5262
COMMUNICATIONS TO T H E
EDITOR
bon-carbon bond cleavage to ethyl l-methylcyclopropanecarboxylate, which then is reduced to the observed product
Vol. 8 1
violet of the carbonyl maximumg as well as on rotatory dispersion data. lo 2a-Fluoro-17a-ethynyltestosterone (1') (m.p. 65", A,, 242 mp, log E 4.28,::A: DEPARTMENT OF CHEMISTRY KESKETHB. WIBERG 243-245', [ a ] D UNIVERSITYOF WASHISGTON 3.00, 5.90 p. Found for C?1H27F02: C, 76.47; H , SEATTLE 5 , WASHINGTON RICHARD P. CIULA 8.28; F, 5.39) was prepared by condensing 17aRECEIVED AVGUST10,1 9 8 ethynyltestosterone with ethyl formate and treating the crude 2-hydroxymethylene compound as described above. STEROIDS. CXXXII.' 2-FLUOR0 AND Since sodio malonic ester is difluorinated by per2 1,2 1-DIFLUORO STEROIDS chloryl fluoride even in the absence of excess base' Sir: it appeared that 21,2l-difluorination, leading to The fluorination of active methylene compounds a hitherto unknown class of steroids, would be with perchloryl fluoride has been demonstrated feasible. The sodio salt of 21 -ethoxalyl-A5-pregrecently. Subsequently, Gabbard and Jensen3 nen-3P-ol-20-one11 in absolute ethanol or benzene utilized this reagent in the steroid series, preparing and in the presence of excess sodium methoxide 2a-fluorocholestanone from cholestanone pyrrolidyl was treated with perchloryl fluoride and then enamine while Kissman, Small and Weiss4 prepared methanolic potassium acetate yielding 21,212a-fluorohydrocortisone from 2-methoxalylhydro- difluoro-A5-pregnen-3~-ol-20-one (m.p. lX3-135", [.ID 54", 5.75 p. Found for CalH30F202: cortisone 20-ketal. 1Ve have prepared a number of 2a-fluor0 steroids C, 71.05; H , S.88; F, 10.28). Oppenauer oxiin the potentially important androstane series by dation gave 21,21-difluoroprogesterone (1'1) (m.p. 241 mp, log E 4.22, reaction of the sodio salt of the appropriate 2- 140-143", [ a ] D 204', ",::A: 5.75, 6.00 p. Found for C?1H2ilF20?:C, hydroxymethylene-3-keto steroid with perchloryl A",: fluoride followed by alkaline cleavage of the result- 72.54; H, 8.10; F, 10.14). In preliminary seven day assays1?in the castrate ant 2-aldehydo-2-fluoro c o m p o ~ n d s . ~ Thus, 2hydroxymethylene-testosteroneJ6 -androstan-17/3-01- rat, I and I1 exhibited 2OY0 and 50yo of the an60" (all rot. in drogenic potency of testosterone with niyotrophic %one (m.p. 125-130°, [ a ] D CHC13), A:.,"," 282 mp, log E 3.94), -17a-methyl- activity about .inycof the standard while conitestosterone,' and 2-hydroxymethylene-lia- pound 111, orally administered in the same assay methylandrostan-17@-01-3-one,~ in benzene solu- was 25y0 as androgenic as methyltestosterone. tion, were reacted successively with sodium Both I and I1 were potent gonadotrophin inhibimethoxide and perchloryl fluoride. Treatment of tors in a IO-day parabiotic rat assay.'? VI was the reaction products with potassium acetate in considerably less active than progesterone in the boiling methanol gave 2a-fluorotestosterone (I) Clauberg assay1*in sharp contrast to the activity (m.p. 140-141", [ a ] $. ~ 131", A",.,"," 242 m p , of 21-monofluoroprogesterone. l 3 log E 4.15, ;:A: 5.90 p. Found for C19Hn;F02: (9) R . N. Jones, D . -4.Ramsey, F. Herling and K. Dobriner, THIS C, 74.27; H, 8.85; F, 5.97): ?aa-fluoroandrostan- JOURNAL, 74, 2828 (1952); B. Ellis and V. Petrow, J . Chein. .To(., 1171) 17B-ol-3-one8 (11) (m.p. 183-18zc, [ a ] D f G3", (1056). See also ref. 2 and 3. (10) C. Djerassi, I. Fnrnaguera and 0. Mancera, Tiin J O C R V X L . 81, 283 mp, log E 1.44, A;:i 5.75 p. Found for 2383 (1959). C21H31FO3.C3H60: C, 70.96; H, 8.91; F, 4.81); (11) H. Ruschig, Bel., 88, 878 (1953). 11 acetate (m.p. 190-193", [ a ] D 56"); 2a(12) Assays by T h e Endocrine Laboratories, Madison, Wisconsin (13) P. Tannhauser, R. J. P r a t t and E. 1'. Jensen, THISJoURh-.41.. fluoro-17a-methyltestosterone (111) (m.p. 168169", [a]D f 1 1 G 0 , "A,: 242 mp, log E 4.22, 78, 2658 (19.50). 5.90 p. Found for CzoH39F02: C, 75.30; RESEARCH LABORATORIES s. A. J. EDW.4RDS H, 8.63; F, 5.50); 2a-fluoro-17a-methylandros- SYNTEX, H. J. RIXCOLD APDO.POSTAL 2679 tan-17P-ol-3-one (IV) (m.p. 193-194", [ a ] D M f x ~ c o D. , F. 46') AE:H283 nip, log E 1.52,::A: 5 . 7 5 ~ . Found for RECEIVED JULY 20, 1959 C20H31F02:C, 74.49; II, 9.69; F, 5.59). The assignment of the 2a-fluor0 configuration rests on the shift in the infrared and lack of shift in the ultra-
+
+
+
+
+
+
(1) Steroids CXXXI. 1. Zderic and D. Ch6vez Lim6n, THISJ O U R NAL,in press (1959). (2, C. E . Inman, E. -2. Tyczkonski, R. E. Oesterling and F. I,. Scott, Erpei,ientia. 14, 33.5 (1938); C. E. I n m a n , R. E. Oesterling and E. A. Tyczkowski, THISJ O U R N A L , 80. 6533 (1958). (3) R . R. Gnhhard a n d E. V. Jensen, J. O Y ~ Chcirr., . 23, 1406 (1958). (4) H. Kissman. A . 11.Small and M. J. U'eiss, THISJ O V R N A L . 81, 1262 (1939). ( 5 ) A number of t h e fluoro hormone analogs reported in this paper have been prepared by E. V. Jensen and co-workers through alternate routes. Their results are published simultaneously p. 525Y. (6) F. Weisenborn, D. Remy and T. L. Jacobs, THISJ O U R N A L , 76, 552 (1954). ( 7 ) H . J. Ringold, E . Batres, 0. Halpern a n d E . Secoechea, i b i d . , 81, 427 (1959). ( 8 ) Dr. E. V. Jensen kindly compared our product with samples prepared through t h e enamine and enol ether (see ref. 5 ) routes and found them t o be identical.
ISOTOPE
SEPARATION
BY ION EXCHANGE
Sir: The sign, magnitude and trend of the separation factors observed by Lee and Begun' provide another example of the close analogy between ion exchange resins and concentrated aqueous solution, because the observed values are very nearly those which would be expected fsom an equilibrium between a dilute aqueous solution and a concentrated aqueous solution of a molality corresponding to that of the exchangers concerned. For two aqueous solutions, e.g., of LiC1, the sepE is given by the ratio of aration factor k = 1
+
(1) D. A. Lee and G 51. Begun, THISJOZ'RXAL, 81, 2332 (lY59)
COMMUNICATIONS TO THE EDITOR
a c t . 5: 1959 activity coefficients
where the bar marks the stronger solution. Equation l takes account of the fact that the solutions are mixtures of two electrolytes, utilizing Br@nsted's rule. The ratio of the activity coefficients in the dilute solution equals about unity. Hence In (1
+
e) N
E
(In 77
- In 78)
(2)
Glueckauf2has shown that the hydration parameters which largely determine the trends of the activity coefficients are a function of the Pauling radius of the cation. Equally-without making use of the hydration concept-the activity coefficients of the aqueous solutions of alkali salts of the same halogen can be directly expressed as functions of the Pauling radii ( r , in hgstrom). The activity coefficients of all alkali chlorides a t 25" for m > 1 are adequately expressed by the equation3
5263
the Li-isotopes, this being the square of the inverse ratio of the masses. If this principle applies generally, the outlook for enriching isotopes of higher mass by pure ion exchange is very dim indeed. One would expect for the exchange on Dowex 50 K39-K4'
A/, = 2 X
lWhA&
and
t
=
1.8 X 10 -5
and for Rbs6-Rbs7 & = 5 X 1 0 - - 6 kand e = 3.5 X ATOMIC ESERGYRESEARCH ESTABLISHMEST HARTVELL, ENCLAXD E. GLUECKAUF RECEIVED JULY 13, 1959 T-CYCLOPENTADIENYL-TCYCLOHEPTATRIENYL VANADIUM
Sir :
The chemical literature contains several examples of the preparation of cyclopentadienyl-metal carbonyl compounds by heating together a metal carbonyl and cyclopentadiene or dicyclopentaIn -/ = 0.60;/% + km (3) diene. However, when molybdenum hexacarwhere bony1 or iron pentacarbonyl is heated with cyclok = 0.270 ((l/r) - 0.38) (4) The differences of the Pauling radii for Li6 and heptatriene, there is no loss of hydrogen to form Lij are known from Thewlis' measurement4 of the tropylium compounds. Instead the complexes ) ~ ~produced. lattice constants of LiaF and Li'F. These data CjHsMo(C0)32 and C ; K B F ~ ( C Oare The only tropylium sandwich compound so far give reported is the ionic compound [ C ~ H ~ M O ( C O ) ~ ] B F ~ , ( T ) L ] ~- ( I ) L , ~= 0.0004 zt 0.00007 ingstrom (5) prepared by a different method. From equations 2, 3, 4, 5 follows Using a different approach to tropylium-sand0.270 % wich chemistry, the reaction between cyclopentam X A r = X r7) = 3 x 10-4% (r~i)* dienyl vanadium tetracarbony15 and cycloheptatriene was investigated. A mixture of CsH5V(C0)d (0.01 mole) and 30 ml. of commercial cycloTABLE I heptatriene was refluxed under nitrogen a t 120' Approx. 103 calculated for for 9 hours. After removal of excess cycloheptaexchanger I O 3 observed LiCl solutions (E) % DVB molality (Z) (ref. 1) from eqn. 6 triene a black residue remained. Sublimation a t 80" yielded oily material which was discarded. 4 3 1.0 0.9 Further sublimation a t 100" afforded 0.82 g. 8 5.5 1.6 1.7 (yield 40%) of purple crystals, C5HjVC7H7, spar12 7.5 2.7 2.3 ingly soluble in organic solvents to give purple 16 9 3.7 2.7 solutions which in air begin to deposit a brown 24 12.5 3.8 3.8 precipitate after a few minutes. Table I shows the agreement between the sepaAnal. Calcd. for C12H13V: C, 69.6; H, 5.8; ration factors E observed,' when the isotopes were separated on various exchanger types of different V, 24.6; mol. wt., 207. Found: C, 69.7, 69.8; crosslinking (divinyl benzene content) by elution H, 5.8, 5.8; V, 24.6; mol. wt., 229 (cryoscopic in with dilute hydrochloric acid solution, and those benzene). The analytical results are in less accord with a calculated for the equilibria between a dilute and a concentrated lithium chloride solution of the composition C5H5VCjHs (C, 69.2; H , 6.25) than same molal concentration as the exchanger. for C5HbVCjHj. However, a clearer demonstraEquation 6 can be used also for the reverse tion of the nature of the new vanadium compound process of calculating the differences in ionic is provided by its magnetic properties. It is radius from observed isotopic separation factors, paramagnetic with X2sp,"I", = 1055 X cm3e.g., for the case of the isotopes NaZ2and Na24.5 mole-l, corresponding to 1.69 B.M., which is On Dowex 50 (Z 6) this gave E = 1.4 X 10-4 equivalent t o one unpaired electron. This cona t 25". Hence from eq. 6 firms the composition as CjHjVCjH; rather than L
-
8 X 10-6 ingstrom
The radii of two adjoining h'a-isotopes would thus differ by only a tenth of the difference between (2) E. Glueckauf, T r a m . Favoday SOL.,SI, 1235 (1955). (3) J. C. Gosh, J. Chem. S o c . , 113, 449 (1918). (4) J. Thewlis, Acta. Crysl., 8, 36 (1955). (5) R. H. Betts. W. E. Harris and M. D. Stevenson, Con. J . of Chem., 34, 65 (1956).
( 1 ) F o r a recent review of compounds wherein an aromatic ring system a n d carbon monoxide groups are simultaneously bonded t o metal atoms, see E. 0. Fischer and H. P. Fritz, "Advances in Inorganic and Radiochemistry" (Eds. H. J. Emeleus and A. G. Sharpe) 1, 55 (1959),Academic Press, Inc., Kew York, N. Y. (2) E. W. Abel, M. A . Bennett, R. Burton and G. Wilkinson, J . Chcm. Soc., 4559 (1058).
(3) R.Burton, M. L. H. Green, E. W. Abel and G. Wilkinson, Chcm. I d . , 1892 (1958). (4) H. J. Dauben a n d L. P. Honnen, THISJOCRNAL, 80, 5570 (1958). ( 5 ) E. 0. Fischer and S . Vigoureux, B e y . , 91, 2205 (1958).
