wv.—7v
_’-——

 

- --x' at "'Qfifliufifium .. .

Lu.

.7. .1.“
fl
0
I

khan...“

'lfl/iLLfl/IJHU’QM ' ' fifilrzefi’zdgiate l
THESIS I? University
'\

 

 

This is to certify that the

dissertation entitled

THE INTERCALATION AND CONVERSION OF NIOBIUM,
TANTALUM, AND MOLYBDENUM CLUSTER CATIONS TO INTER-
LAYER OXIDES AS A NOVEL ROUTE TO PILLARED
MONTMORILLONITES, AND MOLYBDENUM CLUSTER BINDING
TO PROTON EXCHANCSE’I)? ZEOLITE Y

presente

Steven Patrick Christiano

has been accepted towards fulfillment ’ '
of the requirements for ,

PhoDo degreein ChemiStry

, é cfid/

_ M or rofessor

 

 

.n...

Date am / 71/ 73/ m I 'J’f‘Pmm
a / ’

MSU is an Affirmative Action/Equal Opportunity Imu'mrinn 012771

_- ._ V_

 

 

 

 

 

MSU

 

 

RETURNING MATERIALS:
Place in book drop to

 

 

 

LJBRARJES remove this checkout from
—c—. your record. FINES will
be charged if book is
returned after the date
stamped below.
fr 3; .
I gag¢~3u32
. a ‘Efi’
JAN 2 419955

 

 

 

THE INTERCALATION AND CONVERSION OF NIOBIUM,
TANTALUM, AND MOLYBDENUM CLUSTER CATIONS TO INTERLAYER
OXIDES AS A NOVEL ROUTE TO PILLARED MONTMORILLONITES,

AND MOLYBDENUM CLUSTER BINDING TO
PROTON EXCHANGED ZEOLITE Y

By

Steven Patrick Christiano

An Abstract of a Dissertation

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

19814

\
. ,

‘
l‘.

€/'/J"'

ABSTRACT

THE INTERCALATION AND CONVERSION OF NIOBIUM,
TANTALUM, AND MOLYBDENUM CLUSTER CATIONS TO INTERLAYER
OXIDES AS A NOVEL ROUTE TO PILLARED MONTMORILLONITES,

AND MOLYBDENUM CLUSTER BINDING TO
PROTON EXCHANGED ZEOLITE Y

By

Steven Patrick Christiano

The research presented in this dissertation demonstrates
that a novel method for synthesizing metal oxide pillared
montmorillonites was developed. This synthesis utilizes
intercalated metal cluster cations as pillar precursors;
converting them to metal oxide preps in situ by calcina-
tion under vacuum.

Intercalation of the mixed oxidation state cations
Nb601ig’3+ and Ta6Clig’3+ into Na+-montmorillonite was per-

formed through aqueous ion exchange reactions. Ion exchange

isotherms and sodium release measurements were made. Inter-

2+,3+
12

(23 mmoles/lOO g silicate) were characterized

calates containing Nb6Cl

2+,3+
12

by chemical analysis and ESR, UV-visible and infrared spec-

(33 mmoles/lOO g silicate) and
Ta6Cl

troscopy. X-ray diffraction indicated d001 spacings of
18.410.2 A and 18.3:O.3 A for these intercalates, respec-
tively.

Calcination of the intercalates at 120-130°C (2“ h)

Steven Patrick Christiano

and 2A0° (24 h) decomposed the metal clusters to inter-
layer M205 (M=Nb and Ta) deposits. A reaction scheme show-
ing the hydrolysis and oxidation of the intercalated clusters
is prOposed. The niobium and tantalum oxide pillared clays
formed possessed surface areas of 63 and 70 m2/g. The inter-
layer spacings of m9.5 A were maintained up to “00° and 350°,
respectively.

A cationic molybdenum cluster, formed at pH 1.5 after
the argenometric precipitation of the terminal chlorides of
(M06018)C1u°2H20, was intercalated into Na+-montmorillonite
by ion exchange. A plot illustrating sodium release upon

cluster binding is given. The intercalate containing 53.6

)l.3+
2 3.3

characterized by chemical analysis and by UV-visible and

mmoles of Mo6Cl8(OH)2.7(0H per 100 g of silicate was
infrared spectroscopy. X-ray diffraction indicated a dOOl
spacing of 16.6 A.

Calcination of this intercalate at 130° (24 h), 200°
(24 h), and 280° (2“ h) decomposed the cluster. A re-
action scheme involving the hydrolysis and oxidation of the
intercalated cluster is proposed. A biphasic system was
formed containing collapsed smectite and M0020H inter-
layered clay, which possessed an interlamellar separation
of 3.“ A and was stable to 300°.

The pH dependent binding of M06C18(0H)n(0H2)E2::;+
to H+-Y was investigated. Binding was strongly dependent

of the exchange form of the zeolite. Approximately 3% Mo

Steven Patrick Christiano

was bound while a high degree of zeolite crystallinity was
maintained. A binding scheme which involves cluster pro-

tonation by zeolite H+ ions is proposed.

To Pat and Mary, Marsha and Lisa, my
wonderful family. To my good friend
Jack. Especially to Sue, who stood by
me through the many evenings of writing.
Without your love and support I could

not have completed this work.

Mere words cannot express the love that

I have for you.

11

ACKNOWLEDGMENTS

I would like to acknowledge the assistance of the many
people whose labor and ideas contributed to this Disserta-
tion. I especially want to thank Prof. Thomas J. Pinnavaia,
whose kind tutelage contributed greatly to my scientific
growth, and whose chemical knowledge and many insights
proved to be invaluable aids throughout the course of
my work. I would also like to thank Mr. Jailiang Wang for
the contributions he made to this Dissertation. I would
also like to thank the following people: Rasik Raythatha,
Christopher Marshall, Ming-Shin Tzou, Steven Landau, Ivy
Johnson, Emmanuel Giannelis, Kevin Martin, Abbas Kadkhodayan,
and Edward Keller. Your friendship and many midnight
philosophical discussions helped so much.

I would also like to acknowledge Dow Chemical Company
for the Summer Fellowship I received, the National Science
Foundation and Michigan State University Chemistry Depart-

ment for their financial support.

iii

Chapter

TABLE OF CONTENTS

LIST OF TABLES.

LIST OF FIGURES

LIST OF ABBREVIATIONS

I. INTRODUCTION. . . . . . . .
A. Foundations of Catalysis.
B. Smectite Clay Minerals; Structure
and Catalysis . . . . . . .
C. The Chemistry of Pillared Clays
D. Research Objectives . . . . . . . .
E. The Chemistry of Metal Clusters . .
II. EXPERIMENTAL . . . . . . . . . .
A. Materials
B. Synthesis
(Nb6Cll2)Cl2°8H20
2. (Ta60112)Cl2°8H20 . . . . . . .
3. (M06C18)Clu'2H20. . . . . . . .
A. Hydrolysis of (Mo6C18)Clu'2H20.
5. [MO6Cl8(OH)n(H2O)(6-n)](ClOA)(A-n)'
C. Metal Cluster Intercalation

Reactions . . . . . . . . .
1. Nb6012+ 3+ -Montmorillonite.
2. Ta6C12+’ ’3+-Montmorillonite.

iv

Page

vii

. viii

xiv

10
l3
l8
18
20
2O
21
21
22

23

2A
2A
25

Chapter

III.

3. Mo6C18(OH)2 7(H2O)1.3 3+-

Montmorillonite .

Ion Exchange Isotherms.

2+ ,3+

1. M6C112 /Na+-Montmorillonite

Ion Exchange Isotherm

1.3+

+
2. M06Cl8(OH)2 7(H2 O)3 /Na -

Montmorillonite Ion Exchange
Isotherm. . . . . . .

Metal Oxide Pillared Montmorillonites
l. (Nb205)-Montmorillonite
2. (Ta2O5 )-Montmorillonite . . . .

3. Molybdenum Oxide Interlayered
Montmorillonite . . . . . . .

Molybdenum Cluster Binding to H+ -Y
Zeolite . . . . . . . . . . . . .

Physical Methods. . . . . .

THE PILLARING OF MONTMORILLONITE BY
NIOBIUM AND TANTALUM OXIDE AGGREGATES
THROUGH METAL CLUSTER INTERCALATION
AND CONVERSION. . . . ... . . . . .

Metal Cluster Synthesis and Assign-
ment of Oxidation States. . . . . .

Intercalation of Niobium and
Tantalum Cluster Cations Into
Na+-montmorillonite . .

Physical Characterization of

M6Cl2+’ ’3+ -montmorillonites.

The Pillaring of Montmorillonite
by Niobium and Tantalum Oxide
Aggregates. . . . . . . . .

Page

25
26

26

27
28
28

29

29

3O
31

37

37

A5

57

69

Chapter Page

E. Conclusions and Recommendations . . . . . . 89

IV. THE INTERLAYERING OF MONTMORILLONITE
WITH MOLYBDENUM OXIDES BY THE INTER-
CALATION AND CONVERSION OF MOLYBDENUM
CLUSTER CATIONS. . . . . . . . . . . . . . . . 92
A. Intercalation of Molybdenum Cluster

Cations into Na+-Montmorillonites . . . . . 92

B. Physical Characterization of M06018-

(OH)2 7(H20)? +-montmorillonite. . . . . . 10A
C. The Interlayering of Montmorillonite

with Oxides of Molybdenum . . . . . . . . . 110
D. Conclusions and Recommendations . . . . . . 120

V. MOLYBDENUM CLUSTER BINDING TO H+ -Y
ZEOLITE . . . . . . . . . . . . . . . . 122

A. Introduction. . . . . . . . . . . . . . . . 122

B. Binding of the Molybdenum Cluster
by H+ -Y Zeolite . . . . . . . . . . . . . . 127

0. Physical Characterization of
H+
[Mo6C18(OH)3 (H 20)3]0. “7, “53.50'

Y Zeolite . . . . . . . . . . . . . 132
D. Conclusions and Recommendations . . . . . . 1AA
REFERENCES. . . . . . . . . . . . . . . . . . . . . 147

vi

LIST OF TABLES

Metal Oxide Pillared Clays
Absorbtion Band Maxima Published

for Niobium and Tantalum Clusters.
Some Physical Characteristics of
Thermally Treated Niobium and Tanta-
lum Interlayered Montmorillonites.
Surface Areas of Niobium and Tanta—
lum Oxide Pillared Clays .

X-ray Powder Diffraction Data and
Miller Indicies for Na+-Y, H+-Y,

and Molybdenum Cluster, H+-Y

vii

Page

11

MO

75

88

136

Figure

1-3

l-A

3-2

3-3

3-A

LIST OF FIGURES

The structure of the smectite clay
mineral montmorillonite.

A schematic diagram of a metal oxide
pillared clay.

The structure of the M6Cl?: metal
cluster core . . .

The structure of Mo6Clg+ metal
cluster core . . . . . . . . . .
UV-visible absorbtion spectra of (a)

(Nb60112)012'8H2O in aqueous solution

2+,3+

12 -montmorillonite.

and, (b) Nb6Cl

UV—visible spectra of (a) (Ta6C112)Cl2

°8H2O in aqueous solution, (b)

2+,3+
12

(Ta6Cll2)Cl2-8H2O in aqueous solution

Ta601 -montmorillonite, and (0)
following hydrogen peroxide oxidation
to the 3+ oxidation state.

The ion exchange isotherm of Nb6C1

12
into homoionic Na+-montmorillonite

2+,3+

The ion exchange isotherm of Ta6Cll2

into homoionic Na+-montmorillonite

viii

2+,3+

Page

1”

16

38

AA

“7

A9

Figure

3-5

3-6

3-7

3—8

3-9

3—10

Page

Milliequivalents of sodium released

from homoionic Na+-montmorillonite

2+,3+
12

bound. . . . . . . . . . . . . . . . . . . 52

relative to millimoles of Nb6Cl

Milliequivalents of sodium released

from homoionic Na+-montmorillonite

2+,3+
l2

bound. . . . . . . . . . . . . . . . . . . 53

relative to millimoles of Ta6Cl

The x-ray diffraction pattern of

2+,3+
12

33 mmoles of cluster/100 g of clay . . . . 59

Nb6C1 -montmorillonite containing

The x-ray diffraction pattern of

2+,3+
12

23 mmoles of cluster/100 g of clay . . . . 61

Ta6C1 -montmorillonite containing

Infrared spectra of (a) (Nb6C112)012°

8H2O, (b) Na+-montmorillonite, (c)

Nb6Clig’3+-montmorillonite, and (d)

2+,3+
12

at 130° (2A h) and 2A0° (2A h) in

Nb6Cl -montmorillonite treated

vacuo. . . . . . . . . . . . . . . . . . . 63
Infrared spectra of (a) (Ta60112)Cl2-

8H2O, (b) Na+-montmorillonite, (c)

2+,3+
l2
2+,3+
12

thermal treatment at 120° (2A h) and

Ta601 -montmorillonite, and (d)

Ta6C1 -montmorillonite after
2A0° (2“ h) in vacuo . . . . . . . . . . . 6A
ix

Figure

3-11

3-12

3-13

3-1A

3-15

3-16

The BBB spectrum of Nb6C1ig-mont-
morillonite. . . . . . . . .

The ESR Spectrum of Ta6Clig-mont-
morillonite formed by partial air
oxidation.

The differential scanning calorimo-

grams of (a) (Nb6C1l2)C12°8H20, (b)

Na+-montmorillonite, and (c)

2+,3+
12

33 mmoles of cluster/100 g of

Nb6C1 -montmorillonite containing
silicate . . . . . . . . . .

The differential scanning calo-
rimograms of (a) (Ta601l2)C12'

8H2O, (b) Na+—montmorillonite, and

2+,3+
12

taining 23 mmoles of cluster/100 g

(c) Ta6C1 -montmorillonite con-

of clay. . . . . . . . . . . . . .

UV-visible spectra of (a) Nb6C12+’3+-

12
montmorillonite, (b) Nb6CIiZ’3+-mont-

morillonite after thermal treatment
at 130° (2A h) and 2A0° (2“ h) in

vacuo, and (c) Na+-montmorillonite

UV-visible spectra of (a) Ta601fg’3+_

montmorillonite and (b) Ta6ClI2’3+‘

montmorillonite after treatment at 120°

Page

66

67

71

72

76

 

Figure

3-16
3-17

3-18

A-2

14-3

4-“

“-5

(24 h) and 240° (2A h) in vacuo.

Representative x-ray diffractograms

2+,3+_

of thermally treated Ta6C112

montmorillonites

Representative x-ray diffractograms

2+,3+_

of thermally treated Nb6Cll2

montmorillonites
The conductimetric titration of

AgBFu with (MO6C18)Clu°2H O.

2
The isothermal sodium release
upon cluster binding to Na+—mont-

morillonite at pH 1.5.

The x-ray diffractogram of montmoril-

lonite containing 53.6 mmoles of

M06018(OH)2.7(H20)%:§+ per 100 g of

silicate .

The infrared spectra of (a) [M06018-
(OH)n(H2O)(6-n)J(ClOA)(A—n) precipitate,
(b) M06018(OH)2 7(H2O);'§+-montmorillonite,
and (c) Mo6C18(OH)2.7(H2O)%°§+-montmoril-
lonite after thermolysis at 130° (2A h),

200° (2“ h), and 280° (2“ h) under dynamic

vacuum
Transmission UV-visible spectra of

1.3+

xi

Page

77

82

85

97

101

105

107

Figure Page

“-5 montmorillonite, (b) [M06C18(OH)n—

(H2O)(6—n)](ClO“)(“-n) precipitate,

and (c) (M06C18)C1u°2H2O . . . . . . . . . 109
“-6 The x-ray diffraction pattern of

M06018(OH)2.7(H2O)§:§+-montmoril-

lonite after thermolysis at 130° (2“ h),

200° (2“ h), and 280° (2“ h) . . . . . . . 113
“-7 The correlation between the melting

point of a metal oxide and the collapse

temperature of a pillared clay containing

that metal oxide . . . . . . . . . . . . . 119
5-1 A schematic diagram of the structure

of faujasite . . . . . . . . . . . . . . . 12“
5-2 X-ray powder diffraction patterns of

(a) H+-Y, (b) [MO6C18(OH)3(H2O)3]6.A7’
Hg3.5O-Y air dried, and (c) [Mo6C18-

(OH)3(H2O)31;.H7, H53.50-Y after calcina—

tion at “30° for 5 h under flowing

oxygen . . . . . . . . . . . . . . . . . . 13“
5-3 Infrared spectra of (a) H+-Y, (b)

[M06018(OH)3(H20)316.M7, HEB 50-Y’ and

(C) [MO6018(OH)3(H2O)3]0.“7’ $33.50-Y

after calcination at “30° for 5 h under

flowing oxygen . . . . . . . . . . . . . . 139

xii

Figure

5—“

Page
Electron microprobe analysis of
+ +
[MO6C18(OH)3(H2O)3]O.u7’ H53.50_Y
embedded in epoxy resin and cleaved
by an ultramicrotome . . . . . . . . . . . 1“3

xiii

..
'hnv
rib...

—

LIST OF ABBREVIATIONS

 

 

Abbreviation Description

bipy Bipyradine

DMF N,N-Dimethylformamide

DMSO Dimethylsulfoxide

DSC Differential Scanning
Calorimetry

ESCA Electron Spectroscopy for
Chemical Analysis

ESR Electron Spin Resonance

ICP Induction Coupled Plasma

phen g-Phenathroline

Y The aluminosilicate frame-

work of a faujasitic zeo-
lite; [(A102)53.97,
(8102)l38.03]53.97-

xiv

 

 

l

(I)

( f

I. INTRODUCTION

A. Foundations of Catalysis

 

Catalysis has grown to assume tremendous importance
in the large scale production of chemicals and petro-
chemicals. Recently, the area has received heightened
interest due to the ability of catalysts to increase the
rate and selectivity of a reaction, which allows for lower
operating temperatures and pressures. This holds con-
siderable economic impact in View of our limited energy and
feedstock resources.

In general, a catalyst is a substance which influences
the rate of a chemical reaction, by affecting the kinetic
barrier to reaction, but is not itself consumed in the pro-
cess. Catalysts are described as being either homogeneous
or heterogeneous based on their physical state of dispersion
during reaction. Homogeneous catalysts are evenly dispersed,
as gases or as compounds soluble in the reaction medium.
Heterogeneous catalysts are solids under reaction condi-
tions, with the chemical conversion of reactant molecules
occurring at the solid/gas or solid/liquid interfacial
surface of the catalyst. There are many important examples

of industrial processes utilizing each type of catalyst.

There are several homogeneous catalytic processes of
industrial significance. The Wacker processlI3, for ex-
ample, utilizes a soluble Pd/Cu catalytic system in the
production of acetaldehyde by the oxidation of ethylene.

The Oxo processu"6 makes use of a cobalt or rhodium carbonyl
catalyst in the hydroformylation of an a-olefin containing
n carbon atoms, to yield an aldehyde of (n+1) carbon atoms.

7

The Monsanto process utilizes a methyl iodide promoted
rhodium catalyst to produce acetic acid at 99% selectivity,
through the carbonylation of methanol. The Oxirane pro-
cess9 produces propylene oxide by using soluble molyb-
date compounds to catalyze the oxidation of propylene by
organic peroxides.

The majority of commercial catalytic processes utilize
heterogeneous catalysts, however. Some of these processes
are operated on a vast scale. For example, the oxidation
of sulfur dioxide to the trioxide is carried out over a
silica supported vanadium oxide catalyst. This process
allows for the production of greater than 80 billion pounds
of sulfuric acid per year8. The synthesis of ammonia
from hydrogen and nitrogen gases is carried out over a
reduced, promoted iron catalyst at “50°C and high pres—

surellc. Approximately 38 billion pounds of ammonia was

produced by this process in 1981 worldwide8.
The petroleum industry relies very heavily upon hetero-

geneous catalysis to carry out complex hydrocarbon

conversions. Catalytic crackinglo, hydrocrackingll,

hydrodesulfurization (HDS)12, and reforming13, are among
the more important catalytic steps involved in the up-
grading of crude oil to give high yields of lighter distil-
lates and high octane gasoline blends. Approximately 5
million barrels of crude oil can be processed in the United
States per dayloa.
There are many other catalytic processes used commer-
cially today, but the examples above serve to illustrate
the massive scale of operations. In addition, they il-
lustrate the fundamental catalytic functionalities of
reduction/oxidation, hydrogenation and acid/base catalysis.
The past 30 years has seen a great increase in catalytic
research, much of which is devoted to the development of
new catalytic processes. Fundamental to this effort is
the search for new materials which would serve as catalysts
or as catalyst supportslu. The development and testing of

new catalytic materials is essential to the future growth

of catalysis.

B. Smectite Clay Minearls; Structure and Catalysis

Recently, work has refocused attention on layered
silicates, especially the swelling philosilicates, known
as smectite clays, as catalysts or catalyst support
materials. The structure of the smectite clay mineral

montmorillonite is illustrated schematically in Figure 1-115.

 

 

Figure 1-1.

 

AMI-1,0)?”

The structure of the smectite clay mineral
montmorillonite.

The open circles represent oxide ions, the
shaded circles are hydroxide ions, and the
sinusoidally lined circles are water mole-
cules coordinated to the interlayer cations,
M2+. Substitutions within the octahedral
layer are indicated.

n
‘4

 

Smectite clays are composed of negatively charged silicate
layers which, in the natural form, are intercalated by
alkali or alkaline earth cations. The silicate layers are
structurally composed of four sheets of oxide or hydroxide
anions. These anions construct two layers possessing
tetrahedral coordination geometry with a layer of octa-
hedral geometry condensed between them through shared oxide
ions. Hydroxide ions complete the structure of the octa-
hedral layer. These hydroxides are not shared between the
layers, but reside at the bottom of hexagonal rings formed
by tetrahedral units on the silicate interlayer surfaces.
This clay structure is described as being a 2:1 layered
structure based on the tetrahedral to octahedral layer
ratio.

In montmorillonite, the tetrahedral sites are occupied
mainly by silicon (IV) ions and the octahedral holes are
two-thirds filled by aluminum, magnesium, or ferric ions.
This clay is described as being dioctahedral due to the
two—thirds fractional occupancy of the octahedral sites.
Isomorphous substitutions within the layers, such as

Mg2+ for Al3+ within the octahedral layer, or A13+ for

Sil4+ within the tetrahedral layer, will impart a net nega-
tive charge on the lattice. This charge is balanced by
the presence of interlayer cations.

Solvent molecules coordinated to the interlayer cations

and adsorbed on the silicate interlayer surfaces will swell

the structure in a direction perpendicular to the plane of
the layers. This solvent swelling will allow for the total
exfoliation of Na+-montmorillonite in aqueous solution.

The interlayer cations may be substituted with a wide
variety of other cations utilizing ion exchange tech—
niquesl6.

Historically, clays have found large scale use as
heterogeneous catalysts in petroleum processing. Acid
treated montmorillonites, and later halloysites and kaoli-
nites, were used as petroleum cracking catalysts beginning
in the 1930'517. Reduced nickel and iron supported on
hydrofluoric acid treated montmorillonites were used as

18

hydrocracking catalysts in England and Germany during

this period. Acid treated clays have also been utilized

as catalysts for polymerizationlg’20

tionZl.

and aromatic alkyla-

22

More recently, Pinnavaia and coworkers have utilized

the unique properties of solvent swelled layered silicates
containing transition metal compounds as catalysts. These
materials were shown to catalyze a variety of reactions

such as the hydrogenation of alkenes23 2“,25,

26

, alkynes

and dienes , asymmetric hydrogenation327, and the hydro-

formylation of a-olefin528. In these reactions surface
chemical equilibria and spacial restrictions within the
clay interlayer played a major role in controlling reaction

selectivity. Recently, hectorite intersalates containing

v.
0
~

we r-
‘.

H

‘1

¢_

2»

. Q

Q»

 

 

4 a ;. : . .. .. C.
A.» .r.. 2. 0 . r4. QC . .e cc
PM a a 7.“. AU 2.“ AC. r». .54

 

metal chelate complexes have been shown to possess activity
as triphase catalysts29.

Thomas30

and coworkers have reported a number of proton
assisted reactions of organic molecules with transition
metal exchange forms of smectites. The conversion of
alkenes to dialkyl ethers3l, dehydration of primary al-
cohols to produce dialkyl ethers32, the conversion of

thiols to dialkyl sulfides33, and reactions of primary
amines to form secondary amines3u have been reported. Other
workers have also reported similar catalytic results uti-
lizing transition metal exchanged smectites35 and synthetic

swelling claysl8.

C. The Chemistry of Pillared Clays

 

Pillared clays are layered silicates which have been
modified through the intercalation of robust cations.
The large interlayer cations act as molecular props, which
maintain separation of the silicate layers in the absence
of swelling solvent. This creates a material possessing
an extensive two-dimensional interlayer pore structure,
as shown schematically in Figure 1-2. These materials
possess large intracrystalline surface areas suitable for
molecular adsorption and catalysis.
Early pillaring eXperiments by Barrer and coworkers36

dealt with the intercalation cfi‘ alkylammonium ions into

montmorillonite. These bulky cations prOpped the silicate

Si/icafe Sheef

[

Inter/dyer Pore

/

 

   

m h
/ l

 

 

 

 

 

 

Metal Oxide Pillar

Figure 1-2. A schematic diagram of a metal oxide pillared
clay.

This edge-on View shows the presence of highly
dispersed intercalated metal oxide which props
the silicate sheets apart thus creating per-
manent interlayer pores. These pores may
exist as an extensive 2-dimensional network.

PU

..v

v

Ce.

Ii.
an i.

v“

a...
Av

 

p._
7‘

 

layers apart creating permanent interlayer separations
of “ to “.5 A and markedly improved the diffusive and
sorbtive prOperties of the host silicate. The absence
of appreciable hysteresis in the desorbtion of adsorbed
02, N2, Ar, and paraffins, indicated the ease of access

to the interlayer in these compounds37. Layer work by
Shabtai38, and independently by Mortland39, utilized the
intercalation of sterically ridged cations, such as l,“-
diazabicyclo[2,2,2]octane (DABCO), or l,“-diaminoadamantane
to achieve interlayer spacings of approximately 5 A. The
intercalation of metal chelate cations, M(Chel)§+ (where

M equals Fe, Ni; Chel equals phen or bipy) by Loeppert gt

glib

, produced pillared montmorillonites possessing
interlayer spacings of about 8 A. These materials are limited
in their thermal stabilities (ISO-350°) due to the presence
of organic components, however.

The most thermally stable, and thus, most promising
pillared clay systems for catalytic applications, are those
utilizing clusters or chains of inorganic oxides as mol-
ecular props. As shown by Endo, gt al.ul, clays inter-
layered with silica can be synthesized through the hydrolysis
of intercalated tris(acetylacetonato)silicon (IV) cations.
This pillared clay possesses excellent thermal stability
(>500°), but the interlayer spacing was low (m3.0 A), cor-
responding to only one layer of siloxane polymer.

A more common preparative method for pillared clays

utilizes the ion exchange of polynuclear metal hydroxy

An.»

eye

in.’

A
”.4
icon—

5

«Iv

 

 

 
 

.. . . . _. . e . 7 II

n . Y o h . ,\ r. t s « .
L. . A t r: n . .G p. v... .C C; .. . .r .. . r.
P . A i n o e v n L. at .ru wk 05 t .v \U ru.

10

cations formed in aqueous solutions into sodium smectites.
Calcination of the intercalates thus formed, yields the
metal oxide pillared clays. Table 1-1 lists some of the
inorganic pillaring species utilized to date, and some of
the physical characteristics and thermal stabilities of
the metal oxide pillared clays formed. The alumina pil-
lared clay, for example, possesses an interlayer spacing
of approximately 7-10 A and stability to greater than 550°.
This material possesses a surface area of between 500 and
200 m2/g depending on the loading of the pillaring ion.

“5,“6 been utilized to con-

Variable loading has reportedly
trol the intrapillar distances in these materials. This
variability allows for the controlled formation of pores
possessing laterial dimensions averaging from 8-30 fi“6,
which can be designed to suit a particular application.
Pillared clays, expecially the alumina and zirconia
systems, have been tested for catalytic activity. For
example, the alumina and zirconia pillared clays have
been shown to possess the ability to crack middle weight
petroleum fractionsuu. Alumina pillared clays are also
active toward the dealkylation of cumene and B-isonaph-

thaleneus.

D. Research Obiectives

Although the exchange reaction of smectite clays with

polynuclear metal hydroxy cations is a promising route to

,.
.-

 

 

(lv

I

 

ll

.wcfipmoa pmHHHQ so mocovcmaoc npfiz manmfipm> mH mono commasm

n

.osfiu manp pm czocxcz ohm m cam x modam> onem

 

 

 

em on ooomx ea +mmeafimoveamu
N z x
m in- com ma . mo me. Amov Hz
: e+As xmv L
as omm oomA e.mH xgmoamv
. m a x
m: 0mm ooomx m ea e+Aauxmvfio me. Amov hog
3:.m: oozioom oom ma +mfl0mms.A:HAmovstL
2.3
ms.:= ooomuoom ommx AH +smmHAmmovsmAmov30mHHeL
m\ 5 oo m nofiooam meaooaafid
.neom m .sofiaaooom .wdaooom
.mmm< mo .QEoB Hmmmm
oommpsm Umppoqom .xwz

 

 

.nsoao ooaoaaad

mUHxO Hmpmz .HIH mfinme

vs.

GB

h“,

N

M.

‘¢-_

ac
Ala

tiv

 

12

pillared clay synthesis, the process is quite dependent

on the aqueous chemistry of the metal ion used to form

the props. The aqueous chemistry of many transition metal
ions is not amenable to this procedure due to the lack of
formation of a sufficiently large cationic species.

An alternate route to pillared clay formation which
circumvents this limitation is the use of metal cluster
cations as pillar precursors. This unique approach would
involve the electrostatic binding of metal cluster cations
to the clay interlayer and their subsequent modification
in situ to yield metal oxide pillars. This method could
potentially provide better defined interlayer oxide props
and, more importantly, would serve to move the pillaring
process beyond the confines of aqueous metal ion chemistry.
This would expand the number of transition metal oxide pil-
lared clays available for study.

The objective of this study is to utilize metal cluster
cations as precursors to metal oxide props in the forma-
tion of pillared clays. This major objective can be
viewed as being comprised of four intermediate goals.

(1) Develop an understanding of the novel metal cluster
cation-smectite clay interaction, and to utilize this in
the formation of metal cluster intercalate precursory
materials. (2) Design and optimize a reaction scheme
which will carry out the necessary in situ reaction of

the metal cluster to yield an intercalated metal oxide.

l3

(3) Characterize the physical and chemical properties

of the pillared clays formed. (“) Investigate this ap-
proach in a general way, utilizing several metal cluster
cations, in order to assess the general viability of this

synthetic strategem.

E. The Chemistry of Metal Clusters

 

Molecules containing polyhedral groupings of three or
more metal atoms possessing intermetallic bonding are de-
fined51 as metal cluster compounds. The metal cluster com-
pounds, (Nb6C18)C12°8H20, (Ta6C112)C12°8H20, and (M06-
018)Clu°2H2O, were chosen for this study because they con-
tain cationic metal cluster cores. The structure of
M6C1?;, where M is Nb or Ta, is shown in Figure 1-352.

The dark circles represent the metal atoms, the open cir-
cles are the edge bridging halide atoms. X-ray crystal-

lographic data53

show an average metal-metal distance

of approximately 2.9 A for both the niobium and tantalum
clusters. The metal-bridging chloride distance of 2.“3 A
is approximately the same for both clusters, also. The
metal cluster core, M6C122 can adept oxidation states of
n=2,3,“, following reversible one electron oxidation
steps. The niobium cluster core is less easily oxidized
than the tantalum species, which can be oxidized to n=“

by oxygen. The niobium core requires more rigorous oxidiz-

ing conditions, such as contact with chlorine gas, to

Figure 1-3.

1“

  
 

 

 

 

is

“a

   
    

Q Q

The structure of the M6012; metal cluster
core.

M equals Nb or Ta, and n is 2, 3, or “. The
dark circles represent the metal atoms, the
open circles are the edge-bridging chlorides.
Each metal atom possesses an open terminal
binding site.

»3

1v

5.

nbi

fiv

 

 

15

adOpt the “+ oxidation stateSu’SB. The n=3 cation
of M6Cl$g is paramagnetic, exhibiting a magnetic sus-
ceptability close to the spin only value expected for one
electron56. This unpaired electron resides in a non—
degenerate molecular orbital, being equally distributed
on the six metal atoms as shown by electron spin resonance
(esr) spectroscopy of the niobium cluster57.
Each metal atom possesses an Open site through which it
can be coordinated by a ligand in a terminal bonding posi-
tion. A variety of ligands have been coordinated to the
clusters yielding positive, neutral and anionic com-
pounds53’58. The instability of the metal clusters of

niobium and tantalum toward hydrolysis to yield M20 is

5
94.58.59

well established and makes these compounds excel-

lent candidates for the in situ formation of metal oxide

props for pillared clay formation.

The structure of the Mo6C18uu cluster core is shown

in Figure l-“52

. This structure, originally established
by Brosset60, contains an octahedron of molybdenum atoms
(dark circles) with intermetallic bonding, and an average
Mo-Mo distance of 2.61 A. The eight face bridging

chlorides are represented by the open circles. The

average Mo-Clbr distance is 2.“? A61. Each molybdenum
is capable of bonding a ligand terminally. A variety
62-66

of ligands , including alkoxides67 can be bound.

This cluster can exhibit various oxidation states, however,

9"

 

16

 

 

 

 

 

 

Figure l-“. The structure of Mo6Clg+ metal cluster core.

The dark circles represent the metal atoms,
the open circles are the face—bridging
chloride ions. Each metal atom possesses
an open terminal binding site.

17

these are formed under rather pressing conditions. The
oxidation of (Mo6Cl8)Clu with chlorine gas can lead to the
formation of (Mo6Cll2)Cl368. Reduction of the cluster can
occur in the presence of excess phosphine to yield
Mo6C18Cl3(P(Et)3)366.

Detailed studies by Sheldon63’69-72 have shown that
substitution of the bridging chlorides by hydroxide ions
occurs by an 8N2 mechanism, and can lead to the hydrolytic
decomposition of the cluster structure. Potentially this
type of cluster decomposition can be utilized within the

interlayer region of montmorillonite to yield a molybdenum

oxide pillared clay.

C.
r».

r x.
Q»

II. EXPERIMENTAL

A. Materials

 

Niobium pentachloride was obtained from Alfa Chemical
Co., and sublimed at 150°C in vagug prior to use. Cadmium
metal was used in the form of freshly prepared filings.
Niobium metal (Alfa Chemical Co.) was used in the form
of small shot. Tantalum pentachloride (Alfa Chemical Co.)
was sublimed at 190° in X2222 prior to use. Aluminum metal
was in the form of fine turnings. Molybdenum powder (100
mesh) was obtained from Climax Molybdenum Co. Molybdenum
pentachloride was used as received from Alfa Chemical Co.

Natural Wyoming montmorillonite was received in a
spray dried form from the Source Clay Minerals Repository,
University of Missouri. The fraction containing particles
less than 2p was collected through sedimentation from
aqueous suspension for 2“ h. This procedure also served
to remove carbonates and other impurities. The homoionic
sodium exchanged form of the clay was prepared through
treatment with 1.0 M NaCl solution. Excess sodium chloride
was removed by successive washings with distilled water
and centrifugation. The clay was dialysed against dis—

tilled water until free of chloride, and then lyopholized.

18

.L

n. ..

.nu

on

2..

A.»

 

 

 

V

l9

Elemental analysis of the Na+—montmorillonite was per-
formed using Inductively Coupled Plasma (ICP) emission

analysis (vide infra). The unit cell composition was

 

determined to be: N ](Si

30.50[A13.11’ Mgo.50’ Fe0.140 7.91’

A10.09)020(OH)A'

Analysis of a homoionic Ni2+ exchanged form of the
mineral indicated the cation exchange capacity (CEC) of a
typical freeze dried montmorillonite sample to be 60.2 meq/
100 g. The nickel content was measured gravimetrically by
forming inns dimethylglyoxime complex following digestion
of the clay in boiling HCl/HNO3.

Synthetic zeolite Y was obtained from Amoco Petroleum
Research Corporation in a highly crystalline form. The unit
cell formula was determined to be Na53.97 [(A102)53.97
(Si02)138.03] by ICP analysis. The ammonium exchanged
form of this zeolite was produced by ion exchanging twice
with 1.0 M ammonium sulfate solutions for six hours at
70°73. The NHZ—Y was freed from excess salt by repeated
washings with distilled water and centrifugation. The
hydrogen exchange form of the zeolite was produced by
heating the NHZ-Y under flowing N2 at “00° for 2“ h,
followed by “50° for 2“ h or until wet litmus paper failed

to detect NH3 in the effluent stream.

l‘11

LA) .

(D

’r’

4‘

AA

or,

 

; 1.-

20

B. Synthesis

 

1° £fl969l1229l2L§fl29

The niobium cluster salt was prepared following two
different methods. The first method used was that des-
cribed by Harned7u, which utilized the reduction of NbCl5
with cadmium metal at 700°. The metal cluster salt thus
produced was isolated from lo—uN hydrochloric acid solution.
This sample of cluster salt was found by quantitative ESR to
contain an approximately 50/50 mixture of cluster molecules
in the 3+ and 2+ oxidation states. Anal. Calcd. for
(Nb6C1l2)Cl2°8H20: Nb, “6.535 Cl, “1.“3. Found: Nb, “6.71;
Cl, “1.16. UV-visible; Amax = 905, “05, 32“, and 275 nm.

The second synthetic method used was that described by
Koknat gt £1.75. It utilized the conprOportionation of

NbCl with niobium metal in the presence of NaCl at 850°.

5
Dark green crystalline (Nb60112)Cl2'2H2O was quickly iso-

lated from 10"11

N hydrochloric acid solution to limit air
oxidation. In both cases, the electronic solution spectra
matched the published spectral data for this cluster
species. The product synthesized by the second method
possessed absorbtion peaks as assigned76 to the cluster
ion in the 2+ oxidation state; Amax = 905, “00, 32“ and
275 nm. Quantitative ESR measurements on this salt in-
dicated that approximately 10% of the cluster molecules

were in the 3+ oxidation state.

21

2' £E§69$1219l2;2529

The tantalum cluster salt was prepared in good yield

by following the combined procedures of Kuhn and McCarley77
and of Koknat gt g;.75. A mixture of aluminum turnings

and TaCl the latter in slight excess, was placed in a

5,
pyrex tube which was then evacuated and sealed. The re-
duction of the tantalum pentachloride by the aluminum was
carried out in a “00°/200° temperature gradient for 36 h.
Dark green crystals of (Ta6C112)Cl2°8H2O were collected
from lo-uN hydrochloric acid solution under argon atmos-
phere. The electronic solution spectrum matches that des-

cribed in the literature55’77 for the cluster ion in the 2+

oxidation state; Amax = 750, 635, “72, 395, 327 and 282 nm.

3' £M9691819th§§29

The molybdenum cluster chloride was preparedtnzfollowing

the method of Nanelli and Block78. Molybdenum metal powder

and MoCl were placed in a quartz glass tube with a length

5
of 125 cm. The pentachloride was sublimed at 650° in flow-

ing N over the molybdenum metal which was distributed

2
along the length of the tube. The 650° zone was generated
by a tube furnace that was 30 cm in length. This zone was
moved the length of the quartz tube in the direction of

nitrogen flow. The direction of sublimation was reversed

four times. Bright yellow microcrystalline molybdenum

22

cluster was collected by extracting the crude product with
“ 500 m1 portions of “N hydrochloric acid solution. The
intracrystalline HCl was driven off by slowing heating the
salt to 200° under vacuum. Aggt. Calcd. for (Mo6Cl8)Clu°
2H2O: Mo, 55.50; Cl, “1.02. Fgggg: Mo, 55.39; Cl, “1.17.
-l 80

UV-visible: Amax at 320 nm79. IR: 3“0(br) cm

“. Hydrolysis of (M059l8)C1u'2H29

 

An aqueous solution of Mo6C18(OH)n(OH2)E2::;+ was pre—

pared from (M06018)Clu°2H O by the use of a modification

62

2

of a method of Cotton and Curtis A sample of (M06018)Clu°

2H 0 (1.00 g, 9.6“ x lO-Ll moles) was dissolved by boiling

2

in 9““ ml of absolute methanol. The cooled solution was

added drOpwise to a solution containing 0.800 g (3.86 x

10'3 equivalents) of AgClOu dissolved in a mixture of 7“0 m1

of reagent grade methanol and 55 m1 of distilled water. The

presence of approximately 7% water is a modification of the

literature procedure and is necessary to avoid the precipi-

tation of the molybdenum cluster. This precipitate has

been reported to be Ag [Mo 01 ]°nC H OH62’79 The com-
2616 25 -

bined solution was allowed to stir 30 min, after which, the

AgCl precipitate was removed by centrifugation to yield a

clear yellow solution.

The stock solution was stabilized against precipitation

by dilution with 5L of distilled water and the subsequent

23

removal of the methanol through evaporation utilizing a
rotovap apparatus. The pH of the solution was adjusted to
10.0 by addition of either sodium or potassium hydroxide.
The final solution was 1.7 x 10'“ M in molybdenum cluster.

The solution was stored at “° to retard degredative hy-

drolysis.

5. [MO6C_1_8(OH)n_(_QIi22(6_n)_J__(_ngul(u_n)

Portions of the molybdenum cluster stock solution
described above were adjusted to pH 1.0 with 6 M HClOu.
This reaction yielded precipitates which were quite vari-
able in their color (ranging from bright yellow to brown)
and quality. Variability in the quality of perchlorate
precipitates of the metal cluster coordinated by organic
ligands has been reported by Cotton and Curtis62. It
must be noted that these workers found molybdenum cluster
perchlorates to be quite explosive in the presence of
organic ligands. The precipitates obtained from per—
chlorate solution were physically characterized. UV-vis;
Amax at 315 and “00 nm. The infrared spectrum of this pre-
cipitate contained absorbtion bands attributed to per-
chlorate ion (1075 cm'l) and metal cluster coordinated by
water molecules and hydroxide ions (30“0, l“10, 750, 510,

and 3“0 cm-l).

2“

C. Metal Cluster Intercalation Reactions

 

2+,3+

l2 -Montmorillonite

1- 32591

 

In a typical example, 36.9 mg (3.08 x 10.5 moles) of
(Nb60112)C12°8H20 was dissolved with heating and stirring
in 100.0 m1 of distilled water. After being cooled and
filtered, the dark green solution had a pH of “.5. To this
rapidly stirring solution, a suspension of 100.0 mg of
Na+-montmorillonite in 20.0 ml of water was added. After a
reaction time of 6 h at room temperature, the resulting
clay intercalate complex was washed with distilled water
and centrifuged repeatedly to free the silicate from any
excess cluster salt. The dark green product was then lyopho-

lized. A cluster loading of 33 mmole/100 g of silicate was

2+ 3+ +
l2)0.2“ ’ Na0.06[Al3.ll’

Nb, 13.8“, C1,

achieved. Anal. Calcd. for (Nb6Cl

Mgo.50’ F°0.qu(Si7.9l’ A10.09)020(OH)A‘

10.56; Na, 0.13. Found: Nb, 13.8“; Cl, 10.65; Na, 0.13.

 

UV-visible; Amax at 895, “05 and 280 nm.

The determination of an ion exchange isotherm (Eggg
12232) made it possible to obtain predictable loadings
of the niobium cluster on the montmorillonite. The de-
sired cluster equilibrium solution concentration and

loading on the clay was achieved by controlling the ratio

of moles of cluster to montmorillonite equivalents.

25

2+,3+

12 -Montmorillonite

 

2° $2691

The exchange reactions of the tantalum cluster were
carried out in essentially the same manner as those of
its niobium analog. The tantalum cluster is susceptable
to air oxidation under the exchange reaction conditions,
therefore, all manipulations were carried out under argon
atmosphere. For example, “1.6 mg (2.“1 x 10.5 moles) of
(Ta6C112)012°8H20 was dissolved in 100.0 ml of degassed
distilled water. The resulting blue-green solution pos-
sessed a pH of “.5. To this solution was added 50.0 mg of
Na+-montmorillonite suspended in 3“.0 ml of degassed dis—
tilled water. The resulting clay intercalate complex was
successively washed with degassed distilled water and
centrifuged to free the silicate from any excess
cluster salt. The intercalated clay was then lyopholized.
This dried material was quite oxygen sensitive as indicated
by a rapid color change to red upon exposure to air. A

loading of 23 mmole/lOO g of silicate was thus achieved.

2+,3+
12)0.17 ’ N°0.0u[Al3.11’ Mgo.50’

Feoou0](Si7.91, AlO 09)O20(OH)“: Ta, 18.23; 01, 7.1“;
Na, 0.10. Found: Ta, 18.23; Cl, 7.92; Na, 0.095. UV-

Anal. Calcd. for (Ta6C1

visible; Amax at 775, 652 and 355 nm.

3. Mg6gl8(OH)2.7tflzgléZ§+-Montmorillonite

 

Molybdenum cluster intercalation in montmorillonite was

found to be dependent on the cation formation by the

(1"

('f

t D

(D

 

 

26

cluster at low pH. For example, (150.0 mg, 9.03 x 10-5
exchange eq) of Na+-montmorillonite was suspended in 200.0
ml of distilled water. To this suspension was added 1.05

x 10'“

moles of [Mo6C18(OH)n(H20)(6_n)](u—n)+ from stock
solution. Distilled water was added to adjust the cluster
solution concentration to 5 x 10.5 Mhor lower. The pH of
this solution was adjusted to 1.5 by the dropwise addition
of 0.1 M HClOu. This resulted in the formation of a
bright yellow flocculated clay. Ultrasonic agitation (50
watts for 3 minutes) was used to assist the exchange reac—
tion. The bright yellow intercalate was separated from
the colorless supernatant by centrifugation. The inter—
calate was then repeatedly washed with distilled water and
centrifuged until the wash water was pH 7.0. The product

was collected and air dried on a glass slide. This pro-

cedure yielded a cluster loading of 53.6 mmole/lOO g of

1.3+
2)3.330.39'

[A13.11, Mg0.50’ Fe0.uo] (Si7.91, AlO 09)020(OH)u= Mo,
20.35; Cl, 10.03; Na, 0.00. Found: Mo, 20.35; Cl, 7.57;

clay. Anal. Calcd. for [Mo6C18(OH)2.7(OH

Na, 0.0065. UV-visible; A at 330 and “05 nm.

max

D. Ion Exchangg Isotherms

 

2+,3+

12 /Na+-Montmorillonite Ion Exchange Isotherms

1- M621

 

The room temperature ion exchange reaction of M6CII2’3+

and homionic Na+-montmorillonite was measured in strictly

27

analogous fashion for both niobium and tantalum clusters.
Montmorillonite was allowed to equilibrate 18 h with solu-
tions containing known initial concentrations of (M60112)012°
8H20. These solutions contained initial quantities of
cluster salt varying up to six moles per cation exchange
equivalent of the clay. The metal cluster concentration

in each centrifuged sample solution was determined spectro-
photometrically using a Cary 17D SpectrOphotometer. The
absorbance peak measured at “05 and 395 nm for the niobium
and tantalum clusters, respectively, was compared to a
standard solution absorbtion curve. The amount of the metal
cluster bound to the silicate was determined by the dif-
ference between the (M6C112)C12'8H20 present initially,

and the amount remaining in solution at equilibrium.

The quantity of sodium released by Na+-montmorillonite

2+,3+
12

by atomic emission using a Jarrell-Ash spectrometer.

into solution upon reaction with M601 was measured

Measurements were then taken at Amax = 5889.9 A. The
quantity of sodium released from the montmorillonite was
obtained from the total amount measured in each sample
solution by correcting for sodium present in the cluster
stock solution and in a montmorillonite blank.

1.3+ + ,
2. M26918(OH)2.7(OH213.3 /Na -Montmorillon1te Ion

 

Exchange Isotherm

 

The amount of sodium released upon molybdenum cluster

binding at room temperature and pH 1.5 was determined on

28

individual samples containing up to 1.5 moles of cluster
per clay cation exchange equivalent. Samples were stirred
and equilibrated 12 h in polypropylene vials, then centri-
fuged to remove the montmorillonite. Sodium concentra-
tions were measured by atomic emission as described above.
Corrections were applied to compensate for sodium content
in the cluster stock solution, the perchloric acid and the
clay blank. Error bars were calculated for each point
using the standard deviations determined from the five
atomic emission readings for each sample and the estimates

of errors in volume readings.

E. Metal Oxide Pillared Montmorillonites

 

 

1. (Nb2O:)-Montmorillonite
——J

Niobium oxide pillared montmorillonite was prepared
from a clay intercalate containing 33 mmole Nb6Clig’3+/
100 g of silicate. Heating the compound under dynamic
vacuum at 130° for 2“ h, limitedtflmaamount of interlayer
water present, but the cluster integrity was maintained
as evidenced by the Nb:Cl ratio of 1:1.96. Mggl. Calcd.
f°r (Nb60112)6fifi+’ ”30.06‘EA13.11’ “€0.50, Feo.u0]‘
(817.91’ A10.09)O20(OH)u: Nb, 8.72; Cl, 6.65. E2229:
Nb, 8.72; CI, 6.51.

Heating the intercalate lg vacuo at 2“0° for 2“ h

following the 130° treatment, promoted the hydrolytic

29

decomposition and oxidation of the cluster lg situ. This

procedure produced a niobium oxide pillared clay. Anal.
+ +

Feo.uo](Si7.91, A10.09)020(OH)M: Nb, 12.17; Cl, 0.00.

Found: Nb, 12.17; Cl, 0.91.

 

2. (Ta2OE)-Montmorillonite
——J

Montmorillonite pillared by tantalum oxide was prepared

2+,3+
12

The clusterremainedirmact following calcination under

from an intercalate with 23 mmole Ta6Cl /100 g of clay.

dynamic vacuum at 120° for 2“ h, as evidenced by the

)2+,3+
12 0.16 ’

+ O
Nao.0“‘[Al3.1l’ Mg0.50’ Feo.“0](517.91’ A10.09)O20(OH)A°
Ta, 17.90; Cl, 7.01. Found: Ta, 17.90; 01, 7.66.

2.18:1 C1 to Ta ratio. Anal. Calcd. for (Ta6Cl

Heating the intercalate lg vacuo for 2“ h at 2“0° fol-
lowing the l30° treatment promoted the formation of inter-

layer tantalum oxide. gggl. Calcd. for (T8205)0 “8’

+ +
H0.u6’ Nao.ou‘[A13.1l’ “30.50, F°0.u03(517.91’ A10.09)O20‘
(0H),: Ta, 18.35; Cl, 0.00. Found: Ta, 18.35; Cl, 0.A6.

3. Molybdenum Oxide Interlayered Montmorillonite

 

The molybdenum oxide interlayered montmorillonite was

prepared from an intercalate containing 53.6 mmoles of

)l.3+
2 3.3

heating under dynamic vacuum of 130° for 2“ h followed by

Mo6Cl8(OH)2 7(OH /100 g of silicate. The preliminary

30

2“ h at 200°, was designed to limit the amount of inter-
layer water present. Calcination at 280° for 2“ h yielded
the molybdenum oxide interlayered montmorillonite. This
material contained no detectable chloride as determined

by the absence of an AgCl precipitate upon digesting the
product in H280“ and adding alcoholic AgNO3.

F. Molybdenum Cluster Binding to H+-Y Zeolite

 

Molybdenum cluster binding to synthetic fausasitic
zeolite Y is dependent on both the pH of the solution and
the cation exchange form of the zeolite. H+-Y zeolite
(1.00 g, “.70 meq) was suspended in 1.27 L of distilled
water with rapid stirring. To this suspension was added
340.3 ml (5.82 x 10‘5 moles) of M06Cl8(OH)n(OH2)(6_n)-
(010“)(A-n) stock solution. The pH of the solution was
observed to drop from 7.2 to 5.6 during 30 minutes of rapid
stirring of this mixture. The suspension was then heated
to 70° and 1.61 L of l x 10'” M HC104 solution was added
over the course of l h to provide a final pH of “.0. This
mixture was stirred an additional 3 h. The light yellow
solid was separated from the clear supernatant by centrifuga-
tion, and then washed repeatedly with distilled water. gggl.
Calcd. for [Mo6C18(OH)3(OH2)3]O.u7, Hg3.50-[(A102)53097-

(SiO Mo, 2.27; Cl, 1.11. Found: Mo, 2.27; Cl,

2)138.03]‘
0.51.

This material was heated to 200° for 2 h, under flowing

31

N2 as a preliminary drying step. This was followed by a
calcination at “30°C for 5 h. The latter treatment
resultedixlthe apparent decomposition of the cluster and
a series of color changes from yellow to blue (m3“0°) to
white (m“00°). This white color is consistent with the
presence of Mo(VI) oxide. gggl. Calcd. for (M003)3.06’
(SiO

-[(A10 Mo, 2.uu; Cl, 0.00.

+
H53.97 2)53.97 2)138.o33'
Found: Mo, 2.““; Cl, 0.013.

G. ngsical Methods

 

1. X-ranyiffraction

 

X-ray diffraction data were collected utilizing either
a Phillips X-ray diffractometer or a Siemans Crystallo-
flex-“. Both instruments utilized Ni-filtered CuKa radia-
tion (A = 1.5“05 A). Clay samples were generally in the
form of thin films deposited on a glass slide by evapora-
tion of an aqueous suspension at room temperature.
Goiniometer measurements were collected from 2° to 30°
as values of 28. The peak positions were converted to d

spacings using the Bragg equation.

2. Electron Sgin Resonance

Electron spin resonance spectra were obtained at room

temperature using a Varian E-“ spectrometer. Samples

32

were prepared as freeze dried powders or microcrystalline

81

solids. A numerical double integration method was used
to integrate the ESR spectrum for quantitative measure-
ments. The intensity of the sample peak was calibrated
against a CuSOu'5H2O reference using the method of Aasa

and Vanguard82.

3. UV-visible Spectroscopy

 

Electronic spectra were recorded on a Cary 17D or a
Cary 219 spectrophotometer. Solution spectra were
measured versus solvent in matched 1 cm path length
cells. Clay samples were prepared as mulls in Nujol
and held between quartz plates, or dried on quartz slides

and Spread with Nujol.

“. Infrared Spectroscopy

Infrared spectra were measured in the range “000-250
cm-1 utilizing a Perkin-Elmer “57 spectrophotometer.
Samples were prepared as potassium bromide pellets or
were mulled in Nujol and suspended on cesium iodide or

potassium bromide plates.

5. ESCA

Electron spectroscopy for chemical analysis (ESCA)

measurements were furnished by Dr. Luis Matienzo of the

33

Martin-Marietta Corp. These measurements were made using
a Physical Electronics Model 5“8 XPS with the aluminum
Ka12 line as the exciting source and the carbon is line
(28“.6 eV) as a reference. The position of the Nb3d 5/2
line observed from niobium intercalated montmorillonites
was compared to those of Nb205, Nb metal and KNbO3.

6. Surface Area Measurements

 

Nitrogen B.E.T. surface area measurements were per-
formed using a Perkin-Elmer Model 212B sorbtometer with a
thermal conductivity detector. Samples were activated at
elevated temperatures under flowing argon or under vacuum.
Nitrogen was adsorbed onto the sample at liquid nitrogen
temperatures from He/N2 flow mixtures of known composition.
The volume of nitrogen adsorbed was determined by com-
parison of the area of the desorbtion peak to the area
of a nitrogen peak obtained from calibration tubes of
known volume. Surface areas were determined by a three

point B.E.T. method.

7. Thermal Analysis

 

Thermal analysis was performed using a DuPont Model
990 thermal analyser operating in either the differential
scanning calorimetry (DSC) or differential thermal analysis
(DTA) mode. These analyses were carried out at a heating

rate of 10°C/min. A1203 was used as a reference material.

3“

8. Conductimetric Titration

 

Electrolytic solution conductivities were measured

using a Beckman Model RC 1682 conductivity bridge with a

cell containing the 1 cm2 bright platinum electrodes held

approximately 1 cm apart. Solution conductivity was

1

measured in arbitrary units of ohm- multiplied by the

total solution volume in milliliters to compensate for

volume changes during titration. The titration was car-

ried out with the addition of a 6.0“ x 10‘” M methanolic

6

solution of (M06018)C1u'2H O to a “.86 x 10' M AgBFu

2

solution in methanol/water as described above for the
(“-n)+

MO6Cl8(OH)n(H2O)(6-n) stock solution synthesis (Section

3-“). An equilibration time of twenty minutes was allowed

prior to each conductivity measurement. The data points

were fit to lines using a least squares method.

9. Elemental Analyses

 

Inductively coupled plasma (ICP) emission analyses
of Si, Al, Fe, Mg and Na was performed using a Jarrell-
Ash Model 955 Autocomp spectrophotometer at the Michigan
State University Agricultural Experimental Station.
Samples were fused in lithium metaborate at 1000° and
the melt was dissolved in 0.1 M nitric acid. The analyses
of five independent samples was averaged. A National

Bureau of Standards clay sample was used as a calibration

35

standard. Analyses for Nb, Ta, Mo, and 01 were performed by

Galbraith Laboratories, Inc., Knoxville, TN.

10. Electron Micropgpbe Analygis

 

Electron Microprobe analysis was performed using an
Applied Research Laboratories ARL-SM microprobe in con-
junction with a VG HB 501 scanning transmission micro—
scope (STEM). Measurements were made at an X-ray takeoff
angle of 52.5°. The zeolite sample was exchanged with
molybdenum cluster in the usual manner and air dried. The
sample was then embedded in epoxy resin and ultra-thin
(700A) sections were cut by Dr. K. K. Baker of the Michigan
State University Pesticide Research Department. The sam-

8

ple was placed within the microscope at 10' torr for 12

h in order to limit the amount of intracrystalline water

83 to

present in the zeolite. This procedure is reported
help maintain the crystallinity of zeolite samples sub-
jected to electron bombardment during microscopy. Molyb-
denum content was measured using the MoLa emission line,
and chlorine content was determined using the ClKa line.
Quantitative molybdenum and chlorine analyses were made
on two regions of a zeolite particle. The first region
was apparently an interior portion of the zeolite which
had been exposed by cleavage of the particle during sample

preparation. The second region was an exterior portion

of the particle. The relative strengths of the molybdenum

36

and chlorine signals recorded can be directly compared
because the instrument settings and data acquisition
times were identical. The micropore analyses were per-

formed by Mr. V. Shull at the Michigan State University

Physics Department.

III. THE PILLARING OF MONTMORILLONITE BY NIOBIUM
AND TANTALUM OXIDE AGGREGATES THROUGH METAL

CLUSTER INTERCALATION AND CONVERSION

A. Metal Cluster Synthesis and Assignment of Oxidation

States

The niobium Cluster salt was synthesized following two
different literature procedures. The first procedure was

7“

described by Harned , and utilized the reduction of nio-

bium pentachloride with cadmium metal at 700°C. This re-

action can be written75:

6NbC1 + 8Cd

5 7600 Cd2(Nb6Cll2)Cl6 + 6CdC1 . (3-1)

2
The crude cluster product was dissolved in water and the
cadmium dications were precipitated as the sulfide by re-

action with H S. The metal cluster salt, (Nb6C112)C12'

2
8H20, was isolated as dark green crystals from hydrochloric
acid solution. The UV-visible spectrum of this salt in
aqueous solution contains absorbtion bands at Amax equal to
905, “05, 32“, and 275 nm, as shown in Figure 3-la. The
position of these absorbtion maxima, especially the exact

location of the peak around “00 nm, is indicative of the

oxidation state of the cluster.

37

38

 

4
—-----—O
—-—.——-
cur—v

Absorbance—

 

 

 

 

 

 

 

av sd> ‘Mo ab 6&7 to ad; g» up
Mummhmm
Figure 3-1. UV-visible absorbtion spectra of (a) (Nb60112)-
C12'8H2O in aqueous solution and (b) Nb6Clig’3+-

montmorillonite.

Spectrum (a) is of the mixed 2+,3+ oxidation

state cluster salt. Spectrum (b) is of the inter-
calate containing 33 mmoles of cluster/100 g of
clay prepared as a Nujol mull suspended between
quartz glass plates measured versus a mulled
Na+-montmorillonite reference. Absorbance was
measured in arbitrary units.

39

Absorbtion maxima of the electronic spectra published
for the niobium and tantalum salts in the non—anated form
and in the 2+, 3+, and “+ oxidation states are compiled
in Table 3-1. There is considerable disagreement within
the literature as to the exact locations of the absorbtion
bands for the metal clusters in the lower oxidation states.
The positions of the absorbtion maxima included in this
table are taken from relatively recent and reliable
sources.

The results of spectroscopic investigations indicate
several important aspects of the electronic spectra of the
niobium and tantalum clusters. The absorbtion bands at
wavelengths longer than gg. 333 nm show only slight shifts
upon replacement of the bridging chlorides with other
halides. Therefore, these lower energy bands are assigned
to electronic transitions between molecular orbitals de-
rived mainly from metal atom d-orbitals within the cluster
core. The position of these bands is strongly dependent
on the oxidation state of the metal cluster. This is
especially true of the bands at wavelengths longer than
600 nm in the absorbtion spectrum of the tantalum cluster.

The higher energy bands (A < 333 nm) are shifted toward
the red upon changing the bridging halides from chlorine

5“,86,87

to bromine These absorbtion bands are therefore

interpreted as being ligand to metal charge-transfer bands.
Substitution of the terminal ligands results in only slight

shifts in the absorbtion spectra of these Cluster386.

 

“0

 

.+mog Co noaofioseso ofiaoosofihofioon no sofiofiooo g6 sonneaxo

.Uopom Hum oz

9

.Hocmnpo CH vouwcmicoc on on coESmm<m

 

 

 

mm mam.mos.osn 0mm osaogmfifiomeev
mm mmm.mo:.uun.oag.omm omm pmaoflmafioeoev
gg mmm.wmm.ogs.gme.mmg 0mm cmmm.maofimaaoeoev
mm omm.wmm.uuu.osm.oeg 0mm ommm.maonmaaoeeev
gm Nam.---.mms.---.figm.mma mommmo eefiogmanoeczvthsgmmmovL
gm mmm.--:.eme.unu.omm.mem mommmo emfiogmfiaoeoZVMHZJAmmmovL
mm mgm.HHm.oos.mge.mwm.ggw omm Ommm.maogmaaoebzv
on mnm.oam.gmm.mes.wmm.aom on cmmw.maofimaaoenzv
am sgm.omm.emm.ows.moe.aom omm cmmw.maogmaaoeozv
.mom EC nwaK pco>aom UCSOQEOO

 

 

.nsoonsao ssfioosoe one esfiooaz tog socnfiaosa usages seem soapstonoa .Hum oases

“1

The spectrum of the (Nb6C112)C12'8H20 synthesized by
following Harned's method is shown in Figure 3-la, and
possesses a sharp absorbtion band at “05 nm. Comparison
of this peak position to the published data indicates
that it lies between the analogous peak for the cluster

in the 2+ oxidation state which is located at 1m equal

ax
to 396 nm, and that for the cluster in the 3+ oxidation
state located at “26 nm. This suggests that the niobium
cluster salt is present as a mixture of oxidation states,
with the average lying between 2+ and 3+.

Quantitative electron spin resonance (ESR) measure-
ments were performed on the solid niobium Cluster salt to
ascertain the average oxidation state. The intensity of
the resonance absorbtion peak was determined by using a
numerical double integration method81. The integrated
sample peak was compared to the resonance absorbtion peak
of a crystalline CuSOu’5H2O reference using the method of

Aasa and Vangard82

to determine the number of unpaired
electrons present in the cluster salt.

The metal cluster core in the 3+ oxidation state
contains one unpaired electron in a non-degenerate orbital.
The metal cluster salts in the 2+ and “+ oxidation states

56,88 and are not ESR active57. It was

are diamagnetic
determined that approximately 50% of the niobium cluster
molecules were in the 3+ oxidation state in this sample

of (Nb60112)012'8H2O. The remainder of the sample was

 

 

 

. o 1] . . o i . - .. , a] w, _ m
S 8 a T r. .. . E C .. i I. .3 C ..J p c 3.: n. LC 2.: . n... c. I.
S h LL U .1 1T“ QC 1-. + e D» a e L.“ ht h. r. A... .rl. tr» » en ti ha a v
a S d a P . u C u E. QC 8 n. 0 nu. .« 1 0:. hi .d O. .Q e... nt.

 

 

“2

assumed to be in the 2+ oxidation state as indicated by
the UV-visible spectral data. This mixture of oxidation
states may have been the result of incomplete reduction
during the synthesis of the salt, or may be attributed to
air oxidation of the cluster during the isolation step.
This cluster salt containing mixed oxidation states was

used for the majority of the studies for niobium, and will

2+,3+
l2 '

The second procedure used to synthesize the niobium

be referred to as Nb6CI

cluster was described by Koknat gt gl.75. This method
utilized the conproportionation reaction between niobium
pentachloride and niobium metal at 850° in the presence of
sodium chloride. The stoichiometry of this reaction is

shown in Equation (3-2):

1“NbC15 + 16Nb + 20NaCl + SNa“(Nb6Cll2)Cl6 (3-2)
Crystalline (Nb6C112)C12'8H2O was isolated from the crude
product containing the fully anated cluster by recrystalliz-
ing from hydrochloric acid solution. This step was per-
formed quickly to limit air oxidation of the cluster.

The UV-visible solution spectrum of this salt contains

absorbtion bands at Ama equal to 905, “00, 32“, and 275

x
nm. Comparison of these peak positions to Table 3-1 in-
dicates that the majority of the cluster molecules are in
the 2+ oxidation state. Quantitative ESR measurements

confirm that less than 10% of the cluster was present

 

AV
LL.

 

“3

to the 3+ state.

The tantalum cluster salt, (Ta60112)012'8H20, was pre-
pared by following the method of Kuhn and McCarley77,
which utilized the reduction of tantalum pentachloride

with aluminum metal. The reaction stoichiometry is des-

cribed by Equation (3-3).

18TaC15 + 16A1 ”0022000 3(Ta6C112)C12 + 16A1C13 (3-3)
This reaction was performed in a “DO/200° temperature
gradient to promote the convective circulation of the vola-
tile tantalum pentachloride. Dark green crystals of
(Ta6C112)012'8H20 were isolated from hydrochloric acid
solution. The UV-visible spectrum of this salt in a
freshly prepared aqueous solution is illustrated in Figure

3-2a. This spectrum contains absorbtion bands at Ama equal

x
to 750, 635, “72, 395 and 327 nm. The appearance of the
lower energy bands, especially the band at 635 nm, is im-
portant in the assignment of the oxidation state. Comparison
of these absorbtion maxima to those published by Espenson
and McCarley, shown in Table 3-1, indicates that the cluster
salt is in the 2+ oxidation state.

The electronic absorbtion spectrum of the tantalum
cluster salt after reaction with hydrogen peroxide in

aqueous solution (pH “.5) at room temperature is illustrated

in Figure 3-2c. The presence of absorbtion maxima at 900,

““

 

Absorbance —+

 

 

 

 

 

.Kb 469 ab 6&9 in ab 9a? ma)
Wavelength (nm)
Figure 3-2. UV-visible spectra of (a) (Ta6C112)C12°8H20

in aqueous solution, (b) Ta6Clig’3+-montmoril-

lonite, and (c) (Ta60112)C12°8H20 in aqueous
solution after hydrogen peroxide oxidation to
the 3+ oxidation state.

Spectrum (a) is of the cluster in the 2+ oxi-
dation state. Spectrum (b) is of the inter-
calate containing 23 mmoles of cluster/100 g
of clay prepared as a Nujol mull suspended
between quartz glass plates measured versus

a mulled Na+-montmorillonite reference sample.
Absorbance was measured in arbitrary units.

“5

715, “05 and 330 nm indicates that the cluster is in the

3+ oxidation state.

B. Intercalation of Niobium and Tantalum Cluster Cations

 

Into Na+-Montmorillonite

 

The host silicate used for the metal cluster inter-
calation studies was a homoionic sodium exchange form of
a naturally occurring montmorillonite from Wyoming. This
smectite clay possesses the unit cell formula NaO 50—

](Si AlO 09)O20(OH)“’ as de-

[A13.11’ Mgo.50’ Fe0.110 1.91’
termined by inductively coupled plasma (ICP) emission
analysis.

The niobium and tantalum cluster salts dissolve in
aqueous solution to yield Chloride anions and solvated
metal cluster cations. The M6C12; cations will react with
aqueous Na+—montmorillonite suspensions to form unique
metal cluster-layered silicate intercalation compounds.
The metal cluster cations bind to the smectite electro-
statically, displacing the interlayer sodium ions through

an ion exchange reaction. The ion exchange reaction is

written schematically in Equation (3-“).

 

+ n+ n+ +
nNa + M6C1l2 + M6Cll2 + nNa (3-“)

 

 

3' C . 0C I c
an” 2m r u .r u .. C .C a C .y. . C.» «Wu 3.; Y.“ 5b O a «r.
.wm «MU; h. C. n. 6 Y.“ I. r“ PO .r. . mu. 6 u a
C Wile ll b e A.V OVIQ I N0 2. N &

‘FD

 

 

 

“6

wherein the horizontal lines represent the silicate layers
of montmorillonite.

Preliminary experiments suggested that the ion ex-
change reactions of the metal cluster cations with Na+-
montmorillonite exhibited a somewhat anomolous behavior
in comparison to the reactions of simple metal ions. This
behavior necessitated a more detailed study of the ion
exchange characteristics of these metal cluster cations.
One method of characterizing the metal cluster—smectite
interaction is through the measurement of an ion exchange
isotherm. Figure 3-3 illustrates the isotherm measured
2+’3+ into Na+-montmorillonite.

12
The isotherm is plotted as millimoles of Nb601ig’3+ bound

for the ion exchange of Nb6C1

per 100 g of silicate relative to the metal cluster solu-
tion concentration at equilibrium with the clay. The

initial portion of the curve shows a rather strong up-

2+,3+
12

favorably with sodium for the ion exchange sites. At

take of Nb6Cl , indicating that this cation competes

higher equilibrium solution concentrations, the cluster

binding attains a maximum value of approximately 5012

2+,3+
l2

ing is considerably higher than the level expected, which

mmoles of Nb6Cl per 100 g of silicate. This load-
is indicated by the dashed line in the figure. The maxi-
mum binding anticipated was 2“ mmole/100 g of clay, based
on the average charge per cluster of 2.5+, and the cation-
exchange capacity (CEC) of the montmorillonite; 60 meq/

100 g. The cluster binding clearly exceeds this maximum

2%3’
Milimole [NbGCllz] Bound/IOOg Montmorillonite
45 0' m x]
O O C? 9

01
O

N
O

46

 

Q
t

Figure 3-3.

“7

 

l I I I l I 1 L l

l 2 3 4 5

O

20
no“); Molority of [Nbeciz] in Equilibrium with Montmorillonite

2+,3+

12 into

The ion exchange isotherm of Nb601
homoionic Na+-montmorillonite.

The dashed line indicates the level of clus-
ter loading expected based on the average
metal cluster oxidation state and the cation
exchange capacity of the montmorillonite.
The amount of cluster actually bound was
measured spectrophotometrically.

“8

level anticipated.

The ion exchange isotherm measured for the tantalum
cluster is qualitatively the same as that for the niobium
analog. As evident in Figure 3-“, this isotherm features
a very steep initial uptake of the cluster cation which
indicates that initially the tantalum cluster is bound
to the montmorillonite almost quantitatively. Apparently,
the tantalum cluster interacts more favorably with the
montmorillonite than the niobium cluster. The isotherm
curve rapidly reaches a plateau at 50 mmoles of tantalum
cluster bound per 100 g of silicate, above which no addi-
tional cluster uptake is observed. This loading is con-
siderably higher than the level expected. The dashed line
in Figure 3-“ illustrates the maximum loading anticipated
for the Cluster with an average charge of 2.7+. This value
is determined by measuring the amount of sodium released

upon cluster binding (vide infra).

 

Apparently, oxidation of the cluster salt in the solid
state upon only brief exposures to air, and in solution
during the exchange reaction despite attempts to avoid
air exposure has led to the formation of a mixed oxida-

tion state tantalum cluster. This mixed oxidation state

2+,3+
12 '

spectrum of a tantalum cluster intercalate protected from

cluster will be referred to as Ta6Cl The UV-visible

oxygen exposure after lyophilization and mulled in Nujol,

is illustrated in Figure 3-2b. This spectrum contains

m
0
I

8

Bound/IOOg Montmorillonite

0.36
A
O

2
)

U
0

 

Milimoles (Ta 50,2
— n:
C) o

0

Figure 3—“.

 

“9

 

2+,3+

IO“): Moloriiy of {Name} in Equilibrium with Montmorillonite:

2+,3+

12 into

The ion exchange isotherm of Ta6Cl
homoionic Na+-montmorillonite.

The dashed line indicates the level of Cluster
loading expected based on the average metal
cluster charge of 2.7+ as determined by measur—
ing sodium release during exchange, and the
cation exchange capacity of the montmorillonite.
The amount of metal cluster bound was determined
spectrophotometrically.

50

components of the solution spectra of the cluster salt in
the 2+ and 3+ oxidation states. This indicates that the
intercalated cluster possesses an average oxidation state
somewhere between 2+ and 3+, and is consistent with the
average oxidation state of 2.7+ assumed above.

In contrast to the reaction of the tantalum cluster,
quantitative ESR measurements utilizing (Nb6C112)012'
8H20 which contains approximately 10% of the molecules in
the 3+ oxidation state, indicate that no significant oxida-
tion of the niobium cluster occurs during ion exchange re-
actions in solution exposed to the air. The estimate of
the number of unpaired electrons present in a lyophilized
intercalate which had not been exposed to oxygen after
drying was made. This value was used in conjunction with
the number of cluster molecules present in the sample as
determined by elemental analysis to show that no significant
increase in the number of cluster molecules in the 3+ oxi-
dation state had occurred. This result is supported by
the UV-visible spectrum in Figure 3-lb. The spectrum of
Nb6C1ig’3+-montmorillonite is directly comparable to the
solution spectrum of that salt containing mixed oxidation
states. The positions of the absorbtion maxima are un-
changed following the intercalation reaction, therefore, no
oxidation of the cluster cation has occurred.

To better elucidate the phenomena involved in these
ion exchange reactions, the amount of sodium liberated

from Na+-montmorillonite was measured as a function of

51

metal cluster loading. Figure 3-5 illustrates the plot of
milliequivalents of Na+ released per 100 g of montmoril-
lonite versus millimoles of Nb6Cl§:,3+ bound per 100 g of
silicate. The initial portion of this plot is linear,
indicative of a simple ion exchange reaction. The slope
of this line, 2.5 Na+ released/cluster bound, is con-
sistent with the average cluster charge of 2.5+.

This linear relationship is retained until approxi-
mately two-thirds of the interlayer sodium is displaced.
The decreasing scope beyond this point suggests a reduced
charge per cluster. The curve eventually levels out upon
release of one cation exchange equivalent of sodium at
high metal cluster loadings. Interestingly, the niobium
cluster continues to load onto the clay in the absence of
exchangeable interlayer sodium.

The analogous plot showing sodium release upon binding
of Ta6Clig’3+ is shown in Figure 3-6. The initial section
of the plot is assumed to be linear, in direct analogy to
the niobium results. This line was drawn using a least

squares fit to the first four points, with quadruple

weighting on the zero point. This approach was necessary

2+,3+
l2

montmorillonite appears to show considerable error. The

2+,3+
l2

bound. This value for the average oxidation state of the

because the point at “ mmole Ta6Cl bound/100 g of

slope of this line is 2.7 Na+ released per Ta6Cl

tantalum cluster is consistent with the UV-visible data as

discussed above.

52

825

01
O

8 8

Miliequivolents No” Released/IOOg Montmorillonite
— A
O O

 

 

I l l I l l L

IO 2%,? 30 40 50
Miimoies of [MSCIQ] somd/iooq Montmorillonite ‘

9L
(

.
-
.
.

Figure 3-5. Milliequivalents of sodium released from

homoionic Na+-montmorillonite relative to

millimoles of Nb601ig’3+ bound.

The initial linear relationship is indicative
of an ion exchange reaction, with the slope
of this line indicating the liberation of 2.5
Na+ ions per cluster molecule bound.

A 8 m s] m
0 O O O
I I I r I

8‘

Miliequivalents NoiReleosed/IOOg Montmorillonite

20

 

Figure 3—6.

53

   
 

SLOPE- 2.7

l A l l l l l

l I
20 2+5,30 40 so

IO
Milimoles of [TOGCIRJ Bound/l009 Montmorillonite

Milliequivalents of sodium released from

homoionic Na+-montmorillonite relative to

millimoles of T3601i5’3+ bound.

The initial portion of this plot was assumed

to be linear, and was fit to the first four
points by a least squares method, with multiple
weighting to the zero point. The slope of

this line indicates the release of 2.7 Na+

ions per cluster bound.

5“

The tantalum cluster continues to bind to the mont-
morillonite beyond the point where all of the interlayer
sodium has been replaced. Under conditions where the clay
has been fully saturated with the metal cluster cations,
approximately 50 mmoles of M6ClI2’3+ are bound per 100 g
of silicate. This loading is nearly the maximum value
calculated for the binding of a close-packed monolayer of
hydrated metal cluster cations within the clay interlayer.
The diameter of a hydrated cluster molecule is estimated to
be 12.6 A from the anhydrous cluster diameter of 9.8 A
calculated from x-ray crystallographic data53’55’89’90.

At this maximum loading the average cluster charge is
calculated to be 1.3+, assuming the interlayer cluster
molecules simply balance the charge on the silicate layers.

Two reactions are proposed which could effectively reduce
the cluster charge to this level: hydrolysis, Equation
(3-5), and anation with the chloride counter ions in

solution, Equation (3-6).

[M6Cll2°6Hé0]n+ : [M6Cllz(0ii)~5ii201(°'1)+ (3—5)
+ H+
Cl“
[M60112-6H20]n+ : [M6Cll2(Cl)°5H20](n-l)+ (3—6)
+ H20

Hydrolysis occurs when a water molecule which is ter-

minally coordinated to the cluster, releases a proton to

55

solution, to produce a bound hydroxide ion. Hydrolysis is
evident upon dissolution of (M6C112)012'8H2O in water, where-
upon the pH of the solution falls. The anation reaction
scheme, Equation (3-6), proposes that a chloride ion from
solution displaces a coordinated water molecule and binds
to the cluster in a terminal position.

The positions of the equilibria as defined above,
appear to be a function of the cluster loading within
montmorillonite. As cluster loading increases above the
level of the simple ion exchange regime, these equilibria
shift to the right, allowing for the binding of additional
cluster ions. Presumably, the binding of larger numbers
of lower charged cluster cations is favored by the mont-
morillonite. This would better satisfy the rather dif-
fuse negative charge originating from the octahedral
clay layer. It is possible that neutral M6C112(X)n species
could form, however, electrostatic binding to the silicate
would not be possible. The formation of [M6Cll2(X)6](6-n)-
ions by the complete anation of the cluster is a well known
procedure53’55.

The use of the ion exchange isotherms and the plots
of sodium release upon cluster binding makes it possible
to obtain predictable and reproducible loadings of the
metal cluster on the montmorillonite. These plots in-
dicate that a loading of 33 mmoles of Nb6C12+’3+

. l2
mmoles of Ta6Clig’3+ per 100 g of montmorillonite should

or 23

56

result in the release of approximately one cation—exchange

equivalent of sodium from the Clay. Loadings below this

level result in the formation of interstratified materials

as shown by x-ray powder diffraction. Higher loadings

simply result in the clogging of the clay interlayer with

large amounts of metal cluster cations of reduced charge.

Therefore the cluster loadings enumerated above were used for

all physical characterization and pillaring studies.
Elemental analyses confirms that greater than 90% of

the interlayer sodium has been replaced through ion ex-

change by M6ClI2,3+ ions. The anhydrous unit cell formula
for montmorillonite loaded with 33 mmoles of Nb6Clig’3+

1.8+
(X)0.67]0.2u ’
+
Na

0.06‘EA13.11’ “$0.50’ Fe0.u0 7.91, A10.09)°2O(OH)L,.
where X equals OH' or 01-. The average Cluster charge

Per 100 g was established to be [(Nb6C112)2+’3+

1(Si

of 1.8+ can be calculated from this formula based on 0.50
exchange equivalents per unit cell. The anhydrous unit

cell formula for montmorillonite loaded with 23 mmoles of

2+,3+ 2+,3+

per 100 g was established to be [(Ta6C112)
2.7+ +

(X)y]0.l7’ Na0.0u‘[A13.11’ “50.50, F°0.u0](317.9l’ A10.09)’

020(OH)“’ where O i y g 0.3. The average cluster charge

of 2.7+ can be calculated from this formula. The maximum

oxidation state of 3+ for the cluster can be designated

based on the UV-visible spectral data of the intercalates.

 

(.3

(I)

 

J

m

*4)

(I)

(7

('3

57

C. Physical Characterization of M6gli:’3+-Montmorillonites

 

 

X-ray diffraction studies on oriented film samples of

2+,3+
12

of 18.uso.2 A and 18.u10.u A for niobium and tantalum

M6Cl -montmorillonites indicate a mean basal spacing
cluster intercalates respectively. The observed spacing
is near the value expected from the combined thickness of
the silicate sheet (9.6 A) and the diameter of the an-
hydrous intercalated cluster ion, 9.8 A as determined
from x-ray crystallographic data53’88’89. The basal
spacing is slightly less than the value calculated from
the simple addition of the two components, suggesting
that the cluster cation partly fits into the hexagonal
holes on the basal surfaces of montmorillonite.

The diffraction pattern for a Nbéclfg’3+-montmoril—
lonite, with a loading of 33 mmole/100 g of clay is il-

lustrated in Figure 3-7. The x-ray diffraction pattern for

2+,3+
12

g of clay, is shown in Figure 3—8. The observation of at

a Ta6Cl -montmorillonite with a loading of 23 mmole/100
least 6 orders of 002 diffraction is typical of these
materials and indicates a regular ordering of the inter-
layer. Presumably, the degree of hydration and positioning
of the cluster cations is quite uniform throughout the

clay interlayer. The basal spacing and interlayer

ordering appear to be independent of cluster loading above
the levels of these samples. Higher cluster loadings

cause no increase in the basal spacing, thus only a single

 

58

 

.mcofipomLMMHc noopo ponwfin on»

wcfim: popwfizoawo mm; x m.on:.wa mo wcfiomam Hmmmn some one .ooHHm
mmmfiw a co counoaazm EHHM Can» a mo Snow on» CH mm: maqemm one

icoo mpficoaafipoEpsoEi

.mwao mo w ooa\popmsao ho moHoEE mm wCHCHMp

NH

+m.+maomoz no chopped sofioooooefio gosix one

.mim madman

59

-N

m

 

209
m N. new.

“N09

0.

   
 

mim opswfim

mm $2 30

m. x. m. w. om mm ow mm mm on mm

   
  

 

60

.mmcfiomam COHpompmmfiU pocno nonmfin on» wcfimflafiu:
consazoamo mmz m :.on:.ma mo wcfiomam memn some one .opfiam
mmmHm o co nopQOQQSm EHHM CH5» m mo Show on» CH mm: Canaan one

.mmao mo w ooa\nopmsHo mo moHoEE mm wcficfimu

icoo opficoaafipoEpcoEi+m.Wwaomma mo Choppmq coapomLQMHU mmpix one

.wim opzwflm

«28x9

200‘

61

«NSC

 

 

N\

3

«89

Q

mim ogzwfim

mm. 8....on
2 on

L

3.09

D

mm

b

VN

1P

309

QN

ON

$8»

on

Wm.

v0

 

 

 

 

 

 

()

**‘J

(I)

I1

In.“

 

 

 

62

layer of M6ClI2’3+ ions occupy the clay interlayer. For
comparison, a Na+-montmorillonite possesses a spacing of
12.5 A with only two or three orders of diffraction evident
under analogous conditions.

Both metal cluster intercalates retain approximately
10% of their CEC in the sodium form at the cluster loadings
used. The sodium montmorillonite is randomly interstrati-
fied with the cluster exchange form and only makes a
slight contribution to the x-ray diffraction pattern as a
whole. This contribution is manifested as a decrease in
the depth and symmetry of the valley between the 001 and
002 diffraction peaks relative to that of montmorillonite

containing cluster loadings of 50 mmoles/100 g of silicate.

2+,3+
12

consist essentially of the spectrum of the metal cluster

Infrared spectra of the M6Cl -montmorillonites

chloride superposed on that of the silicate. Figures 3-9

and 3-10 illustrate the relevant IR spectra in the region

1000—250 cm'l. 1

The very sharp band at 330 cm- for the

niobium cluster (Figure 3-9a) is also observed for the

Nb6Clig’3+-montmorillonite. The broad band at 3u0-310 cm‘1
for the tantalum cluster (Figure 3-10a) is similarly ob-
2+,3+

intercalate (Figure 3—10c). These
1 1

served in the Ta6Cll2

infrared bands at 330 cm- and 3“0-310 cm- are assigned

to the M-Cl
57.91.

br wagging vibrational mode of the cluster

core These bands roughly coincide with one of the

absorbtion peaks attributed to carbonate impurities present

in the clay sample92.

°/.Transmih‘ance—~

Figure 3-9.

 

 

 

 

 

1000 300 650 I 400 I 250
Mbwmrber (cm')
Infrared spectra of (a) (Nb60112)Cl2°8H20,

2+,3+_
12
2+,3+
12

morillonite treated at 130° (2“ h) and 2“0°

(b) Na+-montmorillonite, (c) Nb Cl
, 6

montmorillonite, and (d) Nb6Cl -mont-

(2“ h) lg vacuo.

Spectra (c) and (d) are of intercalates con-
taining 33 mmoles of cluster/100 g of clay.
All samples were prepared as KBr disks.

‘%Thwmnfifimxr-—v

Figure 3-10.

”55 ' my; ' a») ' 4d) ‘.ab

 

 

 

 

wamunkv amt)

Infrared spectra of (a) (Ta60112)012°8H20,

2+,3+_
12

montmorillonite, and (d) Ta601ig’3+-mont—

(b) Na+-montmorillonite, (c) Ta6Cl

morillonite after thermal treatment at 120°
(2“ h) and 2“0° (2“h) lg vacuo.
Spectra (C) and (d) are of intercalates con-

taining 23 mmoles of cluster/100 g of clay.
All samples were prepared as KBr disks.

65

2+,3+
12

lonites are shown with the solution spectra of the respec-

The UV-visible spectra of the M6Cl -montmoril-
tive Cluster salts in Figures 3—1 (p. 38), and 3-2 (p. ““).
These have already been discussed with respect to the oxida-
tion states of the intercalated metal clusters. The spectra
in Figure 3—1 are virtually identical indicating that the
niobium cluster is unchanged upon intercalation. The
spectrum shown in Figure 3-2b demonstrates that the intact
tantalum has been intercalated, although some shifting of
the average oxidation state has occurred.

The ESR spectrum of Nb6Clig-montmorillonite consists
of a single broad (AH is equal to 600 G) resonance at g
equal to 1.95 as illustrated in Figure 3-11. Line broaden—
ing due to anisotropic interactions within the solid state,
cause the hyperfine splittings expected from equal inter-
action of the unpaired electron in the Nb6Cli; with the
six niobium nuclei (I = 9/2) to be unresolved. This
spectrum is quite similar to that reported for polycrystal-
line [(C2H5)uN]3(Nb6C112)C1657. Figure 3—12 shows the
room temperature ESR spectrum of Ta6C1iz-montmorillonite
formed upon partial oxidation of the metal cluster inter-
calate. This ESR signal at g equal to 1.92 is sufficiently
line broadened to obscure the hyperfine splitting expected
from electron interactions with six equivalent tantalum
nuclei (1 = 7/2). This spectrum is quite similar to that
recently reported for a phosphine derivative of the

58a.

cluster

Figure 3-11.

66

g= |.95

5006

The ESR spectrum of Nb6Clig—montmorillonite.

This spectrum consists of a single broad reson-
ance (AH m 600 G) at g equal to 1.95. The
spectrum of the DPPH reference is shown. The
sample was in the form of a freeze-dried powder
packed into a quartz glass tube. The spectrum
was taken at room temperature.

Figure 3-12 .

67

§7=L92

 

 

 

DHWI

The ESR spectrum of Ta6Clig—montmorillonite
formed by partial air oxidation.

The spectrum consists of a single broad reson-
ance at g equal to 1.92. The spectrum of the
DPPH reference is shown. The sample was in
the form of a freeze-dried powder which was
exposed to the air for several hours and

then packed into a quartz glass tube. The
spectrum was taken at room temperature.

68

The air oxidation of the intercalated metal cluster
cations is promoted by the high acidity in the inter-
layers of dried clay samples. Color changes for the inter-
calated tantalum cluster are suggestive of oxidation to
a “+ cluster within hours of oxygen exposure in a dried
form. Absorbtion bands in the UV-visible spectrum of an
air oxidized sample appear at Amax equal to 750, 505, “25
and 353 nm. The presence of the broad band at 750 nm and
the sharper band at “25 nm supports the assignment of the
“+ oxidation state to the interlayer tantalum cluster in
comparison to the literature values listed on Table 3-1.

Air oxidation of the intercalated niobium cluster to the
3+ oxidation state requires days to occur. A sample of
Nb6Clig-montmorillonite was prepared which contained less
than 10% of the cluster molecules in the 3+ oxidation state
as shown by quantitative ESR measurements. After 3 days
<3f exposure to the air ESR measurements indicate that ap-
Iaroximately 30% of the clusters have been oxidized.

The hydrolysis reaction with interlamellar water main-
tnains the interlayer charge balance during air oxidation
OII‘the intercalated metal cluster. Significantly, the air
Cucidized clusters remain intercalated and maintain the high
irrterlayer spacings. A sample of Ta601ig-montmorillonite
fcufimed upon complete air oxidation of the intercalated

cluster possessed a basal spacing of 17.9 A after approxi-

maiiely'6 months of exposure to air.

69

D. The Pillaring of Montmorillonite by Niobium and Tantalum

Oxide Aggregates

The essential step in the pillaring process is the
formation of a dispersion of metal oxide which is distrib-
uted uniformly throughout the silicate interlayer so as to
form a prOpping system of high structural integrity. Pre—
viously, this has been accomplished through the calcina-
tion of polynuclear hydroxy cations formed in aqueous
solution and placed within the clay interlayers through
ion exchange. The present work makes use of an alternative
approach in which metal cluster cations function as pillar
precursors. This function requires the metal clusters to
decompose lg gltg to produce interlayer metal oxide prOps.

The well establishedBu’58’59 hydrolytic instability of
the M6Cl?; cluster ions suggests that the refractory M205
<3xides may be formed through decomposition of the metal
czluster. The investigation of the niobium cluster hy-
<irolysis by Allen and Sheldonsac led these workers to
Ixropose the formation of a u-hydroxo species, [Nb6Cl(12_n)-
((3H)n1012°mH20 as a precipitate from alkaline solution. A
unore in depth study of the hydrolysis and decomposition of
nixobiwm and tantalum bromide clusters was carried out by
Seluafer and Brnicevicsg. To explain the sudden decrease in
PP1 during base titration, and the apparent cluster decom-

POESition, they proposed the reaction:

70

[M6Br8(OH)u](OH)2- + 5H20 + 3M205 + “H+ + 8Br' + 8H2

(3-7)

The cluster containing both bridging and terminal
hydroxides undergoes hydrolytic decomposition and simul-

taneous oxidation of the metals to yield the pentoxides

with production of hydrogen gas. For this type of reaction

to be applicable to the pillaring of smectite clays, however,
the metal oxides formed must remain within the clay inter-

layers. Any anionic species formed during reaction would

be repelled by the negatively charged silicates, and thus
be expelled from the interlamellar region. Therefore,

intermediates leading to the pillaring oxides must be either

cationic or neutral. This requirement can be met through

the use of carefully controlled reaction conditions.

Differential scanning calorimetry (DSC) was utilized

to observe the thermal transitions of the M6Clig’3+-mont-
Inorillonites and thus delineate the interlayer reactions

(3f the metal cluster. Figures 3-13a and 3-l“a illustrate

tzhe calorimograms of the niobium and tantalum cluster

sualts, respectively. The strongly endothermic transi-

txions observed at approximately 200° are associated with
tile loss of water molecules coordinated to the metal
CJJisters and water within the crystallographic lattice.
Trris assignment is consistent with the results reported
by’ Schafer93. Mass spectroscopy in high vacuum has demon-

Strated the loss of HCl from these metal clusters in the

71

 

e—EhbMHWiEmWEfln—+

 

 

 

 

 

 

Figure 3-13.

0
b
___ 1___~___tg
c
la) zfi an 4a) an

 

The differential scanning calorimograms of

(a) (Nb6C112)C12°8H20, (b) Na+-montmoril-

ig’3+-montmorillonite

lonite, and (c) Nb6C1
containing 33 mmoles of cluster/100 g of

silicate.

These thermograms were measured in the DSC
mode, however the ordinate is measured in
arbitrary thermal units.

72

 

«—EhbMaWiEhmrwn—a

 

 

 

 

Figure 3-1“.

 

The differential scanning calorimograms of

(a) (Ta6Cll2)Cl2°8H2O, (b) Na+-montmorillonite,

and (c) Ta6Cli:’3+-montmorillonite containing

23 mmoles of cluster/100 g of Clay.

Thermogram (d) is a typical baseline measured
with empty pans at sample and reference posi-
tions. The thermograms were measured in the

DSC mode however, the ordinate is measured in
arbitrary thermal units.

73

.914.

temperature range 200—260 The low intensity ende-

thermic peaks in the calorimograms 3-l3a and 3-l“a are

probably associated with less of additional HCl and crys-
tallographic transitions associated with the cluster de-
composition. Infrared spectroscopy indicates that these

clusters are no longer intact following heating up to 500°

1 1

inside the thermal analyzer. The 330 cm- and 3“0-310 cm-

infrared absorbtion bands typical of the M-Cl r vibrational

b
modes are absent.

Figures 3-l3b and 3-l“b illustrate the calorimogram
typical for a Na+-mentmorillenite. The broad endotherm
at approximately 120° is associated with the less of water

adsorbed on the silicatelS. Figures 3-l3c and 3-l“c i1-

2+,3+
12

The bread dehydration endotherms of the montmorillonite and

lustrate the calorimograms of the M6Cl —mentmerillenites.
the metal cluster are clearly evident. At about 260°
weak endotherms are discernable in both figures which are
probably associated with the hydrolytic degredation of
the intercalated metal clusters.

A combination of these DSC results and experimental
observations led to the establishment of a calcination
sequence which provides the necessary conditions for the

formation of interlayer metal oxide preps. Thermal treat-

2+,3+
l2

<3ynamic vacuum for 2“ h was designed to limit the amount

Inent of the M601 -mentmorillenites at l20-l30° under

(If interlayer water present while retaining the integrity

fo the intercalated metal cluster ion. This is an essential

7“
step in the pillaring process, otherwise, the M6ClI2,3+'
montmorillonites quickly collapse at only moderately
elevated temperatures (200°). Further calcination of the
pretreated intercalate lg tgggg at 2“0° for 2“ h, promotes
the decomposition of the cluster within the clay inter-

layer and the successful formation of metal oxide pillars.

2+,3+
l2

lonites following the various calcination steps are correlated

Typical experimental results for M6Cl -montmoril—
in Table 3-2. The observation of a color change after 2“ h
at 2“0° suggests that a decomposition of the metal cluster
has occurred. This conclusion is corroborated by spectros-
copic changes in these materials. Loss of the infrared
absorbtion bands associated with the intact metal clusters
is evident in Figures 3-9d and 3-10d. The UV-visible
spectra of these calcined materials, illustrated in Figures
3-15b and 3-l6b, consist merely of a broad tailing into the
UV. These spectra are in marked contrast to those of the

M6Clig’3+-montmerillonites also illustrated in these figures.

Also, the ESR signals attributed to intercalated M6C1ig
ions are no longer observed in these materials after cal—
cination.

The chlorine to metal ratio determined by elemental
analysis is included in Table 3-2, as an indication of the
integrity of theinetal cluster. The ratio is close to the
‘theeretical value of 2.0 following the thermal treatment

Eit 130°, but is reduced by an order of magnitude after

Calcination at 2“0° under vacuum. The intercalated metal

75

Table 3—2. Some Physical Characteristics of Thermally
Treated Niobium and Tantalum Interlayered
Montmorillonites.

 

 

A. Niobium Interlayered Montmorillonite (33 mmoles of

 

 

 

 

 

Nb6Clig’3+/100 g of Silicate).
Temp., °Ca Time, h Color Cl/Nb d001, A
25 -- Green 2.02 18.“i0.2
130 2A Khaki 1.96 19.3:0.6
2“0 2“ Grey 0.20 19.210.“
325° 6 Grey -——- l8.7:0.6
“00b 8 Grey ---- 18.“,10.0
1450b 2“ Cream --—— 10
B. Tantalum Interlayered Montmorillonite (23 mmoles
Ta601ig’3+/100 g of Silicate).
Temp., °Ca Time, h Color Cl/Ta d001, A
25 -— Blue-Green 2.22 l8.3i0.3
120 2“ Blue-green 2.18 18.9:0.3
2u0 2“ Grey 0.13 19.120.5
350b 5 Grey --—- 19.110.“
“00b “ Grey ---- 9.7

 

 

aAll calcinations were performed under dynamic vacuum.

bThese samples were calcined at l20-130° for 2“ h followed
by 2“0° for 2“ h prior to treatment at these temperatures.

76

 

Awnnme—~

 

 

 

 

 

 

Figure 3-15.

 

 

UV-visible spectra of (a) Nb6Clig’3+—mont-

morillonite, (b) Nb6Clig’3+-mentmerillonite

after thermal treatment at 130° (2“ h) and
2“0° (2“ h) lg vacuo, and (c) Na+-montmoril-

lonite.

Spectra (a) and (b) are of intercalates
containing 33 mmoles of cluster/100 g of
clay. These spectra were measured versus a
Na+-montmerillonite reference. Samples were
mulled in Nujol and suspended between quartz
glass plates. Absorbance was measured in
arbitrary units.

77

 

 

 

Ammwwme—o

 

 

 

 

 

we 4d) in ed) in ad) ab ian
unearth (nm)
F 2+,3+
igure 3-16. UV-visible spectra of (a) Ta6Cl12 -ment-

morillonite and (b) Ta6Clig’3+-montmoril-

lonite after treatment at 120° (2“ h) and
2“0° (2“ h) lg vacuo.

Both intercalates contained 23 mmoles of
cluster/100 g of clay. Samples were mulled
in Nujol and suspended between quartz glass
plates. Spectra were measured against a
mulled Na+-montmorillonite reference sample.
Absorbance was measured in arbitrary units.

78

Clusters have decomposed at 2“0°. The HCl given off during
this decomposition can be detected if the 2“0° calcination
is performed under flowing nitrogen gas. Wet litmus paper
placed in the effluent gas stream rapidly turns red in-
dicating the presence of HCl.

The degredation of the cluster at 2“0° probably occurs
through the hydrolytic displacement of bridging chlorides
similar to the reaction occurring in solution. A reaction

scheme analogous to Equation (3-7) can be written:

 

 

 

 

M 012+ + 15H 0 33190—4 3M 0 + 2H+ + 12HC1 + 8H
6 l2 2 vacuum 2 5 (g) 2(g)’
(3-8)

wherein the horizontal lines represent the silicate layers.
At elevated temperature, water molecules terminally co-
ordinated to the cluster cation and physiserbed on the
silicate interlayer surfaces will react with the inter-
calated M5013: ion. This results in the displacement

of the bridging chlorides by hydroxide ions, presumably by
an 8N2 mechanism and the liberation of HCl. Eventually
hydroxide substitution promotes the oxidative degrada-
tion of the cluster structure to yield the interlayer
Inetal oxide and H2 gas. Protons formed during the re-

.action remain within the montmorillonite to balance the

(Bharge, either as discrete interlayer cations, or as

79

interlayer hydroxides formed upon H+ penetration into the
crystalline lattice of the clay.

Strict stoichiometric control over the number of waters
participating in the reaction is impossible due to the
nature of the silicate, but sufficient water is apparently
available for complete reaction of the cluster to occur.

It is only when a relatively large excess of interlayer
water is present that the reaction leads to collapse of
the pillaring system at low temperatures.

Elemental analysis of a montmorillonite loaded with

2+,3+
l2

calcined lg vacuo at 130° for 2“ h and 2“0° for 2“ h cer—

33 mmoles of Nb6Cl per 100 g of clay which has been

responds to the unit cell formula (Nb205)0 57, H; ““’

”36.06'EA13.11’ “80.50’ F°0.u0](517.91’ A10.09)O20(OH)A°
Analysis of a Ta6Cli:’3+-montmorillonite containing 23
mmoles of cluster per 100 g of clay treated following the
same calcination sequence indicates the anhydrous unit cell
f°rmUIa (T3205)0.u8’ H0.u6’ Na0.0u'[A13.11’ Mgo.50’ F°0.u0]'
(317.91: AlO 09)O20(OH)“’ The interlayer metal oxide has

been formulated as M205 but may exist in an aggregate con-

taining six metal atoms (vide infra). These analyses in-

 

dicate that essentially the original amounts of metal re-
main on the clay and are not volatilized off during the
thermal treatment.

Table 3-2 also lists the mean x-ray diffraction basal
tipacings of the niobium and tantalum interlayered mont-

"HDrillonites following treatment at various temperatures.

80

These average spacings and the standard deviations on the

values are calculated using d values derived from the

001
higher 001 diffractions present. The mean basal spacing
remains essentially constant, within experimental uncer-
tainty, up to 325° and 350° for niobium and tantalum inter-
layered clays, respectively. This is quite remarkable in
light of the drastic structural changes which must occur
in the metal Cluster species during the degredative
hydrolysis and oxidation reactions. These large basal
spacings indicate that an interlayer spacing of approxi-
mately 9.5 A is maintained due to the presence of inter-
calated M205 molecular props. An interlayer expansion of
this magnitude is consistent with the presence of a multi—
nuclear pillar containing three or four planes of oxide
anions. For comparison, at 225° 3 Na+—montmorillonite will
possess a basal Spacing of approximately 10 A, indicative of
a collapsed interlayer.

Representative x-ray diffraction patterns of tantalum
intercalated montmorillonites are shown in Figure 3-17.
One important and consistent feature of these patterns
is that the 002 diffractionijsmore intense than the 001.
This is characteristic of the M6Clgg-montmorillonites with
the intensity difference being more pronounced in the
tantalum cluster intercalates. The large basal spacings
are confirmed using the 003 and 005 reflections, which

indicate that the tantalum oxide pillared clay is still

intact at 350°, despite the lack of an observable 001

 

 

I

Figure 3-17.

81

Representative x-ray diffractograms of

thermally treated Ta601ig’3+—montmori1-

lonites.

These samples were (a) air dried at room
temperature, (b) heated 120° (2“ h) and
2“0° (2“ h) under dynamic vacuum, (c)
treated as in (b) and heated 350° (5 h)
lg vacuo, and (d) treated as in (b) and
heated “00° (“ h) lg vacuo. All samples
contained 23 mmoles of cluster/100 g of
clay. Samples were in the form of thin
films deposited on glass slides.

 

 

 

82

 

 

 

     

(006) (005) (004) (003)

 

a.
N1

a'zaba'ea'safizézbi'ei'efiiéibés

DxMESZQ

Figure 3-17

83

diffraction peak. There is no evidence to suggest stability
of the pillared clay beyond 350°, however. The dif-

fraction peaks observed in Figures 3-l7d, and similarly, in
3-18d, are attributable to the Na+-montmorillonite present as
an m10% impurity in both metal cluster intercalates. These
peaks may also contain H+-mentmorillenite, formed according
to Equation 3-8, which has collapsed in sufficiently ordered
multilayer domains to cause a coinciding x-ray diffraction
pattern.

Figure 3—18 shows the x-ray diffraction patterns for
niobium interlayered montmorillonites. Four orders of
diffraction are evident following calcination at 2“0°,
but after six hours at 325°, the high interlayer spacing
is retained but only two orders of diffraction are evident.
After eight hours at “00° the niobium oxide pillared
structure begins to break down creating a biphasic system
containing collapsed clay with externally deposited Nb205,
and montmorillonite which remains pillared. Longer treat—
ment times and higher temperatures bring about the dis—
order and collapse of the pillaring system through the
fragmentation and migration of the interlamellar metal
oxide. X—ray diffraction patterns assigned95 to bulk

Nb205 have been observed (arrows) from a sample of

2+,3+
12

carefully handled with exclusion of oxygen. This is

Nb6Cl —montmorillenite treated to “50° lg vacuo and

consistent with the suggested formation of interlayer

M205, and demonstrates that the niobium oxide possesses

 

Figure 3-18.

8“

Representative x-ray diffractograms of

thermally treated Nb6Clig’3+-montmeril-

lonites.

2+,3+
12
lonite containing 33 mmoles of cluster/

100 g of clay (a) heated at 130° (2“ h)

and 2“0° (2“ h) under dynamic vacuum, (b)
treated as in (a) and heated at 320° (6 h)
lg vacuo, (0) treated as in (a) and heated
at “00° (8 h) under vacuum, and (d) treated
as in (a) and heated at “50° (2“ h) under
vacuum with exclusion of oxygen during the
sample handling and x-ray diffraction.
Arrows in (c) and (d) designate points of
ingrowth of diffraction peaks attributed

to Nb205. All samples were prepared as

thin films deposited on glass slides.

These samples were Nb6Cl -montmeril—

85

 

ice A

i t ‘ 8.71.68
(OOI)

‘ (002)

(004) (003)

(002) . I9.21.4 A
(OOI)

(004) (003’

 

3102826221222le (6 I“ I2 ioé E I 2
Degrees 29

Figure 3-18

86

sufficient mobility at these temperatures to form a

separate crystalline phase.

2+,3+
12

before and after calcination at 2“0° are consistent with

ESCA measurements96 of Nb6C1 -montmorillonites

the formation of an oxidized niobium species. The Nb3d5/2
binding energy was measured versus several known niobium
samples, and showed a 1.7 eV increase in binding energy

to 208.2 eV, indicating that an increase in the oxidation
state of niobium had taken place during calcination. Defi-
nite assignment of an oxidation number to Nb is not possible
based solely on this data, however.

Despite the major changes in the cluster structure at
2“0°, multiple orders of 001 diffraction are still evident
in the x—ray patterns of both the niobium and tantalum
interlayered systems. This well-ordered interlayer sug-
gests that a fairly uniform distribution of metal species
is retained throughout the hydrolysis and oxidation re-
actions ef the cluster.

Metal oxide moieties of molecular dimensions may be

formed, representing a very high state of dispersion.

The stoichiometry of Equation (3-8) suggests the formation

of a M6015 species. This M6015 would probably be struc-
turally similar to the hexanuclear niobate or tantalate
M6039 anions formed in solution97. The metal atoms of the

proposed M6015 species would be arranged in an octahedron
containing a central oxide anion which coordinates to all

the metal atoms. The remaining fourteen oxide ions would

87

adopt terminal and face bridging positions to complete
the octahedral coordination of each metal atom. These
interlayer oxide species could possibly migrate short
distances to form more condensed islands or Chains of
M2 5, however.

Nitrogen BET surface area data for niobium and tanta-
lum interlayered montmorillonites are listed in Table 3-3.
The numbers in parenthesis are surface areas calculated per
gram of silicate, compensating for the large mass of the
interlayer metals present. These values can be directly
compared to the total interlayer surface area available
in montmorillonite; approximately 750 m2/g.

Both the niobium and tantalum oxide pillared clays
possess available surface areas significantly below those
expected based on their retention of high interlayer

spacings. The M60 moiety, which is assumed to be the

15
pillaring species, can be estimated to be approximately

spherical and 9.“ to 9.6 A in diameter based on measurements
of models. Calculations suggest that only about “0% of the

silicate surface areas should be occupied by M20 at the

5
metal cluster loadings used. Thus, surface areas of 250
m2/g and 230 m2/g for niobium and tantalum interlayered
montmorillonites, respectively, can be calculated. As
shown in Table 3-3, the actual surface areas for these
interlayered montmorillonites, as measured by using a

nitrogen BET method, are approximately one-quarter as

large as expected. Apparently, adsorbent molecules are

88

Table 3-3. Surface Areas of Niobium and Tantalum Oxide
Pillared Clays.

 

 

A. Niobium Interlayered Montmorillonite (33 mmoles of
cluster/100 g of clay).

 

Temp., °C Time, h Surface Area, m2/ga
130 2A 58(68)b
2“0 2“ 6“(78)b
325 6 63(77)b

 

B. Tantalum Interlayered Montmorillonite (23 mmoles of
cluster/100 g of clay).

 

2“0 2“ 32(Li0)b
350 5 70(88)b

 

 

aSurface area per gram of intercalated material.
bValues in parenthesis are surface areas calculated Ber
gram of silicate exclusive of intercalated M6Cllga3 or

M20.

89

unable to penetrate very deeply into the Silicate inter-
layer due to the formation of narrow pores (m5 A) and to
blockages of the pore openings. These blockages may be

due to migrations of metal oxide fragments. The surface
areas measured per gram of silicate are somewhat higher
than the 15 m2/g measured for a freeze dried Na+-montmoril-
lonite sample totally collapsed at 225°. By comparison,
the niobium and tantalum pillared clays may contain some
interlayer pore openings available for adsorbtion. This

is supported by the observation of hysteresis peaks in the

nitrogen desorbtion curve measured at low p/pO.

E. Conclusions and Recommendations

 

The M6C112,3+ metal cluster cations bind to Na+-

montmorillonite through ion exchange for interlayer sodium
ions to form metal cluster-montmorillonite intercalate
compounds. This ion exchange reaction is complicated by
cluster oxidation and by hydrolysis and anation reactions.

The latter two reactions reduce the average charge carried by

2+,3+
12

loadings on the layered silicate than was originally an-

2+,3+_
l2

montmorillonite intercalation reaction that was gained by

the cluster and allow for significantly higher M6Cl
ticipated. The level of understanding of the M6Cl
this study, was sufficient to allow for the successful ap-

plication of these intercalates as precursors for pillared

clay formation.

90

The intercalated metal cluster cation serves as a
pillar precursor, undergoing degredative hydrolysis and
oxidation lg gltg to yield M205 props. These metal oxide
pillared clays are formed through calcination lg tgggg
are are stable to “00° and 350° for niobium and tantalum
interlayered montmorillonites, respectively.

Future work in this area should be concerned with
increasing the surface areas of these materials. This
would allow for greater adsorbtion of reactant molecules
within the interlayer pore structure, and thus, increase
the general utility of these pillared clays as sorbents
and catalyst supports. One possible approach would utilize
smectite clays, either natural or synthetic, with lower
cation exchange capacities as the host silicate. A lower
CEC would necessitate lower metal cluster loadings to
achieve homoionic replacement of the interlayer sodium
ions. This study was performed using a montmorillonite
possessing a relatively low CEC, perhaps only a small
reduction in layer charge is necessary. The use of a
clay possessing a charge which has been Significantly re-
duced by penetration of Li+ ions into the lattice may not
be applicable for pillaring reactions however. This type
treated clay failed to exhibit any ability to be pillared
by zirconyl polynuclear cationsu3.

The formation of the mixed metal Cluster salts

[(CZH“)“N]X[(Ta(6-y)’MOyC112)C16]’ where x is equal to 2

91

or 3 and y is equal to l or 2, has recently been reported
by Meyer and McCarley98. It is quite conceivable that the
mixed metal cluster cationic core contained in these com-
pounds could be intercalated into smectite clays and oxi—
dized lg gltg. This would produce mixed metal oxides con-
taining molybdenum, which might prove to be quite interest-
ing catalytically.

One important limitation to the application of metal
oxide pillared clays in catalysis is their hydrothermal
instability. The collapse of the pillared structure in
the presence of steam appears to be characteristic of metal
oxide pillared clays in general, and is especially true
in the case of niobium and tantalum oxide montmorillonites.
This places severe limitations on the types of catalytic
reactions or regeneration steps these materials may be

subjected to.

IV. THE INTERLAYERING OF MONTMORILLONITE
WITH MOLYBDENUM OXIDES BY THE INTERCALATION

AND CONVERSION OF MOLYBDENUM CLUSTER CATIONS

A. Intercalation of Molybdenum Cluster Cations into Na+-

Montmorillonite

 

To test the general utility of the technique of utiliz-
:Ung intercalated metal cluster cations as metal oxide pillar
precursors in the formation of pillared clays, a brief study
was undertaken using a molybdenum chloride cluster. This
investigation was performed by using the procedure de-
veloped previously for the niobium and tantalum cluster
reaction as a general guideline.

The formation of a thermally stable molybdenum oxide
pillared clay would be highly interesting from the stand-
point of catalysis. This pillared clay would contain a
catalytically active metal oxide in a highly dispersed
state within the unique environment of the clay inter-
layer. Thus, the metal oxide could serve a dual role
as both a structural component of the intercalate and as
a catalytically active agent. Molybdenum species are
active catalysts for a variety of reactions including

99,100

oxidationgg, ammonoxidation , and hydrodesulfuriza-

tion (HDS)12.

92

93

The molybdenum cluster chloride, (Mo6Cl8)Clu'2H20

used in this study was prepared in good yield by fellow-

67. The synthesis

ing the procedure of Nanelli and Block
utilizes the conproportionation reaction between molybdenum
pentachloride and molybdenum metal powder according to
Equation (“—1). The chloride was sublimed over the metal

650°

2.“MoC15 + 3.6Mo-———v (MO6C18)C1“ (“-1)

200°

 

(H3O+)2(MO6C18)C16'6H2O (Mo6C18)Clu°2H2O (4-2)

vacuum
+2HC1 + 6H2O

at 650°C under flowing nitrogen. The metal cluster is
isolated from hydrochloric acid solution and heated to
200° under vacuum to yield the dihydrate salt according
to Equation (“-2).

Initial attempts to intercalate the cluster into mont-
morillonite were thwarted by the limited solubility of the
(Mo6C18)Clu'2H2O in aqueous solution. The cluster is
rather soluble in absolute methanol, but binding of the
cluster to Na+-montmorillonite is limited in this solvent.
Apparently, insufficient swelling of the smectite clay
in methanol ((1001 is 18.“ A) and also, the low extent of
ionization of this salt to yield a cationic metal cluster

core and free chloride anions severely hampers the ion

9“

exchange and intercalation reaction. The bonding of the
terminal anionic ligands is important to the stabilization
of the cluster due to the inability of the Mo6C18 cluster
core to tolerate a buildup of high positive Charge. The
replacement of a maximum of two terminal halide ions in
the presence of a large excess of strongly donating ligands
(DMSO, DMF) is indicative of the importance of terminal
anion binding79.
The lability of four of the twelve halides in the

101 due to

MoX2 salts was recognized as early as 1919
their replacement by hydroxide ions. IsotOpic studies

by Sheldon63 utilizing chlorine-36 indicated that exchange
of the terminal halides occurs rapidly but that exchange

of the bridging halides takes place immeasurably slowly.
Several workers have utilized argenometric precipitation to
remove the labile terminal halides. This procedure is uti-

lized in the preparation of mixed halogen species65, or for

62.79.

replacement by neutral ligands This latter process

would yield compounds of the type Mo6C18L6 A”, where A
is Clog, or [Mo6C18L2JAu, where A is CH3SO3 and L is
DMSO or DMF. This procedure was only marginally suc-
cessful, however. Many other ligands used in these
studies tended to yield products of poor quality. In
addition, the reaction is sensitive to the counterion

79

present with silver. Nitrate ion reportedly causes

precipitation and decomposition of the molybdenum cluster.

95

It should be noted that the perchlorate salts of the clus-
ter coordinated with pyridine, pyridine-N-oxide, and DMF
are violently eXplosive, with the shock sensitivity
diminishing upon removal of the organic ligands62.

Work in this laboratory led to an improvement of the
reaction technique and indicated that the slow addition
of a solution of the molybdenum cluster chloride in ab-
solute methanol to an alcoholic solution of silver per-
chlorate or tetrafluoroborate containing small amounts of
water (%7%) results in the precipitation of AgCl and the
formation of [M06018(OH)n(H20)(6_n)](Li-n)+ ions in solu—
tion. The presence of water is necessary to avoid the
precipitation of the cluster with silver as Ag2[Mo6C18]Cl6

62’79. The water

°nROH which occurs in absolute methanol
molecules bind to the cluster terminally and help to in-
crease its solubility in solution. Some dissociation

of water probably occurs to reduce the net cluster charge

(vide infra).

 

The reaction of silver ion with the terminal halides
of the metal cluster can be monitored by utilizing solu-
tion electrical conductivity measurements. A conducti-
metric titration was performed with AgBFu as the silver
source. Conductivity was measured in arbitrary units with
compensation made for volume changes during titration by
multiplying the measured conductivity (ohm-l) by the total

volume of the solution in milliliters.

96

As shown in Figure “-1, solution electrical conduc-
tivity is plotted relative to the moles of cluster added
per equivalent of Silver. The plot consists of two linear
sections with the titration end point at approximately 0.2“
moles of (Mo6C18)Clu'2H20 per Ag+ equivalent. This value
is within experimental error of the end point expected (0.25
moles of cluster/Ag+ equivalent) from the 1:“ stoichiometry

of the proposed reaction:

CHBOH/HZO
—fir“AgCl(S) +

 

“Ag+ + (MO6018)Clu‘2H20

[M06018(OH)n(H2O)(6_n)](u-n)+ + nH+ . (“-3)

The terminal chlorides are precipitated by silver and
replaced by water molecules in terminal coordination. These
terminally-coordinated water molecules can hydrolyze to form
a bound hydroxide ion and a proton which is released into
solution.

The negative slope of the first portion of the plot
indicates a net decrease in the electrical conductivity of
the AgBFu solution as reaction proceeds. This is con-
sistent with the replacement of four moles of Ag+ by one mole

1(“-n)+ and n protons.

of [(Mo6C18(OH)n(H2O)(6_n)
Electrolytic solution conductivity can be described

‘by Equation (“-“).

CONDUCTIVITY

 

 

0 oil

97

E———7‘|"

l

0.2 03 0.4 01.5 016 0.7

Moles(M06Cle)Cl4 Added/ Equivalents AgBF4

Figure “-1.

The conductimetric titration of AgBFu with
(M06018)Clu°2H2O.

The AgBFu concentration was 6. 93 x 10'5 M
in methanol containing 7% water. The 6.0“
x 10‘ M cluster salt solution in absolute
methanol was added to the silver salt. Con-
ductivity was measured in arbitrary units of
ohm'l mL.

98
A = 0iF(u+ + u_) (“-“)

The equivalent conductance, A, is a function of the frac-
tional ionization of the salt, a, the Faraday constant,

F, and the ionic mobilities of each ion, u102

. The de-
crease in conductivity during the titration can be at-
tributed to the low degree of ionization and low mobility
of the products formed. The molybdenum Cluster salt would
be expected to possess low ionic mobility due to its large
size.

In methanol solution, the protons formed would possess
approximately the same specific conductivity as the Ag+
ions. The observed decrease in conductivity indicates
that the number of protons formed, n, is less than four.
The AgCl precipitate does not contribute to the conduc-
tivity due to its low solubility in the solvent used.

The addition of (Mo6C18)Clu°2H20 beyond the end point
results in a very slight rise in conductivity and is in-
dicative of a small degree of cluster ionization to yield
[(M06C18)(H20)3Cl3]+ and chloride anion in solution. This
observation is consistent with the very low extent of

binding of (Mo6C18)C1u'2H o to Na+-montmorillonite which

2
occurs in methanol solution.

The reaction described in Equation (“-3) is equally
valid if the counterion to silver is perchlorate. AgClOu
was utilized for all additional studies. The [M06018-

(OH)n(H20)(6_n)](u—n)+ species formed in methanol is readily

99

soluble in aqueous solution, therefore, the stock solu-
tions of this species were diluted with water and the
major portion of the methanol was removed by rotary evapora-
tion. Removal of the methanol eliminates the negative
effects this solvent has on the metal cluster intercalation
reaction into montmorillonite and reduces the chance of form-
ing explosive molybdenum cluster perchlorate precipitates
containing methanol.

In aqueous solution the molybdenum cluster possesses
measurable acid/base properties. Sheldon69 has published
an acid titration curve of the M06C18(OH)§— ion formed
at high pH. This curve demonstrates that this molybdenum
cluster species reacts with two equivalents of acid and
is precipitated at a pH of approximately 8.5. This pre-
cipitate is thought to be Mo6C18(OH)u'xH2O69’lO3. The
cluster solution concentration used for this experiment
was not specified, however.

The acid/base properties of the molybdenum cluster
can be utilized for its intercalation into Na+—montmoril-
lonite. The reaction occurs during the drOpwise addition
of 0.1 M HClOu to a montmorillonite suspension which
is 5x.10-5 M or lower in [M06018(OH)n(H20)(6_n)](u-n)+.
The cluster stock solutions were adjusted to a pH of 10.0
during their synthesis to enhance their cold storage
stability. Consequently, the intercalation reaction mix—
ture containing Na+-montmorillonite and cluster stock

solution possessed an initial pH of approximately 8.0.

100

The low Cluster solution concentration is necessary to
avoid the formation of Mo6C18(OH)u-xH2O precipitates upon
acid addition. Apparently, the tetrahydroxo species has
a limited solubility in solution which was not reported
by Sheldon. The slow addition of HClOu to achieve a
final pH of 1.5, leads to further cluster protonation to
form cationic M06018(OH)3(H20)§ species in solution. This
cation will then undergo ion exchange for sodium on the
montmorillonite present in suspension. The rate of acid
addition must be slow enough so that the rate of cluster
protonation is roughly balanced by the rate of cluster
intercalation. The use of a sufficiently slow rate of acid
addition will minimize the formation of [Mo6C18(OH)n-
(H20)(6_n)](ClOu)(u_n)precipitates (where O i n i 3)
which bind to the external surfaces of the montmorillonite.
The details of the molybdenum cluster intercalation
were examined by measuring the number of equivalents of
sodium liberated from homoionic Na+-montmorillonite rela-
tive to the number of moles of [Mo6C18(OH)n(H20)(6_n)](Linn)+
added. These data are plotted in Figure “-2. This plot
contains an initial section which indicates a linear re-
lease of sodium as cluster is added, indicative of an ion
exchange reaction. The slope of this section, 1.05 Na+
released/cluster added, is consistent with the quantita-
tive uptake by ion exchange of a cationic cluster ion as

illustrated in Equations (“-5) and (“-6).

Millequlvolents Na‘ Released/1009 Silicate

101

 

 

O) 1 l 1 l l 1

io zoioioéoéoroeoeoioo

, +
Mlimoies of [MOGC'O(m’nr " Bound/IOOg Silicate '

Figure “-2. The isothermal sodium release upon cluster
binding to Na+-montmorillonite at pH 1.5.

Samples were prepared using a methanol free
cluster stock solution.

102

 

 

Mo6C18(eH)n(H20)E2:2(+ + (n-3)H+ : M06018(OH)3(H2O)§
pH = 10 pH = 1-5 (“-5)
———-pH 1.5
M06018(OH)3(H2O)§ + Me: ————> Mo6C18(OH)3(H20)§ + Na+
(“-6)

The horizontal lines represent the silicate sheets of
montmorillonite. Essentially quantitative uptake of the
cluster by the clay is observed through the range of this
experiment, and up to 180 mmole/100 g silicate loading.
Apparently the initial uptake of cluster is predominantly

by ion exchange, but at higher cluster loadings both cat-
ionic and neutral species are bound. The binding of neutral
[MO6C18(OH)n(H2O)(6-n)](Clo“)(“-n) as ion pairs or in small
crystalline aggregates occurs through hydrogen binding or
physical adsorbtion to the external surfaces of the clay.
This external binding becomes more prevalent as the frac-
tion of ion exchange sites on the montmorillonite occupied
by electrostatically bound cluster increases. The exter-
nally deposited clusters are easily removed by distilled
water washings at pH 7, however. The infrared spectrum

of the molybdenum cluster intercalated montmorillonite after
several washings with distilled water, illustrated in Fig-

ure “-“b (p.101), is free from bands assigned to the C10; ion

103

indicating that any ion pairs originally present have
been removed.

The elemental analysis of a molybdenum cluster inter-
calated montmorillonite prepared by the addition and quan-
titative uptake of 70 mmoles of cluster/100 g of clay
indicated that greater than 99% of the interlayer sodium
ions have been replaced. This analysis also showed that
of the 70 mmoles of cluster originally bound per 100 g of
silicate 53.6 mmoles remained electrostatically bound to
the montmorillonite. Thus 16.“ mmoles of cluster/100 g
of montmorillonite were freed from the Clay during mul-
tiple distilled water washings.

The chemical analysis of the intercalate containing 53.6
mmoles of cluster/100 g of clay corresponds to the
average unit cell formula [Mo6C18(OH)2.7(H20)3.31é:§;—

1(Si The

[Al3.ll’ Mgo.50’ FeO.“0 7.91, A10.09)O20(OH)A°
average charge per cluster can be calculated to be 1.3+
based on this formula. This correlates reasonably well
with the charge of 1.05+ as determined by the sodium
release measurements. A charge of greater than 1.0 sug-
gests that double protonation to produce the dication
Mo6Cl8(OH)2(H20)fi+ occurs in some fraction of the Clusters.
The difference between the average charge per cluster
calculated by using the sodium release data (1.05+) and by
using the analytical results (1.3+) is significant and

suggests the occurrence of a surface equilibrium. Equa-

tion (“-7) shows the reaction of two intercalated monocations

10“

 

 

2Mo6Cl8(OH)3(H2O)§ : Mo6018(0H)2(H20)fi+ + Mo6C18(OH)u(H2O)2

 

 

(“-7)

by proton transfer to produce an intercalated dicationic
cluster and a neutral Cluster in solution. The horizontal
lines represent the silicate layers of montmorillonite. The
position of this equilibrium shifts to the right (as written

above) during the washings with distilled water.

1.3+

B. Physical Characterization of Mosgl8(OH)2 7lM2gl3 3 :

Montmorillonite

 

The x-ray diffraction pattern of montmorillonite inter-
calated with 53.6 mmoles of molybdenum Cluster/100 g of
silicate is shown in Figure “-3. The basal spacing of
16.6 A is large enough to accommodate the thickness of the
silicate sheet (9.6 A) plus the height of the molybdenum
cluster, (7.2 A), as calculated from single crystal x-

60’10“, provided that this cluster

ray diffractometry
cation partly fits into the hexagonal cavities present on
the basal surfaces of montmorillonite. The keying of
cations into these cavities is a fairly well known
phenomenon. It has been observed for partly dehydrated

16 and is proposed to occur

136,37.

interlayer alkali metal cations

with intercalated alkyl ammonium ions as wel

.ooHHm mmwaw a co oopQOQQSm Eafim Umpcmfino cm mm cmnmaona mm3 mamemm one
.mpmofififim co m OOH two +mHmA0mmv~.mAmovaomoz
mo mmaoes m.mm mcficampcoo opficoaafihosucoe go sapwopomLMMHo zaplx one .MIa opswfim

mm 89qu

m w m m 2 ml 3 .2 S om mm vm

- . . p p F .

 

 

105

 

5Q
39

106

The observation of three orders of 00k x-ray diffrac-
tion for the M06018(0H)2.7(H20);:g+-montmorillonite indi-
cates that only limited order exists within the clay inter-
layer. This would suggest that a somewhat uneven distribu-
tion or variable hydration of the molybdenum cluster occurs
within the sample.

The intercalate can be characterized by its infrared

1

spectrum in the region 1900-250 cm- with respect to the

molybdenum cluster precipitated from HClOu solution at pH
1.0. The spectrum of this precipitate, formulated as
[MO6CI8(OH)n(H2O)(6-n)](Clou)(h—n)’ is shown in Figure
H-ha. This spectrum contains bands attributed to the

cluster cation and its coordinated water at 290—330, 785,

l

and 1620 cm' as assigned in detail below. The absorbtion

1

band centered at 1050 cm- is due to the vibrational modes

of 010; 62:1053.

The infrared spectrum of M06Cl8(0H)2 7(H20)§'§+—mont-

morillonite is illustrated in Figure H-Mb. The absorbtion

65,105

band at 290-330 cm'1 is assigned to the Mo-Cl stretch-

ing vibration of the bridging chlorides in the cluster core.

1 106

The bands at 785 and 1620 cm' are attributed , respec-

tively, to the rocking and the wagging modes of water mole-
cules present in the sample, some of which are coordinated

to the cluster core. The (Al,Si)-O deformation bands of

the montmorillonite are present at 465 and 520 cm-1. The

l

intense band at 1025 cm“ is assigned92 to the (Al,Si)—0

107

 

%Transmiflmce —~

 

 

Ién

Figure 4-“.

 

6a)" who T who ' xbo 8d) ' 6a) ' 4d9fizn
Mauumwkmfl

The infrared spectra of (a) [Mo6Cl8(0H)n—
(H20)(6-n)J(ClOH)(M-n) precipitate, (b) M06018—
(0H)2.7(H2O);:§+—montmorillonite, and (c)
Mo6C18(OH)2.7(H2O)%:g+-montmorillonite after
thermolysis at 130° (24 h), 200° (2“ h), and
280° (2“ h) under dynamic vacuum.

In sample (a), 0 < n < 3. Samples (b) and (c)

contained 53.6 mmEles—of cluster/100 g of clay.
All samples were prepared as KBr disks.

108

stretching vibrations of the montrmorillonite92. The medium

1

intensity band at 1u3o cm- is attributed to the C-H bend-

ing modes of methanol107 impurity present in the sample,
probably adsorbed on the montmorillonite interlayers. One
interesting feature of this spectrum is the band at 1700-
1670 cm-1. Presumably this is a portion of the H20 wagging

l

vibrational manifold which is shifted 65 cm” toward higher

energy through interaction with intercalated molybdenum

clusters.

1.3+_
3.3

montmorillonite in a Nujol mull is shown in Figure 4—5a.

The UV-visible spectrum of M06018(0H)2 7(H20)

It contains absorbtion bands at 320 and “05 nm attribut—
able to the presence of the metal cluster. The shoulder

at 220 nm is due to the montmorillonite. The spectrum of
the precipitated [MO6018(OH)n(H2O)(6—n)](ClOH)(M-n) mulled
in Nujol is shown in Figure H-Sb. The spectrum of (Mo6C18)-
Ola-2H20 in methanol solution is illustrated in Figure H—Sc.
This spectrum contains an intense (e 2 H000) absorbtion band
centered at 330 nm which is attributed79’108 to ligand-to-
metal charge transfer from the bridging chlorides.

The spectra in Figure “—5 are very similar to that of
Mo6C18(0H)2- formed in 0.01 E NaOH solution (Amax at 300,
330 nm) as reported by Sheldon63, with the exception that
both bands are shifted considerably toward the red. This
red shift can be interpreted by using the molecular orbital
diagram for the M6X8 cluster calculated by Cotton and

109.

Haas Protonation of the intercalated cluster increases

109

 

 

 

 

 

I.
325
E;
w
200 3b0 450 560 600 760 860 900
wmmMnM(Ww
Figure “-5. Transmission UV-visible spectra of (a) M06C18-

(0H)2.7(H20)§:§+-montmorillonite, (b) [Mo6C18—

(OH)n(H20)(6_n)](010“)(u_n) precipitate, and

(C) (MO6C18)Clu'2H20.

Sample (a) was prepared as a Nujol mull of the
intercalate containing 53.6 mmoles of cluster/
100 g of clay suspended between quartz glass
plates. Spectrum (a) was measured versus
mulled Na+-montmorillonite as a reference.
Sample (b) was prepared by evaporation of
suspension of the precipitate on a quartz
glass plate. Sample (0) was a solution in
methanol. Absorbance was measured in arbi-

trary units.

110

the positive charge residing on each metal atom. The in-
crease in positive charge contracts the metal atomic or-
bitals which decreases the overall cluster bonding. The
diminished intermetallic bonding decreases the energy of
the antibonding molecular orbitals of the cluster bringing
them closer in energy to the atomic orbitals of the
chloride ligands. Therefore the ligand-to-metal charge
transfer energy is lowered resulting in the red shifting

of the UV—visible absorbtion bands.

 

C. The Interlayering of Montmorillonite with Oxides of

Molybdenum

 

The work described above establishes that molybdenum
cluster cations can be intercalated into montmorillonite.
The real interest in this material lies in its use as a
precursor to a molybdenum oxide pillared clay, however.
This application requires the in situ conversion of the
cluster chloride to molybdenum oxide aggregates. A reac—
tion scheme to carry out this conversion was designed to
' utilize in vaggg calcinations similar to the niobium and
tantalum oxide pillaring process described in Chapter IV.

Early experimental results indicated that amodifica-
tion of the preliminary drying step as used for niobium
and tantalum cluster was necessary, however. The initial
thermolysis at 130° for 24 h was supplemented by an

additional 2“ h at 200°. This time period at the higher

111

temperature was designed to more thoroughly eliminate the
excess amounts of interlayer water present in the samples.
The decomposition of the cluster is promoted by calcina—
tion of the intercalate at 280° for 24 h in vagug. This
decomposition is accompanied by a change in the sample
color from brilliant yellow to tan or light blue. If
this calcination is performed under flowing nitrogen an
acidic reaction product, presumably hydrochloric acid,
is detectable by wet litmus paper placed in the effluent
gas stream. In addition, no detectable (AgN03) chloride
was present in a sample of Mo6C18(0H)2.7(H20)%:§+-mont-
morillonite which had been calcined ig_vagug at 280° and
then digested in H280“.

The molybdenum cluster is clearly decomposing during
the final thermolysis step. The infrared spectrum of
the molybdenum cluster intercalate shown in Figure 4-40

1

lacks the broad band at 290-330 cm‘ attributed to the

metal cluster core. The appearance of a band at 900 cm-1

is attributed110

to the Mo=0 stretching frequency of
molybdenum oxides formed upon cluster decomposition.

The appearance of a blue tint in some samples is ac-
companied by the presence of an extremely broad absorb-
tion band centered at approximately 730 nm in the UV-
visible spectrum. The observation of this absorbtion

112

band indicates that an iSOpoly molybdate known as a

molybdenum blue has been formed. These blue oxides

112

contain molybdenum atoms possessing an average oxidation
state between 5.2 and 6.052. Generally these compounds
are anionic and as such, their formation is accompanied

by collapse of the clay interlayer (d equals 9.8 X).

001
The formation of the blue color seems to indicate that

the sample was not thoroughly dried in the preliminary
thermolysis steps.

The formation of a tan color upon calcination is in-
dicative of a sufficient reduction of the interlayer
water prior to the calcination and generally accompanies
the formation of interlayer oxide formation. Unless
specified otherwise the remainder of this discussion sec-
tion refers to samples that were tan in color after cal-
cination at 200°.

The x-ray diffraction pattern of a sample treated at
280° following the optimized thermolysis sequence is shown
in Figure 4—6. The diffraction pattern shows that this
material possesses a range of d001 spacings from 9.8 fi
up to 13.2 fi. This diffraction pattern may be the re-
sult of interstratification, in which a variable inter-
layer spacing is produced by a non-uniform distribution
of interlayer molybdenum oxide aggregates. Alternatively,
this diffraction pattern indicates the formation of a bi-
phasic system consisting of collapsed montmorillonite with
an interlayer spacing of 9.8 K, and a molybdenum oxide

interlayered clay possessing a basal spacing of 13.2 fl.

113

.oofiam
mmmaw m co oopLOQQSm Eafim cfinp m mm oomwompo mm: mHoEMm one .mwao

no w ooa\pmpwsao mo mmHoEE m.mm meadpcoo maamcfiwfipo mpmHMopoucfi one
.Ae :mv comm can .An emv ooom .An :mv come he nensaoesmep seems dpficoa

nafisospsoeu+mumfiommve.mflmovmfioooz no enthuse soepenncofie ensue the

mmwmmmzuonw

o 9 Q 3 2 2 ON NN VN mm 0N

«mm
«to «on

.qmnnm

.ml: mpdwfim

on NM

114

This second interpretation is more consistent with the ob-
servation of the broad peaks present at 5.0 K and 3.2 E,
which are the (002) and (003) reflections of the collapsed
clay. The weak reflection present at 6.6 K is attributed
to the (002) reflection of the expanded phase clay.

The expanded phase is interlayered with molybdenum
oxides which prOp the silicate sheets apart with an inter-
layer Spacing of 3.6 X. This separation indicates that the
interlayer molybdenum oxide, presumably M0020H, is ag-
gregated in the form of sheet or chains slightly greater
in thickness than a sheet of oxide anions. The identifica-
tion of the interlayer species as a Mo(V) oxide, which is
brown, is consistent with both the tan color of the eXpanded
phase clays produced and with the cluster hydrolysis

products prOposed by Sheldon (vide infra). In addition,

 

the series of color changes from bright yellow to tan to
blue observed upon calcination of some pretreated molybdenum
cluster intercalates at 280° strongly suggests an increas-
ing extent of molybdenum oxidation along the series. This
places the upper boundry for the oxidation state of
molybdenum in the tan samples at approximately 5.2+, prior
to the formation of the blue oxides.

The formation of metal oxide from the metal cluster
chloride within the clay interlayer can be examined in
comparison to studies in aqueous solution. The reactions
of the molybdenum halide clusters in alkaline solution

have been studied extensively by J. c. Sheldon7l’72’ll3.

115

He established that the hydroxide was capable of displacing
the bridging chlorides, eventually leading to a decomposi-
tion of the cluster structure and the simultaneous produc-
tion of hydrogen gas. It was determined71 that the re-
placement of the bridging halides follows kinetics first
order in hydroxide concentration and first order in metal
cluster solution concentration. Thus, the replacement of
these chlorides was proposed to occur through reaction

with hydroxide ions by an 8N2 mechanism. The partial
hydrolysis of the cluster as it occurs in alkaline solu-

tion can be written:

[M06018](0H)§‘ + nOH‘ + [M0601(8_n)(0H)nJOH§' + nCl',
(“-8)

where n is equal to l, 2, or 3. Studies on the kinetics of
the decomposition of the molybdenum bromide cluster in
alkaline solution were performed in the presence of H202.
The peroxide served to eliminate the apparently autocatalytic
cluster hydrolysis term in the rate equation. Kinetics
studies indicated72 that the hydrolysis of the chloro-
molybdenum cluster is autocatalytic at elevated (70-90°)
temperatures. The products of this reaction were iden-
tified63’69 as hydrogen gas and Mo(V) oxide.

At elevated temperatures under dynamic vacuum, the

intercalated cluster undergoes an in situ hydrolytic

116

degradation and oxidation to yield interlayer molybdenum
oxide. The degradative hydrolysis occurs through reaction
with water molecules and hydroxide ions coordinated to

the cluster and with water adsorbed on the silicate inter-
lamellar surfaces. The preliminary drying steps at 130°

and 200° will limit the amount of adsorbed water present, but
provides no accurate stoichiometric control over this re—
action. In analogy to Equation (4-8) above, the inter-

layer reaction occurs through the displacement of the u3-
chlorides by hydroxide ions to form face-bridging hydroxides.
Interlayer water molecules would participate in the reac-
tion following deprotonation. The chloride displacement
would occur in a step wise manner until the metal cluster
structure is sufficiently destabilized to promote its
oxidative decomposition to the metal oxide and the simul-
taneous reduction of protons to yield hydrogen gas. The
reaction is summarized in Equation (4—9), in which the

area between the horizontal lines represents the clay inter-

layer region.

 

 

1.3+ 280°

+
Mo6Cl8(0H)2.7(H20)3.3 + 12H20 m 6M0020H + 1.3H

 

 

+ 8HCl

(4-9)

117

The formation of a pillared clay from the molybdenum
cluster intercalate by this reaction demands that the clus-
ter decomposition yield interlayer metal oxide aggregates.
This process is rather sensitive to the presence of
amounts of water in excess of the stoichiometry of Equation
(4-9) due to the tendency of molybdenum oxides to form
anions which migrate from the clay interlayer. In com-
parison, this appears to be somewhat more problematic for
the molybdenum oxides than for the niobium and tantalum
interlayer oxide species.

The best molybdenum oxide interlayered clay produced
to date, contained a mixed phase system. Presumably,
the formation of this biphasic system was due to some
variation in water content within the sample which promoted
collapse of those interlayers which were hydrated to a
greater extent. The formation of and migration by anionic
metal oxide species is an important mode of collapse during
the conversion of a metal cluster intercalate to a metal
oxide pillared clay.

Collapse of the molybdenum oxide interlayered montmoril-
lonite can occur even under the anhydrous interlayer con-
ditions existing after 24 h at 280° in vacuo. The molyb—
denum oxide interlayered clay collapses at the relatively
moderate temperature of 300°, with the oxide migrating
from the interlayer to the outer surfaces of the clay.

There is a general relationship between the bulk

properties of an anhydrous metal oxide and its ability

118

to pillar a smectite clay at elevated temperatures. This
relationship can be illustrated by comparing the melting

128 to the maximum temperature of

point of a metal oxide
stability for a clay interlayered with that oxide reported in
the literature (see Table 1-1 (p.7)). Figure 4-7 illustrates
the linear relationship that exists. A least squares fit

to these data produces the line shown which possesses a
correlation coefficient of 0.75. This is a remarkably good
fit for data gathered from many different sources and

authors utilizing various techniques.

The empirical relationship illustrated by this plot
strongly suggests that the maximum temperature of stability
of a pillared clay is related to the bulk characteristics
of the pillaring metal oxide. This relationship with the
melting point of the metal oxide further suggests that
the mode of collapse of the props may be fundamentally
similar to the process of melting of the bulk oxide al-
though occurring at substantially lower temperatures than
the actual melting point. The decomposition of the metal
oxide pillars probably involves the fragmentation of the
prOps into smaller mobile moieties which easily migrate
from the clay interlayer to deposit on the edge surfaces

of the clay layers.

119

.pac were soc emphaseamo was ms.o no unmaoeccmoo
:ofipmaonnoo < .mme can mo paw monsoon ammoa m mm cswpo was mafia one

.mofixo
Hmpme pmcp wcficfimpcoo zmao Umpmaaaa a mo opzprmQEmp omawaaoo on»

com oofixo Hmpoe w mo pcfioo mafipaoe on» cmmzpmn soapmHmphoo one

816%.: BEE Eat:

 

 

.NI: opswfim

DON

00?

. DOW

KD/Q paw/yd yo agruoxadwal asdouog

120

D. Conclusions and Recommendations

 

The molybdenum chloride cluster (Mo6C18)Clu-2H20 can
be utilized as a pillar precursor for the formation of
an expanded phase montmorillonite containing molybdenum
oxide aggregates. Modification of the metal cluster
salt to the hydrated perchlorate form can be accomplished
by removal of the terminally bound chloride anions through
argenometric precipitation. Intercalation of the hydrated
metal cluster into montmorillonite occurs by ion exchange
for interlayer sodium ions under acidic conditions. These
intercalated clusters undergo in situ hydrolytic decomposi-
tion and oxidation upon calcination at 280° under vacuum.
The proposed reaction produces a molybdenum oxide inter-
layered material possessing an interlayer spacing of up to
3.6 A due to the presence of a layer of molybdenum oxide.
This interlayered material is stable only to the relatively
moderate temperature of 300°, however. The success,
albeit somewhat limited, of the synthesis of a molybdenum
oxide interlayered clay through the reaction of a metal
cluster intercalated montmorillonite demonstrates the
general nature of this reaction and adds credence to the
results described for niobium and tantalum in Chapter
III.

A review of the maximum thermal stabilities of the
metal oxide pillared clays described in the literature,

in conjunction with the collapse temperatures for the

121

niobium, tantalum, and molybdenum oxide interlayered mont—
morillonites lead to the discovery of an interesting rela-
tionship between the melting point of a given metal oxide
and the maximum temperature of stability for the pillared
clay containing that oxide. It is strongly recommended
that additional data points be gathered and added to

Figure 4-7. With further refinements, this type of ap-
proach could lead to estimates of maximum thermal stability

expected for pillared clays prior to their synthesis.

V. MOLYBDENUM CLUSTER BINDING TO H+-Y ZEOLITE

A. Introduction

 

A relatively brief investigation of the binding of the
molybdenum cluster to a hydrogen exchange form of the syn-
thetic faujasitic zeolite Y was performed. The objective
of this study was to utilize the ion exchange chemistry of
the molybdenum cluster develOped for montmorillonite, as
a method of binding the compound to zeolites. Ion ex-
change could potentially place the molybdenum cluster
within the intracrystalline voids of these aluminosilicates.
Calcination of this exchanged zeolite could potentially
create a catalyst containing a well-defined molybdenum
oxide species within the intracrystalline channels of a
zeolite. The rigid zeolite structure might serve to sup-
port and stabilize the molybdenum oxide species without
imposing structural requirements on them as is the case
within pillared clays.

Zeolites are crystalline, hydrated aluminosilicates
of Group I or II elements in their natural forms. These
tectosilicates consist of a three-dimensional framework
of (A1,Si)0u tetrahedra linked together by the sharing

of all oxygens. This framework contains channels or

122

123

interconnected voids within the crystal, which are occupied
by cations and water molecules. The cations are present
to balance the charge on the lattice originating from sub-

stitutions of Al3+ for Siu+

ions. These intracrystalline
cations are ion exchangeable to some degree, depending on
the zeolite framework and the cations in question. The
intrachannel or "zeolitic" water can be reversibly ad-

6

sorbed. However, high vacuum (mlo' torr) and elevated
temperatures (N350°C) are generally necessary to totally
dehydrate the crystal.

There are 34 naturally occurring zeolites and about
100 synthetic types. The synthetic zeolite Y possesses a
crystalline framework similar to the structure of faujasite,
with a silica-to-alumina ratio of 1.5 to 3.0. The struc-
ture of faujasite, illustrated in Figure 5-1, consists
of (Al,Si)Ou tetrahedra assembled into truncated octahedra
known as B-cages. These cages are linked together through
their hexagonal sides to form prisms of double-6-rings.
A void approximately 13 A in diameter, known as the super-
cage, is created within the assembly. Access to the super-
cage is available through a 12-ring opening, shoWn in the
foreground in Figure 5-1, which is 7.4 A in diameter as
determined crystallographically. Molecules as large as
(CuH9)3N, which has a kinetic diameter of 8.1 A, are per-

mitted through this openingllu.

Figure 5-1.

 

124

 

A schematic diagram of the structure of
faujasite.

The verticies indicate the location of
(Al,Si)0h tetrahedra, the midpoints of the
lines indicate the positions of shared oxygens.
The truncated octahedra, known as B-cages, are
linked together through double-6-ring prisms.
A large cavity, known as the supercage, is
thus formed. The supercage is visible in this
diagram through the 7.4 A, l2-ring pore open-
ing in the foreground. Diagram taken from
Reference 11b, p. 168.

125

Recently, there has been a great deal of interest in

placing transition metals, in highly dispersed formslls,

and as aggregates or clustersll6, within the pore structure
of zeolites. It has been suggested117 that intracrystalline
metallic species would possess exceptional catalytic
activity. This activity would be as a result of the high
state of dispersion and the formation of unique metal assem-
bleges made possible through interaction with the zeolite
matrix. The zeolite framework may also impart a size
selectivity to a catalytic reaction based on the relative
rates of diffusion of product or reactant molecules through
the crystal, or by limiting the size of the transition

state complex of the reaction. The Juxtapositioning of
acidic sites on the aluminosilicate and the intracrystalline
metallic catalytic sites may also impart an important bi-

functionality to the catalyst. Rabo _e_t_ 21118 119

and Weisz
were among the first to demonstrate that reduced transition
metal loaded zeolites possessed great promise as catalysts
in petroleum refining. For example, after hydrogen reduc-
tion at 450°, 0.5 wt % Pt dispersed within a Ca-Y zeolite
exhibits high stability to sulfur poisoning in its use as

a reforming catalystll7.

This catalyst actively demon-
strates the size selectivity and bifunctionality expected
of an active metal dispersed within a zeolite.

Recently, much work has dealt with the formation of metal
clusters or metal aggregates within the zeolite intra-

120

crystalline channels The formation of metal clusters

126

and bimetallic species such as Ru-Cu and Ru-Ni indicates
that controlled metallic aggregation can yield extremely
high dispersions of metals and alloys supported on zeo-

120

lites The maintenance of high dispersion is a func-

tion of the preliminary treatmentll6a. It is extremely
difficult, however, to ascertain the exact location of the
metal aggregates; whether they are deposited on the ex-
ternal zeolite surfaces or within the intracrystalline
voids.

Metals and metal clusters can be deposited within zeo-
lites by adsorbtion of the vapor of organometallic or vola-

tile inorganic compounds. The metal carbonyls of M0121,

122 117

Fe , Re, Rul23, Ni, Cr, w, and Mn and the metal

122
cluster carbonyls Fe2(C0)9, Fe3(C0)l2

and Re2(C0)lO,
Ru3(C0)12123 have been adsorbed within zeolites by vapor
deposition. The carbonyl ligands are easily removed by
thermal decomposition to yield active metal species.
Recent work by Lundsfordlzu has utilized the vapor ad-
sorbtion of MoOClu, followed by the decomposition of the
chloride at 400°, to place molybdenum oxides within H+-Y
and ultrastable Y zeolites. These materials were shown to
be active for the selective oxidation of cyclohexene to
the epoxide within a slurry reactor. Similarly, work by
wilhelm125, has utilized the vapor deposition of volatile
molybdenum oxychlorides, formed by reaction of MoO3 with

anhydrous HCl, or of MoO2 with 012, to place 3 to 10 wt %

127

molybdenum loadings on mordenite zeolites. These molyb-
denum containing zeolites are viewed in contrast to com-
mercially available catalysts which are prepared by co-
extrusion or impregnation techniques that deposit the metal
exclusively on the external surfaces of the zeolite.

The ion exchange of molybdenum compounds into zeolites
could potentially produce a more even metal distribution,
and thus, higher catalytic activity, than the vaporization
techniques mentioned above. There are, however, few
cationic forms of molybdenum available for ion exchange,
and these are formed only in strongly acidic environments,
well beyond the acid stability of large pore zeolites.
Therefore, previous attempts at ion exchange of molybdenum
compounds into zeolites have produced poorly crystalline,

low surface area productsl25.

B. Binding_of the Molybdenum Cluster byH+-Y Zeolite

It has been demonstrated in Chapter IV, that the molyb-
denum cluster exhibits the ability to ion exchange into
Na+-montmorillonites at low pH. The molybdenum inter—
calate thus formed, also possessed the desired ability
to undergo the thermochemical transformation to yield
interlayer molybdenum oxide. It is possible to envisage
a similar kind of pH dependent ion exchange into the
pores of a Y type zeolite, and the subsequent decomposi-

tion of the cluster to yield molybdenum oxide trapped

128

within the supercages.

The ion exchange of the protonated molybdenum cluster
into Na+-Y, in direct analogy to the reaction with Na+-
montmorillonite is not possible due to the acidic condi-
tions (pH 1.5) of that reaction. FauJasitic zeolites
possess only limited acid stability, suffering dealumina-
tion and loss of crystallinity at a pH of 3.0 (HClOu; 24 h)
or lower. Therefore, a modified approach was used.

The reaction of M06C18(0H)n(H20)(éfl;r)l)+ from a stock
solution with the hydrogen exchange form of zeolite Y
utilized the zeolite as a source of acid. This reaction
was performed at 70°, with a dropwise (1.5 mL/min) addition
of 0.1 fl HClOu to adjust the solution pH to 4.0. At this
pH and temperature little loss of zeolite crystallinity

was observed (vide infra).

 

Approximately 70% of the molybdenum cluster present in
solution was bound to the zeolite, producing a molybdenum
loading of 2.27 wt% as shown by chemical analysis. In
contrast, a Na+-Y bound only about 12% of the molybdenum
cluster, to produce a loading of 0.39 wt % molybdenum,
under identical conditions. Little, if any molybdenum
was noticeably bound to A1203 or 8102 samples under
similar conditions. No chemical analyses were performed
on these samples, however.

Comparison of the chemical analyses for H+-Y and

Na+-Y indicates that the binding of molybdenum is

129

strongly dependent on the cation exchange form of the zeo-
lite used. This suggests that the molybdenum cluster is
not simply physically adsorbed on the zeolite under reac-
tion conditions. The strong binding of the cluster to

the H+-Y zeolite is probably a result of the acid/base
chemistry of Mo6C18(0H)u(OH2)2 present in solution at pH
4.0. The protonation of the cluster in contact with the
zeolite will result in the electrostatic binding reaction

as shown in Equation (5-1),

§ i

+ \ / + \ ./
MO6018(OH)4(H2O)2 + H /Al\ + Mo6Cl8(0H)3(H20)3 /Al\

§ 1.1)?

where the serpentine lines represent the zeolite alumino-
silicate framework. The protons in a hydrogen form of the
zeolite would be present either as exchangeable cations
bound electrostatically near the aluminum sites or as
hydroxyl groups on silica tetrahedra adjacent to the
aluminum ionslz6. The former case would predominate in
the hydrated zeolite.

The binding of the molybdenum cluster to H+-Y presum-
ably involves protonation and interpenetration of the
metal cluster into the zeolite. The molybdenum cluster

could then occupy the large pores and supercages only,

being too large to enter other cation binding sites. The

130

molybdenum cluster can be viewed as a cubic assembly of
chloride ions that is 7.2 A on a side and 8.6 A on the
facial diagonal. The penetration of this large species
through the 7.4 A, l2-ring pore Opening would prove to
be a difficult step. Experimental observations indicate
that the reaction of the molybdenum cluster with H+-Y at
70° yields a product that is more homogeneous in appearance
than the product of reaction at room temperature, all other
factors being equal. This would be consistent with an
ion exchange reaction of a large cation which requires
high momentum in order to overcome the barrier to penetra-
tion that the tight fit of the pore opening presents.
Assuming that intracrystalline penetration and ion ex-
change has occurred, the cluster cation would be limited to
occupancy of the supercage and the large pores. Similar
restrictions are found for other cations. For example,
the exchange of (C2H5)3NH+, which has a diameter of ap-
proximately 6.9 3, into a Na+-Y is limited to replacement
of only 26% of the sodium ionsllu. The alkyl ammonium
cation is too large to pass through the 2.2 A pore openings
into the B-cage, and thus cannot replace the cations in
these positions. In addition, the ion exchange is limited
by the volume of the cation (168 A3) and the volume avail-

3 11“. Sim-

able in the supercage, approximately 6700 A
ilarly, the maximum loading of the molybdenum cluster
would be severely limited by its diameter and volume of

occupancy (370 A3).

131

Several experiments were performed to test the pos-
sibility of carrying out a cluster decomposition to the
oxide within the zeolite. The quantity of water within
a zeolite after drying under ambient conditions is far
greater than that within a clay intercalate. Therefore,
it was more practical to carry this water away from the
zeolite under flowing nitrogen or oxygen at 200° for 2 h,
rather than under vacuum. The subsequent calcination of
the sample at 430° for 5 h under either gas results in
its color change from yellow to blue (m340°) to white,
suggesting that the cluster structure has decomposed com-
pletely to produce M003. The appearance of the blue
coloration at the intermediate temperature suggests that
the metal cluster has undergone an hydrolysis and oxidation
reaction to yield a molybdenum isopoly blue compound,
similar to that described in Chapter IV. This probably
indicates that the molybdenum oxide has migrated to the
external surfaces of the zeolite. Chemical analysis of
this material indicates that only traces of chloride re-
main following this treatment, indicating that the cluster
has decomposed. The molybdenum is present in essentially
the original quantities as indicated by the anhydrous

unit cell formula (M00 -[(A102)

+
3)3.06’ H53.97 53.97,

(SiO The presence of the full amount of

2)l38.03]'
molybdenum indicates that little is volatilized off the
zeolite in the form of oxides or oxychlorides during

reaction at 430°.

132

c. Physical Characterization of [M06918(0H)3LQH2)3(QH2)3lgou7—

 

+
£53.50'Y Zeolite

 

The reaction at pH 4.0 and 70° between H+—Y and
MO6018(OH)4(OH2)2 in a ratio of 0.67 moles of cluster per
zeolite unit cell produced a light yellow solid. This
stoichiometry was used because it would result in a load-
ing of 5.0 wt % M06C18, assuming quantitative uptake by
the zeolite. Elemental analysis corresponded to an an-
hydrous unit cell formula of [Mo6C18(0H)3(0H2)3];.u7,
Hg3o50[(A102)53.97, (Si02)l38.03]. This, of course, is a
bulk analysis and tells nothing about the actual interrela-
tionship between the molybdenum cluster and the silicate.

As stated above, one of the major difficulties in plac-
ing molybdenum within zeolites by ion exchange is the loss
of crystallinity and surface area by the tectosilicate under
acidic conditions. The zeolite Y used in this study main-
tained excellent crystallinity down to a pH of 4.0 (HClOu)
as judged by x-ray diffraction peak intensities. Excellent
zeolite crystallinity is also maintained in the presence
of the molybdenum cluster at pH 4.0. Figure 5-2 shows
the x-ray diffractograms of freshly prepared H+-Y,
[M06018(OH)3(OH2)338.47’ Hg3.50'[(A102)53.97’ (8102)l38.o3]
air dried, and after a predrying step at 200° (2 h) and
calcination at 430° (5 h) under flowing oxygen. In each
case the x-ray diffraction pattern contains narrow lines

of high intensity. The assignment of Miller crystallographic

133

.mconocmaa ESCHESHM oucfi ooxomo who; scan: mpmozoa app mm condo
mmo mmafimofimv .Nm mmANOH<Vg ou Amado ma N

load mam: moHoEmm lum.mm
.mezxo wcfizoam noon: 3 m pom Doom: pm coaumcfloamo popmm wuom mmm

.ez.mmonmmvamovaomozg on one .oofino new maom.mmm .e:.mmmflommv
umfimovwaowozg ADV awl+m Amv no manoppmo cofipomhmmap mmozoa zmnlx .mlm onswfim

mum opswflm

 

134

 

 

 

 

 

 

 

 

 

 

 

mmmmSflu
m w w m on w.‘ J. M. ms 8 - E .3 8 2. mm. 3. R. 2.
«:2
6%.; 369
:5 But 6'9
l
3. 6?! :65
D and
3.3 $9
3:
.8!
Q
2.2.. 6»! a3:
2R,»
3: 5.?
33
U 3-~ «:9 3.
6v! ax:
:5

«sag

349

135
indicies is based on assignments by Breckllu, and are
identical for each sample as shown in Table 5—1. Zeolite
crystallinity can be judged by the relative intensities
of the (331), (440), (533), (642) and (751) diffraction

peaksl26.

These are changed very little during the re-
action with the molybdenum cluster, and the subsequent oxi-
dation of the material at 430°. Of interest, is the in-
crease of the intensity of the (200) and (311) diffraction
peaks relative to the (331) peak in the H+-Y following
reaction with molybdenum cluster. This increase may

simply be a result of rehydration of the zeolite lattice
during the exchange reaction. The (220) and (311) dif-
fractions of hydrated Na+-Y are of intensities similar

to those of the [Mo6C18(OH)3(0H Y.

2)338.u7’ H83.50“
The presence of exchange cations of different sizes will
shift the unit cell dimensions of the zeolite slightly.
For example, a hydrated zeolite Y possessing a 8102/A102
ratio of 2.43 will exhibit a unit cell dimension varying
from 24.85 to 24.61 A as the exchange cation is varied in
size from tetramethylammonium ion to the H+ ionllu.
Even though the H+-Y zeolite is loaded with approxi—
mately 0.47 molybdenum clusters per unit cell, there is
no change in the unit cell parameters detectable within
the accuracy of the Siemens x-ray diffractometer used for
these measurements. The lack of a change in unit cell

dimensions does not support the theory of ion exchange of

these clusters into the zeolite, but may simply indicate

136

Table 5-1. X-ray Powder Diffraction Data and Miller In-

 

 

 

 

dices for Na+-Yb, H+-Y, and Molybdenum Cluster,
H+—Y.

Hydrated [Mo6018(0H)3(0H2)310.u7
Na+-Yb H+-Y Hg3.50-Y
111:1. L71). an dill .1111 «1121..
111 14.29 111 14.24 111 14.24
220 8.75 220 8.75 220 8.71
311 7.46 311 7.43 311 7.49
333 4.76 333 14-7LI 333 “.7“
440 4.38 440 4.35 440 4.35
620 3.91 620 3.90 620 3.88
533 3.775 533 3.77 533 3.74
711 3.466 711 3.45 711 3.45
642 3.308 642 3.30 642 3.28

731 3.222 731 3.22 731

733 3.024 733 3.02 733 3-00

822 2.917 822 822 2.89

751 2.858 751 2.86 751 2.84

840 2.767 840 2.76 840

911 2.717 911 911

664 2.638 664 2.63 664 2.63
a = 24.73 X a = 24.74 A a = 24.74 A
Si/Al = 2.00 Si/Al = 2.56 Si/Al = 2.56

b

Data taken from Reference 114, p. 369.

137

Table 5-1 (cont'd).

 

 

+
[MO6Cl8(OH)3(OH2)3]0.47’ H53.50‘Y
Calcined 5h at 430°C Under 0

 

2

.051 1.11
111 14.02
220 8.58
311 7.37
331 5.60
333 4.69
440 4.33
620 3.86
533 3.74
711 3.42
642 3.26
731

733 3.00
822 2.88
751 2.82
840 2.74
911 2.67
664 2.60
844 2.56

 

 

24.40 K Si/Al = 2.56

a

138

that there is ample volume within the supercage to contain
the metal cluster ion plus the rest of the hydrated ex-
changeable protons at the cluster loadings used.

Infrared spectroscopy may also be used to investigate
the crystallinity of the zeolite. Work by Flanigen127
has assigned vibrational bands to the structural com-
ponents of the zeolites. Vibrational modes internal to
the (A1,Si)0u tetrahedra are observed at 1250-950, 750-

650 and 500-420 cm‘1

, assigned to the asymmetric and
symmetric stretching and (Al,Si)-O bending modes, respec-
tively. The structure sensitive bands associated with
the crystallinity of the zeolite are attributed to the
vibrational modes external to the tetrahedra. Infrared
bands attributed to vibrational modes of double ring
prisms, pore openings, and the symmetric and asymmetric
stretching of the intracrystalline channels of the zeolite
are observed at 650-500, 420-300, 820-750 and 1050-1150
cm'l, respectively.

Figure 5-3 shows the infrared spectra from 1400 to
250 cm”1 for H+-Y, [Mo6C18(OH)3(OH2)3]S.u7, Hg3.50-Y, and
[Mo6018(0H)3(0H2)3];.u7, Hg3.50-Y after calcination at
430° under oxygen for 5 h. Only very slight differences
between these spectra are noticeable. The bands assigned
to the external lattice modes of the zeolite are strongly
present in all three samples, therefore, the observation

of high crystallinity in these samples by x—ray diffrac-

tion is corroborated by infrared spectroscopy. The

139

 

 

 

 

 

 

11 I "I ‘ ' I
a
I,
U
E
;g
E
8
o
\
7am who ' 0&7 ' db 7' 6&9 ' 4&7 77:0
W0 venurnber (ch9
Figure 5-3. Infrared spectra of (a) H+-Y, (b) [MO6Cl8'

+ +
(OH)3(H2O)330.Q7, H53.50-Y, and (C) [MO6018-
(OH)3(H2O)3]B 47’ Hg3.50-Y after calcination
at 430°C for 5 h under flowing oxygen.

The scale along the top of the diagram indi-
cates the regions assigned to the vibrational
modes internal to the (Al,Si)Ou tetrahedra(I),
and to the structure sensitive vibrations which
are external to the tetrahedra (E). The small
arrows indicate the peaks which indicate the
high degree of crystallinity in these samples.
The band at 900 cm‘1 in (c) is assigned to the
Mo=0 stretching vibration of molybdenum oxide.

 

140

broad Mo-Cl vibrational band of the molybdenum cluster
(340-290 cm'1)65’105 is not observed in Figure 5-3b. This
weak band is obscured by the strong absorbtions of the
zeolite matrix. After calcination the appearance of the

1

peak at 900 cm” indicates the presence of molybdenum

oxide. This band is assigned to the Mo=0 stretching vibra-
t10n110,124.

The observation of perturbations of the external vibra-
tional modes of the zeolite could be indicative of the
presence of intracrystalline molybdenum cluster. Conclusions
based on this type of observation must be made carefully,
utilizing data gathered from a spectrOphotometer of much
higher resolving power than was used for this study,
however.

Nitrogen B.E.T. surface area measurements were carried
out on samples of H+-Y and [Mo6018(0H)3(0H2)3];.u7,

H53.50'Y' Both samples possessed surface areas of ap-
proximately 410 m2/g. The retention of full surface area
during the exchange of the cluster is consistent with the
high crystallinity of the sample. This measurement also
indicates that the cluster is not clogging the pore open-
ings of the zeolite as might be expected from a precipita-
tion onto the external surfaces of the aluminosilicate
crystallites.

UV-visible spectrophotometry was performed using a sample
of [M06C18(OH)3(OH2)3]O.47, HEB 50'Y mulled in Nujol and

supported on quartz glass plates. The spectrum was taken

141

with a Na+-Y mulled in a similar manner in the reference
beam. The spectrum contained merely a broad tailing into
the UV associated with the zeolite. No bands attributable
to the molybdenum cluster were observed in the spectrum.

Electron microprobe analysis was performed on two
regions of a particle that was crystalline in appearance
under the magnification of the electron microscope. This
particle had apparently been cleaved by ultramicrotoming
the sample which was embedded in epoxy resin. This analysis
yielded the plots shown in Figure 5-4. Region 1 (5-4b)
was on the surface of the particle, and contained substantial
amounts of molybdenum and chlorine as indicated by the
number of counts in these regions. Region 2 (5-4a) was on
a cleaved section of the particle which should have been
representative of the internal portions of the zeolite
crystal. These analysis were carried out using identical
instrumental settings and acquisition times.

The analysis plotted in Figure 5-4a suggests that sig-
nificant amounts of molybdenum and chlorine are present
within the zeolite. This microprobe analysis indicates
that a generally even distribution of the cluster is
present and is strong evidence supporting the protonation
and ion exchange of the molybdenum cluster into the pores
of the zeolite as the method of cluster binding. Much care
must be taken in the interpretation of these results and

corroboration should be sought.

Figure 5—4.

142

Electron microprobe analysis of [Mo6C18-
+ +

(OH)3(OH2)3]O 47’ H53.50-Y embedded in

epoxy resin and cleaved by an ultramicro—

tome.

The spot analysis on region 2 which was a
cleaved portion of the particle is shown

in (a). The spot analysis of an exterior
portion of the particle (region 1) is shown
in (b). Instrumental settings and acquisi—
tion times were identical for both measure—
ments.

143

 

 

. 3 Si:
fA/Z

m ° 3
"E .

. - Mo
8 : :

._ AI ' f s;

- Mo
1. .- _.
u .3 '.

 

 

 

IJQD Lam' ' Yzialv

Energy (KeV)
Figure 5-4

144

D. Conclusions and Recommendations

 

Several aspects of this study indicate that further re-
search should be carried out to investigate the molybdenum
cluster binding to H+—Y zeolite. The binding of rather sub-
stantial amounts of the cluster to zeolite appears to be
strongly dependent on the ion exchange form of the zeolite.
Thus, the strong reaction of the cluster appears to be a
function of the acidity of the intracrystalline regions of
the zeolite. It can be estimated that the interior of
hydrated protonic zeolites is approximately equivalent to

114

4.0 M acid This strongly suggests the possibility of

an ion exchange reaction based on pH, as is the case for
the exchange reaction of Mo6C18(OH)n(OH2)(éE;?)+ into
montmorillonite.

In comparison to other methods, this ion exchange tech—
nique has several distinct advantages. This work has been
shown that loading of molybdenum onto the zeolite can be
accomplished with the retention of excellent crystallinity.
The binding of a cluster species to the zeolites allows for
the formation of well defined metal species through the
decomposition of the cluster structure. Although little
corroborating evidence exists, the microprobe analysis
suggests that the molybdenum cluster is located within the
zeolite pore structure. If this is indeed correct, this

would represent the first time a molybdenum species,

especially a metal cluster species, has been ion exchanged

145

into a crystalline zeolite. Clearly this claim should be

verified by additional studies, and corroborating evidence.
The calcination of [Mo6018(0H)3(0H2)3];.u7, Hg3.50—Y

results in the formation of molybdenum oxide supported on

the zeolite, as shown by infrared spectroscopy. The ob-

servation of a light blue color during calcination sug-

gests112 the formation of molybdenum blue compounds prior

to formation of the trioxide. It is most likely that the

anionic isopoly blue compounds would have migrated from the

pores of the zeolite. The presence of intracrystalline water

would most certainly contribute to this process and should

be limited by a preliminary treatment and reaction under

6

high vacuum (mlO- torr). Experiments with reduction

of Pt2+-exchanged Y zeolites indicates that the size of
the platinum metal aggregates formed after reduction at
300° is a function of pretreatment temperature at 10"6
torr. This is apparently due to the presence of intra-
crystalline water which promotes platinum metal migra-
tionllg.

It is recommended that future work on this project be
persued. The initial goal should be to more firmly es-
tablish the ion exchange behavior of the molybdenum clus-
ter into the H+-Y zeolite. The ability of the cluster
to decompose to molybdenum oxides upon calcination is
quite well established, however, this reaction must occur

within the zeolite framework to make full use of the

properties of the zeolite during catalytic application.

146

Many avenues of catalytic research would be opened if this
in situ reaction was established. Reactions such as size
selective oxidation, hydrocracking studies, isomerizations
or alkylations may be possible. These first generation
catalysts could be readily modified by promoters which could

be exchanged into the zeolite prior to the molybdenum cluster

exchange.

REFERENCES

11.

12.

REFERENCES

J. Schmidt, W. Hafner, R. Jira, J. Sedlmeier, R.
Sieber, R. Ruttiger, and H. Kojer, Angew. Chem.,

1959, 7i, 176.

G. W. Parshall, J. Molec. Catal., 1978, m

G. Szonyi, Advan. Chem. Ser., 1968, 10, 53.

 

J:
\o

243.

 

 

D. Evans, J. A. Osborn, and G. Wilkinson, J. Chem.
Soc. (A), 1968, 3133.

C. W. Bird, Chem. Rev., 1962, 62, 283.

 

 

A. J. Chack and R. F. Harrod, Adv. Organometallic
Chem., 1968, 6, 119.

F. E. Paulik and J. F. Roth, U.S. Patent 3,769,329
(1973).

M. Heylin, C. E. News, 1982, 69 (24), 31.

 

 

J. L. Russell, U.S. Patent 3,418,340 (1968).

(a) B. C. Gates, J. R. Katzer, and G. C. A. Schuit,
"Chemistry of Catalytic Processes", McGraw-Hill, Inc.,
New York (1979), pp. 1.

(b) P. B. Venuto and E. T. Habib, Catal. Rev., 1978,
i8 1. ——
"b,

(a) L. F. Hatch, Hydrocarbon Proc., 1969, 58, 77.

 

 

(b) C. N. Satterfield, "Heterogeneous Catalysis in
Practice", McGraw-Hill, Inc., New York (1980), pp. 258.

(c) C. L. Thomas, "Catalytic Processes and Proven
Catalysts", Academic Press, Inc., New York (1970),
pp. 146.

 

(a) A. P. Kreuding, AIChe Symp. Serv., 1972, 70, 138.

Eb) S. C. Shuman and H. Shalit, Catal. Rev., 1970,
245.
m,

 

147

12.

l3.

14.

15.

16.

17.

18.

19.

20.
21.

22.

23.

24.

25.

26.

148

(c) G. S. A. Schuit and B. C. Gates, AIChe J., 1973,
19 417.
3

F. G. Ciapetta, R. M. Dobres, and D. N. Wallace,
Catal. Rev., 1972, 5, 67.

 

J. J. Burton and R. L. Garten, "Advanced Materials
in Catalysis", Academic Press, Inc., New York (1977),
pp. 14.

J. B. Dixon and S. B. Weed, "Minerals in Soil Environ-
ments", Soil Science Society of America, Madison,
Wisconson (1977), pp. 294.

R. E. Grim, "Clay Minerology", 2nd Ed., McGraw-Hill,
Inc., New York (1968), pp. 77-92.

(a) H. Herbst, Erd81 und Teer, 1926, g, 265 and 411.

 

(b) R. C. Hansford, Advan. Catal., 1952, 4, l.

(c) H. H. Vogue in "Catalysis", P. H. Emmett (Ed.),
Reinhold, New York, 1958, 6, 407.

 

H. E. Swift in "Advanced Materials in Catalysis",
J. J. Burton and R. L. Garton (Eds.), Academic Press,
Inc., New York (1977), pp. 209.

B. K. G. Theng, "The Chemistry of Clay—Organic Re-
actions", John Wiley and Sons, Inc., New York (1974),

pp. 269.

R. C. Davidson, F. J. Ewing, and R. S. Shute, Natl.
Pet. News, 1943, 35, 318.

E. A. Zuech, U.S. Patent 3,849,507 (1974).
(a) T. A. Pinnavaia, Science, 1989, 220, 365.

(b) T. J. Pinnavaia, ACS Symp. Ser., 1982, 192, 241.

 

W. H. Quale and T. J. Pinnavaia, Inorg. Chem., 1979,
i8, 2840.

 

P. K. Welty and T. J. Pinnavaia, J. Amer. Chem. Soc.,
1975, 97, 3819.

T. J. Pinnavaia, R. Raythatha, J. G. S. Lee, L. J.
Halloran and J. F. Hoffman, J. Amer. Chem. Soc., 1979,

18%, 6891.

R. Raythatha and T. J. Pinnavaia, J. Organomet. Chem.,
1981, 218, 115.

 

 

 

27.
28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

149

H. M. Chang, Thesis, Michigan State University (1982).

F. Farzaneh and T. J. Pinnavaia, Inorg. Chem., 1983,
2g 2216. ‘“‘“
9

A. Kadkhodayan and T. J. Pinnavaia, J. Molec. Cat.,
in press.

 

 

J. M. Thomas, J. M. Adams, S. H. Graham, and D. T.
B. Tennakoon, Adv. Chem. Ser., 1977, $63, 298.

 

J. M. Adams, J. A. Ballantine, S. H. Graham, R. J.
Laub, J. H. Purnell, P. T. Reid, W. Y. M. Shaman and
J. M. Thomas, J. Catal., 1979, g9, 238.

J. A. Ballantine, M. Davies, H. Purnell, M. Rayanakorn,
J. M. Thomas and K. J. Williams, J. Chem. Soc. Chem.
Commun., 1981, 421.

 

J. A. Ballantine, R. P. Galvin, R. M. O'Neil, H.
Purnell, M. Rayanakorn, and J. M. Thomas, Ibid,

1981, 695.

J. A. Ballantine, H. Purnell, M. Rayanakorn, J. M.
Thomas, and K. J. Williams, Ibid, 1981, 9.

J. M. Adams, A. Bylina, and S. H. Graham, J. Catal.,
1982, 15, 1980.

R. M. Barrer, "Zeolites and Clay Minerals as Sorbents
and Molecular Sieves", Academic Press, Inc., New
York (1978), p. 407-483.

R. M. Barrer and D. M. Macleod, Tran. Faraday Soc.,

1955, 51, 1290.

J. Shabtai, N. Frydman, and R. Lazar, Proc. 6th Int.
Congr. Catal., B5, 1 (1976).

 

 

 

M. M. Mortland and V. E. Berkheiser, Clays Clay Miner.,
1976, 94, 60.

R. H. Loeppert, M. M. Mortland, and T. J. Pinnavaia,
Ibid., 1979, 27, 201.

(a) T. Endo, M. M. Mortland, and T. J. Pinnavaia,
Ibid., 1980, 28, 105.

(b) T. Endo, M. M. Mortland, and T. J. Pinnavaia,
Ibid., 1981, g9, 153.

(c) T. J. Pinnavaia, M. M. Mortland, and T. Endo,
U.S. Patent 4,367,163 (1983).

 

 

 

42.

43.

44.

45.

46.
47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

150

G. W. Brindley and R. E. Semples, Clay Miner., l 77,

i2, 229. '““‘

S. Yamanaka and G. W. Brindley, Clays Clay Miner.,
1979, 27, 119.

R. J. Lussier, J. S. Magee, and D. E. W. Vaughan,
Preprints, 7th Canadian Symp. on Catalysis, Vol. 1,
30th Canadian Engineering Conf., Edmonton, Alberta,
Oct. 19-22, 1980, p. 88-95.

 

 

 

 

J. Shabtai, R. Lazar, and A. G. Oblad, "New Horizons
in Catalysis", 1981, 7, 828.

J. Shabtai, U.S. Patent 4,238,364 (1980).

N. Lahave and U. Shani, Clays Clay Miner., 1978, 26,
116.

 

G. W. Brindley and S. Yamanaka, Amer. Miner., 1979,
93, 830. "“'
S. Yamanaka and G. Yamashita, Clays Clay Miner., 1978,
26 21
mm’ °

 

 

 

S. Yamanaka, G. Yamashita and M. Hattori, Ibid.,
1980, 88’ 281. -——-'

(a) F. A. Cotton, Quart. Rev. Chem. Soc., 1966, 20,
389. __._..__.____ mm

(b) R. B. King, Progress Inorg. Chem., 1972, i5, 287

 

 

 

(c) H. Vahrenkamp, Struct. Bond., 1977, 32, 1.

L. Pauling, "Nature of the Chemical Bond", 4rd. Ed.,
Cornell University Press (1960), pp. 440.

 

C. B. Thaxon and R. A. Jacobson, Inorg. Chem., 1971,
19, 1460. ‘

B. G. Hughes, J. L. Meyer, P. B. Fleming and R. E.
McCarley, Ibid., 1970, 9, 1343.

J. H. Espenson and R. E. McCarley, J. Amer. Chem. Soc.,
196 , 99, 1063.

 

J. G. Converse and A. J. Thomson, J. Chem. Soc.
Dalton, 1972, 2350.

R. A. Mackay and R. F. Schneider, Inorg. Chem., 1967,
6 549.
m,

 

 

59.

60.

61.

62.
63.
64.

65.

66.

67.
68.

69.
70.
71.
72.
73.

74.

151

(a) D. D. Klendworth and R. A. Walton, Ibid., 1981,
£9, 1151.

 

(b) R. A. Field and D. L. Kepert, J. Less Common Met.,

 

1161 13. 378.

(c) R. J. Allen and J. C. Sheldon, Aust, J. Chem.,
1965, 19, 277.

 

 

 

 

N. Brnicevic and H. Schafer, Z. Allg. Chem., 1978,
in. 219-
(a) C. Brossett, Arkiv, Kemi. Min. Geoc., 1946, 999,

 

 

No. 7.
(b) C. Brossett, Ibid., 1947, 999, No.

R. D. Hogue and R. E. McCarley, Inorg. Chem. , 1970,
9,1354.

F. A. Cotton and N. F. Curtis, Ibid., 1965, 9, 241.

 

J. C. Sheldon, Nature, 1959, $99, 1210.

W. M. Carmichael and D. A. Edwards, J. Inorg. Nucl.
Chem., 1967, 99, 1535.

F. A. Cotton, R. M. Wing, and R. A. Zimmerman, Inorg.
Chem., 1967, 9, 11.

A. D. Hamer, T. J. Smith, and R. A. Walton, Ibid.
196 ,ié, 1014.

P. Nannelli and B. P. Block, Ibid., 1968, 7, 2423.

 

 

D. L. Kepert, R. E. Marshall, and D. Taylor, J. Chem.
Soc. Dalton, 1974, 506.

 

J. C. Sheldon, J. Chem. Soc., 1960, 1007.

 

J. C. Sheldon, Ibid., 1960, 3106.

 
 

J. C. Sheldon, Ibid., 1 6 , 4183.

J. C. Sheldon, Ibid., 1 64, 1287.

P. E. Dai and J. H. Lundsford, J. Catal. 198 0, 99,
173

 

H. S. Harned, C. Pauling, and R. B. Corey, J. Amer.
Chem. Soc., 1960, 99, 4815.

 

75.

76.

77.

78.

79.

80.

81.

82.
83.

84.

85.

86.

87.

88.

89.

152

F. W. Koknat, J. A. Parsons, and A. Vongvushrintra,
Inorg. Chem., 1974, 13, 1699.

B. Spreckelmeyer, Z. Anorg. Allg. Chem., 1969, 3
225.

 

am

6.

 

P. J. Kuhn and R. E. McCarley, Inorg. Chem., 1965,
4, 1482.

 

P. Nannelli and B. P. Block, Inorg. Syn., 1976, 9111,
170.

 

J. E. Fergusson, B. H. Robinson, and C. J. Wilkins,
J. Chem. Soc. (A), 1967, 486.

 

D. Hartley and M. J. Ware, J. Chem. Soc. Chem. Commun.,
1967, 912.

 

P. B. Ayscough, "Electron Spin Resonance in Chemistry,"
Meuthan and Co., London (1965), Appendix Five.

R. Aasa and T. Vanngard, J. Mag. Res., 1975, 19, 308.

 

(a) J. Klinowski, J. M. Thomas, M. Audier and s.
Vasudevan, J. Chem. Soc. Chem. Commun., 1981, 570.

 

(b) L. A. Bursill, J. M. Thomas, and K. J. Rao,
Nature, 1981, 289, 157.

R. F. Schneider and R. A. Mackay, J. Chem. Phys.,
1968, 8%, 843.

P. B. Fleming, L. A. Mueller, and R. E. McCarley,
Inorg. Chem., 196 , 6, 1.

 

 

D. J. Robbins and A. J. Thomson, J. Chem. Soc. Dalton,
2350.
3

P. B. Fleming and R. E. McCarley, Inorg. Chem., 1970,
2. 1347. =====

J. G. Converse and R. E. McCarley, Ibid., 1970, 9,
13610 '

 

H
N
N

 

(a) A. Simon, H. G. Schnering, H. Wohrle, and H.
Schafer, Z. Anorg. Allg. Chem., 1965, 339, 155.

(b) A. Simon, H. G. Schnering, and H. Schafer, Ibid.,
1968, $61, 235.

(c) F. W. Koknat and R. E. McCarley, Inorg. Chem.,
1972, 11, 1812.

 

 

 

 

90.

91.

92.

93.

94.

95.
96.

97.

98.

99.

100.

101.

102.

Chem., 1976, 599, 1.

153

P. A. Vaughan, J. H. Sturdivant, and L. Pauling, g.
Amer. Chem. Soc., 1950, za, 5477.

P. B. Fleming, J. L. Meyer, W. K. Grinstaff, and
R. E. McCarley, Inorg. Chem., 1970, 9, 1764.

 

(a) H. Van Olphen and J. J. Fripiat, "Data Handbook
for Clay Minerals and Other Non-Metallic Minerals",
Pergammon Press, New York (1979), pp. 39.

(b) J. A. Gadsden, "Infrared Spectra of Minerals
and Related Inorganic Compounds", Butterworths,
London (1975), pp. 62.

H. Schafer, B. Plautz, and H. Plautz, Z. Anorg. Allg.

 

B. Spreckelmeyer, C. Brendel, M. Dartman, and H.
Schafer, Ibid., 1971, 386, 15.

 

 

R. B. Hahn, J. Amer. Chem. Soc., 1951, 13, 5091.

S. A. Best and R. A. Walton, Inorg. Chem., 1979, 18,

 

 

484. m
(a) W. H. Nelson and R. S. Tobias, Inorg. Chem.,
1963, g, 985.

 

(b) W. H. Nelson and R. S. Tobias, Can. J. Chem.,
1964, 9%, 731.

 

 

J. L. Meyer and R. E. McCarley, Inorg. Chem., 1978,
17, 1867. “—"

 

(a) N. N. Nag, T. Fransen, and P. Mars, J. Catal.,
1981. 6,6,. 77. #—

(b) M. Akimoto and E. Echigoya, Ibid., 1973, 11,

(c) H. Pines, "The Chemistry of Catalytic Hydro-
carbon Conversions", Academic Press, New York (1981),
Chapter 5.

K. Aykan, J. Catal., 1968, 1%, 281.

R. Rosenheim, H. P. Kuhn, Z. Anorg. Allg. Chem., 1910,
66, 1.
’VM

 

D. P. Shoemaker, C. W. Garland, and J. I. Steinfeld,
"Experiments in Physical Chemistry," McGraw-Hill,
New York (1974), pp. 253.

103.

104.

105.

106.

107.

108.

109.

110.

111.
112.

113.

114.

115.

116.

154

R. M. Lindner, K. Haller, and S. L. Helwig, Z. Anorg.
Chem., 1923, 139, 209.

L. J. Guggenberger and A. W. Sleight, Inorg. Chem.,
1969, Q, 2041.

 

D. Hartley and M. J. Ware, J. Chem. Soc. Chem. Commun.
1967, 912.

 

(a) K. Nakamoto, "Infrared and Raman Spectra of
Inorganic and Coordination Compounds", John Wiley
and Sons, 3rd Ed., New York (1978), pp. 227.

(b) I. Nakagawa and T. Shimanuchi, Spectrochim.
Acta, 1964, 99, 429.

R. M. Silverstein, G. C. Bassler, and T. C. Morrill,
"Spectrometric Identification of Organic Compounds",
3rd Ed., New York (1974), p. 208.

 

R. A. Plane and J. P. Hunt, J. Amer. Chem. Soc.,
1957. 79. 3343.

F. A. Cotton and T. E. Haas, Inorg. Chem., 1964, 3,
10.

 

 

 

P. C. H. Mitchell and F. Trifiro, J. Chem. Soc. A.,
1970, 3183.

T. J. R. Weakley, Struct. Bond., 1974, 1%, 131.

 

 

(a) H. So and M. T. POpe, Inorg. Chem., 1972, 11,
1441.

 

(b) J. J. Altenau, M. T. Pope, R. A. Prados, and H.
So, Inorg. Chem., 1979, 19, 418.

 

J. C. Sheldon, Chem. Indust., March 11, 1961, 323.

 

D. W. Breck, "Zeolite Molecular Sieves; Structure,
Chemistry, and Use", John Wiley and Sons, New York
(1974). pp. 287.

(a) M. Boudart and R. A. Della Betta, in "Catalysis",
J. W. Hightower (Ed.), North Holland, Amsterdam (1973),

pp. 1329.

(b) T. E. Whyte, Jr., and R. A. Della Betta,Cata1.
Rev.-Sci. Eng., 1982, 25, 567.

 

(a) D. J. Elliot and J. H. Lundsford, J. Catal.,
1979, 91, ll.

 

116.

117.

118.

119.

120.

121 O

122.

123.

124.

125.

126.

127.

155

(b) P. Gallezot, G. Coudurier, M. Primet, and B.
Imelik, ACS Symp. Ser., 1221, 59, 144.

 

M. Minachev and Y. Isakov, "Zeolite Chemistry and
Catalysis", J. A. Rabo (Ed.), Am. Chem. Soc., Washing-
ton, D.C. (1976), pp. 552.

(a) J. A. Rabo, P. E. Pickert, D. N. Stamires, and
J. E. Boyle, Actes Deuxieme Congr. Internat. Cat-
alyse, Paris, 1961, 2, 382.

 

 

(b) J. A. Rabo, P. E. Pickert, R. L. Mays, Ind.
Eng. Chem., 1961, 53, 733.

 

(a) P. B. Weisz, V. J. Frilette, R. M. Maatman, and
E. B. Mower, J. Catal., 1962, t, 307.

(b) P. B. Weisz, V. J. Frilette, J. Phys. Chem.,
1960, 64, 382.
—’\:’\;

(a) P. Gallezot, A. Alarcon-Diaz, J. A. Dalmon, A.
J. Renouprez, and B. Imelik, J. Catal., 1975, 39,
334.

(b) P. Gallezot and B. Imelik, "Zeolite Chemistry
and Catalysis", J. A. Rabo (Ed.), Am. Chem. Soc.,
Washington, D.C. (1976), pp. 66.

 

 

S. W. Bukata, C. A. Castor, and R. M. Milton, U.S.
Patent 3,013, 988 (1966).

D. Ballivet—chhenko and G. Coudurier, Inorg. Chem.,
1979, gg. 558.

B. Imlik, C. Naccache, Y. Bentaarit, J. C. VeDrine,
G. Coudrier and A. Praciad, "Catalysis by Zeolites",
Elsevier, New York (1980).

 

P. E. Dai and J. H. Lundsford, J. Catal., 1980, 64,
173 and 184.

F. C. Wilhelm, "Preparation and Characterization of
Mo-Containing Zeolite Catalysis", Climax Molybdenum
Co., Report L-287-51, 1977.

Aé P. Bolton and M. A. Lanewala, J. Catal., 1970,
l , 154.
mm

(a) E. M. Flanigen, "Zeolite Chemistry and Catalysis"
J. A. Rabo (Ed.), Amer. Chem. Soc., Washington, D.C.
(1976), pp. 80.

156

127. (b) E. M. Flanigen, H. Khatami, and H. A. Szymanski,
Advan. Chem. Ser., 1971, 101, 201.

J. A. Dean (Ed.), "Lang's Handbook of Chemistry",

128.
12th Ed., McGraw-Hill, New York, 1979.