3 Q‘
‘5‘351‘32 a:
' «I

 

h
gnaw“
a" . J").

. yr 23%..
If? u.
.. A . aw???

. .. .5 . , m&

L um 7.. Anumuai.

49.9"“? at . . ..~ ‘

#1,.

MW . . .,

x

, .mz.

, w
em», 1. Eur

mm». £33
. a... g.»
fun? 1:
«mmfanuum

. Sixtflfififi
ufifiyuus?

4 . . v. EN". .3915...
.. : h..,p.m.f. fiiuflag

.. . .wzrm4é.

V... .

a}... V

 

“.v 3....
r. 5....
\.

:1” 3...“. .1 a

G?

t, \~
. 3&4. 4.
:3... 4f
.
4.?
.4
v... mam.
155...";
V «X Luau
, ‘

 

 

 

 

 

 

 

 

 

 

 

THESIS
2

mo:

This is to certify that the
dissertation entitled

Design, Synthesis and Catalytic Applications
of Mesoporous Silica Molecular Sieves

presented by

Nenzhong Zhang

has been accepted towards fulfillment
of the requirements for

Ph .0. degree in ChemlStl‘y

 

\J

 

/ ”/
Lzyzflyg 1/(zs onazaxnxAAk—-
M

r professor

Date 5/7/99

 

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

 

LIBRARY
Mlchlgan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moo mm.“

DE!

DESIGN, SYNTHESIS AND CATALYTIC APPLICATIONS OF
MESOPOROUS SILICA MOLECULAR SIEVES

BY
Wenzhong Zhang

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1999

DESIGN. f
M]

The dim
snmulazed great
mNPOl’E: ill [hr
of large substran
much larger pen
sits in the inn
reactions. Hexag
Wfitonnecrix ii:
We branching,
and scienriiicali)
Whit sites in:
in mesostrucrure
311d shape of [hi
mmmg siraiecc
Wific pore COE
{OUHEZTOH and h\{

In adding

extremely jmmm

ABSTRACT

DESIGN, SYNTHESIS AND CATALYTIC APPLICATIONS OF
MESOPOROUS SILICA MOLECULAR SIEVES

BY
Wenzhong Zhang

The discovery of the Mobil M418 family of mesoporous molecular sieves has
stimulated great interest in surfactant-directed assembly of inorganic mesostructures. The
mesopores in these molecular sieves provide new opportunities for catalytic conversion
of large substrates in the liquid phase. Even though mesoporous molecular sieves exhibit
much larger pore sizes than conventional zeolites, accessibility to the catalytically active
sites in the framework of the mesopores is still very crucial for diffusion limited
reactions. Hexagonal and cubic mesostructures with uni-dimensional and 3-dimensional
pore-connectivity, respectively, show similar catalytic reactivities, despite differences in
pore branching, most likely both are diffiision limited. Therefore, it is practically useful
and scientifically significant to find some ways to improve the accessibility to the active
catalytic sites inside the mesopores. Regardless of the type of inorganic precursors used
in mesostructure synthesis, the pore size and connectivity of the framework take the size
and shape of the surfactant micelles filling in the structural channels. Therefore, an
promising strategy for the design and synthesis of mesoporous molecular sieves with
specific pore connectivity and accessibility would be to control the surfactant micelle
formation and hydrolysis of the inorganic precursors.

In addition to framework mesopores, complementary textural mesopores are

extremely important for enhancing the accessibility in catalysis. Our first objective was to

tailor W”
claimant
mlartl)" “JG
nucleation l
{166i ll) dill
led to mm"
m to use 0
tailor the f
enhibittng Lt
the flak} su:
mesoscaled
the complcr
accessibilit}
peroxidation
the catalysts.
thingAl-sub:
e“hinted cat
Complementa
the We Chant
What

I
J

and concentra:

,t -
4118.” o i
‘n: the 85

Slim-"n9
“am hfu;

tailor both the framework and the textural porosity of HMS silicas assembled through an
electrically neutral S°I° mechanism. The approach involved control of the solvent
polarity, along with the use of organic auxiliary agent (i.e. mesitylene) to control particle
nucleation rates and pore size. Increasing the polarity of the synthesis medium caused
faster hydrolysis of silicon alkoxide in the presence of the alkyl amine surfactant, which
led to meso-scaled fundamental particles and high textural porosity. Our second objective
was to use organic assembly modifiers such as tartaric, oxalic, gluconic and citric acids to
tailor the formation of MCM-41 through a 3+1" pathway. Well-ordered MCM-41
exhibiting a flake—shaped particle morphology with pore channels running orthogonal to
the flaky surface was obtained. Also, MCM—4l with textural pores in association with
meso-scaled fundamental particles was prepared under careful control. The importance of
the complementary meso-scaled inter-particle (textural) pores for improving the
accessibility of the framework pores in catalysis was demonstrated for the liquid phase
peroxidation of 2,6-di-tert-butyl phenol in the presence of Ti-substituted derivatives as
the catalysts. Also, liquid phase alkylation of 2,4-di-tert-butyl phenol was carried out
using Al-substituted derivatives as the catalysts. By facilitating accessibility, significantly
enhanced catalytic activity was observed for both Al-HMS and Al-MCM-4l with
complementary textural mesopores and meso-scaled fundamental particles comparable to
the pore channel lengths.

Variation of surfactant packing parameter g (= V/aol) by changing the structure
and concentration of surfactant has been reported in literature as an effective approach to
altering the assembled mesostructures. Little is known about chemical modification of

surfactant head group using simple chemicals rather than co-surfactants to ultimately

tailor the r
molecular

successful
distension.
hinthem
ammonium
relatixely

tncorprrnti

introductio

H

mlCt‘ltés a
M's-‘OIAWTOL‘
dlmfilhiom
hexagonal
WIRE“ If;
mediated
Milena.
h‘ 1:03..“ C.

lit ,
m4'veth}ler

tailor the molecular sieves. The concept of altering the structures of mesoporous silica
molecular sieves by organic promoters in electrostatic S'T assembly has been
successfully illustrated. Mesoporous silica molecular sieves with well ordered 3-
dimensional hexagonal, cubic, and uni-dimensional hexagonal phases were
hydrothermally assembled in the presence of organic modifiers using cetyltrimethyl
ammonium bromide as the structure director and sodium silicate as the silicon source at a
relatively low surfactant to silicon ratio (i.e. 0.11). In addition, the concept of
incorporating structural order into mesoporous silica molecular sieves through the
introduction of electrostatic interactions at the interfaces of electrically neutral surfactant
micelles and neutral inorganic precursors has also been successfully demonstrated.
Mesoporous silica molecular sieves with well ordered hybrid domains consisted of 3-
dimensional hexagonal and spherical cubic (3-d-hex-cubic) phases, an uni-dimensional
hexagonal phase, and wormhole structural motifs were assembled using alkyl
polyethylene oxide surfactants and tetraethyl orthosilicate through a new counterion
mediated (N°M*X‘)I° pathway. Our approach explicitly illustrates that silica
mesostructures can be altered using both organic and inorganic modifiers such as
hydroxyl carboxylic acids for quaternary ammonium surfactant and metal salts for alkyl
polyethylene oxide surfactants, respectively. The modifiers most likely bind to the
surfactant head group to cause a significant change in the effective-surfactant-head-
group-area, which alters the packing parameter g and, ultimately, the inorganic

mesostructure.

To my grandparent and parants, my wife and my son.

I Wt
support and
scheme»
example for
met the ye.
if. dellM‘
committee.

lam
[bfi kindness
Wang. Dr. \
Pauly for the

lam.
Oblam XXV
assjs’ifllce ire
and SEM rm;

Mir a
Meme {Or}
5011 Dell W

ACKNOWLEDGMENT

I would like to thank Professor Thomas J. Pinnavaia for his continuously solid
support and invaluable advice over the years in both sciences and life. His attitude and
seriousness toward the pursuing of the truth and beauty of chemistry set a perfect
example for me. I strongly believe that I will greatly benefit from my experiences gained
over the years in his lab. I would also like to thank Professor M.G. Kanatzidis, Professor
J.F. Harrison and Professor S. Garrett for their guidance in serving on my program
committee.

I am so lucky to be a member in such a nice research group. I am very grateful for
the kindness and help from my colleagues. I would especially thank my friends Dr. Zhen
Wang, Dr. Yu Liu, Dr. Seong-Su Kim, Dr. Jialiang Wang, Dr. Tie Lan, and Mr. Thomas
Pauly for their generous help.

I am also very grateful for Dr. J. Wong and Dr. M. Froba for their measurement to
obtain XANES spectra of Ti-substituted mesoporous molecular sieves. Very helpful
assistance from Dr. J. Heckman Jr. and Dr. S.L. Flegler to obtain many attractive TEM
and SEM images is greatly appreciated.

My deepest gratitude goes to my parent, my wife and son for their inspiration and
patience for years to allow me work during weekends and holidays. I am obligated to my
son, Dejia, who sacrificed his enjoyable kid’s time to allow me to stay in the lab rather

than play with him.

vi

LIST OF TABl

LIST OF FlGl'

CHAPTER l f

Abfiil’m

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. iv

LIST OF FIGURES ............................................................................................................. v

CHAPTERI SUPRAMOLECULAR ASSEMBLY OF MESOPOROUS
MOLECULAR SIEVES AND THEIR APPLICATION IN

CATALYSIS ............................................................................................... 1
1.1 Introduction ................................................................................................. 1
1.2 Structural Characteristics of Mesoporous Silica Molecular Sieves
Assembled through Neutral and Electrostatic Pathways ............................. 5
1.3 Synthesis of Mesosporous Silica Molecular Sieves Using Primary
Amine Surfactant as the Structure Directors ............................................... 9
1.4 Catalysis .................................................................................................... 12
1.4.1 General Application of Mesoporous Molecular Sieves ................ 12
1.4.2 Peroxide Oxidation ........................................................................ 13
1.4.3 Ring Opening Polymerization ....................................................... 17
1.4.4 Selective Catalytic Reduction (SCR) of NO ................................. 19
1.4.5 Acid Catalysis ................................................................................ 20
1 .5 References ................................................................................................. 24
CHAPTER 2 MESOPOROUS TITAN OSILICATE MOLECULAR SIEVES
PREPARED AT AMBIENT TEMPERATURE BY
ELECTROSTATIC (S+I’, S+X'l+) AND NEUTRAL (Solo)
ASSEMBLY PATHWAYS : A COMPARISON OF PHYSICAL
PROPERTIES AND CATALYTIC ACTIVITY FOR PEROXIDE
OXIDATIONS .......................................................................................... 27
Abstract ................................................................................................................. 27

vii

 

 

 

l

I l
.l

2.3 ltl
2.4 C.
35 Re
Clltl’lER3 TA
Ol

NE
Abstract
3.1 lntt
3.2 Ext
3.3 Re-
3.4 Dis
3.5 Rel
GWEN [\1
SIE
Ab‘lmcr
4'1 lntr
4.3 EXp
43 Re»
4.3,.

2. 1 Introduction ............................................................................................... 28

2.2 Experimental ............................................................................................. 29

2.2.1 Materials ........................................................................................ 29

2.2.2 Characterization ............................................................................ 31

2.2.3 Catalytic experiments .................................................................... 33

2.3 Results and Discussion .............................................................................. 34

2.3.1 Synthesis ....................................................................................... 34

2.3.2 Characterization ............................................................................ 35

2.3.3 Catalytic results ..... , ....................................................................... 48

2.4 Conclusions ............................................................................................... 52

2.5 References ................................................................................................. 54
CHAPTER 3 TAILORING THE FRAMEWORK AND TEXTURAL MESOPORES
OF HMS MOLECULAR SIEVES THROUGH AN ELECTRICALLY

NEUTRAL (S°I°) ASSEMBLY PATHWAY ........................................... 57

Abstract ................................................................................................................. 57

3. 1 Introduction ............................................................................................... 59

3.2 Experiment ............................................................................ . ................... 61

3.3 Results ....................................................................................................... 63

3.4 Discussion ................................................................................................. 74

3 .5 References ................................................................................................. 8 1
CHAPTER 4 IN-SITU ALUMINATION OF MESOPOROUS HMS MOLECULAR

SIEVES SILICA FOR USE AS ALKYLATION CATALYSTS ............. 82

Abstract ................................................................................................................. 82

4. 1 Introduction ............................................................................................... 83

4.2 Experimental ............................................................................................. 85

4.3 Results and Discussion .............................................................................. 88

4.3.1 Catalyst Synthesis ......................................................................... 88

4.3.2 Catalyst Characterization .............................................................. 89

viii

44
4.5

CHAPTER f

AIM

(J1
I.)

(II
J)

(J!
.L

LII
'JI

4.3.3 Catalytic Results for 2,4-Di-tert-phenol Alkylation .................... 104

4.4 Conclusions ............................................................................................. 107

4.5 References ............................................................................................... 109
CHAPTER 5 TOWARD MCM-4l WITH MESO-SCALED CHANNEL LENGTHS

AND TEXTURAL PORES FOR CATALYSIS ..................................... 111

Abstract ............................................................................................................... 111

5. 1 Introduction ............................................................................................. 1 13

5.2 Experimental ........................................................................................... 1 16

5.3 Results and Discussion ............................................................................ 120
5.3.1 Effect of Organic Acid Addition on Silica Mesostructure

Formation ............................................ . ....................................... 120

5.3.2 Siliceous MGM-41 Type Meso-structure Formation .................. 123

5.3.3 Catalyst Preparation .................................................................. 133
5.3.4 Catalytic Alkylation of 2,4-di-tert-butylphenol with Cinnarnyl

Alcohol ........................................................................................ 141

5.4 Conclusions ............................................................................................. 145

5.5 References ....................................................................................... . ....... 147
CHAPTER 6 INCORPORATION OF STRUCTURE ORDER INTO
MESOPOROUS SILICA MOLECULAR SIEVES ASSEMBLED

VIA A METAL SALT MEDIATED N°(M*X')I° PATHWAY ............... 149

Abstract ........................................................................................................................... 149

6. 1 Introduction ............................................................................................. 150

6.2 Experimental ........................................................................................... 152

6.3 Results and Discussion ............................................................................ 153

6.3.1 Materials Characterization .......................................................... 153

6.3.2 Discussion ................................................................................... 168

6.4 Conclusion ............................................................................................... 173

6.5 References .............................................................................................. 174

Table 1.1
Table 1.2
Table 1.3
Table 1.4
Table 2.1
Tahle 2.2
Table 2.3
Table 3.1
Table 4.1
Table 43

Table 5.1.

T.1
table 5.3.

 

 

Table 1.1

Table 1.2

Table 1.3

Table 1.4

Table 2.1

Table 2.2

Table 2.3

Table 3.1

Table 4.1

Table 4.2

Table 5.1.

Table 5.2.

LIST OF TABLES

Catalytic Peroxidation Activity of Ti-substituted (2 mol%)
Mesoporous Molecular Sieve Silica .......................................................... l4
Lactide polymerization over heterogeneous catalysts ............................... 18

Comparison of the apparent and intrinsic first-order rate constants and
the over all diffusivity for NO SCR reaction over Fe3+ exchanged A1-

HMS and Al-MCM-41 .............................................................................. 20
Cumene Conversions (%) at 300 0C.... ...................................................... 23
Properties of Ti-substituted Mesoporous Silicas Prepared by Ambient

Temperature Synthesis Using Different Templating Pathways ................ 35
Ti K-edge XANES Results for Ti-Containing Materials .......................... 45

Catalytic Activity of Ti-Substituted (2 Mol %) Molecular Sieves
Prepared by Ambient Temperature Synthesis Using Different
Templating Pathways ................................................................................ 49

Physical Parameters for Calcined (650 °C) HMS Molecular Sieve
Silica Prepared by 8010 Assembly in HzO-rich and EtOH-rich Medium .. 75

Physical and Chemical Properties for Al-substituted Mesoporous

Molecular Sieves ..................................................................................... 102
Alkylation of 2,4-di-tert—butylphenol with Cinnarnyl Alcohol
Catalyzed by Al-substituted Mesoporous Molecular Sieves ................... 105

Physical and chemical properties of Al-MCM-4l mesoporous
molecular sieves ...................................................................................... 140

Alkylation of 2,4-di-tert-butylphenol with Cinnarnyl alcohol catalyzed
by Al-substituted mesoporous molecular sieves ..................................... 143

Figare 1.1

Figure 1.2

figure 13

figure 1.4

Figure 1.5

Figure 1.6

figure 2.1

Figare 2)

Figure 1.1

Figure 1.2

Figure 1.3

Figure 1.4

Figure 1.5

Figure 1.6

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

LIST OF FIGURES

Schematic showing of interfacial interactions for surfactant micelles in
cooperatively assembly ............................................................................... 2

X-ray diffraction patterns for mesoporous molecular sieves
assembled via electrostatic (ST) and neutral (S°I°) pathways 6

N2 adsorption and desorption isotherms for mesostructured molecular
sieves assembled via S+l~ and S°I° pathways ............................................... 7

TEM images showing regular order of mesopores for (a)
MOM-41 and (c) MGM-48, but wormhole like channels for (b)
KIT-1 and (d) HMS ........................................................................ 8

TEM image showing wormhole-like framework mesopores
and meso—scaled textural pores indicated by the arrows ........... 10

N2 adsorption and desorption isotherms for HMS prepared in
(A) EtOH-rich medium (H20 : EtOH = 35 : 65 by vol.); (B)
HzO-rich medium (H20 : EtOH = 90 : 10 by vol.) ....................... 11

Powder XRD patterns for calcined pure and Ti-substituted (2 mol %)
silica molecular sieves. (A) MCM-41 and Ti-MCM-41 prepared by
S‘“X'I+ assembly and (B) MCM-4l and Ti-MCM-41 prepared by S+l~
assembly, and (C) HMS and Ti-HMS prepared by S°I° assembly ............ 37

29Si-MAS NMR spectra of as-synthesized pure and Ti-substituted (2
mol %) silica molecular sieves: (A) MCM-41 and Ti-MCM-41
prepared by STXT’ assembly, (B) MCM-41 and Ti-MCM-41 prepared
by STI' assembly, and (C) HMS and Ti-HMS prepared by S°I°
assembly .................................................................................................... 39

UV-VIS spectra of calcined Ti-containing materials ................................. 41

Ti K-edge XANES spectra of reference compounds containing Ti in
tetrahedral and octahedral oxygen environments: (a) anatase, (b) rutile,
(c) MgTiO3, (d) neptunite, and (e) spirotitanate ........................................ 44

Ti K—edge XANES spectra of calcined (a) Ti-HMS prepared by S°I°

assembly, (b) Ti-MCM-4l prepared by STXT’ assembly, (c) Ti-MCM-
41 prepared by STI’ assembly, and (d) TS-l .............................................. 47

xi

 

 

finn3l
Epan
finm33
W
h
d
me34 T
a:
d
p
m

Rants T

fiflm36 X

Hfimt7

gufe .11

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 4.1

N2 adsorption-desorption isotherms for HMS silicas assembled at
ambient temperatures from dodecylamine and TEOS: Curves A and B
are for derivatives obtained from a water-rich (water : ethanol = 90 :
20 (v/v)) and ethanol-rich solution (water : ethanol = 35 : 65 (v/v));
curves C and D are for derivatives prepared from the same water-rich
and ethanol-rich solutions, but in the presence of mesitylene (Mes/8°
= of 1.0). The HK values are the Horvath-Kawazoe pore diameters. ...... 64

N2 adsorption-desorption isotherms for HMS silicas prepared from
tetradecylamine and TEOS in water-rich and ethanol-rich solution.
The labeling of the isotherms is the same as in Figure 3.1 ....................... 65

N2 adsorption-desorption isotherms for a HMS silica prepared from
dodecylamine in water-rich solution in the presence of mesitylene
(Mes/S = 4.5): (A) after calcination at 650 °C with retention of
framework and textural mesopores and (B) after calcination at 1000 °C
with collapse of framework pores but with retention of textural pores.
Insert: Horvath-Kawazoe pore size distribution after calcination at
650 oC ........................................................................................................ 67

TEM images of calcined (650 °C) HMS molecular sieve silicas
assembled from dodecylamine : a. mesoscale fundamental particles
obtained from water-rich solution; b. the spheroid fundamental
particles and macroscale beads-on-a-string texture obtained from
ethanol-rich solution .................................................................................. 69

TEM images of HMS silica prepared from dodecylamine in water-rich
solution in the presence of mesitylene (Mes/S = 4.5) and calcined at
1000 °C to collapse the framework pore structure; (a) Low
magnification image showing the retention of the textural mesopores
and, (b) High magnification image showing the absence of framework
pores .......................................................................................................... 71

X-ray powder diffraction pattern of calcined (650 °C) HMS silicas
assembled from dodecylamine amine in water-rich and ethanol-rich
solution with or without mesitylene as an auxiliary structure director ..... 72

Schematic representation of the structure-directing effects of
mesitylene on HMS assembly. In the water-rich media, the mesitylene
“dissolves” in the hydrophobic central core of the micelle leading to
pore expansion. In the ethanol-rich environment, mesitylene
preferentially “adsorbs” to interfacial head groups, thus increasing
effective head group size and subsequently decreasing the pore size ....... 77

N2 adsorption and desorption isotherms for calcined (650 °C) Al-HMS
molecular sieves with different textural porosity. Each sample was

xii

fimm42

fimm43

P35341651

Figure 4.2

Hym43

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Figure 5.1

prepared at ambient temperature using dodecylamine as the structure
director, TEOS as the silicon source, and using Al(sec-OBu)3 as the
alumination agent, but the alumination agent was added to the reaction
mixture 0, 0.33, and 20 hrs after mixing the TEOS and the surfactant.
The aluminum content in each case is 2.0 mole%. The Horvath-
Kawazoe (H-K) framework pore size, and the ratio of textural to
framework mesoporosity (Vtx/Vfr) is provided for each sample ............... 90

N2 adsorption-desorption isotherms for calcined (650 0C) Al-HMS
molecular seives prepared as described in the caption to Figure 4.1,
except that the alumination agent was sodium aluminate ......................... 91

N2 adsorption-desorption isotherms for calcined (650 0C) 2% Al-HMS
and calcined (540 oC) Al-MCM-41 with zero-textural porosity. The
inset shows Harvoth-Kawazoe pore size distributions for these two
samples. This Al-HMS was prepared using alumnum sec-butoxide as
the alumination agent, which was added to the reaction mixture 20
minutes after mixing the TEOS and the surfactant solution ..................... 93

XRD patterns of calcined 2%Al-HMS molecular sieves prepared as
described in the caption to Figure 4.1 ....................................................... 95

XRD patterns of 2%Al-HMS prepared as described in the caption to
Figure 4.1, except that NaAlO2 was the Al-source ................................... 96

XRD patterns of calcined 2% Al-HMS and Al-MCM-41 with zero-
textural porosity as described in the caption to Figure 4.3 ........................ 97

TEM images showing meso-scaled textural pores as well as worm-
hole like framework meSOpores for calcined 1.5-tex-Al-HMS and 0-
tex-Al-HMS ............................................................................................... 99

27Al—NMR spectra showing predominantly tetrahedral Al-sites (63
ppm) and a small amount (~10%) of octahedral Al-sites (18 ppm) in
as-made Al-HMS prepared using Al(sec-OBu)3 as the alumination
agent. Calcined forms of Al-HMS, as well as calcined Al-MCM-41,
show 70% tetrahedral and 30% octahedral Al-sites ................................ 100

X-ray diffraction patterns for calcined (540 0C) cubic silica meso-
structures formed at 150°C with a reaction composition of 1.0Si :
0.11CTAB : 0.39Na2O : 0.28H+ : 130H2O. The diffraction patterns
for MCM-48 prepared by following Landry’s method as well as the
diffraction patterns for a smaple prepared in the absence of both EtOH
and salicylic acid are included ................................................................. 121

xiii

hmm52

finn53

Hamil

HpmSj

Eflm5f

Emits 5.:

5

Fig-lie 53

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

X-ray diffraction patterns for calcined (540 °C) 3-d hexagonal silica
meso-structures formed under the same conditions as noted in Figure
5.1 with the promotion of citric, oxalic and tartaric acids ....................... 122

N2 adsorption and desorption isotherms for calcined (540 0C) siliceous
MCM-41 prepared under the promotion of gluconic, lactic, oxalic, and
citric acids with a H+INaOH ratio of 0.75 showing framework
mesoporosity as well as clay—like textural porosity similar to that of
montrnorillonite. The Mobil MCM-4l sample was prepared using
sulfuric acid to adjust the pH (i.e. ~10.5) at the very beginning of the
synthesis by control the over-all HTINaOH ratio to 0.625 ...................... 124

TEM images showing (A and B) interwoven tubular MCM-41 formed
upon use of sulfuric acid; (C and D) thin sheet MCM—41 (formed upon
use of citric acid) with pore channels running orthogonal to sheet
surface (the thickness of the sheet estimated to be ~100nm according
to the penetration depth of electron beam accelerated at 120 kV); Both
samples were prepared with an over all H+lNaOH ratio of 0.75 ............ 127

X-ray diffraction patterns for MCM-41 molceular sieves as noted in
Figure 5 .3 ................................................................................................. 128

N2 adsorption and desorption isotherms for MCM-41 molecular sieves
prepared using tartaric acid as the structure promoter under various
conditions showing the controlled formation of significant amount of
HMS-like textural pores as well as clay-like textural pores. Horvath-
Kawazoe pore size distribution plots are included in the inset ............... 130

X—ray powder diffraction patterns for MCM-41 molecular sieves as
noted in Figure 5.6. Notably, the MCM—41 with high textural porosity
exhibits well resolved (100), (110) and (200) reflections ....................... 131

TEM image of MCM-41 showing meso-scaled fundamental particles
with inter particle voids as textural pores and regular hexagonal
framework mesopores running orthogonal to thin sheet surfaces
exemplified by a sample prepared using tartaric acid as the structure
promoter .................................................................................................. 132

N2 adsorption and desorption isotherms for 8.8mol%Al-substituted
MCM-41 molecular sieves (calcined at 540 0C) with clay-like and
HMS-like textural porosity prepared using tartaric acid as the structure
promoter and aluminum nitrate as the follow-up in situ alumination
reagent in the presence of surfactant nriccelles in the pores. (The inset
is the corresponding Horvath-Kawazoe plots showing sharp narrow
distribution of framework mesopores) .................................................... 133

xiv

fgmeSlO

finnill

fined]

Eflmél

fisne63

c

Site 64

Figure 5.10 27Al—NMR spectra for as-made and calcined Al-MCM-41 molecular
sieves prepared using tartaric acid as the structure promoter with
HMS-like and clay-like flaky textural pores ........................................... 136

Figure 5.11 Temperature programmed desorption (TPD) of pyridine pre-adsorbed
on Al-MCM-4l catalysts and purged at 150 0C in N2 stream for 1 h
before the TPD started. This TPD profile was carried out on a thermal
gravity analyzer ....................................................................................... 138

Figure 6.1 TEM images of calcined (600 oC) mesoporous silica molecular sieves
prepared in water using the alkyl polyethylene oxide type surfactant
Brij 56 (C16H33(EO)10H) as the structure director and TEOS as the
silicon source in the presence of CoC12 salt as the structure promoter
under the same reaction conditions (45 °C, 40 h reaction time) except
that the surfactant to Co“ ratios are different. The surfactant to silicon
ratio was 1 to 24 for both samples, but the surfactant to metal ratio was
8.5 for the sample denoted as MSU-C (top photo) and 2.0 for the
sample denoted as MSU-H (bottom photo). These images are provided
as negatives for easier viewing of the pores as black spots. Pore to pore
distance is constantly ~13.7 nm through the specimen for MSU-C. The
insets are the electron diffraction patterns for areas imaged. .................. 154

Figure 6.2 (a). Selected area TEM image of MSU-C; (b). Electron diffraction
pattern of MSU-C; (c). Optic diffraction pattern obtained from the
Fourier-transformation of image (a); ((1). Image obtained by reverse-
Fourier-transformation of the center tetragonal pattern in the optic
diffraction pattern (c); (e). Image obtained from the total reverse-
Fourier-transformation of optic diffraction pattern (c); (f). Image
obtained from the reverse-Fourier-transformation of 7 diffraction spots
selected from pattern (c) in such a way that these 7 spots resembling 7
spots in the real electron diffraction pattern (b). ..................................... 155

Figure 6.3 A TEM segment for MSU-C showing a different orientation of the
structure. The optic diffraction pattern (obtained by Fourier-
transformation of arrow indicated area in the image) is attributable to
both a (01-10)-3-d-hexagonal orientation and a (210) cubic orientation.
These patterns are very similar to the diffraction pattern for meso-
structured silica films with pores running orthogonal to the film
surface, as reported by Brinker et a1. Notably, the pore to pore
distance is ~13.7 nm. ............................................................................... 156

Figure 6.4 X-ray diffraction patterns for the calcined MSU-C and MSU-H
samples that were used to obtain the TEM images shown in Figure 6.1.
Both d-spacings, observed and calculated according to an uni-

XV

Figure 6.5

Figure 6.6

figure 6.7

Figure 6.8

Figure 6.9

81

Ol
tit

 

311

lo;
311
dir
by

111C
8111

C01
131.
Diff
Spa

Figure 6.5

Figure 6.6

Figure 6.7

Figure 6.8

Figure 6.9

Figure 6.10.

dimensional hexagonal phase for MSU-H, are provided for
comparison. ............................................................................................. 159

N2 adsorption and desorption isotherms for calcined MSU-C and
MSU-H. Pore necking hysterisis loop for MSU-C is indicated by the
step at desorption branch, which is ~0.2 unit of relative pressure
behind the adsorption branch. MSU-H exhibits almost identical
adsorption and desorption branches signify the uniformity of uni-
dimensional pores. ................................................................................... 160

TEM images showing the wormhole (top) and 3-d-hex-cubic (lower)
meso-structures assembled at Co2+ : Brij56 molar ratios of 1:13 and l :
8.5 respectively. TEOS was the silicon source and the silicon to
surfactant ratio was 17 to 1 ..................................................................... 161

X-ray diffraction pattern of the calcined (600 °C) meso-structured
silica molecular sieves assembled using Brij 56 as the structure
director and TEOS as the silicon source at a silicon to surfactant ratio
of 15 to 1 under the same reaction conditions (45 °C, 40 h reaction
time), except that the surfactant to Li+ ratio was varied as shown. ........ 163

N2 adsorption and desorption isotherms of as—made meso-structured
silica molecular sieves as described in Figure 6.7. The surfactant in the
mesopores was removed by ethanol extraction (i.e. refluxing in ethanol
for 2 h) prior to the adsorption measurement. Again, the hexagonal
silica exhibits no hysterisis loop, as expected for uniform uni-
dimensional pores, whereas the wormhole silicas show pore necking
hysterisis loops. ....................................................................................... 164

X-ray diffraction patterns for calcined (600 °C) mesoporous silica
molecular sieves prepared using the mixed alkyl polyethylene oxide
surfactant Tergitol 15-S-12 (€12-16H25-33(EO)3-10H) as the structure
director and TEOS as the silicon source under the same reaction
conditions (45 °C, 40 h reaction time and a surfactant to metal cation
ratio of approximately 24 to 1), but different cation valences for the
promotional metal salts. The X-ray patterns show a decrease of d-
spacing from 6.4 to 4.65 nm upon replacing Co2+ and Mn2+ with Cr3+.. 165

N2 adsorption and desorption isotherms for mesoporous molecular
sieves described in Figure 6.9. The Cr3+ mediated structure shows a
pore size 2.0 nm in comparison with a pore size of 3.3 nm for the
structures mediated by Co + and Mn“. ................................................... 166

xvi

Figure 6.11.

Figure 6.12.

Scheme 1 .‘

 

Figure 6.11.

Figure 6.12.

Scheme 1

SEM (top) and TEM (bottom) images of mesoporous silica prepared
using Tergitol 15-S-12 in the presence of Cr3+ as a structure promoter,
showing micrometer sized intergrown spherical particles with flatten
surfaces on the spheres. ........................................................................... 167

”Si NMR spectra for as-made HMS(S°1°), MSU-H, MCM-41(S+X’I*)
and MCM-41(S+I'). MSU-H was synthesized using Brij 56 as the
structure director in the presence of CoCl2. HMS and MCM—415 were
prepared as reported. ............................................................................... 169

Metal salt mediated formation of mesostructured silicas ................ . ....... 170

xvii

SupramOI '

1.1.1ntr0ducti

A major
310-611 research-e
of 11413 matert

KurOda and h”

 

 

 

quatem’an mm
more \ersatile .\
surfactants 18' r a
eiten ed the 5°
assembly meehar
X‘: Cl'. Br' and
further extended

substrate.4 It

me3090mm silica

Another 1

urOl’C‘CUial' siei'es

IT

: tilt the preset

Chapter 1

Supramolecular Assembly of Mesoporous Molecular Sieves and Their
Application in Catalysis

1.1. Introduction

A major advance in the design of mesoporous molecular sieves was provided by
Mobil researchers in reporting ‘ the supramolecular surfactant assembly of a broad family
of M41S materials with uniform pore sizes in the mesopore range 2.0-10.0 nm. Also,
Kuroda and his co-workers 2 described a structurally related mesoporous silica by
quaternary ammonium surfactant rearrangement of a layered precursor (kanemite). The
more versatile Mobil approach was based on the supramolecular assembly of cationic
surfactants (S+) and anionic inorganic precursors (I'). Stucky and his co-workers 3 greatly
extended the STI' electrostatic pathway of Mobil to include a charge-reversed S'l+
assembly mechanism, as well as counterion-mediated STX'I+ and S'M+I' pathways, where
X' = Cl', Br' and M+ = Na”, KT. More recently, counterion-mediated STXT' pathway was
further extended to prepare well ordered cubic mesoporous silica films on glass
substrate.4 It was also extended to protonated nonionic surfactants to synthesis
mesoporous silica with pore size up to 30 nm.5

Another pathway has been developed for the preparation of mesoporous
molecular sieves based on the hydrolysis of an electrically neutral inorganic precursor
(1°) inthe presence of a neutral amine (S°) surfactant as the predominate structure

directing agent. This S°I° pathway was first used to prepare a mesoporous silica

molecular sieve and a Ti-substituted analog. A small amount of protonated amine was

used as aco-s

protonated cu

 

 

Electrostatic it
forces at the ~

An equivalent.

  
  
  

nonionic poly e:
in electrically rt

mesoporous

Mitre 11 SCh

used as a co—surfactant in the original synthesis, 6 but subsequent studies 7 showed that the
protonated co-surfactant component was not needed to achieve framework assembly.
Electrostatic forces do not play an important role in S°I° assembly. Instead, the assembly
forces at the surfactant-inorganic precursor interfaces are based on hydrogen bonding.
An equivalent H-bonding pathway, denoted N°I° has also been demonstrated for
nonionic polyethylene oxide surfactants and 1° precursors. 8 The most recent development

in electrically neutral assembly was a preparation of an entire family of ultrastable vesicle

mesoporous

Supramolecular Assembly Pathways
to MeSOporous Molecular Sieves
MSU ORG

H-Bondlng
(neutral assembly)

 

Burpuoa 0mm

 

Charge matchln
(electrostatic pathways)

Figure 1.1 Schematic showing of interfacial interactions for surfactant micelles in

cooperatively assembly

molecular sic
531‘ pathil}:
Pathwt'S 11“]
workers for the

aeoordinate ct.

metal center. A

 

structures cont:
schematically 11

Other 1
templating area
not intrinsicall
Pit-formed liq
concentration I
0f the surfaces
altering of the
PdeucLs. Hos
difflCllll)’ in I
PTECUrsors 11m

SUFfact
Extensiyeh

d a

“Nantes. i n

ser—
4 4C1” '

33
~

1
.3th Skaters

molecular sieves using neutral gemini surfactants as the structure director. 9 The S°I° and
N°I° pathways are complementary to those electrostatic templating pathways. These
pathways also are distinguishable from the S-I pathway developed by Ying and her co-
workers for the synthesis of meso-structured transition metal oxides. 10 In the S-I pathway
a coordinate covalent bond is formed between the surfactant (usually, an amine) and the
metal center. All aforementioned cooperative assembly pathways afford as-made meso-
structures contain liquid crystal micelles in the framework pores. All the pathways were
schematically illustrated in Figure 1.1.

Other than the cooperative assembly mechanisms, surfactant liquid crystal

templating mechanism have been proposed too.”ll

Although this templating pathway is
not intrinsically different from cooperative assembly pathways, it relies very much on
pre-formed liquid crystal phases to form meso—structured molecular sieves under a high
concentration of surfactant. Diffusion of inorganic precursor molecules to the interfaces
of the surfactant head groups and water molecules in the liquid crystal micelles without
altering of the liquid crystal phase is crucial for obtaining structural order in the final
products. However, the inefficient use of relatively expensive surfactants and the
difficulty in maintaining the liquid crystal phases upon addition of the inorganic
precursors limits the wide-spread use of this templating pathway.

Surfactant-directed assembly of mesostructures by sol-gel methods has been
extensively exploited. Regardless of the type of inorganic precursors used to form meso-
structures, in general, the pore size and connectivity take the size and shape of the

surfactant liquid crystal phase in the structural channels of as-made materials. Therefore,

a good strategy for the design and synthesis of meso-structures would be the over-all

control of it
framework p<
Ill LILlL
alumina 5‘“
molecular sic
metal oxides
mesostructurt
primary 211111
zirconia was
bitWeen surt
could not be
SU'U'CTUTEd tu:
nonionic pl'
mommy}. 5
metal OXIde‘s
removal of St
L‘nlik
meSOmeUS S
has been int,-
Walls are Sill
toward Shirl},

limited X‘l‘ai,

o“, i:

control of micelle formation that determines the final liquid crystal phases in the
framework pores.
In addition to mesoporous silica molecular sieves, structurally stable mesoporous

”"4 many other metal oxides and metallic platinum

alumina, 8'” niobia, ‘0 titania,
molecular sieves ‘5 have been reported. Interestingly, most of these stable mesoporous
metal oxides have made use of electrically neutral or nonionic surfactants. For example,

'6 and niobium oxide 10 have been prepared using

mesostructured zirconium oxide
primary amine surfactants as structure directors. The surfactant in mesostructured
zirconia was easily removed by ethanol extraction because of weak hydrogen bonding
between surfactant molecules and framework walls, however the surfactant in niobia
could not be simply extracted by ethanol because of the presence of dative bonds. Meso-
structured tungsten and iron oxide, as well as other metal oxides were assembled using
nonionic pluronic surfactants.l4 Nevertheless, neutral surfactants produced more
structurally stable mesoporous transition metal oxides. In contrast, most mesostructured
metal oxides prepared using charged surfactants as the structure directors collapsed upon
removal of surfactants. ”"8

Unlike crystalline microporous zeolites, it has been recognized that the
mesoporous silica molecular sieves are built of amorphous silica walls. '9 Although, effort
has been invested in the synthesis of crystalline walled mesoporous molecular sieves, the
walls are still amorphous for most of the mesoporous molecular sieves. An advance
toward synthesis of fully crystalline mesoporous materials was claimed for materials with
limited X-ray order, as indicated by limited number of reflections.14 Intrinsically, the

quasi-crystalline walls of these materials is alike of a mesostructured SnO2, which is a

consequence of particle assembly rather than a surfactant directed assembly.20 Actually,

the assem'r
relative or
relatively 1
worm-hole
patterns. TI
ordered rr

characterist

neutral part

1.2. Structi
thittugh Ni
Four
0116 is an 0
tCTABr a,
prE‘Cllfsor rI
(i.e. Sufiacra
41Prepared
1‘ the Silic
”Mariano;

“Orin-hater,

the assembly forces at the surfactant-inorganic precursor interfaces greatly determines the
relative order of meso-structures. Longer range electrostatic interaction force affords
relatively higher structural order. The shorter range H-bond interaction force affords
worm-hole-like structural motifs that usually exhibit single-line X-ray diffraction
patterns. The key-role of longer range electrostatic interaction in the formation of higher
ordered meso-structures is illustrated in the following section comparing four
characteristic mesoporous silica molecular sieves assembled through electrostatic and

neutral pathways.

1.2. Structural Characteristics of Mesoporous Silica Molecular Sieves Assembled
through Neutral and Electrostatic Pathways

Four characteristic mesoporous molecular sieves are compared below. The first
one is an ordered cubic MCM-48 assembled using cetyltrimethyl ammonium bromide
(CT AB) as the structure director (8*) and tetraethyl orthosilicate (TEOS) as the inorganic
precursor (I) under basic conditions at relatively high concentrations of the surfactant
(i.e. surfactant to silicon ratio at 0.5). 2'22 The second one is an ordered hexagonal MCM-
41 prepared using cetyltrimethyl ammonium as the structure director and sodium silicate
as the silicon precursor through typical ST assembly pathway at relatively low
concentration of the surfactant (i.e. surfactant to silicon ratio at 0.25). 17 The third one is a
worm-hole-like HMS formed by a supramolecular assembly pathway with dodecylamine
as the structure director that is predominately S°I° in character. 7 The fourth one is a
disordered KIT -1 silica assembled via the 8+1" pathway using CTAB, sodium silicate and
EDTA as the reagents as reported by Ryoo et al. 23 These four mesoporous molecular

sieves were chosen because they are structurally distinctive.

Flgllf;

 

order reflectr.

Cubic order 1‘
reflections. is,
12001 retlectii.
smeared hump
intrinsic shon-

KTT-l. electrOs

 

ptochSOf mtgh
density caused 1
EDTA salts-
conditions. 5‘
inevitably 0
electrostatic in
shortened.
Apparen
mesoporous 1
prepared via a1
pathway exhibi
mesoporosity '
y. a
N2 pore-fitting

Piessure 0f 0 3

isothenn for

Figure 1.2 shows the X—ray diffraction patterns of these four samples. Higher
order reflections for hexagonal MGM-41 and cubic MCM-48 are unambiguously shown.
Cubic order is indicated by the presence of (211), (220), (321), (400), (420), and (332)
reflections, whereas hexagonal order is indicated by the presence of (100), (110) and
(200) reflections. In contrast, HMS and KIT -1 both show broad first reflection with
smeared humps at higher 26 angles. The Lack of structural order for HMS indicates the
intrinsic short- range interaction property of H-bonding in the assembly mechanism. For
KIT-1, electrostatic interactions between the surfactant head group and the inorganic
precursor might be weakened due to a decrease in the surfactant head group charge

density caused by association with

 

5* = cur-tamcnypr
s = CuHsNflz

EDTA salts. Under these

conditions, structural disorder

  
   

inevitably occurs as the
51' assembled, 150 °C MGM-48

 

electrostatic interaction range is

Intensity

shortened.

Apparently, those three S‘tammbled,100°c MOM-41

 

mesoporous molecular sieves

' . or ’ .
prepared vra an electrostatic 8+1~ s “”mu‘” '1- HMS

 

 

 

 

 

pathway exhibit only framework Set assembled. 100°C KIT -1
mesoporosity, as indicated by the 4 8 12 16 20
ZGXdeg rees)

N2 pore-filling step at a relative
Figure 1.2 X-ray diffraction patterns

pressure of 0,2 to 0.4. (Fig. 1.3) for mesoporous molecular sieves
assembled via electrostatic (ST) and
The isotherm for HMS neutral (S°I°) pathways.

characteristic
significant N;
figure l.3.ll1r
and 33 A, re
differences ar
size for MC
structurally si'
15 rported t0 1

sizes for 31C)

n——‘

‘IA'IIMA AA-__I__.I I- _ 1.

Al

characteristically exhibits both framework and textural mesoporosity with two

significant N2 uptake steps at relative pressures of 0.3 and 0.7. As shown by the inset in

Figure 1.3, the average pore sizes of HMS, MCM-41 and KIT-1 molecular sieves (29, 30,
and 33 A , respectively) are significantly larger than that of MCM-48 (26 A). Structural
differences are associated with differences in framework pore size. The similarity in pore
size for MGM-41 and KIT-1 may indicate that MCM-41 and KIT-l tend to be
structurally similar, but different from bi-continuous cubic MCM-48, even though KIT-1
is rported to possess interconnected pore channels analogous to MGM-48. Larger pore

sizes for MCM-48 (Ia3d) so far have not been reported. Although the X-ray diffraction

 

 

2000‘

 

 

 

A
a.
'—
(D
“
m ‘
B 1500-
o i
A
'3 a. r
a) 10 20 so 40 so so 70 HMS (S "“12"
'2 WAAAA“‘--
AA‘ 7
o 100 - Av!!! vV‘VIlllsV'V v v v v vv-vwwwc"
.g 0 a 1"" I KIT-1 (8+1)
\\\“
< r
r - a w
w E _ -/,,‘IV’;"'" .ooeoe e o o e o 9 990.10...”
E .wc-a non-41$ r)
= '
— 500 i 9’". . 'orueueueeeeee eee .. «one: to emcee
O 0‘" .0.
‘s
> . .- MOM-48 (3‘1)
o"
N «‘9‘,
z ys‘

 

 

0 "'l"'l"'l"'l"'
0 0.2 0.4 0.6 0.8 1
PIPo

Figure 1.3 N2 adsorption and desorption isotherms for mesostructured
molecular sieves assembled via ST and S°I° pathways.

 

Figure 1.4 TEM images showing regular order of mesopores for (a) MCM-4l and
(c) MCM-48, but wormhole like channels for (b) KIT-1 and (d) HMS.

pattern does
pore size dis:
fundamental .

TEN-l
tfigrue 1.4). '

mmdmm

pore channels

 

 

 

pore channels.
structural dirt}
electrostatic as
Structures asse
limited to the .

matfirials.

13' Synthesis
Sufiactam as
The fir

SUUCIUI‘C direCI

pattern does not explicitly shown higher order reflections, KIT-l exhibits a quite narrow
pore size distribution which is about 3 times narrower than that of HMS. This verifies the
fundamental differences in assembly mechanisms for HMS and KIT-l.

TEM images show the structural versatility for these four molecular sieves.
(Figure 1.4). The image for MCM-4l explicitly shows a typical hexagonal arrangement
of uni-dimensional pores. The image for MCM-48 shows triangle arrangement of the
pore channels running along the cubic (111) direction. Instead of very regulme arranged
pore channels, HMS and KIT -1 both exhibit worm-hole-like channels. It is hard to tell the
structural difference by TEM for HMS and KIT-1. Nevertheless, the long range
electrostatic assembly affords meso-structures with relatively higher order than meso-
structures assembled from electrically neutral surfactants. However, the order is still
limited to the order of pore channels, which is far from the atomic order in crystalline

materials.

1.3. Synthesis of Mesosporous Silica Molecular Sieves Using Primary Amine
Surfactant as the Structure Directors

The first example of HMS silica was prepared at ambient temperature in the
presence of a 13.5:1 molar mixture of dodecylamine and dodecylarnmonium ion as the
structure directing co-surfactants. 6 The product formed under these reaction conditions
exhibited only one resolved XRD reflection, which precluded the assignment of a long
range structure. Selected area electron diffraction studies provided evidence for the
occasional occurrence of very small domains of hexagonal symmetry, but the vast
majority of the sample was highly disordered and lacking in a long range regular

SII'UCIUI'C.

Si
ommmgt
suucture
mmsue

mmgam

953!

O
I. .

ffifi’

ug-q
1“

In"

5..

.1

2’. HIV. -

Subsequent studies revealed that equivalent HMS silicas could be prepared by
omitting the onium ion from the reaction mixture and using only the neutral amine as the
structure director. 7 This S°I° pathway afforded silicas with N2 adsorption properties,
pore sizes, and XRD patterns virtually identical to the original HMS products formed

using a mixture of S° and S+ surfactants. Also, the sparsely occurring small domains of

 

Figure 1.5 TEM image showing wormhole-like framework mesopores and meso-
scaled textural pores indicated by the arrows

hexagonal order were absent. In fact, hexagonal regions are very rarely formed even
when protonated surfactant is present. Instead, the wormhole channel motif with meso-
scaled inter particle voids shown in Figure 1.5 is formed almost exclusively 24 even when
up to 15% of the amine is protonated. The onium ion can be introduced by adding a
protonic acid. Alternatively, the introduction of certain Lewis acid centers, as in the
replacement of some Si4+ sites by Al“, Fe3+ or B“, will result in the formation of some

protonated amine surfactant during the assembly process in order to balance the resulting

10

framework.
Structurally

mesoscalcd I

frarnework. However, this small electrostatic participation of the surfactant is
structurally inconsequential, and neither it alter the wormhole channel motif, nor the

meso-scaled textural pores.

 

 

 

 

 

 

E

'—

(D

0':

B

3

u

o

.a

h

o

m

'u

<

o

E

2

o

>“ HK=29A
z .1

c I I I ' r I I
0 02 0.4 06 08 1
PIP.

Figure 1.6 N; adsorption and desorption isotherms for HMS prepared in (A) EtOH-
rich medium (H20 : EtOH = 35 : 65 by vol.); (B) HZO-rich medium (H20 : EtOH =
90 : 10 by vol.)

Recent studies have shown that particle morphology and textural porosity of HMS

can be judicially controlled by choice of solvent polarity. 23 HMS prepared in higher
polarity water-rich media (i.e. 90:10 H20 : EtOH medium by volume) shows wormhole-
like to sponge-like channels with fractal-like particle texture. This HMS exhibits a

Significant amount of textural mesopores in the range of 150 A to 500 A, built from inter-

growth of r

from the V0

 

ethanol medi
with beads 0
Therefore. re.

but a propertr

1.4 Catalyst
1.4.1 Genera

As pr
have receireg
flames-or},- p
mflg Chen
Complex to [h
also afforded
an excePliona
groups of MC
has been i‘IUd:
thCmESOPOres

Mesop
Woun- A rr
Mun-On 0f :

en
uCOUUIEr d1":
14

growth of meso-scaled fundamental particles. These textural mesopores are different
from the voids formed by the physical packing of particles. HMS prepared in high
ethanol medium (H20 : EtOH = 35 : 65) exhibits much larger fundamental particle size,
with beads on string type of morphology and very little textural porosity. (Figure 1.6)
Therefore, textural porosity is not an intrinsic property of electrically neutral assembly,

but a property generated by the kinetics of the assembly process.

1.4 Catalysis
1.4.1 General Application of Mesoporous Molecular Sieves

As potential sorbents, catalyst supports or catalysts, mesoporous molecular sieves
have received considerable attention. A high density silanol groups on the walls of the
framework provide unique capability for meso-structured silica molecular sieves for
grafting chemistry. A very active Ti-MCM-41 was prepared by grafting of a titanocene
complex to the meso-structured walls of MCM-4l. 25 A grafted Ru-complex on MGM-41
also afforded a super active oxidation catalyst. 26‘” Pd supported on MCM-41 generated
an exceptionally active catalyst for Heck reactions. 28 Grafting chelating ligand to silanol
groups of MCM—41 achieved excellent adsorbent for heavy metals in waste water. 29'” It
has been studied that most of hazardous heavy metals could be remedied by trapping into
the mesopores.

Mesoporous molecular sieves are catalytically useful as long as their pores are
uniform. A relatively high X-ray order is not necessarily beneficial for the catalytic

utilization of mesoporous molecular sieves. Actually, mesoporous molecular sieves still

encounter diffusion limitations, particularly for reactions in condensed media. The

12

accessibility.
extremely ir
The l

silicas. prep.

 

CllCllllCéll COD
pOl}'mCrlZallt
be discussed
Prepared lhrr
meSOporcs. a

faulltate 3CC€

1‘41 Peron

Metal

for the Pero:

CPO-lilies anj

accessibility to the catalytically active centers located in the mesoporous framework is
extremely important for diffusion controlled reactions.

The uses of chemically modified derivatives of mesoporous HMS molecular sieve
silicas, prepared by neutral surfactant assembly pathways, as catalysts for a variety of
chemical conversions, including peroxide oxidations of olefins and phenols, ring opening

polymerization of lactide dimers, selective reduction of NOx, and cumene cracking will

be discussed below. These HMS catalysts are often more active than analogous catalysts
prepared through electrostatic assembly mechanisms. The wormhole-like framework
mesopores, along with the presence of complementary textural mesopores, most likely

facilitate access to the catalytic active centers in the framework.

1.4.2. Peroxide Oxidation
Metal-substituted HMS silicas have received considerable attention as catalysts
for the peroxide oxidation of aromatics to phenols and quinones and of alkenes to

epoxides and diols (see equation 1-3). One of the first reactions to be investigated was

 

OH 0
Eq. (1)
O
OH
HO H
E: + + Eq (2)
—» ° / Eq. (3)
0/ 45 “c o
MMA MPV

l3

 

[he Ti-W

 
   
   
  

correspond”
was ofpamt‘
l. an lndusrr
havebeen in
phenol. mcrh
cpoxide. diol :

The C;
phase oxlda
dCSCrlbed in
Obtained wit
Consideratior
for the mesr

HUS C3121] 3.

Table 1“ ‘
MOI‘~‘C\ular

EWA

S‘FTEne

the Ti-HMS catalyzed oxidation of 2,6—di-tert-butylphenol (2,6—DTBP) to the
corresponding mono-and dinuclear quinones using H202 as the oxidant. 6 This substrate
was of particular interest, because it was too large to access the framework Ti sites of TS-
], an industrial microporous molecular sieve catalyst. 31 Other oxidation reactions that
have been investigated using Ti-HMS as a catalyst includes the conversion of benzene to
phenol, methyl metharcylate to methyl pyruvate and styrene to the corresponding
epoxide, diol and benzaldehyde. 3'2

The catalytic properties of mesoporous Ti-HMS and of Ti-MCM-41 for the liquid
phase oxidations of methylmethacrylate, styrene and 2,6-di-tert-butylphenol are
described in Table 1.1.32 Included in the Table for comparison are the conversions
obtained with microporous TS-l as the catalyst. As expected based on pore size
considerations, the conversions observed for all three substrates are substantially larger
for the mesoporous catalysts than for the microporous catalyst. The S°I°-assembled Ti-

HMS catalyst exhibited consistently greater reactivity than the two Ti-MCM-4l catalysts

Table 1.1 Catalytic Peroxidation Activity of Ti-substituted (2 mol%) Mesoporous
Molecular Sieve Silica

 

 

 

 

Catalyst TS-l Ti-MCM-41 Ti-MCM-41 Ti-HMS
(S‘T) (S+X'I+) (S°I°)
MMA Conv. (mol%) 2.5 4.0 6.2 6.8
Styrene Conv. (mol%) 8.4 10 23 28
2,6-DTBP Conv. (mol%) 5.0 39 22 55
MA is methylmethacrylate.

prepared by electrostatic assembly. The superior performance of Ti-HMS is especially

pronounced in the case of the large 2,6-di-tert-butylphenol substrate.

14

be anri
Iiianiur
coordin
similuri
ngl’t’gd
centers.
superior
aiailihl.
inaccess
Ti-llCh
15). Th
l0 the a

efficienc

silicas 35
Verify 1h;
M0141
active Slit

number 01

The differences in catalytic reactivity between Ti-HMS and Ti-MCM-4l cannot
be attributed to differences in Ti siting. XANES and EXAFS studies showed that the
titanium center adopt primarily tetrahedral coordination in all three catalysts. 32 Also, the
coordination environment is very similar for the three catalysts, as judged from the
similarities in the EXAFS features. Also, UV-VIS adsorption spectra showed no phase
segregation of titania, the spectral features being consistent with site-isolated titanium
centers. Because the framework walls of HMS tend to be thicker than MCM-4l, the
superior reactivity of Ti-HMS cannot be due to an enhancement in the fraction of Ti
available for reaction on the pore walls. Thicker walls should bury more titanium at
inaccessible sites within the walls. The most distinguishing feature between Ti—HMS and
Ti-MCM-41 is the greater textural (inter particle) mesoporosity for Ti-HMS (cf. Figure
1.5). This complementary textural mesoporosity facilitates substrate transport and access
to the active sites in the framework-confined mesopores, thus enhancing catalytic
efficiency compared to MCM-4l meso-structures with little or no textural porosity.

Sayari and his co-workers ”'36 have investigated Ti- and V-substituted HMS
silicas as liquid phase oxidation catalysts for large organic molecules. Their results
verify that the activity of HMS derivatives is typically higher than the corresponding
MCM-41 analogs. He also has emphasized the importance of the accessibility of Ti-
active sites in determining reactivity toward large molecules. He further noted that a
number of oxidation reactions which occured readily over small pore TS-l catalysts, did
not take place in the presence of larger pore Ti-B or ultra large pore Ti-MCM-41 and Ti-
HMS. All these observations suggest that differences in surface hydrophilicity and Ti
redox potential also play a role in determining the reactivity of tetrahedral Ti sites in

these frameworks. On the other hand, TS-l is not known to catalyze the oxidation of

15

JCCIOUC at T

 

acetone oxi.

Kilnl
HUS and
butylhydropt
than Ti-MC.\|

N prepared

temperature.

 

 

 

air were p0
subsequently l
rt-stcsHi r
Tuel suggest

16mm] Poms:
limits] meso;
N331d$0l13ti0n
m1Sand HM
for Meteor;
Process is can
texrural Porog
cam-V315 can

acetone at rates that are competitive with benzene hydroxylation, yet Ti-HMS catalyzes
acetone oxidation readily.37

Kaliaguine and his co-workers 38 have compared the catalytic reactivity of Ti-
HMS and Ti-MCM-41 silicas for the epoxidation of oe-pinene with tert-
butylhydroperoxide as the oxidant. The conversions over Ti—HMS were somewhat lower
than Ti-MCM-41, although epoxide selectivities were similar. Also, Gontier and Tuel
39‘” prepared a series of Ti-HMS using reaction times as short as 15 min at ambient
temperature. Tetrahedral Ti loadings up to 2 wt % and thermal stabilities up to 650°C in

air were possible without forming extra-framework titania. 39

However, they
subsequently found no substantial difference in catalytic reactivity between Ti-HMS and
Ti-MCM-4l for the oxidation of aniline. 40 These results of Kaliaguine and Gontier and
Tuel suggest that the Ti-HMS derivatives used in these studies not possess the high
textural porosity needed for facile access to framework Ti sites. In fact, the absence of

textural mesoporosity in the Ti-HMS catalyst used by Gontier and Tuel was confirmed by
N2 adsorption studies. 4' These reports verify that the interparticle mesoporosity of Ti-
HMS and HMS molecular sieves, in general, is dependent on the reaction conditions used
for framework assembly. In general, textural porosity is formed when the assembly
process is carried out in a water-rich solvent. Alcohol-rich solvents tend to eliminate
textural porosity. Enhanced reactivity for large molecule conversions over HMS
catalysts can be expected only when the textural and framework mesoporosities are

comparable in magnitude.

16

1.43 Ri

achieved
promotir
the total;
potential
Sm IV l-Sl
lactide di
low polyt‘
Sr
umey
Swim-(:3
directing 5
indicated .
a textum
Ta
ellllzttion .-'
a SMOpg
Prepared
those User
Tr.
COm'miO]

and a be

1.4.3 Ring Opening Polymerization

The small crystallite domain sizes and high textural mesoporosity that can be
achieved for HMS derivatives through S°I° assembly may be especially beneficial in
promoting polymerizations and other bulky conversions where diffusion effects can limit
the catalytic effectiveness of larger particle mesostructures. In order to demonstrate the
potential utility of mesostructures for polymerization reactions, it has been shown 42 that
Sn(IV)-substituted HMS is remarkably effective for the ring opening polymerization of [-
lactide dimer to poly (l—lactic acid), abbreviated PLA, with a high molecular weight and
low polydispersity.

Sn-HMS containing 1 mol % tin(IV) was prepared at ambient temperature by S°I°
assembly in ethanolzwater (3:1v/v) using a 100:1 molar mixture of Si(OC2H5)4 and
Sn(iso-C3H7)4 as the inorganic precursors and dodecylamine as the structure
directing surfactant. The N2 adsorption isotherm for the calcined (550 °C) mesostructure
indicated a BET surface area of 886 mzlg, an average framework pore size of 2.7 nm, and
a textural (inter-particle) mesoporosity in excess of the framework mesoporosity.

Table 1.2 reports the conversions of L-lactide dimer to FLA at 130 °C (see
equation 4). Included for comparison purposes are the conversions for pure HMS silica,
a Sn-doped silica gel (1.0 mol% Sn) and pure SnOz. The latter two catalysts were
prepared by hydrolysis of the corresponding alkoxides under condition analogous to
those used to form Sn-HMS.

The polymerization product obtained from Sn-HMS exhibited the highest

conversion as determined by lHNMR (82%), the largest average molecular mass (36000)

and a low polydispersity (1.1). In the case of pure tin oxide as the catalyst, the

17

conversion

lower mole

 

case of Sn-t
centers in t

Lewis-acid :

vs

l

conversion was substantial (73%), but the polymerization product had a much
lower molecular mass (17800) and a high polydispersity (1.7). The low activity in the
case of Sn-doped silica gel may indicate the lack of a suitable dispersion of active metal
centers in the host silica. Sn-HMS, however, clearly combines the reactivity of tin

Lewis-acid sites with the selectivity of a regular mesopore structure in affording PLA

0 o 0
Sn-HMS II II
CH3CH(OH)CO(CH(CH3)CO),,H Eq. (4)
o o 130 °C

Lactide Dimer PLA

 

Table 1.2 Lactide polymerization over heterogneous catalysts“

 

 

Catalyst Conversion (%) PLA Molecular mass
Sn-HMS 82 36 000

HMS 0 --
Sn-doped Silica 22 3 200

Sn02 73 17 800

 

 

“Reaction conditions: 2.00 g ( 13.9 mmol) lactide dimer; 0.1 g catalyst (except for SnOz,
where 0.001 g was used); T = 130 °C; reaction time 72 h. All catalysts were calcined at

550 °C prior to use.

with a reasonably high molecular mass and low polydispersity. It appears that the
ordered pore structure improves average molecular mass and polydispersity values in
comparison to homogeneous catalysts by imposing steric constraints on the propagating
PLA chains and minimizing ‘back-biting’ and intermolecular transesterification

reactions.

18

1.4.4. 5

MCM-
equahc
dil‘t‘ere.
conten'
distribi
were it
0.75 c
mesOpt

0f 00 ll

Sufiace
Fain:
down
“1‘ of

“Jude

1.4.4. Selective Catalytic Reduction (SCR) of NO

Yang et al., 43 recently investigated the activity of Fe3+ exchanged forms of A1-

MCM—4l and Al-HMS for the selective catalytic reduction (SCR) of NO by NH3 (see

equation 5). In order to minimize the number of factors that may cause significant
differences in the catalytic properties between Al-MCM-4l and Al-HMS, the aluminum
contents of both molecular sieves were controlled to around 8%. Also, the pore size
distribution for the two supports was controlled to around 28 A, the BET surface areas
were in the range 800-850 m2/g, and the framework pore volumes were similar (0.62 and
0.75 cm3/g for Al-HMS and A1-MCM-41, respectively). However, the textural
mesoporosity of Al-HMS was comparable to the framework mesoporosity, whereas little
or no textural porosity was present for Al-MCM-41.

Although HMS and MCM-41 are similar both in terms of chemical composition,
surface area, and pore volume, Fe/Al-HMS showed considerably higher activities than
Fe/Al-MCM-4l. The main differences between HMS and MGM-41 are the crystal
domain sizes and the unique textural porosity of HMS. The crystal domain size for HMS
was of the order 150 A, whereas that of MCM-41 was larger by two orders of magnitude,
as judged by XRD line widths and TEM.

The comparison of the estimated apparent and intrinsic rate constants given in
Table 1.3 show that the reaction for the Fe/Al-MCM-41 sample was severely limited by
pore diffusion, but not for HMS molecular sieve. The Thiele effectiveness factor for the
HMS was nearly 1.0, whereas that for the MGM-41 were 0.53 at 350 °C and 0.44 at 400

°C. An overall activation energy of 6.5 kcal/mol for Fe-HMS was in the range for SCR

19

withou
attribut
Thus. I

Al-HM

Table .
the ow
MCM-

Catalyst

\

FéfAl-l

without diffusion limitation. The high effectiveness factor for the HMS catalyst is
attributable to the small domain size and short diffusion path in the framework channels.
Thus, the advantage for the Fe3+ exchanged mesoporous molecular sieves, in particular

Al-HMS, for the SCR reaction is clearly demonstrated.

NO + NH3 03‘3”“ a» N2 + H20 Est-(5)

350 °c

Table 1.3 Comparison of the apparent and intrinsic first-order rate constants and
the over all diffusivity for NO SCR reaction over Fe3+ exchanged Al-HMS and Al-
MCM-4l'l'

 

Catalyst Apparent k (8") Intrinsic k (s’l)£1 Diffusivity (cmzls);l
350 °C 400 °C 350 °C 400 °C 350 °C 400 °C

 

Fe/Al-HMS 56.5 87.5
56.6 87.6 3.95x10'7 4.0x10'7
Fe/Al-MCM-41 30.0 38.5

 

 

éThe intrinsic rate constant and diffusivity are assumed to be the same for both
mesoporous catalysts.

1.4.5. Acid Catalysis

The replacement of silicon by trivalent elements in HMS materials has been
investigated by several groups in an effort to improve Bronsted and Lewis acidity for
catalytic applications. 4448 Although Al-MCM-41 derivatives can be prepared directly by
electrostatic assembly pathways, the as-synthesized materials are structurally sensitive to
calcination. 49 Also, if sodium silicate is used in the synthesis, a post-synthesis treatment

with NH4NO3 is needed to remove residual Na+ ions from the exchange sites. 50 It has

been reported by Corma and his co-workers 5' that Al-MCM-4l undergoes framework

20

dealumin.
dealumin;
dealumin:
proton CXt
In
pathway
molecular
noted earl
of structur
frameworl
létmhedral
interaction
Surfactant
PWSCn'es r
Tu:
pathway it
of Walsh
incOIT‘Orat:
lmponam‘l.
damaging

Man an

dealumination upon calcination to remove the surfactant. Hitz and Prins 52 verified the
dealumination of surfactant-filled Al-MCM—41 upon calcination, but they showed that
dealumination could be minimized by first removing up to 73% of the surfactant by
proton exchange in ethanol.

In contrast to the electrostatic assembly pathways to Al-MCM-41, the S°I°
pathway to Al-HMS derivatives offers a convenient route to acidic mesoporous
molecular sieves with retention of framework aluminum and other trivalent ions. As
noted earlier, the replacement of Si4+ by Al3+ requires the protonation of one equivalent
of structure-directing amine surfactant for every equivalent of Al3+ incorporated into the
framework. However, because only up to 15 mole % of the silicon can be replaced at
tetrahedral sites, the majority of the framework is assembled through H bonding
interactions of the amine surfactant with the silica framework. Consequently, most of the
surfactant (>90%) can be efficiently removed by simple solvent extraction and this
preserves the tetrahedral siting of aluminum centers in the framework.

Tuel and Gontier 44 have recognized the potential importance of the S°I° assembly
pathway for the preparation of HMS derivatives containing trivalent framework elements
of catalytic significance. They found that A134”, Ga”, Fe3+ and 83* all could be
incorporated into the framework of HMS silica at Si/M3+ ratios as low as 10. More
importantly, the neutral amine surfactant could be removed by solvent extraction without
damaging the framework. Also, the remaining small concentration of charge balancing
primary alkylammonium ions associated with the aluminum sites could be removed
either by calcination or by ion exchange without collapsing the framework or removing
the aluminum from the framework. Similar results were obtained for other M3”-

substituted derivatives. Thus, the structural degradation that normally occurs upon

21

calcining .
through S
surfactant.
protonated ~

Mo's

mesoporOUs

 

pathway. L',
synthesis (ti
ethanol-Mate
surfactant). 1
identical to 1»
HMS defiml
”“3: 540 e
cOml‘iifed t0
alUtninum cc
h 13 unlikel}
HMS, Al-Mt
Walls in a“ 1
materials ma
compared m
The Shon int
may Well rm.

may be the n

calcining M3+-substituted mesostructured silicas to remove surfactant can be avoided
through S°I° assembly and the subsequent solvent extraction of the vast majority of the
surfactant. The heat of combustion associated with the removal of the remaining
protonated surfactant is not sufficient to significantly alter the framework structure.
Mokaya and Jones ”'46 also have recognized the advantages of preparing
mesoporous silica and aluminosilicate molecular sieves through a S°I° assembly
pathway. Using a synthetic methodology equivalent to that originally reported for HMS
synthesis (i.e., alkyl amines as the surfactant, alkoxides as the inorganic precursors,
ethanol/water as a solvent, ambient temperature assembly, and solvent extraction of the
surfactant), they prepared a series of derivatives with XRD and textural properties
identical to HMS materials. However, they designated their products as MS rather than
HMS derivatives. Interestingly, their Al3+-substituted products with Si/Al ratios in the
range 5-40 exhibited greater Bronsted acidity and catalytic activity for cumene cracking
compared to Al-MCM-41, amorphous aluminosilicate gel and zeolite HY of similar
aluminum content. 47 In addition, the Al-HMS catalysts were less prone to deactivation.
It is unlikely that the siting and intrinsic acidity of tetrahedral aluminum centers in Al-
HMS, Al-MCM-41, and aluminosilicate gels differ significantly, because the framework
walls in all three materials are amorphous. However, access to the acidic sites in these
materials may differ greatly. What is particularly distinctive of most Al-HMS materials
compared to Al-MCM—41 and amorphous gels is the uniformity of the framework pores.
The short intersecting framework pores, in addition to the small crystallite domain size,
may well facilitate access to the active sites in HMS structures. This improved access

may be the main reason for the superior reactivity of Al-HMS materials as acid catalysts

22

relative to Al-MCM-4l, which have very long channel lengths, and amorphous gels with
irregular, highly constrained pore-structures.

Very recently, van Bekkum and his co-workers 48 have reported modifying Al-
HMS and Al-MCM-41 (Si/Al E 30) to contain entrapped unit cells of ZSM-5. These
modified mesostructures were prepared by ion exchanging the aluminosilicates with
tetrapropylarnmonium cations as MFI structure directors and subsequently digesting the
mesostructures in glycerol at 120°C for 24 h. The resulting nanoporous aluminosilicate

products, designated NPA-l and NPA-2, respectively, exhibited XRD patterns

 

Table 1.4 Cumene Conversions (%) at 300 °C

 

 

 

Catalyst Al-MCM-4l PNA-l Al-HMS PNA-2 ZSM-S
20 min 14.7 41.3 24.8 47.6 95.1
3 hr. 13.6 37.5 26.8 42.4 93.7

 

 

characteristic of Al-MCM-41 and Al-HMS, but reflections characteristic of ZSM-5 were
absent. An FI'IR band at 550-560 cm‘1 in both products was considered to be indicative
of unit cells of ZSM—5 highly dispersed in the mesostructure framework. Further
evidence for the presence of entrapped ZSM-S unit cells was provided by the acid
catalytic activity of NPA-l and NPA-2 toward cumene cracking. As can be seen from
the cumene conversions listed in Table 1.4, the acid catalytic activity of the modified

mesostructures is substantially greater than the parent mesostructures. It is especially

23

noteworthy that PNA-2, obtained by modification of Al-HMS, is more active than the

PNA-l product derived from Al-MCM-41. The superior activity of Al-HMS and NPA-2

relative to Al-MCM-41 and NPA-l may again be attributed to the intrinsically more

efficient access to framework catalytic sites for HMS derivatives.

1.5. References

l.

15.

16.

Beck, J .S.; Vartuli, J .C.; Roth, W.J.; Leonowicz, M.E.; Kresge, C.T.; Schmit, K.D.;
Chu, C.T.-W.; Olson, D.H.; Sheppard, E.W.; McCullen, S.B.; Higgins, J .B.;
Schlenker, J .L J. Am. Chem. Soc. 1992, 114, 10834.

Inagaki, S.; Fukushima, Y.; Kuroda, K. Chem. Commun. 1993, 680.

Huo, Q.; Margolese, D.I.; Ciesla, U.; Feng, P.; Gier, T.E.; Sieger, P.; Leon, R.;
Petroff, P.M.; Schiith, F.; Stucky, G.D. Nature 1994, 368 , 317.

Lu, Y.; Ganguli, R.; Drewien, A.C.; Anderson, M.T.; Brinker, C.J.; Gong, W.; Guo,
Y.; Soyez, H.; Dunn, B.; Huang, M.H.; Zink, J .1. Nature 1997, 389, 364.

Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G.H.; Chmelka, B.F.; Stucky,
6D. Science 1998, 2 79, 548.

Tanev, P.T.; Chibwe, M.; Pinnavaia , TJ. Nature 1994, 368, 321.

Tanev, T.P.; Pinnavaia, T.J. Science 1995, 267, 865.

Bagshaw, S.A.; Prouzet, E.; Pinnavaia, T.J. Science, 1995, 269, 1242.

Kim, S-S.; Zhang, W.; Pinnavaia, T.J. Science 1998, 282, 1302.

Antonelli, D.M.; Ying, J.Y. Angew. Chem, Int. Ed. Engl., 1996, 35, 426.

Attard, G.S.; Glyde, J .C.; Goltner, C.G. Nature 1995, 378, 366.

Bagshaw, S.A.; Pinnavaia, T.J. Angew. Chem. Int. Edn. Eng]. 1995, 34, 2014.
Antonelli, D.M.; Ying, J.Y. Angew. Chem. Int. Edn. Engl. 1995, 34, 2014.

Yang, P.; Zhao, D.; Margolese, D.I.; Chmelka, B.F.; Stucky, G.D. Nature 1998,
396, 152.

Attard, G.S.; Bartlett, P.N.; Coleman, N.R.B.; Elliott, J .M.; Owen, J .R.; Wang, J .H.
Science 1997, 278, 838.

Huang, Y.; McCarthy, T.J.; Sachtler, W.M. Appl. Catal. A. 1996, 148, 135.

24

ii
in
31

ii

Huo.

P.; Fi
1176.
Yoda.
Chen.
Seven
hlonn
Stucltj
liF.S
Zhang
Ryoo.
Zhang
Masch
Kim. S
Liu. 1..
hlehnc
Merci:
Feng,

1997, ,

' Notari

. Zhang

Chem.
Reddg
Redd}
Saran

Rfiddt

17.

18.
19.
20.
21.

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

31.
32.

33.
34.
35.
36.
37.

38.

Huo, Q.; Margolese, D. I.; Ciesla, U.; Demuth, D. G.; Feng, P.; Gier, T. E.; Sieger,
P.; Firouzi, A.; Chmelka, B. F.; Schiith, F.; Stucky, G. D. Chem. Mater. 1994, 6,
1176.

Yada, M.; Machida, M.; Kijima, T. Chem. Commun. 1996, 769.

Chen, C.-Y.; Li, H.—X.; Davis, ME. Microporous Mater. 1993, 2, 17.

Severin, K.G.; Abdel-Fattah, T.M.; Pinnavaia, TJ. Chem. Commun. 1998, 1471.
Monnier, A.; Schuth, F.; Huo, Q.; Kumar, D.; Margolese, D.; Maxwell, R.S.;
Stucky, G.D.; Krishnamurty, M.; Petroff, P.; Firouzi, A.; Janicke, M.; Chmelka,
B.F. Science 1993, 261, 1299.

Zhang, W.; Pinnavaia, T.J. Catal. Lett. 1996, 38, 261.

Ryoo, R.; Kim, J .M.; Shin, C.H.; Lee, J.Y. J. Phy. Chem. 1996, 100, 17718.

Zhang, W.; Pauly, T.R.; Pinnavaia, T.J. Chem. Mater. 1997, 9, 2491.

Maschmeyer, T.; Rey, F.; Sankar, G.; Thomas, J .M. Nature 1995, 378, 159.

Kim, S-S.; Zhang, W.; Pinnavaia, T.J. Catal. Lett. 1997, 43, 149.

Liu, L.J.; Yu, W.Y.; Li, S.G.; Che, C.M. J. Organ. Chem. 1998, 63, 7364.

Mehnert, C.P.; Ying, J.Y. Chem. Commun. 1997, 2215.

Mercier, L.; Pinnavaia, TJ. Adv. Mater. 1997, 9, 500.

Feng, X.; Fryxell, G.E.; Wang, L.-Q.; Kim, A.Y.; Liu, 1.; Kemner, K.M. Science
1997, 276, 923.

Notari, B. Adv. Catal. 1996, 41, 253.

Zhang, W.; Froba, M.; Wang,J.; Tanev, P.T.; Wong 1.; Pinnavaia, T]. J. Amer.
Chem. Soc. 1996, 118, 9164.

Reddy, K.M.; Moudrakovski, I.; Sayari, A. Chem. Commun. 1994, 1059.

Reddy, J .S.; Sayari, A. Chem. Commun. 1995, 2231.

Sayari, A. Chem. Mater. 1996, 8, 1840.

Reddy, J.S.; Liu P.; Sayari, A. Appl. Catal. A 1996, 148, 7.

Zhang, W.; Wang, J.; Tanev, P.T.; Pinnavaia, T]. J. Chem. Soc., Chem. Comm.
1996, 979.

On, T.D.; Kapoor, M.P.; Joshi, P.N.; Bonneviot, L.; Kaliaguine, S. Catal. Lett.
1997, 44, 171.

25

39.
40.
41.
42.
43.

45.
46.
47.
48.
49.
50.
51.
52.

Gontier, S.; Tuel, A. Zeolites 1995, 15, 601.

Gontier, S.; Tuel, A J. Catal. 1995, 157, 124.

Tuel, A.; Gontier, S. Chem. Mater. 1996, 8, 114.

Abdel-Fattah, T.M.; Pinnavaia, T.J. Chem. Commun. 1996, 665.

Yang, R.T.; Pinnavaia, T.J.; Li,W.; Zhang, W. J. Catal. 1997, I 72, 488.

Yue, Y.; Sun, Y.; Xu, Q.; Gao, Z. Appl. Catal. A: 1998, 175, 131.

Mokaya, R.; Jones, W. Chem. Commun. 1996, 981.

Mokaya R.; Jones, W. Chem. Commun. 1996, 983.

Mokaya, R.; Jones, W. J. Catal. 1997, 172, 211.

Kloetstra, K.R.; van Bekkum, H.; Jansen, J .C. Chem. Commun.1997, 2281.
Corma, A. Chem. Rev. 1997, 97, 2373.

Luan, 2.; Cheng, C.F.; He, H.; Klinowski, J. J. Phys. Chem. 1995, 99, 10590.
Corma, A.; Fomés, V.; Navarro M.T.; Pérez-Pariente, J. J. Catal. 1994, 148, 569.
Hitz, S.; Prins, R. J. Catal. 1997, I68, 194.

26

Mt
Temp
Pathv

Abstrat
He:
textural
tempera
incorpor
mdwm
anionic
S‘l‘and
ionic sil
ms N;
(Wheres
degree 0
for the r
isolated
“Wgwj
inds‘tterid
Cage of

GREEN

[0 frames

Chapter 2

Mesoporous Titanosilicate Molecular Sieves Prepared at Ambient
Temperature by Electrostatic (ST, S+X'I+) and Neutral (S°I°) Assembly
Pathways : A Comparison of Physical Properties and Catalytic Activity

for Peroxide Oxidations

Abstract

Hexagonal mesoporous titanosilicates with distinguishable framework charges and
textural mesoporosity, namely, Ti-MCM-4l and Ti-HMS were prepared at ambient
temperature by electrostatic and neutral assembly processes, respectively. Titanium
incorporation at the 2 mol % level was accompanied by increases in lattice parameters
and wall thickness, but the framework pore size remained unaffected. Cross-linking of
anionic framework of as-synthesized Ti-substituted MCM-41 prepared by electrostatic
S+I‘ and S+X‘I+ assembly pathways (where St“ is a quaternary surfactant and I” and I+ are
ionic silicon precursors) was enhanced significantly by Ti-substitution, as judged by 2‘e’Si
MAS NMR. The neutral framework of as-synthesized Ti-HMS formed by $010 assembly
(where S0 is a primary amine and 1° is a neutral silicon preursor) exhibited the same high
degree of cross-linking as the unsubstituted silica analog. UV—VIS and XANES spectra
for the calcined forms of Ti-MCM-41 and Ti-HMS indicated: (i) the presence of site-
isolated Ti species in the framework; (ii) predominantly tetrahedral coordination for Ti,
along with some rehydated five- and six-coordinated sites; (iii) Ti siting that was virtually
independent of the framework assembly pathway. The exceptional catalytic activity in the
case of Ti—HMS, especially toward larger substrates, was attributable to the small
crystallite size and complementary textural mesoporosity that facilitates substrate access

to framework Ti sites .

27

ll lntrt

phase p
hs'droxy
site-isol.
coordin:
small p(
are limi
hexagon
surfacta
POre TS

MC
aSsemtil
ElCCUog‘
r0d-lilte
~)- The:
of [he s
POSltiwef
a neUlra
“Cum 5
“1108111.
With “Cg

frameWO

2.1 Introduction

Titanium silicalite-1 (denoted TS-l) is an effective active catalyst for the liquid
phase peroxide oxidations of alcohols and alkanes 1, the epoxidation of alkenes 2 and the
hydroxylation of aromatics 3. The activity of TS-l arises from the presence of accessible,
site-isolated Ti centers in the silicalite framework that are capable of undergoing a facile
coordination change and forming an active peroxo-titanium complex 4. Because of the
small pore size of the inorganic framework, the substrates that can be oxidized by TS-l
are limited to species having kinetic diameters < 6 A. However, the recently reported
hexagonal mesoporous silica molecular sieves prepared by electrostatic 5'6 and neutral 7’8
surfactant templating pathways offer promising opportunities for the preparation of large
pore TS-l analogs capable of transforming larger organic molecules.

MCM-41 silicas normally are prepared by one of two possible electrostatic
assembly pathways. The S+I‘ pathway originally utilized by Mobil researchersS involves
electrostatic interactions and charge matching between positively charged assemblies of
rod-like micelles of quaternary ammonium surfactants (8+) and anionic silicate species (I
). The second pathway, the so-called counterion mediated S+X’I+ pathway 6, makes use
of the same surfactant cations under strongly acidic conditions in order to assemble a
positively charged silica precursors (1*). In contrast, HMS derivatives were prepared by

a neutral S°I° assembly pathway that involves hydrogen bonding interactions between

neutral S° primary amine surfactants and neutral I° inorganic precursors (e.g., tetraethyl

orthosilicate) 7. The electrostatic assembly pathways afford as-synthesized materials
with negatively charged framework walls, whereas neutral assembly yields neutral

frameworks and walls that are characteristically thicker than those formed by electrostatic

28

assembly. '

chemical prc

Contra
The catalyst
at 408 K. 1
110141 at
greater reac
termed on
the synthes
hydrotherm
concerning
asStttnbleti

In the
HMS can
compare t1

While a,

2‘2 E“(Per

12'] Matt
13-1

by the me

assembly. These fundamental differences give rise to distinguishable physical and

chemical properties for MGM-41 and HMS materials.

Corma et al. have reported the preparation and catalytic activity of Ti-MCM—41 9.
The catalyst was obtained by electrostatic S+I‘ assembly under hydrothermal conditions
at 408 K. We reported the ambient temperature preparation of St‘X‘I+ - assembled Ti-
MCM-4l and S°I° assembled Ti-HMS catalysts ‘0 and found preliminary evidence for
greater reactivity for the Ti-HMS derivative. More recently, several other groups have
reported on the preparation and catalytic properties of Ti-MCM—41. ”‘15 In all instances
the syntheses were accomplished using exclusively the S+I' assembly pathway 5 under
hydrothermal conditions at temperatures above 373K. Thus, relatively little is known
concerning the catalytic activities of titanium-substituted mesoporous molecular sieves
assembled at ambient temperature by electrostatic assembly pathways.

In the present work we investigate the assembly of mesoporous Ti-MCM—41 and Ti-
HMS catalysts via electrostatic and neutral assembly processes at ambient temperature,
compare the physical properties of these products, and elucidate the differences in their

catalytic activities for the peroxide oxidations of large organic molecules.

2.2 Experimental
2.2.1 Materials

TS-l with an analytically determined titanium loading of 2 mol % was synthesized

by the method of Taramasso et al. 3. Tetraethyl orthosilicate (TEOS), tetraisopropyl

orthotitanate (TIPOT) and tetrapropylammonium hydroxide (TPAOH) were used as a

29

source
math
accon

h und

“Tilt

3556

solu
Stirr
hscl;

Si:

source of silica, titanium and template, respectively. The molar composition of the

reaction mixture was 0.022 Ti : 1.0 Si : 0.35 TPAOH : 35 H20. The synthesis was

accomplished by placing the reaction mixture in an autoclave and heating at 443 K for 48
h under static conditions.

Ti-MCM-41 and Ti-HMS derivatives with equivalent titanium loadings (i.e., ~2 mol
%) were prepared using deionized colloidal silica or TEOS as the source of silica, TIPOT
as a titanium precursor, and long chain alkylammonium or alkylamine surfactants,
respectively, as templates. In contrast to the hydrothermal reaction conditions normally
used to prepare Ti-MCM-41 (i.e. autoclaving above 373 K), our Ti-MCM-4l catalysts
were prepared by ambient temperature synthesis via electrostatic S+I' and S+X'I+
assembly. In both of these electrostatic pathways cetyltrimethylammonium cations
(CTMA+) served as the template. For S+I' assembly, a deionized 34 wt % colloidal silica
solution (Aldrich) and TIPOT were added to a solution of template under vigorous
stirring and the pH of the reaction mixture was adjusted to 12 with tetramethylarnmonium
hydroxide (TMAOH). The molar composition of the reaction mixture was 0.020 Ti : 1.0

Si : 0.50 CTMA+ : 1.0 TMAOH : 160 H20. The preparation of Ti-MCM-41 by the
counterion-mediated pathway (S+ X'It) utilized strongly acidic conditions (pH=l.5) in
order to generate and assemble positively charged silicon species from TEOS. In a
typical preparation the template was dissolved in acidic aqueous solution followed by the
addition of a TEOS/TIPOT solution. A higher Ti : Si gel ratio of 1:10 was used for this
pathway in order to prepare a product with the desired Ti:Si ratio of 2 : 98. The use of
excess Ti for this preparation was necessitated by the high solubility of Ti under strongly

acidic conditions. The composition of the reaction mixture was 0.10 Ti : 1.0 Si : 0.20

30

(:1th '. '
aged unde
product.
Ti-ll‘l
100A) as
TEOS/111
vigorous :
0.20 00.1
under vi;
comparat
omitting
thorough?
Corn
01 hexadg
reference

COOTdinat

2.2.2 Ch;

The

CTMA+ : 1.0 HCl : 160 H20. For both electrostatic pathways, the reaction mixture was
aged under vigorous stirring for 24 h in order to obtain the crystalline Ti-MCM-41
product.

Ti-HMS was synthesized via a neutral S°I° templating pathway using dodecylamine
(DDA) as the surfactant and ethanol (EtOH) as a co-solvent. In a typical preparation the
TEOS/1‘ IPOT solution was added to the solution of DDA in water and ethanol under
vigorous stirring. The molar composition of the reaction mixture was 0.022 Ti : 1.0 Si :
0.20 DDA : 9.0 EtOH : 160 H20. The reaction mixture was aged at ambient temperature
under vigorous stirring for 24 h in order to obtain the crystalline product. For
comparative purposes we have also prepared pure silica MCM-41 and HMS samples by
omitting the Ti source in the above preparations. All samples were filtered, washed
thoroughly with water, dried at ambient temperature and calcined at 923 K for 4 h.

Commercially available samples of MgTiO3, anatase, rutile, neptunite and a sample

of hexadecaphenyl-octasiloxy-spiro-(9, 9)—titanate (denoted spirotitanate 16) were used as

reference materials for XANES analysis. Spirotitanate contains Ti atoms tetrahedrally

coordinated to four siloxy oxygens 17.

2.2.2 Characterization
The powder X-ray diffraction (XRD) patterns of all samples were measured on a

Rigaku Rotaflex diffractometer equipped with a rotating anode and Cu-Ka radiation
(2:1.542 A). In general, the diffraction data were collected by using a continuous scan

mode with a scan speed of 2 degrees (ZED/min.

31

state
the s

deter

W-l
retlec
meat
transt'

Specti

Libor

29Si-MAS NMR spectra were measured at 79.5 MHz on a Varian VXR-4OOS solid
state NMR spectrometer equipped with a magic angle spin probe. For each measurement
the samples were placed in zirconia rotors and spun at 4.2 kHz. The quantitative
determination of Q2, Q3 and Q4 sites was accomplished by deconvolution of the spectra.

UV—VIS spectroscopic measurements were carried out on a Shimadzu UV-3101PC
UV-VIS-NIR Scanning Spectrophotometer equipped with an integrating sphere. A

reflection mode with a resolution of 10 nm and BaSO4 reference were used for the

measurements. The collected relative reflection intensity (R..=Rsample/Rrefe,ence) was

transformed into F(R.,..) by using Kubelka-Munk function F(R..) =(1-R..)2/(2R..) 18. All
spectra were plotted in terms of HR...) versus wavelengths.

The X-ray absorption spectra were recorded at the Stanford Synchrotron Radiation
Laboratory with a SPEAR storage ring operating at 3 GeV and 50-100 mA. Wiggler
beam line 102 was used with a Si-(220) double-crystal monochromator. In order to
reduce contributions from higher harmonics the second monochromator crystal was
detuned to 50% of the maximum intensity. The slit before the monochromator as well as
that in front of the first ion chamber was set to 1 mm height. The energy calibration of
the monochromator was checked between every two spectra by measuring the Ti-K edge
(4966 eV) of a Ti metal foil reference. All spectra were recorded at room temperature
using a continuous—scan technique (QEXAFS) 19. The counting time per data point was
0.5 s and the distance between data points was 0.1 eV 20. The pre-edges were normalized
for absorbance by fitting the spectral region from 4900 to 4950 eV with a Victoreen
function and subtracting it as background absorption. In addition, all spectra were

normalized for atomic absorption by using the average absorption coefficient of the

32

region from
preedge pe
XANES Spr
symmetric p
preedge an.
the micro- 8
using a Ster
were measu
filled with a

N2 ads
Omnisorp 3
measuremer

diSIIibutiom

region from 5050 to 5150 eV. The normalization was performed in order to compare the
pre-edge peak intensities, energy positions and peak widths. The normalized Ti K—edge
XANES spectra in the energy range of 4960 to 4980 eV were fitted using a series of
symmetric profile functions. Lorentzian and Gaussian functions were used for fitting the
pre-edge and the ascending absorption edge features, respectively. The XANES data of
the micro- and mesoporous titanosilicates were collected in the fluorescence-yield mode,
using a Stem-Heald type detector filled with Ar gas 21, whereas all reference compounds
were measured in a transmission mode using two ion chambers. The first chamber was
filled with a mixture of He and N2 (ratio 2:1) and the second only with N2.

N2 adsorption-desorption measurements were carried out at 77K on a Coulter
Omnisorp 360CX Sorptometer using a continuous flow measurement mode. Prior
measurements samples were outgassed at 423K and 10'6 Torr for 12 h. The pore size

distributions were calculated by the method of Horvath-Kawazoe 22.

2.2.3 Catalytic experiments

The catalytic performance of all samples was tested for the liquid phase peroxide
oxidation of: (i) methyl methacrylate (MMA) to methyl pyruvate, (ii) styrene to
benzaldehyde, and (iii) 2,6-di-tert-butyl phenol (2, 6-DTBP) to quinones. All reactions
were performed under vigorous stirring in a three-neck glass flask equipped with
condenser and thermometer. The oxidation of MMA and styrene was canied out using
10 mmol of substrate, 30 mg of catalyst, and 10 ml of acetonitrile as a solvent. The

amount of 30 wt % H202, the reaction temperature, and reaction time were as follows: (i)

40 mmol H202, 321 K and 6 h, respectively for MMA oxidation, and (ii) 20 mmol

33

how:
mmc
prod”
a F11
melt

and:

2.31

H202, 321 K, and 3 h, respectively for styrene oxidation. For 2,6-DTBP oxidation,

however, 10 mmol of substrate, 100 mg of catalyst, 10 ml of acetone as a solvent, and 30

mmol 30 wt % H202 were used, and the reaction was carried out at 335 K for 2 h. The

products were analyzed by means of a GC equipped with a SPB-20 capillary column and
a FID. The reaction products were confirmed by GC-MS analysis. The internal standard
method was used for quantitative analysis of the products. The conversion of substrate

and selectivities to the products were calculated through carbon balance.

2.3 Results and Discussion

2.3.1 Synthesis

A major objective of this study was to elucidate the structure-reactivity relationships
for mesoporous MCM-41 and HMS titanosilicates prepared at ambient temperature by
electrostatic and neutral surfactant templating pathways, respectively. In order to
minimize the number of compositional and structural variables for these two classes of
materials and to focus on potential differences in local Ti siting by XANES spectroscopy,
the ambient temperature synthesis was performed in such a way that the Ti loading for
the two classes of materials was 2 mol % and the framework-confined mesopore size was
27-29 A. Cetyltrimethylarnmonium cations were used as templating surfactant for both
S+I' and S+X‘I+ electrostatic preparations.

The ambient temperature synthesis of Ti-HMS materials by the neutral S°I° approach

generally affords materials with larger framework-confined pores than the corresponding

S+I‘ or Si‘X'I+ pathways with surfactants of the same alkyl chain lengths. Thus, in order

to prepare Ti-HMS with an average pore size of 27 A, we selected a primary amine

34

template witl

was accompl

2.3.2 Chara
(a) XR,

catalysts pre

Table 2.1 l
Temgratu

Tizf

template with shorter alkyl chain, namely, dodecylamine. In all cases, template removal

was accomplished by calcination in air at 923 K for 4 hr.

2.3.2 Characterization
(a) XRD. Table 2.1 summarizes the properties of the Ti-MCM-41 and Ti-HMS

catalysts prepared by different templating pathways. Included in the Table for

Table 2.1 Properties of Ti-substituted Mesoporous Silicas Prepared by Ambient
Temrature Synthesis UsingDifferent Templating Pathways

 

 

Parametera MCM-41 MGM-41 HMS
(s+I-) (s+x-I+) (S°I°)
Surfactant C16H33N(CH3)3+ C16H33N(CH3)3+ C12H25NH2
Ti:Si ratio (mol):
Initial gel 2.0:98 10:90 22:97.8
Calcined product 22:97.8 25:97.5 2.4:97.6
d100 (A) 38.1 (36.0)b 36.5 (33.0) 40.2 (36.0)
Unit cell (A) 44.0 (41.5) 42.5 (38.1) 46.4 (41.5)
A (A) 4.1 (8.5) 5.6 (9.6) 0 (3.0)
H-K pore size (A) 29 (28) 27 (26) 27 (26)
FWT (A) 15 (12) 15 (12) 19 (15)
SBET (m2/g) 859 (923) 1354 (1345) 1075 (1108)
Vtotal (cm3/g) 0.70 (0.72) 0.92 (0.95) 1.40 (1.42)
Vertcm3/g) 0.68 (0.70) 0.90 (0.92) 0.68 (0.70)
Vtx (cm3/g) 0.02 (0.02) 0.02 (0.03) 0.72 (0.72)
Vth fr 0.03 (0.03) 0.02 (0.03) 1.06 (1.03)

 

a The unit cell parameter = 2d100\/3 and A is the unit cell contraction upon calcination.
FWT is the framework wall thickness obtained by subtracting the Horvath-Kawazoe (H-
K) pore size from the unit cell parameter. The total liquid pore volume Vtotal was
estimated at a relative pressure of 0.95. The volume of framework confined mesopores,
V2,, was determined from the upper inflection point of the corresponding adsorption step.
The volume for textural porosity, V“, was obtained from the equation Vt, = Via... - V2,.
The data in parenthesis are for the pure silica analogs.

35

comparison art

 

parentheses).
essentially com;
the S‘X‘l+ path
achieve a simil.
(based on Si) fr
5*X'l” pathway

All reactio
but Ti-MCNH
Pflfameters (cf.
Sarnpleg pIEpm
could be 3mm
framework Sylz
henna] [6mm
flalnewmk ‘
mcSOPOFOUS my
electrostatic [er
Surfactant-silk;
interacrjOHS in
harnewOrk “Q1
framewom Th

Stto .
my) Suggegr

comparison are the structural parameters for the pristine silicates (the values in
parentheses). It should be noted that both S+I' and S°I° pathways, allow for the
essentially complete incorporation of the Ti in the product at the 2 mol % level, whereas
the S+X'I+ pathway required a 4-fold excess of Ti in the reaction mixture in order to
achieve a similar Ti loading (2.5 mol%). In addition, the yields of crystalline product
(based on Si) for the S+I' and S° I° pathways were more than 85%, whereas that for the
S+X‘I+ pathway was only about 50%.

All reaction products exhibit similar d100, unit cell parameters and H-K pore sizes
but Ti—MCM-4l and Ti-HMS catalysts differ dramatically by their lattice contraction
parameters (cf. Table 2.1). The significant lattice contraction exhibited by Ti-MCM-41'
samples prepared by the electrostatic S+I' and St‘X'I+ pathways is not surprising and
could be attributed to the cross-linking of the significant amount of non-fully condensed
framework sylanol groups upon calcination (see below). In contrast Ti-HMS prepared by
neutral templating does not exhibit lattice contraction due to its more fully cross-linked
framework. We have previously noted 7’23, that the neutral S°I° pathway affords
mesoporous materials with more fully cross-linked and thicker framework walls than the
electrostatic templating pathways. We have attributed this phenomenon to the absence of
surfactant-silica oligomer charge matching and silica-silica oligomer charge repulsive
interactions in forming the framework walls. As shown by the data in Table 2.1 the
framework wall thickness of all materials increases upon Ti incorporation into the
framework. The changes in unit cell size and framework wall thickness (cf. Table 2.1)

strongly suggest Ti incorporation in the walls of the mesoporous silicate framework. An

36

F: 76

”\k T‘

\

\

35.62. .5 .3 83%... 34...: Ba 9.: Q
.Us. 53888 km E Began. .1624... Ba .36: 3. Ba 32:88 .Cbm E 882.. $202-:
can _YEUE O3 3.63 3.329: 85m $5 BE 3 @2332??? new 2.5 3523 com mEozam max Swazi mm 23mm—

Ameohmouv ®~
3 or m a on my 3 m a on m. or m o

a»..F.P_~—Pphhhblppp upphp-h.—--p_b»». LPL.~l»»-p_-

 

 

 

 

 

 

mi: .1205. 51.20....

 

 

 

 

 

mixfi
5.5.2....» piano—2-... j

 

Mtsueim

 

0— 0m u— +m +— Ix +m

 

 

 

 

 

 

 

 

 

 

37

increa

incorp

substit

:3‘
(b

isot r
produc

larger t

increase in the unit cell size parameter has been also noted 24 to occur upon Ti
incorporation into ZSM-S frameworks.
It is noteworthy that the specific surface areas are very similar for the pure and Ti-

substituted samples. On the basis of comparisons of the full N2 adsorption-desorption

isotherms (not shown), the most significant difference between the Ti-substituted
products prepared by neutral and electrostatic templating pathways is in the substantially
larger textural mesoporosity exhibited by the S°I° assembled Ti-HMS derivative (compare

Vtx and Vm/Vfr ratios in Table 2.1). Ti-HMS exhibits ratio of Vtx/Vfr mesoporosity of

1.06 whereas Ti-MCM-41 samples exhibit negligible small ratios in the range of 0.02-
0.03. The enhanced interparticle mesoporosity of Ti-HMS, which is a consequence of the
smaller particle size afforded by neutral surfactant templating, will be shown later to play
an important role in liquid phase catalytic oxidation reactions of large organic molecules.
Figure 2.1 illustrates the XRD patterns of the calcined Ti-MCM-41 and Ti-HMS
materials together with those for the corresponding pure silica analogs. All materials
exhibit well-defined 100 reflections in their XRD patterns. Hexagonal mesostructures
with greater long range order were obtained at ambient temperature conditions by the
S+X'I+ pathway (Fig. 2.1 A). This is evidenced by the presence of additional relatively
narrow 110 and 200 diffraction lines in the XRD patterns of pure and Ti-substituted
MCM-4l samples. The diffraction patterns of the calcined Ti—MCM-4l and pure silica
analog prepared by S+I' pathway (Fig. 2.1 B) are very similar to these of HMS and Ti-
HMS (Fig. 2.1 C). These patterns generally exhibit 100 reflections accompanied with
broader unresolved higher order reflections. These results are not surprising in view of

the ambient temperature conditions used for the synthesis and the lack of NaOH in our

38

.5583 arm 3 83%... 9.5... Ba m2: Q Ea $3888

.5 3 Base $202-? .85 .152 .6 $3588 Form 3 883:. $202-: .05 .162
A3 .8508 3:60—08 3:3 36 SE S coda—5.0.98-3 can 23 wonmmofiimbm mo 880% £22 wee/758

End
co..- cc 7 ow 7 oo ..- co. co. co w- 03. cu w- 2:. cm. om. cow. 2;. cu... cc 7 an. on.
. . . . _ . . . b _ _ . . _ p .

—hhblhbphl-Ib_blb PbPhPhh_Pbl b Pth

 

29.362 29.5»: 295002

macaw

P-nb—th—bDb—b-h—hbn-Ph

.....m .o -..m .m +can d

39

 

Stl' re

silica

(BSClll

bond:

band

S+I' reaction mixtures. We observed similar lack of long range order for a series of pure
silica MCM-41 materials prepared by ambient temperature synthesis by the S+I'
assembly pathway with surfactants of different alkyl chain lengths 23. The lack of long
range order exhibited by the HMS and Ti-HMS samples was attributed to the weak H
bonding forces that govern the neutral S°I° assembly process and to the corresponding
small scattering domain sizes 7.

(b) 29Si MAS NMR Figure 2.2 shows the 2981 MAS NMR spectra for as-synthesized
samples of pure and Ti-substituted mesoporous molecular sieves prepared by ambient
temperature S+I‘, Se‘X‘I+ and S°I° assembly. In general, three bands centered at chemical
shifts of —92, -100, and -110 ppm were observed for the MCM-4l derivatives. These
bands can be attributed to Si(OSi),((OH)4.x framework units where x=2 (Q2), x=3 (Q3)
and x=4 (Q4), respectively. It is noteworthy that all as-synthesized electrostatically-
templated pure and Ti-substituted MCM-41 samples exhibit a higher fraction of
incompletely cross-linked Q2 and Q3 framework units than the S°I° HMS samples (see
Figure 2.2). In addition, the as-synthesized pure and Ti-substituted HMS samples exhibit
a much higher average ratio of Q4/(Q3+Q2) units (~2.8) than the as-synthesized pure and
Ti-substituted MCM-4l samples prepared by electrostatic templating (~1.0). This
implies that neutral templating allows for the preparation of mesoporous molecular sieves
with more completely cross-linked frameworks. A noteworthy trend is observed by
comparing the spectra of pure and Ti-substituted MGM-41 samples prepared by both
electrostatic templating pathways. Significantly, Ti incorporation leads to a dramatic
increase of the cross-linking of the MCM-41 framework. This is evidenced by the

increase in fraction of Q4 sites at the expense of both Q2 and Q3 framework sites. In the

40

- y...—

case of 1015

significant (1

framework.
An add

conespondi

case of HMS, the increase of the fraction of Q4 sites upon Ti incorporation is not as
significant due to the much greater degree of cross- linking exhibited by the pure silica
framework.

An additional consequence of Ti framework incorporation is the broadening of the

corresponding Q2, Q3 and Q4 peaks in the 29Si MAS NMR spectra. This broadening

 

 

 

 

 

 

210 e 325 a. HMS, S°I°
f b.Ti-MCM-41,
S’X'I”
c. Ti-HMS, 8° 1“
d. Ti-MCM-41, S t l'
e. Anatase
f.TS-1
260
I d
8 G
u. S '
I c
|
b I
I
a
N

 

 

 

200 250 300 350 400 450 500
Wavelength (nm)

Figure 2.3 UV-VIS spectra of calcined Ti-containing materials.

41

could be attributed to the effect of the Ti sites on the chemical environment of the
adjacent Si atoms.

(c) UV-VIS Diffuse Reflectance Spectroscopy. The incorporation of titanium into
MCM-41 and HMS frameworks was further verified by UV-VIS diffuse reflectance
spectroscopy. The corresponding spectra of the Ti-substituted samples are shown in
Figure 2.3, along with those for TS-l, HMS and anatase. The spectrum for TS-l shows
an absorption band at 210 nm, whereas bulk titania (anatase) shows a band at 325 nm.
The band at 210 nm was attributed to ligand-to-metal charge transfer associated with
isolated Ti (IV) framework sites in tetrahedral coordination 25. The broad absorption
band centered at 325 nm is typical for ligand-to-metal charge transfer occurring in bulk
titania. The spectra for the mesoporous Ti-MCM-41 and Ti-HMS derivatives lack the
330 nm band characteristic for segregated titania. This suggests that most of the Ti atoms
in our Ti-MCM-41 and Ti-HMS samples occupy site-isolated positions in the silica
framework.

The spectra for Ti-MCM-41 and Ti-HMS also are clearly different from the
spectrum of the microporous TS-l. This is manifested by the much broader character of
the absorption bands centered at 220 and 260-270 nm. The possibility of some Ti-O-Ti
clustering in the framework cannot be unequivocally precluded as amorphous TiO2/SiO2
gels with various Ti content exhibit absorption bands in the intermediate range of 250-
330 nm 4. However, the 220 nm band clearly signifies that much of the Ti is in site-
isolated form. The slight shift and the increase in the width of the 220 nm band may be
indicative of a Ti in a distorted tetrahedral environment or of the presence of Ti species in

13.14

an octahydral coordination sphere. However, since there is very little shifting of the

42

band toward lower wavelengths upon thermal dehydration, we favor a distorted
tetrahydral environment for most of the Ti sites in Tti-MCM-41 and Ti-HMS. The
XANES data provided below verify this assignment. We believe that this distorted
environment is a direct consequence of the amorphous character of the pore walls (i.e., a
wide variety of Ti-O-Si bond angles).

According to several literature studies, 27 the absorption shoulder at 260-270 nm can
be attributed to the presence of site-isolated Ti atoms in penta- or octahedral
coordination. Similar UV-VIS behavior, indicative of a high fraction of higher
coordinated Ti sites, also has been reported to occur upon hydration of TS-l 27. The
presence of a significant fraction of Ti sites with coordination numbers higher than four
in Ti-MCM-41 and Ti-HMS may be associated with the less crystallographic order in the
pore walls, the much higher accessible surface area, and the larger mesopores. These later
factors, most likely are responsible for the observed enhanced hydration of Ti sites in
MGM-41 and HMS relative to silicalite-1.

It is important to note that the UV-VIS spectra of the electrostatically templated Ti-
MCM-41 and Ti-HMS prepared by neutral templating are very similar and practically
indistinguishable. This implies that the effect of the templating method on the Ti siting in
Ti-MCM-41 and Ti-HMS samples prepared under ambient temperature conditions is
negligible. Therefore, any differences in the catalytic behavior of these materials may

not be attributed to intrinsic differences in Ti-siting.

(d) X -Ray absorption. In order to better elucidate the nature of the Ti-sites in

mesoporous Ti-MCM—41 and Ti-HMS samples we have performed Ti K-edge XANES

measurements 21. For comparison purposes we have also obtained Ti K-edge

43

 

 

annualized Absorption

 

 

 

    

l 1 l 1 I l

4.96 5.0 5.04
E (keV)

 

Figure 2.4 Ti K-edge XANES spectra of reference compounds containing Ti in
tetrahedral and octahedral oxygen environments: (a) anatase, (b) rutile, (c)
MgTiO3, (d) neptunite, and (e) spirotitanate.

44

Spj

Table 2.2 Ti K-edg XANES Results for Ti-Containifl Materials '

 

Sample position normalized FWHMb
(eV) height (eV)
Spirotitanate 4970.2 0.95 0.87
rutile 4971.7 0.20 l .5
anatase 4972.1 0.15 1.7
M8Ti03 4971.5 0.27 0.8b
neptunite 4971 .6 0.33 l .2
TS-l 4970.5 0.52 1.0
(2 % Ti)
Ti-MCM-4l 4970.7 0.30 1.3
(S+I', 2 % Ti)
Ti-MCM-4l 4970.7 0.3 1 1.3
(S+X'I+,2.5 % Ti)
Ti-HMS 4970.8 0.31 1.3
(S°I°, 2.4 % Ti)

 

“The Ti K-edge XANES parameters were obtained by fitting the central
pre-edge peak to a Lorentzian Function. bl'he full width at half maximum,
FWHM, is for the central pre-edge peak.

XANES spectra of reference compounds such as octahedrally coordinated anatase, rutile,
MgTiO3, distorted octahedrally coordinated neptunite, tetrahedrally coordinated
spirotitanate and TS-l. Figure 2.4 shows the Ti K-edge XANES spectra of the reference
compounds. All reference compounds with nearly octahedral Ti site symmetry exhibit
multiple, low intensity pre-edge peaks in the region from 4960 to 4980 eV (cf., spectra a,
b). Distortion of the octahedral geometry (inversion symmetry) leads to an increase in
intensity of the central peak, as shown for neptunite 28 (spectra c and d). In contrast, the

spirotitanate with nearly regular tetrahedral geometry exhibits a single, very high

45

intensity pre-edge peak (spectrum e). These results (see Figure 2.4 and Table 2.2) are in
good agreement with previous studies. 29'” We may conclude, therefore, that tetrahedral
Ti sites give a single, high intensity pre-edge peak, whereas Ti-sites in regular octahedral
symmetry afford multiple, low intensity pre-edge peaks. Although tetrahedral and highly
distorted octahedral Ti sites exhibit similar single pre—edge peak, the two environments
clearly are distinguishable by means of peak intensity and position.

The Ti K-edge XANES spectra of the mesoporous Ti-MCM-41 and Ti-HMS
samples prepared by electrostatic and neutral templating methods, respectively, are
shown in Figure 2.5. Included for comparison is the spectrum for microporous TS-l.
The following features are evident: (i) the spectra for all three samples are very similar
and contain a sharp pre-edge peak, (ii) the energy position of the pre-edge peak for Ti-
substituted MCM-41 and HMS is similar to that for TS-l and spirotitanate, but different
from the reference materials containing octahedral Ti (see Table 2.2), (iii) the two Ti-
MCM-41 samples and Ti-HMS exhibit weaker pre-edge peaks that are slightly shifted
toward higher energies and wider FWHM than those of TS-l and spirotitanate,
suggesting in addition to tetrahedral Ti the probable presence of some Ti higher
coordination sites (see below); (iv) most significantly, the spectral parameters for
calcined Ti-substituted MCM-41 and HMS molecular sieves prepared via different
assembly pathways at ambient temperature are very similar and practically
indistinguishable. These observations are also supported by our UV-VIS results and
show that the Ti siting in mesoporous MCM-41 and HMS materials is independent of the

templating pathway.

46

 

normalized Absorption
fl

 

 

 

 

r l r r

 

L l I l l
4.96 5.0
E (keV)
Figure 2.5 Ti K-edge XANES spectra of calcined (a) Ti-HMS prepared by S°I°
assembly, (b) Ti-MCM-4l prepared by S“X‘I+ assembly, (c) Ti-MCM-

41 prepared by S‘T assembly, and (d)TS-1.

47

 

 

 

 

 

Extending somewhat further our interpretation of the XANES data, we note that
Farges et al. performed ab initio multiple-scattering calculations of the Ti K-edge
XANES spectra for mixtures of reference compounds containing Ti atoms with various
coordination. 33 For instance, they calculated that the normalized height of the central
pre-edge peak varies from 0.6 to 1.0 for 4~fold Ti, from 0.4 to 0.7 for 5-fold Ti and from
0.05 to 0.27 for 6-fold Ti containing compounds. Besides the differences in intensity, the
4- and 6~fold Ti containing compounds were found to differ by energy position of the

pre-edge peak. The energy position of the pre-edge peak for 6cfold Ti containing

compounds was found to be shifted on average of 1.5 eV toward higher energies. 33
Thus, structures containing mixtures of 4-fold Ti with 5-fold or 6-fold coordinated Ti
sites should exhibit pre-edge peaks with lower intensity that are slightly shifted toward
higher energies. Therefore, the coordination state of Ti-sites in our Ti-MCM-41 and Ti-
HMS materials may well be a mixture of 4-, 5- and 6-fold coordinations with tetrahedral
coordination being in the majority. The higher coordination Ti-sites, are most likely

generated through hydration of the tetrahedrally coordinated sites.

2.3.3 Catalytic results

Corma et al. have demonstrated 9 that Ti-MCM-41 catalyzes the epoxidation of
rather small organic molecules such as hex-l-ene and norbomene. We have reported that
Ti-MCM-41 and Ti-HMS are very effective oxidation catalysts for bulky aromatic
molecules that can not be converted over TS-l, such as 2, 6-DTBP 10. More recently,
several groups have elaborated on the preparation, characterization and catalytic activity
of Ti-MCM-41 ”‘15 and Ti-HMS 534. All of the Ti-MCM-4l catalysts investigated to

date were prepared by prolonged hydrothermal synthesis above 373 K using the

48

 

A. N 8.. Mam .m masses 8 2.8038 .80.. a... 8... ON: 2. 0: 2 «OS: 68 .3 0m 8...... cm .3388 m... cm .8 8.8.88 80.23.
0.2.8.. 8.8.8. $22.... a 8 .8288... 982:... .5520: Z .6 .23 .8382. .8 82.2.8.2. a...» 82892808 «02.: 8 882.5
88.2.. 2... 0.82.8 8:88... 8 323828 03.22.88 8 23.8.98 83 0.3828 08:80 9 8.93.3 13.0... 2 >22 a

 

 

m. m... am 8 o 3 o o 68.22
new.
3 .8 mm 2. S. P ma 8 we 22..-...
5-8.9
.e om mm a... .2. MK 8 8 me .8202...
A-~+mv
mm a a... .3 me a. o. 8 o... .5202...
E m n... v. E 4... .K on .22
as .2... as .25 as .2... .28 .2...
.28 8.9.. as .25 .2. .2... 28.9.. .628 as .2... 8 >2). .2. .2... 82.28
3:5 0.>:00 0:0:30 .>:00 0028 BR— ofimxomm OIUSQ .>:00 goo—om .>:00
.9882. NONI 38282.8 mmhabd 82.2.8 0.825 82.28 <22

 

.9335...— wfiaaifioh 0:88.50 mama: 28.—«Fa muggomfioh
2.2.1.2. .3 2.8.8...— 8>2m 81.882 88 82 N0 08353.6-3 .8 .3233. 23.380 QN «Sun.

49

electrostatic S+I‘ pathway originally demonstrated by the pioneering work of Mobil 5.
Here we compare the catalytic activity of Ti-MCM-4l and Ti-HMS samples prepared by
ambient temperature synthesis using both S+I' and S+X'I+ electrostatic templating
pathways and the neutral S°I° templating pathway, respectively. The catalytic
performance of the samples was tested toward olefin and aromatic substrates of different
size, such as the relatively small methyl methacrylate (MMA), styrene and the bulky 2,6-
di-tert-butyl phenol (2,6-DTBP). The catalytic results are summarized in Table 2.3.

The mesoporous Ti-MCM-41 and Ti-HMS exhibit higher catalytic activity than
microporous TS—l for all reactions. The difference in catalytic activity between the
mesoporous and microporous titanosilicates increases with increasing of the substrate
size. The difference in catalytic activity is very small for the relatively small and
elongated MMA molecule (2-2.5 times higher conversion for the mesoporous catalysts).
However, as the substrate molecule becomes larger and more bulky (styrene and 2,6-
DTBP) the difference becomes much more pronounced. Thus, the difference in catalytic
activity among Ti-MCM-41, Ti-HMS and TS-l reaches a maximum in the case of 2,6-
DTBP (4 - 11 times higher conversion). This result is not surprising given the fact that
the Ti-active centers are much more accessible in the larger mesopore size frameworks of
Ti-MCM-4l and Ti-HMS. Due to the small micropore size of TS-l (~6 A), the large 2,6-
DTBP substrate can not penetrate into the framework pores and therefore can not be
transformed (only 4.5 mol % conversion). On the other hand, anatase (bulk titania) was
completely inactive for MMA and styrene oxidation. Its catalytic activity for the

conversion of the large 2,6-DTBP is also negligible (only 4.5 mol %).

50

The oxidation of MMA proceeds with higher selectivity to methyl pyruvate (MPV)
over Ti-MCM-41 and Ti-HMS, but the microporous TS-l affords slightly higher
selectivity to epoxide than Ti-MCM-41 and Ti-HMS (epoxide selectivity not included in
Table 2.3). The tendency of TS—l to oxidize olefins primarily to epoxides is well
documented 35. A similar trend also is observed for styrene oxidation. Generally, Ti-
MCM-4l and Ti-HMS afford higher selectivity to diol, whereas more epoxide is
produced over the microporous TS-l. The fact that the oxidation of MMA and styrene
over Ti-MCM-41 and Ti-HMS proceeds selectively toward the alcohol derivative
(through the epoxide intermediate) could be a consequence of the more hydrophilic
nature of the amorphous pore walls.

Notari et al. have reported that bulk titania or substituted silicas containing Ti-O-Ti
bonds are not suitable for catalytic peroxide oxidation reactions because they selectively
decompose H202.4 The comparison of the results for the H202 decomposition over our
samples reveals a very interesting trend (Table 2.3, last column). In accordance with

Notari et al. our anatase sample affords 13 mol % decomposition of H202. Both TS-l
and Ti-HMS exhibit lower H202 peroxide decomposition activity, whereas Ti-MCM-41
samples prepared by electrostatic 8+1' and S+X'I+ templating methods exhibit a higher
tendency to decompose H202. Especially noteworthy is the higher degree of H202
decomposition by Ti-MCM-41 prepared by the acidic counterion-mediated pathway S+X'
1+. This result is probably related to the pronounced tendency of the acidic S+X'I+
pathway to impede incorporation of Ti atoms in the silicate framework. All of our

syntheses involving the counterion-mediated pathway required at least a two to four-fold

excess of Ti in the initial gel in order to achieve the desired Ti doping. It may be that the

51

 

strongly acid
mesoporous

A furth
exhibits acti
samples for.
pronounced
2.1. Ti-MC
little or n
characteris
meSOporos
usually eq
These dif‘
molecular
szzs d.
liquid pha
meSOPOIOS

large (3er

2'4 Conclu

Hex;lg
Prepared a:
approac h Es

{Elma} S°I°

strongly acidic conditions used in the synthesis limit the incorporation of Ti atoms in the
mesoporous framework.

A further comparison of the catalytic data reveals that S°I° -assembled Ti-HMS
exhibits activity superior to the electrostatically S+I' and S+X'I+-assembled Ti-MCM-41
samples for all reaction systems. This enhanced activity for Ti-HMS is especially
pronounced in the case of the large 2,6-DTBP. As shown by the data presented in Table
2.1, Ti-MCM-41 samples possess predominantly framework-confined mesoporosity and
little or no significant textural mesoporosity. 10 In contrast, Ti-HMS samples
characteristically exhibit significant complementary textural or interparticle
mesoporosity. The ratio of textural to framework-confined mesoporosity of Ti-HMS is
usually equal or higher than 1, whereas that of Ti-MCM-4l is usually close to zero.
These differences in catalytic behavior of Ti-HMS and Ti-MCM-41 mesoporous
molecular sieves are not due to differences in Ti siting, as judged from UV-VIS and
XANES data. We conclude, therefore, that the superior catalytic activity of Ti-HMS in
liquid phase oxidations is most likely due to the presence of complementary textural
mesoporosity that facilitates access of the framework-confined mesopores, especially by

large organic substrates.

2.4 Conclusions

Hexagonal mesoporous titanosilicates containing 2 mol % Ti - substitution can be
prepared at ambient temperature synthesis via three different surfactant templating
approaches (S+I‘, S+X‘I+, and S°I°). Electrostatic S+I' assembly of Ti-MCM-41 and

neutral S°I° assembly of Ti-HMS allow for almost complete incorporation of Ti in the

52

mesoporous framework (at 2 mol % doping level), whereas electrostatic assembly of Ti-
MCM-41 under acidic conditions via a counterion-mediated S+X‘I+ pathway impedes Ti
incorporation. The incorporation of Ti in the anionic framework of MCM-4l materials
and in the neutral framework of Ti-HMS mesopore frameworks is accompanied by an
increase of the lattice parameter and the framework wall thickness. The framework
mesopore size of the MCM-4l and HMS samples remains unaffected by the framework
incorporation of Ti atoms. The degree of framework cross-linking of the electrostatic
S+I' and S+X'I+-templated MGM-41 samples is significantly improved by Ti-substitution
as judged from the 29Si MAS NMR data. The UV-VIS and XANES results for calcined
forms of Ti-MCM-41 and Ti-HMS provide evidence that: (i) Ti species are incorporated
and site-isolated into the framework structure; (ii) Ti atoms are predominantly in a
tetrahedral coordination, but there is the possibility of some Ti sites in five- or six-
coordinated sites . (Higher coordinated Ti-sites most likely are generated by rehydration
of some of the tetrahedrally coordinated sites), and (iii) the Ti siting for the S+I' and S+X'
1+ Ti-MCM-41 and for S°I° Ti—HMS are practically indistinguishable. This implies that
the assembly pathway method has little or no effect on the Ti siting, at least when the Ti-
MCM—4l and Ti-HMS samples are prepared at ambient temperature conditions.

A comparison of the catalytic results for the liquid phase peroxide oxidations of
methyl methacrylate (MMA), styrene and 2,6—di-tert-butylphenol (2,6-DTBP) reveals
that Ti-MCM-4l and Ti-HMS exhibit catalytic activities higher than TS-l for all
reactions. The difference in activity between the mesoporous molecular sieve catalysts
and microporous TS-l increases with increasing substrate size. The difference in activity

is small for the relatively small MMA molecule, but as the substrate becomes larger, as

53

 

in styrene a
pronounced.
centers emb
Due to the .
pnettnte int
li-lhlS ex
electrostatn
especially l
lit-NBS do
be attnbut
thicker fra.
sites Can it
Stmctural
mesoporo
facilitates

e“harming

2'5 Refer

in styrene and, especially 2,6-DTBP, the difference in activity becomes much more
pronounced. This result is not surprising given the much more accessible Ti-active
centers embedded in the larger mesopore size frameworks of Ti-MCM-4l and Ti-HMS.
Due to the small micropore size of TS-l (~6 A) the large 2,6-DTBP substrate can not
penetrate into the framework pores and therefore can not be transformed. S°I° -assemb1ed

Ti-HMS exhibits greater catalytic activity than the Ti-MCM-41 materials assembled by

electrostatic S+I‘ and S+X'I+ pathways The superior performance of Ti-HMS is
especially pronounced in the case of the large 2,6-DTBP. On the basis of UV-VIS and
XANES data, these differences in catalytic behavior of Ti-HMS and Ti-MCM-41 cannot
be attributed to differences in Ti coordination environment. Bercause Ti-I-IMS has
thicker framework walls than Ti-MCM-4l, an enhancement in the fraction of surface Ti
sites can not be responsible for the greater activity of Ti-HMS. The most distinguishing
structural feature between Ti-HMS and Ti-MCM-41 is the greater inter-particle (textural)
mesoporosity for Ti-HMS. These complementary textural mesoporosity most likely
facilitates substrate transport and access of the framework-confined mesopores, thus

enhancing the catalystic efficiency of Ti-HMS.

2.5 References

1. a. Huybrechts, D. R. C.; De Bruycker, L.; Jacobs, P. A. Nature 1990, 345, 240; b.
Esposito, A.; Neri, C. ; Buonomo, F. US Patent 1984 , 4,480,135.

Clerici, M. G.; Romano, U. US Patent 1990, 4,937,216.

Taramasso, M.; Perego, G.; Notari, B. US Patent 1983, 4,410,501.

Notari, B. Stud. Surf. Sci Catal. 1988, 37, 413.

a. Kresge, C. T.; Leonowicz, M. E.; Roth W. J .; Vartuli, J. C.; Beck, J. S. Nature
1992, 359, 710; (b) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.;

9‘5“?!”

54

 

 

 

Margo

Firouz:
l 176.

7. Tanev,

 

8. aAttar.
S. A.; P
9. Corma.
1994.14
10a Time
Pinnava
Materiaj
Eds.).11
11.Franke‘
Related
Weitkan
Elsevjer
1241 Sanka
DentA.
F.;SankE
1331mm,)
14, Remy. 1
Synthesis
PTCSS.
15. 00mm! 3
16' Zeitler, V_

17. HUtsyhOuSe

Kresge, C.T.; Schmitt, K. D.; Chu, C. T-W.; Olson, D. H.; Sheppard, E. W.;
McCullen, E. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114,
10834.

6. a. Huo, Q.; Margolese, D. I.; Ciesla, U.; Feng, P.; Gier, T. E.; Sieger, P.; Leon, R.;
Petroff, P. M.; Schuth, F.; Stucky, G. D. Nature 1994, 368, 317; b. Huo, Q.;
Margolese, D. I.; Ciesla, U.; Demuth, D. G.; Feng, P.; Gier, T. E.; Sieger, P.;
Firouzi, A.; Chmelka, B. F.; Schuth, F.; Stucky, G. D. Chem. Mater. 1994, 6,
1176.

7. Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865.

8. a. Attard, G. S.; Glyde, J. C.; Goltner, C. G. Nature 1995, 378, 366; b. Bagshaw,
S. A.; Prouzet, E.; Pinnavaia, T.J. Science 1995, 269, 1242.

9. Corma, A.; Navarro, M. T.; Perez-Pariente, T. J. Chem. Soc. Chem. Commun.
1994, 147.

10. a. Tanev, P. T.; Chibwe, M.; Pinnavaia, T. J. Nature 1994, 368, 321; b.
Pinnavaia, T. J.; Tanev, P. T.; Jialiang, W.; Zhang, W. in “Advances in Porous
Materials” MRS Symp. Proc. Ser. (Sh. Komameni, J. S. Beck and D. M. Smith,
Eds.), 1995, Vol. 371, p. 53, MRS, Pittsburgh.

11. Franke, 0.; Rathousky, J .; Schulz-Ekloff, G.; Starek, J .; Zukal, A. in “Zeolite and
Related Microporous Materials: State of the Art 1994, Stud. Surf. Sci. Catal.“ (J.
Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich, Eds.), 1994, Vol. 84, p. 77,
Elsevier Sci. BV, Amsterdam.

12. a. Sankar, G.; Rey, F.; Thomas, J. M.; Greaves, G. N.; Corma, A.; Dobson, B. R.;
Dent, A. J. J. Chem. Soc., Chem. Commun. 1994, 2279; b. Maschmeyer, T.; Rey,
F.; Sankar, G.; Thomas, J. M. Nature 1995, 378, 159.

13. Blasco, T.; Corma, A.; Navarro, M. T.; Pariente, J. P J. Catal. 1995, 156, 65.

14. Reddy, J. S.; Dicko, A.; Sayari, A. in “Third International Symposium on
Synthesis of Zeolites and Expanded Layered Compounds” Anheim, 1995, in
press.

15. Gontier, S.; Tuel, A. J. Catal. 1995, 157, 124.

16. Zeitler, V. A.; Brown, C. A. J. Am. Chem. Soc. 1957, 79, 4618.

17. Hursthouse, M. B.; Hossain, M. A. Polyhedron 1984, 3, 95.

55

 

i1 18.1(0m
l9.Ftuh1
20. Frah
211511:

22. Hor
23. Tan
24. van
25. Boc

C at
26. (30.!
27.21. (

l9!
Sm
38. Ca
29. W.
30. Be

31.Y;
33- G:

33. F

C

18. Kortum, G. "Reflectance Spectroscopy", Springer-Verlag, New York, 1969.

19. Frahm, R. Nucl. Instrum. Meth. Phys. Res. 1988, 270, 578.

20. Frahm, R.; Wang, J. Jpn. J. Appl. Phys. 1993, 32, 188.

21. Lytle, F. W.; Greegor, R. B.; Sandstrom, D. R.; Marques, D. R.; Wong, J.; Spiro,
C. L.; Huffman, G. P.; Huggins, F. E. Nucl. Instrum. Meth. 1984, 226, 542.

22. Horvath, G.; Kawazoe, K. J. J. Chem. Eng. Jpn. 1983, I6, 470.

23. Tanev, P. T.; Pinnavaia, T. J. Chem. Mater. 1996, 8, 2068.

24. van Koningsveld, H.; Jansen, J. C.; van Bekkum, H. Zeolite 1990, 10, 235.

25. Boccuti, M. R.; Rao, K. M.; Zecchina, A.; Leofanti, G.; Petrini, G. Stud. Surf Sci.
Catal. 1989, 48, 133.

26. Gontier, 8.; Tue], A., Zeolites 1995, 15, 601.

27. a. Geobaldo, C. F.; Bordiga, S.; Zecchina, A.; Giamello, E. Catal. Lett. 1992, 16,
109; b. Bonneviot, L.; Trong On, D.; Lopez, A. J. Chem. Soc., Chem. Commun.
1993, 685; c. Trong On, D.; Denis, 1.; Lortie, C.; Cartier, C.; Bonneviot, L. Stud.
Surf. Sci. Catal. 1994, 83, 101.

28. Cannillo, E.; Mazzi, F.; Rossi, G. Acta Cryst. 1966, 21, 200.

29. Waychunas, G. A. Am. Min. 1987, 72, 89.

30. Behrens, P.; Felsche, J .; Vetter, S.; Schulz-Ekloff, Jaeger, N. I.; and Niemann, W.
J. Chem. Soc., Chem. Coomun. 1991, 678.

31. Yarker, C. A.; Johnson, P. A. V.; Wright, A. C.; Wong, J.; Greegor, R. B.; Lytle,
F. W.; Sinclair, R. N. J. Non—Cryst. Solids 1986, 79, 117.

32. Greegor, R. B.; Lytle, F. W.; Sandstrom, D. R.; Wong, J.; Schultz, P. J. Non-
Cryst. Solids 1983, 55, 27.

33. Farges, F.; Brown Jr., G. E.; Navrotsky, A.; Gan, H.; Rehr, J. J. Geochimica
Cosmochimica Acta 1995, submitted.

34. Sayari, A.; Karra, V. R.; Reddy, J. S.; Moudrakovski, I. L. in “Advances in
Porous Materials” MRS Symp. Proc. Ser. (Sh. Komameni, J. S. Beck and D. M.
Smith, Eds.), 1995, Vol. 371, p81, MRS, Pittsburgh.

35. Kumar, S. B.; Mirajkar, S. P.; Pais, G. C. G.; Kumar, R; Kumar, R. J. Catal.
1995, 156, 163.

56

 

Abstract
Wat

framework.
neutral (8"
orthosilicatu
= 90 : 10
textural p0;

mitture, w;

Chapter 3

Tailoring the Framework and Textural Mesopores of
HMS Molecular Sieves Through an Electrically
Neutral (S°I°) Assembly Pathway

Abstract

Water : ethanol solvent mixtures of differing polarity have been used to tailor the
framework and textural mesopores of HMS molecular sieve silicas through an electrically
neutral (S°I°) assembly pathway (8° = dodecyl or tetradecylamine; I° = tetraethyl
orthosilicate). Mesostructure assembly from a water-rich solvent mixture, water : ethanol
= 90 : 10 (v/v), afforded wormhole-like framework structures with a complementary
textural pore volume equal in magnitude to the framework pore volume. An ethanol-rich
mixture, water : ethanol = 35 : 65 (v/v), also formed wormhole-like frameworks, but the
textural porosity was less than 20% of the framework pore volume. HMS derivatives
with high textural porosity were comprised of mesoscale fundamental particles that
aggregate into larger particles. In contrast, HMS mesostructures with low textural
porosity were assembled into much larger aggregates of macroscale spheroid to disk-
shaped fundamental particles. The differences in particle textures were attributed to
differences in I° hydrolysis rates and S°I° nucleation and growth rates in the two solvent
systems. The presence of mesitylene in the reaction mixtures resulted in an expansion of
the framework pores under water-rich conditions. Pore contraction, however, was
observed with mesitylene present under ethanol-rich conditions. This versatile structure-
modifying prOperty of mesitylene in S°I° assembly is explained by the solvent-dependent

binding of the aromatic molecules to two structurally distinct and size-altering

57

“dissolved” and “adsorbed” states at the centers and interfacial surfaces of the surfactant
micelle, respectively. Thus, both the framework and the textural pores of HMS silica can
be readily tailored to the needs of a particular materials application through S°I° assembly

by a judicious choice of an appropriate solvent and an auxiliary structure modifier.

58

 

(to

110
by

me

5+
am
Sill.
silt.
Slllr

elec

”II-pr

3.1 Introduction

The discovery of Mobil MCM-41 mesoporous molecular sieves 1 has stimulated
great interest in the surfactant-directed assembly of mesostructures by sol-gel methods.
Thus far, three general assembly pathways have emerged. First, the electrostatic charge
matching pathway between cationic or anionic surfactant micelles and charged inorganic
precursors. Among these, the so-called S+I‘ and S‘I+ pathwayszr3 represent the most
commonly encountered electrostatic assembly pathways. Counterion mediated S+X‘I+
and S'M+I' assembly processes, where surfactant and inorganic reagents are brought
together at the micelle interface through triple ion interactions, represent extensions of the
charge matching pathway3. The second pathway pairs neutral amine surfactants (S°) or
nonionic polyoxyethylene surfactants (N°) with neutral inorganic precursors (1°) through
hydrogen bonding at the S°I° or N °I° interface4'3. The third general assembly route to
mesostructures exploits dative bond formation between donor groups on the surfactant
and a metal acceptor centers in the inorganic precursor9.

The original synthesis of hexagonal MCM-4l silicas was accomplished through a
S+I‘ assembly mechanism using quaternary ammonium ions as the surfactant and silicate
anions as the inorganic precursor]. Related mesoporous structures, designated HMS
silicas, have been obtained through S°I° assembly, wherein S° is an alkylarnine and P is a
silicon alkoxide5. In general, surfactant removal from the neutral frameworks HMS
silicas can be achieved by simple solvent extraction, whereas the displacement of the
electrostatically bound surfactants from the anionic framework of as-synthesized MCM-
41 requires proton exchange or destruction of the surfactant by combustion“). More

importantly, there are significant structural differences between MCM-41 and HMS

59

silicas. MCM—41 derivatives typically exhibit three or more X-ray diffraction lines
indicative of long range hexagonal channel packing. In contrast, HMS silicas show one
or, at most, two broad X-ray peaks as a consequence of a small crystallite domain size
and/or a much lower degree of channel packing order.

Although they are distinguishable with regard to structural ordering, MCM-41 and
HMS molecular sieves both exhibit a sharp step in their nitrogen adsorption isotherms,
corresponding to the presence of a regular mesoporous framework. Owing to the very
small elementary particle size of many HMS derivatives, they can exhibit complementary
textural mesopores, in addition to framework pores. The textural pore volumes for HMS
can be up to 1.5 or more times as large as the framework pore volumes, whereas MCM-
41 exhibits very little textural mesoporosity“. The textural mesopores are important
because they greatly facilitate mass transport to the framework mesopores. For this
reason the catalytic reactivity of HMS is usually superior to MCM-41, especially for
conversions involving large substrates in a liquid reaction medium where reaction rates
are diffusion limited”.

Another potential benefit of S°I° assembly is the possibility of conducting
mesostructure synthesis in media of diverse polarity. Unlike their ionic counterparts, 8°
and 1° reagents are generally soluble in a wide range of solvents. Thus, solvation effects
on the rates of hydrolysis and assembly might be an effective means of controlling
structure. One of the objectives of the present study was to examine the possibility of
tailoring both the framework and textural mesopores of HMS silicas by controlling the

polarity of the reaction medium in which the assembly process is carried out. Our

60

 

approach It
rich and the

We
directing :
equilibriur
nneeHesl3
structure 1
be of gr.

materials

3.2 Expe

approach makes use of two water : ethanol compositions to control polarity, one water-
rich and the other ethanol-rich.

We also have investigated the role of mesitylene as an auxiliary structure-
directing agent in both solvent systems. By using solvation effects to shift the
equilibrium between structurally distinct binding states of mesitylene in the surfactant
micellesl3‘15, we are able to effect an expansion or contraction of the framework pore
structure of HMS. The ability to control both framework and textural mesoporosity can
be of great value in designing HMS materials as catalysts, adsorbents and sensor

materials.

3.2 Experiment

HMS molecular sieves were prepared by S°I° assembly pathways in water :
ethanol solvent mixtures of differing composition and polarity. In both reaction media
tetraethyl orthosilicate (TEOS) served as the neutral silica precursor and dodecylamine
and tetradecylamine were the neutral structure director. In a typical synthesis the
surfactant was dissolved in ethanol, and then the desired amount of water was added
under vigorous stirring to obtain a homogeneous solution. TEOS was added to the
surfactant solution and the mixture was allowed to react under stirring at ambient
temperature for about 20 h. The reactions were carried out in an open beaker in a well
ventilated hood to allow for some evaporation of solvent and concentration of the solid
reaction products. When mesitylene was used as an auxiliary structure director, it was
added to the surfactant solution and stirred for 15 minutes before the addition of TEOS.

All of the HMS reaction products were filtered and dried in air. Although the surfactant

61

 

 

HE:

can be readily removed from HMS mesostructures by solvent extraction, there was no
need to recover the surfactant from the small quantities of products formed in the present
work. Consequently, the as synthesized products were directly calcined at 650 0C in air
for 4 h to simutaneously remove the surfactant and dehydroxylate the framework.

For the purposes of probing the effect of solvent polarity on textural porosity, it
was desirable to form HMS silicas from water—rich and ethanol-rich solutions with
equivalent framework porosities. To achieve equivalent framework pore structures, we
used reagent concentrations in the ethanol-rich system that were twice the concentrations
for the water-rich medium. For HMS assembly in the relatively low polarity, ethanol -
rich reaction medium, where water : ethanol volume ratio was 35 : 65 , the molar
composition of the reaction mixture was 1.0 TEOS : 0.25 Surfactant : 18 EtOH : 34HzO.
For assembly in a relatively high polarity, water - rich solvent mixture, namely, 90 : 10
(v/v) water : ethanol, the molar composition was 1.0 TEOS : 0.25 Surfactant : 10 EtOH :
130 H20.

Powder X-ray diffraction patterns were measured using Cu-Ka radiation (£21542
A) and a Rigaku Rotaflex diffractometer equipped with a rotating anode operated at 45
kV and 100 mA. The scattering and receiving slits were 1/6 degree and 0.3 degree,
respectively.

N2 adsorption and desorption isotherms at -196 0C were obtained on a Coulter
Omnisorp 360CX Sorptometer operated under continuous adsorption mode. Pore size
distributions were calculated from the N 2 adsorption branch using the Horvath—Kawazoe

model. 16

62

 

lOOCX

accelerat
estimate
high ma
about 30
suspensi
minutes.

sectionir

3-3 Resu
l

[Cmperat
mesosm
The Othe
3 ethanol
StruCture
and the e
35 an aUX
Fl
Silicas
respectivg

su .
“Clam

Transmission electron microscopy (TEM) studies were carried out on a JEOL
lOOCX instrument using an electron beam generated by a CeB6 filament and an
acceleration voltage of 120 kV. The resolution of the instrument was about 6 A, as
estimated by indirect measurement of the spherical aberration constant 17 under medium—
high magnification (i.e., 100,000X). Therefore, it was possible to resolve pores above
about 30 A. The specimens were prepared by dipping a carbon coated copper grid into a
suspension (0.1 wt %) of mesoporous material in ethanol that was pre-sonicated for 10
minutes. Attempts to use thin-sectioned specimens were abandoned, because thin

sectioning caused damage and loss of texture-pore information.

3.3 Results

Two versions of HMS silicas were prepared through S°I° assembly at ambient
temperature in reaction media that differed in solvent polarity. In one reaction system the
mesostructures were formed from a “water—rich” solution of 90 : 10 (v/v) water : ethanol.
The other reaction medium was a less polar “ethanol-rich” solution of 35 : 65 (v/v) water
: ethanol. Two S° surfactants, namely, dodecylamine and tetradecylamine, were used as
structure directors. The reaction stoichiometries were the same for both the water-rich
and the ethanol-rich reaction systems (S°/I° = 0.25). When mesitylene (Mes) was present
as an auxiliary structure director, the Mes/8° molar ratio was 1.0 or 45.

Figures 3.1 and 3.2 provide N2 adsorption-desorption isotherms for the HMS
silicas obtained from dodecylamine and tetradecylamine as structure directors,
respectively. The adsorption properties of the mesostructures assembled from the two

surfactants under the same reaction conditions are qualitatively equivalent. As can be

63

 

 

 

 

Figure 3,]

Figure 3.1

 

2000

H
U]
c
?

 

 

 

 

 

N2 volume adsorbed (cc/g, STP)
p—L
2
?

 

 

 

‘ l
, I
500‘ :BOH-rich-t-MES J'
1 HK=21A D
0 0.2 0.4 0.6 0.8 1
PIPo

N2 adsorption-desorption isotherms for HMS silicas assembled at ambient
temperatures from dodecylamine and TEOS: Curves A and B are for
derivatives obtained from a water-rich (water : ethanol = 90 : 20 (v/v)) and
ethanol-rich solution (water : ethanol = 35 : 65 (v/v)); curves C and D are
for derivatives prepared from the same water-rich and ethanol-rich
solutions, but in the presence of mesitylene (Mes/8° = of 1.0). The HK

values are the Horvath-Kawazoe pore diameters.

64

:0 l‘\“\ IGI-IclU-ho Gin-nth! -‘

 

'6

fiflm37

 

s-=c H NH:

14:!

I°=T EOS

     

Hao-rleh + MS

 

N2 volume adsorbed (cc/g, STP)

 

 

 

o' ' '05 ' 'oii . 'ole' ' 'o‘.s' i 1
PlPo

Figure 3.2 N2 adsorption-desorption isotherms for HMS silicas prepared from
tetradecylamine and TEOS in water-rich and ethanol-rich solution. The

labeling of the isotherms is the same as in Figure 3.1.

65

 

seen fr
water-r
positio

(S°=C

have
adsor;

the m1

medic
TEM

BSOC

seen from the isotherms labeled A and B in both Figures, the structures obtained from the
water-rich and ethanol-rich solutions give stepped-shaped isotherms at P/Po < 0.4. The
positions of the steps correspond to Horvath-Kawazoe pore sizes of 28-29A
(S°=C12H25NH2) and 33-35 A (5° = C14H29NH2).

Although the mesostructures obtained from the water—rich and ethanol-rich media
have equivalent frameworks, the textural mesoporosity, as evidenced by the N2
adsorption/desorption in the region P/Po > 0.4, depends dramatically on the polarity of
the medium used for assembly. A textural pore volume even larger than the framework
pore volume can be obtained from the water-rich system, whereas the ethanol-rich
medium generates a mesostructure with little or no textural pores. As will be seen from
TEM images presented below, the high textural mesoporosity for the water-rich system is
associated with the presence of extremely small fundamental particles.

We consider next the effect of mesitylene on the framework pore structure and
textural mesoporosity of HMS silicas assembled from water-rich and ethanol-rich
solutions. The structure mediating properties of mesitylene is manifested in the
adsorption-desorption curves labeled C and D in Figures 3.1 and 3.2. The presence of
mesitylene at Mes/S° = 1 in the water-rich systems causes the adsorption step to be
significantly shifted to higher relative pressure. The shift in the step position corresponds
to a 3-5 A increase in HK pore size. In the ethanol-rich medium, however, the presence
of mesitylene results in a shift of the step position to lower relative pressures,

corresponding to a 5-7 A decrease in HK pore diameter.

66

2
E=~C>

C noncomtm o

.200

Ehm

 

 

figure

dodecy

 

15004 I

 

C12o
Horvath~szoe Plot N
0.3 .
2 ”‘5 o z
.1 g ‘ - 38 A
1000-_ 0.1 2

 

 

 

N2 volume adsorbed (cc/g, STP)

 

 

 

 

 

Orrrlrrrlrirlrrrlrrr
0 0.2 0.4 0.6 0.8 1

PlPo

Figure 3.3 N2 adsorption-desorption isotherms for a HMS silica prepared from
dodecylamine in water-rich solution in the presence of mesitylene (Mes/S = 4.5): (A)
after calcination at 650 °C with retention of framework and textural mesopores and (B)
after calcination at 1000 0C with collapse of framework pores but with retention of
textural pores. Insert: Horvath-Kawazoe pore size distribution after calcination at 650
°C.

67

 

 

depend
solvent
increas.
a mino

under e

from 3
ethanol
3.3, the
pressur

Obtains

b}. the 10

Although mesitylene can substantially expand or contract the framework pores
depending on the polarity of the reaction medium, it does not alter the key role of the
solvent in regulating the textural porosity. As will be shown below, mesitylene actually
increases the textural mesoporosity under water - rich assembly conditions, but it has only
a minor influence on the extremely low textural pore volume of HMS when assembled
under ethanol - rich conditions.

To further probe the influence of mesitylene on the framework pores assembled
from a water-rich environment, we repeated HMS assembly in 90 : 10 (v/v) water :
ethanol at a much higher Mes/8° ratio of 4.5. As shown by the N2 isotherms in Figures
3.3, the adsorption step due to framework pore filling is further shifted to higher relative
pressure as a consequence of a HK pore size (38A) that is 9A larger than the value
obtained in the absence of mesitylene. The insert to Figure 3.3 shows the half width of
the HK pore distribution to be ~10 A, a value typical of HMS materials. On the basis of
the hysteresis loop at higher relative pressure, it appears that the textural pore volume is
substantially increased by ~ 50 % from ~550 ml STP/g in the absence of mesitylene to
~850 ml STP/g by the presence of mesitylene.

In order to verify that the textural pores do indeed arise from interparticle voids
and not from a large pore component of the framework, the HMS silica assembled from
S° = C12H25NH2 at 90: 10 (v/v) H2O : ethanol and Mes/8° = 4.5 was calcined at 1000 °C
to collapse the framework pores. The N2 adsorption/desorption isotherms are given by
curve B of Figure 3.3. Note that the framework pores indeed have collapsed, as signified

by the loss of the pore filling step, but a substantial fraction (>50%) of the textural

68

 

Huma4

 

Figure 3.4

 

TEM images of calcined (650 °C) HMS molecular sieve silicas assembled
from dodecylamine : a. mesoscale fundamental particles obtained from
water—rich solution; b. the spheroid fundamental particles and macroscale

beads-on-a-string texture obtained from ethanol-rich solution.

69

 

porosity i:
structure.
Tl
Structure :
- and ethz
solvent 5)
the space
is no app;
more wor
T:
reaction
reaction
fundame]
aggregate

are Sllgg

porosity is retained. Thus, the textural porosity cannot be a consequence of framework
structure.

The TEM images shown in Figure 3.4 provide insights into the framework
structure and textural mesoporosity associated with HMS materials assembled from water
- and ethanol - rich media. Regular framework pores are readily observed, regardless of
solvent system used to prepare the products. It is quite clear that the pores originate from
the space initially occupied by uniform supramolecular assemblies of surfactant, but there
is no apparent long range order to the pore arrangement. Instead the pore packing motif is
more wormhole - like , perhaps, even sponge-like, in character.

The most important distinguishing feature between the water - and ethanol - rich
reaction products is the particle texture. As shown in Figure 3.4a, the water - rich
reaction mixture yields irregularly-shaped mesoscale fundamental particles. These
fundamental particles aggregate into larger particles. The coastline - like silhouette of the
aggregates, which persists over length scales differing by at least an order of magnitude,
are suggestive of a fractal texture. Textural mesopores comparable in size to the
mesoscale fundamental particles are clearly evident. In contrast, the ethanol-rich medium
forms spheroid to disk-like fundamental particles that are typically 100 nm or larger in
size (see Figure 3.4b). These macroscale fundamental particles are linked into larger
aggregates with a beads-on-a—string texture. The interparticle voids are much larger,
typically in the macropore range (>50 nm).

Figure 3.5 provides TEM images for a HMS derivative assembled from water-rich

solution in the presence of a relatively high concentration of mesitylene (Mes/8° = 4.5)

70

 

F‘é’uIe 3

Figure 3.5

 

TEM images of HMS silica prepared from dodecylamine in water-rich
solution in the presence of mesitylene (Mes/S = 4.5) and calcined at 1000
°C to collapse the framework pore structure; (a) Low magnification
image showing the retention of the textural mesopores and, (b) High

magnification image showing the absence of framework pores.

71

 

FigUre 3

Figure 3.6

Intensity

 

_ d=484 s~=c H NH,

123

|°=TEOS

 

rip-rich; Mods -—..4.s

 

 

l
\

Hzo-rlch; MedS°=1.0

 

tip-rich; MedS°=0

l

 

EON-rich; Meds “=0

 

d=34.0 A

EtOH-rich; MedS °=1.0

 

 

 

 

O 5 1O 15 20
26) (degrees)

X-ray powder diffraction pattern of calcined (650 °C) HMS silicas
assembled from dodecylamine amine in water-rich and ethanol-rich

solution with or without mesitylene as an auxiliary structure director.

72

 

  
   

but calcin-
obtained a
obtained at
destroyed
results desc
the origin 0
Figt
assembled
showed qt
relatively 1
near 50° 2
rich polar
PositiOns

POladty 0‘

but calcined at 1000 0C to collapse the framework mesopore structure. Image (a),
obtained at low magnification, shows the retention of the textural pores. Image (b),
obtained at higher magnification, shows that the framework mesopores indeed have been
destroyed by thermal treatment. This result, which is consistent with the N2 adsorption
results described earlier (cf, Figure 3.3), further identifies the interparticle voids as being
the origin of the textural mesopores.

Figure 3.6 provides the X-ray powder diffraction patterns for HMS derivatives
assembled from C12H25NH2 as the structure director. Structures formed from C14H29NH2
showed qualitatively equivalent diffraction features. The patterns all contain a strong,
relatively broad reflection at 2.0 - 30° 20 and a very weak broad shoulder in the region
near 50° 20. The qualitative form of the patterns is not affected by the water- or ethanol-
rich polarity of the assembly medium or by the presence of mesitylene. However, the
positions of the intense reflection and the weak broad shoulder are dependent by the
polarity of the reaction medium and by the presence of mesitylene.

Table 3.1 summarizes the basal spacings, HK pore sizes, N2 BET surface areas
(8351), total liquid pore volumes (Vt), liquid framework pore volumes (Vfr), the ratio of
textural to framework pore volumes (Vtx/Vfr), and the bulk densities for our HMS
materials. The basal spacings represented by the strong diffraction line are correlated
with the HK pore sizes, even though the framework lacks regular long-range order. The
BET surface areas are in the range 900-1464 m2/g. The incorporation of mesitylene into
the synthesis of HMS from a water - rich medium increases the HK pores size and
decreases the surface area of the mesostructure. Conversely, mesitylene decreases the

pore sizes and increases the surface areas of HMS assembled from ethanol-rich solution.

73

 

 

framework i

ln contrast.
eqUit'alent l
differences
SUbstantiall

Ethanol-rte”

3.4 Disc“S

Th
asstUnbly 1
HMS silic
, which g
ludicious
10 ("I") it
TEM ima;

are 011 a

COUVCTSElt

It is especially noteworthy from the results in Table 3.1 that the total pore volumes
are much larger for HMS derived from a water - rich medium (~1.3 - 1.7 cc/g) than an
ethanol -rich medium (~0.5 - 0.7 cc/g). Yet, the framework pore volumes are confined to
the approximate range 0.5 - 0.7 cc/g. The difference between the total and framework
pore volumes is expressed as the textural pore volume, Vtx. For HMS assembled from a
water-rich medium, the textual pore volume which can be up to 1.6 times as large as the
framework volume, depending on the amount of mesitylene used as a structure modifier.
In contrast, the textural pore volume for HMS obtained from an ethanol-rich medium is
equivalent to only a small fraction (6 - 18 %) of the framework volume. Finally, the
differences in total pore volumes is manifested in the bulk densities, which are
substantially lower for HMS derived from a water-rich medium (0.16 - 0.35 g/cc) than an

ethanol-rich medium (0.53 - 0.67 g/cc).

3.4 Discusion

The results of the present work demonstrate several important advantages of S°I°
assembly for the preparation of mesoporous metal oxide molecular sieves. In the case of
HMS silicas we have shown through N2 adsorption studies that the textural mesoporosity
, which greatly facilitates access to the framework mesopores, can be controlled by a
judicious choice of solvent (£12, Figures 3.1 and 3.2). A water-rich solvent, such as 90 :
10 (v/v) water : ethanol, promotes the formation of mesopore fundamental particles sizes.
TEM images show that the interparticle voids responsible for the textural mesoporosity
are on a length scale comparable to the fundamental particles (c_f., Figure 3.4a).

Conversely, a solvent of lower polarity, namely, 35 : 65 (v/v) water : ethanol, minimizes

74

 

.. ruin: - .Ceu 1‘0 ICNLICCI
_ c. : Edema o_om >9 conceded wmoEw o>o_m 53090:. mEI 60 0mm» noEo‘mo no“ mcoCoEmch‘moncl g 03m _
|.E

 

23:088. .IOEHONI .2 33 oo H on new or u 00 9o; wcozaom notigm 0cm cocémI “.0 20060068 of.

 

 

 

75

 

 

 

mod mod 3.0 8.0 t F F mm Em C. Townes. Bottom
no.0 0 F0 rod 0o.0 omm on flow ouomhos. ”notioum
ofio 0v; No.0 no; 000 me who o.vuom\mo_2 EEONI «Izommzu
0N0 00.0 no.0 00. F 5N0 0v «do 0. Fuomhos. £0:-O~I
no.0 3.0 no.0 om. F moor v0 «.3 ouowhos. EotémI
no.0 o0.0 $0 5.0 33 _.m 0.3 0. Fuomhofi Eocigm
no.0 o F .0 oo.0 oo.0 050w om Yam chowaos. EociOfl
ofio 0o; no.0 mo; $0 on 0.3 o.vuow\mo_2 EotémI «322320
5.0 5; mod no; 000 No 0.9 0.Fuom\mos_ EEONI
00.0 0 P... mod on; mac? 0w 5. 3 Howie—2 Eo:-OmI
83+ who 98 0M8 20 9on 30 mEauos.
x390 =>\5> :> .> hmmm xi 0 5:0on ow
Eaton. cociQm one 20:63..
5 2063mm o_om 3 09390 32.6 o>o_m 8.30206 mi: A00 88 350.3 .2 929530 33095 .P 2an

the textur
There is n
water : er
hysteresis
pores even
Figure 33
application
The
likely is detc

the reaction

 

mesostructur
Virtually win
rapid nuclea
flmdamenral
3.4a). HOWE
and mesosri
temperature
mesostructuI
in SlZe, The
imerpafllcle

The r
manly Of (h

faults in Tab

the textural porosity by forming much larger fundamental particles (9f, Figure 3.4b).
There is no doubt that the textural mesoporosity for HMS assembled in 90 : 10 (v/v)
water : ethanol arises from interparticle pores. The retention of a well expressed
hysteresis loop in the N2 adsorption/desorption isotherm show the retention of textural
pores even after the framework pores have been collapsed by calcination at 1000°C (c_f.,
Figure 3.3). Thus, the textural porosity can be tailored to a particular materials
application by simply controlling the polarity of the solvent for supramolecular assembly.

The relationship between fundamental particle size and solvent polarity most
likely is determined by the relative rates of 1° hydrolysis and supramolecular assembly in
the reaction medium. A water - rich medium leads to rapid nucleation of the
mesostructure. For assembly in 90 : 10 (v/v) water : ethanol, 3 solid product is formed
virtually within minutes of mixing the reagents at ambient temperature. This leads to
rapid nucleation and the formation of irregularly shaped fundamental particles. The
fundamental particles further aggregate into a self-similar agglomerates @11, Figures
3.4a). However, in the lower polarity 35 : 65 (v/v) water : ethanol solvent, 1" hydrolysis
and mesostructure assembly is relatively slow, requiring several hours at ambient
temperature for the onset of product formation. The slower nucleation and growth of the
mesostructure results in spheroid to disk-shaped fundamental particles 100 nm or larger
in size. These fundamental particles interpenetrate to form even larger aggregates with
interparticle voids beyond the mesopore range (c_f., Figure 3.4b).

The role of mesitylene as an auxiliary structure director is highly dependent on the
polarity of the medium in which mesostructure assembly is carried out. As shown by the

results in Table 3.1 for HMS assembly in the water-rich medium, mesitylene enlarges the

76

 

 

 

,5

v A“
I

.,o

I

=0
"-2:
r‘

:3: I}:
‘9 3

Q:
I
*

 

Flgllre 3.7

HMS assembi
central COre ol
mesitylene

head
0
cFOUp SlZe

Micellar Binding States of Mesitylene in S°I° assembly

 

Sfliatemcggunt
1|!th
? Sigicate-surfacunt
"Sb 51 P3 (‘89s ; g
o 4 /0~ v -‘
HI} H “N"? "u [no \0 ‘Sio § 03(0‘35:
QN If“ I ”at“ H ©9© H H0
H s 0 Q
Q“ Ll1111'.v f ”0"” ‘V OHQHQN N G SQ“
N N O
H H H H
u a». M n MES H© 1" “Lb- "‘O
0N W wvwvaO 0" 0N W WNO H
H H moon 31 NM T 0
H .I-I‘PNJJ dchmedium a H
a: g H20/&0H=35:6s H i ° H
H H %H ©§©g rte HHo
H
H). H H '1 H 11%“
MES InH20
nch medium
H2CNE10H=9010
0 \; / 0 \ ’9‘
j ./ 0 Si —
510 § 0 \ .o“
H\ 0H 9 OSi\
Q H "N H H a
Q “H N H 0’ T
"N H .n“\“
osr
ace ‘
H
H ® H
0N N“
H H
H
H
ONH :0
H
4% "NJ!

H
in "ii" HNQH

Silicate-mflactant
micelle expansion

Figure 3.7 Schematic representation of the structure-directing effects of mesitylene on

HMS assembly. In the water-rich media, the mesitylene “dissolves” in the hydrophobic

central core of the micelle leading to pore expansion. In the ethanol-rich environment,

mesitylene preferentially “adsorbs” to interfacial head groups, thus increasing effective

head group size and subsequently decreasing the pore size.

77

 

 

frameworl

with the e
0n the Oil
framework
The
is known I
“dissolved"
the hydroph
between the
the surfactar
states of me
Figure 3.7 fc
HMS asseml
“dissolved“ 1
Of the micell
mesOStI'uCtur
in the medir
Hl'dmgen be
with the pol;
adsorbed Stat.
the sizc of ”1‘
decrease in th.

hydrophobic l

framework pores and increases the textural mesoporosity. This latter result is consistent
with the effect of mesitylene on the pore size of MCM-41 prepared by S+I' assembly 1.
On the other hand, for assembly in the ethanol-rich medium, mesitylene reduces the
framework pore size and decreases still further the already low textural pore volume.

The role of mesitylene on the framework pore structure is intriguing. Mesitylene
is known to bind to surfactant micelles in at least two binding states, namely, in a
“dissolved” state at the hydrophobic center of the molecule, and in an “adsorbed” state at
the hydrophilic micelle / solvent interface 13'15. It is reasonable to expect the equilibrium
between these structurally distinct states to depend on the polarity of the solvent in which
the surfactant micelles are formed. In accordance with the structurally distinct binding
states of mesitylene in surfactant micelles, we propose the mechanism illustrated in
Figure 3.7 for the expansion or contraction of the framework pores by mesitylene during
HMS assembly. In water-rich solvent, in which the solubility of mesitylene is low, the
“dissolved” binding state should be favored over the adsorbed state. In this case the size
of the micelle will be increased and this will be manifested as an enlarged pore in the
mesostructure. But in the lower polarity ethanol-rich solvent, the solubility of mesitylene
in the medium will be increased and this will favor binding at the “adsorption” site.
Hydrogen bonding between the p-electrons of the aromatic and water dipoles associated
with the polar head groups at the micelle-solvent interface is believed to stabilize the
adsorbed state of mesitylene 13. Binding of mesitylene at the micelle surface increases
the size of the effective polar head group. The increase in head group size leads to a
decrease in the radius of curvature and to a reduction in micelle size in order to retain

hydrophobic interactions between the surfactant chains at the micelle center.

78

 

 

Consequer

mesostruc l

In t

that certaii

 

textural me
tetradecylar
sensitive to
the present
mesoporosit
Figures 3.1 .
W e
framework ‘1
Cltannels w;
diffraction F
the results 0
EMS mater
Obtained {Or

Consequently, a framework with a smaller pore size is formed in the assembled
mesostructure.

In our earlier studies of HMS assembly in 50:50 (v/v) water : ethanol we noted
that certain alkylamine surfactants (e.g., dodecylamine) gave derivatives with high
textural mesoporosity, whereas others afforded relatively low textural pore volumes (e.g.,
tetradecylamine) ‘3. It is now apparent that the degree of textural mesoporosity is quite
sensitive to both the solvent polarity and the nature of the S° surfactant. The results of
the present work show that dodecylamine and tetradecylamine afford high textural
mesoporosity when the solvent is highly polar, as in 90 : 10 (v/v) water : ethanol (c_f.,
Figures 3.1 and 3.2).

We have previously described as-synthesized HMS materials as neutral
framework analogs of hexagonal MCM—41 5. Evidence for the hexagonal ordering of
channels was obtained from selected area electron diffraction 4. The X-ray powder
diffraction patterns were attributed to a very small scattering domain size. It is clear from
the results of the present study, however, that the broad diffraction lines characteristic of
HMS materials are not due exclusively to a small scattering domain size. The patterns
obtained for the fine grained HMS particles are indistinguishable from those obtained for
the much larger spheroid to disk-like fundamental particles (of, Figure 3.6).
Consequently, framework disorder, in addition to small scattering domain sizes, plays an
important role in broadening the diffraction lines. On the basis of the TEM images
obtained in the present work (pf, Figures 3.4 and 3.5), we find no evidence for hexagonal
long range channel packing order. Even “disordered" hexagonal channel packing, which

has been previously documented for MCM-4l prepared by S+I' assembly, 19 appears to

79

 

 

be difficu
the reactio
-like HM.“
W0
obtained b)
like framet
electrostatic
ethylenediar
structures is
nearly unifor

0n the basis

 

Work and the
HMS materi;
quantitative d

these and Ollie

be difficult to achieve by S°I° assembly. Additional work is progress to better describe
the reaction conditions that determine the assembly of hexagonal vs. wormhole or sponge
- like HMS framework structures.

Wormhole motifs have been observed for silica and alumina mesostructures
obtained by N010 assembly, where N ° is a polyoxyethylene surfactant 6‘3. Also, sponge-
like framework structures have been described for meSOporous silicas obtained by
electrostatic S+I' assembly in the presence of a structure disrupter (e.g.,
ethylenediaminetetraacetate) 20. Distinguishing between wormhole and sponge-like pore
structures is not a straight forward matter. “Wormholes” imply channel structures with
nearly uniform diameters, whereas “sponges” imply substantially reticulated structures.
On the basis of the relatively narrow HK pore size distributions observed in the present
work and the presence of channel-like voids in the TEM images, we prefer describing
HMS materials as wormhole frameworks. Future work will require more incisive
quantitative descriptions based in part on the X-ray and neutron scattering properties of

these and other highly disordered framework structures.

80

'N‘)

3.5 References

1.

99°.“ng

10.
11.

12.

13.
14.
15.
l6.
17.

18.
19.
20.

Beck, J. S.; Vartuli, J. C.; Roth, W. J .; Leonowicz, M. E.; Kresge, C. T.; Schmitt,
K. D.; Chu, C. T-W.; Olson, D. H.; Sheppard, E. W.; McCullen, B.;

Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834.

Huo, Q.; Margolese, D. I.; Ciesla, U.; Demuth, D. G.; Feng, P.; Gier, T. E.;
Sieger, P.; Firouzi, A.; Chmelka, B. F.; Schuth, F.; Stucky, G. D. Chem Mater.
1994, 6 , 1176.

Huo, Q.; Margolese, D. I.; Ciesla, U.; Feng, P.; Gier, T. E.; Sieger, P.; Leon, R.;
Petroff, P. M.; Schtith, F.; Stucky, G. D., Nature 1994, 368, 317.

Tanev, P. T.; Chibwe, M.; Pinnavaia, T. J. Nature 1994, 368, 321.

Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865.

Bagshow, S. A.; Prouzet, E.; Pinnavaia, T. J. Science 1995, 269, 1242.

Bagshaw, S. A.; Pinnavaia, T.J. Angew. Chem. Int. Ed. Engl. 1996, 35, 1102.
Prouzet, E.; Pinnavaia, T. J. Angew. Chem. Int. Ed. Engl. 1997, 36, 516.
Antonelli, D. M.; Ying, J. Y. Angew. Chem. Int. Ed. Engl., 1996, 35, 426.
Whitehurst, D. M. US. Patent, 1992, 5,143,879.

Schmidt, R. E.; Hansen, W.; Stocker, M.; Akporiaye, D.; Ellestad, O. H. J. Am.
Chem. Soc. 1995, 117, 4049.

Zhang, W.; Froba, M.; Wang, J .; Tanev, P.; Wong, J .; Pinnavaia, T. J. J. Am.
Chem. Soc. 1996, 118, 9164.

Mukerjee, P.; Cardinal, J. R. J. Phys. Chem. 1978, 82, 1620.

Eriksson, J .E.; Gillberg, G. Acta Chem. Scand, 1966, 20, 2019.

Gordon, J. E.; Robertson, J. C.; Thome, R. L. J. Phys. Chem, 1970, 74, 957.
Horvath, G.; Kawazoe, K. J. J. Chem. Eng. Jpn. 1983, I6, 470.

Spence, J .C.H., "Experimental High-Resolution Electron Microscopy ", 1988,
P264, Oxford University Press, New York.

Tanev, P.T.; Pinnavaia, T.J. Chem. Mater. 1996, 8, 2068.

Chen, C. Y.; Xiao, S. Q.; Davis, M. E. Microporus Mater. 1995, 4, l.

Ryoo, R.; Kim, J. M.; Ko, C. H.; Shin, C. H. J. Phy. Chem. 1996, 100, 17718.

81

Chapter 4

In-situ Alumination of Mesoporous HMS Molecular Sieves Silica for
Use as Alkylation Catalysts

Abstract

The in-situ alumination of mesoporous HMS molecular sieves assembled via an
electrically neutral assembly mechanism using primary amine surfactant as the structure
director, tetraethylorthosilicate (TEOS) as the silicon source and NaAlO2 or Al(sec-OBu)3
as the alumination agent was conducted at ambient temperature. The majority of Al-sites
(i.e. >90%) were found in a tetrahedral environment by 27Al-NMR. The textural porosity
was not alternated by alumination of structurally well formed HMS for Al-loadings up to
8 mol%. Al—HMS with 4 different amounts of textural porosity (O—tex, 0.5-tex, l-tex, 1.5-
tex) relative to framework mesoporosity were purposely prepared through control of
solvent polarity and the time of introducing the aluminum precursor following the
addition of TEOS. The results indicate that the alurrrination of HMS at ambient
temperature, using either NaAlO2 or Al(sec-OBu)3, are as effective as the hydrothermal
post synthesis alumination of other mesoporous molecular sieves.1 It is not necessary to
remove the surfactant prior to effectively incorporate aluminum into framework of
mesostructural HMS at ambient temperature. It is also not necessary to use surfactant-free
materials to convert sodium-forms of Al-HMS and Al-MCM-41 into their proton forms
by ammonium-exchange. The total acidity of Al-HMS catalysts measured by TGA
analysis of cyclohexylamine desorption were reasonably equivalent to each other when
the Al-loading was ~2 mol%. Al-HMS with higher textural porosity, regardless of the

aluminum precursor used, showed significantly higher conversion (up to 65%) of the

82

bulky 2,4-di-tert-butylphenol (2.4-DTBP) reacting with Cinnarnyl alcohol to a bulky

flavan (selectivity up to 87%) under very mild reaction conditions (i.e. 60 °C, 6 h).

4.1 Introduction
Mesoporous HMS, as a product of electrically neutral assembly, is distinguished

from other mesoporous molecular sieves such as MCM-4l by exhibiting significant
textural mesoporosity. The complementary textural porosity of HMS, in addition to
framework mesoporosity, was observed in the first HMS sample. 2 Improving and

tailoring textural porosity was reported more recently. 3

The textural porosity of HMS
facilitates diffusion limited catalytic reactions.4 For example, HMS catalysts have better
performance than MCM-4l analogous for both 2,6-di-tert-butylphenol (2,6-DTBP)
peroxidation 5 and NO reduction by NH3.6 No significant physical and chemical property
differences except textural porosity have been found between these two analogs. The
better performance for HMS derivatives can not be attributed to different active sites for
those mesoporous molecular sieves. For cumene cracking, substantially higher activity of
Al-HMS than MGM-41 analogous have been reported.7'8 Higher acidity and smaller
domain size of HMS with apparently shorter diffusion channels might account for the
better performance of Al-HMS in cumene cracking.

Another very important feature of Al-HMS is that the negative charges generated
by substitution of Al into the silica framework can be automatically balanced by the
hydrolysis of the primary amine to its protonated form. Thus, the H-form Al-HMS could

be easily obtained by calcination without tedious ion exchange required for

transformation of sodium-form Al-MCM-4l to H-form. And also, the majority of

83

surfactant in Al-HMS (>90%) can be easily removed by simple ethanol extraction.
Dealumination upon removal of surfactant by calcination can be purposely avoided
through facile ethanol extraction for Al-HMS, but this is more difficult to accomplish for
Al-MCM-41 .9" '

Aluminum substituted mesoporous molecular sieves with mild acidity '0 and large
pores have been recognized as good catalysts for fine chemical syntheses where large
molecular substrates are usually used. ‘2 Al-HMS with high meso-textural pores is
apparently better than other Al-substituted mesoporous molecular sieves in terms of
accessibility to framework active sites. Therefore, it is very important for catalysis to have
HMS derivatives that are rich in textural porosity and have meso-scaled fundamental
particles that limit the pore channel lengths to being comparable to the fundamental
particle size itself. One of the objectives of this report is to prepare Al-HMS derivatives
with large textural porosity, as well as a small (<150 A) fundamental particle size. The
importance of these physical properties to catalysis were verified in the alkylation of 2,4-
di-tert-butylphenol (2,4-DTBP) with cinnamyl alcohol to a flavan. This particular liquid
phase reaction was first reported by Corma and his coworkers.l3 It was reported that this
reaction was not possible to take place inside the microporous channels of H-Y because
the molecular size of the substrates prevented them from accessing acidic centers in the
smaller channels. Even though the conversion of 2,4-DTBP catalyzed by Al-MCM—41
was high, a very significant amount of it was dealkylated to 4-tert-butylphenol (4-TBP,
~25% yield) rather than to flavan.13 In our study, Al-HMS with meso-scaled fundamental

particles and high textural porosity has shown remarkable selectivity (i.e. 86%) to flavan

84

at a significantly higher conversion of 2,4-DTBP (i.e. 65%) without detectable amounts
of 4-TBP in the reaction products.

Another objective is to prepare Al-HMS with cost-effective aluminum sources
such as NaAlO2, because all of the Al-HMS derivatives reported so far were synthesized

using aluminum alkoxide, which is expensive and sensitive to moisture.

4.2 Experimental

Our approach to the synthesis of Al-HMS with high. textural porosity and a small
fundamental particle size using either aluminum alkoxide or NaAlO2 was based on
assembly of HMS framework in a relatively high polarity medium 3 and a follow-up
alumination step. A water medium with 10 vol% ethanol as co-solvent in the presence of
mesitylene (MES) as a textural porosity promoter was used. The synthesis was carried
out using dodecylamine as the neutral surfactant and tetraethyl orthosilicate (TEOS) as
the neutral inorganic precursor. The aluminum precursor was introduced into the
synthesis mixture in a follow—up manner. Three different Al-HMS derivatives were
prepared by control the time of introducing the aluminum precursor. First, add aluminum
precursor solution (Al(sec-OBu)3 in sec-butanol or N aAlO2 in water) immediately after
addition of TEOS to surfactant solution; Second, add aluminum precursor solution 20
minutes after addition of TEOS; Third, add aluminum precursor solution 20 h after
addition of TEOS. The reaction mixture was aged under stirring at ambient temperature
for 20 h after addition of aluminum precursor. The molar composition of our synthesis
mixture was:

0.02~0.08 Al : 1.0 Si : 0.25 Surf. : 1.12 MES : 5.0 EtOH : 130 H20.

85

For comparison, a zero-textural porosity Al-HMS was synthesized using
aluminum alkoxide as the alumination agent in a more non-polar medium comprised of
EtOH/H2O with a volume ration of 50/50. The molar composition of this synthesis was

0.02 Al : 1.0 Si : 0.25 Surf. : 13 EtOH : 42 H2O

The product for each synthesis was filtered and dried in air.

The samples prepared using aluminum alkoxide as the aluminum source were
calcined at 650 0C for 4 h without prior NHf-ion-exchange. The as-made samples
prepared using NaAlO2 as the aluminum source were ion—exchanged at ambient
temperature for 4 h using 0.1 M NH4NO3 solution with a solution to solid ratio of 50
ml/g. The obtained product was dried in air and calcined at 650 °C for 4 h.

The Al-MCM-4l was prepared using 27wt% sodium silicate solution as the
silicon source, and cetyltrimethyl ammonium bromide as the structure director. The pH
was initially adjusted to 10.5 using sulfuric acid and adjusted again to 10 after the
reaction mixture was heated at 100 0C for 1 day. Subsequently, the reaction mixture was
heated at 100 °C for another day. Then, A12(N03)3 aqueous solution was added at ambient
temperature under stirring to the reaction mixture, which was statically aged at 100 0C for
another 2 days. Aluminum nitrate was chosen as the aluminum precursor over other
aluminum sources such as aluminum sulfate and sodium aluminate. The purpose for this
choice is to avoid the undesired acidic sites generated by residual anions (i.e. 8042') in the
case of aluminum sulfate and the structural alternation caused by pH increase in the case
of sodium aluminate. The molar composition of the reactant in this synthesis was:

0.02 Al : 1.0 SiO2 : 0.25 Surf. : 0.39 Na2O : 0.24 H2SO4 : 83 H20

86

The product was filtered and washed thoroughly with water. The filter cake was dispersed
in 0.1 M NH4NO3 solution for 4 h at ambient temperature to have it ion-exchanged. Then
it was filtered and dried in air before calcination at 540 °C for 6 h.

The powder X-ray diffraction pattern of the samples was recorded on a Rigaku
Rotaflex Diffractometer using Cu K, radiation 0c: 1.542 A). N2 adsorption and
desorption isotherms were measured on an ASAP 2000 sorptometer at -196 0C. The
sample was evacuated under 10'5 torr at 150 °C over night before the measurement. 27A1
NMR spectra were recorded on a Varian VXR-4OOS spectrometer with a 7-mm zirconia
rotor, and a spinning frequency of 4 kHz. TEM images were taken on a JEOL 100CX
with a Ce85 gun that was operated at an acceleration voltage of 120 kV. The specimen
was loaded onto a holey carbon film that was supported on a copper grid by dipping the
grid into a solid sample suspension in ethanol. The total acidity of a catalyst was
measured by means of TGA analysis of cyclohexylamine desorption in the temperature
region 240 to 420 °C. Cyclohexalymine was pre-adsorbed on the catalyst and the sample
was then purged with N2 at 200 °C for 1 h before temperature programmed desorption
was started.

Alkylation of 2,4-di-tert-butylphenol (2,4—DTBP) with cinnamyl alcohol was
carried out in a 100 ml round bottom flask. A 1.0mmol quantity of 2,4-DTBP, 1.0mmol
cinnamyl alcohol and 50 ml of isooctane as the solvent were used. A 500mg quantity of
catalyst was added when the reaction mixture was heated to 60 °C. The reaction was kept
at 60 °C under magnetic stirring for 6 h. After reaction, the catalyst was filtered and
washed with dichloromethane. The dichloromethane was separated from the filtrate by

distillation at ~30 °C under reduced pressure so that heat-induced side reactions could be

87

prevented. The reaction products were qualitatively confirmed using a GC-MS. They
were quantitatively analyzed by an internal standard method using a GC with a FID
detector and a 15 m long SPB-l type capillary column. In order to obtain the response
factor of the flavan relative to internal standard that was 2,4-di-tert-butylbenzene, pure
flavan (i.e. ~95%) was purified from a large batch reaction using a TLC approach. The
conversion was calculated by dividing the amount of 2,4-DTBP consumed by the amount
initially present. The selectivity toward flavan was calculated by dividing the amount of
falvan formed by the amount of 2,4-DTBP consumed. The yield of flavan was obtained
by dividing the amount of flavan formed by the amount of 2,4-DTBP added. The
conversion of cinnamyl alcohol was virtually complete in most reactions because

'4 such as

cinnamyl alcohol alone could transform into many undesirable products
cinnamyl aldehyde, 1,5-di-phenyl pentadiene, and polymeric aldehyde. The latter

products formed a tar on the catalyst, which required calcination in air to regenerate the

catalyst.

4.3 Results and Discussion
4.3.1 Catalyst Synthesis

In order to obtain acidic sites for catalysis, aluminum is usually incorporated into
the framework of microporous zeolite or mesoporous silica molecular sieves under
hydrothermal conditions in situ.15'16 Alumination of mesoporous silica MCM-41 by
sodium aluminate in a previous report showed that the aluminum species were
predominantly incorporated into the framework near the pore openings, if the alurrrination

was carried out in the presence of surfactant molecules inside the mesoporous channels.17

88

This is a particularly significant problem for well-ordered large particle sized MCM-41.
The interpretation for this result might be as following. First, the surfactant molecules in
the uni-dimensional channels of MCM—41 limit the diffusional migration of aluminum
species to deep inside the channels. Second, the interfacial electrostatic charge matching
‘8 between framework and surfactant micelles generated a large electrostatic repulsion to
negatively charged aluminum species, therefore, the aluminum was predominantly
incorporated at the opening of the pores. In contrast, the framework of HMS assembled
through an electrically neutral mechanism was easily aluminated at ambient temperature
even in the presence of surfactant micelles in the pores. A reasonable interpretation for
this might be that the diffusion resistance was minimized for aluminum species to migrate
into the meso-scaled pores with relatively short lengths in meso-scaled fundamental
particles to become incorporated into the neutral framework.

For sodium forms of mesoporous molecular sieves, previous reports very much
rely on ammonium exchange of the (calcined) surfactant free materials and further
calcination to generate the H-forrn.'9'20 However, we found that it is not necessary to
remove the surfactant before ammonium exchange to obtain effective catalysts. Sodium
form of Al-HMS and Al-MCM-41 were transformed into their ammonium forms easily
through NHi-exchange at ambient temperature in the presence of surfactant molecules

without noticeable NH4+ diffusion limitations.

4.3.2 Catalyst Characterization

Figure 4.1 and 4.2 show N2 adsorption-desorption isotherms of Al-HMS prepared

using Al(sec-OBu)3 and NaAlO2, respectively. The textural porosity increases with the

89

 

 

2000

~ 2%AI-HMS

L Time of AI(OBu)3

h Addition (h)
A 20
n. 1500 r V .
:7, . H-K (A) "‘ '
55 t 1.5
>
g . 33A
3 1000 " ”3
m
2 10
0 33A
5
O
>
ZN

LlllllklllLlllllL

 

 

 

 

O 0.2 0.4 0.6 0.8 1
Relative Pressure PlPo

Figure 4.1. N2 adsorption and desorption isotherms for calcined (650 0C) Al-HMS
molecular sieves with different textural porosity. Each sample was prepared at ambient
temperature using dodecylamine as the structure director, TEOS as the silicon source, and
using Al(sec-OBu)3 as the alumination agent, but the alumination agent was added to the
reaction mixture 0, 0.33, and 20 hrs after mixing the TEOS and the surfactant. The
aluminum content in each case is 2.0 mole%. The Horvath-Kawazoe (H-K) framework
pore size, and the ratio of textural to framework mesoporosity (Vthfr) is provided for

each sample.

90

 

I

2%AI-HM S

 

2000 F
' Time of NaAlO2

& ' Addition (h)
«'72
0'5
2 1500
i
m
3
w 1000
E
2
O
>
ZN

 

e e e e :rgoeerezeee

soc t" W

0.5

 

I; l 1 l l l l l l l l l l l l l I

O 0.2 0.4 0.6 0.8 1
Relative Pressure PlPo

Figure 4.2. N2 adsorption-desorption isotherms for calcined (650 °C) Al-HMS
molecular seives prepared as described in the caption to Figure 4.1, except

that the alumination agent was sodium aluminate.

91

delay-time for introduction of the aluminum precursor after addition of TEOS. It is
obvious to conclude from Figure 4.1 and 4.2 that the textural pore volume depends very
closely on the time of the aluminum precursor addition to the synthesis medium. The
highest textural pore volume relative to framework pore volume, was achieved from an
in-situ alumination of siliceous HMS with the surfactant micelle still in the mesopore
channels. Al-HMS with high textural porosity as high as ~l.5 times of framework
mesoporosity (denoted 1.5-tex-Al-HMS) was obtained when the alumination agent was
added '20 h after the hydrolysis of the TEOS in the presence of the surfactant at ambient
temperature. Due to the fact that most of the framework skeleton of HMS is formed
within first 20 minutes after addition of TEOS to the surfactant solution, structural
alternation would be minimized if aluminum precursor was added 20 minutes after
addition of TEOS. Actually, a Al-HMS with the textural porosity equivalent to the
framework mesoporosity, (denoted 1.0-tex-Al-HMS) was obtained at this alumination
time. Interestingly, when the aluminum precursor was added immediately after TEOS in
an aluminum-silicon co-incorporation manner, the Al-HMS exhibits much lower textural
porosity with smaller meso-scaled textural pores indicated by the uptake of N2 at
relatively smaller P/Po (i.e. 0.7). The ratio of textural to framework pore volume (VuNfr)
was ~0.5, hence, denoted 0.5-tex-Al-HMS. In addition, the step in the isotherm indicative
of N2 filling of framework mesopores at P/Po ~0.3 was smeared, and the framework pore
volume and pore size was decreased. This suggests that aluminum incorporation at the
initial stage of the assembly deteriorate the formation of framework, especially for neutral
S°I° assembly. This result is in agreement with Al-MCM—41 which exhibits a much

broader (100) reflection, and weakened higher order reflections, in comparison with

92

Figure 4.3.

 

 

 

 

 

 

 

 

7001
600:
p: :
(I) .
5 500:
3 .
3 I
E 400:
h .
o ..
m .-
8 3001
a, .
E .
a .
3 200‘.
> - s
N V
Z t
1 00-. g \“\\1)).V.V.vxv;v_v_v_v_
- 0 20 40 60 30 100
‘ (A)
o I I I l I I T r r I U l I I I r I I j
0 0.2 0.4 0.6 0.8 1
Relative Pressure PIPo
N2 adsorption-desorption isotherms for calcined (650 °C) 2% Al-HMS and

calcined (540 °C) Al-MCM-41 with zero-textural porosity. The inset
shows Harvoth-Kawazoe pore size distributions for these two samples.
This Al-HMS was prepared using alumnum sec-butoxide as the
alumination agent, which was added to the reaction mixture 20 minutes

after mixing the TEOS and the surfactant solution.

93

siliceous MCM-4l.2' Our observations on the effect of co-assembly of aluminum with
silicon in HMS is in agreement with the results of Al-HMS prepared by Jones and his co—
worker.7 They mixed aluminum alkoxide and TEOS before adding the mixture to the
surfactant solution. The total pore volume of their mesostructure was significantly
reduced from 1.68 mllg for the pure silica analogous to 0.65 ml/g by incorporation of just
~2.5 mol% Al into the framework. Interestingly, the Al-HMS prepared by Jones et al did
exhibit higher acidity than Al-MCM-41 analogous although textural porosity was not
observed. We verified their result by preparing a Al-I-IMS in a relatively less polar
solvent medium with a EtOH/H2O ratio of 50/50 (see below for discussion about acidity).
As showed in Figure 4.3, this Al-HMS exhibits no hysteresis loop for textural pores in the
isotherm and so does the Al-MCM-41. Actually, the N2 isotherm for Al-MCM-41
exhibits very sharp framework mesopore filling step at relative pressure of 0.3.
Consequently, the pore size distribution is relatively sharp and narrow also. In contrast, 0-
tex-HMS exhibits a somewhat wider pore size distribution. The differences in framework
structure, hexagonal vs. wormhole, are reflected in the widths of the pore size
distribution.

Figure 4.4 and 4.5 show the powder X-ray diffraction patterns of Al-HMS
prepared by addition of Al(sec-Obu)3 and NaAlO2 respectively to the TEOS-
dodecylamine hydrolysis mixture after various reaction times. Three Al-HMS samples
with different textual porosity prepared using aluminum alkoxide as the aluminum
precursor show nearly identical d-spacing at 4.2 nm (Fig. 4.4). Relatively higher d-
spacing (52 A) along with larger pore size (41 A) was obtained for Al-HMS prepared

using NaAlO2 (Fig. 4.5). This may indicate that sodium cations facilitates the pore size

94

 

 

 

 

 

 

 

 

 

 

 

41.7 A 2%AI‘HMS
{l s°=c H NH,I°=TEOS,
12 25 2
Calcined at 650 °C
E /
in
c
.9
5
Time of AI(OBu)3
addition (h)
24
0.33
0
..,....,s...,....
0 5 10 15 20

29 (degrees)

Figure 4.4. XRD patterns of calcined 2%Al-HMS molecular sieves prepared as
described in the caption to Figure 4.1.

95

 

 

 

 

 

 

 

51.75.
f 2%Al-HMS
s°=c H NH,I°=TEOS
12 25 2
Calcined at 650 °C
.5 ..
c on
0 o
'E <-
Time of NaAIO2
addition (h)
24
0
WW],
0 5 10 15 20

20 (degrees)

Figure 4.5. XRD patterns of 2%Al-HMS prepared as described in the caption to
Figure 4.1, except that NaAlO2 was the Al-source.

96

 

 

 

 

 

 

 

 

510‘
' 2%Al-Loading
41d~
g 310 -
m -i
C
o
«H
E
210 . Al-MCM-41(O-tex)
.1
11d-
' AI-HMS (o-tex)
"'l""l""l
O 5 10 15 20

29 (degrees)

Figure 4.6. XRD patterns of calcined 2% Al-HMS and Al-MCM41 with zero—textural
porosity as described in the caption to Figure 4.3.

97

enlargement by exchange into the interface of surfactant and framework wall as proposed

22 It worth pointing out that this kind of pore

for larger pore MCM-41 synthesis.
enlargement process takes place at elevated temperature for MCM-4l, but it occurs for
HMS at ambient temperature. Regardless of the type of aluminum precursors, the sample
prepared by the addition of an aluminum precursor immediately after addition of TEOS
caused the intense reflection to be broadened. This loss of wormhole framework structure
is in agreement with the N2 adsorption-desorption isotherms (cf. Fig. 4.1 and 4.2).

The XRD patterns for 0-tex-Al-HMS and Al-MCM-41 are showed in Figure 4.6.
O-tex-Al-HMS exhibits equivalent reflection to that of pristine HMS.3 Al-MCM-4l
exhibits well resolved hexagonal X-ray reflections. Apparently, the structural order of
both HMS and MCM-4l were fully retained since both samples were prepared through a
follow-up in-situ alumination. Even though they were assembled from a different
mechanism, the total pore volumes for both of them are virtually the same (c.f. Fig. 4.3).

As stated earlier, it is very important to have mesoporous molecular sieves with
small fundamental particle size so that the porous channels can be shortened and
diffusion resistance minimized. It was found that meso-scaled fundamental particles are
always associated with high textural porosity in pristine HMS. To confirm this for Al-
HMS, we examined three derivatives prepared using either Al-alkoxide or NaAlO2 by
TEM. As exemplified in Figure 4.7A, fractal-like particles with a worm-hole-like
structural motif prevail for both 1.0-tex-Al-HMS and 1.5-tex-Al-HMS. Irregularly shaped
inter-particle voids (~150 to ~500 A) are present between aggregates of the fundamental

meso-scaled particles ~100 to ~500 nm in size. These voids are the textural pores

responsible for the hysterisis loop observed in the high textural pore materials at partial

98

 

Figure 4.7. TEM images showing meso-scaled textural pores as well as worm-hole
like framework mesopores for calcined 1.5-tex-Al-HMS and 0—tex-Al-
HMS.

99

Figure 4.8.

  
 
  
  

1.5-tex-AI—H MS(650 °C)

Al-MCM-41 (540 °C)

1.5-tex-H MS (NaAlO 2)

O-tex-AI-H MS (AI(OBu) 3)

 

 

 

 

1 .5-tex-Al-H MS (AI(OBu) 3)

_‘

 

 

 

 

 

 

 

 

I I I I I I I l V ' I ' ' I I '7 I I I

400 200 0 -200 -400 ppm

27Al-NMR spectra showing predominantly tetrahedral Al-sites (63 ppm)
and a small amount (~10%) of octahedral Al-sites ( 18 ppm) in as-made
Al-HMS prepared using Al(sec-OBu)3 as the alumination agent. Calcined
forms of Al-HMS, as well as calcined Al-MCM-41, show 70% tetrahedral
and 30% octahedral Al-sites.

100

pressures above 0.9. Unlike the textural porosity-rich sample, Al-HMS with zero-textural
porosity shows sub-micro-sized beads on string type of spheroidal particles (Fig. 4.7B).
Although the particle size for a 0.5-tex-Al-HMS is about 2 times larger (not shown) than
1.0-tex-Al-HMS, but they are still about 5 times smaller than that of O-tex-Al-HMS.

27Al-NMR spectra are shown in Figure 4.8. Obviously, the vast majority of
aluminum species in as-made Al-HMS are in a tetrahedral coordination
environment. These aluminum sites possess a chemical shift at ~63 ppm. The NMR
spectra differences between high and low textural porosity samples are practically
indistinguishable. Al-HMS prepared using Al-alkoxide shows a slightly higher content of
octahedral aluminum species (c.f. the shoulders at around 16 ppm) than the one prepared
using NaAlOz, possibly because of aluminum oxide clusters generated by fast hydrolysis
of aluminum alkoxide. Calcined Al-MCM-41 and clacined Al-HMS exhibited about 30%
octahedral (indicated by 18 ppm shoulders) and 70% tetrahedral aluminum sites that
possesses a chemical shift position at 63 ppm also. Nevertheless, the completion of the
alumination of HMS at ambient temperature suggests that the surfactant molecules inside
the framework pores be disregarded by aluminum precursors. As a consequence of a
neutral interfacial interaction (H-bonding) between the surfactant and the framework,
anionic aluminum precursor species is allowed to migrate easily through the neutral
interfacial gaps in the absence of charge repulsion, and to become incorporated into the
framework.

The amount of acid of each catalyst was measured through TGA analysis of the
amount of cyclohexalamine (CHA) desorbed in the temperature range 240 to 420 °C. As

shown in Table 4.1, among all five Al-HMS samples regardless of textural porosity and

101

aluminum precursor, the average amount of acidity was 0.34 mmol CHA/g. This accounts

up to 100% of the loading of Al, if it is assumed that each Al generates one acid site. This

indicates that both tetrahedral and octahedral Al-sites in Al-HMS contribute to the total

acidity, which suggest that octahedral Al-sites in Al-HMS be isolated, possibly grafted

Al-sites rather than aluminum oxide clusters. In agreement with the results reported by

Jones et al, the acidity observed for our 2 mol% Al-HMS is about 3 times higher than that

of 2 mol% Al-MCM-41 (i.e. 0.10 CHA mmol/g). The difference in the acidity strongly

Table 4.1. Physical and chemical properties for Al-substituted mesoporous molecular

 

 

 

sieves
Aluminum Saar Pore Volumesa (egg) H-K Pore Acidityb
Catalyst Source (m2/ g) V, Vf, Vtcx Size (nm) (mmol lg)

0-tex-Al-HMS Al(OBu)3 1234 0.86 0 0 2.9 0.37
0.5-tex-Al-HMS Al(OBu)3 968 1.08 0.67 0.41 2.8 0.36
1.0-tex-Al-HMS Al(OBu)3 976 1.38 0.73 0.65 3.3 0.38
1.5-tex-Al-HMS Al(OBu)3 1020 1.77 0.80 0.97 3.3 0.30
1.5-tex-Al-HMS NaAlOz 978 1.95 0.80 1.15 4.1 0.30
Al-MCM-4l A1(NO3)3 1070 0.87 0 0 3.1 0.10

 

 

 

3. The total pore volume (Vt), including the textural pore volume (van), was measured on
the isotherm at P/Po = 0.99. The framework pore volume (Vfr) was measured at P/Po =
0.5. b. The total acidity for these catalysts were obtained by means of TGA through
measurement of cyclohexylamine deorption between 230 to 420 °C.

102

suggest that about 2/3 of Al-sites for Al-MCM-41 be inaccessible by cyclohexylamine.
As long as a 27Al-NMR spectrum of Al-MCM-41 does not exhibit a significantly higher
number of octahedral Al-sites, the difference in acidity between Al-MCM-4l and Al-
HMS with the same Al-loading can not be attributed to large amount of extra-framework
aluminum sites. It appears more likely that the lower total acidity in the case of Al-MCM-
41 is due to lower accessibility of Al-sites buried in the ~12 A thick walls. The possibility
of Na+ blocking the acid sites was excluded by the fact that a low Na/Al ratio of 0.1 was
observed by ICP elementary analysis of this Al-MCM-41. The acidity for all Al-HMS
samples are not considerably different from each other. Therefore, the performance
differences in acid catalyzed reaction for Al-HMS molecular sieves can not be attributed
to the intrinsic difference of acid sites (see below).

In summary, Al-HMS with high textural porosity, small fundamental particle size,
and wormhole framework porous structure was prepared by carefully controlling the
addition of aluminum precursor during the syntheses. In order to minimize the framework
deterioration caused by Al incorporation and to maintain the high textural porosity, the
aluminum precursor was introduced into the reaction mixture after the formation of the
silica framework (i.e., after 20 h of TEOS hydrolysis in the presence of the primary amine
structure directors). It worth to pointing out that both aluminate anion and
Al(OH)x(OBu)3-,, type of hydrolysis molecules (i.e., possible aluminum species involved
in the alumination process) can diffuse into the surfactant-filled framework mesopores

and became incorporated into the silica framework at ambient temperature.

103

4.3.3 Catalytic Results

As discussed before, four type of Al-HMS molecular sieves with different amount
of textural porosity were prepared through control of the reaction conditions. In order to
compare the catalytic activity for the Al-substituted mesoporous molecular sieves, the
acid catalyzed alkylation of 2,4-di-tert-butylphenol was chosen because bulky reagents

are completely inaccessible to the acid centers inside the micropores of H-Y zeolite 13.

 

The
OH
\ ..... 0
Acid catal st 0
0 * V > O
Cin-OH
2,4-DTBP

importance of diffusion might be manifested in this large molecule reaction. In the first
report of this reaction using mesoporous MCM-41 as the catalyst, the reaction was carried
out at 90 °C for 24 h.13 We found that it is not necessary to carry out this reaction for that
long, because one of the two reactants, cinnamyl alcohol, is nearly 100% consumed after
6 h at even 60 °C. Most of the by—products from this reaction were transformed from
cinnamyl alcohol alone in the presence of catalyst. Three products were identified from
the conversion of 2,4-di-tert-bytylphenol by GC-MS analysis. These three products
account for up to ~90% of consumed 2,4-di-tert-butylphenol. The major product is flavan.

Two by-products were 4-tert-butylphenol and 4,6-di-tert-butyl benzofuran. Since these

104

two by-products only account for less than 5% of consumed 2,4-di-tert—butylphenol, they

are not listed in Table 4.1. Interestingly, 4—tert-butylphenol, the product of dealkylation

Table 4.2. Alkylation of 2,4-di-tert-butylphenol with cinnamyl alcohol catalyzed by

Al-substituted mesoporous molecular sieves3

 

 

Aluminum Conv. (%) Selectivity for Yield of
Catalyst Source 2,4-DTBP Flavan (%) Flavan (%)
0-tex-Al-HMS Al(OBu)3 50.8 71.3 36.2
0.5-tex-Al-HMS Al(OBu)3 59.6 77.9 46.4
1.0-tex-Al-HMS A1(0Bu)3 60.9 77.1 46.9
1.5-tex-Al-HMS A1(0Bu)3 61.9 83.2 51.5
1.5-tex-Al-HMSb NaAlOz 65.3 86.8 56.6
Al-MCM-4lb A1(N03)3 47.2 71.8 33.9

 

 

 

3. Each catalyst contains 2 mol% Al; the reaction was carried out using 10.0 mmol of each
reactant, 500 mg of catalyst , in 50 ml of isooctane at 60 °C for 6 h.
b. These catalysts were ion-exchanged with 0.1 M NH4N03 before calcination.

of 2,4-DTBP was barely detectable for most Al-HMS catalysts. However, this was a
much more apparent by-product for conventional Al-MCM-4l as the catalyst. As shown
in Table 4.2, the conversion of 2,4-DTBP and the selectivity as well as the yield of the
desired product flavan all increase with the textural porosity of Al-HMS. The selectivity
to fiavan was un—precedently high (i.e. ~87%).

It is particularly noteworthy that 0-tex-Al-HMS and Al-MCM-41 exhibit very

similar catalytic activity, but the activity is much lower than their HMS analogs with high

105

textural porosity. These results signify the importance of textural porosity in facilitating
the access of substrate molecules to acid sites. Our MCM-41 sample exhibited almost the
same flavan yield as reported by Corma et al for Al-MCM-41 with approximately the
same Al-loading. Corma’s Al-MCM-41 catalyst was NH4+-exchanged after removal of
the surfactant by calcination whereas our Al-MCM-41 was NH“-exchanged prior to
removal of the surfactant. An important conclusion from the similarity in catalytic
activity would be that our approach for ammonium exchange prior to removal surfactant
is as efficient as normal ammonium exchange for surfactant-free MCM-4l molecular
sieves.

The physical and chemical properties for all the catalysts, as shown in Table 4.1,
are approximately similar to each other except for the BET surface area and pore volume.
Even though both of the BET surface area and the framework mesopore volume for 0.5-
tex-Al-HMS are smaller than that of 0-tex-Al-HMS, the catalytic conversion and the
yield of flavan for the former catalyst are higher. Therefore, the activities of these Al-
HMS catalysts are not directly related to BET surface area and mes0pore volume.
Indistinguishable differences in Al-siting indicated in 27Al-NMR (Fig. 4.8) agrees very
well with the amount of total acidity for all Al-HMS catalysts measurement by thermal
gravity analysis (T GA) of cyclohexlamine desorption. As shown in Table 4.1, the
average total acidity for all Al-HMS with 2% Al-loading was about 0.34 mmng, which
was very close to the theoretical acidity of 0.33 mmng. Therefore, it is impossible to
attribute the differences in catalytic activity to intrinsic differences in Al-siting among Al-
HMS catalysts. However, Al-HMS generally exhibit three times higher total acidity than

analogous MCM-4l, which is in agreement with the results reported. 7 The acidity

106

difference between Al-MCM-41 and Al-HMS with the same Al-loading might be
indicative of the fundamental differences in their assembly mechanisms rather than the
residual Na“, because the amount of Na+ left in this Al-MCM-41 was only 10 mol% of
the total Al.

The results in Table 4.2 unambiguously suggest that textural porosity indeed
improve the catalytic conversion of bulky substrates in liquid reactions. In addition to
textural pores, the Al-HMS samples with high textural porosity also exhibit very thin (i.e
< 100 nm, this is the maximum specimen thickness for electron beam accelerated at 120
kV to go through easily to obtain good quality TEM pictures) and small fundamental
particles (~ 150 nm). Consequently, this feature of Al-HMS would provide limited pore
channel lengths that should be comparable to the fundamental particle size. Thus the
diffusion resistance of bulky molecules would be minimized and this enhances the
accessibility of catalytic active centers in the framework channels and favors the

maximum conversions.

4.4 Conclusions

Al-substituted HMS molecular sieves with different amounts of textural porosity
relative to the framework mesoporosity were prepared through control of solvent polarity
and the time of introduction of the aluminum precursor to the developing wormhole
frameworks. Alumination of HMS molecular sieves assembled in the presence of neutral
amine surfactant at ambient temperature using NaAlOz or Al(sec-Obu)3 was as effective
as literature-reported hydrothermal aluminum incorporation processes for MCM-41

molecular sieves. The purpose for in-situ follow-up alumination is to avoid alternation of

107

structural features (i.e. high textural porosity etc.) desirable for condensed phase catalysis.
Ammonium exchange of sodium-form Al-HMS and Al-MCM-4l at ambient temperature
in the presence of surfactant in porous channels was also effective enough to achieve
sufficient active H-form mesoporous catalysts. The complementary meso-scaled textural
pores as well as meso-scaled fundamental particles of mesoporous molecular sieves were
very important to improve the catalytic alkylation of bulky 2,4-DTBP with cinnamyl
alcohol to flavan under very mild conditions (60 °C, 6h). The conversion of 2,4-DTBP
and the yield of flavan both were dramatically improved in comparison with Al-HMS and

Al-MCM-41 catalysts without textural pores.

108

4.5 References:

1.

10.

ll.

12.

l3.

14.

15.

16.

17.

Ryoo,R.; Jun, 8.; Kim, J.M.; Kim, MJ. J. Chem. Soc., Chem. Commun. 1997,
2225.
Tanev, P.T.; Chibwe, M.; Pinnavaia, T.J. Nature 1994, 368, 321.
Zhang, W.; Pauly, T.R.; Pinnavaia, T.J. Chem. Mater. 1997, 9, 2491.
Pinnavaia, T.J.; Zhang, W. Stud. Surf. Sci. Catal. 1998, 117, 23 (Elsevier Science
B.V.).
Zhang, W.; Froba, M.; Wang, J.; Tanev, P.T.; Wong, J.; Pinnavaia, T.J. J. Am.
Chem. Soc. 1996, 118, 9164.
Yang, R.; Pinnavaia, T.J.; Li, W.; Zhang, W. J. Catal. 1997, 172, 488.
Mokaya, R.; Jones, W. J. Chem. Soc., Chem. Commun. 1996, 983 and J. Catal.
1997, 172, 211.

Kloetstra, K.R.; van Bekkum, H.; Jansen, J.C. J. Chem. Soc., Chem. Commun.
1997, 2281.
Tue], A.; Gontier, S. Chem. Mater. 1996, 8, 114.
Corma, A.; Fomes, V.; Navarro, M.T.; Perez-Pariente J. Catal. 1994, 148, 569.
Hitz, S.; Prins, R. J. Catal. 1997, 168, 194.

Corma, A. Chem. Rev. 1997, 97, 2373; and Sayari, A. Chem. Mater. 1996, 8,
1840.
Armengol, E.; Cano, M.L.; Corma, A.; Garcia, H.; Navarro, M. J. Chem. Soc.,
Chem. Commun. 1995, 519.
Armengol, E.; Corma, A.; Garcia, H.; Primo, J. Appl. Catal. A: General 1995,
126, 391.
Szostak, R.; Molecular Sieves Principles of Synthesis and Identification; Van
Nostrand Reinhold: New York, 1989.
Beck, J.S.; Vartuli, J.C.; Roth, W.J.; Leonowicz, M.E.; Kresge, C.T.; Schmitt,
K.D.; Chu, C.T.-W.; Olson, D.H.; Sheppard, E.W.; McCullen, S.B.; Higgins, J .B.;
Schlenker, J .L. J. Am. Chem. Soc. 1992, 114, 10834.

Ryoo, R.; Ko, C.H.; Howe, R.F. Chem. Mater. 1997, 9, 1607.

109

18.

19.
20.

21.

22.

Janicke, M.T.; Landry, C.C.; Christiansen, S.C.; Kumar, K.; Stucky, G.D.;
Chmelka, B.F. J. Am. Chem. Soc. 1998, 120, 6940.

Luan, Z.H.; Cheng, C.F.; He, H-Y.; Klinowski, J. J, Phys. Chem. 1995, 99, 10590.
Gunnewegh, E.A.; Gopie, S.S.; van Bekkum, H. J. Mol. Catal. A: Chemicals
1996, 106, 151.

Luan, Z.H.; Cheng, C.F.; Zhou, W.Z.; Klinowski, J. J. Phys. Chem. 1995, 99,
1018.

Corma, A.; Kan, Q.; Navarro, M.T.; Perez-Pariente, J.; Rey, F. Chem. Mater.
1997, 9, 2123.

110

Chapter 5

Toward MCM-4l with Meso-scaled Channel Lengths and Textural
Pores for Catalysis

Abstract

The concept of altering structures of mesoporous silica molecular sieves through
introducing organic promoters to bind to the head group of charged surfactants in the
electrostatic S+I' assembly has been successfully illustrated. Mesoporous silica molecular
sieves with well ordered 3-dimensional hexagonal, cubic, and uni-dimensional hexagonal
phases were selectively assembled using cetyltrimethyl ammonium bromide as the
structure director and sodium silicate as the silicon source under a relatively low
surfactant to silicon ratio (i.e. 0.11) at 150 °C. Intrinsically, the structural variation was
strategically realized through surfactant-organic promoter association that leads to the
variation of surfactant micelle packing parameters (g = Vlad) and, ultimately, the
inorganic meso-structures. Organic promoters such as salicylic, tartaric, oxalic, and citric
acid were found effective to alter meso-structures from hexagonal to either 3-
dimensionally hexagonal or cubic. These promoters were also effective at mild
hydrothermal condition (i.e. 100 °C) for synthesis of particular MCM-41 with a flaky
morphology. These flakes exhibit mesopore channels running orthogonal to the flaky
sheets with meso-scaled lengths. Considerably well ordered MCM-41 molecular sieves
with meso-scaled fundamental particles and high textural porosity were also prepared by
careful control of the addition of organic acids.

Consequently, a follow-up in situ alumination of the flaky and textural pororosity

MCM-4l in the presence of surfactant micelles inside the pores affords well incorporated

111

Al-sites with over 95% are in tetrahedral coordination even for Al-loading up to
8.8mol%. However, up to a third of these Al-sites are transformed into octahedral sites
upon the surfactant removal by calcination. Al-sites judged from 27Al-NMR and the total
acidity measured from TPD of cyclohexylamine are indistinguishable from each other for
similarly loaded Al-MCM-41 catalysts regardless of particle morphology. Consequently,
any differences in catalysis for these Al-MCM-41 catalysts can not be attributed to any
differences in the intrinsic acid siting. The number of Al-sites with stronger acidity
indicated by the desorption of pyridine at 350 °C was proportionally increased with Al-
loading, but the number of Al-sites with weaker acidity indicated by the desorption of
pyridine at 217 oC remained the same.

The results of the studies of 2,4-DTBP alkylation with cinnamyl alcohol proves
that the accessibility of the active Al-sites in flaky and textural porous Al-MCM-41 with
short channel lengths is significantly facilitated. At least 30% higher conversion of 2,4-
DTBP and 50% higher yield of flavan are obtained relative to Al-MCM-4l catalyst
without a flaky or textural porous morphology. This alkylation reaction is likely
catalyzed by the mild acid sites since increasing the number of stronger acid sites by

increasing the Al-loading does not improve the activity considerably.

112

5.1 Introduction

M4lS materials are ordered, mesoporous silicas and aluminosilicate molecular
sieves with pore size in the range of 20 — 100 A.1 Among M41S family members,
hexagonal MCM-41 and cubic MCM-48 are very promising catalysts and catalyst
supports for chemical reactions involving large molecular substrates.2 The pore channels
for MCM-48 are bi-continuous, branched and tortuous.3 The pore channels for MCM-4l,
however, are uni-dimensional. Uni-dimensional channels may invoke counter-stream
diffusion resistance, because the reactants need to diffuse in and the products need to
diffuse out of the tubular channels.4 The significance of this particular drawback for
MCM-4l has not yet attracted attention.

Even though MCM-48 exhibits attractive 3-d pore connectivity, the difficulty to
synthesize hetero-atom incorporated cubic meso-structures severely limited the study of
this particular meso-structure and far fewer reports have been published.5'8 Actually, the
synthesis of Al-MCM-48 is even more problematic, most of the time the obtained
materials lack catalytic properties that are distinctive from MCM-4l. Other than the 3-d
pore connectivity, MCM-48 has not been found to be fundamentally different from
MCM-4l.7 Therefore, for the purpose of catalysis it is desirable to simplify MCM-48
synthesis or to improve the accessibility to the uni-dimensional channels of MCM-41.

MCM-4l silica molecular sieves have been synthesized under various conditions.
Two distinguishable electrostatic assembly pathways have been reported using positively
charged quaternary ammonium surfactants (S+) as the structure director. One is the
surfactant—inorganic precursor charge matching mechanism (SIT) working under basic

conditions using silicate species (I') and the other is the counterion mediated assembly

113

mechanism (S+X'l+) working under acidic conditions using protonated silicon alkoxide
species (1+). 9 The later one encounters significant difficulties for the synthesis of metal-
substituted derivatives since the catalytically active metal-centers are barely incorporated
into the silica framework under the severe acidic conditions (i.e. pH < 2).10 Although it is
possible to synthesize MCM—41 under ambient conditions via the ST assembly pathway,
the structure obtained, however, is lacking in thermal and hydrothermal stability.ll
Therefore, most MCM-41 samples for catalysis, especially Al-substituted analogs, were
synthesized under hydrothermal conditions through the ST pathway."12

To prepare MCM-41 with well ordered hexagonal structure via S‘T pathway,
relatively high concentrations of surfactant and high temperatures (> 100 °C) were
adopted by Mobil scientists.l Unlike most zeolitic structures that normally crystallize
under very basic conditions (for example, NaY forms at pH above 12), mesoporous
MCM-41 structures with amorphous silica framework walls usually form under less basic
condition (i.e. pH <11) to minimize the dissolution of amorphous silica walls. Therefore,
MCM-41 synthesis under basic conditions using sodium silicate as the silicon source has
to use some kind of acid to lower the pH. Ryoo and his co-worker reported a step-by-step
acid titration approach to obtain well ordered MCM-41 using acetic acid over 4 days at
100 °C.13 White and his co-worker later reported that sulfuric acid was the best acid for
adjusting pH during MCM—4l synthesis by following this titration approach.14 Similarly,
Mon and his co-workers'5 prepared a hierarchical tubular MCM-41 structure through a
day by day addition of sulfuric acid at 100 °C. The function of acid, other than adjusting

pH in these preparations, is not clear yet.

114

In general, meso-structures form through surfactant micelle assembly involving
inorganic reagents cross-linking and condensing around the micelles to build a
framework. As reviewed by Behrens16 the interfacial interactions between surfactant
micelle and inorganic precursors are very crucial for assembly of meso-structures. The
interactions primarily determine the final phase of the meso-structures.l7 Actually, in the
classical surfactant micelle chemistry,""20 the interaction energy is the key factor for the
surfactant packing parameter g (g = V/(aol)). If g is around 1/2, surfactant micelles tend to
form a hexagonal phase. A cubic phase forms for g values between 1/2 and 2/3. If the g
value approaches 1/3, a 3-dimensional (3-d) hexagonal or a spherical cubic phase is
favored. Therefore, to tailor the structure and pore conductivity of MCM-41, the easiest
way would be to control interfacial interactions.

The first objective of the present study was to alter the interfacial interaction to
obtain versatile mesostructures using organic acid as the structure promoter. The second
objective of our study is to slightly alter the interfacial interaction using organic acid as
structure modifier so that meso-scaled fundamental particles and channel lengths could
be obtained for MCM-41. Organic acid was chosen, because salicylic acid was known to
associate with the head group of cetyltrimethyl ammonium bromide to form helical
fibers.21 By using carboxylic acid, our strategy is to minimize the association energy to
such a point that it is not strong enough to cause a phase transition but towards a phase
transition, allowing the channel lengths and the orientation of particle growth to be
tailored.

The third objective was to prepare a well-ordered MCM-41 aluminosilicate with

textural porosity. Davis and his coworkers reported 22 that the textural porosity for MCM-

115

41 was closely related to the disorder caused by surfactant-rod-micelle-entangling
because a significant amount of N2 uptake in the isotherm for textural pores was
accomplished without seeing higher order X—ray reflections in the powder patterns. Our
results indicate that the textural porosity is not necessarily associated with the structural
disorder. However, it is closely associated with meso-scaled fundamental particles.
Fundamental particles aggregate to form meso—scaled inter-particle voids as the textural
pores.

Textural porosity along with meso-scaled fundamental particles indeed facilitates
the access to catalytic centers inside mesopore channels.23 In order to shown the
importance of textural porosity for catalysis, alkylation of 2,4-di-tert-butylphenol with
cinnamyl alcohol in the presence of MCM—41 aluminosilicate molecular sieves as the
catalyst was conducted under very mild conditions (i.e. 60 °C, 6 h). Significantly higher
conversion (i.e. >60%) of 2,4-DTBP and higher yield of the desired product flavan
(~50%) was obtained for Al-MCM-41 with high textural porosity and meso-scaled

fundamental particles relative to the analogs without textural porosity.

5.2 Experimental

MCM-48 and 3-dimensional hexagonal mesoporous silicas were synthesized at
150 °C through surfactant directed assembly in the presence of organic promoters such as
salicylic, citric, tartaric and oxalic acids. A typical synthesis was started by mixing a
sodium silicate aqueous solution as the silicon source and cetyltrimethyl ammonium
bromide as the structure director at ambient temperature. The structure promoters were

added to the mixture at ambient temperature under vigorous stirring for 15 minutes

116

before the mixture was transferred to a Teflon lined autoclave to age at 150 °C for 15 h.
The composition for the synthesis was following: 1.0Si : 0.11CT AB : 0.39Na20 :
0.28H+ : 130HzO. In the case of salicylic acid, the promoter to surfactant ratio was
controlled to 1, which also contributed part of 0.28 mole H" per mole of silicon and the
rest part of H” was compensated by sulfuric acid. In the case of citric, tartaric and oxalic
acid, all of the H+ were came from these promoters.

MCM-4l sample was prepared using 27wt% sodium silicate solution as the
silicon source and cetyltrimethyl ammonium bromide as the structure director. An
organic acid such as tartaric, oxalic, citric, gluconic, and salicylic acid was used as the pH
adjuster, as well as the structure modifier. Sulfuric acid was included for comparison.
Three types of MCM-4l with different textural porosity were prepared by controlling the
amount and the time for acid modifier addition. The molar composition for the synthesis
of MCM-4l with textural porosity was:

0.02 — 0.10 Al : 1.0 SiOz : 0.25 Surf. : 0.39 NaZO : 0.58 H+ : 83 H20
To control the pH as well as the association energy between the organic acid and the
surfactant head group, part of the acid was added to the sodium silicate aqueous solution
before the surfactant was added. The remaining part was added later. In a typical
experiment, 83% of the total amount of acid shown in the aforementioned composition
formula was added to the sodium silicate solution at the beginning, and the remaining
17% was added after 1 day heating at 100 0C. The MCM-4l obtained exhibits textural
pores similar to the textural pores formed by aggregates of layered clay particles. A high

textural porosity MCM-41 was synthesized by the addition of 100% of the required

117

amount of acid in one-lot after mixing the surfactant and sodium silicate solution, and
then heating at 100 0C for 2 days.

A typical Mobil MCM-4l without textural porosity was prepared using sulfuric
acid at 100 0C for 2 days. The amount of sulfuric acid was 17% less than organic acid
used in the aforementioned compositions. The acid was added at the very beginning and
the pH of this synthesis was ~10.5.

For siliceous MCM-4l, the product was filtered after 2 days aging at 100 °C. For
Al-MCM-41, A12(NO3)3 aqueous solution was added at ambient temperature under
stirring to the as-made siliceous MCM-41 in its mother liquor, and then it was statically
aged at 100 0C for another 2 days. Actually, this was still an in situ alumination process
in a follow-up manner. Aluminum nitrate was chosen as the aluminum precursor over
other aluminum sources such as aluminum sulfate and sodium aluminate. The purpose for
this choice is to avoid the undesired acidic sites generated by residual anions (i.e. 8042')
in the case of aluminum sulfate and structural alternation due to pH increase in the case
of sodium aluminate.

The product was filtered and washed thoroughly with water. The filter cake was
ion-exchanged in 0.1 M W 03 solution at ambient temperature for 4 h if needed. Other
wise, the filter cake was dried in air and directly transformed into the final catalyst
without prior NHf-exchangment by calcination at 540 °C for 6 h. It is possible that the
Na+ for neutralization of the negative framework charges generated by Al-incorporation
might be replaced by charged surfactant molecules inside the mesopores. In this case, A]-
MCM-41 prepared using sodium silicate may not require ammonium exchange to

transform it into a catalytically efficient H-form.

118

The powder X-ray diffraction patterns of the samples were recorded on a Rigaku
Rotaflex Diffractometer using Cu K0 radiation (71,: 1.542 A). N2 adsorption and
desorption isotherms were measured on an ASAP 2000 sorptometer at -196 °C. The
sample was evacuated under 10-5 torr at 150 °C over night before the measurement. 27Al
NMR spectra were recorded on a Varian VXR-400S spectrometer with a 7-mm zirconia
rotor and a spin frequency of 4 kHz. TEM images were taken on a JEOL IOOCX with a
CeB6 gun that was operating at acceleration voltage of 120 kV. The specimens were
loaded onto a holey carbon film that was supported on a copper grid by dipping the grid
into a solid sample suspension in ethanol.

Alkylation of 2,4-di-tert-butylphenol (2,4-DTBP) with cinnamyl alcohol was
carried out in a 100 ml round bottom flask. A 1.0 mmol quantity of 2,4-DTBP, 1.0 mmol
cinnamyl alcohol and 50 ml isooctane as the solvent were used. A 500 mg quantity of
catalyst was added when the reaction mixture was heated to 60 °C. The reaction was kept
at 60 0C under magnetic stirring for 6 h. The catalyst was filtered and washed with
dichloromethane after the reaction. The dichloromethane was separated from the filtrate
by distillation under reduced pressure at ~30 °C, so that heat-induced side reactions
could be prevented. The reaction products were qualitatively confirmed using a GC-MS.
They were quantitatively analyzed by an internal standard method by using a GC with a
FID detector and a 15m long SPB-l type capillary column. The conversion was
calculated by dividing the amount of 2,4-DTBP consumed by the amount initially
present. The selectivity toward flavan was calculated by dividing the amount of falvan
formed by the amount of 2,4-DTBP consumed. The yield of flavan was obtained by

dividing the amount of flavan formed by the amount of 2,4-DTBP added. The conversion

119

of cinnamyl alcohol was virtually complete in most reactions because cinnamyl alcohol
alone could transform into many undesirable products 24 such as cinnamyl aldehyde, 1,5-
di-phenyl pentadiene, and polymeric aldehyde. The latter product formed a tar on the

catalyst, which required calcination in air to regenerate the catalyst.

5.3 Results and Discussion
5.3.1 Effect of Organic Acid Addition on Silica Mesostructure Formation

It has been known for a decade that salicylic acid can associate with head group
of cetyltrimethyl ammonium surfactant to form remarkable helical micelle fibers.2| Some
other a— and B-hydroxyl carboxylic acids interact with charged surfactants in basic
medium to form similar micelle fibers too. It has also been known that the surfactant
packing parameter g can be varied upon variation of the surfactant head group size and
the alkyl chain length. Actually, variation of g parameter to achieve new meso-structures
has been shown quite successful in the preparation of SBA-2. 25 Landry and his co-
worker recently reported 26 that the g parameter of cetyltrimethyl ammonium bromide
could be varied to allow formation of cubic MCM—48 instead of hexagonal MCM-4l by
using 3 mole of EtOH per mole of silicon. Presumably, association of carboxylic acid to
the head group of surfactant would affect the g parameter also. We report here that
carboxylic acids indeed promote cetyltrimethyl ammonium bromide surfactant to form 3-
d hexagonal as well as cubic silica meso-structures under the reaction conditions
normally give rise to uni-dimensional MCM-41.

As shown in Figure 5.1, a MCM-48 like cubic silica meso-structure was formed

in stead of MCM-41 upon addition of 1 mole of salicylic acid per mole of surfactant. The

120

Figure 5.1.

 

2" S”I' Pathway

S" = CTAB
l' = Sodium Silicate
$71- = 0.12

 

EtOH, MOM-48

Intensity
420
332

 

No EtOH, MGM-48
Salicylic acid/CTAB = 1.0

 

100

110
200

No EtOH, MGM-41

 

 

No EtOH, MOM-41
Salicylic acid! CTAB = 3.0

L L .l 1 l J I 1 1 l

O 5 1O 15 20
29 (degrees)

 

 

 

 

X-ray diffraction patterns for calcined (540 oC) cubic silica meso-
structures forrned at 150°C with a reaction composition of 1.0Si :
0.11CT AB : 0.39Na2O : 0.28H+ : 130H2O. The diffraction patterns for
MCM-48 prepared by following Landry’s method as well as the
diffraction patterns for a smaple prepared in the absence of both EtOH

and salicylic acid are included.

121

 

S‘T Pathway
3-d Hexagonal Pam"

100
002

 

 

 

 

 

fl 8" = CTAB
l' = Sodium Silicate

$71. = 0.12

> s

-I-' .- . . .

'5', o o. Cltrlc acid/CTAB =1.0

c 3: 37-

d) —— -

4.0

E
Oxalic acid/CTAB = 1.5

Tartaric acid/CTAB = 1.5

 

 

 

l l l l l l l l l l l L l

2 4 6 8 1o
26 (degreeS)

 

Figure 5.2. X-ray diffraction patterns for calcined (540 0C) 3—d hexagonal silica meso-
structures formed under the same conditions as noted in Figure 5.1 with

the promotion of citric, oxalic and tartaric acids.

122

diffraction patterns of MCM-48 prepared through Landry’s procedure and a comparison
MCM-4l sample prepared in absence of salicylic acid and ethanol are included in Figure
5.1 too. Interestingly, too much salicylic acid deteriorates the meso-structure while 3
mole of salicylic acid per mole of surfactant was used. This indicated that the association
2' of carboxylic acid with the charged surfactant head group results in diminished long
range 8+1~ electrostatic interaction. Consequently, the structure was less ordered judged
by X-ray diffractions. Figure 5.2 shows the well resolved XRD patterns for 3-d
hexagonal silica formed upon the addition of 1 mole of citric acid per mole of surfactant.
Oxalic acid and tartaric acid shown similar promotion effects. These results strongly
suggest that carboxylic acids can function as promoters to change the surfactant packing
parameter and ultimately cause the silica meso-structure to change from uni—dimensiona
hexagonal to 3-d hexagonal and even cubic phases under certain reaction conditions. In
comparison to the formation of conventional MCM-41, 3-d hexagonal and cubic meso-
structures were formed under relatively lower surfactant to silicon ratio (i.e. 0.11 vs.
0.25) and H+ to NaOH ratio (i.e. 0.5 vs. 0.75) at 150 °C under the promotion of organic

acids.

5.3.2 Siliceous MCM-41 Type Mose-structure Formation

Mou and his co-workers proposed '5 that a hierarchical tubular MCM-41 formed
through a sheet-like intermediate. The tubular MCM-4l formed at 100 °C by ultimately
reaching an over all H+ to Si ratio of 0.57 when sulfuric acid was added day-by-day one
time a day over 4 days. Our original goal was to synthesize MCM-41 with thin sheet

morphology so that the accessibility of framework channels would be greatly enhanced.

123

Figure 5.3.

 

   

 

   

Gluconic Acid
6? 1 : : : : : . it”;
('3 1000, Lactic Acid
as I
a: M”
g ,' Oxalic Acid
800‘

 

' Citric Acid

' Sulfuric Acid
Mobil MOM-41

N Volume adsorbed
a:
O
9

 

Montmorillonite ....‘

‘:.: . . - o ‘ . ‘

   

 

 

0 0.2 0.4 0.6 0.8 1
PlPo

N2 adsorption and desorption isotherms for calcined (540 °C) siliceous
MCM-41 prepared under the promotion of gluconic, lactic, oxalic, and
citric acids with a H+/NaOH ratio of 0.75 showing framework
mesoporosity as well as clay-like textural porosity similar to that of
montrnorillonite. The Mobil MCM-41 sample was prepared using sulfuric
acid to adjust the pH (i.e. ~10.5) at the very beginning of the synthesis by
control the over-all H‘VNaOH ratio to 0.625.

124

In order to form and isolate the “sheet intermediate” organic acid was purposely added
one time a day, and the mixture was aged only for two days instead of 4 days at 100 0C.
The over all H+ to NaOH ratio was controlled to 0.75. MCM—41 with characteristic N2
adsorption-desorption hysterisis loop similar to that for textural pores of layered
montmorillinite was obtained (c.f. Figure 5.3). The isotherm for a typical MCM-41
prepared using sulfuric acid does not show any textural pores in contrast. TEM and SEM
images of this sample showed tubular MCM-41 particles as the dominant form (Figure
5.4A and 5.4B) along with minor potion of flake-like particles (Q0%). The flaky
morphology in the MCM-41 samples prepared using organic acids, however, was the
major morphology (c.f. Figure 5.4C and 5.4D), no tubular particles were observed. The
pores of flaky particles were funning orthogonal to the flaky-surface. Obviously, the pore
channel lengths were limited to the thickness of the flakes. That is estimated to be less
than 100 nm, because the penetration depth for an electron beam accelerated at 120 kV is
about 100 nm.27 These uni-dimensional pores with short channel lengths should be
beneficial for diffusion limited catalytic reactions.

Actually, MCM-4l molecular sieves with clay-like flaky features indicated by the
presence of textural porous hysteresis loop in the isotherms (Figure 5.3) were prepared
using many organic acids while keeping the over-all H+lNaOH ratio at 0.75. All of them
exhibited a ~3 A width at half of maximum for the Harvoth-Kawazoe pore size
distribution. They all exhibit very well resolved hexagonal X-ray patterns with (100),
(110), (200), (220) and (300) reflections clearly seen (Figure 5.5). Although White and
his co-workers claimed 14 that a well ordered MCM-41 was obtained through tedious

sulfuric acid titration during 4 days heating at 100 °C, the sample prepared by following

125

 

 

 

Figure 5.4A and 5.48

126

Figure 5.4

 

 

TEM images showing (A and B) interwoven tubular MCM-41 formed
upon use of sulfuric acid; (C and D) thin sheet MCM-41 (formed upon use
of citric acid) with pore channels running orthogonal to sheet surface (the
thickness of the sheet estimated to be ~ 100nm according to the penetration
depth of electron beam accelerated at 120 kV); Both samples were
prepared with an over all HVNaOH ratio of 0.75.

127

 

11d

 

 

 

  

 

 

 

81¢“
IIIIIIIIIIIIIIIIIIIIIIIIIII]
_ 5 1O 15 20 25 30
a. 610 . 29
7, Gluconic Acid
r:
0 .
H
5 3
4 104 M Oxalic Acid

 

210‘— k Citric Acid

Lactic Acid

—
O IITIIfj’TIIITrTIIIIIIIIIIIIIII

0 5 10 15 20 25 30
26) (degrees)

 

 

 

 

Figure 5.5. X-ray diffraction patterns for MCM-41 molceular sieves as noted in

Figure 5.3.

128

their procedures using sulfuric acid does not exhibit significantly superior order over the
sample prepared using organic acid. TEM images (c.f. Figure 5.4) for flaky MCM—41
show primarily sub-micrometer sized thin (< 100nm) sheets with pore channels running
orthogonal to the sheet surface. However the particle size of MCM-4l reported by White
and Ryoo et al are in micrometers. 28

Textural porosity HMS molecular sieves with inter particle voids of aggregated
meso-scaled fundamental particles normally form upon fast hydrolysis and nucleation of
hydroxylated neutral silicon species in the presence of alkylamine surfactant.29 In order
to facilitate the nucleation of silicate species in MCM-41 synthesis, we purposely added
an amount of organic acid equivalent to a H+lNaOH ratio of 0.75 in one lot instead of
two to allow the pH drop down to ~9 immediately, and then let the reaction mixture to
age at 100 0C for 2 days. As illustrated in Figure 5.6, MCM-41 with a significant textural
pore volume was produced in this way. The same type of synthesis performed at ambient
temperature, however, afforded MCM-41 with lower framework as well as textural pore
volume. Accordingly, the X-ray reflection patterns for this particular sample only showed
a broad (100) reflection with a hump at higher 20 angles (c.f. Figure 5.7). This signifies
the importance of temperature in nucleation and assembly of meso-structures.

In agreement with HMS materials,28 the high textural porosity of MCM-4l was
correlated with the low bulky density (~0.2 g/cc), which is only half of the value of the
MCM-41 without textural pores. This suggests that the high textural porosity MCM-41
samples contain large amount of fluffy tiny particles. In addition, Figure 5.7 shows that
MCM-41 with textural porosity still exhibits X-ray patterns with well resolved (100),

(110) and (210) reflections. This is in contrast to the report by Davis and his co-

129

 

 

 

Figure 5.6.

MGM-41 assembled with
Tartaric Acid modification

 

 

 
 
  
  

 
 
 
 

 
  
     

 

  

 

  

 

 

 

1600- pH10.5,then 9,
‘ 3 100°C
1400 - E I pH1os.1oo°c
E ‘ ' a. pH9,100°C
m 120020105413“,
03 - H.5_.___
E i 10 20 so 40 so so 70 80
V 1000': H-Kpore
3’ 800; (‘-
3 ‘ pH=10.5, then 9, r.t.
(U 4 lac-ocvm'”
or:
g 500 pH =1o.5, then 9,100 °c
'5
> 400 .-- - e e e e e : 2292;919:91"
z“l " o
pH=10.5,100 c
200 “:‘."g‘..... u
airwmwuwr...
0 0.2 0.4 0.6 0.8 1
PIPo

N2 adsorption and desorption isotherms for MCM-41 molecular sieves
prepared using tartaric acid as the structure promoter under various
conditions showing the controlled formation of significant amount of
HMS-like textural pores as well as clay-like textural pores. Horvath-

Kawazoe pore size distribution plots are included in the inset.

130

 

 

 

Calcined (540 °C) MGM-41
Assembled Via S‘I' Pathway
With Tartaric Acid Modification

 

 

 

61d_

51dj

41d- n. ....................

: i I‘5 10 159 20 215 310

g- ; pH = 10.5, then 9,100 °c
2 31d-
2 -
5

210— pH=10.,5100 °C

 

 

4
110 LU pH= 9,100 °C

pH= 105 then9 rt.
0 l 'l""l"‘7

0 5 10 15 20 25 30
26 (degrees)

 

 

 

Figure 5.7. X-ray powder diffraction patterns for MCM-41 molecular sieves as noted
in Figure 5.6. Notably, the MCM-4l with high textural porosity exhibits

well resolved (100), (110) and (200) reflections.

131

Figure 5.8.

 

TEM image of MCM-4l showing meso-scaled fundamental particles with
inter particle voids as textural pores and regular hexagonal framework
mesopores running orthogonal to thin sheet surfaces exemplified by a

sample prepared using tartaric acid as the structure promoter.

132

22

workers, which stated that disordered MCM-41 with single X-ray reflection and a N2
hysteresis loop in the isotherm at relative pressure larger than 0.7 was a consequence of
rod-like surfactant micelle entanglement. Therefore, X-ray disorder of MCM-4l is not
necessarily associated with textural porosity.

In accordance with the samples prepared using citric acid, oxalic acid, gluconic
acid and acetic acid, the sample synthesized through two step addition of tartaric acid
also exhibited a nice X-ray diffraction pattern containing 5 well resolved reflections
(Figure 5.7). In addition, they all exhibited a nice flaky-type hysterisis in the isotherm
(c.f. Figure 5.6). These results suggest that the “intermediate sheets” to tubular MCM-41
were likely captured to form flaky MCM-4l under the promotion of those carboxylic acid
promoters.

Similar to HMS, MCM-41 with high textural porosity does exhibit meso-scaled
fundamental particles (Figure 5.8), which are aggregated into fractal-like larger particles.
Since this sample was not thin-sectioned, but dusted on the copper grid for taking TEM
images, the specimen thickness may be considered as the particle thickness. Therefore,
the thickness of these fundamental particles would be estimated to be less than 100 nm,
also. Moreover, as long as the pore channels are running orthogonal to the particle

surface, the channel length of this particular MCM-4l would be estimated as ~100 nm,

comparable to the particle thickness.

5.3.3 Catalyst Preparation

Normally, Al-MCM-41 is synthesized through an aluminum-silicon co-assembly

process. However, Al-MCM-41 prepared in this way loses higher 29 angle X-ray

133

reflections most of the time. The higher the Al-loading, thebroader the (100) reflection.
In order to obtain well ordered Al-MCM-41, a follow-up in-situ alumination in the
presence of surfactant micelles in the pores was reported by Ryoo and his co-workers.28
However, this particular sample indicated inhomogeneous alumination of the framework
with more aluminum incorporated near the pore openings than deep inside the channels.
Actually, Ryoo’s MCM-41 was synthesized through an acid titration approach which
gave rise to a well ordered MCM-4l with a relatively large particle size (even larger than
2 pm”). Therefore, the migration of aluminum species to deep inside of the channel is
very likely hindered in the presence of surfactant molecules. Moreover, the electrostatic
interfacial interaction would generate an electric field to prevent negatively charged
aluminum species from accessing the negatively charged framework. Hence, the
alumination would not be fully completed. To overcome this difficulty and make the
alurrrination of the pore walls more homogeneous in the pores, using aluminum nitrate as
the alumination agent for our MCM-41 with meso-scaled fundamental particles and
meso-scaled channel lengths should have advantages over other MCM-41 samples.

The follow-up in-situ alumination of MCM-41 with flaky particles was only
successful for Al-loading up to 9%. The flaky particle morphology was lost in attempts to
introduce 20% aluminum. At these higher Al loading, the characteristic hysteresis loop
responsible for inter particle voids in the isotherms was diminished. Seemingly, the flaky
particles were stacked into larger particles. This suggests that flaky sheets are more likely
an intermediate of tubular MCM—41.

In contrast to intermediate flaky sheets, MCM-41 with textural porosity showing

meso-scaled fundamental particles is stable to the follow—up in situ alumination treatment

134

even for an attempt to introduce 20% aluminum. Elemental analysis of the sample with

20% initially Al-content in the synthesis mixture revealed that the actual Al-loading for

 

 

    
 

 

 

 

 

 

 

 

1 500
A
‘ A
1 A
A j g
g . E Textural
(D 1 ‘t‘
651 ooo~
B .
3 ~ .
cl IIIIIIIIIII'III'IIIITIT "
1 0 2 4 6 8 10 12 ‘f‘
h A"
O 1 H-K Pore Size (nm) ‘ ‘ A .1 A" ,2 LA-
.3 AA-A\“‘ " I
< . " Al-MCM-41 (Textural) ,
at . ‘
E 500- ..-‘
3 o/‘L
3 ' 2'1"“
> ‘ ‘§““
N " A9“ /
Z . f
. T I I l I I I j I I I I I I I I I I I

 

0 0.2 0.4 0.6 0.8 1
PIPo

Figure 5.9. N2 adsorption and desorption isotherms for 8.8mol%Al-substituted MCM-
41 molecular sieves (calcined at 540 °C) with clay-like and HMS-like
textural porosity prepared using tartaric acid as the structure promoter and
aluminum nitrate as the follow-up in situ alumination reagent in the
presence of surfactant miccelles in the pores. (The inset is the
corresponding Horvath-Kawazoe plots showing sharp narrow distribution

of framework mesopores).

135

  
   
   
  

Al-MCM-41

Octfl’etr
8.8% AI

(T extural porous,
540 °C)

21179

88%AI
(Flaky, 54o °c

 
   
    

 

 

 

 

8.8%AI
0/1 00 (Tex, as-made)
8.8%AI
0,100 (Flaky, as-made)
' I r ' ' I ' ' ' I ' 'A‘ I ' ' .fl‘,
400 200 0 -200 -400

mm

Figure 5.10. 27Al-NMR spectra for as-made and calcined Al-MCM—41 molecular sieves
prepared using tartaric acid as the structure promoter with HMS-like and

clay-like flaky textural pores.

136

the final product was only 10%. In fact, the 20% aluminum introduced in the form of
aluminum nitrate decreased reaction pH too much to allow the incorporation of Al into
the pore walls taking place effectively. As shown in Figure 5.9, this particular Al-MCM-
41 still exhibits a good textural pore volume up to 50 % of the total pore volume and well
resolved hexagonal X-ray patterns. Not only did tartaric acid afford Al-MCM-4l with
textural pores, but also oxalic acid, citric acid, gluconic acid and lactic acid afforded Al-
MCM-41 with quite good textural pores.

27Al-NMR spectra show that over 95% of total aluminum sites in the as-made
samples are in tetrahedral coordination (Figure 5.10). This result indicated that the
follow-up in situ alumination was as effective as in situ silicon-aluminum co-assembly
and post-synthesis implantation30 in the absence of surfactant. 2“’Al-NMR spectra of our
samples indicates dealumination upon calcination occurred regardless of high or low
textural porosity. Approximately a third and a fifth of the total aluminum-sites were
transformed from tetrahedral coordination to octahedral coordination upon calcination for
2%Al- and 8.8%Al-MCM-4l, respectively. This verifies the dealumination reported by
Corma 3' and Prins 32 et a1. Notably, there are no observable differences in 27Al-NMR
spectrum for MCM-4l with and without textural porosity. In conclusion both intra-
frarnework and extra-framework aluminum species are approximately identical from
sample to sample for all calcined MCM-4l samples with the same Al-loading regardless
of the textural porosity. Interestingly, the ratio of tetrahedral to octahedral Al-sites for the
calcined samples increased with increasing Al-loading. This was further confirmed by

temperature programmed pyridine desorption. As shown in Figure 5.11, the number of

137

Figure 5.11.

 

 

 

1 2 - Pyridine-TPD profile
1 _
8.8%Al-MCM-41
0.8 F 0.44 mummy/g
’6
\ _
3 0'6 0.17 mmol Pylg
1?
0.4 " 1“
1‘ III \I
0.2 _ : “\ I’ K
‘\i' .‘ \ 1",, ‘\
O _ 0.11 mmOI PYIgai"""'\./ 5..., -,-.
2.0%AI-MCM-41

200 250 300 350 400 450 500 550 600
Temperatu retC)
Temperature programmed desorption (TPD) of pyridine pre-adsorbed on
A1-MCM-41 catalysts and purged at 150 °C in N2 stream for l h before the

TPD started. This TPD profile was carried out on a thermal gravity

analyzer.

138

weaker acid sites indicated by pyridine desorption peak at 217 °C remained the same, but
the number of stronger acid sites indicated by pyridine desorption peak at 350 oC
increased by 4 fold when the Al-loading increased 4 times. Assuming the surfactant
micelles are homogeneously distributed through the pores, the sample without prior
NHX-exchangement but with higher Al-loading would require higher concentration of
Na+ cations at the pore opening to balance the negative framework charges generated by
tetrahedrally incorporated A1 if most of the Al-sites were at the pore opening as reported
by Ryoo et al. 28 Hence, non-linear increase of acidity value with Al-loading would be
expected. Notably, our two samples were not subjected to NHf-exchangment. A linear
increase of the number of stronger acid sites with increasing Al-loading strongly suggest
that aluminum was not likely just incorporated at pore openings, but rather
homogeneously incorporated into the framework through the pores.

In agreement with the results of Jones et al,33 our previous study also found that
2%Al-MCM-4l exhibited quite low acidity (0.1 mmol CHA/g, compared with 3.4 mmol
CHA/g for 2%Al-HMS) possibly due to inaccessible Al-sites buried in MCM—41
framework walls.34 In order to achieve an acidity for Al-MCM-4l more or less
comparable to Al-HMS, aluminum-loading for most of the MCM-41 was increased to
8.8%, which gave acidity levels of 0.44 mmol CHA/g, as measured by TGA analysis (c.f.
Table 5.1). This value is also comparable to that of the HY from Nth available from
Aldrich. As shown in Table 5.1, all Al-MCM-41 samples with similar Al-loading
regardless of surface area, porous volume and textural porosity are very similar to each
other in total acidity. Aluminum nitrate may provide nearly neutral aluminum

hydroxylated species in the alumination medium with a pH at ~9. Therefore, the follow-

139

Table 5.1. Physical and chemical properties of Al-MCM-41 mesoporous molecular

 

 

 

 

 

Al-MCM-lsillwf:S Si/Al S351 Pore Volume (cc/g)b Pore Size Acidity"

Catalyst ratio (m2/ g) Vt Vf, V,” (nm) (mmol lg)
H-Y zeolited 2.80 670 0.30 0.30 0 0.72 0.50
Mobil 50.0 1070 0.87 0.87 0 3.3 0.10
(Sulfuric Acid) 11.1 1051 0.92 0.86 0.06 3.2 0.47
Flaky (Citric Acid) 11.1 903 0.91 0.73 0.14 3.6 0.42
Flaky (Oxalic Acid) 11.1 990 0.87 0.71 0.16 3.5 0.43
Flaky (T artaric Acid) 11.1 922 0.87 0.71 0.16 3.5 0.44
Tex-porous (Tartaric) 50.0 927 0.89 0.72 0.17 3.5 0.14
Non-tex (Tartaric)° 1 1.1 967 0.83 0.83 0 3.5 0.44
Tex—porous (Tartaric) 11.1 1045 1.46 0.82 0.64 3.6 0.43
Tex-porous (Citric) 1 1.1 101 1 1.38 0.72 0.66 3.5 0.43
Tex-porous (Oxalic) 11.1 1023 1.13 0.71 0.42 3.5 0.43

 

 

3. All Al-MCM-41 samples were not NRC-exchanged except the sample prepared
according to Mobil’s method. b. The total pore volume (V,) and the framework pore
volume (V2,) were calculated from the isotherm at P/Po 0.99 and 0.5, respectively. The
textural pore volume (Vm) is the difference between V. and V2. °. The total acidity for
these catalysts were obtained by means of TGA through measurement of
cyclohexylamine deorption between 240 to 420 °C. d. This H-Y was obtained through
calcination (540 °C) of a NH4-Y with a Si/Al ratio of 2.8 and containing 2.5 wt% Na2O,
which was bought from Aldrich. °. This sample has no textural pores prepared using 17%
less amount of tartaric acid comparable to Mobile’s MCM-41.

140

 

up in situ alumination using aluminum nitrate as the reagent in the presence of surfactant
molecules in the pores are likely incorporated homogeneously into the framework all the
way through the pores. One more reasonable conclusion derived from the data in Table
5.1 and Figure 5.10 and 5.11 is that any differences in catalysis for those Al-MCM-4l

with equivalent acidity would not be attributed to any differences in intrinsic acid siting.

5.3.4 Catalytic Alkylation of 2,4-di-tert-butylphenol with Cinnarnyl Alcohol

As discussed in last section, a family of Al-MCM-41 with textural pores was
prepared using various organic acids through a general method. Like the preparation of
Al-HMS with textural pores, the follow-up in situ alumination of MCM-41 with textural

pores using aluminum nitrate is efficient enough to incorporate tetrahedral Al into the

OH
\ CH2OH 0
+ m Acid catalyst> I o

Cin-OH

 

2,4-DTBP
framework in the presence of surfactant molecules. In order to evaluate the catalytic

performance for acid-sites, alkylation of 2,4-di-tert-butylphenol was conducted over Al-
MCM-41 catalysts with various textural pores. The importance of accessibility to active
sites in diffusion limited catalytic reaction might be manifested by these Al-MCM-41
catalysts because of their different framework pore accessibility. In the first report 35 of
using mesoporous MCM-41 as the catalyst, this reaction was carried out at 90 °C for 24
h. As we stated,34 it is not necessary for running this reaction for 24 h because one of the
two reactants, cinnamyl alcohol is nearly 100% consumed after 6 h run at even 60 0C.

Most of the by-products (about 10% of the total products) from this reaction were from

141

 

 

 

cinnamyl alcohol alone. Three products were identified for conversion of 2,4-di-tert-
butylphenol by GC-MS analysis. These three products account up to ~90% of the
consumed 2,4-di-tert-butylphenol. The major product is flavan. Two by-products were 4-
ten-butylphenol and 4,6—di-tert-butyl benzofuran. Since these two by-products only
account for less than 5% of consumed 2,4-di-tert-butylphenol for all Al-MCM-41
catalysts, they were not listed in Table 5.2. Interestingly, 4-tert-butylphenol, the product
of dealkylation of 2,4-DTBP was barely detectable for most Al-MCM-41 catalysts, but
obviously detected for HY zeolite and textural pores-free Al-MCM-41 prepared using
tartaric acid that exhibited quite low flavan selectivity (normally, 11.1% and 46% for HY
and textural pores-free Al-MCM-41, respectively).

As shown in Table 5.2, the traditional liquid phase alkylation catalyst, sulfuric
acid was not at all effective for this reaction. Even though 2 times more protons from
sulfuric acid (30 mg) was used in comparison to the acidity of 500 mg 8.8%Al-MCM-41
catalyst (~0.45 mmol/g), the conversion of 2,4-DTBP, selectivity and yield of the desired
flavan were still insignificant. This may signify the importance of mild acidity observed
for aluminum substituted mesoporous molecular sieve catalysts in use for this type of fine
chemical alkylation. It seems that only one of those two types of acid sites on these A]-
MCM-41 catalyst detected by temperature programmed pyridine desorption (c.f. Figure
5.11) is effective for this alkylation reaction. Re-use of the not-regenerated catalyst gave
rise to a conversion and selectivity only ~15% lower than the values observed for the
fresh catalyst. Actually, the number of weaker acid sites with a pyridine desorption peak
at 217 °C still maintain the same after one cycle of the reaction. And the stronger acid

sites indicated by pyridine desorption at 350 °C was no longer available for pyridine

142

 

 

Table 5.2. Alkylation of 2,4-di-tert-butylphenol with cinnamyl alcohol catalyzed by

Al-substituted mesoporous molecular sieves

 

 

Al-MCM41 Conv. (%) Selectivity of Yield of

Catalyst 2,4-DTBP Flavan (%) Flavan (%)
30 mg H2804a 25.5 36.9 9.4
HY 13.3 11.1 1.5
Mobilb 47.2 71.8 33.9
(Sulfuric Acid) 50.0 75.2 37.6
Flaky (Citric Acid) 64.4 76.9 49.5
Flaky (Oxalic Acid) 61.9 71.2 44.1
Flaky (Tartaric Acid) 69.2 84.0 58.5
Tex-porous (Tartaric)b 69.7 79.5 55.4
Non-tex (Tartaric)c 46.3 45.9 21.3
Tex-porous (Tartaric) 65.8 83.5 54.9
Tex-porous (Citric) 64.4 79.5 5 1.0
Tex-porous (Oxalic) 63.0 79.0 50.0

 

 

*. Each catalyst contains 8.8 mol% Al except indicated samples; the reaction was carried
out using 10.0 mmol of each reactant, 500 mg of catalyst , in 50 ml of isooctane at 60 °C
for 6 h. a. The reaction was carried out at 90 °C for 24 h instead of 60 °C for 6 h . b.
These sample contains 2.0 mol% A]. c. This sample has no textural pores (c.f, Figure
5.6).

143

adsorption because they were totally covered by polymeric aldehyde formed from
cinnamyl alcohol condensation. Increasing the Al-loading leads to an increase of the
number of stronger acid sites with tetrahedral coordination (c.f. Figure 5.10 and 5.11).
However, the conversion and selectivity for alkylation does not improve with the increase
of Al-loading (c.f. Table 5.2 for Tex-porous-tartaric-entries). This indicates that the Al-
sites with mild acidity is crucial for this alkylation reaction in liquid phase.

The accessibility of active sites is another crucial factor for this bulky alkylation
reaction. The poor performance of HY zeolite as alkylation catalyst is simply due to the
inaccessibility of the active Al-sites in the micropores even though the amount of acidity
for HY is approximately equivalent to that of 8.8%Al-MCM-41. In contrast to the
conventional Mobil Al-MCM-4l catalyst with particle sizes of a few micrometers, the
catalyst with a thin flaky particle morphology and short pore channels exhibited much
higher 2,4-DTBP conversion and flavan yield (i.e. at least by 30%). Also, Al-MCM-41
prepared using sulfuric acid in the same way as the organic acid promoters to adjust the
pH two times over 2 days, exhibits primarily a micrometer sized tubular particle
morphology and an only slightly improved catalytic activity. However, the conversion of
2,4-DTBP, and the selectivity and yield to the flavan all increase with increasing textural
porosity. The conversion of 2,4-DTBP (i.e. 66%) for texturally porous Al-MCM—4l
molecular sieves was almost as good as texturally porous Al-HMS.34 The selectivity to

flavan was quite remarkable also (i.e. ~83%).

144

5.4 Conclusions

Well ordered MCM-41 molecular sieves with flaky morphology and short uni-
dimensional pores orthogonal to flaky-sheet surface, as well as meso-scaled fundamental
particles with high textural porosity were prepared through controlled addition of organic
acids such as tartaric, citric, oxalic, gluconic, lactic and salicylic acid. Structural
promotion effects of the organic acids were explicitly demonstrated by the formation of
3-d hexagonal and cubic structure at 150 0C (instead of 100 °C for MCM-41) and
surfactant to silicon ratio of 0.11 (instead of 0.25 for MCM-41) and H+ to NaOH ratio of
0.49 (instead of 0.75 for MCM—4l). Intrinsically, this is most likely a consequence of the
variation of surfactant packing parameter g. The g value is likely varied through the
surfactant-organic acid association to cause the phase change in formation of 3-d
hexagonal and cubic meso—structures. On the other hand, slightly variation of surfactant
packing parameters under conditions for MCM-4l preparations upon organic acid
promotion leads to the formation of flaky as well as texturally porous silica molecular
sieves. Follow-up in situ alumination of the flaky and textural porous MCM-41 in the
presence of surfactant micelles in the pores afforded well incorporated Al-sites with over
95% are in tetrahedral coordination even for Al-loading up to 8.8mol%. However, up to a
third of these Al-sites are transformed into octahedral sites upon removal of the surfactant
by calcination. The Al-sites judged from 27Al-NMR and the total acidity measured from
temperature programmed desorption of cyclohexylamine are in-distinguishable for
similarly loaded Al-MCM-41 regardless of particle morphology. Consequently, any
differences in catalysis for these Al-MCM-41 catalysts with equivalent acidity can not be

attributed to any differences in the intrinsic acid siting. The number of Al-sites with

145

stronger acidity indicated by desorption of pyridine at 350 °C was proportionally
increased with Al-loading, but the number of Al-sites with weaker acidity indicated by
desorption of pyridine at 217 OC remained the same.

The results of the studies of 2,4-DTBP alkylation with cinnamyl alcohol proves
that the accessibility of the active Al-sites in flaky and textural porous Al-MCM-41 with
short channel lengths is significantly facilitated. At least 30% higher conversion of 2,4-
DTBP and 50% higher yield of flavan are obtained relative to Al-MCM-41 without a
flaky or textural porous morphology. This alkylation reaction seems to be catalyzed by
the mild acid sites since increasing the number of stronger acid sites by increasing the Al-

loading does not improve the activity considerably.

146

5.5 References:

l.

199°.‘19‘

10.

11.

12.

13.
14.
15.

16.
17.

Beck, J.S.; Vartuli, J.C.; Roth, W.J.; Leonowicz, M.E.; Kresge, C.T.; Schmitt,
K.D.; Chu, C.T.-W.; Olson, D.H.; Sheppard, E.W.; McCullen, S.B.; Higgins, J .B.;
Schlenker, J .L. J. Am. Chem. Soc. 1992, 114, 10834.

Corma, A. Chem. Rev. 1997, 97, 2373; and Sayari, A. Chem. Mater. 1996, 8,
1840.

Alfredsson, V.; Anderson, M.W. Chem. Mater. 1996, 8, 1141.

Yang, R.; Pinnavaia, T.J.; Li, W.; Zhang, W. J. Catal. 1997, 172, 488.

Monnier, A.; Schuth, F.; Huo, Q.; Kumar, D.; Margolese, D.I.; Maxwell, R.S.;
Stucky, G.D.; Krishnamurty, M.; Petroff, P.; Firouzi, A.; Janicke, M.; Chmelka,
B.F. Science 1993, 261, 1299.

Anderson, M.W. Zeolite 1997, I9, 220.

Zhangw, W.; Pinnavaia, T.J. Catal. Lett. 1996, 38, 261.

Romero, A.A.; Alba, M.D.; Klinowski, J. J. Phys. Chem B 1998, 102, 123.

Huo, Q.; Margolese, D. I.; Ciesla, U.; Feng, P.; Gier, T. E.; Sieger, P.; Leon, R.;
Petroff, P. M.; Schuth, F.; Stucky, G. D., Nature 1994, 368, 317.

Zhang, W.; Froba, M.; Wang, J.; Tanev, P.T.; Wong, J.; Pinnavaia, T.J. J. Am.
Chem. Soc. 1996, 118, 9164.

a. Stucky, G.D.; Monnier, A.; Schuth, F.; Huo, Q.; Margolese, D.; Kumar, D.;
Krishnamurty, M.; Petroff, P.; Firouzi, A.; Janicke, M.; Chmelka, B.F. Mol. Cryst.
Liq. Cryst. 1994, 240, 187; b. Tanev, P.T.; Pinnavaia, T.J. Chem. Mater. 1996, 8,
2068.

a. Kloetstra, K.R.; van Bekkum, H.; Jansen, J.C. Chem. Commun. 1997, 2281; b.
Luan, Z.H.; Cheng, C.F.; He, H-Y.; Klinowski, J. J, Phys. Chem. 1995, 99,
10590.

Ryoo, R.; Kim, J .M. Chem. Commun. 1995, 711.

Edler, K.J.; White, J .W. Chem. Mater. 1997, 9, 1226.

a. Lin, H.-P.; Mou, C.-Y. Science 1996, 273, 765; b. Lin, H.-P.; Cheng, S.; Mou,
C.-Y. Chem. Mater. 1998, 10, 581.

Behrens, P. Angew. Chem. Int. Ed. Engl. 1996, 35, 515.

Huo, Q.; Margolese, D.I.; Stucky, G.D. Chem. Mater. 1996, 8, 1147.

147

 

 

18.

19.
20.

21.
22.

23.

24.

25.
26.
27.

28.
29.
30.
31.
32.
33.
34.
35.

36.

Israelachvili, J .N .; Mitchell, D.J.; Ninharn, B.W. J. Chem. Soc., Faraday Trans. 2
1976, 72, 1525.

Hyde, S.T. Pure Appl. Chem. 1992, 64, 1617.

Henriksson, U.; Blackmore, E.S.; Tiddy, G.J.T.; Soderman, O. J. Phys. Chem.
1992, 96, 3894.

Fuhhop, J-H.; Helfrich, W. Chem. Rev. 1993, 93, 1565.

Davis, M.E.; Chen, C.-Y.; Burkett, S.L.; Lobo, R.F. Mater. Res. Soc. Symp. Proc.
1994, 346, 831.

Pinnavaia, T.J.; Zhang, W. Stud. Surf. Sci. Catal. 1998, 117, 23 (Elsevier Science
B.V.).

Armengol, E.; Corma, A.; Garcia, H.; Primo, J. Appl. Catal. A: General 1995,
126, 391.

Huo, Q.; Leon, R.; Petroff, P.M.; Stucky, G.D. Science 1995, 268, 1324.

Gallis, K.W.; Landry, C.C. Chem. Mater. 1997, 9, 2035.

Flegler, S.L.; Heckman, Jr., J .W.; Klomparens, K.L. “Scanning and Transmission
Electron Microscopy- An Introduction ” 1993, (W .H. Freeman and Company),
New York.

Ryoo, R.; Ko, C.H.; Howe, R.F. Chem. Mater. 1997, 9, 1607.

Zhang, W.; Pauly, T.R.; Pinnavaia, T.J. Chem. Mater. 1997, 9, 2491.

Ryoo,R.; Jun, S.; Kim, J .M.; Kim, M.J. Chem. Commun. 1997, 2225.

Corma, A.; Fomes, V.; Navarro, M.T.; Perez—Pariente J. Catal. 1994, 148, 569.
Hitz, S.; Prins, R. J. Catal. 1997, 168, 194.

Mokaya, R.; Jones, W. Chem. Commun. 1996, 983 and J. Catal. 1997, 172, 211.
Zhang, W.; Liu, Y.; Pinnavaia, T.J. J. Phys. Chem. in preparation.

Armengol, E.; Cano, M.L.; Corina, A.; Garcia, H.; Navarro, M. Chem. Commun.
1995, 519.

Janicke, M.T.; Landry, C.C.; Christiansen, S.C.; Kumar, K.; Stucky, G.D.;
Chmelka, B.F. J. Am. Chem. Soc. 1998, 120, 6940.

148

Chapter 6

Incorporation of Structural Order into Mesoporous Silica Molecular
Sieves Assembled Via a Metal Salt Mediated N°(M*X')I° Pathway

Abstract

The concept of incorporating structural order into mesoporous silica molecular
sieves through electrostatic interactions between electrically neutral surfactant micelle
and neutral inorganic precursors has been successfully illustrated. Mesoporous silica
molecular sieves with well ordered domains consisted of a hybrid 3-dimensional
hexagonal and spherical cubic phases (3-d-hex-cubic), as well as an uni-dimensional
hexagonal structure and a wormhole structural motif were assembled using alkyl
polyethylene oxide surfactant as the structure director and tetraethyl orthosilicate as the
silicon source through a new counterion mediated N°(M”X')I° pathway controlled by
judicial choice of a surfactant to metal ratio, and also by choice of a metal salt from the
group of €002, MnCl2, NiCl2, CuCl2, ZnCl2, FeCl2, FeCl3, RhCl3, Cer and LiCl.
Structural variation was strategically mediated through surfactant-metal complex
formation that leads to variation of surfactant packing parameters (g = V/aol) and
ultimately the inorganic meso-structures. The 3-d-hex-cubic structure is favored by “2+”
metal cations such as Coz+, Ni2+ and Mn2+. Hexagonal structure is favored by both “1+”
and “2+” metal cations such as Li+, C02": Cu2+, Zn“, and Fe“. No structural order is
faVored by “3+” metal cations such as Cr”, Fe3+ and Rh“. It seems that 3-d-hex-cubic
structure formation prefers a lower interfacial charge density with a surfactant to metal
ratio in the range of 4 to 12. Hexagonal structure formation is favored by a relatively

higher interfacial charge density with a lower surfactant to metal ratio in the range of l to

149

3. Wormhole structure are obtained near zero interfacial charge density with a quite high

surfactant to metal ratio above 17.

6.1 Introduction
Hexagonal MCM-41 and cubic MCM-48 molecular sieves have been primarily

assembled via an electrostatic ST pathway relying on the lyotropic properties of the

3 Analogous hexagonal and cubic silica meso-

quatemary ammonium surfactants."
structures have been also prepared through liquid crystal templating using alkyl
polyethylene oxide surfactants in the presence of significant amounts of mineral acid. 4
Consistently, hexagonal mesoporous silica with pore size up to 300 A was recently
obtained using big triblock copolymer nonionic surfactants via an electrostatic S”X'I+
assembly mechanism in the presence of large amount of mineral acid.5 Mesoporous silica
and alumina molecular sieves with wormhole channel motif, however, were
cooperatively assembled via an electrically neutral N"Io pathway with H-bonding
interactions between nonionic surfactant (N°) and neutral inorganic reagent (1").6'8
Although the neutral assembly pathways possess environmentally benign advantages
over the electrostatic assembly pathways, our previous studies only obtained disordered
wormhole structures. Actually, as a consequence of weak H-bonding assembly, both S°I°

"0 and N°I° assembly lead to silica6 and alumina7 mesoporous molecular sieves with

9
wormhole channels. Nonetheless, chemical control of the surfactant assembly process for
formation of desired mesoporous molecular sieves is still a challenge. Structurally well

designed mesoporous molecular sieves are desirable for use as special sorbents, catalyst

supports and catalysts, functional ceramics, sensors and separation-membrane precursors.

150

Therefore, it would be chemically significant and scientifically interesting to produce
ordered meso-structures from neutral assembly by a facile method.

It has been known that the interfacial interactions between surfactant micelles and
inorganic precursors play an important role in directing the formation of meso-
structures.ll Longer ranged electrostatic interaction usually affords relatively higher
order of mesostructures in comparison with the shorter ranged H-bonding interaction that
leads to X-ray disordered but uniformly porous meso-structures.12 Our strategy to
incorporate higher structural order into mesoporous silica molecular sieves assembled
using nonionic surfactant and neutral inorganic precursors is based on partial coulomb
interactions at the interface between surfactant micelles and inorganic precursors without
altering the fundamental feature of H-bonding assembly. Polyethylene oxide has been
known to form helical complexes with Li+, Cu2+ and Zn2+.l3'15 In this work we examine
the capability of alkyl polyethylene oxide surfactants to complex metal cations as an
effective approach to incorporate coulomb forces at the interfaces between surfactant and
hydroxylated neutral silicon species to form relatively well ordered mesoporous silica
molecular sieves.

Here we report a versatile novel method to prepare mesoporous silica molecular
sieves. Mesostructures such as a hybrid structure consisted of 3-d hexagonal and
spherically cubic (3-d-hex-cubic), an uni-dimensional hexagonal, and a wormhole-like
meso-structures were synthesized through judicious choice of metal cations and control
of the surfactant to metal ratio. We also purposely allow the assembly to take place in a
neutral medium to maintain the H-bonding feature through a metal salt mediated neutral

assembly pathway (N°(M’X')I°). The strategy involves variation of surfactant molecular

151

packing parameter g (i.e. g = V/aol) '6 through formation of metal complex which varies
the effective head group area of the surfactant substantially to cause a phase change of

the assembled products.

6.2 Experimental

Structurally ordered mesoporous silica molecular sieves were synthesized using
tetraethylorthosilicate (TEOS) as the silicon source and nonionic surfactant including Brij
56 (C16H33(EO)10H)2 Tergitol (Crz-I6H25-33(EO)8-IOH) and Triton X-100
(C3H17Ph(EO)ioOH) as the structure directors in the presence of metal salts such as
CoCl2, NiCl2, MnCl2 and LiCl. The reactions were carried out at a neutral pH and a
temperature between 35 to 60 °C in a reciprocal thermal bath under very gently shaking
for 40 h. The product was filtered and dried in air. The obtained 3-d-hex-cubic and uni-
dimensional hexagonal mesoporous silicas were denoted as MSU-C and MSU-H,
respectively. The molar composition for MSU-C synthesis was

1008i : 2-7Surf : 0.2-0.8MX : 8500H2O.
MSU-C was only formed in the presence of “2+” metal salts especially in favor of €002,
NiCl2 and MnCl2. The molar composition for MSU-H preparation was
100Si : 2-7 Surf : l-7MX : 8500H2O.

MSU-H was assembled using the same reagents, under the same reaction conditions for
MSU-C, except for the surfactant to metal ratio. MSU-H was also formed using LiCl,
CuCl2 and ZnCl2 in place of CoCl2. As-made material was either refluxed in ethanol for 2

h or calcinded at 600 °C for 6 h to remove the surfactant from their mesopores.

152

The powder X-ray diffraction pattern of the samples was recorded on a Rigaku
Rotaflex Diffractometer using Cu K0 radiation (2.: 1.542 A). N2 adsorption and
desorption isotherms were measured on an ASAP 2000 sorptometer at -l96 °C. The
sample was evacuated under 10'5 torr at 150 °C over night before the measurement. 29Si
NMR spectra were recorded on a Varian VXR-400S spectrometer with a 7-mm zirconia
rotor, at a spin frequency of 4.2 kHz. TEM images were taken on a JEOL IOOCX with a
CeB6 gun that was Operating at an acceleration voltage of 120 kV. The specimen was
loaded onto a holey carbon film that was supported on a copper grid by dipping the grid

into a suspension of the solid sample in ethanol.

6.3 Results and Discussion
6.3.1 Materials Characterization

Figure 6.1 shows TEM images of MSU-C and MSU-H. MSU-C (top image)
exhibits cavity type pores with an entrance pore diameter of ~5.5 rim and inter-pore
distance of ~13.7 nm. MSU-H exhibits MCM-41-like uni-dimensional channels with a
diameter of about 4.0 nm. Selected area electron diffraction patterns (insets) confirms
uni-dimensional hexagonal structure for MSU-H. However, the diffraction patterns for
MSU-C is complicated. No explicit structure assignment could be achieved for this
diffraction pattern. To assign this diffraction pattern more accurately, a series of patterns
were generated using an image-processing program, NIH Image.

For the sake of comparison, the TEM image of MSU-C and its electron diffraction
pattern are shown in Figure 6.2a and 6.2b, respectively. The optic diffraction pattern 2c

obtained through Fourier-transformation of the image 28 shows two interdigitated

153

 

Figure 6.1. TEM images of calcined (600 °C) mesoporous silica molecular sieves
prepared in water using the alkyl polyethylene oxide type surfactant Brij 56
(C16H33(EO)10H) as the structure director and TEOS as the silicon source in the presence
of €002 salt as the structure promoter under the same reaction conditions (45 °C, 40 h
reaction time) except that the surfactant to Co2+ ratios are different. The surfactant to
silicon ratio was 1 to 24 for both samples, but the surfactant to metal ratio was 8.5 for the
sample denoted as MSU-C (top photo) and 2.0 for the sample denoted as MSU-H
(bottom photo). These images are provided as negatives for easier viewing of the pores as
black spots. Pore to pore distance is constantly ~13.7 nm through the specimen for MSU-
C. The insets are the electron diffraction patterns for areas imaged.

154

Figure 6.2.

fihfiiifi§Ot
*‘%*Q§
‘3"Qfifidt
##‘rfiitti
3i§ittaer

iUQ‘QQth
8Q.i"fitta
.‘Q‘Qfifi‘

 

(a). Selected area TEM image of MSU-C; (b). Electron diffraction pattern
of MSU-C; (c). Optic diffraction pattern obtained from the Fourier-
transformation of image (a); (d). Image obtained by reverse-Fourier-
transforrnation of the center tetragonal pattern in the optic diffraction
pattern (c); (e). Image obtained from the total reverse-Fourier-
transformation of optic diffraction pattern (c); (1). Image obtained from the
reverse-Fourier-transformation of 7 diffraction spots selected from pattern
(c) in such a way that these 7 spots resembling 7 spots in the real electron

diffraction pattern (b).

155

Figure 6.3.

 

A TEM segment for MSU-C showing a different orientation of the
structure. The optic diffraction pattern (obtained by Fourier-transformation
of arrow indicated area in the image) is attributable to both a (01—10)-3-d-
hexagonal orientation and a (210) cubic orientation. These patterns are
very similar to the diffraction pattern for meso—structured silica films with
pores running orthogonal to the film surface, as reported by Brinker et al.”
Notably, the pore to pore distance is ~13.7 nm.

156

hexagonal pattems and a central tetragonal pattern. This is indicative of a faulted hybrid
structure that combines both 3-dimentional hexagonal and cubic structural symmetries.
Reverse Fourier-transformation of the center tetragonal pattern gave rise to image 2d,
which is obviously different from original image 2a, but it shows the same inter-pore-
distance of 13.7nm as for cavity pores. Reverse Fourier-transformation of the whole optic
pattern 2c generated a nice cavity image 2e with cylindrical pores around the cavity-
shaped pores. This image is more like the observed image as shown in 2a. Surprisingly, if
we just reverse-fourier-transform 7 reflection spots chosen from the pattern in 2c in such
a way that they are arranged similar to the observed electron diffraction pattern 2b, the
resulted image (21) clearly show 6 cylindrical pores in a distorted hexagon around the
cavity-shaped pores. This is very much like the observed image 2a. Since only ideal 3-d
hexagonal and cubic hybrids can generate diffraction pattern as shown in 2c, image 2f
suggests that the obtained MSU-C is more like a distorted structural hybrid than a pure
phase. These TEM images are similar to those of a meso-structured silica film reported
by Brinker et al.'7 It seems that the hybrid meso—structure is dynamically evolving under
certain assembly conditions. Actually, we also observed a segment of our MSU—C sample
by TEM (Figure 6.3), which shows cavity pores arranged in distorted cubic fashion. The
optic diffraction pattern (from Fourier-transformation of the image section pointed by an
arrow) is consistent with either a (210)-cubic orientation or a (01-10)-3-d-hexagonal
orientation.

In addition to the hybrid ordering, there are disordered wormhole structural
domains that co-exist with the well ordered structural domains within the same particle

for both MSU-H and MSU-C. The disordered domains presumably resemble the ordered

157

counterparts in porous connectivity. Three-dimensional porous connectivity for the 3-d-
hex-cubic structure and uni-dimensional porous conductivity for the hexagonal structure
should be retained through the whole particle for MSU-C and MSU-H, respectively.

The X-ray powder diffraction patterns of MSU-H further confirms the hexagonal
feature by showing (100), (110) and (200) reflections (Figure 6.4). The XRD pattern for
MSU-C only exhibits a single reflection peak. The intense peak at low 29 angle exhibits
a d-spacing of 6.1 nm. By assuming this peak to be a cubic (210) reflection, the
calculated unit cell parameter is 13.7 nm which is consistent with the inter-pore-distance
observed by TEM for this sample (c.f. Figure 6.1). Lacking higher order reflections in the
X-ray, the diffraction pattern also is in consistent with the presence of disordered 3-d
wormhole pores.

N2 adsorption and desorption isotherms of MSU-C (Figure 6.5) indeed exhibit a
characteristic hysterisis loop for necked cavity pores. In reasonable agreement with the
TEM measurement, the effective pore size calculated using the Horvath-Kawazoe model
is 6.0 and 4.3 nm for MSU-C and MSU-H, respectively. The pore size for MSU-C is
about 2 times larger than the pore diameter of a MCM-48 (i.e. 2.5 nm)18 assembled using
cetyltrimethyl ammonium bromide which has the same alkyl chain length as Brij 56. A
high surface area of 1103 m2/g and a pore volume of 1.06 cc/g for MSU-C strongly
suggest that the over-all quality of this silica molecular sieve is comparable to a
bicontinuous cubic MCM-48.18 Similar to that of MCM-41, the isotherm for MSU-H

showing no hysterisis loop is consistent with uni-dimensional hexagonal pores.

158

Figure 6.4.

Intensity

 

NIM‘X‘) I °

N0 = 016113350) 10”
l° = TEOS
MIX. = (30612

 

MSU-C
INF/Co” = 8.5
MSU-H 7
Wee” = 2
(hkl) duplnm dale/um

100 6.14 6.14
110 3.53 3.54

L: 3.02 3.07
| |

7 I I I T f l I I I I

 

 

 

 

 

 

. . . , . .
0 2 4 6 8 10
29 (degrees)

X-ray diffraction patterns for the calcined MSU-C and MSU-H samples
that were used to obtain the TEM images shown in Figure 6.1. Both d-
spacings, observed and calculated according to an uni-dimensional

hexagonal phase for MSU-H, are provided for comparison.

159

 

      
    

E 800i I°=TEOS MSU-c
N°Ic6*=as

17.

65

B

3

HK = 4.3 nm
I~I°Ic62+ = 2

 

N Volume Adsorbed
.5
O
9

200

 

 

0 0.2 0.4 0.6 0.8 1

Figure 6.5. N2 adsorption and desorption isotherms for calcined MSU-C and MSU-H.
Pore necking hysterisis loop for MSU-C is indicated by the step at
desorption branch, which is ~0.2 unit of relative pressure behind the
adsorption branch. MSU-H exhibits almost identical adsorption and

desorption branches signify the uniformity of uni-dimensional pores.

160

 

Figure 6.6.

TEM images showing the wormhole (top) and 3-d-hex-cubic (lower)
meso-structures assembled at C02+ : Brij56 molar ratios of 1:13 and 1 : 8.5
respectively. TEOS was the silicon source and the silicon to surfactant

ratio was 17 to 1.

161

In accordance with previously reported MSU-X silicas, only a wormhole
structural motif was formed when little or no metal salt was used to assembly a
mesostructure at a quite high surfactant to metal ratio of 17 (Figure 6.6, top). However,
when this ratio was reduced to 8.5, a nice 3—d-hex—cubic structure was reproduced (Figure
6.6, bottom). The details of this marvelous image reveals a very high resemblance to a
spherical cubic surfactant meso-phase in space group of Im3m showing cavity type
linkages which has been observed previously.19

Figure 6.7 shows X-ray diffraction patterns of mesostructures formed at various
surfactant to Li+ ratios. Li+ was chosen because there are many reports on Li+-
polyethylene oxide complexes.”21 Relatively well ordered hexagonal silica with higher
29 angle reflections was formed at a Li+ISurf ratio of 1 to 2. Disordered silicas with
single reflection were formed at Li+/Surf ratio of 0 or greater than 3. Interestingly, no
cubic or hybrid meso-structure was obtained under any circumstances for Li“, K+ and
Na". The N2 adsorption and desorption isotherms (Figure 6.8) without hysterisis loops
confirm the presence of uni-dimensional hexagonal pores for the silica with hexagonal
reflections. In contrast, the N2 isotherms for the disordered silicas that exhibit a single X-
ray reflection show hysterisis loops indicating significant pore necking (cavity-shaped
worrnholes), and probably 3-dimensional pore connectivity.

As expected, the effect of polyethylene oxide complex formation with “3+” metal
cations are different from those observed for complexes with “1+” and “2+” metal ions.
The experiment performed for Fe”, Cr3+ and Rh3+ indeed showed very different results
from those with Fe“ and Co”. All of the silicas prepared in the presence of “3+” metal

cations exhibit wormhole structure regardless of the surfactant to metal ratio. Also, their

162

 

 

 

 

  
 
   
 
 
  
  

 

 

 

 

 

  

 

810‘,
710 T
610 ‘j Li’lBri]56=5
{a 510 i
a .I
3 .
g 410 i Ll’lBri156=2
: -
o .
l:- 3104; (th) duplnm dalclnm
* (100) 5.52 5.52
1 (110) 3.15 3.18
. (200) 2.77 2.75
210 : LIVBI'I]56=1
1 10 i Ll’lBri]56=0.5
j Ll’lBrl]56=0
OT'I'H'I'H'I'H'I'H'I

1 2 3 4 5 6 7 8
29 (degrees)

Figure 6.7. X-ray diffraction pattern of the calcined (600 °C) meso-structured silica
molecular sieves assembled using Brij 56 as the structure director and
TEOS as the silicon source at a silicon to surfactant ratio of 15 to 1 under
the same reaction conditions (45 °C, 40 h reaction time), except that the

surfactant to Li" ratio was varied as shown.

163

 

 
 
    
 
   
  

s = 836 m’lg
— BET
1000 4 HK = 4.8 nm
* Lr/Bri] 56 = 1
Ethanol Extracted

(cc/g, STP)

aoOj

      
     
  
 

_ 2
Sag—892mlg
HK=4.3nm

N Volume Adsorbed
at
O
‘1

HK = 5.5 nm

Li’lBri] 56: 0
Ethanol Extracted

 

 

II I IlfiIflfi I II

0 0.2 0.4 0.6 0.8 1
PIPo

Figure 6.8. N2 adsorption and desorption isotherms of as-made meso-structured silica

 

molecular sieves as described in Figure 6.7. The surfactant in the
mesopores was removed by ethanol extraction (i.e. refluxing in ethanol for
2 h) prior to the adsorption measurement. Again, the hexagonal silica
exhibits no hysterisis loop, as expected for uniform uni-dimensional pores,

whereas the wormhole silicas show pore necking hysterisis loops.

164

51d

MSU-X Calcined at 600 °C
N?M*x1l °
n° = r ItoI 15-s-12
M‘X‘ = Metal Chloride
is
310‘ g
«5
II

Intensity

zid‘

11047

Figure 6.9.

 

 

5d

 

 

 

 

 

 

 

 

 

0‘
0 5 10 15 20
29 (degrees)

X-ray diffraction patterns for calcined (600 °C) mesoporous silica
molecular sieves prepared using the mixed alkyl polyethylene oxide
surfactant Tergitol 15-S-12 (C12-16H25-33(EO)3-10H) as the structure director
and TEOS as the silicon source under the same reaction conditions (45 °C,
40 h reaction time and a surfactant to metal cation ratio of approximately
24 to 1), but different cation valences for the promotional metal salts. The
X-ray patterns show a decrease of d-spacing from 6.4 to 4.65 nm upon
replacing Co2+ and Mn“ with Cr“.

165

 

1 000

    

0
N1M*X‘)l o MSU X Calcined at 600 C
N° = Tergitol 15-S-12
.o = reos N°IMn2’ = as

800 .. MT = Metal Chloride

 

E

(0

:5

E 600"
‘6

ID

2
a, 400‘ o,
E

2

O

>

z 200'

 

 

 

Figure 6.10. N2 adsorption and desorption isotherms for mesoporous molecular sieves
described in Figure 6.9. The Cr3+ mediated structure shows a pore size 2.0
nm in comparison with a pore size of 3.3 nm for the structures mediated
by Co2+ and Mn“.

166

Figure 6.11.

 

2.5 urn

 

 

SEM (top) and TEM (bottom) images of mesoporous silica prepared using
Tergitol 15-S-12 in the presence of Cr3+ as a structure promoter, showing
micrometer sized intergrown spherical particles with flatten surfaces on

the spheres.

167

pore size is at least 7 A smaller than the silicas prepared in the presence of “1+” or “2+”
metal cations. For example, the mesoporous silica formed in the presence of Cr3+ exhibits
a much smaller d-spacing (i.e. 4.6 nm), in addition to smaller mesopores (~2.0 nm), in
comparison with the materials promoted by Mn“ and C021 (Figure 6.9 and 6.10).
Actually, all of the mesoporous molecular sieves formed with Rh“, Fe”, and 0'“ as
promoters exhibit smaller pores (~2.0 nm). These same metal cations promote the
formation of intergrown spherical particles, many with flatten surfaces (Figure 6.11).
Notably, these mesoporous molecular sieves were synthesized under the same reaction
conditions using Tergitol 15-S-12 (a mixture of C12 to C16 alkyl polyethylene oxide
surfactant) as the structure director. The different pore structure and morphologies
obtained for "3+" and "2+" cations strongly suggest that metal complex formation play an

essential role in determining the phase and pore connectivity in the final meso-structure.

6.3.2 Discussion

The possibility that metal cations may be incorporated into the silica framework is
excluded, because the transition metal along with the surfactant can be easily removed by
ethanol extraction. Chemical analysis confirms that no residual metal is detectable for
ethanol extracted samples. Therefore, first, the metal cations are very likely complexed
by surfactant head groups. Second, the assembly process in the presence of metal salt
maintains neutral H-bonding interfacial interactions that are weak enough to allow
removal of surfactant by solvent extraction. Consequently, MSU-C and MSU—H are very

likely assembled through a new metal complex mediated N°(M*X')I° mechanism.

168

Figure 6.12 shows 29Si NMR spectrum for as-made MSU-H, a representative
silica mesostructure prepared by a metal-salt mediated pathway. Typical 29 Si NMR
spectra for HMS obtained by S°I° assembly and two MCM-4l samples obtained by ST

and S+XT’ pathways are also included for comparison. The spectra for both MSU-H and

  
  

HMS (S°I°)

MSU-H (N‘lM’X) 1°)

  
 

MGM-41 (560*)

 
 
 

MGM-41 ($1)

  

 

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

-40 -60 -80 -100 -120 -140 -160ppm

Figure 6.12. 29Si NMR spectra for as-made HMS(S°I°), MSU-H, MCM-41(S"X‘1") and

MCM-41(S+l'). MSU-H was synthesized using Brij 56 as the structure

director in the presence of CoCl2. HMS and MCM-41s were prepared as

reported.”4

169

 

 

 

Scheme 1

170

MCM-41 (S+X'F) exhibit three well resolved resonance bands at -92, -100, and -110 ppm
representing 2Q, 3Q and 4Q silicon cross-linking connectivity. Apparently, the spectrum
for MSU-H is very much like the spectrum for MCM-4l formed though the counterion
mediated S+X'1" pathway, but different from either HMS (S°I°) or MCM-41(S’T). This
resemblance strongly supports that MSU-H was assembled through a new counterion-
mediated N°(M*X')I° pathway.

By forming complexes with metal cations, we incorporate longer ranged coulomb
interactions into the surfactant micelles that cooperatively assemble with silica species to
form either ordered uni-dimensional hexagonal or 3—d-hex-cubic structures. As illustrated
in Scheme 1, small Li+ likely form intra-molecular complex with polyethylene oxide head
group in a helical conformation.l3 The rod-like micelles formed from this surfactant-
metal complex should be more rigid and easily aligned into a long-range hexagonal
inorganic-organic mesophase with the help of interfacial coulomb interactions between
the micelles and inorganic precursors. However, for larger Co”, which is well-known for
adopting versatile coordination numbers, intra— or inter-surfactant molecular complexes
are likely to form upon changing the surfactant to metal ratio (SIM). The formation of
helical intra-molecular surfactant-metal complex may be more favorable as a result of
positive charge repulsion at a lower SIM ratio (i.e. l/l). As a consequence of competitive
complexation, inter-molecular surfactant-metal complex may be more favorable at a
higher S/M ratios (i.e. 8.5/1). In the case of inter-molecular complex formation, a larger

effective head group area (a0) would result a significant decrease in surfactant micelle

packing parameter g to about 1/3 (g = V/aol). Therefore, in forming a thermodynamically

171

favored product, spherical micelles (at low surfactant concentration, i.e. 0.027 M) could
pack into a body centered Im3m cubic meso-structure.22'23 Noticeably, an ideal Im3m
cubic structure is fundamentally different (c.f. Scheme 1) from bicontinuous Ia3d cubic
MCM-48 and Pm3n cubic SBA—l assembled using cetyltriethyl ammonium bromide via
electrostatic 8”)!“+ pathway.23 In the case of intra-molecular complex formation,
surfactant micelles with a relatively smaller head group area and larger micelle packing
parameter g at about 1/2, rod-like surfactant-complex micelles may pack cooperatively
into hexagonal meso—structures.

Meso-structure formation is known to be driven by non-polar and polar
interactions in a medium with suitable polarity. Previous work has revealed that meso-
structure phases depended on the lyotropic properties of the surfactant.3 Under the
cooperative assembly conditions used in the present work, the initial surfactant
concentration was far less than needed to form liquid crystal phases.24 However, the
surfactant molecules were eventually occluded in the framework pores as if they were in
liquid crystal phases. The structure of the assembled product was primarily determined
by the surfactant structure and reaction conditionsf"23 For example, cubic MCM-48 was
synthesized at higher surfactant concentrations than hexagonal MCM-4l.3 3-d hexagonal
SBA-2 was synthesized using double headed quaternary ammonium gemini surfactant
rather than single headed surfactant.” Similarly, hexagonal, 3-d hexagonal and cubic
mesoporous silicas were made using different polyethylene and polypropylene oxides
block copolymer surfactants in the presence of concentrated mineral acids.26 And worm-
68

hole like mesoporous silicas were made using nonionic surfactant in neutral medium.

Obviously, these strategies focused on changes of surfactant structure or surfactant

172

 

concentration or co-surfactant ”'28

to change the packing parameter (g) and this
effectively varied the phase of inorganic meso-structure. Our approach by changing the
surfactant packing parameter through chemical complex formation between surfactant
head group and metal cations that ultimately change the assembled meso-structure is
new. Although the possibility of forming complexes between polyethylene oxide
surfactant and transition metal salt has been demonstrated for the formation of meso-
structured transition metal oxides,29 the metal salt has not been found to be influential in
affecting the structure of the final product. In our synthesis approach, however, the metal
cations were used just as structure promoters rather than a framework precursor. In

addition, none of the reported meso—structures prepared using nonionic surfactants have

exhibited a TEM image similar to ours for the hybrid 3-d-hex-cubic structure.

6.4 Conclusion

Modification of polyethylene oxide surfactant head groups through complexation
with transition and alkaline metal cations was carried out for silica mesoprous molecular
sieve synthesis. The variation in meso-structure upon variation of the valence of metal
cations was quite dramatic. Mesoporous silica molecular sieves with either ordered 3-d-
hex-cubic or uni-dimensional hexagonal domains were synthesized by judicious choice of
metal cations. The 3-d-hex-cubic structure was favored by “2+” metal cations such as
Co“, Ni2+ and Mn”. The uni-dimensional hexagonal structure were favored by both
“1+” and “2+” metal cations such as Li", C021, Cu2+, Zn“, and Fe“. Wormhole structure
were favored by “3+” metal cations such as Cr“, Fe3+ and Rh“. It seems that the 3-d-
hex-cubic structure prefers a lower interfacial charge density with a surfactant to metal

ratio in the range of 4 to 12. The uni-dimensional hexagonal structure favors a relatively

173

higher interfacial charge density with a lower surfactant to metal ratio in the range of 1 to
3. The wormhole structure favors a zero or low interfacial charge density with a quite

high surfactant to metal ratio, above 17.

6.5 Reference

1. Kresge, C.T.; Leonowicz, M.E.; Roth, W.J.; Vartuli, J.C.; Beck, J .8. Nature 1992,
359, 710.

2. Beck, J.S.; Vartuli, J.C.; Roth, W.J.; Leonowicz, M.E.; Kresge, C.T.; Schmitt,
K.D.; Chu, C. T.-W.; Olson, D.H.; Sheppard, B.W. J. Am. Chem. Soc. 1992,
114, 10834.

3. Vartuli, J.C.; Schitt, K.D.; Kresge, C.T.; Roth, W.J.; Leonowicz, M.E.; McCullen,
S.B.; Hellring, S.D.; Beck, J.S.; Schlenker, J.L; Olson, D.H.; Sheppard, B.W.
Chem. Mater. 1994, 6, 2317.

4. Attard, G.S.; Glyde, J .C.; Goltner, C.G. Nature 1995, 378, 366.
Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G.H.; Chmelka, B.F.;
Stucky, G.D. Science 1998, 279, 548.

6. Bagshaw, S.A.; Prouzet, E.; Pinnavaia, T.J. Science 1995, 269, 1242.

7. a. Bagshaw, S.A.; Pinnavaia, T.J. Angew. Chem. 1996, 108, 1108; Angew. Chem.
Int. Ed. Eng. 1996, 35, 1106. b. Zhang, W.; Pinnavaia, T.J. Chem. Commun.
1998, 1185.

8. Prouzet, E.; Pinnavaia, T.J. Angew. Chem. 1997, 109, 533; Angew. Chem. Int. Ed.
Eng. 1997, 36, 516.

9. Tanev, P.T.; Pinnavaia, T.J. Science 1995, 267, 865.

10. Zhang, W.; Pauly, T.R.; Pinnavaia, T.J. Chem. Mater. 1997, 9, 2491.

11. Behrens, P. Angew. Chem. Int. Ed. Engl. 1996, 35, 515.

12. Zhang, W. “Ph.D. Dissertation in Michigan State University”, Chapter 1, 1999.

13. Armand, MB. in Polymer Electrolyte Review - I (Eds.: J .R. MacCallum, C.A.
Vincent), Elsevier Applied Science, London and New York, 1987, p l.

174

14.

15.
16.

17.

18.
19.
20.
21.
22.
23.
24.

25.
26.

27.
28.
29.

Jr. Bailey, F.E.; Koleske, J.V. in Alkylene Oxides and Their Polymers, Marcel
Dekker, New York, 1990.

McBreen, J.M.; Yang, X.-Q.; Lee, H.—S. J. Electrochem. Soc. 1996, 143, 3198.
Israelachvili, J .N .; Mitchell, D.J.; Ninham, B.W. J. Chem. Soc., Faraday Trans. 2
1976, 72, 1525.

Lu, Y.; Ganguli, R.; Drewien, C.A.; Anderson, M.T.; Brinker, C.J.; Gong, W.;
Guo, Y.; Soyez, H.; Dunn, B.; Huang, M.H.; Zink, J .1. Nature 1997, 389, 364.
Zhang, w.; Pinnavaia, T.J. Catal. Lett. 1996, 38, 261.

Ober, C.K.; Wegner, G. Adv. Mater. 1997, 9, l7.

Striebel, K.A.; Wen, S.J.; Ghantous, DJ. J. Electrochem. Soc. 1997, 144, 1680.
Tajitsu, Y. J. Mater. Sci. 1996, 31, 2081.

Hyde, S.T. Pure & Appl. Chem. 1992, 64, 1617.

Huo, Q.; Margolese, D.I.; Stucky, G.D. Chem. Mater. 1996, 8, 1147.

Huo, Q.; Margolese, D. 1.; Ciesla, U.; Demuth, D. G.; Feng, P.; Gier, T. E.;
Sieger, P.; Firouzi, A.; Chmelka, B. F.; Schiith, F.; Stucky, G. D. Chem. Mater.
1994, 6 , 1176.

Huo, Q.; Leon, R.; Petroff, P.M.; Stucky, G.D. Science 1995, 268, 1324.

Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B.F.; Stucky, G.D. J. Am. Chem. Soc.
1998, 120, 6024.

Gallis, K.W.; Landry, C.C. Chem. Mater. 1997, 9, 2035.

Chen, H.; Huang, L.; Li, Q. Chem. Mater. 1997, 9, 2685.

Yang, P.; Zhao, D.; Margolese, D.I.; Chmelka, B.F.; Stucky, G.D. Nature, 1998,
396, 152.

175