1380 A. G. STRENG Vol. 85

(CH3)3Sn:-, and, by the trends in coupling constants tion of H for CH3in the coupling Sn-CH3 (starting with observed in the neutral stannanes, substitution of a (CH3)4Snand extrapolating to the methylstannanes) . hydrogen for a methyl group in the trimethyltin anion There have been other observations of non-additivity would be expected to increase the SnCH3 coupling, as of substituent effects, such as the increasing positive observed. Due to rapid solvolytic proton exchange, deviations obtained for C-I€ or Si-H coupling, by the tin-bonded hydrogen atom would not be expected substitution of (particularly F) in CH4 or to split the tin-methyl resonance into a doublet, nor SiH4,20or for the coupling Sn-CH3 upon substitution of would it be possible to observe a resonance for this C1 for CH3 in (CH,)&I.~ A discussion of possible proton as distinct from the solvent protons in such explanations for these deviations is given by Juan and dilute solutions. Gutowsky.*Oh Substituent Effects on Sn-H and Sn-CH3 Coupling.- The effect of substituting CH3 for H on the coupling Acknowledgment.-The authors wish to thank Sn-H in SnH4 and the methylstannanes is not strictly Professor Daniel E. Kivelson of this Department and additive, as illustrated in Fig. 1. The coupling for Dr. J. R. Holmes of the Allied Chem. Corp., hforris- (CHd3SnH strongly deviates from the extrapolated town, pi. J., for valuable discussions; Rlr. Richard straight line, and is larger than might have been ex- Gillespie for assistance in obtaining magnetic resonance pected if the effects were additive. The first three spectra on the 40 Mc. instrument; and the Ethyl points are nearly linear, except for a slight alternation Corporation for a sample of (CH3)4Pb. in positions which, though small, we nevertheless feel (20) (a) K Muller and P I Rose, J Aril Chcm Soc , 84, 3973 (1962), is real. B similar pattern of behavior, including a (h) (J' discussion and references in C Juan and H S Gutowsky, J Chem slight alternation, is observed for the effects of substitu- Phys , 37, 2198 (1962)

[CONTRIBUTED BY THE RESEARCH INSTITUTE OF TEMPLEUNIVERSITY, PHILADELPHIA 44, PENNA.] The Chemical Properties of ' BY A. G. STKENG RECEIVEDDECEMBER 5, 1962

Dioxygen difluoride has a remarkably high oxidizing power, even at very low temperatures. The reactivity of OzFt with Clz, Br2, I*, P, S, and their lower , as well as with the fluorides of nitrogen, with HCl, HBr, H?S and with some other compounds was studied. Formation of intermediate addition products was observed in the reactions with ClF, BrF3, SF4, HCl and HBr. Most intermediates have only a transitory existence.

Introduction Reagents Used Dioxygen difluoride, 02Fz,is the second member of Dioxygen difluoride was prepared directly from the the - family, consisting of OFs,2a OZF,,2h elements by the method described el~ewhere.~It was 03Fs3and O~FZ.~It is an orange-yellow solid which stored in a Pyrex glass cylinder, frozen at 90°K. melts at 109.7'K. to an orange-red liquid. Dioxygen monofluoride, ClF, nitryl fluoride, N02F, difluoride is sufficiently stable at temperatures below its trifluoride, PF3, and silicon tetrafluoride, but decomposes rapidly into O2 and F2 SiF4,were also prepared by the well known methods.'O-13 at temperatures close to its normal boiling point, 216'K. All other reagents were the commercial products of Dioxygen difluoride has been known since 1933, when the highest available purity, supplied by Stauffer Ruff and Menzel achieved its preparation.2h Since Chemical Co. General Chemical Division of Allied then a number of investigations have been devoted to Chemical Co., Pennsalt Chemicals Corp., E. I. du the methods of preparation of 02F2and to its thermo- Pont de Nemours and Co., Matheson Co., Air Reduc- dynamic and physical properties6-8; but very little tion Co. and others. Most of these reagents were fur- has been published about its chemical properties. Only ther purified by fractional distillation. recently, when this paper was ready for typing, there appeared a very interesting article, by R. T. Holzmann Experimental Technique and M. S. Cohen Ilnorg. Chem., 1, 972 (1962)], The experiments were carried out either in a closed vacuum describing the reaction of 02F2with tetrafluoroethylene. system made of Pyrex glass or in one made of Kel-F, fitted with Since dioxygen difluoride is stable only at low tem- stainless- valves and T-pieces. Mercury manometers with a protective layer of Kel-F oil were used for measuring the peratures, its chemistry was studied in a temperature pressure. region which is substantially below the usual range of The experimental conditions were varied according to the chemical studies. requirements of each combination of reagents. In most cases, 02F~was first frozen on the walls or on the Raschig-ring packing An exploratory study of the behavior and reactivity of the reaction vessel. The second reagent was then added in of dioxygen difluoride with various substances was small portions by vacuum distillation and condensed above the made in order to obtain information on the basic chem- 02F2. The vessel was then slowly warmed to the temperature of istry of 02F2 and its ability to form addition products. reaction. The solid and liquid products remained in the re- action vessel, and the gaseous products were removed for analy- (1) This paper describes a part of the wurk sponsored by the Office [IC sis. The reaction vessel was then cooled again to the original Naval Research, under Contract Sonr-3085(01). ____ ~ (2) (a) P. Lebeau and A. Ilamiens, Compl. rend., 185, 652 (1927); (h) (9) A. G Streng and A. V. Grosse, "Addition and Substitution Com- 0. Ruff and W Menzel, Z,anoyg. zi. allgem. Chem.. 211, 204 (1933). pounds of Oxygen Fluorides," Second Annual Progress Report for the Office (3) S. Aoyama and S. Sakuraha, J. Chem. Soc. Japan, 69, 1321 (1938). of Naval Research, Contract Nonr-3085(01), Research Institute of Temple (4) A. V. Grosse, A. G. Streng and A. 11. Kirshenhaum, J. Am. Chem. University, Philadelphia 43, Pa., January 19, 1962. Soc., 83, 1004 (1961). (10) 0. Ruff, E. Ascher and F. Lads, Z. anorg. alleem. Chem., 176, 256 (.5) 0. Ruffand W. Menzel, 2. anorg. 11. allgem. Chem., 217, 85 (1931). (1928) (0) H. J. Schumacher and P. Frisch, %. physik. Chem., B37, 1 (1937). (11) Georg Brauer, "Handbuch der Praparativen Anorganischen Chemie ," (7) A. 1). Kirshenhaum and A. V. Grosse, J. Am. Chem. Soc., 81, 1277 Ferdinand Enke Verlag, Stuttgart, 1954. (19.59). (12) L. I.ehouche, W. Fischer and W. Biltz, Z. am%'e. allgem. Chrm., 207, (8) A. I). Kirshenbaum and A. G. Streng, J. Cheiii. Phys., 35, 1410 64 (1932). (1961) (13) 8. Ruff and E. Ascher, ibid., 196, 413 (1931). May 20, 1963 CHEMICALPROPERTIES OF DIOXYGENDIFLUORIDE 1381 temperature and a new portion of reagent was added. The melting point of 02F2 and the excess of OzFZ pumped procedure was repeated until all the OZFZwas consumed. The reaction products (gaseous, liquid and solid) were measured and off. , and water analyzed. were the reaction products. The amounts of the prod- If O~FZwas used as liquid, it was condensed in the bottom of the ucts indicate that the reaction proceeded in accordance reaction vessel. In some cases reactions were performed with with the summary equation 0sFz dissolved in a suitable solvent. Two-limb apparatus was used for some of the reactions between two liquid reagents or 0zFz + 5H. +2HF 4- 0.5HzOz + H20 (1) their solutions. The reagents were condensed separately in the Traces of Hi04 were also formed. No or other two limbs and one liquid was added to the other by tilting the apparatus. intermediate products have been found. If a reaction between liquid OZFZ(or its solution) and a gas was studied, the gas, usually diluted with an inert gas, was Reaction with Cl,, C1F and HC1 bubbled through the liquid. The reagents and the reaction products were measured care- A rapid addition of chlorine to dioxygen difluoride fully. In the gaseous and in the liquid phase, they were meas- cooled to about 140'K. caused a violent explosion. ured volumetrically; in the solid phase, by weight. The However, when small portions of CIz were added slowly reaction products were identified by chemical methods, by de- to OzF2cooled to 130'K., a violet intermediate product termination of their physical constants and by infrared spectrog- raphy. did form, together with ClF3.16 Here, the first step Liquid nitrogen, , various Freons and Dry was most probably the fluorination of C12 to CIF were used as refrigerants. OzFz Cls ----f 02 2C1F It must be emphasized that inadequate cooling or a fast addi- + + (2) tion of reagents to O~FZand oice vena caused explosions. After ClF was formed, it reacted further with 02F2, as The Reactivity of with Organic Compounds described below. 02F2 The study of the reaction of OsF2 with C1F showed Being a high energy oxidizer, dioxygen difluoride that if the reaction is carried out without special pre- reacted vigorously with organic compounds, even at cautions at temperatures above 140'K., the two sub- temperatures close to its melting point. It reacted in- stances react violently with heat evolutions following stantaneously with solid ethyl , producing a blue the equation flame and an explosion. When a drop of liquid 02F2 OzF2 CIF +02 ClFa 30.1 kcal. was added to liquid , cooled at 90°K., a white + + + (3) flame was produced instantaneously, which turned The ClF abstracts the from 02F2, forming ClF3 green upon further burning. When 0.2 ~m.~of liquid and liberating 02. Simultaneously, due to the heat of 02Fz was added to 0.5 ~m.~of liquid CH4 at 90°K., a reaction 3, a part of the OZF,decomposes to 02 and Fz. violent explosion occurred. l4 02Fz +O2 + FP + 4.73 kcal. (4) When added to Dry Ice, dioxygen difluoride did not In the reaction products, fluorine was determined by react and was only absorbed by the solid. Addition of the Moissan method by absorbing it with mercury in a acetone to this mixture resulted in sparking accompanied gas buret. Oxygen was determined by absorption in an by an explosion. alkaline solution of pyrogallol. The volume of chlorine A 2% solution of O2F2 in HF reacted violently with a trifluoride was measured as liquid and also as gas after flash with benzene at 195°K.15 vaporization and the compound identified by the melt- Reactivity with , Water and Hydrogen ing and boiling points and by infrared spectrum. When, however, the reaction between 02F2 and C1F Liquid dioxygen difluoride reacted vigorously when was carried out at moderate temperatures (119-130'K.) added to solid anhydrous ammonia at temperatures and with a slow addition of ClF, a third reaction took close to 110'K. It caused explosions when added to place, forming an intermediate compound of the ele- ice at 130-140'K. and reacted also with traces of water mentary composition (02CIF3). in accordance with the when dissolved in HF containing HzO, at 195'K.: scheme the brown color of the solution disappeared and 02 gas escaped. n02Fz + nClF +(02C1Fs),, (5) In view of the high reactivity of hydrogen atoms at The extent of each of the reactions 3, 4 and 5 depends low temperatures, it was considered of interest to study upon the reaction conditions. their reaction with 02F~. It was hoped that since The intermediate compound, dioxygen chlorine hydrogen atoms are likely to abstract fluorine, forming trifluoride (referred to simply as O2ClF3), has an in- HF, intermediate species, either radicals or others, tense violet color and is a very strong oxidizer. Its might form under suitable conditions. The H-atom properties were described elsewhere. generator used was described elsewhere.15 Since this intermediate product is energy rich, it Dioxygen difluoride was condensed in the form of a decomposes rapidly if the reaction proceeds too vio- ring on the walls of a U-tube cooled to 77'K. There lently, or even under mild condition, if impurities are was no reaction between 02Fz and molecular hydrogen present. The product is solid at 195'K. and in the when the gas was pumped through to the U-tube at absence of impurities, excepting CIF3, was kept at this 77'K. at the rate of 1.25 l./hr. and with P = 1 mm. temperature for more than a year without noticeable Atomic hydrogen, however, reacted with OzF2 at the decomposition. same conditions, forming a white solid. Several times At first, it was thought that the intense violet color during the experiment, the H-atom generator was might be due to the formation of mixed with some turned off and the 02F2warmed to its melting point to unreacted 02F2 allow it to separate from the layer of reaction products, 302Fs -I- 3C1F --f 20, f 3ClFx which covered the OsFz and prevented further reaction. (6) After about two-thirds of the 02Fz was consumed, the Although ozone is deep blue and the reaction product generator was turned off, the U-tube warmed to the violet-blue, ozone and liquid OzF2form a violet-blue solution. (11) A. D. Kirshenbaum, J. G. Aston and A. V. Grosse, Final Report, It was found, however, that O3 does not dissolve in Contract No. DA-3G-034-ORD-2250, Research Institute of Temple Uni- versity, Philadelphia 44, Pa., November 18, 1958. ClF3 and does not form an addition product with it. (1.5) A. G. Streng and A. V. Grosse, "Addition and Substitution Com- In the absence of OzFz and in the presence of CIF,, pounds of Oxygen Fluorides," First Annual Progress Report for the Office ozone, as expected, retains its characteristic deep blue of Saval Research, Contract Nonr-3085(01), Research Institute of Temple University, Philadelphia, Pa., January 3, 1901. (IS) A. G. Streng and A. V. Grosse, Adrr, in Chem. Series, 86, 159 (1962). 1382 A. G. STRENG Vol. 85 color. Ozone can be isolated easily from such a mix- hydrous HF at 190-195'K. The solution changed ture either by extraction with liquid Oz,which im- color and became deep blue like that of ozone. The mediately gives a blue solution, the ClF3 remaining question might well be raised as to whether in the practically undissolved, or by high-vacuum distillation. decomposition of OzClF3 the oxygen is liberated as O3 At a total pressure of 12 p and at a temperature of or a new type of compound is formed which has a color 158'K. the violet compound does not distil. Ozone, on similar to that of ozone. It is of interest to mention the other hand, with a vapor pressure of about GOO mm. here that more than 3.5 wt.% of ozone dissolves in at l58'K. distils readily. HF at 195'K. at pressures less than one atmosphere to Following is a typical example of the preparation of form a solution similar in appearance to the one de- the violet product: 1.130 g. of OzFz was distilled under scribed above. vacuum into a reaction vessel of about 100 ~m.~volume, Freon 12, CCIJ?,, Freon 13, CClF3, perfluoropropane, melted, distributed evenly on the walls of the lower half CaFs, , C103F, hydrogen fluoride, of the reaction vessel by rotation and frozen at 90'K. HF, and , OFz, were tried as diluents. The stoichiometric amount (1: 1 ) or 0.880 g. of There was no formation of the violet intermediate C1F (measured as a gas) was added in portions of about product when Freon 12 or OFz was used as the solvent 100 mg. After each addition, the reaction vessel for OZFZand cooled ClF gas (diluted by He or 02) was warmed to 119'K. (melting point of ClF) and then was bubbled through the solutions. With Freon 13 slowly to 130'K. The violet compound formed rapidly or C103F as the solvent, the violet product formed in and, simultaneously, white solid ClF3 also, with some small amounts and collected on the bottom of the evolution of 02and small and varying amounts of Fz. reaction vessel. Its stability, however, in the pres- The reaction vessel was then cooled again to 90'K., ence of these compounds was low and OzClF3 decom- the evacuated and collected for analysis, a fresh posed completely at about 195'K. With HF, solu- portion of ClF added and the cycle repeated until all the tions of about 23% of OZC1F3 were obtained.17 A 02Fz was consumed. difficulty arises, however, from the fact that the solu- Assuming that the heat of formation of 02ClF3is tion of the violet compound in HF is stable at 190- about half-;.e., 15 kcal. per mole-of the total heat of 195'K. only under an oxygen pressure of two atmos- reaction 3, one can readily understand that overheating pheres. For this reason, the removal of HF and the can lead to the decomposition of the violet compound isolation of the violet compound are complicated. Attempts to lower the melting point of HF by the 02CIFa +02 + ClF3 (7) addition of KFls and thereby to increase the stability The relative extents of the reactions 4, 5 and 7 can of the solutions of the violet compound gave no im- be determined by analysis of the gases for 02and Fz, provement. On the contrary, in the presence of KF since only reaction 4 leads to elementary fluorine and the solutions of the violet compound in HF lost their reaction 5 proceeds without evolution of 02or Fz. color much faster than normally, most probably owing In the example above, only negligible traces of Fp to the formation of K(ClF4). were found, while the amount of 02evolved was 205.8 Dioxygen , OzC1F3, was formed ~m.~at N.T.P. or 0.294 g. Thus, all the oxygen lib- also in the reaction of OzF2 with HCI at 130-140'K. erated was due to reaction 7 corresponding to 56.9% The analysis of the reaction products showed that the by weight of the OzFz used. The remainder, or 43.1y0 reaction proceeds in accordance with the equation by weight, of OzFz combined, following equation 5, with ClF. Thus, the yield of OzClF3 was 43.1% of theory. 202F2 + HC1 OsCIF3 f HF + 02 (8) After decomposition of OzClF3, the total amount of As an example, 63.0 mg. of OzF2 was treated with 17.0 ClF3 produced was 1.493 g. as determined by weight. mg. of HCl. The oxygen evolved was determined by The over-all material balance is given in Table I. absorption in an alkaline pyrogallol solution. The TABLEI hydrogen fluoride formed was combined with NaF and determined by titration with 0.1 N NaOH after de- MATERIALBALANCE OF THE OzF2+ C1F EXPERIMENT composing the violet compound and distilling off the Reagent used, Products obtained, ClF3. The additional Ozevolved upon decomposition g. g. of OzC1F3 was determined separately. The chlorine 02Fs 1.130 OzCIF:: 0.869 CIF 0.880 02 294 trifluoride was measured as a gas and identified by the Fz ,000 infrared spectrum. The yield of OzC1F3, calculated CIF:, .84i from the amount of O2 evolved upon decomposition of the violet compound, was about 41%. Total 2.010 2.010 The suggested reaction steps are

Much effort was exerted to find the conditions neces- OzFz + 2HC1 +2HF f 02 + CI? (9) sary to minimize reactions 3 and 4, ;.e., to increase the 02F2 Cly +2ClF 02 (2) yield of 02ClF3. With C3Fs as a diluent, the yield of + + the violet compound was increased to 81% of theory, 202F2 + 2C1F ---f 20zClF3 (5) but the stability of 02C1F3 in the presence of C3FS 40zFz 2HC1 +202C1F3 + 2HF 202 (8) was found to be low. It was found also that in the + + presence of traces of water, nitrogen or nitrogen An excess of HC1 and a rise of temperature above oxyfluorides, the violet compound decomposes quickly 140'K. caused fast decomposition of 02ClF3. or does not form. At 130 and 140'K., HC1 is solid (m.p. 158.9'K.) If the violet compound, OzClF3, is warmed to about but it has a vapor pressure of about 10 mm. at 140'K. 140'K. in the presence of ClF and ClF3,it changes into a Thus, the reaction actually takes place between gaseous greenish-blue compound (or mixture of compounds), HC1 and gaseous or liquid 02Fz. The violet compound which exists only over a very narrow temperature range. partly deposited on the walls, but also dissolved in the The greenish-blue compound dissolves in C1F at about liquid O2Fz. Solutions containing up to 8.5% by weight 125-130°K., to form a greenish-blue solution, but the of OzC1F3 in O2F2 were obtained. color disappears in about three to five minutes. Prob- (17) The concentrations of 02CIFa in solutions were determined by meas- ably another blue compound was formed while pumping uring 0%evolved after the decomposition of OzClFa. off the oxygen from a violet solution of OsCIFs in an- (IS) G. H. Cady, J. Am. Chem. SOC.,66, 1431 (1934). May 20, 1963 CHEMICALPROPERTIES OF DIOXYGENDIFLUORIDE 1383

Finally, the violet product formed also when pure Reaction 11 is more difficult to control than liquid ClF3 in a quartz tube under pressure of -2 atm. reaction 5. O2 was irradiated with light at 195'K. At approximately the same conditions, dioxygen However, if the 02 pressure is only 15 mm. or O3 is difluoride reacted with a mixture of Br2, BrF and BrF3. used instead of 02, the violet product does not form. Ruff and Menzellg and Braida20 reported that upon No reaction was observed between OzF2 and ClF3 mixing Br2 and BrF3, an intermediate species, BrF, in the solid state at 90°K., or in the liquid state at formed, but pure BrF was not isolated, owing to its temperatures up to 190'K. dissociation into Br2 and BrF3. Fischer and co- Reaction with Br2, Fluorides and HBr worker~~~-*~found that in the gas phase the reaction Brt BrF3 3BrF (12) Liquid 02F2, at temperatures close to its melting + -+ point (109.?'K.), reacted vigorously when added to took place to varying extents. solid bromine cooled to 90'K. In the experiments with 02F2,a product obtained by When liquid BrF3cooled to its melting point (282'K.) mixing BrF3 with 10% of Brz was used. This product was dropped onto solid 02F2 cooled to 90°K., a spon- reacted with OZFsbetween 90' and 130°K., forming a taneous reaction occurred with evolution of heat and dark-brown (violet shaded) intermediate, which de- gas. Analysis of the reaction product showed that the composed to BrF3 and BrF5 at temperatures above reaction proceeded in accordance with the equation 130'K. If the reaction was carried out at tempera- OtF2 + BrF3 d BrFj + O2 + 46.1 kcal. (10) tures above 130'K., it proceeded rapidly and directly to BrF3, BrF5 and 02,without forming any colored The 02 was identified by the usual method of gas intermediates. analysis and the by determina- The colored intermediate products formed with BrF3 tion of its melting point (= 211.9'K.), boiling point and Br2-BrF-BrF3 mixture have only a transitory (= 313.7'K.) and (= 3.09 g./~m.~at 212'K.). existence and attempts to stabilize them were not SUC- Under milder conditions, in some experiments a cessful. Moreover, these reactions are not always brown-violet intermediate product was obtained. The reproducible. formation of this product was, however, not always The study of the reaction between OzF? and HBr reproducible for unknown reasons. In these experi- showed that when a small amount of 02F2reacts with an ments, BrF3 was condensed on the walls of the reaction excess of HBr at about 130'K. the reaction proceeds vessel at 90'K. A thin layer of 02Fzwas then con- according to the equation densed on the BrF3. The bath temperature was raised slowly and at about 130'K. the reaction between BrF3 02F2+ 2HBr +2HF + Br2 + 02 (13) and OzF2 began, forming a violet-brown compound, Dioxygen difluoride was condensed on Rashig rings in with some gas evolution. Analysis showed that the the reaction vessel. HBr was added at 90'K., allowed gas consisted mainly of oxygen (with a small amount of to condense above the 02F2and the vessel was then fluorine). The colorless liquid reaction product was slowly warmed to 130'K. Dioxygen difluoride melted identified as BrF5. The small amount of F2 was due to and its vapor reacted with the excess of HBr, liberating the partial decomposition of 0*F2. An example of a Br2and 02. The free bromine, when warmed to 140'K., weight balance of reactants and products is given in reacted with OzFr forming colorless BrF5. No addi- Table 11. tional colored products were formed. When, however, TABLEI1 the HBr layer was condensed lower and contacted with an excess of liquid 02F2at 130'K., a violet compound MATERIALBALANCE OF THE BFo 02F2EXPERIMENT + similar in appearance to 02BrF5,formed. The meas- Reacted 854 mg. BrFs + 452 mg. O?F? urements and the analysis of the reaction products indi- Reaction products cated that with an excess of 02F2 the reaction Reaction products obtained expected most probably proceeds in accordance with the equa- 02evolved during reacn. tion 27.0 c~n.~or 38.6 mg. 30$2 + HBr --f 02BrF6f HF + 02 (14) 02evolved upon decompn. of the interniediate colored The colored intermediate product decomposed at product higher temperatures, forming BrFj and liberating 106.0 ~m.~or 151,s mg. ... gaseous 02. O? total Reactivity with IZand IF6 133.0 or 190.1 mg. 144.6 ~tn.~or A spontaneous reaction occurred when liquid 02F2 206.6 mg. cooled to about llO'K. was added rapidly to BrFj total crystals cooled to 90'K. There was no visible reaction 0.37 cm1.3at m.p. or 1129.4 mg. between S-50yo solutions of 02F2in Freon 13 (CClF3) 1143 mg. and iodine at temperatures up to 195'K. Ft total Iodine pentafluoride, IF5, in contrast to its chlorine Traces ... and bromine analogs (ClF and BrF3), did not react Total with O2F2over the temperature range of 90-195'K. 1333 1 mg. 1338 .O Only a slow decomposition of 02F2 to O2 and FZtook mg. place. Under more drastic conditions the formation The violet-brown compound began to decompose at of iodine heptafluoride, IF,, will probably take place. 130'K. and decomposed completely at 170'K. to oxy- Reaction with Phosphorus and PF3 gen and BrF6. The formation of the colored intermediate product With red phosphorus, 02F2reacted vigorously when proceeded analogously to the O*F2 + ClF reaction, in added rapidly at about 1 10°K. accordance with the equation (19) 0. Ruff and W. Menzel, 2. aftorg. ailnetn. Chem., aoa, 60 (1931). (20) 0. Ruff and A. Braida, ibid., 914, 87 (1933). OzFz f BrFz -+ OZBrF) (11) (21) J. Fischer, K. I). Stennenberg and R. C. Vogel, J. Am. Chem. sa., This reaction was always accompanied by some gas 76, 1487 (1954). (22) J. Fiacher, J. Bingle and R. C. Vogel, ibid., 78, 902 (19SG). evolution due to decomposition of OrBrF5 to BrFb (23) I<. I). Stennenberg, I<. C. l'ogel and J. Fischer, ibid., 78, 1320 and O2and to partial decomposition of O2F?to O?and F2. (IY5i) 1384 A. G. STRENG Vol. 85

Phosphorus trifluoride, PF3, reacted with OzFz at violence, but the intermediate colored product either 126'K. forming PF5 and O2 forms in very small amounts or does not form at all. OzFz 4- PF3 +PFs + 02 (15) A 12.8% by weight O?F2 solution was used, for At the same time some of the liberated oxygen reacted example. It was frozen at 90'K. and a gaseous SF4- with PF3, forming a white solid which was fairly stable C103F mixture (1 : 1) was added in small portions, each at 0'. This compound was the only solid reaction containing 20-100 mg. of SF4. After the addition of product; it was not the well known POF3, which melts each portion, the reaction vessel was warmed to 130'K. at 233.4'K. and boils at 233.8'K., but rather an inter- Only traces of violet compound formed, which deposited esting polymer of POF3 on the walls of the reaction vessel above the OzF2-Cl03F mixture. The formation of the violet compound nOpFz + 3nPF3 +nPF5 + 2(POF3), (16) took place only at temperatures of 90-116'K. At On standing at 0' or at room temperature, the 130'K. a slow visible reaction between OzF2 and SF4 polymer depolymerized completely to POFs began, with evolution of O2 gas and formation of SFs. Further warming to 160-170°K. increased the reaction (POF3), e tZ( POF3). (17) rate and caused decomposition of the colored compound. If the amount of PF3 added in one portion was larger Also, decomposition of 02F2to O2and F2was noted. than 50 mg. or if the compounds were warmed quickly, The use of a smaller amount of diluent gave a larger the reaction proceeded with flame. quantity of the colored intermediate product, but it Reactivity with N02F, NF3 and N2F4 could not be preserved. The formation of the colored intermediate product probably proceeds similarly to the Dioxygen difluoride is soluble in nitryl fluoride, formation of the colored products in the 02FL + C1F NOSF, at 195'K., forming a very fluid orange solution. and O?F?f BrF3 reactions A slow decomposition of 02F2 takes place at this tem- perature, but no reaction with NO?F was observed. 02F2+ SF4 +OzSF6 (19) There was no visible reaction between nitrogen tri- This reaction is even more difficult to control than fluoride, NF3, and OzF2 at 130-140'K. Tetrafluoro- reaction 11. In most cases it proceeds directly to the hydrazine, NzF4, reacted with 02Fz at 170'K., forming formation of SFs and O2in line with eq. 18 and leads to NF3 and 02, but without the formation of any colored an explosion. With perchloryl fluoride, C103F, as a intermediate product. solvent, explosions could be avoided in some cases (at Our experience with ?JzF4 showed that upon dis- temperatures below 116'K.), but the yield and the tillation, either in Pyrex glass or in a Kel-F system, the stability of the intermediate product were very low. tetrafluorohydrazine itself formed a dark violet com- With HgS, dioxygen difluoride reacted in accord with pound which looks like the violet compound formed by the equation the reaction between O2F2and C1F. It was made clear 402F~+ H2S +2HF + 402 + SF6 f 432.9 kcal. (20) however (before using N2F4 for the reaction with 02F2), The formation of all these reaction products was es- that this violet compound is not of the same type as tablished analytically. It was observed that a very that obtained in the 02Fz+ C1F reaction. The violet slow reaction starts in the gas phase at 130'K. The compound formed by NzF4 could be distilled, for ex- vapor pressure of O2F2at this temperature is about 1 ample, from one vessel into another, together with N2F4 mm., and of H2S about 0.5 mm. When the reaction without decomposition, whereas 02CIF3 is non-volatile. vessel was further warmed slowly to 196'K., the rate According to Johnson and C~lburn,?~condensation of the reaction increased. In contradistinction to the of cold gaseous N2F4 (- 30') at relatively high pres- 02F2+ HC1 and 02Fa + HBr reactions, the 02Fz sures gives a water-white liquid. However, if the gas + H2S reaction gave no colored intermediate com- is condensed at elevated temperatures and low pres- pound. sures, the liquid obtained has a color varying from light blue to blue-black. The color is said to be due Behavior of OzF2with Some Other Substances to trace amounts of nitrosodifluoramine, NF2N0.24a Liquid dioxygen difluoride at temperatures close to Reaction with S, SF4and H2S its melting point reacted vigorously when added to charcoal cooled to 90'K. It did not appear to react, When added rapidly to cooled to SO'K., liquid however, with beryllium powder, quartz fiber and chro- 02Fz reacted instantaneously with a flash. mium trioxide, even upon warming to room tempera- , SF4,reacted violently with con- ture. centrated OzFzat about 130°K., forming SFs and 02 Solid SiF4 (m.p. 183'K.) did not react with liquid 02F2 + SF, --f SF6 4- 02 + 121.5 kcal. (18) OzF2. Approximately 2070 of SiF4 is soluble in liquid The solutions were used, therefore, to prevent too 02F2at 16O'K. without noticeable reaction. At about violent reactions. Sulfur tetrafluoride diluted with 196'K., OzFz decomposed into O2 and Fa,while SiF4re- C103F (2:3 by volume) was distilled into the reaction mained unchanged. Similarly, CF4did not react with tube containing frozen 02F2 at 90'K. Traces of an OzFz. intermediate violet-purple compound formed immedi- In contact with a Pt sheet covered with PtF4, dioxy- ately on the walls of the reaction vessel, even at 9O'K. gen difluoride exploded at 160'K. Most of the SF4 and C103F condensed on the walls Explanatory Remarks above the 02F?. The tube was then warmed above 126'K. At about 13O'K. the SF4-C103F mixture be- Experiments showed that dioxygen difluoride has a gan to melt and flow down to the OzF2. More purple- remarkably high oxidizing power. With most of the violet compound formed and the reaction went out of substances tested, it reacted at cryogenic conditions. control with an explosively violent evolution of gas and The reactions tend naturally to proceed to completion, heat. The excess of unreacted 02F2 remained on the i.e., to form the most stable reaction products. With bottom of the shattered reaction tube. The reaction CIF, BrF3 and SF4,however, as well as with HC1 and between highly diluted O2Fz and SF, proceeds without HBr, highly colored intermediate products were formed (24) 17. A. Johnson and C. B. Colbtlrn, J. Am. Chei71. Soc., 83, ,3013 when the reactions were carried out with the necessary

(I !IO 1 ) , precautions. Reactions with BrF3 and HBr were more ('14a) C. B. Colhurn and F. A. Johnson, 1mrp. Chem., 1715 (lW2). difficult to control than those with CIF and HC1, and May20, 1963 IODINE-CATALYZEDISOMERIZATION OF OLEFINS 1385

TABLEI11 evolved: the more energy liberated, the more difficult it THE HEATSOF FORMATIONOF COMPOUNDS INVOLVED AND THE is to quench the reaction and to freeze and stabilize the HEATSOF REACTIONSCONSIDERED intermediate compound. Com- AHm, Refer- The heats of formation of the compounds involved pound kcal./mole ence Reaction and the heats of reactions to be considered are given in 0zFz +4.73 =t0.3 25 OzF2 + C1F - CIFB + OZ Table 111. + 30.1 kcal. It can be seen that the reactions of OzFz with ClF and CIF -13.510 fO.ll 26 HCl evolve the smallest amounts of energy. In fact, it C1F3 -38.869fl.0 26 is much easier to slow down these reactions and obtain BrF - 20 27 2OZFz + BrF - BrFs -I- the intermediate product 02ClFa. More energy is 202 + 95.7 kcal. evolved in the reactions with bromine fluorides and HBr BrFj -64.8 28 OzF2+ BrF3 - BrFj 4- OZ and they are more difficult to control. Still higher + 46.1 kcal. amounts of energy are liberated in the reactions with BrFb -106.2 28 sulfur compounds and attempts to obtain the intermedi- SF4 -171 i + 2.5 29 ate 02SFo failed in most cases. Performing these reactions at lower temperatures SFs -288.5&0.7 26 for longer times may give better results. HzS -4.815 26 40~F2 + HzS -* SFs + In the experiments described in this paper, the reac- 2HF + 401 + 432.9 tions were carried out for only a few hours at tempera- kcal. tures up to 130-140'K. Since the AH'S of most of the HBr -8.66 f 0.05 26 3OZFz+ HBr - BrFj + expected reactions are high, much could be gained if the HF + 30, + 176.9 reactions were carried out at lower temperatures over kcal. periods of days or weeks. A series of such reactions has been started. It is found that OzF2 reacts with C1F HCl -21.97i0.09 26 202Fz + HCI -P CIS + HF + 202 + 91.5 kcal. even at 90°K., although the reaction proceeds very HF -65.14f0.03 26 slowly. The formation of the violet compound was noticed only after 3 days. The amount of the colored their intermediate product, O2BrF6,has only a transi- intermediate product seems to increase with time. A tory existence. Still more difficult to control were the further study of the 02F2+ C1F reaction and of the re- reactions with SFd and HzS,with the result that in most action of OzF2with other reactants at 77' and 9OOK. for cases the analogous intermediate, OzSFG, did not form. long periods of time is planned. These facts are in agreement with the amounts of energy Acknowledgment.-The author wishes to thank Dr. (23) A. D. Kirshenbaum, A. V. Grosse and J. G. .4ston, J. Am. Chem. Soc., 81, 6398 (1959). A. V. Grosse for his helpful suggestions, Dr. W. F. (26) JANAF Thermochemical Tables, The Dow Chemical Co., Midland, Faragher for reading and commenting on the manu- Mich., 1960. script and Mrs. L. V. Streng and Mr. A. D. Kirshen- (27) H. Brodersen and H. J. Schumacher, Z, Salurforschung, 2a, 368 baum for their contributions in the experimental work. (1947) (28) L. Stein, J. Phys. Chem., 66, 288 (1962). Acknowledgment is also due to the Office of Naval (29) DuPont's Information Bulletin: "Sulfur Tetrafluoride Technical." Research for financial support.

[CONTRIBUTION FROM THE DEPARTMENTOF CHEMISTRY, UNIVERSITY OF SOUTHERN CALIFORNIA, LOS ANGELES 7, CALIF.] The Iodine-catalyzed, Positional Isomerization of Olefins. A New Tool for the Precise Measurement of Thermodynamic Datal

BY SIDNEYW. BENSONAND A. N. BOSE RECEIVEDDECEMBER 20, 1962

It is shown that in the temperature range of 200-300°, small amounts of 12 will bring about relatively rapid positional as well as geometrical isomerization of olefins in a homogeneous gas phase reaction. There is no other chemical reaction in the system. Applied to butene-1 and -2 this permits very precise measurements of the equilibrium constants and hence the free energy differences. For the reactions butene-1 e trans-butene-2 (11') and cis-butene-2 e trans-butene-2 (Y) it is found that at 508"K., KIV = 3.48 and KIT= 1.63. The less accurate values estimated from the API tables are Klv = 4.58 and Kv = 1.74. Using KI~values at 300' we calculate 4H1r = -3.1 4~ 0.2 kcal./mole, 4.5'1"= -3.6 & 0.4 e.u. These are appreciably different from the API values of -2.i kcal./mole and -2.2 e.u. It is proposed that the API values of the entropy of butene-1 be raised by 1.4 e.u. The values of AH" and 4.5'~are in excellent agreement with API values. It is suggested that the IZcatalysis in addition to providing a valuable tool for olefin synthesis may also be used to obtain in a relatively simple manner very precise values of the differences in thermodynamic properties of olefins and their parent hydrocarbons. The application to keto-enol equilibria and other unsaturates may also be possible.

Introduction 1 The reaction of organic iodides (RI) with HI goes I + KI zK + I? 9 stoichiometrically to yield RH + Ia in the temperature I range 250-320°.2a,b :3 K + HI~RH+ I (11) RI HI I_ RH I* + + (1) A The mechanism of the reaction is atomic, proceeding through an I-atom attack on R12 For the saturated hydrocarbons, step 4 is very slow in the above temperature range so that reaction I (1) This work has been supported by Grants from the U. S. Atomic Energy Commission and the National Science Foundation. goes to essential completion (>99%). For the unsatu- (2) (a) R. A. Ogg, J. Am. Chem. Sac., 66, 526 (1934); (b) S. W. Benson rated hydrocarbons with a-H atoms, the allylic reso- and H. E. O'Iieal, J. Chem. Phys., 94, 514 (1961). nance interaction reduces the a(C-H) bond energy by