cm−1) | λ PL /nm | τ /ns | Φ PL /% | λ PL /nm | Φ PL/% air/N2f | τ avg /ns | |
a E HOMO/LUMO = −(Eox/Ered + 4.8) eV,21 where Eox is the anodic peak potential and Ered is the cathodic peak potential calculated from DPV in relation to Fc/Fc+ as the internal reference and referenced versus SCE (0.46 V vs. SCE)22,23 in DCM with 0.1 M [nBu4N]PF6 as the supporting electrolyte. b At 298 K, values quoted are in solutions under N2, which were prepared by three freeze–pump–thaw cycles (λexc = 420 nm for 1, 480 nm for 3a, and 500 nm for 3b and 3c). c λ exc = 374 nm. d At 298 K, values quoted are in solutions under air. e Thin films were prepared by spin-coating 5 wt% doped samples in 4,4′-bis(N-carbazolyl)-1,1′-biphenyl (CBP) (λexc = 340 nm). f Photoluminescence quantum yields of the thin films were determined using an integrating sphere (λexc = 305 nm or 340 nm) under air and N2 atmosphere at 298 K. g Time-resolved PL decays were recorded at 298 K under an O2-free atmosphere (λexc = 374 nm). Average lifetime (τavg = ∑Aiτi2/∑Aiτi), where Ai is the pre-exponential for lifetime (τavg). | |||||||||
1 | −5.77 | −3.33 | 480 (39.4) | 521 | 8.3 | 42 | 552 | 33/34 | 17.3 |
3a | −5.49 | −3.40 | 409, 540 (33.1) | 580 | 7.6 | 83 | 610 | 68/68 | 7.0 |
3b | −5.41 | −3.37 | 463 (20.9), 548 | 659 | 8.9 | 22 | 654 | 34/35 | 11.6 |
3c | −5.23 | −3.41 | 458 (25.7), 568 | 741 | 0.8 | 1 | 700 | 7/7 | 3.1 |
We performed photophysical measurements of compounds 1 and 3a–c to understand their ground- and excited-state behaviour. Compound 1 shows a single strong absorption band at 480 nm (ε = 39.4 × 103 M−1 cm−1) in toluene (Fig. 5). DFT predicts this band to reflect a mixture of locally excited (LE) π–π* and charge transfer (CT) transitions because of the ambipolar nature of PF-NMI, which correlate with the high oscillation strength (f = 0.5305) of the S0 → S2 transition; however, the oscillation strength of S0 → S1 is very low 0.0655 (Fig. S11 and Table S1†) and so this band likely reflects a combination of these two transitions. In compound 3a, the low-energy absorption band is bathochromically shifted at 540 nm (ε = 33.1 × 103 M−1 cm−1) and has the same profile as that in 1 where the transitions are of mixed LE/CT character (f = 0.4397). The absorption profiles of 3band 3c are more complex. The low energy bands at 548 nm and 568 nm, respectively, reflect a CT transition from the donor to PF-NMI (Fig. S11 and Table S1†). The high energy bands (and associated high-energy shoulder) at 463 nm (ε = 20.9 × 103 M−1 cm−1) for 3b and at 458 nm (ε = 25.7 × 103 M−1 cm−1) for 3c (Table 1) are assigned to the absorption of the PF-NAI core as the profile grossly aligns with that of 1 and the band is only modestly blue-shifted by about 20 nm.
The parent compound 1 shows (Fig. 5) structured photoluminescence in toluene at λPL of 521 nm, indicative of emission from an LE state, and has a ΦPL of 42%. Compound 3a has a similar emission profile that is red-shifted at λPL of 581 nm, aligning with that predicted by DFT; however, the ΦPL is much higher at 83%. The emission profiles of 3b and 3c are distinctly red-shifted and unstructured at 659 nm (deep-red) and 741 nm (NIR), reflecting emission from a CT state (Fig. 5 and Table 1). The ΦPL decreases to 21.6% and 1.1% for 3b and 3c, respectively (Table 1); this is due to the lower oscillator strength associated with the CT transition and to the energy gap law.24,25 All emitters showed mono-exponential decay kinetics with photoluminescence lifetimes (τPL) of 8.3 ns for 1, 7.6 ns for 3a, 8.9 ns for 3b and 0.8 ns for 3c in toluene (Table 1 and Fig. S13†). All emitters exhibit positive solvatochromism; however, the degree of solvatochromism significantly increases from 1 (65 nm) to 3c (149 nm), which demonstrates that 1 and 3a have emissive excited states of mixed LE/CT character while the excited states of 3b and 3c are of CT character (Fig. S12†). We also investigated the PL behaviour of these emitters in the solid state (Fig. 5) by doping 5 wt% 1 and 3a–c in 4,4′-bis(N-carbazolyl)-1,1′-biphenyl (CBP). Surprisingly, the emission of compounds 1 and 3a was approximately 30 nm red-shifted to 552 and 610 nm (the structured emission, though less resolved is still present) while the ΦPL decreased (34% for 1 and 68% for 3a) relative to the measurements in toluene (Table 1). In contrast, the emission of 3b and 3c was blue-shifted to 654 and 700 nm, respectively, while the ΦPL increased to 35% for 3b and to 7% for 3c. All compounds showed multi-exponential decay kinetics, with average lifetimes τPL of 17.3 ns for 1, 7.0 ns for 3a, 11.6 ns for 3b and 3.1 ns for 3c (Table 1).
In summary, we report a one-pot cascade cyclization resulting in fully conjugated pyrrolizine-fused NMI scaffolds, which upon facile post-functionalization with a set of electron-rich substituents deliver red and NIR emissive properties (<700 nm) both in solution and in the solid state. The demonstrated design should enable further modifications, opening a new avenue for their use as NIR-emissive materials in optoelectronic devices which is the current subject of our studies.
Footnotes |
† Electronic supplementary information (ESI) available: Experimental procedures for the syntheses of materials, spectroscopy data of new compounds, single crystal X-ray crystallographic data, cyclic voltammogram, the copies of NMR spectra of new compounds, and theoretical calculation details. CCDC 2216025–2216027, 2290709 and 2290710. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3qo01914d |
‡ These authors contributed equally. |
This journal is © the Partner Organisations 2024 |