NMR studies of rate processes and conformations—XVIII

Tetrahedron.

Vol. 26. pp. 4227 to 4240.

Pergmon

Rem

1970.

RiMed

in Gmat Britain

NMR STUDIES OF RATE PROCESSES AND CONFORMATIONS--XVIII HINDERED

NITROGEN INVERSION IN 5-MEMBERED HETEROCYCLIC AMINE!? J. M. LEHN and J. WAGNER

Institut de Chimie, I rue Blake Pascal, Strasbourg-67, France? (Received in the UK 29 January 1970; Acceptedfor

publication 21 April 1970)

A&met- Barriers to nitrogen inversion have been studied in a number of 5-membered cyclic amities by variable temperature NMR spectroscopy. Electronegative substituents (Cl, Br, N, 0) on nitrogen lead to a marked increase in inversion barrier as compared to the parent compounds where nitrogen is only linked to carbon. The substituent effects observed in these compounds are compared with those obtained in aziridine derivatives. It is found that the effects are much larger in the latter cases. The structural effect on substituent effects due to the Smembered ring + 3-membered ring” structural change is estimated to be of t.lx order of 6 kcal/mole. Steric and solvent effects ate also discussed. R&m&- Les barritres d’inversion de I’azote dans un certain nombre d’amina cycliques a cmq chaIrions ont ttt ttudi&ea par RMN a difftrentes temperatures. La presence de substituants plus Ckctrontgatifs que le carbone (Cl, Br, N, 0) lib directement a I’azote produit une augmentation notabk de la barribre d’inversion. Les effets de substituant obtenus pour ces composes ont ttt compares a ceux observes pour da dtrivts de I’aziridine; ces effets sont beaucoup plus importants dans ce demia CBS.La modification structurale “cycle a cinq chainons + cycle a trois chafnons” donne lieu a un effet structural sur les effets de substituant, estimt a environ 6 k&/mole. Quelques interactions steriques et des effets de solvant sont aussi d&&s. BARRIERS to nitrogen inversion may be considered as useful probes for understanding structural effects in organic molecules. The rates of nitrogen inversion in alkylamines or in unstrained cycloalkylamines are generally high, and the observation of temperature dependent changes in the PMR spectrum is limited to low temperature. when changes are at all observable. The process may be slowed down by linking electronegative heteroatoms (N,O,Cl,Br,F) to the N atom or including it into a strained cyclic system (aziridines, azetidines). When both effects are present (or two heteroatoms2) nitrogen inversion becomes very slow and stereoisomers arising from slowly inverting nitrogen sites may even be isolated (see Refs in 2 -4). We present here a study of hindered nitrogen inversion in a series of tertiary pyrrolidine and A3-pyrroline derivatives (I-XI) including compounds where the N atom is only linked to C atoms.* Compounds XII6 and XIII” have been included for comparison. l

Previous paper in this series: see reference 1.

t Laboratoite associc au CNRS. l A study of N-chloropyrrolidine has been published recently. Changes in the low temperature p.m.r. spectrum of N-methyl-pyrrolidine have also been reported5.

4227

4220

J. M.

hHN

and J. W~mam

CI

G-i -x

I: II: III: IV: V:

X-CD, x-cl X - Br X-ND, X-OH

\’ m

N-Cl

N--x

VIII: X = CH3 IX: X-Cl

VI: X-CH3 VII: x-cl

,si I

I/

CH,

N,-

-CH,

l

XI

X

C6H,CH,

C6H,C+N

CH, p-cH3 CH,

-CH, CH,

RCH3

CH,

c

H

‘

3

5

N-X

*’

C,H,CH;

GH, XVII

XVIII

XIX: X - NH, XX: X=OH

NMR studks of

rateprocessesand conformations-XVIII

4229

RESULTS

Preparation ofthe substrates. The N-Me (or N-CD3) compounds I and XI have been obtained by reduction of the corresponding N-COOMe derivative with LAH (or LAD). The N-Cl and N-Br compounds II, III, VII, IX, X have been obtained by direct halogenation of the corresponding secondary amines using either sodium hypochlorite or N-chloto (or N-bromo) succinimide The N-ND1 IV amine has been obtained by LAH reduction of the N-nitroso derivative followed by exchange with D,O. Pyrolysis of the corresponding N-ethyl-N-oxide leads to the hydroxylamine V. (for more details see Experimental). variable temperature PMR spectra. All the compounds I-XI (except VIII) display temperature dependent PMR spectra. The most pronounced effects are observed for the CH, groups in a position to the inverting nitrogen site. The two protons of an a-CH, group form an A, (or AA’) system at room temperature and become an AB system at low temperature. Further coupling to vicinal protons leads to more complex spectral patterns. In the case of II (in CHFCl,) and V different signals are also observed for the gem-dimethyl groups at low temperature. The spectrum of II at - 40” and at - 120” is shown in Fig. 1. Free energies of activation AGZ for the observed kinetic process may be calculated for the coalescence temperature T, using the Eyring rate equation AG,+ = 4.57 T,

T

+ log (Av;t, + arh,>+

1

with a transmission coefficient f = 1, and where Av~ and J, are respectively the chemical shift difference and the coupling constant for the protons forming the observed AB system The calculations are based on the parameters obtained for the CH, protons in position 2 for I-V and XI and in positions 2 and 5 for VI-VII-IX-X. In compounds I-V and IX-XI these protons give an AB system at low temperature and an A, system at room temperature (coupling with the other protons is too small to be resolved). In the case of the coalescence of the two Me singlets in II and V, the last term in the above equation becomes log TJAv. Because signals are more or less overlapping and are more or less complex in the various compounds the accuracy in the .determination of the parameters varies from one compound to another. TaDle 1 lists the main results obtained in this study : JAB, AvAB,T, AC:. Results obtained for VII’, XII6 and XIII’” have also been added. DISCUSSION

Free energies and enthalpies of activation for nitrogen inversion. In order to compare the effects of substituents on the nitrogen inversion barrier in compounds I-XIII, enthalpies and entropies of activation AH* and AS’, are required. It is however well documented”” that systematic errors and inaccuracies in the determination of exchange rates by lineshape methods lead to erroneous AH’ and, even mucb more so, AS’ values ; values of the free energy of activation are however much more reliable. On the other hand, it is reasonable to expect that an intramolecular process like nitrogen inversion should have a small AS* term (presumably of the order of 5 eu or less). Large AS* values obtained at various places through the literature are probably due to systematic errors.

4230

J. M.

b -40

L.EHN

and J. WAONJXR

a

d

-40’

FIG. 1. 60 MHz NMR spectrum d N-Chloro-3,3dimcthyI-pyrrolidine II (in CHFCI,): (a) total spectrum at -40”; (b) cnlargad spectrum d the CH,(2) singlet and CH,(S) (triplet) protons at -40°C; (c) same region as in (b) but at - 120”.

Furthermore, when observing the same process in similar compounds, one would not expect AS* to vary very much from one case to another. Thus, for all cases discussed, AH’ values have been calculated from the available AG’ values (generally AGZ at coalescence temp) using values of f 5 eu for AS’ (Table 1). AS+’ values of this magnitude have been found in the racemization of optically active oxaziridinesr ’ (+ 5 eu; XVIII). A complete lineshape analysis of the temperature dependence of the CH, protons in XII leads to a AS* value of - 6.5 eu6. A AS* value of +4*2 eu has been obtained for the epimerixation ofa N-methoxy-1,2_oxazolidine derivative.2* Although the absolute AH* values obtained in this way may be in error by a few kcal/mole, the relative values from one compound to another are expected to be l Recent accurate studies of ring inversion in cyclohexane12 and d internal rotation processes” yield small or nearly zero activation entropies.

also

NMR studies of rate process*l and conformations- XVIII

4231

appreciably more reliable.* When comparing two compounds we admit that it is valid to consider only the AH” values calculated using the same AS“ ( + or - 5 eu) for both The difference in AH* from AS” = + 5 eu to AS* = - 5 is of course smaller the lower the temperature at which AC: is determined. The activation enthalpies may be considered as being equal to the potential barrier for the activation process.4* I4 Substituent eJects on inversion barriers. Table 1 shows that the barriers to nitrogen inversion increase markedly when the nitrogen site is bonded directly to atoms more electronegative than carbon (N,O,Cl). Such barrier raising has also been observed recently in other systems, especially in 3- and 4-membered cyclic molecules (see Ref 2-5 and refs therein). It is seen that the increase A, (where A1 = (AIf’ of the compound considered)--(AZP of the parent compound)) of the inversion barriers within a series of structurally similar molecules is the larger for the more electronegative substituent (see Table 3 ; a supplementary column has been added to Table 3 for A1 factors calculated from AH* values obtained by taking AS’ = 0 eu) Several steric, inductive and electrostatic effects may be invoked when one is trying to trace down the specific substituent properties which give rise to the observed barrier increase.” However the description of the “origin” of such substituent effects on inversion barriers depends on the theoretical language used. Non-empirical quantum mechanical studies of nitrogen inversion m azlridme and oxaziridine16 indicate that the N-O bond certainly plays a role in the increase of the inversion barrier. Semi-empirically, one may say that electron attraction away from the inverting site along a cr bond (i.e. a substituent with a - I inductive effect) leads to an increase in the inversion barrier, in agreement with earlier expectations based on valence bond and hybridization arguments.“** The presence of two heteroatoms linked to nitrogen should lead to an even more important barrier increase; indeed a barrier of 29.2 kcal/mole has been found in a 5-membered cyclic amine, where nitrogen is attached to two 0 atoms.’ Explaining the origin of the effect of an heteroatom is by no means a simple and clear task. For instance the effect of an adjacent N atom is three times smaller in IV than in XII; several factors, other than those one might attribute to the N atom itself, may play a role in differentiating these two compounds (steric factors, rotation about the N-N bond . . . ). Similar considerations apply to the other compounds, so that it may seem wiser in a qualitative or semiquantitative discussion to take the substituent effects as they are without trying to relate them to a specific substituent “property”. Another question to be asked when studying substituent effects is the following: are there structural effects on substituent effects? and if yes, how does. molecular structure affect substituent effects In Table 2 we have collected literature data about nitrogen inversion in a relatively homogeneous series of aziridine derivatives (XIV-XVIII) and we have calculated l Using AS* = f 10 eu rather than f 5 cu changes AH’ values appreciably, but leads only to small changes (a few tenth d a kcal/mole) in the variatiom from one compound to another. l * It should k noted that suds a formulation is purely phtnomcnological Elcctronegativity increae may run paraM to barriu inbut cannot k considaxd as tk ot9gin of the bama increase, which depmds on all typea of interactions present in tk ground and tranaiticlllstata. (see re$.4).

4232

J. M. IZHN and J. WAONER

corresponding AH+ values in a similar way as for compounds I-XII (Table 1) using A.P = +5 eu (see also above).+. In Table 3 we have listed the increase in barrier AZ in a given compound (XV-XVII) with respect to the parent system (XIV). It is then possible to compare A1 and AZ for a given type of substitution in the reasonably comparable series of 5-membered (I, II, XII, XIII) and 3-membered (XIV-XVIII) ring compounds. It is clear from the A1 and AZ values listed in Table 3 that the barrier increase due to a given substituent is much larger in the case of 3-membered cyclic molecules than in 5-membered ones. Furthermore the difference AA = AZ-A1 (Table 3) indicates that going from 5-membered to 3-membered cyclic structures increases the effect of a given substituent by a grossly “constant” factor of UI 6 f 1 k&/mole. Thus, the “5-membered ring + 3-membered ring” structural change leads to a ca + 6 kcal/mole structural efict on substituent e&cts.t This result might indicate that there is an uccumulatioe effect and perhaps even an amplificative-accumdation effect when several barrier increasing factors act on the same nitrogen site; the effect of the second or third structural factor being amplified by the first or the lirst two factors In the case of aziridines, the influence of electronegative substituents may then be considered as being amplified by the strain effect of the 3-membered ring (i.e. by the changes in the various electronic and nuclear interactions brought about by the 3-membered cyclic structure with respect to the 5-membered ring). Similarly, it seems possible that the very high barrier (29.2 k&/mole; A1 = 29.2 - 8.2 = 21 kcal/mole) found for an N-methoxy-1Joxazolidine derivative2 incorporates an amplification by the first 0 atom (as XIII; A1 = 8.5 kcal/mole) of the effect of the second 0 atom (which would amount to: 21 - 8.5 - 12.5 kcal/mole).$ One may then speculate that an N-methoxyoxaziridin (or even more so trifluoroamine NFJ) should have a very high barrier to nitrogen inversion (of the order of 40-50 kcal/mole).§ Using the present results one may also try to estimate an approximate value of the inversion barrier in the recently prepared N-amit~o~~ or in the unknown N-hydroxy (or N-alkoxy2’) aziridines (XIX and XX). The A2 factor for these two compounds are cu 7 (XIX) and 11.5 (XX) k&/mole (Table 3), leading to activation enthalpies of grossly 2627 and 30-32 kcal/moles for XIX and XX respectively. It may also be noted that the barriers obtained for the 5-membered N-Cl II, N-ND, IV, and N-OH V, compounds (Table 1) are very similar to those found for acyclic (AGZ ( - 75”) = 9.7 kcal/mole)?2 N-aminodiamines : N-chlorodibenzylamine benzylamine (AGZ ( - 95’) = 8.5 kcal/mole)23 and N-methoxybenzylmethylamine E. = 129 kcal/mole)“ll the

) In the case whem large AS*vaIua have bear reported the activation parameters have been recalculated using AS” = *5 eu Another N-chloro-a&dine derivative has been studial recently,‘s but in this cast the presence d additional steric effects is expected to lower tht inversion barrier. t One could think that the structure effect on substituent effects might be proportional to the barrier height, but the results an not accurate enough to permit such a conclusion to be drawn. t Other structural features may howeva also contribute to this effect.’ 8 A barrier of 5659 kcal/mole has been estimated for NF3.19

1In tk last two cases where hindered rotation about the N-N the results have beai interpreted in terms of nitrogen inversion.

or N-O

bonds might also be operative,

CHFCI,

II

sienal

CHA2)

VI

CHFCI,

DzO

CH,(2) (-125)

CHFCI,

IV

t-129

‘% CH,W5l

CH,P) f-8)

CH,M (-58) CH,

CH,(2) (-116)

(-116)

CFCI,

V

DATA

?I&)

CH,(2) f-130) CH,(2) (-1m

(at T”C)

ObSCrVd

III

CFCl,

CHFQ

Solveat

I

Compound

TAB~B 1. SPECIIWL

1 I.6 f 02

-.

9-8 f 02

+7*4 -107 f 3 69*2

6-8 f 02

f 3 30*5

-25

-11*3

-98+3

+ 3

f 3

f 4

329 f 04

55 f 01

27.5 f 04

356 f D5

8.4 f 02

-87*3

367 f 0.5 -98

-97f3

2-7 f 01

355 f 1.5

-79

-117

T.

“C

INVERSION

366 f OS

54.3 f 1

Hz

FOR NITRODBN

AVAB

PARAMTEM

m*1

8.7 f 03

Hz

JAB

AND ACTWAlTON

AH+

104 99

95 f 02 9.0 f 02

14-9 f 03 79 f 02

15.0 f 04

13.2 f 02

13.0 f 02

8.5 f 02

163 8.7

165

144

143

94

34

104

9.4 f 02

8.5 f 03

8-2

(As’= +5cu) kcal/molc

I-XIII

7-4 f 03

AG:

IN CYXfPOUNDs

13.5 7.1

13.5

129

(A&v= -5cu) kcal/mok

AH’

!z

_

H ,$

P

3

CFCls

CHFCl,

CH2’J2

cm3

X

XI

XII<

XIII’

CH2@

CH,(3J) (-68)

CHJ2.5) CHJ2.5) CHd2.5) (-120) CH,(2+) (-110) CH,(2) (-120)

-

9.1 f 0.1

9.7 f o-2

146 f 04

14.5 f 05

407 f 0.5

79f05

98 & 05

9.8 f 07

26.3

156 f 0.5

42”

+

11.1 f 0.2

86 ho.2

&8 & 0.2

8.5 & @4

104

- 45”

-1oof3

-92f3

-98fS

-64” b

* !kcRd7a.

’ See ReE6. Complete l&shape analysis Aff* = 9.5 kcal/mole; IL‘S’= -6.5 eu.

* !SeeRd5.100 MHz spectra witb doubk irradiation. Complete lineshape analysis gave E, = 134 kcal/mok and log A = 144 (-A$ * No line broaden&g observed down to - 135”.

ma3

CF,a, CHFCl,

VII’ VIII IX

= 17-5eu).

17.2

12.2

9.5

9.7

99

11.4

14.0

10.0

1.7

74

7.6

9.4

% 2;

fi z e a -

NMR studies of rate prowsses and confommtions- XVIII

4235

(XIII)

(IV) (V)

:o

-NH, -OH

1.2 6.2

lJ.4’ 1.1 5.7

7.r

3.7

0 21)

0 2.2 4.0

0

+ku

Ai kcal/mole

19 5.4

6Q

3.4

0 1.8

-5eu

(xIXY (xxy

(XVII)

(XVI)

(XIV) (XV)

INVERSION

7.2 12.2

14.5

9.0

0 I.5

+5eu

Ah:kcal/molc

NITROGEN

Compound]AS*= .’

3. STRWTURALEFFSCISON

7.1 Il.7

140

90

0 I.5

0

BARRIERS

70 il.4

13.5

&5

0 7.5

-5eu

As*=

-

6‘0

5-O

0 5.3

+5eu

-

63

5.3

0 5.5

0

AAt kcal /mole

1 Corrected foa the absent of the gcmdimetbyl groups in XIII as compared to I by subtracting @5 kcal/moie. * Obtained from A, = AA f A, (AA _ 6 kcal/mole; sa text).

a Change in AH”with respod to the N-CH, compound I; A1 = AHTX) -82 or A, = M(X) -66. * Change in AW with respect to the N-&H, compound XIV; A, = A?f’Qt) -200 or A2 = W(X) - 174. ,1Oaiisusad,thebbvalutschangebyltss~anO5 c change in AH* from the five-membered ring to the three-membered ring series; AA = A, -A,.IfA.P= kcal/mole.

(XII)

:N-R

-

(II)

(I)‘-

Compoun+.SC=

-Cl

4H3

Substituent on Nitrogen

TABLE

-

66

5.1

0 5.1

-5eu -.--

i 4

L

z r

NMR studies of rate processes and conformations--XVIII

4231

Steric eflects. It is seen that the presence of a double bond in the A’-pyrrolines (VIII-IX) lowers the nitrogen inversion barrier (by cu 2 k&/mole in IX) with respect to VI and VII. The same holds for the benzo derivative X (Table 1). In addition to a possible homoconjugation of the nitrogen lone-pair with the n-system, this barrier lowering may arise from the negative strain effect due to the introduction of a double bond in the 5-membered ring: opening of the C2-C3-C4 and C3-C& angles from tetrahedral (in VI, VII) to 120” (in VIII-X) also tends to open the CI-N--C3 angle and thus to bring the stable pyramidal state nearer to the planar (C2-N-C5 w 120”) transition state. The gemdimethyl group in I-II lowers the inversion barrier with respect to VI-VII by ca 05-l kcal/mole (Table l), presumably because of steric interaction with the substituent on nitrogen which destabilizes the pyramidal state with respect to the transition state. The slight barrier increase (of the order of 05 kcal/mole; Table 1) observed for XI with respect to VI might arise from a pinching effect due to the smaller C-S-C angle (cc 100°-1050) as compared to the tetrahedral C-C-C angle, although several other “effects” may be operative. Solvent effects. It may also be noted that the inversion barrier of II determined in CHFC& is cu 05 k&/mole higher than in CFCl,. This barrier raising effect is similar to that observed for chloroform solutions as compared to more inert solvents.25 Thus, in CHFC&, IX and X would probably have AGZ values of 90 and 9.3 kcal/mole respectively. Changing from organic (CDCl,) to water solutions leads generally to a marked increase in nitrogen inversion barrier. 25 This is also seen in the case of compound V ; changing from CDC13 to D,O solutions brings the coalescence temperature from - 11”to + 30”, which corresponds to a barrier increase of co 2 kcal/mole.* A 1.3 k&mole solvent effect has also been observed for XIII.“’

EXPERIMENTAL M.pe were obtained on a Kofler block and are uncorrected. Elemental analyses were performed at the Strasbourg Divisicu of tk microanalytical laboratory of tk Centre National de la Recherche Scientihque. NMR spectra were measured with a Varian A-60 spectrometer equipped with tk variabk temp accessory. The t&ripsdown to - 50” were measured by means of tk Varian MeOH and glycol samples and shift-temp correlaticlncharts; tk ten\ps lower than - 50”have becn meas.uredusinga copper-constantan thermocouple inserted directly in a spinning sampk tuk; their absolute values am correct to within f2”, Both techniques give tk sark temp within f 1” in their common range Spectra calibration was e&ted at every temp with a Hewlett-Packard 200 CD wide-rans oscillator and a Hewlett-Packard 5212 A frequency counter. During tk recording of tk spectra a low enough radiofrquency field was used so as to avoid differential saturation effects. Chemical shifts are in ppm downfield from internal TMS Coupling constants are in Hertz (Hz). The following symbols are used fmdescribing tk NMK spectra:s: smgtet; d: doublet; 1: triplet; m: multiplet; b: broad. Dimethyl-3,3-succinimfdewas prepared according to tk literatures3 from 2,2dimethylsuccink acid and aqueous ammonia; m.p. 105”(lit.” mp. 106”) Dimethyl-3,3-pyrrolidinewas prepared by tk general method l When sucir a solvent eNect is takm into account tk inversion barriers measured here for I and VI agree reasonably well with those obtained by Holtzman and Saundersz6 in water solution for some Smembered cyclic amine& using tk method developed by Saunders and Yamadas’

4238

J. M.

LWN

and J. WAGNER

of reductiou d the previous imide with LAH in diethyl ether soln The product, obtained as colourless liquid, was dried over molecular sieves and distilled, b.p 100-105°, yield 70%; NMR spectrum (CDCI,, 34’); NH, s, 6 1.97; CHr(2k $ d 258; C&H,), s.6 1.05; CH,(4), t, d I%, J = 7.25 Hz; CH,(S), f 6 2.97, J = 7x25 Hz (Found: C, 72.83; H, 1302; N, 14.14. Calal for C,H,,N: C, 7266; H, 13.21; N, 14.12%)

Compound I was obtained by LAD reduction in diethyl ether soln of the N-carbometboxy-3,3dimethylpyrrolidine which was prepared by treating 3,3dimethylpyrrolidine (3 9) with methyl chloroformate (4 g) in metbanolic Na,CO,aq (HIO/MeOH = l/25; 6 g suspended Na,CO,) The carbomethoxy derivative was purified by distillation (b.p. PS-9?“/13 mm, yield 8@$ and displayed the following siguak in the NMR spectrum (CDCl, 34”): COOCH, $6 3.66; CH,(2& d, 6 3-11; C(CH,), s, 6 1%; CH,(4k f 6 1.63, J = 7.0 Hz;CH,(f~m,63~3.(Found:C,61~14;H,944;N,P~l2CalcdforCsH,,NO~:C,61~12;~9~62;N,~91~). Compound I was then obtained by treating the carbomethoxyadduct (1.5 8) with LAD (04 gl in diethyl etha sola (30 ml) It was distilled from Na and stored over molecular sieves (yield 60%); NMR spectrum (CD& 34”): CHr(2), s d 233; C(CH,),. s, 6 la; CH,(4), f S 1.62 J = 70 Hz; CH,(5), f 6 2.62, J = 70 I-Ix. N-chforo-3.3-df~~yfpyrroIidine (II) was obtain& by treating 3,3_dimethyl pyrrolidine with NaOCl according to the general method. 33 The product, obtained as a colourleas liquid, was puriftat by ~tillation under reduced press and at a temp ~60”; NMR spectrum (CDCI, 24”): CH,(Z), a 6 2P6; C(CH,), s, 6 1.13; CH,(4), t, 6 1.67, J = 715 Hz; CHs(5k t, d 3.22 J = 7.15 Hz (Found: C, 54.26; II, 9.05; N, 1063. Calal for C,H,rNCI: C, 53.93; II, P’X; N. 1048%). N-Eromo_3,3dfmethyfp~pyrrolidilce(III) was obtained from 3,3dimethylpyrrolidiue using N-Rromosuccinimide in abs diethyl ether and cooling to - 1Q -1s”. The soln was filtered and the ether evaporated whik cooling; then the resulting mixture was extracted with CFCI, at a temp ~0”. The compound was too unstable for analysis and decomposed at temp >O”; NMR spectrum (CFCl, -40”): CH,(2), s, d 290; C(CH,), s, 6 1.12; CH,(4), t, 6 154, J = 70 Hz; CH,(Sk t, d 3.17, J = 7a I-Ix. N-Amino_3,3dimerhyfpyrroffdine (IV) An aqueous soln of 3.3dimethylpyrrolidim: was treated with NaNOs and AcGH to produce the NNitroso-3,3dimethylpyrrolidin+ according to the method of Marckwald.34 The prodti obtained was purified by distillation to give a yellow oil (bp. 102-RN”/13 mm, yield 70!!& NMR spectrum (CDt&, 34”) : CH,(2), two f 5 333; 6 399, J = 1.2 Hz; C(CH3)r, two s, 6 l-14; 6 1.15; CH,(4), two m, centered 6 1.87; CH,(Sk two m, d 3.65; two rotamers, 38% and 62% (Found: C 56.42; Ii, 9.30; N, 21.74. Calcd for C,H,,N,O: C, 56.22; II, 9.44; N, 21.86%). The N-Nitroso-3,3dimethylpyrrolidine (3.3 g) was then reduced with LAH (1.2 g) in abs diethyl ether soln. The product obtained, N-amino-3,3dimethylpyrrolidiue distilled as a colourless liquid (b.p 80-82”. 80 mm, yield 75%); NMR spectrum (CDCl, 34”): N-NH,, b s, d 312; CH,(2), s, 6 2.47; C&H,),, s 6 l+JP; CHs(4k t d, d l-59, Jt = 7.3 Hz; J2 = Q8 I-Ix; CH,(5), t d, S 2.75, J’, = 7.3 Hz; fr = 09 Hz (Found: C 63.17; II, 12.30. Calcd for C6HIINI: C, 63.11; H. 12.36; N, 2453%). The Ndeutero product was obtained by treatment with D,O, extraction with CHsClr, drying over Na2S0, and molecular sieves; NMR spectrum (CHFCI,, -10”): CH,(2), s, d 2.44; C(CH,)s. s, 6 l-08; CH,(4), f S 1.58, J = 73 Hz; CH,(5), t, 6 2.71, J = 7.3 Hz N-hydroxy-3.3dfmetftyfpyrroffdins (V) was prepared according to the method described3’ for the preparation of N-hydroxypyrrolidine N-ethyl-3,3_dimethylpyrrolidine was treated with H,O, and the N-oxide obtained was pyrohxed at 145-150”~15mm The crude product obtained was further purified by GLFC on a 3’ Carbowax B M column (10%) on chromosorb W at 7O-809 The product oollectuI by GLPC was redistilled and displayed the following signals in tbe NMR spectrum (CDC13. 34”): CH,(2k s d 28: C&H,),, a d 1.1; CHz(4b t, 6 163, J = 74 Hz; CHr(5I t, 6 397, J = 79 Hz (Found: C, 62.88; Ii, 1l-30; N, 12.33. Calcd for CeH,,NO: C 62.57; II, 11.58; N, 12.16%). N-methyfpyrrofidine (VI) is a commercial product and was redistilled and dried over molecular sieves before use. N-C~loro~y~of~i~ (VII) was prepared from pyrrolidine in the same way as II. N-Metiryf-A,-pyrrofine (VIII) was preparai according to the literature;‘s b.p. 78-80” (lit.37 b.p. 79-80“). N-chforod,-pyrroffne (IX) was prepared by treating A,-pyrroline (Eastman) in diethyl ether soln at - 10, - 15”, with Nchlorosuccinimide in the same way as described for III. The product was too unstable to be

NMR studies of rate processes and conformations-

XVIII

4239

isolated in pun form and analyzed and was obtainai as a solution in CFCl,; NMR spectrum (CFCI,, -40”): CH,, s, 6 3.87; CH, s, d 5.83. N-Chlorodihydroisoindondole (X) Dibydroisoindok (isoindolinc) WBSprepared according to tk literature” (b-p. lOO-102/15 mm (lit.‘* b.p. 11So/30 mm)). The N-chlorodcrivativc was obtained by chlorination d tbcisoindoline with N-&lore succinimidc in the SBM way as for IX. It was also too unstabk to be isolatal in pure form and analysed; NMR spectrum (CFCl,, -20”): CHI, s, 6 431; CH, s, 6 7.11. 344ethylthiazolidine (XI). N-carbomethoxythiazolidinc was prepared by reacting methyl chlorofonnate with thiaxolidine in benzene solo in presence d triethylaminc; the crude produd obtained showed an NMR spectrum in agreement with the structure ((CDCl,, 34’): CH1(2), s, b 444; COOCH,, s, 6 3.71; CH,(4), t, d 3.71, J = 6.2 Hz; CH,(5), t, d 2.97, I = 6.2 Hz) and was used without further purification. Reduction with LAH in dicthylether yielded XI, b.p. 6>70”/60 mm (lita b.p. 151-152”); NMR spectrum (CDCI,. 34”): CH,(2), s, d 403; N-CH,, s, b 2.31; CH,(4wH,(S), m, S 2.97.

REFERENCES ’ A Feigcnbaum and J. M. Leha Bull. Sot. Chfm Fr. 3724 (1969) ’ K. Milller and A Eschenmoser, Helo. Chin Acta 521823 (1969) ’ ” 1. M. L&n and J. Wagner. Chem Comm 148 (1969); ’ S. J. Brois, J. Am Chem Sot. 90,504 508 (1968) * J. hi. Lehn, Forzschr. Chem Forsch, Springa Vcrlag in press (1970). ’ J. B. Lambert and W. L Oliver, Tetrahedrorr Letters 6187 (1968) 6 B Dietrich, J. M. Lehn and P. Linscheid, unpublished results ’ ” F. G. Riddell, J. M. Lehn and J. Wagner, Chem Comm 1403 (1968); J. M. Lehn and J. Wagner, unpublished results; b D. L Griffith and B. L Olson, Chem Comm 1682 (1968); c M. Rabat F. B. Jon- Jr, E H Carlson, E Banucci and N. A L&cl, to be published s A Allcrhand, F. M. Chen and H S. Gutowsky, J. Chem Phys. 4% 3040 (1965) ’ A Allerhand, H. S Gutowsky, J. Jonas and R A Me&r, J. Am Ckm Sot. 88.3185 (1%) lo G. Bins& Topics in Stereochemistry (Edited by E. L. Eliel and N. L Allinger) Vol 3; p 97. Interscience, New York (1968) ” F. Montanari, I. Moretti and G. Terre, Chem Comm. 1086 (1969) I2 F. A L Anet and A. J. R. Bourn, 1. Am Ckm Sot. 89,760 (1967) I3 A Mannschreck, A Matthcus and G. Rissman, J. Mel Spectry. 23, 15 (1967); A Jaeschke. H Muensch, H G. Schmid, H Fricbolin and A Mannschreck, Ibid 31, 14 (1969); M Rabinovitz and A Pines, J. Am Chem Sot. 91, 1585 (1969); K. I. Dahlquist and S. Forsen, J. Phys Ckm 4124 (1969) ‘* S Glasstone, K J. Laidlcr and H Eyring Tk Theory 4/ Rate Processes, pp 194, 198 McGraw-Hill, I&w York, N.Y. (1941) ” F A L Anet, R D. Trcpka and D. J. Cram, J. Am Chem Sot. 89, 357 (1967) I6 J’. M. Lehn, B. Munsch, Ph. Milli& and A Veillard, 7heorer CMm Acta 13, 313 (1969); Pk Millit. B. L&y, J. M. Lchn, B. Munsch, to be published. I’ H A Bent, Chem Rev. 61275 (1961) ‘s D. Felix sod A Eschenmoscr. Angew. Chem 80,197 (1968) I9 G. W. Kocppl, D. S. Sagatys, G. S Krishnamurty and S. L Miller, J. Am Chem Sot 89. 3396 (1967); corrected value for NF3 obtained from the authors ” S J. Brois, Tetrakdrcn Letters 5997 (1968) 21 S. J. Brois. J. Am Ckm. Sue.. 92, 1079 (1970) ” J. M. L&I and J. Wagner. unpublished results ” M. J. S. Dewar and B. Jcnningg 1. Am Ckm Sot 91.3655 (1969) ” D. L Griffith and J. D. Roberts Ibid 87,4089 (1965) ” J. E. Anderson and J. M. Lchn, Ibid 89.81 (1967) 26 L Holtzman and M. Saunders, private communication 27 M Saunders and F. Yamada, J. Am Ckm. Sot. 85.1882 (1963) ** M: Jautclat and J. D. Roberts, Ibid. 9l, 642 (1969)

42.40

J. M.

&HN

and J. WAGNER

a’ R G. Kostyanovsky, Z E Samojlova and L L Tchcrvin, Tetrahedron Letters 719 (1969) ” A Mann&reck and W. Seitq Angew. Chem 81,224 (1969) 3’ A Maxmschreck, J. Lins and W. Scitz, Lit-b@ Ann 727,224 (1969) ” S. S G. Sircar, J. Chem Sot. 1252 (1927) ” A F. Grack and R E Meyer, J. Am Chom Sot So, 3939 (1958) ‘* W. Marckwald and Alb. Frhr. v. DrostcHueLsho~ Chen Ber. 3& 3261(1898) ” J. Thesing and W. Sirrcnbcrg lb&i. 9& 1749 (1959) 36 J. M. Bobbitt, L. H. Amundscn and R L Steiner, J.‘Org. Chem 25,223O (1960) ” G. Ciamician and P. Maguaghi, Chem Ber. 18,725 (1885) ‘a J. Born&n, S C Lashua and A P. Boissel, J. Org. Chem 22,125s (1957) 39 E D. ~Bcrgmaxmand A Kaluszyner, Rec. ‘Ifou. Chtm 78,289 (1959)

Note added inproofThe free energies of activation at coalescence temperature haw btcn reported recently for N-mcthylpyrrolidisc (J. R Lambert and W. L Oliver, jun., J. Am Chem Sot. 91.7774 (1969)) and for methyldibcnzylaminc (M. J. S. Dewar and W. B. Jennings, Tetndwdroll Letters 339 (1970); C. H Bushweller and J. W. O’Ncil, J. Am Chem Sot. 972159 (1970)).