вход по аккаунту


Coordination Chemistry and Catalysis. Investigations on the Synthesis of Cyclooctatetraene by the Method of W. Reppe

код для вставкиСкачать
In addition to thermal isomerization, pyrolytic reactions
also make it possible to convert alkenes into compounds
(0.g. diols and triols) in which the functional groups are
bound to certain carbon atoms depending on the boron
heterocycles formed as intermediates.
The pyrolytic transformations of organoboranes are not
limited to compounds containing B-C and B-H bonds.
Well-defined B-N heterocycles can also be prepared
from organic compounds containing boron-nitrogen
bonds [60]. In the case of compounds containing oxygen
in addition to boron, the possibilities of preparing the
corresponding B - 0 heterocycles appear to be slight. I n
this case, boroxines are mostly formed instead of heterocycles containing boron and oxygen as well as carbon.
Received, September 25th and October 4th, 1963
[A 338/137 IF1
German version: Angew. Chem. 75, 1079 (1963)
[60] R . Koster and K . Iwasaki: New Compounds Containing
Boron-Nitrogen Bonds. International Symposium on BoronNitrogen Chemistry, Durham, N.C. (U.S.A.), April 1963. Preprints of Papers. p. 123.
Coordination Chemistry and Catalysis
Investigations on the Synthesis of Cyclooctatetraene by the Method of W. Reppe
The relationship between the structure and the catalytic activity of nickel(l1) complexes
in the synthesis of cyclooctatetraene by the method of W. Keppe is discussed. The cyclotetramerization o f acetylene takes place within labile Ni(II)-acetylene x-complexes.
Inhibition tests have made it probable that four molecules of acetylene are grouped
around the nickel ion in the transition state, in a configuration which favor$ the formation
of the eight-membered ring.
The synthesis of cyclooctatetraene discovered in 1940
by Reppe [I J belongs to the most fascinating reactions of
modern chemistry. The simple starting material acetylene is converted by means of certain nickel(I1) complexes into cyclooctatetraene, the unsaturated hydrocarbon which Willstutter [2] managed to prepare only in
small quantities from pseudo-pelletierine at the beginning of the century by an extremely tedious and laborious
sequence of reactions. Reppe's synthesis permits production of this compound on a technical scale. Since benzene is the main product of uncatalysed thermal polymerization of acetylene, it is particularly surprising that
largely selective cyclotetramerization takes place under
substantially milder conditions in the presence of nickel(11) complexes. Reppe [l] suggested that the formation of
the eight-membered ring occurs within labile nickelacetylene complexes. More detailed investigations on
the mechanism of this reaction were not reported
until very recently. The attempts made by various
authors to explain the reaction remained necessarily
speculative, since decisive new knowledge of the complex chemistry of nickel was only obtained in recent
years. The observation [3,4] that cyclooctatetraene was
[ l ] W. Reppe, 0. Schlichting, K . Klager, and T.Toepel, Liebigs
Ann. Chem. 560, 1 (1948).
[2] R. Willstutter et al., Ber. dtsch. chem. Ges. 44, 3428 (191 I),
and subsequent papers.
[3] G.N. Schruuzer, J. Amer. chem. Soc. 81, 5310 (1959).
[4] G. N. Schrauzer, Chern. Ber. 94, 1403 (1961).
Angew. Chem. internat. Edit. 1 Vol. 3 (1964)
1 No. 3
also formed in the presence of certain Ni(0) complexes,
e. g. bis(acrylonitrile)Ni(O), was the starting point of our
own investigations, first reports on which appeared in
1961 and 1962 [5,6]. Meanwhile, further data have been
obtained, SO that the present comprehensive report contains hitherto unpublished material.
I. Conditions and Scope of Reaction
The formation of cyclooctatetraene occurs between 80
and 120 "C (optimum at 85-95 "C) at an acetylene pressure of 15-25 atm. The Ni(I1) complexes most commonly
used as catalysts are the acetylacetonate, cyanide, salicylaldehydate and the enolates of acetoacetic esters [ 1,7].
Cyclooctatetraene is obtained in 70 % yield (calculated
on the quantity of acetylene used) together with varying
amounts of benzene (up to 15 %), a resinous polymer,
and a dark brown, insoluble, cuprene-like mass ("niprene") [I]. Small quantities of styrene,l-phenyl- 1,3butadiene, vinylcyclooctatetraene [1,8,9], and traces of
azulene are also formed.
[ 5 ] G . N. Schrauzer, Angew. Chem. 73, 546 (1961).
[6] G. N . Schrauzer and S. Eichler, Chem. Ber. 95, 550
171 N . Hugihara, J. chem. Soc. Japan, pure Chem. Sect. 73, 323
(1952); Chem. Abstr. 47, 10490i (1953).
[8] L. E. Craig and C. Larrabee, J. Amer. chem. SOC.73, 1191
191 A . C . Cope and S. W. Fenfoi?,J. Amer. chem. SOC.73, 1195
The reaction has no induction period and the catalysts
need not be subjected to any contact development. All the
catalysts known up to the present are relatively shortlived. Of the side-reactions which deactivate the catalysts,
that leading to the formation of niprene is particularly
The reaction is essentially confined to unsubstituted
acetylene. Under the same conditions, substituted acetylenes always yield linear polymers or derivatives of benzene. However, a few mono- and disubstituted cyclooctatetraene derivatives have been produced catalytically by mixed cyclooligomerization of acetylene with
monosubstituted and disubstituted alkynes. Up to now,
only o n e substituted alkyne could be incorporated into
the eight-membered ring at a time [lo, 111, to give the
compounds ( l a ) to ( l e ) .
the paramagnetism is maintained; this status prevails
until the distance A‘ becomes so large that thermal promotion to a higher levcl is no longer possible.
d islor led
NI ion
planar lield
0, 1 0
3dYzr z
Scheme 1. Tetragonal distortion of octahedral complexes.
11. Properties of the Catalysts
All hitherto known catalytically active Ni(I1) complexes
have octahedral or planar structures. In octahedral complexes, the 3d level is split to give one triply degenerate
(t2,) and one doubly degenerate (e,) term. Since there
are eight 3d electrons available in the Ni(l1) ion, the eg
level contains only two electrons and the complexes are
therefore paramagnetic. The degeneracy of the levels
is destroyed by any distortion of the octahedral symmetry, or by unequal changes in the ligand strength
[I 21. Particular importance is attached to the cases where
the coordination octahedron is distorted axially or the
ligand strength is increased tetragonally on the octahedral base. In the extreme case. in which the axial ligands are completely removed, or where the tetragonal
ligand strength is greatly increased, the splitting of the
levels becomes that of a diamagnetic planar complex
(Scheme 1).
Examples of these two extreme types of complexes are
the paramagnetic bis(acetylacetonato)Ni(II) with its octahedral configuration in the solid state, and planar, diamagnetic bis(dimethylglyoximato)Ni(II).
When the
distortions of the octahedral symmetry are only small,
[lo] A . C. Cope and H . Campbell, J . Arner. chem. Sac. 73, 3536
(1951); 74, 179 (1952).
[ I l l A . C. Cope and D. S . Smith, J. Amer. chem. Soc. 74, 5136
(1952); A . C. Cope and D . F. Rugen, ibid. 75, 3215 (1953); A . C.
Cope and R . M . Pike, ibid. 75, 3220 (1953).
[ I21 For general information on Ni(I1)-complexes, see, for
example, the review articles: C.J . Ballhausen: Introduction to
Ligand Field Theory. McGraw-Hill, New York 1962, pp. 261 et
seq.; J. R. Miller, H . J . Enieleus, and A . G . Sharper Advances in
lnorganic Chemistry and Radiochemistry. Academic press, New
York 1962, Vol. 4, pp. 133 et seq.
A large number of Ni(I1) complexes lie in the immediate
vicinity of the singlet-triplet transition point; either of
the two states may then become the ground state, since
the energy difference between them is very small. Complexes of this type exhibit markedly anomalous magnetic
behavior. Particular examples of such compounds are
the Ni complexes of N-alkylsalicylaldimine. Initially, we
shall consider only complexes containing straight-chain
alkyl groups.
The N-methyl derivative (2) crystallizes in two diamagnetic low-temperature modifications, and forms a
buff colored, insoluble, and paramagnetic high-temperature modification (3), in which the coordination number of the nickel ion is probably 6 [13,14]. However, in
this and in other cases, the transition from the planar
diamagnetic form into the paramagnetic modification
takes place immediately on dissolution. The paramagnetism of the dissolved complex may be caused either by
solvation, association, or a combination of these two ef-
1131 C. M . Harris, S. E. Livingstone, and I . E . Reege, Austral.
J. Chem. If, 331 (1958).
[I41 L. Sacconi, P. Paoletti, and R . Cini, J. Amer. chern. SOC.80,
3583 (1958).
Angew. Chem. internat. Edit. Vol. 3 (1964) 1 No. 3
fects. With pyridine, for example, bis(N-methylsalicylaldiminato)Ni(II) forms a paramagnetic I :2-adduct ( 4 )
shown [ 18-20] that the analogous complexes with
branched hydrocarbon residues also form tetrahedral
structures. Thus, the bis-(N-cyclopenty1)-, bis-(N-cyclohexy1)-, and bis-(N-cyclooctylsalicylaldiminato)Ni(ll)
complexes are diamagnetic and planar in the solid state,
but more or less strongly paramagnetic in solution. In
this case, the spin decoupling is caused not only by association, but also by the rearrangement into the tetrahedral configuration. Although the formation of tetrahedral complexes of Ni(I1) is not energetically favored,
this configuration may still be realized if the ligands
carry bulky substituents. Thus, bis-(N-t-butylsalicylaldiminato)Ni(II) (51, exists almost completely in the
tetrahedral form both in solution and in the solid state
[18]. The tetrahedral bis-(N-alkylsa1icylaIdiminato)Ni
complexes undergo relatively rapid ligand exchange and
are also active catalysts for the formation of cyclooctatetraene, a property which will be described in greater
However, the same complex is also paramagnetic in noiicoordinating solvents, but the magnetic moment depends
on the concentration and the temperature. Molecularweight determinations indicate the presence of associated species in which the nickel ion achieves the axial
perturbation by interaction with the oxygen atoms of adjacent molecules [16]. The anomalous magnetic behavior was traced back to the dependence of the degree
of association on temperature, concentration, and solvent [17].
Tn the synthesis of cyclooctatetraene, the only complexes
that are catalytically active are those which are octahedral, or which can easily assume this configuration
(Table 1). Complexes with extremely strong tetrahedral
coordination are inactive, since in contrast to the octahedral complexes, they are largely inert with respect to
1 Weak ligand field
moderately rapid to rapid
very slow
Magnetic behavior
Ligand exchange
Catalytic activity
Ni .___
octahedral and planar
Ni-salicyl aldehyde
I Ni-acetylacetonate
ligand-exchange reactions. Ligand-exchange reactions
are still possible with planar nickel complexes with
moderately strong coordination, since the energy required for their conversion into the octahedral type of
complex is only small. The spin decoupling of
Ni(I1) complexes takes place with participation of polar
solvents [18] and leads to a weakening of the original
nickel-ligand bonds, A somewhat
representa?ion of a ligand-exchange is shown in Scheme 2.
111. Transition State and Rate-Determining Step
If the formation of cyclooctatetraene is to occur within
a nickel-acetylene complex, the original coordinate
nickel-ligand bonds must first be broken and new such
bonds must be formed with acetylene molecules. It was
therefore assumed [5,6] that the loosening of coordinate
metal-ligand links is the rate-determining step for both
Whereas N-alkylsalicylaldiminato qomplexes with
straight-chain alkyl groups may occur only in the octahedral and the planar configuration, it has recently been
[IS] H . C . Clark and A . L. Odell, J. chem. SOC. (London) 1955,
343 I .
[I61 R. H. Holm, J. Amer. chem. SOC.83,4683 (1961); J. P . Packler and F. A . Cotton, ibid. 82, 5005 (1960); 83, 2818 (1961).
[I71 R . H . Holm and T. M . McKinney, J . Amer. chem. SOC. 82,
5506 (1960).
[I81 L. Sacconi, P. Paoletti, and M. Ciampolini, J. Amer. chem.
SOC.85,411 (1963).
Angew. Chem. internat. Edit. 1 Vol. 3 (1964) 1 No. 3
chelate-forming agent
Scheme 2. Exchange mechanism between a monodentate and a
bidentate ligand.
unidentale ligand
the ligand-exchange reaction and the cyclooctatetraene
synthesis. This appears to be indeed the case and may
be illustrated by two examples: The paramagnetic and
kinetically labile bis(salicylaldehydato)Ni(II) is a good
catalyst for the formation of cyclooctatetraene, whereas
the bis(salicylaldiminato)Ni(II) is kinetically inert and
virtually ineffective as a catalyst.
[I91 L. Sacconi, P. L . Orioli, and M . Ciampolini, Proc. chem.
SDC.(London) 1962,255.
[20] R.H. Holm and K. Swaminathan, Inorg. Chem. 2,181 (1963).
As early as 1952, Hngihnrn 1211 suspected a connection
between the magnetic properties and the catalytic activity
of Ni(I1) complexes in the synthesis of cyclooctatetraene,but
an interpretation of his results has only recently become
Since the polar solvent participates in the ligand exchange reactions of Ni(I1) chelates (the reactions proceed
more rapidly in tetrahydrofuran, for example, than in
benzene). The formation of cyclooctatetraene also depends on the polarity of the medium [5,6]. The yields are
generally larger in dioxan than in tetrahydrofuran, and
are least in benzene. More strongly solvating solvents
than dioxan or tetrahydrofuran are unsuitable, since the
acetylene may no longer displace the solvent molecules
that are clustered around the Ni(I1) ion; thus, no cyclooctatetraene is formed in solvents such as pyridine,
benzonitrile, or water. Four coordination sites can be
made available in the nickel complexes by cleavage of
the coordinate linkages, and these can be occupied by
acetylene molecules. In the resultant, approximately
octahedral complex, these molecules can assume a steric
configuration which favors the formation of cyclooctatetraene (Scheme 3).
Fig. 1. Magnetic moments and catalyiic activities of bls-(N-alkylsallcylaldiminato)Ni(II) chelates for the synthesis of cyclooctatetraene in
various solvents [61, in relation to the size of the alkyl groups. Experimental conditions: Reaction temperature: 95 “C, acetylene pressure: 25 arm;
amount of catalyst: 5 mmoles (with benzene as solvent: 10 mmoles);
reaction time: 15 hours.
B : in benzene
C: in tetrahydrofuran
D: in dioxan
Upper ordinate: magnetic momect at 20 “C [Bohr magnetonsl
Lower ordinate: yield of cyclooctatetraene [gl
Scheme 3. Assumed transition state in the formation of
cyclooctatetraene on ”1)
The basic strengths of the lower straight-chain alkylamines alternate. Sacconi et al. [22] were able to show
that the magnetic moments of the bis-(N-alkylsalicylaldiminato)Ni(Jl) chelates alternate in parallel with the
basic strengths of the alkylamines. The yield of cyclooctatetraene from acetylene with these chelates as catalysts also varies with the effective ligand field strengths.
The N-methyl compound is the most strongly paramagnetic chelate of this series and gave the highest yields
[ 5 , 6 ] . This result shows that the same rate determining
step is involved in the formation of cyclooctatetraene
and in the apparently entirely unrelated ligand exchange
reaction. This step is most probably the reaction of the
catalyst with the first acetylene molecule, during which
the coordination number of the Ni(1I) ion may vary
from four to six, with the simultaneous conversion into
the paramagnetic form. This view is supported by the
results of recent measurements of the kinetics of the reaction between bis-(N-alkylsalicylaIdiniinato)Ni(II) and
dimethylglyoxime. The reaction, in which bis(dimethy1glyoximato)Ni(II) is the product, is of second order. The
rate is controlled by the addition of the first molecule of
dimethylglyoxime to the original Ni(1I) complex, which
would also explain why small differences in the effective
ligand strengths have such a pronounced effect on the
rate [23].
1211 N . Hagiliara, .I.chem. SOC. Japan, pure chem. Sect. 7Y, 323
(1952); Chem. Ahstr. 47, 10490i (1953).
[22] L. Sacconi, P. Paolrtti, and G . DeIre, J. Amer. chem. SOC.
79, 4062 (1957).
[23] G. N. Schranzer and P . Glockner, unpublished work.
Fig. 2. Relative rate of ligand exchange of dimethylglyoxime w:th
bis(N-alkylsalicylaldiminato)Ni(II) complexes in tetrahydrofuran [6J.
Ordinate: Relative rate OF exchange (N-methyl compound
IV. Synthesis of Benzene by Inhibition of the
Formation of Cyclooctatetraene
Since the rate of formation of cyclooctatetraene is controlled by a slow ligand-exchange process, no evidence
regarding the structure of the transition state can be
gained from kinetic measurements. If the transition state
is in Eact that represented by Scheme 3, the synthesis of
cyclooctatetraene should be effectively inhibited by electron donors. When the reaction is carried out with acetyIene and the usual catalyst in the presence of triphenylphosphine (molar ratio 1: 1 to the catalyst), no cyclooctatetraene is produced, but instead, benzene is obtained in high yield [5,6]. Benzene formation is not inAngew. Chem. internat. Edit. 1 Val. 3 (1964)
No. 3
hibited by the addition of even large amounts of triphenylphosphine. Triphenylphosphine is evidently capable of blocking a coordination site in the complex in
the manner represented by Scheme 4A. The interaction
of triphenylphosphine with the catalysts was also directly demonstrated by preparative methods [23]. The
reaction of triphenylphosphine with Ni(CN)2 in alcohol
yields a pale yellow, diamagnetic bis(tripheny1phosphine)nickel cyanide. The triphenylphosphine is only
loosely bonded in this compound, and is re-eliminated in
refluxing benzene. It is a powerful catalyst for the polymerization of acetylene, but produces only benzene. The
analogous complex of the more strongly basic diphenylethylphosphine is more stable and a much less effective
trinierization catalyst [23].
water per mole of catalyst are required to reduce the yield of
cyclooctatetraene to one-half, whereas only 2 moles of
pyridine are necessary to achieve the same result.
In the synthesis of benzene on triphenylphosphinepoisoned catalysts, the same process is rate-determining
as in the absence of the inhibitor. The dependence of the
yield of benzene on the magnetic moments of the bis(Nalkylsalicylaldiminato)Ni(Il) complexes is shown in
Figure 3.
Fig. 3. Depcndence of the yield of benzenz (in tetrahydrofuran at 95 "C
and a CzHz pressure of 25 atm) on the magnetic moments of the
bis-~Ar-alkylsalicylaIdiminato)Ni(II)chelates. Left-hand ordinate (for
curve A): yield of benzene [g] per 5 mmoles of catalyst in 15 hours.
Right hand ordinate (for curve B): Magnetic moments
[Bohr magnetonsl of the catalyst complexes in dioxan.
Scheme 4. Inhibition of the formation of cyclooctatetraene (A) by the
association of competing electron donors with the Ni(I1) ion (cf. Scheme
3). With bifunctional electron donors, even benzene is no longer formed
V. Influence of the Structure of the Catalyst
Inhibitors which block not only one, but two cis-positions in the active complex, e.g. ap'-bipyridyl, 1,lOphenanthroline, or 1,2-bis(diphenylphosphino)ethane
[(C6H5)*P-C2H4-P(C6H5)2], completely poison all catalysts. Even benzene is no longer formed, and no acetylene is consumed. The effect of basic additives increases
in the order:
(Dioxan) < trimethylamine < pyridine < triphenylphosphine < diphenylethylphosphine < tributylphosphine.
The order of this series permits certain conclusions to
be drawn regarding the bonding of the acetylene in the
active complex. Relatively weak o-bonds with only small
x-bond contributions are formed between the acetylene
and the Ni(I1) ion, since the formation of cyclooctatetraene is inhibited even by trimethylamine or pyridine, both
of which are o-donors but at most only weakly n-bonding ligands. Yamazaki and Hagzhara [24] recently established a similar series: dimethylaniline < pyridine <
Bi(C6Hd < Sb(C6Hd3 < AS(C6Hh < P ( C 6 H h
Under the conditions of the cyclooctatetraene synthesis,
the acetylene behaves towards the Ni(I1) ion as a ligand
with a weak field. The Ni(I1) ion should therefore remain paramagnetic in the transition state and should
exist in an approximately octahedral ligand field.
Water is also an inhibitor, but its effect is much weaker than
that of the weakly x-bonding pyridine. In tetrahydrofuran,
with bis(acetylacetonato)Ni(II) as catalyst, about 50 moles of
[24] H . Yamazaki and N . Hagihara, J. chem. SOC. Japan, ind.
Chem. Sect. 61, 21 (1958); Chem. Abstr. 53, 1 8 8 8 5 e (1959).
Angew. Chem. internat. Edit.
VoI. 3 (1964) 1 No. 3
The N-alkylsalicylaldimine-Ni(I1)
chelates with branched
alkyl residues, which exist in equilibrium with tetrahedral structures, are also active catalysts for the formation of cyclooctatetraene. The yields of cyclooctatetraene are again roughly proportional to the rate of
ligand exchange with dimethylglyoxinie. It is noteworthy
that benzene is produced in somewhat larger quantities
(up to 30 % relative to the yield of cyclooctatetraene)
than with the analogous catalysts containing straightchain alkyl groups. Evidently the bulkier branched substituents partially inhibit the formation of the transition
state required for the cyclotetramerization of acetylene.
The bis-(N,N'-dialkylaminotroponiniinato)Ni(Il) complexes (6) recently described contain tetrahedral chelates with stronger ligand fields than those of the Nalkylsalicylaldimines. Compound (6a) is diamagnetic
and planar; (6b) is paramagnetic and, again for steric
reasons, tetrahedral [ 2 5 ] . Both complexes undergo
ligand exchange only slowly in inert solvents, and react
with acetylene to yield only traces of cyclooctatetraene
[23]. Thus, the catalytic activity of Ni(I1) complexes for
the cyclization of acetylene is essentially a function of
the effective ligand-field strengths.
Complexes of type (7) with n = 2, 3 , 5 , and 6 are not at
all or only slightly active. For n = 4, a relatively large
quantity of benzene (about 40 7; relative to the yield
of cyclooctatetraene) is formed in addition to cyclooctatetraene. Additional chelate rings are presumably
formed between the nickel ion and the oxygen atoms of
the side-chains ; these may be exceptionally stable for
n = 4, and thus block one coordination site in the transition state much more frequently (cf. Fig. 4) [23].
reaction [26] (Scheme 5 ) . The formation of niprene is a
metal-catalysed polymerization of acetylene, brought
about either by the nickel catalyst itself or by its decomposition products (metallic nickel).
Scheme 5. Formation of the by-products styrene. phenylbutadiene. and
VII. Formation of the New Carbon-Carbon Bonds
Fig. 4. Yields of cyclooctatetraene (curve A) and benzene (curve B) with
catalysts of type (7), in tetrahydrofuran at 95 “C and a CzHr pressure
of 25 atm.
Ordinate: Yield [g] per 5 mmoles of catalyst in 15 hours.
Abscissa. Chain length [number of carbon atoms].
Complete inactivity results if the mobility of the ligands
around the Ni(I1) ion is restricted, as with chelates of
type (8) in which n = 6 . At n = 7 and n = 8, the com-
To explain the above results one has to assume that
cyclooctatetraene and benzene are formed within labile
acetylene-Ni(I1) x-complexes. I n the transition state the
acetylene molecules appear to be both activated and
drawn closely together. Activation of the acetylenic
carbon atoms presumably occurs with formation of the
n-complex, whereby the sp-hybridized carbon atoms
acquire alarger amount of p-character. This would facilitate their rehybridization to sp2, which gives the acetylene molecule the reactivity of a biradical. It is likely
that the formation of the products does not involve any
further intermediate, even if the new carbon-carbon
bonds are not all formed simultaneously.
VIII. The Cyclobutadiene Hypothesis
plexes again become slightly active and yield small
quantities of cyclooctatetraene [23].
VI. Formation of By-products
As shown by the inhibition experiments, the cyclooctatetraene synthesis can be easily “switched over” to yield
only benzene. It is therefore not surprising that small
quantities of benzene are almost always produced as a
by-product. The formation of styrene, phenylbutadiene,
and vinylcyclooctatetraene can be attributed to mixed
cyclooligomerization of acetylene with vinylacetylene
and higher linear oligomers of acetylene, which are
formed in small quantities under the conditions of the
[25] D. R . Eaton, W. D. Phillips, and D. J. Caldwell, J. Amer.
chem. SOC.85, 397 (1963). The author is indebted to Dr. R. E.
Benson, Wilmington, Del. (U.S.A.), for two samples o f these
Another aim of the present work was to search for some
indication of the intermediate formation of cyclobutadiene. No such indication was found; on the contrary,
some of the observations were incompatible with this
hypothesis. E. D . Bergmunn [27] was the first to suggest
the possibility that cyclooctatetraene is formed from
acetylene by dimerization of two molecules of cyclobutadiene. Kuri and Shidu [28] subsequently postulated the
formation of cyclobutadiene as an intermediate in the
photo-polymerization of acetylene, which yields benzene plus a small quantity of cyclooctatetraene. In their
famous paper on the possible existence of x-complexes
of cyclobutadiene, Longuet-Higgins and OrgeI [29] also
expressed the opinion that cyclooctatetraene is obtained
I261 E. C . Herrick and J. C . Sauer, U.S.-Patent 2661520; Chem.
Ahstr. 48, 6162f (1954).
[27] E. D. Bergmann: The Chemistry of Acetylene and Related
Compounds. Interscience, New York 1948, p. 93.
[28] Z. Kuri and S . Shida, Bull. chem. SOC.Japan 25, 116 (1952);
Chem. Zhl. 127, 10689 (1956).
[29] H . C . Longuet-Higgins and L. E. Orgel, J. chem. SOC.(London) 1956, 1969.
A n g e w . Chem. internat. Edit. 1 Vol. 3 (1964)
/ No. 3
from two molecules of cyclobutadiene that are formed
on the nickel catalyst. Criegee and Schroeder were able
to verify this hypothesis by thermolysis of tetramethylcyclobutadiene-NiC12 [30], but a temperature of 190 "C
was required. Although traces of cyclooctatetraene are
actually formed in the reaction of "cyclobutadiene-
1,l-dicyano- and 1,1,2-tricyanoethylene derivatives
which have meanwhile been isolated - [32] also give small
quantities of benzene and cyclooctatetraene. In these
complexes, the catalytically active particle is the zerovalent nickel atom, which rapidly loses its activity [4,33]
because it is difficultly resolvated by acetylene.
Scheme 6 . Reactions of bis(acrylonitrile)nickel(O).
mercury" with bis(acety1acetonato)nickel it-; is not
certain whether these really arise by dimerization of
cyclobutadiene [31].
As a consequence of its electronic structure, cyclobutadiene must be considered a strongly n-bonding ligand.
Intermediate cyclobutadiene-Ni(I1) complexes would
therefore be diamagnetic and kinetically more stable
than the parent catalyst molecules. The rate of cyclooctatetraene formation would then be controlled by the consecutive reactions of the cyclobutadiene complex, a fact
which cannot be reconciled with the experimental evidence.
The results of the inhibition experiments are likewise
incompatible with the cyclobutadiene hypothesis. There
is no logical explanation for the assumption that the
cyclobutadiene formed in catalytic quantities on the
nickel catalyst should dimerize selectively to cyclooctatetraene, despite the large excess of acetylene present,
yet in the presence of triphenylphosphine should yield
almost exclusively benzene. It must be concluded that
the formation of cyclooctatetraene in Reppe's synthesis
does not involve cyclobutadiene as an intermediate.
IX. The Formation of Cyclooctatetraene
on Ni(0) Complexes
Bis(acrylonitrile)nickel(O) is a coordinately unsaturated x-complex, and reacts with acetylene to form
heptatriene nitrile, benzene, and cyclooctatetraene [3,4]
(Scheme 6).
However, the yields are only approximately stoichiometric. On reaction with acetylene, the complexes of the
[30] R . Criegee and G. Schroder, Liebigs Ann. Chem. 623, 1
[31] M . Avram, E. Marica, J. Pogay, and C. D . Nenitrescu, Angew. Chem. 71, 616 (1959).
Angew. Chem. internat.
Edit. VoL 3 (1964) 1 No. 3
Electron donors such as triphenylphosphine stabilize
atomic nickel [4,33] and thereby catalyse the formation
of both heptatriene nitrile and benzene [4]. The yield of
benzene can be considerably increased in this way, but
the formation of cyclooctatetraene is completely suppressed. Trialkylphosphines also inhibit the formation of
benzene on Ni(0) catalysts, since with the latter they form
more stable complexes [34] than, for example, triphenylphosphine. Ni(0) complexes of the type Ni(PX3)4 (where
X = C1 or F) cyclize acetylenemonocarboxylic esters
mainly to benzene derivatives, but small quantities of
1,2,4,6- and 1,3,5,7-tetraalkoxycarbonylcyclooctatetraenes are also formed [35]. The yields of cyclooctatetraene and benzene in the reaction of acetylene with the
Ni(0) complexes of unsaturated nitriles, dinitriles, and
trinitriles were found to be independent of the polarity
of the solvent. They remain practically the same in benzene, tetrahydrofuran, and dioxan [34]. It can therefore
be concluded that in these reactions the catalytically
active species is the unchanged nickel atom.
The author is deeply indebted to Prof. W. Reppe for
his valuable advice and discussions, and to the Badische
Anilin- und Sodafabrik AG for the generous support of
this work. On behalf of his colleagues and himself, he
further wishes to thank Pro$ E. Wiberg for the help
given at the Miinchner Institut. Thanks are also due to the
Deutsche Forschungsgemeinschaft for financial support.
Received, May 27&, 1963
[A 342/147 IE]
German version: Angew. Chem. 76, 28 (1964)
[32] G. N . Schrairrer, S . Eichler, and D . A . Brown, Chem. Ber.
95,2755 (1962).
[33] G. Wifke, Angew. Chem. 75, 10 (1963); Angew. Chem. internat. Edit. 2, 105 (1963).
[34] G . N . Schrauzer and S . Eichler, unpublished work.
[35] J . R . Leto and M . F. Leto, J. Amer. chem. SOC.83, 2944
Без категории
Размер файла
654 Кб
chemistry, investigation, synthesis, coordination, catalysing, method, cyclooctatetraene, reppe
Пожаловаться на содержимое документа