or the simplification of complex ESR spectra possible. They observe that bis-nitrophenylmercury species re duced to give stable bis-radical anions which showed well-defined doublet spectra with moderately narrow lines. Mono-nitrophenylarylmercury com pounds gave radical anions with simi lar spectra. In addition to the ex pected ESR signal from species with 200 Hg present, satellite spectra of an intensity about 15% of the main sig nal were observed. These satellite spectra have a shape identical with the main spectra and are attributable, Dr. Dessy says, to those species con taining 1 9 9 Hg. The hyperfine splitting constants for Hg, in each successive position of the ring with respect to —N0 2 are shown below, along with the hydro gen coupling constant for these same positions in the nitrobenzene radical anion N02 H 3 3g
S^-^SsS
\T
\f
^\~x^ J^
Η
δ 34S
H 103
g
Hg llg
3.8g H 42g Hg
It appears, Dr. Dessy adds, that a McConnell-type equation can be written for H g hyperfine interactions of the form: a = p250g. (forH, a = P 25g.) Up to the present, measurement of low spin densities or simplification of complex spectra involved exhaustive computer techniques, deuterium sub stitution, or ENDOR techniques. These procedures all have serious limitations, however.
Unexpected perfluoro reduction mapped • FLUORINE An unexpected reduction reaction of completely fluorinated (perfluoro) aromatic com pounds with organomagnesium (Grignard) reagents has produced a salient in fluorine chemistry which could re sult in a new synthetic route for use ful fluorine compounds. Dr. W. L. Respess mapped out the chemistry of the reaction and reported a possible radical-anion mechanism at Atlantic City. The work was done by Dr. Respess and Dr. Christ Tamborski at the Air Force Materials Lab oratory, Wright-Patterson Air Force Base, Ohio. Basically, the new reaction starts with a perfluorinated aromatic com 50 C&EN SEPT. 23, 1968
pound such as hexafluorobenzene and an organomagnesium reagent such as diethylmagnesium in tetrahydrofuran solvent. As in reactions with other nucleophilic reagents, including am monia, alkylamines, and alkyllithium, the hexafluorobenzene takes on an ethyl group to produce an alkylated substitution product, ethylpentafluorobenzene. But the hexafluorobenzene also gives a second reaction product with the diethylmagnesium—a two-ring compound, 4,4 / -diethyloctafluorobiphenyl. This, Dr. Respess states, is the first example of a perfluoroaromatic starting compound undergoing two different types of reactions with one nucleophilic reagent. To plumb the reaction further, the chemists replaced the single-ring hexa fluorobenzene with the more reactive double-ring substrate, decafluorobiphenyl. The reaction of decafluorobiphenyl .and ethylmagnesium bromide in tetrahydrofuran yields a mixture of products. In the initial reaction stages two alkylated products predominate. However, once these have formed, intermediate products appear leading to the hydrogen-substituted 4-hydrononafluorobiphenyl, 4,4'-dihydrooctafluorobiphenyl, and 4-ethyl-4'-hydrooctafluorobiphenyl. Zeroing in on specific parts of this reaction, -the chemists reacted the monoalkylated product, with a sixfold excess of ethylmagnesium bromide. The reaction is complete within four hours and produces a striking pre dominance (86%) of the reduction product, 4-ethyl-4 / -hydrooctafluorobiphenyl. Similar reduction of 4-hydrononafluorobiphenyl by ethylmagnesium bromide also takes place. The major product (45') is the reduced 4,4'dihydrotetrafluorobiphenyl. A smaller fraction (15') of 4-ethyl-4 / -hydrooctafluorobiphenyl also results. The question is now how the unex pected reduction occurs. Dr. Respess and Dr. Tamborski ruled out an alkylsubstituted ring as the precursor to the original biphenyl they discovered by showing no reaction between ethylpentafluorobenzene and ethylmagne sium bromide. They then showed that the reduc tion of 4-ethylnonafluorobiphenyl with ethylmagnesium bromide went through an organometallic (MgBr-substituted) intermediate. Addition of trimethylchlorosilane to the reaction mixture produces a derivative, 4-ethyl-4 , -trimethylsilyloctafluorobiphenyl, formed as expected from an organometallic intermediate. The chemists examined several pos sible mechanisms for formation of the intermediate. They discounted catal ysis by impurities in the magnesium
Proposed radical-anion path
after finding that ordinary Grignard and triply sublimed (purified) mag nesium gave similar products. A second possibility, simple metal halogen exchange of the starting ma terials, also appeared unlikely, since the consequent ethyl fluoride was not detected. Thirdly, reduction through a hydrogen-substituted perfluoroaro matic intermediate resulting from ac tion by a beta-hydrogen in the alkylmagnesium compounds was eliminated because alkylmagnesium starting com pounds without a beta-hydrogen also gave the reduction products. Dr. Respess and Dr. Tamborski are currently investigating a radical anion mechanism through electron spin res onance experiments. This mechanism seems plausible from the high colora tion in all the reactions with ethyl magnesium bromide. The anions could result from electron transfer from the Grignard reagent to the per fluoroaromatic compound. With such a reaction path through organometallic species, the novel re duction could open new synthesis routes to perfluoroarylmagnesium in termediates from the more readily available perfluoroaromatic com pounds. This would be especially true if concurrent alkylation of the fluoroaromatic material were mini mized—a development Dr. Respess and Dr. Tamborski term "very con ceivable." In some of their experiments »to date on the problem of minimizing alkyla tion, the chemists have varied the or ganomagnesium starting material. The new Grignard reagents-methylmagnesium chloride, isopropylmagnesium bromide, allylmagnesium bromide, and phenylmagnesium chloride-represent different degrees of nucleophilic char acter and reductive capabilities, the two chemists add.
