.\T...‘. .7"
‘1‘?" ~ . ‘;‘

1.4-.":-|L§|~\
h‘I‘I‘rbL‘w
.3.

 

N
. I ‘ - ‘_.‘,‘;‘.,‘ . . . .
V V ' n a" .‘L.\A1’::‘A . .‘. —

 

 

x

yv~ (.417; .

....’-....r’
.

 

Km”.
MAR:

q} r2
I V:
‘a’a‘nrwvm

i??? '- -

-. .,

Mr“ V.

I. 4714"
I

‘0

 

 

' #97)“ f

 

o

 

p.57.
”52‘:

‘~1

”‘1.!',‘I:".'l' ,7
-r. i;

R V
M.
'1‘.

 

”‘4.
,1-“
4-. .

.a.1 ‘-

g

 

.v < m- . - .nr‘cmfi.nmusu

 

 

 

,
...: .. ,..-.--,
:r. .;-._,I;,_,..
.
.

 

’
,.
-2: :“7‘ ,,_ hI;._,.:r_...... W3,
.—
. :1" :g,
- by”! ‘ un—mr‘. '-
5.... 4..
:..-'i;.... 7:55am
..

9 “up: .-.-: -.-...
a. v...

   

 

.. ..
.. m... .... 1v .~
.4....-:7.‘~:.V-: .V. ‘

L3
SIUTY

J l\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\\l

31w2

This is to certify that the

dissertation entitled
Intercalation and Pillaring Reactions of

Layered Alkali Metal Silicates

presented by

James Steven Dailey

has been accepted towards fulfillment
of the requirements for

Eh D degree in ChemiStry

% MW;
. 19f» professor

 

Datey/é /7?/

MSU is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

 

1“ § _
; HEEARV l
éMmhis-‘an 5:53.,
5 University,

L_

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

__ fl

 

 

 

 

lit:
_T_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
_ cMchnS-q

 

INTERCALATION AND PILLARING REACTIONS OF LAYERED
ALKALI METAL SILICATES

by

James Steven Dailey

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1991

ABSTRACT

INTERCALATION AND PILLARING REACTIONS OF LAYERED
ALKALI METAL SILICATES

by

James Steven Dailey

The goal of this work was to design and synthesize pillared
microporous materials from intercalation complexes of the layered
alkali metal silicates. By extrapolating knowledge gained from other
layered silicate clays to the alkali metal silicates, new classes of
silicate intercalation compounds have been developed. Two routes
were used to intercalate robust pillars between the layers of the
layered alkali metal silicates, direct ion exchange of pillar cations for
alkali cations, and the intercalation of siloxane reagents within

preintercalated derivatives of H+-magadiite.

The cationic metal cage complex (l,3,6,8,10,13,16,19
octaazabicyclo[6.6.6]eicosane)cobalt(III), Co(sep)3+, has been

investigated as a potential pillaring reagent for Na+-magadiite
(Nal..,Si1 4027.9(OH)1 97.6 H20) a synthetic layered sodium silicate.

Reaction of Na+-magadiite with aqueous solutions of Co(sep)Cl3 at
25°C resulted in the binding of Co(sep)3+ cations to the external
crystallite surface of the layered silicate. In contrast, an intercalated

product exhibiting a 17.6 A basal spacing was generated by reaction

at 100°C. 29Si MAS NMR and FT-IR spectroscopy indicate that

Co(sep)3+ intercalated reaction products retain the magadiite layer
structure. Moreover, scanning electron micrographs of the reaction
products showed retention of the original particle morphology,
suggesting a topotactic intercalation. However, during intercalation,
some of the Co(sep)3+ underwent an unusual demetalation reaction
leaving a combination of C001) and Co(sep)3+ between the layers.
Nitrogen surface area analysis showed that only a small amount of
microporous surface existed in the Co(sep)3+ intercalated derivative,
suggesting that most of the interlayer space is "stuffed" with cobalt

species.

Reaction of the silicic acid H+-magadiite, H3_lSi14 029,5-H20, with
an excess of an n-alkylamine such as octylamine afforded high basal
spacing derivatives in which the silicate layers were separated by
layers of solvated alkylammonium cations ((1001: 34 A). The
properties of the alkylammonium silicate as a precursor ‘for the
intercalation of hydrolyzable reagents such as tetraethylorthosilicate,
TEOS, has been investigated. Neat TEOS readily intercalates at room
temperature and partially polymerizes in a molecularly regular
fashion in the H+-magadiite galleries, giving rise to ordered basal
spacings in the range 23.3-28.1 A. The mechanism for
polymerization appears to occur by a topochemical process.
Calcination of the TEOS hydrolysis products yielded layered
magadiite derivatives regularly intercalated by silica pillars. Silica
pillared derivatives exhibited gallery heights of 9.5-14.5 A and
microporous surface areas of 480-670 mzlg depending on the TEOS

reaction stoichiometry. The silica pillared products contain interior

surface silanol groups of the type Q3 HOSi(OSi)3 and Q2 (HO)ZSi(OSi)2
as confirmed by proton cross polarization 29Si magic angle spinning
NMR spectroscopy. These surface silanol groups should prove to be
useful as grafting sites for metal centers for catalysis or sensor

applications.

In memory of my father, William J. Dailey.

ACKNOWLEDGMENTS

I would like to thank my research advisor Professor Thomas J.
Pinnavaia for supplying patient guidance and support during this
course of study. Special thanks go to Professor Carl H. Brubaker for
his helpful comments as second reader. I would also like to extend
my deepest gratitude and heart-felt thanks to all the Pinnavaia
group members past and present, who made graduate school an
enjoyable learning experience. Ahmad Moini and Laurent Michot
deserve special thanks for the advise, encouragement and friendship
they have given over the years. Also, Lucy, Niasha and Tendai
Michot I thank you all from the bottom of my heart for helping to
make my last two years happy and bright. A very special thanks

goes to Jayantha Amarasekera for helping me with my job search.

Financial support given by the Department of Chemistry
Michigan State University, the National Science Foundation and the
National Institute of Environmental Health Sciences Superfund Grant

is gratefully acknowledged and appreciated.

I want to give special thanks to all my friends back home
especially Brian "Red" Hansen and Chip "Charlie" Laufer for always

having faith in the "Vector”.

To my family I extend my love and thanks for being so
supportive. I especially want to thank my mother, sister,
Grandparents, Aunts and Uncles who have been behind me always

and who have my unending affection and gratitude.

Most importantly I am grateful to my wife Cindy for all the
love and understanding she has given me over the years. Her
friendship and patience during this time were instrumental to the
completion of this dissertation. To Cindys' family , Terry Sr., Mary,
Bill and Terry Jr., thank you for taking me into your home and
making me feel a part of your family. Your love and support over

the last few years has helped me greatly in my studies.

TABLE OF CONTENTS

 

C ha p t er P a g e
LIST OF TABLES _ ,_ v _- x
LIST OF FIGURES ___________ - - -- - - ..... - xii

 

CHAPTER I. Layered Alkali Metal Silicates: An Overview .1

 

 

 

 

 

 

 

 

 

A. Introduction ...-....1
B. Syntheses - - -- - 2
C Structures of the Alkali Metal Silicates ............................................... 6
D. Morphology .............................................................. 19
E Intercalation Properties- .................................................... 19
l. Cation Exchange Properties ........................................................... 19
2. Alkylammonium Derivatives as Reaction Precursors ........ 25
3. H+-Silicates _ - - _ -- J6
4. Intercalation Complexes of H+-Silicates - -.27
5. Structure of H+-silicates -- -31
6. Silylation of Interlamellar Silanol Groups....- 33
7. Pillared Layered Silicates .33
F. List of References ....................................................................................... 36

vii

Chapter Page

CHAPTER II Intercalative Reaction of a Cobalt(III)
Cage Complex, Co(sep)3+, with Magadiite,

 

 

 

 

a Layered Sodium Silicate .......................................... 41

A. Introduction ................................................................................................. 41
B. Experimental. ........ - - ........................... 42
1. Preparation of Na+-magadiite ....................................................... 42

2. Preparation of Co(sep)Cl3 .............................................................. 44

3. Co(sep)3+-magadiite Reactions .................................................... 45

4. Physical Measurements .................................................................. 45

C Results and Discussion--- ......... --------46
1. Preparation of Na+-magadiite ...................................................... 46

2. Reaction with Co(sep)3+, ................................................................. 51

D. Conclusions - - ................................................................................. 70
B. List of References ....................................................................................... 73

CHAPTER III Silica Pillared Layered Silicates Derived
From the Polymerization of
Tetraethylorthosilicate in the Galleries of

 

 

Alkylammonium Exchange Magadiite ............ 75

A. Introduction ----- -- - - ---75
B. Experimental ------------------ - -- ------------------- 77
1. Synthesis of H+-magadiite ............................................................. 78

2. Synthesis of Octylammonium-m‘agadiite .................................. 78

3. Synthesis of Silica Intercalated Magadiites ............................. 78

4 . Physical Measurements .................................................................. 79

viii

 

C Results and Discussion - - ----------.- ........................ 80

 

1. Na+- and H+-magadiite.- - -- -- --80
2. Octylamine/Octylammonium Magadiite .................................. 84
3. TEOS/magadiite Reaction Products ............................................ 85
4. Mechanism of Intercalation. ...................................................... 107
D. Conclusions ................................................................................................. 113
E List of References .................................................................................... 115

ix

Table

1.1.

1.11.

1.111.

11.1.

11.11.

11.1“.

111.1.

LIST OF TABLES

 

 

 

 

 

Page
Layered Alkali Metal Silicates and their Proton
Exchange Forms -- - - ----- -- -- ----.-.---3
29Si MAS NMR Data for Various Samples of Na+-
magadiite and H+-magadiite--- - ---15
Possible guest molecules for interlamellar adsorption
by crystalline silicic acids and the gallery height
(47)- - -------- -- ------ - -- -- -- ----------- 28
Chemical Compositions of Natural and Synthetic
Magadiite ----50
Compositions (wt %) of Products Prepared by
Reaction of Na+-magadiite with Co(sep)3+ at 25°C and
100°C ------ - - -56
Surface Area Analysis of Intercalated Derivatives ............ 62
Basal Spacing and Gallery Heights of Magadiite
Reaction Precursors and TEOS Reaction Products ............... 90

III.II. Analytical Data ................................................................................... 91
III.III. Surface Area Analysis (m2/g) of Pillared Derivatives ..... 93

III.VI. 29Si MAS NMR Chemical Shift Values and Q3/Q4
ratios for H+-magadiite and the TEOS/magadiite

reaction products .......................................................................... 102

xi

LIST OF FIGURES

 

 

 

Figure Page
1.1. Schematic representation of a silicate layer which is
composed of continuous sheets of Q3 corner sharing
tetrahedra. Structure derived from the crystal
SITUCtuI'C Of H28i205, I (64) ................................................................. 7
1.2. Variation in the folding of Si205 units with cation
charge to radius ratio in a) Li28i205, b) HZSiZOS—I, c) a-
NaZSiZOS, and d) KHSiZOS-II (9) -- -- - -8
1.3. Polyhedral representation of the crystal structure of
synthetic makatite (22) ..................................................................... 10
1.4. Polyhedral representation of the crystal structure of
K28i409. a) view perpendicular to the layer, b) view
parallel to the layer (23) - -- - 11
1.5. Schematic representation of the interlayer space of A)
Na+-magadiite and B) H+-magadiite (30)"- 14
1.6. Postulated layer structure for Na+-magadiite. a)

verticcs of the magadiite layer represent silicon
positions; oxygen positions are between adjacent
silicon pairs. The fourth coordination position of
silicon is not shown. b) Polyhedral representation of
the layer structure of magadiite; verticcs represent
oxygen positions (24) ------ ---- -- 17

 

1.7. Representation of some intercalation chemistry of the
layered alkali metal silicates (47) -----

B

xii

1.8.

1.9.

1.10.

1.11.

11.1.

11.2.

11.3.

11.4.

Arrangements of n-alkylammonium cations within a
layered silicate. Cation chain length was varied while

keeping layer charge constant. a) CZH 5NH3+, b)

C8H17NH3+, 311d C) C12H25NH3+ ......................................................

Basal spacing of n-alkylammonium derivatives of
layered silicates vs number of carbon atoms in the
alkyl chain for kanemite x (2), KHSi205 E] (37),
magadiite O (13), and kenyaite 0(17). All spacings
were determined for derivatives in equilibrium with

RNH3+ solutions ...................................................................................

Basal spacing (d001 A) versus number of carbon
atoms (NC) in the alkyl chain for n-alkylamine

derivatives of H+—Kanemite El (2), HZSi205-10(37),
HZSiZOS-II x (49), and H+-magadiite o (38). All basal

spacings were derivatives under mother liquid ..................

Interlamellar grafting of trimethylsilyl groups into the
intracrystal surface of silicic acids by reaction of
trimethylchlorosilane (TMCS) with the N-
methylformamide (NMF) solvated form of the silicic
acid. The unsolvated silicic acid is unaccessable for

reaction with TMCS (55) - - ---------- -

Schematic representation of the cobalt sepulchrate

 

complex. ---------- - - - _- _ - -_

X-ray diffraction pattern (Cu-Ken) of an oriented film
sample of Na+-magadiite - - ..............

Thermogravimetric analysis curves obtained under
flowing argon for samples (A) Na+-magadiite, (B)

..23

..24

..30

43

---47

Co(sep)3+-mag(25°C) and (C) Co(sep)3+-mag(100°C) ............. 49

X-ray diffraction patterns for oriented film samples of
the reaction products obtained after successive

xiii

11.5.

11.6.

11.7.

11.8.

11.9.

111.1.

111.2.

111.3.

exchange of Na+-magadiite with 0.12 N Co(sep)Cl3 at
25°C-- -------- -----

X-ray diffraction patterns for oriented film samples of
the reaction products obtained after successive

exchange of Na+-magadiite with 0.12 N Co(sep)Cl3 at

100°C- - - - _ - - --

2S’Si Magic Angle Spinning NMR spectra for (A) Na+-
magadiite, (B) Co(sep)3+-mag(25°C), and (C) Co(sep)3+-

mag(100°C)---- ----- - -------------------

52

..--54

---58

 

Infrared spectra (2% w/w KBr pellet) of (A) Na+-
magadiite, (B) Co(sep)3+-mag(25°C) and (C) Co(sep)3+-
mag(100°C)--- ----------------------- -

 

Scanning electron micrographs at X6600 and X1500
magnification for (A) Na+-magadiite, (B) Co(sep)3+-

mag(25°C), and (C) Co(sep)3+-mag(100°C) ..............................

Proposed positions of Co(sep)3+ ions in the pleated
folds of magadiite. C024“ cations are placed at the
ridges which form between folds. Vertices of the
magadiite layer represent silicon positions; oxygen
positions are between adjacent silicon pairs. The

fourth coordination position of silicon is not shown ..........

X-ray diffraction patterns for film samples of (A) Na+-
magadiite, and (B) H+-magadiite

...64

...71

- -..-81

 

Thermogravimetric analysis curves obtained under
flowing argon for (A) Na+-magadiite, and (B) H’-
magadiite -- -

X-ray diffraction patterns for film samples of
octylammonium-magadiite (A) sample air-dried at
25°C and (B) solvated by excess octylamine

1”
U)

 

xiv

111.4.

111.5.

111.6.

111.7.

111.8.

111.9.

111.10.

111.11.

111.12.

X-ray diffraction patterns for film samples of the
uncalcined TEOS/magadiite reaction products that
resulted from treatment of octylamine solvated
octylammonium-magadiite with
tetraethylorthosilicate. The TEOS/magadiite mole

ratios were (54:1), (90:1) and (153:1) .....................................

X-ray diffraction patterns for film samples of
TEOS/magadiite reaction products after calcination.

Calcined in air at 360° C for four hours ....................................

Nitrogen adsorption/desorption isotherms for the
calcined TEOS/magadiite reaction products produced
with a ratio of (54:1), (90: l) and (1:531) moles
TEOS/mole magadiite-= - --------------

Schematic representation for the variation of pillar

...87

..89

-- =...95

 

size with the change in reaction stoichiometry

Infrared spectra for KBr pellets (0.25 wt %) of (A) H+-
magadiite, (B) air-dried octylammonium-magadiite, (C)
TEOS/magadiite reaction product (90:1) air dried, and
(D) TEOS/magadiite (90:1) reaction product after

calcining at 360°C ..............................................................................

29Si MAS NMR spectra of H+-magadiite and calcined
TEOS/magadiite reaction products produced with a
ratio of (54. l), (90: 1), and (153: 1) moles TEOS/mole
magadiite

29Si MAS NMR spectra of Na+-magadiite, H+-magadiitc
and uncalcined TEOS/magadiite reaction products
produced with a ratio of (54: 1), (90' 1), and (:1531)

 

moles of TEOS/mole magadiite-"-

29Si CP-MAS NMR spectra of calcined TEOS/magadiite
reaction products produced with a ratio of (54:1),
(90:1), and (153:1) moles TEOS/mole magadiite.

...96

100

Products calcined at 360°C in air for four hours ................. 105

29Si CP-MAS NMR spectra of H+-magadiite, and
uncalcined TEOS/magadiite reaction products produced

XV

111.13.

111.14.

111.15.

with a ratio of (54:1), (90:1), and (153:1) moles of

TEOS/mole magadiite ....................................................................... 106

Scanning electron micrographs of H+-magadiite at A)
X6000 and B) X600 magnification

Scanning electron micrographs at X6000 magnification
for A) air-dried octylammonium-magadiite, B)
TEOS/magadiite reaction product (54:1) air-dried, C)

calcined TEOS/magadiite (54:1) .................................................

Schematic representation for the intercalation and
pillaring of octylamine solvated octylammonium-

magadiite .............................................................................................

xvi

-108

.110

.112

CHAPTER I

Layered Alkali Metal Silicates: An Overview

A. Introduction

A large number of naturally occuring alkali metal silicate
minerals with layered structures have been discovered in alkaline

lakes around the world over the last 25 years. They include the

anhydrous sodium silicate natrosilite (Na28i205) (1) and the hydrous
sodium silicates kanemite (NaHSi205-5H20) (2), makatite
(Na28i408(0H)2-4H20) (3), magadiite (Na28i14029-11H20), and
kenyaite (Na28i200410xH20) (4). Also a naturally occuring crystalline
silicic acid, silhydrite, was discovered in a Trinity county (CA)
magadiite deposit (5,6). Synthetic forms of these and other alkali
metal silicates are readily obtainable. These layered silicates also
possess interesting ion exchange and intercalation properties. By
extrapolating knowledge gained from other layered silicate clays to
the alkali metal silicates, it should be possible to design new classes
of silicate intercalation compounds with novel catalytic properties.
The proton exchange forms of the alkali silicates represent new
families of crystalline silicic acids with layered structures. Owing to
the presence of interlamellar SiOH groups, the layered silicic acids
may prove to be excellent hosts for the immobilization of metal

complexes. In anticipation of the future catalytic application of these

1

2

materials, we review here the synthesis, structure, and intercalation
properties of the layered alkali metal silicates and their

corresponding layered silicic acid forms.
B . S y n th e s e s

Alkali silicates have been synthesized by two general methods,
namely, by high temperature reaction of SiOz with alkali metal
carbonate melts, or by hydrothermal reaction of aqueous alkali
hydroxide solutions with SiOz. Both the molten salt and
hydrothermal routes can generate anhydrous or hydrous alkali
silicates, depending on the reaction conditions. Table 1.1. lists the
compositions of typical alkali metal silicate phases and their acid

exchange forms, where known.

or and 13 phase sodium disilicates have been synthesized by
fusing Na2C03 and Si02 in a 1:2 molar ratio at 500°C-860°C. The
lower temperature polymorph p-Na28i205 forms between 600-
700°C, while a-Na28i205 crystallizes in its pure form between 750°C
and 860°C (7-9). Potassium disilicate, K28i20 5, was readily prepared
at 800°C by fusing K2CO3 and SiOz in a 1:2 molar mixture (10). The
more silicious K28i409 can be synthesized by heating a X20-4Si02
glass (10) at 400°C in the presence of small amounts of H20 (9).

The hydrothermal synthesis of layered alkali silicates has been
accomplished by the reaction of aqueous NaOH, KOH or LiOH with SiOz
between 60°C and 300°C. The higher reaction temperatures, 200°C-
350°C resulted in anhydrous alkali silicates such as Na28i20 5 (ll).
Wey et. al (12) synthesized KHSi20 5 by reaction of a ternary mixture

 

m F <2 - o F .5223 omzmsmozamx
a F emirsoaaf - a F .522; cures-03m?
A F cures-03mm: - A FF; .5223 F2539; carrsoaammz
om omxxsuozamz - 2.3.2 .5222 22:835. ONIFFauoEmum-z
o F omszoaaf - e F-m F .522; curetoymumz
- <2 mm 2.2.: .5925 Fezmxmé camemfovaovammz
#3.. F 2 seem“: co m F5 .522; #3me
3 <2 mm mm :2: momma:
- <2 m m o F .m .582; ..53me
- <2 - o F :2: momma-x
Fe =. .= ._ sow-Bu: - m .5222 6:229; omzmnomazmz
cm >_ soy-mm: mm m6 :2: mo~_m~m2u
mime ._ ._ sow-am: _ 3 NF :9: mo~_m~mze
8:232". - coco-.82“.
oSFoEFm 6.32:5 8:222; .8265.
222:0 ea“. .22 22:” .926 28556 ozofiim 28:6 :22

 

.mFEom oweanoxm 220.5 :2: new «235m .802 =8=< coho-$4

.: 035,—.

4

of KOH, Si02, and H20 in a 1.1:2:20 molar ratio at 300°C. The
naturally occuring hydrous sodium disilicate kanemite .
(NaHSi205-5H20), has been obtained as an alteration product of other
sodium silicates such as Na2$i20 5, makatite (Na28i409-5H20) or
magadiite (Na28i14029-11H20) (2). Kanemite has also been derived
from NaOH-SiOz mixtures prepared in methanol subsequently dried,
heated to 700°C and dispersed in H20 (2). The other hydrous layered
alkali metal silicates were formed most readily at temperatures
between 60°C and 175° (13,14).

A layered sodium octosilicate (Na28i8017-XH20) has been

prepared by heating a mixture of 30% "Ludox" colloidal silica solution
and NaOH (3) in sealed glass bottles at 100°C for one to three weeks
(15,16). Mixtures with a Si02/Na20 molar ratio of 4.6:1 resulted in

an opaque crystalline material with an empirical formula of

8Si02:Na20:9H20. A layered tetrasilicate with a compositional range
close to that of makatite has been synthesized by Baker et a1. (11).
The material produced at 70°C from a 1:2:17 NaZO:Si02:H20 reaction
mixture exhibited a final composition of 3NazO-I3Si02-11H20.
Another synthetic preparation reported in the literature, utilized a
ternary system with a NaOH:Si02:H20 ratio of 2:4:20, but no other

details were given (17).

Synthetic Na+-magadiite (Na28i14029-11H20), was first
prepared by the crystallization of sodium silicate solutions at 100°C
over a three to four week period (15). Lagaly et.al. (13) studied in
detail the range of ternary compositions that resulted in crystalline

samples of Nat-magadiite (13). Na+-magadiite forms over a range of

5

reaction compositions, but the best samples were obtained by
reaction of a 2:9:75 molar ratio of NaOH:Si02:H20 at 100°C in sealed
glass ampules. At certain compositions magadiite was formed in
combination with kenyaite (Na28i20041-xH20) (17). The phase
boundary between these two products has also been investigated
(14). Reaction mixtures with Si02/NaOH $16 yielded Na+-magadiite
as an intermediate in the formation. of kenyaite. These mixtures
result in quartz formation at longer reaction times, indicating a
magadiite to kenyaite to quartz transformation most likely occurs
(17). The alkali cation also has an effect on which silicate will form.
Sodium ions favor the formation of Na+—magadiite, while potassium
ions favor K+-kenyaite formation (17). More recently, a new
potassium silicate hydrate with a composition, K20014Si02-8H20,
similar to K+-magadiite was prepared which possesses enhanced
intracrystalline reactivity (18). Various other alkali-kenyaite
materials containing Li, Cs, and Sr were also prepared under special
conditions (19). The effect of high levels of anions other than OH' on
the synthesis of magadiite and kenyaite indicated that the
kenyaite/magadiite ratio decreased with an increase in the atomic
number of the halide used. The presence of F' increased the
conversion to kenyaite but also enhanced the conversion of kenyaite
to quartz (14). The increase of hydrothermal reaction temperatures,
from 100°C to between 125° - 175°C, dramatically decreased reaction

times for the synthesis of magadiite and kenyaite(l4).

C. Structures of the Alkali Metal Silicates

In general, layered alkali silicates are composed of silicate
layers separated by alkali cations. Hydrated forms may also contain
H20 between the layers usually as a hydration sphere around the
interlayer cation. Adjacent layers are held together by electrostatic
interactions between the layer and cation or by hydrogen bonding

with interlayer H20 .

The crystal structures of certain alkali silicates have been
determined (cf. Table 1.1.). The layered structures of the alkali
disilicates, Li28i205, Q-Na2Si205,p-Na2Si205, and KHSi205 are
composed of continuous sheets of Si04 tetrahedra which corner share
with oxygens from adjacent tetrahedra to form six membered rings.
Each Si04 tetrahedron shares three oxygens with adjacent tetrahedra.
The remaining apical oxygen of the tetrahedron is directed above or
below the plane of bridging oxygens in an alternating fashion, as
shown in Figure 1.1. Since each Si04 tetrahedron in the Si205 layer
carries a negative charge, one monovalent cation or a proton is
required to satisfy layer charge neutrality. Due to the flexibility of
layers formed by corner sharing of $0,, tetrahedra a change in cation
size or charge can alter the degree of layer folding. For instance,
Figure 1.2. illustrates the decrease in folding of the tetrahedral layer
as the cation size increases from Li+< H+< Na+< 10. It appears that the

smaller cations are more enfolded within the tetrahedral layer.

A layered structure was corroborated for KHSi205 and

kanemite (NaHSi20503H20), by 29Si MAS NMR spectroscopy. The 2S’Si

 

1.1. Schematic representation of a silicate layer which is composed
of continuous sheets of Q3 corner sharing tetrahedra. Structure
derived from the crystal structure of HZSiZOS, I (64).

1.2. Variation in the folding of Si205 units with cation charge to
radius ratio in a) Li28i205, b) HZSiZOS-I, c) a-Na28i205, and d) KHSi205-
II (9).

9
MAS NMR spectrum of KHSi205 indicated one resonance characteristic
of a Q3 OSi(OSi)3 type environment (20). The hydrous sodium

disilicate kanemite exhibited two resonances, one at -99 ppm
relative to TMS characteristic of a Q3 OSi(OSi)3 type environment and
a weaker resonance near -110 ppm assigned to the presence of a Q4
Si(OSi)4 type environment (21). The Q4 resonance grew in intensity
with sample age and this was thought to result from the

condensation of adjacent layers.

The layered structure of synthetic makatite
(Na28i408(0H)204H20) was found to be based on a continuous sheet of
Q3 tetrahedra condensed to form six membered rings. But instead of
unbranched zweier chains as in the disilicates, makatite was built up
of unbranched vierer single chains. In addition, the makatite
structure was found to be composed of highly folded [Si204(0H)]°n'
layers which contain six membered rings of Q3 tetrahedra, as shown
in Figure 1.3. (22). More highly condensed alkali silicates such as
K28i409, although still built up of unbranched zweier single chains,
have layers composed of both four and six membered rings of $0,,
tetrahedra, as illustrated in Figure 1.4. Both Q3 and Q4 type

tetrahedra were present in this structure in an equal ratio (23).

The crystal structures for magadiite (Na2Si14029-11H20),
octosilicate (Na28i8017-xH20), and kenyaite (Na2Si20041-xH20), are
unknown. Structural information has been obtained using various
techniques such as thermal analysis, X-ray powder diffraction,
infrared spectroscopy, 1H, 23Na MAS NMR, and especially, 29Si MAS
NMR spectroscopy.

10

Al;

0(5‘

 

 

,: 55(3)

01101' are) I
unemucneo wenen

LA , ‘ 3-5-...
5 45
S

1.3. Polyhedral representation of the crystal structure of synthetic
makatite (22).

 

 
 

 

 

 

1.4. Polyhedral representation of the crystal structure of K28i409.
a) view perpendicular to the layer, b) view parallel to the layer.(23)

12

The X-ray powder diffraction pattern of octosilicate vacuum-
dried at 30°C indicated a basal spacing of 11.0 A. 293i MAS NMR
spectroscopy of sodium octosilicate indicated the layer was composed
of Q3 OSi(OSi)3 and Q4 Si(OSi)4 type silicon environments (24, 25).
Cross polarization resulted in an enhancement of the Q3 silicon
environment which confirmed the presence of Q3 silanol HOSi(OSi)3
groups (25). Infrared spectroscopy was used to elucidate structural
similarities between sodium octosilicate and various zeolites (24).
This study indicated the presence of structural blocks in sodium
octosilicate which contain five membered rings of tetrahedra. Two
models for the layer structure of octosilicate were proposed (24,25).
For instance, Schwieger et al. (25) condensed two makatite layers to
produce a model with the correct Q3/Q4 ratios and approximate basal
spacing. This model unfortunately did not contain the five
membered ring blocks shown present by the IR study. Garces et a1.
(24) proposed that sodium octosilicate has layers which consist of six
membered ring sheets of Si04 tetrahedra and structural blocks
composed of four, five, and six membered rings which protrude from
these sheets. The calculated layer thickness for this model would be
approximately 11A, close to the basal spacing of the hydrated form

of sodium octosilicate.

X-ray powder diffraction studies of natural and synthetic
samples of Na+-magadiite indicated a basal spacing of 15.6 A (26-
28). The basal spacing was consistent with silicate layers ~ll.2 A
thick plus a layer of hydrated sodium cations which fill the gallery
space (26,29). The X-ray diffraction patterns of synthetic Na+-

13

magadiite samples heated to between 25°C and 300°C showed that
the interlayer H20 is lost by 120°C. Also, Nat-magadiite heated to
165°C reversibly re-hydrates. Above 165°C there is a loss of
structural integrity (28). New structural information has been
obtained from thermally treated samples of naturally occurring
Trinity Center magadiite using a combination of 1H NMR spectroscopy
and infrared spectroscopy (30). Both techniques indicated the
presence of a large concentration of silanol groups in Na+-magadiite
~2Na+l30H. 23Na MAS NMR spectroscopy of these samples also
indicated two Na environments, which were assigned to interlayer
Na+ and NaOH (28). The proposed interlayer structure, as shown in
Figure I.5A., indicates that the negative charge of the layer is
balanced in part by either Na+ ions or protons. In addition, a bilayer
arrangement of H20 and ion pairs of Na+ OH' were thought to coexist
within the interlayer. An earlier 29Si MAS NMR study of proton
exchanged natural Trinity County Na+-magadiite, (31), indicated a

ratio of 1.7:] for the Q"':Q3 environments of silicon, corresponding to
5.2 Q3 HOSi(OSi)3 groups per Si14 unit cell.

The reported ratios for the Q‘4:Q3 silicon environments, deduced
from 29Si NMR spectroscopy for synthetic and natural magadiite
samples varies greatly. As shown by the data in Table 1.11. the
quantitative use of 29Si MAS NMR data requires that values for the
spin-lattice relaxation times, T1, be known in order to avoid
saturation due to insufficient spin relaxation between pulses. Spin-
lattice relaxation times are especially long for synthetic Na+-

magadiite with values of 160 seconds and 280 seconds, being

14

 

T.

' ...63

15.511 %‘®;® m ® @ @H A
®H®
W®®

I
I
I
i—-—.-4 magadllIe
~11 A

 

1

 

 

e ' “
11.2A AflA Us We HA ':11 He Ha HA B

 

 

1.5. Schematic representation of the interlayer space of A) Na+-
magadiite and B) H+-magadiite (30).

l 5
Table 1.11. 29Si MAS NMR Data for Various Samples of Na+-magadiite

and H+-magadiite

 

 

Magadiite Q4/Q3 Q3 sites/unit cell ref
Sample

Na+,Synthetic 2.0 4.6 2 6
Nat, Synthetic 1.0 7.0 25
Na+, Synthetic 2.5 4.1 25
Na+, Synthetic 3.0 3.5 24
Na+, Synthetic 2.9 3.6 3 2
H+, Synthetic 3.6 3.0 3 2
H+,Synthetic 3 .0 3 .5 2 4
Na+, Natural 1.2 6.3 25
H+, Natural 1.7 5.2 3 1

observed, for the Q3 and Q4 silicon environments respectively (32).
The differences in Q3 and Q4 silicon relaxation times for synthetic
Na+-magadiite in this case were attributed to 2S’Si-IH dipolar

relaxation of the Q3 silicon environments due to coupling to

interlayer H20 (32). On the basis of thermal analysis (weight loss)
and Na20 content, the empirical formula for this synthetic Na+-
magadiite sample was found to be Na1.7Si14027.9(0H)1.907.6H20 (32).
The 29Si MAS NMR spectra of this synthetic sample reveals a Q4:Q3
ratio of 2.85:1 which is equivalent to 3.6 Q3 sites per Si14 unit. This
agreed closely with the empirical formula for Na+-magadiite which
indicated 3.6 [Na+, 0H] Q3 groups per Si14 unit. Experiments
performed with delay times < 5 T1 in this case resulted in spectra
which exhibited artificially enhanced Q3 silicon resonances. Previous

29Si MAS NMR studies, Table 1.11., were performed with delay time

16

values between 4 and 20 seconds, much to short for complete
relaxation of Si environments subjected to pulse widths of 90°. This
may explain the discrepancies in Q4:Q3 ratios observed by various
authors. The presence of paramagnetic impurities in natural samples
of magadiite may provide a mechanism for more efficient relaxation

but synthetic samples require careful relaxation studies.

Two models for the layer structure of Na+-magadiite have been
reported (24, 25). The model proposed by Garces et al. (24)
combines information obtained from X-ray powder diffraction,
infrared spectroscopy and 29Si MAS NMR. The model consists of five
membered ring blocks of tetrahedra protruding from both above and
below a continuous sheet of six membered ring Q3 tetrahedra, as
shown in Figure 1.6. The model put forward by Schweiger et al. (25)
was based on X-ray basal spacing and 29Si MAS NMR data. In this
latter model the layer is formed by connecting three makatite-type
layers to form an (Sin 026)?’ unit cell (25). The proposed structure
resulted in a framework composed of four, six and eight membered
ring blocks of tetrahedra. This model is incompatible with infrared
results reported by Garces et a1. (24) which indicated the presence in
N a+-magadiite of five and six membered ring tetrahedral blocks,

similar to those reported present in pentasil zeolites.

Two orders of basal plane X-ray reflections at 19.8 A and 9.9 A
have been reported for synthetic and natural Na+-kenyaite samples
(17). Synthetic K+-kenyaite also is reported to exhibit basal
reflections near these values (17,25). Thermal treatment of

synthetic Na+-kenyaite at 120°C under vacuum resulted in a

17

 

 

1.6. Postulated layer structure for Na+ -magadiite. a) vertices of the
magadiite layer represent silicon positions; oxygen positions are
between adjacent silicon pairs. The fourth coordination position of
silicon is not shown. b) Polyhedral representation of the layer
structure of magadiite; verticcs represent oxygen positions (24)

18

decrease in basal spacing to 17.7 A. This decrease in basal spacing
was attributed to the loss of a monolayer of H20. Synthetic Na+-
kenyaite remained crystalline to 600°C, but a transformation to
quartz occured at 700°C. The retention of the 17.7 A basal reflection
for hydrous Na+-kenyaite suggested that the van-der Waals
thickness of the layer was near this value (17). As stated previously,
the synthesis of kenyaite resulted in the formation of magadiite as a
byproduct (14). This suggested that magadiite was a possible
intermediate for kenyaite formation. The possibility that layers of
magadiite condense to form the kenyaite layer was discussed
previously (17). This mechanism is not likely because X-ray
diffraction studies of suspensions taken from kenyaite syntheses

show no phase intermediate between magadiite and kenyaite (17).

The 29Si MAS NMR spectra of synthetic Na+-kenyaite indicated
silicon resonances characteristic of Q3 OSi(OSi)3-type and Q4 Si(OSi)4-
type environments (25). The observed Q4:Q3 ratio of 4.1:1 for
kenyaite supports the evidence that kenyaite contains more highly
condensed layers than Na+-magadiite. Unfortunately, spin-lattice
relaxation times for synthetic kenyaite were not determined and this
puts into question the reported Q4:Q3 ratio. 0n the basis of basal
spacing, chemical composition, and Q4:Q3 ratios a model structure was
derived by Schweiger and co-workers for synthetic kenyaite
(25.28.33). The combination of 5 makatite layers resulted in a
theoretical thickness of 25.1 A and a Q4:Q3 ratio of 4:1. The 23Na
MAS NMR spectrum of synthetic kenyaite with proton to sodium

cross polarization revealed the presence of two Na environments

19

(34). As in synthetic Na+-magadiite, these environments were
assigned to the presence of Na+ interlayer cations and to physisorbed

NaOH ion pairs.
D. Morphology

Scanning electron micrographs of synthetic Na+-octosilicate
indicate a crystallite morphology composed of rectangular plates, 3-4
pm on a side and - 400 A thick (15,16,24). The crystallites of
magadiite and kenyaite, both natural and synthetic derivatives, show
a more unusual particle morphology composed of intergrown plates
which form rosette-like spherical nodules (17,24,32). Variations in
synthetic reaction conditions, such as changes in the temperature of
crystallization from water glass or in the alkali metal cation caused

alterations in the particle morphology of kenyaite (17,19).
E. Intercalation Properties
1. Cation Exchange Properties

In general, the alkali cations of layered alkali metal silicates
can be replaced by three classes of cations: (i) hydrated alkali metal,
alkaline earth or transition metal cations, (ii) organic cations and (iii)
hydronium ions. The proton exchange forms in most cases retain the
layered structure of the parent alkali metal silicate and, in addition,
possess interesting intercalation and ion exchange properties. The
various ion exchange and intercalation reactions are depicted in

Figure 1.7.

20

Alkali Metal Silicate

 

  
 

0 ever. 0 fe'e'e'e To
:10 :EeimQ‘OEeigeie ‘16:.

    
 
   

   
   

   

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 110+ 0+
Cationic . 32+ M“ ”3
Surfactant
4.
eg. CnHZn+lNHB
- -H H. H H
L' 111' “1‘1““ 5,, Silicic Acid
I if g 1 W.
‘ neat alkylamine
.- H l I l

 

 

 

 

 

 

 

 

 

 

 

 

1.7. Representation of some intercalation chemistry of the layered
alkali metal silicates (47).

21

The ion exchange of Na+ ions for NH4+, Li+, K+, Mg2+, Pb2+, Ag+,
C021", Ni2+ and Cu2+ in synthetic sodium octosilicate produced the
corresponding alkali, alkaline earth or transition metal silicate. Also
the exchange reactions were 85-10096 complete and no change in
crystallite shape or size occurred upon exchange, indicating the
reactions were topotactic (16). Synthetic Na+-magadiite also
exchanges sodium ions for other inorganic ions such as Ca2+, NH4+
and Ni“ (35). Owing to its ion exchange capability, synthetic Na+-

magadiite is an effective builder for low phosphate detergents (36).

KHSi205 (37), kanemite NaHSi205°5H20 (2), Na+-magadiite
NazSi14029-9H20 (13,38), kenyaite Na2Si20041-xH20 (l7), and sodium
octosilicate Na2Si3017-xH20 (16) all reacted with aqueous
alkylammonium chloride solutions to produce intercalated
derivatives composed of silicate layers expanded by RNH3+ ions. For
instance, KHSi20 5 reacted exclusively with aqueous solutions of n-
alkylammonium chlorides, RNH3+ with (R=CH3 to Can), to produce
intercalated onium ion derivatives with expanded basal spacings
(37). Kanemite, magadiite and kenyaite intercalated quaternary
ammonium cations such as dimethyldialkylammonium and
trimethylalkylammonium, in addition to n-alkylammonium cations.
Na+-magadiite, in particular, reacted with n-alkylpyridinium
(R=C10H21-C18H37) cations. Iler found that sodium octosilicate
intercalates both small and large chain quaternary ammonium
cations, but only a monolayer of cation formed between the layers

(16).

22

In general the basal spacing varies with the size, shape and
spatial orientation of the interlayer cation. In addition, the
configuration or conformation of the alkyl chain also effects the basal
spacing. The effect of chain conformation on basal spacing has been
studied previously for kanemite, magadiite and other layered
systems (2,38-41). The general orientation of the chains within the
interlayer depends on various factors such as the area of the alkyl
chain and the charge on the layer. A monolayer of cations oriented
parallel to the silicate layer will result if one half the unit cell basal
area, UBA/2 2 CA, where CA=(the cross sectional area of the cation) X
(# of cations /unit cell). If CA exceeds UBA/2 the orientation of the
alkyl chains will change from parallel to perpendicular or some tilt
angle in between. Figure 1.8. illustrates the change in interlayer
orientation for the alkylammonium ions as the alkyl chain length
and, hence the area of the intercalated ion, increases. After
perpendicular monolayers, further enlargement of the alkyl chain
results in a bilayer arrangement. Analogous interlayer structural
orientations were observed if the layer charge changed while

keeping the alkyl chain length constant.

For example, basal spacings for n-alkylammonium derivatives
of KHSi20 5, kanemite, magadiite and kenyaite increase linearly as the
alkyl chain length increases. Figure 1.9. shows a plot of basal spacing
versus alkyl chain length for these silicates with various types of
cations. N-alkylammonium intercalates of KHSi205 and kanemite
with carbon chain length between 7-10 exhibited basal spacings

characteristic of bilayer arrangements with tilt angles less than 90°.

23

 

 

 

 

 

 

 

 

1.8. Arrangements of n-alkylammonium cations within a layered
silicate. Cation chain length was varied while keeping layer charge

constant. a) CszNH3+, b) C8H17NH3+, 311d C) C12H25NH3+.

24

 

 

 

 

o magadiite D KH31205
<> kenyaite X kanemite

7O ,.. l 1 I l l I l I l 1 l 1 1 l I -i

E o 3

60 j o ‘-

. .

50 F O 1

r o 1

E 5 I

. 8 ..

i 40 ~— 0 '1
5 3 :
e 30 1- 6 —‘
'u ; 6 j
I [:1 I

20 j T

r 1

F '1

1° 2' c3 c: f

0 ” 1 1 1 J 1 1 L 1 1 1 1 1 1 1 1 ‘

4 6 8 10 12 14 16 18 20

No

1.9. Basal spacing of n-alkylammonium derivatives of layered

silicates vs number of carbon atoms in the alkyl chain for kanemitex
(2), KHSi205 U (37), magadiite 0 (l3), and kenyaite 007). All

spacings were determined for derivatives in equilibrium with RNH3+

solutions.

25

For kanemite, a change in chain length from 10 to 212 resulted in
perpendicular bilayer formation, as shown in Figure 1.9.
Dimethyldialkyl ammonium derivatives of kanemite exhibit swelling
or enhanced basal spacings to bilayer arrangements when brought
into contact with alkanol or alkyl amine reagents (2). Analogous
bilayer arrangements were observed for n-alkylammonium
montmorillonite derivatives solvated by n-alkylalcohols and n-
alkylamines (39). More recently, Fripiat et al. (20) have shown that
intercalation of either ethylammonium or heptylammonium ions
within KHSi205 resulted in the hydrolysis and partial condensation of
adjacent layers yielding an interstratified structure. The possibility
that similar hydrolysis reactions have occurred in alkylammonium—

kanemite intercalates has not been investigated previously.

The basal spacings of n-alkylammonium intercalates of
magadiite, when ploted versus carbon chain length, as shown in
Figure 1.9., indicated an increase of 2.6 A of basal spacing per carbon
atom in the chain. This result is indicative of a bilayer structure
within the interlayer with the alkyl chains oriented perpendicular to
the layer in an all trans configuration (13). When the kenyaite basal
spacings were normalized with respect to layer thickness, it became
apparent that the interlayer structure of kenyaite (17.7 A layer

thickness) was composed of bilayers with tilt angles < 90°.
2. Alkylammonium Derivatives as Reaction Precursors

Various long chain alkylammonium exchange forms of alkali

metal silicates have been used as precursors fOr subsequent

26

intercalation reactions. For instance, ethylammonium and
heptylammonium ion intercalated derivatives of KHSi205 have been
exchanged with A113O4(OH)24(H20)127+ solutions (42). The pillared
materials which resulted, although of low crystallinity, exhibited BET
surface areas in the range of 400 mz/g. Approximately two thirds of
this surface was attributed to micropores. Fripiat et a1. further
postulated that this material was composed of an interstratified
network of open pillared galleries and siloxane-bridged galleries. In
addition, a dodecyltrimethylammonium exchange form of magadiite
has been utilized as a precursor for the intercalation of acrylonitrile
monomer (43). Heat treatment of the resultant acrylonitrile-
C12H25N(CH3)3-magadiite complex at 1300°C in flowing Ar produced
p-SiC (43). Treatment of the same material at 1400°C under flowing
N2 resulted in p-Si3N3 formation (44). The interlayer SiOH groups in
alkylammonium derivatives of magadiite and kenyaite were found to
be reactive centers for the derivatization by silane reagents (45). For
example, dodecyltrimethylammonium complexes of magadiite and
kenyaite reacted readily with trimethylchlorosilane or
diphenylmethylchlorosilane to form intercalated derivatives (45,46).
29Si MAS NMR and 29Si CP-MAS NMR of these derivatives indicated

the presence of Q1R3Si-O type environments.
3. H+-Si|icates

Many layered alkali metal silicates reversibly exchange
interlayer cations for protons by ion exchange to form layered silicic

acids. In some cases the proton exchange form is favored even at

27

high pH (47). The reversibility of the proton exchange reaction also

allows for cation exchange of other alkali cations for protons (47).

The stability of alkali metal silicates towards acid has been
related to the size and charge of the alkali metal cation. For instance,
treatment of Li28i205 with acid did not result in loss of Li+ from the
silicate structure (48). Other layered silicates were found to be

unstable under acidic conditions. For example, the layer structures

of BaSi205 and K28i205 were altered to X-ray amorphous products

with general composition ZSiOonZO by acid treatment (48). In most

cases, however, the proton exchange form is stable and the layer

structure is retained ( cf. Table 1.1.).
4. Intercalation Complexes of H+-Silicates.

Layered silicic acids intercalate various polar organic guest
molecules. Table I.III. summarizes the various families of organic
molecules that have been found to intercalate in the galleries of
H28i205, H+-makatite, and H+-magadiite. However, not all silicic acids
possess the same reactivity. For instance, although all silicic acids
intercalate alkylamines, only the more highly condensed forms such
as H+-makatite, H+-magadiite, and H+-kenyaite imbibe other polar
organic molecules. The mechanism of intercalation involves
disruption of the hydrogen bonds which hold the layers together. In
order for this to occur the guest compound must possess a high
molecular dipole moment 2 3.5 Debyes (47). In addition, the polar
organic guest compound must have strong acceptor sites for

hydrogen bond formation.

28

Table I.III. Possible guest molecules for interlamellar adsorption by
crystalline silicic acids and the gallery height (47).

 

 

 

Group Examples Gallery Height (11)
H Si205 H-makatite H-magadiite
Short chain fonnamide . 1.7 2.4
amides dimethyl formamide 4.9 3.6 5.5
aoetamide 3.8 3.8 2.9
diethyl aoetamide 6.1 5.4 5.9
Urea and urea -- -- 4.3
derivatives N,N-dimethyl urea -- 5.2 5.3
N,N'-diethyl urea - 6.3 4.3
Sooxidcs dimethylsulfoxide 3.6 6.7 4.6
N -oxides trimethylamine—N-oxide - 3.7 4.7
pyridine-N-oxide 4.9 3.7 4.6
3-picoline-N-oxide 5. l 4. 3 5.6
Amatic pyridine 4.7 6.8 5 3
bases imidazole -- 6.1 3.2
pyrazine - 3.5 -
quinoline -- 10.7 6.1
Alkylamines hexylamine 17.1 18.4 18.4
decylamine 26.9 29.3 29.0
benzylamine - 12.7

 

--:no reaction

29

Another group of molecules which preferentially intercalate
into silicic acids are organic bases with pr < 9. In this case the
orientation of the interlayer molecules was discussed elsewhere

(2,17,37,38,47,49).

The layered silicic acids derived from, kanemite, KHSi20 5, or-
Na28i205 and Na+-magadiite all intercalate n-alkylamine molecules to
form structures similar to the n-alkylammonium derivatives
discussed earlier. The basal spacings for these derivatives also
increases in a linear fashion as the alkyl chain length increases.
Figure 1.10. indicates that all the layered silicic acid forms possess
approximately the same interlayer onium ion structures. The
interlayer structure in each case corresponds to bilayers of
alkylamine molecules with tilt angles of 2 60°. H+-magadiite
preintercalated with dimethylsulfoxide reacted with excess n-
alkylamines to form bilayer structures. Derivatives with alkyl chain
length S 10 resulted in chain tilt angles of 56° relative to the silicate
layer. Longer chains, 2 12, arrange themselves perpendicular to the
layer. The H+-magadiite intercalation complexes also retained ~ 2
moles of DMSO/ 14Si02 (38). Lagaly et al. (38) have reported
previously that the direct reaction of alkylamines and H+-magadiite

without preintercalation by dimethylsulfoxide resulted in only a

small increase in basal spacing (d001=14 A). The presence of
interlayer H20 in the H+-magadiite phase (d001=13.2 A) used by
Lagaly et a1. (38) may impede the intercalation of n-alkylamines.

However, related work reported that reaction of H+-magadiite,

dehydrated phase (d001=1l.2 A), with excess octylamine resulted in

30

 

 

 

 

0 H-magadllte D H-kanemite
0 H-KHSIZOS X H-aNaZSIZOS
70 I I I I I I I I I I I I I I I I I I I
’ ‘1
. o I
60 — O _.
Z O 1
. o .
50 :- O X .
: 0 O x :
40 ”' o X -
< . O Q g D ..
A _ O ..
5 30 - O 5 U _.
9, I 0 0 g 2
'° : g 5 :
20 r g; ‘.
. 0 5 :
10 E5 25 ‘:
0 l l J l l l l 1 1 l l l l l 1 1 l l
0 2 4 6 8 10 12 14 16 18 20
Nc

1.10. Basal spacing (d001 A) versus number of carbon atoms (NC) in
the alkyl chain for n-alkylamine derivatives of H+-Kanemite E] (2),
HZSiZOS-I (>(37), HZSiZOS-II X (49), and H+-magadiite O (38). All
basal spacings were derivatives under mother. liquid.

31

an intercalated derivative with 34 A basal spacing (50).
5. Structure of [Pt-silicates

In layered silicic acids, adjacent layers are connected by
hydrogen bonds between silanol groups or between silanol groups
and interlayer water molecules (51). Strong chemical bonding in the
two dimensions, which make up the layer, in combination with the
inert nature of the silicate layer towards acid allow for retention of
the layer structure during proton exchange. Crystal structures of
various disilicic acid exchange forms indicate, (c.f. Table 1.1.), that the
structures change only slightly after proton exchange. 29Si MAS NMR
data for acid-hydrolyzed KHSi205 (20) and kanemite, NaHSi205-3H20
(21) indicated that both materials were sensitive to hydrolysis. In

both cases Q3 SiOH groups from adjacent layers condensed to form Q4
Si(OSi)4 type environments. Similar NMR results were obtained for

11231205 derived from a-Na28i205 (52).

More highly condensed acids such as those derived from
makatite, octosilicate, magadiite and kenyaite have unknown
structures. The X-ray powder patterns for these silicic acids were
indexed by comparison with the parent alkali silicate (47). Small
changes in the a- and b- axes after proton exchange were revealed,

indicating some distortion of the silicate layers.

The X—ray powder diffraction pattern for H+-magadiite vary

depending on preparation. Both hydrated (d001=13.2 A) and
dehydrated (d001=11.2 A) forms of H+-magadiite have been obtained

by altering the titration conditions . H+-magadiite, like its sodium

32

analogue, exhibited 298i MAS NMR spectra indicative of both Q3
HOSi(OSi)3-type and Q4-Si(OSi)4 type environments (27,31). Various
workers have published Q4:Q3 ratios for H’r-magadiite that range
from 1.7-3.6, as shown in Table 1.11. The differences in Q3:Q4 ratios
can be attributed to inadequate relaxation of the spins associated
with the Q4 silicon environment as compared to the Q3 environment
(32). The characterization of H+-magadiite by infrared spectroscopy
and 1H NMR spectroscopies has indicated the presence of two types
of OH groups. As shown in Figure I.5B. one set of SiOH groups is

believed to be involved in strong hydrogen bonds between adjacent
layers (OHB) and the second set of SiOH have been attributed to

weakly hydrogen bonded groups (OH A) (53). The (OHA) groups point
towards tetrahedral ring holes in the adjacent layer. 1H NMR
indicated that the average distances between neighboring protons
was ~ 2.5 A. This small H-H distance required that the protons be
from adjacent layers. Thermal analysis of H+-magadiite in
combination with 29Si MAS NMR suggested that interlayer siloxane
bond formation occured even at 230°C (27). This contradicted an
earlier thermal study that indicated the onset of Si-O-Si bond
formation to be 400°C (29).

The titration of Na+-kenyaite suspensions with dilute acid (0.1
M HCl) resulted in exchange of Na+ by protons (17). The basal
spacing of the Na+-kenyaite, decreased from an initial value of 19.8
A to 17.6 A following proton exchange. A comparison of layer
thickness for H+-magadiite and H+-kenyaite indicated that kenyaite

possessed a thicker silicate layer, 11.2 A versus 17.6 A respectively.

33

6. Silylation of Interlamellar Silanol Groups

The interlayer SiOH groups of H+-magadiite, H+—kenyaite, and
on-stizO 5 are accessible, for reaction, with silane grafting reagents
when the galleries were preintercalated by polar organic molecules
(54). Preintercalation of the silicic acid separates individual layers,
thus allowing the organosilane access to the interlayer SiOH groups.
Direct reaction of layered silicic acids with trimethylchlorosilane
resulted in grafting of (CH3)3-Si groups to the external surfaces only.
However, as illustrated in Figure 1.11., preintercalation of the layered
silicic acid by either dimethylsulfoxide (DMSO), N-methylformamide,
or N,N-dimethylformamide followed by reaction with either
hexamethyldisilazane or trimethylchlorosilane results in the
silylation of interlayer SiOH groups. X-ray powder diffraction of
these materials indicated gallery heights of ~ 8 A. Infrared
spectroscopy and chemical analyses indicated that the silylation
reaction was accompanied by displacement of the preintercalated
polar organic molecules (55). The 29Si MAS NMR spectra of H-
magadiite treated in this fashion indicated the presence of (CH3)3Si-
environments. Also, from these data it was estimated that 30% of
the SiOH groups present in H+-magadiite had reacted with

organosilane (31).
7. Pillared Layered Silicates

Various authors have tried to generate pillared materials by

utilizing the intercalation chemistry and intracrystalline reactivity of

34

7W , Z/%

on
B
on on
I. |_
A 1" 0.. [MS

 

1_ 1_

1.11. Interlamellar grafting of trimethylsilyl groups into the
intracrystal surface of silicic acids by reaction of
trimethylchlorosilane (TMCS) with the N-methylformamide (NMF)
solvated form of the silicic acid. The unsolvated silicic acid is
unaccessable for reaction with TMCS (55).

35

the layered silicic acids. For example, Sprung et al. (56) reported on
the pillaring of H+-magadiite with polyhedral silicate species. The
polyhedral three dimensional silicate species used as pillaring
reagents were synthesized by cyclic polymerization of either
phenyltrichlorosilane or cyclohexyltrichlorosilane. The direct
reaction of H+-magadiite with these pillaring reagents resulted in
partially pillared materials. The resultant products when calcined at
350°C exhibited surface areas between 100-200 m2/g and basal
spacings of ~ 15.6 A.

More recently Landis et. al. (57,58) expanded on Rojo and Ruiz-
Hitzkys work (54,55) and produced pillared microporous derivatives
of H+-magadiite, H+-makatite and H+-kenyaite. The pillaring
procedure they employed utilizes a pro-swelling step which
separates the layers and expands the interlayer space allowing
access for siloxane pillaring reagents. For example, preintercalation
of H+-magadiite by octylammonium ions or octylamine molecules
from aqueous solutions followed by treatment with
tetraethylorthosilicate, TEOS, resulted in the intercalation and
hydrolysis of TEOS between the layers of H1. -magadiite. After
calcination at 538°C in air for several hours this material exhibited a

nitrogen BET surface area of 530 m2/g.

10.

ll.

12.

13.

14.

36

LIST OF REFERENCES

I. M. Timoshenkov, Y. P. Menshikov, L. F. Gannibal and I. V.
Bussen, Zapiski: W. B17, 104 (1976).

K. Beneke, and G. Lagaly: W62, 763 (1977) .
R. A. Sheppard and A. J. Gude: Wm. 55, 358 (1970).

J. L. McAtee, R. House and H. P. Eugster: W53, 2061
(1968).

A. J. Gude, and R. A. Sheppard: W57, 1053 (1972).

K. Beneke and G. Lagaly: W12, 363 (1977).
F. Leibau: Aim—Quilt“, 395 (1961) .

I. Williamson and F. P. Glasser: W7, 127
(1966).

H. Schweinsberg and F. Liebau: W387, 241
(1972).

F. C. Kracek, N. L. Bowen and G. W. Morey: W33,
1857 (1929).

C. L. Baker, L. R. Inc and J. H. Wills: W72. 5369
(1950)

R. Wey and A. Kalt: W26 ser D, 1 (1967).

G. Lagaly, K. Beneke, and Armin Weiss: Proc. Int. Clay Conf.
Madrid 1972 p.663-673,Madrid 1973

R. A. Fletcher, and D. M. Bibby: W. 35, 318 (1987).

15.
l6.
17.
18.

19.

20.

21.
22.
23.

24.

25.

26.

27.

28.

29.

30.

31.

37
L. McCulloch: W74. 2453 (1952).

R. K. Iler: W 19, 648 (1964).
K. Beneke, and G. Lagaly: Wing. 68, 818 (1983) .

K. Beneke and G. Lagaly: AM”, 224 (1989).

K. Beneke, H.-H. Kruse and G. Lagaly: W518,
65 (1984).

2. Q. Deng, J. F. Lambert and J. J. Fripiat: W5 1,
375 (1989).

I. D. Johnson: Ph. D. Thesis Michigan State University, (1985).
H. Annehed, L. Falth, F. J. Lincoln: W- 159, 203 (1982).
H. Schweinsberg and F. Liebau: M. B30, 2206 (1974).

J. M. Garces, S. C. Rocke, C. Crowder, and D. L. Hasha: W
Minn“, 409 (1988).

W. Schwieger, D. Heidemann and K. H. Bergk: W
W22, 639 (1985).

G. W. Brindley: W54, 1583 (1969).

T. J. Pinnavaia, I. D. Johnson, and M. Lipsicas: 1, Solid. State,
M63, 118 (1986).

A. Brandt, W. Schwieger, K. H. Bergk, P. Grabner, and M. Porsch:
W24. 1 (1989)-

G. Lagaly, K. Beneke: W60, 650 (1975).

J. M. Rojo, E. Ruiz-Hitzky and J. Sanz: M27, 2785
(1988).

J. M. Rojo, J. Sanz, E. Ruiz-Hitzky, J. Serratosa: ZAML—AHL
m540/54l, 227 (1986).

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

38

J. S. Dailey and T. J. Pinnavaia: Submitted to J. Inclusion
Phenomena

A. Brandt, W. Schwieger, and K. H. Bergk: W
Minerals“. 564 (1937)-

R. K. Harris and G. J. Nesbitt: M78, 245 (1988).

W. Schwieger, K. Bergk, K. P. Wendlandt, and W. Reschetilowski:
Z._Ch§m.25, 228 (1985).

K. Bergk, W. Schwieger, P. Renger, P. Martin, D. Kaufmann, R.
Dpitz, and D. Muth: Patent DD 221198 A1.

A. Kalt, B. Perati, and R. Wey: ”International Clay Conference
1978; M. M. Mortland, and V. Farmer eds., Elsevier: Amsterdam,
1979; p501.

G. Lagaly, and K. Beneke: W60642 (1975).

A. Weiss: W2. 134 (1963).
G- Lagaly: W22. 43 (1986).
G- Lagaly: W15. 575 (1976).

Z. Deng, J. F. H. Lambert and J. J. Fripiat: Water, 1,
640 (1989).

Y. Sugahara, K. Sugimoto, T. Yanagisawa, Y. Nomizu, K. Kuroda,
and C. Kato: W195, 117 (1987).

Y. Sugahara, Y. Nomizu, K. Kuroda and C. Kato: W 95,
124 (1987).

T. Yanagisawa, K. Kuroda and C. Kato: MS, 167
( 1988).

T. Yanagisawa, K. Kuroda and C. Kato: W61,
3743 (1988).

G- Lagaly: WWII. 105 (1979).

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.
60.

61.

62.

63.

39
F. Wodtcke and F. Liebau: W335, 178 (1965).

G. Lagaly, K. Beneke, P. Dietz, and A. Weiss: mm
W13, 819 (1974)

J. S. Dailey and T. J. Pinnavaia: in preparation.

G. Lagaly, H. M. Reckert and H. H. Kruse: ”Crystalline Silicic
Acids” in Chemical Reactions in Organic and Inorganic
Constrained Systems eds. R. Setton, Reidel (1986).

D. Heidemann, A. R. Grimmer, C. Hubert P. Starke and M. Magi: Z_,
W528. 22 (1985)-

J. M. Rojo, E. Ruiz-Hitzky, J. Sanz and J. M. Serratosa: m
ChimiLMinmlem, 807 (1983).

E. Ruiz Hitzky and J. M. Rojo: W287, 28 (1980).

E. Ruiz I-Iitzky, J. M. Rojo, and G. Lagaly: W521.
263, 1025 (1985).

R. Sprung, M. E. Davis, J. S. Kauffman and C. Dybowski: Mug,
W29, 213 (1990).

M. E. Landis, P. Chu, I. D. Johnson, G. W. Kirker, and M. K. Rubin:
U. S. Patent 4 859 648, (1989)

M. E. Landis, B. A. Aufdembrink, P. Chu, I. D. Johnson, G. W.
Kirker, and M. K. Rubin: W113, 3189 (1991).

F. Liebau: mailman. 120, 427 (1964).

Y. Hubert, A. Kalt, J. L. Guth and R. Wey: W282
Set. D., 405 (1976).

Z- Johan and G. F- Maglionc: MW 95.
371 (1972)

A. K. Pam and D. W. J. Cruickshank: W324, 13 (1968).

A. x. Pant: W324, 1077 (1968).

40

64. M. T. LeBihan, A. Kalt and R. Wey: WM
Mallow“. 15 (1971)-

65. F. Liebauzmmmmgnlit, 389 (1961).

CHAPTER II

Intercalative Reaction of a Cobalt(III) Cage Complex,
Co(sep)3+, with Magadiite, a Layered Sodium Silicate.

A. Introduction

Pillared layered materials have attracted widespread interest
recently due in part to their potential utility as catalysts and
molecular sieving adsorbents [1,2]. Normally, pillared microporous
solids are derived from materials with ion exchange capabilities.
Examples of layered ionic materials that have been examined
previously for their pillaring properties include, 2:1 phylosilicate
clays, layered double hydroxides and zirconium phosphates[1].
Another class of ionic layered minerals, namely, the hydrous sodium
silicates have been identified as having intercalation properties
suitable for producing pillared derivatives [2]. In an effort to
broaden the diversity of pillared materials, we have been
investigating the pillaring reactions of lamellar hydrous sodium

silicates. Examples of this unique family of silicates include,
magadiite (Na28i14029-9H20), kenyaite (N328i20041'10H20)' and

kanemite (NaHSi205-3H20). All are naturally occurring minerals first
found in lake beds at Lake Magadi, Kenya [3]. These compounds also
can be conveniently prepared in the laboratory via hydrothermal
synthesis [4-6]. It has been shown that the hydrated interlayer

sodium ions of these materials can be exchanged by large

41

42

alkylammonium cations to form organic derivatives [5-7]. These
alkylammonium derivatives of magadiite and kenyaite have been
used as precursors for reaction of interlayer silanols by organo

silanes [8].

Certain alkylammonium exchange forms of smectites show
considerable promise as selective adsorbents for the removal of
priority pollutants from contaminated ground waters [9,10]. Similar
adsorption properties might be expected for organo cation exchange
forms of layered silicates. However, in contrast to smectite clays,
layered sodium silicates have a relatively high layer charge and are
not easily swellable. The highly charged layers require highly
charged cations to induce porosity. No examples of magadiite
intercalated by a robust cation of charge 3* or larger have been

reported previously.

In this work, we examine the possibility of preparing a pillared
derivative of magadiite using a high charged cation of known
dimension. The cation chosen for this purpose was cobalt
sepulchrate, [Co(sep)]3+, a highly stable cobalt cage complex derived
from the capping of trisethylenediaminecobalt(III) cation [11]. The

structure of this complex is illustrated in Figure 1.1.
B. Experimental
1. Preparation of Na+-magadiite

Synthetic Na+-magadiite was prepared by the reaction of NaOH
and Si02 under hydrothermal conditions according to a previously

43

 

 

 

Co(sepulchrate) 3 +

11.1. Schematic representation of the cobalt sepulchrate complex.

44
described method [4]. A suspension of Davisil 62 SiOz (12.0 g, 0.20

mol) in 60 mL of 1.11 M_ NaOH (0.067 mol) was allowed to digest at
150°C for 46 hours in a Teflon-lined stainless steel bomb. The solid

Na+-magadiite reaction product was separated by centrifugation,
washed twice with 200 mL of deionized H20 in order to remove

excess NaOH and air dried at 40°C.

2. Preparation of Co(sep)Cl3

The synthesis of (l, 3, 6, 8, 10, 13, 16, 19
octaazabicyclo[6.6.6]eicosane)cobalt(III) chloride, [Co(sep)]Cl3, was

accomplished by using the method described by Creaser et al. [11].
To a stirred suspension of [Co(en)]C13-H20 (18.0 g, 0.05 mol) and
Li2CO3 (25 g, 0.34 mol) in deionized H20 (125 mL) were added 37.9%
aqueous formaldehyde (550 mL, 7.5 mol) and 30% aqueous NH4OH

(157.4 mL, 2.5 mol diluted to 550 mL). The solutions were added
separately by addition funnel over a period of three hours. The

mixture was allowed to stir for an additional 30 minutes, and Li2CO3

was filtered from the red solution. The pH of the filtrate was

adjusted from 9 to 2.5 by the addition of concentrated HCIO4, and the

LiClO4 that formed was filtered from the solution. The filtrate was
pasSed through a Dowex 50W-X8 column (100-200 mesh), and the
column was washed with five liters of 0.2 M trisodium citrate. The
resin was then washed with H20 and 1 M_ HCl to remove Na+. The
resultant orange Co(sep)3+ cation was eluted from the column with 3
M_ HCL (~ 4 liters). The orange solution was reduced to dryness on a

rotoevaporator and recrystallized from water by the addition of

45
acetone. Anal Calcd for: CoC12N8H3oCl3: Co, 13.05; C, 31.91; N, 24.81;

H, 6.69; CI, 23.54. Found: Co, 12.41; C, 30.76; N, 23.72; H, 6.96.
3 . Co(sep)3+-magadiite Reactions

Co(sep)3+ exchange forms of magadiite were prepared by the
addition of synthetic Na+-magadiite (0.50 g, 4.7rt10'4 mol) in 10 mL
of deionized H20 to c:o(sep)c:13 (0.914 g, 2.02rt10'3 mol) in water (40
mL). Reaction mixtures were maintained either at 25°C for 72 hours
or at 100°C for 24 hours. After each reaction period, the solid was
centrifuged and reexchanged with 50 mL of fresh 0.04 M Co(sep)Cl3
solution. The exchange solutions were replaced up to four times.

Each reaction product was recovered by centrifugation and washed
with deionized H20 until the orange solid was Cl' free.

4. Physical Measurements

Basal spacings were determined from 001 X-ray reflections
using a Rigaku Rotaflex diffractometer equipped with Cu Ka
radiation. Samples were prepared by depositing a suspension of the
washed solid on a glass microscope slide and allowing the suspension
to air dry at 40°C. The extent of intercalation was followed after

each exchange by measuring the change in basal spacing.

29Si MAS NMR experiments were performed on a Varian 400
VXR solid state NMR spectrometer operated at 79.5 MHz. A Doty
multinuclear MAS probe equipped with zirconia rotors was used for
all measurements. 29Si spin lattice relaxation times were determined

experimentally by inversion recovery experiments. The 298i

46

relaxation time and 90° pulse width for Na-magadiite were found to
be 280 seconds and 5.8 us respectively. The Q3 and Q4 silicon
environments exhibited relaxation times of 160 seconds and 280
seconds respectively. A total of 12 scans were recorded for each
sample. The sample spinning rate was approximately 5 kHz. A delay

time five times as large as T1 for the Q4 sites was used in order to

obtain quantitative results.

Nitrogen adsorption isotherms at liquid N2 temperature were

determined on a Quantachrome Quantasorb Jr. using ultra pure N2 .as

the adsorbate and Helium as the carrier gas. All samples were
outgassed at 120°C under flowing He for 12 hrs. Surface areas were

determined using the BET equation and the t-plot method [12].

FTIR spectra were obtained by use of an IBM IR44
spectrometer and the KBr pressed pellet technique. All pellets

contained 2% sample by weight.

Thermogravimetric analyses were performed using a Cahn TG
system 121 analyzer. All samples were heated to 1000°C at a

heating rate of 5°C/min.
C Results and Discussion
1. Preparation of Na+-magadiite

Base hydrolysis of silica gel at 150°C under autogenus pressure
produced a well crystallized Na+ -magadiite product. The X-ray
diffraction pattern for an air dried film sample, shown in Figure 11.2.,

exhibited 001 reflections corresponding to a basal spacing of 15.6 A.

47

 

 

 

 

 

 

15.6%
>.
:‘i
tn
1':
o
«in!
E
m
>
O;
2
0
CK
¢ 7.7x 5.21
III'III'II—IIIII'III'III' III III III III Ifi

 

2 6101418222630343842
Degrees 2 Theta

11.2. X-ray diffraction pattern (Cu-Kat) of an oriented film sample of
Na+-magadiite.

48

Despite the planar structure of Na+-magadiite and efforts to produce
well oriented films, several hkl reflections with h¢k¢0 were
observed due to the rosette-like morphology of the aggregated
platelets. The broad bump in the X-ray pattern centered near 20
degrees 26 (d=4.43 A) may be indicative of a quasi-crystalline silica
impurity. The peak positions for this synthetic product agreed
closely with values reported previously for synthetic and natural

magadiite [3,13-15].

The sodium content of synthetic Na+-magadiite depended in
part on the extent of hydrolysis that occured during the washing
procedure. Initial washes removed excess NaOH associated with the
material. However, repeated washes leached Na+ cations from the
interlayer and caused layer hydrolysis. The chemical composition of
Na+-magadiite was obtained by combining the results of
thermogravimetric analyses and chemical analyses. Air dried Na+-
magadiite that had been well washed to remove excess NaOH, Figure
II.3a., lost 13% of its total weight as water below 200°C. An
additional 1.6% was lost between 200°C- 1000°C. The weight loss
above 200°C was assigned to the dehydroxylation of SiOH groups.
The combination of NaZO content (4.98%) and weight loss provided

an empirical composition for synthetic Na+-magadiite of
Nal.78i1402.7.9(OH)1.9 7.6 1120. These data agreed well with those put

forth by Lagaly et a1. [7] and Garces et a1. [14] that indicated an ideal

unit cell of NaZSi 91120. Other workers report compositions

14029
shown in Table [1.1. that vary between 2.13 and 1.96 sodium cations

per 14 silicon atoms. The samples synthesized in this work (cf.,Table

49

 

Weight (Z)

 

 

 

82

 

 

r100 '2éor 400' 550' 700' '850' 1000

Temperature (°C)

11.3. Thermogravimetric analysis curves obtained under flowing

argon for samples (A) Na+-magadiite, (B) Co(sep)3+-mag(25°C) and (C)
Co(sep)3+-mag(100°C).

50

Table 11.1. Analytical Composition of Synthetic and Natural Magadiite

 

 

Sample
Na

Na+-magadiite,synthetic 6.58
unwashed, this work (1.00)
N a+-magadiite, synthetic 4.98
washed, this work (1.00)
Na+-magadiite, synthetic 5.58
Lagaly (1975) (1.00)

Na+-magadiite, natural 5.6
McAtee (1968) (1 -00)

N a+-magadiite, synthetic 6.56
Garces (1988) ( 1.00)

N a+-magadiite, synthetic 6.31
Garces (1988) ( 1.00)
H+-magadiite, synthetic >.01

washed,this work

 

Weight Percent Atomic Ratios

(moleratio)

Si H Total Na Si HzO_
75.8 15.5 97.88 2.35 14 10.9
(11.9) (8.11)

80.2 14.6 99.78 1.69 14 8.5
(16.6) (10.0)

74.9 18.2 98.68 2.02 14 11.3
(13.4) (3.26)

77.4 15.2 98.2 2.13 14 9.2
(13.8) (2.71)

84.8 7.8 99.16 2.09 14 4.29
(12.9) (1.19)

87.0 8.1 101.4 1.96 14 4.35
(14.2) (4.42)

93.9 6.1 100 0 14 2.0

51

11.1.) straddle these compositions, depending on post hydrothermal
treatment of the product, especially the number of washings. The
water content also varied dramatically depending on differences in
drying conditions. The low sodium content of our washed product,
1.69 Na+ per 14 Si, was a direct result of hydrolysis caused by the

extensive washing with water.
2. Reaction with Co(sep)3+

The basal spacings of the products obtained by ion exchange
reaction of Co(sep)3+ with Na+-magadiite depended greatly on the
reaction temperature. For reactions carried out at 25°C, there is little
or no change in the basal spacing of the host material as shown by
the XRD patterns in Figure 11.4. After three treatments with 0.12N
Co(sep)“, the magadiite basal spacing increased slightly from an
initial value of 15.6 A to ~15.9 A. The 001 peak became less
symmetric with a possible shoulder developing on the low angle side
of the diffraction peak, and the hkl reflections in the region 24-30°

26 decreased in intensity.

The color of the reaction product obtained at 25°C was orange
indicating that some Co complex was bound to the magadiite surface.
Despite the change in diffraction pattern and the color change, it was
unlikely that significant intercalation had occurred. The change in
basal spacing was far too small to be directly attributed to
intercalation of Co(sep)3+. The change in hkl intensities in the 24-30°
region most likely was due to a breakdown in the rosette

morphology of the magadiite as a result of prolonged stirring of the

52

 

15.63

 

1st exchange

 

Relative IntenSIty

2nd exchange

 

    

 

3rd exchange

 

T” ' 1 ‘ I ' I ' 1 ' 1
2 6 10 14 18 22 26 30 ‘34 38
Degrees 2 Theta

 

 

11.4. X-ray diffraction patterns for oriented film samples of the

reaction products obtained after successive exchange of Na+-
magadiite with 0.12 N Co(sep)Cl3 at 25°C.

53

crystals during the exchange reaction. Multiple exchange reactions at
25°C most likely resulted in the Co(sep)3+ being bound at the
external surfaces of the crystals. Chemical analysis, to be presented

below, verified this conclusion.

In contrast to the results obtained at 25°C, the successive
treatments of Na+-magadiite with fresh 0.12 N [Co(sep)]Cl3 solutions
at 100°C resulted in the steady increase in the basal spacing from
15.6 A before exchange to 17.6 A after the 3rd exchange reaction, as
shown in Figure 11.5. Further exchange did not result in additional
expansion of the structure. The increase in basal spacing indicated
intercalation of [Co(sep)]3+ between the layers of magadiite. On the
basis of an 11.2 A thickness for magadiite in the absence of hydrated
Na+ ions, the observed spacing corresponded to a gallery height of -
6.4 A. This value was in agreement with the gallery height reported
for smectite clays pillared by Ir(diamsar)3+, an oval-shaped metal

cage complex with dimensions similar to Co(sep)3+ [16].

The compositions of the products obtained by ion exchange
reaction of Na+-magadiite with Co(sep)3+ at 25°C and 100°C were
deduced from TGA and chemical analysis. TGA curves for the
reaction products produced at 25°C and 100°C, henceforth designated
Co(sep)3+-Mag(25°C) and Co(sep)3+-Mag(100°C), respectively, are
shown in Figure II.3b. and II.3c. Included in the Figure for
comparison purposes are the TGA curves obtained for Na+-magadiite
before exchange with Co(sep)3+. The product produced at 25°C

exhibited an initial weight loss of 8.5 wt % between 20°C - 200°C
assigned to adsorbed H20. In addition, the 2.0 wt % loss from

54

 

 
 

1 6.68
2nd exchange

  

1 7.511

 

 

3rd exchange

W

17.63

M | 4th exchange

I'I'I'I'I'I'I'I'I'lf
2 6101418222630343842
Degrees 2 Theta

 

CT

Relative Intensity

 

 

 

 

 

11.5. X-ray diffraction patterns for oriented film samples of the

reaction products obtained after successive exchange of Na+-
magadiite with 0.12 N Co(sep)Cl3 at 100°C.

55

200°C - 350°C was attributed to dehydroxylation. Above 350°C the
weight loss (4.5 %) was probably due to the oxidation by air of the
sepulchrate ligands of Co(sep)3+. Elemental analyses, given in Table
11.11., indicated a total of 4.4 wt % percent carbon, and nitrogen, in

good agreement with the TGA results.

The product prepared at 100°C exhibited a weight loss of 8.0 %
below 350°C ,which we assigned to the loss of H20 and hydroxyl
groups. The additional 10 % weight loss from 350°C-1000°C was
probably due to the loss of sepulchrate ligand by desorption and/ or
oxidation. Elemental analyses for total carbon and nitrogen was (9.8
wt %) in close agreement with the 10 % weight loss observed by TGA
above 350°C.

On the basis of the data summarized in Table 11.11. for
Co(sep)3+-Mag(25°C), there were approximately 0.18 cobalt atoms {per
Sil4 unit. The N/Co molar ratio was 6.63, in reasonable agreement
with the expected N/Co molar ratio of 8 for Co(sep)3+. A substantially
lower N/Co ratio of 4.1 was obtained for Co(sep)3+-Mag(100°C). The
C/N molar ratio was 1.6, in close agreement with the expected C/N
molar ratio of 1.5 for Co(sep)3+. This suggested that Co(sep)3+ cations
and uncomplexed cobalt cations were present in this material.
Cobalt(11), the most stable uncomplexed form of cobalt in aqueous
solution, is known to form a blue complex with SCN' in ethanolic
solution [17]. The supernatant recovered after leaching Co(sep)3+-
Mag(100°C) with 6 M HCl became deep blue when ethanolic KSCN was
added, indicating the presence of Co(II). However, Co(sep)3+-

Mag(25°C), gave a negative SCN' test for Co(II), indicating that

56

Table 11.11. Compositions (wt %) of Products Prepared by Reaction of
Na+-magadiite with Co(sep)3+ at 25°C and 100°C.

 

Co(sep)3+-Mag(25°C) Co(sep)3+-Mag(100°C)

 

 

N820 6.1 0.3
$02 82.7 80.5
Co 1.0 4.2
C 2.7 5.6
H 1.97 2.35
N 1.6 4.2
H20 a 8.5 8.0
Total 102.6 103.3
Chemical A B
Formula

 

A. 110.46861’.'(,[<:o(sep)10333511 4o29 -l.1 H20,

B- Ho.loc°0236[C°(5°P)103393114029'1AHZO
1'Weight loss below 200°C from TGA.

57

essentially all the cobalt was bound as the Co(sep)3+ complex. Thus,
the composition of Co(sep)3+-Mag(25°C) was consistent with Co(sep)3+
present mainly at the external surfaces of the layered silicate.
Whereas Co(sep)3+-Mag(100°C) was an authentic intercalation
compound that contained a mixture of Co(sep)3+ and uncomplexed
Co(II), formed as a result of demetalation of Co(sep)3+.

The 29Si MAS NMR spectra for Na+-magadiite, Co(sep)3+-
Mag(25°C) and Co(sep)3*-Mag(100°C) are shown in Figure 11.6. Na“-
magadiite exhibited two general Si environments, namely Q3 type
HOSi(OSi)3 or Na+[OSi(OSi)3] sites and Q4 type Si(OSi)4 sites [2]. A
single Q3 peak was centered at -99.8 ppm and three Q4 resonances
occured at -110.7 ppm, -111.7 ppm and -114.2 ppm respectively.

The ratio of integral intensities for Na+-magadiite, Q3/Q4=0.35,

corresponded to 3.63 Q3 sites/Si”. These data were in close
agreement with the Q3/Si14 ratio obtained from the analytical

composition (Q3/Si14=3.6).

The Q3 and Q4 resonances in Co(sep)3+-Mag(25°C) exhibited
chemical shifts similar to those found for Na+-magadiite, but the lines
are somewhat broadened. For instance, the shoulder present at -
110.7 ppm in Na+-magadiite was unresolved in Co(sep)3+-Mag(25°C)
due to line broadening. The increase in linewidth was even more
apparent in Co(sep)3+-Mag(100°C). The presence of paramagnetic
Co“ in the latter product undoubtedly contributed to the spectral
broadening. A comparison of relative integral intensities for the 29Si

resonances indicated that the Q3/Q4 ratio was very similar for Nat

58

412.1

 

 

410.7

414.2

 

 

 

 

10c

-92 -98 -106 -114 pp.

 

11.6. 29Si Magic Angle Spinning NMR spectra for (A) Na+-magadiite,
(B) Co(sep)3+-mag(25°C), and (C) Co(sep)3+-mag(100°C).

 

59
magadiite (Q3IQ4=0.35), Co(sep)3+-Mag(25°C) (Q3/Q4=0.27), and
Co(sep)3+-Mag(100°C) (Q3/Q4=o.26). These data confirmed that the
layer structure of magadiite remained intact after intercalation of
[Co(sep)]3+. Earlier workers have published Q3/Q4 ratios for Na+-
magadiite that range from 0.254 [2,15,18]. However, a careful

2S’Si environments

evaluation of the relaxation times for the various
has not been undertaken until now. We have found that the spin-
lattice relaxation times for the Q3 site is 160 s and ~ 280 s for the Q4
sites. The differences in Q3 and Q4 relaxation times can be attributed
to the presence of H20 in the interlayer. The importance of 29Si-IH
dipolar relaxation via water molecules has been shown for other
layered materials [19]. The close proximity of interlayer water to the
Q3 29Si sites of Na+-magadiite allows for their relaxation via 29Si-lH

dipolar relaxation and, in turn, reduces the relaxation time of the Q3

site greatly as compared to the Q4 site.

Infrared spectra for Na+-magadiite and the Co(sep)3+ reaction
products are shown in Figure 11.7. The IR spectrum for Na+-
magadiite, Figure(11.7c), agreed well with those reported elsewhere
[15,20]. The bands centered at 3500 cm'1 and 1650 cm'1 have been
assigned previously to the stretching and bending frequencies,
respectively, of intercalated water (20). The bands from 1500 cm'1
- 400 cm'1 have been attributed to the stretching and bending
frequencies of the SiO4 units that make up the layer (15). Garces et

a1. [15] have assigned the bands atabout 1225 cm'1 to five
membered rings of SiO‘4 tetrahedra that are present in certain

zeolites.

60

 

 

 

 

 

 

 

 

 

1 225
., l
0
C
8
6 2650 1 ‘65
8 . c 1
<2
G)
.2
4.!
2
0
0:

BO

 

A _AL_

IIII IIII IIII'IIII IIII IIII IIII

4000 3500 3000 2500 2000 1500 1000 500
Wavenumbers (cm")

 

 

 

 

11.7. Infrared spectra (2% w/w KBr pellet) of (A) Na+-magadiite, (B)
Co(sep)3+-mag(25°C), and (C) Co(sep)3+-mag(100°C).

61
The IR spectrum of Co(sep)3+-Mag(100°C) exhibited all the

peaks characteristic of magadiite, along with weak absorptions from
Co(sep)“. Weak broad 1R bands at 3100 cm"1 , 2960 cm'1 and 2850

1

cm' were assigned to the C-H stretching frequencies of the ligand.

1 was characteristic of the C-H

Also, a weak absorption at 1465 cm'
asymmetric bending modes of Co(sep)“. The characteristic
magadiite absorptions in the 1500 cm'1 - 400 cm'1 region remained
unchanged upon intercalation of Co(sep)3+ cation, indicating that the

layered structure was not changed during intercalation.

Nitrogen adsorption isotherms over the partial pressure range
of 0< P/Po < 0.4 were obtained at 496°C for Na+-magadiite, and the

)3+

Co(sep reaction products. The data were treated by using the BET

equation [12] in order to obtain surface area values. The equivalent
surface areas were determined from the BET monolayer volume Vm
(BET), and the microporous volumes determined from the t-plots (see
Table II.111.). A very small microporous surface area, (7 mzlg), was
observed for Co(sep)3+-Mag(25°C). The virtually non-existent

)3+ intercalation

microporosity was consistent with little or no Co(sep
for the reaction product obtained at 25°C. Somewhat more
surprising was the low microporous surface area for Co(sep)“-
Mag(100°C), (8.7 mzlg). Although Co(sep)3+ had been intercalated,
apparently insufficient space was available to allow access to the
nitrogen molecule (kinetic diameter 3.6 A). That is, the mixed
Co(sep)3+ and Co“ intercalation product appeared to be "stuffed”

with little free volume between the metal complex pillars.

62

Table 11.111. Surface Area Analysis of Intercalated Derivatives

 

 

Sample 3357 Vm cma liq 9'1 Nonmicroporous
ng-l Surface ng;
Na-magadiite 24 O 24
H-magadiite 45 0 45
Co(sep)-Magadiite 33 0.0025 24
25°C (7 m29-1)
Co(sep)-Magadiite 57 0.0031 48
100°C (8.7 m29-1)

 

Na+-magadiite can adopt a particle morphology composed of
silicate layers intergrown to form spherical nodules resembling
rosettes [15]. The proton exchange form of magadiite also has this
characteristic particle morphology [7]. Scanning electron
micrographs of the Co(sep)3+ reaction products, shown in Figure
11.8b., and 11.8c., revealed the same morphology as the starting Na+-
magadiite, Figure 11.8a., except that the size of the rosettes were
reduced by the attrition caused by vigorous stirring during reaction.
This result indicated that the Co(sep)3+ in Co(sep)3+-magadiite(100°C)
was intercalated in a topotactic fashion. Close inspection of the

micrographs also indicated the absence of other crystalline phases.

The demetalation of Co(sep)3+ that occurred upon reaction of
the complex with Na+-magadiite at 100°C is particularly noteworthy.
Sargeson and his coworkers have reported previously that it was not

possible to remove the metal from the Co(III) complex [11].

 

63

11.8A Scanning electron micrographs at X6600 and X1500
magnification for Na+-magadiite

 

1.8U CEUS8

1
B
B
B

lSKU X6888

 

XUW

65

11.8B. Scanning electron micrographs at X6600 and X1500

magnification for Co(sep)3+-mag(25°C)

 

 

67

11.8C. Scanning electron micrographs at X6600 and X1500
magnification for Co(sep)3+-mag(100°C)

 

 

X1500

 

 

 

 

69

)3+ is very stable and remains intact even after prolonged

Co(sep
treatment in boiling 12 M HCl. Elemental analyses suggested that
during the intercalation of Co(sep)“, one half of the metal centers
were no longer complexed by the sepulchrate ligand and was
confirmed by qualitative analysis for Co(II). The stability of
Co(sep)Cl3 was tested in basic solution under reaction conditions
identical to those used to obtain the 100°C intercalation product.
Under these conditions no Co(II) was detected using the SCN'
qualitative test. Demetalation seemed to be a direct result of the
interaction of Co(sep)3+ with magadiite. Treatment of Co(sep)2+ with
acid leads to release of Co(II) from the sepulchrate ligand [11], but an
analogous reaction does not occur for the 3* cation. The mechanism
of decomplexation on the magadiite surface may have involved
reduction of the Co(sep)3+ complex by water followed by rupture of
the cage and release of Co(II). The ovalshaped Co(sep)3+ is oriented
with with its long axis parallel to the silicate sheet, as judged from
the 6.4 A gallery height obtained by X-ray diffraction. Magadiite
might facilitate the reduction of Co(sep)3+ by distorting the bicyclic

complex in the intercalated state.

The relatively sharp X-ray diffraction pattern observed for
Co(sep)3+-magadiite(100°C) indicated that the material was regularly
intercalated. That is, both the Co(sep)3+ and Co“ cations appear to
be mixed within each interlayer in a uniform manner. A previously
proposed model structure for Na+-magadiite suggested that the
magadiite layer has a corrugated surface [15]. A plausible

arrangement in this case could involve placement of the larger

70
Co(sep)3+ cations between the pleated folds of the layer with the Co2+
cations adopting positions at the ridges as illustrated in Figure 11.9.
Confirmation of this structure will have to await future studies of the

magadiite structure.
D. Conclusions

The reaction of Co(sep)3+ with Na+-magadiite at 25°C resulted
in a product that contained Co(sep)3 1' cations bound to the external
crystallite surfaces of the layered silicate. In contrast, reaction at
100°C resulted in penetration of the metal complex between the
layers of magadiite. The magadiite structure was retained even after
prolonged reaction with Co(sep)3+ at 100°C. Moreover, the particle
morphology for the Co(sep)3+ intercalated derivative was the same as
the starting Na+-magadiite suggesting a topotactic intercalation.
However, concomitant with intercalation at 100°C, was the partial
demetalation of Co(sep)3+ which resulted in a mixed ion intercalate
containing both Co“ and Co(sep)3+ between the layers. N2 surface
area analysis indicated that only a small amount of microporous
surface existed in the Co(sep)3+ intercalated derivative, indicating
that most of the interlayer space was stuffed with little or no
available microporosity between pillars. Thus, Co(sep)3+ is not
suitable for the pillaring of magadiite, even though this cation has
been successfully used previously as a pillaring agent for related
ionic structures such as smectite clays [16]. In the case of the
layered sodium silicates it seems that pillaring is difficult to achieve
by direct ionic exchange. Future work will focus in part on the use of

robust organic cation intercalated derivatives of magadiite for

71

 

 

 

11.9. Proposed positions of Co(sep)3+ ions in the pleated folds of
magadiite. Coz+ cations are placed at the ridges which form between
folds. Vertices of the magadiite layer represent silicon positions;
oxygen positions are between adjacent silicon pairs. The fourth
coordination position of silicon is not shown.

 

accessing the intracrystalline

molecules from aqueous solution.

72

space of magadiite by organic

10.

ll.

12.

73

LIST OF REFERENCES

I. V. Mitchell,(ed.) : Pillared Layered Structures: Current Trends
and Applications, Elsevier, New York (1990).

T. J. Pinnavaia, 1. D. Johnson and M. Lipsicas: When],
63, 118 (1986).

H. P. Eugster: Seienee 157, 1177 (1967).

R. A. Fletcher, and D. M. Bibby: W35, 318
(1987).

K. Beneke and G. Lagaly: W68, 818 (1983).
K. Beneke and G. Lagaly: W152. 763 (1977).

G. Lagaly and K. Beneke: Am._MineL60, 642 (1975).

T. Yanagisawa, K. Kuroda, and C. Kato: MS,
167 (1988).

M. M. Mortland, S. Shaobai and S. A. Boyd: WM,
581 (1986).

S. A. Boyd, S. Shaobai, J. F. Lee, and M. M. Mortland: Wey
Minna“, 125 (1988).

I. 1. Creaser, R. J. Geue, J. MacB. Harrowfield, A. J. Herlt, A. M.

Sargeson, M. R. Snow, and J. Sprinborg: II._Am,__C_hem.__5mg.104,
6016 (1982) .

S. J. Gregg. and K. S. W. Sing: MW
W, 2nd edition, Academic press, London (1982).

IIT.

13.

14.

15.

16.

17.

18.

19.

20.

74
G. W. Brindley: Am,_MjneL54, 1583 (1969).

J. L. McAtee, R. House, H. P. Eugster: Am._M’meL53, 2061
(1968).

J. M. Garces, S. C. Rocke, C. E. Crowder, D. L. Hasha: We):
Mimi“, 409 (1988).

F. Tsvetkov and J. White: W110. 3183 (1988).
W. L. Masterton, E. J. Slowinski, and C. L. Stanitski: Chemjeel
13111121118112. Alternate Edition, 5th edition, CBS College Publishing
(1983).

W. Schwieger, D. Heidemann, and K. Bergk: W22,
639 (1985).

P. F. Barron, P. Slade, and R. L. Frost: W89. 3305
(1985).

J. M. Rojo, E. Ruiz-Hitzky and J. Sanz: W27, 2785
(1988).

J"

CHAPTER III

Silica Pillared Layered Silicates Derived From the
Polymerization of Tetraethylorthosilicate in the Galleries of
Alkylammonium Exchange Magadiite

A. Introduction

Pillared layered materials have attracted widespread interest
due in part to their catalytic and molecular sieving properties (1,2).
In an effort to broaden the diversity of pillared materials, we have

been investigating the pillaring reactions of the lamellar hydrous
sodium silicate Na+-magadiite (NaZSi O -9H20). Other examples of

14 29
this unique family of silicates include, kenyaite (Na28i20041-10H20),
and kanemite (NaHSi205-3H20). All are naturally occurring minerals
first found in lake beds at Lake Magadi, Kenya (3). These compounds

also can be conveniently prepared in the laboratory by hydrothermal

synthesis (4-6).

Intercalative ion exchange reactions of Na+-magadiite with
robust cations as pillaring agents is restrictive due to the relatively
high layer charge and non-swellable nature of the layered silicate.
Recent attempts to pillar Nat-magadiite by ion exchange of cobalt
sepulchrate, [Co(sep)]3+, a metal cage complex noted for its stability

in solution (7), afforded instead a non-microporous intercalated

75

Int-—

76

derivative in which one half of the metal centers were no longer

complexed by the sepulchrate ligand (8).

It has been shown that the hydrated interlayer sodium ions of
Na+-magadiite and related layered silicates can be exchanged by
large alkylammonium cations to form organic derivatives (5,6,9).
Analogous onium ion derivatives have been formed by the reaction
of the proton exchange from of the layered silicate with alkylamines
(10). Alkylammonium derivatives of magadiite and kenyaite have
been used as precursors for the reaction of interlayer silanols by
organo silanes. For example, dodecyltrimethylammonium intercalates
of magadiite and kenyaite react with trimethylchlorosilane to form

silane grafted derivatives (11).

In a related strategy to form pillared derivatives, the
interlayer SiOH groups of H+-magadiite have been investigated for
reaction with silane grafting reagents. Ruiz-Hitzky et a1. (12) have
found that if H+-magadiite is preintercalated with polar organic
molecules such as dimethylsulfoxide, trimethylsilyl groups can be
grafted to the silanol groups of the gallery surfaces. More recently,
Sprung and Davis (13) have used polyhedral silicate species derived
from phenyltrichlorosilane as pillaring reagents for H+-magadiite.
The direct reaction of H+-magadiite with this pillaring reagent
resulted in partially pillared derivatives of H+-magadiite. The
resultant calcined products exhibited surface areas in the range 100-

200 mzlg and basal spacings near 15.6 A.

77

The goal of this work was to utilize the intercalation chemistry
of magadiite to generate pillared microporous materials. The
strategy used to form pillared derivatives involves the
preintercalation of H+-magadiite by long-chain alkylamine molecules,
followed by the intercalation of a siloxane pillaring reagent, such as
tetraethylrothosilicate, which can graft with the interlayer SiOH
groups and also self polymerize to form Si02 aggregates. Recently
Landis et al. (14) reported that n-alkylammonium-magadiite samples
produced from aqueous solution react with tetraethylorthosilicate to
produce high surface area pillared materials (BET surface area = 530
m2/g). Although their strategy is similar to ours differences exist in
the preparation of the n-alkylammonium derivatives which affect
the basal spacings and surface areas of the final materials. Up to this
point no in depth characterization of these synthetic products has
been reported. Also, a systematic evaluation of the dependence of
the physical properties of the resultant materials on different

synthetic proceedures have not been investigated.
B. Experimental

1. Synthesis of H+-magadiite.

H+-magadiite was prepared by titration of Na+-magadiite with
0.1 N HCl by adaptation of the method of Lagaly et. al. (10). A

suspension composed of of Na+-magadiite (18.5 g) and 460 mL of
deionized H20 was titrated slowly with 0.1 N HCl. The solution pH

was lowered to 1.9 over a period of 12 hours and maintained at a

78

value of 1.9 for an additional 12 haurs. The pH of the solution was
measured continuously by using a Fisher Accumet Digital pH meter.
Solid H+-magadiite was separated by centrifugation and washed with

deionized H20 until Cl' free. The resultant product was dried in air

at 40°C.

2. Synthsis of Octylammonium-magadiite.

A reactive gel of octylammonium magadiite was formed by
allowing H+-magadiite (0.5 g, 0.57 mmol) to react at room
temperature with excess octylamine (2.0 g, 15 mmol). During
octylamine addition, the H+-magadiite solid absorbs the liquid amine
immediately forming a gray gelatinous mixture that will not flow.
The resultant gel was used without further treatment for all

reactions.
3. Synthsis of Silica Intercalated Magadiites.

Silica intercalated derivatives of magadiite were prepared by
the reaction of neat tetraethylorthosilicate, TEOS, with a gel
composed of octylammonium-magadiite solvated in excess
octylamine. To three 150 ml Erlenmeyer flasks each containing the
previously described octylamine H+-magadiite gel (2.5 g, 0.57 mmol
H+-magadiite) were added three different aliquats of TEOS (6.3 g, 31
mmol), (10.3 g, 51 mmol), and (17.7 g, 87 mmol). The reaction flasks
were covered and the mixtures were stirred vigorously for 24 hours.
The siloxane intercalated reaction products were separated by

centrifugation, and dried from ethanol in air at 40°C. Calcined

 

——

 

79

derivatives were prepared by heating the air-dried solids in air at

360°C for four hours.
4. Physical Measurements

Basal spacings were determined from the 001 x-ray reflections
of oriented film samples using a Rigaku Rotaflex diffractometer
equipped with Cu Ka radiation. Samples of Na+-magadiite, H+-
magadiite and the uncalcined TEOS/magadiite reaction products were
prepared by depositing on a glass slide a suspension of the solids and
allowing the suspension to air dry at 40°C. Samples of the calcined
silica intercalated magadiite were prepared for X-ray diffraction
analysis by heating the uncalcined TEOS/magadiite reaction products
on glass microscope slides to 360°C in air for four hours. The basal
spacing of octylammonium-magadiite solvated by excess octylamine
was obtained by smearing a thin film of the gel across a glass
microscope slide and then recording the diffraction pattern of the

wet sample.

29Si MAS NMR experiments were performed on a Varian 400
VXR solid state NMR spectrometer operated at 79.5 MHz. A Bruker
multinuclear MAS probe equipped with zirconia rotors was used for
all measurements. The 2S’Si relaxation times for Na+-magadiite, H+-
magadiite, and the uncalcined and calcined TEOS/magadiite reaction
products were determined by the inversion recovery method. 2S’Si
MAS spectra were obtained by using 4.6 us 90° pulse widths. A total

of 12 scans were accumulated for each sample. The spinning rate

was 5 kHz. Delay times approximately five times as large as T1 were

 

 

80

used in order to obtain quantitative integral intensities. Cross
polarization experiments were carried out with delay times of 10 s

and contact times of 1000 11s.

Nitrogen adsorption/desorption isotherms were determined on
a Quantachrome Autosorb Sorptometer at liquid N2 temperature

using ultrahigh purity N2 and He as adsorbate and carrier gases,

respectively. All samples were outgassed at 150°C under vacuum for
12 hours. Surface areas were determined using the BET equation

and the t-plot method (15).

Fourier transform infrared spectra were obtained on an IBM
1R44 spectrometer using the KBr pressed-pellet technique. All
pellets contained 0.5% sample by weight.

Thermogravimetric analyses were performed using a Cahn TG
System 121 thermogravimetric analyzer. All samples were heated to

1000°C at a heating rate of 5°C/min.
C. Results and Discussion
1. Na+- and H+-magadiite.

The basic hydrolysis of silica gel at 150°C under autogenus
pressure produced a well-crystallized Na+-magadiite product. The X-
ray diffraction pattern of an air dried film sample, shown in Figure
111.1A., exhibited several 001 reflections corresponding to a basal
spacing of 15.6 A. Despite the planar structure of Na+-magadiite and

efforts to produce well oriented films, several hkl reflections with

 

wiry-rm!

 

81

 

 

 

 

 

15.6 R

3‘
'71
c
0
E 12.4 A
c1
:5 5.6 R B'
o
T, W
D:

d 7.7 R 5'2 R A-

2 61014182226303438
Degrees 2Theta
111.1. X-ray diffraction patterns for film samples of (A) Na+-

magadiite, and (B) H+-magadiite. 22

 

 

 

 

82

h¢k¢0 were observed due to the rosette-like morphology of the
aggregated platelets. The peak positions for this synthetic product
agree closely with values reported previously (16,17) for synthetic

and natural magadiite.

The slow titration of Na+-magadiite with 0.1 N HCl resulted in
the exchange of sodium ions for protons in the layer structure (18).
The x-ray diffraction pattern of an air dried H+-magadiite film,
shown in Figure III.IB., exhibits 001 reflections corresponding to a
basal spacing of 11.2 A, in agreement with earlier work (18). This
decrease in basal spacing relative to the sodium form indicates a loss
of interlayer H20 upon replacement of Na+ by H+. Also, the general
broadening of the diffraction peaks indicates that greater stacking

disorder occurs upon proton exchange.

The sodium content of synthetic Na+-magadiite depends in part
on the extent of hydrolysis that occurs upon washing the product.
Washing with water initially removes excess NaOH associated with
the material. However, repeated washing leaches Na+ cations from

the interlayer and causes hydrolysis.

The chemical composition of the Na+-magadiite used in this
work was obtained by combining the results of thermogravimetric
analyses and chemical analyses. As shown by the thermogravimetric
curve in Figure III.2., air dried Na+ -magadiite that has been well
washed to remove excess NaOH, loses 13% of its total weight as water
below 200°C. An additional 1.6% is lost between 200°C- 1000°C. The
weight loss above 200°C is assigned to the dehydroxylation of SiOH

 

_——#_—

 

83

100-

 

98‘

96-

 

 

Weight (7:)
(O
N
l

90-
88-4 A.

86-

 

 

84

 

100. '250' .400 550' '7001 '850' 1000

Temperature (°C)

111.2. Thermogravimetric analysis curves obtained under
flowing argon for (A) Na+-magadiite, and (B) H+-magadiite.

 

mum

84

groups. By combining the NaZO content (4.98%) and weight loss, we

obtained an empirical composition for synthetic Na+-magadiite of
Na1.7Sil 4027.9(OH)1.9 - 7.6 H20. This result is in general agreement

with the ideal formula of Na28i14029 ~9H20 put forth by Lagaly et

al. (9), Garces et al. (17) and McAtee et a1. (19). The samples
synthesized in this work (cf.,Table 11.1.) straddle these compositions,
depending on post hydrothermal treatment of the product, especially
the number of washings. The water content can also vary
dramatically depending on drying conditions. The low sodium
content of our washed product, 1.7 Na+ per 14 Si, was a direct result

of hydrolysis caused by the extensive washing with water.

Thermal analysis of H+-magadiite, as shown by the curve in
Figure III.2B., indicated an initial weight loss below 300°C of 2.1%
due to the desorption of H20. The 4.0% weight loss above 300°C was
attributed to the elimination of OH groups from the structure. The

water loss together with the virtual absence of sodium, is in
agreement with an approximate unit cell composition of H4Si14O30.

H20. This formula agrees well with the compositions reported by

other workers (10,13).
2. Octylamine/Octylammonium Magadiite.

Lagaly and co-workers (10) have previously reported that H+-
magadiite intercalated by dimethylsulfoxide reacts with alkylamines
to form ordered bilayers of alkylamine molecules between the
silicate layers . Analogous gallery ordering is known to occur for n-

alkylammonium montmorillonites swollen by n-alkylalcohols or n-

 

 

 

85

alkylamines (20). The air-dried product resulting from the direct
reaction of octylamine with H+-magadiite exhibits a basal reflection
indicative of a 2.9 A gallery height or a 14.1 A basal spacing as
shown by the diffraction pattern in Figure 111.3A. Thus,
octylammonium cations were intercalated with the chains oriented
parallel to the silicate layer. The X-ray diffraction pattern of air
dried octylammonium-magadiite changes dramatically upon
solvation by excess octylamine, as shown in Figure “1.38. The
gallery height for the amine-solvated gel was 22.8 A, which indicated

the formation of bilayers of octylamine between the layers.
3. TEOS/magadiite Reaction Products.

The products obtained by the reaction of the
octylamine/octylammonium-magadiite gel with TEOS was dependent
on reaction stoichiometry. Depending on the molar ratio of TEOS to
magadiite, products with different basal spacings were obtained.
Figure 111.4. illustrates the X-ray diffraction patterns for the
uncalcined siloxane intercalates isolated from reaction mixtures
containing 54, 90, and 153 moles TEOS/moles of magadiite. These
patterns indicate basal reflections corresponding to basal spacings of
23.3 A, 25.2 A and 28.1 A respectively. Thus there was an apparent
increase in basal spacing with increasing amount of TEOS used in the
reaction. Since the layer thickness of magadiite is ~ll.2 A, the basal
spacings correspond to gallery heights of 12.1 A, 14.0 A and 16.9 A,
respectively. These gallery heights decrease from the initial gallery

height of 22 A observed for the octylamine/octylammonium-

E-

86

 

 

 

 

 

 
 

 

 

 

 

54.08
3‘ 1
g 14.1 R A.
E
a, 6.98 E
.2
1.6
o
a:
I 17.08 B
11.58 '
‘ I 1 I ' I ' I
0 4 81216202428323640
Degrees 2Theta
111.3. X-ray diffraction patterns for film samples of

octylammonium-magadiite (A) sample air-dried at 25°C and (B)
solvated by excess octylamine.

87

 

 

 

 

 

 

 

 

 

 

 

28.1 X
3‘
'6
E, 13.9 X (153:1)
g 25.2 X
.9.
0
D:
J L (90:1)
2 4 6 8101214161820
Degrees 2Theta
111.4. X-ray diffraction patterns for film samples of the

uncalcined TEOS/magadiite reaction products that resulted from
treatment of octylamine solvated octylammonium-magadiite with
tetraethylorthosilicate. The TEOS/magadiite mole ratios were (54:1),
(90:1) and (153:1).

 

88

magadiite precursor gel, so that the gallery shrinks by ~ 540 A
when octylamine is replaced by TEOS.

The X-ray diffraction patterns shown in Figure 111.5. for the
products calcined in air at 360°C, indicate a further decrease of 2-3 A
in gallery height for each derivative. Also, there is a decrease in the
degree of order along the c-axis, as evidenced by a broadening of the
basal reflection of each derivative. Table 111.1. summerizes the d
spacings and gallery heights for the starting materials along with
those for the calcined and uncalcined TEOS/magadiite reaction

products.

The chemical compositions of the products derived from
reaction of octylamine/octylammonium~magadiite with TEOS were
obtained by combining TGA and C, H, N, chemical analyses. All
nitrogen was assumed to be due to the presence of octylammonium
cation and all excess carbon was attributed to the presence of
residual alkoxide (C2H50-), associated with polymerized siloxane.
Table 111.11. provides the elemental compositions of H+-magadiite,
octylammonium-magadiite and the uncalcined TEOS/magadiite
reaction products as a function reaction stoichiometry. It is
noteworthy that the percentage of octylamine present in the
uncalcined reaction products decreases as the amount of TEOS used
in the reaction increases. This may indicate that neutral amine is
replaced to a larger extent as the amount of TEOS present in the
reaction mixture increases. The three reaction products also contain
residual alkoxide. It is noteworthy that the amine content of the

samples is similar to the amount present in the octylammonium—

 

'H'HY

18‘! “If?"

89

 

A 25.9 R

 

 

22.4 8

Relative Intensity
k

(90:1)

 

 

fix (54:1)

I I ' I ' I ' I ' I ' I '

4 6 8 10 12 14 16 18 20
Degrees 2Theta

 

 

111.5. X-ray diffraction patterns for film samples of
TEOS/magadiite reaction products after calcination. Calcined in air at
360° C for four hours.

 

 

 

90

Table 111.1. Basal Spacing and Gallery Heights of Magadiite Reaction

Precursors and TEOS Reaction Products

 

Sample Basal Spacing, A GalleryAHeight,

 

N a+-magadiite

H+-magadiite
C8H17NH3+-magadiite
(air dried)
C8H17NH3+-magadiite
(amine solvated)
TEOS/magadiite
(56:1)
TEOS/magadiite
(92:1)
TEOS/magadiite
(156:1)
TEOS/magadiite
(56:1, calcined 360°C)
TEOS/magadiite
(92:1, calcined 360°C)
TEOS/magadiite
(156:1, calcined
360°C)

15.6 4.4
12.4 1.2
14.1 2.9
34.0 22.8
23.3 12.1
25.2 14.0
28.1 16.9
20.7 9.5
22.4 11.2
25.9 14.7

 

rum
. —‘_

 

91

magadiite derivative, indicating that only strongly bound

octylammonium cations remain after treatment with TEOS.

The nitrogen adsorption/desorption isotherms shown in
Figure 111.6. were obtained for calcined TEOS/magadiite reaction
products. The data were plotted according to the BET equation (15)
in order to obtain surface area values. The equivalent surface areas

for H+-magadiite and the calcined derivatives determined from the

Table 111.11. Analytical Data

 

 

Sample %Sioza %CgH17NH2b %C;H5-ob %H2oc %Total
H-magadiite 93.9% —- m 6.1% 100%
(Edged) 84.4% 11.1% -- 5.0% 100%
TEOS/magadute"

(56:1) 79.0% 13.5% 2.4% 2.9% 97.8%
TEOS/magadiite

(92:1) 77.7% 12.4% 3.8% 4.8% 98.7%
TEOS/magadiite

(l_56:1) 80.4% 10.4% 3.3% 4.5% 98.6%

 

1‘Based on weight retained after heating to 1000°C

bValues obtained by carbon, hydrogen, nitrogen analyses.
cObtained by weight loss below 200°C
dReaction mixture stoichiometry in moles TEOS/moles magadiite

monolayer volume Vm (BET) are listed in Table III.III.. Microporous

volumes for the samples were determined by using the t-plot

method (15). The data listed in Table 111.111 indicate H+-magadiite

 

 

92

 

 

 

 

 

 

 

 

 

 

 

 

2501
2001 (154:1) 6
7 8
Q:
1 s 5 5 s
150: Q Q Q) 5 ”‘3‘; 13
1 o
1 0
lm‘ ' I I j I I
’53 250
b .
53 1 3
F 1 (901)
<9 200. ‘ Q,
a . 669%
S . Q Q” 9 Q Q h b 5 ‘3
'6': 150: .0
a '2 «P
E 1
_=_ :
a 1m I I ' I I Y
>
250
I o
3 (54:1) . o
. O O O
: ...%%%%°°°°°°
1 q:
100‘ - , - , . . . . .
0.0 0.2 0.4 0.6 0.8 1.0
P/Po
111.6. Nitrogen adsorption/desorption isotherms for the

calcined TEOS/magadiite reaction products produced with a ratio of
(54:1), (90:1) and (153:1) moles TEOS/mole magadiite.

 

——

l 1mm!"

93

has a total surface area of 45 mzlg due exclusively to adsorption at
non-porous external surfaces. The calcined products obtained from
the TEOS/magadiite reaction exhibited dramatically larger surface
areas between 520-680 m2/g, depending on the concentration of
TEOS used in the reaction mixture. The interlayer space or
microporous surface accounts for the majority of porosity introduced
into these materials. The amount of external and mesoporous
surface, determined by using the parallel pore method (21) vary
little between samples. This increase in surface area can be

attributed to the pillaring of magadiite by intercalated aggregates of

Table III.III. Surface Area Analysis (m2/g) of Pillared Derivatives

 

 

Sample SEEP Smm Smic S
H+-magadiite 45 45 0 4 5
TEOS/magadiite
(56:1) 620 635 590 45
TEOS/magadiite
(92:1) 680 705 670 35
TEOS/magadiite '
(156:1) 520 525 480 45

 

1‘ Definitions of abbreviations: SBET is the N2 BET surface area.
Sum], Smic’ and S are the total, microporous, and non-

microporous surface areas obtained from t-plots of the
adsorption data

 

IrTT—umfl
. - 's 1 . g]

 

94

silica. The relationship between the microporous surface area, Table
111.111., and the basal spacing, Table 111.1., indicate there is a change
in porosity of the calcined TEOS/magadiite reaction products as the
TEOS stoichiometry changes. For example, an optimum surface area
is obtained at an intermediate doo1 value. This effect can be
explained by assuming an increase in pillar size as the gallery height
increases. Thus, as the pillars increase in size the lateral spacing
between pillars will decrease , Figure 111.7. An intermediate value of
the basal spacing would yield an optimum porosity only if there
exists the best compromise between pillar height and lateral pillar
spacing. The infrared spectra of calcined and uncalcined
TEOS/magadiite (90:1) reaction products are compared in Figure
111.8., with the spectra for octylammonium-magadiite and H+-
magadiite. Both the calcined TEOS/magadiite (90:1) product and H+-
magadiite exhibit the same stretching bands between 4000-400 cm:
1. H+-magadiite exhibits infrared bands similar to the sodium form
(18). The broad overlaping bands centered at 3445 cm"1 were
assigned to the stretching frequencies of the OH group of either
silanol groups or water associated with the silicate (18). The band

centered at 1632 cm"1 was attributed to the bending frequencies of
H20 (18). The bands from 1500 cm'1-400 cm'1 were attributed to

the stretching and bending frequencies of the SiO4 units that make

up the layer (17). For Na+ -magadiite, Garces et al. (17) have

attributed the bands at 1237 cm'1 to five membered ring blocks of
SiO4 tetrahedra. In addition, the bands at 1210 cm-1 and 1175 cm-1

were attributed to five, six and four membered ring block structures

 

 

11:";

 

95

TEOS/magadiite Smic

 

 

£1 . 54 .510 milg

 

 

# l

 

 

 

 

 

 

11121 90 600 mzlg

1

T 2

14.7.1 153 475 m /g

1

111.7. Schematic representation for the variation of pillar size

with the change in reaction stoichiometry.

 

 

96

 

 

 

Relative Absorbanco

 

 

 

 

 

I l I 1 I I I l T I I I I l
4000 3500 3000 2500 2000 1 500 1 000 500
Wavenurnbera cm-t

111.8. Infrared spectra for KBr pellets (0.25 wt %) of (A) H+-
magadiite, (B) air-dried octylammonium-magadiite, (C)
TEOS/magadiite reaction product (90:1) air dried, and (D)
TEOS/magadiite (90:1) reaction product after calcining at 360°C.

 

 

97

similar to those found in the zeolites epistilbite and dachiardite (17).
Comparison of the infrared spectrum of Na+-magadiite with the
spectra obtained for the calcined TEOS/magadiite (90:1) product and
H+-magadiite indicate an almost identical match of stretching
frequencies for the region between 1500 cm'1 and 400 cm'l. It is
possible to attribute by analogy the stretching frequencies near 1225
cm'1 to the presence of block structures composed of a combination
of five, six and four membered rings of S104 tetrahedra in both the
calcined TEOS/magadiite (90:1) product and H+-magadiite. In turn,
the layer structure of H+-magadiite and the calcined TEOS/magadiite
(90:1) product are composed of the same building blocks as Na+-
magadiite. In addition, the presence of absorptions at ~1225 cm°1
indicates there is retention of the layer structure upon pillaring by
silica. The infrared spectrum of the uncalcined TEOS/magadiite
(90:1) reaction product exhibits IR bands which match the
octylammonium-magadiite spectrum closely Figure 111.8B-C. In
addition to the bands characteristic of the magadiite layers, weak
infrared bands at 2962 cm'l, 2929 cm'l, 2873 cm'1 and 2859 cm'1
were observed for the uncalcined TEOS/magadiite (90:1) product.
These bands are attributed to the C-H asymmetric and C-H
symmetric stretching frequencies of the intercalated octylammonium
cation. Weak broad bands centered at 1537 cm'1 are assigned to the
N H 3+ deformation (bending) frequencies of the ammonium cation. A
weak. absorption at 1469 cm'1 results from the C-H asymmetric
bending of the alkyl chain. A weak intensity band at 923 cm'l, here
assigned to the Si-O stretching of SiOH groups in octylammonium-

magadiite were also found to be present in freshly precipitated silica

 

 

98

gels near 950 cm‘1 (22) and in hydrolyzed polysiloxane polymers
near 890 cm'1 (23). This indicates first that octylammonium cation
remains even after intercalative reaction with TEOS and secondly,
that the bands associated with layer structural blocks remain intact
after intercalation of the siloxane species. Also, there is a shift in the
absorption associated with the Si-O stretching of the SiOH groups
from 923 cm‘1 to 958 cm’l. This most likely is due to the the
overlaping Si~O stretching frequencies of the intercalated siloxane
which exhibit higher frequency peaks in polysiloxane polymer 1104
cm'l, (24). After calcination, the bands which were present in the
infrared spectrum of the uncalcined TEOS/magadiite (90:1) product
near 2900 cm'l, 1460 cm"1 and 1539 cm"1 dissapear. The loss of
these bands results from the oxidation and desorption of
octylammonium cation and ethoxide groups present in the material.
In addition, the absorption near 958 cm'1 in the uncalcined material
shifts to 965 cm'1 after calcination indicating the formation of Si-O-

Si linkages (22).

The 2S’Si MAS NMR spectra of the calcined TEOS/magadiite

reaction products, shown in Figure 111.9., all indicate two general

silicon environments, namely a Q3 type HOSi(OSi)3 environment near

401.5 ppm and Q4 type Si(OSi)4 environments near -111 ppm and -

114.5 ppm. The relaxation times for the calcined TEOS/magadiite
reaction products were found to be 44 s i 5 s for the Q3
environments and 141 s :1: 11 s for the Q4 environments. The 29Si
MAS NMR spectra of the uncalcined TEOS/magadiite reaction

products, shown in Figure 111.10., exhibit Q3 and Q4 resonances at

 

tram-w

 

99

 

 

 

 

  

.1105
Q3/Q4 =
414.4
-111.1
.1013
(153:1, calcined
) TEOS/mag. 0.18
111.1
. 401.4
(90:1, calcmed) TEOS/mag. O 17
.922 '
.110
.1015
(54:1, calcined) TEOS/mag. 11 7 0.21
0.28

 

 

 

H+-magadiite U

111.9. 29Si MAS NMR spectra of H+-magadiite and calcined
TEOS/magadiite reaction products produced with a ratio of (54:1),
(90:1), and (153:1) moles TEOS/mole magadiite.

 

FW"

100

410.8

93/94 =

 
  
 
  
     
  

 

 

 

(153:1) TEOS/mag. 0.26
111M

(90:1) TEOS/mag. 0.29
(54:1) TEOS/mag. 0.30

113.7

400.4

.111.3
H+-magadiite 0.28

413.9

-99.3

Na+-magadiite 0.36

111.10. 29Si MAS NMR spectra of Na+-magadiite, H+-magadiite
and uncalcined TEOS/magadiite reaction products produced with a
ratio of (54:1), (90:1), and (153:1) moles of TEOS/mole magadiite.

101

approximately the same chemical shift values as the calcined
products, Table 111.VI., with only slight differences. Calcination
results in a slight upfield shift of all chemical shift values probably
due to increases in Si-O-Si bond angles. Also, Q3/Q4 ratios for the
calcined TEOS/magadiite products are lower than the Q3/Q4 ratios for
the uncalcined derivatives due to the condensation of Q3 silanol
groups. During calcination Q3 silanol groups which are in close
proximity condense to form siloxane bonds. Even after calcination a
large amount of Q3 silanol groups still remain. The absence of Q0, Q1,
and Q2 silicon resonances in the 29Si MAS NMR spectra of both the
calcined and uncalcined TEOS/magadiite reaction products indicate
TEOS has polymerized to form silicate species which are composed of
Q3 and Q4 silicon environments. Also, because the Q3/Q4 ratios for
the products are the same as for H+-magadiite it follows that the
Q3/Q4 ratio for the intercalated silicate species is close to that of H+-

magadiite.

The 29Si MAS NMR spectra of the calcined and uncalcined
TEOS/magadiite reaction products all exhibit chemical shifts and
relative intensities which closely match H+-magadiite and Na+-

magadiite. For instance, Na+-magadiite exhibits two general Si
environments, namely Q3 type (HOSi(OSi)3) or Na+[OSi(OSi)3]) and Q4

type Si(OSi)4 environments (18,25). The 29Si MAS NMR spectrum of

N a+-magadiite, Figure 111.10., exhibits a single Q3 peak centered at -
99.8 ppm and three Q4 resonances occur at 410.7 ppm, 411.7 ppm
and 414.2 ppm respectively. Previous workers have published

Q3/Q4 ratios for Na+-magadiite (17,25). However a careful

 

111111-11

 

102

Table 111.1v. 2981 MAS NMR Chemical Shift Values and Q3/Q4 ratios
for H+-magadiite and the TEOS/magadiite reaction products.

 

 

 

 

Tm‘ Sample ChemicalShift (ppm) Q3/Q4
02*.03 ()4
H-Magadiitg -100.4 410.8, -113.7 0.28
Wigs“ -99.9 410.7, 413.9 0.30
mosfififia‘m‘" -99.9 410.8. 413.9 0.29
150$???“ -99.9 410.8, 413.7 0.26
TEOS/magadiite
(56,1,calcincd -1015 411.1, -1144 0.21
360°C)
TEOS/magadiite
(92,1,calcincd -92.2*.-101.4 411.1, 414.7 0.18
360°C)
IEOS/magadiite
(156.1,calcincd -101.3 410.6, 414.4 0.17
360°C)
CP-MAS
15031233899 401.6 -1137
TEOS/magadiite
(92:1) 401.6 -1132
TEOS/magadiite
(156.1) 905*. -101.2 412.7
TEOS/magadiite
(56:1, calcined -90.8*, 401.4 ~11 1.9
T508613?
gadiite _ _ _
(92:1,calcined 919". 101.2 111.7
360°C)
IEOS/magadiitc
(156:1, calcined 91.3". -101.6 412.9
360°C)

1‘ Reaction Stoichiomeu'y ratio in moles TEOS/mole H-magadiite

 

103

evaluation of the relaxation times of the various 29Si environments
has not been undertaken until now. The differences in Q3 and Q4
relaxation times for Na+-magadiite, 160 s and 280 s respectively are
attributed to the presence of H20 in the interlayer. The importance
of 29Si-IH dipolar relaxation via water molecules has been shown for
other layered materials (26). The close proximity of interlayer water
to the Q3 2S’Si sites of Na+-magadiite allows for there relaxation via
29Si-lH dipolar relaxation and in turn reduces the relaxation time of
the Q3 site greatly as compared to the Q4 site. H+-magadiite like Na+-
magadiite, and the calcined and uncalcined TEOS/magadiite reaction
products exhibits two general silicon environments, namely Q3 type
HOSi(OSi)3 and Q4 type Si(OSi)4 environments (25). The 29Si MAS
NMR spectrum of H+-magadiite, Figure 111.10., exhibits a single Q3
resonance at 400.4 ppm and Q4 resonances at 410.8 ppm and -
113.7 ppm. In general the chemical shift values for H+-magadiite
closely match those for Na+-magadiite and the reaction products.
The only noticeable difference between Nat-magadiite and H+-
magadiite was the linewidth. This increase in linewidth after
exchange of protons for sodium ions may be a result of unaveraged
dipolar interactions of the 29Si nucleus with 1H. A careful evaluation
of the relaxation times, T1, for the Q3 and Q4 silicon environments in
H+-magadiite indicate a marked decrease as compared to Na+-
magadiite. Relaxation times for H+-magadiite are 95 s for both the Q3
and Q4 environments. The relative integral intensities of H+—
magadiite indicate a Q3/Q4 ratio of 0.28 which agrees closely with the
Q3/Q4 ratio of the uncalcined TEOS/magadiite reaction products.

 

.—

warm-r

 

104

Based on a Si14 unit cell, the Q3/Q4 ratio of H+-magadiite converts to
3.08 silanol groups/ unit cell. The combination of this information

and thermal analysis results in a more accurate unit cell formula for
H+-magadiite Of Na.0.ozsi14026.5(OH)3.08°1.5H20.

The 29Si CP-MAS spectra of the calcined TEOS/magadiite
reaction products, Figure 111.11, indicate a new resonance at ~-91
ppm in addition to the enhanced Q3 resonance near 101.5 ppm.
These resonances are assigned to Q2 (HO)28i(OSi)2 29Si environments.
The cross polarization 29Si spectra of the uncalcined TEOS/magadiite
reaction products are shown in Figure 111.12. These products all
exhibit a Q3 peak near 401.6 ppm and Q4 resonances at
approximately 413 ppm. In each case the Q3 resonance has been
enhanced indicating the presence of silanol groups. These Q3
HOSi(OSi)3 groups may be associated with either the layer or the
siloxane pillar. The uncalcined TEOS/magadiite (153:1) reaction
product exhibits a weak intensity peak near -90.5 ppm. This
indicated the presence of Q2 (I-IO)ZSi(OSi)2 2S’Si environments. The
general absence of Q2 environments in the uncalcined derivatives
indicate the Q2 298i environments are present as ethoxide rather than
hydroxyl groups which accounts for the lack of enhancement.
Calcination results in combustion of the ethoxide groups. Water,
produced in the combustion or present in the atmosphere may
rehydrate these silicon environments resulting in Q2 silanol
environments which would be enhanced by cross polarization.

Proton cross polarization of 2S’Si nuclei in H+-magadiite, Figure 111.12.,

results in the enhancement of the Q3, HOSi(OSi)3, resonance centered

 

1‘12””!

 

105

401.6

-9l.3

 

 

(15311. calcined) TEOS/mag. ' 4101-2

1

412.9

 

-9l.0

‘—
——f

(90:1, calcined) TEOS/mag.

 

l
” 411.7

 

 

7.

-90.8

(54:1, calcined) TEOS/mag

 

 

 

1
71/»

111.11. 2S’Si CP-MAS NMR spectra of calcined TEOS/magadiite
reaction products produced with a ratio of (54:1), (90:1), and (153:1)
moles TEOS/mole magadiite. Products calcined at 360°C in air for
four hours.

 

‘REM

401.2
412.7
(153:1) TEOS/mag.
-101.6
w
413.2
(90:1) TEOS/mag. 401.6
413.7

(54:1) TEOS/mag, 401.

NJ

H+-magadiite

  
 
  

111.12. 29Si CP-MAS NMR spectra of H+-magadiite, and
uncalcined TEOS/magadiite reaction products produced with a ratio
of (54:1), (90:1), and (153:1) moles of TEOS/mole magadiite.

107

at -101.6 ppm. The Q4 Si(OSi)4 resonances at -111.7 ppm and -114.4
ppm are not enhanced due to the absence of hydroxyl groups at

these sites.

N a+-magadiite is known to adopt a particle morphology
composed of silicate layers intergrown to form spherical nodules
resembling rosettes (17). The proton exchange form also has this
characteristic particle morphology Figure III.IBA. (27). Scanning
electron micrographs of octylammonium-magadiite indicate a loss of
the initial particle morphology Figure III.I4A. It appears that the
treatment of H+-magadiite with octylamine followed by air drying
results in a break up of the spherical nodules with a concomittant
random arrangement of the platelets. Scanning electron micrographs
of the uncalcined and calcined TEOS/magadiite products, shown in
Figure 111.143., and III.14C., respectively, exhibit the same
morphology as ocytylammonium-magadiite. The calcined and
uncalcined TEOS/magadiite derivatives exhibit particle sizes between
5p-l400u, much larger than the particle size of H+-magadiite, as
shown in Figure 111.138. The similarity between the morphology of
octylammonium-magadiite and the calcined and uncalcined
TEOS/magadiite reaction products indicated that intercalation of TEOS

occured in a topotactic fashion.
4. Mechanism of Intercalation.

Reaction of H+-magadiite and octylamine forms a derivative
with a bilayer structure between the layers of H+-magadiite. The

bilayer structure was composed of both octylammonium cations and

108

111.13. Scanning electron micrographs of H+-magadiite at A)
X6000 and B) X600 magnification.

 

 

c all)

.14.”.
Q...
U
E
C
U
.14....
B
.1

8841

 

8E898

8U

l

2
4
8
DH.

18KU H6888

 

110

111.14. Scanning electron micrographs at X6000 magnification for
A) air-dried octylammonium-magadiite (top left), B)
TEOS/magadiite reaction product (54:1) air-dried (top right), C)
calcined TEOS/magadiite (54:1) (bottom).

ooamx 33H

.1
.. g1... 141%
p

3444.3 5

 

112

 

 

 

 

 

 

 

 

 

111.15. Schematic representation for the intercalation and
pillaring of octylamine solvated octylammonium-magadiite.

113

neutral octylamine Figure 111.15. Reaction of this intercalated
derivative with tetraethylorthosilicate results in the replacement of
the neutral octylamine with TEOS by a diffusion process. As TEOS
replaces octylamine the layer SiOH groups react with the siloxane
groups to form siloxane bonds and EtOH. Further hydrolysis of the
intercalated siloxane results during air drying and storage of the
siloxane intercalated derivative. Calcination results in combustion of
octylammonium cations and ethoxide groups which still remain in
the material. The loss of organic matter during calcination creates
voids within the material and accounts for the microporous surface

areas Figure 111.15.

The high surface area and crystalline nature of the silica
pillared derivatives indicate a high degree of cooperativity in pillar
formation. There are four possible driving forces for the formation
of descrete pillar domains. For example; i) charge localization within
the layer may result in the formation of descrete domains of
alkylammonium cations which separate TEOS domains, ii) domains of
surface SiOH groups could function as reaction sites for the siloxane
reagent resulting in pillar domains, iii) a highly corrugated layer may
result in layer structure control over pillar formation, iv) solvation
effects between the alkylammonium cations, the layer and TEOS may
result in segregation of domains. In addition, the possibility exists

that two or more of the above phenomena may work in combination

114

D. Conclusions

The reaction of H+-magadiite with octylamine results in an
octylammonium intercalated derivative swollen by neutral
octylamine. This derivative has a basal spacing of 34 A when
solvated by excess octylamine, but the basal spacing collapses to 14
A after drying. We have demonstrated the utility of this material as
a precursor, in the solvated state, for the intercalation of
hydrolyzable reagents such as TEOS. Reaction of this precursor with
TEOS results in the intercalation and polymerization of TEOS at
specific sites on the H+-magadiite surface. The mechanism for
polymerization appears to occur by a topochemical process with the
H-magadiite surface acting as a template. The final reaction product,
after calcination, is composed of layers of magadiite separated by
regularly spaced silica pillars. By adjustment of the TEOS
stoichiometry final products were obtained that possess gallery
heights from 9.5 A to 14.9 A and microporous surface areas from
480 - 670 m2/g. In general the surface areas of our materials are
higher than those reported by Landis et a1. (14). The gallery heights
increase with increasing TEOS concentration. However, the
microporous surface areas attained a maximum and then decreased
which may indicate a change in the pillar size. Interior surface
silanol groups of the type Q3 HOSi(OSi)3 and Q2 (HO)zSi(OSi)2 are
present in the final pillared products. These silanol groups should

prove useful as grafting sites for catalytic metal centers.

 

10.

ll.

12.

13.

115

LIST OF REFERENCES

F. Figueras: W30, 457 (1988).

D-E-W- Vaughan. In EersncctixcLiLMcleculaLSiexeSciencz
W.H. Flank and T.E. Whyte,Ed.;American Chemical Society,
Washington, D C 1988; p308.

H.P. Eugster: MIST 1177 (1967).

R. A. Fletcher, and D. M. Bibby: W35, 318
(1987).

K. Beneke and G. Lagaly: W68, 818 (1983).

K. Beneke and G. Lagaly: W62, 763 (1977).

I. I. Creaser, R. J. Geue, J. MacB. Harrowfield, A. J. Herlt, A. M.
Sargeson, M. R. Snow, and J. Sprinborg: W 104,
6016 (1982) .

J. S. Dailey, T. J. Pinnavaia, Submitted to J. Inclusion
Phenomena.

G. Lagaly and K. Beneke: W60, 642 (1975).
G. Lagaly, K. Beneke and A. Weiss: W60, 650 (1975).

T. Yanagisawa, K. Kuroda, and C. Kato: W 5, 167
(1988).

E. Ruiz-Hitzky, J. M. Rojo, and G. Lagaly: WM
263, 1025 (1985).

R. Sprung, M. E. Davis, J. S. Kauffman and C. Dybowski: ImLEng.
W29. 213 (1990)-

 

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25

26.

27.

116

M. E. Landis,'B. A. Aufdembrink, P. Chu, 1. D. Johnson, G. w.
Kirker, and M. K. Rubin: IL._Ame[._§.2hem._S_Qo;;_.113, 3189 (1991).

S. J. Gregg, and K. S. W. Sing, "Adsorption, Surface Area and
Porosity", 2nd edition, Academic press, London (1982).

G. W. Brindley: W54, 1583 (1969).

J. M. Garces, S. C. Rocke, C. E. Crowder, D. L. Hasha: W
Miner. 36, 409 (1988).

J. M. Rojo, E. Ruiz-Hitzky and J. Sanz: W27, 2785
(1988).

J. L. McAtee, R. House and H. P. Eugster: W53, 2061
(1968).

A. Weiss: W2. 134 (1963)-

J. F. Delon and R. Dellyes: W265, 1161
(1967).

M. Hino and T. Sato: W44, 33 (1971).

H. Ishida, " Adhesion Aspects of Polymeric Coatings." K. L.
Mittal, ed. Plenum; New York, 1983, p. 45.

Chiang, Ishida: W74, 396 (1987).

T. J. Pinnavaia, I. D. Johnson and M. Lipsicas: L_Qf_S_Qlid_S_tate
M63, 118 (1986).

P. F. Barron, P. Slade, and R. L. Frost: M39. 3305
(1985).

G. Lagaly: WWII. 105 (1979)-

 

 

 

        

 

HICHIGRN STA

I Lulu) mm Willi»

29300902

      

 

 

 

 

llllll“