/ / MAGNETIC RESONANCE IN CHEMISTRY, VOL. 26, 564--570 (1988)

~# --------------~--------~~-~-~--~-~--~~-~--~~~

Diastereoisomers with Three Neighbouring Phenyl Groups

XF-Hindered Phenyl and Formylmethylamino Group Rotations in 3-(Formylmethylamino )-1,2,3-triphenylpropyl Chlorides

V. S. Dimitrov, V. B. Kurteva, M. J. Lyapova, B. P. M!khova and I. G. Pojarliefft Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, 1040 Sofia, Bulgaria

Unusually high free energies of activation for the rotation of phenyl groups, 12.1 and 11.1 kcal mo1- 1, are observed

in the (lR*, 2S*, 3S*) and (lR*, 2R*, 3S*) isomers of 3-(formylmethylamino)-1,2,3-triphenylpropyl chloride, and for the rotation of the formylmethylamino group (11.7 kcal mol- 1

) in the Z rotamer of the former isomer. These high barriers can be related to the interdependence of the rotation of such large planar groups in the strongly biased preferred conformations. The barriers for rotation of the formyl group are in the usual range (22.5 and 20.6 kcal mol- 1) for (lR*, 2S*, 3S*) and (lR*, 2R*, 3S*) isomers, respectively.

KEY WORDS 1H NMR Hindered rotation Phenyl, formylmethylamino and formyl groups 3-Formylmethylamino-1,2,3-triphenylpropyl chlorides

INTRODUCTION

At ambient temperature, the 1 H NMR spectra of a series of N-acylated 3-methylamino-1,2,3-triphenyl­propyl derivatives indicated the presence of hindered rotations of the phenyl groups and around the N-alkyl bond. Characteristic patterns of behaviour were observed depending on the configuration of the dia­stereoisomers, and we now report a more detailed study of two typical representatives, (1R*,2S*,3S*)- and (1R* ,2R*, 3S*)-3-( formylmethylamino )-1,2,3-triphenyl­propyl chlorides (designated ul and Ju, respectively, according to Ref. 2; I stands for like, e.g. R,R, and u for unlike, e.g. R,S):

H Ph Me ~h H \ \ '

\

' Ph \

N Cl Ph ........._CHO

' ' ' ' ' ' ' ' ' ' ' ' N-Me H Cl Ph H Ph H H I

CHO 1 ul 2 lu

As is usual for N,N-dialkylformamides, 3 the spectra of each isomer indicated the presence of the two E-Z rotamers with respect to the amide bond, more or less evenly populated. Slow rotation around the alkyl-N bond can be safely assumed to involve interconversion between two rotamers, each with the H-C-3 bond in

* For Part X, see Ref. 1. t Author to whom correspondence should be addressed.

0749-1581/88/070564--07 $05.00 © 1988 by John Wiley & Sons Ltd

plane with the amide group, since there is ample evi­dence that similar conformations are preferred for sec­ondary alkyl groups attached to sp2 hybridized atoms.4

Hence, in total, the following four isomers can be expected to result from the two hindered rotations:

H Me _\ j

2- r \_ C Ph /0

H 3 E-A

H Me

-t-j er Ph :,-H

0 5 Z-A

H H \_0

_\ N~ c2' r \

Ph Me

4 E-B 0

H >--H _\ N

er r \ Ph Me

6 Z-B

The forms arising from rotation around the N-C-3 bond are denoted A or B, in a similar manner as in Ref. 4.

The phenyl groups can be assumed to undergo a 180° flip, which is of course a pseudorotation, observable because of the large difference in the chemical shifts of the ortho and even the meta protons of some of the groups.

At the outset, the observation of slow sec-alkyl-N rotations in the formamido derivatives was unexpected, as these have low barriers and are observed at relatively low temperatures. Liden et al. 5 found AG* value (150 K) of 7.8 kcal mo1- 1 for N,N-diisopropylformamide, whereas with N-methyl-N-isopropylformamide no such rotation could be observed even at 133 K, although the

Received 7 December 1987 Accepted (revised) 8 March 1988

HINDERED PHENYL AND METHYLFORMAMIDO ROTATIONS IN TRIPHENYLPROPANES 565

question was left open as to whether this is due to a lower barrier or a strongly biased single conformation. In the case of vicinal groups in 2-thiothiazole deriv­atives, the same workers found similar barriers for neighbouring methyl and isopropyl and for two neigh­bouring isopropyl groups.6 No less unusual is the observed slow phenyl group rotation, as in simple deriv­atives of the PhCHR2 type 7 the barrier is in the region of 2 kcal mo! - 1 even in such relatively rigid compounds as phenylcyclohexane8 and, consequently, not detect­able on the NMR time scale.

The study of the hindered rotations in the title com­pounds is of particular interest since, as will be shown below, N-acyl-N-methylamino-1,2,3-triphenylpropyl derivatives are present in a single, strongly biased con­formation with respect to the rotations involving the propane skeleton, thus allowing a less ambiguous inter­pretation of the steric effects involved.

EXPERIMENTAL

Materials

3-Formylmethylamino-1,2,3-triphenylpropyl chlorides. These were prepared by formylation of the parent 3-methyl­amino chlorides9 with N-formylimidazole formed in situ. To a solution of 3 mmol of N,N-carbonyldi­imidazole in dry THF, 3 mmol of 100% formic acid were added with ice cooling and stirred for 1 h. A solu­tion of 1 mmol of a mixture of the two diastereoisomers with the same relative configuration at C-2, C-3 of 3-methylamino-1,2,3-triphenylpropyl chloride in dry THF was then added, and allowed to attain room tem­perature. The solvent was removed and the residue extracted with dichloromethane-water. The organic layer was washed with water, dried over MgS04 ,

evaporated and the residue recrystallized several times from isopropanol or ethyl acetate-hexane.

(lR*, 2S*, 3S*) isomer (ul). The starting material was a mixture of ul and ll isomers in a ratio of 1.6: 1. The recrystallized product was obtained in 43% yield; m.p. 192-193 °C; one spot on TLC (diethyl ether- hexane, 3: 2); IR (CHC1 3), vctto 1660 cm - l; MS (CI), 364 (22), 328 (12), 269 (6), 238 (1.4), 210 (2), 180 (43), 148 (57), 125 (4), 120 (28), 91 (0.1). C23H 22ClNO requires N, 3.85; Cl, 9.74; found, N, 3.97; Cl, 9.83%.

(IR*, 2R*, 3S*) isomer (lu). The starting material was a mixture of Ju and uu isomers in a ratio of 6.9: 1. The recrystallized product was obtained in 48% yield; m.p. 180-182 °C; one spot on TLC (diethyl ether-hexane, 3: 2); IR (CHCl3), vCHo 1660 cm - l; MS (CI), 364 (38), 328 (21), 269 (5.4), 238 (3), 210 (5), 180 (32), 148 (42), 125 (4), 120 (11), 91 (3). C23H 22CINO requires N, 3.85; Cl, 9.74; found, N, 4.14; Cl, 9.99%.

Note. The configurations of the formyl derivatives are based on the yields of the recrystallized products, which are compatible only with the assumption that the iso­lated compounds are obtained from the major com­ponent of the initial mixtures.

1 H NMR spectra

These were recorded on a Spectrospin WM-250 spectrometer with an Aspect 2000 Data System, using 4-5 mg ml- 1 solutions and TMS as internal standard. Typical recording conditions were spectral width 4.0 kHz, pulse width 3 µs (flip angle ca 50°), pulse repeti­tion time 2.05 s, 256 scans, 16K of data memory. The accuracy of the chemical shifts was better than ± 0.002 ppm and coupling constants were accurate to at least ±0.5 Hz.

Kinetic runs

The rates of interconversion of the Z-E rotamers were determined by monitoring the change of the popu­lations with time, until equilibrium, by integration of the methyl signals. The rate constants were determined from a least-squares treatment of the equation

where the subscripts 0, t and e refer to zero, t and equi­librium time, respectively, for the rotamer undergoing stereomutation.

The rates of conversion kz-E and kE-z were obtained from the relationships

k = kz-E + kE-z

and

The error in k was less than 6%; assuming an error of ca 5% in K (the values are close to unity), kz-E and kE-z should be accurate to ca 10%.

Variable-temperature measurements

These were made in the usual manner on a Spectrospin WM-250 spectrometer by recording spectra at 5 or 10 K intervals. The temperature readings were calibrated from the chemical shifts of MeOH.

ASSIGNMENT OF SPECTRA

At ambient temperature both isomers, ul and Ju, showed two sets of signals for the proton resonances (Fig. 1 and Table 1), as expected for dialkylformamides in which the E and Z isomers are known to be of similar stability.3 At higher temperatures in DMSO, signals of protons in the corresponding position of the molecule coalesced at different temperatures, depending on the shift difference, thus confirming that the pairs of signals are due to rotamers of a single compound.

The protons of the propane skeleton gave rise to AMX spectra (JAM= 0) with H-2 absorbing at highest field as a quartet. As seen in Table 1, in all cases a large coupling is observed with one of the remaining protons and a small coupling with the other. The large coupling

... lA•••

__L_.__ I I - \..l. . '·

b I I 7 ' s 4 ,

<

0.. '--~--"'--~~~-'--~~~--"-~~~--"'--~~~s.._~~~-"-~~~__J

Figure 1. 1 H NMR spectra of 3-formylmethylamino-1,2,3-triphenylpropyl chlorides in DMSO-d6 : (a) ul isomer at 294 K; (b) ul isomer at 413 K; (c) lu isomer at 293 K; (d) lu isomerat413 K.

d

c

(

V>

°' °'

:< ,.,, ti -~ -..., ~ 0 < t>i ..., ;i.. t""'

/

i'

,.,. HINDERED PHENYL AND METHYLFORMAMIDO ROTATIONS IN TRIPHENYLPROPANES 567

Table 1. 1H chemical shifts (ppm) from TMS and coupling constants (Hz) of 3-(formylmethylamino )-1,2,3-triphenylpropyl chlorides

Configuration a Solvent T (K) C!::t_, H-1 H-2 H-3 C!::l_O J(12) J(23)

ul CDCl 3 213 2.58b 5.06 3.85° 6.54 7.70 2d 12.4 2.54 5.05 5.38 8.33 2d 11.9

292 2.56 5.01 e 3.941 6.38 9 7.68 2.6 2.52 3.84 5.33 8.29 2.6 11.8

DMSO-d6 294 2.54 5.00h 4.30' 6.26 9 7.62 3d 2.38 4.38 5.49 8.36 3d 11.7

413 2.55 5.06h 4.26" 6.19 7.63 3.3 11.9 2.42 5.38 8.31 11.6

lu CDCl 3 233 2.95 5.34h 4.32 6.06 8.25 2.8 13.0 2.53 4.34 5.03 8.55 2.8 12.0

293 2.90 5.33h 4.30 6.08 8.22 3.0 12.7 2.55 4.31 5.02 8.54 2.8 12.0

DMSO-d6 293 2.86 5.50h 4.56 5.82 8.13 3.3 12.8 2.43 4.65 5.37 8.59 5.1 11.6

413 2.86; 5.46° 4.52° 5.90 8.21; 4.7 11.8 2.54; 5.31 8.60;

a The top line refers to the Z rotamer and the bottom line to the E rota mer. bZ-A form at 2.90 ppm. c Overlapping quartets. d Approximate value. •Common signal. 'Broad quartet. 9 Broad doublet. h Overlapping doublets. ; Broad singlet.

was assigned to that with H-3, from previous observa­tions that acylation in 3-methylamino-2,3-diphenyl­propanols10 and in 3-methyl-l,2,3-triphenylpro­panols11 strongly increased J(23). This assignment corresponds to the single conformations 1 and 2, since in the absence of other strong interactions only two conformations, each with a large and a small J value, avoid the prohibited 1,3-parallel interactions in 1,3-disu bsti tu ted 1,2,3-trip hen y I propanes. 12 This assign­ment is strongly supported by the following consider­ations. As already mentioned, the preferred conformations for rotation around the N-C-3 bond should be of type A or B (formulae 3-6).4 This has the important consequence that, irrespective of the relative configuration of the fragment C-2, C-3, u or I, a substit­uent at C-2 which is anti to H-3 will also be in a 1,3-parallel position with respect either to CHO in form A or to Me in form B and hence, will give rise to severe steric hindrance. This is the reason why conformations with anti H-2 and H-3 [large J(23)] are strongly pre­ferred. The steric interactions involved are readily rec­ognised in formulae 7 and 8, showing fragments with I-configuration at C-2, C-3. With both isomers, ul and Ju, the chemical shift difference between the E-Z rota­mers is much larger for H-3 than for H-1, which sup­ports the assignment in view of the magnetic anisotropy of the ~arbonyl group. 13

Apart from some aromatic protons, all other signals in the spectra of the Ju isomer are sharp, whereas with the ul isomer this is true for only one of the rotamers. For the second rotamer of the ul isomer the signals from H-3 and H-2 are broad, indicating exchange with a third form, although no further signals could be posi­tively identified (Fig. 1). On decreasing the temperature of a solution of the ul isomer in CDCl3 , however, the Me signal at 2.575 ppm showed the characteristic pattern of behaviour for exchange between two sites with strongly differing populations (Fig. 2), exhibiting a small peak at 2.90 ppm. Simultaneously, the chemical shift of the H-3 proton changes from 6.546 at 221 K to 6.28 ppm at 335 K which, on assuming a ca 0.1 popu­lation of the minor rotamer at the higher temperature, indicates a chemical shift of less than 4 ppm for H-3 of the latter. As seen from formulae 3-6 H-3 should be strongly deshielded 13 in Z-B, which is observed for the major form of the rotamer exhibiting slow methyl group exchange. The minor form should be Z-A. The rotamer not exhibiting slow methyl group exchange is then the E rotamer with respect to the amide bond; this is probably mainly in the E-B form as H and 0 have similar steric behaviour with respect to A-B type preferences. 5 •

14

Dreiding models indicate that the formyl proton in the E-B rotamer is in the deshielding zone of 2-Ph and 3-Ph, while it is closer to the shielding zone of 3-Ph in the Z-B rotamer; this can explain the considerable chemical shift difference.

The assignment of the rotamers of the Ju isomer as Z-B and E-B is tentative and is based on similar con­siderations.

568 V. S. DIMITROV ET AL.

Z-A a

Z-B

E-B

b

2.5

Z-B

Z-B E-B E-B

c

2.5 2.5

Figure 2. Signals of the methyl protons in the 1 H NMR spectra of the ul isomer of 3-formylmethylamino-1,2,3-triphenylpropyl chlor ­ide in CDCl 3 at (a) 213 K; (b) 253 Kand (c) 278 K.

DETERMINATION OF ROTATIONAL BARRIERS

Attempts to determine the barriers of rotation for the formyl group by variable-temperature measurements in DMSO were abandoned, because the compounds began to decompose above 400 K before coalescence of the well defined signals of Me or CHO. The rates of conver­sion of the Z rotamers into the equilibrium mixture could be measured directly because of the slow stereo­mutation, and the two isomers crystallized as Z rota­mers. The free energy of activation was determined from

AG* = RT[ln(kB/h) - ln(k/T)] (1)

where the symbols have their usual meanings. The barrier to rotation around the alkyl-N bond in

the Z rotamer of the ul isomer was determined by the approximation of Anet and Basus, 15 for exchange between two sites of very unequal populations ( > 10: 1 ), from the maximum broadening of the Me signal at 2.575 ppm which exchanged with the minor signal at 2.90 ppm. The rate at maximum broadening is given by

The maximum broadening was determined by inter­polation of the temperature dependence of the band width and the chemical shift difference at this tem­perature, bv, by extrapolation of the temperature depen­dencies of the chemical shifts at slow exchange. In the latter temperature interval, 213- 228 K, the population

\._

ratio remains almost constant (0.03: 0.97), and this was used in Eqn (2) at 253 K.

Barriers to rotation of a phenyl group could be deter­mined in two cases. The slow rate of interconversion of the E- Z rotamers allowed samples of the ul isomer to be obtained on rapid cooling which contained the Z rotamer with insignificant amounts of the E rotamer. In the low-temperature spectra the phenyl group reson­ances could be partially analysed. Two isolated peaks integrating for one proton were observed at 213 K: a triplet (meta proton) at 7.06 ppm (J ::::: 7.5 Hz) and a doublet (ortho proton) at 6.33 ppm (J = 7.2 Hz). A triplet at 7.49 ppm and a doublet at 7.79 ppm, partially overlapping with other bands, could also be discerned. On increasing the temperature the triplets coalesced at ca 243 K, as judged by the disappearance of the triplet at 7.06 ppm and the corresponding change in the inte­grals of the other bands. The coalescence of the ortho proton absorptions could be determined more clearly, since the averaged signal appeared in a spectral position free from other signals. After coalescence the integral of the band at ca 7.8 ppm was reduced from that of four protons to three protons.

The low-temperature spectra of the same isomer com­prising the equilibrium mixture of the E and Z rotamers showed the same signals at 6.33 and 7.06 ppm which were distorted; these integrated for one proton with respect to the sum of the rotamers. On varying the tem­perature the signals underwent changes which were very similar to those observed with the pure Z rotamer, indi­cating a similar barrier for the same phenyl group in the E rotamer. "'

Owing to the relatively rapid stereomutation in the Ju isomer, only spectra of the mixture of the rotamers could be recorded. At lower temperatures in CDC1 3 the phenyl resonances underwent complex changes which could not be analysed. In CD2Cl 2 , however, where spectra could be recorded down to 183 K, a distinct coalescence process of two ortho protons [doublets at 6.07 and 7.48 ppm (J = 7.5 Hz)] was observed, giving finally a single doublet at 6.83 ppm at 286 K. The coalescence temperature of 248 K was determined as the mid-point between the temperature of the spectrum at which the two separate signals could still be dis­cerned and that at which the averaged signal appeared (223 and 263 respectively, at 10 intervals), and is prob­ably accurate to 10 K. The integrals of the ortho protons identified them as belonging to the major isomer. The rate of rotation of the phenyl groups at the coalescence temperature was determined using the approximation of Gutowsky and Holm16 for exchange between equally populated sites without coupling:

k = nb~ )2

(3)

where bv is the chemical shift difference in the absence of exchange assumed to be equal to that at slow exchange.

As the barriers to rotation were obtained with varying degrees of accuracy with regard to the determi­nation of both the rate and the temperature, an estimate of the possible error limits seemed appropriate. The fol­lowing equation for the error in AG* can be obtained

HINDERED PHENYL AND METHYLFORMAMIDO ROTATIONS IN TRIPHENYLPROPANES 569

Table 2. Rate constants and free energies of activation of rotation in 3-formylmethylamino-1,2,3-triphenylpropyl chlorides

Configuration Process T (K) K. k (s- 1 ) f>.G" (kcal mol-')

ul Z-+E 290 0.84 7.1±10-5 22.5 ± 0.13 E-+Z 290 8.5 x 10- 5 22.4 ± 0.13

Z-B -+Z-A 253 0.03• 14• 13.4• Z-A-+Z-B 253 453 11 .7 ± 0.14

Ph 269 803 12.1 ± 0.45

lu Z-+E 290 0.94 1.8x10-3 20.6±0.12 E-+Z 290 1.9x10-3 20.6±0.12

Ph 248 780 11.1 ± 0.45

•Approximate value owing to uncertainty in the population of the minor com­ponent.

from Eqn (1) in the usual manner after a minor approx­imation:17

(<J,.,,6 ,)AG*)1

= {[ln(k8/h) ~ ln(k/T)] 2 }

x (<JJk)1 + (<lr/T)1 (4)

As can be readily seen, when <JJk = 0 the relative error in Tis directly equal to that of AG*. As the error in T for the macrokinetics, for the Anet and Basus approx­imation and for the coalescence determinations can be safely assumed to be better than ± 1, ± 3 and ± 10 K, respectively, these contributions amount to 0.3% (290 K), 1.2% (253 K) and 4.0% (248 K). The contribution of <JJk, the first term in Eqn (4), is much less significant. A very important feature of Eqn (4) is its insensitivity to variations in the k/T ratio. If the latter changes numerically from 2.5 x 10- 7 to 5, the contribution of <JJk to <J46 ,,/AG* increases by less than a factor of 2. Hence the maximal error in k of 4%, claimed by Anet and Basus15 for their approximation when the ratio of the major to the minor form is at least 10: 1, contributes less than 0.2% to. the relative error in AG* within the above k/T limits (<Jr/T = 0). The contribution of the ca 10% error in the rate constants from the macrokinetics is similarly less than 0.5%. In the coalescence measure­ments <JJk = <J0Jbv, and in our actual determinations this can arise from the temperature dependence of the chemical shift difference. Since these differences are very large, a 10% error is probably a safe margin. The actual contributions to <J46 ,,/AG* are, of course, smaller because of the relatively large value of the denominator of the first term ..

The results obtained for the rates and rotational bar­riers are summarized in Table 2 (the errors quoted are the maximal limits evaluated above).

DISCUSSION

The most interesting results from this study are the unusually high barriers observed for the rotation of the phenyl groups of 11-12 kcal mo1- 1 when compared with barriers in simple compounds with the phenyl group attached to a secondary alkyl group. These, as already mentioned, are around 2 kcal mo! - 1. The most

likely reason for the high barriers in the triphenyl deriv­atives is that the phenyl rotations are in some way coupled with those of the formylmethylamino group, the latter having intrinsically higher barriers of rotation.

The phenyl groups whose rotational barriers could be determined are characterized by a very large difference in the chemical shifts of the two ortho protons. This is expected for a phenyl group gauche to another phenyl group with one ortho proton in the shielding and the other in the deshielding zone caused by the second benzene ring. In the preferred conformations of the two isomers studied :

9 ul-Z-B (A)

10 lu-Z-B(A)

such groups are 1-Ph and 2-Ph in the ul isomer and 1-Ph and 3-Ph in the Ju isomer (for 2-Ph in Ju the effect of the two flanking phenyl groups will be partially cancelled). As can readily be seen from molecular models the rotations of the large planar groups, phenyl or formylmethylamino, involve considerable librations of the other planar groups in a vicinal or geminal posi-

570 V. S. DIMITROV ET AL.

tion. Such a coupling should be further enhanced in these congested molecules, as the other modes of releas­ing steric strain in the transition state for rotation such as torsional and bond angle deformation are expected to become more difficult. All these arguments suggest that the hindered rotations observed involve 2-Ph in the ul isomer and 3-Ph in the Ju isomer. Although with ul the values obtained for AG* of the phenyl and for­mylmethylamino group rotations are similar, these rota­tions are not necessarily fully concerted but, undoubtedly, the crossing of one of the barriers involves considerable 'climbing up' the other. The slow for­mylmethylamino group rotation in the ul isomer is observed in the rotamer assigned as Z. On rotation, steric hindrance will involve an oxygen atom in the Z rotamer or a hydrogen atom in the E rotamer, and this should contribute to higher barriers in the former.

No hindered methylformylamino group rotation could be observed in the Ju isomer in the temperature range studied. This could be due to a lower barrier

"­and/ or strongly biased conformations with respect to the N-alkyl rotation. For the formylmethylamino group rotation the difference in environment is that gauche to the formamido group there is a phenyl group in the ul isomer, whereas it is C-1 in the Ju isomer (formulae 1 and 2). On the assumption that the B forms are pre­ferred, a strongly biased A:¢ B equilibrium would suggest that interference with the methyl group is more severe with C-1, while a lower barrier would suggest the opposite. This cannot be answered from the available evidence; for example, the lower barrier to rotation for the formyl group (the E-Z interconversion) in the Ju isomer could be due to increased ground-state strain in the first case, as is usual with amides, or to smaller transition state strains in the second case.

Acknowledgement

We thank Professor J. Sandstrom for helpful discussions.

REFERENCES

1. V. B. Kurteva, M . J. Lyapova and I. G. Pojarlieff, lzv. Otd. Khim. Nauki, Bulg.Akad. Nauk. 21, in press (1988) .

2. D. Seebach and V. Prelog, Angew. Chem. 21, 654 (1982). 3. W. E. Stewart and T. H. Siddal, Ill, Chem . Rev. 70, 517

(1970). 4. U. Berg, T. Liljefors, C. Roussel and J. Sandstrom, Acc. Chem.

Res. 18, 80 (1985). 5. A. Liden, C. Roussel, T. Liljefors, M. Chanon, R. E. Carter, J .

Metzger and J . Sandstrom, J. Am. Chem. Soc. 98, 2853 (1976) .

6. C. Roussel, A. Liden, M. Chanon, J. Metzger and J . Sand­strom, J. Am. Chem. Soc. 98, 2847 (1976) .

7. W. J . E. Parr and . T. Schaefer, Acc. Chem. Res. 13, 400 (1980) .

8. T. Schaefer, W. Niemczura and W. Danchura, Can. J. Chem. 57, 355 (1979).

9. V. B. Kurteva, M. J . Lyapova and I. G. Pojarlieff, J. Chem. Res. (S) 398, (M) 3311 (1986) .

10. M. Poneva, V. Dimitrov, B. Jordanov, A. Orahovatz and B. Kurtev, lzv. Otd. Khim. Nauki, Bulg. Akad. Nauk 6, 671 (1973); Chem. Abstr. 80, 145279j (1974).

11. M. J. Lyapova, I. G. Pojarlieff and B. J. Kurtev, Dok/. Baig. Akad. Nauk 34, 1513 (1981); Chem. Abstr. 97, 38192e (1982) .

12. D. C. Best and C. A. Kingsbury, J . Org. Chem . 33, 3253 (1968) .

13. H. Paulsen, K. Todt and H. Ripperger, Chem . Ber. 101, 3365 (1968); K. Todt and H. Paulsen, Fresenius Z. Anal. Chem. 235, 29 (1968) .

14. I. Pettersson and J. Sandstrom, Acta Chem. Scand., Ser. B 38, 397 (1984).

15. F. Anet and V. Basus, J . Magn. Reson. 32, 339 (1978) . 16. H. S. Gutowsky and C. H. Holm, J . Chem. Phys. 25, 1228

(1956). 17. G. Binsch, in Dynamic Nuclear Magnetic Resonance Spec­

troscopy, edited by L. M. Jackman and F. A. Cotton, p. 45. Academic Press, New York (1975).