close

Вход

Забыли?

вход по аккаунту

?

Conformational Aspects of Many-Membered Rings.

код для вставкиСкачать
obtained in anhydrous dimethylformamide at 150 'C
(Scheme 2,d).
e) R e a c t i o n w i t h A m i n e s
Poly-y-hydroxycrotonolactonereacts with amines in
aqueous media at 20 to 50°C to give the ammonium
salt of the polymeric y-hydroxy acid (see Scheme 2,el).
In organic solvents, e.g. in an excess of the amine,
higher reaction temperatures are required, and the
polymeric y-hydroxy carboxamide is then formed
(Scheme 2,ez).
f) R e d u c t i o n w i t h L i t h i u m
Tetrahydridoaluminate
Polymeric crotonolactone can be smoothly reduced
with LiAIH4 in tetrahydrofuran to give a water-soluble
polymer of 2-butene-l,4-diol (Scheme 2,f).
W e are grateful to Professor Korte for his interest in this
work.
Received: December Zlst, 1965; revised August 25th, 1966 [A 549 IE]
German version: Angew. Chem. 78, 1093 (1966)
Translated by Express Translation Service, London
Conformational Aspects of Many-Membered Rings1*]
BY PROF. J. DALE
UNION CARBIDE EUROPEAN RESEARCH ASSOCIATES, BRUSSELS (BELGIUM)
[**I
The stereochemistry of the various chemical groupings which constitute the structural units
of ring systems isfirst reviewed. It is then shown that the strain-free possibilities even of
large rings are restricted, and that the ring conformations theoretically derived generally
correspond to those found by physical methods. Variations of physical properties and of
cyclization yield in homologous series are explained on a conformational basis.
Introduction
Although the conformational analysis of aliphatic sixmembered rings has had a remarkable development [% 31
over the twenty years following HasseZ's demonstration[ll by electron diffraction that the chair is the
normal stable form and that axial and equatorial
substituents can be distinguished, the development has
been much slower in other ring systems. In larger rings
the understanding of conformational details was virtually lacking until Dunirz's recent demonstration 141 by
X-ray diffraction of a unique conformation of the tenmembered ring skeleton in a variety of cyclodecane
derivatives. It is true that synthetic difficulties, certain
anomalies in physical properties, and abnormal transannular reactions of medium (eight- to eleven-membered) rings had been rationalized vaguely as being due
to compression in the interior of the ring and to the
[*I For previous reviews see J. Sicher in: Progress in Stereochemistry. Butterworths, London 1962, Vol. 3, p. 202; R. A .
Raphael, Proc. chem. SOC.(London) 1962, 97; V . Prelog, Pure
appl. Chem. 6 , 545 (1963).
[ * *] Present address: Chemical Institute, University of Oslo,
Blindern (Norway).
[l] 0. Hassel, Tidsskr. Kjemi, Bergves. Metallurgi 3, 32 (1943).
[2] D. H . R. Barton and R. C . Cookson, Quart. Rev. 10,44 (1956).
[3] E.L.Elie1: Stereochemistry of Carbon Compounds. McGrawHill, New York 1962.
[4] E. Huber-Buser, J. D . Dunitz, and K . Venkatesan, Proc. chem.
SOC.(London) 1961, 463.
lo00
proximity of atoms across the ring, but the conformational details remained obscure.
The nature of the unique Dunitz conformation of cyclodecane (see below) is such that, as in the cyclohexane
chair, the carbon skeleton follows quite closely the
diamond lattice; however, since this involves too close
an approach between some of the hydrogen atoms,
a compromise is reached between angle opening, torsion
of bonds, and hydrogen repulsion.
As shown in Fig. 1, a C-C-C angle opening of 5 "
demands only an energy of 0.5 kcal/mole, of lo", 1.9
kcal/mole, while eclipsing a single bond in saturated
hydrocarbons costs 2.8 kcal/mole [51. Oscillation of the
C-C-C angle in a CH2 chain (normally 112 ") between
100 and 124 O, therefore, is as easy as rotation about a
single C-C bond.
Crystal structures of some other medium and large
ring compounds have been determined NJ and have
confirmed the view that the angular (Baeyer) and torsional (Pitzer) strain are well balanced and, if physically
possible, are both totally absent. Knowledge of the
most stable orientation about single bonds thus becomes particularly important, since the shape of a large
ring depends much more on the orientation of the chain
members about each bond than on the exact value of
the bond angles.
[5] J. B. Hendrickson, J. Amer. chem. SOC.83, 4537 (1961).
[6] Earlier summaries: J . D. Dunirz and V. Prelog, Angew.
Chem. 72, 896 (1960).
Angew. Chem. internat. Edit.
Vol. 5 (1966) 1 No. 12
I
I
3-
I
I
120" 124O"I
w
-
1 -
-
-
-50"
-60"
-40"
-30"
-20"
0"
-10"
10"
20"
30"
LO"
50"
60"
rn
Fis. 1. Change in the potential energy (AE, in kcal/mole) of polyrnethylene chains with deviations from-the
equilibrium values of the valency
angle (inner curve) and the dihedral angle (outer curve) The numbers on the curves are the actual angles.
Some Basic Stereochemical Facts
The most reliable information about the basic stereochemistry of the various molecular units comes from
direct physical methods such as X-ray diffraction,
electron diffraction, and microwave spectroscopy
(rotational spectra). Other spectroscopic methods, such
as electronic spectra (ultraviolet and visible), vibrational
spectra (infrared and Raman), and nuclear magnetic
resonance spectra, can also yield useful information
but are rarely of comparable precision. Conclusions
have also been drawn from dipole moments and
thermodynamic data.
The stable conformations of a C-C single bond in a
saturated carbon chain are illustrated in Fig. 2 "1, which
shows the potential wells for the staggered positions,
-2
2
cp 60"
180"
300"
360"
a
m
240"
120"
=
b
C
d
C
b
4
Fig. 2. Dependence of the potential energy (in kcallmole) of a tetramethylene chain on the conformation of the central C-C bond 171. The
dihedral angle is denoted by 0 .
a and c, eclipsed conformations; b gauche conformation;
4 anti conformation.
[7] M . Kobelt, P. Barman, V. Prelog, and L. Ruzicka, Helv. chim.
Acta 32, 256 (1949).
Angew. Chem. internat. Edit.
Vol. 5 (1966) J No. I 2
separated, even in the case of ethane, by surprisingly
high and theoretically unsatisfactorily explained barrier
peaks at the eclipsed positions [S]. The lower energy of
the anti (or trans) as compared to the two gauche conformations leads to a tendency for long paraffin chains
to be essentially all-anti and extended at low temperatures, and this is unfavorable for cyclization. If the
chain carries an alkyl branch, the adjacent C-C bond
does not distinguish between the branch and the main
chain, so that the probability of having a "kink" in the
main chain, necessary for cyclization, increases ; this is
the simplest explanation for the gern-dimethyl effect 191
in cyclization processes. Carbon-carbon single bonds
next to almost any group other than CH2, CH3. or
heavy atoms (higher halogens), take up gauche as easily
as anti conformations; thus a bend of the main chain
becomes more probable next to double bonds, triple
bonds, carbonyl groups, hetero-atoms, etc., and cyclization is facilitated 1101,
Fig. 3 shows how the stereochemistry of most of the
structural elements and the adjoining atoms occurring in
ring systems can be derived from that of saturated
hydrocarbons by assuming that hydrogen atoms can be
replaced with no essential stereochemical change by a
lone electron pair in atoms like oxygen, nitrogen, and
sulfur, and that double bonds can be considered as two
bent bonds [111. The barriers of rotation (kcal/mole) in
the illustrated or in similar compounds are given at the
bond in question. It is remarkable how similar they are,
except those next to multiple bonds or to hetero-atoms
engaged in partial conjugation (ester, amide), where the
[8] D. J . Millen in : Progress in Stereochemistry. Butterworths,
London 1962, Vol. 3, p. 138.
[9] F. G . Bordwell, C . E. Osborne, and R . D . Chapman, J. Amer.
chem. SOC.81, 2698 (1959); N. L. Allinger and V. Zalkov, J. org.
Chemistry 25, 701 (1960).
[lo] J. Dale, J . chem. SOC.(London) 1963, 93.
[ l l ] L. Puuling, Proc. nat. Acad. Sci. USA 44, 211 (1958).
1001
Ibi
Icl
for dimethylhydrazine 1231. For the disulfide group it has
been assumed [24,251 that the normal dihedral angle is 90 O,
although in some carbon-substituted open-chain systems
like L-cystine[261 and its hydrochlorider271 it is as low as
74 and 79 '. In a six-membered 1,Zdithiane derivative, a
perfect gauche value of 60' was found[251. In linear polysulfidesI281, too, the dihedral angles are all close to gauche,
varying from 61 to 82
'.
Id1
lei
I0
lhi
Ikl
Fig. 3. Derivation of the stereochemistry of some structural elements
from that of a saturated hydrocarbon. The numbers are energy barriers
(kcal/mole) for rotation about the respective bonds.
values become much lower so as to make the chain
more flexible. For the ether linkage (Fig. 3,e) both the
rotational barrier and the conformation have been
determined on dimethyl ether [121; f o r the sulfide linkage
(k) on dimethyl sulfide [131; and the barrier in secondary
amines (h) on dimethylamine [141.
Although only the anti conformation is shown in Fig. 3, it is
likely that the gauche and anti conformations of the bond
between carbon and hetero-atom in homologous compounds
have comparable energies. Thus in crystalline polyoxymethylene [I51 all C - 0 bonds are gauche: in polyethylene
glycol[l5] they are all anti (and the C-C bond gauche). The
situation is less clear when two hetero-atoms are adjacent
(Fig. 3; d, g, j). Thus it has been assumed generally that the
torsional angle in hydrogen peroxide is 90 to 100 '. However,
the observation that one of the possible values for the dihedral angle in hydrogen peroxide dihydrate [I61 is 68 O and
that in sodium oxalate perhydrate1171 it is 180' indicates
that it is a normal anti-gauche system; this is supported by
the presence of perfectly chair-shaped six-membered tetraoxa rings in sodium peroxoborate
and in various dimeric
cycloalkanone peroxides [191. With hydrazine 1201 there is
agreement about a threefold barrier and that gauche and
anti forms coexist; even in bishydrazinezinc chloride [211
the torsional angle is 74". The situation is similar in tetrafluorohydrazine [*21, but no conclusions could be reached
[12] P.H.Kasai and R.J.Myers, J. chem. Physics 30, 1096 (1959);
U. Blukis, P. H . Kasai, and R. J . Myers, ibid. 38, 2753 (1963).
[13] L. Pierce and M . Hayashi, J. chem. Physics 35, 479 (1961).
[14] W. G . Fateley and F. A . Miller, Spectrochim. Acta 18, 977
(1962).
[15] S. I. Mizushima and T. Shimanouchi, J. Amer. chem. SOC.86,
3521 (1964).
[16] I. Olovsson and D . H . Templeton, Acta chem. scand. 14, 1325
(1960).
[17] B. F. Pedersen and B. Pedersen, Acta crystallogr. 16, A 75
(1963).
[18] A. Hansson, Acta chem. scand. IS, 934 (1961).
[19] P. Groth, Acta chem. scand. 18, 1301, 1801 (1964); 19, 1497
(1965).
[20] D . W. Scott, G. D . Oliver, M . E. Gross, W. N. Hubbard, and
H . M . Huffman, J. Amer. chem. SOC.71, 2293 (1949); E. L. Wagner and E. L. Bulgordy, J. chem. Physics 19, 1210 (1951).
[21] A . Ferrari, A . Braibanti, and G . Bigliardi, Acta crystallogr.
16, 498 (1963).
221 D . R . Lideand D. E. Mann, J. chem. Physics 31,1129 (1959).
[
Both hydrogens o f a CH2 group can be replaced by one
bivalent group to form a double bond without disturbing
the stereochemistry of the carbon chain, as shown for
the ketone in Fig. 3,c. This basic carbon skeleton is the
predominant one in methyl ethyl and diethyl ketone [291,
and the same stereochemistry and low rotational barrier
next to the carbonyl group has been established in
numerous other carbonyl compounds, like aldehydes [30-321, acetyl derivatives 133,341, and acyl halides[309331. The barrier next to the carbonyl group
(Fig. 3,c) was determined on acetone [14,351, that of the
C C bond
, ~
on propionaldehyde [321. For a double bond
between two carbon atoms, the situation is similar
(Fig. 3,a), but the torsional barrier in adjacent single
bonds is only moderately reduced in mono- and transdisubstituted 1361 and l,l-disubstituted [371 olefins, whereas it is much lower (0.6 to 1.4 kcal/mole) in cis-disubstituted olefins [361. It is noteworthy that even in cis2-butener381 the conformation derived from Fig. 3,a is
strictly obeyed even if this implies the closest possible
hydrogen-hydrogen distance; conformers of higher
l-alkenesC391 follow the same pattern. Next to a C=N
double bond the basic stereochemistry is the same, as is
shown in the case of N-methylenemethylamine [401,
with a torsional barrier for the adjoining single bond of
1.97 kcal/mole.
~
[23] W . H . Beamer, J. Amer. chem. SOC.70, 2979 (1948).
[24] J. Donohue, A . Caron, and E. Goldish, J. Amer. chem. SOC.
83,3748 (1961).
[25] 0. Foss, K. Johnsen, and T . Reistad, Acta chem. scand. 18,
2345 (1964).
[26] B. M. Oughton and P. M . Harrison, Acta crystallogr. 12,396
(1959).
[27] L. K. SteinrauA J. Petersen, and L. H . Jensen, J. Amer. chem.
SOC.80, 3835 (1958).
[28] S. C . Abrahams and E. Grison, Acta crystallogr. 6, 206
(1953); S. C . Abrahams, ibid. 7, 423 (1954).
[29] C. Romers and J. E. G. Cveutrberg, Recueil Trav. chim.
Pays-Bas 75, 331 (1956).
[30] E. B. Wilson, Proc. nat. Acad. Sci. U.S.A. 43, 816 (1957).
[31] R. J . Abraham and J. A. Pople, Molecular Physics 3, 609
(1960).
[32] S. S. Butcher and E. B. Wilson, J. chem. Physics 40, 1671
(1964).
[33] L. C. Krisher and E. B. Wilson, J. chem. Physics 31, 882
(1959); 33, 304 (1960).
[34] 0. L. Stiefvafer and J. Sheridan, Proc. chem. SOC.(London)
1963, 368.
[35] J . D . Swalen and C . C . Cosfain, J. chem. Physics 31, 1562
(1959).
[36] R. A. Beaudet, J. chem. Physics 38, 2548 (1963); 40, 2705
(1964).
[37] V. W. Laurie, J. chem. Physics 34, 1516 (1961); 39, 1732
(1963).
[38] T. N . Sarachmann, quoted in [36].
[39] A . A. Bothner-By and C . Naar-Colin, J. Amer. chem. Soc.83,
231 (1961); A . A . Bothner-By, C . Naar-Colin, and H . Giinther,
ibid. 84, 2748 (1962).
[40] J.T. Yardley, J . Hinze, and R. F. Curl, J. chem. Physics 41,
2562 (1964).
Angew. Chem. internat. Edit.
/
Vol. 5 (1966)
/ No.
12
If an oxygen atom is next to the double bond, enol
ethers are obtained whose conformation should be as
in Fig. 3,f. Esters show a strong preference for the
planar anti conformation (unfortunately often referred
to as cis in formates) [41-431, and it seems reasonable to
attribute this to partial double bond character in the
C-0 bond next to the carbonyl group. Except for
formic acid [441 and its ethyl ester [451, there is no evidence
for another conformation in open chains; a planar
cis-ester would give eclipsed single bonds, and a nonplanar gauche-ester would not be able to acquire
double-bond character. It has been pointed out c461 that
in lactones with a six-membered ring the ester group is
cis-planar, but the examples quoted are all fused-ring
systems favoring a boat conformation. In methyl vinyl
ether a second (unstable) form has been observed 1431,
presumably because resonance stabilization in the planar
anti form is less important in a C=C-0 than in an
O=C-0 system; it is interesting that the latter is
not planar cis, but nonplanar gaache. The partial
double-bond character is probably also responsible
for the lowered torsional barriers in the adjacent
single bonds of esters (Fig. 3,f), although these
were not determined on higher esters, but on acetic
acid [301 and methyl formate 1421. The crystal structures
of aliphatic carboxylic acids [471 and dicarboxylic
acidsr48,491 also demonstrate that torsion of the C-C
bond next to the carbonyl group is particularly easy;
the carboxyl group is usually oriented as shown in
Fig. 3,f, in or close to the plane of the methylene chain,
but sometimes it is twisted by up to 30” by packing
forces 1481.
The N-monosubstituted amide group (Fig. 3,i) is
completely analogous to the ester group. The anticonformation is strongly favored [50,511, and again only
in formamides is there evidencer51l for another less
stable isomer, supposed to be planar cis. The torsional
barriers are not known, but they seem to be low in the
single bonds on both sides. Thus there are small
deviations from the ideal anti and gauclie values (180
and 60”) in the N-C(or) bond of acetamide-like compounds 152,531 and in the %-helix of proteins 1541, and a
stronger deviation in the C-C(x) bond of the cx-helix [541.
- ~-._
~
[41] J . K. Wilmshurst, J. molecular Spectroscopy I , 201 (1957).
[42] R. F. Curl, J. chem. Physics 30, 1529 (1959).
[43] N. L. Owen and N. Sheppard, Proc. chem. SOC.(London)
1963, 265.
[44] T. Miyazawa and K.S.Pitzer, J.chern.Physics 30,1076 (1959).
[45] D. Tabuchi, J. chem. Physics 28, 1014 (1958).
[46] A. M. Mathieson, Tetrahedron Letters 1963, 81.
[47] F. J. Strieter, D. H. Templeton, R. F. Scheuermann, and R.L.
Sass, Acta crystallogr. 15, 1233, 1240, 1244 (1962).
[48] C. H. MacGillavry, C. Hoogschagen, and F. L. J. Sixma,
Recueil Trav. chim. Pays-Bas 67, 869 (1948).
[49] J. D. Morrison and J. M . Robertson, J . chern. SOC.(London)
1949,980,987, 993, 1001.
[50] S . I. Mizushima, T. Simanouti, S. Nagakura, K. Kurataiii,
M.Tsuboi, H. Baba, and 0. Fujioka, J. Amer. chem. SOC.72, 3490
(1950), J . E. Worsham and M . E. Hobbs, ibid. 76,206 (1954).
[51] L. A. LaPlanche and M.T.Rogers, J. Amer. chem. SOC.86,
337 (1964).
[52] M. Bailey, Acta crystallogr. 8, 575 (1955).
[53] J . Donohue and R.E.Marsh, Acta crystallogr. 15, 941 (1962).
[54] T. Mivazawa, J. Polymer Sci. 55, 215 (1961).
Angew. Chem. internat. Edit.
/
Vol. 5(1966) I No. 1 2
In less stable protein structures the deviations are much
greater [541. Since the double-bond character of the C-N
bond (of formamide [55l) is only 25 %, the cis-hydrogen
on nitrogen (Fig. 3,i) is not expected to be coplanar with
N-C=O ; deviations from planarity have been observed
in N-acetylglycine [531 and formamide [551.
A similar situation arises in conjugated diolefins in which
the central single bond has a preference for the transoid
(s-trans or s-anti) conformation shown in Fig. 3,1, since it has
double bond character. “Resonance” being no doubt less
important here than in amides and esters, the “other form”
may even be non-planar gauche instead of planar cis. Only
the transoid (anti) conformer has been found in butadiene
itself by electron diffraction [561 and vibrational spectra [571,
and a vain search for the cis-conformer by microwave
spectroscopy has been reported 1581. Similarly, fluoroprene [581
and isoprene [591 are planar anti, and the well-known linear
shape of all-trans-polyenes and carotenoids implies also
anti-conformation of the single bonds. Isomerization of
hydrocarbons having two double bonds in a 1,3-dimethylcyclohexane 1601 or a decalin [611 skeleton shows a preference
for positions where the conjugated diene can assume an
anti-conformation. On the other hand, when in rings of more
than five members a cisoid conformation is structurally
imposed, there exists no evidence that the diene is planar; on
the contrary, the low intensity of ultraviolet absorption of
1,2-dimethylenecyclohexane1621 suggests a non-planar “gauchoid” conformation. The crystal structure of calciferol[631
shows that an analogous “cis”-diene system is non-planar
with a torsional angle of 54 O , close to the gauche conformation, but here, of course, steric hindrance may play a role[*].
Steric hindrance also blurs the picture in the p-ionone ring
of vitamin A acid
and lower homologues 165.661, but it is
interesting that in one of the crystal structures the ring
double bond is anti and almost coplanar with the polyene
system of the chain [651, whereas in the other cases [64,661 it is
gauchoid with torsional angles of ca. 35 and 80’. When a
C - C double bond is conjugated with a n aldehyde group,
as in acrolein, anti-preference for the enclosed single bond
has been established[67], but when it is conjugated with a
carboxyl group, almost planar cisoid conformations have
been observed [64,651, presumably because the torsional
barrier is further reduced by resonance in the hydrogenbonded carboxyl group. Delocalization of the double bonds
[55] C. C. Costain and J . M . Dowling, J. chern. Physics 32, 158
(1960).
[56] A . Almenningen, 0. Bastiansen, and M. Traetteberg, Acta
chern. scand. 12, 1221 (1958).
[57] D. J. Marais, N. Sheppard, and B. P. Stoichefl, Tetrahedron
17, 163 (1962).
[58] D. R . Lide, J. chem. Physics 37, 2074 (1962).
[59] R.J. W. CeFkvre and K.M.S.Sundaram, J. chern. SOC.(London) 1963,3547; D. R. Lide and M . Jen, J. chem. Physics 40,252
(1964).
[60] 0. H. Wheeler, J. org. Chemistry 20, 1672 (1955).
[61] R . B. Bates, R. H . Carnighan, and C . E. Staples, J. Amer.
chern. SOC.85, 3030 (1963).
[62] W. J. Bailey and H . R. Golden, J. Arner. chem. SOC.75,4780
(1953); A.T.Blomquist and D.T.Longone, ibid. 79, 3916 (1957).
[63] D. Crowfoot Hodgkin, B. M . Rimmer, J. D. Dunitz, and K . N.
Trueblood, J. chem. SOC.(London) 1963, 4945.
[*I Added in proof: A recent microwave investigation led to a
non-planar diene system also for 1,3-cyclohexadiene itself. The
dihedral angle is only 18 ’. S. S . Butcher, J. chem. Physics 42,1830
(1965).
[64] C. H. Stam and C. H . MacGiIlavry, Acta crystallogr. 16, 12
(1963).
[65] E. L. Eichhorn and C. H. MacGillavry, Acta crystallogr. 12,
872 (1959).
[66] B. Koch and C. H. MacGillavry, Acta crystallogr. 16, A 48
(1963).
[67] J . Fine, .
I
H
.
. Goldstein, and J. W.Simmons, J. chem. Physics
23, 601 (1955).
1003
must also be held responsible for the free or nearly free
internal rotation of methyl substituents in aromatic systems,
e.g. in methylpyridines [681. Similarly, methyl groups next
and to cylindrito trigonal boron, as in trimethylboron
cally symmetrical sp carbon atoms, as in dimethylacetylene [701,
are essentially free to rotate.
Cycloalkanes a n d Simple Substituted Derivatives
The only strain-free saturated carbon skeletons are
those which can be derived from the diamond lattice “01.
This means that only even-membered carbon rings c6
and higher can have skeletons without torsional strain,
but the 8-, lo-, and 12-membered (medium) rings are
strained because of repulsion between hydrogen atoms
pointing “inwards” (Fig. 4 and 5).
In rings larger than the 14-membered ones, many of the
strain-free skeletons that can be derived using space-filling
models, which do not allow torsional or angular strain 1101,
are supposedly unstable because they either contain more
gauche-bonds than the necessary minimum or have too open
structures that d o not fill space efficiently. Adopting this
horror vacui postulate, one arrives at the conclusion [lo] that
only the 14-, 18-, 22-, 26-, etc. membered rings can have the
ideal “rectangular” conformation shown in Fig. 4, with two
parallel long chains roughly at van der Waals distance linked
at each end by Cp bridges. The 16-, 20-, 24-, etc. membered
rings must have C3 bridges if they are to be without torsional
strain, or may prefer to collapse, adopting some strain in
order to have C2 bridges and better internal van der Waals
contacts (Fig. 5 ) . Cyclododecane may change similarly to
avoid the three hydrogen-hydrogen repulsions (Fig. 5).
Structural determinations have confirmed the first three
conformations in Fig. 4. The ring conformations of
several derivatives of cyclodecane have been found to
be almost identical and close to the ideal skeleton shown
in Fig. 4. The crystal structures determined so far
include two different modifications of trans-l,6-cyclodecanediammonium dichloride 1711, the corresponding
cis-isomer [721, cyclodecylammonium chloride 1731, and
tvans-l,6-dibromocyclodecane
[741. The observed maximum C-C-C angle opening (to 120”) and deviation
from a staggered position (up to 29”) both represent
strains of about 1.3 kcal/mole (Fig. 1)and are therefore
balanced. The crystal structure found for cyclotetradecane1751, as well as for the structurally equivalent 1,8diazacyclotetradecane dihydrobromide 1761, is very close
to the quasi-rectangular one derived in Fig. 4,third row;
in the latter compound, the cationic nitrogen atoms
occupy the corner positions. For several higher cycloalkanes, early X-ray results [771 suggested two parallel
straight polymethylene chains at van der Waals distance,
but gave no details of the folding at each end. The recently determined crystal structure of cyclotetratriacon-
:rll
+o+ 0
= o
o+-
Lf-j
0 + o + o
0
Fig. 5. Conformational possibilities of cycloalkanes ( C H Z ) ~The
~.
carbon atoms in the projections on the right-hand side are marked with
reference to a plane through the corner atoms;
denotes above, denotes below the plane.
+
Fig. 4. Conformations of strain-free and compact cycloalkanes of
general formula (CH2)4n+z. Right: actual conformation: left: hypothetical planar conformation. D and I. substituents are indicated by and - - -, respectively. g, Bond, with gauche conformation; a, bond
with anti conformation.
[68] D. W. Scott et a/., J. physic Chem. 67, 680, 685 (1963).
[69] L. S. Bartell, B. L. Carroll, and J. P . Guillory, Tetrahedron
Letters 1964, 705.
[70] R. Kopelntan, J. chem. Physics 41, 1547 (1964).
1004
0
[71] E. Huber-Buser and J. D . Dunitr, Helv. chim. Acta 43, 760
(1960); 44, 2027 (1961).
[72] J. D . Dunitr and K . Venkatesan, Helv. chim. Acta 44, 2033
(1961).
[73] M . H . Mladeck and W . Nowacki, Helv. chim. Acta 47,1280
(1964).
[74] J.D.Dunitz and H. P. Weber, Helv. chim. Acta 47, 951 (1964).
[75] B. A. Newman (Bristol), unpublished.
[76] J. D . Dunirz and E. F. Meyer !Zurich), unpublished.
[77] A . Muller, Helv. chim. Acta 16, 155 (1933).
Angew. Chem. internat. Edit. 1 Vol. 5 (1966)
No. I 2
- +
0
tane ((234) has confirmed this 1751 and established that
the folding consists of the bond sequence
. . . gauche, gauche, anti, gauche, gauche,. . .
as shown in Fig. 4. Since the cycloalkanes of the series
( C H Z ) ~ ~ therefore,
+~,
seem to possess well-defined
centrosymmetrical conformations, L and D substituents
(on either side of the hypothetical flat ring which the
Fischer and Haworth projections are based upon) have
been marked in Fig. 4 in order to facilitate the discussion of steric relationships.
In the series of cycloalkanes (CH& (Fig. 5) the situation is less clear. Only for cyclododecane are conclusive
X-ray results available [781. Its approximately square
conformation (Fig. 6,a) although deviating systematically from the diamond lattice, seems to be typical for
this ring size, since it is also found in azacyclododecane
la /
torsional strain, and hydrogen repulsion, showed (821 that
several cyclooctane conformations have comparable energies[*], and electron diffraction data on the gas could not
be explained by any one single model, but fairly well with a
mixture of several models [831 [*I.
The lowest member, cyclobutane, provides a striking
demonstration of the necessity of the C-C single bonds
to get away from the completely eclipsed conformations
of a planar structure even at the expense of further
reduction of the already very strained C-C-C angles
to below 90 ’. Thus cyclobutane itself 1841 is non-planar,
and the angle between the planes of the two halves is
29 in bromocyclobutane (851.
The inherent strain in odd-membered cycloalkanes can be
distributed in various ways without, or with only little,
difference in energy. This uncertainty, or travelling of a n
“imperfection”, has been termed pseudo-rotation and was
first recognized in cyclopentane [861. When substituents are
introduced, certain symmetries, “envelope” and “halfchair” [86,871 are preferred. The detailed calculations by
Hendrickson have led to a deeper understanding of pseudorotation also in the cyclohexane flexible form (51 (between
“classical” and “twisted” boat), in the two conformational
families (“boat” and “chair”) of cycloheptane [51, and in
medium rings [8*L The undetermined nature of odd-membered rings is also reflected in the crystal structure of cyclononylammonium bromide 1881, which contains two similar
conformers with C-C-C, angles up to 124 and deviations
up to 44 from the staggered position of the bonds.
The alternation in conformational stability between
even- and odd-membered rings is reflected in the melting
points (sometimes a transition point), particularly of
the larger rings [lo]. Singly substituted cycloalkanes as
well as saturated macro-heterocycles behave sin+
larly [lo].
Cyclic Ketones
Fig. 6. Crystal structure of (a) cyclododecane [941 and (b)
azacyclododecane hydrochloride [951.
8
hydrochlorider791, with the cationic NH2 group at a
corner position (Fig. 6,b). In both compounds the
torsional strain is similar to that found in cyclodecane
derivatives, but the maximum bond angle is smaller
(116-117 ”). In higher homologues ( C H Z ) ~the
~ same
amount of strain might be distributed over more bonds,
and the regular increase of the longest period in the
crystal lattice 1771 suggests that the torsionally strained
C2 bridged compact conformations are preferred to
the unstrained C3 bridged open conformations.
For cyclooctane the conformational situation is still undecided; evidence has been obtained both for the saddle
form[lo,8ol derived from the diamond lattice (shown in
Fig. 5 ) , and for a crown form [811. In the isoelectroniccation of
azacyclooctane hydrobromide, a twisted crown is indicated [61.
Machine computation, taking into account bond angle,
[78] J. D. Dunirz and H . M . M. Shearer, Helv. chim. Acta 43, 18
(1960).
[79] J. D. Dunitz and H . P. Weber, Helv. chim. Acta 47, 1138
(1964).
[80] J. Dale, I . Laszlo, and W. Ruland, Proc. chern. SOC.(London)
1964, 190.
[81] N. L. Allinger, S. P. Jindal, and M . A. DaRooge, J . Amer.
chem. SOC.27,4290 (1962).
Angew. Chem. internat. Edit.
1 Vol. 5 (1966) 1 No. 12
The cycloalkanones show a remarkable similarity with
the cycloalkanes in melting point behavior [lo]. The
CH2-scissoring vibration also follows the same splitting
pattern through the medium ring region in the two
series 1891, and the higher symmetrical cyclic diketones
(c24, C28) show the same feature of two parallel long
[82] J. B. Hendrickson, J . Amer. chem. SOC.86, 4854 (1964);
cf. also K . B. Wiberg, ibid. 87, 1070 (1965).
[ * ] If the value 112”, instead of 109.5
for the preferred
C-C-C angle had been used, the result might have been different.
[83] 0. Bastiansen and H . Jensen (Oslo), unpublished.
[*I Added in proof: Recently, a hybrid saddle-crown conformation
(boat-chair form) has been established for the eight-membered
ring in dimeric cyclooctanone peroxide (P. Groth, Acta chern.
scand. 19, 1497 (1965)) and in cyclooctane-1,Z-truns-dicarboxylic
acid (J. D. Dunitz and A. Mugnoli, Chem. Commun. 1966, 166).
On the other hand, the low-temperature NMR-data reported for
perfluorocyclooctane (A. Peake, J. A. Wyer, and L. F. Thomas,
Chem. Commun. 1966, 95) can, in the writer’s opinion, only be
accounted for by a saddle conformation.
[84] A. Almenningen, 0. Bastiansen, and P. N . Skancke, Acta
chem. scand. 15, 711 (1961).
[85] W. C. Rothschild and B. P. Dailey, J . chem. Physics 36,2931
(1962).
[86] K . S. Pitzer and W. E. Donath, J. Amer. chern. SOC.81, 3213
(1959).
[87] F. V . Brutcher, T. Roberts, S.J.Barr, and N.Pearson, J. Amer.
chem. Soc. 81, 4915 (1959).
[88] R. F. Bryan and J . D. Dunitz, Helv. chim. Acta 43, 3 (1960).
[89] C. Chiurdoglu, T . Doehard, and B.Tursch, Bull. SOC.chim.
France 1960, 1322.
O,
1005
chains at van der Waals distances in the crystalL771,
although details of the end folding and the location of
the carbonyl groups are not known. In medium rings
one would expect the carbonyl groups to occupy
positions where transannular hydrogen repulsion could
be relieved. In fact, the ultraviolet absorption of cyclodecane-l,2-dione indicates an anti-planar dicarbonyl
system 1901, which only fits the cyclodecane skeleton
when two “interior” hydrogen atoms are removed
(Fig. 7).
0
111
skeleton (Fig. 4). In fact, the 10-membered lactone is
the first of the series of saturated monolactones[99J to
be solid at room temperature (m.p. 32°C). The discontinuities in boiling point, dipole moment, and rate
of hydrolysis as function of ring size indicate that the
transition from the less stable cis-ester configuration is
completed in the 10-membered ring [1OOJ. The low
melting point (2 “C) [lo11 of the 12-membered monolactone is probably due to the considerable departure
from planarity (155-163 ” instead of 180 ”) in the antibonds. These bonds, where most of the Pitzer strain is
concentrated, are the only places where the anti-ester
group can be fitted into the molecule. In the 14-membered monolactone (m.p. 28 “C) [*011 there are three
possibilities of accommodating a strictly planar antiester group (Fig. 4).
The simplest ring systems containing two ester groups
are the dilactones (2) (Fig. 8) and, with one ester group
Fig. 7. Suggested conformation of cyclodecane-l,2-dione [90].
In view of the low torsional barrier in bonds adjacent
to carbonyl groups, eclipsed bonds occur more easily;
thus cyclobutanone is planar [911, cyclopentanone should
prefer the “half-chair” and not the “enve1ope’’ conformation [871, and even 3,3,5,5-tetramethylcyclopentane
1,2-dione is planar 1921. In cyclohexanone the chair form
should be of lowest enthalpy [86,931, having the single
bonds adjoining the C=O group in their preferred
conformation. However, the flexible boat form will be
less Pitzer-strained than in cyclohexane, and since its
entropy is favorable [941, the equilibrium may contain
the boat form[95]. Most of the established cases of
cyclohexane boat forms in non-condensed systems are
ketones 1951. Neither cis- nor trans-2,5-dimethylcyclohexane-l,4-dione occur in the chair form [961. Cyclohexane-1,Cdione itself has a substantial dipole moment
in solution[97J, and occurs as a distorted boat in the
solid [9*1.
In larger rings, however, the flexibility introduced with the
keto-group does not seem to manifest itself. When there is
more than one carbonyl group in the ring, dipole interaction
may stabilize or destabilize particular conformations [lo].
Lactones a n d Cyclic Polyesters
The planar anti- or trans-ester system can only be
accommodated in rings of ten and more members by
replacement of two C H 2 groups of the cycloalkane
[90] W . H . Urry, D . J.Trecker, and D . A . Winey, Tetrahedron
Letters 1962, 609.
[91] A . Bauder, F. Tank, and H . H. Giinthard, Helv. chirn.’Acta
46, 1453 (1963).
[92] L.C.G.Goaman and D.F.Grant,Tetrahedron 19,1531 (1963).
[93] W . D. Cotrerill and M. J. T. Robinson, Tetrahedron 20, 765,
777 (1964).
[94] N , L. Allinger and L . A . Freiberg, J. Arner. chem. SOC.82,
2393 (1960).
[951 J. Dale, J. chern. SOC.(London) 1965, 1028.
[96] R. D. Stolow and M. M. Bonaventura, Tetrahedron Letters
1964, 95.
[97] 0. Hassel and E. Naeshagen, Tidsskr. Kjerni, Bergves. Metallurgi 10, 81 (1930).
[98] P. Grorh and 0 . Hassel, Acta chern. scand. 18, 923 (1964).
1006
Fig. 8. Conformations of dilactones (top) and cyclic diesters (bottom).
Left: odd number of CHI groups between ester groups; right: even
number of CHZgroups [102].
turned around, the symmetrical diesters (3). In both
series, the ester groups should tend to form the bridges
between two parallel polymethylene chains at van der
Waals distance, as the necessary gauche bonds at the
corners are then obtained at no extra expense. When the
number of CH2 groups in each half is odd, there is no
strain, but when it is even, there must be skeletal strain
except if three-atom bridged open (hence unstable)
conformations are adopted. The alternation of melting
points (Fig. 9) of the dilactones[991 confirms this conclusion; the alternation may be accentuated in this case
by dipole interaction [lo]. In the series of symmetrical
diesters (3) [102J, the alternation has disappeared or
become inverted (Fig. 9); although it is expected[lol
that dipole interaction should in this case counteract
[99] M . Stoll and A . Rouve‘, Helv. chim. Acta 18, 1087 (1935).
[loo] R. Huisgen and H. Ort, Tetrahedron 6, 253 (1959).
[loll H . Hunsdiecker and H. Erlbach, Chem. Ber. 80,129 (1947).
[lo21 J. Dale, J. chern. SOC.(London) 1965, 72.
Angew. Chem. internat. Edit. J Vol. 5 (1966) J No. 1 2
n.3
N.10
L 5 6 7 8 9 10 11 12 13 11, 15 i 6 17
22
26
30
31
35
14
18
they form two series, n odd and n even, in each of which
there is a fixed increment of one unit cell dimension,
corresponding to the increasing length of straight polymethylene chains as the series is ascended. The crystal
structure of jacobine bromhydrin (4) 11041, which contains a 12-membered diester ring, confirms the similarity
with the parent hydrocarbon and the orientation of the
mg
Fig. 9. Melting points of
diesters [1021.
(0) dilactones
and ( 0 ) symmetrical cyclic
Abscissa: Upper row number of CH2 groups between ester groups in
one half of the molecule, n ; lower row, total number of ring
members, N .
the alternation, the effect should be noticeable only in
the lower members where the dipoles are close. The
dipole moments [I021 of the dilactones and symmetrical
141
Fig. 11. Crystal structure of jacobine bromhydrin [1041.
3
I-
0'.
n.0
N.
* -
"
1 2
rn
"
3
10
i,
planar anti-ester groups (Fig. ll), although at one
corner the ring is heavily strained by fusion with two
5-membered rings.
"
"
"
"
Lactams, Cyclic Amides, Peptides, a n d
Depsipeptides
"
5 6 7 8 9 10 11 12 13 14
18
22
26
30
14
Fig. 10. Dipole moments of (0)dilactones, ( 0 ) symmetric cyclic diesters,
and ( 0 )open-chain diesters H&O&-(CH~),-CO~C~HS [102].
Abscissa: Upper row, number of CH2 groups between two ester groups,
n; lower row, total number of ring members, N.
**
diesters (Fig. 10) are in agreement with the predictions
from the models (2) and (3). Also, X-ray results on the
dilactones[lo31 show that, apart from the lowest members (number of CH2 groups, n,five or six in each chain),
The introduction of one or more amide groups into a
cycloalkane should have the same conformational consequences as discussed for the ester group, but the intramolecular dipole interactions are expected to be more
pronounced. Intermolecular hydrogen bonding should
increase the heat of fusion and obscure subtle variations
in the conformational part of the entropy of melting
and, consequently, in the melting point alternation [lo].
The smallest of the simple monolactams which has a
//c
H
0
/5a/
H
/5c/
Fig. 12. Conformations of dilactams (top) and cyclic diamides (bottom) with odd (left) and even (middle and right) number of CHI groups
between the amide groups in each half of the molecule [l09].
~11041 J. Fridrichsons, A . M . Mathieson, and D. J . Sutor, Acta
[lo31 C. S. D. King (Brussels), unpublished.
Angew. Chem. internat. Edit. / Vol. 5 (1966)
crystallogr. 16, 1075 (1963).
No. 12
1007
high melting point [105,1061 is the nine-membered ring,
which is also the first to allow the stable anti- (or trans-)
amide configuration in the solid [1061; however, no
melting point alternation is observed [I051 in higher
monolactams. Dielectric constants 11071 and infrared
and ultraviolet spectra 11061 show that the ten-membered ring is the smallest one present exclusively with
an anti-amide group in solution, as was-also observed
for the anti-ester group.
The accentuated melting point alternation predicted 1101
for the series of macrocyclic dilactams of the type (54
(Sc), and not the open strain-free ones (Sd). The same
result was obtained [I091 from an infrared study of
association of cyclic diamides in dilute solution.Whereas the very stable 14-membered caprolactam dimer
showed conformational homogeneity at least up to its
melting point (348 “C), the 18-membered homologue
has also an unstable crystal modification with C=O
vicinal anti bonds [1131. This increasing flexibility with
ring size is likewise seen in the 26-membered dilactam,
which seems to contain not exclusively anti bonds
within the polymethylene chains [1131.
In the higher cyclic caprolactam oligomers the number
of anti bonds increases [1131, as is expected for stretched
double chains, and the same is true for the cyclic oligoamides of adipic acid and hexamethylenedianiire
(Nylon 6,6). The lowest member of this latter series,
the 14-membered diamide, has the C=O vicinal bonds
only in gauche orientation [1131. At 150 “C the N-vicinal
bonds are also gauche, but become increasingly anti at
lower temperatures, so that at very low temperatures
the only strain-free conformation, (6) (Fig. 14), is
I
I
n.3
N
LA5m
10
I
4
I
I
I
,
5
6
7
8
14
18
,
9
22
,
,
10
11
26
Fig. 13. Melting points of (e) dilactams [lo81 and (0) symmetrical
cyclic diamides [log]. The figure differs from that in [lo91 in that the
melting point of the highest dilactam (212°C) [1101 has been included,
and the melting point of the lowest diamide (233-234°C) [ I l l ] has
been corrected.
Abscissa: Upper row, number of CH2 groups between the amide groups
in each half of the ring, n ; lower row: total number of ring
members, N .
and (5c) (Fig. 12) has been confirmedl1osl (Fig. 13),
as has the attenuated alternation expected for the
symmetrical cyclic diamides, (56) and (5e), 11091 (Fig.13);
the difference between the two series disappears as the
amide dipoles become separated by longer saturated
chains. By studying the CHz-scissoring infrared bands
assigned 11121 to the gauche- and anti-conformations on
either side of the amide group, and by using polarized
infrared radiation, it was shownIll31 not only that the
14-, 18-, and 26-membered dilactams have gauche bonds
in both the N and C=O vicinal positions and, therefore, have conformations of the type ( 5 4 , but also that
this holds for the 12- and 16-membered dilactams, so
that these have the compact strained conformations
11051 L. Ruzicka, M . Kobelt, 0 . Hafiiger. and V . Prelog, Helv.
chim. Acta 32, 544 (1949).
11061 R. Huisgen, H . Brade, H . Walz, and I. Glogger, Chem. Ber.
90, 1437 (1957).
11071 R . Huisgen and H . WaL, Chem. Ber. 89, 2616 (1956).
[lo81 M . Rothe, Angew. Chem. 74, 725 (1962).
11091 J. Dale and R . Coulon, J. chem. SOC.(London) 1964, 182.
[110] H . Zahn, H. D . Stolper, and G. Heidemann, Chem. Ber. 98,
3251 (1965).
[ I l l ] G. I. Glover and H . Rapoport, J. Amer. chem. SOC.86,3397
(1964).
[112] G . Heidemann and H . Zahn, Makromol. Chem. 62, 123
(1963).
[I1 31 G. Heidemann et a / . (Aachen), unpublished.
1008
rti1
Fig. 14. Suggested low-temperature conformation of hexamethylene
adipamide.
adopted. The inherent well-known instability [I141 of
this isomer, compared with the caprolactam dimer of
the same size, seems to stem from the presence of two
extra non-vicinal gauche bonds and a less favorable
dipole orientation.
Melting point alternations in homologous series of
oligolactams have already been interpreted in detail [lo]
for the case of E-aminocaproic acid [1151 and w-aminoundecanoic acid [1161. Recently, data have become
available also for p-alanine 11081,y-aminobutyricacid [lo8],
and 8-aminovaleric acid 11171, and the following generalization can be made: In the general series
high conformational stability, hence high melting points
and high synthetic yields are expected when the number
of ring atoms, x(m + 2), is 14, 18, 22, 26, 30, etc. In
addition, it is advantageous when the gauche bonds
forming corners are not from the interior of the polymethylene chains, but those adjoining the amide groups,
and this is only possible when m is odd (both “bridges”
consist then of amide groups) or with m = 4 (one
“bridge” only is an amide group). It can be seen from
[I 141 D . Heikens, P . H. Hermans, and H . A . Veldhoven, Makromol.
Chem. 30, 154 (1959).
[115] H.Zahn and J.Kunde, Liebigs Ann. Chem.618,158 (1958).
11161 H. Zahn, H . Roedel, and J . Kunde, J. Polymer Sci. 36, 539
(1 959).
[117] M.Rothe and R.Hossbach, Makromol.Chem.70,150 (1964).
Angew. Chem. internat. Edit.
Vol. 5 (1966) No. 12
n
Table 1. Melting points (“C) of oligolactams (6a) and of cyclic peptides
( m = I ) . Number of ring atoms, N
(m
2 ) x . The underlining signianti-amide; - - - -, ris-amide).
fies conformational stability (--,
+
~
rn
-
1
3
2
I
4
H
5
X
-
I
24
2
283
(5)
(10)
~
3
243
4
255
5
6
7
(15)
(20)
H
(25)
295
(30)
Fig. 16. The hypothetical flat cyclohexaglycyl molecule (full line)
compared with the puckered cyclooctadecane conformation (broken
line).
(35)
8
(40)
9
(45)
10
(50)
m-g
0
(60)
(70)
Table 1 that t h e melting points of conforniationally
favored compounds are in fact higher thar. those of their
immediate neighbors.
The cyclic peptides (m = 1) decompose at or before the
melting point. The remarkable stability of cyclohexaglycyl[1181, and the general preferenceclo? 1191 for the formation,
sometimes by doubling reactions, of hexa- and deca-peptides.
thus finds a surprisingly simple explanation without the
usual assumptions of intramolecular hydrogen bonding and
dipole interaction. In the recently determined crystal structure of cyclohexaglycyl hemihydrate [1201 the pseudo-cell
contains two different centrosymmetrical molecules with
planar anti-amide groups (Fig. 15). In one molecule (a) there
proposed “191 conformations. The second molecule (b) has
no internal hydrogen bonds, but instead is hydrogen-bonded
with water molecules. It may perhaps be concluded that in
aqueous solution n o internal hydrogen bonds exist, in accord
with the finding [I211 that N-methylacetamide is not associated
in aqueous solution.
When cc-amino acid residues other than glycine occur
in cyclic peptides, size and orientation (D or L) of the
substituents must also be considered. Fig. 17, formula
Fig. 17. Idealized conformation of a cyclic hexapeptide with
and - - - -,respectively.
substituents marked
D
and
L
(7), shows the idealized conformation of a cyclic hexapeptide with D- and L-substituents indicated; in the
corners bulky substituents of any configuration can be
accommodated, but the two remaining substituents must
have opposite configuration if they are bulky. The frequent occurrence of D-amino acids in natural cyclic
polypeptides 11191 and depsipeptides [ 1221, in contrast
to their absence in open-chain polypeptides, and the
high synthetic yield obtained in the cyclization of oligopeptides with mixed configuration, may, in part, be
linked with this conformational feature.
Fig. IS. Crystal structure of cyclohexaglycyl hemibydrate [120].
(a) Conformation with two weak intramolecular hydrogen bridges;
(b) conformation without intramolecular hydrogen bridge. The two
regions of high electron density in the center are water molecules above
and below the ring.
are two weak internal hydrogen bonds, and the conformation
is intermediate between the “crowded”, flat, Pitzer-strained
molecule and the “rectangular” puckered cyclooctadecane
conformation (Fig. 16), and is quite different from other
[I181 C . H . Batnfordand F. J . Weymourh, J . Amer. chem. Soc. 77,
6368 (1955).
[119] R. Schwyzer, Record chem. Progr. 20, 147 (1959).
[I201 I. L. Karle and J. Karle, Acta crystallogr. 16, 969 (1963).
Angew. Chem. internut. Edit. / Vol. 5 (1966)
No. 12
The strong tendency for doubling in the cyclization of
tri- and pentapeptides with formation of cyclic hexaand decapeptides was explained by Schwyzev [1191 as
being due to a preceding hydrogen-bond association of
two molecules. However, a concerted reaction of both
ends of two peptide chains seems unlikely, and it has
already been shown [I211 that peptide association is
insignificant in water solution and weak in tetrahydrofuran and dioxane, while doubling occurs even in water
and methanol[l231. Recently, it has been demon[121] I . M . Klotr and J. S. Franzen, J. Amer. chern. SOC.82, 5241
(1960); 84, 3461 (1962).
[122] Yu. A.Ovchinnikov,V.T. lvanov, A . A. Kiryushkin, and M.M.
Shemyakin in G.T.Young: Peptides. Proc. 5. Europ. Syrnp. Oxford, Sept. 1962. Pergamon Press, London 1963, p. 207.
[123] Yu.A.Ovchinnikov, V.T.lvanov, A. A. Kiruyshkin, and M . M .
Shemyakin, Doklady Akad. Nauk SSSR 153, 122 (1963).
1009
stratedr123-1251 that tripeptides and tridepsipeptides,
which cannot associate lengthwise, nevertheless undergo
doubling on cyclization. Thus, glycyl-L-prolyl-glycine [124,1251 gives the cyclic hexapeptide as easily as
does diglycyl-L-proline11241, although the prolyl unit
carries no N H group; further, the tridepsipeptide glycylglycolyl-glycine, having an ester linkage instead of one
of the amide links [1251, and even N-methylated tridepsipeptides [1231, cyclize with formation of double molecules. Doubling (and tripling) occurs also with tetradepsipeptides 11231, N-methylated or not, but the yields
of tetra-, octa- and dodeca-depsipeptides are generally
lower and not significantly different [1231, presumably
because none of these compounds can have strain-free
conformations (Table 1). Shemyakin [I231 proposed on
this basis a reaction sequence with a dilution-dependent
final cyclization step.
The configurational sequence of the amino acids in
oligopeptides (and of hydroxy acids in depsipeptides)
has a profound influence on the cyclization yield. Thus,
Schwyzer has reported [I261 that while gramicidin S
itself, which contains two D-phenylalanyl residues,
was synthesized in good yield, the corresponding all+
cyclodecapeptide, with two L-phenylalanyl residues,
could not be obtained. Similarly, the cyclic pentapeptide
was obtained in 57 % yield from glycyl-L-leucyl-glycylD-leucyl-glycine as against 41 % for the L,L-isomer 11271,
and in 39 % yield from glycyl-D-leucyl-L-leucyl-glycylglycine compared with 12 % for the L,L-isomer [1281. The
cyclic hexapeptide, cyclophenylalanyl-glycyl-glycylphenylalanyl-glycyl-glycylwas mainly the meso-form,
whether obtained from racemic tripeptides or diastereomeric hexapeptides [I291 ;when synthesizedseparately 11291,
the yield of the meso-form (D,L)was higher than for the
optically active isomers (D,D and L,L). The same phenomenon was observed in the cyclization of glycyltyrosyl-glycyl-glycyl-benzylhistidyl-glycine
11301; when
both asymmetric centers had the L-configuration, the
yield of cyclic hexapeptide was 31 %, whereas it was
58 % (2/3 containing D-tyrosyl) when the tyrosine was
racemic. In these cases there is enough room for the
few bulky substituents at corner positions [cf. (7)], so
that the yield variation must be due to differences in
shape of the peptide chain before the final cyclization
step. Such differences are indicated by the observation 11281 that open-chain peptides containing both Land D-amino acids have a smaller dielectric increment,
hence a smaller distance between -NH3+ and -COO-,
than sterically homogeneous peptides. Although at[124] M. Rothe, K . D . Steffen, and I. Rothe, Angew. Chem. 75,
1206 (1963); Angew. Chem. internat. Edit. 3, 64 (1964).
[125] R . Schwyzer, J . P . Carridn, B. Gorup, H. Nolting, and
A . Tim-Kyi, Helv. chim. Acta 47,441 (1964).
[I261 R . Schwyrer in: Ciba Foundation Symposium on Amino
Acids and Peptides with Antimetabolic Activity. Churchill, London 1958, p. 171.
[127] G . W . Kenner, P . J.Thomson, and J. M.Turner, 3. chem.
SOC.(London) 1958, 4148.
[128] P.M. Hardy, G . W .Kenner, and R . C. Sheppard, Tetrahedron
19, 95 (1963); J. Beacham, V.T. Ivanov, G. W. Kenner, and R . C .
Sheppard, Chem. Commun. 1965, 386.
[129] R. Schwyzrr and A.Tun-Kyi, Helv.chim.Acta 45, 859 (1962).
[130] K . D . Kopple, R. R. Jarabak, and P . L. Bhatia, Biochemistry
2, 958
1010
(1963).
tempts have been made at a conformational analysis of
such peptide (and depsipeptide) chains [128,1311, these
are not convincing because of admitted arbitrary
assumptions, and lack of data for the conformational
preferences about the C(cr)-CO bond, and especially the
C(cr)-N and C(cc)-O bonds. The only conclusion that
seems certain is that the extended chain is favored by
steric homogeneity, and that each opposite configuration introduces a bend, which cannot be clearly specified. This is particularly well demonstrated by a series
of cyclization experiments [1311 with diastereomeric
tetradepsipeptides (N-methylvalyl-cr-hydroxyisovalerylN-methylvalyl-x-hydroxyisovalericacid); seven sterically heterogeneous isomers gave the cyclotetradepsipeptide in yields from 40 to 75 %, while the sterically
homogeneous isomer gave only 8 %, accompanied by
,Ei11818i
(81
Fig. 18. Proposed conformation of cyclic tetradepsipeptides, based on
the conformation of cyclododecane. D and L substituents are marked
and
-,respectively.
~
~
--
the doubling product (13 %). From Fig. 18, (8), it is
seen that the isopropyl side groups must all be at
“corners” and have enough room in any of these
isomers, so that the ring-strain in the end products can
hardly differ.
Cyclic Olefins and Annulenes
The introduction of a double bond into a cycloalkane
makes strain-free conformations impossible for any
ring size, and it has been pointed out [lo] that almost all
known cycloalkenes are liquids, reflecting their conformational instability. The exact stereochemistry of
the cycloalkenes is not known, except for derivatives of
cyclohexene and for cyclopentene [1321; the latter is not
planar, but has the 4-CH2 group bent out of the plane,
presumably to avoid totally eclipsed 3,4- and 4,5bonds, although some eclipsing in the 2,3- and 1,5bonds is thereby introduced. Dihedral angles in single
bonds adjoining the double bond have been calculated
from NMR coupling constants for cis-cycloalkenes up
to Clo and for trans-cyclodecene 11331; the considerable
deviations from the “correct” angles of 60 and 180 in
7 to 10 membered rings suggests that the departure
from ideality is concentrated in these bonds with their
lower torsional barrier. Since three-membered rings
approach the stereochemistry of the double bond, the
[131] Yu. A . Ovchinnikov, V.T. Ivanov, A . A . Kiryuskin, and
M.M.Shemyakin, Doklady Akad. Nauk SSSR 153, 1342 (1963).
[132] G. W . Rathjens, J. chem. Physics 36, 2401 (1962).
[133] G. V. Smith and H . Kriloff, J. Amer. chem. SOC.85, 2016
(1963).
Angew. Chem. internat. Edit.
1 Vol. 5 (1966) No.
12
crystal structures of cycloalkenimine derivatives [134,1351
[Fig. 19 and 20) may give an indication of the conformations of cis-cyclooctene and trans-cyclododecene,
Table 2. Heats of hydrogenation [1361 and thermodynamic data fof
cis,trans-equilibria [141] of cycloalkenes.
1
Cyclohexene
Cycloheptene
Cyclooctene
Cyclononene
Cyclodecene
4.4,7,7-Tetramethylcyclodecene
5, S,S,S-Tetramethylcyclodecene
Cycloundecene
Cyclododecene
32.2
26.5
24.0
25.4
27.1
25.9
23.0
23.6
20.7
22.0
-9.2
-2.9
--3.3
-3.4
AF
AH
(kcal/mole)
AS
(e.u.)
-4.04
-1.86
-2.9
-3.6
3.0
-4.7
0.1
-0.4
-1.5
-2.4
0.67
0.49
OH
Fig. 19. The crystal structure of cis-9,9-dimethyl-9-azoniabicyclo[6.1.0]nonane iodide [1341.
Fig. 20. The crystal structure of trans-13,13-dimethyl-l3-azoniabicyclo[lO.l.Oltridecane iodide [1351.
respectively [*I. In the latter case, trans-l3,13-dimethyl13-azoniabicyclo[lO.l.O]tridecane iodide (Fig.20) [1351, it
is interesting to note the strong resemblance of the 12membered ring to the conformation of cyclododecane
it self.
The heat of hydrogenation gives an idea of the relative
strain in a cycloalkene compared with that of the corresponding cycloalkane, and also of the relative stabilities
of cis- and trans-isomers. The data [I361 in Table 2 show
that in cyclooctene the trans-isomer is much more
strained than the cis-isomer and higher trans-olefins ;
this strain is also reflected in the notable dipole momentL1371 (0.82 D), as compared with that of transcyclodecene (0.15 D). The strain renders the molecule
rigid as is demonstrated by the resolution into optical
isomers [1381. Up to cyclodecene the cis-isomers have
lower enthalpies than the trans-isomers; however, in
[134] L. M . Trefonas and R . Majeste, Tetrahedron 19, 929 (1963).
[135] L. M.Trefonas and J. Couvillion, J. Amer. chem. SOC.85,
3184 (1963).
[*I Added in proof: That such an analogy does indeed hold has
now been demonstrated by the crystal structure of laurencin
( A . F. Cameron, K. K. Cheung, G . Ferguson, and J. M . Robertson,
Chem. Commun. 1965, 638). In this molecule a substituted
5-oxa-cis-cyclooctane ring has a conformation very close to that
shown in Fig. 19.
[136] R. B. Turner and W . R . Meador, J. Amer. chem. SOC. 79,
4133 (1957).
[137] N . L. Allinger, J. Amer. chem. SOC.79, 3443 (1957); 80,
1953 (1958).
[138] A . C , Cope, C . R . Ganellin, H . W.Johnson, T . V . Van Auken,
and H . J. S. Winkler, J. Amer. chem. SOC.85, 3276 (1963).
Angew. Chem. internat. Edit.
Vol. 5 (1966) j No. I 2
the symmetrical tetramethyl derivative (Table 2) this
situation is reversed [1391. An explanation on the basis
of the Dunitz conformation of cyclodecane, with all
the methyl groups at two corners, has been proposed [1391;
a trans double bond fits into the symmetrical position
better (although not at all well!) than a cis double bond,
which becomes highly strained (Table 3). This is also
borne out by the observation [I401 that the corresponding acetylene is reduced by sodium in liquid ammonia
to a cis-olefin, with the double bond shifted to the
neighboring position.
The enthalpy differences between cis- and trans-cycloalkenes are in good agreement with values obtained by
equilibration of medium size cycloalkenes (Table 2) [1411.
In the 11- and 12-membered rings the trans-isomer is preponderant, but surprisingly in the latter case this is not due
t o a lower enthalpy, but t o the more favorable entropy of
the trans-isomer. In larger rings the equilibrium is largely
o n the trans-side 11421.
When two double bonds are present,strain-free skeletons
become possible in particular cases, namely when the
two double bonds are diametrically placed, have the
same configuration, and are linked by odd-membered
polymethylene chains [lo] (Fig. 21). In the larger rings,
.II
I
I
\
/
' '
Ib)
,
\
Id)
Fig. 21. Conformations of frans,trans- and cis,cis-cyclotetradeca-l,Sdiene (a and b), of cis,cis-cyclodeca-l,6-diene( c ) , and of cyclohexa1,4-diene (d).
[139] P . Goebel, J. Sicher, M . Svoboda, and R . B.Turner, Proc.
chem. S O C . (London) 1964,237.
[140] M . Svoboda, J. Sicher, and J. Zuvada, Tetrahedron Letters
1964, 15.
[141] A . C . Cope, P . T . Moore, and W. R . Moore, J. Amer. chem.
S O C . 82, 1744 (1960).
[142] A . J. Hubert, unpublished.
1011
like the 14-membered ring, the trans,trans-isomer (a) is
favored because hydrogen interactions destabilize the
cis&-isomer (b). In the 10-membered ring, the trans,
trans-isomer of type (a) would have too formidable a
x-electron repulsion, which can be avoided in the cis&isonier‘(c); the 6-membered ring necessarily has a planar
conformation. As no absolutely perfect conformation
is possible for the conjugated cycloalkadienes (see below)
one might expect that these at least partially deconjugate
on equilibration to give preferentially diametric isomers
in the case of (4n 2)-membered rings.
+
The experimental facts support these deductions. The
melting points of diametrical cis,cis-cycloalkadienes
:j
f
-/o! \/
-
60
I I
N: 6
I
8
I
,
,
,
,
,
,
,
,
torsionally strain-free Conformation in Fig. 21 ,c has
been proposed for cis,cis-cyclodeca-l,6-diene11491 and
for the cis,cis-l,6-dioxocyclodeca-3,8-diene
[1501. The
1,5-isomers, on the other hand, must be strained, as
evidenced by the fact that both the trans,trans- [1511 and
cis,trans-l,5-isomers 11521 undergo valence isomerization to divinylcyclohexane. In cyclonona-l,5-diene even
the cis&-isomer undergoes this kind of valence isomerization 11531. Higher cyclic diolefins have been isomerized at 200 “C with triethylborane 11541. Extensive
deconjugation does take place; there remains only 10 %
of conjugated isomer in the 12-membered ring, 25 % in
the 13-membered, and in higher rings up to the 22-membered it never surpasses 50%. Among the non-conjugated isomers the expected preference for the diametrical isomer in 14- and 18-membered, and lack of
preference in 12- and 16-membered rings was confirmed 11541. The situation is complicated by the concurrent &,trans-isomerization. With strong bases at
room temperature the result is similar, but the preferences are more accentuated [1421. The much simpler
problem of cis,trans-isomerization without migration
of positional isomers has recently been studied [I441 at
room temperature. The results confirm (Table 3) the
10 12 14 16 18 20 22 ?L ?fi
Table
3.
Cis,trans-isomer equilibria of cycloalkadienes
Fig. 22. Melting points of the cis&-cycloalkadienes
C
C (cH~),-cH=cH-
( c H ~ ) C~H- = C H J
[143] J. Dale, A. J. Hubert, and G. S. D . King, J. chem. S O C .
(London) 1963, 73.
[144] J. Dale and C . Moussebois, J. chem. SOC.(London) [C]
1966, 264.
[145] B. Franrus, J. org. Chemistry 28, 2954 (1963).
[I461 H. Gerding and F. A. Haak, Recueil Trav. chim. Pays-Bas
68, 293 (1949); H.D.Stidham, Spectrochim. Acta 21, 23 (1965).
[147] D . Devaprabhakara, C . G . Cardenas, and P. D . Gardner,
J. Amer. chem. SOC.85, 1553 (1963).
[148] A . C . Cope, C . F. Howell, and A . Knowles, J. Amer. chem.
SOC. 84, 3190 (1962).
1012
~
~
z
)
n
-
~
~
=
~
~
-
(
brought
~ ~ 2 about
) n - ~ ~ = ~ ~
by U.V. irradiation in the presence of diphenyl disulfide
Abscissa: Upper row, total number of ring members, N ; lower row,
number of CHI groups between two vinrlene groups, n [1431.
(Fig. 22) show the required alternation11431, and the
corresponding trans,frans-isomers are alternately solids
and liquids [1441. Equilibration of cyclohexadiene 161,1451
gives 31 % of the 1,4-isomer, which shows that its
energy is comparable with that of the statistically
favored conjugated isomer. The vibrational spectra of
the 1,Cisomer indicate a planar structure 11461. Equilibration of cycloocta-l,5-diene gives almost quantitatively
the conjugated isomer 11471 since all cis,cis-isomers
must be Pitzer-strained; the cis,trans-l,5-isomer is also
strongly Baeyer-strained and so rigid that it preserves
its asymmetry even under pyrolytic conditions [1481. The
9-membered ring approaches the behavior expected for
the 10-membered ring; equilibration of cyclononadienes
thus gives 94 % of the most “diametrical” cis,cis-1,5isomer and only 6 % of the 1,3-isomer[1471, while
equilibration of cyclodecadienes gives almost quantitatively the “diametrical” cis,cis-l,6-isomer [1421, even
though this position is disfavored statistically. The rigid,
(
Ring atoms
(2n
4)
+
3
I 10
4
12
5
14
6
16
7
18
8
20
9
:22
I
[144].
Temp.
( “C)
25
75
25
1
I
75
25
75
25
75
25
75
25
75
25
75
-
8.3
92
85
29
38
71
68
42
47
66
57
3.8
8.0
30
35
6.4
12
58
51
22
27
48
45
31
37
96
92
70
57
2.0
2.8
13
I1
6.9
5.1
10
7.8
2.8
5.9
strong preference for the trans,trans-isomer (although
cispans is statistically favored) in the 14-, 18-, and
22-membered and the lack of preference in 16- and
20-membered rings. By comparing the equilibria at 25
and 75°C (Table 3) it also becomes clear that these
preferences are due to a low enthalpy in spite of an
unfavorable entropy, hence indicate rigid, stable conformations.
[149] C.A.Grob and P. W.Schiess, Helv. chim.Acta 47, 558 (1964).
[150] C . A. Grob and P. W. Schiess, Helv. chim. Acta 43, 1546
(1960).
[151] C . A. Grob, H. Link, and P . W. Schiess, Helv. chim. Acta
46, 483 (1963).
[152] P. Heimbach, Angew. Chem. 76,859 (1964); Angew. Chem.
internat. Edit. 3, 702 (1964).
[153] E. Vogel, W . Grimme, and E. DinnP, Angew. Chem. 75,
1103 (1963); Angew. Chem. internat. Edit. 2, 739 (1963).
[154] A.J. Hubert and J.Dale, J.chem. SOC.(London) 1963,4091.
Angew. Chem. internat. Edit. 1 VoI. 5 (1966)
/ No. 12
The macrocyclic conjugated cycloalkadienes have been
isolated from equilibrium mixtures of various positional
isomers [I551 and shown to have exclusively the cis,transconfiguration from the 12- to the 16-membered ring;
in the 18-membered and higher rings the trans,transisomer occurs also. Models show that a relatively strainfree conformation with a planar transoid cis,trans-diene
system first becomes possible in the 14-membered ring
(Fig. 23,a) ; experimentally, cyclotetradeca-l,3-diene is
all-trans is by far the highest melting[1573, may be a
reflection of its particular conformational stability. A
related structure with three trans double bonds in an
11-membered ring is present in humulene, and the
conformation found 11581 in its silver nitrate adduct,
Fig. 25. Crystal structure of humulene (all-trans-l , I ,4,8-tetramethylcycloundeca-3.7,lO-triene)[ I 581.
I
la1
(lo), (Fig. 25), has main features in common with the
12-membered ring, ( 9 ) ; most of its inherent strain is
located in that part of the ring where one CHz-group is
missing.
Among the cyclononatrienes the all-cis-l,4,7-isomer
seems to be the most stableL1591 and its conformation
has been proposed[1591 to be a rigid crown (Fig. 26),
I
mm
Ibl
Fig. 23. Suggested conformations for (a) cis,frans-cyclotetradeca-1,3diene and (b) trans,trans-cyclooctadeca-1,3-diene[155]. Atoms marked 0
are situated in the plane of the diene system, those marked 1 above,
those marked 2 further above, while atoms marked -1 lie below the plane.
found to be the first to have a full intensity ultraviolet
absorption [1551. Similarly, the 18-membered ring is the
smallest which can accommodate a planar transoid
transpans-diene system in an essentially strain-free
conformation (Fig. 23,b) [1551.
With three double bonds in a ring, “trigonal” ideal
conformations with symmetrically placed trans double
bonds, (9) (Fig. 24), become possible, but only for the
d
/9J
Fig. 24. Crystal structure of all-trans-cyclododeca-1,5,9-triene
[156].
first member, all-trans-cyclododeca-l,5,9-triene,has
such a strain-free conformation been established [I561 by
X-ray methods. Higher rings (C18, c24, etc.) would have
intolerably large holes in the interior. The fact that
among the known cyclododeca-1,5,9-trieneisomers, the
~~~
__
[155] A.J. Hubert and J. Dale, J.chem. SOC.(London) 1965, 6674.
[156] G. Allegra and I.W.Bassi, Atti Acad. naz. Lincei, Rend.
CI. Sci. fisiche, mat. natur. 38, 72 (1962).
Angew. Chem. internal. Edit. / Vol. 5 (1966) No. 12
rn
/
11601.
Fig. 26. Crystal structure of all-cis-cyclonona-l,4,7-triene
subsequently confirmed by X-ray methods [1601. The
single bonds must contain some Pitner strain and the
crystal structure [I601 reveals that some strain is also
introduced by transannular hydrogen repulsion ; this is
manifested in the heat of hydrogenation [1601. Equilibration of cyclooctatriene, on the other hand, shows that
the conjugated all-cis-l,3,5-isomer is by far more stable
than the 1,3,6-isomer [1611, but it is not known whether
the triene chromophore is planar. In cycloheptatriene
it has been shown by spectral methods that not only are
the protons of the CH2 group non-equivalent [1621, but
that the molecule has a boat conformation[1631 with a
non-planar triene system, (12) (Fig. 27), as was confirmed recently by electron diffraction [1641; X-ray
11 571 H . Breil, P. Heimbach, M . Kroner, H . Miiller, and G . Wilke,
Makromol. Chem. 69, 18 (1963).
[158] A . T . McPhail, R. I. Reed, and G. A . Sim, Chem. and Ind.
1964, 976; J. A . Hartsuck and I . C. Paul, ibid. 1964, 977.
[159] P . Radlick and S. Winstein, J. Amer. chem. SOC.85, 344
(1963); K . G. Untch, ibid. 85, 345 (1963).
[160] W . R . Roth, W. B. Bang, P. Goebel, R. L. Sass, R. B. Turner, and A . P . Yu,J. Amer. chem. SOC.86, 3178 (1964).
[161] D . S . Glass, J . Zirner, and S. Winstein, Proc. chem. SOC.
(London) 1963, 276; W. R. Roth, Liebigs Ann. Chem. 671, 25
(1964).
[162] C. IaLau and H. deRuyfer, Spectrochim. Acta 19,1559 (1963).
[163] F. R. Jensen and L. A. Smith, J. Amer. chem. SOC.86, 956
(1964).
[164] M . Traetteberg, J. Amer. chem. SOC.86, 4265 (1964).
1013
diffraction on a derivative has led[1651 to the same
result. The dihedral angle between planes of adjacent
double bonds, 54 O, calculated from electron diffraction
data [1641, is so close to 60 that it must be concluded
rn
Fig. 27.
Conformation of cycloheptatriene i n the gas phase
11641.
macrolides [I681 containing an all-trans triene in a 19membered ring and an all-trans tetraene in a 24-membered ring.
In rings consisting of a closed loop of conjugated double
bonds, like benzene and the annulenes, the main interest
has centered on problems of aromaticity [1731, or double
bond localization versus delocalization (alternating or
identical bond lengths). In addition to electronic factors,
formulated in the well-known (and sometimes abused)
Hiickel rule, and steric factors, such as the possibility
of having a planar structure with correct bond angles,
conformational factors may also play a role. Thus in
benzene, as well as in the flat[l73,174l molecule of
[18]annulene, (14) (Fig. 28), the localized structure
that the 2,3 and 4,5 single bonds prefer gauchoid to
planar cisoid conformations. That this is not caused by
angle strain is shown by the fact that the ring is strictly
planarc1661 as soon as the double bonds become delocalized (e.g., 6 x-electrons on 7 atoms in the cycloheptatrienylium cation).
The cycloalkatetraenes of general formula
H
H
H
H
(14)
mimij
Fig.
28. Hypothetical planar localized double bond structures of
benzene and [18lannulene.
with two cis@-diene chromophores placed diametrically, show again melting point alternation with the
number, n, of methylene groups in the connecting
saturated chains [1551; the more rigid conformations of
the high-melting members with odd n are also manifested by the sharper vibrational structure of the ultraviolet absorption bands.
Longer conjugated double bond systems, other than
the annulenes (see below), have only been encountered
as part of macrocyclic lactones, both natural (macrolide
antibiotics [I679 and synthetic [1681. If one extrapolates
from the model of Fig. 23,b, one is led to expect that an
extended all-trans triene requires a 22-membered, a
tetraene a 26-membered, and a pentaene a 30-membered ring in order to be “comfortable”. It is interesting
that all macrolide polyene antibiotics have ultraviolet
spectra indicating an all-trans all-anti structure [1671. In
only a few cases, however, is the ring size known; thus
pimaricin 11691 has a tetraene in a 26-membered ring,
and lagosin 11701, filipin 11711, and fungichromin [I721 have
pentaene chromophores in 28-membered rings. The
latter must therefore have a certain tension. Similar
tensions must exist also in simple synthetic polyene
[165] R . E. Davis and A . Tulinsky, Tetrahedron Letters 1962, 839.
[166] W. G. Fafeley and E. R . Lippincott, J. Amer. chem. SOC.77,
249 (1955).
[167] W. Oroshnik and A . D . Mebane, Fortschr. Chem. org. Naturstoffe 21, 17 (1963).
[168] L. D . Bergelson and Yu. G. Molotkovsky, Izvest. Akad.
Nauk SSSR, otdel khim. Nauk 963, 105.
[169] 0. Ceder, Acta chem. scand. 18, 726 (1964).
[170] M . L. Dhar, V.Thn/ler, and M . C . Whiting, J. chem. SOC.
(London) 1964, 842.
[171] 0 . Ceder and R . Ryhage, Acta chem. scand. 18,558 (1964).
[172] A . C. Cope, R . K . Bly, E. P . Burrows, 0 . J. Ceder, E. Ciganek, B. T. Gillis, R . F. Porter, and H . E. Johnson, J. Amer. chem.
SOC.84, 2170 (1962).
1014
(Kekule structure) would have three eclipsed planar
cisoid single bonds ; since non-planar strain-free conformations with gauchoid single bonds are not possible
here, the above bonds can only be avoided by delocalization. This may constitute an additional driving force for
delocalization in the [4 n + 21 annulenes which also are
aromatic according to Hiickel. For cyclooctatetraene
([8]annulene), [16]-, and [24]annulene (15), (16), (17),
(Fig. 29), etc., strain-free non-planar conformations
with perfect anti and gauche single bonds linking
localized double bonds are possible, so that there is no
tendency for delocalization ; just these rings happen
I151
Fig. 29. Tub-shaped conformations of [8nlannulenes; (15) established
for gaseous cyclooctatetraene [1751; (16) suggested for [16lannulene
and (17) for [24lannulene.
[173] F. Sondheimer, Pure appl. Chem. 7, 363 (1963).
[174] J. Bregman, F. L . Hirshfeld, D . Rabinovich, and G . M . J.
Schmidt, Acta crystallogr. 19, 227 (1965).
Angew. Chem. internat. Edit.
1 Vol. 5
(1966)
No. I 2
also to be non-aromatic according to Hiickel. For cyclooctatetraene the structure (15) has been established by
electron diffraction”751, and from the data a dihedral
angle of 57” between adjacent double bonds can be
calculated, which is close to the perfect gauche conformation. The same dihedral angle is found also in the
crystal structure of the carboxylic acid derivative 11761.
That angle strain in the planar structure cannot be
decisive for the non-planarity is indicated by the fact
that when the electronic situation requires delocalization of the double bonds, as in the radical anion[1771
and the dianion [177,1781,the ring becomes planar. X-ray
dataL1791 show that [14]annulene does not have alternating bonds and that the carbon atoms are arranged
in such a way that overcrowding in the interior is
relieved, mainly by a serious in-plane distortion of the
pyrene-like skeleton (Fig. 30). The relatively high melting point11801, 136”C, indicates that this may be a
, , , . o
1
28
Fig. 30. Crystal structure of [141annulene
11791.
unique conformation, and the NMR spectrum below
-40°C shows the presence of a ring current[lsl],
generally taken as a sign of aromaticity 11731. The coalescence of all N M R signals to one line at 20 “C indicates
that the ring is nevertheless flexible enough to interconvert “inner” and “outer” hydrogen atoms [*811.
For [lSIannulene the X-ray data [173,1741 indicate a
coronene-like, practically flat conformation (Fig. 31)
and again no alternation in bond length. In [18]annulene
trisulfide[1821 there seems to be no ring current; the
three cisoid single bonds are now part of “aromatic”
thiophene rings, and delocalization of the double bonds
over the large ring is no longer necessary.
[175] 0.Bastiansen, L. Hedberg, and K. Hedberg, J. chern. Physics
27, 1311 (1957).
[176]D . P. Shoemaker, H . Khdler, W. G. Sly, and R . C. Srivastava,
J. Arner. chern. SOC. 87,482 (1965).
[I771 T. J . Katz and H . L. Strauss, J . chem. Physics 32, 1873
(1960); T. J. Katz, J. Arner. chern. SOC.82, 3784, 3785 (1960).
[178] H. P. Fritz and H. Keller, Z. Naturforsch. 166, 231
(1961).
[1791 J. Bregman, Nature (London) 194, 679 (1962).
[180] F. Sondheimer and Y. Gaoni, 5 . Arner. chern. SOC.82, 5765
(1960).
[181] Y . Gaoni, A . Melera, F. Sondheimer, and R . Wolovsky,
Proc. chern. SOC. (London) 1964,397.
[182] G. M . Badger, J. A . Elix, and G. E. Lewis, Proc. chem. SOC.
(London) 1964,82.
Angew. Cliem. internat. Edit.
1 Vol. 5
(1966) J No. 12
Fig. 31. Crystal structure of [lXIannulene
[173,1741.
The melting point of [lSIannulene is over 230 “C ~ 8 3 1
and thus still higher than that of [14]annulene. A more
rigid conformation of [lSlannulene is suggested also
by the higher temperature, 110 “C, required for the
NMR signals of outer and inner protons, strongly split
by the ring current at room temperature, to coalesce to
one sharp line [1811. [16]Annulene [I739 1841 melts already
at 93 “C, and its single NMR line[173], indicating a
“flexible” structure, is at the position observed for conjugated polyenes. The non-planar conformation (16)
with alternating bonds is in accord with its properties,
and the mixing of cis and trans double bond protons in
the NMR spectrum can be explained by valence isomerization. For [20]annulene, not even non-planar conformations can be strain-free and accordingly this is a
liquid 11851. The NMR spectrum of [24]annulene [I861 is
similar 11731 to that of [16]annulene, indicating conjugated localized double bonds. The structure (17)
seems to fit the ultraviolet spectrum, the chromophore
being considered as two weakly coupled planar halves.
Cyclic Acetylenes
One triple bond alone introduces the same conformational disorder into an even-membered cycloalkane as
does one double bond, and it has been pointed out [lo]
that all known cycloalkynes are liquid. The requirement
of the acetylenic grouping that four carbon atoms lie on
a straight line evidently leads to a rapid increase of
tension with diminishing ring size. Thus in the equilibrium at 100°C between a cyclic acetylene and the
corresponding cyclic allene (with only three collinear
atoms) the percentage of acetylene changes from 74 %
for the 11-membered, to 35 % for the 10-membered,
and to only 7 % for the 9-membered ring 11871.
[183] F. Sondheimer, R. Wolovsky, and Y.Amiel, J. Arner. chem.
SOC. 84,274 (1962).
[184] F. Sondheimer and Y. Gaoni, J. Arner. chem. SOC. 83,4863
(1961).
[185] F. Sondheimer and Y. Gaoni, J. Amer. chern. SOC. 83, 1259
(1961); 84, 3520 (1962).
[186]L. M . Jackman, F. Sondheimer, Y.Amiel, D . A . Ben-Efraim,
Y . Gaoni, R. Wolovsky, and A . A . Bothner-By, J. Amer. chem. SOC.
84, 4307 (1962).
I1871 W.R.Mo0reandH.R. Ward,J.Arner.chern.Soc.85,86(1963).
1015
Cyclooctyne has been isolated [1881, and the enormous
strain manifests itself in an explosive reaction with
phenyl azide; nevertheless, the triple-bond character is
essentially retained since the infrared absorption is at
4.53 p as in higher homologues. The transient existence
of lower-membered cycloalkynes (C7, c6, C5) has been
inferred from reaction products [1F91, but such intermediates may perhaps involve non-linear excited “acetylenes” having no triple bond character.
More subtle conformational details are revealed by the heats
of hydrogenation of some tetramethyl derivatives of cyclodecyne [1391. Again, a n explanation is possible o n the basis
of the Dunitz conformation of cyclodecane with gem-dimethyl groups at two corners; the symmetrically located
triple bond in 5,5,8,8-tetramethylcyclodecyneis more easily
accommodated and gives a smaller heat of hydrogenation
than the asymmetrically placed triple bond in the 4,4,7,7tetramethyl derivative.
Two acetylenic bonds diametrically placed in 14-, 18-,
22-membered rings etc. can give strain-free ring conformations, whereas in 12-, 16-, and 20-membered rings
this is not possible, and a strong melting-point alternation (Fig. 32, lower curve) is observedt1431. By application of the postulate that a minimum of gauche bonds
are accepted, conformations of the type (18) (Fig. 33)
were proposed [I431 for the former series and seemed to be
supported by unit cell measurements [1433. However, the
\
f19J
L
Fig. 33. Conformations of cyclotetradeca-1,S-diyneas proposed (18)
[I431 and as found in the crystal (19) [103].
solution. Recently, it has been shown 11901 that with
strong bases in dimethyl sulfoxide at 40°C the triple
bonds migrate (through intermediate allenes), and the
equilibrium composition (Table 4), approached from
Table
4. Isomerization of
cycloalkadiynes
with potassium t-buroxide in dimethyl sulfoxide at 40 “C [190].
Ring atoms
(m+n+4
200 180 -
160
13
%
14
m+n
%
1 ;:::
15
16
80-
18
a
%
f 60-
20
LO -
m+n
%
20 -
n=3
I 11
I
4+
5+7
46
6+ 8
29
7+ 9
38
5+
11
6+
31
I 37
[a] Statistical probability ‘/z.
0-
N = 10
6+61al
9
7+7[al
49
8+8!al
3+
0
4+ 7
16
5+6
80
m+n
m+n
%
m+n
3 i 6
9
4 i 6
9
5+5[al
91
%
-
e
4+ 5
91
m+n
~
1L0 -
u
Distribution of positional isomers at “equilibrium”
4
5
6
1L
7
18
8
9
22
10
11
26
different pure isomers, confirms that the diametrical
isomer is much preferred in 14- and 18-membered
rings, but is a minor constituent in 16- and 20-membered
I
\
-
Fig. 32. Melting points 11431 of cycloalkadiynes (x = 1) and
cycloalkatetraynes (x 2) of the type
C(CH~)”(c=c),-(CI<~),,-(c~c),
1
Abscissa: Upper row, number of CH2 groups between the acetylene
groups, n ; lower row, total number of ring members, N,
for x = 1.
complcte structure determination of cyclotetradeca-1,8diyne has revealed 11031 that the conformation (19)
(Fig. 33), with two unfavorable gauche bonds, is preferred in the crystalline state. Perhaps this conformation
packs better; type (18) may still be preponderant in
[I881 A . T. Blomquist and L . H. Liu, J. Amer. chern. SOC.75,2153
(1953).
[189] G.Wifrig ef nl., Chem. Bet. 94, 3260, 3276 (1961); 96, 329,
342 (1963).
1016
\
\ ,
‘\
-
-
121J
brh3i:
Fig. 34. Proposed conformations for the 1,E-isomer (20) and the
1.7-isomer (21) of cyclobexadecadiyne [1901.
[190] A. J.Hubert and J.DnIe, J.chern.Soc. (London) 1965,3118.
Angew. Chem. internat. Edit.
/ VoI. 5 (1966) / No. 12
rings; the odd-numbered rings (C13 and C15) resemble
the 14-membered ring. For the isomers expected for
16- and 20-membered rings, the conformation (20) was
a pviovi a good candidate since a strain-free skeleton is
possible, although it is not very compact and has two
extra gauche bonds ; the unexpected preference for a still
less symmetrical isomer can be explained by assuming it
to have the conformation (21) (Fig. 34) which, although
a little strained, is more compact and has only one extra
Raurlie bond L1901.
explains the instability and the anomalous ultraviolet
spectrum [1911.
The stereochemistry of the fully conjugated ethylenicacetylenic rings, the dehydroannulenes, has been elucidated by NMR methods[173,1861 and in a few cases
confirmed by X-ray methods; thus both 5,11,17-tridehydro[l8]annulene [I931 and 1,8-didehydro[l4]annulene[1941 have the planar structures (24) and (25)
(Fig. 36, 37).
H
Table 5. Synthetic yield, melting point, and equilibrium distribution
of cycloalkadiyne isomers 11901.
+
5 -1- 5
57
98
90
6+ 6
23
-3
10
c14, m
n
Yield in synthesis (%)
M.P. ("C)
% in equilibrium
Cv,, m -:- n
Yield in synthesis (%)
M.P. ( " C )
% in equilibrium
4+ 6
18
30
10
5+ I
10
28
44
3t 7
1
liq.
ca. 1
4+ 8
13
30
38
j
&
+
2 t 8
7
12
0
3 -I- 9
-
$ ) <H
[=I
H
H
9
Table 5 shows that there is a parallel between a high
melting point and preponderance of cycloalkadiynes in
the isomer equilibrium "901. The synthetic yield is seen
to deviate somewhat from this pattern, because it also
is dependent on the conformation of the open chain
immediately before cyclization.
Among the numerous known polyacetylenic cyclic compounds containing diacetylenic groupings [143,1731 the
conformation of only one series has been investigated,
namely, the symmetrical cycloalkatetraynes
(24)
Fig. 36. Crystal structure of 5,11,17-tridehydro[l8lannulen~[193].
..........................
............................
__
__
1251
55Ezj
Fig. 37. Crystal structure of 1,8-didehydro[l4lannulene11941.
Alternations in yield, stability, and melting point 11431
(Fig. 32) led to the proposal of conformations analogous to those for the corresponding cycloalkadiynes,
although these might have excessive interior holes. For
the lowest member n = 3, X-ray data [1911 indeed indicate
a chair-like shape (22) (Fig. 39, later [1921 also deduced
for 1,6-dioxacyclodeca-3,8-diyne,(23) ; the too short
transannular distance between the diacetylenic groups
1231
'0
An ortho-phenylene group in a ring compound should
have essentially the same conformational consequence
as a cis double bond,although the absence of a rotational
barrier in the adjacent single bonds and the bulk of the
aromatic ring may have some effect. Accordingly, most
known benzocycloalkenes are liquid [lo] or relatively
low-melting [1431. Too few diametrical dibenzocycloalkadienes are known to permit general conclusions ; in
the case of dibenzocycloocta-1,5-dienethe crystallo-
[A548.jJ
Fig. 35. Crystal structure of cyclotetradeca-1,3,8,IO-tetrayne
(22) 11911
and of 1,6-dioxacyclodeca-3,8-diyne (23) [192].
~~
Rings Containing Phenylene Groups
Fig. 38. Crystal structure of dibenzocycloocta-1,5-diene[I951
~
[191] F. Sondheimer, Y .Amid, and R . Wolovsky, J. Amer. chem.
SOC.79, 6263 (1957).
[192] F. Sondheimer, Y. Gaoni, and J . Bregman, Tetrahedron
Letters 1960, No. 26, p. 25.
Angew. Chem. internat. Edit. 1 Vol. 5 (1966) 1 No. I 2
11931 N . A. Bailey and R . Mason, Proc. chem. SOC.(L,ondon)
1964, 356.
[194] N . A. Bailey and R . Mason, Proc. chern. SOC.(London)
1963, 180.
1017
graphic data require a center of symmetry 11951 as in the
conformation (26) (Fig. 38). The crown conformation
proposed [I961 for cyclotriveratrylene, a derivative of
tribenzocyclonona-l,4,7-triene,is analogous to that
itself (11).
found [I601 for all-cis-cyclonona-l,4,7-triene
Crown conformations, with the benzene rings alternately
above and below a plane, have been found to be the
only ones possible with molecular models for evenmembered oligo-ortho-phenylenes [I971; with tetraphenylene such a conformation is directly derivable
from that of cyclooctatetraene, (15) [1751.
m-Phenylene groups should be able to replace without
strain three successive CH2 groups in the middle of a
straight chain of five or more CH2 groups, such as can
only occur in 14-membered and higher cycloalkanes.
Too few examples are known with one or two such
groups in a macrocyclic ring; only for [2,2]metacyclophane [I981 has the conformation been established by
X-ray methods and shown to have the centrosymmetrical anti-structure (27) (Fig. 39) and not syn; however,
short saturated chain in these lower members stretches
over and is in close contact with one face of the aromatic
ring, as shown by the shift of the NMR signals of the
CH2-groups by the ring current in [9]- and [lo]-paracyclophanes [205,2061 and, less, in [12]paracyclophane 12051,
and by transannular reactions on solvolysis of [9]paracyclophane txylates [2061. The rigidity of these structures is also demonstrated by the resolution [2071 of an
aromatic carboxylic acid derivative of [lO]paracyclophane into the optical antipodes.
With two p-phenylene groups placed diametrically, the
[n,n]paracyclophanes, an expected conformational stability for odd n, and instability for even n, is borne out
by the observed [2081 melting point alternation for n > 3
(Fig. 40) in complete analogy with the diametrical
300
~
250
4 200 \
-
Y
I
,150
~
E
100
-
50
-
Fig. 39. Crystal structure of [2,2]metacyclophane [19Sl.
this 10-membered ring is too rigidly strained (deformed
benzene rings) to illustrate any general principle. The
closest transannular distance between aromatic carbon
atoms is only 2.69 A, compared with 3.4 8, for graphite.
Similar centrosymmetric conformations have been
found for the 4,12-dimethyl derivative [I993 and the
analogous diaza derivative, di(pyridine-2,6-dimethylene) [2001.
The p-phenylene group is linked linearly like the acetylene and diacetylene groups. Again, the known rings
containing one such group, the [nlparacyclophanes, are
liquids as are the cycloalkynes [lo]. The smallest member
known, the [8]paracyclophane [201J, has its benzene ring
bent from the planar configuration, as evidenced by the
abnormal ultraviolet spectrum. The [9]paracyclophane [2021 and higher homologues 12031 have more
normal spectra; nevertheless, even [lo]- and [12]paracyclophane have slightly bent benzene rings (2041. The
[195] W . Baker, R. Banks, D . R. Lyon, and F. G . Mann, J. chern.
SOC. (London) 1945,27.
[196] A . S. Lindsey, Chern. and Ind. 1963, 823; H . Erdtman,
F. Haglid, and R. Ryhage, Acta chern. scand. 18, 1249 (1964).
[197] G. Wittig and G. Lehmann, Chern. Ber. 90, 875 (1957).
[198] C . J. Brown, J . chern. SOC. (London) 1953, 3278.
[199] A . W . Hanson, Acta crystallogr. 15, 956 (1962).
[200] W . Baker, K . M . Buggle, J. F. W . McOmie, and D . A . M .
Watkins, J . chern. SOC. (London) 1958, 3594.
[201] D . J. Cram and G. R. Knox, J. Arner. chern. SOC. 83,2204
(1961).
[202] D . J. Cram and M . F. Antar, J. Arner. chern. SOC. 80, 3109
(1958).
[203] D . J. G a i n , N . L. Allinger, and H . Steinberg, J. Arner.
chern. SOC. 76, 6132 (1954).
[204] N . L. Allinger, L . A . Freiberg, R . B. Hermann, and M . A .
Miller, J. Arner. Chern. SOC. 85, 1171 (1963).
1018
n = 2
3
1,
5
6
7
8
9
Fig. 40. Melting points of [n,nlparacyclophanes (lower curve) and of
cage compounds of the type (28). in which two benzene rings are
symmetrically linked via three methylene bridges (upper curve) [208].
Abscissa: Number of CHI groups between the benzene rings.
cycloalka-diynes and -tetraynes (Fig. 32). Although the
polymethylene chains do not need to have the maximal
number of anti-bonds (cf. the acetylenic counterparts
already discussed), this is nevertheless indicated to be
the case by the similarity with the melting point curve
of the multimacrocyclic cage analogues (28) (Fig. 40)
[2081, which have three equivalent polymethylene
bridges linking two benzene rings in 1,3,5-position,
since here no alternative strain-free conformation for
even n with extra gauche bonds is -possible. Further, in
NMR spectra only the y and 6 CH2 groups show an
abnormal chemical shift; when the chains are straight,
these are the only ones that can direct their protons in
between the two benzene rings. The crystal structure is
known only for the two lowest members of the [n,n]paracyclophanes (n = 2 and 3), but these are severely
strained, as evidenced by the anomalous ultraviolet and
infrared spectra [203,2091; the two aromatic rings avoid
[205] J. S. Waugh and R. W. Fessenden, J . Amer. chern. SOC.79,
846 (1957).
12061 D . J. Cram and M . Goldstein, J. Arner. chern. SOC. 85, 1063
(1963).
[207] A.T. Blomquist and B. H . Smith, J. Arner. chern. SOC. 82,
2073 (1960).
[208] A . J. Hubert and J. Dale, J. chern. SOC. (London) 1965,
3160, and unpublished work.
[209] D . J. Cram and H. Steinberg, J. Arner. chern. SOC. 73, 5691
(1951).
Angew. Chem. internat. Edit.
/ Vol. 5
(1966)
1 No. 12
coming too close by bending into a boat shape. In the
case of [2,2]paracyclophane the molecule has really no
freedom to choose, and the crystal structure, (29)
(Fig. 41) [2101, is as expected; it demonstrates, however,
mz
Fig.
/29J
41. Crystal structure of [2,2]paracyclophane [ZIOI
that the strain is well distributed over the whole molecule and that the distance between the two aromatic
planes, which is normally 3.4 8, (graphite), is here
3.09 8, in the middle and only 2.83 A between linked
atoms. The [3,3]paracyclophane has somewhat more
freedom, and the strong x-electron repulsion causes the
aromatic rings not only to bend but also to be displaced
from a centered position, (30) (Fig. 42) [2111. When the
m
Fig.
f301
42. Crystal structure of [3,3lparacyclophane 12111.
repulsion is diminished by complex formation on one
side [2121, the structure presumably falls back into the
centered position whereby bond angle strain is released;
this is assumedr2121 to be the driving force for the
exceptionally strong tendency of (30) for complex
formation with tetracyanoethylene.
A large number of other (unsymmetrical) paracyclophanes have been prepared by Cram et al., and their
physical properties studied, but information of conformational interest is limited to the distance between
the two benzene rings. In the case of [l,n]paracyclophanes the synthetically available member [2131 with
the shortest polymethylene chain linking the 4- and
#-positions of the diphenylmethane group is the one
with n = 7, but spectral anomalies[z131 appear also for
n = 8 and 9 and are noticeable even for n = 12, indicating
that the saturated chain draws the two benzene rings
closer together and even deforms them. The limiting
ring size for restricted rotation of the phenylene group
in paracyclophanes is indicated by the resolution ~ 1 4 1of
an aromatic carboxylic acid derivative of [3,4]paracyclophane and the failure to resolve[2141 the corresponding derivative of [4,4]paracyclophane.
[210] C. J. Brown, J. chem. SOC.(London) 1953, 3265.
[211] P . K. Gantzeland K . N . Trueblood, Acta crystallogr. 18,958
(1965).
[212] L . A. Singer and D . J . Cram, J. Amer. chem. SOC.85, 1080
(1963).
[213] D. J . Cram and M. F. Antar, J. Amer. chem. SOC.80, 3103
(1958).
[214] D.J . Cram, W . J. Wechrer, and R. W. Kierstead, J . Amer.
chem. SOC.80, 3126 (1958).
Angew. Chem. internat. Edit,
Vol. 5 (1966) J No. 12
Highly Heterogeneous Ring Systems
Ring systems composed of several different chemical
groups (aromatic nuclei, multiple bonds, carbonyl
groups, heteroatoms, etc.) have been extensively studied.
However, such chemical and spectroscopic studies have
been concerned mainly with the stereochemistry of
parts of the ring, for example the question of planarity
of chromophoric conjugated groups, or with transannular interactions of two such groups, or with the
dependence of synthetic yield and resolvability on ring
size, and no conclusions about the shape of the whole
ring as such have been drawn.
Some X-ray results of such compounds merit mention;
thus, the crystal structure of 1,2;7,8-dibenzocyclododeca-1,7-diene-3,5,9,11-tetrayne
12151 reveals clearly
curved diacetylene groups and well distributed angle
strain, and the diolefin of [2,2]paracyclophane has a
structure [2161 strikingly similar to that of the parent
compound (29) [2101. Of particular interest is the crystal
structure of rifamycin B [217,2181, in which an aliphatic
bridge containing a diene conjugated with an amide
group, an enol, and various substituents, spans over a
naphthalenic system. Here again, abnormal chemical
shifts of the NMR signals of two CH-CH3 groups
indicate that these are influenced by an aromatic ring
current. The crystal structure 12171 also reveals planar
anti-amide groups, a correct stereochemistry next to
double bonds, and a planar transoid cis,trans-diene.
The latter feature seems to be general with such rings;
thus, conjugated diene systems formed in the isomerization of macrocyclic hydrocarbons containing a
phenylene group and two triple bonds prefer the cis,
trans-configuration and the anti-conformation [2191.
Several porphins have been examined by X-ray methods;
perhaps unexpectedly, almost all of them show deviations
from planarity attributable t o steric hindrance. It has also
been concluded [2201 that these ring systems must be quite
flexible, since the triclinic [2201 and tetragonal[2211 crystal
forms of tetraphenylporphin have different conformations;
the copper salt also is non-planar 12221. Similarly, nickel
etioporphyrin I is non-planar [2231, while nickeletioporphyrin
I1 is flat [224J. The related phthalocyanines are planar [2251.
The crystal structure of cyclotetramethylene tetranitramine [226J shows this molecule to contain an 8-membered
puckered centrosymmetric ring, but this structure is too
unusual t o justify further discussion.
[215] W . K . Grant and J. C. Speakman, Proc. chem. SOC.(London) 1959, 231.
[216] C. L. Coulter and K. N . Trueblood, Acta crystallogr. 16,
667 (1963).
[217] M . Brufani, W . Fedeli, G. Giacomello, and A . Vaciago, Experientia 20, 339 (1964).
[218] W . Oppolzer, V. Prelog, and P . Sensi. Experientia 20, 336
(1964).
[219] J.Dale and A.J. Hubert, J.chem.Soc. (London) 1963, 5475
I2201 S . Silvers and A.Tulinsky, J.Amer.chem.Soc. 86,927 (1964).
[221] J. L. Hoard, M . J. Hamor, and T. A. Hamor, J. Amer.
chem. SOC.85, 2334 (1963).
[222] E. B. Fleischer, J. Amer. chem. SOC. 85, 1353 (1963).
[223] E. B. Fleischer, J. Amer. chem. SOC.85, 146 (1963).
[224] M . B. Crute, Acta crystallogr. 12, 24 (1959).
[225] J. M . Robertson, J. chem. SOC.(London) 1936, 1195.
[226] P. F. Eiland and R . Pepinsky, Z . Kristallogr., Kristallgeometr., Kristallphysik, Kristallchem. 106, 273 (1955).
1019
Inorganic Rings
The term “inorganic ring” is limited here to compounds
whose molecular entity is preserved in solution and
which do not contain carbon as a ring member but
may contain organic groups as substituents. A wide
variety of such compounds is known 12271; interestingly,
there is a natural tendency not only for the formation
of 6-membered rings, but also a preference for 8-membered rings not paralleled in organic ring systems. Both
planar and chair-shaped 6-membered inorganic rings
are known 12271, and even chair-shaped isocyclic hexamers of sulfur 1241 and of diphenyltin 12281 have been
described.
Among the 8-membered rings, the crystal structure of
orthorhombic sulfur, Sg[2291, as well as of the azaderivatives heptasulfurimide S7NH ‘2301, hexasulfurdiimide(S3NH)z [2301 and tetrasulfurtetraimide (SNH)4 12311
reveal closely similar crown conformations with dihedral angles of 99 O, much wider than in open chains 1281.
On the other hand, when the nitrogen atoms carry no
hydrogen, as in sulfur nitride, (SN)4 [*321, the ring
“chooses” a saddle conformation with the nitrogen atoms
at the “corners” and dihedral angles close to 60 O, but
the valence description of this structure poses problems [2321. With a fluorine substituent on each sulfur
atom, as in tetrathiazyl fluoride, (NSF)4 [2331, the ring
assumes again a saddle conformation, but this time
with the sulfur atoms at the corners, and the SN bonds
are alternately long and short. With only tetrahedrally
linked ring atoms, as in cyclic oligomers of dimethyl(phosphino)borine, the situation is less confusing; thus
the trimer 12341 is a chair-shaped 6-membered ring and
the tetramer 12351 a saddle-shaped 8-membered ring with
phosphorus at the corners, suggesting normal staggered
single BP bonds.
Eight-membered rings containing alternating silicon and
oxygen (cyclosiloxanes), silicon and nitrogen (cyclosilazanes), or phosphorus and nitrogen (cyclophosphonitriles) as ring atoms, do not seem to prefer definite
conformations,in keeping with the extreme flexibility 12361
of such chains. Thus, the crystal structure of octamethyl
cyclotetrasiloxane [237lrevealsa centrosymmetric 8-membered chair, while the analogous crystalline octamethyl
~-
12271 H . Garcia-Fernandez, Bull. S O ~chim.
.
France 1961, 2453 ;
1963,416, 677.
[228] D. H . Olson and R.E.Rundle, Inorg. Chem. 2, 1310 (1963).
[229] S. C . Abrahams, Acta crystallogr. 8, 661 (1955).
[230] J. Weiss, Z . anorg. allg. Chem. 305, 190 (1960).
[231] E. W. Lund and S . R. Svendsen, Acta chem. scand. 11, 940
(1957); R. L. Sass and J.Donohue, Acta crystallogr. 11, 497
(1958).
[232] D. Clark, J. chem. SOC.(London) 1962,1615; B. D . Sharma
and J. Donohue, Acta crystallogr. 16, 891 (1963).
[233] G. A . Wiegers and A. Vos, Acta crystallogr. 14, 562 (1961).
12341 W. C . Hamilton, Acta crystallogr. 8, 199 (1955).
[235] P. Goldstein and R . A. Jacobson, J. Amer. chem. SOC.84,
2457 (1962).
[236] A . C . Chapman, N. L. Paddock, D . H. Palm, H.T. Searle, and
D . R . Smith, J. chem. SOC.(London) 1960, 3608.
[237] H. Steinfink, B. Post, and I . Fankuchen, Acta crystallogr. 8 ,
420 (1955).
1020
cyclotetrasilazane [2381 is a mixture of two conformations, one being identical with the above chair, the other
something between a saddle and a boat. Octamethylcyclotetraphosphonitrile ~ 3 9approaches
1
a perfect saddle
with the phosphorus atoms at the corners, as does the
analogous octakis(dimethy1amino)cyclotetraphosphonitrile 12401 [*I, while the tetrameric phosphonitrilic chloride 12411 has a more flattened ring intermediate between
a boat and the above type of saddle. The corresponding
fluoride 12421 finally is planar or close to planar; in this
latter case another crystalline form exists at low temperature 12421.
Since the above data suggest a certain tendency for these
ring bonds to be “staggered” and for bulky substituents
to prefer corner positions, it is of interest to examine
whether the conformations of larger rings can also be
derived from the diamond lattice[lol. It turns out that
three-atom bridges are required, not only in order to
have the substituted atoms (Si, P) at corner positions,
but also because the “hole” is needed to accommodate
substituents from those Si or P atoms which are not at
the corners. This type of conformation becomes first
possible for 16-membered rings (octamers), and it is
gratifying to find that this is exactly the conformation
Me
Me
Me
1311
Fig.
43.
Crystal structure of hexadecamethylcyclooctasiloxane[243].
established by X-ray methods 12431 for hexadecamethylcyclooctasiloxane (31) (Fig. 43). Analogous conformations are possible for 20-, 24-, etc. membered rings, but
not for 14-,18-, 22-, etc. membered ones, so that a melting point alternation would be expected. In the series
of cyclic phosphonitrile fluorides 12361 (Fig. 44) such
an alternation is observed, but the alternation continues
also to the lower members both in this series and for the
cyclic phosphonitrilic chlorides 12441 and the cyclic
~~
-
(2381 G. S. Smith and L. E. Alexander, Acta crystallogr. 16, 1015
(1963).
[239] M . W. Dougill, J. chem. SOC.(London) 1961, 5471.
[240] G . J. Bullen, Proc. chem. SOC.(London) 1960, 425.
[*I Added in proof: Octamethoxycyclotetraphosphonitrile is even
closer to the ideal saddle shape (G. B. Ansel/ and G. J. Bullen,
Chem. Commun. 1966, 430).
[241] R. Hazekamp, T . Migchelsen, and A. Vos, Acta crystallogr.
15, 539 (1962).
[242] H . M . McGeachin and F. R.Tromans, J. chem. SOC.(London) 1961, 4777.
[243] L. K. Frevel and M . J. Hunter, J. Amer. chem. SOC.67,2275
(1945).
[244] L. G. Lund, N. L. Paddock, J . E. Proctor, and H. T . Searle,
5. chem. SOC.(London) 1960, 2542.
Angew. Chem. internat. Edit.
/ Vol. 5 (1966) / No.
12
120
100
~
-
80 -
1
u
20e
0-
Q
x
-20
-
-40
-
-60
-80 -
.1
N.6
X=3
8
6
10
5
12
6
lL
7
16
8
18
9
v
20
10
22
11
Fig. 44. Melting points of (0)cyclic phosphonitrile chlorides r2441, of
cyclic phosphonitrile fluorides [236], and of (A) cyclic dimethylsiloxanes [245].
interaction. The striking similarity between the melting
point curves in Fig. 44 may be a suggestion of isoelectronic structures, that is, tetrahedral positively charged
phosphorus and “tetrahedral” negatively charged nitrogen. The additional ionic bonding might explain
the shortening of the otherwise identical P-N
bonds without hindrance to the almost “free” rotation
which is suggested by the extreme flexibility [236,2461 of
these compounds. The bonding in cyclic phosphonitrile
halides has for a long time remained controversial; the
suggestion of electron delocalization and aromatic
character 12471 has been vigorously contested [248,*491,
one of the main counter-arguments being constancy [236,244,2491 of the far-ultraviolet absorption
along each of these series.
Received: April 20th, 1965; revised August 17th, 1966
(0)
Abscissa: Upper row, total number of ring members, N ; lower row,
number of monomer units, X .
dimethylsiloxanes
(Fig. 44) [*I This is somewhat
surprising, since an analogous conformation for the
12-membered ring would have a severe transannular
[?45] W . Patnode and D. F. Wilcock, J. Amer. chem. SOC.68, 358
(1946).
[*I Added in proof: Alternations of other properties of phosphonitrile derivatives with ring size have recently been observed,
e.g., complex formation with hexamethylbenzene (S. K. Das,
[A 548 IE]
German version: Angew. Chem. 78, 1070 (1966)
R. A . Shuw, B. C. Smith, and C. P.Thakur, Chem. Commun.
1966, 33) and basicity (C. E. Brion, D. J. Oldfield, and N. L.
Paddock, Chem. Commun. 1966, 226).
[246] A. C. Chapman and N. L . Paddock, J. chem. SOC.(London)
1962, 635.
[247] D. P. Craig and N.L.Puddock, Nature (London) 181, 1052
(1958); D.P.Craig, M.L.Hefferman, R.Mason, and N.L.Paddock,
J. chem. SOC.(London) 1961, 1376.
[248] M . J. S. Dewar, E. A. C. Lucken, and M . A. Whitehead,
J. chem. SOC.(London) 1960, 2423.
12491 B. Lakatos, A . Hess, S. Holly, and G . Horvdth, Naturwissenschaften 49, 493 (1962).
Inorganic, Organometallic, and Organic Analogues of Carbenes
BY DR. 0. M. NEFEDOV AND (IN PART) DR. M. N. MANAKOV [*I
N. D. ZELINSKII INSTITUTE OF ORGANIC CHEMISTRY OF THE
USSR ACADEMY OF SCIENCES, MOSCOW (USSR)
This review deals with inorganic, organometallic, and organic compounds, as well as with
elements, that on the basis of their electronic structure and their reactions can be regarded
formally as analogues of carbenes. These “carbene analogues” include in particular the
compounds of monovalent boron, aluminum, nitrogen, and phosphorus; those of divalent
silicon, germanium, tin, and lead; atomic oxygen; atomic sulfur; and atomic selenium.
The preparation and chemical properties of the carbenes and their analogues are compared.
1. Introduction
The chemistry of carbenes has undergone a remarkable
development in recent years [1-61. The Unusual reactions
of these compounds are attributable to the fact that they
contain a carbon atom with a sextet Of electrons, two of
[*I
New address: D. I. Mendeleev Institute of Chemical Technology, Moscow.
[I] I. L. Knunyants, N. P. Gambaryan, and E. M. Rocklin, Usp.
Chim. 27, 1361 (1958).
[2] Ph. Miginiac, Bull. SOC.chim. France 1962, 2000.
[3] E. Chinoporos, Chem. Reviews 63, 235 (1963).
141 W . E. Parham andE. E.Schweirer,Org. Reactions13,55 (1963).
Angew. Chem. internat. Edit.
1 Vol. 5
(1966)
/ No. I 2
which do not participate in the bonding. Depending on
the preparation of the carbenes, these electrons may be
in the unpaired (triplet) or paired (singlet) state. Owing
to the unsaturation, most carbenes are electrophilic “1.
BY addition of two electrons to complete the octet, the
carbene carbon atom reverts to the tetravalent state,
with formation of one (sp2) or two (sp3) new bonds.
~
[5] W . Kirmse: Carbene Chemistry. Academic Press, New YorkLondon 1964.
[6] J. H i m : Divalent Carbon. Ronald Press, New York 1964.
[7] Some so-called nucleophilic carbenes appear to be exceptions; see, e . g . , H.-W . Wunzlick, Angew. Chem. 74, 129 (1962);
Angew. Chem. internat. Edit. I, 75 (1962); W . Kirmse, Liebigs
Ann. Chem. 666, 9 (1963).
1021
Документ
Категория
Без категории
Просмотров
1
Размер файла
2 121 Кб
Теги
many, conformational, aspects, membered, ring
1/--страниц
Пожаловаться на содержимое документа