Direct Fluorination of Steroidal Olefins to cis-Vicinal Difluorides

atoms are located at the corners of a regular tetrahedron. From both two-dimensional Fourier maps and stereo- chemical considerations, Corradini11 ded...
0 downloads 0 Views 267KB Size
1822

b Fe,(GO),, Figure 1. Molecular configuration of Fe8(CO)12.

CO,(CO),z

Figure 2. Molecular configuration of CO~(CO)~~.

triangle are in excellent agreement with the correspondThe idealized molecular configuration of CaVsyming Fe-Fe distances in HFe3(CO)ll- (Eiz., 2.685, 2.696, metry (Figure 2) consists of an apical CO(CO)~group and 2.511 A with individual esd's of 0.003 A). )~ A two-dimensional X-ray diffraction study of Cod- coordinated by Co-Co bonds to a C O ~ ( C Ofragment containing three identical CO(CO)~groups situated (CO),z by Corradini" showed that the four cobalt at the corners of an equilateral triangle with the cobalt atoms are located at the corners of a regular tetrahedron. atoms linked to one another by both symmetrical bridgFrom both two-dimensional Fourier maps and stereoing carbonyl groups and Co-Co bonds. The twelve chemical considerations, Corradini deduced an icosacarbonyl groups are arranged in an icosahedral array hedral molecular arrangement for the twelve carbonyl about the tetrahedron of cobalt atoms. This molecular groups. However, the lack of resolution in the indiconfiguration is in agreement with that proposed by vidual carbonyl groups on the Fourier projections Corradini. The six independent Co-Co bond lengths coupled with a crystalline disorder did not make his geoare equivalent within experimental error and have an metrical assignment of the carbonyl groups unambiguaverage value of 2.49 A (individual esd, 0.02 A). ous. Since Corradini" did not specify any statistical' A salient feature is that the approximate molecular details, the reliability of his limited structural analysis configuration of Fe3(C0)1z can be formally derived could not be assessed. Furthermore, other possible alfrom that of C O ~ ( C Oby ) ~ the ~ abstraction of a basal ternative molecular models involving different arrangemetal atom from the tetrahedron of cobalt atoms. The ments of carbonyl groups mbsepently were proposed relative disposition of the other three metal atoms with for C O ~ ( C O1 2), ~ ~ to the icosahedron of carbonyl groups remains The reinvestigation of the structure of C O ~ ( C O ) ~respect ~ essentially invariant except that this more regular polywas prompted by our three-dimensional X-ray studies of Rh4(C0)1214and Irl(CO)1215which showed Ir4(CO)12 hedron in Co4(CO)1zcollapses to some extent in Fe3(CO)lz in order that the twelve carbonyl ligands be to be different from Rh,(C0)I2 and Co4(CO)1z. The tetracoordinated to only three iron atoms. meric configuration of Ir4(CO)12was found to possess approximately Td cubic symmetry and to be stabilized Acknowledgments. We are pleased to acknowledge by only iridium-iridium bonds. Although the lattice the financial support of the National Science Foundaparameter data of Rh4(C0)12 are similar to those of tion. The use of the CDC computers at the WisCo4(CO)12,attempts to determine the complete strucconsin Computing Center was made possible through ture of Rh4(C0)1zwere not successful due to a crystal the partial support of NSF and WARF through the disorder presumably of a different nature from that of University of Wisconsin Research Committee. We Co4(CO)12as indicated by the different observed space also wish to dedicate the structure of Fe3(C0)1z in group symmetries for the two compound^.'^ Howmemory of the late Professor Robert E. Rundle of ever, Beck and Lottes16 have shown that Rh4(C0)1Zand Iowa State University for his interest and encourageCo4(CO)12must have the same over-all molecular strucment in this work. ture from the resemblance of their infrared spectra. Chin Hsuan Wei, Lawrence F. Dahl Orthorhombic crystals of Co4(CO)12 were prepared, Department of Chemistry, Unicersity of Wisconsin and the crystal data were found to be in agreement with Madison, Wisconsin 53706 those of Corradini. Three-dimensional photographic Receiced January 19, 1966 intensity data consisting of 529 independent reflections were taken with Mo K a radiation. After considerable difficulty the entire structure was solved and refined to Direct Fluorination of Steroidal Olefins to R1and RZvalues of 12.6%. cis-Vicinal Difluorides P. Corradini, J . Chem. Phys., 31, 1676 (1959). U. Kriierke and W. Hiibel, Chem. Ber., 94, 2829 (1961). D. L. Smith, J . Chem, Phys., 42, 1460 (1965). C. H. Wei, G. R. Wilkes, and L. F. Dahl, unpublished research. G. R. Wilkes, Ph.D. Thesis, University of Wisconsin, Madison, Wis., 1965; G. R. Wilkes and L. F. Dahl, to be published. (16) W. Beck and K. Lottes, Chem. Ber., 94, 2578 (1961). (11) (12) (13) (14) (15)

Journal of the American Chemical Society / 88:8 / April 20, 1966

Sir : It has been shown that elemental fluorine will add smoothly to olefins. The substrate was diluted with an (1) R. F. Merritt and F. A. Johnson, J . Org. Chem., in press.

1823 position 4 where there is only one vicinal proton. inert solvent and the addition was conducted at low The vicinal H F coupling constants could not be obtemperatures (usually - 78 "). This procedure has tained since the appropriate region was obscured by the been successful for conjugated olefins such as indene,' acenaphthylene, stilbene, and substituted phenanabsorption of the proton geminal to the chlorine atom. The stereochemistry is assigned cis by analogy with the threnes,2 as well as for a,P-unsaturated acid halides. cholestenone case and the other examples. cis addition was the predominant process with stilbene, acenaphthylene, and the phenanthrenes. 1 , 2 The only Acknowledgment. This work was sponsored under other alternate procedures utilize PbFd3 and perhaps Army Ordnance Contract DA-01-021 AMC-11536 (Z). C6H5IFZ4 and are neither as general nor as simple as one We are grateful to Mrs. Inella G. Shepard and Mr. would prefer. Morris Howard for technical assistance. Direct fluorination of A4-cholesten-3-one proceeds (10) The cis addition to cholesteryl chloride also is supported by the smoothly in CC13F at -78" to produce the cis-4,5following observations. Fluorination of cholesteryl acetate gives a low yield (10-20%) of 5a,6a-difluorocholestan-3P-ol acetate, mp 117-1 18",11 difluoride in yields of 60-70x. The crude adduct F19nmr peaks at + +178.9 and +194 (doublet, geminal JHFS 46 mixture was purified from unreacted cholestenone by cps). This acetate was converted by the method of Barnes and Djerassi" silica gel chromatography followed by crystallization mp 171-172",11 F19 nmr peaks at 4 to 5o,6a-difluorocholestan-3-one, + 174.9and + 196 (doublet, geminalJEF S 46 cps). Direct fluorination from methanol and affords the adduct as colorless has not proceeded satisfactorily. crystals, mp 187-188". Anal. Calcd for CZ7HA4F20: of (1A6-cholesten-3-one 1) C.S. Barnes and C. Djerassi, J . Am. Chem. Soc., 84, 1962 (1962), C, 76.73; H , 10.49; F, 8.99. Found: C, 76.54; H, report mp120-121 for the 5a,6a-difluorocholesteryl acetate and mp 173-1 74" for 5a,6a-difluorocholestan-3-one. 10.76; F, 9.40. The infrared spectrum of the difluoride contained a Richard F. Merritt, Travis E. Stevens carbonyl absorption at 5.72 p indicative5 of an aRohm and Haas Company, Redstone Research Laboratories Huntsville , Alabama fluorine substituent. The usual C-F absorption was noted between 8.7 and 9.6 p. Dehydrofluorination of Received December 17, I965 the adduct (sodium methoxide in methanol) gave 4fluoro-4-cholesten-3-one, mp 102-103 0.6 2,CDinitrophenyl Phosphate The F19 nmr spectrum contained two multiplets centered at 4 170.1 and 207.4' in an integrated ratio Sir: of 1 : 1. The low-field (4 170.1) multiplet was too The synthesis of 2,4-dinitrophenyl phosphate has complex to interpret and assigned the fluorine atom at eluded the best efforts of a series of investigators. This C-5. The higher field group (CY to the carbonyl) was a has led some of them to suppose that the ester must pair of doublets with couplings of 47 and 12 cps. be exceptionally labile, a point of some interest in conThe 47 cps coupling is indicative* of a geminal J H F nection with its possible role in the uncoupling action of and the minor (12 cps) value is assigned as JFF. dinitrophenol in oxidative phosphorylation. The proton nmr spectrum contains two doublets We have prepared 2,4-dinitrophenyl phosphate in centered at 6 5.0 which are assigned the single C4 near-quantitative yield by the debenzylation of the diproton geminal to a fluorine atom. The geminalJHF benzyl ester. This is readily prepared from the phenol of 47.0 cps is readily apparent. The JHFvicinal then and dibenzyl phosphorochloridate in dry ether, in the remains with a value of 32.6 cps. This large J H F is presence of 1 mole of 2,6-lutidine. After refluxing for sufficient to assign the stereochemistry of the adduct as 1.5 hr the solution is filtered hot to remove 2,6-lutidine cis.g It has been showng that J H F (axial-axial) = hydrochloride, and the ester crystallizes on cooling. 23.4-25.4 cps and JHF (equatorial-axial) = 4.9-1 1.7 cps Recrystallization from ether gives colorless crystals, mp in the glycopyranosyl fluorides. trans-Diaxial orien65-66". tation of C4-hydrogen and C5-fluorine atoms demands Two grams of the triester is suspended in 50 ml of dry cis orientation of fluorine atoms. ether and dry HBr passed slowly into the solution at The isolated double bond of cholesteryl chloride room temperature. After 1.5 hr HBr is no longer could also be smoothly fluorinated to the 5,6-difluoride, absorbed, and a yellow oil has separated. After remp 103-104". Anal. Calcd for C27H45C1F2:C, 73.19; moval of the solvent, and traces of HBr, in vacuo, the H, 10.24; F, 8.58. Found: C, 73.17; H, 10.55; residue is dissolved in a large volume of dry ether and F, 8.41). The F19 nmr spectrum contained a broad 2,6-lutidine added until the solution just turns yellow band at 4 +179.7 assigned to the fluorine atom at (1.1-1.2 g). A voluminous white precipitate of the C5 and a complex doublet (J = 40 cps) at #I 194.5. mono-2,6-lutidinium salt is formed. After filtration, The 40 cps value is suggestive of geminal H F coupling' and recrystallization from ethanol, this has mp 142" and is assigned as the fluorine atom at Ce. The specdec. The yield is about 90%. Anal. Calcd. for trum is too complex to assign this fluorine atom t o C I ~ H ~ ~ N ~C,O 42.1; ~ P : H, 3.77; N, 11.32; P, 8.35. Found: C, 41.7; H, 4.00; N, 11.49; P, 8.29. (2) R. F. Merritt, unpublished results. One of us has shown previously3 that a plot of the (3) A. Bowers, P. G . Holton, E. Denot, M. C. Loza, and R. Urquiza, J . Am. Chem. Soc., 84, 1050 (1962). logarithms of the rate constants for hydrolysis of sub1,2v10

O

+

(4) P. G. Holton, A. D. Cross, and A. Bowers, Steroids, 2,71 (1963). ( 5 ) L. J. Bellamy, "The Infrared Spectra of Complex Molecules,"

John Wiley and Sons, Inc., New York, N. Y., 1957. (6) S. Nakanishi, R. L. Morgan, and E. V. Jensen, Chem. Znd. (London), 1137 (1960),report mp 100-lOlo,y c - 0 1688 cm-1 for this steroid. Our sample had yc-o Nujol at 1686 cm-1 and an Fl9 nmr peak at 9

+ 140.2.

+

(7) F19nmr data are given in values of (ppm from CClsF as internal standard). (8) J. A. Pople, Mol. Phys., 1, 216 (1958). (9) L. D. Hall and J. F. Manville, Chem. Ind. (London), 991 (1965).

(1) (a) M. Rapp, Ann. Chem., 224, 156 (1884); (b) V. H. Parker, Biochem. J., 69, 306 (1958); (c) R. Wittmann, Chem. Ber., 96, 771 (1963); (d) R. Azerad, D. Gautheron, and M. Vilkas, Bull. SOC.Chim. France, 2078 (1963). (2) (a) W. F. Loomis and F. Lipmann, J . Biol. Chem., 201, 357 (1953); (b) F. Hunter in "Phosphorus Metabolism," Vol. 1, W. D. McElroy and B. Glass, Ed., Johns Hopkins Press, Baltimore, Md., 1951,p 297. (3) A. J. Kirby and W. P. Jencks, J . Am. Chem. Soc., 87, 3209

(1965).

Communications to the Editor