134?. .
€343".
. =

,. we“.

.. 31.1...
a a
Y
1....»
3 .- , at
..

.531. HERE
{37.1.9545
til-.1 1 3‘

a S
11.... ..
5:11,!vi .7
I ,‘ i\> V
.. 711$.» 1.. ‘ .
It Vn“ .
, 1 . . .
V . ‘ .‘ 3...:
L?! i. I. :1
£41.“: tawfi I
x .. s
.. . -Yl
5.9:: -1
..

“FF .3.

$.33“.

«(51.5.3
3 .3?- I.

t .l
2...?! I.)
2.

$1.1!
Iv. :f.

1 £11.13.
{"4315 v2.

 

3.335; lfi.‘
1: J .
.mzx..!......h...au.. v.
.. .n.‘ b5 “thrift; ! .
1.. a? a
, a.
.

inn.» aw".
WK . “to.
an.

.. v
.017 ,
.le.

. .

v.i.i..2.....

I 2: v1.3.
#1.}.
2‘5

.12: it...
2:, n

he»...
7“ .3917. 4‘4. ,
.9 :3.5sz2 00.}...
Ia>s~\ V .2329...
.Ivkpill...’ .

 

r.'—‘eV

 

lllllHIUIIllilllNllllllllllllllllllllllllllif'lilllllls

293 01389 3700

This is to certify that the

dissertation entitled

DESIGN AND CHARACTERIZATION OF MIXED TRANSITION
METAL OXIDE THIN FILM GAS SENSORS
DERIVED THROUC.‘ SOL- GEL CHEMISTRY

presented by

Lee H. Hoffman

has been accepted towards fulfillment
of the requirements for

”.5. degree in Chemi StY‘y

 

%/ ML

 

Major professor

Date 3‘02 7*f0/

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

 

 

 

LIBRARY

Mlchlgan State
University

 

 

 

.—

. ..... . .,,.. U ..
PLACE ll RETURN BOXto monthl- chockwtftom your record.
TO AVOID FlNESntumonorbdmdato duo.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Maneuv- Action/Equal Oppoctunlty lmtltwon
Wain-9.1

DESIGN AND CHARACTERIZATION OF
MIXED TRANSITION METAL OXIDE THIN FILM GAS SENSORS
DERIVED THROUGH SOL-GEL CHEMISTRY

By

Lee W. Hoffman

A THESIS

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

MASTER OF SCIENCE

Department of Chemistry

1996

ABSTRACT

DESIGN AND CHARACTERIZATION OF
MIXED TRANSITION METAL OXIDE THIN FILM GAS SENSORS
DERIVED THROUGH SOL-GEL CHEMISTRY

By

Lee W. Hoffman

Several surface and bulk characterization techniques have been used to study the
effects that temperature and loading have on the structure of vanadium oxide/titanium
oxide thin film conductimetric sensors. This Work establishes the crystal structures of the
oxides present in the film as well as the oxidation state and the coordination of vanadium.

The mixed oxide thin films were characterized using FTIR, XRD, UV-Vis and
XPS. The TiOz crystal structure changes from amorphous to anatase to rutile as the
annealing temperature is increased. The presence of vanadium lowers temperatures at
which phase transformation occurs, even at low loadings. An increased particle size was
observed as temperature was increased. XPS results show the change in oxidation state of
the surface vanadium ion as temperature as well as a function of loading. A migration of
the vanadium species to the surface increased as the annealing temperature increased or
the amount of bulk vanadium increased. The coordination and the oxidation state of bulk
vanadium ion were obtained through UV-Vis. The TiOz band gap remains unchanged as
the vanadium loading is increased. However, the T102 band gap is lowered as temperature

increases and finally collapses at high temperatures.

To my wife and my son.

iii

ACKNOWLEDGMENTS

When I applied to MSU for graduate studies in chemistry, I planned on furthering
my knowledge in this discipline. I had no idea, however, that my knowledge base in many
other areas would grow at the same time I as that in chemistry.

As a result, I have many to thank and, as always, first I will thank my family and
foremost on this list is my wife, Dawn. She has been a source of steadfast support
throughout the time I have been working towards this goal. My son, Stephan, has
compromised his desire to have dad there at every moment to “do things” when, instead,
much of the time I was absorbed in school. My family patiently waited for this time to
come when we can ‘start our new life.’ My parents, too, deserve a “thank-you” for their
support in anything I have tried.

Next, I want to thank all the faculty I received help and guidance from at Michigan
State University. There are a few in particular that I would like to mention. Dr. Allison
helped persuade my decision to attend MSU for graduate studies. When I interact with
him, it seems he understands that I am human. That was nice. His assistance these last
few months were needed to bring this project to an end, and I really appreciate everything
he has done. Also, without Dr. Jeffrey Ledford, the projects described in this thesis would
not have been possible. Dr. Ledford was a great source to go to when it seemed like

everything I was doing was a waste. I wish he could have been here through the end and I

iv

hope the best for him in the future. I would also like to thank Dr. Paul Hunter. It was a
great pleasure teaching under his direction.

The Ledford group deserves many thanks. Paul, you were a great companion, you
were there when I needed an ear, sometimes responding with helpful opinions (albeit not
always the opinion I would have preferred!). Although our friendship will last, our
personal day-to-day interaction will be missed. Mike, you showed me the power of a
laugh, especially at myself. Per, Greg and Kathy, your experience with sol—gel chemistry,
sensor characteristics, and all the methods used to characterize our films I relied on and
are thus an integral part of my research. I want to thank Greg and Per as well for helping
me through ‘crisis’ situations in data manipulation and presentation. I would also like to
thank many of the current and former Dantus group members. Your support was there
when I needed it. And Carl, whom I was mistaken for several times (and vice versa), I am
glad our friendship evolved (Stephan and Dawn too).

My final thanks goes out to Jim. I met Jim about a year and a half ago, at the point
when I was feeling like my future was pointless. Throughout the time we have known
each other, I have been able to put direction back into my life. Like I told him, “I am in
the driver’s seat now, and whether I am going to see what is ahead of me or what is

behind me is my choice.”

TABLE OF CONTENTS

Page
List of Tables ................................................................................................................ viii
List of Figures ................................................................................................................ ix
Chapter 1: Methods to Synthesize and Characterize Thin Film Conductimetric Gas
Sensors ........................................................................................................ 1
1.1 Introduction ................................................................................................... 1
1.2 Metal Oxides and Conductimetric Gas Sensing ............................................... 2
1.2.1 Metal Oxide Semiconductor Surfaces ............................................... 2
1.2.2 Conductivity in Metal Oxides ........................................................... 4
1.2.3 Reactions at Oxide Surfaces ............................................................. 8
1.3 An Overview of the Sol-Gel Method ............................................................ 10 -
1.3.1 Transition Metal Sol—Gel Chemistry ............................................... 11
1.3.2 Metal Carboxylate Thin Film Synthesis Via Carboxylic Acid
Complexation ............................................................................... 16
1.3.3 Synthesis and Properties of Titanium Oxide Thin Films ................. 18
1.3.4 Synthesis and Properties of Vanadium Pentoxide Gels .................... 21
1.4 X-Ray Photoelectron Spectroscopy (XPS) ................................................... 23
1.5 Ultraviolet-Visible Spectroscopy (UV-Vis) ................................................... 26
1.6 X-Ray Diffraction (XRD) ............................................................................. 29
1.7 List of References ......................................................................................... 31

Chapter 2: Effects of Annealing Temperature on the Structure of Vanadium
Oxide/Titanium Oxide Thin Films Derived from Metal Carboxylate Sol-Gel

Processing ................................................................................................ 36
2.1 Introduction ................................................................................................. 36
2.2 Experimental ................................................................................................ 39.
2.3 Results and Discussion ................................................................................ 42
2.4 Conclusions ................................................................................................. 56
2.5 List of References ....................................................... - .................................. 58

Chapter 3: Effects of Doping Levels on Vanadium Oxide/Titanium Oxide Thin Film

Conductimetric Sensors .............................................................................. 60

3.1 Introduction ................................................................................................. 60
3.2 Experimental ................................................................................................ 62
3.3 Results and Discussion ................................................................................ 64
3.4 Conclusions ................................................................................................. 75
3.5 List of References ......................................................................................... 78
Chapter 4: Future Work ................................................................................................ 80
4.1 Sensor measurements ................................................................................... 81
4.2 Oxidation Reactions ..................................................................................... 83
4.3 Reduction Reactions ..................................................................................... 85
4.4 List of References ......................................................................................... 88

List of Tables
Table Page
Table 2.1. Particle size of TiOz crystals in V/I‘ i thin films. Calculations
made using FWHM of respective XRD feature. ............................................ 46

Table 2.2. Binding energies of V 2pm features as a function of calcination
temperature. ................................................................................................. 5 1

Table 3.1. Particle size of TiOz(anatase) as determined through line broadening
calculations of XRD diffraction patterns. ...................................................... 66

Table 3.2. Measured Band gaps measured for V/I' i films. A film containing
TiOz(anatase) crystals is included for reference. ............................................ 71

Table 3.3. XPS data for films derived from sols containing between 2.5 and
20% V. *Determined from V 2pm / Ti 2pm intensity ratios. ........................ 73

viii

Figure

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 2.1.

Figure 2.2.

Figure 2.3.

Figure 2.4.

Figure 2.5.

Figure 2.6.

List of Figures

Page

Band model of semiconductor. X is the distance through the

semiconducting material and f is the value of the Fermi function.

(From Madou, M. J.; Morrison, S. R. Chemical Sensing with

Solid State Devices; Academic Press: San Diego, CA, 1989.) ........................ 3

Accumulation and depletion of electrons resulting in a space-charge
layer formation. (a) distribution of charges and (b) band scheme near
conduction band edge. EC = conduction band edge; 8;: = Fermi level;
S = surface states; eVs = serface potential; x = distance into
semiconducting material. (From Cox, P. A. The Electronic Structure
and Chemistry of Solids, Oxford University Press, Oxford, 1987.) ................ 5

XPS process ............................................................................................... 24

FTIR spectra of (a) VIP i film annealed at 400°C and (b) Ti film
annealed at 400°C ....................................................................................... 43

XRD patterns measured for VIP i films annealed at (a) 300°C,
(b) 400°C, (c) 500°C, and ((1) 600°C ........................................................... 45

UV-Vis spectra collected for primarily amorphous thin films
containing the following crystalline phases (a) TiOg (anatase),
(b) TiOz (anatase) and T102 (rutile), and (c) V205 ....................................... 48

UV-Vis spectra obtained for V/T i films annealed at (a) 300°C,
(b) 400°C, (c) 500°C, and ((1) 600°C ........................................................... 50

Vanadium 2pm photoelectron spectra collected for Vfl‘ i films
(a) ascast, films annealed at (b) 300°C, (c) 400°C, ((1) 500°C, and
(e) 600°C and a reference spectra of V205 (f) .............................................. 53

Variation in XPS detennined V/T i atomic ratio as a function of
annealing temperature ................................................................................. 55

Figure 3.1. XRD diffraction patterns obtained for V/T i atomic ratios of:
(a) 0.026. (b) 0.053, (c) 0.111, (d) 0.176, (c) 0.250 ..................................... 65

Figure 3.2. UV-Vis spectra for reference films containing (a) TiOz(anatase)
01' (b) V205. ................................................................................................ 68

Figure 3.3. UV-Vis spectra for films containing V/T i atomic ratios of
(a) 0.026, (b) 0.053, (c) 0.111, (d) 0.176 and(e) 0.250. ............................... 69

Figure 3.4. XPS results showing the effect of vanadium loading on the feature
due to V2p3n. Films were derived from sols with (a) 0.026,
(b) 0.053, (c) 0.111, (d) 0.176, and (e) 0.250 V/T i atomic ratios ................. 72

Figure 3.5. Comparison of surface and bulk V/I‘ i atomic ratios.
Experimental data represented by circles. Spuares represent a plot for

equal bulk and surface ratios ....................................................................... 74
Figure 4.1. Schematic of sensor array ............................................................................ 82
Figure 4.2. Cell used to make sensor measurements. ..................................................... 84

Chapter 1

Methods to Synthesize and Characterize Thin Film
Conductimetric Gas Sensors

1.1 Introduction

The semiconducting properties of transition metal oxides are well understood, and
sensors based on TiOz are becoming increasingly popular. Current interest in the chemical
sensing properties of these oxides involves issues relative to the monitoring of air
pollutants [1]. To date, most sensor studies have focused on single crystals of the metal
oxides. Recently, however, the development of sol-gel chemistry has led to new sensor
materials based on metal oxide thin films. TiOz-based sensors have been studied [2-5] and
VIP i oxide materials have received attention in catalysis studies for the oxidation of
hydrocarbons [6.7] and the reduction of NOx compounds [8-10]. However, sensor studies
utilizing the properties of the sol-gel derived V/l‘ i system have not been investigated.
Thus the goal of this work is to characterize vanadium oxide/titanium oxide
conductimetric sensors produced via sol- gel chemistry.

Before sensor properties can be studied and/or evaluated, an appropriate system
must be developed. In Vfl‘ i catalyst studies, it has been shown that both the amount of
vanadium present [10-12] and the temperature at which the catalyst is prepared [13,14]
affect the material’s structure. It would therefore be appropriate to investigate the affect
these two factors have on the structure of 501- gel produced vanadium oxide/titanium oxide

thin films. A variety of spectroscopic techniques (XPS, XRD, and UV-Vis) will be used

2
to characterize both the surface and bulk structures of the films. Prior to describing the

results of the study, properties of and reactions occurring on semiconducting metal oxide
surfaces will be discussed, along with background on the spectroscopic techniques used

for the characterization of these thin films.

1.2 Metal Oxides and Conductimetric Gas Sensing

1.2.1 Metal Oxide Semiconductor Surfaces

A semiconductor is a non-metallic solid that conducts electricity through thermal
excitation of electrons across an energy gap. A band model of a semiconductor surface is
represented in Figure 1.1. The energy (or band) gap EC is the energy separating the
valence band Ev from the conduction band EC. Applying this to metal oxides, the valence
band is the highest occupied molecular orbital (HOMO) of oxygen while the conduction
band is the lowest unoccupied molecular orbital (LUMO) of the metal ion [15]. The
Fermi energy E; represents the energy of the highest filled level. For a semiconductor, E;
is defined to lie in the energy gap between the valence and conduction bands of the
electron distribution at the point at which a level has an equal chance of being occupied or
empty. The stoichiometric metal oxide has no electron holes in the ground state (contains
no defects) [15]. Thus, at any given temperature, the number of electrons in the
conduction band of a stoichiometric metal oxide must equal the number of holes in the

valence band resulting in the Fenni energy level placed midway between the two bands.

AB A5

 

 

 

 

 

 

 

A BC
__._____EF_.___.____
E0
V
EV
4 D
1 0 x
f

Figure 1.1. Band model of semiconductor. X is the distance through the semiconducting
material and f is the value of the Fermi function. (From Madou, M. J.;
Morrison, S. R. Chemical Sensing with Solid State Devices; Academic Press:
San Diego, CA, 1989.)

However, the presence of impurities can affect semiconductors in two ways [15]. First,
n-type semiconducting materials result when the number of electrons in the conduction
band is greater than the number of holes in the valence band with the Fermi energy level
closer to the conduction band. On the other hand, p-type semiconducting materials result
when conditions are opposite resulting in the Fermi energy moving closer to the valence
band.

Redox properties of transition metal oxides allow these materials to be non-
stoichiometric (or contain defects) [16]. Since these oxides can be easily reduced, oxygen

atoms may be removed resulting in the reduction of the metal ion. The d electrons of the

4
metal ion associated with the oxygen vacancies occupy electronic states within the

bandgap [16]. Surface defects play an important role in the catalytic processes of
semiconducting transition metal oxides. Through thermal treatment, TiOz thin films to be
used as chemical sensors can be made slightly oxygen deficient (adding electrons to the
conduction band), resulting in n-type sensing properties [5].

Surface electronic states, caused by defects (O vacancies) or adsorbed species,
modify the band energy near the surface of n—type semiconductors. When the surface of a
semiconductor accepts a charge, a space-charge layer of equal but opposite charge is
formed. Figure 1.2 shows the influence electron donors or acceptors have on the surface
of an n-type semiconductor. The adsorption of electron acceptors such as oxygen
removes electrons from the semiconductor. The removal of electrons forms a depletion
region in the space-charge layer. The adsorption of electron donors such as hydrogen, on
the other hand, will form an accumulation region by donating electrons to the space-
charge layer. Thus, as a depletion layer is formed, the removal of electrons decreases the

carrier density and the conductance decreases.

1.2.2 Conductivity in Metal Oxides

Metal oxide conductimetric gas sensors utilize changes in conductivity resulting
from the chemisorption or reaction of the analyte gas with the oxide surface.
Chemisorption of the analyte gas changes the surface potential, the grain boundary
potential and contact potential of the oxide surface through electron transfer [17]. The

surface of the metal oxide contains a finite number of electron donors (e.g. adsorbed

 

 

 

 

a)
Depletion Accumulation
/ ionized donors [electrons
— / + _ /
_ ++ + _ '-
_ ++ . . . + _ —
vacuum _ + + semrconductng material + _ —
_ + + _ -
b)

 

 

8—

 

 

 

 

 

Figure 1.2. Accumulation and depletion of electrons resulting in a space-charge layer
formation. (a) distribution of charges and (b) band scheme near conduction
band edge. EC = conduction band edge; Ep = Fermi level; S = surface states;
eV, = surface potential; x = distance into semiconducting material [From Cox,
P. A. The Electronic Structure and Chemistry of Solids; Oxford University
Press; Oxford, 1987].

6
hydrogen) and/or electron acceptors (e.g. adsorbed oxygen). If we consider a surface

saturated with adsorbed oxygen atoms (electron acceptors), a space-charge layer will form
near the surface. This layer extends into the interior of the semiconductor through the
extraction of electrons from the lattice. Conductivity of this space-charge region is altered
by a change in the surface concentration of the ionosorbed oxygen.

The bulk conductivity 0,, of a semiconductor can be expressed as:

0b = "(min + Ppr (1-1)
where nb and pb are the bulk concentrations of the electrons and holes, respectively and q
is the electronic charge. (Conductivity is generally expressed as mhos cm".) The
mobilities of the electrons and holes are represented by it. and up, respectively. The
mobility of the carrier is its drift velocity per unit electric field and is inversely related to
temperature resulting in decreased mobility (lower conductivity) with increased
temperature according to the relation below:

11 o. T” (1.2)
Bulk conductance of a semiconducting material is given by:

G, = obWt/L (1.3)
where W is the width of the film, L is the length, and t is the thickness. The overall
conductance of the semiconducting metal oxide sensor is influenced by both surface as
well as bulk components G. and Ch:

G = Go + G. (1-4)

The calculation for the surface conductivity 0, is different from that of the bulk

conductivity in that it does not need to take into account the presence of holes (electron

7
deficiency) in the oxide lattice. For n-type metal oxide semiconductors such as TiOz,

changes in surface conductivity Ac, can be determined using the following equation [18]:
A6. = qllsAn. (1-6)

The number of surface carriers n, for semiconductors is expected to increase as the
temperature increases according to the following relation [15]:

n, oc exp(-EG /2kT) (1.7)
Note that the band gap Ea for most semiconductors is large compared to kT. The
predicted Arrhenius-like behavior results in an activation energy equal to one half the band
gap of the semiconducting material.

Since we are considering a surface, the conductance is independent of thickness.
Surface conductance is therefore expressed in units of mhos or “mhos per square” because
the conductance between the sides of a square plate is independent of the size of the
square. The change in surface conductance is expressed as:

AG, 2 Ao,W/L (1.8)
where W is the width of the film and L is the length (the distance between contacts).

There are three generally accepted mechanisms describing the performance of
metal oxide sensors based on a barrier to electron transfer: 1) the analyte, reacting with
lattice oxygen, will increase the number of defects which in turn increase the number of
donor electrons, 2) electrons are injected into the semiconducting surface by adsorption of
an electron donating analyte, or 3) oxygen atoms adsorbed on the surface react with the

analyte resulting in electrons being injected into the metal oxide surface when the products

8
desorb [16,19]. Studies of TiOz sensors in air have used the third mechanism to describe

surface reactions of the sensor with analyte gases [20,21].

1.2.3 Reactions at Oxide Surfaces

Morrison has used a simple model to approximate the sensor response of n-type
semiconductors to a reducing analyte [22]. Morrison showed the variation of the surface
charge is due to the interaction of oxygen atoms present on the surface with the analyte

gas. Several simple reaction steps were postulated:

k.
C' + 02 '—. 02-
.. (1.9)
e- + 02' -—’ 20 '
1‘2 (1.10)
R+02- —»R02+e-
ks (1.11)
R + 0- —’ R0 + e-
k4 (1.12)

where the k’s are rate constants, R represents the reducing gas, and e' is an electron in the
conduction band. Only the first equation is considered to be frrst order since the [02] is
assumed to be constant. Since gas sensors normally operate at elevated temperatures
where Og’ can be removed from the surface [23], the first equation is reversible. The
second reaction is assumed to be irreversible since the reducing gases react readily with
0'. Products of the third and fourth reactions desorb resulting in irreversible reactions. k3

is considered to be small so it will be omitted from the discussion below.

9
The total charge on the surface, N., can be calculated from the concentration of

oxygen species on the oxide surface. We can calculate the concentration of the 02'
species and the 0' species with the steady state approximations:
[02-] = klns [02] / (k1 + kzns) (1.13)
[0'] = (2km. / k4[R])(k1n,[02] / {k1 + k2nsi) (1.14)
where n, is the electron concentration. The total charge, N., can be calculated from these
two values using the following equation:
N. = {k1n,[02] / (k1 + kzn,)} {1 + 2k2n. / k4IR]} (1.15)
Morrison also showed through Schottky’s and Boltzmann’s relations an exponential
increase of 11s with an increase in N.. Thus, in the equation (1.15), 11, changes significantly
with changes in [R] while N, stays constant. ‘ Equation (1.15) may be rewritten in a
quadratic from which n. can be determined:
{2ktkztozr mm] in? + {kilozi — 1.2m... = k-rNt (1.16)
Using this equation, the relationship of 11s as a function of [R] can be used to determine the
resistance, p, as a function of the partial pressure PR of the reducing agent:
p=0t/n, and PR=7[R] (1.17)
with or and y constants. For thin films, the removal of electrons from the surface of the
material dominates the material’s change in conductivity [24]. The end result is a

logarithmic relationship of conductance on [R].

10
1.3 An Overview of the Sol-Gel Method

Sol—gel chemistry has been used to prepare ceramic materials [25-28]. The
materials produced through sol-gel processes may be classified as follows: sol - a
dispersion of colloidal particles (usually in the range of 1-100 nm) in a liquid, and gel - a
polymeric chain, whose average length is greater than a rrricrometer, composed of a rigid
network of particles with pores on the submicrometer scale [52].

Advantages the sol-gel process has over conventional methods include easy
preparation of multicomponent materials, the use of lower processing temperatures than in
traditional ceramic preparations, and the sol and gel rheological properties that allow for
fibers, films or composites to be prepared using various techniques such as spinning, dip-
coating or impregnation [28]. With sol-gel chemistry, one can control the synthesis from
the molecular precursor to the final product, which allows for the fabrication of materials
that cannot be obtained through conventional “powder” methods. The properties of SOI-
gel materials have led to their use in a wide range of applications including the fabrication
of optics [27], coatings [26,29], electronic devices [30-32], and sensors [33-36].

The sol-gel process described here involves the use of transition metal alkoxide
prcursors to produce metal oxide materials. Transition metals may exist as mixed valences
which allows electrons to move from low to high valence states [37]. This leads to oxides
of these metals being used as: semiconducting coatings [38], switching devices [39], or

electrochromic devices [40]. Since transition metal oxide gels are particle hydrates, they

11
can exhibit high proton conductivities [41] and some of them behave as inorganic ion

exchangers [42].

1.3.1 Transition Metal Sol-Gel Chemistry

Mechanisms describing the formation of silicon-based sols and gels have been
established [43-47], followed by the development of transition metal sol-gel chemistry; but
up till now transitional metal sol-gel chemistry is less understood than silicon sol-gel
chemistry. Although many early sol-gel processes involved the use of metal nitrates or
halides, the hydrolysis of metal alkoxides has received recent attention. A metal alkoxide
may be prepared by using one of three reagents [48]:

a) a pure metal

M +nROH -—> M(OR)., + n/2H2 (1.18)

b) a metal hydroxide
M(OH),, + nROH —> M(OR)n + nH20 (1.19)

or c) a metal chloride
MClx + nNaOR —> MORn + nNaCl (1.20)

Metal alkoxides are soluble in a range of organic solvents, most notably in
solutions of their parent alcohol. Such solubility is critical to the sol-gel process. Since
our synthesis involved the hydrolysis of an alkoxide, it is important that the alkoxide be

soluble in a water-miscible solvent such as alcohol.

12
Most sol-gel preparations involve the hydrolysis of a metal alkoxide followed by

condensation and polymerization. Metal hydroxide precipitates are formed from the
hydrolysis of metal alkoxides in the presence of water [49,50]. Adding water to the

alkoxide causes the formation of a M-OH hydroxo group [28], as represented below:

H R

\

l 2 3
H-|O + M-OR ——> O: —> M-OR —-> HO-M ‘— O —’ M-OH + ROH
H

a H H
< > (b) ‘°’ ‘d’ (1.21)

Steps 1 and 2 show a nucleophilic attack of the metal atom by a water molecule. The
electronegative strength of the alkoxo group greatly influences this attack, and will be
discussed later. Proton transfer results in the intermediate labeled (c). The final step of
this mechanism is the loss of the alcohol to form a metal hydroxide. The rate at which this
reaction occurs is referred to as the rate of hydrolysis (kn).

Reactions ( 1) and (2) occur via 8N2 nucleophilic substitution. The metal center’s
ability to act as a Lewis acid has a major influence on the substitution reaction. The Lewis
acidity of the metal is affected by the coordinative unsaturation and oxidation state of the
metal. With increased acidity, proton transfer (to give intermediate c) will occur with a
lower activation energy. The electrophilicity of the metal can also affect the k" of
transition metals. More electropositive metals are more susceptible to nucleophilic attack.
Electronegative ligands increase kH by increasing the electrophilicity of the metal [51].

Another way to influence the reactive nature of the metal during hydrolysis is
through steric hindrance. Larger alkoxy groups will slow the rate of nucleophilic

substitution. Thus, the arrangement of the alkoxide ligand can be used to affect the

13
reaction rate during nucleophilic attack. Bulky alkoxide ligands act as better leaving

groups and k“ is greater when the alkoxy group is more highly branched. The influence of
the ligand on the rate of hydrolysis follows the sequence: primary < secondary < tertiary
[28].

The use of non-complexing solvents also influences the rate of hydrolysis. Non-
complexing solvents may cause the alkoxide to oligomerize which in turn causes the
reduction of k" due to additional energy that will be required to dissociate the complex
[51]. Overall, however, the hydrolysis reaction will be favored when the electrophilic
character of the metal is strong and when the leaving ability of the alcohol is high [28,52].

Condensation is initiated by the formation of hydroxyl groups formed during the
hydrolysis step, thus it is difficult to separate hydrolysis from condensation. Condensation
occurs via three different mechanisms: alcoxolation, oxolation, and olation [28]. In
alcoxolation, a metal oxygen backbone is formed when an alcohol group is removed:

M-OH + M-OR —> M-O—M + ROH (1.22)
This reaction occurs via an 8N2 mechanism, similar to that of the hydrolysis reaction.
Oxolation occurs in a similar manner to reaction 1.22 above, however it involves the
removal of a hydroxyl group:
M-OH + M-OH —) M-O-M + H20 (1.23)
If the coordination of the metal is less than fully satisfied, olation can occur with alkoxides
by the formation of hydroxyl bridges between the metal atoms after the removal of a water

or alcohol molecule:

14

‘1’ i
M-OH + M-OR —> M-O-M + ROH (1.24)
The kinetics of olation are fast when compared to other condensation mechanisms due to
the lack of a proton transfer step in the mechanism [28].

For metal alkoxides, the metal’s electrophilic character as well as its degree of
freedom (amount of coordinative unsaturation) can be used to explain the fact that
hydrolysis and condensation occur more rapidly for transition metal alkoxides than for
silicon alkoxides. The electrophilic character of some metal alkoxides M(OR)n (expressed
as the partial charge 5(M) on the metal atom) [28,53] together with their degree of

unsaturation (expressed by the difference N-n where N is the coordination number usually

found in the oxide) are as follows:

 

Alkoxide am) N - n
51(0Pr‘). +0.32 0
more), +0.60 2
more). +0.64 3.4
Ce(0Pr‘)4 +0.75 4

Increasing the electrophilic character and/or increasing the degree of unsaturation will
increase the reactivity of a metal center of a metal alkoxide towards nucleophilic
substitution.

The hydrolysis and condensation reaction rates for silicon alkoxides are slow and
are generally increased using an acid or base catalyst. However, the case for alkoxides

with a transition metal center is different - hydrolysis and condensation reactions will often

15
form precipitates unless they are slowed. The slowing of hydrolysis is most often

accomplished through the introduction of complexing ligands that can be bound to the
metal center. These ligands slow condensation in two ways: 1) the reactivity of the metal
center is lowered by decreasing its electropositive strength and/or steric hindrance and 2)
the number of groups available for hydrolysis is lowered. Organic acids and B-diketones
are the reagents typically used [51,54-62]. Organic acid ligands are stronger ligands than
alkoxides making the metal center more stable with respect to hydrolysis. The role of
organic acids in the control of titanium alkoxide hydrolysis has received much attention in
the literature [56-58,62-66].

Condensation can be controlled through modification of the water/metal alkoxide
ratio. Through determination of the weight percent of TiOz in the polymer, Yoldas et al.

[67] have shown that modifying the water ratio,

h _ IHzOI
' [M(OR)z] (1.25)

will modify the structure of the final product. Yoldas [68] also found that an increased
oxide content of the hydrolysis product is observed when the water ratio is increased.
Livage [28] used this finding to derive the following three domains: When h<1,
alcoxolation and oxolation condensation reactions occur primarily. As a result, alkoxy
bridges are formed and gelation cannot occur since there is no excess water. When 151152
(2 is the oxidation state of the metal), competition between alkoxolation and oxolation
results and chain polymers are formed. Alcoxolation is favored for h<2; but when h

approaches 2, hydrolysis may not go to completion and oxolation may result. When h>z,

16
precipitation occurs quickly and condensation occurs via olation as contrasted to

oxolation resulting in polymers formed by crosslinking [28].

From the microstructural point of view, silicate sol-gel chemistry differs from non-
silicate sol-gel chemistry in three ways: (1) due to the d-orbital configuration of transition
metal alkoxide compounds, ligand field stabilization can be used to obtain varied
structures [67]; (2) the structure of the polymer obtained can vary widely since the
coordination numbers of transition metals range from 2-8; and (3) since metal alkoxides
are more electropositive, they are often oligomeric. The size of the metal as well as the
size and nature of the alkoxide ligand will determine the steric demands on the metal
alkoxide. These demands will influence the extent to which the metal alkoxide will form
oligomers [48]. The oligomers are then used as building blocks in the resulting polymer

through suitable chemical modification.

1.3.2 Metal Carboxylate Thin Film Synthesis Via Carboxylic Acid
Complexation
It is essential to control k" when producing transition metal oxide thin films via the
sol-gel process [28,49,51,69,70]. This control allows the materials to be made from
alkoxide precursors without forming precipitates. The extent of hydrolysis is lowered by
replacing the more reactive alkoxide ligands with chemical additives such as B-diketones

[71], organic acids [56,61,72], and glycols [48].

17
Carboxylic acids can act both as chelating ligands as well as oxygen donors

(especially towards oxophilic metals). The alkoxide group is replaced by a carboxylic acid
as shown in the reaction below [28]:

M(OR)., + yXOH —-) [M(OR).,-,(OX),] + yROH (1.26)
Carboxylate ligands behave as terminating reagents on the metal. There are four possible
arrangements the carboxylate ligand could form with the metal: ionic, unidentate and
bidentate bridging, or chelating [73]. The number and arrangement of carboxylate ligands
accepted by a metal center will depend on the metal’s size, charge, and number of
available sites.

After a carboxylic acid has been added to an alkoxide precursor, addition of water
initiates hydrolysis and condensation. Hydrolysis is slowed as the carboxylate ligand is a
poorer leaving group than the alkoxide[64]. The increased size of the ligand will slow the
rate of hydrolysis even further by inhibiting attack through increased steric hindrance [74].

The chemical control of hydrolysis and condensation of metal alkoxide-carboxylate
chemistry can be demonstrated in the reaction of titanium isopropoxide with acetic acid.
Partial charge calculations show that the HOAc group in the intermediate, shown below,
to be negatively charged (8mm = -0.7) while the Pr’OH group is positively charged
(5Pr’OH = +0.1). With the coordination of Ti not being satisfied and a positive partial
charge on Ti, AcOH displaces Pr’OH through nucle0philic substitution in the following
condensation reaction:

Ti(OPri)4 + AcOH —+ [Ti(OPr’)4(OAc)H] —> [Ti(OPri)3(OAc)]., + PrOH (1.27)

18
the acetate groups form bidentate bridging ligands resulting in an oligomeric species where

Ti expands its coordination from four to six [74]. Hydrolysis of the oligomeric species
leads to:

Ti(OPr’)3(OAc) + H20 -> Ti(OH)(OPr’)2(OAc) + PrOH (1.28)

The continuity of thin films has been studied by Berglund, et al. [63,64,75]. They

showed that careful choice of appropriate acidzmetal (R,) and waterzmetal (R...) ratios is

necessary for producing a good film. Their work resulted in the formation of thin and

transparent continuous films that show controlled shrinkage and were not brittle.

1.3.3 Synthesis and Properties of Titanium Oxide Thin Films

Titanium dioxide plays a significant role in sensing materials. TiOz-based materials
exhibit surface adsorption properties that are sensitive to water vapor and other gases
[76,77]. The hydrolysis of titanium isopropoxide occurs rapidly and can be carried out
completely before the start of condensation if the conditions are acidic. These conditions
lead to transparent monolithic gels which can be used to produce crystalline TiOz in the
anatase or rutile form [67]. (Anatase and rutile are two naturally occurring crystal
modifications of T102 - rutile the most common.) Under basic conditions, carboxylate
ligands remain attached to the Ti atom causing hydrolysis and condensation to occur
simultaneously. As a result, crystallization is prevented and amorphous powders are
obtained where the morphology depends on the pH [74].

Titanium alkoxide precursors have been used in many recent studies involving sol

and gel formation. Ti(OR)4 (where R = Bu“, Bu’, Et, Pri, Pr“) has been used with

19
inorganic acid catalysts (HCl, HNO3) to produce monolithic TiOz gels [67,78]. The

hydrolysis ratio h of these systems was controlled such that 1<h<4. T102 gels were
obtained through control of the hydrolysis ratio, while sols of these systems were
produced through use of a proper acidzalkoxide ratio [67,79-81].

Severin et al. [66] have also described the production of T102 thin films via the
sol-gel process. The titanium isopropoxide precursor was modified by mixing it with
valeric acid followed by the addition of water. Sols produced from the valeric acid
modified alkoxides have better adhesion to glass and greater integrity than sols produced
using other carboxylic acids [63]. Following the reaction scheme previously described, a
complexation and hydrolysis reaction of Ti(OPr’)4 results in the formation of a metal oxide
backbone:

I l l l
— Ti- OH + -Ti— O -H —> —Ti— 0 —Ti - + H20 (1.29)
l | l |

l | | l
— T1- OPr + —Ti— O —H —) —Ti— 0 —Ti — + PrOH (1.30)
l l l |

An oligomeric structure where the Ti coordination from 4 to 6 has been observed
through chemical modification of the alkoxide precursor with organic acids [56,66,72].
Both bidentate and bridging carboxylate ligands have been observed [56,66].

Heat treatment of valeric acid modified titanium alkoxide thin films at moderate
temperatures (250°C - 600°C) removes the carboxylate ligands leading to an increase in

the titanium-oxygen backbone concentration [82]. At higher temperatures, crystalline

20
anatase (600°C) and rutile (800°C) regions appear in the film. In all cases the film is

primarily amorphous. Crystalline phases in TiOz thin films prepared through other sol gel
techniques have also been reported [83]. These studies showed that doping lowers the
temperature needed for phase transformation of TiOz.

Although thin TiOz films have not been utilized in conductimetric sensing, there
has been extensive sensor research on employing TiOz both as a single crystal and in
polycrystalline forms. The kinetics and thermodynamics of the reaction of the T102
surface with analyte gases are influenced by surface defects [84]. These defects
significantly affect the charge transfer between the semiconductor and electrolyte [85].
Dopants in the form of tri- or pentavalent cations create defects (oxygen vacancies) into
the TiOz lattice [5].

TiOz sensors have been used to monitor the combustion of fuel in engines through
the detection of oxygen [2-5]. Bulk defects of these sensors are thermodynamically
controlled [5] and can be used at high temperatures to determine the oxygen concentration
over a wide range of partial pressures. Tri- and pentavalent cations are used to control the
sensitivity and conductivity of the sensor. The response of Pt-doped TiOz sensors
prepared at low temperatures has been attributed to Schottky diode characteristics
between the metal and the semiconducting material [2].

TiOz-based sensors for the detection of nitrogen-containing compounds have also
been studied [86,87]. The reaction mechanism of these sensors varies with the dopant
added. For MoO3/I' 102 sensors, sensor response to ammonia has been attributed to

oxidation involving lattice oxygen [86]. Boccuzzi et al. [87] have shown that for TI02 or

21
Ru/Ti02, the adsorption site for nitrogen containing compounds was the Lewis acid site

with Ti“.
1.3.4 Synthesis and Properties of Vanadium Pentoxide Gels

The catalytic, electronic, and ionic properties of V205 [88-91] are interesting from
the standpoint of chemical sensing. The basicity of bridging oxygens as well as the strong
Lewis acidity associated with V“ provides reaction sites for the oxidation of
hydrocarbons, NO. reduction, and dehydration of alcohols [88,92,93].

A common way to prepare vanadium alkoxides involves the heating of an
ammonium metavanadate mixture under reflux in benzene [94]:

NH4VO3 + 3ROH —-) VO(OR)3 + 2H20 + NH; (1.31)
where R = OPr“, OAm‘. Monomers or dimers can be formed depending on the size of the
alkoxide group. The structures of vanadium alkoxides was studied by Sanchez et al. [95].
Large, bulky groups such as OAmt prevent coordination expansion of the vanadium atom
and produce a monomeric structure via alkoxy bridging. VO(OPr”) forms a dimeric
species via edge sharing of alkoxy bridges to give:

0
l

ii R (II I

’\g’ \OR OR ’V< >V

R
/0\ ,OR
\0’ ”‘OR
R

O

The high electrophilicity of the vanadium atom causes the vanadium alkoxides to
be very reactive towards hydrolysis. Alkoxy groups become harder to remove as

hydrolysis proceeds, thus a large amount of water is needed to completely remove all

22
groups [96]. A large water to metal ratio produces a highly cross-linked oxide network

[91] that has a structure close to that of the crystalline oxide [96]. However, not all
alkoxy groups are removed when hydrolysis ratios are low (11 S 3). In this case, a gel is
formed which can best be called an alkoxo oxide [VzOs.,.(OR),.]n [95]. Oligomers are
obtained from vanadium alkoxides reacted with low hydrolysis ratios (h<1) [91].

Since V5+ is a strong oxidizing agent, V205 gels always contain reduced V“ ions
[97]. This results in a mixed valence compound whose semiconducting properties arise
from electron hopping between the two valence states. Methods of further reduction of
vanadium pentoxide gels include exposure to air and dehydration as well as reaction of the
gel with the solvent [91]. Along with affecting the hydrolysis reaction, the size of the
alkoxide group attached to vanadium also affects the magnitude of reduction. Only 1% of
the V” centers were reduced to V“ when VO(OAm‘)3 was used as a precursor while
10-15% of v“ was reduced when (01%")3 was used [95).

Much of the work surrounding vanadium pentoxide xerogels (a gel formed when
the liquid is removed at ambient pressure by thermal evaporation causing shrinkage) has
involved the study of its electronic conducting properties and their dependence on
temperature and humidity [91]. Vanadium’s interesting catalytic chemistry along with the
newer Opportunities provided by sol-gel chemistry have led to an increased interest in
using vanadium pentoxide xerogels for chemical sensors.

There has been a wide use of V205 in conductimetric sensing. Inubishi et al. [98]
utilized the influence of relative humidity on the proton transfer conductivity of the oxide

to construct a humidity sensor. By doping the vanadium xerogel with various metal

23
oxides, the structure of the gel was changed. Inubishi reported that changes in structure

resulted in a change of the electronic structure, and sensor response was attributed to
these changes in electronic structure. Glezer and Lev [99] used amperometric glucose
sensors to examine the proton conducting properties of V205, while polymer composite
sensors were studied by Rushau et al [100]. Utilizing both the catalytic activity and
electronic properties of vanadium oxide, high temperature conductimetric gas sensors
have been developed for analytes such as hydrocarbons, alcohols, NO, and N02 [101,

102].

1.4 X-Ray Photoelectron Spectroscopy (XPS)

Photoelectron spectroscopy provides quantitative and qualitative information
about the surface structure and composition of a material using the photoemission of
electrons [103-105]. X-ray photoelectron spectroscopy (XPS) is one of the most
powerful surface sensitive techniques for the study of electronic structure of filled levels
of surface atoms. XPS can probe both the valence and core level electrons of a material.

Light consists of photons with energies given by the expression:

E = hv (1.32)
and solids contain a frequency threshold below which light cannot eject electrons from a
particular metallic surface. The minimum energy required to remove an electron from the

metal surface, etp, can be calculated from this threshold frequency, vc:

eq) : th (1.33)

24
where (1: denotes the work-function of the material and e is the magnitude of the electron

charge.

Above the threshold frequency, the emitted electrons will have a maximum kinetic
energy that corresponds to the difference between the incident photon and the work
function. For the maximum kinetic energy the emitted electrons must be at the Fermi
level, that is, the highest lying occupied electronic state of the metal. Photons emitted
below the maximum kinetic energy signify electrons emitted from a level having a binding
energy E3. Therefore, the kinetic energy of the ejected electron is the difference between
the energy of the incident photon and the binding energy of the core electron (Figure 1.3),

taking into consideration the work-function ew of the spectrometer:

E1. = hv - E3 - cm (1.34)

.llll
1L

   
    

 

Figure 1.3. XPS process.

25

Photoelectron spectroscopy thus allows the determination of the relative population of
electronic states of atoms.

Binding energies of electrons in molecules will depend on the environment. An
increase in the electronegativity of the ligands about a central atom causes a displacement
of the valence electrons from the atom towards its ligands. The core electrons therefore
encounter a stronger attraction since fewer electrons surround the nucleus resulting in an
increase in binding energy [105]. Two balancing factors - the oxidation state of the atom
and the Madelung potential of the lattice - cause a chemical shift that changes the binding
energy of the core electrons. The subsequent chemical shift of the core electrons from the
atomic binding energy can reveal information about the binding environment of an atom.

The XPS signal intensity varies for each particular element and has been defined as
follows [106]:

1r = lolliGD(€i)A-(€t) (135)
where 10 is the X-ray flux, n, is the number density of atoms i, 0' is the photoexcitation
probability of electrons corresponding to a particular atomic orbital of species i, D(e.-) is
the fraction of electrons detected by the analyzer, and Me.) is the mean free path of the
emitted electron.

In order to obtain quantitative information about a material, relative intensities of

the XPS features are compared. The following relationship has been derived [107]:

1]] ll o.21)(£2)A'T(82)
—=— 1.36
112 12 C"10(Er))‘r(€r) ( )

26
where mm; is the relative concentration of species 1 and 2 with 11/12 representing the

relative intensities. The remaining values of the equation are assumed to be constant and
contribute to a “sensitivity factor” particular to the material being studied. These
sensitivity factors can be derived theoretically using previously determined values for 0'
[106] and ME) [107] with D(ez)/D(el) measured experimentally or deduced from design
specifications. Values for sensitivity factors may also be derived empirically [103]. To
derive these empirical values, XPS intensities for a known material (say pure TiOz) are
obtained and used to calculate 11/12. From this value, along with 711/112 calculated from the

known stoichiometry, sensitivity factors can then be calculated.

1.5 Ultraviolet-Visible Spectroscopy (UV-Vis)

UV-Vis measurements involve the partial absorption of electromagnetic radiation
[108,109], causing a decrease in the transmitted power of electromagnetic radiation
through a sample. When absorption of a photon occurs, the energy of this photon is equal
to the energy difference between the ground state and one of the excited energy levels of
the absorbing species. Since this difference in energy is specific for a given species,
absorption spectra are useful as qualitative and quantitative measures of a substance.

During absorption, a species M is modified by the absorption of a photon, to an
excited state M* as shown below:

M + hv —9 M‘ (1.37)

27
In a short time (on the order of nanoseconds) the excited species may lose this energy,

most commonly in the form of heat:

M‘ -+ M + heat (1.38)
Relaxation can also occur through photochemical decomposition of M' to other products
or by fluorescent or phosphorescent radiation emission. The lifetime of the excited species
is generally small and its concentration is therefore negligible.

For inorganic molecules, the radiation absorbed may be the result of transitions
between filled and unfilled d orbitals. There are several factors that will affect the
differences of energy between these orbitals : 1) the position of the element in the
periodic table, 2) its oxidation state, and 3) the nature of the ligand bonded to the element
[110]. The environment surrounding the transition metal strongly influences the photon
absorption.

The electronic transitions involved with the absorption process occur in the
partially filled d orbitals of transition metals. Electrostatic repulsion between the electron
pairs of donors and the d orbital of the metal ion results in the splitting of d-orbital
energies. Energy transferred to an electron in a d orbital of lower energy to an orbital of
higher energy can be caused by the absorption of a photon.

There are five representations for the electron-density distributions or “orbitals” of
d-electrons: dxy, (1..., d)... (1,2,2 , and dzz. The first three are formed between the spatial
axes x, y, and z, and consequently, the dxy, d“, and dfl orbitals have their maximum
densities between the axes. The other two representations have their electron densities

directed on the axes.

28

For a molecule in an octahedral formation, the It donor atoms will exert a repulsion
on the orbitals increasing the energy of these orbitals causing them to become destabilized.
The repulsion for the d , dxz, and (1,, orbitals will be equal considering their axes do not
match with the bonding axes and differ only in orientation. There is a greater effect,
however, on the d12 and d.2.,2 since their charge densities lie along the axes, resulting in a
smaller increase in energy. When considering a tetrahedral structure of atoms around the
central metal ion, similar arguments to those made for the octahedral structure can be
made. However, under tetrahedral conditions, the relative increase of energies for the two
groups (dxy, du, and dyz or dzz and dx2-,2) is reversed.

The visible spectrum can also be used to measure the band gap in semiconducting
materials [111]. The electrons in a semiconductor occupy allowed energy bands resulting
from the interactions of atomic energy levels. As stated earlier, the highest occupied band
is referred to as the valence band (By), the lowest unoccupied is referred to as the
conduction band (EC), and the energy between these two is the energy (band) gap (Ea).
Absorption of incident radiant energy can excite the electrons from the valence to
conduction band. The minimum amount of energy needed for absorption to occur
indicates the band gap energy of the semiconducting material. Thus, when the energy of a

quantum (hv) equals E0, absorption of light will occur:

EG :— (1.39)

When the wavelength (AG) is measured in run, the calculated band gap is usually measured

in units of eV.

29
1.6 X-Ray Diffraction (XRD)

Since X-rays have wavelengths on the order of Angstroms they are well suited to
study the structure of materials [108]. XRD is a bulk characterization technique and can
be used for particle size estimations, determination of bulk phases, and quantification of
the extent of bulk transformations.

X-ray photons are elastically scattered by atoms in a periodic lattice in the material
being studied. The lattice spacing is derived from the diffraction of X-rays by crystal
planes using the Bragg equation [112]:

n}. = 2d sin(9) ; _ n = 1,2,... (1.40)
where it is the wavelength of the X-rays, d is the distance between two lattice planes, 9 is
the angle between the incoming X-rays and the normal to the reflecting lattice plane, and n
is called the order of reflection. The X-ray photon will only reach the detector when
equation (1.40) is satisfied. At all other instances, destructive interference prevents the X-
ray photons from being detected. Since the angle of the incoming X-rays will be the same
as the reflection angle of the X-rays with respect to the normal to the lattice plane, the
angle measured by the diffratometer is 20. By measuring the angle, 29, at which the
detected X-rays are scattered, the lattice spacing, d, of the crystal can then be calculated.
From this spacing, the corresponding compound or crystallographic phase of a particular
compound can then be determined.

XRD data can also provide information on the particle sizes in materials being

studied. Poorly crystallized materials containing large particles show broad features with

30
peaks narrowing as a more perfect crystal is obtained. Particle sizes can be determined

using full width half maximum (FWHM) measurements with Scherrer’s equation [113]:

IO.

L =
[3 cost)

 

(1.41)

where 7» is the X-ray wavelength, [3 is the peak width, 0 is the angle between the X-ray
and the normal to the surface, and K is a constant related to the crystalline shape and the
way L and B are defined. K is assumed to be ~O.9 for the Scherrer equation [114].

Particle sizes determined through the line-broadening technique should be used for
relative measures of average particle size. XRD line widths are influenced by the
distribution of particles, internal strain, and instrumental factors leading to line broadening.
Thus absolute particle sizes cannot be determined using line width measurements [115].

X-ray diffraction provides important information about the influence mixing has on
the crystalline phase of vanadium and/or titanium in sol-gel derived thin films after being
annealed. Since XRD does not detect amorphous species, the point at which crystalline
species are formed (if they are formed), as well as the occurrence of phase changes within

the crystal can be determined.

31

1.7 List of References

9’39?"

1.0.
11.
1.2.
13.
14.
15.

16.

17.
18.
19.

20.

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

Meixner, H.; Gerblinger, J.; Fleischer, M. Sens. Act. 1993 15-16, 45.

Xu, Y.; Yao, K.; Zhou, X.; Cao, Q. Sensors and Actuators, 1993, 313-14, 103.
Wu, M. T.; Yao, X.; Yuan, Z. H.; Sun, H. T.; Wu, W. C.; Chen, Q. H.; Xu, G. Y.
Sensors and Actuators, 1993, 313-14, 491.

Haseqawa, S.; Sasaki, Y. Matsuhara, S. Sensors and Actuators, 1993, 313-14,
509.

Kimer, U., Schierbaum, K. D.; Gtipel, W.; Liebold, B.; Nicoloso, N.; Weppner,
W.; Fischer, D.; Chu, W. F. Sens. Act., 1990, 31, 103.

Rao, C. N. R.; Raju, A. R.; and Vijayamohanan, K., Gas-sensor Materials in New
Materials, S.K. Joshi, C.N.R. Rao, R Rsruta and S. Nagakura (eds.)), New Delhi,
India, 1992, pp. 1-37.

Martin, C.; Rives, V.; Sanchez-Escribano, V.; Busca, G.; Lorenselli, V.; and
Ramis, 6., Surf. Sci. 1991 251/252, 825.

Tufano, V. and Turco, M., Appl. Catal. 1993 32, 9.

Turco. M.; Lisi, L.; Pirone, R.; Appl. Catal. 1994 32, 133.

Went, G. T.; Leu, L.-J., Rosin, R. R.; and Bell, Alexis T.; J. Catal. 1992 134, 492.
Busca, G.; Centi, G.; Marchetti, L.; Trifiro, F. Langmuir 1986 2, 568.

Nickl, J.; Schltigl, R.; Baiker, A.; Kndzinger, H.; Ertl, G. Catal. Lttr. 1989 3, 379.
Vejux, A. and Courtine, P.J. Solid State Chem. 1978 23, 93.

Saleh, R.Y.; Wachs, I.E.; Chan, S.S.; Chersich, C.C. J. Catal. 1986 98, 102.

Cox, P. A. The Electronic Structure and Chemistry of Solids, Oxford University
Press: Oxford, 1987.

Madou, M. J.; Morrison, 8 R. Chemical Sensing with Solid State Devices;
Academic Press: San Diego, CA, 1989.

Geistlinger, H. Sens. Act. 1993 I7, 47.

Janata, J. Principles of Chemical Sensors, Plenum Press, New York, 1989.
Morrison, S. R. Sens. Act. 1982 2, 329. Advani, G. N.; Nanis, L. Sens. Act. 1982
2, 201.

Chang, S. C. in Chemical Sensors, Analytical Chem. Symposia Series 17; Elsevier,
Amsterdam, 1983 (Fukuoka Conference); eds. T. Seiyama, K. Fueki, J. Shiokawa,
S. Suzuki, p.78.

Thorton, E. W.; Harrison, P. G. J. Chem. Soc. Faraday Trans. 1 1975, 71, 461.
Morrison, 8. R. Sens. Act. 1987 11, 283.

Morrison, 8. R. Surf. Sci. 1971 27, 586.

Cooper, R. B.; Advani, G. N.; Jordan, A. G. J. Electron. Mater. 1981 10, 455.
Reuter, H. Adv. Mater. 1991 3, 569.

Reuter, H. Adv. Mater. 1991 3, 258.

Ulrich, D. R. J. Non Cryst. Sol. 1988 100, 174.

Livage, J.; Henry, M.; Sanchez, C. Frog. Solid St. Chem. 1988 I8, 259.
Kozhukharov, V.; Trapalis, CH.; Samuneva, B.; Kirilova, E. J. Mater. Sci. Lett.
1992 I I , 1206.

30.
31.

32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.

43.

45.

46.

47.

48.

49.

50.
51.

52.
53.

54.
55.

56.
57.

32

Grammatico, J. P.; Porto Lopez, J. M. J. Mater. Sci. Mater. Elect. 1992, 3, 82.
Mosset, A.; Gautier-Luneau, 1.; Galy, J.; Strehlow, P.; Schmidt, H. J. Non-Cryst.
Solids 1988, 100, 339.

Greismar, P.; Papin, G.; Sanchez, C.; Livage, J. J. Mater. Sci. Lett. 1990, 1288.
Inubishi, A.; Masuda, S.; Okubo, M.; Matsumoto, A.; Sadamura, H.; Suzuki, K. in
High Tech Ceramics, P. Vincenzini, Ed., Elsevier Science Publishers, Amsterdam,
1987, 217.

Yoshimura, N.; Sato, S.; Itoi, M.; Taguchi, H. Sozai Busseigaku Zasshi 1990, 3,
47.

Levy, D. J. Non Cryst. Solids 1992, 14762148, 508.

Lev, 0. Analysis 1992, 20, 543.

Austin, I.G.; Mott, N.F. Adv. Phys, 1969, 18, 41.

Bullot, J.; Gallais, O.; Gauthier, M.; Livage, J. Appl. Phys. Lett., 1980, 36, 986.
Bullot, J.; Gallais, O.; Gauthier, M.; Livage, J. Phys. Stat. Sol. A, 1982, 71 , K1.
Chemseddine, A.; Morineau, R.; Livage, J. Solid State Ionics, 1983, 9-10, 357.
Barboux, P.; Baffler, N.; Morineau, R.; Livage, J. Solid State Ionics, 1983, 9-10,
1073.

Aldebert, P.; Baffler, N.; Legendre, J.J.; Livage, J. Rev. Chim. Min., 1982, 19,
485.

Brinker, C.J.; Clark, D.E.; Ulrich, D.R., Eds.; Mat. Res. Soc. Symp. Proc. 32,
Materials Research Society, Pittsburgh, 1984.

Brinker, C.J.; Clark, BE; Ulrich, D.R., Eds.; Mat. Res. Soc. Symp. Proc. 73,
Materials Research Society, Pittsburgh, 1986.

Brinker, C.J.; Clark, D.E.; Ulrich, D.R., Eds.; Mat. Res. Soc. Symp. Proc. 121,
Materials Research Society, Pittsburgh, 1988.

Zelinski, B.J.J.; Brinker, C.J.; Clark, BE; Ulrich, D.R.,Eds.; Mat. Res. Soc. Symp.
Proc. 180, Materials Research Society, Pittsburgh, 1990.

Hampden-Smith, M.J.; Klemperer,W.G.; Brinker,C.J., Eds; Mat. Res. Soc. Symp.
Proc. 271, Materials Research Society, Pittsburgh, 1992.

Bradley, D.C., Mehrotra, R.C.; Gaur, D.P. Metal Alkoxides, Academic Press,
London, 1978.

Zheng, H.; Colby, M.W.; Mackenzie, J.D. Mater. Res. Soc. Symp. Proc. 1988,
121 , 537.

Guglielmi, M.; Carturan, G. J. Non-Cryst. Solids 1988, 100, 16.

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

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

Livage, J.; Henry, M. in Ultrastructure Processing of Advanced Ceramics,
Mackenzie, J.D.; Ulrich, D.R., Eds., Wiley, New York, 1988, 183.

Mehrotra, R.C.; Gaur, D.P.; Bohra, R. Metal B-diketonates and Allied
Derivatives, Academic Press, London, 1978.

Mehrotra, R.C.; Bohra, R. Metal Carboxylates, Academic Press, London, 1983.
Doeuff, S.; Henry, M.; Sanchez, C.; Livage, J. J. Non-Cryst. Solids, 1987, 89, 206.
Atik, M.; Aegerter, M.A. J. Non-Cryst. Solids, 1992, 14732148, 813.

58.

59.
60.
61.
62.
63.

65.

66.

67.
68.

69.
70.

71.

72.

73.
74.
75.
76.
77.
78.

79.
80.

81.
82.
83.
84.

33

Chaumont, D.; Craievich, A.; Zarzycki, J. J. Non-Cryst. Solids, 1992, 14732148,
127.

Debsikar, J.C. J. Non-Cryst. Solids, 1986, 87, 343.

Kundu, D.; Ganguli, D. J. Mat. Sci. Lett., 1986, 5, 293.

Doeuff, 8.; Henry, M.; Sanchez, C.; Livage, J. J. Non-Cryst. Solids, 1987, 89, 84.
Leaustic, A.; Riman, R.E. J. Non-Cryst. Solids, 1991, I35, 259.

Gagliardi, C.D.; Dunuwila, D.; Berglund, K.A. in Mat. Res. Soc. Symp. Proc. 180;
Zelinski, B.J.J.; Brinker, C.J.; Clark, D.E.; Ulrich, D.R.,Eds.; Materials Research
Society, Pittsburgh, 1990, 801.

Gagliardi, C.D.; Dunuwila, D.; Van Vlierberge-Torgerson, B.A.; Berglund, RA.
in Mat. Res. Soc. Symp. Proc. 271, Hampden-Smith, M.J.; Klemperer,W.G.;
Brinker,C.J., Eds.; Materials Research Society, Pittsburgh, 1992, 257.

Van Vlierberge-Torgerson, B.A.; Dulebohn, J.I.; Berglund, K.A. in Chemical
Processes of Advanced Materials, Hench, L.L.; West, J.K., Eds.; John Wiley &
Sons, Inc., 1992, 415.

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

Yoldas, B. E. J. Mat. Sci. 1986, 21, 1087.

Mehrotra, R.C. in Recent Advances in Inorganic and Radiochemistry, Sharpe,
A.G.; Emeleus, H.J., Eds., Vol. XXVI (Academic Press, New York), 269.
Melpoder. S. M.; Cotrain, B. K. Mater. Res. Soc. Symp. Proc. 1988, 121 , 811.
Babonneau, F.; Leaustic, A.; Livage, J. Mater. Res. Soc. Symp. Proc. 1988, 121,
317.

Schmidt, H.; Rinn, G.; NaB, R.; Sporn, D. Mater. Res. Soc. Symp. Proc. 1988,
121, 743.

Sanchez, C.; Babonneau, F.; Doeuff, 8.; Leaustic, A. in Ultrastructure Processing
of Ceramics, Glasses, and Composites; Mackenzie, J. D., Ulrich, D. R., Eds.; John
Wiley & Sons; New York (1988), p.77.

Schmidt, H.; Seiferling, B. in Mat. Res. Soc. Symp. Proc., 1986, 73, 739.

Sanchez, C; Livage, J. New J. Chem. 1990, 14, 513.

Gagliardi, C. D.; Berglund, K. A. Mat. Res. Soc. Symp. Proc. 1989, 155, 127.
Teller, H. L. Sens. and Act., 1983, 4, 679.

Ichinose, N. Am. Ceram. Soc. Bull., 1985, 64, 1581.

Prassas, M.; Hench, L. L. in Ultrastructure Processing of Ceramics, Glasses and
Composites, L. L. Hench and D. R.Ulrich Eds., Wiley, New York, 1984, 100.
Livage, J. Mat. Res. Soc. Symp. Proc., 1986, 73, 717.

Doeuff, 8.; Henry, M.; Sanchez, C. in Better Ceramics Through Chemistry II, C.J.
Brinker, D.E. Clark and D. R. Ulrich Eds., Mater. Res. Soc. Symp. Proc., 73,
MRS, Pittsburgh, 1986, 653.

Komareni, C.; Roy, R.; Breval, E. J. Am. Ceram. Soc., 1985, 68, C41.

Severin, K. G.; Ledford, J. S. manuscript in preparation.

Hoffman, L.W., Ledford, J .S. manuscript in preparation.

deel, W.; Kimer, U.; Wiemhtifer, H. D.; Rocker, G. Solid State Ionics, 1988, 28-
30, 1423.

85.

86.
87.
88.

89.
90.

91.
92.
93.
94.
95.

96.

97.
98.

99

100.
101.
102.
103.
104.
105.
106.

107.
108.

109

110

111.

112

113.

34

Hamasaki, Y.; Ohkubo, 8.; Murakami, K.; Sei, H.; Nogami, G. J. Electrochem.
Soc., 1994, 141(3), 660.

Raju, A. R.; Rao, C. N. R. Sensors and Actuators, 1993, 321, 23.

Boccuzzi, F.; Guglieliminotti, E. Sensors and Actuators, 1993, 321, 27.

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

Livage, J.; Holivet, J. P.; Tronc, E. J. Non-Cryst. Solids, 1990, 121, 35.

Livage, J.; Barboux, P.; Badot, J. C.; Baffier, N. in Better Ceramics Through
Chemistry 111; Brinker, J. C.; Clark, D. E.; Ulrich, D. R. Eds.; Mat. Res. Soc.
Symp. Proc., 1988, 121, 167.

Livage, J.; Chem. Mater., 1991, 3, 578.

Pomonis, P. J.; Vickerman, J. C. Faraday Discuss. 1981, 72, 247.

Carrizosa, 1.; Muneura, G. J. Catal. 1977, 49, 189.

Prandtl, W.; Hess, L. Anorg. Allg. Chem, 1958, 296, 36.

Sanchez, C.; Nabavi, M.; Taulelle, F. in Better Ceramics Through Chemistry III;
Brinker, J.C.; Clark, BE; Ulrich, D.R. Eds.; Mat. Res. Soc. Symp. Proc., 1988,
121, 93.

Nabavi, M.; Sanchez, C.; Livage, J.; Eur. J. Solid State Inorg. Chem, 1991, 28,
1173.

Livage, J. Mater. Res. Bull. 1991 26(11), 1173.

Inubishi, A.; Masuda, S.; Okubo, M.; Matsumoto, A.; Sadamura, H.; Suzuki, K. in
High Tech Ceramics; Vincenzini, P. Ed., Elsevier: Amsterdam, 1987; 217.

Glezer, V.; Lev, O. J. Am. Chem. Soc. 1993, 115, 2533.

Rushau, G. R.; Newnham, R. E.; Runt, J.; Smith, B. E.; Sensors and Actuators
1989, 20, 269.

Rao, C. N. R. Talanta 1992, 115, 2533.

Rao, N.; van den Bleek, C. M.; Schoonman, J. Solid State Ionics 1993, 59, 263.
Briggs, D.; Seah, M. P. Practical Surface Analysis, 2"‘1 ed. Wiley: New York,
1990.

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

Werthiem, G. K. in Solid State Chemistry Techniques, Cheetham, A. K.; Day, P.;
eds.; Claredon Press; Oxford, 1987.

Scofield, J. H. J. Electron Spectrosc. Relat. Phenom. 1976, 8, 129.

Penn, D. R. J. Electron. Spectrosc. Relat. Phenom. 1976 9, 29.

Skoog, 1D. A.; West, D. M.; Holler, F. J. Analytical Chemistry: An Introduction,
4'” ed., Saunders College Publishing, Orlando, FL, 1990.

Drago, R. S. Physical Methods for Chemists, 2“d ed., Saunders College Publishing,
Orlando, FL, 1992.

Niemantsverdriet, J. W. Spectroscopy in Catalysis, VCH Publishers, New York,
1993.

Ibanez, J. G.; Solorza, O.; Gomez del-Campo, E. J. Chem. Ed. 1991 68(10), 872.
Bragg, W. L. Proc. Comb. Phil. Soc., 1912, 17, 43,

Scherrer, P. Gdtt. Nachr. 1918 2, 98.

35

114. Klug, H. P.; Alexander, L. E. X-Ray Diffraction Procedures for Polycrystalline
and Amorphous Materials, 1‘ Ed. Wiley: New York, 1954.
115. Cohen, J. B. Ultramicroscopy 1990 34, 41.

Chapter 2

Effects of Annealing Temperature on the Structure of
Vanadium Oxide/Titanium Oxide Thin Films Derived from
Metal Carboxylate Sol-Gel Processing

2.1 Introduction

The semiconducting properties of titanium dioxide have been widely studied. A
large portion of current research involving TiOz is being directed towards the development
of chemical sensors based on its semiconducting properties [1-3]. An important step in
the development of chemical sensors is to study the active phase of the Vfl‘ i oxide system
with respect to its semiconducting properties. Before sensing properties can be studied,
suitable materials must be developed.

This chapter presents the first of a two-part study conducted to investigate the
influence of vanadium on the structure of titanium oxide thin films. Titanium dioxide films
d0ped with 10 mole % vanadium were prepared via the sol-gel technique. Preparation
involved the hydrolysis of metal alkoxides using excess valeric acid at room temperature.
The structures of these films were changed by varying the temperature at which the films
were annealed. Surface and bulk film structures were studied using XPS, XRD, FTIR and
UV-Vis spectroscopies. Results obtained by XRD agree with data presented in earlier
studies indicating that the presence of vanadium promotes the crystallization of titanium

oxide [4]. No crystallization of vanadium was found at elevated temperatures, as evident

36

37
through XRD patterns and FTIR spectra. Results obtained by XPS indicate that
temperature has an influence on the amount as well as the phase of vanadium present in
the surface of these films. UV-Vis findings support XPS findings, as well as provide
information on the coordination of vanadium atoms in the films. UV-Vis data will also
show that annealing temperature significantly affects the band gap of TiOz in the films.

The properties of VIP i catalysts prepared by impregnating the TiOz support with
vanadia species have been studied [5]. Many of the recent studies using V/I‘ i catalysts
have involved the oxidation of hydrocarbons [6,7] and the reduction of nitric oxide by
ammonia [8-10]. The activity of these catalysts has been attributed to the surface phase of
the vanadia species. High surface area and controlled porosity (which are hard to control
with traditional catalyst preparation methods) influence the quality of the sensor. Sol-gel
technology offers more control over these properties of the resulting material, giving
materials prepared via sol-gel chemistry advantages over materials produced through
traditional methods. In addition to controlled surface area and porosity, the electronic
properties can be enhanced through the mixing of metal oxides [11] or by annealing the
oxide materials at various temperatures [12]. Thus, sol-gel methods using titanium and
vanadium alkoxides to produce mixed metal oxides may prove to be advantageous for the
preparation of chemical sensing materials [13-15].

The oxide network obtained through sol-gel preparations is attractive since
homogeneous multi-component systems can be achieved through the mixing of precursors
[17]. Temperatures for processing can be low, making chemical modifications easier and

low processing temperatures can lead to a wide range of new properties [1 6]. Therefore,

38

much attention has been directed towards understanding the hydrolysis and condensation
of silicon and transition metal alkoxides [17,18]. Modifying the alkoxides with carboxylic
acids before hydrolysis increases the quality of the films produced from alkoxide
precursors. Slower condensation rates are obtained by carboxylic acid complexation of
the metal alkoxide precursors. The slower rates are important for producing polymeric
gels as opposed to solid oxide precipitates [16]. Valerie acid modified alkoxides produce
sols from which films can be cast that are uniform, transparent, and that adhere well to
quartz substrates [19,20].

It has been proven that both the temperature at which these systems are treated as
well as the amount of vanadium present affect the final structure of the product. Machej
et al. [21] have reported that the metal oxide-oxide surface interaction of V/T i catalysts is
affected by different annealing temperatures. The structure and composition of WT i
catalysts have been studied previously as a function of temperature [4]. These authors
found that the structural transformation of the T102 support from anatase to rutile and the
morphology of the supported vanadium phase were functions of temperature. The
impregnation of the support with the vanadium salt results in the formation of vanadyl
oxalate at low temperatures. Raising the calcination temperature initially leads to the
presence of V205 crystals before the supported vanadia phase reacts with the
TiOa(anatase) to produce a VxTil-,.Oz(rutile) solid solution.

To date, work involving vanadium oxide/titanium oxide thin films prepared though
the sol-gel process has not appeared in literature. The goal of the research presented in

this chapter was to study the effect of annealing temperature on the surface and bulk

39
phases of vanadium oxide/titanium oxide thin films produced via sol-gel technology. In
the following chapter, the effect vanadium loading has on film structure is presented.
Results from these two studies will be used to determine the feasibility of sol-gel produced

vanadium oxide/titanium oxide thin films for use as conductimetric sensors.

2.2 Experimental

Materials. Titanium (IV) isopropoxide (97%, Aldrich Chemical Co.), valeric acid
(99+%, Aldrich Chemical Co.), and vanadium triisopropoxide oxide (Gelest Inc.) were
used without further purification. Distilled, deionized water was used for all preparations.

Solution Preparation. The method used to prepare the films was adapted from
one previously deve10ped for titanium compounds [20]. Since both metal precursors are
easily hydrolyzed, solutions were prepared in a glovebox under continuous N2 purge. In
the glovebox, the reactions were carried out at room temperature in capped vials. The
sols were prepared such that there was 10 mole % vanadium in titanium. The
metal:water:acid ratio was 1:1.5z9. Vanadium triisopropoxide was added to titanium (IV)
isopropoxide followed by the addition of valeric acid and then water. The solution was
vigorously shaken with a vortex mixer following the addition of each reactant. Solutions
were prepared in duplicate.

When the two metal precursors were initially mixed, no apparent change was

observed and the solution remained colorless. When the acid was added, an immediate

40
color change to bright yellow was observed. Finally, when water was added, there was a
further color change to a dark yellow-orange.

Film Preparation. Solutions were deposited on a quartz substrate, spun for a
minimum of 5 minutes and air-dried overnight. Next, the films were dried for 6 hours at
100°C in a tube furnace under the flow of dry air (100 cc/min, medical grade, AGA Gas).
Each film was then annealed at the desired temperature for 24 hours under the same
conditions (dry air, 100 cc/min). Film preparations from each solution were carried out in
triplicate. An as-cast film, meaning unheated, from each solution was analyzed in order to
confirm solution/film reproducibility.

Reference films were prepared containing vanadium oxide and titanium oxide
separately. These films were prepared in a manner similar to those containing both
vanadium and titanium. The film containing vanadium oxide was annealed at 500°C for 4
hours. The titanium oxide films were annealed at 600°C to obtain films containing crystals
primarily in the anatase phase and at 1000°C to obtain a mixture of anatase and rutile
crystals, although the films still remained highly amorphous. The titanium films were also
annealed for 4 hours.

Fourier Transform Infrared Spectroscopy (FT IR). FTIR spectra were collected
from films that were spin coated on KBr windows, using a Mattson Instruments Galaxy
3020 Fourier transform infrared spectrometer. Data were obtained in mid-IR region
(4000 - 400 cm") with a resolution of 2 cm". Data acquisition was performed using an

Enhanced First software package.

41

X-ray Photoelectron Spectroscopy (XPS) Films cast on quartz slides were
analyzed with a VG Microtech spectrometer using a Clam2 hemispherical analyzer. Data
were collected with a Mg anode operated at 15 kV and 20 mA (300W) and an analyzer
pass energy of 50 eV. The core regions for four elements were scanned with each film: C
ls, O ls, Ti 2pm and V 2pm. Binding energy charge corrections for films containing T1
were made with reference to the Ti 2pm peak (458.5 eV [22]). Corrections for reference
films of V205 were made with reference to C 1s (284.6 eV). The binding energies given
have a precision of i 0.1 eV. Quantitative measurements were performed using sensitivity
factors determined previously [23]. For quantitative determination of the V 2pm region,
the contribution from the 0 Is satellite was subtracted out from the area attributed to V
2pm before calculations were made. Curve fitting was performed using a 20%
Lorentzian-Gaussian mix Voigt function [24].

X -ray Diffraction (XRD). All films were analyzed with a Rigaku XRD
diffractometer using Cu K01 radiation (1. = 1.5418 A). The X-ray tube was operated at 45
kV and 100 mA. The diffractometer was scanned at a rate of 0.07 deg/min (deg = 20).
Divergence and scattering slits were both set at 1°. Since quartz shows a broad
absorbance contributing to data at the lower end of the scanned region, information for a
blank quartz slide was obtained and subtracted from the XRD data collected for each film.

Ultraviolet-visible spectroscopy (UV-Vis). A Unicam UV-Vis spectrometer was
used to collect UV-Vis spectra. The incident beam was normal to the surface of the film
using a blank quartz slide as a reference. Data were collected using Vision software in

absorbance mode with a bandwidth of 2 nm and a data interval of 1 nm.

42

2.3 Results and Discussion

FT IR. Normalized spectra representative of the VIP i and Ti films calcined at
400°C are shown in Figure 2.1. No features attributable to valerate ligands are present
indicating they have been removed during calcination. The titanium oxide thin film
spectrum (Figure 1b) contains a major peak at 448 cm'l that is attributed to a titanium-
oxygen stretch. The spectrum recorded for the VIP i film (Figure 1b) also contains the
band at 447 cm'1 due to the titanium-oxygen stretch. The feature at 944 cm’1 is indicative
of an absorbance from a partially hydrated surface V=O. The broad absorption centered
at 796 cm" is attributed to my stretching modes of the v-0-v bonds [25]. The
shoulders at 890 cm'1 for the V=O stretch and 715 cm'1 for the VO-V stretch suggest the
presence of a reduced vanadium species. Using data reported Nyquist and Kagel [26], the
features occurring at 796 and 890 cm’1 can be attributed to oxides of V5+ while the
shoulders on both peaks appearing at a slightly higher frequency can be attributed to
oxides of V‘”. Thus, IR data indicates the presence of more than one oxidation state for
vanadium with V5+ being more abundant.

V205 crystals have two characteristic absorbances at 1020 and 940 cm", attributed
to the vV=O of crystalline V205 and the symmetric stretch of V,,Oy clusters, respectively
[27]. Since no feature at 1020 cm'1 is observed, our FTIR data suggests that no
crystalline V205 is present in the film. Busca et al. [27] report that surface vanadium will
be amorphous at low loadings, and amorphous and crystalline for higher loadings. Wachs

[28] showed that crystalline V205 IR absorption always occurs at 1020 cm'1 and surface

Aboorbance

 

 

43

 

 

 

 

 

1000 000 800 700 800 500 400

Frequency (cntU

Figure 2.1. FTIR spectra of (a) V/l' i film annealed at 400°C and (b) Ti film annealed at

400°C.

44

vanadium oxide species will adsorb between 800-995 cm'1 for a hydrated surface. Thus,
our results suggest only a partially hydrated surface vanadium species is formed without
crystalline vanadium oxide.

XRD. XRD patterns of the V/I’ i films are shown in figure 2.2. The range selected
(20 = 24.00 - 29.00) shows the most intense peaks for both TiOz anatase (20 = 25.30) and
rutile (20 = 27.45). Intense peaks for penta- and tetra-valent vanadium oxides also occur
in this region: 20 = 26.21 for V205 and 27.88 for V204. Background due to the quartz
substrate has been subtracted from each of the patterns. The as-cast film (not shown) and
the film annealed at 300°C (2.2a) were featureless. At 400°C (2.2b) a broad feature
centered at 20 =25.3 is present. After annealing the film at 500°C (Figure 2.2c), the
feature at 20 = 25.3 narrows and a new feature at 20 = 27.5 can be observed. Finally at
600°C (2.2d), the feature at 20:25.3 continues to narrow and the feature at 20:27.5
increases significantly.

At the lowest temperatures, figure 2.2a, the absence of any feature in the XRD
patterns suggests that the film is amorphous. The feature in the 400°C calcined film at 20
= 25.3 (2.2b) suggests the formation of TiOz(anatase) crystals while features due to both
anatase (20 = 25.3) and rutile (20 = 27.5) appear when the film is annealed at 500°C
(2.2c). The relative concentration of TiOz(rutile)/TiOz(anatase) increases as the annealing
temperature is further increased to 600°C (2.2d). Information provided by XRD has
shown the formation and phase transformation of TiOz crystals in our films. Our data do

not show formation of V205 crystals.

Intensity

 

 

 

24 25 26 27 28 29
Degrees (29)

Figure 2.2. XRD patterns measured for V/Ti films annealed at (a) 300°C, (b) 400°C, (c)
500°C, and (d) 600°C.

46

As observed by the inset in figure 2.2, narrowing of the TiOz(anatase) XRD
feature occurs with increased annealing temperature. The narrowing of the anatase
feature suggests that the particle size of this phase grows with temperature. The particle
sizes for each of the films, as determined through the Scherrer equation [29], are
presented in table 2.1. The mean particle size of Ti02(anatase) significantly increases from
143.4 A after initially being detected at 400°C to 545.0 A when the film is annealed at
600°C. Although the relative intensity of the TiOg(rutile) feature significantly increases
with temperature (figs. 2.2c and 2.2d), the mean particle size remains constant at

approximately 542 A.

Table 2.1. Particle size of TiOz crystals in V/Ii thin films. Calculations were made using

 

 

 

FWHM of respective XRD feature.
anatase rutile
temp peak pos. FWHM size (A) peak pos. FWHM size (A)
as cast --- --- --- --- --- ---
300 C --- --- --- --- -—-
400 C 25.4 0.578 143.4 --- --- ---
500 C 25.3 0.184 451.3 27.5 0.154 542.8
600 C 25.4 0.153 545.0 27.6 0.153 542.4

 

 

 

 

The data collected by XRD shows the formation of TiO2 crystals, which agrees
with the information obtained from infrared spectroscopy. XRD also provides information
about the resulting phase of T102. The predominance of the peak at 20 = 27 for films
calcined at 600°C (spectrum (1) suggests a significant formation of Ti02(rutile) as

compared to that of TiOg(anatase). Saleh et al. [4] studied 7% VII i catalysts calcined

47

over the temperature range 110-750°C. In their work, the vanadia species were
introduced through wet impregnation to Ti02(anatase). A significant amount of the
anatase phase was transformed to rutile at 700°C. For samples heated between 350 and
650°C, Saleh et al., reported the appearance of a very weak peak which they attributed to
crystalline V205. Busca et al. [27] did not observe crystalline V205 in their XRD patterns
collected for a 9% w/w sample calcined at 720°C. However, Busca reported the presence
of a V205 peak in an XRD pattern when increasing the V205 loading to 15% w/w. Earlier
studies performed on sol-gel derived TiOz thin films [30] indicate the phase
transformations from amorphous to anatase and anatase to rutile occur at temperatures of
600° and 1000°C, respectively. In comparison, our data show the presence of vanadia
species influences the crystallization and phase transformation of titanium.

Wachs et al. [31] reported that crystalline V205 is formed only after the complete
formation of a monolayer over the Ti02(anatase) support. Saleh et al. [4] took this
further by stating that crystalline V205 initiates the formation of V,T“r.-x02 (rutile) from the
supported vanadia phase on TiOz(anatase). In our studies of thin films derived from sol-
gel chemistry, we have shown that the initial transfonnation of amorphous Ti to anatase,
and then further transformation of anatase to rutile at a lowered temperature did not
require the presence of crystalline V205.

UV-Vis. For reference, figure 2.3 contains UV-Vis data collected from highly
amorphous TiO; thin films containing anatase (2.3a) or anatase and rutile (2.3b) crystal
structures. Determination of the crystal phase present in each film was obtained from

XRD data collected on these films. The broad absorbance below 350 nm in both 2.3a and

48

Absorbance

 

 

 

 

 

l I l
200 400 600
Wavelength (nm)

Figure 2.3. UV-Vis spectra collected for primarily amorphous thin films containing the
following crystalline phases (a) T i0; (anatase), (b) T102 (anatase) and TiOz

(rutile), and (c) V205.

49

2.3b can be attributed to an electronic transition from the valence to conduction band of
T102 [32]. Conduction band edges were determined for TiOz anatase and rutile to be 3.57
eV and 3.36 eV, respectively. A band gap of 3.44-3.51 eV has been reported for a fully
crystallized rutile thin film [33]. The band gap for the material in the present study is
slightly less than that of rutile indicating a poorly crystallized or highly amorphous thin
film [34,35]. Figure 2.3 also contains the spectrum from a film primarily composed of
V205 (2.3c). The absorbances observed for the vanadium oxide film are a result of
charge transfer from vanadium to terminal oxygen [33]. Schraml-Marth et al. [36] report
tetrahedrally coordinated vanadium(V) has a charge transfer transition in the 285-335 nm
region while the charge transfer transition for octahedrally coordinated vanadium(V)
occurs in the 400-500 nm region. The differences in energies in our spectrum of vanadium
oxide suggest differences in the coordination of vanadium, where absorbances at 400 and
265 nm represent vanadium in its +5 oxidation state with octahedral and tetrahedral
coordinations, respectively.

Figure 2.4 shows the UV-Vis data collected for films calcined at the indicated
temperatures. The spectra for both samples calcined at 300°C and 400°C are similar. A
strong absorbance due to T102 is observed below 350 nm and the shoulder at 390 nm is
attributed to octahedrally coordinated vanadium(V). A broad, weak absorption in this
spectrum occurs between 478 and 900 nm. Schraml-Marth et al. [36] have suggested that
octahedrally coordinated vanadium(IV) absorbs with a broad transition between 625 and
770 nm. Flynn and Pope [37] attributed a peak near 1100 nm to an “intervalence charge

transfer transition” for a mixed valence compound. Thus, we can attribute this broad peak

Absorbance

50

 

 

-

 

J j l l I l
200 400 600 800

 

Wavelength (nm)

Figure 2.4. UV-Vis spectra obtained for VIP i films annealed at (a) 300°C, (b) 400°C, (c)

500°C, and ((1) 600°C.

51

at the lower energy to vanadium(IV) in an octahedral environment as well as charge
transfer between V(IV) and V(V). The band edge of T102 forboth the films calcined at
300°C and 400°C is observed at 3.54 eV, which is similar to that of Ti02(anatase). The
film calcined at 500°C shows an increased absorption due to octahedrally coordinated
V(IV). In addition, the band edge for T10; is shifted to 3.41 eV, which indicates a shift
towards the band edge indicative of TiOg(rutile). The film calcined at 600°C (Figure 2.4d)
shows a dramatic increase in the absorption due to V(IV) in an octahedral environment.
This is accompanied by a decrease in the size of the feature between 360 and 440 nm.
Although the absorption between 360 and 250 nm due to T10; is still present, a collapse
of the band edge is observed for the 600°C films that can be attributed to a formation of
the solid solution VxTil-,,Og.

XPS. XPS findings have been summarized in table 2.2. A small feature due to
adventitious carbon was observed in all spectra at 284.6 eV. The 0 1s spectra measured

for the films consists of an intense feature due to backbone oxygen (530.0 eV) while a

Table 2.2. Binding energies of V 2pm features as a function of calcination temperature.
V205 refers to data collected from a vanadium oxide standard.

 

 

 

Sample Binding Energy V 2p3/2 FW HM V 2p3/2
V/I‘ i ascast 515.1 2.435
3009C 516.4 3.122
4009C 516.6 2.718
5009C 517.0 2.668
6009C 517.0 2.846
V205 517.1

 

 

52

shoulder appears at ~2 eV higher binding energy indicating a small amount of water
adsorbed on the surface.

The affect of temperature on the V 2pm binding energy is presented in figure 2.5.
Along with the spectra obtained from films calcined at various temperatures, a spectrum
showing the V 2pm feature from an as-cast film (Figure 2.5a) as well as the spectrum
collected from a V205 standard (Figure 250 have been included. The V 2pm binding
energy of 517.1 eV observed in our V205 standard corresponds to the generally accepted
value for vanadium in its +5 oxidation state [38]. The XPS spectrum collected from the
valerate films shows a V 2pm centered at 515.2 eV indicating the presence of a reduced
vanadium species [38]. XPS data show the V 2pm feature shifting to higher binding
energies as the annealing temperature is increased to 500°C, suggesting that the surface
vanadium species is being oxidized. At 600°C, however, a shift to lower binding energy
of the vanadium feature suggests some of the fully oxidized vanadium is now being
reduced.

The FWHM of the V 2pm feature was also studied as a function of temperature.
The results indicate an initial increase of the linewidth from 2.44 eV in as-cast samples to
3.12 in samples calcined at 300°C. Further annealing resulted in a decreased FWHM for
films calcined at 400 and 500°C while films calcined at 600°C showed a slight increase in
FWHM over that from 500°C calcined films.

The V 2pm feature from as-cast films is primarily due to the presence of both V4+
and V3+ [38]. The relative ratio of V4+ to V”, calculated from the relative areas of the

fitted peaks, is approximately 2:1. The shift of the V 2pm peak to higher binding energies

53

 

Nomialized Intensity

 

1 1 1 11 1 II 1 1 J
520 518 516 514 512

 

Binding Energy (eV)

Figure 2.5. Vanadium 2pm photoelectron spectra collected for VII" i films (a) ascast, films
annealed at (b) 300°C, (c) 400°C, (d) 500°C, and (e) 600°C and a reference
spectra of V205 (t).

54

and the subsequent (initial) increase in FWHM upon initial calcination of the thin films
suggests the vanadium species are primarily in their 4+ and 5+ oxidation states. Further
shift of the V 2pm to an even higher binding energy with a continuous decrease in
linewidth (with the exception of the 600°C calcined film) suggests that the V in the films
continue to oxidize to the most stable state of V'”. At 600°C, the increase in the FWHM
of V 2pm feature indicates an increased presence of V” in the surface of the film To our
knowledge, results indicating the V 2pm binding energy shifting as a function of
temperature have not appeared previously in literature, although the shifting of V 2pm
binding energy as a function of loading has been studied [39]. This additional presence of
V"+ at high temperatures could support an explanation of the formation of a solid solution
of VxTrl-,.Oz(rutile).

The V/T i surface ratio obtained from XPS analysis of samples heated between 300
and 600°C is presented in figure 2.6. The ratio of 0.1 from samples heated to 300°C
doubled to 0.2 when the sample was heated to 400°C. The V/T i ratio doubled again when
the sample was heated to 500°C (0.4). Finally, a significant increase in the ratio to 1.1 is
observed when the films were calcined at 600°C. There are two situations associated with
such a growth of the surface V/T i ratio. First, the crystal structure of TiOz will influence
the surface ratio of V/I‘ i in the film. Growth of TiOz particles will lead to phase
separation. Vanadium is known to spread well on the TiOz surface, coating the outside of
the crystals. The XRD data presented earlier showed the films were completely

amorphous at 300°C, leaving the V species well dispersed in the amorphous TiOz (XRD

also showed no formation of crystalline T102). At 400°C, XRD showed the formation of

55

 

1.20

1.00 ‘

0.80 ‘

0.60 ‘

V/T i Atomic Ratio

0.40 a

0.20 ‘

 

 

 

 

0.00 i i
300 400 500 60C

Calcination Temperature (9C)

Figure 2.6. Variation in XPS determined Vfl‘ i atomic ratio as a function of annealing
temperature.

56

small amount of TiOg(anatase), making it likely that the V species was dispersed on these
crystals. The formation of the larger TiOz(rutile) phase at 500°C makes it possible for the
V species to spread, further increasing the value of V/I‘i as the V ‘coats’ the TiOz
particles. The dramatic increase in the VIP i ratio observed from the 600°C calcined film
(1.1) compared to that of 500°C film (0.4) correlates well with the information obtained
through XRD. The significant increase of Ti02(rutile) phase at 600°C allows the
amorphous V species to spread well on the surface of the TiOz crystals thus significantly
reducing the amount of Ti observed by XPS.

The second consideration involves the oxidation state of vanadium. As discussed
above, the increase in annealing temperature uSed to treat the thin films causes oxidation
of the vanadium species. The increased oxidation occurs in parallel with the migration of
the vanadium species to the surface of the film. Saleh et al. [4] showed a trend for the
V/T i ratio increasing over the temperature range 110-650°C. A similar trend for
temperatures between 300 and 600°C was also apparent in our samples. However, the

change for this ratio was more than double in magnitude as compared to results presented

by Saleh et al.
2.4 Conclusions
Bulk and surface characterization techniques have allowed us to investigate the

structure of V/Ti mixed oxide thin films as a function of temperature. Temperature affects

both the titanium structure and the vanadium phase within the films. We have shown the

57

initial transformation of amorphous TiOz to TiOz(anatase) and further transformation from
the anatase to rutile structure occurring at lower temperatures when vanadium was
introduced. Introduction of vanadium resulted in a shift of the TiOz band gap at low
temperatures and a collapse of the band gap at higher temperatures. Through the
annealing process, the surface vanadium was oxidized through V(IV) to primarily V(V).
The V/T i ratio on the surface of these films also increased with increased annealing
temperature, indicating the migration of the vanadium species to the surface at higher
temperatures. XRD and IR indicate the migration with increased temperaturedoes not
coincide with the formation of V205 crystallites . UV-Vis along with XPS data indicate
that the bulk of the vanadium coexists in the oxidation states V"+ and V”. Our studies
indicate the matrix VxTrl-.02(rutile) can be obtained without the presence of crystalline
V205 when starting with metal alkoxide precursors.

The results of our work are similar to those already presented. However, with the
use of sol-gel technology, films can be obtained at lower temperatures with phase
transformations being carried out to a greater extent - giving better control over the entire

process. Also, the mechanism by which phase transformation occurs has been altered.

58

2.5 List of References

1. Satade, K.; Katayama, A.; Ohkoshi, H.; Nakahara, T.; Takeuchi, T. Sens. Act. 3
1994 20, 111.

2. Bocuzzi, F.; Guglielminotti, E. Sens. Act. 3 1994 21, 27.

3. Shimizu, Y.; Kazushi, F; Takao, Y. Egashira, M. Sens. Act. 3 1993 13-14, 623.

4. Saleh, R. Y.; Wachs, I. E.; Chan, S. S. and Chersich, C. C.; J. Catal., 1986 98,
102.

5. Zhang, Z. and Henrich, V. E., Surf. Sci., 277 (1992) 263. Gaior, M.; Gasior, 1.;
and Grzybowska, G.; Appl. C atal. 1984 10, 87.

6. Rao, C. N. R.; Raju, A. R.; and Vijayamohanan, K., Gas-sensor Materials in New
Materials, S.K. Joshi, C.N.R. Rao, R Rsruta and S. Nagakura (eds.)), New Delhi,
India, 1992, pp. 1-37.

7. Martin, C.; Rives, V.; Sanchez-Escribano, V.; Busca, G.; Lorenselli, V.; and
Ramis, G., Surf. Sci. 1991 251/252, 825.

8. Tufano, V. and Turco, M., Appl. Catal. 1993 32, 9.

9. Turco. M.; Lisi, L.; Pirone, R, Appl. Catal. 1994 32,133.

10. Went,G. T.; Leu, L.-.,J Rosin, R. R.; and Bell, AlexisT.; J. Catal. 1992 134,492.

11. Yamazoe, N. Sens. Act. 3 19915, 7. .

l2. Morrison, S. R. Sens. Act. 1987 12, 425.

13. Yoldas, B.E. J. Non-Cryst. Solids 38/39 (1980) 81.

14. Wark, T. A.; Gulliver, E. A.; Jones, L.C.; Hampden-Smith, M.J.; Rheingold, A. 1...;
Huffman, J. C. in Better Ceramics Through Cemistry IV, Zelinsky, B. J. J.;
Brinker, C. J.; Clark, D. E.; Ulrich, D. R., Eds.; Mater. Res. Soc. Proc. 180;
Pittsburg, PA (1990) 61.

15. Wellbrock, U.; Beier, W.; Frischat, G.H. j. Non-Cryst. Solids 1992 147,148, 350.

16. Sanchez, C.; Livage, J. new J. Chem. 1990 I4, 513

17. Livage, J.; Henry, M.; Sanchez, C. Prog. Solid State Chem. 1989 I8, 259.

18. Brinker, C. J.; Scherer, G. sol-Gel Science; Academic Press: San Diego, 1990.

19. Gagliardi, C. D.; Berglund, K. A. in Better Processing Science of Advanced
Ceramics, Askay, K. A.; McVay G. L.; Ulrich, D. R., Eds.; Mater. Res. Soc. Proc.
180; Pittsburg, PA (1986) 127.

20. Gagliardi, C. D.; Dunuwila D.; Berglund, K. A. in Better Ceramics Through
Chemistry IV, Zelinsky, B. J. J.; Briker, C.J.; Clark, D. E.; Ulrich, D. R., Eds.;
Mater. Res. Soc. Proc. 155; Pittsburg, PA (1990) 801.

21. Machej, T.; Haber, J.; Turek, A. M.; and Wachs, I. E., Appl. Catal. 1991 70, 115.

22. Wagner, C. D.; Riggs, M. W.; Davis, L. E.; Moulder, J.; Muilenberg, G. E.
Handbook of X -Ray and Photoelectron Spectroscopy, Physical Electronics
Industries, Eden Prairie, MN, 1979.

23. Severin, K. G.; Ledford, J. S. unpublished results.

24. Software provided by Dr. Andrew Proctor, University of Pittsbugh, Pittsburgh,
PA.

25. Cristiani, C.; Forzatti, P. and and Busca, G.; J. Catal. 1989 116, 586.

26.

27.
28.
29.

30.
31.

32.
33.

34.
35.

36.
37.
38.
39.

59

Nyquist, R. R. and Kagel, R. 0. Infrared Spectra of Inorganic Compounds: 3800-
45 cm"; Academic Press: New York, 1971.

Busca, G.; Centi, G.; Marchetti, L. and Trifiro, F. Langmuir 1986 2, 568.

Wachs, I. E., J. Catal. 1990 124, 570.

Klug, H. P.; Alexander, L. E. Diffraction Procedures for Polycrystalline and
Amorphous Materials, 1“ Ed., Wiley: New York, 1954.

Severin, K. G.; and Ledford, J. S., unpublished results.

Wachs, I. E.; Saleh, R. Y.; Chan, S. S. and Chersich, C. C.; Appl. Catal. 1985 15,
339.

Iwaki, T.; Miura, M. bull. Chem. Soc. me. 1971 44, 1754.

Radecka, M.; Zakrzewsda, K.; Cztematsek, H.; Stapinski, T. Appl. Surf. Sci. 1993
65/66, 227.

So, H.; Pope, M. T. Inorg. Chem. 1972 II, 1441.

Iwamoto, M.; Furukawa, J.; Matsudami, K.; Takenaka, T.; Kagawa, S. J. Am.
Chem. Soc. 1983 105, 3719.

Schraml-Marth, M.; Wokaun, A.; Baiker, A. J. Catal. 1990 124, 86.

Flynn, C. M., and Pope, M. T. J. Amer. Chem. Soc. 1986 92, 2715.

Sawatzky, G. A. and Post, D.; Phys. Rev. 3 1979 20(4), 1546.

Nickl, J.; Schlogl, R.; Baiker, A.; Knozinger, H. and Ertl, G.; Catal. Lett. 1989 3,
379. »

Chapter 3

Effects of Doping Levels on Vanadium Oxide/Titanium Oxide Thin Film
Conductimetric Sensors

3.1 Introduction

The previous chapter discussed the effects of annealing temperature on the
structure of sol-gel produced vanadium oxide/titanium oxide thin films. Continuing this
two-part investigation of the influence vanadium has on the structure of titanium oxide
thin films, we present here the effects of changing vanadium doping levels. Again,
preparation involved the hydrolysis of metal alkoxides using excess valeric acid at room
temperature. Data shows that the structure of the thin films could be changed by altering
the vanadium doping level. XPS data show the dependence of the vanadium oxidation
state and the V/T i surface atomic ratio on bulk V concentration. No crystalline vanadium
oxide was detected by XRD in any sample up through a 0.025 bulk V/T i atomic ratio.
Crystalline TiOz(anatase) was observed in all films, with the relative concentration and
particle size of TiOz crystals showing a dependence on the vanadium loading. UV-Vis
data also show that the amount of vanadium present in the film does not affect the TiOz
band gap.

TiOz-based metal oxide sensors have been the focus of studies to evaluate their
properties under various conditions including exposure to oxygen [1-5], humidity [6], and
NO. compounds [7]. Much evidence has shown that mixed metal oxides interact and
provide different properties than that of the metal oxide alone [8-10]. The V205/l'i02
system is widely used in industry because of its catalytic properties. These properties

60

61
include the reduction of nitric oxide in the presence of ammonia [11-13] and the oxidation

of hydrocarbons [14,15]. The catalytic activity and selectivity has been found to be
dependent on the amount of surface vanadium present. The optimal activity occurs when
a monolayer of vanadium is formed [13]. The presence of excess vanadium results in the
formation of crystalline V205. However, a moderate amount of this oxide species does
not significantly change catalytic activity, since crystalline V205 shows poor selectivity
towards these analytes [16].

Using sol-gel chemistry, mixed metal oxides can be obtained from metal alkoxide
precursors [17-19]. The alkoxide precursors undergo an inorganic polymerization
reaction resulting in a metal oxide backbone. The mixing of alkoxide precursors allows
homogeneous, multicomponent systems [20] leading to glasses at lowered temperatures.
The rheological properties of the resulting materials can be utilized to produce materials
through a wide variety of techniques [21-23]. It has been shown that the rate of the
hydrolysis and condensation reactions can be controlled by the introduction of carboxylic
acids [20,24]. Carboxylic acids affect the rates by complexing with the alkoxide precursor
and slowing the condensation reaction [25]. Stable and uniform thin films that adhere well
to quartz have been produced through the chemical modification of the alkoxide
precursors with valeric acid [26,27].

The effects of vanadium loading on the properties of the mixed oxide have
previously been studied [28-34]. It is widely accepted that vanadium ‘wets’ crystalline
TiOz up to the point of a monolayer [34] and a mechanism for this wetting (or migration)
has been proposed [35]. All systems previously studied have involved crystalline TiOz

(primarily anatase) as a starting material combined with vanadium either by mechanical

62
methods or through wet impregnation. There has not been a reported study of the V/T i

system synthesized via sol-gel chemistry. A purpose of the present research was,
therefore, to study how vanadium loading affects the structure of the resulting VZOJI‘ioz
system in materials prepared through 501- gel techniques.

Uniform vanadium oxide/titanium oxide thin films were prepared from vanadium
triisopropoxide oxide, titanium (IV) isopropoxide and valeric acid. Films were prepared
with varied doping levels of vanadium and characterized using a combination of surface
and bulk methods. X-ray photoelectron spectroscopy (XPS) was used to provide
information about the chemical species present in film surfaces and their relative
concentrations. X-ray diffraction (XRD) provided information about the crystal structures
in the films. Ultraviolet-visible spectroscopy (UV-Vis) was used to determine the
coordination of the vanadium species present and detemrine the band gap of TiO; in each

film

3.2 Experimental

Materials. Valeric acid (99+%, Aldrich Chemical Co.), Titanium (IV)
isopropoxide (97%, Aldrich Chemical Co.), and vanadium triisopropoxide oxide (Gelest
Inc.) were used without further purification. Distilled, deionized water was used for all
preparations.

Solution Preparation. Sols were prepared similarly to the method previously

developed for titanium [36]. Since both metal precursors are easily hydrolyzed, solutions

63
were prepared in a glovebox under continuous N2 purge. The vanadium loadings were

varied to produce sols with VII“ i atomic ratios of 0.26, 0.53, 0.111, 0.176, and 0.250.
Vanadium triisopropoxide was added to titanium (IV) isopropoxide followed by the
addition of valeric acid and then water. The reactions were carried out at room
temperature in capped vials. The metal:water:acid ratio was 1:1.5 :9. The solutions were
vigorously shaken with a vortex mixer following the addition of each reactant and were
prepared in duplicate.

Film Preparation. The solutions were deposited on a quartz substrate and spun
for a minimum of 5 minutes and air-dried overnight. Films were then annealed at 400°C
for 24 hours in a tube furnace under the flow of dry air (100 cc/min, medical grade, AGA
Gas). Film preparations from each solution were carried out in triplicate.

Reference films were prepared containing vanadium oxide or titanium oxide
separately. The solutions and films were prepared in similar manner to those containing
both vanadium and titanium. The film containing vanadium oxide was annealed at 500°C
for 4 hours. The titanium oxide film was annealed at 600°C for 4 hours.

X -ray Photoelectron Spectroscopy (XPS) Films cast on quartz slides were
analde with a VG Microtech spectrometer using a ClamZ hemispherical analyzer. Data
were collected with a Mg anode operated at 15 kV and 20 mA (300W) and an analyzer
pass energy of 50 eV. The core electron regions for four elements were scanned with
each film: C 1s 0 ls, Ti 2pm and V 2pm. Binding energy charge corrections for films
containing Ti were made with reference to the Ti 2pm peak (Ti 2pm = 458.5 eV [37]).

Corrections for reference films of V205 were rrrade with reference to C 1s (284.6 eV).

64
The binding energies given have a precision of i 0.1 eV. Quantitative measurements from

XPS data were performed using the sensitivity factors determined previously [38]. For
quantitative determination of the V 2pm region, the contribution from the 0 1s satellite
was subtracted out from the area attributed to V 2pm. The V/I‘ i surface ratio was
calculated from the core level features of each film. Curve fitting was performed using a
20% Lorentzian-Gaussian mix Voigt function [39].

X-ray Diffraction (XRD). All films were analyzed with a Rigaku XRD
diffractometer using Cu K01 radiation (1. = 1.5418 A). The X-ray tube was operated at 45
kV and 100 mA. The diffractometer scanned at a rate of 0.07 deg/min (deg = 20).
Divergence and scattering slits were both set at 1°. Background due to the quartz
substrate has been subtracted from each of the patterns.

Ultraviolet-visible spectroscopy (UV-Vis). UV-Vis spectra was collected with a
Unicam spectrometer. The incident beam was normal to the surface of the film using a
blank quartz slide as a reference. Data were collected using Vision software in absorbance

mode with a bandwidth of 2 nm and a data interval of 1 nm.

3.3 Results and Discussion

XRD. The XRD patterns for the various atomic ratios are presented in Figure 3.1.
The only feature observed with XRD was due to TiOz(anatase), thus the range presented
is between 20 values of 24.50 and 26.50. The formation of TiOz(anatase) crystals from a

completely amorphous TiOz film requires a temperature of at least 600°C [38]. However,

Intensity
I I I
>> 0‘
M
a 0

 

Degrees (29)

Figure 3.1. XRD diffraction patterns obtained for V/T i atomic ratios of: (a) 0.026, (b)
0.053, (c) 0.111, (d) 0.176, (c) 0.250.

66
we have shown the presence of vanadium influences the crystallization of amorphous

titanium at lower temperatures [40]. The broad feature appearing at the lowest loading
indicates the presence of even small amounts of vanadium will initiate crystal formation of
Ti02(anatase). As the loading is increased, an increase in the relative intensity of the
anatase peak is observed suggesting the increased presence of vanadium further promotes
the formation of TiOz(anatase). Vanadium crystals are not evident in the XRD patterns.
Along with the increased intensity of the TiOz(anatase) diffraction pattern,
narrowing of this feature’s FWHM was also observed. A decrease in the linewidth of an
XRD diffraction pattern is indicative of the formation of smaller particles [41]. Table 3.1
shows the particle sizes of TiOz(anatase) that were calculated from FWHM values of
XRD patterns using the Scherrer equation [41]. These calculations show an increase in
the T102 particle size as the vanadium loading increases. Calculations also suggest a
rrrinimum particle size of ~120 A could be obtained at 400°C through control of vanadium

loading.

Table 3.1. Particle size of TiOz(anatase) as determined through line broadening
calculations of XRD diffraction patterns.

 

 

 

Bulk V/Ti Atomic Ratio FWHM Particle Size (A)
0.026 0.65 125
0.053 0.68 120
0.111 0.61 133
0.176 0.44 187
0.250 0.36 224

 

 

67
Bond et al. [33] observed XRD lines due to vanadium oxide for V/T i mixed oxide

catalysts containing a vanadium loading greater than 0.029 V/Ti atomic ratio. Saleh et al.
[16] observed a trace amount of crystalline V205 in a catalyst with a 0.111 V/I' i atomic
ratio. Our results indicate that the vanadium loading can be significantly increased beyond

previously studied V/T i systems without the formation of vanadium oxide crystals.

UV-Vis. For reference, Figure 3.2a contains UV-Vis data collected from an
amorphous TiO; thin film containing anatase crystals. Confirmation of the presence of
crystals in the film was obtained through XRD analysis of the film. The broad absorbance
below 350 nm can be attributed to an electron transition from the valence to conduction
band of T102 [42]. Figure 3.2 also contains a spectrum from a film primarily composed of
V205 (3.2b). The absorbances observed for a vanadium oxide film are the result of a
charge transfer from vanadium to a terminal oxygen [10]. Schraml-Marth et al. [43]
report tetrahedrally coordinated vanadium(V) shows a charge transfer transition in the
285-335 nm region while the charge transfer transition for octahedrally coordinated
vanadium(V) occurs in the 400-500 nm region. The differences in the energies of the
absorbances in our spectrum of vanadium oxide suggest differences in coordination of
vanadium, where absorbances at 400 and 265 nm represent vanadium in its +5 oxidation
state with octahedral and tetrahedral coordinations, respectively.

Figure 3.3 presents the UV-Vis spectra obtained from films with varied atomic
ratios of V/T i. The strong absorbance below 350 nm in all spectra can be attributed to
TiOz. There is also a broad feature appearing in the T102 anatase film in the higher

wavelength region between 400 and 1000 nm. Tetrahedrally coordinated vanadium(V)

68

Normalized Absorbance

 

 

l 1 l l l a I I
200 400 600 800

 

Wavelength (nm)

Figure 3.2. UV-Vis spectra for reference films containing (a) TiOg(anatase) or (b) V205.

A .UUUJ US“ 0W

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 1 4 L 1
400 600 800 1000
—- e
d
C
b
_ a
l 1 J J J I I I l l
200 400 600 800 1000
Wavelength (nm)
Figure 3.3. UV-Vis spectra for films containing V/I‘ i atomic ratios of (a) 0.026, (b)

0.053, (c) 0.111, (d) 0.176 and(e) 0.250.

70
absorbs ultraviolet radiation below 350 nm. The strong TiOz feature also absorbing in this

region will not allow for this species to be differentiated in our spectra since the signal for
the vanadium species will be much weaker. The feature above 400 nm shifts and becomes
broader with the introduction of vanadium. There are two possibilities for vanadium ionic
states that could contribute to this broad absorption feature. First, octahedrally
coordinated V(V) will absorb between 400 and 500 nm and second, V(IV) in an
octahedral environment will absorb between 770 and 625 nm. At the lowest atomic ratios,
the vanadium ions remained in the +5 oxidation state, as evidenced by the increased
absorbance between 400 and 500 nm. When the vanadium loading was increased, the
absorbance near 500 nm began to shift to a higher wavelength. This shift can be attributed
to an increased presence of V(IV) caused by the interaction of the vanadium species with
anatase crystals of T102. When the vanadium loading is increased to 0.176 V/Ti, a shift
back to lower wavelength is observed. This shift is attributed in an increased presence of
vanadium that is fully oxidized to V5+ and is octahedrally coordinated. This trend
continues for the film derived from a sol containing a 0.250 V/Ti atomic ratio.

Another interesting feature that comes from our studies of semiconducting
materials that can be detected through UV-Vis spectroscopy is the band gap of T102. The
band gap of TiOz can be measured from the tail at the high wavelength end of the UV-Vis
absorbance. Accurate measurements of this tail have been discussed by Ibanez et al. [44].
The band gap was determined for the film containing TiO_a,(anatase) (Figure 3.2a) to be
3.57, slightly greater than the gap of 3.4 eV reported by Tang et al. for single crystal

TiOg(anatase) [45]. The band gap values for films containing atomic ratios between 0.026

71
and 0.250 as well as the value representing a primarily anatase film are presented in Table

3.2. The reported values have a relative standard deviation <1%. The TiOz band gap
remains constant as the V/I‘ i atomic ratio is increased, suggesting that the band gap is not
affected by the amount of vanadium present in our materials. Gautron et al. [46] observed
a significant reduction of the TiOz band gap to to near 2 eV when V02 was introduced

into TiOz.

Table 3.2. Measured Band gaps measured for V/Ti films. A film containing TiOz(anatase)
crystals is included for reference.

 

 

 

Bulk V/T i Atomic Ratio Band gap (eV)
TiOz 3.6
0.026 3.6
0.053 ‘ 3.6
0.111 3.5
0.176 3.5
0.250 3.5

 

 

XPS. The shifting of V 2pm binding energy as a function of loading has been
studied by various groups (see, for example, ref. [32]). An increased binding energy of
the V 2pm feature is observed as a result of the increased vanadium content in the film
(Figure 3.4). When the V/T i ratio is 0.026, a binding energy of 516.0 eV is observed.
According to previously published data on vanadium oxides [47], a feature occurring in
this region can be assigned to the vanadium ion primarily in its 4+ oxidation state. Since
free vanadium ions are easily reduced under vacuum, it is conceivable that the vanadia
species on the film surface is reduced prior to analysis with XPS (which occurred at a

pressure of ~10'9 torr). The binding energy of V 2pm increases by 0.5 eV when the WT i

Mir-ml l zed I ntenel ty

72

 

 

518 518 514
Bi ndl 119 Energy (eV)

Figure 3.4. XPS results showing the effect of vanadium loading on the feature due to

V2p3n. Films were derived from sols with (a) 0.026 (b) 0.053, (c) 0.111, (d)
0.176, and (e) 0.250 V/T i atomic ratios.

73
atomic ratio is increased to 0.053. This feature can be attributed to the presence of both

v“+ and v5+ in the surface of the film [47]. The v 213.,2 binding energy increases more
gradually when the mole ratio is increased further. The resulting shift to a higher binding
energy (516.9 eV for 0.250 V/T i atomic ratio) suggests the amount of V5+ increases in the
surface of the film with increased loading. The surfaces of films with the highest
vanadium content contain primarily V”.

The 0 Is core level remained unchanged when the loading of vanadium was
altered (Table 3.3), assuring that the V 2pm peak shift was not an artifact of the binding
energy correction. For all samples, the oxygen peak contained a strong feature due to the
backbone species centered at 529.8 eV with a shoulder appearing at 531.6 eV due to

hydroxyl groups bonded to TiOz on the surface of the film.

Table 3.3. XPS data for films derived from sols containing between 2.5 and 20% V.

 

 

 

Surface* V/I i
V/T i Atomic Ratio Atomic Ratio BE 0 1s BE V 2pm
0.026 0.023 529.8 516.0
0.053 0.079 529.8 516.5
0.111 0.189 529.8 516.7
0.176 0.310 529.8 516.8
0.250 0.395 529.8 516.9

 

 

*Determined from V 2pm / Ti 2pm intensity ratios.

When altering the V/Ii mole ratio of the alkoxide precursors, the surface
composition of the resulting metal oxide thin film changes as shown in Figure 3.5. At the
lowest vanadium loading, the surface atomic ratio in the film is close to (within

experimental error) the bulk atomic ratio. However, when the bulk ratio was increased

74

 

 

00"“ .
e

0.3—
0
r1 - O
0
r
0
=3

0.2—
3 e
3 o
0
0
I
‘ d
L
a

e
0.1-
e
- /O
o o— 9/
1 ' r ' 1 . 1 1 r . 1
0. 00 0. 06 0. 10 0. 15 0. 20 0. 25

Bulk V/‘I'i Atom’c Ratio

Figure 3.5. Comparison of surface and bulk V/T i atomic ratios. Experimental data
represented by circles. Squares represent a plot for equal bulk and surface
ratios.

75
above 0.025, the V/I‘ i surface ratio was higher than the bulk V/T i ratio. XRD data

presented earlier showed the increased concentration of Ti02(anatase) crystals as a result
of increased vanadium loading with the particle size also increasing with vanadium
loading. A higher surface ratio than the bulk ratio is attributed to the V species in the
V/I‘i mixture ‘spreading’ across the TiOz crystals. The increased presence of TiOz crystals
along with the increased particle size of these crystals provides a mechanism for the
observed migration of the vanadium species to the film’s surface. The migration of
vanadium is consistent with earlier results for catalysts derived through wet impregnation
of the vanadium species on a TiOg(anatase) support [48]. The V/l' i surface ratio increases
consistently up to the film containing a 0.111 bulk ratio. At higher vanadium content,
however, the rate of increase of the surface ratio shows non-linearity. Stranick et al. [49]
showed that this decreasing non-linearity of the surface ratio vs. bulk ratio plot indicates a
poorly dispersed dopant on the support, so our non-linear relation between bulk and
surface ratios suggests a poorly dispersed vanadium phase. This poorly dispersed phase is
attributed to the formation of crystalline V205. The fact that peaks are not observed in
XRD diffraction patterns indicates the V205 particles are smaller than 0.3 nm (the

detection limit for XRD [50]).

3.4 Conclusions

The use of sol-gel chemistry has allowed the formation of Ti02(anatase) at low

loadings, an increased amount of anatase crystals with increased vanadium content, and

76
further increased vanadium loading to the point suggesting the formation of crystalline

vanadium oxide. Materials in the current study have shown a change in the observed
monolayer coverage compared to materials obtained through traditional catalyst
preparation techniques.

We have shown in the previous chapters that temperature affects both the surface
and bulk structure of sol-gel derived vanadium oxide/titanium oxide thin films [40]. The
present study has shown the level of bulk vanadium in WT i thin films also affects the
film’s surface and bulk structures. Both vanadium and titanium crystal structures change
as the bulk atomic ratio changed. While vanadium loading mainly affects the particle size
of the TiOz(anatase) crystals, the vanadium species present in the film changed from
completely amorphous at low loadings to amorphous and crystalline at higher loadings.
XPS data suggests the vanadium ions present in the surface of the film are also influenced
by the bulk vanadium concentration.

The surface atomic ratio of V/T i was observed by XPS as the bulk ratio was
changed. Two conclusions can be made from the change in surface V/l‘ i ratio: First, the
non-linearity of the bulk vs. surface atomic ratios indicates a crystalline vanadium species
was formed although not detected through XRD. Second, the trend of the increasing
surface V/I‘ i ratio being greater than the increased vanadium doping suggests the vanadia
species spreads over the TiOa crystals, consistent with observations made by Haber et al.
[35]. Although the band gap of our films differs from that of crystalline T102, the gap was

not affected by the d0pin g level of vanadium in the films.

77
Vanadium loading can be used to control the surface and bulk phases of both

transition metals in a V/T i oxide system. The use of sol-gel technology to prepare these
mixed metal-oxide thin films has allowed us to change the surface structure and

morphology from that previously defined with technical catalysts.

78

3.5 List of References

1. Kimer, U.; Schierbaum, K. D.; deel, W. Leibold, B.; Nicoloso, N .; Weppner, W.;
Fischer, D.; Chu, W. F. Sens. Act. 1990 31, 103.

2. deel, W.; Kimer, U.; Wiemhtifer, H. D.; Rocker, G. Solid State Ionic 1988 28-
30, 1423.

3. Xu, Y.; Yao, K.; Zhou, X.; Quanxi, C. Sens. Act. 3 1993 13-14, 492.

4. Hasegawa, 8.; Sasaki, Y.; Matsuhara, S. Sens. Act. 3 1993 13-14, 509.

5. Wu, M. T.; Yao, X.; Yuan, 2. H.; Sun, H. T.; Wu, W. C.; Chen, H. Xu, G. Y.
Sens. Act. 3 1993 13-14, 491.

6. Nenov, T.; Yordanov, S.; Sens. Act. 3, 1992 8, 117.

7. Boccuzzi, F. Guglielminotti, E. Sens. Act. 3, 1994 21 , 27.

8 Stencel, J. M.; Makovsky, L. E.; Sardus, T. A.; de Vries, J.; Thomas, R.; Mouljin,
J. A. J. Catal. 1984, 90, 314.

9. Soled, S.; Murrel, L. L.; Wachs, I. E.; Mc Vicker, G. B.; Sherman, L. G.; Chan, S.
S.; Despenziere, N. C.; Baker, R. T. K. in Solid State Chemistry in Catalysis;
Brasselli, R. K., Brazkil, J. F., Eds.; American Chemical Society: Washington,
DC, 1985; p.165.

10 Wachs, I. E.; Saleh, R. Y.; Chan, S. S.; Chersich, C. C. Appl. Catal. 1985, 15,
339. ‘

1.1 Tufano, V. and Turco, M. Appl. Catal. 32 (1993) 9.

12. Turco. M.; Lisi, L.; Pirone, R.; Appl. Catal. 32 (1994) 133.

13. Went, G. T.; Leu, L.-J., Rosin, R. R.; and Bell, A. T.; J. Catal. 134 (1992) 492.

14 Rao, C. N. R.; Raju, A. R.; and Vijayamohanan, K., Gas-sensor Materials in New
Materials, 8. K. Joshi, C. N. R. Rao, R Rsruta and S. Nagakura (eds.)), New
Delhi, India, 1992, pp. 1-37.

15. Martin, C.; Rives, V.; Sanchez-Escribano, V.; Busca, G.; Lorenselli, V.; and
Ramis, 6., Surf. Sci. 251/252 (1991) 825.

16 Saleh, R. Y.; Wachs, I. E.; Chan, 8. S.; Chersich, C. C. J. Catal. 1986, 98, 102.

17 Yoldas, B. E. J. Non-Cryst. Solids 38, 39 (1980) 81.

18. Wark, T. A.; Gulliver, E. A.; Jones, L. C.; Hampden-Smith, M. J.; Rheingold, A.
L.; Huffman, J. C. in Better Ceramics Through Cemistry IV, Zelinsky, B. J. J.;
Brinker, C. J.; Clark, D. E.; Ulrich, D. R., Eds.; Mater. Res. Soc. Proc. 180;
Pittsburg, PA (1990) 61.

1.9. Wellbrock, U.; Beier, W.; Frischat, G. H. j. Non-Cryst. Solids 147,148 (1992)
350.

20 Livage, J.; Henry, M.; Sanchez, C. Prog. Solid State Chem. 18 (1989) 259.

21. Sakka, S.; Kamiya, K.; .J. Non-Cryst. Solids 1982 48, 31.

22. Dislich, H.; Hinz, P. J. Non-Cryst. Solids 1982 48, 11. .

23. Clark, D. E. in Science of Ceramic Chemical Processing, Hench, L. L. and Ulrich,
D. R. Eds., Wiley: New York, 1986, p.237.

24 Brinker, C. J.; Scherer, G. sol-Gel Science; Academic Press: San Diego, 1990.

25. Sanchez, C; Livage, J. New J. Chem. 1989 I8, 513.

26.

27.

28
29
30
31
32

33

34.

35

36.

37.

38.

39.

40.
41.

42.
43.

45.

46

47
48

49.
50.

79

Gagliardi, C. D.; Berglund, K. A. in Better Processing Science of Advanced
Ceramics, Askay, K. A.; McVay G. L.; Ulrich, D.R., Eds.; Mater. Res. Soc. Proc.
180; Pittsburg, PA (1986) 127. b)

Gagliardi, C. D.; Berglund, K. A. in Better Processing Science of Advanced
Ceramics, Askay, K. A.; McVay G.L.; Ulrich, D. R., Eds.; Mater. Res. Soc. Proc.
180; Pittsburg, PA (1986) 127.

Wachs, I. E.; Saleh, R. Y.; Chan, S. S.; Chersich, C. chemtech 1985, 756.

Went, G. T.; Leu, L. J.; Bell, A. T. J. Catal. 1992, 134, 479.

Inomata, M.; Mori, K.; Miyamoto, A.; Murakami, Y. J. Phys. Chem. 1983, 87,
761.

Lietti, L.; Forzatti, P.; Ramis, G.; Busca, G.; Bregani, F. Appl. Catal. 3 1993, 3,
13.

Nickl, J.; Schldgl, R.; Baiker, A.; Kniizinger, H. and Ertl, G.; Catal. Lett., 3 (1989)
379.

Bond, G. C.; Sarakany, A. J.; Parfitt, G. D. J. Catal. 1979, 57, 476.

Wachs, I. E.; Jehng, J. M.; Hardcastle, F. D. Solid State Ionics 1989 32/33, 904.
Haber, J .; chhej, T.; Servicka E. M.; Wachs, I. E. Catal. Lttr. 1995, 32, 101.
Severin, K. G.; Ledford, J. S.; Torgerson, B. A.; Berglund, K. A. Chem. Mater.
1994 6, 890.

Wagner, C. D.; Riggs, M. W.; Davis, L. E.; Moulder, J.; Muilenberg, G. E.
Handbook of X -Ray and Photoelectron Spectroscopy, Physical Electronics
Industries, Eden Prairie, MN, 1979.

Severin, K. G.; Ledford, J. S. unpublished results.

Software provided by Dr. Andrew Proctor, University of Pittsbugh, Pittsburgh,
PA.

Hoffman, L. W.; Ledford, J. S. manuscript in preparation.

Klug, H. P.; Leroy, E. A. X -Ray Dijfraction Procedures for Polycrystalline and
Amorphous Materials Wiley-Interscience: New York, 1974.

Iwaki, T.; Miura, M. bull. Chem. Soc. me. 1971 44, 1754.

Schraml—Marth, M.; Wokaun, A.; Baiker, A. J. Catal. 1990 124, 86.

Ibanez, J. G.; Solorza, 0.; Gomez del-Campo, E. J. Chem. Ed. 1991 68(10), 872.
Tang, H.; Lévy, G.; Berger, H.; Schmid, P. E. Phys. Rev. 3 1995 52(11), 7771.
Gautron, J.; Lemasson, P.; Poumellec, B.; Marucco, J-F. Solar Energy Mat. 1983,
9, 101.

Sawatzky, G. A. and Post, D.; Phys. Rev. B 1979 20, 1546.

Kihenslci, J.; Baiker, A.; Glinski, M.; Dollenmeier, P.; Wokaun, A. J. Catal. 1986
101, 1.

Stranick, M. A.; Houlalla, M.; Hercules, D. M. J. Catal. 1987 103, 151.
Niemantsverdriet, J. W. Spectroscopy in Catalysis VCH: New York, 1993.

Chapter 4

Future Work

In the previous two chapters 1 have shown that the chemical state and structure of
the surface species in V/T i thin films can be controlled through loadings and annealing
temperatures. By increasing the temperature, the surface V” i ratio increases. This
increase in surface ratio is attributed to the ‘wetting’ of the vanadium species on the T102
crystals as they are formed - with larger crystals forming at higher temperatures. The
amount of vanadium present in the bulk also affects the surface V/T i ratio. Again, larger
TiOz crystals are formed with increased bulk V/l' i ratio, allowing the vanadium species to

spread on these crystals.

It is desirable to correlate the information derived from the present studies with the
activity of these thin films to develop a more complete understanding of the VIP i metal
oxide thin film conductimetric sensor. As outlined below, the device to conduct sensor
measurements has already been developed. The next step will be to detemrine the system
to be used for initial sensor measurements. Through catalysis studies, the active site of the
VIP i system in the oxidation of hydrocarbons is attributed to the reducibility of the surface
vanadium species [1]. Correlating this information with results from chapter 2 pertaining
to the surface V/I' i ratio, the best system to use for initial sensor measurements would be a
thin film containing 0.111 bulk V/I’ i atomic ratio prepared at 600°C. This film contains a

high surface V/F i ratio with no indication of crystalline vanadium formed and is therefore

80

81
presumed to show high surface activity and selectivity towards surface reations with

reducing analyte gases such as hydrocarbons.

4.1 Sensor measurements.

Since we are interested in a systematic investigation of the structure-reactivity correlations
of V/I’ i metal oxide sensor materials, a system must be developed to carry out the sensor
measurements. In parallel with the studies described previously, we have developed a
device to make conductivity measurements using the semiconducting properties of our
metal oxide thin films. The preparation of the thin film on an alumina substrate has been
adapted from earlier work published by deel and co-workers [2], and is shown
schematically in Figure 4.1. The oxide substrate we chose to use was polished alumina.
The most commonly used electrodes are gold; however, the adhesion of the oxide surface
to the gold breaks down at high temperatures. Platinum, with the use of a titanium
adhesion layer, however, has been shown to withstand high temperatures [2]. When
titanium is deposited, a titanium oxide layer approximately 10-20 A thick will form on the
alumina substrate. Further deposition results in the formation of a Ti metal layer on the
T102. Platinum can then be deposited on the Ti metal layer with good adhesion. Our
method to deposit the thin metal films involved physical vapor deposition of Ti and Pt. A
tungsten boat centaining the particular metal was resistively heated by applying a current
to the melting point of that metal. The deposited Ti layer was approximately 100 A thick

while the Pt layer for the electrodes was approximately 500-800 A thick. The Pt heater

82

11mm >

 
  
  

Pt electrodes with
Ti interface

4-point conductimetric sensor array, top view _ . _
semiconducting thin film

Pt electrodes
/.
\

m

 

<——-—-A1203 or 8102

 

 

\Trmterface

cross sectional view
Pt heater

Figure 4.1. Schematic of sensor array.

83
also used a Ti adhesion layer with the platinum approximately 2—3 pm thick. Gold wires

were bonded to the platinum surfaces to serve as electrical contacts. A schematic diagram
of the cell used to make the sensor measurements is shown in figure 4.2. A gas of interest
may be delivered to the cell from cylinders using mass flow controllers. Control and data
acquisition are handled through a PC that is interfaced with the flow controllers and

electronics of the cell.

A known current will be applied across two opposite electrodes on the sensing
substrate with a resistance measured across the other two. When a particular analyte gas
is passed over the surface of the thin film, reaction of this analyte gas with the surface with
change the conductance of the surface (as outlined in chapter 1), thus changing the
resistance of the thin film. The sensor response to a particular analyte can then be
characterized assuming a linear relationship between concentration of the analyte gas and

resistance of the film.

4.2 Oxidation reactions.

Initial experiments would involve a study of the effect of reducing gases such as
CO or CH: on sensor response. As stated in the introduction, reducing gases are expected
to react with adsorbed oxygen on the surface thereby changing the conductance of the
film’s surface. We have shown that many different surface phases can be obtained in our
thin films via vanadium loading and/or annealing temperature. However, the ease in

reduction of our surface V by analyte gases is not known. The oxidation of hydrocarbons

84

thermocouple leads

heater leads

 

gas flow

 

 

 

 

 

 

 

 

Figure 4.2. Cell used to make sensor measurements.

 

electrode leads

85
over vanadia catalysts supported by titanium oxide has been studied extensively [1,3-7].

In one particular study, the partial oxidation of methanol by vanadia species supported on
TiOz, Deo and Wachs [1] suggest that the activity of the surface vanadium oxide phase is
independent of loading when the vanadium oxide content is kept below monolayer
coverage. These authors concluded that the surface reactivity of vanadium oxide/titanium
oxide is correlated with the reducibility of the supported vanadium oxide. Gasior at al. [3]
studied the effect of the T102 support phase (anatase or rutile) on the selective oxidation
of o-xylene to phthalic anhydride. The results of their study led to the following
conclusions: anatase-containing samples 1) show a higher activity and selectivity in the
oxidation of o-xylene, 2) have a lower degree of reduction of the vanadia species during
the catalytic oxidation, and 3) have a lower observed dehydration/dehydrogenation ratio
than rutile-containing samples. The surface states responsible for the catalytic oxidation of
hydrocarbons over vanadia/titanium dioxide have been well established. However, our use
of sol-gel technology has produced V/I‘ i thin films with surface structures different from
those used in previous studies. It would therefore be interesting to study the surface

phases responsible for sensor activity resulting from exposure to various reducing gases.

4.3 Reduction reactions.

Another interesting study would involve the response measurement of the surface
of a VII i film in the presence of various NOx compounds. We have shown the films to be

stable up through an annealing temperature of 600°C. This would allow for a study of

86
automobile exhaust to determine the amounts of NOx compounds via conductimetric

sensing if the sensor response to NO, has been developed. The use of vanadia/titania
catalysts for the reduction of NO. compounds in the presence of ammonia has also
received much attention [8-18]. Early studies by Inomata at al. [8] have shown that the
surface oxygen species V=O speeds up the NO-NH; reaction. By studying the activity of
the reduction of NO over vanadia on varied supports, these authors have shown the NO-
NH3 reaction to be structure insensitive. Thus, a higher amount of surface activity can be
attained through a higher concentration of V=O species. T0psoe at al. [12] have recently
reported the catalytic activity of NO. reduction over vanadia/titania catalysts is due to two

different reactions occurring on the surface, denoted as acid or redox, as reported below:

NH: v’to-"H-th o-V"

v-"*-o-"*H.N~-Hov“ “20
NO

v-‘to *HaN-N "'H-O-V‘”

Vs+-O“"’H3N-I\EO H—O-V‘”
N2 + HzO

Acid Redox

87
The adsorption of ammonia occurs at the Bronsted acid sites or VSI-OH. After a partial

transfer of the acidic proton, NO can react with the intermediate to produce the released
products, N2 and H20. Oxidative regeneration of the reduced vanadium species in

V‘”-OH to fully oxidized V=O species completes the cycle.

In view with information presented in the last two sections, the field of chemical
sensors utilizing properties of the vanadia/titania system provides many opportunities for
research. The control of oxidation states, crystallinity, and surface structures of both
vanadium and titanium (leading to a change in the concentration of V=O) provides an
interesting starting point for the study of conductimetric sensor responses of

vanadia/titania thin films with respect to variousanalyte gases.

88

4.4 List of References

ONUIPD) Nr—t

00>!

10
11
12

13.

14

15.
16.

17.

18

Deo, G; Wachs, I. E. J. Catal. 1994, 146, 323.

Schierbaum, K.D.; Vaihinger, 8.; deel, W.; Van Den Vlekkert, H.H., Kloeck, B.
De Rooij, F. Sens. Act., 1990, 31, 171.

Gasior, M.; Gasior 1.; Grzybowska, B, Appl. Catal. 1984, 10, 87.

Rusiecka, M.; Grzybowska; Gasior, M.; B, Appl. Catal. 1984, 10, 101.

Busca, G.; Elmi, A.; Forzatti, P. J. Phys. Chem. 1987, 91, 5263.

Wachs, I. E.; Saleh, R. Y.; Chan, S. S.; Chersich, A. C. Appl. Catal. 1985, I5,
339.

Deo, G; Wachs, I. E. J. Catal. 1994, I46, 335.

Inomata, M.; Miyamoto, A.; Ui, T.; Kobayashi, K.; Murakami, Y. Ind. Eng. Chem.
Prod. Res. Dev. 1982, 21, 424.

Wong, W.C.; Nobe, K. Ind. Eng. Chem. Prod. Res. Dev. 1984, 23, 564.

Baiker, A.; Dollenmeier, P.; Glinski, M.; Reller, A. Appl. Catal. 1987, 35, 351.
Went, G.T.; Leu, L.-J.; Rosin, R.R.; Bell, A.T. J. Catal. 1992, 134, 492.

Topsoe, N.-Y.; Topsoe, H.; Dumesic, J.A. J. Catal. 1995, 151, 226.

Topsoe, N.-Y.; Dumesic, J .A.; Topsoe, H. J. Catal. 1995, 151, 241.

Handy, B.E.; Maciejeweski, M.; Baiker, A. J. Catal. 1992, 134, 75.

Schneider, M.; Scharf, U.; Wokaun, A.; Baiker, A. J. Catal. 1994, 150, 284.
Ciambelli, P.; Bagnasco, G.; Lisi, L.; Turco, M.; Chiarello, G.; Musci, M.; Notaro,
M.; Robba, D.; Ghetti, P. Appl. Catal. 3 1992, 1, 61.

Tufano, V.; Turco, M. Appl. Catal. B 1993, 2, 9.

Turco, M.; Lisi, L.; Pirone, R.; Ciambelli, P.. Appl. Catal. B 1994, 3, 133.

.. hr”
twv‘ 7n

no

“Fart: 39...“: -4
M -‘ . g . . .
I I

”one

11.".-
1 *5».

 

$44.3

um
u...
1 . . .\

*.

 

.-
m: -

. r
I-J.
3:15.
we r
3'20?
.3. é'c' '1

4
a .

1.44.

t ;’
r .

tag; at ii
I‘ ‘5

areas
.1
‘a- I“ i 1

.t
m:
.3!”

was i r»
v 'I . . ‘
1 2» c “31.?“ ‘ 93 ‘ . :1 ’1‘
{a . . g .7 ‘0'! 1'
"1.50,: A
a'a‘:.'.;z"1l
‘1 .9"

.3?-
.i‘,‘

.3»
.3. , m.“
“(I‘d-n

-. .-
'13..

gel
.—
s

~>-
1

n».
e»
'5

In.

1‘ u.
7w}
M ‘
u...
...

n

’bh'nv

'2‘: :.

4..
v s
1"...“

'e
r

"'1 'I'iilzi
" 4'4 1:

.
4'.) 1,"

.,.
a“
u
v-
7'.
“.9
an
"4..
nuyv-
~>t

Om

urn-e-

a».
‘ tau-x”
”an”

t‘u

.443.
"Jr? '9»
.
"i

v

23»
. -.. 4:...“
.... ”use“,
and...» VA; fl“.
. -.
£114

‘hw

.‘Sv
5.531 .
I! d.

.. t

m
WA
“I .‘r
. ...
..
“2‘3.-
4

yum

"‘- m
at:
.
us.
£2.
.

~
tn.
'1}
..
.

 

I g a! [1
ii‘iigg‘igii
’ q

1
.. . g
g .
A I 111‘111'3

+ .. 1
, n
‘3‘."-
1.23
a...

-
.5...»
‘45-:
u...
3:...

_.
,o

i
i l

' i '4' . 1'. .
int {titties-iii »

 

1"
a: