US2519983A

US2519983A - Electrochemical process of making fluorine-containing carbon compounds

Info

Publication number: US2519983A
Application number: US62496A
Authority: US
Inventors: Joseph H Simons
Original assignee: Minnesota Mining and Manufacturing Co
Current assignee: 3M Co
Priority date: 1948-11-29
Filing date: 1948-11-29
Publication date: 1950-08-22
Anticipated expiration: 1967-08-22

Description

` Aug. 22, 1195o .Li-1'. slMoNs ,519,983 ELEc'rRocmucAL PRocEss `oF MAKING FLUoRmE-cou'mumc cARBon-couPouNns Filed Nov. 29, 1948 www? Patented ug. 1950 ELECTROCHEMICAL PROCESS OF MAKING FLUORINE-CONTAININ G CARBON POUNDS COM- Joseph H. Simons, State College, Pa., assigner to Minnesota Mining & Manufacturing Company, St. `Paul, Minn., a corporation of Delaware Application November 29, 1948, Serial No. 62,496 20 Claims. (Cl. 204-62) This application is a continuation-impart of my copendingv application Ser. No. 677,407 (tiled June 17, 1946), since abandoned. The latter was led as a continuation-impart of the following prior applications which thereafter were abandoned in its favor: Ser. Nos. 384,729 (tiled March 22, 1941), 569,265 (led December 21, 1944), and 626,434 (led November 2, 1945).

This invention relates to my discovery of a basically new process of making fluorocarbons and other fluorine-containing carbon compounds. The process is simple in operation, does not involve the use or formation of free iluorine at any stage, and has great versatility in enabling the direct production of many. types of product compounds using organic starting compounds which are readily available and relatively inexpensive. It is an electrochemical process and its utility has been demonstrated by extensive pilot plant operations employing a 200G-ampere cell, as well as by many laboratory experiments.

This process was publicly disclosed in the oral presentation on September 15, 1948, at the Portland, Oregon, session of the 114th meeting of the American Chemical Society, of ve papers am thored by me and my assistants. Abstracts of these papers had been previously published in the advance Abstracts of Papers volume issued by the society (pp. 42-0 to 45-0). News items relating thereto and briefly describing the process have been published in the societys magazine Chemical and EngineeringNews, vol. 26, p. 2428 (Aug. 16, 1948) and pp. 2878-9 (Sept. 27, 1948). These papers were subsequently published in the Journal of the Electrochemical Society, vol. 95, No. 2, pp. 47-67 (Feb. 1949). y

Briefly stated, this new and useful electrochemical process involves electrolyzing a liquid hydrogen fluoride (HF) solution containing a fluorinatable organic starting compound, at an electrolyzing potential which is insuilicient to generate free fluorine under the existing conditions, but -which is suiilcient to cause the production of uorine-cont'aining carbon compound products at a useful rate. A wide variety of organic compounds are soluble in anhydrous liquid hydrogen `fluoride and provide adequate electrolytic conductivity. Use can also be made` jof soluble organic and inorganic conductivity additives to permit of electrolyzing liquidvhydrogen iluoride solutions thereof admixed with relatively insoluble organic starting compounds, such as alkanes, which do not provide adequate conductivity. Pure anhydrous liquid hydrogen fluoride, per se, is non-conductive. The electrolyte is free from water in more than a small proportion but need not be anhydrous in a strict sense. v

Excellent results can be obtained with simple single compartment electrolytic cell arrangements. No diaphragm is needed between electrodes. Fluorination can be completed in one step to directly obtain fully fluorinated product compounds which are relativelyv insoluble and either evolve with the cell gases or settle to the bottom of the cell from which they can be drained, depending upon their boiling points.

The process is suitable for continuous as well as batch operation. The cell can be operated at atmospheric pressure. The cell and the cathodes can be made of iron or steel, and the anodes of nickel, and such cells have been satisfactorily operated at 5 to 8.volts, D. C., in producing a wide variety of product compounds.

The reaction mechanism is not fully understood but the electrolyzing process apparently transforms the organic starting compounds at or adjacent to the anode and may be regarded as an anodic process. Hydrogen is evolved at the cathode, being derived from the hydrogen iluoride:

v and will also evolve in the cell by derivation from fully saturated product compounds.

other hydrogen-containing compounds which may be present. Fluorine is not evolved, and the reaction mechanism does not involve the formation of molecular (free elemental) uorine as an intermediate agent.

.Product compounds are obtainable (both cyclic and non-cyclic) which have the same number of y .a,s1o,ees A l 3 Y tainable from cyclic starting compoimds as the result of bond cleavagel and nuorine addition. Also. the use oi' polycarbon cyclic and non-cyclic starting compounds generally results in appreciablc yields of product compounds having fewer carbon'atoms in-the molecule, due to fraa'inenta-4 o'f the carbon skeleton and iiuorine addition.

' having a greater number of. carbon atiunaV than the starting compounds are obtainableas the result of oollpling'of formed in` the solution. Ihekinds and relative .of productcompoimdsiobtainable in lany'giv'en case will dependupon the lstarting compound andthe operating By properselection it is possible toobtain excellent fcwerormorecarbonatomstbanthepareut hydrocarbon groups can also be obtained from auch compounds. A feature of the present process isthat it provide a simple and economical procedure for making normally liquid fluorocarbonsghavingnveormorecarbonatomsinthe molecule.

Y The formation of iiuorocarbons is accompanied ily: .the formation of iluorocarbon hydrides and partially fluorinated hydrocarbons, due to incomplete hydrogen remacement in the case of some molecules. 'The proportions depend upon u condition. The fiuorocarbon monohydrides and dibydrides are of particular value because they yieldsofdcsired productawhichcanbe I by fractional distillation from each other andfrom low-yield by-product compounds.

Y In commercial operations, use canbe made of the production of multiple' products in `building up have a high degree of thermal stability and chemical inertncss and are non-combatible, and

l yet thehydrogen atoms offer points of attack for of iiuorocarbon derivatives. Thus ytheaeiiuorocarbon hydrides can be thermally stocks offmanyuseful compounds from afmuch lmore limited number of starting compounds.

In somecascs resinous material is formed in the electrodes. Y

N9 (June 1946) of which I was a iso-author. VBaturad nuorocarbons and iiuorocarbon derivatives are obtainable in good yields ranging and brominated to yield the corresponding iiuorocarbon chlorides and bromides. as in 'a paper published in the Journal of vAmerican Chemical Society, voLB, pp. 968- from the simplest compound, Cm, which is a lowboiling'gas. up to high-boiling compounds havvIo briei'ly indicateat thisv point 'I of the process in'enabling the productionv of iiuorocarbon derivatives, mention is made (for .illustration and notlimitation) of the following types designated byi generic-formulae wherein l?. represents saturated iiuo'rocarbon radicals (cyclic or non-cyclic) consistingsolely of carbon'and aucune atoms: R'on" (obtainable from ethers).

R'R"R."'N (obtainable vfrom-tertiiu'yg amines), R'R."NP (obtainable from secondary amines), RNE (obtainable from-primary amine's),R.COli'l (obtainable from monocarboxylic acids),

R'COOR.'

.substantial yields of duorocarbons having the same number of carbonatomsas the starting compounds. For example, alkanes. alcohols and mcnocarboxylic acids lcan be employed to yield uorocarbons having the same number of carbonv In addition, tluorocarbons having fewer and more carbon-atoms are produced. l'iuorocarbons havingfewer carbon atoms thanthe vstarting compound can also-be produced from :hetero compounds wherein carbon atoms are linked by non-carbon atoms, asin the'case of secondary'and 'tertiary amines; the

groups being released to form coriiuorocarbons. vll'iuoi'ocarbons'having lo -ing eight or more carbon atoms in the molecule and having boiling points above 100 C.

The fiuorocarbon and related compounds ob- Aminable by this process have utility for many purposes. The saturated i'iuorocai'bonsI nuorocarbon hydrides, fiuorocarbon ethers, and trifiuorocarbon amines, for example. have a high degree of thermal stability and chemical inertv ness. they do not burn, and they have uniquely Y low refractive indices and dielectric constants as compared with ordinary organic compounds. The liquid compounds (containing five or more carbon" atoms) have exceptionally low surface tensions and viscosities, coupled with high densitics. These and other compounds can be used for such purposes (depending upon boiling' points) as refrigerants, ilre extinguishers, inert solvents and diluents. heat exchange iluids, turbine impellents, dielectrics, hydraulic mechanism iluids, and lubricants. The products of this process can also be used as-intermediates in the synthesis of other chemical compounds. In particular, the fluorocarbon acid uorides (RCOF) provide the reactive starting compounds formaking a vast number of derivative compounds containing a tluorocarbon radical and having unique properties on that account.

All organic compounds potentially capable of reacting with iiuorine can be employed to produce ilumine-containing carbon compound products by the present process. Fluorine is the most chemically active substance. No organic material is completely resistant to iluorine except carbon tetrailuoride (GF4). Even the polycarbon saturated fiuorocarbons can react by carboncarbon bond cleavage and iiuorine addition, the stable end product being CFs.

As a practical matter inrespect to the commercial utilization of the process, the hydrogencontaining organic compounds which contain at least one hydrogen atom bonded to acarbonatom in the molecule are of main interest. These can all be used for 'making iluorocarbons. and other iiuorine-containlng carbon compounds which Vcontain at least-one fluorine atom bonded to a carbon atom in the molecule. The sub-classes of these organic shrting compounds of particular value for making iluorocarbons, and fluorocarbon derivatives of chief commercial interest. are the hydrocarbons (which contain only carbon and hydrogen), and the compounds which are oxygen-containing or nitrogen-containing in addition to being hydrogen-containing. Exemplary of the latter are the cyclic and non-cyclic organic acids, alcohols; ethers, esters, ketones, aldehydes, nitriles, amines, amides, etc. These oxygen-containing and nitrogen-containing organic starting compounds are soluble in anhydrous liquid hydrogen fluoride and provide adequate conductivity, no conductivity additive being needed. The hydrocarbons (which are relatively insoluble) can be employed when use is made of a conductivity additive, as explained in some detail later on. These sub-classes embrace the readily available and relatively inexpensive alkanes, alcohols, ethers, carboxylic acids (including the anhydride forms), and amines, both cyclic and non-cyclic, which can all be emciently employed in the preparation of iluorocarbons and of other ilumine-containing carbon compounds.

Conjoint use can be made of4 two or more organic starting compounds and in some cases this will result in improving the operating emciency. For example, the efilciency of the electrolysis of a solution of hexyl ether can be improved by including propionic acid as a second organic starting compound. Y

The proportion of the organic starting material relative to the hydrogen fluoride can vary over a wide range and it can be either a minor component or a major component of the solution.

As previously mentioned, this process does not involve the use or generation of free (molecular) fluorine at any stage. It is fundamentally different from processesA wherein free fluorine is reacted with organic compounds which are dissolved in or mixed with liquid hydrogen fluoride, as by introducing gaseous fluoride into the reaction zone. For any given cell arrangement and charging stock there is a minimum cell voltage at and above which molecular fluorine (free elementary fluorine) is formed from the hydrogen fluoride, but the resultant effects are quite different from those obtained in the present process and can be easily recognized. When the cell voltage is raised to this critical `value or higher, the onset of free fiuorine formation manifests itself unmistakably by a marked change in product formation due to the rapid and powerful decomposition and fragmentation effects on the organic compounds in the cell; extensive corrosion of the electrodes occurs; some gaseous iluorine escapes and can be detected in the exit gas mixture even when present in trace amounts; and minor and even major explosions occur. Satisfactory yields of fluorinated compounds having a substantial number of carbon atoms in the molecule cannot be obtained.

The present process is based upon my discovery that liquid hydrogen fluoride solutions containing organic compounds can be eiilciently electrolyzed to produce fluorocarbons'and other iluorine-containing carbon compounds, by using a cell voltage which in any given case is below the minimum voltage required for free fiuorine generation. For example, whenrusing nickelanode types of cells for which operating voltages in the range of 5 to 8 are quite satisfactory. it has been found that the voltage must be increased to at least to 12 before evidence of free iluorine generation occurs.

The minimum theoretical cell voltage required for the generation of free iluorine (Fn) vis ap-g.v

proximately 3 volts. The reason why total cell voltages higher than this can be used in actual practice ofthe process without causing free fluorine formation is because a substantial part of the total cell voltage (measured across external anode and cathode connections) is utilized under the existing conditions in overcoming the resistances of the electrodes, the resistance of the electrolyte between the electrodes, and the blocking resistances of electrode polarization films.

In an experiment using a carbon-anode cell it was found that electrolysis of a hydrogen fluoride solution of acetic acid resulted in the formation of fiuoromethane (CHsF) at a cell voltage as low as about 0.5 volt. In another experiment using a nickel-anode cell, a hydrogen fluoride solution of acetonitrile (CHsCN) was electrolyzed for five days at a cell voltage of 1.7 volts and the products included CFaCN, CFaH, GF4, CzFsI-I and CzFe. These cell voltage values were very substantially lower than the minimum cell voltage (approximately 3 volts) required for generation of molecular iluorine even under ideal conditions. Hence these experiments conclusively demonstrate the utility of electrolyzing potentials which are insuillcient to generate free iluorine and yet are sufiicient to cause the production of fluorinecontaining carbon compounds.

In a review paper written by me and published in 1931 (Chemical Reviews, vol. 8), long prior to the discovery of the present process, I referred to and summarized previously published experimental data of Fredenhagen and Cadenbach on the solubilities and equivalent conductivities cf various inorganic and organic compounds dissolved in anhydrous liquid hydrogen fluorideA (p. 223, et seq.). The well-known laboratory procedure for measuring conductivities of solutions employs a small cell provided with platinum electrodes and utilizes a rapidly reversing current (A. C.) of 1,000 cycles or more persecond to substantially eliminate electrode polarization effects andy to cause momentarily formed electrode products to revert with eachalternation of electrode polarity rather than to escape from the electrodes as free products. In such experiments the current is passed, for each measurement of voltage versus current flow, during a very short time interval (a few minutes at most and often a few seconds or less), for otherwise accurate values for a desired temperature cannot be obtained due to heating of the electrolyte and for other reasons.

Fredenhagen and Cadenbach stated in their paper (Z. physikal. Chem., Abt. A., vol. 146, pp.

245-) that they employed a conductivity cell vessel made of a gold-platinum alloy, which served as one electrode, and the other electrode was a platinum capillarytube extending through the stopper down into the vessel. The cell was a small one and contained 20.7 cc. of solution (approximately 1.3 cubic inches) when making a measurement. That they employed alternating current is made clear by the absence of any description oi?y an unconventional procedure and by their use of a telephone receiver in balancing the bridge. Nothing was said to suggest that the total time of current flow for any given solution amounted to more than a few minutes at most, even when several measurements were made at intervals. Each solution had served its purpose and was discarded when the measurement work had been completed. There was no suggestion that detectable trace amounts of freeA ananas trolyte resistance, as well as altering the com--v position of the solution being tested. These authors speculated as to the manner 'of ionic dissociation of the organic compounds when dissolved in the anhydrous liquid hydrogen fluoride, and on the possible formation of reaction products resulting therefrom, in commenting upon the various conductivity values, since the magnitudes of the latter would depend upon what it was whose conductivity was actually being measured. This had reference only to substances formed as the result of dissolving the original organic compounds in the anhydrous liquid hydrogen fluoride and not to the formation of free electrolysis products.

A mere scientific conductivity determination is not a useful process for making electrolysis products from a solution of a starting compound 8 notoniyfluorocarbomandiiuoroearbonhydridn but also nuorine-contaimng carbon cunpmds whichcontainoxygenornitrogenatomsinths molecule.

Commercial anhydrous liquid hydrwm finoridenormallycontainsatraoeofdissolvedwatsr, and traces of water are readily absorbed whm .fthereiscontact withamoistatmosphere. Buch watercanbereadilyremovedorreduoedtoavery low value by passing current through the liquid hydrogen fluoride prior to adding the starting compound, the water beingv consumed by the formation of hydrogen, oxygen nuoride (Oh) and oxygen which evolve as gases. In actual practice, the first few hours of electrolysis of the hydrogen fluoride solution containing the soluble organic starting compound will generally cause the proportion of water to diminish to 0.1% orless anditisnotnecessarytosubiectoommerciai anhydrous hydrogen fluoride to a pruiminary water-removal treatment. The expression anhydrous liquid hydrogen fluoride as used whose conductivity is being measured, and the objectives are antithetical. It will be understood that no attempt is being made herein to claim mere conductivity measurements.

When laboratory size cells are used in performing batch experiments on the present process, the runs continue for periods of hours (generally for a day or more). Likewise the commercial operation of the process using large cells requires electrolyzing the cell charge for a period of many hours in order to obtain a commercially desirable yield ratio of product compounds from the starting material, and a cell will normally be continuously operated for days before being shut down, the hydrogen fluoride and starting material being replenished during operation.

Alternating currents can be usefully employed 4inV practicing the present process when polarizable electrodes (e. g., nickel electrodes) and a relatively low frequency (e. g., 60 cycles) are used. In this case each electrode alternately functions as an anode and as a cathode, but the operating conditions permit the release of free electrolysis products at the electrodes. The use of direct currents is greatly preferred. The efilciency of alternating currents is much inferior and greater heating of'the electrolyte is4 produced. In the case of normal direct current operation each cathode and anode electrode continuously functions as such at a uniform voltage (which may be varied during the run for optimal operation), and the cathodes and anodes can be made of .different materials adapted to produce the moet emcient results. Pulsating vunidirectional currents, -unfiltered rectified A. C., and superimporsed A. C. onD.C.,canalsobeusedandaretoberegarded as types of direct currents.

As previously mentioned, pure anhydrous liquid hydrogen fluoride is non-conductive but a wide variety of organic starting compounds are soluble therein and provide adequate electrolytic conductivity to permit the passage of eective electrolysing currents. The starting oompoimds of chief commercial interest in this respect are the oxygen-containing and nitrogen-containing organic compounds which have at least one hydrogen atom attached to a carbon atom in the molecule, of which various examples have already been given. and from these them een be obtained u herein is employed, except when otherwise stated. to include liquid hydrogen fluoride which omtains traces of water notexceeding 1% by weight. Idquid hydrogen fluoride containing 1% water has a low conductivity, being about 0.00 mho (reciprocal ohm) per centimeter cube at 0 C.

The conductivity when there is a 0.1% water oonicnt is about 0.006 mho.

When the proportion of dissolved water in the liquid hydrogen fluoride is increased substantially above 1% there is a material decrease in the operating eillciency of the process in various respects. While a small proportion of water can be used to advantage as a means of increasing the conductivity, a point shortly to be discussed in connection with the use of conductivity additives, the presence of water in proportions ex-v ceeding 10% by weight of the total of water and hydrogen fluoride results in such marked inemciency that it should fbe' avoided for economical operation of the process and the obtaining of desirable yields of iluorocarbon compounds. Water in amount substantially exceeding 10% also gives rise to serious explosion hasards due to the evolution of readily explosive gas mixtures containing substantial proportions of oxygen fiuoride and oxygen mixed with hydrogen. Some organic starting compounds may contain dissolved water in the commercial forms employed (as in the case of ethyl alcohol) and in such cases the starting compounds should be employed in proportions which will not result in there being more than 10% by weight of water relative to the total weight of water. and hydrogen fluoride. It will be understood that the generic term liquid hydrogen fluoride." when not restricted by the anhydrous" adjective, includes not only anhydrous liquid hydrogen fluoride but also solutions thereof which contain small proportions of water in excess of 1% as above indicated.

Some organic compounds, such as acetic add, react with liquid hydrogen fluoride to form an equilibrium mixture containing water and the anhydride of the starting compound, such as acetic anhydride. In such cases the theoretical total amount of water capable of being formed from the starting compound may exceed the 10% value previously mentioned without causing an uneconomically low operating emciency, since it is not the mere equivalent of free water Iadded allatonceiotheliquidhydrogeniluorlde.-

Liquid hydrogen fluoride containing a small proportion of dissolved water, as described above. is also entirely different in other respects from the aqueous solutions of hydrogen fluoride which are sold in bottles and are commonly referred to as hydroiluoric acid. These contain less than 50% by weight of hydrogen uoride as normally sold. Such solutions cannot be kept in glass or metal containers and if introduced into electrolysis cells such as are used in the present process (for example, a steel cell having nickel anodes and iron cathodes) would cause rapid corrosion of the metals contacted even in the absence of current flow. Whereas the boiling points of hydrofiuoric acid solutions containing 50% by weight of hydrogen fluoride or less are in the range of 100 to 114 C., the boiling point of pure anhydrous liquid hydrogen fluoride is 1.0 to a maximum value of about 1.26 at about 75% hydrogen fluoride and then rapidly decreases to a value of 1.0 for anhydrous liquid hydrogen fluoride (measured at 0 C). The conductivity decreases as the hydrogen fluoride percentage increases above 75%, and rapidly so as the percentage increases above 90%. These data are of interest as showing that liquid hydrogen fluoride, both when anhydrous and when containing a small proportion of dissolved water, as employed in the present process, differs radically in physical as well as chemical properties from hydroiluoric acid solutions which contain a large proportion of water and which cannot be employed in the present process.

The present Iprocess is not limited to the use of organic starting compounds which are soluble in anhydrous liquid hydrogen fluoride and provide adequate electrolytic conductivity to permit of effective electrolysis. I have made the surprising discovery that the necessary electrolytic conductivity can be provided by a third component lute which can be either organic or inorganic, so that organic starting compounds which are insoluble in the liquid hydrogen fluoride can be effectively employed in admixture with liquid hydrogen fluoride solutions containing a dissolved conductivity additive. This feature of the generic process greatly extends its versatility and usefulness. Thus it makes possible the use of very cheap organic compounds, such as the hydrocarbons, as starting materials for producing uorocarbons and uorocarbon hydrides.

Examples of hydrocarbon starting compounds which can be effectively utilized when a conductivity additive is employed, are the alkanes (both cyclic and non-cyclic). The aromatic compounds, such as benzene and toluene, can also be employed. Moreover, this procedure permits of good yields of fiuorocarbon compounds having the same number of carbon atoms as the starting compound. Thus hexane can be used for making CsFn (tetradecailuorohexane), a normally liquid iluorocarbon. As in the procedures utilizing soluble organic starting compounds, the

electrolyzing potential employed is insufficient to process is not anequivalent of direct fluorination processes but is an entirely different kind of process which does not involve the use or evolution of free (molecular) iluorine and which produces product compounds that cannot be obtained in significant proportions by direct fiuorination with free fluorine.

The exact mechanism of the electrolyzing process remains a, mystery. The gases and heat levolved in the cell vigorously agitate the cell contents so that a mixture of the insoluble organic starting compound and the liquid hydrogen fluoride electrolyte solution is present at the anode, and it is a fact that the existing conditions cause the formation of fluorine-containing carbon compounds, including fluorocarbons, at the anode.

The conductivity additives fall into two general groups, organic and inorganic.

'I'he organic conductivity additives are compounds which can themselves be employed as starting compounds, and numerous examples have already been indicated. The most important are the oxygen-containing and nitrogencontaining organic compounds which contain at least one hydrogen atom attached to a carbon atom of the molecule. Illustrative examples are: acetic anhydride, acetic acid, propionic acid, caproic acid, methyl alcohol. diethyl ether, triethyl amine, pyridine, and 2-fluoropyridine, which have all been employed in experiments wherein hydrocarbon starting compounds were converted into iluorocarbons and fluorocarbon hydrides. The proportion ofthe conductivity additive to the hydrogen fluoride can be relatively small, e. g., 1/2 to 5%, but much higher proportions can be used. These additives are themselves converted during operation of the process, and hence the iiuorine-containing carbon compound reaction products are derived in part from the soluble conductivity additive and in part from the insoluble organic starting compound. Thus,

in effect, there are two types of organic starting vcompounds utilized in this procedure, one being soluble and serving also as a conductivity additive and the other being insoluble. For example, when hexane is employed in admixture with liquid hydrogen fluoride which contains dissolved pyridine, both the hexane and the pyridine are converted to produce fluorine-containing carbon product compounds, including fluorocarbons. Another example of an organic conductivity additive is mercurio cyanide, which also illustrates organic starting compounds that do not contain hydrogen.

The inorganic conductivity additives are inorganic compounds which are soluble in liquid hydrogen fluoride and provide effective electrolytic conductivity. The soluble uorides, such as sodium fluoride and potassium fluoride, can be used and have the advantage of not decomposing and not being consumed. Ammonia can be used but is gradually consumed, yielding nitrogen trifiuoride (NFa) which is a low-boiling relatively insoluble gas and is evolved. Water can be used in limited amount. The proportion of Water in the liquid hydrogen uoride solution should not exceed a small proportion because otherwise there will be poor efficiency and a poor yield of desired compounds. Eiective results can be obtained using as little as 1 to 2% of added water. The water is gradually consumed, yielding oxygen fluoride (OFn), carbon dioxide (CO2) oxygen and hydrogen, and must be replenished during extended runs. Some organic conductivity additives also suena provide water so that their use actually involves both organic and inorganic additives. Thus acetic acid is dehydrated by hydrogen fluoride to produce acetic anhydride and water, and alcohols may contain a small amount of dissolved water. Further illustrative examples are sulfuric acid and phosphoric acid. The illustrative inorganic additives named above have all been employed in experiments wherein hydrocarbon starting compounds were converted into fluorocarbons and fluorocarbon hydrides, including significant yields of fully fluorinated product compounds having the same number of carbon atoms as the starting compound. Thus Calm has been made from octane, and C-:Fu and Crl'u from toluene, for example.

Various combinations of two or more conductivity additives can be employed. Thus 'an organic additive and an inorganic additive can be jointly employed.

In some cases the soluble conductivity additive (depending upon the kind and amount used) will function as a mutual solvent so as to cause an otherwise insoluble organic starting material to be appreciably soluble in the electrolyte solution. This action does not appear to be necessary to the operation of the process, since useful results are obtained notwithstanding the insolubility of the starting compound.

The principle of using a conductivity additive can also be employed to increase the efficiency even when the organic starting compound is sumciently soluble so that it could be used without such expedient. This is helpful in the case of starting compounds of limited solubility in anhydrous liquid hydrogen fluoride and which do not provide an efficient conductivity.

Another expedient is to change the operating temperature or pressure or both to cause a more favorable solubility relation. Thus a starting compound which is poorly soluble or relatively insoluble in liquid hydrogen fluoride at 0 C. mayI be made much more soluble therein by substantially increasing the temperature. while the solubility of normally gaseous compounds can be increased by using lower temperatures.

The accompanying drawing is a diagrammatic sectional elevation of an illustrative apparatus for carrying out the process and also serves as a flowsheet. The drawing has been simplified in the interest of clarity by omitting heat-insulating Jackets and coverings, couplings, etc., and by employing conventionalized representations as commonly used in schematic chemical drawings. 'I'he apparatus as shown is suitable for laboratory experiments and can also in larger form be employed for pilot plant or small-scale commercial use. A large commercial plant would necessarily be designed on more complex lines, as chemical engineers will understand, but would embody the basic principles of the invention as illustrated and dcribed.

Referring to the drawing, the heart of the apparatus is the electrolysis cell I, which is provided with a cooling jacket 2 having an inlet I and outlet l for flow of a coolant liquid (such as an aqueous solution of ethylene glycol or methanol) which is pumped through the jacket and recycled through a refrigerator (not shown), so as to maintain a desired cell operating tempera-v ture. An operating temperature in the neighborhood of 0 C. is convenient and permits of simple operation at substantially atmospheric pressure. The cell has a removable cover i which can be Suspended fromthecovex-soastobelocated withinthecellwhenthecoverisinpositionis the electrode pack consisting of a series of anode plates and an alternating series of intermeshed cathode plates. An extreme outward anode plate l and cathode plate 1 are shown, the other plata being to the rear and not visible in the drawing. The anode plates are connected and held in spaced relation by a cross bar connector l to which the tabs of the plates are welded and similarly the cathode plates are Joined by connector I. The connectors are suspended from anode conductor rod Il and cathode conductor rod Il. respectively. These rods pass through the cell cover, being insulated therefrom and scaled by concentric insulating material, and external connection is made to them from the power source.

The cell body can be conveniently made of steel. Iron cathode plates and nickel anode plates are suitable, and the cross bars and conductor rodscah be made of the same metals as the electrode plates served by them. The plates can be closely spaced, a spacing distance between l adjacent anode and cathode surfaces of the order of V. to '/4 inch being quite satisfactory. Teflon" (polytetrafiuoroethylene) is a very satisfactory non-corroding material for insulating and sealing the conductor rods from the cover and for providing a gasket between the cover and body flange of the cell.

The bottom of the cell is provided with an outlet pipe llhaving a valve Il, used for draining the cell. This can be used during operation for draining liquid products which settle to the bottom of the cell.

The cell cover is provided with a vertical inlet pipe Il (which connects to the bottom of the hydrogen fluoride condenser shortly to be described) and a .charging inlet pipe ll, having a valve Il. connects thereto to permit of introducing into the cell the liquid hydrogen fluoride, the organic starting compound, and any other substance (such as a conductivity additive) which is to be present. lThis inlet pipe extends down into the cell a sufficient distance so that its lower end will be well below the surface of the liquid solution during normal operation, thereby making a liquid seal. It is located in front of the electrode pack and is spaced therefrom.

The cell cover is also provided -with a vertical outlet pipe Il, for gaseous products evolved `in the cell. which connects to the hydrogen fluoride condenser Il located above the cell. This condenser has a jacket I9 to permit of circulating a coolant liquid for maintaining a desired low temperature. The pipe il extends down into the condenser short of the bottom, and may be in the form of a spiral for better heat transfer. vThe bottom of the condenser vessel is connected, through valve 2l, with the aforesaid vertical inlet pipe I4 of the cell. This permits the liquid condensate to drain back into the cell.

The top of this cmdenser is connected by tube Il to the reactor vessel 2l, the tube extending down into the vessel, which is filled with a 15% .sodium hydroxide solution. The top of the reactorisinturn connectedbytube Iltothedrier vessel 24, the tube extending down into the vessel. I'hisvesselisfilledwithadryingmedium,suchas potassium hydroxide pellets, for removing water from the gaseous mixture passing therethrough.

The top of the drier is connected by tube Il. having valve Il, to the product condenser vessel 2l, the tube extending down short of the bottom.

bolted to an outward top flange of cel! 7.5 The tube, m be pilalled so as to improve heat transfer. This vessel is provided with a removable liquid-air cooling jacket 28, for maintaining in the condenser a very low temperature sufficient for condensing even such low-boiling product compounds as carbon tetrafluoride (GF4) and fiuoroform (CFaH), but insuiflcient to condense the accompanying hydrogen.

The top of this condenser is provided with an outlet tube 29, having a valve 30, for conducting the hydrogen gas to the gas meter 3|, the latter being used for measuring the hydrogen production. The outlet tube is provided with a branch tube 32, having a valve 33, which connects on the upstream side of valve 30. During

normal operation valve

26 and 30 are open and valve 33 is closed. However,

valves

26 and 30 can be closed and valve 33 opened to withdraw the vcondensate compounds, the latter being vaporized by warming in the absence of the liquid air; and during this period the gaseous products from the cell canv be diverted to an alternate recovery system (not shown) if continuous cell operation is in progress. The withdrawn products can be led to a fractional distillation system (not shown) for separating the various product compounds which have been formed.

The gaseous mixture evolved from the cell will normally contain some hydrogen uoride (evaporated in the cell) in addition to hydrogen and to volatile fluorne-containing carbon product compounds. If an oxygen-containing organic starting compound is employed this will also result in the formation of some CO2. The hydrogen fluoride condenser i8 is maintained at a temperature of about minus 40 C. and serves to condense out most of the hydrogen fluoride, which is returned to the cell. The other constituents of the gas mixture are substantially insoluble in the liquid hydrogen fluoride condensate and are carried along in vapor phase in the hydrogen gas stream. Vapors from any higher-boiling iluorine-containing carbon product compounds formed in the cell will also condense and return to the cell; such compounds in the main settling to the bottom of the cell.

'I'he sodium hydroxide reactor 22 serves to remove from the gas stream any residual HF and also any C02 which is present. 'I'he drier 24 serves to remove water vapor which has been picked up in the reactor. The fluorocarbons and other fluorine-containing carbon product compounds are separated from the hydrogen by being frozen out in the product condenser 21, from which they can be withdrawn in the manner already indicated.

Other means can be used for separating the hydrogen from the purified gaseous product compounds. A condenser cooled by a mixture of solid-CO2 and acetone can be included ahead of the liquid-air condenser to preliminar-ily condense out the higher-boiling product compounds. Instead of using a liquid-air condenser, the gaseous mixture can be compressed and refrigerated so aS to condense out the product compounds. This can be done in several stages so as to first condense out the bulk of the higher-boiling compounds and then, at a. lower temperature or higher pressure (or both), condense out the remaining product compounds. The'hydrogen can also be separated by chemical means, as by passing the gas stream through heated copper oxide.

I'he aforesaid purifying and recovery system has a wide application but can be readily modified as desired. For example, an aqueous calcium chloride bubbler can be inserted in series before been found generally suitable for most efllcient the sodium hydroxide bubbler for removal of residual hydrogen fluoride from the gas stream. Oxygen fluoride (OFn), formed when water is present in the electrolyte solution of the cell, can be removed from the gas stream by bubbling through an aqueous potassium-sulfite-iodide solution contained in a reactor vessel included in the train ahead of the drier. When nitrogen-containing organic starting compounds are being processed, the gas stream will contain some nitrogen trifiuoride (NFa). This is a stable low-boiling gas and can be separated from the product compounds by fractionation.

A further alternative -is to remove the residual traces of hydrogen fluoridefrom the gas stream leaving the hydrogen fluoride condenser by passing through a vessel packed with sodium fluoride pellets. It is not necessary to remove CO2, OF2 and NFa from the gas stream, and the use of aqueous purifying solutions can be omitted. The gas stream from the sodium fluoride vessel can go directly to the product condenser in this case as no drying is needed. The CO2 (and any OF: and NF.: present) can be separated from the products by distillation. This procedure is especially useful when reactive fiuorocarbon derivatives (such as iluorocarbon acid fluorides made from carboxylic acids) are contained in the gas stream.

The principal non-volatile products of interest are insoluble'in the hydrogen fluoride electrolyte solution and settle to the bottom of the cell from which they can be drained. The immiscible liquid drained from the cell may include normally solid compounds dissolved therein. This liquid can be washed with potassium hydroxide solution to remove traces of hydrogen fluoride and then fractlonally distilled. Soluble product materials which remain in the electrolyte solution can be recovered by distillation following the run. Some product compounds (of intermediate volatility) will be recovered in part from the gaseous mixture evolved from the cell and in part from the immiscible liquid which is drained from the bottom of the cell.

The body of the electrolysis cell can be used as a cathode, being yconnected to the cathodes in the cell, or, as is generally preferable, it can simply iloat, not being electrically connected or grounded.

The cathode elements do not need tobe made of iron or steel. Any metallic conducting material which resists the corrosive action of the hydrogen fluoride solution can be used, including copper, magnesium, aluminum, nickel, nickel alloys'l etc. Anode elements can be made of nickel alloys as well as of nickel, or of Monel metal, silicon carbide, carbon, etc.

Using iron cathodes and nickel anodes, operating cell voltages in the range of 5 to 8 volts have results, with 5 to 6 volts usually being the optimum. Current densities as high as 20 amperes or higher per square foot of anode surface are obtainable. A 50-ampere cell, having about 350 sq. in. of anode surface, has proved quite satisfactory for general laboratory research use and permits of making substantial quantities of product compounds for study of properties and for preliminary use evaluation. The compactness of the cells per unit of power is an advantageous feature of the process. Operating temperatures in the range of about minus 20 C. to

plus C. have prog/ ed highly effective. vWithin 5 this range, temperatures of about 0 C. to 20 C.

asians l readily permit of convenient operation at about atmospheric pressure (the boiling point of hydrogen fluoride being 19.5 C. and being raised by dissolved compounds), and the necessary cooling can be readily effected.

It is to be understood, however, that the invention is not restricted to the operating conditions noted above. Under appropriate conditions cell voltages as low as about $5 volt can be used. While higher cell voltages than those mentioned can be used under some circumstances without causing free iluorine generation, arrangements utilizing cell voltages not exceeding 8 volts are much preferred since the use of a higher voltage means that much energy is being lost in overcoming current blocking conditions in the cell. A wide range of operating temperatures and pressures can be employed. Thus temperatures as low as minus 80 C. and higher than plus 75 C. have actually been used in experiments. The freezing point of pure hydrogen iluoride is minus 83 C. and a still lower freezing point results when it contains dissolved material. Thus the liquid state oi' liquid hydrogen fluoride solutions can be maintained at very low temperatures. Temperatures higher than about C. generally require operation at superatmospheric vapor pressures in order to elevate the boiling point of the solution. However, the necessary pressure can be minimized and even atmospheric pressure can be utilized by employing solutes which markedly lower the hydrogen fluoride vapor pressure, as by dissolving potassium or sodium iluoride in relatively large proportion in the hydrogen fluoride. Superatmospheric cell pressures are needed in some cases to maintain the starting compound in liquid phase or to prevent rapid evaporation, even though a low operating temperature is employed. This is especially true in the case of low-boiling starting compounds which are not dissolved in the liquid hydrogeniluoride. To take an extreme illustration, ethane has a boiling point of minus 88 C. at atmospheric pressure and hence .the cell must be operated at an elevated pressure in order to maintain ethane in a liquid state in the cell at temperatures above this value.

It will be evident from the foregoing description that this invention pertains to a basically new unit process in chemistry and is not a mere process for producing a particular type of compound from a particular type of starting material. Practically all organic materials can be utilized as starting compounds, whether or not soluble in liquid hydrogen fluoride, and the process provides a new general method by which they can be transformed into a great variety of fluorine-containing carbon product compounds, including many novel compounds not hitherto made by any other process.

In all cases the basic generic principle of the present invention is employed, namely, the useful electrochemical preparation of iluorine-containing carbon product compounds from iluorinatable organic starting compounds, by electrolyzing for a period of hours a current-conducting mixture of liquid hydrogen fluoride solution and the organic starting material (either dissolved or admixed), which is free from water in more than a small proportion, at a temperature and pressure at which a liquid state is maintained and at an electrolyzing potential which is insuillcient to generate free iluorine under the existing conditions, but which is suillcient to cause the production of iluorine-contalning carbon compound products at a useful rate, and recovering one or more iluorine-containing carbon compound products of the process.

In order to further illustrate the invention by specific examples, the following table is presented as a convenient summarization of the results of a large number of experiments. This table is not intended to set forth all compounds that have been made or can be made, and serves for illustration rather than limitation. It lists a considerable number of different iluorine-containing carbon product compounds which have been made by this process from soluble organic starting compounds dissolved in anhydrous liquid hydrogen iluoride. Boiling point values for most of the compounds are given in the left-hand column. These are to be regarded as approximate values. Many of these values were determined at about 740 mm. pressure and have not been corrected to standard condition (780 mm.) values. and hence are somewhat lower (l to 3) than the latter. These compounds were separated and clearly identiiled by physical and chemical properties. In all necessary cases the'content of fluorine and of other non-carbon elements present was determined by analysis. In some cases where compounds of the same series or. of an analogous series had previously been obtained and identified, it was not considered necessary to chemically analyze the product compounds for unambiguous identincation; the physical and chemical properties, together with knowledge of the starting compound employed. being quite suilicient to enable definite identiilcation. In some cases derivatives were made and identified, further corroborating the identication of the cell product. Infrared ab? sorption spectra determinations have been made on various series of these compounds to supply further identincation corroboration. Several compounds were analyzed by the mass spectrograph method.

Under the listing of each product compound -there is listed in the indented column one or more HF-soluble starting compounds from which it was made. No attempt has been made at completeness in the latter respect-thus carbon tetrailuoride (GF4) and iluoroform (CPH) were presumably produced ln substantially all experiments. The product compounds are those which were regarded as of particular significance in the experiments on the listed starting compounds. Some starting compounds have produced significant yields of two or more types of compounds, as for example, the ethers have yielded both iluorocarbons and iluorocarbon oxides, the amines have yielded fully fluorinated nitrogen-containing carbon compounds as well as iluorocarbons, the carboxylic acids have yielded iluorocarbon acid iluorides as well as iluorocarbons, etc. Most of these product compounds have been, and all of them could be, produced with the type of cell arrangement which has been described in connection with the drawing, utilizing iron cathodes and nickel anodes and cell voltages below 8 volts D. C., and operating temperatures in the neighborhood of 0 C. The solubilities of these starting compounds in anhydrous liquid hydrogen iluoride at 0 C. permitted use of starting solutions containing at least about a 1 to l0 ratio by weight of the starting compound to the hydrogen iluoride and in many cases a substantially higher ratio was actually used. 'I'hus in some experiments using acetic acid, the initial amount thereof exceeded that of the hydrogen iluoride.

19 September 10, 1948) and 70,154 (tiled January 10,v 1949). Ser. Nos. 38,751 and 38,752 have since issued as Patents Nos. 2,490,098 and 2,490,099, dated December 6, 1949, and Ser. No. 39,999 as Patent No. 2,500,388, dated March 14, 1950.

The following examples are based on actual experiments and further serve to illustrate the invention. Additional detailed examples have been set forth in my aforesaid parent applications, Ser. Nos. 384,729, 569,265, 626,434, and 677,407, to which reference may be made, but in the interest of brevity have not been set forth in the present specification,

Example I In this experiment use was made of a 100- ampere cell. The casing was a rectangular steel box having a removable steel cover plate bolted to the flanged top of the casing. The conductor rods were insulated and sealed from the cover plate by means of Saran (polyvinylidene chloride) tubing, which was also employed in sheet form as a gasket between the cover and the casing ilange. Eight anode plates of nickel and seven interleaving cathode plates of iron were used, spaced so that the distance between adjacent electrode surfaces was 0.29". The dimensions of each plate were 18" x 111/2" x The distance between the outer anode plate surfaces and the cell walls was 0.40". The cell casing lserved as a cathode element and was connected to the cathode plates. The anode plates had a total operating area of 23 sq. ft. (21,400 sq. cm.) The cell was maintained at a temperature during operation of from minus C. to plus 10 C. by means of a chilled brine bath.

This cell was charged with 4.45 kg. (50.5 mols) of n-butyric acid (approximately pounds) and with 27.5 kg. (1375 mols) of commercial anhydrous liquid hydrogen fluoride (approximately 61 pounds). The applied cell voltage was 5.3 volts, which produced an average current of 100 amperes, the average current density being about 4.4 amps/sq. ft.

The exit gas mixture from the cell was conducted through a tube (connected to the cover plate) to a condenser where the bulk of the HF was condenser and drained back to the cell, then through a sodium iluoride tube for further removal of HF, then through a potassium hydroxide bubbler to remove CO2 and any residual HF, a potassium iodide and sodium thiosulfate bubbler to remove OFz, a liquid air trap to condense the iluorocarbon compounds, and lastly a wet test meter for measuring the quantity of hydrogen.

The materials collected in the liquid air trap during a portion of the run were transferred to a gas holder over water. The weight of this puried condensate was 745 grams, produced during an operating period of 54.4 hours, with 232 faradays being passed. This material was fractionated into nine portions with yields as shown in the following table. Molecular weights of the fractions were obtained by using a gas density balance and applying the ideal gas law; they represent approximate weighted averages in the case of fractions boiling over an appreciable range. The weight amount of each fraction was obtained from the pressure of the gas when the fraction was allowed to vaporize into a receptacle of known volume, applying the ideal gas law. The rst cut was a mixture of GF4, CFsH and CzFe. Other cuts consisted of relatively Pure single compounds, as shown.

Formula Bolling range, Quantity, Mol. Cut u C grams Identity wt. o(

m 80 H0 60 #7..... -16 t0 -4 50 f8.-." -4 t0 -2 30 152 CaFsHIN.. 152 nu.-. -2 t0 +18 30 150 Residue...

It will be evident from the above table that the gaseous mixture contained a large proportion of iluorocarbon compounds having three carbon atoms in the molecule, i. e. CaFa (octailuoropropane), CaFHH (heptauoropropane) and CaFaHa (hexafluoropropane) After several runs using this same charge (without replenishment of the butyric acid) the following materials were collected from the gas mixture: Boiling below -38 C., 1141 grams; at approximately 38 C., 1988 grams; between -38 and 17 C., 182 grams; between -17 and -4 C., 1183 grams; above -4 C., 972 grams. The residual solution in the cell. after an extended period of electrolysis, was found to contain 2.27 kg. (5 pounds) of organic material which contained about 40% by weight of combined uorine. Thus non-gaseous fluorine-containing carbon compounds were also produced by the process and remained in the cell.

Other experiments have demonstrated the formation from butyric acid of a substantial yield of CiFi@ (decafluorobutane), having a boiling point close to 0 C., as well as other compounds, including the tluorocarbon acid fluoride compound CaF'zCOF.

Example II A small laboratory cell was used. The cell body was constructed of a one-foot length of iron pipe (3" diameter) having an iron plate brazed to one end to form the bottom. An iron rod (1/2 dia.) was brazed to the center of the bottom and extended upwardly to near the top. 'I'he pipe and this center rod constituted the cathodes. An iron cover plate was secured to a. ange at the upper end of the pipe. Through the center of the cover passed an anode conductor support rod sealed and insulated therefrom, to which was fastened a cylinder of sheet nickel (1l/2" dia.) surrounding the iron center rod and forming the anode. Saran" was used for sealing and gasketing. An outlet tube was connected to the upper end of the pipe for withdrawing gaseous products. The cell was cooled in an ice bath to maintain an operating temperature in the neighborhood of 0 C.

The cell was charged with 400 grams (20 mois) of commercialanhydrous liquid hydrogen fluoride, 93 grams (1 mol) of toluene, and 5 grams (0.3 mol) of added water, serving as a conductivity additive. The toluene was substantially insolublein the hydrogen fluoride solution but became admixed therewith during operation.

Electroylsis proceeded at 5 volts and 1.1 amperes for a period of 12 days, during which 88 liters of gas were evolved which yielded 11 grams of condensate in a'liquid air trap. The condensate was distilled to yield a 0.5 gram portion boiling below C., a 0.5 gram portion boiling at i5 about 0 C., and a principal fraction of 10 grams dryness.

21 boiling above room temperature. The latter was redistilled in a rractionating column and the following fractions were separated:

Weight (grams) ABoiling range 0.):

57-60 1.6 60-63 1.6 63-64 3.2 Residue 0.7

The 6364 C. fraction had a molecular weight of 383, corresponding to a 7-carbon iluorocarbon. l

Using the same cell arrangement, the formation of iiuorocarbons from n-octane, using water as a conductivity additive, was demonstrated. The cell was charged with 550 grams (27.5 mols) of commercial anhydrous liquid hydrogen nuoride, 160 grams (1.4 mols) of n-octane, and 10 grams (0.6 mol) of added water. An electrolysis proceeded during the 600 hour run period, a total of 440 additional grams of hydrogen fluoride and 60 grams of water were added from time to time to replenish the hydrogen fluoride and to maintain the conductivity. The cell voltage varied from 6.5 to 7.7 volts and the cell current varied in the range of 1.5 to 3.0 amperes. A total of 47.3 faradays passed. Fractionation of the product mixture demonstrated that the original octane had been (to a large extent) converted to fluorocarbon compounds, of which more than one-third was CsFla. Substantial yields of Cf/Fm and CGFM were also obtained.

The production of iluorocarbons from n-octane was also demonstrated in experiments using ammonia, mercurio cyanide, and pyridine,` as conductivity additives. In the absence of the noctane, a solution f mercurio cyanide in anhydrous liquid hydrogen fluoride yielded GF4 and NFs when subjected to a`cell potential of about 7.5 volts, and with the liberation of free mercury.

Example III A cell of the type described in the preceding example was charged with 205 grams (5 mols) of acetonitrile (Cl-BCN) and 785 grams (39.3 mols) of anhydrous liquid hydrogen uoride (which initially contained a substantial trace of Water). Current was passed for ve days at an average of 5.5 amperes. The cell voltage averaged I1.7 volts and remained close to this value during this ve day period. Fluorocarbon compounds Were formed during the entire run. The

identified products included OFz, CFsH, GF4,

CzFsH, CaFe andCFsCN. Y

The CzF5H and CFaCN appeared to form an azeotropic mixture and separation by distillation was not accomplished. This mixture fraction (boiling in the range of -73 to 48 C.) 'was chemically treated to identify the CFsCN in the following way: The gaseous mixture was passed through 50% aqueous potassium hydroxide at 60-70" C., then into a standard HCl acid solution, through a drying tube, and,recondensed in a liquid air trap. Titration of the acid indicated that 3.2 grams of CFaCN had been hydrolyzed, equivalent to 5.1 g. of CFaCOOK. `The basic solution was neutralized with HC1 and evaporated to The residue was then dried and was subjected to an analytical procedure which demonstrated the presence `of 5.1 g. of CFsCOOK The gas from the liquid air trap was -found to have a molecular weight range of 118 to 124 and was mainly CzFsH (formula wt. 120).

22 Eample IV A small copper cell having a nickel anode was charged with acetic acid and KF3HIF in the ratio of 1 to 3.4 by weight. Current was passed through the cell at about 8 volts and 9 amperes until 6 faradays had passed. The operation was conducted at .atmospheric pressure and a temperature averaging 75 C. This operating pressure was possible, despite the elevated temperature, due to the low vapor pressure of the complex of potassium fluoride and hydrogen iiuoride, which remained in a non-boiling liquid state. The products included CF4 andl CFaH.

A similar experiment, using a cell charge of sodium acetate and KF3HF in the ratio of 1 to 6.8 by weight, was conducted under similar conditions and with similar results.

Y Example V Use was made of an iron laboratory cell containing nickel anodes and iron cathodes, operating at atmospheric pressure and a temperatureof 0 C. The initial cell charge consisted of 7,500 grams of highly anhydrous liquid hydrogen iiuoride and 700 grams of dried pyridine. Additional pyridine was added during the run to make a total of about 2,000 grams. The pyridine employed was dried over solid sodium hydroxide and then distilled. Current was passed for a period of four days, at an average cell voltage of approximately 6.0. A total of 447 faradays was passed.

Theliquid cell product mixture was drained from the bottom of the cell, treated with lime to remove residual HF', and was fractionally distilled. There was obtained about 250 grams of a liquid fraction which was identiiied as relatively pure CsFmNF (undecauoropiperidine), a cyclic compound having the following measured properties:

Boiling point (at about 740 mm.) C 48 Refractive index (at 20 C.) 1.281 Density (grams/cc. at 20 C.) 1.744 Surface tension (dynes/cm. at 20 C.) 13.6 Molecular Weight (from vapor density) 285 Per cent F 73.4 Per cent N 4.87

The values for the pure compound as calculated g from the formula are: molecular weight, 283; Per cent F, 73.1; Per cent N, 4.95.

Other products included Z-iiuoropyridine, C5F12 and N'Fa.

The CsFlz (dodecafluoropentane) is formed by cleavage of the ring (eliminating N) and Iiuorine replacement and addition to produce the saturated open chain. Substantially all of the CsFrz formed from pyridine is of the normal, straightchain type. A highly purified sample (obtained from another experiment in which pyridine was used) having a molecular weight value (determined from vapor density) of 288, in agreement with the formula weight of 288, and which was estimated from the freezing point range to be at least 99.5 mol per cent pure, was found to have the following physical properties:

Use was made of an iron-cathode nickel-anode cell of the type shown in' the drawing, operating at atmospheric pressure and a `temperature of about C. The initial cell charge consisted of 1800 grams of anhydrous liquid hydrogen fluoride and 400 grams of di-n-butyl ether. Additional ether was added during the run to replenish that consumed. The cell voltage was in the range of 4 to 6 volts and produced a current density of about amperes per square foot of anode surface.

At the end of 60 hours, 1355 grams of immiscible liquid was recovered from the bottom of the cell, and this was washed with potassium hydroxide solution to remove traces of HF, and was then fractionally distilled to yield a liquid fraction which was identified as relatively pure normal C4F9-O-C4Fs (di-n-nonaiiuorobutyl ether) This fraction had the following measured properties:

Boiling point (at 741 mm.) C 100.7 Refractive index (at C.) 1.261 Density (grams/cc. at 34 C.) 1.689 Dielectric constant (at 20 C.) 1.82 Molecular weight (from vapor density) 459 The formula weight of the pure compound is 454.

This compound has boiling point, refractive index, and density values which are slightly but appreciably lower than those of the normal CsFia uorocarbon (mol. wt. 438) which have been found to be 104 C., 1.267, and 1.765, respectively, and may be compared with the values given above.

This butyl ether starting compound also produces a substantial yield of the C4F1o fluorocarbon, and lesser yields of lower fluorocarbon compounds. Minor yields of uorocarbon ethers -aving fewer than eight carbon atoms are also produced. Some material is also produced which is higher boiling than the C4F9-O-C4F9 product. There is some evidence for the formation CaFis, presumably formed by the combination of C4Fg radicals in the solution.

The iiuorocarbon ethers are so stable that direct analytical determination of the oxygen content is extremely diiiicult. The formation of such compounds from hydrocarbon ether starting compounds has been further substantiated by homologous series runs down to and including the production of CFa-O-CFs and from dimethyl ether and diethyl ether, respec tively. These products were found to have boiling points of minus 59 C. and plus 1 C., respectively, and the measured molecular weights (vapor density method) were the same as the formula weights. Their identification was further coniirmed by mass spectrograph analyses which admit of no doubt since the mass numbers of the ionic fragments were measured with ample precision.

Having described various embodiments and examples of my novel process for purposes of illustration rather than limitation, what I claim is as follows:

1. A new and useful electrochemical process of making fluorine-containing carbon compounds from uorinatable organic starting compounds, which comprises electrolyzing for a period of hours a current-conducting mixture of the organic starting material in liquid hydrogen fluoride free from water in more than a small proportion as herein described, at a temperature and pressure at which a liquid state is maintained and at a cell voltage which is insufllcient to gen- 24 erate i'ree uorine under the existing conditions but which is suflicient to cause the production of fluorine-containing carbon compound products at a useful rate, and recovering at least one fluorine-containing carbon compound product of the process.

2. A new and useful electrochemical process of making carbon compounds which have at least one fluorine atom bonded to a carbon atom in the molecule, which comprises electrolyzing for a period of hours a current-conducting mixture of liquid hydrogen fluoride free from water in more than a small proportion and an organic starting compound having at least one hydrogen atom bonded to a carbon atom in the molecule, at a temperature and pressure at which a liquid state is maintained and at a cell voltage which is insuicient to generate free fiuorine under the existing conditions but which is suilicient to cause the production of fluorine-containing carbon compound products at a useful rate, and recovering a carbon compound product of the process which has at least one iiuorine atom bonded to a carbon atom in the molecule.

3. A new and useful electrochemical process of making carbon compounds which have at least one fluorine atom bonded to a carbon atom in the molecule, which comprises electrolyzing a current-conducting mixture of liquid hydrogen fluoride free from water in more than a small proportion and an organic starting compound having at least one hydrogen atom bonded to a carbon atom in the molecule, by passing direct current through the solution for a period of hours at an electrolyzing potential which is insumcient to generate free iiuorine under the existing conditions but which is sufiicient to cause the production of uorine-containing carbon compound products at a useful rate, the temperature and pressure being such as to maintain a liquid state, and recovering a carbon compound product of the process which has at least one uorine atom bonded to a carbon atom in the molecule.

4. A process according to claim 3 wherein an oxygen-containing starting compound is employed which has at least one hydrogen atom attached to a carbon atom in the molecule.

5. A process according to claim 3 wherein`a carboxylic acid starting compound is employed.

6. A process according to claim 3 wherein a nitrogen-containing starting compound ls employed which has at least one hydrogen atom attached to a carbon atom in the molecule.

7. A process according to claim 3 wherein an amine starting compound is employed.

8. A process according to claim 3 wherein a hydrocarbon starting compound is employed.

9. A process according to claim 3 wherein an alkane starting compound is employed.

10. A new and useful electrochemical process of making uorine-containing carbon compounds from iluorinatable hydrogen-containing organic starting compounds which are readily soluble in anhydrous liquid hydrogen nuoride to form electrolytically conducting solutions, which comprises electrolyzing in a nickel-anode cell a current-conducting solution of the organic starting material in anhydrous liquid hydrogen fluoride by passing direct current through the solution for a period of hours at an electrolyzing potential which is insuicient to generate free iiuorine under the existing conditions but which is suflicient to cause the production of fiuorinecontaining carbon compound products at a useful rate, the temperature and pressure being such 25 as to maintain a liquid state, and recovering at least one uorine-containing carbon compound product of the process.

l1. A process according to claim wherein the temperature is maintained in the range of about minus to plus 80 C., and the total cell potential does not exceed 8 volts.

12. A new and useful electrochemical process of making uorine-containing carbon compounds from fluorinatable organic starting compounds, which comprises electrolyzing for a period of hours a current-conducting solution essentially consisting of liquid hydrogen fluoride containing a dissolved conductivity additive and'also containing an admixed undissolved organic starting compound and which is free from water in more lthan a small Iproportion as herein described, at

a temperature and pressure at which a liquid state is maintained and at an electrolyzing potential which is insuiiicient to generate free fluorine under the existing conditions but which is suiiicient to cause the production of luorinecontaining carbon compound products at a useful rate, and recovering at least one fluorinecontaining carbon compound product of the process. i

13. A process according to claim 12 wherein a soluble organic compound is employed as a conductivity additive. 14. A process according to claim 12 wherein a soluble oxygen-containing organic compound which contains at least one hydrogen atom bonded to a carbon atom in the molecule is employed as a conductivity additive.

15. A process according to claim 12 wherein a soluble monocarboxylic acid is employed as a conductivity additive.

16. A process according to claim v12 wherein a soluble nitrogen-containing organic compound which contains at least one hydrogen atom bonded to a carbon atom in the molecule is employed as a conductivity additive.

17. A process according to claim 12 wherein a soluble inorganic com-pound is employed as a conductivity additive.

18. A process according to claim l2 wherein a soluble fluoride salt is employed as a, conductivity additive.

19. A new and useful electrochemical process for making uorine-containing carbon cornpounds, which comprises electrolyzing in a nickel-anode cell a current-conducting solution essentially consisting of liquid hydrogen fluoride containing a dissolved conductivity additive and also containing an admixed and undissolved iiuorinatable hydrogen-containing organic starting compound and which is free from water in more than a small proportion as herein described, by passing direct current through the solution for a period of hours at an electrolyzing potential which is insuicient to generate free uorine under the existing conditions but which is sufcient to cause the production of ilumine-containing carbon compound products at a useful rate, the temperature and pressure being such as to maintain a liquid state, and recovering at least one uorine-containing carbon compound product of the process.

20. A process according to claim 19 wherein anhydrous liquid hydrogen fluoride, and a hydrocarbon starting compound, are employed.

JOSEPH H. SIMONS- REFERENCES CITED The following references are of record in the le of this patent:

UNITED STATES PATENTS Number Name Date 2,186,917 Gaylor Jan. 9, 1940 2,220,713 Grosse et al. Nov. 5, 1940 OTHER REFERENCES Simons: Chemical Reviews, vol. 8 (1931), Dp. 223-230.

Getman et a1.: Outlines of Theoretical Chemistry, 6th ed., (1937), pp. 367, 368, 473.

Porter: Chemical and Metallurgical Engineering (July 1946), p. 107.

Claims

1. A NEW AND USEFUL ELECTROCHEMICAL PROCESS OF MAKING FLUORINE-CONTAINING CARBON COMPOUNDS FROM FLUORINATABLE ORGANIC STARTING COMPOUNDS, WHICH COMPRISES ELECTROLYZING FOR A PERIOD OF HOURS A CURRENT-CONDUCTING MIXTURE OF THE ORGANIC STARTING MATERIAL IN LIQUID HYDROGEN FLUORIDE FREE FROM WATER IN MORE THAN A SMALL PROPORTION AS HEREIN DESCRIBED, AT A TEMPERATURE AND PRESSURE AT WHICH A LIQUID STATE IS MATAINED AND AT A CELL VOLTAGE WHICH IS INSUFFICIENT TO GENERATE FREE FLUORINE UNDER THE EXISTING CONDITIONS BUT WHICH IS SUFFICIENT TO CAUSE THE PRODUCTION OF FLUORINE-CONTAINING CARBON COMPOUND PRODUCTS AT A USEFUL RATE, AND RECOVERING AT LEAST ONE FLUORINE-CONTAINING CARBON COMPOUND PRODUCT OF THE PROCESS.