Benzothiazole Nickelation: An Obstacle to the Catalytic Arylation of Azoles by Cyclopentadienyl Nickel N-Heterocyclic Carbene Complexes

Abstract

NiCp^†L(NHC)]⁽⁺⁾ complexes (Cp^† = Cp (η⁵-C₅H₅), Cp* (η⁵-C₅Me₅); NHC = N-heterocyclic carbene; L = Cl⁻ or NCMe) have been tested as pre-catalysts for the direct arylation of benzothiazole in the presence of an alkoxide. Only the pentamethylcyclopentadienyl derivative, [NiCp*Cl(IMes)] (IMes = 1,3-bis(2,4,6-trimethylphenylimidazol-2-ylidene), enabled low conversion to the desired coupling product with phenyl iodide as the electrophilic coupling partner. In contrast, all cyclopentadienyl complexes proved to be inactive. ¹H NMR studies of the “catalytic” reaction mixtures demonstrate that they cleanly convert to an unreactive C(2)-benzothiazolyl derivative, whose identity has been confirmed by an independent synthesis and characterization. The latter constitutes a potential energy well that quenches all further reactivity, and provides a rare example of C(2)-metallated azolyl complex.

1. Introduction

The recent trend to use 3d transition metal catalysts, that is driven by economic and environmental concerns, has led to a re-assessment of nickel’s reactivity [1,2,3,4]. In particular, nickel catalysts have emerged as promising candidates for C–H bond functionalization [5,6,7,8], and nickel-catalyzed reactions between (hetero)arenes and aryl halides [9,10,11,12], phenol derivatives [13,14,15,16,17,18], aryl or alkenyl esters [19,20,21] or even alkyl halides [22,23] have been shown to be reliable atom-economical alternatives to palladium-catalyzed cross-couplings. In this context, we have recently reported that the unsaturated T-shaped nickelacycle 1, that bears a mono-anionic C,C-NHC-cyanoalkyl chelate (NHC = N-heterocyclic carbene), shows moderate activity for the coupling of benzothiazole with aryl iodides (Scheme 1) [24]. In contrast, the corresponding cyclopentadienyl nickelacycle 2 that does not possess any potentially available free coordination sites, unless an eventual Cp ring slippage occurs [25,26], proved totally unreactive under the same conditions (Scheme 1). This suggested to us that an unsaturated nickel center and/or labile ligands were necessary to observe benzothiazole arylation [24]. We therefore wondered whether Cp^†Ni [Cp^† = Cp (η⁵-C₅H₅), Cp* (η⁵-C₅Me₅)] complexes bearing a monodentate NHC ligand as well as a readily labile halide or acetonitrile ligand could catalyze such direct biaryl couplings. This family of complexes has indeed been proven to constitute a very versatile class of pre-catalysts [27,28,29,30,31,32,33] that notably catalyze the α-arylation of ketones [34]. Herein, we report our findings, which show that, under the typical conditions used for the direct coupling of azoles with aryl halides or pseudo-halides (i.e.: under harsh reaction conditions with an alkoxide base [5,9,10,12,24]), apart from a Cp* derivative which shows a little activity with aryl iodides, these complexes form a stable C(2)-nickelated azolyl complex that arrests any further reactivity.

2. Results and Discussion

2.1. Catalytic Studies

Four half-sandwich Ni(NHC) complexes, which have already been demonstrated to catalyze a variety of reactions [27,28,29,30,31,32,33], were chosen for the present study: complexes 3 [35,36,37] and 4 [27,37] bearing a Cp ring, a chloride and, respectively, the 1,3-bis(2,4,6-trimethylphenylimidazol-2-ylidene (IMes) and 1,3-bis(2,6-diisopropylphenylimidazol-2-ylidene (IPr) as NHC ligands, complex 5 [38], the Cp* analogue of 3, and complex 6 [29], the cationic analogue of 3 that bears a labile acetonitrile ligand instead of a chloride (Figure 1).

Catalytic attempts focused on the reaction of benzothiazole with p-bromotoluene or iodobenzene under conditions similar to those established for complex 1 [24] and Itami’s Ni(OAc)₂/bipyridine or 1,1′-bis(diphenyphosphino)ferrocene system [9,12], i.e.: in 1,4-dioxane at 120–140 °C in the presence of LiOtBu (1.5–2 equiv.) and 5 mol% of 3–6 for 16 h (

Table 1. Attempted coupling of benzothiazole with p-bromotoluene or iodobenzene in the presence of [NiCp^†L(NHC)]⁽⁺⁾ complexes.¹

Entry	Catalyst	X	R	Conv. [%] 2
1 3	3	Br	Me	0
2	3	I	H	0
3 4	4	I	H	0
4	5	I	H	13
5	6	I	H	0

¹Reaction conditions: benzothiazole (1 equiv.), p-bromotoluene or iodobenzene (1.5 equiv.), LiOtBu (2 equiv.), 3–6 (5 mol %) in 1,4-dioxane at 140 °C for 16 h. ² Conversion to the desired coupling product established by GC; average value of two runs. ³ Run at 120 °C. ⁴ Run with 1.5 equiv. of LiOtBu.

). No conversion to the desired hetero-biaryl compound was observed with 3 with either p-bromotoluene or iodobenzene (entries 1 and 2), and variations of the base (NaOtBu, KOtBu, Cs₂CO₃ or KOAc) and solvent (toluene or THF) gave no better results (not shown in Table 1). Similarly, the other two Cp complexes 4 and 6 did not catalyze the coupling of benzothiazole and iodobenzene under such reaction conditions (entries 3 and 5). Only the more electron-rich and more sterically crowded Cp* complex 5 enabled a conversion of 13% at 140 °C (entry 4). However, this was still far from the activity observed with complex 1 (Scheme 1) that proved to have a narrow reaction scope and showed a very moderate activity with most substrates [24].

As we were intrigued by the incapacity of the CpNi(NHC) complexes to catalyze these direct couplings, we attempted to establish the fate of the latter under the reaction conditions. To our surprise, the ¹H NMR spectra of the reaction media (after removal of the insoluble and volatiles—see Section 3) of reactions run with 3 under various conditions of solvent (THF, toluene or 1,4-dioxane), base (LiOtBu, NaOtBu, KOtBu) and temperature (from 90 to 140 °C) systematically indicated the presence of unreacted benzothiazole and p-bromotoluene or iodobenzene, as well as the full and clean conversion of 3 to a novel CpNi(IMes) complex 7 bearing a C(2)-metallated benzothiazolyl group (see Figures S1 and S2, and Scheme 2), whose identity was confirmed by an independent synthesis (vide infra). The latter was indeed characterized by typical signals for its IMes and Cp groups, but which are shifted compared to those of 3, as well as by two apparent triplets centered at 7.11 and 6.97 ppm, which integrate for one proton each relative to the Cp and IMes signals and correspond to the H(5) and H(6) protons of a benzothiazolyl unit. However, the signals of the H(4) and H(7) overlapped with those of unreacted benzothiazole and were not unambiguously observed. Complex 7 was also observed in the attempted couplings with the cationic complex 6 (Figure S3), highlighting a general behaviour for these CpNi(NHC) species thereof.

Interestingly, although C(2)-benzoxazolyl or benzothiazolyl complexes are often proposed as intermediates in C–H/C–X couplings between heteroarenes and aryl, alkenyl or alkyl electrophiles [5,10,12,14,16,17,19,20,22,23], they remain very rare. We are indeed aware of only four related PCN-Pd [39,40,41] and NNN-Ni [23] pincer species, which have been isolated with a benzothiazolyl ligand and demonstrated to be intermediates in similar couplings. Thus, although it obviously constitutes a potential energy well, complex 7 is an interesting example of azolyl trapping in such (attempted) direct coupling. Furthermore, it is to our knowledge only the second C(2)-benzothiazolyl-nickel species reported to date [23].

2.2. Independent Synthesis and Characterization of the Benzothiazolyl Complex 7

To confirm the identity of 7, complex 3 was treated with a small excess of benzothiazole (1.2 equiv.) in the presence of potassium bis(trimethylsilyl)amide (KHMDS) as a base in toluene at room temperature (Scheme 3). The resulting complex was isolated in 80% yield as an air-stable green-brown powder after work-up. Full characterization by ¹H and ¹³C{¹H} NMR, HRMS spectroscopies and CHN microanalyses confirmed the expected formulation. Thus, for instance, the ¹H NMR spectrum of 7 no longer displays the H(2) proton of benzothiazole, but displays the H(5) and H(6) protons as two apparent triplets at 7.12 and 6.98 ppm, as seen in the spectra of the catalytic attempts (vide supra), and the H(4) and H(7) protons as two doublets at 7.58 and 7.51 ppm (Figure S4). The Cp and IMes protons are found at values that are typical for NiCp(IMes) compounds [27,28,29,30,31,32,33,35,36,37]. Regarding the ¹³C{¹H} NMR spectrum, the C(2) carbon of the benzothiazolyl moiety is observed at 176.0 ppm, which is significantly downfield compared to the corresponding carbon of free benzothiazole (153.8 ppm) [42], and thus confirms its metallation (Figure S5) [23]. In addition, it is worth mentioning that the carbene carbon signal of 7 is downfield shifted (177.9 ppm) when compared to that of 3 (165.9 ppm [35]), which suggests a decrease in the Lewis acid character of the metallic center caused by an increase in the σ-donor ability of the ancillary ligand, [43,44,45,46]. This is similar to what is observed with the phenyl and methyl derivatives: [NiCpPh(IMes)] (181.2 ppm) [34] and [NiCpMe(IMes)] (187.4 ppm) [35].

2.3. Crystallographic Study of 7

The X-ray diffraction study of a single crystal of complex 7 corroborated the NMR data and confirmed the molecule’s structure (Figure 2 and Appendix A). Significant bond distances and bond angles are given in

Table 2. Key bond distances (Å) and angles (°) of 7 with Esd’s in parentheses.

Bond or Angle	7
Ni–C(1)	1.873(2)
Ni–C(2)	1.875(2)
Ni–Cpcent 1	1.774
Ni–CCp av 2	2.140
C(1)–Ni–C(2)	94.26(10)
C(1)–Ni–Cpcent	135.3
C(2)–Ni–Cpcent	130.4
Ni–(C(1)–Cpcent–C(2))	0.037

¹ Cp_cent = centroid of the Cp group. ² Average Ni–C distance to the Cp ring.

and crystallographic data and data collection parameters are listed in Table S1. Complex 7 crystallizes in the orthorhombic chiral space group P2₁2₁2₁. The Ni–C(1) [C(1) = the carbene carbon atom] and Ni–C(2) distances are respectively 1.873(2) and 1.875(2) Å, which fall slightly below the range observed for related NiCp(IMes) complexes bearing a third C-bound ligand, such as [NiCp(IMes)Ph] (Ni–C(1): 1.875(2), Ni–C(2): 1.908(2) Å) [34] and [NiCp(IMes)(IMe)](BF₄) (Ni–C(1): 1.899(3), Ni–C(2): 1.906(3) Å) (IMe = 1,3-dimethylimidazol-2-ylidene) [47]. The Ni–C(2) distance is also significantly shorter than that observed in the only other fully characterized Ni–C(2)-benzothiazolyl complex (1.936(8) Å), that was just reported [23]. The C(1)–Ni–C(2) angle of 94.26(10)° falls in the same range as observed for [NiCp(IMes)Ph] (95.35(9)°) [34] and [NiCp(IMes)(IMe)](BF₄) (96.91(11)°) [47]. Finally, when one takes into account the plane formed by the Cp centroid, C(1), and C(2), the nickel atom is almost in a planar environment (Table 2, last entry), as is typically observed in similar two-legged piano-stool CpNi(NHC) complexes [27,28,29,30,31,32,33,35].

2.4. Investigation of the Reactivity of the Cp* Complex 5

The little activity observed with the Cp* complex 5 (Table 1, entry 7) raised the question whether the observed coupling reaction goes through a benzothiazolyl complex similar to 7. To answer this question, complex 5 was treated with benzothiazole and KHMDS similarly to 3. A quick reaction occurred, but the corresponding benzothiazolyl complex 8 could only be isolated in 12% yield (Scheme 4) and we have been unable to obtain it analytically pure owing to its relative instability in solution that leads to insoluble materials and/or paramagnetic species. The metallation of benzothiazole was nevertheless confirmed by the absence of the H(2) proton in the ¹H NMR spectrum (Figure S6) and the downfield shift of the C(2) carbon from 153.8 to 180.9 ppm in the ¹³C{¹H} NMR spectrum (Figure S7).

Complex 8 was then treated with 1.5 equiv. of iodobenzene in 1,4-dioxane at 140 °C for 16 h. The reaction led to a massive decomposition of the complex together with a small amount of the expected coupling product (<15% of the identified species) as inferred from the ¹H NMR analysis of the crude reaction mixture that shows a very complex mixture. This illustrates the less stable (or more reactive) character of the Cp* derivative 8, when compared to its Cp analogue 7, and suggests the possible involvement of such a benzothiazolyl complex in the coupling reaction [23,39,40,41], provided that it does not constitute a potential energy well like the Cp derivative 7.

A possible explanation for the enhanced reactivity of 8 may come from both its more electron-rich character, as shown by its downfield shifted carbene carbon in the ¹³C {¹H} NMR spectrum: 187.4 ppm in 8 vs. 177.9 ppm in 7 [43,44,45,46], and its more significant steric crowding. Indeed, in a hypothetic Ni(II)/Ni(IV) catalytic cycle, as proposed by Chatani et al. for a related arylation reaction of Csp²–H bonds with aryl iodides [48], the presence of a bulky electron-donating ligand such as Cp* may (i) favor the oxidative addition of the aryl halide through electronic donation, and (ii) subsequently enhance the reductive elimination of the coupling product due to its steric bulkiness [49]. Further developments of this family of complexes for similar cross-couplings might therefore privilege species with more electron-donating and/or more bulky ligands [50].

3. Material and Methods

3.1. General Comments

All reactions were carried out using standard Schlenk procedures under an atmosphere of dry argon. Solvents were dried and distilled under argon. Solution NMR spectra were recorded at 298 K on a Bruker Avance I 300 MHz spectrometer operating at 300.13 MHz for ¹H and at 75.47 for ¹³C{¹H}. A DEPT ¹³C spectrum was recorded for complex 7 to ease the ¹³C signals assignment. The chemical shifts are referenced to the residual deuterated or ¹³C solvent peaks. Chemical shifts (δ) and coupling constants (J) are given in ppm and Hz, respectively. GC analyses were carried out with n-dodecane as an internal standard, using an Agilent 7820A instrument equipped with a HP-5 column (cross-linked 5% phenyl silicone gum, 30 m × 0.32 mm × 0.25 μm). Conversion was determined by considering instrument response factors. The GC conditions were: (i) 50 °C for 5 min, (ii) increase in temperature at a rate of 10 °C/min until 150 °C, (iii) isotherm for 5 min, and (iv) temperature increase at 20 °C/min until 240 °C; retention times: iodobenzene: 8.7 min; dodecane 11.6 min; benzothiazole: 12.0; 2-phenylbenzo[d]thiazole: 23.4 min. The retention time of 2-phenylbenzo[d]thiazole was determined with a pure sample of the latter obtained by using 1 as catalyst (Scheme 1) [24]. Elemental analyses were performed by the Service d’Analyses, de Mesures Physiques et de Spectroscopie Optique, Institut de Chimie, UMR 7177, Université de Strasbourg. High-resolution mass spectra were recorded on a Bruker micrOTOF mass spectrometer by the Service the Spectrométrie de Masse, Institut de Chimie, UMR 7177, Université de Strasbourg. Benzothiazole, haloarenes and n-dodecane were distilled and stored under argon before use. [NiCpCl(IMes)] (3) [35,36], [NiCpCl(IPr)] (4) [27,37], [NiCp*Cl(IMes)] (5) [38], and [NiCp(NCMe)(IMes)](PF₆) (6) [29] were prepared according to the published methods.

3.2. Typical Procedure for the Attempted Heteroarylations of Aryl Halides

An oven-dried Schlenk tube containing a magnetic stir bar was loaded with 3 (10 mg, 0.0216 mmol, 5 mol %), LiOt-Bu (69 mg, 0.861 mmol, 1.9 equiv.), benzothiazole (50 µL, 0.459 mmol, 1.0 equiv.), iodobenzene (75 µL, 0.664 mmol, 1.4 equiv.) or p-bromotoluene (109 mg, 0.637 mmol, 1.4 equiv.), n-dodecane (15 µL, 0.066 mmol), and 1,4-dioxane (3 mL), and sealed. The Schlenk tube was then put into an oil bath that was heated up to 140 °C. After 16 h, the reaction medium was cooled down to room temperature, and a sample (0.1 mL) was removed for GC analysis. The volatiles were then removed under gentle vacuum, and the resulting residue was extracted with diethyl ether and filtered over a small plug of silica (Merck Silica Gel 60-mesh size 40–60 μm). Diethyl ether was then evaporated and the solid residue analysed by ¹H NMR in CDCl₃.

3.3. Synthesis of [NiCp(C₇H₄NS)(IMes)] (7)

KHMDS (1.2 mL, 0.600 mmol, 0.5 M in toluene) was added to a solution of 3 (200 mg, 0.431 mmol) and benzothiazole (56 µL, 0.514 mmol) in toluene (12 mL) at room temperature, resulting in a very rapid change of color from red to green. After 12 h, the reaction medium was filtered through a Celite pad and concentrated in vacuo to give a brownish powder that was washed with pentane (3 × 10 mL). Re-dissolution of this solid in toluene (10 mL) and THF (0.5 mL) and filtration through neutral alumina with toluene and toluene/THF (20:1) as successive eluents then afforded a green solution that was evaporated to dryness to give 7 as a green-brown powder (193 mg, 0.343 mmol, 80%). Anal. Calcd for C₃₃H₃₃N₃NiS: C, 70.48; H, 5.91; N, 7.47. Found: C, 70.55; H, 5.98; N, 7.80. HR-MS (ESI): m/z [M + H]⁺ calcd for C₃₃H₃₄N₃NiS 562.1821, found 562.1789. ¹H NMR (CDCl₃, 300.13 MHz): δ 7.58 (d, ³J = 8.1, 1 H, C₇H₄NS), 7.51 (d, ³J = 7.8, 1 H, C₇H₄NS), 7.12 (m, 1 H, C₇H₄NS), 6.98 (m, 1 H, C₇H₄NS), 6.91 (s, 2 H, NCH), 6.89 (s, 4 H, m-H_Mes), 4.83 (s, 5 H, C₅H₅), 2.38 (s, 3 H, p-CH₃), 2.10 (s, 6 H, o-CH₃). ¹³C{¹H} NMR (CDCl₃, 75.47 MHz): δ 177.9 (NCN), 176.0 (NCS), 155.1 (C₆H₄-C(3a)), 140.6 (C₆H₄-C(7a)), 138.8 (p-C_Mes), 136.8 (ipso-C_Mes), 135.6 (o-C_Mes), 129.1 (m-C_Mes), 123.5 (NCH), 122.8, 120.4, 119.5 and 118.6 (C₆H₄-C(3,4,5,6)), 91.4 (C₅H₅), 21.3 (p-CH₃), 18.5 (o-CH₃).

3.4. Synthesis of [NiCp*(C₇H₄NS)(IMes)] (8)

KHMDS (0.9 mL, 0.450 mmol, 0.5 M in toluene) was added to a solution of 5 (236 mg, 0.447 mmol) and benzothiazole (48 µL, 0.442 mmol) in toluene (12 mL) at room temperature, resulting in a very rapid change of color from red to olive. After 2 h, the volatiles were evaporated under vacuum to give a brown residue that was extracted with toluene and loaded on the top an alumina pad (4 × 3 cm) that was eluted (i) with toluene to yield a yellow fraction, and (ii) with a toluene/THF mixture (20:1) to yield an olive fraction. The latter was then evaporated to dryness to give 8 as an olive solid (34 mg, 0.0538 mmol, 12%). ¹H NMR (C₆D₆, 300.13 MHz): δ 7.92 (d, ³J = 8.1, 1 H, C₇H₄NS), 7.76 (d, ³J = 7.8, 1 H, C₇H₄NS), 7.24 (m, 1 H, C₇H₄NS), 7.02 (m, 1 H, C₇H₄NS), 6.81 (s, 4 H, m-H_Mes), 6.00 (s, 2 H, NCH), 2.23 (s, 3 H, p-CH₃), 2.09 (s, 6 H, o-CH₃), 1.43 (s, 15 H, C₅Me₅). ¹³C{¹H} NMR (C₆D₆, 125.77 MHz): δ187.4 (NCN), 180.9 (NCS), 157.1 (C₆H₄-C(3a)), 141.3 (C₆H₄-C(7a)), 138.6 (o-C_Mes), 138.2 (p-C_Mes), 137.1 (ipso-C_Mes), 129.7 (m-C_Mes), 124.2 (NCH), 123.5, 120.9, 120.3 and 119.8 (C₆H₄-C(3,4,5,6)), 101.7 (C₅Me₅), 21.5 (p-CH₃), 19.4 (o-CH₃), 10.6 (C₅Me₅).

3.5. Crystallographic Studies

Single crystals of 7 suitable for X-ray diffraction studies were harvested from a batch of crystals obtained at room temperature from a toluene/pentane (1:3) solution. Diffraction data were collected at 173(2) K on a Bruker APEX II DUO Kappa CCD area detector diffractometer equipped with an Oxford Cryosystem liquid N₂ device using Mo-Kα radiation (λ = 0.71073 Å). The crystal-detector distance was 38 mm. The cell parameters were determined (APEX2 software [51]) from reflections taken from three sets of twelve frames, each at ten seconds exposure. The structures were solved using direct methods with SHELXS-2014 and refined against F² for all reflections using the SHELXL-2014 software [52]. A semi-empirical absorption correction was applied using SADABS [53] in APEX2 [51]: transmission factors: T_min/T_max = 0.7423/0.9540. All non-hydrogen atoms were refined with anisotropic displacement parameters, using weighted full-matrix least-squares on F². Hydrogen atoms were included in calculated positions and treated as riding atoms using SHELXL default parameters. A summary of crystal data, data collection parameters, and structure refinements is given in Table S1.

4. Conclusions

In summary, we have shown that apart from the Cp* derivative 5, which shows a little activity, all other tested half-sandwich Ni(NHC) of general formula [NiCpL(NHC)]⁽⁺⁾ complexes are incapable of catalyzing the direct coupling of benzothiazole with aryl halides in the presence of an alkoxide as a base. They instead form a C(2)-metallated benzothiazolyl complex that constitutes a potential energy well that quenches all further reactivity. Although benzothiazolyl complexes are often proposed as possible intermediates in such biaryl couplings, the isolated complex, [NiCp(C₇H₄NS)(IMes)] (7), is to our knowledge a very rare example of fully characterized C(2)-benzothiazolyl-nickel complex.