close

Вход

Забыли?

вход по аккаунту

?

Intramolecular 1 3-Dipolar Cycloaddition Reactions.

код для вставкиСкачать
Volume 15
-
Number 3
March 1976
Pages 123-180
International Edition in English
Intramolecular 1,3-Dipolar Cycloaddition Reactions
By Albert Padwar*]
The intramolecular 1,3-dipolar cycloaddition reaction of suitably functionalized 1,3-dipoles
represents a general scheme for the synthesis of novel fused ring heterocycles. Such reactions
of a number of 1,3-dipoles are summarized and the general outline and potential analogies
for these reactions noted. While the immediate aim of this review is to survey and correlate
published work, it is hoped that general and specific points in need of study will be revealed
and will stimulate further work in this fertile field.
1. Introduction
Huisgen et al. were the first to recognize fully the general
concept and scope of 1,3-dipolar cycloadditions[’ -51, a reaction of considerable scope for the synthesis of five-membered
heterocyclic rings. 1,3-Dipolar cycloadditions are bimolecular
in nature and involve the addition of a 1,3-dipole to a multiple
bond system leading to five-membered heterocycles, according
to Scheme 1.
A 1,3-dipole is basically a system of three atoms over which
are distributed four 71 electrons as in the ally1 anion system.
The three atoms can be a wide variety of Combinations of
C, 0, and N. The dipolarophile can be virtually any double
or triple bond. The term “1,3-dipole” arose because in valence
bond theory such compounds can only be described in terms
ofdipolar resonance contributors, as shown for diazomethane.
[*] Prof. Dr. A. Padwa
Department of Chemistry,
State University of New York at Buffalo
Buffalo, New York 14214 (USA)
Angew. Chem. Int. Ed. Engl.
Vol. 1 5 ( 1 9 7 6 ) No. 3
The extreme 1.3-dipolar forms with their complementary nucleophilic and electrophilic centers readily explain the tendency
to undergo addition to 71 bonds. Indeed, it must be possible
to write 1,3-dipolar forms for all such species which undergo
this type of addition.
Huisgen et a/. have systematically studied the mechanism
of 1,3-dipolar cycloadditions[’ -3]. In the great majority of
1,3-cycloadditions, the reaction rate is not markedly influenced
by the dielectric constant of the solvent medium in which
the reaction is conducted. The independence of solvent polarity[‘], the very negative entropies of activation[71,and the stereospecificity and regiospecificity[21point to a highly ordered
transition state. In most instances of 1,3-dipolar cycloaddition
reactions, when two isomers are possible as a result of the
use of unsymmetrical reagents, one isomer usually predominates, often to the exclusion of the other[8,9!
The principal question that arises when considering the
regiospecificity of 1,3-dipolar additions is whether the two
new o bonds formed on addition of the 1,3-dipole to the
dipolarophile are formed simultaneously or one after the other.
The mechanism that has emerged from Huisgen’s group is
123
that of a single-step, four-center, "no-mechanism'' cycloaddition, in which the two new bonds are both partially formed
in the transition state, although not necessarily to the same
extent[' 31. A symmetry-energy correlation diagram reveals
that such a thermal cycloaddition reaction is an allowed process['- ". "!An alternative mechanism that has been proposed
is a two-step process involving a spin-paired diradical intermediate"'].
The regioselectivity of 1,3-dipolar cycloaddition reactions
has, until recently, been a most difficult phenomenon to
explain. Rationalizations of regioselectivity based on a concerted transition state model have invoked both electronic
and steric effects"
13]. A solution to the vexing problem
of regioselectivity in 1,3-dipolar cycloadditions has recently
been proposed by Houk et al.['4-'71, who used the frontier
orbital method for rationalizing the effect of substituents on
rates and regioselectivity of 1,3-dipolar cycloadditions.
According to the frontier orbital treatment of 1,3-dipolar cycl~additions['~'I, the relative reactivity of a given 1,3-dipole
toward a series of dipolarophiles will be determined primarily
by the extent of stabilization afforded the transition state
by interaction of the frontier orbitals of the two reactants.
Reactions have been
into three types depending
on whether the dominant interaction is between the highest
occupied molecular orbital (HOMO) of the dipole and the
lowest unoccupied molecular orbital (LUMO) of the dipolarophile, or the dipole LUMO and the dipolarophile HOMO,
or whether both these interactions are of equal significance.
The energies of both HOMOS are increased by the presence
of electron-donating (including alkyl) and conjugating substituents and the energies of both LUMOs are decreased by
electron-withdrawing and conjugating substit~ents['~''I.
This frontier orbital model also deals with the problem
of regioselectivity of addition. In order to predict regioselectivity, it becomes necessary to determine the relative magnitudes
of the coefficients in the H O M O and LUMO of the 1,3-dipole
and dipolarophile. The favored cycloadduct will be that formed
by union of the atoms with the largest coefficients. Further
mechanistic discussion is to be found in the original literature.
In spite of the copious literature dealing with bimolecular
1,3-dipolar cycloaddition reactions, intramolecular examples
have received only a minimum of attention. 1,3-Dipoles bearing a functional group able to behave as a dipolarophile
are extremely interesting substrates. In fact, the intramolecular
cycloaddition of a properly functionalized 1,3-dipole represents a general scheme for the synthesis of novel fused ring
heterocycles (see Scheme 2).
- 3 j
d
Scheme 2
groups['g! All these reactions are usually carried out by heating the two reagents together in an inert solvent, and the
products are often easily isolated in high yield. It should
be pointed out that the cycloadducts are not stable in all
cases and sometimes undergo interesting transformations.
Whereas the cycloaddition of a simple alkene to a simple
nitrone produces a simple isoxazolidine, a fused bicyclic isoxazolidine (2) can be formed when the alkene and nitrone
moieties are suitably arranged in the same molecule. This
useful synthetic idea was first investigated by LeBel et al.["],
using molecules in which the nitrone was separated by a
propylene or butylene chain from the alkene. The unsaturated
nitrone ( I ) was not isolated but'was generated either by
oxidation of an N-alkenylhydroxylamine or by condensation
of an unsaturated aldehyde with N-methylhydroxylamine.
H~C-N-CHZ(
QH
CHz)sCH=CHz
0
HzC=CH(CHz)sC,//
When the cycloaddition was carried out with trans- and
cis-5-heptenal, two different isoxazolidines, (3a) and (3 b),
respectively, were obtained. It is apparent from these results
that the intramolecular addition is kinetically controlled and
that the configuration of the olefin is retained in the product
isoxazolidine. Internal cycloaddition to give the cis fused bicy-
1
1
3
0
9
HE-NHOH
CH3
0
%
I
CH3
(3a)
clo[3.3.0]octane skeleton is preferred since closure to the highly
strained trans fused isomer would involve a transition state
of much higher energy.
A preference was found for the formation of the (fused)
bicyclo[3.3.0]octane system ( 4 ) , but the (bridged) bicyclo[3.2.l]octane system (5) appeared when bond formation
between the 2-position of the alkene and the a-position of
the nitrone was hindered by substitution in those positions["].
In the present paper, the findings dealing with intramolecular t,3-dipolar cycloadditions will be reviewed.
2. Nitrones
1,3-Dipolar cycloadditions of nitrones have been reported
with a variety of alkynes, alkenes, isocyanates, isothiocyanates,
thiocarbonyl compounds, phosphoranes, sulfenes, and sulfinyl
124
For example, when only one of the substituents R', RZ,
R3,and R4 is a methyl group and the others hydrogen atoms,
only the fused product ( 4 ) is formed, but when R ' = R Z = C H 3
Anguw.'Chem. Int. Ed. Engl. 1 Vol. IS (1976) No. 3
and R 3 = R 4 = H approximately equal amounts of fused ( 4 )
and bridged ( 5 ) products are formed. The presence of methyl
groups in these positions (i.e. R ' = R Z = C H 3 ) forces them
to be strongly eclipsed in the transition state leading to the
normal fused product ( 4 ) . In this case, the severe steric interactions allow for the competing formation of the bridged product is}.
isoxazolidine into nitrone (I 0) followed by ring closure to
give a product with the opposite stereochemistry. This isomerization represents an example of the reversibility of the intramolecular cycloaddition of an unsaturated nitrone. Retro 1,3dipolar additions have been previously noted with intermolecular examples of nitrone-olefin ad duct^['^] and provides reasonable analogy for the above interconversion.
R'
R'
R
syn -( 1 2 )
syn-(12)
anti-(12)
?rR
H
R
H
H
cis
The cycloaddition reaction was extended to include the
homologous nitrone (7)[22J.Mercuric oxide oxidation of N-6heptenyl-N-methylhydroxylamine (6) afforded a mixture of
cis- and trans-N-methyl-8-oxa-7-azabicyclo[4.3.0]nonane(8)
(ratio 2: 1) as well as lesser quantities of a product having
the bicyclo[4.2. llnonane structure ( 9 ) .
OH
I
H3C-N-CH2(CH2)4CH=CH2
%
16 I
1
(7)
CHzl
(8)
(9)
Under conditions of kinetic control, the cis ring fusion
[cf. (S)] predominated in the condensation of 6-heptenal
with methylhydroxylamine. However, the major product
(1 1 a ) derived from (+)-citronella1 and N-substituted hydroxylamines was shown to have the trans ring fusion. In this
300 "C
m
-0%
I
R
(10)
+
o
m
I?
R
I
R
(Ila)
(Ilb)
major product
minor product
case the stereochemistry was found to be temperature dependent; increasing proportions of ( I 1 b ) were obtained at high
reaction temperatures[231.At 300 "C the trans fused isoxazolidine ( I 1 a ) undergoes partial isomerization to the cis isomer.
This process was rationalized by thermal conversion of the
Angew. Chem. Int. Ed. Engf. 1 Vui. 15 (1976) Nu. 3
RR
0-N
M
H
H R
trans
In order to rationalize the stereochemical results, LeBel
suggested that the transition state leading to cis ring fusion
requires the potential six-membered carbocyclic ring to adopt
a twist conformation. A slightly deformed chair arrangement
will produce the trans-isomer. Examination of models indicates
that orbital overlap for 1,3-addition is favorable for a cis
ring juncture for both syn- and anti-configurations of the
intermediate nitrone (I 2). Effective overlap in the transition
state for formation of a trans-fused product is essentially precluded with an anti nitrone, and only the more favorable syn
arrangement can lead to product. It is quite reasonable to
assume that the ground state equilibrium composition of the
nitrone favors the anti configuration. If the barrier to syn+anti
isomerism is of the order of magnitude of the activation energies for intramolecular addition, then cis product will be
favored. O n the other hand, if the barrier to rotation in
the nitrone is low, then the ratio of products will depend
only on the difference in free energy levels of the respective
transition states (Curtin-Hammett principle). In the transition
state leading to trans isomer, a significant interaction develops
between the 1-H and the 5-methylene group which allows
the seemingly less favorable twist arrangement of the tetramethylene chain leading to cis-isomer to compete successfully.
This would account for the observed isomer ratio of 2.1
(cisjtrans) in the condensation of 6-heptenal with methylhydroxylamine. With citronellal, the trans product is more stable
than the cis compound, and product development control
predominates.
LeBel et al. further demonstrated the utility and synthetic
scope of the intramolecular 1,3-dipolar cycloaddition of
nitrones by preparing a variety of polycyclic isoxazolidines[21. 2 5 - 2 7 ]. Examples include the syntheses of compounds
(141, ( 1 5 ) , (171, (191, f21), ( 2 3 ) , and ( 2 5 ) .
These ring closure reactions were suggested to be initiated
by electrophilic attack of the nitrone on the olefin, the nitrone
carbon atom thus resembling a carbenium ion center[25! No
125
with bridged structures such as (32) and (35) were also
formed in some cases[28-301.
HaC-NHOH
9-8
0
Also, the ether oxygen can be replaced by a formamido
group which results in the formation of the related tetrahydroquinoline product (37)[281.
H3C'
1
CHO
(36)
1
CHO
(37)
When the alkene functionality is attached to the nitrone's
nitrogen atom, only bridged structures are formedr3']. In the
case of N-3-butenyl nitrone (38), the terminal alkene carbon
atom takes part in CC bond formation.
distinction was made between a one-step process or a two-stage
mechanism involving a rapid second step.
Oppolzer et aL.[zs-301have extended the intramolecular 1,3dipolar cycloaddition reaction to include nitrone-alkenes
separated by chains that generally incorporate a benzene ring
and a heteroatom to produce isoxazolidines ( 2 7 ) and (29).
In most of the cases examined the isoxazolidine was found
to contain a cis-ring junction between the heterorings. Adducts
126
Cycloaddition fails for the N-butenyl system where R = CH3,
presumably owing to considerable steric hindrance in the
transition state. The preferential formation of (40) rather
than ( 4 1 ) has been attributed to steric destabilization of the
transition state for the formation of the latter c ~ m p o u n d ~ ~ ' ~ .
Inspection of Dreiding models of nitrone ( 3 8 ) reveals that
significant o overlap between 0 and C-4 and between C-2
and C-3 in transition state ( 3 9 ) cannot develop simultaneously
unless the latter assumes product-like geometry[311.The preference for (40) could also be explained by electronic factors.
The formation of this isomer corresponds to the electronically
favored addition direction of a nitrone to a monosubstituted
ethylene'32!
Angew. Chem. Int. Ed. Engl.
Vol. 15 ( 1 9 7 6 ) No. 3
,=*om
I CH3
CH3
(43j
(42a). n = 2
(42b). n = 3
(42c), n = 4
I
I
C-5 atom of the cis-cyclodecenone ring are too far apart
to permit the internal cyclization reaction to occur.
Bapat et ~ l . [ ~ have
' ] recently described the thermal behavior
of a cyclic nitrone, 5-allyl-3,3,5-trimethyl-l-pyrroline
N-oxide
(52), which gives both a Cope rearrangement product ( 5 3 )
and an internal cycloaddition product ( 5 4 ) when heated in
an inert solvent[3'! The proportions of the two rearranged
products varied as a function of the thermal conditions used.
A
/
n.4
H
CH3
H3C
CHs
The intramolecular cyclization of allenic nitrones was found
to afford several novel bicyclic
Treatment of
5,6-heptadien-2-one with N-methylhydroxylamine gave a transient nitrone (42a) which produced the unsaturated bicyclic
isoxazolidine (43) by intramolecular cycloaddition at the terminal double bond of the allenic function. Cyclization of
the homologous nitrone (42 b ) gave bicyclic adduct (44) as
well as the stereoisomeric isoxazolidihes (46) and (47). The
data indicate that the exo-methyleneisoxazolidine (45) is
initially formed by addition of the nitrone to the internal
bond of the allenic group. Subsequent acid catalyzed addition
of ethanol produced the ethers (46) and (47). A related
set of reactions occurred with acetylenic nitrone (49)[33'.
Heating nitrone (52) in boiling toluene for 1 h gave the Cope
rearrangement product (53) as the only isolable product.
The rearranged nitrone (53) was slowly converted into cycloadduct (54) in boiling xylene, and this cycloadduct could
most conveniently be prepared by heating (52) in boiling
xylene for 5 h. Thus the formation of cycloadduct (54) can
be attributed to cyclization of the rearranged nitrone (53).
The intramolecular dipolar addition of 5-allyl-1-pyrroline
N-oxide (55) has been found to give a cycloadduct (56)
which is a useful precursor of the tropane skeleton13"! This
conversion was achieved by rnethylation of (56) with methyl
iodide to give the quarternary salt (57) which gives pseudotropine ( 5 8 ) upon reduction with lithium tetrahydridoaluminate.
N-0
(49)
Treatment of 7,8-nonadien-2-one with N-methylhydroxylamine gave nitrone (42 c ) which cyclized to bicyclic adduct
(48) by addition of the nitrone moiety to the internal carboncarbon double bond of the allenic function. Thus the product
formed in these intramolecular cyclizations depends on the
separation of the allene and nitrone groups.
Reaction of the l(l0)-unsaturated trans-5-0~0-5,lO-secosteroid (50) with N-methylhydroxylamine has been reported to
give a non-isolable nitrone which undergoes spontaneous transannular cyclization with formation of the N-methylisoxazolidine derivative (52) in over 60% yield[34]. The geometry
of the trans-cyclodecenone ring in the modified steroid com-
An elegant approach to the ring skeleton of histrionicotoxin
involving an intramolecular cyclization of a nitrone moiety
with an activated olefin has recently been attempted by two
groups of
381. Cyclization of nitrone ester (6 0 )
was considered to offer high promise for producing the desired
ring system (61) of histrionicotoxin. However, treatment of
nitro-ketone (59) in aqueous methanol with zinc and
ammonium chloride gave only cycloadduct (62) via spontaneous cyclization of a transient nitrone intermediate (60).
Chemical evidence for the structure of (62) was obtained
pound is favorable for transannular 1,3-dipolar cycloaddition.
In contrast, the diastereomeric l(l0)-unsaturated cis nitrone
system does not undergo intramolecular 1,3-dipolar addition
since the double bond between C-I and C-10 and the trigonal
Angrw. Chem. I n t . Ed. Engl.
1 Vol. IS
(1976) No. 3
127
by cleavage of the nitrogen-oxygen bond with zinc and acetic
acid to the hydroxy lactam (63). The formation of (62)
rather than (62) from the nitrone cyclization was attributed
to the greater steric interactions in the transition state leading
to (61)[371.
It is interesting to note that cyclization of the closely related
unsaturated nitrone (65) resulted in the formation of both
regioisomers (67) and (68)[381.Nitrone (65) as well as smaller
quantities of the isomeric nitrone (66) were generated by
oxidation of the hydroxylamine derivative (64). The major
cycloadduct (67) was thermally isomerized to the thermodynamically more stable compound (68) having a bicyclo[3.2.l]octane partial structure, presumably by way of a 1,3dipolar cycloreversion. Reduction of the nitrogen-oxygen bond
of (68) afforded cis-l-azaspiro[5.5]undecan-8-ol (70) which
contains the skeleton of histrionicotoxin (without side
chains)I37. *I.
6
CH=N-NHTs
(75)
(64)
Y
ni\
I
HO
LiAIHn
H
(70)
3. Diazoalkanes
1,3-Dipolar cycloadditions of diazoalkanes to multiple
bonds was first observed in 1888 by B ~ c h n e r [ ~ "Since
.
the
initial report, many examples have appeared in the literature" - 5 ] . There are, however, only a few reports dealing with
the intramolecular addition of diazoalkanes to carbon-carbon
double bonds. Kirrnse and Dietrich reported that the thermolysis of (71) gave 1-pyrazoline (72) in 80% yieldf4']. On
further heating, the initially formed I-pyrazoline lost nitrogen
to give the corresponding cyclopropane (73). In contrast
to the thermal results, photolysis of (71) produced a mixture
of (73) and (74).
N=N
f 74)
(731
The major product of the reaction of 4-cycloheptenecarbaldehyde tosylhydrazone (75) was found to be 5,6-diazatricy128
(76)
when the lithA similar internal cycloaddition
ium salt of tosylhydrazone (77) was heated at 120°C. The
cycloadduct obtained (78) (81 %)was converted into cyclocopacamphene (79) on further photolysis. The facile intramolecular 1,3-dipolar cycloaddition of the intermediate olefinic
diazoalkane to give (78) is presumably related to the proximity
I
OH
cl0[5.3.0.O~~~]dec-5-ene(76)[4'1.This product was formed by
intramolecular 1,3-cycloaddition of the diazomethane part
of the molecule to the double bond. The authors noted that
the double bond has a considerable activating influence on
the decomposition of the tosylhydrazone anion and that the
cycloaddition of the diazomethane moiety occurred much
faster than the competing reaction of carbene formation. Normally pyrazoline formation with unactivated double bonds
is a very sluggish reaction['! The significant lowering of the
temperature at which the tosylhydrazone anion decomposed
was attributed to double-bond participation[411.
Y
Y
of the two reacting functional groups and to the fairly strained
nature of the olefinic double bond.
Eost et
441 have recently developed a simple approach
to the 7,8-diazatetracyclo[3.3.0z~4.03~6]oct-7-ene
system (82).
Irradiation of diazo compound (80) generated the tetracyclic
framework (82) in reasonable yield. The success of the photol~
2
1
.
~
~
~
9
ysis of (80) in producing (82) is quite interesting in light
of Hoffunn's calculations regarding the geometry for singlet
carbene addition~["~I.
These calculations indicate that for a
concerted cycloaddition the p orbital of the carbene overlaps
in a 0 manner with one end of the ethylene n bond. This
lopsided configuration is unattainable in the carbene derived
from (80) without severely distorting the rigid bicyclic framework. The authors suggested that the formation of (82) might,
in fact, proceed by initial 1,3-dipolar addition of the diazo
group to the double bond to produce pyrazoline (81). Extrusion of nitrogen from this transient intermediate could then
account for the formation of (82). The photoconversion
(80) --t (81) would have to occur from some high vibrational
level of the ground state since photochemical 1,3-dipolar cycloadditions are symmetry forbidden.
Angew. Chem. Int. Ed. Engl.
/
Vol. 15 ( 1 9 7 6 ) No. 3
In 1935 it was briefly reported that 3-diazopropene slowly
decomposes at room temperature to give p y r a z ~ l e [471.
~ ~Sub.
sequent investigations showed that this is a general reaction
of this class of diazo c o m p o ~ n d s ~ ~The
~ - ~small
~ 1 . p value
for the cyclization of l-aryl-3-diazopropenes, the rate enhanceArCH=CH-CHN2
-
(83)
Ar
Ar
N
azide other than hydrogen azide adding to a nitrile was
observed with 2'-azido-2-biphenylcarbonitrile (92)C5"'. Thermolysis of (92) resulted in the formation of tetrazolophenanthridine (93). This transformation can best be described as
an intramolecular 1,3-dipolar cycloaddition reaction.
Q-pLc+p
L$
CN N3
I
(84)
/
ment by aryl conjugation with the carbon-carbon double bond,
and the relative insensitivity of the reaction rate to structural
changes led Hart and Brewbaker'"' to conclude that cyclization of 3-diazoalkenes to pyrazoles is an intramolecular, concerted 1,3-dipolar cycloaddition.
A similar reaction has been reported['!' to occur upon
thermal decomposition of the bis-diazo ketone (86). The formation of a trimer ( 8 9 ) in this reaction suggested the intermediacy of 3,Sdiphenyl-I ,2-diaza-4-cyclopentadienone(88). The
formation of (88) was proposed to involve an electrocyclic
I
KN9N
(85)
(931
(92)
The synthesis and thermal decomposition of various olefinic
azides (94) to give l-azabicycIo[3.1.O]hexanes ( 9 6 ) and cyclic
imines ( 9 5 ) has been described by L o g ~ t h e t z d ~The
~ J .azides
were decomposed under a variety of conditions. When the
thermolysis was carried out in a hydrocarbon solvent at 8 0 T ,
25 "C
ring closure of the bis-diazo ketone (86) to give (87), followed
by rapid disengagement of nitrogen. The conversion of ( 8 6 )
into (87) represents another example of the intramolecular
cycloaddition of vinyldiazomethanes to p y r a ~ o l e s-['O1.
~~
4. Azides
Organic azides are well known to behave as 1,3-dipoles
in thermal cycloaddition reactions[' -31. The first example of
this reaction was observed by Michael in 1893'"'. With azides,
intramolecular cycloadditions have been occasionally
reported[53*
541, but systematic data are available only for
a series of azid~alkenes[~~!
Uhle reported that the displacement of the p-toluenesulfonatefunction of the pseudodiosgenin
derivative ( 9 0 ) with potassium azide in dimethylformamide
is followed by 1,3-dipolar cycloaddition to the enol ether
olefinic bond to furnish a triazoline derivative (91 )Cs31.
the decomposition products comprised a mixture of cyclic
imines and l-azabicyclo[3.l.O]hexanes. On the other hand,
when the olefinic azides were allowed to stand at 25°C for
2 months, the isomeric triazolines ( 9 7 ) were formed in quantitative yield. Decomposition of the triazolines under conditions
identical with the decomposition of the corresponding azides
gave ( 9 5 ) and (96). On the basis of these results, it was
suggested that the triazolines are intermediates in the decomposition of all the olefinic azides studied. Ally1 azide and
4-azido-I-pentene cannot undergo internal 1,3-dipolar cycloaddition to give a triazoline similar to (97) because of
the high degree of strain in such a structure. As a result, their
decomposition is very sluggish and no products analogous
to ( 9 5 ) and ( 9 6 ) are formed. The mechanism proposed to
explain the formation of ( 9 5 ) and ( 9 6 ) is outlined below.
A nitrene intermediate was completely excluded in these reactions.
(941, n = 2
198)
The intramolecular 1,3-dipolar cycloaddition of aryl azides
bearing alkenyl, alkynyl, and nitrile groups has recently been
reported by Fusco et a1.[561.
The decomposition of ( 9 9 ) carried
out in refluxing benzene gave aziridines (101). The formation
of these products involves initial intramolecular cycloaddition
Although less reactive than CC triple bonds, CN triple
bonds are known to undergo 1,3-dipolar cycloadditions with
some organic azides"'. The first authenticated case of an
Angew. Chem. Inr. Ed. Engl.
1 Vol. 15 ( 1 9 7 6 ) No. 3
-
1991
R' = R 2 = II;
1
R2
(1001
R' = H , R Z = CH,;
(101)
R' = P h , R 2 = H
129
leading to the unstable triazoline system ( l o o $ The intervention of such an intermediate was deduced by carrying out
the decomposition of ( 9 9 ) at room temperature in hexadeuteriobenzene and monitoring the reaction progress by NMR
analyses. In addition to the signals of the starting azide and
the final product, the spectrum clearly showed a set of signals
corresponding to triazoline ( 1 00).
Thermolysis of the corresponding alkyne ( 10 2 ) gave triazole
( 1 0 3 ) as the only decomposition product. This is the first
reported example of the intramolecular cycloaddition of the
azido group to an acetylenic function. While the transformation of azide ( l 0 2 b ) was complete after 3 h, under the same
conditions compound (102 a ) entirely disappeared after
15min. The greater reactivity of (102a) in comparison to
(102 b ) was somewhat unexpected considering that the conjugated alkynes are usually better dipolarophiles than the unconjugatedL5'1. The bulky phenyl substituent present in (102b)
apparently hinders the approach of the reactant groups and
consequently decreases the rate of cycloaddition.
NN
,
ilO2a), R = H
(IOZb), K = P h
/103aJ, H = H
1103b), R = P h
investigation, Hall et ul.[591suggested that the cyclization
of o-azidobenzophenones (106) to 3-aryl-2,l -benzoisoxazoles
(107) can best be explained as proceeding uia an intramolecular 1,3-dipolar cycloaddition across the carbonyl function.
The following mechanism was proposed to account for the
kinetic observations.
5. Azomethine Imines
These 1,3-dipoles are normally not isolable, but their facile
preparation in situ from storable precursors makes them easy
to use. One convenient procedure for the synthesis of a variety
of azomethine imines involves the reaction of 1,2-disubstituted
hydrazines with carbonyl compounds[31.Huisgen's group has
shown that the 1,3-dipoles prepared in this fashion are readily
intercepted with olefinic dipolarophiles to give pyrazolidinesL31. O p p o l z ~ [extended
~~]
this reaction to include systems
which contained the olefin and azomethine imine portions
in the same molecule[61.6 L 1 . For example, treatment of o-allyloxybenzaldehyde (108) with N-methyl-N'-phenylacetylhydrazine gave cycloadducts (110) and ( I l l ) . This reaction presumably proceeds through the intermediacy of azomethine
imine (109) which undergoes a subsequent intramolecular
cycloaddition.
The ring closure of the related azidonitrile system (104)
has also been studied. The formation of tetrazole derivatives
(105) was interpreted in terms of an intramolecular 1,3-dipolar
cycloaddition. In intermolecular terms, only nitrile groups
activated by electron-withdrawing substituents have been
shown to behave as dipolarophiles toward azides['l. The
N=N
3NJQ
X-CHZCN
I
fI05)
X = 0, CH2
mutual ortho disposition of the two interacting groups in
(104) provides a favorable stereochemical relationship for
intramolecular cycloaddition. Some activating effect by the
oxygen atom could be responsible for the greater reactivity
in the case X = 0.
Ortho-substituted phenyl azides are known to decompose
much more rapidly than the corresponding meta or para
isomers. This rate enhancement has been noted primarily
in those systems where the ortho substituent has some type
of a$-unsaturation and in which the decomposition leads
to cyclization. Several workers[581have suggested that the
reaction does not involve a nitrene intermediate but instead
proceeds via a concerted mechanism in which cyclization and
loss of nitrogen are concurrent. On the basis of a kinetic
Ar
OCH ZPh
CXJN
I
(104~
H3C,$,g-C
r
Ar
1
Similarly, reaction of the related acetylenic derivative ( 1 1 2 )
with N-acetyl-N'-methylhydrazine gave cycloadduct (I 1 3 ) in
high yieldr621.
+ HsC-NH-NH-COCH3
O-CH zC Z CR
A
Oppolzer[621also found that the ether oxygen could be
replaced by a formamido group, which results in the formation
of the related cycloadduct (1 15), X = NCHO.
Ar
0
(1061
- N,
Ar
a
P
(107)
130
AngeM,. Chem. Inr. Ed. Engi. / Vol. 15 (1976) No. 3
The cyclization was extended to include the reaction of
N-benzyl-N-(4-pentenoyl)hydrazine (1 16) with benzaldehyde
to give cycloadduct (1 I 7)I6I1. Similarly, treatment of ( I 18)
with formaldehyde produced ( 1 19). Both sets' of reactions
proceed through the intermediacy of an azomethine imine.
/COCH3
N
~ F - N H - C
OCH~
+ CHzO
f
CHzPh
E
NH-NH-CHZPh
+
PhCHO
&yA!J~Ph
__+
6. Nitrile Imines
(117)
o".:,
0
0
CHzPh
N-Acyl-N'-alkylhydrazines react with an aldehyde only on
the alkyl substituted nitrogen, thus permitting a regioselective
preparation of pyrazolidines with different N-substituents.
Oppolzer showed that the intramolecular I,3-dipolar cycloaddition reactions of these azomethine imines represent a convenient and simple method for the synthesis of some novel
diazabicyclic ring systems. For instance, reaction of (120)
as well as the homologous olefin (122) with formaldehyde
gave the 1,7-diazabicyclo[2.2.l]heptane (221) and 1,7-diazabicyclo[3.2.l]octane (223) in high yield[6'!
Nitrile imines are compounds containing the grouping
e e
R-N-N-C-R'.
The most convenient method for generation of this class of 1,3-dipoles involves the action of tertiary
bases on hydrazidoyl halides such as (129) and (131)[63!
Here again the 1,3-dipole cannot be isolated and the external
dipolarophile must be present in the reaction mixture from
the outset. Treatment of (129) with base leads to a nitrile
imine which can be readily intercepted by the adjacent acetylenic functionality present in the molecule to give the pyrazole
(130)[641.Similarly, treatment of (131) with base affords cy-
0
,COCH3
N
When the condensation was carried out between (122)
and benzaldehyde, a mixture of two structural cycloadducts
was obtained. In this case, cyclization of the transient azomethine imine (124) occurs spontaneously and produces a mixture
of ( 1 2 5 ) and (126).
(124)
7. Nitrile Ylides
Nitrile ylides are a young but thoroughly investigated class
of 1,3-dipoles[' 3! Access to this class ofdipoles can be realized
by (a) treatment of imidoyl halides with base'"], (b) thermal
or photochemical elimination of phosphoric acid ester from
2,3-dihydro-l ,4,2h5-oxazaphospholes[681,
and (c) photolysis of
2 H - a ~ i r i n e s l701.
~ ~The
. greatest opportunity for structural variation is offered by the latter route. P u d w ~ [and
~ ~ S~hrnid['~]
]
have independently shown that irradiation of 2H-azirines leads
to irreversible ring opening and the formation of nitrile ylides
as intermediates. These species may be intercepted with a
wide variety of dipolarophiles to form five-membered ring
heterocycles.
Whereas the cycloaddition of 2H-azirines with simple olefins
produces I-pyrrolines, a rearranged isomer can be formed
when the alkene and azirine moieties are suitably arranged
in the same molecule. This intramolecular cycloaddition was
first observed by Padwu et al.[71* using 2-vinyl substituted
azirines. Irradiation of (233) in benzene afforded a 2,3-disub-
kPh
(126), minor product
(1251, major product
The related N-acetyl-N'-(0-vinylbenzy1)hydrazine ( 1 27)
undergoes a similar intramolecular cycloaddition to produce
the partially hydrated 2,5-imino-2-benzazepine (128) in 40%
yield.
Angew. Chem. l n l . Ed. Engl.
cloadduct (132). These ring closures are particularly interesting
in that they involve cycloadditions with unconjugated alkynes,
substrates which are generally unreactive toward nitrile
imines[''! Furthermore, while the usual orientation of monosubstituted acetylenes in the intermolecular cycloadducts with
nitrile imines is such as to afford 5-substituted pyraz o l e ~ [65.
~ ~ *obtainment of the 4-substituted pyrazole (132)
reveals that in a properly chosen substrate the regiospecificity
of monosubstituted acetylenes can be inverted.
1 Vol. 15
( 1 9 7 6 ) No. 3
131
stituted pyrrole (134) while thermolysis gave a 2,5-disubstituted pyrrole (135). Photolysis of (136) proceeded similarly
and gave 1,2-diphenylimidazole(137) as the exclusive photoproduct. This stands in marked contrast to the thermal reaction
of (136) which afforded 1,3-diphenylpyrazole (138) as the
only product.
Ph
(138)
The evidence obtained clearly indicates that the photorearrangements of (133) and (136) proceed by a mechanism
which involves a nitrile ylide intermediate. This conclusion
was reached by carrying out the irradiation of (133) in the
presence of a trapping agent (140). Under these conditions,
the formation of the 2,3-disubstituted pyrrole (134), which
is formed in high yield in the absence of a trapping reagent,
is entirely suppressed. Intramolecular cycloaddition of the
nitrile ylide intermediate (139) followed by a 1,3-sigmatropic
hydrogen shift of the initially formed five-membered ring readily accounts for the formation of (134) as the final product.
A
1139)
H
1
HZC=CH-CO~CH,
P
h
trans-isomerizationabout the CC double bond. The preference
for cyclization of (143) to a seven-membered ring was attributed to stereoelectronic factors. Closure of the linear dipolar
intermediate (143) obtained from (142) occurs more easily
via a seven-membered transition state and leads to the preferential formation of benzazepine (144). Cyclization of the nitrile
ylide derived from the trans isomer to a seven-membered
ring is precluded on structural grounds and formation of
2,3-diphenylpyrrole (146) occurred instead.
Nitrile ylides have been classified by Huisgen" - 3 1 as nitrilium betaines, a class of 1,3-dipoles containing a central
nitrogen atom and a CN n-bond orthogonal to the 4n-ally1
system. Among the possible resonance forms of a nitrile ylide,
a carbene structure can be envisioned which makes conceivable
a 1,l-cycloaddition of this 1,3-dipole. Recently, Padwa and
Carlsen reported the first example of such a cy~loaddition['~!
w
(140)
R
C02CHs
The thermal transformations observed with these systems,
on the other hand, have been rationalized in terms of an
equilibration of the 2H-azirine with a transient vinyl nitrene
( 1 4 1 ) which subsequently rearranges to the 2,5-disubstituted
pyrrole.
(141)
In contrast to the photochemical results encountered with
(133), the presence of a cis-styryl side chain in the 2-position
of the azirine ring (142) leads to ring expansion and gives
benzazepine (144). The 2-[2-(a- and P-naphthyl)vinyl1-3-
phenyl-2H-azirines behaved similarly and proceeded with
complete regiospecificity.The photolysis of the isomeric transstyrylazirine (145) took an entirely different course and produced 2,3-diphenylpyrrole (146) as the major product. The
results indicate that opening of the azirine ring followed by
intramolecular cyclization proceeds at a faster rate than cis132
These authors found that irradiation of 2-allyl-2-methyl-3phenyl-2H-azirine (147) gave 3-methyl-I-phenyl-2-azabicyclo[3.1.0]hex-2-ene (148) as the exclusive photoproduct. Photolysis of (147) in the presence of excess dimethyl acetylenedicarboxylate resulted in the trapping of the normal nitrile
ylide and afforded cycloadduct (149) in high yield. Under
these conditions, the formation of (148), which is produced
in quantitative yield in the absence of a trapping reagent,
was entirely suppressed.
The photoreactions of the closely related methyl-substituted
azirines (150) and (152) were also examined. Irradiation
of (150) produced azabicyclohexene ( 1 5 1 ) while photolysis
of (152) gave the analogous compound (153) as the primary
photoproduct. Upon standing at room temperature, (153)
epimerized to the thermodynamically more stable exo-isomer
( 1 54).
Theformationofthe photoproducts (148), (15I),and (153)
was found to be completely suppressed when the irradiations
were carried out in the presence of methyl acrylate (140).
The only products formed under these conditions were the
usual I-pyrrolines.
Angew. Chem. I n t . Ed. Engl.
1 Vol. 15 (19761 No. 3
sonable explanation to account for the observed stereoselectivity is that the 1.1-cycloaddition process occurs by initial
attack of the carbene carbon on the terminal position of
the double bond. Such an attack will generate a six-membered
ring dipole ( 1 5 6 ) which contains a secondary carbenium ion
Padwa and C~rlsen[’~]
have pointed out that bimolecular 1,3dipolar additions proceed via a “two-plane’’ orientation
complex in which the dipole and dipolarophile approach each
other in parallel
In order to achieve this type
of transition state, the linear nitrile ylides derived from the
irradiation of these allylazirines must first bend. This involves
disruption of the orthogonal TC bond at some modest energy
cost but leaves the allyl anion n system undisturbed. The
cycloaddition of allylazirines (147), (150), and (152) with
added dipolarophiles proceeds in this fashion and affords
1-pyrroline derivatives [e. 9. (149)] as the primary cycloadducts. Inspection of molecular models of the allyl substituted
nitrile ylides indicates that the normal “two-plane’’ orientation
approach of the linear nitrile ylide and the allyl 71 system
is impossible as a result of the geometric restrictions imposed
on the system. Product formation is possible, however, if
the linear nitrile ylide undergoes rehybridization to give a
species of bent geometry which subsequently undergoes 1,lcycloaddition with the neighboring double bond. The most
favorable transition state for the 1 ,l-cycloaddition reaction
is one in which the rc-orbitals of the nitrile ylide and olefinic
double bond are orthogonal. This orthogonality could, in
principle, permit the occurrence of a concerted orbital symmetry-allowed [,2, + n2,] process and thus accommodate the
formation of the thermodynamically less favored endo isomer
(153).
More recent work[741,however, has shown that irradiation
of the cis-substituted azirine ( 1 5 5 ) also produced azabicyclohexene (253). This observation indicates that the 1,l-cycloaddition process is stereoselective but not stereospecific. A rea-
CH3
(15-5)
Anyew. Chnn. lnt. Ed. Enyl. f Vil.
11-53)
IS ( 1 9 7 6 ) Nu. 3
as well as an azaallyl anion portion. Collapse of this new
1,3-dipole to the thermodynamically favored exo product
( 1 5 4 ) will result in a severe torsional barrier on closure.
O n the other hand, collapse to the thermodynamically less
favored endo isomer (153) moves the phenyl and methyl
groups increasingly further apart and would account for the
formation of the less stable product.
Supporting evidence for this suggestion was obtained by
irradiating the isomeric 3-methyl-2H-azirine system ( 157).
Photolysis of this azirine results in the quantitative formation
of azabicyclohexene (148). This is the same azabicyclohexene
that was formed from the irradiation of azirine (147). A
control experiment showed that ( 1 4 7 ) and ( 1 5 7 ) were not
interconverted under the photolytic conditions.
h
Ph
e
CH3
(148)
The formation of ( 1 4 8 ) ft‘om (157) is perfectly consistent
with the generation of a six-membered ring dipole (160).
Thus, cyclization of the initially formed carbene ( 1 5 8 ) generates (160) which closes to azabicyclohexene (148). The same
six-membered ring dipole is also formed from azirine (147).
Thus the stereoselectivity observed with azirines ( 1 5 2 ) and
(155) as well as the regiospecificity encountered with azirines
(147) and (157) can be attributed to a two-step cyclization
path which involves a six-membered ring dipole. The formation
of endo-azabicyclohexene ( I 5 3 ) from methylazirine ( I 61 ) provides additional support for this interpretation.
When the chain between the azirine ring and the alkene
end is extended to three carbon atoms, the normal mode
of 1,3-intramolecular dipolar cycloaddition occurs. For
example, irradiation of azirine (162) gave I-pyrroline ( 1 6 3 )
133
8. Carbonyl Oxides
Carbonyl oxides (Criegee zwitterions) have never been isolated, but indirect evidence for their existence is abundantL78,’?These unusual species are usually generated by
ozonolysis of a suitable olefin, although more recent evidence
has been presented for their formation by reaction of a carbene
with oxygen[80! Criegee presented evidence which indicates
that ozonides may be formed by the 1,3-dipolar addition
of carbonyl oxides to carbonyl compounds[78’.Some recent
work by Criegee’s group has shown that the ozonolysis of
1,2-disubstituted 1 -cyclopentenes, e. g. ( I 71),containing different substituents in the 4-position of the ring produces carbonyl
oxides such as ( 1 7 2 ) as reaction
These
dipoles undergo intramolecular 1,3-dipolar cycloaddition with
each of the two functional groups present. Thus two different
ozonides are formed, in the present case (173) and (174).
in quantitative
In this case, the methylene chain
is sufficiently long to allow the dipole and olefinic portions
to approach each other in parallel planes.
(162)
i
J
(163)
H2C =CH
’ In a related case, Schmid et a!. reported on the photoisomerization of 2-isoxazoline ( 1 6 4 ) to oxazoline (167)”61. The reaction was proposed to proceed via a transient azirine (165).
This intermediate was not isolated but was suggested to
undergo rapid ring opening to nitrile ylide (166) which cyclized to the observed photoproduct oia an internal 1,3-dipolar
cycloaddition reaction.
(173)
‘H3
dH3
(174j
9. Nitrile Oxides
2,3-Dipoles of this class can be conveniently generated by
the base treatment of hydroxamic acid chlorides or by oxidation of aldoxirnes[’*’. Numerous cycloadditions involving
nitrile oxides had already been described when the general
concept of 1,3-dipolar cycloadditions was formulated by
Huisgen et ~ 1 . ~ Oxidation
~1.
of 2-allyloxybenzaldehyde oxime
( I 75) by nitrogen dioxide has recently been reported to give
the fused ring compound ( I 77),the product of an intramolecular cycloaddition of the intermediate nitrile oxide (176)[831.
( 1 66)
The related 4-phenyl-2-oxa-3-azabicyclo[3.2.O]hepta-3,6diene (168) system undergoes a similar photochemical rearrangement to produce 2-phenyI-1,3-oxazepine (I 70)[”1. This
reaction has been proposed to involve an azirine intermediate
(169) which subsequently cyclizes to the observed product.
Ph
When thechain between the azirine ring and olefinic portion
is reduced to two carbon atoms, no cycloadduct was obtained.
In this case, the “two-plane” orientation approach of the
linear nitrile ylide and the 7c system is too difficult to attain
and other competing processes occur[751.
134
H
(167j
0 0
- N-OH
&
C
C=N-O
H zC H =C H
O-CHzCH=CHz
Nitrile oxides generally undergo bimolecular 1,3-dipolar
cycloaddition with terminal double bonds to give 5-substituted
2 - i s o ~ a z o l i n e s [ ~8~s l~. Obtainment of the 4-substituted 2isoxazoline (177) in the intramolecular cycloaddition of ( I 7 5 ) ,
indicates that geometrical factors can force the reaction to
occur in the opposite manner from that normally encountered.
This observation led Garanti et al. to examine the intramolecular 1,3-dipolar cycloaddition reaction as a function of chain
length between the dipole and dipolarophile[86! Treatment
of aldoxime ( 1 7 8 ) with nitrogen dioxide gave a mixture of
(179) (17%) and ( 1 8 0 ~ )(2%).
With the next higher homologs, the reaction afforded compounds (180b) and (180c) together with a sizable quantity
of uncharacterized resinous material. These results clearly
849
Angriw Chrm. Int. Ed. Enyl.
Vol 15 (3976) N O . 3
0-N
indicate that the chain length between the dipole and dipolarophile groups has a substantial effect on the intramolecular
cycloaddition. Thus, while aldoxime ( 275) gave ( I 7 7 ) in 42 ‘%
yield, formation of (179) from (178) occurred to a minor
extent (17 %) and no internal adduct could be obtained from
the next higher homologs. The formation of the large ring
compounds (18Ua)--(I80c) involves an initial intermolecular cycloaddition to give long chain intermediates which are
capable of undergoing a subsequent intramolecular ring closure to the final product.
10. Conclusion
Intramolecular 1,3-dipolar cycloaddition is an extremely
versatile and important reaction. The range of synthetic possibilities which it opens for the construction of fused heterocycles
is extremely large. A number of dipoles which undergo this
reaction have been summarized and the general outlines and
potential analogies for these reactions noted. Further work
will be needed to clarify some of the mechanistic features.
The development of general synthetic routes leading to new
heterocyclic systems using this reaction offers considerable
challenge and promise.
Received: August 13, 1975 [A 98 IE]
German version: Angew. Chem. XX. 131 (1976)
~~
[I]
[4]
171
[XI
[Y]
[ IO]
[ I I]
[I21
1131
[I41
[ 151
[I61
1171
[ 181
[ 191
[20]
1211
[22]
~
R. Hiiisgm, Angew. Chem. 75. 604, 742 (1963): Angew. Chem. fnr.
Ed. Engl. 2, 565, 633 (1963).
ii. J . Org. Chem. 33, 2291 (1968).
11. R . GI-ushej, and J . Sarirr in S. Piiroi: The Chemistry of
nterscience. New York 1964, pp. X06ff.
L. E. Siiiiih, Chem. Rev. 23, 193 (1938).
6, Chem. Rev. 69, 345 (1969).
e i i , G. Szeimies, and L. Miihius, Chem. Ber. 100. 2494 ( 1967);
en, H . Srongl, H . J . Sturin, and H . Wirgmhokr, Angew. Chem.
73. 170 (1961).
R . Huisgni, L. Miihiiis, G . Mullw. H . Slaiigl, G. Szeiiiiies, and J . M .
C/eriion. Chem. Ber. 98, 3992 (1965).
R. Hiiis!qoi. H . Gorrliurdr. and R . Guo.shey. Chem. Ber. 101. 536 (1968).
M . Chrisri and R. Huisgeii, Tetrahedron Lett. 1968. 5209.
A . E d d l . R. Hiiiyjcw. R. S I I S ~ I ~ ~ UG.I ~ W~illhillich,
II,
D G r a s h q . and
E. Spimllrr. Chem. Ber. 100. 2 I92 ( I 967).
R. Hofliiioiiii and R. B. U+)oiln80ril,Acc. Chem. Res. I , 20 (1968).
R. A . Firesrone, J . Or&. Chem. 33, 2285 (1968): 37, 2181 (1972); J.
Chem. Soc. A lY70, 1570.
R . Hiiiqeii. R . S U Y ~ I I ~ Uand
I I I ~ ,K . Buiige, Chem. Ber. 105. 1324 (1972).
K . N . H a r k . J. Am. Chem. Soc. 94, 8953 (1972).
K . N . Hoiik. J . Siiiia, R. E. Duke, Jr.. R. W Siruziev, and J . K . Grurgr,
J . Am. Cheni. Soc. 95, 7287 (1973).
K . A. Houk. J . Sinis, C. R . Mitr.5, and L. J . Lushus, J. Am. Chem.
Soc. Y5. 7301 (1973).
J . Siiii.5 and K . N. H U I I ~J.. Am. Chem. Soc. YS. 5798 (1973).
R. S i i . ~ t i i i ~ ~ iTetrahedron
ti,
Lett. 1971, 2717.
D. Si. C. Black. R. F. Crozio.. and I/: C. Ducis, Synthesis 1975. 205.
N. A . LrBrl and J . J . W1iuiig. J. Am. Chem. Soc. 81, 6334 (1959).
N . A . LeBtd. Trans N. Y. Acad. Sci. 27, 858 (1965).
N A. LrBel. M . E. Post, and J . J . Whiiiig. J. Am. Chem. Soc. X6.
3759 (1964).
A i i g n a . C h m . liir.
Ed. Eiigl. J Vol. 15 f l y 7 6 j Nu. 3
1231 N . A . LrBel and 7: 4 . Lujinr\s. Tetrahedron Lett. 1966. 2173.
[24] G. R. Drlpierre and M . Liiiiiihtw J . Chem. Soc. 196.1. 4693.
[25] N . -1. L e B d G. M . J . S/iisiirc:ii/,, and L. A. Spirrlock. .I. Am. Chcm.
Sou. 84. 4360 (1962).
1261 N. A. LeRel, N . D. OjAu, J . R. Mrnke. and R . J . N n i l t i t v l . J Org.
Chem. 37, 2896 (1972).
1271 N . A . Lt,BrI and E. G. Buiiiicci. J . Org. Chem. 36, 2440 i197 I ).
[28] U:Oppdro. and K . Keller, Tetrahedron Lett. 1970. I 117.
1291 W Oppolzrr and H . P. Webrr, Tetrahedron Lett. 1970. 1121.
1301 M: Oppulrcr and K . Kelln., Tetrahedron Letr. IY70, 4313.
[3 I ] U: C. Liiiiimii, Jr.. J . Am. Chcm. Soc. YI. 2820 (1969).
[32] R. Hirisgtw, R. G r o s h n , H . Hair& and H . Seidl, Chem. Ber. 1 0 1 ,
2043 (1968).
[33] N . A. LeBel and E. Banwci. J. Am. Chem. Soc. Y2. 5278 (1970).
[34] M. L. Miliuilocic~.L. L o r m ~and
~ , Z . Miik,siinorii. Tetrahedron 29, 26%
[ 1973).
1351 J . B. Bapat, D. Sr. C . Black. R. F . C. Broivii. and C. Ichkx. Aust.
J . Chem. 25, 2445 (1972).
[36] J . J . 7iifriridlu and E. J . 7~i~hiil.ski.J . Chcm. Soc. Chem Conimuii
I Y 73, 720.
[37] J . J . T~furie110 and E. J . Tr?hulski. J . Org. Chcm. 3Y. 3378 (19741.
[38] E. Giissiiiger. R . IinhoL and H. Welrrli. Hclv. Chim. Acta 5X, 96 (1975).
[39] E. Biichiwr, Ber. Dtsch. Chem. Ges. 2I. 2637 (1x88).
[40] W Kiriii.sr and H . Dicrriclr, Chem. Ber. 100, 2710 (1967).
[41] M. S c J i ~ w r ~A,. B ~ s o l d ,and E. R. N d s u n , J . Org. Chem. 30. 242.5
(1965).
[42] E. Pirrs, R. W Bririoii, R. .I. K e z i t w , and R. D. Sidlic,. Can. J . Chem.
49, 2623 (1971).
[43] B. M . Fusr and R. M . Car!.. J. Am. Chem. Soc. Y3, 5572 (1971).
[44] B. 1Z.I. Trosr. R. M. C o r ~ .P. H . S i d d e r . and H . B. Neuholil. J . Am.
Chem. Soc. 9.5, 7813 (1973).
[45] R. Hoffinaiiii. J. Am. Chem. Soc. YO. 1475 (196X).
1461 D. W . 4 d n i i i . ~ o i iand J . K r i i i i e r . J. Chem. Soc. 1935, 286.
1471 C. D. Hurd and S. C . Lui, J. Am. Chem. Soc. 57. 2656 (1935)
[48]
1491
[50]
[51]
[52]
[53]
[54]
A . Ledwirh and D. P o r r j . J. Chem. SOC.B lY67, 41
D. Y Curtiti and S. M. Grrher, J. Am. Chem. Soc. 74, 4052 (1952).
J . L. Brewhukw and H. H u r t , J. Am. Chem. Soc. Y I . 71 I (19691.
8. M. Trosr and P. J . W/iitiiiuti, J . Am. Chem. Soc. 96. 7421 (1974).
A . Micl1iicl. J . Prakt. Chem. [2] 4s. 94 (1893).
F . C. UMr. J . Org. Chein. 32. 1596 (1967).
P. A. S. S i ~ i i r h ,J . M . Clqg. and J . H. H d i . J . Org. t h e m . 23. 524
( 1958).
[ 5 5 ] A . L. Logorhriis, J . Am. Chcm. Soc. 87, 749 (1965).
[ 5 6 ] R. Fiisco, L. Guriinri, and G. Zeccki, J. Org. Chem. 40, 1906 (1975).
[57] J . Busride. J . Humelin, F . G s i r r . and Y L’oQuuii~/. Bull. Soc. Chim.
Fr. 1973, 2555.
[58] L. K . D w l l and J . E. Krnip, J. Chcm. Soc. B 1Y6X, 976.
[59] J . H . Hall. F . E. Behr, and R. L. R i d . J. Am. Chcm. Soc. Y4. 4952
( 1972).
[60] W Opp01;er. Tettahedron Lett. 197% 2199.
[61] W Oppolzcv, Tetrahedron Lett. 1972. 1707.
1621 W O p p i k r r . Tetrahedron Lett. 1970, 3091.
[63] R. Huisgrri, M. Sriild, G. Wfillhillii.liand
, H. Kiiiipfi’r. Tetrahedron
17. 3 ( 1962).
[64] R Fiisco. L. Guraiiri, and G. Zecchi. Tetrahedron Lett. IY74. 269.
[65] R. Huisgeii, H . Kiiupfi’r, R. Siisrviann, G. Wiillhillich, and 1/ Wehn.iiilor/cr,
Chem. Ber. 100, 1580 (1967).
1661 S. Morroci.lri, A . Riicii, and A . Zoiiorurri, Tetrahedron Lett. IY70. 3215.
[67] R . Huisgerr, H . Stuirgl. H . J . Siiirni, and H. Wii[qeiihofw, Angew Chem.
74, 31 (1962): Angew. Chem. Int. Ed. Engl. I , 50 (1962): R. Hiiisgoi.
H . Sraiigl. H . J . Sliirin. R. Ruob, and K . Bunyr, Chem. Ber. 105. 1258
(1972): K . Buiigr. R. Hui.\geii, and R . Rooh. ihiil. 10.5. 1296 (1972).
1681 K . Biirqw and J . F h i . Chem. Ber. 105. 3814 (1972).
1691 A. Puiiwu, M . Dhurun. J . Simdaiiu/% and S. 1. Wern~nrtd.Pure Appl.
Chem. 33, 269 (1973).
1701 P. Clirris, 7: Doppliv. N . Gakis, M . Geuqurukis, H . Gie:riitlriiitin., P.
Gilgcw, H . Hf‘iiii(jiirIiwr. B. Jiicksuii. M . Miierk?, N . S. Nurii.si~~ihiiii.
H . J . RosoiLruiii, A . Wiinderli. H . J . Huiiscii. and H. S d i n i r d , Pure
Appl. Chem. 33. 339 (1973).
[71] A. Poihru, J . Siiiolaiiof, and A . TPrnipo., Tetrahedron Lett 1Y74. 29:
A. Piirlwn and J . Smu/anif/; ihiil. 1974. 33.
1721 A. Prulwo, J . Siirolonoff, and A . Treniper, J . Am. Chem. Soc. 97. 4682
( 1975).
[73] A . Piirlivii and P. H . J . Curl.\m. J . Am. Chcm. Soc. 97, 3862 (1975).
1741 A. Pudir.a and P. H. J . Curlsen, unpublished results
1751 .4. Purliw and N . Krrrnigara, unpublished results.
[76] H. Giicriidiinwr, H . Heiinyiirriiiv. B. Jadsoii. 7: W r i i k / ~ vH.
, J . Hriii.scii,
and H . Schniid, Helv. Chim. Acta 36. 261 1 (1973).
[77] 7: M i i k i i r and H S i i k u n a , Tetrahedron Lett. 1973. 1835.
17x1 R. Crirgee in . I . 0 Edwards: Peroxide Reaction Mechanisms. Interscience. New York 1962. p. 32; Angew. Chem. 87, 765 (1975): Angeh.
Chcm. Int. Ed. Engl. 14, 745 (1975).
1791 P. S. Boiley. Chem. Rev. S8, 925 (1958).
1801 P. D Biirrfp!r and 1 G. Trrr,.[ui-. J. Am. Chem. SOC.84. 3408 (1962).
135
R. Crieyrr, A. Bonriu, and H . Keul, Chem. Ber. 108, 1642 (1975).
[82] C. Grundmann and P . Griinanger: The Nitrile Oxides. Springer. Berlin
1971.
[83] R. Fusco, L. Goranti. and G . Zrcrhi, Chim. Ind. (Milan) 57, 16 (1975).
[Sl]
K . Bast, M . Christl, R. Huisym, W Mack, and R. Susrrnan,i, Chem.
Ber. 106. 3258 (1973).
[SS] A. Dondoni and G. Barbaro. J. Chem. SOC. Perkin Trans. 2, lY73, 1769.
[86] L. Garanri, A. Sufa, and C . Zrcchi, J. Org. Chem. 40, 2403 (1975).
[84]
The Fischer-Tropsch Synthesis: Molecular Weight Distribution of
Primary Products and Reaction Mechanism
By G . Henrici-Oh6 and S. Olivd*]
An evaluation of literature data concerning the production of hydrocarbons and alcohols
by the Fischer-Tropsch procedure indicates that the molecular weight distribution of the primary
products can be described by the Schulz-Flory equation (“normal” or “most probable” distribution). It follows that the highest attainable selectivity for a given molecular weight range
can be calculated. A reaction mechanism is suggested.
1. Introduction
The oil embargo, and the escalation in price of crude oil,
have brought about a considerable increase in the price of
oil-based raw materials such as paraffins, olefins, alcohols,
etc. Although the eventual exhaustion of petroleum resources
has been predicted for a considerable time, the present shortage
actually occurred several years earlier than expected, and
the chemical industry, with its current reliance on oil-based
raw materials, is particularly affected.
As oil becomes more and more expensive, chemists revert
to coal as a source of carbon which had almost been forgotten
in the course of the years. Some fifty years ago, however,
German chemists were already confronted with the same problem. Germany, possessing huge stocks of coal, but completely
inadequate oil resources, was forced to undertake the transformation of coal into liquid hydrocarbons. As early as 1924,
Franz Fischer published a book on this subject[’]. In the
succeeding decades two processes were simultaneously developed to great industrial importance : the catalytic hydrogenation ofcoal at high pressure (Bergius process), and the catalytic
synthesis of hydrocarbons from carbon monoxide and hydrogen, obtained from coal and water, at normal and medium
pressure (Fischer-Tropsch process). During World War I1
this development reached its zenith.
After 1945 the situation changed. European pre-embargo
prices of oil and coal were comparable (about $30-40 per
ton), but coal has an unfavorable TOE (ton of oil equivalent):
1 ‘/2 tons of coal supply the same energy as 1 ton of oil.
Moreover, the ease of automating transport systems and feedlines for refineries, production plants, industrial and domestic
heating, etc. has helped oil to oust the dirty and inconvenient
material coal. Even at present-day prices of about $100 per
ton of petroleum uerms $70-80 per ton of coal, the production of gasoline from coal does not appear immediately attractive. However, one of the sound effects of the embargo shock
has been to create a feeling of responsibility for the future.
Apart from the search for other sources of energy such as
[*] Dr. G. Henrici-Olive and Prof. Dr. S. OlivC
Monsanto Research S. A.
Eggbuhlstrasse 36, CH-8050 Zurich (Switzerland)
136
nuclear, geothermal, water, and wind power, scientific and
industrial investigations turn back to coal.
The present study describes some possibilities and limitations of the synthesis of hydrocarbons and alcohols from
CO and Hz(Fischer-Tropsch synthesis). Moreover, an estimate
is made of how much product of a given degree of polymerization may be expected in the most favorable case. The basis
of this estimate is the observation that products obtained
by this process exhibit a characteristic distribution of chain
length, as we found during the evaluation of literature data.
We have applied concepts originating in the field of molecular
catalysis with defined transition metal complexes and from
polymer chemistry to mechanistic details reported by various
authors, and suggest a reaction mechanism for the FischerTropsch synthesis.
2. Development and Present Status of the FischerTropsch Synthesis
At the beginning of the synthesis of hydrocarbons from
C O and H2 stands the classical methane synthesis of Sabatier
and Senderens in 1902r2],but it was not until 1922 that Fischer
and Tropsch obtained their first patent on “Synthol”, a mixture
of oxygen-containing derivatives of hydrocarbons, produced
from C O and H2 at over 100atm and at 40O0C, with alkalitreated iron shavings as catalyst[3! There are several competent
review articles describing the further development of the
Fischer-Tropsch synthesis[41,which we do not intend to duplicate. We shall restrict ourselves to some highlights, with
emphasis on the present status.
In 1925 Fischer and Tropsch succeeded in directing the
synthesis to give predominantly hydrocarbons; they patented
a normal pressure process, operating in the temperature range
250-300”C, and using Fe/ZnO or Co/Cr203 as catalyst[’].
In 1936/37 Fischer and Pichler found that working at medium
pressure (5-30 atm), with cobalt[61or iron[’] catalysts, substantially improved yield and catalyst lifetime. Using a ruthenium catalyst, and working at high pressure (1000-2000atm)
and temperature (140-20O0C), Pichler et al. could orient
the synthesis to high melting linear paraffins having molecular
*I; the “polymethylenes” obtained are
weights up to i06[4d,
practically identical with Ziegler polyethylene.
Angen. C h r m Int. Ed. Engl. f Vol. 15 ( 1 9 7 6 ) No 3
Документ
Категория
Без категории
Просмотров
1
Размер файла
1 243 Кб
Теги
intramolecular, reaction, cycloadditions, dipolar
1/--страниц
Пожаловаться на содержимое документа