Hm

llllllllllllllllllllllllll

This is to certify that the

thesis entitled

A New Microporous Silica:

Its Intercalation in Magadiite
presented by

Astrid Baviére
has been accepted towards fulfillment

of the requirements for

Wdegree in Chemistry

L/W Jim

professor

 

[Date.__lngL_3£L,_l£U12__._

0-7639 MSU is an Affirmative Action/Equal Opportunity Institution

A NEW MICROPOROUS SILICA:
ITS INTERCALATION IN MAGADIITE

By

Astrid Baviére

A THESIS

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

MASTER OF SCIENCE

Department of Chemistry

1992

 

 

ABSTRACT

A NEW MICROPOROUS SILICA:
ITS INTERCALATION IN MAGADIITE

By

Astrid Baviere

The study of magadiite evidenced the structural importance of water
in the intergallery space. Removal of water from Na-magadiite by heat
treatment is equivalent to proton exchange with regard to effects on
stacking order. The loss of order in H-magadiite was assigned to a
rearrangement of the gallery space due to the absence of water molecules.

The acid-catalyzed hydrolysis-polymerization of tetraethylortho-
silicate in the presence of alkylamines generates precursors, which, when
calcined, afford highly microporous silica. The pore size is controlled by
the alkylamine chain length.

The polymerization of tetraethylorthosilicate into alkylamine-swollen
layered silicic acid affords materials with large surface areas, due to the
silica-intercalated porosity. The hydrolysis is catalyzed by the interlayer
protons. Alkylamines are used to swell the layers in order for the protons

to be accessible, but they also play a role in designing the pores of the

silica-intercalated magadiite.

 

To my family

 

ACKNOWLEDGMENTS

I would like to acknowledge here advisor Prof. T. J. Pinnavaia for
the opportunity I was given to realize this project. I really appreciated his
comments and suggestions as well as his everyday good spirit. Also, thanks
should go to the members of my Guidance Committee, Prof. H. A. Eick
and Prof. G. J. Blanchard for their helpful advice on writing this thesis. I
am also appreciative of Prof. H. A. Eick's help on X—ray analysis and
electron diffraction.

I want to say a big 'Thank you' to the Pinnavaia's group members,
who made these two years in the laboratory enjoyable, even if, sometimes,
chemical smell could be detected, glassware was missing, but it was all fun.
I really want to thank all of them for being understanding when I was
monopolizing the Macintosh for drawing my figures. I especially want to
thank Dr. Laurent Michot for his research spirit and for trying to
communicate it through his friendly behavior. I will also remember Dr.
Antonio Lara for caring about me coming back late from the Chemistry
Building. No words can describe Dr. Wang and the nice feeling just to
know that his smiling face was around. Also, thanks should go to Dr. Jay
Amaratsekara and Dr. Ravi Kukkadapu for helping me to transfer files, for

discussing science and keeping me informed on sport events (especially

basketball).

 

I will also remember all my friends that made my two years at
Michigan State University enjoyable. Among them, a special thank goes to
my Summer '92 roommates, Raul (the future president of the Philippines)
and Takayo, for being there when I needed breaks and for making me taste
Filipino and Japanese food. I also want to acknowledge Arvind and Sanjay
for their presence. I will always remember Petra for making fun of my
French accent. Also to be cited are Hyon, the little Korean doll who wanted
to fly and Greg the Greek who wanted to teach her how to.

Furthermore, special thanks should go to the 'French delegation'.
Among them to Anne, with two '6' in her last name, the first person I met
when I arrived at Lansing Airport. Other members of the connection
include Betty, Sophie, Sylvie, Pascal l and 2, and Frédéric. Frédéric will
always be 'The French Cook' to me and his home 'The French Party
House'.

Finally, but not least, I want to thank Jean-Rémi for being there
everyday: to help me in the laboratory, to correct the first draft of this
thesis, to paste the captions under the figures, to believe in me and much

more...

vi

 

TABLE OF CONTENTS

LIST OF TABLES xi
LIST OF FIGURES xv
ABBREVIATIONS xxvi

 

 

 

RESEARCH OBJECTIVES 1

 

EXPERIMENTAL
mm

Na-magadiite

 

 

 

H—magadiite
Amine-solvated magadiite

 

Siloxane-intercalated magadiite

 

Silica-intercalated magadiite

 

TEOS-derived silica precursor

TEOS-derived silica
Analysis
Chemical analysis

Elemental analysis

 

 

 

 

oaso‘o‘mmm-c-hwm

 

vii

 

C, H, N analyses

Thermal analysis

 

 

 

Thermogravimetric analysis

 

Differential scanning calorimetry

 

Structural analysis

 

X-ray powder diffraction
Nitrogen adsorption-desorption

 

Infrared spectroscopy

 

29Si nuclear magnetic resonance

 

Scanning electron microscopy

 

OOOOOOOO\I\I\I\I\)\IO\

Transmission electron microscopy

 

MAGADIITE CHARACTERIZATION

 

 

 

 

 

 

 

 

Introduction 9
R 1 di s i n 11
Syntheses 1 1
Chemical compositions 12
X-ray powder diffraction 21
Infrared spectroscopy 36
29Si nuclear magnetic resonance 42
Nitrogen adsorption-desorption 49
Structure discussion 50

 

QQIICLELQLI 56

 

viii

 

ACID CATALYZED TEOS POLYMERIZATION IN THE PRESENCE
OF ALKYLAMINES:
SYNTHESIS OF A NEW POROUS SILICA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Intr tion 57
Background 58
jlfhc sol-gcl proccssing of silicates 60
Hydrolysis 60
Condensation-polymerization 61
Gelation 61
Drying 61
Densification 62

1 i i 62
Reaction conditions 62
Chemical compositions 64
Infrared spectroscopy 69
29Si nuclear magnetic resonance 77
Nitrogen adsorption-desorption 81
X-ray powder diffraction 87
Transmission electron microscopy 92
Conclusion 97

 

POLYMERIZATION OF TEOS INTO LAYERED SILICIC ACID:
A NEW POROUS SILICA INTERCALATED IN MAGADIITE LAYERS

 

 

Introdoctio 98
l i i 100
Synthesis 100

 

 

 

 

Alkylammonium/alkylaminemagadiite 100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

X-ray diffraction 101
Chemical compositions 108
29Si nuclear magnetic resonance 112
Conclusion 1 14
Experimental conditions for TEOS polymerization in H-magadiite . 115
Thermogravimetric analysis 115
Chemical compositions 117
X-ray powder diffraction 119
Nitrogen adsorption 121
Infrared spectroscopy 126
29Si nuclear magnetic resonance 132
Conclusion 135
Alkylamine and TEOS to magadiite molar ratio efiects .............. 136
Chemical compositions 137
29Si nuclear magnetic resonance 139
X-ray powder diffraction 143
Nitrogen adsorption-desorption 150
Conclusion 156
CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK ..... 158
APPENDICES
Appendix A: Electron diffraction analysis 161
Apocnoix B: Scherrer equation 163
Apocnoix C: Gas adsorption analysis 164

 

 

 

LIST OF REFERENCES 169

 

xi

 

 

 

TABLE 1.

TABLE 2.

TABLE 3.

TABLE 4.

TABLE 5.

TABLE 6.

TABLE 7.

TABLE 8.

TABLE 9.

TABLE 10.

LIST OF TABLES

Na-magadiite formulae

 

Composition of the samples formed at various stages in
the acid titration of Na-magadiite ...........................

H-magadiite formulae

 

X-ray diffraction data for Na-magadiite and H-
magadiite

 

Selected X-ray powder diffraction reflections for Na-
magadiite at various temperatures and for H-magadiite

Basal spacings and crystallinity for Na/H-magadiite
samples prepared by acid titration of Na-magadiite.
The samples correspond to those identified in the
titration curve shown in Figure 1 ...........................

 

Magadiite infrared data

Magadiite hydroxyl stretch FTIR data under vacuum at
150°C

 

Magadiite solid state 29Si nuclear magnetic resonance
data

 

Various magadiite 29Si NNHI data ..........................

12

26

30

31

36

40

46

 

TABLE 11.

TABLE 12.

TABLE 13.

TABLE 14.

TABLE 15.

TABLE 16.

TABLE 17.

TABLE 18.

TABLE 19.

TABLE 20.

TABLE 21.

 

Magadiite N2 adsorption-desorption data

oooooooooooooooo

Experimental and theoretical data for layered Na-
silicate structures according to Schwieger eta1.18

Experimental conditions for the acid-catalyzed TEOS
polymerization in the presence of alkylamines .........

Chemical composition from elemental analysis for
TEOS-hydrolysis products formed in the presence of
alkylamines

 

Chemical composition from thermogravimetric
analysis for air-dried TEOS-hydrolysis products
formed in the presence of alkylamines

0000000000000000000

Comparison of chemical composition from elemental
analysis and thermogravimetric analysis for air-dried
TEOS-hydrolysis products formed in the presence of
alkylamines

 

Infrared data of the TEOS-hydrolysis reactants and
products

 

298i NMR data of the dried and calcined products
(450°C) TEOS-hydrolysis products formed in the
presence of alkylamines

 

N2 adsorption data of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines

 

X-ray diffraction data of the air-dried and calcined
products

 

Amine and ammonium chloride salt dimensions

xiii

49

50

63

67

68

73

77

85

91

91

 

 

TABLE 22.

TABLE 23.

TABLE 24.

TABLE 25.

TABLE 26.

TABLE 27.

TABLE 28.

TABLE 29.

TABLE 30.

TABLE 31.

TABLE 32.

 

X-ray diffraction data for various
alkylammonium/alkylamine magadiites ...................

Chemical composition from elemental analysis for the
air-dried alkylamine-intercalated magadiites

Number of amines present per unit cell for the air-
dried alkylamine-intercalated magadiites .................

Validity of the C, H, N analysis

ooooooooooooooooooooooooooooo

Various cation-magadiites 29Si N MR data

Chemical composition from thermogravimetric
analysis for unwashed and ethanol-washed octylamine-
derived siloxane-intercalated magadiites

oooooooooooooooo

Chemical composition from elemental analysis for
unwashed and ethanol-washed octylamine-derived
siloxane- and silica-intercalated magadiites at various
calcination temperatures

 

X-ray diffraction data of unwashed and ethanol-
washed octylamine-derived siloxane- and silica-
intercalated magadiites at various calcination
temperatures

 

N2 adsorption data of unwashed and ethanol-washed
octylamine-derived silica-intercalated magadiites at
various calcination temperatures

oooooooooooooooooooooooooooo

Infrared data of unwashed and ethanol-washed
octylamine-derived siloxane- and silica-intercalated
magadiites

 

Experimental conditions for the study of the
polymerization of TEOS into H-magadiite

ooooooooooooo

xiv

103

108

110

111

112

116

118

119

124

129

 

 

TABLE 33.

TABLE 34.

TABLE 35.

TABLE 36.

Chemical composition from elemental analysis for
unwashed alkylamine-derived siloxane- and silica-
intercalated magadiites at the TEOS:amine:magadiite
molar ratio = 100:27zl

 

X-ray diffraction data of unwashed alkylamine-derived
siloxane- and silica-intercalated magadiites ............

N2 adsorption data of unwashed alkylamine-derived
silica-intercalated magadiites

 

Calculated alkylamine-derived silica-intercalated
magadiite microporous surface area assuming the
interlayer space is stuffed with TEOS-derived silica .

XV

138

143

154

155

 

FIGURE 1.

FIGURE 2.

FIGURE 3.

FIGURE 4.

FIGURE 5.

FIGURE 6.

FIGURE 7.

 

LIST OF FIGURES

Magadiite titration curve
(3 g of Na-magadiite in 75 mL of deionized water,
HC10.1 M, addition rate: 3 mL/min) ......................

Magadiite thermogravimetric analysis curves
a. Na-magadiite, b. sample D as identified in the
titration curve shown in Figure 1, c. H-magadiite

Magadiite differential scanning calorimetry spectra
(heating rate: 5°C/min)
a. Na-magadiite (4.5 mg), b. H-magadiite (4.7 mg) ..

Schematic illustration of the structural definitions of
basal spacing, gallery height and layer thickness .....

Na-magadiite X-ray powder diffraction pattern (pwd)
(the bottom pattern is an expanded scale of the top
pattern)

 

Na-magadiite X-ray powder diffraction pattern (film)
(the bottom pattern is an expanded scale of the top
pattern)

 

H-magadiite X-ray powder diffraction pattern (pwd) .

xvi

13

14

20

22

23

24

FIGURE 8.

FIGURE 9.

FIGURE 10.

FIGURE 11.

FIGURE 12.

FIGURE 13.

 

Na-magadiite electron rnicrographs

3. Scanning electron micrographs (left: x3,600:
1cm=2.78rtm, right: x15,000: 1cm=0.667um)

b. Transmission electron micrograph ( x48,000:
1cm=0.208p.m)

c. Electron diffraction micrograph (camera
length=83cm: 1cm=1.53 )

 

Na-magadiite X-ray powder diffraction patterns (pwd)
at various temperatures

a. room temperature, b. 75°C, c. 150°C, d. 300°C

e. back to room temperature

 

X-ray powder diffraction patterns (film) for mixed
Na/H-magadiite samples prepared by acid titration of
Na-magadiite. The samples correspond to those
identified in the titration curve shown in Figure 1

Unwashed Na-magadiite X-ray powder diffraction
pattern (film)

 

Magadiite infrared spectra (KBr pellet, room
temperature)

a. Na-magadiite, b. sample D as identified in the
titration curve shown in Figure l, c. H-magadiite

Magadiite hydroxyl stretch FI‘IR bands under vacuum
at 150°C

top: pristine pressed pellet (a: 20 mg ; b: 10 mg)
bottom: 1 mg of sample in a 100 mg KBr pressed
pellet

a: Na-magadiite, b: H- g iiite

 

xvii

28

29

32

34

37

39

FIGURE 14.

FIGURE 15.

FIGURE 16.

FIGURE 17.

FIGURE 18.

FIGURE 19.

FIGURE 20.

FIGURE 21.

FIGURE 22.

 

Magadiite solid state 29Si nuclear magnetic resonance
spectra

(delay time: 1200 seconds for Na-magadiite and 600
seconds for sample D and H-magadiite, line
broadening: 40, 12 scans)

a. Na-magadiite, b. sample D as identified in the
titration curve shown in Figure 1, c. H-magadiite

Magadiite N2 adsorption-desorption curves ; outgassing
was carried out under vacuum at 150°C

left: N2 adsorption-desorption isotherm

right: t-plot

a. Na-magadiite, b. H-magadiite ............................

Structure of makatite and the postulated structures for
octosilicate, magadiite and kenyaite(Schwieger et a1. )18
left: top view of the makatite single- -sheet surface16

right: edge- on view

 

Magadiite structure postulated by Garces et a1.22

left: edge-on view

right: top view

‘ possible arrangement for the level ......................

Hypothetical proposed sodium site in the magadiite ..

Ion exchange pillaring

 

Alkoxide polymerization pillaring .........................

Thermogravimetric analysis curves for air-dried
TEOS-hydrolysis products formed in the presence of
alkylamines

a. HA100, b. OA100, c. DA100 ............................

Infrared spectra of the starting materials (liquid film
between KBr plates)
a. TEOS, b. alkylamine: octylamine ......................

xviii

43

48

51

53

55

59

59

67

70

FIGURE 23.

FIGURE 24.

FIGURE 25.

FIGURE 26.

FIGURE 27.

FIGURE 28.

 

Infrared spectra of the air-dried TEOS-hydrolysis
products formed in the presence of alkylamines (KBr
pellet)

a. HA100, b. OA100, c. DA100 ............................

Infrared spectra of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines (KBr pellet)

a. CHA100, b. COA100, c. CDA100 .....................

Infrared spectra of silica gel Davisil-62 (KBr pellet)

298i NMR spectra of the air-dried TEOS-hydrolysis
products formed in the presence of alkylamines

(delay time: 600 seconds, line broadening: 140, 12
scans)

a. HA100, b. OA100, c. DA100 ............................

298i NMR spectra of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines

(delay time: 600 seconds, line broadening: 140, 12
scans)

a. CHAIOO, b. COA100, c. CDA100 .....................

N2 adsorption curves of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines prepared at an initial TEOS to amine
molar ratio of 7.5 ; outgassing was carried out under
vacuum at 150°C

left: N2 adsorption-desorption isotherm

right: t-plot

a. CHA200, b. COA200, c. CDA200 ....................

71

72

76

78

79

82

FIGURE 29.

FIGURE 30.

FIGURE 31.

FIGURE 32.

N2 adsorption curves of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines prepared at an initial TEOS to amine
molar ratio of 3.7 ; outgassing was carried out under
vacuum at 150°C

left: N2 adsorption-desorption isotherm

right: t—plot

a. CHA100, b. COA100, c. CDA100 .....................

N2 adsorption curves of the calcined (450°C) TEOS-
hydrolysis products formed in the presence of
alkylamines prepared at an initial TEOS to amine
molar ratio of 1.9 ; outgassing was carried out under
vacuum at 150°C

left: N2 adsorption-desorption isotherm

right: t-plot

a. CHASO, b. COASO, c. CDA50 ...........................

X-ray powder diffraction (pwd) patterns of the air-
dried and calcined (450°C) TEOS-hydrolysis products
formed in the presence of alkylamines prepared at an
initial TEOS to amine molar ratio of 7.5

left: air-dried products

right: calcined products

a. HA200 and CHA200, b. OA200 and COA200

c. DA200 and CDA200

 

X-ray powder diffraction (pwd) patterns of the air-
dried and calcined (450°C) TEOS-hydrolysis products
formed in the presence of alkylamines prepared at an
initial TEOS to amine molar ratio of 3.7

left: air-dried products

right: calcined products

a. HA100 and CHA100, b. OA100 and COA100

c. DA100 and CDA100

 

83

84

88

89

 

FIGURE 33.

FIGURE 34.

FIGURE 35.

FIGURE 36.

FIGURE 37.

FIGURE 38.

X-ray powder diffraction (pwd) patterns of the air-
dried and calcined (450°C) TEOS-hydrolysis products
formed in the presence of alkylamines prepared at an
initial TEOS to amine molar ratio of 1.9

left: air-dried products

right: calcined products

a. HA50 and CHA50, b. OA50 and COA50

c. DA50 and CDA50

 

CDA50 TEM pictures (3x190,000: 1cm=175A) .....

X-ray diffraction patterns of alkylammonium/
alkylamine-magadiites gels

top: alkylarnmonium/alkylamine-magadiites in amine
suspension

bottom: air-dried alkylammonium/alkylamine-
magadiites

a. hexylammonium/alkylamine-magadiite

b. octylammonium/alkylamine-magadiite

c. decylammonium/ alkylamine-magadiite ..............

Alkylamine intercalation into silicic acids, according to
the number of carbon atoms in their chain

a: monolayer, b: gauche-block bilayer

c: perpendicular bilayer

 

Air-dried decylamine-intercalated magadiite solid state
298i NMR spectrum

(delay time: 1200 seconds, line broadening: 40, 12
scans)

 

Thermogravimetric analysis curves for octylamine-
derived siloxane-intercalated magadiites

a. unwashed octylamine-derived siloxane-intercalated
magadiite, b. octylamine-derived siloxane-intercalated
magadiite washed once with ethanol .....................

90

93

102

106

113

 

FIGURE 39.

FIGURE 40.

FIGURE 41.

FIGURE 42.

X-ray diffraction patterns of unwashed and ethanol-
washed octylamine-derived siloxane- and silica-
intercalated magadiites

left: unwashed octylamine-derived siloxane- and silica-
intercalated magadiites (film)

right: ethanol-washed octylamine-derived siloxane-
and silica-intercalated magadiites (pwd)

a. siloxane-intercalated magadiite, b. silica-intercalated
magadiite 100°C, c. silica-intercalated magadiite
250°C, d. silica-intercalated magadiite 450°C, e. silica-
intercalated magadiite 800°C ..............................

N2 adsorption curves of unwashed octylamine-derived
silica-intercalated magadiites at various calcination
temperatures

left: N2 adsorption isotherm

right: t-plot

a. silica-intercalated magadiite 100°C, b. silica-
intercalated magadiite 250°C, c. silica-intercalated
magadiite 450°C, d. silica-intercalated magadiite
800°C

 

N2 adsorption curves of ethanol-washed octylamine-
derived silica-intercalated magadiites at various
calcination temperatures

left: N2 adsorption isotherm

right: t-plot

a. silica-intercalated magadiite 100°C, b. silica-
intercalated magadiite 250°C, c. silica-intercalated
magadiite 450°C, (1. silica-intercalated magadiite
800°C

 

Dependence of the microporous and non-microporous
surface areas on the calcination temperature

left: unwashed octylamine-derived silica-intercalated
magadiites

right: ethanol-washed octylamine-derived silica-
intercalated magadiites

 

120

122

123

125

 

 

FIGURE 43.

FIGURE 44.

FIGURE 45.

FIGURE 46.

FIGURE 47.

Infrared spectra of unwashed octylamine-derived
siloxane- and silica-intercalated magadiites (KBr
pellet, room temperature)

a. siloxane-intercalated magadiite, b. silica-intercalated
magadiite 100°C, c. silica-intercalated magadiite
250°C, d. silica-intercalated magadiite 450°C, e. silica-
intercalated magadiite 800°C ..............................

Infrared spectra of ethanol-washed octylamine-derived
siloxane- and silica-intercalated magadiites (KBr
pellet, room temperature)

a. siloxane-intercalated magadiite, b. silica-intercalated
magadiite 100°C, c. silica-intercalated magadiite
250°C, (1. silica-intercalated magadiite 450°C, e. silica-
intercalated magadiite 800°C ..............................

Hydroxyl stretch FI‘R bands under vacuum at 150°C
for unwashed octylamine-derived silica-intercalated
magadiite (450°C)

(1 mg of sample in a 100 mg KBr pressed pellet)

298i NMR spectra of unwashed octylamine-derived
siloxane- and silica-intercalated magadiites

(delay time: 600 seconds, line broadening: 140, 12
scans)

a. siloxane-intercalated magadiite, b. silica-intercalated
magadiite 100°C, c. silica—intercalated magadiite
250°C, d. silica-intercalated magadiite 450°C, e. silica-
intercalated magadiite 800°C ..............................

29Si NMR spectra of ethanol-washed octylamine-
derived siloxane- and silica-intercalated magadiites
(delay time: 600 seconds, line broadening: 140, 12
scans) -

a. siloxane-intercalated magadiite, b. silica-intercalated
magadiite 100°C, c. silica-intercalated magadiite
250°C, d. silica-intercalated magadiite 450°C, e. silica-
intercalated magadiite 800°C

 

xxiii

127

128

131

133

134

FIGURE 48.

FIGURE 49.

FIGURE 50.

2S’Si NMR spectra of unwashed amine-derived
siloxane-intercalated magadiites with initial
TEOS:amine:magadiite molar ratio = 100:27:1

(delay time: 600 seconds, line broadening: 140, 12
scans)

a. hexylamine-derived: HAMAG100, b. octylamine-
derived: OAMAG100, c. decylamine-derived:
DAMAGlOO

 

298i NMR spectra of unwashed amine-derived silica-
intercalated magadiites with initial
TEOS:amine:magadiite molar ratio = 100:27:1

(delay time: 600 seconds, line broadening: 140, 12
scans)

a. hexylamine-derived: CHAMAGIOO, b. octylamine-
derived: COAMAGlOO, c. decylamine-derived:
CDAMAGlOO

 

X-ray powder diffraction (pwd) patterns of unwashed
alkylamine-derived siloxane- and silica-intercalated
magadiites with initial TEOS:amine:magadiite molar
ratio = 50:27:1

left: unwashed alkylamine-derived siloxane-
intercalated magadiites

right: unwashed alkylamine-derived silica-intercalated
magadiites

a. hexylamine-derived: HAMAGSO and CHAMAGSO,
b. octylamine-derived: OAMAGSO and COAMAGSO,
c. decylamine-derived: DAMAG50 and CDAMAGSO

140

141

144

FIGURE 51.

FIGURE 52.

FIGURE 53.

FIGURE 54.

X-ray powder diffraction (pwd) patterns of unwashed
alkylamine-derived siloxane- and silica-intercalated
magadiites with initial TEOS:amine:magadiite molar
ratio = 100:27:1

left: unwashed alkylamine-derived siloxane-
intercalated magadiites

right: unwashed alkylamine-derived silica-intercalated
magadiites

a.hexylamine-derived:HAMAG100 and CHAMAGlOO
b.0cty1amine-derived:0AMAG100 and COAMAGlOO
c.decylamine-derived:DAMAG100 and CDAMAGlOO

X-ray powder diffraction (pwd) patterns of unwashed
alkylamine-derived siloxane- and silica-intercalated
magadiites with initial TEOS:amine:magadiite molar
ratio = 200:27:1

left: unwashed alkylamine-derived siloxane-
intercalated magadiites

right: unwashed alkylamine-derived silica-intercalated
magadiites

a.hexylamine-derived:HAMAG200 and CHAMAG200
b.0cty1amine-derived:OAMAGZOO and COAMAG200
c.decylamine-derived:DAMAG200 and CDAMAGZOO

Dependence of the basal spacing of the unwashed
alkylamine-derived siloxane- and silica-intercalated
magadiites on TEOS:magadiite molar ratio

left: unwashed alkylamine-derived siloxane-
intercalated magadiites

right: unwashed alkylamine-derived silica-intercalated
magadiites

a. hexylamine-derived: HAMAG and CHAMAG

b. octylamine-derived: OAMAG and COAMAG

c. decylamine-derived: DAMAG and CDAMAG

Schematic representation of the role of alkylamine on
the basal spacing of unwashed alkylamine-derived
siloxane- and silica-intercalated magadiites ............

145

146

147

 

FIGURE 55.

FIGURE 56.

FIGURE 57.

FIGURE 58.

FIGURE 59.

FIGURE 60.

FIGURE 61.

 

N2 adsorption curves of unwashed alkylamine-derived
silica-intercalated magadiites with initial
TEOS:amine:magadiite molar ratio = 50:27:1"

left: N2 adsorption isotherm

right: t-plot

a. hexylamine-derived: CHAMAGSO, b. octylamine-
derived: COAMAGSO, c. decylamine-derived:
CDAMAGSO

 

N2 adsorption curves of unwashed alkylamine-derived
silica-intercalated magadiites with initial
TEOS:amine:magadiite molar ratio = 100:27:1

left: N2 adsorption-desorption isotherm

right: t-plot

a. hexylamine-derived: CHAMAGIOO, b. octylamine-
derived: COAMAGlOO, c. decylamine-derived:
CDAMAGlOO

 

N2 adsorption curves of unwashed alkylamine-derived
silica-intercalated magadiites with initial
TEOS:amine:magadiite molar ratio = 200:27:1

left: N2 adsorption isotherm

right: t-plot

a. hexylamine—derived: CHAMAG200, b. octylamine-
derived: COAMAG200, c. decylamine-derived:
CDAMAG200

 

Electron diffraction analysis parameters .................
Adsorption-desorption isotherm ..........................

BET curve

 

t-plot

 

151

152

153

161

164

166

167

ABBREVIATIONS

DSC: Differential Scanning Calorimetry
DTA: Differential Thermic Analysis

F'I‘IR: Fourier Transform Infra Red

NMR: Nuclear Magnetic Resonance

SEM: Scarming Electron Microscopy

TEM: Transmission Electron Microscopy
TEOS: Tetraethylorthosilicate Si(OCH2CH3)4
TGA: Thermo Gravimetn'c Analysis

TMS: Tetrarnethylsilane Si(CH3)4

XRD: X-Ray Diffraction

b: broad (XRD)

film: film with preferred 001-orientation (XRD)
n: narrow (XRD)

pwd: randomly oriented powder (XRD)

s: strong (FTIR)

sh: shoulder (FI‘IR)

vs: very strong (FI‘IR)

vs.: versus (NMR)

vw: very weak (FI‘IR)

w: weak (FTIR)

xxvii

RESEARCH OBJECTIVES

There is a considerable interest in obtaining microporous and
mesoporous materials of controlled sizes. The outstanding properties of
zeolites as adsorbent and catalysts have encouraged researchers to search
for new microporous and mesoporous materials, especially with pore
Openings larger than 10 A. Layered materials can lead to porous materials
when the sheets are propped apart by intercalative species. The general
pillaring method is to insert gallery guests that are sufficiently robust to
expand the host layer and laterally separated to provide a two-dimensional
pore structure.1 A new route to create microporous materials has been
recently discovered that involves the hydrolysis and condensation-
polymerization of metal alkoxides in the gallery of alkylammonium-
exchanged derivatives of layered host.2-3o4 Unlike conventional pillaring
methods}.6 the use of metal alkoxides allows for the formation of the
pillars in situ, reducing the number of steps in the pillaring process. This

very promising new approach produces porous materials, but the processes

 

2

involved are still not understood. The goal of this work is to gain a better
understanding of the reaction in order to predict pore sizes depending on
the reaction conditions. To this end, magadiite has been intercalated by the
reaction of tetraethylorthosilicate (TEOS) in presence of alkylamines as
swelling agents.

Na-magadiite (Na2 Si14 O23 (OH)2 . 8 H20), and more particularly,
its acid form, H-magadiite has been chosen in this study as a starting
layered material. Indeed, magadiite belongs to the hydrous sodium silicate
minerals family. Conclusions drawn for magadiite are likely to be
applicable to other members of the family. Furthermore, magadiite is
easily synthesized7 by hydrothermal reaction and its swelling properties
have been extensively studied.8 However, as magadiite structure is still
unknown, we first decided to carefully study Na-magadiite, and its acid
form H-magadiite, to better elucidate the interlayer space where the
intercalation will occur.

The silica intercalation reaction makes use of in situ hydrolysis and
condensation-polymerization of TEOS. After obtaining some knowledge
about the nature of the layered host, we investigated the sol-gel process
involved during the intercalation of TEOS by conducting alkylammonium-
mediated TEOS hydrolysis experiments in the absence of magadiite.

Finally, we studied the intercalation of magadiite by TEOS

hydrolysis.

 

 

EXPERIMENTAL

SYNTHESIS
Na-magadiite

Na-magadiite was prepared by the reaction of NaOH and Si02 under
hydrothermal conditions according to the procedure of Fletcher and
Bibby.9 The starting molar ratios were 1:3:50 for NaOH:Si02:H20.
Deionized water (300.0 mL, 16.7 mol) was used to dissolve sodium
hydroxide (13.3 g, 0.33 mol) in a Teflon-lined l-liter Parr reactor.
Davisil-62 silica (60.0 g, 0.99 mol) was added to the 1.1 M NaOH solution.
The suspension was heated to 150°C at a rate of 1°C/min, and stirred at that
temperature for 42 hours. The solid Na-magadiite product was separated
by centrifugation after 10 min at 10,000 rev/min. It was then washed once
with 200 mL of deionized water before being redispersed in deionized

water and air-dried, at room temperature, on a polyethylene film.

H-magadiite

H-magadiite was obtained by titration of synthetic Na-magadiite with
HCl according to the method described by Lagaly et a1..10 Na-magadiite
was dispersed in water at a ratio of 25 ml of deionized water per gram of
Na-magadiite (9<pH<10). The suspension was titrated with a 0.1 M
solution of hydrochloric acid to lower the pH to less than 1.9. The titration
rate was about 3 m1/min. The pH of the suspension was maintained at a
value of 1.9 for a period of one day. H-magadiite was separated by
centrifugation after 10 min at 10,000 rev/min and washed with deionized
water until the supernatant was free of chloride ions, according to the
silver nitrate test. Three washings with 30 mL of deionized water per gram
of magadiite were necessary. The solid was then redispersed in deionized
water and air-dried, at room temperature, on a polyethylene film. The
mass of Na-magadiite titrated ranged from 2 g (1.9’1‘10'3 mol) to 42 g
(3.9*10‘2 mol).

Amine-solvated magadiite

Three amines were used to prepare amine-solvated magadiite:
hexylamine, octylamine and decylamine. The amine was added to H-
magadiite in an approximately 27:1 aminezmagadiite molar ratio to form a
gel. For the silica-pillaring experiments, the resulting amine-solvated
magadiite was freshly prepared and used after 5 to 10 min of aging in the
excess amine. For the amine-swelling experiments, the amine was allowed
to react with H-magadiite for one day in a sealed vial before being air-

dried on a glass plate at room temperature.

5
Siloxane-intercalated magadiite
The wet amine-solvated magadiite gel was mixed with
tetraethylorthosilicate (TEOS) in three different TEOS:amine:magadiite
molar ratios: 50:27:1, 100:27:1 and 200:27:1. The suspensions were stirred
for one day before being separated by centrifugation after 10 min at
10,000 rev/min. The siloxane-intercalated magadiite reaction products
were redispersed in ethanol and air-dried on a glass plate at room
temperature. Where indicated, the product was washed once with absolute
ethanol. The mass of H-magadiite used for these preparations ranged from
0.5 g (5.6”‘10'4 mol) to 20 g (220*10'4 mol).

Silica-intercalated magadiite

To obtain a silica-intercalated magadiite, the siloxane-intercalated
magadiite was calcined at 450°C for 4 hours in a programmable oven at a
heating rate of 5°C/min. The powder was slowly cooled to room
temperature. A preliminary survey of calcination temperatures at 100°C,
250°C, 450°C and 800°C indicated that the optimal surface area was
achieved at 450°C.

TEOS-derived silica precursor

An alkylamine/alkylammonium solution was prepared by adding the
desired amount of concentrated hydrochloric acid to the neat liquid amine.
Tetraethylorthosilicate (TEOS) was then added to the ammonium/amine
solution in three different TEOS:amine:HCl:H2O molar ratios: 50:27:2z7,
100:27:2:7 and 200227227. The suspension was stirred for one day before
being separated by centrifugation after 10 min at 10,000 rev/min. The

6

solution containing the TEOS-derived silica precursor was then air-dried

on a glass plate at room temperature to obtain a white powder.

TEOS-derived silica
To obtain a TEOS-derived silica, the TEOS—derived silica precursor
was calcined at 450°C for 4 hours in a programmable oven at a heating rate

of 5°C/min. The powder was slowly cooled to room temperature.

ANALYSIS
Chemical analyses
Elemental analysis (Na, Si)

A 50-mg of sample was fused with 300 mg (6*10'3 mol) of lithium
borate for 10 min at 1000°C in a preignited graphite crucible. The
resultant molten glass was transferred to 50 ml of 6% wt nitric acid
solution. This solution was stirred until dissolution of the glass was
complete and then further diluted to 100 mL with deionized water. The
solution was analyzed by induced coupled plasma atomic emission
spectrometry (ICP), at the Inorganic Laboratory of the Michigan State
University Toxicology Department, on a Jamell-Ash atom-comp

instrument.

C, H, N analyses
Carbon, hydrogen and nitrogen analyses were performed at the
University of Illinois Microanalysis Laboratory by oxidation of the

elements and chromatographic detection of the gases produced.

 

Thermal analyses
Thermogravimetric analysis (TGA)

The thermogravimetric curves were obtained on a Cahn TG system
121 analyzer. The starting temperature was held at 30°C for 5 min. The
sample was then heated to 800°C at 5°C/min. The furnace cooling fan was
switched off at 500°C.

Differential scanning calorimetry (DSC)
The calorimetric data were obtained on a DuPont-9900 thermal

analyzer, at a heating rate of 5°C/min .

Structural analysis
X-ray powder diffraction (XRD)

The basal spacings were determined by X-ray powder diffraction,
using a Rigaku Rotaflex diffractometer equipped with a rotating copper
anode producing Ka radiation. The current and voltage used to operate the
anode were 45 mA and 100 kV, respectively. The scan speed was set to
2°/min. The temperature dependences of the XRD patterns were performed
in situ in a specially designed high temperature aluminum attachment.
Samples were prepared by spreading the powder on powder glass holders,
or, when indicated, by air-drying the sample suspension on a microscope

slide.

Nitrogen adsorption-desorption (surface and porosity)

Nitrogen adsorption-desorption experiments were performed on an

Omnisorb 360 CX sorptometer after outgassing the samples for 12 hours at
the specified temperature (75°C, 100°C, 150°C and 300°C) under vacuum.

 

8

Helium gas was used for volumetric calibration. Surface area values and

pore data were obtained by the BET equation and the t-plot method.ll

Infrared spectroscopy (FTIR)

The Fourier transform infrared spectra were recorded on an IBM
IR44 spectrophotometer, using the KBr pressed pellet technique or by
preparing pressed pellets of the pure material. Some experiments were run,
under a flow of helium gas, under reduced pressure, at 150°C. The sample

was cooled in situ before recording the spectrum.

2”Si nuclear magnetic resonance (NMR)

The solid state 298i 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 zircon rotors was used for the
measurements. The spectra were obtained using a 4.6 us 90° pulse width, a
4 kHz spinning rate, a delay time of 600 s and by accumulating 12 scans,

unless otherwise specified.

Scanning electron microscopy (SEM)
The scanning electron microscopy images were obtained on a JEOL

JSM-25 microscope at the Michigan State University Center for Electron

Optics.

Transmission electron microscopy (TEM)

The transmission electron microscopy images and the electron
diffraction patterns were recorded on a JEOL JEM-lOOCX microscope at
the Michigan State University Center for Electron Optics.

MAGADIITE CHARACTERIZATION

INTRODUCTION

The hydrous sodium silicate minerals family includes four members:
kanemite (Na H Si2 O4 (OH)2 . 2 H2O),12‘l4 makatite (Na2 Si; 03 (OH)2 . 4
H2O),12-l6 magadiite (Na2 Si” 023 (OH)2 ‘. 8 H2O)7o12.15-17 and kenyaite
(Na2 Si22 O44 (OH)2 . 9 H2O).13.15-13 Some of these minerals have been
synthetically prepared by early workers.19-20 Magadiite and kenyaite were
first found in 1967 by Eugster, near Lake Magadi, Kenya.17 Makatite was
found in 1969 by Sheppard and Gude, at the same place”; its name comes
from the Masai“ (Kenya) word emakat, which means soda, in reference to
the high sodium content of the mineral. Kanemite was found in 1971 by
Maglione, near Lake Chad, Chad, in the Kanem region.12

Other silicates are more or less related to this silicate mineral family.

These include: natrosilite (Na2 Si2 05)l3.2l which is an anhydrous sodium
silicate mineral ; sodium octosilicate (Na20 . 8 SiO2 . 9 H2O)163233 which

is only known as a synthetic hydrous sodium silicate, and some synthetic

 

10
potassium silicates such as K H Si2 05,24 K2 Sig 029 . x H20 and K2 Si2o O41
. y H2035

The above sodium silicate minerals have been studied in their natural
and synthetic forms. They are easily synthesized by hydrothermal reaction
of aqueous sodium hydroxide with silica at various SiO7/Na20 molar
ratios. Reaction composition, reaction temperature and reaction time
determine the final product.8»22,25 Among this family of silicates, only
magadiite has a natural silicic acid analog, named silhydrite.2-6 The
materials exhibit a broad range of properties, such as, the sorption of
interlamellar water and polar organic molecules, cation exchange of
internal sodium cations, intracrystalline swelling, grafting and
transformation into crystalline layered silicic acids by proton
exchange.7.8-13J6,25v27-32 Those specific properties promote their application
as adsorbents, catalysts, catalyst supports, cation exchangers or molecular
sieves.

Owing to their intercalation properties and potential use for material
application, the layer structure of these silicates has been well studied. The
negative charge of the layers is compensated by sodium cations, which are
solvated by water molecules. Adjacent layers are held together by
electrostatic interactions between the layers and the gallery cation and/or
by hydrogen bonding with interlayer water. 14.16.330.33.“ However, except
for makatite,16 their crystal structure is still unknown due to the lack of
suitable single crystals. Recently, some structures have been proposed,2235
based on spectrosc0pic data, such as infrared spectroscopy and solid state
1H, 23Na, 2S’Si nuclear magnetic resonance.

In this work, we obtain a better understanding of the magadiite

composition and structure by carefully studying Na-- a” hire, H- o iiite

 

 

 

 

11

and the titration process that leads to the formation of the proton-

exchanged silicic acid derivative.

RESULTS AND DISCUSSION
Syntheses
Like the other sodium silicate minerals, Na-magadiite is easily

synthesized by hydrothermal reactionz9

4.7 NaOH + 14 SiO2 + 233 H2O -> Na2 Si14 O23 (OH)2 . 8 H2O + NaOHaq
(150°C, 42 hours)

The suspension containing Na-magadiite is centrifuged. If the white powder
is redispersed in water and air-dried without further treatment, unwashed
Na-magadiite containing excess NaOH is obtained. However, if the white
powder is washed once with 200 mL of water before being redispersed in
water and air-dried, Na-magadiite, more or less close to the ideal
composition is obtained (Table 1). The magadiite acid form H-magadiite is
obtained by acid titration of Na-magadiite. Figure 1 illustrates the acid
titration curve for Na-magadiite and the samples that were analyzed at

various stages of the titration.

 

12

TABLE 1. Na-magadiite formulae

 

 

 

 

 

 

 

 

sample formula
1 N 212.0 114 (OH)2 . . H2
this work unwashed Na43 S114 023 (OH)4,3 . 87 HRS
synthetic
this work synthetic Na23 Si14 O23 (OH)2_3 . 7.8 H20
this work synthetic Na1_9 Si14 023 (0H)1,9 . 7.0 H20
this work synthetic N315 Sil4 a: (OI-1)] g . 5.5 H20
Dailey4 unwashed N323 114 ( H)2.3 - . H2
synthetic
Rojo et al.34 *namml: CA Na23 Si14 028 (OH)2.3 . 89 H20
Garces et 211.22 *natural: CA N822 Sim 028 (OH)2.2 . 46 H20
Lagaly et al.7 *natural: CA Nam Si14 O23 (OH)2] 10.3 H2O
McAtee et 8136 *namral: CA N821 Si14 023 (OH)2,1 . 7.5 H20
Garces et al.22 synthetic N321 $14 028 (OH)2.1 - 3-2 H20
Lagaly ct 31.7 synthetic N820 Si“ 023 (OH)2,0 . 10.3 H20
Schwieger et a1.18 53'"me N320 $14 028 (OH)2.0 - 91 H20
McAtee ct aL35 *nann‘alt 0R Nazo $14 028 (OH)2.0 . 8-3 H20
McAtee et al.35 *namml: CA N320 Si14 028 (OI-02.0 . 32 H20
McCulloch19 synthetic N a2.o Si14 023 (OH)2.o . 7 .8 H20
Garces et a1.22 synthetic N320 Si14 028 (OH)2.0 . 34 H20
Eugster” *natural: K Na1_9 Si14 023 (0H)1,9 . 7.8 H20
Dailcy“ Synthetic N313 Si14 028 (0H)1.7 . 76 H20
: Kenya, CA: California, OR: Oregon

 

Chemical compositions
The formulae have been determined using the elemental analysis and
the thermogravimetric data. From the general Na-magadiite formula:
Nan Si14 Ox (OH)y . 2 H2O
we are going to obtain the formulae ‘for Na/H-magadiites (Table 2):
Nan Hh Sig 023 (0H),. . 2 H2O

 

13

 

 

 

 

14 v u u I
13 '
12" '
11- -
1017A ‘
9. .B -
8' C D ‘
PH7- l3mm“?!church...”E ‘
6' 93'F ‘
5' n "
4‘ n -
3- is -
2- q‘b‘u‘xnflllun' -4
1-
0 l l l
0 1 3 4

mol H+/mol magadiite

FIGURE 1. Magadiite titration curve
(3 g of Na-magadiite in 75 mL of deionized water,
HCl 0.1 M, addition rate: 3 mL/min)

The molar mass M; is:

Mt = n*23.0 + 14*28.1 + x*l6.0 + y*(16.0+1.0) + z*(2*1.0+16.0)

The elemental analysis provides the molar ratio of silicon to sodium. Let's
consider the ratio 14m.

The thermogravimetric curve (Figure 2) shows that the weight loss can be
divided into three main steps : room temperature to 170°C, 170°C to 450°C
and 450°C to 800°C. We assigned the mass loss below 170°C to physisorbed

 

14
and interstitial water (z), and the loss between 170°C and 800°C to
condensation-dehydroxylation of silanols groups (y). The assignment of the
high temperature limit for water loss differs slightly from those found in
the literature. Eugster17 took 110°C, Rojo et a1. 200°C30-33 or 250°C34 and
Dailey4 200°C for water loss from Na-magadiite and 300°C from H-
magadiite. We limited the water loss to temperature below 170°C because it

matches both Na-magadiite and H-magadiite thermogravimetric curves.

 

 

% weight remaining

 

 

 

80 l r r 1 r r 1
0 100 200 300 400 500 600 700 800
temperature °C

FIGURE 2. Magadiite thermogravimetric analysis curves
a. Na-magadiite, b. sample D as identified in the

titration curve shown in Figure 1, c. H-magadiite

15
Let's consider Z, the weight loss below 170°C:
Nan Si14 Ox (OH), . 2 H2O --------- > Nan Si14 Ox (OH),
z=1s0*fi;
For Y, the weight loss between 170°C and 800°C, we have:
Nan Si14 Ox (OH),

--------- > Nan Si14 O(x+y/2)

Y=90*fi;

The fourth equation needed to solve the four-unknown system is derived
from electroneutrality:

0 = n*1 + 14*4 + x*(-2) + y*(-1)
The solutions to this set of equations are:

31*n + 60.l*14

1-Y-Z
M
_ _t
z-Z*18
y=Y*M'§‘
n+4H4-y

16
The parameter x is then adjusted to 28, with h = 28-x and y' = y-28+x

to obtain the final formula: Nan Hh Si14 O23 (OH),- . 2 H2O

The formulae found for Na-magadiite by this method (Table 1) are
in good agreement with those generally found in the literature. We can
notice that Lagaly et al.7 generally find high water contents while Garces et
a1.22 find low water contents for both natural and synthetic samples. The
chemical compositions depend on two parameters: the sample by itself and
the analytical method. The synthetic route and washing procedure,737 as
well as the conditions used to store the product will give products with
various compositions. Too much sodium can be explained by some
occluded NaOH molecules ; a low sodium content can be understood as the
consequence of hydrolysis due to washing with water.34 Those variations of
sodium content in comparison to the theoretical value have also been
observed for kanemite.l3 Enhanced water content can be explained by an
increased hydrophilicity of the surfaces. A low water content can only be
understood by cross-linkage between the layers which impedes the water
molecules from occupying the gallery. We can see, for our samples, that
the amount of sodium can be either smaller or larger than the theoretical
value of 2. The large amount of sodium can be explained by the sodium
hydroxide being trapped between the sheets during the synthesis. The small
amount of sodium is due to too much washing. That is why the unwashed
Na-magadiite: Na4.3 Si14 O23 (OH)4_3. 8.7 H20 (that can be alternatively
written as: Na2_o Si14 O23 (OH)2,0. 8.7 H20. 2.3 NaOH) looses more than

half its sodium content upon washing. Then, the washed product Na-

 

17
magadiite: Nam Si“ 023 (OH)1,9. 7.0 H20 has a smaller amount of sodium

in comparison to the expected value of 2.

According to Figure 1, the titration of Na-magadiite occurs in two
steps. The first inflexion point occurs at pH=8.5, for the addition of 0.3
mol H+/mol magadiite (14%), and the second one occurs at pH=4.3, for the
addition of 2.2 mol H+/mol magadiite (100%). This two-step titration
curve is characteristic of the hydrous sodium silicates: for kanemite,l3
pH=8.5 (70%) and pH=5 (100%) and for sodium octosilicate,23 pH=7.0
(50%) and pH=4.5 (100%). Rojo et al.34 found for magadiite: pH=9 (5%)
and pH=5 (100%). This difference in percent exchanged at the first
inflexion point is due to the variation in the amount of occluded sodium

hydroxide present in between the layers.

TABLE 2. Composition of the samples formed at various stages in the acid
titration of Na-magadiite

 

 

 

 

 

 

sample formula composition %H+
N320: Si02: H20 exchange

A- Nam no.0 3114 023 ton)” . 7.0 H20 0.95 : 14 “z 7.95 ' 0.0
B Nam H03 Sig 023 (OH)1.5 . 6.4 H20 0.70: 14 : 7.30 12.5
C N313 H04 Sig 023 (OH)1.7 . 5.9 H20 0.65 : l4 : 6.95 23.6
D Nao_7 H1_1 Si14 023 (0H)13 . 3.4 H20 0.35 : 14 z 4.85 61.1
E NaoJ H1_7 Si14 O23 (OH)1,3 . 1.0 H20 0.05 : 14 : 2.75 94.4
F NaoJ H13 Si14 O23 (OI-D19 . 0.8 H20 0.05 : 14 : 2.65 94.7
G:

 

 

 

N H1, Si1402 (0mm.0.7 H20 0.00: 14 22.50 100.0
' A: N a-magadute, U: H-magadrrte

Several trends can be noted from the compositions isolated from the

titration of Na-magadiite (Table 2). The amount of hydroxide remains

18
almost constant while the sodium content keeps on decreasing with
increasing added H+. The amount of sodium, hydrogen and water are
correlated : the disappearance of sodium and water is compensated by the
appearance of hydrogen. In the first stage of the reaction, corresponding to
the transition from sample A to B (Figure 1), the amounts of sodium and
hydroxide decrease in relatively the same proportions, while the hydrogen
content remains almost constant. This corresponds to the titration of the
sodium hydroxide trapped in between the layers during the Na-magadiite
synthesis.34 Upon further titration to pH~6.6, the amount of hydroxide
remains constant, but the sodium amount decreases sharply while the
hydrogen amount increases (compare samples B to E). The plateau on the
titration curve is assigned to the sodium-proton exchange reaction. At the
same time, the water content exactly follows the decrease in sodium: the
water molecules leave the solid together with the sodium cations. It is an
evidence for the presence, at this point, of water molecules only in the
form of an hydration sphere around the sodium cations. The last step
(samples E to G) corresponds to the second equivalence point. The
hydroxide concentration remains constant at its initial value. Gradually, the
sodium content reaches 0 while the hydrogen content reaches the initial
sodium value, and, the water content still decreases a little. During this
step, the exchange process is completed, but the water content decreases
from 1.0 to 0.7. Our chemical composition for H-magadiite has a relatively
low water content in comparison to the ones found in the literature (Table

3).

19
TABLE 3. H-magadiite formulae

 

 

 

 

 

 

sample Tormula
Si02 : H20

this work 14 : 2.5

Bailey“ 14 : 2.0

Rojo et 31.33 14 : 3.7

2E1, et al.23 14 : 5.4

 

According to the Na-magadiite formula, there are 1.9 moles of
sodium per Si14 unit. However, according to the titration curve, 2.2 moles
of H+ are required to titrate one mole of Na-magadiite (0.3 moles for
interlayer sodium hydroxide and 1.9 moles for sodium cations). There is
thus a 14% deviation in the sodium content, which might be due in part to
experimental errors and to titration of dissolved CO2. Contamination by
C02 is plausible since the titration was carried out over a period of one

day.

To better understand the various slopes before 170°C on the Na-
magadiite TGA curve, we ran some DSC experiments up to 300°C. The
Na-magadiite and H-magadiite DSC curves are displayed in Figure 3. H-
magadiite DSC curve only exhibits a weak endothermic phenomenon
starting at about 75°C. This thermal process is assigned to water loss from
external surfaces. Its weak intensity reflects the H-magadiite low water
content: 14 SiO2 . 2.5 H20. The Na-magadiite DSC curve shows five
distinct endothermic phenomena at the approximate onset temperatures of
120°C, 155°C, 180°C, 205°C and 240°C. Kanemite DTA analysis12 to
800°C displayed three endothermic peaks at 160°C, 220°C and 600°C and

20

 

155°C

189°C 205°C 240°C

   

-045.

heat flow (W/g)

-i.O-‘

 

‘04-

715°C

heat flow (W/g)
s s

-O.4~

~05 . 0 fl 4 . . .
c so 100 150 200 250 ' I Ysco

temperature °C

 

 

 

FIGURE 3. Magadiite differential scanning calorimetry spectra
(heating rate: 5°C/min)
a. Na-magadiite (4.5 mg), b. H-magadiite (4.7 mg)

21
one broad exothermic peak at 655°C. The endothermic peaks have been

assigned respectively to elimination of water molecules, dehydroxylation of
silanol groups and a further dehydration ; the exothermic peak has been
assigned to the recrystallization of the amorphous product. Makatite DTA
analysis” to 1000°C displayed five endothermic peaks at 80°C, 100°C,
185°C, 530°C and 810°C and one broad exothermic peak at 675°C. So, like
makatite, Na-magadiite displays three peaks below 200°C which have been
assigned to water desorption. The 120°C and 155°C peaks are tentatively
assigned to external water and excess water in the interlayer space,
respectively. The sharp and intense peak at 180°C can be assigned to the
loss of water from the sodium hydration sphere. The last two peaks at
205°C and 240°C might be due to structural rearrangements accompanied
by dehydroxylation, because of the stabilizing effect loss of the hydration

sphere.

X-ray powder diffraction

A careful study of the oriented powder X-ray diffraction patterns of
Na-"Mg'uiiite, H-m°g°diite and intermediate samples (samples B to F) has
been achieved. The basal spacing, or d-spacing, or d001, given by the X-ray
diffraction analysis, is the sum of the layer thickness (11.2 A)38 and the
gallery height (Figure 4).

22

 

layer thickness d 001
, d-spacing
gallery gallery height basal spacing

 

FIGURE 4. Schematic illustration of the structural definitions of
basal spacing, gallery height and layer thickness

Na-magadiite X-ray powder diffraction patterns have been recorded
using two preparatory methods: 1) a more or less randomly oriented
powder (pwd) and 2) an air-dried sample on a microscope slide in order to
make a film with preferred 001-orientation (film). The patterns are
displayed in Figures 5 and 6. The spacings are gathered with the indexed
reference38 in Table 4. Also displayed in Table 4 are the spacings for H-
magadiite (pwd). The H-magadiite X-ray powder diffraction pattern is

shown in Figure 7.

23

cm

3533 m8 2: no 28m caveman“. 5 mm E33 8885 28
€39 E839 couoflumv c333 harx Sufiawmfiéz .w mMDOE

on O».

p u - b -

ow op

0

am
pm

 

 

23w; _ Sn.”

:3
86
can

:v 2L.

wfin

 

 

 

 

 

 

03.
N
o
Kirsuoiur axnepr o g Kirsueiur QAIJBPJ o
m

<od~

 

 

 

O
O
0
V3
F‘

24

9533 QB 05 no 28m @3530 :a 3 88:3 8980 23
$55 Ease negate Sega 3-x 32238-2 .o mmaoa

 

 

 

 

 

 

 

 

 

 

g misusing omelet 0 § Kirsumur oAnejor O
m M

cm
ow om 0v Om ON A: o
VNwA /: v
P. mod
_ m ohm
owwfi 25 ”in
36
mm
c Q _ o_m . 0.? L pm _ o.N . 0.? . o
‘ lllt|<l j
36 3. b
< 0.2

 

 

     

 

 

838 58::— cone—“.56 c3309 wasx oummvamafii .5. mmDOE

cm
on 9. pm

coda SON

 

NV.

36
2.6
:6

 

ann

 

< 92

O

 

 

§ linsuowt simmer O
M

 

 

 

26

TABLE 4. X-ray diffraction data for Na-magadiite and H-magadiite

 

film—

 

 

 

HHp—ap—sp—ap—A

b-‘Lfi

7.78

5.18

1.941
1.824
1.722

1.635
1.558

 

Na
d (A) {-71%1
5.

9

10

3—5

D—iIIHIIp—n

 

pwd

 

002 5.59

003 3.70
3.42
3.19

hkO 2.503

2.206

1.847
1.801

1.592

 

 

H
hkl d (A) 21%]
001 1 . l

45

14
39

96

 

 

27

Table 4 compares the spacings for the two Na-magadiite X-ray
powder diffraction sample preparations. The randomly oriented powder
sample affords for more crystallographic data, but the film preparation
allowed us to resolve a peak at 3.19 A that was obscured by a 3.15 A
reflection. From the randomly oriented powder pattern in Figure 5, we see
that the two peaks overlap because of the broad base of the sharp 3.15 A
peak, but the line could not be determined. Some refection lines for Na-
magadiite are missing in comparison with the lines found in the X-ray
powder diffraction handbook data,39 but, interestingly, there is one extra
reflection line at 4.11 A. We first thought it could be due to an impurity,
but this spacing corresponds closely to the measured a=‘b=4.16 A value for
the electron diffraction pattern shown in Figure 8c (Appendix A). The
electron diffraction pattern also gave us a [3 angle of 93.94°. Furthermore,
that reflection line was also present in the Na-magadiite patterns taken at
different temperatures (Figure 9 and Table 5) and in the H-magadiite
pattern. We thus tentatively assigned this reflection, as well as the 2.503 A
line, to an hk0 reflection line. Unfortunately, we were not able to index
satisfactorily our Na-magadiite X-ray powder diffraction data with the
measured parameters a=b=4.16 A, C: 15.6 A and B=93.94°. Thus, in our
following study, we will still use the reference values found by Brindley38
instead of the ones found by Garces et a1.22 and Eugster17. For a
monoclinic system, Garces et al.22 found for its hypothetical unit cell
a=27.50 A, b=9.20 A, c=7.52 A, B=101°, while Brindley33 used a=b=7.25
A, c=15.69 A, B=96.80°, even though he did not try the possibility for the
unit cell to be triclinic. Earlier, Eugster17 assumed a tetragonal symmetry
to obtain a=b=12.62 A and c=15.57 A. The Na-magadiite X-ray powder

diffraction pattern is characterized by a cluster of five peaks between 3.63

 

FIGURE 8. Na-magadiite electron micrographs
a. Scanning electron rnicrographs (left: x3,600: 1cm=2.78um, right:
x15,000: 1cm=0.667p.m)
b. Transmission electron micrograph (x48,000: 1cm=0.208um)
c. Electron diffraction micrograph (camera length=83cm: 1cm=1.53A)

 

29

 

relative intensity
9%

 

 

 

 

FIGURE 9. Na-magadiite X-ray powder diffraction patterns (pwd) at
various temperatures

a. room temperature

b. 75°C

c. 150°C

d. 300°C

e. back to room temperature

 

30
A and 3.15 A.37 With the resolution of the 3.19 A reflection thanks to the
film preparation, this cluster is more precisely made of six peaks. In
contrast, the H-magadiite X-ray powder diffraction pattern is characterized
by an intense reflection line at 3.42 A. The exceptional intensity of this line
can be understood by considering the Na-magadiite patterns taken at

various temperatures (Figure 9 and Table 5).

TABLE 5. Selected X-ray powder diffraction
reflections for Na-magadiite at various temperatures and
for H-magadiite

 

 

 

th)
Na-25°C Na-75°C Na—150°C Na-300°C H—25°C
14.4 14.4
dool 15.6 14.4 12.4 12.1 12.0
hk0 4.11 4.11 4.11 4.15 4.11
3.63 3.58
3.56 3.52 3.49 *
3.45 3.44 3.43 3.44 3.42
3.31 3.27
3.15

 

 

 

 

 

 

 

It should to be noticed that our dam-values for Na-magadiite at
various temperatures are somewhat different from the ones found by
Lagaly et al..7 The values reported by Lagaly et al. are: up to 100°C
d001=15.6 A, from 100°C to 200°C d001=13.5 A, from 200°C to 400°C
d001=11.6 A, from 400°C to 500°C d001=11.5 A. Na-magadiite is converted
to quartz at 500°C and to tridymite at 700°C while H-magadiite is
converted to cristobalite at 900°C.7 As can be seen from the dooj-values in
Table 5, by heating Na-magadiite to 150°C or beyond, we eliminate the

water molecules, and so the (1001 value decreases. Heating the Na-magadiite:

 

31
Nam Si14 O23 (OH)1,9 . 7.0 H20 decreases the amount of water present in
between the layers as does exchanging the sodium cations by protons to
obtain the H-magadiite: 14 SiO2 . 2.5 H20. We can see on the Na-magadiite
heating experiences that the five reflection lines shift and get overlayed to
one another. Thus, we can assume that the intense 3.42 A line on the H-
magadiite pattern is the sum of several overlaying lines. An other
interesting point, though not understood, can be noted from the Na-
magadiite pattern at various temperatures. A 14.4 A line appears after
heating at 75°C. Perhaps this reflection is hidden by the 15.6 A dom line at
room temperature. In that case, this 14.4 A line could be assigned to a hk0

reflection.

The X-ray diffraction patterns of the Na/H-magadiite compounds
formed by acid titration of Na-magadiite are displayed in Figure 10. Table
6 provides the basal spacing for each derivative and the scattering domain
size or crystallinity along the c-axis. The latter quantity was determined
from the Scherrer equation (Appendix B).

TABLE 6. Basal spacings and crystallinity for Na/H-magadiite samples

prepared by acid titration of Na-magadiite. The samples correspond to
those identified in the titration curve shown in Figure 1

 

 

sample %H+ dfl crystallinity number of layers
exchange ( ) (A) orderly stacked

unwashed Na-magadiite 15.6 400 26

A‘ 0.0 15.6 400 26

B 12.5 15.6 400 26

C 23.6 15.6 400 26

D 61.1 15.6 and 13.4 340and-

E 94.4 12.4 50 4

F 94.7 12.4 50 4

0‘ 100.0 12.0 50 4

 

 

 

 

 

 

 

‘ A: Na-magfle, G: H—magadiite

 

32

 

 

 

 

 

 

 

 

 

15.6 A sampleB 15.6 A samplec
it 1 1 - oil . . - -
15.6 A 881119ch sampleE
g:
2
2 12.4 A
s
E .
2
2 [13.4 A
sampleF sampleG
12.4 A ‘2-0 A
0 io 20 30 0 10 A 3‘0 3'6
20 20

 

FIGURE 10. X-ray powder diffraction patterns (film) for mixed Na/H-
magadiite samples prepared by acid titration of Na-magadiite. The samples
correspond to those identified in the titration curve shown in Figure 1

 

33

A 15.6 A basal spacing and a high crystallinity along the stacking
direction are associated with Na-magadiite, while H-magadiite exhibits a
12.0 A basal spacing and a lower crystallinity, as judged by the width of
the door reflection. For the Na-magadiite structure, the basal spacing and
crystallinity. are identical from the unwashed Na-magadiite (Figure 11)
with a composition of Na4,3 Si14 O23 (OH)4,3 . 8.7 H20 to sample C with a
composition of Na1_3 Ho_4 Si14 O23 (OH)1_7 . 5.9 H20. The gallery height is
insensitive to variations in Na+ and H20 contents over this range. For the
H-magadiite structure, the basal spacing varies only slightly with the water
content, while the crystallinity remains the same: compare samples E (Nam
H1.7 Si14 028 (0H)1.8- 1-0 H20) to G (His Si14 028 (0H)1.8 - 0.7 H20).
Interestingly, as was already noticed by Rojo et al.,34 sample D (Nao_7 Hm
Sig 023 (OH)1,3 . 3.4 H20) is the only compound displaying two phases ;
the 15 .6 A peak represents the Na-magadiite structure, while the shoulder
at about 13.4 A represents H-magadiite. The crystallinity of the 15.6 A
phase is of the same order as the previous ones (340 A versus 400 A), but,
the crystallinity of the 13.3 A phase could not be determined because of
poor resolution. When two thirds of the sodium have been exchanged
(sample D), the Na-magadiite structure remains unchanged for most layers.
The Na-magadiite structure is stable over a wide range of sodium and
water content. This stability might be due to the sodium cations stabilized
by their hydration sphere. This large entity is able to prop apart the layers
even though some of the sites are already exchanged: the intercalated
sodium cation surrounded by its hydration sphere is robust. Furthermore,
as can be seen from the data for the unwashed Na-magadiite (Nam Si” 023
(OH)4,3 . 8.7 H20), the intercalated space can accommodate some extra

molecules like water and sodium hydroxide in excess of the ideal formula

 

34

AEEV E033 8:255. covkom mark oamfiawmfiéz @3335. .2 EDGE

mm
Om

 

 

 

 

 

O

Airsuoiur oAnejor

Gown

 

35
Na2,o Si“ 023 (OH)2,o . 8.0 H20. For the H-magadiite structure, the amount
of gallery water contributes to the basal spacing. That is, the basal spacing
found for H-magadiite (12.0 A) is not equal to the layer thickness (11.2
A)33 because of the presence of some water molecules. The presence of the
two phases for sample D with 61% H+ exchanged is a further evidence for
the assignment of sodium to proton exchange with retention of structure on
the titration plateau (.. pH=7). The two distinct phases present in the pattern
can be understood thanks to the magadiite morphology and the assumption
of a diffusion process for the proton exchange. Na-magadiite and H-
magadiite have a rosette-like structure (Figure 8a),2.4.8.18.20.22.29.38 It is
formed by the stacking of thin square plates (Figure 8a and b). This
morphology has also been seen for most of the elements of the hydrous
sodium silicate family.12-17»24 The layers on the edges of the rosette might
be completely converted into the acid form, while only some sodium
cations have already been exchanged in the internal layers. Even though the
layer charge is no longer exclusively balanced by sodium cations, the

gallery is still propped by the dimensions of the hydrated sodium cations.

The crystallinity, or coherent domain is associated with the sharpness
of the peaks and the number of reflection lines. Peak broadening and
disappearance of reflection lines are the consequences of stacking
disorder.l3»34 The Na-magadiite structure crystallinity is equal to 400 A,
corresponding to about (400/15.6) 26 layers orderly stacked. The H-
magadiite structure crystallinity equals 50 A, corresponding to about
(SO/12.0) 4 orderly stacked layers. This loss of crystallinity, as well as the
retention of the sodium structure up to some extent of proton exchange, has

also been noted for kanemitel3 and octosilicate.23 There is thus a

 

36
reorganization of the Na-magadiite structure, with retention of the crystal
morphology, upon proton exchange. This change in crystallinity is
indicated by the disappearance of most of the hkl reflections. This same
effect on the X-ray diffraction patterns could also be seen upon heating Na-
magadiite (Figure 9). Thus, the loss of crystallinity for H-magadiite is
associated with the loss of the water molecules, which play an important

role in Na-magadiite stability and structure .

Infrared spectroscopy

In order to obtain a better appreciation of the structure changes
occurring upon converting Na-magadiite to H-magadiite, we studied next
our samples (Na-magadiite, mixed Na/H-magadiite sample D and H-
magadiite) by infrared spectroscopy (Table 7). Figure 12 provides the
FTIR spectra for those samples.

TABLE 7. Magadiite infrared data

Na-
magadiite 443 579 706 811 949 1060 1205 1664 3470

sampleD 443 450 579 706 957 12001653 3458

magadiite 450 579 706 975 1200 3440

    

A:
B: double ring22~37

c: Si-O symmetric suetch22-37 and Si-O-H beud40
D: lattice vibration23

E: Si-O asymmetric stretchzzv37

F: 0-H bend12—33'37

G: O-H samba-2133:3590

 

 

 

relative absorbance

 

 

 

 

l I I l T I I
1800 1600 1400 1200 1000 800 600 400
Wavenumber (cm‘ 1)

 

relative absorbance

 

 

 

 

 

I l I I
4000 3800 3600 3400 3200 3000
Wavenumber (cm‘l)
FIGURE 12. Magadiite infrared spectra (KBr pellet, room temperature)
a. Na-magadiite c. H-magadiite
b. sample D as identified in the titration curve shown on Figure 1

38

The infrared data obtained for Na-magdiite are close to those
reported by Garces et a1.22 and by Scholzen et al..37 By comparing his data
with the 1225 cm-1 characteristic vibration of some zeolites, Garces et al.22
assigned bands at 1237, 1210 and 1175 cm-1 to the presence of Si04 units
forming five-member rings. He also assigned the 576 and 546 cm-1 bands
to type B structural blocks analogous to those found in high silica zeolites.
The 949 cm-1 band in the Na-magadiite lattice vibration region has also
been reported by Borbély et al.23 for octosilicate. The 975 cm-1 vibration is
known to be a Si-OH stretching band for isolated silanolsfl0 The O-H bend
vibration at 1664 cm-1, and the 1172 and 460 cm-1 bands of the Si-O stretch
and bend vibrations, respectively, have been noted for kanemite.12 The two
latter bands might be due to the same kind of SiO4 tetrahedra. The Na-
magadiite and H-magadiite absorption lines are similar, yet distinguishable.
Furthermore, sample D displays the features characteristic of the two
structures. H-magadiite has a smaller number of SiO4 tetrahedra bands (see
Si-O bend and Si-O stretch regions). However, the layer double ring
structure is identical. Also, there are changes in the O-H bend and stretch
regions. The lower number of SiO4 tetrahedra types in H-magadiite can be
due to the conversion of O3Si-O- Na+ into O3Si-O-H sites that were already
present in the Na-magadiite structure. The changes in the O-H
environments on passing from Na-magadiite to H-magadiite are due to the
loss of the hydration sphere around the sodium cation, to the creation of
new silanol groups and to the formation of siloxane bridges. Many bands
are present in the O-H stretch region of both Na-magadiite and H-
magadiite, but all the hydroxyl peaks are broadened and shifted to lower
wavenumbers because of the presence of water molecules that create

hydrogen bonds.

 

 

3590

3632

relative absorbance

 

 

 

4000 3800 3600 3400 3200 3000
Wavenumber (cm'l)

 

1 .
3595

relative absorbance

 

3720

 

 

 

4000 3800 3600 3400 3200 3000
Wavenumber (an-1)

FIGURE 13. Magadiite hydroxyl stretch Fl'IR bands under vacuum at
150°C
top: pristine pressed pellet (a: 20 mg ; b: 10 mg)
bottom: 1 mg of sample in a 100 mg KBr pressed pellet
a: Na-magadiite, b. H-magadiite

40

We attempted to eliminate physisorbed water molecules by running
some experiments under vacuum at 150°C (Figure 13 and Table 8).

TABLE 8. Magadiite hydroxyl stretch FI‘IR data under
vacuum at 150°C

 

Two preparative methods, which give complementary information,
have been used. As a KBr pellet, the Na-magadiite reacts and looses all its
water and silanol groups at 150°C. This reaction is reversible: when the
pellet is put back into atmosphere and set back into the instrument, the O-H
stretches appear again. The pristine Na-magadiite pellet gives more
information: upon loosing its extra physisorbed water molecules, four
peaks can be determined at 3744, 3719, 3595 and 3497 cm-1. The KBr
pellet gives more information in the case of the H-magadiite sample: three
peaks can be identified at 3632, 3590 and 3447 cm-1. However, when the
pristine H-magadiite powder is compressed into a pellet in absence of KBr,
the water cannot escape the sample. The peaks are then overlayed by the
broad water peak, but an extra peak appears at 3720 cm-1 and the 3632 and
3590 cm-1 peaks can be observed. The multiple O-H stretching bands in the
4000-3000 cm-1 region are indicative of O-H bonds in different
environments. Several O-H groups can be present in magadiite: physically

adsorbed water, which is removed at 150°C under vacuum, interlayer

 

41
water molecules participating to the structure and silanol groups. Silanol
groups can be on the internal or external surfaces, hydrogen bonded with
other silanol groups or with water molecules. By comparison of the
magadiite spectra taken before and after treatment at 150°C, we can assign
the 3291 cm-1 shoulder to external physisorbed water. Upon complete
departure of water, the Na-magadiite silanol groups are accessible and very
reactive toward substitution when the pellet is diluted by KBr. Possible

reactions are as follow:

150°C, vacuum

roomtemperature
Si-O-H+KBr ---------------- >Si-Br+KOH ---------------- >Si--OH+KBr
HO +HO

' 2 2
01’
150°C, vacuum room temperature
Si-O- H + KBr ---------------- > Si- 0 K+ + HBr -------------> Si-O-H + KBr
-H20 +H20

However, when the pristine Na-magadiite is pressed to a pellet in the
absence of KBr, only external water can desorb. In that case, the structure
does not change and the silanol groups can clearly be seen. For H-
magadiite, when the pellet is pure, the water molecules cannot desorb at all
and the spectrum displays a broad peak. However, when the sample is
prepared as a KBr pellet, water can escape and then the silanol groups can
be determined. There is thus a completely different behavior for Na-
magadiite and H-magadiite in the presence of KBr at 150°C. The
substitution reaction which occurs for Na-magadiite is a diffusion process
that can not happen in the small H-magadiite gallery. The 3744 cm-1 band
disappears upon proton exchange. This might be due to non hydrogen-

 

42

bonded internal silanol groups, due to the presence of gallery sodium
cations and large gallery height. After proton exchange, some hydrogen
bonds arise from neighboring silanols because the average H-H distance
between O-H groups decreases from ~4 A in Na-magadiite to ~25 A in H-
magadiite.34 Then, a new band at 3632 cm-1 appears which is due to silanols
from adjacent layers interacting with each other. The sharp 3744 cm-1 band
has also been observed on the surface of silica for isolated silanols.40 The
3719 cm-1 band might be due to isolated silanols on the external surfaces ;
this band is shifted to lower wavenumbers due to some interparticle
contact. This band is very weak in the pristine H-magadiite spectrum and
completely absent in the KBr pellet sample of H—magadiite. As for Na-
magadiite, reaction with KBr can occur because this silanol is outside the
gallery. The 3595 and 3590 cm-1 bands might be due to hydrogen-bonded
silanols while the 3497 and 3447 cm-1 bands might be due to interlayer
water molecules. It should be noted that Rojo et al.33 found different OH
stretching bands for H-magadiite at 150°C depending on whether the
starting Na-magadiite was natural or synthetic. For the natural mineral, the
bands were located at 3640, 3380, 3260 and 3180 cm-1 while for the
synthetic derivative, he found just two bands at 3620 and 3440 cm-1.
Conversion of his natural Na-magadiite to H-magadiite gave two types of
silanols: some (H3) interacting with the opposite layer (3380, 3280 and
3180 cmil) and the others (112,) not (3660-3640 and 3490 cm'l).33-34

298i nuclear magnetic resonance
To obtain more information on the Si-OH groups, we studied next
the layer structure of the magadiite by solid state 29Si nuclear magnetic

resonance (Figure 14 and Table 9).

 

 

 

[unliivaHHIHHIHIilrvnliilrlintlnii'vHIIIHTTTWT'
-75 ‘85 ‘95 -105 '115 DDm

FIGURE 14. Magadiite solid state 2S’Si nuclear magnetic resonance spectra
(delay time: 1200 seconds for Na-magadiite and 600 seconds for sample D
and H-magadiite, line broadening: 40, 12 scans)

a. Na-magadiite c. H-magadiite

b. sample D as identified in the titration curve shown on Figure 1

 

44

TABLE 9. Magadiite solid state 29Si nuclear magnetic resonance data

 

sample S1 chemical shfis (ppm vs. TNE) Q3/Q4 per Sim unit cell

 

 

 

 

 

 

 

 

<23 04 mic <23 o
Na—magadiite -§.5 -110.0 -111.4 -114.1 0.39 . 10.1
D -99.8 -110.6 -111.6 -ll4.4 0.32 3.4 10.6
H-rnagadiite -101.2 -110.9 -lll.7 -114.3 0.29 3.2 10.8

 

 

298i NMR allows for the differentiation of SiO4 tetrahedra environments
and relative abundance of the environments. The spectra of magadiite
exhibits Q3 and Q4 sites, according to Lippma and a1. nomenclature.“
Silicon atoms chemical shifts are governed by their bonding angles and
length.37 Q4 sites mean the silicon atoms are coordinated to four O-Si units.
They are represented as Si(OSi)4 units. The Q3 silicon atoms are
coordinated to three O-Si units, the last bond being made up of O-H or O'
Na+ groups. We can express a Q3 environment as (H-O)S_i(OSi)3 or (Na+O‘
)Si(OSi)3-

The spectra for Na-magadiite, sample D and H-magadiite are similar.
Upon proton exchange, the Q3 peak broadens with a full width at half
maximum in the range 1.6-3.8 ppm and shifts from -99.5 ppm to -101.2
ppm. This broadening and shifting of the peaks are a consequence of
hydrogen bondings between the layers.30 Indeed, a hydrogen-bond between
layers will make the two silicon tetrahedra which are sharing the proton
more Q4-1ike. For the three samples, the Q4 peak is divided into three
components at -110.9, -111.7 and -114.3 ppm, with relative intensities of
982100265 for H-magadiite. For the sample D, the chemical shifts are
-110.6, -111.6 and -114.4 ppm, with relative intensities of 89:100255. For

45

Na-magadiite, the shifts are at -110.0, -111.4 and -114.3 ppm, with relative
intensities of 70:100:52. The three H-magadiite Q4 peaks have been missed
by some authors4tl4t30 (Table 10) because the first two peaks are of the
same intensity. An excessive smoothing of the spectra will only allow the
determination of one peak positioned in between the two peaks of the same
intensity. As the magadiite structure is still unknown, the three Q4 peaks
have not yet been assigned. There are four possible Q4 sites: Q4(4Q4),
Q4(3Q4,1Q3), Q4(2Q4,2Q3) and Q4(1Q4,3Q3). The -110.9 ppm peak is
sensitive to the presence of O-Na+, because, after proton exchange, it shifts
and its intensity increases to equal the -111.7 ppm peak. The -111.7 and
-114.3 ppm peaks do not change upon proton exchange. We may postulate
that they are Q4 silicon tetrahedra that are preferentially connected to Q4
rather than Q3 silicon tetrahedra: Q4(4Q4) or Q4(3Q4,1Q3).

The Q3/Q4 ratio is equal to 0.39 for Na-magadiite. This ratio can be
translated, for a unit cell of 14 silicon atoms, into 10.1 Q4 sites and 3.9 Q3
sites. This ratio is in good agreement with the Na-magadiite formula found
by chemical analysis : Na1_9 Si14 O23 (OH)1,9 . 7.0 H20. This latter formula
is showing 1.9 Si-O-Na+ groups and 1.9 Si-O-H groups, for a total of 3.8
Q3 silicon sites. The Q3/Q4 ratio is equal to 0.32 for the sample D, and
equal to 0.29 for H-magadiite. This latter ratio can be translated, for a unit
cell of 14 silicon atoms, into 10.8 Q4 sites and 3.2 Q3 sites. The Q3/Q4 ratio
depends on the condensation of Q3 SiOH groups. If there was no layer
condensation, the Q3/Q4 ratio would be constant. In our case, it appears that
0.7 (18%) silanols condensed per unit cell upon proton exchange. As the
Q3/Q4 ratio decreases from Na-magadiite to H-magadiite and the intensity
of the Q4 peak at -110.9 ppm increases, we can assign the latter peak to Q4

sites formed by condensation of silanol groups. Condensation of silanols

46

groups also occurs during the proton exchange of kanemite.42 Indeed, as
kanemite is made of a single sheet of SiO4 tetrahedra, its 29Si NMR
spectrum only displays a Q3 peak. However, upon proton exchange, a Q4
peak appears due to some silanol condensation.42

Various Q3/Q4 ratio have been reported in the literature (Table 10).
Dailey4 noticed that Na-magadiite has unusually long relaxation times: 160
seconds for Q3 and 280 seconds for Q4 ; for H-magadiite, both Q3 and Q4
relaxation times are 95 seconds. To obtain quantitative analysis, the delay
time should be five time larger than the relaxation times for a 90° pulse.
Scholtzen et a1.37 state that the delay time should be equal or greater than
60 seconds. A ‘short relaxation time will enhance the Q3 signal relative to

the Q4 signal.

TABLE 10. Various magadiite 2981 NMR data

 

sample ref Si chemical shifts (ppm vs. TMS) Q3/Q4 delay

 

 

 

Na-magadiite 22 -99.1 -109.4 -110.6 -1 13.2 1 10
Na-magadiite 22 -99.7 -109.5 -111.2 -113.6 0.83 10
Na-magadiite 18 ~101.0 -112.2 -1 14.8 -117.5 0.48 20
Na-magadiite 37 -99.2 -110.1 -111.0 -113.5 0.43 7
Na—magadiite 37 -99.3 -110.0 -1 l 1.2 -113.9 0.42 7
Na-magadiite 4 -99.3 -110.1 -111.3 -113.9 0.36 ~1250

4

 

 

 

 

 

 

 

 

Na-rnagadiite 14 -102 -113 0.33

Na-ma adiitc 31 -99.0 -110.7 7
H-magadute 30 T98 7(53 0.37 4
H-magadiite 14 -105 ~111.5 0.3 4
H—magadiite 4 -1004 -110.8 -113.7 0.28 ~1250

 

Sample D (Nao_7 Hm Sig 023 (OH)1,3 . 3.4 H20) does not exhibit the Q3/Q4

ratios or chemical shift values of Na-magadiite or H-magadiite. The

 

47
observed values are intermediate between those of Na-magadiite and H-
magadiite. So, the chemical shifts and Q3/Q4 ratio depends on the amount of
protons in between the layers.

Generally speaking, we can say that the Na-magadiite and H-
magadiite exhibit the same layer framework based on NMR spectra. But, it
is somehow surprising that the Na-magadiite Q3 peak is narrower than H-
magadiite one, even though, in the Na-magadiite case, the Q3 sites are
created by two kinds of tetrahedral silicon units, namely, (Na+0‘)Si(OSi)3
and (HO)Si(OSi)3. The broader H-magadiite Q3 peak may be the result of
reduced interlayer order due to some silanol condensation and the

formation of hydrogen bonds.

23Na and 1H NMR experiments have also been carried out by other
workers.293334 Rojo et al.34 found two kinds of sodium in Na-magadiite at
+6.7 and -1.8 ppm vs. NaCl. He assigned the +6.7 ppm peak to NaOH
because it disappeared completely at the beginning of the titration. The -1.8
ppm peak, which decreases to zero upon complete titration, was assigned to
N a+~OSiE. On the basis of relaxation time studies, Lagaly29 found two kinds
of hydrogen in H-magadiite. One due to water molecules, the other one to
silanol protons. Rojo et al.34 found three different proton environments in
Na-magadiite and two in H-magadiite. Upon heating, Na-magadiite and H-
magadiite lose two and one proton environment, respectively. These
resonances were assigned to water molecules. For Na-magadiite, the two
kinds of water can be external physisorbed water molecules and interlayer

water molecules in the hydration sphere of Na+.

48

£635-: a .aeeemeesz .e
833 "Eur .Eofiofl noun—euovéoueag NZ £2
er2 E .552; coca:
So Ego was» 9.6398 u 833 councgouéoueag NZ Sagas: .2 MMDOE

 

 

 

 

 

 

 

 

 

 

 

Q:
o— w o v N c
gcfim 1 u u q 1 a q u c
c._ wd ed To Nd ed
p - n - o
u“ n W- . .
. - on m
. m
- 9: m . . S
. an 1
S
.. .02 fl . .a .
82w
4.8 PE t . cu
o.— ad ad v.0 Nd o.o
e m.
J v A
on m
A
01 t a .
8“ n <4 cm
a ...._
o2 m .. .
. m
p - n h n b SN
6 . . . . . . . CV

 

 

 

(Bloc) dis sumIOA 990.10sz

49
Nitrogen adsorption-desorption
We also studied the magadiite surfaces by gas adsorption-
desorption11 (Appendix C). The isotherms and t-plots are displayed on
Figure 15 and the data are gathered in Table 11.

TABLE 11. Magadiite N2 adsorption-desorption data

sample temperature surface surface
°C

Na-magadiite 150 21 21

 

Na-magadiite exhibits a total surface area of 21 mZ/g, with no micropores,
regardless of the outgassing temperature over the range 75-300°C. At
various outgassing temperatures, H-magadiite shows a higher total surface
area of 70 mZ/g, due to the presence of micropores. Dailey4 found a
surface area of 24 m2/g and 45 m2/g for Na-magadiite and H-magadiite,
respectively, outgassed at 120°C. Lagaly et al.8 found a surface area of 83
m2/g for H-magadiite which be assigned to interparticle pores. It is also
worth noticing that the surface area found for H-kanemite is of the same
magnitude: 50 mZ/g after outgassing at 700°C.42 According to the t-plot, it
is unlikely that the microporosity we observe for H-magadiite arises from
interparticle pores. Although the basal spacing decreases with increasing

outgassing temperatures (Table 5: from 14.4 A to 12.1 A), the Na-

 

50
magadiite surface area does not change. The increase in surface area upon
proton exchange (dom=12.0 A), which arises due to the presence of

micropores, indicates a different organization of the layers.

Structure discussion

The recent development of some analytical techniques such as FTIR
and solid state NMR have allowed some authors13v22 to propose some
structure for magadiite. The two models assumed the formation of layers

by superposition of three tetrahedra SiO4 sheets.

Schwieger et a1.18 made use of the single-sheet makatite structure to
derive structures for octosilicate, magadiite and kenyaite. The crystal
structure of makatite has been determined, in 1982, by Annehed et a1..16 It
is made of continuous sheets of highly folded Q3 tetrahedra, condensed to
form six membered rings (Figure 16). The study is based on basal spacing
values, found by X-ray powder diffraction, and on high resolution solid
state 29Si nuclear magnetic resonance spectroscopy. The derived structures
consist of two, three and five combined makatite-monolayers (Figure 16

and Table 12).

TABLE 12. Experimental and theoretical data for layered Na-silicate
structures according to Schwieger et al.13

 

 

 

 

experimental model
srhcate Simafi thl Q3/Q4 m0 -' 9%” Q3/Q’4
molar ratio ( ) ratio molar ratio ( ) ratio
4 9.05 0° 4 9.10 0°
8 11.0 1.12 8 14.1 1.00
14 15.8 0.48 12 19.1 0.50
20 19.7 0.25 20 29.2 0.25

 

 

 

 

 

 

51

 

FIGURE 16. Structure of makatite and the postulated stuctures for
octosilicate, magadiite and kenyaite (Schwieger and a1“)

left: top view of the makatite single-sheet surface16

right: edge-on view

52
Table 12 clearly indicates that experimental basal spacing data are smaller
than theoretical ones. The basal spacing derived from the model is the sum
of the makatite layer thickness (5.02 A), times the number of makatite
monolayers, plus the makatite gallery height (4.08 A):18

doc; = n*5.02 + 4.08
n=1,2,3,5 for makatite, octosilicate, magadiite and kenyaite, respectively.

The difference in basal spacings between experimental and theoretical
values was explained by a condensation of Si-OH units between the
makatite-monolayers. However, a discrepancy between experimental and

theoretical Si/Na ratio for magadiite was pointed out, but not explained.

The hypothetical magadiite structure developed by Garces et a1.22 is
based on X-ray powder diffraction data, infrared and NMR spectroscopies
and comparison with mordenite- and pentasil-group zeolites. The structure
is based on the experimentally derived value of Q3/Q4 ratio equal to unity
and on the presence of an infrared peak at 1225 cm-1 characteristic of
zeolites containing five-member rings. The Garces model consists of
'layers of sixemember rings of tetrahedra and blocks containing five-
member rings attached to both sides of the layers' (Figure 17). However,

this model does not take into account the non-equivalent Q4 seen in NMR.

The Schwieger and Garces models for Na-magadiite do not agree
with one another. This is mainly due to the different values found for the
Q3/Q4 ratios. The 20-second delay time taken by Schwieger et al.18 and the
10-second delay time taken by Garces et al.22 are too short for the spin

relaxation times of 160 seconds and 280 seconds for Q3 and Q4,

 

53

 

FIGURE 17. Magadiite structure postulated by Garces et a].22
left: edge-on view

right: top view

‘ possible arrangement for the level

54

respectively.4 The quantitative use of the NMR data was thus, most
probably, incorrect due to insufficient delay time between pulses.

Unfortunately, we are not able to propose a Na-magadiite structure
that will account for our NMR data. However, on the basis of our results
for the structure stabilizing effect of the hydration sphere (chemical
composition, DSC) around the sodium cation, we can postulate a well-
defined site for the sodium cation and its hydration sphere in the interlayer
space. Indeed, an imperfect layer stacking orientation along the c-axis due
to the disappearance of the stabilizing effect of the sodium hydration sphere
upon proton exchange will affect the scattering of the X-rays (XRD). Such
disorder has no consequence on the sheet lattice vibrations (FI'IR, 29Si
NMR) and the morphology of the product (SEM, TEM). This loss of
stabilization upon proton exchange will allow some stacking disorder that
will create some intragallery micropores and then increase the surface area
(gas adsorption). Thus, the disorder seen by some techniques is not due to a
change of the layer structure, but to a change of the interlayer stacking
arrangement. Furthermore, when the water molecules disappear, the layers
become close to one another. This allows some silanol condensation, giving
rise to the formation of siloxane bridges (NMR). Those bridges could
explain the disappearance of some of the silanol absorption bands (FTIR).

It is also noteworthy to notice that, ideally, the sodium to water ratio
is equal to 4 for the Na-magadiite: Na2,o Si“ 023 (OH)2,o . 8.0 H20. So, the
sodium cation can be located in the center of 4 water molecules. Moreover,
we can recall the similarities between the 185-190°C endothermic peaks
assigned to the loss of the hydration sphere on makatite15 and magadiite. In

the makatite structure,l6 sodium cations are either in the center of an

55
octahedra formed by water molecules, or in the center of a distorted
trigonal bipyramid formed by two water molecules and three oxygen
atoms, one coming from a layer, the two other from the opposite layer. So,
derived from that known makatite structure, we can postulate that the
sodium cations occupy positions within the layer that might be octahedral
sites created by 6 oxygen atoms: 4 from the water molecules, 1 as an
hydroxide from a silanol and the last one bearing the negative charge
(Figure 18). This importance of the hydration sphere around the sodium
cation might as well be relevant for the other members of the hydrous
sodium silicate minerals family, as H-kanemite was also found to be less

organized than its sodium counterpart.42

n/O\$/O\$/O\$/0\

'H
H o_ri H d m
H23 Naif-(1H H2): ' .pO‘H

H-or' 5 oil 1117' ta‘~o’H
H/ QH \H H/ (')- \H

1
Si Si Si Si
\ 0/ \ O/ \ O/ \ 0/

FIGURE 18. Hypothetical proposed sodium site in the magadiite

56
CONCLUSION
The acid titration of Na-magadiite is not a simple cation exchange
reaction. The replacement of Na+ by H+ is accompanied by a structural
change in the interlayer space and a condensation of the interlayer silanols
groups. Although there is retention of the layer framework structure, the

reaction is not rigorously topotactic.

Upon proton exchange, as well as upon heating Na-magadiite, the
amount of gallery water decreases and the layer stacking structure becomes
less organized. This stacking disorder is due to the changes in the interlayer
space. Thus, the water molecules might occupy specific sites in between the
layers and play a stabilizing role even after the replacement of two thirds
of the sodium cations. The Na-magadiite stability is promoted by these
water molecules that form the hydration spheres around the sodium

cations.

 

 

ACID CATALYZED TEOS POLYMERIZATION
IN THE PRESENCE OF ALKYLAMINES:
SYNTHESIS OF A NEW POROUS SILICA

INTRODUCTION

A new approach has been recently developed for the metal oxide
pillaring of certain types of lamellar solids.“ The conventional pillaring
synthesis45»46 takes advantage of ion exchange properties of the host. The
approach can be described as a two-stage reaction. In a first step, a
polycationic pillaring species is intercalated by ion exchange in presence of
the layered material. Secondly, the newly intercalated pillar is converted to
a nanoscopic metal oxide aggregate by calcination (Figure 19). In the new
approach, a condensation-polymerization reaction of an alkoxide M(OR),,
(eg., tetraethylorthosilicate Si(OCH2CH3)4 (TEOS)), is carried out in the
gallery of an acid form of the layered material. The gallery height of the
acidic layered material is expanded by an intercalation of alkylamine and

protonated alkylammonium cations ; the M(OR)x molecules can then access

57

58
the interlayer space and proton-catalyzed hydrolysis can occur to form a
metallo-organic polymer. The product is finally calcined to eliminate the
organic molecules and to stabilize the intercalating species as an oxide
(Figure 20). The final product displays extremely high surface areas.

In this work, we elucidate the mechanism of that new pillaring
reaction that yields such microporous materials. To achieve this goal, we
performed experiments in the same conditions as for the pillaring
reactions, but without the presence of the layered material. The chemical
effect of the layered compound was mimicked by the addition of acid. The
studied reaction in that part is thus an acid catalyzed TEOS polymerization

in presence of alkylamine.

BACKGROUND

Many references can be found in the literature concerning TEOS
polymerization/‘7'6o Most of them deal with the TEOS/water/solvent
ternary system in presence of acid or base catalyst.53 The most studied
solvent is ethanol. This system is studied in order to better understand the
sol-gel processing of metal alkoxides.47 Their polymerization enable the
synthesis of glasses, ceramics and compositesSO-Szt-i6 at low temperature.5559
They can be produced as homogeneous and pure51 powders, films, fibers
and monolithic products.55t57-59 The studies are done in order to obtain a
fundamental knowledge about the sol-gel processes: hydrolysis,
condensation-polymerization, gelation, drying and densification.47 The
investigators studied in particular the effects of temperature, TEOS
concentration, solvent, pH, catalyst, amount of water50i53i55v59 on the
reaction rate and mechanism49-51v55-59 in order to derive models. They

usually use nitrogen adsorption, small angle scattering, liquid 1H, 13C, 29Si

59

exchange

 

Z=H,R,M

FIGURE 20. Alkoxide polymerization pillaring

60

NMR, gas chromatography, Raman spectroscopy, density and viscosity
measurements48v50-60 to characterize sol-gel processes.

Our work was oriented toward the understanding of the silica
formed in between the layer of magadiite by hydrolysis of TEOS in the
presence of alkylamine, its condensation-polymerization and calcination at
450°C. To our knowledge, no TEOS polymerization has been studied in

presence of alkylamines.

THE SOL-GEL PROCESSING OF SILICATES47

The sol-gel process involves a solution or sol that undergoes a sol-gel
transition. At that point, the one-phase solution becomes a two-phase
system due to destabilization, precipitation or supersaturation. The
mechanism is not yet understood, but can be divided into five steps:
hydrolysis, condensation-polymerization, gelation, drying and
densification. The overall chemical reaction can be express in the case of
TEOS as: A

Si(OCH2CH3)4 + 2H2O ----- > SiO2 + 4 CH3CH2OH

Hydrolysis
In the hydrolysis reaction of metal alkoxides such as TEOS, water
molecules displace alkoxy (eg. ethoxy) groups to create silanols:
aSi(OCH2CH3) + H2O ----- > ESiOH + CH3CH2OH
The reaction is an electrophilic substitution in presence of acid, but a
nucleophilic substitution in a basic solution. The complete hydrolysis would
produce silicic acid Si(OH)4, but some condensation-polymerization occurs

simultaneously.

 

61
Condensation-polymerization
The way in which the species hydrolyze and polymerize determines

the propagation of the linking units:

zSi(OCH2CH3) + ESiOH ----- > ESiOSiE + CH3CH2OH

aSiOH + ESiOH ----- > ESiOSiE + H2O
Under acidic condition, the first hydrolysis causes the subsequent
hydrolysis of the same unit to be more difficult, so that linear polymer
growth will be favored. Furthermore, acid catalyzed reactions in solutions
containing a low water concentration produces linear polymers, while
solutions containing high water concentration produces cross-linked
polymers or branched clusters. Under basic conditions, subsequent
hydrolysis of the alkoxy groups of the starting species is easier, so that

branched cluster growth is promoted.

Gelation

Gelation is related to the diffusion of the oligomer species and the
viscosity of the solution. It can occur before the end of hydrolysis and
condensation-polymerization. Gelation is defined by the time-to-gel
parameter. Dilute solutions exhibit long time-to-gel, while concentrated
solutions have short time-to-gel. Increasing temperature also decreases
time-to-gel. The sol-gel transition is reached when the one-phase liquid
becomes a two-phase system composed of a solid and a liquid that can be

converted into a two-phase system of a solid and a gas.

Drying
The solvent phase can be removed by low-temperature evaporation,

to form xerogels or by hypercritical evacuation, to create aerogels.

62
Condensation-polymerization can still occur during the solvent removal.
Drying provokes shrinkage, but the gel is not said to be dry until it is

subject to heat treatment.

Densification
Densification occurs during the final heat treatment: trapped water
and organic molecules are eliminated, residual organics are oxidized and

silanols are condensed to form siloxane bridges and form pure silica.

RESULTS AND DISCUSSION
Reaction conditions

The reaction conditions for the hydrolysis of TEOS in the presence
of alkylamines are summarized in Table 13.

It is worth noticing that the water quantity is less than stoichiometric
for complete hydrolysis and polymerization of TEOS. Each TEOS
molecule requires two water molecules for complete polymerization to
form silica, whereas the actual amounts used were in the range 0.04-0.16
mol H2O/ mol TEOS. The samples were furthermore exposed to
atmosphere moisture during the air-drying process. Such low H20/TEOS
ratios were selected in order to be as close as possible to the conditions for
the silica intercalation into H-magadiite (H2 Si” 029 . 1.5 H20). The
amount of water used in our preparations is the smallest possible, as water

is coming from the concentrated hydrochloric acid (38% wt).

63

TABLE 13. Experimental conditions for the acid-catalyzed TEOS
polymerization in the presence of alkylamines

 

molar molar molar
sample amine calcina wt (g) wt (g) wt (g) ratio ratio ratio
tion amine HCl TEOS TEOS/ TEOS/ TEOS/

 

(4 hrs) conc. amine H+ H2O
HA50 hexyl none 3.12 0.260 12.0 1.9 22 6.4
CHASO 450°C
OASO octyl none 3.97 0.265 12.0 1.9 21 6.2
COA50 450°C
DA50 decyl none 4.82 0.270 12.0 1.9 21 6.1
CDA50 450°C

 

HA100 hexyl none 6.22 0.462 48.1 3.7 49 14
CHA100 450°C
OA100 octyl none 7.94 0.486 48.0 3.7 46 14
COA100 450°C
DA100 decyl none 9.66 0.498 48.0 3.7 45 13
CDA100 450°C
HA200 hexyl none 3.12 0.257 48.1 7.5 R 26
CHA200 450°C
OA200 octyl none 3.97 0.251 48.1 7.5 90 26
COA200 450°C
DA200 decyl none 4.84 0.243 48.1 7.5 93 27
CDA200 450°C

 

 

 

 

 

 

 

 

The polymerization occurs immediately when TEOS is completely
added to the amine/HCl/water mixture. The clear TEOS solution turned
white, except for the HA50 sample which remained clear. During the rapid
(~l day) air drying process on glass plates, the samples form homogeneous
films of thin white powder. After calcination at 450°C, even under flowing
air, the thin powder turns brown due to residual carbon. The CDA samples
were browner-than the COA samples, which were also browner than the
CHA samples. The yields of CHA100, COA100 and CDA100/TEOS exceed
to 30%. They have been approximated as:

Wt calcined product wt TEOS _ wt calcined product Wt TEOS
100 * M3i02 / MTEos ‘ 100 * 60.1 208.33

64
The reacting solutions containing the white particles were allowed to
air dry without further treatments. New phenomena occured during the
addition of ethanol as solvent to the hydrolyzed TEOS/amine solution.
After centrifugation of the solutions containing the gel, the liquid mother
was poured away and ethanol was added to the white gel. The gel, instead
of becoming dispersed, dissolved. Upon air drying the clear ethanol

solution, a thin white powder was formed.

Chemical compositions
The elemental analyses for TEOS-hydrolysis products, before and
after calcination, are listed in Table 14. The results derived from

thermogravimetric analysis (Figure 21) are displayed in Table 15.

TABLE 14. Chemical composition from elemental analysis for TEOS-
hydrolysis products formed in the presence of alkylamines

Ht Sid“!

0A100 28 6.5 . 25 37 37
0 .

1 14
42 56 1.0

COA100 0.86 1.3 .
2.1 1.4 41

 

** calculated to the method explained in the text

65
The C, H, N results from the chemical analyses are given in ppm (pg/l) for
C, H and N. By assuming a density of 1 g/l for the analyzed solutions, we

can convert those results to %wt.

The amine content has been calculated from the %wt N, given by the
analysis:

nA: moles of nitrogen in 100 g = moles of amines in 100 g

= %wt N _ m
MN _ 14
%wt amine = nA * MA MA = 101.19 g/mol for hexylamine
MA = 129.25 g/mol for octylamine
M A = 157.30 g/mol for decylamine

The ethoxy-ethanol content has been calculated using the %wt C left after
subtraction of the carbon present in the amine:

mate is defined as the number of moles of ethoxy groups and ethanol
molecules trapped per 100 g of compound.

‘7 tC ‘7 tC
nuto=<°§4VC 'C*nA)/2=(0v1v2 'C*nA)/2

 

 

c = 6 for hexylamine
c = 8 for octylamine
c = 10 for decylamine

%wt ethoxy = nEto * M1510 = nEtO * 45

The water content has been calculated using the %wt H left after deduction

of the hydrogen present in the amine and the ethoxy-ethanol groups:

66

mug is defined as the number of moles of water created by silanols
condensation and by desorption of water per 100 g of compound.

%wt H
nH20=( MH

 

tH
-h*nA-5*neo)/2=(%‘I

 

-h*nA-5*nato)/2
h = 15 for hexylamine

h = 19 for octylamine

h = 23 for decylamine
%wt H20 = nH20 * MH20 = nH20 *18

The silicon content in 100 g of compound has been calculated, assuming
that after the C, H, N analysis, the remaining product is silica Si02.

%wt Si02 = 100 - %wt amine - %wt ethoxy - %wt H20

If 118102 and 11$ represent the number of silica molecules and silicon atoms,
respectively, in 100 g of compound:

%wt SiOz = n5i02 * MSi02 = n5i02 * 60.1
%wt SiOz
nSi = nSiO2 = 601
%wt Si = nSi * MSi = nSiOz * 28.1

Finally, the total oxygen content in 100 g of compound is determined to
complete the elements weight % to 100:

%wtO =100-%wtC-%WtH-%th-%thi

67

 

O

UJ-hUIQQ
COO

% weight remaining
0

 

'—‘ N
O O
I

 

 

I l l J l l I

0 100 200 300 400 500 600 700 800
temperature °C

 

0

FIGURE 21. Thermogravimetric analysis for air-dried TEOS-hydrolysis
products formed in the presence of alkylamines
a. HA100, b. OA100, c. DA100

The %wt Si02 determined from thermogravimetric analysis has been
calculated assuming the remaining mass left after heating at 800°C is pure
silica (Table 15).

TABLE 15. Chemical composition from thermogravimetric
analysis for air-dried TEOS-hydrolysis products formed in
the presence of alkylamines

 

sample first from... second from... %wt
%wt loss to...°C %wt loss to...°C SiO2
HA100 12 30 to 250 16 250 to 800 72
OA100 33 30 to 300 12 300 to 800 55
DA100 62 30 to 225 19 225 to 800 19

 

 

 

 

 

 

 

68

The thermogravimetric curves show a two-step weight loss. The first
increment increases with the amine length and can be assigned to some
amine loss. The second increment is almost constant and might be due to
the elimination of some bonded amines and to the loss of water from
silanols condensation. The calcination temperature has been chosen
according to those thermogravimetric curves: there is no more weight loss

above 450°C.

The chemical compositions found by chemical analysis calculations

and thermogravimetric analysis agree in general terms (Table 16).

TABLE 16. Comparison of chemical composition from
elemental analysis and thermogravimetric analysis for air-
dried TEOS-hydrolysis products formed in the presence of

 

 

 

 

alkylamines ~
sample chemical analyses TGA
%wt volatiles %wt $02 %wt volatiles %wt 8102
HA100 28 72 28 72
OA100 47 53 45 55
DA100 71 29 81 19

 

 

 

 

 

The amine content in the samples increases with the length of the alkyl
chain due to lower vapor pressures. Surprisingly, there are almost no
ethoxy groups left after the polymerization. The polymerization went all
the way through to pure silica. Presumably, water from the atmosphere
during air-drying contributes to the hydrolysis process. Water is present
before and after calcination. After calcination, water is present as

physisorbed water molecules in the pores. The presence of 6-9 wt% water

69
for HA100 and OA100 before calcination is more questionable. Indeed, as
was noticed earlier, there are not enough water molecules in the initial
reaction mixture to entirely polymerize TEOS. We thus should expect the
polymerization to continue until no water remains... The water present in I
the air-dried gel might be due to physisorbed external water which adsorbs

from the atmosphere after drying and densification.

Infrared spectroscopy

The presence of amine and some water, as well as the absence of
ethoxy groups, can also be verified from the infrared spectra (Figures 22,
23 and 24). The data are gathered in Table 17.

The bands in the 3000-2500 cm-1 region are due to symmetric and
antisymmetric stretching vibrations of CH2 and CH3 groups.61 In the
octylamine spectrum (Figure 22b), the bands at 3370 and 1600 cm-1 have
been assigned to N-H stretching and NH2 bending, respectively.61 The
bands at 1469 and 955 cm-1 are due to C-C stretching and C-C bending,

respectively.

 

0

§

'8

O

U)

.D

63

0)

>

'5

E

2 a.
b.

 

I I 1 I I I I I 1 I I I l I

1900 1700 1500 1300 1100 900 700 500
Wavenumber (cm'l)

 

 

 

 

relative absorbance

 

 

 

I l I I l I I I I l I I I l

 

4000 3800 3600 3400 3200 3000 2800 2600
Wavenumber (cm'l)

FIGURE 22. Infrared spectra of the starting materials (liquid film between
KBr plates)
a. TEOS, b. alkylamine: octylamine

 

relative absorbance

 

 

 

l I l J I I I I I I I I I

1900 1700 1500 1300 1100 900 700 500
Wavenumber (cm‘l)

 

 

 

relative absorbance

 

I I I l I I I I I I I I I I

 

 

4000 3800 3600 3400 3200 3000 2800 2600
Wavenumber (cm‘l)

FIGURE 23. Infrared spectra of the air-dried TEOS-hydrolysis products
formed in the presence of alkylamines (KBr pellet)
a. HA100, b. OA100, c. DA100

 

relative absorbance

 

 

 

I I I I I I l I I l I I I

1900 1700 1500 1300 1100 900 700 500
Wavenumber (cm‘l)

 

 

relative absorbance

 

 

 

l I I I I I I I I I I l I I

 

4000 3800 3600 3400 3200 3000 2800 2600
Wavenumber (cm'l)

FIGURE 24. Infrared spectra of the calcined (450°C) TEOS-hydrolysis
products formed in the presence of alkylamines (KBr pellet)
a. CHA100, b. COA100, c. CDA100

73

TABLE 17. Infrared data of the TEOS-hydrolysis reactants and
products

 

sample wavenumber (cm'l)

 

TEOS 2977 28 2 1485 1394 1296 1171 1084 966
2930 1443 1366 1107 796

 

octylamine 3370 2W 2872 1600 14K) 1382 955
2928 2857 824

 

Davisil-62 3440 1631 l 105 975
silica 3275 803

 

HA100 3428 2959 2872 1631 1469 1394 1074 95?
2928 2857 1382 796

 

 

0A100 3407 2959 2872 1631 1469 1394 1222 1152 1070 959
2928 2857 1600 1382 1212 796
1521 724

 

 

DA100 3333 2959 2872 1651 1489 1394 1212 1163 1070 818
2928 2857 1631 1469 1382 1152 796
1576 1435 1346 759

 

CHA100 3434 1631 1080 53
COA100 3275 803
CDA100 567

 

 

 

 

 

 

 

Before calcination of the TEOS-hydrolysis products (Figure 23), the
different compounds exhibit the alkylamines' characteristic C-H vibrations
at 2959, 2928, 2872 and 2857 cm], and other alkyl groups modes at 1469,
1382, 1306 and 724 cm:1 bands (Figure 22). As was noticed during the

analysis of the chemical compositions, the amount of amine incorporated

74
into the TEOS-hydrolysis products after air-drying increases from
hexylamine to decylamine. All three air~dried amine-containing hydrolysis
products exhibit the same four bands in the OH vibration region with the
same relative intensities ; those bands and their relative intensities are the
same as the ones found for octylamine. The three characteristic peaks of
the ethoxy groups (Figure 22) at 2977, 2930 and 2892 cm-1 cannot be seen.
This is a further evidence for complete hydrolysis. HA100 and OA100
spectra display a broad hydroxyl and water peak centered at about 3420
cm-1, but the DA100 spectrum shows a small peak at 3333 cm-1 that might
be assigned to N- H vibration even though the octylamine spectrum only
displays a peak at 3370 cm 1. Again, the chemical composition values are
correlated with the infrared data: only HA100 and OA100 exhibit some
water molecules. In the region below 1600 cm'l, a greater number of peaks
appears in the DA100 spectrum than in the OA100 spectrum which also
exhibits more peaks than the HA100 spectrum. This is the region expected
for both Si-O and Si-N vibrations. Thus, the peaks in the DA100 spectrum
which cannot be assigned to alkylamine, TEOS and which are not present
in the HA100 spectrum might be due, not only to some new kinds of Si-O
bonds, but also to Si-N vibrations. Indeed, silicon-nitrogen compounds
have been known for a long tirne.52-63 An easy way to synthesize them is to
react silyl chloride with ammonia. Some of their infrared bands were
assigned twenty five years ago.54.65 For the cyclic (H3SiN-SiH2)3, Wells and
Schaeffer64 assigned the bands at 945 (vs, sh) cm-1 to Si-N-Si deformation
and the bands at 992 (vs, sh) and 1018 (s, sh) cm-1 to Si-N stretching.
Biirger and Sawodny65 assigned the Si-N stretching vibrations to the bands
at 721 (vs) cm'l, 741 (vs) and 686 (vs) cm-l, 731 (vs) and 626 (vs) cm-l,

710 (vs) cm-1 for the following dimethylamino-chlorosilanes: C13 Si

75

[N(CH3)2L C12 Si [N(CH3)2]2 C11 Si [N(CH3)2]3 and Si [N(CH3)2]4,
respectively. For dimethylamino—methylsilanes, Si-N stretching vibrations
were assigned to bands at 699 (vs) and 582 (vw) cm-1 for (CH3)2 Si
[N(CH3)2]2 and to 679 (s) cm-1 for (CH3) Si [N(CH3)2]3. For diethylamino-
methylsilanes, Si-N stretching vibrations were assigned to bands at 599
(vw) cm-l, 690 (s) and 600 (vw) cm'l, 675 (s) and 625 (vw) cm-1 for
(CH3)3 Si [N(C2H5)2]. (CH3)2 Si [N(C2H5)2]2 and (CH3) (C1) Si [N(C2H5)2]2.
respectively. For all these compounds, N-Si-N deformation bands were
found to be situated below 400 cm-1.

After calcination of the TEOS-hydrolysis products (Figure 24), the
infrared spectra for the three compounds look identical to the spectrum for
silica (Figure 25). The amines are almost entirely eliminated, as evidenced
by the absence of bands in the C-H vibration region. Water is present in
large amounts in the calcined TEOS-hydrolysis products: there is a broad
O-H stretching band centered at 3434 cm-1 with a shoulder at 3275 cm-1
and an O-H bending vibration at 1631 cm'l. As will be determined by the
porosity study, the water molecules are physisorbed in the pores. The Si-O
vibrations are assigned to bands at 1080, 953, 803, 567 and 460 cm~l. The
1080 cm-1 peak is the most intense : it is also accompanied by a large

shoulder toward higher wavenumbers.

The infrared data correlate well the chemical compositions calculated
from chemical analysis and thermogravimetric data. After hydrolysis,
there are no remaining ethoxy groups or ethanol molecules trapped in the
air-dried product ; the amine content increases with the amine chain length.
The alkylamine is completely oxidized after calcination. Only in HA100

and OA100 is some water present. After calcination, water molecules are

 

relative absorbance

 

I

 

I I I I I I I I I

 

1900

1500 1300 1100 900 700

Wavenumber (cm-1)

1700

 

relative absorbance

 

 

I I I I I I I I I I I I I

 

4000 3800 3600 3400 3200 3000 2800 2600

Wavenumber (cm'l)

FIGURE 25. Infrared spectra of silica gel Davisil-62 (KBr pellet)

77

present in the three compounds. Infrared analysis provides'two new pieces
of information: after calcination, the three products are similar and are

silica-like.

2981 nuclear magnetic resonance

The differences between the air-dried products and the similarities of
the calcined compounds can also be verified from solid state 298i nuclear
magnetic resonance. The spectra are displayed on Figures 26 and 27 and

the data are collected in Table 18.

TABLE 18. 29Si NMR data of the dried and calcined
products (450°C) TEOS-hydrolysis products formed in the
presence of alkylamines

sample

OA100 - 99.3 ~109.9
- -11

COA100 Z1021 . :1093
1

-101. -1

 

Pouxviel et al.51 has determined the liquid 2981 NMR chemical shifts
for various Qn types of silicon tetrahedra.41 The Qn peak position depends
on the degree of condensation of the silicon tetrahedra. Q0: -72 to -82 ppm;
Q1: -82 to -89 ppm ; Q2: -90 to ~96 ppm and Q3: -100 to -104 ppm.

The air-dried TEOS-hydrolysis products before calcination (Figure
26), exhibit 29Si NMR spectra containing two separated peaks. The Q3 and

 

 

IlllllIIIIIIIIIIIIIITIIIIIIIIllllll
—80 -90 —100 -120 ppm

FIGURE 26. 29’Si NMR spectra of the air-dried TEOS-hydrolysis products
formed in the presence of alkylamines
(delay time: 600 seconds, line broadening: 140, 12 scans)

a. HA100, b. OA100, c. DA100

79

 

 

TIT—IIITIIIITFITIIIHhIIIIIIIIIIIIT—l

~80 -90 -100 -120 ppm

FIGURE 27. 29Si NMR spectra of the calcined (450°C) TEOS-hydrolysis
products formed in the presence of alkylamines
(delay time: 600 seconds, line broadening: 140, 12 scans)

a. CHA100, b. COA100, c. CDA100

80

Q4 peak assignments have been done according to their chemical shifts. The
presence of Q3 peaks is correlated with the silanols seen on the infrared
spectra. The first resonance is situated around -100 ppm and is assigned to
silicon tetrahedra in a Q3 environment: (SiO)3S_iO-H. The second line at
about -110 ppm can be assigned to silicon tetrahedra in a Q4 environment:
Si(OSi)4. The approximate Q3/Q4 ratio was obtained by deconvolution of
the peaks ; the numerical value should be handled with care due to the poor
resolution of the peaks. Peak height ratios can be qualitatively compared
(Table 18). The intensity of the Q3 peak increases from sample HA100 to
DA100. Thus, the amine plays a role in determining the arrangement of the
hydrolyzed silicon centers.

The calcined (450°C) TEOS-hydrolysis products (Figure 27), exhibit
similar 29Si NMR spectra. The Q4 peak is shifted slightly toward lower
values at about -109 ppm, while the Q3 forms a shoulder on the Q4 peak at
about -102 ppm. The change in the 2S’Si NMR chemical shifts on calcination
suggests the formation of siloxane bridges by silanols condensation that

creates Q4 sites from Q3 silicons:

(SiO)3SiO-H + (SiO)3SiO-H ----- > (SiO)3Si-O-Si(OSi)3 4» H2O
Q3 Q3 Q4 Q4

The cross linkage of units and the increasing number of Q4 sites change the
overall environment of every Q3 which is surrounded by more Q4 than
before calcination. They are still Q3 sites, but with more Q4 near-neighbor
environments ; the shielding of Q3 sites by Q4 sites increases and therefore

the Q3 resonance is shifted upfield.

81

13C and 1H solid state nuclear magnetic resonance analysis would
give us more information about the presence of ethoxy groups or ethanol
molecules and the silanols.

The 29Si NMR analysis provides proof that the amine affects the
hydrolysis and is probably acting as a template in deciding Q3/Q4 ratios and
overall structure. Increasing the length of the alkyl chain creates more Q3
sites in the dried products. However, after calcination, the spectra of the
three compounds look similar, regardless of the amine used in the

hydrolysis step.

Nitrogen adsorption-desorption

As the hydrolysis of TEOS inside layered materials creates highly
microporous products, 'we decided to study the products of our acid-
catalyzed TEOS polymerization in the presence of alkylamine by nitrogen
adsorption. This measurement allowed us to determine adsorption-
desorption isotherms and t-plots to calculate the total surface area and the
porosity (Appendix C). Furthermore, Strawbridge et a1.48 showed that gels
made by acid-catalyzed TEOS polymerization exhibit high surface areas.
For example, a gel synthesized with a TEOS to H2O molar ratio of 0.5 in
presence of ethanol and concentrated HCl has a microporous volume of
0.36 cm3/g. He assigned this porosity to interstices caused by particles
packing during the gelation. The adsorption-desorption isotherms and the t-
plots are displayed in Figures 28, 29 and 30 according to the TEOS to
amine molar ratio used in the hydrolysis step. The numerical data are

gathered in Table 19.

82

8950 .o dom<OU .a 68.420 .a
83; ”Ewe .Eofiomm 829823-888»? «Z “co—
er2 8 8:8? 89:. :5 888 83 wcmmmeweso ” mg. Co
28 88E 05.8 8 mom; 3:5 8 a 8888 8583? Co 8888 2: E 888
8888 8308.3-mOME. 0.63 8528 05 .8 838 8:883 NZ .3 550E

S w e234 N o

 

 

 

 

 

 

 

 

 

 

 

 

. q i . u u . 4 O . . . cam
. . o H m o o o v6 ad ad
.. t on .a . a . _ . . . . c
. m .. 8N
. 1 9: fiM .
. I . 8 .. 1 8?
l L om— we 1 u
\. o
u L I. u n u _ u . u . t 80
hm... . 1 8m m o a w o so 4 o 8 8o
- ”w 4 ml. .n h
- . o3 m . 8m
0‘ 9 '\b.
1 . S
r 41 1 8m fl 31 r 8v
.0... . w; _ . . . coo
. / p u u n u n
t e - com Met 3 no 8 so 8 8
v .0 . q . . . . u . 1 O
.11 . as. . .
. . com
h o .
1 . omv . Se
. . . . . t _ . 8w . . . . _ . r . 80

 

 

 

 

 

(8/33) as oumIOA poqrospe

83

8:50 .o .8350 .p .oo_<:0 .a
83-. 3an .8285 8ch898888..“ «2 ”co—
er2 8 5:8? 855 So 8an 83 w5mmmwe8 u hm Co
58 :85 058 8 momfi. 3:5 8 a 3808 8583“ Co 8888 05 5 888
32608 mmmboézsmomfi. 0.592 8528 05 .8 8?:5 86883 NZ .3 ”.5505
2 m e 5: a e N o

 

 

a . . . a . a _ 1 O O
. OJ ad 06 am vd Nd Cd
I 1 CW .fl 4 . . a . . . q u I O
. . con
i 1 OS f t. 8».

8

i .. om—

 

.
.

o
o 8
w

loom

 

i own

i
'8

18m

 

 

(8/30) (us oumIOA poqrospe
§ .
3 (us oumIOA poqrospe

.. omm

 

 

i 18».

t owv

 

 

 

 

 

 

 

em<QU .o .ow<OU .9 .On<:u a
SE; ”Em: .Euofiofl coufiomuwéoueawm «2 ”a2
Uoofl 8 8259» “one: So 8an 83 wfimmawSO " a; mo
2:: SEE 05:8 8 mom; REE 5 “a @8829 85:86:“ .3 8535 05 E @258
32605 mam—EuEéOmF AUocnvv 350:8 05 mo 32:0 5:983 «2 6m @533.»

a 093; N o

 

84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. . u ‘ . ‘ u ‘ O awn—\nm
. o; wd 0.0 v.0 Nd 0.0
on .a . . . q . . . . o
. 1 SN
J Sfi v ‘ l‘l‘llll\\”
.. B ‘\ 1 §
1 Chg m. w . 8
m . m p _ - . . _ h n O . AOXvO m-
o._ m6 ed v.0 N. o.
% 1 SN D. ‘ q . . . . . . ‘ o W.
A .5 p
. m. . A
h 1 o2 m 1 8m 01
o 9 . m
.~ A S \l‘l‘l‘l‘l‘l‘l‘l‘l‘K a
O 1 8m fl 1 §S
. . . \W 4 nu
IIII p L . . - . . . _ . 08M“
1.... - own a 3 ad ad Yo 3 ed fl
.0 4 u q q 4 q . d 1 o a
2.1. - o8 .
8m
1 owv 8v
. . . . . . . . COW . . . . . . . F . 80

 

 

85

TABLE 19. N2 adsorption data of the calcined (450°C) TEOS-hydrolysis
products formed in the presence of alkylamines

COA200 1110 1110 970 0.39

COA100 1270 1270 1100 0.45

COA50 1220 1160 1030 0.41
1140 11 1230 920

area, : area, #3

     

volume
* Sumis derived from V” by thet- -plot method; Sp(2)=S.oml- Sm”.
** the C value is an energetic constant depending on the chemical behavior of the surface.

The surfaces of the silica products exhibit different chemical
behavior for each amine, regardless of the initial TEOS to amine ratios.
Indeed, the C value found by the BET method is decreasing from
hexylamine- to decylamine-derived products: C~160, 110, 70 for
hexylamine-, octylamine- and decylamine-derived products. This is further
evidence for the role of the amine during the TEOS-hydrolysis process.

Octylamine- and decylamine-derived products exhibit, for the three
TEOS to amine molar ratio, about the same total surface area at ~1200
mZ/g. The total surface areas calculated from the BET equation43 and the t-
plot method44 agree relatively well. The microporous surface areas Sum
were determined by the classic de Boer et al.44 t-plot method. We
compared Spa) with 8,1(2), the microporous surface area obtained by
subtraction of the non-microporous surface area Snow. from the total

surface area Stotal- With that method, we found a microporous surface area

 

86

of ~1000 m2/g for the octylamine- and decylamine-derived products at the
three different TEOS to amine molar ratios. The disagreement between
Su(l) and Spa) may be due to the fact that the micropores are so large (XRD
patterns on Figures 31 to 33 display 29 to 37 A lines) that they do not
match typical micropores (pore < 20 A) models, but are too small to be
treated by mesopores (20 < pore < 500 A) theory and so do not exhibit
hysteresis. A better microporous surface area might be obtained from
immersion microcalorimetry. Indeed, the Harkins-Jura method“ would
allow for a precise determination of the external surface area which would
be subtracted from the total surface area. However, the microporous
volume is always larger for the decylamine derived products than for the
octylamine ones. Since the surface areas are identical, but the pore volumes
are different, the pores sizes are different. The decylamine-derived
products exhibit larger micropore sizes than those of octylamine, because
their microporous volume is larger. According to Strawbridge et al.48,
when the two lines derived from the two slopes on the t-plot curve intersect
at high t-values, the pore sizes are larger than when the intersection occurs
at low t-values. Adsorption of molecules with known kinetic diameters
would give a more precise idea of the pore sizes. Thus, regardless of the
TEOS to amine molar ratio, octylamine- and decylamine-derived products
display high surface areas, due mainly to microporosity. The microporous
surface areas are identical for both amine derived products, but in the
decylamine case, the pores size is larger because the microporous volume is
larger.

The hexylamine-derived products exhibit higher surface areas when
the TEOS to amine molar ratio decreases. For the lowest TEOS to amine

molar ratio, the surface areas are about equal to the octylamine and

87

decylamine derived products ones, but the microporous volume is smaller.
Again, for an identical surface area, the pore volume and the pore size
depend on the amine length: the longer the amine, the larger the pore size.
Thus, the size of the alkylamine chain in the TEOS-hydrolysis product
serves as a pore size template. The surface area values depend on the
amount of hexylamine present: the more hexylamine, the higher surface
area. This dependence of surface area on the initial amount of hexylamine
is related to other differences on the hexylamine-derived products. Indeed,
as the hexylamine vapor pressure is higher than the vapor pressure for
octylamine and decylamine, it tends to evaporate more readily during the
air-drying process. As we saw before, the amine directs the pore structure
of the final silica product. So, if the hexylamine leaves the system before
hydrolysis and polymerization are completed, it will be less effective in
influencing the size of the final micropores. When the initial amount of
hexylamine is increased, it takes more time for the amine to evaporate
upon air-drying. Therefore, it can not play a structural role as do

octylamine and decylamine.

X-ray powder diffraction

The X-ray diffraction patterns (Figures 31, 32 and 33) are somehow
surprising. The calcined samples exhibit one broad peak or no peak. The
air-dried sample spectra display one broad peak and/or one sharp peak.
The data are gathered in the Table 20.

88

oom<QU Ea 8N<Q .o .8950 EB oom<O .p .oom<:U v5 oom<m .m
82:55 @0528 ”EM: £2605 BEER ”ao—
m& me one 3.08 05:8 9 mOmF RES 5
“a Bangui 35:83:? we 8:033 05 5 3.58 82605 Embonanmonfi. Goonvv
8528 EH BEES 05 mo mEozaa €33 cognate eonaoa 52-x 4m mmDOE

 

 

 

e m_wss_m_° e_¢_hsw.w_c
.a
d . 3&1
<22
m m
. e .n
a m. m
m. <5. m.
m m
, m.
<8: n .m
.0
.0
<3“: <32
. _ _ _ _ q _ _ q _ _ _ _ _ _ _ _ _

 

 

 

 

 

89

oo~<QU 0:0 00:5 .0 .8350 0:0 8:5 .0 .8220 0:0 83.: .0
303003 005300 ”Em: £00005 00.50-00 £2
Em 00 0:8 0208 05:8 00 wow... 325 :0
“a 02320 00580150 00 0000005 06 E 00:08 2000000 mmmboeumnmomi. 609va
005200 000 0010-0? 05 00 2:2qu €33 003000006 80300 $0-0m .Nm mmDOE

 

 

 

 

 

 

@N
c—_m_w¢_mw_w_c 2_w_wpw_fi_$
<wd—
.0
.0 .
m .n
in
m < <3”: m
e n.
m. .A.
a
m. / m.
m an
<3. m. <2: m.
.n
.0
4mm». <—.Nm
_________ _____.__

 

 

 

 

ow<Q0 0:0 035 .0 690.00 0:0 on<O .0 .0350 0:0 09‘: .0
£00005 0050000 "Em: £00005 0000-00 £0—
0.“ 00 0000 000:. 0580 00 mom; 000:: :0
00 0000505 005800030 .00 0000005 05 E 00:08 £00005 mum—000300."? 60¢va
0050—00 000 0000-000 05 00 0:00:05 038 000000020 00305 00-x .mm MMDOE

 

 

 

 

 

 

 

2 0 0004 0 o S m 000.. 0 a
_ _ _ _ _ _ b _ . _ h _ _ _ b _ _
. .a
$0..
0 .0 J m.
9 n m.
m. .0 m
9. m.
00.. m m
.0 m. <00: m.
.M
.0
<37.
.o
<2: <0 00
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -7!

 

91

TABLE 20. X-ray diffraction data of the air-
dried and calcined products

31 : COA200

34 - 25.4 COA100
37 - 32.1 CDA100

34 - COA50
44 -

 

The sharp peak at about 19.8, 25.4 and 32.3 A for the air-dried
hexylamine, octylamine and decylamine derived products, respectively, are

due to ammonium chloride salt formation (Table 21).

TABLE 21. Amine and ammonium
chloride salt dimensions

 

amine len ammonium chloride

 

 

 

 

 

( ) basal spacing (A)
hexylamine 9.2 19.8 (sharp)
octylamine 11.7 25.5 (sharp)
decylamine 14.2 29.3 (sharp)

 

The presence of the ammonium salt depends on the length of the amine and
so on its vapor pressure. The decylammonium salt is more often present
and in larger quantities than the octylamine and hexylamine ones. The
quantity of amine remaining in the air-dried products might also be related

in some extent to the atmospheric conditions and the humidity used to air-

92
dry as some ammonium bicarbonate salt can also be formed. In all cases,

the ammonium salts are completely removed after calcination.

Transmission electron microscopy

The presence of the broad diffraction peaks at high d-spacings
suggests some layered organization of the polymer that remains stable even
after calcination. Strawbridge et al.48 TEM pictures of a gel synthesized
with a TEOS to H20 molar ratio of 0.5 in presence of ethanol and
concentrated HCl has a layer structure but with no fine structure. It
correlated with the fact that under low water content, gelation results from
the entanglement of linear chains created during the condensation-
polymerization reaction.67 We have analyzed our sample by transmission
electron microscopy (Figure 34) in an attempt to detect some repeating
units of dimension equivalent to those found by X-ray diffraction studies.
We seek evidence for layers, ribbons or spheric particles.

Some layered structure is evidenced on the edges of the particles.
Unfortunately, we can not really draw more conclusions from those
pictures. The material seems to have some structure due to the presence of
atomic resolution, but no general feature can be seen. Shades are apparent
on some pictures, maybe due to interconnecting ribbons. But, up to now,

no conclusion can be made: a more intense study should be completed.

 

 

 

FIGURE 34a. CDA50 TEM picture (3x190,000: 1cm:

175

 

A)

94

 

 

 

 

1cm=175A)

FIGURE 34b. CDA50 TEM picture (3x190,000

95

 

 

 

 

 

FIGURE 34c. CDA50 TEM picture (3x190,000: 1cm=175A)

96

 

FIGURE 34d. CDA50 TEM picture (3x190,000: lcm=175A)

97

CONCLUSION

The acid catalyzed hydrolysis-polymerization of tetraethylortho-
silicate in presence of alkylamines generates air-dried precursors which,
when calcined are converted to highly microporous silica. The high total
surface area and layered structure has been sometimes noticed for acid or
base catalyzed polymerization of TEOS in the presence of a solvent.
However, the study of our samples demonstrated the influence of
alkylamines on the pore structure and exhibited its directing role during
the polymerization. We can predict that using a higher amine alkyl chain,
we will be able to reach the domain of mesopores. This method could then
be used to create silica with determined micropores or mesopores sizes at
low temperature and without the use of a solvent. Thus, it seems we created
a new form of porous silica. Additional experiments still have to be done to
obtain more information on the pore size of these materials. Information
can be obtained from adsorption isotherms of molecules with known
kinetic diameters and from the study of transmission electron micrographs.
Finally, with a better knowledge of this new silica, application studies could
be made, eg. by substituting silicon atoms by aluminum atoms to obtain

some porous acidic alumina-silica.

POLYMERIZATION OF TEOS INTO LAYERED
SILICIC ACID: A NEW POROUS SILICA
INTERCALATED IN MAGADIITE LAYERS

INTRODUCTION

Hydrous sodium silicate minerals are layered materials. The layers
consist of SiO4 tetrahedra units linked together. The sheets bear negative
charges which are compensated by sodium cations surrounded by an
hydration sphere. Adjacent layers are held together by electrostatic
interactions between the layers and the gallery cation and/or by hydrogen
bondings with interlayer water.14o16v22.30.33v34 The acid derivatives, called
silicic acids, are obtained by proton exchange. Those solid acids retain the
layer structure, but some rearrangements occur in the gallery due to the
loss of interlayer water molecules (see section on magadiite
characterization). The hydrous sodium silicate minerals and their acidic
analogs are starting reactants for the synthesis of new layered materials.

This synthesis of new lamellar compounds results from the intercalation of

98

 

 

99
new species in the interlayer space. It is carried out by cation exchange of
sodium or protons, by grafting or polymerization into the silicic acid
galleries.

Na-magadiite (Na2 Si14 023 (OH)2 . 8 1120)17 and its acid form H-
magadiite have been extensively studied. Indeed, Na-magadiite is easily
prepared by hydrothermal reaction.9 It has an intermediate layer thickness
(11.2 A)38 among the hydrous sodium silicate minerals family. Kanemite
(Na H Si2 O4 (OH)2 . 2 H20)12 and makatite (N212 Si4 Og (OH)2 . 4 H20)15
are lower condensed silicates while kenyaite (Naz Si22 O44 (OH)2 . 9 H20)17
is more condensed. Recently, magadiite cation exchange has been
demonstrated for lanthanum,32 cesium32 and cobalt sepulchrate.4 Ruiz-
Hitzky et al.68~69 and Yanagisawa et al.31 reported the grafting of
organosilanes and alkyldisilazane monomers after expansion of the H-
magadiite layers by polar organic solvents or by organoammonium
intercalation, respectively. Sprung et al.70 reported direct grafting of
organosilane oligomers. Several species have been polymerized into H-
magadiite, among them are acrylamide,23-7l acrylonitrile,7i caprolactam28
and tetraethylorthosilicate.2-4

The work of Sprung et al.70 and Landis et al.3 both yielded silica-
pillared magadiite, but they used different pillaring approaches. Sprung et
al.70 used a traditional-derived approach where the pillaring species were
prepared separately: an alkyltrichlorosilane was allowed to polymerize
before being allowed to interact with a H-magadiite suspension at 0°C. The
oligomers were then grafted onto the layers. Landis et a1.3 invented a new
pillaring procedure: the pillaring species were introduced as monomers in
swelled organoammonium magadiite solution and polymerized in situ at

25°C-80°C. The latter method resulted in larger surface areas.

100
On the basis of our understanding of magadiite chemistry, together
with our knowledge of the acid catalyzed polymerization of TEOS in the
presence of alkylamines, we decided to study the intercalation-

polymerization of TEOS in alkylammonium/alkylamine swelled magadiite.

RESULTS AND DISCUSSION
Synthesis

The intercalation-polymerization of TEOS in alkylammonium
magadiite can be divided into several stages. First, Na-magadiite
synthesized by hydrothermal reaction9 is proton exchanged10 with HCl to
obtain its acid-derivative: H-magadiite. The layers of the silicic acid are
swelled by the addition of alkylamine. TEOS is added to that amine-
solvated alkylammonium/alkylamine magadiite. Polymerization occurs
between the layers in the presence of alkylamine to form a siloxane-
intercalated magadiite. The remaining alkylamines and ethoxy groups are
removed by calcination and the structure is stabilized as silica-intercalated
magadiite.

We first carried out amine swelling experiments to better understand
the effect of the amine on the H-magadiite structure. We then established
the experimental parameters such as washing procedure and calcination
temperature. Finally, we studied the effect of amine chain length and TEOS

concentration on the intercalation-polymerization reaction.

Alkylammonium/alkylamine-magadiite
TEOS intercalation-polymerization reaction in magadiite galleries is
initiated by amine intercalation. As was revealed by the work of Lagaly et

al.,7 amines readily swell H-magadiite. There are several steps in the

101

swelling mechanism. Firstly, the adsorbate penetrates between the silicate
layers, causing a lattice expansion and displacing the initial gallery
molecules. For the acid derivative of magadiite, the driving force for
amine penetration is protonation. That is, the amine enters the gallery and
reacts with a proton to form an ammonium cation. Secondly, solvatation of
the resulting onium ions by neutral amine molecules results in an energy
gain from van der Waals interactions, allowing the layers to separate. The
amine chain length is an important parameter for the stabilization of the
final ammonium/amine-magadiite.

The alkylammonium/alkylamine-magadiites have been studied under
different conditions using various alkylamines. The alkylamines used were
hexylamine, octylamine and decylamine. The alkylammonium/alkylamine-
magadiites were studied as amine-solvated products or air-dried
compounds. As the amine-solvated products are unstable intermediates, we
have only been able to record their X-ray diffraction patterns. However,
the air-dried compounds, formed by evaporation of excess amine at room
temperature, have been characterized by X-ray diffraction, C, H, N
analyses and 29Si NMR.

X-ray powder diffraction
X-ray diffraction analysis has been our main analytical tool for

investigating gallery expansion. The data are gathered in Table 22.

102

 

 

 

 

 

 

 

 

 

 

 

 

l l I l l I l l l
41.3A
32.011
2. C.
E 34.011
§ 25111
E J
b
30:“ 20.2.4
I.
)‘1'12'9é'l'1o
J 1 l l l l I l 1
37.711
.3
9
”D
I
1’
C.
b. W.
, 1 Ask
> ' i ' 11 2's (1 ' s ' 10

FIGURE 35. X-ray diffraction patterns of alkylammonium/alkylamine-
magadiites gels
top: alkylammonium/alkylamine-magadiites in amine suspension
bottom: air-dried alkylammonium/alkylamine-magadiites
a. hexylammonium/hexylamine-magadiite
b. octylammonium/octylamine-magadiite
c. decylammonium/decylamine-magadiite

103

TABLE 22. X-ray diffraction data for various
alkylammonium/alkylamine magadiites

 

The alkylammonium/alkylamine-magadiite patterns in amine
suspension are displayed in Figure 35. The narrow peaks at 20.2 A, 25.1 A
and 32.0 A for the hexylamine, octylamine and decylamine samples,
respectively, are inconsistent with the presence of swelled magadiite. The
sharpness of this peak suggested the presence of alkylammonium salts
formed by reaction of the amine with atmospheric C02 or an atmospheric
acid contaminant. To confirm this hypothesis, we first crystallized the
ammonium chloride salt by adding concentrated hydrochloric acid to the
amine on a microscope slide. The X—ray diffraction patterns exhibited
narrow peaks at 19.8 A, 25.5 A and 29.3 A, for hexylammonium,
octylammonium and decylammonium chloride, respectively (Table 21).
The positions of these narrow peaks are close to those found on the
alkylammonium/alkylamine-magadiite pattems. On the basis of the amine
dimensions (Table 21), the ammonium salt crystallizes in double layers. In
H-magadiite, no counter-anions other than the silicate layers are present for
alkylammonium salt formation. Furthermore, the salt peak intensity is

dependent on the aging time of the glass slide: the older the preparation,

 

104
the less intense the narrow peak. Those two facts support the hypothesis of
a reaction related to the atmosphere. We, thus, decided to run a blank
experiment with octylamine and amorphous silica. Amorphous silica is
made of SiOz grains that do not exhibit any acid-base properties. The silica
gel served only as a solid support to spread the octylamine on a vertical
microscope slide. The pattern obtained contained a narrow peak at 25.4 A.
When the slide was allowed to air-dry, the powder did not stick anymore to
the microscope slide: octylamine had evaporated and nothing was linking
the silica grains together anymore. No X-ray diffraction pattern could be
recorded. This blank experiment confirmed the assignment of the narrow
peaks in the alkylammonium/alkylamine-magadiite XRD patterns. They are
attributed to the reaction of the amine with atmospheric C02 and water to

yield a bicarbonate ammonium salt according to the following reactions:

RNHzajq)
H20(g)m + (302(3) <=======> [H2CO3] <=======> [RNH3+, HCO3 ](solid)
<---- unstable < ----- bicarbonate
drying drying ammonium salt

As the excess liquid amine evaporates in air, the equilibrium shifts to the
left. So, the bicarbonate ammonium salt disappears on aging, as it has been

observed for our various samples.

The relatively broad XRD peaks at 30.9 A, 34.0 A and 41.3 A for
the amine-solvated samples are assigned to onium ion exchange form of
magadiite swollen by hexylamine, octylamine and decylamine, respectively.
They arise due to amine intercalation into H-magadiite. The gallery heights
(as defined in Figure 4) are 19.7 A, 22.8 A and 30.1 A for hexylamine,

 

105
octylamine and decylamine magadiites, respectively. As stated by Lagaly,29
alkylamines are arranged in double layers within magadiite. Lagaly
extensively studied amine intercalation in various silicic acids. The
magadiite intercalation of alkylamines having up to 5 carbon atoms in their
chains, results in an alkylamine monolayer parallel to the silicate layers
(Figure 36a). The resulting basal spacing is 14.4 A. For alkylamines
containing 6 to 9 carbon atoms, intercalation of gauche-block alkylamine
bilayers occurs in magadiite (Figure 36b). The resulting basal spacings
range from about 30.0 A for hexylamine to about 34.2 A for nonylamine.
For longer alkylamines, the adsorbate arranges in bilayers perpendicular
to the silicate layers (Figure 36c). The resulting basal spacing for the

decylamine intercalated magadiite is about 40.0 A.

The air-dried samples exhibit different basal spacings. The gallery
heights, as well as the crystallinity (Appendix B), decreases upon drying.
The basal spacings collapse to 14.3 A, 14.3 A and 37.7 A for hexylamine,
octylamine and decylamine magadiites, respectively. Hexylamine and
octylamine belong to the same amine category as defined by Lagaly et al..
Upon evaporation of the excess amine, the interlayer molecules rearrange
to form an alkylamine monolayer parallel to the silicate layers
characterized by a 14.3 A basal spacing. In order to rearrange into a
monolayer, excess amine has to be removed and, thus, van der Waals
interactions have to be overcome. Although the decylamine-solvated sample
already belonged to a different category, the particularity of the
decylamine intercalation is better distinguished when the sample is air-
dried: the basal spacing is maintained at 37.7 A after one month at room

temperature. For the air-dried hexylamine and octylamine samples, the

106

 

FIGURE 36. Alkylamine intercalation into silicic acids, according to the
number of carbon atoms in their chain

a: monolayer

b: gauche-block bilayer

c: perpendicular bilayer

107
basal spacing collapsed after a few hours. The explanation is certainly
related to the longer chain length of decylamine. It results in a higher
boiling point for decylamine: 216-218°C. Hexylamine and octylamine
boiling points are 131-132°C and 175-177°C, respectively. Its vapor
pressure is certainly lower. So, decylamine molecules can not readily
escape from the magadiite layers. Also, in a bilayer conformation, the
decylamine has 8 or 4 more carbon atoms than the hexylamine and
octylamine, respectively. Those additional carbons create more van der

Waals interactions that must be overcome to evaporate the excess amine.

Van der Waals interactions between intercalated amine chains are
related to crystallinity. The greater the van der Waals interactions, the
better the crystallinity. As the number of carbon atoms in the alkylamine
chain increases, from 6 for hexylamine, to 8 for octylamine and to 10 for
decylamine, the size of the scattering domain along the layer stacking
direction of the amine-solvated alkylammonium/alkylamine magadiite
increases from 260 A to 370 A and to 590 A, respectively. H—magadiite
exhibits a scattering domain of 50 A. When the amines lie parallel to the
layer, and so, when van der Waals interactions are minimal, the air-dried
hexylamine and octylamine-magadiites scattering domain sizes are identical
at ~l90 A. The gallery order for amine-intercalated magadiite is still better
than that for H-magadiite. For air-dried decylamine-magadiite, even if the
amine did not evaporate completely, the degree of order changed owing to
a decrease in the number of van der Waals interactions. The crystallinity
went down to 340 A. This value is lower than the initial one, but higher
than those of the air-dried hexylamine and octylamine-magadiites. This is

in good agreement with our general conclusion, namely, for amine-

108

intercalated magadiite, van der Waals interactions between the chains and

crystallinity are related.

Chemical compositions

In order to understand better this phenomenon, we studied the
carbon, hydrogen and nitrogen contents. The data are presented in Table
23.

TABLE 23. Chemical composition from elemental analysis for the air-
dried alkylamine~intercalated magadiites

 

 

 

%wt element %wt com sifion
air-dried %wt %wt %wt %wt %wt %wt ¢%wt %wt
amine-intercalated c‘ H‘ N' 81“ o" 8102“ amine” H20“

 

 

 

 

hexylamme 6.3 1.9 l .2 40 50 86 8 6
octylamine 7.8 2.1 1.1 40 49 85 10 5
decylamine 56 11 6.2 10 17 20 70 10

 

‘ measured from C, H, N analysis
" calculated according to the method explained in the text

The C, H, N results from the chemical analyses are given in ppm (ug/l) for
C, H and N. By assuming a density of 1 g/l for the analyzed solutions, we
can convert those results to %wt.

The amine content has been calculated from the %wt N, given by the
analysis:

nA: moles of nitrogen in 100 g = moles of amines in 100 g

109

 

%wt amine = nA * MA MA = 101.19 g/mol for hexylamine
MA = 129.25 g/mol for octylamine
MA = 157.30 g/mol for decylamine

The water content has been calculated using the %wt H left after deduction
of the hydrogen present in the amine:
nHzO is defined as the number of moles of water created by silanol

condensation and by desorption of water per 100 g of compound.

%L‘ZLH-mnA)/2=(%“:tH-h*nA)/2
h = 15 for hexylamine
h = 19 for octylamine
h = 23 for decylamine

%wt H20. = 111120 * M1120 = 111120 * 18

 

 

DH20=(

The silicon content in 100 g of compound has been calculated, assuming
that after the C, H, N analysis, the remaining product is silica SiOz.
%wt SiOz = 100 - %wt amine - %wt H20

IfnSioz and nSi represent the number of silica molecules and silicon atoms,
respectively, in 100 g of compound:
%wt 3102 = nsroz * M5102 = nsroz * 60-1
‘7 wt SiO
I151 = nSioz = W

%wt Si = my * MSi = n5i02 * 28.1

110
Finally, the total oxygen content in 100 g of compound is determined to

complete the elements' weight % to 100:
%wtO =100-%wtC-%th-%th-%thi

As the swelling experiments are intercalation reactions, it is
therefore possible to calculate the number of amines remaining per Si14
unit after air-drying (Table 24).

TABLE 24. Number of amines present per unit cell
for the air-dried alkylamine-intercalated magadiites

 

 

air—dried
amine-intercalated hexylamine octylamine decylamine
magadiite
Si 14 l4 l4
amine 0.8 0.8 18

 

 

 

 

The validity of the chemical analysis can be checked: the carbon
content should match the amine content as the carbon only comes from the

amine aliphatic chain (Table 25).

c = 6 for hexylamine
c = 8 for octylamine
c = 10 for decylamine

l l 1
TABLE 25. Validity of the C, H, N analysis

 

 

 

air-dried
amine-intercalated hexylamine octylamine decylamine
ma adiite
%wt 0 12 0.52 0.65 4.6
c * mm, 0.50 0.61 4.4
% error 4 7 5

 

 

 

 

As the ideal H-magadiite formula is H2 Si” 023 (OH)2 . x H2O
(alternatively written: Si14 O26 (OH)4 . x H2O), the maximum number of
ammonium molecules produced by the acid-base reaction between amines
and hydrogens from silanol groups and from the interlayer protons is 4. It
appears (Table 24) that air-dried hexylamine- and octylamine-intercalated
magadiites lose not only all the neutral amines, but also some of the
ammonium molecules. In the case of a C6 or C3 chain, the alkylammonium
interaction with the layers is not strong enough to maintain the layers apart
in the absence of neutral amine. To balance the charges, the
alkylammonium cations deprotonate before leaving the layers. They
dissociate into amine and protons, and leave the protons between the layers
to balance the charges.

As for the air-dried decylamine-intercalated magadiite,
alkylammonium cations and neutral amines still form a bilayer, we took
into account geometrical considerations. Using molecular graphics to
simulate molecule configuration and size, we find the distance between two
hydrogen atoms linked to the nitrogen to equal 1.747 A. By projection and
geometrical considerations, we calculated the radius of the amine head

surface: m~l A (1.747/2*cos30°). Thus, the surface area required per

amine in the vertical position is at least 3.2 A2 (nrNZ). However, there are

112
some unoccupied spaces. Assuming a closest-packed arrangement, we
consider the stacking hexagonal. Accounting for the unoccupied spaces, we
find a needed surface per amine of at least 4 A2 ([2*l/tan30°]2*3/9).
According to Brindley,38 the lattice parameters in the sheet plane are
a=b=7.25 A. So, the available space for an amine double layer is 105 A2
(atthSheets) per Si14 unit. If, as has been calculated from chemical
analyses, there are 18 decylarnines per Si14 unit, they would occupy, in the
closest-packed configuration, 72 A2 (l8amines*4A2). So, when the
decylamine-intercalated magadiite is air-dried, the remaining decylarnines
are not in a close-packed vertical configuration. This fact can account for

the loss in crystallinity and basal spacing.

29Si nuclear magnetic resonance

The air-dried decylamine-intercalated magadiite is stable and exhibits
a large basal spacing. We thus have been able to obtain 29Si nuclear
magnetic resonance spectra and to determine precisely the Q3/Q4 ratio
without any interlayer interaction (Table 26). The air-dried decylamine-
intercalated magadiite solid state 29Si NMR spectrum is displayed in Figure
37.

TABLE 26. Various cation-magadiites 298i NMR data

 

 

 

 

 

sample Si chemical shifts (ppm vs. TMS) Q3/Q4 per unit cell

Q3 Q4 1311°_ Q3 Q4

Na-magadiite -99.5 -ll0.0 -lll.4 -ll4.l 0.39— . 1 .l

H-magadiite -101.2 -110.9 -lll.7 -ll4.3 0.29 3.2 10.8
air-dried

decylamine-intercalated ~100.8 -111.6 -114.9 0.29 3.2 10.8
magadiite

 

 

 

 

 

 

113

 

 

IllllllllllllIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

—90 -100 -110 -120 ppm

FIGURE 37. Air-dried decylamine-intercalated magadiite solid state 2981
NMR spectrum
(delay time: 1200 seconds, line broadening: 40, 12 scans)

114

The air-dried decylamine-intercalated magadiite basically exhibits
the same NMR spectrum as H-magadiite (Figure 14c). The Q3/Q4 ratio, as
defined in the section on magadiite characterization is identical. It proves
that the Q3/Q4 ratio calculated from the H-magadiite spectrum is not
perturbed by the hydrogen bondings between the layers. It also tells us that
the amine intercalation only separates the layers and does not have any
effect on the layer structure. This is a noticeable difference in comparison
to the effect of aqueous alkyltrimethylammonium intercalation into
kanemite.42-73 As kanemite is constituted of a single sheet of SiO4
tetrahedra, its 2S’Si NMR spectrum only displays one resonance at -97.2
ppm42 or -101 ppm” assigned to Q3 silicon atoms. However, upon
intercalation of dodecyltrimethylammonium42 (hexadecyltrimethyl-
ammonium),73 the Q3 peak shifts to higher values: -100.3 ppm (-100.8
ppm) and one Q4 peak appears at -109.2 ppm (-110.2 ppm). This behavior
did not occur for Na-magadiite and K-kenyaite treated in the same
conditions.31 This difference was said to be due to the deformation of the
single SiO4 tetrahedral sheet that allows for natural condensation upon

alkyltrimethylammonium intercalation.42

Conclusion

To conclude on this set of experiments, it is worth noticing that the
three alkylamines exhibit similar magadiite swelling behavior in the
presence of excess amine. They incorporate a bilayer configuration inside
the gallery when fully solvated by excess amine. The intercalation is
initiated by an acid-base reaction between the amines and hydrogen ions
from silanol groups and interlayer protons. The structure is stabilized by

van der Waals interactions between the aliphatic chain of the

115

alkylammonium cations and the neutral amine. Air-dried hexylamine- and
octylamine-intercalated magadiite exhibit a monolayer structure. However,
air-dried decylamine-intercalated magadiite still retains an amine bilayer
configuration. Therefore, it would not be surprising if the products
resulting from TEOS polymerization in pre-swollen magadiite exhibit

properties dependent on the alkylamine reactant.

Experimental conditions for TEOS polymerization in H-
magadiite

An important parameter in the final design of the silica-intercalated
magadiite is the calcination temperature. Indeed, heat treatment stabilizes
the system, in particular, by eliminating the organic molecules. We thus
carried out experiments at various calcination temperatures to establish the
best calcination temperature for the further studies. As one purpose of the
heat treatment is to eliminate organic molecules, we also carried out
syntheses where an ethanol washing procedure was included before
calcination. In this set of experiments, the silica-intercalated magadiite was
synthesized from 5.04 g (5.6”‘10-3 mol) of H-magadiite, 20.2 g (0.16 mol)
of octylamine and 120 g (0.58 mol) of tetraethylorthosilicate. The

TEOS:octylamine:magadiite molar ratio was 103:2821.

Thermogravimetric analysis

In order to choose a suitable calcination temperature, we obtained
first thermogravimetric analysis curves (Figure 38). The numeric data are
gathered in Table 27. The %wt SiO2 determined from thermogravimetric
analysis has been calculated assuming the remaining mass left after heating

at 800°C is pure silica.

 

\O
C
l

% weight remaining
00
o
l V

 

70 '

l l

l l

 

 

 

0 100 200 300 400 500600700

temperature °C

800

FIGURE 38. Thermogravimetric analysis curves for octylamine—derived
siloxane-intercalated magadiites
a. unwashed octylamine-derived siloxane-intercalated magadiite

b. octylamine-derived siloxane-intercalated magadiite washed once

with ethanol

TABLE 27. Chemical composition from thermogravimetric analysis
for unwashed and ethanol-washed octylamine-derived siloxane-
intercalated magadiites

 

 

 

 

octylamine-derived first second third %wt
siloxane-intercalated %wt loss: %wt loss: %wt loss: SiO2
magadiite external water octylamine dehydroxylation
unwashed l . 6 20 2.7 76
ethanol-washed 3.7 3. l 7.9 85

 

 

 

117

The thermogravimetric curves show a three-step weight loss. The
first one from room temperature to about 150°C is assigned to external
water loss. It is almost identical for the unwashed and the ethanol-washed
samples. The second weight loss from ~150°C to ~350°C is assumed to be
due to the octylamine elimination. This second step is much larger in the
case of the unwashed sample, suggesting that washing, indeed, removes the
amine. The last slope occurring between 350°C and 800°C is assigned to
silanol dehydroxylation. We thus studied samples formed at the following
calcination temperatures: 100°C, after water elimination ; 250°C, after
partial amine evaporation ; 450°C, when amine elimination is complete ;

800°C, after partial dehydroxylation.

Chemical compositions
The assignments of the various weight loss have been confirmed by
C, H, N analysis (Table 28).

As was found from the thermogravimetric analysis, the unwashed
octylamine-derived siloxane-intercalated magadiite contains much more
octylamine than its washed counterpart. The octylamine content found by
the two methods are in agreement: 20% and 3.1% from the
thermogravimetric analysis, 18% and 3.2% from the C, H, N analysis, for
the unwashed and washed samples, respectively. Also, the comparison of
the SiO2 content quantity measured by thermogravirnetry (Table 27: 76%
and 85%) and the one found by calculation from the C, H, N analysis
(Table 28: 75% and 85%) proves the reliability of our calculations.

 

118

TABLE 28. Chemical composition from elemental analysis for unwashed
and ethanol-washed octylamine-derived siloxane- and silica-intercalated
magadiites at various calcination temperatures

 

 

 

 

 

 

%wt element %wt composition
sample %wt %wt %wt %wt %wt %wt %wt %wt %wt
C' H' N‘ Si“ O” amine” ethoxy” H2O" SiOz"

unwashed 14 4.0 2.0 35 45 18 1.5 5.8 75
1(X)°C 13 3.4 1.7 35 47 16 2.0 7.0 75
250°C 5.8 2.1 0.66 39 52 6.1 2.4 8.1 84
450°C 0.78 1.6 0.09 40 57 0.83 0.31 13 86
800°C 0.32 1.4 0.14 41 57 1.3 0.00 11 88
washed 4.4 1.8 0 35 4O 54 3.2 3.7 7.5 85
1(X)°C 5.8 2.0 0 83 40 52 7.7 0.17 7.3 85
250°C 4.4 1.7 0 23 4O 53 2.1 5.3 6.3 86
450°C 0.37 0.99 0.14 43 55 1.3 0.00 7.2 92
8(X)°C 0.17 0.57 0.07 45 55 0.65 0.00 4. 3 95

 

 

‘ measured from C, H, N analysis

" calculated according to the method explained in the section acid-catalyzed TEOS
polymerization in the presence of alkylamines

Furthermore, the water composition is found to be larger in the case of the
washed sample by the two analysis: from Table 27, we find a total water
composition of 4.3% and 11.6% ; Table 28 gives the water %wt of 5.8%
and 7.5%. The amount of octylamine is always larger in the unwashed
samples, even at various calcination temperature. This quantity decreases
when the calcination temperature is increased, but octylamine is still
present even at 800°C. However, the ethoxy-ethanol groups are completely
eliminated after calcination at 450°C.

119
X-ray powder diffraction
We next look at the X-ray diffraction data (Table 29) to determine if
the layer structure is retained. The X-ray diffraction patterns for the
unwashed and ethanol-washed octylamine-derived siloxane- and silica-

intercalated magadiites are displayed in Figure 39.

TABLE 29. X-ray diffraction data of unwashed and ethanol-
washed octylamine-derived siloxane- and silica-intercalated
magadiites at various calcination temperatures

 

 

 

unwashed ethanol-washed
samvle (1001 (A) gallexy door (A) gallery
height (A) he1ght (A)

siloxane-intercalated 36.0 24.8 18.9 7.7
silica-intercalated 100°C 34.7 23.5 18.9 7.7
silica-intercalated 250°C 34.5 23.3 18.9 7.7
silica-intercalated 450°C 33.2 22.0 17.1 5.9
silica-intercalated 800°C 31.3 20.1 15.9 4.7

 

 

 

 

 

The octylamine-derived siloxane- and silica-intercalated magadiites
display large basal spacings (Table 29) which are confirmed by the
presence of the dooz peaks (Figure 39: left). Upon calcination, the gallery
height (= basal spacing - layer thickness = dom - 11.2) slightly decreases up
to a calcination temperature of 250°C. But, when the ”temperature is
increased to 450°C and 800°C, the doc” decreases further. As it has been
seen by thermogravimetric and chemical analyses, the amine is essentially
eliminated at 450°C. Thus, up to 250°C, the alkylamine might still play a
role in propping apart the layers. Then, the basal spacing decreases due to
the stacking of the layer onto the gallery species (450°C) and drops further
because of the condensation-dehydroxylation of the gallery species onto the

layers (800°C). Ethanol-washing not only eliminate some of the amines

120

 

 

 

 

 

 

 

 

 

 

 

Illlllll lllLlllLl
.‘E‘
§ .‘é a.
.5 g
g .3 L‘—
‘5 .—
2 2 b.
2
C.
JLa
\jLC;
lllllllll lTlrjllll
0 2 4 6 810
28 0 2 4266 810

FIGURE 39. X-ray diffraction patterns of unwashed and ethanol-washed
octylamine-derived magadiites at various calcination temperatures

left: unwashed octylamine-derived siloxane- and silica-intercalated
magadiites (film)

right: ethanol-washed octylamine-derived siloxane- and silica-intercalated

magadiites (pwd)
a. siloxane-intercalated magadiite b. silica-intercalated 100°C
c. silica-intercalated 250°C d. silica-intercalated 450°C

e. silica-intercalated 800°C

 

121

(Table 28) but it also decreases the basal spacing. As some amines are
eliminated, they can not play their role as well as if they are more
numerous. But, this phenomenon can be understood if we recall the acid-
catalyzed TEOS polymerization in the presence of alkylamine (see the
section on acid catalyzed TEOS polymerization in the presence of
alkylamines). Indeed, the silica-product of the polymerization was found to
dissolve in ethanol and to crystallize upon evaporation of the solvent. The
same phenomenon, most probably, also is occurring for the siloxane
formed inside the magadiite layers. However, we do not know how well the
silica-product recrystallizes and also, at that point, the recrystallization

occurs in the absence of amine.

Nitrogen adsorption

As we studied the TEOS polymerization inside the H-magadiite
layers in order to be able to tailor new microporous solids, we ran some
nitrogen adsorption experiments. The N2 adsorption isotherms are
displayed in Figures 40 and 41. The surface areas data are gathered in
Table 30 (Appendix C).

122

Uooow c8a=§3£1ao=mm .0

0°09. 3385-85“ d 0.63 uaafiesséommm .o

012: 353235-33 A 3:23: Baneastfiéfi a
83; ”Emm— .EB€8_ .8598? «2 £2
“2332—88 conga—8 32:; 8 mamgwafi
333285-85... vo>rov$5§boo @3335. me 323 .8398? «Z .9. EDGE

A3 a
3 w c v a co

 

 

 

 

 

 

 

.l\ 1
1 1 on
m
o .
a 11..
1 8H m.
m. cw
m 8.
S
1 C? nu L 92
. m .. .. 8N
_ .o n 1 onm
1 com 1
1 com
1 own
1 b 1 p 1 p 1 - 1 OWN _ u p - 8V

 

 

 

 

(S/oo) (us omnIOA poqrospe

 

123

Uooow waafieafiéofim .o
0.63 naafieonfiéoam 6 000nm can—uonafiéofiw .o
009: cam—«285-8311. .n 8:392: 333885-28on .a
83-. an»: .Euofiofl 5:983 «2 £2
mocanmEB 183523 ESE; 3 mazcmwafi c8£a88£
135m uo>touoEE£boo 60:33-33”. .«o «958 52883 «2 .31 MMDOE

 

 

 

 

 

 

 

 

 

 

 

3s a
2 w o #1 N o
1 - 1 . 1 1 1 . 1 C
1 mm
B
p
m PE
N. CA wd ed v.0 Nd o5
1 CW ”1 1 . 1 . 1 q 1 . 1 O
o
f m1 _ . 3.808.388.888081‘1
.1 .n \\ on
S
.. 1 m. "u 0803003000030
.n m; Run-III: .o
. . o 1 o2
1.... . o2 .
o
32
r .
. 1 h 1 h 1 n 1 WNfi 1 . 1 . 1 . 1 - 8N

(8/00) as amnIOA peqrospe

124

TABLE 30. N2 adsorption data of unwashed and ethanol-washed
octylamine-derived silica-intercalated magadiites at various calcination
temperatures

 

 

 

 

 

 

octylamine-derived 133T43 t-plot‘“
Silica-mm S [on] S mu] Snort-u Spa) i 811(2) . V“
magadiite m2/g m2/g mzlg mZ/g mzlg liq “113/g
unwashed precursor
1(X)°C 70 50 50 0 0 0
250°C 670 690 100 680 590 0.24
450°C 760 790 100 770 690 0.27
800°C 620 630 90 610 540 0.21
washed precursor
1(X)°C 40 40 40 0 0 0
250°C 250 280 80 200 200 0.07
450°C 320 380 50 300 330 0.10
800°C 90 90 80 20 10 0.01

 

 

 

S”: microporous surface area, 8mm-“ : non-microporous surface area, V“: microporous
volume
" Spa) is derived from V“ by the t-plot method ; Sum =Smm _ Sm“.u

As we already saw in the section on acid-catalyzed TEOS
polymerization in the presence of alkylamines, the total surface areas
calculated from the BET equation43 and the t-plot method“ agree relatively
well. However, the unwashed samples' microporous surfaces determined by
the classic de Boer et a1.“4 t-plot method are always higher than the
difference between the total surface area and the non-microporous surface
area. This might be due to the fact that the micropores are so large (see
XRD patterns on Figure 39: left, display gallery height of 20 to 23 A) that
they do not match typical micropores (pore < 20 A), but are too small to
be assigned to mesopores (20 < pore < 500 A). On the other hand, SH“) and
8,10) values for the ethanol-washed samples (XRD patterns on Figure 39:
right, display gallery height of 5 to 8 A) are in agreement.

 

800

125

 

surface area (m2/g)
#1 u Ox \1
e a as $

é

microporous surface

 

non microporous surface
............................. .

 

 

 

   

400
microporous surface .
' 300
- 200
non microporo surface - 100
1,. ---------------- 1
0

 

 

 

160 250 450 800
calcination temperature °C

160 250 450 800
calcination temperature °C

FIGURE 42. Dependence of the microporous and non-microporous surface

are

as on the calcination temperature

left: unwashed octylamine-derived silica-intercalated magadiites
right: ethanol-washed octylamine-derived silica-intercalated magadiites

126

Figure 42 is a plot of the microporous (Sum) and non-microporous
(Snomu) surface area data from Table 30 as a function of the calcination
temperature. Clearly, as we eliminate the octylamine upon heating from
100°C to 450°C, the pores are unplugged and the nitrogen gas molecules
can access them. So, the pores are shaped by the octylamine. However,
when the calcination temperature is increased to 800°C, the total surface
area decreases, almost certainly due to a collapse of the structure. The non-
microporous surface area is about the same for each samples: ~80 m2/g. It
is independent of the calcination temperature and on the washing
procedure. For the two washing procedures, the maximum surface area is
obtained for the 450°C calcination temperature. Also, we can see that
ethanol-washing decreases the total surface areas. This result can be related
to the depletion of amine and the loss of gallery height seen in Table 29.
So, either the microporous surface area is dependent on the size of the
gallery height, or after dissolution of the siloxane in ethanol, the silica

recrystallizes in a different way.

Infrared spectroscopy

We set the experimental parameters to 450°C for the calcination
temperature with no ethanol-washing according to structural analyses. We
then tried to relate these physical behaviors to chemical properties. First,
we carried out infrared spectroscopy experiments. The spectra are

displayed in Figures 43 and 44. The data are gathered in Table 31.

 

127

 

relative absabmce

 

 

 

l

l l l l I I l l
1300 1100 900 700 500
Wavenumber (curl)

 

1 1 1
1900 1700 1500

 

relative abwrbancc
?

 

 

 

4000 3800 3600 3400 3200 3000 2800 21500
Wavenumber(cm"')

FIGURE 43. Infrared spectra of unwashed octylamine-derived siloxane-
and silica-intercalated magadiite (KBr pellet, room temperature)
a. siloxane-intercalated magadiite b. silica-intercalated 100°C
c. silica-intercalated 250°C d. silica-intercalated 450°C
e. silica-intercalated 800°C

128

relative absorbance

 

1900 1700 1500 1300 1 I“) 900 700 500
Wavenumber (an'l)

 

relative absorbance

 

l l l l l I l

l l I
3600 3400 3200 3000
Wavenumber (curl)

 

 

1 l l
4000 3800 mo 26C!)

FIGURE 44. Infrared spectra of ethanol-washed octylamine-derived

siloxane- and silica-intercalated magadiite (KBr pellet, room temperature)
a. siloxane-intercalated magadiite b. silica-intercalated 100°C
c. silica-intercalated 250°C (1. silica-intercalated 450°C
e. silica-intercalated 800°C

129

TABLE 31. Infrared data of unwashed and ethanol-washed octylamine-
derived siloxane- and silica-intercalated magadiites

 

octylamine- wavenumber (cm'l)
derived
silica- .

in l ted V(O-H) 5(1-120) V(Sl-O)
magadiite V(CH2.CH3) V(C-C)

 

 

 

 

unwashed 3620 2959 2872 1631 1469 12 107 821 613
precursor 3440 2928 2857 1200 796 576

 

 

1(X)°C 3620 2959 2872 1631 1469 1233 1078 975 821 619
3440 2928 2857 1200 796 576

 

250°C 3440 1631 1233 1078 975 821 613
2928 2857 1194 796 576

 

450°C 3440 1631 1226 1078 975 821
796 576

 

800°C 3440 1631 1226 1078 969
803 588

 

 

washed 3650 2959 2872 1631 1233 1078 975 821 613
precursor 3440 2928 2857 1200 796 576

 

 

1(X)°C 3658 2959 2872 1631 1233 1078 975 821 613
3440 2928 2857 1200 796 576

 

 

250°C 3650 1631 1233 1078 775 821 613
3440 2928 2857 1200 796 576

 

 

 

 

 

 

 

 

 

130

 

 

 

 

 

450°C 3658 1631 7 75
3440 803 576
542
483
450

8(X)°C 3669 1631 1W8 969
3440 803 592
450

 

 

 

 

 

 

 

 

There are only small differences in the infrared spectra of the
unwashed (Figure 43) and ethanol-washed samples (Figure 44). The only
difference, as well as the one between the calcination temperatures is the
amount of octylamine. There is almost no octylamine present (Figure 22b :
2959, 2928, 2872, 2857 and 1469 cm-1) in the ethanol-washed samples.
However, in the unwashed samples, the amine is almost completely
removed at 450°C. In the Si-O vibration region, we can recognize the
bands originating from the magadiite layer structure: 1233-1223, 1200,
1078, 975, 821, 613, 576-579, 542 and 483 cm-1 (Figure 12). The bands at
1237 and 1210 cm-1 for Na-magadiite were assigned by Garces et al.22 to
the presence of SiO4 units forming five-member rings. Also, the 579-576
and 542 cm-1 bands represent the presence in the structure of double
rings.2237 The Si-O asymmetric stretch vibration at 1078 cm-1 is the most
intense band in the Na-magadiite and H-magadiite spectra as well as on the
octylamine-derived siloxane- and silica-intercalated magadiite.The 975 cm-1
vibration is a Si-OH stretching band for isolated silanols.40 This set of
infrared experiments tells us that the silicate structure is retained after the
polymerization of TEOS inside the magadiite layers. To get information

about the silanols, we performed some infrared measurements

 

131

 

l l l l l l
3525
3719
3744

3670

relative intensity

 

 

 

4000 3800 3600 3400 3200 3000

Wavenumber (cm'l)

FIGURE 45. Hydroxyl stretch FI‘IR bands under vacuum at 150°C for
unwashed octylamine-derived silica-intercalated magadiite (450°C)
(1 mg of sample in a 100 mg KBr pressed pellet)

132

under vacuum at 150°C to eliminate the water molecules filling the pores

(Figure 45).

Four hydroxyl stretch bands can be seen at 3744, 3719, 3670 and 3525 cm-
1. By comparing with the data for Na-magadiite and H-magadiite samples
(Table 8), we find that the sharp bands at 3744 and 3719 cm-1 were already
present in the pure Na-magadiite pressed pellet sample. These two bands,
however, were absent in the Na-magadiite KBr pellet and H-magadiite
samples. These sharp peaks are indicative of isolated silanols. They are
regenerated during the intercalation reaction. The two broad peaks at 3670
and 3525 cm-1 could be assigned to the two 3595 and 3497 cm-1 peaks from
the Na-magadiite structure. They are shifted to higher wavenumbers

because of the presence of the intercalated silica.

29S1 nuclear magnetic resonance
To obtain further information for the silanol groups, we conducted
some 29Si nuclear magnetic resonance experiments. The spectra are

displayed in Figures 46 and 47.

By comparing the spectra with those of H-magadiite and the products
of the acid-catalyzed reaction of TEOS in presence of alkylamines, we can
see that the peaks shapes and positions are close to those of H-magadiite.
The retention of the magadiite structure found by infrared spectroscopy is
confirmed by 29Si nuclear magnetic resonance. We can also notice, for the
unwashed and washed samples, that the Q3 peak relative to (SiO)3§iO-Z
(Z¢Si) silicon environment41 is decreasing in intensity and broadening as

the calcination temperature is increased. It even displays a tail to lower

133

 

W
-90 ‘100 “110 '120 me

FIGURE 46. 29Si NMR spectra of unwashed octylamine-derived siloxane-

and silica-intercalated magadiites

(delay time: 600 seconds, line broadening: 40, 12 scans)
a. siloxane-intercalated magadiite b. silica-intercalated 100°C
c. silica-intercalated 250°C d. silica-intercalated 450°C
e. silica-intercalated 800°C

134 ‘

 

 

I'llI1117]]IIIIYTIVITTIIIIIIIIIIII|IIIIIVIII‘VYT1_I

--90 —100 -110 -120 ppm

FIGURE 47. 29Si N MR spectra of ethanol-washed octylamine-derived

siloxane~ and silica-intercalated magadiites

(delay time: 600 seconds, line broadening: 40, 12 scans)
a. siloxane-intercalated magadiite b. silica-intercalated 100°C
c. silica-intercalated 250°C (1 silica-intercalated 450°C
e. silica-intercalated 800°C

 

135
field, suggesting the presence of the Q2 silicon environment (SiO)2S_i(O-Z)2
(MD. This tail is confirmed on the 450°C samples, and sometimes the
250°C ones, by the presence of a peak at about 92 ppm. However, upon
calcination at 800°C, the Q2 peak weakens and the Q3 peak gets shifted
toward the Q4 peak: they can not be differentiated anymore. So, when the
octylamine-derived silica-intercalated products are calcined at too a high
temperature, the magadiite layers start becoming destroyed. The
asymmetry of the Q4 silicon environment S_'1(0Si)4 peak suggest several
contributions to that kind of tetrahedra. Because of those overlaying Q4
peaks as well as the shallow Q2 peak and the overlapping of the Q3 and Q4
peaks, the ratio of the different silicon environment by deconvolution is
complex and can lead to various conclusions. Up to this point, only
qualitative conclusions can be drawn: upon increasing the calcination
temperature, more silanol groups (Q3) condense to form siloxane bridges

(Q4) ; the magadiite layer is altered at 800°C.

Conclusion

In conclusion, X-ray diffraction analysis and nitrogen adsorption
analysis helped us to choose the optimum experimental parameters: in the
further study we did not include a washing procedure during the synthesis
and the calcination temperature was set to 450°C. Indeed, ethanol-washing,
even though apparently not changing the general chemical features of the
products, dissolves the product of the TEOS polymerization. As it
recrystallizes in absence of alkylamine upon evaporation, the arrangement
inside the layers is disturbed. According to 29Si nuclear magnetic
resonance, calcining the siloxane-intercalated silica at too a high

temperature, not only converts the siloxane into silica, but also partly

136
destroys the magadiite structure. That is why, upon increasing the
calcination temperature up to 450°C, the pores are freed from octylamine
and the surface area increases. But then, with no more amine in the
structure above 450°C, the surface area decreases due to a collapse of the

structure.

Alkylamine and TEOS to magadiite molar ratio effects

The synthesis of new porous materials is a fast developing field, in
particular in the silicon domain.2-4v42.7°v73174 The goal of those studies is to
tailor microporous and mesoporous compounds in a wide range of specific
and uniform pore sizes. Indeed, zeolites only exhibit homogeneous pore
sizes in the range 4 A to 15 A. Thus, crystalline materials with well-
defined pore sizes in the range 15 A to 30 A are not available for industrial
applications. Different approaches have been used. One route is to use a
layered starting material. Porous material is created either by polymerizing
inside the layers of metal alkoxides M(OR)X,2-4 either by grafting oligomer
species70 or by generating partial condensation of the layers.42v73 A more
recent procedure is to polymerize metal alkoxide M(OR)x in the presence
of an onium ions.74 In the preceding chapter on the acid-catalyzed TEOS
Si(OCH2CH3)4 polymerization in the presence of alkylamines, we studied
the synthesis of microporous materials with sizable pore sizes. After
establishing the experimental parameters, we are going to study a parallel
approach making use of layered materials. We are then going to relate the
products to those from the acid-polymerization of TEOS in the presence of
alkylamine.

In order to vary the pore sizes, we studied the polymerization of

TEOS into H-magadiite in the presence of three different amines:

137

hexylamine, octylamine and decylamine. The increase in the amine length
is expected to increase the pore size in the c-axis direction. We also used
various TEOS:amine:magadiite molar ratios of 200:27:1, 100:27:1 and
50:27:1, as defined in Table 32.

TABLE 32. Experimental conditions for the study of the
polymerization of TEOS into H-magadiite

 

 

 

 

 

sample amine calcination molar ratio
temperature TEOS amine magadn?

HAMAGSO hexylamine none 50 27 l
CHAMAGSO 450°C 50 27 1
OAMAGSO octylamine none 50 27 1
COAMAGSO 450°C 50 27 1
DAMAG50 decylamine none 50 27 1
CDAMAGSO 450°C 50 27 1
HAMAGlOO hexylarrTrne none 100 27 l
CHAMAGIOO 450°C 100 27 1
OAMAGlOO octylamine none 100 27 l
OOAMAGlOO 450°C 100 27 1
DAMAGlOO decylamine none 100 27 l
CDAMAGIOO 450°C 100 27 1
HAMAG200 hexylamine none 200 27 l
CHAMAG200 450°C 200 27 1
OAMAGZOO octylamine none 200 27 1
COAMAGZOO 450°C 200 27 1
DAMAGZOO decylamine none 200 27 1
CDAMAG200 450°C 200 27 1

 

 

 

 

 

 

Chemical compositions
We first studied the effect of the amine on the chemical composition

of the siloxane- and silica-derived magadiites (Table 33).

138
TABLE 33. Chemical composition from elemental analysis for unwashed

alkylamine-derived siloxane- and silica-intercalated magadiites at the
TEOS:amine:magadiite molar ratio = 100:27:1

H‘ O“

OAMAGIOO 14' 3.5 . 35 45 18 1.5 5.8 75

31 43

COAMAGIOO 0.78 1.6 . 40 57 0.83 0.31 13 86
41

 

” calculated according to the method explained in the section acid-catalyzed TEOS
polymerization in the presence of alkylamines

Contrary to our findings for the TEOS-derived silica precursor
(Table 14), the change in the amine dimension does not greatly affect the
chemical composition of the siloxane-intercalated magadiites. The only
significant change is in the increasing presence of ethoxy—ethanol groups
with the increase of the amine chain length. It should be noticed, from a
comparison of Tables 14 and 33, that the amount of alkylamine present is
much lower than in the case of the polymerization without any layered
support. In some extent, this can be understood by the presence of the
silicate layers which contribute to the %wt $02. After calcination, most of
the amine is removed as well as most of the ethoxy groups which are most
probably converted to hydroxyl groups. The resulting chemical

composition is close to those found for the TEOS-derived silica (Table 14).

139

2S‘Si nuclear magnetic resonance

To further investigate the ethoxy-ethanol and silanol groups, we
conducted some 2S’Si nuclear magnetic resonance experiments (Figures 48
and 49).

To obtain the correct Q4:(Si0)4fii, Q3:(SiO)3S_iO-Z and Q2:
(SiO)zS_i(O-Z)2 (Z¢Si) peak ratios,41 the spectra were recorded with a time
interval of 600 seconds. This delay time is about four times larger than the
relaxation times found to be 141 :t 11 s for a Q4 environment and 44 i 5 s
for a Q3 environment.4 Each unwashed amine-derived siloxane-intercalated
magadiite spectra exhibits one Q3 peak at -100.6 ppm and a Q4 peak
centered at -110.9 ppm. However, the HAMAG10O Q3 peak is broader than
the OAMAGIOO and DAMAGlOO ones. Moreover, as the Q4 to Q3
intensity ratio is about the same for the three samples, we can say that the
hexylamine-derived sample possesses more silanols (Q3) than the two other
samples. After calcination at 450°C, the unwashed amine-derived silica-
intercalated spectra exhibit less defined Q3 and Q4 peaks due to the Q3 peak
upfield shift and smaller intensity ; the separation of those two peaks is
improved when the amine chain length is increased. Also, as we saw in
Figure 46 and on the samples studied at 360°C,4 the resolution of the peaks
gets worse as the calcination temperature is increased. However, a Q2 peak
appears at about -91 ppm. Thus, upon calcination at 450°C, rearrangements
occur: some silanol groups (Q3) are converted to siloxane bridges (Q4) by
condensation, but, surprisingly, some Q2 silicon environments appear. The
presence of those Q2 peaks in the calcined samples and their absence in the
uncalcined samples have been checked by Dailey4 through proton cross-

polarization. Proton cross-polarizationenhances the resonance of nuclei

 

140

 

lllllllllllllllllll'llllllllllllIIIIIIIIIIIIIT—Tm
-90 -100 —110 -120 ppm

FIGURE 48. 29S1 NMR spectra of unwashed amine-derived siloxane-
intercalated magadiites with initial TEOS: aminezmagadiite molar ratio =
100:27:1 (delay time: 600 seconds, line broadening: 140, 12 scans)

a. hexylamine-derived: HAMAGIOO

b. octylamine-derived: OAMAGlOO

c. decylamine-derived: DAMAGIOO

 

141

 

llllllllll'llllllIllIIIIITITTIIIIIIIIIIIIIIIIITW
-90 -100 -110 -120 ppm

FIGURE 49. 29Si N MR spectra of unwashed amine-derived silica-
intercalated magadiites with initial TEOS: aminezmagadiitc molar ratio =
100:27:1 (delay time: 600 seconds, line broadening: 140, 12 scans)

a. hexylamine-derived: CHAMAGIOO

b. octylamine-derived: COAMAGIOO

c. decylamine-derived: CDAMAGlOO

 

/

Illllllllllllllllllllllllllllllllllllllllllllm
-90 —100 —110 -120 ppm

FIGURE 48. 29Si N MR spectra of unwashed amine-derived siloxane-
intercalated magadiites with initial TEOS: aminezmagadiite molar ratio =
100:27:1 (delay time: 600 seconds, line broadening: 140, 12 scans)

a. hexylamine-derived: HAMAG100

b. octylamine-derived: OAMAGIOO

c. decylamine-derived: DAMAGlOO

141

 

 

WTIIIIIIIIIIIIIIIIIIYT]llll'llllllllllllllllllq

—90 -100 -110 -120 ppm
FIGURE 49. 29Si NMR spectra of unwashed amine-derived silica-
intercalated magadiites with initial TEOS: aminezmagadiite molar ratio =
100:27:1 (delay time: 600 seconds, line broadening: 140, 12 scans)
a. hexylamine-derived: CHAMAGlOO
b. octylamine-derived: COAMAGIOO
c. decylamine-derived: CDAMAGIOO

142

neighboring protons. Dailey“ proposed the presence of Q2 environments in
the uncalcined samples as ethoxys: (SiO)fii(O-CH2CH3)2 instead of silanols:
(SiO)7§i(O-H)2 in the calcined ones. The Q2 peak lack of enhancement was
then due to an absence of coupling because of the increasing number of
bonds between the two interacting atoms. Pouxviel et al.51 took the 29Si
NMR spectra of two commercial TEOS. The pure §i(O-CH2CH3)4 with no
bridging oxygen (Q0) peak appears at -82 ppm ; disilicic and trisilicic ester
ending groups (Q1) SiO_S_i(O-CH2CH3)3 peak were located at -89 ppm ;
peaks at -95.2 ppm and -96.4 ppm were assigned to cyclotetrasilicic and the
middle group (Q2) of linear trisilicic ester ; peaks at -103 ppm and -104
ppm corresponded to branched groups (Q3) of polysilicic ester ; Q4 silicon
atoms were expected to be located at -110 ppm. Thus (SiOhSi(O-CH2CH3)2
Q2 peaks for oligomer species are expected to be located at about -96 ppm.
However (SiO)7§i(O-CH2CH3)2 Q2 peaks for polymers are expected to be
shifted upfield due to a greater shielding. This shift might be the
explanation for the absence of Q2 peaks for the uncalcined samples. Even
though the Q2 silicon environments due to ethoxy groups should be more
numerous in the uncalcined samples, they might be overlapping the Q3 peak
due to silanols. The comparison of the intercalated magadiite samples
(Figures 48 and 49) and the TEOS-derived silica samples (Figures 26 and
27) is the proof for the magadiite layer retention effect during the TEOS
polymerization. Indeed, the intercalated magadiite 29Si NMR spectra are
more similar to the Na-magadiite and H-magadiite (Figure 14) ones than to
the TEOS-derived silica spectra.

143

X-ray powder diffraction

The alkylamine and TEOS:amine:magadiite molar ratio effects can
be easily followed by X-ray diffraction. The patterns of the unwashed
alkylamine-derived siloxane-r and silica-intercalated magadiites are
displayed according to the initial TEOS:amine:magadiite molar ratio in
Figures 50, 51 and 52. The basal spacing and gallery height (= basal
spacing - layer thickness = basal spacing - 11.2) are gathered in Table 34.

A summary graph is represented in Figure 53.

TABLE 34. X-ray diffraction data of unwashed
alkylamine-derived siloxane- and silica-intercalated
magadiites

 

As the X-ray diffraction patterns exhibit one 001 peak and
sometimes a 002 reflection, only one phase is present in our unwashed
alkylamine-derived siloxane- and silica-intercalated magadiites. It proves
that TEOS polymerized inside the magadiite layers. If it had polymerized
outside the layers, we would get a two-phase X-ray diffraction pattern: one
for the H-magadiite at 12.0 A and one for the TEOS-derived silica. The

144

83121.8 23 8323 "815353185. .0
303,28 Ea 803,20 6015915111138 .12
83.2.1.5 E... 832...: 69118-058515: .1
05:39:: vfiaocouqm-aoam ovarian-05850:“ 3583:: ”59..
85835 @8u3885-28xomm “overflow-0525353 06:33:: ”:2
36% u 811
Ha 05 8: «wage—5:8” E5 :53 «8:335 383885-326 was -23on
coweoiwmsaia .Bammmwem 1838 see 83856 1838 18-x .on $598

 

 

S w o 8 1 N o 2 w o 8 1 N c
_ _ _ 5 _ _ _ _ 5 q _ — _ — q _ —
<92
.1 <53 .1
<5? a <98 m
w w.
m. . .1. a.
m. <12 m.
m m
5.8 a <~§ .o
J, .0
_ _ _ F_<Np WN _ _ _ _ _ _ _ _<V_MV_ _

 

 

 

 

145

8_O<S<Qo Ea 8~O<S<Q ”358-051.5508 .0
853250 use 85.1.2110 ”315555558 .12
85.1.2195 use 88.11211: 6218885113112 .1
8:3qu cam—8825-851.“ eaten-055533 60:33:: ”Eat
smegma: 313835-0833 358-0553.? 81.83:: 52

392: u one

852: oficawafino5fiflm0mh 355 .53 853mg: nos—«835-826 new -2835
agree-05853? 3583:: no «538 839 9588th .3633 13.-x An 350E

2

con

v

a

c

 

u

u q q

<92

 

 

<9:

‘I

u

q

q _

 

 

 

51391191!!! 01111219.:

2

 

 

 

 

» e 8 e a o
q _ . q q d 4 _ -
<9: .1
<98
<3:
.1.
m.
<98 m.
<E~ n
o
_ _ _ _ _ _ _ <G_~V

 

 

146

88<2<oo as 88<2<n 62535552. .0
8NO<E<OU Ea oomo<2<o 63:865.:«380 .n
8NO<E<IU ES 8NO<E<I ungtoubEEa—axoa .a
Bacamme uSfluSBEéSm» co>toubfifis~9=¢ 3:335. ”Em:
3:338 @8532256553 co>toviEa§u=a 3:335 £2
finance“ u 2::
SEE oEvawaEHoEEflmOm—F RES 5:» “8:392: BEEoBETaom—E can -28on
uo>toc飧§fia 8593:: no 33an 833 838.5% 3952— >E-X .Nn mMDOE

@N @N
o— w o v N o OH x c v N O

— d _ _ — _ _ _ _ — q q q _ — _ _

45.2 $3

 

 

 

.6
£11me aApelaJ

Kuwaiti} ”pep:

 

 

 

 

 

 

 

 

147

 

 

 

 

 

 

 

 

50 50
42.2 A‘\, _ _ c
40- 39.0 A . ‘ c. - 40
2‘5 30-
DD
.5
=3
5
g 20- - 20
10- - 10
0 0

 

 

 

5'0 160 260 5'0 160” 260.
TEOS/magadiite molar ratio TEOS/magadiite molar ratio

FIGURE 53. Dependence of the basal spacing of the unwashed alkylamine-
derived siloxane- and silica-intercalated magadiites on TEOS:magadiite
molar ratio
left: unwashed alkylamine-derived siloxane-intercalated magadiites
rigth: unwashed alkylamine-derived silica-intercalated magadiites

a. hexylamine-derived: HAMAG and CHAMAG

b. octylamine-derived: OAMAG and COAMAG

c. decylamine-derived: DAMAG and CDAMAG

148
alkylamine effect is readily seen on the unwashed alkylamine-derived
siloxane-intercalated magadiites. As the amine chain length is increased, the
basal spacing and gallery height are increased. The basal spacing for a
given amine is independent of the TEOS:amine:magadiite molar ratio.
Whatever this ratio is, the unwashed alkylamine-derived siloxane-
intercalated magadiites display a basal spacing of about 27.8 A, 36.2 A and
42.2 A for the hexylamine-, octylamine- and decylamine-derived samples,
respectively. These dam values closely correspond to the basal spacings
found for the amine-solvated alkylammonium/alkylamine magadiites (Table
22): 30.9 A, 34.0 A and 41.3 A for hexylamine, octylamine and
decylamine, respectively. Thus the basal spacings determined for the
unwashed alkylamine-derived siloxane-intercalated magadiites reflect the
amine-solvated alkylammonium/alkylamine magadiites spacings. The slight
variations might be due to the different methods of sample preparation.
The XRD patterns for the amine-solvated alkylammonium/alkylamine
magadiites were recorded for gels spread onto a microscope glass slide
while those for the unwashed alkylamine-derived siloxane-intercalated
magadiites were recorded as oriented powders. However, upon calcination
at 450°C, two types of behavior can be determined. Except for the
octylamine- and decylamine-derived samples at the lowest TEOS:magadiite
molar ratio, the basal spacing of the unwashed alkylamine-derived silica-
intercalated magadiites is independent on the initial TEOS:magadiite molar
ratio. The basal spacing is only slightly decreased to 24.5 A, 34.9 A and
39.0 A, for hexylamine, octylamine and decylamine, respectively. This fact
demonstrates the importance of the choice of the amine for the TEOS
polymerization along the c-axis direction. Regardless of the

TEOS:magadiite molar ratio chosen, the gallery height will not be larger

149

 

FIGURE 54. Schematic representation of the role of alkylamine on the
basal spacing of unwashed alkylamine-derived siloxane- and silica—
intercalated magadiites

than the length of the alkylamine bilayer. The octylamine- and decylamine-
derived products at the initial TEOS:magadiite molar ratio of 50:1 samples
exhibit a different behavior. Upon calcination at 450°C, the basal spacing
drops to about the same 24.8 A value for the two samples. The fact that this
latter behavior only occurs for the longest amines and lowest TEOS
addition leads us to the conclude it was due to the limited polymerization of
TEOS, as shown on Figure 54. For the octylamine- and decylamine-

derived samples, the TEOS-derived silica needs to fill the gallery space

150

opened by the amine, namely 24.0 A and 30.0 A for octylamine'and
decylamine, respectively, in comparison of 15.6 A for hexylamine. As the
TEOS polymerization is an equilibrium reaction and depends on the TEOS
concentration, for the small added amount of TEOS (TEOS:magadiite
molar ratio = 50:1), polymerizing species are not available in large enough
quantity to fill that gap. Thus, as we said earlier, before calcination, the
layers are still pushed apart by the amine ; but after removal of the amine
by heat treatment, the layers are stacked onto the polymer which is shorter

than the double layer amine.

Nitrogen adsorption-desorption

X-ray diffraction analysis showed that the amine chain length has an
effect on the c-axis dimension of the TEOS polymerization inside the
magadiite galleries. However, at sufficiently large amount of TEOS, no
effect was observed on the basal spacing. TEOS stoichiometry could have
an effect on the filling of gallery space along the other two directions,
namely along the a-axis and b—axis. So, some changes should occur in the
surface area analysis. We thus performed some nitrogen adsorption-
desorption experiments. The curves for the unwashed alkylamine-derived
silica-intercalated magadiites are displayed according to the initial
TEOS:amine:magadiite molar ratio in Figures 55, 56 and 57. The surface
areas data are gathered in Table 35.

n—- A-

 

151

owO<E<QU 60>tov6§§oon .o
onO<2<OU fic>to~¥o§§oo .n .omogéo 60>tonéfifim3xo: .a
8:: ”Eur .5558_ 5:808“ «2 £2
fihmnon "one 308 Sanamafiuoqgumomk 325 .23 mofivmwafi
8339—8510:? uo>t$o§>xfi 8513:: m0 «258 noun—0mg NZ .nw mMDOE
Q— wd ed 16 N6 9w

 

 

 

 

 

 

 

 

A<v a . .5 .
S w o 4 N co 9:
. 08
Y L
. . on e
a
A
, « . 9: p m.
I... G “I. m
a . m a
a S
s u
- - O? W W)
. o 8
10%.. 4 9w (
1111 (
- .o .. com .
p . . . . . p b OWN

 

 

 

 

 

 

 

 

152

9

855230 6253525108 .0
8821218 6253552on .2 883215 6253525352 a

8:3 ”Ewt .gofiofi :OEEOmOcéOSEOmE «2 ”co—

:582 nose 32: 9223385525208. was 53 “222mg
cos—«833-852.. co>tov65=a§u=a 3:335. .«o 3E3 2.03883 «2 6n may—DOE
c." 3 ed to ad od

:2
02.4

 

 

u 1 u 1 -

On

OS

02

com

Own

8m

 

 

emm

(3/30) as GUIHIOA paqrospe

 

r

.a

d 1 d d d 1 o

 

 

 

 

 

- n b h n b -

 

wd 0.0 v.0 Nd O.

 

 

d 1 d

 

 

 

(3/90) d.LS owntm mom

 

153

885218 "Biocsessaou .o

8NO<2<OU ”60>toVoEEa—boo .n .8NO<2<=U 60>tovoEEm—2xo: .u
83; ”Emma .gofiofl .5ch81 «2 £2
fihmuoom noun.— HfloE Bmuawafinofifianmomrr 3:5 .23 3:82“qu

358285-85... ogre—c-0583}? @3335. .«o «9:3 8:883 «Z Sn EDGE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o; wd ed Yo. Nd Qw
3c 2 .
S w e v 82
a q q a o
.. con
Om . u . u - .9 u . . 00m
1 o H w o o o v o N o o
W: . a u a . 9 . q C
.9 .
a C
a. . .9 9 9.
I M 9
. w m 99.
a .
. 82 m %8
9. .9, 9
.000 OW
. 0 8m
. _ . . 8N

 

(B/oo) (us aumIOA paqxospe

154

TABLE 35. N2 adsorption data of unwashed alkylamine-derived silica-
intercalated magadiites

 

 

 

 

 

alkylamine-derived BET43 t-plot‘“
Silica-mm S total S [on] Snon-” Sua) . 811(2) . V"
magadnte mZ/g m2/g mZ/g m2/g mz/g liq cmB/ g
WmSOA
CHAMAGSO 320 370 90 270 280 0.10
COAMAGSO 560 580 100 560 480 0.20
CDAMAGSO 550 570 100 580 470 0.20
TEOS:magadiite=100:1
CHAMAG 100 440 500 130 350 370 0.12
COAMAGIOO 760 790 100 770 690 0.27
CDAMAGIOO 820 840 130 1070 710 0.38
TEGSInagadiite=20021
CHAMAGZOO 270 310 100 200 210 0.07
COAMAGZOO 540 540 100 510 440 0.18
CDAMAGZOO 570 590 1 10 530 480 0.19

 

 

 

 

 

S“: microporous surface area, 8mm-” : non-microporous surface area, V”: microporous
volume
* Sum is derived from V“ by the t-plot method ; Su(2)=8m.1 - Sm.“

The total surface areas found by the BET equation43 and t-plot“
method agree relatively well (Appendix C). The non-microporous surface
areas are about the same ~100 m2/g for all the samples, regardless the
amine used and the TEOS concentration. The high total surface areas are
due to the presence of micropores. The unwashed octylamine- and .
decylamine-derived silica-intercalated magadiites exhibit almost the same
surface areas while the hexylamine-derived samples always exhibit surface
areas half as large. The similarity of the octylamine- and decylamine-
derived surface areas has already been noted for the TEOS-derived silica
(Table 19). Assuming the silica gallery species and magadiite silicate layers
have the same density, we calculated the expected microporous surface

areas if the gallery space is completely stuffed and if the microporosity is

155

only coming from the TEOS-derived silica. The results are displayed in
Table 36.

TABLE 36. Calculated alkylamine-derived silica-intercalated
magadiite microporous surface area assuming the interlayer space is
stuffed with TEOS-derived silica

 

  
  

    

 

 

 

 

 

 

 

 

 

T19: alkylamine-derived
alkylamine-derived ratiowm = TEOS-derived silica-intercalated ma "te
silica-intercalated dom/ silica T36": T35:
magadiite gallery height spa) ‘ m2/g calculated sum ‘ mZ/g
(T able 34) (Table 19) 3,0) * mZ/g (Table 35)
TEOS :magadiite=50:1
CHAMAGSO 0.53 870 460 280
COAMAGSO 0.63 1030 650 480
CDAMAGSO 0.60 920 550 470
TEO§:magadiite=100 l
CHAMAGIOO 0.56 650 360 370
COAMAGIOO 0.68 1100 750 690
CDAMAGIOO 0.71 1110 790 710
TE5§:magadiite=200:1
200 0.54 710 380 210
COAMAGZOO 0.68 970 660 440
CDAMAG200 0.7 l 990 . 700 480
'Suzmicroporous surface area, Smn_u:non-microporous surface area, Spa) =Smm -
non-u

” T36 = ratiowmis t T19

Calculated microporous surface areas assuming a stuffed interlayer
space (T36) give a higher microporous area than the one obtained
experimentally from the t-plot method (T35). Thus, the interlayer space is
not completely stuffed with the highly microporous TEOS-derived silica.
Some holes might be present, possibly as large pores between entities that
are acting like pillars. However, a difference with the TEOS-derived silica

can be noticed: the surface areas are dependent on the initial

156

TEOS:amine:magadiite molar ratio. The total surface area reaches a
maximum for the ratio 100:27:1 and the surface areas are about the same
for the 50:27:1 and 200:27:1 ratio. The surface areas increase from the
TEOS to amine molar ratio = 50:1 to 100:1 reflects the increase in the
basal spacings. The maximum surface area values for an identical gallery
height (Table 34) for the TEOS to magadiite molar ratios = 100:1 and
200:1 can be understood in terms of different mechanisms of TEOS
polymerization. The interlayer space available for TEOS polymerization is
constant, but there are more polymerizable species, the TEOS-derived
silica formed at the 200:1 ratio might be more dense due to enhanced

TEOS polymerization on the internal surface of the pores.

CONCLUSION

The polymerization of TEOS in the galleries of layered silicic acid
affords highly microporous materials. The polymerization is done in
presence of alkylamines that swell the magadiite layers. Indeed, for TEOS
polymerization to start inside the layers, the TEOS molecules have to reach
the interlayer protons. As the diameter of TEOS is about 9 A, the
molecules can not access the H-magadiite protons because the interlayer
space is less than 1 A. When H-magadiite layers are swelled by
alkylamines, the protons are converted into ammonium cations in presence
of excess amine, and the gallery space is expanded, depending on the amine
chain length. The polymerization of TEOS can then start. However, as we
saw in the case of acid-catalyzed TEOS polymerization in the presence of
alkylamines, the amine used to swell the layers will also play a further role
in the polymerization. As ethanol washing dissolves the TEOS-derived

precursors, as well as the siloxane-intercalated magadiite, and the surface

157

areas conclusions are identical, we conclude that the magadiite layers are
intercalated by our new porous TEOS-derived silica. The calcination
temperature used to obtain silica-intercalated magadiite from the siloxane
precursor has been found to be optimized at 450°C. Indeed, below 450°C,
some amine is still present in the structure, and above 450°C, the magadiite
structure starts collapsing or rearranging. The alkyl chain length of the
amine affects the basal spacing of the final silica-intercalated magadiite.
Indeed, when enough TEOS species are present, the limiting factor for the
polymerization growth in the c-axis direction is the amine bilayer space
created between the magadiite layers. However, when too many TEOS
molecules are present, the silica-intercalated magadiite looses some of its

microporosity, probably due to a densification of the silica-intercalated

species.

CONCLUSIONS
AND
SUGGESTIONS FOR FUTURE WORK

The study of Na-magadiite, H-magadiite and the conversion of the
sodium to the proton form leads us to a better understanding of the
structural importance of the intergallery space. Indeed, for various sodium
contents Na-magadiite exhibits substantially longer range order along the
stacking axis in comparison to H-magadiite. Furthermore, the stacking
order of Na-magadiite decreases as water is removed from the gallery
upon heating. Removal of water from Na-magadiite by heat treatment is
equivalent to proton exchange with regard to effects on stacking order. In
the known makatite structure, water molecules connect consecutive layers
as part of the hydration sphere around the sodium cation and are situated at
specific locations in the gallery space. We thus assigned the loss of order in
H-magadiite to a rearrangement of the gallery space due to the absence of

water molecules. This assignment could be confirmed by a study of other

158

159
members of the hydrous sodium silicate minerals family where the
stabilizing effect of the gallery water molecules on the structure should also
be manifested as a general trend for the whole family. An additional
improvement to elucidate the Na-magdiite structure would be the synthesis
of single crystals and the X-ray crystallographic determination of the

structure.

The acid-catalyzed hydrolysis-polymerization of tetraethylortho-
silicate in the presence of alkylamines generates precursors, which, when
calcined, afford highly microporous silica. The pore size is controlled by
the length of the amine alkyl chain. Further studies with longer amines
would probably lead to pore sizes in the domain of mesoporosity. This
method could then be used to create silica with determined micropores or
mesopores sizes. More information on the final silica product could be
obtained from adsorption isotherms of molecules with known kinetic
diameters and from the study of transmission electron micrographs. A
better knowledge of the reaction would be gained by liquid 29Si NMR
studies on the supernatant and photon correlation spectroscopy to
determine particle growth. Porous acidic alumina-silica for catalytic
purposes could then be synthesized by substitution of silicon atoms by

aluminum atoms.

The polymerization of tetraethylorthosilicate into alkylamine-swollen
layered silicic «acid affords materials with large surface areas, due to the
silica-intercalated porosity. The hydrolysis is catalyzed by the interlayer
protons. Alkylamines are used to swell the layers in order for the protons

to be accessible. However, the alkylamines also play a further role in

160

designing the pore sizes and the basal spacing of the final silica-intercalated
magadiite. The optimum calcination temperature of 450°C is used to obtain
silica-intercalated magadiite from the siloxane precursor. The surface areas
exhibited a maximum value for a TEOS:magadiite molar ratio of 100:1.
This optimal surface area could be explained by a difference in the density
of the intercalated species. Density experiments and measurements of the
unreacted silicon centers left in solution by liquid 2598i NMR studies would
bring a better understanding of the silica-intercalated magadiite

compounds.

APPENDICES

161

APPENDIX A

Electron diffraction analysis”

The electron diffraction lattice dimension are given by the Bragg's

relation and the small angle approximation.

electron beam

 

sample

 

 

D

FIGURE 58. Electron diffraction analysis parameters

The Bragg's relation gives:

A

3d = sine ~ tg0 = g for 0~0

162

nlL
so, d= 2D

 

- d is the lattice dimension

- L is the camera length -> we used L=83 cm

- A is the electron wavelength -> for a 100 kV voltage, L=0.037 A

- D is the measured length on the diffractogram -> between two
consecutive points, D=0.369 cm

- n is the order -> between two consecutive points, n=l

thus, d = 4.16 A

163

APPENDIX B

Sghemr equation76

This equation is used to calculate the crystallinity of a material using
its X-ray diffraction spectrum.

k Ag“
2 1th 0050

thl =
- Dim: crystallinity or coherent scattering domain (A)
- k: constant close to unity -> if we take k=1, we are going to obtain
proportional values for the crystallinity
- Ac“: copper K0L radiation wavelength (in nm) -> 7.0, = 1.542 nm

- 1m: corrected half-height width of the sample hkl peak
1'1,“2 = 1"m2 + 1mg2
I'm: measured half-height width of the sample hkl peak
1"hk1: measured half-height width of the reference hkl peak
for the phlogopite reference, 1"th = 0.0023 radian

- 0: half the angle (in radian) at which the hkl peak occurs

164

APPENDIX C

Wits

The gas adsorption analysis experimental data is the adsorption-
desorption isotherm. Usually, the isotherm is set at 77 K, the temperature
of liquid nitrogen. It is a plot of the adsorbed volume: Vads in cc/g at STP
versus the relative pressure: P/Po (Figure 59).

- P0 is the saturating pressure of the adsorbed gas.

Vads

 

 

 

P/Po

 

FIGURE 59. Adsorption-desorption isotherm

 

165
Three kinds of pores are defined:

- micropores: pores inferior to 20 A in diameter
- mesopores: pores whose diameter is comprised between 20 A and 500 A

- macropores: pores superior to 500 A in diameter

As the sides of the micropores are close to one another, the
interaction energy is large and the adsorption will be rapid. The
microporous data can be derived from the t-plot method using the first
points of the isotherm.

The presence of mesopores can be noted if there is an hysteresis in
the region of the last points on the adsorption isotherm. Indeed, as
mesopores are larger than micropores, they are filled with liquid due to
capillarity condensation. They will thus exhibit a delay time during the
desorption that will create the hysteresis. The determination of the
mesopores is derived from the Kelvin's law.

The presence of macropores cannot be detected from gas adsorption.
The determination of the macropores, as well as the mesopores, is done by

mercury porosity.

The total surface area can be calculated by applying the BET

theory.43 The BET equation converts the isotherm adsorption curve to a

 

straight line by changing the yards to Vads (11304,) (Figure 60).

The BET theory is valid up to P/Po=0.35

, . 1 _ 1 lgll E
BET equation. Vads (P043) - C Vm + C V",l Po

 

166

P
Vads(Po-P)
A

 

0 ) P/Po
0.35

 

 

 

FIGURE 60. BET curve

The slope: givlfi and the intercept: C {In of the straight line allow for

 

the determination of C and Vm.

@1321.)

- C is an energetical constant: C = k exp R T

with E1: filling energy of the first layer of gas adsorbate
EL: liquefaction energy of the gas adsorbate
- V,n is the monolayer capacity. It is given in cm3 of adsorbate / g of
adsorbent.

The BET surface area is equal to the total surface area. It is given in m2/g.

_N_A
3351‘ = Vxn V__N28” GNZgu = Vm * 437

- NA is the Avogadro' 8 number -> 6.02 1023 molecules/mol

- V1.12,” is the volume of one mole of gas -> 22414 cm3/mol

167

- 6N2“, is the average area occupied by a molecule of nitrogen in the

completed monolayer -> 16.2"‘1020 mZ/molecule

The microporosity can be determined by using the t-plot method.“4
The t-plot is a representation of the adsorption isotherm, where the x-axis
is converted to t-values. The t-values are function of P/Po; as our surfaces
are more or less constitued of silicon oxide, we used t-values derived
experimentally by de Boer from the study of non microporous oxides. The
t-plot is valid up to P/Po=0.4 This method provides the total surface area

(Stow), the non-microporous surface (Snomu) and the microporous volume
(Vu).

 

 

 

 

 

=e.)

FIGURE 61. t-plot

The total surface area (Sum) in m2/g is calculated from the first slope on
the t-plot:

168

N
SW = lst slope at VIN—2A- 6m,” a: «11.2 = lst slope at 15.47

gas
- ¢N2 is the thickness of a single molecular layer of nitrogen molecules ->
3.54 '0' 10-10 m

The non microporous surface area (Show) in m2/g is calculated from the

second slope on the t-plot:

N
sum 11 = 2nd slope * {IN—2:; 6N2,“ * ¢N2 = 2nd slope * 15.47

The gaseous microporous volume (V11) in cm3/g is given by the intercept of
the second slope. This gaseous volume can be converted to a liquid volume
in liq cm3/g:

. _V”(gas) MN2_
Vualq) - VN2gu * PNz ‘Vp(gas) * 0.00154

- pm is the density of liquid nitrogen -> 0.81 g/cm3

- MN; is the molar mass of nitrogen gas -_> 28 g/mol

The microporous surface (Sn) in m2/g can be derived from this

microporous volume:

N
s, = Vu(gas) 9 V—L. om,” = Vu(gas) :- 4.37
N23”

LIST OF REFERENCES

169

LIST OF REFERENCES

T. J. Pinnavaia. Science, 220, 365 (1983).

M. E. Landis, C. Pochen, I. D. Johnson, G. W. Kirker, M. K. Rubin
and B. Cynwyd. US patent 4,859,648 (1989).

M. E. Landis, B. A. Aufdembrink, P. Chu, I. D. Johnson, G. W.
Kirker and M. K. Rubin. Journal of the American Chemical
Society, 113, 3189 (1991).

J. S. Dailey. PhD thesis Michigan State University (1991).

T. J. Pinnavaia, M. S. Tzou, S. D. Landau and R. H. Raythatha.
Journal of Molecular Catalysis, 27, 185 (1984).

R. Burch and C. I. Warburton. Journal of Catalysis, 97, 503 (1986).

G. Lagaly, K. Beneke and A. Weiss. The American Mineralogist,
60, 642 (1975).

10

11

12

13

14

15

l6

17

18

19

170

G. Lagaly, H.-M. Riekert and H. H. Kruse. Chemical Reactions in
Organic and Inorganic Constrained Systems, Setton edition,
Reidel Publishing Company, 361 (1986).

R. A. Fletcher and D. M. Bibby. Clays and Clay Minerals, 35(4),
318 (1987).

G. Lagaly, K. Beneke and A. Weiss. Proceedings of the International
Clay Conference, Madrid, 663 (1972).

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

Z. Johan and G. F. Maglione. Bulletin de la Socie’te’fiancaise de
Mine’ralogie et de Cristallographie, 95, 371 (1972).

K. Beneke and G. Lagaly. The American Mineralogist, 62, 763
(1977).

T. J. Pinnavaia, I. D. Johnson and M. Lipsicas. Journal of Solid State
Chemistry, 63, 118 (1986).

R. A. Sheppard and A. J. Gude. The American Mineralogist, 55, 358
(1970).

H. Annehed, L. Filth and F. J. Lincoln. Zeitschrifi fur
Kristallogaphie, 159, 203 (1982).

H. P. Eugster. Science, 157, 1177 (1967).

W. Schwieger, D. Heidemann and K.-H. Bergk. Revue de Chimie
Minérale, 22, 639 (1985).

L. McCulloch. Journal of the American Chemical Society, 74, 2453
(1952).

20

21

22

23

24

25

26

27

28

29

30

171

V. G. Il’in, L. F. Kirichenko, Y. K. Ivanov, V. Y. Kondrachuk and
Z. Z. Vysotskii. Doklady Physical Chemistry, USSR. English
Translation, 174, 408 (1967)

I. M. Tirnishenkov, Y. P. Menshikov, L. F. Gannibal and I. V.
Bussen. Zapiski Vsesoyuznogo Mineralogicheskogo
Obschchestva, 104, 317 (1975). (abstract in M. Fleischer, The
American Mineralogist, 61, 339 (1976).

J. M. Garces, S. C. Rocke, C. E. Crowder and D. L. Hasha. Clays
and Clay Minerals, 36(5), 409 (1988).

G.Borbély, H. K. Beyer, H. G. Karge, W. Schwieger, A. Brandt and
K.-H. Bergk. Clays and Clay Minerals, 39(5), 490 (1991).

M.-T. Le Bihan, A. Kalt and R. Wey. Bulletin de la Sociéte’
F rancaise de Minéralogie et de Cristallographie, 94, 15
(1971 ).

K. Beneke and G. Lagaly. The American Mineralogist, 74, 224
(1989).

A. J. Gude and R. A. Sheppard. The American Mineralogist, 57,
1053 (1972).

R. K. Iler. Journal of Colloidal Science, 19, 648 (1964).

G. Lagaly, K. Beneke and A. Weiss. The American Mineralogist,
60, 650 (1975).

G. Lagaly. Advances in Colloid and Interface Science, 11, 105
(1979).

J. M. Rojo, J. Sanz, E. Ruitz~Hitzky and J. M. Serratosa. Zeitschrift
fuer Anorganische und Allgemeine Chemie, 540-541, 227
(1986).

31

32

33

34

35

36

37

38

39

4O

41

42

172

T. Yanagisawa, K. Kuroda and C. Kato. Bulletin of Chemical Society
of Japan, 61(10), 3743 (1988).

K.-H. Bergk, C. Schuetz and W. Schwieger. Silikatlechnik, 41(7),
241 (1990).

J. M. Rojo, E. Ruiz-Hitzky, J. Sanz and J. M. Serratosa. Revue de
Chimie Minérale, 20, 807 (1983).

J. M. Rojo, E. Ruiz-Hitzky and J. Sanz. Inorganic Chemistry, 27,
f 2785 (1988).

J. B. Peri. The Journal of Physical Chemistry, 70(9), 2937 (1966).

J. L. McAtee, R. House and H. P. Eugster. The American
Mineralogist, 53, 2061 (1968).

G. Scholzen, K. Beneke and G. Lagaly. Zeitschriftfiter
Anorganische und Allgemeine Chemie, 597, 183 (1991).

G. W. Brindley. The American Mineralogist, 54, 1583 (1969).
Powder diffraction file. Inorganic volume, set ?, JCPDS

B. A. Morrow and A. J. McFarlan. The Journal of Physical
Chemistry, 96, 1395 (1992).

E. Lippma, M. Magi, A. Samoson, G. Engelhardt and A.-R.
Grummer. Journal of the American Chemical Society, 102,
4889 (1980).

T. Yanagisawa, T. Shimizu, K. Kuroda and C. Kato. Bulletin of
Chemical Society of Japan, 63(4), 988 (1990).

43

45

46

47

48

49

50

51

52

53

54

173

S. Brunauer, P. H. Emmet and E. Teller. Journal of the American
Chemical Society, 60, 309 (1938).

J. H. de Boer, B. c. Lippens, B. G. Linsen, J. c. P. Broekhoff, A.
Van den Heuvel and T. J. Osinga. Journal of Colloid and
Interface Science, 21, 405 (1966).

T. J. Pinnavaia, M. S. Tzou, S. D. Landau and R. H. Raythatha.
Journal of Molecular Catalysis, 27, 185 (1984).

R. Burch and C. I. Warburton. Journal of Catalysis, 97, 503 (1986).
L. C. Klein. Annual review of Materials Science, 15, 227 (1985).

I. Strawbridge, A. F. Craievich and P. F. James. Journal of Non-
Crystalline Solids, 72, 139 (1985).

J. B. Blurn and J. W. Ryan. Journal ofNon-Crystalline Solids, 81,
221 (1986).

K. C. Chen, T. Tsuchiya and J. D. Mackenzie. Journal of Non-
Crystalline Solids, 81, 227 (1986).

J. C. Pouxviel, J. P. Boilot, J. C. Beloeil and J. Y. Lallemand.
Journal ofNon-Crystalline Solids, 89, 345 (1987).

C. W. Turner and K. J. Franklin. Journal of Non-Crystalline Solids,
91, 402 (1987).

K. Karnaiya, T. Yoko and H. Suzuki. Journal of Non-Crystalline
Solids, 93, 407 (1987).

J. F. Quinson, N. Tchipkam, J. Dumas, C. Bovier and J. Serughetti.
Journal of Non-Crystalline Solids, 100, 231 (1988).

 

55

56

57

58

59

60

61

62

63

65

66

174
A. H. Boonstra and T. N. M. Bernards. Journal of Non-Crystalline
Solids, 105, 207 (1988).

J. C. R0 and I. J. Chung. Journal of Non-Crystalline Solids, 110, 26
(1989).

T. Kawaguchi and K. Ono. Journal of Non-Crystalline Solids, 121,
383 (1990).

M. T. Haris, R. R. Brunson and C. H. Byers. Journal of Non-
Crystalline Solids, 121, 397 (1990).

A. H. Boonstra andJ. M. E. Baken. Journal of Non-Crystalline
Solids, 122,171 (1990).

M. J. van Bommel, T. N. M. Bernards and A. H. Boonstra. Journal
of Non-Crystalline Solids, 128, 231 (1991).

H. A. Szirnanski and R. E. Erickson. Infrared band handbook, 2nd
edition, Plenum Press, London, (1970).

A. Stock and C. Somieski. Berichte der Deutschen Chernischen
Gesellschaft, 54, 740 (1921).

A. Stock. Hydrides of Boron and Silicon, Cornell University Press,
Ithaca, N.Y., 30 (1933).

R. L. Wells and R. Schaeffer. Journal of the American Chemical
Society, 88, 37 (1966).

H. Burger and W. Sawodny. Spectrochimica Acta, 23(A), 2827
(1967).

W. D. Harkins and G. Jura. Journal of the American Chemical
Society, 66, 1362 (1944).

67

68

69

70

71

72

73

74

75

76

175

C. J. Blinker, K. D. Keefer, R. A. Assink, B. D. Kay and C. S.
Ashley. Journal of Non-Crystalline Solids, 63, 45 (1984).

E. Ruiz-Hitzky and J. M. Rojo. Nature, 287, 28 (1980).

E. Ruiz-Hitzky, J. M. Rojo and G. Lagaly. Colloid and Polymer
Science, 263, 1025 (1985).

R. Sprung, M. E. Davis, J. S. Kauffman and C. Dyboski. Industrial
and Engineering Chemistry Research, 29, 213 (1990).

T. Yanagisawa, C. Yokoyarna, K. Kuroda and C. Kato. Bulletin of
Chemical Society of Japan, 63(1), 47 (1990).

Y. Sugahara, K. Sugilnoto, T. Yanagisawa, Y. Nomizu, K. Kuroda
and C. Kato. Yogyo Kyokai Shi, 95, 117 (1987).

T. Yanagisawa, T. Shimizu, K. Kuroda and C. Kato. Bulletin of
Chemical Society of Japan, 63(5), 1535 (1990).

a. J. S. Beck. US patent 5,057,296 (1991).

b. C. T. Kresge, M. E. Leonowicz, W. J. Roth and J. C. Vartuli. US
patent 5,098,684 (1992).

c. C. T. Kresge, M. E. Leonowicz, W. J. Roth and J. C. Vartuli. US
patent 5,102,643 (1992).

J. R. Fryer. The chemical application of transmission electron
microswpy, Academic Press (1979).

Scherrer equation

fi _\\

 

 

STQ E UNIV ,

     

l H!)

H‘
93007929064