'..1 M151 "'S:" ‘11'
"11:..1 1 '111"'[ '1“;

'111111 l'>' [3.331131
I.

   
 

~' 11111211512

32-11; 11:11...“ 11,,
'1‘ ‘13:,“ 1‘.

1". , 11:
':.\1".l::f'/ ""9 H '1. ‘111ll11;""’1|t .Etfifll'
'.-1.1"," "‘1'” " 111211 f. '11

1.“.',!1'~1"" ' “1'12: 1"1"1"'11
'1”'.111'€7' 3'7“” 11 ”1'21“"
3":2'" H." ‘;;'.::‘ ""',",131
11 117;?

11311.

1" 1" "

'11;1'§||.i113'
“1'1'."-1g'1'1;'11‘{"'1

‘v

. 1 ’1314; ‘1' 1111-111
"' '1'" 1,1." 1'1 211% '1'."'-" "W11 11"" 1- 1""'1
1.” 1‘ ‘

, M1

. 1"";
1"':'."11 1“)“ 1"}; 'i'i‘lil'

a—d

WV."-
11,1'1 “1:1" .1] ‘1". . ‘ ..
1:. ' 1 1.1321.

.212 2'111'1
$715.17, 1;

‘.1"‘11'1"111"1111‘1'1 ' ' .'.'.-s1"" '3113” 11"“-

"' .111 ‘ 1131111111351! """1 '1‘ 1111*!

1; .1
1'. 1‘21 1'
"2"‘1 1'; 1"

--—v~ pry-1’ w wv—q

-r.-‘- ....

 

 

“wait: 7?

w*-~-k2. 2 1

§ zii

1 151

. 1.!

'l -;2 1 ',
1 1:13

' ; 1'1:

‘ 11W?

.

1““ a
"a? ..

.93..

r355

.. ”4.4...- 1
. ‘.‘:t““‘- --—:4

 

”1‘11”“ 11'1"! 'L"' 11' fi
111'111"1'1"1'111'1""'"'~'"'11.. .
.1 1'1 '
' 1.2.11'2 V 1‘ {11:1, q'im'id'
L" .'.','1 "S l" 1114' 1" “J!” ","21: ;.1.1.11 '1'; 11$" :21?“ 2.;1111I1‘1

' """" :S 1;;

1'."

'1?"'

”.1

d;|;."
11'
"" :1 .1

_»

w
.u

-.-n ~u~

.IE
, .

S.
13"
1:

1
-;F:
1

.

J

<..

‘1'»
22:1

"J?

'- ....

..
.wm' 13‘s“.
"" < - «:5.»
“$22.21...-m.
Juana."

”-.‘

~..
—~

~. .-
.- .v‘vv .-
—.~ ~1~>

'5
3.

»~
7' .

1 151”;
' 1111
"1 '6' 1"'"3'

53:,
1v

‘8?"

.. n1

1) 1S“ 1'11-

Tax
“7'5!

LulJIA »
lunch:

v
.1

211 .
" '4'“ '12; mi:

@5133 5'
'7‘ 1

.2
"1L?" .
f‘ a» -----2")
m, :“ 2 2.”;
‘ 3:32:55“.
.

P"1) ~11;
1.111“ :2
5"‘1252 'r _”2'

t
11

.27}: :1- 4

“1-4

fi.»..-.

‘1'}...
fiffifi‘

... -.,,
14-.
.-..__. ~.~

,-— u w..- _

42.5.

»
1-1"

“ 12' - J...’
‘ tut " .
' 512'“-.. ,. -:
n w"... “a £9.
‘33" -4:
f. r 2
. -
A ‘2
.J‘ x,
2:

:L': "
m»- v

at"

1.
H11.

"',:"'i '11}

2 .1.
L. ' ”3'- 1'3“
:22: 21:39
‘Vfi: 2‘”

*7? I 3|..-

"v\

, _--'x :3:

_‘v-~
“r

 

.~

~
,2

1:1,
.21.. .
”F.

 

llllllllallllllllllllllllllIlllllllllllllIllllllllllllllllllllll

31293 01698 7335

This is to certify that the

dissertation entitled

PREPARATION AND CHARACTERIZATION OF SOL-GEL

DERIVED METAL OXIDE THIN FILMS AND POWDER?-

FOR COATINGS AND CATALYSIS
presented by

Per A. Askeland

has been accepted towards fulfillment
of the requirements for

 

Ph rD_ degree in Chgmistry

' ' /
a“- 0. «£1.;L,TK._’MJ
\

Major professor

 

Date fiwé 4777

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Michigan State
Unlverslty

 

 

 

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

DATE DUE I MTE DUE DATE DUE
O

“ 1 A
l ,:

 

 

A
a. w~ T

 

 

 

 

 

 

 

 

 

 

 

 

 

1/” WWW.“

PREPARATION AND CHARACTERIZATION OF SOL-GEL DERIVED METAL
OXIDE THIN FILMS AND POWDERS FOR COATINGS AND CATALYSIS.
By

Per A. Askeland

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1997

a}

ABSTRACT

PREPARATION AND CHARACTERIZATION OF SOL-GEL DERIVED METAL
OXIDE THIN FILMS AND POWDERS FOR COATINGS AND CATALYSIS.

By
Per A. Askeland

Sol-gel chemistry is a wet chemistry technique involving the hydrolysis and
condensation of metal alkoxides to form gels, films, or powders. By controlling the
relative rates of hydrolysis and condensation the structure and morphology of sol-gel
derived materials can be controlled to fit the need of a desired application. This
dissertation examines the design of metal oxide coatings and catalysts prepared by sol-
gel chemistry.

Organically modified aluminum oxide films were prepared by reacting
aluminum alkoxide solutions with valeric acid or 2,4-pentanedione (acac). The effect of
the ligand-to-metal ratio on film quality was examined by visual inspection. Fourier
transform infrared spectroscopy (F TIR) was used to examine the variation in the bulk
structure of the films as a function of ligand-to-metal ratio. The ligand-to-metal ratios at
the surface of the fihns were measured using X-ray photoelectron spectroscopy (XPS).

Ceria-alumina catalysts were prepared by the hydrolysis and condensation of
cerium and aluminum valerate precursors. FTIR measurements of the uncalcined

catalysts suggest that some mixing occurs between the cerium and aluminum precursors.

Powder X-ray diffraction (XRD) measurements reveal that the catalysts are amorphous

after calcination at 500 °C. Diffraction patterns for CeOz and y—Ale3 are observed

following calcination at 900 °C. XPS measurements revealed that cerium was dispersed
following calcination at 500 °C, but did not remain so after calcination at 900 °C. The
surface areas of the catalysts were reasonably high following calcination at 500 °C, but
decreased significantly after calcination at 900 °C.

Ceritun and titanium mixed oxide thin films were prepared from cerium and
titanium valerate precursors by the sol-gel method. Changes in the film structure due to
variation of the Ce/T i atomic ratio and calcination temperature was measured by XRD.
Surface segregation of cerium was measured by XPS. All films showed a large surface
enhancement of cerium following calcination at 800 °C.

Cerium and titanium oxide thin films were prepared by calcining valerate films
at 400 and 600 °C. Structures of the films are determined with XRD and XPS. The
reactivities of these films are evaluated using XPS analyses before and after in situ H2
reduction. For all films the extent of Ce reduction was found to increase with reduction
temperature. The extent of reduction decreased with increased cerium loading and
crystallinity. The crystallinity of TiOz did not have an effect on the extent of Ce

reduction.

Acknowledgments

It goes without saying that this work would have never been possible without
the help and support of numerous people. I would like to thank my advisors Dr. Jeff
Ledford and Dr. Kris Berglund for their guidance, encouragement and support. I
would also like to thank Dr. Simon Garrett and his research group, Todd and Lili, for
their patience in allowing me to finish my experiments.

Most thanks go to the other surviving orphans; Kathy Severin, Greg Noonan,
Paul Park, and honorary orphan Dana Spence. Without their advice, criticism,
encouragement, and friendship I don’t believe I would have ever made it. I would also
like to thank former Ledford group members Ed Townsend, Mike Thelen, Tom Curtis,
Mark Waner and Jeff Rasimas who made the journey much more enjoyable.

I’m not quite sure how Sara and Matt managed to live me and still keep their
sanity, but fortunately for me they were there and with further treatment they should be
able to reenter society. Likewise I have to thank Sue, Chris, Michelle, Tim, JP, Al,
Al, and Jen for their fi'iendship. I need to thank Dag’s (Brenda, Ebru, Randy and T.)
for being my home away from home and the good people and CMSC for being my
new home away from home. So as the last orphan, it’s time for me to turn out the

lights and say, of course, get to work.

TABLE OF CONTENTS

LIST OF FIGURES
LIST OF TABLES
CHAPTER 1
INTRODUCTION .......................................................................................
1.1 Introductory Remarks .............................................................
Sol-Gel Chemistry .................................................................
1.2.1 Overview of Sol-Gel Chemistry
1.2.2 Sol-Gel Preparation of Catalysts ...................................
1.3 Catalytic Systems Studied in this Text ................................... .
1.3.1 The Role of Cerium Oxide in Catalysis ........................
1.3.2 The Role of the Catalytic Support ...............................
1.4 Material Characterization Techniques ......................................
1.4.1 X-ray Photoelectron Spectroscopy (XPS) ....................
1.4.2 XPS of Cerium Oxides .................................................
1.4.3 Powder X-ray Diffraction ............................................
1.4.4 Surface Area Measurements (BET) ..............................
1 .5 References .................................................................................
CHAPTER 2
PREPARATION AND CHARACTERIZATION OF ORGANICALLY
MODIFIED ALUMINUM OXIDE THIN FILMS ..........................................
2.1 Abstract ................................................................
2.2 Introduction ...........................................................
2.3 Expenmental .......................
2.4 Results and Discussion ...........................................................
2.4.1 Unmodified Alumina .....................................................
2.4.2 Acetylacetone modified Alumina ..................................
2.4.3 Valerie Acid Modified Alumina ....................................
2.5 References .................................................................................
CHAPTER 3
SOL-GEL SYNTHESIS OF CERIA-ALUMINA CATALYSTS ....................
3.1 Abstract .....................................................................................

3.2 Introduction.................................................. . . . . . . . . . . . . . . . . . .

cobw—

11

14
18
18
24
26
27
32

33

33
34
35
36
38
38
43
47

49

49
50

3.3 Experimental ............................................................................. 52

3.4 Results and Discussion ............................................................. 54
3.4.1 Uncalcined Materials ................................................... 54
3.4.2 500Cey Catalysts ......................................................... 58
3.4.3 900Cey Catalysts .......................................................... 63
3.5 References ................................................................................. 67
CHAPTER 4
SYNTHESIS AND CHARACTERIZATION
OF CeOzl'I‘iOz THIN FILMS .......................................................................... 70
4.1 Abstract ..................................................................................... 70
4.2 Introduction ............................................................................... 71
4.3 Experimental ............................................................................. 72
4.4 Results and Discussion ........................................................... . 73
4.5 References ................................................................................. 89
CHAPTER 5
H2 REDUCTION OF SOL-GEL DERIVED CERIUM
AND TITANIUM OXIDE THIN FILMS ....................................................... 91
5.1 Abstract .................................................................................... 91
5.2 Introduction ............................................................................... 92
5.3 Experimental ............................................................................. 93
5.4 Results and Discussion ............................................................. 96
5.4.1 Standard Materials ........................................................ 96
5.4.2 10% Ce/T i Films ........................................................... 98
5.4.3 50% Ce/T i Films ........................................................... 103
5.4.4 H2 Reduction Studies .................................................... 103
5.5 References .................................................................................. 1 1 1
CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS ................................. 113
6. 1 Conclusions .................................................................... 1 13
6.2 FutureDirections............................................_. ................. 115
6.3 References ...................................................................... 1 17

vi

LIST OF FIGURES

Figure 1.1 Scheme for the reduction of Ce02 by H2 ......................................
Figure 1.2 Schematic diagram of the formation of various alumina phases.
Figure 1.3 Phase diagram for TiO2 ................................................................
Figure 1.4 XPS spectrum of a 50% Ce02/TiO2 thin film ..............................
Figure 1.5 C Is spectrum of an acac modified alumina film .........................

Figure 1.6 Cerium 3d XPS spectra of Ce02 (A) and Ce(III)
acetylacetonate (B) ......................................................

Figure 2.1 FTIR spectrum of an unmodified alumina film ............................

Figure 2.2 FTIR spectra of acac modified alumina films with ligand to
metal ratios of 0.5 (a), 1.0 (b), 1.5 (c), 2.0 (d), and 3.0 (e) ...........

Figure 2.3 Extent of acac ligation versus ligand:aluminum alkoxide
solution ratio ..............................................................

Figure 2.4 FTIR spectra of valeric acid modified alumina films with acid
to metal ratios of 0.5(a0, 1.0 (b), 2.0 (c), 3.0 (d), and 6.0 (e) .......

Figure 2.5 Extent of valerate ligation versus valeric acidzaluminum
alkoxide solution ratio ...................................................................

Figure 3.1 F TIR spectra (4000-400 cm") 0 f as prepared films with CezAl
ratios ofa) 0:1, b) 1:10, c) 1:1, and (1) 1:0 ....................................

Figure 3.2 FTIR spectra (3 800-3100 em") 0 fas prepared turns with CezAl
ratios ofa) 0:1, b) 1:10, c)1:1, and d)1:0 ....................................

Figure 3.3 XRD patterns of catalysts calcined to 500 °C for CezAl ratios of
a) 0:1,b) 1:3. c) 1:2, and d)1:1 .....................................................

Figure 3.4 Variation of the Ce 3d/Al 2p intensity ratio as a function of the
Ce:Al atomic ratio for the SOOCey (I) and 900 Cey (O) ............

Figure 3.5 XRD patterns of catalysts calcined to 900 °C for Ce:Al ratios of
a) 0:1, b) 1:10, c) 1:4, and d) 1.3 ......................................

vii

12

15

19

21

23

25

36

38

40

43

46

55

57

6O

62

Figure 3.6 Variation of the Ce02 crystalline size and BET surface area
as a function of the CezAl atomic ratio for the 900 Cey catalysts

Figure 4.1 Powder X-ray diffraction patterns of the 10% Ce02/TiO2 films
calcined at a) 400 0C, b) 600 °C, c) 800 oC, and d) 1000 0C .........

Figure 4.2 Powder X-ray diffraction patterns of the 20% Ce02/1' iO2 films
calcined at a) 400 °C, b) 600 °C, c) 800 0C, and d) 1000 oC .........

Figure 4.3 Powder X-ray diffraction patterns of the 30% Ce02/TiO2 films
calcined at a) 400 °C, b) 600 °C, c) 800 °C, and d) 1000 °C .........

Figure 4.4 Powder X-ray diffraction patterns of the 40% Ce02/TiO2 films
calcined at a) 400 °C, b) 600 °C, c) 800 OC, and d) 1000 °C .........

Figure 4.5 Powder X-ray diffraction patterns of the 50% Ce02/TiO2 fihns
calcined at a) 400 °C, b) 600 °C, c) 800 °C, and d) 1000 °C .........

Figure 4.6 The variation of Ce surface enhancement versus the bulk Ce/T i
atomic ratio for the 8OOCTy (denoted as open squares) and
lOOOCTy (denoted as closed circles) catalysts ..............................

Figure 5.1 Cerium 3d XPS spectra of Ce02 (A) and Ce(III) acac (B).........

Figure 5.2 Ce 3d XPS spectra of Ce02/TiO2 films: (A) 400CTO.1, (B)
600CTO.1, (C) 400CT1.0, and (D) 6OOCT1.0 ...............................

Figure 5.3 Extent of photoreduction versus analysis time for a 10% CezTi
oxide film calcined at 400 °C ........................................................

Figure 5.4 XRD spectra of (A) 600CTO.1 and (B) 600CT1.0 ........................

Figure 5.5 Cerium 3d XPS spectra of a 400CTO.1 film before reduction (a)
and following H2 reduction at 200 °C (b), 250 °C (c), 300 °C ((1),
and 350 0C (e) ................................................................................

Figure 5.6 Extent of Cerium reduction versus reduction temperature for the
400CTO.1 (closed squares), 600CTO.1 (closed circles),
400CT1.0 (open squares), and 6OOCT1.0 (open circles) films .....

Figure 5.7 Cerium 3d XPS spectra of a 600CTO.1 film before reduction (a)

and following H2 reduction at 200 °C (b), 250 0C (c), 300 0C (d),
and 350 ° ............................................................

viii

66

75

77

8O

83

86

88

97

99

100

102

105

107

108

Table 1.1

Table 1.2

Table 3.1

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 5.1

C.

LIST OF TABLES

Effects of the rates of hydrolysis and condensation on film quality

Effect of drying techniques on the thermal properties of alumina

gels .........................................................................
Variation of BET surface area as a function of Ce loading and

calcination temperature ..................................................

Average particle size and surface Ce/T i atomic ratios for the
xCTO.11 films as a function of calcination temperature .................

Average particle size and surface Ce/T i atomic ratios for the
xCTO.25 films as a function of calcination temperature ..................

Average particle size and surface Ce/T i atomic ratios for the
xCTO.43 films as a function of calcination temperature .................

Average particle size and surface Ce/T i atomic ratios for the
xCTO.67 fihns as a function of calcination temperature. . . . .

Average particle size and surface Ce/T i atomic ratios for the
xCT1.0 films as a function of calcination temperature ...................

Observed and predicted oxygen loss for 10% and 50% CezTi
films calcined at 400 and 600 oC following H2 reduction at 350

60

75

79

82

83

85

110

Chapter 1
Introduction

1.1 Introductory Remarks.
Metal oxides are materials with useful properties for catalysis and sensors [1].
The ionic nature of the metal-oxygen lattice provides a substrate which facilitates the
dissociative adsorption of molecules. Cations and anions on the surface provide sites for
Lewis acid/base interactions. Since many transition metals can exist in a variety of
oxidation states, there is also an opportunity for redox chemistry. Finally, the presence of
ionosorbed oxygen and reactive lattice oxygen can assist in many oxidation reactions [1].
The choice of a catalyst for any reaction should involve the careful selection of the above
properties to ensure the optimal selectivity for a desired product. Ideally, the preparative
technique should be flexible enough to tailor the homogeneity, porosity and reactivity
(selectivity) of the catalyst. Until recently metal oxide catalysts have been prepared by
coprecipitation and sintering of metal salts or by chemical vapor deposition (CVD).
However, coprecipitation does not offer the control necessary to easily produce
homogenous mixtures, while CVD does not lend itself well to the production of highly
porous materials [2].
Sol-gel chemistry is an attractive wet chemical alternative to conventional
preparative techniques for metal oxides [3, 4]. The sol-gel process involves the
controlled hydrolysis and condensation of metal alkoxides to form metal oxide gels, thin

films or powders. Advantages over traditional techniques include the easy preparation of

multicomponent systems, lower processing temperatures, and expanded rheological
properties of the materials. Sol-gel chemistry allows control of the synthesis from the
precursor to the product, allowing for more controlled production of unique materials. At
the relatively low processing temperatures, organic functionalities can be incorporated
into the gel for the production of novel materials.

The project described herein is devoted to the design of catalytically useful
materials prepared from sol-gel chemistry. The first phase of the project studies the
manufacture of organically modified aluminum oxide thin films. Aluminum alkoxide by
itself reacts rapidly with water to form aluminum oxide powders. However, by replacing
the alkoxide ligands with less reactive species, the rate of hydrolysis is lowered and gels
and thin films can be produced. This work involves characterizing the bulk of the fihn
with Fourier transform infrared spectroscopy (FTIR) and the surface of the film with X-
ray photoelectron spectroscopy (XPS) to examine the effect of varying ligand-to-metal
ratios on film quality.

Aluminum oxide supported cerium oxide has been used in automobile exhaust
catalysts to promote the water-gas shifi reaction, suppress the CO inhibition effect during

CO oxidation and remove CO and NOX simultaneously with oxygen storage ability.

Cerium oxide has also been used as a promoter in base metal oxide catalysts. The second
part of this thesis involves tailoring the reactivity of aluminum and cerium alkoxides to
form alumina/ceria catalysts with high surface areas. A series of sol-gel derived

Ce02/A1203 catalysts calcined at 500°C and 900°C were prepared where the Ce/Al

atomic ratio was varied from 0 to 1.0. The catalysts were characterized by powder X-ray

diffraction (XRD), XPS and FTIR. The surface areas were determined by BET surface
area techniques.

The third and fourth sections of the thesis involve characterizing and studying the
reactivity of Ce02/TiO2 thin films. Mixtures of Ce02 and TiO2 are used in coatings,
catalysis, and solar energy cells. The materials were prepared using sol-gel chemistry and
characterized as a function of calcination temperature with XPS and XRD. Cerium oxide
reactivity was studied by H2 reduction of the materials at elevated temperatures between

200 °C and 350 °C. The reactivity was evaluated by XPS analysis of the reduced films.

1.2 Sol-Gel Chemistry

The properties of catalysts and catalytic supports are often dependent on their
preparation method. For instance, the dispersion and size of an active phase, the
homogeneity of a multicomponent mixture, and the porosity and surface area of catalysts
are all functions of the preparation method and the precursors used. Sol-gel chemistry is
a versatile method for the preparation of catalytic materials. The wide variety of synthetic
parameters available in sol-gel chemistry provides excellent control of the structure and
properties of the final catalyst. Sol-gel chemistry can proceed through an alkoxide or
colloidal route depending on the type of molecular precursor used [5,6]. The work
described herein involves alkoxides as precursors, so the alkoxide route will be the focus
of this discussion.
1.2.1 Overview of Sol-Gel Chemistry

Sol-gel chemistry is a two step process involving hydrolysis and condensation of a

metal alkoxide. The process begins with hydrolysis, whereby water reacts with the metal

alkoxide to replace the alkoxide ligand with a hydroxyl group. An example of hydrolysis

is the reaction of tetramethylorthosilane (TMOS) with water as shown in equation 1.1 [3]

Si(OCH3)4 + 4(HZO) —> Si(OH)4 + 4CH,OH (1.1)
Condensation immediately follows hydrolysis as the hydroxyl ligands react to form
metal-oxygen bonds and liberate water (1.2), thus continuing the cycle.

Si-OH+HO-Si—>Si—O—Si+HZO (1.2)
As hydroxylated metal alkoxides begin to network a gel is formed. The rate constants for
hydrolysis (kg) and condensation (kc) are a function of temperature, water-to-metal ratio,
pH, solvent and precursor composition [3].

Typically, the rate of hydrolysis will increase with temperature and water-to-metal
ratio (R) [3,7]. Another common method of controlling kn is by changing the alkoxide
ligand. Mackenzie et a1. [8] found that kn decreases with the increasing length of the
alkoxy group due to the increased hydrophobicity and steric protection provided by the
longer hydrocarbon chain. Oftentimes, an acid or base is used as a catalyst during the sol-
gel process. At acidic conditions hydrolysis occurs at a faster rate than condensation
while under basic conditions condensation is more rapid. The relative rates of hydrolysis
and condensation will have a direct bearing on the final structure of the gel. When kH is
large with respect to kc, the resulting gel is predominately unbranched. A highly
branched gel is obtained when kn is smaller than kg. The effect of the rates of hydrolysis
and condensation on gel quality has been generalized by Livage and is shown on the

following table [6].

Table 1.1 Effects of the rates of hydrolysis and condensation on film quality [6].

 

Hydrolysis Condensation Result
slow slow colloids/sols
fast slow polymeric gels
fast fast colloidal gel or gelatinous
precipitate
slow fast precipitation

 

 

 

Sol-gel chemistry with transition metal alkoxides is more complicated due to the
greater positive charge of the transition metal and its ability to undergo coordination
sphere expansion. Hydrolysis begins by the addition of water to the metal alkoxide

according to the mechanism proposed by Livage et a1. [4]

H H R
R0- M + O/——> R0 —M--O/ ——> \ -- M—OH

(a) \H (b) \H (c) \(d

M— OH + ROH
(1.3)
where the overall reaction is a nucleophilic substitution. Step (a) involves the
nucleophilic attack of a water molecule at the metal atom. The activation barrier for this
step depends on the coordination number of the metal and its oxidation state. The larger
the difference between the two; the lower the activation barrier. In transition state (b) the

metal coordination sphere expands by one. An intermolecular proton transfer leads to (c).

The activation barrier for this step depends on the acidity of the proton. The complex can
then separate at step (d) to give a metal hydroxide and an alcohol. Overall, the reaction
will be favored when the electrophillic character of the metal is strong and when the
leaving ability of the alcohol is high [3, 4].

The rate of hydrolysis for transition metals can be tailored by changing the
coordination and/or electrophillicity of the metal. A dominant factor that affects the
hydrolysis rate is the nature of the solvent and the concentration of electrolytes [3]. Non-
complexing solvents may cause the alkoxide to oligomerize. This reduces the rate of
hydrolysis since additional energy is required to dissociate the complex [6]. Additions of
electronegative ligands such as halides increase the electrophillicity of the metal and
therefore increases kH [6].

Like silicon alkoxides, the rate of transition metal alkoxide hydrolysis slows upon
increasing the chain length of the alkoxide ligand. The rate of hydrolysis is also modified
by complexation of the precursor with various chelating agents such as carboxylic acids,

geminal diols, and B-diketonates [9]. The increased metal-ligand bond strength following

the addition of a more electronegative ligand may also inhibit hydrolysis [9]. A larger
ligand might also slow hydrolysis due to hydrophobic repulsion and steric protection of
the metal. I

Condensation begins as soon as hydroxyl groups are generated. Condensation
may involve three competitive mechanisms: alcoxolation, oxolation and olation [4].
Alcoxolation involves the reaction of a metal hydroxide and alkoxide to remove an

alcohol and form a metal-oxygen oligomer.

H H

l l | l
—M—O + M-OR —>—M—O--M-OR

/.

M—O-M-“O/ —> M—O —M + ROH
\H
(1.4)

Oxolation involves the reaction of two metal hydroxides and the loss of a proton.

l l |
—M—O + M-OH —> —M—O--M-OH
I I

M—O —M "'O< -—> M—O —M + H20
H (1.5)
Olation occurs with alkoxides when the coordination of the metal is not saturated. It is

characterized by the removal of a water or alcohol molecular and the formation of

hydroxyl bridges between the metal atoms.

(1.63)
/H i‘
M-OH + M--O —> M_O_M + H20
\H
H H
/ I
M-OH + M-O\ —> M—o—M + HOR
H (1.6b)

Olation is fast compared to other condensation mechanisms due to the lack of a proton
transfer step (step (c) in equation 1.3) in the mechanism [4].
Studies with titanium alkoxides have shown that the waterzmetal ratio can control

the mechanism of condensation [10]. This was determined by monitoring the weight

percent of TiO2 in sol-gel derived titanium polymers. The higher percentage of TiO2 the
higher the complexity of titanium polymer. Yoldas [10] found that the weight percent of
the oxide increased with water ratio. This and similar findings led Livage to derive the
following generalization. For low water ratios (h<l), alcoxolation and oxolation lead to
the formation of alkoxy bridges is preferred and as a result polymers are rarely obtained. .-
For hydrolysis conditions where 1 < h< n (n being the coordination number of the metal)
a polymer may be formed. However cross-linking will be low and a species such as the

following will be formed:

xM(0H)(OR) —> —M — O —M — O —1JI— + xROH

For systems where h>n, extensive cross-linking will results which will lead to an

increased oxide content of the polymer.

1.2.2 Sol-Gel Preparation of Catalysts

Many catalysts and catalyst supports are multicomponent mixtures. Tailoring the
reactivity of the alkoxides can control the homogeneity of the final material. The
reactivity of two alkoxides must be fairly well matched to achieve a homogenous
mixture. If there is a difference in reactivities, the more reactive alkoxide tends to form a
core onto which the less reactive component attaches [1 1]. In general, matching the
precursor reactivities can be accomplished by modifying the precursor with a different

alkoxy group or other ligand to slow the rate of hydrolysis, ‘prereacting’ a less reactive

alkoxide with water to give it a head start (commonly known as prehydrolysis), or by
using bimetallic alkoxide precursors [11].

The use of chelating agents and different alkoxy groups to reduce kn has been
described previously. An example of prehydrolysis for the synthesis of multicomponent
oxides has been outlined by Walther et al. [12] and Schraml-Marth et a]. [13] for the
development of TiO2-SiO2 catalyst supports. Simultaneous hydrolysis of both alkoxides
produces a gel consisting of segregated TiO2 and SiO2 phases. When the silicon alkoxide
is hydrolyzed before the titanium alkoxide, a homogenous mixture of TiO2 and SiO2 is
obtained. A final method for the preparation of multicomponent oxides is the use of
bimetallic alkoxide precursors. Narula [14] has prepared homogenous Ce02/A1203 and
La2O3/Al203 catalysts by hydrolyzing Ce/Al and La/Al bimetallic alkoxide precursors.
~ Though Ce and La are well dispersed throughout the alumna, the rare earth concentration
is limited to the stoichiometry of the precursor.

The surface area and porosity of a sol-gel derived catalyst can be controlled by
modifying the precursor chemistry, or the aging, and drying of the gel [11]. As described
previously, the relative rates of hydrolysis and condensation have a dramatic effect on the
final structure of the gel [4]. When hydrolysis is rapid versus condensation the resulting
gel is poorly branched and microporous. When condensation is more rapid than
hydrolysis, the resulting gel is highly branched with a larger pore size.

The structure of the gel will continue to change if it is allowed to age in its
solvent. Several processes can occur through aging including condensation, syneresis,
and coarsening [3]. Condensation will continue to occur within the gel network as long

as neighboring hydroxyls are within a close enough proximity to react. Over time, the gel

10

should become more crosslinked. Syneresis is the shrinkage of the gel following
extensive condensation and expulsion of liquid from the pores. Coarsening is the
irreversible decrease in the surface area through dissolution and reprecipitation of the gel
[3].

As the solvent within the pores of the gels is removed by evaporation, the gel
network is deformed by large capillary forces which results in shrinkage. The pore
structure of the gel can be maintained during drying by minimizing or eliminating the
capillary pressure of the liquid-vapor interface within the pore [11]. The capillary
pressure (P) is defined as

P = 20' cos(0)/r (1.8)
where 0' is the surface tension, 6 is the contact angle between the liquid and solid and r is

the pore radius. One method to reduce the capillary pressures is the use of solvents with
low surface tensions [15, 16]. More commonly, various drying methods are used to
remove the liquid-vapor interface altogether. Removing the solvent at low temperatures
and pressures form xerogels, which usually have a high surface area, but are microporous.

Supercritical drying results in an aerogel, a material with high surface area and porosity.

1.3 The Role of Cerium Oxide in Catalysis

Cerium oxide has played an important role in many catalytic systems as an active
phase, a promoter, or a support. Cerium oxide is a key component in three-way catalysts
for the treatment of exhaust gas from automobiles [17-19]. In addition, Ce02 is used in
the removal of SOx from fluid catalytic cracking flue gases [20] and is an important

constituent of several oxidation catalysts [21-27]. Cerium’s use in three-way catalysis is

11

a result of beneficial precious metal-ceria interactions and the activity of the Ce‘w/Ce3+
redox couple. The redox couple is useful due to its ability to reversibly change from
Ce02 under oxidizing conditions to Ce2O3 under reducing conditions. Other useful
properties of ceria include enhancement of precious metal dispersion, promotion of the
water-gas shift reaction, and improvement of the thermal stability of the support [28].

A great deal of insight concerning the redox chemistry of ceria has been gained by
studying the reduction of ceria and ceria containing catalysts by H2. Temperature
programmed reduction (TPR) studies performed by Fiero et a1. [29], have found that H2
reduction is a two step process involving the chemisorption of H2 followed by the
removal of oxygen. Hydrogen chemisorption is an activated process on Ce02 leading to
the formation of hydroxyl groups and is represented by the following reaction.

Ce02 + x/2 H2 -—> CeO2H, (1.9)
At temperatures above 200 oC oxygen anions begin to be removed as desorbing water
molecules

CeO2Hx —-> Ce02.” + x/2 H2O (1.10)
or with hydroxyl groups reacting with hydrogen to produce water.

CeO2Hx + x/2 H2 —> Ce02, + x H2O (1.11)

F iero and coworkers have claimed that H2 diffuses into the bulk of the ceria to
form metal bronzes [29]. However, Bemal et a1. [30] have proven that ceria reduction is
a surface phenomena. As a result, a kinetic model for ceria reduction has been proposed
which is shown on Figure 1.1 [31]. The reaction begins with the dissociative
chemisorption of hydrogen to form hydroxyl groups (1). Next, an oxygen vacancy is

formed which reduces two neighboring Ce4+ ions (2). This is followed by desorption of a

12

water molecule (3). The surface of the material is then ‘healed’ by diffusion of surface
anionic vacancies into the bulk material (4).

The progress of reduction is sensitive to the surface area of the sample and bulk
reduction only occurs when all surface sites are fully reduced. TPR experiments on Ce02 ‘
reveal two distinct reduction peaks at approximately 500 and 750 °C [32]. The 500 °C
peak is assigned to the removal of ‘surface capping’ oxygen anions attached to a surface
Ce4+ ion in an octahedral coordination. The 750 °C peak is due to the removal of a bulk
oxygen anion bound to two or more Ce4+ ions.

The surface area of Ce02 is observed to have a strong effect on the
oxidation/reduction cycle of Ce. Yao and Yu Yao [32] reported that the relative area of
the aforementioned 500 oC reduction peak increases with increasing Ce02 surface area.
Johnson and Mooi [33] related the number of capping oxides to the cerium oxide particle
size. Assuming that ceria crystallite exist as cubes of n oxide ions on a side and the ionic
radius of the oxide ion is 1.40 A, the crystallite size (a) is related to n by

a=2.80n (inA) (1.12)

The number of capping oxide ions (0c) can be determined by Equation 1.13.

_ 2(6n2 —12n+8)

C n3

0 (1.13)

 

(1)

(2)

(3)

(4)

 

 

l3

 

 

 

 

 

 

 

 

 

Figure 1.1 Scheme for the reduction of Ce02 by H2. Reduction begins by the

dissociative chemisorption of H2 (1) to produce two hydroxyl groups (2).
The removal of one oxygen atom to form water (3) produces an oxygen
vacancy and reduces two neighboring cerium atoms. The surface is healed
(4) by migration of the oxygen vacancy to the bulk of the material [31].

14

The number of capping oxides will then increase with decreasing crystallite size. Thus,
minimizing the Ce02 particle size is desirable to ensure high catalytic activity at lower
temperatures.

When supported on alumina or silica, the reductive properties of ceria are
dependent on the extent and nature of its interaction with the support [26, 32, 34, 35].
TPR studies of Ce02/A1203 catalysts reveal that the shape of the TPR curve is dependent
on the crystallite size of Ce02 with the high temperature peak attributed to bulk oxide
removal only observed at high Ce02 loadings [32]. The position of the high temperature
peak was also found to shift to higher temperatures with increased Ce loading. At low
Ce02 loadings, a CeAlO3 phase is observed by XPS and Raman spectroscopy following
H2 reduction [32]. The Ce‘H/Ce3+ redox cycle for low loadings is described as a flip-flop
between a dispersed Ce02 phase under oxidizing conditions and CeAlO3 under reducing
conditions.

Recently, Ce02-TiO2 mixed oxides have been explored as catalytic supports [36-
38]. Ce02 is easily reduced in ceria-titania mixtures and titania has been shown to
stabilize the Ce3+ ion [36]. Other mixed cerium oxides include Ce02-ZrO2 [39-44],
Ce02-HfO2 [40], Ce-Pr mixed oxides [45], and Ce-Tb-La mixed oxides. Insertion of Zr
or Hf atoms into the fluorite structure of Ce02 facilitates reduction by increasing bulk
oxygen mobility[44]. Combining a mixed valence rare earth oxide such as Pr60n or
Tb407 with Ce02 increase the oxygen storage capabilities of Ce02 and lowers the
reduction temperature of surface oxides due to the formation of intrinsic and extrinsic

lattice defects [45].

 

15

1.4 Material Characterization Techniques
1.4.1 X-ray Photoelectron Spectroscopy (XPS)

Photoelectron spectroscopy utilizes the photoemission of electrons to obtain
quantitative and qualitative information about the structure and composition of the
surface of a material [46-48]. In ultraviolet photoelectron spectroscopy (UPS), ultraviolet
radiation is used as a source to obtain information concerning the valence band of a
material. However, in XPS X-ray radiation can obtain information concerning both
valence and core electrons. An analyzer measures the kinetic energy of photoemitted
electrons.

Though the absorption depth of X-rays is large, ejected electrons from the bulk of

the material are attenuated due to inelastic scattering [48]. The mean free path (A) for

photoemitted electrons in a solid is determined from Equation 1.14.

2=__E__
a(lnE+b)

(1.14)
where E is the incident photon energy and a and b are parameters which separately
account for the concentrations of valence and core electrons in the material [49]. Ninety
five percent of the photoelectrons which reach the detector emerge from within 31 of the
surface. Typically, this limits the sampling depth of XPS to a maximum of 100 A

The binding energy (133) of a photoemitted electron is determined from the

following equation [48]

EB=hV “(EK+¢spcc) (1-15)

l6

 

 

 

 

 

 

 

Al Anode
O ls
Ti 2p
Ce 3d
C Is
Ce 4d
I I 1
1200 800 400 0
Binding Energy (eV)

Figure 1.2 XPS spectrum of a 50% Ce02/TiO2 thin film. The spectrum was obtained
using an Al anode operated at 300 W. The binding energies are referenced
to the aliphatic C Is peak at 284.6 eV.

17

where EK is the kinetic energy of the ejected electron and ospcc is the work function of the

spectrometer. The binding energy is characteristic of the atomic orbital from which the
electron is emitted. Binding energies differ from ionization energies by final state effects
due to relaxation processes that occur simultaneously with photoemission. Since each
element has a unique set of binding energies associated with it, the elemental composition
of a surface can be determined from its XPS spectrum. Figure 1.2 shows the
photoemission spectrum of a 50% CeO2zTiO2 powder. Peaks shown are assigned to Ce,
Ti, 0 and C.

Information concerning the binding environment of an atom can also be
determined by XPS. Figure 1.3 shows the carbon ls spectrum for an acac modified
alumina thin film. Three distinct peaks are discerned and assigned to methyl or
methylene carbon (284.6 eV), alkoxy carbon (286.7 eV) and carboxylate carbon (288 eV).
As the electronegativity of neighboring elements increase, electrons are more tightly
bound and the binding energy shifts to higher energy. Thus, atoms in different binding
environments will show different binding energies for the atom. The relative
concentrations of the different types of carbon can be determined from their relative peak
areas.

The intensity of the signal for a particular element i (1,) is defined by Equation
1.16

Ii =IonioiD(8i)}“i(8i) (1°16)

where 10 is the flux of the X-ray, n, is the concentration of species i, 0'; is the

photoionization cross-section of species i, D(ei) is the efficiency of the analyzer and Ma)

18

 

C Is
Mg Anode

 

 

 

I l l
282 280

 

 

Binding Energy (eV)

C Is spectrum of an acac modified alumina film. The solid lines represent

Figure 1.3.
the deconvoluted spectra and individual C ls peaks.

19

is the mean free path of the emitted electron [50]. Because of difficulties in measuring 10

and D(ei), absolute concentrations are seldom measured with XPS. However, if the

relative concentration (C) of two elements A and B are calculated:
-0.5
EA. =I_A_ ___°B"BS§M (1.17)
C B IB oAAAe A ‘

only the kinetic energy dependence of D(ei) is retained. The intensity of a peak is taken
to be its area following background subtraction. The term in the right bracket is the
sensitivity factor. Sensitivity factors can be determined theoretically using cross sections
determined by Scofield [50] and mean free paths from a universal curve [51]. However,

mean free paths are material dependent and the cf“ term is instrument dependent.

Sensitivity factors can also be determined experimentally by analyzing standard materials

of known composition.

1.4.2 XPS of Cerium Oxides

The Ce 3d photoelectron spectrum is very complicated due to a variety of splitting
and relaxation effects that occur during the photoemission of core electrons [52-58]. The
XPS Ce 3d spectra of Ce02 and Ce(III) acetylacetonate is shown on Figure 1.4a and b,
respectively. The lines labeled v" and u" (where n is a tick mark representing individual
satellite peaks) correspond to the 3d5/2 and 3d3/2 states, respectively. The different lines
are the result of different final state electronic configurations. Following the ejection of a
core level photoelectron, the unoccupied Ce 4f orbital relaxes to the point where mixing

can occur with valence 0 2p orbitals. As a result three different final state electronic

20

 

 

 

 

J l l

 

920 900 880

Binding Energy (eV)

Figure 1.4 Cerium 3d XPS spectra of Ce02 (A) and Ce(III) acetylacetonate (B). The
peaks labeled v0 (110) and v’ (u’ ) refer to final states Ce(III)[4f]l O[2p]5
and Ce(III)[4f]2 O[2p]4 , respectively. The features labeled v (u) and v’ ’
(u’ ’ ) refer correspond to a mixing of the Ce(IV)[4f]l O[2p]5 and
Ce(IV)[4f]2 O[2p]4 final state configurations. The v’ ’ ’ (u’ ’ ’ ) peak
corresponds to the Ce(IV)[4f]o O[2p]6 final state.

21

configurations can exist: 4f0, 4f‘, and 4f. The different electronic configurations lead to
different binding energy satellites in the 3d spectra. Well-screened states are shifted to
lower binding energies and poorly screened states are shifted to higher binding energies.
The v’” (u”’) satellite corresponds to a Ce(IV)[3d]9[5d6s]°[4f]0 O[2p]6 final state while
the v (u) and v ' ’(u ’ ') features are assigned to a mixing of the final Ce(IV)[3d]9[5d6s]°[4f]l
O[2p]5 and Ce(IV)[3d]9[5d6s]°[4f]2 O[2p]4 configurations.

Cerium (III) compounds have an initial state 4fl configuration. Thus, relaxation
following the photoemission of core electron results in only two different final states 4fl
and 4f2. The v0 (up) and v’ (u') features are assigned to the Ce(III)[3d]9[5d6s]°[4f]l
O[2p]6 and Ce(III) [3d]9[5d6s]0[4f]2 O[2p]5 final states, respectively. The different final
state configurations for Ce(III) and Ce(IV) compounds are quite valuable in
distinguishing between the two valence states. Typically, the amount of Ce(III) can be
deduced by measuring the area ratios of the v0 (110) and v’ (u ’) peaks to the total area of
the Ce 3d spectra. However, it is difficult to accurately curve fit all ten peaks associated
with mixed valence cerium oxides. In this work, the Ce(IV) content was determined by
taking the ratio of the u”’ peak area to the total area of the Ce 3d spectrum.

Cerium 3d spectra are further complicated by a reduction of the surface cerium, a
phenomena observed with many Ce(IV) containing compounds[59-64]. This process,
termed photoreduction, is attributed to many effects including X-ray flux [65], integrated
X-ray dose [66], secondary electrons from the X-ray source [67], sample charging [61],

temperature [68] and high vacuum [69]. The extent of photoreduction is also affected by

22

sample morphology, as described by Park and Ledford [64] who noted that crystalline

Ce02 reduces at a less rapidly than amorphous ceria.

1.4.3 Powder X-ray Diffraction

As X-rays pass through a single crystal, they are scattered by atoms in the crystal
lattice. Interferences (both constructive and destructive) occur because the spacing
between the atoms is on the same order of magnitude as the wavelength of the incident X-
rays The angle of reflection is related to the incident wavelength by the Bragg equation

nA=2dsin0 (1.18)
where A is the incident wavelength, n is a series of integers (l, 2, 3...), and d is the
interplanar spacing.

Powder XRD uses a polycrystalline sample, typically a powder. Ideally, all
possible orientations of all possible lattice planes are present in a powder sample and the
incident radiation is simultaneously scattered for all lattice planes. A diffraction pattern
for a particular powder is obtained by rotating the detector thorugh all angles of
reflection. XRD peaks appear only if the angle of incidence for the X-ray satisfies the
condition Braggs law. At all other angles, destructive interference occurs. The relative
intensities of the reflections depend on the kind of atoms and their arrangement between
the lattice planes. A powder diffraction pattern therefore is a reasonably complete display
of the diffraction from a compound. There is a particle size limitation to powder XRD.
If the sample contains amorphous phase or very small crystallites (< 3 nm), no diffi'action
peaks will be Observed due to destructive interference in scattering directions where the

X-rays are out of phase [70].

23

The peak width of a single large crystal is very narrow, and the width of the peak
is made up primarily of instrumental broadening factors. However, as when the size of
the crystallites decreases to less than 100 nm, pure diffraction broadening becomes a
major factor in the breadth of a peak [70]. Pure diffraction broadening increases with
decreasing crystallite size until a point is reached where peaks become so broad that they
are no longer distinguished from the background (typically for crystallites smaller than
Burn). The mean crystallite size (3) of crystal particles is determined from the Scherrer
Equation [70].

K)»
BsinO

 

a: (1.18)

where A is the X-ray wavelength and K is the particle shape factor, taken as 0.9. The

dimension d is defined as the effective thickness of the crystallite in a direction
perpendicular to the reflecting planes. On the basis of this definition there is only a small

dependence of K upon the crystallite shape. [3 is the full width at half maximum in

radians. 9 is the angle between the beam and normal on the reflecting plane (half of 20).

1.4.4 Surface Area Measurments (BET method)

The adsorption of a particular molecular species from a gas or liquid phase onto
the surface of a solid is the principal method for measuring the total surface area of
porous structures. If one can control the conditions of monolayer coverage and calculate
the cross section of the molecule, then the surface area of a sample can be obtained from

the volume of the adsorbed molecules. The molecules used for measuring surface area

 

24

should be inert, small, spherical and easy to handle at the required temperature. Though
not spherical, N2 is usually used due to its availability and low cost.

The most common method for measuring the surface areas of catalysts is that
developed by Brunauer, Emmett, and Teller (called the BET method) [71]. They
extended the Langmuir mechanism [72] to multimolecular layers. Langmuir assumed
that an adsorption site on the solid surface accommodates one adsorbed molecule and the
adsorption sites are uniform regardless of surface coverage, so that the adsorption
probability (heat of adsorption) is the same at all sites. He also proposed that the rate of
evaporation equals the rate of condensation for the first layer under equilibrium condition.
For multimolecular layers, Brunauer, Emmett, and Teller assumed that the adsorption rate
was proportional to the vacant sites of the lower layer and that the desorption rate was
proportional to the adsorbed molecules in that layer. They also assumed that the heat of
adsorption for all layers except the first one is equal to the heat of condensation of the
molecule. The summation over an infinite number of adsorbed layers gives the following

equation;

P _ 1 +(C-1)P

—_ (1.19)
V(Po-P) va v...CP.

 

where V is the volume of gas adsorbed at pressure P, Vm is the volume of gas adsorbed in
monolayer, P0 is the saturation pressure of adsorbate gas at the experimental temperature,
C is a constant related exponentially to the heats of adsorption and condensation of the

gas. The BET equation yields a straight line when PN(Po-P) is plotted versus P/Po. The

25

BET plot is usually found to be linear in the range P/Po = 0.05-0.35, and this range is
usually used for surface area measurements. The deviation from linearity increases at
higher P/Po values (> 0.35) due to multilayer adsorption and/or pore condensation. The
experimental error increases at lower P/Po (< 0.05) due to the small amount of molecules

adsorbed. The slope (S) is expressed as

 

S = £12., (120)
VaC
and the intercept (I) is expressed as
I = 1 (l 21)
VmC '

Solving the equations (1.20) and (1.21) for Vnrl gives

1
Vm= — 1.22
S+I ( )

The surface area of the catalyst can be calculated from Vm if the cross sectional area of an
adsorbed molecule is known. Practically, the BET measurement can be performed with
the value of the weight adsorbed on the sample (W & Wm) instead of volume (V & Vm)
[73]. From the PVT relationship the amount of adsorbate adsorbed on the sample can be

calculated from calibrated integrator counts caused by thermal desorption of the adsorbate

26

W= A VealPM

1.23
Acal RT ( )

 

 

where W is the mass of adsorbate adsorbed on the sample, A is the sample integrator
counts, A“. is the calibration integrator counts, Veal is the calibration volume (cm3), P is

the ambient pressure, M is the adsorbate molecular weight (g), R is the gas constant (82.1

. _'nq

mL atrn deg'l mol"). Therefore equation (2) and (5) can be rewritten as

 

 

 

 

P = 1 +(C-1)P (1.24)
W(Pc — P) WmC WnCPe
W = -—1— (1 25)
m S + I °
The total. surface area of the sample (St) is determined from following equation;
WmNAcs
St= M (1.26)

where Wm is the weight of adsorbate adsorbed at a coverage of one monolayer, N is
Avogadro’s number, Acs is the cross sectional area of adsorbate molecule (N2 = 0.162
nmz), M is the adsorbate molecular weight (g). The specific surface area S is given by

following equation

S = , S‘ (1.27)
welght of sample (g)

 

1.5

10

10.

ll.

13.

14.

15.

l6.

l7.

18.

27

References.

Kung, H. H. Transition Metal Oxides: Surface Chemistry and Catalysis, Elsevier:
Amsterdam, 1989.

Gugliemi, M.; Carturan, G. J. Non-Crystalline Solids 1988, 100, 16.

Hench, L. L.; West, J. K. Chem. Rev. 1990, 90, 33.

Livage, J .; Henry, M.; Sanchez, C. Prog. Solid St. Chem. 1988, I8, 259.
Gugliemi, M.; Carturan, G. J. Non-Crystalline Solids 1988, 100, 16.

Sanchez, C.; Livage, J .; Henry, M.; Babonneau, F. J. Non-Crystalline Solids 1988,
100, 65.

Aelion, R.; Loebbel, A.; Eirich, F. J. Am. Chem. Soc. 1950, 72, 5705.

Mackenzie, J. D.; Chen, K. C.; Tsuchiya. T J. Non-Crystalline Solids 1986, 81,
227.

Sanchez, C.; Livage, J. New J. Chem. 1990, 14, 513.
Yoldas, B. E. J Mat. Sci. 1986, 21, 1087.
Ward, D. A.; Ko, E. 1.1nd. Eng. Chem. Res. 1995, 34, 421.

Walther, K. L.; Wokaun, A.; Handy, B. E.; Baiker, A. J. Non-Crystal]. Solids
1991, 134, 47.

Schraml-Marth, M.;Walther, K. L.; Wokaun, A.; Handy, B. E.; Baiker, A. J. Non-
Crystall. Solids 1991, 143, 93.

Narula, C. In Better Ceramics through Chemistry V; Mater. Res. Soc. Symp.
Proc.; Hampden-Smith, M. J .; Klemperer, W. G.; Brinker, C. J ., Eds; Materials
Research Society: Pittsburgh PA, 1992; Vol 271, 567.

Smith, D. M.; Desphande, R.; Brinker, C. J. Ibid, 567.

Desphande, R.; Smith, D. M.; Brinker, C. J. Ibid, 553.

Summers, J. C.; Ausen, S. A. J. Catal. 1979, 58, 131,

Kim, G. Ind. Eng. Chem. Prod. Res. Dev. 1982, 21 , 267.

‘_ —.‘-—-—-—-—'

 

19.

28.

29

30.

31.

32.

34.

35.

36.

28

Gandhi, H. S.; Piken, A. G.; Shelef, M.; Delosh, R. G. SAE Paper 760201 1976,
55.

Bhattacharyya, A. A.; Wolterrnann, G. M.; Yoo, J. S.; Karch, J. A.; Cormier, W.
E. Ind. Eng. Chem. Res. 1988, 27, 1356.

Liu, W.; Flytzani-Stephanopoulos, M. J. Catal. 1995, 153, 304.
Irnarnura, S.; Uematsu, Y.; Utani, K.; Ito, T. Ind. Eng. Chem. Res. 1991, 30, 18.
Brazdil, J. F.; Grasselli, R. K. J. Catal. 1983, 79, 104.

Zamar, F .; Trovarelli, A.; de Leitenberg, C.; Dolcetti, G. J. Chem. Soc. Chem.
Commun. 1995, 965.

Yu, Z.; Yang, X.; Lunsford, J. H.; Rosynek, M. P. J. Catal. 1995, 154, 163.

Haneda, M.; Mizushima, T.; Kakuta, N.; Ueno, A.; Sato, Y.; Matsuura, S.;
Kasahara, S.; Sato, M. Bull. Chem. Soc. Jpn. 1993, 66, 1279.

Amrikhanova, A.; Krichevskii, L. A.; Katalitskii, A. D. Kinet. Catal. 1994, 35,
838.

Trovarelli, A. Catal. Rev. 1995, 439.
Fiero, J. L. G.; Soria, J.; Sanz, J.; Rojo, J. M. J. Sol. State Chem. 1987, 66, 154.

Bemal, S.; Calvino, J. J .; Cifredo, G. A.; Gatica, J. M.; Omil, J. A. P.; Pintado, J.
M. J. Chem. Soc. Faraday Trans. 1993, 89, 3499.

El Fallah, J .; Boujana, S.; Dexpert, H. Kiennemann, A.; Majerus, J .; Touret, 0.;
Villain, F.; Le Normand, F. .1. Phys. Chem. 1994, 98, 5522.

Yao, H. C.; Yu Yao, Y. F. J. Catal. 1984, 86, 254.

Johnson, M. F. L.; Mooi, J. J. Catal. 1987, 103, 502.: J. Catal. 1993, 140, 612.
Shyu, J. 2.; Weber, W. H.; Gandhi, H. S. J. Phys. Chem. 1988, 92, 4964.
Nunan, J. G.; Robota, J.; Cohn, M. J.; Bradley, S. A. J. Catal. 1992, I33, 309.

Dauscher, A.; Hilaire, L.; Le Normand, F.; Muller, W.; Maire, G.; Vasquez, A.
Surf Inter. Anal. 1990, I6, 341.

37.

38.

39.

40.

41.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

29

Dauscher, A.; Wehrer, P.; Hilaire, L. Catal. Lett., 1992, 14, 171.
Guglielminotti, E.; Boccuzzi, F. J. Mol. Catal. A., 1996, 104, 273.

Ranga Rao, G.; Kaspar, J .; Meriani, 8.; di Monte, R.; Graziani, M. Catal. Lett.,
1994, 24, 107.

Zanar, F.; Trovarelli, A.; de Leitenburg, C.; Dolcetti, G. J. Chem. Soc, Chem.
Commun, 1995, 965.

de Leitenberg, C.; Trovarelli, A.; Zamar, F.; Maschio, S.; Dolcetti, G.; Llorca, J.
J. Chem. Soc., Chem. Commun, 1995, 2181.

Balducci, G.; Fomasiero, P.; Di Monte, R.; Kaspar, J .; Meriani, S.; Graziani, M.
Catal. Lett., 1995, 193.

Fomasiero, P.; Di Monte, R.; Ranga Rao, G.; Kaspar, J .; Meriani, S.; Trovarelli,
A.; Graziani, M. J Catal., 1995, 151, 168.

de Leitenberg, C.; Trovarelli, A.; Llorca, J.; Cavani, F.; Bini, G. Appl. Cat. A.,
1996, 161.

Logan, A. D.; Shelef, M. J. Mat. Res, 1994, 9, 468.

Briggs, D.; Seah, M. P. Practical Surface Analysis, 2nd ed. Wiley: New York,
1990.

Woodruff, D. P.; Delchar, T. A.; Modern Techniques of Surface Science,
Cambridge University Press: New York, 1989.

Wertheim, G. K. in Solid State Chemistry Techniques, Cheetharn, A. K.; Day, P.;
eds.; Clarendon Press: Oxford, 1987.

Reich, T.; Yarzehmski, V. G.; Nefedov, V. I. J. Electron Spect: Relat. Phenom.
1985, 46, 255.

Scofield, J. H. J. Electron Spects Relat. Phenom. 1979, 8, 129.
Seah, M. P.; Dench, W. A. Surf Interface Anal. 1979, 1, 2.

Burroughs, P.; Harnmet, A.; Orchard, A. F.; Thornton, G. J. Chem. Soc. Dalton
Trans. 1976, 1686.

53.
54.
55. .

56.

57.

58.
59.
60.

61.

62.

64.

65.

66.

67.

68.

69.

70.

3O

Thornton, G.; Dempsey, M. J. Chem. Phys. Lett. 1981, 77, 409.
Koelling, D. D.; Boring, A. M.; Wood, J. H. Solid State Comm. 1983, 4 7, 227.
Fujirnori, A. Phys. Rev. B, 1983, 28, 4489.

Le Normand, F.; El Fallah, J .; Hilaire, L.; Legare, P.; Kotani, A.; Parlebas, J. C.
Solid State Comm. 1989, 71, 885.

Wuilloud, E.; Delley, B.; Schneider, W. -D.; Baer, Y. Phys. Rev. Lett. 1984, 53,
202.

Kotani, A.; Jo, T.; Parlebas, J. C. Adv. Phys. 1988, 37, 37.
Paparazzo, E. Surf Sci. 1990, 234, 253.
Paparazzo, E. J. Vac. Sci. Technol. A 1991, 9, 1416.

El Fallah, J .; Hilaire, L.; Romeo, M.; Le Normand, F. J. Electron Spectrosc. Relat.
Phenom. 1995, 73, 89.

Dauscher, A.; Hilaire, L.; Le Normand, F.; Muller, W.; Maire, G.; Vasquez, A.
Surf Inter. Anal. 1990, 16, 341.

Dauscher, A.; Wehrer, P.; Hilaire, L. Catal. Lett., 1992, 14, 171.
Park, P. W.; Ledford, J. S. Langmuir 1996, 12, 1794.

Copperthwaite, R. G.; Lloyd, J. J. Electron Spectrosc. Relat. Phenom. 1978, 14,
159.

Batista-Leal, M, Lester, J. E. ,Lucchesi, C. A. J. Electron Spectrosc. Relat.
Phenom. 1977, 11,333.

Johnson, C. E. In Electronic States of Inorganic Compounds: New Experimental
Techniques; NATO Advanced Study Institutes, University of Oxford, 1974; Day,
P. Ed.; D. Reidel Publishing Company: Dordrecht-Holland, 1975; 409.

Hirokawa, K.; Honda, F.; Oku, M. J. Electron Spectrosc. Relat. Phenom. 1975, 6,
333.

Bumess, J. H.; Dillard, J. G.; Taylor, L. T. J. Am. Chem. Soc. 1975, 97, 6080.

Nuffield, E. W. X-ray Difiiraction Methods, Wiley: New York, 1966.

71.

73..

31

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

Langmuir, I. J. Amer. Chem. Soc. 1916, 38, 2221.

Manual of Quantasorb Jr. Quantachrome 1985.

Chapter 2

Preparation and Characterization of Organically Modified Aluminum
Oxide Thin Films.

2.1 Abstract

Organically modified almninum oxide thin films were prepared by the sol-gel
method. Either valeric acid or 2,4-pentanedione (acac) was added to solutions of
aluminum iSOpropoxide in an isopropanol/toluene mixture. The effect of various
acid:metal ratios (R.) or acac:metal ratios (R) on the quality of films spin cast from these
solutions was examined by visual inspection. A combination of Fourier transform
infrared (FTIR) spectroscopy and X-ray photoelectron spectroscopy (XPS) was used to
determine film structure and composition including ligand coordination and extent of
ligation. The chelating ligands used in this work dramatically affect the aluminum oxide
backbone and the hydroxyl environment of the films. An increase in chelator-to-metal
ratio was found to lead to an increase in the extent of ligation for both valerate and acac
modified precursors. The increase in ligation leads to a constant decrease in aluminum
oxide and hydroxyl absorbances for the acac modified films. The general appearance of
the spectra was found to change from an unmodified film (RL=0) at low ligand ratios to
an Al(acac)3 type material with increasing RL. The increase in valerate ligation leads to a
change in aluminum oxide coordination as well as a change in the hydroxyl absorbances.
The organic modifiers used in this work improved film quality by decreasing the stiffness

of the aluminum oxide backbone because of a decreased extent of crosslinking.

32

33

2.2 Introduction

Sol-gel chemistry is a proven route for producing aluminum oxide gels, glasses,
coatings, and powders. With sol-gel chemistry, hydrolysis and condensation of metal
hydroxyl or alkoxide ligands produce metal oxide polymers. Control of the rate and
extent of hydrolysis at the precursor stage allows for the controlled production of unique
materials [1]. Aluminum sol-gel chemistry was developed by Yoldas et al. [2-4] who
produced transparent gels by acid peptiziation of alkoxides at high water to metal ratios.
The sols produced by the Yoldas method consist of alumina particles whose size and
shape are a function of acid composition and concentration.

The tension in the liquid within the pores of a film is supported by the solid pore
walls. As the film drys, the increased capillary pressures generated by solvent
evaporation pulss the pores rapidly inward, causing the film to shrink [5]. If the film
adheres to the substrate, the reduction in volume decreases the film thickness. When the
film begins to solidify to the point where the stresses in it can no longer be relieved by
flow, cracks may develop [5]. Cracking can be reduced by decreasing the thickness of the
film or by a reducing the size of the polymers of the alumina sol [5]. Decreasing the

stiffness of the material can also reduce cracking. Complexing agents such as B-

diketonates [6-8], carboxylic acids [6,9-11] and glycols [6] have been used to obtain new
precursors which may produce less rigid films.

While the effects of chelating agents on the structure of the gels have been well
documented, the effect of chelating agents on the film quality has not been extensively
examined. In particular, the effect of ligand concentration on the structure of aluminum

oxide films is not well known. It is the goal of this chapter to examine the influence of

34

acetylacetone and valeric acid on film quality, composition, and structure using X-ray

photoelectron spectroscopy (XPS) and Fourier transform infrared spectroscopy (FTIR).

2.3 Experimental

Materials. Aluminum isopropoxide (Gelest, Inc.), 2,4-pentanedior'1e (99%,
Aldrich Chemical Company), isopropanol (Fischer), toluene (Fischer) and valeric acid
(99+%, Aldrich Chemical Company) were used without further purification. Distilled,
deionized water was used for all synthesis.

Film Preparation. Film preparation is based on procedures described by
Gagliardi et al. [12] for the production of transition metal carboxylate films. Aluminum
isopropoxide is dissolved in a isopropanol/toluene (20% isopropanol v/v) mixture which
became clear after heating to approximately 80°C. Through various experimental trials,
the 20% isopropanol/toluene mixture was determined to be an execellent solvent for
aluminum isopropoxide. An appropriate amount of valeric acid or 2,4-pentanedione
(acac) was then added to the aluminum isopropoxide solution to obtain chelator to metal
ratios of 0.5, 1, 2, 3, 6, 9, and 12 for valeric acid (denoted as RA) and 0.5, 1, 1.5, 2, and 3
for acac (denOIed as RL). An unmodified allunina film was prepared without the presence
of a chelating ligand. Water was added to this mixture using water to metal ratio of 1.5:1.
Gels began to form after the addition of water with the onset of gelation being delayed by
an increase in the ligand to metal ratio. Films were prepared by spin coating the final
solution onto a clean quartz slide or 13 mm KBr disk. Film quality on quartz was
determined from visual inspection of the film’s color, transparency, and cracking. A high

quality film was transparent with few visible cracks.

35

Fourier Transform Infrared Spectroscopy (FTIR). FTIR spectra of films cast
on KBr crystals were obtained using a Mattson Instruments Galaxy 3020 Fourier
transform infrared spectrometer. The mid-IR region (4000-400 cm!) was examined with
a resolution of 2 cm". Data acquisition and processing were performed using an
Enhanced First software package.

XPS Analysis. X-ray photoelectron spectra were obtained with a Perkin-Elmer
surface science instrument equipped with a Model 10-360 precision energy analyzer and
an omnifocus small spot lens. All spectra were collected using a Mg anode (1253.6 eV)
operated at a power of 300W (15 kV and 20mA) with an analyzer pass energy of 50 eV.
Three regions were scanned in each film: C Is, 0 Is and Al 2p. Binding energies were
referenced to adventitious carbon (C Is = 284.6 eV) and were measured with a precision
of i 0.15 eV. Sensitivity factors were empirically derived from the analysis of Al(acac)3
(Aldrich Chemical Co.) and A1203 (sapphire, Harrick Scientific Corp.). XPS peaks were
fit with 20% Lorentzian-Gaussian mix Voigt profile functions using a nonlinear least-
squares curve-fitting program [13]. Each reported result is based on the analysis of at
least three films. It must be noted that the use of XPS to evaluate film composition
assumes that the films are homogenous.

2.4 Results and Discussion
2.4.1 Unmodified Alumina Films cast from unmodified solutions (without the addition

of chelating agents) were not transparent and were extensively cracked. Infrared results

(Figure 2.1) showed the presence of strong phonon peaks at 850 and 590 cm'l suggesting

 

36

 

 

n l . l . l r 1 . L
4000 3500 3000 2500 2000 ' 1500

F reauencv (cm-1)

Figure 2.1 FTIR spectrum of an unmodified alumina film.

 

37

an extensive aluminum oxide backbone. The 850 cm'1 absorbance has been described as

 

 

 

“u.

more difficult to assign, Tarte has ascribed absorbances in this region to vibrations of the
A106 octahedra [14] while Schroeder and Lyons have attributed absorbances in this
region to the tetrahedral configuration [15]. SIEEEPIL‘IBQCIERWLEEEES is also present
which may lead to rigid adhesion to the substrate as well as strong contact between
alumina chains. It is likely that film cracking is due a highly crosslinked aluminum oxide
backbone creating a very rigid material.

2.4.2 Acetylacetone modified Alumina The infrared spectra of the acac modified films
(with RL = 0.5 ,1, 1.5, 2, and 3) are presented in Figure 2.2. Films modified with acac
show C=o (at 1598 and 1455 cm"), C=C (1529 cm"), and Al-O-C (485 cm")
absorbances which are consistent in peak position and relative intensity with reported
values for Al(acac)3 [16]. The similarity of acac peak positions for all ligand ratios
suggests acac binding is independent of RL.

At low acac loadings, an intense hydroxyl peak, which has a maximum intensity
at 3450 cm", is assigned to hydroxyl ligands in a hydrogen-bonding environment. The
absence of the characteristic O-H bend for water at 1620 cm'1 suggests the films are not
extensively hydrated and 7 intramolecular hydrogen bonds are present between hydroxyl
groups or hydroxyl-grOups linked to backbone or ligand oxygens. As the acac loadings
increase, the hydroxyl absorbance shifts to higher energy and decreases in intensity. The

shift to a higher wavenumber is consistent with a change to a non-hydrogen bonding

environment [17]. As the ligand ratio increases, hydroxyl groups are replaced by acac

 

'PM '1" l '

38

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8, A.; l
izwm I“
c l
N J

 

 

 

fi/\v~¥ J
W

I T I T r j I l I l I l I

. 1
4000 3500 3000 2500 2000 1500 1000 500

 

 

 

Frequency (cm-1)

Figure 2.2 FTIR spectra of acac modified alumina films with ligand to metal ratios of
0.5 (a), 1.0 (b), 1.5 (c), 2.0 (d), and 3.0 (e).

4me ‘-.—“ ‘a-“fi

39

ligands, and the hydroxyl concentration of the film decreases as well as the number of
hydrogen bonding sites. As the acac concentration in the precursor solution increases the
backbone Al-O absorbances in the F TIR spectrum decreases in intensity. The peak
positions of the two phonon-modes do not change indicating the Al-O bonding does not
change. Thus the decrease in the backbone intensity is assigned to a decrease in the
extent of polymerization of the film. Nass and Schmit [8] observed a similar decrease in
alumina particle size with increasing acac ratio.

The XPS C Is spectra for the acac modified films consists of three peaks: a peak
at 284.6 eV due to alkyl ligand and adventitious carbon, a peak at 286.7 eV assigned to
carbon bound to one oxygen, and a carboxylate carbon at 288.7 eV. The carboxylate
feature is due to carbonate contamination of the film. The 286.7 eV peak is assigned to
either alkoxide or acac ligands, however the absence of alkoxide absorbances in FTIR
spectra for the aluminum acac films suggests that the this carbon is primarily due to the
diketonate. The extent of ligation of aluminum in the films was determined by
calculating the atomic ratio (from the relative areas of the 286.7 eV C Is and Al 2p peaks)
and applying the appropriate sensitivity factors. The extent of ligation is plotted against
the ligand:aluminum ratio in the film solution in Figure 2.3 and was found to increase to
0.8 at RL=2. The extent of ligation at RL=3 (not shown) is very low due to sublimation of
the fihn under the ultra-high vacuum conditions.

The A1 2p binding energy for the RL=O.5 film is approximately 74.1 eV. This is

within range of the reported values for aluminum oxide (74.0-74.7 eV) [18-21]. The peak

40

 

 

 

 

 

 

 

1.00 __
.. O
a 0.75
.2
*5
.29 ._l_
v-1 __
Q-l .
O
E
O.)
3?
LLl
0.50 -
0.25 . . . .
0 2

AcAc :Aluminum Ratio

Figure 2.3 Extent of acac ligation versus ligand:aluminum alkoxide solution ratio.

.1. its arr-l1

 

41

position does not change with increasing ligand ratio. Since the extent of ligation
increases with ligand ratio, it is somewhat surprising that the Al 2p peak position does not
shift to the value reported for Al(acac)3 (73.1 eV) [18]. There are two possible
explanations for this. The first is that since the extent of ligation of the aluminum centers
is rather low (0.8 at RL=2.0), the environment around the aluminum atoms is not changed
enough to cause a significant chemical shifi. It may also be, that if there is a distribution
in the extent of ligation, the high end of that distribution may be too volatile to withstand
UHV conditions. Thus, only the higher oxide content portion of the film will be
analyzed.

Films cast from the acetylacetone modified alkoxide precursors onto quartz slides
were crack free from 15R152. At low ligand ratios, large Al-O absorbencies are observed
suggesting a large extent of polymerization consistent with highly crosslinked aluminum
oxide polymers. In addition, the large hydroxyl absorbance indicates that the polymers
may be connected via a network of hydrogen bonds. Thus, we anticipate these films to
be very rigid and susceptible to cracking under drying conditions. As RL increases, the
acac ligands limit the extent of hydrogen bonding and the extent of polymerization.
Therefore the films consist of smaller and less crosslinked polymer chains. The films
should become more flexible and able to cope with the large Capillary force observed
during drying. The upper limit of the crack-free boundary is most likely due to the
formation of Al(acac)3-like particles. The RL=3 films exhibit very poor adhesion to

quartz and sublime at low pressures.

42

2.4.3 Valerie Acid Modified Alumina Valerie acid IR absorbances (Figure 2.4) at
2960, 2930, and 2860 cm'1 appear for all acid ratios. These bands have similar frequency,
shape and relative intensity to C-H modes of valeric acid and indicate the presence of
valerate ligands. Less intense valerate absorbances are observed at 1315, 1243, and 1109
cm]. Unreacted valeric acid is identified by the dimeric valeric acid carbonyl stretch at
1711 cm". Valerie acid may be present in the fihns due to insufficient drying or it may
be trapped in the films during drying.

For the lowest acid ratio film (RA=0.5, Figure 2.4a ) the asymmetric and
symmetric carboxylate stretches of the valerate ligand are observed at 1572 and 1464 cm'1
[22]. The coordination mode of the valerate ligands was determined by comparing the
asymmetric and symmetric frequency splitting with sodium valerate (Av=141 cm") [23].
The observed frequency splitting (Av=108 cm") is smaller than ionic valerate, and the

ligand coordination is assigned to be bidentate [22]. Shoulders on the low wavenumber
side of the symmetrical stretch and the asymmetry of the asymmetrical stretch indicate
that there may be a distribution of carboxylate ligands. The assignment of each
carboxylate absorbance is impossible as the different asymmetric stretches cannot be
resolved. However, it is likely that the low acid films consist of both monodentate and
bidentate valerate ligands. A

As the acid ratio increases, the low wavenumber shoulders of both carboxylate
stretes decrease in absorbance until a single set of peaks is observed (RA=6.0, Figure
2.4e). The absorbancies of the asymmetric and symmetric stretches occur at 1589 and

1469 cm", respectively. The frequency splitting of these peaks increases to 120 cm'1

43

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I l ' F ' 1 ' I T l ' l ' I
4000 3500 3000 2500 2000 1500 1000 500

Frequency (cm-1)

Figure 2.4 FTIR spectra of valeric acid modified alumina fihns with acid to metal ratios
of 0.5 (a), 1.0 (b), 2.0 (c), 3.0 (d), and 6.0 (e).

 

 

44

which is still indicative of bindentate valerate coordination. From the F TIR data it is
difficult to assess if the ligand coordination is bidentate bridging or chelating. Proposed
structures for hydrolyzed aluminum acetate sols are aluminum oxide chains or rings
consisting of an aluminum oxide backbone with one bidentate chelating acetate ligand per
aluminum center [22]. It is likely that the valerate ligands coordinate similarly to the
acetate ligand and we would expect the carboxylate binding to be bidentate chelating.

A single, broad hydroxyl absorbance is observed at 3450 cm'1 for the RA=0.5 film
(Figure 2.4a). A high frequency shoulder of the asymmetric carboxylate at 1620 cm'1 is
assigned to an O-H bend of water and indicates that the film is partially hydrated. It is
likely that the hydroxyl absorbance is a combination of O-H stretches in water and
aluminum hydroxyls hydrogen bonded to the water. As the acid ratio increases, the broad
hydroxyl peak decreases in intensity and narrower peaks at 3690 and 990 cm’1 appear.
These absorbances have been reported as bridging hydroxyls in a non-hydrogen bonding
environment [24]. As the acid ratio increases, the peak at 1620 cm'l disappears due to an
increase in the hydrophobic character of the film with increasing valerate concentration.

Unlike the acac modified films, the Al-O backbone absorbances of the acid
modified films do not resemble the unmodified film. Instead, new absorbances appear at
660 and 500 cm]. Though a change in Al coordination is obvious from these results, the
assignment for these new absorbances is difficult. We believe that these are possibly due
to a shift toward octahedral binding of the aluminum centers [14].

The XPS of the C Is feature revealed peaks at 284.6 eV from alkyl carbon and
288.2 eV from carboxylate carbon. As before, the alkyl carbon is a combination of

adventitious and ligand species, while the carboxylate carbon is a combination of ligand

45

species and carbonate contamination of the surface. No carbon features were observed at
286 eV suggesting that alkoxide ligands are not present at the surface of the films. The
extent of valerate ligation was determined by dividing the carboxylate C Is (288.2 eV)
peak by the A1 2p peak. The extent of valerate ligation is plotted against the acid ratio on
Figure 2.5. For an unmodified film the extent of ligation is 0.3 which is assigned to
carbonate contamination. The extent of ligation increases from approximately 0.9 for
Ra=1 to 1.7 for R.=3 after which it plateaus for higher acid loadings and approaches 2.

Valerate modified alkoxides produced crack-free films when R133. XPS results of
valerate modified films have shown that the extent of ligation increases until at RA=3 it
begins to plateau. As the extent of ligation increases, crosslinking of aluminum oxide
particles should decrease and the rigidity of the film should decrease as well. This is
verified by the FTIR results where an immediate loss of backbone Al-O features was
observed upon addition of valeric acid. RA=3 also coincides with the shift to a non-
hydrogen bonded hydroxyl group. These isolated hydroxyl groups may be “protected”
enough by the valerate ligands to prevent linkage with other aluminum oxide polymers

thereby creating a more flexible film.

46

 

2.2 -

2.0 - I”
b

1.8 a .5-

 

 

1.6-
1.4-
1.24

1.0- E

0.8 -

Extent of Ligation

0.6 -

0.4 -

0.2 -

 

 

 

Valerie Acid : Aluminum Ratio

Figure 2.5 Extent of valerate ligation versus valeric acidzaltuninum alkoxide solution
ratio

2.4

10.

ll.

12.

l3.

14.

15./,-

’16.

47

References

Livage, J.; Henry, M.; Sanchez, C. Prog. Solid St. Chem. 1988, 18, 259.
Yoldas, B. E. J. Appl. Chem. Biotechnol. 1973, 23, 803.

Yoldas B. E., Amer. Ceram. Soc. Bull. 1975, 54, 286.

a 1 kr‘hfl

Yoldas B. E., Amer. Ceram. Soc. Bull. 1975, 54, 289.

Brinker, C. J.; Hurd, A. J .; Sehunk, P. R.; Frye, G. C.; Ashley, C. S. J. Non-
Crystalline Solids, 1992, 147&148, 424.

Tayaa, H.; Mosset, A.; Galy, J. Eur. J. Solid State Inorg. Chem,. 1992, 29, 13.

frag-«mathn’. J.;-4'14
.J- 1 ~_

Babonneau, F.; Coury, L., Livage, J. J. Non-Crystalline Solids, 1990, 121, 153.
Nass, R.; Schmidt, H. J. Non-Crystalline Solids, 1990, 121 , 329.

Ayral, A.; Phalippou, J.; Droguet, J. C. In Better Ceramics Through Chemistry 111;
Brinker, C. J .; Clark, D. E.; Ulrich, D. R.; MRS Sympositun Series 121; Materials
Research Society; Pittsburgh; 1988; p 239.

Rezgui, S.; Gates, B. C Chem. Mater. 1994, 6, 339.

Rezgui, 8.; Gates, B. C.; Burkett, S. L.; Davis, M. E. Chem. Mater. 1994, 6, 2390
Gagliardi, C. D.; Dunuwila, D.l; Berglund. K. A. In Better Ceramics Through
Chemistry IV; Zelinsky, B. J. J ., Brinker C. J ., Clark, D. E.l Ulrich, D. R., Eds;
Mater. Res. Soc. Proc. 180; MRS: Pittsburgh, PA, 1990; p 801.

Software prodied by Dr. Andrew Proctor, University of Pittsburgh, Pittsburgh,
PA.

Tarte, P. Spectrochim. Acta, 1967, 23A, 2127.
Schroeder, R. A.; Lyons, L. L. J. Inorg. Nucl. Chem. 1966, 28, 1155.

Nakarnoto, K.; McCarthy, P. J .; Armand, R.; Martell, A. E. J. Am. Chem. Soc,
1961, 81, 1066.

l7.

18.

19.

20.

48

Colthup, N. B.; Daly, L. H.; Wiberley, S. E. Introduction to Infrared and Raman
Spectroscopy, 3rd ed.; Academic Press: San Diego, 1990.

McGuire, G. E.; Schweitzer, G. K.; Carlson, T. A. Inorg. Chem, 1973, 12, 2450.

Handbook of X-ray Photoelectron Spectroscopx Wagner, C. D.; Riggs, W. M.;
Davis, L.E.; Moulder, J. F.; Muilenburg, G. E., Eds.; Perkin Elmer, 1979. "

Nefedov, V. 1.; Gati, D.; Dzhurinskii, B. F.; Sergushin, N. P.; Salyn, Ya. V. Zh.
Nerog. Khim. 1975, 20, 2307.

Nefedov, V. 1.; Salyn, Ya. V.; Leonhardt, G.; Seheibe, R. J. Electron Spectros,
1977, 10,121.

Mehrotra, R. C.; Bohra, R. Metal Carboxylates, Academic Press: London, 1983.
Severin, K. G.; Ledford, J. S. Chem. Mater., 1994, 6, 890.

Chukin, G. D.; Seleznev, Yu., L. Kinet. Catal. 1989, 30, 55.

 

:1.-
I

lgurru "'1‘. 1271131.“. no a- a.

Chapter 3

Sol-Gel Synthesis of Ceria-Alumina Catalysts

3.1 Abstract

Cerium and aluminum mixed oxide powders were prepared from cerium and
aluminum valerate precursors by the sol-gel method. The Ce/Al atomic ratio was varied
from 0 to 1.0 and the powders were calcined at 500 °C (designated 500Cey) and 900 °C
(designated 900Cey), where y is the Ce/Al atomic ratio. The presence of bridging cerium-
aluminum hydroxyl groups for the uncalcined catalysts as determined by Fourier
transform infrared (FTIR) spectroscopy suggests that some mixing of the two metals
occur during preparation. The calcined catalysts were characterized by X-ray
photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and BET surface area
measurements. For the 500Cey catalysts, XPS measurements indicate that cerium is
uniformly dispersed at low loading ( y 5 0.33). XRD results are consistent with this as
crystalline Ce02 is evident only in the powder patterns of catalysts with higher cerium
loadings. The surface area of all cerium containing 500Cey catalysts is greater than that
of the alumina support. It is increased by 50% for the 500Ce0.l catalyst and nearly
doubled for all catalysts with y20.25. In contrast, the 900Cey catalysts contain Ce02
crystallites for all ceritun loadings and the surface area of these catalysts decreases with
increasing cerium concentration. Additionally, cerium was found to inhibit the transition

to Ot-Al203 which occurs for the unpromoted A1203 support.

49

50
3.2 Introduction

Rare earth oxide additives have been widely utilized as structural promoters for
heterogeneous catalysts [1-9]. Rare earth promotion increases the resistance of y-alumina
to thermal loss of surface area and improves the dispersion and thermal stability of
precious metal catalysts [2-9]. Addition of cerium oxide to noble metal-based automobile
exhaust gas catalysts promotes the water-gas shift reaction [10,11], suppresses the CO
inhibition effect during CO oxidation [12-14], and removes CO and NOX simultaneously
[4-7,15-17]. Cerium has also been used as a promoter in base metal oxide catalysts [18-
21]. Bedford and LaBarge [19] have used copper oxide impregnated, high surface area .
ceria and alumina catalysts to remove CO, NO, and hydrocarbons from automotive
exhaust gas.

Nunan et al. [22] observed that Pt binding with Ce02 increases with decreasing

Ce02 particle size for Ce02/y-AI2O3 supports. Therefore, it is desirable to produce a
highly dispersed Ce02 phase or small Ce02 crystallites in the preparation of Ce02/y-
Al2O3 catalysts. Most published studies of Ce02/y-Al2O3 supports have examined 7-

Al2O3 supports impregnated with cerium (III) nitrate solutions. It is generally reported
that impregnated catalysts with low cerium loadings (< 5 wt%) consist of small Ce02
crystallites and Ce+4 ions dispersed in a CeAlO3-like phase [23,24]. Catalysts with high
cerium loadings (> 5 wt% Ce02 per 100 mz/g y-Al2O3) are known to contain large Ce02
particles [23,24].

Sol-gel chemistry is a synthetic technique which offers control of both the catalyst

structure and catalytic properties through modifications of the precursor chemistry.

51

Haneda et al. [25] used sol-gel techniques to prepare alumina supported cerium catalysts
by adding cerium nitrate dissolved in ethylene glycol to bohemite sols. These catalysts
exhibited a greater surface area, dispersion, and lower crystallite size compared to
analogous impregnated catalysts. Narula [26] has prepared aluminum supported cerium
catalysts from heterobirnetallic ceriurn/alumimun precursors. These catalysts had a high

surface area (~220 m2/ g) and were amorphous up to 700 °C. The limitation to using this

method is that the CezAl ratios are constrained to the stoichiometry of the bimetallic
precursors.
Previous studies of Ce02/A1203 catalysts have focused primarily on the properties

of impregnated y-alumina materials. Little effort has been devoted to systematically

investigating the effect of cerium content on the structure of Ce02/A1203 catalysts
prepared using sol-gel chemistry. The present work is part of a program to investigate the
effects of rare earth oxide promoters on the structure and activity of transition metal oxide
based emission control catalysts.

In this study, Ce02/A1203 materials were prepared using a metal carboxylate sol—
gel approach [27-32]. The object of these experiments is to provide a homogenous CezAl
catalyst with high surface area and small Ce02 crystallites. We believe that a
homogenous starting mixture of the two components is a gOod path for achieving these
goals. Sol-gel chemistry is an ideal preparative route for this as it not only provides for
good mixing but potentially allows for the formation of mixed metal oxide bonds to
reduce segregation at elevated calcination temperatures. One anticipated difficulty with
the sol-gel route is the rapid rate of hydrolysis of the two alkoxides which could lead to a

heterogeneous starting material. Complexation of metal alkoxides with carboxylic acids

52

has been shown to limit the rate of hydrolysis by replacing alkoxide ligands with more
difficult to remove carboxylate ligands [33,34]. In our approach, valeric acid is added to
the mixture of metal alkoxides in an attempt to slow hydrolysis and promote better
mixing of the two metals prior to precipitation. The resulting gels are then examined
with Fourier transform infrared spectroscopy (FTIR) to determine if mixing occurs
between the ceria and alumina precursors. The calcined powders are examined with X-
ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and BET surface area
measurements to determine the influence of cerium content on the chemical state and

dispersion of the alumina supported cerium oxide.

3.3 Experimental

Materials. Cerium(IV) methoxyethoxide (18% w/v in methoxyethanol, Gelest),
aluminum isopropoxide (Aldrich Chemical Company), isopropanol (Fisher) and valeric
acid (99+%, Aldrich Chemical Company) were used without further purification.
Distilled, deionized water was used for all synthesis.

Catalyst Preparation. Cerium methoxyethoxide was added to aluminum
isopropoxide to produce cerium to aluminum atomic ratios of 0:1, 1:10, 1:4, 1:3, 1:2, and
1:1. Valerie acid was added to these mixtures to obtain acid to metal ratios of 9. Dilution
with isopropanol (alcohol-to-metal ratio = 10) and heating to 50 °C with vigorous mixing
produced clear yellow solutions. Water was added (water-to-metal ratio = 1.5) and the

solutions were allowed to dry in air for 24 h followed by heating at 50 °C until the

majority of solvent and unreacted acid were removed. The dried product was a

translucent yellow-orange powder. The catalysts were then calcined at 500 or 900 °C in

53

air for 16 h to produce light yellow powders at low CezAl ratios and darker yellow
powders at higher Ce:Al ratios. Catalysts calcined at 500 °C are designated “500Cey,”
where y is the Ce/Al atomic ratio. “900Cey” designates catalysts calcined at 900 °C.

Fourier Transform Infrared Spectroscopy (FTIR). FTIR spectra were
obtained using a Mattson Instruments Galaxy 3020 Fourier transform infrared
spectrometer. Spectra were obtained in the mid-IR region (4000-400 cm") with a
resolution of 2 cm". Data acquisition and processing were performed using Mattson
Enhanced First software package. Samples were prepared by spin-casting cerium-
aluminum solutions onto 13 mm KBr discs.

BET Surface Area. Surface area measurements were performed using a
QuantaChrome Quantasorb Jr. Sorption System. Approximately 0.1 g of the catalyst was
outgassed in a N2/He mixture (5% N2) at 350 °C for 1 h prior to adsorption
measurements. The measurements were made using relative pressures of N2 to He of
0.05, 0.08, and 0.15 (N2 surface area = 0.162 nmz) at 77 K.

X-ray Diffraction. X-ray powder diffraction patterns were obtained with a
Rigaku XRD diffractometer employing Cu K01 radiation (A = 1.5418 A). The X-ray was
operated at 45 kV and 100 mA. Diffraction patterns were obtained using a scan rate of
0.25 deg/min with divergence and scatter slit widths of 1/6°. Catalyst powders were
mounted in the cavity of a circular silicon sample holder (C-12, Dow Coming). The
mean crystalline sizes ((1) of the Ce02 particles were calculated from XRD line

broadening measurements using the Scherrer equation [3 5]:

d=IOc/[3c050 (3.1)

54

where A is the X-ray wavelength, K is the particle shape factor, taken as 0.9, and B is the

full width at half maximum (F WHM), in radians, of the Ce02 <111> line.

XPS Analysis. XPS data were obtained using a Perkin-Elmer Surface Science
instrument equipped with a magnesium anode (1253.6 eV) operated at 300 W (15 kV, 20
mA) and a 10-360 hemispherical analyzer operated with a pass energy of 50 eV. Typical
operating pressures in the analysis chamber were 1x10’8 Torr. Catalyst samples were
analyzed as powders dusted onto double-sided sticky tape. Binding energies for the
catalyst samples were referenced to the Al 2p peak (74.5 eV). XPS binding energies were

measured with a precision of i 0.15 eV, or better.

3.4 Results and Discussion
3.4.1 Unealcined Materials

The F TIR spectrum of the aluminum valerate material (prior to calcination) is
shown on Figure 3.1a. It is assumed that the films cast from the precursor solutions are
representative of the uncalcined powders. The three bands observed at 2960, 2933 and
2875 cm'1 are assigned to the methyl and methylene C-H stretches of the valerate ligands.
Peaks indicative of aluminum triisopropoxide are absent frOm the spectrum suggesting
that alkoxide ligands were completely removed following complexation with valeric acid

and hydrolysis. Similar results are observed for all materials studied in this work.

 

55

 

 

 

 

 

 

 

 

      

 

 

 

d

c

b

a
'T‘l'l'l'lTl

4000 3500 3000 2500 2000 l 500 1000 500

Frequency (cm'l)

Figure 3.1. FTIR spectra (4000-400 cm") of as prepared films with CezAl ratios of a) 0:1, b)
1:10, c)1:1, and d)1:0.

56

The asymmetric and symmetric carboxylate stretches of the valerate ligands
appear at 1590 and 1470 cm'l respectively. Carboxylate ligands can bind to metals via
monodentate or bidentate coordination (either chelating or bridging) [36]. The
coordination mode of the valerate ligands is determined by comparing the asymmetric

and symmetric frequency splitting with that of sodium valerate (Av=l4l cm") [27]. The
observed frequency splitting (Av=122 cm") is smaller than ionic valerate, and the ligand

coordination is assigned as bidentate [36]. From the FTIR data, it is uncertain whether
the carboxylate ligand coordination is bidentate chelating or bridging.

An expanded view of the hydroxyl region (3200-3800 cm") is shown on Figure
3.2. There is a single sharp peak at 3692 cm'l (Figure 3.2a) characteristic of a hydroxyl
group in a non-hydrogen bonding environment. This peak and the sharp peak at 990 cm'l
(Figure 3.1a) are characteristic of a bridging Al-O-Al hydroxyl group [37]. The low
frequency region of the FTIR spectrum reveals three peaks at 670, 580, and 500 cm'1
which are assigned to Al-O vibrations. The aluminum valerate film is amorphous, which
makes the exact assignment of the Al-O peaks difflcult. However, the peaks are within
the correct wavelength range reported for A106 octahedra [38]. This agrees with 27Al
NMR results for acetic acid modified alumina gels that reveal an octahedral coordination
about the Al centers [39].

The FTIR spectrum of a cerium valerate film is shown on Figure 3.1d. The
asymmetric and symmetric carboxylate frequencies for this material appear at 1543 and

1418 cm'1 respectively. Again, the observed frequency splitting (Av=125 cm") is smaller

 

 

3640 cm’l

 

c
3670 cm"

 

3692 cm'l

 

 

J

l ' l I I
3750 3500 3250

 

 

 

Frequency (cm '1)

Figure 3.2. FTIR spectra of the hydroxyl region (3800-3100 cm!) of
as prepared films with CezAl ratios of a) 0:1, b) 1:10, c) 1:1,
and (1) 1:0.

 

58

than ionic valerate and the carboxylate ligands are assigned to a bidentate coordination to
cerium. There are two peaks in the hydroxyl region (Figure3.2d) at 3400 cm‘1 and 3640
cm". The peak at 3400 cm'1 is assigned to a hydroxyl group in a hydrogen-bonding
environment while the peak at 3640 cm'I is assigned to a bridging Ce-O-Ce hydroxyl
absorbance [40].

As the cerium loading increases from the Ce0.1 material (Figure 3.1b) the Ce
carboxylate absorbances increase in intensity. The relative intensities of the aluminum
and cerium carboxylate absorbances are approximately equal in the Ce1.0 film spectrum
(Figure 3.1c). Neither aluminum nor cerium carboxylate frequencies change with cerium
content, indicating that cerium loading has little effect on valerate coordination. For
Ce0.1 (Figure 3.2b), an absorbance at 3670 cm'1 appears which is assigned to a bridging
Ce-O-Al hydroxyl ligand. The aluminum hydroxide absorbance at 990 cm'1 splits to two
peaks at 995 and 960 cm", which are assigned to the bridging aluminum hydroxyls and
bridging aluminum-cerium hydroxyl absorbances, respectively. For the Cel.0 material
(Figure 3.2c), the bridging Ce-O-Ce absorbance at 3640 cm'1 is observed. Additionally a
new absorbance at 680 cm'1 is observed for all Ce:Al films. This absorbance is not
observed in the aluminum and cerium films and is assigned to aluminum-cerium metal

oxide bond

3.4.2 500Cey Catalysts.
The BET surface area of each catalyst as a function of the Ce/Al atomic ratio is
given in Table 3.1. The surface area of the alumina support is 74 mz/g. At 500 °C the

surface areas of the catalysts are increased by 50% (with respect to the alumina support)

59

after the first addition of cerium and nearly doubled for all further additions of cerium.
The increase in surface area is a direct result of the presence of cerium. The enhancement
observed with our catalysts is attributed to cerium atoms inhibiting the sintering of
aluminum oxide by occupying potential aluminum sites. The surface areas of the
catalysts are lower than comparable catalysts prepared by Narula [26] and Haneda et al.
[25], and this is likely to be a result of different drying techniques. The 500Ce0.2 catalyst
was remade using vacuum drying to give a surface area of ~180 mZ/g, which is
comparable to the surface area of a similar catalyst prepared by Narula [26].

The XRD pattern of the 500Ce0 material (Figure 3.3a) reveals the <220> line for
y—Al2O3. The breadth and low intensity of this feature suggests that the alumina is largely
amorphous with small crystallites of y-A12Og, The effect of cerium loading on the cerium
crystallinity of the catalysts is illustrated in Figures 3.3b-d. With loadings of y<0.5, Ce02
patterns are not observed which means that the cerium is amorphous. However, as the
cerium loading increases the <111>, <200>, <220>, and <311> lines of Ce02 become
evident as seen in Figures 3.3c and 3.3d. The calculated average particle size of Ce02 is
18 and 36 A, respectively. After calcination at 500 °C the <220> line of y-Al2O3 is
observed for all of the catalysts suggesting that though cerium may have an effect on the
surface area, there is no evidence that Ce changes the crystal structure of A1203.
Diffraction patterns for CeAlO3 are not observed for any of the catalysts examined in this

study.

60

 

 

 

C602
<1 1 1>
C602
<200> Ce02
<31 1>
d
C
b
'Y-A1203 <220>
a

 

 

Degrees (20)

Figure 3.3. XRD patterns of catalysts calcined to 500 0C for CezAl ratios of a) 0:1, b) 1:3, c)
1:2, and d)1:1.

61

Variation of the Ce 3d:Al 2p intensity ratio as a function of the CezAl atomic ratio
for the 500Cey catalysts is shown in Figure 3.4. The CezAl intensity ratios measured for
the catalysts increase linearly with CezAl atomic ratio except at the highest cerium
loading. Using the Kerkhoff-Moulljn model [41], the linear region of this curve is
assigned to a dispersed cerium oxide phase on the surface of the catalyst. The negative
deviation fi'om linearity at high loading is assigned to the formation of Ce02 crystallites at
the catalyst surface. These findings are consistent with the XRD results for the 500Cey

catalysts which show a Ce02 crystalline phase only at high cerium content.

Table 3.1. Variation of BET surface area as a function of Ce loading and calcination

 

 

temperature
CezAl Atomic Ratio Surface Area 500 °C (mZ/g) Surface Area 900 °C (mz/g)
0:1 74 55
1 :10 l 10 55
1:4 132 ‘ 40
1 :3 130 34
1 :2 133 30
l :1 123 15

 

 

 

 

62

 

40 r

354

t»
O
J

N
U]
1

J

L

.l
15 I;

Ce(3d):Al(2p) XPS Intensity Ratio
8
l
I

p—a
O
l

 

 

 

I I I I f

l
0.2 0.4 0.6 0.8 1.0
Ce:A1 Ratio

Figure 3.4. Variation of the Ce 3d/Al 2p intensity ratio as a function of the Ce:A1
atomic ratio for the 500Cey (I) and 900 Cey (O)

63
3.4.3 900Cey Catalysts.

The XRD pattern (Figure 3.5a) of the alumina support after calcination at 900 °C
shows the presence of Ot-Al203. This is consistent with a transition from the y- to Ot-
Al203 phase as described for alumina xerogels [42]. Particle size calculations show that
the alpha phase exists as large crystallites with a particle size of about 28.0 run, while the
gamma phase has a particle size of approximately 5.0 nm. The surface area decreases
approximately 25% to 55 m2/g (Table 3.1) when compared to the 500Ce0 material. This
decrease is typical for the gamma to alpha alumina phase change.

After calcination at 900 °C the <111>, <200>, <220>, <311>, and <222> lines for
Ce02 are present for all Ce containing catalysts (Figure 3.5b-f). The variation of Ce02
particle size with cerium loading is shown on Figure 3.6 (denoted as triangles). The
average crystallite size increases from 10 to 17 nm as the Ce:A1 ratio increases from 0.1
to 1. The 900Cey surface areas decrease (Table 3.1 and Figure 3.6 denoted as circles)
from 55 mz/g to 10 mz/g over the same range of atomic ratios. This decrease is attributed
to the increasing Ce02 particle size.

For the 900Ce0 material, peaks corresponding to the 01- and y-Al203 phase are
observed (Figure 3.5). However, the alumina in the cerium promoted catalysts is largely

amorphous. Only for the lowest cerium loadings 05.25) are peaks evident due to y-A1203

crystallites. The intensity of these peaks decreases with increased cerium loadings
(Figure 3.5b-d). At higher cerium loadings A1203 is apparently amorphous. The

mechanism employed by the cerium oxide to stabilize ‘y-Al203 and amorphous A1203 is

uncertain. Previous work by Schaper et al. [43] on the addition of La203 to 'y-Al203

 

 

 

 

 

 

 

 

CcOz
<11 l>
CcOz
<220> CcOz
d
c H
b
0
ct ,(
O
a .Y o

 

 

 

Degrees (2 0)

Figure 3.5. XRD patterns of catalysts calcined to 900 °C for Ce:A1 ratios of a) 0:1, b) 1:10,
c) 1:4, and d) 1:3. 01 and 7 refer to the or and 7 phases of alumina respectively.

“PM“???

 

65

indicated a 100 0C increase in the y- to Ot-Al203 transition. This increase was attributed to
a decrease in the nucleation rate caused by the formation of a LaA103 surface species.
However, for the case of cerium addition it is unlikely that Ce4+ would be reduced to form
CeAlO3 when calcined in air. Likewise XRD patterns for CeAlO3 are not observed. An
alternative explanation is the Ce-O-Al bonds observed in the uncalcined material hinder
the crystallization of A1203 particularly at higher cerium loadings. However, further work
is needed to elucidate this mechanism.

The Ce 3dzAl 2p intensity ratio measured for 900Cey (Figure 3.4) catalysts
approach an intensity ratio of approximately 10 with increasing cerium content. From the
Kerkhoff-Moulljn model this curve indicates that Ce02 forms crystallites on the surface
[41]. Thus, cerium oxide is no longer dispersed following calcination at 900 °C. This
result is supported by the XRD findings which show Ce02 crystallites for all cerium
catalysts calcined at 900 °C.

Based upon the results of the 500Cey catalysts, the sol-gel route is an effective
method for the preparation of Ce02/A1203 catalysts. The cerium oxide phase is dispersed
to very high cerium loadings (Ce:A1 atomic ratios of 1:2) and the surface areas of the

catalysts remain high (~130 mz/g) especially since the drying techniques used in this work

should lead to a significant loss in surface area. Afier calcination at 900 °C there is a
significant loss in both surface area and Ce dispersion as the ceria and alumina phases
begin to form crystallites. This process may be hindered by incorporation of textural

promoters such as La, Ba, or Sr.

66

 

 

 

 

I I I I I I I l I l
. O A .
50-
- 16
N .. .
a a
c, 40- a, 5,3
0 a -c
2- <—'~ __,. - 14 a
d) '. ....
8 ~ .‘~. .1 8
‘g -. A; $43.
V) '. . g
5 30- V. A
m 3
. - 12
20- ,A' .
.l "
A O _ 10
' I ' I ' l ' l ' l
0.2 0.4 0.6 0.8 1.0

Ce:A1 Ratio

Figure 3.6. Variation of the Ce02 crystalline size (0) and BET surface area (0) as a
function of the Ce:A1 atomic ratio for the 900 Cey catalysts.

3.5

10.

11.

12.

13.

14.

15.

16.

l7.

l8.

19.

67

References

Trovarelli, A. Catal. Rev. 1996, 39, 439.

Yao, Y. F. Yu; Kummer, J. T. J. Catal. 1987, 106, 307.

Su, E. C.; Rothschild, W. G. J. Catal. 1986, 99, 506.

Su, E. C.; Montreuil, C. N.; Rothschild, W. G. Appl. Catal. 1985, 17, 75.
Herz, R. K. Ind. Eng. Chem. Prod. Res. Dev. 1981, 20, 451.

Yao, H. C.; Yu-Yao, Y. F. J. Catal. 1984, 86, 254.

Gandhi, H. S.; Piken, A. G.; Shelef, M.; Delosh, R. G. SAE paper No. 760201,
1976.

Dictor, R.; Roberts, S. .1. Phys. Chem. 1989, 93, 5846.

Le Normand, F.; Hilaire, L.; Kill, K.; Krill, G.; Maire, G. J. Phys. Chem. 1988,
92, 2561.

Kim, G. Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 267.

Herz, R. K.; Sell, J. A. J. Catal. 1985, 94, 166.

0h, 8. H. J. Catal. 1990, 124, 477.

Yao, Y. F. YuJ. Catal. 1984, 87, 152.

Oh, S. H.; Eickel, C. C. J. Catal. 1988, 112, 543.

Herz, R. K. Am. Chem. Soc. Symp. Ser. 1982, 178, 59.

Jin, T.; Okuhara, T.; Mains, G. J .; White, J. M. J. Phys. Chem. 1987, 91, 3310.
Cho, B. K.; Shanks, B. H.; Bailey, J. E. J. Catal. 1989, 115, 486.

Agarwal, S. K.; Spivey, J. J.; Butt, J. B. Appl. Catal. 1992, 81, 239.

Bedford, R. E.; LaBarge, W. J. US. Patent 5,063,193 Nov. 5, 1991.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

68

Lu, G.; Wang, R. Cuihua Xuebao 1991, 12, 314.
Tian, Y.; Fu, Y.; Lin, P. Cuihua Xuebao 1994, 15, 189.
Nunan, J. G.; Robota, H. J.; Cohn, M. J.; Bradley, S. A. J. Catal. 1992, 133, 3310.

Shyu, J. 2.; Otto, K.; Watkins, W. L. H.; Graham, G. W.; Belitz, R. K.; Gandhi,
H. S.J. Catal. 1988, 114, 23.

Shyu, J.Z.; Weber, W.H.; Gandhi, H.S. J. Phys. Chem, 1988, 92, 4964.

Haneda, M.; Mizushima, T.; Kakuta, N.; Ueno, A.; Sato, Y.; Matsuura, S.;
Kashara, K.; Sato, M. Bull. Chem. Soc. Jpn., 1993, 66, 1279.

Narula, C. In Better Ceramics Through Chemistry V, Hampden-Smith, M. J.;
Klempere, W. G.; Brinker, C. J., Eds; Mater. Res. Soc. Symp. Proc. 271; MRS:
Pittsburgh, PA, 1992.

Severin, K.G.; Torgerson, B.V.; Berglund, K.A.; Ledford, J .S. Chem. Mater.
1994, 6, 890.

Noonan, G. 0.; Ledford, J. S. Chem. Mater. 1995, 7, 1117.

Severin, K.G.; Ledford, J .S. Langmuir 1995, 11, 2156.

Dunawila, D.; Berglund, K.A. Chem. Mater. 1994, 6, 1556.

Gagliardi, C.D.; Dunuwila, D.; Berglund, K.A. in Better Ceramics Through
Chemistry IV, Zelinsky, B.J.J.; Brinker, C.J.; Clark, D.E.; Ulrich, D.R., Eds;
Mater. Res. Soc. Proc. 180; MRS: Pittsburgh, PA, 1990; p. 801.

Gagliardi, C.D.; Berglund, K.A. in Processing Science of Advanced Ceramics,
Aksay, K.A.; McVay, G.L.; Ulrich, D.R., Eds; Mater. Res.Soc. Proc. 155:
Pittsburgh, PA, 1986; 127.

Gagliardi, C. D.; Dunuwila, D.; Van Vliergerge-Torgenson, B. A.; Berglund, K.
A. in Better Ceramics Through Chemistry V; Hampden-Smith, M. J .; Brinker, C.
J .; Klemper, W. G., Eds; Mat. Res. Soc. Symp. Proc. 271: Pittsburgh, PA, 1992:
257.

Sanchez, 0; Livage, J. New J. Chem. 1990, I4, 513.

35.

36.

37.

38.

39.

40.

41.

42.

43.

69

Klug, H. P.; Alexander, L. E., X-ray Diflraction Procedures for Polycrystalline
and Amorphous Materials, lst.ed.; Wiley: New York, 1954.

Mehrotra, R. C.; Bohra, R. Metal Carboxylates; Academic Press, New York,
1983.

Chukin, G. D.; Seleznev, Yu. L. Kinet. Katal. 1989, 30, 55.
Tarte, P. Spectrochim. Acta 1967, 23A, 2127.

Rezgui, 8.; Gates, B. C.; Burkett, S. L.; Davis, M. E. Chem. Mater. 1994, 12,
2390.

Laachir, A.; Perrichon, V.; Badri, A.; Lamotte, J. Catherine, E.; Lavalley, J. C.; El
Fallah, J .; Hilaire, L.; 1e Normand, F .; Quemere, E.; Sauvion, G. N.; Touret, 0. J.
Chem. Soc. Faraday, Trans. 1991, 8 7, 1601.

Kerkhof, F. P. J. M.; Moulljn, J. A.; J. Phys. Chem. 1979, 83, 1612.

Mizushima, Y.; Hori, M. J. Non-Crystalline Solids, 1994, 167, 1.

Seahper, H.; Doesburg, E. B. M.; Van Reijen, L. L. Applied Catalysis, 1983, 7,
211.

Chapter 4

Synthesis and Characterization of Ce02/T 102 Thin Films

4.1 Abstract

Ceritun and titanium mixed oxide thin films were prepared from cerium and
titanium valerate precursors by the sol-gel method. The Ce/T i atomic ratio was varied
from 0 to 1.0 and the calcination temperature of the films was varied from 400 °C to 1000
0C. The films are noted as xCTy where x is the calcination temperature and y is the Ce/T i
atomic ratio. The films were characterized by X-ray photoelectron spectroscopy (XPS)
and powder X-ray diffraction (XRD). All of the films were amorphous following
calcination at 400 °C. Anatase T102 and Ce02 phases were observed with the 600CTy
films for low cerium loadings (1250.25). Unidentified phases, attributed to the
incorporation of Ce within the titanium oxide lattice, were observed for intermediate
cerium loadings (025350.67). The anatase to rutile TiO2 transition was observed after
calcination at 800 0C. The transition of unidentified phases to rutile TiO2 was observed at
1000 °C. Following calcination at 800 °C, cerium oxide was largely segregated to the
surface of the film. The enhancement of Ce02 at the surface of the film decreases with

increasing cerium loading.

70

71

4.2 Introduction

Sol-gel derived mixed cerium oxide and titanium oxide films have been utilized
as optical coatings [1-4], solar cells [5-6], and catalysts [7-10]. Coatings of Ce02/TiO2
produce yellow films with optical cutoffs between 380-400 nm [1]. These coatings

exhibited acid resistance when coated onto aluminum foils [2].

Ceria and ceria-titania films have been developed as counter electrodes for
electrochromic devices. Lavrencic Stangar, et al. [5] compared the lithium storage and
electrochromic properties of Ce02 and Ce02-TiO2 films. Solar transmittances of both
coatings were comparable with optical cutoffs occurring below 380 nm. Lithium
insertion and charge storage were determined with cyclic voltarnmetry to determine each
film’s utility as a counter electrode for electrochromic devices [5]. Ce02 coatings have a
higher electrochemical reversibility and a greater lithium storage capacity for films
thicker than 50 nm. However, Ce02-TiO2 films have a faster response time. Both films
exhibited electrochromism in the UV spectral range [5].

Ceria-titania mixed oxides are also useful in catalysis due to the ability of TiO2 to
stabilize the Ce3+ cation [7]. Ceria promoters on TiO2 supports favor precious metal
dispersion and are more active in the NO-CO reaction giving N2 and C02 at low
temperatures [11]. Ce02/TiO2 mixtures are better hydrocarbon reforming catalysts than
the individual oxides [9]. Ceria-titania mixtures have also been used for the
photodecomposition of H20 and photomethanation of C02 [10].

Though much work has been done examining the properties of sol-gel derived

Ce02/TiO2 films, relatively little work examining their structure. This chapter is an

72

attempt to remedy this by examining the effect of cerium loading and calcination
temperature on the bulk and surface of Ce02/TiO2 films by powder X-ray diffraction

(XRD) and X-ray photoelectron spectroscopy (XPS), respectively.

4.3 Experimental

Materials. Cerium(IV) methoxyethoxide (18% w/v in methoxyethanol, Gelest
Inc.), titanium(IV) isopropoxide (Aldrich Chemical Company), and valeric acid (99+%,
Aldrich Chemical Company) were used without further purification. Distilled, deionized
water was used for all synthesis.

Film Preparation. The methods used to synthesize films were based on
techniques developed previously for tin and titanium oxide thin films [12,13]. Films
were prepared with cerium to titanium atomic ratios of 1 :10, 1:5, 1:3, 1:2 and 1:1. Atomic
ratios of valeric acidzwaterzmetal were 9:1 5:] respectively. Reactions were carried out at
'room temperature in capped vials in a N2 purged glove box. Valerie acid was added to
the alkoxide followed by water. A vortex mixer was used to vigorously stir solutions
following the addition of each reactant. The prepared solutions were a dark orange/red
and were stable for several weeks. Freshly prepared solutions were dispersed on cleaned
and dried quartz slides, spun for five minutes and air-dried overnight. Films were
calcined at 400 0C, 600 0C, 800 0C or 1000 °C in a muffle furnace for 24 h. Films are
designated “xCTy” where x is the calcination temperature and y is the CezTi atomic ratio.

X-ray Diffraction (XRD). XRD diffraction patterns of calcined films on quartz

slides were obtained with a Rigaku XRD diffractometer employing Cu K01 radiation

73

(7t=1.541838 A). The X-ray was operated at 45 kV and 100 mA. Diffraction patterns

were collected using diffraction and scattering slit widths of 1°. The background due to
the quartz substrate was removed from the patterns by subtracting the XRD pattern of a
blank quartz slide. Peak widths and locations were determined using a non-linear least
squares curve fitting program [14] and assuming Gaussian line shapes. Mean particle
sizes were calculated using the Scherrer equation [15] assuming no line broadening due
to stress.

X-ray Photoelectron Spectroscopy (XPS). Materials were analyzed with a VG
Microtech spectrometer using a Clam2 hemispherical analyzer. All XPS spectra were
collected using an Al anode (1486.3 eV) operated at a power of 600 W (15 kV and 40 mA
emission current) with an analyzer pass energy of 50 eV. Four regions were scanned for
each film: C 1s, 0 Is, Ti 2p, and Ce 3d. Binding energies were referenced to adventitious
carbon (C 1s = 284.6 eV) and were measured with a precision of :1; 0.1 eV. Quantitative
XPS calculations were performed using experimentally derived sensitivity factors
previously determined in this laboratory. XPS peaks were fit with 20% Lorentzian-
Gaussian mix Voigt functions using a non-linear least squares curve fitting program [14].

Values reported have a standard deviation of <10%.

4.4 Results and Discussion
10% Ce/T i Films

The XRD patterns for the xCTO.11 films are shown on Figure 4.1. The average
particle size of each phase and the surface CezTi atomic ratios for each film are reported

in Table 4.1. The 400CTO.11 films are amorphous (Figure 4.1a) with a surface Ce/T i

 

74

atomic ratio of 0.4. Two peaks are observed for the 600CTO.11 film. a peak at 25.5
degrees assigned to the <101> plane of anatase TiO2 and a broad peak at 28.6 degrees
assigned to the <111> plane of Ce02. The average particle size of the TiO2 and Ce02
phases are 15 and 3 nm, respectively. It should be noted, however, that the Ce02 particle

size is near the limit of detection for XRD [16].

 

 

 

Table 4.1. Average particle size and surface Ce/T i atomic ratios for the xCTO.11
films as a function of calcination temperature.
Calcination Anatase Rutile Ce02 Ce/T i
Temperature Particle Size Particle Size Particle Size Atomic
(°C) (Inn) (11111) (11m) Ratio
400 -- -- -- 0.40
600 15 -- 3 0.70
800 36 47 21 23
1000 -- 78 57 40

 

 

 

 

 

 

The surface Ce/Ti atomic ratio of the 600CTO.11 film increases to 0.7. This
increase is attributed to the depletion of amorphous titania as the anatase TiO2
crystallizes. A fraction of the surface Ti atoms will be transferred to the bulk of the
titania crystallites and removed from the sampling depth of the instrument. Since only
the surface of the film is examined by XPS, these Ti atoms are essentially “lost” to the

instrument and the Ce/T i surface atomic ratio will increase.

75

 

 

 

Rutile
«:i] l U.»

l

78 nm

C602
<1 11>

 

 

1000 0C

 

 

Figure 4.1.

I
25 26 27 28 29 30 31
Degrees (26)

Powder X-ray diffraction patterns of the 10% Ce02/TiO2 films calcined
at a) 400 0C, b) 600 °C, c) 800 °C, and d) 1000 °C.

76

Three peaks are apparent in the diffraction pattern of the 800CTO.11 film (Figure
4.1b): the <101> anatase peak, the <111> Ce02 peak, and a peak at 27.5 degrees (20)

assigned to the <110> plane of rutile TiO2. The presence of both anatase and rutile TiO2
indicate that the phase transition to rutile TiO2 is occurring. The average particle sizes of
the anatase and rutile TiO2 phases are 36 and 47 nm, respectively. The Ce02 particle size
increases to 21 nm. The surface Ce/Ti atomic ratio increases to 23 indicating a dramatic
surface enhancement of cerium (vide infia). Only the rutile TiO2 and Ce02 peaks are
observed in the diffraction pattern for the 1000CTO.11 film (Figure 4.1d). The absence of
the anatase peak suggests that the phase transition to rutile TiO2 is complete. The particle
size of the rutile phase increases to 78 nm and the Ce02 particle size increases to 57 nm.

The surface Ce/Ti atomic ratio increases to 40.

20% Ce/T i Films.

The XRD patterns for the xCTO.25 films are shown in Figure 4.2. The average
particle sizes of each phase and the surface Ce/T i atomic ratios for each film are reported
in Table 4.3. The 400CTO.25 film is amorphous (Figure 4.2a) with a surface Ce/Ti

atomic ratio of 0.7.

 

 

 

 

   

 

77
US
77 nm
(:60:
Rutile <1 l ]>
<1 10>
1000 0C
60 um ‘
u! 36 nm
Anatase
<101> 800 0C
44 nm
600 0C
39 n 5 III
33 um
I l r l I I I l' I I I I I

 

Figure 4.2.

Degrees (29)

Powder X-ray diffraction patterns of the 20% Ce02/Ti02 films calcined
at a) 400 °C, b) 600 °C, c) 800 °C, and d) 1000 °C.

 

78

 

 

 

Table 4.2. Average particle size and surface Ce/T i atomic ratios for the xCTO.25
film as a function of calcination temperature.

Cale. Anatase u] + u2 Rutile u3 Ce02 Ce/T i .
Temp. Part. Size Particle Size Part. Size Part. Part.Size Atomic
(°C) (um) (run) (run) Size (nm) Ratio

(run)
400 -- -- -- - -- 0.7
600 33 39 -- -- 5 0.8
800 44 51 36 -- 15 31
1000 - -- 83 77 60 70

 

 

 

 

 

 

 

 

There are four peaks in the diffraction pattern of the 600CTO.25 film (Figure
4.2b): the <101> face of anatase TiO2, the <111> plane of Ce02, and two unidentified
peaks at 26.1 and 26.8 degrees (20) (designated 111 and u2, respectively). The peak
separation and the relative intensities of these peaks are similar for every occurrence of ul
and u2, and we believe that these two peaks represent one phase. The peak positions do
not match any previously reported phase of titanium oxide or cerium oxide, nor do they
match peak positions previously reported for CeTiO3, Ce2Ti~05, Ce2Ti207, or CeaTi9024
[17]. We believe that these peaks are the result of incorporation of cerium within the
titanium oxide lattice. It is possible that interactions with the quartz substrate also
influence the structure of this phase. The average particle size of the anatase and Ce02

phases are 33 and 5 nm, respectively. The average particle size of the u] and u2 phase

79

(based on the 111 peak width) is 39 nm. The surface Ce/T i atomic ratio is 0.8 which is
approximately that observed for the 400C025 film.

Five peaks are present after calcination at 800 °C (Figure 4.2c): the <101> anatase
peak, the ul and u2 features, the <1 11> Ce02 peak, and the <111> peak for rutile TiO2.
The relative intensity of the anatase peak decreases significantly suggesting that anatase
TiO2 is changing to rutile or the u] and u2 phase. The Ce02 particle size increases to 15
nm and the surface Ce/T i atomic ratio increases to 31.

A new peak at 26.4 degrees (20) (designated u3) is observed in the XRD pattern

of the 1000CTO.25 film (Figure 4.2d) in addition to the <110> peak of rutile TiO2 and the
<111> peak of Ce02. The identity of the u3 phase is also uncertain but may be due to the
incorporation of cerium into the rutile TiO2 phase. The average particle sizes of the rutile
and Ce02 phases are 83 and 60 nm, respectively. The particle size of the u3 phase is 77
nm. The surface Ce/T i atomic ratio increases to 70 nm.

30% Ce/Ti Films.

The XRD patterns for the xCTO.43 films are shown on Figure 4.3. The average
particle sizes of each phase and the surface Ce/T i atomic ratios for each film are reported
in Table 4.3. The 400CTO.43 film is amorphous (Figure 4.3a) with a surface Ce/T i
atomic ratio of 1.0.

Four peaks are observed in the XRD pattern of the 600CTO.43 film (Figure 4.3b):

the aforementioned peaks assigned to the ul and u2 phase, and two new unidentified

peaks at 25.8 and 26.6 degrees (20) (designated u4 and u5, respectively). The peak

I‘.I—r.‘ ' ' -

 

80

 

CbCh
<111>
Rutile
<110>
d u4
u5
ul
u2
C
b

 

 

 

I l I j I l I I I I T
24 25 26 27 28 29 30
Degrees (20)

Figure 4.3. Powder X-ray diffraction patterns of the 30% Ce02/I102 films calcined at
a) 400 °C, b) 600 °C, c) 800 °C, and d) 1000 °C.

cm. .’ .':.)-_‘::‘T..?'~.._: ..

81

separation and relative intensities of the two peaks are similar to the 111 and u2 peaks.
Thus, the u4 and u5 phase is similar to the 111 and u2 phase, but shifted to a larger d
spacing to account for an increased incorporation of cerium. The average particle size of
the ul (and u2) and u4 (and u5) phases are 67 and 103 nm, respectively. Ce02 is not
observed at 600 °C. The surface Ce/T i atomic ratio is 1.4. The increase in the Ce/Ti
atomic ratio is most likely due to the crystallization of titanium oxide. From the large
size of the titania crystallites, a larger Ce/T i atomic ratio would be anticipated. However,
if cerium is also incorporated into the titania lattice a moderate increase would be
expected.

The 111, u2, u4 and u5 features are all observed in the diffraction patterns of
800CTO.43 films (Figure 4.3c). The particle size of the u] and u2 phase remains

constant. The Ce/T i atomic ratio increases to 41.

Table 4.3. Average particle size and surface Ce/T i atomic ratios for the xCTO.43
films as a function of calcination temperature.

 

 

 

Calcination ul + u2 u4 + u5 Rutile TiO2 Cc02 Ce/T i

Temperature Particle Size Particle Size Particle Size Particle Size Atomic
(°C) (um) (um) (um) _ (um) Ratio
400 -- -- -- -- 1.0
600 67 103 -- -- 1,4
800 65 63 -- -- 41
1000 -- 78 73 58 53

 

 

 

 

 

 

 

 

 

82

Peaks due to the ul and u2 phase are not present in the XRD pattern of the
1000CTO.43 film (Figure 4.3d). Diffraction patterns for the u4 and u5 phase, the <111>
face of Ce02, and the <110> face of rutile TiO2 are observed. The particle size of the u3
and u4 phase increases to 78 nm. The particle size of Ce02 and rutile TiO2 are 58 and 73

nm, respectively. The surface Ce/T i atomic ratio increases to 53.

40% Ce/Ti Films.

The XRD patterns for the xCTO.43 films are shown in Figure 4.4. The average
particle sizes of each phase and the surface Ce/T i atomic ratios for each film are reported
on Table 4.4. The 400CTO.67 film is amorphous (Figure 4.4a) with a Ce/T i atomic ratio
of 1.4. Three peaks are observed for diffraction pattern of the 600CTO.67 film (Figure
4.4b), peaks at 26.1 and 26.8 degrees assigned to the ti] and u2 peaks, respectively, and
the <111> face of Ce02. The average particle size of the 111 and u2 phase is 69 nm. The
Ce02 particle size was calculated at 14 nm and the surface Ce/Ti atomic ratio is 1.4.

The same three peaks are observed in the XRD pattern for the 800CTO.67 film
(Figure 4.4c). The particle size of the ul and u2 phase increased to 79 nm and the Ce02
particle size increased to 30 nm. The surface Ce/T i atomic ratio increases to 43. Only
two features are observed in the XRD pattern of the 1000CTO.67 film (Figure 4.4d), the
u5 peak and the <110> face of rutile TiO2. Surprisingly, diffraction patterns for

crystalline Ce02 are not observed. The surface Ce/T i atomic ratio increases to 84.

f -4"'*' Q

 

83

 

u3

Rutile
<1 10>

 

 

 

C602
c <111>

 

 

 

Degrees (20)

Figure 4.4. Powder X-ray diffraction patterns of the 40% Cc02fl’102 films calcined at
a) 400 °C, b) 600 °C, c) 800 °C, and d1 1000 °C.

84

Table 4.4. Average particle size and surface Ce/T i atomic ratios for the xCTO.67 film
as a function of calcination temperature.

 

 

 

 

 

 

 

 

 

 

Calcination ul + u2 Rutile TiO2 u3 Ce02 Ce/Ti

Temperature Particle Size Particle Size Particle Particle Size Atomic
(°C) (nm) (nrn) Size (nm) Ratio

(H!!!)

400 -- -- -- -- 1.4
600 69.0 -- -- 14.3 1.4
800 78.8 -- -- 29.6 43
1000 -- 73.1 84 - 84
50% Ce/T i Films.

The XRD patterns for the xCT1.0 films are shown in Figure 4.5. The average
particle sizes of each phase and the surface Ce/T i atomic ratios for each film are reported
in Table 4.5. The surface Ce/T i atomic ratio of the amorphous 400CT1.0 film. Only the
<111> face of Ce02 is observed in the diffraction pattern of the 600CT1.0 film (Figure
4.5a). The average Ce02 particle size (Table 4.2) was calculated as 7 nm. It is interesting
to note that, unlike all other Ce loadings, no crystalline phaseiof TiO2 was observed in the
600CT1.0 films. The surface Ce/T i atomic ratio decreases to 3.0 which is due to the
depletion of amorphous ceria to form Ce02 crystallites.

Two peaks are present after calcination at 800 °C (Figure 4.5c), the <111> Ce02

feature and a peak at 25.5 degrees assigned to the <101> face of anatase TiO2,

85

respectively. The Ce02 particle size increases to 40 nm. The particle sizes of the anatase
phase is 34 nm. The surface Ce/T i atomic ratio increases to 67.

The <110> plane for rutile TiO2 is observed for the 1000CT1.0 film (Figure 4.5d).
The anatase features are no longer present indicating that the anatase to rutile transition
has occurred. The Ce02 particle size increases to 110 nm and the rutile particle size is 70
nm suggesting that this is a very crystalline film containing large crystallites. The surface

Ce/T i atomic ratio increases to 89.

 

 

 

Table 4.5. Average particle size and surface Ce/I’ i atomic ratios for the CTO.1
film as a function of calcination temperature.
Calcination Anatase Rutile Ce02 Ce/T i
Temperature Particle Size Particle Particle Size Atomic
(°C) (nrn) Size (nm) (nm) Ratio
400 -- -- -- 3.6
600 -- -- 7 3.0
800 34 -- 40 67
1000 -- 70 110 89

 

 

 

 

 

 

86

 

 

 

 

 

 

d

c

b

a
I l l I J
24 26 28 3O 32

Degrees (20)

Figure 4.5. Powder X-ray diffraction patterns of the 10% Ce02/TiO2 films calcined
at a) 400 °C, b) 600 °C, c) 800 °C, and d) 1000 0C.

87

Table 4.6 describes a rudimentary phase diagram for the Ce02/T 102 system
examined in this work. All 400CTy films are amorphous. Anatase TiO2 was only
observed at low cerium loadings 050.25) and at the highest cerium loading (y=1.0). As
the cerium loading increased, the ul (and u2) and u4 (and u5) phase appears in place of
the anatase phase as cerium atoms become incorporated into the titania lattice. Rutile
TiO2 appears at 800 0C only for films which contain anatase TiO2. The 111 and u2 phase
does not undergo a transition to rutile TiO2 until higher temperatures. The u4 and u5
phase appears not to undergo phase transitions within the calcination temperatures used in
this study.

An enrichment of ceria at the surface of the films is evident from the large surface
Ce/I' i atomic ratios of the 800CTy and lOOOCTy films. The surface segregation of Ce02
is due to a minimization of the overall surface energy of the film by migration of the
lower surface energy component (Ce02) to the surface [18]. To normalize the Ce/T i
atomic ratio for each film, the Ce surface enhancement was calculated by dividing the
surface Ce/Ti atomic ratio by the Ce/T i atomic ratio used for film preparation. The Ce
surface enhancement versus the bulk Ce/T i atomic ratio for the 800CTy and lOOOCTy
films are shown in Figure 4.6. The surface enhancement of the 800CTy films decrease
from 205 at y=0.11 to 66 at y=1. Larger surface enhancements are observed after
calcination at 1000 °C. A surface enhancement of 354 is observed for the 1000CTO.11

film which decreases to 89 for the 1000CT1.0 film.

l-rm‘m.R—Vn_£‘ . x 1959‘ —‘ Z mgr-as nah-F1

i

88

 

 

 

 

 

 

 

 

350- 9 1:1 800CTy
o lOOOCTy
300-
, o
E
250-
8
g .
"E. 200m D
m
g .
.66
E 150-
V)
E ‘ D o ‘
'c
Q) 100-
c.) '3 o
. 1:1 1:1
50-
.1
O I l I l I l I I I I
0.2 0.4 0.6 0.8 1.0
Ce/Ti Bulk Ratio .

Figure 4.6. The variation of Ce surface enhancement versus the bulk Ce/T i atomic
ratio for the 800CTy (denoted as open squares) and lOOOCTy (denoted as
closed circles) catalysts

 

if.“

4.5

10.

ll

13.

14.

15.

16.

89

References

Makishima, A.; Kubo, H.; Wada, K.; Kitarni, Y.; Shimohira, T. Comm. Am.
Ceram. Soc. 1986, 69, C-127.

Makishima, A.; Asami, M.; Wada, K. J. Non-Crystalline Sol. 1988, 100, 321.
Makishima, A.; Asami, M.; Wada, K. J. Non-Crystalline Sol. 1990, 121 , 310.

Sainz, M. A.; Duran, A.; Femanedez-Navarro, J. M. J. Non-Crystalline Sol. 1990, p
121, 315

Lavrencic Stangar, U.; Ore], B.; Grabec, 1.; Kalcher, K. Sol. Energ. Mat. Sol. Cells
1993, 31,171.

Keomany, D.; Poinsignon, C.; Deroo, D. Sol. Energ. Mat. Sol. Cells 1994, 33,
429.

 

]E

Dauscher, A.; Hilaire, L.; Le Normand, F.; Muller, W.; Maire, G.; Vasquez, A.
Surf Inter. Anal. 1990, 16, 341.

Dauscher, A.; Wehrer, P.; Hilaire, L. Catal. Lett., 1992, 14, 171.
Dauscher, A.; Maire, G. J. Mol. Catal. 1991, 69, 259.

Ogura, K.; Kawano, M.; Yano, J.; Sakata, Y. J. Photochem. Photobiol. A.:Chem.
1992, 66, 91.

Guglielminotti, E.; Boccuzzi, F. J. Mol. Catal. A 1996,104, 273.

Severin, K. G.; Ledford, J. S.; Torgerson, B. A.; Berglund, K. A. Chem. Mater.
1994, 6, 800.

Severin, K. G.; Ledford, J. S. Langmuir 1995, 11, 2156.

Software provided by Dr. Andrew Proctor, Univeristy of Pittsburgh,
Pittsburgh, PA.

Scherrer, P. Gan. Nochr. 1918, 2, 98.

Nuffleld, E. W. X-ray Diffraction Methods, Wiley: New York, 1966.

17.

18.

90

Preuss, A. Gruehn, R.; J. Solid State Chem. 1994, 110, 363.

Hofrnann, S. in Surface Segregation Phenomena, Dowben, P. A., Miller, A. ed.
CRC Press: Boca Raton, 1990, 107.

_ ‘iifi. NIT)?» tip

r 137 Thimi fifl‘p
l

Chapter 5

H2 Reduction of Sol-Gel Derived Cerium and Titanium Oxide Thin
Films

5.1 Abstract

Cerium and titanium oxide thin films are prepared by calcining valerate films at
400°C and 600 °C (films are denoted as “400CTy” and “600CTy” where y is the CezTi
atomic ratio). Structures of films prepared with y values of 0.1 and 1.0 are determined
with X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS). Films
calcined at 400 °C show no evidence of crystallinity, 600CT1.0 films contain partially
crystalline cerium oxide, and 400CT1.0 films contain both crystalline titanium and
cerium oxides. The reactivities of these films are evaluated using XPS analyses before
and after in situ H2 reduction. For all films the extent of Ce reduction was found to
increase with reduction temperature and is dependent on the sample composition and
structure. The 400CTO.1 films were the most reactive with ~77% of the cerium reduced
at the surface following H2 reduction at 350 0C. The 400CT1.0 and 600CTO.1 films were
less reactive with extents of reduction of 31% and 27%, respectively. The 600CT1.0 film
was least reactive with only 10% of the surface cerium reduced following reduction at

350 °C. Titanium was not found to be reduced in any of the films examined in this study.

91

arm-LE; SrA‘IIJiS- ' -

92

5.1 Introduction

In the past several years, cerium oxide has become an important component as a
promoter for three-way catalysts [1-10]. Cerium oxide (Ce02) can maintain precious
metal dispersion as well as increasing the thermal stability of the support (typically
alumina or silica). Additionally, the Ce‘W-Ce3+ redox couple undergoes a transformation
to Ce203 under reducing conditions and Ce02 under oxidizing conditions. This allows for
good oxygen storage capabilities[l 1]. Much effort has focused on the understanding and
tailoring of the chemistry of Ce02 and mixed cerium oxides.

Temperature programmed reduction (TPR) studies performed by Fiero et al. [2]
have found that Ce02 reduction by H2 is a two step process involving the chemisorption
of H2 followed by the removal of oxygen. Hydrogen chemisorption is an activated
process on Ce02 leading to the formation of surface hydroxyl groups. At temperatures
above 200°C oxygen anions are removed as water. Yao and Yu Yao [1] described H2
reduction of Ce02 as the removal of surface capping oxygen anions at lower temperatures
(~500 °C) followed by the removal of lattice oxygen at higher temperatures (~750 0C).
The extent of reduction is related to the sample morphology and increases with Ce02
surface area[1,12,13-15]. This is due to the increase in concentration of surface capping
oxygen anions at higher surface areas.

A great deal of work has focused on Ce02/A1203 materials as supports for three-
way catalysts. Recently, however, other mixed cerium oxides such as Ce02/Zr02 [16-
21], Ce02/Hf02 [17], Ce/Pr mixed oxides [22], and Ce/Tb/La mixed oxides have shown
promise as catalysts and catalytic supports. Insertion of Zr or Hf atoms into the fluorite

structure of Ce02 facilitates reduction by increasing bulk oxygen mobility[21].

—‘.h __ ‘L h.‘
' v
m
I
_e
I

 

93

Combining a mixed] valence rare earth oxide such as Pr60 11 or Tb407 with Ce02 increase
the oxygen storage capabilities of Ce02 and lowers the reduction temperature of surface
oxides by introducing intrinsic and extrinsic lattice defects [22].

Another material that has shown promise as a catalytic support is Ce02-TiO2
mixed oxides [23-25]. Cerium oxide is easily reduced in ceria-titania mixtures and titania
has been shown to stabilize the Ce3+ ion [23]. The structures of sol-gel prepared Ce02-
Ti02 catalysts are very dependent on the preparation technique. Dauscher et al. [24] have
determined that catalysts prepared by simply drying a CeCl3 and Ti(i0C3H7)4 yield an
undetermined precursor which is converted to Ce02 following calcination at 400 °C. The
same precursors prepared under basic conditions give a CeTiO3 precursor that becomes
amorphous following calcination.

This chapter presents a study of the reactivity of Ce02-TiO2 thin films prepared
from cerium and titanium valerate precursors. The previous chapter in this dissertation
has found that Ce loading and calcination temperature have drastic effects on the structure
of Ce02-TiO2 films. The structure of the films as a function of cerium loading and
calcination temperature are determined by X-ray diffraction (XRD) and X-ray
photoelectron spectroscopy (XPS). The reactivity of the films is examined by XPS

following in-situ H2 treatments.

5.2 Experimental
Materials. Cerium(IV) methoxyethoxide (18% w/v in methoxyethanol, Gelest

Inc.), titanium(1V) isopropoxide (Aldrich Chemical Company), and valeric acid (99+%,

 

lf‘figtl. '..-..‘I'-‘.'» -' '

94

Aldrich Chemical Company) were used without further purification. Distilled, deionized
water was used for all syntheses.

Film Preparation. The methods used to synthesize films were based on
techniques developed previously for tin and titanium oxide thin films [26,27]. Films
were prepared with cerium to titanitun atomic ratios of 1:10 and 1:1. Atomic ratios of
valeric acidzwaterzmetal were 9: 1 .511 respectively. Reactions were carried out at room
temperature in capped vials in a N2 purged glove box. Valerie acid was added to the
alkoxide followed by water. A vortex mixer was used to vigorously stir solutions
following the addition of each reactant. The prepared solutions were a dark orange/red
and were stable for several weeks. Freshly prepared solutions were dispersed on cleaned
and dried quartz slides, spun for five minutes and air-dried overnight. Films were
calcined at either 400 °C or 600 °C in a muffle furnace for 24 h. Films calcined at 400 °C
are designated “400CTy,” where y is the Ce/T i atomic ratio and “600CTy” designates
catalysts calcined at 600 °C.

X-ray Diffraction (XRD). XRD diffraction patterns of calcined films on quartz
slides were obtained with a Rigaku XRD diffractometer employing Cu Ker radiation
(Ar-1.541838 A). The X-ray was operated at 45 kV and 100 mA. Diffraction patterns

were collected using diffraction and scattering slit widths of 1°. The background due to
the quartz substrate was removed from the patterns by subtracting the XRD pattern of a
blank quartz slide. Peak widths and locations were determined using a non-linear least

squares curve fitting program [28] and assuming Gaussian line shapes. Mean particle

 

95

sizes were calculated using the Scherrer equation [29] assuming no line broadening due
to stress.

In Situ Heat Treatments. Oxidation and reduction of materials were carried out
in a reactor attached to the surface science instrument. The reactor was equipped with a
heater cartridge, a thermocouple, and temperature controller (Omega). Materials were
oxidized by heating for 2 h under a flow (100 cm3/min) of a dried mixture of 20% 02
(U .H.P., Purity Gas Co.) in He (99.995%, AGA Gas Co.). Reductions were carried out
under pure H2 (99.95%, AGA Gas Co.) which has been passed through water and oxygen
filters. Reductions were carried out for 30 min. at temperatures from 200 °C to 350 °C
Samples were reoxidized at 300 °C for two hours between each reduction to minimize
the cumulative effects of photoreduction (vide infra). Samples were allowed to cool to
room temperature before the reactor was evacuated. The sample was then transferred to
the main chamber for XPS analysis without intervening exposure to ambient air.

X-ray Photoelectron Spectroscopy (XPS). Materials were analyzed with a VG
Microtech spectrometer using a Clam2 hemispherical analyzer. All XPS spectra were
collected using an Al anode (1486.3 eV) operated at a power of 600 W (15 kV and 40 mA
emission current) with an analyzer pass energy of 50 eV. Four regions were scanned for
each film: C Is, 0 ls, Ti 2p, and Ce 3d. Binding energies were referenced to adventitious
carbon (C Is = 284.6 eV) and were measured with a precision of i 0.1 eV. Quantitative
XPS calculations were performed using experimentally derived sensitivity factors
previously determined in this laboratory. XPS peak shapes were fit with 20% Lorentzian-
Gaussian mix Voigt functions using a non-linear least squares curve fitting program [28].

Values reported have a standard deviation of <10%. Because of the potential error

 

Flu. '| r

96

induced by photoreduction, the Ce 3d spectrum was acquired in the first 2.5 min of
analysis.
5.4 Results and Discussion

5.4.1 Standard Materials. The Ce 3d photoelectron spectrum is very
complicated due to a variety of splitting and relaxation effects that occur during
photoemission of core electrons [30-36]. The XPS Ce 3d spectra of Ce(III)
acetylacetonate and Ce02 are shown in Figure 5.1a and b, respectively. The lines labeled
v" and u" correspond to the 3d5/2 and 3d3/2 states, respectively. The different lines are the
result of different final state electronic configurations. The Ce 3d spectrum of the Ce(IV)

standard (Figure 5.1a) contains three 3d5;2 features at 882.8 eV (v), 889.4 eV (v”), and
898.7 eV (v"’) and three 3d3/2 features at 900.9 eV (u), 908.7 eV (u"), and 916.5 eV (u”’).
The Ce 3d spectrum of the Ce(III) standard consists of the v0 and v’ 3d5/2 features at 881.5
and 885 eV respectively as well as the 3d3/2 features uo and u’ at 901 and 903.7 eV
respectively. The u"’(v”’) and u”(v”) are not observed for trivalent nor metallic cerium

compounds and are only characteristic of Ce(IV) compounds. Hence, it is possible to
distinguish Ce(IV) from reduced cerium based of the Ce 3d spectrum.

The optimal method to quantitate the Ce(IV) contentfrom the Ce 3d spectrum is
to determine the area fraction of the v”(u") and v”’(u”’) satellites. However, accurately
resolving the various satellites (particularly in mixed valent cerium compounds) is quite
difficult. Therefore to approximate the amount of Ce(IV) in each sample the area ratio of

the u’” feature to the total Ce 3d area was determined. A calibration curve was

 

 

 

 

 

l | I

 

920 900 880

Binding Energy (eV)

Figure 5.1 Cerium 3d XPS spectra of Ce02 (A) and Ce(HI) acetylacetonate (B). The
peaks labeled v0 (110) and v’ (u’ ) refer to final states Ce(III)[4f]l O[2p]5
and Ce(III)[4f]2 O[2p]4 , respectively. The features labeled v (u) and v’ ’
(u’ ’ ) refer correspond to a mixing of the Ce(IV)[4f]l 0[2p]5 and
Ce(IV)[4f]2 O[2p]4 final state configurations. The v’ ’ ’ (u’ ’ ’ ) peak
corresponds to the Ce(IV)[4f]0 0[2p]6 final state.

98

constructed from composite spectra of the Ce(III) and Ce(IV) standards [28]. A

completely oxidized film has a u’” area fraction of 15% which is within experimental

error of previously reported values for Ce02 [3 7, 38].

5.4.2 10% Ce/Ti Films. The X-ray diffraction pattern for the 400CTO.1 film is
featureless suggesting an amorphous material. The Ti 2pm and 2p3/2 features in the
titanium 2p XPS spectrum of this film are centered at 463.7 and 458.1 eV, respectively,
which is indicative of TiO2. Two peaks are observed in the 0 ls region, one at 529.6 eV
which is assigned to the metal oxygen backbone (the 0 Is peaks for cerium and titanium
oxides occur at the same binding energy) and a shoulder at 531.0 eV that is assigned to
hydroxyl oxygen. The area fraction of the hydroxyl peak is approximately 15% of the
total oxygen 1s spectrum.

The shape of the Ce 3d spectrum (Figure 5.2a) is similar to that of the Ce(IV)

standard (Figure 5.1a), however, v’ and u’ features at 885 and 904 eV are also present
indicating a partially reduced surface. From the u’” area fraction (12.9%) it was

estimated that approximately 11% of the surface cerium was reduced. The reduction is
associated with a rapid photodegradation of Ce02, a phenomena osbserved with many
Ce(IV) containing compounds [23, 24, 39-42]. This process, termed photoreduction, is
attributed to many effects including X-ray flux [43], integrated X-ray dose [44],
secondary electrons from the X-ray source [45], sample charging [41], temperature [46]
and high vacuum [47]. Dauscher et al. [23, 24] previously observed photoreduction in
Ce02-TiO2 powders. It was determined that an amorphous 50% ceria-titania powder was
approximately 50% reduced after 400 min. of analysis time. The 4000.1 film reduces at

much greater rate (Figure 5.3) as the surface ceria of the film is 95% reduced following

 

99

 

 

 

l l l
% KI) 8%

Binding Energy (eV)

 

Figure 5.2 Ce 3d XPS spectra of Ce02/TiO2 films: (A) 400CTO.1, (B) 600CTO.1,
(C) 400CT1.0, and (D) 600CT1.0. The number prior to CT refers to the
calcination temperature (in centigrade) and the number following CT
refers Ce:Ti atomic ratio of the precursor.

 

100

 

 

 

 

 

1.0
O
o
0.8 - . 0
. ..
.. O l.

E 0.6 -
3
'8
m d O
c:
.2 .
r: 0.4 -
is?
H
LI...

0.2 -

O
0.0 I I I I I I I l I I
0 5 10 15 20 25
Analysis Time (min)

Figure 5.3 Extent of photoreduction versus analysis time for 10% Ce:Ti oxide film
calcined at 400 °C. The fraction reduced is determined from the amount of
Ce(IV) reduced following exposure to the X-ray.

101

25 minutes of exposure to the X-ray. The discrepancy between the two samples is
probably influenced by the operating power of the two X-rays (300 W vs. 120 W).
Additionally, the difference in sample composition should also effect the rate of
reduction.

Crystalline Ce02 is observed in the films prepared by Dauscher et al. [23] which should
photoreduce to a lesser extent than our amorphous films. The Ce/T i atomic ratio was
determined to be 0.40, indicating an approximate four fold enhancement of cerium at the
surface. Similar surface enhancements were observed by Dauscher [23].

The XRD pattern of the 600CTO.1 film (Figure 5.4a) has peaks at 25.5 and 28.5
degrees. The peak at 25.5 degrees is assigned to the anatase <101> face of TiO2. The
peak width is 0.58 degrees, which corresponds to a mean particle size of 15.1 nm. The
less intense feature at 28.5 is assigned as the <111> peak of Ce02. This peak is very
broad and the Ce02 particle size was determined to be 3.4 run, very close to the size
limitation of XRD (3 am).

After calcination at 600 °C, the v’(u’) features are not as intense as in the
400CTO.1 film spectrum. Examination of this film before reduction revealed that the
extent of photoreduction (after 2.5 min. of analysis) is approximately one half (6%) of the
400CTO.1 film and is attributed to the formation of Ce02 crystallites that are less
susceptible to photoreduction [42]. The Ce/Ti ratio grew to 0.7 indicating an incrase in
the surface enhancement of cerium. This enhancement is attributed to a depletion of
amorphous TiO2 as it crystallizes. A fraction of the Ti atoms will be lost within the bulk

Additionally, the difference in sample composition should also effect the rate of

...,.l

Ar. u— . v. d“; —.

 

102

 

 

 

 

I l I l
b
a
I l l |
24 26 28 30
p Degrees 20

Figure 5.4 XRD spectra of (A) 600CTO.1 and (B) 600CT1.0.

 

 

103

reduction. of these crystals, beyond the sampling depth of the instrument. Since only the
surface of the TiO2 crystallites is sampled, the Ce/T i ratio will increase.

5.4.3 50% Ce/Ti Films. The XRD pattern for the films calcined at 600 °C is
found in Figure 5.4b. As with the 10% Ce/T i film, the materialtis completely amorphous
after calcination at 400 °C. However, after calcination at 600 °C a large peak at 28.9
degrees is assigned to the <111> peak of Ce02. From the width of this peak, the mean
particle size of the Ce02 crystallites was determined to be 7.4 nm. It is noteworthy that
the higher Ce loading inhibits the formation of the anatase TiO2 phase observed in the
10% Ce/Ti film (Figure 5.4a).

The cerium 3d spectra for the 50% Ce/T i films calcined at 400 and 600 °C are
shown in Figure 5.2c-d. Approximately 9% of the surface cerium for the 400CT1.0
sample is photoreduced following 2.5 min. of exposure. The 600CT1.0 film is only 2%
reduced due to the formation of Ce02 crystallites as described earlier. The Ti 2pm and
2133/2 features appear at 463.7 and 458.1 eV, respectively, indicating that the surface Ti
exists as Ti(IV). There is a dramatic enhancement of the Ce/T i ratio (Ce/Ti=5.45)
following an increase to the 50% loading. Increasing the calcination temperature from
400 to 600 0C decreases the Ce/Ti ratio from 5.45 to 4.45. This is due to a decrease in Ce

dispersion following Ce02 crystallite formation.

5.4.4 H2 Reduction Studies. Figure 5.5 shows the Ce 3d spectra for the
400Ce0.1 film following H2 treatments from 200 0C to 350 oC. As the reduction

temperature increases (Figure 5.5b-d), features corresponding to the u’(v’) satellites arise

 

 

 

 

 

 

104
l l l
e)
i'.‘
d)
C)
b)
a)
l I l
920 900 880
Binding Energy (eV)

Figure 5.5 Cerium 3d XPS spectra of a 400CTO.1 film before reduction
(a) and following H2 reduction at 200 °C (b), 250 °C (c), 300 °C
((1), and 350 °C (e).

105

accompanied by a decrease in the u”(v”) and u”’(v”’) satellite intensities, as is
characteristic of a reduction to Ce(III) at the surface. Following the H2 treatment at 350
°C, the Ce 3d spectrum (Figure 5.5e) strongly resembles the Ce(III) standard (Figure 5.1b)
indicating almost complete surface reduction. The variation in the extent of

reduction(determined by the decrease of the u’” area of the untreated sample) versus

reduction temperature is shown in Figure 5.6 (denoted as closed squares). The percentage
of the cerium in the +3 oxidation state increases from 3% after reduction at 200 0C to
77% following H2 treatment at 350 oC.

The extents of reduction for the 400Ce0.1 films following reduction at 350 °C
were approximately twice those reported for the reduction of high surface area Ce02
catalysts[48, 49]. We believe that this is the due to the lack of an extensive Ce-O-Ce
lattice associated with highly dispersed Ce02. Thus, we anticipate greater distribution of
reaction sites as well as a larger number of uncoordinated sites available for reaction.
This is similar to TPR results by Yao and Yu Yao [l] for Ce02/A1203 catalysts where the
high temperature peak associated with bulk oxygen removal is not observed when Ce02
is highly dispersed.

Figure 5.7 examines the Ce 3d spectra of the 600Ce0.1 film following H2
treatments. As with the 400Ce0.1 film, the 600Ce0.1 film shows signs of cerium
reduction with increasing H2 treatment temperature. The extent of reduction is much less
than the aforementioned 400Ce0.1 film (Figure 5.6 open squares) with approximately
30% of the surface cerium reduced at 350 °C. A decrease in the extent of cerium

reduction is observed following an increase in cerium loading from 0.1 to 1.0. The

“A ‘1..—
.' If

 

106

 

 

 

 

I l l
920 900 880

 

Binding Energy (eV)

Figure 5.7 Cerium 3d XPS spectra of a 600CTO.1 film before
reduction (a) and following H2 reduction at 200 °C (b), 250
°C (c), 300 °C (d), and 350 °C (e).

107

 

 

 

 

80
I
60 -
c:
.9
3 40 a
7?) I
04
e\° . :1
o
20 - I a
. o 0
1:1 0
a O
O T I I I 1 I I
200 250 300. 350

H2 Reduction Temperature (°C)

Figure 5.6 Extent of Cerium reduction versus reduction temperature for the 400CTO.1
(closed squares), 600CTO.1 (closed circles), 400CT1.0 (open squares), and
600CT1.0 (open circles) films.

 

108

400Ce0.5 film is only 27% reduced (Figure 5.6 closed circles) following H2 treatment at
350 °C. Though the 400Ce0.5 films are noncrystalline we would still anticipate a higher
concentration of bulk cerium oxide bonds and, likewise, a decrease in the availability of
reaction sites. An additional decrease in reacitivity is observed with 600Ce0.5 films: The
extent of reduction for this film (Figure 5.6 open circles) following H2 treatment at 350
°C is approximately 10%. This decrease is again attributed to the formation of crystalline
Ce02

Examining the change in oxygen content of the films following reduction (Table
5.1) can corroborate the extent of reduction. The oxygen content is based on 0 Is area
ratios versus the combined Ce 3d and Ti 2p area ratios. The predicted oxygen loss is
determined from the fraction of cerium reduced (as shown in Figure 5.6) and assumes that
Ce(IV) to Ce(III) is the only reduction which occurs. The change in oxygen content for
the 400Ce0.5, 600Ce0.l and 600Ce0.5 films are all within the error limits of the
predicted values. The loss observed for the 400Ce0.1 film is approximately twice as
large as prediced and can be attributed to the removal of carboxylates or other surface
contaminants.

The titanium 2pm and 2pm peak positions do not change over the course of the
experiments nor do new peaks corresponding to Ti(III) appear for any of the films,
suggesting that titanium is not reduced under these conditions. From previous
experiments in our laboratory, pure titanium oxide films calcined at 400 and 600 °C were

unreactive towards H2 reduction over the temperature range presented in this work [50].

 

In 17.5.

Table 5.1

109

Observed and predicted oxygen loss for 10% and 50% Ce:Ti
films calcined at 400 and 600 °C following H2 Reduction at 350 °C.

 

 

 

 

 

Ce Reduced Expected Observed
Film Ce/T i Ratio at 350 °C Oxygen Oxygen
(%) Loss(%) Loss(%)
400CTO.1 0.4 76.8 2.6 5.5
600CTO.1 0.7 30.8 3.2 3.1
400CT1.0 3.6 26.7 8.8 9.0
600CT1.0 3 .0 10.4 7.8 7.6

 

 

 

 

 

 

 

The films prepared with 10% and 50% cerium loadings calcined at 600 °C
exemplify the two structural extremes that can be observed for Ce02zTiO2 materials. The
600CTO.1 film consist of very small Ce02 crystallites and crystalline TiO2, while the
600CT1.0 films consist of amorphous Ti02 and crystalline Ce02. As predicted by
Johnson and Mooi [12], the extent of reduction decreased with increasing Ce02 particle
size. The effect of the TiO2 structure seems to be negligible in these films. The XPS
results indicate that the titanium oxidation state remains unchanged following reduction.
Therefore we believe that the titanium oxide matrix itself has very little effect on the
chemistry of the Ce02 other than maintaining a high dispersion of Ce02 at low loadings

and low calcination temperatures.

 

5.4

59

10.

11.

l3.

14.

15.

16.

17.

110

References

Yao, C. H.; Yu Yao, Y. F. J. Catal. 1984, 86, 254.

Fiero, J. L. G.; Soria, J.; Sanz, J.; Rojo, J. M. J. Sol. State Chem. 1987, 66, 154.
Trovarelli, A. Catal. Rev. :Sci. Eng. 1996, 39, 439.

Yao, Y. F. Yu; Kummer, J. T. J. Catal. 1987, 106, 307.

Su, E. C.; Rothschild, W. G. J. Catal. 1986, 99, 506.

Su, E. C.; Montreuil, C. N.; Rothschild, W. G. Appl. Catal. 1985, I 7, 75.

Herz, R. K. Ind. Eng. Chem. Prod. Res. Dev. 1981, 20, 451.

Gandhi, H. S.; Piken, A. G.; Shelef, M.; Delosh, R. G. SAE paper No. 760201,
1976.

Dictor, R.; Roberts, S. J. Phys. Chem. 1989, 93, 5846.

Le Normand, F.; Hilaire, L.; Kili, K.; Krill, G.; Maire, G. J. Phys. Chem. 1988,
92, 2561.

Miki, T.; Ogawa, T.; Ueno, A.; Matsuura, S.; Sato, M. Chem. Lett. 1988, 565.
Johnson, M. F. L.; Mooi, J. J. Catal. 1987, 103, 502.: J. Catal. 1993, 140, 612.

El Fallah, J .; Hilaire, L.; Romeo, M.; Le Normand, F. J. Electron Spec. Rel. Phen.
1995, 73, 89.

TschOpe, A.; Liu, W.; Flytzani-Stephannopoulos, M.; Ying, J. Y. J. Catal. 1995,
15 7, 42.

Bruce, L. A.; Hoang, M.; Hughes, A. E.; Tumey, T. W. J. Catal. 1996, I34, 351.

Ranga Rao, G.; Kaspar, J .; Meriani, S.; di Monte, R.; Graziani, M. Catal. Lett.,
1994, 24, 107.

Zanar, F.; Trovarelli, A.; de Leitenburg, C.; Dolcetti, G. J. Chem. Soc, Chem.
Commun, 1995, 965.

 

18.

19.

21.

22.

23.

24.

25.

26.

27.

31.

32.

33.

34.

111

de Leitenberg, C.; Trovarelli, A.; Zamar, F.; Maschio, S.; Dolcetti, G.; Llorca, J.
J. Chem. Soc, Chem. Commun, 1995, 2181.

Balducci, G.; Fomasiero, P.; Di Monte, R.; Kaspar, J .; Meriani, S.; Graziani, M.
Catal. Lett., 1995, 193.

Fomasiero, P.; Di Monte, R.; Ranga Rao, G.; Kaspar, J .; Meriani, S.; Trovarelli,
A.; Graziani, M. J. Catal., 1995, 151, 168.

de Leitenberg, C.; Trovarelli, A.; Llorca, J .; Cavani, F.; Bini, G. Appl. Cat. A.,
1996, 161.

Logan, A. D.; Shelef, M. J. Mat. Res., 1994, 9, 468.

Dauscher, A.; Hilaire, L.; Le Normand, F.; Muller, W.; Maire, G.; Vasquez, A.
Surf Inter. Anal. 1990, 16, 341.

I mm“?! 5“ J‘E j I. as“ m “I
O m

Dauscher, A.; Wehrer, P.; Hilaire, L. Catal. Lett., 1992, I4, 171.
Guglielminotti, E.; Boccuzzi, F. J. Mol. Catal. A., 1996, 104, 273.

Severin, K. G.; Ledford, J. S.; Torgerson, B. A.; Berglund, K. A. Chem. Mater.
1994, 6, 800.

Severin, K. G.; Ledford, J. S. Langmuir 1995, 11, 2156.

Software provided by Dr. Andrew Proctor, Univeristy of Pittsburgh,
Pittsburgh, PA.

Scherrer, P. Gan. Nochr. 1918, 2, 98.

Burroughs, P.; Hammet, A.; Orchard, A. F.; Thornton, G. J. Chem. Soc. Dalton
Trans. 1976, 1686. .

Thornton, G.; Dempsey, M. J. Chem. Phys. Lett. 1981, 77, 409.
Koelling, D. D.; Boring, A. M.; Wood, J. H. Solid State Comm. 1983, 47, 227.
Fujimori, A. Phys. Rev. B, 1983, 28, 4489.

Le Normand, F .; El Fallah, J .; Hilaire, L.; Legare, P.; Kotani, A.; Parlebas, J. C.
Solid State Comm. 1989, 71 , 885.

 

 

35.

36.

37..

38.

39.
40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

112

Wuilloud, E.; Delley, B.; Schneider, W. -D.; Baer, Y. Phys. Rev. Lett. 1984, 53,
202.

Kotani, A.; Jo, T.; Parlebas, J. C. Adv. Phys. 1988, 37, 37.

 

Shyu, J .Z.; Otto, K; Watkins, W. L. H.; Graham, G. W.; Belitz, R K.; Gandhi,
H. S. J. Catal., 1988, 114, 23.

Graham, G. W.; Schmitz, P. J .; Usmen, R. K.; McCabe, R. W. Catal. Lett.,
1993, 17, 175. F.

Paparazzo, E. Surf Sci. 1990, 234,253.

Paparazzo, E. J. Vac. Sci. T echnol. A 1991, 9, 1416.

 

El Fallah, J .; Hilaire, L.; Romeo, M.; Le Normand, F. J. Electron Spectrosc. Relat.
Phenom. 1995, 73, 89.

1pm..- .‘I‘.» 1:8! ' .
.1

Park, P. W.; Ledford, J. S. Langmuir 1996, 12, 1794.

Copperthwaite, R. G.; Lloyd, J. J. Electron Spectrosc. Relat. Phenom. 1978, 14,
159.

Batista-Leal, M.; Lester, J. E.; Lucchesi, C. A. J. Electron Spectrosc. Relat.
Phenom. 1977, 11, 333.

Johnson, C. E. In Electronic States of Inorganic Compounds: New Experimental
Techniques; NATO Advanced Study Institutes, Unverisity of Oxford, 1974; Day,
P. Ed.; D. Reidel Publishing Company: Dordrecht-Holland, 1975; 409.

Hirokawa, K.; Honda, F.; Oku, M. J. Electron Spectrosc. Relat. Phenom. 1975, 6,
333.

Bumess, J. H.; Dillard, J. G.; Taylor, L. T. J. Am. Chem. Soc. 1975, 97, 6080.
Laachir, A.; Perrichon, V.; Bardri, A.; Lamotte, J .; Catherine, E.; Lavalley, J. C.;
El Fallah, J .; Hilaire, L.; le Normand, F.; Quemere, 13.; Sauvion, G. N.; Touret, O.
J. Chem. Soc. Faraday Trans. 1991, 87, 1601.

Perrichon, V.; Laachir, A.; Bergeret, G.; Frety, R.; Toumayan, L.; Touret, O. J.
Chem. Soc. Faraday Trans. 1994, 90, 773.

Severin, K. Sol-Gel Derived Metal Oxide Thin Films: Design of New Materials
for Chemical Sensing, Ph. D. Dissertation. 1996.

Chapter 6

Conclusions and Future Directions

6.1 Conclusions.

The effect of chelator to metal ratio on film quality was examined for valeric
acid and acetylacetonate modified aluminum oxide thin films. The chelating ligands
used in this work dramatically affect the aluminum oxide backbone and hydroxyl
environment of the films. An increase in the acaczmetal ratio resulted in a shift to a
non-hydrogen bonding environment of the hydroxyl group and a decrease in the Al-O
backbone intensities. Quality films were produced from 15R <2, which occur at a
range where the acac ligands limit the extent of interchain hydrogen-bonding and the
extent of polymerization. At higher ligand ratios, the films become cracked due to the
formation of Al(acac)3-like particles.

Valerie acid modified films were crack-free at acid ratios of at least 3. This
value coincides with a shift to non-hydrogen bonded hydroxyl groups which may limit
the extent of cross-linking. It is believed that both organic modifiers used in this work
improve film quality by decreasing the stiffness of the aluminum oxide backbone due to
a decreased extent of crosslinking.

The effect of the Ce/Al atomic ratio on the surface area and cerium dispersion of
Ce:A1 mixed oxide powders was examined in Chapter 3. The powders were prepared

by the sol-gel method from cerium and aluminum valerate precursors. The presence of

113

 

114

bridging cerium-aluminum hydroxyl groups for the uncalcined catalysts suggest that
some mixing of the two metals occurs during preparation. XPS measurements indicate
that the cerium is uniformly dispersed following calcination at 500 °C with the
exception of very high cerium loadings where crystalline Ce02 is observed. The surface
areas of most cerium containing catalysts are nearly twice that observed for the alumina
support. This enhancement is attributed to cerium atoms inhibiting the sintering of
aluminum oxide by occupying potential alumina sites.

Following calcination at 900 °C, XRD patterns reveal that Ce02 crystallites are
present in all cerium containing catalysts. The average crystallite size increases from 10
to 17 nm as the Ce:A1 atomic ratio increases from 0.1 to l. The surface area decreases
from 55 mZ/g to 10 mzlg over the same range of atomic ratios. This decrease is
attributed to increasing Ce02 particle size.

The effect of cerium loading and calcination temperature on the surface and bulk
structures of sol-gel derived cerium/titanium mixed oxide thin films was examined in
Chapter 4. All films were amorphous after calcination at 400 °C. After calcination at
600 0C, anatase Ti02 and Ce02 phases were observed for films with low cerium
loading. Unidentified phases attributed to the incorporation of cerium within the
titanium oxide lattice were observed for higher cerium loadings. A moderate surface
enhancement of cerium was observed for all films calcined at 400 and 600 °C.

Rutile TiO2 was observed for low cerium loadings (Ce:Ti 5 0.25) following
calcination at 800 0C. A mixture of unknown phases was observed for the 800CTO.43

films and a mixture of unknown phases and Ce02 was observed for the 800CTO.75

' .wu awn-n:
. .

 

LEA“.- .~ .'

115

films. Anatase Ti02 and Ce02 was observed in 800CT1.0 film, suggesting that at high
cerium loadings, the anatase to rutile transition is shifted to higher temperatures. Rutile
TiO2 is observed in all films calcined at 1000 °C, a new unidentified peak is observed
for the 1000CTO.25 and 1000CTO.75 films which is attributed to incorporation of Ce
within the rutile Ce02 lattice. Films calcined at 800 and 1000 0C, have a very large
surface enhancement of cerium which decreases with increasing cerium content.

The cerium loading and cerium and titanium oxide crystallinity on the reactivity
of the 10% and 50% Ce:Ti films calcined at 400 and 600 °C, was examined by in-situ
H2 reduction. Only cerium was observed to reduce. The extent of cerium reduction was
observed to decrease with increasing Ce02 particle size and Ce loading. The decrease
with increasing Ce02 crystallite size is attributed to a decrease in the concentration of

the more reactive surface capping oxygen ions.

6.2 Future Directions

To be an effective catalytic support, the Ce02 phase of ceria-alumina catalytic
supports must be either well dispersed or exist as small crystallites. In Chapter 3 we
have demonstrated that the sol-gel process is an excellent method for producing these
materials. The ceria phase was well dispersed up to cerium loadings of approximately
70 wt% following calcination at 500 °C. However following calcination at 900 °C,

cerium is less dispersed and the surface areas of the catalysts decrease significantly. It

 

116

would be of great technological interest to maintain the high cerium dispersion and
surface areas of the catalyst at high Operating temperatures. From this work there seems
to be two possible routes to achieve this.

The first route involves the use of a different organic chelating agent or different
chelator to metal ratios to promote better mixing at the precursor stage. Work by Narula
[1] has shown that bimetallic cerium-aluminum precursors are less crystalline after
calcination at 900 °C. It would follow, that a better mixing of the two metals by our
method should also decrease the crystallinity. The homogeneity of the starting materials
may be influenced by the type of chelator used and the chelatorzmetal ratio. We believe
that if hydrolysis and condensation of the two precursors are slowed, adequate mixing of
the two metals may occur. A second route is the incorporation of textural stabilizers
such as La, Sr, or Ba. These elements limit alumina sintering by the formation of an
aluminate compound such as LaAlO; [2] or by hindering the y- to or- A1203 transition.
It is anticipated that these elements should maintain a high surface area for Ce02-A1203
catalysts at high calcination temperatures.

The structure of Ce02-TiO2 thin films was described in Chapter 4. Previously
unidentified crystal phases resulting from the incorporation of Ce into a TiO2 lattice
were observed for many of the films. It is uncertain whether these phases are the result
of good mixing of the precursors or due to the influence of the support. Characterizing
powders prepared with the same Ce:Ti atomic ratios as the films may elucidate the
effect of the support material on film structure. Additionally, the effect of mixing on

film structure can be explored by preparing the films using methods that would

 

 

 

117

theoretically prevent good mixing (such as separate polymerization of the cerium and
titanium alkoxides).

The effect of Ce02 and anatase TiO2 crystallinity and particle size on the
reactivity of CeO2zTiO2 films towards H2 reduction was examined in Chapter 5. Many
other variables can be explored such as TiO2 and Ce02 crystal structure. In addition,
other reactions can be examined such as CO and methane oxidation. Finally, Dauscher
et al. [3, 4] discovered that CeO2zTiO2 films become conductive upon ceria reduction.

This would suggest that CeO2zTiO2 thin films have applications for chemical sensors.

6.3 References

 

1. Narula, C. In Better Ceramics Through Chemistry V, Hampden-Smith, M. J .;
Klempere, W. G.; Brinker, C. J ., Eds.; Mater. Res. Soc. Symp. Proc. 271; MRS:
Pittsburgh, PA, 1992

2. Schaper, H.; Doesburg, E. B. M.; Van Reijen, L. L. Appl. Catalysis, 1983, 7,
211.

3. Dauscher, A.; Wherer, P,; Hilaire, L. Catal. Lett., 1992, 14, 171.

4. Dauscher, A.; Hilaire, L.; Le Normand, F.; Muller, W.; Maire, G.; Vasquez, A.
Surf Inter. Anal. 1990, I6, 341.

 

 

 

 

 

 

. . I t t it! . t 'lltllll
‘ .uyuoh .V. «It Nb . s. 00!..- .
taatlv .I-Jtlto‘lnltlllll
. {Ill-Ir airbuhvvyv.

 

 

. «I

 

 

 

. . Va!
1.. -,cc<t‘V1!. 1‘43;
. 1.0«1 ilk. '- . . .II t.
. elicit...‘ ”N155" \.
. in}. t1 .

. . , .r. 1-141%.1I1...
TIOI. 1:4.
90. 0.111104

.J/

t!’

‘lé -

vita

. “Wright
: ' ,1.le
H: 1.;, .. .

til. .IAIID . . .
vest .ub .- v neuritaflv a

mam .

W

I\0
. .. «£11.. ..
cl‘d Oll-

it:

In .
.Q

xi "I

. A. I.

v w \ rt:

» V v.

. I‘ll >7.

.
Nu,

. . . _ . .. . . . . :de
vikfid. .. .3. .Mukufl .nn. .. . .. . . . . . .. . .. . . .
Ema .. a in... $3.... .
.. . . ya Ema...

4a.... T...

in

..l J...
fifth”... .
K and v.9? . . . t
r- i an... r. . . . «n.3, 5
2? J., r . tart... _. 6...... 1