Carbon-13 Magnetic Resonance Spectra and the Stereochemistry of Some Substituted

Thomas Gibson The Procter & Gamble Company, Miami Valley Laboratories, Cincinnati, Ohio 45247, USA

Using selective deuteriation and broad-band and off -resonance proton decouplhg techniques, complete signal assignments are made for the 13C NMR spectra of a number of ketones and unsaturated esters with the bicyclo[2.l.l]hexane ring systems.

INTRODUCTION

Thermal isomerization of the ester l a has been shown to give the isomer 2a, presumably by homolytic cleav- age of the C-1-C-5 bond.' The configuration about the double bond in l a and 2a was established by photoisomerization to 3a and 4a, respectively. The chemical shift differences of the easily identified pro- tons on C-3 were in accord with the configurations as shown. In seeking to confirm this assignment through the use of 13CNMR, we found that the signal for C-3 could not be unambiguously established. The carbon resonances for the CH, groups at C-3 and C-6 in la-4a could not be resolved into an unambiguous pattern that was internally consistent, and thus provide independent evidence for the configurations about the double bonds. It appeared that deuteriation at C-3 would solve this problem by the well known reduction in intensity of the resonance of a fully deuteriated carbon, owing to a longer T, relaxation time. In this

R A

I 2

3 4

a . R = H b. R=D

paper a method is reported for the synthesis of the selectively trideuteriated esters 1b-4b, which not only solves this problem, but also allows complete assign- ment of all carbon resonances in these compounds. The spectral data for some related ketones are also included.

RESULTS AND DISCUSSION

The original method used for the synthesis of l a was modified by the use of deuteriated carvone 5 as the starting material.' Exhaustive exchange of the acidic protons in carvone gave 5, in which approximately 95% of d, material was present ('HNMR analysis). Irradiation in methanol gave the tetradeuteriated car- vonecamphor 6, which was not isolated but was con- verted directly to 7. Prolonged irradiation of car- vonecamphor in methanol has been shown to provide the non-deuteriated analog of 7,'~~ and it has also been established that photolysis of the a,o -dideuterio analog of 6 results in the transfer of a deuterium atom to the 5-end0 position of the product.' The gas- chromatographic behavior of 7 was identical with that of its non-deuteriated co~n te rpa r t .~ Hydrolysis gave the crystalline acid 8, which provided spectral data in full accord with the assigned structure. Bromination of 8 was carried out with PBr3/Br,, followed by methanolysis, to give a mixture of epimeric bromo esters 9, in which no deuterium was present at the a -carbon atom. Presumably, complete exchange oc- curred under the acidic conditions of the bromination. Dehydrobromination under mild conditions gave a single unsaturated ester product, lb , as previously observed for 1a.l Comparison of the 'HNMR spec- trum of l b with that of l a showed the expected disappearance of the two-proton quartet previously ascribed to the C-3 protons in la . In addition, the doublet signal appearing at 61.16 in la, assigned to the methyl group on C-5, now appeared as a slightly broadened singlet. The presence of deuterium at C-5 was also indicated by the absence of the one-proton quintet at 61.65, previously assigned to the proton at C-5 in l a . Esters l a and l b also showed identical

CCC-0030-4921/83/0021-0482$02.50

482 ORGANIC MAGNETIC RESONANCE, VOL. 21, NO. 8, 1983 0 Wiley Heyden Ltd, 1983

13C NMR AND STEREOCHEMISTRY OF SOME SUBSTITUTED BICYCL0[2.1 .l]HEXANES

D

5 6

1

8 7

I

9 Ib

behavior on GLC and HPLC analysis. With l b in hand, the complete set of isomers l b 4 b could be generated. This was done as previously described, by thermal rearrangement of l b to 2b, and by photo- chemical rearrangement of l b and 2b to 3b and 4b, respectively. In all cases, reactions and products were entirely consistent with those of the non-deuteriated analogs.

The ketones 10-13 were available from a previous study.4 The ketones 14 and 15 were prepared from l b and 2b, respectively, by oxidation with KMnO,/NaIO,.' These ketones showed chromato- graphic and spectroscopic behaviour in complete accord with their non-deuteriated analogs 12 and 13. Details for all of the above transformations and spectroscopic data are given in the Experimental section.

The I3C NMR chemical shifts for all eight esters, obtained from the broad-band, proton-decoupled spectra, are presented in Table 1. Spectra for com- pounds la-3a were also obtained using off -resonance decoupling, and the spectrum of 4a was determined under completely coupled, NOE enhanced conditions, allowing the substitution patterns for all eleven car- bons to be determined. This allowed assignment of

signals to C- 1, the only saturated quaternary carbon present, appearing as a singlet in the coupled spectra.

Differentiation of signals for carbons 2 and 10 was made on the basis of the upfield shift observed upon isomerization of the C-5-methyl from the exo to the endo configuration. This isomerization would not be expected to affect the remote carbonyl carbon (C-lo), but should shift the signal for C-2, although its direc- tion and magnitude are difficult to predict, since there are few available data on comparable compounds in this ring system.

Assignment of signals for carbons 3-6 is made by combination of the off-resonance data with the results of deuteriation. Thus, the triplet at 639.5 in l a broadens to a barely detectable multiplet in l b and is therefore due to C-3. The doublet at 650.7 also undergoes a similar change, and is thus derived from C-5. The remaining doublet at 637.0 in l a and 36.6 in l b can be therefore assigned to C-4, and the triplet signals at S41.8 and 41.7 to C-6. Similar arguments provide the results shown in Table 1 for the remaining six esters. The chemical shift differences observed at C-2 upon isomerization at (2-5 also appear at C-3 and C-6 and occur in an entirely consistent manner throughout the whole series, therefore strengthening the assignments made. Signals due to C-9 and C-11 (the ester methyl group) were assigned by off- resonance decoupling and comparison with literature values for similar compound^.^ The above data now permit assignment of configuration about the double bond, on the basis of the shifts induced at C-3 upon photoisomerization. Thus, the signal at 639.5 in l a shifts to 41.2 in 3a, the downfield shift being entirely consistent with that expected for the configurations as written. For example, in the cis isomer 16, the signal for C-4 appears 5.9ppm upfield from that of the

-7=/Co2H C0,Me H O P

16 17

corresponding carbon in the trans isomer.6 The effect is less pronounced in the acid 17, where Ascis-trans = -2.3ppm,' a value in good agreement with that ob- served for la-3a. Thus the deuteriation/photoiso- merization corroborates the double bond configuration of these compounds, as previously deduced from 'H NMR spectral data.*

The assignment of the signals to the two methyl groups was worked out as follows. It would seem

Table 1." -C NMR chemical shifts of compounds 1-4

Compound C-1 c-2 c-3 c 4 c-5 c-6 c-7 C-8 c-9 c-10 c-11

l a 2a 3a 4a l b 2b 3b 4b

57.6(s) 172.6(s) 58.7(s) 170.6(s) 58.3(s) 164.8(s) 59.9(s) 162.4(s) 57.4 172.5 58.6 170.5 58.2 164.6 59.8 162.3

39.5(t) 37.0(d) 50.7(d) 34.8(t) 36.3(d) 49.9(d) 41.2(t) 36.4(d) 50.5(d) 37.1 (t) 35.8(d) 50.7(d)

36.6 36.0 36.1 35.5

41.8(t) 45.4(t) 42.5(t) 45.w 41.7 45.2 42.4 45.8

12.4(q) 10.7(q) 14.l(q) 10.4(q) 15.3(q) 10.9(q) 17.6(q) 10.7(q) 12.4 10.6 14.0 10.2 15.3 10.7 17.6 10.7

105.4(d) 108.2(d) 109.2(d) 1 12.4(d) 105.4 108.2 109.4 112.5

167.8(s) 167.6(s) 166.7(s) 166.8(s) 167.8 167.6 166.7 166.8

50.7h) 50.8(q) 50.9(q) 50.9(q) 50.7 50.8 50.8 50.8

a Chemical shift in ppm downfield from TMS, CDCI, solution. s = singlet, d = doublet, t = triplet, q = quartet.

ORGANIC MAGNETIC RESONANCE, VOL. 21, NO. 8, 1983 483

T. GIBSON

unlikely that isomerization at the double bond would have an appreciable effect on the chemical shift of the e m methyl group in the 1 and 3 series. In fact, there is also little effect on C-5 and C-6. In the conversion of l a to 3a, one of the methyl groups shifts downfield by either 2.9 or 4.6ppm. If the latter is true, the other methyl must be shifted upfield by 1.5 ppm, which seems unlikely. This suggests that the C-8 methyl gives rise to the signal which occurs in the region of 10.7 ppm in both compounds, while the C-7 signal is shifted downfield by 2.9 ppm. A closer inspection of the spectra of la-4a revealed that the upfield signal in each case displayed a somewhat wider band width, suggesting additional coupling. Since the C-7 methyl is attached to a quaternary carbon, there is no possibility

for two-bond CH coupling. The proton at C-5, how- ever, could couple to C-8 to the extent of 4 - 6 H ~ . ~ This was confirmed for 4a by inspection of the com- pletely coupled, NOE enhanced spectrum, which showed that the signal centered at 10.7ppm is a quartet of doublets with coupling constants of 124 and 6 Hz. Assignment to all the remaining compounds follows on the basis of the above arguments.

Effects similar to those observed in the esters are apparent in the ketone data in Table 2. Again, off- resonance decoupling and deuteriation allows unam- biguous assignments to all signals except the methyl groups in 12-15. The assignments given are based on the increased band width of one of the signals, pre- sumably due to C-8, in the partially decoupled spectra

Table 2." =CNMR chemical shifts of compounds 10-15 Compound

0 10

0 d 11

0 & 12

0 13

0 D 14

c-1 c-2 c-3 C-4 C-5 C-6 C-7 C-8

56.l(d) 214.4(s) 40.8(t) 35.9(d) 41.0(t) 41.0(t)

64.2(s) 215.4(s) 37.7(t) 35.l(d) 49.8(d) 44.0(t) I l . l (q ) 10.8(q)

62.7 214.7 34.9 40.6 9.7 10.3

'hD 64.1 215.5 34.8 43.9 11.1 10.7

0 15

a Chemical shifts in ppm downfield from TMS. CDCI, solvent. s=singlet, d=doublet, t = triplet, q = quartet.

484 ORGANIC MAGNETIC RESONANCE, VOL. 21, NO. 8, 1983

13C NMR AND STEREOCHEMISTRY OF SOME SUBSTITUTED BICYCL0[2.1.1]HEXANES

of 12 and 13. The downfield shift of 1.4ppm for C-7 which results from the exo-endo shift is consistent with effects observed in the esters.

Inspection of the data in the tables shows that the signals for C-1, C-2, C-4 and C-8 in the deuteriated compounds occur at slightly shielded positions com- pared with those for these carbons in the non- deuteriated analogs. The effect is especially pro- nounced at C-4, and can be explained on the basis of a cumulative P-deuterium isotope effect. The spectra were not obtained with sufficient accuracy to deter- mine the amount of this effect quantitatively, but these shifts lend qualitative support to the signal assign- ments made on the basis of other effects.

It is interesting that the exo-endo methyl change (12 + 13, 14 .+ 15) results in a slight downfield shift of the carbonyl carbon signal, whereas methyl y- effects at bicyclic keto groups are generally slightly shielding,' as they are at C-2 in the esters 1-4. Another exception to this general observation occurs in 7-syn-methylbicyclo[2.2.l]heptan-2-one, which shows a deshielding of the carbonyl signal by 0.8 ppm with respect to the 7-anti compound.'' It is apparent that it is not always possible to make structural assign- ments based simply on chemical shift changes in these types of compounds, without some additional informa- tion about the relative orientations of the carbon atoms involved.

EXPERIMENTAL

Infrared spectra were determined for the neat com- pounds on a Perkin-Elmer Model 257 spectrometer; 'H NMR spectra were recorded in CDC1, solution on a Varian HA-100 spectrometer and mass spectra on a Kratos MS-30 spectrometer. I3C NMR spectra were obtained on a Varian CFT-20 spectrometer at 20 MHz with a resolution of *2Hz. The coupled spectrum of 4a was obtained on a Jeol FX-270 spectrometer at 67.8 MHz. All spectra were run in CDC1, solution (50-100 mg ml-') and were measured in ppm from TMS as an internal reference.

Tetradeuteriated acid (8)

Sodium metal (1.Og) was dissolved in 50.6g of CH,OD, 30.3g of D,O were added, then 15.0g of d-carvone. The mixture was heated to reflux (1 h), cooled, diluted with diethyl ether, extracted twice with saturated sodium chloride solution and the ether solu- tion dried over magnesium sulfate. Removal of solvent gave 12.7 g of deuteriated carvone, which showed 85% exchange of the four protons on C-3 and C-5 by 'HNMR analysis. Repetition of the above process gave 9.8 g (64%) of ca 95% carvone-d,. This material was dissolved in 500ml of anhydrous methanol and irradiated through Vycor with a 450 W mercury lamp for 17 h. The solvent was removed by distillation, followed by vacuum distillation of the residue to give 6.45 g (55%) of the crude methyl ester 7 , b.p. 56- 62 "C (4.4 mmHg). This material was -95% pure, and

showed a GLC retention time identical with that of the non-deuteriated analog. The infrared spectrum showed C-D bands at 2145 and 2215cm-', and the 'H NMR spectrum showed two methyl singlets at 60.90 and 1.07 ppm. Hydrolysis of the ester in 50 ml methanol/20 ml H,O with 4.1 g of KOH gave 6.4 g of crude acid. Crystallization from formic acid gave 3.3 g of acid 8 (57%), m.p. 49-51 "C (lit. m.p. 50-52 "C for the undeuteriated acid).2 Vacuum sublimation of 107 mg gave 104 mg of pure 8, m.p. 50 "C, which showed infrared bands at 2220, 2150 and 1708cmp', 'HNMR signals at 60.95 (3H, s, H-7), 61.07 (3H, brd. s, H-8), 1.10 ( lH, H-6-endo), 1.92 (IH, dd, J = 7 . 0 and 2.8Hz, H-6-exo), 2.05 (lH, d, J=2 .8Hz , H-4) and 2.22 (2H, m, H-2 and H-9). The mass spectrum gave a molecular ion at m/z 172.1357; calc. for C,,H,,D,O,, 172.1401.

Bromoester (9)

To 3.2 g of crude acid 8 in a 100 ml round-bottomed flask were added 9.0 of PBr,. After stirring for ap- proximately 15 min a clear solution was obtained to which 11.0 g of Br, was added in three portions over 20min. The mixture was heated on a steam bath for 3 h, cooled and carefully quenched with 30 ml of anhydrous methanol. The resulting solution was di- luted with 200ml of diethyl ether, extracted with water (3 x 50 ml) and saturated NaHCO, (3 X 50 ml) and dried over MgSO,. Filtration, removal of solvent on a rotary evaporator and vacuum distillation gave 4.6 g (93%) of 9, b.p. 75-78 "C (1 mmHg). This ma- terial showed a single peak on GLC with the same retention time as the non-deuteriated analog and, like the latter material,' the NMR spectrum showed a pattern indicative of the presence of an epimeric mix- ture of bromides.

exo-E-Trideuterio ester (lb)

Dehydrobromination of 3.0g of 8 with sodium methoxide in methanol was carried out as described for la' (60 "C, 4 days). Kugelrohr distillation gave 1.67 g (80%) of lb , b.p. 90-110 "C (1 mmHg); a pure sample was obtained by preparative HPLC on a What- man 10/50 Magnum PAC column (95:5 hexane- diethyl ether, 4.0 ml min-l) followed by Kugelrohr distillation. This material showed urn= 2208, 2145, 1710 and 1664cm-', 230nm ( E 14300), 'HNMR 61.08 (3H,s), 1.17 (3H,s), 1.27 ( lH, d, J = 8 Hz), 2.21 (2H,m), 3.72 (3H,s) and 5.65 ( lH, s) and showed a molecular ion at m/z 183.1323; calc. for CllH13D302, 183.1338.

endo-E-Trideuterio ester (2b)

The ester lb , 0.1735 g, was sealed in a small ammonia-washed ampoule which had been dried at 110 "C for approximately 1 week, and heated in an oil bath at 175°C for 18h.' HPLC analysis showed an 84: 16 ratio of 2b: l b , identical with that observed

ORGANIC MAGNETIC RESONANCE, VOL. 21, NO. 8, 1983 485

T. GIBSON

with the non-deuteriated analog.' Purification by pre- parative HPLC (of lb), followed by Kugelrohr distilla- tion, gave 0.1104 g of 2b (64%). This material showed v,,, 2215, 2160, 1720 and 1664 cm-', 226 nm ( E 10 500), 'HNMR 60.58 (3H, s), 1.13 (3H, s), 1.25 (lH, d, J=6.4Hz), 1.51 (lH, dd, J=2 .8 and 6.2Hz), 2.35 (1H, d, J=2.8Hz), 3.72 (3H,s) and 5.65 (lH,s) and a molecular ion at miz 183.1321.

exo-Z-Trideuterio ester (3b)

A solution of 0.4041 g of l b in pentane was irradiated with a mercury resonance immersion lamp (254nm) for 150 min. HPLC analysis showed a 31 : 69 mixture of 3b:lb, identical with that observed with the un- deuteriated analogs. Preparative HPLC followed by distillation gave 81 :4mg (20%) of pure 3b, v,, 2210, 2150, 1725 and 1656cm-l, hzzH 234nm ( E

16500), 'HNMR61.14 (3H,s), 1.36 (3H,s), 1.44 (lH,m), 2.15 (lH,s), 2.18 (lH, m), 3.68 (3H, s) and 5.67 (lH, s) and a molecular ion at m/z 183.1323.

endo-Z-Trideuterio ester (4b)

The ester lb, 0.3085 g, was pyrolyzed as previously described to give an 83 : 17 mixture of 2b : lb. This was combined with 0.1OOg of pure 2b, dissolved in hex- ane, flushed with argon and irradiated for 2 h with a mercury resonance lamp. HPLC analysis showed a mixture of 4b : 3b: 2b: l b in the ratio 25 : 5 : 61 : 9, similar to that obtained with the non-deuteriated analog. Preparative HPLC followed by Kugelrohr dis- tillation gave 64mg of 4b, showing urn, 2220, 1733 and 166Scm-', AEtoH 233 nm ( E 11 SOO), 'H NMR 60.69 (3H, s), ?? (2H, m), 1.45 (3H, s), 2.29 (1H,d,J=3Hz), 3.67 (3H,s) and 5.83 (lH,s) and a molecular ion at m/z 283.1341. The coupled 13C NMR spectrum showed signals at 6166.8 (s),

162.4 (s), 112.4 (d, J = 157Hz), 59.9 (s), 50.9 (9, J = 146Hz), 50.7 (d, J = 142Hz), 45.9 (t, J = 142Hz), 37.1 (t,J=133Hz), 35.8 (d, J=153Hz), 17.6 (4, J = 127 Hz) and 10.7 (qd, J = 124 and 6 Hz) ppm.

1, ex0 -5 -Dhethy1-3,3,5- trideuteriobicyclo[2.2.l]hexan-Z-one (14)

To a solution of 0.10 g of KMnO,, 2.20 g of NaIO, and 0.80g of K,CO, in 50ml of water was added 0.283 g of ester l b in 1 ml of THF. The mixture was stirred vigorously at room temperature for 18 h, ex- tracted twice with diethyl ether and dried over MgSO,. GLC analysis showed only ketone 14, identi- cal in retention time with the non-deuteriated com- pound 12. Preparative GLC gave 110.3mg (54%) of pure 14, showing v,, 2220, 2150 and 1750cm-', 'HNMR 61.01 (3H, s), 1.17 (3H, s), 1.56 (lH, d, J=7.5Hz),2.41(1H,d,J=3.5Hz)and2.47(1H,dd, J=7.5 and 3.5Hz) and a molecular ion at mlz 127.1082; calc. for C8H9D30, 127.1076.

l-endo-5-Dimethyl-3,3,5- trideuteriobicyclo[2.l.l]hexan-2-one (15)

Oxidation of 2b, 0.303g, in the same manner as that used for lb, gave 26.2 mg (12%) of pure 15 after preparative GLC. This material showed v,, 2220, 2160 and 1750cm-', 'HNMR 60.77 (3H, s), 1.05 (3H, s), 1.58 (lH, d, J=7.0Hz), 1.81 (lH, dd, 3.0 and 7.0 Hz) and 2.60 (lH, d, J = 3.0 Hz) and a molecular ion at m/z 127.1088.

Acknowledgments

The author expresses his appreciation to Dr Fouad Ezra and Mr Jack Wendel for invaluable assistance with the I3C spectra.

REFERENCES

1. T. Gibson, J. Org. Chem. 46, 1073 (1981). 2. J. Meinwald and R. A. Schneider, J. Am. Chem. SOC. 87,

3. T. W. Gibson and W. F. Erman, J. Org. Chem. 31, 3028

4. T. W. Gibson and W. F. Erman, J. Org. Chem. 37, 1148

5. J. 6. Stothers, Carbon-13 Nuclear Magnetic Resonance

6. T. A. Bryson, Tetrahedron Lett. 4923 (1973).

5218 (1965).

(1966).

( 1 972).

Spectroscopy. Academic Press, New York (1972).

7. E. Lippmaa, T. Pehk, K. Anderson and C. Rappe, Org. Magn.

8. F. W. Wehrli and T. Wirthlin, Interpretation of Carbon-13

9. G. E. Langford, H. Auksi, J. A. Gosbee, F. N. Machlachlan

10. J. 6. Grutzner, M. Jautelat, J. 6. Dence, R. A. Smith and J.

Reson. 2, 109 (1970).

NMR Spectra, p. 54. Heyden, London (1976).

and P. Yates, Tetrahedron 37, 1091 (1981).

D. Roberts, J. Am. Chem. Soc. 92, 7107 (1970).

Received 8 October 1982; accepted 4 March 1983

486 ORGANIC MAGNETIC RESONANCE, VOL. 21, NO. 8, 1983