Key words
abnormal Peterson olefination - allylsilane - one-pot synthesis
Allylsilanes are versatile allylic anion equivalents.[1]
[2] Unlike most allyl organometallic species such as allylic Grignard reagents, allylic rearrangement is not a concern for allylsilanes. Formation of the new bond via electrophilic reaction occurs regionspecifically at the γ-position of the allylsilane (Scheme [1]).[1e]
Scheme 1
Many allylsilane syntheses were reported during the 1990s.[2] Most popular are transformations of alkenes; whereas transformations of carbonyl compounds into allylsilanes are scarce, although one obvious example utilizes the Wittig protocol to convert ketones or aldehydes.[2e] However, this method requires multiple synthetic steps starting from triphenylphosphine to furnish the corresponding allylsilane, and removing the triphenylphosphine oxide by-product is challenging (Scheme [2]).[3]
Scheme 2
Another example uses the Peterson olefination[4] protocol to convert an ester into the desired allylsilane via loss of trimethylsilanol from the initial bis-adduct (Scheme [3]).[2d]
Scheme 3
The standard Peterson olefination converts β-hydroxysilanes into desilylated olefins under either acidic or basic conditions. Battiste and Kwan reported the use of alumoxanes in vinylsilane synthesis.[5] The methodology employs a Peterson protocol to convert non-enolizable ketones into the corresponding vinylsilanes. Since the silyl group is intact during the elimination, these are referred to as abnormal Peterson olefinations. Doxsee and Kwan extended the vinylsilane synthesis, starting from non-enolizable aryl aldehydes via the abnormal Peterson olefination.[6] Instead of using alumoxanes as amphoteric reagents, Doxsee and Kwan used the Tebbe reagent for the elimination step. Kwan’s subsequent work revealed that an additional Lewis acid such as the titanium in Tebbe’s reagent assists the deoxygenation.[7] We report herein the use of the abnormal Peterson protocol to convert enolizable ketones into allylsilanes in a convenient two-step, one-pot reaction. Alkylating enolizable ketones and aldehydes is often challenging because it can give undesired products such as self-aldol condensation products. Several synthetic methods have been developed to alkylate enolizable ketones and aldehydes effectively.[8] Recognizing the ability of Zn2+ to cleave carbon–oxygen bonds,[9] we have employed the alkylation method, using ZnCl2, developed by Ishihara et al.[8d] for our first reaction step with slight modification.
In Ishihara’s method, the (CH3)3SiCH2 group (20 mol%) was used as a non-transferable ligand in the catalytic alkylzinc(II) ate complex and to enhance its catalytic efficacy in Grignard reactions (nucleophilic additions to carbonyl group). Reaction of excess TMSCH2MgCl (≥ 110 mol%) with acetophenone led to the formation of the corresponding β-hydroxysilane. We decided to examine the scope of the reaction of excess TMSCH2MgCl further with other aryl enolizable ketones to yield the corresponding β-silylalkoxides.
Enolizable aryl ketones were added to a heterogeneous mixture of 130 mol% trimethylsilylmethyl magnesium chloride (1 M in diethyl ether), 10 mol% ZnCl2 and 170 mol% LiCl (Scheme [4]). The original protocol called for the use of THF as solvent. Our preliminary study focused on the reaction with acetophenone by subjecting acetophenone to the reaction protocol with and without the use of THF as solvent. The reported protocol called for 2 hours stirring at 0 °C after dropwise addition of ketone also at 0 °C. Our results suggested that the reaction can be allowed to stir overnight (ca. 12 h) without significant formation of side-products, even in the absence of THF solvent. Diethyl ether from the (trimethylsilyl)methyl magnesium chloride solution (1 M) appears to suffice as solvent to facilitate the formation of β-silylalkoxides.
Scheme 4
A mixture of hexanes was also examined as a possible solvent. While no significant difference was observed in the first reaction step (converting the enolizable ketones into their corresponding β-silylalkoxides), the subsequent elimination reaction generally required longer reaction times and higher temperatures.
To carry out the abnormal Peterson reaction of the β-silylalkoxides, diisobutylaluminum chloride was added followed by THF. The resulting mixture was transferred by using a cannula into a sealed tube under argon and the mixture was heated at 130 °C for 15 hours (Scheme [5]), followed by washing with saturated aqueous sodium tartrate and work-up. The results are summarized in Table [1].
Scheme 5
Table 1 Two Step, One-Pot Synthesis of Allysilane from Aryl Methyl Ketones
|
Entry
|
Aryl methyl ketones
|
Yield (%)a
|
|
|
Allysilane
|
Olefin
|
|
|
|
|
|
1
|
X = 4-H
|
58[10]
|
8[11]
|
|
2
|
X = 4-F
|
71[10]
|
0[12]
|
|
3
|
X = 4-Cl
|
63[13]
|
0[14]
|
|
4
|
X = 4-Br
|
68[15]
|
0[12]
|
|
5
|
X = 4-OCH3
|
43[16]
|
32[12]
|
|
6
|
X = 4-CH3
|
43[16]
|
7[14]
|
|
7
|
X = 4-CH2CH3
|
64[16]
|
11[17]
|
|
8
|
X = 2-CH3
|
multiple products: aldol products + allylsilane + olefin
|
a References for 1H NMR spectroscopic comparison.
The Ishihara study[8d] briefly examined the use of the trimethylsilylmethyl group ((CH3)3SiCH2–) as the alkylating agent. We modified the reported protocol slightly and used it to alkylate substituted aryl methyl ketones in the first step, yielding the corresponding β-silylalkoxides. Heating the reaction mixture containing the silylalkoxide derived from acetophenone (Table [1], entry 1) reaction in THF (without adding diisobutylaluminum chloride) simply produced the corresponding β-hydroxysilane. This indicates that additional amphoteric reagent, diisobutylaluminum chloride, is required for the abnormal Peterson olefination to yield the corresponding allylsilanes. Diisobutylaluminum chloride was used rather than less bulky diethylaluminum chloride since, in the acetophenone reaction (entry 1), the latter yielded a greater amount of olefin product.
p-Haloaryl methyl ketones afforded the corresponding allylsilanes in acceptable yields with high chemoselectivity. Reactions of electron-rich ketones (Table [1], entry 5–7) afforded a mixture of the corresponding allylsilanes and olefins. It is worth noting that the most electron-rich ketone (entry 5) produced the highest amount of the corresponding olefin. To determine in which step the corresponding olefin is produced in the two-step, one-pot synthesis, we ran a series of alkylation reactions to produce β-hydroxysilanes for each ketone. Reactions of electron-rich ketones yielded a mixture of β-silylalcohols and olefins while p-haloaryl methyl ketones yielded primarily β-silylalcohols. Alkylation of o-methylaryl methyl ketone gave multiple products such as aldol addition products, olefins, and β-silylalcohols.
Applying the developed protocol with 2-methylacetophenone afforded multiple products (Table [1], entry 8). To understand the system better, we applied Ishihara’s protocol in attempts to convert 2-methyl-, 2-fluoro-, and 3-methylacetophenone into their corresponding β-hydroxysilanes (Scheme [6]). Repeated reactions yielded multiple products from these three ketones. 1H NMR spectroscopic and TLC analyses indicated a minimum of three components in the mixture: recovered ketone, self-aldol condensation product, and β-hydroxysilane. Thus, our attempts to prepare allylsilanes were ineffective with these three ketones. We also examined the developed protocol with 4-nitroacetophenone but, given the vulnerability of the nitro group to nucleophilic attack from the Grignard reagent, it is not surprising that the reaction failed to produce the corresponding allylsilane product.[18]
Scheme 6
In conclusion, we have employed and slightly modified the method for alkylation of enolizable ketones developed by Ishihara to alkylate a series of methyl aryl ketones with a trimethylsilylmethyl ((CH3)3SiCH2–) group to generate β-silylalkoxides as the first step of our one-pot allylsilane synthesis protocol. The second step entails addition of diisobutylaluminum chloride with heating, leading to the abnormal Peterson olefination reaction, affording allylsilanes in fair yield.[19] This one-pot synthesis works particularly well with p-haloaryl methyl ketones.
