Open Access Article
Tienan Jin
*b,
Sho Aidaa,
Takuma Satoa and
Masahiro Terada
a
aDepartment of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan
bResearch and Analytical Center for Giant Molecules, Graduate School of Science, Tohoku University, Sendai 980-8578, Japan. E-mail: tetsuo.kin.a6@tohoku.ac.jp
First published on 12th February 2026
Bis(polycyclic) aromatic enes (BPAEs), comprising two polyaromatic units bridged by an ethylenic linker, exhibit distinctive electronic properties, yet heteroatom-containing analogues like azafulvalenes remain scarcely explored despite their tunable π-conjugation and redox versatility. Herein, we report an efficient Pd(II)-catalyzed intramolecular diarylation of alkynes that enables the selective synthesis of previously inaccessible azafulvalene-based BPAEs through dual C–H bond activation. This strategy adopts N-aryl and 2-aryl-substituted biarylalkynyl indole scaffolds, enabling modular access to structurally diverse azafulvalene-based BPAEs with precisely incorporated nitrogen atoms. Mechanistic investigations and DFT calculations revealed a pivalate-assisted concerted metallation–deprotonation (CMD) pathway, with the syn-insertion step as the rate-determining process. The addition of trifluoroacetic acid significantly lowers the activation barrier of this key step, enhancing the overall reaction efficiency. The synthesized azafulvalene-based BPAEs exhibit broad and red-shifted absorption bands extending up to 650 nm, low-lying LUMO levels (−3.47 to −3.63 eV), and narrow HOMO–LUMO gaps, consistent with donor–acceptor electronic structures predicted by DFT and TD-DFT calculations. This work establishes a modular and atom-economical synthetic approach for constructing π-extended nitrogen-containing fulvalene frameworks and highlights the potential of azafulvalene-based BPAEs as promising building blocks for advanced organic optoelectronic materials.
C) bridge. Their rigid and planar π-backbones combined with inherently twisted alkene moieties impart unique electronic, optical, and structural properties, making them attractive candidates for functional organic materials.1–19 Among these, fulvalene-based BPAEs such as 9,9′-bifluorenylidene (99′BF) and its π-extended analogues have been extensively investigated due to their high electron affinity and strong electron-accepting ability, enabling applications in organic photovoltaics and electronic devices (Scheme 1a).1–6 The twisted alkene geometry and electronically diverse π-systems of such compounds also support dynamic properties, including molecular switching and oligoradicaloid behavior.5–12
![]() | ||
| Scheme 1 Pd-Catalyzed diarylation of alkynes via C–H activation for synthesis of bis(polycyclic) aromatic enes (BPAEs). | ||
In contrast, BPAEs incorporating heterofulvalene units, where one or both rings contain heteroatoms such as nitrogen, oxygen, or sulfur, have received far less attention.13,14 Tetrathiafulvalene-derived BPAEs are known for their excellent electron-donating character and wide use in advanced electronic materials, while six-membered heterofulvalene analogues such as dihydroacridine, xanthene, and thioxanthene derivatives exhibit unique photoresponsive and redox-active behaviors (Scheme 1a).15–20 Particularly intriguing are azafulvalene-based BPAEs, in which nitrogen atoms are embedded within the fulvalene core. This underexplored subclass offers promising features such as extended π-conjugation, enhanced electron-donating ability, and tunable redox properties. When combined with electron-deficient fluorenylidene units, azafulvalene frameworks could enable novel optoelectronic functionalities. However, despite the success of conventional synthetic methods, including Barton–Kellogg olefination1 and transition-metal-catalyzed cyclization,21–26 no general approach has been established for constructing azafulvalene-based BPAEs.
Transition-metal-catalyzed diarylation of alkynes has recently emerged as a powerful tool for building multifunctionalized alkenes and complex polycyclic frameworks.22,27–39 In particular, palladium-catalyzed intramolecular diarylation, often involving C–H bond activation, offers an efficient and atom-economical route to π-extended bispolycyclic enes. For example, Pd(0)-catalyzed intramolecular diarylation of diaryl-tethered alkynes bearing C–Br and C–H functionalities efficiently furnishes BPAE skeletons (Scheme 1b).36–39 More recently, a Pd(II)-catalyzed dual C–H activation of bis(biaryl) alkynes has enabled the direct synthesis of 99′BF-type frameworks without halide prefunctionalization (Scheme 1c).22 These strategies exemplify the increasing importance of multiple C–H activation as a step- and atom-economical method for building complex π-extended scaffolds. Despite these advancements, the development of modular and general synthetic methods for structurally diverse BPAEs, especially heteroatom-containing variants, remains a synthetic challenge.
Building on these advances, we report a Pd(II)-catalyzed intramolecular diarylation of alkynes that accomplishes selective activation of two distinct aromatic C–H bonds in N-aryl- and 2-aryl-substituted biarylalkynyl indole scaffolds (Scheme 1d). This strategy delivers the first general and modular access to azafulvalene-based BPAEs with controlled nitrogen arom incorporation. The resulting π-extended molecules display low-lying LUMO levels, narrow bandgaps, and long-wavelength absorption, underscoring their promise as next-generation organic optoelectronic materials.
:
1 mixture of PdCl2 and PivOH as catalyst and MnO2 as oxidant at 80 °C for 6 h, delivering product 2a in 83% yield (Table 1, entry 1). Altering the PdCl2
:
PivOH ratio to 1
:
1.5 or 2
:
1 under otherwise identical conditions lowered yields to 63% and 71%, respectively, indicating an optimal 1
:
1 ratio. No reaction occurred in the absence of PivOH (entry 2), with 1a recovered nearly quantitatively, demonstrating that PivOH is essential for the C–H activation step. Other Pd(II) halide salts, PdBr2 and PdI2, showed poor catalytic activity (entries 3 and 4). PdCl2 complexes bearing ligands, PdCl2(CH3CN)2 and PdCl2(PPh3)2, afforded 2a in moderate yields (entries 5 and 6). Pd(OAc)2 and Pd(OPiv)2 were completely ineffective, while Pd(TFA)2 gave 2a in 20% yield (entries 7–9). These comparisons highlight the unique effectiveness of the PdCl2/PivOH combination. Addition of trifluoroacetic acid (TFA) further improved the yield of 2a to 88% with 10 mol% TFA and 96% with 20 mol% TFA (entries 10 and 11). However, the PdCl2/TFA system without PivOH gave only a low yield of 16% (entry 12), confirming that both PivOH and TFA contribute positively but are not interchangeable. Other Brønsted acids such as AcOH and MeSO3H also increased conversion (entries 13 and 14). In contrast, the presence of base (CsOPiv) completely suppressed diarylation (entry 15). Finally, replacing MnO2 with alternative oxidants (2 equiv.), such as AgOPiv (0%), Ag2O (5%), CuCl2 (0%), or CuO (5%), resulted in no reaction or only trace to low yields. Collectively, these results underscore the uniquely effective role of MnO2 in combination with PdCl2, highlighting the superior catalytic performance of the PdCl2/PivOH/MnO2 system in promoting the present C–H activation and diarylation process.
| Entry | Cat. Pd(II) | Additive (mol%) | 2ab (%) | 1ab (%) |
|---|---|---|---|---|
| a Conditions: 1a (0.1 mmol), Pd (10 mol%), PivOH (10 mol%), additive (10 mol%), and MnO2 (5 equiv.) in DMAc (0.125 M, 0.8 mL) at 80 °C for 6 h.b 1H NMR yield determined using CH2Br2 as an internal standard. Isolated yield is shown in parentheses.c Using 15 mol% of PivOH.d Using 5 mol% of PivOH.e In the absence of PivOH. | ||||
| 1 | PdCl2 | 83, 63,c 71d | 9, 22,c 20d | |
| 2e | PdCl2 | trace | 99 | |
| 3 | PdBr2 | 16 | 78 | |
| 4 | PdI2 | 2 | 98 | |
| 5 | PdCl2(MeCN)2 | 56 | 40 | |
| 6 | PdCl2(PPh3)2 | 48 | 40 | |
| 7 | Pd(OAc)2 | trace | 99 | |
| 8 | Pd(OPiv)2 | trace | 99 | |
| 9 | Pd(TFA)2 | 20 | 80 | |
| 10 | PdCl2 | TFA (10) | 88 | 11 |
| 11 | PdCl2 | TFA (20) | 96 (94) | 0 |
| 12e | PdCl2 | TFA (10) | 16 | 84 |
| 13 | PdCl2 | AcOH (10) | 90 | 4 |
| 14 | PdCl2 | CH3SO3H (10) | 85 | 9 |
| 15 | PdCl2 | CsOPiv (100) | 0 | 89 |
Next, we investigated the substrate scope of various 2-alkynyl indoles bearing different functional groups using the optimized catalytic system (Table 2). Notably, the TFA additive showed a prominent effect on the reaction efficiency. For example, 2-alkynyl indoles with a phenyl group or a strongly electron-donating 4-methoxyphenyl group on the N-atom afforded the corresponding products 2b and 2c only in moderate yields in the absence of TFA, likely due to partial decomposition of the indole substrates. In contrast, the addition of 10 mol% TFA significantly improved the efficiency, affording 2b and 2c in 90% and 72% yields, respectively. This beneficial effect of TFA was even more evident for indole substrates bearing electron-withdrawing groups. For instance, substrates with CF3 or Cl substituents at the para-position of the N-phenyl ring gave poor yields without TFA, whereas the addition of TFA dramatically enhanced the efficiency, affording 2d and 2e in good to high yields. Likewise, the substrate bearing an electron-withdrawing CN group at the 3-position of the indole ring and an electron-donating p-tolyl group on the N-atom exhibited an outstanding TFA effect, delivering 2f in 99% yield. Notably, the gram-scale synthesis of 2f also proceeded efficiently without obvious decrease in the isolated yield. The BPAE structure of 2f was unambiguously confirmed by X-ray crystallographic analysis. As expected, the molecule exhibited a highly twisted geometry, with a dihedral angle of approximately 30° between the two polyaromatic rings, despite their connection through a C
C double bond. Furthermore, diarylation of indole substrates bearing electron-withdrawing CN or CO2Me substituents at the 3-position of the indole ring in the presence of TFA produced the corresponding products 2g–i in moderate to high yields. In contrast, substrates with a phenyl substituent at the 3-position of the indole ring afforded products 2j and 2k only in moderate yields, regardless of TFA, likely due to competitive diarylation between the 3-phenyl and biphenyl groups. The reaction was also compatible with substrates bearing CH3 or OCH3 substituents at the 5-position of the indole ring, affording 2l and 2m in 94% and 87% yields, respectively. Gratifyingly, the [5]helicene-conjugated BPAEs 2n and 2o were successfully synthesized via the diarylation of the corresponding N-naphthyl- or binaphthyl-alkyne-substituted indole substrates. In contrast, the substrate lacking a substituent at the 3-position of the indole ring failed to deliver the desired product, presumably because the high nucleophilicity at C3 promotes unproductive side reactions, leading to substrate decomposition.
Two intermolecular competitive kinetic isotope effect (KIE) experiments were carried out to probe the dual C–H activation process (Scheme S1 in the SI). In the reaction of a 1
:
1 mixture of 1f and 1f-d7, where the deuterium atoms were incorporated on the N-p-tolyl group of the latter, a low KIE value of 1.0 was obtained under the standard conditions, both in the presence and absence of TFA (Scheme 2a). Similarly, the reaction of a 1
:
1 mixture of 1f and 1f-d5, bearing deuterium atoms on the biphenyl moiety of the latter, gave slightly higher KIE values of 1.4 and 1.3, affording 2f-d4 as an equimolar mixture of syn- and anti-diarylation products (Scheme 2b). These consistently low KIE values suggest that neither of the two C–H activation steps is likely to be rate-determining.
The most plausible reaction pathway (Path A), together with alternative pathways B and C, is illustrated in Scheme 2c and d. To gain deeper mechanistic insight into this transformation, density functional theory (DFT) calculations were performed at the M06/def2-SVP level for the reaction of 1b in order to identify energetically favorable pathways, including key intermediates (IMs) and transition states (TSs) (Scheme 2d and e). The observation that a 1
:
1 PdCl2
:
PivOH ratio gives the best activity, together with the poor catalytic performance of Pd(OPiv)2 and Pd(TFA)2 (Table 1), indicates that the catalytically active species is generated in situ. We therefore propose that the first step is ligand exchange between PdCl2 and PivOH to afford a Cl–Pd–OPiv complex, with concurrent loss of HCl (Scheme 2c). This active Pd(II) species can coordinate to the alkyne moiety of 1b, affording the π-complex IM1A (Path A). Subsequently, a pivalate-assisted C–H palladation occurs at the 2′-position of the biphenyl unit in IM1A through a concerted metallation–deprotonation (CMD) mechanism, leading to the formation of aryl–Pd complex IM2A via transition state TS1–2A (Scheme 2c and e). Alternatively, a competing pathway (Path B), involving C–H palladation at the ortho-position of the N-phenyl ring in Pd-complex IM1B, was also considered (Scheme 2d, see Fig. S1 in SI). However, the calculated Gibbs free energy (ΔG) for TS1–2B (+8.0 kcal mol−1) is higher than that of TS1–2A (+5.6 kcal mol−1). Furthermore, the activation free energy (ΔG‡) for TS1–2A (+5.6 kcal mol−1) is lower than that for TS1–2B (+6.5 kcal mol−1). These results clearly indicate that the initial C–H palladation at the biphenyl unit (via TS1–2A) is both kinetically and thermodynamically more favorable than the corresponding process at the N-phenyl ring (via TS1–2B), thereby validating Path A as the dominant pathway. Elimination of HCl from IM2A furnishes intermediate IM3A, which subsequently undergoes an intramolecular syn-insertion of the Pd–aryl bond into the alkyne through transition state TS3–4A with a relatively high activation free energy (ΔG‡ = 14.8 kcal mol−1). This step affords the thermodynamically stable vinylpalladium complex IM4aA. An alternative pathway (Path C), previously proposed by our group,22 involves CMD-type C–H activation of IM3C to generate a biaryl–Pd intermediate IM4C via TS3–4C, followed by alkyne insertion (Scheme 2d). DFT calculations reveal that this pathway is both kinetically and thermodynamically unfavorable, as indicated by the high activation free energy of TS3–4C (ΔG‡ = 16.3 kcal mol−1) and the low stability of IM4C. Conformational rotation of the N-phenyl ring converts IM4aA into IM4bA, which then undergoes a second, rapid pivalate-assisted CMD C–H activation at the ortho-position of the N-phenyl ring via TS4–5A with a low ΔG‡ of 2.1 kcal mol−1, generating the six-membered palladacyclic intermediate IM5A. Subsequent elimination of PivOH from IM5A yields intermediate IM6A, which undergoes reductive elimination via TS6–7A with a moderate activation barrier (ΔG‡ = 9.2 kcal mol−1), delivering IM7A. Decoordination of IM7A furnishes the final product 2b together with a Pd(0) species. It is well established that MnO2 reacts with HCl to generate Cl2, MnCl2, and H2O. PdCl2 can be formed by oxidation of Pd(0) with Cl2 in the presence of HCl.40 On this basis, we propose that the Pd(0) species generated in the final step is reoxidized to PdCl2 by the in situ–generated Cl2 under acidic conditions, thereby regenerating the active Pd(II) catalyst and completing the catalytic cycle (Scheme 2c). Additionally, for unsymmetrical products such as 2f-d4 (Scheme 2b), the proposed syn-insertion mechanism predicts preferential formation of the syn-isomer. Partial post-reaction isomerization to the anti-isomer can nevertheless occur, resulting in a mixture of syn- and anti-diarylation products, consistent with observations reported in our previous study.22
Among the calculated steps, the syn-insertion process via TS3–4A exhibits the highest activation barrier, identifying it as the rate-determining step of the catalytic cycle. To further elucidate the role of TFA as an additive, additional DFT calculations were performed on the corresponding trifluoroacetate-bound Pd species (Scheme 2e). The calculated ΔG for the TFA-bound intermediates and transition state, IM3A′ (−4.6 kcal mol−1), TS3–4A′ (+7.4 kcal mol−1), and IM4aA′ (−25.0 kcal mol−1), are all markedly lower than those of the corresponding pivalate-bound species (IM3A, −0.7 kcal mol−1; TS3–4A, +14.1 kcal mol−1; IM4aA, −18.9 kcal mol−1). Notably, the ΔG‡ of TS3–4A′ is 2.8 kcal mol−1 lower than that of TS3–4A, indicating that coordination of TFA to the Pd center significantly stabilizes both the intermediate and transition-state species. These results suggest that TFA facilitates the syn-insertion step by lowering the associated energy barrier, thereby enhancing the overall reaction efficiency.
Encouraged by the successful diarylation method for constructing azafulvalene-based BPAEs, we next explored the synthesis of BPAEs incorporating positionally distinct azafulvalene cores. To this end, indole substrates 3, bearing N-alkynyl biphenyl and 2-aryl substituents, were designed to afford the corresponding azafulvalene-based BPAEs 4 (Table 3). Electron-withdrawing substituents such as CN or CO2Me were introduced at the 3-position of the indole ring to enhance the stability of the N-alkynyl moiety. Optimization studies revealed that the addition of TFA as an additive in the PdCl2/PivOH/MnO2 catalytic system significantly improved the reaction efficiency. For instance, the inclusion of 10 mol% TFA furnished product 4a in 86% yield within 3 h, whereas only 76% yield was obtained after 12 h in the absence of TFA. Indoles bearing electron-donating Me or MeO substituents at the para-position of the 2-phenyl ring exhibited negligible electronic effects, affording the corresponding products 4b and 4c in similarly high yields. Replacing the CN group with a CO2Me substituent at the 3-position of the indole ring had little influence on the reactivity, and product 4d was obtained in high yield. Furthermore, activation of the aromatic C–H bond at the 2-position of the naphthyl group on the indole ring proceeded smoothly, producing the π-extended BPAEs 4e and 4f in moderate yields. Notably, an N-alkynyl indole bearing a 2-benzothienyl substituent was also compatible with the current catalytic system, delivering the linear heteroacene-conjugated product 4g in 75% yield, regardless of the presence or absence of TFA.
| a Conditions: 3 (0.1 mmol), Pd (10 mol%), PivOH (10 mol%), TFA (10 mol%), and MnO2 (5 equiv.) in DMAc (0.125 M, 0.8 mL) at 80 °C. Isolated yields of 4 after silica gel chromatography are shown. |
|---|
![]() |
The UV/Vis absorption spectra of selected newly synthesized BPAEs in dilute chloroform are shown in Fig. 1a, and the corresponding longest absorption maxima (λmax) and absorption onsets (λonset) are summarized in Fig. 1b. The azafulvalene-based BPAEs exhibit broad absorption bands extending to 550–650 nm, with both λmax and λonset significantly red-shifted compared to the fulvalene-based analogue 99′BF.1 Notably, compounds 2f, 2n, and 2o, which possess an N-atom positioned away from the olefin bridge, display further bathochromic shifts compared to 4a and 4g containing an enamine motif. The frontier molecular orbital energy levels were estimated from cyclic voltammetry measurements (Fig. 1b). The new BPAEs exhibit slightly higher-lying HOMO levels (−5.40 to −5.53 eV) and markedly lower-lying LUMO levels (−3.47 to −3.63 eV) than those of 99′BF (HOMO: −5.58 eV; LUMO: −3.37 eV),1 resulting in narrower HOMO–LUMO gaps and consequently longer wavelength absorptions. These experimental observations are consistent with the results of density functional theory (DFT) calculations performed at the B3LYP/6-31G(d,p) level (Fig. 1c). DFT-based orbital contour analyses indicate that the HOMOs of 2f, 2n, 4a, and 4g are delocalized from the N-atom-containing polyaromatic ring to the olefin bridge, whereas their LUMOs are predominantly localized on the azafulvalene core, excluding the N-atom. In contrast, the HOMO of 2o is primarily localized on the helical polyaromatic segment, while its LUMO resides on the azafulvalene core. These distributions suggest a donor–acceptor type electronic structure that facilitates intramolecular charge transfer upon excitation. Time-dependent DFT (TD-DFT) calculations at the CAM-B3LYP/6-311+G(2d,p) level further indicate that, for compounds 2f, 2n, 4a, and 4g, the lowest singlet excited state (S1) arises predominantly from a HOMO → LUMO transition (see Tables S1–S5, SI). In contrast, compound 2o exhibits a dark S1 state with negligible oscillator strength corresponding to a HOMO−1 → LUMO transition, while its bright S2 state mainly involves the HOMO → LUMO transition.
CCDC 2496062 contains the supplementary crystallographic data for this paper.41
| This journal is © the Partner Organisations 2026 |