CBr4 + BrF3 > CBr3F (large amount) Bromine

tetrahalides, except C F 4, and a number of fluorocarbon chlorides and bromides have been produced (16).

The reactions are summarized in the following equations:

~ 2 5 °

1. CC14 + BrF3 • CF2C12 + CFC13

CCI4

+ BrF3

• CF3CI

(good yield) + CF2C12 Autoclave

0°

2. CBr4 + BrF3 > CBr3F (large amount) Bromine

solution C F 2Br2 (small amount) CBr4 + BrF3 • CF3Br (94%)

Methforyl bromide can also be prepared from CBr4 by reaction with antimony trifluoride and bromine at 180 to 220° (438).

The photochemical bromination or chlorination of fluorocarbon iodides yielded the corresponding bromides and chlorides. When á,ů-diiodides were used, small amounts of bromoiodo and chloroiodo deriva­

tives were found (169). Iodine monobromide and iodine monochloride have been added to C 2F4 to form CF2ICF2Br and CF2ICF2C1 (25).

Bis (methforyl) mercury has been cleaved with chlorine or bromine at 100 to 150° with the formation of CF3C1 and CFsBr (124). These methods are not generally used as preparative methods, but may be valuable in special cases.

The carbon halogen bond distances (C—X) have been determined for the substances C F 3X by microwave studies (389, 392, 393). This is shown in Table VIII with the (C—X) distances in CHsX given for comparison.

TABLE VIII

Carbon-Halogen Distance, dcx ^{in CF}Z^{X with d}cx ^{in CH}ZX Given for Comparison

X Angle FCF* dcx (Angstrom) C F3^X dcx (Angstrom) C H3^X

F 109° 28' 1.34f 1.385

Cl 108° 1.740 1.779

Br 108° 1.908 1.936

I 108° 2.134 2 . 1 3 9

* Values of the angle of 108° were assumed to be the same as the measured value in C F3^{H .}

f T h e C—F distance in C F3H is 1.332 A (136). A redetermination of the C—F distance in CF* by electron diffraction seem desirable; neutron diffraction measurements (4) gave a value of 1.33 A.

As the table indicates, this distance is shorter in the fluorocarbon halides than in the hydrocarbon halides but decreases as the halogen goes from fluorine through chlorine and bromine to iodine.

The ultraviolet absorption spectra of CF3Br, CF2Br2, and CHF2Br and their gaseous refractive indices have been measured by Davidson (104). C F 3Br and CHF2Br have their absorption maxima at about 208 and 207 ταμ, about the same place as methyl bromide, 204 đŔě. The peak for C F 2Br2 is at a longer wave length, 227 đŔě. The fluorocarbons are much more transparent to ultraviolet, the extreme transmission limit of octforane being at about 156 τημ. The r e f r a c t i v e indices of CF3Br, CF2BrH, and CF2Br2, expressed as (^em, - 1) × 10

4

, are 9.78, 9.87, and 15.13, respectively. The values are calculated for standard conditions.

The near ultraviolet absorption spectra of some chlorofluoro com­

pounds have been measured (252). The spectra appear continuous and

366 J. H. SIMONS AND T. J. BRICE

obey Beer's Law. Whenever a hydrogen atom is replaced by a fluorine atom in a halomethane, the absorption shifts to shorter wave lengths.

In ethylenic compounds, however, the introduction of fluorine causes the absorption peaks to shift to longer wave lengths. In the methane series the absorption is believed to correspond to dissociation which becomes more difficult when fluorine is introduced. Lâcher et al. believe that absorption in the ethylene series is associated with electron excitation rather than bond rupture and that the fluorine atoms make the process of excitation easier.

Measurements of the magnetic susceptibilities of halogen derivatives of methane and ethylene have been reported (255, 256). I t was found necessary to use the concept of increments due to interactions between adjacent nonbonded atoms as well as increments due to the atoms alone to obtain a satisfactory correlation of the data.

Few attempts to use the saturated fluorocarbon chlorides or bromides as chemical intermediates have been reported. Apparently there is little tendency for the C—Br or C—CI bond to cleave and form free radicals, the reaction which is of major importance for fluorocarbon iodides.

Replacement reactions by nucleophilic reagents have not been reported and are probably very difficult to do at best. An attempt to prepare alkforylsilicon compounds from CF₃C1 and copper-silicon alloys was unsuccessful (231).

An interesting series of reactions starting with monochlorocyclo-hexforane has been described (430). This compound was reduced to hydrocyclohexforane. Conversion to cyclohexforene resulted from treat­

ment with 3 3 % KOH. Both of these reactions are reported to occur with ease. The cyclohexforene was then oxidized to ( — C F₂C F₂C 0₂H )₂.

The simultaneous reduction and pyrolysis of chlorofluorocarbons is reported to be a method of preparing many olefins and hydrogen contain­

ing compounds (37). For example, CF₂C1₂ and an equimolar quantity of hydrogen in a platinum tube at 685° gave a mixture of CF₂HC1, C F₂= C F₂, and C F₂H₂. Similarly CF₂C1CF₂C1 yielded C F₂= C F₂, C F₂H C F₂H , CF_?HCF₂C1 and C F₂H₂.

CF₃Br and C F₂B r₂ are of considerable interest as fire extinguishers.

They are very efficient, and the low toxicity of their decomposition prod­

ucts makes them particularly useful in confined places (75).

FLUOROCARBON IODIDES

W T

hile fluorocarbon chlorides and bromides have been known for a relatively long time, the preparation of fluorocarbon iodides is a relatively recent event. Enough research has been done to show that they are useful synthetic intermediates; this is not now true of the fluorocarbon chlorides

and bromides.

367 Five methods of preparation have been described. Of these, the reac­

tion of the silver salts of fluorocarbon acids with iodine at 100° is probably the most important (61, 161, 168, 175, 183, 195, 404).

* C 0₂A g + I₂- » *I + C 0₂ + Agi

The yields are nearly quantitative. Both mono- and diiodides can be pre­

pared from the appropriate acids. Sodium, potassium, mercuric, lead, and barium salts may also be used, though under more drastic conditions (161). Silver glutarforate forms, in addition to the diiodide, a compound

0 = C — Ď

/

\

believed to be C F₂C F₂C F₂; it could result from the internal loss of Agi from I ( C F₂)₃C 0₂A g (183).

The reagents must be anhydrous and organic diluents should be avoided, since at one stage a powerful iodinating agent, believed to be ÖĎĎĎÉ, is formed. Reaction of this intermediate with water or hydro­

carbon regenerates the acid ; the over-all reactions are :

6 ^ C 0₂A g + 3 H₂0 + 3 I₂- > 6 * C 0₂H + 5AgI + A g I 0₃

* C 0₂A g + RH + I2 —• * C 0₂H + RI + Agi

C F₃I may be prepared in low yield from CF₃COI by pyrolysis in a stream of iodine (161).

Three other methods that do not require fluorocarbon acids have been found. Carbon tetraiodide and tetraiodoethylene were converted to C F₃I and C₂F₆I by carefully controlled reactions with I F₅ (16). It was later found that I C₂F J , which can be made from C₂F₄ and iodine, formed C₂F₆I when treated with I F₅ (169). The method is rather limited since only a few other iodocarbons, such as hexaiodobenzene, are known.

Fluorocarbon olefins, such as ethforylene, react directly with I F₅ to yield iodides, in this case iodoethforane (405).

C F₃I can itself be used as the starting material for the preparation of longer-chain iodides. When it was reacted with C₂F₄, C₃F₇I , C₆F₇I , and higher fluorocarbon iodides were obtained (155). The reaction was initiated either photochemically or thermally (200°). The reactions are believed to be:

C F₃I CF₃- + I

CF₃- + n ( C₂F₄) -> C F₃( C F₂C F₂)_n. C F₃( C₂F₄)_n. + C F₃I -» C F₃( C₂F₄) J + CF₃

-Since compounds of the type C F₃( C₂F₄)_nC F₃ were not found, the chain-terminating step C F₃( C₂F₄) - + C F₃I —• C F₃( C₂F₄)_nC F₃ + I oc­

curred to a negligible amount.

In a similar manner I C F₂C F₂I reacted with C₂F₄ to form I ( C₂F₄)₂I in

368 J. H. SIMONS A N D T. J. B R I C E

high yield; further reaction with C2F4 gave diiodides having up to nine

— C F 2 C F 2 — units (169).

Fainberg and Miller (125) have prepared C F₂I — C C 1 = C F₂ and C F₂I — C F = C F₂ by the reaction of K I in acetone with C F₂= C X — CF₂C1 ; C F 2 C I — C F = C F C 1 and CF₂C1—CC1=CC1₂ did not react. These facts suggested that the reaction did not involve the allylic carbon atom directly. An initial attack by the I~ on the terminal vinyl C F₂ group, with the elimination of CI" by an allylic shift, would explain the difference in the reactivities of the two sets of compounds. The terminal CFC1=

group would be less susceptible to attack by X~ than the C F₂= group.

The moniodides are dense, colorless substances; C F₃I boils at —22°

and the boiling point increases about 30° for each additional — C F₂— group. Upon exposure to light they acquire a pink color from iodine liberated by photochemical decomposition. I C F₂C F₂I readily splits out iodine, reverting to ^C2F4 and I₂, but the longer chain á,ů-diiodides are about as stable as the monoiodides.

The absorption spectra of methforyl iodide and ethforyl iodide were found to be continuous in the near ultraviolet, probably due to decom­

position to fluorocarbon and iodine radicals (17). The quantum yield of decomposition products was low but could be increased by the addition of oxygen, chlorine, cyanogen, or ethylene. The increase was attributed to the removal of one of the radicals by reaction with the added substance.

Most of the known reactions of the fluorocarbon iodides are free radical reactions in which the carbon-iodine bond is first cleaved to form fluorocarbon and iodine radicals; these radicals then initiate other reac­

tions. The reactions of C F₃I with C 2 F4 already considered are examples of this.

Fluorocarbon iodides have been added to ethylene in a telomerization process similar to that with C₂F₄; the products were CF₃(CH₂CH₂)nI.

Compounds having η = 1 to 3 were isolated, with ç = 1 predominating (155). The classes of compounds C F₃( C H₂C H₂)_nC F₃a n d CF₃(CH₂CH₂)nH were not found. However, when the initial reaction products, C F₃ -( C₂H₄)_nI , were heated above 240° or irradiated for a long time in the presence of excess C₂H₄, C F₃C₂H₆ and C F₃C₄H₉ were obtained.

The mono- and diiodides were chlorinated or brominated thermally or photochemically; the corresponding mono- or dichlorides or bromides were produced (169). The diiodides also yielded small amounts of chloro-iodides and bromochloro-iodides. The chloro-iodides reacted with fluorine to form the expected fluorocarbons.

The addition of C F₃I to acetylene forms C F₃C H = C H I , an olefin mentioned previously.

When methforyl iodide was heated at 220° with yellow phosphorus,

369 ( C F₃)₃P , ( C F₃)₂P I , and C F₃P I₂ were produced (32). ( C F₃)₃P burns in air and is quantitatively hydrolyzed to fluoroform by dilute alkali. Arsenic also reacted with C F₃I , and (CF₃)₃As, (CF₃)₂AsI, and C F₃A s I₂ were isolated. They all hydrolyze readily, liberating fluoroform.

The reaction of C F₃I with sulfur at 205° produced C₂F_eS₂; it was assumed to have the structure CF₃—S—S—CF₃. CSF₂, thiocarbonyl fluoride, was also found and in increasing amounts as the temperature was raised. At 220° it was the predominant product.

The known versatility of Grignard reagents in organic reactions has led to numerous attempts to prepare them from fluorocarbon iodides in the vain hope that an analogous reaction in fluorochemicals would be equally useful. The chief difficulty is that the iodides do not react with magnesium at ordinary temperatures in the absence of an organic solvent, so far as is known; in the presence of ethyl ether, reaction occurs readily but the Grignard reagent (presumably) formed extracts hydrogen from the solvent to form C F₃H (in the case of CF₃I) or is otherwise destroyed.

Henne and Francis (198) have avoided the difficulty by carrying out the preparation of the Grignard at —80° in the presence of carbon dioxide, so that carbonation occurred immediately. In this manner, they syn­

thesized C₃F₇C 0₂H from C₃F₇I in 4 5 % yield. Haszeldine (165) has made an extensive study of the formation of these Grignard reagents.

In an earlier paper Emeleus and Haszeldine (123) described in detail unsuccessful attempts to prepare fluorocarbon derivatives of magnesium, zinc, cadmium, lead, arsenic, lithium, and gallium. They did succeed in preparing mercurials. The mercurials first prepared were C₂F₅H g I and CF₃HgI. Two methods of preparation were described. C F₃I and excess mercury were heated in the absence of a solvent to 260 to 290° ; the yield of CF₃HgI was 22%. Prolonged irradiation at 150° gave an 80% yield of CF»HgI.

CF₃HgI is a white crystalline solid readily soluble in organic solvents and also quite soluble in water. The mercurial decomposed slowly in water, depositing mercury (I) and (II) iodides. When it was treated with aqueous potassium iodide fluoroform was liberated:

C F₃H g I + 3 1 - + H₂0 -» C F₃H + O H " + ( H g l₄)

-Because of the slowness of the reaction with water, aqueous solutions of C F₃H g I could be used to prepare other mercurials. The hydroxide CF₃HgOH was prepared by treating a water solution of CF₃HgI with silver oxide. The free base formed salts with acids; CF₃HgCl, CF₃HgBr, and C F₃H g N 0₃ were prepared in 80 to 90% yields from CF₃HgOH in this manner. All of these derivatives are water soluble, crystalline solids, and their aqueous solutions have measurable conductivities.

370 J. H. SIMONS A N D T. J. BRICE

CF₃HgOH is a weaker electrolyte than CH₃HgOH, but a stronger one than C₆H₆HgOH. The conductivity of C F₃H g N 0₃ approaches that of a strong electrolyte; this high conductivity was attributed to hydrolysis of the nitrate which is strongly acidic in water solution.

Bis (methforyl) mercury was also synthesized by Emeleus and Haszel­

dine (124). The best method of preparation was to shake a mixture of C F₃I and cadmium amalgam at room temperature; 40% yields of C F₃H g C F₃ were obtained. Amalgams of copper and silver were also used successfully. C F₃H g C F₃ is a white crystalline solid with a pungent, choking odor. When it was heated with a halogen at 100 to 150°, C F₃X was obtained in 90% yields. Mercury (II) halides at 170° convert C F₃ -H g C F₃ to CF₃HgX. When the dimercurial was heated above 170° in glass, it decomposed to silicon tetrafluoride, carbon dioxide, and mercury.

CF₃HgI was found to be stable up to about 300°, at which point it, too, decomposed.

C F₃H g C F₃ is readily soluble in water; the solution has a low but definite conductivity. The solution is neutral and quite stable. A crys-tallographic examination (124) indicated that the unit cell is cubic (a = 8.11 A; four molecules of C F₃H g C F₃ per unit cell) with mercury atoms located in the 000, 1/2 1/2 0, 1/2 0 1/2, and 0 1 / 2 1 / 2 positions.

The molecule is linear.

The reaction of C F₃I with solutions of KOH in alcohol, acetone, or ether yielded C F₃H ; C F₃I did not react with dilute aqueous alkali at room temperature (18). Attempts to replace the iodine atom with OH, N H₂, CN, and N 0₂ by treatment with nucleophilic reagents have been unsuc­

cessful. C F₃I did undergo an exchange reaction with sodium iodide in ethyl alcohol solution; the reaction was followed kinetically using radio­

active iodide as a tracer (18). The reaction was found to be first order with respect to C F₃I , zero order with respect to iodide. The t a c t i o n , and the reaction with alcoholic KOH as well, have been postulated to proceed through the ions CF₃~ and I+, with the formation of these ions being the rate-determining step in the exchange reaction.

The unsaturated fluorocarbon iodide C F₂= C F C F₂I reacts with bromine or lithium bromide in the equivalent of nucleophilic replacement of I by Br:

Acetone

C F₂= C F C F₂I + LiBr • C F₂= C F C F₂B r

Room temperature

C F₂= C F C F₂I + Br₂ > C F₂= C F C F₂B r

in the dark

These reactions are believed to proceed by an initial nucleophilic attack on the C F₂= group followed by a shift of the double bond and elimina­

tion of I~, rather than by direct attack on the C F₂I group (126).

371 C F₂= C F C F₂I can be coupled by zinc in dioxane to form C F₂= C F -C F₂C F₂C F = C F₂ and reduced by zinc and methanol to C F₂= C F C F₂H .

FLUOROCARBON ALDEHYDES AND KETONES

At the time Volume I was written there was no known fluorocarbon aldehydes and very few aldehydes containing fluorocarbon groups. Since then both types of compounds have been prepared and by several differ­

ent methods.

The first preparation of fluorocarbon aldehydes was reported by Husted and Ahlbrecht (230). They found that acetforic and butyrforic acids could be reduced to a mixture of the 1,1-dihydroalcohol and the aldehydrol with lithium aluminum hydride :

LiAlH4

C F₃C 0₂H > C F₃C H₂O H + C F₃C H ( O H )₂

The aldehyde could subsequently be recovered from the aldehydrol by dehydration with phosphorus pentoxide or sulfuric acid. The aldehyde was also isolated from the reduction products of butyrf or amide.

Henne, Pelley, and Aim (206) attempted to prepare CF₃CHO by a variety of methods. The reduction of CF₃CN by L1AIH4 at low tempera­

tures was successful. They were unable to prepare it by any of the follow­

ing methods: (a) the pyrolysis of C F₃C 0₂H and formic acid over man-ganous oxide, (6) the reduction of C F₃C O N H₂ with hydrogen over platinum or copper chromite, (c) the reduction of CF₃CN with stannous chloride, (d) CF₃COCl and L1AIH4, (â) CF₃COCl and H₂ (Rosenmund reduction), (/) CF₃COCl and sodium hydride or borohydride, (g) the catalytic reduction of CF₃COSR over Raney nickel, (h) the hydrolysis of CF₃CHC1₂, (t) the oxidation of C F₃C H₂O H .

Shechter and Conrad oxidized C F₃C H₂C H₃ in the vapor phase at 437 to 462° with nitric acid and obtained a 20% yield of CF₃CHO (390).

Acetforaldehyde is a low-boiling ( — 19 to —18°) compound having many of the chemical properties of organic aldehydes. Both the aldehyde and its hydrate form the 2,4-dinitrophenylhydrazone, m.p. 150-151°C.

CF₃CHO dissolves slowly in water to form the aldehydrol, is oxidized by Tollens reagent or potassium permanganate to acetforic acid, and is cleaved by alkali to fluoroform and formic acid (haloform reaction).

Acetforaldehyde has the property of polymerizing to a clear glass-like solid. The resin is slightly soluble in acetone and ether, but insoluble in water, chloroform, carbon disulfide, and carbon tetrachloride. When heated the polymer is converted back to the monomer; the polymer can be decomposed by aqueous alkali and by acid, and again the monomer

is obtained.

372 J. H . SIMONS A N D T. J . BRICE

Butforaldehyde has chemical properties similar to those of CF3CHO (230, 314). I t also forms a stable solid monohydrate and a diacetate.

Methyl magnesium iodide reacts with C3F7CHO to form the secondary alcohol C3F7CH(CH3) OH which can be dehydrated to the olefin C3H7CH

= C H 2 in about 80% yield via the acetate.

The reducing action of Grignard reagents on fluorocarbon carbonyl compounds has been studied (292). The carbonyl compounds were

ÖĎÇĎ , ^ C O C H 3, and ^ C 0 2C H 3. All Grignard reagents having 0-hydro-gen atoms gave reduction. Methy and phenyl Grignard rea0-hydro-gents did not. The amount of reduction is frequently large and increases with increasing size of the fluorocarbon group.

When the fluorocarbon group is not adjacent to the carbonyl group, organic chemical synthetic methods can be used to prepare the aldehyde.

C F 3CH2CHO can be made by the dichromate oxidation of CF3CH2-CH2OH, for example (206), or the hydration of C F 3C = C H (173, 203).

CF3CH2CH2C1 has been used as a starting material in the preparation of a number of compounds (288) ; the Grignard reagent prepared from it was reacted with ethyl orthoformate and the product hydrolyzed to form C F 3CH2CH2CHO.

The catalytic fluorination of methyl ethyl ketone and cyclopentanone has been used to prepare the corresponding fluorocarbon ketones as well as ketones containing one and two hydrogen atoms (222).

The permanganate oxidation of ( C F 3)2C=CC12 yielded hexafluoro-acetone in 60% yield (208). The oxidation was carried out in neutral or acid media to avoid alkaline cleavage of the ketone. The oxidation of isobutforene also yielded this ketone (60). Since the reaction can be carried out in good yield, the oxidation of branched chain fluorocarbon olefins appears to be a promising method of preparing fluorocarbon ketones. Little is known about the chemistry of these ketones. They form stable hydrates and semicarbazides and they are cleaved by alkali; that is all the information that has been reported to the present.

Mixed fluorocarbon-organic ketones have been prepared by well-known procedures and some general properties determined, as has been described in Volume I, p. 482; see also (65). Simons, Black, and Clark (402) have prepared a series of fluorocarbon aromatic ketones following the synthesis used for the first member of this series (414). The acid chlorides were condensed with the aromatic compounds using either aluminum chloride or aluminum bromide as the condensing agent. The following ketones CF3COCeH5, C2F5COC6H5, C4FeCOCeH5, C 5F n-COCeH5, p-CF3COCeH4CH3, p-C2F6COCeH5CH3, P-C3F7COC6H4CH3, p-C4F9COCeH4CH3, and P-C5F11COC6H4CH3 gave a reported 2,4-dinitrophenylhydrazone derivative. C3F7COC6H5 gave a reported

semi-373 carbazone derivative, although the authors state that the ketones yielded both derivatives. No derivative was prepared from 2,4-(CH₃)₂CeH₃ -COCsFn.

Nes and Burger (324) have prepared methforyl benzyl ketone by condensing ethylacetforate with phenylacetonitrile to form a-phenyl-acetforylacetonitrile and then hydrolyzing and decarboxylating this compound. This ketone, CF₃COCH₂CeH₅, was reported to have been prepared by the hydrolysis of the addition product of C F₃C N and benzylmagnesium chloride as well as by the reaction of CF₃COCi and C₆H₅CH₂ZnCl (233). Nes and Burger showed that the product of these reactions was actually o-acetforyl toluene, formed by a rearrangement of the benzyl group during reaction.

Methforyl benzyl ketone and its ketoxime were catalytically reduced and several products obtained. Reduction of the ketone with hydrogen in the presence of ammonia gave C F₃C H N H₂C H₂C₆H O (72%); in ethanolic ammonia, the aromatic ring was also reduced to form 1,1,1-trifluoro-2-amino-3-cyclohexylpropane. Reduction of the ketoxime in ethanolic hydrogen chloride resulted in a mixture of C F₃C H N H₂C H₂

Methforyl benzyl ketone and its ketoxime were catalytically reduced and several products obtained. Reduction of the ketone with hydrogen in the presence of ammonia gave C F₃C H N H₂C H₂C₆H O (72%); in ethanolic ammonia, the aromatic ring was also reduced to form 1,1,1-trifluoro-2-amino-3-cyclohexylpropane. Reduction of the ketoxime in ethanolic hydrogen chloride resulted in a mixture of C F₃C H N H₂C H₂

In document Fluorocarbon Chemistry BY J. H. SIMONS Fluorine Research Center, The University of Florida, Gainesville, Florida (Pldal 32-101)