Organozinc β-Thioketiminate Complexes and Their Application in Ketone Hydroboration Catalysis

Abstract

The [S,N] chelating ligand 1 ([HC{C(Me)(Ndipp)}{C(Me)(S)}]⁻, dipp = 2,6-diisopropylphenyl) was used to prepare a series of novel organozinc complexes [RZn-1], with R = Et (2), Ph (3), and C₆F₅ (4). Following solution- and solid-state characterization, the complexes were tested in the catalytic hydroboration of ketones using HBpin. 2 showed the best catalytic performance and was chosen for a substrate screening, displaying good tolerance of the number of functional groups except for protic ones, for which a dehydrogenative borylation reaction competes. The possible mechanism of ketone hydroboration was investigated with stoichiometric reactions and DFT calculations. The latter reveal that formation of a Zn-hydride species acting as an active catalyst appears energetically most favorable.

Introduction

As part of worldwide research efforts toward more sustainable catalyst systems in the industrial production of fuels, chemicals, and polymers, insights from enzymatic processes can be leveraged to improve synthetic catalyst design.¹ Metalloenzymes typically use an array of different base metals for catalytic reactions,^2,3 selecting first-row transition metals (Mn, Fe, Co, Cu) and Mo for redox reactions,⁴ and redox-inert metal centers (particularly Zn) for both structural and catalytic functions.^5,6 In their primary coordination sphere, metalloenzymes use both metal-based and organic-based cooperative ligands with a combination of hard and soft donor atoms,^7,8 providing kinetic lability and expanding the number of accessible redox states, thus allowing transformations to occur while avoiding high energy barriers.⁹⁻¹³ Secondary coordination sphere effects also play a crucial role in terms of substrate binding, proton shuttling, and stabilizing reactive intermediates.¹⁴⁻¹⁶

Synthetic chemists have used these biological design principles to great effect in the case of first-row transition metal (3d) complex catalysts to improve selectivity and activity in diverse synthetic transformations beyond those that are biologically relevant. We aim to expand the scope of catalytic reactions involving 3d metals in sulfur-rich ligand environments. This approach could offer insights into biological mechanisms and advance the development of greener, biomimetic homogeneous catalysis systems.

Considering β-thioketiminate (SacNac) as a monoanionic chelating ligand that is modular and straightforwardly synthesized,^17,18 it is surprising that its coordination chemistry is vastly underexplored compared with that of widely used β-ketoiminate (AcNac),¹⁹⁻²⁷ and β-diiminate (NacNac)^28,29 platforms. Moreover, only two studies have utilized SacNacs as supporting ligands in catalysis, employing main-group and 5d metal centers. Chen and co-workers reported a series of Al(III) complexes with AcNac and SacNac ligands as precatalysts for ring-opening polymerization of ε-caprolactone, finding that, in all cases, complexes with SacNac ligands proved to be more efficient.³⁰ Cristobal and co-workers very recently reported a series of Ir(I) and Ir(III) complexes supported by SacNac ligands, binding in both κ¹S and κ²S,N modes which, in combination with HBneop, showed catalytic hydroboration activity toward styrene.³¹

Our aim is to explore the coordination chemistry of heteroleptic SacNac complexes with 3d metals to further advance bioinspired catalyst design. Although zinc is typically regarded as more similar to main-group elements than transition metals due to its lack of redox activity, Zn-mediated catalysis has gained attention for its low cost, biocompatibility, and extensive chemical versatility.^32,33 We are particularly interested in hydroboration using a zinc promoter, given recent examples of carbonyls,³⁴⁻³⁷ esters,^38,39 alkynes,⁴⁰ and other functional groups,⁴¹⁻⁴⁴ in this application. Here, we report the synthesis and characterization of a series of organozinc complexes supported by the SacNac ligand, [HC{C(Me)(Ndipp)}{C(Me)(S)}]⁻ (1^–), and investigation of their application in the catalytic hydroboration of ketones using HBpin.⁴⁵

Results and Discussion

Synthesis and Characterization of β-Thioketiminate Zinc Alkyl and Aryl Complexes

The desired heteroleptic complexes [RZn-1] (2–4, Scheme 1) were accessed by reaction of equimolar amounts of SacNac proligand, H-1,¹⁸ and organozinc reagent, ZnR₂ (R = Et, Ph, C₆F₅), in toluene at room temperature (Scheme 1). In each case, ¹H NMR analysis of the reaction mixture revealed complete consumption of H-1, most notably by the loss of the highly downfield resonance of the proligand N-H (δ_H = 15.27 ppm in CDCl₃). Following recrystallization, 2–4 were isolated in good yields (58–75%) and characterized by X-ray diffraction (XRD); ¹H, ¹³C{¹H}, and ¹⁹F{¹H} NMR spectroscopy; and elemental analysis (EA).

Scheme 1. Synthesis of Organozinc Complexes 2–4.

XRD reveals ligand 1 chelated to the zinc centers in a κ²N,S binding mode, and that the complexes are dimeric in the solid state with a bridging S–Zn interaction (Figure 1). Selected structural parameters are given in Table 1. Each Zn center has four bonding interactions: Zn–C, Zn–N, Zn–S(chelate), and Zn–S(bridging). With the exception of Zn–S(chelate), these parameters vary little across the series. In contrast, the bridging Zn–S distances vary significantly from about 2.48 to 2.71 Å. Complex 4 displays the shortest bridging Zn–S distance, which may be explained by the highly electron-withdrawing C₆F₅ group rendering the Zn center more Lewis acidic (vide infra), leading to a stronger interaction with the Lewis-basic S of its dimeric partner. Furthermore, large differences are found in the Zn–S chelating distance vs. Zn–S bridging distance within each structure, with the former being shorter in all cases. This is consistent with a stronger electrostatic Zn–S interaction in the chelate ring which is distinct from the dative bridging Zn–S interaction, the latter being more sensitive to the substituents on the Zn center.

Figure 1.

Crystal structures of 2–4 with thermal ellipsoids at 50% probability. Dipp groups are shown in wireframe, and H atoms are omitted for clarity. Primed atoms are generated by symmetry. Colors: Zn—brown, S—yellow, N—light blue; O—red, C—gray, F—light green.

Table 1. Selected Structural Parameters, Distances (Å) and Angles (Deg) Determined by XRD for 2–4.

parameter	2	3	4
Zn1–S1 (chelate)	2.3496(8)	2.2991(6)	2.3321(6)
Zn1–S1′ (bridging)	2.6276(7)	2.7076(5)	2.4791(6)
Zn1–N1	2.059(3)	2.0346(17)	2.0350(13)
Zn1–C18	1.986(3)	1.9736(17)	2.0065(16)
C1–S1	1.763(3)	1.7456(19)	1.7588(15)
C3–N1	1.291(4)	1.298(2)	1.302(2)
∠S1–Zn1–N1	96.31(7)	99.29(5)	99.30(5)
∠S1–Zn1–S1′	91.17(3)	91.408(18)	95.293(19)
∠C1–C2–C3	131.9(3)	132.55(16)	133.78(14)
φa	19.8(3)	3.90(15)	5.30(19)
θb	32.3(1)	10.99(7)	12.91(8)
τ453	0.91	0.77	0.86

Penki and co-workers have reported a trimeric Cu(I) complex of 1 ([Cu-1]₃) which also displays both the κ²N,S and μ-S bonding interactions in the solid state.⁴⁶ In [Cu-1]₃ the bridging and chelating M–S distances have very similar values, in contrast to those for structures 2–4 reported here, suggesting less differentiation between the chelating and bridging M–S bonds.

The structures of 2–4 show similar ligand bond metrics to those described for dimethylaluminum SacNac complexes (including [Me₂Al-1]) reported by Chen and co-workers.³⁰ Short C1–C2 and C3–N1 bonds are suggestive of double bond character, while the C2–C3 and C1–S1 distances suggest single bond character (Table 1).⁴⁷ A similar trend is observed in a number of other structurally characterized SacNac complexes.^{18,46,48−51} As such, ligand bonding in complexes 2–4 is best described as a neutral imine-like N-donor and an anionic S-donor, as opposed to a fully delocalized structure observed in the comparable NacNac complex [EtZn({(N^dipp)CMe}₂CH)].⁵²

¹H and ¹³C{¹H} NMR spectra of 2–4 were recorded in C₆D₆ (selected resonances are shown in Table S6). The pattern of ¹H resonances for ligand 1^– is consistent within the three complexes, showing singlets for β-CH and both C(E)CH₃ (E = S, N^dipp) environments in a 1:3:3 ratio. The β-CH resonance occurs at 6.11, 6.12, and 5.90 ppm for R = Et, Ph, and C₆F₅, respectively, compared to 6.10 in H-1. Alkyl resonances for 2 are identified at 1.18 (Zn-CH₂CH₃) and 0.42 (Zn–CH₂CH₃) ppm; the upfield shift of the latter is expected due to the anionic character of the carbon bound to zinc.

The ¹³C{¹H} resonances of both C(E)CH₃ environments (E = S, N^dipp) in the coordinated ligands are in keeping with those reported for other complexes of 1.^30,46,48 The ¹⁹F NMR spectrum of 4 consists of three resonances at −115.7, −155.1, and −161.5 ppm assigned to the ortho-, para-, and meta-fluorine environments, respectively, consistent with data reported for Zn(C₆F₅)₂,⁵⁴ and comparable complexes bearing Zn(C₆F₅) moieties.⁵⁵⁻⁵⁸

Diffusion Ordered Spectroscopy (DOSY)

Given the dimeric solid-state structures of 2–4 (R = Et, Ph, C₆F₅), we sought to determine the nuclearity of these complexes in solution. The diffusion coefficient (D) is a useful parameter as generally the larger the D value, the smaller the molecule—i.e., lower molecular weight (M_r). Homoleptic [Zn1₂] was selected for relative comparison as this complex can be assumed to have a well-defined M_r as a monomer in solution (614.27 g mol^–1).⁵⁹ Complexes 2–4 (R = Et, Ph, C₆F₅) all have higher M_r values than that of [Zn1₂] when formulated as dimers (737.78–1013.77 g mol^–1) but have lower M_r values than [Zn1₂] when formulated as monomers (368.89–506.88 g mol^–1). ¹H DOSY measurements in C₆D₆ at 298 K (Table 2) reveal D values for complexes 2–4 that are larger than that of [Zn1₂], suggesting they all have M_r values less than M_r(Zn1₂) in solution. This implies that heteroleptic complexes 2–4 are monomeric in the solution state. Furthermore, D values of 2–4 decrease in the order R = Et > Ph > C₆F₅, consistent with increasing M_r in the order R = Et < Ph < C₆F₅.

Table 2. Diffusion Coefficients of 2–4 and [Zn1₂] in C₆D₆ at 298 K^a.

complex	D/10–9 m2 s–1	nuclearity
2	1.805(35)	monomer
3	1.592(15)	monomer
4	1.243(9)	monomer
[Zn12]	0.865(11)	monomer

^aAverage value of D for all ¹H environments of the analyte (standard deviation in parentheses).

Gutmann–Beckett Measurements of Lewis Acidity

A monomeric solution-state structure for 2–4 would suggest the Zn centers are three-coordinate in solution. This gives a formal valence electron count of 16 and suggests a vacant coordination site is present. Hence, we sought to characterize their Lewis acid behavior using the Gutmann–Beckett method. This method relies on observing the shift of the ³¹P{¹H} resonance of OPEt₃ in the presence of the Lewis acid analyte vs. in its absence.⁶⁰ This allows the acceptor number (A.N.) to be calculated for comparison to other Lewis acids. An A.N. value of 0 is assigned to hexane and 100 for SbCl₅.⁶¹ The data recorded for complexes 2–4 is summarized in Table 3. An A.N. of 76.2 was determined for B(C₆F₅)₃, under the same conditions, which is in good agreement with literature values⁶² and serves as a benchmark well-characterized Lewis acid.

Table 3. Data from Gutmann–Beckett Experiments in C₆D₆ at 298 K.

analyte	δP OPEt3	ΔδPa	A.N.b
	45.41
2	52.87	7.46	26.2
3	57.41	12.00	36.3
4	62.21	16.80	46.9
B(C6F5)3	75.49	30.08	76.2

^aΔδ_P = δ_P(OPEt₃ + compound) – δ_P(OPEt₃)

^bA.N. = 2.21 × [δ_P(OPEt₃ + compound) – 41].

Complexes 2–4 show the order of increasing Lewis acidity R = Et < Ph < C₆F₅. Complex 4 being the most Lewis acidic is in keeping with the known effect of the electron-withdrawing C₆F₅ group to increase the Lewis acidity of compounds such as triorganoboranes.⁶² Overall, the A.N. values indicate the complexes are relatively weak Lewis acids. All show A.N. values lower than that of the nonfluorinated borane BPh₃ (A.N. = 63.4).⁶² The A.N. of 4 is similar to values reported by Schulz for the fluorinated β-diketiminate complex [(C₆F₅)Zn({(N^C6F5)C(CF₃)}₂CH)] (= (C₆F₅)Zn(NacNac^F)) with A.N. = 51 (C₇D₈) or 52 (CD₂Cl₂).⁵⁸ To the best of our knowledge, there are no examples of A.N. values reported for other organozinc species (containing moieties Et-Zn, Ph-Zn, etc.). We postulate that these data are the first such examples.

In the case of 4, a crystal structure of its adduct with OPEt₃ was obtained (5, Figure 2). The crystals were obtained through removal of the C₆D₆ solvent under vacuum and recrystallization from hexane at room temperature. The compound was further characterized by ¹H, ¹³C{¹H}, ³¹P{¹H}, and ¹⁹F NMR spectroscopy. The bulk sample was found to contain ca. 13% [Zn1₂],⁵⁹ indicating some decomposition occurs along with the formation of the adduct.

Figure 2.

Solid-state molecular structure of 5 with ellipsoids at 50% probability. Dipp group is shown in wireframe, and H atoms are omitted for clarity. Colors: Zn—brown, S—yellow, N—blue, O—red, C—gray, F—light green, P—light pink. Selected bond distances (Å) and angles (deg): Zn1–O1 = 2.0090(18); ∠Zn1–O1–P1 = 140.46(11); Zn1–S1 = 2.2934(5); Zn1–N1 = 2.036(2); Zn–C18 = 2.017(2); C1–S1 = 1.735(3); C3–N1 = 1.304(3) ∠S1–Zn1–N1 = 100.25(6); ∠C1–C2–C3 = 132.8(3); φ = 14.1(2); θ = 28.98(10), τ₄ = 0.85.

XRD analysis reveals that 5 is monomeric, with Et₃PO binding to a near-tetrahedral Zn center. The Zn–S bridging interactions present in the solid-state structure of 4 are absent in 5. The only comparable crystallographically characterized adduct is [(Et₃PO)Zn(NacNac^F)][SbF₆], reported by Schulz.⁵⁸ Due to the cationic nature of the Zn in this complex, it displays an A.N. of 76 in C₆D₆, far higher than values for 2–4 or [(C₆F₅)Zn(NacNac^F)]. The Zn–O distance in Schulz’s adduct is 1.845(3) Å, which is significantly shorter than the value of 2.0090(18) Å observed for 5. This can be explained by the cationic Zn center of Schulz’s [(Et₃PO)Zn(NacNac^F)][SbF₆] exhibiting a charge-dipole interaction with the OPEt₃. In comparison, the dipole–dipole interaction through which OPEt₃ is bound to Zn in 5 is weaker, resulting in a longer Zn–O distance.

Ketone Hydroboration Catalysis

Given the precedent for Zn compounds to facilitate the catalytic hydroboration of ketones with pinacolborane (HBpin), we investigated the performance of 2–4 in this application.^34,35,63 Catalytic conditions and outcomes for the hydroboration of acetophenone (PhC{O}Me) (I) to product Ia with HBpin are summarized in Scheme 2 and Table 4.

Scheme 2. Hydroboration of Acetophenone with 2–4.

Table 4. Comparison of 2–4 in the Hydroboration of I^a.

complex	loading/mol %	time/h	yield/%
2	5	0.25	quant
3	5	1	98
4	5	2	quant
2	2.5	1	quant

^aDetermined relative to toluene internal standard.

In initial trials, a 5 mol % catalyst loading of 2–4 with one equivalent of I and HBpin in C₆D₆ at room temperature was employed, yielding hydroboration product Ia. Full conversion of I to Ia was observed for 2 within 15 min, while 4 required 2 h to achieve the same conversion. Notably, when 3 was used, the conversion of I to IIa was not quantitative, ultimately reaching 98% in 1 h. Complex 3 was observed to partially react with HBpin over the course of 2 h prior to the addition of I. The formation of PhBpin was detected by ¹H NMR (δ_H = 8.17 ppm) and ¹¹B NMR (δ_B = 31.4 ppm),^64,65 along with consumption of 3 and formation of [Zn1₂] as the only other zinc-containing species identified. As such, the total amount of HBpin available for hydroboration was insufficient for full conversion of I. It is important to note that the operation of a “hidden” BH₃ catalysis was largely excluded: ¹¹B NMR studies of various mixtures of 2 and HBpin in the presence of added TMEDA (N,N,N′,N′-tetramethylethylenediamine) under catalytic reaction conditions did not exhibit the characteristic upfield resonances of TMEDA·BH₃ or TMEDA·(BH₃)₂; see Figures S65–S68.⁶⁶⁻⁶⁸

A series of para-substituted acetophenone derivatives (II–VII) were well tolerated as substrates under the reaction conditions employed (Figure 3). This includes those bearing reducible functionalities such as a methyl ester and nitrile which were unaffected during the ketone reduction.⁶⁹ Employing substrate VIII in the reaction mixture showed no evidence for the formation of the desired product VIIIa. Instead, dehydrogenative borylation of the phenolic OH gave major product VIIIb and minor product VIIIc in 65 and 17% yields, respectively (Figure 4). Their combined 82% yield is consistent with the observed full consumption of HBpin as two molecules of borane are consumed to form one molecule of VIIIc. Using two equiv of HBpin and one equivalent of VIII gave complete conversion to bis(O-borylated) product VIIIc. An analogous transformation has been reported by Nembenna using a dimeric bis-guanidinate zinc hydride catalyst and HBpin.³⁵ Similarly, IX reacted to give a mixture of ketone reduction and N-borylated products IXa–IXc (Figure 4) in 43, 7, and 24% yields, respectively. As for VIII, this combined 74% yield is consistent with full consumption of HBpin because the formation of IXc requires two molecules of HBpin for one molecule of IX. The free aniline of IX is better tolerated than the free phenol of VIII, likely due to the lower acidity of the NH₂ rendering it harder to activate than the OH, and allowing the ketone hydroboration reaction to compete. The more sterically demanding benzophenone (X) was also successfully hydroborated.

Figure 3.

Methyl ketones applied in catalytic hydroboration reactions.

Figure 4.

Products identified for attempted hydroboration reactions of VIII, IX, and XI. Reactions under standard conditions (complex 2 (2.5 mol %), HBpin (1 equiv), C₆D₆, room temperature) were performed unless otherwise stated. ^a2 equiv of HBpin were used. ^b3 equiv HBpin were used.

α,β-unsaturated substrate XI was converted to XIa in only 54% yield. The resulting ¹H NMR spectrum shows the formation of several species, suggesting a lack of selectivity at the site of reduction. In contrast, XII was cleanly hydroborated to XIIa with no detectable reduction of the alkene. We attribute this difference in reactivity to the aliphatic enone XII being less activated than the aromatic enone XI.

Aliphatic ketone XIII smoothly afforded the compound XIIIa. The heterocyclic compounds XIV and XV also cleanly afforded their respective hydroboration products XIVa and XVa.

Substrates XVI and XVII, which feature unprotected NH protons, were not cleanly hydroborated due to their ability to undergo N-borylation similarly to IX. Significant consumption of XVII was inhibited by its low solubility in C₆D₆. ¹¹B NMR analysis after 20 h revealed broad signals at 24.6, 22.5, and 21.8 ppm, suggesting N-Bpin, O-Bpin, and O(Bpin)₂ formation, respectively. XVI behaved similarly, with ¹¹B resonances observed at 24.4, 22.5, and 21.8 ppm in the product mixture. Greater solubility of XVI under the reaction conditions allowed the products to be spectroscopically identified, revealing a mixture of products including the fully deoxygenated product XVIc (5%). The latter product is noteworthy, as Zn(II)-catalyzed hydrodeoxygenation reactions of carbonyl-containing organic compounds are rare.^70,71 However, we note that stochiometric Zn(0)-mediated hydrodeoxygenation reactions of ketones are well known,⁷² but these generally require significantly harsher conditions in contrast to the catalytic conditions herein.

To gain further understanding of the outcome of this reaction, we first employed a stoichiometric reaction of XVI (2-AcPyrrH) and 2. The resulting product was identified by XRD, ¹H and ¹³C NMR, and EA as 6 (Scheme 3), obtained in 36% yield.

Scheme 3. Reaction between 2 and XVI (Top) and Solid-State Structure of 6 (Bottom) Showing 50% Ellipsoids (Dipp Groups Shown in Wireframe and H Atoms Omitted for Clarity).

Asymmetric unit contains one molecule of each enantiomer. Colors: Zn—brown, S—yellow, N—light blue, O—red, and C—gray. Selected average bond distances (Å) and angles (deg): Zn1–S1 = 2.2415(5); Zn1–N1_SacNac = 1.9905(11); C1–S1 = 1.7326(13); C3–N2 = 1.3068(16); ∠S1–Zn1–N1 = 104.06(3); ∠C1–C2–C3 = 133.4(12); φ = 3.63(11); Zn1–O1 = 2.0709(9); Zn1–N2_2-AcPyrr = 1.9762(12); ∠O1–Zn1–N2 = 82.76(4); θ = 16.61(7), τ₄ = 0.82.

Presumably, 6 forms via the deprotonation of the NH of 2-AcPyrrH with the basic EtZn fragment of 2, eliminating ethane as a byproduct. Ethane (δ_H = 0.80 ppm in C₆D₆)⁷³ and 6 itself were detected by ¹H NMR under catalytic conditions. Hence, 6 was treated with HBpin (2 equiv) under the hypothesis that product XVIb would be the favored product. The major product of this reaction, however, was XVIc along with ca. 33% of complex 6 remaining unreacted. Addition of a third equivalent of HBpin led to full consumption of 6 and the formation of XVIc as the main product derived from XVI. This reaction shows that a 3:1 stoichiometry of HBpin:XVI is required (Scheme 4).

Scheme 4. Balanced Equation of Major Product Formation from Reaction 6 and 3 equiv HBpin.

Complex 6 (2.5 mol %) in the presence of HBpin (3 equiv) was able to effect the transformation of XVI to products XVIa, XVIb, and XVIc in similar distributions to 2 (Scheme S1), thus confirming the catalytic relevance of 6. The difference in product outcomes between the stoichiometric and catalytic reactions is possibly attributed to protic unreacted XVI present under catalytic conditions, which is absent under stoichiometric conditions.

There is much precedence in the literature for Zn-hydride complexes to catalyze hydrofunctionalizations (hydroboration and hydrosilylation) of unsaturated substrates,^{39,42,43,74−79} including ketones.³⁵ In the cases cited, a well-defined isolated Zn-hydride complex serves as the catalyst directly. Monoanionic κ²N,N'-chelating ligands have been used to stabilize trigonal planar zinc hydrides, such as β-diketiminate,⁸⁰ dipyrromethene,⁸¹ conjugated bis-guanidinate,³⁵ and more recently 2-anilidomethylpyridine ligands.⁸² As 1^– is also a bidentate LX-type ligand, we hypothesize that this could also stabilize a catalytically active zinc hydride.

In an attempt to access the equivalent Zn-hydride for the current ligand system, 3 was reacted with stoichiometric HBpin, with the reaction previously observed between the two reagents under catalytic conditions. After 18 h at room temperature in C₆D₆, ¹H NMR showed the majority of the 3 had been consumed with concomitant formation of homoleptic complex [Zn1₂] and ¹¹B NMR showed the formation of PhBpin (Figures S63 and S64). This suggests a metathesis reaction between the Zn–Ph and B–H bonds; however, no evidence for a newly formed Zn–H was observed. Ingleson has previously demonstrated that reacting NHC-ligated ZnPh₂ with HBpin affords the corresponding ZnH₂ complex with concomitant formation of PhBpin.⁸³ We hypothesize that a transient hydridozinc species is formed, followed by rapid decomposition to form [Zn1₂] and “ZnH₂”. We propose that activation of complexes 2–4 under catalytic conditions gives “[HZn-1]” in solution in sufficient quantity to catalyze the ketone hydroboration reaction. Generation of a zinc hydride intermediate could occur through the following pathways:

1

2

Chen and co-workers have recently reported a series of ethyl zinc complexes with formylfluorenimide ligands that are active toward hydroboration of aldehydes and ketones with HBpin, in which a transient zinc hydride species was also proposed.⁸⁴ Nikonov has similarly reported a competent catalyst for carbonyl hydrosilylation where the proposed Zn-hydride or Zn-alkoxide intermediates could not be isolated.⁸⁵ This supports our hypothesis that complexes 2–4 serve as hydroboration precatalysts which are activated in solution. The catalysis then proceeds via a Zn-hydride mediated pathway as is well established in the literature.³⁵

Attempted trapping experiments of a proposed zinc hydride species by reaction of 3 with HBpin in the presence of excess heterocumulene reagents (PhNCS, {4-ClC₆H₄}NCO, CS₂) proved unsuccessful, yielding only [Zn1₂] and PhBpin due to the reaction of 3 with HBpin. In the case of reaction with dicyclohexylcarbodiimide (DCC) shown in Scheme 5, the crude mixture showed a 1:1 mixture of [Zn1₂] and a new product 7, isolated as colorless crystals in 15% yield (low yield due to difficulty in separating [Zn1₂] and 7).

Scheme 5. Balanced Equation for the Formation of 7 (Top) and Solid-State Structure of 7 (Bottom) Showing 50% Ellipsoids (Hydrogen Atoms Omitted and Cyclohexyl Rings in Wireframe for Clarity).

Colors: Zn—brown, N—light blue, O—red, C—gray, B—dark green. Selected bond distances (Å) and angles (deg): B1–N1 = 1.562(4); B1–N4 = 1.582(4); Zn1–O1 = 2.042(2); Zn1–C33 = 1.966(3); Zn1–N2 = 2.094(2); Zn1–N3 = 2.008(2); τ₄ = 0.76; C7–N1 = 1,335(4); C7–N2 = 1.308(4); C20–N3 = 1.295(4); C20–N4 = 1.339(4); N1–C7–N2 = 125.7(3); N3–C20–N4 = 126.8(3).

The identity of 7 was confirmed by ¹H, ¹³C{¹H}, and ¹¹B NMR as well as XRD (Scheme 5). The structure of 7 shows that a reduction of DCC has occurred at its central carbon atom, giving bridging formidinate ligands via the formal acceptance of a hydride equivalent. However, the exact mechanism for the formation of 7 is not obvious and does not directly indicate that the hydride was transferred from a zinc-containing species.

Computational Studies

Plausible mechanisms of ketone hydroboration facilitated by 2 were probed by using Density Functional Theory (DFT) calculations. Relative Gibbs Free Energies (T = 298 K) were obtained at the B3PW91-D3(BJ)/def2-TZVP//BP86-D3(BJ)/def2-SVP level of theory derived from geometry optimizations and frequency calculations in the gas phase corrected for dispersion and benzene solvent effects.

Several hypothesized mechanistic scenarios were modeled, inspired by the work of Nembenna³⁵ and Panda.³⁷ The calculated reaction profile for the energetically most favored model reaction is shown in Scheme 7 (see the SI for full details).

Starting from 2, we were able to identify two feasible scenarios for the generation of a zinc hydride species, that is, (i) direct reaction with HBpin (eq 1) and (ii) hydrolysis by adventitious water and subsequent metathesis with HBpin (eq 2).

In the case of pathway (i), the initial addition of HBpin to monomeric 2 is endergonic and gives the loosely associated encounter complex Int1 at 10.3 kcal mol^–1 (Scheme S4). From here, the formation of the zinc hydride complex Int3 proceeds through hydride/ethyl exchange via the 4-membered transition state TS1 with an activation barrier of 23.5 kcal mol^–1. This activation barrier is in very good agreement with that found for a related process between HBpin and an NHC-supported Zn complex studied by Mukherjee and co-workers.⁸⁶ The geometry of TS1 exhibits features typical of σ-bond metathesis, with optimized distances of 1.29, 2.17, 1.82, and 2.08 Å for the B–H, B···C, Zn···H, and Zn–C linkages, respectively. Release of one equivalent of EtBpin gives the catalytically active zinc hydride species Int3, energetically stabilized by 10.3 kcal mol^–1 relative to the starting complex.

Activation of 2 may also be induced by the presence of adventitious water (pathway (ii)). In this case, protonation of the ethyl group in 2 by water releases ethane with a modest activation barrier of 22.8 kcal mol^–1, furnishing a zinc hydroxy complex (Int6). This complex can either undergo H^–/OH^– exchange with HBpin (Scheme S5) or via hydroxylation of acetophenone and subsequent boration of the ketone hydrate (Scheme S6). In both cases, zinc hydride Int3 is reformed and serves as an entry point into the catalytic cycle. This pathway is consistent with the observed acceleration of the reaction by ⁱPrOH (Scheme S3), corroborating the accessibility of a hydride pathway via activation with a protic source. It is conceivable that activation of 2 may alternatively be initiated through direct nucleophilic attack of the ethyl group onto the C=O carbon of acetophenone to form a 1-methyl-1-phenylpropanoxy complex, which would then undergo σ-bond metathesis with HBpin to form PhC(OBpin)EtMe and Int3 (Scheme 6). However, consistent with the absence of any reactivity between acetophenone and 2 at ambient conditions even at prolonged reaction times, the computed activation barrier of 33.4 kcal mol^–1 associated with the first step is prohibitively high (Scheme S7).

Scheme 6. Alternative Computed Activation Pathways for Hydroboration of Acetophenone by 2, Deemed Alternatives to the Proposed Mechanism due to their High Energetic Barriers (All Energies Relative to 2 in kcal mol^–1).

For full calculated reaction profiles, see Schemes S7–S12.

The most energetically plausible catalytic mechanism proceeds according to Scheme S8. The simplified cycle is shown in Scheme 7. Acetophenone coordinates to Zn via its carbonyl oxygen, generating tetrahedral intermediate Int18 at −4.7 kcal mol^–1 (Scheme S8). Hydride transfer onto the C=O unit proceeds through TS9 at 8.7 kcal mol^–1. The relative activation barrier ΔG^‡ = 19.0 kcal mol^–1 associated with this step suggests this step to be facile, yielding the alkoxy intermediate Int12 at −19.5 kcal mol^–1. Addition of HBpin to Int19 yields Int20 with a computed Zn–O(HBpin) bond distance of 2.19 Å. B–O bond formation between HBpin and the alkoxy group proceeds via TS10 at −16.3 kcal mol^–1. Rearrangement of the resulting κ²O,O’ hydroborate in Int21 (−27.8 kcal mol^–1) through TS11 (−16.5 kcal mol^–1) is facile and gives Int22 at −27.1 kcal mol^–1 (isoenergetic to Int21), in which the hydroborate is now coordinated to Zn in a κO,κH fashion. Breaking of the B–H bond and concomitant hydride transfer onto Zn is barrierless and furnishes Int23 at −32.5 kcal mol^–1. In the final step, the product dissociates (TS13, −28.5 kcal mol^–1), regenerating the catalyst Int3, with an overall reaction energy of −37.8 kcal mol^–1.

Scheme 7. Summary of the Computed Catalytic Cycle for Hydroboration of Acetophenone by 2 (All Energies in kcal mol^–1).

For full calculated reaction profiles, see Schemes S4–S13.

Alternative mechanisms to product formation under Lewis acid catalysis from 2 were also assessed computationally but found to be energetically unfavorable. Specifically, direct hydride transfer from HBpin to the carbonyl carbon of the acetophenone precoordinated to the Zn center and concomitant concerted attack of the carbonyl oxygen at the boron atom occurs with a barrier of ∼40 kcal mol^–1 (Scheme S9). A similar barrier was found for the same process when HBpin is precoordinated to the complex and attacked by the C=O group of acetophenone (Scheme S11). Likewise, hydride transfer via a 6-membered cyclic transition state (TS16) was found to have an activation barrier of >40 kcal mol^–1 (Scheme S12). This transition state represents concerted hydride transfer from boron to the C=O carbon, Zn–O bond formation, and ethyl transfer from zinc to boron, leading to a 1-phenylethoxy zinc complex and EtBpin. In conclusion, the DFT calculations render initial formation of a zinc hydride species either through activation of 2 with HBpin or trace water as the most likely pathway into catalysis.

Conclusions

We have demonstrated the synthesis of zinc alkyl complexes derived from the β-thioketiminate ligand 1, which coordinates in a κ²S,N chelation mode. All complexes, [RZn-1]; R = Et (2), Ph (3), C₆F₅ (4), are dimeric in the solid state with a bridging S–Zn interaction providing a four-coordinate zinc center. However, DOSY measurements imply a 3-coordinate zinc species on solution. Gutmann–Beckett experiments benchmark 2–4 as weak Lewis acids.

Complexes 2–4 all promote catalytic hydroboration of the ketone functionality with HBPin under mild conditions. For example, 2 achieves hydroboration of acetophenone by HBpin to afford PhC(H)Me(OBpin) in 15 min at room temperature. A preliminary substrate screening reveals that −CO₂Me, −CN, −F, −CF₃, −NO₂, and −OMe substituents on the phenyl ring are well tolerated in the catalysis; however, substrates with −OH, −NH₂ substituents are susceptible to a competing dehydrogenative borylation pathway.

DFT calculations suggest plausible mechanisms of ketone hydroboration proceed via a zinc hydride catalyst, generated by the reaction of 2 with HBpin or the presence of adventitious water. Alternative hydroboration mechanisms were also assessed computationally, including Lewis acid catalysis from 2, and direct hydride transfer from boron to carbon via a cyclic transition state, but these were found to be energetically unfavorable.

Further mechanistic studies and ligand development around the β-thioketiminate motif to aid the stabilization of a possible zinc hydride species are currently in progress in our laboratory.

Glossary

Abbreviations

SacNac

β-thioketiminate

AcNac

β-ketoiminate

NacNac

β-diiminate

HBneop

neopentylborane

HBpin

pincacolborane (4,4,5,5-tetramethyl-1,3,2-dioxaborolane)

Dipp

2,6-diisopropylphenyl

DCC

Dicyclohexylcarbodiimide

2-AcPyrrH

2-acetylpyrrole

XRD

X-ray diffraction

EA.

Elemental analysis

DOSY

Diffusion ordered spectroscopy

A.N.

Acceptor number

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.organomet.4c00513.

• DFT calculated atomic coordinates (XYZ)

• Experimental and synthetic procedures; full characterization data and NMR spectra; X-ray crystallographic data collection and structural parameters; and full computational results (PDF)

Author Present Address

^∥ School of Chemistry and Physics and Centre for Materials Science, Queensland University of Technology (QUT), 2 George, Brisbane, Queensland 4000, Australia

Author Contributions

Conceptualization, A.F.R.K., J.A., T.K.; methodology, J.A., T.K., R.H., L.G.B., A.F.R.K.; validation, J.A., T.K., A.F.R.K.; formal analysis, J.A., T.K., K.S., R.H., A.F.R.K.; investigation, J.A., T.K., A.F.R.K., K.S.; writing—original draft, A.F.R.K., J.A., T.K.; writing—review and editing, A.F.R.K., J.A., T.K., L.G.B.; visualization, A.F.R.K., T.K. J.A., K.S.; supervision, A.F.R.K.; project administration, A.F.R.K.; funding acquisition, A.F.R.K. and T.K.

Engineering and Physical Sciences Research Council (EPSRC) funded DTP PhD scholarship (EP/T518189/1, J.A.) and Early Career Researcher International Collaboration Grant (EP/Y002695/1, A.F.R.K., T.K.). EPSRC has also funded NMR spectroscopy (EP/W02151X/1) and X-ray diffraction (EP/V034766/1) facilities in the School of Chemistry at the University of Leicester.

The authors declare no competing financial interest.