Compounds of antimony (Sb) have recently received increased attention and development as multifaceted ligands for transition metals with direct metal-Sb bonding [1][2][3][4][5],as remote, redoxactive modifiers of classical ligand types [6], and anion sensing [7][8][9][10][11][12].In particular, we were intrigued by the two examples developed by the by the Gabbaï group for fluoride capture studies (A and B, Figure 1) [13,14]. These molecules contain two stiborane units that cooperate in capturing the fluoride anion. The stiboranes were obtained via oxidation of the trivalent antimony precursors with o-chloranil (3,4,5,6-tetracholoro-o-quinone). Compound B was found to be superior to A, ostensibly because the central oxygen in A can provide only a repulsive interaction with fluoride, whereas the central C-H in B can provide an attractive C-H…F hydrogen bonding interaction. Because of our past studies with diarylamine-centered pincer ligands [15], we surmised that a bis(stiborane) molecule analogous to A and B, but with a central NH unit might possess even more affinity for capturing fluoride. The Gabbai group reported a relevant compound C in 2023, but did not explore its oxidation with o-chloranil [16]. We recently reported transition metal complexes (e.g., E) of a PNSb ligand D [17]. In this work, we endeavored to synthesize an "SbNSb" analog in order to oxidize to a bis(stiborane) by the Gabbai o-chloranil method and assess its fluoride affinity. Although this synthetic approach was successful, the targeted bis(stiborane) underwent an irreversible isomerization that prevented us from studying its fluoride affinity and which we present in this report.
Treatment of 1 with three equivalents of butyllithium, followed by addition of two equivalents Ph2SbCl and an aqueous workup, led to the isolation of 2 in 31% yield as a free-flowing, white solid (Scheme 1). 1 H and 13 C{ 1 H} NMR analysis was consistent with the expected C2v symmetry.
The presence of the N-H moiety was confirmed by the observation of an 1 H NMR resonance at δ 5.68 ppm in C6D6 and of the N-H stretch by IR spectroscopy at 3338 cm -1 .
Treatment of 2 with o-chloranil in toluene caused an initial color change of the colorless solution to orange, then finally to yellow within 5 minutes. After removal of volatiles and washing with pentane, 3a was isolated as a pale-yellow solid in 83% yield. Characterization by 1 H NMR spectroscopy in C6D6 revealed a downfield shift of the N-H resonance from δ 5.68 ppm to 9.03 ppm, along all other resonances under the expected C2v symmetry. It was noted that longer reaction times (in toluene solvent) led to the formation of precipitate which we later determined to be an isomerization product. In addition, the use of chloranil that had been stored at ambient temperature for an extended period of time appeared to lead to a faster, unavoidable formation of this new product. We were able to prepare it reproducibly via treatment of 2 with o-chloranil in THF for 16 h, resulting in a 92% yield after workup. This new compound was identified as 3b, a cyclized and ionized isomer of 3a. 1 H NMR analysis in DMSO-d6 revealed the presence of a downfield N-H resonance at δ 9.79, several much sharper Ph resonances, and a single resonance for the Ar-CH3 hydrogens.
Previous examples of ionizing rearrangements of catecholate-containing stiboranes to stibonium antimonate salts.
Compounds 3a and 3b can be observed as distinct, simultaneously present species by 1 H NMR spectroscopy in a variety of solvents (vide infra) at ambient temperature. In addition, a variable temperature NMR study of 3b detected no changes at 0, -20, or -40 °C compared to ambient temperature. These observations rule out the possibility that 3a and 3b are in rapid equilibrium on the NMR timescale, and thus the isomerization of 3a into 3b is irreversible. The observed redistribution of Sb substituents represents a non-redox disproportionation of a five-coordinate, pentavalent Sb with one bidentate catecholate and three aryl substituents into a tetraarylstibonium with a four-coordinate Sb, and an antimonate with a six-coordinate Sb. Analogous ligand redistribution reactions of related Sb(V) tris(aryl)/catecholates have been previously observed to occur (Scheme 2) [18,19]. In those examples, the isomerization was favorable in the more polar solvents (chloroform or acetone), but reversible in the less polar toluene. In our case, the isomerization is likely more favorable because of entropic reasons (a single molecule of 3a contains both five-coordinate Sb centers as opposed to two molecules of F or H engaged in the isomerization).
To understand whether the isomerization rate was affected by the solvent nature, 3a was dissolved in a series of solvents and its isomerization into 3b was tracked via 1 H NMR spectroscopy. The isomerization reactions appeared to follow a first-order decay (Figure 2) and the relative rates were in the following order: THF-d8 > CD2Cl2 > CDCl3 ≈ C6D6 > DMSO-d6. The isomerization was essentially complete in THF in <1 d and in CD2Cl2 in ca. 3 d, but was not complete in the other solvents even after a week. These findings do not allow to make a generalized interpretation based on either the solvent polarity or the coordinating ability of the solvent: DMSO and THF would be the two most polar and most coordinating, but they are at the opposite ends of the rate range. In addition, we cannot exclude that the transformation is catalyzed by adventitious impurities, such as water, or traces of an additional Sb Lewis acidic species present.
Because of the easily occurring rearrangement to 3b, we did not pursue the studies of fluoride capture with 3a. However, we wish to investigate the fluoride affinity of the cation in 3b. In order to exchange the anion in 3b, it was first treated with one equivalent of CsF in THF causing an immediate change to a clear, homogeneous solution. After the removal of volatiles, and extraction with PhF, 4 was isolated in 84% yield as a free-flowing white powder (Scheme 1). Given the high isolated yield of 4, we assume by stoichiometry that the by-product of its formation is the Cs salt of the antimonate anion in 3b, but we made no effort to characterize it. The 19 F NMR spectrum in CDCl3 featured a single resonance at δ -100.4 ppm, confirming the presence of the Sb-F bond. The N-H proton resonated at 6.68 ppm in CDCl3 and at 8.31 ppm in THF-d8. The solid-state structure of 4 (vide infra) is dissymmetric, but the 1 H and 13 C{ 1 H} NMR spectra of 4 in THF-d8 were consistent with a time-averaged C2v symmetry. One of the 1 H NMR aromatic resonances, however, was distinctly broad. In CDCl3, this 1 H resonance was missing, presumably broadened beyond detection, and the 13 C{ 1 H} spectrum contained broader resonances, as well. These observations indicate that the symmetrizing fluxionality in 4 is faster in THF than in chloroform.
Reaction of 4 with Me3SiOTf led to the release of an equivalent of free Me3SiF and formation of 5, which was characterized by multinuclear spectroscopy as well as X-ray diffractometry. 1 H NMR observations support the identity of the cation with a downfield N-H resonance at δ 9.55 ppm (in CDCl3), along with all expected aromatic and benzylic (δ 2.22 ppm, 6H) resonances. 5 exhibits a singlet in the 19 F NMR spectrum at δ -76.9 ppm corresponding to the triflate anion. Structural analysis revealed a hydrogen bonding interaction between the N-H and an oxygen atom of the triflate anion (N--O distance of ca. 2.91 Å) (Figure 4).
utilized by the Gabbai group in order to determine the equilibrium fluoride binding constant, Kf, of various antimony-based Lewis acids [20,21]. Spectrophotometric data collected from the titration of a 3.2 x 10 -5 M THF solution of 5 with a 0.01 M THF solution of [ n Bu4N][F] (Figure 3) could be accurately fitted to a 1:1 binding isotherm showing formation of the fluoride adduct, 4 (See supporting information). Compound 5 was determined to have a large fluoride binding constant of Kf > 10 7 , which is consistent with other tetraarylstibonium species in organic media [22]. M tetrabutylammonium fluoride (TBAF) in THF. Sb1-C1, 2.160(3); Sb1-C8, 2.106(3); N1-C6, 1.390(5); N1-C9, 1.383(4). Selected distances in 5 (Å): Sb1-C1, 2.093(5); Sb1-C12, 2.094(5); N1-C6, 1.376(7); N1-C7, 1.392(7); O2-H1, 2.039(14).
The structures of 3a, 3b, 4, and 5 were determined by single-crystal X-ray diffractometry (Figure 4, Table 1). Compound 3a features two stiborane moieties flanking a central NH unit (Figure 4, top left). One of the stiborane sites coordinates a molecule of THF (d(Sb-O) = 2.522(3) Å), and the other appears to feature a weak N→Sb interaction (2.970(4) Å). The Sb-O bond distance is consistent with other dative O→Sb(V) bonds [24]. The aryl-antimony bond distances appear to be typical and show little variation [25,26]. The structure of 3b (Figure 4, top right) contains wellseparated antimonate anion and a stibonium cation. Compared to the bis-stiboranyl amine isomer 3a, the Sb-CPh bond length in the antimonate anion is typical (2.131(3)). The structures of the cation in 4 (Figure 4
In summary, we discovered that although oxidation of a bis-2-stibinoaryl amine (2) with chloranil did lead to the desired molecule (3a) with two Sb(V) Lewis acids flanking a central NH site, isomerization of 3a to an ionic, tetraarylstibonium antimonate isomer (3b)
prevents the study and use of 3a as a potential fluoride capture agent. The isomerization proceeds with the cyclization of the diarylamine core. It is possible that the use of a different diarylamine framework that is restricted from such isomerization by the presence of additional covalent linkages, perhaps analogous to the known polycyclic diarylamido-centered PNP pincer ligands [16,[28][29][30][31] could provide a promising alternative.
Unless otherwise stated, all experiments were carried out using standard glovebox and Schlenk line techniques under a dry argon atmosphere. C6D6 was dried over NaK, benzophenone, and 18-crown-6, distilled, and stored over molecular sieves in an argon glovebox prior to usage. All other deuterated solvents were degassed and stored over molecular sieves. Diethyl ether, tetrahydrofuran, and pentane were dried and deoxygenated using a PureSolv MD-5 solvent purification system and were stored over molecular sieves in an argon-filled glovebox. Ph2SbCl was prepared via reaction of neat SbPh3 and SbCl3 in a 2:1 ratio [32].
Compound 1 was prepared via an adapted literature procedure [33]. NMR spectra were recorded on a Varian iNova 500 (
31 P{ 1 H} NMR, 202.276 MHz; 13 C{ 1 H} NMR, 125.670 MHz; 1 H NMR, 499.678 MHz, 19 F NMR, 470.385 MHz) and Bruker 400 spectrometer ( 1 H NMR, 400.09 MHz; 13 C NMR, 100.60 MHz, 19 F NMR (376.498 MHz) spectrometers in given solvents. Chemical shifts are reported in ppm (δ). 13 C{ 1 H} and 1 H NMR spectra were internally referenced to residual solvent resonances [34]. In reporting spectral data, the following abbreviations were utilized: s = singlet; d = doublet; t = triplet; dd = doublet of doublets; m = multiplet. Infrared spectra were collected on an Agilent CARY FT-IR spectrometer. Elemental analyses were performed by CALI, Inc.
(Highland Park, NJ, USA). Mass spectrometry was performed by the Laboratory for Biological Mass Spectrometry at Texas A&M University.
In a 20 mL scintillation vial, equipped with a magnetic stir bar, SbPh3 (4.90 g, 13.9 mmol) and SbCl3 (1.58 g, 6.93 mmol) were combined and left to stir. The solids gradually became a homogenous, golden oil that was left to stir for 16 h before being removed from the stir plate. The oil was left to stand until it had solidified, after which the solid was removed from the vial to give the product as an off-white solid. Yield: 6.47 g (100%). 1 H NMR (400 MHz, C6D6): δ 7.47 -7.42 (m, 4H, Sb-Ph), 7.10 -6.99 (m, 6H, Sb-Ph). 13 C{ 1 H} NMR (101 MHz, C6D6): δ 145.4 (s), 134.7 (s), 130.0 (s), 129.3 (s). Note: The oil must be left to solidify to yield a pure product. 13 C{ 1 H} NMR spectroscopy analysis of the oil reveals a mixture of SbPhxCl3-x species (see SI for more details). 13 C{ 1 H} NMR (101 MHz, C6D6) for SbPh3: δ 138.9 (s), 136.6 (s), 129.2 (s), 128.9 (s). 13 C{ 1 H} NMR (101 MHz, C6D6) for SbCl2Ph: δ 153.3 (s), 132.8 (s), 131.2 (s), 129.4 (s).
To a 200 mL Schlenk flask equipped with a magnetic stir bar was added 1
(2.18 g, 6.13 mmol) and 50 mL diethyl ether. To this stirring solution was slowly added a 2.5 M solution of n-butyllithium in hexanes (7.44 mL, 18.6 mmol) causing the solution to become slightly yellow. The solution was allowed to stir for 3 h, and then Ph2SbCl (3.83 g, 12.3 mmol) was added as a suspension in 3 mL of THF, causing the solution to become a cloudy, bright yellow. After the addition was completed, the suspension was allowed to stir for 16 h. To the suspension was then added 1 mL of degassed H2O, causing the mixture to become a light yellow, homogeneous solution, which was then allowed to stir for 1 h. Solvent was then removed under reduced pressure to yield a waxy yellow residue which was dissolved in a minimal amount of fluorobenzene and filtered through a frit containing Celite and silica. The fluorobenzene was removed in vacuo to give a waxy yellow solid which was washed with 20 mL of isooctane and dried. The resultant solid was dissolved in a minimal of diethyl ether, layered with isooctane, and recrystallized at -37 o C over 12 h. The solvent was decanted, and the solids were washed with isooctane then pentane and dried, providing the product as a white, free-flowing solid. Yield: 1.407 g (31%).
1 H NMR (400 MHz, CDCl3): δ 7.39 -7.35 (m, 8H, Sb-Ph), 7.31 -7.25 (m, 12H, Sb-Ph), -6.91 (m, 4H, Ar CH), 6.72 (d, 2 H, JHH = 8 Hz, 2H, Ar CH), 5.38 (s, 1H, N-H) 2.13 (s, 6H, benzylic CH3). 13 C{ 1 H} NMR (101 MHz, CDCl3): δ 147.1, 138.0, 137.5, 136.5, 132.6, 132.0, 130.9, 129.0, 128.6, 120.5, 20.9 (benzylic CH3). ATR-IR (cm -1 ): νN-H-3338 cm -1 HRMS (ESI + ) m/z [M+H] + : calc'd for C38H33NSb2: 748.0766, found 748.0719. 4.4. Synthesis of 3a: To a 20 mL scintillation vial equipped with a magnetic stir bar was added 2 (65 mg, 0.0857 mmol) and o-chloranil (42 mg, 0.171 mmol), followed by 4 mL of toluene to give a red solution that gradually turned a pale yellow over the course of 15 minutes. The solvent was removed under reduced pressure, and the remaining solid was rinsed with pentane (1 × 2 mL) before drying further to reveal the product as a pale-yellow solid (88 mg, 83%). White, single crystals suitable for X-ray diffraction were grown from a THF/pentane mixture at -38 o C. 1 H NMR (C6D6, 400 MHz): δ 9.03 (s, 1H, N-H), 7.72 (br s, 8H), 7.22 (d, JHH = 8 Hz, 2H, Ar CH), 7.07 -6.88 (m, 16 H), 1.78 (s, 6H, benzylic CH3). 13 C{ 1 H} NMR (C6D6, 101 MHz): δ 147.3, 144.7, 135.0, 134.5, 134.3, 133.9, 133.7, 132.0, 129.9, 129.3, 125.7, 124.7, 121.4, 117.1, 20.5. ATR-IR (cm -1 ): νN-H-3193 cm -1 . 4.5. Synthesis of 3b: To a 20 mL scintillation vial equipped with a magnetic stir bar was added 2 (64 mg, 0.0857 mmol) and o-chloranil (42 mg, 0.171 mmol), followed by 6 mL of THF to give an orange solution. The solution was left to stir overnight (ca. 16 h), during which the color changed to very pale yellow. The solvent was removed under reduced pressure, and the residue was washed with Et2O (2 × 2 mL) before drying further to reveal the product as a pale yellow solid (97 mg, 92%). 1 H NMR (400 MHz, DMSO-d6): δ 9.79 (s, 1H, N-H), 7.81 (m, 4H), 7.65 (br, 8H), 7.55 (m, 4H), 7.43 (m, 6H), 7.36 (d, J = 8 Hz, 2H, Ar CH), 7.29 (d, J = 8 Hz, 2H, Ar CH), 2.24 (s, 6H, benzylic CH3). 13 C{ 1 H} NMR (101 MHz, DMSO-d6): δ 147.5, 146.0, 144.7, 144.2, 134.8, 134.8, 134.3, 132.5, 132.2, 130.4, 130.3, 130.1, 129.0, 119.0, 118.2, 117.2, 116.0, 115.2, 105.3, 20.1. ATR-IR (cm -1 ): νN-H-3334 cm -1 . Elem Anal Found (calc): C: 49.46 (49.47) H: 3.12 (3.15). 4.6. Variable temperature NMR study of 3b: To a J. Young tube was added 3b (17.0 mg, 0.013 mmol) and THF-d8 resulting in a clear, colorless solution. 1 H NMR spectra taken at variable temperatures (25, 0, -20, and -40 o C) revealed no changes in solution symmetry on the NMR timescale.
To a J. Young tube was added 11 mg of 3a, followed by 1.5 µL of mesitylene as an internal standard. Afterwards, 500 µL of the appropriate solvent was added to each tube. The tube was shaken to ensure complete solvation of the material, and then analyzed by 1 H NMR spectroscopy (ca. 10 min after mixing).
The tubes were continued to be monitored by 1 H NMR spectroscopy at various timepoints. All concentrations (including initial [3a]0) of isomers were determined by integrating the aryl-CH3 resonances against the mesitylene CH3 resonance. For samples where the ionic isomer had limited solubility, DMSO-d6 was added after the final timepoint to confirm its presence and ratio to the neutral isomer.
To a 20 mL scintillation vial equipped with a magnetic stir bar was added 3b (122 mg, 0.098 mmol) and 4 mL THF. To this stirring suspension was then added CsF (15 mg, 0.099 mmol), causing the suspension to form a homogeneous solution. The resultant solution was then stirred for 16 h, and then volatiles were removed under reduced pressure, forming an off-white residue. The residue was then extracted with 5 mL PhF and filtered through a plug of Celite to give a clear, colorless solution. Volatiles were removed in vacuo to provide the product as a freeflowing, white powder (46 mg, 87%). 1 H NMR (400 MHz, CDCl3): δ 7.92 -7.85 (m, 4H, SbPh2), 7.45 -7.37 (m, 6H, SbPh2), 7.12 (dd, J1 = 8 Hz, J2 = 2 Hz, 2H), 6.82 (d, J1 = 8 Hz, 2H), 6.68 (br s, 1H, N-H), 2.20 (s, 6H, Ar-CH3). 19 F{ 1 H} NMR (376 MHz, CDCl3): δ -100.4. 1 H NMR (400 MHz, THF-d8) δ 8.31 (s, 1H, N-H), 7.93 7.84 (m, 4H, SbPh2), 7.58 -7.46 (m, 2H, Ar CH), 7.41 -7.30 (m, 6H, Sb-Ph), 7.08 (dd, J1 = 8 Hz, J2 = 2 Hz, 2H, Ar CH), 6.92 (d, J = 8 Hz, 2H, Ar CH), 2.14 (s, 6H, benzylic CH3). 13 C{ 1 H} NMR (101 MHz, THF-d8) δ 148.1 (s), 138.4 (s), 138.1 (s), 137.3 (s), 132.6 (s), 130.7 (s), 129.0 (s), 128.8 (s), 117.9 (br s), 117.3 (s), 20.44 (s). 19 F{ 1 H} NMR (376 MHz, THF-d8): δ -105.1HRMS (ESI + ) m/z [M+H] + : calc'd for C26H23FNSb: 491.0851, found 491.0847.