Cycloadditions of two thiocarbonyl ylides with α,β-unsaturated esters and nitriles: Steric course and mechanism

TETRAHEDRON

Pergamon

Tetrahedron 58 (2002) 507±519

Cycloadditions of two thiocarbonyl ylides with a,b-unsaturated

esters and nitriles: steric course and mechanism^q

²

³

Rolf Huisgen,^pGrzegorz Mloston, Henry Giera and Elke Langhals^§

È È È

Department Chemie der Ludwig-Maximilians-Universitat Munchen, Butenandtstr. 5-13 (Haus F), D-81377 Munchen, Germany

Dedicated to Professor JuÈrgen Sauer on the occasion of his 70th birthday

Received 24 September 2001; accepted 5 November 2001

AbstractÐIn the probably concerted cycloadditions of the sterically hindered thiocarbonyl ylide 1 with fumaronitrile, maleonitrile, and

dimethyl fumarate, the dipolarophile con®guration is retained whereas retention/inversion 99:1 for dimethyl maleate (51 times less reactive

than fumarate) signals a small involvement of a two-step pathway. The latter becomes dominant when two acceptor groups stabilize the

anionic terminus of a zwitterionic intermediate. Nonstereospeci®c cycloadditions of 1 with dimethyl 2,3-dicyanofumarate (16, retention/

inversion 60:40) and dimethyl 2,3-dicyanomaleate (17, 76:24) were observed. Special conditions were required to avoid a preceding cis,

trans isomerization, 16Y17, catalyzed by thiadiazoline 4, the precursor of 1. In the case of the related thiocarbonyl ylide 37, this catalysis

1. Introduction

last decades on borderline crossings, i.e. on the promotion of

the two-step process by suitable choice of reactants. The

model systems can be targeted on the formation of diradical

or zwitterionic intermediates. The experimental approach is

based on the interception of intermediates, violation of

stereoretention for cis-, trans-isomeric substrates, and on

kinetic phenomena.

Concerted formation of two new s-bonds vs. two-step reac-

tion via an intermediate is a fundamental question of

cycloaddition chemistry. Experiment and theory support

the conclusion that the majority of Diels±Alder reactions

(DAR)²and 1,3-dipolar cycloadditions^3ause the concerted

pathway. Since the advantage of the concert in terms of

activation energy is limited, much effort was spent in the

In the framework of Sustmann's reactivity model,⁴we

Scheme 1.

^qSee Ref. 1.

Keywords: 1,3-dipolar cycloadditions; thiocarbonyl ylides; stereospeci®city; zwitterionic intermediates.

Corresponding author. Tel.: 149-89-2180-7712; fax: 149-89-2180-7717; e-mail: rolf.huisgen@cup.uni-muenchen.de

p

²

Present address: Section of Heteroorganic Compounds, University of Lodz, Narutowicza 68, PL-90-136 Lodz, Poland.

Present address: Fachhochschule MuÈnchen, Fachbereich 05, Lothstr 34, D-80334 MuÈnchen, Germany.

Present address: Consortium fuÈr Elektrochemische Industrie, Wacker Chemie GmbH, Zielstattstr. 20, D-81379 MuÈnchen, Germany.

³

§

PII: S0040-4020(01)01147-4

508

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

regioisomeric cycloadducts

7 and 8 were formed

(Scheme 2).¹¹The ratios, 7a/8a65:35 and 7b/8b80:20,

re¯ect the nucleophilicities at both termini of 1. In

accordance with a concerted addition, no polymerization

of the vinyl monomers was observed.¹¹The cycloadditions

of thiocarbonyl ylides 9 and 10 with methyl acrylate and

acrylonitrile furnished exclusively the 3⁰-substituted

thiolanes, corresponding to 7.^12,13Thus, the higher steric

demands of 1 enforced some formation of 8.

In the reaction of 1 with methyl methacrylate as solvent, the

cycloadduct (3⁰/4⁰-substitution now 20:80) was accom-

panied by 2.6% of polymer.¹¹Thanks to the ampli®cation

effect associated with polymerization, very small quantities

of intermediates in cycloadditions can be elegantly detected

by this trapping technique.¹⁴

Scheme 2.

studied reactions of nucleophilic 1,3-dipoles (high

p-HOMO energies) with electron-de®cient dipolarophiles

(low p-LU energies) in the hope of stabilizing a zwitterionic

intermediate. Thiocarbonyl ylides⁵are closely related to the

allyl anion, and tetra-acceptor-substituted ethylenes were

chosen as dipolarophiles. It turned out that a further con-

dition has to be ful®lled to divert the cycloaddition from the

orthodox concerted pathway: massive steric hindrance at

one terminus of the 1,3-dipole.⁶

When 4 was reacted with dimethyl fumarate (1.1 equiv.) in

abs. THF at 408C (N₂elimination with t_1/286 min) for 8 h,

the ¹H NMR analysis with weight standard indicated 95% of

the trans-cycloadduct 11 (Scheme 3). The signals of cis-

adduct 12 were missing in the ¹H NMR spectrum

(500 MHz), although comparison with the ¹³C-satellites of

11 would allow to detect as little as 0.03% of 12. Thus, the

retention of trans structure amounts to .99.97%.

The easily available and storable 1,3,4-thiadiazoline 4 is the

precursor of thiocarbonyl ylide 1. A preceding paper dealt

with the capturing of the 7-membered cyclic ketene imine 5

in the reaction of 1 with tetracyanoethylene (TCNE); 5 was

in situ converted by water to the lactam 6 (Scheme 1) and

by methanol to the corresponding methyl imidate.^7,8The

zwitterionic intermediate 2 enters into the reversible

formation of 5 and the irreversible closure of the thiolane

ring. Products 6 and 3 were obtained in 65:35 ratio.

cis-Disubstituted ethylenes are weaker dipolarophiles than

the trans compounds.^3bIn the product obtained with

1.1 equiv. of dimethyl maleate, 27% of thiirane 15, formed

by the electrocyclic ring closure of 1,¹⁵signaled the incom-

pleteness of the interception with the dipolarophile. The

cycloadduct consisted of 12 (70%) and 11 (,1%).

The commercial sample of the liquid dimethyl maleate

contained 0.34^0.03% of dimethyl fumarate. By cyclo-

addition with 0.12 equiv. of diphenyldiazomethane (k_trans

/

Here we report on the steric course of the 1,3-cycloadditions

of thiocarbonyl ylides 1 and 37 with electrophilic cis-, trans-

isomeric dipolarophiles. The results of two preliminary

communications^9,10are supplemented with new experi-

mental material.

k_cis36)¹⁶and subsequent distillation, the fumarate content

is expected to be reduced to 0.004%. When this puri®ed

maleate sample (neat, 3.4 equiv.) was reacted with 4, the

¹H NMR analysis (500 MHz, ¹³C-satellite technique)

revealed cis/trans12/1198.9:1.1. The minor violation

of the dipolarophile retention suggests some involvement

of a pathway via an intermediate which is capable of

rotation.

2. Reactions of 2,2,4,4-tetramethyl-3-thioxocyclo-

butanone S-methylide (1) with mono- and di-acceptor-

substituted ethylenes

The cycloaddition of 1 to a mixture of dimethyl fumarate

and dimethyl maleate provided the competition constant,

k51, i.e. 1 reacts 51 times faster with fumarate than with

maleate. Recently we reported k65 for the same trans/cis

pair vs. 9 and discussed the mechanistic signi®cance of such

relative reactivities.¹²Competition experiments of fumaro-

nitrile and maleonitrile for 1 furnished k2.6. The different

order of magnitude suggests the importance of steric

hindrance for the diester pair.

When the not isolable 1 was set free from its precursor 4 in

an excess of methyl acrylate or acrylonitrile as solvent, the

Both fumaronitrile and maleonitrile are crystalline and easy

to purify. The fumaronitrile adduct 13 was obtained in 93%

yield without 14 becoming visible.

The recrystallized sample of maleonitrile contained

0.058^0.006% of fumaronitrile. The preparative reaction

with 4 (benzene, 408C) furnished 88% of the cis-adduct

14 and 11% of thiirane 15. In the subsequent test on

Scheme 3.

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

509

Table 1. Isomerization of 17 (0.11 M) to 16 (equilibrium 17/1612:88) and

of 19 to 18 in CDCl₃at room temperature; catalysis by thiadiazolines

(0.10 M) and its inhibition with acid (7.6 mM H₂SO₄in CDCl₃as solvent)

stereospeci®city, 4 was reacted with 2 equiv. of maleonitrile

(neat), and the H NMR analysis as above afforded 13 and

1

14 in a ratio of 0.13:99.87. The fumaronitrile content

(0.058%) of the dipolarophile entered the competition

with k2.6 and was responsible for 0.080% of 13 (based

on 4). Thus, the retention in the cycloaddition with dimethyl

maleate amounts to 99.95%, and the residual 0.05% is

insuf®cient evidence for a second pathway.

Thiadiazoline

Time (min)

17/16

19/18

0

60

250

0

70

270

4

98:2

91:9

98:2

89:11

98:2

84:16

92:8

39:61

81:19

98:2

63:37

98:2

54:46

96:4

41H₂SO₄

21

211H₂SO₄

22

221H₂SO₄

231H₂SO₄

The spectra of 11±14 are consistent with the structures. The

cyclobutanone carbonyl gives rise to IR bands at 1775±

96:4

95:5

75:25

93:7

92:8

61:39

77:23

1

1791 cm²¹and d (¹³C) 216±219. The H signals of the

a

four ring protons of 11 and 13 are resolved at 400 MHz,

and the assignments rest on a synopsis of the parameters for

a multitude of thiolanes. Four different ¹H and ¹³C signals of

C±Me are the consequence of chirality. In the mass spectra,

the elimination of dimethylketene accounts for the base

peaks. Both fragments can appear as radical cations, as

93% of m/z 70 (C₄H₆O¹) for 13 and 81% for 14 testify.

Further losses of CO₂Me, HCO₂Me and HCN, respectively,

are noticeable and possibly lead to thiophene radical cations.

No reversion of the original cycloaddition was observed.

a

16 crystallizes.

catalyzed the slow equilibration, and 16/1788:12 was

attained in 3 d from both sides.¹²We were alarmed by

some cis, trans isomerization, 16Y17, which occurred

before the cycloaddition with 1. It turned out that the pre-

cursor of 1, the spirothiadiazoline 4, catalyzed the isomeri-

zation of the dipolarophile at room temperature, i.e.

conditions where the formation of the thiocarbonyl ylide,

4!11N₂, was still slow. The conversion 17!16 in 0.1 M 4

in CDCl₃reached 9% in 60 min and 19% in 250 min

(Table 1) while no isomerization was noticed without 4.

3. Reactions of 1 with 2,3-dicyanofumaric esters and 2,3-

dicyanomaleic esters

The catalytic activity of several thiadiazolines increased in

the sequence 4,21,22,23 (Scheme 4) and showed no

correlation with the rate of N₂elimination. The thioether

function of the cycloadducts exerted no catalysis. The

reversible formation of a type 20 complex appears to be

responsible for the isomerization 16Y17. Experiments

with 2,3-bis(tri¯uoromethyl)maleonitrile, i.e. another tetra-

acceptor-substituted ethylene, established the catalytic

activity of 1-pyrazolines and other cyclic azo compounds.¹⁷

When the dipolarophile bears two electron-attracting sub-

stituents at the same C-atom, the stability of the zwitterionic

intermediate takes a quantum jump. Such intermediates

were intercepted in the reactions of 1 with TCNE⁸and

arylidenemalononitriles.¹¹The cycloadditions of 1 with

dimethyl 2,3-dicyanofumarate (16) and dimethyl 2,3-di-

cyanomaleate (17) are nonstereospeci®c. However, to

quantify the inversion during the cycloaddition, it has to

be checked whether cis, trans isomerization occurs before

or after the cycloaddition step.

We observed that strong acid inhibited or even prevented

the thiadiazoline catalysis, and found that 0.0076 M sulfuric

acid in CDCl₃was optimal. In this medium 0.11 M cis-

diester 17 was not isomerized in the presence of 0.10 M

thiadiazoline 4 or 21 (Table 1); thus, 7 mol% of H₂SO₄

preserved the stereostability of 17. Less complete was the

protection in the case of thiadiazoline 22 and it failed for 23

which differs from 4 by the absence of the carbonyl group.

The tetra-acceptor-substituted ethylenes 16 and 17 are fairly

thermostable, but sensitive to base: suspended KF in CDCl₃

The diisopropyl esters 18 and 19 were found to be resistant

to the catalysis by thiadiazoline 4 (Table 1) whereas 22 and

23 still initiated some cis, trans isomerization, 18Y19. The

pair 18/19 was included in our study of the steric course of

cycloaddition.

A higher reaction temperature (808C) helped in curbing the

detrimental catalysis of cis, trans isomerization. The thia-

diazoline 4 enters into competing pathways: bimolecular

catalysis of the process 16Y17 and unimolecular extrusion

of N₂. The rate constant of the latter is expected to rise faster

with increasing temperature.

The reactions of 4 with 16 and 17 produced mixtures of

trans- and cis-thiolanes, 26 and 27, which were dif®cult to

separate, but modest amounts were obtained pure (Scheme 5).

The d_Hof the diastereotopic 5⁰-H₂offered an empirical

Scheme 4.

510

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

Scheme 5.

criterion of trans, cis assigment: 3.56 and 3.64 (J 11.4 Hz)

for 26 (trans), and 3.43 and 3.87 (J12.1 Hz) for 27 (cis).

The AB spectrum (small Dd) was observed for all trans-

adducts of 16 and 17 and the AX type for all cis-adducts

obtained with the thiocarbonyl ylides which were formed

from thiadiazolines 4, 21±23 and several more. For one of

those cis, trans pairs X-ray analyses proved the structures.¹⁸

Obviously, the thiolane conformations of trans- and cis-

adducts are only little in¯uenced by the substituents in

2⁰-position. The diisopropyl esters 28 and 29 follow the

same rule: Dd0.06 for the trans and 0.47 ppm for the

cis-structure.

thermodynamic stability (see later) of the cis-cycloadducts.

The lack of cis, trans isomerization in the unconsumed

dipolarophile indicates that the zwitterions 24 and 25 do

not dissociate back to the reactants (Scheme 5).

The cis, trans ratios of Table 2 are results of kinetic control.

The cycloadducts 26±29 were stable under the reaction

conditions. No stereoisomerization of trans-adduct 26 was

noticed in benzonitrile at 1398C (7 h).

The last column of Table 2 refers to a competition

experiment of 16 and 17 for thiocarbonyl ylide 1. The

evaluation took care of the isomerization rates, which

were observed in separate experiments. It is no surprise

that kk₁₆/k₁₇4.7 is smaller than for the pair dimethyl

fumarate and maleate, k51. Competition constants of

5.6 and 65 were found for the reactions of thiobenzo-

The stereochemical tests were run in 7.6 mM H₂SO₄in

CDCl₃and ¹H NMR analyzed without workup in the

presence of a weight standard (Table 2). In the experiments

at 808C (10 min), the excess of 16 and 17 was not iso-

merized, and the product ratios were virtually identical for

dimethyl esters and diisopropyl esters:

Table 2. Reactions of thiocarbonyl ylide 1 with dimethyl 2,3-dicyanofu-

marate (16) and dimethyl 2,3-dicyanomaleate (17) in 7.6 mM H₂SO₄in

CDCl₃(1 mL); ¹H NMR analysis of steric course

trans/cis-Adduct

Dimethyl ester 16 (trans)

Dimethyl ester 17 (cis)

Diisopropyl ester 18 (trans)

Diisopropyl ester 19 (cis)

26/2760:40

26/2724:76

28/2961:39

28/2925:75

Experiment

1

2

3

4

5

4 (mmol)

20.3

26.3

23.3

34.7

20.9

100.1

57.5

29.0

76.1

80

10 min

26

33

44:46

25

6

90

16 (trans) (mmol)

17 (cis) (mmol)

Temperature (8C)

Reaction time

26 (trans) (%)

27 (cis) (%)

32.7

40

10 h

16

47

25:75

23

1

87

118.9

80

10 min

21

66

24:76

4

95

The preponderance of retention over inversion shows that

rotational equilibrium of the assumed zwitterionic inter-

mediates 24 and 25 was not reached. As a consequence of

24Y25, the ratio k_cycl/k_rotis smaller than the ratio of

retention/inversion. The inversion was higher in the experi-

ments with the trans-dipolarophiles 16 and 18 than in those

with the cis-compounds 17 and 19. There could, but must

not be, a connection of the kinetic preference with the higher

40

10 h

45

22

68:32

9

80

10 min

31

21

60:40

19

11

Ratio 26/27

Dione 34 (%)

S,S-Acetal 35 (%)

All products (%)

76

82

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

511

26±29 (Scheme 5). The previous discussion⁸needs no

repetition. Ketene imine 30 might occur in diastereo-

isomers since the allenic bond system is equivalent to one

stereocenter. In noteworthy contrast to the tricyano-

substituted ketene imine 5, the ester-bearing 30 was not

intercepted by methanol.

3.1. Cycloadditions of 2,2,4,4-tetramethylcyclobutane-

thione S-methylide (37)

Scheme 6.

phenone S-methylide (9) with the same two pairs of

Thiocarbonyl ylide 37 differs from 1 by the absence of the

keto group, and its cycloadditions to 16±19 showed devia-

tions in several respects. Thiadiazoline 23, the precursor of

37, was the most active catalyst in the cis, trans isomeriza-

tion, 16Y17 (Table 1). A small concentration of strong acid

did not prohibit the catalysis by 23. The cis-form 17 in

CDCl₃(7.6 mM in H₂SO₄) reached 16/1761:39 in the

presence of 0.10 M 23 in 1 h. The diisopropyl esters 18/19

were somewhat more resistant.

dipolarophiles.¹²

The yields of cycloadducts 26±29 were lower (52±87%,

Table 2) for the experiments run in the presence of H₂SO₄

than for those in THF without acid (up to 94%). Dione 34

and dimethyl dithioacetal 35 were found as side-products,

and Scheme 6 shows a conceivable pathway of their forma-

tion via the protonated species 33.

When 4 was reacted with 16 or 18 in THF containing 2 vol%

of water, the spirolactams 31 and 32 emerged with yields of

25 and 30%, respectively, alongside the thiolanes. The

Thiadiazoline 23 was prepared from thione 36 and diazo-

methane. The ®rst-order rate constant of N₂elimination was

smaller by 30% than that of 4 (xylene, 408C, t_1/21.6 h). The

electrocyclization of 37 gave rise to thiirane 38 (81%).

1

lactams appeared in the H NMR spectra as 2:1 mixtures

of diastereoisomers A and B which probably differ in the

con®guration at C-5⁰. Form A isomerized to B on the silica

gel layer. The lactams show the amide I frequency, e.g. at

1685 cm²¹for 31, well separated from the ester carbonyl at

1753 and the cyclobutanone absorption at 1787 cm²¹; NH

signals are visible in IR and ¹H NMR spectra. Four different

¹³C-signals for C±Me demonstrate chirality.

The reaction of 23 with dimethyl 2,3-dicyanofumarate (16,

1.2 equiv.) in CH₂Cl₂(408C, 15 h) furnished the thiolanes

39 and 40 in 87% yield (Scheme 7). The structural assign-

ment con®rmed the empirical rule, based on Dd(5⁰-H₂) and

described above: AB for 39 (trans, d 3.51 and 3.59) and AX

for 40 (cis, d 3.39 and 3.85). Similar parameters were

observed for the pair of diisopropyl esters, 41 and 42. The

3-H₂are likewise diastereotopic, and the AB-spectra of 39

and 40 are resolved in the 400 MHz spectra, e.g. d 1.60,

1.65 for 39.

In the trapping of an intermediate by water, the reaction of 1

with 16 corresponds to that observed with TCNE, in which

lactam 6 was formed (Scheme 1).⁸Thus, the rotameric

zwitterions 24 and 25 reversibly cyclize to the ketene

imine 30 and close irreversibly the thiolane ring to give

1

The H NMR analytical tests on the steric course were

carried out at 808C without acid. Virtually identical cyclo-

adduct ratios, 39/4055:45, were observed for the reactions

of 23 with 16 and 17. In the second experiment the 20%

excess of 17 was largely isomerized to 16 after the reaction.

Thiocarbonyl ylide 1 reacts 4.7 times faster with 16 (trans)

than with 17 (cis). Since a similar rate ratio k₁₆/k₁₇appears

probable in the reactions with 37, at least some rotation must

take place on the level of the zwitterions 43 and 44. Other-

wise the nearly 1:1 ratio of trans- and cis-adduct, 39 and 40,

would not be understood. However, the competing equili-

bration 16Y17 which is catalyzed by the precursor 23 does

not allow to quantify the nonstereospeci®city of the

cycloaddition itself. Nevertheless, full equilibration of the

zwitterions 43 and 44 is supposed.

When the pure cycloadducts 39 and 40 were heated in sepa-

rate experiments in benzonitrile at 1398C, an equilibrium of

39/4031:69 was established from both sides. The absence

of side-reactions allowed the kinetic measurement of the

reversible ®rst-order reactions. The equilibrium is attained

with a half-life (ln 2/(k₃₉1k₄₀)) of 12.8 h (benzonitrile,

1398C). There is no doubt that the above tests on the steric

course of cycloaddition (CDCl₃, 808C, 5 min) re¯ect kinetic

control.

Scheme 7.

512

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

The zwitterions 43 and 44 are the logical intermediates in

the adduct isomerization. In contrast, adduct 26 which

contains the 1-oxo group did not isomerize in benzonitrile

at 1398C. It is tempting to ascribe the difference in thermo-

stability to the lower energy level of the thiocarbonylium

zwitterions 43/44, compared with that of 24/25. The

intracyclic ®eld effect of the keto group destabilizes 24/

25. The cycloadduct of thiobenzophenone S-methylide (9)

to 16 equilibrates even in acetonitrile at 808C to trans/

cis30:70.¹²The ring-opening pro®ts from the conjugation

of the thiocarbonylium ion with two phenyl groups.

maleate which is 51 times less reactive than the trans-

isomer vs. 1, the retention is not complete, and 1.1% of

trans-adduct 11 implies a low-grade involvement of a

two-step mechanism. Fumaronitrile is superior to maleo-

nitrile only by a factor of 2.6 in the addition of 1. Retention

was observed for the reaction of 1 with maleonitrile.

In the cycloadditions of 1 with dimethyl (and diisopropyl)

2,3-dicyanofumarate and 2,3-dicyanomaleate (16±19), the

borderline from the concerted to the two-step pathway is

crossed. Two acceptor substituents stabilize the anionic

terminus of the intermediate to such an extent that the

concerted mechanism does not even appear to be partially

involved. Retention still preponderates over inversion

(Table 2), i.e. cyclization and rotation of the zwitterionic

intermediates are in keen competition.

Diisopropyl 2,3-dicyanomaleate (19) is less prone to the

thiadiazoline catalysis of cis-, trans-isomerization than

dimethyl ester 17 (Table 1). In the reactions of 23 with 18

and 19 (CDCl₃, 808C), trans/cis ratios of 41/4263:37 and

49:51, respectively, were observed, i.e. a slight preference

for retention.

Successful validation of a two-step mechanism requires

evidence that the loss of stereochemical integrity occurs

during the cycloaddition and not before or afterwards. In

this respect, the reactions of thiocarbonyl ylides 1 and 37

with 16±19 offer textbook examples and warnings. The

thiadiazoline precursors 4 and 23 catalyze the cis, trans

isomerization of the dipolarophile, and that more effectively

for the dimethyl esters 16 and 17 than for the diisopropyl

esters 18 and 19. Only the catalysis by 4, and not that by 23,

was counteracted with some mol% of strong acid. As a

consequence, the ratio of retention/inversion was deter-

minable for the cycloadditions of 1, whereas in the case of

37 the con®gurational losses before and during the cyclo-

additions could not be separated.

In the mass spectra of 39±42, the molecular peaks are very

small. A large peak is [M2C₄H₈]¹; the isobutene elimi-

nation is analogous to that of dimethylketene in the MS of

26±29. The smaller fragments are identical in both series,

and their molecular formulae show that alkoxycarbonyl

functions break off easier than cyano groups. Most of

the radical cations ®t molecular formulae of thiophene

derivatives.

4. Discussion and conclusions

ªThe pure and simple truth is rarely pure and never

simpleº.ÐOscar Wilde

Kinetic control was con®rmed for the reactions of both 1

and 37 with 16±19. An interesting difference emerged: the

cycloadducts of 1 are stable in benzonitrile at 1398C, while

those of 37 slowly equilibrated. These conditions are more

drastic than those of the cycloaddition experiments.

Full con®gurational retention of dipolarophile and 1,3-di-

pole is mandatory for concerted cycloadditions. Retention

could also be the outcome when the intermediate of a two-

step pathway undergoes cyclization much faster than

rotation (k_cycl.k_rot). For distinguishing the two cases, it is

helpful to determine the analytical limit of the missing

isomer in arti®cial mixtures of cis- and trans-adducts. It

has been shown, e.g. that diazomethane addition to methyl

angelate proceeds with .99.997% retention (GC analysis),¹⁹

and .99.992% resulted for the 1,3-addition of N-benzyl-

The zwitterions 24 and 25 do not only close the ®ve-

membered ring of 26±29. A reversible closure of the

seven-membered ring was revealed by the partial inter-

ception of ketene imine 30 by water, giving rise to lactams

31 and 32. The desirable evidence that the trapping of the

intermediate does not in¯uence the formation rate of the

latter was not procurable here, however. The liberation of

the thiocarbonyl ylide from its precursor is always rate-

determining.

idene-3-pyrazolidone to (E)-b-nitrostyrene (HPLC).²⁰

stereospeci®city of .99.89% was elegantly concluded

from an equilibration study of 3,4-dihydroisoquinoline

N-oxide and (E)-b-nitrostyrene with their cycloadduct.²¹

A

The steric course of a 1,3-dipolar cycloaddition has rarely

been used as evidence for a two-step pathway. The group of

A free energy consideration teaches that the pertinent rota-

tional barriers of hypothetical zwitterionic intermediates for

the three examples mentioned should exceed DG^±of its

cyclization by .6.2, .5.6, and .4.2 kcal mol²¹, respec-

tively. Rotations about C(sp³)±C(sp²) bonds have notor-

iously low barriers. The high barriers inferred discredit the

assumption of an intermediate and offer a criterion for

concertedness.^3c

The retention of .99.97% in the cycloaddition of 1 with

dimethyl fumarate, described above, would correspond to

DG^±_rot2DG^±_cycl.5.0 kcal mol²¹for a hypothetical (but

rejected) intermediate. In the addition of 1 with dimethyl

Scheme 8.

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

513

Sauer recently presented kind of pendant to the system

described here, namely the nonstereospeci®c cycloaddition

of an electron-de®cient 1,3-dipole to an electron-rich

dipolarophile.²²The N-dicyano-azomethine ylide 45 (and

a second related model) reacted with (E)-1-dimethyl-

amino-1-propene to afford trans- and cis-adduct 46 in

48:52 ratio (Scheme 8). Both adduct structures were

elucidated by X-ray, and kinetic studies with substituent

variation were in accordance with LU(1,3-dipole)±HO-

(dipolarophile) control.²³

Reaction of 1 with 1.1 equiv. of dimethyl fumarate (DF).

Thiadiazoline 4 (396 mg, 2.00 mmol),^11,30was dissolved in

the stirred suspension of DF (317 mg, 2.20 mmol) in abs.

THF (5 mL) and heated to 408C for 8 h. After removal of the

solvent at the rotary evaporator and dissolving the residue in

CDCl₃, the weight standard (sym-tetrachloroethane, d 5.92)

1

was added. Comparison of the H NMR integral at d 3.93

(3⁰-H) with that of the standard indicated 95% of 11.

Puri®cation by PLC (acetone/petroleum ether 2:8) gave a

colorless oil (378 mg, 60%) that crystallized on cooling.

Fine needles were obtained from pentane, mp 58±608C.

IR (KBr): n 1171 cm²¹s, 1200s, 1216s (C±O), 1331m,

1366m, 1388m, 1437m; 1748vs. (CvO, ester), 1775s

(CvO, ketone). ¹H NMR (400 MHz): d 1.25 (s, Me),

1.31 (s, 2Me), 1.38 (s, Me), 3.08, 3.12 (ddd, AB of

The borderline region between concerted and two-step DAR

has been the topic of extensive exploration.²The occurrence

of open-chain 1,3-dienes in s-cis and s-trans conformations

confers new complexity. Only the s-cis form is amenable to

the concerted DAR whereas both s-cis and s-trans confor-

mations can undergo two-step reactions with dienophiles

which afford (412) or (212) cycloadducts, respectively.

Bartlett's classic studies on concerted and biradical path-

ways in cycloadditions of dienes with ¯uorinated and

chlorinated ethylenes may be recalled.²⁴Among recent

examples, the studies of the Sustmann group are quoted.

(E)-1-Dimethylamino-1,3-butadiene reacted with fumaro-

nitrile and maleonitrile with retention, whereas 16 and 17

furnished the same trans-adduct; a preceding isomerization,

17Y16, restricted conclusions.²⁵The DAR of 1,4-bis-

(dimethylamino)-1,3-butadiene with 16 and 17 at 2508C

is accompanied by an electron-transfer mechanism.²⁶

0

ABMX, J_{5 A,5 B}11.6 Hz, J_{4 ,5 A}6.7 Hz, J_{4 ,5 B}8.0 Hz,

5⁰-H_A, 5⁰-H_B; the 5⁰-H_Bpart is long range-split by 0.5 Hz),

3.45 (ddd, M part of ABMX, 4⁰-H), 3.93 (d, X part, J_{3 ,4}

0

3.5 Hz, 3⁰-H), 3.75, 3.77 (2s, 2MeO). ¹³C NMR (20.2 MHz):

d 20.8, 22.7 (2q, 2Me), 23.0 (q, 2Me), 32.9 (t, C-5⁰), 52.2,

52.5 (2d, C-3⁰, C-4⁰), 53.1, 53.3 (2q, 2MeO), 62.3, 66.3,

68.6 (3s, C-2, C-3, C-4), 172.1, 173.2 (2s, CvO, ester),

219.6 (s, CvO, ketone). MS (408C); m/z (%): 314 (0.3,

M¹), 283 (4, [M2MeO]¹), 255 (2, [M2CO₂Me]¹), 244

(100, [M2C₄H₆O]¹; ¹³C calcd 12/found 13; ¹³C₂1³⁴S 5.1/

5.8), 212 (37, [2442MeOH]¹), 184 (35, [2442HCO₂Me]¹,

¹³C 10.0/10.7), 143 (5), 125 (34, [1842CO₂Me]¹), 91 (4),

85 (6), 70 (15, C₄H₆O¹), 59 (8, MeOCuO¹). Anal. calcd

for C₁₅H₂₂O₅S (314.39): C 57.30, H 7.05, S 10.20; found: C

57.44, H 7.07, S 10.22.

Nonstereospeci®city was observed for (212) cycloadditions

via tetramethylene zwitterions: the reactions of cis/trans

isomeric enol ethers and thioenol ethers with TCNE²⁷or

2,2-bis(tri¯uoromethyl)ethylene-1,1-dicarbonitrile.²⁸Further

criteria, such as structure-rate relationship, the dependence

of rate on solvent polarity, reversibility of the ®rst step, and

a wealth of interception reactions were in accordance with

zwitterionic intermediates.²⁸Rotation about the acceptor

bond was demonstrated by the (212) cycloadditions of

trans- and cis-1,2-bis(tri¯uoromethyl)ethylene-1,2-dicarbo-

nitrile with methyl vinyl ether.²⁹

(b) Stereospeci®city test. DF (1.01 g, 7.0 mmol) was

dissolved in dimethyl succinate (6 mL) by stirring at 418C

for 1 h. After further stirring with 4 (407 mg, 2.05 mmol) at

418C for 8 h, solvent and excess of DF were removed by

bulb to bulb distillation at 0.2 Torr. ¹H NMR analyses of the

residue in C₆D₆with sym-C₂H₂Cl₄showed a nearly quanti-

tative yield of 11. In the 500 MHz spectrum in C₆D₆, the 2s

at d 1.228 and 1.278 (2Me) of 11 show low-frequency ¹³C

satellites (0.555%) at 1.096 and 1.149 ppm, corresponding

1

to J_CH133 and 129 Hz. It was concluded from their

integrals that the presence of 0.03% of 12 should give rise

to a discrete peak at 0.985 ppm (s, Me). Since no signal was

visible in that region, the concentration of 12 must be

,0.03%.

5. Experimental

5.1. General

5.1.2. Dimethyl 2,2,4,4-tetramethyl-1-oxospiro[cyclobu-

tane-3,2⁰-thiolane]-3⁰,4⁰-cis-dicarboxylate (12). (a) Reac-

tion of 1 with 1.1 equiv. of dimethyl maleate (DM). The

experiment in THF at 408C was run as described above

IR: Perkin±Elmer 125 or Beckmann FT model IFS 45.

1

NMR: Bruker WP80CW (80 MHz) for H and WP80DS

(20 MHz) for ¹³C (multiplicities by comparison of

¹H-decoupled with off-resonance spectra), Varian XR

400S for ¹H (400 MHz) and ¹³C (100 MHz). Solvent:

acid-free CDCl₃, if not otherwise stated. The precision of

1

for the isolation of 11. The H NMR analysis of the crude

product with standard indicated 70% of 12 (d 2.81±3.37,

4⁰-H15⁰-H₂) and 27% of spirothiirane 15 (d 2.57, 3⁰-H₂);

closer inspection of the adduct fraction revealed a content of

,1% of 11. Thiolane 12 was obtained after PLC as a color-

less oil (329 mg, 52%) which crystallized after 2 weeks at

58C, mp 54±568C (methanol). IR (KBr): n 1177 cm²¹s,

1202s, 1242s (C±O); 1743s (CvO, ester), 1780s (CvO,

1

the H NMR analysis with weight standard and machine

integration depends on the signals and amounts in the

average to ^4% (relative). The result was improved by

using several signals and repeated integrations (4±6

times). MS (EI, 70 eV): AET 902 or Finnigan MAT 90;

intensities of isotope peaks are given in the form, e.g. ¹³C

calcd (%)/found (%). PLC: Preparative thick-layer chromo-

graphy on 2 mm of silica gel 60 PF (Merck).

1

ketone). H NMR (80 MHz): d 1.21, 1.30, 1.36, 1.39 (4s,

4Me), 2.82±3.37 (m, 4⁰-H15⁰-H₂), 3.67, 3.69 (2s, 2MeO),

3.69 (d, J10.6 Hz, 3⁰-H); (500 MHz, C₆D₆): 0.987, 1.19,

1.25, 1.30 (4s, 4Me), 2.63 (m, 5⁰-H₂), 3.27, 3.34 (2s, 2MeO),

3.44 (t, 4⁰-H), 3.64 (d, 3⁰-H). ¹³C NMR (20.2 Hz): d 20.7,

5.1.1. Dimethyl 2,2,4,4-tetramethyl-1-oxospiro[cyclo-

butane-3,2⁰-thiolane]-3⁰,4⁰-trans-dicarboxylate (11). (a)

514

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

22.6, 23.0, 23.6 (4 q, 4Me), 30.3 (t, C-5⁰), 50.8, 53.1 (2d, C-

3⁰, C-4⁰), 51.9, 52.2 (2q, 2MeO), 61.8, 65.8, 67.6 (3s, C-2,

C-3, C-4), 170.3, 170.5 (2s, 2CvO, ester), 219.4 (s, CvO,

ketone). MS (308C); m/z (%): 314 (5, M¹), fragments

similar to MS of 11. Anal. calcd for C₁₅H₂₂O₅S (314.39):

C 57.30, H 7.05, S 10.20; found: C 57.09, H 7.11, S 10.21.

(e) Stability test. Thiadiazoline 4 (0.71 mmol) was decom-

posed in the solution of 12 (0.57 mmol) in abs. THF (1 mL)

at 418C for 8 h. After PLC (hexane/acetone 7:3), no 11 was

1

found in the H NMR spectrum (500 MHz).

5.1.3. 2,2,4,4-Tetramethyl-1-oxospiro[cyclobutane-3,2⁰-

thiolane]-3,4-trans-dicarbonitrile (13). Fumaronitrile

(FN, 2.20 mmol) was reacted with 4 (2.00 mmol) in THF

1

(b) Puri®cation of dimethyl maleate (DM). The H NMR

1

spectrum (80 MHz) of a commercial quality (neat liquid)

shows at d 6.76 the s (vinyl-H) of DF; with 1,3,5-tribromo-

benzene as weight standard (d 7.69), a content of 0.34^

0.03% of DF was analyzed. 25.0 g (173 mmol) of this

sample was stirred in an ice-bath and treated with 4.00 g

of diphenyldiazomethane (20.6 mmol) in two portions, until

the deep-red color disappeared. Distillation at 868C/15 Torr

gave DM with an estimated DF content of ,0.01% (NMR

test as above). The concentration of DF should be reduced

from 0.34 to 0.0041% (i.e. below the analyt. limit of the

80 MHz spectrum) on the basis of Eq. (1); diphenyldiazo-

methane reacts at 408C with DF 36 times faster than with

DM.¹⁶

(4 mL) at 418C for 8 h, and H NMR analysis of the crude

product with standard indicated 93% of 13. Evaporation of

solvent and excess dipolarophile, re¯uxing of the residue in

ethanol (3 mL), and storing for 1 day at 158C afforded 13

(341 mg, 69%) as colorless crystals, mp 112±1138C. IR

(KBr): n 971 cm²¹m, 1017m, 1245m, 1368m, 1386m,

1

1459s, 1468s; 1791vs (CvO), 2247m (CuN). H NMR

(400 MHz, CDCl₃): d 1.31, 1.45, 1.58, 1.68 (4s, 4Me), 3.18

0

(dd, J_{5 A,5 B}12.0 Hz, J_{5 A,4}4.0 Hz, 5 -H_A), 3.57 (dd,

J_{5 A,5 B}12.0 Hz, J_{4 ,5 B}8.6 Hz, 5⁰-H_B), 3.82 (ddd, 4⁰-H),

0

3.93 (d, J_{3 ,4}2.0 Hz, 3 -H); (500 MHz, C₆D₆): d 1.13,

1.14, 1.29, 1.32 (4s, 4Me, signals used for stereospeci®city

test). ¹³C NMR (100 MHz, DEPT): d 21.0, 22.7, 22.88, 22.90

(4Me), 33.8 (C-5⁰), 36.9 (C-4⁰), 42.3 (C-3⁰), 62.5, 66.2, 68.8

(C-2, C-3, C-4), 117.7, 118.3 (2CN), 216 (CvO). MS

(608C); m/z (%): 248 (1, M¹), 221 (0.3 [M2 HCN]¹), 178

(100, [M2C₄H₆O]¹), 163 (7, [1782Me]¹), 151 (18,

[1782HCN]¹), 136 (5), 125 (4), 85 (5), 70 (93, C₄H₆O¹).

Anal. calcd for C₁₃H₁₆N₂OS (248.34): C 62.87, H 6.49, N

11.28, S 12.91; found: C 63.03, H 6.60, N 11.44, S 12.93.

(c) Stereospeci®city test. The puri®ed DM (840 mg, 5.83

mmol) and freshly recrystallized 4 (334 mg, 1.68 mmol)

were heated to 418C without solvent for 8 h. After removal

of the excess DM at 408C/10²³Torr and PLC (hexane/

1

acetone 7:3; no separation of 11 and 12), the H NMR

spectrum (500 MHz, C₆D₆) at high attenuation showed the

s of 2C±Me of 11 at d 1.382 and 1.408. Comparison of the

integrals (gravimetry of cut-outs) with those of two high-

frequency ¹³C-satellites at 1.374 and 1.430 ppm, each

0.555% of the C±Me singlets of 12 at d 1.247 and 1.303

led to 1.28 and 1.09% of 11 in two recordings with different

spinning frequencies. Furthermore, comparison of the area

at d 1.408 (11) with those of the 32-fold diminished

integrals of 12 at 1.186, 1.247, and 1.303 ppm (3Me)

provided 1.05% of 11 in the isomer mixture. A middle

value of 1.14% of 11 appeared reasonable. Calculated

with k51 (see below), the DF content (2.37 mmol) in the

puri®ed DM should generate 0.014% of 11, a negligible

amount.

5.1.4. 2,2,4,4-Tetramethyl-1-oxospiro[cyclobutane-3,2⁰-

thiolane]-3,4-cis-dicarbonitrile (14). (a) Puri®cation of

maleonitrile (MN). A sample of MN which was prepared

from maleamide by treatment with POCl₃,³²contained 20%

1

fumaronitrile (FN), as shown by the H NMR spectrum in

MeOD. Three recrystallizations from ethanol furnished MN

as colorless rods, mp 32±32.58C (31±328C).³²According to

the H NMR analysis (500 MHz, MeOD), the FN content

1

was reduced to 0.058%. The magnetic nonequivalence of

the 2 ole®nic H³³generates a ¹³C-satellite spectrum of AA⁰X

type with ¹J_CH185 Hz and ³J_CH11.4 Hz. The integrals of

the high-frequency ¹³C-satellites at d 6.742 and 6.765 of

MN were compared with that of the FN singlet at d

6.705; the sum of the two satellites corresponds to 0.555%

of the vinyl-H signal of MN.

(d) Competition constant of DF and DM for thiocarbonyl

ylide 1. DM (10.00 mmol) and DF (0.506 mmol) were

stirred at 418C, until DF was dissolved. After addition of

thiadiazoline 4 (0.409 mmol) and 7 h at 418C (without

solvent), 8.8 mL of N₂(94%) was eliminated. The excess

of DM and DF was removed at 408C (bath)/10²³Torr by

Kugelrohr distillation. The ¹H NMR analysis (80 MHz,

C₆D₆) with sym-C₂H₂Cl₄yielded 0.265 mmol of 11 (d at

3.91, 3⁰-H) and 0.141 mmol of 12 (d at 3.64, 3⁰-H), together

99%. The 1.14% portion of 11 in the product from DM (see

above) requires correction to 0.262 mmol of 11 and

0.143 mmol of 12 to re¯ect the relative dipolarophilic

activities (k51) which were evaluated by Eq. (1)³¹with

(A)_o(DF)₀, (B)₀(DM)₀, etc.

(b) Cycloaddition with 1.1 equiv. of MN. The reaction with

4 (2.00 mmol) was carried out as usual, and ¹H NMR-analy-

sis showed 92% of 14. Trituration of the residue with

pentane gave colorless crystals (382 mg, 77%), mp 140±

1428C (methanol). IR (KBr): n 1031 cm²¹m, 1447m,

1465m; 1786s (CvO), 2248m (CuN). ¹H NMR (80

MHz): d 1.27 (s, 2Me), 1.32, 1.59 (2s, 2Me), 2.85±3.40

0

(m, ABC, 4⁰-H15⁰-H₂), 3.88 (d, J_{3 ,4}4.5 Hz, 3 -H); (500

MHz, C₆D₆, used for stereospeci®city test): 0.54, 0.86, 1.08,

1.25 (4s, 4Me), 1.94 (ddd, 5⁰-H_A), 2.21 (dd, 5⁰-H_B), 2.76 (t,

4⁰-H), 2.91 (d, 3⁰-H). MS (708C); m/z: 248 (9, M¹), 233 (0.3,

[M2Me]¹), 205 (1), 178 (100, [M2C₄H₆O]¹), 163 (7,

[1782Me]¹), 151 (21, [1782HCN]¹), 125 (5), 70 (81,

C₄H₆O¹). Anal. calcd for C₁₃H₁₆N₂OS (248.34): C 62.87,

H 6.49, N 11.28, S 12.91; found: C 63.14, H 6.65, N 11.10, S

13.05.

0

k_A

k_B

logꢀA₀2 logꢀA

e

k

logꢀB₀2 logꢀB

e

logꢀA₀2 logꢀA₀2 ꢀA±Adduct

e

ꢀ1

logꢀB₀2 logꢀB₀2 ꢀB±Adduct

e

(c) Reaction with 2 equiv. of MN in NMR tube. 4

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

515

(0.22 mmol) and MN (0.44 mmol) in C₆D₆(0.5 mL) were

reacted at 408C for 16 h. With sym-C₂H₂Cl₄as standard,

88% of 14 (d 1.25, Me) and 11% of thiirane 15 (d 0.76

for 2Me) were analyzed.

1.39 for 19, 1.44 for 18) was suitable for ¹H NMR analysis

(80 MHz). Time-dependent ratios of 17/16 and 19/18 are

listed in Table 1. The catalytic activity of the thiadiazolines

increases in the sequence given in Table 1. The ef®ciency of

4 was similar in octane, C₆D₆, and CDCl₃, but as a conse-

quence of the low solubility of 16 the octane solution

became turbid after several minutes, and 16 began to

crystallize from the C₆D₆solution after 3 h.

(d) Stereospeci®city test. 4 (1.14 mmol) and MN

(2.20 mmol) in C₆D₆(1 mL) were reacted 8 h at 408C.

After removal of the solvent and Kugelrohr distillation of

excess MN, the colorless crystalline residue was H NMR-

1

analyzed (500 MHz) with 50 mg/mL of C₆D₆. The integrals

of the ¹³C-satellites of 14 at d 0.954 and 1.377 were

compared with those of two s at 1.134 and 1.144 ppm

(2Me) of 13 and gave a ratio of 13/140.13:99.87. The

reliability was checked with the 500 MHz spectrum of an

arti®cial mixture of 1410.80% of 13 (result 0.77%). The FN

content (0.058%, see above) of the MN sample competes

with MN for 1,3-dipole 1 (k2.6) and allows to expect

0.080% of 13 besides 14. Thus, only 0.05% of 13 are

unaccounted.

(c) Inhibition of thiadiazoline catalysis by strong acid. Conc.

H₂SO₄(0.5 mL) was vigorously shaken with 20 ml of

CDCl₃and stored for 20 h. The clear CDCl₃phase was

carefully decanted from the yellow drops of H₂SO₄into a

Schlenk ¯ask, kept under argon for another 10 h, and then

decanted into a second Schlenk ¯ask for storing. Titration

(1.0 mL12 mL of H₂O) with 0.01N NaOH indicated

7.6 mM H₂SO₄in CDCl₃. When C₆D₆was subjected to

the same treatment, 2.0 mM H₂SO₄in C₆D₆was obtained.

Methanesulfonic acid and tri¯uoromethane-sulfonic acid

likewise inhibited the thiadiazoline catalysis.

(e) Competition constant of FN and MN for thiocarbonyl

ylide 1. FN (2.12 mmol) and MN (3.39 mmol) were com-

peting for 1 which was generated from 4 (0.804 mmol) in

The experiments with 4 described above were repeated in

7.6 mM H₂SO₄in CDCl₃as medium (Table 1). The isomeri-

zation catalysis by 4 and the tetrasubstituted thiadiazoline

21 was prevented by 7.6 mol% of H₂SO₄. The protection

was incomplete in the case of 22, and much less so for 23.

1

C₆D₆(1.5 mL) at 408C. The H NMR analysis (80 MHz)

indicated 0.395 mmol of 13 and 0.244 mmol of 14, together

80%. The FN content (0.058%) in MN has to be taken in

account for the evaluation of k2.74 with Eq. (1). A second

experiment with FN (3.08 mmol), MN (5.16 mmol), and 4

(0.278 mmol) provided (k)2.52; middle value 2.6.

(d) Diisopropyl 2,3-dicyanomaleate (19) by photoisomeri-

zation of 18. The photoreaction 16!17 was carried out by

Gotoh et al.³⁴in the presence of 1 equiv. of benzene-1,4-

dicarbonitrile. We found that direct irradiation with a high-

pressure mercury arc was effective and saved the trouble of

removing the sensitizer.¹²Similarly, the irradiation of 18

(7.50 g) in CH₂Cl₂(300 mL) for 18 h gave 19/1875:25,

and fractional crystallization from hexane furnished 1.60 of

19 (4±5% content of 18). A NMR-pure sample showed

mp 51±528C. The mixed fractions were used for the next

photoexperiment. IR (KBr): n 1073 cm²¹m, 1101s, 1198s,

1282vs 1299vs (C±O), 1739s (CvO), 2235vw (CuN). ¹H

NMR (CDCl₃): d 1.37 (d, J6.0 Hz, 4Me), 5.17 (sept,

J6.0 Hz, 2CH); (C₆D₆): 0.90 (d, J6.1 Hz, 4Me), 4.72

(sept, 2CH). ¹³C NMR (20.2 MHz): d 21.3 (q, 4Me), 73.7

(d, 2CH), 111.4 (s, 2CN), 125.7 (s, CvC), 157.0 (s,

2CvO). Anal. calcd for C₁₂H₁₄N₂O₄(250.25): C 57.59, H

5.64, N 11.20; found: C 57.49, H 5.69, N 11.24.

5.1.5. cis,trans-Isomerization of 2,3-dicyanofumaric

esters and 2,3-dicyanomaleic esters. (a) Stability in solu-

tion. Dimethyl dicyanomaleate¹²(17, 0.10 mmol) in 1 mL

of solvent was stored at 208C for 18 h in the dark under N₂in

1

the NMR-tube and H NMR-analyzed by the s of MeO

which appears for 16 at d 4.03 and for 17 at 3.94

(CDCl₃). The small d difference requires recording with

5 Hz/cm and calibration with arti®cial mixtures. In CDCl₃,

octane, and C₆D₆the original content of ,2% of 16 was not

increased, whereas it reached 70% in CD₃CN. In a second

series, the solutions of 17 were heated at 808C in closed

NMR-tubes for 30 min: 17/1690:10 in CDCl₃(equilibr.

12:88),¹²91:8 in C₆D₆and in octane, 22:78 in CD₃CN

(probably close to equilibrium). The solubility of 16 is

lower than that of 17 in all tested solvents, e.g. in CDCl₃

8.50 mg/mL and in MeCN 38.2 mg/mL at 208C. It is not

clear whether the isomerization in CD₃CN is a function of

solvent polarity or a consequence of impurities. We

observed that small amounts of strong acids in CD₃CN

conc. H₂SO₄was superior to CF₃CO₂HÐstabilized 17.

No isomerization of 17 (0.10 mmol) was noticed in

16 mM H₂SO₄in CD₃CN (1 mL) after 18 h at rt and 2 h

of re¯uxing.

5.1.6. Dimethyl 3⁰,4⁰-dicyano-2,2,4,4-tetramethyl-1-oxo-

spiro[cyclobutane-3,2⁰-thiolane]-3⁰,4⁰-trans-dicarboxyl-

ate (26) and cis-isomer 27. (a) Reaction with 16 without

acid and isolation of 26 and 27. The suspension of 16

(427 mg, 2.20 mmol) in abs. THF (6 mL) was stirred in a

408C-bath for 30 min, 4 (396 mg, 2.00 mmol) was added,

and 16 dissolved during 8 h of stirring at 408C. The solvent

was evaporated, and volatile side-products (e.g. 15) were

removed at 408C/10²³Torr. The H NMR analysis with

1

(b) 1,3,4-Thiadiazolines as isomerization catalysts. The cis-

dicarboxylic esters 17 and 19 as the minor component in the

equilibria with 16 and 18 were chosen to compare the cata-

lytic activity of several thiadiazolines and their inhibition by

acid. In volumetric ¯asks (1 mL), 17 (0.11 mmol, contains

2% of 16 as above) and 19 (0.11 mmol, 5% content of 18,

see later) were treated with freshly recrystallized thiadiazo-

lines (0.10 mmol each) in CDCl₃and stored at rt in the dark.

The Me groups of iPrO appear as doublets; the left branch (d

sym-C₂H₂Cl₄was based on the C±Me singlets and showed

26/2752:48 and 94% yield. PLC did not separate the

isomers. A top fraction of pure 26 (17%), mp 151±1538C,

came from hot ethanol. Fractional crystallization from

methanol and, ®nally, from ether at 2788C gave pure cis-

adduct 27 (5%), mp 132±1338C. A cycloaddition experi-

ment carried out in acetonitrile afforded 26/2748:62 and

offered a richer source for the isolation of 27.

516

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

(b) Data of 26 (trans). IR (KBr): n 1247 cm²¹s, 1258s

(C±O), 1749s (CvO, ester), 1794s (CvO, ketone),

ampli®cation, integrations in such a multi-component spec-

trum are problematic and high precision is not expected.

The excess of 17 is (qualitatively) observed, but no 16 is

visible. For trans-thiolane 26, the Me signals indicated

23.9 mmol and that of 5⁰-H₂18.4 mmol.

1

2250vw (CuN). H NMR (400 MHz): d 1.43, 1.47, 1.56,

1.85 (4s, 4Me), 3.56, 3.64 (AB, J11.4 Hz, 5⁰-H₂), 3.95,

3.97 (2q, 2MeO). ¹³C NMR (20.2 MHz): d 23.3 (q, 2Me),

24.1, 24.5 (2q, 2Me), 36.3 (t, C-5⁰), 54.7, 55.0 (2s, 2MeO),

60.6, 62.6 (2s, C-3⁰, C-4⁰), 67.7, 68.9, 70.2 (3s, C-2, C-3,

C-4), 114.5, 116.0 (2s, 2CN), 163.7, 163.8 (2s, CvO, ester),

217.3 (s, CvO, ketone). Anal. calcd for C₁₇H₂₀N₂O₅S

(364.41): C 56.03, H 5.53, N 7.69, S 8.80; found: C

55.79, H 5.68, N 7.96, S 8.82.

The corresponding data for cis-adduct 27 were 69.8 and

62.0 mmol, bringing the yield of cycloadducts to about

87% with 26/2724:76. Furthermore, 3.7 mmol of 34 and

4.5 mmol of 35 were analyzed. Thus, 95% of the initial

thiadiazoline 4 were accounted for. Prior to the evaluation,

an arti®cial mixture of 16, 17, and the four products in the

same solvent was analyzed with fair results; 16 (85%)

showed the largest deviation. In run 3 of Table 2 (17,

408C), some 16 was observed in the unconsumed dipolaro-

phile.

(c) Data of 27 (cis). IR (KBr): n 1258 cm²¹s, br (C±O),

1745, 1753s (CvO, ester), 1790s (CvO, ketone), 2245vw

1

(CuN). H NMR (400 MHz): d 1.15, 1.64, 1.83, 2.03 (4s,

4Me), 3.43, 3.87 (AX, J12.1 Hz, 5⁰-H₂), 3.91, 3.95 (2s,

2MeO). ¹³C NMR (100 MHz, DEPT): d 22.2 (Me), 23.0

(2Me), 23.8 (Me), 35.4 (C-5⁰), 54.8, 55.1 (2MeO), 58.3,

60.9 (C-4⁰, C-3⁰), 64.6, 70.4, 71.1 (C-2, C-3, C-4), 114.9,

116.5 (2CN), 163.3, 164.0 (2CvO, ester), 216.2 (CvO,

ketone). MS (608C, high resolution data of many peaks

with CMASS on MAT 95Q provided molecular formulae;

R#500); m/z (%): 364 (0.09, M¹), 333 (1.1, [M2MeO]¹),

305 (1.7, [M2CO₂Me]¹), 294 (100, [M2C₄H₈O]¹,

C₁₃H₁₄N₂O₄S¹; ¹³C 14.5/15.0, ¹³C₂1³⁴S 5.4/5.9), 262 (2,

[2942MeO]¹), 249 (4, [2942CO₂H]¹, C₁₂H₁₃N₂O₂S¹),

235 (4, [2942CO₂Me]¹, C₁₁H₁₁N₂O₂S¹], 209 (6,

[2352CN]¹), 194 (7, C₉H₈NO₂S¹), 193 (10), 192 (7,

C₁₀H₁₂N₂S¹), 191 (49, [2352CO₂]¹, C₁₀H₁₁N₂S¹), 183

(22, C₈H₉NO₂S¹), 177 (19, C₉H₇NOS¹), 176 (10,

C₁₀H₈OS¹), 161 (6, C₈H₅N₂S¹), 150 (18, C₈H₈NS¹), 149

(6, C₈H₇NS¹), 125 (8, C₆H₅OS¹), 70 (19, C₄H₆O¹), 59 (12,

MeOCuO¹), 41 (9, Allyl¹). Anal. calcd for C₁₇H₂₀N₂O₅S

(364.41): C 56.03, H 5.53, N 7.69; found: C 55.74, H 5.29, N

7.55.

(f) Competition of 16 and 17 for thiocarbonyl ylide 1.

Experiment no. 5 of Table 2 followed the above procedure.

The amounts of unconsumed 16 and 17 were corrected for

signal overlap, and application of Eq. (1) furnished (an

unreliable) k4.4. More trustworthy is the determination

from the yields of 26 and 27 (in mmol):

logꢀ16₀2 logꢀ16₀2 x

k

4:7

logꢀ17₀2 logꢀ17₀2 y

where x and y are functions of the ratios of 26/27 observed in

separate experiments with 16 and 17 (No. 2 and 4, Table 2):

0:60x 1 0:24y ꢀ26 15:3;

x 18:9

y 15:3

e

0:40x 1 0:76y ꢀ27 19:2;

e

5.1.7. Dimethyl 6⁰-cyano-2⁰,3⁰,4⁰,5⁰,6⁰,7⁰-hexahydro-

2,2,4,4-tetramethyl-1,4⁰-dioxospiro[cyclobutane-3,2⁰-(1,3)-

thiazepine]-5⁰,6⁰-dicarboxylate (31). (a) Interception of

ketene imine 30 with water. Thiadiazoline 4 (2.00 mmol)

and 16 (2.20 mmol) in THF (5 mL, contained 2 vol% of

H₂O) were reacted at 508C (bath) for 5 h. After the usual

(d) Thermal stability of cycloadducts. Sample of pure 26

and 27 were re¯uxed in toluene for 2 h, and the H NMR

1

analysis did not reveal any change. In another test, pure

trans-adduct 26 (90 mg) and octamethylcyclotetrasiloxane

(9 mg, weight standard) in freshly distilled benzonitrile

(0.4 mL) were heated in a sealed NMR-tube under

1

workup, the H NMR analysis with trichloroethylene as

1

argon at 1398C for 7 h; the H NMR spectrum indicated

beginning decomposition, but the CMe signals did not

reveal 27.

standard indicated the thiolanes 26 (d 1.85, Me) and 27

(2.03, Me) in 71% yield with 26/2752:48 as well as the

lactams 31A (4.32, 5⁰-H) and 31B (4.42, 5⁰-H) in 25% yield

with 31A/31B,2:1. PLC (CH₂Cl₂/acetone 95:5) furnished

26127 (485 mg, 67%), as the ®rst zone, followed by 31B

(130 mg, 17%) which crystallized from methanol at 48C, mp

190.5±1918C. The conversion 31A!31B took place on the

silica gel layer. IR (KBr): n 1034 cm²¹m; 1214112381

1261s, br. (C±O), 1685s (CvO, amide-I), 1753s (CvO,

ester), 1787s (CvO, ketone), 2255vw (CuN), 3240sh, br.

(N±H). ¹H NMR: d 1.33, 1.41 (2s, 2Me), 1.48 (s, 2Me), 3.43

(s, 7⁰-H₂), 3.78, 3.88 (2s, 2MeO), 4.34 (s, 5⁰-H), 6.85 (s, br.,

NH). ¹³C NMR (100 MHz, DEPT): d 19.7, 21.8, 22.6, 23.4

(4Me), 37.2 (C-7⁰), 45.5 (C-6⁰), 52.6 (C-5⁰), 53.3, 54.7

(2MeO), 66.6, 69.0, 69.8 (C-2, C-3, C-4), 161.7, 166.1,

167.7 (3 CvO, ester, amide), 216.6 (CvO, ketone). MS

(1458C); m/z (%): 382 (0.04, M¹), 367 (0.85, [M2Me]¹),

322 (100, [M2HCO₂Me]¹), 312 (21, [M2C₄H₆O]¹, ¹³C

3.0/3.3, ¹³C₂1³⁴S 1.1/1.3), 263 (5), 222 (20), 219 (18),

217 (21), 204 (14), 203 (14), 189 (16), 183 (18), 178 (15),

125 (7), 70 (11, C₄H₆O¹), 69 (17, C₄H₅O¹), 59 (13,

(e) ¹H NMR analysis of steric course. The following signals

of the 360 MHz spectrum were suitable for machine

integration: 16, s d 4.03 (MeO); 17, s 3.94 (MeO, joint

integration with one MeO of 26); 26 (trans), 3s at 1.47,

1.56, 1.85 (3Me), AB of 5⁰-H₂at 3.56, 3.64; 27 (cis), 4s

(4Me), d at 3.43 (5⁰-H_A); dione 34, s 1.32 (4Me); dithio-

acetal 35, s 2.05 (2SMe); thiirane 15, s 2.57 (3⁰-H₂).

Weight standards: 2,6-dimethylnaphthalene, s 2.47 (2Me),

as-C₂H₂Cl₄, s 4.28 (2H).

Experiment 4 of Table 2 serves as example. Freshly recryst.

4 (19.8 mg, 100 mmol) and pure 17 (23.1 mg, 119 mmol) in

7.6 mM H₂SO₄in CDCl₃(1 mL) were heated in a closed

NMR tube under argon for 10 min at 808C. After releasing

the N₂pressure at 2788C, 2,6-dimethylnaphthalene

(2.67 mg, 17.1 mmol) was added, and the 360 MHz

spectrum recorded. Despite sectionwise expansion and

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

517

CO₂Me¹). Anal. calcd for C₁₇H₂₂N₂O₆S (382.43): C 53.39,

H 5.80, N 7.33; found: C 53.72, H 5.74, N 7.09.

25.3 mmol) and 18 (8.45 mg, 33.8 mmol), dissolved in

7.6 mM H₂SO₄in CDCl₃(1 mL), was heated in a closed

NMR tube at 808C (bath) for 10 min. After cooling and

1

(b) No interception with methanol. After the reaction of 4

with 16 in CDCl₃, containing 2 vol% of methanol (10 min,

808C), the ¹H NMR spectrum showed 26/2760:40, but no

imidate signals. Even in methanol as solvent (3 h, 658C)

only the thiolanes were formed.

adding sym-C₂H₂Cl₂as standard, the 360 MHz H NMR

spectrum showed the undisturbed 5⁰-H₂signals of 28 (AB)

and 29 (AX). The ratio 28/29 was 61:39 and the yield 56%.

The analogous reaction of 4 with 19 (1.4 equiv.) afforded

28/2925:75 and 49% yield. The side-products were not

investigated. In both reactions, an isomerization of the

excess 18 or 19, respectively, was not noticed.

5.1.8. Diisopropyl 3⁰,4⁰-dicyano-2,2,4,4-tetramethyl-1-

oxospiro[cyclobutane-3,2⁰-thiolane]-3⁰,4⁰-trans-dicarb-

oxylate (28) and cis-isomer 29. (a) Isolation of cyclo-

adducts 28 and 29. Diisopropyl 2,3-dicyanofumarate (18,

450 mg, 1.80 mmol) in octane (5 mL) was vigorously

stirred in a 1208C bath. The solution of 4 (396 mg,

2.00 mmol) in octane (5 mL) was dropwise added in

20 min. After removal of the solvent in vacuo, the residue

was taken up in hot 2-propanol and afforded a ®rst-fraction

(90 mg, 12%) of 28 as colorless prisms, mp 129±1308C

(diisopropyl ether). The isomer separation of further

fractions by PLC failed. An analogous experiment was

carried out with 19 and furnished 29 (210 mg, 28%)

as the ®rst crystal fraction, mp 125±1268C, from

2-propanol.

5.1.9. Diisopropyl 6⁰-cyano-2⁰,3⁰,4⁰,5⁰,6⁰,7⁰-hexahydro-

2,2,4,4-tetramethyl-1,4⁰-dioxospiro[cyclobutane-3,2⁰-

(1,3)-thiazepine]-5⁰,6⁰-dicarboxylate (32). Reaction with

1

18 in THF 12 vol% of H₂O as described for 31. The H

NMR spectrum (C₆D₆, trichloroethylene as standard)

showed thiolanes 28 and 29 (about 45%), thiirane 15 (s,

2.60, 4%), as well as lactams 32A (s 4.37, 5⁰-H, 20%),

and 32B (s 4.55, 5⁰-H, 10%). PLC (CH₂Cl₂and a second

run in CH₂Cl₂/acetone 95:5) was accompanied by isomeri-

zation of 32A!32B; the oil crystallized from pentane, mp

147±1488C. IR (KBr): n 1032 cm²¹m; 1104s, 1209s,

1233m, 1259m, 1286m (C±O); 1377s, 1390s; 1682vs.

(amide I), 1745vs (CvO, ester), 1788s (CvO, ketone),

2250vw (CuN), 3100w, 3280m, br. (N±H). ¹H NMR

(80 MHz): d 1.00±1.60 (m, 8Me), 3.39 (s, 7⁰-H₂), 4.24 (s,

5⁰-H), 5.07, 5.08 (2sept, J6.1 Hz, 2OCHMe₂), 7.44 (s, br.,

NH), (C₆D₆): d 0.62±1.55 (m, 8Me), 3.28, 3.47 (AB,

J15.0 Hz, 7⁰-H₂), 4.52 (s, 5⁰-H), 4.89 (sept, J6.1 Hz,

2OCHMe₂), 7.93 (s, NH). Anal. calcd for C₂₁H₃₀N₂O₆S

(438.53): C 57.51, H 6.90, N 6.39, S 7.31; found: C

57.87, H 7.28, N 6.25, S 7.29.

(b) Data of 28 (trans). IR (KBr): n 1099 cm²¹s,

125311269s (C±O), 1745sh, 1750s (CvO, ester), 1784s

(CvO, ketone), 2260vw (CuN). H NMR (400 MHz):

1

1.35, 1.38, 1.39, 1.43 (4 d, J6.4 Hz, 2 pairs of diastereo-

topic Me, 2iPr), 1.43, 1.52, 1.57, 1.88 (4s, 4Me), 3.53, 3.59

(AB, J11.2 Hz, 5⁰-H₂), 5.17, 5.20 (2sept, J6.4 Hz, 2CH

of 2iPr). ¹³C NMR (20.2 MHz): d 21.0, 21.1, 21.2, 21.5 (4q,

4Me of 2iPr), 23.4, 23.5, 24.2, 24.6 (4q, 4Me at C-2,

C-4), 36.1 (t, C-5⁰), 60.9, 62.9 (2s, C-2, C-4), 67.5, 68.9,

70.3 (3s, C-3, C-3⁰, C-4⁰), 73.4, 74.1 (2d, 2CH of 2iPr),

114.7, 116.4 (2s, 2CN), 162.7 (s, 2CvO, ester), 217.7 (s,

CvO, ketone). Anal. calcd for C₂₁H₂₈N₂O₅S (420.52): C

59.98, H 6.71, N 6.66, S 7.63; found: C 60.08, H 6.68, N

6.54, S 7.65.

5.1.10. 1,5⁰-Dihydro-2,2,4,4-tetramethylspiro[cyclobutane-

1,2⁰-(1,3,4)-thiadiazole] (23). (a) 2,2,4,4-Tetramethyl-

cyclobutanethione (36). The methylation of cyclobutanone

by the procedure with MeI and KOH in DMSO³⁵furnished

2,2,4,4-tetramethylcyclobutanone in 49% yield, bp 128±

1308C (1288C).³⁶For the conversion to thione 36, the ketone

(5.00 g) in abs. MeOH (40 mL) and trimethyl orthoformate

(7.2 mL) was treated with HCl and H₂S at 08C for 12 h. The

usual workup afforded 36 (5.04 g, 89%) as an orange oil. ¹H

NMR: d 1.26 (s, 4Me), 2.24 (s, 3-H₂).

(c) Data of 29 (cis). IR (KBr): d 1099 cm²¹s, 1254s, 1266s

(C±O), 1742s, 1757s (CvO, ester), 1788s (CvO, ketone),

2240vw (CuN. H NMR (360 MHz): d 1.15, 1.64, 1.88,

1

2.03 (4s, 4Me), 1.38 (d, J6.0 Hz, 2CHMe₂), 3.38, 3.85

(AX, J_gem11.9 Hz, 5⁰-H₂), 5.08, 5.12 (2sept, J6.0 Hz,

CHMe₂). ¹³C NMR (20.2 MHz): d 21.3 (br., q, 2CHMe₂),

22.2, 22.8, 23.0, 23.9 (4 q, 4Me), 35.2 (t, C-5⁰), 58.1, 61.3,

64.3, 70.7, 71.4 (5s, C-2, C-3, C-4, C-3⁰, C-4⁰), 73.5, 74.3

(2d, 2OCHMe₂), 115.2, 117.1 (2s, 2CN), 162.4, 162.9 (2s,

2CvO, ester), 216.2 (CvO, ketone). MS (608C); m/z (%):

420 (0.06, M¹), 361 (2.4, [M2iPrO]¹, ¹³C 0.47/0.44), 350

(47, [M2C₄H₆O]¹, ¹³C 10.9/11.3, ¹³C₂1³⁴S 3.5/4.0), 319

(13), 264 (7, [3502CO₂C₃H₆]¹, C₁₃H₁₆N₂O₂S¹), 237 (19,

[2642HCN]¹, ¹³C 2.5/2.8), 222 (100, [2642C₃H₆]¹,

C₁₀H₁₀N₂O₂S¹, ¹³C 11.1/12.5, ¹³C₂1³⁴S 5.0/5.5), 195 (93,

[2372C₃H₆]¹, C₉H₉NO₂S¹), 179 (22, [2372OC₃H₆]¹),

177 (41, [2642CO₂iPr]¹, C₉H₉N₂S¹), 150 (16, [1772

HCN]¹), 125 (8), 70 (26, C₄H₆O¹), 59 (4, C₃H₇O¹), 43

(71, iPr¹). Anal. calcd for C₂₁H₂₈N₂O₅S (420.52): C

59.98, H 6.71, N 6.66, S 7.63; found: C 60.03, H 6.88, N

6.55, S 7.68.

(b) Preparation of thiadiazoline 23. The reaction of 36 with

diazomethane in diethyl ether proceeded as described for

4.^11,30Twice recrystallized from pentane at 2788C, 23

1

(57%) was obtained colorless, mp 9±108C. H NMR: d

1.03, 1.23 (2s, 4Me), 1.94, 2.37 (AB, J11.6 Hz, 3-H₂),

5.58 (s, 5⁰-H₂). Anal. calcd for C₉H₁₆N₂S (184.30): C

58.65, H 8.75, N 15.20, S 17.40; found: C 59.11, H 8.97,

N 15.81, S 17.40.

(c) Kinetics of N₂extrusion. The nitrometric technique

followed the earlier description.³⁷Triple measurements in

each solvent provided values for 10⁴k₁(s²¹) at 40^0.28C:

1.19 (xylene), 1.14 (THF), 0.71 (acetonitrile).

(d) 2,2,4,4-Tetramethylspiro[cyclobutane-1,2⁰-thiirane] (38).

The above thermolysis solutions afforded the oily 38 (81%),

bp 90±1108C (bath)/12 Torr. H NMR (80 MHz): d 1.00,

1

1.17 (2s, 4Me), 1.87, 1.95 (AB, J8.5 Hz, 3-H₂), 2.48 (s,

(d) Steric course of addition of 1 with 18 and 19. 4 (5.01 mg,

3⁰-H₂). ¹³C NMR (20.2 MHz): d 27.41, 30.2 (2q, 4Me),

518

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

27.53 (t, C-3⁰), 36.3 (s, C-2, C-4), 47.6 (t, C-3), 66.1 (s, C-1).

MS (208C); m/z (%): 156 (34%, M¹), 141 (20, [M2Me]¹),

109 (15, [1412S]¹), 100 (100, [M2C₄H₈]¹), 85 (95,

[1002Me]¹), 67 (32, C₅H₇¹), 55 (19, C₄H₇¹). Anal. calcd

for C₉H₁₆S (156.28): C 69.16, H 10.32, S 20.52; found: C

69.42, H 10.32, S 20.52.

117.0 (2s, 2CN), 164.1, 164.7 (2s, 2CvO). Anal. calcd for

C₁₇H₂₂N₂O₄S (350.43): C 58.26, H 6.32, N 8.00, S 9.15;

found: C 58.56, H 6.29, N 7.80, S 9.16.

(f) Equilibration kinetics of cycloadducts 39 and 40. 39

(0.312 mmol) and octamethylcyclotetrasiloxane (OMCTS,

0.0151 mmol) in benzonitrile (0.4 mL) were sealed in an

NMR tube under argon and heated to 1398C. Based on

OMTCS, the integrals of s d 1.20 (Me of 39) and s 0.83

(Me of 40) were used for the analysis. After 74.3 h, the

equilibrium 39/4031:69 was reached from both sides

within the analyt. limits; no loss by decomposition was

noticeable. Evaluation of 15 integrals each for 39 and 40

was based on the integrated rate Eq. (2) for reversible ®rst-

order reactions and gave 10⁵£(k₃₉1k₄₀)1.51^0.03 s²¹

(benzonitrile, 1398C). Partitioning afforded k₃₉1.06£

10²⁵s²¹and k₄₀0.45£10²⁵s²¹

5.1.11. Dimethyl 3⁰,4⁰-dicyano-2,2,4,4-tetramethylspiro-

[cyclobutane-1,2⁰-thiolane]-3⁰,4⁰-trans-dicarboxylate (39)

and cis-isomer 40. (a) Cycloaddition of 37 with 16.

Thiadiazoline 23 (750 mg, 4.07 mmol) and 16 (871 mg,

4.49 mmol) in CH₂Cl₂(20 mL) were re¯uxed for 15 h.

After evaporation of the solvent, the residue was triturated

with CDCl₃, and the excess of 16 was ®ltered. The ¹H NMR

analysis with sym-C₂H₂Cl₄indicated 39 (2s at 1.36, 1.40,

2Me, 1.64 mmol, 40%) and 40 (2s at 1.01, 1.96, 2Me,

1.93 mmol, 47%), i.e. in the ratio 46:54. Adduct 40

(378 mg, 27%) crystallized from methanol at room temp.,

followed on cooling to 08C by 39 (200 mg, 14%, colorless

needles). Further fractions consisted of mixtures which were

partially separated manually. Solubilities in methanol at

208C: 20.1 mg of 39 per mL, 15.7 mg of 40 per mL.

ꢀk₃₉1 k₄₀t logꢀc₀2 c_e=ꢀc_t2 c_e

ꢀ2

The concentrations, i.e. the integrals, ®t well the conversion

curves which were calculated with above parameters.

5.1.12. Diisopropyl 3⁰,4⁰-dicyano-2,2,4,4-tetramethyl-

spiro[cyclobutane-1,2⁰-thiolane]-3⁰,4⁰-trans-dicarboxyl-

ate (41) and cis-isomer (42). (a) Reaction of 37 with 18.

The solution of 23 (4.14 mmol) in octane (5 mL) was

dropped into the stirred suspension of 18 (4.14 mmol) in

octane (10 mL) at 808C in 15 min. After further 15 min at

808C, the ®ltered solution was evaporated in vacuo. A

partial separation of the thiolane mixture was achieved by

repeated PLC (pentane/ether 9:1, 4 times). Bulb-to-bulb

distillation of the top fraction (240 mg) at 2008C/10²²

Torr gave pure 42 as a viscous oil which crystallized after

several months at 2248C, mp 38±418. The third PLC

fraction (280 mg) was enriched in 41 which crystallized

from diisopropyl ether, mp 100±1018C.

(b) Variation of solvent. Analogous experiments were

carried out in acetonitrile at 408C (39/4027:43, 88%

yield) and in toluene at 408C (39/4063:37, 90%). A

cycloaddition of 23 with 16 (1.4 equiv.) in CDCl₃

(0.5 mL) was carried out in the closed NMR tube (10 min,

808C-bath) and gave 39/4056:44.

(c) Reaction of 37 with 17. The experiment in the NMR

tube, carried out at 808C with 23 (0.10 mmol) and 17

(0.12 mmol) in CDCl₃(0.5 mL), provided 39/4055:45.

The excess of 17 was largely isomerized to 16. According

to Table 1, 23 was the most active catalyst in the equili-

bration 16Y17, and the catalysis was hardly suppressed in

7.6 mM H₂SO₄in CDCl₃.

(b) Data of 41 (trans). IR (KBr): n 1100 cm²¹s, 1257s, br.

(C±O), 1745s (CvO), 2245vw (CuN). ¹H NMR (400

MHz): d 1.32, 1.35, 1.39, 1.41 (4 d, J6.4 Hz, 4Me of

2iPr), 1.35, 1.46, 1.50, 1.74 (4s, 4Me), 1.59, 1.64 (AB, J

11.6 Hz, 3-H₂), 3.49, 3.55 (AB, J11.4 Hz, 5⁰-H₂), 5.13,

5.18 (2sept, J6.4 Hz, 2CH of 2iPr). ¹³C NMR (20.2

MHz): d 21.0, 21.2 (2£), 21.5, 27.9, 28.1, 30.8, 31.1 (8 q,

8Me), 36.1 (t, C-5⁰), 42.4, 46.6 (2s, C-2, C-4), 49.5 (t, C-3),

60,5, 63.5 (2s, C-4⁰, C-3⁰), 73.0, 73.2 (2d, 2CHMe₂), 74.8 (s,

C-1), 115.0, 116.9 (2s, 2CN), 163.1, 163.3 (2s, 2CvO). MS

(708); m/z (%): 407 (0.05, M¹11), 391 (0.01, [M2Me]¹),

350 (43, [M2C₄H₈]¹, ¹³C 8.1/8.4, ¹³C₂1³⁴S 2.6/2.8), 347

(5, [M2iPrO]¹, C₁₈H₂₃N₂O₃S¹, ¹³C 1.1/1.1, ¹³C₂1³⁴S 0.33/

0.35), 305 (3, [3472C₃H₆]¹, ¹³C 0.5/0.6), 264 (6,

[3502CO₂±C₃H₆]¹, C₁₃H₁₆N₂O₂S¹, ¹³C 0.9/1.0; this peak

as the following ones occur also in the MS of 29), 237 (33,

C₁₂H₁₅NO₂S¹, ¹³C 4.5/4.6), 222 (88, [2642C₃H₆]¹,

C₁₀H₁₀N₂O₂S¹, ¹³C 9.7/10.4, ¹³C₂1³⁴S 4.4/4.6), 195 (100,

[2372C₃H₆]¹, C₉H₉NO₂S¹, ®ts isopropyl cyanothiophene-

carboxylate¹, ¹³C 10.0/10.8, ¹³C₂1³⁴S 4.4/4.6), 194 (6), 178

(17), 177 (44, [2642CO₂iPr]¹, C₉H₇NOS¹), 176 (17), 161

(12, C₈H₅N₂S¹), 150 (20, C₈H₈NS¹), 149 (7), 109 (9,

C₅H₃NS¹, ®ts cyanothiophene¹), 80 (8), 43 (55, C₃H₇¹),

41 (22). Anal. calcd for C₂₁H₃₀N₂O₄S (406.53): C 62.04,

H 7.44, N 6.89, S 7.89; found: C 62.02, H 7.25, N 6.87, S

7.91.

(d) Data of 39 (trans). Mp 122±1238C. IR (KBr): n

1179 cm²¹m, 1248s, 1263s (C±O), 1746s (CvO),

2249vw (CuN). H NMR (400 MHz): d 1.36, 1.40, 1.48,

1

1.72 (4s, 4Me), 1.60, 1.65 (AB, J11.6 Hz, 3-H₂), 3.51,

3.59 (AB, J11.5 Hz, 5⁰-H₂), 3.91, 3.94 (2s, 2MeO). ¹³C

NMR (20.2 MHz): d 27.6, 27.9, 30.7, 31.0 (4 q, 4Me), 36.2

(t, C-5⁰), 42.6, 46.4 (2s, C-2, C-4), 49.5 (t, C-3), 54.2, 54.8

(2q, 2MeO), 60.2, 63.1, 74.6 (3s, C-4⁰, C-3⁰, C-1), 114.8,

116.5 (2s, 2CN), 164.0, 164.3 (2s, 2CvO). MS (708C, for

assignments below m/z 294, see MS of 27); m/z (%): 351

(0.04, M¹11), 350 (0.05, M¹), 319 (2.6, [M2MeO]¹), 294

(100, [M2C₄H₈]¹), 249 (3), 235 (4), 208 (4), 194 (5), 192

(7), 191 (46), 183 (32), 177 (15), 176 (9), 150 (17), 125 (8),

109 (5), 59 (15), 55 (4, C₄H₇¹), 41 (9, Allyl¹). Anal. calcd

for C₁₇H₂₂N₂O₄S (350.43): C 58.26, H 6.32, N 8.00, S 9.15;

found: C 58.35, H 6.24, N 7.89, S 9.17.

(e) Data of 40 (cis). Mp 1418C. IR (KBr): n 1251

cm²¹11260s, br. (C±O), 1429s; 1742s, 1753s (CvO),

2246vw (CuN). H NMR (400 MHz): d 1.01, 1.55, 1.78,

1

1.96 (4s, 4Me), 1.71, 1.77 (AB, J11.1 Hz, 3-H₂), 3.39,

3.85 (AX, J12.2 Hz, 5⁰-H₂), 3.86, 3.90 (2s, 2MeO). ¹³C

NMR (20.2 MHz): d 26.7, 27.7, 29.1, 29.4 (4 q, 4Me), 35.1

(t, C-5⁰), 40.6, 47.3 (2s, C-2, C-4), 48.1 (t, C-3), 54.3, 54.7

(2q, 2MeO), 57.0, 61.2, 75.1 (3s, C-4⁰, C-3⁰, C-1), 115.5,

R. Huisgen et al. / Tetrahedron 58 (2002) 507±519

519

(c) Data of 42 (cis). IR (®lm): n 1102 cm²¹s, 1249vs, br.

(C±O), 1377m, 1390m; 1752vs (CvO), 2245vw (CuN).

¹H NMR (400 MHz): d 1.04, 1.55, 1.83, 1.96 (4s, 4Me),

1.33, 2£1.364, 1.375 (4 d, J6.3 Hz, 4Me of 2iPr), 1.70,

1.76 (AB, J11.1 Hz, 3-H₂), 3.35, 3.83 (AX, J12.2 Hz,

5⁰-H₂), 5.08, 5.11 (2sept, 2CH of 2iPr). ¹³C NMR

(20.2 MHz): 21.14, 21.23, 21.29, 21.38, 26.6, 27.7, 29.0,

29.6 (8 q, 8Me), 34.8 (t, C-5⁰), 40.5, 47.4 (2s, C-2, C-4),

48.2 (t, C-3), 56.9, 61.7 (2s, C-4⁰, C-3⁰), 73.0, 73.6 (2d,

2CHMe₂), 115.9, 117.6 (2s, 2CN), 163.2, 163.8 (2s,

2CvO); C-1 covered by CDCl₃signal. Anal. calcd for

C₂₁H₃₀N₂O₄S (406.53): C 62.04, H 7.44, N 6.89, S 7.89;

found: C 62.13, H 7.34, N 6.95, S 7.88.

11. Huisgen, R.; Penelle, J.; Mloston, G.; Buyle Padias, A.; Hall,

H. K. J. Am. Chem. Soc. 1992, 114, 266.

12. Huisgen, R.; Li, X.; Giera, H.; Langhals, E. Helv. Chim. Acta

2001, 84, 981.

13. Mloston, G.; Huisgen, R.; Huber, H.; Stephenson, D. S.

J. Heterocycl. Chem. 1999, 36, 959.

14. (a) Hall, H. K. Angew. Chem., Int. Ed. Engl. 1983, 22, 440.

(b) Hall, H. K.; Buyle Padias, A. Acc. Chem. Res. 1990, 23, 3.

15. Huisgen, R.; Mloston, G.; Fulka, C. Heterocycles 1985, 23,

2207.

16. Huisgen, R.; Sturm, H.-J.; Wagenhofer, H. Z. Naturforsch

1962, 17b, 202.

17. These experiments of E. Langhals will be published in another

context.

18. Giera, H. PhD Thesis, University of Munich, 1991.

19. Bihlmaier, W.; Geittner, J.; Huisgen, R.; Reissig, H.-U.

Heterocycles 1978, 10, 147.

(d) NMR tests on steric course. Thiadiazoline 23

(0.269 mmol) and 18 (0.350 mmol) in CDCl₃(0.75 mL)

were reacted in the closed NMR tube at 808C for 5 min.

After cooling, sym-C₂H₂Cl₄(0.203 mmol) was added, and

20. (a) Huisgen, R.; Weinberger, R. Tetrahedron Lett. 1985, 26,

5119. (b) Weinberger, R. PhD Thesis, University of Munich,

1989.

1

the H NMR analysis (AB spectra of 5⁰-H₂) showed 41/

4263:37 and a yield of 89%. An analogous experiment

21. Burdisso, M.; Gamba, A.; Gandol®, R.; Pevarello, P.

Tetrahedron 1987, 43, 1835.

22. (a) Weber, A.; Sauer, J. Tetrahedron Lett. 1998, 39, 807.

with 19 provided 41/4249:51, also with 89% yield.

Acknowledgements

È

(b) Bohm, Th.; Weber, A.; Sauer, J. Tetrahedron 1999, 55,

9535.

We express sincere thanks to the Fonds der Chemischen

Industrie, Frankfurt, for the support of our work. G. M. is

grateful to the Alexander von Humboldt Foundation for a

stipend. Furthermore, we thank Helmut Huber for excellent

NMR spectra, Reinhard Seidl and Dr Werner Spahl for the

MS, and Helmut Schulz and Magdalena Schwarz for the

elemental analyses.

23. Elender, K.; Riebel, P.; Weber, A.; Sauer, J. Tetrahedron

2000, 56, 4261.

24. (a) Bartlett, P. D. Science 1968, 159, 833. (b) Bartlett, P. D.

Quart. Rev. (Chem. Soc.) 1970, 24, 473. (c) Bartlett, P. D.;

Mallet, J. J.-B. J. Am. Chem. Soc. 1976, 98, 143.

25. (a) Sustmann, R.; Rogge, M.; NuÈchter, U.; Bandmann, H.

Chem. Ber. 1992, 125, 1647. (b) Sustmann, R.; Rogge, M.;

NuÈchter, U.; Harvey, J. Chem. Ber. 1992, 125, 1657.

26. (a) LuÈcking, K.; Rese, M.; Sustmann, R. Liebigs Ann. 1995,

1129. (b) Reese, M.; Dern, M.; LuÈcking, K.; Sustmann, R.

Liebigs Ann. 1995, 1139.

References

1. 1,3-Dipolar cycloadditions, Part 123. For Part 122 see Ref. 8.

2. Review: Sauer, J.; Sustmann, R. Angew. Chem., Int. Ed. Engl.

1980, 19, 779.

27. Review: Huisgen, R. Acc. Chem. Res. 1977, 10, 117; see also

p 199.

28. (a) Huisgen, R.; BruÈckner, R. Tetrahedron Lett. 1990, 31,

2553. (b) BruÈckner, R.; Huisgen, R. Tetrahedron Lett. 1990,

2557, 2561.

3. Review: (a) Huisgen, R. 1,3-Dipolar Cycloaddition Chemistry;

Padwa, A., Ed.; Wiley: New York, 1984; Vol. 1, pp. 1±176.

(b) Huisgen, R. 1,3-Dipolar Cycloaddition Chemistry; Padwa,

A., Ed.; Wiley: New York, 1984; Vol. 1, pp. 126±128.

(c) Huisgen, R. 1,3-Dipolar Cycloaddition Chemistry;

Padwa, A., Ed.; Wiley: New York, 1984; Vol. 1, pp. 63±76.

4. Sustmann, R. Pure Appl. Chem. 1974, 40, 569.

29. (a) Urrutia Desmaison, G. PhD Thesis, University of Munich,

1986. (b) Huisgen, R. Adventure Playground of Mechanisms

and Novel Reactions. In Pro®les, Pathways, and Dreams,

Seeman, J. I., Ed.; American Chemical Society: Washington,

DC, 1994; pp. 180±182.

5. Review: (a) Kellogg, R. M. Tetrahedron 1976, 32, 2165.

(b) Huisgen, R.; Fulka, C.; Kalwinsch, I.; Li, X.; Mloston,

30. Diebert, C. E. J. Org. Chem. 1970, 35, 1501.

31. (a) Ingold, C. K.; Shaw, F. R. J. Chem. Soc. 1927, 2918.

(b) Huisgen, R. Angew. Chem., Int. Ed. Engl. 1970, 9, 751.

32. Linstead, R. P.; Whalley, M. J. Chem. Soc. 1952, 4839.

33. Kalinowski, H. O.; Berger, S.; Braun, S. ¹³C-NMR-Spectro-

scopy; George Thieme: Stuttgart, 1984 p 425.

34. Gotoh, T.; Buyle Padias, A.; Hall, H. K. J. Am. Chem. Soc.

1986, 108, 4920.

È

G.; Moran, J. R.; Probstl, A. Bull. Soc. Chim. Belg. 1984,

93, 511. (c) Mloston, G.; Heimgartner, H. Polish J. Chem.

2000, 74, 1503.

6. Huisgen, R.; Langhals, E.; Mloston, G.; Oshima, T.; Rapp, J.

J. Heterocycl. Chem. 1987, 24, S1.

7. Huisgen, R.; Mloston, G.; Langhals, E. J. Org. Chem. 1986,

51, 4085.

35. Langhals, E.; Langhals, H. Tetrahedron Lett. 1990, 31, 859.

36. Paquer, D.; Reffet, D.; Vazeux, M. Rec. Trav. Chim. Pays-Bas

1978, 97, 284.

8. Huisgen, R.; Mloston, G.; Langhals, E. Helv. Chim. Acta

2001, 84, 1805.

9. Huisgen, R.; Mloston, G.; Langhals, E. J. Am. Chem. Soc.

1986, 108, 6401.

37. Huisgen, R.; Kalvinsch, I.; Li, X.; Mloston, G. Eur. J. Org.

Chem. 2000, 1685.

10. Mloston, G.; Langhals, E.; Huisgen, R. Tetrahedron Lett.

1989, 30, 5373.

Article Doi

Cycloadditions of two thiocarbonyl ylides with α,β-unsaturated esters and nitriles: Steric course and mechanism

DOI: 10.1016/S0040-4020(01)01147-4

Source and publish data:

Authors:

Article abstract of DOI:10.1016/S0040-4020(01)01147-4

Full text of DOI:10.1016/S0040-4020(01)01147-4

Products guided by the article

R&D Labs maybe for 35234-87-8

Relevant to this article

Hot Product