A novel and efficient synthetic route to perfluoroisobutyronitrile from perfluoroisobutyryl acid

A novel synthetic route to perfluoroisobutyronitrile from perfluoroisobutyryl acid which has mild conditions and low toxicity is described. This study introduces detailed synthetic protocols and characterization including GC-MS, 13C NMR and 19F NMR spectroscopy of perfluoroisobutyryl acid, perfluoroisobutyryl chloride, perfluoroisobutyl amide and perfluoroisobutyronitrile. Besides, this route is superior to the established patent and shows potential application in high voltage electrical equipment.


Introduction
The gaseous compound sulfur hexauoride (SF 6 ) has a high dielectric strength (DS) and a good arc extinguishing performance. These features allow it to be widely used as a mainstream insulating gas which was previously used in GIS (Gas Insulated Switchgear), GIL (Gas Insulated Lines) or other large power equipment. However, it is also extremely harmful to the environment and the atmosphere because it takes a long time to decompose when it has been in the air. The GWP (Global Warming Potential) of SF 6 is around 23 000 times that of CO 2 . It is expected to result in a stronger greenhouse effect.
The Kyoto Protocol, the supplement of the United Nations Framework Convention on Climate Change has declared SF 6 as one of the six limiting gases. In order to overcome this issue, participating countries are urgently striving to nd a new alternative compound with a low GWP as well as high DS that can completely replace SF 6 . [1][2][3] On-going research in General Electric (GE), America has led to the emergence of some new insulating gases like CF 3 I, c-C 4 F 8 , i-C 5 F 10 O, and i-C 4 F 7 N. There are handful reports about the synthesis of i-C 4 F 7 N, furthermore most of them aren't practical to apply in industry. [4][5][6][7][8][9][10][11] The gas i-C 4 F 7 N, when mixed with CO 2 , exhibits excellent properties, which surely be one of the rst choices to replace SF 6 . Besides, the GWP of i-C 4 F 7 N is only 2210, which is far lower than that of SF 6 . Under the same pressure, insulation strength and arc extinguishing performance of i-C 4 F 7 N/CO 2 can reach 90% or higher than that of SF 6 .
Recently, GE has built a 420 kV i-C 4 F 7 N/CO 2 GIL. 3 M (Minnesota, Mining and Manufacturing) has issued a patent on the synthetic route to i-C 4 F 7 N (Scheme 1a). 12 According to their protocol, methyl peruoroisobutyrate was used as the initial starting material which was made to react with NH 3 to form peruoroisobutyl amide. The dehydration of the per-uoroisobutyl amide resulted in the formation of per-uoroisobutyronitrile. The dehydration of peruoroisobutyl amide to produce its corresponding nitrile is an established synthetic strategy.
Sureshbabu had reported the general protocol for the synthesis of Boc-amino nitriles (Scheme 1b). 13 The method was also found to be suitable for the synthesis of peruoroisobutyronitrile.
Earlier attempts by Ishikawa to prepare organometallic compounds from peruoroalkyl iodides and carbon dioxide yielded peruoroalkyl acid as the organic product. 14 Ishikawa had explained the mechanism for the formation of per-uoroalkyl acid and reported the 19 F NMR of a series of per-uoroalkyl acids. However, the 19 F NMR data of C 3 F 7 COOH had certain inaccuracies. Ishikawa had taken C 8 F 17 COOH as an example, and explained the detailed procedure of its preparation. Even so, the physical properties of C 3 F 7 COOH and C 8 F 17 COOH such as the boiling point or melting point are very different. In fact, it was difficult to obtain peruoroisobutyryl acid following the general procedures given by Ishikawa. To the best of our knowledge, only a handful of reports are available regarding the synthesis of peruoroisobutyryl acid. One of the prime reasons behind this is the most of the researchers focused on the formation of peruoroisopropylzinc iodide, but actually the concentration of CO 2 also performed an important factor in the reaction. The scenario here is somewhat different from other organometallic reactions. We had even tried to purchase peruoroisobutyryl acid from some commercial chemical companies, but interestingly, none of them were successful in synthesizing it, in spite of following the reported protocols. The other reason for this compound being so less studied is the fact that only a little was known about its practical value previously. However, considering Kyoto Protocol, per-uoroisobutyronitrile has a widespread application in GIL and other large-scale electrical equipment. Hence, it is highly likely for peruoroisobutyryl acid to become a new raw material for the preparation of peruoroisobutyronitrile.
So far, no study exists which reports the synthesis of per-uoroisobutyronitrile from peruoroisobutyryl acid. Besides, there are no reports on the properties of peruoroisobutyryl chloride. This study would be the rst of its kind to report the details.

Results and discussion
Our attention was drawn to a novel synthetic route to i-C 4 F 7 N from a safer raw material where peruoroisobutyryl acid was initially converted to the corresponding chloride which was a highly reactive intermediate (Scheme 2). The synthetic route to peruoroisobutyryl acid (1) was shown in Fig. 1. In our work, the peruoroisopropyl iodide was found to react readily with zinc powder and carbon dioxide in an aprotic solvent under an ultrasound condition. In additions the reaction could also take place under high pressure. In this study, we had used the two methods to prepare 1 (detail data and results of various conditions were shown in ESI †) and listed their conditions and yield.
For this reaction, carbon dioxide was found to react with zinc compounds only under ultrasound and high pressure which is 1 MPa or higher. Increasing the concentration of CO 2 in the system determined the success of the step. Ultrasonic condition or high pressure has been proved to be indispensable, as the desired product could not be obtained under normal pressure without ultrasound or high pressure. Nevertheless, when the solution and dry ice were placed in a closed cylinder with a pressure of 1 MPa or higher, the formation of per-uorocarboxylic acid salt was observed. The summarized results were shown in Table 1.
Peruoroisopropyl iodide was found to react readily with zinc powder in solution to form peruoroisopropyzinc iodide, Scheme 2 The synthetic route in this study.  which further reacted with carbon dioxide under ultrasound or high pressure to give a nal yield of nearly 70% 1. It was found that the synthesis at the reux temperature was easier when the halide atom was iodide rather than bromide. On the other hand, the thermal stability of peruoroisopropylzinc iodide in solution was in striking contrast to the instability by per-uoroisopropyzinc bromide. 15 The amount of metal powder used in the reaction ranged from 1 to 4 equivalents of the raw material. It is evident that the yield of the reaction reached a plateau when the amount of the metal reached to more than 4 equivalents of the halogen in the raw material. Besides, the excessive metal powder in the solution rendered the separation step more difficult. In contrast, if the amount of the metal was less than 1 equivalent, the nal solution was found to have a substantial amount of the starting material, since the metal powder was not enough to react with the entire reactant. In order to maximize the rate of conversion, and reduce the complexities during the separation, the most suitable amount should be 3 equivalents of the raw material. The zinc powder (diameter ranged from 1-3 mm) should be fully activated by 1 mol L À1 HCl (aq) and washed by acetone and ethanol. Then it was stored in the dry vacuum case until to be used.
Aprotic solvents such as N,N-dimethylformamide (DMF), tetrahydrofuran (THF), or diethyl ether used have been studied. DMF was found to be a convenient solvent because it could be hydrolyzed directly by water. Diethyl ether has an added advantage because its boiling point is lower than per-uoroisopropyl iodide. 16 As a result, no remarkable amount of peruoroisopropyl iodide was lost at the reux temperature. Owing to the high saturation vapor pressure of diethyl ether, the combination of the solution and carbon dioxide became difficult. When using THF as the solvent, we found that per-uoroisopropyl iodide could not be dissolved in it. The purication of the solvent by molecular sieve was benecial for the reaction. Direct use of anhydrous solvent was also a convenient method and gave satisfactory results.
The concentration of raw material in the solvent ranged from 1 mol L À1 to 3 mol L À1 . The solvent-free method seems to be unsuitable because of the instability of peruoroisopropyl iodide which evaporates quickly under ultrasound or high pressure due to the generation of heat. This would obviously result in the loss of a massive amount of raw material, thereby lowering the yield. If a very high quantity of the solvent is used, the concentration of raw material will become less than 1 mol L À1 , which is not desirable. A high quantity of the solvent would also require a larger quantity of hydrochloric acid in the next step while lowering of the pH to enable the acidication of the peruorocarboxylic acid salt.
We found that the most suitable reaction temperature was 60 C. In order to prove the most suitable temperature, we carried out the reaction over a temperature gradient. The corresponding results are summarized in ESI. † The result illustrated that the reaction could not take place at low temperatures. The main impurity was found to be the raw material itself when the temperature was below 60 C. Below 60 C, most of the 1 will decompose to form hexauoropropylene (HFP) in the distillation step. 17 This would ultimately reduce the quantity of 1. In contrast, if the temperature exceeded 70 C or higher, most of raw materials was lost during the reaction. Furthermore, the solution could not maintain the required concentration of carbon dioxide, as a result of which high conversion rates could not be obtained. In addition, the stability of the peruorocarboxylic acid metal salt decreases at higher temperatures. Therefore, it was necessary to nd an optimum temperature that could avoid thermal decarboxylation an effect that was accentuated by the decomposition to form HFP or dimers and trimers irreversibly.
Ultrasound has been already proven to be an essential requirement for the combination of carbon dioxide and metal compounds in promoting the reaction because ultrasound forms cavities that can readily dissolve carbon dioxide in solution. The pressure was also found to be an important factor in this study. It was found that the reaction could take place even without ultrasound, if the pressure was greater than 1 MPa. This observation could be justied on the basis of the fact that high pressures could reasonably increase the concentration of carbon dioxide in the solution. Either way, both the methods aimed at increasing the dissolution of the carbon dioxide in the solution.
The peruoroisobutyrylzinc compound was so stable that it could be stored at room temperature with hardly any other reactions occurring. However, heating over 100 C initiated its decomposition to form HFP. Hydrolysis of per-uoroisobutyrylzinc proceeded rapidly with the addition of water at room temperature. The products included Zn(OH)I and 1,1,1,2,3,3,3-heptauoropropane (C 3 F 7 HF). C 3 F 7 HF was a gas phase at room temperature while the Zn(OH)I formed a precipitate in the sodium hydroxide solution. The precipitate of Zn(OH)I disappeared when then pH of the solution was made acidic. The reaction of peruoroisobutyrylzinc compound with CO 2 could take place under ultrasound or high pressure to form uoroisobutyryl acid salt. The salt was converted into 1 aer the addition of hydrochloric acid in order to adjust the pH to be less than 2. The nal pure 1 was obtained aer being separated, dried and distilled.
The choice of solvent in the step concerning the synthesis of 2 was versatile. The solvents used could be DMF or THF, and interestingly (Table 2), the reaction could also progress without the presence of any solvent. Different solvent showed different results. For example, DMF was found to catalyze this particular synthesis. Addition of high amounts of DMF made the reaction more vigorous. If THF was used as solvent, the vapour pressure of the solvent needed to be considered. This is because THF would be present even in the distillate, although it could be removed in next step of synthesizing the peruoroisobutyl amide (3). The possible reason is that the boiling point of THF is lower than that of the amide. The absence of any solvent was acceptable. But storing the puried product would be difficult due to the vapour pressure and low boiling point.
The synthesis of 2 didn't take place at low temperature. Contrarily, the high temperature was also not favorable for the reaction because of the low boiling point and high vapor pressure of 2. So the most suitable temperature was determined to be the initial reux temperature. 18 These properties also suggested that the storage of 2 was difficult. Therefore, a separation device was used to collect 2, which was made to drip into a methanol ammonia solution for the synthesis of 3. This apparatus allowed the reaction to proceed without any pause, which otherwise would have been essential if the advanced purication of 2 had to be done. This set up reduced the loss of peruoroisobutyry chloride and improved the reaction efficiency. It therefore became convenient to purify the product. Once the 3 was prepared, the synthesis of peruoroisobutryl amide and peruoroisobutyronitrile was easy. Their synthesis could be proceeded using the routine procedures for amination and dehydration. The dehydration process of 4 in particular is very well established, and hence we have not included too many details here.
Compared to this study, the Scheme 3 as the mainstream synthetic route of methyl peruoroisobutyrate had many disadvantages and dangers. First, the raw material was chloroformate which was a highly toxic liquid that had been forbidden in many places. Second, the reaction in the gasliquid phase in the next step wasn't easy to be controlled and operated. In fact, there was another way to synthesize C4. The electrolytic uorination was also an effective route to prepare methyl peruoroisobutyrate. But the route was more dangerous and had too many by-products to be separated. Our raw materials were easier to be purchased, stored and prepared. Furthermore, it had low chemical toxicity and good stability. Obviously, the methodology in Scheme 2 was simpler and safer to be used in the laboratory and its efficiency and yield were also no lower than any other methods.

General remarks
Peruoroalkyl halides were purchased from SHANG FLUORO. Zinc powder was purchased from Macklin Company and its diameter ranged from 1-3 m, purity was over 99.99%. Nearly 99.999% pure nitrogen gas and carbon dioxide gas were purchased from the WuHanShi XiangYun Industry Co. Ltd. All the solvents were of analytical grade and used without further purication, unless otherwise stated. Ultrasonication was performed on a GT SONIC-P3 ultrasonic apparatus. Y2UR-250 type high-pressure reactor used in this study was purchased from ShangHai YanZheng Instrument Co. Ltd.
NMR spectra of 1 and 2 were recorded on a Bruker Avance-III NMR spectrometer ( 1 H: 400 MHz, 19 F: 376 MHz, 13 C: 100 MHz) with reference to an internal ( 1 H, 13 C: SiMe 4 ) and an external standard ( 19 F: CDCl 3 ). NMR spectra of 3 and C4 were recorded on at temperature 298 K on a Bruker Avance-III 500 MHz spectrometer equipped with a 5 mm BBFO probe, a 19 F frequency of 470.59 MHz and a 13 C frequency of 125.76 MHz. All 19 F spectra were recorded with a recycle delay of 2 s, sweep width of 242 ppm, 220k acquisition data points, 4 scans. All 13 C spectra were recorded with a recycle delay of 3 s, sweep width of 242 ppm, 32k acquisition data points, 256 scans. All NMR spectra were processed with 1 Hz line broadening.
GC-MS were carried out on Varian 450-GC Gas Chromatograph and Varian 320-MS TQ Mass Spectrometer. The Gas Chromatograph was equipped with a 30 m and 0.250 mm, 0.25 mm df, VF-5 column.

Synthesis of peruoroisopropyzinc iodide
19.5 g (0.3 mol, diameter 1-3 mm) zinc powder and 40 mL DMF was added to a round bottom ask containing a tube for introducing gas, a thermometer, a spherical condensing tube and a drop funnel. Nitrogen was introduced into the apparatus to prevent the system from air. 30 mL DMF solution containing 29.6 g (0.1 mol) peruoroisopropyl iodide was added to the aforesaid mixture. Then the solution was heated at 60 C for 2 h which included the feeding dropping time of about 1 h. To lter the metal powder, the peruoroisopropylzinc iodide was obtained in DMF.

Synthesis of peruoroisobutyryl acid (1) by ultrasonic method
To a round bottom ask containing a tube for introducing carbon dioxide gas, a thermometer, a spherical condensing tube and a drop funnel, was added metal powder and 40 mL DMF. The gas-introducing tube was inserted into the solution to ensure a better contact between carbon dioxide and the solution. The ultrasound was kept on until the reaction was over. The speed of introducing carbon dioxide was about 0.1 L min À1 . To the ask, 30 mL DMF solution containing 0.1 mol raw material was added into 19.5 g (0.3 mol, diameter 1-3 mm) zinc powder and 40 mL DMF. Then the solution was heated 60 C for 2 h which included a dropping time of about 1 h. The solution was ltered to remove the remaining metal powder. Aer the reaction, 100 mL of 6 N hydrochloric acid was added for the hydrolysis, and pH was adjusted to be less than 2. The solution was now separated in two phases. Aer the lower organic layer was distilled, dried, ltered and collected, deep brown per-uoroisobutyryl acid was obtained.

Synthesis of peruoroisobutyryl chloride (2)
To take into consideration the properties of peruoroisobutyry chloride, a water separation reux device was used. 10.8 g (0.05 mol) peruoroisobutyryl acid along with and 12.3 mL (0.15 mol, 3.0 equivalents) THF as solvent and 10% mol DMF as catalyst was added to a round bottom ask which included a thermometer, a drop funnel and downward distilling head with a serpentine reux condenser. Notably, the cooling medium in the condenser was better below À10 C. 5 mL (0.06 mol, 1.2 equivalents) of oxalyl chloride was dropped into the solution in 2 h at 40 C. As the reaction progressed, a massive colorless peruoroisobutyryl chloride was obtained below reux condenser. The yield was nearly 60%.
Colorless liquid; 13 (3) 11.6 g (0.05 mol) peruoroisobutyryl chloride solution was dropped into 22 mL 7.0 M (0.15 mol NH 3 in solution) NH 3 -$MeOH. The temperature was controlled below 20 C. The reaction time was around 1 h. Aer the reaction, the main solid by-product was ammonium chloride, which was henceforth ltered. There was still a small amount of ammonium chloride dissolved in methanol. Methanol was removed by distillation. To add chloroform into solution at room temperature and lter all solid by-product. Aer being kept overnight in the refrigerator, peruoroisobutyl amide crystals were found to precipitate out. The yield was 90%.

Synthesis of peruoroisobutyryl amide
Colorless crystal; 13

Synthesis of peruoroisobutyronitrile (C4)
10 g peruoroisobutyryl amide and 12 mL pyridine was put in 20 mL DMF. Then 13.8 mL triuoroacetic anhydride was dropped into the solution. The temperature was controlled below 0 C. The reaction time was around 3 h. The 4.5 g per-uoroisobutyronitrile was obtained in ice trap. The yield was 49.5%.

Conclusions
In summary, we have developed a novel method for the synthesis of peruoroisobutyronitrile from peruoroisobutyryl acid which is a safer and more convenient route than established patent. The compound is introduced as the new environment-friendly insulating gas as the replacement of SF 6 in high voltage electrical equipment, with the aim of abiding by Kyoto Protocol. Peruoroisobutyryl acid could be synthesized with ultrasound and high pressure over 1 MPa at 60 C. It reacted with (COCl) 2 at 40 C to form peruoroisobutyryl chloride that then react with NH 3 to get peruoroisobutyryl amide. Raw materials and intermediates are characterized by GC-MS, 13 C NMR 19 F NMR spectra in the study.

Conflicts of interest
There are no conicts to declare.