High-yielding method for preparation of carbocyclic or N-containing heterocyclic β-keto esters using in situ activated sodium hydride in dimethyl sulphoxide

ABSTRACT It was found that NaH suspension in DMSO was highly activated when reacted with an alcohol. The in situ generated NaH/alkoxide mixture permitted very rapid and complete deprotonation and acylation of various cyclic ketones with alkyl carbonates at ambient temperature. Activated NaH/alkoxide in DMSO is particularly effective in Dieckmann condensations, where it affords 5- and 6-membered carbocyclic or N-containing heterocyclic β-keto esters in high yields. A heterocyclic Dieckmann condensation was performed on a molar scale, demonstrating the scalability of the procedure. Besides, DMSO is non-toxic, relatively inexpensive and environmentally benign solvent. GRAPHICAL ABSTRACT

Therefore, this work was aimed to develop an efficient and general preparative method for both the ketone acylation and Dieckmann condensation, avoiding hazardous solvents and elevated temperatures.

Results and discussion
In connection with our research projects, we needed multigram quantities of various carbocyclic and heterocyclic β-keto esters (Tables 1 and 2). However, available literature procedures for ketone acylation with alkyl carbonates using NaH in benzene (8) or NaH/dry KH in THF (9) resulted, in our hands, in slow conversions, difficult to separate mixtures and low to moderate yields. In addition, benzene and dry KH are particularly hazardous chemicals. While Dieckmann condensation of various amino-diesters (Table 2) using alkoxide bases often afforded heterocyclic β-keto esters in 70-90% yields, it required 3-5 h of reflux in toluene or xylenes. More significantly, under the reaction conditions, a number of the starting amino-diesters (e.g. 2a, 2d, 2e, Table 2) formed thick gels, greatly reducing the isolated yields and purity of the products. Therefore, we sought to develop a more efficient and general preparative method for both the ketone acylation and Dieckmann condensation.
To that end, the acylation of carbocyclic ketones with alkyl carbonates in DMSO was reexamined, using NaH as a base. Initial attempts resulted in exceedingly slow deprotonation/acylation rate at ∼20°C, while at ∼40-50°C, the reaction proceed erratically and with delayed exotherm, often yielding numerous side products. Further experiments involved addition of various co-solvents, solid alkaline alkoxides (MeONa, EtONa, t-BuOK) and/or free alcohols. The activity of NaH suspension in DMSO was highly enhanced after treatment with certain amounts of the alcohol (MeOH, EtOH, allyl, t-BuOH). Except for t-BuOH, the alcohols were chosen according to the alkyl carbonate used, to prevent transesterification. Subsequent addition of ketone/alkyl carbonate mixture resulted in immediate deprotonation/ acylation at ∼20°C, as evidenced by vigorous H 2 evolution and aliquot analysis. Variations in molar amounts of NaH and alcohol relative to the ketone gave the optimal NaH/alcohol/ketone molar ratio of ∼2.5:1.0:1.0, respectively. The relative amount of NaH can be reduced to ca. 2.0-2.1 equiv, without affecting the yields; however the overall reaction rate was decreased (by 20-50%). It appears that the alcohol has a dual role, providing alkoxide as the kinetically more active base as well as the activation of the NaH surface. However, the complete conversion of NaH to the alkoxide with stoichiometric quantity of alcohol greatly reduced the isolated yields and purity of β-keto esters. Hence, the mixture of NaH and in situ formed alkoxide in DMSO was found to be a particularly effective base system, providing both rapid and irreversible deprotonation of ketones, Scheme 1. Attempted use of LiH failed to effect any reaction.
In experiments where a mixture of alcohol, ketone and alkyl carbonate was added to a stirred NaH/ DMSO suspension, a significant reaction delay was observed, followed by the vigorous exotherm. The results confirm that the NaH suspension must be activated with alcohol, before the addition of reactants. Under these conditions, complete conversion of ketones was achieved and the corresponding β-keto esters were isolated in excellent yields, on a 20-160 mmol scale (Table 1). Side products of undetermined structures were detected only occasionally, when the internal temperature exceeded ∼35°C. Use of ethyl formate instead of alkyl carbonate provided β-formyl derivative 3i.

Conclusion
In conclusion, the use of in situ generated NaH/alkoxide mixture in DMSO allows for a high-yielding and efficient procedure for the Dieckmann condensation and alkoxycarbonylation of ketones. It is inexpensive and suitable method for large scale preparations of various β-keto esters. In comparison with previously reported procedures, advantages of the method described are numerous; expediency, high yields, the operational simplicity, and in addition, use of DMSO as non-toxic, nonflammable and environmentally benign solvent which could be quantitatively removed by simple aqueous extraction. Other solvents used (petroleum ether, PhMe, MeOAc or EtOAc) are also of low toxicity. Scheme 1. Acylation of cyclic ketones with alkyl carbonates using activated NaH. Scheme 2. Dieckmann condensation of diesters using activated NaH/DMSO.

General information
1 H NMR spectra were recorded at 200 MHz or 500 MHz and 13 C NMR/APT at 50 or 126 MHz as indicated in each case. Data are reported as chemical shifts in δ (ppm). The chemical shift references were as follows: 1 H NMR (7.26 δ, CHCl 3 or 0.00 δ, TMS) and 13 C NMR (77.00 δ, CDCl 3 ). 1 H and 13 C NMR spectra were Bruker Avance 500, or Varian Gemini 2000 instruments. All high-resolution mass spectra were obtained by ESI-ToF and recorded on Agilent Technologies 6210-1210 TOF-LC-ESI-HR/MS instrument in positive mode. All experiments were monitored by thin-layer chromatography on silica and/or by 1 H NMR. Reagent-grade solvents and commercial precursors (declared H 2 O content <0.1%) were used without further purification. BHT refers to 3,5-di-tert-butyl-4-hydroxytoluene. NaH was used exclusively as a 60% mineral oil suspension. Typically, the mineral oil was washed with petroleum ether prior to the reaction, using the reverse filtration under positive pressure of Ar. In molar-scale experiments, the washing was omitted and the mineral oil was removed during the work-up. The reaction mixture was workedup with an aqueous solution of tartaric acid, followed by pH adjustment to pH 9-10 with 20% K 2 CO 3 solution. Low volatile compounds (amino-diesters, most β-keto esters and heterocyclic ketones were vacuum dried to remove H 2 O and solvents (20-50°C, 0.1-0.5 mm Hg, 30-60 min) prior to the spectral analysis and determination of the yields.

Typical procedure for the ketone acylation
The compounds 3a-3i were prepared according to the typical procedure for 3a.