953

Table I. Average Molybdenum- Distances for the pi autodiffractometer using Nb-filtered Mo 6 radiation and 8-28 scans. cu-M0sO26~-ion in ~~-[(~-C~H~)(C~H~),P]~MO~~~~.CH~CN.H~~The four molyWenum atoms of the asymmetric unit were located using direct methods (MULTAN) and the remaining atoms by standard difference Fourier techniques. The resulting structural parameters have been refined TypeU Distance, A to convergence (R= 0.031 for 461 1 independent reflections having 28-n < 43' (the equivalent of 0.50 limiting Cu KE sphere) and I > 38(1)) using 1.696 (3,2, 3) unit-weighted fulknatrix least-squares techniques with anisotropic tfwmal parameters for all nonhydrogen atoms and isotropic thermal parameters 1.904 (3,4,9) for all hydrogen atoms. Refinement is continuing with those reflections 2.425 (3, 19, 56) having 28~< 59'. 1.783 (3, 3, 5) (6) F. A. Schroeder, Acta Crystallogr.. Sect. 6, 31, 2294 (1975). (7) In ref 2a an analogous but far geater variation is noted. In view of the lower 1.708 (3) precision obtained in the study reported in ref 2a, it is difficult to assess the significance of this greater variation. (1 See Figure la for labeling scheme. The first number in paren- (8) K. Matsumoto, A. Kobayashi, and Y. Sasaki, Bull. Chem. SOC.Jpfl., 48, theses following an average value is the root mean squared value of 1009 (1975). (9) KihlborglO in his study of Moos was apparently the first to note that certain the estimated standard deviation for an individual datum. The second molybdate structures containing octahedrally coordinated molybdenum and third numbers, when given, are the mean and maximum deviations with marked offcenter displacement toward an octahedral edge may be from the averaged value, respectively. best viewed as containing tetrahedral units. In the case of MoO3, he de- scribed infinite drains of comer sharing, distorted Moo4 tetrahedra which are only weakly associated to yield six-coordinate molybdenums. formula (M004~-)-)z(M06018)which characterizes the weak (IO) L. Kihlborg. Ark. Kemi, 21, 357 (1963). interactions between the two Mo042- ions and a ring of six (11) J. Berkowitz, M. G. Inghram, and W. A. Chupka, J. Chem. phys., 26,842 distorted Moo4 tetrahedra sharing This represen- (1957). corner^.^ (12) We refer here to the pioneering work of Baker and Tsigdinos on oxygen tation is also in accord with the existence of stable rings and chromium exchange in (Cr(OH)s3-)(Mo6018):L. C. W. Baker in "Ad- (Moo3)*, n = 3,4, and 5, studied in the vapor phase by Ber- vances in the Chemistry of Coordination Compounds", S. Kirchner, Ed., MacMillan. New York, N.Y., 1961, and references therein. kowitz et al. (13) M. Filowitz and W. G. Klemperer, J. Chem. Soc., Chem. Commun., 233 The simple relationship between kinetic lability and low (1976). (14) R. Strandberg, Acta Chem. Scad., 27, 1004 (1973). bond order established here is consistent with previous ex- (15) J. K. Stalick and C. 0. Quicksall, lnorg. Chem., 15, 1577 (1976). change studies12and may be extended in an attempt to predict (16) K. Barkigia, L. M. Rajkovic, M. T. Pope, and C. 0. Quicksall, J. Am. Chem. the kinetic behavior of other polyaxomolybdates. Several Soc., 97,4146 (1975). (17) For approximate bond distances in P-MO~O~~~-,see ref 3a. heteropolyanions may be viewed as structurally related to the (18) Camille and Henry Dreyfus Teacher-Scholar. CY-MO~O~~~-ion and represented as (AsO~~-)~(MO&I~),'~ (CsHsAs032-)2(Mo60~8),'3(P043-)2(Mos0~*),14 V. W. Day,* M. F. Fredricb ( CH3P032-)2( Mo501 =,),I5 and ((CH3)2As02-) (OH-)- Department of Chemistry, University of Nebraska (Mo4012),l6 which in turn implies potential lability with re- Lincoln, Nebraska 68508 spect to dissociation of hod3-, C6H~As03~-,FQ3-, CH3P032-, and (CH3)2As02-, respectively. One may spec- W. G. Klemperer,* W.Shum ulate further into the more general case where a large po- Department of Chemistry, Columbia University lyoxomolybdate cluster may be dissected into smaller clusters New York, New York 10027 by breaking only weak bonds. For example, @-Mo~O~~~-may Received October 18, 1976 be accurately represented by ((02-)(Mo4012)2 since the 8- Mo~O~~~-cluster can be dissected into two (02-)(Mo4012) subunits by breaking only bonds whose lengths exceed 2.22 AI7 and hence have bond orders less than 0.2.6 This fact points Intercalate of Tetrafluoride with Graphite toward a mechanistic pathway for the reaction of ((02-)- (Mo4012)2 with (CH&As02H to form the ((CH&As02-)- Sir: (OH-)(Mo4012) ion mentioned above. We are currently at- The compounds ' and xenon oxide tet- tempting to verify some of these speculations using dynamic rafluoride2 have previously been shown to intercalate in I7O NMR techniques. graphite. While these compounds react directly with graphite, Acknowledgments. V.W.D. and W.G.K. acknowledge the reacts only in the presence of hydrogen flu- donors of the Petroleum Research Fund, administered by the oride to yield an intercalate of variable compo~ition.~We have American Chemical Society, for partial support of this re- now found that xenon tetrafluoride also forms intercalates with search. Also, W.G.K. acknowledges the National Science graphite. Foundation for financial support and V. W.D. acknowledges Weighed quantities of graphite and excess XeF4 were al- a generous grant of computing time from the University of lowed to react in preweighed Kel-F reaction vessels for periods Nebraska Computing center. W.G.K. and W.S. are grateful of up to 3 weeks at room temperature. Reactions were gener- to Mr. I. Miura for obtaining NMR spectra. ally complete after about 10 days, their slowness probably due to the low vapor pressure of XeF4. The excess XeF4 was then References and Notes pumped off until the reactor attained constant weight or did (1) W. G. Klemperer and W. Shum, J. Am. Chem. Soc.,98,8291 (1976). not lose weight at rates exceeding 2 mg/h. Stoichiometries (2) For the structure of (r-M08O~6~-see: (a) J. Fuchs and H. Hartl, Angsw. were calculated on the assumption that total weight gains were Chem.. lnt. Ed. €fig/., 15, 375 (1976); (b) M. F. Fredrich. V. W. Day, W. Shum, and W. G. Klemperer, Am. Cryst. Assoc. Summer Meeting, 1976, due to XeF4. The latter were extrapolated back to zero paper M5. pumping time. (3) For the structure of P-Mo8026'- see (a) L. 0. Atovmyan and 0. N. Krase The graphite used was either BDH powder or GTA grade chka, J. Struct. Chem. (USSR).13,319 (1972); (b) I. Linqvist. Ark. Kemi, 2, 349 (1950). Grafoil from Union Carbide Co. Considerable variations in (4) M. Filowitz, W. G. Klemperer. L. Messerle, and W. Shum, J. Am. Chem. stoichiometry were observed. The results with powder based Soc., 98, 2345 (1976). Although this reference considers complexes containing only octahedrally coordinated molybdenum, assignments of on nine different reactions were C28.3*2.4XeF4, while those resonances to 0, and OC in the present work are made using anal- with Grafoil were C41*1XeF4 based on four experiments. In reasoning. (5) Compound 3 crystallizes in the centrosymmetric "clinic space goup, one case, however, a stoichiometry of C17.8XeF4was obtained CZI (No. 15) with a = 28.096 (6) A, b = 14.313 (2) A, c = 27.1 16 even after pumping for 27 h. Neither xenon nor carbon fluo- (5) 1, P = 121.32 (1)'. and Z= 4 (fmhmits as given above). Diffracted rides were liberated during the course of the reaction, and intensities were measured on a spherical crystal having pr = 0.28 for 12 892 independent reflections having 28- < 59' (the equivalent of samples showed no visible changes. analyses (Table 1.2 limiting Cu KE spheres) on a computer-controlled fourcircle Syntex I) correspond roughly to stoichiometries obtained from weight

Communications to the Editor 954

dled outside the vacuum line, although part of the intercalated fluoride is slowly hydrolyzed, releasing HF. The residue is neither shock-sensitive nor exhibits appreciable oxidizing power. Preliminary investigations show that the intercalate may also be useful as a moderate fluorinating agent in organic chemistry. Experiments with aromatic systems gave the fol- lowing results: With benzene, the products were monofluoro- benzene (26%), p-difluorobenzene (1 l%), and starting mate- rial (63%). Phenanthrene6 yielded mainly 9-fluorophenan- threne (22%) and 9,9,1O-trifluoro-9,1O-dihydrophenanthrene

,W IW rW, , m Iwm, , ,W, (19%). As in the case of XeF2' fluorination of benzene proceeds Ilm*.lu-.c,L, only in the presence of HF catalyst. Figure 1. Thermogravimetric analysis of C26.7XeFd: sample weight, 25.7 mg; heating rate, 4O/min; atmosphere . Dashed line corresponds to References and Notes calculated final weight after complete decomposition. (1) H. Selig, M. Rabinovitz, I. Aganat, C. H. Lin. and L. EM,J. Am. Cbm. Soc., 98, 1601 (1976). (2) H. Selig and 0. Gani, lnorg. Nucl. Chem. Lett., 11, 75 (1975). (3) M. Rabinovitz, H. Selig, I. Agranat, C. H. Lin, and L. Ebert, unpublished ob- servations. Table I. Analysis of C,XeF4 (4) H. D. Frame, Chem. Phys. Lett., 3, 182 (1969). (5) W. Riidorff and U. Hofman, Z. Anorg. Allg. Chem., 238, 1 (1938). mequiv mequiv of I (6) (a)M. Zupan and A. Pollak,

C40XeF4 0.182 0.355 1.95 H. Selig,* M. Rabinovitz, I. Agranat, C. H. Lin C24.3XeF4 0.264 0.525 1.98 16.6 15.2 Institute of Chemistry, The Hebrew University of Jerusalem C26.7XeF4 0.241 0.534 2.21 16.6 14.5 Jerusalem, Israel C32.sXeF4 0.249 0.500 2.00 11.9 12.7 L. Ebert Exxon Corporation Linden, New Jersey 07036 increases. Analyses of oxidizing power of the intercalates Received October 4, 1976 showed only 2 equiv of iodine liberated per mole of intercalate (Table I). This indicates that the XeF4 is intercalated in the form of several species, some bound more tenaciously to the graphite than others. Indeed, preliminary wide line 19Fnuclear magnetic resonance spectra reveal the presence of both XeF4 Conjugate Addition of B-l-Alkynyl-9-borabicyclo[3.3.1]- species (a doublet-singlet centered at +29 ppm with respect nonanes to cu,&Unsaturated Ketones. A Convenient to CFC13 with doublet separation (3600 f 320 Hz) and XeF2 Synthesis of 7,bAcetylenic Ketones species (a singlet centered at +148 ppm) possessing derivative Sir: extremum line widths of 0.1 to 0.15 G).4 Consideration of in- tegrated absorption mode peak heights leads to a ratio of B- l-Alkynyl-9-borabicyclo[3.3.llnonanes (B-1 -alkynyl- F(XeFz)/F(XeF4) of 2.3 to 3.0. No proton resonance could 9-BBN) (l),readily prepared by the reaction of boron tri- be observed. fluoride diethyl etherate with the corresponding lithium methyl X-ray patterns of the intercalates obtained with copper ra- alkynyldialkylborinate,l undergo a smooth 1,4-addition in diation showed the complete absence of the original strong pentane at room temperature to methyl vinyl ketone (MVK) graphite line at d = 3.35 A and the appearance of a new pat- and related ketones capable of adopting a cisoid conformation. tern containing very broad, diffuse lines at approximately 3.80 Hydrolysis of the initially formed enol borinate intermediates and 3.24 A for C28XeF4 and at 3.65 and 3.28 A for C40XeF4. (2) provide the corresponding 4-alkynyl-2-butanones (3) in If one were to assume the 3.8 A diffraction line to be [003] high yields (eq 1). reflection of a second stage compound of I, = 11.4 A and the A 3.65 A line to be the [004] reflection of a third stage compound of I, = 14.6 A, as is the case in a number of graphite acid salts,5 one would predict the ratio of carbon atoms to xenon atoms in a single layer to be 13-14/1, a result consistent with the pas- sage of second stage C28XeF4 to third stage C40XeF4. 1 Thermogravimetric analyses were run on a Mettler ther- moanalyzer (Figure 1). The sample begins to lose weight im- IUI.-CCH2CH&BCH3 mediately, because C26XeF4 is unstable at room temperature. The first weight loss step is complete at around 150 OC and 3 corresponds to conversion to a higher stage compound of The use of alkynyl copper reagents to effect the 1,Caddition stoichiometry C4&eF4, thus confirming the preparative results of an alkynyl group to an a$-unsaturated ketone is precluded with Grafoil. The third stage compound decomposes rapidly by the tenacity with which copper binds alkynyl ligands.2 One above 300 OC, and decomposition is complete above 600 OC. possible solution to this difficulty is the temporary transfor- No xenon fluorides were liberated upon heating, xenon gas and mation of the acetylene to a vinylstannane derivative. Thus, carbon tetrafluoride being the only gases detected. After de- Corey has reported that the cuprate derived from trans-1,2- composition, the residue again shows the characteristic bis(tri-n-butylstanny1)ethylene reacts as a nucleophilic ethynyl graphite line at 3.35 A. group equivalent in the conjugate addition to a,B-unsaturated The graphite-xenon tetrafluoride intercalate is easily han- ket~nes.~Unfortunately, the generality of this reaction has not

Journal of the American Chemical Society / 99:3 / February 2, 1977