W .

w .

_ _
_

 

 

 

  

,

R

 

 

 

 

 

 

 

 

 

      

                

       

 

   

 

. .. .. ... 4 av..2.....:
. . r .. ‘ .9. 3.. I .V .H. .
.. v .H .... .5 ............
.. .m W... 0
-. e R.
. M ”“6”“ L...
.. , t .. D A fie
......e...
. . 127?»..1 . .. , .. 132:... .u .rm. tn... :wwfafllfc - .._.. ..,. .. Q”.- ~ . caviar. ..... . .-
... .. I... .. _ .t. . J: ... a». r 3.. .:........... I...:.:...b.::......ulnbw-:.nx:.e 5. 1...... avrmwfimwfiz mfi:«...§*§. . .I

 

n

.rr‘

1293 00848 3897

WWW tummmmflunmnrmna“ H, W

This is to certify that the
thesis entitled
POLYMER-SUPPORTED CATALYSTS

presented by

LeRoy Carl Kroll

has been accepted towards fulfillment
of the requirements for

Pth 1 degree in ChemiStrX

 

 

Date .AJ-lng't 9. 1974

0-7639

 

 
   
       

    
 
 
 

nuns & Suns
800K muaiavsuc '1
LIBRARY amoens ;
SPIIIOPOII. IICIIIGA] E
‘{A

  

 

L_____..'

 

:1th I 7 "f3
M0%}l§3

42‘

ABSTRACT
POLYMER-SUPPORTED CATALYSTS
By
LeRoy Carl Kroll

Homogeneous catalysts have been limited from devel-
oping into large-scale commercial chemical industry tools by
several factors. Two of these factors have been their diffi-
cult removal from solution and their tendency to form aggre-
gates of low solubility. It was proposed that the attachment
of ligands present in a homogeneous catalyst to an inert poly-
meric support might provide a means of relieving both of these
problems. If the supported ligand could be used to form
analogs of a homogeneous catalyst on the polymer. then the
filtration of the polymer would make removal of the catalyst
from solution simple. If the supporting polymer had a fairly
rigid matrix, then metal-metal bond formation or other types
of aggregation might be prevented.

Four questions summarize the primary interests of
this research:

1. Could a working analog of a homogeneous catalyst
be supported within a polymer matrix?

2. Would the support used produce predictable alter-
ations in the selectivity of a supported catalyst?

3. Could new, more active catalysts be synthesized
by supporting their precursors on.a rigid polymer matrix and

then activating them?

LeRoy Carl Kroll

4. Could chelation by the polymer-supported ligands
be controlled and/or prevented?

Styrene-divinylbenzene copolymer ”beads" were chosen
as the polymer support. Initial investigations used phos-
phination of this support via Scheme 1 to create a polymer-
supported ligand suitable for the complexation of Wilkinson's
Catalyst,3 Rhc1(P¢3)3. to the polymer. This complexation was

Scheme 1 SnClLPl
a

ClCHzOCHZC 3

 

 

successfully achieved by the eqilibration of phOSphinated
copolymer beads with Wilkinson's Catalyst, giving an active
catalytic polymer which could be reused many times without
serious loss of activity.q'5

Supported Wilkinson's Catalyst catalyzed the hydro-
genation of substrates of large size or high polarity (e. .,
Ag—cholestene or allyl alcohol respectively) at a much
slower rate than non-supported Wilkinson's Catalyst. This was
attributed to the effects of the polymer support, which was
nonpolar and which disPlayed size selectivity due to its
pore-containing structure. The pore structure restricted the
diffusion of substrates into the interior of the catalyst
beads, where the majority of the catalyst was .

The supported titanocene system deve10ped by C.

6

Gibbons was used to reduce a variety of hydrocarbon sub-

strates, but oxygen-containing compounds were found to be

LeRoy Carl Kroll
inactive to hydrogenation with that catalyst.

The supported titanocene system was investigated to
determine if titanocene, which is normally a polymeric aggre-
gate in its homogeneous hydrogenation form, was activated by
attachment to a polymer support. Comparisons of the catalytic
reduction rate per mmbl of titanocene present indicated that
at least a 25-fold activation had occurred in supporting the
catalyst. This was attributed to the ability of the rigid
supporting matrix used (20% divinylbenzene) to keep metal
atoms apart, preventing aggregation and the accompanying loss
in activity.

The failure of supported titanocene to fix nitrogen

7 or Vol'pin-Shur8 condi-

to ammonia using either van Tamelin
tions would also support the belief that metal association
was uncommon in the supported system, because dimeric titano-
cene is necessary for nitrogen fixation to ammonia.9 Vol'pin
and Shur reported that monomeric titanocene was involved in
the formation of aniline by titanocene when phenyllithium was
reagent used to reduce Ti(IV) to Ti(II) .8 When a sample of
supported titanocene was run through Vol'pin-Shur nitrogen
fixation procedures with some homogeneous titanocene present,
an unusually large amount of aniline was formed. This seemed
to verify the concept that supported titanocene was largely
monomeric.

Chelation studies used phosPhinated copolymer equil-
ibrated with [RhCl(COE)2]2 (003 = cyclooctene) or with

RhC1(CO)(P¢ ) followed by analysis for free ligand. The
3 2

LeRoy Carl Kroll
studies tended to confirm the presence of chelation in 2%
divinylbenzene copolymer systems while 20% divinylbenzene
c0polymer supports restricted chelation due to the polymer

matrix's rigidity.lo

BIBLIOGRAPHY

1. K. W. Pepper, H. M. Paisley. M. A. Young, J. Chem. Soc.. A,
1253: 4097-

2. C. Tamborski. et a;., J. Qrg. Chem.. 22, 619 (1962).
3. C. O'Connor, G. Wilkinson, J. Chem. Soc., A, 1268, 2665.

4. R. H. Grubbs, L. C. Kroll, J, Amer, Chem, Sgc., 23, 3062
(1971).

5. R. H. Grubbs L. C. Kroll, E. M. Sweet. J. Macromol. Sci.-
ersln... 4,2. 101+? (1973).

6. a; C. Gibbons, M. S. Thesis. Mich. State Univ., 1972.
( b) R. H. Grubbs, et al.. J. Amer. Chem. Soc,, 25, 2373
1973 .

7. E. E. van Tamelin, et a;.. ibid., 82, 5707 (1968).
8. M. E. Vol'pin and V. B. Shur,in9"0rganometallic Reactions".

Vol. I. E. I. Becker, M. Tsutsui, Eds., Wilry-Interscience,
New York, N.Y., 1970, p 102.

9. R. H. Marvich, gt_al., J, Amer. Chem, Soc.. 9%. 1219 (1972).
10. J. P. Collman, et a1., ibid., 93, 1798 (1972).

POLYMER-SUPPORTED CATALYSTS

By

LeRoy Carl Kroll

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1974

To the glory of God
Not because He needs it.
but because the God
Who wants everyone to know Him personally (John 17:3)
is worthy of all glory.

ii

ACKNOWLEDGEMENTS

My sincere thanks and praise go to my preceptor.

Dr. Robert H. Grubbs. His thoughts. suggestions, encourage-
ment, and most of all his patience have been prime movers in
this work.

The contributions of many of the faculty members at
Michigan State, who have been of the utmost importance in
the development of my philosophy of science and its practical
application, are also gratefully acknowledged.

The character of "the Grubbs Group" or rather the
characters in the Grubbs Group, have provided me with unique
memories. From our early, pre-grant days of chemical and
equipment ”grubbing" to our present state of moderate af-
fluence, they have remained as bizarre as ever.

Many other grad students, undergrads, stockroom men,
and especially the secretaries (who run this whole place
anyway) were sources of much appreciated aid. I'd express
my thanks more individually. but there is a paper shortage.

Incredibly great thanks go to those many joint heirs
in.Christ who have directly helped in the typing of this
thesis or who would have if I had needed them. Agape strikes
again.

A special smile goes to the lady in white.

iii

TABLE OF CONTENTS

Page
INTRODUCTION............................................ 1
RESULTS AND DISCUSSION..................................12

Supported Wilkinson's Catalyst........................12
Supported Titanocene..................................23
Hydrogenation studies...............................23
Nitrogen fixation studies...........................28
Chelation Studies.....................................33
[RhCl(COE)2]2 studies...............................34
Rh01(co)(P¢'3)2 studies..............................39
GENERAL SUMMARY AND SUGGESTED FURTHER RESEARCH..........H3
EXPERIMENTAL.................o..........................46
Introduction..........................................46
Bead Syntheses........................................47
Supported Wilkinson's Catalyst Batches A and B......47
Supported WilkinsonNs Catalyst Batches C and D......#9
Supported Wilkinsonfis Catalyst Batch: E.............50
Benzyl-supported titanocene dichloride..............51
l-Phenyl-l-cycIOpentene.............................52
l-Phenyl-l.3-cyclopentadiene........................53
Phenyl-supported titanocene.........................54
Hydrogenations........................................58

Batch.A supported Wilkinson's Catalyst hydrogen-

ationsu............................................

iv

TABLE OF CONTENTS (Continued)

Page

Batch B supported Wilkinson's Catalyst
hydrogenations.......................................60

Batch C supported Wilkinson's Catalyst
hydrogenatlons.......................................6O

Batch D supported Wilkinson's Catalyst
hydrogenations.......................................61

Batch.E supported Wilkinson's Catalyst
hydrogenatlons.......................................62

Benzyl-supported titanocene hydrogenations...........63
Phenyl-supported titanocene hydrogenations...........65
Wilkinson's Catalyst hydrogenations..................66
lOfifiPalladium on Charcoal hydrogenations.............67
Homogeneous titanocene hydrogenations................67
Nitrogen Fixations.....................................68

van Tamelin fixation using benzyl-supported
titanocenecooo0000000000000co00000000..coco-00000000068

Vol'pin-Shur fixation using benzyl-supported
titanocene...........................................69

van Tamelin fixation using homogeneous titanocene....70
Vol'pin-Shur fixation using homogeneous titanocene...7l
Chelation studies......................................72
[Rhc1(coa)2]2stud1es.................................72
Rh01(c0)(P¢'3)2 studies...............................?5
BIBLIOGRAPHY.............................................78

LIST OF TABLES

Table Page

1 BATCH D SUPPORTED WILKINSON‘S CATALYST
COMPARATIVE RATES yg. WILKINSON'S CATALYST......15

2 BATCH E SUPPORTED WILKINSON'S CATALYST
COMPARATIVE RATES xg. WILKINSON'S CATALYST......18

3 BATCH E SUPPORTED WILKINSON'S CATALYST
SOLVENT AND SIZE EFFECTS........................21

4 BENZYL-SUPPORTED TITANOCENE HYDROGENATIONS......23
VOL'PIN-SHUR NITROGEN FIXATIONS.................30

PHOSPHINE BEADS CHELATION STUDIES
[RhCl(COE)2]2 EQUILIBRATION.....................36

7 PHOSPHINE BEADS CHELATION STUDIES
[RhCl(COE)2]2 EQUILIBRATION.....................38

8 PHOSPHINE BEADS CHELATION STUDIES
RhCl(CO)(P¢5)2 EQUILIBRATION....................41

vi

INTRODUCTION

Homogeneous catalysis by transition metal com-
plexes has been a recognized process since the early

1 and has recently experienced growing application.

1950's.
Heinemann has estimated that a three-fold increase in the
use of homogeneous catalysts in industry has occured in
the last ten years.2 Homogeneous catalysts have several
advantages over heterogeneous ones:
1. simpler mechanistic investigation.
2. greater selectivity.
3. better mass and heat transfer characteristics.
There is one great disadvantage of homogeneous catalysts
relative to heterogeneous catalysts. however, as pointed
out by Cotton and Wilkinson:3
For separation of the products from reactants
and catalyst, heterogeneous systems have great
practical advantage over homogeneous ones....
This problem of the separation of a homogeneous catalyst
from the solution in which it has served as a catalyst has
been the major hindrance to the greater employment of
homogeneous catalysts in commercial processes. Methods

which remove the catalyst almost invariably destroy its
activity. requiring a separate regeneration step. An

2

example of this is the Oxo Process, which uses the homo-
geneous catalyst HCo(CO)u. The Oxo Process is the most
widely used homogeneous catalytic process in industry
today.” This process converts alkenes to aldehydes in
high yields as shown in Scheme 1. To remove the catalyst
from the process stream, however. it must be precipitated
by treatment with alkali. then regenerated by acidifi-
cation and extraction with an organic solvent.5 This is
an unusually simple process, for most homogeneous cat-
alysts would not survive such harsh treatment, and most
could not be so readily regenerated after removal from

solution.

\—\CO-Co(H2)(C0)3 HCo (co)3

H2
fCO-Co(CO)3 Co(CO)3

\ /

CO (00)“

The Oxo Process

Scheme 1

3

New'research in the sciences has often been the
result of the observation of natural processes. The
research reported in this thesis is an example of newg
work stimulated by the consideration of natural cata-
lytic systems, the enzymes. Active sites on enzymes are
protected and isolated from the surrounding medium by
the general hydrophobic nature of the supporting protein.6
Substrates are able to diffuse into the active sites.
however. and then diffuse out again after reaction. .As
is well known, the activity of enzymes as catalysts is
very high -- the supporting protein does not seem to
hinder the ability of the active site to Operate, and may
even.enhance it to some extent by keeping deactivating
substances away.

Support of a homogeneous catalyst within a syn~
thetic matrix might duplicate some of the advantages of
an.enzyme. The supported catalyst might display its
former activity. and yet be modified by the nature of
the supporting matrix in such a way that a change in its
selectivity would be observed. A hydrOphobic matrix.
for example. might allow hydrocarbon substrates to reach
the catalyst but hinder polar substrates. If the matrix
was a material which could be easily filtered, then re-
moval of the supported catalyst would be a very simple
process which would cause no deactivation of the catalyst.

The idea of a supported catalyst is not new. of

course. but the concept of supporting a homogeneous

u
system.primarily;gijhin the support is quite new. .At
the time of this work's early stages. a few reports of
homogeneous catalysts supported primarily within a matrix
were published. Haag and Whitehurst at Mobil Oil Corp-
oration attached [RhCl(CO)2]2 to a diphenylphosphine-
substituted styrene-divinylbenzene copolymer and catal-
yzed the hydrogenation of l-hexene with it;7a The same
‘workers used a sulfonated ion exchange resin; Amberlyst
15. to support [Pd(NH3)u]++ in an ionic and a reduced
form. The resins were used to reduce alkenes and alkynes
in catalytic hydrogenations?b Lazcanc and Germain used
.Amberlyst.A27, a strongly basic ion.exchange resin. to
support a low concentration of [PdClu]". This was used
to reduce cyclohexene, styrene, and nitrobenzene. Both
ketones and aromatic systems were unaffected.8 They
also found that their catalyst could be reused at least
eight times without the formation of any metallic pal-
ladium.

For the research reported here. it was decided to
use styrene-divinylbenzene copolymer beads as the sup-
porting matrix for the preparation of enzyme-like cat-
alysts. Beads ranging in size from 30 to #00 mesh and
with a divinylbenzene content of two or twenty percent
were used. The divinylbenzene content is a good approx-
imation of the amount of cross-linking in the copolymer.
The cross-linking gives the copolymer a pore-containing
structure and thereby affects the ability of some

5
substrates to diffuse into the inner area of the beads.

Since many useful homogeneous catalysts have tri-
phenylphcsphine ligands, it was decided to use a similar
ligand to functionalize the polymer. The route proposed
for the attachment of this ligand is outlined in Scheme 2.
The copolymer was first chloromethylated. using the
method of Pepper. Paisley. and Young.9 and then treated
with lithiodiphenylphosphide. prepared by the ”direct
method“ cf’Tamborski.‘gt_al. 1°

SnCl ZzPLil
©‘O Tonga—Leg; WHzcl "THP‘ ®"O‘CI'Izwz

Scheme 2
Once phosphine ligands were attached to the c0polymer
beads. it was felt that the ligands could be used to dis-
place those of a labile metal complex in solution with the
polymer. giving a new metal complex attached to the beads.
Wilkinson’s Catalyst.11

first attempts because it is difficult to remove from sol-

RhCl(P¢é)3. was chosen for the

uticns without deactivating it. Furthermore. the mech-
anisms of its reactions have been studied in great detail.
Not only would the successful support of such a catalyst
open the possibility of reusing it many times. but it was
expected that if most of the supported catalyst was within
the interior of the polymer beads.12 then its selectivity

might also be improved. This effect might be of two types,

6

size and polarity. Because of the pore structure of the
polymer beads. a catalyst supported within the interior
of the beads might be inaccessable to large substrates.
such as steroids. while others of somewhat smaller size
might display differential rates of hydrogenation based
on their bulk. Because the copolymer is also nonpolar,
substrates which contain polar substituents. such as allyl
alcohol. might be reduced more slowly by the supported
catalyst than by the homogeneous form. It has already
been observed that a similar polymer excludes water quite
well because of its nonpolar nature.13

A possibility which develops, once one has estab-
lished the validity of the concept of a matrix supported
catalyst. is the creation of new. highly activated catalysts.
This is eSpecially true if matrix binding sites are dis-
tributed in a controllable manner. Transition metal com-
plexes must have a site of coordinative unsaturation in

1“ The formation of such a site of

order to be reactive.
unsaturation in many homogeneous complex systems leads to
dimerization or polymerization of the active species to pro-
duce insoluble precipitates. Such aggregates are much less
reactive than the original monomeric Species might have
been. By attachment of the precursor complex to a fairly
rigid matrix, it should be possible to form an active cata-
lyst without aggregation because of the mobility restric-
tions of the matrix-supported ligand to which the metal is

bound. Substrates in solution might then diffuse in to

7

react with the metal and then diffuse out again afterwards.
modeling the behavior of an enzyme system.quite closely.

.An example of a potentially useful complex which
tends to polymerize in solution is titanocene, Tisz Cp =
cyclOpentadienide anion). This is formed when the cor-
responding dichloride is treated with a reducing agent.
such as sodium naphthalide. a Grignard reagent. or an
organolithium reagent. C. Gibbons has successfully coord-
inated an analog of TiCp2012 to the styrene-divinylbenzene
copolymer beads using a benzyl linkage to give 1 on the
polymer.15 He found that treatment of the polymer-sup-
ported catalyst with butyllithium in hexane produces an
active hydrogenation catalyst for olefins. In the anal-
agous homogeneous system, Brintzinger has determined that
most of the titanocene is present as an insoluble polymeric
hydride.16'17 If the monomeric hydride complex is more ac-
tive as a hydrogenation catalyst than the polymer hydride.
then the supported titanocene may be much more active, pro-

vided that the support matrix keeps the active metal centers

a. a a. Hf. ”\

T1012 Tin Ti-H

3H
RN
M

8
apart. The active Species on the polymer may be either
g or 3. both of which are known for the decamethyl analog
of titanocene,17b

sible for steric reasons. The formation of 3 should be

a complex in which dimerization is impos-

reversible. leading to g in the presence of hydrogen, if
the decamethyl analog is a reasonable model compound. It
is possible that the formation of 3 might be less rever-
sible than the carbon-hydrogen insertion reaction in the
decamethyl complex. so that removal of the benzylic meth-
ylene group from a position adjacent to the cyclopenta-
dienyl ring might activate the supported complex more.

One goal of the research reported here was the
development of the synthetic route in Scheme 3. as a means
of testing the proposal that removal of the benzylic meth-
ylene group might activate the supported titanocene more.

   

Scheme 3
Two other goals related to the use of supported
titanocene as a hydrogenation catalyst were the verifi-
cation of the presence of most of the titanocene within
the interior of the polymer and the further investigation
of substrates which might be hydrogenated using this

9

catalyst. To achieve the first goal. samples of whole
and ground beads were used as hydrogenation catalysts.
If the majority of the catalytic sites lay within the in-
terior of the polymer. the grinding of the beads would be
eXpected to increase the rate of hydrogenation, because
the substrates would no longer have to diffuse through
the pore structure of the polymer beads to reach the cat-
alyst. Comparison with homogeneous catalyst samples
should also determine if there is an activation of the
catalyst by supporting it on a rigid polymer matrix. The
second goal. further evaluation of the substrates acti-
vated for hydrogenation by the supported_catalyst. can
be easily met by testing the desired substrates. Sub-
strates of interest include dienes. alkynes, unsaturated
ketones, and aromatic systems.

Titanocene has also been used as a nitrogen fix-
ation vehicle by several researchers. Dimeric titanocene

S}?b.18.l9 so it

seems to be required for these fixation
would seem that if the supported titanocene system pre-
vents dimer formation then no nitrogen should be fixed

by the supported system. This may be tested by using the
conventional procedures for nitrogen fixation and testing
for the formation of ammonia. While both the van Tamelin18
and Vol'pin-Shur nitrogen fixation systems require dimeric
titanocene for ammonia formation. Vol'pin and‘Shur19 have
reported that aniline is formed in their nitrogen fix-

ation.method from a monomeric titanocene species when

lo
phenyllithium is used to reduce titanocene dichloride to
titanocene. If a nitrogen fixation is attempted using
supported catalyst with some homogeneous titanocene
preSent, then the amount of aniline formed should in-
crease relative to the amount of ammonia formed.

The question of whether or not the polymer matrix
is rigid enough to prevent nearby ligands from chelating
supported complexes arose during our work. Several re-
searchers have worked on this problem with varied results.
Some have found that chelation or other interactions of
polymer-supported species occurs only when a high degree
of substitution, over 1 mmol per g of beads, or when the
substrates used are able to bond to sites over 10 Rapart
are presentf‘o'm’22 Others favor chelation as a major

factor even at 1 mmol per g of beads levels of substi-

tutione 23
A study of supported titanocene's nitrogen fix-

ation behavior. as mentioned above. would give some val-
uable information on this subject. However, other eXper-
iments more specifically designed to answer the question
of whether or not chelation is a major factor could be
run also. The treatment of phosphinated capolymer beads
with solutions of [RhCl(COE)2]2 (COE s cyclooctene) or
RhCl(CO)(P¢'3)2 followed by analysis of the solutions for
free COE or PCB should give a fairly accurate measure of
the amount of chelation on the polymer. provided that the
stoichiometry of the ligand displacement is known.

11
The primary areas of investigation covered by this
work. then, were (1) whether polymer support of homogen-
eous catalysts was feasible, (2) whether such supported
catalysts displayed altered selectivity compared to their
homogeneous analogs, (3) whether new. more active cata-
lysts could be synthesized using this approach, and (h)
whether chelation by polymer-supported ligands would be

a prObleMe

RESULTS AND DISCUSSION

Supported Wilkinson's Catalyst

The first research goal was the successful sup-
port of a known homogeneous catalyst, Wilkinson's Cata-
lyst, on a styrene-divinylbenzene copolymer matrix. The
attachment of a diphenylphOSphine moiety to this was car-
ried out using the method outlined in Scheme 2 (page 5).
Phosphination of the copolymer such that the product had
0.5 to 1.0 mmol of phoSphine per g of beads was sought.

In the first actual phosphinated beads prepar-
ation. 200-400 mesh Zfidivinylbenzene-styrene copolymer
beads were used. These were chloromethylated with SnCln
and (31015001120159 and were found to have 0.9 meq of
chloride per g of beads by analysis. The analytical tech-
nique used was to reflux the polymer in pyridine and then
determine the chloride content of the pyridine using stan-
dard Volhard analysis.24

by the method of Tamborski. gt_al.. was then used to phos-

LithiodiphenylphOSphide prepared

phinate the copolymer. Repetition of the chloride anal-
ysis procedure after phosphination resulted in only 0.05
meq of chloride per g of beads remaining on the beads. and
an elemental analysis indicated the presence of 0.62 mmol
of phosphine per g of beads.

12

13

A sample of the phosphinated beads was equili-
brated with an excess of Wilkinson's Catalyst in aceto-
nitrile for eight days. After removal of the solvent and
repeated rinses with fresh acetonitrile, the beads were
vacuum dried and tested for their activity as hydrogen-
ation catalysts. These 'Batch A' catalyst beads cata-
lyzed the hydrogenation of cyclohexene at a rate of
0.0“6 ml of hydrogen per minute.

In an attempt to create catalyst beads of a higher
level of activity, another method of preparation was tried.
More of the phoSphinated copolymer beads were used, but
this time they were treated with [RhCl(03H6)2]2 in a sol-
vent of THF-ethanol.25 After three days of stirring at
ambient temperatures, the golden-yellow beads were analyzed
and were found to have 0.127 mmol of rhodium per g of beads.
Hydrogenation tests were performed using l-hexene. cyclo-
hexene, and .gicholestene as substrates. The steroid was
used to determine if most of the catalytic sites were within
the interior of the beads. since it was expected that its
large size would restrict its ability to diffuse through
the pore structure of the beads to the catalytic sites. The
observed rates, which were 1.12, 0.173. and 0.003 ml of
hydrogen per minute for l-hexene. cyclohexene, and
[lg-cholestene respectively. seem to verify the idea
that most of the catalyst is within the interior of the
beads. In order to investigate the nature of the substrate

size selectivity with the supported catalytic system. a

14
larger batch of beads was prepared for use instead of
these 'Batch B' beads.

Some 200-400 mesh Zfidivinylbenzene-styrene beads.
chloromethylated by T. K. Brunck. were phosphinated with
lithiodiphenylphosphide. The beads had 1.0 meq of chloride
per g of beads before phosphination and 0.3 meq of chloride
per g after phosphination. so their estimated phOSphine
content was 0.7 mmol per g of beads. Catalyst 'Batch 0'
and 'Batch D' were prepared by the equilibration of por-
tions of these phosphinated beads with Wilkinson's Catalyst
in benzene. Batch C. which weighed about 2 g. was used
for only two cyclohexene reductions before it stopped func-
tioning. The reduction rates of those two reductions were
0.575 and 0.025 ml of hydrogen per minute.

Batch D. which weighed about 3 g. was used for a
number of reductions. with a variety of substrates. as
listed in Table 1. In these reductions. cyclohexene was
used as a reference substrate; it was reduced before and
after every other substrate to provide a reference rate for
the determination of the relative reduction rate of each
olefin. For Aa-cholestene. cyclohexene was injected into
the system after one day of steroid reduction measurements.
It was expected that the rates would parallel the size of
each substrate, if most of the catalyst was supported within
the interior of the polymer beads. The expectations seem
to have been.well fulfilled, for the reduction rates do in
fact closely parallel the size of the substrates. The

15

Shampoo 93. mm .3830 no Saomaammfioa :33 as.»

0|

cpssfls use moment»: we as no.0 fl .opssas non demons»: no as 3..

 

 

 

 

 

.m

as e.m em mxmnmvaosm .meeee so seesaw m m m

Wmmaveto 8.0 1033.3 $0 «0.0 eseemeaocolmq

0H.0 0m.0 m0.0 0:.0 meeeoooaoS0

0H.0 0H.0 00.0 0a.0 eceooeoeoaoso

0H.0 mm.0 mm.0 50.0 eceoeeapooua

0N.0 00.H NH.H Hm.m usesem.a

13.330 $0 W093 004 004 esoxecoaoso

mxmumvaunm — meeem m mumvausm _ meeem empmoe

moses.» ovsaomnm Roz moans o>3mdom opsuvmnsm

 

 

 

 

smwg<a<0 m.zomzHest
..mu mma<e onaopnmm m>Ha<m<msoo
aqueseo m.zomszAHz anamommom 0 moa<m

H amfia

w.

16
same reductions were run using homogeneous Wilkinson's
Catalyst to reduce the substrates to determine Which rate
differences were primarily due to the catalyst itself and
not the diffusion restrictions of the polymer. These are
also listed in Table 1. Table 1 includes both the relative
rates, using cyclohexene as the reference compound, and the
actual measured rates of reduction. The rate of cyclohexene
reduction given is the mean of six measurements, but when
the relative reduction rate of each substrate on the beads
was calculated, the cyclohexene rate used for comparison
was the mean of the reductions run immediately before and
after the other substrate. For the steroid, the relative
rate was determined by injecting cyclohexene into the
system while the steroid was being reduced and measuring
the new reduction rate. The reduction rate before injection
of cyclohexene was then subtracted, to give a cyclohexene
reduction rate of 0.309 ml of hydrogen per minute.

The data of Table 1 clearly indicate that size
selectivity is a major factor in the supported system under
investigation, with larger substrates increasingly less
reactive due to their inability to diffuse into the polymer
to the catalytic sites. This behavior is substantially
different from ordinary heterogeneous catalysts. as may
be seen when the relative rates of reduction of cyclohexene
and AZ-cholestene with the heterogeneous catalyst 10%
Palladium on Charcoal are compared. These relative rates

are also listed in Table 1.

17

The fairly low reduction rates of the Batch D
beads prompted the preparation of a larger batch of new
catalyst beads using supported Wilkinson's Catalyst, known
as 'Batch E'. These were prepared as the beads of Batches
C and D had been, but were of a 30-80 mesh size. so that
filtration Operations with these beads were simpler. The
improved handling characteristics of these beads may have
contributed to their greater rhodium content, which was
determined to be 0.33 meq of rhodium per g of beads.

A ten gram sample of the Batch E beads was used
in an extensive series of reductions with over 75 reductions
catalyzed without a major loss in activity. .After an initial
activity drOp from 3.7 ml of hydrogen per minute to 2.2 ml
of hydrogen per minute for 1-hexene. the beads lost only
10$ of their activity over the next 30 reductions. For
these reductions. which are listed in Table 2, the standard
substrate used was l-hexene. This was reduced before and
after each other substrate and the relative rates listed are
based on comparisons with these rates. Homogeneous Wilkin-
son's Catalyst was again used for comparison reduction rates.
The reduction pattern of the substrates used in this study
duplicated that of the substrates in the study using the
Batch D beads: substrates of larger bulk were reduced more
slowly than smaller substrates.

Among the alkenes listed in Table 2 are two with
hydroxy groups in their structures. allyl alcohol and

u-pentene-l-ol. These were chosen to further investigate

18

3.3de non scwonuhc mo as 0.10»... .ovssds non moment»: «0 as a:

.oSouson chum Cw osoo msowpossoh 90:90 Add .Hosmnvmnososson H.H u Pso>aom

0|

 

 

 

 

 

 

 

 

p
as 0.H pa nxmumvaoem_xmeeen so masses 0 0H m
.3..- m 0 . a .i..- on .0 W0-H-osepcem-0
an .0 00 .H S .H 00 .H .0. 0:380:
0m.~ 00.H 00.0 00.H H0-H-eceesem-:
mm.: 00.m 00.0 00.0 Hocooae Haaa<
00.: 50.0 00.0 00.0 ecooovoaua
mm.a mw.H 00.0 um.a ososnonsoz
00.~ 00.H 50.0 mn.0 esexee-a-asreesaus-m.m.m
0m.m 0a.m as.0 0H.H ecexeeoaoso
00.0 ma.~ 00.H 00.H esexoxua
nfimumvaosm . memom mfimumvaonm modem
moves oesaompm semi meson o>apmaom opmspmpsm

 

 

BmHH<H<o m.zomZH%HH3

.mn mma<m onaosnmm m>Ha<m<m200
0. Swiss 920sz3; anamfimsm n roses

N Hands

 

19

the selectivity effects of the polymer support on the
catalyst. If the nonpolar matrix exhibits substantial
interaction with the substrates, reduction rates of the
polar alcohols should be considerably slower than when
they are reduced by homogeneous Wilkinson's Catalyst.

Such an effect is observed for allyl alcohol, which is
reduced 25 to 509%slower than l-hexene using the supported
catalyst but at the same rate as l-hexene using homo-
geneous Wilkinson's Catalyst. For h-pentene-l-ol. very

little relative rate change is noted.

The effect of increasing the polarity of the sol-
vent mixture was investigated using 1a1 benzene-ethanol
as the solvent for bead reductions. Under these condi-
tions, 4-pentene-l-ol showed a substantial rate difference
from that of l-hexene. The l-hexene rate seemed to in-
crease by 107éwhile the fl-pentene-l-ol rate decreased by
50 fl. most probably due to a greater preference of the al-
cohol for the more polar solvent mixture. The l-hexene
would tend to migrate more into the polymer because of its
nonpolar nature, resulting in an increase in reduction rate.
The effects of polarity on substrate selectivity seem to be
less dramatic than those of size, but nevertheless a clear
effect is present.

One problem encountered in the use of the Batch E
beads was caused by their activity. It was found that for
some of the more reactive substrates, such as l-hexene.

the reduction rate observed was not limited by the rate

20
of difquion of the substrate into the beads alone, but
also by the ability of hydrogen to diffuse into the sol-
vent. This was detected when it was observed that the
rate of hydrogen uptake was dependent on the rate of
stirring. When the amount of beads used was reduced to
2 g. no changes in rates were observed when the stirring
speed was varied over its entire range, even with l-hexene
as the substrate.

A new set of reductions. using just l-hexene,
cyclohexene, and cyclooctene, was run using the 2 g sample
of Batch.E beads. These three alkenes were used because
it was felt that their sizes varied enough to indicate
trends in size selectivity. Solvents of pure benzene and
1:1 benzene-ethanol were used for all three alkenes. The
results are listed in Table 3. All substrates undergo at
least a 100% increase in reduction rate on going to the
more polar solvent mixture, indicating that the preference
of the nonpolar substrates for the nonpolar polymer matrix
is a larger factor than the size restrictions of the poly-
mer's pores. A drop in the relative rate of the two cyclic
alkenes compared to l-hexene does seem to have occured,
probably due to the pore shrinkage which occurs in the more
polar solvent.

Figure l is a plot of the relative reduction rates
observed for l-hexene. cyclohexene, and cyclooctene in the
data of Tables 1.2, and 3. As can be seen in Figure 1. the
general trend among the supported catalyst systems is toward

21

TABLE 3

BATCH E SUPPORTED WILKINSON'S CATALYST-a-
SOLVENT AND SIZE EFFECTS

 

 

nation Relative rate

 

 

Substrate Solvent Hydroge
used used rate-

l-Hexene CH 0.67
Cyclohexene CH 0.53
Cyclooctene CH 0.12
l-Hexene CH-EtOH 2.97
Cyclohexene CH-EtOH 2.19
-Cyclooctene CH-EtOH 0.36

1.00
0.79
0.18
#.43
3.27
0.5»

 

 

 

 

 

 

 

‘5 2 g sample of beads
1’- in ml of hydrogen per minute 10.05ml of hydrogen per min.
2 g Batch 8 +
5" 10 g Batch 8 o
2 g Batch - x;
2 g éEtOH = i
[4,.
3...
Relative
Rate
2.4
1..
‘l-Hexene Cyclohexene Cyclooctene
FIGURE 1

RELATIVE REDUCTION RATES. BATCHES D AND E

 

22
slower rates for larger substrates. Although the 10 g
sample of Batch E seemed to reduce cyclohexene a bit
faster than l-hexene. in general the reduction rates of
different substrates follow the eXpected pattern.

The greater selectivity of the beads. when the
solvent is more polar and the pores are therefore smaller.
can be seen in the increase in the slope of the curve in
going from the 2 g Batch E beads in benzene (x) to 1:1
benzene-ethanol (*) .
figlpggy of gtudies 9n suppgztgd Wilkinson's Cgtglyst, The
above research indicated that:

1. Support of a functioning homogeneous catalyst
on.a styrene-divinylbenzene copolymer matrix is possible.

2. Such a system has most of the supported cata-
lyst within the interior of the matrix used.

3. The supported catalyst displays size and polar
selectivity toward substrates which is not observed with
the homogeneous catalyst.

h. The supported catalyst may be reused many times
without regeneration. in contrast to the homogeneous equiv-

alent.

23
WW

fiydzgggngtign_stugigs‘ The attachment of dicyclopenta-
dienyldichlorotitanium (IV) . also known as titanocene

dichloride, to a styrene-divinylbenzene copolymer via a
benzylic linkage has been accomplished by C. Gibbons.15

He found that the supported complex could be reduced with
alkyllithium reagents and used as a hydrogenation catalyst.
A similar reaction had already been observed for the homo-

geneous system.16

Two further experiments were done with
these beads relating to their use in hydrogenations.

In the first study, the ability of the supported
titanocene to catalyze the hydrogenation of a variety of
substrates was tested. These are listed in Table 4. For
each of the reductions recorded. 0.015 1(1001 mmol.ofisup-

ported titanocene was used to reduce 0.5 ml of substrate.

TABLE 4
BENZYL-SUPPORTED TITANOCENE HYDROGENATIONS

 

 

 

 

Substrate Hydrogenation rateg
(m1 of Hz/ minute )

1".‘5-0 yclooctadiene 1. 5‘}
1.3-Cyclooctadiene 1.81

Styrene 2.05
leHexyne Polymer
3~Hexyne 1.26

Benzyl acetate No Rxn. (?)
Vinyl acetate N0 Rxn. (?)
3-cholestene-R-one N0 Rxn. (?)

 

 

5 1'0.10m1 of Hz/minute

2h

Both octadienes were readily reduced to their com-
pletely saturated counterparts. as was 3-Hexyne. These
three reductions were monitored by gas chromatography to
detect any isomerization or partially reduced interme-
diates present during the course of the reduction. No
isomerization or partially reduced products was detected.
The formation of a cloudy solution of polymer when l-Hexyne
reduction was attempted is understandable. since the homo-
geneous system is known to polymerize terminal alkynes.27
Styrene forms ethylbenzene quantitatively. The three other
substrates tested all contain oxygen and thus could quite
possibly deactivate the titanocene, since the reaction of
oxygen-containing functional groups with low-valent titanium

28 Titanocene might also be acting as a nucleo-

is known.
phile in these systems. leading to unreactive products with
the substrates.29 The former reaction is more common for
titanium. The homogeneous titanocene system has been used
to catalyze alkene, diene. and alkyne hydrogenation as well.
but it has been reported to polymerize styrene in an excess
of hydrogen.l6’27
One pr0perty of the homogeneous titanocene system
is its ability to reform TiCpZCl2 upon treatment with H01.17b
If such a reaction could be carried out with the supported
titanocene system, it could be reused. To test the pos-
sibility of this approach to making the supported titanocene
”recyclable", a sample of the pale grey active polymer was

exposed to air, at which exposure it immediately became tan.

25
the color of deactivated supported catalyst. This was then
treated with dry HCl gas while suspended in absolute ethanol.
The beads became pink (the color of the beads before acti-
vation) within five minutes. After hexane rinsing and drying
under vacuum, the beads were activated with butyllithium and
used for the reduction of a sample of l-hexene. The reduc-
tion rate observed on the first use of the beads was 1.75 ml
of hydrogen.per minute: the beads after recycling produced
a rate of 1.82 ml of hydrogen per minute. Regeneration of
the catalyst appears to be possible.

One other experiment done with the benzyl-supported
'titanocene system used sodium naphthalide to activate the
beads. This was to verify the usefulness of the same re-
ducing agents used for the activation of the homogeneous
titanocene system. A reduction using l-hexene gave a rate
of 0.45 ml of hydrogen per minute. Although less active
than the butyllithiumractivated beads. these beads were
still quite good as a catalyst.

Confirmation of the presence of most of the titan-
ocene within the interior of the beads. as was done for the
Wilkinson's catalyst beads. was desired. A different method
was used for the titanocene beads. however. If most of the
catalyst is within the interior of the beads, then sub-
strates will have to pass through the pores of the beads
before being reduced. As was observed for the Wilkinson's
Catalyst beads. such a passage limits the rate of reduction
because the diffusion into the beads requires time. If the

26
head structure is destroyed, however, a great rate in-
crease should be observed, because diffusion through the
pores to get to the catalytic sites will not be necessary.
Identical quantities of cyclohexene were reduced using sam-
ples of the same amount of beads with and without grinding
the beads in a mortar and pestle before activation. The
whole beads had a reduction rate of 88.7 ml of hydrogen per
minute per meq of'titanium: the rate for the ground beads
was 714 ml of hydrogen per minute per meq of titanium. The
belief that most of the supported titanocene was within the
beads seemed to be clearly supported by this result.

The question of whether or not titanocene was acti-
vated by supporting it on the polymer matrix was considered
and an answer sought by comparing the reduction rate on non-
supported and supported titanocene. Two reductions were
run using homogeneous titanocene with the same procedure
and quantities as had been used in the whole bead-ground
bead rate comparisons. The reduction rate observed was 28
m1 of hydrogen per minute per meq of titanium. This was
clearly much slower than the supported titanocene reductions.
A one to two hour induction period was also observed when
the homogeneous system was used. During this. the reduction
rate was only 20960f the maximum value. The supported system
had the same reduction rate from the beginning of catalyzed
reductions until less than 5%of the substrate was left. at
which time the rate began to dr0p off.

27

Since the possible presence of 3 (page 7) was felt
to be a way in which the benzyl-supported titanocene might
be deactivated. the synthetic process of Scheme 3 (page 8)
was applied to some 20%diviny1benzene-styrene copolymer
beads. The synthetic conditions were first developed using
non-polymeric systems because analysis of the products is
simpler when they are not on the polymer support.30 The
synthetic procedure worked for the non-polymeric attempts,
and was then applied to the polymer. Bromination of the
first batch of beads run.was done using the method afOlah?l
with FeCl3 and bromine in nitromethane. This yielded a
highly brominated polymer with over 3 meq of bromide per
g of beads. The polymer contained some residual iron salts.
however, which proved difficult to remove. After lithiation
and treatment with cyclopentenone, only 0.058 meq of titanium
per g of beads were successfully attached to the polymer.
When used as a catalyst, however. the beads proved to be more
active than the benzylic-supported titanocene beads. with a
reduction rate of 106 ml of hydrogen per meq of titanium
for cyclohexene reduction. This is a 13%increase relative
to the reduction rate found with the whole beads having
benzylic-supported titanocene. This preparation also has
the advantage of not requiring the use of chloromethylethyl
ether to functionalize the basic polymer. This ether has
been found to be a potent carcinogen.

Another preparation of ”phenyl-supported" titanocene

used the same lithiation and cyclopenteneone reaction

28
sequence as the preceding batch, but the bromination was
done using BF3 as the catalyst. A lower degree of brom-
ination was obtained, but the final product had a much
higher percentage of titanium substitution, 0.546 meq of
titanium per g of beads.
Nitrgggg fixatign studies. Homogeneous titanocene has
been used for nitrogen fixations by several different
workers. The method of van Tamelin uses sodium naphtha-
lide in THF with TiCp2012 to give ammonia which may be
trapped out of a nitrogen stream passed through the reac-
tion solution.18 In the method of Vol'pin and Shur, TiCp2C12
in diethyl ether is treated with an organolithium or
Grignard reagent under 100-200 atmoSpheres of nitrogen.19
Hydrolysis is necessary to form ammonia in the Vol'pin-Shur
system, and the formation of arylamines is observed when
aryllithium reagents are used to reduce titanocene dichloride.
Both nitrogen fixation procedures were run using the sup-
ported titanocene as the source of titanocene. Homogeneous
titanocene reactions were also performed to determine the
correct conditions for nitrogen fixation and to verify the
validity of the analytical methods used.

Researchers in titanocene nitrogen fixation are all
agreed that dimeric titanocene is involved in the formation
of reduced nitrogen.]'7"l9 It would seem that if the sup-
porting matrix in the polymer-supported titanocene system
is rigid enough to prevent association of the supported

titanocene, then no ammonia formation should be observed in

29
fixation attempts using the beads. This was found to be
the case. Although homogeneous titanocene produced ammonia
in both the van Tamelin and Vol'pin-Shur systems when attempted
fixations were run, no significant amounts of ammonia were
ever produced by the supported titanacene beads.

Analytical methods differed for the two systems.

For the van Tamelin method. the nitrogen stream was passed
through a gas washing tower containing aqueous H280“. This
was then boiled down and either simply tested for ammonia
content with Nessler's Reagent32 after being made basic or
else titrated with a standard NaOH solution if the H280“
solution was of a known concentration. For the Vol'pin-Shur
fixations. Nessler's Reagent or Kjeldahl analysis was used.
For the latter. the acidic hydrolysis mixture was made basic.
steam distilled into boric acid solution. and titrated with
standardized HCl solution to quantitatively determine the
amount of ammonia formed.33 Since phenyllithium was used as
the reductant in the Vol'pin—Shur fixations, aniline yields
were determined by benzene extraction of the steam distil-
late and gas chromatographic determination of the aniline
using a durene standard.

No ammonia was ever detected using van Tamelin's
method with polymer-supported catalyst only. When the
homogeneous system was used. ammonia was clearly detected,
but when a combination of supported catalyst and homo-
geneous catalyst were used, no ammonia could be detected,

even though both quantitative and qualitative methods of

30
analysis were used. Many problems were encountered with
the nitrogen flow system and evaporation of the solvent
despite the inclusion of a THF bubbler before the reac-
tion flask, so the Vol'pinPShur method was used for most
of the bead nitrogen fixation experiments.

Table 5 summarizes a series of nitrogen fixations
run using the Vol'pinsShur method of nitrogen fixation.
Systems using homogeneous titanocene were run three times
and were found to give ammonia on all three runs according
to Nessler's Reagent. Four runs were then done with quan-
titative analysis of the products. giving the first four
items in Table 5. The first two used identical quantities
of reagents: the third used twice as much reducing agent.
and the fourth was a blank run under argon. Two bead

nitrogen fixations follow. These were analyzed quantita-

TABLE 5
VOL'PIN-SHUR NITROGEN FIXATIONS

 

 

 

 

 

 

 

 

 

Titanocene ., Reactants used PBBducts foundg
form mmol Ti I mmol Cli mmol NH3 mmol CNH2
Homogeneous 2.41 17.0 0.41 0.02
Homogeneous 2.40 17.0 0.32 0.02
Homogeneous 2.40 34.0 0.66 0.03
Homogeneous 2.40 17.0 0.01 0.00
Supported 2.16 17.0 0.02 0.00
Supp. + Homog. 4.40 34.0 0.80 0.09

 

3 3 0.01 mmol

 

31
tively also. The first used the same quantities of
reagents as the first two homogeneous reactions, while
the second used equal amounts of homogeneous and supported
titanium with as much reducing agent as the third homo-
geneous reaction.

The failure to detect any significant amounts of
ammonia in trials using only supported catalyst was
expected, as mentioned earlier. because dimeric titano-
cene is necessary and dimerization of the supported ti-
tanocene is probably difficult. The results tend to con-
firm further the idea that this polymer support can keep
the active metal centers apart. Another attempted nitro-
gen fixation using the supported system was done with
Nessler's Reagent used to test for ammonia. The results
were also negative.

If dimeric titanocene is indeed necessary for the
fixation of nitrogen, attempted fhlation using homogeneous
titanocene plus supported titanocene might lead to the for-
mation of some dimer pairs by supported and homogeneous
titanocene molecules. increasing the yield of’ ammoniaa
This is observed, as seen in the last item of Table 5.
Another interesting and significant fact can be seen in this
last item of Table 5. The ratio of ammonia to aniline in
the homogeneous systems is $20: in the mixed homogeneous-
supported titanocene system it is 9. If, as Vol'pin and Shur
propose, their system forms aniline by a monomeric titano-

cene reaction (with some form of fixed nitrogen possibly

32
from a former dimer) then in the mixed system, there is a
much greater chance that monomeric formation of aniline
will occur, because only half of the supported titanocene
is dimerized with homogeneous titanocene at best. The 50%
increase in the amount of aniline formed is then quite
reasonable. This result adds further credence to the
indications of reduced aggregate formation found in the
hydrogenation experiments with supported titanocene.

§pmp§py_p§_§tudies on supported pitanocene. The titan-

ocene studies in hydrogenation and nitrogen fixation.have

 

established the following:

1. Supported titanocene reduces dienes and inner
alkynes to alkanes without isomerization or detectable
formation of intermediate monalkenes.

2. Terminal alkynes and oxygen-containing sub—
strates are not reduced. though some reaction may occur.

3. Organolithium reagents or sodium naphthalide
both activate supported TiCp2Cl2 for hydrogenation cata-
lysis.

4. Supported titanocene may be reconverted to
its dichloride by treatment with HCl gas. -

5. Removal of the benzylic methylene group from
a position adjacent to the supported cyclopentadienide
results in increased activity.

6. Supported titanocene is mostly within the
interior of the beads in this system.

33
7. Aggregation of titanocene supported on the

copolymer is difficult or impossible.
8. Supported titanocene is active in nitrogen
fixation only when homogeneous titanocene is present in

solution as well.

Chelation Studies

Although some of the above results would seem to
indicate that the supported catalysts are held far enough
apart so that metal-metal interactions are difficult if
not impossible, the question of whether or not polymer-
attached ligands are mobile enough to chelate with sup-
ported complexes and thereby deactivate them is a recur-
ring onezo'ZB. It was decided that the equilibration of
a labile complex with the phOSphinated 00polymer, fol-
lowed by the measurement of the amount of freed ligand,
'would give some information relating to this question.

Two different complexes were chosen for these studies.

the rhodium complex of cyclooctene (COE), a very labile
dimeric complex, and rhodium Vaska complex, a less

labile monomeric complex. The use of Wilkinson's Catalyst
itself proved impractical because at the concentrations
necessary for accurate free triphenylphosphine determina-
tion dimerization occurred with the release of triphenyl-
phosphine (Reaction 1).

Reaction 1 2 RhC1(PC3)3 :_—--_ [RhC1(PC3)2]2 + 2 P913

The extent of ligand displacement in both systems

used was measured by gas chromatography. This proved

\

34
fairly simple for the COE dimer reactions but difficult

for the Vaska complex reactions, which had to be run near
the lower limits of detection for triphenylphOSphine. Thus
the results in the Vaska system are less reliable.

[RhCl(COE)2]2 studies.‘ The COE dimer complex was used in

 

reactions with 2%and 20%divinylbenzene-styrene c0polymer
beads. The 2%beads used had 0.8 mmol of phosphine per g
of beads, while the 2076beads used had either 0.9 or 0.4
mmol of phosphine per g of beads. In all the cases tested,
supported phOSphine was present in at least a four-fold
excess relative to the complex.

Twenty percent divinylbenzene beads with 0.9 mmol
of phOSphine per g of beads were equilibrated with the COE
complex (phOSphine:complex = 6.8 mmol :l.41mm01) in benzene
and sampled at two hours, one day, and nineteen days after
solvent addition. Free COE levels detected were 2.45, 2.90,
and 2.10 _-I_-0.301mnol at the reSpective times. The solution
was removed and treated with excess triphenylphosphine to
free any COE in solution as the rhodium complex.34 The anal-
ysis which followed indicated 2.25 mmol of'COE were present.
so virtually no rhodium.COE complex was present was present
in the solution: it had coordinated to the polymer. The
overall results indicated that half the complexed COE had
been freed, presumably diSplaced by phosphine ligands. The
degree of bidentate chelation indicated was about 70%.

Samples of the above beads were treated with tri-
phenylphosphine to prove that 3 mmol of COE were still on them.

35
One sample tested was treated with triphenylphosphine under

nitrogen: a second sample was treated with triphenylphos-
phine under hydrogen. The former was analyzed for free
COE after three hours: the latter was analyzed for cyclo-
octane after one and five hours. Both samples were ex-
pected to contain 0.08 mmol of COE. The first produced
0.07 mmol of COE and the second produced 0.08 mmol of
cyclooctene (both 1 0.01 mmol of hydrocarbon).

A further test was run using the same batch of
phosPhinated copolymer beads that was used in the above
equilibration as well as some of the 27%beads containing
0.8 mmol of phosphine per g of beads. A blank containing
[RhC1(COE)2]2 and two standards containing the complex and
triphenylphosphine were also run. After injection of ben-
zene, the bead-containing reactions were analyzed one day
later. while the others were analyzed after seven hours.
The amounts of reagents used and the results are listed in
Table 6. Note that two other standards using the rhodium
complex and triphenylphoSphine were run and are also in-
cluded in Table 6.

As may be seen from Table 6. free COE was present
only when a phosphine of some form was present. All of the
complex still in solution in these tests was removed before
analysis by passing the solution sample through a short
column of alumina. which held the highly polar complex
but allowed COE to pass freely. The amount of COE that

was freed, however, was 25 to 509%m0re than would be

36

00009 Ho 0 Hon oswnmmong mo floss 0.0
00009 mo w non «sasmmosm Mo 0088 0.0

ol

 

 

 

 

 

 

m.
H085 no 00 H «M
00.0 M00 000.0. 000.0 .0
00.0 000 000.0 000.0 .0
00.0 000 00.0 000.0 0
00.0 .00 00.0 000.0 0
NOeO a». “OOH QGQG III... WQHeO .3
00.0 0.00000 0.5000 00.0 130.0 0
00.0 0.300 00000 00.0 000.0 0
H0 .0 0.00.000 0300 00.0 000.0 .0
00900900 vsomona «schema acomonml.
m. moo ooh... Hoes 90.23023 .00 such 05.000303 Hons moo .005 cams—um
[ “VII“

 

 

 

2000500000000 000303305
0000000 2033000 00400 020000000

0 amfiu.

 

37
eXpected from a 1:1 stoichiometric disPlacement of COE by

the phosphine component in the homogeneous standard systems.
Thus an accurate determination of the degree of chelation
would seem.to be difficult to obtain, though the chelation
indicated should be the upper limit of that actually occur-
ing. From the results in Table 6. the degree of chelation
indicated in samples 1, 2. and 3 respectively was 156. 15a.
and 107%bidentate chelation. Chelations of over 100%
indicate tridentate chelation. Although samples 1 and 3 used
virtually the same amount of supported phosPhine and the same
amount of complex, the latter had considerably less chela-
tion indicated. This is reasonable, since the zofidivinyl-
benzene beads used in sample 3 should have a less mobile
matrix than the 2%divinylbenzene beads used in samples 1
and 2.

A third chelation study was made using 20%divinyl-
benzene beads with only 0.4 mmol of phOSphine per g of
beads. Patchornik found that self-condensations of enoliz-
able esters on polymer supports were very low when substi-
tution was kept low,20 so it was hoped that a lower degree
of phOSphination would decrease the amount of chelation
occuring. Samples of these beads were equilibrated with
the COE complex while two more standards using the complex
plus triphenylphosphine were also run. Both were sampled

at 3 and 50 hours; Table 7 lists the results.

Note that a very high degree of chelation, beyond
expectations for this system. was indicated. At the end of

38

0000: no w n09 osanmmosm no 0053 0.0 m

 

 

 

 

 

 

 

 

005. 00.00.... 0.
000.0 000 .0 M00000 000000 0.0 000.0 0
000.0 000.0 M00000 005000 0.0 000.0 0
000.0 000.0 000 000.0 000.0 0
000.0 000.0 000 000.0 000.0 0
a. a _ a. .

oovoovov acouonm Hammonn psomeun
”00000.ngnu 00:0 000000000 00 0000 000000000 005a 000 002a 000500
[ IIUhH 11!?

20300050000 0000003350”.
0040000 20200000 00000 03000000

0 munda

39
fifty hours. the beads in sample 3 analyzed as having 161146
bidentate chelation and those in sample 4 as having 210%
bidentate chelation. No larger excess of supported phos-
phine was present in this experiment than in the previous
ones. The only difference was the method used to prepare
the phosphinated polymer itself. Whereas the previous
experiments used phosphinated copolymer prepared via the
chloromethylated copolymer. this batch used polymeric
phosphine prepared from directly brominated beads which
were treated with butyllithium followed by chlorodiphenyl-
phosphine. It may be that this method produces a polymer
having "bunched" phoSphines on the polymer which are able
to chelate more than.the phosphine ligands on the other
polymers.
RhC31(CO)(P¢3)2 studies. Two samples of phoSphinated 20%

 

divinylbenzene-styrene copolymer beads containing 1.25 mmol
of phosphine per g of beads were equilibrated with the
rhodium Vaska complex. A standard reference sample con-
taining triphenylphosphine and the complex was also run.
but the formation of solid material made its results diffi-
cult to use. The solutions were analyzed after passage
through an alumina column.to remove any complex still in
solution. The Vaska complex was found to release all its
triphenylphOSphine when injected into the gas chromatograph.
so all but the last analyses used no alumina columns. The
results of the analyses are recorded in Table 8. After the
experiment was completed. the beads used in sample 3 were
washed, dried. and submitted for microanalysis. This indi-

cated the presence of 0.0”3 mmol of rhodium per g of beads.

#0
Based on the Vaska complex not present in solution. 0.025
mmol of rhodium per g of beads had been expected, with an
upper limit of 0.03# mmol of rhodium per g of beads if all
the complex present had been coordinated to the polymer.
Clearly the microanalytical results are unreliable.

While the results obtained in this experiment had
an estimated error of 15-20% (because the work was at the
lower limits of detection of triphenylphOSphine) they also
seemed to verify the presence of bidentate chelation. Indeed,
because virtually all of the triphenylphOSphine present
was detected either as free phoSphine or dissolved complex,
bidentate chelation of the supported complex seems to be
100 76. In both samples 2 and 3 a very large molar excess of
phoSphine was present, and that may have contributed to the
amount of chelation observed, since if even 5 or 10%of the
copolymer phosphine ligands were near enough to chelate a
metal. then with time an equilibrium state favoring the
chelated metal would be approached. It should be considered
also that the degree of chelation calculated for the COE
system assumes no breakage of the chloride bridges in the
dimer, although this is quite possible. If chloride bridge
breakage were a common occurance. then the tridentate che-
lation occuring could be less than that determined by
assuming that the bridging remains intact. This can be
understood when it is recognized that two separate rhodium
atoms coordinated completely to polymeric ph03phine ligands
would give a chelation percentage of 3007£by the method

which assumes chloride bridging is intact, while the actual

41

mAnumZOUIQE 02,080.00 9608.0 3. 02005.00 305.05. ewmmmam .0303 H
mAnumZoozosm 02,003.00 90200.0 3. 00000300 smacks». mwmmmmm 0.00.000.

'UIOI

osasnmonm wovaommsmauogdon 00000 0.0.0 + Namumvgovaosm 00000 30.0
0000032000 03.000090050800000 00000 0.0 + NAmumzooznfim 0200 0006 w.
300.000.000.00 00000 020.090 0 mum 0200 0.00.00 + 0003002030050 0800 03.0

 

 

 

 

 

 

m.
0000 00.0 H m

«Maw 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000.0 .0.

Q .o I O C O C O I O
Woma 00 0 00 0 00 0 00 0 00 0 00 0 00 0 000 0 m0
-2. .2. 2.. .2. -2. :3 000.0 ...... 00000.. 00
0.00100 31 00 001 00 00 T0 00 00 00 ha 004 .0010. 0 008.00 1
«03.03.00 mum 090.0 00000 J mum 00000 0.309 0095.0
[‘ 114%.“; ll - dill ‘ F

 

 

 

000000000008 0 0 00.3 0 030000
0000000 00002088 00000 0020000000

m mania

1+2
percentage would be 200$ , since every rhodium atom would
have a tridentate chelation form (100% would represent com-
plete bidentate chelation). This computational problem
only arises beyond the bidentate chelation level. how-
ever. so that if one major species is present in the sys-
tems using COE complex treatment. the predominance of bi-
dentate chelation is confirmed.
W. The overall pattern of the
chelation studies tends to confirm the presence of biden-
tate chelation in systems which have more mobile poly-
mer matrices (Mg. 2%divinylbenzene copolymer systems)
while indicating that in less mobile polymer matrix sys-
tems (1.94,, 20%divinylbenzene copolymer systems) which
have the ligand well distributed throughout the polymer,
chelation may be effectively kept at a low level. This
would correlate well with the results of eXperiments with
the supported titanocene system which seem to favor a low
level of interaction between adjacent metal centers on

the 20$ divinylbenzene matrix used.

GENERAL SUMMARY AND SUGGESTED
FURTHER RESEARCH

This research was directed toward answering four
major questions, as outlined in the introduction. The
answers obtained are:

l. Polymer support of a homogeneous catalyst is
feasible.36

2. Altered selectivity is displayed by these sup-
ported catalysts. based on the size and the polarity of the
substrates used.36’37

3. New catalysts of greater activity than their
homogeneous analogs can be synthesized by this method of
polymer support.38

4. Chelation may be a problem in polymeric sys-
tems which have flexible (low cross-linked) matrices and/or
high degrees of ligand substitution.38

As with any research. many new ideas have been pro-
moted by this work, and problems which still need much
effort to solve them have been encountered. This research-
er suggests that more work might be profitably directed
toward the following problems:

1. Analysis of the supported systems. This has
been the greatest problem in this work. because the poly-

mer matrix often blocks or overpowers analyses of the sup-

ported metals. One recent reviewer has pinpointed this as
1+3

4#
the biggest problem in studies of supported systems.30
Application of more sensitive methods such as Fourier
transform far infrared Spectroscopy and electron micro-
probe studies have been undertaken by some of my coworkers
and seem to be yielding useful information. The use of
color changes and microanalyses has reached its limits in
most of our work.

2. More complete determination of the factors
involved in chelation and ways to control them would seem
to be a possible means of producing more active catalysts,
and is again dependent upon the improvement of analytical
methods.

3. The nature of the supported titanocene system
would seem to be a fruitful area of investigation. This
system is extremely active in hydrogenation reactions, and
it might be possible to use it in the production of aryl
amines if a small amount of homogeneous titanocene could
be used to provide a catalytic nitrogen fixing system.

4. Now that a method of putting cyclopentadienide
ligands onto the polymer without chloromethylation is
available. the investigation of other ”sandwich” complexes
supported on the polymer might be undertaken. Although in
this researcher's opinion more reactive metal complexes
than titanocene (which is very reactive) might attack the
support. it is possible that many new or improved catalytic
processes could be carried out on the supported system.

It is certainly true that greater effective catalyst concen-
.trations are attainable using this approach.

1+5

5. The development of a "flow-through” system for
catalytic reactions is a long-term pipe dream for this
method of catalysis. With methods such as the use of
cyclic ethers as a hydrogen source for alkene reductions
with Wilkinson's Catalyst currently being reported in the
1iterature.35 the possibility of such flow systems becomes
more likely and should be investigated.

6. Further investigation of the selectivity of the
supported systems as well as kinetic studies of the bead
reaction are obvious research efforts one could suggest.

Both of these areas are already under investigation.

EXPERIMENTAL

Introduction

All NMR spectra were run on a Varian T—60 spectrom-
eter using intermal TMS in a solvent of deuteriochloroform.
Spectra are given in 6 units as parts per million down-
field from tetramethylsilane. Mass Spectra were run on a
Hitachi PerkineElmer EMU-6 by L. Guile. All gas chromato-
graphs were taken on a Varian.Aerograph Model 90-P using
%” columns. Microanalyses were done by Spang Microanaly-
tical Laboratory or Galbraith Laboratories (Rh analyses)
except for titanium analyses, which were done by E.S.
Chandreasekaren.39

Lithium reagents were purchased from.Alfa Inorgan-
ics. PhOSphines were supplied by Pressure Chemical Co.
All polymer beads used in the reactions were gifts of
The Dow'Chemical Company. All solvents used were reagent

grade, purchased from either J.T. Baker or Mallinckrodt.

46

1+7
e d e

Supported Wilkinson's Catalygt Batches A and B. A 15 g
sample of 2%DVB (Divinylbenzene)-styrene copolymer beads

 

was treated with 5 g of SnClu and 125 ml of ClCHZOCI-{ZCH3
according tho the method of Pepper. Paisley, and Young.9
The stirred reaction was halted after five hours by the
addition of 200 ml of 1:3 dioxane-H20. After stirring
overnight. the liquid was drained off and four lal dioxane-
320 rinses (100 ml each) were carried out. The following
sequence of solvent washes was then performed:

1. 3:2 dioxane-H20 5. Pure dioxane

2. 3:1 dioxane-105ml 6. 1:1 dioxane-benzene

3. 9:1 dioxane-10 $1101 7. Benzene

4. 9:1 dioxane-H20 8. Benzene
The damp. pale pink beads were vacuum dried. When dry.
they were white. Volhard analysiszu indicated a chloride
substitution of 0.9 meq of chloride per g of beads.

A 12 g sample of the chloromethylated beads (10.8
meq of chloride) was refluxed with a solution of lithio-
diphenylphosphide prepared the previous day from 8.02 g
of chlorodiphenylphosphine (#0 mmol) and l.# g of lithium
metal clippings (200 mmol) in 100 ml of THF.10 After 36
hours, the reaction was terminated by the addition of 100
ml of aqueous saturated NH4Cl solution. After stirring for

1.5 hours. the solution was drained off and the beads were

rinsed with 50 ml aliquots of the following solvents:

48

1. 1:1 dioxane-10 fiHCl 6. 3:1 dioxane-benzene
2. 1:1 dioxane-10 $HC1 7. 1:1 dioxane-benzene
3. 1:1 dioxane-H20 8. 1:3 dioxane-benzene
4. 3:1 dioxane-H20 9. Benzene
5. Dioxane 10. Benzene

After vacuum drying, Volhard analysiszu indicated the pres:
ence of 0.05 meq of chloride per g of beads. An elemental
analysis was also performed.

Agglxsig; 89.35 $10 7.74 % H 1.96 % P

This is 0.62 mmol of phoSphine per g of beads.

Batch A of the supported Wilkinson's Catalyst beads
was prepared by refluxing 1.37 g of the phosphine beads
(0.85 mmol of phosphine) with 1.07 g of Wilkinson's Cat-
alyst (1.16 mmol) in 100 ml of acetonitrile for eight days.
Some crystals of the dimer of Wilkinson's Catalyst.
[RhCl(P¢5)2]2, were observed. The solvent was removed and
the beads were rinsed with .acetonitrile and vacuum dried.

Batch B of the supported Wilkinson's Catalyst beads
was prepared by a different route. Prepene was bubbled
through 2.6 g of RhC13'3H20 (10 mmol) in 100 ml of 1:1 THF-
ethanol for three days to form [Rhc1(03H6)2]2.25 A 2.06 g
sample of phosphine beads (1.27 mmol of phoSphine) was then
added to the solution a week later. The supported phOSphine
ligands diaplaced some of the propene ligands, producing
Wilkinson's Catalyst on the beads. After three days, the
liquid was removed and the golden-yellow beads were rinsed

with three 10 ml THF rinses and three 10 m1 benzene rinses.

49
After vacuum drying, a sample of the beads was taken for
elemental analysis.
Ana : 86.38 75 c 7.54 9% H 1.46 0 P 1.31%Rh
This is 0.4? mmol of phosphine and 0.127 meq of rhodium per g

of beads.

Supported Wilkinson's Catalyst Batches C and D. Lithiodi-
phenylphOSphide prepared from 11 g of chlorodiphenylphos-

 

phine (50 mmol) and 1.8 g of lithium metal clippings (257
mmol) in 50 m1 of M10 was used to phosphinate 7 g of 2%
DVB-styrene copolymer beads which had been chloromethylated
by T. K. Brunck (7 meq of chloride; these beads had 1.0 meq

of chloride per g by Volhard analysiszu)

Syringe techniques were used to work with these
beads. because their 200-400 mesh size had been found to
clog frit pores. with Batches A and B. The supernatant solu-
tion was removed with a syringe after stirring for one day
at ambient temperatures. Five THF rinses (50 ml each) were
performed, accompanied by ten minutes of stirring between
each solvent addition and removal. Saturated aqueous NHuCl
solution was injected into the bead flask to destroy any
residual phospho-lithium reagent. The following 50 ml
rinses were then carried out:

1. 1:1 dioxane-H20 5. Dioxane
2. 1:1 dioxane-10%HCl 6. 3:1 dioxane-benzene
3. 5:1 dioxane-lOflHCl 7. 1:1 dioxane-benzene

4. 10:1 dioxane-10%HCl 8. Benzene

50
The beads were vacuum dried and analyzed for chloride con-
tent using the Volhard method.2u A chloride content of 0.3
meq per g beads was found, indicating a maximum phosphine
content of pg; 0.? mmol of phosphine per g of beads.

Batch C of the supported Wilkinson's Catalyst beads
was prepared by equilibrating 2.25 g of the phOSphinated
beads (1.6 mmol of phosphine) with 1.78 g of Wilkinsons
Catalyst in 30 m1 of benzene. These were stirred for one
week, then drained and repeatedly rinsed with benzene. Two
days of rinsing were necessary before the rinse solutions
had no more coloration. The beads were then vacuum dried.

Batch D of the supported Wilkinson's Catalyst beads
was prepared by 3.2 g of phosPhinated beads (2.2 mmol of
phOSphine) with 3.22 g of Wilkinson's Catalyst (3.5 mmol)
in #0 ml of benzene under argon for one month. Rinsing

with benzene was carried out for four days. The beads were

then vacuum dried.

Supported Wilkinson's Catglygt Batch E. Treatment of 20 g
of 30-80 mesh, 296DVB-styrene capolymer beads with 160 ml
of ClCHZOCH20H3 and u g of SnClu was carried out at room
temperature, for a total of five hours. The liquid was
drained off through a fritted filter, and 150 m1 rinses of
aqueous dioxane, aqueous dioxane plus 575H01, and dioxane
(twice) were performed. Vacuum drying followed.

The entire batch of beads was treated with lithio-
diphenylphosphide made from 2.8 g of lithium metal (400 mmol)
and 17.6 g of chlorodiphenylphosphine (80 mmol) in 300 ml of

51
THE. This was stirred at reflux for one day, then cooled
and hydrolyzed with 100 ml of saturated aqueous NHQCl after
100 m1 of the reaction solution had been removed. After
three hours. the aqueous solution was removed, and the fol-

lowing 150 ml rinses were performed:

1. 1:1 THF-H20 5. THF

2. 1:1 THF-55HC1 6. 1:1 THF-benzene
3. 1:1 THF-H20 7. Benzene

4. 7:3 THF-H20 8. Benzene

The beads were then vacuum dried.
All of the phosphinated beads were then equilibrated
with 13.2 g of Wilkinson's Catalyst in 350 m1 of benzene
for twelve days. They were then rinsed repeatedly with
benzene over a two week period ( at least two rinses each
day ). The deep-red beads were vacuum dried, and a sam-
ple was taken for elemental analysis.
W 82.90760 6.71%H 3.30%P 1.399501 3.43% Rh
This is 0.33 meq of rhodium per g of beads.

Benzyl-supported titanocene dichloride, Following the
procedure deve10ped by C. Gibbons,15 a sample of beads

with the benzyl-supported form of titanocene dichloride was
prepared. A 12 g sample of 30-80 mesh 2%divinylbenzene-
styrene capolymer beads was treated with 100 m1 of
ClCHZOCHZCH3 and 3.5 g of SnClu for five hours at ambient
temperatures. The beads were then washed with aqueous di-

oxane containing pa. 5%}{01 three times. aqueous dioxane

52
four times. and dry dioxane five times. Vacuum drying
followed.

Sodium diapersion (1.1g, #8 meq) was allowed to react
with a ml of freshly-distilled cyclopentadiene (50 mmol) in
150 ml of THF under argon. The chloromethylated beads were
added and the solution was stirred for three days. To ter-
minate the reaction, 20 ml of THF containing 2 m1 of 1075
HCl was injected with stirring. The solution was drained
off and several more rinses were carried out. using atotal
of 400 m1 of THF containing 10 m1 of 10%HCl. Three more
rinses with 100 ml aliquots of dry THFwere followed with
vacuum drying.

The golden-yellow beads were suSpended in 50 ml of TH?
and 25 ml of 2.# M methyllithium in diethyl ether was added.
The reaction was cooled in a dry ice-acetone bath for five
hours and then the solvent was removed, followed by four
#0 m1 THF rinses. The beads were vacuum dried. TiCpCl3 40
(13.5 g, 60 mmol) was then added to the dry beads. Benzene
(150 ml) was then added and the solution was stirred. The
beads, which should have had pg. 10 meq of active sites,
began to turn dark pink within ten minutes. The solution

was drained after eighteen hours and six dioxane rinses were

performed, followed by vacuum drying.

L-Phepygizcyclopentene, This was synthesized in a model .of
the proposed reaction of Scheme 3 (page 8) before attempting.

the reaction on the polymer support.

53
To 5 ml of'Eastman grade cyclopentanone dried over

“A molecular sieves in 10 m1 of diethyl ether (“.75 g, 57
mol) 1:0 ml of 1.6 M phenyllithium in diethyl ether (61Pmmol)
was added slowly over an hour. with stirring and cooling in
an ice bath to maintain a solution temperature below 10°C.
After fifteen.minutes of further stirring, 20 ml of H20 was
added. The organic layer was separated. washed with satur-
ated sodium chloride and aqueous NHucl. and dried over anhy-
drous MgSOu. The yellow solution was reduced to a volume

of 5 ml on a rotary evaporator. Distillation gave a frac-
tion collected at 230-2h0°C which weighed 0.5 g. Its NMR
spectrum matched that reported for the title compound by

“1 7.2-7.6 (m). 6.18(t). 2.3-2.8 (m),
and 1.8-2.2 (m), area ratio = 5:1:4:2.

Ketley and McClanahan:

l-Phenyl-l.3-cy910pentgdiene, A method similar to that

reported by Riemschneider, gp_gl.. was used.42 The title

compound was synthesized as the second model for Scheme 3
(page 8) before applying the Scheme to the polymer support.
2-Cyc10pentene-l-one was made via its ethylene

glycol ketal according to the method of DePuy. §3_§;.u3
The ketal was converted to the ketone by treatment with one
equivalent of 0.02%aquebus oxalic acid, and was distilled
just before use (b.p. = 150-15500). The ketone's NMR spec-
trum consisted of multiplets at 7.85, 6.20.2.70, and 2.35

with respective relative areas of l:l:2:2.

Dropwise addition of 9 m1 of 2-cyclopentene-1-one

54

(9 g, 0.11 mol) to 100 ml of 2Mphenyllithium in benzene
(0.20 mol) was done over a fifteen minute period while the
reaction was kept at less than 10°C in an ice bath. The
solution was stirred for thirty minutes and then 50 m1 of
H20 was slowly added to it. The benzene layer was sep-
arated and dried over molecular sieves overnight at 0°C.

Removal of the benzene on a rotary evaporator left
a yellow viscous liquid which was distilled at aSpirator
vacuum. The fraction collected a 160°C (p§.l g) closely
matched the NMR spectrum expected for the dimer of l-phenyl-
1.3-cyc10pentadiene, based on comparison to the Spectrum of
the dimer of cyclopentadiene itself. The Spectrum con-
sisted of multiplets at 7.5, 6.3. and 1-4 (areas 10:2:8).
The mass spectrum had a very strong parent peak at m/e = 142
(calculated for l-phenyl-l.3-cyclopentadiene is 142.08).
Major peaks also occured at m/e==l4l.115,40,44,41,39.57 and
69 (decreasing magnitude)._

Riemschneider reported that 2-phenyl-l,3-cyclopenta-
diene formed when 2-hydroxy-2-phenyl-l.3-cyc10pentadiene was
distilled and rapidly rearranged to 1-pheny1-l.3-cyclopenta-

42

diene and dimerized. The tendency of the product to yellow

and harden in air which was reported was also observed.

Phenyl-supported titanocene dichloride. The first synthesis
of the title system began with the bromination of 22 g of 20%

DVB-styrene copolymer beads using 110 g FeCl3 and 32 g of
bromine (0.20 mol) in 200 ml of nitromethane. The reaction

55
was stirred in the dark and monitored by measuring its

oxidizing ability. Periodically. 1.00 10.01 m1 aliquots
of the solution were removed and added to 1.7 g of K1 in
25 m1 of H20: the free I2 formed was then titrated with
standard Na28203 using a starch indicator.33b The reac-
tion had ended after two days. The beads were then rinsed
with 150 ml portions of:

l. 5%}{01 4. 1:5 5%HCl-THF

2. H20 (twice) 5. TH? (three times)

3. 1:1 5%HCl-THF
The polymer was then washed with THF in a soxhlet extractor
overnight and vacuum dried. An elemental analysis was done.
Apélxfiifip 67.92 % c 6.58 % H 25.34 % Br
This is 3.2 meq of bromide per g of beads.

A 26.3 g sample of the beads (84 meq bromide) in

25 ml of benzene was treated with 100 m1 of 2.4 M butyl-
lithium in hexane for four days. The beads were then washed
with three 50 ml THF rinses and suSpended in 40 ml of THF.
As the suspension was stirred. 8.2 g of 2-cyc10pentene-1-one
(0.10 mol) was added. Stirring was continued for one day.
The beads were then washed with 75 ml rinses of 1:1 benzene-
ethanol. 1:1 benzene-lfiethanolic H01, H20. and absolute
ethanol (three times). The beads were then vacuum dried,
followed by overnight extraction with dioxane in a soxhlet
extractor. They were analyzed.
Analysis: 84.10 %»C 7.82 1 H 4.68 % Br
Thus, 0.6 meq of bromide per g of beads is still present.

56

The cyclopentadiene-containing polymer (10 g) was
treated with 50 m1 of 2.4 M methyllithium in diethyl ether
with an additional 50 ml of diethyl ether added. The evo-
lution of methane gas was monitored with a gas buret until
it stapped, at which time pg. 700 ml of gas had been evolved
(30 mmol). Three THF rinses were carried out, and the beads
were then stirred with 11 g of TiCpClB (50 mmol) in 125 ml
of THF for twenty days. The beads were treated with THF
in a soxhlet extractor for two days. vacuum dried. and ana-
lyzed for titanium content.39
Ana 8 0.058 meq of titanium per g of beads
These beads were used for all the phenyl-supported titano-
cene reactions reported in this thesis.

The second synthesis of the title system followed
the same general pattern as the first synthesis. but with
a new bromination procedure. 118 g of 20%DVB-styrene beads
was brominated using 7 g BF3 and 16 g bromine (0.10 mol of
both) in 500 ml of nitromethane. The reaction was car-
ried out for one day in the dark. At the end of the reac-
tion period, pa. 5%of the beads floated at the top of the
solution, while the remainder settled to the bottom. It
was concluded that the less dense beads were unbrominated
(and therefore less dense) so they were siphoned off and
discarded. The remaining beads were rinsed with nitro-
methane. methanol. methylene chloride. and then vacuum dried.
Apglygis: 85.54%0 7.62%H 6.78%Br

This is 0.8 meq of bromide per g of beads.

57

Treatment of 20.18 g of the brominated beads (17 meq
of bromide) with 50 ml of 1.9 M.butyllithium in hexane (95
mmol) and 100 ml of hexane was done for 36 hours. The
resulting light tan beads were rinsed five times with THF
and then 75 ml of THF was added to suspend the beads.Slow
addition of 2.0 g of 2-cyclopentene-l-one (24 mmol) to the
cooled solution poduced an immediate change in the color of
the beads to white. but they gradually yellowed during the
three days of stirring which followed. Two wet THF rinses
and seven dry THF rinses preceded vacuum drying of the beads
with gentle heating (50°C). An analysis was run.

sis 89.04 % C 7.90 7‘ H less than 0.1% Er

The entire 20 g batch of cyclopenteneone-treated
beads was suSpended in 100 ml of diethyl ether in a dry ice
bath and treated with 50 ml of 0.8 M methyllithium in ether
for three days. They were rinsed three times with diethyl
ether and then vacuum dried. TiCpCl3 (5.6 g, 0.025 mol) was
added to the beads with 200 ml of benzene. This was stirred
for three days. washed with three benzene rinses, and extracted
with THF overnight in.a soxhlet extractor. The vacuum dried
beads were non-uniform dark red-pink in color. A titanium
analysis was run on these beads also.39
Analysis: 0.546 meq of titanium per g of beads

The bromination using BF3 seems to produce much more
supported titanocene in the product. possibly because no

residual iron salts are left on the polymer.

58

Hydrogenations

All solvents used were reagent grade, further purified
and deoxygenated by refluxing and distillation under nitrogen
with sodium (or potassium)-benz0phenone in the distillation
flask. Ethanol was dried, deoxygenated, and purified by
distillation from sodium ethoxide and diethyl phthalate under
nitrogen. Solvents were kept dry and oxygen-free by storage
under nitrogen and transferal with syringes.

11

Wilkinson's Catalyst was prepared from freshly

recrystallized triphenylphosphine (Pressure Chemical Company)

and RhCl '3H20 (Engelhard Industries). Titanocene dichloride

was used3as purchased from Alfa Inorganics. Naphthalene was
used as purchased from Eastman Kodak.

All of the liquid substrates were distilled from
sodium under nitrogen, except for allyl alcohol, which was
distilled from diallyl phthalate and sodium alloxide under
nitrogen, and l-hexyne. which was distilled from molecular

sieves under nitrogen. Ag-cholestene was prepared from

cholesterol by the method of Douglas. et al..uu while benzyl

 

acetate was prepared from sodium acetate and benzyl chloride
in dimethylformamide, a very simple and straightforward
reaction. The 4-cholestene-3-one used was made by K. Patel.
The apparatus used, refered to below as "the hydro-
genator". consisted of a normal atmOSpheric pressure hydro-
genation apparatus. with gas bursts of 250 or 50 ml used for

volume measurements. The leveling bulb was placed so that at

59

any time except when volume readings were actually being made
the system was operating at a slight negative pressure, thereby
preventing leaks from giving false hydrogen uptake data. The
process refered to as “hydrogen flushing" consisted of a mini-
mum of three vacuum cycles of better than 5 torr vacuum with
alternate hydrogen addition to remove virtually all oxygen

from the system before solvent addition.

Two additions were made to the hydrogenator during
the course of the research project: a temperature bath for
the reaction flask, kept at 25 10.590. and SAM. SAM was an
automated hydrogen addition system, which replaced the burst-
leveling bulb system of uptake measurement. SAM automatically
allowed a fixed amount of gas to enter the system whenever a
chosen partial vacuum was reached, giving an electric pulse
with each aliquot. The pulse was fed to a strip-chart
recorder. and thus gave a measure of the time at which each
aliquot was introduced. Whenever SAM and/or the bath have
been used. the Specific expermental description mentions it.

Alkenes were purchased from Aldrich Chemical Co. and
from Chemical Samples 00.: alkynes were purchased form.Far-
chan Research Laboratories.

All reductions were stirred using magnetic stirrers
with stir bars of the appropriate size in the reaction ves-
sels.

The benzyl-supported titanocene catalyst beads used
for hydrogenations were prepared by C. Gibbonss and W. Bonds
and had 0.37 meq titanium per g of beads.39

60

Batch.g guppgpted Wilkinsogfs Qapglygt hydrogepgpigpg. These

were run in a Schlenk tube, using 3 m1 of benzene which was
injected after hydrogen flushing of the system. After thirty
minutes of equilibration, 5 ml of cyclohoxene was injected and
the hydrogen volume was measured over a four day period. The
reduction rate was 0.046 ml of hydrogen per minute. The beads

were then rinsed and dried under vacuum.

Batch B supported Wilkinson’s Capglyst hydrogenations. These
beads were also tested in a Schlenk tube. For the first trial.

a solvent mixture of 30 m1 of THF and 5 ml of benzene was
injected into the Schlenk tube after hydrogen flushing. After
thirty minutes, 0.5 m1 of 1-hexene was injected and the hydro-
gen uptake was measured for twenty-seven hours. A reduction
rate of 1.12 ml of hydrogen.per minute was observed. The
beads were then rinsed with TH? and benzene and vacuum dried.

Reductions of l.0¢0.l M initial concentrations of
alkenes in benzene were carried out for cyclohexene.£§-choles-
tene. and mixtures of the two olefins using 10 ml total solu-
tion volumes. Cyclohexene was observed to be reduced at 0.173
ml of hydrogen per minute. while 0.003 ml of hydrogen per
minute was the steroid's rate. The beads were rinsed with

benzene andvacuum dried between each use.

Batch 9 suppopped Wilkinsgnfg Qapglyst hydrogenatigps, All
of the batch C beads were placed in a 100 ml Schlenk tube

and connected to the hydrogenator. After hydrogen flushing.

61
benzene (10.5 ml) was injected into the tube via a septum in
the sidearm. After 9;. one hour of stirring under hydrogen.
1.2 ml of cyclohexene was injected (initial olefin concen-
tration = 1.0 M). The hydrogen volume was monitored for
several hours. The solvent was removed by syringe and the
beads were vacuum dried. This reaction was repeated. The
first reduction had a rate of 0.575 ml of hydrogen per minute:
the second was 0.025 ml of hydrogen per minute.

ngch D supported Wilkinson’s Catglyst pydpogenatigpg. The

entire batch was placed in a Schlenk tube and flushed with
hydrogen after attachment to the hydrogenator and 9.5 ml of
benzene was injected: after stirring for one hour, 1.0 ml of
cyclohexene was injected and the hydrogen volume was measured
with time until the reduction was complete. The beads were
rinsed with benzene three times and vacuum dried.

After returning the Schlenk tube to the hydrogenator
and hydrogen flushing, 5 m1 of benzene was injected: 1.61 g
of'té-cholestene in.2 ml of benzene was injected after one
hour (initial olefin concentration 0.7 M) and hydrogen volume
yp. time readings were taken for two days. Cyclohexene (1.0
ml) was then injected, and the uptake rate was monitored for
Six more hours. The beads were rinsed with benzene and vacuum
dried.

The entire batch was then transfered to a Specially-
designed 300 m1 round-bottomed flask equipped with two sidearms.
Both sidearms had stopcocks, with one having a fritted filter

62
where it met the flask wall. Thus all solution additions and
removals were done without having to remove the flask from
the hydrogenator.

The beads were used for the reduction of several
alkenes, to determine the relative rates of reduction of each.
All alkenes were at 1.0 30.1 M initial concentrations in ben-
zene. Each alkene was injected into the beads already equil-
ibrated under hydrogen for one hour. and hydrogen uptake was
measured for at least two hours (results in Table 1, page 15).
Cyclohexene was used as a reference. run before and after the
reduction of every other alkene. Cyclohexene. l-hexene.
cyclododecene,26 l-octadecene, and pig-cyclooctene were

reduced.

Batch E supported Wilkinson's Catalygt hydpogenations.
A 10 g sample of the beads was used in the same Specially-

designed flask which had been used for the Batch D beads.
immersed in the 25°C bath. Reductions of l-hexene, allyl
alcohol. 4-pentene-l-ol. 3.5,5-trimethyl-l-hexene. cyclo-
hexene, norbornene, and l-dodecene were done. l-hexene was
used as the reference compound, reduced before and after each
other alkene. The substrates were all reduced in benzene,
using a total solution solution volume of 20 ml and initial
olefin concentrations of 1.0 10.1 M. l-hexene and 4-pentene-
l-ol were also reduced in 1:1 benzene-ethanol, using the same
volumes and substrate concentrations as used for plain benzene.

The hydrogen uptake was measured every fifteen minutes for

63
the first 1.5-2.0 hours of the reduction. The results are
in Table 2 (page 18).

Another series of reductions was done using the same
reaction flask with only 2 g of beads and 10 ml total solu-
tion volumes. All of these measurements were done with the
stirring bar's Speed varied frequently during the course of
the reduction and volume readings made every five minutes,
to confirm that the rate of hydrogen diffusion into the solu-
tion was not a limiting factor in the relative rates measured.
l—hexene, cyclooctene, and cyclohexene were used in plain
benzene and 1:1 benzene-ethanol for this reaction series.
all at 1.0‘10.1 M initial concentrations, with the reduction
flask in the 25°C water bath. The results of these reductions

are in Table 3 (page 21).

WW Following the
general method of C. Gibbons,15 a sample of benzyl-supported

titanocene dichloride beads weighing 0.040 10.003 g was
weighed into a 100 m1 sidearm round-bottomed flask. After
attachment to the hydrogenator and hydrogen flushing, the
beads were treated with 0.18 ml of 2 M butyllithium in hexane
and 5 m1 of hexane. After stirring for three hours. 0.5 ml
of substrate was injected and hydrogen volume readings were
taken every five to ten minutes for two to four hours. The
substrates tested were 1.5-cyclooctadiene, 1.3-cyclooctadiene,
1-hexyne, 3-hexyne. benzyl acetate, vinyl acetate. styrene,

and 4-cholestene-3-one. For the latter four, two 5 ml hexane

64
rinses were carried out before the substrate was tested
in another 5 ml of hexane. The two cyclooctadiene and the
3-hexyne reductions were periodically monitored via gas
chromatography (5', 20% NaCl on Silica Gel at 240°C) to see
if intermediate monoolefins were present. The results are
listed in Table 4 (p23).

One sample of the benzyl-supported titanocene di-
chloride beads was tested using sodium naphthalide as the
reducing agent. For this test. the same procedure as fol-
lowed above was used, except that 1 m1 of 0.23 M sodium naph-
thalide in.THF was added to the beads in 5ml of THF. After
stirring for two hours, three 6 ml hexane rinses were car-
ried out and 5 ml of hexane was added. After another hour,
0.5m1 of 1-hexene was injected and the hydrogen volume was
read for 3 hours. The rate observed was 0.45 ml of hydrogen
per minute.

Another set of reductions was done with these beads,
using SAM to monitor the reduction. 0.036 g of beads were ‘
treated with 1 ml of 2 M butyllithium in hexane and Sml of "
hexane. and were then used to reduce 0.5 m1 of cyclohexene
after an hour. The reduction.was monitored to completion.-
The same reduction procedure was repeated using 0.040 g of
the beads which had been thoroughly ground in a mortar and
pestle before use. The whole beads produced a rate of 88.7 ml
of hydrogen per minute per meq of titaniumgthe ground beads
produced one of 714 ml of hydrogen per minute per meq of
titanium.

65

Begd regeneratign, A 0.043 g sample of the benzyl-supported
titanocene dichloride beads used above was weighed into a

100 m1 sidearm round-bottomed flask, attached to the hydro-
genator. and flushed with hydrogen: 5 ml of hexane and 0.13 ml
of 2.2 M butyllithium in hexane was injected, stirred for
11.5 hours. and then 0.5 ml of l-hexene was injected. The
hydrogen volume was measured every five minutes for one hour
after injection. The rate observed was 1.75 ml of hydrogen
per minute. .

The flask was removed from the hydrogenator, the sol-
vent was removed, and 5 m1 of absolute ethanol was added.

The beads went from pale grey (active catalyst) to pale brown
(inactivated catalyst) immediately. Dry HCl was then bub-
bled through the solution. The beads became pink in less than
two minutes. the color they are before activation with a redu-
cing agent. Four hexane rinses and vacuum drying were pare
formed.

The HCl treated beads were now run through exactly
the same cycle of hydrogenation which they had been run
through on.their initial use. with the same amounts of rea-
gents and the time intervals as used earlier. The reduction
was monitored for seventy minutes. at which time the reduce
tion.was complete. The rate observed was 1.82 ml of hydrogen

per minute.
W A 0.036 s camels

of the beads was placed in,a 50 ml sidearm round-bottomed
flask and attached to the hydrogenator. After hydrogen

66
flushing, 0.25 ml of 2.4 M butyllithium in hexane and 20 ml
of hexane were injected. .An hour later, two 10 ml hexane
rinses were carried out. The beads were then stirred in 10 ml
of hexane for 15 minutes, followed by the injection of 0.25 ml
of cyclohexene. The hydrogen volume was measured every five
minutes for an hour and then at two and 2.5 hours after in-
jection. The beads were somewhat ground up during the course
of the reduction. A rate of 106 ml of hydrogen per minute

per meq of Ti was observed.

Wi1kgpsop's Cgtg1ys§ hydrgggpgtgopg, A series of reductions
were run.using RhCl(Pd3)311 at 2.4 10.1 mM concentration in

benzene with alkene concentrations of 0.70 10.02 M initially.
In each reduction, the dry catalyst was weighed into a 25 ml
sidearm pear-shaped flask, attached to the hydrogenator,
flushed with hydrogen, equilibrated with hydrogen in benzene
for an.hour. and then the substrate was injected. Hydrogen
volume was monitored for at least four hours. The substrates
used were -—cholestene, cyclohexene. cyclododecene?6 l-hexe
ene. cyclooctene. and l-octadecene. All except the steroid
were reduced at least twice. The results are recorded in
Table 1 (page 15).

Another series of reductions was performed. using
1.00 10.05 mm Wilkinson’s Catalyst concentrations and
1.00‘10.05 M initial olefin concentrations. Benzene was used
as the solvent for reductions of l-hexene. cyclohexene, cyclo-
octene. l-dodecene. 3.5,5-trimethyl-l-hexene, allyl alcohol.

norborene, and 4-pentene-l-ol: a mixture of equal volumes of

6?
benzene and ethanol was used as a solvent for reductions of
l-hexeneand 4-pentene-l-ol as well. The results of all
these reductions are recorded in Table 2 (page 18).

W A 0.187 a sample

of 10% Palladium on Charcoal was weighed into a Schlenk
tube. and 1.61 g of 80% Ag-cholestene (20% cholestane) was
added. Benzene (5 ml) was added and the system was attached
to the hydrogenator and flushed with hydrogen. The rate

of hydrogen consumption was monitored until the reduction
was complete. A reduction rate of 0.55 ml of hydrogen per
minute had been observed: 0.? m1 of cyclohexene was now

, injected and its rate of hydrogenation was also measured.

A rate of 1.30 ml of hydrogen per minute was observed for

cyclohexene.

W A 0-0125 a sample
of TiCp2012 (50 mmol) was weighed out into a 100 ml sidearm

round-bottomed flask, attached to the hydrogenator. and
flushed with hydrogen: 1 ml of 2.0 M butyllithium was then
injected. followed by 5 ml of hexane. This was stirred for
thirty minutes, and then 0.5 ml of cyclohexene was added.

The reduction was monitored using SAM, with the reaction
vessel immersed in the 25°C bath. A rate of reduction of

28 ml of hydrogen per minute per meq of titanium was observed
after an initial induction period of an hour which had a
rate of 5.6 ml of hydrogen per minute per meq of titanium.

68
The reaction was rerun using 0.0145 g of TiCpZCl2
(58 mmol): all other quantities and procedures were the same.
A reduction rate of 28 m1 of hydrogen per minute per meq
titanium was again observed, as was an induction period of

an hour.

Nitrggep Fixgpiops

All solvents were dried and degassed by distillation
from sodium (or potassium) and benzophenone under nitrogen.
Naphthalene was used as purchased from Eastman Kodak.
TiCpZCl2 was used as purchased from Alfa Inorganics. The nitro-
gen gas used was high purity dry nitrogen used as sold by
Airco. All supported titanocene used was prepared by C. Gibbons

and W. Bonds, with 0.86 meq of titanium.per g of beads.

=11 en: .3 _‘ .1 '. 3.1.: 9e... -:...o:.o_qn=-. .. a.“ ::=. A
0.81 g sample of naphthalene (6.3 mmol) was reacted with 0.155g
of sodium metal (6.? mmol) in 25 ml of TM? in a Schlenk tube
with a magnetic stir bar under argon for twelve hours: 1.305 g
of beads (1.12 mmol of titanocene dichloride) was then added
with an additional 10 ml THF. Nitrogen saturated with THF was
then bubbled through the system into a gas washing tower filled
with 100 m1 of 2051H280u for one day. The washing tower solu-
tion was then boiled down to 50 ml and a 10 ml sample of it
was made basic with NaOH solution. This was tested with
Nessler's Reagent32 for the presence of ammonia. Wet litmus

and odor tests were also done. No ammonia was detected.

69
The reaction was repeated, using 1.81 g of naphtha-
lene (14 mmol). 0.65 g of sodium (28 mmol). 75 ml of THF.
0.252 g of TiCp2C12 (1 mmol) and 1.34 g of beads (1.15 mmol
of titanium) with the same gas passage system. The 0.0838 N
H2504 solution used in the washing tower was boiled down to

50 ml after passing nitrogen through the system for eight
hours. After the addition of sufficient distilled water to

bring the acid solution back to its original volume. it was
titrated with 0.101 N NaOH using bromcresol green as the
indicator; 41 ml of base was required,indicating no ammonia

formation.
Vg1'p1n-Shpp 21xgtigp,ugipg bepzyl-gpppgpted p;pgppcgpe, All

fixations were done using 1000-1500 psi nitrogen in an auto-
clave previously flushed with nitrogen, at ambient tempera-
tures with continual stirring during the pressurized period
of the reaction. To 2.6 g of beads (2.2 mmol titanocene
dichloridé put into the autoclave's glass insert before the
autoclave was sealed 10 ml of 1.7 M phenyllithium in 7:3
benzene-diethyl ether (17 mmol) and 100 ml of diethyl ether
were added. The system.was pressurized and stirred for two
days, then depressurized and acidified with 20 ml of absolute
methanol and 20 m1 of 20% H280“. After stirring the results
ing solution for two hours in the autoclave. the solution was
removed and boiled down to 25 m1. It was made basic with
aqueous NaOH and tested for ammonia using Nessler's Reagent32
and odor tests. The solution was also extracted with 2 ml of
benzene and subjected to gas chromatography to detect any

7O
aniline present (55 Carbowax at 150°C). No ammonia or anil-
ine was detected.

The above reaction sequence was repeated using 2.7 g
of beads (2.2 mmol) with 10 ml of 1.7 M phenyllithium
(17 mmol) and 100 ml of diethyl ether. Analysis for ammonia
and aniline consisted of steam distillation of the basified
methanol-and-acid-treated reaction mixture into 4% boric
acid?3a The boric acid solution was extracted with 2 ml of
benzene and this extract was dried over NaOH and gas chro-
matographed using a durene standard to determine the amount
of aniline formed. The boric acid solution was then titrated
with 0.0482 M H01 using bromcresol green as the indicator to
determine the amount of ammonia formed. A total of 0.40 ml
of standard H01 was needed, indicating the formation of
0.02 mmol of NH3: no aniline was detected.

The reaction was repeated a third time, using 0.60 g
of TiCpZCl2 (2.4 mmol), 2.47 g of beads (2.1 meq of titanium),
80 m1 of diethyl ether, and 20 m1 of 1.7 M phenyllithium
(34 mmol). The same steam distillation, boric acid collec-
tion and H01 titration,338 and gas chromatography analysis
procedures were used as in the previous reaction. The for-
mation of 0.80 mmol of ammonia and 0.09 mmol of aniline were
indicated by the analyses run.

van.1gpg11n fixapigp;p§ipg hqppgeneoug tgpgppcepg, .A 0.54 g
sample of Ticp2c12 (2.1 mmol) was placed in a Schlenk tube,

~capped with a septum, and flushed with argon. .Addition of

71
120 ml of 0.1 M sodium naphthalide in THF obtained from
W. Bonds was then performed and the reaction.was stirred
rapidly for ten minutes. A syringe needle was then inserted
through the septum.so that it entered the solution. Nitro-
gen passed through the needle could escape the Schlenk tube
via the tube's sidearm, pass through a tube and into a gas
washing tower filled with 120 ml of 20%H2804. After one
day of nitrogen passage, the tower's contents were boiled
down to 50 m1, cooled, and then a 5 ml sample was made basic
with aqueous NaDH and tested for ammonia with Nessler's
Reagent?2 wet litmus, and by odor testing. Ammonia was

detected.
Vgl'pin-Shur f;§gtipn psipg ngppggnegug t1tan9¢gne,. The

following procedure was deed for repeated trials: while

the assembled autoclave was flushed with nitrogen, 0.6010.02g
of TiCp2012 (2.4 mmol) was weighed into a Schlenk tube and
treated with 8 m1 of 1.7 M phenyllithium (13.6 mmol) and

100 ml of diethyl ether under argon: this was stirred for
one hour and then transfered to the autoclave using a syringe.
The autoclave was pressurized to 1200-1500 psi nitrogen for
two to three days. After release of the pressure, the system
was worked up with 20 m1 of absolute methanol and 20 m1 of
209M230“, which was stirred for thirty minutes after addition.
The solution was then removed and boiled down to 20 ml. It
was then made basic and analyzed for ammonia. Three runs

were analyzed using Nessler's Reagent?2 Ammonia was present

72
in all three runs. Two more runs were done, using the
Kjeldahl method of quantitative ammonia anaiysis?3a A last
run was made using 20 m1 of phenyllithium solution. The
results are recorded in Table 5 (page 30).

A blank was run to test the Kjeldahl method's
steam distillation-boric acid collection system for back-
ground base levels. The usual reaction sequence and quan-
tities of reagents were used, but the system was never ex-
posed to nitrogen. The analysis indicated the presence of
0.01 mmol of ammonia (0.20 ml of 0.0482 M H01 were required
for titration of the boric acid solution). No aniline was

indicated by gas chromatographic testing.

Chelgtion Studies

[Rh01(00E)2]2 studies. All analyses were done using either

 

a 5', 5 5 SE-30 on.Chromosorb N gas chromatography column

at 125°C or a 5', 3 % SE-30 on Varaport-30 column.at 120° .
The latter was slightly better. For each analysis, 0.5 m1
of the solution.was withdrawn from the reaction flask and
passed through a one inch column of alumina in a disposable
pipet tube to remove the complex from solution, leaving only
free COE. All solvents were dry and oxygen-free, and all
reactions were carried out under nitrogen or argon.

Some phoSphinated 20 % DVB beads were prepared for
equilibration studies by reacting them with lithiodiphenyl-
phOSphide made from 7.5 ml of chlorodiphenylphosphine (40 mmol)
and 1.5 g of lithium metal (210 mmol) in 250 m1 of THF. The

73
beads used (20.0 g) had already been chloromethylated by
C. Gibbons, who determined that they contained 1.03 meq of

24 After phOSphination, less than

chloride per g of beads.
5 % of the original chloride was still present, meaning a
maximum phosphine content of‘pg. 0.9 mmol per g of beads.

To 7.64 g of the above phOSphinated beads (6.8 mmol
of phOSphine) in a Schlenk tube. 1.01 g of [Rm1(cos)2]2
(1.41 mmol, 5.64 mmol of'COE) and 0.298 g of durene were
added. The solution was stirred and sampled at two hours,
one day, and nineteen days after the addition of solvent.
The free 003 detected was 2.45, 2.90, and 2.10 mmol at the
reapective sampling times. The solution was removed from
the beads after the last sample and three 5 ml benzene washes
were performed. The rinses were added to the solution removed
earlier and the entire benzene solution was treated with
0.925 g of triphenylphosphine (3.5 mmol) to release 00E from
any of the complex which might still be in solution. The
triphenylphosphine-treated liquid was then analyzed for free
00E: 2.25 mmol was detected. The beads were further rinsed
and then vacuum dried. A 0.2159 g sample of the dried reddish
beads was treated with 0.0434 g of triphenylphosphine (0.166
mmol) and 0.0200 g of durene under nitrogen in 3 ml of benzene
for three hours, then sampled for free 00E content. Gas
chromatographic analysis indicated 0.07 mmol of free COE was
present. {Another sample weighing 0.2073 g was treated with
0.0392 g of triphenylphosphine (0.150 mmol) and 0.0129 g of
durene in 5 ml of benzene under hydrogen and analyzed for

74
cyclooctane after one and five hours. After one hour, 0.06
mmol of cyclooctane was detected: at five hours, 0.08 mmol
was present.

A sample of 2% Divinylbenzene-styrene copolymer beads
phophinated by E. Sweet having at least 0.8 mmol of phosphine
per g of beads were also tested for chelation, with another
sample of the 20% cross-linked beads used above. A blank
sample and two samples using triphenylphoSphine were also

used, as follows:

 

 

[Rh01(COE) J PhOSphine Component
§emele. sinursne. cg: defiei, __efezme_ as

1 0.0161 0.0870 0.121 2% beads 0.8292
2 0.0199 0.0954 0.133 2% beads 1.0697
3 0.0208 0.0795 0.111 20% beads 0.7129
4 0.0064 0.0350 0.049 -- none ~-
5 0.0101 0.0390 0.054 P05 0.0708
6 0.0144 0.0447 0.062 P05 0.1294

Each sample was treated with benzene after flushing with

argon: 3 ml was used for samples 1-3 and 1 ml for samples

4-6. The samples were stored under argon, periodically Shaken,

and analyzed after 7 hours for samples 4-6 and after one day

for samples 1-3. The results are listed in Table 6 (page 36).
Two more reference reactions were run, using triphenyl-

phosphine and [Rh01.(00E)2]2 in 1.5 m1 of benzene under argon.

Sample 1' used 0.0152 g durene, 0.1020 g [Rh01(00E)2]2 (0.142

mmol), and 0.0571 g P05. Sample 2' used 0.0148 g durene,

0.0875 g [Rh01(00E)2]2 (0.122 mmol), and 0.0685 g P05. These

were stirred for twelve hours under argon and then sampled

75
for 00E. See Table 6 (page 36) for the results.

Another sample of phOSphinated beads was prepared by
treatment of 20 g of brominated 20% Divinylbenzene-styrene
copolymer beads (16 mmol of bromide) with 40 ml of 2.2 M
butyllithium in hexane under nitrogen for one day, followed
by four 50 m1 THF rinses and treatment with 1.7 g of chloro-
diphenylphoSphine (7.7 mmol) in 75 ml of THF for twelve hours.
These beads were then rinsed with 50 ml aliquots of: l) Satu-
rated aqueous NHuCl (twice) 2) Water (twice) 3) 1:1 THF-water
4) 2:1 THF-water 5) TH? (four times) followed by vacuum
drying. They were then used in the following experiment:

 

 

[Rh01(00E) ] PhoSphine Component
Sample ELQBEQDQ .ds ___kam1 8
1 0.0202 0.1759 0.245 P05 0.0577
2 0.0301 0.1423 0.198 P05 0.1466
3 0.0345 0.1540 0.236 beads 5.0
4 0.0220 0.1425 0.198 beads 3.8

10 ml toluene was added to each flask after nitrogen flushing
and analyses were carried out after three and fifty hours.

The results are listed in Table 7 (page 38).

Rh01(CO)(P0:)2 studies. A 15 g sample of 20%1Diviny1benzene-

 

styrene copolymer beads was treated with 100 ml chloromethyl-
ethyl ether and 4 g of stannic chloride for five hours, then
washed with dioxane-5% M01 three times, aqueous dioxane twice,
and soxhlet extracted overnight with dry dioxane. These were
then treated with lithiodiphenylphosphidelo made from.l.40 g
of lithium metal clippings (0.20 mol) and 8.0 ml of chlorodi-
phenylphosphine (0.04 mol) in 100 m1 of THF under nitrogen

76
for two days and then worked up with deoxygenated solutions
(100 ml each) of: 1) Saturated aqueous NH401, 2) Water
(twice), 3) 1:1 Water-THE, 4) 1:2 Water-THE, 5) THF (four
times). They were vacuum dried for two days and analyzed.
A sis 87.82 % C 7.36 % H 3.82 % P 1.04 %»Cl
This is 1.25 mmol of phosphine per g of beads. This batch
of beads was used for the following chelation study!

- > . PhOSphine RhCl(CO)(P¢ )1?
§gpp1g g Tetrgcggang g fopp g p292
1 0.0482 0.0648 P05 0.3046 0.441
2 0.0487 4.8 beads 0.1777 0.257
3 0.0929 8.2 beads 0.1954 0.283

Each sample was placed in a vial which was sealed and flushed
with nitrogen: 20 ml of benzene was added to samples 1 and 2
and 25 ml of benzene was added to sample 3. The vials were
then sampled (0.5 ml each time) for gas chromatography on

a 9" 6%SE-30 on Chromosorb w column at 190°C. Analyses Were
performed at 1,.15, 40, 65, 90, 140, and 165 hours after sol-
vent addition. A sample of the solution in vial 2 was taken
at 175 hours and analyzed before and after passage through

a one inch alumina column to remove any'Rh01(00)(P¢'3)2 in
solution. The same test was done with the sample 3 solution
after 216 hours of elapsed time. The results of all the
analyses of this system are given in Table 8 (page 41).

The beads of sample 3 were washed with benzene, vacuum dried,

and submitted for elemental analysis.
M 87-33 9‘0 7.33 %H 3:81.731” 0.70%01 0-445311

77
The analysis indicates the presence of 1.23 mmol of phos-

phine and 0.043 meq of rhodium per g of beads,

A test of the amount of triphenylphosphine freed
by the Vaska complex on injection into the gas chromato-
graph was performed. A solution of 0.0174 g of Vaska
complex, m:c1(00)(P03)2 (0.0252 mmol, 0.0504 mmol of
triphenylphosphine) and 0.0122 g of triphenylphoSphine
(0.0466 mmol) was prepared. Toluene (20 ml) and a stan-
dard (tetracosane, 0.0118 g) were added and the closed
system was stirred under nitrogen. Gas chromatographic
analysis was carried out on the solution after one and
three days (9” SE-30 column at 190°C). For both analyses,
the solution was analyzed with and without passage through
a one inch column of alumina in a disposable pipet tube.
The alumina removed all of the Vaska complex from solution,
but allowed free passage of triphenylphosphine in solution.
The one day sample analysis indicated the presence of
0.097 mmol of triphenylphoSphine before alumina treatment
and 0.044 mmol of triphenylphOSphine after alumina: the
three day sample analysis indicated the presence of 0.094
mmol of triphenylphoSphine before alumina treatment and
0.047 mmol of triphenylphOSphine after alumina. All of the
above are 1 0.005 mmol of triphenylphOSphine. The results
seem to clearly indicate that all of the triphenylphOSphine
on the rhodium Vaska complex is lost on injection into the

gas chromatograph, and is detected and measured accurately.

BIBLIOGRAPHY

1.

BIBLIOGRAPHY

A. E. Martell and M. E. Calvin, ”Chemistry of Metal

Chelate Compounds", Prentice-Hall, Englewood Cliffs, N. J.,
1952. Ch 8.

2.
3.

H. Heinemann, C e e o , M§y~1221, 286.

F. A. Cotton and G. Wilkinson, "Advanced Inorganic Chem-

istry, A Comprehensive Text”, 3rd Ed., Interscience Publishers,
New York, N. Y., 1972, p 770.

4.
5.
o.
7.

R. F. Heck, Adv, 0rganometa11ic thm., 4, 243 (1966).
H. Lemke, Hydrocapppn Process., 45. 27 (1966).
B. H. J. Hofstee, Nat e, 213, 42 (1967).

a W. 0. Haag, e3 g1., Belgian Patent 721,686 (1968).
b W. 0. Haag, D. D. Whitehurst, Ger. Offen. 1.800.371

and 1,800,380 (1970).

8.

R.L.Lazcan0. J~E- Germain. Bulll_§2£a_921eamflaaa 1223-

1869.

9.

K. w. Pepper, H. M. Paisley, M. A. Young, J, thm, 899,,

A: 1253: 4097-

10.
11.
12.
13.

C. Tamborski, gj_31., J, Qrg, Chem,, 22, 619 (1962).

C. O'Connor, G. Wilkinson, 1. Chem. 390,, A, 1268, 2665.
R. B. Merrifield, Sc1gpce, 159, 178 (1965).

D. C. Neckers, D. A. Kooinstra, G. W. Green, M

Chem, Sgc,, 24, 9284 (1972).

14.

15.
16.

a; J.P. Collman, c ts Chem. Res., 1, 136 (1968).
b R. F. Heck, d., g, 10 19 9 .

C. Gibbons, i. S. Thesis, Michigan State University, 1972.
Y. Tajima, E. Kunioka, J, ng, Chem,, 33, 1689 (1968).
a) R. H. Marvich, H. H. Brintzinger, J. Amer, Chem. 890,,

17.
93, 2046 (1971).

b) J. E. Bercaw, R. H. Marvich, L.G. Bell, H. H. Brint-

zinger, 1big., 94, 1219 (1972).

78

79
18. E. E. T 1i , 1., ’b'd 707 (1968).
g; E. E. 2 2321.2. 1+1: 1:11: 3;: is. (19.9).
c) E. E. van Tamelin. 21.21... 113.51... 2 . 5251 (1970).

19. M. E. Vol'pin and V. B. Shur in "Organometallic Reactions”,
Vol. I, E. 1. Becker, M. Tsutsui, Eds,, Wiley-Interscience,
New York, N. Y., 1970, p 102.

20. a) A. Patchornik, M. A. Kraus, g. Amer, Chem, Sgg,, 22,
7587 (1970).
b)A. Patchorkik, M. A. Kraus, ibid., 23, 7325 (1971).

210 8; Co Co LeZnOff. Jo Ye Wong, E. 0 J. C 9111 . 59. 2892 (l9;2)e
b C. C. Leznoff, J. Y. Wong, 1131.0... 1, 2 52, 3756 (1973

22. S. L. Regen, D. P. Lee, 1, Amer, Chem, Sgc,, 2fi, 294 (1974).
23. a) J. P. Collman, L. S. Hegedus, M. P. Cooke, J.R. Norton,
G. Dolcetti, D. V. Marquardt, ibid., 4, 1798 (1 72).

o) J. P. Collman, 0. A. Re'e'd", b 3., 95, 2 8 (1973).

24. J. M. Steward, J. D. Young, ”Solid Phase Peptide Synié
thesis”, W. H. Freeman, San Francisco, 0a., 1969', p 55.

25. R. Cramer, Inorg. Chem“ 1, 722 (1962).

26. 30 % gig-isomer ,’ 70 % m-isgmer based on gas chrom-
atographic analysis (5' SE-30 at 150 C).

27. K. Shikata K. Nishino, K. Azuma,‘ Y. Takegami, 1&9ng
We 5. 358.(1965).

28. a) K. B. Sharpless, R. P. Hanzlik, E.E. van Tamelin,
Am 1'. an. S C . 28. 209 (1963).
b) E. E. van Tame in, M. A. Schwartz, 1b1d.,§z, 3277 (1965).

29. a; D.F. Shriver, ct . hem s , 2, 231 (1970).
b C. N. Schrauzer, 1214-: 1, 97 (19 8).

c) J. P. Collman‘ M. R. MacLaury W
3019 (1974). ' ' '2; '
30. 2. M. Michalska D. E. Eebster‘ W
65 (1974). ' ' ' 10.
31. G. 018.11: 21.1.9.1.“ W: 52: 1039 (1964)-
32. R. C. Weast, e , Eds., "The Handbook of Chemistry and

Phységs", Vol. 45, C emical Rubber 00., Cleveland, Ohio, 1964,
p D- e

80

33. a) D. A. Skoog and D. N. West, "Fundamentals of Analy-
tical Chemistry;’Holt, Reinhart, and Winston, New York,

N0 Y0: 1963: P 3550
b) 0 e Cit . p462.

34. L. Porri, A. Lionetti, G. Allegra, A. Immiezi, Qhem,
gamesna. 1225. 336-

35. T. Nishiguchi, K. Fukuzumi, J. Amer. Chem. Soc,, 26,
1893 (1974)-

36. R. H. Grubbs, L. C. Kroll, J. Amer. Chem. 800;, 23,
3062 (1971)

37. R. H. Grubbs, L. C. Kroll, E. M. Sweet, J. Macromo1.
Sci,-Chem., AZ, 1047 (1973).

38. Re Ho Grubbs. et 8.1., J. Amer. Chemo 509;. 25. 2373'
(1973).

39. The analysis procedure followed consisted of drastic
oxidation of the beads with hot H280“ and (NH )2 SO , then
burning off all material except T102, whichw S 2th n weighed.

40. R. D. Gorisich, J. Amer. Shep, Soc., 62, 4211 (1960).

41. A. D. Ketley, J. L. McClanahan,J .Org, Chem,, 39, 942,
(1965).

42. R. Reimschneider, A. Reisch, H. Horak, Mgnatsh. Chem,,
219 829 (1960)-

43. C. H. DePuy, B. W. Ponder, J. D. Fitzpatrick, J. r .
.5294: 22: 3508 (1964)

44. G. H. Douglas, P. S. Ellington, G. 0. Meakins, R.
Swindells, J, Chem, Sgc., 1252, 1720.

45. D. Evans, J. A. Osborn, G. Wilkinson, Inor S th.,
XI. 99 (1968).

"I111111111111“