close

Вход

Забыли?

вход по аккаунту

?

Nitrenes and the Decomposition of Carbonylazides.

код для вставкиСкачать
ANGEWANDTE CHEMIE
VOLUME6
NUMBER11
NOVEMBER 1967
PAGES 897-101 2
Nitrenes and the Decomposition of Carbonylazides
BY W. L W O W S K I ~ l
Gratefully dedicated to Professor Karl Freudenberg
The decomposition of organic carbonylazides can lead to the forniation of nitrenes.
Ethoxycarbonylnitrene is formed in the photolytic and thermal decomposition of ethyl
azidoformate and by ci-elimination from N-(p-iiitrobenzenesu~onyloxy)urethari.Both
of’ the possible electronic states of this nitrenc take part in intermolecular reactions. Pure
singlet nitrene is formed by u-elimination from the urethan and on thermal decomposition
of ethyl azidoformate, but changes so rapidly into the triplet form that the reactions
of both forms are observed. Singlet ethoxycarbonylnitrene undergoes selective and
stereospecific insertion into C- H bonds and adds stereospecifically to olefins. Triplet
ethoxycarbonyltiitrene, however, does not undergo insertion into C -H bonds, and adds
to olefins with complete loss of the geometric corijiguration. By following quantitatively
the stereospecificity of’the addition reuction and by selective interception of the triplet
and singlet forms of the riitrene, it can be shown that the photolysis of ethyl azidoformate
leads directly to nitrene of which one third is in the triplet state. In the decomposition
of aryl- and alkylcarbonylazides (acid azides), the removal of nitrogen is accompanied
by a synchronous rearrangement to isocyanates (Curtius rearrangement). In this system,
nitrenes are obtained o d y by photolysis. They add to doirbie bonds and undergo very
selective insertion into C-H bonds, but do not rearrange at a measurable rate to isocyanates. The photolytic Curtius rearrangement is also a concerted reaction.
1. Introduction
of electrons, and the other two contain one electron
each.
Interest in the study of electron-deficient, uncharged
intermediates has increased greatly in the past 15
years. A great deal of work is being carried out in
particular on the carbenes, but their nitrogen analogues
have also attracted increasing attention. These compounds contain a nitrogen atom having only six
electrons in its valence shell. Two of these electrons
form the bond with the single ligand, while the other
four can be arranged in two ways, i.c. as two pairs of
electrons on a nitrogen atom with a low-energy empty
orbital (singlet nitrene) or as one electron pair plus
two electrons with parallel spins (triplet nitrene). In
the triplet state, one of the three low-energy orbitals
that are not used for bonding is occupied by a pair
The terms “singlet” and “triplet” are borrowed from
spectroscopy, but the two electronic states must not
be confused with the spectroscopic states. The very
short-lived spectroscopic states have the geometric
configuration of the ground state (Franck-Condon
principle). The long-lived states of c.g. the nitrenes or
carbenes, on the other hand, assume the energetically
most favorable conformation, which can differ very
appreciably from that of the ground state. States
separated from the ground state by forbidden transitions can be sufficiently long-lived to enter into
intermolecular reactions before crossing to the electronic ground state.
[*I Prof.
W. Lwowski
New Mexico State University, Research Center
Las Cruces, N . M . 88001 (U.S.A.)
Angew. Chern. infernat. Edit. 1 Vol. 6 (1967) f No. I I
In the study of the mechanisms of potential nitrene
reactions, therefore, one is concerned, not only with
the questions of whether a nitrene is in fact present
and how it behaves cheniically, but also with the
897
question of the electronic state of the reacting nitrene.
It is possible (and does in fact happen; see below) that
two different electronic states of a nitrene enter into
intermolecular reactions, in which they behave quite
differently.
Unfortunately, there is as yet no agreement o n nomenclature.
Molecules R-N are described as azenes, azylenes, imenes,
imines, imidogen derivatives, and nitrenes [1,21. These names
are all open to objections[l.21. The name “nitrene” is used
in the present article because of the analogy with “carbene”.
It is not intended to present a survey of the entire
nitrene field, which has been the subject of recent
reviews 11-31. Instead, the present article will deal with
some aspects of the chemistry of the carbonylnitrenes,
R-CO-N, connected with our investigations on the
decomposition of carbonylazides.
The carbonylazides (acid azides) were discovered by
Curtius [4,51, who also observed their rearrangement [61
to isocyanates on thermal decomposition. As early as
1896, Stieglitz 171 explained the Curtius rearrangement
(path A) as involving a nitrene intermediate, similar
to Tiemann’s formulation 181 of the Lossen rearrangement (path B).
R-CO-NH-OAcyI
Curtius later coined the term “rigid azides” 191 for azides that
d o not rearrange, but form“short-lived intermediates R-N<”,
which react intermolecularly. Thus Curtius wrote that sulfonylazides form the residues Ar-SOZ-N<,
which exhibit
“unmistakable analogies” t o the residues R-CH<, which
are nowadays known as carbenes [lo]. Though Stieglitz’s
views on the mechanism of the Curtius rearrangement appear
to be incorrect (Section 6), it is obvious that Tiemann, Srieglitz, and Curtius had a clear conception of carbenes and
nitrenes as early as seventy years ago. They simply lacked
the means of proving conclusively the existence of these
species. Moreover, at that time there was no theoretical
foundation for an understanding of the reactivity of the
nitrenes.
The first nitrene whose existence was verified by
physical methods was NH. This nitrene can be prepared by the decomposition of hydrazoic acid, and can
be detected spectroscopically in the gas phase. Some
of its reactions have been elucidated (for reviews
see [2,111).
[l] See L. Horner and A . Christmann, Angew. Chem. 75, 707
(1963); Angew. Chem. internat. Edit. 2, 599 (1963).
[2] See R. A . Abramovitch and B. A . Davis, Chem. Rev. 64, 149
(1964).
[3] W. Kirmse, Angew. Chem. 71, 537 (1959).
[4] T. Curtius, Ber. dtsch. chem. Ges. 23, 3023 (1890).
[5] Cf. A . Dorapsky, J. prakt. Chern. 125, 1 (1930).
[6] T. Curtius, Ber. dtsch. chern. Ges. 27, 778 (1894).
[7] J. Stieglitz, Amer. chern. J . 18, 751 (1896).
[8] F.Tiemann, Ber. dtsch. chem. Ges. 24, 4162 (1891).
[9] T. Curtius and F. Schmidt, Ber. dtsch. chern. Ges. 55, 1571
(1922).
[lo] T. Curtius, Z . angew. Chem. 26, 111, 134 (1913).
1111 P. A . S . Smith: Open Chain Nitrogen Compounds. Benjamin, New York 1965, Vol. 1, p. 4.
898
2. Reactions of Nitrenes and Azides
The assumption that the first step in the decomposition
of an azide is the removal of nitrogen appears satisfactory, since the great heat of formation of the
nitrogen molecule and its suitability as a leaving group
(e.g. in the decomposition of the diazonium salts) are
well known.
+ R-N
However, this does not mean that the activation
energy for the process (1) must necessarily be low and
that the nitrene hypothesis may be accepted without
question for all azide decompositions. In fact, for all
the azide decompositions that can be explained on the
basis of nitrenes, it is also possible to offer other
reasonable mechanisms. This can be illustrated by
nitrene and azide mechanisms for the formation of
rearrangement, addition, and insertion products (see
eq. (2)-(7) on facing page).
In fact, the heat-induced Curtius rearrangement appears to proceed via an azide mechanism (eq. ( 3 ) ) and
not via a nitrene (eq. (2)) (see Section 6). With reference
to the formation of aziridines from olefins and decomposing azides (eq. (4),(5)), it has been known for
a long time that triazolines can be prepared by addition of azidey, and that they can decompose to form
nitrogen and aziridines [121.
The nitrene mechanism (6) seems particularly attractive for insertion reactions; however, evidence favoring the azide mechanism (7) can also b e produced.
Thus one would expect that a radical pair such as that
in (7) can dissociate, the nitrogen radical then yielding
nitrogen and R-CO-NH2 by abstraction of another
H atom. Amide (or amine in the case of alkylazides
R‘-N3) is in fact always found as a by-product.
Similar arguments can also be advanced in every other
case for the distinction between azide and nitrene
mechanisms, i.e. the reaction mechanism cannot be
deduced entirely from the nature of the products.
Kinetic measurements provide further evidence concening the thermal decomposition of azides. It may
be assumed that the reactions of a nitrene will be faster
than its formation (see eq. (8)).
The solvent used as the substrate participates only
indirectly (by solvation) in the formation of the
nitrene, and should have only a relatively small in____
[12] L. W o f , Liebigs Ann. Chem. 394,30,68 (1912); K . Alder and
C . Stein, ibid. 485, 211 (1931); R . Huisgen, Angew. Chem. 72,
370 (1960); G . Szeimies and R . Huisgen, Chem. Bet. 99, 491
(1966); Shinogi and Co., Netherlands Pat. Appl. 6 514295
May 6,1966; Chem. Abstr. 65,16941a (1966). For a review, see
R . Huisgen, G . Grashey, and J. Sauer in S. Patai: The Chemistry
of Alkenes. Interscience, New York 1964, p. 835ff.
Angew. Chem. internat. Edit.
Val. 6 (1967) 1 No. I 1
be applied to photolytic decompositions, in which the rate of'
decomposition of the azide is determined by the light flux
and the quantum yield.
The existence of a number of nitrenes has been demonstrated by physical methods. The triplet state of
the simplest member of the series, NH, can be identified by its optical absorption spectrum [21, and the
+
R-CO-N
R-CO-N
+
R-CO-N~
H-C?
1
-.
R-CO-N
2
R-CO-P~, j
-Nz
slow
R-co-N-N=N:C) R-co-N-N=N-C) etc
+ .cf
I
R-CO-NH-Cc
R-N
(4)
R-CO-NH-CC
.+
R-co-N-N~
H
R-N3
Q
>%
excitation
+
substrate
- - fast
--+
+
Nz
products
fluence on the rate or temperature of the decomposition of the azide. However, the solvent (substrate) is
involved in an azide reaction, and its chemical nature
must have a decisive influence on the rate of the
reaction.
Waters and Walker[131 observed two types of kinetic
behavior in the thermal decomposition of some aryland alkylazides in various solvents. In one group of
solvents the rate of decomposition (at 132 "C) is
independent of the nature of the solvent (mineral oil,
tetrachloroethylene, etc.); the authors conclude that
the reaction proceeds by a nitrene mechanism in these
solvents. In other solvents (benzene, indene, styrene)
the rate of decomposition (again at 132 "C) is higher,
and is determined by the solvent, indicating that a n
azide mechanism is involved. For example, the ratio
of the rates of decomposition of phenylazide in
mineral oil and in indene is 1:225. Indene evidently
reacts directly with the azide. D. S. Breslow[141
studied the thermolysis of a number of azidoformates
in aliphatic and aromatic solvents as well as in
ketones and alcohols. The very slight differences in the
rate of decomposition indicate a nitrene mechanism.
Strictly speaking, the procedure described shows only that
the solvent is not involved in the rate-determining step. It
tells us nothing about the nature of the intermediate, which
could be a valence tautomer of the azide or of the nitrene
rather than a nitrene. Unfortunately, the procedure cannot
I131 P . Walker and W . A. Waters, J. chem. SOC.(London) 1962,
1632.
1141 T . J. Prosser, A . F. Marcantonio, C. A . Genge, and D. S.
Breslow, Tetrahedron Letters 1964, 2483.
Angew. Chem. internat. Edit. VoI. 6 (1967) / No. 11
(7)
disappearance of N H in mixtures with ethylene has
been followed spectroscopically [151. Most of the
available data relate to triplet nitrenes at very low
temperatures, particularly ESR measurements in
glasses between 4 and 80 "K 1161.
Though measurements of this type show the existence of
nitrenes under the conditions in question, they do not answer
questions about the chemical properties of nitrenes, and tell
us nothing about the mechanisms by which azides decompose
at room temperature and in liquid solutions. Moreover, the
measurements provide information only about the electronic
ground state of the nitrenes observed. It is quite possible,
however, that a nitrene is formed in a higher electronic state,
e.g. the lowest singlet, and enters into chemical reactions in
this state before it has a chance to cross to the electronic
ground state. I t will be shown in Section 4.1 that this is in
fact what happens to ethoxycarbonylnitrene.
3. Ethoxycarbonylnitrene
The photolytic [17,181 o r thermal [18,191 decomposition
of ethyl azidoformate [201 in various solvents yields
products whose formation can be explained by the
1151 D . W . Cornell, R . S. Berry, and W . Lwowski, J. Amer. chem.
SOC.88, 544 (1966).
[16] GSmolinsky, E. Wasserman, and W.A. Yager, J. Amer.chem.
SOC.84,3220 (1962); G.Smolinsky, L.C.Snyder, and E. Wasserman
Rev. mod. Physics 35,576 (1963); E. Wasserman, L.C.Snyder, and
W. A.Yager, J. chem. Physics 41, 1763 (1964); E. Wasserman,
GSmolinsky, and W.A. Yager, J.Amer.chem.Soc. 86, 3166 (1964);
R . M . Moriarty, M . Rahman, and G.J.King, ibid. 88, 842 (1966);
For UV-spectra, see: A . Reiser, H . Wagner, and G. Bowes, Tetrahedron Letters 1966, 2635; A . Reiser, G. Bowes, and R . J. Horne,
Trans. Faraday SOC. 62, 3162 (1966).
[17] W . Lwowski and T. W. Mattingly, jr.. Tetrahedron Letters
1962,277.
1181 W. Lwowski and T. W . Mattingly, jr., J. Amer. chem. SOC.
87, 1947 (1965).
[19] R.J.Cotter and W.F.Beach, J. org. Chemistry 29, 751 (1964).
[20] M . 0. Forster and H. E. Fierz, J. chem. SOC.(London) 93,
81 (1908).
899
assumption of a nitrene intermediate, RO-CO-N, or
by azide mechanisms. A number of reactions of this
type are shown in Scheme 1.
\I
:C-NH-COOCzH5
?>N-COOCzH,
C
RO-NH-COOCzH5
The following quantitative comparison of the C-H insertion productsconfirms our belief in a common intermediate in the decomposition of (2) and of ethyl azidoformate.
Insertion into the C-H bonds of 2-methylbutane gives
four isomeric products (3)-(6). Division of the ratios
of the quantities (determined by gas chromatography)
by the number of hydrogen atoms of each type gives
the relative reactivities, which are shown in TabIe 1.
The two methyl groups on C-2 in 2-methylbutane have
the same reactivity. Despite somewhat different reaction conditions, the figures differ only slightly. We
regard this as confirhation of a common reactive
intermediate, which must be ethoxycarbonylnitrene.
Scheme 1. Decomposition of ethyl azidoformate in hydrocarbons I17,18,21-23,271, olefins [17,181, alcohols 124,251,
benzene [21,26,281, amines 126,291, and nitriles [30,311.
To settle the question of the existence of ethoxycarbonylnitrene as an intermediate in the reactions in
Scheme 1, we prepared the nitrene by an independent
route [21 321. N-(p-Nitrobenzenesulfony1oxy)urethan
( I ) can be converted even by weak bases into its anion
(2), which splits off a p-nitrobenzenesulfonate ion in a
first-order reaction. The half-life of the reaction (in
dichloromethane at room temperature) is about 7
seconds 1331. The other fragment formed is ethoxycarbonylnitrene. In the presence of the solvents indicated in Scheme 1, the same products are formed as
from ethyl azidoformate.
Table 1 also contains data on the thermolysis of ethyl azidoformate in 2-methylbutane, which were determined independently by D . S . Breslow [231, and which agree well with
our results. The differences in the numerical values for nitrenes
prepared by different methods are probably due to the
fact that the very different leaving groups (N2 and
@03S-C6H4-N02) in the photolysis and in the a-elimination
are still located in the neighborhood of the nitrene. The
difference in the values for the photolysis and the thermolysis
of the azidoformate is probably due mainly t o the difference
in the reaction temperatures (30 "C and 100 "C respectively).
0
C Z H ~ O O C - N - O - S O ~ - C ~ H ~ - N O z (9)
(2)
C~H,OOC- N
I
+ 04s - c,H,- NO^
I211 W. Lwowski, T. J. Maricich, and T. W. Mattinglp, jr., J.
Amer. chem. SOC.85, 1200 (1963).
[22] W. Lwowski and T. J. Maricich, J. Amer. chern. SOC.86,3164
(1964).
[23] M. F. Sloan, T. J. Prosser, N. R. Newburg, and D. S. Breslow, Tetrahedron Letters 1964, 2945.
I241 R. Puttner and K. Hafner, Tetrahedron Letters 1964, 3119.
[251 W. Lwowski, R. DeMauriac, T . W. Mattingly, j r . , and E.
Scheiffele, Tetrahedron Letters 1964, 3285.
(261 K . Hafner and C. Konig, Angew. Chem. 75, 89 (1963); Angew. Chem. internat. Edit. 2,96 (1963); K. Hafner, D . Zinser, and
K . L . Morifz, Tetrahedron Letters 1964, 1733.
[27] H. Nozaki, S. Fujira, H . Takaya, and R. Noyori, Tetrahedron
23, 45 (1967).
[28] W. Lwowski and R. Johnson, Tetrahedron Letters 1967, 891.
[29] W. Lwowski and L . SeIman, Abstracts 150th Meeting Amer.
Chem. SOC.,Sept. 1965, Abstract S 25.
1301 W. Lwowski, A. Hartenstein, C. DeVita, and R . L . Smick,
Tetrahedron Letters 1964, 2497.
[31] R. Huisgen and H. Blaschke, Liebigs Ann. Chem. 586, 145
(1965).
[32] W. Lwowski and T. J. Maricich, J. Amer. chem. SOC.87,3630
(1965).
[33] W. Lwowski and T. J. Maricich, unpublished.
900
Table 1. Relative reactivities of ethoxycarbonylnitrene with tertiary (30).
secondary (20). and primary (lo) C-H bonds.
I
Nitrene source
Hydrocarbon
and solvent
Azide
photolysis
30:20: 10
100% 2-Methyl-
37
Azide thermolysis
3 0 : 20 :
10
a-Elimination
30: 20: l o
-~
9
1
butane
36 10 1
38 mole-% 2Methylhutane in
dichloromethane
lOO%3-Methyl- 16 5.3 1
hexane [341
27 mole-% 314 5.2 1
Methylhexane in
dichloromethane 1341
30 I231
10 1231 1
27
11
1
25
8
1
-
-
17
6
1
-
18
6.6
I
13.5
4.7
1
Similarly, good agreement was found in the mixtures
obtained with cyclohexene [18,21,321 and with nitriles 1301.
Such a marked agreement cannot be coincidence, and
it must therefore be accepted as certain that the decomposition of ethyl azidoformate and a-elimination
reactions of ( I ) proceed via a common intermediate.
~
1341 W. Lwowski and J . M . Simson, unpublished; J. M . Simson, Dissertation, Yale University 1967.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) /No.I 1
4. Singlet a n d Triplet Ethoxycarbonylnitrene
4.1. Addition to C=C Double Bonds
Following the detection of ethoxycarbonylnitrene as a
reactant, there arose the question of its electronic
state at the moment of reaction. Wasserman[351
recently showed by spin-resonance measurements at
the temperature of liquid helium that the ground state
of ethoxycarbonylnitrene is thk triplet. However, it
may be assumed that the nitrene is formed in the
singlet state, particularly in a-elimination from ( I )
and in the thermolysis of azides. Spontaneous spin
reversal is practically unknown in reactions of this
type, though it is not impossible on the grounds of
any natural law. Nothing is known about the lifetime
of the nitrene in the singlet state, except that it is quite
capable of entering into intermolecular reactions in
solution (see below).
Attempts to determine the spin state of ethoxycarbonylnitrene spectroscopically and at the same time to
follow its reactions in the gaseous state by flash photolysis have so far been unsuccessful [361. In the absence
of a reaction partner, the nitrene (presumably as the
triplet) decomposed, with a life time of between
3x10-7 and 10-5 sec, to alkoxy and isocyanate
radicals.
The problem was finally solved by a method developed
by Skell for carbenes [371, which proved very succesfulC381 in spite of many critical arguments, though
theoretical details still remain to be clarified [391.
SkeN has postulated that singlet carbenes add to C=C double
bonds in a single step, and hence stereospecifically. Triplet
carbenes, on the other hand, should first react with the olefin
to form a n open-chain diradical, the ring closure of which
proceeds more slowly than rotation about the former C=C
double bond. This free rotation leads to the establishment
of an equilibrium between conformers, and finally to the
formation of a mixture of geometrically isomeric cyclopropanes. In the extreme case, the composition of this
mixture depends only on the position of the equilibrium
between the conformers, and not OR the nature (cis or trans)
of the starting olefin. Scheme 2 shows the application of
Skell's method t o nitrenes.
We had observed that the addition of ethoxycarbonylnitrene to cis- and rruns-4-methyl-2-pentene is only
partly stereospecific. Both ethyl cis- and trans-2-isopropyl-3-rnethylaziridine-1-carboxylateare formed in
both cases. The main product is the aziridine whose
configuration corresponds to that of the olefin used.
If the nitrene is prepared by a-elimination in dichloromethane solution containing 33 mole- % of cis-olefin
(7), the total aziridine yield [based on ( I ) ] is 57 %, of
[35] E. Wasserman (Bell Laboratories), personal communication.
1361 D . W . Cornell, R. S. Berry, and W. Lwowski, J. Amer. chem.
SOC.87, 3626 (1965).
[37] P. S. SkeN and R. C. Woodworth, J. Amer. chem. SOC. 78,
4496 (1956); R . C. Woodworth and P.S.SkeIl, ibid. 81, 3383 (1959).
[38] See P. P. Gaspar and G . S. Hammond in W . Kirmse: Carbene Chemistry. Academic Press, New York 1964, p. 259ff.
[39] R . Hoffniann, Trans. New York Acad. Sci., Ser. 11, 28, 415
(1966).
Angew. Chem. internat. Edit. 1 VoI. 6 (1967) / No. I !
(9.L trans
(S)?cis
which 92 % is (8) and 8 % is (9). Hujner[401 found a
similar situation in the photolysis of methyl azidoformate in cis- and trans-2-butene. If it is assumed
that all the nitrene is formed in the singlet state and
that Skrll's method is applicable, the formation of
singlet n i t r e n e
cis -
cis - a z i r i d i n e
olefin
t r i p l e t nitrene
cis-
cisoid
cis -aziridine
t r an soid
trans -aziridine
olefin
triplet nitrene
trans
-
olefin
k,j
k,
< k,
zz k,
Scheme 2. Application of Skrll's method to the addition of singlet and
triplet nitrenes to cis- and trans-olefins.
products resulting from nonstereospecific reactions can
be explained by the additional assumption that part of
the singlet nitrene changes into triplet nitrene before
adding to the olefin. This (hypothetical) transition
should be a first-order reaction and the addition to the
olefin a second-order reaction ; this hypothesis can be
checked by the use of different olefin concentrations.
The lower the olefin concentration, the greater should
be the quantity of nonstereospecifically formed aziridine. This is the case, as can be seen from Table 2 1411.
The values can be quantitatively explained by a simple
kinetic model, which is shown in Scheme 3. According
to this model the precursor P [ethyl azidoformate or
( I ) ] forms only singlet nitrene, which then enters into
competing reactions leading either to triplet nitrene
[40] K . Hafner, W. Kaiser, and R . Puttner, Tetrahedron Letters
1964, 3953.
[41] W. Lwowski and J . S . McConaghy, J. Arner. chem. SOC. 87,
5490 (1965).
90 1
1
1
Table 2. Stereospecificity of the addition of ethoxycarbonylnitrene to
cis- and trans-4-methyl-2-pentene in dichloromethane.
Olefin
concentration
(mole- %)
i
Total
of
~
%)
( % in mixture)
57
50
33
10
5
3.3
2.5
1.5
trans-Aziridine
from cis-olefin
7.8
17.5
26
34
37.5
43
38
24
cis-Aziridine from
trans-olefin
( % in mixture)
Figure 1 shows the results of the experiments with the
nitrenes produced by or-elimination. As required by Scheme 3,
a straight line is obtained if k5.!k4 is taken to be 0.015. The
slope of the line is 0.036 = 0.004. This indicates that the
addition of the singlet nitrene to 4-methyl-2-pentene is about
30 times as fast as the singlet-triplet transition, i.e. k3/k2 =
30. (Note however that k3 is a second-order rate constant,
whereas kz is a first-order constant.)
2.6
-
8.0
-
12.3
67
100
33
10
5
3.3
1.5
78
60
1.O
t
100
33
70
10
58
43
5
2.5
1.5
0
1
55
1
50
54.5
4
k2
L
L
,
nitrene
% byproducts
non-stereospecifically
formed aziridine
Scheme 3. Possible reactions between nitrene and olefin
The stereospecific path (k3) leads to the formation of
(8) from cis-(7) and ( 9 ) from trans-(7). In the reaction
of the triplet nitrene (k4), the same mixture of (8) and
(9) is formed both from cis- and from trans-(7). The
composition of the mixture of aziridines depends only
on the position of the equilibrium between the conformers of the open-chain diradical intermediate (cf.
Scheme 2). In our case the mixture consists of 80 % of
trans-aziridine (9) and 20 % of cis-aziridine (8) [421. If
it is assumed that k~ is much smaller than k3, and that
k 2 and k3 are of a similar order of magnitude, a steady
state assumption for the concentration of the triplet
nitrene leads to the following relationship (eq. (11))1421.
T / S = kz (kj [olefin]
+ k5/k4)-1
20
10
30
LO
d
[Olefin]+ k51k,
14.7
3
stereospecifically
formed aziridine
/
1
(unimolecular) or to stereospecific combination with
the olefin and formation of aziridine (bimolecular).
The triplet nitrene adds on nonstereospecifically to the
olefin or is consumed in side reactions. Since all the
triplet nitrene is formed from the singlet nitrene, side
reactions of the singlet nitrene do not affect the
distribution between the paths k 3 and k 4 .
singlet
nitrene
0
/b
(11)
where T is the yield of aziridine mixture (80 % trans + 20 yo cis)
formed by nonstereospecific reaction; S is the yield of stereospecifically formed aziridine; for k2, kj, k4, and k5 see Scheme 3.
____
[42] J. S. MeConaghy and W. Lwowski, J. Amer. chem. SOC.89,
2357 (1967).
Fig. 1. Experimental results of the addition of the ethoxycarbonylnitrene obtained by a-elimination to the 4-methyl-2-pentenes.
For explanations see eq. (1 1).
The thermolysis of ethyl azidoformate in dichloromethane
solutions of cis- and trans-4-methyl-2-pentene gives similar
results. Corresponding to the higher temperature, k2/k3 is
greater, i.e. 0.08 instead of 0.036. The straight line for the
data obtained in the thermolysis of azide, like the line in
Figure 1, passes through the origin. This means that both
the a-elimination and the thermal decomposition of ethyl
azidoformate 1431 yield only singlet nitrene, and that the
triplet nitrene is formed from singlet nitrene.
The photolysis of ethyl azidoformate with light of wavelength
2537 A[431, on the other hand, leads directly to a mixture
of about 1,/3 triplet and 213 singlet nitrene. A plot similar to
that in Figure 1 gives a straight line that does not pass through
the origin, but intersects the ordinate at TjS = 0.44. This
corresponds to a triplet component of 0.44/1.44 = 30.5 % in
the nitrene formed. At 12 "C, the dope of the line is 0.035.
The mathematical evaluation of the photolytic reactions is
basically the same as for the thermal reaction and for the
or-elimination, but owing to the complete absorption of light
it is necessary to use expressions other than kl, and the fact
that triplet nitrene is formed directly requires that a steady
state condition also be applied to the singlet nitrene concentration.
The quantitative agreement of the experimental data
with the requirements of Scheme 3 does not in itself
conclusively prove the correctness of our assumptions
concerning the electronic states of the nitrenes and the
mechanisms of their reactions. Other reaction mechanisms could fit in with the same kinetic laws, and intermediates other than singlet and triplet nitrene could
add stereospecifically or non-stereospecifically to
olefins. However, our ideas are supported by independent observations:
Both the triplet nitrene and the singlet nitrene can be
intercepted by reactions in which only one of the two
types of nitrene takes part. Skell assumes a diradical
~~~
..
1431 J. S.McConaghy and W. Lwowski, J. Amer. chem. SOC.89,
4450 (1967).
Aiigew. Chem. infernal. Edit.1 Val. 6 (1967) / No. I 1
intermediate for the triplet reaction in his scheme for
carbenes, which has been applied here to nitrenes
(Scheme 2). The rate of formation of this intermediate
should be greater for more strongly stabilized radicals,
since this stabilization should already be reflected in
the transition state leading to the diradical.
The triplet nitrene can be intercepted with a-methylstyrene. The reaction of a solution containing 3.3
mole- % of cis-4-methyl-2-pentene in dichloromethane
[production of nitrene by a-elimination from ( I ) ]
gives absolute yields of 25 % of ethyl cis-2-isopropyl-3methylaziridine-1-carboxylateand 1 3 % of the trans
isomer. When 3.3 mole- "/o of a-methylstyrene is added
to the reaction mixture, only a trace of the transsubstituted aziridine is formed, while the yield of cisaziridine falls to 16 %. The reaction is thus almost
completely stereospecific. When 1.5 mole-% each of
cis-4-methyl-2-pentene and a-methylstyrene are used,
the.aziridine mixture obtained contains only 2.1
of
the trans compound, as opposed to 43 % of the trans
compound in the absence of the triplet interceptor. It
can be calculated from these results that triplet
ethoxycarbonylnitrene reacts 86 times as rapidly with
a-methylstyrene as with cis-4-methyl-2-pentene. The
nature of the reaction product supports the assumption
of a diradical intermediate 1421 ; thus a-methylstyrene
gives, not an aziridine, but ethyl N-(2-phenyl-2propeny1)carbamate (10).
+
olefin
95 "/, cyclohexane), the total yield of aziridines is
15
of which 30 % are trans compound and 70 % cis. With
dichloromethane as the solvent, a 43 % yield of an aziridine
mixture containing 26 % of trans compound would be
expected. Cyclohexane evidently intercepts only the singlet
nitrene; the triplet nitrene produced by photolysis remains
and leads to the formation of a large quantity of transaziridine and to higher yields. It is possible to calculate
from these data the proportion of triplet nitrene formed
directly during photolysis. The calculation gives a result of
31 %, which agrees almost exactly with the value (30.5 %)
obtained from the experiments in olefin-dichloromethane
solutions.
x,
The mechanism of the direct photochemical formation
of triplet nitrene has not yet been elucidated. However,
experiments on the acetophenone sensitized decomposition of ethyl azidoformate indicate that the triplet
azide is not the precursor of the triplet nitrene [Is].
4.2. Insertion into C-H
Bonds
The insertion of nitrenes into C-H bonds may in
principle proceed in one or two steps. In one-step
processes, the atoms involved may assume electric
charges in the transition state (eq. (13), (14)), or they
may remain entirely or substantially uncharged (eq.
(12), eq. (14a), cf. eq. (6)).
i
'cH~- NH -COOC~H,
r 10)
Singlet ethoxycarbonylnitrene can also be selectively
intercepted. It undergoes very rapid insertion into
C-H bonds, a reaction that has not so far been
observed with triplet ethoxycarbonylnitrene[441. If
ethoxycarbonylnitrene produced by photolysis is
allowed to react with cyclohexene (0.2 and 100
mole- % in dichloromethane), the ratio of the yield of
aziridine to the yield of insertion products (insertion
into the 4-position and the more active 3-position of
cyclohexene) is 58 :1 and 5 :1 respectively 1451. The
failure of triplet ethoxycarbonylnitrene to undergo
insertion into C-H bonds is also confirmed by the
experiments with 3-methylhexane (see Section 4.2).
When ethyl azidoformate is photolysed in cyclohexane, a
large part of the singlet nitrene is intercepted as cyclohexylurethan. The directly formed triplet nitrene remains, and
can add nonstereospecifically to added cis-4-methyl-2-pentene
At a n olefin concentration of 5 %, this gives a 33 % yield of
a n aziridine mixture consisting of 54.5 % of trans-substituted
and 44.5 % of cis-substituted aziridine. On the other hand,
if the nitrene is formed by a-elimination (again in 5 % cis. .-
-
[44] However, triplet cyanonitrene reacts with C-H bonds: A . C.
Anastassiou, J. Amer. chem. SOC.88, 2322 (1966).
[45] W. Lwowski and F. P . Woerner, J. Amer. chem. SOC. 87,5491
(1965).
Angew. Chem. internat. Edit.
Vol. 6 (1967)
f No. I 1
In a two-step reaction, insertion on asymmetric carbon
atoms should lead to at least partial loss of optical
activity.
Srnolinsky 1461 found that in intramolecular insertions
with formation of five-membered rings (2-ethyl-2methylindoline and 4-ethyl-4-methyloxazolidin-2-one),
the optical activity is at least partly retained. It was
later shown 1471 that 4-ethyl-4-methyloxazolidin-2-one
is formed with almost complete retention of configuration. We [481 have studied an intermolecular
insertion reaction because various substrate concentrations can be used in such a system, and it is
therefore possible to search for insertion reactions of
triplet nitrene. As was mentioned in Section 3,
ethoxycarbonylnitrene and 3-methylhexane react to
form all the possible insertion products. Starting with
optically active 3-methylhexane, we separated the ethyl
N-(1-ethyl-1-methylbuty1)carbamate ( I I ) by gas chromatography, and found that this product is formed
.
.
[46] G. Smolinsky and B. I. Feuer, J. Amer. chem. SOC.86, 3085
(1964).
[47] S. I . Yamada, S. Terashima, and K . Achiwa, Chem. pharmac.
Bull. (Tokyo) 13, 751 (1965).
[48] W. Lwowski and J. Simson, Abstracts, 153th Meeting Amer.
chem. SOC.,April 1967, Report 0 163; I. Simson, Dissertation,
Yale University 1967.
903
with complete retention of configuration. The absolute
configurations of the starting material[491 and of the
product 1481 are known.
y
C,H,OOC-N
+
I
Ethoxycarbonylnitrene can be prepared particularly
easily and in various ways, since the starting materials
exhibit only a small tendency to undergo the Curtius
or Lossen rearrangements [251. Alkyl- and arylcarbonyl
azides rearrange readily to isocyanates. Nitrenes are
formed from these azides only on photolysis, and not
on thermal decomposition (see below). Edwards 1511
first observed intramolecular insertion into C-H bonds
in his work on atisine and in the decomposition of
acid azides 151,521. These cyclizations are valuable in
preparative work 153,541.
:zH5
5
H-C-CH,
5. Other Carbonylnitrenes
+
C2H500C-HN-C-CH3
I
C3H7
s (+)
R (+) (11)
The optical purity of (11) is independent of the experimental conditions, as can be seen from Table 3.
Table 3. Stereospecificity of the insertion of ethoxycarbonylnitrene
into the tertiary C-H bond of 3-methylhexane.
Method of nitrene
production
a-elimination
azide thermolysis
azide photolysis
azide photolysis
azide photolysis
3-Methylhexane Observed
in CH2CIZ
rotation value
(mole- %)
(11) ( “1
1.2
100
1.2
26.8
1 00
+ 1.67
+ 1.69
+ 1.74
+ 1.71
-k 1.67
rl
Retention of
configuration
( %)
qH3
98 f 7
995 5
102* 3
100* 6
97+ 5
H
Since the proportion of triplet nitrene must be very
much higher at low substrate concentrations (1.2
mole- %) than in pure substrate, whereas the rotation
value remains unchanged, it must be concluded that
the triplet nitrene either inserts with retention of configuration or does not take part in the reaction at all.
Since the yield of insertion product decreases very
rapidly with decreasing substrate concentration, we
feel (taking earlier results [451 into account) that triplet
nitrene does not take part in the reaction I441.
It may therefore be concluded that the insertion of
ethoxycarbonylnitrene into the C-H bonds of simple
hydrocarbons is a one-step process. However, the onestep singlet mechanism need not apply to all C-H
bonds. It is known [21 that nitrenes can remove hydrogen atoms from other molecules. If this led to the
formation of a radical pair with a low activation
energy of combination, a two-step insertion mechanism
such as that found for cyanonitrenes[441 could occur [271.
The nature and size of the charge distribution in the
transition state of a one-step insertion are not fully
known, The marked preference for tertiary rather
than secondary and primary C-H bonds (Table 1) is
compatible both with a positive charge and with
“radical character” of the carbon atom involved. In
experiments with norbornane, whose C-1 and C-4
atoms accept positive charges only with difficulty,
smooth insertion into these positions has been observed. The average reactivity of all the other C-H
bonds in norbornane (per C-H bond) is only about
1.7 times as great as that of the C-H bonds on C-1
and C-41501. Thus these carbon atoms do not assume
a strong positive charge.
[49] K. Freudenberg and W. Hohmann, Liehigs Ann. Chem. 584,
54 (1954).
[SO] W. Lwowski and J. Reed, unpublished.
904
On photolysis of benzoyl azide in dimethyl sulfoxide,
dioxane/water, or acetic acid, Horner [55,561 found
products that can be attributed to intermolecular
reactions of benzoylnitrene:
0
CsHs-CON3
+ Nz
+ C~HS-CO-N=S(CH~)Z
+N2
+ CsHs-O-NHOH
C ~ H S - C O N ~H3C-COOH
+
+-
+ CH3SOCH3
CsHs-CON3 + H20
N2
+ C~HS-O-NH-O-COCH~
Products resulting from further reaction of phenyl
isocyanate were also formed.
When acetyl azide was photolysed in benzonitrile and
in phenylacetylene, Huisgen 1571 obtained 2-methyl-5phenyl-1,3,4-oxadiazole and 2-methyl-5-phenyl-l,3oxazole respectively, i.e. the formal products of 1,3cycloaddition of acetylnitrene to the triple bonds of
the solvents.
~
[51] J . ApSimon and 0 . E. Edwards, Proc. chem. SOC.(London)
1961,461; Canad. J. Chem. 40, 896 (1962).
[52] See reference 121, pages 171 and 172.
[53] W. L. Mejer and A . S . Levinson, Proc. chem. SOC.(London)
1963, 15; J. org. Chemistry 28, 2859 (1963).
[54] R . F. C. Brown, Austral. 3. Chem. 17,47 (1964).
[55] L. Horner and A . Christmann, Chem. Ber. 96, 388 (1963).
[56] L.Horner, G.Bauer, and J.Ddrges, Chem.Ber.98,2631 (1965).
[57] R . Huisgen and J.-P. Anselme, Chem. Ber. 98, 2998 (1965).
Angew. Chem. internat. Edit. Vol. 6 (1967) 1 No. I 1
assumption that the first step is the formation of a
carbonylnitrene R-CO-N, which partly rearranges
into isocyanate and partly forms nitrene products
R-CO -NXY resulting from intramolecular and
intermolecular reaction.
While trying to find a model system for the study of
the mechanism of the photolytic Curtius rearrangement [581, we observed the cyclizations of photolytically produced (with light of wavelength 2537 A)
alkylcarbonylnitrenes indicated in Table 4. In agreement with Edwards’ results 151,521, 8-lactams are
formed in preference to y-lactams [591.
NCO
CON3
I
CH2
I
YH2
CH2
I
5:
,c,
h
HzV
NH
I
H2C<$CH-R
N2 +
O\
+
CH2
C-NH
I
I
y
1
+
I
H2
yH2
R
H2C,C/C,HYH2
HZ
R
2
yH2
yH2
R
On photolysis of pivaloyl azide, pivaloylnitrene is
formed in a yield of at least 47 %. The nitrene adds to
the double bond of cyclohexene and undergoes insertion into C-H bonds[59,601. Four isomers of (12) are
obtained from 2-methylbutane in an overall yield of
25 ”/.
The photolysis of pivaloyl azide in cyclohexane gives
N-cyclohexylpivalamide in 20 ”/, yield, and 0.5 ”/, of
pivalamide. The quantity of gas evolved is always
Table 4. Photolysis of n-alkanecarboxylic azides i n dichloromethane and cyclohexane.
~~
Starting material
R-CON3
where R =
Photolysis in dichloromethane:
Yield of intramolecularly formed products
Photolysis in cyclohexane:
Yield of intramolecularly
formed products
Yield of N-cyclohexyl-n-alkanecarboxamide
23.5 % 5-Methyl-2-pyrrolidinone
7.8 % 5-Methyl-2-pyrrolidinone
8.1 %
4.7 %
1.9 %
1 1 . S % 8-Lactam
22.7 % 6-Methyl-2-piperidinone
I
21 % 6-propyl-2-piperidinone
The preferential formation of the six-membered
lactams agrees with the assumed three-membered ring
transition state in insertions into C-H bonds (eq. (6)).
The transition state for the formation of the 8-lactani
would be a six-membered ring with a fused threemembered ring. The assumption of a linear transition
state for the insertion (-N . . . . . .H . . . . .C-) requires
a distorted seven-membered ring with three atoms
arranged in a line for the formation of the s-lactam;
this is probably a much less favorable arrangement.
1.4 %
j
11.5 % S-Lactam
greater than would correspond to the theoretical yield
of nitrogen. Analysis showed that isobutylene is also
formed. (The isobutylene was determined as its mercuric nitrate complex 1611.) For example, photolysis in
6. Pivaloylnitrene and the Mechanism of the
Curtius Rearrangement
Edwards [521, Horner 1561, Huisgen C571, and we [601 observed that (unrearranged) nitrene products are obtained only on photolysis, and not on thermolysis, of
alkane- and arenecarboxylic azides. No nitrene products can be found on thermolysis, in spite of careful
searches, whereas they are often formed in large
quantities on photolysis. This shows that no nitrenes
are formed during the thermal Curtius rearrangement;
the reaction evidently proceeds in a single step with
simultaneous removal of the nitrogen and migration
of the alkyl or aryl group. Thus the Stieglitz formulation (2) [71 is invalid at least for the thermal rearrangement, and it must be assumed that the reaction
proceeds as indicated in eq. (3). In the photolytic
reaction, on the other hand, it seemed a reasonable
. __ ..
1581 L . Horner, E. Spietschkn, and A . Gross, Liebigs Ann. C h e m .
573, 17 (1951).
[59] S.Linke and W. Lwowski, unpublished.
[60] W . Lwowski and G. T. Tisue. J. Amer. chem. SOC.87, 4022
(1965).
Angew. Chem. infernut. Edit. J Yol. 6 (1967)
1 No. I 1
neopentane gives a 1 3 ”/, yield of isobutylene. Thus if
the pivaloylnitrene is not intercepted in some other
way, it appears to decompose into isobutylene and
isocyanic acid. The isocyanic acid, together with part
of the isobutylene, forms a polymeric oil, whose I R
spectrum shows C-H absorption (at 2994 cm-1) as
well as 0 - H or N-H (broad, about 3290 cm-1) and
carbonyl absorptions (broad, about 1686 cm-1). Including the oil and t-butyl isocyanate, the total yields
were always greater than 90 %, based on the pivaloyl
azide used [59J.
The selectivity of the insertion of pivaloylnitrene into
C--H bonds is greater than that of ethoxycarbonylnitrene, and is 160:9:1 for the tertiary, secondary, and
primary C-H bonds in 2-methylbutane 1601.
-~
. -
1611 G. Dettigis, C. R . hebd. Seances Acad. Sci. 126,1043 (1898);
C . D . Hurdand A . R . Goldsby, J. Arner.chem. SOC.56,1815 (1934);
C. E. Storr and T . Lane, Analytic. Chem. 21, 572 (1949).
905
The yield of t-butyl isocyanate (on photolysis with
light of wavelength 2537 8, at -10°C) is always
39-42 "/o. This provides a clue to the mechanism of
the rearrangement.
If pivaloylnitrene were the precursor of the isocyanate,
the quantity of isocyanate obtained should decrease
with increasing ability of the solvent to intercept
nitrenes. In neopentane (which contains only unreactive primary C-H bonds), only 0.2 % of Nneopentylpivalamide is obtained, together with 0.7 %
of pivalamide, 40 % of t-butyl isocyanate, 13.5 %
of isobutylene, and polymeric oil. In cyclopentane,
which contains the much more reactive secondary
C-H bonds, 1 3 % of insertion product, 41 % of isocyanate, and 0.5 % of pivalamide are obtained,
together with isobutylene and polymer. In cyclohexene,
in which the addition and insertion products are
formed in a total yield of 47 %, the isocyanate is
formed in 41 % yield. Thus the isocyanate cannot
have been formed from the nitrene, but must be
formed directly from excited pivaloyl azide. The
photolytic Curtius rearrangement is therefore also a
concerted reaction.
906
It is interesting to note that pivaloylnitrene does not rearrange
into isocyanate. On paper, there appears to be nothing to
prevent such a rearrangement, and it might be assumed that
such a reactive intermediate would react intramolecularly
much more readily than intermolecularly. Nevertheless, the
life of pivaloylnitrene must be appreciable (extending over
many collisions), since it would otherwise be impossible to
explain its high selectivity coupled with relatively good yields
(e.g. in insertion into C-H bonds).
Our understanding of the finer details of the chemical
behavior of nitrenes (and of electron-deficient intermediates in general) is still incomplete. Further
theoretical and experimental effort will be necessary
before it will be possible to make reliable predictions
about the behavior of such intermediates.
I a m very gratefiil to my co-workers for their untiring
eflorts and for the numerous contributions they have
made to the planning of new experiments. Our work
would not have been possible without fhe generous
financial suppcrt of the National Institutes of Health
and the National Science Foundation (U.S.A.).
Received: March 9th, 1967
[A 602 1El
German version: Angew. Chem. 79,922 (1967)
Translated by Express Translation Service, London
Angew. Chem. internat. Edit.
Vol. 6 (1967) 1 No. I 1
Документ
Категория
Без категории
Просмотров
2
Размер файла
926 Кб
Теги
decompositions, nitrene, carbonylaziden
1/--страниц
Пожаловаться на содержимое документа