1
Introduction
Yoichi Hoshimoto received his M.Sc. and Ph.D. from Osaka University in 2013. He then joined the Frontier Research Base for Global Young Researchers, Osaka University as a tenure-track assistant professor (2013–2018). Subsequently, he was promoted to associate professor in the Department of Applied Chemistry, Faculty of Engineering, Osaka University (2018). Since 2023, he has been recognized as an Outstanding Young Researcher in the Center for Future Innovation (CFi), Faculty of Engineering, Osaka University. He is one of the Thieme Chemistry Journals Awardees 2024.
Triarylboranes (BAr3), including both homoleptic and heteroleptic species, are typical Lewis acids that are widely used as catalysts, activators, sensors, and bio-imaging agents.[1] In the field of main group catalysis in particular, recent progress in the field of frustrated Lewis pairs (FLPs)[2] has led to a significant structural diversification of halogenated triarylboranes beyond the archetypical B(C6F5)3. This structural diversification can be achieved by the introduction of substituents to the Ar groups at the ortho-, meta-, and/or para-positions, with respect to the boron center, as part of a strategy to control the Lewis acidity of BAr3.[3]
[4] These strategies to control the Lewis acidity focus on regulating the accessibility (a kinetic aspect) and energy (a thermodynamic aspect) of the empty p orbital at the boron center.
Strategies that substitute meta-F and/or para-F atoms in B(C6F5)3 with more or less electron-withdrawing substituents have been applied to prepare more or less electrophilic BAr3 derivatives through regulation of the electron affinity at the boron center.[1a]
[b]
[3]
[5] Strategies that regulate the steric repulsion between a Lewis base (LB) counterpart (front strain; Figure [1], left) through modulation of the size of the ortho-substituents have also been widely explored.[6] Alternatively, the Lewis acidity of BAr3 can be modulated by regulating the intramolecular repulsion between the Ar groups of tetrahedral LB–borane adducts (back strain; Figure [1], right). In this context, the classical concept of back strain refers to the repulsion between ortho-substituents.[7] Moreover, Hoshimoto and co-workers recently discussed the concept of ‘remote’ back strain based on the repulsion and supportive non-covalent interactions (NCIs) between meta-substituents themselves and/or between meta-substituents and substituents in the LBs.[8]
[9]
[10] During our efforts to develop an effective method to finely tune the (catalytic) reactivity of BAr3, we discovered that only a limited number of substituents had been introduced at the meta-positions for the derivatization of BAr3.
Figure 1 A schematic illustration of front strain and (remote) back strain generated between BAr3 and Lewis bases (LBs).
Hence, a review that summarizes the structure and use of meta-substituted BAr3 will be a worthwhile addition to previously reported reviews that have predominantly focused on the derivatization of BAr3 via substitution at the ortho- and para-positions.[1a]
[b] Moreover, we previously confirmed that over 80% of ortho- and meta-substituents in BAr3, along with 50% of para-substituents, consist of H, F, and CH3 groups, based on our analysis of the 98 homoleptic BAr3 compounds synthesized up to and including 2020 (as found using SciFinder in February 2024) (Figure [2]). In particular, F or CH3
ortho-disubstituted compounds were found to have been frequently explored for the kinetic protection of the vacant p orbital on boron by regulation of the front strain. Conversely, the impact of meta-disubstitution on the Lewis acidity of BAr3 compounds has been less well studied, probably due to the limited number of synthetic routes to meta-F2- or meta-(CH3)2-substituted species.[9]
[11] This Short Review thus aims to summarize the structures and applications of meta-substituted homoleptic and heteroleptic BAr3 species and to shed light on the importance of such structural modifications in the context of regulating the reactivity of BAr3 compounds.
Figure 2 Analysis of the substituents introduced at the ortho-, meta-, and para-positions in 98 homoleptic BAr3 species (as found using SciFinder in February 2024); the relative ratio of each substituent type is given, e.g., for a BAr3 including a 2,4,6-trimethyl group, two ortho- and one para-CH3 groups are counted.
2
Scope of this Review
This Short Review analyzes both homoleptic and heteroleptic BAr3 compounds whose synthesis had been reported up until February 2024. BAr3 species that have merely been explored theoretically are not included. In addition, boranes that have meta-H, -F, and -CH3 substituents are beyond the scope of this Short Review , even though the corresponding homoleptic species were considered when Figure [2] was prepared, to avoid significant overlap with previous reports.[1a]
[b] Furthermore, several BAr3 compounds with 2,6-dimethylaryl groups (e.g., mesityl groups) and ortho-bridged planar structures have also been omitted, given that these compounds have already been summarized in other critical reviews.[1c]
,
[12]
[13]
[14]
2.1
The Electronic and Steric Influence of meta-Substituents
A fundamental and highly effective approach to modulate the Lewis acidity of BAr3 is to regulate the energy levels of the unoccupied p orbital on the boron atom, i.e., modulating the intrinsic electrophilicity. Unsurprisingly, chemists have explored the introduction of (strongly) electron-donating or -withdrawing substituents at the meta- and/or para-positions of the Ar groups. The introduction of the strongly electron-withdrawing CF3 group in 3,5-(CF3)2C6H3 (ArF) has stimulated the curiosity of many chemists.[15]
[16]
[17]
[18]
[19]
[20] In 2012, Ashley and co-workers demonstrated the synthesis of BArF
3 (B1
) on a practical scale via a reaction between a Grignard reagent including the ArF ligand and BF3·OEt2.[15a] As expected, the Lewis acidity of B1
was confirmed experimentally, using the Gutmann–Beckett method and Et3P=O as a probe for a 31P NMR analysis, to be higher than that of B(C6F5)3. Conversely, when trans-crotonaldehyde was employed as a probe for the 1H NMR analysis (Child’s method), B1
was found to exhibit lower Lewis acidity than B(C6F5)3. Ashley and co-workers also explored the reactivity of an FLP comprised of B1
and 2,2,6,6-tetramethylpiperidine (TMP) with H2, which afforded the salt [HTMP][μ-H(B1
)2] (Scheme [1a]).[15a] In contrast, Blagg, Lawrence, and Wildgoose reported that the heterolytic cleavage of H2 did not occur when FLPs of TMP and B(2,5-(CF3)2C6H3)3 (B2
) or B(2,4-(CF3)2C6H3)3 (B3
) were employed.[17a] The authors also rationalized the increased electrophilicity of B1
relative to B2
or B3
based on the analysis of the formal reduction potentials of all three compounds. Using B1
in another catalytic system, Gagné and co-workers showed that different products were obtained when a specific combination of B1
/EtMe2SiH or B(C6F5)3/Et2MeSiH was used in the catalytic reduction of natamycin (Scheme [1b]).[18`]
[b]
[c]
[d]
Scheme 1 (a) Heterolytic cleavage of H2 via the combination of TMP and B1
–B3
, with their formal reduction potentials (vs. [FeCp2]0/+ [V]) and (b) the reduction of a bioactive compound using B1
or B(C6F5)3.
Pápai, Soós, and co-workers also investigated the electronic effects of meta-substituents through a comparison of the Lewis acidity of a series of heteroleptic BAr3 compounds (B4
–B15
) based on their hydride ion affinity (HIA) and the Gutmann–Beckett method (Figure [3]).[11] These authors showed that the replacement of the meta-H atom with an F atom significantly enhances the Lewis acidity, whereas a replacement of a meta-Cl atom in either the Mes or 2,6-Cl2-aryl moieties results in negligible changes.
Figure 3 Comparison of the theoretical HIA (kcal mol–1) and relative Lewis acidity (%LA) of B4
–B15
; %LA values were determined using the Gutmann–Beckett method with Et3P=O and are calculated with respect to B(C6F5)3 (%LA = 100).
The steric effects imparted by meta-substituents have been examined from three main perspectives: (1) the buttressing effect; (2) London dispersion forces; and (3) remote back strain as a sum of the electronic/steric repulsion and NCIs. For example, Wada and co-workers synthesized B(3-Br-2,6-(MeO)2C6H2)3 (B16
) through the direct bromination of air-stable B(2,6-(MeO)2C6H2)3 with N-bromosuccinimide (NBS) (Scheme [2]).[21] While B(2,6-(MeO)2C6H2)3 forms isolable adducts with primary amines and ammonia, B16
does not form such adducts with the same amines. The authors attributed this reactivity difference to the buttressing effect caused by the meta-Br atoms, i.e., the Br atoms push the adjacent MeO groups closer to the boron atom thus hindering access of the amines to the boron center. The electron-withdrawing nature of the Br atoms was not considered in this case.
Scheme 2 Synthesis of B16
.
Slootweg and co-workers recently proposed that London dispersion forces play a critical role in stabilizing encounter complexes formed between BAr3 and LBs (Scheme [3a]).[22a] The authors observed the formation of 1:1 co-crystals comprised of B(3,5-
t
Bu2C6H3)3 (B17
) and N(3,5-
t
Bu2C6H3)3. Based on theoretical calculations, the authors concluded that interaction energies between the meta-
t
Bu groups in B17
and those in N(3,5-
t
Bu2C6H3)3 are significantly larger than the corresponding energies formed between B1
and N(3,5-
t
Bu2C6H3)3. Hansen, Paradies, and co-workers further studied the importance of dispersion forces based on a combined experimental and theoretical approach.[22b] These authors expanded the discussion to include the B(3,5-R2C6H3)3 (R =
t
Bu, B17
; Me, B18
) and P(3,5-R′2C6H3)3 (R′ =
t
Bu,
i
Pr, Me) pairs and concluded that the Lewis acid/base adducts generally become more stable as the size of the dispersion-energy donor increases, albeit that their stability is sensitive to the solvation conditions (Scheme [3b]). In this study, the reactivity of B(3,5-
i
Pr2C6H3)3 was not explored due to difficulties associated with synthesis and stability.
Scheme 3 London dispersion energies stabilize (a) the amine-borane encounter complex involving B17
, and (b) the phosphine-borane adducts involving B17
and B18
; the reported P–B bond lengths (Å) and association energies of the Lewis pairs (ΔG
exp in kcal mol–1) are shown.
Scheme 4 Comparison of theoretical parameters for the formation of Et3P=O–B
n
(n = 19–21) adducts, i.e. relative Gibbs energies (ΔG° in kcal mol–1 with respect to [B
n
+ Et3P=O]), and deformation energies (E
DEF in kcal mol–1). The chemical shifts in 31P NMR, δP, obtained in the reaction between B19
–B21
and Et3P=O (0.3 eq.) in CH2Cl2 are also given.
From a different perspective, Hoshimoto and co-workers also explored how the physical and electronic properties of meta-substituents increase/decrease the stability of borane-LB adducts, and hence, prevent/promote the dissociation of LBs from the adducts to generate free BAr3 (or FLP species) (Scheme [4]).[8]
[9]
[10] The authors focused on the concept of ‘remote’ back strain (Figure [1b]), which is defined as the sum of repulsive (steric/electronic) and attractive (NCIs) interactions that occur between the meta-substituents of the tetrahedral BAr3 units in the borane-LB adducts (or a transition state of the borane deformation). In general, these meta-substituents must be effectively separated in the free trigonal planar BAr3 structure.[9] To quantify the remote back strain, homoleptic boranes bearing 2,6-F2-3,5-allyl2C6H3 (B20
) and 2,6-F2-3,5-TMS2C6H3 (B21
) were synthesized for a comparison with B(2,6-F2C6H3)3 (B19
), as these boranes exhibit nearly identical intrinsic Lewis acidity (the LUMO energy levels) and front strain toward LBs. The relative Gibbs energy values (ΔG°) for the formation of an adduct with a LB (the LBs used in the corresponding work were Et3P=O, H2O, CO, THF, and NMe3) and the deformation energy (E
DEF),[3a] which is an energetic penalty paid for the geometrical change at the boron center upon adduct formation, were evaluated for each of these three boranes. For example, when Et3P=O was used as the LB, the E
DEF values increased in the order B19
< B21
< B20
, which is consistent with the trend determined via the Gutmann–Beckett method. However, the ΔG° values showed a different trend and increased in the order B21
< B19
< B20
. This discrepancy was rationalized by considering the multiple NCIs formed between the meta-TMS groups themselves and the meta-TMS and P-Et groups in the Et3P=O–B21
adducts. Finally, the authors concluded that repulsion and the NCIs generated between meta-substituents are essential for estimating and regulating the remote back strain for fine-tuning the catalytic activity of BAr3. It should also be noted that B20
is a rare example of a liquid BAr3.
In another report, meta-substituents were used as a tool for monitoring the stereoisomerization of BAr3. Mislow and co-workers demonstrated that the stereoisomerization in B(3-
i
Pr-2,4,6-Me3C6H)(2,6-Me2C6H3)2 proceeds via a two-ring flip mechanism, as confirmed by temperature-dependent 1H NMR spectroscopy based on the diastereotopic nature of the meta-
i
Pr group.[23]
2.2
Molecular Transformations Mediated by meta-Substituted Boranes
Some meta-substituted BAr3 compounds have been applied as aryl-transfer reagents. Frohn and co-workers studied the migration of an aryl group from an electrophilic BAr3 species, such as B(C6F5)3 and B(3-CF3C6H4)3 (B22
), to XeF2 in CH2Cl2, eventually affording [ArXe][ArBF3] (Ar = C6F5 or 3-CF3C6H4) (Scheme [5a]).[24] In 2019, Melen, Wirth, and co-workers developed an aryl-transfer reaction from BAr3 to various α-aryl-α-diazoacetates (Scheme [5b]).[25] They found that the number of aryl groups that are transferred depends on the Lewis acidity of BAr3, e.g., B(C6F5)3 and B(3,4,5-F3C6H2)3 can transfer all three Ar groups, whereas B(3,4-Cl2C6H3)3 (B23
) can only transfer one of its three 3,4-Cl2C6H3 groups. Ishida, Iwamoto, and co-workers also reported on the migration of aryl groups from BAr3, including B(C6F5)3, B1
, and BAr*3 (Ar* = 3,5-
t
Bu2-4-MeOC6H2; B24
), to a dialkylsilanone, which was proposed to proceed via the formation of a Si=O–B adduct (Scheme [5c]).[26] This report nicely demonstrates that BAr3 can enhance the reactivity of an unsaturated silicon center upon adduct formation, which is often seen in the activation of carbonyl compounds by Lewis acids.
Scheme 5 Aryl-transfer reactions from BAr3 to (a) XeF2, (b) a diazo ester, and (c) a dialkylsilanone.
Chiu and co-workers reported heteroleptic BAr3 species B25
and B26
, which bear para-OH groups (Scheme [6]).[27] Trivalent boron compounds such as these contain Lewis basic and Brønsted acidic functional groups and it is not always facile for these moieties to co-exist within the same compound. Interestingly,
t
Bu groups were introduced at the meta-positions of bulky 1,3,5-R3C6H2 (B25
: R = Me (Mes); B26
: R =
t
Bu (Mes*)) group at the boron center. Moreover, these authors explored the oxidation of these boranes and confirmed that boryl analogues of the galvinoxy radical, such as [B25
]•–
and [B26
]•–
, were generated. The introduction of boron decreases the quinoidal character of the phenoxyl radical and activates the open-shell species.
Scheme 6 Oxidation of B25
and B26
(Mes* = 1,3,5-
t
Bu3C6H2).
Bourissou and co-workers reported the preparation of the ortho-phenylene bridged phosphine-borane compound B27
, which contains two ArF groups at the boron center (Scheme [7]).[28] Based on NMR, single-crystal X-ray diffraction, and DFT analyses, an intramolecular coordination interaction between the phosphorus and the boron centers was proposed. The catalytic reactivity of B27
was evaluated in the dehydrogenation of acyclic and cyclic (di)amine-boranes.
Scheme 7 Dehydrogenation of cyclic amine-boranes catalyzed by B27
.
One of the most extensively explored fields in main group catalysis is the reduction of unsaturated compounds with BAr3 catalysts.[2]
[29] Needless to say, meta-substituted BAr3 have also contributed to the significant development of this important field. For example, Pápai, Soós, and co-workers reported the catalytic hydrogenation of aldehydes, ketones, and enones using H2 and B12
–B15
.[30]
The use of hydrosilanes as a reductant is another practical example of the main-group-catalyzed reduction of unsaturated compounds. For example, Ingleson and co-workers pioneered the development of a main-group-catalyzed reductive alkylation of amines with carbonyl compounds and hydrosilanes in the presence of B(C6F5)3, BPh3, or B(3,5-Cl2C6H3)3 (B28
).[31] In particular, the in situ generation of B28
from Na[B(3,5-Cl2C6H3)4] enabled the transformation of primary amines whose conjugate acids span pK
a values of 10.6 to 18.5 in MeCN (Scheme [8]).[31a]
Scheme 8 Reductive alkylation of amines catalyzed by B28
, generated from Na[B(3,5-Cl2C6H3)4] in situ.
Between 2017 and 2018, Hoshimoto and co-workers and Soós and co-workers independently reported the BAr3-catalyzed reductive alkylation of amines with aldehydes and H2, where H2O is generated as the sole byproduct.[7]
[32] In the former case, a catalyst-controlled reaction system that generates an active FLP species comprising B13
and THF was extensively applied to the reductive alkylation of multisubstituted aniline derivatives. However, the direct use of amino acids was still found to be challenging even under harsh reaction conditions.[32] Meanwhile, Soós and co-workers constructed a substrate-controlled system, that furnishes an FLP from B(2-Cl-6-FC6H3)(2,6-Cl2C6H3)2 and in situ generated imine intermediates, which was predominantly applied to the functionalization of N-alkyl amines.[7] To further expand the utility of such BAr3-catalyzed reductive functionalization methods using H2, Hoshimoto and co-workers recently demonstrated an in silico assisted strategy to significantly shorten the lengthy trial-and-error processes usually used for the optimization of BAr3 (Scheme [9]).[33] In this study, B29
–B38
were prepared for the construction of an in silico library of BAr3 for the collection of the experimental parameters required for machine learning. Eventually, the optimal reaction system was discovered to be B34
and 4-methyltetrahydropyran (MTHP) and this was successfully applied to the reductive alkylation of aniline-based amino acids and C-terminal-protected peptides.
Scheme 9 (a) Heteroleptic BAr3 species B29
–B38
designed for the machine-learning-assisted optimization of borane catalysts and (b) B34
-catalyzed reductive alkylation of amino acids and peptides using H2. a 60 atm H2; MTHP = 4-methyltetrahydropyran.
Recently, Hoshimoto and co-workers also disclosed a conceptually novel approach for the direct use of ‘crude’ H2 (a gaseous mixture of H2, CO, CO2, and/or CH4) for the catalytic hydrogenation of unsaturated molecules (Scheme [10a]).[8]
[9]
[10] Given that a huge amount of H2 will be produced through the production of crude H2 from hydrocarbon resources (i.e., natural gas, biomass, or food waste), the development of a technology bypassing the energy- and cost-intensive multistep purification processes of crude H2 will be valuable.[34] The authors found that the hydrogenation of 2-methylquinoline (MeQin), used as a model liquid organic hydrogen carrier (LOHC), proceeded in the presence of 0.1 mol% of B(2,6-Cl2C6H3)(2,6-F2-3,5-X2C6H)2 [X = F (B13
), Cl (B39
), Br (B40
), and ArF (B41
)] under solvent-free conditions.
Scheme 10 (a) Simplified schemes of typical contemporary routes for H2 purification for the hydrogenation of unsaturated compounds; (b) direct use of crude H2 for the catalytic hydrogenation of MeQin using B13
and B39
–B41
; and (c) direct use of crude H2 for the catalytic hydrogenation of 1-naphthaldehyde using B42
–B44
. a 1,4-Dioxane used as the solvent.
The catalyst turnover number (TON) increased in the order B13
(1000) < B41
(1340) < B39
(1400) < B40
(1520) (Scheme [10b]).[8] It should be noted here that the intrinsic Lewis acidity of B13
, B40
, and B41
is nearly identical, and thus, such a significant difference in TON should be attributed to the size of the meta-substituents (e.g., the degree of their remote back strain). B40
and B41
were also applied to the catalytic hydrogenation of 2,6-lutidine in the presence of gaseous mixtures of H2/CO (40/4 atom each) and H2/CO2 (40/4 atom each) to afford 2,6-dimethylpiperidine. Finally, by taking advantage of the catalytic activity of B41
in the dehydrogenation of 2-methyl-1,2,3,4-tetrahydroquinoline (H4-MeQin), the authors demonstrated a molecular-based H2 purification via the hydrogenation of MeQin with crude H2 and the subsequent dehydrogenation of H4-MeQin to afford highly pure H2. Subsequently, Hoshimoto and co-workers also demonstrated that B20
could be successfully applied to the catalytic hydrogenation of unsubstituted quinoline under mixed gas (H2/CO/CO2) conditions.[9]
The aforementioned approach for the direct use of crude H2 has recently been expanded to the catalytic hydrogenation of carbonyl compounds. In this case, a BAr3 compound of the type, B(2,6-F2-3,5-X2C6H)3, was used, and the alcohol yield was found to increase when the meta-substituents were changed from X = F (B42
) to Cl (B43
) to Br (B44
) (Scheme [10c]).[10] Notably, this trend in reaction efficiency is consistent with the increase in the E
DEF values. Therefore, the increased remote back strain seems to provide higher reaction efficiency by preventing the formation of an adduct with the LBs involved in the system. Importantly, the formyl groups in the aromatic and aliphatic aldehydes that also contain halogen and olefinic substituents can be selectively hydrogenated under mixed gas conditions. Furthermore, B44
catalyzed the hydrogenation of undec-10-enal in the presence of a gaseous mixture of H2, CO, CO2, and CH4 (76/0.2/20/3.2 molar ratio), which is produced from CH4 via desulfurization, stream reforming, and CO-shift conversion processes in industry. These examples showcase (i) the power that can be extracted from BAr3 catalysis via the fine-tuning of their Lewis acidity based on meta-substitution and (ii) the advantages of BAr3 relative to the simple alternative of transition metal catalysts that require the use of purified H2.
Figure 4
meta-Substituted BAr3
B45
–B48
with 2,4,6-Me3-aryl units.
2.3
Other Examples of meta-Functionalization of BAr3
The introduction of Mes groups into BAr3 can significantly increase their stability toward LBs such as H2O. For example, in 1960, B(2,4,6-Me3-3,5-(NO2)2C6)3 (B45
) was prepared by treatment of B(Mes)3 with a mixed acid under cryogenic conditions (Figure [4]).[35] Later, Ashley, Wildgoose, Slootweg, and co-workers used B45
to accomplish the homolytic H2 cleavage through a one-electron reduction in the absence of an external Lewis base.[36] In 1981, Wilson and co-workers reported the synthesis of B(3-MeOC6H4)(2,4,6-Me3C6H2)2 (B46
) and B(3-ClC6H4)(2,4,6-Me3C6H2)2 (B47
).[37] More recently, Ito and co-workers demonstrated the synthesis of B(3-Br-4-MeC6H3)(2,4,6-Me3C6H2)2 (B48
) via the reaction between 2,4-dibromo-1-methylbenzene and Ph2MeSi–BMes2 in the presence of Na(O
t
Bu).[38]
