. . .MI; v-

-,.v~u.
h". u ‘ ’n‘:.:4 l— I...
"-.fTL‘: § \ ‘ I -

I ,. ‘u:.)fiu.u-n'x..

Li7:';ft5‘ ‘52" I A 1‘ I54;
3"” -. ‘ “ 2‘.‘\:.‘I.,\=i,s.;
“21‘1":

:Nu \l

rid-mm a. .357 . ,‘ ;.::
'1' ‘ ‘ “' "'1 I: ‘n’ I h“ ~5I’L‘.‘: W, 2).C.I \ a:
'1 II' .a, “Inf“. "Hard-m1“ f“) «, . . I . 9 .59 L‘
I x ‘ . -. I. . ', ' " ' ’ ““
§-~I;~':_‘ 1-! , . I ~ _ : JM ‘ .1. (213311?) ‘ J’ " j, v . .: ‘; ”8%. ‘ . ‘7“ “'27..
‘ , . . . A . u I
35739:»

mm In:
. -..
n In ~L~ v.

I

 

Rm

x
-IIINJI‘

' a
. .quI I .
,, ‘b'flp.

3"‘M
.«u,
luv" (inn-61

‘ r v.“

.

’33:“;

I. w...
Mr.

W; 55“?
“314* .1 2:1. I’"
. 3" ‘ fit-r

“val i..:
u. *3
“"2’I:£"~~.uu

JKE'I
“In

\lq‘f‘g 3:3»
‘ 1 0 1‘ “may;

Iv.“

Mufti-‘31.” H“
I

(M

I‘ : -.
J 4
ARI.» . I v’.;u:;5-','.' duo-I1
I
1“,.” . v'qllfo‘vl ..
IL I

‘.’ ‘ - ‘ .‘ I ‘. I r,‘ “ ... ’l,‘ :M,.,:/',:‘
“amt .-‘ ,, . I . : ..,. . ,-.I.'I 7. ”4ng
;Il ~ A ; -

 

LBI BRARIE

IllllllillllllllllllmlllIllililHlll‘IUlllHlU |

 

 

 

 

HUI

 

 

3 1293 00885
.:.'l"."-'..'" \g‘ 2. 3" .
._, -- - This IS to certify that the _.
---- -. dissertationentitled -

SURFACE—WEIEQ-EIMQGQLLfll-ér- COEFLING
REACTIONS AND NalMONTMORILLONITE INTERCALATION
PRODUCTS

presented by

Leighta Maureen Johnson

has been accepted towards fulfillment
of the requirements for

Ph.D. Chemistry

degree in

 

 

(I’m recraoIM/Mocw
fljor p‘gI’asor
Date—C: 40% y} /'77/

MS U is an Affirmative Action/Equal Opportunity Institution 042771

 

PLACE OI RETURN BOX to remove this checkoutfl'om your record.
To AVOID FINES return on or before date duo.

DATE DUE DATE DUE DATE DUE

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative ActiorVEqud Opportunity IMIIUIOI'I

7..
I.

SURFACE-MODIFIED IMOGOLITE: COUPLING REACTIONS AND
Na-MONTMORILLONITE INTERCALATION PRODUCTS

BY

Leighta Maureen Johnson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1991

ABSTRACT

SURFACE-MODIFIED IMOGOLITE: COUPLING REACTIONS AND
Na-MONTMORILLONITE INTERCALATION PRODUCTS

BY

Leighta Maureen Johnson

The tubular aluminosilicate, imogolite, having a
diameter of 23 A, a length of 2000 to 3000 A and an
internal channel diameter of #8.? A was the material of
interest in this study. Imogolite intercalates as a
regular monolayer between the layers of Na-montmorillonite
clay. The intercalated tubes align in van der Waals
contact with one another and the clay layer. Therefore,
the microporosity of this tubular silicate layered
silicate, TSLS complex was restricted to the internal tube
channels.

The focus of this work was concerted catalyst design.
Modification of the external surfaces of imogolite was
investigated as a possible route for enhancement of the
microporosity of the TSLS complex. Surface modification
was envisioned to promote lateral separation between the
Na-montmorillonite intercalated surface-modified tubes.
Removal of the surface-modifying agent by calcination might
yield highly microporous products.

The internal and external surfaces of imogolite were
silylated by allowing gamma-aminopropyltriethoxysilane

(APS) to react with imogolite suspensions in acidic

solution. The hydrolytic lability of the APS-modified
imogolite was examined by monitoring the depletion of
siloxane polymer from the tubular surfaces by dialysis
against water at ambient pH. Bimodal depletion behavior
was observed for APS-modified imogolite. This behavior was
attributed to differences in the hydrolytic lability for
siloxane polymer bound at the internal SiOH surfaces and
the external Alon surfaces. Covalent bonding between APS
and imogolite 8101-! sites was detected in the 1H decoupled
2951 MAS NMR spectrum of the APS-modified material.
Polymer hydrolysis prevented the formation of ordered APS-
modified imogolite intercalated products.
Gamma-aminopropylphosphonic acid (APP) was found both
to chemisorb to the external Ala-120)+ imogolite surface and
to intercalate within the tube channel. Highly ordered
washed products were obtained from APP-modified imogolite
intercalated Na-montmorillonite at neutral pH. 29Si and 1H
decoupled 31F MAS NMR, FTIR, x-ray diffraction, elemental
analysis and adsorption. measurements were used to
characterize the material. The products were identified as
a quadruple layer of close-packed imogolite tubes
intercalated between the layers of Na-montmorillonite.
Reactions of APP with Na-montmorillonite yielded ordered
intersalate compounds of APP or APP/APP' depending upon the

reaction pH.

To my family and Phil

iv

ACKNOWLEDGMENTS

I would like to thank Dr. T.J. Pinnavaia for his
guidance and financial support' throughout the course of
this work. I truly respect his skill as a teacher. Now I
can appreciate many of the things he taught me, although I
didn’t know I was learning them at the time. The outcome
of the chemistry was unpredictable, but I could always rely
on his special ability to provide an alternate point of
viewu I hope to continue to have the opportunity to
exchange ideas in the future.

I am very grateful to Dr. Rick for his efforts as the
second reader of this manuscript. I appreciate his
commitment and cooperation under the time constraints
involved.

I would also like to express my gratitude to my
husband, for his efforts to allow me to concentrate more
fully on my work.

I would like to thank Kermit Johnson and Dr.
Yesinowski for their technical assistance and consultation.
I would also like to thank the members of the Pinnavaia
group for their helpful discussion and support.

Finally, I would like to thank Ahmad Moini and Kirk

Schmitt of Mobil Research and Development Corporation for

V

providing critical 3113 NMR spectra when the equipment at

MSU was being repaired.

vi

LIST OF
LIST OF
Chapter
A.
B.
C.
D.
E.
F.
G.
H.
I.

Chapter
Polymer

TABLE OF CONTENTS

 

 

 

 

 

TABLES
FIGURES
I Introduction
Historical Significance
Imogolite
Na-Montmorillonite
Tubular Silicate Layered Silicate, TSLS
Research Objectives
APS-modified Imogolite/Na-montmorillonite____
Phosphorus Acid Coupled Alumina Surfaces
APP-modified Imogolite/Na-montmorillonite____
List of References
II Silylation of a Tubular Aluminosilicate
(Imogolite), by Reaction with Hydrolyzed gamma-

aminopropyltriethoxysilane (APS).

A.
B.
C.
D.
E.

Chapter

A.

B.

Introduction

 

Experimental

 

Results

 

Discussion

 

List of References

 

Page

ix

10

11

15

17
19
20
31

36

III Hydrolytic Lability of silylated Imogolite,
and Intercalation of the silylated Tubular, Aluminosilicate

in Montmorillonite.

Introduction

 

Experimental

 

vii

39

40

EQQE

 

 

 

C. Results 44
D. Discussion 61
E. List of References 67
Chapter IV Reaction of Gamma-aminopropylphosphonic acid

with Imogolite and Na-montmorillonite.

 

 

 

 

A. Introduction 69
B. Experimental . 71
C. Results and Discussion 73
D. Conclusions 103
E. List of References 105

 

viii

LIST OF TABLES

 

 

Table Page
II.I. FTIR vibrational frequencies (cm-1) for
hydrolyzed APS and hydrolyzed APS-imogolite reaction
products 28
111.1. 29Si chemical shifts for silicon environments

in pristine APS polymer, APS-modified Imogolite and APS-
modified gamma-alumina 57
IV.I. Acid dissociation constants and 31F chemical
shifts for gamma-aminopropylphosphonic acid species in
aqueous solution and isolated in the solid state 75
IV.II. Selected IR Frequencies (cm'l) for solid APP
recovered from solution at different pH values 76

 

IV.III. Selected IR Frequencies (cm-1) (KBr pellet)

 

 

 

for APP-modified systems 82
IV.IV. 1H decoupled 3lP MAS NMR resonances observed

for APP-modified systems 83
IV.V.. Elemental Analysis results for the APP-coupled
Na-montmorillonite and APP-coupled imogolie intercalated
montmorillonite samples 92
IV.VI. Ratio of the areas obtained by deconvolution

analysis of 298i NMR resonances at -79 ppm (imogolite)
and -93 ppm (Na-montmorillonite), and the corresponding
imogolite:c1ay wt. ratios 100

ix

LIST OF FIGURES

 

Figure Rage
Figure 1.1. Idealized structural representation of the
cross section of the imogolite tube. 2
Figure 1.2. Idealized structure of Na-montmorillonite.

Silicon atoms occupy the tetrahedral positions in the
oxygen framework, aluminum or magnesium atoms occupy

the octahedral sites. Mn+ represents the interlayer Na
cation. 6

Figure 1.3. Schematic representation of TSLS. A 8

Figure II.1. Cross-sectional view of the tubular
aluminosilicate imogolite (adapted from reference 22).

The inner and outer tube diameters are ~ 8.2 A and 23 A,
respectively. Tube lengths of the synthetic derivative
used in the present work are typically from 250 to

350 nm. The empirical formula, as read from outer to

inner atomic planes is (HO)3A1203Si(OH). 18

Figure 11.2. FTIR spectra (KBr disks) of (A) APS polymer
prepared by hydrolysis and drying in air; (B) APS polymer
prepared by hydrolysis in acetic acid at pH 3.6 and air
drying; and (C) synthetic imogolite. 22

Figure 11.3. FTIR spectra (KBr disks) of silylated
imogolite reaction products formed at formal APS : Al
molar ratios of (A) 0.5 : 1, (B) l : l, (C) 2 : l, (D)

2.5 : 1. 24

Figure 11.4. FTIR difference spectra for the region
between 800 cm'1 and 2000 cm'1 obtained by digital
subtraction of the imogolite spectrum in Figure II.2C

from the spectra for the APS-imogolite reaction products

in Figure 11.3. 26

Figure 11.5. 2981 MAS NMR spectra: (A) the pure

siloxane polymer obtained by hydrolysis in aqueous

acetic acid at pH 3.6; (B) the pure siloxane polymer
obtained by hydrolysis in water: (C) the acid-

hydrolyzed APS-imogolite reaction product obtained

at a formal APS : Al ratio of 0.5 : 1.0. 29

Figure Bege
Figure III.1. FTIR absorbance spectra (KBr pellets)
of the substrates (A) imogolite and (B) gamma-alumina._ 47

Figure III.2. FTIR absorbance spectra (KBr pellets)
of APS-modified imogolite samples isolated after dialysis
for the time intervals indicated. 49

 

Figure III.3. FTIR absorbance spectra (KBr pellets) of
APS-modified gamma-alumina samples isolated after
dialysis for the time intervals indicated. 50

Figure III.4. Plot of log(wt. % APS) versus dialysis
time for the hydrolyzed APS polymer, APS-modified
imogolite and APS-modified gamma-alumina. 53

Figure 111.5. 13 decoupled 2931 MAS NMR (79.459unz)
spectra of (A) 20 wt. % APS-modified imogolite and

(B) 24 wt. % APS-modified gamma-alumina samples using

a 3 us pulse width, a relaxation delay time of 20 s,

and a sample spinning rate of 5 kHz. 55

Figure III.6. X-ray diffraction patterns of oriented

films of (A) Na-montmorillonite, (B) 57 wt. % APS-
modified Na-montmorillonite, (C) sample in (B) washed

once with H20, and (D) sample in (B) washed twice

with H20. 59

Figure III.7. X-ray diffraction patterns of oriented

films of (A) 7 wt. % APS-modified imogolite suspension
combined with a Na-montmorillonite suspension in a wt.
ratio of 2.5:1 imogolite to clay, (B) sample in

(A) washed once with water and (C) sample in (A) washed
twice with water. 60

Figure III.8. Schematic representation of a cross
sectional view of the imogolite structure. The tube

walls are composed of five tiers of atoms. Reading

from the inner to the outer wall, the tier compositions
are H0, Si, 0, Al, and 0H, respectively. The diameter

of the channel, 8.7 A, is sufficiently large to

accomodate APS as a silylating agent. 64

Figure IV.1 1H decoupled 31F MAS NMR spectrum'

(161.903 MHz) of solid gamma-aminopropylphosphonic

acid as received, obtained with a pulse width of 4.8

usec, a relaxation delay time of 60 seconds and a

sample spinning rate of 4KHz. 80

Figure IV.2 x-ray diffraction patterns of (A) APPMONTA,
(B) APPMONTA washed once with water and (C) APPMONTA
washed twice with water. 87

xi

figure Pege
Figure IV.3 X-ray diffraction patterns of (A) APPMONTB,

(B) APPMONTB washed once with water and (C) APPMONTB
washed twice with water. 89

Figure IV.4 FTIR difference spectra (KBr Pellets) of
(A) APPMONTA and (B) APPMONTB. 9O

 

Figure IV.S X-ray diffraction patterns of (A) APPIIMSS
and (B) APPIIMSS after heating at 130°C in air for 48 h. 95

Figure IV.6 x-ray diffraction pattern of APPIIM75. 97

Figure IV.7 1H decoupled 31P MAS NMR spectrum

(145.76 MHz) of APPIIM75, obtained with a 3 usec

pulse width and a relaxation delay time of 62 sec.
(Spectrum courtesy of Kirk Schmitt, Mobil Research

and Development Corporation). 98

xii

CHAPTER I
Introduction
A. HISTORICAL SIGNIFICANCE

A wide variety of applications are known to utilize
the unique surface chemistry of aluminas. The
effectiveness of aluminas as adsorbant materials has led to
the application of these oxides in chromatographic
separations processes.1 The thermal stability and high
surface area of aluminum oxides has prompted their use as
catalysts and catalyst supports.2r3 Surface modification
of aluminas has improved their performance in each of these
applications.“'6 The work described here incorporates the
use of coupling agents for surface modification, in order
to’ develop superior' microporous materials for catalysis

applications.

B. IMOGOLITE

The external aluminum hydroxide surfaces of the
tubular aluminosilicate, imogolite, were selected for
surface modification in this study. Figure 1.1 shows a
schematic representation of a cross section of the
imogolite tube. The tube diameter is ~23 A, and the

internal cavity diameter is 8.7 A.7 The length of the

Figure I.3.

 

Schematic representation of TSLS.

00
.0
oo
.0

Si
Al '

OH

3

imogolite tubes, as synthesized by the method described by
Farmer and Frasers, was determined by electron microscopy.
The tube lengths ranged between 2000 and 3000 A. Imogolite
has also been characterized by FTIR and 29Si MAS NMR
spectroscopy.9rlo

Certain physical and structural properties of
imogolite make it attractive as a potential catalyst
material. Imogolite is a very high surface area material
due to the internal cavity and small particle size. The
hydroxylated nature of both the internal cavity and the
external surfaces of the tube, give rise to water-clear
aqueous imogolite suspensions under conditions of reduced
pH and concentrations <0.4 wt. %. Particle aggregation is
initiated at elevated concentration and pH, resulting in
gelation of the imogolite sol particles. A more complete
review of the imogolite literature has been presented in
previous work.11

~The adsorption properties of imogolite has been
studied by several different groups. Theng et 31.12
investigated the pH dependent adsorption of Na and Cl ions
on natural imogolite. Natural imogolite was found to
possess the highest positive surface charge at pH 4.
Chloride adsorption occurs at Al-OH—Al sites on the
external surfaces of the tubes, which become protonated at
low pH. A chloride adsorption capacity of 3S meq/100 g was
measured at pH 4. The imogolite surface charge was

estimated at one positive charge per 700 A2, or one

4

positive charge per 40 aluminum. Phosphate adsorption onto
imogolite was also determined. The capacity of imogolite
to adsorb phosphate was only 12 meq/100 9, much lower than
that of chloride. Phosphate was presumed to adsorb to
defect sites or Al(0H)H20 sites at the tube edges more
readily than external Al-OH-Al sites, resulting in a
reduced adsorption capacity.

Clark and McBride13 also measured chloride adsorption
as a function of pH. Results obtained from positive
surface charge measurements on imogolite using NaCl
differed from those obtained using‘ NaClO4. Adsorption
capacities obtained from measurements of negative surface
charge were independent of adsorbate. Anion dependent
intercalation of the neutral salts was proposed to explain
the anomalous results. In a separate paper, Clark and
McBride14 reported obtaining anomalous results from
experiments involving the adsorption of Cu(II) and Co( II)
onto imogolite. The amount of Cu(II) adsorbed as a
function of the concentration of Cu cations in solution,
approached a plateau at a much lower value than the
corresponding Co(II) adsorption curve. Again, it was
necessary to consider intercalation of Cu(No3)2 ion pairs
to explain the anomaly. The solution stability of the
cupric nitrate ion pairs is much greater than the stability
of the corresponding cobalt nitrate ion pairs in solution.
Therefore, intercalation effects were only observed for Cu

ion adsorption. The potential for imogolite to intercalate

5

neutral ion pairs should be recognized when imogolite

interacts with charged species.

C. Na-MONTMORILLONITE

Na-montmorillonite clay was used to improve the
performance of imogolite as a catalyst material, and to aid
the characterization of surface modified imogolite
products. The layered structure of Na-montmorillonite
coupled with its capacity to undergo ion exchange makes
this clay mineral ideal for these applications. The
idealized structure of Na-montmorillonite is shown: in
Figure I.2. The idealized chemical formula for Na-
montmorillonite is Mn+x/n°yH20[Al4_ngx](Si8)020(OH)4. The
layer is composed of four planes of oxygens which define
upper and lower sheets of tetrahedral sites and a central
sheet of octahedral sites. This 2:1 ratio of tetrahedral
to octahedral sheets classifies Na—montmorillonite as a 2:1
phyllosilicate. Pinnavaia15 provides an excellent review
of the‘ structure, properties and classifications of clay
minerals. The cation exchange properties of Na-
montmorillonite clay arise from the presence of hydrated
exchangeable cations between the clay layers. These
interlayer cations balance the net negative charge on the
clay layer, originating from the substitution of octahedral
Al3+ by 1192+. Imogolite and surface modified imogolite,
having a net surface positive charge, can be intercalated

between the layers of Na-montmorillonite through the

 

 

 

 

 

 

 

 

Figure I.2. Idealized structure of Ila-montmorillonite.
Silicon atoms occupy the tetrahedral positions in the
oxygen framework, aluminum or magnesium atoms occupy the
octahedral sites. Mn+ represents the interlayer Na cation.

7

replacement of the hydrated Na cations. The identification
of surface-modified imogolite intercalated montmorillonite
reaction products may aid in the characterization of

surface-modified imogolite reaction products.

D. TUBULAR SILICATE LAYERED SILICATE, TSLS

The utility of imogolite asa catalyst material was
compromised, by an unfavorably low' dehydroxylation onset
temperature of 350°C. The resistance of the tubular
structure to dehydroxylation has been improved by the
intercalation of the individual imogolite tubes as‘ a
regular monolayer between the layers of Na-montmorillonite
clay.16 Dehydroxylation of the tubular silicate layered
silicate, or TSLS complex, is initiated at >400°C. The
TSLS material consists of aligned rows of imogolite tubes
in van der Waals contact with adjacent tubes along the
length of the interlayer and with the clay layer in the
direction perpendicular to the interlayer. This is
illustrated schematically in Figure I.3. The microporosity
observed for ‘this ‘material is therefore limited. to the
channels within the tubes, and the pores formed by packing

the cylinders against planar sheets.17

E. RESEARCH OBJECTIVES
The goal of this work was catalyst design; increasing
the microporosity of the pillared clay catalyst material,

TSLS, through surface modification of the imogolite pillar

 

Schematic representation of TSLS.

Figure I.3.

9

prior to intercalation. Controlled absorption of a
coupling agent to the external surfaces of imogolite as a
regular monolayer, was envisioned to introduce bulk between
the individual tubular units, forcing lateral separation
between the tubes in a surface-modified-imogolite
intercalated Na-montmorillonite product. Removal of the
surface ‘modifying agent. by calcination, might introduce
permanent porosity between the imogolite tubes.

The absence of van der Waals contact between
neighboring tubes in the surface-modified-imogolite
intercalated material may result in reduced dehydroxylation
onset temperatures for the new material. A superior
catalyst material might still be isolated from the high
temperature decomposition products of imogolite. The
decomposition :mechanism. proposed. by' MacKenzie et al.,18
suggests that the individual imogolite tubes would undergo
cleavage of the oxygen linkages along their midsection,
followed by subsequent flattening and condensation of the
two fragments. Decomposition of laterally separated, Na-
montmorillonite intercalated imogolite tubes in this way,
might result in a layered aluminosilicate pillared smectite
clay; Consequently, the calcined product of laterally
separated imogolite within the galleries of Na-
montmorillonite might lead to the formation of an entirely

new microporous material.

10
F. APS-MODIFIED IMOGOLITE/Na-MONTMORILLONITE

Surface modification of imogolite was first attempted
using a 2 wt. % solution of the coupling agent gamma-
aminopropyltriethoxysilane (APS). The behavior of this
coupling agent in solution and its interaction with
hydroxylated surfaces has been previously described.11
Several criterion must be satisfied in order for a surface
modified imogolite to perform as envisioned. Most
importantly, a net. positive charge 'must be retained to
avoid aggregation of the tubes. This requires a reaction
pH no greater than 7 and an imogolite concentration
<0.4 wt. %. APS was attractive as a potential surface
modifying agent because of its ability to uniformly couple
to hydroxylated surfaces. This behavior is promoted by APS
hydrolysis at concentrations of 2 wt. % at pH 4 in aqueous
solution because the hydrolyzed APS exists in chiefly
monomeric form (as the trisilanol). Under these
conditions, the aminofunctionality is protonated, promoting
coupling' between the silanol groups and the positively
charged imogolite surface. APS coupled to imogolite in
this way would maintain a positive surface charge, enabling
intercalation of the surface coupled. products into the
galleries of the Na-montmorillonite clay. .

Gamma-ammoniumpropylsiloxane polymer coupling to the
imogolite surface was confirmed by 29Si MAS NMR and Fourier
Transform Infrared spectroscopy (FTIR) from air-dried

products obtained by treating 0.1 wt. % imogolite

ll
suspensions with 2 wt. % APS solutions hydrolyzed at pH 3.6

in the presence of acetic acid.19 The hydrolytic
instability of the APS—modified imogolite products was
subsequently demonstrated by depletion of the
organosiloxane polymer from suspensions of APS-modified
imogolite during dialysis against deionized water. The APS
depletion rate was determined by monitoring the APS
concentration of the air-dried products by FTIR at
intervals throughout dialysis.

The isolation of APS-modified imogolite intercalation
products might have been possible despite the somewhat
labile nature of the imogolite-polymer interface.
Intercalation of the imogolite tubes was known to be rapid,
and the interaction between the clay layers and the APS-
modified imogolite could have had a stabilizing effect on
the complex. Attempts to intercalate APS-modified
imogolite between the layers of Na-montmorillonite clay
were unsuccessful. The reaction products obtained
reflected the dynamic nature of the APS-coupled imogolite

material.

G. PHOSPHORUS ACID COUPLED ALUMINA SURFACES

Phosphonic acids were selected as surface modifying
agents in an effort to obtain a surface modified imogolite
product with increased stability. Phosphonic acids have
been successfully coupled to alumina surfaces from aqueous

solution. The adsorption of phosphorus acids on alumina

12

has shown to provide increased resistance to corrosion.
Surface coupling between phosphorus acids and alumina has
been. characterized by Infrared (IR), Inelastic Electron
Tunneling (IET) and Auger Electron spectroscopy.

Ramsier et al.20 studied the adsorption from aqueous
solution of several phosphorus acids [phosphonic acid (PA),
phosphinic acid (PPA), hydroxymethyl phosphonic acid (HMPA)
and nitrilotris(methylene)-triphosphonic acid (NTMP),
(N[CH2P(O)(OH)2]3)] on alumina by IR and IETS. The absence
of P=0 bands in the spectra of the adsorbed species led to
the conclusion that these acids were chemisorbed to : the
alumina surface through resonance stabilized structures of
a symmetric phosphonate (phosphinate) group. The presence
of P-H bands in the spectrum of the adsorbed NTMP, however,
suggested that this acid dissociates upon adsorption by P-C
and C-N bond cleavage. The PA and HMPA decomposition
byproducts of the NTMP were adsorbed to the alumina surface
while the amine and alcohol decomposition byproducts were
evolved as gases.

Templeton and Weinberg”!22 have identified
temperature dependent molecular and dissociative adsorption
pathways in gas phase adsorption studies of phosphonate
esters on alumina. Dimethyl methylphosphonate
(CH30)2(CH3)P(O), DMMP, dissociates through P-0 bond
cleavage to fOrm the methyl methylphosphonate species which
adsorbs to the alumina surface through its two remaining

P-O units.

13
White et al.23 and Davis et al.24 addressed the

relationship between the ability of phosphorus acids to
adsorb to alumina surfaces and the corrosion resistance
performance of the phosphorus acid-coupled material. White
et a1. utilized IETS, FTIR and molecular electrostatic
potential calculations to determine the relative strengths
of adsorption of PA, HMP, MPA and NTMP on alumina. They
were able to correlate the best hydration inhibitor, NTMP,
with the strongest adsorption and the poorest hydration
inhibitor, MPA, with the weakest adsorption. Davis
et al.24 examined the IET spectra of NTMP adsorbed Onto
aluminum coupons from aqueous solution. The NTMP
adsorption occurred by the stepwise displacement of water
or hydroxyl groups on the surface by the phosphonate 'legs’
of the NTMP. Davis et al. also monitored the surface of
the NTMP-coupled aluminum coupons during exposure to
humidity in an effort to understand the corrosion process.
Corrosion was also found to be a stepwise process,
initiated by the slow dissolution of the NTMP from the
surface and followed by the rapid hydration/corrosion of

the freshly exposed aluminum.

H. APP-MODIFIED IMOGOLITE/Na-HONTHORILLONITE
Gamma-aminopropylphosphonic acid (APP), the phosphonic

acid analog to APS, was chosen for its ability to

preferentially orient itself with respect to the imogolite

surface to maintain a positive surface charge in the pH

14

range of interest. The species exists as a neutral
zwitterion +NH3(CH2)3PO3H‘ in the solid state, as
determined from x-ray crystallography.2-5 In solution, the
relative concentrations of APP species are dictated by
three acid dissociation constants. Appleton et al.26
determined the effect of pH on the 31P and 1H NMR spectra
of various aminoalkylphosphonic acids in solution,
including APP. FTIR has also played an important role in
the structural identification of phosphonic acids.
Diagnostic IR bands allowing differentiation between APP
species in solid samples have been assigned 2 by
Garrigou-Lagrange et al.27 among others.28'31

The surface coupled products obtained from the
reaction of imogolite with APP at 0.43:1 (w/w)
APP:imogolite at pH 7.5 and pH 3.5 have been characterized
by 31P MAS NMR and FTIR. The well-ordered products
obtained by reaction of Na-montmorillonite with 5.9 moles
APP/clay equivalent at pH 3.9 and 7.5 have also been
characterized by x-ray diffraction, elemental analysis and
31p NMR, in order to distinguish them from APP-modified
imogolite intercalation. products. The ordered products
obtained from the reaction of APP-modified imogolite at pH
7.5 and 0.43:1 (w/w) APP:imogolite with Na-montmorillonite
at 2.5:1 (w/w) unmodified imogolite to clay after one
washing have also been analyzed. ' The ordered products
isolated were not the expected APP-modified imogolite

intercalated complexes.

LIST OF REFERENCES

10.

11.

12.

13.

14.

15.

16.

LIST OF REFERENCES

Handbook of Chromatography, V. 2: Zweig, G.: Sherma,
J., Eds.: CRC Press: Cleveland, OH, 1972.

Pines, H.: Haag, W. J. Am. Chem. Soc. 1960, 82, 2471.

Maciver, D.S.: Tobin, R.B.; Barth, R.T.: J. Catalysis
1963, 2, 485.

Peri, J.B.: Hannan, R.B. J. Phys. Chem. 1960, 64,
1526.

Peri, J.B. J. Phys. Chem. 1965, 69, 211.

Peri, J.B. J. Phys. Chem. 1965, 69, 220.

Cradwick, P.D.G.; Farmer, V.C.: Russell, J.D.: Masson,
C.R.: Wada, K.: Yoshinaga, N. Nature Phys. Sci. 1972,
240, 187.

Farmer, V.C.: Fraser, A.R. Dev. Sedimentol. 1978, 27,
547.

Barron, P.F.; Wilson, M.A.: Campbell, A.S.: Frost,
R.L. Nature 1982, 299, 616.

Goodman, B.A.: Russell, J.D.: Montez, B.: Oldfield,
E.: Kirkpatrick, R. J. Phys. Chem. Miner. 1985, 12,
342.

Johnson, L.M. M.S. Thesis, Michigan State University,
1988.

Theng, B.K.G.: Russell, M.: Churchman, G.J.; Parfitt,
R.L. Clays Clay Miner. 1982, 30, 143.

Clark, C.J.: McBride, M.B. Clays Clay Miner. 1984,
32, 291.

Clark, C.J.: McBride, M.B. Clays Clay Miner. 1984,
32, 300.

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

Johnson, I.D.: Werpy, T.A.: Pinnavaia, T.J. J. Am.
Chem. Soc. 1988, 110, 8545.

15

17.

18.

19.

20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.

31.

16

Werpy, T.A;; Michot, L.J.; Pinnavaia, T.J. In Novel

Materials in Heterogeneous Catalysis: Baker, R.T.K.:
Murrell, L.L., Eds.; American Chemical Society:
Wash1ngton, D.C., 1990.

MacKenzie, K.J.D.; Bowden, M.B.: Brown, I.W.M.;
Meinhold, R.B. Clays Clay Miner. 1989, 37, 317.

Johnson, L.M.: Pinnavaia, T.J. Langmuir 1990, 6, 307.

Ramsier, R.D.: Henriksen, P.N.; Gent, A.N. Surf. Sci.
1988, 203, 72. '

Templeton, M.K.: Weinberg, W.H. J. Am. Chem. Soc.
1985, 107, 97.

Templeton, M.K.: Weinberg, W.H. J. Am. Chem. Soc.
1985, 107, 774.

White, H.W.: Crowder, C.D.; Alldredge, G.P. J.
Electrochem. Soc. 1985, 132, 773.

Davis, G.D.: Ahearn, J.S.: Venables, J.D. J. Vac.
Sci. Technol. 1984, A2, 763.

Glowiak, T.: Sawka-Dobrowolska, W. Acta. Cryst. 1980,
B36, 961.

Appleton, T.G.: Hall, J.R.: Harris, A.D.: Kimlin,
H.A.; MCMahan, I.J. Aust. J. Chem. 1984, 37, 1833.

Garrigou-Lagrange, C.: Destrade, C. J. Chim. Phys.
1970, 67, 1646.

Destrade, C.: Garrigou-Lagrange, C. J. Chim. Phys.
1970. 67, 2013.

Menke, A.G.: Walmsley, F. Inorg. Chim. Acta 1976, 17,
193.

Fenot, P.: Darriet, J.; Garrigou-Lagrange, C.:
Cassaigne, A. J. Mol. Struct. 1978, 43, 49.

Sahni, S.K.: van Bennekom, R.: Reedijk, J. Polyhedron
1985, 4, 1643.

CHAPTER II
Silylation of a Tubular Aluminosilicate Polymer
(Imogolite), by Reaction with Hydrolyzed gamma-
aminopropyltriethoxysilane (APS)
A. INTRODUCTION

Chemical modification of surfaces is an area of
intense interest from both fundamental and practical points
of view.1'5 A variety of analytiCal techniques which have
been recently developed enable the surface chemical
interactions to be studied and the chemical species present
on a surface to be identified.6'9 This molecular-level
information is essential for an understanding of ‘the
macroscopic properties of a surface, such as adhesion,
resistance to chemical attack, and fracture.

The tubular aluminosilicate polymer imogolite,
(H0)3A1203Si(0H), has been targeted for surface
modification in the present study. The imogolite structure
is shown in cross-section in Figure II.1. The molecular
size of the tubular unit (- 23 A diameter, - 8 A internal
channel and 250 - 350 run length) gives rise to a large
surface area of 900 mz/g.10 The hydroxylated external
surface imparts water solubility to the tubes at low
concentration (< 0.4 wt. %) and acidic pH. Treatment of
the hydroxylated surfaces with organosilanes may allow for
a more hydrophobic surface, permitting extraction of the
tubular units into organic solvents. The ability to
control the solubility properties of imogolite using

surface modification techniques could lead to broader
17

18

00 Si
00 .M
CM) 0
.0 0H

 

Figure II.1. Cross-sectional view of the tubular
aluminosilicate imogolite (adapted from reference 22). The
inner and outer tube diameters are ~ 8.2 A and 23 A,
respectively. Tube lengths of the synthetic derivative
used in the present work are typically from 250 to 350 nm.
The empirical formula, as read from outer to inner atomic
planes is (H0)3A1203Si(0H) o

19

applications of this novel inorganic polymer in materials
science.

The reaction of imogolite with hydrolyzed gamma-
aminopropyltriethoxysilane (APS) was investigated as a
means of achieving the desired surface modification, in
part, because of the proven utility of hydrolyzed APS for
the organofunctionalization of inorganic oxide

surfaces.11'12

B. EXPERIMENTAL

Imogolite was synthesized according to the method
described by Farmer and Fraser13 and purified by dialysis
against deionized water. The hydrolyzed APS-imogolite
products were prepared by addition of a 2 wt. % aqueous APS
solution to a 0.1 wt. % suspension of purified imogolite
with subsequent stirring overnight. The 2 wt. % hydrolyzed
APS solution was prepared by reaction of APS (Petrarch
Systems) with water over a period of half an hour in the
presence of enough acetic acid to achieve a pH of 3.6. A
low pH is desirable for the coupling reaction because under
these conditions imogolite remains soluble and APS
undergoes rapid and complete hydrolysis. The amount of APS
incorporated in the coupling reaction was based on the
amount of aluminum used in the imogolite synthesis. Since
the synthesis of imogolite is not quantitative, typically
in < 50 % yield, the amount of APS present in the coupling

reaction was always in excess of monolayer coverage, even

20
when the formal APS : Al ratio was at the lowest formal
value of 0.5 : 1.0.

FTIR vibrational spectra for KBr-pellet samples were
obtained on an IBM model IR4OS spectrometer equipped with
an IBM IR44 work station. Difference spectra for
imogolite-bound APS were obtained by weighted digital
subtraction of the imogolite spectrum from the spectra of
the .APS-imogolite reaction products. Subtractions were
performed over the frequency range 4000 cm'1 to 400 cm'l.
Baseline corrections were not needed. Spectra generated in
transmittance ‘mode gave difference spectra in agreement
(3; 2 cm'l) with those generated in absorbance mode. The
validity of the difference routine was checked by comparing
the difference spectrum for APS on APS-modified silica gel
with the difference spectrum reported by Chiang et al.14
Within the spectral resolution of our instrument
(1 2 cm'l), the frequencies obtained for' our difference
spectrum were identical to those obtained by Chiang et al.

29Si MAS NMR spectra were obtained at 35.76 MHz on a
Bruker WH-180 spectrometer equipped with a Doty magic angle

spinning probe.

C . RESULTS

When hydrolyzed APS was mixed with imogolite at
pH 3.6, there was no visible indication of reaction. Both
reaction components remain dissolved and the reaction

mixture remained colorless. An initial attempt to separate

21

silylated imogolite from unreacted APS by dialysis was
inconclusive, because of the hydrolytic instability of the
silylated imogolite products. Complete and facile removal
of the coupling agent was achieved after three days of
dialysis at room temperature. Despite the hydrolytic
instability of the system, FTIR and MAS 29Si NMR
experiments demonstrated that there is indeed binding
between the hydrolyzed APS and imogolite and that the
resulting products are not simply homogeneous mixtures of
the two reaction components.

Shown in Figure II.2A is the FTIR spectrum of APS
polymer prepared by air-drying a film on a glass slide.
The spectrum in Figure II.28 represents air-dried APS which
has been hydrolyzed in the presence of acetic acid at
pH 3.6. The pair of bands near 1130 cm’1 and 1030 cm"1 in
both spectra are assigned to the asymmetric stretching
frequencies of Si-O-Si linkages in the polymer. The
observed frequencies of the bands in the 1300 cm’1 to
1600 cm"1 region reflect both the nature of the counter ion
associated with the protonated aminofunctionality and the
degree of protonation resulting from the hydrolysis
conditions. The partially protonated aminosilane polymer
formed by hydrolysis of APS in air exhibits carboxylate
stretching frequencies characteristic of bicarbonate ions
[vas(c02‘), 1638 cm“; vs(c02"), 1336 cm‘1115 in addition
to onium ion stretching frequencies [vas(NH3+), 1580 cm'l:

vs(NI-I 3+), 1484 cm'l] in accord with earlier assignments for

22

 

I
2932

TRANSMITTANCE

 

 

 

 

I I I I I I

I
4000 3000 2000 1000
WA V EN U M B E R S

 

o
u
u:
o
l
\\

111111111
‘\..
_____.—.

TRANSMITTANCE
9 9
5 0|
8 8
1

I "m" 11

 

 

..4 \
1573;;1404 £1129

 

4000 aobo 2000 1000

 

 

 

 

 

WAVENUMBERS
1 (o)

m

0 0.60:
z .
.5 0.50-4
: I
5 0.4054
2 i
< .
a: 0.304 '
"' I

I I I I I
4000 3000 2000 1000
WAVENUMBERS

Figure II.2. FTIR spectra (KBr disks) of (A) APS polymer
prepared by hydrolysis and drying in air: (B) APS polymer
prepared by hydrolysis in acetic acid at pH 3.6 and air
drying: and (C) synthetic imogolite.

23
the hydrolyzate of APS at ambient pH.1“'116’18 The poly-

(gamma-ammoniumpropyl ) siloxane acetate formed by
hydrolysis of APS in the presence of aqueous acetic acid,
however, in addition to [vas(NH3+), 1573 cm’l] exhibits
characteristic bands due to the presence of both
bicarbonate [vas(C02'), 1638 cm’l: VS(CO 2‘), 1336 cm‘l) and
acetate [vs(C02"), 1404 cm‘l; 6(OCO), 653 cm '1] 19 counter
ions.

The FTIR spectrum of imogolite shown in Figure II.2C
is characterized by the presence of prominent water bands
near 3500 cm"1 due to the OH stretch and near 1630 cm’1 due
to the HOH bending vibration. The two bands located. at
995 cm'1 and 940 cm"1 are those characteristic of the
Si-O-Al stretching vibrations of the tubular silicate.20
These latter bands are consistent with the presence of
grafted orthosilicate units in imogolite.

The FTIR spectra of the hydrolyzed APS-imogolite
reaction products contain several interesting features.
Figure II.3 provides the spectra of the products obtained
at formal APS : Al mole ratios of 0.5 : 1, 1 : 1, 2 : 1 and
2.5 : 1. The position of the vas(NH3+) frequencies
(1564 - 1572 cm’l) and VS(C02') frequencies (1410-
1414 cm'l) are similar to those observed in the pure
polymer formed by acid hydrolysis (1573 and 1404 cm'l,
respectively). Two vas(Si-O-Si) bandsl4 due to the
coupling agent are found above 1000 cm'l: one band occurs

at 1130 cm"1 and the other occurs at 1030 cm'l. These

24

.e " m.~ Ace .H u m Roy .H n a Ame .H “ m.o Ace no madame
nmaoe H< u mm< HmEHOH um pmeuou muoncoua :Owuomou oufiHoooaH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

omume>eem no Amxmeu umxe muuomem «Hem .n.HH manage
mcwm532m><>> mmw0232m><3
coo p coon coon 9600 690— ach can 8°C
1 _ P L A p _ p L rib..- _- L- u _.|.. ..»IL1I~.1L- » b L[ r _ _ h _ _tu —II.T 8.6
an: .32 Tan... 1 8.: £2
«a 2: W W «.3 -
.698
I, - . N - x
2: // .2 I 3. o S «no. .-
# /\\ .1/ r. m 7
\ Hone... _ .eee
. / M n , ;Jfl
“83 W .
IfH 3 a. not.
Q 1 x
I u -- - I-3-...InIIIIu-III. -, . 3.. L 02.6 ills-cu li-l - l!lll!llll|ull:|lui.tL 02.6
mmwm532m><>> mmm0532w><3
cco p coca coon ocov coop coca can 800
._ ._ _ -, _ _. hi! _Ll- r!!— o.-_ _ .- - _ a; r 00nd Ti. . _-.-._ p r P a Li I b L til—1&2“...
- I. -
iv: 89 - H
(J _ - V
g/ 9! . 3.6 N
on: )/ (. 7 H m
mead V
I — / I w
2 /. cod 3..
.8 -

 

BONVLLIWSNVHI

BONVJJJWSNVHJ.

25

bands are more easily observed in the spectra of products
containing higher APS concentrations. The assignment of
these two bands to mainly Si-O-Si stretch frequencies is
consistent with the presence of a network of highly
polymerized silane covering the imogolite tubes. A
monolayer of APS monomers and dimers linked to the surface
should have been represented by a single band near
1016 cm’1.14

It. was desirable to obtain further evidence for a
coupling reaction between the hydrolyzed APS and imogolite,
and to eliminate the possibility that the observed spectra
may be due to a physical mixture of imogolite and APS
polymer. This was accomplished through the analysis of
difference spectra obtained by digital subtraction of the
imogolite spectrum (cf., Figure II.2C) from spectra
obtained for the reaction products in Figure II.3. The
difference spectra are shown in Figure II.4.

The region between 1600 cm"1 and 1100 cm'1 in the
difference spectra is of special interest. The band in the
1129 - 1143 cm“1 region is an asymmetric Si-O-Si stretch
arising from the self-condensation of hydrolyzed APS units.
Since imogolite does not contain siloxane bonds, the
Si-O-Si stretch in the silylated reaction products arises
exclusively from the surface-bound polymer. In the four
resultant spectra this band is shifted to lower wavenumbers
with increasing concentration of APS. The positions of

other bands such as Vas(C02-) and vs(C02') due to

26

.m.HH musowm afl.muosoouq :ofluomou wuwaowoawummd gnu
you muuomam on» Eouu U~.HH mugoflm CH aguuooam wuwaoooaw on»
no cofluomuunsm anaemic >3 cwcfimuno Iao ooom pew Tao oom
newswon :ofimou we» now sunbeam wocmumuufic mHEm .¢.HH gunmen

 

 

 

 

 

 

 

 

 

 

mmmm532m><>> mmmm522m><>>
can: OGN— OOQP camp 25— 90°F OONP 8: 80¢ coop
LII _--_lp _.-,-._.FL-LLIPLIF. bl, _-_L .3. _ _ . ._ . _ . _ _ .. _u_L .. LL. . _-_ . .

3n. flan: «on. f

as; 2! T

Tomd I. x-

I u T

I V r.

- m

yeee -

r m r

r I. .-

I u l

11.05 C N rl

/r o I

. a I

-wnnlea.ee .:-- I. -

mmwmfiazm><>> mmmm532m><3
O2: coup OGQF coop camp 600— camp 90' — Sup 80 w

LI»! .175 . .LLeEIFL-Ll “mm” F. ._ LL. mand FLI— P _ t p _ . _ _ _ . p the. LLnfi. 2.6

2: I . I. .u.

n 8.. W 23 r
n N no: H MFA.

.1!de S . .

n W .\_ .

Wuhan“ H
n W I 36

mm; D «a. -

n 3 -
w 8.6 -- Q I 8.6

 

 

 

 

 

 

EONVLIJWSNVHJ.

BONVLLIWSNVHJ.

27

bicarbonate ion and VS(CO 2") due to acetate ion remain
unshifted in the difference spectra.

Table II.I lists the relevant FTIR frequencies
observed for the reaction products obtained at different
APS : Al molar ratios. The frequencies for the air-
hydrolyzed and acid-hydrolyzed APS polymer as an isolated
solid are also included. The 14 cm‘1 shift in vaS(Si—O-Si)
from 1129 cm‘1 in the pure acid hydrolyzed polysiloxane to
1143 cm'1 in the 0.5 : 1.0 APS : Al reaction, product is
significant. Although the 14 cm"1 shift is deduced from
difference spectra, the magnitude of the shift is much
larger than the 2.0 cm'1 instrumental resolution. A shift
in this band is also observed in the unsubtracted spectra
as shown in Table II.I, but in the opposite direction.
This is due to the fact that, especially at low APS
concentrations, the band position is shifted by
interference from the broad imogolite Si-O-Al band. (cf.,
Figures II.3A and 11.33.)

29Si MAS NMR spectroscopy was used to further probe
this system in an effort to obtain additional insights into
the nature of .APS bound. to imogolite. ‘Pure synthetic
imogolite exhibits a single resonance at -79 ppm, in
agreement with the shift reported for naturally occurring
imogolite.21'22 As shown in Figure II.5A, a single
resonance due to fully cross-linked. Rsi(OSi)3 sites is
observed for acid-hydrolyzed APS,23 but the siloxane line

occurs at -69 ppm, about 10 ppm downfield from imogolite.

28

Table 11.1. FTIR vibrational frequencies (cm-1) for hydrolyzed
APS and hydrolyzed APS-imogolite reaction products.a

 

 

 

Formal + - b -c
APS:A1 vas(NH3) vs(C02) vs(C02) vas(SiOSi)
Ratio
W
1.0 : 0.0d 1580 - 1336 1131
1.0 : 0.0e 1573 1404 1336 1129
2.5 : 1.0 1564 1410 1339 1126
2.0:1.0 1566 1412 1335 1127
10:10 1568 1413 1339 1120:
0.5 : 1.0 1572 1414 1342 1116
W13
25:10 1563 1410 1339 1137
20:10 1562 1410 1338 1137
1.0 : 1.0 1564 1413 1340 1141
05:10 1567 1415 1342 1143

 

a All reactions were carried out in aqueous acetic acid at pH 3.6
except where otherwise indicated.

b This is the symmetric carboxylate stretching frequency of the
acetate counter-ion.

c This is the symmetric carboxylate stretching frequency of the
bicarbonate counter-ion. _

d This ratio is representative of the pure APS polymer formed
by hydrolysis and drying in air.

c This ratio is representative of the pure APS polymer formed

by hydrolysis in acetic acid solution at pH 3.6 and subsequent
evaporation in air.

29

.o.e " m.o do oeuma H4 a med Hmauou e um emcemuno unseoua
:ofluocou duzoooefilmmm cd~>aouc>nlcfiom on» ADV “Honda cw
mwm>aouc>c >n cdcfimuno uda>aog dcmonHm muse ecu Amy «o.n an
an ounce ofiudom meadows aw mwm>aoup>s >n vacuum-.50 Magda
dcexoim when an» :3 "sunbeam mzz m4: “ma 6.: 0.55.?—

 

 

 

Eae End End
cap- o 8.... a can. cop- o :2
Jud..4.....qq..— —-J - . — .u . «J ‘1‘ — q _ _ _

rm-

 

 

om-

 

 

 

 

 

30

This chemical shift difference readily distinguishes the
silicon environments in imogolite and hydrolyzed APS.
Interestingly, if pure APS is hydrolyzed in water rather
than in aqueous acetic acid, then three resonances are
observed as shown in Figure II.5B. We assign the lines at
-69, -60, and -51 ppm to RSi(OSi)3, RSiX(OSi)2, and
RSix2(OSi) end groups, respectively, where X represents an
~01! or, less likely, an unhydrolyzed -0C21-15 group. These
assignments are in agreement with ‘those given by
Caravajal24 et al., among other workers.23v25’26

Figure II.5C illustrates the 2gsi MAS NMR spectrum of
the imogolite-APS reaction product obtained at a formal
APS:Al ratio of 0.5 : 1.0. At this reaction ratio, the APS
is in excess of ideal monolayer coverage, as judged by the
relative intensities of the siloxane and imogolite
resonances. The line at -80 ppm corresponds to the 03(3Al)
environment of imogolite. Significantly, the organosilane
region exhibits two silicon resonances at -70 and -61 ppm.
Since 29Si chemical shifts of siloxanes move upfield with
increasing degree of cross-linking to neighboring silicon
sites, we attribute the -70 ppm resonance to a fully cross-
linked. RSi(OSi)3 site analogous to that in pure acid-
hydrolyzed APS, whereas the -61 ppm line is assigned to a

less cross-linked RSiX(OSi)2 environment.

31
D. ' DISCUSSION

The hydrolysis of APS in acidic aqueous solution
(pH 3.6) is known to afford mainly monomer and dimer
species.1 Thus, the silanol is much more stable toward
. self-condensation to higher molecular weight polysiloxanes
than other organosilanes. The solution stabilization is
thought to be due to an interaction between the
aminofunctional group and either the Silicon center or the
hydrogen of one of the silanols.l4127"29 This interaction
protects the silanols against further condensation and
inhibits polymerization.29

The limited condensation of silanols between
hydrolyzed APS units prevents precipitation of the
organosilane coupling agent from solution. But the
evaporation at room temperature of an hydrolyzed APS
solution containing acetic acid at pH 3.6 does result in
the formation of a siloxane polymer, as indicated by the
appearance of siloxane stretching frequencies at 1129 cm'1
and 1035 cm'"1 (cf. Figure II.2B) and in the appearance of
only fully cross-linked RSi(OSi)3 sites in the 29Si MAS NMR
spectrum at -69 ppm (cf. Figure II.5A). However,
hydrolysis of APS in absence of acetic acid affords a less
polymerized and lower cross-linked product as indicated by
2981 MAS NMR lines at -60 and -51 ppm indicative of
RSiX(OSi)2 and RSiX2(OSi) groups (cf. Figure II.58) which

are not fully cross-linked.

32

The appearance of a strong IR band near 1568 cm‘1
indicates the presence of NH3+ groups in the APS-imogolite
products. Earlier studies of hydrolyzed APS on polished
iron29 and aluminum30 mirrors and on high surface area
silica gel14 have resulted in the assignment of a similar
band to amino groups H—bonded to surface silanols.
However, analogous vibrations in aminosiloxanes have also
been unequivocally assigned to the formation of
[NH3+][HCO3'] groups due to reaction of the amino group
with atmospheric C02 and water.16'18 In our APS-imogolite
reaction system the pH of the reaction mixture was adjusted
to 3.6 with acetic acid. Thus, the amino group should be
fully protonated during the surface coupling and surface
polymerization reactions, in accord with the observed
spectrum.

We attribute the increase in Si-O-Si frequency with
decreasing hydrolyzed APS surface coverage to an increasing
fraction of Si-O-Al bonding formed by the condensation' of
the siloxane to Al-OH sites on the external surfaces of the
imogolite tubes. At high surface concentrations of APS
most of the bound APS should resemble the pure polysiloxane
in bonding and structure. Thus, the vibrational
characteristics of the siloxane polymers far from the
surface will be similar to those for the APS polymerized in
absence of the surface. As surface coverage decreases, a
greater fractien of the polymer interacts with the

imogolite surface, most likely due to Si-O-Al coupling, and

33

the Si-O-Si frequency shifts to higher energy. We were
unable to observe directly the Si-O-Al frequency which
should result from the coupling of hydrolyzed APS polymer
to the imogolite surface. The tubular aluminosilicate
itself contains a strong, broad absorption near 940 cm"1
due to Si-O-Al linkages (cf. Figure II.2C) and this
absorption obscures weaker bands due to siloxane polymer
coupling.

No evidence is found for the coupling of hydrolyzed
APS to the intra-tubular Si-OH surfaces (cf. Figure II.1).
Since the inner diameter is 8.2 A, hydrolyzed APS _in
monomeric or dimeric form might be accessible to the Si-OH
groups. However, initial coupling to Si-OH groups near
channel openings should block the interior regions of the
channel from further coupling. Furthermore, all surface-
coupled APS is lost after three days of dialysis against
deionized water at room temperature. If coupling to the
Si-OH surface occurred, we would expect these bonds to be
stable to hydrolysis, comparable in stability to other
silicate surfaces functionalized by APS.31 For these
reasons it is unlikely that the interior surfaces of
imogolite are silylated.

The 29Si MAS NMR spectrum of the 0.5 : 1.0 APS : Al
reaction product (cf ., Figure II.5C) indicates that the
surface-bound polymer is less extensively cross-linked than
the pure APS polymer prepared under analogous reaction

conditions. Two resonances at -70 and -61 indicate the

34

presence of fully cross-linked Rsi(osi)3 groups and
partially cross-linked RSiX(OSi)2 groups, respectively. In
contrast, acid hydrolyzed APS alone contains only fully
cross-linked RSi(OSi)3 environments. Apparently, the
immobilization of the polymer at the imogolite surface
limits the degree of cross-linking. However, the
imogolite-bound APS polymer is more extensively polymerized
and cross-linked than the pure polymer formed by hydrolysis
in absence of acetic acid (cf. Figure II.5B).

Although the FTIR and MAS NMR spectroscopic data are
consistent with covalent coupling of a well cross-linked
APS polymer to the Al-OH groups of imogolite, the surface
modified products are hydrolytically unstable. Dialysis of
the hydrolyzed APS-imogolite complexes results in the
complete loss of polymer from the imogolite surface after
three days of reaction. The hydrolysis of polymer far from
the imogolite surface is expected, because the pure polymer
itself also undergoes rapid hydrolysis (within hours) under
equivalent dialysis conditions. However, even the APS
coupled to the imogolite surface is hydrolytically
unstable. No siloxane polymer, even at the sub-monolayer
level, could be detected by FTIR for the products after
several days of dialysis against deionized water.

Although aminosilane coupling agents are useful for
the organo-functionalization of silanol mineral surfaces,
they lack the desired hydrolytic stability for

organofunctionalization of the curvi-linear Al-OH surfaces

35

characteristic of imogolite. Future studies are planned
which will address the question of whether the topology of
the imogolite surface contributes to the hydrolytic
instability of the coupled aminosiloxane bonds relative to

other types of Al-OH surfaces.

LIST OF REFERENCES

10.

11.

12.

13.

14.

15.

LIST OF REFERENCES

Plueddemann, E. P. silane Coupling Agents; Plenum:
New York, 1982; Chapter 3.

Sung, C. S. P.; Lee, S. H.: Sung, N. H. Polym. Sci.
Technol. 1980, 12B, 757.

Kulkarni, R. D.: Goddard, E. D. Int. J. Adhes. Adhes.
1980, 1, 73.

Furukawa, T.; Eib, N. K.; Mittal, K. L.;
Anderson, H. R., Jr. J. Colloid Interface Sci. 1983,
96, 322. .

Alexander, J. D.: Gent, A. N.: Henriksen, P. N. J.
Chem. Phys. 1985, 83, 5981.

Garbassi, E.: Occhiello, E.: Bastioli, C.: Romano, G.
J. Colloid Interface Sci. 1987, 117, 258.

Mittal, K. L., Ed. Symposium on Adhesion Aspects of
Polymeric Coatings; Plenum: New York, 1983.

Leary, H. J., Jr.: Campbell, D. S. Org. Coat. Appl.
Polym. Sci. Proc. 1983, 46, 433.

Furukawa, T.: Eib, N. K.: Mittal, K. L.;
Anderson, H. R., Jr. Surf. Interface .Anal. 1982, 4,
240.

Egashira, K.: Aomine, S. Clay Sci. 1974, 4, 231.

Leyden, D. E., Ed. Silanes, Surfaces and Interfaces:
Gordon and Breach: New York, 1985.

Leyden, D. E.: Collins, W. T., Eds. silylated
Surfaces: Gordon and Breach: New York, 1978.

Farmer, V. C.: Fraser, A. R. Dev. Sedimentol. 1978,
27, 547.

Chiang, C. H.: Ishida, H.; Koenig, J. In J. Colloid
Interface Sci. 1980, 74, 396.

Colthup, N. B.: Daly, L. H.: Wiberly, S. E.
Introduction to Infrared and Raman Spectroscopy:
Academic: New York, 1975.

36

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

37

Ishida, H. In Symposium on .Adhesion Aspects of
Polymeric Coatings; Mittal, K. L., Ed.: Plenum: New
York, 1983.

Boerio, F. J.: Williams, J. W. Proc. 36th Annu. Conf.
Reinf. IPlast./Compos. Inst“, Soc. .Plasta Ind. 1981,
Section 2-F.

Naviroj, 8.: Koenig, J. L.: Ishida, H. Proc. 37th
Annu. Conf. Reinf. Plast./Compos. Inst. , Soc. Plast.
Ind. 1982, Section 2-C.

Nakamoto, K. Infrared and Raman Spectra of Inorganic
and Coordination Compounds: Wiley: New York, 1986.

Cradwick, P. D. 6.: Farmer, V. C.: Russell, J. D.;
Masson, C. R.: Wada, K.; Yoshinaga, N. Nature Phys.
Sci. 1972, 240, 187.

Barron, P. F.; Wilson, M. A.; Campbell, A. 3.;
Frost, R. L. Nature, 1982, 299, 616. .

Goodman, B. A.: Russell, J. D.: Montez, 8.;
Oldfield, E.: Kirkpatrick, R. J. Phys. Chem. Minerals
1985, 12, 342.

Sudholter, E. J. R.: Huis, R.; Hays, G. R.;
Alma, N. C. M. J. Colloid Interface Sci. 1985, 103,
554.

Caravajal, G. S.: Leyden, D. E.: Maciel, G. E. In
Silanes, Surfaces and Interfaces: Leyden, D. E., Ed.:
Gordon and Breach: New York, 1985, p. 283-303.

De Haan, J. W.: Van Den Bogaert, H. M.: Ponjee, J. J.:
Van De Ven, L. J. M. J. Colloid Interface Sci. 1986,
110, 591.

Rudzinski, W. E.: Montgomery, T. L.: Frye, J. 3.;
Hawkins, B. L.: Maciel, G. E. J. Chromatogr. 1985,
323, 281.

Plueddemann, E. P. Proc. 24th Annu. Conf. Reinf.
Plast./Compos. Inst., soc. Plast. Ind. 1969, Section
l9-A.

Plueddemann, E. P. In .Analysis of Silicones:
Smith, A. L., Ed.: Wiley: New York, 1974.

Boerio, F. J.: Armogan, L.: Chang, S. Y. J. Colloid
Interface Sci. 1980, 73, 416.

Boerio, F. J. Org. Coat. Plast. Chem. ACS Preprints,
1981, 44, 625.

38

31. Morrall, S. W., Ph.D. Thesis: Colorado State
University, 1984.

CHAPTER III
Hydrolytic Lability of silylated Imogolite, and
Intercalation of the Silylated Tubular, Aluminosilicate in
Montmorillonite.

A . INTRODUCTION

The modification of oxide surfaces by coupling with
gamma-aminopropyltriethoxysilane (APS) and other
functionalized organosilane coupling agents is finding
useful applications in the areas of electrochemistry,
chromatography, catalysis, and composite materials.1’6 In
some systems, however, such as those involving Fe203 and
A1203 surfaces, the performance characteristics of Ithe
organosilane-oxide interface is limited by the lability of
the surface bonds in humid or aqueous environments.7'9

We recently have been concerned with the surface
chemical properties of the novel tubular aluminosilicate
imogolite.10 This compound has a molecularly regular outer
diameter of 23 A and an inner channel diameter of 8.7 A.11
The outer surface consists of a gibbsite-like array of
Al-OH groups, whereas the inner surface is composed of a
regular distribution of SiOH groups. Our earlier FTIR and
29Si MAS NMR studies provided evidence for the binding of
APS polymer to the external alumina-like surfaces of
imogolite. I

Organosilane imogolite derivatives are potentially
useful pillaring reagents for the synthesis of a new class
of microporous tubular silicate-layered silicate (TSLS)

intercalation compounds.12 Pristine imogolite itself
39

40

funCtions as a pillaring reagent by intercalating into
smectite clays at the monolayer level and forming regular
intercalates that exhibit several orders of 001 x-ray
diffraction. However, the intercalated tubes are laterally
in van der Waals contact: consequently, the microporosity
of the TSLS complex is restricted primarily to the internal
channel of the tubes. By increasing the Size of the tubes
through the coupling of organo groups, one might expect to
increase the lateral separation between tubes and thereby
generate additional microporosity in a TSLS complex upon
removing the organo groups through calcination.

The results of the present study demonstrate in part
that APS-modified imogolite intercalates into
Na-montmorillonite, but the resulting TSLS complex does not
exhibit regular x-ray basal spacings. In an effort to
better understand the surface chemistry' of APS-modified
imogolite, we have undertaken a study of the hydrolytic
lability of this material. Also, we have included in the
study the hydrolysis of APS-modified gamma-alumina in order
to differentiate the hydrolytic properties of the external
AlOH surface of imogolite from those of an internal SiOH

Surface.

B. EXPERIMENTAL
Materials

Imogolite was synthesized according to the method
described by Farmer and Fraserl3 and purified by dialysis

against deionized water. Gamma-alumina ("Catapal B") was

41

obtained from Vista. Na-montmorillonite (SWy-l, Source
Clay Minerals Repository) was purified by sedimentation.
APS was purchased from Petrarch Systems and used without

further purification.

Silylation Reactions

Silylation reactions with APS in aqueous solution were
carried out using 0.17 wt. % and 0.08 wt. % suspensions of
gamma-alumina and imogolite, respectively. The
suspensions were adjusted to pH 4.2 with acetic acid before
the addition of APS. A 2 wt. % solution of APS was
adjusted to pH 4.2 with acetic acid and allowed to stir for
0.5 h. The desired amounts of APS solution were then added
to the mineral suspensions, and the mixtures were allowed
to stir over night. The amount of organosilane added to
each suspension equaled 78 and 85 wt. % of the air-dried
solid. imogolite and. gamma-alumina, respectively; These
quantities of APS used in the Silylation reactions exceeded
the amounts needed for monolayer coverage of the
substrates. A 0.5 wt. % solution of pure APS in deionized
water also was prepared, and the pH was adjusted to 3.7

with 1.0 M acetic acid.

Hydrolysis of APS-Imogolite

The silylated alumina and imogolite suspensions were
divided into 25 mL aliquots and placed into tubular
cellulose dialysis membranes (Spectrapor-Spectrum Medical

Industries) having a molecular weight cutoff of 12,000 to

42

14,000. The membrane tubes were subsequently placed into
5 L of deionized water. A membrane tube was removed every
two hours for the first eighteen hours, and at longer
intervals thereafter. Each time a sample was removed, the
deionized water was replaced. With the exception of the
hydrolyzed APS, the dialyzed suspensions were air-dried,
and the amounts of bound APS were analyzed by Fourier
Transform Infrared (FTIR) spectroscopy. The depletion of
pristine APS polymer from the dialysis membrane was
determined by allowing the solution to evaporate in air and

weighing the residue recovered.

gana-Aninopropyltriethoxysilane-montmorillonite Complex

A 1.33 wt. % Na-montmorillonite suspension was mixed
with a known volume of a 1 wt. % APS solution that had been
adjusted to pH 3.7 by the addition of 1.0 M acetic acid.
The mixture was allowed to dry in air to afford a composite
product containing 57 wt. % APS. The product was then
washed with water and air-dried.
gam-A-inopropyltriethoxysilane-Inogolite-Montnorillonite
Complex

A 0.1 wt. % suspension of imogolite was combined with
a known volume of a 1 wt. % APS solution that had been
adjusted to pH 3.7 by the addition of 1.0 M acetic acid.
The concentration of APS was made equal to 7 wt. % of the
imogolite concentration. A volume of 1.33 wt. %
Na-montmorillonite was added to the APS-treated imogolite

suspensions in a weight ratio of 2 . 5:1

43

imogolite:montmorillonite. 'The mixture was stirred
overnight, collected by centrifugation, resuspended in

water and air-dried.

Physical Measurements

x-ray diffraction patterns were obtained for oriented
films formed by air-drying the suspensions on the surface
of a glass slide. A Rigaku' Rotaflex model RU-200BH
diffractometer equipped with a copper target was used to
measure basal spacings.

FTIR vibrational spectra of air-dried products were
recorded over the frequency range from 4000 cm'12 to
400 cm'1 on an IBM model IR 408 instrument equipped with an
IR44 workstation. The samples (40 mg) were weighed on an
analytical balance and pressed into 2 wt. % KBr pellets.
The concentration of APS in each sample was determined from
the absorbance ratios of selected APS and substrate
stretching frequencies. The absorbance of the Vas(NH3)+
vibration at 1570 cm'1 was used to determine APS
concentrations for the APS-modified imogolite and gamma-
alumina samples. The intensities of the silane vibrations
were normalized with respect to the V(SiOAl) stretch at
954 cm"1 for imogolite and the 1070 cm'1 v(AlOAl) vibration
in gamma-alumina. APS Concentrations for representative
samples determined in this manner were confirmed by
elemental analysis. The concentration data were

reproducible to within +/-10 percent of the reported value.

44
The 1H decoupled MAS 29Si NMR spectra were obtained on

a Varian VXR400 spectrometer equipped with a Doty probe.
The 2gsi spectra were generated at a frequency of
79.459 MHz incorporating a 30 degree pulse of 3 usec, a
delay time of 20 seconds and a sample spinning rate of
5 KHz. All chemical shifts were reported with reference to
TMS.

Carbon and hydrogen elemental analyses were performed
by Galbraith Laboratories in Nashville, Tennessee. Silicon
and aluminum elemental analysis was done at the inorganic
chemistry laboratory of the Department of Toxicology: at
Michigan State University.

Three-point nitrogen BET surface areas for the
materials were measured on a Quantachrome Quantasorb Jr.

sorptometer.

C. RESULTS

Gamma-aminopropyltriethoxysilane (APS) is known to
hydrolyze in dilute aqueous solution to the
organotrisilanol and an organosiloxane dimer.7 Allowing
the solution to evaporate in the open atmosphere leads to
the protonation of the amino group by carbonic acid, and to
the condensation of APS units to form a gamma-
ammoniumpropylsiloxane polymer. When an oxide surface is
exposed to a hydrolyzed APS solution, the APS initially
becomes hydrogen-bonded to the surface hydroxyls.7
Dehydration normally results in the formation of a surface-

coupled organosiloxane polymer. The characterization of

45
the. surface-bound polymer can be accomplished using FTIR
and 2981 MAS NMR spectroscopy.

FTIR and 29Si MAS NMR measurements from previous
work10 have provided evidence for the formation of a
surface-bound polysiloxane on imogolite upon evaporating a
suspension of the substrate in 2.4 wt. % APS at pH 3.6 in
air at ambient temperature. This siloxane polymer,
however, was readily dissociated from the imogolite surface
upon dialysis against distilled water.

In the present work we have monitored by FTIR
spectroscopy the APS concentration throughout the dialysis
of the APS treated-imogolite suspension in order to
determine the hydrolysis rate. The dialysis rate for the
imogolite-bound polymer was compared with the dialysis rate
for the pristine polymer under analogous conditions. As
expected, the diffusion of pristine polymer through the
dialysis membrane was much faster than the rate of loss for
the polymer coupled to imogolite. For comparison purposes,
we also examined the dialysis rate for the silylated
surface of gamma-alumina, an amorphous oxide having surface
AlOH units. A comparison of the hydrolysis behavior could
then be made for siloxane polymer bound to the bifunctional
SiOH and AlOH surfaces of imogolite and the monofunctional
surfaces of alumina having only AlOH sites.

The APS concentrations for the siloxane-modified
imogolite and gamma-alumina samples were determined from

the absorbances of selected IR bands for KBr pellets of the

46

air-dried samples. The concentration of pristine APS
polymer was determined by weighing the residue collected
from the air-dried samples after each dialysis interval.

The FTIR spectra for the starting imogolite and gamma-
alumina substrates are shown in Figure III.1. Imogolite
(Figure III.1A) exhibits two bands due to the presence of
water, namely, the V(OH) stretching frequency near
3500 cm’1 and the 6(HOH) deformation band at 1630 cm‘l.
There also are two bands at 995 cm"1 and 940 cm"1 due to
V(SiOAl) stretching modes that are indicative of the
isolated orthosilicate units present in the imogolite
structure.15 The gamma-alumina FTIR spectrum, (Figure
III.1B), contains prominent hydroxyl stretches in the
3000 cm’1 to 3700 cm"1 range and a sharp V(AlOAl) band, at
1070 cm'l.

The FTIR absorbances characteristic of pristine APS
polymer, formed by hydrolysis in the presence of acetic
acid at pH 3.7 in aqueous acetic acid and evaporation of
the solution in air at room temperature, were analyzed in
detail in our previous work.10 Two vas(SiOSi) stretching
modes were assigned to absorbances at 1130 and 1030 cm'l.16
The vas(CH2) frequency of the propyl group and the
alkylammonium ion stretching frequency, Vas(NH3+)r were
assigned to bands at 2926 cm'1 and 1573 cm’l, respectively.
Acetic acid, along with carbonic acid formed from
atmospheric C02, caused protonation of the amino group in

the polymer. Very strong absorbances due to the

47

 

 

 

995 940
A
m ( )
o
c
o
.o
L.
o
0)
.o
<
o
>
O;
2 1070
a:
0: (B)

 

 

 

 

 

 

rIrrrTrfi firfir'
4000 3500 3000 2500 2000 1500 1000 500
Wavenumber

Figure III.1. FTIR absorbance spectra (KBr pellets) of the
substrates (A) imogolite and (B) gamma-alumina.

48

counterions occurred at 1400 cm’1 for the acetate VS(RC02')
stretch and at 1638 cm'1 and 1336 cm'1 for the bicarbonate
vas(C02’) and vs(C02‘) absorbances, respectively.

We consider next the FTIR properties of APS adsorbed
on imogolite and gamma-alumina at initial loadings of 78
and 85 wt. % respectively. These loadings are in excess of
the amounts necessary for monolayer coverage. On the basis
of the N2 BET surface areas for imogolite (500 mz/g) and
gamma-alumina (290 mz/g), a hydroxyl group surface
concentration of ~12-14/100 A2, and coupling of one RSiO3
moiety to 3 surface OH groups, complete monolayer coverage
should occur at 46 and 31 wt. % APS, respectively. Of
course, ‘these ‘values :represent. upper' loading’ limits for
ideal monolayer coverage because they are based on
estimates that disregard the surface occupied by the organo
groups and the formation of siloxane polymer. Vibrational
frequencies due to the polysiloxane were readily observed
in the FTIR spectra of the initial APS-modified samples.
Spectra obtained after dialysis periods of 0, 4, 12 and 50
hours are shown for the APS-modified imogolite and gamma-
alumina systems in Figures III.2 and III.3, respectively.

The siloxane was identified in the spectra of the
initial samples by a band in the 1120 to 1140 cm'1 region
which was assigned to the asymmetric stretching frequency
of SiOSi linkages in the polymer.16 An alkylammonium ion
stretching frequency, vas(NH3+) , similar to that observed

in the pure polymer, was found near 1570 cm'l. The acetic

49

 

 

 

 

a:
o
c
o
.o
L.
8
.o 4 hrs
42
‘3
2.3 12 hrs
3
a)
DC
50 hrs

 

 

 

4000'35'00'3o'oo‘ 25'00'20'00' 1300110'00‘ 560
Wavenumber

Figure III.2. FTIR absorbance spectra (KBr pellets) of
APS-modified imogolite samples isolated after dialysis for
the time intervals indicated.

50

 

 

 

 

1571 1136
2933 “.04 .

0 hrs '

8

c 4 hrs

0

.o

L.

O

m

.o

<

.g 12 hrs

4.:

2

1D

0:
50 hrs

 

 

 

4000' 3530'30'00'251003600' 1 500' 1 600' 560
Wovenumber

Figure III.3. FTIR absorbance spectra (KBr pellets) of
APS-modified gamma-alumina samples isolated after dialysis
for the time intervals indicated.

51

acid responsible for the initial protonation of the amine,
gives rise to the strong acetate VS(RCOZ') band near
1400 cm'l. The amine also reacts with carbonic acid formed
by atmospheric C02. Vibrations due to the presence of
bicarbonate ion were visible at 1638 cm"1 due to
vas(C02'), and 1336 cm"1 due to VS(C02'). The sharp band
near 2930 cm'1 has been assigned to the Vas(CH2) frequency
of the propyl group.

As dialysis time was increased, a sharp reduction in
siloxane absorbances relative to the substrate absorbances
was observed for each of the APS-modified substrates. . A
comparison of the FTIR spectra for the pristine substrates
in Figure III.1, with the APS-modified substrates in
Figures III.2 and III.3, provides verification for the
retention of significant amounts of siloxane polymer by
both substrates after 12 hours of dialysis. The presence
of the gamma-aminopropyl group is especially visible in the
VaS(CH2) stretching region near 2930 cm'1 and in the
alkylammonium ion stretching frequency, Vas(NH3+)v near
1570 cm'l. The latter absorbance was utilized in the
determination of the siloxane concentration. After twelve
hours of dialysis, the APS concentration was 40 and 24
wt. % for the imogolite and gamma-alumina substrates,
respectively. These values begin to approach the values
expected for ideal monolayer coverage for the two

substrates.

52

Plots of the log [wt. % APS] versus dialysis time for
the pristine APS, APS-modified imogolite and APS-modified
gamma-alumina systems are Shown in Figure III.4. Although
these data are not suitable for detailed kinetic treatment,
the relative APS labilities can be expressed empirically by
the time required to hydrolytically remove half of the
polymer (t1/2)' For pristine APS, the t1/2 value estimated
from the Figure is only 1.7 hours. The monomeric and
dimeric units that form upon APS hydrolysis are depleted
rapidly through the pores of the dialysis membrane. It is
particularly significant that the depletion of the pristine
APS from the dialysis membrane is at least five times as
fast as the depletion of polymer from APS-modified
imogolite and gamma-alumina. Thus, the polymer depletion
in these latter cases is limited by the hydrolytic
displacement of the polymer from the substrate surfaces.

Bimodal depletion behavior was observed for the
dialysis of both APS-modified imogolite and gamma-alumina.
At‘ the initial stages of dialysis where the APS surface
concentration is in excess of a monolayer, the depletion
rates were similar (t1/2 = 10 h). However, at longer
dialysis times the depletion rate decreased for the APS-
modified imogolite and gamma-alumina. At polymer loading
below monolayer levels, <46 wt. % for imogolite, the
depletion of polymer from APS-modified imogolite slowed to
t1/2 = 63 h, whereas the depletion rate for the APS-

modified gamma-alumina increased to t1/2 = 2.4 h.

.ncwsgamleEEmm cdfluwcosnmmc
can muwaouoefl cdfimflpoelmmc .umexaon mm< pdu>aouc>s on» you
was» memeemee memum> Amea w .useeoe 66 node .e.HHH seamen

$505 65:.

53

 

 

nzu .fim .Nv .#n .#N Aw— Le
b D b D b D 000
14x0

m 4 .

4

m m 4 4 Ind

4 .
Imu—

4

o o m .

o o o o o a 4 1m.—
» o
ec_E:_<lmn—< m m m o ..
3.588.192 o a .

md< ‘4 n.N

 

 

(de z) 501

54

In an effort to characterize the polymer adsorbed on
imogolite and gamma-alumina, the modified oxides were
investigated by 2gsi MAS NMR. The spectrum for APS-
modified imogolite at a 20 wt. % concentration is Shown in
Figure III.5A. This sample was prepared by direct reaction
with APS at pH 3.7, and did not undergo dialysis. However,
dialyzed samples showed the same characteristic NMR spectra
as non-dialyzed samples at comparable APS loading. This
spectrum contains a very intense resonance at -79 ppm due
to the HOSi(OAl)3 environment of the unmodified imogolite
orthosilicate units.]-7v18 The envelope of resonances in
the -45 ppm 'to ~70 ppm region. contains a shoulder at
-48 ppm and two resolved resonances at -58 ppm and -68 ppm.
These three resonances were assigned to the following
silicon sites of the surface bound polymer: RSi(OH)2(OM),
RSi(OH)(OM)2 and RSi(OM)3 where M is an adjacent Si site in
the polymer, or an Al or Si site at the imogolite surface.
These assignments are in accord with the chemical shifts
observed for APS-modified silica surfaces.:'-9"23 A fourth
broad resonance at -90 ppm is assigned to silylated
silicon sites in the imogolite structure. The 11 ppm shift
upfield is consistent with the conversion of some imogolite
Si(OH)(OAl)3 sites at the inner surface of the tubes to
Si(OSi)(OAl)3 sites by reaction with APS.

The 1H decoupled 29Si MAS NMR spectrum of APS modified
gamma-alumina at a 24 wt. % concentration is shown in

Figure III.5B. This sample also was prepared by the direct

55

 

 

 

_59 -67

60

(B)

 

-10'-30 -50 -70 -90 -120 ppm

Figure 111.5. 1H decoupled 2931 MAS NMR (79.459 MHz)
spectra of (A) 20 wt. % APS—modified imogolite and (B) 24
wt. % APS-modified gamma-alumina samples using a 3 us pulse
width, a relaxation delay time of 20 s, and a sample
spinning rate of 5 kHz.

56
reaction with APS at pH 3.7 without undergoing dialysis.

Resonances due to the surface-coupled polymer at -50 ppm,
-59 ppm and -67 ppm were assigned to RSi(OH)2(OM),
RSi(OH)(OM)2 and RSi(OM)3 (M = Si, Al) sites, respectively.
Here, the lines are sharper than the corresponding lines
for APS-modified imogolite. The reduced linewidth suggests
that the APS siting is less heterogeneous on the
monofunctional alumina surface.

Table III.I summarizes the 2981 MAS NMR chemical shift
assignments for APS-modified imogolite and gamma-alumina.
Included in the table are the shifts for pristine polymer
prepared at pH 3.7 and at the ambient pH value of 10. The
values for the oxide bound polymer are shifted slightly
downfield from those observed for the pristine polymer. It
is noteworthy that the extent of APS crosslinking on the
metal oxide surfaces at pH 3.7 is less than that observed
for pristine polymer prepared at the same pH. In fact, the
distribution of the organosilicon sites on the imogolite
and gamma-alumina surfaces are similar to those observed
for APS polymerized at ambient pH 10.

We next investigated whether APS-modified imogolite
could be intercalated into the galleries of a Na-
montmorillonite clay to form a tubular silicate layered
silicate (TSLS) intercalation complex. The APS-modified
imogolite intercalation in Na-montmorillonite was
anticipated because the positive charge provided by the APS

polymer should allow for electrostatic interaction between

57

On In 3 3835 803

SEES 82E. 5 .2 Lo mm .1. E a

 

- - as- co-
1 I act I
- - we- 3.
ca- as- 46. an-

2 me a
cone—08.22—

_m- mm<

On an H
castes—Don

- mm<

seems—34.1»
c n - comb—ooEéA—a‘

nos—owes.—
” .4 - comm—58-93.

 

 

38% 5.499195 52er 42939.9. Azofizoema 5.34m

 

85:52-». comm—ooEén—e. can cam—owes: 35.58.92
.88on mm< 0553 5 8:08—8:28 zoom—mm com «£5 3382.0 53 .3: 03¢.

 

58

the negatively charged clay layers. However, hydrolyzed
APS itself may compete with APS-modified imogolite for
intercalation in montmorillonite. Therefore, prior to
investigating the reaction of APS-modified imogolite with
Na-montmorillonite, we examined the reaction of the clay
with hydrolyzed APS at pH 3.7. Figures III.6A and III.6B
compare the x-ray diffraction patterns for Na-
montmorillonite before and after treatment with sufficient
hydrolyzed APS solution to give a 57 wt. % APS
concentration at pH 3.9. A highly ordered single phase was
obtained with a 21.3 A d001 spacing, as compared with a
12.7 A spacing for the air-dried Na-montmorillonite. Upon
washing the product with deionized water, the d-spacing for
the air-dried product decreased to 18 A and finally to
14.4 A as shown by the x-ray patterns in Figures III.6C and
III.6D.

Figure III.7 shows the x-ray diffraction pattern of
the air-dried product obtained by the reaction of a 7 wt. %
APS-modified imogolite suspension with a Na-montmorillonite
suspension at a weight ratio of 2.5:1 imogolite:c1ay.
Also, included are the x-ray diffraction pattern of the
product after one and two washings with water. The
unwashed product exhibits very diffuse reflections near 15
and 34 A. Washing the product with water improves the
crystallographic ordering and leads to an imogolite-
intercalated montmorillonite with a 35.6 A d001 spacing.

This spacing is identical to the value obtained for TSLS

59

 

 

 

 
  
   

12.7 A

Na-mont
E (A)
(n ’ I “ 7‘“ ‘
E.
}_ 21.3 A
Z
DJ APS-moat. unwashed
2 (B)
E A _ -11- 4..-
DJ
0: 14.4 A

APS-moat. washed 1X

9

 

14.4 A

APS-moat. washed 2X
(D)

3 5 7 9 11 13 15
DEGREES TWO THETA

 

  

 

 

Figure III.6. x-ray diffraction patterns of oriented films
of (A) Na-montmorillonite, (B) 57 wt. % APS-modified Na-
montmorillonite, (C) sample in (B) washed once with H20,
and (D) sample in (B) washed twice with H20.

60

 

   

(A) unwashed

37.4

15's (B) washed 1 X

 

RELATIVE INTENSITY

 

washed 2X

 

 

 

 

2 4 6 6 1B 12 14
DEGREES TWO THETA

Figure III.7. X-ray diffraction patterns of oriented films
of (A) 7 wt. % APS-modified imogolite suspension combined
with a Na-montmorillonite suspension in a wt. ratio of
2.5:1 imogolite to clay, (B) sample in (A) washed once with
water and (C) sample in (A) washed twice with water.

61
formed by intercalation of pristine imogolite into Na-
montmorillonite.11 Clearly, the TSLS reaction products are
easily distinguishable by x-ray diffraction from APS-
intercalated montmorillonite.

TSLS products were also obtained using APS-modified
imogolite at concentrations of 15 and 30 wt. %. These TSLS
products behaved similarly to the material obtained using
7 wt. % APS. The unwashed products exhibited highly
diffuse x-ray patterns. Subsequent washings yielded
products with x-ray diffraction patterns characteristic of
ordered systems having basal spacings of 35 to 37; A,
typical of pristine imogolite intercalated montmorillonite

products.

D. DISCUSSION

Within experimental uncertainty, identical depletion
rates (t1/2=10 hr) were observed for APS-modified imogolite
and APS-modified gamma-alumina at loadings corresponding to
multilayer APS coverage. This result suggests that the
structure of the APS polymer far from the substrate surface
is similar in both systems. However, substrate dependent
depletion rates were observed for polymer loadings below
approximately 48 wt. % APS. The half-life for APS
depletion from APS-modified imogolite . decreased
dramatically (t1/2 = 63 hours) at low APS loadings compared
with that observed for higher loadings. On the other hand,
the rate of APS removal from APS-modified gamma-alumina

increased (t1/2 = 2.4 hours) at low APS loadings (of.

62

Figure III.4). These differences in depletion rates at low
APS loadings suggest that the APS binding mechanism depends
on the nature of the substrate.

The nature of the APS polymer—surface coupling was
probed by 1H decoupled 2gsi MAS NMR. 1H decoupled 29Si MAS
NMR results from an APS-modified gamma-alumina sample
having an APS loading of 24 wt. %, cf. Figure III.58,
showed the presence of RSi(OH)2(OM), RSi(OH)(OM)2, and
RSi(OM)3 resonances, where M = Si, Al linkages. Resonances
having chemical shifts and relative intensities of pristine
APS polymerized at ambient pH dominate the spectrum.
Therefore, few covalent polymer linkages to the gamma-
alumina surface exist at submonolayer APS loadings on APS-
modified gamma-alumina. The effects of isolated covalent
or non-covalent polymer-surface coupling are indirectly
demonstrated by the extent of crosslinking observed. In
the presence of gamma-alumina at pH 3.5, APS polymerizes to
a moderately crosslinked form characteristic of pristine
APS polymerized at pH 10. Under conditions of reduced pH,
in the absence of gamma-alumina, pristine APS polymerizes
to a fully crosslinked form, characterized by the presence
Of a single RSi(OSi)3 resonance. These results do not
allow’ the identification of the binding’ mechanism, but
together with the hydrolysis behavior, they suggest that
polymer crosslinking is inhibited by the isolation of APS

units immobilized on the alumina surface.

63

298i MAS NMR was used to explain the reduced polymer
depletion rate for APS-modified imogolite samples at low
APS loading. The 1H decoupled 29Si MAS NMR spectrum of a
20 wt. % APS-modified imogolite sample, cf. Figure III.5A,
exhibits resonances due to RSi(OH)3_x(OM)x, where
M = Si, Al. In contrast to the APS-modified gamma-alumina
spectrum, the chemical shifts and relative intensities of
resonances observed were unique to the APS-modified
imogolite, suggesting a large percentage of polymer-surface
linkages. The chemical shifts of the resonances in this
spectrum were Shifted downfield from typical pristine
polymer resonances. The intensity of the RSi(OH)(OM)2
resonance was enhanced, compared with the other polymer
resonances in the APS-modified material. A resonance due
to polymer coupled to imogolite silicon sites (-90 ppm) of
the form (AlO)3(SiO)RSi(OM)2 gave direct evidence for
covalent binding of the polymer to the internal SiOH
surfaces of imogolite. The channel, 8.7 A in diameter,
accessible to hydrolyzed APS units from the ends of the
tubes, is depicted in a schematic representation of a cross
section of the imogolite structure in Figure III.8.
Deconvolution analysis of the resonances in the 20 wt. %
APS-modified imogolite 29Si MAS NMR spectrum from Figure
III.5A, indicate that 36 percent of the imogolite silicon
sites were coupled to APS. The mole ratio of APS silicon
sites to coupled imogolite silicon sites was found to be

approximately equal to 1, indicating that SiOH surface

64

Cibbsito b

I imogolite 217/11 ' I

I". .- .“‘| 11'

' C
C O (91% 01.51.23.732:
. O C Q o .

 

 

on Si
00 AI
00 o
.0 OH

 

Figure III.8. Schematic representation of a cross
sectional view of the imogolite structure. The tube walls
are composed of five tiers of atoms. Reading from the
inner to the outer wall, the tier compositions are HO, Si,
0, Al, and OH, respectively. The diameter of the channel,
8.7 A, is sufficiently large to accomodate APS as a
silylating agent.

65

coupling is favored. over' AlOHi surface coupling at this
polymer loading. Both the NMR and the hydrolysis data Show
that at low' polymer loading, APS polymer’ is covalently
bonded to SiOH sites within the imogolite channel in APS-
modified imogolite.

The reaction of Na-montmorillonite in aqueous
suspension with APS-modified imogolite at pH 3.7 yielded a
TSLS product with a very diffuse x-ray pattern. Air-dried
Na-montmorillonite normally exhibits a pattern with several
rational orders of 001 reflections corresponding to a basal
spacing of 12.7 A. Therefore, the diffuse pattern observed
for the APS-modified TSLS complex is indicative of the
formation of an interstratified intercalate. Washing the
interstratified product with water results in the loss of
some APS and the formation of ordered TSLS products having
a 35 A basal spacing. This basal spacing corresponds to
the sum of the van der Waals thickness of the clay layer
(10 A) and the hydrated imogolite tubes (25 A). Although
APS remains bonded to the imogolite tubes, as evidenced by
characteristic vibrations at 2926 cm'1 and 1570 cm'l, the
presence of a 35 A basal spacing for the APS-modified
imogolite/clay intercalate precludes the possibility of APS
binding to the external AlOH surface of the tubes. This
result provides further evidence for the binding of APS in
the internal channel of imogolite at low APS loadings.

The lability of APS at the external AlOH surfaces of

imogolite complicates the use of APS-modified imogolite as

66

a precursor to microporous TSLS derivatives. Thus far we
have been unable to utilize the functionalized tubes as a
means of uniformly spacing the tubes as pillars in the
galleries of smectite clay. A more viable approach
currently being investigated, involves the use of
organophosphates or organophosphonic acids coupling agents
for modification of the external AlOH surfaces of
imogolite. The results obtained from these latter studies

will be reported elsewhere.‘

LI ST OF REFERENCES

10.

11.

12.

13.

14.

LIST OF REFERENCES

Moses, P.R.; Wier, L.M.: Lennox, J.C.: Finklea, H.O.;
Murray, R.W. Analyt. Chem. 1978, 50, 576.

silylated Surfaces: Leyden, D.E.: Collins, W.T., Eds.;
Gordon and Breach: New York, 1980.

Silanes, Surfaces and Interfaces: Leyden, D.E., Ed.;
Gordon and Breach: New York, 1985.

Chemically Modified Oxide Surfaces: Leyden, D.E.;
Collins, W.T.: Lochmuller, C.H., Eds.; Gordon and
Breach: New York, 1990.

Sung, C.S.P.: Lee, S.H.: Sung, N.H. Polym. Sci.
Technol. 1980, 12B, 757.

Garbassi, F.: Occhiello, E.: Bastioli, C.: Romano, C.:
J. Colloid Interface Sci. 1987, 117, 258.

Plueddemann, E.P. silane Coupling Agents: Plenum: New
York, 1982.

Kulkarni, R.B.: Goddard, E.D. Int. J. Adhes. Adhes.
1980, 1, 73.

Furukawa, T.: Eib, N.H.: Mittal, R.L.: Anderson,
H.R., Jr. J. Colloid Interface Sci. 1983, 96, 322.

Johnson, L.M.: Pinnavaia, T.J. Langmuir 1990, 6, 307.

Cradwick, P.D.G.: Farmer, V.C.: Russell, J.D.: Masson,
C.R.: Wada, K.: Yoshinaga, N. Nature Phys. Sci. 1972,
240, 187.

Johnson, I.D.: Werpy, T.A.: Pinnavaia, T.J. J. Am.
Chem. Soc. 1988, 110, 8545.

Farmer, V.C.: Fraser, A.R. Dev. Sedimentol. 1978, 27,
547.

Maciver, D.S.: Tobin, H.H.: Barth, R.T. J. Catalysis

‘ 1963, 2, 485.

67

15.

16.

17.

18.

19.

20.

21.

22.

23.

68

Caravajal, G.S.; Leyden, D.E.: Maciel, G.E. In Silanes
Surfaces and Interfaces: Leyden, D.E., Ed.; Gordon and
Breach: New York, 1985: pp 283-303.

Sudhoter, E.J.R.: Huis, R.: Hays, G.R.: Alma, N.C.M.
J. Colloid Interface Sci. 1985, 103, 554.

Barron, P.F.: Wilson, M.A.: Campbell, A.S.; Frost,
R.L. Nature 1982, 299, 616.

Goodman, B.A.; Russell, J.D.: Montez, B.: Oldfield,
E.: Kirkpatrick, R. J. Phys. Chem. Miner. 1985, 12,
342.

Cradwick, P.D.G.; Farmer, V.C.: Russell, J.D.: Masson,
C.R.; Wada, K.: Yoshinaga, N. Nature Phys. Sci. 1972,
240, 187.

Chiang, C.H.: Ishida, H.: Koenig, J.L. J. Colloid
Interface Sci. 1980, 74, 396.

Engelhardt, G.: Michel, D. High-Resolution Solid-State
NMR of silicates and Zeolites: Wiley: New York, 1987;
p 211.

De Haan, J.W.; Van Den Ven, L.J.M. J. Colloid
Interface Sci. 1986, 110, 591.

Rudzinski, W.E.: Montgomery, T.L.: Frye, J.S.:
Hawkins, B.L.: Maciel, G.E. J. Chromatogr. 1985, 323,
281.

CHAPTER IV

Reaction of Gamma-aminopropylphosphonic Acid with Imogolite
and Na-montmorillonite

A. INTRODUCTION

Imogolite is a tubular aluminosilicate with a
molecularly regular outer diameter of 23 A, an internal
channel diameter of 8.7 A and a length of 2000 to 3000 A.1
We have recently shown that the tubular structure can be
directly intercalated as a regular monolayer into the
galleries of smectite clays to form a new class of
microporous pillared clays.2 Nitrogen adsorption
measurements and selective adsorption studies showed that
the microporosity of the material originated primarily from
the void volume provided by the imogolite intra-tube
channel.3 Relatively little void volume was provided
between the tubes due to their tendency to cluster in van
der Waals contact. The imogolite intercalated
montmorillonite complex was denoted tubular silicate-
layered silicate or TSLS. TSLS H-bonding interactions, and
the close packing of the imogolite tubes resulted in
thermal stability enhancement. of the intercalated tubes
(>400°C) relative to pure imogolite (350°C).

More recently, we have become interested in modifying
the external surfaces of imogolite as a means for the
rational design of new TSLS complexes. Specifically, the
intercalation of a uniformly surface-modified imogolite
pillar into the galleries of Na-montmorillonite was thought

to introduce lateral separation between the tubes, thereby
69

70

increasing the microporosity of the complex. Our first
attempts at the surface modification of imogolite utilized
organosilane coupling agents which readily form covalent
bonds ‘with. hydroxylated. surfacesw4' The results of ‘the
coupling reaction between hydrolyzed gamma-
aminopropyltriethoxysilane (APS) and the hydroxylated
surfaces of imogolite have been reported elsewhere.5
Hydrolyzable aminoalkylsilanes were attractive reagents due
to their ability to afford monomeric trisilanols having
protonated amino groups from acidic solutions. These
protonated amino groups ensure that the APS-modified
imogolite retains the net positive charge necessary for
subsequent intercalation reactions. The coupling between
APS and the AlOH sites on the external surfaces of
imogolite was found to be too labile to allow the formation
of an ordered, APS-modified TSLS complex.

In the present work, we investigated the surface
reaction chemistry of imogolite and gamma-
aminopropylphosphonic acid. Effective coupling has been
demonstrated between phosphorus acids and alumina
surfaces.6"9 The aminoalkylphosphonic acids, like the
aminoalkylsilanes, possess protonatable amino groups to
allow the coupled imogolite surface to retain a net
positive charge. The products obtained by the reaction of
gamma-aminopropylphosphonic acid with Na-montmorillonite

were characterized in addition to the products of the

71

reaCtion between gamma-aminopropylphosphonic acid-modified

imogolite and Na-montmorillonite.

B . EXPERIMENTAL

Imogolite was synthesized according to the method
described by Farmer and Fraser10 and purified by dialysis
against deionized water. The gamma-aminopropylphosphonic
acid was purchased from Sigma and used without further
purification. Na-montmorillonite, (SWy-l, Source Clay
Minerals Repository) with a cation exchange capacity of
88 meg/100 g, was purified by sedimentation.

Reactions of gamma-aminopropylphosphonic acid with Na-
montmorillonite

An 18 mM solution of APP in deionized water was
prepared at an ambient pH of 3.9. A portion of a
1.33 wt. % aqueous Na-montmorillonite suspension was added
to enough APP solution to achieve a 1:1 (w/w) ratio of
clay:APP, corresponding to 5.9 moles APP per equivalent of
clay. The final pH of the suspension was 5.3. After a
reaction time of 16 hours, the suspension was poured onto a
large glass plate and allowed to evaporate in air at room
temperature. The unwashed reaction products were recovered
from the plate. A second reaction of APP with Na-
montmorillonite was carried out under identical
stoichiometric conditions, but. the pH of the 18 mM APP
solution was adjusted to 7.5 with 1 M NaOH. After a
reaction time of 16 hours, the suspension was allowed to

air dry and the unwashed product was recovered.

72

Reactions of gamma-aminopropylphosphonic acid with
imogolite

A 0.1 wt. % suspension of imogolite was combined with
an amount of solid APP sufficient to achieve an
APP:imogolite weight ratio of 0.43:1, corresponding to 1
mole of APP per mole of aluminum. Half of the resulting
suspension at pH 5.3 was adjusted to pH 2.6 with 1M HCl.
The remaining portion was adjusted to pH 7.5 with 1M NaOH.
The suspensions were allowed to stir overnight, poured onto
glass plates and air-dried at room temperature.

Reactions of gamma-aminopropylphosphonic acid treated
imogolite with Na-montmorillonite

A volume of 1.33 wt. % Na-montmorillonite suspension
was added to a suspension of APP treated imogolite, as
prepared above at pH 7.5, to obtain a weight ratio of 2.5:1
imogolite to montmorillonite. The mixture was stirred
overnight. The reaction products were then separated from
the supernatent by centrifugation, resuspended in deionized
water and air-dried.

Physical Measurements

FTIR spectra of solid samples prepared as 2 wt. % KBr
pellets were taken over the frequency range from 4000 cm"1
to 400 cm"1 on a Nicolet model IR42 spectrometer with a
resolution of 4 cm‘l. Difference spectra were obtained by
weighted digital subtraction of absorbance spectra over the
entire frequency range. Baseline corrections were not

needed.

73
The 1H decoupled 31P MAS NMR and 2931 MAS NMR spectra

were obtained on a Varian VXR 400 spectrometer equipped
with a Doty probe. The 31P spectra were generated at a
frequency of 161.903 MHz with a 90 degree pulse of 4.8
microseconds, and a relaxation delay time of 60 seconds at
sample spinning rates of 2 to 4 kHz. Chemical Shifts were
reported with reference to 85% H3PO4. The 2gsi spectra
were generated at a frequency of 79.459 MHz with a 90
degree pulse of 6.7 microseconds and a relaxation delay
time of 120 seconds. Chemical shifts were reported with
respect to TMS.

X-ray diffraction patterns were obtained from oriented
films of the suspensions air-dried on the surface of a
glass slide. A Rigaku Rotaflex model RU-200BH
diffractometer equipped with a copper target (Cu Ka
radiation) and the Rigaku D/Max powder diffraction system
was used to measure basal spacings.

Elemental analyses were performed by Galbraith
Laboratories in Nashville, Tennessee.

Three-point nitrogen BET surface areas were measured

on a Quantachrome Quantasorb Jr. sorptometer.

C. RESULTS AND DISCUSSION

Gamma-aminopropylphosphonic acid (APP)
Gamma-aminopropylphosphonic acid (APP) is known from

x-ray crystallographic analysisl1 to adopt a neutral

zwitterionic structure, +NH3(CH2)3PO3H‘, in the solid

74

state. In solution, four Species at different levels of
protonation can be obtained. The stepwise acid
dissociation constants for the three protonated species
reported by Appleton et al.12 are listed in Table IV.I. A
reference to the abbreviations adopted for the
identification of these species is also included in Table
IV.I. The species of interest are identified by the
letters APP followed by the net charge on the molecule.
For example, +NH3(CH2)3PO3H2 is denoted APP+, and
+NH3(CH2)3PO32' is designated APP'.

The characterization of the APP treated materials was
approached in this manner. First, the equilibrium solution
APP species and the APP species isolated as reaction
products were identified. Next, the APP reactant and
product species were compared. Finally, a description of
the chemical interactions between the APP and mineral
reactants which might be responsible for the observed
differences was provided” The complementary structural
information provided by FTIR and NMR spectroscopy allows
differentiation between the four APP species, as well as
the identification of surface-bound species. The important
FTIR absorbances and 31P chemical shifts for each APP
species isolated as a solid were identified.

Table IV.II lists selected FTIR. absorbances for a
2 wt. % KBr pellet of the solid gamma-aminopropylphosphonic
acid as received. The ambient pH of an 18 mM solution of

this material was 3.6. FTIR absorbances also were reported

75

Table IV.I. Acid dissociation constants and 31F chemical
shifts for gamma-aminopropylphosphonic acid species in
aqueous solution and isolated in the solid state.

31P Chemical Shifts (ppm)

APP SPECIES Abbrev. Solutiona Solid Stateb pKaa
if"?3ESQ;3138;};‘-"LIES?""VEST-"m""3Smumfé
+NH3(CH2)3P03H' APP 23.85 28 6.8
+NH3(CH2)3P032' APP‘ 20.55 24 11.0

NH2(CH2)3P032' APPZ' 22.65 26 -

aThese values obtained from reference 12.
hese values are from this work.

76

mac—co.— .e ESE—ob .m $55.05... .>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82.5.7. n a... ”week; b2, u a.» ”EB—02: u E 333 u 3
.2 2
Echo 233.858— Bmwfl Bane. -NAE< EOaZ
a8... .62. 322 3%. IE... n.»
33 .ano 8:2: 2%.: 58.: 52: 3%: .522 .582 .336. .72 man
38. em
263 Eng... 8.: :58. 22: $8: .542 3:2 .332 .3... .535.
3.8. an... en
2mm... 82.68. So. :52: 3mm: 2%: .542 38. .53....— +AE< .0:
$5.: 2.6.: 3.2;
2.0.52 .85., .28.. Goa: 205.. 3.5.. .85.. Sum; :58 3.25 89.2 Shun“; me
. Q
.329.
In 328:6 H .8238 :8: “.2382 an... 2.8 Low $-58 38:25on E ensue—om .=.>_ 63.2.

77
for ‘the solid. gamma-aminopropylphosphonic acid recovered
from aqueous solution at pH 2.6 in the presence of HCl, and
at pH 7.5 and 12 in the presence of NaOH. The
identification of APP species in solution at pH values of
2.6 and 7.5 was necessary for the characterization of
subsequent reaction products. The identification of
equilibrium species isolated at. pH 12 was included to
confirm diagnostic bands. The assignments of the
absorption bands in Table IV.II parallel those made by
Garrigou-Lagrange and Destrade13 among other workers.14'17
The bands at 1160 cm'1 and 925 cm"1 correspond to the
V(P=O) and V(POH) stretching frequencies diagnostic of
neutral APP species. These bands are present in the
spectrum of the solid as received, and in the spectrum of
the sample isolated from a solution which had been adjusted
to pH 2.6 using HCl. These absorbances were also observed
in the spectrum of the solid isolated from a solution which
had been adjusted to pH 7.5 using NaOH. The composition of
this sample was a mixture of both neutral APP and APP"
species. As a result, a more complex spectrum was observed
for the sample isolated at pH 7.5 than the spectra observed
for the samples isolated at pH 3.6 and 2.6, composed
chiefly of neutral APP. The APP‘ fraction of the sample
formed at pH 7.5 was responsible for the additional very
strong absorbances observed at 1070 cm'1 and 970 cm'l,
diagnostic of symmetric V(PO3) vibrations. These

assignments were confirmed by bands in the spectrum of the

78

solid isolated from a solution which had been adjusted to
pH 12 with NaOH. This sample exhibited absorbances due to
the APPZ' and APP' species. Almost all traces of neutral
APP disappeared. This was best illustrated by the absence
of a strong V(P=O) vibration, typical of highly asymmetric
P03 groups, in the spectrum. Intense bands diagnostic for
symmetric ‘V(PO3) groups ‘were observed. at 1070 cm”1 and
970 cm’l. A new band, also corresponding to a V(PO3)
vibration at 1053 cm'l, could not be unambiguously assigned
to either a symmetric or an asymmetric vibration.

A Single anomalous band was included in Table IV.II.
The band at 1240 cm”1 was assigned to a 6(OH) vibration.
Other workersls'7 also observed this band. We tentatively
assigned the band to a 6(OH) vibration here, because the
intensity' of the Iband. decreases ‘with increasing' pH, as
indicated in table IV.II.

1H decoupled 31P MAS NMR spectra confirm the
identities of the APP species derived from the FTIR results
for the four samples isolated at different pH. Table IV.I
lists the solution and solid state (this work) 31P chemical
shifts for APP at each level of ionization. The chemical
shift values reported for the solution species were
determined by Appleton et 81.12 Chemical shifts obtained
for the solid samples led to the assignment of 33, 28,_and
24 ppm to the APP+, APP, and APP“ species, respectively.
The chemical shifts observed for the species isolated as

solids were shifted downfield approximately 3 ppm from the

79

literature solution values. These Shifts may arise from
stronger hydrogen bonding between NH3+ and PO3H’ groups
because of their close proximity in the crystalline state.

The 1H decoupled 31p MAS NMR spectrum of the APP solid
as received is shown in Figure IV.1. Two major phosphonic
acid environments with chemical shifts of 33 and 28 ppm,
and three minor environments having chemical shifts of 5.2,
0.7 and 0.4 ppm probably due to phosphate impurities, were
resolved in this spectrum. The intense resonances were
assigned to APP+ and neutral APP, respectively. A 31p
chemical shift of 25.2 ppm has been reported18 for neutral
APP. The hygroscopic nature of the APP solid may give rise
to slight variations in chemical shift with hydration. The
chemical shift of the APPZ‘ species reported in Table IV.I
was not determined, but was estimated at 26 ppm from the
solution value.
Reactions of APP with imogolite

Evidence for coupling between APP and imogolite was
obtained from 1H decoupled 31P MAS NMR spectra of the air-
dried APP-modified imogolite reaction products. An amount
of APP sufficient to provide an equimolar ratio of
phosphorus to aluminum was used to form these products.
The APP solid was added directly to imogolite suspensions
which were immediately adjusted to pH 2.6 (APPIMOGA) and pH
7.5 (APPIMOGB) with HCl and NaOH, respectively. These pH
values were chosen to maximize the concentration of a

single phosphonic acid species in solution while

80

28

 

33 ".7

 

 

W

70 50 30 10 -10 ppm

Figure IV.1 1H decoupled 31P MAS NMR spectrum (161.903

MHz) of solid gamma-aminopropylphosphonic acid as received,
obtained with a pulse width of 4.8 usec, a relaxation delay
time of 60 seconds and a sample spinning rate of 4KHz.

81

maintaining the structural integrity of the imogolite (and
Na-montmorillonite in subsequent reactions). In this way,
the products formed from neutral APP and imogolite
reactants (APPMONTA) could be compared with those formed
between APP" and imogolite (APPMONTB). This information
may elucidate the role of pH in the formation of the
ordered products obtained by reaction of Na-montmorillonite
suspensions with APP—modified imogolite suspensions at pH
7.5.

FTIR difference spectra were obtained for the APP-
modified imogolite reaction products at each pH. Digital
subtraction of the imogolite FTIR spectrum from the APP-
modified imogolite spectra eliminated interference from
strong imogolite SiOAl bands in the APP diagnostic region.
A list of the important vibrations obtained from these
difference spectra is given in Table IV.III. Diagnostic
bands for both symmetric P03 and asymmetric P03 groups were
observed for APPIMOGA and APPIMOGB samples, suggesting the
presence of multiple APP species in the products.

Two different phosphorus sites were detected in the
31P MAS NMR spectrum for both of the APP-modified imogolite
samples. The phosphorus chemical shifts for the APP
treated clay systems are listed in Table IV.IV. Two
phosphorus chemical shifts, at 23 ppm and 19 ppm, were
observed for the APP-modified imogolite at pH 2.6. The
23 ppm resonance was shifted upfield #5 ppm from the

chemical shift obtained for the solid APP neutral species

82

wee—ooh d ””523 .w $5.685 .>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

52.8.; n :m "wees... be; u m> ”985 u m "Eamon... u E ”ace? u 3
3N3 .3 age A...

is: 2.2.: 2:2. .63. LE... 22:92
38 .2 .3 3mm. LE< anz 2.

some .38 232.28.: .32: 32. .382 .58. .72 $202.5.
.. m 8&8. .. a 314.2 on

258 258.282: .52: new: 52: 83.2 522 .33. me... «.5028...

a. 1:2 $10.3

52a 8.3 2:8. .86.: 3.2. .532 382 .72.. 8025.2

3.3. mammal

some as... 8.8... 2...... >53. 532 .35.: an... 2.82.9.4
$6.: 3.8.: 3.2: =

2.9.: .88., .6: aces 22:32: .85. 81.: $9.. 3.2: See. Emma...” easem—

 

HEQALA vogue—adage. com 90:3 .55 3-88 865305 E ensue—om

.=—.>— 035.

83

Table IV.IV. 1H decoupled 31P MAS NMR resonances observed for
APP-modified systems.

Dominant 311’ Chemical Shifts (ppm)

' Species

APPIMOGA APP 23 19
pH 2.6
APPIMOGB APP/APP' 20 16
r H 7.5
‘ APPMONTA APP
‘ pH 5.3 .

APPMONTB APP/APP'
pH 7.5

APPIIMSS APP-

H 7.5

 

 

 

 

 

 

 

 

 

a pH was adjusted using aqueous 1M HCl or IM NaOH solutions.

84
at 28 ppm, cf. Table IV.I. This chemical shift difference

suggests that APP has been coupled to the external surfaces
of the imogolite tubes.

The chemical shift of the APP coupled to the external
AlOH surfaces of imogolite at pH 2.6 was assigned to the
23 ppm resonance. The resonance at 19 ppm was assigned to
neutral APP in close proximity to the internal SiOH
surfaces of imogolite. 31F chemical shifts for silicon
coupled phosphorus sites are observed downfield from those
of aluminum coupled phosphorus sites. Interestingly, the
chemical shift of the externally coupled APP occurs near
the solution value (cf. Table IV.I) for the neutral APP,
the dominant species in solution under the reaction
conditions. Chemisorption of the negatively charged
phosphonate to surface imogolite A1(HZO)+ sites, abundant
at the reaction pH, may mimic the neutral APP phosphorus
environment in solution, resulting in a solution-like
chemical shift for the surface-bound APP. The surface
immobilization of neutral APP‘ leads to the dissociation of
APP+ to reestablish equilibrium. The neutral APP, with a
length of :36 A, can diffuse into the 8.7 A internal
imogolite cavity. The intercalation of neutral salts into
the imogolite channel has been observed by other workers.19
Changes in bond lengths and bond angles of the P-o
linkages, known to effect drastic changes in chemical
shift,18 may be caused by neutral APP confinement within

the hydroxylated channel. In view of these facts, the

85

19 ppm chemical shift observed for the APP species proposed
to occupy the cavity is plausible. The 29Si MAS NMR
spectrum of the APP-modified imogolite products shows a
single resonance at -79 ppm due to the uncoupled imogolite
silicon sites. This confirms the 31P NMR result. Covalent
bond formation between APP and the SiOH sites located
within the imogolite channel would have prompted a more
dramatic chemical shift difference between the resonances
of the solid APP and the APP-modified imogolite.

The APPIMOGB sample prepared at pH 7.5 exhibited
characteristics similar to those of the APP-modified
imogolite prepared at pH 2.6. The 31P MAS NMR spectrum
contains two phosphorus environments of approximately
equivalent intensities, having chemical shifts of 20 and
16 ppm. At pH 7.5, the imogolite surface retains a net
positive charge, the isoelectric point being pH 9.5. The
dominant APP species in solution under the reaction
conditions were APP" and neutral APP. The resonance
having a chemical shift of 16 ppm was assigned to the
neutral APP intercalated into the imogolite channels in a
manner analogous to the APPIMOGA products. Chemisorption
of the APP" species to the positively charged Al(H20)+
sites through the phosphate group as observed in the
APPMONTA products was still achievable. Charge balance
could be maintained by Na cations in solution. The 20 ppm
chemical shift was assigned to the APP' bound to the

external imogolite surface at pH 7.5. This chemical shift

86

is equivalent to the chemical shift of the APP" species in
solution (cf. Table IV.I), as observed in the APPIMOGA
products. This was again attributed to the effects of APP
immobilization on the imogolite surface.

APP reactions with Na-Iontnorillonite

The pH dependence of surface coupling between APP and
imogolite was coupled with information about APP reactions
with Na-montmorillonite. This was a necessary step toward
the characterization of the reaction products of Na-
montmorillonite and gamma-aminopropylphosphonic acid-
modified imogolite. The addition of a Ila-montmorillonite
suspension to an 18 mM APP solution in an amount necessary
to achieve 5.9 mole APP/clay equivalent at ambient pH
yielded APPMONTA. Similarly, the addition of a Na-
montmorillonite suspension to an 18 ml! solution of APP
which had been adjusted to pH 7.5 with NaOH, in an amount
necessary to achieve 5.9 mole APP/clay equivalent gave rise
to APPMONTB. These products were identified by elemental
analysis, FTIR and 18 decoupled 31F MAS NMR as intersalate
complexes. Multiphase reaction products, including an
unreacted clay phase, were isolated from reactions
performed with reduced APP/clay ratios at either pH.

Shown in Figure IV.2A is the powder x-ray diffraction
pattern of an oriented film sample of the APPMONTA reaction
product. The pH of the solution increased from 3.9 to 5.3
as the reaction proceeded. The 19 A basal spacing was

unchanged after the film was heated at 130'C in air for

87

 

 

 

 

 

 

 
   
 

A 19A
(7')
E unwashed
,_ (A)
3 9A
LIJ
2
E W A
19A 11.511
LIJ
a: (B)
washed 1X
11.4}.
(C)
washed 2X

 

 

r ' T ' T w l r
1.5 3.5 5.5 7.5 9.5 11.5 13.5
DEGREES TWO THETA

Figure IV.2 5 X-ray diffraction patterns of (A) APPMONTA,
(B) APPMONTA washed once with water and (C) APPMONTA washed
twice with water.

88
36 h. X-ray diffraction patterns of the product after

washing with water are shown in Figures IV.ZB and IV.2C.
The intercalated phosphonic acid was almost entirely
removed after the second washing, as indicated by the
appearance of a pure clay phase with a basal spacing of
11.4 A.

The intercalated product formed at pH 7.5 exhibited a
basal spacing of 23 A, appreciably larger than the 19 A
spacing for the product formed at ambient pH. The x-ray
diffraction pattern of an oriented film of APPMONTB is
shown in Figure IV.3A. The basal spacing was reduced to
21 A after the film was heated at 130°C in air for 36 h.
Washing the 23 A phase with water resulted in almost
complete displacement of this phase by a pure Na-
montmorillonite phase with a basal spacing of 11.3 A
(of. Figures IV.38 and IV.3C).

FTIR difference spectra of the 23 A and the 19 A
reaction products of APP and Na-montmorillonite were
obtained by digital subtraction of the Na-montmorillonite
FTIR spectrum from the spectra of the APP-modified clay
samples. This eliminated interference from the SiOSi and
SiOAl clay bands in the region where diagnostic APP bands
are observed. Figure IV.4A shows the difference spectra
obtained over the frequency range from 4000 to 400 cm'l.

The absorbances observed in the difference spectra are

listed in Table IV.III.

 

 

 

 

 

 

 

 

  
 
   
 

”23A

5

E

'_ 1 1 .5 A

.Z.

.1 J

2 _ _ unwashed (A)

3— A

CK
washed 1X (B)
~ - ..1

1 1.3 A

washed 2X (C)

 

 

 

1.5 ' 35.55 535 ' 7T5 ' 91.5 11.5 13.5
DEGREES TWO THETA

Figure IV.3 X-ray diffraction patterns of (A) APPMONTB,
(B) APPMONTB washed once with water and (C) APPMONTB washed
twice with water.

90

 

 

(A)

Relative Absorbance

 

 

 

 

I I I j I I l I I
4000 3500 so'oo 25'00 zo'oo 15'00 1000 500
Wavenumber

Figure IV.4 FTIR difference spectra (KBr Pellets) of
(A) APPMONTA and (B) APPMONTB.

91

The bands observed in the spectrum of the 19 A
intercalate are analogous to those observed for the neutral
APP species (of. Table IV.II) isolated at ambient pH. The
neutral APP species is the dominant species in solution at
the reaction pH. Bands diagnostic of asymmetric P03 groups
located at 1155 cm'1 and 1134 cm'1 due to the V(P=O) and
V(P03) vibrations, respectively, are observed. A very
strong V(POH) band at 930 cm"1 also indicates the presence
of a neutral APP species in the APPMONTA sample.

FTIR evidence for neutral APP is further substantiated
by the 1H decoupled 31P MAS NMR spectrum of the APPMOMTA.
One phosphorus resonance was observed with a chemical shift
of 25 ppm. Although shifted 3 ppm upfield from the
chemical shift assigned to neutral APP in the solid state,
cf. Table IV.I, there is sufficient agreement considering
possible effects of the close proximity of the clay layers.
31P chemical shifts for the APP coupled systems are listed
in Table IV.IV. The presence of the neutral APP species
within the clay galleries was also supported by the
elemental analysis data.

The elemental analysis results for the APP coupled
systems are reported in Table IV.V. These results indicate
that there are 6.24 APP molecules and 0.88 Na+ cations per
montmorillonite sheet face (Sis). A single 31P NMR
resonance indicates that each APP is equivalent. Layer

charge balance is maintained in the APPMONTA system by the

92

Table IV.V. Elemental analysis results for’ the .APP-
coupled Na-montmorillonite and APP-coupled imogolite
intercalated montmorillonite samples.

Mblecules or atoms per

SAMPLE montmorillonite Sis unit
APPMONTA 6.24 APP

0.88 Na

APP/Na = 7
APPMONTB 6.00 APP

3.92 Na

APP/Na = 1.5
APPIIMSS 1.76 APP

1.60 Na

APP/Na,=

.1
APP/A11"?Ohm = o . 028

93

Na+ cations, requiring the intercalated APP units to be
electrically neutral.

The bands observed in the FTIR difference spectrum of
the 23 A phase synthesized at pH 7.5 parallel those
observed in the spectrum of the pristine APP isolated at pH
7.5, except for the absence of the V(P=O) band. Figure
IV.4B shows the FTIR difference spectrum obtained for
APPMONTB. Important bands are listed in Table IV.III.
FTIR results suggest that both neutral APP and APP"
species, species dominant in solution under the reaction
conditions, are prominent in this sampLe. The absence of
the V(P=0) band, known to be very sensitive to its
environment,20 is not unusual.

The 31p NMR spectrum of APPMONTB surprisingly
contained only one phosphorus resonance with a 23 ppm
chemical shift. A single resonance may have been observed
because of proton exchange between the neutral APP and APP"
that is too rapid to allow differentiation between the two
species on an NMR time scale. In this event, the observed
chemical shift should be intermediate between the
resonances of the two exchanging sites, assuming equivalent
residence time for the proton at each site. The chemical
shift of the neutral APP residing in the gallery was
25 ppm, cf. Table IV.IV. Assuming the effect of the clay
layer on APP" is similar to that observed for APP, the
'chemical shift of intercalated APP" should be 21 ppm,

(shifted 3 ppm upfield from solid APP"). The chemical

94

shift expected for fast proton exchange between these two
sites is 23 ppm.

The presence of both neutral APP and APP" species in
APPMONTB is supported by elemental analysis results,
cf. Table. IV.V. These results indicate the availability
of sufficient Na", after utilizing 0.88 moles to balance
the layer charge, to accommodate only 3 moles of APP" per
Sig unit. Since there are 6 moles of APP per Sig unit,
three moles of APP are required to be electrically neutral.
APP-modified Inogolite/Na-Iontnorillonite System

The x-ray diffraction pattern of the product obtained
by reacting APP-modified imogolite with a .Na-
montmorillonite suspension at pH 7.5 and washing once with
water, is shown in Figure IV.5. The products were isolated
from. reactant ratios of 0.43:1 (w/w) APP:imogolite and
2.5:1 (w/w) unmodified imogolite:c1ay. The APP
concentrations, pH and washing procedure were critical to
single phase formation. The x-ray diffraction pattern of
this product, designated APPIIMSS, is characterized by a
single reflection at a two theta value corresponding to
55 A. No reflections were visible in x-ray diffraction
patterns of the unwashed material. Pristine APP was found
to be x-ray amorphous.

A second reaction was performed at pH 8 on a larger
scale. The relative concentrations of the reactants
remained the same as in the previous reaction. The x-ray

diffraction pattern of the products obtained from this

95

 

55A

 

 

L ' (A)

 

RELATIVE INTENSITY
('—

 

 

(B)

 

 

 

0.5 ' 2f5 ' 435 ' 6:5 ' 835 '10.5 12.5 14.5
DEGREES TWO THETA

Figure IV.5 x-ray diffraction patterns of (A) APPIIMSS
and (B) APPIIM55 after heating at 130°C in air for 48 h.

96

reaction is shown in Figure IV.6. These products will be
referred to as APPIIM75 due to the 75 A reflection
exhibited by the material. A minor second phase was also
observed in these reaction products, exhibiting several
orders of reflection, having a basal spacing of 36 A. A
36 A basal spacing is typical of an unmodified imogolite
intercalated montmorillonite (TSLS) phase. This basal
spacing corresponds to the sum of the van der Waals
dimensions of the clay layer (10 A) and the hydrated
imogolite tubes (z25 A). The APPIIM75 material has a
composition comparable to that of the APPIIMSS material,
but a different distribution of components, as a result of
secondary phase formation.

Elemental analysis results from the APPIIMSS sample,
listed in Table IV.V, indicate an APP:Na mole ratio of 1.1.
This ratio, coupled with the reaction pH, suggests the
presence of APP" in the complex. The 1H decoupled 31P MAS
NMR spectrum of APPIIM75 shown in Figure IV.7 supports this
conclusion. The spectrum consists of an intense phosphorus
resonance with a 21 ppm chemical shift, a value
intermediate between the 20 ppm resonance observed for the
externally bound APP" in APPIMOGB and the 23 ppm resonance
observed for the fast exchanging APP/APP" in APPMONTB,
cf. Table IV.IV. A weak resonance was also observed in
this spectrum at 3.2 ppm, probably due to the impurities in
the starting material. The absence of a 16 ppm resonance

in the 31P NMR spectrum of this material indicates that

97

 

 

 

F75}.
E
U)
2
LIJ
F-
E
LIJ
2
p.
S
LIJ
CK
37A

 

 

 

0.5 ' 235 . 4T; 6T5 . 535 315.535.5115
DEGREES TWO THETA

Figure IV.6 X-ray diffraction pattern of APPIIM75.

98

21

 

 

 

 

ppm

Figure -IV.7 1H decoupled 31F MAS NMR spectrum
(145.76 MHz) of APPIIM75, obtained with a 3 usec pulse
width and a relaxation delay time of 62 sec. (Spectrum
courtesy of Kirk Schmitt, Mobil Research and Development
Corporation).

99

washing removes APP from the internal channels of the
tubes. FTIR difference spectra obtained for these samples
provided no diagnostic bands because of low APP
concentrations and the absence of an appropriate reference
spectrum.

Additional information on the APPIIM samples was
provided by elemental analysis and the 29Si MAS NMR
spectrum for the material. The 29Si MAS NMR spectrum was
obtained for the APPIIM75 sample. Table IV.VI lists the
ratio of imogolite silicon (chemical shift -79 ppm)21r22 to
clay silicon (chemical shift -93 ppm). The areas
corresponding to each 295i resonance were determined from
deconvolution analysis of the spectrum. An imogolite to
clay ratio of 9:1 (w/w) was calculated from the NMR data,
using 12 unit cells per imogolite ring.

The imogolite:c1ay ratios calculated for the 75 A
products were used to determine that there are 1.76 moles
of APP per Sis in the APPIIMSS sample. The formation of an
APPMONT phase required 6 moles of APP per Sis. Therefore,
the increase in basal spacing from 36 A in the TSLS complex
to 55 A in the APPIIMSS sample is not likely to be due to
the formation of an imogolite intercalated APPMONTB phase.
Similarly, the concentration of APP in the APPIIMSS sample
is insufficient to provide for an increased basal spacing
due to APPIMOGB intercalated montmorillonite. The

APP:imogolite ratio allows for only ten percent of

100

Table IV.VI. Ratio of the areas obtained by deconvolution
analysis of 2gsi NMR resonances at -79 ppm (imogolite) and
-93 ppm (Na-montmorillonite), and the corresponding
imogolite:c1ay wt. ratios.

Siim09/51913Y IMOGOLITE/CLAY

SYSTEM pH NMR RATIO wt. RATIO
TSLS 1 wash 4 0.67:1 1.5:1
TSLS 1 wash 8 3.53:1 8:1

APPIIM75 1 wash 8 3.93:1 9:1

101

monolayer coverage where one APP is coupled with each
imogolite Al.

Clearly, APP plays an entirely different role in the
formation of these materials. Further scrutiny of the 298i
MAS NMR spectrum allows for correlation of the data for the
APPIIMSS system. The 9:1 (w/w) imogolite:c1ay ratio of the
reaction products is larger than the 2.4:1 (w/w) ratio
provided by the initial reaction conditions. This means
the clay has been preferentially removed in the washing
step. A substantial reduction in . APP concentration also
was observed, as evidenced by the change from .43:1 (w/w)
APS:imogolite under the initial reaction conditions to
.028:1 (w/w) in the reaction products.

The expansion of the basal spacing beyond 36 A must be
caused by multilayer intercalation of imogolite tubes. The
APPIIMSS material with a 55 A basal spacing possesses a
45 A gallery height. This allows the intercalation of a
double layer of imogolite tubes with a 23 A diameter. The
experimentally determined 9:1 (w/w) imogolite:c1ay ratio,
however, is consistent with a material in which a quadruple
layer of imogolite tubes arrange themselves between the
Clay layers. Correlation of the x-ray diffraction
patterns, NMR and elemental analyses results is possible
only by assignment of the 55 A reflection to a (1002 rather
than a d001 reflection. The corresponding 110 A d001
reflection results in a gallery height of 100 A. This
gallery height is consistent with the intercalation of a

102
quadruple layer of tubes, as predicted by the imogolite to

clay ratios.

This description was verified by probing the
dehydration behavior of the APPIIM55 material. The x-ray
diffraction pattern of APPIIM55 after heating at 130°C in
air for 48 hours, is shown in Figure IV.58. Heating caused
a dramatic decrease in crystallinity and an increase in the
d001 reflection to a basal spacing of 70 A or a 60 A
gallery height. This spacing is consistent with a
quadruple layer of tubes which have lost an external layer
of water and have been partially collapsed by dehydration.

The characterization of these materials as multiple
layers of imogolite tubes between montmorillonite sheets is
further supported through nitrogen adsorption data. The
surface areas of APPIIMSS and TSLS were measured and found
to be equivalent (2:240 mz/g). The available surface area
for both materials was limited to the internal channels of
the tubes and the spaces between tubes arising from their
close packing between the clay layers. Therefore, surface
area should be independent of the number of rows of tubes
stacked between the clay layer.

Confirmation of the intercalation of close-packed
imogolite multilayers was also obtained by calcination
experiments. The APPIIM55 material should retain an
appreciable surface area after it has been calcined at
350°C due to stabilization of the intercalated tubes toward

dehydroxylation. Imogolite which has not been

103

intercalated, however, undergoes dehydroxylation at this
temperature, resulting in a reduction in surface area
following calcination.

A poorly ordered material with an imogolite:c1ay ratio
of 8:1 (w/w) was formed under reaction conditions in
absence of APP, but otherwise identical to those used to
synthesize APPIIM75. This material, like APPIIM55, had an
imogolite composition in excess of that required to obtain
a monolayer of intercalated tubes. Unlike APPIIM55, this
material exhibited weak x-ray diffraction reflections
typical of the TSLS complex. The poorly ordered reaction
products obtained in absence of APP, were composed of TSLS
and disordered imogolite phases. The surface area of this
material was equivalent to that obtained for highly ordered
TSLS complexes. The effect of calcination on the APPIIMSS
material was compared with the effect of calcination on the
disordered TSLS material. The surface area of APPIIMSS
after calcination at 350'C was 192 mz/g. The surface area
of the disordered TSLS material after calcination at 350'C

was only 105 mz/g.

D. CONCLUSIONS

Highly ordered products were obtained from reactions
between APP-modified imogolite and Na-montmorillonite.
specific reactant concentrations, pH and washing procedures
were critical to single phase formation. These reaction

products were characterized by 2981 and 31P MAS NMR, FTIR,

104

x-ray diffraction, elemental analysis and nitrogen
adsorption measurements. These results provided a
description of the material as quadruple layers of close-
packed imogolite tubes intercalated between the layers of
Na-montmorillonite. The role of APP in the reaction
appears to involve ordering small bundles of imogolite
tubes in solution. Selective removal of Na-montmorillonite
from the reaction with washing occurs both in the presence
and in the absence of APP.

A The reaction of APP with Na-montmorillonite results in
the formation of an intersalate complex. At an ambient
reaction pH of 5.3, the neutral APP is the species which
occupies the gallery. At pH 7.5, in the presence of NaOH,
APP and APP" species, found to undergo rapid proton
exchange, occupied the gallery.

APP was found to chemisorb to the Al(I-120)+ sites on
the external imogolite surface. Dialysis of APP-coupled
imogolite samples against deionized water promoted
desorption of APP from the surface. APP was also found to
simultaneously intercalate into the channels of the

imogolite tubes.

LIST OF REFERENCES

10.
11.
12.
13.
14.

15.

LIST OF REFERENCES

Cradwick, P.D.G.; Farmer, V.C.; Russell, J.D.: Masson,
C.R.: Wada, R.: Yoshinaga, N. Nature Phys. Sci. 1972,
240, 187.

Johnson, I.D.: Werpy, T.A.: Pinnavaia, T.J. J. Am.
Chem. Soc. 1988, 110, 8545. '

Werpy, T.A.; Michot, L.J.; Pinnavaia, T.J. In Novel
Materials in Heterogeneous Catalysis; Baker, R.T.K.;
Murrell, L.L., Eds.; American Chemical Society:
Washington, D.C., 1990.

Plueddemann, E.P. Silane Coupling Agents; Plenum: lNew
York, 1982. '

Johnson, L.M.: Pinnavaia, T.J. Langmuir 1990, 6, 307.

Davis, G.D.; Ahearn, J.S.; Venables, J.D. J. Vac.
Sci. Technol. 1984, A2, 763.

White, R.W.; Crowder, C.D.; Alldredge, G.P. J.
Electrochem. Soc. 1985, 132, 773.

Templeton, M.K.: Weinberg, W.H. J. Am. Chem. Soc.
1985, 107, 774.

Ramsier, R.B.; Henriksen, P.N.; Gent, A.N. Surf. Sci.
1988, 203, 72.

Farmer, V.C.: Fraser, A.R. Dev. Sedimentol. 1978, 27,
547.

Glowiak, T.; Sawka-Dobrowolska, W. Acta. Cryst. 1980,
B36, 961.

Appleton, T.G.; Hall, J.R.; Harris, A.D.; Kimlin,
H.A.: McMahan, I.J. Aust. J. Chem. 1984, 37, 1833.

Garrigou-Lagrange, C.: Destrade, C. J. Chim. Phys.
1970, 67, 1646.

Destrade, C.: Garrigou-Lagrange, C. J. Chim. Phys.
1970. 67, 2013.

Menke, A.G.: Walmsley, F. Inorg. Chim. Acta 1976, 17,
193.

105

16.

17.

18.

19.

20.

21.

22.

106

Fenot, P.; Darriet, J.: Garrigou-Lagrange, C.:
Cassaigne, A. J. Mol. Struct. 1978, 43, 49.

Sahni, S.K.; van Bennekom, R.: Reedijk, J. Polyhedron
1985, 4, 1643.

Harris, R.K.; Merwin, L.M.; Hagele, G. Mag. Res.
Chem. 1989, 27, 470.

Clark, C.J.; McBride, M.B. Clays Clay Miner. 1984,
32, 300.

Thomas, L.C. Interpretation of the Infrared Spectra
of Organophosphorus Compounds; Heyden: London, 1974.

Barron, P.F.; Wilson, M.A.; Campbell, A.S.: Frost,
R.L. Nature 1982, 299, 616.

Goodman, B.A.; Russell, J.D.; Montez, 8.; Oldfield,
E.: Kirkpatrick, R. J. Phys. Chem. Miner. 1985, 12,
342.

"Illllllllllllliflllllr