Boron clusters, such as carboranes, possess three-dimensional (3D) aromaticity and hold immense potentials in materials [
1–
15], catalysis [
16–
19], drug design [
20–
31], and organometallic complexes [
32–
40]. Hence, the functionalization of boron clusters is of vital importance. However, the presence of multiple inert B–H bonds in similar chemical environments poses a considerable challenge for the synthesis of carborane derivatives [
41]. Until now the selective and efficient B–H bond functionalization of carboranes has received tremendous research interest [
42–
82]. Over the past decades, metal-mediated B–H bond activation has become a prevalent strategy for the functionalization of carboranes owing to the robust capability of the metal that enables weakening of a B–H bond through the formation of a B–M bond for further transformations (
Fig. 1a) [
42–
66]. This approach has facilitated the synthesis of carborane-based functional molecules with distinctive physicochemical properties, such as boron cluster-extended heterocycles (
Fig. 1b) [
64–
66]. Despite the effectiveness of this approach, external ligands, additives or harsh reaction conditions are generally required [
42–
66]. In contrast, metal-free tactics are capable of overcoming these limitations, therefore being highly attractive [
67–
83].
It is well recognized that favorable non-covalent interactions play a critical role in synthetic chemistry, particularly in controlling regioselectivity or site-selectivity. Given the established significance of halogen bonding in activating inert bonds [
84–
89], this study explores metal-free B–H bond functionalization through non-covalent interactions with halogen-containing molecular motifs. In doing so, we identified a type of boron cluster-based non-covalent bond, defined as the cage···Ⅰ(Ⅲ) interaction, arising from the interaction between a
nido-boron cluster and a hypervalent iodine unit (
Fig. 1c). Single-crystal X-ray diffraction (SCXRD) studies and spectral data indicated the formation of a well-defined electron donor-acceptor complex between the
nido-carborane and the hypervalent iodine unit.
Further theoretical calculations demonstrated that this bonding involves significant charge transfer, which reduces the negative charge on the B(9)–H of
nido-carborane, thus facilitating subsequent B–H bond functionalization and transformations. Finally, the metal-free and site-selective B–H functionalization of carboranes has been achieved to lead to the facile formation of B–N, B–P, B–O, and B–S bonds (
Fig. 1d). Additionally, a new category of boron cluster-extended skeleton in the form of 2D-3D fused hetero-polycyclic structure has been generated (
Fig. 1d). Such an advanced framework enables the facile construction of novel functional molecules that show aggregation-induced emission (AIE) or oxygen sensing.
Carboranes are boron cluster species featuring three-dimensional aromaticity [
41,
90], which differs from planar aromatic rings. Recently, we have demonstrated that the
nido-carborane cluster can serve as a novel bonding unit for the construction of non-covalent interactions, such as the cage···
π interaction [
91]. Given the fact that the research on non-covalent interactions involving boron clusters remains limited [
91–
93], we are interested in the continuous exploration of new types of boron cluster-based non-covalent interactions. On the other hand, iodine is an important element, which has been extensively involved in functional molecules [
94–
97]. Notably, the manipulation of the bonding type of iodine would influence the electronic structures of iodine-containing compounds. To investigate the possible cage···Ⅰ(Ⅲ) interaction between a boron cluster and an iodine-containing unit, the
nido-carborane was chosen to interact with an electrophilic iodine compound. Along this line, the hypervalent iodine-supported
nido-carborane derivatives of
3–1 and
3–2, as shown in
Figs. 2a and
b, were synthesized through ion exchange reactions (see Supporting information for synthesis details).
With these compounds in hand, we began to analyze the crystal structures to shed light on the structural characteristics of the cage···Ⅰ(Ⅲ) interaction as depicted in
Figs. 2a and
b. The
nido-carborane anion is surrounded by several hypervalent iodine cations in both
3–1 and
3–2 due to the electrostatic interactions, as shown in the packing structures (Figs. S18 and S19 in Supporting information) with the carborane cage oriented towards the backside of the V-shaped organoiodine(Ⅲ) moieties. Moreover, various intermolecular contacts such as C–H···B–H are observed in the crystal structures, collectively endowing a robust non-covalent bonding network for the formation of the cage···Ⅰ(Ⅲ) interaction. With careful inspection, the cage···Ⅰ(Ⅲ) interaction with a bonding distance of 4.813 Å is identified for
3–1 (measured from the centroid of the boron cage to the iodine atom), alongside B–H···I and B···I contacts with distances of 2.896 and 3.547 Å (
Fig. 2a), respectively. A shorter cage···Ⅰ(Ⅲ) bonding distance of 4.753 Å is detected for
3–2, with B–H···I interactions at 2.741 and 2.958 Å, and B···I interactions at 3.401 and 3.966 Å (
Fig. 2b), respectively. These indicate that the cage···Ⅰ(Ⅲ) interaction can be influenced by the substituents of the aryl hypervalent iodine derivatives.
To further confirm the electronic interaction between the
nido-carborane and the hypervalent iodine unit, we investigated the UV–vis absorption spectra of
3–2 in both methanol and the crystalline state (
Figs. 2c and
d). The potassium salt of
nido-carborane (
1) and diphenyliodonium trifluoromethanesulfonate (
iodonium-2) were selected as control compounds. As shown in
Fig. 2c, no absorption peak over 250 nm was detected for
1 in methanol. In contrast, the absorption peaks around 260 nm for
iodonium-2 and
3–2 were observed, which correspond to the
π→
π∗ transitions of the phenyl substituents. Notably,
3–2 exhibits a new charge transfer absorption band in the 320–400 nm due to the electron donating-accepting interaction between
nido-carborane and hypervalent iodine groups (
Fig. 2c). The charge transfer absorption band is more clear for
3–2 in the crystalline state (
Fig. 2d). The spectral results were further validated by the theoretical calculations using the time-dependent density functional theory (TD-DFT) methods (Fig. S24 in Supporting information). Moreover, results from Job's plot indicates a 1:1 stoichiometric association between the iodonium ion and
nido-carborane anion under the experimental conditions (Fig. S13 in Supporting information). These observations indicate efficient electronic communication between the
nido-carborane and the hypervalent iodine cation. Due to the unique electronic structures of boron clusters [
98–
102], further theoretical calculations are conducted to gain a deeper understanding of these intermolecular interactions.
The interactions between the boron cage and the hypervalent iodine moiety for
3–1 and
3–2 have been analyzed by means of the ZORA-BLYP-D3(BJ)/TZ2P dispersion-corrected functional (see Supporting information for computational details). In particular, the two conformations taken from the X-ray structure correspond to the vertex of the V-shaped iodine cation interacts with the cage, depicted in
Fig. 2e. Energy decomposition analysis (EDA) reveals that the interaction between
nido-carborane and hypervalent iodine is mainly electrostatic (Table S6 in Supporting information), with ∆
Velstat much larger than ∆
Eoi in all cases (%elstat = 67–73 and%oi = 10–21). The nature of the interaction in terms of%elstat and%oi is very similar to that found in systems showing non-covalent cage···
π interaction [
91]. It has to be considered a non-covalent interaction despite the presence of a certain percentage of orbital (covalent) interactions. In vacuo, this positive charge is favored by directly interacting with the cage, thus avoiding certain repulsion between the boron cage and the negatively charged aromatic phenyl substituents of iodine fragment. This is further supported by the molecular electrostatic potential isosurfaces depicted in
Fig. 2f. On the other hand, this bonding structure is also favored by the attractive orbital interactions. The most relevant interaction is between the highest occupied molecular orbital (HOMO) orbital of the
nido-cage and the lowest unoccupied molecular orbital (LUMO) of the cationic iodine molecule (
Fig. 2g). As a consequence of this interaction, we observed certain charge transfer from the
nido-cage to the iodine unit in
3–1 and
3–2. For completeness, the non-covalent interactions (NCI) isosurfaces [
103] depicted in
Fig. 2h allow us to better see the interactions for systems
3–1 and
3–2.
The above results indicate the cage···I(Ⅲ) interaction-induced charge transfer, which can modulate the electronic structure of carborane, thus might be useful for B–H bond functionalization. In light of this, we chose the hypervalent iodine reagent PhI(OAc)
2, which can dissociate in solution to generate the (PhIOAc)
+ species [
104,
105], to both promote the formation of the cage···Ⅰ(Ⅲ) interaction and control the strength of a cage···Ⅰ(Ⅲ) interaction. Theoretical calculations support the presence of a cage···Ⅰ(Ⅲ) interaction between the (PhIOAc)
+ group and
nido-carborane group, as shown in the NCI isosurfaces (
Fig. 3a). Notably, the charge transfer leads to a significant decrease of the charge in the carborane cage from −0.575 to −0.145, and from −0.087 to −0.003 for the H (B(9)) atom (
Fig. 3a and Fig. S8a in Supporting information). The electronic effect can be utilized for further B–H bond functionalization and the formation of the B–I(Ⅲ) bond, which is similar to the classic concerted metalation-deprotonation (CMD) in metal-mediated B–H functionalization [
42–
66]. Along this line, we investigated the reaction of
nido-carborane and PhI(OAc)
2 in a solution of acetic acid and water. As a result, the B–I coupled product
3 was generated as confirmed by nuclear magnetic resonance (NMR) and high-resolution mass spectroscopy (HRMS) (
Fig. 3b). This species is stable at a lower temperature and its preparation can be easily scaled up to a gram scale in a rapid manner (
Fig. 3b). The subsequent substitution of B–I coupled products by nucleophiles, such as pyridine (
4), triphenylphosphine (
5), and dibenzyl sulfide (
6), gave rise to the corresponding carborane derivatives in moderate to excellent yields under mild conditions (
Fig. 3c). The compound
5 was further characterized by single-crystal X-ray diffraction, which unambiguously confirms the B(9) selectivity. Moreover, when a phenyl group is grafted on the carbon vertex of
nido-carborane, the reaction preferentially occurs at the B(11)–H site. As exemplified by the crystal structure of
7, the B(11)–H site is substituted, which is distant from the phenyl substituent. This should be attributed to the steric hindrance. The result differs from the oxidation coupling of
nido-carborane in our previous report [
74], which gave rise to a mixture of B(9)- and B(11)-functionalized products. For comparison, here the exclusive selectivity has been realized. Hence, the B–I coupled species could be used as a diverse reagent for further transformations.
Furthermore, theoretical calculations were conducted to elucidate the possible reaction pathway (Figs. S7 and S9 in Supporting information). The formation of the cage···Ⅰ(Ⅲ) interaction in the complex (1-I) is thermodynamically favorable, as indicated by an exergonic change of −6.0 kcal/mol. As the formation of the cage···Ⅰ(Ⅲ) interaction significantly reduces the negative charge of the carborane cage and the B(9)–H bond (Fig. S7), the B(9)–H bond functionalization is much more favored to form a B–I bond (compound 3), accompanied by a further energy drop of 45.7 kcal/mol (Fig. S7). The further nucleophilic substitution by pyridine leads to a continued energy drop of 32.5 kcal/mol (Fig. S9). The calculations support the B–I containing compound as an intermediate or a reagent for subsequent synthetic applications. In addition, the intramolecular charge transfer effect from the carborane (HOMO) to the cationic substituent (LUMO) for compounds 3–7 has been demonstrated by theoretical calculations (Fig. S28 in Supporting information).
It is well known that benzoxazole is a crucial structural motif extensively utilized in the fields of biomedicine and photoelectric materials [
106,
107]. Given the distinctive electronic structure and enhanced physical properties exhibited by boron cage extension systems [
64–
66] compared to 2D aromatic ring-extended molecular structures, we attempted to construct the carborane-incorporated benzoxazole analogs with 2D-3D fused structures by the cage···Ⅰ(Ⅲ) interaction-mediated synthetic protocol. Despite the challenges associated with the direct B–O cross-coupling of boron cluster [
42–
83], the aforementioned synthetic objective might be reached by the cage···Ⅰ(Ⅲ) interaction-induced B–H bond functionalization and subsequent intramolecular cyclization. To our delight, the boron cage extended structure was achieved by using the amide-bearing
nido-carborane
1a as the sole starting material, PhI(OAc)
2 as the organoiodine(Ⅲ) reagent and MeOH as the solvent at room temperature for 10 min, affording
2a in 86% isolated yield (
Fig. 4). The molecular structure of
2a was confirmed by SCXRD analysis to show an expected 3D carborane-fused benzoxazole. With the optimized conditions (Table S1 in Supporting information), we proceeded to explore the substrate scope. We were pleased to find that the
nido-carborane substrates bearing
para-substituted phenyl groups with various electron-donating substituents such as methyl (
2b, 2n, and
2o), methoxy (
2c),
n-propyl (
2d), tertbutyl (
2e), trifluoromethoxy (
2f) were well-tolerated, affording the desired products in excellent yields. In comparison, the electron-withdrawing substituents including halogen (
2h–
2k) and the ester group (
2l) led to slightly decreased yields (67%–83%), which is attributed to the reduced nucleophilicity of the amide oxygen. Steric hindrance also affects the production, as indicated by a lower yield of
2m. The substrates containing polycyclic and heteroaryl substituents, such as naphthyl (
2p), pyridyl (
2q), and thienyl (
2r), are well tolerated. If the phenyl group is replaced by an alkyl substituent, the reaction also proceeds smoothly in a synthetically useful yield (
2s). Additionally, the reaction is compatible with substrates bearing a chromophore like TPE (
2t) or a pharmacophore derived from camphanic acid (
2u). Moreover, a new 2D-3D fused structure with a six-membered heterocyclic ring (
2v) was successfully synthesized by using the substrate bearing an embedded carbon atom. However, such cyclization does not occur when two electron-withdrawing substituents of fluorine appear in the
ortho-substitution of phenyl (
2w), indicating the electronic effect of the substituent on this reaction. Thus, a possible mechanism of the 5-
exo-type annulation was proposed (Fig. S12 in Supporting information), which involves the cage···Ⅰ(Ⅲ) interaction-induced B–H bond activation, the formation of the crucial iodonium intermediate and intramolecular nucleophilic substitution. In contrast to our previous metal-catalyzed annulation to yield 2D-3D fusion [
64–
66], here the B–O closure is achieved by using a cheap organic compound of PhI(OAc)
2 instead of a noble metal catalyst.
The photophysical properties of the cyclized products were investigated for potential applications. To avoid the influence of the phenyl unit,
2s was selected as a model for UV–vis spectral analysis (
Fig. 5a). In contrast to compounds
1 and
1s, the cyclized molecule of
2s shows a new absorption peak around 270 nm, which arises from the charge transfer in the 2D-3D fused structure. The photoluminescence (PL) spectra of
2c, 2g and
2l reveal that the substituent of the new molecular skeleton is able to evidently tune the energy level of the excited state. Specifically, the electron-donating substituents on the phenyl ring leads to a blueshift of the emission (as exemplified by
2c, λ
em = 500 nm) in comparison to
2g (λ
em = 531 nm), whereas the electron-withdrawing substituent induces a redshift of emission (as exemplified by
2l, λ
em = 579 nm) (
Fig. 5b). In addition, the TPE-derived compound
2t not only displays the AIE property [
108,
109] but also exhibits yellow emission in the aggregate state (
Fig. 5c), distinct from the blue emission of TPE. Furthermore, to illustrate the versatility and practicality of our methodology, we synthesized a new type of iridium(Ⅲ) complex (
Ir-2x) by using the carborane-based 2D-3D fused compound
2x as a new ligand. The complex also shows oxygen sensing (
Fig. 5d). Therefore, the molecular design described here significantly broadens the range of carborane-based phosphorescent iridium(Ⅲ) complexes [
110,
111].
We have developed a new boron cluster-based cage···Ⅰ(Ⅲ) interaction between nido-carborane anion and hypervalent iodine cation. This interaction significantly modifies the electronic structure of the boron cluster, thus facilitating selective B–H functionalization to generate a reactive and isolable B–I(Ⅲ) coupled intermediate. Here, the cage···I(Ⅲ) interaction in the intermediate is reminiscent of the B–M bond observed in transition metal-mediated B–H bond activation. Furthermore, the B–I(Ⅲ) containing intermediate is the crucial synthon to be used for construction of B–X (X = N, P, O, S) bonds. More importantly, using this reaction protocol, a new type of carborane-fused hetero-polycyclic compounds featuring a B–O bond has been synthesized. These unprecedented 2D-3D fused structures demonstrate distinct photophysical properties with application potentials.
PDF
Full
75
43
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
