A. W. Petrocelli
General Dynarnics/Electric Boat Groton, Connecticut and D. 1. Kraus University of Rhode Island Kingston
1
II
The inorganic Superoxides
In 1810 Gay Lussac and Thenard (I) observed that "potassium combines with twice or even three times as much oxygen as it requires to pass into the state of potash." This was the first recorded evidence of the fact that oxygen can enter into chemical combination in a form other than O=. It was not. however, until 1861 that the first really careful quantitative study of these apparently "abnormal oxides" was undertaken. I n that year A. V. Harcourt (2) obtained, by the combustion of pure potassium metal in air, a yellow compound for which he was able to assign the formula K204. Harcourt's observations were verified in 1894 by Holt and Sims (3). Between the years 1905 and 1907 E. Rengade (4-9, using essentially the same method of preparation employed by Harcourt, reported the preparation of Rb202,Rb203,and Rhz04,and the corresponding cesium compounds. De Forcand (7,8), in 1910 and 1914, repeated Rengade's preparative work and also determined the heats of formation and the melting points of the various higher oxides of potassium, rubidium, and cesium. His values for these quantities are listed in Table 1. De Forcand's values for the heats of formation were obtained from experiments performed on the heat of dissociation of the oxide in concentrated sulfuric acid, and are still regarded as the most reliable. Bichowsky and Rossini (9) have accepted de Forcand's data in their compilation of the heats of formation of chemical substances. In addition to the preparation of the higher oxides of the alkali metals by combustion of the pure metals in air or oxygen, another method, involving the rapid .oxidation of the metals in liquid ammonia, was introduced in 1893 by Joannis (10). I n 1926 IG-aus and Whyte (11) perfected the method and prepared K202 and Kz04. The reaction of hydrogen peroxide with the alkali and alkaline earth metal hydroxides has also been studied as a possible synthetic route to superoxides. The question of whether a reaction occurs between HzOzand H02 or the ion of its acid dissociation, 0%(the superoxide ion) is in debate. George (12) concludes on the basis of experiments of KOz dispersed in Table 1. Heats of Formation and Melting Paints of the Higher Oxides of Potassium, Rubidium, and Cesium
Oxide
146
/
-AH, keal/mole
Journal o f Chemical Education
Hz02solution that no reaction occurs. The validity of this conclusion has been called into question because the experiment dealt with a heterogenous system and mechanisms involving reactions of H102 and perhydroxyl (HO?) continue to be suggested. The supposition that HOz is a rather strong acid and the reaction is in fact with 0%ion is widely held. I n 1952, two Russian investigators, I. A. Kazarnovskii and A. B. Neiding (IS), proposed an unusual reaction involving the format~on of a compound, K202.2H102 (potassium peroxide dihydroperoxidate). They claimed that this compound, formed by the reaction of KOH and H202, can be kept without marked decomposition a t the temperature of solid CO2 (-78°C). However, at higher temperatures they claimed that it disproportionates into the superoxide, KO?, and water. Kazarnovskii and Neiding proposed the following as the main reaction for the disproportionation:
-
3K202.2H202 2KOn
+ 4 (KOH.H*O) + 301
(1)
They proposed that the main reaction (I), consists of an oxidation of the K?02by Hz02according to: KaOz.2H20
-
-
2K01
+ 2H10
(2)
followed by the reaction: 2KOz
+ 3H,O
-
Z(KOH.H.0)
+ l1/?0a
(3)
As a mechanism of the main reaction they write, K.++o.- (H.O.).
KtO.-
+ KtOH- + H202 + OH
(4)
wherein the colorless diamagnetic ion 02-gives up an electron and forms the yellow paramagnetic species 02-. The OH radical then reacts with H20zto give HOs thus
-
H?O*
+ OH
KOH
+ H02
and :
HO.
+ HnO
(5)
+ Ha0
(6)
KOz
The Russians arrived a t most of the above conclusions on the basis of magnetic measurements. They followed the build up of paramagnetism in the reaction mixture as a result of the formation of the paramagnetic KOz and subsequent decrease in paramagnetism resulting from the reaction of the KOZ with H 2 0 to form the hydroxide and oxygen. Recent work on this reaction at General Dynamics/Electric Boat by Petrocelli has resulted in the isolation of the KOz in verv " hinh - vields " and purity (14). Work by Symons (I5), Wilmarth (16), and Forscheimer and Taube (17) indicates that in reactions between ueroxides and hvdroxides. the ozonide suecies (Oa-) is important. The Russian mechanism, however, fails to recognize this role of 03-in the formation of KO2 and this impasse needs to be resolved.
It is to be noted that up to this point no mention of the superoxides of lithium, sodium, or hydrogen has been made. Lithium superoxide (LiOz) has not as yet beeu isolated. Claims have been made in the Russian literature for the isolation of Li02 but they have not as yet been substantiated (18). Thompson and Kleinberg (19) have offered convincing evidence that LiO, does form when lithium is rapidly oxidized in liquid ammonia a t -7S°C. A bright lemon-yellow solution is formed, which is characteristic of superoxide formation. I n addition they found that the absorption spectra of oxidized liquid ammonia solutions of lithium, sodium, and potassium show all three solutions to have absorption bands close to the same wavelength, 350 mp. The similarity of the absorption spectra of the lithium solution to that of the sodium and potassium solutions, which are known to contain the respective superoxides, supports the claim that lithium forms a stable superoxide in liquid ammonia solution. However, evaporation of the ammonia did not result in the successful isolation of solid LiOa. Sodium superoxide (NaO?) was prepared for the first time in 1949 by Kleinberg, et al. (20). They obtained maximum yield of sodium superoxide by subjecting the peroxide to a temperature of 490°C a t an oxygen pressure of 298 atmospheres for a period of 100 hr. I n 1955, George (21, 28) demonstrated, by means of paramagnetic resonance techniques, that pale yellow samples of commercial sodium peroxide contained approximately 10% sodium superoxide. Margrave (23) has reported some thermodynamic data on NaOl which are quite interesting and cast some light on the reasons for the difficulty of obtaining NaO? by a direct oxidization process. Margrave considered, th~rmodynamically,the decomposition of NaOp (s), i.e.,
Using the low-temperature heat capacity data of Todd (24) on NaO? and NanOzhe established the free etlergy functions for the two solids. The thermodynamic data for O2 (g) are well known. The resdts of his theoretical calculations predict that the equilibrium oxygen pressure is 4 atm at 29S°K and 40 20 atm at 600°K. Some actual experimental work carried out by Friedman and Margrave (25) indicates that the equilibrium oxygen pressure is actually around 30 atm at 600°K. The case of the higher oxides of hydrogen is even more interesting and will be briefly considered here. The importance of obtaining hydrogen snperoxide (Hop)as a stable compound is obvious. Not only would it provide a nearly ideal condensed state storage medium for breathing-oxygen, but it also would probably be a very effective high-energy oxidizer in the rocket propellant field. Most claims for the isolation of hydrogen superoxide, which will be written as Hs04 in order to avoid confusion with the perhydroxyl radical HOI, have come from Russian investigators (26-28). All such claims appear to have been effectively refuted by recent work (89-31). Benson (29) has proposed, on the basis of thermodynamic and kinetic evidence, that HzOl can not exist. This question is of such importance in the realm of theoretical and practical chemistry that it is still being actively studied by a small number of investigators.
*
The tendency of the alkali metals to form higher oxides upon direct combustion with oxygen increases with the ionic radius of the alkali metal ion. In Table 2 are listed the ionic radii of the metal ions and the oxides formed by combustion in an excess of oxygen. The size of the anions increases in the order O=, 0%-, and 02=. These facts clearly suggest that increasing cation size is essential to promote crystal stability by permitting inclusion of the larger 02= and 02-ions (3Z). Table 2. Stable Forms of the Alkali Metal Oxides Produced b y Direct Combustion of the Metals in an Excess of Oxvaen ,
--
Ionje radius Metal
Oxide farmed
The structure of the superoxide ion is of interest and should be considered. For many years following their discovery the superoxides were considered to be tetroxides having a formula as written in Table 1, i.e., M?04. Siuce sulfur forms tetrasulfide compounds the above formulation for the superoxides appeared quite reasonable. The electronic structure of the "tetroxide ion" was pictured as shown in Figure l a (33).
Figure 1. Electronic structure of 04-(tetroxide ion), 02- (superoxide ion), and O3-(~esquioxideion), reading from left to right.
I n 1931 Pauling (34) introduced the concept of the three-electron bond and suggested the possibility that the snperoxide ion might indeed be an 0%-species sperivs cnntniniug R i h r w - d ~ ~ r tmnd. o ~ i l ' h r strurihuwn i t t Fiuure 1 b. I'tdine rcnsoned run, of this . ~ ~ ion - is~ that if the formulation of the superoxide ion Ts actually as shown in Figure l b then the compounds containing this ion should display paramagnetism by virtue of the fact that the three-electron bond contains one unpaired electron. On the other hand one should expect no paramagnetism if the oxygen is in the form of a tetroxide ion. Following a suggestion by Pauling, Neuman (35), in 1934, made a study of the magnetic properties of potassium superoxide and found that it is indeed paramagnetic. Final confirmation of the existence of the 02-ion resulted from the X-ray studies of Kossatochkin and Kotov (36), in 1936 and Helms and Klemm (37, 38) in 1939. Their X-ray data demonstrated that a t room temperature superoxides possess a tetragonal facecentered lattice of the calcium carbide type. Carter, Margrave and Templeton (39) later showed that KO2 undergoes a transition in the region 60 to 100°C, to a cubic form. The elementary cell contains four alkali metal atoms and eight oxygen atoms. The distance ~
~~~~~
-
Volume 40, Number 3, March 1963
/
147
Toble 3.
Surnrna~yof lmporlont Properties of Various Inorganic Air Revitalization Moterials
Compound
Formula
Lithium ozonide Sodium ozonide Potassium ozonide Lithium superoxide Sodium superoxide Potassium superoxide Cdeium superoxide Sodium chlorllte (candles) Hydrogen peroxide Hydrogen superoxide
LiOs NaOs KO8 Lion NaO. KO2 Ca(O& NaCIOa &Or H~OI
Lb Ol1lb eam~ound 0.73 0.56 0.46 0.61 0.43 0.34 0.46 0.40 0.47 0.73
Use as air revitalization system
Lb generated base/lb com~ound
source
CO? ~ i e kUD
Remarks
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yes Yes Yes Yes Yes Yes Yes No No No
1 2 3 4 5 6 7 8 9 10
0 2
0.44 0.56 0.65 0.62 0.73 0.79 0.71 None None None
7. Produced in low vield. ~ i & vdesirable due to hieh mkltine t CdOH),. - ~. o i n of . .8. Widely used as emergency 0. source. 9. Liquid solutions. Requires auxiliary ehemied for C03 removal. 10. Essentially unsubstantiated Russian claims. Highly attractive as high energy oxidizer.
between adjacent oxygen nuclei in the superoxide ion is 1.28 + 0.07 A. This inter-atomic distance is ip good agreement with the calculated value of 1.24 A (33), expected for the single bond plus the three-electron bond. At the present time, therefore, there is little doubt that the electronic structure of the superoxide ion is as illustrated in Figure lb. One of the interesting and important questions concerning superoxide chemistry has been whether or not in the course of the decomposition of the superoxides a stable compound Mz03is formed. Many workers (7, 40, 41) have reported the preparation of the oxide M203. The proposed structure of the sesqnioxide ion is shown in Figure lc. In 1933 Centerszwer and Blumenthal (41) reported results of a thermal decomposition study of the 1M02, Md&, and MzOz compounds of potassium, rubidium, and cesium. They measured equilibrium pressures a t various temperatures in the range 198 to 580°C. They did not, however, report equilibrium pressure data as a function of sample composition at a given temperature. Thus, pressure-composition isotherms cannot be constructed from their data and a definite conclusion as to which oxides form in the course of the thermal decomposition of the superoxide cannot be reached. I n addition the dissociation pressures which they report are extremely high as compared to the values obtained in a subsequent similar study made by Kaaarnovskii and Raikhshtein (42). Possible reasons for the apparent abnormality of Centerszwer's results have been advanced by Petrocelli (43). Helms and Klemm (57,58) carried out X-ray studies of samples which were analyzed as K2O3. They found that this material does not contain any 0,- ion, and is actually a mixture of superoxide and peroxide, KOz KzOz. Further, in 1947, it was quite definitely established by Kaaarnovskii and Raikhshtein (42) that in the thermal decomposition of potassium superoxide below its melting point, K203 is not formed as an intermediate oxide. Their results, which disagree with those of Centerszwer, showed that on the removal of oxygen from the superoxide a t a constant temperature, the pressure remains constant until the composition of
+
148
/
Journal of Chemicol Education
the peroxide, IZ201,is reached, where a sudden drop in pressure occurs. On the basis of this evidence it seems quite conclusive that ILO, does not form. This study was repeated in part by Petrocelli (45) and the results were found to be in good agreement with those of IZazarnovskii and Raikhshtein. Present-day interest in alkali and alkaline earth superoxides is quite high, due primarily to the potential which these compounds possess as inorganic air revitalization materials. Table 3 summarizes some of the important properties of compounds useful for air revitalization purposes. A review of the chemistry of the inorganic ozonides has been recently published (44). Acknowledgment
The authors wish to acknowledge Robert Chiarenselli for his assistance in the preparation of this article. Literature Cited
(1) GAYLUBSAC, J. I,., AND THENARD, I,. J . , Reehercke~Physicn Chemipes, 3, 132 (1811 I. (2) HARCOVRT, A. V., J. Cheat. Soc., 14, 267 (1862). , (1894). (3) HOLT,W., AND SIMS.W. E., J . Chem. ~ o c .65,432 (4) RENGADE. E.. Cornst. R d . . 140.1183 11905). i 5 j REKGADE; E.; i b i d , 142, 11'49, i533 (lb06). ' (6) RENG~DB, E., Ann. Chin&.Phys., 11,345 (1907). ( 7 ) DE FORCAND, R., Com.pl. JIML~.,158,843 (1914). (8) DE FORCAND, P*., ihid , 150,1399 (1910). (9) BICHOWSEY, F. R., A N D KOESINI,F. O., "The Thermochemistry of the Chemical Substances," Reinhold Pnhlishing Co., New York, 1936. (10) JoaNn~s,A., Compt. Rend., 116,1370 (1893). (11) K R A ~ SC. , A,, AND WHYTCE. F., J . Am. Chcm. Soc., 4 8 , 1781 (1926). ~ OFaradnulay ~ Y Sm.,2,196 (1947). (12) GEORGE, P.: ~ ~ S C Z L S S of I., A N D NEIDING,A. B., Dokhdy, Amad. (13) K*ZARNOVSKI~, Xauk. SSSR, 86,717 (19v2). A. W., General D~gnmics/Electric Boat (14) PFTROCELLI, Report 17413-62023, Drc. 31, 1960, 55 pp. M. C. R.,AND TOWNSEND, M. E., J . Chem. Roe., (15) SYMONS, 1959, 263. , R., SCRWARTZ, N., A N D WILMARTH, W.K., (16) ~ I U L ~ A N OC. J . Phys. Chem., 63,358 (1959). (17) FORSHEIMER, 0 . L.,I N D TAUBE,H., J. Am. Chem. Soc., 76, 70QQ1 1 954) -, (18) VOLNOY, I. I., AND SHATUNINA, A. K.,IZU& Akad. Nouk. S~L"ROtde2,Khim. Navk., 1957,762. \---
(19) THOMPSON, J. K., AND KLEINBERG, J., J. Am. Chenh. Soc., 73. 1243 fl(1.511.
J. E., INGRAM, D. J. E., SYMONS, M. C. R., (22) BENNETT, GEORGE, J. P., Phzl. Alaquzine,. 46,. 443 . P.,. AND GRIFFITH, (1955). J. L., J. CHEM.EDUC.,32,520 (1955). (231 MARGRAVE, S. S., J. Am. Chem. Rne.: 75, 1229 (1953). (24) TODD, P. J., A N D MARGRAVE, J. I,., Unpublished work, (25) FRIEDMAN, Unirersity of Wisconsin, 1954. (26) KOBOZEY, N. I., SKOHOKHODOY, I. I., SEKWSOV,L. I., A N D EREMIN, Ye. N., Zhur. Fii. Khim., 30,2580 (1956). (27) KOBOZEY, x. I., ~BOROKHODOV, I. I., NEKRASOV, L. I., A K D MAKAHOVA, E. I., ibid., 31, 1843 (1957). .4. I., ibid., 31,515 (1957). (28) GORBAXEV, N W., , J. Chem. Phys., 33, 306 (1960). (29) ~ N S O 8. J. G., AND RUTLEDGE, P. V., wa1117e, 174, 2013 (30) MARSHALL, (1959). , 1685 (31) GIGUERE,P. A,, AND CHIN, D., J. Chem. P ~ u s .31, (1959).
(32) REMY,H., "Treatise on Inorganic Chemistry," Elsevier Publishinz Co.. New York. 1956. PAULING, "kaktre of the Chemical Bond, " Cornell University Press, New York, 1948. PAUI,ING, L., J. AWL.C.5en~.Soc., 53, 3225 (1931 ). NEUMAN, E. W., J. Chem. Phgs., 2 , 3 1 (1934). KOSSATOCHKIN, W., AND K o ~ o v ,W., J. Chem. Phvs., 4, 458 (1936). . . (37) HELMS,H., A N D KLEMM,W., Z . Anow. Allyern. Chem., 241, 07 (1939). W., ibid., 242,201 (1939). (38) HELMS,H., AND KLEMM, J. L., A N D TEMFLETON, P. H., (3:)) CARTER,G . F., MARGRAVE, Arta Cryst., 5,851 (1952). (40) RENGADE, FJ., Ann. Chem. Phys., ll,348(1907). M., AND BLUMENTHAL, M.. Bull. Intcm. (41) CENTERSZNER, Acad. PolonaisP, Clnsse x i . wulh. nnt., 19338, 499 (1933). I., A N D KAIKHSHTEIN, S. I., J . Phys. Chem (42) KAZARNOVSKII, (U.8.SR.), 21,245 (1937). A. W., DWS.Abst., X X I , 1081 (1960). (43) PETROCELLI,
c,
Volume 40, Number 3, March 1963
/
149