close

Вход

Забыли?

вход по аккаунту

?

Mechanistic Aspects of Olefin Metathesis.

код для вставкиСкачать
Mechanistic Aspects of Olefin Metathesis
By N. Calderon, E. A. Ofstead, and W. A, Judy[*]
In this paper views regarding the nature of the transition state in olefin metathesis are
discussed. Earlier concepts that assumed a pairwise scrambling mechanism have been recently
challenged. Metathesis studies involving acyclic as well as cyclic olefins have been reported
which strongly suggest a chain process that may involve a carbene-to-metallocycle interconversion
as the pathway responsible for transalkylidenation. Catalytic systems that display a low apparent
metathesis activity on terminal olefins display a unique specificity in cross reactions with
internal olefins. Critical experiments indicate that terminal olefins rapidly undergo regenerative
metathesis preferentially. Results suggest that in the transient metallocycle, the most stringent
steric requirements occur at the j3-carbon, although the origin of this steric demand is not
obvious.
1. Introduction
The results of key experiments which extablished the site
of bond cleavage in olefin metathesis['-31 generated an extensive quest for a mechanism that could rationalize the various
aspects of a remarkable new catalytic reaction. Ideally, if
one is to obtain a coherent description of a catalytic process
on an atomic scale, it is necessary to establish: a) the structure
of the catalytic site, b) logistic aspects, i. e., pathways by which
incoming reactants replace outgoing products, and c) the
nature of the transition state and the intimate electronic transformations that occur along the reaction coordinate.
Characterization of catalyst sites in processes that are promoted by transition metals is not easily accomplished. Many
diverse catalysts promote olefin metathesis. These encompass
uni- and multicomponent systems derived from various metals
having different oxidation states, and may be homogeneous
or heterogeneous in nature. Hence, the assignment of a particular structure as being the "true" description of the catalytic
site is virtually impossible.
The relevance of ligand exchange processes to the logistics
aspect of olefin metathesis is clear. An outstanding feature
of certain metathesis catalysts is their capacity to promote
rapid reactions at very low concentrations. Extremely fast
rates are common, and these could not be realized if an
efficient route did not exist for the exchange of incoming
and outgoing olefin molecules on the active site of the metal.
Using rhodium complexes, C r ~ r n e r demonstrated
[~J
that ligand
exchange processes may proceed by either SN2 or SN1
mechanisms, depending primarily on the ligand field surrounding the metal and the availability of coordination sites.
Views regarding the nature of the transition state in olefin
metathesis are in a state of flux. Until recently, most proposed
schemes for the transalkylidenation process assumed a pairwise scrambling that required an initial bisolefin-metal entity
(1 ) bearing two olefinic ligands in a cis configuration about
M
M
the metal. Subsequently, a controversy developed over the
detailed nature of the transformation of this complex (1)
into the corresponding product complex (2) [Eq. (a)]"].
Thus, concepts such as quasicyclobutane['], tetramethylenemetal
and metallocycl~pentane[~]
were proposed
in attempts to describe the intervening transition state. The
fact that a bisolefin-metal complex of type (1 1 has never
been isolated for any active metathesis catalyst, lent all of
the aforementioned schemes a certain degree of vulnerability.
Lately, several publications disclosed results that appear to
be incompatible with the reaction pathway depicted in Eq.
(a). Rather, they point to a non-pairwise chain mechanism
that favors a single olefin molecule complexed to a carbenebearing metal undergoing interchange via a metallocyclobutane intermediate [Eq. (b)][**].
R'CH-M
R'CH=CHR
-
R'CH
M
RICH
C HR
/I
I1
The intent of this article is to present an up-to-date account
of recent developments toward understanding the mechanism
of olefin metathesis, and to compare the relative merits of
the two basic schemes. In addition, a discussion is included
on a recently discovered characteristic of catalysts which display a low apparent activity with terminal'olefins. These have
been found to react with mixtures of terminal and internal
olefins resulting in the selective formation of cross metathesis
products.
2. Vulnerabilities in Pairwise Scrambling Schemes
2.1. Polymerization of Cycloolefins
Do[goplosk[81suggested that the formation of a high molecular weight product during the initial stage of a cycloolefin
metathesis polymerization was inconsistent with a pairwise
scrambling mechanism. Indeed, high molecular weight
polymers are commonly observed during early stages of cy[*] Here and in the following equations the ligands to metal atom M external
to the olefin have in most cases been omitted.
['I
Dr. N. Calderon, E. A Ofstead, and W. A. Judy
The Goodyear Tire & Rubber Co., Research Division
Akron, Ohio 44316 (USA)
Anyew. Chem. I n r . Ed. Enyl.
J !.'?A
1 5 ( 1 9 7 6 ) No. 7
[**I For sake ofclarity, hereand in the following equations the carbene-metal
bond has usually been drawn as a double bond; for the character of this
bond see Section 4.
401
cloolefin polymerizations. For example, the polymer obtained
at 6% conversion from cyclooctene using a WCI,/C,H,OH/
C,H,AlCl catalyst had a molecular weight of about 20000&
300000~y1.
(This product was accompanied by low molecular
weight macrocyclic oligomers not included in the molecular
weight value.) This phenomenon requires that cycloolefin polymerizations propagate by some type of chain process rather
than by a stepwise growth process. Such a process is nevertheless reconcilable with a pairwise scheme when certain restrictions are invoked["? The pairwise mechanism [Eq. (a)] could
exhibit a chain reaction character if the rate of the transalkylidenation step were much faster than that of the exchange
W
I
i-
NCH
absence of acyclic olefins,propagation proceeds by macrocyclization as in Eq. (g)ll31.
W
W
W
RCH
CHR'
Linear, open chain macromolecules would be obtained by
cross metathesis with acyclic olefins as depicted in Eq. (h)
where the acyclic alkylidene moieties become chain ends:
W
Rcf
I1
RICH
\HRl
+ RCH=CHR
CHR'
step [Eq. (c)], and if in the propagation step the equilibrium
between (3) and ( 4 ) were to lie far on the side of the bidentate
complex [Eq. (d)].
On the other hand, if reaction (c) were faster than reaction
(a), a transfer-to-monomer process would predominate in the
early stages of cycloolefin metathesis, resulting initially in
a preponderance of low molecular weight cyclic oligomers
and essentially no high molecular weight product.
Chauuin et al., who were first to suggest non-pairwise transalkylidenation via a carbene-metal intermediate [see Eq. (b)]I1 'I,
recently examined the effect of conversion on molecular weight
distributions ofpolymers and copolymers obtained from cyclopentene and cyc1ooctene"'I. The monomer-polymer and
polymer-oligomer equilibria were presented as "back-biting''
processes. Hence, for cyclopentene the propagation-depropagation equilibrium involved an intramolecular carbene-metal
interchange as depicted in [Eq. (e)].
In a similar fashion, the formation of macrocyclic oligomers
in cyclooctene polymerization is illustration in Eq. (0.
Two significant points evolve from the non-pairwise scheme
illustrated here:
a) The high molecular weight polyalkenamers are linear,
even when acyclic olefins are absent in the reaction.
b) Cyclic oligomers are not formed via intermolecular condensation of two smaller rings.
In contrast to the above scheme, cycloolefin metathesis
via a pairwise interchange process predicts that, in the complete
402
Furthermore, this mechanism affords cyclic oligomers by
both inter- and intramolecular metathesis.
Studies to determine whether polyalkenamers are linear
or macrocyclic have been reported by
By adding
varied amounts of chain scission agents to a cycloolefin metathesis and examining their effect on product molecular weight,
it was possible to establish the acyclic nature of the high
molecular weight polyoctenamer. Figure 1 is a plot of I/D',
the reciprocal of the number-average degree of polymerization
(determined by vapor phase osmometry), versus T, the ratio
of reacted 3-hexeneto cyclooctene. The three lines are theoretical behavior curves for three different cases. Curve A is for
the random scission of a mixture of polymer chains which
is entirely acyclic, curve B is the theoretical expression for
the scission of a mixture of rings and chains, and curve C
the case of random scission of cyclic macromolecules. The
mathematical derivation of the respective expressions that
0.151
/
L
0
I
0.05
010
015
rFig. 1. Dependence of the reciprocal of the number-average degree of polymerization, D , o n the mole ratio of reacted 3-hexene to cyclooctene, I. Theoretical
curves are A: all linear chains; B: ring-chain equilibrium; and C: all cyclic
polymers.
Angrw. Chem. Int. Ed. Engl.
1 Vol. I S ( 1 9 7 6 ) N o . 7
produce these curves is fully described in ref.
The agreement between the experimental points and curve B is striking.
It lends considerable support to the assertion that, in the
absence ofany deliberately added acyclic olefin, polyoctenamer
as prepared under the conditions used in that study consisted
of high molecular weight linear chains (SO--SS%
by wt.),
while the remaining low molecular weight species were cyclic
oligomers having an average molecular weight of about 500.
Extending this work to polypentenamer, Scott“ examined
the effect of conversion on the weight-average and numberaverage molecular weights, as well as the molecular weight
distribution. This study revealed that polypentenamer was
linear in structure as was polyoctenamer.
The experimental verification that high molecular weight
polyalkenamers are linear and not macrocyclic has a limited
significance for the mechanism of olefin metathesis. It is compatible with the non-pairwise interchange concept, such as
the carbene-metal scheme proposed by Herisson and Chauuinl”’, but it does not exclude entirely the pairwise transalkylidenation scheme. It is recognized that the final, necessary
and sufficient condition for ring-chain equilibria, as observed114-161and described
is the presence of a trace
amount of chain scission reactant, typically an acyclic olefin.
Thus, the pairwise scrambling mechanism can be reconciled
with the formation of linear chains even when end-groups are
not introduced deliberately into the system, if one takes into
account (a)the possible in situ formation of chain ends during
the reaction of the organoaluminum catalyst component with
the tungsten salt; or (b) the presence of very small amounts
of olefinic impurities.
2.2. Formation of Cyclic Oligomers
Three basic features were established in a
related
to the formation of oligomers during cycloolefin metathesis:
1. The oligomers are macrocyclic.
2. The population of macrocyclics is controlled primarily
by the double bond frequency along the polymeric chain.
3. The oligomeric macrocyclics and their corresponding
high molecular weight polyalkenamer are interconvertible.
As stated earlier. whereas pairwise interchange schemes
allow formation of macrocyclic oligomers by two routes, intramolecular “pinching-off of a cyclic entity from a higher macromolecule and intermolecular condensation of two smaller
rings, the carbene-metal scheme provides only one route for
cyclic oligomer formation, namely, intramolecular “back-biting” [Eq. (e) and (f)]. By determining quantitatively the relative
abundance of the various macrocyclics produced during the
metathesis of 1,5-cyclooctadiene, it was demonstrated that
these are produced exclusively by an intramolecular process117. 181
1S-Cyclooctadiene is known to yield two series of macrocyclics. One contains “whole” multiples of the monomer
[(CsHl ,),I, the others may be described as “sesqui-oligomers”
[(C8H12)n-C4H6r where n>2]. Table 1 presents the relative
concentrations of the respective macrocycles-the “whole”
.. .) and the “sesqui-oligomers”
oligomers (C16rC24,C32
(C, 2, CZo,C28.. .). The results indicate that the molar ratios
of the various “sesqui-oligomers” relative to C16, as well as
the molar ratios of the “whole” oligomers to C16, remain
essentially constant throughout a conversion range of 1.3-toAngew. Chem. lnt. Ed. Engl. !Vol. I5 ( 1 9 7 6 ) N o . 7
98.5 %.(The only variant is the c12/c16ratio at 98.5 ”/, conversion, which has been rationalized” ‘1.)
To appreciate the significance of these results one must
recognize that, irrespective of the mechanism invoked, the
C1z oligomer( 1,5,9-~yclododecatriene)
cannot be produced by
any intermolecular process, since the smallest precursor available is a C8 ring. O n the other hand the c16 ohgomer (1,5,9,13cyclohexadecatetraene) might be produced, at least in principle, by inter- as well as intramolecular processes. The fact
that a significant amount of the C l z product was present
even at the lowest conversions measured, and that its relative
abundance remained constant throughout the course of polymerization is indicative that only one pathway for the formation of cyclic oligomers is operative, namely, an intramolecular
process. This argument applies also for the higher oligomers,
as clearly indicated in Table 1.
Table 1. Molar ratio of cyclic oligomers (C8HIz)”and sesquioligomers
(CsHtd-C,H,
to cyclohexadecatetraene ( C , , / C , , = 1.0) (after Ref. [l8]).
1.3
3.2
7.9
9.5
19.5
36.3
51.1
61.2
98.5
0.47
0.51
0.55
0.48
0.52
0.52
0.51
0.53
0.93
0.65
0.63
0.63
0.62
0.59
0.56
0.58
0.54
0.65
-
-
0.32
0.27
0.28
0.31
0.32
0.33
0.30
0.27
0.18
0.24
0.25
0.27
0.18
0.26
0.19
0.23
-
0.10
0.10
0.13
0.13
0.10
0.14
Although the data presented above are quite compatible
with a non-pairwise metathesis mechanism, they still d o not
entirely discount the pairwise reaction pathway. The data
can be accommodated with a pairwise mechanism by invoking
the following kinetic assumptions:
In Eq. (i), C8 denotes 1.5-cyclooctadiene monomer which
adds to a (C,),,, polymer yielding (C8),+, with a rate constant
for this step of k,. Since (C8), is structurally indentical to
polybutenamer, it may also be represented as (C4)2m.This
macromolecule at all times is reversibly interconverted to
the respective cyclic oligomer series (C4)”.If k, 4k,, conditions
result wherein oligomers are formed via the intramolecular
route exclusively.
2.3. Cross Metathesis of Acyclic Olefins with Cycloolefins
Cross metathesis studies between acyclic and cyclic olefins,
conducted for the purpose of examining telomer compositions, have been reported on different occasions[”. 7-221. Perhaps the most relevant work with respect to pairwise us.
non-pairwise mechanisms is the one disclosed recently by
K a t ~ [ ~The
~ ] telomer
.
distribution produced during the early
stages of the cross metathesis of cyclooctene with a mixture
of trans-2-butene and trans-2-octene, in the presence of a
molybdenum-based catalyst, was monitored. In addition to a
403
C,, the cross product of 2-butene and 4-octene, two symmetrical (Clz and C16)and one unsymmetrical (c14)dienes were
formed [Eq. (j)].
CHs-CH=CH-C3H7
C6
+
c
\CH=CH-C H~
CH-CH-CHs
ClZ
+C
CH=CH-CH,
CHZCH-C~H~
+
C
C H=C H-C 3H 7
C H =C H-C 3H7
1,7-Octadiene is a substrate which is uniquely suitable for
mechanistic studies. There is good evidence to indicate that:
a) 1,7-octadiene does not undergo intermolecular metathesis
according to Eq. (k), and b) the cyclohexene produced during
intramolecular cyclization does not participate in further
metathetic processes. These features were exploited by
G r u b b ~ [who
~ ~ ]studied
,
the metathesis of a mixture of 1,7-octadiene with [ 1,1,8,8-D4]-1,7-octadiene,
in an attempt to distinguish between a non-pairwise chain mechanism and the
pairwise scheme. The non-pairwise path is visualized as shown
in Eq. (m) and (n).
/(CH2)4,
/ (CH2)4,
CH
TCHz + SCHzH
/I
CHz
- I
W-CH
I
CH2-CHz
CH
I1
CHz
-
c16
c14
171
(In addition, one would expect formation of higher telomer
homologs. These were not discussed.) Katz's contention was
that, if any of the painvise scrambling schemes were to prevail,
the ratios of the unsymmetric to symmetric dienes, c14/c12
and C14/C16, would be zero initially. The experiments revealed
that the C14/C12 and c14/c16 ratios were significantly different
from zero at the initial stage of the reaction; in fact, the
extrapolated values at zero time suggested that the unsymmetric diene was favored initially. This kinetic study strongly
supported a non-pairwise reaction scheme, such as the carbene-metal concept, but it did not entirely exclude pairwise
schemes. Katz conceded that the discriminating merit of his
experiment holds true only if one assumes that the transalkylidenation step in pairwise scrambling mechanisms [Eq. (a)]
is the rate determining step. However, if the exchange step
were rate determining [Eq. (c)] his data could be accounted
for by either scheme.
It can be demonstrated that for an equimolar mixture of
(7) and [1,1,8,8-D4]-(7) one expects to obtain a statistical
distribution (I :2: 1) of [D4]-, [Dz]- and [Do]-ethylene at
low conversion even in the absence of ethylene redistribution
reactions. On the other hand, the pairwise path predicts different distributions of the deuterated ethylenes, depending on
the relative rates of metathesis versus exchange. The painvise
scheme is presented in Eq. (0) and (p).
2.4. Metathesis of a,w-Dienes
Nonconjugated a,w-dienes, when treated with a metathesis
catalyst, may undergo either metathetical polycondensat i ~ n [ ' ~forming
]
open chain oligomers with terminal double
bonds [Eq. (k)], or they may undergo intramolecular metato form a cycloolefin [Eq. (I)]. Both processes will
yield ethylene.
The factor which determines which process will prevail
at equilibrium depends primarily on the value of n, which
correlates with the relative thermodynamic stability of a given
I-(CHz),CH=CHJ
sequence in a ring versus a chain. For
example, D ~ l Z ' A s t a [metathesized
~~]
1,4-pentadiene and 1,5hexadiene exclusively to higher molecular weight linear oligomers because the corresponding monomeric cycloolefins
A
p + H 2
D2C DzC-CH2
p CH~=CH(CH~),,CH=CH~
74
are highly strained. On the other hand, Kr011[~~J
was able
to convert 1,7-octadiene to the stable product cyclohexene
with 99 % selectivity.
404
The di-deutesated and tetra-deuterated bidentate bisolefin
complexes produced according to Eq. (p) will then metathesize
; , i d ~ ~ ~ ~ - 3 . ' v L T ~ n t ~ ~ - ~ ~ ~ ~ c
ance with Eq. (0).
If intramolecular exchange were rapid and k,, $ k , (i. e.,
the transalkylidenation step were rate determining), then one
would expect a 1 :0: 1 distribution of [D4]-, [Dz]- and [Do]ethylene in the initial stage of reaction. Thus, as a cyclohexene
is displaced by a l,7-octadiene [Eq. (o)] the new complex
rapidly eliminates ethylene and forms a bidentate species [left
Angew. Chrm. In[. Ed. Engl.
1 Vol. 15 ( 1 9 7 6 ) No. 7
side of Eq. (p)]. If on the other hand k,*kk,,, the resulting
bisolefin complex would undergo transalkylidenation [right
side of Eq. (p)], giving rise to the formation of some CH2=CD2
in addition to C2H4. Grubbs calculated 1 : 1.6 : 1 as being the
limiting [D4]-, [Dz]-, [Do]-ethylene ratio for the process
depicted in Eq. (0) and Eq. (p)when k,$ kcx. Experimentally,
the distribution was substantially different from either limiting
value. The abundance of the respective ethylenes was closer
to a statistical distribution, consistent with a non-pairwise
mechanism. A significant finding was that the
[D4] :[Dz] :[Do] distribution of the residual 1,7-octadiene
did not vary from start to finish in this particular metathesis
experiment. This strongly indicates that essentially no regenerative intermolecular scrambling of methylenes among terminal
double bonds of diene molecules took place. Furthermore,
examining the [D4]-, [D2]-, and [Do]-ethylene distribution
before and after spiking the reaction mixture with [Do]-ethylene led Grubbs to conclude that no significant degree of
scrambling among the ethylenes took place under the prevailing reaction conditions.
Heretofore, a discussion of four experimental observations,
mentioned quite often as evidence for a non-pairwise scheme
for the transalkylidenation step, has been presented. Of these,
Grubb’s
stands alone as the one that cannot be readily
accounted for by pairwise schemes.
3. Evidence for Carbene-Metal Involvement
Although Pettitc61 did not apply “carbene” nomenclature
to his pairwise transition state proposal, his description
obviously alluded to an involvement of carbenes in olefin
metathesis. Since then there have been several experimental
disclosures that could be interpreted as providing evidence
for carbene involvement in the reaction. Cardidz61 observed
some disproportionation when he reacted a mixture of unsaturated amines with a rhodium phosphane catalyst [Eq. (9);
R=C6H5, R’=P-CH3-C6H4].
As a side product, a rhodium(1) complex with one half
of the electron-rich olefin was isolated, for which the structure
(8) was proposed. (The analogous compound with R instead
of R has also been found.)
Subsequently, O ’ N e i / / [ 2demonstrated
8~
in an elegant fashion
that a M O ( C O ) ~ / A ~catalyst
~ O ~ was capable of converting
ethylene into propylene directly:
O’Neill attempted to outline a mechanism for this unique
3 C2+2C3 process. The suggested most “straightforward pathway” involved a splitting of one ethylene into methylenes,
and addition of each of the latter to another ethylene, followed
by rearrangement of the resulting trimethylene into propylene.
It must be stressed that, although the catalyst for this process
is formally a metathesis catalyst, the two reactions are qualitatively distinct. Any conceivable mechanism for the 3 C 2 - + 2 C 3
reaction must reconcile migration of hydrogens. Selective olefin metathesis reactions do not involve hydrogen migrations.
Hence, a carbene mechanism for a 3 Cz - + 2C3 cannot be reasonably “transplanted” to olefin metathesis.
Dolgoplosk[8]was able to demonstrate that an active catalyst
for cycloolefin metathesis can be generated by co-reacting
WC16 or WCI4 with phenyldiazomethane. In the absence of
cycloolefin monomer, a reaction analogous to O’Neill‘s [Eq.
(r)] was observed, and phenyldiazomethane was decomposed
to nitrogen and stilbene. Dolgoplosk proposed a sequence
as depicted in Eq. (t):
CeHsCH=WX.
+ C ~ H ~ C H N+I N2 + Ce,HsCH=CHCeH5 + WX,
(t)
In the presence of cycloolefin monomer a very high molecular
weight polyalkenamer is formed. The mechanism adopted
for this polymerization is identical to Chauuin’s,wherein propagation-depropagation equilibrium [Eq. (e)] occurs at an active
carbene-metal chain-end,
Recently, C ~ s e y [ ’ ~ isolated
’
(C6H5),C=W(CO),, a
carbene-metal complex free of a heteroatom bonded
to Ccarbene.Until this report, it was generally believed
that stable metal-carbene complexes required an electron donor heteroatom directly attached to the Ccarbene.
Following Casey’s work, S c h r o ~ k [ ~ ’‘1*demonstrated that “unstabilized” carbene complexes of tantalum, such as
(CH3)3CHC=Ta[CH2C(CH3)3]3
and
HzC=Ta(CSH5)2(CH3),can be prepared, isolated and characterized.
Upon heating (C6H5)2C=W(C0)5with isobutene (1 OO”C,
2 h) a mixture of products was
indicating a net
transfer of a [CH,] group from isobutene to the diphenylcarbene [Eq. (u)].
B
PhjP,
N
Cl;Rh=C<
N
PhsP
3
(8)
,CH2 \
(CH,)Z.C-C(C~H~)Z.
+
(CsH5)2CZCHz + W ( C 0 ) s
(U)
I
R
O”ei[l[271observed that a CoO/Mo03 supported catalyst,
which is active in propylene metathesis, readily decomposed
diazomethane into ethylene and N,:
It was suggested that the same sites that are active in metathesis
“also selectively convert adsorbed methylenes into ethylene”.
Anyew. Chem. Int. Ed. Engl. J Vol. I5 ( 1 9 7 6 ) No. 7
The mechanism proposed for this reaction [Eq. (v)] accommodates both the observed cyclopropanation and the methylene transfer from isobutene to the diphenylcarbene (the carbony1 ligands have been omitted).
Casey’s remarkable experiment, when first carried out, was
stoichiometric rather than truly catalytic, and the element
of reversibility was lacking. Very recently K a t ~ ‘ has
~ ~shown
]
that ( C ~ H S ) ~ C = W ( C O
can
) ~in fact act as a metathesis catalyst.
405
Ph
Ph
Me
Ph
Me
place. For example, in the reaction of (CH3)2Znwith wc16
carried out in a deuterated solvent, methane free of deuterium
was produced. It has been proposed that a-hydrogen elimination from the tungsten-methyl group occurs [Eq. (x) or (y)],
leading to the formation of the initial carbene-tungsten
complex.
The existence of a W-CH, ~t W(=CH2)H equilibrium,
proceeding by migration of an a-hydrogen from a o-bonded
alkyl to the metal, has also been proposed by Greea[381 in
connection with the reversible rearrangement [Eq. (z)] :
Ph
Me
4. Possible Origins of Carbene-Metal Species
The variety of metals and their derivatives and combinations
which demonstrate metathesis activity is very broad. If the
non-pairwise carbene-to-metallocycle mechanism is to be
accepted, it becomes of interest to examine potential routes
to the formation of the initial carbene-metal entity. For this
purpose it is convenient to classify metathesis catalysts into
two major categories: a) catalyst combinations that do not
involve an organometallic component, and b) combinations
that involve organometallic cocatalysts. To the first category
belong the classical heterogeneous supported catalysts such
as MOO,, W(CO)6, and Re207 on alumina. Combinations
such as WC16/AIX3, or the photochemically induced
w(co)6/cc14which was recently disclosed independently by
two research groups[343351, are also included in this category.
Typical systems belonging to the second category are unmodified and alcohol-modified WC16/R,,,AlCl, (m+ n = 3), MoC15j
R ,A], WCI,/RSn, M o [ ( P ~ , P ) ~ C ~ ~ ( N O,A1
) ~ ],C1
/ R and
WC16/R2Zn.As pointed out by Dolgoplosk['], one may speculate that activecatalysts of the first category undergo initiation,
or in other words produce the original carbene-metal complex,
by a process analogous to the one proposed by
where a methoxycarbene-platinum entity was detected when
certain acetylene-platinum complexes were reacted with methanol. The analogous transformation of an olefin-metal is illustrated schematically in Eq. (w).
,
It must be emphasized that this proposal is highly speculative, and no experimental evidence exists at present in support
of such a process in actual metathesis reactions.
Catalyst systems that are activated by organometallic cocatalysts, presumably forming transition metal intermediates
containing at least one alkyl group which is o-bonded to
the metal, are prone to transform into carbene-metal entities.
M ~ e t t e r t i e s r371
~ ~observed
'
that, during the reaction of WCls
with metal alkyls, a significant evolution of alkanes takes
406
Recent work by Schrock with organotantalum compounds
suggests the feasibility of both intramolecular and intermolecular a-hydrogen abstractions as straight-forward routes to isolable carbene-tantalum complexes which are not stabilized
by a heteroatom. X-ray and NMR analyses were e m p l ~ y e d [ ~ ~ J
to determine the structureofCH,=Ta(C,H,),CH,
and to estimate the energy barrier to rotation about the Ta==CH, double
bond. The complex, as illustrated in Fig. 2, has two eclipsed
cyclopentadiene rings. The H2C-Ta bond length is shorter
than the H3C-Ta bond, indicating some double bond character for the methyleneligand. The methylene ligand is oriented
perpendicular to the C-Ta-C
plane. These observations
are in accord with the concept of a Ccarbene having an sp2
hybridization, wherein its vacant pzorbital can overlap preferably with a hybridized Ta orbital in the C - T a x plane.
Q
_...-.
..__.
Fig. 2. Structure of the complex CH2=Ta(C5H&CH3.
The barriers of rotation for =CH2, =CHC(CH3)3, and
=CHCsH5 around the respective Ccarbene-Ta bonds were
estimated from temperature-dependent NMR a n a l y s e ~ [ ~the
~l;
free energies of rotations, AG *, are 2 21.4, 16.8, and 19.3 kcalj
mol for the respective carbene ligands. These high values
directly reflect the significant double bond character of carbene-metal complexes that bear no heteroatom substituents
on the carbene carbon atom.
Evidence for an alternate route for the formation of the
original carbene-metal entity was submitted recently by Far0na[40*41! Using Re(C0)5CI in combination with either
C2H5A1C12or CH3A1Cl2,the homo-metatheses of 1,7-octadiene and 4-octene was investigated. Using highly sensitive
GC-Mass Spectroscopy techniques, he was able to monitor
trace amounts of anomalous olefinic products which were
proposed as resulting from the first-formed carbene. The firstformed olefins in the metathesis of 1,7-octadiene when
C2H5A1C12cocatalyst was employed were identified as 1butene and 1,7-decadiene, whereas when 4-octene substrate
was used, minute amounts of 3-heptene were found. Accordingly, Farona concluded that the original carbene-metal must
have been a propylidene-Re entity. In an analogous experiment
Anyew. Chem. Inr. Ed. Enql. J Vol. 15 ( 1 9 7 6 ) No. 7
obtain propylene, 3-heptene, 1 -butene and 2-hexene. The data
point to the fact that, although 1-pentene is sluggish towards
self-metathesis, it readily reacts with 2-pentene to produce
the unsymmetric cross products. The results also suggest that
I-pentene, in fact, inhibits the self-metathesis of 2-pentene
to the extent that the.tendency of 2-pentene to yield cross
products with 1 -pentene is eight times greater than its tendency
y1 /oA1C12+___,
C2H5AIC12
(C0)4Re=C,
to self-metathesize.This observation was recently reconfirmed
- [C~H,AICI,P
C ZH5
by M u e t t e r t i e ~ ' ~using
~ ] a RA1Cl2/WCl6/ROH catalyst on
2-pentene and 1-nonene substrates. A possible hypothesis for
this observation is that terminal olefins are, in fact, more
reactive than internal olefins, but that they prefer to undergo
regenerative metathesis via either metallocycle ( 9 ) or ( I 0).
This phenomenon has been recently reconfirmed by K ~ t z ' ~ ~ ]
and M ~ e t t e r t i e s [ ~Evidently,
~'.
transition states ( I f ) and ( f 2)
are not favored, since they are incompatible with regenerative
metathesis.
using CH3AICI2cocatalyst, the first-formed olefins suggested
the formation of an ethylidene-Re carbene initially.
The proposed mode of formation of the original propylidene-Re complex as postulated by F a r o n ~ [ is~ ~illustrated
]
in Eq. (a, a):
Y1
( C 0 ) 4 R e C 0 + C2H5A1C12 -+
0
(C0)4Re-CHC2H5
+
OAlClZ
The presence of an unusual by-product, octa-I ,7-dien-3-one,
was claimed as supporting evidence for this scheme.
In conclusion, whereas formation of the original carbenemetal can be explained in catalyst systems that involve metal
alkyl components, no experimental work exists to suggest
a pathway for carbene formation in catalyst systems that
do not employ organometallics. Evidence supports the contention that sp2-bonding of carbene ligands which are not stabilized by heteroatom substituents is substantial, and the carbene-metal bond should be viewed as a double bond in the
proposed carbene-to-metallocycle mechanism.
5. Structural Selectivity in Cross-MetathesisReactions
Earlier, we disclosed[", 8 , 2 0 1 that certain catalysts which
display a low apparent metathesis activity with terminal olefins, when employed on mixtures of terminal and internal
olefins, lead to the selective formation of cross metathesis
products. Table 2 presents the relative distribution of the
various cross metathesis products of I-pentene with cis-2-pentene after 25 % of the original 2-pentene had been consumed
(50 % of theoretical equilibrium). Self-metathesis of either olefin will lead to symmetric products; ethylene and 4-octene
from 1-pentene, and 2-butene and 3-hexene from 2-pentene.
The major products from the cross metathesis were the unsymmetric ones. Depending on the orientation of the incoming
olefin with relation to a specific carbene-metal site, one can
Table 2. Cross metathesis of I-pentene and cis-2-pentene (after Ref. [IS]).
CaH,=CH>
+ C2H5CH=CHCH,
Relative concentrations [a]
Symmetric
CH2=CH2
CSH7CH=CHC3H7
CH~CHZCHCH~
C2HsCH=CHC2H5
Unsymmetric
1
1
2
2
CHBCH=CH2
C2HSCH=CHC3H,
C2HSCH=CH*
CH~CHZCHC~H~
[a] Values at 50 % of theoretical equilibrium.
Angew. Chem. Int. Ed. Engl.
1 Val. I5 ( 1 9 7 6 ) N o . 7
4
4
12
12
When terminal and internal olefins are co-reacted, regenerative metathesis prevails, thus inhibiting self-metathesis of the
internal olefin. Terminal olefins occupy the catalyst, but occassionally an internal olefin enters the complex, whereupon
metathesis yields an unsymmetric product.
The regenerative methathesis of terminal olefins was confirmed experimentally by examining the product distribution
in the cross metathesis of -pentene with [Dl o,-l -pentene
(Table 3).
Table 3. Cross metathesis of I-pentene and [Dl0]-l-pentene; GC/MS data.
C8Hi6,C,H,D, and C,D,, formed only m trace amounts (after Ref. [18]).
Compounds
m/e
initially
C~DTCD=CD~
C~D~CD=CHZ
C3H7CH=CD2
C~HTCH=CH~
80
78
72
70
0.1074
-
0.8926
Relative concentrations
after
equilibrium
l0min
(calc.)
0.0395
0.0685
0.0771
0.8148
0.01 15
0.0959
0.0959
0.7967
The data in Table 3 indicate that after a 10 minutes' reaction
the original mixture of C3H7CH=CH2 (m/e=70) and
C,D,CD=CD, (m/e=80) underwent about 70% of the theoretical scrambling. Only trace amounts of 4-octenes were
formed.
In an attempt to resolve whether regenerative metathesis
of terminal olefins proceeds via transition state ( 9 ) or (fO),
the route of telomer formation was examined in the cross
metathesis of I-pentene and cyclopentene. This has been reported[18.201 to produce preferentially the homologous series
of unsymmetric polyenes of the general formula
C,H,CH[=CH-(CH2)3-CH=],CH2
(Cia,
C 151
Czo...etc.),along with minor amounts of the two symmetric
homologous series, CH,[=CH-(CH 2)3-CH=]xCH2
407
(C,, CI2, CI7, C2, etc.) and C3H7CH[=CH-(CH2),CH=],CHC3H7 (C,,, C,,, C,,, C,, etc.)[']. Figure 3
shows a chromatogram of a [I-pentene + cyclopentene]
reaction product mixture. The main peaks are attributed to
the unsymmetric homologous compounds.
If regenerative metathesis were to proceed oia intermediate
(9), it is evident that the eventual incorporation of cyclopentene units to form the unsymmetric polyenes will follow
sequence (ab):
On the other hand, if one assumes that the regenerative
metathesis of I-pentene proceeds via transition state ( l o ) ,
then an accounting of the production of the selective unsymmetric polyenes will require the participation of an "unfavored"
transition state of type ( I I ) [see Eq. (ac)].
Since two types of carbenes are invoked here, the course
of reaction in Eq. (ac) may be energetically less favorable
than that of Eq. (ab), which invokes entirely monosubstituted
carbenes throughout. However, the results do not unequivocally exclude either scheme.
Studies involving cross-reactions of ethylene with internal
olefins might be of value, since types (11) and (12) would
now be unavoidable during the formation of cross-products,
and kinetic effects might become discernible.
Some insight into the influence of steric effects on the course
of metathesis reactions was revealed in the reaction of l-pentene with 2-pentene discussed previously. For a reaction
sequence analogous to that in Eq. (ab), cross-metathesis
would
be
preceded
by
formation
of
C3H7CH=W(CH3CH=CHC2H5),which is capable of yielding two isomeric metallocycle intermediates. These will subsequently afford either 2-hexene and I-butene [Eq. (ad)] or
3-heptene and propylene [Eq. (ae)].
C~H~CH-CHZ
I
W-
0
20
10
i [minl--+
Fig. 3. Gas chromatogram of polyenes obtained in the 1-pentene
tene cross metathesis (after Ref. [18]).
CHz-CH,
1
1
W -CH'
+ cyclopen
The data in Table 2 suggests that, for the catalyst system
(C2H5)2AICl/WC16/ClC2H40H, reaction sequence (ad) is
significantly favored over that of (ae). One must attribute
this difference in ease of formation of the respective products
to steric hindrance. On the basis of reaction scheme (ae),
the implication appears to be that steric interactions involving
substituents located on the P-carbon in the transient metallocycle ring are more severe than those for the a-carbon substituents, although the origin of this steric demand is not obvious.
Further evidence for preferential formation of a metallocycle
where the least substituted carbon is preferentially the p-carbon was given recently by
In model compound
studies involving a tungsten compound which contained a
bulky diarylcarbene ligand, transfer of an alkylidene unit from
a reacting olefin resulted in highly selective transfer of the
least substituted alkylidene unit, as described previously in
Eq. (v). The authors pointed out that the new metal-carbene
species which resulted after metathesis was preferentially that
C,H,CH=CH2
(CH2)3
C H Z - C H C ~ H+~C~H,CH=CH~
CH2-CHC3H7
W-CHCH3
W-CHCH,
I
p] In
reactions of cyclooctene with a-olefins. Herisson and Chauuin also
found high selectivity at low conversions [I I& .ilthough a lesser selectivity
was observed by Lal and Smith at higher ~\?in;crrions of the cyclooctene
~481.
408
1
CH C 3 H 7
I
-
1
1
-CzHsCH=CH?
+ C,H,CH=CH~
-CsH+H=CHCH3
*
(ad
CHz-CHCsH,
I
1
W-CH2
Angew. Chem. Int. Ed. Engl.
1 Vol. I S (1976) No.
7
which possessed the most stable (i. e., highly substituted) carbene.
In considering the steric effects observed in the reaction
of I-pentene with 2-pentene, one can also invoke a scheme
which evolves from structure ( 9 ) [analogous to Eq. (ac)].
The product distribution in Table 2 now indicates pathway
(ag) to predominate, and is consistent with the notion that
the greatest steric interactions occur nearest the transition
metal. Here, crowding due to adjacent ligands about the metal
would appear to be a plausible source of these steric effects.
6. Conclusions
Despite numerous new and sophisticated experiments continually being brought to bear on questions regarding the
chemistry of olefin metathesis, a number of challenging problems remain unsolved.
The concensus of recent publications is that a chain process,
presumably involving a metal-carbene species, is operating,
yet arguments in support of the pairwise scheme continue
to rise Phoenix-like. Most recently, G a ~ s m a n presented
[~~~
new data to show that, with an appropriately chosen diolefin,
quantitative interconversion of the diolefin to a cyclobutane
structure could be accomplished using a typical olefin metathesis catalyst. Previously, the failure to observe cyclobutanes
during metathesis reactions had been given as one argument
against the plausibility of the “quasicyclobutane” transition
state. It had been expected that, since two isolated double
bonds are energetically comparable to a cyclobutane, both
forms should be present under metathesis conditions according
to the pairwise scheme. Gassman points out that the determining factor may be the relative energies of a pair of coordinated
double bonds (substantial stabilization) us a coordinated cyclobutane (very little stabilization), rather than the relative stabilities of the free species.
Mechanisms invoking other intermediates such as the tetrahedral trimethylene model (where all three methylenes are
equally related to the metal), as described by Haines and
Leigh in their recent excellent and comprehensive
may also adequately explain metathesis reactions, and cannot
be discounted. In fact, several authors have suggested that
perhaps more than one pathway for metathesis may be possible, and in a given situation would be determined by the
type of catalyst used.
Because of the tremendous variety in reagents which serve
to generate metathesis-active species, reconciliation of a common mechanism for metathesis in these diverse systems represent a real challenge. Certainly, the catalysts appear quite
tolerant of diversity in ligands about the transition metal,
with the possibility of bimetallic species being suggested occassionally.
The origin of stereospecificity in metathesis reactions
remains a mystery. Homogeneous catalysts rather universally
show a weak cis-directing tendency in the reaction of cis-olefins.
demonstrated that regardless of the catalyst used,
cis-2-penteneafforded about 55 % cis-3-hexene and 60 % cis-2butene initially. When cycloolefins are employed, however,
wide variations in polymer structure are known to occur,
depending upon the catalyst used. Apparently the growing
polymer chain exerts a pronounced influence on the stereospecificity of the metathesis process. This phenomenon may be
Angew. Chum. In!. Ed. Enyl. J Vol. 15 (1976) N o . 7
due to multiple coordination of the polymer chain, which
cannot occur in the reaction of acyclic olefins.
Received: March 16, 1976 [A 117 IE]
German version: Angew. Chem. 88,433 (1976)
[1] N . Calderon, H . I: Chen, and K . W Scott, Tetrahedron Lett. 1967,
3327.
[2] N . Calderon, E . A . Ofstead, J . P. Ward, W A . Judy, and K . W Scott,
J. Am. Chem. SOC. 90,4133 (1968).
[3] J . C. Mol, J . A . Moulijn, and C. Boelhouwer, Chem. Commun. 1968,
633.
[4] R. Cramer, J. Am. Chem. SOC.86, 217 (1964).
[5] C. P. C. Bradshaw, E. J . Howman, and L. Timer, J. Catal. 7, 269
(1967).
161 G. S. Lewandos and R . Pettit, Tetrahedron Lett. 1971, 789.
[7] R. H. Grubbs and 7: K . Brunck, J. Am. Chem. SOC. 94, 2538 (1972).
[XI B. A . Dolgoplosk, K . L. Makouetsty, 7: G . Golenko, I: J! Korshak,
and E . 1. Timyakoua, Eur. Polym. J. 10, 901 (1974).
[9] N . Calderon, E. A . Ofstead, and W A . Judy, Paper presented at the
Central Regional Meeting of the American Chemical Society, Akron,
Ohio, May 1968.
[lo] N . Calderon, Acc. Chem. Res. 5 , 127 (1972).
[11] J . L. Herisson and I: Chauuin, Makromol. Chem. 141, 161 (1970).
[12] J . P. Arlie, I: Chauuin, D. Commereuc, and J . P. Souflet, Makromol.
Chem. 175,861 (1974).
1131 K . W Scott, N . Calderon, E. A . Ofstead, W A . Judy, and J . P . Ward.
Adv. Chem. Ser; 91, 399 (1969).
[14] K . W Scott, N . Calderon, E. A. Ofstead, W A . Judy, and J . P . Ward,
Rubber Chem. Technol. 44, 1341 (1971).
[I51 K . W Scott, Polym. Prepr. Am. Chem. SOC., Div. Polym. Chem. 13,
874 (1972).
[16] E. A . Ofstead and N . Calderon, Makromol. Chem. 154, 21 (1972).
1171 N . Calderon, Paper presented at the Japan-US. Seminar on Unsolved
Problems in Ionic Polymerization, Hakone (Japan), October, 1974.
[I81 W J . Kelly and N . Calderon, J. Macromol. Sci.-Chem. A 9 , 91 1 (1975).
[19] E. A . Zuch, W B. Hughes, D. H . Kubicek, and E. 7: Kitterman, J.
Am. Chem. Sac. 92, 528 (1970).
[20] W J . Kelly, Am. Chem. SOC.,Div. Pet. Chem. Prepr. 17, H32 (1972).
[21] R. Rossi, P. Diuersi, A . Lucherini, and L. Porri, Tetrahedron Lett.
1974, 879.
[22] 7: J . K a t z and J . McGinnis, J. Am. Chem. SOC.97, 1592 (1975).
[23] G . Dall‘Asta, G . Stigliani, A . Greco, and L. Motta, Polym. Prepr. Am.
Chem. SOC.,Div. Polym. Chem. 13, 910 (1972); see also Chim. Ind.
Milan 55, 142 (1973).
[24] W R . Kroll and G. Doyle, J. Catal. 24, 356 (1972).
[25] R . H . Grubbs, P. L. Burk, and D. C . Carr, J . Am. Chem. SOC. 97,
3265 (1975).
[26] D. J . Cardin, M . J . Doyle, and M . F. Lappert, J. Chem. SOC. Chem.
Commun. 1972, 927.
[27] P . P . O’Neill and J . J . Rooney, J. Chem. SOC.Chem. Commun. 1972,
104.
[28] P. P. O’Neill and J . J . Rooney, J. Am. Chem. SOC.94, 4383 (1972).
[29] C. P . Casey and 7: J . Burkhardt, J. Am. Chem. SOC.95, 5833 (1973).
[30] R. R . Schrock, J. Am. Chem. SOC. 96, 6796 (1974).
[31] R. R . Schrock, J. Am. Chem. SOC. 97,6577 (1975).
I
.
Burkhardt. J. Am. Chem. SOC. 96. 7808 (1974).
[32] C. P. Casev and 7: .
M . H . Chisholm and H . C. Clark, Chem. Commun. 1970,763.
A . Agapiou and E . McNelis, J. Chem. SOC. Chem. Commun. 1975,
187.
P. Krausz, F. Garnier, and J . E . Dubois, J. Am. Chem. SOC.97, 437
(1975).
E. L. Muetterties and M . A . Busch, J. Chem. SOC. Chem. Commun
1974, 754.
E. L. Muetterties, Inorg. Chem. 14, 951 (1975).
N . J . Cooper and M . L. H . Green, J . Chem. SOC. Chem. Commun.
1974, 761.
L. J . Guggenberger and R . R . Schrock, J. Am. Chem. SOC.97, 6578
(1975).
M . F. Farona and W S. Greenlee, J. Chem. SOC. Chem. Commun.
1975, 759.
W S. Greenlee and M . F. Farona, Inorg. Chem., in press.
M . 7: Mocella, M . A . Busch, and E. L. Muetterties, J. Am. Chem.
SOC.98, 1283 (1976).
J . McGinnis, 7: J . K a t z , and S. Hurwitz, J. Am. Chem. SOC. 98, 605
(1976).
C. P. Casey, H . E. Tuinstra, and M . C. Saeman, J. Am. Chem. SOC.
98, 608 (1976).
P. G . Gassman and 7: H . Johnson, J. Am. Chem. SOC. 98, 861 (1976).
R. J . Haines and G . J . Leigh, Chem. SOC. Rev. 4, 155 (1975).
J . M . Basset, J . L. Bilhou, R . Mutin, and A . Theolier, J. Am. Chem.
SOC.97, 7376 ( I 975).
J . Lal and R. R . Smith, J. Org. Chem. 40. 775 (1975).
~I
409
Документ
Категория
Без категории
Просмотров
2
Размер файла
923 Кб
Теги
mechanistic, metathesis, aspects, olefin
1/--страниц
Пожаловаться на содержимое документа