A STUDY CONCERNING THE MECHANISM
OF THE OLEFIN METATHESIS REACTION USING THE

METATHESIS OF LABELED 2,8—DECADIENES
By

Charles Richard Hoppin

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

1978

ABSTRACT
A STUDY CONCERNING THE MECHANISM
OF THE OLEFIN METATHESIS REACTION USING THE
METATHESIS OF LABELED 2,8-DECADIENES
By

Charles Richard Hoppin

The mechanism of the olefin metathesis reaction has
been under intensive investigation in recent years.
Previous experiments that have studied the metathesis of
deuterium labeled diolefins have indicated that a non-
pairwise carbene mechanism is operating rather than the
originally proposed pairwise schemes. However, these
studies did not precisely account for possible isotopic
scrambling in the reactants or products that would, if
occurring, lead to ambiguous results.

This study investigates the metathesis of isotopically
labeled 2,8-decadienes which allows the monitoring of all
of the competing metathesis processes and accounts for the
scrambling of the labels in the starting material and
products.

Initially the metathesis of isomerically labeled cis,
cis—2,8-decadiene was investigated. This reaction,
catalyzed by several transition metal based non—supported

catalysts, yields cyclohexene and cis- and trans-Z-butene

Charles Richard Hoppin
irreversibly in high yield and selectivity. In the early
stages of the reaction the formation of cis-2-butene is
favored over trans-2-butene, and then the butenes are
isomerized by a metathesis process to give a thermo-
dynamically stable mixture eventually. At the same time,
the starting material isomerizes by a similar process.

The two isomerizations have been related by graphical
methods, and a kinetics model accounts for this relation-
ship.

The results of the studies of the metathesis of
isomerically labeled decadienes are used to interpret the
results obtained from the metathesis of mixtures of dO— and
d6-2,8—decadiene. It was found that the 2—butenes can be
sampled at conversions of starting material of SAO% which
will give isotopic distributions that will distinguish
between the non-pairwise and pairwise schemes. The non-
pairwise mechanism has been implicated from the results
obtained.

Some evidence was also found for possible initiation
and termination modes of the olefin metathesis reaction.
An initiating metal-methylene species was trapped with one
end of 2,8—decadiene to yield propylene. The source of
each fragment of propylene was determined by deuterium
labeling. Ethylene was also observed and is believed to

arise from a slow dimerization of the metal-methylene.

To

Bonnie

"I know a passion still more deeply charming
That fever'd youth e'er felt; and that is love,
By long experience mellow'd into friendship."

-Thomson

ii

.n .

.1

2..

T,
r:

r: .rh
a: a c

ACKNOWLEDGMENTS

I would like to take this opportunity to express my
appreciation for the guidance and patience extended to me
by my research preceptor, Dr. Robert H. Grubbs, and to
wish him and his family well on the occasion of their move
to Pasadena and Caltech.

I would also like to recognize the members of my
research committee, Dr. Michael N. Rathke, Dr. Carl H.
Brubaker and Dr. Richard H. Schwendeman, who have provided
helpful input into this work.

My fellow graduate students, in particular Dick
Stuart, Chip Millard and Rich Woodbury, have provided
much friendship and memorable moments and have made
graduate school enjoyable.

Also, I would like to mention that my parents,
brothers and sister as well as my in-laws have provided
a good deal of support and encouragement throughout my
academic career.

Finally, there are few words to express my fondest
appreciation for my wife, Bonnie, who typed and edited
this dissertation. Her patience and support have made
this work possible.

111

TABLE OF CONTENTS

Page

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . V
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . Vi
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . 1
RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . “2
Synthesis of Olefins . . . . . . . . . . . . . . . U2
Catalysts and Metathesis Reactions . . . . . . . . A3
Metathesis of cis,cis—2,8-decadiene . . . . . . . . AA
Deuterium Labeling Studies . . . . . . . . . . . . 7O

Initiation and Termination of Olefin Metathesis . . 81

EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . 88
APPENDIX
PROCEDURE FOR THE CALCULATION OF THE POSSIBLE
d3/d6-2-BUTENE RATIOS ARISING FROM A PAIRWISE
MECHANISM . . . . . . . . . . . . . . . . . . . . . 97

LIST OF REFERENCES . . . . . . . . . . . . . . . . . . .103

iv

O\U'l

LIST OF TABLES

Typical Olefin Metathesis Catalytic Systems

Predicted Ethylene Distributions for the
Metathesis of a 1:1 Mixture of I? and I}

Extrapolated t/c-Cu Values at 0% Conversion . .
Approximated Relative Rate Constants
Calculated t/c-Cu and t/c-C10 Values

Experimental Data from the Metathesis of a

l : 1.21 d6 : dO Mixture of cis,cis-2,8-decadiene

Calculated d3/d6-CA Values by a Pairwise Scheme

Adjusted d3/d6-Cu Values Calculated by a Pairwise
SCheme O C O O O O O O O O O O O O O O O O O 0

Reaction of 2,8-decadiene Using Catalysts A and B
with Olefin/Metal = 150—170 . . . . . . . . .

Page

16
A6
66
68

7O
75

77

86

CDNOUW

\O

ll.

l2.

13.

IA.

15.
l6.

17.

LIST OF FIGURES

Transalkylation
Transalkylidenation
The pairwise quasicyclobutane scheme

Metallocyclopentane mechanism for olefin
metathesis .

The non-pairwise carbene chain mechanism

Cross product metathesis studies . . . . . . . .
Proposed mechanism of 1,7-octadiene metathesis
Formation of a tungsten metallocyclobutane

Proposed formation of a metal carbene from an
olefin and an active metathesis catalyst

Formation of initiating metal carbene .

Proposed formation of an initiating metal
carbene from a non-supported catalyst system and
1,7-OCtadiene o o o o o o o o o o o o o o o o

Degenerate and productive metathesis processes

Metallocyclobutane model for explaining
stereospecificity . . . . . . . . . . . . . . .

Predicted orientation of the approach of a cis
olefin to a metal carbene . . . . . . . . . .

Synthesis of d6-cis,cis-2,8-decadiene . . . . . .

Change in the t/c-2-butene ratio during the
metathesis of cis,cis-2,8-decadiene . . . . . . .

Isomerization of 2-butenes and starting material
relative to the amount of conversion of starting
material . . . . . . . . . . . . . . . . . . . .

vi

Page

ll
13
17
26

26
3O

32

A7

l8.
19.

20.

21.

22.

23.

2A.

25.

26.

27.

28.

The relationship between t/c—Cu and t/c-C10

Competing metathesis processes during the
metathesis of cis,cis-2,8-decadiene

Relative concentrations of acyclic olefins
during the metathesis of cis, cis- 2, 8— decadiene
with catalyst A . . . . . . . . . . . . . . .

Comparison of t/c-C ratios from the metathesis
of cis—2-decene and cis,cis-2,8—decadiene

Possible reactions of the 08 metal carbene 1}

The four primary reactions of .3? with cis, cis-
2, 8- decadiene . . . . . . . . .

The eight secondary reactions of .1? with trans- -C10
and trans- 2- butene . . . . . . . . . . . .

Comparison of the calculated and the experimental
relationships between t/c-Cu and t/c-ClO . .

Observed d3/d6-Cu values during the metathesis of

a mixture of dO- and d6- 2, 8- decadiene at
particular t/c- ‘CA ratios . . .

Observed d /d6'CA values compared to the

calculated values for the non—pairwise (carbene)
and pairwise schemes . . . . . . . . . . . .

Origin of propylene from the interaction of 16
with 2, 8- decadiene . . . . . . .

Vii

Page
52

53

55

56
59

6O

62

69

72

79

82

INTRODUCTION

One of the most interesting reactions in organic
chemistry is the olefin metathesis reaction which involves
a unique transition metal catalyzed transalkylidenation
between olefins. Since its discovery a large volume of
research directed towards the elucidation of the mechanism
and exploitation of the versatility of the metathesis
reaction has been reported and reviewed.1 Some of the
general features of the reaction will be outlined along
with a review of recent published results that are pertinent
to the results presented here.

In 1964, Banks and Bailey2 reported that propylene
disproportionates over a heterogeneous catalyst yielding
2-butene and ethylene (equation 1).

2 CH2=CHCH3 -—5 CH3CH=CHCH3 + CH2=CH2 (1)

<?-_‘—‘
MOO3/Al203
150°, 30 atm.

Later, Calderon3

found several homogeneous systems capable
of metathesizing a wide variety of linear olefins. At about
the same time, similar catalysts were found to promote the
ring opening polymerization of cycloolefins such as cyclo-

pentene (equation 2).”

 

> {CH=CH-<CH2>3an (2)

It was soon determined that these two processes, dispropor-
tionation and ring-opening polymerization, were specific
cases of a new general type of reaction. The term "olefin
metathesis" was then proposed to describe this reaction.5

The olefin metathesis reaction is initiated by a wide
variety of transition metal based catalytic systems. These
catalyst recipes have been grouped into two general types
based on their preparation and handling.

The first group of catalysts are the "supported
catalysts". These typically involve transition metal
compounds impregnated on an inert, high surface area support
and have found some use in industrial processes.6 Some
typical systems are shown in Table 1. Although most of
these catalysts require high operating temperatures and
pressures, the rhenium oxide based system can promote
metathesis at ambient temperature and pressure.7

This body of work deals exclusively with the second
type of catalytic systems referred to as the "non-supported
catalysts". Typical examples are also given in Table 1.
These systems generally contain a mixture of a transiton
metal complex of either tungsten, molybdenum or rhenium with
a main group metal alkyl "co—catalyst" in an inert non-polar
solvent. Some catalytic systems require promoters such as

12

oxygen or ethanol. The resulting catalysts are capable of

Table 1. Typical Olefin Metathesis Catalytic Systems

 

 

 

 

 

Supported Reference
Re2O7/AI2O3 7
MoC06/A1203 8
CoO-MoO3/A12O3 9
WOB/SiO2 lO
MoO3/A1203 2,11
Non-Supported Reference
WC16/EtAlcl2/EtOH 3,12
WC16/nBuLi l3
WCl6/SnMeu l“
Mo(NO)2012(P¢3)2/ 15
M83A12Cl3

Re(CO)5C1/EtA1012 l6
W(CO)6/CClu/hv l7

WOClu/MeZZn 18

 

 

 

v I

Cu

 

A

promoting metathesis in very low concentrations. Typical
reactions are conducted with olefin/metal ratios between

19 has found that the

two hundred and five hundred. Hughes
rates of reaction are generally first order in both catalyst
and olefin under these conditions. He has also found that
the energy of activation is under 10 kcal/mole. This
implies that entropy differences are important contributors
to product distribution since the difference in the ground
state energies between most olefins is small. The
metathesis of 2-pentene illustrates this point. By using
several different catalysts an equilibrium is quickly

reached between the products and reactants shown in

equation 3
= —-——'. = =
2 CH3CH CHC2H5 _ CH3CH CHCH3 + C2H5CH CHCZHS (3)

based solely on a statistical distribution of olefins.19
Thus, in the case of most olefins, conversions of starting
material to products can be at the most 50%.

As seen in Table 1 there are a wide variety of
catalytic systems differing in oxidation states, ligands and
co-catalysts that can all promote the same reaction. It is
then difficult to describe the nature of true catalytic site
especially when the non-supported catalysts may be either
homogeneous or heterogeneous. Most of these catalysts are
quite sensitive to oxygen or moisture and usually do not
tolerate the presence of functional groups other than the

carbon-carbon double bond.20 A discussion describing recent

v.

v-r -

>7

n:

 

5
results regarding the interaction of transition metal
complexes with main group compounds in metathesis systems
will be presented later. At this point an outline of the
important mechanistic studies published recently regarding
the actual rearrangement step of the metathesis process
will be presented.

The first mechanistic consideration of the olefin
metathesis reaction was to determine whether the rearrange-
ment involved a transalkylation (Figure 1) or a trans-
alkylidenation (Figure 2). The process illustrated in
Figure l is a redistribution of alkyl groups around an
ethylene nucleus. The second possibility requires an
unprecedented carbon-carbon double bond cleavage and

scrambling of alkylidene units (Figure 2).

I
R'CH=CH-1—R' R'CH=CH R'
' *> + ‘
I
R2-1—CH=CHR2 R2 CH=CHR
I

 

2

Figure l. Transalkylation

l
RCH*CHR RCH CHR

I
ll +||

I
R'CHTCHR' R'CH CHR'

V

Figure 2. Transalkylidenation

6

Calderon21

devised a simple experiment from which he
concluded that a transalkylidenation is operative. He
studied the metathesis of a mixture of fully labeled and

unlabeled 2-butene according to equation A

CD3CD=CDCD3 + CH3CH=CHCH3 :;==:=:CH3CH=CDCD (A)

3
l

'b

and found only one new olefin 1. This product can only
arise via a carbon-carbon double bond cleavage. Similar
studies performed with other catalysts using Cl“ labeled
olefins reached the same conclusion.22

The first mechanistic schemes proposed to account for
this unique process generally called for the scheme shown
in Figure 3. This involves the initial formation of a bis-
olefin metal complex 2 followed by a rearrangement to a new
bis—olefin complex A through a "quasi-cyclobutane"
transition state 1 where M denotes the active catalyst.23
Most of the support for such a concerted rearrangement was
based on theoretical discussions that attempted to
illustrate how low lying metal d-orbitals could promote a

non-allowed Woodward-Hoffman 2 + 2-cycloaddition.214

Pettit25

found that cyclobutanes are unreactive to
metathesis conditions and are never found as products. As
an alternative to the quasicyclobutane model he prOposed a

tetramethylene metal species 2.

2 RCH=CHR' + M ‘___—-" M

Figure 3.

k

8X

RCH=CHR'

 

RCH=CHR'

2

RCH--- HR'
|\ /‘

\
I \ , I
l ‘M '

’ I
I / \

\

é’ ‘
R H---—CHR'

8U)

 

 

 

RCH CHR' RCH=CHR
k
ex
*—
RCH CHR' R'CH=CHR'
A +
’h
M

The pairwise quasicyclobutane scheme

RCH CHR
>14
A
RCH CHR
5
’b

Grubbs26 proposed that a metallocyclopentane inter-

mediate may be involved in.a pairwise scheme as Shown in
Figure A. Experimental evidence supporting such a scheme
was found by preparing a tungsten metallocyclopentane and
subjecting it to metathesis conditions. Through appropriate
labeling studies it was discovered that the olefins produced
could be best explained by a.rearrangement of intermediates
such as 6 and 1. There have been several precedents for
such metallocyclopentanes, and they are presently being
investigated extensively.27 However, one of the objections
to all of the pairwise schemes mentioned here was the
questionable existence of bis—olefin complexes 2 and A.
Such complexes have never been isolated from metathesis
reactions and are rare in transition metal chemistry.l
During the time the pairwise schemes were being
presented, experimental evidence was found that seemed to
contradict these schemes. Dolgoplosk28 found that induction
periods are required for ring—opening polymerizations, and
high molecular weight products are observed in the early

stages of the reaction. These are both qualities of a chain

reaction. Scott29 discovered that these high molecular

w.

¢

.>\

T

Q»

,3
.r9

“.2
P

h A
.C

 

9
weight polyalkenamers are linear and that the low molecular
weight oligomers are macrocyclic. It was also found that
these macrocycles did not result from an intermolecular

reaction between two smaller rings which a pairwise scheme

requires.
RCHTCHR' R' R RCH=CHR
I v— <——- ' M w——- If
R
RCH=CHR' R' R' R'CH=CHR'
2 6 7 A
"\J 'b ’L ’b

Figure A. Metallocyclopentane mechanism for olefin
metathesis

The pairwise scheme can still conceivably account for
these seemingly contradictory results. The formation of
linear polymers in ring-opening polymerizations can be
explained by the reaction of a trace acyclic olefin
impurity with a large macrocycle. Theoretically only one
molecule would be capable of opening a ring with a
molecular weight of several thousand. The accumulation of
high molecular weight products in the early stages of the
reaction can be explained by a kinetics argument. Referring
back to Figure 3, one need only consider the respective
rates of the olefins onto and off of the metal center (ke )

x
and the rate of the actual metathesis rearrangement step

IO

(kmet)' If kex<<kmet the olefins would remain attached to

the metal during several metatheses before dissociating
from the metal. In this way high molecular weight products
are accumulated very quickly. This possibility, once
referred to as the "sticky-olefin" hypothesis, has been

19 30

challenged by Hughes and recently by Katz who maintain

that the metathesis step is rate determining, i.e.,

k <<k

met ex'
A new type of mechanistic scheme was proposed by
Chauvin31 the better to explain the kinetics and product

distributions typical of cycloolefin polymerizations. The
scheme shown in Figure 5 differs from the other schemes in
that it is a non-pairwise chain reaction involving odd
number carbon intermediates. The first step involves the
formation of an initiating metal carbene 8 (also referred
to as a metal alkylidene) that reacts with a single olefin
creating a species that may resemble a metallocyclobutane§%
This three carbon metallocycle is then cleaved along a
different axis to give a new propagating metal carbene 1p
and a new olefin. During the past few years much work,
including the study presented here, has been directed
toward the distinction between the pairwise schemes and
this non-pairwise pathway.

Calderon and Kelly32 have used this scheme to account
for the product distributions of the macrocyclic oligomers
from ring-opening polymerizations. As mentioned earlier it

was found that these products are not formed from the

l1

 

MCH2R
t R' R"
R'CH=CHR" R'CH=CHR"
M=CHR ._____: —-"'
M=CHR ' M
8
’b
9
’b
R'CH
ll CHR"
M=CHR' + R"CH=CHR : M—ll
CHR
10
’b

 

 

R" !
__R"CH=CHR' R"CH=CHR'
"’ M=CHR"= ' M
R!

11

R"
i HR.

M=CHR" + R'CH=CHR' i—; M——-
CHR'

Figure 5. The non-pairwise carbene chain mechanism

l2
intermolecular reaction from two smaller rings, and so a
metal carbene "back-biting" scheme was proposed

(equation 5).

 

R M-T—R M=CHR
M ----- E L(
z (CH2)r1 "—’ CH2)n'—" + (5)
(CH2)nCH=CH&E) [(CH2)nCH=CH;; "' (CH2)n

(CH2)nCH=CH;%

Katz3O conducted several experiments that produce
results favoring the odd-carbon carbene scheme. He studied
the cross products formed at low conversion between cyclo-
octene and symmetrical olefins. In a typical study
(Figure 6) the concentrations of the ClA cross product 11
were significant at very low conversions. A pairwise scheme
would predict that only the 012, C6 and 016 products will be
formed initially. Katz found that the CIA/016 and the
ClA/Cl2 ratios were approximately 2:1 which matches the
random distribution of alkylidenes predicted by the Chauvin
scheme. However, as Katz admits, this distinction is only
valid if the metathesis step is rate determining. As in the
other experiments, the other extreme, kex<<kmet’ can be used

to explain the results by a pairwise scheme.

+ CH3CH=CHCH

Figure 6.

3

+

13

C3H

7

CH=CHC H —->CH CH=CHC H

3 7 3 3 7

-CHCH3
—CHCH3

"CHCH

C
lA

1N1

"CHC3H7

—CH3H7

C16

Cross product metathesis studies

1A
Work in this laboratory has exploited the unique
metathesis of 1,7-octadiene to distinguish between the two

schemes. As shown in equation 6,

6—. 0 + CH2=CH2 + CH
— M (6)

(< 1%)
(98%)

1,7—octadiene undergoes an intramolecular cyclization to
form cyclohexene and ethylene in high yield and selectivity.
Cyclohexene is one of the few simple olefins that remarkably
does not undergo further metathesis.33 With any other

a,w diolefin the resulting cyclic olefin formed polymerizes
to form polyalkenamers as shown for 1,6-heptadiene in

equation 7.

<: —p + CH2=CH2 (7)
M

M\\

{——CH=CH ( CH2 )§—)n

Grubbs314 has reported the metathesis of mixtures of
terminally labeled and unlabeled l,7-octadienes with three
different catalysts giving cyclohexene and labeled ethylenes

as shown in equation 8.

l5

2 =
_ CH2 CH2 (do)
CH2
+
12
—-’ + CH =CD (d ) (8)
2 2 2
+ M
+
—-CD2 CD2=CD2 (du)
—CD2
13

It is possible to predict the distribution of the labeled
ethylenes according to either scheme and compare the
calculated values with those obtained experimentally. All
of the pairwise schemes, regardless of the particular key
intermediate, can be considered together. If a pairwise
scheme were operating, then the two extreme cases mentioned
earlier must be considered. Referring once more to

Figure 3, if the metathesis step is rate determining,
kmet<<kex’ then it is easy to see that d2-ethylene cannot
be formed in the initial stages of the reaction. Assuming
a 1:1 mixture of 12:13, then the d0:d2:du ethylene ratios
would be 1:0:1. The second extreme for the pairwise schema
kex<<kmet’ involves a much more difficult calculation.

Using an iterative scheme described elsewhere, a l:l.6:l

ratio for the d :d2zdu ethylene is predicted. These num-

O
bers arise from the intramolecular nature of the reaction.

The carbene scheme predicts a random distribution of the

I

16
methylene units, or a 1:2:1 distribution. The key
distinction then is between the "sticky-olefin" pairwise
scheme, where kex<<kmet’ and the random distribution for the
carbene scheme. Any intermediate pairwise scheme where kex
is comparable to kmet will fall between the two extremes.
The predicted values for the ethylene distributions for the
reaction where 13:13 is 1:1 are summarized in Table 2.

Table 2. Predicted Ethylene Distributions for the
Metathesis of a 1:1 Mixture of 13 and k3

 

 

 

Calculated
Scheme d0 d2 d“
Pairwise
kmet<<kex l O l
kex<<kmet l 1.6 l
Carbene l 2 l

 

 

 

 

 

 

By using different ratios of 12:13 and different
catalysts, Grubbs found that the ethylene distributions at
low conversions most closely match the predicted carbene
ratios. Katz35 conducted a very similar study to this and
came to a similar conclusion. The mechanism was postulated
to follow the scheme shown in Figure 7. The results account
for the propagation steps of the chain mechanism proposed by
Chauvin. In this case the chain carrying species is the

metal-methylene moiety. The initiation and termination

—CH

-CH

-M=CH2 +M=CH

 

—CH

17

 

1 CH
+M=CH D -M——|| 2
C

——._2.- —. H2
-M=CH2 -CH —CH
2 2
+ CH2=CH2
+ M=CH
2
etc.

Figure 7. Proposed mechanism of l,7-octadiene metathesis

18
steps will be considered later.

There are some important restrictions that must be
placed on the study of the l,7—diene systems. It is
obvious that if there is an intermolecular scheme that
scrambles the labels in the starting material upon
immediate contact with the metal then isotopic distribution
in the ethylenes would be the same for either mechanism.
During Grubbs' study of the low conversion metathesis of
13:19 mixtures, the starting material was isolated and
found to have undergone minimal scrambling. At higher
conversions significant scrambling occurred via a degen-
erate metathesis, but was not studied precisely.

Another possible complication is that the labeled and
unlabeled ethylenes produced,may scramble via metathesis or
other pathways at the metal center before dissociating from
the catalyst, again giving ambiguous results. To check for

this possibility d ethylene was passed through the

0
reaction vessel, and the new ethylene mixture was isolated.
If there were a secondary scrambling scheme, the external
ethylene would change the equilibrium between the deuter-
ated ethylenes. This was not found since the d2:du ratio
remained the same. However, this is an external probe, and
. it is not certain whether the ethylene can be properly
mixed with the catalyst to achieve realistic results.

A new set of experiments was proposed to more

precisely account for the isotopic scrambling in the

starting material and products. It was decided to turn to

19
the study of the metathesis of mixtures of unlabled lA
m

and labeled 15 2,8-decadienes according to equation 9.
m

---CHCH3
-CHCH3 CH3CH=CHCH3 (do-Cu)
1A (do-Clo) +
+m -———> .] + CD3CH=CHCH3 (d3-Cu)
--CHCD3 + (9)
"CHCD3 CD3CH=CHCD3 («16-04)

15 (as-Clo)
'b

In a study analogous to the studies of the l,7-octadiene
system the 2-butenes were isolated during the reaction and
analyzed for isotOpic distribution. At the same time the
starting material was also isolated and examined for
scrambling. By understanding the isotopic scrambling
processes competing with productive metathesis, d0:d3:d6-
2-butene ratios were obtained which satisfy the restric—
tions placed on such studies mentioned earlier. In order
to follow this scrambling process the metathesis of
isomerically pure cis,cis-2,8-decadiene was initially
studied in some detail. The isomeric scrambling, or
isomerization, is then compared to the isotopic scrambling.
Results were also obtained from the 2,8-decadiene
metatheses that pertain to the initiation and termination

steps of the olefin metathesis reaction. At this point a

2O
discussion is presented on published results dealing with
these processes and the involvement of metal carbenes in
the reaction. This discussion will be followed by a
review of some of the important stereochemical studies of
metathesis reported in the literature.

As stated earlier, the results from the l,7-octadiene
studies indicated that the propagating steps in the olefin
metathesis reaction seem to involve a metal carbene odd—
carbon interchange. The involvement of carbenes in the
metathesis reaction has been implied by other recent
studies.

36

Cardin isolated a heteroatom stabilized rhodium

carbene 16 during the metathesis of the amines shown in

equation 10.

[Hi [Hr Hi

R

N
(P¢3)2Rh=< ]

H N

R

21

O'Neill37 and Dolgoplosk28 independently found that
diazo compounds decompose under metathesis conditions to
nitrogen and an olefin. Dolgoplosk also generated an
active catalyst from WC16 and phenyldiazomethane and
proposed that the active catalyst is a metal carbene
species.

Casey has synthesized some non-heteroatom stabilized
tungsten carbenes and has studied their reactions with

38

olefins. The diphenyl tungsten carbene 1] reacts with
isobutylene to give a new olefin 29 and a cyclopropane 1?

(equation 11).

¢
¢\ >L—j< [w< ]
=w(co) + ——b (b —D
¢/C 5 A 1:00
8

A? 1.,

An intermediate metallocyclobutane 18 was proposed as a
common intermediate for both products with the formation of
19 via a Chauvin type of pathway. No other olefins were
produced in this stoichiometric reaction implying that the

most substituted carbene is favored due to an electrOphilic

22
character of the carbene carbon.
Although this reaction is irreversible Katz39 has
reported that 17 is a clean catalyst for the metathesis of
many different olefins. However, others have been unable

to repeat this work.“0

More recently Casey“l has studied the reactions of
some unsymmetrical metal carbenes. Preparation of a
methylphenyl carbene %} was conducted at low temperature,

but when warmed to room temperature decomposed to styrene

and some cyclopropanes (equation 12).

¢ ¢ ¢ ¢ ¢ CH
(C0) w=c< ———-—> \-__— +\A( +‘A( 3 (12)
5 C CH3 ‘1’

H Room
Temp. A7%
21 26%

’b

Although it was first believed that the cyclopropanes
resulted from an intermediate metallocycle from the
reaction of the carbene with styrene, it was found that
external olefins are not incorporated in the products.
Thus 2} decomposes faster than reaction with olefins.
Casey postulated a different scheme that did not involve
metallocyclobutane intermediates. Caseyuz also reported
that the reaction of 23 with isobutylene (and other

olefins) forms almost exclusively the cyclopropane 2§ and a

trace of stilbene according to equation 13.

23

_/H
(CO)SW-C\¢ + )L 47) A +¢/=/b+ W(CO)5 (13)

a; 2‘3 (95%) (trace)

The difference in reactivity and product distributions
between the reactions of 17 and 12 with isobutylene has
been noted by Casey.L12 He has interpreted these results
as evidence that 13 must lose a CO ligand before olefin
coordination can occur because of steric effects. This
process affords a carbene olefin complex that can
reversibly form a six coordinate metallocyclobutane which

yields metathesis products (equation 1A).

.00 q) )k

(CO) w=C ——~ (CO) w=C’ —-—>
5 \¢ 1‘ \cb (emuw (1A)
17 \
“ ¢
\\_.
4/—

The less sterically hindered 22 adds an olefin without

N
loss of CO to form a seven coordinate metallocycle which
then eliminates W(CO)5 directly to give a cyclopropane

(equation 15).

2A

A
(CC)5w=C’fl -———-e- A >ZC§vp (15)
I I ____‘<;

 

¢ (co>5w—J(

22
w(co)5

Gassmanl43 has illustrated that cyclopropanes can be
converted completely to olefins by a metathesis catalyst.
He has also trapped a carbenoid species in very low yields
with several Michael acceptors. In contrast to Casey's
studies, these results imply a nucleOphilic character of
the metal carbene carbon.

Schrockuu has isolated many tantalum and niobium
alkylidenes. X-ray and NMR analysis of 25”5 Indicated that

there is a high barrier of rotation in the metal-carbon

bond implying the presence of a true double bond.

CH3
Cp2Ta/

§§CH2

2A
m

The chemistry of these metal carbenes most closely
resembles that of phosphorus ylides.“6 Propylene was found
to react in high yield with a neopentylidene %? to give a

new olefin in high yield (equation 16).

25

=/ Ta
Cp2Cl2Ta=CHC(CH3)3 -—9 -—9 M
25 CH3 95%

’b

=-/

26 (16)
'1;
CH
Ta 3
-——ap '\\
0%
26a
’b

However, this reaction is irreversible, and there are no
known tantalum or niobium based metathesis catalysts.
Schrock also says that the preference of 16 over 26a could
be explained by either electronic or steric arguments.

All of these model metal carbene studies have shown
that carbenes may initiate metathesis and have been reacted
with olefins to give products from a pathway believed to
involve metallocyclobutanes.

Much work has also been conducted in the study of

“8 has studied the

stable metallocyclobutanes. Green
formation of tungsten and molybdenum cyclobutanes formed by
reduction of a tungsten w-allyl complex (Figure 8). Treat-
ment of these metallocycles with acid yields the propylene
hydride complex shown. From these results Green has

postulated a scheme whereby an initiating carbene species

is formed from an olefin and the active catalyst (Figure 9).

26

 

2NC1 \ I _
[Cp WHLi]u ———:- pr ___,~. Cp w PF
2 2 W 2 6
dil HCl
NaBHu
+>\ _ HBFu
Cp2w\ BF)4 (f Cp2
H H2O

Figure 8. Formation of a tungsten metallocyclobutane

>____,>C\/H + >=>< 7/r\<__>(:3<

[M]?— —" -—‘ [M]- H

[M=C<]

Figure 9. Proposed formation of a metal carbene from
an olefin and an active metathesis catalyst

Puddephattu9 has reported the isomerization of a

platinum metallocycle which forms an equilibrium mixture in

the relative amounts shown in equation 17.

C1 C1
Py Py
>P = Py>Pt ¢
Py I
C1 C1

27
He also reports that the data is most consistent with a
concerted process without formation of an intermediate
carbene complex. Thermolysis of these metallocyclobutanes
yields quantities of propenylbenzenes and phenylcyclo-
propane. Styrene or ethylene was not observed from a
metathesis process.

The platinum metallocyclobutanes have been character-
ized by x-ray crystallography. These results indicate that
the ring is puckered significantly and that there is a
tetrahedral distortion around the platinum center.50

Another group51 has studied the in situ formation of

tungsten metallocycles from the reaction of tungsten halides

with the dianion shown in equation 18.

¢ ¢ ¢ ¢
wc1 + ¢>!/\sfi¢‘--*’ $K<A>>H '—“—" $¢C§k¢
6 w
C114

+ (18)

¢>p=CH2
¢
The metallocycles decompose to give mostly cyclopropanes
and quantities of 1,1—diphenylethylene, the expected
metathesis product.

Although most of the metallocyclobutanes described
above yield cyclopropanes in varying yields, cyclopropanes
are never found in significant quantities in metathesis
systems. According to Gassman's work, if cyclopropanes are
formed, they are converted to olefins by some other

process . “3

28

Mango52

has recently published a paper in which he
used a classical thermodynamics argument to question the
complete validity of the Chauvin scheme. He assumes that
the proposed pathway is an equilibrium catalytic reaction.
In other words, the distribution of products should be
based only on the difference in the respective free energies
of the products and reactants. If olefin A metathesizes

to a new olefin B, then cyclopropanes should also originate
from a common intermediate in the Chauvin scheme. Mango
argued that since the differences in free energies between
cyclopropanes and olefins are not great a significant
amount of these cyclopropanes should be found in the
reaction mixture. He concluded that whereas metal carbenes
do react with olefins to give cyclopropanes and some
metathesis products, these model studies only resemble the
actual chemistry of olefin metathesis.

Judging from the results of Casey it can be argued
that under metathesis conditions only six coordinate
metallocyclobutanes are present. These decompose to give
olefins exclusively. The formation of cyclopropanes would
arise from some other intermediate such as a seven
coordinate metallocycle which may not be available during
metathesis. This would invalidate Mango's first assumption
that cyclopropanes and olefins arise from a common

intermediate.

29

There has been some evidence presented that implies
that preformed metal carbenoids may arise from a catalyst
mixture and act as the metathesis initiating Species. It
should be noted that most of the non—supported systems
require an alkylating co-catalyst (Table 1). It is
generally believed that the alkyl group from the main group
metal co-catalyst is transferred to the transition metal,
and a carbene is eventually formed.

53

Muetterties examined several tungsten hexachloride
based catalysts. In particular he found that when WCl6 is
treated with dimethylzinc, methane is evolved. When a
deuterated solvent is used no deuterium incorporation in the
methane is observed. Muetterties proposed that dimethyl-
zinc methylates the tungsten in two steps yielding the
dimethyl tungsten species 2]. This species undergoes a
rapid a-elimination followed by a reductive elimination of
methane to yield a methylene tungsten carbenoid 28 as shown

in Figure 10. IchikawaSu

also reported similar results
from other WCl6 based systems and found that the main group
metal must be capable of donating a carbanion to tungsten.
Muetterties53 doubts that a simple methylene species
such.as 28 is a likely metathesis precursor because of its
expected high reactivity and instability under metathesis
conditions. He proposed that some bridged bis-metal

species such as a? may be a more accurate description of the

initiating species.

30

IR
4’1‘cr’ \y
29
W

He noted that all metathesis catalysts involve halides and
most often contain chloride ligands. This bridged model
would better facilitate the required a-hydride elimination,

Muetterties claims.

[CluW—CH2] + CH“
28

’b

Figure 10. Formation of initiating metal carbene

Tebbe55 has also noted types of compounds similar to
a? with organo-aluminum and tantalum hydride interactions.
Benzce’56 studied the infrared spectra of several non-
supported catalysts in solution and found evidence for
Lewis-salt adduct formation between the transition metal

complex and the co-catalyst.

31
Farona57 reported the isolation of products which may
have resulted from the reaction of an initiating metal
carbene with an olefin. He studied the metathesis of

l,7-octadiene catalyzed by Re(CO)5Cl/EtAlcl The product

2.
mixture was found to contain, besides cyclohexene, catalytic
amounts of l-butene and l,7-decadiene. Farona proposed

that these products originate from the reaction of a three

carbon rhenium carbene 19 with one end of l,7-octadiene

(equation 19).

Re(CO)5C1 + EtAlCl2

—-» [Re=CHCH,,CH3]

30
q, _
E;

[Re=CH21 + <— [Re=CH2] +=A +
1 etc. [::::: 31 I |
’b

These products are formed initially along with the
propagating carbene 13 which continues on to produce
cyclohexene after the first catalytic turnover. The
formation of 19 was proposed to originate from the scheme
shown in Figure 11. Compound 13 was found in small amounts

in the product mixture.

32

 

 

1 i1
(CO)uReCO + C2H5AlCl2 ——=> (CO);4 e-C&A1Cl2
C2H5
- C H AlCl Cl AlCl
(CO)uRe- —0A1CI2 < 2 5 2 (CO)uRe- 2
i H 2H5
2 5
H2
(C0)uRe fi 3 [(CO)uRe=CHCH CH3]
I H 2
02H ORE
5| H- HR +
A10120 \J

Figure 11. Proposed formation of an initiating metal
carbene from a non-supported catalyst system
and l,7-octadiene

33

The stereochemistry of the olefin metathesis reaction
has been studied and some stereospecificity has been
observed. For example, the ring—opening polymerization of
cyclopentene yields a linear polypentenamer with >95% cis
double bond content.58 The metathesis of acyclic olefins,
however, generally shows moderate specificity only in the
early stages of the reaction. Initially, cis olefins give
cis products and trans olefins give trans products. As the
reaction proceeds, the products and starting material
isomerize, eventually yielding a thermodynamic mixture of
isomers no matter which isomer is originally present. For
example, Hughes59 found that metathesis of either pure cis-
2-pentene or pure trans-2—pentene eventually gives the same
equilibrium mixture of 2—pentenes, 2-butenes and 3-hexenes
(equation 20).

\../\

o -->CH CH=CHCH CH + CH CH=CHCH + CH CH CH=CHCH CH
r 3 2 3 3 3 3 2 2 3
~\4¢~\// t/c=A.5 t/c=2.9 t/c=6.6

(20)

However, in the early part of the metathesis of cis-2-
;pentene the trans/cis-2-butene ratio is less than one.

In most stereochemical studies of the olefin metathesis
.reaction, the change in the trans/cis ratios (t/c) is
plotted against time or percent conversion of starting
Inaterial. The latter plots are generally non-linear

overall, but are roughly linear until about AO percent

3A
conversion of starting material.l3’6O This section of the
curve is usually extrapolated to 0% conversion, and a t/c
ratio is obtained that reflects the stereospecificity
during the first catalytic turnovers. These values change
significantly with changes in the catalytic systems. In
general, molybdenum catalysts are more specific than

61 The specificity is also

60a t62

tungsten based catalysts.

the co-catalys and

60a

dependent on traces of oxygen,
the concentration of the catalyst. These factors have
made any systematic comparison of catalyst specificity
difficult. However, some general trends have been noticed,
and consistent results may be obtained from any particular
catalyst.

Several groups63 found that the olefin reactivity
decreases in the following order: ethylene>termina1 mono-
substituted>internal disubstituted>trisubstituted>>tetra-
substituted. This order simply reflects the ease with which
an olefin can complex with the active catalyst. Among
terminal monosubstituted olefins, the alkyl group length
does not affect the rate of metathesis.60b For example,
l-pentene will metathesize at about the same rate as
l-hexene.

The metathesis of internal olefins has been studied
extensively. It was found that cis isomers react faster

13221’603’62 Very pure trans olefins

than trans isomers.
will metathesize only after an induction period of varying

length. This induction period is lost when traces of

35

terminal or cis olefins are present.6u

Apparently trans
olefins are not capable of participating in the initiating
steps of metathesis.

There is good evidence that the isomerization of
internal olefins as shown in equation 20 involves a

19,21,60a,62 The isomerization process

metathesis reaction.
is referred to as a "degenerate" metathesis because no new
olefins are formedJ The normal "productive" metathesis
produces olefins with a different number of carbons.
Terminal olefins undergo degenerate metathesis much faster
than productive metathesis. This fact was found only by
labeling studies because this degenerate process will not
exhibit isomerization.65 Simple disubstituted olefins
undergo degenerate and productive metathesis at about the
same rate.66 This characteristic should then be expected
for the metathesis of 2,8-decadienes.

The difference between the degenerate and productive
metathesis processes has been explained using a metal
carbene scheme and is illustrated in Figure 12. When R' is
hydrogen (terminal olefins) the degenerate metathesis is
favored, or kd>>kp, and when R' is a linear alkyl then
kdzkp. The large difference in rates exhibited by terminal
olefins has been attributed to either the preferred transfer
<1f the alkyl group to the least hindered end of the olefin
or to the formation of a more substituted carbene.65 The

isomerization of internal olefins can be seen to arise from

the degenerate pathway. Any stereochemical study, such as

36
the 2,8-decadiene experiments presented here, must account

for these competing processes.

kp R R RCH=CHR
RCH=CHR ' + M=CHR = I I = +
M M=CHR'

 

RJ—M "—""..—— M=CHR + R ' CH=CHR

Figure 12. Degenerate and productive metathesis processes

Two different arguments have been presented to
describe the origin of the specificity in the metathesis
of internal olefins. Caseyu2 has discussed the selectivity
in terms of the steric interactions in an intermediate
puckered metallocyclobutane. By using the model shown in
Figure 13, where L represents a relatively large substitu—
ent and S a smaller substituent, Casey proposed that the
1,3-diaxial interaction must be minimized.

This model successfully predicts that cis olefins will
give cis products. The formation of a trans olefin must
arise from a metallocycle that has a larger 1,3-diaxial
interaction indicated by the asterisk in Figure 13. The
same model will predict that trans olefins will preferen-
tially metathesize to new trans olefins. This cyclobutane

model has also been used by Caseyul’uz to predict the

37
stereospecificity found in cyclopropane formation. Katz67
has used a similar model to explain the specificity

exhibited in ring-opening polymerizations.

cf

L

S

W H L
L -—#b ,._‘
zz’IE’llé’ H H S
S
l
/L

S + W C cis
H’ ‘H \S
k
013 \ H
H
W L L S
kl>k2 S H L
*S
trans

Figure 13. Metallocyclobutane model for explaining
stereospecificity

65 has chosen a different model that considers

Basset
only the approach of the olefin to the metal carbene as the
important contributor to stereospecificity. This model
focuses on the interaction of the empty metal pz orbital
with a w-bond of the approaching olefin. An intermediate
metallocyclobutane may still be involved, but is not impor-
tant for steric considerations. Basset's model, shown in
Figure 1A, illustrates the case where a cis olefin
approaches a metal carbene containing two ligands (X) that
are involved in determining the orientation of the approach

of the olefin. Basset assumes that the metal carbene is in

a fixed configuration where the alkyl group is opposite to

38
one of the ligands. The olefin will then approach in a way
that minimizes the steric interaction between the olefin
alkyl groups and the metal ligands. This model also
successfully predicts that cis products will be formed from
cis olefins. The same holds true for consideration of
trans olefin metathesis. Basset observed that when X is
changed from chlorine to the larger bromine atom the

specificity increases.

BK \\ HI \\H
C“ R1 \H ’I’C= “
ll ‘ X \C‘“ R’ ~Rl
X714“ «— II -——- +
R2/C’MH
{X H

Figure 1A. Predicted orientation of the approach of a cis
olefin to a metal carbene
Basset65 has used this model in a study of the

metathesis of cis- and trans-2-pentene at low conversion.
As seen in equation 20, 2-pentene, as well as other methyl
olefins, yield 2-butene as one product. This means that
these stereochemical studies can be compared with the 2,8-
decadiene studies presented here. Basset found that cis-2-
pentene metathesizes to give a trans/cis-Z-butene (t/c-Cu)
ratio of 0.78 at 0% conversion. The same specificity was
noted for the other product, 3-hexene. This slight prefer-

ence for the cis products indicates that the model shown in

39
Figure 1A exhibits only subtle steric constraints. The

specificity increases when heterogeneous catalysts are used
(t/c-Cu=.A), and this is attributed to surface phenomena.
The metathesis of trans-2-pentene yields an initial t/c-Cu
ratio of 2.0. All of these ratios were found using similar
tungsten based catalysts and are averages of several runs.
As the metathesis of internal olefins such as cis-2-
pentene proceeds the products and starting material
isomerize by the competing degenerate process shown in
Figure 12. Eventually the thermodynamically stable mixture
is obtained (equation 20). Basset observed that the trans/
cis-2-pentene ratio changes linearly with the t/c-Cu ratio
(or the t/c-3—hexene ratio) as the reaction proceeds. By
using a kinetics argument based on the carbene scheme, this
linear relationship was derived semi—empirically. This
derivation involved the consideration of thirty-two
possible interactions of the two possible metal carbenes,
M=CHCH3 and M=CHCH2CH3, with each isomer of the three
olefins present in the reaction mixture. Basset derived
rate equations for the formation of each isomer and made
estimates of the relative rate constants for each inter-
action from experimental data. He then used these rate
constants to calculate the relative amounts of t/c-2-
pentene and t/c-2-butene at a particular stage in the
reaction. As expected, the calculated relationship was
valid only at low conversions of starting material (~10%),

'when the competing processes are not as important.

A0
The studies of the metathesis of cis,cis-2,8-decadiene

presented here will be compared to the results Basset
obtained. The metathesis of cis,cis-2,8-decadiene is
expected to give cyclohexene, trans-2—butene and cis-2-

butene (equation 21).

(: -_’ .+\=/+\/\ (21)

cis,cis—2,8-decadiene

At the same time the starting material should also
isomerize and must be accounted for. Similar stereo-
chemical plots will be reported, and a kinetics model
derived in a way analogous to Basset's will be presented.
This reaction, however, will cut the possible competing
metathesis processes from thirty-two to only twelve because
it involves an irreversible reaction to cyclohexene. This
means that the 2,8-decadienes will eventually be completely
consumed, and only the 2-butenes will metathesize in a
degenerate isomerization process.

The results of the studies on isomeric scrambling will
be used to interpret the experiments dealing with the
deuterium labeled 2,8-decadienes described earlier
(Figure 9). The isotopic scrambling in the starting
material and products must be accounted for according to
the restrictions mentioned earlier concerning such labeling

studies. The scrambling of the isotopic labels should be

Al

an entirely analogous reaction as the scrambling of the
isomeric labels (isomerization) because they are both
degenerate processes. These results will give an indica-
tion of how accurate the sampling of the labeled 2-butenes
is in distinguishing between the pairwise and non-pairwise
schemes. The labeled starting materials were also
exploited for experiments concerning the initiation and

termination of the olefin metathesis reaction.

RESULTS AND DISCUSSION

Synthesis of Olefins

 

The study of the metathesis of 2,8-decadienes required
the synthesis of isomerically and isotopically pure com-
pounds. The synthetic scheme adopted for terminally
labeled d6-cis,cis-2,8-decadiene is outlined in Figure 15.
The first step involved the dilithiation of commercially
available l,7-octadiyne and subsequent treatment with fully
labeled methyl iodide. The labeled diyne was isolated and
characterized and then hydrogenated stereospecifically with
Lindlar's catalyst (5% Pd on BaSOu) in pyridine. The
material isolated was found to have a high cis content
(97 i 2%) and complete (299%) deuterium incorporation.

The corresponding dO—cis,cis-2,8-decadiene was
prepared in the same way except that unlabeled methyl
iodide was used. The dO-trans,trans-2,8-decadiene isomer
is commercially available and found to have >99% trans
content. Reduction of 2-decyne with Lindlar's catalyst

afforded cis-2-decene in high yield and specificity.

A2

A3

 

2n-BuLi
H—E-(CH2)u-E-H 0° ,5, [Li-E-(CH2)u-E-L1]
l 2CD3I,O°+R.T.
CD3-E—(CH2)u-E—CD3 (73%)
H2(l atm)

‘(,/’//5%/Pd on BaSOu

Pyridine, 3 hours

3 CD3

(82%)

Figure 15. Synthesis of d6-cis,cis-2,8-decadiene

Catalysts and Metathesis Reactions

 

The non-supported catalytic systems used in these
studies are typical of those described in the literature.
The Mo(N0)2Cl2(P¢3)2/Me3Al2Cl3 system, referred to as
catalyst A, was used extensively because of its high
selectivity for metathesis and relatively long lifetime.15
The mixture produced when Al/Mo = 7 and the concentration
of molybdenum is .01M,is homogeneous in chlorobenzene,
and is light brown. The solution is allowed to
"incubate", or stir without any substrate, for thirty
minutes at room temperature before adding the olefin. The
catalyst solution must be kept under argon at all times
since brief exposure to oxygen will destroy the activity.

The catalyst solution may be frozen with dry ice and be

stored for several days with only slight loss of activity

AA
upon thawing. While incubating, methane and traces of
ethylene and HCl are observed in the gas above the
solution. The olefin : metal ratio is typically 150-200
1, but may be higher if desired.

The other system studied extensively was the mixture
of WCl6/SnMeu,lu referred to as catalyst B. In this case
a Sn/W ratio of 2 is used, and the solution in chloro-
benzene is mostly homogeneous exhibiting a dark reddish-
brown color. Most W016 based systems are believed to be

18 Incubation of this

partially heterogeneous in nature.
catalyst lead to a decrease in activity, and so the olefin
was added prior to the addition of the co—catalyst, SnMeu.
Methane is detected over the catalyst solution, and in the
absence of an olefin, ethylene is found in trace amounts.
Olefin to tungsten ratios of 100-150 were used.
All of the metathesis reactions were conducted in

sealed tubes under argon. The reactions were continuously
monitored by withdrawing small liquid and gas samples and

analyzing via glc. Deuterium content of olefins was

indicated by mass spectral analysis.

Metathesis of cis,cis-2,8-decadiene

 

The metathesis of cis,cis-2,8-decadiene yields cyclo-
hexene and a thermodynamic mixture of cis- and trans-2-

butene in high yield and selectivity (equation 22).

A5

C: —> + CH3CH=CHCH3 (t/c=
A or B, room 2.9)

temp., 3-A hrs. 90% 90%

C: +@

t’c-Clo t,t-Clo

 

c,c-ClO

(22)
5% total

After only a few hours at room temperature conversions as
high as 95% could be obtained with either the molybdenum
(A) or tungsten (B) based catalysts mentioned before. The
yield of cyclohexene is usually about 90% while the
remainder of the cis,cis-diene is isomerized to the two
other possible 2,8-diene isomers. Only a trace of a 016
olefin arising from an intermolecular reaction between two
dienes was observed, and no other C10 isomers were found in
the product mixture.

As with the metathesis of other isomerically pure
olefins, the metathesis of this cis,cis-diene shows
moderate stereospecificity at low conversions giving
predominantly cis-2-butene. As the reaction proceeds,
trans-2-butene production increases until the thermodynamic
equilibrium is reached, or the t/c-2-butene ratio (t/c-Cu)
is about 2.9. During these reactions small gas samples
were withdrawn at several time intervals and analyzed for

68

the gaseous t/c-Cu ratio. Other gaseous products

observed, such as propylene, will be considered later. The

A6
t/c-Cu ratios obtained were plotted against time, and as in
other stereochemical studies, the curve is linear. The
curve for the cis,cis-diene metathesis using catalyst A is
shown in Figure 16. Extrapolation of this curve to
t==0 gives the t/c-Cu ratio at 0% conversion. In this
case, a value of about .25 i .03 is obtained which is
reproducible. Other olefins were metathesized and extra—
polated t/c-Cu values found. These values are given in

Table 3 along with examples of the work of Basset.

Table 3. Extrapolated t/c-Cu Values at 0% Conversion

 

 

Olefin Catalyst t/c-Cu (0%)
cis,cis-2,8-decadiene A .25 i .03
cis,cis-2,8-decadiene B .61 1 .0A
cis-2-decene A .21 i .03
cis-2-decene B .62 1 .0A
cis—2-pentene B .73 i .0565
trans,trans-2,8-decadiene A 2.0 i .2
trans,trans-Z,8-decadiene B 1.9 i .2
trans-2-decene A 2.1 i .269
trans—2-pentene B 2.065

 

 

 

A7

cacaomomoum.m
Imwoamfio no mflmmnpwpme on» wcfipso ofipmu osmosplmuo\p on» CH mwcmco .mH opsmfim

AmmpchEv mEHB

oma 03H OMH omH oaa OOH om ow as on om on om om OH
_._a_..___.____ o

 

 

 

A8

It is difficult to compare the results given here with
the work of others due to the complex dependency of the
specificity of metathesis reactions on the conditions
employed. Nonetheless, it was surprising to discover the
relatively high degree of specificity exhibited during the
metathesis of cis,cis-2,8-decadiene with catalyst A.
Initially it was assumed that this phenomenon was due to
the intramolecular nature of the reaction. However, the
intermolecular metathesis of the corresponding mono—olefin,
cis—2-decene, with the same catalyst exhibited about the
same specificity at 0% conversion which implies that the
double bonds in the diene are acting independently.

Differences in catalyst specificity can also be noted
in Table 3. Molybdenum catalysts are generally known to be
more specific than tungsten based catalysts,61 but differ-
ences in co-catalysts may also be involved. Basset also
used a molybdenum catalyst similar to the one studied here
and found a moderate increase in stereospecificity,65 but
this was not as dramatic a difference as shown here between
catalysts A and B. The specificity observed with catalyst
B is comparable to the results found by Basset using the
same catalyst for the metathesis of cis-2-pentene. He also
studied many other tungsten based systems, and all
exhibited about the same specificity. These results seem
to indicate that cis—2-pentene, cis-2-decene and cis,cis-

2,8-decadiene exhibit roughly the same specificity in the

A9
early part of the reaction. This further implies that the
length of an alkyl group on one side of a methyl olefin
does not markedly affect the stereospecificity.

The metathesis of trans,trans—2,8-decadiene was also
studied. These reactions generally displayed induction
periods varying from a few minutes to about thirty minutes.
Conversion to cyclohexene and 2-butene then proceeds at
approximately half the rate as the cis,cis-isomer.

Addition of a trace of l-hexene eliminated the induction
period, but did not increase the rate of conversion. These
results confirm other studies that concluded that trans
olefins react slower than cis olefins and that trans
olefins do not participate in the initiating steps of

metathesis.6oa

A preference for trans-2—butene was found
at low conversions, and after about 30% conversion the
thermodynamic mixture of 2-butene isomers was observed.
The metathesis of cis,cis-2,8-decadiene by using the
molybdenum based catalyst was studied in more detail.
Liquid aliquots were withdrawn in addition to the gas
samples as the reaction progressed. The liquid samples
were analyzed for cyclohexene and for isomerization and
conversion of the starting material. The changes in the
trans/cis ratios for the 2-butenes (t/c-Cu) and the 2,8-
decadienes (t/c-Clo) were plotted against percent conver-

sion of starting material and are shown in Figure 17. The

t/c-Clo ratio was calculated by considering the

50

Hwfipmume wcfippmpm mo coampm>coo mo
pesosm on» Op O>Hpmfimm HmHmOpme mcfipumpm cam mmcmpsnnm mo coapmufinmEomH .NH onswfim

coampm>coo w

 

OOH om om os om om o: om om OH
_ a n _ _ _ _ _ _ . O o
O 0 . H.
O
l N.
o o as
i m.
o D
U i a.
O D
D .. m.
nH . o\p

1 m

a . a.

a - a.

O

.3.
J 04
D 030.63 030.03 - HA
- m;

 

51

contribution of the three possible 2,8—decadiene stereo-
isomers to the total trans and the total cis content of the

C mixture.

10
It should be noted that these isomerization plots
are linear until about 50% conversion. The two processes
can be related by plotting t/c-Cu against t/c—C10 values
observed at the same time. This plot is shown in Figure 18.
The linearity of this plot implies that the two isomeriza-

tion processes result from similar reaction pathways.

It has been well established that the isomerizations
of olefins observed during metathesis reactions are
degenerate metathesis reactions. These were described
earlier in Figure 12. The intramolecular metathesis of
cis,cis-2,8-decadiene, then, involves several competing
metatheses. These reactions are the productive metatheses
of the 2,8—decadienes which yield cyclohexene and 2-butenes
irreversibly, and the reversible degenerate metatheses
which isomerize the double bonds in the diene and 2-butene.
Figure 19 illustrates these competing processes.

The measured t/c-Cu values can now be seen to arise
from a combination of three processes. Initially, only
cis,cis-decadiene is present which metathesizes to give a
A : 1 preference of cis-2-butene over trans-2-butene
(Table 3). As the starting material isomerizes to give an

increasing amount of trans double bonds, the ensuing

productive metathesis will afford an increasing amount of

52

oao o
- \s as
m:
o-o\a
cmmzpmn magma
Ofipwam
a one

OH
o-o\s

 

 

.mH

 

madmfim

53

Productive metathesis processes:

c:~o+\-/+vx
(X—ORM
CC—OHM

Degenerate metathesis processes:
1’ 1’
—— ‘

\-=/—-\/\

F

Figure 19. Competing metathesis processes during the
metathesis of cis,cis-2,8-decadiene

5A
of trans-2-butene. Also, once the concentration of the
2-butenes begins to increase they will undergo a reversible
degenerate metathesis eventually yielding the thermodynamic
mixture. The plots in Figure 17 and Figure 18 reflect
t/c-Cu values that are a composite of these three
processes. The latter plot will later be accounted for
with a kinetics model which considers all of these
processes.

The final plot, shown in Figure 20, for the metathesis
of cis,cis-2,8-decadiene indicates the relative amounts of
olefins in the reaction mixture at a particular point in
reaction time. It can be seen, for example, that the
production of cis-2-butene begins to level off late in the
reaction reflecting the dominance of the degenerate isomer-
ization to the trans isomer. The amount of trans-C10
begins to level off only after most of the cis isomer is
converted which reflects the slower rate of trans olefin
metathesis mentioned previously.

The metathesis of cis-2-decene with catalyst A
exhibits somewhat similar behavior. Figure 21 shows the
comparison of the t/c-Cu values obtained from the mono-
olefin and the diolefin metathesis reactions versus percent
conversion. The curve for the metathesis of cis-2-decene
is linear only until about 20% conversion, and this
linearity matches the plot obtained from the diene meta-

thesis during the same amount of conversion. The

55

< pm>HOOMO Sufi: OCOHOmOOpummemHo.mHo mo
mfimospmpoa on» wcfipsw mcflmmao OHHomom mo mCOHpmppsmocoo O>HpmHmm .Om mpswflm

Amouzcflev CEHB

Dom 02m ONN 00m O®H OOH 03H ONH OOH ow cm 0: ON
_ _ A 41 q d . . . II O
_s__ _ alas
fig 1

 

O x x x x 3
4
D

[34X

0 q 1 ON
D w D I om
4 1 o:
O
I om
1 cm
4 O 1 E
D sons 0 .. om
4 soup 0 .. cm
x 36..» Oi

O OHOIO

 

OOH

COHpfimoQEOO
&

56

wcmHomOOclw.mleo.mHo
can OsmooclmlmHo mo mHmocpmuOE one Eopm mOHpmp :0|o\p mo somedeoo .Hm opstm

COHmcH®>QOO R

 

OOH om om ON. OO cm on om om OH
_ _ _ _ _ _ _ A _ _ o
AVmcmomolmano 1 H.
HumcmHUmOwUImamlmHomeo
I. N-
m
d < 1 m.
D .
D4 D a
D 1 m. 1J
D 1o}
4 1 o.
D
I N.
D 1 m.
4
l m.
1 H.H
gmH

 

57

difference between the curves at higher conversions results
from the inter- or intramolecular nature of the reaction.
As with other simple mono-olefins, cis-2-decene will
metathesize to two new products, 2-butene and 7-hexadecene,
which can then undergo further metathesis back to the
starting material. All of the competing metathesis
reactions are reversible during mono-olefin metathesis
unlike those during diene metathesis (Figure 19). In
effect, that means that there are more ways for the olefins
to metathesize and thus more ways to isomerize. It should
be remembered that cis-2-decene is never completely
consumed, and the maximum conversion is dictated by the
thermodynamic equilibrium between trans- and cis-2-decene.

By using these stereochemical results it should be
possible to devise a kinetics model in a way analogous to
that described by Basset6S which will supply an explanation
for the origin of the relationships shown in Figure 18 and
Figure 20. The model is based on a non-pairwise metal
carbene scheme, the evidence for which will be presented in
the next section describing the isotopic labeling studies.
Only the propagating steps of the olefin metathesis
reaction will be considered at this point.

The first assumption to be made is that there is an
equal concentration of the two possible metal carbenes
formed during the reaction. This is because there seem to

be only small differences in steric effects exhibited by

58
most methyl olefins. The C8-carbene 19 irreversibly
cyclizes to form cyclohexene, cis- and trans-2-butene (CA)

and the 02-carbene 15 according to equation 23.

§\” RCH=CHCH3 RCHTCHCH3
_ T’ — 4? M=CHCH3
35
33 ca 311 r» (23)
m m letc.

(R = C7Hl3 or CH3)

The irreversibility of equation 23 arises from the
observation that cyclohexene does not metathesize. It was
also observed that no C16 olefin was formed from an inter-
molecular reaction between 19 and a molecule of decadiene.
This process is illustrated in Figure 22, where R is equal
to the remainder of the decadiene (or C8), and implies that
k>>k". It is then also reasonable to assume that this

relationship is valid when R = CH Another assumption

3.
made is that once 1? forms it will preferentially exchange
olefins to complex with the readily available intra-
molecular carbon-carbon double bond to form 1? rather than
exchange with an olefin from the bulk solution. This means

that k>>k'.

59

 

RCH=CHCH
A, . 3
..__ 1"
J + __
+RCH=CHCH3
33
(\1
(R = C7H13 or CH3)

k k"
—JA

_L

3A

 

Figure 22. Possible reactions of the C8 metal carbene 33
’\1

60
With the assumption that k>>k', one need only consider
twelve possible reactions of 15. Initially, only four
primary reactions are possible during the early stages of

the reaction, and these are shown in Figure 23.

 

 

R CH H k H C CH R
\=/ 3 + M=CH/C 3 10 g) 3\=/ 3 + M=CH/
35
”b
R CH k H C
E/ 3 + M=CH\ 2° 3» 3\=\: + M=CH/R
CH H
3 3
R CH H C k R CH H
\=/ 3 + 3\CH=M _.L> E/ 3 + M=CH/C 3

 

R\ /CH3 kAc R / 3
+ H=M > \= + M=CH
H3C/C H3

Figure 23. The four primary reactions of 35 with cis,cis-
2,8-decadiene m

Since the double bonds in the starting material react
independently, the rest of the decadiene is represented as
R. The four possibilities shown arise from the different
ways the olefin can coordinate with the carbene. This
scheme does not reflect on the actual mode of interaction
discussed in the introduction, such as the intermediacy of
a puckered cyclobutane, but is simply a kinetics model.

The first two reactions are the productive metathesis

61
reactions which yield the 2-butenes and carbene i} which
then cyclizes exclusively to cyclohexene and is not
considered further. The latter two reactions are degen-
erate reactions which can reform the cis-ClO isomer (k3c)
or isomerize to the trans isomer (kuc)' As the new olefins
are produced eight secondary reactions must be considered
and are shown in Figure 2“.

Rate equations can then be written for the formation
of trans—2-butene (t-Cu), cis-2-butene (c-Cu) and "trans"-
decadiene (t-Clo). The 02 metal carbene is simply written
as [carbene] for simplification. The rate of production of

cis—2-butene is derived from Figure 23 and Figure 2“ and is

shown in equation 24.

6c-Cu

 

at = [carbene](klc[c-Clo] + klt[t'ClO] - k6CEC-Cu]

+ k5tEt—Cuj) (2a)

Similar equations are derived for t-Cu (equation 25) and

t—Clo (equation 26).

 

6t-C
u _
6t - [carbene](k20[c-Clo] + thEt-Clo] + k6CEC’Cu]
- kstEC-Cu3> (25)
Gt-Clo
—Et——- = [carbene](kuc[c-Clo] - (klt + k2t + k3t)[t-Clo])

(26)

H C CH

3\=/3

Figure 24.

M=CH

62

H3Q\==;H3

CH3
H3C
R CH3
R\=‘
CH3
H3q CH3
CH3
H3

M=CH

M=CH

The eight secondary reactions of 35 with

trans-C

10

and trans-2-butene

’b

63

Equation 2“ can be rewritten as

 

 

 

 

 

dc-Cu [t-Clo] [c—Cu J
= [carbene][c-Clo](klC + klt——————— - k6C———————
5t [C’Clo] [C’Clo]
[t-Cu J
+ kSt———————J.
[0-010]
Inserting a = [C'CHJ/[C-CIO] gives equation 27.
dc-C [t-C JEt-C J
u = [carbene][c-C10](klc + a(k1t 10 u
6t [t-Cu JEC—Cu]
[t-Cu1) ) ( )
-k +k____ 27
6c 5t
[c-Cu]
In a similar way equation 25 can be rewritten as
equation 28.
(St—CLl [t-ClOJEt-Cu]
= [carbeneJEC-Clo](k2 + a(k2t
6t c [t-Cu ][c-Cu]
[t_C”]> > (28)
+ k - k ——————
6c 5t[c-Cu]

Dividing equation 28 by equation 27 gives an equation
representing the relative amounts of trans- and cis-2-

butene (equation 29).

su

 

 

 

 

 

 

( [t-ClOJEt-Cu] [t-Cu])
t-C k + a k + k — k
u 2C 2t[t‘Cu JEC‘Cu] 6C Sttc-Cu]
_ = E ][ J [ J (29)
t-C t-C t-C
lo M u
c-C k + a(k - k + k -—————0
“ 1° ltEt-C ][c-C J 60 5t[c-C ]
u M M
In a somewhat analogous way equation 30
[t-Cu][t-C ]
2 10
t"C (1k -0. (k +k2 +kt)
10 UC lt t 3 [c-CAJEt-Cu J
— [t-C JEt-C J [t—C ]
10 u u
c-C k + a (k - k + k ——————)
10 lo ltEt-Cu JEc—Cu] 6 5t[c-Cu]
= t-Cloa (30)
C-Cu

is derived for the relative amounts of trans- and cis-
decadiene.

The relative values of the rate constants can be
approximated from the experimental data presented earlier.
The extrapolated t/c-Cu value at 0% conversion found for
the metathesis of cis,cis-2,8-decadiene should correspond
to the relative rates of k2C and klc’ This means that
k2c/klc = .25, and similarly from the studies of the
metathesis of the trans,trans-diene, k2t/klt = 2.0
(Table 3). Since the rate of metathesis of the trans,
trans-diene is about half that of the cis,cis-diene,

equation 31 follows.

klt + k2t + k3t + kut = .5(klc + k2c + k3c + kuc) (31)

65
There is no way to find the values for k3C or kut from
these experiments. However, the size of the alkyl group
does not seem to have an effect on the reactivity, and so
k3c can be approximated as being equal to klc and kut set
roughly equal to k2t' At low conversions it was found that

the rates of production of trans-2-butene and trans-C
70

10
This observation implies that kuc

were about the same.
is approximately equal to k2c'

There have been only a few studies conducted
concerning the metathesis of trans olefins, but the
approximations that Rut/k3t = 2.0 and kut = k2t are not
unreasonable. Finally, because 2-butene has an inherent C2
axis, it may coordinate with a metal carbene in two degen—
erate ways to give the same product. This degeneracy
means that the number of possible alignments of 35 with
2-butene to give a particular product must be doubled
compared with the alignment of 3? with an unsymmetrical
olefin, or one end of 2,8-decadiene, leading to the same

product. By using this information the relationships

shown in equation 32 are implied.

= 2 = 2
kBC klc’ k6c k

20’ k5t 2klt and k6t = 2k2t (32)

The relative rate constants were calculated by using the
approximations made above, equation 31 and equation 32,
and by arbitrarily assigning k = 1.0. The results are

lc
shown in Table h.

66
Table A. Approximated Relative Rate Constants

klc = k3c = 1.0 k1t = k3t = .21

k2c = kuc = .25 k2t = kut = k5t = .A2
kSC = 2.0 k6t .8u

R60 = .50

At low conversions the concentrations of t-C10 and
t-Cu are about the same, and the t/c-Cu ratio is taken to
be .25. These values along with the values in Table A are
put into equation 29 and equation 30 to yield equation 33

and equation 3“ respectively.

 

 

 

 

t-Cu .25 + .50
= (33)
c-Cu l - .3Ua
t-C 25a - 21a2
10 = (3A)

Recall that a = c-Cu/c-Clo which is one way of
measuring the amount of productive metathesis. These
equations predict that the maximum amount of cis-2-butene
present will occur when (1 = l/.34 = 3.0. This value is in
rough agreement with the curve in Figure 20. It is also
predicted that the amount of trans-C10 will be maximized
when a = l/.2l = 4.76. The experimental agreement with
this value is not as good, probably due to the high con-
version needed to observe this occurrence. Most of the

reactions conducted gave inconsistent results after long

67
reaction times due to the loss of catalytic activity. The
equations also predict that the amount of trans-2-butene
will never level off, a result of the irreversibility of
the diene metathesis.

To calculate a curve for t/C’Cu versus t/c-Clo,
various values of a are fed into equation 33 and equation
3“, and the resulting ratios are given in Table 5. This
curve is plotted and shown in Figure 25. Figure 25 also
contains the experimental plot shown earlier in Figure 18
for conversions of less than 50%. The calculated curve
illustrates a linear relationship as predicted by this
model. The slopes of the two curves are in good agreement,
especially when one considers the number of assumptions
made for this calculation. As a check, the points found
experimentally in Figure 25 are related to the amount
of conversion, and the corresponding 0 value is found.
these a values are in general agreement with the calculated
curve. More accurate curves could perhaps be found by
varying the relative rate of trans and cis olefin meta-
thesis, since only a rough approximation was employed in

this model.

68

Table 5. Calculated t/c-Cu and t/c-Clo Values

 

 

 

a t/c-Cu t/c-C10
0 .25 0
.05 .27 .012
.10 .31 .024
.15 .34 .034
.20 .38 u045
.30 .44 .062
.40 .52 .077

 

In summary, a kinetics model was devised, based on a

non-pairwise carbene scheme, which successfully accounted

for the experimental results obtained from the intra-

molecular metathesis of cis,cis-2,8-decadiene. The number

of competing metathesis reactions to be considered was

reduced to twelve from the thirty-two needed in mono-olefin

studies. The next section describes studies involving the

metathesis of isotopically labeled decadienes and uses the

results of the "isomerically labeled" decadiene metathesis

reactions presented here.

 

69

afionoxp cam sonoxu

coozpmn mafiszOHumHoh Hmpcmefipmaxo map pew popmHSonO map mo comapmqeoo

oaono\p

NH. OH. mo. 00.
— _ _ w _

II... D Hmpcoefipoqu

.Illl. nu UmpmHSOHmo

no. mo.

 

 

.mm mczwfim

70

Deuterium Labeling Studies

Mixtures of d6- and dO-cis,cis—2,8-decadiene were
metathesized using catalyst A. At several stages of the
reaction, gas and liquid samples were withdrawn and
analyzed via glc. The cis-2-butene was isolated at several
t/c-Cu ratios by preparative glc and submitted for mass
spectral analysis. The starting material was analyzed for
isomerization, and a portion was isolated and also
submitted for mass spectral analysis. In this manner,
d3/d6 ratios were found for the cis-2-butenes and the 2,8-
decadienes. Table 6 lists the values obtained for the

experiment where dO/d6-cis,cis-2,8-decadiene = 1.21 + .02.

Table 6. Experimental Data from the Metathesis of a
l : 1.21 d : d Mixture of cis,cis-2,8-

 

 

decadiene 0
t/c-Cu d3/d6'04 d3/d6-Clo % Conversion
.28 i .04 2.35 i .1 0.0 i .02 5 t 2
.38 i .04 2.4 i .1 .12 i .02 15 t 3
.5u 1 .05 2.u 1 .1 .32 1 .03 us 1 3
.74 i .05 2.4 i .1 .60 i .05 60 t 5
1.2 1 .10 2.4 i .1 --— 83 i 4

 

 

 

Since the t/c-Cu ratios change in a linear relation-
ship in the early part of the reaction with respect to

conversion (Figure 17) these t/c values can be used as an

71

internal measure of the extent of metathesis, differing
from the l,7-octadiene labeling studies reported earlier in
which all probes were external to the system. In those
studies it was not certain whether the ethylenes isolated
were obtained before any of the possible equilibrations of
the isotopic labels could take place. Isolation after such
an equilibration would give ambiguous results. By plotting
the d3/d6-2-butene ratios against the t/c-Cu ratios it
should be possible to extrapolate the d3/d6 values to 0%
conversion, or t/c-Cu = .25. This extrapolated value
should reflect a number obtained during pre-equilibrium
conditions. The graph shown in Figure 26 yields a value
for d3/d6'C4 of 2.4 i .l. The calculated ratio for a
carbene scheme is 2.42 : l, and the expected value for a
pairwise scheme was calculated to be 1.90 : l at 0%
conversion, or when there has been no scrambling of labels.

A more precise study was conducted in order to follow
the scrambling of the isotopic labels in the starting
material. This process, which is illustrated in equation
35, is very similar to the isomeric scrambling (isomeriza-
tion) discussed in the previous section and shown in

equation 36.

72

mOHpmp zono\p LwHSOHpme pm mcmfivmomoum.mlmo
pom lop mo mpszfie m mo mfimoanme map mafipzc mCSHm> zoumo\mn ©o>pmmno

:

 

 

 

0|o\p
m.H o.H m. m. a. m.
_ _ _ a _ _ o
_
_ - To
_
_ 1 o.H
_ m
.l 0H
_
czam> mmfizafima prdfizoamo “—
_ n o.m
HI— ] HIV 3 a
_
a .:
COHmLm>coo :RQ:

.om opswfim

I©U\MU

73

._CHCD3 +M= CHCH3 -M= CHCD --CHCH3
"CHCD3 "CHCDB
D3—

 

 

d6'ClO d3-Clo (35)
CH3 H3q_
-— +M=CHCH -M= CHCH /’ CH
3_ L_lc 3 3
— F— H3: —
CH3 H3 CH3
(36)

Both processes are degenerate metathesis reactions. The
increasing d3/d6-C10 ratios shown in Table 6 result from
this scrambling process (equation 35). No other scrambled
products were detected which could have arisen from some
secondary isomerization process such as a metal-hydride
reaction. Theoretically, if the labels in the starting
material were completely scrambled, then a (10 : d3 : d6-
C10 ratio of (1.21)2 : 2(1.2l) : 1 would be expected from
this experiment. As seen in Table 6 the fully scrambled
d3/d6-Clo ratio of 2.42 : l is never approached even at
higher conversions. The curve of d3/d6-C10 versus conver-
sion is similar to the t/c-C10 versus conversion (Figure 17)
in that it is linear to about 50% conversion, and then the

rate of scrambling increases. At 45% conversion the

74
isotopic scrambling is (.32/2.42)(100) or 13% complete.
This slow rise in isotopic scrambling during the early
stages of the reaction is entirely analogous to the
behavior exhibited by the isomeric scrambling and is due to
the intramolecular nature of the reaction.

It should be recalled from the restrictions presented
in the introduction that if the labels in the starting
material are scrambled sufficiently the pairwise scheme and
the carbene scheme will predict a random distribution of
labels in the products. Calculations were undertaken for
the pairwise scheme at several d3/d6-Clo ratios from zero
to .5 and were accomplished by using an iterative scheme
similar to ones described in previous reports.3u In this
instance, however, the probability of a metal-butene
species interacting with the newly formed d3-decadiene is
added on in three more terms of each calculation. These
calculations are shown in more detail in the Appendix. As
the amount of d3-Clo increases, the amount of d3-2-butene
increases. When the labels have completely scrambled in
the starting material, or d3/d6-C10 = 2.42, then d3/d6-C4
will also be 2.42. The calculated values are shown in
Table 7. These ratios were derived by only considering the
butenes arising from the starting material and should be

valid at very low conversions.

75

Table 7. Calculated d /d6-Cu Values by a Pairwise Scheme

 

 

3
d3/d6-C10 Calculated d3/d6-Cu
0 1.900
.116 1.926
.196 1.933
.301 1.988
.519 2.014

 

 

To adjust these calculated values to be more realistic
the interaction between two butene molecules must also be
considered. This means that the total d3/d6'c4 ratio
according to the pairwise scheme will be made up of two
processes. The first process is the result of productive
metathesis of the scrambled starting material, shown in
Table 7, and the second process is the degenerate
scrambling between the butenes. When the concentration of
these butenes increases the latter process will begin to
contribute more heavily to the total d

/d6-C ratio which,

3 4
of course, implies that the values in Table 7 are too low.
Required now is an approximation of the extent of the
independent scrambling of the isotopic labels in the 2-

b t t f - .
u enes a each new value 0 d3/d6 Clo

 

76
The experimental evidence given earlier, and in other
reports, indicates that size of the alkyl groups in simple
disubstituted olefins does not markedly affect the rate of
metathesis. This means that the isotopic scrambling in the
starting material (equation 37) can be related to the
scrambling in the butenes (equation 38) in a way similar

to that presented in the previous section.

 

C7CH = CHCH3 k C7CH = CHCD3
+ ——-> + (37)

X = CHCD3 X = CHCH3

CH3CH = CHCH3 k' CH3CH = CHCD3
+ > + (38)

X = CHCD3 X = CHCH3

The examples shown in these equations illustrate possible
interactions required for scrambling. In these cases the

X can be a metal for the carbene scheme or an alkylidene
for any pairwise scheme. Because of the C2 axis inherent
in the symmetrical 2-butenes it is reasonable to assume
that k' = 2k (cf. equation 32). This means that when the
labels in the starting material have scrambled to an extent
determined by k in equation 37, then the 2-butenes will
have undergone twice as much scrambling. For example, when
d3/d6-Clo is .12 (Table 6), then the starting material has

undergone (.l2/2.42)(100) = 5% scrambling. The butenes

77
produced will have undergone approximately 10% scrambling
independently. The adjusted d3/d6-C,4 calculated values
are shown in Table 8. These were calculated by taking the
d3/d6’C4 values derived from the starting material
(Table 7) and allowing the butenes to partially scramble to
the random distribution according to the approximated

extent of scrambling. For example, when d /d6-C is .12,

3 10

10% of the corresponding d3/d '04 ratio from Table 7 (.193)

6
is allowed to scramble completely and then is added back

into the remaining 90% of the d /d6-Cu ratio (1.926 -

3
.193 = 1.733) that arises directly from the starting

material by the productive metathesis.

Table 8. Adjusted d3/d6_c4 Values Calculated by a Pairwise

 

 

Scheme
% Isotopic
Scrambling
d3/d6-Clo d3/d6'c4 C10 C4 Adjusted d3/d6-C4

0 1.900 0 0 1.900
.116 1.926 5 10 1.970
.196 1.933 8 16 1.997
.301 1.988 12 24 2.070
.519 2.014 21.5 43 2.145

 

 

 

78

The adjusted calculated d3/d6-CA values from Table 8
are plotted against the d3/d6-ClO ratios, and the curve is
shown in Figure 27 along with the experimental values from
Table 6. It was shown earlier that the rate of isomeric
scrambling begins to increase after about 50% conversion.
It should then be expected that the isotopic scrambling
will behave similarly meaning that this calculated
pairwise curve will not be as accurate after about 40-50%
conversion of starting material, or when d3/d6-Clo is about
.35+.4. After this point the approach to the random
distribution of isotopic labels will change. The
calculated carbene curve is, of course, a horizontal line
at d3/d6 = 2.42.

Figure 27 has important implications. The difference
between the two calculated curves clearly indicates that
if the butenes are sampled at conversions 540% of starting
material, the values obtained for the isotopic distribution
will, within experimental error, unambiguously distinguish
between the two possible schemes. There are two data
points well within this limit, and both indicate that a
carbene scheme is operative. Another experiment was
undertaken with the initial do : d6'C10 ratio of 2.68 : 1,
and a large sample of 2-butenes was isolated after about
5-6% conversion. In this case, because of the larger 2-
butene sample isolated, a (10 : d : d6-Cu ratio was

3
obtained. This ratio is given in the experimental section

79

moEmsom mmfizpfima paw Amcmnmmovmwfizpfiwauco:

03p pom mosam> popmHSono map 0» memQEov mmsfim> nonmC\mU Um>pomno .mm mpswfim

 

 

oaouoexmo
w. No 00 mo :0 m. No H. O
14 . . _ _ _ d _ o
L m.
IIII mmfizpfimalco:
0 mmfispfima
”popmasoamo . o.H
D Hmucmefihoqu
soums\me
. m.H
coflmgo>coo mom
0 O O mvo.m
O
D u D JD.
1 m.m

 

80
along with the calculated ratios for both schemes. The

carbene scheme is again implicated. The previous
experiment (Table 6) was continuously monitored and
allowed only small samples to be taken, and because of the
errant background in the mass spectrometer only the
d3/d6'c4 ratio was found.

These isotopic labeling studies have taken advantage
of the isomerization studies presented earlier to allow a
calculation of a pairwise curve that accounts for the
scrambling of the isotopic labels in the products and
starting material. These processes must be considered in
any such labeling study according to the restrictions
discussed in the introduction. The experimental evidence
given here has indicated that a non-pairwise metal carbene
mechanism is a more accurate description of the olefin
metathesis process rather than any pairwise scheme.
Although other studies have reached the same conclusion,
the studies presented here most accurately account for all
of the competing metathesis processes. Values for the
isotopic distribution in the produCts were obtained that
were known to have arisen during pre-equilibrium
conditions. These experiments satisfy the restrictions

placed on such labeling studies.

81
Initiation and Termination of Olefin Metathesis71

As mentioned earlier, small amounts of methane and
ethylene are produced when Mo(N0)2Cl2(P¢>3)2 is treated with
Me3A12013 or when WCl6 is mixed with Sn(CH3)u. When a
methyl olefin, in particular 2,8-decadiene, is added to
either catalyst solution propylene is initially produced
followed by the usual metathesis products such as 2-butene.
The propylene could arise from an isomerization of the
methyl olefin to a terminal olefin which would then react
with a two-carbon metal carbene. However, this possibility
seems unlikely since no other C10 isomers were found and
since these catalysts are not known to promote such
isomerizations. A more probable source of the propylene
is from the reaction of a preformed methylene metal species
36 with one end of 2,8-decadiene. This process is shown in
Figure 28.

The propylene observed can be seen to be produced
directly by path a in Figure 28 along with the two-carbon
propagating metal carbene 3?. Path b is also possible and
forms 3? and a new diene 37. This new diene 3] should
cyclize rapidly through path c by interacting with 35.

This secondary process will also produce propylene. At
this stage the concentration of 3? should be much higher
than the concentration of 36, and so ethylene is not

expected. In fact, the ethylene observed is produced only

from the catalyst solution because the addition of the

82

+CH4 + (CH2=CH2)

MO(NO)2C12L2/MG3A12C13 (A)

or ——% [M=CH2]
WC16/SnMeu (B) 36
C—CHCH3
‘CHCH3

RCH=CHCH3

M=CH
a
b

R CH3
M 1:15
M
CH2=CH l

\CH

3 [::::i::2

2
H3

/CH3
—M CHCH3 + CH
I ll 35
—CHCH3 M w
33
l q. ‘6/[Mzi5 CHCH3]
[M CHCH31 +.
35
m letc. + CH2=CH

‘\CH3

Figure 28. Origin of propylene from the interaction of
36 with 2,8-decadiene

83
olefin does not increase its concentration. Terminal
olefins are known to be more reactive than internal
olefins. This observation implies that the terminal end
of a] should react exclusively with 1?. Compound 17 was
never observed in any of these studies.

The yield of propylene produced from the reaction of
2,8-decadiene and catalyst A was found to be about 90%
based on Mo using n-propane as an internal gas standard.
Using catalyst B the yield is only 8-10% based on tungsten
when this catalyst is allowed to incubate. The yield of
propylene can be increased, as shown later, but these facts
seem to indicate that the observed propylene is formed in
catalytic amounts.

Propylene was not observed when 2,8-decadiene was
treated with WCl6/Sn¢u, ¢w01 /AlC1 or Mo(NO)2Cl2(P¢3)2/

3 3
EtAlCl . Some propylene was observed when WCl6/SnBuu was

2
used, but this was suppressed by addition of n-propyl
acetate. This particular catalyst system is known to
promote some isomerization of starting material via some
pathway other than metathesis. ‘

Deuterium labeling studies were undertaken to
determine the source of each fragment of the propylene.
Because d6—2,8-decadiene was readily available it was
treated with either catalyst A or catalyst B, and the

propylene was isolated and analyzed via mass spectroscopy.

Both propylene samples exhibited very similar fragmentation

8M
patterns with an M+ peak of m/e=45. These spectra closely

resemble a known spectrum of 3,3,3-d -l-propene.72

3
Another study was then aimed at the labeling of the
co-catalyst. Several unsuccessful attempts were made to
prepare fully deuterated methyl aluminum sesquichloride.
Fully labeled tetramethyl tin was then prepared by reaction
of d3-methyl magnesium iodide with tin (IV) chloride. When
unlabeled 2,8-decadiene was treated with the WCl6/Sn(CD3)u
system the isolated propylene exhibited a fragmentation
pattern in the mass spectrum that is similar to the known
-propene pattern (m/e=uu, M+). Finally, d

l,l-d -propene

2 5
was isolated (m/e=u7, M+) when d6-2,8-decadiene was treated
with the same catalyst. These results confirm the scheme

in Figure 28, and equation 39

1/2 CD =CD
_ fl 2 2
WC16 + Sn(CD3)l4 T’[M-CD2]
CD“ -—CHC§3
*
-CHCH3 (39)
cfi
cg — *3
CH
D>C=CH/ 3 + ‘7— 3
D * M
[M=CHCH3]

follows the isotopic labels during the reaction of deuter-

ated catalyst B with labeled diene (indicated by asterisks).

85

The methane and ethylene produced from the mixture
of W016 and Sn(CD3)u were isolated, and both were found to
be fully deuterated. The origins of these species are
shown in equation 39. Methane production probably
originates from a reductive elimination of a dimethyl metal
species as shown in Figure 9. The fact that it is fully
deuterated indicates that a free radical decomposition,
which has been observed with other metal-methyl species,73
is not likely in these systems. The ethylene appears to
arise from a slow dimerization between two metal methylene
species. Similar dimerizations have been noticed by

”6 and Casey“2 (cf. equation 13). It should be

Schrock
recalled that ethylene production is absent when the olefin
is added prior to the co-catalyst. This means that the
metathesis reaction is much faster than carbene dimeriza-
tion. The dimerization process is one mode of termination
for the metathesis catalysts while most of the active
species are probably destroyed by oxygen or other
impurities in the system.

The activity of such methylene—metal precursor
catalytic systems may be optimized using these results.
The final yields of the catalytically produced products
(methane, ethylene and propylene) from the metathesis of

2,8-decadiene are shown in Table 9 for the two catalysts

and different conditions.

86

Table 9. Reaction of 2,8-decadiene Using Catalysts A and B
with Olefin/Metal = 150-170

 

 

Catalyst Conditions Yields Based on Metal (i 5%)
CH“ CH2=CH2 CH2=CHCH3
A Incubation 1 hr., 95% H% 85-90%
Reaction at room
temperature
B Incubation 20 min., 70% 2% 8-10%
Reaction at room
temperature
B No incubation, 75% 0% U5%
Reaction at room
temperature
B No incubation, 75% 0% 65%
Reaction at 50°C

 

 

 

The activity of these catalysts can be enhanced by
optimizing the yield of methane and propylene and
minimizing the production of ethylene. This may be done
with other, cheaper, 2-olefins and should allow rapid
optimization of the reaction conditions. This method may
also be used in a more detailed study of the differences
between methylene-metal species formed from different
catalytic systems. It was found, for example, that the
rate of formation of prOpylene is faster with catalyst A
than with catalyst B. Also, the rate of propylene
formation with either catalyst is slower than metathesis

indicating some differences in reactivity between 1? and 3b.

87
In conclusion, these studies have shown that a

preformed transition metal methylene species is most likely
responsible for the initiation of olefin metathesis with
some catalysts. The chemistry of this species appears to
be different from the propagating species and undergoes a
slow dimerization. This initiating species may be a
bridged bis-metal species such as g9, and this is presently

being investigated by others.

EXPERIMENTAL

General

All metathesis reactions were conducted in oven-dried
glass tubes sealed with rubber caps and were alternately
evacuated and flushed with argon several times on a vacuum
line. The solutions of reagents and solvents were stored
under argon and were transferred with oven-dried, argon-
flushed syringes.

Analytical glc was conducted with a Varian Series 1400
FID chromatograph with either a 20' x l/8" Durapak column
(A) at 60° or a 32' x l/8" 5% Carbowax 20M/Chromosorb W (B)
at loo-150°C. Preparative glc was conducted with a 10' x
1/4" 7% paraffin wax/Chromosorb w column at 80° for gas
samples and a 16' x l/H" 5% DC-SSO/Chromosorb G column at
120° for liquid samples, both using a Varian 90-P chromato—
graph. NMR analyses were obtained using a Varian T—60
spectrometer and 5 values were recorded relative to TMS.
Infrared spectra were obtained with a Perkin-Elmer 237B and
calibrated with a polystyrene standard. Mass spectra were
conducted with a Hitachi EMU-6E spectrometer.

The methyl aluminum sesquichloride (Me3A12Cl3) was
purchased from Ethyl Corporation and used as a 1M solution
in chlorobenzene. The light-green Mo(N0)2Cl2(P¢3)2 was

88

89

7“ and was

prepared according to the procedure of Cotton
stored in a desiccator. Tetramethyl tin and dl2-tetra-
methyl tin were prepared according to the procedure of
Edgell and Ward.75 The tungsten hexachloride was purchased
from Ventron and used as a .05M solution in chlorobenzene
under argon.

Chlorobenzene was distilled from CaH2 or P205 under
argon into a flask with a side arm with a two-way stopcock
and stored over AA sieves. A five percent forecut was
discarded during the distillation. Commercially available
t,t-2,8-decadiene (Chemical Samples) was stored over sieves
and found to be >99% isomerically pure by glc analysis with
column B. All olefins were activated prior to reaction by

passing through basic Woelm alumina under an atmosphere of

argon.

Preparation of l,1,1,10,10,10-d6-2,8-decadiyne

 

In a flame-dried, argon flushed 250 ml three neck
flask fitted with a gas inlet and a stir bar were placed
3.7 g (35 mmol) l,7-octadiyne and 50 ml distilled hexane
(CaH2). The solution was cooled in an ice bath, and 35 ml
of a 2.0M n-BuLi/hexane solution was added dropwise over
30 minutes with stirring. A white precipitate formed, and
after complete addition the slurry was stirred for a few
minutes at room temperature. The volatiles were removed in
gagug, and the remaining white solid dissolved in 100 m1

oxygen-free THF (benzophenone/Na). The flask was fitted

90
with a Friedrichs condenser and an addition funnel, and a
solution of 10.0 g (69 mmol) CD31 (Aldrich) in 25 ml
oxygen-free THF was added over 15 minutes at room tempera-
ture and then stirred overnight. The contents were poured
into ice-water, and the aqueous layer extracted with three
50 ml portions of n-pentane. The combined extracts were
washed with 50 ml dilute aqueous HCl (5%), 50 ml water
and 50 ml saturated sodium bicarbonate and were dried over
MgSOu. After filtering, the pentane was distilled off and
the remaining light yellow liquid distilled (b.p. 85-87°/
10 mm) to give 3.7 g (75%) glc pure product: IR (neat)
21uo cm—1(CEC), no 3300 cm-1(CEC-H); 1H NMR(CDC13) 5
l.5(m,H,-Cfi2-), 6 2.1(m,M,CEC—Cfl2—), no peak at 6 1.8 for

terminal methyl resonances.

 

Preparation of 1,1,1,10,10,lO-de—cis,cis-2,8—decadiene

In a dry 250 ml flask with gas inlet and exit were
placed 50 ml freshly distilled pyridine (BaO), 50 mg 5%
Pd on BaSOu (Ventron) and 3.0 g (21.3 mmol) of d6-2,8-
decadiyne. Dry hydrogen was then slowly passed through the
stirring mixture with a sintered glass frit for several
hours at room temperature. The progress of the reaction
was easily monitored by periodic glc analysis of the
reaction mixture with the 5% DC-550 column described before.
Upon complete conversion of the diyne, the reaction mixture
was passed through a short pad of Celite and the filtrate

concentrated to about 30 ml by carefully distilling off the

91

pyridine using a short Vigreaux column. The concentrate
was taken up in 100 ml ether and washed with 50 ml 10% aq.
HCl and two 50 ml portions of brine. The ether solution
was dried over MgSOu, and after filtering the ether was
removed by distillation. The remaining clear liquid was
distilled (80-83°/40 mm) to give 2.52 g (82%) product.
Analytical glc analysis using column B at 105° showed that
the total trans content was about 2 t .5% or 98% total cis
content. 1H NMR(CDC13) 5 1.6(m,u,—cg2-), 5 2.2(m,u,=CH-
CR2), 6 5.5(m,u,=cg); IR (neat) 1u5o cm'1(c=C), no CEC
stretch; mass spectrum (70ev) m/e (source) 1““ (m+), 126

(m+-CD3), 108 (m+-2CD3) (no significant peaks m/e 126+1MA).

Preparation of cis,cis-2,8-decadiene

In the same way described above 5.0 g (37.3 mmol)
2,8-decadiyne (Chemical Samples), 100 mg 5% Pd on BaSOu
and 150 ml freshly distilled pyridine were stirred under a
flow of dry hydrogen at room temperature. The product was
obtained in the same way as above (b.p. 78-82°/37 mm)
3.65 g (71%) clear product. GLC analysis (column B)
indicated 98 i 1% cis content. lH NMR(CDC13) 5 1.6(m,u,-
C§2-), 6 l.8(d,6,CflB), 6 2.2(m,u,=CH—Cfi2-), 6 5.5(m,fl,=C§);
IR (neat) luso cm'1(c=c), 725 cm'1(cis C=C), no 020, 965
(vw)(trans C=C); mass spectrum (70ev) m/e 138 (m+), m/e 123
(m+-CH3).

92
Metathesis of cis,cis—2,8-decadiene with Catalyst A

The procedure for the metathesis of cis,cis-2,8—
decadiene is typical of the procedures used for metathesis
reactions.

In an oven-dried 20 mm x 120 mm glass tube with a
magnetic stir bar was placed 10 mg (.013 mmol)
Mo(NO)2Cl2(P¢3)2. The tube was then fitted with a rubber
serum cap and alternately evacuated and flushed with argon
on a vacuum line several times. To this was added 3 ml
dry chlorobenzene and .1 ml of the 1M Me3Al2Cl3 solution
via syringe. The resulting clear brown solution was
stirred for thirty minutes at room temperature. Meanwhile,
.20 g (1.45 mmol) cis,cis-2,8-decadiene and 50 pl (.26
mmol) n-decane (internal liquid standard) were passed under
argon through 2-3 g basic alumina (Woelm) with the aid of
2 m1 chlorobenzene into a small vial. This solution was
then added to the incubated catalyst solution via syringe.
A gas solution of n-propane in argon, 5 m1 (.026 mmol) was
added with a gas syringe. Gas (~.2 m1) and liquid (~.3 m1)
samples were withdrawn at several intervals. The gas
samples were analyzed immediately on column A using
appropriate calibrations from standard samples. Liquid
aliquots were quenched with a few drops of water, passed
through Florosil into screw cap vials and stored at 0°

until glc analysis with column B.

93

Reproducible results could also be obtained by running
several small reactions simultaneously, quenching each at a
specified time with water and analyzing the gas and liquid

from the sealed reaction vessel.

Metathesis of cis,cis-2,8-decadiene with Catalyst B

In a 20 mm x 120 mm oven-dried, argon flushed tube
fitted with a rubber cap was placed .5 ml of the .05M WCI6
solution, .20 g (1.U5 mmol) activated cis,cis-2,8-
decadiene, 50 ul n-decane and 70 ul of a .75M SnMeu in
chlorobenzene solution in that order. Analytical glc
analysis was conducted in a similar way as above, except
that liquid samples were quenched with a few drops of a
dilute aq. NHMOH solution.

Metathesis of a 1.21 i .02:1 Mixture of doid6-cis,cis-2,8-
decadiene

 

This reaction was conducted by the method described
above for catalyst A. To the incubated catalyst was added

.153 g of a 1.21 1 .02:1 mixture of d and d6 decadiene

0
(determined by weight and mass spectral analysis). At four
time intervals, 10 ml gas samples were withdrawn and
replaced with an equal volume of argon. The gas was
injected into the preparative gas column mentioned above
and the cis-2-butenes isolated with a glass trap cooled in
liquid N2 at the exit port. The trapped butenes were then
submitted for mass spectral analysis. The d3/d6-Cu ratios

listed in Table 6 were found by comparing the peak heights

9A
at m/e = 59 (d3-Cu) and m/e = 62 (d6-Cu). Liquid aliquots
were withdrawn and the 2,8—decadiene isolated by prepara-
tive glc with the DC-550 column mentioned above. These
samples were analyzed via mass spectroscopy, and the peak
heights at m/e = 138 (do-C10), m/e = 1A1 (d3-Clo) and
m/e = 1AA (d6-C10) were compared. No other peaks in this
region were noted. The d3/d6-C10 ratios were found and are
recorded in Table 6. The t/c-Cu ratio was determined at
each gas and liquid sampling and are also listed in Table 6.

Metathesis of a 2.68 i .05:1 mixture of doid6-cis,cis-2,8-
decadiene

 

 

 

In the same way as described above, .120 g of a 2.68 i
.05:1 mixture of do:d6-cis,cis-2,8-decadiene was added to
catalyst A. After ten minutes two 20 m1 gas samples were
withdrawn from the tube and the reaction immediately
quenched with .5 m1 H20. The 2-butenes were isolated in
the same way and the collection tube submitted for analysis.
A d0:d3:d6 ratio of 1 : .72 1 .06 : .18 t .03 was found.
The calculated carbene ratio is l : .75 i .OA : .14 i .02,
and the ratio for the pairwise scheme is 1 : .6A t .02
.17 i .02. The t/c-Cu ratio was .31 i .0A which corre-
sponds to about 6-8% conversion of starting material.

Using n-decane as an internal liquid standard, the yield of

cyclohexene was found to be 5 i 2% based on decadiene.

95

Isolation of 3,3,3—d -1-propene

 

3
To a solution of catalyst A was added .15 g d6-2,8-

decadiene and the solution stirred overnight. The
propylene was isolated with the preparative glc gas column
in a trap cooled in liquid and submitted for mass spectral
analysis. Mass spectrum (70ev) m/e (relative intensity);
Found: A6 (87), A5 n+(1000), AA (913), A3 (696), A2 (391),
A1 (696); Literature72: A6 (29), A5 m+(1000), AA (956),
A3 (635), A2 (260), A1 (A96). Found (15ev): A6 (9A), A5
m+(1000), AA (170), A3 (9A), A2 (75), A1 (38).

A similar reaction was conducted with a solution of
catalyst B: mass spectrum (15ev) A6 (107), A5 m+(1000),
AA (285), A3 (107), A2 (71), Al (36).

Isolation of 1,1,3,3,3-d5-1-propene

 

To a solution of .013 mmol WC16 in chlorobenzene was
added .15 g d6-2,8-decadiene and .26 m1 of a 1M Sn(CD3)u
in chlorobenzene in that order. After stirring overnight,
the propylene was isolated as above. Mass spectrum (l3ev)
m/e (rel. intensity), A7 m+(639), A6 (203), A5 (1000),

AA (A53), A3 (81A), A2 (5A6).

Isolation of 1,1-d2-1-propene
Using the same deuterated catalyst as above, .15 g
dO-2,8-decadiene was metathesized and the propylene

isolated. Mass spectrum (50ev) m/e (rel. intensity);

96
Found: AA m+(750), A3 (1000), A2 (A29), A1 (786), A0 (21A);

Literature72 (70ev): AA (797), A3 (1000), A2 (370), A1
(366), A0 (506).
Isolation of du-methane and du-ethylene from a Solution of
WCl6 and Sn(CD3)4

A solution of .013 mmol WC16 and .026 mmol SnMeu in
chlorobenzene was stirred for two hours at room tempera-
ture. Using the gas preparative glc column methane was
isolated in a trap containing activated charcoal at liquid
N2 temperature. The ethylene was trapped in a long, spiral
glass tube at the same temperature. Methane mass spectrum
(18ev)76: m/e (rel. intensity) 20 m+(1000), 18 (620),
16 (80); Literature72 (50ev): 20 (1000), 18 (830), 16
(125); Ethylene mass spectrum (15ev): 32 (not oxygen)(1000),
30 (trace), 28 (120); Literature72 (70ev): 32 (1000), 30
(618), 28 (6A0).

APPENDIX

APPENDIX

PROCEDURE FOR THE CALCULATION OF THE
POSSIBLE d /d6-2-BUTENE RATIOS ARISING
FROM PAIRWISE MECHANISM

The procedure outlined here for the calculation of
the predicted isotopic distribution between the 2-butenes
(do‘CA=BO: d3—C4=B3, d6’CA=B6) arising from the metathesis
of a mixture of dO- (DO) and 96' (D6) 2,8-decadiene is
analogous to those reported by Grubbs3u and Katz.35 In
this instance, however, the interaction of the three
possible metal-butene complexes (M-BO, M-B3, M-B6) with the
scrambled starting material (D3) is also considered.

The model shown below illustrates schematically the
possible butenes and metal-butene complexes resulting from
a pairwise metathesis between any of the three butenes and
any of the three decadienes. As explained earlier this

calculation considers the case where k >>kex in Figure

met

3 because this yields the greatest amount of scrambling.

Possible B's Resulting M-B's

 

 

3
“ 2-3 1-A
l l-M-
1-3 2-A
A l 2
1-2 3-A

97

98

All of the possible combinations, where these four
numbers can represent labeled (CHCD3) or unlabeled (CHCH3)
ethylidene units, are considered, For example, the
metathesis between fully labeled decadiene (D6) and un-
labeled 2-butene (B0) is represented when 1=A=—CHCD3 and
2=3=-CHCH3. It can be seen from the model shown above,
for instance, that the probability of forming Bo from this
interaction (2-3) is 1/3 and the probablity of forming B3
(1—3 and 1-2) is 2/3. There is no chance of forming B6 in
this case. The probabilities for the formation of any of
the three butenes or the three metal-butene complexes may
be found for all of the nine possible interactions in Just
this way. These probability factors (Pn) for each species
are given below in a probability matrix generated from the

nine interactions.

P e
n valu s

 

a 2 11-2 Eli. 9.3. 91 11-131 M_-§2 8:9...
1 D6 M-B6 l O O l 0 O
2 D6 M-B3 1/3 2/3 0 2/3 1/3 0
3 D6 M—BO O 2/3 1/3 1/3 2/3 0
A D0 M-B6 1/3 2/3 0 0 2/3 1/3
5 DO M-B3 0 2/3 1/3 0 1/3 2/3
6 DO M-BO O 0 l O O 1
7 D3 M-B6 2/3 1/3 0 1/3 2/3 0
8 D3 M-B3 1/6 2/3 1/6 1/6 2/3 1/6
9 D3 M-BO 0 1/3 2/3 0 2/3 1/3

99

Using these Pn values the production of the metal-
butene complexes or the 2-butenes can be found by the

general equation A1.

d[B]

dt

or = Pl[D6][M-B6] + P2[D6][M-B3] + P3[D6][M-BO]
d[M-B] + PAEDOJEM-B6] + P5[D0][M—B3] + P6[DO][M-BO]

+ P7[D3][M-B6] + P8[D3][M—B3] + P9[D3][M-BO]
(A1)

The first step in the calculation for a particular
DO : D3 : D6 ratio is to find the relative M—B values using
equation A1. This is accomplished by using an iterative
scheme whereby any M-BO : M—B3 : M—B6 is initially assumed,
and these values are placed in the three equations for the
formation of the three metal-butenes. Solving these
equations yields new relative values for the complexes
which are then fed back into the three equations. This
procedure is repeated until a converging ratio is found.
The resulting relative M-B values are then put into the
three equations describing the formation of B0’ B3 and B6.
This affords the relative ratios for the labeled butenes
predicted by the pairwise scheme.

The following table gives the assumed DO : D3 : D6

ratio and the calculated values for the relative amounts

of metal-butene complexes and the butenes. The B /B6, or

3

lOO

d3/d6-Cu, values are also shown in Table 7.

 

 

Case Do : D3 : De M-Bo M-B, M281 gigs, R145,
1.21 : 0 : 1.0 1.33 1.15 1.00 1.896 1.A08

2 1.15 : .11: .95 1.32 1.23 1.00 1.926 1.389

3 1.11 : .18: .92 1.31 1.29 1.00 1.933 1.371

A 1.06 : .27: .88 1.31 1.36 1.00 1.988 1.365

5 .98 : .A2: .81 1.29 1.50 1.00 2.01A 1.338

The B3/B6 ratios calculated above do not account for
any independent isotopic scrambling of the butenes via a
degenerate metathesis process. To adjust these values a
portion of the relative amounts of each butene x[B], where
x is a fraction between zero and one, is allowed to scramble
independently to the statistical distribution [B]'. These
scrambled portions are then added back into the remainder
of the labeled butenes arising only from the scrambled
starting material. Thus, the adjusted relative butene
values [BJadj for each of the three butenes are given in

equations A2, A3 and AA.

[36],,J = [B61 - XEB5] + [861' (A2)
[B3Jadj = [B3] - X[B3] + [333' (A3)
[BOJadj = [BO] ' X[B0] + [BO], (AA)

The factor x is determined by the approximated extent

Of independent scrambling in the 2-butenes. This

101
approximation originates from the assumption that
k' = 2k in equations 37 and 38.
The total hydrogen content [BH] and the total
deuterium content [BD] in the portion of the butenes
allowed to scramble will not change. These values are

found by equations A5 and A6.
[EH] = XEBoj + X/2EB3] (A5)

[B

D] x[B6] + x/2[B3] (A6)

If [BHJ/[BD] = A, the statistical distribution between
the scrambled portion of 2-butenes will be A2 : 2A : l
V . 1 . t
for [BO] . [B3] . [B6] .
The total amount of scrambled material must be
equal to the total butene portion prior to independent

scrambling as shown in equation A7.
[BOJ' + [B3J' + [361' = x([B0 + B3 + B6) (A7)

Using equations A5, A6 and A7 a correction factor C can
easily be found,
([301 + [B31 + [361)x

A2 + 2A + 1

 

C =_

which will allow calculation of the [B]' values as

follows;

[B0]' = A2C ; [333' = 2AC ; [361' = c

102
These [BJ' values are substituted into equations A2,

A3 and AA to yield equations A8, A9 and A10.

[B6Jadj [B6] A x[B6] + 0 (A8)
[B3JadJ = [B3] — x[B3] + 2AC (A9)
[BO]adj = [BO] - x[BO] + A2C (A10)

Various values of x are approximated and the
adjusted [B3J/[B6] values are calculated. These values

(d3/d6-Cu) are given in Table 8.

 

LIST OF REFERENCES

10.
ll.
12.
13.

LIST OF REFERENCES

a) N. Calderon, E.A. Ofstead, W.A. Judy, Angew. Chem.,
Int. Ed. Engl., 15, A01(1976); b) J.C. M01 and J.A.
Moulin, Advan. Catal., 2A, 131(1975); c) R.J. Haines
and G.J. Leigh, Chem. Soc. Rev. A, 155(1975); d) R.H.
Grubbs in "New Applications of OPganometallic Reagents
in Organic Synthesis", ed. D. Seyferth, Elsevier Scien-
tific, N.Y., 1976, p. A23; e) T.J. Katz, Advan.
Organomet. Chem., 16, chap. 9(1977).

 

 

 

 

 

R.L. Banks and G.C. Bailey, Ind. Eng. Chem., Prod. Res.
Develop., 3, 170(196A).

 

N. Calderon, H.Y. Chen, K.W. Scott, Tetrahedron Lett.,
3327(1967).

G. Natta, G. Dall'Asta, I.w. Bassi, G. Carella,
Makromol. Chem.,gl, 87(1966).

 

a) K.W. Scott, N. Calderon, E.A. Ofstead, W.A. Judy,
J.P. Ward, Advan. Chem. Ser., 91, 399(1969); b) E.
Wasserman, D.A. Ben-Efraim, R.'Wolovsky, J. Amer. Chem.
sog., 90, 3286(1968).

 

H.W. Wittcoff, Chem. Tech., 75A(1977).

 

R.L. Banks, TOpics Current Chem., 25, 39(1972).

R.L. Burwell and A. Brenner, J. Mol. Catal., l, 77
(1975). m

a) C.P.C. Bradshaw, E.J. Johnson, L. Turner, J. Catal.,
, 269(1967); b) A.J. Moffat and A. Clark, ibid., 17,
64(1970). '———- m

 

R. Westhoff and J.A. Moulin, ibid., A6, AlA(l977).

D.L. Crain, ibid., L}, 110(1969).

 

J. Lal and R.R. Smith, J. Org. Chem., A0, 775(1975).
m

J.L. Wang and H.R. Menapace, ibid., 33, 379A(1968).
’b

103

1A.

15.
l6.
17.

18.

19.

20.

21.

22.

23.

2A.
25.

26.

27.

28.
29.

30.

31.

10A

P.B. van Dam, M.C. Mittelmeijer, C. Boelhouwer, J;
Chem. Soc. Chem. Comm., 1221(1972).

 

E.A. Zeuch, ibid., 1182(1968).
M.F. Farona and W.S. Greenlee, ibid., 759(1975).

P. Krausz, F. Garnier, J.E. Dubois, J. Amer. Chem.
Soc., 97, “37(1975)-
m

M.T. Mocella, R. Rovner, E.L. Muetterties, ibid., 98,
A689(l976). “

 

W.B. Hughes, ibid., .3, 532(1970).

Some exceptions: a) R. Nakamura and E. Echigoya,

Chem. Lett., 1227(1977); b) R. Nakamura, S. Matsumato,
E. Echigoya, ibid., 1019(1976).

N. Calderon, E.A. Ofstead, J.P. Ward, W.A. Judy, W.A.
Scott, J. Amer. Chem. Soc., 90, Al33(l968).
m

 

a) J.C. Mol, J.A. Moulin. C. Boelouwer, J. Chem. Soc.
Chem. Commun., 633(1968); b) A. Clark and C.J. Cook,
J. Catal., 15, A20(1969).

 

 

N. Calderon, Accounts Chem. Res., 5, 127(1972) and
references therein. W

 

F.D. Mango, Chem. Tech., 758(1971).

 

G.S. Lewandos and R. Pettit, Tetrahedron Lett., 789
(1971).

R.H. Grubbs and T.K. Brunck, J. Amer. Chem. Soc., 9&,
2538(1972).

R.H. Grubbs, A. Miyashita, M. Liu, P.L. Burk, ibid.,
100, 2A18(1978) and references therein.

 

 

 

B.A. Dolgoplosk, gt él-2 Eur. Polym. J., 10, 901(197A).
’b

K.w. Scott, Polym. Prepr. Am. Chem. Soc. Div. Polym.
Chem., 13, 87A(l972).
m

 

T.J. Katz and J. McGinnis, J. Amer. Chem. Soc., 99,
1903(1977). ”

J.L. Herisson and Y. Chauvin, Makrmol. Chem., 1A1, 161
(1970). ”

 

 

32.

33.

3A.

35.
36.

37.
38.

39.

A0.
A1.

A2.

A3.
AA.

A5.

A6.
A7.

A8.

A9.

50.

105

w.J. Kelly and N. Calderon, J. Macromol. Sci. Chem.,
Ag, 911(1975).

P.B. Marshall and B.J. Ridgewall, Eur. Polym. J., 5,
29(1969)- m

R.H. Grubbs, D.D. Carr, C. Hoppin, P.L. Burk, J. Amer.
Chem. Soc., 98, 3A78(l976).

 

 

 

T.J. Katz and R. Rothchild, ibid., 97, 2519(1975).

D.J. Cardin, M.L. Doyle, M.F. Lappert, J. Chem. Soc.
Chem. Commun., 927(1972).

 

 

P.P. O'Neil and J.J. Rooney, ibid., 10A(1972).

C.P. Casey and T.J. Burkhardt, J. Amer. Chem. Soc.,
96, 7808(197A).

J. McGinnis, T.J. Katz, S. Hurwitz, ibid., 98, 605
(1976)- m

 

Private communication between C.P. Casey and R. Grubbs.

C.P. Casey, L.D. Albin, T.J. Burkhardt, J. Amer. Chem.
800-, 99, 2533(1977).

C.P. Casey and S. Polichnowski, ibid., 99, 6097(1977).
‘———' m

 

P.G. Gassman and T.H. Johnson, ibid., 98, 6055(1976).
R.R. Schrock, ibid., 96, 6796(197A).

L.J. Guggenberger and R.R. Schrock, ibid., 97, 6578
(1975). m

R.R. Schrock, ibid., 2?, 5399(1976).

S.J. McClain, C.0. Wood, R.R. Schrock, ibid., 99, 3519
(1977). m

M. Ephritikine, M.L.H. Green, R.E. MacKenzie, J. Chem.
Soc. Chem. Commun., 619(1976).

 

R.J. Puddephatt, M.A. Quyser, C.F.H. Tipper, ibid.,
626(1976).

R.O. Gillard, M. Keeton, R. Mason, M.F. Pilbrow, D.R.
Russel, J. Organometal. Chem., 33, 2A7(1971).
'b

51.

52.
53.
5A.

55.
56.

57.

58.

59.

60.

61.

62.

63.

6A.
65.

66.

67.

68.

106

M. Ephritikine, J. Levisalles, H. Rudler, D.V.
Villemin, J. Organometal. Chem., 12A, 01(1977).

 

 

F.D. Mango, J. Amer. Chem. Soc., 9?, 6117(1977).

 

E.L. Muetterties, Inorg. Chem., If, 951(1975).

K. Ichikawa, T. Takagi, K. Fukuzumi, Transition Met.
Chem., 1, 5A(1976).

 

 

F.N. Tebbe, J. Amer. Chem. Soc., 95, 5A12(1973).

L. Bencze’, J. Pallos, L. Marko, J. Mol. Catal., 2,
139(1977). m

 

 

W.S. Greenlee and M.F. Farona, Inorg. Chem., 15,
2129(1976). W

F. Haas, K. Nfitzel, G. Pampus, D. Thiesen, Rubber
Chem. Technol., A3, 1116(1970).
m

 

W.B. Hughes, J. Chem. Soc. Chem. Commun., A31(1969).

 

a) J.M. Basset, G. Coudurier, R. Mutin, H. Praliaud,
J. Catal., 3A, 152(197A); b) G. Doyle, ibid., 30,
118(1973)- m ”

T.J. Katz and W.H. Hersh, Tetrahedron Lett., 585(1977).

 

J.L. Wang, H.R. Menapace, M. Brown, J. Catal., 26,
A55(l976). ‘—-——-—’ m

w.B. Hughes, Chem. Tech., A86(1975) and references
therein.

 

 

J.M. Basset, J. Amer. Chem. Soc., 93, 7376(1975).

J.L. Bilhou, J.M. Basset, R. Mutin, W.F. Graydon,
ibid., a9, AO83(1977).

C.P. Casey and H.E. Tuinstra, ibid., 100, 2290(1978)
and references therein. ”

S.J. Lee, J. McGinnis, T.J. Katz, ibid., 98, 7818
(1976)- m

Much of the 2-butene formed during these reactions
remains dissolved in the solvent and appropriate glc
calibrations were employed for the analysis of the
gaseous 2-butene.

69.

70.

71.

72.

73.

7A.

75.

76.

107

R.K. Stuart, unpublished results.

At conversions 10% the rates of formation of t-Clo
and t-Cu were found to be .15 i .0A mol olefin
mol catalyst”1 min-1.

R.H. Grubbs and C.R. HOppin, J. Chem. Soc. Chem.
Commun., 63A(l977).

 

E. Stehagen, S. Abrahamson, F.W. McLafferty eds.,
"Atlas of Mass Spectral Data", vol. 1, Interscience,
N.Y., 1969.

R.R. Schrock and G.w. Parshall,Chem. Rev., 76, 2A3
(1976). ”

F.A. Cotton and B.F.G. Johnson, Inorg. Chem., 3,
1609(196A). W

 

 

W.F. Edgell and C.H. Ward, J. Amer. Chem. Soc., 75,
1169(1953). ”

 

Some isotopic exchange with water in the mass spec-
trometer was observed.

"IA11111111711!"AAA