The Versatile Reaction Chemistry of an Alpha-Boryl Diazo Compound

Abstract

The first α-boryl diazo compound that is capable of engaging in classic synthetic organic diazo reaction chemistry is described. The diazomethyl-1,2-azaborine 1, which is a BN isostere of phenyldiazomethane, is significantly more stable than phenyldiazomethane; its reaction chemistry ranges from C–H activation, O–H activation, [3+2] cycloaddition, and halogenation, to Ru-catalyzed carbonyl olefination. The demonstrated broad range of reactivity of diazomethyl-1,2-azaborine 1 makes it an exceptionally versatile synthetic building block for the 1,2-azaborine heterocyclic motif.

Main group element substituted diazo compounds have been well studied and utilized as versatile reagents in organic synthesis. Illustrative examples include trimethylsilyldiazomethane (Si)¹ and Ohira–Bestmann reagent² (P). In contrast, α-boryl diazo compounds are rare, with only about a handful of examples reported to date (Scheme 1, top).³ Lewis acidic boranes have been demonstrated to decompose diazo compounds;⁴ thus all the reported α-boryl diazo compounds either contain quaternized boron or involve trigonal planar boron atoms attached to two π-donating heteroatoms (O, N) to minimize boron’s Lewis acidity (see Scheme 1, top). Despite the broad applicability of diazo compounds as versatile reagents in synthetic organic applications,⁵ no instances of organic reaction chemistry have been described for α-boryl diazo compounds, presumably due to their perceived highly sensitive nature toward reaction conditions for organic synthesis. We envisioned that when the boron atom of an α-boryl diazo species is embedded in an aromatic scaffold such as in diazomethyl-1,2-azaborine 1⁶ (Scheme 1, bottom), sufficient stability could still be achieved to enable diverse reaction chemistry of the diazo functional group. In this communication, we demonstrate that diazomethyl-1,2-azaborine 1 is an exceptional example of the α-boryl diazo family of compounds that is remarkably stable and capable of engaging in a wide range of diazo activation modes, in some cases with distinct reaction selectivities, thus rendering 1 a powerful synthetic building block for 1,2-azaborine chemistry.⁷ In view of emerging applications of BN heterocycles in materials chemistry,⁸ biomedical research,⁹ and organic synthesis,¹⁰ the development of new versatile 1,2-azaborine building blocks is a significant objective.

Scheme 1.

Novel α-Boryl Diazo Compound with Diverse Reaction Chemistry

Diazomethyl-1,2-azaborine 1 can be prepared in a two-step procedure from commercially available starting materials. Treatment of N-TBS-B-Cl-1,2-azaborine with AgOTf furnishes the B-OTf adduct,¹¹ which is more electrophilic toward nucleophilic attack by TMSCHN₂ to generate the target compound 1 with concomitant release of TMSOTf (Scheme 2, top). It is worth noting that the outcome of this process contrasts with the reaction observed between TMSCHN₂ and other B–X-containing compounds, where insertion chemistry with elimination of N₂ and retention of the TMS group is typically observed.^12,13 Diazomethyl-1,2-azaborine 1 is the BN isosteric analogue of phenyldiazomethane; thus comparisons are warranted: (1) Stability: We determined that diazomethyl-1,2-azaborine 1 is a stable oil. When stored at room temperature as a neat compound, no decomposition of 1 has been observed for over a month. In stark contrast, phenyldiazomethane is an unstable and potentially explosive liquid that decomposes readily at room temperature.¹⁴ (2) IR: The presence of the diazo functional group in 1 is indicated by an intense band at ν = 2062 cm⁻¹ for the asymmetric CN₂ stretching vibration. In comparison, the CN₂ vibration for phenyldiazomethane¹⁵ is at 2053 cm⁻¹.¹⁶ The higher CN₂ vibrational frequency observed for 1 in comparison to phenyldiazomethane is consistent with some additional stabilization of the negative charge at the CN₂ carbon position.¹⁷

Scheme 2.

Synthesis and Comparative Characterization of Diazomethyl-1,2-azaborine 1^a

^aReported isolated yields are the average of two runs. TBS: tert-butyldimethylsilyl.

Undaunted by the lack of precedent of synthetic applications of α-boryl diazo compounds, we investigated the reaction chemistry of diazomethyl-1,2-azaborine 1 in the context of diversity-oriented synthesis^18,19 using 1 as a universal building block. The first reaction we examined is the catalytic cross-metathesis of a diazo compound and a carbonyl to form an alkene.²⁰ In the presence of Cl₂Ru(PPh₃)₃ as the catalyst, diazo compound 1 undergoes metathesis with benzaldehyde to furnish BN stilbene derivative 2 with PPh₃ as the stoichiometric reductant (Scheme 3, eq 1).²¹ We were able to isolate the corresponding Ru-α-borylalkylidine complex 3 (Scheme 3, eq 2) and demonstrate its chemical competency as a catalyst for the transformation (Scheme 3, eq 3). Gratifyingly, we are able to grow single crystals of complex 3 from a pentane/dichloromethane solution that are suitable for X-ray diffraction analysis. To the best of our knowledge, complex 3 is the first isolated Ru-α-borylalkylidine complex.^22,23 Not many trans-bistriphenylphosphine dichlororuthenium alkylidene complexes have been crystallographically characterized, and we were unfortunately unable to obtain the structure of the corresponding carbonaceous Ru-benzylidine complex for direct comparison due to its facile decomposition in solution.^24,25 Instead, we used the reported Cl₂(PPh₃)₂Ru-vinylcarbene 4 as the reference compound.²⁶ While most structural parameters are quite similar between 3 and 4, we note a few features for the Ru-α-borylalkylidine 3 that are distinct from the reference complex 4 (Scheme 3, table): (1) The two Ru–P bonds in 3 (2.484 Å, 2.326 Å; Δ = 0.158 Å) are significantly more nonsymmetrical than the ones observed in 4 (2.400 Å, 2.372 Å; Δ = 0.028 Å). (2) The ∠Ru–C1–B angle (124.1°) in 3 is smaller than the corresponding angle in 4 (131.2°). (3) The Ru–C1–B plane in 3 is somewhat perpendicular to the plane of the BN-heterocyclic ring (∠Ru–C1–B–N = 135.8(9)°). This is in contrast to compound 4, where the alkenyl group is more aligned to conjugate with the Ru alkylidine (∠Ru–C1–C–C = −158.6(4)°).

Scheme 3.

Ru-Catalyzed Carbonyl Olefination with Diazomethyl-1,2-azaborine 1 via Ru-α-borylalkylidine Intermediate 3

Next, we investigated the Cu-catalyzed C–H activation of terminal alkynes in the presence of the α-boryl diazo compound 1. Fu et al. reported a simple CuI-catalyzed procedure that couples terminal alkynes with diazoesters and diazoamides to furnish 3-alkynoates (Scheme 4, eq 4).²⁷ In a number of cases, the corresponding allene isomer was observed as a minor byproduct. Fox and co-workers subsequently optimized the conditions for allene formation for α-alkyl-α-diazoesters with terminal alkynes and determined that the initially formed alkynoate intermediates isomerize to the allenoate products in the presence of the stoichiometrically employed K₂CO₃ base (Scheme 4, eq 5).²⁸ Thus, it appears that the alkynoate species is the kinetically preferred product under Cu catalysis that subsequently can isomerize to the allenoates in the presence of a base. Surprisingly, when we adapted the protocol developed by Fu and treated diazomethyl-1,2-azaborine 1 with terminal alkynes in the presence of catalytic amounts of CuI in MeCN, and in the absence of any base, the corresponding allene 5 is formed exclusively instead of the expected internal alkyne (Scheme 4, eq 6).

Scheme 4.

Cu-Catalyzed C–H Activation of Terminal Alkynes with Diazomethyl-1,2-azaborine^a

^aYields are the average of two runs.

The formation of allene 5 is indicated by the two diastereotopic Si–Me signals of the N-tert-butyldimethylsilyl (TBS) group in the ¹H NMR, which is consistent with the existence of an axially chiral stereogenic unit. The distinct product selectivity of CuI-catalyzed alkyne–diazo coupling for α-boryl diazo compound 1 again highlights its unique electronic structure that is likely underlying the observed selectivity.²⁹

We also note that in the absence of a viable external C–H source and in the presence of the Rh₂OAc₄ catalyst an intramolecular C–H insertion occurs at the methyl or t-Bu group (on the N-TBS moiety), respectively, yielding the corresponding bicyclic 1,2-azaborine compounds 6a and 6b with a strong preference for the formation of the fivemembered 6a (eq 7).³⁰

In addition to C–H activation chemistry, diazo compounds are known to engage in uncatalyzed O–H insertion chemistry.³¹ This reactivity feature has been utilized in chemical biology applications, where stabilized diazo compounds have been developed to label proteins and nucleic acids via O-alkylation.³² For an α-boryl diazo compound the high oxophilicity of boron may present an additional challenge to successfully O-alkylate suitable oxygen-based nucleophiles such as carboxylic acids. Gratifyingly, we determined that diazomethyl-1,2-azaborine 1 reacts quite rapidly at room temperature with carboxylic acids to form esters in acetonitrile in the absence of a catalyst. As can be seen from Scheme 5, aryl and alkyl carboxylic acids are suitable substrates. An electrophilic alkyl bromide functional group is also tolerated (Scheme 5, entry 7c).

Scheme 5.

O-Alkylation of Carboxylic Acids with Diazomethyl-1,2-azaborine^a

^aReported isolated yields are the average of two runs.

To explore additional reactivity types that diazo compounds can engage in, we investigated the [3+2] cycloaddition of diazomethyl-1,2-azaborine 1 with alkenes. A diazo compound is typically considered electron-rich and thus reacts as a nucleophile in normal-electron-demand cycloadditions with electron-deficient dipolarophiles.³³ We are pleased to find that diazomethyl-1,2-azaborine 1 undergoes regioselective [3+2] cycloaddition reactions with α,β-unsaturated esters, ketones, and nitriles in the presence of a Pd catalyst to presumably initially form intermediates 8′, which then subsequently undergoes a formal 1,3-boryl shift to furnish adducts 8 (Scheme 6). In contrast, phenyldiazomethane preferentially forms cyclopropane compounds with elimination of N₂ (see Supporting Information for experimental details). As an additional comparison, while phenyldiazomethane reacts relatively cleanly with styrene to furnish 1,2-diphenylcyclopropane in the presence of cobalt(II) tetraphenylporphyrin,³⁴ diazomethyl-1,2-azaborine 1 remains mostly unreactive, producing minor amounts of the intramolecular C–H activation product 6a (see Supporting Information for experimental details).

Scheme 6.

[3+2] Cycloaddition Reaction with Diazomethyl-1,2-azaborine^a

^aReported isolated yields are the average of two runs.

Finally, we envisioned that diazomethyl-1,2-azaborine 1 can serve as a precursor to the halomethyl building blocks 9 and 10 (Scheme 7). Benzyl halides are versatile intermediates in diversity-oriented synthesis in medicinal chemistry,³⁵ and compounds 9 and 10 represent the direct BN isosteres of benzylic halides. While examples of halomethyl BN naphthalenes have been reported,³⁶ to the best of our knowledge, such a building block for the monocyclic 1,2-azaborine heterocycle has not been prepared. Importantly, the annulation method involving potassium chloromethyltrifluoroborate and 2-aminostyrene to yield the chloromethyl BN naphthalenes³⁶ is not applicable to the monocyclic benzene-type BN heterocycle. We envisioned that similar to how O-alkylation proceeds with protonation of the α-boryl carbon followed by nucleophilic attack of the oxygen nucleophile,³¹ treatment of 1 with a suitably matched proton/halide source should yield the corresponding halomethyl 1,2-azaborines 9 and 10.³⁷ After screening a number of conditions, we determined that 1-ethynyl-4-nitrobenzene as the proton source in combination with CuI as the iodide nucleophile works well for the iodination of diazomethyl-1,2-azaborine 1 (Scheme 7, eq 8). Similarly, the combination of bromoacetic acid as the proton source and cetrimonium (hexadecyltrimethylammonium) bromide is suitable for converting 1 to the bromomethyl 1,2-azaborine 10 (Scheme 7, eq 9).³⁸

Scheme 7.

Synthesis of Halomethyl 1,2-Azaborine Building Blocks from 1^a

^aReported isolated yields are the average of two runs.

We chose bromomethyl 1,2-azaborine 10 to demonstrate the ability of halomethyl 1,2-azaborines to serve as an electrophilic building block in a variety of coupling and substitution reactions.^36,39 As can be seen from Scheme 8, compound 10 can engage in Suzuki–Miyaura cross-coupling reaction to produce BN diarylmethanes 11. Compound 10 also reacts with a number of nucleophiles such as thiocyanate, azide, amine, and phthalimide anion to furnish the corresponding adducts 12–15 in moderate to good yields, further demonstrating the diversity of compounds that can be accessed with diazomethyl-1,2-azaborine 1 as the universal precursor.

Scheme 8.

Bromomethyl 1,2-Azaborine 10 as a Versatile Electrophilic Building Block^a

^aReported isolated yields are the average of two runs.

In summary, we synthesized the first α-boryl diazo compound that is capable of engaging in classic synthetic organic diazo reaction chemistry, including C–H activation, O–H activation, [3+2] cycloaddition, halogenation, and Rucatalyzed carbonyl olefination. Furthermore, we showed that diazomethyl-1,2-azaborine 1, a BN isostere of phenyldiazomethane, is a stable compound and that its corresponding Ru carbene complex 3 exhibits bonding features that are distinct from an analogous Ru-alkylidine complex. Overall, we believe that diazomethyl-1,2-azaborine 1 as a new member of the α-boryldiazo compound family and as a remarkably versatile 1,2-azaborine building block will advance both the basic science of diazo chemistry and the multifaceted chemistry of 1,2-azaborine heterocycles.