2856
COMMUNICATIONS TO THE EDITOR
--Position
TABLE I K4Fe (CN)6-----
KaFe(CN)s---Widthu
Position
Width"
+
a
Vol. XT,
-0.7 f 0.3h 3.0 i 0.2 -11.6 i 0.2' 2.5 0.2 +0.9 i 0.5" +50 i 0.5oc . . . Extrapolated to zero concentration. Relative to X14 in pure CHsCX; 32'.
is known, while the ferrocyanide is diamagnetic. T h e measurements of the chemical shifts might lead to evaluation of the spin density distribution in the paramagnetic ion and from the line widths the mechanism and kinetics of the,electron exchange between the para- and diamagnetic species can be determined. The position and width of NI4 resonance in solutions of K3Fe(CN)e were previously measured by Shulman2 and his results are in fair agreement with ours. Shulman attributes the paramagnetic shift t o anisotropic hyperfine interaction. The N L 4and C13 line positions and widths of approximately 0.5 31 aqueous solutions of the pure compounds are given in Table I. Results are expressed in gauss, and measured in a magnetic field ( H o ) of about 14 kgauss. -4 bulk susceptibility correction was not applied since calculation shows it to be rather small (less than the line width). Shifts to a higher magnetic field are indicated as positive. The position of the N I 4 and CI3 resonances in the paramagnetic species is a linear function of the reciprocal of the absolute temperature and the slope is given in the last column of Table 1. The results in Table I indicate: (1) T h e line width of N14 in K3Fe(CN)e is smaller than in K 4 F e ( C x ) 6 , which means t h a t i t is completely determined by quadrupolar relaxation, since, otherwise, one would expect additional broadening due to the unpaired electron in K3Fe(CN)6. The electrical field gradient a t the N nucleus seems to be larger in K4Fe(CN)6,which might be attributed either to its stronger complexation with counter-ions,3which affects the symmetry of the ion, or to the higher total charge of the anion. ( 2 ) Relative to line positions in the diamagnetic species, NI4 resonance in K3Fe(CN)e is shifted to lower field while the C13 resonance is shifted, considerably more, to higher field. Calculation of the hyperfine coupling constants A i from the contact shifts4 gives AS)^ = 0.78 gauss and ilcll = - 11.8 gauss (taking g = 2 and S = l,/J. Assuming equal signs and similar magnitudes of the Q-values in the expression --1i = Qpi, we deduct that the spin density of the CI3 would be approximately 15 times higher than on the N14 with opposite signs. T h e linear dependence of C13 and NI4 resonances on the reciprocal absolute temperature shows that the shift is mainly due to the unpaired electron (paramagnetic shift). In aqueous solutions of mixtures of K4Fe(CN)6 and K3Fe(Cr\j)e.one N14 resonance is observed; hence, the electron exchange rate is fast compared to the reciprocal of the chemical shift. The line width is dependent on the concentration of the ferri- and ferrocyanide. At constant total concentration maximum width is obtained in approximately 1:1 mixtures. T h e position of the common line is a linear function of the mole fraction of the ferri- and ferro- in the mixture. The rate of electron exchange was evaluated from the measured line width using the fast exchange approximation for an unequal doublet with different "natural" T,values.5 Assuming the reaction to be described b y : I
( 1 ) F W G r a y and W M Birse, J . C h e m Soc., 106, 2707 (1914). ( 2 ) R G Shulman, J Chem P h y s , 2 9 , 94.5 (1988) 13) S . Sutin, Anjz Reo .\‘uti Sci , 12, 299 (1962) (4) I> R . Eaton. A D Josey, W' D . Phillips and K E Benson, Discussions F a r a d a y S o c , 34, 77 (1962) ( 5 ) J A Pople, W G . Schneider, and H J Bernstein, "High Resolution Nuclear Slagnetic Resonance," hlcGraw-Hill Book C o , 19.59, Chapter I O .
Chemical shift
Temperature dependence KsFe(CN)a (10-80O)
10.9 f 0.2 + 3 . 3 X l o 3 gauss deg. 49.2 i 0.4 -13.7 X lo3 gauss deg. Relative to C13 in pure C6H6; 32'.
rate = ~[F~(CN)G-~]]F~(CN)~-~], the value of k was found t o be (6.0 f 0.7) X 104set.-' .If-' a t 3 2 O . The range of ferri- and ferrocyanide concentrations used was 0.3-0.6 M ; the error given is the standard deviation from 14 measurements. N o analogous measurements with C13 line width were obtained because of the low S / N ratio. The rate constant obtained is in fair agreement with that obtained b y M'ahl and co-workers,3,6 who studied the electron exchange by isotope labeling techniques. The advantage of n.m.r. in this study is obvious, in view of the elimination of the uncertainty in the "zero-time exchange" occurring with the isotope labeling technique,6 while the disadvantage lies in the high concentrations t h a t must be used. The catalytic effects of various alkali positive ions on the exchange and the temperature dependence of the reaction are presently under study. Measurements were performed on a Varian DPBO spectrometer equipped with a V4311 1.3-Mc. fixed frequency unit for NI4 measurements and a V321OA variable frequency unit operating a t 15.0 Mc. for C'" measurements. Samples were contained in 13-mm. o.d. tubes. In temperature dependence studies h' mm. o.d. tubes were used in a specially constructed dewar insert. For some C13 measurements K4Fe(CN)eand K3Fe(CN)6were prepared from CI3-enriched7 K C N by a procedure siniilar to that given by Thompson.' (6) A . C Wahl, Z Elekivochem , 6 4 , 90 (1960). ( 7 ) T h e authors wish t o thank Professor G Stein of t h e Hebrew University, Jerusalem, for kindly providing K C S enriched t o 1 5 a t o m 70 C ' ? ( 8 ) K . C Thompson, J A m Chem Soc , T O , 1045 (1948)
DEPARTMEKT O F CHEMISTRY A . 1.OEWENSTEIS TECHSIOS-ISRAEL INSTITUTE O F TECHSOLOGY M. SHPORER HAIFA,ISRAEL G. SAVON RECEIVED J U L Y 8, 1963
Long-Range Spin-Spin Coupling of Protons in Isomeric 6,7-Dichloroestrone Methyl Ethers
Sir: Spin-spin coupling of protons, attached to carbon atoms two or more bonds apart, has been observed in a number of systems containing unsaturation.' The success of theoretical treatments of such unsaturated systems2 is evident from the verification of some of their predictions b y experiments.' T h e theory of spin-spin splitting does not provide a reliable basis for the prediction of long-range spin-spin coupling in saturated systems. We have observed a strong spin-spin coupling between protons attached to carbon atoms two u-bonds (1) For leading references, see (a) E I . Snyder and J . I ) Roberts. J A n t . C h e m SOC.,84, 1.582 (lYOZ), (b) I) J Collins, J . J . Hobbs, and S Sternhell, Telrahedron Letters, 197 11963); (c) T A. Wittstruck, S. K hlalhutr:i and H . J. Ringold, J . A m . Chem. S o c , 86, 169!4 ( l Y K 3 ) ; (d) J. R Holm?. a n d I ) Kivelson, ibid , 83, 2959 (1961); ( e ) J. A. Elvidge and R G . Foster. J Chein Soc.. ,590 (1963), (f)E. W Garbisch in Abstracts of Papers presented at the 145th National Sleeting of the American Chemical Society, New York. N , Y , Sept , 19f33, p . 3 6 Q ; r e f . 7f and earlier references cited therein. (2) (a) M .Karplus, J C h e m P h y s . . 33, 1812 ( l Q ( i 0 ) ;(b) R . A Hoffman, MOL.P h y s . , 1, 326 119.58), and subsequent papers, seeref l a ; ( c ) M Barfield and D . M . G r a n t , J A m . Chem. SOC.,86, 1899 (19ri3). (3) ( a ) See ref l a , footnote 6 ; ( h ) D. R . Ilavies a n d J I ) Roberts, J . A m . Chcm. Soc., 84, 22.52 ( 1 9 6 2 ) ; I C ) F A I, Anet, ibid , 84, 747 (1Wi2). ( d ) K B Wiberg, B R Lowry. and R . G . S i s t , i b i i l , 84, 1394 l l 9 l i 2 ) ; le) J hfeinwald and Y . C hleinwald, i b i d . , 86, 2514 (19li3), and earlier references cited therein; i f ) J . I hlusher and E J . Corey. 7 e t r a h e d m u 18. 791 !1002), have developed the theory of "virtual' long range spin-spin coupling. which does not apply t o the findings reported b y u s . see f w t n r , t e 0
COMMUNICATIONS TO THE EDITOR
Sept. 20, 1963
2857
TABLE I N.M.R.SPECTRA OF 4-H, 6-H, 7-H, A N D 8-H Ia
6P-H
7p-H 4-H
OF
Ia
AxD
Ib
1 JHHI (c.P.s.)
7
4.46" 5.43" 2 82;0 2 . g b
Jeg, i g Jsg, 8 g
Jsp. 4 Jig,
Ib 6a-H 7p-H 4-H
8g
4.0a 1. g b G0.6' Id 2.2' Go.5E
4.72" J6al 7 g 5.44a Jca, 1 Jip, 8 g = I p 3 14;a 3.2' Determined by spin decoupling a Measured a t 60 Mc./sec. Estimated ( a t 100 Mc./sec.) of 4-H from 6-H, curve B, Fig. 1. from pattern width ( a t 60 Mc./sec.) of 6-H and values of J6g,7g and J F , ~ , ~ Confirmed @. by decoupling-see curves C and D, Fig. 1. Estimated from pattern width ( a t 60 Mc./sec.) of 7-H and value of J&,p. e Estimated from line width ( a t 60 Mc./sec.) by spin decoupling; see of 6-H and value of J ~ ~ , 7 gConfirmed . Determined by spin decoupling ( a t curves F and G, Fig. 1. 100 Mc./sec.) of 4-H from 6 - H ; curve G, Fig. 1. 0 Estimated from pattern width ( a t 60 Mc./sec.) of 7-H and value of Jea.ig. Confirmed by spin decoupling; see curves F and H , Fig. 1.
'
apart in a rigid system, wherein the protons are in close proximity. The two isomeric methyl ethers, of 6 a ,7a-dichloroestrone (Ia) and of 6P,7a-dichloroestrone ( I b ) , 4 gave n.m.r. spectra, Table I, which
x
C1 Ia, RI= C1, R2 = H C1 b, Rn = H , R2
could be completely analyzed by means of multiple resonance spin-spin d e c o ~ p l i n gFig. , ~ 1.6 The coupling constants of the vicinal protons 6-H, 7-H, and 8-H were in the approximate range for gauche protons in a rigid system in the half-chair conformation of ring B , 7 %Teak long-range couplings were observed between the aromatic (4-H) protons and the benzylic quasi-axial (6P-H) proton in Ia, and the corresponding quasiequatorial (6a-H) proton in Ib. The most dramatic feature of the observed spectra was the unmistakable presence of the long-range coupling, J = 1.8 c.P.s., between the quasi-axial (6P-H) and axial (8P-H) protons of Ia, and the absence of a corresponding observ( 4 ) Compound I a , m . p . 164O, " : :X 282 (ZZOO), sh. 288 mp ( e 2000); 289 mrr ( e 24001, were isolated a s by-products of t h e a n d I b , m.p. 143', oxofluorination ( M . S e e m a n a n d Y. Osawa, J . A m . Chem. Soc , 85, 232 (1963)) of 6-dehydroestrone methyl ether. Correct analytical values were obtained for both compounds. T h e details of their synthesis a n d structure proof will be reported in a forthcoming paper. (5) ( a ) R . Freeman a n d D H. Whiffen, M o l . P h r s . , 4 , 321 (1961). (b) We are grateful t o D r . E . A. Pier of Varian Associates, Palo Alto, California, f o r performing t h e spin decoupling experiments. (c) T h e first-order analysis is significant because Y I - vj J l j for all protonsin t h e X A B C system, (6) T h e coupling constants reported represent absolute values of J . Their relative signs are as yet unknown. (7) C i . ( a ) M. Karplus, J . Chem. P h r s . , 30, 1 1 (1959); (b) K . L. Williamson a n d W . S Johnson, J . A m . Chcm. SOL.,8 8 , 4 6 2 3 (19611, a n d F. J . Schmitz a n d W . S.Johnson, Tetrahedron Leltcrs, 6 4 7 (1962); ( c ) J. W. Clark-Lewis a n d I,. M . Jackman, P r o c . Chem. Soc., 166 (1961); ( d ) A. D. Cross, J . A m . Chcm. S o c . . 84, 3206 (1962); (e) E D. Garbisch, J . Org. Chem.. a?, 4249 (19621, estimated coupling constants in cyclohexenes of unknown degree of ring deformation. T h e effect of ring size and conformations of cycloalkenes on coupling constants was studied b y ( f ) G. V. Smith a n d H. Kriloff, J . A m . Chem Soc , 86, 2016 (19631, a n d (8) P. Laszlo and P. R . Schleyer, ibid., 86. 2017 (1963). Quantitative correlations of coupling constants and dihedral angles between vicinal protons are n o t of general applicability, but de. pend upon rigidity of t h e system as well a s environmental effects, see ref 7 b ; ( h ) R . U. Lemieux, J. D Stevens, a n d R R. Frazer, C a n . J . Chem., 40, 1955 (19621, also in relation t o t h e findings of F . A. L. Anet, J . A m . Chcm S O L ,84, 10R3 (1962); ti) K . L. Williamson, ibtd., 85, 516 ( l 9 6 3 ) , (j) A. Nickoii, M . A Castle, R: H a r a d a , C E . Berkoff, and R. 0. R'illiams, ibid , 86, 2183 (1963):
I
L A
TI4.46
5.43 7PH
6PH
T, 4.72 6q-H
5.44 76H
rb
Io
I >>
1
Fig. 1.-3.m.r. spectra of 6a,ia-dichloroestrone methyl ether ( I a ) (curves A-E) and of Gp,7n-dichloroestrone methyl ether ( I b ) (curves F-H) in the 4.4-5 5 7 region (6-H and 7-H signals) a t 100 Mc./sec. (CDC13 solution, T M S internal standard)
Curve
A B C
Decoupling in adlation, p p m
Curve
None 1 56 downfield (4-H) 1 54 downfield (4-H) and 2 58 upfield (8-H)
D E F G H
Decoupling irradiation ppm
2 60 upfield (8-H) 1 58 upfield (8-H) none 1 52 downfield (4-H) 2 03 upfield ( 8 - H )
able coupling between the quasi-equatorial ( 6 a - H ) and axial (8P-H) protons of I b These findings define a
COMMUNICATIONS TO THE EDITOR
2S5S
steric requirement for the novel type of long-range spinspin coupling between 1,3-quasi-axial-axial protons.8-1na (8) T h e steric requirement is defined primarily in terms of the juxtaposition of Bp-H a n d 88-H, although t h e juxtaposition of the benzylic OB-H a n d the aromatic system cannot he a p v i o i i disregarded as a n element of the steric requirement T h e weak, a n d oterically unspecific, couplings between both benzylic (GwH and (jp-H) protons with the aromatic (4-H) proton could be regarded a s evidence for the lack of strong r-7 interactions; this situation m a y be contrasted with t h e strong and sterically specific interactions between vinylic ( 4 - H ) a n d ailylii (8P-H) protons, see ref I b and c T h e juxtaposition of RB-H a n d 80-H is akin t o the geometric arrangements, ref. 3d. Fig 2 a n d 3 , favorable for enhanced spin-spin coupling on the basis of overlap integrals of the C-H bond orbital wave functions calculated b y Wiberg, el a i , ref. 3d. Our findings therefore appear t o accommodate t h e predictions from this t r e a t m e n t based on Fermi contact potential In view of environmental effects, other t h a n geometry, which may influence JHH values, ref. 7 , i t is n o t yet warranted t o consider this a conclusive experimental confirmation of the theory. ( Y ) T h e concept of "virtual long range spin-spin coupling, developed by Musher and Carey, ref. 3f, describes a "coupling" between nonvicinal protons, which occurs when, f o r example. protons B and C , in a closed linear 3-spin system, are strongly m o u g h coupled so t h a t proton A can "see" the spin states of C although i t experiences no coupling strength from C , i . e . , J A C = 0 T h e X A B C system described b y us becomes a closed linear ABC system on decoupling of 4 - H , F i g 1 iB) As demonstrated by us, J A C # 0, and this coupling is "real," not "virtual " ( I O ) (a) S u t p o r t e d b y American Cancer Society G r a n t P-286 B ; (b) Post-doctoral Fellow.
KOSKELLPARKMEMORIAL INSTITUTE BUFFALO 3, XEM-YORK RECEIVED J U L Y 9, 1963
\Ye hereby wish to report the development of the first successful electrochemical method of reducing aromatic compounds selectively to either dihydro or tetrahydro products. A simple electrolytic cell shown in Fig. 1 was used.
B ' platinum electrodes; B, asbestos, D. glass seal.
C, condensers,
LYhen aromatic compounds dissolved in methylamine were reduced in this cell wifhout the asbestos divider (?.e., when the anode and cathode compartments were not separated) excellent conversions to dihydro products resulted. a'hen the same aromatic compound
8
Aromatich
Yield,
Alkylcyclohexenes,
IXhydrobenzenes,
Recovered aromatic,
7cc
c/c
7c
%
if
Sir:
divider with
TABLE I ELECTROLYTIC REDUCTIOSS O F VARIOUS -AROMATIC HYDROC A R B O S S I S METHYLAMISE-LITHIUM CHLORIDE S O L U T I O S S TV'ITH A N D WITHOUT CELLDIVIDER"
Benzened 49 4 95 1 Benzenee 49 100 .. ... Toluened 64 5 94 .. Toluene" 44 86 14 Eth ylbenzened 73 4 93 ... Ethylbenzene€ 63 96 . . 4 Cumened 82 6 78 13 Curnene' 75 89 . . 11 t-Butylbenzened 85 5 68 20 t-Butylbenzene' 81 53 ... 47 In those runs where the asbestos divider was used, 0.4 mole of LiCl was employed in each compartment. With n o divider, 0.8 mole of LiCl was used. About 50,000 coulombs was passed through the solution in each case. The amperage varied between about 1.2 t o 2 and the time of reduction was about 7-9 The amount of aromatic in every instance hr. in each case. The yield in each case was calculated so as t o was 0.1 mole. YOSHIOO S A W A ~ ~take ~ into account the amount of aromatic recovered. The M . XEEMAN per cent conversion would hence be lower. Without asbestos cell divider. With asbestos cell divider.
An Electrochemical Method of Reducing Aromatic Compounds Selectively to Dihydro or Tetrahydro Products
Fig 1 -.A,
VOl. 85
0
without divider
( 1 ) A recent publication by H IT' Sternberg, R > l a r k b y , and I \Vender, J Eluctrockt~iii S r i c , 110, 4 2 5 1 I'J63), disclcises the successful electrachemical reduction of benzene and tetralin A mixture of reduction pruducts was obtained and no increased selectivity was noted when an "anode compartment" v a s used
was reduced wzth the asbestos divider in place, equally excellent conversions to tetrahydro products resulted. Thus, benzene can be reduced to either cyclohexene or 1,4-dihydrobenzene b y the simple device of carrying out the electrolytic reduction in the presence or absence of the cell divider. Table I lists the results obtained when a representative series of aromatic hydrocarbons was reduced electrolytically both in the presence and absence of the cell divider. An electrolyte consisting of lithium chloride was used in each case. A11 of the electrolyses were carried out under similar conditions. The preparation of 2,5-dihydroisopropylbenzene and the isopropylcyclohexenes which follows can be considered as typical. An electrolysis cell (Fig. 1) 1TO mm. in length by 100 mm. in diameter was fitted with two Dry Ice condensers, platinum electrodes ( 2 x 5 cm.), and an asbestos divider. Cumene (12 g., 0.1mole) was placed in thecathode compartment. Lithium chloride (1Tg.)0.4 mole) and 450 ml. of anhydrous methylamine were placed in each compartment. -4 total of 50,000 coulombs (2.0 a m p . ; 90 v.) was passed through the solution in 7 hr. The solvent was then allowed to evaporate and the mixture was hydrolyzed by the slow addition of water. After extraction with ether, drying, and removal of solvent, there was obtained 9.0 g. (757,) of product boiling a t 149-153'. Analysis by v.p.c. (Xerograph, 14 ft. P,P' oxydipropionitrile column; 75'; 16 p.s.i. He) showed the presence of 89% isopropylcyclohexenes and 11Tc cumene When the same cell was used without the divider, 12 g. (0.1 mole) of cumene was reduced, using 34.0 g. (0.8 mole) of lithium chloride and 900 ml. of methylamine. Again, 50,000 coulombs (2.0 a m p . ; 85 v.) was passed through the mixture in T hr. Using the same work-up as described earlier, there was collected 9.8 g. (S2%) of product boiling a t 152-15.7'. Analysis by V.P.C. showed the presence of (3% isopropylcyclohexenes, 3 7 , of an unidentified diene, 13% cumene, and 78% 2,5dihydroisopropylbenzene. U-hen the reductions are carried out in the presence of the asbestos cell divider, the amine solution in the cathode compartment becomes deep blue in color. This color is highly reminiscent of the blue color ob-
