.23.?
>1
.sfifi
.mvfimm flufim
3,. J. v. .a .
,.J3a .. . ..
. gm .wmm‘...

.1

[3.51.11 .11
firming ;
3%
4A
;
x

..‘..... .. ‘
....x:‘v.1~sll . 2.»:-

 

 

This is to certify that the

dissertation entitled

Surfactant-Mediated Assembly of Crystalline
Mesoporous Aluminas: Synthesis, Characterization,
and Application in Hydrodesulfurization

presented by

Randall Wayne Hicks

has been accepted towards fulfillment
of the requirements for

Ph.D. . Chemistry
degree in

fl] P

 

 

J l 2 2002
Date uy ’

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-1277l

 

LIBRARY
Michigan 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6101 c-JCIRC/DateDuo.p65-p.15

 

SURFACTANT-MEDIATED ASSEMBLY OF CRYSTALLINE MESOPOROUS
ALUMINAS: SYNTHESIS, CHARACTERIZATION, AND APPLICATION IN

HYDRODESULFURIZATION

By

Randall Wayne Hicks

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2002

ABSTRACT
SURFACTANT-MEDIATED ASSEMBLY OF CRYSTALLINE
MESOPOROUS ALUMINAS: SYNTHESIS, CHARACTERIZATION, AND
APPLICATION IN HYDRODESULFURIZATION
By

Randall Wayne Hicks

Aluminas are utilized in many industrial applications, including as adsorbents,
abrasives, ceramics, catalysts, and catalyst supports‘. Many different phases of
aluminas exist, but due to its favorable combination of surface and textural
properties, the most important phase is y-alumina. Surface areas and pore
volumes of conventional aluminas are typically less than 250 m2/g and 0.5 cc/g,
respectively. Performance in catalytic applications is limited in part by these
properties.

As has been shown in Silica chemistry, the use of surfactants to aid in the
assembly of a mesostructure leads to improvement in textural properties? Similar
advances are anticipated in alumina chemistry, but to date, little progress has
been made in this area. Since the disclosure of mesostructured aluminas in
19963, all but a couple of subsequent reports have described compositions of
mesostructured aluminas with amorphous walls. This limits their thermal and
hydrothermal stability, and thus their potential use in catalytic applications.

The present work describes the synthesis of the first members of a new family
of crystalline mesostructured aluminas prepared from the hydrolysis of aluminum

sec-butoxide in the presence of either polyethylene oxide (PEO) or amine

surfactants as porogens. Members of this family exhibiting the y-AI203 phase,
denoted MSU-y aluminas, have surface areas, pore Sizes, and pore volumes in
excess of 400 mz/g, 15 nm, and 1.5 cc/g, respectively‘. In addition, forms of
these aluminas have expressed thermal and hydrothermal stability.
Mesostructured boehmiteS, MSU-B aluminas, have also been obtained using
Similar synthetic methods. In the synthesis of either alumina, the key step is the
formation of an MSU-S/B surfactant-boehmite precursor, which is converted to
the desired final product through calcination.

To demonstrate their usefulness, MSU-y aluminas have been utilized as
catalyst supports in dibenzothiophene hydrodesulfurization reactions.
Conversions and selectivities superior to that of a commercial catalyst have been
achieved. Their success in this capacity can be traced to the stability of the y-
AI203 phase, the ability to effectively disperse the active catalytic component, and

increased accessibility to the active sites.

References

(1) Misra, C. American Chemical Society Monograph, Vol. 184: Industrial
Alumina Chemicals, 1986.

(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, E. W.; et al. J. Am.
Chem. Soc. 1992, 114, 10834-10843.

(3) BagShaw, S. A.; Pinnavaia, T. J. Angew. Chem, Int. Ed. Engl. 1996, 35,
1102-1105.

(4) Zhang, Z.; Hicks, R. W.; Pauly, T. R.; Pinnavaia, T. J. J. Am. Chem. Soc.
2002, 124, 1592-1593.

To my wife, Jennifer, for without her sacrifice this work would not be possible.

ACKNOWLEDGEMENTS

There are many people to thank for their contributions toward the completion
of my doctoral degree during my time at MSU. First, I express gratitude to my
research advisor, Dr. Thomas J. Pinnavaia, for his support, both intellectual and
financial, over the last six years. There are two specific items that need mention.
First, thanks for encouraging me to stay in the Ph. D. program when I was
prepared to leave with a master’s degree. I’m glad that I stayed. Second, thanks
for not objecting to my desire to teach. I could have exclusively conducted
research, but I feel that I’m better for having done both, even if it meant a longer
tenure at MSU. So thank you for letting me do it my way.

Thanks to the department, especially DrS. Paul Hunter and Kathy Hunt, for
allowing me to teach over the last year. I hope that the department found it as
good an experience as I did. My success in the job market this year was largely a
consequence of that opportunity.

Thanks to Chris Marshall at Argonne National Laboratory for the generous
donation of time and facilities. The data obtained from the catalytic testing
performed in his labs was crucial to the completion of this work.

Thanks to former Pinnavaia group members Peter Tanev for introducing me
to aluminas and Tom Pauly for valuable discussions concerning my project.
Thanks to Zhaorong Zhang for our collaborations over the last 2+ years. Having
also worked on it, you’re the only other person that truly knows how frustrating

this project has been. To the rest of the current group members, thanks for

making lab life more enjoyable. You can now all feel free to steal my lab
equipment and mess up my bench as much as you’d like.

Thanks to three friends that I met at MSU, Pete LeBaron, Paul Szalay, and
Bill Scanlon. Graduate school would have been nearly intolerable without the
good times that we had over the years.

Thanks to my family for their support, even if they weren’t sure exactly what I
was doing here. To all my parents, Charles (Linda) Hicks and Neva (Ron) Penix,
thanks for preparing me for this experience. Thanks too, to my mother-in-law,
Leone Beck, who has always provided words of encouragement.

Thanks to my brother, Ryan, and my other friends back home, Rick & Jess,
Mark & Shelley, Vern, Frog, and John for support and much needed stress relief.
I only wish that we had more time to spend together Since I’ve been here.

Most importantly, thanks to my wife and my buddy, Jennifer. I am eternally
grateful for the love and support that you have given me. I can’t express here in
words what that has meant. We’ve both had to make sacrifices so that I could
finish what I started, more so for you, particularly over this last year. But it will all

be worthwhile in the end. And soon, it‘wiIl be our time again.

vi

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................... x
LIST OF FIGURES ........................................................................................ xi
LIST OF ABBREVIATIONS ........................................................................... xix
Chapter 1: Introduction .............................................................................. 1
1.1. Classification of Aluminas ........................................................... 1
1.2. Transition Aluminas ..................................................................... 5
1.3. Hydrodesulfurization .................................................................... 9
1 .4. Mesostructures ............................................................................ 18
1.5. Mesostructured Aluminas ............................................................ 21
1.6. Stabilization of Aluminas ............................................................. 23
1.7. Objectives and Rationale ............................................................ 26
1.8. References .................................................................................. 28

Chapter 2: Synthesis and Characterization of Mesostructured

Boehmite ............................................................................... 31

2.1 . Introduction .................................................................................. 31
2.2. Experimental ............................................................................... 34
2.2.1. Synthesis ....................................................................... 34

2.2.2. Characterization ............................................................ 35

2.3. Results and Discussion ............................................................... 36

vii

2.3.1. Formation of MSU-S/B .................................................. 36

2.3.2. General Properties of MSU-B Aluminas ........................ 39
2.3.3. Peptization ..................................................................... 54
2.4. Mechanistic Aspects of MSU-B Alumina Synthesis .................... 55
2.5. Conclusions ................................................................................. 58
2.6. Future Directions ......................................................................... 59
2.7. References .................................................................................. 61

Chapter 3: Synthesis and Characterization of Mesostructured y-

AI203 ........................................................................................ 63
3.1. Introduction .................................................................................. 63
3.2. Experimental ............................................................................... 65
3.2.1 . Synthesis ....................................................................... 65
3.2.2. Characterization of MSU-ry Aluminas ............................. 67
3.3 Results and Discussion ................................................................ 68
3.3.1. MSUq Aluminas from PEO Surfactants ........................ 68
3.3.2. MSU-y Prepared from Amine Surfactants ...................... 85
3.3.3 Stability of MSUJY Aluminas ............................................ 93
3.4. Conclusions ................................................................................. 103
3.5. Future Directions ......................................................................... 105
3.6. References .................................................................................. 106

viii

Chapter 4: Hydrodesulfurization Over MSU-y Supported Catalysts ....... 108

4.1. Introduction .................................................................................. 108
4.2. Experimental ............................................................................... 1 10
4.2.1. Sample Preparation ....................................................... 110
4.2.2. Reactor Design and Experimental Conditions ............... 113
4.2.3. Characterization ............................................................ 1 15
4.3. Results and Discussion ............................................................... 116
4.3.1. Catalytic Reactions ........................................................ 116

4.3.2. Structural and Textural Properties of MSUJY Catalysts. 119

4.3.3. Relationships Between Catalyst Structure

and Performance .......................................................... 128

4.4. Conclusions ................................................................................. 139

4.5. Future Directions ......................................................................... 139

4.6. References .................................................................................. 142
Chapter 5: Concluding Remarks ............................................................... 144

Table 2.1.

Table 3.1 .

Table 3.2.

Table 3.3.

Table 3.4.

Table 3.5.

Table 4.1 .

Table 4.2.

LIST OF TABLES

Synthetic parameters and textural properties of MSU-B
Aluminas ..................................................................................

Synthetic parameters and textural properties of MSU-y
aluminas prepared from aluminum sec-butoxide and PEO
surfactants ...............................................................................

Comparison of textural properties of calcined MSUJY formed
at 500 °C and calcined MSU-6 alumina formed at 800
°C .............................................................................................

Textural properties of MSU-y aluminas prepared through the
calcination of MSU-S/B surfactant-boehmite compositions
and MSU-B boehmites at 500 °C ............................................

Textural properties of doped MSU-y aluminas prepared from
aluminum sec-butoxide and P84. In each synthesis, the
dopant was administered as a metal salt which was
dissolved in the surfactant solution prior to addition of the
aluminum sec-butoxide ............................................................

Textural properties of MSU-y aluminas prior to and after
treatment at 600 °C in a 20% steam atmosphere for 5 hr .......

Results of HDS reactions over MSUJy catalysts exhibiting
distinctly different morphologies ..............................................

Textural properties of MSU-y alumina HDS catalysts ..............

44

72

87

92

94

104

117

127

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 2.1 .

Figure 2.2.

Figure 2.3.

Figure 2.4.

Figure 2.5.

LIST OF FIGURES

Family of alumina compounds .................................................

Transformation sequence observed upon the thermal
dehydration of aluminas (adapted from Wefers and Misra) .....

Activity of transition metal sulfides in DBT HDS plotted as a
function of periodic position of the metal (adapted from
Pecoraro and Chianelli) ...........................................................

Cross-sectional (upper) and three-dimensional (lower)
depiction of the boehmite double-layer. The thickness of the
double layer is ~ 12 A ..............................................................

Schematic of interactions leading to formation of MSU-S/B.
Each slab represents one boehmite double layer. Lines on
the boehmite layers represent hydroxyls forming hydrogen
bonds to the surfactant micelles. Areas of overlapping
boehmite layers indicate cross-linking through condensation
of adjacent hydroxyls ...............................................................

Representative low angle XRD patterns for the as-
synthesized surfactant-boehmite composite, denoted MSU-
S/B, prepared from aluminum sec-butoxide and
dodecylamine in an alcoholic/aqueous medium as described
by line 3 in Table 2.1 and the surfactant-free analog, MSU-B,
obtained by calcination of the MSU-S/B composite at 325 °C
for 10 hr ...................................................................................

Representative wide angle XRD patterns for MSU-S/B and
MSU-B aluminas Shown in Figure 2.3. Crystallographic
planes contributing to the peaks are noted ..............................

Nitrogen adsorption — desorption isotherms and BJH
adsorption pore size distributions (inset) for MSU-B aluminas
differing in surfactant concentration. The percentage given
corresponds to the mass percent of dodecylamine in the
surfactant solution. The three isotherms, offset by 200 cc/g
for clarity, correspond from top to bottom to samples
described in lines 3, 8, and 19 in Table 2.1. PSD curves are
displayed in the same order as the isotherms .........................

xi

4

6

12

33

38

40

41

43

Figure 2.6.

Figure 2.7.

Figure 2.8.

Figure 2.9.

Figure 2.10.

Nitrogen adsorption - desorption isotherms and BJH
adsorption pore Size distributions (inset) for MSU-B aluminas
prepared from aluminum sec-butoxide and dodecylamine at
ambient temperature in alcoholic/aqueous solvents of
varying ethanol content. The three isotherms, offset by 200
00/9 for clarity, correspond to MSU-B aluminas described in
lines 1-3 of Table 2.1. PSD curves are displayed in the same
order as the isotherms .............................................................

Nitrogen adsorption — desorption isotherms and BJH
adsorption pore Size distributions (inset) for MSU-B aluminas
prepared from aluminum sec-butoxide and dodecylamine at
different temperatures. The four isotherms, offset by 200
cc/g for clarity, correspond to MSU-B aluminas described in
lines 3-6 of Table 2.1. PSD curves are displayed in the same
order as the isotherms .............................................................

Nitrogen adsorption - desorption isotherms and BJH
adsorption pore size distributions (inset) for MSU-B aluminas
prepared from the hydrolysis of aluminum sec-butoxide in
the presence of different amine surfactants as porogens.
The Six isotherms, offset by 200 cc/g for clarity, correspond
to MSU-B aluminas described in lines 7-12 of Table 2.1.
PSD curves are displayed in the same order as the
isotherms .................................................................................

Bright field TEM images (A and C) and electron diffraction
patterns (B and D) of MSU-B materials prepared from
aluminum sec-butoxide and dodecylamine as described in
Table 2.1, (A) line 8 and (B) line 3. The open framework
results from the random orientations of fundamental
particles. Electron diffraction patterns, (B) and (D),
corresponding to bright field images (A) and (C), respectively
Show diffuse rings, Signifying the polycrystalline nature of the
material. The reflections contributing to the pattern are
labeled in (D) ...........................................................................

27Al NMR Spectrum of MSU-B prepared from aluminum sec-
butoxide and dodecylamine prepared as described in Table
2.1, line 3. The resonance at 7 ppm is consistent with the
structure of boehmite, in which all aluminum ions are in
octahedral environments .........................................................

xii

46

47

49

52

53

Figure 2.11.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Wide angle XRD patterns of MSU-S/B prepared from
aluminum sec-butoxide and dodecylamine at 20 °C Showing
evolution of the boehmite phase over time. The presence of
weak reflections assignable to boehmite after only 1 minute
Show that the formation of MSU-B does not proceed through
an MSU-X type of intermediate with amorphous framework
walls .........................................................................................

Representative wide angle powder XRD patterns of an MSU-
S/B surfactant-boehmite composition made from aluminum
sec-butoxide and the PEO surfactant P84 under reaction
conditions described in Table 3.1, line 3 and the
corresponding MSUJy alumina prepared from calcination of
the as-synthesized MSU-S/B at 500 °C for 4 hr ......................

Nitrogen adsorption-desorption isotherm and BJH adsorption
pore Size distribution (inset) of an MSU-Iy alumina prepared
from aluminum sec-butoxide and the PEO surfactant P84 as
described in Table 3.1, line 3. Included for comparison is the
same data for an alumina prepared without surfactant as
described in Table 3.1, line 20. lsotherms are offset 200 cc/g
for clarity and PSD curves are displayed in the same order
as the isotherms ......................................................................

27Al NMR of calcined (500 °C) MSU-y alumina prepared from
aluminum sec-butoxide and P84 as described in Table 3.1,
line 3. The two resonances at 6.7 ppm and 72 ppm are
assigned to aluminum in octahedral and tetrahedral
coordination, respectively. The peaks are in the approximate
75:25 ratio consistent with y-AI203 ...........................................

Bright field TEM image of (A) calcined (500 °C) MSU-y
alumina prepared from aluminum sec-butoxide and P84 as
described in Table 3.1, line 3. (B) Lower magnification image
of the same product and the corresponding electron
diffraction pattern, with diffuse rings assignable to the [311],
[400], and [440] crystallographic planes. (C) Dark field image
of the same particles as in (B), but formed from an ~79 arc of
the [440] diffraction ring. The crystal orientations contributing
to the [440] ring appear aS bright sports in the image .............

xiii

57

69

71

74

76

Figure 3.5.

Figure 3.6.

Figure 3.7.

Figure 3.8.

Figure 3.9.

Nitrogen adsorption-desorption isotherms of MSU-y aluminas
prepared from aluminum sec-butoxide and P84 as described
in Table 3.1, lines 3 and 6-10 showing the effect of solvent
polarity on the textural properties. The volume percent of 2-
butanol in the solvent is labeled for each isotherm. Isotherms
are offset 300 cc/g for clarity. The BJH adsorption pore size
distributions (inset) are in the same order as the isotherms...

Nitrogen adsorption-desorption isotherms and BJH
adsorption pore size distributions (inset) for MSUJY aluminas
prepared from aluminum sec-butoxide and P84 as described
in Table 3.1, lines 3 and 11-14 showing the effect of
surfactant concentration (mass P84 / mass solvent) on the
textural properties. The mass percent of P84 is labeled for
each isotherm. Isotherms are offset 300 cc/g for clarity. The
PSD curves are displayed in the same order as the
corresponding isotherms .........................................................

Nitrogen adsorption-desorption isotherms and BJH
adsorption pore Size distributions (inset) for MSU-ry aluminas
prepared from aluminum sec-butoxide and differing PEO
surfactants as described in Table 3.1, lines 3 and 11-14.
Isotherms are offset 200 cc/g for clarity. The PSD curves are
displayed in the same order as the corresponding isotherms.

Nitrogen adsorption-desorption isotherm and wide angle
powder XRD pattern (inset) of a calcined (800 °C) MSU-6.
The initial MSU-S/B surfactant-boehmite intermediate was
prepared from aluminum sec-butoxide and the PEO
surfactant P84 under conditions described in Table 3.1, line
13 .............................................................................................

TEM image of MSU-y prepared through the calcination of
MSU-B boehmite at 500 °C for 4 hr. The initial MSU-S/B
surfactant-boehmite composition was prepared from
aluminum sec-butoxide and dodecylamine (DDA) at a
reaction stoichiometry of AlzDDAzEtOHzH20 = 5:1:68.0:73.4
and a reaction temperature of 20 °C. The surfactant free
MSU-B boehmite phase was obtained by calcining the MSU-
S/B precursor at 325 °C for 10 hr. The electron diffraction
pattern (inset) demonstrates the crystallinity of the
framework of the mesostructure, as indicated by the diffuse
rings attributable to the [311], [400], and [440]
crystallographic planes of Y-AIan ............................................

xiv

80

82

83

86

9O

Figure 3.10.

Figure 3.11.

Figure 3.12.

Figure 3.13.

Figure 3.1 4.

Nitrogen adsorption-desorption isotherms and BJH
adsorption pore size distributions (inset) for MSU-y aluminas
prepared through the calcination (500 °C) of amine-mediated
MSU-S/B surfactant-boehmite precursors (samples A and C)
and surfactant free MSU-B precursors (samples B and D).
The AlzDDAzEtOHzH20 reaction stoichiometry was
5:1 :68.0:73.4 for all samples. Reaction temperature was 20
°C for samples A and B and 45 °C for samples C and D.
Isotherms are offset 300 cc/g for clarity. The PSD curves are
displayed in the same order as the corresponding isotherms.

Wide angle powder XRD patterns of MSU-y aluminas
prepared from aluminum sec-butoxide and P84: pristine and
doped with either lanthanum or cerium. Peaks marked with
an asterisk are attributable to a cerium aluminate phase,
CeAl03. The doped MSU-y aluminas were prepared through
the calcination (500 °C) of doped MSU-S/B precursors
prepared as described in Table 3.1, line 13 ............................

Nitrogen adsorption-desorption isotherms and BJH
adsorption pore Size distributions (inset) of MSUJY aluminas:
pristine and doped with either lanthanum or cerium. The
doped MSUJY aluminas were prepared through the
calcination (500 °C) of doped MSU-S/B precursors prepared
as described in Table 3.1, line 13. lsotherrns are offset 400
cc/g for clarity. The PSD curves are in the same order as the
corresponding isotherms .........................................................

TEM images of MSU-y aluminas prior to (A,C,E) and after
steaming (B,D,F) at 600 °C for 5 hr in a 20% steam
atmosphere. Images A and B are for pristine MSUq. Images
C, D and E, F are for MSUq doped with 5% La3+ and 5%
Ce‘“, respectively. The MSUJY aluminas were prepared as
described in the caption to Figure 3.11. The scale bar
represents 100 nm in all figures ..............................................

Nitrogen adsorption-desorption isotherms and BJH
adsorption pore size distributions (inset) of pristine and
doped MSU-y aluminas after steaming for 5 hr at 600 °C in a
20% steam atmosphere. Samples were prepared as
described in the caption to Figure 3.11. Isotherms are offset
300 cc/g for clarity. The PSD curves are in the same order

as the corresponding isotherm ................................................ .

91

96

97

100

101

Figure 3.15.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

Wide angle powder XRD patterns of pristine and doped
MSU-y aluminas after steaming for 5 hr at 600 °C in a 20%
steam atmosphere. Peaks marked with an asterisk are
attributed to CeAlOa, cerium aluminate. Samples were
prepared as described in the caption to Figure 3.11 ...............

Schematic of the HDS reactor employed in the study. The
liquid feed of dibenzothiophene in hexadecane was
introduced to the gas feed, a mixture of hydrogen and
nitrogen. The two leads are passed through the preheater
(set to 350 °C), where all feed is vaporized. Then the feed
reactants are passed over the catalyst bed at 400 °C. Liquid
products are condensed and collected while gas products
are scrubbed in 2M NaOH prior to venting ..............................

Reaction pathways of DBT in HDS reactions. The upper
pathway involves direct desulfurization leading to the
formation of biphenyl (BP). The lower pathway involves a
hydrogenation step prior to desulfurization, leading to the
formation of cyclohexylbenzene (CHB) ...................................

Wide angle XRD patterns for lath-12 MSUJY alumina taken at
various stages of the catalytic testing process. The y-Al203
phase is evident throughout the process and the lack of
peaks assignable to M082 indicate that the active catalyst is
present in small domains and is well dispersed on the
suppon .....................................................................................

Wide angle XRD patterns for scaffold-12 MSU-y aluminas
taken at various stages of the catalytic testing process. The
y—AI203 phase is evident throughout the process and the lack
of peaks assignable to M082 indicate that the active catalyst
is present in small domains and is well dispersed on the
suppon .....................................................................................

Wide angle XRD patterns for the commercial Crosfield
catalyst, taken at various stages of the catalytic testing
process. The decrease in intensity of the peaks assignable
to y-AI203 in the pattern of the sulfided sample may be due
to the coating of the alumina surface by small domains of
M082. Peaks from y—Al203 and M032 in the pattern of the
Spent sample indicate that M082 particles may aggregate
sufficiently to cause scattering and expose the surface of the
suppon .....................................................................................

102

114

118

120

121

122

Figure 4.6.

Figure 4.7.

Figure 4.8.

Figure 4.9.

Figure 4.10.

Nitrogen adsorption - desorption isotherms and BJH
adsorption pore size distributions (inset) for lath-12 MSUay
catalysts taken at various stages of the catalytic testing
process. The pore volume steadily decreases as the catalyst
iS loaded, sulfided, and spent. Isothems are offset 100 cc/g
for clarity and PSD curves are displayed in the same order
as the corresponding isotherms ..............................................

Nitrogen adsorption - desorption isotherms and BJH
adsorption pore size distributions (inset) for scaffold-12
MSUJy catalysts taken at various stages of the catalytic
testing process. The pore volume decreases slightly as the
catalyst is loaded, sulfided, and spent while the pore size is
increased and the distribution broadened over the same
conditions. Isothems are offset 0, 100, 100, and 200 cc/g
from bottom to top for clarity and PSD curves are displayed
in the same order as the corresponding isotherms .................

Nitrogen adsorption — desorption isotherms and BJH
adsorption pore size distributions (inset) for the commercial
Crosfield catalyst taken at various stages of the catalytic
testing process. The pore volume decreases slightly as the
catalyst is sulfided and spent. However, the pore Size and
distribution remain relatively unaffected over the same
conditions. Isothems are offset 100 cc/g for clarity and PSD
curves are displayed in the same order as the corresponding
isotherms .................................................................................

Bright field TEM images of the scaffold-12 MSU-ry catalyst
(A) after loading and (B) after using in an HDS reaction for 3
hr at 400 °C and 400 psi. The scale bar represents 100 nm.
The inset iS an approximately 30 nm x 30 nm area of the
catalyst. That the framework of the alumina is largely
unaffected by the conditions of the HDS reaction provides
evidence of its stability .............................................................

Bright field TEM images of the lath-12 MSU-y catalyst (A)
after loading and (B) after using in an HDS reaction for 3 hr
at 400 °C and 400 psi. The scale bar represents 100 nm.
The inset is an approximately 30 nm x 30 nm area of the
catalyst. The local pore structure (inset) is compromised to
some degree, as is Shown by the decreased void space in
(B) as compared with (A). The long range order of the
alumina laths is lost after use in the HDS reaction ..................

124

125

126

130

132

Figure 4.11.

Figure 4.12.

Figure 4.13.

Bright field TEM images of the commercial Crosfield catalyst
(A) after loading and (B) after using in an HDS reaction for 3
hr at 400 °C and 400 psi. The scale bar represents 100 nm.
The inset is an approximately 30 nm x 30 nm area of the
catalyst. Although the local pore structure (inset) appears to
withstand the conditions encountered in the HDS reaction, it
appears that there is some aggregation of the fundamental
particles after use ....................................................................

The Daage-Chianelli rim-edge model of the M082 catalyst
particle. Sulfur removal takes place on rim and edge Sites
while hydrogenation of the aromatic ring iS restricted to rim
Sites only ..................................................................................

Influence of morphology on the ratio of rim sites to total sites
available on the edges of M082 particles. While increasing
the stack height leads to a Slight increase in the overall
number of edge Sites, the more pronounced effect is the
decrease in the relative number of rim sites ............................

xviii

134

136

138

LIST OF ABBREVIATIONS

4H-DBT Ring hydrogenated analog of DBT, dibenzothiophene
BET Brunauer Emmett Teller
BJH Barrett Joyner Halender
BP Biphenyl
S-BuOH sec-butanol, 2-butanol
CHB Cyclohexylbenzene
DBT Dibenzothiophene
DDA Dodecylamine
EO Ethylene oxide
EPA Environmental Protection Agency
EtOH Ethanol
GC-MS Gas chromatography - mass spectrometry
h-bond Hydrogen bond
HDA Hexadecylamine
HDS Hydrodesulfurization
HMS Hexagonal mesoporous Silica
HOMO Highest occupied molecular orbital
M418 Mobil family of mesoporous SilicaS
MOM-41 Mobil composition of matter 41
MSU-tS Mesoporous alumina comprised of a 0-AI203 framework
MSU-y Mesoporous alumina comprised of a y-AI203 framework

xix

MSU-B
MSU-SIB
MSU-X
nm

NMR
OA

P/Po
PEO

PO

ppm
PSD
s-BuOH
surf
TalA

Tal TetA
Tal TriA
TEM
TMS

v/v

XRD

Mesoporous alumina comprised of a boehmite framework
Surfactant-boehmite composition, precursor to MSU aluminas
Wormhole mesoporous aluminas (or Silicas)

nanometer

Nuclear magnetic resonance

Octylamine

Relative pressure, P = pressure, P0 = saturation pressure
Polyethylene oxide

Propylene oxide

parts per million

Pore Size distribution

sec-butanol, 2-butanol

Surfactant

Tallow amine

Tallow tetraamine

Tallow triamine

Transmission electron microscopy

Transition metal sulfide

volume ratio

X-ray diffraction

Chapter 1

Introduction

1 .1. Classification of Aluminas

Aluminas, a class of compounds comprised of aluminum, oxygen, and often
hydrogen, are wonderfully interesting and diverse. The name alumina iS broadly
applied to all members of this group that includes the aluminum hydroxides,
aluminum oxide hydroxides, and aluminum oxides. The properties of these
compounds vary greatly, from powders that are relatively soft to larger crystals of
extreme hardness, inert to readily reactive, free flowing to viscous, catalytically
active to inactive. Regardless of nature, it is interesting that all of these aluminas,
when heated to temperatures high enough and for periods of time long enough,
are eventually converted to d-alumina or corundum, the only stable form of
alumina.

Aluminas are commercially important chemicals. The world’s primary source
of aluminum is from the mineral bauxite, an impure hydrous ore of aluminum and
other metals. Bauxite is refined to useful alumina chemicals through the Bayer
process. The ore is first treated with hot concentrated sodium hydroxide to
extract the aluminum hydroxides and remove the impurities. Crystalline
aluminum hydroxide is then precipitated, washed, and thermally dehydrated to
produce aluminum oxide‘. While most of the alumina produced is used in the
production of metallurgical grade aluminum, a large amount is devoted strictly to

alumina chemicals. According to the International Aluminum Institute, over 40

million tonnes of alumina was produced in 20012. Not surprisingly, aluminas find
uses in a variety of applications including as fillers in plastics and paper, fire
retardants, adsorbents, abrasives, ceramics, and perhaps most importantly, as
catalysts and catalyst supports.

As mentioned above, a range of composition is observed within the family of
aluminas. They may be crystalline or amorphous. The crystalline aluminas are
aluminum hydroxides, AI(OH)3, aluminum oxide hydroxides, AlOOH, and
aluminum oxides, A|203. There are three forms of the trihydroxides: gibbsite,
bayerite, and nordstrandite. Of these, gibbsite is the most common and is the
product of the Bayer process. In its structure, aluminum ions occupy 2/3 of the
octahedral vacancies in the close packed lattice of hydroxide ions. Hydroxide
ions are aligned so that the sequence of layer stacking is AB...BA...AB...BA.
Bayerite is not found in nature, but has been prepared synthetically. Its structure
is similar to that of gibbsite, but the layer sequence is AB...AB...AB. The third
trihydroxide, nordstrandite, has a structure that is a composite of the previous
ones. Alternating double layers of gibbsite and bayerite are stacked to form the
structure‘. The best method to distinguish between these different crystalline
phases is x-ray diffraction (XRD). In addition to the crystalline trihydroxides,
amorphous AI(OH)3 also exists.

Two other members of the alumina family are aluminum oxide hydroxides,
AlOOH, boehmite and diaspora. Boehmite occurs as a component of bauxite and
can also be synthesized through neutralization of aluminum salts or hydrolysis of

aluminum alkoxides around the boiling point of water". The boehmite structure is

comprised of extended chains of AlOOH dimers. The chains form a double layer
in which hydroxyls in one layer are positioned over the depressions between
adjacent hydroxyls in the other layer. The oxygen ions are arranged in a cubic
close packed lattice. Diaspore is found in nature associated with geologically
older bauxite deposits and metamorphic rocks and it is therefore believed that
high pressures and temperatures are required for its formation. Its structure is
Similar to that of boehmite in that the basic unit is the chain of AlOOH dimers,
however, the oxygen atoms are arranged in a hexagonal close packed lattice. In
denoting aluminas, the common crystallographic designation assigns a prefix of
a- to the hexagonal phases and a prefix of y— to the cubic phases. Therefore,
boehmite is often represented as y—AIOOH and diaspore as a-AIOOH.

The only thermodynamically stable oxide of aluminum is corundum, a-Alzos. It
can be prepared by several methods, the most common being thermal
degradation of other aluminas at temperatures over 1000 °C. Corundum iS also
found in nature in igneous and metamorphic rocks. The 0- designation is
consistent with its hexagonally close packed oxygen lattice in which aluminum
ions occupy 2/3 of the octahedral holes. ln gem quality crystals of corundum,
trace impurities of chromium lead to the formation of rubies while iron and
titanium impurities lead to sapphires. The family of these alumina compounds is

summarized in Figure 1.1.

 

Aluminas

 

 

 

 

 

Hydroxides Oxides
Gibbsite, y-AI(OH)3 Aluminum Oxide, o-AI203
Bayerite, d-AI(OH)3 (Corundum)

Nordstrandite, A|(OH)3

 

 

 

 

 

 

 

V

Oxide Hydroxides
Boehmite, v-AIOOH

Diaspore, a-AIOOH

 

 

 

 

Figure 1 .1. Family of alumina compounds.

1.2. Transition Aluminas

AS mentioned previously, all aluminas eventually are converted to corundum
upon heating at high temperatures. The appropriately named transition aluminas
are disordered crystalline phases formed at temperatures of 250 — 800 °C during
the transformation to a-Al203 at approximately 1100 °C. At least seven different
transition aluminas have been reported: p (rho), X (chi), n (eta), y (gamma), K
(kappa), B (theta), and 0 (delta). The particular phase observed depends upon
both the precursor and the temperature to which that precursor is heated.
Gibbsite is converted first to X at a temperature of 350 °C, then to K around 750
°C while the other trihydroxides proceed though an n phase at lower
temperatures, then to 6 at temperatures over 800 °C. All of them are transformed
to 0 beyond 1100 °C. The oxide hydroxide phases exhibit different transformation
sequences than those of the trihydroxides. Boehmite is converted to v around
500 °C, to 0 above 750 °C, and then to 6 above 900 °C before reaching a above
1000 °C. The behavior of diaspore is interesting in that it converts directly to a at
a temperature of 500 °C. The similarity in the hexagonal packing of oxygen
atoms in the dense structures of these phases, requiring therefore only minimal
rearrangement upon transformation, is thought to be responsible for this
observation. Within the group of transition aluminas, the nearly anhydrous 6, K,
and 0 phases are more ordered than the other forms, as their formation requires
higher temperatures‘. The transformation sequences are summarized in Figure

1.2.

GibbsiteI—pl Chi |——+ | Kappa |Alpha |

|Boehmite| —> | Gamma | Delta |Theta |Alpha]

| Diaspore | —'—P | Alpha |

 

100 200 300 400 500 600 700 800 900 1000 1100
Temperature °C

Figure 1.2. Transformation sequence observed upon the thermal dehydration of
aluminas (adapted from Waters and Misraa).

Despite the number of different phases, the structures of the transition
aluminas are fairly Similar with respect to each other. This in part explains the
observed transformation sequence. X-Al203 exhibits a highly disordered
hexagonal packing of oxygen anions while the high temperature form of k- Al203
is also hexagonal but more ordered. n- Al203 has a tetragonally distorted spinel
structure in which 21 1/3 of the 24 cation vacancies are occupied by aluminum
ions. The 6- AI203 structure is similar to that of the lower temperature n form. The
structure of y- Al203 is also of the spinel type, but it can be distinguished from
that of the n form by its Significantly larger tetragonal distortion. 6- Al203 is Similar
to v, but it exhibits slightly greater long-range orders. In general, the identification
of transition aluminas is best performed through XRD and 27Al nuclear magnetic
resonance (NMR). Each transition alumina has a slightly different XRD pattern,
the higher temperature forms exhibiting sharper peaks than the others. 27Al NMR
can be used to distinguish the distribution of aluminum ions in the octahedral and
tetrahedral vacancies of the oxygen lattice.

As the transition aluminas are disordered phases, dehydration upon heating
and subsequent rearrangement of the lattices lead to the formation of porosity
and internal surface area. While there iS a loss of mass upon heating, the particle
sizes remain the same, resulting in the development of porosity. In the
dehydration of the trihydroxides, surface areas can reach as high as 350 m2/g at
lower temperatures before rapidly and continually declining toward the

conversion to a- Al203. For the oxide hydroxides, there iS also an initial increase

in porosity leading to surface areas of 90-100 m2/g at a temperature of 500 °C
that decreases to 15-20 m2/g upon heating at about 550 °C‘.

Upon the thermal dehydration of aluminas, the surfaces of the transition
aluminas are altered extensively. The condensation of surface hydroxyls and
their subsequent removal as water create a rough surface, exposing
coordinatively unsaturated aluminum ionS and leaving the remaining hydroxyl
ions in various environments. This leads to acidic and basic Sites on the surface
of the alumina. While they are also used as adsorbents in liquid drying, water
purification, and chromatographic columns, the combination of textural properties
(pore volume and surface area) and surface properties leads to activity in
catalytic applications. Consequently, transition aluminas are also referred to as
activated aluminas.

Of all of the different phases, y-Alea is perhaps the most important and most
widely utilized transition alumina. It is formed upon thermal dehydration of
boehmite at temperatures ranging from 450 to 750 °C as described by the
following equation:

2 AlOOH -> Al203 + H20
y- AI203 has been used as a catalyst in alcohol dehydration reactions to form
alkeneS and ethers and in isomerization reactions. y- A|203 has also proven to be
extremely useful as a catalyst support in numerous reactions due to its
combination of heat resistance, chemical inertness, and favorable textural and
surface properties. y- AI203 can be produced with surface areas ranging from 185

— 250 m2/g with pore volumes typically around 0.20 cc/g"3.

Two primary methods are employed in the preparation of supported catalysts,
impregnation and coprecipitation. In the incipient wetness impregnation method,
a salt solution of the active species iS deposited on the alumina and then
thermally degraded to form the active catalyst. In the coprecipitation method, salt
solutions of both the aluminum and active component are precipitated
Simultaneously, again followed by thermal degradation to the produce the active
catalyst. High surface area is a desirable property because a larger amount of
active catalyst can be loaded onto the support. Many of the reactions in which

alumina supported catalysts are used are important in the petroleum industry.

1.3. Hydrodesulfurization

One such reaction is the hydrodesulfurization (HDS) reaction, a crucial
component of petroleum refining. Heteroaromatic sulfur organic molecules,
common components of crude oil deposits, must be removed before fuels can be
cleanly consumed. Fuels high in sulfur, when burned, produce acidic sulfur oxide
species that lead to the production of smog and acid rain. Federal regulations put
forth by the EPA are calling for a 97% reduction in sulfur content in fuels, from
500 to 15 ppm sulfur, by the year 2010‘. At the same time that these constraints
are put into place, lower sulfur content crudes are being depleted, making it
necessary to rely upon feedstocks that contain larger amounts of sulfur. In
addition, the presence of sulfur has a detrimental effect on other catalysts

involved in downstream catalytic processing of feeds. Therefore, development of

improved methods for sulfur removal is pertinent both environmentally and
economically.

In the HDS reaction, sulfur is removed from organic molecules as H28. Many
studies of HDS processes are performed using thiophene or dibenzothiophene
(DBT) as substrates. These molecules, as well as substituted and polyaromatic
analogs, are representative of the sulfur containing compounds found in crude oil
stocks. Sulfur removal in these larger molecules is especially challenging as
steric effects come into play. Catalytic studies involving dibenzothiophene are
therefore better indicators of potential performance using actual oil feedstocks.
Reactions are usually performed in a catalytic reactor at temperatures ranging
from 350 — 400 °C and pressures of 400 psi.

The catalysts employed in HDS reactions are transition metal sufides (TMS).
Many TMS Show activity in HDS reactions, but because of their considerably
lower cost, the HDS catalysts typically employed in industrial processes are
either molybdenum or tungsten sulfides, combined with either nickel or cobalt
sulfides, loaded on a y— Al203 support. Some classic works by Chianelli and
others""14 over the last 20 years explain why these catalysts are effective, making
them the preferred choice for use in the petroleum industry.

Much work has been done toward Characterizing transition metal sulfide
catalysts. Several structure/function relationships have been derived as models
for catalytic activity and selectivity have been proposed to explain the nature of
these TMS materials. The primary factor in determining the utility of a TMS

catalyst is in the electronic effect that is observed in these catalysts. That is,

10

correlations to and trends in activity in HDS can be made based upon the
position of the metal in the periodic table. Pecoraro and Chianelli measured the
activity of TMS catalysts in DBT HD85. When the activity is plotted against the
periodic position, a “volcano” plot is observed (Figure 1.3), where a maximum in
the activity is seen for Group VIII metalSS. This trend was explained based upon
the metal-sulfur bond strengths. As first described by Sabatier in 1911, in order to
be an effective catalyst in a given reaction, a Species must form bonds of
intermediate strength with the substrate7. In the case of HDS, the metal-sulfur
bond should be strong enough that the substrate bind to the metal, but weak
enough that the catalyst allows H28 to desorb from the active Site. If the heat of
formation (M-S bond strength) is too strong, then the active site is in essence
poisoned by the substrate. If too weak, then the substrate does not bind tightly
enough to lead to C-8 bond cleavage. The transition metals in the center of the
d-block meet this requirement and are Show the greatest activity in HDS. The
activity for the 4d and 5d metals is considerably higher than the 3d metals which
are relatively inactive in HDS. Additionally, the 3d TMS do not exhibit a volcano
plot behavior, indicating that factors other than periodic position have an
influence on catalytic activity as well.

Using transition metals in anionic octahedrally coordinated M86” clusters,
Harris and Chianelli carried out calculations of the electronic structure of TMS in
an attempt to better explain the observed HDS activitiesa. Several factors were
identified to be responsible for these observations. Included were the degree of

covalency of the TMS, a measure of MS d-p orbital interaction and covalent

ll

 

I: O 1*"t Row Ru ' :
Ib A 2Ind Row
I 3rd ROW OS
10" .
I 1‘
b d
: :1

U I
1 I

mmol"s'1
d
°..
‘1

Mo

I Inn]

   
  

Molecules of DBT Converted

Q
I
1
2]; ._ . . . ....I

 

 

Co
10“ , Ni -
E W ‘2',
I 1
I- -I
F .

 

Periodic Position

Figure 1.3. Activity of transition metal sulfides in DBT HDS plotted as a function
of periodic position of the metal (adapted from Pecoraro and Chianelli)?

12

bond strength of the M-8 bond. The bond strength decreases on going across a
period as more electrons are placed in antibonding orbitals. At the same time
though, the degree of covalency increases as M-S orbital mixing increases.
These factors are in line with the Sabatier principle. A third factor was identified,
the symmetry and occupation of the highest occupied molecular orbital (HOMO).
This effect is seen in bonding aspects of TMS. The activity is best when the
HOMO in these clusters is a metal-based t2g set of antibonding orbitals. The
bonding to a thiophene molecule can be through the 8 atom in a o-bonding mode
or through the ring in a TT-bonding mode. For 3d metals, the o-bonding
interactions are generally weak as the lone pair of electrons on the sulfur that is
donated to the metal in bond formation is tied up in aromaticity. For 4d metals,
there is a o and Tr component to the bonding, thought to be a requirement for
effective bonding of thiophene to transition metals. This is further supported by
the fact that thiophene can act as a Tr-acceptor ligand and backbonding from the
metal to an antibonding TT' orbital on thiophene is observed, weakening the C-8
bond and thereby increasing activity. These three factors were grouped into a
Single activity parameter that correlated well with activity.

As mentioned previously, although other TMS exhibit greater activity in HDS
catalysis, it is the M082 that is the industry workhorse. When promoted by a
second metal such as Co, there is an increase in the electron density, and hence
activity, of the primary metal. This is the primary function of the promoter. A
model Similar to the previous one designed for unpromoted catalysts took into

account a second 3d metal, using MoM’Sg'“ clusters to study the electronic basis

13

for promotiong. AS the metal was varied across the period, the t2g set of orbitals
decreases in energy relative to the higher 4d t2g set. Upon reaching cobalt in the
series, the 4d t2g set is at an energy between the t2g and eg sets of the 3d metal.
Cobalt has seven d electrons, Six of which entirely fill the t2g set while the seventh
is now placed in the t2g set of Mo, formally donating to M0 and increasing its
electron density. The same effect is observed for nickel. All of the copper orbitals
lie below the Mo orbitals and electrons are withdrawn from the molybdenum in
this case. Other 3d metals had no effect. While this model does not predict the
exact location of the promoter atoms, it is assumed that they are linked to the
molybdenum atoms through a bridging sulfur atom. Furthermore, as electrons
are transferred from antibonding Co or Ni orbitals to antibonding Mo orbitals, the
MOS bond iS weakened as the 00-8 bond is strengthened. A sulfur shared
between the two metals will have an intermediate M-S strength and, in accord
with Sabatier’s principle, Should exhibit maximum activity.

There are several possible locations available to the promoter atom in the
M082 catalyst. If the catalyst is supported on, for example, alumina, the cobalt
may be embedded in the upper layer of the support, a phase denoted as
CozAl203. Cube-like domains of 00983 have also been observed as a component
of catalysts. Topsoe and Clausen identified a third phase as well, the so-called
Co-Mo-S structure, a mixed Co-Mo sulfide phase in which the three elements are
in close proximity to one another”. Through the use of Mossbauer spectroscopy,
the type and quantity of cobalt structures were determined. It was found that the

Co-Mo«8 structure was the catalytically active phase as trends in activity

14

correlated favorably with the amount of this phase present in the catalyst. It was
determined by use of analytical electron microscopy that the most common
position of the cobalt is on the edge of an M082 layer. This position for Co iS now
generally accepted and has led to the view that M082 itself can be treated as a
support for the promoter, allowing for a high content of cobalt and improved
activity.

In addition to electronic considerations in catalyst performance, a geometric
effect is also observed. This has to do with the nature of the crystal structure of
the catalyst. For example, RuS2 has an isotropic structure whereas M082 has an
anisotropic structure. The ruthenium catalyst’s structure is cubic and there exists
a correlation between HDS activity and the surface area of the catalyst. On the
other hand, the molybdenum structure is a layered one in which the sheets are
stacked only 2 - 10 nm thick, but extend 100 - 300 nm in a direction
perpendicular to the layer stack. Accordingly, activity does not correlate to
surface area of the M082 catalyst. While the basal planes of this material
contribute greatly to the surface area, they do not have an effect on activity in
HDS. However, oxygen chemisorption correlates well with HDS activity for both.
In the case of the M082, oxygen was Shown to accumulate on edges of the
particles, indicating that the edge planes contain the catalytically active sites".

While the electronic and geometric effects have an impact on the activity of an
HDS catalyst, the crystal structure also plays a role in selectivity. Daage and
Chianelli considered the two-site nature of the edges of M082 particles in

developing the rim-edge model‘z'“. In HDS of DBT, predominantly two products

15

are formed: biphenyl (BP) and cyclohexylbenzene (CHB). BP is formed from
direct desulfurization of DBT while CHB is formed when DBT iS first
hydrogenated before undergoing desulfurization. M082 was prepared under
various thermal conditions, leading to samples with varying degrees of
crystallinity. The layer stack height was determined by measuring the width of the
002 peak in the powder XRD pattern. It was found that as the particles became
increasingly crystalline and the stack height increased, there was a severe
decrease in the rate constant for the hydrogenation reaction. It was proposed
then that hydrogenation occurs only on rim Sites, those which are at the edge of
the basal planes. Desulfurization occurs on both rim and edge Sites, the latter
located at the edge of internal layers. A convenient feature of this model is that
the number of rim sites is‘ dependent only upon the stack height, not the shape of
the particles. Thus, selectivity is governed by height within the particle, activity by
the diameter of the particle.

The active sites of the HDS catalyst are sulfur vacancies at the edges of the
M082 particles. In promoted catalysts, it is likely that the active Sites are sulfur
vacancies near cobalt that is in the Co-Mo-S phase. A clear mechanism to
explain HDS catalysis has yet to be developed. Information used in developing a
mechanism has come from solid-state chemistry, organometallic chemistry, and
surface science studies. Many model systems use thiophene as the substrate.
Prins et al. suggest the following as a possibility“. Adsorption of the thiophene
sulfur atom onto the active site occurs first, followed by a four-electron reduction

process to yield butadiene and hydrogen sulfide. The electrons for the reduction

16

are supplied either by four one-electron or two two-electron oxidations of
molybdenum. In any model, thiophene (or DBT) can bind to the active site in two
manners: one-point end-on adsorption through sulfur or Side-on adsorption
through a carbon-carbon double bond. The proposed Side-on adsorption
mechanism involves the formation of persulfide ions. One support for this mode
is that persulfide ions are found in the structures of RuS2, 0382, two of the
highest activity TMS catalysts, and in the sulfides of cobalt and nickel, the
common promoters of activity.

For the larger DBT molecule and its substituted analogs, steric hindrance may
become a factor in deciding which mode of adsorption is most likely. Daage and
Chianelli address this in their report of the rim-edge model”. The HDS reaction
proceeds through a vertical (8 n‘) adsorption mode that can occur on all edge
sites. The hydrogenation reaction progresses through a flat adsorption mode that
requires a larger surface area of the catalyst. This is unlikely to occur on the
edge in general, but at the rim, where part of the molecule will dangle in space.
For thiophene, end- on adsorption is common, but for larger molecules such as
DBT, a planar adsorption is preferred.

While the theoretical studies described above are based on unsupported
catalysts, it is a common practice to use a catalyst support in actual HDS
reactions. Alumina is the support of choice in this reaction for several reasons.
The material can withstand the temperatures and pressures involved in the HDS
reaction. The support itself is inert and is not involved in the reaction. This is

important in terms of selectivity and efficiency, where acidity can lead to cracking

l7

reactions and increased consumption of costly hydrogen. Additionally, the
surface area of conventional y- AI203 is high enough that the catalyst can be
dispersed effectively, leading to a larger number of catalytically active sites. The
use of an alumina support is also cost effective in two ways. First, alumina itself
is inexpensive. Second, the use of the support allows for the dilution of the metal,
allowing for a less expensive catalyst. One concern with supported catalysts is
that the active Sites of the catalyst that are in contact with the support may be
inaccessible, thus lowering the activity of the catalyst. If the generally accepted
rim-edge model is envisioned and the M082 interacts with the support through
the basal planes, then the Sites that are blocked would be rim Sites, on which
hydrogenation occurs. Therefore, selectivity would be improved. However, if the
M082 particles attach to the support in other manners, then edge sites could

become buried, lowering the activity.

1 .4. Mesostructures

Since the disclosure of mesostructured Silica molecular sieves by Mobil
researchers in 199215”, there has been much work done in the development of
these materials. The original work employs cationic surfactant micelles to
template silicate oligomers into mesostructures. Removal of the surfactant
creates the final products, the M418 family of materials. These materials are
characterized by their uniform and tunable pore sizes from ~15 - 100 A, large

pore volumes, and high surface areas in excess of 1000 m2/g. The principal

18

member of this family, MGM-41, possesses hexagonal pores with long-range
order. Therefore, the mesoporous molecular sieves can be viewed as large-scale
analogs to microporous zeolites.

Noting the impact that mesostructured Silicas have had in catalysis, Similar
improvements may be anticipated with the use of mesostructured aluminas.
Particularly, these aluminas could be utilized as catalyst supports in HDS. The
larger pore volumes and surface areas of these materials when compared to
conventional aluminas would allow for better dispersal of catalyst and increased
access to active Sites. Additionally, larger amounts of catalyst may be loaded
onto the support without the blockage of pores. With this additional loading of
catalyst and increased number of accessible active Sites, the loss of active some
Sites due to contact with the support can be tolerated.

Following Mobil’s work, Stucky and coworkers identified the important steps in
the formation of mesostructured Silicas through electrostatic pathways". First,
there is multidentate binding of Silicate oligmers to the surfactant. Second, the
oligomers polymerize in the interface region. Third, charge matching between
inorganic and organic Species dictates the extent to which the reaction proceeds.
These factors were then used to extend MobiI’S original synthetic pathway into
four general electrostatic pathways leading to mesoporous materials”. In
addition to Silica, mesostructures of antimony, tungsten, iron, lead, zinc, and
aluminum oxides could also be prepared from these pathways, although not all

materials were thermally stable.

l9

Work from the Pinnavaia group established two additional pathways based
upon hydrogen bonding interactions between neutral inorganic precursors and
neutral primary amine surfactants”. The materials afforded differed from their
predecessors. Both materials exhibited uniform pores, but, owing to the relatively
weaker hydrogen bonding forces that govern the assembly of the materials in
latter work, long range order was not observed. These disordered Silicas, named
HMS, were characterized by the wormhole structural motif that arose.
Additionally, the wall thickness of materials prepared in this manner was greater
than that of materials synthesized from electrostatic pathways, where wall
thickness is determined by charge matching. The neutral assembly pathway
imposes no such restriction. The textural properties, however, were comparable
to those of the previously described materials. Additionally, surfactant molecules
could be removed (and recycled) from the aS-synthesized HMS materials through
solvent extraction, as opposed to calcination which is required to remove
surfactant from electrostatically prepared materials. A study by Tanev20
compared the properties of the two types of materials and reported that HMS
materials, due to the thicker framework walls, exhibit improved thermal stability.
The smaller particle Size of HMS leads to the formation of complimentary textural
porosity, facilitating access to framework-confined mesopores.

Another neutral assembly pathway, similar to the previously mentioned one,
was also discovered”. The composition of materials formed from these neutral

pathways could also be specified by selecting the appropriate inorganic

20

precursor. Dative bonding pathways to tantalum and niobium oxide mesoporous
materials were proposed by Antonelli and Yingzz'za.

In the ten years that have passed since MobiI’S initial report, the development
and proliferation of mesostructured Silica materials has been widespread. To
quantify this explosion of research in this field to some extent, consider that a
literature search for reports of mesoporous Silica Since 1992 will result in over
1,800 references while a similar search for mesoporous aluminas will yield less
than 8% of that amount. The 8% is exaggerated to some degree, as many of the
references describe aluminas prepared by conventional, non-templating,
methods or alumina in mixed metal oxide materials. The results of the search are
especially discouraging, noting the extensive use of aluminas in catalysis. It is
also a testament to the elusiveness of the ability to synthesize these materials.
While the increased utility of mesoporous Silicas in catalytic applications has

already been realized“, a Similar increase in the use of mesoporous aluminas is

anticipated in the near future.

1.5. Mesostructured Aluminas

Many aluminas are porous by nature. The transition aluminas formed through
thermal dehydration of aluminum hydroxides are examples. However, the pore
size in these materials is not uniform; a broad distribution of sizes is observed.
This leads to necking that can limit access to the pores. Mesoporous aluminas

assembled through templating pathways offer narrower pore Size distributions

21

and improved textural properties. Bagshaw and Pinnavaia first reported a
mesoporous alumina synthesized in this manner”. A series of aluminas, titled
MSU-X, were synthesized from hydrolysis of aluminum sec-butoxide in the
presence of nonionic surfactants. After calcination, these materials exhibited
surface areas up to 535 m2/g and pore volumes up to 0.68 cc/g with uniform pore
sizes. The wormhole frameworks were thermally stable to temperatures of 500
°C. The 5-coordinate aluminum centers detected by 27Al NMR spectrum, a
property of amorphous aluminas, suggests that Lewis acid sites may be present
in these materials.

Other reports of mesoporous alumina quickly surfaced following Bagshaw’s
initial communication. Yada and coworkers synthesized alumina mesophases
from aluminum nitrate and sodium dodecyl sulfate26'27. A lamellar phase can be
converted into a disordered hexagonal phase during hydrolysis of the aluminum
salt by urea. Although the hexagonal phase is not evident by XRD after
calcination at 600 °C, a low angle peak is observed, indicating that a disordered
mesoporous phase is present.

A report from Davis’ group described the formation of mesoporous aluminas
prepared from the hydrolysis of aluminum alkoxides in the presence of carboxylic
acids as structure directors”. Although the materials were calcined at relatively
low temperatures for Short durations, surface areas up to 710 mz/g were
reported. The pore size exhibited by these materials was near 20 A regardless of

surfactant Chain length.

22

Reporting a generalized synthesis strategy leading to large pore mesoporous
metal oxides, Stucky and coworkers provided an example of a mesoporous
alumina”. Nonionic surfactants were used to template the oxide from an
aluminum Chloride precursor. The homogenous sol formed was allowed to gel in
an open Petri dish at 40 — 60 °C. The resulting alumina, once calcined, showed
an average pore Size of 140 A with a pore volume of approximately 0.35 cc/g.
The framework walls of the alumina were amorphous.

A contribution from Gabelica and coworkers indicated a novel approach to
mesoporous aluminas”. Mixing of anionic and cationic surfactants was employed
to achieve better charge matching between the inorganic and organic phases.
Aluminum salts and Keggin ions were precipitated in the presence of these
surfactant mixtures to yield aluminas with surface areas up to 810 m2/g, an
unprecedented value. Remarkably, there has been no further work reported
stemming from this group’s initial report. The pore Size and pore volume reported
for the high surface area alumina were 27 A and 0.61 cc/g, respectively. The
material was comprised of a regular arrangement of pores with amorphous walls,

as judged by XRD and the 27AI NMR.

1 .6. Stabilization of Aluminas

All of the previously described aluminas are comprised of amorphous
frameworks. While these materials were generally thermally stable up to 500 °C,

they may not remain stable under certain catalytic conditions where higher

23

temperatures and pressures and hydrothermal conditions are encountered. Over
time and with extended use under these conditions, alumina particles can grow in
Size and Sinter, suffering structural and textural modification. Transition alumina
phases are only metastable and will eventually be converted to d-Al203 under
extreme thermal conditions. A concomitant loss of textural properties is observed
with this structural transformation. To combat these effects, aluminas have been
doped with small percentages of lanthanides, typically 1 — 5% lanthanum and
cerium. This has been carried out primarily by two methods. Coprecipitation of
the two oxide precursors or incipient wetness impregnation of lanthanide
containing salts can afford the stabilized materials. There are also two
explanations for the resulting stability of the doped materials. Shaper31 and
Beguin32 concluded that a lanthanum aluminate compound covered the surface
of the alumina. Oudet33 proposed that a LaAI03 perovskite type structure
behaves similarly. In each case, the stabilization is due to strong interactions
between the lanthanide structure and the alumina, inhibiting the conversion to o-
AI203. Alternatively, it was proposed by Burtin34 that stability was due to the
occupation of lanthanide ions in cationic vacancies of the spinal type lattice of
alumina. Braun et al. also concluded Similarly in their doping work”. In this
manner, the lanthanide ions fortify the alumina structure and limit the amount of
diffusion of aluminum ions on conversion to corundum. Other metal ions have
been used in stabilization, however, lanthanum and cerium have consistently
shown the best results. Cerium exhibits two common oxidation states, Ce3+ and

Ce“. Oxidation to the higher state can lead to the formation of CeO2 and its

24

leaching from the alumina. Lanthanum exhibits only the La3+ oxidation state and
is therefore a better stabilizer. The stabilization of MSU-X aluminas has been
achieved through lanthanide doping”. Indeed, textural properties were improved
in doped aluminas as compared to the pure aluminas.

In addition to the stabilization of amorphous aluminas, it is of interest to
synthesize mesoporous crystalline aluminas in an effort to improve catalytic
properties. Owing to the crystallinity, these materials exhibit inherent stability
when compared to their amorphous analogs. The first report of such an alumina
was recently made by Pérez-Pariente and coworkers”. In a synthesis Similar to
that of MSU-X alumina, aluminum alkoxides were hydrolyzed with nonionic
surfactants and dipropylamine to yield a thermally stable, crystalline mesoporous
alumina. The synthesis was altered to include thermal aging of the freshly made
gel. Upon calcination, mesoporosity was retained within the crystalline product.
The XRD pattern of the calcined product Showed peaks assignable to a transition
alumina phase, confirming a crystalline phase. Additionally, unlike many of the
amorphous aluminas, there was no peak in the 27Al NMR corresponding to 5-
coordinate aluminum centers. Surface areas in excess of 500 mzlg and pore
volumes of 0.91 cc/g were attained. The degree of crystallinity and the textural

properties were compromised after calcination at 700 °C.

25

1.7. Objectives and Rationale

AS can be inferred from the reports summarized above, the synthesis and
application of mesoporous aluminas are areas that are ripe for exploration and
development. One objective of this study is to synthesize novel mesoporous
aluminas. Alumina is perhaps the world’s most widely used catalyst support. As
such, new materials with improved structural and textural properties would be a
welcome addition to the alumina family. Noting the unstable nature of the
amorphous aluminas, the focus here is on crystalline aluminas. To date, only one
example of a crystalline mesostructured alumina has been reported. Although
these materials promise to be more stable than their amorphous relatives,
corundum, o-Al203, is the only phase of alumina that can be considered truly
stable. Therefore, doping of lanthanides and transitional metals into the alumina
frameworks in an effort to improve thermal and hydrothermal stability, with
retention of textural properties, will also be explored. The doping of ions into the
frameworks of these aluminas can also stabilize a particular phase. This is
important in many catalytic applications where the elevated temperatures
required for catalysis initiate phase transitions of the alumina, accompanied by
degradation of textural properties. Additionally, the incorporation of transition
metal ions may introduce catalytically active centers into the alumina. Whatever
strategy is employed for stabilization, the goal is to produce high performance

mesoporous aluminas.

26

Finally, another goal is to demonstrate the value and utility of these newly
designed aluminas. Especially of interest is the ability of these materials to
function as catalyst supports in hydrodesulfurization reactions. Crystalline
aluminas offer the necessary stability under the conditions used in HDS reactors.
Aluminas with improved properties give hope that more effective and efficient
routes to sulfur removal will be found, enabling the petroleum and automotive
industries to meet the demands of government and society. This iS an
increasingly crucial need as high quality oil feedstocks are consumed, leaving

heavier and dirtier stocks as the primary fuel source.

27

1.8. References

(1)

(2)
(3)

(5)
(6)
(7)
(8)
(9)
(1 0)

(11)

(12)

(13)

(14)

(15)

(16)

Misra, C. American Chemical Society Monograph, Vol. 184: Industrial
Alumina Chemicals, 1986.

International Aluminum Institute, 2002.

Wefers, K.; Misra, C. "Oxides and Hydroxides of Aluminum," Alcoa
Laboratories, 1987.

In New York Times: New York, NY, 2000.

Pecoraro, T. A.; Chianelli, R. R. J. Catal. 1981, 67, 430-445.
Chianelli, R. R. NATO AS! Sen, Ser. C1983, 105, 361-378.
Sabatier, P. Ber. Deutsch Chem. 693. 191 1 , 44, 2001 .

Harris, 8.; Chianelli, R. R. J. Catal. 1984, 86, 400-412.

Harris, 8.; Chianelli, R. R. J. Catal. 1986, 98, 17-31.

Topsoe, H.; Clausen, B. S. Catal. Rev. - Sci. Eng. 1984, 26, 395-420.

Tauster, S. J.; Pecoraro, T. A.; Chianelli, R. R. J. Catal. 1980, 63, 515-
519.

Daage, M.; Chianelli, R. R. J. Catal. 1994, 149, 414-427.
Daage, M.; Chianelli, R. R.; Ruppert, A. F. Stud. Surf. Sci. Catal. 1993, 75,
571-584.

Prins, R.; De Beer, V. H. J.; Somorjai, G. A. Catal. Rev. - Sci. Eng. 1989,
31, 1-41.

Beck, J. 8.; 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.; et al. J. Am.
Chem. Soc. 1992, 114, 10834-10843.

Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. 0.; Beck, J. 8.
Nature (London) 1992, 359, 710-712.

28

(17)

(18)

(19)

(20)

(21)

(22)

(23)
(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

Monnier, A.; Schuth, F.; Huo, O.; Kumar, 0.; Margolese, D.; Maxwell, R.
8.; Stucky, G. D.; Krishnamurty, M.; Petroff, P.; et al. Science
(Washington, D. C., 1883-) 1993, 261, 1299-1303.

Huo, 0.; Margolese, D. l.; Ciesla, U.; Feng, P.; Gier, T. E.; Sieger, P.;
Leon, R.; Petroff, P. M.; Schueth, F.; Stucky, G. D. Nature (London) 1994,
368, 317-321 .

Tanev, P. T.; Pinnavaia, T. J. Science (Washington, D. C.) 1995, 267,
865-867.

Tanev, P. T.; Pinnavaia, T. J. Chem. Mater. 1996, 8, 2068-2079.

Bagshaw, 8. A.; Prouzet, E.; Pinnavaia, T. J. Science (Washington, D. C.)
1995, 269, 1242-1244.

Antonelli, D. M.; Ying, J. Y. Angew. Chem, Int. Ed. Engl. 1996, 35, 426-
430.

Antonelli, D. M.; Ying, J. Y. Chem. Mater. 1996, 8, 874-881.
Corma, A. Chem. Rev. (Washington, D. C.) 1997, 97, 2373-2419.

Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem, Int. Ed. Engl. 1996, 35,
1 102-1 105.

Yada, M.; Machida, M.; Kijima, T. Chem. Commun. (Cambridge) 1996,
769-770.

Yada, M.; Hiyoshi, H.; Ohe, K.; Machida, M.; Kijima, T. Inorg. Chem. 1997,
36, 5565-5569.

Vaudry, F.; Khodabandeh, 8.; Davis, M. E. Chem. Mater. 1996, 8, 1451-
1464.

Yang, P.; Zhao, D.; Margolese, D. |.; Chmelka, B. F.; Stucky, G. D. Chem.
Mater. 1999, 11 , 2813-2826.

Valange, 8.; Guth, J. L.; Kolenda, F.; Lacombe, 8.; Gabelica, Z.
Microporous Mesoporous Mater. 2000, 35-36, 597-607.

Schaper, H.; Doesburg, E. B. M.; Van Reijen, L. L. Appl. Catal. 1983, 7,
211-220.

Beguin, B.; Garbowski, E.; Primet, M. Appl. Catal. 1991, 75, 119-132.

29

(33)

(34)

(35)

(35)

(37)

Oudet, F.; Courtine, P.; Vejux, A. J. Catal. 1988, 114, 112-120.

Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M. Appl. Catal. 1987, 34,
239-254.

Braun, 8.; Appel, L. G.; Zinner, L. B.; Schmal, M. Br. Ceram. Trans. 1999,
98, 77-80.

Zhang, W.; Pinnavaia, T. J. Chem. Commun. (Cambridge) 1998, 1185-
1186.

Gonzalez-Pena, V.; Diaz, l.; Marquez-Alvarez, C.; Sastre, E.; Perez-
Pariente, J. Microporous Mesoporous Mater. 2001, 44-45, 203-210.

30

Chapter 2

Synthesis and Characterization of Mesostructured Boehmite

2.1. Introduction

Much effort has been invested in the development of alumina phases for
industrial use. Heavy interest lies in the production of transition aluminas with
favorable surface and textural properties. The primary target phase is y-Al203.
This phase is achieved by thermal dehydration of boehmite‘. Noting the
importance of y-AI203 as a catalyst and catalyst support, one can appreciate that
boehmite is also a commercially important chemical. However, comparatively few
reports on the synthesis of boehmite can be found in the literature.

Aside from its most important use as the precursor to y-Al203, boehmite has
other uses as well. Boehmite has served as a component In the synthesis of
aluminum containing anionic clays or layered double hydroxides2 and carboxylate
alumoxanes3. It has been used infrequently as a catalyst support and as a
catalyst itself. The lack of applications of boehmite as a catalyst stems from the
widespread use of pseudoboehmite in this capacity. The differences in these two
materials are discussed below.

Boehmite is an aluminum oxide hydroxide (AlOOH). The structure of boehmite
is comprised of close packed oxide and hydroxide ions in which double chains of
AlOOH molecules extend along the a-axis. These units are stacked in the

direction of the b-axis where hydroxyls of one layer are located over the

31

depressions between adjacent hydroxyls in the other layer (Figure 2.1)“.
Hydrogen bonding holds the double layers together.

Pseudoboehmite, or gelatinous boehmite, can be classified as alumina
structurally Similar to that of boehmite, but having a greater water content and
smaller particle Size than boehmite. Aluminum alkoxides are generated as
byproducts in the formation of linear alcohols by the Ziegler process. Hydrolysis
of these alkoxides results in the production of pseudoboehmite, available
commercially under the name of Catapal‘. This material has a high surface area
and is easily converted to a useable sol form by peptizing with dilute (typically
nitric) acids. In this respect, it is discernable from boehmite. When deciding which
material to use in applications where a soluble aluminum source is desired, this
property makes pseudoboehmite the material of choice.

Not surprisingly, the terms boehmite and pseudoboehmite are often
incorrectly used in the literature. In addition, both have also been referred to as
aluminum hydroxides, causing even more confusion. There has been some
debate as to the actual differences in the two materials. Boehmite contains 15%
H2O by mass where pseudoboehmite has been found to contain up to 30% H2O.
The location of the excess water in the structure was originally thought to be
contained in the region between the double layers. However, d-Spacings found
for the 020 peak in the XRD patterns of the materials were not large enough to
accommodate such an excess volume of water“. This leads to the currently
accepted idea that pseudoboehmite is structurally identical to boehmite, differing

only in its smaller crystallite size and higher water content. In this model

32

OH OH OH OIH CiH

(1H I I —Al— -—A1— —Al—
Ti I‘i‘i—i“? I _i_.I_I._I_
—:I—O—:l— O—TI—O—Al—O _.T

”on, “on l‘l_,0H,\ III/OH‘R

“(1H \f'oIH (1H OIH 'OlH' (1H
— —AI— —Al— —Al— —?1— (r—Tl— (I—Tl—

‘I i I I I I _ _ _
— A1—(|)— Tl—O— ITI—O— Tl—O— Ald—O [I] o

OIH OH OH OH OH OH

12A

 

Figure 2.1. Cross-sectional (upper) and three-dimensional (lower) depiction of
the boehmite double-layer. The thickness of the double layer is ~ 12 A.

33

of pseudoboehmite, the increased number of terminal hydroxyl groups accounts
for the excess water. The smaller the crystals, the greater the number of
hydroxyls present.

Traditionally, boehmite is synthesized by the neutralization of aluminum salts
under hydrothermal conditions“. Unlike the y-Al203 and the other transition
aluminas, boehmite iS relatively non-porous. Mesoporous aluminas prepared in
the presence of surfactants have been reported in the literature, however, all but
a few describe materials that are amorphouss’”. Similarly, the use of surfactants
to direct the formation of mesostructured boehmite is reported here. The novelty
of this approach lies in the choice of the inorganic precursor and the surfactants
employed in the synthesis and the method of surfactant removal. The use of
aluminum sec-butoxide allows for formation of a boehmitic alumina phase during
synthesis. Primary and tallow amine surfactants decompose at a temperature
that allows for their removal via calcination without the concomitant conversion of

boehmite to y-Al203.

2.2. Experimental

2.2.1. Synthesis

The synthesis of mesostructured boehmite was executed as follows: The
surfactant, a primary or tallow amine (see footnote to Table 2.1), was dissolved in
a solution of ethanol and water (75/25 v/v) at ambient temperature. To this

solution, aluminum sec-butoxide was added under vigorous stirring by the use of

34

a magnetic stirrer. The mixture was either allowed to stir at ambient temperature
or was transferred to a temperature controlled shaker bath for 20 hours in a
sealed vessel. Precipitates formed upon hydrolysis of the alkoxide were
recovered by vacuum filtration, washed with water and ethanol, and allowed to
dry at ambient temperature. The final surfactant-free products were obtained by
calcining the aS-synthesized products in air at 325 °C for 10 hours. Alternatively,
surfactant could be removed by ethanol extraction at 80 °C. While the amounts of
the starting materials were varied, a typical synthesis had the following molar
ratios:

5.0 AI(8-BUO)3 : 1.0 Surfactant : 28.7 Ethanol : 31.0 Water

2.2.2. Characterization

Powder diffraction patterns were collected on a Rigaku Roteflex
Diffractometer using Cu K, radiation (A = 0.154 nm). Intensity was calculated in
counts per second every 0.02 degrees 26. Sweep rates were varied from 2 — 8
degrees 26/min.

Nitrogen adsorption — desorption isotherms were collected on a Micromeritics
TriStar 3000 sorptometer at 77K. Samples were degassed at least 12 hours at
150 °C under a vacuum of 1043 torr prior to analysis. Surface areas and pore
sizes were determined by the BET15 and BJHl6 methods, respectively.

TEM images were collected on a JEOL 100CX microscope with a CeBs
filament at an accelerating voltage of 120 kV. Samples were prepared by

sonicating a suspension of the powder in ethanol for 20 min. One to two drops of

35

the suspension were then placed on a carbon coated holey film that was
supported on a 3mm, 300 mesh copper grid.

27Al NMR spectra were collected on a Varian 400 VRX solid — state NMR
spectrometer. Samples were placed in a zirconia rotor and Spun at 4 kHz with a
pulse width of 0.5 us and a pulse delay of 0.5 3. Chemical Shifts were referenced

to AI(H2O)63+ which is assigned a value of 0 ppm.

2.3. Results and Discussion

2.3.1 . Formation of MSU-SIB

Addition of the aluminum sec-butoxide to the surfactant solution results in
rapid hydrolysis of the alkoxide, instantly forming a precipitate of a surfactant —
boehmite composition, denoted MSU-SIB. The conversion of the alumina species
is as follows:

Al(O-‘°’Bu)3 + 2H2O —-> AlOOH + 3 s-BuOH

It has been reported that the formation of boehmite requires hydrothermal
conditions‘. Although these reactions are performed at ambient temperature, two
reasons may help to explain the presence of boehmite as a product of the
reactions. First, the hydrolysis of the alkoxide is exothermic and a considerable
amount of heat is released. The volume of the solution in the reaction mixture iS
not great enough to disperse this heat and initially the temperature in the reaction
vessel is much greater than ambient temperature. Second, and more importantly,

by providing multiple Sites for the formation of hydrogen bonds between amine

36

head groups and boehmite hydroxyls, the use of the surfactant aids in the growth
of particles. Boehmite particles will terminate with an exposed layer of hydroxide
groups at the surface. These groups readily form h-bonds with the surfactant
head groups (Figure 2.2). In contrast, hydrolysis of aluminum alkoxides in the
absence of surfactant leads to the formation of pseudoboehmite. Indeed, this is
one method of commercial production of pseudoboehmite.

Upon calcination, the surfactant is removed and the as-synthesized MSU-SIB
iS converted to a mesostructured boehmite, denoted MSU-B. Noting the
transformation sequence upon dehydration of aluminas, the conversion of
boehmite to y-Al203 occurs at temperatures of 400 °C‘. Consequently, if
calcination is performed above 400 °C, the structural integrity of the boehmite
phase is compromised. The use of primary and tallow amines is key in this
regard because they will degrade at temperatures below 400 °C and can
therefore be removed at lower calcination temperatures. Also upon calcination,
particles may Sinter leading to growth of larger crystals at the expense of smaller
ones. This helps to ensure that the oxide framework is comprised of boehmite
and not pseudoboehmite. .

Calcination could also be performed at higher temperatures. In fact, MSU-B
materials were stable to calcination at 350 °C for short durations. At longer
periods or at 400 °C, the onset of conversion from boehmite to y—Al203 was
observed as judged by the appearance of peaks assignable to y-Al203 in the

wide angle XRD patterns. Therefore, milder conditions were preferred and

37

 

 

      
    
 

 

F 3
L J [ Irfir

 

 

 

 

r l
IIIITETIIII

«at m

//‘\\ /\"".

. . [\\,-
P

3 4

1" ‘ ‘

VV "rat-5?.»
9.5:, r

W °” 5“”

Figure 2.2. Schematic of interactions leading to formation of MSU-SIB. Each
Slab represents one boehmite double layer. Lines on the boehmite layers
represent hydroxyls forming hydrogen bonds to the surfactant micelles. Areas of
overlapping boehmite layers indicate cross-linking through condensation of
adjacent hydroxyls.

38

calcination was performed at the lowest temperature possible to preserve the
boehmite phase while still removing the surfactant.

Calcination is a convenient method of removing the surfactant, although the
surfactant is destroyed in the process. Surfactant removal and recovery can be
achieved through ethanol extraction if desired. However, the small amounts of
primary and tallow amines used and their costs, compared to other more
expensive surfactants such as polyalkylene oxides, do not warrant their recovery.
Calcination also ensures that hydroxyls on the boehmite surface are condensed,

eliminating water and fortifying the structure of MSU-B.

2.3.2. General Properties of MSU-B Aluminas

MSU-B aluminas were characterized by powder XRD, nitrogen adsorption,
TEM, and 27Al NMR. Figures 2.3 and 2.4 Show the low angle and wide angle
powder XRD patterns, respectively of aS-synthesized MSU-SIB and MSU-B
materials, prepared under conditions as described in Table 2.1, line 3. From
peaks in the wide angle patterns it is evident that the alumina is present in the
boehmite phase. This occurs invariably in all samples prepared in the method
described above. However, while a low angle peak is observed for this sample,
this is not always the case. As discussed in greater detail below, MSU-B
materials derived from alkoxide are not formed through a true templating
mechanism. Therefore, low angle peaks that are typically taken as evidence of
mesophase formation due to pore-to-pore correlation distances in the synthesis

of mesoporous materials cannot be relied upon here.

39

 

6.2 nm

4,-- MSU-B

Intensity

 

 

 

 

 

8 10

Degrees (2(9)

Figure 2.3. Representative low angle XRD patterns for the as-synthesized
surfactant-boehmite composite, denoted MSU-SIB, prepared from aluminum sec-
butoxide and dodecylamine in an alcoholic/aqueous medium as described by line
3 in Table 2.1 and the surfactant-free analog, MSU-B, obtained by calcination of

the MSU-SIB composite at 325 °C for 10 hr.

40

 

020

120
MSU-B

 

. 140

031 051
200 231

 
    

 

 

Intensity

 

 

 

 

 

Degrees (20)

Figure 2.4. Representative wide angle XRD patterns for MSU-SIB and MSU-B
aluminas Shown in Figure 2.3. Crystallographic planes contributing to the peaks

are noted.

41

In the particle assembly of these materials, the morphology of the oxide formed

does not necessarily reflect the shape of the surfactant micelle. Rather, it is
believed that a low angle peak, when observed, iS due to the thickness of
fundamental particles. While the observance of the peak itself is not always
reproduced from one reaction product to another under identical reaction
conditions, the d-spacings extracted from the peak position are not. The height of
the boehmite double layer is approximately 12.2 A. The recurring values of the
observed d-Spacings of calcined samples are approximately integer multiples of
12 A. It appears that 5 — 7 double layers per particle is most common as
suggested by the position of the low angle reflection that occurs most frequently
at 26 values corresponding to d-Spacings of 60, 72, and 84 A. There is more
variance in the d-Spacings of the MSU-SIB samples, but there may be some
water or alcohol intercalated between the layers that causes the repeat distance
to differ from that of the theoretical value of 12 A.

The nitrogen adsorption — desorption isotherms of MSU-B materials are
Shown in Figure 2.5. These isotherms correspond to aluminas prepared with
varying surfactant concentrations under conditions described in lines 3, 8, and 19
of Table 2.1. The isotherms for MSU-B most Closely resemble the Type IV

”"8 and associated with mesostructured materials. The

isotherm defined by Sing
lack of a well-defined step in the adsorption branch indicates that the pore Size
distributions are somewhat broad. This is in line with the particle assembly
pathway leading to formation of the material. Although not well defined, the

isotherms indicate that MSU-B materials have more uniform porosity as

42

 

1000

       
 

 

 

     
  

 

 

 

2 a
2
8004 1’:
A ‘ 4% DDA
O'l .
B . oungflfit'oflflisflfizo
3 Pore Diameter (nm)
'0 6004
d)
n 4
L
O .
g 9% DDA
< -
q, 4004
E _
2
o _.
> -
0% DDA
200* (Blank)
o"'l"‘l"'l"'lr"
0 0.2 0.4 06 08 1
PIP
0

Figure 2.5. Nitrogen adsorption — desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-B aluminas differing in surfactant concentration.
The percentage given corresponds to the mass percent of dodecylamine in the
surfactant solution. The three isotherms, offset by 200 cc/g for clarity, correspond
from top to bottom to samples described in lines 3, 8, and 19 in Table 2.1. PSD
curves are displayed in the same order as the isotherms.

43

Table 2.1. Synthetic parameters and textural properties of MSU-B aluminas.

 

 

 

 

 

b Pore Pore
Al/SurI/EtOH/H2O Temp 833 ,
Surfactant“ . 2 Sue" Volume
Molar RatIo (°C) (m lg)
(nm) (cc/0)
1 DDA 5 2 1 220.1 : 196 20 380 5 0.59
2 DDA 5 : 1 :42.6 2 138 20 428 3 0.72
3 “DDA” 5 :“’i”:‘68.0”:“73.4“””20“””4‘55”" ’7 4.5 '7 ‘ 0195“”
'4 "”‘bD‘A 5?'1‘"‘“:”58I0':75.4"“ 45 “ ”"4977W3.5‘ 084
5 DDA 5 2 1 268.0 2 73.4 65 460 5 0.82
6 DDA 5 2 1 2 68.0 273.4 95 451 4 0.69
" 7 DA 5: 1 : 20.0 : 21.5 20 325 5 0.47
8 DDA 5 : 1 :28.7 2 31.0 20 378 3 0.59
9 HDA 5 : 1 237.4 2 40.4 20 391 3.5 0.59
10 Tal A 5 : 1 :46.2 249.9 20 410 2.5 0.60
11 Tal TriA 5 : 1 264.0 2 69.0 20 429 2 0.60
12 Tal TetA 5 : 1 272.8 2 78.7 20 436 2.5 0.61
“1'37 OA 2.2: 1 503221.577 ~'20 #4527 5" “Ciel—T”
14 DDA 3.1 2 1 :20.02 21.6 20 418 5 0.63
15 HDA 4.0 : 1 220.0: 21.6 20 434 3 0.64
16 Tal A 4 0: 1 220.0: 21.6 20 410 3 0.60
17 Tal TriA 5.9 : 1 220.0: 21.6 20 462 3 0.63
18 Tal TetA 6.9 : 1 :20.0: 21.6 20 459 3 0.61
"1‘9“"Nofiemw 570728.77513—7720 322 "4" l ”0.71””7

 

a DDA = dodecylamine, OA = octylamine, HDA = hexadecylamine, Tal A = tallow
amine, Tal TriA = tallow triamine, Tal TetA = tallow tetraamine
The tallow amine structure is generalized by the following formula:

R-HN-(CH2CH2CH2NH)x-H

Where R = a mixture of alkane groups ranging from C12-C13 and x = 0, 1, 2, or 3
for tallow amine, diamine, triamine, and tetraamine, respectively.

b Ambient temperature taken as 20 °C

° BJH pore diameter as measured from the adsorption branch

compared to the sample prepared without surfactant. The shape of the isotherm
of the blank sample does not proceed through an inflection point at any partial
pressure above 0.2. In this manner, its behavior most closely resembles a Type II
isotherm, exhibited by non-porous materials.

A summary of textural properties of MSU-B aluminas iS provided in Table 2.1.
Included also are the textural properties for one sample prepared in absence of
surfactant. From the data, it can be seen that surface areas of MSU-B aluminas
in the approximate range of 380 — 460 m2/g, pore Sizes 2.0 — 5.0 nm, and pore
volumes 0.60 - 0.95 CC/g. Entries in lines 1-3 of the table correspond to samples
in which the surfactant solution becomes higher in ethanol content. Based upon
XRD evidence, more surfactant is incorporated into the mesostructure at higher
ethanol content. Free surfactant peaks appear around 20 degrees 26 in the wide
angle powder XRD patterns (not Shown) of precipitates borne out of low ethanol
content solvents. Judging by the textural properties of these samples, surface
area, pore size, and pore volume all increase at higher ethanol content. The
more surfactant incorporated into the mesostructure, the greater the void left
upon its removal. The isotherms are provided in Figure 2.6.

The third through sixth entries in Table 2.1 are for samples of identical
stoichiometry aged at different temperatures. Figure 2.7 provides the isotherms
for these aluminas. While the surface areas seem random, both pore size and
volume steadily decrease with an increase in aging temperature. At higher

temperatures, IeSS surfactant iS incorporated into the mesostructure.

45

 

(NM!)

1000 1 K 75% EtOH

V V I 1' ‘T‘T YYYYYYYYYYY

   

1

 

 

00

O

O
I

o 5 10 15
Pore Diameter (nm)

   

Volume Adsorbed (cc/g)
a:
O
O

 

 

 

_ 50% EtOH
400—

‘ 25% EtOH
2006

O I I I I I I I I r I I I I r I I . . .
0 0.2 0.4 0.6 0.8 1
PIP
0

Figure 2.6. Nitrogen adsorption - desorption isotherms and BJH adsorption pore
Size distributions (inset) for MSU-B aluminas prepared from aluminum sec-
butoxide and dodecylamine at ambient temperature in alcoholic/aqueous
solvents of varying ethanol content. The three isotherms, offset by 200 cc/g for
clarity, correspond to MSU-B aluminas described in lines 1-3 of Table 2.1. PSD
curves are displayed in the same order as the isotherms.

 

d
N
O
O
1 l 1 1
dV/dD

 

 

 

IIIIIIIIIIIIIIII

   

Pore Diameter (nm)

 

Volume Adsorbed (cc/g)

 

 

 

Figure 2.7. Nitrogen adsorption — desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-B aluminas prepared from aluminum sec-
butoxide and dodecylamine at different temperatures. The four isotherms, offset
by 200 cc/g for clarity, correspond to MSU-B aluminas described in lines 36 of
Table 2.1. PSD curves are displayed in the same order as the isotherms.

47

Filling of pores also occurs at lower partial pressures for products of higher
temperature reactions, signifying larger particle growth.

Entries in lines 7-18 of Table 2.1 are for aluminas prepared from different
amine surfactants under otherwise identical processing conditions. Nitrogen
adsorption - desorption isotherms for these MSU-B samples are provided in
Figure 2.8. The aluminum to surfactant molar ratio was held constant at 5:1 and
the surfactant solution in each case was 9% by mass surfactant (lines 7-12).
Then this series of experiments was repeated, maintaining a constant aluminum
to surfactant mass ratio (lines 13-18). It is interesting that the pore size and pore
volume are approximately constant regardless of the size, mass, or molar mass
of the surfactant. The surface areas do vary, but not predictably. There is a
steady increase with increasing surfactant size in the constant molar ratio
products, but they do not correlate well with surfactant size in the latter case. in
traditional templated syntheses, pore sizes and volumes increase with increasing
size of the surfactant molecule. That the pore size remains nearly constant under
a range of synthetic conditions gives support to a particle assembly mechanism.

The porosity of MSU-B boehmites arises from the voids between individual
boehmite particles. As noted above, the pores of the oxide framework do not
mimic the size or shape of the surfactant. So, although regular micelles may be
formed, the boehmite particles may be ordered randomly about the micelles,
leading to irregularly shaped pores. Again, noting the difference in these
materials and silica materials prepared through supramolecular templating, there

may be instances when the observed d-spacing is actually smaller than the

48

 

 

     
 

 

 

 

 

 

 

 

1.00:, k

‘3

“'0

j k TaiTetA
A i ................
m _ ll 3 1'0 1T5 2'0
3 1200 _ Pore Diameter (nm) Tal TriA
3 q ,_.
'o i
.2 _ TalA
h ..
O 1
'8 J HDA
< 800—:
d) d/,
g _: DDA
'5 */ f
> 2 //

400: / 0A
OIIIIII'FT“l‘r"'Tr
0 02 0.4 06 08 1
PIP
0

Figure 2.8. Nitrogen adsorption - desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-B aluminas prepared from the hydrolysis of
aluminum sec-butoxide in the presence of different amine surfactants as
porogens. The six isotherms, offset by 200 cc/g for clarity, correspond to MSU-B
aluminas described in lines 7-12 of Table 2.1. PSD curves are displayed in the
same order as the isotherms.

average pore size. in materials prepared from templating synthesis, pore sizes
are smaller than the d-spacings because the d-spacing is indicative of the repeat
distance of both pore and framework wall. However, in MSU-B, the XRD peak is
most likely due to the particle thickness and the pore size is due to void spaces,
between particles. Therefore, the fact that the small angle XRD d-spacings do
not correlate with the pore size is understandable.

The presence of a scaffold morphology was identified through TEM. Bright
field images of representative MSU-B aluminas are shown in Figure 2.9. The
open framework resulting from the random orientation of fundamental boehmite
particles can be seen. it is estimated that average particle thickness in these
samples is 3 — 7 nm. Electron diffraction patterns are also provided for these
materials. The presence of diffuse rings confirms the polycrystalline nature of the
materials. Many small crystals in all orientations give rise to the diffuse rings.
This is in contrast to a single crystal that would show distinct points in the pattern.
The patterns contain four rings containing the strongest reflections of boehmite.

27Al NMR spectra were obtained to verify the coordination environment of the
aluminum in MSU-B) (Figure 2.10). 2"Al NMR chemical shifts are referenced to
[Al(H20)6]3", which is assigned a value of 0 ppm. Octahedral (6-coordinate),
tetrahedral (4-coordinate), and 5-coordinate aluminum will have shifts of less
than 10 ppm, 65 — 75 ppm, and approximately 35 ppm, respectively. In boehmite,
all aluminum has octahedral coordination. The shift observed for MSU-B at 7

ppm is in agreement with a boehmite structure. The small peak coincident with

50

Figure 2.9. Bright field TEM images (A and C) and electron diffraction patterns
(B and D) of MSU-B materials prepared from aluminum sec-butoxide and
dodecylamine as described in Table 2.1, (A) line 8 and (B) line 3. The open
framework results from the random orientations of fundamental particles.
Electron diffraction patterns, (B) and (D), corresponding to bright field images (A)
and (C), respectively show diffuse rings, signifying the polycrystalline nature of
the material. The reflections contributing to the pattern are labeled in (D).

51

 

 

52

7.00

 

 

 

[lTTIllTTTIrrFTrUIIllllllrllllliiiTlTlrl]

200 150 100 50 0 -50 -100 -150 -200
ppm

Figure 2.10. 27Al NMR spectrum of MSU-B prepared from aluminum sec-
butoxide and dodecylamine prepared as described in Table 2.1, line 3. The
resonance at 7 ppm is consistent with the structure of boehmite, in which all
aluminum ions are in octahedral environments.

53

the spinning sideband may signify a small amount of Al in a tetrahedral
environment, possibly due to the onset of the boehmite to gamma alumina phase

transition.

2.3.3. Peptization

Boehmite and pseudoboehmite can be differentiated from each other based
upon their ability to be dispersed in dilute acidic solutions. The small particle size
and high water content of pseudoboehmite will cause this form of boehmite to
dissolve and form a sol in such solutions. This form of boehmite is said to be
peptizable. Boehmite, on the other hand, does not behave in this manner. MSU-
B materials were tested for peptizability to determine the identity of the oxidic
framework in these materials.

Small amounts of as-synthesized MSU-SIB, calcined MSU-B, and ethanol
extracted MSU-B materials were added to 0.05M HNOa. in no instance did the
materials peptize, as judged by the settling of the materials. It is possible that in
the case of the as-synthesized samples, the presence of surfactant restricts
access to the boehmite framework. If this is the case, then it cannot be
concluded that the framework in the MSU-SIB materials is boehmite. However,
after surfactant removal, neither calcined nor ethanol extracted MSU-B materials
peptize, indicating that these surfactant free materials truly are boehmites.

The sample prepared without surfactant was also tested for peptizability. it
was determined that the as-synthesized material is peptizable. That small particle

pseudoboehmite is formed in this synthesis is not an unreasonable result when

54

the hypothesis that the presence of surfactant allows for particle growth is
employed. After calcination, however, this product also does not peptize.
Therefore it is believed that calcination leads to sintering and growth of the
particles that governs the formation of boehmite as opposed to pseudoboehmite.
Calcination may play a dual role in the synthesis of MSU-B. Not only does this
lead to surfactant removal, but it also helps to ensure that boehmite is the phase

that is formed.

2.4. Mechanistic Aspects of MSU-B Alumina Synthesis

Bagshaw and Pinnavaia first reported the synthesis of mesoporous alumina,
named MSU-X aluminas. It was prepared using aluminum alkoxide and PEG
surfactants, but under different synthetic parameters and calcination procedures
than those used in the present work. In the MSU-X synthesis, the alkoxide to
surfactant ratio is greater than the ratio used here and the two reagents are
mixed homogenously prior to the hydrolysis step. The amount of water used in
the synthesis is considerably less as well. The addition of the water is done over
a 10-minute period, after which the reaction mixture forms a gel. in the MSU-B
synthesis, the alkoxide is added to an aqueous solution containing the surfactant.
The rate of hydrolysis and precipitation in this procedure is faster and a
precipitate rather than a gel is formed. Accordingly, the morphologies of the two
aluminas afforded are different. Amorphous MSU-X exhibits a wormhole motif
where the crystalline MSU-B has a scaffold morphology. It is believed that the

faster hydrolysis and precipitation in the MSU-B reactions are responsible for the

55

difference in morphology. In the MSU-B synthesis, boehmite particles are formed
first. These particles are directed to form a mesostructure and hydroxyls between
individual boehmite particles are condensed upon calcination, leading to the
scaffold structure. In this manner, the surfactant is not used as a rigid template,
but as a generic porogen used to impart increased porosity and surface area to
the alumina. Under the controlled hydrolysis procedure used in the MSU-X
synthesis, alkoxide molecules have time to interact with the surfactant molecules
prior to hydrolysis, allowing for the formation of the amorphous wormhole
structure. In this manner, the formation of MSU-X can be considered a templating
mechanism whereas the formation of MSU-B is governed by a particle assembly
mechanism. Particle assembly syntheses leading to other mesostructured oxides
have been reported, including Sn02'9, ZrOzzo, and mixed metal oxides”.

In the synthesis of mesostructured y-Al203 from aluminum salt precursors, it is
believed that the alumina phase first formed has amorphous framework walls like
MSU-X, which are converted to the crystalline boehmite walls of MSU-SIB upon
thermal aging”. However, in the syntheses of MSU-S/B from aluminum alkoxide
as reported here, no evidence exists to support such a claim. Inspection of the
wide angle powder XRD patterns of the MSU-SIB shows that at least in a
disordered state, formation of a boehmite phase occurs in the first minute after
addition of the alkoxide to the surfactant solution. Within 4 hours, the boehmite
phase is well organized and within 12 — 24 hours, the reaction and condensation

is complete, as shown in Figure 2.11.

56

 

 

 

lntensi

 

1m

 

 

rillllllllllfilTIITIIIIITIIITTYITIjT]

1o 20 30 4o 50 60 7o 80
Degrees (20)

Figure 2.11. Wide angle XRD patterns of MSU-SIB prepared from aluminum
sec-butoxide and dodecylamine at 20 °C showing evolution of the boehmite
phase over time. The presence of weak reflections assignable to boehmite after
only 1 minute show that the formation of MSU-B does not proceed through an
MSU-X type of intermediate with amorphous framework walls.

57

2.5. Conclusions

Reported here is what is believed to be the first mesostructured boehmite
material, MSU-B, prepared from the hydrolysis of aluminum sec-butoxide in the
presence of amine surfactants as porogens. The key in the synthesis strategy is
the formation of the MSU-SIB surfactant/boehmite precursor using surfactants
that are easily vaporized at low temperature calcination. it has been shown that
these materials are synthesized through a particle assembly pathway as
opposed to the templating mechanisms often used in the preparation of
mesostructured materials. Additionally, an MSU-X intermediate is not formed, as
boehmite is precipitated immediately. This affords products having a scaffold
morphology with an open framework comprised of boehmite walls. This open
framework allows for the realization of materials having surface areas of up to
497 m2/g, pore sizes of up to 6.8 nm, and pore volumes of up to 0.95 cc/g. The
materials are stable to calcination at 350 °C without conversion of boehmite to
transition alumina phases. Although surfactant can be removed by ethanol
extraction, calcination is the preferred method as it ensures that the inorganic
framework is boehmitic in nature. Further evidence that the MSU-B materials
consist of walls of the boehmite phase was provided from the lack of the ability to
peptize these materials.

The textural properties of MSU-B materials could be controlled only slightly by
altering the synthesis parameters. Increasing the mass of surfactant used in the
surfactant solution leads to increased surface areas and pore volumes. Surface

areas are relatively independent of the reaction temperature, but pore volumes

58

decrease with increasing temperature, signifying that less surfactant is
incorporated into the framework at higher temperatures. At a constant aluminum
to surfactant molar or mass ratio, however, increasing the surfactant size does
not correlate to an increase in pore size or volume. In fact, pore volumes and
pore sizes remain nearly constant over this series of experiments.

The textural properties of MSU-S/B materials are not controlled by surfactant
size. Also, the mass of surfactant used in any given synthesis does not match the
observed pore volume of the surfactant free product. Additionally, the peak in the
low angle region of the powder XRD patterns is not reproducible. From these
observations, it is concluded that MSU-S/B surfactant-boehmite compositions are
formed through a particle assembly pathway rather than a supramolecular

templating pathway.

2.6. Future Directions

Pseudoboehmite is actually the industrially preferred form of the aluminum
oxide hydroxide because it is conveniently dispersed into a sol and easily
introduced into other formulations or used in other applications. While the present
work has focused on the synthesis of boehmite and avoidance of
pseudoboehmite, a method to produce pseudoboehmite from a less expensive
precursor would be of considerable interest. Similarly, it would be interesting to
synthesize MSU-B materials from other aluminum sources as well.

Although pseudoboehmite can be formed without the aid of a surfactant, it

would interesting to know if a mesostructured (pseudo)boehmite sol could be

59

prepared. Another application of pseudoboehmite is that of a vaccine adjuvant23.
A mesostructured sol may provide for increased loading of vaccine, effectively
lowering the amount of aluminum administered per dose.

Boehmites are more often used as precursors to catalysts than as catalysts
themselves. Finding suitable catalytic applications for MSU-B would be a logical
direction in which to proceed. Since these materials are stable only under
relatively lower temperatures, they may be of utility as catalyst supports for
organic conversions under mild conditions.

There also exists another aluminum oxide hydroxide, a-AIOOH, diaspore. This
polymorph is formed at high temperatures and under high pressures. If it is
feasible to form a mesostructured diaspore, then this would also be of great
interest. Because diaspore is structurally similar to a-Al203, corundum, it is
converted directly to corundum at 550 °C. Prepared from other aluminas,
temperatures in excess of 1000 °C are required for this transformation. Thus a
mesostructured diaspore may lead to a mesostructured G-AI203 phase with
significantly improved textural properties.

Finally, while the target of the syntheses described in this work was
mesostructured boehmite, it is possible to calcine MSU-SIB (and MSU-B)
materials at elevated temperatures, converting them to mesostructured y—Al203.

This is one of the synthetic approaches discussed in the following chapter.

60

2.7. References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)
(11)

(12)

(13)

(14)

Misra, C. American Chemical Society Monograph, Vol. 184: Industrial
Alumina Chemicals, 1986.

Stamires, D.; Brady, M. F.; Jones, W.; Kooli, F. In US; (Akzo Nobel N.V.,
Neth.). Us, 2002, pp 12 pp., Cont.-in-part of US. Ser. No. 21,839,
abandoned.

Callender, R. L.; Harlan, C. J.; Shapiro, N. M.; Jones, C. D.; Callahan, D.
L.; Weisner, M. Fl.; MacQueen, D. B.; Cook, Fl.; Barron, A. Fl. Chemistry of
Materials 1997, 9, 241 8-2433.

Wefers, K.; Misra, C. "Oxides and Hydroxides of Aluminum,” Alcoa
Laboratories, 1987.

Yoldas, B. E. Am. Ceram. Soc., Bull. 1975, 54, 289-290.

Baker, B. R.; Pearson, R. M. J. Catal. 1974, 33, 265-278.
Guzman-Castillo, M. L.; Bokhimi, X.; Toledo-Antonio, A.; Saimones-
Blasquez, J.; Hernandez-Beltran, F. Journal of Physical Chemistry B
2001, 105, 2099-2106.

Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem, Int. Ed. Engl. 1996, 35,
1102-1105.

Vaudry, F.; Khodabandeh, S.; Davis, M. E. Chem. Mater. 1996, 8, 1451-
1464.

Yada, M.; Kijima, T. Zeoraito1997, 14, 104-111.

Valange, S.; Guth, J. L.; Gabelica, 2. Stud. Surf. Sci. Catal. 1998, 117,
517-518.

Cabrera, 8.; El Haskouri, J.; Alamo, J.; Beltran, A.; Beltran, D.; Mendioroz,
S.; Marcos, M. D.; Amoros, P. Adv. Mater. (Weinheim, Ger.) 1999, 11,
379-381.

Gonzalez-Pena, V.; Diaz, l.; Marquez-Alvarez, C.; Sastre, E.; Perez-
Pariente, J. Microporous Mesoporous Mater. 2001, 44-45, 203-210.

Zhang, Z.; Hicks, R. W.; Pauly, T. R; Pinnavaia, T. J. Journal of the
American Chemical Society 2002, 124, 1592-1593.

61

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309-
319.

Barrett, E. P.; Joyner, L. G.; Halender, P. P. Joumal of the American
Chemical Society 1951 , 73, 373.

Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.;
Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603-619.

Rouquerol, F.; Rouquerol, J.; Sing, K. Adsorption by Powders and Porous
Solids; Academic Press: London, 1999.

Severin, K. G.; Abdel-Fattah, T. M.; Pinnavaia, T. J. Chemical
Communications (Cambridge) 1998, 1471 -1472.

Pacheco, G.; Zhao, E.; Garcia, A.; Sklyarov, A.; Fripiat, J. J. Journal of
Materials Chemistry 1 998, 8, 219-226.

Wong, M. S.; Jeng, E. S.; Ying, J. Y. Nano Letters 2001, 1, 637-642.

Zhang, Z.; Pinnavaia, T. J. Journal of the American Chemical Society
2002, submitted.

Hem, S. L.; White, J. L. PHARMACEUTICAL BIOTECHNOLOGY FIELD
1995, 6, 249-276.

62

Chapter 3

Synthesis and Characterization of Mesostructured y-A1203

3.1. Introduction

Transition aluminas are disordered crystalline phases formed through the
thermal dehydration of aluminum hydroxides and oxyhydroxides‘. These oxides
are used as adsorbents and catalysts or catalyst supports in many chemical
processes, including the cracking, hydrocracking, and hydrodesulfurization of
petroleum feed stockse. At least seven transition aluminas have been reported.
Of these, y-Al203 is perhaps the most important tor catalytic applications. The
utility of y-Al203, as well as other transition aluminas, can be traced to a favorable
combination of textural properties (i.e., surface area, pore volume and pore size)
and acid - base characteristics.

The structure of y—Al203 is based on that of a tetragonally distorted defect
spinel (MgAl204) structure. The unit cell of spinel contains 8 formula units. The 32
oxygen atoms are arranged in a cubic close packed lattice. in spinel, the 8
aluminum ions occupy octahedral positions and the 8 magnesium ions occupy
tetrahedral positions. The presence of trivalent aluminum ions (in y—Al203) in lieu
of divalent magnesium ions (in spinel) necessitates that only 5 1/3 of the original
8 tetrahedral sites be occupied per unit cell in the y—AI203 structure. Overall, the

ratio of the number of aluminum ions in octahedral to tetrahedral sites is 3:1.

63

Recently, density functional theory calculations have revealed the presence of
hydrogen in the spinel lattice of y-Al2033.

The y-Al203 phase is formed upon the dehydration of the aluminum
oxidehydroxide boehmite at temperatures ranging from 400 to 700 °C. During this
process, porosity is generated as water is eliminated from the structure, creating
cracks and void space in the alumina. Along the progression from hydroxide
phases to oxide phases, there is an initial rise in surface area and porosity.
However, heating at temperatures above 500 °C leads to diminished textural
properties. Conventional forms of y—Al203 typically exhibit a BET surface area
below 250 mZ/g and a pore volume less than 0.50 cc/gz. The properties of y-
Al203 for applications in catalysis and adsorption are determined in large part by
these textural parameters.

In view of the recent advancements realized for mesostructured forms of
silicas4'5, similar improvements may be anticipated in the properties of
mesostructured aluminas. However, comparatively limited progress has been
made in this area. Since Bagshaw’s initial disclosure of the supramolecular
synthesis of mesostructured alumina from aluminum alkoxides and PEG
surfactantss, others have synthesized similar materials using both ionic7'9 and

10’“ surfactants as structure directors. A common feature of all these

non-ionic
mesostructured aluminas is the amorphous framework walls that limit their
thermal and hydrothermal stability, greatly compromising their usefulness in
potential catalytic applications. Recently, a mesostructured alumina containing a

transition alumina phase has been reported”. However, the true composition of

this material remains unclear without evidence of a boehmite precursor, which,
as described below, is the crucial step leading to the formation of mesostructured
y-AI203.

This work reports on the synthesis, characterization, and stability of crystalline
mesostructured y-aluminas prepared from the hydrolysis of aluminum sec-
butoxide in the presence of either PEO or amine surfactants as porogens. The
novelty of the synthesis is the resulting formation of a mesostructured
surfactant/boehmite composition, denoted MSU-SIB. Upon thermal
transformation, MSU-SIB is converted to crystalline, mesostructured y-Al203,
denoted MSU-y. It will be demonstrated that these materials are structurally
stable and afford textural properties exceeding those of conventional transition

aluminas.

3.2. Experimental

3.2.1. Synthesis

The assembly of the MSU-SIB surfactant-boehmite mesostructure was
achieved by hydrolysis of aluminum sec-butoxide in an alcoholic aqueous
solution of the desired surfactant. Syntheses could be performed over a wide
range of reaction conditions. Two general routes to mesostructured y-Al203 are
identified in sections 3.2.1.1 and 3.2.1.2 below. The surfactants selected for

study are as follows:

65

Surfactant Formula

Pluronic P84 (EO)19(PO)39(EO)19
Pluronic P65 (EO)19(PO)30(EO)19
Pluronic L64 (EO)13(PO)30(EO)13
Pluronic L61 (EO)2(PO)30(EO)2

Tergitol 15-s-9 C7H15CH2((EO)90H)CyH15
Tergitol 15-S-12 C7H15CH2((EO)120H)CyH15

3.2.1.1. Synthesis from PEO Surfactants

A typical MSU-y synthesis using a PEO surfactant as a porogen was carried
out as follows. Pluronic P84 and aluminum sec-butoxide were dissolved in 2-
butanol. To this solution water was added to hydrolyze the alkoxide. The solution
was stirred on a magnetic stirrer until homogenous, sealed in a glass vessel, and
placed in a 100 °C oven under static conditions for a period of ~20 hours. The
resulting MSU-S/B was recovered by vacuum filtration, washed with 2-butanol
and water, and air dried. Conversion to MSU-y was achieved by calcining the as-
synthesized MSU-S/B at 500 “C for 4 hours. Alternatively, during the synthesis,
the water, alcohol, and surfactant could be mixed first, adding the alkoxide last.
Reaction mixtures in syntheses involving lesser amounts of solvent would
become quite thick. in these instances, a laboratory grade blender was used to

mix the reactants.

66

3.2.1.2. Synthesis from Amine Surfactants and MSU-B Materials

A second method used to synthesize MSU-y materials is to calcine either the
MSU-SIB surfactant-boehmite compositions made from amine surfactants or
surfactant free MSU-B boehmites, described in the previous chapter, at

temperatures ~500 9C. Both methods afford mesostructured MSU-y aluminas.

3.2.1 .3. Synthesis of Doped MSU-y Aluminas

MSU-y aluminas were also doped with metal ions. In these syntheses, the
desired amount of dopant was administered as a salt which was dissolved in the
2-butanol and water surfactant solution. La(N03)3 - 9 H20 and (NH4)2Ce(N03)6
were used to supply La and Ce, respectively, in the doping reactions. The
processing of the reaction mixtures was performed as described in section

3.2.1.1 above.

3.2.2. Characterization of MSU-y Aluminas

Powder diffraction patterns were collected on a Rigaku Rotoflex
Diffractometer using Cu K, radiation (A = 0.154 nm). Intensity was calculated in
counts per second every 0.02 degrees 26. Sweep rates were varied from 2 - 8
degrees ZBlmin.

Nitrogen adsorption - desorption isotherms were collected on a Micromeritics
TriStar 3000 sorptometer at 77K. Samples were degassed at least 12 hours at
150 °C under a vacuum of 10’6 torr prior to analysis. Surface areas and pore

sizes were determined by the BET13 and BJH“ methods, respectively.

67

TEM images were collected on a JEOL 1OOCX microscope with a 0936
filament at an accelerating voltage of 120 kV. Samples were prepared by
sonicating a suspension of the powder in ethanol for 20 min. One to two drops of
the suspension were then placed on a carbon coated holey film that was
supported on a 3mm, 300 mesh copper grid.

27Al NMR spectra were collected on a Varian 400 VRX solid — state NMR
spectrometer. Samples were placed in a zirconia rotor and spun at 4 kHz with a
pulse width of 0.5 ps and a pulse delay of 0.5 s. Chemical shifts were referenced
to AI(H20).33+ which is assigned a value of 0 ppm.

Steam testing was performed by passing 20% H20 in nitrogen over the
alumina at 600 ‘-’C for 5 hours. Exposed aluminas were air dried before further

analysis.

3.3 Results and Discussion

3.3.1. MSU-y Aluminas from PEO Surfactants
3.3.1.1. General Properties of Materials

Representative wide angle powder XRD patterns of an as-synthesized MSU-
S/B surfactant-boehmite composition made from aluminum sec-butoxide and P84
and the corresponding MSU-y alumina are presented in Figure 3.1. In the case of
these materials, peaks assignable to boehmite and y-Al203 for MSU-SIB and
MSU-y, respectively, are evident. As was the case for the surfactant-free MSU-B

boehmite materials described in the previous chapter, a low angle diffraction

68

 

MSU-y

440
311 40°

 

 

Intensity

MSU-SIB

 

 

 

 

 

 

 

 

 

 

 

illlllfiTlrrlrlIIIWTIUWITIIYTjIUIIII

10 20 30 4o 50 so 70 so
Degrees (26)

Figure 3.1. Representative wide angle powder XRD patterns of an MSU-SIB
surfactant-boehmite composition made from aluminum sec-butoxide and the
PEO surfactant P84 under reaction conditions described in Table 3.1, line 3 and

the corresponding MSU-y alumina prepared from calcination of the as-
synthesized MSU-S/B at 500 9C for 4 hr.

69

peak in the XRD patterns was not always observed. However, when low angle
peaks were observed, the d-spacings of these peaks have recurring values in
multiples of ~8 A. The most common values, 48 and 56 A, correspond to an
average of 6 to 7 building units in the particle.

Nitrogen adsorption — desorption isotherms and pore size distributions of a
representative MSU-y alumina and a similar alumina prepared without surfactant
(blank) are shown in Figure 3.2. it can be seen that the pore volume and pore
size are increased in the MSU-y as compared to the material prepared in
absence of surfactant. Porosity in conventional y-AI203 is generated as cracks
and voids form upon the dehydration of the oxidehydroxide precursor, so it is
reasonable to expect some adsorption in the blank sample. However, use of
surfactants as porogens imparts considerably larger pore volumes and sizes, as
is reflected in the isotherms. Typical surface areas, pore sizes, and pore volumes
for MSU-y aluminas range from 300 — 380 m2/g, 10 — 20 nm, and 0.75 — 1.6 cc/g,
respectively. These same values for the blank sample are 259 m2/g, 5 nm, and
0.47 cc/g, which are in line with those expected for textural properties of
conventional y-aluminas. The textural properties of these materials made from
PEO surfactants are summarized in Table 3.1.

In addition to the XRD data, further evidence of the presence of y-Al203 in the
framework walls was provided by the 27Al NMR spectrum of MSU-y. As described
in section 3.1, the unit cell of y-Al203 contains aluminum ions in both octahedral
and tetrahedral coordination, in a ratio of 3:1. Accordingly, 27AI NMR spectra of y-

Al203 show peaks consistent with this distribution of aluminum ions”.

70

 

1 000

dVIdD

      

 

- o 5 1o 15 20 25 so
Pore Diameter (nm)

Volume Adsorbed (cc/g)

Blank

 

 

 

PIP

Figure 3.2. Nitrogen adsorption-desorption isotherm and BJH adsorption pore
size distribution (inset) of an MSU-y alumina prepared from aluminum sec-
butoxide and the PEO surfactant P84 as described in Table 3.1, line 3. included
for comparison is the same data for an alumina prepared without surfactant as
described in Table 3.1, line 20. Isotherms are offset 200 cc/g for clarity and PSD
curves are displayed in the same order as the isotherms.

71

Table 3.1. Synthetic parameters and textural properties of MSU-y aluminas
prepared from aluminum sec-butoxide and PEG surfactants.

 

 

 

 

 

 

Al/Surl/s-BuOH/HZO sag Pore Size” Pore Volume
Surfactant

Molar Ratioa (m2/g) (nm) (cc/g)
1 P84 100: 1 :27 : 544 334 15 1.17
2 P84 75 : 1 :27 : 544 332 17 1.23
3 P84 66: 1 :27 : 544 378 13 1.09
4 P84 50 : 1 :27 : 544 365 15 1.25
5 P84 25 : 1 :27 : 544 366 17 1.39
6 P84 66 : 1 :249 : 544 339 10 0.76
7 P84 66 : 1 : 160 : 544 341 13 0.93
8 P84 66 : 1 : 107 : 544 381 4, 17 0.84
3 P84 66 : 1 :27 : 544 378 13 1.09
9 P84 66 : 1 : 12 : 544 342 14 1.06
10 P84 66 :1 :0 : 544 325 10 0.92
3 P84 66: 1 :27 : 544 378 13 1.09
11 P84 66 : 1 :46 : 933 355 11 1.14
12 P84 66 : 1 :65 : 1322 338 11 1.04
13 P84 66:1 :103 2100 338 12 1.12
14 P84 66: 1 :218 : 4433 314 9 0.85
15 15-S-9 66 : 1 : 17 : 285 367 20 1.63
16 15-S-12 66 : 1 : 14 : 232 368 20 1.66
17 L61 66 : 1 :41 :670 279 12 0.78
18 L64 66 :1 :70 : 1160 300 10 0.76
19 P65 66 : 1 :83 : 1360 324 6 0.69
20 None 66 : 0 : 65 : 1322 259 5 0.47

 

‘ Due to the large molar masses of PEO surfactants, molar ratios of aluminum alkoxide to PEO
surfactants are much higher than when amine surfactants are used. The mass ratios, however,

are comparable (section 3.3.1.2).

b BJH pore diameter as measured by the adsorption branch

72

The 27Al NMR spectrum of MSU-y is provided in Figure 3.3. Two resonances
are observed, one at 72 ppm and another at 6.7 ppm, corresponding to the four-
and six- coordinate aluminum, respectively. From the relative intensity of these
peaks, it can be seen that there is good agreement between the observed
tetrahedral occupancy with the theoretical values. Furthermore, in the NMR
spectra of amorphous aluminas, an additional resonance at ~35 ppm is often
observed, which is attributable to 5- coordinate aluminum centers. The lack of
any such resonance in the spectrum of MSUJY is evidence that no appreciable
amount of amorphous alumina exists in this material.

Perhaps the best evidence for the confirmation of a mesostructured alumina
with crystalline walls is attained from TEM. Bright field and dark field images of
MSU-y prepared from P84 and aluminum sec-butoxide are shown in Figure 3.4.
The same fiber like morphology evident in MSU-B is also seen here. This
indicates that there is retention of morphology on going from the boehmite to the
gamma phase. From the image in Figure 3.4 A, it appears that the randomly
oriented fibers have a length of approximately 30 nm and an average thickness
of 5 nm. The porosity of MSU-y alumina is defined by the space between the
overlapping fibers.

In part B of Figure 3.4, a lower magnification image and the corresponding
selected area electron diffraction pattern is provided for an MSU-y prepared from
aluminum sec-butoxide and P84 as described in Table 3.1, line 3. The presence

of the rings in the diffraction pattern, corresponding to hkl reflections of [311],

73

6.72

 

71 .96

 

I 11" [III 1111 III III]
I I I I W

IIII II IIIIIIIII
200 150 100 50 0 -50 -100 -150 -200
mm

Figure 3.3. 27AI NMR of calcined (500 9C) MSU-y alumina prepared from
aluminum sec-butoxide and P84 as described in Table 3.1, line 3. The two
resonances at 6.7 ppm and 72 ppm are assigned to aluminum in octahedral and
tetrahedral coordination, respectively. The peaks are in the approximate 75:25
ratio consistent with y-Al203.

74

Figure 3.4. Bright field TEM image of (A) calcined (500 °C) MSU-y alumina
prepared from aluminum sec-butoxide and P84 as described in Table 3.1, line 3.
(B) Lower magnification image of the same product and the corresponding
electron diffraction pattern, with diffuse rings assignable to the [311], [400], and
[440] crystallographic planes. (C) Dark field image of the same particles as in (B),
but formed from an ~7° arc of the [440] diffraction ring. The crystal orientations
contributing to the [440] ring appear as bright sports in the image.

75

 

76

[400], and [440], is a direct consequence of the polycrystalline nature of the
material. Their presence confirms the crystalline nature of these materials. Dark
field imaging in TEM is accomplished by using only diffracted electrons to form
images. Additionally, an aperture can be inserted into the electron beam to select
only electrons that are diffracted by specific crystallographic planes. Part C of
Figure 3.4 displays such an image, the dark field image of the same particles as
in Part B, created from ~79 arc of the [440] diffraction ring of MSU-y. The crystal
orientations contributing to the [440] ring of the diffraction pattern appear as
bright spots in the dark field image. The appearance of these bright spots

uniformly across the image confirms that y—Al203 nanoparticles comprise the

framework walls in MSU-y.

3.3.1 .2. Variations of Synthetic Parameters

As was discussed in the previous chapter for mesoporous boehmites
prepared from aluminum sec-butoxide and amine surfactants, these new MSU
aluminas are synthesized through a particle assembly pathway. Consequently,
controlling the textural properties of the products as is practiced with templated
mesostructured materials is not straightfonrvard. Although the pore size
distributions do not respond to differences in surfactant size, the y-Al203
compositions reported here can be made over a range of synthetic conditions
while exhibiting textural properties in excess of for conventional aluminas. The

textural properties are summarized in Table 3.1.

77

The aluminum to surfactant ratio was altered to range from ratios Al/P84 =
100 to Al/P84 = 25. It should be noted that the molar ratios of Al to surfactant
employed in the syntheses leading to the formation of MSU-y aluminas are
considerably greater than in the syntheses leading to the formation of MSU-B
boehmites. The molar masses of the PEO surfactants are generally much larger
than those of the amine surfactants. While the reported molar ratios differ greatly,
the mass ratios of the two reagents are comparable. For example, at AVP84 =
66, the mass of aluminum sec-butoxide to the mass of P84 = 3.91. For
comparison, an AVDDA = 5 corresponds to an aluminum sec-butoxide to DDA
mass ratio of 6.65.

While the textural properties of the alumina prepared at Al/Surf =66 appear to
be out of line, there is a general increase in both surface area and pore volume
with decreasing Al/Surf ratios (Table 3.1, lines 1-5). There is no apparent trend in
pore size. While aluminas prepared at other ratios appear better, most syntheses
were carried out at an Ai/Surf = 66 ratio for two reasons. First, reaction mixtures
at high surfactant concentrations (low Al/Surf ratios) tend to be very thick, sticky,
and difficult to process. Second, at the time of this work, the origin of the low
angle diffraction peak was unclear, and aluminas exhibiting a low angle peak
were expected to have more desirable textural properties than those that did not.
The differences in textural properties for reaction products that exhibit a low
angle XRD reflection and those that do not are not substantial. Perhaps there
exists a certain ratio where there is an optimum coverage of surfactant by

alumina, above which any additional alumina is not mesostructured into new

78

fibers, but rather goes toward increasing the thickness of existing MSU-y fibers.
Void space that would contribute to pore volume would, above this ratio, be
sacrificed as alumina fibers thickened. Nevertheless, syntheses carried out at
this ratio proved fruitful.

Recently, the synthesis of y-Al203 nanofibers prepared from NaAlOz and PEG
surfactants was reported”. Molar ratios of aluminum to surfactant in syntheses
leading to the y-AI203 fibers were ~ 2:1, much lower than the ratios employed in
the current work. Although no claims of a mesostructure were made, fibers of
comparable length and thickness to those reported here were obtained.
Additionally, the large amount of surfactant used in the syntheses provided for
the formation of an alumina with pore volumes in the range of 1.41 - 1.95 cc/g. A
direct relationship between mass of surfactant used in the synthesis and pore
volume of the resulting product was noted in Zhu’s work. These findings are not
in agreement with observed trends in the syntheses of MSU-B boehmites
reported in the previous chapter and in the syntheses of MSU-y aluminas, as
described in the following paragraphs.

Solvent polarity was also adjusted to examine the effect on the textural
properties of MSU-y aluminas. From the nitrogen adsorption - desorption
isotherms and pore size distributions in Figure 3.5 and the data in Table 3.1,
lines 6 — 11, it can be seen that the textural properties go through a maximum for
aluminas prepared at a water: 2-butanol ratio of 4 : 1 (80/20 v/v). It could be that
as the polarity decreases at high a 2-butanol content, the surfactant ceases to

form micelles capable of directing the structure. At the other extreme, of low 2-

79

 

 

 

 

8
‘ S 0%I
In
2000 - /77
a _ /////‘19%
B d 0 5 10 15 20 25 30 :/// ,f/
3 : Pore Diameter (nm) .._...,.- , 20%
.8 1500‘
P I
L
O
m
'5
<
0)
E
2
O
>

 

 

 

PIP

Figure 3.5. Nitrogen adsorption-desorption isotherms of MSU-y aluminas
prepared from aluminum sec-butoxide and P84 as described in Table 3.1, lines 3
and 6-10 showing the effect of solvent polarity on the textural properties. The
volume percent of 2-butanol in the solvent is labeled for each isotherm.
isotherms are offset 300 cc/g for clarity. The BJH adsorption pore size
distributions (inset) are in the same order as the isotherms.

80

butanol content, the total amount of solvent may be decreasing to point that is
not conducive to mesostructure formation.

Nitrogen adsorption - desorption isotherms and pore size distributions for
MSU-y aluminas prepared from aluminum sec-butoxide and P84 solutions of
varying concentration are shown in Figure 3.6. It was observed that the surface
areas, pore volumes, and pore sizes decreased slightly with decreasing
concentration of the surfactant employed in the synthesis. The effect of
surfactant concentration is most pronounced at a concentration of 5%. Aside
from ease of processing the reaction mixture, there seems to be little advantage
in diluting the reaction mixture. The surfactant concentration does not appear to
be a critical factor affecting the textural properties and pore structure. This is the
same phenomenon that was observed in the syntheses of MSU-SIB precursors
and MSU-B boehmites from aluminum sec-butoxide and amine surfactants as
described in the previous chapter.

Finally, MSU-y aluminas could be prepared with other nonionic PEO
surfactants in which Figure 3.7 shows the nitrogen adsorption — desorption
isotherms and the corresponding pore size distributions for these aluminas. It
seems odd that the smallest molecular weight surfactants yield aluminas with the
largest pore volumes and sizes. However, in no case does the amount of
surfactant used in the synthesis correspond to the pore size or volume observed
in the surfactant free products. Again, the lack of a direct relationship between
surfactant properties and textural properties of MSU-y aluminas is evidence of a

particle assembly pathway.

81

 

 

    

  
 

a N 2

. , [5%
’15 I /
\ _
3 150° 0W.
v _ Pore Diameter (nm) _
13 fl/x"/”
01
.9
L-
O
m
'0
<
0)
E
2
O
>

 

 

 

Figure 3.6. Nitrogen adsorption-desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-y aluminas prepared from aluminum sec-
butoxide and P84 as described in Table 3.1, lines 3 and 11 -14 showing the effect
of surfactant concentration (mass P84 / mass solvent) on the textural properties.
The mass percent of P84 is labeled for each isotherm. isotherms are offset 300
cc/g for clarity. The PSD curves are displayed in the same order as the
corresponding isotherms.

82

 

2000

 

 

Volume Adsorbed (cc/g)

 

 

 

 

PIP

Figure 3.7. Nitrogen adsorption-desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-y aluminas prepared from aluminum sec-
butoxide and differing PEO surfactants as described in Table 3.1, lines 3 and 11-
14. isotherms are offset 200 cc/g for clarity. The PSD curves are displayed in the
same order as the corresponding isotherms.

83

3.3.1.3. Effect of Calcination Procedure

As described in the previous chapter, the lower boiling points of the amine
surfactants allowed for their removal from the as-synthesized MSU-S/B
surfactant-boehmite compositions at mild temperatures with the retention of the
boehmite phase during MSU-B synthesis. Such is not the case with the PEO
surfactants. Temperatures above 400 9C are required for surfactant removal via
calcination, causing the boehmite in the MSU-SIB hybrid composition to be
converted to y—Ai203.

y-Al203 is but one of at least seven identified transition aluminas. Calcination
at elevated temperatures may be a general route to transition aluminas. The
transformation sequence of thermal dehydration of aluminas dictates that 6-Al203
is the next phase to form once y-Al203 is heated at temperatures above ~750 QC‘.
MSU-SIB aluminas prepared from PEO surfactants were calcined at 800 9C for
2hr to achieve the conversion to a mesostructured 6-Al203, denoted MSU-6. This
transition alumina also has a structure based on the spinel structure and
consequently, evolution from y-A|203 to 6-Al203 can be performed without
significant changes in the pore structure of the alumina‘m’. Therefore, it is not
unreasonable to expect the retention of desirable textural properties in these
materials, as evidenced from the nitrogen adsorption — desorption isotherms
(Figure 3.8). Due to the similarities in 6- and y-AI203, it is somewhat difficult to
distinguish between these two transition aluminas. XRD diffraction can be used
to some extent to differentiate phases, but the patterns of all transition aluminas

are strikingly similar. A better tool is to use is 27Al NMR, however in this case, site

84

occupancy of Al3+ ions is nearly identical. The inset of Figure 3.8 provides the
wide angle XRD pattern of the reaction product formed through the calcination of
an MSU-SIB surfactant-boehmite composition at 800 QC. The case could be
made that these peaks are assignable to 6—Al203. A comparison of textural
properties of these two mesostructured aluminas (Table 3.2) reveals that while
the pore volumes and sizes are comparable, there is a marked decrease in the
surface area. This could be due to sintering of y-Al203 fibers upon transformation
to 6-Al203. The 27AI NMR spectra (not shown) looks nearly identical to that of the
MUS-v alumina. Due to the lack of definitive proof of 0-A1203 phase formation, it
could be argued that the material claimed here to be a mesostructured 6-AI203
may only be a y- Al203 whose pore structure is beginning to collapse.
Regardless, this material exhibits exceptional textural properties having been

calcined at 800 9C.

3.3.2. MSU-y Prepared from Amine Surfactants

MSU-y aluminas could also be prepared from amine surfactant-boehmite
compositions synthesized as described in the previous chapter. This was
achieved by directly calcining MSU-S/B surfactant-boehmite precursors, or the
surfactant-free MSU-B boehmite aluminas that had been previously calcined, at
temperatures sufficient to initiate conversion to y-AI203 (>400 ‘2C). The resulting
products exhibited XRD patterns consistent with y-AI203, signifying the crystalline
nature of these materials. As expected, these materials have the same

morphology as other MSU-y aluminas. The bright field TEM image and

85

 

700

 

 

'I
600- a:
"a
“C
‘2
-.s
500-
400-

Volume Adsorbed (cc/g)

 

MSU-8

 

10 20 30 40 50 60 70 00
Degrees (26)

 

 
      

 

 

0 I I I I I I I I I I I l I I I I I T T

0.6

0 0.2 0.4
PIP

O

0.8

1

Figure 3.8. Nitrogen adsorption-desorption isotherm and wide angle powder
XRD pattern (inset) of a calcined (800 QC) MSU-6. The initial MSU-SIB
surfactant-boehmite intermediate was prepared from aluminum sec-butoxide and
the PEO surfactant P84 under conditions described in Table 3.1, line 13.

86

Table 3.2. Comparison of textural properties of calcined MSU-y formed at 500 9C
and calcined MSU-6 alumina formed at 800 90a

 

 

Calcination . b
Surface Area Pore Size Pore Volume
Temperature ( 2/ ) ( ) (cc/ )
rn nm

(QC) 9 9
1 500 306 18 1.14
800 251 16 0.98
2 500 338 12 1.04
800 249 16 1.00

 

‘3 The initial MSU-SIB surfactant-boehmite intermediate was prepared from
aluminum sec-butoxide and the PEO surfactant P84 under conditions described
in Table 3.1, line 13.

b BJH pore diameter as measured by the adsorption branch

87

corresponding electron diffraction pattern shown in Figure 3.9 for an MSU-y
prepared from an amine-mediated MSU-B boehmite verified that an open
framework was created upon the assembly of the small crystalline particles into a
mesostructure. Nitrogen adsorption — desorption isotherms (Figure 3.10) also
show behavior similar to MSU-y aluminas derived from PEO surfactants. The
textural properties of these materials are summarized in Table 3.3. The pore
volumes and pore sizes are approximately equal (at a given temperature) for
MSU-y aluminas prepared directly from MSU-SIB or from previously calcined
MSU-B. The surface areas are greater for MSU-y aluminas prepared through the
calcination of MSU-SIB than for MSU-B precursors. Perhaps there is some
structural degradation of MSU-B caused by adsorption of water onto the surface,
rendering a poorer quality MSU-y to be formed upon a second calcination. it was
observed that pore volumes for MSU-B materials decreased at higher aging
temperature (Table 2.1). That trend is observed here as well. Finally, it is noted
that data provided in Table 3.3 for the MSU-Iy aluminas from MSU-SIB and MSU-
B indicate that the textural properties are enhanced compared to those in MSU-y

aluminas prepared from PEO surfactants.

88

 

Fig
boe
COIT
at a
tem
by c
Dam
IIIGS
and

Figure 3.9. TEM image of MSU-y prepared through the calcination of MSU-B
boehmite at 500 °C for 4 hr. The initial MSU-S/B surfactant-boehmite
composition was prepared from aluminum sec-butoxide and dodecylamine (DDA)
at a reaction stoichiometry of Al:DDA:EtOH:H20 = 5:1:68.0:73.4 and a reaction
temperature of 20 °C. The surfactant free MSU-B boehmite phase was obtained
by calcining the MSU-SIB precursor at 325 °C for 10 hr. The electron diffraction
pattern (inset) demonstrates the crystallinity of the framework of the
mesostructure, as indicated by the diffuse rings attributable to the [311], [400],
and [440] crystallographic planes of y—Ai203.

89

 

9O

 

dVIdD

 

  

a .1

3 1200 -

3 Porse Diameter (Iism)

'U .

O

.O -

L

O -

8

o .

E

2 .1

O

> -
400 -

 

 

 

o
0
re
.°
4:.
0
ca
o
on
—L

Figure 3.10. Nitrogen adsorption-desorption isotherms and BJH adsorption pore
size distributions (inset) for MSU-y aluminas prepared through the calcination
(500 9C) of amine-mediated MSU-S/B surfactant-boehmite precursors (samples
A and C) and surfactant free MSU-B precursors (samples B and D). The
AlzDDAzEtOHzH20 reaction stoichiometry was 5:1:68.0:73.4 for all samples.
Reaction temperature was 20 9C for samples A and B and 45 ‘2C for samples C
and D. Isotherms are offset 300 cc/g for clarity. The PSD curves are displayed in
the same order as the corresponding isotherms.

91

Table 3.3. Textural properties of MSU-y aluminas prepared through the
calcination of MSU-SIB surfactant-boehmite compositions and MSU-B boehmites
at 500 9C.”

Surface Area Pore Sizeb Pore Volume

 

Precursor
("IQ/9) (nm) (cc/9)
A MSU-S/B 440 7 1 .07
B MSU-B 385 7 0.97
C MSU-S/B 529 3 0.86
D MSU-B 431 4.5 0.89

 

a The MSU-S/B and MSU-B precursors were prepared from aluminum sec-
butoxide and dodecylamine. The AlzDDAzEtOHzH20 reaction stoichiometry was
5:1:68.0:73.4 for all samples. Reaction temperature was 20 ‘2C for samples A and
B and 45 9C for samples C and D.

b BJH pore diameter as measured by the adsorption branch

92

3.3.3 Stability of MSU-y Aluminas

3.3.3.1 Doped MSU-y Aluminas

it is known that the incorporation of certain elements into y—Al203 has a
stabilizing effect on the structure. While transition metals have been doped into
aluminas toward this end, more commonly either lanthanum or cerium is used as
the dopant. It has been demonstrated that these ions stabilize the y-A1203 phase

”2‘ on the

toward conversion to d-Al203 by forming a metal aluminate species
surface of the alumina. Due to the nature of the spinel structure of y-Al203 and
other transition aluminas, metal ions can also stabilize alumina by filling the
vacancies in the iatticezz'za. Most studies are aimed at delaying the temperature
at which y-Ai203 is converted to G-A1203 or other transition aluminas. Many
important industrial catalytic applications, of alumina are performed in a steam
environment, so aluminas must also be stable under these conditions if they are
to function effectively. A study of the ability of lanthanum to stabilize alumina in a
steam environment suggested that the lanthanide aluminate mechanism is at
work here as well“.

MSU-X aluminas were doped with rare earth metals to improve their thermal
stability”. While MSU-y aluminas have proven to be thermally stable, improving
their hydrothermal stability would increase the possibility for effective use in
catalytic applications. It was also possible to dope the MSU-y aluminas of the

present work with cerium and lanthanum. As shown in Table 3.4, MSU-y

aluminas could be doped with an assortment of metal ions with the products

93

Table 3.4. Textural properties of doped MSU-y aluminas prepared from
aluminum sec-butoxide and P84. in each synthesis, the dopant was administered
as a metal salt which was dissolved in the surfactant solution prior to addition of
the aluminum sec-butoxide.

 

 

Surface Area Pore Sizea Pore Volume
Dopant 2

(m /0) (nm) (cc/0)
Cu 437 1 0 1 .05
Ni 426 1 0 1 .08
Cr 438 3-1 5 0.96
Co 390 3-1 5 0.88
Fe 380 1 0 0.93
Mn 392 7 0.81
Ba 322 1 3 0.91
La 386 7 0.68
Ce 41 6 9 0.99

 

8‘ BJH pore diameter as measured by the adsorption branch

94

exhibiting favorable textural properties. While the introduction of transition metals
may introduce catalytically active sites to MSU-y aluminas, steam testing (below)
was conducted for the lanthanum and cerium doped materials, keeping with
conventional methodologies.

It was concluded that in amorphous MSU-X aluminas, direct incorporation of
rare earths into the framework was a feature of the formation. The method of
incorporation in MSU-y, however, may involve both the formation of an aluminate
coating and the incorporation of the dopant into vacancies of the y-Al203 lattice.
Peaks assignable to a CeAlOa phase in the wide angle XRD pattern of 5%
cerium doped MSU-y indicates that formation of an aluminate surface species is
involved during the incorporation process (Figure 3.11). However, no such peaks
are evident in the pattern of the 5% lanthanum doped material, indicating that
direct substitution into the spinel lattice of y-Ai203 may also be involved.

Nitrogen adsorption — desorption isotherms (Figure 3.12) indicate that the
pore volume of MSU-y is greater in the pristine sample than in either doped
variation, although the pores are smaller and more uniform in the doped
materials. The TEM images of these materials (Figure 3.13 A) reflect this as well;
larger pores are evident in the clean sample (A) than in the lanthanum doped (C)

or cerium doped (E) sample.

3.3.3.2. Steam Testing of MSU-Iv Aluminas
These three aluminas were tested for hydrothermal stability by exposing each

to a 20% steam atmosphere at 600 SC for 5 hr. Interestingly, the pristine MSU-y

95

 

 

 

 

 

intensity

 

Pristine

 

 

 

 

 

jIII rIII IIII IIII IIII IIII IIII
I I I I I I I

10 20 30 40 50 60 70 80
Degrees (26)

Figure 3.11. Wide angle powder XRD patterns of MSU-y aluminas prepared from
aluminum sec-butoxide and P84: pristine and doped with either lanthanum or
cerium. Peaks marked with an asterisk are attributable to a cerium aluminate
phase, CeAlOa. The doped MSU-y aluminas were prepared through the
calcination (500 9C) of doped MSU-SIB precursors prepared as described in
Table 3.1, line 13.

96

 

    

 

 

  

 

 

 

. 5% Ce
1400' a
- 2
>
. 1:
A 1200-
m - m
E I 0 P10 DIO :0 40 50
z 1000- ore ame er (nm) 77
g I 5% La
3 u
in 800-
.0 .
< 2
“E’ 600-
2 I
g I
400-
200_ Pristine
o I I I I I I I l l I I l I l I I I I I
0 0.2 0.4 0.6 0.8 1
P1P
0

Figure 3.12. Nitrogen adsorption-desorption isotherms and BJH adsorption pore
size distributions (inset) of MSUJY aluminas: pristine and doped with either
lanthanum or cerium. The doped MSU-y aluminas were prepared through the
calcination (500 °C) of doped MSU-SIB precursors prepared as described in
Table 3.1, line 13. isotherms are offset 400 cc/g for clarity. The PSD curves are
in the same order as the corresponding isotherms.

97

alumina appears to have the best hydrothermal stability. it retains 85% of the
original surface area while the lanthanum doped and cerium doped MSU-y
aluminas retain only 57% and 55%, respectively.

A major concern under these conditions is the sintering of the nanoparticles. A
comparison of TEM images of the three samples before steaming (Figure 3.13)
reveals that there is more void space between fundamental building units of the
framework in the undoped alumina(A). it may be that the greater distance
between particles keeps them from sintering. The particles of 5% La doped (C)
and Ce doped (E) aluminas may be more easily sintered as they are initially in
closer proximity. The TEM images of the steamed aluminas (B,D,F) support this
view. While still lath-like in morphology, the particles appear shorter and thicker
in all cases, but more so for the doped aluminas than for the clean sample.

Sintering leads to loss of surface area, which is observed in all cases. it is
possible that doping of aluminas leads to the formation of micropores in the
structures, which contribute greatly to the surface area. Upon steaming and
subsequent sintering, micropores are lost as larger pores are formed. It is noted
that the average pore sizes in the doped aluminas increase after steaming
(Figure 3.14).

The y-Al203 phase is preserved for all materials. Electron diffraction patterns
reveal the diffuse rings, confirming crystallinity of the frameworks. However, this

Structural integrity of the y—Al203 phase is greatest in the pure alumina as judged

by XRD patterns (Figure 3.15). While all samples exhibit peaks consistent with y-

98

Figure 3.13. TEM images of MSU-y aluminas prior to (A,C,E) and after steaming
(B,D,F) at 600 °C for 5 hr in a 20% steam atmosphere. Images A and B are for
pristine MSU-y. Images C, D and E, F are for MSU-y doped with 5% La3+ and 5%
09‘“, respectively. The MSU-y aluminas were prepared as described in the
caption to Figure 3.11. The scale bar represents 100 nm in all figures.

99

 

 

 

~ 5%.!

.5
O
O
O

l

dVIdD

    

 

a .

E 800' 0 10 20 30 40 50 5%La
v .. Pore Diameter (nm)
'5 .

0

Q .

'5

in 600 _.

'U .

< .

0

E 1

3 400-

O .

>

Pristine

 

 

 

Figure 3.14. Nitrogen adsorption-desorption isotherms and BJH adsorption pore
size distributions (inset) of pristine and doped MSU-y aluminas after steaming for
5 hr at 600 9C in a 20% steam atmosphere. Samples were prepared as described
in the caption to Figure 3.11. Isotherms are offset 300 cc/g for clarity. The PSD
curves are in the same order as the corresponding isotherm.

lOl

 

 

 

 

Intensi

 

 

Pristine

 

 

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIITIIII1
10 20 30 40 50 60 70 80
Degrees(2®)

Figure 3.15. Wide angle powder XRD patterns of pristine and doped MSU-y
aluminas after steaming for 5 hr at 600 9C in a 20% steam atmosphere. Peaks
marked with an asterisk are attributed to CeAi03, cerium aluminate. Samples
were prepared as described in the caption to Figure 3.11.

102

A1203, the peak intensity is greatest in the pristine alumina. Peaks assignable to
the CeAl03 phase are again evident for the cerium doped sample.

Traditionally, lanthanides are stabilizers of aluminas toward conversion to d-
Al203. in the case of MSU-y aluminas, they may also provide stabilization toward
d-Al203, but the textural properties are compromised in comparison to the
pristine alumina. Perhaps compared against a pure alumina synthesized without
the aid of a surfactant, the lanthanides would improve hydrothermal stability.
Ironically, as the pore structure is relatively unchanged after steaming, the most
open framework appears to be the most stable. The results of the study are

summarized in Table 3.5.

3.4. Conclusions

Synthesis of the first mesostructured y—Ai203 has been accomplished through
a particle assembly pathway from the hydrolysis of aluminum alkoxide in the
presence of electrically neutral nonionic PEG and amine surfactants as
porogens. The formation of an MSU-S/B precursor is key in. attaining these
mesostructures. It has been demonstrated that these aluminas possess textural
properties superior to those of conventional aluminas, exhibiting surface areas,
pore sizes, and pore volumes in excess of 400 m2/g, 15 nm, and 1.00 cc/g,
respectively. The lack of a correlation between the size, mass, or molar mass of

surfactant employed in the synthesis with the textural properties of the surfactant

103

Table 3.5. Textural properties of MSU-y aluminas prior to and after treatment at
600 9C in a 20% steam atmosphere for 5 hr.al

Surface Area Pore Sizeb Pore Volume

 

Sample Condition 2
(m lg) (nm) (cc/0)
Pristine Fresh 371 13 1.1 1
Steamed 31 6 1 2 0.85
5% La Fresh 386 7 0.68
Steamed 220 12 0.77
5% Ce Fresh 416 9 0.99
Steamed 228 14 0.77

 

a The samples were prepared as described in the caption to Figure 3.11.

b BJH pore diameter as measured by the adsorption branch

104

free products is evidence that these materials are not assembled through a
supramolecular templating pathway. The open framework resulting from the
assembly of fundamental alumina fibers was verified through TEM imaging.
MSU-y aluminas can be doped with transition metals and lanthanides without
compromising the structure or surface area of the MSU-y alumina. Additionally,
these materials show exceptional thermal and hydrothermal stability, up to 600
9C for 5 hr in a 20% steam environment. These properties make MSU-y excellent

candidates for catalysts or catalyst supports.

3.5. Future Directions

While one example of a mesostructured higher temperature transition alumina
was shown here, it would be of interest to synthesize more of these phases. Pore
structures in these materials seem to remain intact at high temperatures, leading
one to inquire into the possibility of making a mesostructured d-Alzog, or at least
an d-Al203 that demonstrates marked improvements in textural properties.

As was the case for mesostructured boehmite, finding a suitable catalytic
application for MSU-y is also of interest. However, noting that y-A1203 is perhaps
the world’s most widely used catalyst support, the possibilities for making an
impact in this area are enormous. One such endeavor is the topic of the next

chapter.

105

3.6. References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

Wefers, K.; Misra, C. "Oxides and Hydroxides of Aluminum," Alcoa
Laboratories, 1987.

Misra, C. American Chemical Society Monograph, Vol. 184: Industrial
Alumina Chemicals, 1986.

Sohlberg, K.; Pennycook, S. J.; Pantelides, S. T. Journal of the American
Chemical Society 1999, 121, 7493-7499.

Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S.
Nature (London) 1992, 359, 710-712.

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.; et al. J. Am.
Chem. Soc. 1992, 114, 10834-10843.

Bagshaw, S. A.; Pinnavaia, T. J. Angew. Chem, Int. Ed. Engl. 1996, 35,
1102-1105.

Vaudry, F.; Khodabandeh, S.; Davis, M. E. Chem. Mater. 1996, 8, 1451-
1464.

Yada, M.; Machida, M.; Kijima, T. Chem. Commun. (Cambridge) 1996,
769-770.

Valange, S.; Guth, J. L.; Kolenda, F.; Lacombe, S.; Gabelica, Z.
Microporous Mesoporous Mater. 2000, 35-36, 597-607.

Cabrera, 8.; El Haskouri, J.; Alamo, J.; Beltran, A.; Beltran, D.; Mendioroz,
S.; Marcos, M. D.; Amoros, P. Adv. Mater. (Weinheim, Ger.) 1999, 11,
379-381.

Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D. Chem.
Mater. 1999, 11 , 2813-2826.

Gonzalez-Pena, V.; Diaz, l.; Marquez-Alvarez, C.; Sastre, E.; Perez-
Pariente, J. Microporous Mesoporous Mater. 2001, 44-45, 203-210.

Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309-
319.

Barrett, E. P.; Joyner, L. G.; Halender, P. P. Journal of the American
Chemical Society 1951, 73, 373.

106

(I5)

(15)

(17)

(18)

(19)

(20)
(21 )

(22)

(23)

(24)

(25)

John, C. 3.; Alma, N. C. M.; Hays, G. R. Appl. Catal. 1983, 6, 341-346.

Zhu, H. Y.; Riches, J. D.; Barry, J. C. Chemistry of Materials 2002, 14,
2086-2093.

Wolverton, C.; Hass, K. C. Physical Review B: Condensed Matter and
Materials Physics 2001 , 63, 024102/024101-024102/0241 16.

Krokidis, X.; Raybaud, P.; Gobichon, A.-E.; Rebours, B.; Euzen, P.;
Toulhoat, H. Journal of Physical Chemistry B 2001, 105, 5121-5130.

Schaper, H.; Doesburg, E. B. M.; Van Reijen, L. L. Appl. Catal. 1983, 7,
211-220.

Beguin, B.; Garbowski, E.; Primet, M. Appl. Catal. 1991, 75, 119-132.
Oudet, F.; Courtine, P.; Vejux, A. J. Catal. 1988, 114, 112-120.

Mizukami, F.; Maeda, K.; Watanabe, M.; Masuda, K.; Sano, T.; Kuno, K.
Stud. Surf. Sci. Catal. 1991, 71, 557-568.

Braun, S.; Appel, L. G.; Zinner, L. B.; Schmal, M. Br. Ceram. Trans. 1999,
98, 77-80.

Schaper, H.; Amesz, D. J.; Doesburg, E. B. M.; Van Reijen, L. L. Appl.
Catal. 1984, 9, 129-132.

Zhang, W.; Pinnavaia, T. J. Chem. Commun. (Cambridge) 1998, 1185-
1186.

107

Chapter 4

Hydrodesulfurization Over MSUdy Supported Catalysts

4.1. introduction

The hydrodesulfurization (HDS) reaction presents a challenge in the oil
refinement process. HDS is the removal of sulfur from sulfur-containing organic
molecules by treatment with H2 at elevated temperatures and pressures. HDS
has been performed in refineries since the 1950’s. Since that time, however,
improvement of the process has been constantly targeted. When burned, fuels
that are high in sulfur release SOx pollutants into the atmosphere, contributing to
smog and acid rain. in addition, the presence of sulfur in automobile exhaust
reduces the effectiveness of catalytic converters. Sulfur, if not removed, also has
a detrimental effect on downstream catalysts used in the refinement process.
Providing more motivation, the EPA has currently set up guidelines to restrict the
amount of sulfur allowed in diesel products to 15 ppm by the year 2010‘. This
corresponds to a 97% reduction over a ten year period. Complicating matters is
that relatively clean crude oil stocks have been consumed, leaving only heavier
crudes laden with an increased amount of sulfur. Clearly, advances in HDS
technology must be made to satisfy both economic and environmental concerns.

Transition metal sulfides (TMS) have proven to be effective catalysts for HDS.
Within this group of compounds, there are select metals whose sulfides are

particularly active. The position of the metal in the periodic table and the

108

morphology of the metal sulfide play a role in determining the relative catalytic
activity in the HDS reaction. A number of review articles discuss periodic trends
of TMS in HDS catalysisz's. Ruthenium disulfide is perhaps the most active TMS.
However, molybdenum disulfide is the catalyst of choice for commercial use. The
lower activity of M082 is compensated by its reduced cost. In addition, the activity
of TMS can be promoted by the presence of another metal sulfide, commonly
those of cobalt and nickel. The role of the promoter is not entirely clear, but it has
been demonstrated that they increase catalytic activity by increasing the electron
density on the primary metal or lowering the Mo-S bond strength9, allowing for
the creation of more active sites once sulfur atoms are removed by hydrogen.
Catalytically active sites on M082 are characterized by sulfur vacancies, or
coordinatively unsaturated metal ions, at the edge planes of the TMS. Utilizing a
support can help to disperse the catalyst and expose more of these edge planes.

Studies of catalytic activity have been performed for many different supports

1012 13,14

ranging from carbon , mesoporous synthetic clays and aluminosiiicates15
17, and zeolites”. Still, alumina is the most common support employed for HDS
catalysts. Alumina is stable under the conditions encountered during the reaction.
It is inert; hence it does not interfere with the HDS reaction. Aluminas are
inexpensive, are readily available, and exhibit desirable surface and textural
properties. Many aluminas are porous, allowing for better dispersal of catalyst
and increased access to the catalytically active sites. The transition alumina y-

Al203 is one such alumina with the features described above; it is the most

109

common support used in HDS chemistry. In fact, many commercial catalysts are
mixtures of molybdenum and cobalt sulfides on a y-Al203 support.

The use of mesoporous supports may lead to improved pathways to sulfur
removal. The increased surface area and pore volume in these materials should
allow for better dispersal of the catalytic phases and better access to the active
sites on these catalysts. Additionally, the pore sizes in these supports can
accommodate larger sulfur containing substrate molecules such as
dibenzothiophene-based molecules, which mimic more closely the sulfur bearing
contaminates in real oil stocks. To this end, mesostructured MSU-y aluminas
have been used in the present work as catalyst supports in HDS reactions. MSU-
y aluminas exhibiting different particle morphologies have been recently
synthesized‘g. Aluminas with scaffold or lath morphology could be attained from
different aluminum containing precursors. Here, the difference in morphology is

used as a basis for comparison of performance in HDS reactions.

4.2. Experimental

4.2.1. Sample Preparation

The MSU-y aluminas of scaffold morphology were synthesized as described in
the previous chapter. MSU-y with lath morphology was prepared similarly, but
from hydrolysis by ammonia of either aluminum cations or the Al13 oligomeric

cation. The MSU-y aluminas used in this study were prepared as follows.

110

4.2.1.1. MSU-y Exhibiting Scaffold Morphology

A 10% mass solution of Pluronic P84, a nonionic PEO surfactant, was
dissolved in 2-butanol and water (20/80 v/v) under stirring on a magnetic stirrer.
To this solution, aluminum sec-butoxide was added and stirred until
homogenous, to achieve an overall reaction stoichiometry Al:P84:s-BuOH:H20 of
66:1:103:2100. The reaction mixture was removed from stirring and allowed to
age at 100 °C for 20 hr in a sealed vessel to yield the MSU-SIB surfactant-
boehmite precursor. This intermediate was recovered by vacuum filtration,
washed with water and 2-butanol, and air dried. The surfactant free MSU-y
alumina was obtained through calcination of the MSU-SIB material at 500 °C for

4 hr.

4.2.1.2. MSU-y Exhibiting Lath Morphology

The assembly of a MSU-SIB mesostructure was achieved by aging an
aqueous mixture of a [Al13O4(OH)24(H20)12]Cl-, and P84 at 80 °C for 6 - 10 h and
then hydrolyzing the mixture with concentrated NH4OH. The precipitate formed
upon hydrolysis was aged at 80 °C for 6 h, then at 100 °C for 24h, and air dried.
The resulting mixture of MSU-SIB and NH4C| was calcined at 325 °C for 3h, then

at 550 °C for 4h, to form the surfactant free MSU-y alumina.

4.2.1 .3. Catalyst Loading
The MSU-y alumina supports were loaded with catalyst by incipient wetness

impregnation of molybdenum and cobalt salts followed by thermal degradation to

111

the oxides. The supports were loaded with the volume of solution necessary to fill
90% of the available pore volume. For example, to achieve a 6% loading of Mo
on an alumina having pore volume of 1.0 cc/g, 0.90 mL (per gram of Al203) of a
solution of ammonium heptamoiybdate tetrahydrate would be required. The
amount of salt dissolved in this solution is determined by the desired loading. in
this case, a 6% Mo loading would require that 0.06 gram of Mo per gram of
A1203, or 0.110 gram of the Mo salt, be impregnated. The requisite volume of this
solution was added dropwise to the MSU-y alumina, with intermittent agitation by
either manual or mechanical shaking, to provide even wetting. The M0
impregnated alumina was calcined at 400 °C for 2 hr to decompose the salt to
the oxide.

The pore volume of this intermediate material was obtained through nitrogen
adsorption. The cobalt was loading was determined to provide a 3:1 molar ratio
of Mo:Co. Cobalt (ll) nitrate hexahydrate was impregnated onto the support in the
same manner as described for the Mo loading. Another calcination at 400 °C for
2 hr yielded the catalyst in its non-active, oxidic form.

Conversion to the catalytically active sulfidic form was achieved by heating
the materials to 400 °C for 1 hr in a tube furnace while a stream of 10% H28 in
nitrogen was passed over the material. In each case, this resulted in the
formation of a black powder that was readied for use by pressing into a pellet.

In this study, the lath and scaffold aluminas were separately loaded with 6%,
9%, and 12% M0 by mass. These samples are denoted by their morphology and

loading. For example, the scaffold alumina with a 9% Mo loading is referred to as

112

Scaffold-9. The loading of the commercial Crosfield 465 catalyst was in the range

of 10-20% M0 by mass. It was sulfided as received.

4.2.2. Reactor Design and Experimental Conditions

The HDS pilot plant used in this study is housed in the Chemical Technology
Division of Argonne National Laboratory, Argonne, IL. The automated,
continuous-flow unit is made primarily of 316 stainless steel tubing. Liquid feed in
all experiments was 0.8 wt% S as dibenzothiophene (DBT) in hexadecane. This
feed is representative of the kind of sulfur present in a middle distillate oil. The
gas feed is a hydrogen and nitrogen mixture. Both the liquid and gas feeds are
separately pumped through the system, eventually mixing together prior to
entering the preheater furnace. At this point, the liquid is vaporized and all feed
exits the preheater in the gas state. The feed is next passed through the reactor
furnace and over the catalyst bed. The bed contains approximately 19 of catalyst
diluted with 29 SiC, resting between plugs of glass wool. The reaction products
are swept out of the reactor and separated. Liquid products are condensed and
collected while gas products, mainly H28, are passed through a NaOH scrubber
solution before being vented. A schematic of the system is provided in Figure
4.1.

The typical reaction conditions were as follows. All liquid lines were heated to
35 °C to prevent the hexadecane from plugging the line. The preheater furnace
was set to 350 °C and the reactor furnace to 400 °C. The nitrogen flow was set at

200 cc/min and hydrogen at 500 cc/min. The system was purged and the liquid

113

 

 

 

Vent

 

 

 

 

 

 

 

Catalyst
V ‘—
Preheater V Bed

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Liquid NaOH
Collection

 

Figure 4.1. Schematic of the HDS reactor employed in the study. The liquid feed
of dibenzothiophene in hexadecane was introduced to the gas feed, a mixture of
hydrogen and nitrogen. The two feeds are passed through the preheater (set to
350 °C), where all feed is vaporized. Then the feed reactants are passed over
the catalyst bed at 400 °C. Liquid products are condensed and collected while
gas products are scrubbed in 2M NaOH prior to venting.

114

feed was introduced. The system pressure was set to 400 psi and was controlled

by a back pressure regulator. Liquid products were collected at desired intervals.

4.2.3. Characterization

Catalysts were characterized by powder x-ray diffraction (XRD), nitrogen
adsorption, and transmission electron microscopy (TEM). XRD patterns were
obtained on a Rigaku Rotoflex diffractometer with Cu K, radiation of 1.54 A.
Nitrogen adsorption — desorption isotherms were collected on a Micromeritics
TriStar 3000 sorptometer at 77K. Samples were degassed at least 12 hours at
150 °C under a vacuum of 10'6 torr prior to analysis. Surface areas and pore
sizes were determined by the BET” and BJH21 methods, respectively. TEM
images were collected on a JEOL 1OOCX microscope with a CeBe filament at an
accelerating voltage of 120 kV. Samples were prepared by sonicating a
suspension of the powder in ethanol for 20 min. One to two drops of the
suspension were then placed on a carbon coated holey film that was supported
on a 3 mm, 300 mesh copper grid.

Liquid collections from the HDS reactor were diluted in hexane for GC-MS
analysis. The products were separated on a DBS—MS column and analyzed using
an HP 5890 GC-MS Series it Plus. The conversion was determined as the
percentage of DBT converted to products. The selectivity was defined as the

percentage of product that is biphenyl.

115

4.3. Results and Discussion

4.3.1 . Catalytic Reactions

in all cases, the product mixture of the HDS reactions contained biphenyl,
cyclohexylbenzene (CHB), 4H-DBT (the ring hydrogenated product of DBT), and
unreacted DBT. Reaction pathways are provided in Figure 4.2. From the
conversion and selectivity data presented in Table 4.1, it can be seen that
conversions roughly increase with metal loading. It is interesting that for both
aluminas, the 9% loaded samples show a drop in conversion from that of the 6%
loaded samples. As expected, conversions are highest for the 12% loaded
samples, with a maximum of 87% for the scaffold-12 catalyst. Comparing
selectivities, those of the scaffold catalysts are consistently higher than those of
the lath catalysts. As the 12% loaded samples showed the highest conversions
and were the most selective, these catalysts were studied in more detail. it
should be noted, however, that both MSU-y catalyst at all loadings show results
comparable or superior to those obtained from the commercial catalyst tested
under similar conditions. The Crosfield catalyst had a conversion of 77% and a
selectivity for biphenyl of 61%, for a percent yield of biphenyl of 47%. The MSU-y
lath-12 and scaffold-12 catalyst led to biphenyl yields as high as 49% and 84%,

respectively.

116

Table 4.1. Results of HDS reactions over MSU-y catalysts exhibiting distinctly
different morphologies.

 

 

 

 

 

Time DBT Biphenyl Biphenyl

Samplea (hr) Conversionb Selectivityc Yield
% % %
1 53 43 23
Lath -6 2 79 62 49
3 69 62 43

W W 1 WWWéz’W "WS‘t—BWW "“3—6WW
Lath -9 2 66 58 38
3 64 60 38

W W ' W W1“ W W72W 63 ”W45W’"
Lath-12 2 78 61 48
3 77 64 49
W 1 " W 59”” WW52WWW W " W 31 ‘
Scaffold-6 2 76 73 55
3 74 89 66

WW 1 ' A W 61 W _ _ W 94 W W 57W—
Scaffold-9 2 67 89 60
3 67 93 62

WW W W W1" 73 99 W W72W‘
Scaffold-1 2 2 87 93 81
3 86 98 84

 

‘ The number refers to the loading of Mo, expressed as a mass percentage of material
b percentage of initial DBT converted

° percentage of biphenyl in product

117

DBT 3P
\ W“ _
02:1.) —- /_\ \ /

I
(IQ-Gk)

Figure 4.2. Reaction pathways of DBT in HDS reactions. The upper pathway
involves direct desulfurization leading to the formation of biphenyl (BP). The
lower pathway involves a hydrogenation step prior to desulfurization, leading to
the formation of cyclohexylbenzene (CHB).

118

4.3.2. Structural and Textural Properties of MSU-y Catalysts

Structural and textural properties were measured at several stages throughout
the catalyst preparation and testing processes. Powder XRD patterns and
nitrogen adsorption - desorption isotherms were obtained for each catalyst after
each of the following: alumina synthesis, metal loading, sulfiding, and catalytic
testing. The catalysts at each of these stages are referred to as pristine, loaded,
sulfided, and spent, respectively. The properties of a commercial catalyst,
Crosfield 465, are included for comparison.

The wide angle powder XRD patterns for the lath-12, scaffold-12, and
Crosfield catalysts are provided in Figures 4.3 — 4.5, respectively. The structural
integrity of the MSU-y aluminas remained intact over the duration of the
processing and HDS reaction; the y—Al203 phase is evident in each pattern. The
Crosfield 465 catalyst shows different behavior. After sulfiding, the alumina peaks
become increasingly broad, possibility due to the formation of mixed phases, or
more likely, due to coating of the alumina surface by highly dispersed M082,
leading to a decrease in the y-Al203 peak intensities. Restacking of the M082
may occur in the spent catalyst, exposing the y—Al203 support. The pattern of the
spent catalyst shows peaks that are assignable to both the y-A1203 and poorly
crystalline M082 phases. The absence of reflections consistent with M082 in any
of the MSU-y catalysts’ diffraction patterns indicates that the active catalytic
phase is present in small domains and is well dispersed on the supports. Similar

behavior (not shown) is observed in the lower metal content samples as well.

119

 

Lath

Spent

 

 

Sulfided

    

intensi

 

Pristine

 

 

 

 

IIIIIIIIIIIIIIIIIIIITIIIIIIIIIIIIlI—l

10 20 30 40 50 60 70 80
Degrees (26)

Figure 4.3. Wide angle XRD patterns for lath-12 MSU-y alumina taken at various
stages of the catalytic testing process. The y-Al203 phase is evident throughout
the process and the lack of peaks assignable to M082 indicate that the active
catalyst is present in small domains and is well dispersed on the support.

120

 

Scaffold

    

Sulfided

 

 

intensi

 

 

 

Pristine

 

 

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIITI

10 20 30 40 50 60 70 30
Degrees (26)

Figure 4.4. Wide angle XRD patterns for scaffold-12 MSU-y aluminas taken at
various stages of the catalytic testing process. The y-Al203 phase is evident
throughout the process and the lack of peaks assignable to M082 indicate that
the active catalyst is present in small domains and is well dispersed on the
suppon.

121

 

Crosfield

Spent

 

 

    

Sulfided

 

intensi

 

Loaded

 

 

 

 

 

 

IIIIllIIIIIIII'IIIIIIWIIIIIIIIITITI

10 20 30 40 50 60 70 30
Degrees (26)

Figure 4.5. Wide angle XRD patterns for the commercial Crosfield catalyst, taken
at various stages of the catalytic testing process. The decrease in intensity of the
peaks assignable to y-Al203 in the pattern of the sulfided sample may be due to
the coating of the alumina surface by small domains of M082. Peaks from y-A1203
and M082 in the pattern of the spent sample indicate that M032 particles may
aggregate sufficiently to cause scattering and expose the surface of the support.

122

Upon inspection of the nitrogen adsorption — desorption isotherms of the 12%
Mo loaded samples, Figures 4.6 and 4.7, it is clear that the metal loading,
sulfiding, and catalytic testing have a deleterious effect upon the pore structure
and textual properties of these catalysts. As the active catalytic phase is
introduced onto the support, the pore size, pore volume, and surface area will
naturally decrease. As shown in Table 4.2, in the lath-12 series, the pore volume
is reduced to 25% of its original value after use in the HDS reaction. The
scaffold-12 catalysts are the most resistant to the conditions imposed by this
process as they retain approximately 45% of their original pore volumes. The
effect is more pronounced at higher metal loading in each sample. A similar trend
is observed in the surface areas of these materials. Again, the scaffold-12
alumina exhibits the highest degree of structural integrity, retaining approximately
65% of its original surface area while the lath-12 retains only 15%.

in the case of the lath aluminas, at any point after loading the metal, the pores
appear to collapse, leaving only a small amount of porosity. The pore size
distributions are so broad that it is not possible to determine a mean pore size for
the catalyst at any point after loading (inset, Figure 4.6). in the scaffold-12
catalysts, the pore size distribution shifts to a smaller average pore size
throughout the process (inset, Figure 4.7). Comparing the MSU-y catalysts to the
commercial catalyst, the scaffold-12 catalyst exhibits adsorption behavior most
similar to the Crosfield catalyst (Figure 4.8). The textural properties of the lath,

scaffold, and commercial catalyst are summarized in Table 4.2.

123

 

  
   
 
    
    

 

 

 

 

 

 

 

 

 

800: Lath
700: a
. E
..>, L
A 600j X;
9 : 1m
8 . ii 10 20 30 40 50
v _ Pore Diameter (nm)
500
'0 ..
0 .
e .
" a e I
8 400: Pristine
U -l
< .
0 -I
E 3001
3 . Loaded
'5 t
> ..
200:
1 Sulfided
100:
Spent /
5.......,............
0 0.2 0.4 0.6 0.8 1

PIPO

Figure 4.6. Nitrogen adsorption - desorption isotherms and BJH adsorption pore
size distributions (inset) for lath-12 MSU-y catalysts taken at various stages of the
catalytic testing process. The pore volume steadily decreases as the catalyst is
loaded, sulfided, and spent. lsothems are offset 100 cc/g for clarity and PSD
curves are displayed in the same order as the corresponding isotherms.

124

 

Scaffold

W

 

    
   
 
    
    

. 0 10 20 30 40 50
Pore Diameter (nm)

Pristine

Loaded

Volume Adsorbed (cc/g)

 
 
  

200; Sulfided
Spent

 

 

 

PIPO

Figure 4.7. Nitrogen adsorption — desorption isotherms and BJH adsorption pore
size distributions (inset) for scaffold-12 MSU-y catalysts taken at various stages
of the catalytic testing process. The pore volume decreases slightly as the
catalyst is loaded, sulfided, and spent while the pore size is increased and the
distribution broadened over the same conditions. Isothems are offset 0, 100, 100,
and 200 cc/g from bottom to top for clarity and PSD curves are displayed in the
same order as the corresponding isotherms.

125

 

- Crosfield

Spent

dVIdD

      
    
    

1

Sulfided

 

b

O

O
l

0 10 20 30 40 50
Pore Diameter (nm)

1

: Loaded

Sulfided

Volume Adsorbed (cc/g)
0.1
O
O

 

 

 

PIP

Figure 4.8. Nitrogen adsorption - desorption isotherms and BJH adsorption pore
size distributions (inset) for the commercial Crosfield catalyst taken at various
stages of the catalytic testing process. The pore volume decreases slightly as the
catalyst is sulfided and spent. However, the pore size and distribution remain
relatively unaffected over the same conditions. lsothems are offset 100 cc/g for
clarity and PSD curves are displayed in the same order as the corresponding
isotherms.

126

Table 4.2. Textural progerties of MSUj alumina HDS catalysts.

 

 

 

 

 

 

 

 

 

 

 

BET Surface BJH Pore
_ Pore Volume
Sample Condition Area Size — Ads.
(cc/9) 2
(m lg) (nm)
Pristine 0.60 309 1 0.0
Loaded 0.56 234 --
Lam's Sulfided 0.35 158 --
2- _ 2 - Spent __. -928” 140 _ __
Pristine 0.60 309 1 0.0
Loaded 0.39 21 3 --
Lath‘g Sulfided 0.35 159 --
a - z -,_S_P_9m ___ __Q.2_-L_ 2 75 _ - _.:: __
Pristine 0 60 309 10.0
Loaded 0.33 1 82 --
Lath” Sulfided 0.31 127 --
g 2_ 2 Spent _ _Q.15 _ 47 --
Pristine 1 .23 368 13.0
Loaded 0.97 354 1 1.0
scafi°'d'6 Sulfided 0.77 270 1 1.0
2 v j W - (VS ent 22-22-. ____0.51_ _ _ 223 8.0
Pristine 1 .23 368 13.0
Loaded 0.58 335 4.0
scam'd'g Sulfided 0.43 271 4.0
_ SpenL ___ _ _ _0.44 - 249 4.0
Pristine 1 .23 368 13.0
Loaded 0.98 350 12.0
scafi°'d'12 Sulfided 0.58 261 9.0
‘ Spent _ 0.56 238 9.0
Loaded 0.55 223 7.5
Crosfield Sulfided 0.24 100 1 1 .0
Spent 0.31 147 8.0

 

127

it is understandable that the scaffold and commercial catalyst have similar
adsorption behavior as they exhibit similarities in morphology as well. Figures 4.9
through 4.11 provide a series of TEM images comparing the particle morphology
of the loaded catalysts with that of the spent catalysts. The morphology of the
scaffold-12 and commercial catalysts is largely unaffected after use in the HDS
reaction. However, the long range order in the lath-12 catalyst has been
destroyed to some extent. The insets in each case show an approximately 30 nm
by 30 nm area. It can be seen that the local pore structure in each catalyst
remains nearly unchanged. The decrease in pore volume, as determined from

the isotherms, is reflected in these images.

4.3.3. Relationships Between Catalyst Structure and Performance

While both MSUJY catalysts were effective in HDS reactions, the scaffold
aluminas do show better performance than the lath aluminas. This can be
explained by the differences in morphology of the alumina support. The pores in
the lath aluminas are slit-like and run in the direction of the fundamental particles
comprising the framework. Contrast that to the scaffold aluminas, where the
pores are void spaces between randomly oriented fundamental particles. This
leads to two different M032 particle sizes when supported on the aluminas.

The layered structure of M082 is highly anisotropic. The sheets of M082 can
extend a thousand-fold the distance in the stack direction, leading to the
formation of a large, but catalytically inactive, basal plane. it has been realized

that only the edge planes of the material contain the catalytically active sites.

128

Figure 4.9. Bright field TEM images of the scaffold-12 MSU-y catalyst (A) after
loading and (B) after using in an HDS reaction for 3 hr at 400 °C and 400 psi.
The scale bar represents 100 nm. The inset is an approximately 30 nm x 30 nm
area of the catalyst. That the framework of the alumina is largely unaffected by
the conditions of the HDS reaction provides evidence of its stability.

129

 

 

130

Figure 4.10. Bright field TEM images of the lath-12 MSU-y catalyst (A) after
loading and (B) after using in an HDS reaction for 3 hr at 400 °C and 400 psi.
The scale bar represents 100 nm. The inset is an approximately 30 nm x 30 nm
area of the catalyst. The local pore structure (inset) is compromised to some
degree, as is shown by the decreased void space in (B) as compared with (A).
The long range order of the alumina laths is lost after use in the HDS reaction.

131

 

 

Figure 4.11. Bright field TEM images of the commercial Crosfield catalyst (A)
after loading and (B) after using in an HDS reaction for 3 hr at 400 °C and 400
psi. The scale bar represents 100 nm. The inset is an approximately 30 nm x 30
nm area of the catalyst. Although the local pore structure (inset) appears to
withstand the conditions encountered in the HDS reaction, it appears that there is
some aggregation of the fundamental particles after use.

133

 

 

While the basal plane contributes greatly to the surface area of the catalyst, the
surface area does not correlate with activity in HDS”. The edge planes contain
the sulfur vacancies that are the active sites in HDS.

Comparing the structures, there is a good match in morphology between the
pore structure in the MSU-y lath catalyst and M082. Still, the pores will limit the
extent of the growth of M082 in the direction perpendicular to the planes, leading
to the formation of truncated slabs of the active sulfide on the MSU-y laths. This
leads to highly dispersed, small domains of MoSz.

The pore structure of the MSUJY scaffolds is not as ordered as in the MSU-y
laths. Additionally, the fibers of alumina constituting the framework do not provide
a good area for M082 to grow into large sheets. Consequently, the growth of
M082 should be greater in the layer stacking direction on MSU-y scaffolds than
on MSU-y laths. For the same molybdenum loading, a more highly stacked M082
particle will provide a greater number of edge sites than will a lesser stacked
particle. This is one contributing factor to the greater conversion on MSU-y
scaffolds than on MSU-y laths. Secondly, due to the nature of the open
framework in the scaffold catalyst, access to these catalytically active edge sites
may be increased as well.

While the conversions on MSU-y catalysts are comparable, the selectivities
are much better for the scaffolds than for the laths. The differences in selectivity
are addressed by the rim — edge model of the M032 catalys123'24. If the M082
particle is viewed as a stack of discs, then the edge sites are divided into two

categories, distinguished by their location within the stack (Figure 4.12). Sites on

135

:I edge

It layers

basal

 

Figure 4.12. The Daage-Chianelli rim-edge model of the M082 catalyst particle.
Sulfur removal takes place on rim and edge sites while hydrogenation of the
aromatic ring is restricted to rim sites only“.

136

the top and bottom discs are designated rim sites and all others as edge sites.
The taller the stack, the higher the edge / rim ratio. In the development of this
model, Daage and Chianelli found that hydrogenation rates were inversely
proportional to the stack height. It was therefore concluded that hydrogenation
takes place only on the rim sites where sulfur removal (HDS) occurs on both
sites.

On MSU-y catalysts, a greater stack height on the scaffold catalysts is in
accord with the increased selectivity. Having shorter but wider particles of M032
on the lath catalysts also agrees with the model. While it is argued above that
merely increasing the stack height will lead to a higher conversion due to an
increased number of edge sites, it is clear that particle size effects have a greater
influence on selectivity than conversion. As an example, consider the blocks in
Figure 4.13 to represent M082 particles. Keeping the overall amount of M032 the
same, but chopping the slab in half and stacking the two pieces on top of each
other, a larger number of edge sites are created. The more pronounced effect is
that the ratio of rim sites to rim plus edge sites is decreased. This is in line with
the data; selectivities are much better on scaffolds than on laths, while

conversions are comparable at a given loading.

137

 

\3
fix.

100

 

Total Edge Sites: 660
Rim / (Rim + Edge) = 0.66

   
   

 

 

10 j
6

 

50

 

Total Edge Sites: 720

Rim / (Rim + Edge) = 0.33

 

Figure 4.13. influence of morphology on the ratio of rim sites to total sites
available on the edges of MoS2 particles. While increasing the stack height leads
to a slight increase in the overall number of edge sites, the more pronounced

effect is the decrease in the relative number of rim sites.

138

4.4. Conclusions

MSU-y aluminas were used effectively as catalyst supports for HDS catalysis.
These crystalline mesostructured aluminas are capable of dispersing up to a
12% loading of molybdenum and able to withstand the harsh conditions
encountered during the reaction. MSU-y aluminas with lath morphology had
conversions are selectivities comparable to a commercial catalyst while the
aluminas with scaffold morphology had conversions and selectivities superior to
the commercial catalyst. The MSU-y alumina supports may lead to more
economical catalysts, as greater conversions and selectivities are obtained with
lesser quantity of active catalyst.

The difference in performance of the two MSU-y aluminas is attributed to
morphological matching with the catalytically active MoS2 phase. The laths
promote growth of MoS2 in a sheet-like morphology while scaffolds lead to
increased stacking in a direction perpendicular to the layers of MoS2.
Additionally, the open framework and larger pore volume of the scaffold alumina
as compared to the lath alumina may provide easier access to catalytically active
edge sites of M082. Viewed in the rim-edge model of MoS2 catalysts, these

results are reasonable.

4.5. Future Directions
This initial venture into utilization of MSUJy aluminas in HDS reactions was
fruitful, but there are some aspects that can be explored further. Can MSU-y

aluminas be loaded with an increased mass of molybdenum? Some industrial

139

catalysts contain upwards of 20% molybdenum. The method of incorporation
employed in this study was incipient wetness impregnation. Supports are loaded
based upon their pore volume. The limiting factor is the solubility of the metal salt
in the volume available with which to work. The higher pore volume of the
scaffold materials may be better able to accommodate a higher loading.
Othentvise, an alternative method will be needed.

HDS reactions were limited to 3 hour runs. What is the longevity of the MSU-y
catalyst? Can it remain on stream for extended periods of time? The average life
of industrial catalysts may be approaching twp years. Are MSU-y catalysts
poisoned after a certain time? If so, can they be recycled or regenerated?

How do MSU-y catalysts perform with other substrates? As oil crudes become
increasingly dirtier, larger molecules with sterically hindered sulfur atoms, such
as 4,6-dimethyldibenzothiophene become more prevalent. The larger molecules
present the toughest challenge in HDS. How do MSU-y aluminas perform with
real oil feeds?

There are reports in the literature employing lanthanide doped aluminas as
supports for catalysts”. Noting that similar doping of MSU-y aluminas actually
decreased the hydrothermal stability of the materials, it is possible that such an
experiment would be of little value here.

Finally, there has been a report of a new HDS catalyst claiming to have
superior activity as compared to any catalyst currently in existence. This new
catalyst, marketed under the name Nebula, was developed jointly by Akzo-Nobel,

ExxonMobil, and Nippon Ketjen“. Nebula catalysts are not based upon an

140

alumina support as many of it predecessors are. These catalysts are already
being used in European refineries to produce low sulfur diesel fuels.
Consequently, there may be a shift in direction in research and development of
HDS catalysts. What fate that holds for alumina based catalysts is yet to be

determined.

141

4.6. References

(1 I
(2)
(3)
(4)
(5)

(6)
(7)
(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)
(16)

(17)

(18)

in New York Times: New York, NY, 2000.

Chianelli, R. R. NATO ASI Ser., Ser. C 1983, 105, 361-378.

Topsoe, H.; Clausen, B. S. Catal. Rev. - Sci. Eng. 1984, 26, 395-420.
Chianelli, R. R. ACS Symp. Ser. 1985, 279, 221-234.

Prins, R.; De Beer, V. H. J.; Somorjai, G. A. Catal. Rev. - Sci. Eng. 1989,
31, 1-41.

Harris, 8.; Chianelli, R. R. Theor. Aspects Heterog. Catal. 1990, 206-255.
Chianelli, R. R.; Daage, M.; Ledoux, M. J. Adv. Catal. 1994, 40, 177-232.

Hensen, E. J. M.; De Beer, V. H. J.; Van Santen, R. A. NATO ASI Ser.,
Ser. 31998, 60, 169-188.

Nielsen, L. P.; Christensen, 8. V.; Topsoe, H.; Clausen, B. S. Catalysis
Letters 2000, 67, 81 -85.

Hensen, E. J. M.; Brans, H. J. A.; Lardinois, G. M. H. J.; De Beer, V. H. J.;
Van Veen, J. A. R.; Van Santen, R. A. J. Catal. 2000, 192, 98-107.

Farag, H.; Whitehurst, D. D.; Sakanishi, K.; Mochida, i. Catal. Today1999,
50, 9-17.

Ledoux, M. J.; Michaux, 0.; Agostini, G.; Panissod, P. J. Catal. 1986, 102,
275-288.

Carrado, K. A.; Xu, L.; Marshall, C. L.; Wei, 0.; Seifert, S.; Bloomquist, C.
A. A. Stud. Surf. Sci. Catal. 2000, 129, 417-424.

Carrado, K. A.; Marshall, C. L.; Brenner, J. R.; Song, K. Microporous
Mesoporous Mater. 1998, 20, 17-26.

Yue, Y.; Sun, Y.; Gao, Z. Catalysis Letters 1997, 47, 167-171.
Song, 0.; Reddy, K. M. Appl. Catal., A1999, 176, 1-10.

Ramirez, J.; Contreras, R.; Castillo, P.; Klimova, T.; Zarate, R.; Luna, R.
Appl. Catal., A 2000, 197, 69-78.

Okamoto, Y. Catalysis Today 1997, 39, 45-59.

142

(19)

(20)

(21)

(22)

(23)

(24)

Zhang, Z.; Hicks, R. W.; Pauly, T. R.; Pinnavaia, T. J. Journal of the
American Chemical Society 2002, 124, 1592-1593.

Tauster, S. J.; Pecoraro, T. A.; Chianelli, R. R. J. Catal. 1980, 63, 515-
519.

Daage, M.; Chianelli, R. R.; Ruppert, A. F. Stud. Surf. Sci. Catal. 1993, 75,
571-584.

Daage, M.; Chianelli, R. R. J. Catal. 1994, 149, 414-427.

Miciukiewicz, J.; Laniecki, M.; Domka, F. Catalysis Letters 1998, 51, 65-
68.

lnternet publication currently available on the ExxonMobil website:
http://www.exxonmobiI.com/refiningtechnologies/pdf/h7_nebula.pdf

143

Chapter 5
Concluding Remarks

A Commentary

As mentioned in the introduction, there has been little progress made in the
development of mesostructured aluminas relative to other metal oxides. At the
start of this work, all known mesostructured aluminas were amorphous in nature.
This work has demonstrated a considerable achievement in reporting at least two
new crystalline forms of mesostructured alumina. Alumina has uses in numerous
applications as a catalyst support or as a component in catalyst system. HDS
was just one arena in which the utility of these new materials could have been
demonstrated. The improvement in textural properties, along with the stability of
these materials, provides great potential to MSUJY aluminas for use in a variety of
catalytic applications.

To fully reach this potential, eventually the syntheses reported here will need
to be commercialized. That leads to one important drawback regarding these
materials. They are made from relatively expensive precursors of polyalkylene
oxides and aluminum alkoxide. Will other applications prove it necessary to scale
up this production of this material? Are the increased costs balanced by
increased performance? These are questions that will be addressed at some
time in the future.

On a personal note, this project has been very challenging. There were times
that l wondered if anything would come out of it. Of course, in the end it has been

144

very rewarding. I am satisfied with the contributions that i have made to build
upon existing knowledge. It has only been ten years since Mobil’s initial
disclosure of the M418 family of materials. So while the entire field of
mesostructured materials is relatively new, the subset of mesostructured
aluminas is in infancy.

I’m excited that I have had the opportunity to create some of the pioneering
work in this field. MSU and the Pinnavaia group have been at the forefront of this
from the beginning. The first mesostructured alumina and now the first crystalline
mesostructured aluminas have developed in these labs. Finally, I am proud to be
in the lineage of scientists that have contributed to this project. It started with
Steve Bagshaw, continued with Peter Tanev, Wenzhong Zhang, myself, and is
now left in the capable hands of Zhaorong Zhang (and future additions to the
Pinnavaia group). There is still progress to be made in this field and i look

fonrvard to the future contributions from this group.

145

    
  

IIItI'IIjflIjIIIII'IIjIIIIjIIIjIII

3