Synthesis of 1-Azatriene Complexes of Tungsten: Metal-Promoted Ring-Opening of Dihydropyridine

Abstract

For nearly a century, chemists have explored how transition-metal complexes can affect the physical and chemical properties of linear conjugated polyenes and heteropolyenes. While much has been written about higher hapticity complexes (η⁴–η⁶), less is known about the chemistry of their η² analogues. Herein, we describe a general method for synthesizing 5,6-η²-(1-azatriene) tungsten complexes via a 6π-azaelectrocyclic dihydropyridine ring-opening that is promoted by the π-basic nature of {WTp(NO)(PMe₃)}. This study includes detailed spectroscopic and crystallographic data for the η²-dihydropyridine and η²-1-azatriene complexes, both of which were prepared as single regio- and stereoisomers.

Introduction

Since Reihlen and co-workers first reported the complex Fe(CO)₃(η⁴-1,3-butadiene) in 1930,¹ chemists have been fascinated by the ability of transition metals to influence the physical and chemical properties of conjugated polyenes.² The vast majority of these studies have focused on η⁴-, η⁵-, and η⁶-coordinated polyenes, with some of the most significant developments historically involving iron.^2,3 By comparison, much less is known about the properties of their η²-bound counterparts. Of the few complexes reported that involve nonaromatic polyenes, most are cyclooctatetraene complexes of Mn,^4,5 Cu,⁶ and Ni.⁷ One reason for this dearth of linear η²-polyene complexes is the increased likelihood of constitutional- and stereoisomers, particularly if the metal complex is chiral.⁸ As an example, a complex formed between 1,3-pentadiene and an asymmetric metal complex can form up to 24 isomers (12 pairs of enantiomers).⁸ Thus, we were intrigued when we unexpectedly discovered that certain η²-dihydropyridine tungsten complexes could undergo an electrocyclic ring-opening to generate single diastereomers (racemic mixture) of η²-coordinated linear trienes and azatrienes (Figure 1B).⁹ This Zinke-König-like ring scission was possible when a substituent (X) alpha to the nitrogen was capable of π-donation, thereby stabilizing the carbocation generated by C–N ring scission. We questioned whether it was possible that the π-donating properties of tungsten alone could promote such a ring-scission, without the need for a π-donor substituent. The ring-opening could occur via an η²-allyl species that, following an allyl shift,¹⁰ would form an azatriene complex with the metal fragment intact (Figure 1C). If the dihydropyridine (DHP) complex could be prepared as a single regio- and stereoisomer, a single regio- and stereoisomer of the azatriene could result.

Figure 1.

(A) Generic representation of the Zincke-König reaction. (B) Tungsten-mediated Zincke-König reaction with a C or N π-donor (X). (C) Proposed electrocyclization featuring an η²-allyl intermediate.

Previous efforts have demonstrated that a complex of {WTp(NO)(PMe₃)} ([W]; Tp = trispyrazolylborate) and N-acetylpyridinium triflate (4D) is a valuable precursor to functionalized tetrahydropyridines,¹¹⁻¹³ isoquinuclidines,¹⁴ and trienyl enamides.⁹ This complex (4D) is generated by the initial exchange of WTp(NO)(PMe₃)(benzene) (1) with pyridine-borane to form complex 2 as a mixture of distal (D) and proximal (P) coordination diastereomers (3:1 ratio; Figure 2A).¹⁵ This mixture then undergoes oxidative BH₃-removal in acidic acetone to form the parent complex, [WTp(NO)(PMe₃)(η²-pyridinium)]OTf (3D and 3P). Owing to metal–ligand backbonding,¹⁶ the pyridine nitrogen is highly nucleophilic, allowing for its rapid acetylation in the presence of acetic anhydride. Fortuitously, gently heating this solution (55 °C) improves the coordination diastereomer ratio (cdr) from 3:1 to 10:1, thus maximizing the diastereomer with partial positive charge distal to the PMe₃ (Figure 2, Panel B).¹⁰ This work demonstrates the tosylation of 4 via an analogous process that enriches before subsequently undergoing an array of nucleophilic additions followed by a heat-induced ring opening to form azatrienes η²-bound to a [W] fragment.

Figure 2.

(A) Previously reported synthesis of 4D and the analogous synthesis of 5D from 1. (B) Heat-induced isomerization of 5P to 5D.

Results and Discussion

While the acetylpyridinium complex 4D has shown broad synthetic utility,^9,11−14 the corresponding C2-alkylated DHP complexes (i.e., no potential to π donate) failed to undergo the ring-opening shown in Figure 1. Hence, we sought a more electron-withdrawing functional group for the nitrogen that could still effect high DHP stereoselectivity (cdr). We settled on the N-tosylated pyridinium complex 5D to conduct our investigation.

In a procedure analogous to the synthesis of the N-acetylated complex (4), Ts₂O was stirred in a solution of 3 and 1,2-dichloroethane. Upon addition of lutidine, a change in the J_WP (³¹P NMR spectrum; ¹⁸³W) from ∼294 Hz (consistent with 3D) to 279 and 286 Hz, indicated that tosylation had occurred.¹⁷ After 10 min at room temperature, the cdr had changed from 3:1 to ∼6:1 (³¹P NMR). Similar to its acylated derivative 4D, the distal stereoisomer (5D) is thermodynamically favored and can be maximized through mild heating (55 °C) (Figure 2B).¹⁰ The resulting 11:1 mixture would reach as high as 20:1 (¹H NMR) after undergoing an aqueous workup and precipitation. The structural assignment of this compound was supported with 2D NMR and high-resolution mass spectrometry (HRMS) data and yields for 5D were as high as 89% when carried out on a multigram scale (10 g). Of note, the ¹H and ¹³C NMR data indicated that the counterions triflate (–OTf) and tosylate (–OTs) were both present in a mixture in the final product. A cyclic voltammogram demonstrated that this complex has an E_pa of +1.18 V (N,N-dimethylacetamide (DMA), 100 mV/s), just slightly positive of the acetylated analogue, 4D.¹⁵ However, the ν_NO values for 4D and 5D (FTIR/ATR) are 1611 and 1607 cm^–1, respectively, an observation that together with the electrochemical data suggests that the tosyl group on 5D is not significantly more electron-withdrawing than the acetyl group on 4D. Fortunately, this feature would ultimately not preclude our intended goal (vide infra).

Nucleophilic Additions to the DHP

The protection of the η²-bound pyridine with a tosyl group activates the iminium carbon for a wide range of chemoselective nucleophilic addition reactions. Nucleophilic addition to N-activated pyridinium salts is well-established,¹⁸ and while a wide range of 1,2-DHPs can be prepared via this method, these syntheses can be fraught with regioselectivity issues when the pyridine lacks a directing group (Figure 3A). In the case of complex 5, the coordination of the pyridinium at C3 and C4 removes the possibility of C4 addition, and the bulky nature of the tungsten fragment [W] exclusively directs these nucleophiles anti to the metal at the C2 carbon.¹⁹

Figure 3.

(A) Pyridinium salts (R = electron-withdrawing group) undergo selective addition reactions when coordinated to [W]. (B) DHP complexes 6D–13D are generated by incorporating various nucleophiles to 5D at C2 anti to the metal. All syntheses were conducted at room temperature under an inert N₂ atmosphere. ^aZnMe₂, THF. ^bBnMgCl, THF. ^cMBA, Zn°, THF. ^dMTDA, TEA, DCM. ^eallyl bromide, Zn°, THF. ^fpropargyl bromide, Zn°, THF. ^gNaCN, MeOH. ^hNaBH₄, MeOH. (C) ORTEP/ellipsoid diagrams of 7D and 11D were acquired via SC-XRD.

A range of carbon nucleophiles were chosen for the synthesis of DHP complexes shown in Figure 3. Yields of DHP complexes 6D–13D range from 43–84%, the compounds being synthesized as racemic mixtures (Figure 3B). Nucleophiles successfully added included dimethylzinc (6D), benzyl magnesium chloride (BnMgCl) (7D), methyl bromoacetate (MBA) with zinc powder (8D), and methyl trimethylsilyl dimethylketene acetal (MTDA) (9D). Barbier reactions were conducted in the syntheses of 10D and 11D, using allyl bromide and propargyl bromide, respectively. Finally, cyanide and hydride were incorporated through the addition of NaCN and NaBH₄ (12D and 13D). The structures of 6D, 7D, and 11D were determined via single-crystal X-ray diffraction (SC-XRD; Figure 3C), confirming their regio- and stereochemical assignments.

DHP Ring Opening

In addition to the successful preparations of DHP complexes shown in Figure 3, we attempted to synthesize a 2-nitromethyl-substituted DHP complex (14D). Using reaction conditions identical to those reported by Harrison et al. for the analogous acetyl-pyridinium system,¹⁹ excess nitromethane and triethylamine (TEA) were combined in a DCM solution of 5D and stirred for 40 min. ¹H NMR spectra revealed the presence of a lone diamagnetic complex characterized by several unexpected signals including a downfield doublet at 8.60 ppm and two doublet-of-doublets at 7.29 and 6.27 ppm. We anticipated that deprotonation alpha to the nitro-group might occur (14′ in Figure 4) analogous to what Harrison et al. observed for malononitrile (Figure 1B). However, a full 2D NMR analysis indicated that 15D was instead the desired 1-azatriene complex 15D (Figure 4A). Unfortunately, paramagnetic impurities were also present and attempts to cleanly generate 15D or 14D by modulating the solvent, base, equivalence, temperature, or time were ineffective. However, other DHP complexes, including 6D–8D and 11D, were found to undergo this ring-opening with gentle heating (55 °C for 24 h), forming the desired 1-azatrienes 16D–19D (Figure 4B). These compounds were characterized by similar downfield ¹H NMR features as observed for 15D but could be cleanly generated. Crystal structures of 17D and 19D not only confirm the proposed azadiene connectivity but show an E,E azatriene stereochemistry. This is consistent with H–H coupling for the uncoordinated alkene (∼15 Hz) and NOE data, which supports a trans-vicinal-hydrogen orientation (Figure 4). Bond lengths for the azatriene ligand are unremarkable, being largely in agreement with the crystal data of organic analogues²⁰⁻²³ but with a slight shortening (∼0.02 Å) of the C4–C5 and C2–C3 bonds. This observation is consistent with metal-backbonding into the azadiene π-system.²⁴

Figure 4.

(A) Synthesis of 15D via the 14D intermediate. (B) Scope of DHP ring-opening to form 15D–19D.

We anticipate that the driving force behind this reaction is the generation of a 1-azatriene ligand, which is an excellent π-acceptor for the π-basic [W] fragment.¹⁶ For example, DFT calculations indicate that the conversion of 6D to 16D is −2.8 kcal/mol. Typically, the reverse reaction—the electrocyclic ring closure— is the spontaneous reaction,²⁵ unless the nitrogen has a strong withdrawing group and there is a π-donating substituent at C2 of DHP (see Figure 1).²⁶ For example, DFT calculations indicate that conversion of the dihydropyridine ligand of 6D is 0.7 kcal/mol more stable than the azatriene ligand of 16D (Supporting Information). The purported mechanism behind this transformation (Figure 5) involves breaking of the C2–N bond (I → III) followed by 6π-electron-rearrangement concomitant with the “η²-allyl shift”¹⁰ to render an η²-azatriene complex with the terminal alkene (C5, C6) η²-bound to the metal. Breaking the C2–N bond likely requires a conformational change from the synclinal conformation observed in crystal structures (I_s in Figure 5, Panel B) to an antiperiplanar transition state (II in Panel B) that allows the developing p orbital on C2 to align with the W–C3–C4 π system.¹⁰ Between the ground-state conformation and the transition state lies a higher energy conformation (I_a) that approximates the transition state. From this purported geometry, the R substituent on C2 (which was anti to the tungsten in the dihydropyridine I) is brought into a trans configuration relative to C4 (Figure 5, Panel B; III). The rearrangement from IV to V also locks in the trans stereochemistry of the C4–C5 bond (labeled as C3–C4 in the product 1-azatriene). Thus, a single stereoisomer of each azatriene complex is observed. We note that while the ester 8D readily ring-opens, the bulkier analogue 9D is stable, even with prolonged (∼3 days) heating at elevated temperatures (∼80 °C). A plausible explanation for this observation is a steric interaction in the requisite antiperiplanar conformation (I_a in Figure 5) between the geminal dimethyl moiety and the [W] fragment. Approximating the R group of 9D with a t-butyl group, DFT calculations find that the conformation leading to the transition state is 15.7 kcal/mol above the synclinal ground state, whereas for R = Me (i.e., 6D), the energetic cost is only 8.3 kcal/mol (Figure 5). This steric interaction can also be viewed in the crystal structure of 17D, where R = Bn (see Figure 5, Panel C).

Figure 5.

(A) Proposed mechanism for achieving E/E stereochemistry (B) Newman projection showing development of E stereochemistry and steric strain for large R. (C) ORTEP (50% ellipsoids) of 17D and 19D.

Interestingly, 12D and 13D fail to readily ring-open; even after prolonged periods (∼3 days) of elevated heating (∼80 °C), other than starting material, only decomposition products were observed. This observation is thought to be a result of the reduced hyperconjugation that would stabilize the positive charge at C2 of the allyl species III in Figure 5 and, hence, the transition state II.¹⁰ Furthermore, when a 6:1 cdr mixture of 7D and 7P was heated overnight (55 °C), a crude ¹H NMR spectrum showed that only the distal form (7D) ring-opened to form 17D, while the proximal isomer (7P) remained unreacted. Previous studies have demonstrated a thermodynamic preference to orient carbenium carbons distal to the PMe₃.¹⁰ This experiment suggests that the asymmetric unit {WTp(NO)(PMe₃)} stabilizes the transition state (II in Figure 5) of distal DHP complexes significantly more than for the proximal diastereomers.

Attempts to oxidatively decomplex the 1-azatriene complex 18D using CAN, DDQ, or O₂/silica were unsuccessful;^11,27 CAN and DDQ both resulted in observable degradation of the complex with no identifiable free organics, while O₂/silica rendered only starting material.

Several noteworthy comparisons can be drawn to other instances of DHP ring-opening. Zincke first observed this reactivity when combining a secondary amine with a dinitrophenylpyridinium salt. Other instances of azatrienes spontaneously formed from substituted DHP salts involve 1,2-DHPs substituted at C2 with other heteroatoms,²⁸ aryl or alkyne R groups,^29,30 or amino groups³¹ that serve as π donors (vide supra).²⁶ Examples of metal-coordinated azatrienes are rare and almost always involve sigma coordination through a heteroatom. As an example, Cui et al.²¹ combined 2-picoline-N-oxide with a magnesium hydride complex and observed two different ring-opening products, depending on the solvent. These N-oxoazatrienes were bound through the oxygen. Wolczanski et al. demonstrated that pyridine could be opened to a binuclear niobium alkylidine complex.³² Several reports of η⁴-azatrienes have emerged, but they are cyclic and typically involve a larger π system.³³⁻³⁵ The only previous report of an η²-bound azatriene that we are aware of involves the complex MoTp(NO)(DMAP)(N-methyltrifluoropicolinium), which upon hydride reduction to the corresponding DHP and exposure to triflic acid (HOTf), evolves into a Mo-coordinated 1-azatriene analogous to those in Figure 4.²⁴ The CF₃ group of DHP and protonation of nitrogen were shown to be essential for the ring-opening to occur. The lack of azatriene formation from N-acetyled [W]-DHPs remains mysterious. While our own data is ambiguous (vide supra), Piettre and co-workers have made a strong case that sulfonyl groups are more withdrawing than acetyl groups.³⁶ As stated earlier, a more withdrawing sulfonyl group would be better suited for stabilizing the proposed intermediates III and IV (Figure 5A).

Concluding Remarks

Over several decades, our research group has endeavored to demonstrate the ability of an η²-coordinated transition metal (Os, Re, Mo, W) to effect organic transformations of aromatic molecules not accessible by conventional synthetic methods.^3,16,37,38 An impediment to investigating parallel chemistry with linear polyenes has been in the preparation of such complexes as single regio- and stereoisomers. For example, {WTp(NO)(PMe₃)} and the 1-azaheptatriene ligand of 16D could form 12 different isomers (racemic) in which the metal was dihapto-coordinated.⁸ Many of these isomers would likely form with competitive rates and would have similar free energies,^8,16 hence one would need to rely on tedious separation methods or fortuitous solubility differences to obtain a single isomer. This study outlines a possible strategy for forming such linear polyene complexes as single isomers, starting with an aromatic precursor where metal linkage isomerization is facile¹⁶ and then ring-opening to prepare the linear polyene system. Analogous approaches could be envisioned with pyrroles, furans, diazenes, and even cyclic polyenes of the form C_nH_n (e.g., benzene, cycloheptatrienyl, cyclooctatetrene), where initial isomer formation can be minimized.¹⁶ Efforts in our lab continue toward these goals.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.organomet.4c00108.

• Synclinal and antiperiplanar conformations of tBu and Me, and organic ligands of 6D and16D (XYZ)

• ¹H and ¹³C NMR spectra of selected compounds and crystallographic information for compounds 5D–19D (PDF)

The authors declare no competing financial interest.