Palladium-catalyzed intramolecular aerobic oxidative cross-coupling of BH/CH between o-carborane and arenes

Abstract

An efficient Pd-catalyzed regioselective intramolecular aerobic oxidative dehydrocoupling of BH/CH between o-carborane and arenes has been achieved with the construction of a series of five-, six- and seven-membered rings under mild reaction conditions. Control experiments indicate that B–H activation proceeds preferentially over the aryl C–H. These new polyarene–carborane conjugates have potential applications in materials as demonstrated by pyrene fused o-carborane that exhibits unique dual-phase emission, intramolecular charge transfer (ICT), and aggregation-induced emission (AIE) properties.

Efficient Pd-catalyzed regioselective intramolecular aerobic oxidative coupling of BH/CH has been achieved with the construction of a series of five-, six- and seven-membered rings, facilitating potential applications in carborane-based luminogens.

Introduction

Carboranes are a class of polyhedral boron hydride molecular clusters in which one or more of the BH vertices are replaced by CH units.¹ Their structural features, such as spherical geometry and three-dimensional electron delocalization, make them valuable components in various applications, ranging from functional materials to pharmaceuticals.^2,3 The o-carborane unit is known for its distinct electronic characteristics, including steric electronic conjugation and position-dependent electron-donating and accepting properties.⁴ Due to the electronic interactions between the carborane unit and the attached π-conjugated substituents, o-carboranes have gained significant interest as a versatile boron element-block for constructing optoelectronic materials, especially for their remarkable ability to suppress aggregation-caused quenching (ACQ). This feature is mainly observed in carbon vertex-substituted derivatives.⁵ However, the limited functionalization methods for boron vertex-arylated o-carboranes restrict their further applications.⁶ Among the various methodologies developed for o-carborane B–H activation, transition-metal-catalyzed BH/CH oxidative coupling is regarded as the most direct and efficient method for the formation of C(sp²)–B bonds. For instance, Ir-catalyzed intermolecular oxidative CH/BH cross-coupling of carboranyl carboxylic acid has been demonstrated with thiophene⁷ and aryl substrates bearing various directing groups including carboxy,⁸ amide,⁹ and nitrogen-containing heterocycles (Scheme 1a).¹⁰ Additionally, Fe-catalyzed intramolecular CH/BH oxidative cyclization of o-carborane with 8-aminoquinoline-based auxiliaries leads to the formation of C,B-substituted carborane-fused phenanthroline derivatives (Scheme 1b).¹¹

Scheme 1. Transition metal-catalyzed dehydrogenative arylation.

These reactions generally require high temperatures and anhydrous/anaerobic reaction conditions for successful B–H activation. As part of an ongoing project in our laboratory, we aim to develop efficient and environmentally friendly methods for cage functionalization to synthesize boron-arylated o-carboranes with potential optoelectronic applications. We report here a Pd-catalyzed intramolecular aerobic oxidative dehydrocoupling of BH/CH between o-carborane and arenes with the construction of a series of five-, six- and seven-membered rings under mild reaction conditions (Scheme 1c).

Results and discussion

Our investigation began by examining the BH/CH oxidative coupling of 1-(2-methoxy-1-naphthyl)-2-benzyl-o-carborane (1aa) in the presence of Pd(OAc)₂ catalyst (10 mol%), H₂O₂ (1.0 equiv.) and TFA (1.0 equiv.) in HOAc at 40 °C in open air, giving acenaphtheno-o-carborane (2aa) with the construction of a boron-containing five-membered ring in 70% NMR yield (Table 1, entry 1). Screening of copper oxidants proved that Cu(OTf)₂ was the optimal choice, generating 2aa in 96% yield (Table 1, entries 2–4). Compared with Pd(OAc)₂, Pd(TFA)₂ and Pd(acac)₂ gave slightly lower yield of 2aa, whereas Pd(MeCN)₄(BF₄)₂ and PdCl₂ showed poor or no catalytic activity (Table 1, entries 5–8). Solvent was also very crucial for this reaction. The use of toluene, THF, and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as the solvent gave the cross-coupling product in 16–30% yields. Trifluoroacetic acid (TFA) offered a reduced yield of 78% (Table 1, entries 9–12). While the use of Pd(TFA)₂ as catalyst and TFA as solvent led to a 90% yield of 2aa (Table 1, entry 13). Further screening indicated that lowering the catalyst loading or oxidant amount resulted in reduced yields of 2aa (Table 1, entries 14 and 15). Lowering reaction temperature to 25 °C led to poor reaction efficiency (Table 1, entry 16). In view of the yield of 2aa, entry 4 was chosen as the optimal reaction conditions.

Table 1. Optimization of the conditions for 5-membered ring construction^a.

Entry	Catalyst	[O]	Solvent	Yieldb (%)
1	Pd(OAc)2	H2O2c	HOAc	70
2	Pd(OAc)2	Cu(OAc)2	HOAc	8
3	Pd(OAc)2	Cu(OPiv)2	HOAc	7
4	Pd(OAc)2	Cu(OTf)2	HOAc	96
5	Pd(TFA)2	Cu(OTf)2	HOAc	90
6	Pd(acac)2	Cu(OTf)2	HOAc	88
7	Pd(MeCN)4(BF4)2	Cu(OTf)2	HOAc	8
8	PdCl2	Cu(OTf)2	HOAc	N.R.
9	Pd(OAc)2	Cu(OTf)2	Toluene	16
10	Pd(OAc)2	Cu(OTf)2	THF	17
11	Pd(OAc)2	Cu(OTf)2	HFIP	30
12	Pd(OAc)2	Cu(OTf)2	TFA	78
13	Pd(TFA)2	Cu(OTf)2	TFA	90
14d	Pd(OAc)2	Cu(OTf)2	HOAc	78
15e	Pd(OAc)2	Cu(OTf)2	HOAc	72
16f	Pd(OAc)2	Cu(OTf)2	HOAc	44

^aAll reactions were carried out on 0.1 mmol scale in 1 mL of solvent.

^bNMR yield using 1,1,2,2-tetrachloroethane as an internal standard.

^c1 equiv. of TFA was added.

^d5 mol% Pd(OAc)₂.

^e0.5 equiv. Cu(OTf)₂.

^f25 °C, 12 h.

Under such optimized reaction conditions, the scope of this reaction was investigated and the results were compiled in Table 2. We first investigated the substrate generality of 2-methoxy-1-naphthyl group, affording the BH/CH dehydrocoupling products 2aa–2an in 73–92% isolated yields. The mild reaction conditions were tolerant of a variety of functional groups including F, Cl, CO₂Me, C(O)Me and CN, and no obvious electronic effect was observed in 2aa–2an. Notably, the reaction of substrate 1aa proceeded smoothly on 1.0 mmol scale with a slightly lower efficiency (86% yield, 334 mg). Moreover, steric hindrance of 2-substitutions on naphthyl group did not affect the reaction efficiency (2ao–2as). 1-(1-Naphthyl)-2-benzyl-o-carborane 1as also worked well, giving acenaphtheno-o-carborane product 2as in 81% isolated yield, while variation of the R³ group at 2-position of naphthyl had little effect on the reactivity for the formation of products 2ao–2ar. 1-(1-Naphthyl)-2-benzyl-o-carboranes (1a) bearing –OMe or –F substituents at different position of the naphthyl group offered the corresponding products (2at–2aw) in 69–86% isolated yields. For 2at–2aw, electron-donating units on naphthyl group generally offered better yields of 2a than those of electron-withdrawing substituents. However, heteroaryl containing substrate 1ax showed poor reactivity towards the BH/CH oxidative coupling between o-carborane and thienyl groups to afford 2ax in only 16% isolated yield due probably to the interactions between Pd and S atoms. On the other hand, different substituents at cage C(2) were also evaluated, leading to the corresponding products 2ba–2ga in moderate to excellent isolated yields. The results indicated that steric factor has an influence on the regioselectivity, as both B(4)- and B(3)-coupling products (2ha and 3ha) were obtained with a molar ratio of 6 : 1 in the reaction of C(2)–H substrate 1ha.

Table 2. Substrate scope of 5-membered ring construction^a,b.

^aAll reactions were carried out on 0.2 mmol scale in 2 mL of HOAc.

^bIsolated yields.

^cThe reaction was carried out on 1.0 mmol scale in 10 mL of HOAc.

^d12 h.

^e60 °C.

^f80 °C, 24 h.

Furthermore, 1,2-di(1-naphthyl)-o-carborane (4) was also compatible with this five-membered ring construction process to afford B(4,7)- and B(4,11)-cross-coupling products 5 and 6 in 11% and 45% yields, respectively. The molecular structures of 2aa, 2ha, 5 and 6 were further confirmed by single-crystal X-ray analyses, supporting the unambiguous assignment of the substituted vertices (Scheme 2).

Scheme 2. Intramolecular dehydrogenative coupling of 1,2-dinaphthyl-o-carborane.

We then examined the feasibility of six-membered ring construction via BH/CH oxidative coupling using 1-(2-biphenyl)-2-benzy-o-carborane (7a) as a model substrate. Treatment of 7a with 1 equiv. of Cu(OTf)₂ in the presence of 10 mol% of Pd(OAc)₂ catalyst in HOAc at 40 °C for 12 h afforded the desired product 8a in 88% yield (Table 3, entry 1, standard conditions). Screening of other solvents and oxidants did not give better results (Table 3, entries 2–7). Decreasing the catalyst loading to 5 mol% resulted in a decreased yield of 8a (Table 3, entry 8). It is worth noting that lowering the amount of copper salt to 0.5 equiv. gave the product in 91% yield. Finally, using 0.1 equiv. of Cu(OTf)₂ as the cocatalyst and air as the oxidant led to 90% yield of 8a at 60 °C (Table 3, entries 9–11).

Table 3. Optimization of conditions for 6-membered ring construction^a.

Entry	Variations from the “standard conditions”	Yieldb (%)
1	None	88
2	HFIP instead of HOAc	26
3	THF instead of HOAc	13
4	Cu(OAc)2 instead of Cu(OTf)2	12
5	Cu(OPiv)2 instead of Cu(OTf)2	9
6	2 equiv. AgOTf instead of Cu(OTf)2	20
7	1 equiv. H2O2/TFA instead of Cu(OTf)2	8
8	5 mol% Pd(OAc)2	59
9	0.5 equiv. Cu(OTf)2	91
10	0.1 equiv. Cu(OTf)2	75
11	0.1 equiv. Cu(OTf)2, 60 °C, 5 h	90

^aAll reactions were carried out on 0.1 mmol scale in 1 mL of HOAc.

^bNMR yield using 1,1,2,2-tetrachloroethane as the internal standard.

With the optimized conditions in hand, we investigated the scope and limitation of this 6-membered ring construction (Table 4). The results indicated that both electron-donating and -withdrawing substituents at 4′- and 6′-positions of biphenyl group were well tolerated, leading to the desired products (8a–8l) in moderate to very good isolated yields (68–85%). In addition, the mild reaction conditions were tolerant of various functional groups, including halogens (8e, 8f, 8k, 8l) and ester (8g). For C_cage-connected aromatic ring, biphenyls with substituents at 6- and 4-positions (8m, 8n and 8q–8t) gave very high coupling efficiency (75–88%). However, those with 5-Me, 5-Ph, or 3-OMe substituents delivered the corresponding products (8o, 8p and 8u) in relatively lower yields (61–65%), while the use of 5-F substituent led to no reaction. Excellent yields were observed for C(2)-Me and -Ph substituted o-carborano-phenanthrene products (8v and 8w). With C(2)-H substrate 7x, both 1,4- and 1,3-o-carborane fused phenanthrenes (8x and 9x) were isolated in 77% and 9% yields, respectively, probably for steric reasons.

Table 4. Substrate scope of 6-membered ring construction^a,b.

^aAll reactions were carried out on 0.2 mmol scale in 2 mL of HOAc.

^bIsolated yield.

^c12 h.

^d20 mol% Cu(OTf)₂, 12 h.

12-Dibiphenyl-o-carborane (10) also worked to afford the desired B(4,7)- and B(4,11)-crosscoupling products 11 and 12, respectively, in the same 38% isolated yields in the presence of 20 mol% Pd(OAc)₂ and 20 mol% Cu(OTf)₂ in HOAc at 60 °C (Scheme 3). The B(4)-, B(3)-, B(4,7)- and B(4,11)-regioselectivities in 8a, 8x, 9x, 11 and 12 were further confirmed by single-crystal X-ray analyses.

Scheme 3. Intramolecular dehydrogenative coupling of 1,2-dibiphenyl-o-carborane.

This strategy worked also for the synthesis of o-carborane fused seven-membered rings (Scheme 4). In the presence of 10 mol% of Pd(OAc)₂ and 1 equiv. of Cu(OTf)₂, the intramolecular BH/CH oxidative coupling of 1-(2-(benzyl)phenyl)-o-carboranes 13 afforded the corresponding coupling products 14a and 14b in 25% and 41% yields, respectively. Instead of the methylene unit, substrates with X = O- or NPh-bridging moiety between the two phenyl groups showed poor reactivities for the seven-membered ring formation. In addition, such oxidative cross-coupling was compatible with 1-(2-(1-naphthyl)phenyl)-o-carborane substrate 15 with the employment of 2 equiv. of Cu(OTf)₂ to give 16 in 30% yield. Low yields may be related to unfavorable entropy and enthalpy strains in the formation of 7-membered rings.

Scheme 4. Intramolecular dehydrogenative 7-membered ring construction.

To gain some insight into the reaction mechanism, several control experiments were carried out. The oxidative coupling reactions did not occur in the absence of palladium catalyst (Scheme 5a). The yield of 2aa was dramatically dropped to 7% without Cu(OTf)₂ (Scheme 5b). In contrast, replacement of air with N₂ in the reaction system offered only a 21% yield of 2aa under standard reaction conditions (Scheme 5c), suggesting that oxygen may act as an oxidant with the help of a Cu(ii) cocatalyst.¹² To gain additional information regarding the initial step of the reaction, treatment of 1aa with 20 mol% Pd(OAc)₂ under non-oxidative conditions using DOAc as the solvent led to 19–88% deuteration efficiency on the ten boron vertexes (Scheme 5d). These results indicated that B–H activation is more favorable than the aryl C–H activation in this Pd-catalyzed process, which is calculated [at the B3LYP level of theory in conjunction with the Lanl2dz basis set and the corresponding Hay–Wadt effective core potential (ECP) for Pd and standard 6-31+G** basis set for all remaining atoms] to be thermodynamically favorable with ΔG = −9.9 kcal mol⁻¹ (Fig. 1).

Scheme 5. Control experiments. (a) Reaction without Pd(OAc)₂. (b) Reaction without Cu(OTf)₂. (c) Reaction under N₂ atmosphere. (d) Palladium catalyzed BH deuteration in DOAc.

Fig. 1. Transition states calculated at the B3LYP/6-31+G**/Lanl2dz level of theory.

On the basis of the aforementioned experimental results and literature work,¹³ a plausible reaction mechanism is proposed in Scheme 6. Electrophilic attack of the Pd(ii) center at the B(4)–H affords an intermediate A, which undergoes intramolecular C–H bond activation to give the six-membered palladacycle B. It is noted that this reversible B–Pd bond formation will take place at all the ten BH vertexes of o-carborane cage, which follow the order B(3,6)–H ≪ B(4,5,7,11)–H < B(8,10)–H < B(9,12)–H relied on the differences in vertex charge.¹⁴ Among these, only the B(3) and B(4)–Pd can approach the aromatic C–H due to the substrate configuration. The excellent B(4)-regioselectivity can be attributed to the electronic effect of the boron cage as well as the steric effect of C(2)-substituents. Reductive elimination affords the final products 2aa with the release of Pd(0), followed by Cu(ii) promoted the aerobic oxidation of Pd(0) to Pd(ii).¹²

Scheme 6. Proposed reaction mechanism.

To further demonstrate the synthetic utility of the above intramolecular BH arylation protocol, 1-pyrenyl-o-carborane 17 was subjected to the palladium-catalyzed aerobic oxidative cross-coupling to afford the corresponding pyrene incorporated 1,4-disubstituted o-carborane 18 in 89% yield (Scheme 7). A single-crystal X-ray structure of 18 is shown in Fig. 2. The dihedral angle between the C_cage–C_cage bond and the pyrene moiety is about 43°. In the crystal packing diagram, there is no intermolecular π⋯π interaction due to the bulkiness of o-carboranyl unit, resulting in the suppressed ACQ effect. Moreover, one-dimensional continuous B_cageH⋯π interactions followed by the linear packing structure is observed.¹⁵

Scheme 7. Synthesis of carborane-fused polycyclic aromatic compound.

Fig. 2. (a) Molecular structure of 18. Hydrogen atoms are omitted for clarity. (b) Packing structures of 18.

The photophysical properties of 18 were subsequently investigated. The UV/vis absorption spectrum of 18 in solution was recorded and compared to those of pyrene and substrate 17.¹⁶ As shown in Fig. 3a, the absorption spectra of 17 and 18 exhibited bathochromic shifts relative to pyrene, which can be attributed to the involvement of the o-carborane unit in an extended π-conjugation.¹⁷ On the other hand, the dual-emission bands were observed in the emission spectra of both 17 and 18. While both 17 and 18 displayed emission peaks near 400 nm, their photoluminescence (PL) spectra differed significantly from pyrene, showing broad emission bands with peaks in the range of 570 to 600 nm (Fig. 3b). Furthermore, the B–C bond coupling in 17 caused a red-shifted emission, which was due to the increased electron-withdrawing effect of the bisubstituted o-carborane.¹⁸ To explore the dual-emission mechanism in 18, changes in the absorption and PL spectra were monitored in solvents with varying polarities (Fig. 3c and d). No major peak shifts were observed in the absorption spectra, which could be attributed to the minimal solvent effect in the ground state. In contrast, as the solvent polarity increased, substantial bathochromic shifts were observed in the broad emission band at longer wavelengths in the PL spectra, suggesting that the luminescence near 600 nm in 18 is originated from the intramolecular charge transfer (ICT) state.^16,18 The fluorescence near 385 nm was largely unaffected by solvent polarity, displaying a mirror-image relationship with the absorption spectrum in the longer-wavelength region, thereby confirming its assignment to the locally excited (LE) state.¹⁶

Fig. 3. (a) UV-vis absorption spectra of pyrene, 17 and 18 in THF (c = 1.0 × 10⁻⁵ M). (b) The normalized PL spectra of pyrene, 17 and 18 in THF (c = 1.0 × 10⁻⁵ M), the excitation wavelength is 340, 360 and 360 nm, respectively. (c) UV-vis absorption spectra of 18 in different solvents (c = 1.0 × 10⁻⁵ M). (d) The normalized PL spectra of 18 in different solvents (c = 1.0 × 10⁻⁵ M, λ_ex = 360 nm).

Subsequently, the aggregation and solid-state emission properties of 18 were examined (Fig. 4 and Table 5). In a study involving THF/water mixtures, a THF solution of 18 emitted yellow light at 590 nm due to its ICT emission. As the water fraction (f_w) in the mixture increased from 0% to 60%, the emission intensity gradually decreased, and the spectrum exhibited a continuous red shift, which can be attributed to the ICT effect.¹⁹ With further increases in f_w, the emission intensity significantly increased again, accompanied by a gradual blue shift. In the aqueous mixture with f_w of 99%, the emission shifted to 567 nm. Additionally, a higher quantum efficiency (94%) was observed for the powder sample compared to the THF solution and the aggregated state, indicating that 18 exhibits both aggregation-induced emission (AIE) and AIE enhancement (AIEE) properties. For this arylated o-carborane with a fused structure at the neighboring carbon and boron atoms for fixing molecular conformation, the elongation of the C_cage–C_cage bond in the excited state, followed by nonradiative decay, was proposed as the primary mechanism for emission quenching in solution.²⁰

Fig. 4. (a) Emission spectra of 18 in the binary THF/H₂O solution with different volumetric ratios (f_w, vol%) of water (c = 100 μM, λ_ex = 360 nm). (b) Dependence of I/I₀ ratios of 18 in THF with different water fractions. (c) Photographs of 18 in THF/H₂O solution with various fractions of water (f_w) taken under UV light (365 nm).

Table 5. Summary of emission properties of 18.

	λ em a (nm)	Φ PL b (%)	τ F (ns)
THF	384, 404, 590	17.7	6.0
99% H2O	567	34.5	9.8
Solid	570	94.1	7.9

^aExcited at 360 nm.

^bDetermined as an absolute value.

Conclusions

In summary, an efficient intramolecular aerobic Pd-catalyzed BH/CH oxidative coupling has been achieved, leading to the facile synthesis of previously unavailable carborane-fused five-, six- and seven-membered ring structures under mild conditions. A plausible reaction mechanism including sequential electrophilic B–H and C–H activation, and reductive elimination was proposed, which is supported by the deuterium labeling experiment and DFT calculations. The facile synthesis has potential application in carborane-based luminogens.

Data availability

The experimental procedures and additional data can be found in the ESI.† Crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre, under deposition number 2406471 (2aa), 2406472 (2ha), 2406473 (5), 2406474 (6), 2406475 (8a), 2406476 (8x), 2406477 (9x), 2406478 (11), 2406479 (12), 2406480 (14b), 2406481 (16), and 2406482 (18). Copies of the data can be obtained free of charge from the CCDC viahttps://www.ccdc.cam.ac.uk/structures/.

Author contributions

Z. Q. and Z. X. directed and conceived this project. Z. W. and J. Y. conducted the experiments. J. Z. did the theoretical work. All authors discussed the results and wrote the manuscript.

Conflicts of interest

There are no conflicts to declare.