Activation of hydrogen and carbon monoxide by transition metal complexes

MISSING IMAGE

Material Information

Title:
Activation of hydrogen and carbon monoxide by transition metal complexes
Physical Description:
viii, 212 leaves : ill. ; 28 cm.
Language:
English
Creator:
Miller, James G., 1957-
Publication Date:

Subjects

Subjects / Keywords:
Transition metal compounds   ( lcsh )
Hydrogen   ( lcsh )
Carbon monoxide   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1985.
Bibliography:
Bibliography: leaves 203-210.
Statement of Responsibility:
by James G. Miller.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000878306
oclc - 14878361
notis - AEH6049
sobekcm - AA00004880_00001
System ID:
AA00004880:00001

Full Text










ACTIVATION OF HYDROGEN AND CARBON MONOXIDE
BY TRANSITION METAL COMPLEXES














By

JAMES G. MILLER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1985






























To my parents, Mary and George Miller, and my wife,

Nan, whose constant support and encouragement made this

work possible.














ACKNOWLEDGEMENTS


I wish to express my deep appreciation to my research

director, Professor Russell S. Drago, for his guidance,

motivation, and understanding during my years at the

University of Illinois and the University of Florida.

A very grateful thank you goes to Professor Glenn C.

Vogel for his encouragement and friendship.

I would like to thank the members of the Drago research

group, both past and present, for their help, discussions

and friendship. Among the group Mike Desmond, Dean Oester,

Pete Doan, Jim Stahlbush and especially Keith Weiss deserve

special thanks. Robert King and Chris Schirmer are

acknowledged for their invaluable assistance with the GC

mass spectrometry and electrochemistry studies,

respectively.

Financial support by the University of Illinois, the

University of Florida and Professor Drago in the form of

teaching and research assistantships is gratefully

acknowledged.

Finally, a very special thanks go to Ginger Solano and

my wife Nan for their patience and help in preparing this

manuscript.


111i








TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS . . iii

ABSTRACT. . . vi

CHAPTER

I. INTRODUCTION. . . 1

II. ACTIVATION OF MOLECULAR HYDROGEN. . 4

Introduction. . . 4

Experimental. . . 11

Materials. . 11
Instrumentation. . 12
Synthesis. . . 13
Thermodynamic Measurements . 14
Hydrogen Gas Flow System . 15

Results and Discussion. . 24

Conclusion. . . 46

III. IMMOBILIZED HOMOGENEOUS CARBON MONOXIDE
REDUCTION CATALYSTS. . 48

Introduction. . . 48

Experimental. . . 59

Reagents. . 59
Instrumentation. . 59
Fixed Bed Flow Reactor . 60
Synthesis. .. . 63

Results and Discussion. . 73

Polymer Supported [Ru(CO)313]" and
[HRu3(CO)11]-. . 73

Synthesis and characterization. 73
Catalysis . ... 83

Covalently Supported Ir4(CO)11 83
Synthesis . ... 83
Infrared characterization ... 86
Catalysis . ... 94
CH3CI catalyst development. 102

iv








Halogen sources . ... 126
GC Mass spectrometry. . 131
Mechanism . 142

Conclusion. . . 146

IV. IMIDAZOLE FUNCTIONALIZED SILICA GEL ... 149

Introduction. . . 149

Experimental. . ..... 153

Materials. . 153
Instrumentation. . 153
Synthesis. . . 154

Results and Discussion. . 158

Synthesis Imidazole Support. ... 158
Hydroformylation . 167

Conclusion. . . 169

V. SOLUBLE BIMETALLIC COMPLEXES. ... 171

Introduction. . . 171

Experimental. . . 173

Materials. . . 173
Instrumentation. . 173
Synthesis. . . 177

Results and Discussion. . 181

Ligand Preparation. . 181
Metalomers, Synthesis and Characterization 182
Electrochemistry . ... 195

Conclusion. . . 199

VI. SUMMARY AND CONCLUSION. . 201

REFERENCES . . .. 203

BIOGRAPHICAL SKETCH. . 211









v








Abstract of Dissertation Presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


ACTIVATION OF HYDROGEN AND CARBON MONOXIDE
BY TRANSITION METAL COMPLEXES

By

James G. Miller

December 1985

Chairman: Russell S. Drago
Major Department: Chemistry

Reported here are four studies related to the catalytic

processes involving the activation of carbon monoxide and/or

molecular hydrogen by transition metal complexes. The first

study investigates the process of hydrogen activation by

square-planar rhodium (I) complexes of the formula

(P(p-tolyl)3)2RhClB where B is P(p-tolyl)3,

tetrahydrothiophene (THTP), pyridine and N-methylimidazole.

Enthalpies were determined for the reaction of the

P(p-tolyl)3 and THTP complexes with H2 forming cis

dihydride complexes of the formula (P(p-tolyl) 3)2RhClBH2

Attempts to determine enthalpies for the pyridine and

N-methylimidazole complexes were unsuccessful.

The metal-hydrogen dissociation energies determined

from these enthalpies were found to be insensitive to the

ligand variation around the rhodium center.

In the second study, ruthenium and iridium homogeneous

CO reduction catalysts were immobilized on silica gel and

alumina. Ionic attachment of the two ruthenium anions








[Ru(CO)3 3]- and [HRu (CO)11 ] to ammonium

iodide functionalized silica gel, as demonstrated by

infrared showed no syngas (CO and H2) conversion up to

1750C and 1 atm pressure.

A physical mixture of A1C13 and Ir4(CO)11

covalently attached to a phosphine silane functionalized

silica gel or alumina was found to selectively produce

CH3C1 from syngas under very mild conditions, 250C, 1

atm pressure. Replacement of AlC13 as the chloride source

by addition of HCl(g) to the syngas feed enabled production

of CH3C1 with 99% selectively at temperatures up to

1000C. The novel chemistry was demonstrated with other

halogen sources: aqueous HC1, C12 and HBr (producing

CH3Br). The presence of a Lewis acid and the phosphine

silane linkage appeared to be important for their catalytic

properties.

In the third study a 4(5) substituted imidazole

functionalized silica gel support was synthesized to

heterogenize a homogeneous rhodium imidazole

hydroformylation catalyst. The synthesis utilizes Schiff

base reactions between an aminosilane functionalized silica

gel, histamine and 1,4-dihydroxy-2,5-dibenzaldehyde. The

rhodium metalated support was active for the conversion of

1-hexene but suffered severe metal leaching.

In the fourth study a series of soluble bimetallic

complexes of the formula [M(II)(DIOX)BF 2]2 (M = Cu, Ni,

Zn, Pd; DIOX = bis-4-tert-butyl-2,6-diformyl-phenol dioxime)


Vii









was synthesized and characterized. These complexes are

formed from analogous insoluble 0-H-*"0 bridged complexes by

reaction with BF *Et 0. Electrochemical studies were

performed on the Cu(II) BF2 capped complex in acetone and

DMF and illustrated strong metal site interactions.


viii














I. INTRODUCTION


The field of catalysis is one of the most active and

intensely pursued areas of chemistry research. The

overriding goal of this research, driven primarily by

its economic importance, is to develop better and more

efficient catalysts.

Reported here are the results of four studies all

related to catalytic processes (i.e. hydrogenation, carbon

monoxide reduction, hydroformylation) involving the

activation of carbon monoxide and/or molecular hydrogen by

transition metal complexes. These studies focus on two

different aspects of catalysis research: 1) gaining an

understanding of chemical transformations taking place in a

catalytic cycle and 2) developing new and novel transition

metal catalysts. In the first study thermodynamic data

were obtained on the activation of molecular hydrogen by

square-planar rhodium(I) complexes of the formula

[P(p-tolyl)3]2RhClB (B = P(p-tolyl)3, THTP, Py, MeIm).

These results provide needed information on the strength of

the metal hydrogen bonds formed upon activation of molecular

hydrogen by a typical hydrogenation catalyst, "Wilkinson's"

catalyst and the sensitivity of these bonds to variation

in the ligand environment around the metal center. These

1









results also give insight into adverse effects of additives

such a pyridine to the catalytic system.

In the second study immobilized homogeneous transition

metal catalysts were synthesized and tested for their use in

developing selective carbon monoxide reduction catalysts

which operate under mild reaction conditions. A low

pressure and temperature catalytic system was discovered and

described for the conversion of syngas (CO + H2) and HC1

selectively to chloromethane using a bifunctional supported

iridium cluster catalyst. The active catalyst precursor

consists of Ir4(CO)11 covalently attached to a phosphine

silane functionalized silica gel or alumina and can be

operated in the presence or absence of an A1C13

co-catalyst.

The third study, which relates very closely to the

immobilized catalysts used in the previous study, involves

the development of a 4(5) substituted imidazole

functionalized silica gel support. The new support was

designed to heterogenize a reported homogeneous rhodium

imidazole hydroformylation catalyst of the formula

[Rh(imidazole)] 23'3 The synthesis and characterization

of this new support are reported along with preliminary

tests on the hydroformylation activity of the rhodium metal

loaded support.

The fourth and final study reported deals primarily

with the synthesis and characterization of new soluble

bimetallic macromolecules which might be capable of






3


facilitating bifunctional catalysis similar to that

demonstrated in the second study. These complexes

synthesized are of the formula [M(DIOX)BF2]2 (DIOX =

bis-4-tert-butyl-2,6-diformyl-phenol dioxime and M = Cu, Ni,

Zn, Pd) and contain two metal centers in close proximity

allowing metal site interactions.














II. ACTIVATION OF MOLECULAR HYDROGEN

Introduction



Activation of molecular hydrogen by metals is an

important fundamental process with relevance to many

catalytic systems (i.e. hydrogenation of olefins,4

hydroformylation,5 and reduction of carbon monoxide by

H2 6) and to storage of H2 by dissolution in metals.7

In order to develop better and more efficient

catalysts, it is important to determine what metal hydrogen

bond dissociation energies are needed to affect desired

chemical transformations. For example, a transition metal

hydride complex will not be a good hydrogenation catalyst if

its M-H bond strengths are so large that it is unable to

easily transfer hydrogens to an olefin.

To date, very little thermodynamic data have been

reported in the literature regarding metal hydrogen bond

dissociation energies and their dependence on various

ligands in the metal coordination sphere. The reaction of a

series of square planar rhodium (I) complexes of the formula

[P(p-tolyl)312RhClB (B=P(p-tolyl),, THTP, Py, MeIm)

with molecular hydrogen forming six coordinate rhodium

(III) dihydride complexes (Equation 1) provided an excellent

opportunity for this type of study.








[P(p-tolyl)3I2RhClB + H 2-- [P(p-tolyl)3I2RhC1B(H2) (1)




The complex [P(p-tolyl)3]2RhC1B, (where B=P(p-tolyl)3)

often termed "Wilkinson's catalyst," has been extensively

studied since it was first reported to be a good

hydrogenation catalyst by Wilkinson et al. in 1966. To

this date, a number of studies and mechanistic schemes have

been published.812 The most recent was proposed by

Halpern et al. and is shown Figure 2-1 where

L=P(p-tolyl)3 or P(C6H5 )3, S = solvent, and /C =C

represents some non-sterically hindered unconjugated olefin.

In concentrations of up to 0.1M excess phosphine the

predominant catalytic pathway is not through direct

hydrogenation of RhC1L3 I but through a dissociative

pathway where RhC1L3 dissociates a phosphine generating

the unsaturated three coordinate species RhC1L3 II. This

was found to activate H2 10 times faster than RhC1LL

generating another coordinately unsaturated complex

RhClL3(H)2 III which then reacts with an olefin giving

the six coordinate complex IV. Stepwise hydrogen transfer

to the olefin produces the alkane product and also

regenerates the three coordinate complex II. The catalytic

cycle then continues along the pathway contained in the

dotted circle.

A number of studies have been published attempting to

correlate changes in the ligand environment around the




























Figure 2-1. Mechanistic scheme proposed for the
hydrogenation olefins by "Wilkinson's
Catalyst."













L L

Rh L
CI L


L S
-S -L


Rh
CI L


L H
C Rh
ClI I L
.1= 1
f^ '-'


-c-C-
H H


L H H

Rh

S


S


m









rhodium center to catalytic activity. Most of these studies

involved changing all three phosphines with other phosphines

and other 2 electron donating bases. When a series

of para-substituted aryl phosphines were used, it was found

that catalytic activity increased as the electron donating

ability of the phosphine increased. This could be due to

the more electron donating phosphines making the rhodium

more basic improving its ability to activate H2 (as well

as to back bond into olefins).

When alkyl phosphines were used, which are even more

electron donating than aryl phosphines, catalytic activity

was diminished. These results can be interpreted in a

number of ways. First, the alkyl phosphines could be making

the rhodium center so basic that the rhodium hydride and

rhodium olefin complex are becoming too stable. Second,

steric factors may also be playing an important role. The

less bulky more electron donating phosphines may be binding

too tightly to the rhodium center preventing the necessary

phosphine dissociation steps that provide important

coordinative unsaturation. When the bulky aryl phosphines

were used,19 crowding around the metal center may be

increasing the tendency for ligand dissociation.

Further studies have used N, S, As and Sb

donors 6,17'20,21 around the rhodium center. Also, it was

reportedly that large excesses of pyridine, acetonitrile

and various sulfur donors added to solutions of

RhCl(P(C6H5)3)3 greatly reduce catalytic activity.








Due to the complexities of a catalytic system it is

difficult to predict exactly what reactions are being

affected by variations imposed on the system unless each

catalytic reaction step can be studied separately. This

study will hopefully give some insight into one segment of

the catalytic reaction, that is the effects of ligand

variation on the activation of molecular hydrogen by rhodium

complexes with structures similar to Wilkinson's.

A convenient and fairly simple method for making

derivatives with one ligand around the rhodium center of

Wilkinson's catalyst being varied was found making use of

the dimer bridge cleavage reaction shown in Equation 2.

Calorimetric studies on this22 23


P Cl P P B

Rh Rh + 2B (--> 2 Rh (2)

P Cl P P Cl

(P = P(p-tolyl)3)


and two other chloro bridged dimer systems [RhCl(CO)]2 24

and [RhCl(COD)2 225 have previously been reported by

the Drago research group.

The bridge cleavage reaction of the phosphine dimer

([RhCl(P(p-tolyl)3)212) was found to occur with nitrogen,

phosphorus and sulfur donors. Attempts using oxygen donors

were unsuccessful. By way of H and P NMR the

geometric structure of all the monomeric base adducts formed

were found to be the same; that is the base occupied a









position cis to the chlorine, as shown in Equation 2. Heats

obtained for the bridge cleavage reaction were incorporated

into the E, C and W correlation 26,27 (Equation 3) and EA

and CA parameters were generated for the three coordinate

acid (P(p-tolyl)3)2RhCl.



AH = EAEB + CA Cg + W (3)



The monomeric base adducts generated in solution were

also shown to react with molecular hydrogen at ambient

temperature and pressure (Equation 1). The six coordinate

rhodium (III) dihydride complexes formed (shown below) were

found to have the same geometric structure, again using 1H

and P NMR. The two inequivalent hydrides end up cis to



H

P H

Rh

Cl P

B





each other and the two equivalent phosphines occupy trans

positions.

This study deals specifically with the thermodynamic

aspects of the activation of molecular hydrogen by

Wilkinson's catalyst and its derivatives. Determination of








theAH for Equation 1 will provide us with a better

understanding of the influence of donor strength on the

activation of molecular hydrogen and on the average Rh-H

bond strength.

In addition, the resulting thermodynamic data could

possibly be entered into the E, C and W correlation and EA

and CA parameters could be determined for the five

coordinate acid "(P(p-tolyl)32)RhCl(H)2." By comparing

these E and C parameters to the E and C parameters of the

three coordinate acid "(P(p-tolyl)3)2RhCl," the effect of

oxidative addition of H2 on the acidity of the rhodium

center could be quantitatively determined.




Experimental



Materials

Toluene was dried and degassed by first refluxing over

CaH2 for 24 hours, collected by distillation under N2

and then subjected to freeze pump thawing (a minimum of 5

cycles using a mercury diffusion pump). It was then stored

in an inert atmosphere box until needed. Rhodium

trichloride 3H20 was purchased and used without further

purification.

All bases used in this study were purchased and

purified by the following methods. Tri-p-tolylphosphine was

recrystallized in hot absolute ethanol followed by drying








under vacuum. Pyridine (Py) was stored over KOH pellets

overnight, then distilled from BaO under N2 (middle

fraction taken, boiling range 112.0 112.50C).

Tetrahydrothiophene (THTP) was fractionally distilled over

CaH2 at atmospheric pressure. N-methylimidazole was

fractionally distilled over CaH2 at 10 mmHg pressure and

the middle fraction taken. Dimethylthioformamide was dried

over 4A molecular sieves. All liquid bases were degassed by

freeze pump thawing.

Ethylene, hydrogen and nitrogen gases used were

obtained from Linde (Union Carbide); the hydrogen and

nitrogen were of ultra-high purity ( 2 ppm 02, 3 ppm

moisture).

High capacity oxygen traps were purchased from L.C.

Company, Inc. These were designed to remove 02 to below

0.1 ppm with a gas flow of less than 3 liters per minute.

The traps contain a Mo based indicator which is blue-green

when activated and grey when spent. The traps were

periodically regenerated by placing them in a Lindberg

(model 123-8) tube furnace at 3750C while passing H2 gas

through them at a rate of 30 cc/min for 20 minutes.



Instrumentation

All air sensitive manipulations were performed in a

Vacuum Atmosphere Co. model HE-43-2 inert atmosphere box,

in specially designed Schlenck-ware glassware. Elemental








analysis was performed by the Microanalytical Laboratory,

University of Illinois, Urbana, Illinois.

Electronic absorption spectra were taken on a Cary

14-RI spectrometer equipped with a Varian constant

temperature chamber in the sample compartment. The chamber

was coupled to a Braun Thermoboy circulating temperature

bath which will maintain a constant temperature to

+0.2 0C. For temperatures below room temperature a cold

slush bath was added to the circulating system. The bath

consisted of 1/4" x 10' copper tubing coil suspended in a

dewar containing any number of solvent/CO2 or solvent/N2

slushes27 depending on the temperature desired. Methanol

was used as the circulating liquid at low temperatures to

prevent freezing and the sample compartment was continuously

purged with N2 to prevent moisture from condensing on the

sample cell.



Synthesis

Di-u=-chlorotetrakis(ethylene)dirhodium(I),

[Rh(C2H4)2Cl]2, was prepared according to a method

reported by Cramer.28 A solution of 0.6 g (2.28 mmoles)

RhCl 3 3H 0 dissolved in 1 ml of H 0 and 7.0 ml

methanol was added to a 125 ml Parr pressure bottle

containing a magnetic stir bar. The mixture was then purged

5 times with ethylene and finally pressurized to 20 psi.

After 15 hours of stirring the solution had changed color

and an orange-yellow precipitate had formed. The solid was








filtered and washed with anhydrous methanol in a N2 filled

glove bag and dried for 2 hours over P205. (Caution,

vacuum drying will result in the loss of ethylene from the

complex.) The complex can be stored indefinitely under N2

at 0 0C.

Di-c--chlorotetrakis(tri-p-tolylphosphine)dirhodium(I),

[Rh(P(p-tolyl)3)2C1 ]2 was prepared by combined methods

reported by Tolman et al.8 and Wilkinson et al. In an

inert atmosphere box 0.266 g [Rh(C2H4)2C1]2, 0.85g of

freshly recrystallized P(p-tolyl)3 and 17 ml toluene were

placed in a 50 ml round bottom flask equipped with a Vigreux

column. The resulting solution was refluxed with stirring

for 3 hours. While still warm, 35 ml of hexanes were added

and the solution was allowed to cool. The orange

precipitate that formed was filtered, washed with hexanes

and dried for 12 hours under vacuum. This complex is found

to decompose slowly with time and can be purified by

repeated recrystallization in toluene followed by addition

of hexanes. Analysis: C, 67.6; H, 5.7; Cl, 5.0.

Theoretical: C, 67.5; H, 5.7; Cl, 4.8.



Thermodynamic Measurements

Spectrophotometric titration was used to determine the

equilibrium constant of the following reaction:


(P(p-tolyl)3)2RhClB + H2 1===^ (P(p-tolyl)3)2RhClB(H)2


(4)








Due to extreme 02 sensitivity of the above reaction a

special hydrogen gas flow system was used to vary the

hydrogen partial pressure. A series of preset H2/N2 gas

mixtures were bubbled through a 0.1 cm UV-vis cell

containing a toluene solution of (P(p-tolyl)3)2RhClB.

Absorbance changes of the hydrogenated complex in the

visible range (350-650nm) were recorded after equilibrium

was reached with each hydrogen partial pressure.

Equilibrium constants were determined with the

Gasuptake program developed by Beugelsdijk29 in which the

best K and (Eo-E) were simultaneously determined. By

use of the best fit (Eo-E) calculated for each temperature

equilibrium constants were generated for each absorbance

change. These K's were then used in a van't Hoff analysis

to determine AH. In determination of 90% confidence

intervals forAH, a degree of freedom was subtracted for

each temperature used to account for holding the best fit

(Eo-E) constant.



Hydrogen Gas Flow System

A specially designed gas flow system was assembled as

shown in Figure 2-2. Hydrogen and nitrogen gas was first

passed through L.C. Company 02 traps removing 02 up to

0.1 ppm. The gases then enter a Matheson (Model 7352)

rotometer fitted with no. 610 tubes where the gases are

accurately mixed to the desired N2/H2 ratios. The

no. 610 tubes are the most sensitive flow meters available












































































4J


V)
U,

O




re











Ci


















ZI
z m






0.



CC



-0

--0
0




h-

--


CzJ Cj
2 m


a) L.
C (U
co

o :
r- 0


I I


I I,


I I


I '

I V
0r
I __


0)




I-
SCD


c
0

a 0


cr
0)



c


0 C
C
0











D









from Matheson and are essential for accurate determination

of N2/H2 ratios.

The mixed gas leaving the rotometer enters a brass

manifold made up of Swaglok hardware including five gas

valves. From here the gas is directed to a specially

designed bubbler containing a toluene solution of the

rhodium chloride dimer. The bubbler saturates the gas with

toluene so that when it is bubbled through the toluene

solution in the UV-vis cell, there is no appreciable

evaporation. The function of the oxygen sensitive rhodium

chloride dimer in the bubbler was to act as a final 02

trap.

The gas now saturated with toluene was bubbled through

the bottom of a highly modified 0.1 cm pathlength barrel

UV-vis cell (Figures 2-3, 2-4) situated in the thermostated

source compartment of a Cary 14 spectrophotometer. The

special cell contained a toluene solution of the solution

generated monomeric rhodium base adduct. Gas exiting the

cell was bubbled through a mineral oil bubbler which

maintained a constant internal pressure in the system. A

mercury monometer present in the system was used to

determine the internal pressure.

Solutions of the various monomeric base adducts were

generated by adding enough of each base to a rhodium dimer

solution to cause greater than 95% cleavage of the dimer.

The amount of each base added was based on equilibrium

constants reported by Farris30 and Hoselton23 for the




























Figure 2-3. UV-cell.

A. 1mm pathlength barrel cell
B. #7 0-ring joint
C. Teflon needle valves
























B






1 mm
I I


Front


Side




























Figure 2-4. UV-vis cell holder.

A. Thermostated cell holder
B. UV-vis cell
C. Aluminum cell holder
D. Rubber liner











bridge cleavage reaction. The general operating procedure

of the hydrogen gas flow system is described below in an

example experiment (refer to Figure 2-2).

The UV-vis cell and the toluene bubbler were removed

from the flow system and placed in an inert atmosphere box.

The cell was filled with a toluene stock solution that was

2.02 x 10-3M in [Rh(P(p-tolyl)3)2C112 and 8.111 x 102M

in P(p-tolyl)3. The toluene bubbler was loaded with

toluene that contained a spatula or two of

[Rh(P(p-tolyl)3)2C1]2. The teflon valves were closed

on both the cell and the bubbler. The cell and bubbler were

then removed from the inert atmosphere box and placed back

into the flow system.

Having all the valves closed except for valves #3, 6

and 9, the system was evacuated by opening valve #10 which

was connected to a vacuum pump. After closing valve #10,

the system was then purged with N2 by opening valve #8.

The evacuation-purge cycle was repeated a minimum of five

times. Following the final purge with N2, valves #8 and

#6 were closed and valves #4, 5 and 11 were opened.

Valve #8 is opened fully and the rate of N2 bubbling

was now controlled by the rotometer valve. Valve #9, which

is the major bypass valve, is slowly closed until a

satisfactory bubble rate is achieved in the toluene

bubbler. Too high a rate may cause toluene to foam over

into the built in trap.









Valves on the cell, #1 and #2, were then opened fully.

Valve #3, a minor bypass valve, was slowly closed until the

desired cell bubble rate was achieved, about 1 bubble per

second. Too rapid a bubble rate may cause evaporation.

After the rhodium monomer solution had come to equilibrium

with the N2 in the system, the valves on the cell were

closed and the visible spectrum was recorded.

Valve #3 was opened fully and the rotometer valves were

adjusted to give the first desired hydrogen partial

pressure. The bubbling through the toluene bubbler was

adjusted again by valve #9 and the system was allowed to

purge for 30 minutes. Following purging, the cell valves

were opened and bubbling through the cell solution was

adjusted using valve #3. Once the cell solution had come to

equilibrium, indicated by no further change in the

absorbance spectrum, the cell valves were closed and the

visible spectrum was recorded. This same procedure was

followed for each additional hydrogen-nitrogen gas mixture.

The rotometer was calibrated prior to the experiment

using a soap bubble flow meter connected to valve #7.





Results and Discussions



Fairly extensive thermodynamic and structural studies

have previously been done by Farris30 and Hoselton on

the bridge cleavage reaction (Equation 2) of the chloro








bridged dimer [RhCl(P(p-tolyl)3)212 with various

sulfur, nitrogen and phosphorous donor bases. Equilibrium

constants and heats of reaction obtained for the reaction

are shown in Table 2-1.

Based on these results monomer solutions used for the

reaction with hydrogen (Equation 1) were made by adding a

sufficient quantity of base to a rhodium dimer solution to

cleave the dimer by greater than 95%. The ratios of base to

dimer concentration suggested by Farris were as follows:

P(p-tolyl)3, 40:1; THTP, 60:1; pyridine, 40:1. The

resulting monomer solutions were then loaded into the cell

of the specially designed flow system.

Great care was taken to exclude 02 while preparing

the monomer solutions because of their extreme oxygen

sensitivity. Fresh monomer solutions were made before each

experiment in an inert atmosphere box. The rhodium dimer

which was also oxygen sensitive had to be recrystallized

periodically to maintain its purity.

Previously, Farris30 had attempted to determine heats

for the reaction of the various base monomers with molecular

hydrogen using a variable temperature spectrophotometric

titration method utilizing a static gas system. This proved

unsuccessful mainly due to problems with oxygen

contamination.

An alternative to a static type system is a gas flow

type system which was developed for this study (Figure 2-2)

and described in detail in the experimental section. The








Table 2-1. Thermodynamic Results
[RhC1(P(p-tolyl)3)212
24+20C.


for the Interaction of
with Various Bases at


a a b
Base AH K K (spec)


c d e
P(p-tolyl)3 -4.7 + 0.3 Large

Pyridine -4.9+0.2 Large 183+22

THTP -1.9+0.2 (1.0+0.4)x102 -

N-methylimidazole -6.6+0.2 (2.4+2.0)x104




a Determined by Calorimetry.23 b Determined
spectrophotometrically.30 c Units of Kcal/mole of monomer
formed. d Reported errors are standard deviations.
e Keq. is unknown but assumed to be very large for purpose
of calculating AHav.








main advantage of this system is that a positive pressure of

H2 and N2 is maintained throughout an experiment which

reduces the chance of 02 leaking into the system from the

outside atmosphere. One disadvantage of a flow system is

that care must be taken to prevent solvent evaporation due

to continuously bubbling the various H2:N2 gas mixtures

through the cell solution. This potential problem was

solved first by maintaining a slow continuous bubble rate

through the cell solution and second, by placing a toluene

bubbler in the gas line before the cell. This aided by

prior saturation of the H2:N2 gas with toluene

minimizing any solvent evaporation due to bubbling gas

through the cell.

In order to test the flow system a toluene solution of

air stable tetraphenylporphyrin solution was placed in the

cell of the flow system. After 2 hours of continuously

bubbling N2 through the solution, there was no significant

change in its UV-visible spectrum indicating that the

toluene bubbler was effectively preventing solvent

evaporation.

The changes that occurred in the electronic spectra

upon bubbling various partial pressures of hydrogen through

solutions of (P(p-tolyl)3)2RhClB are illustrated in

Figures 2-5 and 2-6 for the bases P(p-tolyl)3 and THTP

respectively. In both cases the shoulder (centered at

430nm for P(p-tolyl)3 and 420nm for THTP) on the side of



























Figure 2-5. UV spectrum of a toluene solution of 1.86xO13M
[RhCl(P(p-tolyl)3)212, and 7.47x10O-2
P(p-tolyl)3 at 35.00C in equilibrium with
various pressures of H2. (1: PH2 = 0.00 atm,
2: PH2 = .023 atm, 3: PH2 = .045 atm,
4: PH2 = .075 atm, 5: PH2 = .145 atm,
6: PH2 = .363 atm, 7: PH2 = .909 atm.)














1.0







0.8







0.6


U
C
.0
0

< 0.4







0.2


Wavelength, nm


P(p-tolyl )3


350 400 450 500 550



























Figure 2-6.


UV spectrum of a toluene solution of 2.98x10-3
[RhCl(P(p-tolyl)312, and S.79xlO-2M THTP at
10.00C in equilibrium with various pressures
of H2. (1: PH2 = 0.00 atm, 2: PH2 = .015 atm
3: PH2 = .102 atm, 4: PH2 = .217 atm
5: PH2 = .965 atm.)























THTP


Wavelength, nm


0.4





0.2





0


400 440 480 520 560 600


640








a large charge transfer peak decreased in intensity as the

concentration of H 2 increased.

An isobestic point was observed in the P(p-tolyl)3

system at about 350nm probably indicating that only two

chromophors were present in solution and that the rhodium

monomer was being cleanly converted into the six coordinate

cis-rhodium dihydride complex. This provided further

evidence supporting what had previously been shown by NMR

studies of the reaction. (Caution must be taken not to rely

solely on the appearance of an isobestic point. An

additional experimental method should be used to determine

the actual reaction taking place.)

The spectral data obtained were analyzed using the

Rose-Drago31 equation (Equation 5)



K-1= PH2 /Aojb (Eo-E) 1 (5)

A Ao


which was modified for the hydrogen uptake experiment.32

In this equation PH2 is the hydrogen partial pressure

bubbled through the solution, b is the pathlength of the

cell, A is the initial concentration of the acid (Rh

monomer), A is the absorbance of the cis dihydride complex,

Ao is the absorbance of the acid solution (with no H2

present), (Eo-E) is the difference in the molar absorptivity

between the hydrogenated and monohydrogenated complex and

K-1 is the inverse of the equilibrium constant for the
K is the inverse of the equilibrium constant for the


reaction.








The spectral data and the best determined K and (Eo-E)

values at various temperatures are listed in Table 2-2 for

the bases P(p-tolyl)3 and THTP. An equilibrium constant

previously reported by Tolman et al. of 40 + 3 atm at

250C corresponds well with the results of this study where

at 30.10C a value of 29.9 + .26 atm-1 was obtained.

Attempts to study the pyridine adduct produced

inconsistent results. This system, unlike the previously

described base adduct systems, was not as well behaved and

presented a number of complications. The first complication

encountered was the extreme 02 sensitivity of the pyridine

system. Additional precautions to exclude 02 had to be

taken during the generation of the monomer solutions and

when flushing the flow system.

Initial spectral results established that the

equilibrium constant for the pyridine system was at least an

order of magnitude greater than was observed for the

P(p-tolyl)3 and THTP adducts. This was a problem because

the lowest possible H2 partial pressure that could be

accurately delivered by the rotometer was converting most of

the pyridine monomer into the hydrogenated form.

To overcome this problem, another Matheson rotometer

was added to the flow system. The rotometers were connected

in series as illustrated in Figure 2-2. Rotometer 1 diluted

the H2 with N 2. The mixture then entered the second

rotometer reducing the H 2 concentration with N2 once

again. The double dilution feature of the system allowed








Table 2-2. Thermodynamic Data for the Reaction of
RhClB(P(p-tolyl)3)2 + H2k-4RhC1B(P(p-tolyl)3)2H2.


Temp (oC)


P(p-tolyl)3


9.0


30.1

35.0

44.0

53.8


113.7

29.9

24.4

12.2


7.27


b c
AH = 11.0 + .5 Kcal/mole


THTP


10.0

15.0

20.6

30.0

40.0

50.0


77.4

47.6

33.9

21.5

13.2

5.91


AH = 11.6 + 1.0 Kcal/mole


a Best fit K. b Determined by van't Hoff analysis,
based on generating an equilibrium constant for each
absorbance change after the best Eo-Ewas determined and
held constant. c 90% confidence interval.








for the accurate delivery of the very low H2:N2 ratio

gas needed.

A problem encountered which was mainly responsible for

the inconsistent results obtained was the observation of a

slow secondary reaction occurring during the course of the

hydrogen uptake experiment. As seen in the other base

systems, reduction in the shoulder absorbance was observed

on exposing the pyridine monomer solution to various H2

partial pressures. Once the solution had come to

equilibrium with a hydrogen partial pressure, the shoulder

absorbance increased slowly with time instead of remaining

constant. This phenomenon is attributed to a slow secondary

reaction taking place, possibly due to the presence of a

large excess of pyridine in solution with the newly formed

rhodium dihydride complex. (As in the pyridine

experiments, similar problems were encountered with the

1-methylimidazole adduct experiments.)

The two systems that could be investigated

(P(p-tolyl)3 and THTP) fortunately provided a substantial

variation in the strength of binding to the Rh(I) center.

Enthalpies of hydrogen binding were determined by use of the

van't Hoff equation where a plot of -RlnK vs. 1/T yields a

straight line with a slope of AH and an intercept of AS.


-R Ln K = AH AS (6)
t~








Due to the large error limits involved with

calculating AH from a plot of only five or six points

corresponding to K at each temperature, a method suggested

by Breese33 was employed to utilize all of the data

obtained in the equilibrium constant experiments. His

method involves calculating the best K and (E -E) at each

temperature. The best (E -E) at a specific temperature is

then substituted into the Rose-Drago equation and an

equilibrium constant is calculated for each absorbance

change. Thus a number of equilibrium constants were

determined at each temperature (Table 2-3) instead of only

one. The end result was a AH with a much smaller error

limit. Care must be taken in determining standard

deviations to reduce the degrees of freedom by one for every

temperature to account for fixing the value of (E -E).

The van't Hoff plots obtained are shown in Figures 2-7 and

2-8 respectively for the P(p-tolyl)3 and THTP systems.

The enthalpies obtained are shown in Table 2-2. These

enthalpies can be combined with the heat of dissociation of

H2 to produce the average metal-hydrogen bond dissociation

energies (Table 2-4).



RhClB(P(p-tolyl)3)2(H)2 --- RhClB(P(p-tolyl)3)2 + H2 (7)

H2 --->2H- (104.2 kcal mole-1) (8)


RhClB(P(p-tolyl)3)2(H)2--- -RhClB(P(p-tolyl)3)2 + 2H-


(9)











Table 2-3. Thermodynamic Data on
RhCl(P(p-tolyl)3)2B + H2+- RhC1(P(p-tolyl)3)2BH2.



B (Rh)(M) PH (atm) Abs Kc


P(p-tolyl)3


Temp. = 9.00C
(Eo-E)a = 1209
Ka = 113.7
CSD = 3.8
MSD/CSDb = 1.4


Temp. =







Temp. =


30. 10oC
(Eo-E) = 951.3
K = 29.9
CSD = .26
MSD/CSD = 1.5



35.0OC
(Eo-E) = 1095
K = 24.4
CSD = .85
MSD/CSD = 1.5


Temp. = 44.OOC
(Eo-E) = 1007
K = 12.2
CSD = .19
MSD/CSD = 1.7



Temp. = 53.80C
(Eo-E) = 1047
K = 7.27
CSD = 3.7
MISD/CSD = 1.9


4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3


4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3


3.72 x 10-3
3.72 x 10-3
3.72 x 10-3
3.72 x 10-3
3.72 x 10-3


4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3


4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3
4.05 x 10-3


.0246
.0490
.0813
.156
.392
.981


.0237
.0474
.0786
.151
.379
.948


.0227
.0454
.0752
.145
.363


.0224
.0449
.0741
.143
.359
.898


.0423
.0701
.135
.338
.844


.356
.421
.444
.465
.477
.483


.158
.227
.270
.317
.354
.371


.135
.221
.267
.318
.364


.087
.141
.194
.263
.334
.371


.090
.139
.217
.310
.358


108.5
125.0
119.7
120.6
96.2
74.1


29.3
30.2
29. 8
30.7
29.9
27.4


21.8
26. 1
25.3
24.6
23. 1


12.1
11.8
12.2
12.7
12.6
11.2


6.37
6.96
7.78
8.04
6.41









Table 2-3. Continued


Temp. =






Temp. =





Temp. =






Temp. =






Temp. =






Temp. =


PH 2(atm) Abs


a Best fit K and (Eo-E).


b CSD is conditional standard


deviation and MSD is marginal standard deviation for best
fit K. c Equilibrium constant for each absorbance change
after best fit (Eo-E) was determined and held constant.


THTP


(Rh)(M)


10.00C
(Eo-E) = 847
K = 77.4
CSD = .99
MSD/CSD = 1.2


15.00C
(Eo-E) = 667
K = 47.6
CSD = .94
MSD/CSD = 1.4

20.60C
(Eo-E) = 505
K = 33.9
CSD = .68
MSD/CSD = 1.5


30.00C
(Eo-E) = 541
K = 21.5
CSD = 1.1
MSD/CSD = 1.8


40.0C
(Eo-E) = 442
K = 13.2
CSD = .31
MSD/CSD = 2.2


50.00C
(Eo-E) = 883
K = 5.91
CSD = .31
MSD/CSD = 2.2


5.66
5.66
5 66
5.66
5.66


6 54
6.54
6.54
6 54


4 75
4.75
4.75
4.75
4.75
4.75

5 14
5.14
5.14
5.14
5.14


5.04
5.04
5.04
5.04
5.04


5.98
5.98
5.98
5.98
5.98


10-3
10-3
10-3
10-3
10-3


10-3
10-3
10-3
10-3


10-3
10-3
10-3
10-3
10-3
10-3

10-3
10-3
10-3
10-3
10-3


10-3
10-3
10-3
10-3
10-3


10-3
10-3
10-3
10-3
10-3


.0147
.102
.217
.371
.966


.0147
.102
.217
.371


.0144
.0456
.0993
.212
.363
.946

.0142
.0986
.210
.359
.932


.0136
.0946
.202
.344
.898


.0131
.0912
.194
.331
.864


.256
.423
.453
.464
.474


.178
.365
.397
.412


.078
.144
.188
.211
.222
.231

.058
.195
.229
.245
.262


.030
.113
.159
.180
.189


.018
.190
.283
.350
.440


78.0
73.6
79.0
81. 2
90.9


47.1
50.0
46.2
45.3


33.6
33.0
36.4
34.4
34.3
27 6

18.5
23.8
22.2
20.6
17.5


12.1
12. 1
15.0
16.53
9.3


2.69
6.18
5.95
5.93
5.80


~




























Figure 2-7.


Van't Hoff plot for the P(p-tolyl)3 system.








0


Ln K vs. 1/T




0























0

LEGEND
=P(p --tolyi)3





3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7
1/T x 10-3




























Figure 2-8. Van't Hoff plot for the THTP system.












Ln K vs. 1/T


--Ap-


LEGEND
* = T'IITP


1.0 3.1 3.2 .3.3 3.4 3.5 3.8 3.
1/T x 10-3


__











Table 2-4. Average M-H Bond Energies.


Compound


AH (av.M-H) Kcal/mole


Rh(P(p-tolyl)3)3C1(H)2 57.6 + .3

Rh(P(p-tolyl)3)2(THTP)C1(H)2 57.9 + .5
a
Ir(CO)XL2H2 57-61
b
Co-H 39 + 6


a L. Vaska and Werneke.34
Armentrout.35


b J. Beauchamp and








Comparing the values obtained for the two base

systems, it is immediately apparent that within

experimental error the average metal-hydrogen dissociation

energies are identical. Thus the rhodium-hydrogen bond

strength was found to be insensitive to varying one ligand

around the rhodium center.

Based on the observation that two vastly different

bases (P(p-tolyl)3 and THTP) had little or no effect on

the Rh-H bond strength, it is probably safe to assume that

other donor bases, such as pyridine, would also have little

effect on the Rh-H bond dissociation energy. Thus the

adverse effects the addition of pyridine (mentioned in the

introduction) has on hydrogenation by Wilkinson's catalyst

is probably due to factors other than those affecting the

catalyst's metal hydrogen bond strength.

A variable temperature 1HNMR study reported by

Hoselton et al.22 showed that that following kinetic

process was taking place in the presence of excess pyridine:




H K H
P I H P i H
Rh Rhf (10)
C r' P B | P
B Cl
B = Py


The pyridine which is trans to a hydride was shown to

exchange with free pyridine in solution. A proposed

intermediate in the reaction was a trigonal bipyramid








"Rh(P(p-tolyl)3)2C1(H)2" species. The incoming

pyridine could attack a position trans to either hydride

which would account for the observed averaging of the two

inequivalent hydride resonances with increasing temperature.

A similar kinetic process was reported for an analogous

Wilkinson's catalyst system, Rh(P(C6H5 )3)3C1(H)2, by

Tolman et al. The first order rate constants observed for

exchange of pyridine and P(C6H5)3 for the similar

intermediate were 50 sec and 1000 sec respectively at

25 0C. This ligand exchange process serves a very

important role in the proposed mechanism of Wilkinson's

catalyst in that it provides an important open coordination

site for the binding of an olefin. The fact that the rate

of ligand exchange involving pyridine is an order of

magnitude smaller than that observed for P(CH5 )3 may

account for added pyridine having adverse effects on

catalytic activity due to efficient competition for the

olefin binding site by pyridine.

The metal hydrogen bond energies determined can be

compared to others reported in the literature (Table 2-4).

In a study34,36 of the influence of ligand variation in

the system Ir(CO)XL2 (where L is a series of phosphines

and X is Cl Br and I-), enthalpies of

hydrogenation of 2.7 to 18.1 kcal mole-1 were measured

corresponding to average metal hydrogen bond energy

variations of 53 to 61 kcal mole-1. The gas phase neutral








Co-H bond dissociation energy, determined by ion beam
351
techniques,35 was reported as 39+6 kcal mole





Conclusion



The main purpose of this study was to investigate the

process of hydrogen activation by transition metals and to

understand what effect ligand environment has on the metal

hydrogen bond dissociation energy.

The enthalpy of hydrogen activation by

Rh(P(p-tolyl)3)2ClB was obtained for B = P(p-tolyl)3

and THTP. Attempts to determine enthalpies for pyridine and

N-methylimidazole systems were unsuccessful. The enthalpies

that were obtained were used to determine average metal

hydrogen bond dissociation energies. This provided a good

estimate of the hydrogen bond energies associated with a

good hydrogenation catalyst (Wilkinson's catalyst) and with

one of its derivatives. It also greatly increased the

amount of available data presently in the literature.

The two bases THTP and P(p-tolyl)3 fortunately were

substantially different so that the effect of ligand

variation could be observed. The equilibrium constants

obtained for Equation 1 parallel the basicities of the two

bases toward the rhodium (I) center. However, the

enthalpies obtained for the two systems were remarkably

independent of the donor strength.37 Thus the average








metal-hydrogen bond energy was found to be insensitive to

differences in the two ligands. Assuming this result could

be extended to the pyridine system, the adverse effects of

adding pyridine to the catalytic system are probably not due

to pyridine's effect on the catalyst's metal hydrogen bond

strength but more likely due to pyridine blocking important

coordination sites on the catalyst.

A final outcome of this study was that a convenient

method was developed for the determination of enthalpies of

hydrogen activation.














III. IMMOBILIZED HOMOGENEOUS CARBON MONOXIDE
REDUCTION CATALYSTS

Introduction


Immobilization of homogeneous transition metal

catalysts on various organic or inorganic supports, often

termed "hybrid38 or "HETMETCO"39 (heterogenized

transition metal complex) catalysts, is presently a very

exciting and active field of catalysis research. The over-

riding goal in this area of research is to combine the

advantages of heterogeneous and homogeneous catalysts, while

minimizing their inherent deficiencies. A fair number of

articles have been published reviewing the growing volume of

reported studies involving synthesis, characterization and
38-44
catalytic properties of supported homogeneous catalysts

A list showing the major advantages of heterogeneous

vs. homogeneous catalysts is given in Table 3-1.

Heterogeneous catalysts, viewed by many as being first

generation catalysts, have been widely used in industry for

many years. One of the major advantages of these catalysts

is their ease of separation from reaction products. Other

advantages are that heterogeneous systems are generally more

thermally stable and less corrosive than homogeneous

systems. A major drawback that has hindered the development

of heterogeneous catalysts has been the limited number of








physical techniques available to characterize these

systems. Although numerous studies over the years have

revealed, little information about the actual active

catalyst species, recent advancements in surface techniques

such as SEM, ESCA, XPS, Auger etc. have provided a better

understanding of the nature of these catalysts.


Table 3-1. Advantages of
Catalysts.


a Homogeneous Versus Heterogeneous


Homogeneous vs. Heterogeneous


1. more active 1. separation of catalyst
from product

2. reproducible 2. minimizes reactor
corrosion
3. milder operating
conditions 3. more thermally stable

4. characterization

5. more selective


Homogeneous catalysts, termed second generation

catalysts, are in many ways superior to heterogeneous

systems. These catalysts are generally more active, more

selective, and because these catalysts are derived from

discrete soluble transition metal complexes they lend

themselves to study by a large number of physical methods.

They have the advantage that their properties can be widely

varied by changing the ligand environment around the metal








center and because the composition of the catalyst or

catalyst precursor is known these systems are very

reproducible.

Industrially, homogeneous catalyst systems are much

harder to handle then heterogeneous catalyst systems. One

major problem involves the separation of the often precious

metal catalyst from the reaction products. Usually, unless

the reaction products can be easily distilled from the

reaction mixture, such as in the Monsanto45 process for

the carbonylation of methanol to acetic acid, the

catalyst-product separation can be a very costly process.

Other disadvantages of homogeneous systems are that they are

generally more corrosive than heterogeneous systems, and the

reaction medium is limited to the solvents in which the

catalyst is soluble.

Immobilization of homogeneous catalysts may not only

combine the advantages of homogeneous and heterogeneous

catalysts but has the potential for some new desirable

features. One feature is site isolation of catalytic

species. This would prevent catalyst site interactions

which are often a mechanism for catalyst deactivation.

There is also the potential for site interaction between

similar or different catalytic sites placed in close

proximity to one another. This could possibly mimic some of

the chemistry demonstrated by enzyme systems.

In this study attention will be focused on the use of

immobilized homogeneous catalysts in the area of carbon








monoxide reduction catalysts. Known homogeneous ruthenium

and iridium CO reduction catalysts have been ionically or

covalently attached to inorganic supports. The development

of these catalysts was geared toward discovering catalysts

capable of selectively producing chemicals from mixtures of

CO and H2 gases, referred to as "syngas," under very mild

reaction conditions. The supports chosen were silica gel

and alumina functionalized by various organosilanes.

In the course of this study a novel catalytic system

for the selective conversion of syngas and HC1 to

chloromethane was discovered (Equation 1).



H2 + CO + HC1 catalyst) H20 + CH3C1 (1)



This reaction occurs under extremely mild conditions, 250C

and 1 atmosphere pressure, and has the potential to be an

industrially significant process.

Processes involving the conversion of syngas, such as

the Fischer-Tropsch synthesis46 described in Equation 2,

have been known and extensively studied and reviewed since

the early 1900's.47-50 In the last decade concerns over

dwindling petroleum supplies have sparked a renewed interest




CO + H2 catalysts) Hydrocarbon Products (2)
(Alkanes, Alkenes, Oxygenates)








in the utilization of syngas as a viable source of chemicals

and motor fuels. The Fischer-Tropsch process, which

operates at elevated temperatures and pressures, utilizes a

heterogeneous catalyst usually consisting of one of the

three metals Fe, Co and Ru. The major disadvantage of the

Fischer-Tropsch synthesis which has plagued its development

since its discovery is its inherent lack of selectivity.

Early studies51 ,52 demonstrated that the product

distribution produced by traditional Fischer-Tropsch

catalysts obeys the simple polymerization model below

(Schutz-Flory distribution), Equations 3 and 4. This model

assumes chain growth occurs by incorporation of single

carbon fragments and that the probability of chain growth is

Wn = n (1 o< )2 o< n-1 (3)

rp (4)
rp + rt


independent of chain length. The term Wn represents the

weight fraction of a certain carbon number n, is defined

as the probability of chain growth and is equal to the rate

of polymerization divided by the sum of the rate of

polymerization and the rate of termination, and (1-oc)

represents the probability of chain termination.

In order to improve selectivity one needs to deviate

from this type of Schultz Flory kinetics. Recently, there

have been a number of reports53 of product distributions

deviating significantly from this polymerization model.

Product distributions have been altered by using new metal








loading techniques,54 selective catalyst poisoning55

(i.e. partial sulfiding) and the use of shape selective

supports such as zeolites.5658 Although these reports

look promising, further work needs to be done to establish

whether these effects are experimental artifacts or are due

to actual mechanistic effects.

Due to their potential for greater product selectivity,

homogeneous carbon monoxide reduction catalysts have been a

subject of active research in the last decade. A number of

systems involving primarily cobalt59 60 rhodium61 62 or

ruthenium63,64 have been developed for the production of

mixtures of oxygenated products (i.e. methanol,

methylformate, ethylene glycol) but most operate under

extreme conditions, with pressures usually exceeding 1000

atmospheres.

More recently several homogeneous systems have been

developed which operate under moderate conditions (below

1000 atmospheres). Knifton at Texaco reported an active

catalyst for the direct synthesis of ethylene glycol from

syngas at 430 atmospheres and 140-2500C. The catalyst

consisted of a number of ruthenium precursors dissolved in a

highly polar molten quaternary phosphonium salt. The active

catalyst species based on IR and IH NMR studies was

believed to be the anion [HRu3(CO) 11

Two ruthenium based systems were reported by

Dombek66,67 at Union Carbide. The first involved the

monomer Rh(CO)5, evidenced by high pressure IR and








believed to be the active catalytic species for the

conversion of syngas to methanol at 2300C and 340

atmospheres pressure. The addition of carboxylic acids was

found to promote the formation of glycol esters.

The second system involved a substantially more active

ruthenium catalyst promoted by ionic iodide promoters and

polar solvents. At 2300 C and 850 atmospheres H2 and CO,

the primary product was free ethylene glycol. The catalyst

precursor, Ru3(CO)12 was thought to be converted to the

two anions [Ru(CO)3 3]~ and [HRu3(CO)1] by the

reaction below (Equation 5) in a 2:1 ratio which was found

to be the optimum ratio for maximum ethylene glycol

production. The absence of either one of the anions results

in an inactive system.


Ru3(CO)12 + 31- + H2---)2[HRu3(CO)11] + [Ru(CO)3I3]- + 3CO
(5)


In 1977 Muetterties and coworkers68,69 reported a

novel homogeneous system which converted syngas to light

saturated hydrocarbons under very mild reaction conditions,

1800C and 1-2 atmospheres pressure (Equation 6). The

catalyst consisted of the cluster Ir4(CO)12 in a NaCl-A1C13


1r4(CO)12
CO + 3H2 ----I4C1 >Cl to C4 alkanes, (CH3Cl) (6)
AlC13-NaCl melt
1800C, 1-2 atm


melt solvent and is the first reported homogeneous system

capable of producing non-oxygenated products from syngas.








This report was most significant in that it demon-

strated the importance of Lewis acid interactions with metal

bound CO, such as demonstrated below, allowing the CO to be

reduced under mild conditions. Studies by Collman et al.70



M--CO***-Al Cl3



further substantiated the above finding but also showed that

methylchloride was a reaction product, indicating A1C13

was being consumed during the reaction.

A similar system reported by Muetterties and Choi71

showed that Os3(CO)12 in a BBr3 melt was also an active

catalyst for the production of light alkanes under similar

conditions and in this case methylbromide was detected as a

reaction product.

Methylchloride is an important commodity chemical and

its uses and corresponding consumption in 1977 are listed

in Table 3-2.



Table 3-2. Consumption of Methyl Chloride 1977.


Uses Thousands of Metric Tonsa


Methylchlorosilanes 175
Tetramethyl Lead 51
Butyl Rubbers 23
Other 28
Total 277


a Data obtained from "Chemical Economics Handbook".72








The two principal commercial processes for

producing chloromethane are the chlorination of methane and

the methanol-hydrogen chloride process shown below

(Equations 7 and 8). The chlorination of methane process



> 4000C

C12 + CH4 )CH3C1 + HC1 (7)



> 2800C

CH30H + HC1 )CH 3C (8)



can be accomplished with or without a catalyst. Typical

catalysts for the process are KC1, CuCl, and CuC12 melts.

Although the primary product of the process is chloromethane

other multichlorinated methanes are produced as well.

Presently, the preferred process is the methanol-HC1 route,

which is very selective for chloromethane, up to 97.8%.

Typical catalysts for this process are gamma alumina or

phosphoric acid on activated carbon.

An alternative synthesis reported in 1977 by
74
Vannice at Exxon demonstrated that syngas plus a

chloride source, such as C12 or HC1, could be used to

produce chlorinated hydrocarbons at 200 to 10000C and 0.1

to 500 atmospheres pressure. The system consisted of a well

dispersed supported catalyst containing Pt/Re, Pt/Ir, Pt,

Ir, or Re on an acidic support such as alumina. The

predominant product was methyl chloride but other products








such as multichlorinated methanes, ethene, propane and butyl

halides were also produced.

Over the years a large number of supports, both organic

and inorganic, have been utilized to heterogenize transition

metal catalysts. A fairly complete list40 is shown in

Table 3-3.



Table 3-3. Materials Used to Support Metal Complexes.


Inorganica

silica

zeolites

glass

clay

metal oxide
(i.e. A1203,
Ti02, MgO)


Organica

polystyrene

polyamines

silicates

polyvinyls

polyallyls

polybutadiene

polyamino acids

urethanes

acrylic polymers

cellulose

cross-linked dextrans

agarose


a List obtained from Hartley and Vezey.40




In the work presented here both functionalized silica

gel and alumina supports were used to ionically attach known








homogeneous CO hydrogenation catalysts. The choice of these

supports over other available supports was based on their

thermal stability, their intrinsic acidic nature, and the

large variety of silanes (containing functional groups) that

can be bond to the surface for metal attachment.

The surface of these supports consists of hydroxyl

groups which can react with silanes of the formula

R3Si(CH2) B (R=C1, OEt, OMe) by way of the simple

condensation reaction shown in Equation 9.75



H (OH)3-y
H + (OEt)3SiCH2B ---- -Si(CH2B + 3 ETOH (9)
-OH



The symbol B represents any number of heteroatom groups

(i.e. -P(Ph)2, -NH2' -SH) linked to the surface by a

variable length organic chain. The letter Y represents the

number of Si-O-Si bonds to the surface, which is dependent

on the concentration of organosilane being used.76 These

attached organosilanes can be used directly or modified to

support transition metal complexes. The B groups used in

the work presented here were -P(Ph)2 and -N(CH3)3+I.












Experimental



Reagents

All solvents and reagents were used as purchased unless

otherwise stated. Triethoxypropylaminosilane and

2-(diphenylphospho)ethyltriethoxysilane were purchased from

Petrarch Chemical. Iridium trichloride 3H20, Ir4(CO)12 and

Rh6(CO)16 were purchased from Strem. Triruthenium

dodecacarbonyl, Rul3 and [RhCl(CO)2]2 were purchased

from Alfa. Sodium chloride was dried prior to use in an

oven at 1400C.

Hydrogen was purchased from Air Products. Both carbon

monoxide grade C.P. 99.5% and hydrogen chloride technical

grade 99.0% gas were purchased from Matheson.

Silica gel was purchased from W.R. Grace and was

Davison grade #62. Its wide pore diameter was 14 mm, pore

volume was 1.1 cm 3/g and had a specific area of 340 m2

Alumina used was Acid, Brockman Activity I, mesh 80-200.

ZnO was purchased from Aldrich.



Instrumentation

All air sensitive manipulations were performed in a

Vacuum Atmosphere Co. model HE-43-2 inert atmosphere box.

Elemental analyses were performed by the Microanalytical

Laboratory, (C,H and N) University of Florida, Gainesville,








Florida, and by Galbraith Laboratories Inc. (Ir and P by

ICP) Knoxville, Tennessee.

GC analyses were performed on two different gas

chromatographs. One was a model 3700 FID/TC Varian gas

chromatograph equipped with a 1/16 inch x 1 meter stainless

steel column packed with Chromosorb P supported diethylene

glycol adipate (DEGA)and a 1/16 inch x 10 foot stainless

steel Poropak Q column. The other was a model 940 FID

Varian gas chromatograph equipped with a 1/16 inch x 8 foot

stainless steel Poropak Q column.

Infrared spectra were recorded on a Perkin Elmer 283B

spectrometer and also on a Nicolet 20DX using diffuse

reflectance (spectra performed by Nicolet).

GC mass spectrometry was performed on a AEI MS30 mass

spectrometer using a KOITOS DS55 data station and equipped

with a PYE Unicam 104 GC containing a 1/4 inch x 5 foot

Poropak Q column. Samples were run by Dr. R. King at the

Microanalytical Laboratory, University of Florida,

Gainesville, Florida.



Fixed Bed Flow Reactor

A specially designed gas flow system was assembled as

shown in Figure 3-1. Hydrogen and carbon monoxide first

enter a Matheson (model 7352) rotometer fitted with No. 610

tubes where the gasses were accurately mixed to the desired

H2CO ratio. When needed, HC1 gas was passed through a

Matheson #601 flow meter where it was combined with the































































0


co
0)
0


O
1-e








x









00
1-




40



















I 0r
I rI


C-
w L








already mixed syngas. The amount of HC1 added was regulated

by teflon needle valve #1. Problems were encountered with

maintaining an even flow of HC1, and a pressure release

bubbler (D) placed before the flow meter proved helpful.

The gasses then were passed through the catalyst bed which

consisted of a powder catalyst packed on top of a glass frit

and held in place with a plug of glass wool. The gasses

leaving the reactor tube were then passed through a

CO2/acetone cold trap and exited finally through a mineral

oil bubbler. The reactor contained two septa for syringing

gas samples for analysis before and after the reactor. The

flow of gasses through the catalyst was regulated by the

teflon needle valve #2. As this valve was closed, a greater

amount of the feed gas was forced through the reactor tube

as opposed to exiting through the mineral oil bubbler (C).

The reactor was designed so that the reactor could be

assembled in an inert atmosphere box. The bypass arm B was

designed for purging the system.

Surrounding the reactor tube was a Lindberg (model

123-8) thermostated tube furnace used to regulate the

reactor temperature. The entire system was made of glass

or teflon tubing due to the corrosive nature of HC1 gas.



Synthesis

Silica Gel Bound Propylaminosilane (silica gel)^/,^NH2

The functionalized support was prepared by a method

reported by Leyden et al.77 In a 125 ml Erlenmeyer flask,








10 g of silica gel was combined with 50 ml toluene. After

stirring for 10 minutes, 5 ml of triethoxypropylaminosilane

was added. The mixture was stirred for an additional

30 minutes, filtered and the solid was washed with copious

amounts of toluene. The solid was air dried followed by

heating in an oven at 800C for 24 hours. Analysis: C,

3.74%; H, 1.08%; N, 1.40%. (Theoretically represents

1.0 mmole silane per gram silica gel.)



Trimethylammonium Iodide Salt of Silica Gel Bound

Propylaminosilane (silica gel )','N(CH3) 31

The functionalized support was prepared by a method

reported by Zombek.78 In an Erlenmeyer flask equipped
with an addition funnel and blanketed with N2, 10 g of

propylaminosilane bound silica gel, 1.13 g 2,6-lutidine and

80 ml DMF were combined. A solution ( 30 ml) made up of

5.67 g Mel in DMF was added dropwise to the mixture while

stirring. Stirring was continued for 15 hours at ambient

temperature. The reaction mixture was filtered and the

solid washed with acetone and benzene followed by vacuum

drying (0.1 mm Hg) for 24 hours. Analysis, C, 4.52%; H,

1.30%; N, 1.22%; I, 8.37%. (Theoretically represents

.68 mmole trimethylpropylammonium iodide groups per gram.)



Silica Gel Bound Phosphine Silane (silica gel)/'\P(Ph)2

The synthetic method was patterned after a preparation

reported by Schrader and Studer.79 Silica gel (5.0 g)








was dried under vacuum at 3250C for 4 hours. In a two

neck 30 ml round bottom flask equipped with a reflux

condenser and a rubber septum in a nitrogen filled glove

bag, the silica gel and 100 ml of 3:1 benzene:p-dioxane

(each distilled from CaH2) were combined. The apparatus

was flushed with N2 for 1/2 hour while vigorously

stirring. The stirring rate was reduced, 0.45 ml of

2-(diphenylphospho)ethyltriethoxysilane was syringed in and

the mixture was refluxed for 12 hours. After cooling, the

solid was filtered in a glove bag, washed with benzene and

dried under vacuum for 24 hours at 250C.



Silica Gel Bound Phosphine Silane (High Loading)

The same procedure was used as described above for

(silica gel) P(Ph)2 except the amounts of reagents used

were 2 g silica gel, 100 ml 3:1 benzene:p-dioxane and 0.9 ml

2-(diphenylphospho)ethyltriethoxysilane.



Alumina Bound Phosphine Silane (alumina)x\ -NP(Ph)2

Alumina (10 g Acid Brockman Activity I mesh 80-200) was

washed with dilute HC1 and dried in an oven at 900C for

12 hours. In a 500 ml two neck round bottom flask equipped

with a reflux condenser and a rubber septum, the alumina and

200 ml of xylenes were combined. The mixture was degassed

by bubbling N2 through the suspension while stirring for

1.5 hours, followed by syringing in 1.5 ml

2-(diphenylphospho)ethyltriethoxysilane and refluxing for








12 hours under N2. After cooling the solid was filtered,

washed with xylenes, toluene, and absolute ethanol followed

by drying at 1000C for 24 hours.



[NEt4][Ru3H(CO)11]

A procedure was used similar to the synthesis reported

by Lewis et al.80 All manipulations were performed in an

inert atmosphere box, or with Schlenk-ware techniques. In

a flask, 0.3291 g Ru3(CO)12, 0.1035 g NaBi4 and 50 ml

tetrahydrofuran (degassed, distilled over CaH2) were

combined. After stirring for 30 minutes the solution

turned deep red. The mixture was filtered through a frit

and the THF was removed under vacuum. The reddish residue

was then combined with 20 ml of degassed methanol followed

by the addition of 0.132 g tetraethylammonium bromide

dissolved in 5 ml methanol. The volume was reduced to

10 ml followed by cooling overnight in a CO2/acetone

bath. The precipitated solid was filtered, washed with

a small amount of -78 C methanol and dried under vacuum

for 24 hours. Analysis: C, 29.38%; H, 3.53%; N,

1.73%. Theoretical: C, 30.75%; H, 2.85%; N, 1.90%.

(Problems were encountered during analysis due to 02

sensitivity.)







[Cs][Ru(CO)3)3]

A procedure was used similar to that reported by

Griffith and Cleare.81 In a 100 ml round bottom flask

equipped with a reflux condenser, 1.1 g anhydrous Rul3,

8.3 ml hydroiodic acid and 8.3 ml formic acid were

combined. The mixture was refluxed and stirred for 12

hours. After cooling, the mixture was filtered and 0.83 g

CsCl was added to the filtrate. The solvent was removed on

a rotary evaporator, the residue was extracted with absolute

ethanol and the solvent was then removed again. Analysis:

C, 4.2%; I, 54.8%. Theoretical: C, 6.2%; I, 54.5%. The

brownish- yellow residue was recrystallized in absolute

ethanol.



Silica Gel Functionalized [RuI3(CO)31] and [HRu3(CO)111

In a 50 ml Erlenmeyer flask 0.1136 g of the ammonium

iodide salt of silica gel bound propylaminosilane (8.37% I),

0.0222 g Cs[RuI3(CO)3] and 20 ml of tetrahydrofuran were

combined. The purple mixture was heated while stirring and

remained at reflux temperature for 10 minutes. After

cooling, the light maroon solid was filtered and washed with

THF. The resulting solid was brought into an inert

atmosphere box and combined with an intensely maroon colored

solution consisting of[NEt4][HRu3(CO)11] dissolved in

5 ml of THF. After the mixture was stirred for 10 minutes,








the light maroon colored solid was filtered, washed with THF

and dried under vacuum at 250C.



Ir(CO)2Cl(p-toluidine)

A procedure reported by Klabunde was used. In a

100 ml round bottom flask equipped with a reflux condenser

and a gas dispersion tube, 1.015 g IrCl 33H20, 0.290 g

anhydrous LiC1, 40ml 2-methoxyethanol and 4.5 ml H 20 were

combined. The apparatus was purged with N 2 through the

gas dispersion tube for 15 minutes. The N 2 was switched

to CO and a slow bubble rate was maintained while refluxing

and stirring for 10 hours. The mixture was cooled and while

purging with N2, 0.377 g of p-toluidine was quickly

added. The mixture was stirred for 15 minutes and added to

145 ml of H20. The purple precipitate was filtered in

air, washed with 90 ml H20 and dried under vacuum for 1

hour. The solution was filtered followed by removal of the

solvent under vacuum. Analysis: C, 30.12%; H, 2.53%; N,

3.90%; Theoretical: C, 27.92%; H, 2.24%; N, 3.35%.



(Silica Gel) '- P(Ph)2Ir4(CO)11

The method used was derived from a preparation reported

by Schrader and Studer.79 In an inert atmosphere box in

a 500 ml Parr pressure bottle set-up (described in the

experimental section of Part Three) a degassed solution of

76.5 ml 2-methoxyethanol and 2.7 ml H 20 was combined with








0.055 g Ir(CO)2Cl(p-toluidine), 18 g mossy zinc (acid

washed and chopped into small chunks) and the entire yield

of the resin resulting from the preparation of silica gel

ound phosphine silane ( 5 g). The system was pressurized to

40 psi ten times with argon followed by ten times with CO.

While stirring, the system was pressurized to 55 psi with CO

and heated to 900C. (Note: Good stirring was essential to

obtain an active catalyst. A 1/2 inch star stir bar,

purchased from Fisher Chemical Co., worked well.) After 24

hours the system was cooled and the slurry was decanted from

the mossy zinc in air, followed by filtering. The yellowish

support was washed with THF and dried under vacuum for 24

hours at 250C. (Heating while drying may result in

decomposition of the supported carbonyl cluster.) Analysis:

Ir, 0.23% (obtained by ICP).



(Silica Gel)--P(Ph)2Ir4(CO)11 (High Loading)

The same procedure was used as above for

(silica gel)\ -NP(Ph)2Ir4(CO)11 except the pressure bottle

set-up was loaded with 0.110 g Ir(CO)2Cl(p-toluidine),

153 ml 2-methoxyethanol, 5.4 ml H20, 36 g mossy zinc and

the entire yield from the preparation of high loading silica

gel bound phosphine silane. Analysis: Ir, 0.69%.







(Alumina)../N P(Ph)21r4(CO)11

The same procedure was used as above for

(silica gel)>Nz P(Ph)2Ir4(CO)11 except the reagents

loaded in the pressure bottle set-up were: 3 g alumina bound

phosphine silane, 0.056 g Ir(CO)2Cl(p-toluidine), 76.5 ml

2-methoxyethanol, 2.7 ml H20 and 18.0 g mossy zinc.

Analysis: Ir, 0.75%: P, 0.84%. (The theoretical Ir4 : P

ratio is 1:27.)



Ir4(CO)12 on Silica Gel (Physically adsorbed, 2.5% Ir)

The method used was derived from a preparation reported

by Howe.83 In a 100 ml round bottom flask, 2.0 g silica

gel (untreated), 0.072 g Ir4(CO)12 and 50 ml cyclohexane

were combined. After stirring for 3 hours the solvent was

removed under vacuum and the resulting solid was dried for 6

hours at 250C.



Ir4(CO)11P(Ph)3

The synthesis was derived from a preparation reported

by Shapley and Stuntz.84 In a 250 ml Parr pressure bottle

1 ml H20 was combined with 30 ml 2-methoxyethanol. Carbon

monoxide was bubbled through the solution for 15 minutes by

way of a syringe needle. Quickly, 0.155 g (0.397 mmole) of

Ir(CO)2Cl(p-toluidine), 0.026 g (0.99 mmole) P(Ph)3,

3.0 g mossy zinc (acid washed) and a stir bar were added. A

pressure head was placed on the bottle and the system was








purged with 60 psi CO ten times. Under 60 psi CO, the

mixture was heated while stirring to 900C. After 30

minutes the mixture was cooled, the pressure released and

the mixture filtered. The yellow filtrate was evaporated to

dryness under vacuum.

The residue was dissolved in 50 ml of 5:1

hexanes:methylene chloride solution and added in portions to

a 14" long, 1/2" diameter column packed with 230-400 mesh

silica gel. The yellow solution separated into two yellow

bands upon elution with 5:1 hexanes:methylene chloride.

Each band was collected and the solvent was removed under

vacuum. The first band off the column yielded a yellow

crystalline solid, Ir4(CO)11P(Ph)3, and the second

band yielded a yellow oil Ir4(CO)10(P(Ph)3)2'

Infrared analysis of both compounds matched well with

literature48 reported spectra.



(OEt)3Si(CH2)2P(Ph)2Ir4(CO)11 and
[(OEt)3Si(CH2)2P(Ph)212Ir4(CO)10 Mixture

The preparation used was reported by Schrader and

Studer.79 In an inert atmosphere box in a 500 ml Parr

pressure bottle, a degassed solution consisting of

100 ml 2-methoxyethanol and 3.2 ml H20, 0.20 g

IrCl(CO)2(p-toluidine) and 4.5 g mossy zinc (acid

washed) were combined. A pressure head was placed on the

bottle and it was purged with 40 psi ten times with argon

followed by ten times with CO. Under 55 psi of CO the








system was heated to 900C while stirring. After 12

hours the mixture was cooled, brought into an inert

atmosphere box and filtered. The solvent was removed from

the filtrate under vacuum. The IR spectrum of the

resulting yellow residue showed characteristic IR bands

for both iridium silane complexes. No separation was

attempted.



(Alumina)\,/\ P(Ph)2IrCl(CO)P(Ph)3

The procedure used was derived from a preparation

reported by Evans et al.85 In a 100 ml round bottom

flask, 0.11 g of Ir(CO)Cl(P(Ph)3 2, 1.0 g alumina bound

phosphine silane and 55 ml benzene (dried over CaH2 and

distilled) were combined. After the mixture was stirred at

ambient temperature for 72 hours under N2, the support was

filtered, washed with benzene and dried under vacuum.



IrCl(CO)(P(Ph)3)2 on Alumina (Physically adsorbed)

In a 100 ml round bottom flask, 0.11 g Ir(CO)Cl(P(Ph)3)2,

2.0 g alumina (acid washed) and 50 ml benzene were combined.

After stirring at ambient temperature for 120 hours under

N2, the solvent was removed by vacuum. The support was then

dried under vacuum for 12 hours.







Pyrolyzed Rh6(CO)16 on ZnO (0.6% Rh)

The preparation used was reported by Ichikawa. In

a 300 ml round bottom flask, 0.042 g Rh6(CO)16, 4.0 g

ZnO and 200 ml acetone were combined. The mixture was

stirred under N2 for 24 hours and then the solvent was

removed under vacuum. The resulting solid was placed in a

glass tube and heated to 1600C under vacuum for 2.5 hours.




Results and Discussion



Polymer Supported [Ru(CO)313]- and [HRu3(CO)111]

Synthesis and characterization

The support used to attach the ruthenium anions

[Ru(CO)3 13 and [HRu3(CO)11]- was an aminopropylsilane

functionalized silica gell that had been modified to a

trimethylpropylammonium iodide form.78 The synthetic

procedure for this support is shown in Figure 3-2.

Having a sufficient concentration of quaternary

ammonium iodide groups on the silica gel surface was an

important consideration. This would allow the ruthenium

anions to interact with one another once ion exchanged onto

the support. The ruthenium anion interactions are suspected

to be important because in Dombek's67 homogeneous system

both anions had to be present in solution for catalysis and

neither anion was a good CO reduction catalyst.


























Figure 3-2. Synthetic route for the preparation of
trimethylpropylammonium iodide functionalized
silica gel.














Step 1


OH + (OEt)3Si^sVtNH2


Toluene
-----&


O-SilNv"-NH2
t/


Step 2


O-SiY-/'-NH2 +


DMF
CH31 ---' O-Si-N -N(CH3)31








Based on an infrared study, Nyberg and Drago87

reported that the concentration of organosulfide groups

needed on a silica gel support to allow the formation of the

rhodium carbonyl dimer shown below was 0.20 mmole S per gram

of silica gel.

Elemental analysis of the quaternary ammonium iodide

modified support showed that the support contained 8.37%




S

(co)2Rh Rh(Co)2

S



iodine which translates theoretically to 0.63 mmole

quaternary ammonium iodide groups per gram of silica gel.

Thus the concentration of quaternary ammomium iodide sites

should be more than sufficient to allow site interaction.

The anions [Ru(CO)3 3]F and [HRu3(CO)11]"

were first synthesized separately as shown by Equations 10

and 11, and isolated as the cesium and tetraethylammonium

salts respectively.



CsCl
RuI3 + HI + HCOOH ---- ----Cs[Ru(CO)313] (10)
reflux


THF [NEt4]Br
Ru3(CO)12 + Na[BH4] ---- NEt4 [NEt4][HRu3(CO) 11] (11)
250C








A successful method for ion exchanging the two

ruthenium anions onto the support utilized two separate ion

exchange reactions in tetrahydrofuran. In the first ion

exchange reaction, the modified silica gel support was

stirred in a Cs[Ru(CO)31 3 THF solution which contained

a quantity of [Ru(CO)3 3]~ ions equal to one half the

number of theoretically available quaternary ammonium iodide

sites. The infrared spectrum of the isolated support after

stirring for 10 minutes at 25 0C showed no evidence of

[Ru(CO)3 3] being present (Figure 3-3b). Stirring

the support for 15 minutes under reflux conditions resulted

in an infrared spectrum (Figure 3-3c) containing two fairly

intense bands at 2091 and 2021 cm-1 (Table 3-4). These

correspond well with the CO stretching bands observed at

2100 and 2027 cm-1 for the KBr spectrum of Cs[Ru(CO)3 13

(Figure 3-4a). (The 2027 cm-1 value is the average

frequency of the two close bands at 2033 and 2021 cm-1).

The resulting [Ru(CO)3 3]- exchanged support was then

contacted with degassed, intensely maroon colored THF

solution containing the dissolved salt [NEt4][HRu3(CO)11 .

After 10 minutes of stirring at 250C the isolated support

contained three main infrared bands (Figure 3-4c) in the CO

stretching frequency; 2099, 2036 and 1987 cm -. These

infrared bands correspond well with what would result from

adding together the two spectra of the separate anions. The

infrared result fairly conclusively shows that the modified

silica gel support contains both the anions [Ru(CO)3 I3
























Figure 3-3. Infrared spectra in the range 2500-2600 cm-1
of a) (silica gel)/-/\ N(CH3)3I, (nujol mull);
b) [RuI3(CO)3]- exchanged at 250C, (nujol
mull); c) [RuI3(CO)3]- exchanged at reflux
temperature, (nujol mull).















a
























8--
b/







































2500


1800


1600


v I


2000


CM -1







80










E



0 CY





--4
444
t





0 00
Oc cc
V) <-) 0)
v_ li C oo



SU)


r0 0
U -
* -4 U) '-C -4

<3 C- C ) p


V E i 0 I-I r- U
NO r O O r,




N ) ON 0 0
-4 CD C) C) CD










00
oC E z 1 -




M 3 ^-t -- t nT, or










CZ












0- U CC

rr-4
Cl. +









E l I I






0 7 C)
S3 ., U














>V ) C) C/) Y 0)0
C=O '- I) 0 o
E- U S I CO -






03 o C -co






















Figure 3-4. Infrared spectra in the range 2500-1800 cm-1
of a) Cs[RuI3(CO)3j, (KBr);
b) [NEt4][HRu (CO)l1], (in CH3CN);
c) [Rul3(CO)3 and [HRu3(CO)-1
on (silica gel)/\/ N(CH3)3', nujol mull).





































































2500


2000

CM-'


a^


c








and [HRu3(CO)11] but the relative concentration of

each is unknown.



Catalysis

Approximately 0.5 grams of the ruthenium substituted support

was placed in a reactor tube in an inert atmosphere box, and the

tube was then placed in the flow reactor system described in the

experimental section. On passing a 3:1 H2:CO mixture through

the catalyst at atmospheric pressure and a flow rate of 10

cm /minute no products were observed at temperatures below

1500C. At 1500C a small amount of methane was detected along

with some CO2, which is often produced as a byproduct of

methanation or Fischer-Tropsch by way of the water gas shift

reaction. On raising the temperature to 175 0C there was little

or no increase in methane production.



Covalently Supported Ir4(CO)11

Synthesis

The method used to attach Ir4(CO)x (where x = 10

or 11) clusters onto the surface of silica gel or

alumina involved covalently bonding them to surface

bound phosphine silane groups. The synthetic procedure

used was derived from work published by Schrader and

Studer79 and is shown in Figure 3-5. The silane,

2-(diphenylphosphino)ethyltriethoxysilane was first bound to

the surface of the support by reaction of the ethoxy groups

on the silane with the hydroxyl groups on the support

























Figure 3-5. Synthetic route for the preparation of (silica
gel or alumina)\"P(Ph)2Ir4(CO)11.













Step 1


H + (OEt)3S- P(Ph)














Step 2





T + Ir(CO)Cl(P-toluidine)


Benzene
Dioxane
Reflux


















CO
Zn
H20-----
HpO


-O-Si sP(Ph) + 3 ETOH

/-


.P(Phlr4(CO)11


2-methoxyethanol








surface. The next step involved assembling the tetrairidium

clusters from the monomer Ir(CO)2Cl(p-toluidine) during

covalent attachment to the surface bound phosphine groups.

Through a fairly extensive FTIR study of the metal loaded

silica gel Studer and Schrader79 demonstrated that the

primary species functionalized on the surface was a mixture

of the following two clusters, --PIr4(CO)11 and

P2Ir4(CO) 10
It was found that complete mixing in the cluster

supporting step was an important factor in obtaining an

active catalyst. Attempts to double the synthesis procedure

(using 10 g of support) resulted in severe stirring problems

and thus an inactive catalyst.

The three primary catalyst formulations investigated in

this study are shown in Table 3-5 along with their elemental

analyses.



Infrared characterization

A fairly extensive Fourier transform infrared study was

conducted by Studer and Schrader79 on tetrairidium

clusters bound by 2-(diphenylphospho)ethyltriethoxysilane

functionalized silica gel. The characteristic carbonyl

stretching bands reported for both the mono-phosphine

substituted (silica gel)vN^P(Ph)2Ir4(CO)11 and the

diphosphine substituted (silica gel) (V\ P(Ph) 2)2Ir4(CO)10

are listed in Table 3-6. Also included is a summary of

















Table 3-5. Elemental Analysis of the Primary Catalyst
Formulations.


Catalyst Wt% Ir P H C



1. (Silica Gel)/^P(Ph)2Ir4(CO)11 .23 -


2. (Silica Gel)v/"P(Ph)2lr4(CO)11 .69 6.40 1.34


3. (Alumina)v/\P(Ph)2Ir4(CO)11 .75 .84 -
















VL)




-4 0): f
0-3 00

1)- 4 dl) ~ 0




- 00-4 00 0 0c 0 C
C:) C) CO -4 0o C)



4-4














04
x 03



0 U e O

S CCD








44- O
C-1 0 cc U S
U as -4 C)-O
3 C') a' 0U 01 (3, z









13 S o S o -4o
4 a4 -












o 0













-41
) u ) -3





















p Rt -4
C1 0 U) C o <

F-uI
H- I SI









I SO d --








spectra taken of the polymer bound tetrairidium clusters and

molecular analogs used in this study.

The majority of the spectra were taken on a Perkin

Elmer 283B grating spectrometer except for the spectra of

Ir4(CO)x (x=10 or 11) bound to phosphine functionalized

alumina. These spectra were done by Nicolet on a Nicolet

20DX FT infrared spectrometer, using the diffuse reflectance

technique.

Spectra obtained for silica gel bound tetrairidium

clusters at two different iridium concentrations, 0.23% and

0.69%, are shown in Figure 3-6. At the low iridium

concentration (0.23%) the carbonyl stretching frequencies

due to the tetrairidium cluster were almost

indistinguishable from the background. The infrared

spectrum of the 0.69% Ir support showed more pronounced

carbonyl stretching bands. However accurate measurement of

these bands was difficult due to the limitations of the IR

spectrometer. Therefore it was difficult to determine if

the tetrairidium cluster was mono or diphosphine

substituted.

The spectra taken on the analogous alumina system

(Ir = 0.75%) were more conclusive. Figure 3-7 (a,b) shows

the FTIR spectra of the phosphine silane supported alumina

and metal supported phosphine silane functionalized

alumina. The spectrum resulting from the subtraction of

spectrum (a) from (b) is shown in (c). The remaining three

bands, 2114, 2060 and 2004 cm-1 of the tetrairidium

























Figure 3-6. Infrared spectra (nujol mulls) in the range
2500-1600 cm-1 of (silica gel)v^P(Ph)21r4(CO)11
a) 0.23% Ir; b) 0.69% Ir.

























































2500


2000


i U


1800


1600


CM -1


-- --
























Figure 3-7. Fourier transform infrared spectra in the range
4000-400 cm-1 (diffuse reflectance) of
a) (Alumina)v P(Ph)2,
b) (Alumina)'/P(Ph)2Ir4(CO)11, c) b-a.