EXTRACTION OF ANIONSFROM MOLTEN LiN03-KN03 was unsuccessful. Based on present estimates, the reaction should be exothermic by about 20 kcal/mol. A further attempt to observe this reaction seems warran ted. From our previous estimate6 of the lattice energy of NF4fF- as 147 kcal/mol and accessory data cited therewith and above, we estimate the enthalpy of formation of that substance to be -4 d= 10 kcal/mol. Decomposition t o NF3 and FSa t 25" is therefore exothermic by 26 f 10 kcal/mol. Nevertheless, the formation of NF4+F- a t low temperatures is thermodynamically reasonable since NF4+ is stable with respect to its likely decomposition products, F- is stable, and their union to form NF4+F- is favored by the lattice energy. The crystal is probably only kinetically stable, even a t low temperatures, with respect to decomposition to nitrogen trifluoride and fluorine. The formation of NF4+F- in the irradiation of mixtures of nitrogen trifluoride and fluorine a t low temperatures therefore appears to be a reasonable possibility. The dependence of the yield on the mole ratio of fluorine to nitrogen trifluoride suggests that the rate of the reaction depends on the limited solubility of the latter substance in the former. The yield should be capable of enhancement by use of a stirred reactor. The alternative possibility for the product of this reaction,34 NF4fHFz- via a possible hydrogen fluoride impurity, seems unlikely for several reasons. (1) Precautions were taken to remove hydrogen fluoride and water from the reagents. ( 2 ) If the product had been NFdfHFz-, we are confident that we would have ob(34) We are grateful to a referee for pointing out t o us this possibility
Inorganic Chemistry, Vol. 11, No. 7 , 1972 1701 served H F + in the mass spectrum of the decomposition products; we saw no trace of it. (3) The decomposition temperature of NF4+HF2- is given by Tolberg, et al.,35as about 230°K, whereas our product decomposed below 143'K. Further investigation of the radiation-induced reaction between nitrogen trifluoride and fluorine a t low temperatures appears to be merited. Additional confirmation of NFj as a product could be obtained by improved mass spectrometry designed to detect fluorine in the decomposition products and by laser-Raman spectroscopy, which was not available to us when the experiments described herein were done. Acknowledgments.-The work reported here was supported in large part by the Advanced Research Projects Agency under Contract No. DA-31-124-ARO (D)-54, monitored by the U. S. Army Research Office, Durham, N. C. The preparation of significant quantities of NF4BF4 was supported in part by a subcontract from the Dow Chemical Company, under Contract No. F04611-67-C-0025 supported by the U. S. Air Force. We are grateful t o Dr. G. C. Sinke and the Thermochemical Laboratory of Dow Chemical Company for permission to quote from their thermochemical results prior to publication. We are grateful also to Dr. W. E. Tolberg of Stanford Research Institute for useful discussions and to Dr. K. 0. Christe of the Rocketdyne Division, North American Rockwell, Inc., for discussions and for assistance with the spectroscopic characterization of NF4BF4. (35) W E Tolberg, R T. Rewick, G. R Zeilenga, M. P Dolder, and M E. Hill, submitted for publication.
CONTRIBUTION FROM THE DEPARTMENT OF CHEMISTRY, MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASSACHUSETTS 02139
Extraction of Some Anions from Molten Lithium Nitrate-Potassium Nitrate by Tetraoctylphosphonium Nitrate in Polyphenyl or 1-Nitronaphthalene Solvent BY ZOILO C. H. TAN*'
AND
J. W. IRVINE, JR.
Received October 29, 1971 Tetraoctylphosphonium nitrate (TOPN) in a polyphenyl or 1-nitronaphthalene solvent has been used to extract perrhenate, chloride, and A4gClz-ions from a eutectic molten salt mixture of LiNOa-KN03 at 150" The distribution coefficients of perrhenate and of chloride with TOPN in polyphenyl are comparable t o those obtained previously with tetraheptylammonium nitrate (THAN), while the distribution coefficient of AgClz- in TOPN is greater. Distribution coefficients using 1-nitronaphthalene as the organic solvent are somewhat higher for Reo4- both with the phosphonium salt and with the ammonium salt The dependence of the perrhenate distribution on the temperature and solute concentration was also studied with both pblyphenyl and 1-nitronaphthalene. As in the case with THAN, the distribution of the anions is interpreted in terms of a simple anion-exchange equilibrium followed by polymerization of some species in the organic phase. The equilibrium constants for the anion-exchange and the dimerization constants were derived from the distribution data. Both quaternary salts polymerized strongly even a t low solute concentrations although polymerizatioll in 1-nitronaphthalene I S less than that in polyphenyl. Also, the polymerization of the ammonium salt is stronger than that of the phosphonium salt in either solvent.
Introduction Tetraheptylammonium nitrate (THAN) has been used to extract simple and complex anions from a LiN03-KN03 eutectic melt.2 The distribution of the ( 1 ) Reseaich Laboratories, Eastman Kodak Co , Rochester, N. Y. 14650 (2) I J Gal, J Mendez, and J. w Irvine, Jr., I n w g C h m . 7 , 985 (1968).
anions between the melt and the extractant has been interpreted in terms of a simple anion-exchange eqUilibrium followed by polymerization of some species in the organic phase. I t was of interest to extend these studies to Other Organic having ion-exchange characteristics.
1702 Inorganic Chemistry, Vol. 11, No. 7 , 1972
For an extension of this work, a new extractant, tetran-octylphosphonium nitrate (TOPN), was synthesized. A eutectic mixture of o-terphenyl, m-terphenyl, and biphenyl (designated as polyphenyl") or l-nitronaphthalene was used as the organic solvent. The eutectic mixture of LiN03-KN03 was the molten salt phase in all studies. The distribution of the anionic species Re04-, C1-, and AgClz- between the nitrate melt and TOPN in polyphenyl solvent was studied. The distribution of Reo4- was also studied with 1-nitronaphthalene as the organic solvent in an effort to study the effect of variation in the dielectric constant of the solvent on extraction. The extraction of Reo4with 1-nitronaphthalene as solvent was also determined with THAN as the extractant. Experimental Section The preparation of THAN, polyphenyl solvent, and eutectic nitrate melt and the procurement of the radioactive tracers '8@Re,38Cl,and llomAgare described elsewhere.2 Preparation of Tetra-n-octylphosphonium Nitrate.-The method followed for the intermediate iodide is based on that suggested by Jerche13 and refined by E l h a n a r ~ . ~Tri-n-octylphosphine (M and T Chemicals Inc.) and 1-iodooctane (Eastman White Label) were mixed in the molar ratio of 1.1 : 1. Absolute ethyl alcohol was added to the mixture a t the ratio of 1.5 ml of alcohol to 0.1 mol of tri-n-octylphosphine used. The mixture was stirred and refluxed for 5 hr at -110'. Subsequent cooling a t room temperature yielded a white crystalline mass. The tetra-noctylphosphonium iodide was recrystallized several times from analytical grade ethyl acetate until it was pure white; after being dried under vacuum for 48 hr, it had a measured melting point of 8849.5". Seventy-three grams of the iodide salt was dissolved in 180 ml of reagent grade toluene. The solution was shaken with 150 ml of 4 N NaN03 solution for about 30 min with an automatic shaker. After discarding the aqueous phase, the organic solution was shaken again with freshly prepared 4 N h-aN03 solution. The shaking process was repeated a t least 20 times or until the content of iodide in the washed solution was negligible. The final traces of iodide in the organic solution were removed by adding 25 ml of 0.1 N AgN03 solution. After separating the aqueous phase and precipitate, the organic phase was filtered and washed with water until tests showed that Ag+ and I- were absent. The solution was evaporated to dryness and then dried under vacuum for 48 hr. The solid product, recrystallized twice from a mixture of ethyl acetate and petroleum ether (bp 30-60') and dried under vacuum for 48 hr, had a measured melting point of 65-66.5'. The yield was 80%. Anal. Calcd for C32H&03P: C, '70.4; H, 12.55; N, 2.56; P, 5.67. Found: C, 69.65; H, 12.75; N, 2.52; P, 5.77. 1-Nitronaphthalene.-This compound was Eastman White Label grade with a melting point of 56.5-58.5". Prior to use, it was melted and passed through a heated column of activated alumina (Alcoa, Grade F-1, 14-28 mesh). Procedure for Determining Distribution Coefficients.-The experimental technique, using radioactive isotopes, was described previously ,z In all extraction processes, an equilibration time of 60 min, sufficient to obtain equilibrium, was chosen. The molal distribution coefficient, D , of an anionic species was calculated as D =
ZOILOC. H . TANAND J. W. IRVINE, JR. TABLE I DISTRIBUTION COEFFICIENT, D , OF Re04- A N D C1- BETWEEN MOLTEN LiN03-KN03 ASD TOPN IX POLYPHENYL AS A FUNCTION OF INITIAL TOPh' COSCENTRATION C i ~ ~ CiTOPN, m
0 4.89 x 1.00 x 2.02 x 3.00x 5.00 x 8.00 x 1.00 x 2.00 x 3.00 X 5.00 x 8.00 x 1.00 x Q
ture 150 +
DR~o~-
~
DCI-
