March 20, 1963
COMMUNICATIONS TO THE EDITOR
825
very important that there be available a theoretical approach capable of treating such compounds and semiquantitatively predicting their stabilities. We wish to report that an adaptation of the simple molecular orbital theory is applicable to the problem, and to give the results of calculations of the stabilities of some inert gas halides. The energy of an electron in a localized two-center molecular orbital, $ = N ( + A A+G), is given by the equation6 w = % (1 $. x)gA (1 - x2)'"@ f % ( 1 - x)qC (1) where x = (1 - X z ) , / ( l A*) is defined as the bond ionicity, qa and qG are coulomb integrals for the halogen 9 and the central atom, respectively, and p is the exchange integral; 2 W measures the energy of the procCH3 ess l/nGA, -+ l/nG"+ A+ 2e-. From W and data available in the literature, the energy of formation of gaseous GA, from the elements is calculated. The coulomb integrals are approximated by expressions conI11 taining the bond ionicity, the ionization potentials and 0 the valence state preparation energy (v.s.p.e.)of the inert gas atom, the ionization potential and electron affinity CH3 I of the halogen, the G-A bond distance, and a geometric [cHz-C H ~ -C H - ic--H ~ factor.6 The G-A exchange integral is approximated 0 by the p for the appropriate interhalogen compound, using the equation of Pauling and Sherman,' and corIV recting for v.s.p.e. of the inert gas atom and van der ( E 15)900))268 (20,000), 335 (250). Both compounds Waals repulsion. have identical n.m.r. absorptionlo a t 6 1.6 for the Cg, Xenon tetrafluoride has a square planar configuraC, methylenes; 6 2.3 for the C6, CS methylenes; 6 2.0 tion,8a and XeFz is linear.8b The same geometries for the C2-CH3; 6 1.2 for the side-chain methylenes and were assumed for other possible inert gas halides GA, methenes; and 6 0.8 for the side-chain methyls. and GA,. The experimental valuesa of R were used for Chlorobiumquinone (I) is unique among natural the Xe compounds. Bond lengths of the other inert vitamins K in having a double bond conjugate with the gas halides were estimated from the Pauling covalent quinone'' and one carbon less than the normal multipleradiig of the halogen A and of the halogen adjacent to of-five carbons in the side chain. Although a vinyl the inert gas in the periodic table. Other estimates naphthoquinone has not been found previously, it has were obtained from the known bond 1engths'O of the been postulated as an intermediate in oxidative phoshalides of elements immediately preceding the inert gas phorylation.'? in the periodic table. Since for the tetrahalides the The absence of coenzyme Q and plastoquinone in C. calculation is rather sensitive to the exact value asfhiosulfatophilurn is of interest, as is the fact that addisumed for R,calculations were made for a range of tion of a mild oxidizing agent during the extraction reasonable values. The valence states were assumed to substantially increases the yield of chlorobiumquinone. be p2dsfor the tetrahalides and ps for the dihalides, and This is not the case for the accompanying vitamin the v.s.p.e. values were estimated from spectroscopic K2(35),and suggests that chlorobiumquinone is present data. mostly in a reduced form. Best values of the energies and ionicities are shown in (9) hI. Tishler, I,. F. Fieser a n d N . L . Wendler, J . A m . Chem. SOC.,62, Table I. The model predicts considerable stability for 28063 (1940). the known compound XeF4, and somewhat lower stai10) I n deuteriochloroform with internal tetramethylsilane. (11) A quinone recently detected spectrophotometrically in Sarcina lrcteu bility for XeF,. Krypton tetrafluoride is expected to may be of this t y p e [ D . H. L. Bishop, K. P. P a n d y a a n d H . K. King, Bioexist, while KrF2, XeC14and ArF4 are marginal. ( h e m . J . 83, GO6 (19G2)I. The method predicts the intuitively expected increase (12) I. Chmielewska, Biochim. Biophys. Acta, 39, 170 (1960). in stability of the halides with increasing atomic num(13) Fellow of t h e Consejo Nacional d e Investigaciones Cientificas y TPcnicas de la Republica Argentina, 1962. ber of the inert gas, and greater stability for fluorides DEPARTMEKT OF CHEMISTRY BENJAMIN FRYDMAN' than for the corresponding chlorides. More quantitaUSIVERSITYOF CALIFORNIA HENRYRAPOPORT tively, a change in the central atom from xenon to krypBERKELEY, CALIFORXIA ton or even argon is expected to entail a smaller deRECEIVED JANUARY 11, 1963 crease in stability than a change of ligand from fluorine to chlorine. Calculations were not made for bromides BOND ENERGIES AND IONIC CHARACTER OF INERT GAS and iodides, since the trends in the calculated results HALIDES' strongly suggest that none of these would be stable. Sir: A fluoride of radon has been prepared.? Insufficient Much interest has been aroused by the recent demoninformation about the ionization potentials of radon is ~ t r a t i o n ~that - ~ a t least ssme of the so-called inert (6) R . G. Pearson, S. Yamada and H . B. G r a y , Pvoc. 7th I C C C , G.5 (1962); gases are able to form true chemical compounds. It is R . G. Pearson a n d H. B. G r a y , Inorg. Chem., in press.
tion coefficients supports the molecular formula of chlorobiumquinone. All these structural assignments were confirmed and further features established by a detailed study of chlorobiumquinone's n.m.r. spectrum in comparison with those of vitamin Kncao),plastoquinone and coenzyme Qlo, presented in Table 11. Thus structure I is uniquely defined for chlorobiumquinone. Further confirmation was found in a comparison of octadecahydrochlorobiumquinone (111) (C46HaoOz; Found, C, S2.9; H , 11.8) with hexahydrovitamin K1 (IY).$ In the ultraviolet, I11 has X ~ ~ ~ ' 260 " " e mp ( E 16,(100)268 (20,100), 333 (250) ; IV absorbs a t 260 mp
+
+
+
+
(1) Acknowledgment is made t o the National Science Foundation for support of this research. ( 2 ) H . H . Claassen, H. Selig a n d J. G . M a l m , J . A m . Chem. Soc., 84, 3593
(1962).
(3) C . L. Chernick, el ul., Science, 188, 136 (1962). (4) P. R . Fields, L. Stein a n d M. H . Zirin, J. A m . Chcm. S O L ,84, 4164 (1962). (5) J. L. Weeks, C. L. Chernick a n d M . S. Matheson, ibid., 84, 4612 (19G2).
+
(7) L. Pauling a n d J. Sherman, J . A m . Chem. SOL.,69, 1450 (1937). (8) (a) J. A. Ibers a n d W. C . Hamilton, Science, 139, 106 (1963); (b) H . Levy and P. A. Agron, .I. A m , Chem. SOC.,8 8 , 241 (1963); we t h a n k Dr. W . C. Hamilton for communicating these results t o u s prior t o publication. (9) L, Pauling, "The N a t u r e of t h e Chemical Bond," Cornell University Press, I t h a c a , pi. Y.,1960, p. 224. (10) "Interatomic Distances," T h e Chemical Society, London, 1958. (11) C. E . Moore, "Atomic Energy Levels," K a t l . Bur. S t d . Circ. 467, Washington, D. C., 1949, 1952, 1958.
826
COMMUNICATIONS TO THE EDITOR TABLE I
Molecule
XeF4 KrF4
R, A. 1.92 1.72 1.72 1.78 1.78 1.60 1.64 1.40 2.30 2.34 2.11 2.11 1.98 1.72 2.30
Energy of formation from elements, kcal./mole Ionicity, G(g) n/2Az(g) -C GAn (9) %
58 49 55 47 52 53 51 33 45 44 37 40 63 63 54
+ - 123.1 - 23.5 - 53.0" - 0.5 - 25.8"
Dissociation energy per bond, kcal./mole I/nGAn(g) + 1/nG(d A(g)
+
49.6 24.4 31.7" 18.8 25.1" ArF4 30.7 11.0 53.7 5.3 SeF4 507.1 -108.1 XeC14 50.9 15.8 61.1 13.3 KrC14 137.4 5.8 132.3" - 4.6" XeF2 11.1 24.2 KrF2 5.3 16.1 XeC12 42.9 7.1 a Calculated using an extrapolated value of 52.5 e.v. for the fourth ionization potential of Kr.
-
-
available to permit calculations for halides of this element. However, the trends in Table I make it fairly certain that if the data were available, calculation would predict RnF4 and RnF2,and perhaps RnC14 as well. Another interesting feature is the sharp break in the sequence of predicted energies between the halides of argon and neon. While the decrease in calculated bond energy of the fluorides is slow enough from xenon to argon so that ArF4 is still a possibility, the sudden drop between argon and neon virtually guarantees that neon halides cannot exist. The present results indicate that the inert gas halides fit naturally into the sequence of halides of the nonmetallic elements in their higher valence states, and that no special bonding assumptions are needed. DEPARTMENT OF CHEMISTRY JAMESH . WATERS COLUMBIA UNIVERSITY HARRYB. GRAY SEW YORK27, h-.Y. RECEIVED JASUARY7, 1963 T H E BONDING IN T H E INERT GAS-HALOGEN COMPOUNDS-THE LIKELY EXISTENCE O F HELIUM DIFLUORIDE
VOl. 85
Among these discussions, there has appeared a revival of the molecular orbital description proposed by PimentelsJ' and, we feel, the model provides a simple and sufficient basis for explaining the existence of such compounds as XeF2. The ease of extension of the scheme to such compounds as and XeF4 has been amply pointed out by ~ t h e r s and ~ J ~need ~ ~ not ~ be reproduced here. There is, however, one other aspect of the molecular orbital description offered by Pirnentel' that deserves consideration. The essential similarity of the molecular orbital descriptions of HFZ- and 1 3 - was noted' and remarked upon again by Pimentel and McClellan. lo Experimental support for this connection has also appeared" and we are encouraged to explore its implications in reference to inert gas chemistry. In particular, the compound HeFz can be expected to be stable. The molecular orbital description of HF2-, based upon the halogen axial 2p orbitals and the hydrogen Is orbital, need not be repeated here and its applicability to the isoelectronic molecule HeFz is obvious. We can, however, make some comparisons between HFz- and HeF? that may aid in searching for this interesting molecule. A rough estimate of the infrared spectrum of HeFz can be based upon that of HF2-. The vibrational frequencies and a set of force constants for HF2are shown in Table I.I2 These force constants transferred to HeFz lead to the predicted frequencies shown in the second row of Table I. We feel, however, that TABLE I THEVIBRATIONAL POTENTIAL FUNCTION AKD FREQUENCIES OF H F z - (EXPERIMENT) AND HeF2 (PREDICTED) Frequency, cm. -1 V,
Y2
Y3
Bond (Raman (Infrared (Infrared active) length, active) active)
HF?HeF2 HeF2
600 600 585
1230 640 6i3
1425 743 1600
A
1.13 1.13 1.08
Force constant' jr frr
2.31 2.31 3.47
1.72 1.72 0.35
fa
0.28 .28 28
jF = bond stretching force constant (in millidynes/Angstrom) ; frr
= bond stretching interaction force constant (in millidynes/
Angstrom); fa = angle bending force constant (in millidyneKngstromlradian).
the spectrum of XeFz is relevant here. Smith has conc l ~ d e d that ' ~ the ratio frr/fr is much smaller for XeF2 than for the trihalide ions. We do not share his view that this difference vitiates the molecular orbital bonding description, for a reasonable rationale can be formulated to explain it. The molecules XeFz and IClz-, though isoelectronic (in bonding electrons), involve quite different formal charge distributions. The proposed molecular orbitals tend to place somewhat less than one electron charge on the central atom and somewhat more than 1V2 electron charges on the terminal atoms. In the case of IC%-, the formal charge implication is that the excess charge of the ion is distributed on the chlorine atoms and the iodine atom has a charge near zero. The case of XeF2 contrasts since the central atom must have a significant positive formal charge to balance the negative charge placed on the terminal atoms. This difference can be expected to tend to strengthen the bond somewhat because of the electrostatic attractions, raising f r . 1 4 At the same time, the
Sir: In 1951, Pimentel discussed the bonding in trihalide ions in terms of a simple molecular orbital description.' In this widely ignored paper, the applicability of the bonding scheme to other molecular species was recognized and, in fact, the existence of inert gas-halogen compounds was predicted. "It i s to be expected that a rare gas could f o r m complexes with halogens."' At about the same time, Hach and Rundle2discussed bonding of trihalides in similar terms, though without any . reference to possible inert gas compounds. Since then there has appeared significant support for this molecular orbital description of the trihalides through electric quadrupole resonance studiesa. and infrared studies6 The quadrupole coupling constants in both IC12- and Ic14- confirm the proposal that d orbitals do not contribute significantly to the bonding. The recent preparation of inert gas compounds has (7) L. C. Allen a n d naturally stimulated much interest in their b ~ n d i n g . ~ - ~(19fi2).
W. d e w . Horrocks, J r . , J. A m . Chem. S O L . , 84, 4344
(8) K. S. Pitzer, Srieirce, 139, 4 1 1 (1963). (1) G . C . Pimentel, J. Chem. Phys., 19, 446 (1951). (2) R . 5. Hach a n d R. E. R u n d l e , J. Am. Chem. Soc., '73, 4321 (1951). (3) C. D . Cornwell a n d R . S . Yamasaki, J. Chem. Phys., 2'7, 1060 (1957). (4) R . S. Yamasaki a n d C. D. Cornwell, ibid., 30, 1265 (1959). ( 5 ) W. B. Person, G. R. Anderson a n d J. D. Fordemwalt, ibid., 36, 908 (1961). (0) L. C. Allen, Science, 138, 892 (1962).
(9) R. E. R u n d l e , J . A m . Chem. Soc., 86, 112 (1963). (IO) G. C. Pimentel a n d A. L. McClellan, "The Hydrogen Bond," W. H. Freeman a n d Co., l Q G 0 , p. 343. (11) W . B. Person, R . E. Humphrey and A. I. Popov, J . A m . Chem. S O L . , 82, 29 (1960). (12) I,. H. Jones and R . A. Penneman, J. Chem. P h y s . , 2 2 , 781 (1954). (13) D. F. Smith, ibid., 38, 270 (1963).
