Dichlorotrifluoromethoxyacetic Acid: Preparation and Reactivity

We describe the first gram scale preparation of the reagent dichlorotrifluoromethoxyacetic acid. This stable compound is obtained in five steps starting from the cheap diethylene glycol. The reactivity of the sodium salt of this fluorinated acid was also tested and allowed the preparation of new amides.


Introduction
Since the seminal preparation of the trifluoromethoxy group by Yagupolskii in 1955 [1], the interest in this very specific organic moiety has grown continuously, in particular for life sciences purposes [2][3][4][5][6][7][8][9]. Such interest can be explained by the conjunction of its multiple advantages: the "pseudohalogen" character of this entity which makes it comparable to a fluorine atom in terms of electronic properties, and the deep modifications of the conformation as well as the physico-chemical behavior induced in molecules linked to this group [10][11][12][13][14]. The unrivaled and promising properties brought by this ether function are in deep contrast with the synthetic difficulties to prepare it [15][16][17]. Major and recent progresses have been made in either the direct introduction of the trifluoromethoxy moiety [18][19][20][21][22][23] (often through a nucleophilic pathway) or in its preparation from alcohols or phenols [24][25][26][27][28]. There is nevertheless still an urgent need for new methods able to selectively introduce this moiety at a late stage of a synthetic procedure. The design of new reagents enabling the grafting of this substituent should be a highly valuable addition to the presently existing methods. Based on our ongoing research project in this field, we thought that related trifluoromethoxy group-bearing molecules should be easily accessible based on previous work of our laboratory [29][30][31][32].
In this communication we describe the preparation of the sodium salt of dichlorotrifluoromethoxyacetic acid, its attempted use in chlorotrifluoromethoxycarbene generation and trapping thereof, as well as the concomitant preparation of some interesting new nitrogen-based trifluoromethoxy-bearing building blocks.

Results and Discussion
The planned synthesis of the target dichlorotrifluoromethoxyacetic acid, 3 (Scheme 1) was based on initial chlorination of the trifluoromethoxy ester 1 previously described by our group [24]. Thus, exhaustive chlorination of ester 1, dissolved in CCl 4 , under UV irradiation in an intermittent stream of dichlorine easily afforded the perchlorinated ester 2. 19 F NMR spectra of this ester exhibited clear evidence for partial restricted rotation as shown by the presence of one sharp (δ = −54.4 ppm) and one very broad peak (δ = −54.5 ppm) instead of the expected two sharp signals. Further saponification of one equivalent of ester 2 with sodium hydroxide gave two equivalents of the unknown targeted acid 3 after acidification of the reaction medium, gratifyingly making use of both trifluoromethoxy groups present in the starting molecule. Free acid 3, which tenaciously retained diethyl ether solvent, could only be partially characterized and was then transformed in a final step. The derived sodium salt 4 of acid 3 could however be readily isolated in acceptable overall yield from ester 1 after simple treatment of a diethyl ether solution of acid 3 with a stoichiometric amount of sodium bicarbonate followed by thorough drying under high vacuum.

Results and Discussion
The planned synthesis of the target dichlorotrifluoromethoxyacetic acid, 3 (Scheme 1) was based on initial chlorination of the trifluoromethoxy ester 1 previously described by our group [24]. Thus, exhaustive chlorination of ester 1, dissolved in CCl4, under UV irradiation in an intermittent stream of dichlorine easily afforded the perchlorinated ester 2. 19 F NMR spectra of this ester exhibited clear evidence for partial restricted rotation as shown by the presence of one sharp (δ = −54.4 ppm) and one very broad peak (δ = −54.5 ppm) instead of the expected two sharp signals. Further saponification of one equivalent of ester 2 with sodium hydroxide gave two equivalents of the unknown targeted acid 3 after acidification of the reaction medium, gratifyingly making use of both trifluoromethoxy groups present in the starting molecule. Free acid 3, which tenaciously retained diethyl ether solvent, could only be partially characterized and was then transformed in a final step. The derived sodium salt 4 of acid 3 could however be readily isolated in acceptable overall yield from ester 1 after simple treatment of a diethyl ether solution of acid 3 with a stoichiometric amount of sodium bicarbonate followed by thorough drying under high vacuum. With compound 4 in hand, we then studied the opportunity to generate the chlorotrifluoromethoxycarbene 6 (Scheme 2). Most fluorinated carbenes are known for their electrophilic character. They usually react with electron rich functionalities [33][34][35]. The presence of an oxygen atom adjacent to the carbenic center in the carbenic species 6 (Scheme 2) we planned to generate, was however susceptible to alter this normal behavior [36][37][38]. We thus attempted the trapping of the derived carbene 6 with a wide panel of variously substituted olefins with either electrophilic (trichlorofluoroethene, dimethylbutadiene, etc.) or nucleophilic (enol ethers, etc.) character. Whatever the conditions employed (solvents, temperature, aromatic, double or triple bonds as a trap), no trace of the desired chlorotrifluoromethoxymethylated compounds was obtained. In some cases, the only perfluorinated With compound 4 in hand, we then studied the opportunity to generate the chlorotrifluoromethoxycarbene 6 (Scheme 2).

Results and Discussion
The planned synthesis of the target dichlorotrifluoromethoxyacetic acid, 3 (Scheme 1) was based on initial chlorination of the trifluoromethoxy ester 1 previously described by our group [24]. Thus, exhaustive chlorination of ester 1, dissolved in CCl4, under UV irradiation in an intermittent stream of dichlorine easily afforded the perchlorinated ester 2. 19 F NMR spectra of this ester exhibited clear evidence for partial restricted rotation as shown by the presence of one sharp (δ = −54.4 ppm) and one very broad peak (δ = −54.5 ppm) instead of the expected two sharp signals. Further saponification of one equivalent of ester 2 with sodium hydroxide gave two equivalents of the unknown targeted acid 3 after acidification of the reaction medium, gratifyingly making use of both trifluoromethoxy groups present in the starting molecule. Free acid 3, which tenaciously retained diethyl ether solvent, could only be partially characterized and was then transformed in a final step. The derived sodium salt 4 of acid 3 could however be readily isolated in acceptable overall yield from ester 1 after simple treatment of a diethyl ether solution of acid 3 with a stoichiometric amount of sodium bicarbonate followed by thorough drying under high vacuum. With compound 4 in hand, we then studied the opportunity to generate the chlorotrifluoromethoxycarbene 6 (Scheme 2). Most fluorinated carbenes are known for their electrophilic character. They usually react with electron rich functionalities [33][34][35]. The presence of an oxygen atom adjacent to the carbenic center in the carbenic species 6 (Scheme 2) we planned to generate, was however susceptible to alter this normal behavior [36][37][38]. We thus attempted the trapping of the derived carbene 6 with a wide panel of variously substituted olefins with either electrophilic (trichlorofluoroethene, dimethylbutadiene, etc.) or nucleophilic (enol ethers, etc.) character. Whatever the conditions employed (solvents, temperature, aromatic, double or triple bonds as a trap), no trace of the desired chlorotrifluoromethoxymethylated compounds was obtained. In some cases, the only perfluorinated Most fluorinated carbenes are known for their electrophilic character. They usually react with electron rich functionalities [33][34][35]. The presence of an oxygen atom adjacent to the carbenic center in the carbenic species 6 (Scheme 2) we planned to generate, was however susceptible to alter this normal behavior [36][37][38]. We thus attempted the trapping of the derived carbene 6 with a wide panel of variously substituted olefins with either electrophilic (trichlorofluoroethene, dimethylbutadiene, etc.) or nucleophilic (enol ethers, etc.) character. Whatever the conditions employed (solvents, temperature, aromatic, double or triple bonds as a trap), no trace of the desired chlorotrifluoromethoxymethylated compounds was obtained. In some cases, the only perfluorinated molecule detected was the dichlorotrifluoromethoxymethane 5 [39]. Even if very volatile, we were able to isolate it by careful distillation (from a crude mixture with diglyme as solvent), yielding small amounts of pure product 5. The latter was characterized by NMR. The generation of the expected carbene 6 was assumed to proceed in a two-step pathway: first a decarboxylation followed by the elimination of a chlorine atom. The presence of the compound 5 was a formal proof of the success of the first step but also of the inability of the carbanionic intermediate to evolve into a carbenic species. Its final reprotonation (from a proton coming from the reaction medium) delivered then the neutral molecule 5. In order to assess the shelf stability and utility of salt 4, we tried to use it in the preparation of some amide derivatives 7, using an already described method for chlorodifluoroacetic acid (Scheme 3) [40]. molecule detected was the dichlorotrifluoromethoxymethane 5 [39]. Even if very volatile, we were able to isolate it by careful distillation (from a crude mixture with diglyme as solvent), yielding small amounts of pure product 5. The latter was characterized by NMR. The generation of the expected carbene 6 was assumed to proceed in a two-step pathway: first a decarboxylation followed by the elimination of a chlorine atom. The presence of the compound 5 was a formal proof of the success of the first step but also of the inability of the carbanionic intermediate to evolve into a carbenic species. Its final reprotonation (from a proton coming from the reaction medium) delivered then the neutral molecule 5. In order to assess the shelf stability and utility of salt 4, we tried to use it in the preparation of some amide derivatives 7, using an already described method for chlorodifluoroacetic acid (Scheme 3) [40].

Scheme 3. Preparation of various dichlorotrifluoromethoxy acetamides.
As shown by the results depicted in the Scheme 3, these derivatives were isolated in relatively modest but satisfactory yields comparable to those obtained with the sodium salt of chlorodifluoroacetic acid [40]. Aliphatic, benzylic and aromatic amines were suitable for this transformation. Obviously, sodium salt 4 exhibited sufficient stability for normal handling. To the best of our knowledge, none of the compounds 7a-e have been described so far. The reagent 4 is consequently a new building block for the introduction of a trifluoromethoxy moiety.

General Information
Each reaction was carried out under an argon atmosphere in freshly distilled solvent, unless otherwise noted. All chemicals were purchased from commercial sources (Sigma-Aldrich, Saint-Quentin Fallavier, France; ABCR, Karlsruhe, Deutschland or Alfa Aesar, Haverhill, MA, USA) and were used without further purification. Organic solvents were purchased from Merck (Darmstadt, Deutschland) and Carlo Erba (Val-de-Reuil, France). NMR spectra were recorded on AC-200 and AC-300 spectrometers (Bruker, Wissembourg, France). Reported coupling constants and chemicals shifts were based on a first order analysis. Internal reference was the residual peak of CHCl3 (7.26 ppm) for 1 H (200 MHz), central peak of CDCl3 (77.1 ppm) for 13 C (50 MHz) spectra, and internal CFCl3 (0 ppm) for 19 F (188 MHz) NMR spectra. Chemical shifts are reported in parts per million (ppm) and constants J in hertz (Hz). Mass spectra (MS) in the positive ion mode (ESI+) were obtained on a Xevo Q-Tof instrument (WATERS, Guyancourt, France). IR spectra were recorded on a Nicolet 400SD spectrophotometer (Thermo Fisher, Villebon-sur-Yvette, France).

Scheme 3. Preparation of various dichlorotrifluoromethoxy acetamides.
As shown by the results depicted in the Scheme 3, these derivatives were isolated in relatively modest but satisfactory yields comparable to those obtained with the sodium salt of chlorodifluoroacetic acid [40]. Aliphatic, benzylic and aromatic amines were suitable for this transformation. Obviously, sodium salt 4 exhibited sufficient stability for normal handling. To the best of our knowledge, none of the compounds 7a-e have been described so far. The reagent 4 is consequently a new building block for the introduction of a trifluoromethoxy moiety.

General Information
Each reaction was carried out under an argon atmosphere in freshly distilled solvent, unless otherwise noted. All chemicals were purchased from commercial sources (Sigma-Aldrich, Saint-Quentin Fallavier, France; ABCR, Karlsruhe, Deutschland or Alfa Aesar, Haverhill, MA, USA) and were used without further purification. Organic solvents were purchased from Merck (Darmstadt, Deutschland) and Carlo Erba (Val-de-Reuil, France). NMR spectra were recorded on AC-200 and AC-300 spectrometers (Bruker, Wissembourg, France). Reported coupling constants and chemicals shifts were based on a first order analysis. Internal reference was the residual peak of CHCl 3 (7.26 ppm) for 1 H (200 MHz), central peak of CDCl 3 (77.1 ppm) for 13 C (50 MHz) spectra, and internal CFCl 3 (0 ppm) for 19 F (188 MHz) NMR spectra. Chemical shifts are reported in parts per million (ppm) and constants J in hertz (Hz). Mass spectra (MS) in the positive ion mode (ESI+) were obtained on a Xevo Q-Tof instrument (WATERS, Guyancourt, France). IR spectra were recorded on a Nicolet 400SD spectrophotometer (Thermo Fisher, Villebon-sur-Yvette, France).

Conclusions
An easy access to the sodium salt 4 of dichlorotrifluoromethoxyacetic acid was devised. Attempted trapping of chlorotrifluoromethoxycarbene generated by decarboxylation of this salt with alkenes failed presumably because of the poor reactivity of this carbene under the conditions used for its formation. Nevertheless, salt 4 proved sufficiently stable for the preparation of new trifluoromethoxylated-bearing amide synthons 7a-e. Improved precursors of trifluoromethoxycarbene are under current development in our laboratories. We are studying in particular the preparation of chlorofluorotrifluoromethoxyacetic acid.