From Symmetric Glycerol Derivatives to Dissymmetric Chlorohydrins

The anticipated worldwide increase in biodiesel production will result in an accumulation of glycerol for which there are insufficient conventional uses. The surplus of this by-product has increased rapidly during the last decade, prompting a search for new glycerol applications. We describe here the synthesis of dissymmetric chlorohydrin esters from symmetric 1,3-dichloro-2-propyl esters obtained from glycerol. We studied the influence of two solvents: 1,4-dioxane and 1-butanol and two bases: sodium carbonate and 1-butylimidazole, on the synthesis of dissymmetric chlorohydrin esters. In addition, we studied the influence of other bases (potassium and lithium carbonates) in the reaction using 1,4-dioxane as the solvent. The highest yield was obtained using 1,4-dioxane and sodium carbonate.


Introduction
Glycerol (1,2,3-propanetriol), is one of the most versatile and valuable chemicals known to man. Glycerol has over 1,500 known end uses, including applications as an ingredient or processing aid in cosmetics, toiletries, personal care products, pharmaceutical formulations and food [1]. Traditionally, it is obtained as a by-product in four different processes: soap manufacture (saponification), fatty acid OPEN ACCESS production (hydrolysis), fatty ester production (transesterification) and microbial fermentation. It can also be synthesized from propylene oxide [2].
Production of glycerol as a by-product has increased rapidly during the last decade as a consequence of biodiesel production [3,4]. Biodiesel is a common term for the different alkyl fatty esters formed by transesterification of vegetable oils or fats with an alcohol, usually methanol, with or without a catalyst. It has been implemented as an alternative to transportation fuels in Europe, the United States, Brazil and other countries as a result of increased interest in using renewable energy sources to reduce greenhouse gas emissions and to alleviate the depletion of fossil fuel reserves [5,6].
In the biodiesel process, glycerol is usually generated at the rate of one mole of glycerol for every three moles of synthesized methyl esters; approximately 10% of the total product by mass. Thus, the increase in biodiesel production will result in an accumulation of glycerol for which there is currently insufficient conventional uses. This situation has created a glut in the glycerol market and a drop in its price.
Crude glycerol generated during the transesterification process leading to biodiesel contains impurities such as methanol, water, inorganic salts (catalyst residue), free fatty acids, unreacted mono-, di-and triglycerides, methyl esters and a variety of other organic materials, depending on the biodiesel process [7,8], hence it must be purified before subsequent use in conventional applications.
In this paper, we report the results obtained when two solvents (1,4-dioxane and 1-butanol) and different bases are used to prepare dissymmetric chlorohydrin esters from symmetric 1,3-dichloro-2propyl esters obtained from glycerol.

Results and Discussion
Initially, we studied the influence of combining two solvents and two bases in the synthesis of dissymmetric chlorohydrin esters. Thus, several 1,3-dichloro-2-propyl esters 1a-d were treated with water and an organic (1-butylimidazole) or inorganic (sodium carbonate) base in two different solvents: 1,4-dioxane or 1-butanol ( Table 1). Most of the esters used were α,α-dimethyl-substituted (versatic) esters. Yields of the reactions carried out with sodium carbonate and 1,4-dioxane were usually higher than the corresponding reactions with sodium carbonate and 1-butanol, except in the reaction with 2-chloro-1-(chloromethyl)ethyl 2,2-dimethylpropionate (1b), where the yield was the same. A similar tendency was observed using 1-butylimidazole as the base, but this time the exception was in the reaction with 2-chloro-1-(chloromethyl)ethyl 2-methyl-2-phenylpropionate (1d).
In the reactions with 2-chloro-1-(chloromethyl)ethyl hexadecanoate (1a), the best solvent was 1,4dioxane. There was no difference between the use of either sodium carbonate or 1-butylimidazole. The product, 2a, was not observed when 1-butanol was used. Probably, because this not an α,α-dimethylsubstituted ester, and it is not stable in such a solvent.
We propose the mechanism shown in Scheme 1 for the formation of 3-chloro-2-hydroxy-1-propyl esters. The transposition process observed here has already been described by our group in the preparation of glycidyl palmitate [13]. This mechanism is compatible with the experimental results. Similar behaviour was described by Leggetter and Brown studying the influence of substituents on the opening of substituted 1,3-dioxolanes [28,29]. Therefore, the preferred cleavage of C2-O3 bond might be explained by electron-withdrawing substituents, which should produce on C4 a higher positive charge density than on C5. This will provoke the formation of the indicated chlorohydrin ester instead of the other putative regioisomer.
In addition, we studied the influence of other bases using 1,4-dioxane as the solvent because it had been the best solvent considering all the tested compounds. Our aim was to improve the obtained yields. Two esters were chosen as model compounds. One was the palmitic ester because it does not possess α,α-dimethyl-substitution. The other was α,α-dimethylphenyl acetate. Considering that there was a small difference in the yield of compounds 2a and 2d when organic or inorganic bases were used in 1,4-dioxane, the effect of sodium, potassium and lithium carbonates on 3-chloro-2-hydroxy-1propyl ester formation was checked. Scheme 1. Mechanism proposed to explain the formation of 3-chloro-2-hydroxy-1-propyl esters using water and base as reagent.
Satisfactory yields were obtained using a cheap base such as sodium carbonate ( Table 2), independent of the acyl compound used. When the effect of each base on the transformation of dichloro esters 1a and 1d was studied, the percentages of 3-chloro-2-hydroxy-1-propyl esters 2a and 2d were always higher in the transformation of 2-chloro-1-(chloromethyl)ethyl hexadecanoate (entry a) than in the transformation of phenyl acetate derivatives, irrespective of the base used. The yield differences with the inorganic bases may be explained by differences in water solubility of these salts. Sodium and potassium carbonate solubility increases with increasing temperature and therefore the availability of hydroxyl groups which act as nucleophile. On the other hand, potassium cations are more basic than sodium cations. Yields in the reactions with potassium carbonate are lower because 3-chloro-2-hydroxy-1-propyl ester was transformed into an epoxide (result not shown). These compounds were not observed when sodium carbonate was used. Lithium carbonate has behaves differently because its solubility decreases with increasing temperature.
Finally, a multigram-scale synthesis of 3-chloro-2-hydroxy-1-propyl 2,2-dimethylpropanoate 2b was carried out, considering that the best yields were always obtained using the 2,2dimethylpropanoate ester 1b. 1,4-Dioxane and sodium carbonate were used to transform 5.05 g of 1b into compound 2b. A yield of 2.9 g of 2b (62% yield) was obtained after purification by distillation, showing that the method is fully applicable to obtain gram quantities of dissymmetric chlorohydrins from the corresponding symmetric esters.

Procedure for the syntheses of 2-chloro-1-(chloromethyl)ethyl esters 1
Carboxylic acid (1 mmol), glycerol (184 mg, 2 mmol) and chlorotrimethylsilane (540 mg, 5 mmol) were added to a reaction vial fitted with a PTFE-lined cap. The mixture was heated at 80 °C for 48 h. After cooling, an organic solvent was added, and the mixture was washed three times with water. The organic layer was dried over anhydrous MgSO 4 , and the solvent was evaporated under vacuum. The residue was purified by dry flash column chromatography on silica gel.

Procedure for the syntheses of 3-chloro-2-hydroxy-1-propyl esters 2
A solution of 2-chloro-1-(chloromethyl)ethyl ester (1 mmol), water (29 μL, 1.6 mmol) and either sodium carbonate (106 mg, 1 mmol) or 1-butylimidazole (124 mg, 1 mmol) in dried solvent (1,4dioxane or 1-butanol (3 mL)) was heated at 115 °C for 48 h in a capped reaction vial. After cooling, dichloromethane was added, and the mixture was washed three times with water. The organic layer was dried over anhydrous MgSO 4 , and the solvent was evaporated. The crude compound was analyzed by 1 H-NMR to determine the yield of the corresponding product using N,N-dimethylformamide (DMF) as an internal standard. Further purification by dry flash column chromatography on silica gel (hexane/ethyl acetate) gave the desired product.

Procedure for the syntheses of 3-chloro-2-hydroxy-1-propyl esters 2a and 2d using different inorganic bases
A solution of 2-chloro-1-(chloromethyl)ethyl ester (1 mmol), water (29 μL, 1.6 mmol) and a carbonate (1 mmol) in dried 1,4-dioxane (3 mL) was heated at 115 °C for 48 h in a capped reaction vial. After cooling, dichloromethane was added and the mixture was washed three times with water. The organic layer was dried over anhydrous MgSO 4 , and the solvent was distilled. The crude compound was analyzed by 1 H-NMR to determine the yield of the corresponding product using DMF as an internal standard. Further purification by dry flash column chromatography on silica gel (hexane/ethyl acetate) gave the desired product.

Conclusions
1,3-Dichloro-2-propyl esters prepared from glycerol can be transformed into the corresponding 3chloro-2-hydroxy-1-propyl esters. These dissymmetric chlorohydrin esters are potentially valuable compounds and can be used as building blocks. With the methodology described above, different 3chloro-2-hydroxy-1-propyl esters were synthesized with all studied compounds; better reaction conditions were obtained when using sodium carbonate as base and 1,4-dioxane as solvent. In addition, when different inorganic carbonates were used, sodium carbonate gave the best results. The best procedure developed allows to prepare 3-chloro-2-hydroxy-1-propyl 2,2-dimethylpropanoate on a multigram-scale.