Traceless Acetal-Directed Catalytic Hydrosilylation of Propargyl Acetates Harnessing the π-Acidic Catalyst

Abstract

Traceless acetal-directed, α-specific syn-hydrosilylation of propargyl alcohols has been developed, enabling a synthesis of β-silyl allylic alcohols. An introduction of inexpensive, readily accessible acetal as a traceless, two-atom tether directing group (DG), along with a π-acidic catalyst, facilitates the proximal, α-selective syn-hydrosilylation of a broad spectrum of primary to tertiary propargyl alcohols. Notably, the utilization of a highly fluorinated, π-acidic Rh(I)/P(C₆F₅)₃ catalyst allows rapid cyclizing syn-addition of a silicon–metal species across a C–C triple bond via a strong M-π interaction. A postmodification of the resulting cyclic silyl acetal not only removes the DG, rendering it traceless, but also introduces a functional group to the silicon moiety, enhancing the versatility and utility of the products.

Graphical Abstract

INTRODUCTION

Vinylsilanes are stable and virtually nontoxic.^1–4 Their significance lies in their role as crucial building blocks for the synthesis of bioactive molecules and functional materials.^1,2,4 The versatile nature of vinylsilanes allows for either independent or simultaneous conversion of both the alkene and silyl moieties into several important functional groups. These functional group (FG) transformations encompass a wide range of reactions, including stereoselective electrophilic substitution,^5,6 cycloaddition,⁷ ene reaction,⁸ oxidation,^9–11 sigmatropic rearrangement,^12,13 and a variety of organometallic reactions (e.g., Heck reaction,¹⁴ Denmark–Hiyama cross-coupling,^4,15 Murai reductive ortho-alkylation,¹⁶ hydroester-ification,¹⁷ and olefin metathesis^18–20). In particular, β-silyl allylic alcohols (i.e., proximal hydroxy vinylsilanes) have served as important latent α-hydroxy ketones and multisubstituted olefins through oxidation^9–11 and cross-coupling^4,15 stereoselective electrophilic substitution reactions,^5,6 respectively.

Despite significant advancements in transition-metal-catalyzed alkyne hydrosilylation to access such synthetic building blocks, alkyne hydrosilylation performed on unsymmetrical alkynes often results in formation of an inseparable mixture of regio- and stereoisomers. In particular, intramolecular α-regioselective hydrosilylation of propargylic hydrosilyl ethers 1 remains challenging. It simply arises from the instability of putative oxasilacyclobutanes 2 via 4-exo-dig cyclization instead of giving 5-endo-dig cyclization product 3 (Figure 1a).^16,21 To address this issue, Fürstner achieved innovative Ru-catalyzed, intermolecular α-selective anti-hydrosilylation of propargylic alcohols 4 to give α-hydroxy vinylsilanes 5 (α/β = 60:40 to 100:1) through the interligand interactions (Figure 1b).²² Tomooka and Denmark independently reported Pt-catalyzed, α-selective syn-hydrosilylation of propargylic alcohol derivatives by introducing a dimethylvinylsilyl and siloxane-directing group (DG) to the substrate, respectively, to improve the proximal α-selectivity,^23,24 in contrast to the preferential distal functionalization in most intermolecular hydrosilylation (Figure 1c).²⁵ The former approach employs elevated temperature conditions for primary and secondary propargyl alcohols, while the latter experiences inefficient installation of the siloxane DG, which is mainly associated with the use of a large excess of tetramethyldisiloxane reagent, a requisite purification step, and a limited scope (primary propargylic alcohol).

Figure 1.

Hydrosilylation of propargyl alcohols. FG = functional group, DG = directing group. FG = functional group.

Herein, we report our approach to α-selective syn-hydrosilylation of propargyl alcohols using a traceless acetal DG and a π-acidic Rh(I)/P(C₆F₅)₃ catalytic system to provide β-silyl allylic alcohols 13 (Figure 1d). The design aspects of our traceless DG strategy are 4-fold: (1) an efficient preparation of an acetal DG from inexpensive and readily accessible esters via Ir-catalyzed ester hydrosilylation without the need for purification, (2) the development of a two-atom, acetal DG [cf., siloxane (Si–O) in 8] which conceptually delivers hydrogen and a silicon moiety to unsymmetrical alkyne in a regio- and stereoselective manner via 6-exo-dig cyclization (cf., 4-exo-dig cyclization with propargylic hydrosilyl ethers 1),^26–33 (3) the use of a π-acidic catalyst for establishing a strong M-π interaction which enables rapid syn-silyl metalation (int-I), leading to an expansion of the scope of propargylic alcohols by encompassing all substrate classes—primary, secondary, and tertiary propargylic alcohols and heterocycle-containing alkynes, and (4) postmodification of the resulting silacycles for the facile removal of the acetal DG along with the concomitant diversification of the silyl group in a single reaction vessel.

RESULTS AND DISCUSSION

Given the limited literature on the α-hydrosilylation of tertiary alcohols vis-à-vis primary and secondary alcohols, we selected a model substrate to investigate the potential impact of steric hindrance on α-versus β-hydrosilylation. Hydrosilyl acetal 11a holding the two-atom tether was conveniently prepared through chemoselective Ir-catalyzed ester hydrosilylation over alkyne hydrosilylation within 10a in near quantitative yield without requiring purification (Table 1).²⁶ Specifically, under the catalytic conditions developed, the internal alkyne adjacent to the ester remained intact. With 11a in hand, we set off to survey the catalytic conditions to effect α-selective alkyne hydrosilylation. Unfortunately, the various Ir catalysts that we examined gave lower yields. Next, a range of supporting ligands, including phosphines, phosphite, pentamethylcyclo-pentadienyl (Cp*), and tris(pyrazolyl)borates (Tp), were examined with variations in their electronic and steric capabilities in the presence of a rhodium precatalyst (Table 1). This study showed that the use of a rhodium catalyst, assembled through the combination of the Rh^I precatalyst and highly fluorinated phosphine ligand P(C₆F₅)₃, resulted in an excellent yield (92%) of 12a with excellent regio- and stereoselectivity (entry 6).

Table 1.

Evaluation of the Alkyne Hydrosilylation Catalyst^a

entry	supporting ligand	yield (%)b	α/β c	E/Z d
1	PPh3	81	only α	only E
2	P(4-MeOPh)3	84	only α	only E
3	PPh2Et	83	only α	only E
4	JohnPhos	84	only α	only E
5	PCy3	78	only α	only E
6	P(C6F5)3	92	only α	only E
7	P(OPh)3	77	only α	only E
8	Cp*−e	90	only α	only E
9	Tp(CH3)2K	82	only α	only E
10	TpCF3,PhNa(THF)	88	only α	only E
11	Tp(CF3)2Na(THF)	85	only α	only E

^aConditions: (i) 10a (0.20 mmol), THF (1 M); (ii) supporting ligand (1.2 mol % for phosphine ligands, 0.4 mol % for Cp*⁻ and Tp ligands), THF (0.4 M).

^bDetermined by ¹H NMR spectroscopy utilizing an internal standard (mesitylene).

^cDetermined by GC/MS analysis and ¹H NMR spectroscopy.

^dDetermined by NOE experiments.

^e[RhCp*Cl₂]₂ was used. TMS = trimethylsilyl.

To gain a better understanding of the high efficiency observed in the reaction with the highly fluorinated phosphine P(C₆F₅)₃, in situ ¹H NMR spectroscopy was employed to monitor the progress of the hydrosilylation reactions of 11a with the ligands listed in Table 1. As shown in Figure 2a, the rate of hydrosilylation was largely correlated with the electronic nature of the ligands, with electron-poor ligands showing faster reaction kinetics than their electron-rich counterparts. Generally, the rate order followed the trend: triaryl phosphine > monoalkyldiaryl phosphines > dialkylmonoaryl phosphines > triaryl phosphite > trialkyl phosphines. A similar trend was observed for the electronic properties of the Tp ligand series: [Tp(CF₃)₂]⁻ > [TpCF₃,Ph]⁻ > [Tp(CH₃)₂]⁻. In particular, the electron-poor, highly fluorinated phosphine ligand P(C₆F₅)₃ exhibited a substantially faster kinetic profile. First-order kinetic plots of ln ([M₀]/[M_t]) versus time for the hydrosilylation showed that the reaction with P(C₆F₅)₃ proceeded nearly 5 times faster than with tris-(pentaprotiophenyl)phosphine P(C₆H₅)₃ (Figure 2b).³² The results suggest that the electron-withdrawing nature of the highly fluorinated P(C₆F₅)₃ ligand enhances the Lewis acidity of the rhodium center. Consequently, the π-acidic rhodium center augments the Rh-π interaction, which in turn facilitates rapid silyl-rhodation and ultimately accelerates catalytic turnover. Overall, a π-acidic Rh(I)/P(C₆F₅)₃ catalyst is synthetically advantageous for acetal-directed, α-specific syn-hydrosilylation of propargyl acetates, enabling shorter reaction times and higher yields.

Figure 2.

(a) Kinetic profiles of the alkyne hydrosilylation of 11a with ligands. The data sets were obtained by in situ ¹H NMR spectroscopy in C₆D₆ at 20 °C using an internal standard (mesitylene). 11a (0.20 mmol), supporting ligand (1.2 mol % for phosphine ligands, 0.4 mol % for Cp*⁻ and Tp ligands), THF (0.4 M). [RhCp*Cl₂]₂ (0.2 mol %) used. (b) First-order kinetic plots of ln ([M₀]/[M_t]) versus time at 25 °C with different ligands. M₀: initial concentration, M_t: concentration at a given time.

Based on our preliminary kinetic studies along with literature precedents,^32,33 the plausible overall mechanisms for the three-step sequence involving (i) Ir-catalyzed hydrosilylation of esters, (ii) alkyne hydrosilylation harnessing the π-acidic Rh catalyst, and (iii) the activation–functionalization of silacycle 12 are depicted in Figure 3. First, Brookhart’s a binuclear silylene-bridged iridium dimer enables chemoselective hydrosilylation of esters 10, affording hydrosilyl acetals 11.²⁶ While a modified Chalk–Harrods mechanism is generally accepted in alkene/alkyne hydrosilylation chemistry, our preliminary mechanistic investigations failed to identify a Rh–H species using extensive ¹H NMR spectroscopy. Instead, we propose an oxidative addition of a Si–H bond in hydrosilyl acetals 11 to the Wilkinson-type, electron-poor rhodium catalyst, followed by reductive elimination of HCl, generating the requisite π-acidic Rh^I–silane int-I. The establishment of a strong rhodium-π interaction between a Rh^I–Si moiety and a C–C triple bond within int-I, coupled with rapid silyl rhodation, is key to the successful rapid cyclizing, 6-exo-syn alkyne hydrosilylation to produce the vinylsilyl rhodium species int-II. A subsequent bimolecular rhodium-π interaction, followed by σ-bond metathesis within int-II, gives the product (E)-dioxasilinanes 12 and regenerates int-I, completing the catalyst cycle. To assess the potential protonation of int-2 by HCl, we introduced a stoichiometric amount of base, such as Hünig’s base or NaHCO₃, to the hydrosilylation reaction mixture. However, neither base significantly affected reaction rates or yields compared with the base-free reaction. While a protonation mechanism remains plausible, these results further support the proposed metathesis mechanism. Upon addition of an anionic external FG to 12, a putative, penta-coordinated silicon species int-IV is formed. The succeeding irreversible ring-opening fragmentation of int-IV produces β-silyl allylic alcohols 13 upon protonation with acid, along with the elimination of a corresponding aldehyde which can further react with the anionic FG. This approach not only removes a DG but also diversifies the newly introduced silyl moiety in a single pot, improving the scope of organosilanes.

Figure 3.

Proposed mechanisms for traceless acetal-directed, proximal, α-selective syn-hydrosilylation of propargyl alcohols harnessing the π-acidic Rh catalyst.

With the optimized conditions in hand, the impact of acetal tethers on cyclization was studied (Table 2). Sterically and electronically tuned hydrosilyl acetals (11a–11f) were accessed via Ir-catalyzed ester hydrosilylation and evaluated for alkyne hydrosilylation. Consistent with our findings in the context of arene C–H silylation,^28,30 the sterically more accessible esters 10a–10d (R = H, Me, α-chloro, ⁱPr) smoothly underwent sequential Ir and Rh catalysis, involving ester and alkyne hydrosilylation, respectively, leading to the formation of (E)-dioxasilinanes 12a–12d. Ester hydrosilylation of sterically demanding ester 10e (R = ^tBu) to afford 11e exhibited slower kinetics (12 h), but the rate of alkyne hydrosilylation of 11e was comparable to that of substrates 11a–11d (ca. 0.5 h). While electron-deficient trifluoromethyl ester 10f underwent ester hydrosilylation at significantly slower rates (48 h), alkyne hydrosilylation was achieved in 1 h, yielding 12f. Together, the π-acidic Rh catalyst facilitated fast cyclization of all types of hydrosilyl acetals with excellent regio- and stereoselectivity.

Table 2.

Scope of Hydrosilyl Acetals^a

^aConditions: (i) 10 (0.2 mmol), THF (1 M); (ii) THF (0.4 M).

^bIsolation yield.

^cDetermined by GC/MS analysis and ¹H NMR and NOE spectroscopy.

^dYield was determined by the reaction of 12e and 12f with MeLi.

Encouraged by the high regio- and stereoselectivity achieved with structurally diverse hydrosilyl acetals, we investigated the scope of traceless acetal-directed, α-selective hydrosilylation of propargyl alcohols to produce β-silyl allylic alcohols 13 via (E)-dioxasilinanes 12 (Table 3). Trimethylsilyl-substituted primary (10g), secondary (10h, 10i), and tertiary propargyl acetates (10a) provided 12g–12i and 12a in good to excellent yields with excellent α-regio- and stereoselectivity. Notably, the primary propargyl acetate 10g exhibited slower hydrosilylative cyclization kinetics and lower yield compared to secondary and tertiary propargyl acetates, presumably due to the absence of the Thorpe–Ingold effect. Cycloalkynol derivatives (10j–10m), ranging from three- to six-membered rings, tolerated the reaction conditions and generated 12j–12m. Furthermore, alkyl-, phenyl-, and naphthyl-substituted propargyl acetates (10n–10p) uneventfully underwent, producing the corresponding (E)-dioxasilinanes (12n–12p) in good yields. Notably, the enyne substrate 10q, containing a cyclohexenyl substituent, was compatible with the reaction conditions, leading to the formation of 12q (79% yield). Heterocycle-substituted propargyl acetates were then examined. The reaction of benzofuran-substituted propargyl acetate 10r afforded 12r (77% yield). However, the reaction of pyridine-substituted propargylic acetate 10s posed a significant challenge. Although a nearly quantitative conversion of 10s to hydrosilyl acetal 11s was observed under the conditions of Ir-catalyzed ester hydrosilylation, an additional extensive optimization was required for the Rh-catalyzed alkyne hydrosilylation step. Instability of both hydrosilyl acetal 11s and cyclic acetal 12s was observed under the standard Rh-catalyzed hydrosilylative cyclization conditions. However, we were pleased to discover that modifying the reaction conditions to a shorter reaction time (15 min) at higher temperature (100 °C) vis-à-vis 30 min at room temperature for all other substrates provided 12s. The yield of 12s was determined after the nucleophilic addition reaction with MeLi to produce β-silyl allylic alcohols 13s (70% yield, three steps from 10s). Finally, under identical conditions, secondary and tertiary bis-propargylic acetates 10t and 10u delivered 13t and 13u via (E)-dioxasilinanes 12t and 12u in good yields, respectively. The traceless nature of this reaction sequence was demonstrated by the addition of nucleophile (MeLi) to (E)-dioxasilinanes,^30,34 which removed the acetal-directing group, uncovered α-hydroxy group, and modified the silyl moiety in the same vessel to provide β-silyl allylic alcohols (13a, 13g–13u).

Table 3.

Scope of Propargyl Acetates^a

^aConditions: (i) 10 (0.2 mmol), THF (1 M), rt; (ii) THF (0.4 M), rt; (iii) MeLi (2.2 equiv), THF, −78 °C to rt.

^bIsolation yield.

^c5 mmol of 10a was used.

^dDetermined by GC/MS analysis and ¹H NMR spectroscopy.

^eDetermined by ¹H NMR spectroscopy utilizing an internal standard (mesitylene).

^fYields of 12s were not determined due to their partial instability. Conditions for 12s: (i) 10s (0.2 mmol), THF (1 M), rt; (ii) THF (0.4 M), 100 °C, 15 min nd: not determined.

Having demonstrated the scope of nucleophiles that add to cyclic silanes in our earlier work,^30,34 we next sought to examine the traceless acetal-directed, α-selective dual hydrosilylative cyclization of diyne 14 to generate bis-silyl diene (Scheme 1a). A three-step sequence involving consecutive ester and alkyne hydrosilylations of propargyl bisacetate 14, followed by MeLi addition, furnished functionalized 1,6-dihydroxy bis-silyl diene 15 (71% yield). Dioxasilinane 12n, prepared from the sequential Ir and Rh hydrosilylations of 10n, was capable of directly engaging in stereoretentive Pd-catalyzed cross-coupling to afford vinylarene 16 (Scheme 1b).^4,15

Scheme 1. Synthesis of Bis-Silyl Diene and Cross-Coupling of Vinylsilane^a.

^aConditions: (i) 14 (0.2 mmol), (ii) [Ir(coe)₂Cl]₂ (0.1 mol %), H₂SiEt₂ (3 equiv), THF (1 M), rt; (ii) [Rh(nbd)Cl]₂ (0.2 mol %), P(C₆F₅)₃ (1.2 mol %), THF (0.4 M); (iii) MeLi (4.4 equiv), THF, −78 °C to rt. ^bConditions: 10n (0.2 mmol), (ii) [Ir(coe)₂Cl]₂ (0.1 mol %), H₂SiEt₂ (3 equiv), THF (1 M), rt; (ii) [Rh(nbd)Cl]₂ (0.2 mol %), P(C₆F₅)₃ (1.2 mol %), THF (0.4 M); PhI (2 equiv), Pd₂(dba)₃ (2.5 mol %), TBAF (4 equiv), THF, rt to 40 °C.

To demonstrate the versatility of this strategy, we investigated the late-stage hydrosilylative modification of bioactive phenyl-substituted-17α-ethinylestradiol (EE) derivative 17 (Scheme 2). Under the reaction conditions developed, 17 underwent two consecutive Ir- and Rh-catalyzed hydrosilylation to give (E)-dioxasilinanes 20 via 19. Upon the addition of MeLi to 20, vinylsilane estradiol 18 was directly synthesized (overall 78% yield, 3 steps).

Scheme 2. Late-Stage Hydrosilylative Modification of the 17α-Ethinylestradiol (EE) Derivative^a.

^aConditions: (i) 17 (0.2 mmol), (ii) [Ir(coe)₂Cl]₂ (0.1 mol %), H₂SiEt₂ (3 equiv), THF (1 M), rt; (ii) [Rh(nbd)Cl]₂ (0.2 mol %), P(C₆F₅)₃ (1.2 mol %), THF (0.4 M); (iii) MeLi (2.2 equiv), THF, −78 °C to rt.

CONCLUSIONS

In summary, we have demonstrated the design and synthesis of a traceless two-atom tether acetal DG for the highly regio- and stereoselective catalytic α-selective syn-hydrosilylation of propargyl alcohols, leading to β-silyl allylic alcohols. The acetal DG was conveniently prepared via Ir-catalyzed ester hydrosilylation with excellent FG tolerance and a high yield. A π-acidic rhodium catalyst enabled the proximal, α-selective syn-hydrosilylation of a broad spectrum of primary to tertiary propargyl alcohols, resulting in (E)-dioxasilinanes with excellent regio- and stereoselectivity. The steric and electronic impacts of the hydrosilyl acetal structure on the α-selective, syn-hydrosilylative cyclization were investigated, where the π-acidic Rh catalyst facilitated fast cyclization of all types of hydrosilyl acetals with excellent regio- and stereoselectivity, presumably due to the establishment of a strong Rh-π interaction. This catalytic protocol demonstrated reasonably broad FG compatibility and scope, including the pyridine-substituted propargylic acetate. The scope of the resulting silyl moiety was expanded by postmodification of the resulting cyclic silyl acetal, introducing a new FG to the silicon moiety. This protocol was applied to a synthesis of conjugated bis-silyl 1,3-diene and stereoretentive Pd-catalyzed cross-coupling of dioxasilinanes to afford vinylarene. Finally, the versatility of this method was demonstrated in a late-stage hydrosilylative modification of a bioactive, complex molecular framework of a 17α-ethinylestradiol (EE) derivative.

METHODS

General Procedure for Traceless Acetal-Directed Catalytic Hydrosilylation of Propargyl Acetates.

[Ir-(coe)₂Cl]₂ (0.1 mol %) and propargyl acetates 10 (1 equiv) were dissolved with THF (1 M) in a flame-dried vial. Dihydrosilane (3 equiv) was added to the mixture. The septum on the vial was replaced by a screw cap with a Teflon liner. The reaction mixture was stirred for 12 h at rt. The volatiles were removed in vacuo to afford hydrosilyl acetals 11, which were directly used for a subsequent reaction without further purification. [Rh(nbd)Cl]₂ (0.2 mol %) and P(C₆F₅)₃ (1.2 mol %) were dissolved in THF (0.4 M). The crude hydrosilyl acetals 11 were added to the mixture in one portion. The septum on the vial was replaced by a screw cap with a Teflon liner, and the mixture was stirred at rt. The reaction progress was monitored via GC–MS spectrometry. The volatiles were removed in vacuo, and the resulting mixture was dissolved with pentane, filtered through a pad of Celite, and concentrated in vacuo to afford crude dioxasilinanes 12, which were purified by MPLC (hexanes/EtOAc). To a flame-dried vial was added a solution of dioxasilinanes 12 (1 equiv) in THF (0.2 M), and the mixture was cooled to −78 °C. MeLi (1.6 M in Et₂O, 2.2 equiv) was added slowly to the reaction mixture. After being stirred for 30 min at −78 °C, the reaction mixture was quenched with saturated aqueous NH₄Cl and extracted with Et₂O. The volatiles were removed to afford crude material, which was purified by MPLC (hexanes/EtOAc) to afford β-silyl allylic alcohols 13 as a colorless oil.