CHEMISTRY

A. I. TITOV, G. N. VEREMEEV, V. V. SMIRNOV, and O. D. SHAPILOV

A NEW REACTION FOR REPLACING AN ALCOHOLIC HYDROXYL BY FLUORINE AND ITS APPLICATION

(Presented by Academician I. N. Nazarov, 16 X 1956)

One of the problems in the study of organofluorine compounds is the search for methods for the direct replacement of alcoholic hydroxyls by fluorine. The generally known reactions for obtaining alkyl halides—in particular, the action of hydrogen fluoride and fluorine-containing phosphorus compounds on alcohols—proved to be of little use for this purpose (¹, ²). General indications of progress in this area (³, ⁴) require verification.

In 1942, one of us and A. N. Baryshnikova demonstrated the possibility of replacing an alcoholic hydroxyl by fluorine in one stage, using as an example the conversion of ethylene chlorohydrin into 1,2-fluorochloroethane by boiling it with a mixture of benzenesulfonyl fluoride and potassium fluoride. Our investigations established the generality of this reaction, which proceeds according to the equation:

[
\mathrm{ROH} + \mathrm{R'SO_2F} + 2\mathrm{KF} = \mathrm{RF} + \mathrm{R'SO_3K} + \mathrm{KHF_2}
\tag{1}
]

and proved its mechanism. It was found that the reaction proceeds through the following stages.

In the first stage of the interaction, formation of an alcoholate occurs according to the scheme:

[
\mathrm{ROH} + 2\mathrm{KF} \rightleftarrows \mathrm{ROK} + \mathrm{KHF_2},
\tag{2}
]

which is confirmed by the solubility of potassium fluoride in alcohols, the alkaline character of these solutions, the ease with which the more “acidic” halohydrins enter into reaction (1), the comparability of the dissociation constants of (\mathrm{H_2F_2}), water, alcohols, and other facts. In particular, we consider that the formation of ethylene oxide discovered by I. L. Knunyants and co-workers upon the action of potassium fluoride on ethylene chlorohydrin (⁵):

[
\mathrm{CH_2Cl{-}CH_2OH} + 2\mathrm{KF}
\ \xrightleftharpoons[-\mathrm{KHF_2}]{}\
\mathrm{CH_2Cl{-}CH_2OK}
\ \xrightarrow[-\mathrm{KCl}]{}\
\begin{matrix}
\mathrm{CH_2} & - & \mathrm{CH_2} \
& \backslash & / \
& \mathrm{O} &
\end{matrix}
\tag{2′}
]

is indisputable proof of the intermediate formation of an alcoholate in the reaction investigated by us. We also observed the formation of ethylene oxide in the preparation of ethylene fluorohydrin by the Hofmann method (⁶).

Next follows acylation of the alcoholate by sulfofluoride with formation of an alkyl sulfonate:

[
\mathrm{ROK} + \mathrm{R'SO_2F} \to \mathrm{R'SO_2OR} + \mathrm{KF},
\tag{3}
]

which was proved directly by isolation of these esters from the reaction mixture in the first phases of the synthesis. Quite recently, as became known to us after completion of the investigation, Pattison (⁷), in a very general form, pointed to the use of potassium fluoride for obtaining certain alkyl sulfonates in its presence by the action of sulfofluorides on alcohols. Partial formation of sulfonates is also possible upon the action of a sulfofluoride on alcohols without participation of potassium fluoride.

In the last stage, alkylation of potassium fluoride takes place:

[
\mathrm{R'SO_2OR + KF \to RF + R'SO_2OK,}
\tag{4}
]

as was previously known from the work of Helferich ((^8)), Razumovskii and Fridenberg ((^9)), and has again been confirmed by us in a number of examples.

Summation of the left- and right-hand sides of equations (2), (3), and (4) leads to the final scheme of synthesis (1).

In parallel with reaction (1), side processes may occur: formation of ethers, unsaturated compounds, their polymerization, etc.

The new method was successfully applied to the preparation of alkyl fluorides and their substituted derivatives.

Of particular interest are the experiments on preparing 1,2-difluoroethane by the new reaction. In recent works by Henne ((^{10},\,^{11})), it was described as an unstable compound with b.p. 10–11°, decomposing at 0° into butadiene and hydrogen fluoride and readily hydrolyzed by water to ethylene glycol. These properties seemed paradoxical even among organofluorine compounds. According to Henne’s report, it seemed that one could not expect successful synthesis of 1,2-difluoroethane by the new reaction. However, its preparation presented no particular difficulties.

1,2-Difluoroethane proved to be a quite stable substance with b.p. 26° and possessed the usual properties of fluoroparaffins, in particular resistance to hydrolysis. Its b.p. is regularly situated in the series:

[
\mathrm{CH_2Cl—CH_2Cl\ (84^\circ),\quad CH_2Cl—CH_2F\ (52^\circ),\quad CH_2F—CH_2F\ (26^\circ).}
]

The same difluoroethane was synthesized by alkylation of KF with β-fluoroethyl benzenesulfonate.

Reaction (1) in the case of ethylene chlorohydrin, along with the main product, 1,2-fluorochloroethane, led to appreciable formation of 1,2-dichloroethane and, apparently, 1,2-difluoroethane. The formation of dichloroethane is due to alkylation of potassium chloride by the β-chloroethyl benzenesulfonate formed as an intermediate, arising as a result of the reaction of the same ester and ethylene chlorohydrin with KF, for example:

[
\mathrm{CH_2Cl—CH_2OSO_2C_6H_5 + KF \to CH_2F—CH_2OSO_2C_6H_5 + KCl}
\tag{5}
]

[
\mathrm{CH_2Cl—CH_2OSO_2C_6H_5 + KCl \to CH_2Cl—CH_2Cl + C_6H_5SO_3K}
\tag{6}
]

[
\mathrm{CH_2F—CH_2OSO_2C_6H_5 + KF \to CH_2F—CH_2F + C_6H_5SO_3K}
\tag{7}
]

The ease of occurrence of reactions 6 and 7 has been proved directly. These conclusions are also applicable to explaining the formation of dichloroethane upon alkylation of potassium fluoride with β-chloroethyl benzenesulfonate in Razumovskii’s experiments.

We give a description of several experiments.

I. 20 g of methyl alcohol, 80 g of benzenesulfonyl fluoride, and 58 g of potassium fluoride were boiled for 7 h with a reflux condenser while stirring, and the methyl fluoride evolved was collected over 50% CaCl₂ (7.5 l—more than 60% of theory). The gas purified with H₂SO₄ was identified by molecular weight (34.0) and fluorine content (55.5%). The latter was determined by repeatedly bubbling the gas through an alcoholic solution of caustic potash, measuring the decrease in volume, and titrating (F^-).

The reaction with ethyl alcohol proceeded with greater difficulty; among the side products, ethyl ether and ethyl benzenesulfonate were isolated. Alkylation of KF with ethyl sulfonate also gave identical (C_2H_5F) with an admixture of olefins.

II. 32 g of ethylene chlorohydrin, 80 g of benzenesulfonyl fluoride, and 64 g of KF were heated in a bath (180–190°) with a dephlegmator ensuring distillation of 1,2-difluoroethane (26–28°). The yield of product was about 50%. After drying and redistillation, 1,2-difluoroethane had the following properties: b.p. 26–26.2°; (d_4^{10} = 1.024); (n_D^{12} = 1.3014); (M) 65.82 (cryoscopic); 66.10 (by Meyer) (theor. 66.05); F 56.6% (theor. 57.5).

Elimination of the fluoride ion in aqueous and alcoholic NaOH solution proceeded slowly (at 70° over one hour, 1–2%) and somewhat faster in the presence of acids (up to 6%). Like dichloro- and fluorochloroethane, it was mixed with fuming HNO₃ and separated upon dilution with water.

β-Fluoroethyl benzenesulfonate—b.p. 161–162° (8 mm), $d_4^{13}=1.3497$, $n_D^{15}=1.5104$. F was determined by titration after boiling 1 g of the ester for half an hour with 150 ml of 2% aqueous-alcoholic NaOH solution—7.9% (theor. 9.3). It is interesting that the ester mixed with 1–10 parts of CCl₄, but the solution separated upon further dilution. Alkylation of KF with the ester gave the above-described 1,2-difluoroethane in a yield close to the theoretical. Recently Edgell and Parts (¹²), without isolating 1,2-difluoroethane in substantia, showed the probability of its formation from the infrared spectrum of the vapors in an analogous reaction of fluoroethyl toluenesulfonate with potassium fluoride.

III. 40 g of ethylene chlorohydrin, 60 g of C₆H₅SO₂F, and 50 g of KF were heated for 6 h in a bath (180–200°) with removal of the reaction products (25 g); the following fractions were isolated from them: 1st, 25–40°, 1.3; 2nd, 45–60°, 12.9 g; 3rd, 65–75°, 2.5; 4th, above 75°, 8 g. From the 2nd fraction, 1,2-fluorochloroethane (I) was isolated: b.p. 51–52°, $d_4^{16}=1.184$, $n_D^{16}=1.3955$, % Cl 43.0; % F 24.9. N-fluoroethyl-3-nitrophthalimide (II), m.p. 105°, was obtained from I by heating with potassium nitrophthalimide to 180–200° (10 h); no fluorine was eliminated in this reaction. II obtained from I synthesized by Razumovskii’s method was identical. From the higher fractions, 1,2-dichloroethane was isolated; it was also formed upon heating β-chloroethyl benzenesulfonate with potassium chloride (b.p. 82°, $d_4^{20}=1.253$, $n_D^{20}=1.4465$). From the residue after the main reaction had been carried out, β-chloroethyl benzenesulfonate was isolated (b.p. 174° at 6 mm, $d_4^{17}=1.361$, $n_D^{17}=1.531$).

Received
31 VIII 1956

CITED LITERATURE

¹ Gilman’s Organic Chemistry, N. Y., 1943.
² I. L. Knunyants, O. V. Kil’disheva, Usp. Khim., 15, 868 (1946).
³ W. Klatt, Zs. anorg. Chem., 222, 289 (1935).
⁴ O. Sherer, Angew. Chem., 52, 457 (1939).
⁵ I. L. Knunyants, O. V. Kil’disheva, ZhOKh, 19, 101 (1949).
⁶ F. W. Hoffman, J. Am. Chem. Soc., 70, 2596 (1948).
⁷ F. L. M. Pattison, RZhKhim., No. 48976 (1955); Nature, 4433, 737 (1954).
⁸ B. Helferich, A. Gnüchtel, Ber., 74 B, 1035, 1807 (1941).
⁹ V. V. Razumovskii, A. E. Fridenberg, ZhOKh, 19, 93 (1949).
¹⁰ A. Henne, M. Renoll, J. Am. Chem. Soc., 58, 882 (1936).
¹¹ A. Henne, T. Midgley, J. Am. Chem. Soc., 58, 889 (1936).
¹² W. F. Edgell, L. Parts, J. Am. Chem. Soc., 77, 4899 (1955).