CHARACTERIZATION STUDY OF PLATINUM-DOPED STANNIC OXIDE
CERAMICS FOR METHANE SENSING IN AIR
ZUHAIRI BIN IBRAHIM
A thesis submitted in fulfilment of the
requirements for the award of the Degree of
Doctor of Philosophy
Faculty of Science
Universiti Teknologi Malaysia
JUNE 2005
PSZ 19:16(Pind. 1/97)
UNIVERSITI TEKNOLOGI MALAYSIA
BORANG PENGESAHAN STATUS TESIS Š
JUDUL:
CHARACTERIZATION STUDY OF PLATINUM-DOPED
STANNIC OXIDE CERAMICS FOR METHANE SENSING
IN AIR.
SESI PENGAJIAN :
2004/2005
ZUHAIRI BIN IBRAHIM
Saya
(HURUF BESAR)
mengaku membenarkan tesis (PSM/Sarjana/Doktor Falsafah)* ini disimpan di Perpustakaan
Universiti Teknologi Malaysia dengan syarat-syarat kegunaan seperti berikut :
1. Tesis adalah hakmilik Universiti Teknologi Malaysia.
2. Perpustakaan Universiti Teknologi Malaysia dibenarkan membuat salinan untuk tujuan
pengajian sahaja.
3. Perpustakaan dibenarkan membuat salinan tesis ini sebagai bahan pertukaran antara
institusi pengajian tinggi.
4. **Sila tandakan (9 )
√
SULIT
( Mengandungi maklumat yang berdarjah keselamatan atau
kepentingan Malaysia seperti yang termaktub di dalam AKTA
RAHSIA RASMI 1972)
TERHAD
( Mengandungi maklumat TERHAD yang telah ditentukan
oleh organisasi/badan di mana penyelidikan dijalankan)
TIDAK TERHAD
___________________________
(TANDA TANGAN PENULIS)
Disahkan oleh
(TANDATANGAN PENYELIA)
PROF. MADYA DR. ZULKAFLI OTHAMAN
Nama Penyelia
Disahkan oleh
Alamat Tetap:
97, JALAN MEWAH RIA 4/4, TAMAN
BUKIT MEWAH, 81200 JOHOR BAHRU,
(TANDATANGAN PENYELIA)
PROF. DR. MOHD MUSTAMAM ABD KARIM
Nama Penyelia
Disahkan oleh
JOHOR.
Tarikh: 16.6.2005
_____________________________
(TANDATANGAN PENYELIA)
DR. DIANE HOLLAND
Nama Penyelia
CATATAN : * Potong yang tidak berkenaan.
** Jika tesis ini SULIT atau TERHAD, sila lampirkan surat daripada pihak
berkuasa/organisasi berkenaan dengan menyatakan sekali sebab dan tempoh tesis ini
perlu dikelaskan sebagai SULIT atau TERHAD.
Š Tesis dimaksudkan sebagai tesis bagi Ijazah Doktor Falsafah dan Sarjana secara
penyelidikan, atau disertasi bagi pengajian secara kerja kursus dan
penyelidikan atau Laporan Projek Sarjana Muda (PSM).
“I/We* hereby declare that I/we have read this thesis and in my/our*
opinion this thesis is sufficient in terms of scope and quality for the award
of Doctor of Philosophy”
Signature
:
Name of Supervisor I
:
P.M. DR. ZULKAFLI OTHAMAN
Date
:
16.6.2005
Signature
:
Name of Supervisor II :
PROF. DR. MOHD MUSTAMAM
ABD KARIM
Date
:
Signature
:
16.6.2005
Name of Supervisor III :
DR. DIANE HOLLAND
Date
16.6.2005
*Delete as necessary.
:
BAHAGIAN A – Pengesahan Kerjasama*
Adalah disahkan bahawa projek penyelidikan tesis ini telah dilaksanakan melalui kerjasama
antara __________________________ dengan __________________
Disahkan oleh :
Tandatangan
:
…………………………………....…
Nama
:
………………………………………
Jawatan :
………………………………………
Tarikh : …………
(Cop rasmi)
* Jika penyediaan tesis / projek melibatkan kerjasama.
BAHAGIAN B – Untuk Kegunaan Pejabat Sekolah Pengajian Siswazah
Tesis ini telah diperiksa dan diakui oleh:
Nama dan Alamat
Pemeriksa Luar
: Prof. Dr. Radzali Othman
School of Materials & Mineral Resources Engineering
Universiti Sains Malaysia
Engineering Campus
14300 Nibong Tebal
Penang
Nama dan Alamat
Pemeriksa Dalam I
: Prof. Madya Dr. Karim Deraman
Fakulti Sains
UTM, Skudai
Pemeriksa Dalam II
: Prof. Madya Dr. Yusuf Wahab
Fakulti Sains
UTM, Skudai
Nama Penyelia Lain
(jika ada)
:
Disahkan oleh Penolong Pendaftar di Sekolah Pengajian Siswazah:
Tandatangan
:
……………………………………… Tarikh : ………….
Nama
:
GANESAN A/L ANDIMUTHU
ii
DECLARATION
I declare that this thesis entitled “CHARACTERIZATION STUDY OF PLATINUMDOPED STANNIC OXIDE CERAMICS FOR METHANE SENSING IN AIR” is the
result of my own research except as cited in references. The thesis has not been accepted
for any degree and is not concurrently submitted in candidature of any other degree.
Signature
:
Name
: Zuhairi Bin Ibrahim
Date
: 16.6.2005
iii
This thesis is dedicated to my
my beloved wife (Rohanin Ahmad) and my dearest son (Faisal Zuhairi).
Thank you for being with me all along.
iv
ACKNOWLEDGEMENTS
I would like to express my thanks to all my supervisors; Assoc. Prof. Dr. Zulkafli
Othaman (UTM), Prof. Dr. Mohd Mustamam Abd Karim (UPSI, Universiti Pendidikan
Sultan Idris) and Dr. Diane Holland (University of Warwick, Coventry, England) for
being very resourceful, inspiring, supporting and understanding during my study. Thank
you too to Prof. Dr. Mohd Rahim Sahar (UTM) for getting me started, for his
supervision and for being a very helpful whilst he was the head of the Physics
Department. I am also grateful to Assoc. Prof. Dr Khairi Saidin who have provided me
with the platinum powder and Assoc. Prof. Dr. Rosli Hussin for providing the aid for the
construction of the probe in the SECS. To all the staffs in the Physics Department
(UTM) thank you for visiting, caring and the encouragement you gave all along.
My progress would be slow without the ever helpful hands of Mr. Mohd Jaafar
bin Mohamed Raji, Mr Lee Siak Kuan and Mr. Chin Keng Kwang. I am also grateful to
Dr. Johari Adnan for getting the internet facilities going. My special thanks to friends I
made Dr. John Ojur Dennis and Dr. Agus Steyo Budi who were always there when
needed. To all the friends I had and made during the time of my study, you made my
days went on like there was always something to look forward. I am ever so grateful to
Allah Almighty who chose me as his visitor to Mecca and Madinah via my Hajj 2003.
v
ABSTRACT
Pure SnO2 and Pt-SnO2 ceramics were prepared by the dry-pressing method
using a pressure of 40 MPa and sintered at various temperatures between 100-000oC
from a mixture of powders of (100-x)SnO2.xPt (0 ≤ x wt % ≤ 5). The electrical
properties of the ceramics were studied using a home-made Sensor Element
Characterization System (SECS) and an Impedance Analyzer. The sensing probe of the
SECS was modified so it was much slimmer with most of the electrical connections
concealed and could measure either the bulk or surface resistance of the ceramic. The
optimum composition for detecting methane in air was 0.5 wt.% Pt-SnO2 sintered at
1000oC and the optimum operating temperature was at 400oC. The resistance of the 0.5
wt.% Pt-SnO2 in 25000 ppm methane decreases from ~ 54.0 kΩ to ~ 4.6 kΩ at
temperatures of 200oC up to 440oC respectively. The activation energies were between
0.30 eV and 0.45 eV for temperatures between 200oC and 400oC. The corresponding
conductance (G) decreased with Pt loading and the gas partial pressure (p) or methane
flow rate (χ). As such, it indicates that the doped SnO2 is an n-type semiconductor. The
conductance power law takes the form G ~ p-0.5 and this indicates that the chemisorbed
ions on the doped ceramics depended only on temperature. The conductance (G)methane concentration (c) takes the form G = kc0.35. A linear relationship ln G = 0.35 ln
c – 11.9 was obtained when plotting ln G against ln c. The relative conductance change
(∆G/G) and the square root of methane concentrations (c½) obey the relationship ∆G/G
= 0.08c ½ which indicates the doping with 0.5 wt.% Pt increased the sensitivity of the
base material (SnO2) to methane by a factor of 133. The response and recovery times
were affected by the methane flow rate, operational temperature, level of doping with
values between 30 s up to 154 s and between 600 s up to 1317 s respectively. The
doping of Pt at 0.1 wt.% up to 5 wt.% in SnO2 produced ceramics with densities of
7.01g/cm3 up to 7.03 g/cm3 which exceeds the full density of pure SnO2 (6.90 g/cm3).
The strength and stability were indicated from the doped SnO2 measurements of Vickers
hardness (10 GPa and up to 19 GPa), Young modulus (20 GPa and up to 55 GPa) and
Bulk modulus (20 GPa and up to 80 GPa) for Pt loadings between 0.1wt.% and 2.5
wt.%. High resolution X-ray diffraction showed that the mean crystallite size ranges
between 25 nm and 55 nm for Pt loadings from 1 wt% up to 5 wt.% in SnO2. The strain
in doped samples could not be eradicated by either sintering at high temperature
(1000oC) or high Pt loading (5 wt.%). X-ray photoemissions spectroscopy (XPS),
Mössbauer and nuclear magnetic resonance (NMR) analysis showed that the doped
SnO2 has additional chemical environment (compared to pure SnO2) can be attributed to
the ease of detecting methane in air via electrical measurements.
vi
ABSTRAK
Seramik Timah Oksida tulen dan timah oksida yang didop dengan Platinum telah
disediakan dengan kaedah Tekanan Kering dengan menggunakan tekanan 40 MPa dan
disinter pada suhu antara 100-1000oC daripada campuran dalam bentuk bedak
berkomposisi (100-x)SnO2.xPt (0 ≤ x %berat ≤ 5). Pencirian elektrik bahan tersebut
dilakukan dengan mengunakan alat yang dibina dinamakan Sistem Cirian Elemen
Sensor (SECS) and Penganalisa Impedans bagi mengesan gas metana di udara. Prob
pengesan Sistem Cirian Elemen Sensor diubah agar ia lebih langsing dengan sambungan
elektriknya terlindung dan boleh mengukur rintangan padu atau rintangan permukaan
seramik. Adunan optimum untuk mengesan metana di udara adalah 0.5 % berat Pt-SnO2
dan suhu operasi optimumnya pula ialah 400oC. Rintangan elektrik bagi 0.5 % berat PtSnO2 di dalam 25000 bahagian per juta metana di udara susut dari ~ 54.0 kΩ ke 4.6 kΩ
pada suhu 200oC hingga ke 400oC, masing-masing. Konduktans (G) pula susut dengan
tambahan Pt dan tambahan tekanan separa gas (p) atau kadar aliran metana (χ). Dengan
itu tertunjuk bahawa SnO2 yang didop ialah semikonduktor jenis-n. Hukum kuasa
konduktans dinyatakan dalam bentuk G ~ p-0.5 dan ini menunjukkan ion-ion yang
diserapkimia pada seramik yang didop hanya bersandar kepada suhu. Hubungan antara
konduktans (G) dan kepekatan metana (c) adalah dalam bentuk G = kc0.35. Hubungan
linear ln G = 0.35 ln c – 11.9 diperolehi bila memplot ln G lawan ln c. Perubahan relatif
konduktans (∆G/G) dan punca ganda dua kepekatan metana (c1/2) mematuhi hubungan
∆G/G = 0.08 c½ , yang menunjukkan 0.5 % berat Pt meningkatkan kepekaan bahan asas
(SnO2) kepada metana dengan faktor sebanyak 133. Masa respons dan masa pemulihan
dipengaruhi oleh kadar aliran metana, suhu operasi, amaun dopan dengan nilai- nilai
30 s hingga 154 s dan antara 600 s hingga 1317 s, masing-masing. Mengedop Pt dari
0.1 % berat sehingga 5 % berat dalam SnO2 menghasilkan seramik dengan ketumpatan
7.01 g/cm3 hingga 7.03 g/cm3 yang melebihi ketumpatan penuh SnO2 (6.90 g/cm3).
Kekuatan dan kesetabilan SnO2 yang didop ditunjukkan oleh ukuran dari kekerasan
Vickers (10 GPa sehingga 19 GPa), Modulus Young (20 GPa sehingga 55 GPa) dan
Modulus Pukal (20 GPa sehingga 80 GPa) bagi tambahan Pt dari 1 % berat sehingga 5
% berat dalam SnO2. Pembelauan sinar-X resolusi tinggi menunjukkan min saiz kristalit
berada dalam julat 25 nm sehingga 55 nm untuk tambahan Pt dari 1 % berat sehingga 5
% berat dalam SnO2. Spektroskopi fotopancaran sinar-X (XPS), Mössbauer dan
salunan-magnetik-nuklear (NMR) menunjukkan SnO2 yang didop memiliki suasana
kimia tambahan (berbanding dengan SnO2 tulen) yang mungkin menjadi atribut
mudahnya mengesan metana di udara melalui pengukuran elektrik.
vii
TABLE OF CONTENTS
CHAPTER
1
TITLE
PAGE
Title page
i
Declaration
ii
Dedication
iii
Acknowledgements
iv
Abstract
v
Abstrak
vi
Table of Contents
vii
List of Tables
xiii
List of Figures
xv
List of Symbols
xxv
List of Appendices
xxviii
INTRODUCTION
1
1.1
General introduction to gas sensing
1
1.1.1
Methane gas
1
1.1.1.1 Anthropogenic methane sources
2
1.1.1.2 Natural methane sources
3
Ceramics
4
1.1.2.1 Electronic ceramics
4
1.1.2
1.2
Justification for research
6
1.2.1
Methane gas and global warming
7
1.2.2
Methane gas explosions
7
1.2.3
The importance of methane sensing in Malaysia
8
viii
2
1.3
Scope of study
10
1.4
Statements of hypothesis
11
1.5
Objectives of the study
12
1.6
Thesis plan
12
GAS SENSORS REVIEW
15
2.1
Overview
15
2.2
Methane gas sensing
22
2.2.1
Pellistors
23
2.2.2
Semistors
23
2.2.3
Planar heat wire
24
2.2.4
Pd-SiC or Pt-SiC Schottky diodes
24
2.2.5
TWPF –Thermal Wave Pyroelectrics Film
24
2.2.6
Metal oxides
25
2.3
Stannic oxide as a sensing element
2.4
Problems and improvements with SnO2-based methane
2.5
3
25
sensor
27
Current research trend on SnO2-based methane sensor
29
SENSING MECHANISMS AND MODELS
31
3.1
Sensing parameters
31
3.2
Sensing mechanisms
31
3.2.1
Defect formation
32
3.2.2
Depletion layer formation
32
3.2.3
Polycrystalline materials
34
3.2.4
Grain size effects
36
3.2.5
High temperature sensing
36
3.2.6
Catalyst doping
36
3.2.6.1 Fermi energy control
37
3.2.6.2 Spillover mechanism
40
Current status on methane sensing mechanisms
40
3.3
ix
3.4
4
Methane detection
40
EXPERIMENTS AND MEASUREMENT TECHNIQUES
45
4.1
Introduction
45
4.2
Ceramics preparation
45
4.3
Bulk density and porosity
48
4.4
SECS-Sensor Element Characterization System
48
4.4.1
49
Response and recovery time
4.5
Impedance spectroscopy
50
4.6
Elastic modulus
51
4.6.1
The ultrasonic mechanical characterization
system (UMC)
52
4.6.2
Young modulus
55
4.6.3
Bulk modulus
56
4.7
Vickers hardness
4.8
Brunauer-Emnett-Teller (BET) and Barrett-Joyner-
56
Halenda (BHJ)
57
High resolution X-ray diffraction (HRXRD)
58
4.10 Scanning electron microscope (SEM) and EDAX
60
4.11 Transmissions electron microscope (TEM)
60
4.12 Atomic Force Microscope (AFM)
61
4.9
4.13 Fourier Transform Infra-Red Spectroscopy (FTIR) and
FT-Raman Shift spectroscopy
61
4.14 X-ray photoelectron spectroscopy (XPS)
63
4.15 Mössbauer spectroscopy
64
4.16 Nuclear Magnetic Resonance (NMR)
65
4.17 Differential Thermal Analysis (DTA)
65
4.18 Thermal Gravimetric Analysis (TGA)
66
x
5
SECS-OPTIMUM OPERATING TEMPERATURE AND
OPTIMUM COMPOSITION
72
5.1
Measurements from SECS
72
5.1.1
Improvements on the SECS
72
5.1.2
Resistance (R), conductance (G) and
sensitivity (S/) determination
74
5.1.3
Effects of RL on RS in dry synthetic air
77
5.1.4
Effects of RL on RS in 25000 ppm methane
5.1.5
5.2
5.3
6
(in air)
79
Reproduciblity
80
Optimum operating temperature and
composition
81
Conclusion
85
RESISTANCE, CONDUCTANCE AND SENSITIVITY
86
MEASUREMENTS
6.1
Introduction
86
6.2
Resistance in air
87
6.3
Resistance in methane
89
6.4
Conductance in air and methane
92
6.5
Conductance-time dependent
97
6.6
Conductance-power law
98
6.7
Conductance-methane gas concentrations
100
6.8
Sensitivity-time dependent
101
6.8.1
Relative conductance and sensitivity
103
6.8.2
Effects of flow rates
104
6.9
7
Discussion and conclusion
105
RESPONSE AND RECOVERY TIME
110
7.1
Response and recovery time
110
7.2
Effect of flow rate
110
xi
8
7.3
Effect of methane concentrations
114
7.4
Effect of operational temperature
116
7.5
Effects of Pt loadings
118
7.6
Conclusion
119
MICROSTRUCTURE AND PHYSICAL
120
PROPERTIES ANALYSIS
8.1
Introduction
120
8.2
TGA
120
8.3
Density
122
8.4
True porosity
123
8.5
BET and BJH
128
8.6
Vickers hardness
133
8.7
Elastic modulus
135
8.8
High-resolution X-ray diffraction (HRXRD)
136
8.8.1
Starting powders
136
8.8.2
Pt-SnO2 ceramics
140
8.8.2.1 Sintering effects on pure SnO2
140
8.8.2.2 Effects of Pt loadings on pure SnO2
141
8.8.2.3 Intensity ratio (I211/ I220)
147
8.8.2.4 Intensity ratio (Ihkl/ I110)
149
8.8.2.5 Induced strain calculations
151
8.9
Raman-Shifts spectroscopy
159
8.11 FTIR spectroscopy
161
8.11 Surface analysis
166
8.11.1
EDAX
166
8.11.2
SEM and AFM
171
8.11.3
X-ray photoemissions spectroscopy
(XPS) analysis
172
8.11.4
Mössbauer spectroscopy analysis
177
8.11.5
NMR analysis
180
xii
9
8.12 Summary and conclusion
181
CONCLUSIONS
183
9.1
Sumamary and findings
183
9.2
Recommendations
187
REFERENCES
APPENDICES
PUBLISHED PAPERS
188
226-235
236
xiii
LIST OF TABLES
TABLE NO.
2.1
TITLE
Development in the research field of gas sensors
PAGE
17
1960s – 1980s.
2.2
Development in the research field of gas sensors
18
late 1990s until 2001.
5.1
Load resistance and sensor resistance with voltage
78
supply VC = 20 V in synthetic air.
5.2
Load resistance and sensor resistance with voltage
79
supply VC = 20 V in 25 000 CH4 in air.
5.3
Sensitivity of various composition of Pt-SnO2
82
ceramics at operating temperatures 250-430oC.
5.4
Sensitivity of 0.1-1.0 wt.% Pt in SnO2 at operating
84
temperature of 400oC extracted from Figure 5.10.
6.1
Activation energy for various Pt loadings in SnO2
88
in intrinsic and extrinsic regions.
6.2
Resistance of 0.5 wt.% Pt-SnO2 in 25 000 ppm CH4.
90
6.3
Conductance and gas flow rate.
99
6.4
Flow rate and maximum sensitivity.
105
7.1
Response and recovery time for various flow rates.
112
7.2
Response and recovery time for various
114
CH4 concentration.
7.3
Response and recovery time for various operating
temperature.
116
xiv
7.4
Response and recovery time for various Pt
118
loadings in SnO2.
8.1
Distance dhkl calculated using Bragg formula for
154
0.5 wt.% Pt-SnO2 sintered between 700-1000oC.
8.2
Summary of microstructure/physical properties
of Pt-SnO2.
181
xv
LIST OF FIGURES
FIGURE NO.
1.1
TITLE
Microstructure of a fine ceramic showing the
PAGE
5
grain and grain boundary of a typical ceramic.
3.1
Formation of surface depletion layer in SnO2.
33
Ec is the conduction band energy, Ef is the Fermi
level energy and Ev is the valance band energy
and ES is the surface barrier energy.
3.2
Model of a typical compressed powder and
34
Schottky barrier formation.
3.3
Well-sintered polycrystalline formation of
35
undepleted neck region.
3.4
Effect of noble metal in SnO2-Fermi energy control.
37
3.5
Band model for catalyst and n-type semiconductor.
38
Ecs is the surface conduction band edge of the n-type
semiconductor. Ev is the valance band and Ef is the
Fermi energy.
3.6
Depletion region formed by a well-dispersed catalyst.
38
3.7
Spillover of oxygen and hydrogen on a catalyst
39
doped semiconductor.
3.8
Chemical model of spillover mechanism of SnO2
40
doped noble metal.
3.9
Stoichiometric SnO2 (110) surface with bridging
oxygen atoms on the top.
41
xvi
3.10
Non-stoichiometric (reduced) SnO2 (110) surface with
42
in-plane oxygen only.
3.11
Reaction scheme: methane with oxygen on a sputtered
43
SnO2 film at 774 K.
4.1
Preparation of the Pt-SnO2 ceramics.
46
4.2
Sensor Element Characterization System (SECS).
49
4.3
Configuration of sample between electrodes.
49
4.4
Schematic diagram for the determination of response
51
and recovery time. The quantity in the vertical axis
can be VL, RS, G or S/.
4.5
Impedance spectrometer at the Physics Department,
52
University of Warwick, England.
4.6
Ultrasonic Mechanical Characterization (UMC).
53
(a) simple pulse ultrasonic system,
(b) envelope of pulse echo train, (c) echo as seen
on an oscilloscope.
4.7
Pulse echo overlap waveforms.
55
4.8
Acoustic wave propagation in sample.
56
4.9
The average length of the Vickers diagonals.
59
4.10
Autosorb Micromeritic (Model ASAP 2010).
59
4.11
BET classification of absorption isotherms.
60
4.12
Scanning Electron Microscope (JEOL, model
63
JSM-6100) at the Physics Department,
University of Warwick, England.
4.13
Transmission Electron Microscope (JEOL Electron
64
Microscope (JEM), model 2000FX) at the Physics
Department, University of Warwick, England.
4.14
Atomic Force Microscope (ARIS, model 3300),
at the Physics Department, University of Warwick,
England.
65
xvii
4.15
Equipment used for FTIR and FT-Raman spectroscopy
67
at Institute of Ibnu Sina, Universiti Teknologi Malaysia.
4.16
Mössbauer spectrometer (schematic diagram).
70
5.1
The sensor probe used in the GSCS (left) and SECS
73
(right).
5.2
Electrical circuit for sensor resistance measurement.
74
5.3
Typical data collection from SECS, graph of VL
75
versus time (t). Methane gas was introduced at
t = 100 s and cut-off at t = 300 s. The sample used
was 0.5 wt.% Pt-SnO2 and operating at 400oC.
VC = 20 V and RL = 700 Ω.
5.4
Corresponding sensor resistance (RS) versus time (t)
76
graph. Methane gas was introduced at t = 100 s and
cut-off at t = 300 s.
5.5
Corresponding conductance (G) versus time (t)
76
graph.
5.6
Corresponding sensitivity (S/) versus time (t) graph.
77
5.7
Graph of RS against RL in dry synthetic air.
78
5.8
Graph of RS against RL in 25 000 ppm CH4.
80
5.9
Graph of load voltage (VL) against time (t) showing
81
the reproducibility feature.
5.10
Sensitivity curves (S/) at various operating temperature
82
(T).
5.11
Graph of sensitivity (S/) against Pt loading (W*) at
84
operating temperature of 400oC.
6.1
Graph of resistance (R) against temperature (T)
87
in air.
6.2
Arrhenius plot for samples at temperatures between
o
50-450 C.
88
xviii
6.3
Graph of Activation energy (EA) against Pt loadings
89
(W*) for temperatures between 21-200oC and
200-450oC.
6.4
Plot of resistance (R) in 25 000 ppm methane
90
against temperature (T) for 0.5 wt.% Pt in SnO2.
6.5
Arrhenius plot for 0.5 wt.% Pt in SnO2 in 25 000
91
ppm methane.
6.6
Graph of conductance (G) against temperature (T)
92
in air.
6.7
Graph of conductance (G) against temperature (T)
93
in air (for doped samples only).
6.8
Graph of conductance (G) against temperature (T)
94
in methane for 0.5 wt.% Pt-SnO2.
6.9
Graph of conductance (G) against 1000/T in methane
95
for 0.5 wt.% Pt-SnO2.
6.10
Graph of conductance (G) against inverse temperature
96
(1000/T) of a thick film sensor.
6.11
Graph of conductance (G) against time (t) in
97
methane for 0.5 wt.% Pt-SnO2 at temperatures
300, 330, 350, 380 and 400oC.
6.12
Graph of ln G against ln χ.
99
6.13
Graph of conductance (G) against CH4
100
concentration (c).
6.14
Graph of ln G against ln c.
101
6.15
Graphs of sensitivity (S/) against time (t) of
102
0.5 wt.% Pt-SnO2 in air and in 25 000 ppm CH4
at operating temperatures 350– 440oC at flow rate
of 400 sccm.
6.16
Graph of relative conductance change (∆G/G)
against square root of concentration (c1/2) for pure
SnO2 and 0.5 wt.% Pt-SnO2 sintered at 1000oC.
103
xix
6.17
Sensitivity curves (S/) against time at flow rates
105
(χ).
6.18
Graph of intensity (I) against binding energy (EB).
106
XPS spectrum of the modified sample showing
Pt(0) and Pt(2) states.
6.19
Comparison of sensor resistance (RS) for 0.5 wt.% Pt-
107
SnO2 in air.
7.1
Sensitivity (S/) curves against time (t).
111
7.2
Response and recovery time (t) at various
112
flow rates (χ).
7.3
Sensitivity (S/) against time (t) at various CH4
113
concentrations.
7.4
Graph of response and recovery time (t) against CH4
114
concentration (c).
7.5
Sensitivity (S/)-time (t) curves at operating
115
temperatures 250-400oC.
7.6
Response time (t) against operating temperatures (T).
116
7.7
Graph of sensitivity (S/) against time (t) for
117
various Pt loadings.
7.8
Response and recovery time (t) against Pt
118
loadings (W*).
8.3
TGA of unsintered 0.5 wt.% Pt-SnO2 powder.
121
Graph of weight loss (W) against temperature (T).
8.2
Graph of density (ρ) of pure SnO2 against
122
sintering temperature (T).
8.3
Graph of density (ρ) of doped SnO2 against
123
Pt loadings (W*) sintered at 1000oC.
8.4
Graph of true porosity (p/) of pure SnO2 against
124
sintering temperature (T).
8.5
Graph of true porosity (p/) of doped SnO2 against
Pt loadings (W*) sintered at 1000oC.
124
xx
8.6
SEM micrographs. A: Pure SnO2 sintered at 500oC
125
and B: Pure SnO2 sintered at 700oC.
8.7
Graphs of true porosity (p/) and bulk density (ρ)
126
o
against Pt loadings (W*) sintered at 1000 C.
8.8
SEM micrographs. A: 0.1 wt.% Pt sintered at 1000oC
127
and B: 0.5 wt.% Pt sintered at 1000oC.
8.9
Graph of BET specific surface (S) area and particle
128
size (R/) versus sintering temperature.
8.10
Graph of BET specific surface (S) area and particle
129
size (RX) versus Pt loadings (W*) sintered at 1000oC.
8.11
Adsorption/Desorption curve from BET analysis.
130
8.12
TEM: A fresh Pt powder; B fresh SnO2 powder.
131
8.13
Pore size distribution by the BJH method.
132
8.14
Graph of Vickers hardness (HV) against
133
temperature (T).
8.15
Graph of Vickers hardness (HV) and bulk density (ρ)
134
versus Pt loading (W*) sintered at 1000oC.
8.16
Graph of velocity (v) of transverse/longitudinal wave
135
o
versus Pt loadings (W*) sintered at 1000 C.
8.17
Graph of bulk, Young and shear modulus versus
135
Pt loadings (W*) sintered at 1000oC.
8.18
XRD pattern of fresh pure SnO2 powder. Plot of
137
intensity (I) against Bragg angle (2θ).
8.19
Typical TEM images of fresh pure SnO2 powder.
137
8.20
XRD pattern of fresh Pt powder.
138
8.21
Typical TEM images of Pt powder: A; scale bar
139
50 nm, B; scale bar 20 nm.
8.22
XRD pattern of pure SnO2 sintered at temperatures
500-1000oC. Diffractograms show the plot of
intensity (I) against Bragg angle (2θ).
140
xxi
8.23
Calculated mean crystallite size (Rx) of pure SnO2
141
against sintering temperature (T).
8.24
XRD pattern of Pt-SnO2 at Pt loadings 0.5 – 5.0 wt.%
142
sintered at 1000oC. Diffractograms show the plot of
intensity (I) against Bragg angle (2θ).
8.25
Mean crystallite size (Rx) of Pt-SnO2 at Pt
143
loadings (W*) sintered at 1000oC.
8.26
Mean crystallite size (Rx) of 0.5 wt.% Pt-SnO2
144
against sintering temperatures (T).
8.27
XRD of doped (3 wt.% Pt) and undoped sample
144
sintered at 1000oC in the (101) direction.
Diffractograms show the plot of intensity (I)
against Bragg angle (2θ).
8.28
Dispersion of 0.5 wt.% Pt. The white speckles
145
are Pt clusters against the dark background of SnO2.
8.29
Broad shoulder formation of Pt (111) at high angle
146
side of 3 wt.% Pt-SnO2 sintered at 1000oC.
8.30
Peak shifts of 0.01o to lower angle side of doped
146
o
(3 wt.% Pt-SnO2) sintered at 1000 C with respect
to peak of undoped SnO2 sintered at 1000oC in the
(101) direction.
8.31
Graph of intensity ratio (Ihkl/I110) against sintering
148
temperature (T) for pure SnO2.
8.32
Graph of intensity ratio (I211/I220) against Pt loadings
148
o
(W*) sintered at 1000 C.
8.33
Graph of intensity ratio (I211/I220) against sintering
149
temperature (T) for 0.5 wt.% Pt-SnO2.
8.34
Graph of intensity ratio (Ihkl/I110) against sintering
150
temperature (T) for pure SnO2.
8.35
Graph of intensity ratio (Ihkl/I110) against sintering
temperature (T) for 0.5 wt.% Pt-SnO2.
150
xxii
8.36
Graph of intensity ratio (Ihkl/I110) against Pt
150
loadings (W*) sintered at 1000oC.
8.37
Graph of mean strain (<e2>1/2) of the atoms in
152
plane (hkl) in the normal direction to the plane
against sintering temperature (T) for undoped SnO2.
8.38
Graph of mean strain (<e2>1/2) of the atoms in
152
plane (hkl) in the normal direction to the plane
against sintering temperature (T) for 0.5 wt.%Pt-SnO2.
8.39
Graph of mean strain (<e2>1/2) of the atoms in
153
plane (hkl) in the normal direction to the plane
against Pt loadings (W*) sintered at 1000oC.
8.40
Distortions δc and δa against sintering temperature
155
for pure SnO2.
8.41
Distortions δc and δa against sintering temperature
156
(T) for 0.5 wt.% Pt-SnO2.
8.42
Distortions δc and δa versus Pt loadings (W*)
157
o
sintered at 1000 C.
8.43
Raman Shift spectra of the undoped SnO2 dry-
159
pressed ceramics sintered at temperatures 100-1000oC.
8.44
Raman Shift spectra; plot of intensity (I) against
160
Raman shift (δ) of the Pt-SnO2 dry-pressed
ceramics sintered at 1000oC.
8.45
FTIR spectra of the fresh SnO2 powder.
161
8.46
FTIR spectra of the fresh Pt powder.
162
8.47
FTIR spectrum of Pt- SnO2 ceramics sintered at
163
o
1000 C at various Pt loadings.
8.48
FTIR absorption spectrum of Pt-SnO2 ceramics
165
sintered at 1000oC at various Pt loadings after
exposure to 25 000 ppm CH4 at 400oC.
8.49
EDAX spectrum of undoped (pure SnO2) sintered at
1000oC.
167
xxiii
8.50
EDAX spectrum of 0.5 wt.% Pt-SnO2 sintered at
167
1000oC.
8.51
SEM micrographs of 0.5 wt.%Pt- SnO2 sintered at
168
1000oC. The white speckles are Pt metal clusters and
the dark background is SnO2.
8.52
SEM micrographs 1-3 of pure SnO2 sintered at various
170
sintering temperature and SEM micrographs 4-6 of pure
SnO2 doped at various Pt loadings and sintered at 1000oC.
8.53
AFM topography of 0.5wt.% Pt-SnO2 sintered at
171
1000oC. Scanned area 14x14 (µm)2 in the x-y
direction with maximum height-z direction, 2.55 µm.
8.54
XPS spectra of 3 wt.% Pt-SnO2 dry-pressed ceramic
172
o
sintered at 1000 C. Plot of intensity (I) against
binding energy (EB).
8.55
O 1s of doped (3 wt.% Pt-SnO2) and undoped
173
(SnO2) sintered at 1000oC.
8.56
3d5/2 and 3d3/2 of 3 wt.% Pt-SnO2 dry-pressed ceramic
174
sintered at 1000oC.
8.57
Graph of of intensity (I) against binding energy (EB).
175
Pt 4f 7/2 and 4f5/2 of 3 wt.% Pt-SnO2 dry-pressed
ceramic sintered at 1000oC.
8.58
Peak ratio Pt 4f 7/2 and 4f5/2 of 3 wt.% Pt-SnO2 dry-
176
pressed ceramic sintered at 1000oC.
8.59
XPS valance band of the pure and doped SnO2
177
o
sintered at 1000 C.
8.60
Mössbauer spectrum for pure SnO2 sintered at 1000oC
178
8.61
Mössbauer spectrum of 3 wt.% Pt-SnO2 sintered
178
at 1000oC.
8.62
Combined Mössbauer spectrum of undoped (pure SnO2)
o
and doped (3wt.% Pt-SnO2) sintered at 1000 C.
179
xxiv
8.63
Chemical shift (δ) with respect to SnCl2 solution
for pure SnO2 and 3wt.% Pt-SnO2 sintered at 1000oC.
180
xxv
LIST OF SYMBOLS
A
-
area of cross-section
A/
-
absorbance
Å
-
Angstroms
B
-
bulk modulus
β
-
Full Width Half Maximum
c/
-
velocity of light
c
-
gas concentration
cps
-
count per second
χ
-
flow rate of methane
dV
-
average length of the Vickers diagonals
δ
-
shifts (Raman, chemical)
δa
-
distortions in the lattice parameter a
δc
-
distortions in the lattice parameter c
e
-
electronic charge
E
-
elastic modulus
Eg
-
forbidden band gap energy
EC
-
conduction band energy
EV
-
valence band energy
ED
-
ionization energy of donors
EA
-
ionization energy of acceptors
Ea
-
activation energy
Ef
-
Fermi energy
eVS
-
work function of an electron
εo
-
permittivity of free space
xxvi
εr
-
relative permittivity
F
-
force
F/
-
structure factor
G
-
conductance in methane
Go
-
conductance in air
HV
-
Vickers hardness
hkl
-
Miller indices
h
-
Planck constant
Io
-
intensity of incident beam
IT
-
intensity of transmitted beam
K
-
absorption coefficient
kB
-
Boltzmann constant
∆L
-
increase in length
L
-
original length
l,z
-
thickness
λ
-
wavelength of X-ray radiation (e.m. radiation)
µ
-
shear modulus
Ni
-
net density of ions in the space charge region
n
-
refractive index of a medium
NS
-
negative surface charge
ρ
-
bulk density
p/
-
porosity
P
-
load
Po
-
partial pressure of oxygen
p
-
partial pressure
p/
-
porosity
R/
-
particle size
R
-
electrical resistance
RX
-
mean crystallite size
RF
-
radio frequency
S/
-
sensitivity
xxvii
Š
-
selectivity
S
-
specific surface area
sccm -
standard cubic centimetre per minute
σ
-
electrical conductivity
θ
-
diffraction angle
/
θ
-
phase angle
T
-
transmission
VC
-
voltage supply
VL
-
voltage across load resistor
VS
-
voltage across sensor
vL
-
longitudinal velocity
vS
-
shear velocity
W
-
weight loss
W*
-
Pt loading
W1
-
weight in air
W2
-
weight in toluene
Y
-
Young modulus
xxviii
LIST OF APPENDICES
APPENDIX
TITLE
PAGE
A
Tin oxide powder specification
226
B
Platinum powder specification
227
C
SECS
228
D
Data from SECS
232
E
Calculation of FWHM, β
235
1
CHAPTER 1
INTRODUCTION
1.1
General Introduction to Gas Sensing
A sensor is a form of transducer which converts physical or chemical
quantity into an electrical quantity for the purposes of measurement. A transducer
is a device which converts one form of energy into another. A gas sensor is then a
chemical sensor whose sole purpose is to determine the gas composition and
concentration via an electric signal. The use of sensors has increased as it was
necessary where environmental, health and safety issues are concerned to improve
the quality of life (Brailsford and Logothesis, 1998). For example, low level toxic
gases emission from exhaust systems could only be possible if high efficient
sensors are realised (Ogita et al., 2001).
1.1.1
Methane gas
Methane gas is colourless, odourless and lighter than air. The methane gas
is a molecule which is made up of 1 carbon atom and 4 hydrogen atoms. Natural
sources of methane include wetlands, grass hydrate, termites, oceans and
freshwater bodies. Human related (anthropogenic) activities like fossil fuel
production, animal husbandry, rice cultivation, biomass burning and waste
management also release methane into the atmosphere and alter the atmospheric
2
composition. Also, almost all (95%) of the methane emissions are from coal in
underground mines. This is still the main danger in coal mines all over the world.
Methane gas sensing is difficult because it is colourless and odourless. A crude
method of methane sensing is by its pungent smell when an additive such as
mercaptan is added in low dosage. Mixtures of methane and air between 5 to 15%
methane when ignited can burst into flame and explode (Leer, 1992). This will
then cause widespread fire and can claim many lives.
1.1.1.1 Anthropogenic methane sources
Amongst the anthropogenic sources of methane are landfills, natural
gas and oil systems, domesticated livestock, coal mining, livestock manure, rice
cultivation, biomass burning and wastewater treatment. Under anaerobic
conditions (without oxygen) landfills and open dumps decompose and generate
methane. The volume of methane generated depends upon the waste mass and the
moisture content. One of the primary component of natural gas is methane which
escapes to the environment during the production, processing, storage,
transmission and distribution stages. The fact that the gas is found in conjunction
with oil means that the production, refinement, transportation and storage of crude
oil is also considered as a source of methane. Cattle, buffalo, sheep, goats are
amongst the ruminant animals kept as domesticated livestock. These animals
produce methane as part of their digestive processes. It is in their large forestomach or rumen that microbial fermentation takes place where the feed is
converted into products that can be digested by the animal. The byproduct of the
microbial fermentation is methane which is eructed or released by the animal.
Human too produces methane via their digestive processes but the emissions from
this source is insignificant compared to the case of livestock. Methane which is
trapped in coal deposits and in the surrounding strata is released during coal
mining operations. Methane is also emitted during the combustion of coal.
Reducing the emissions of methane from coal mining is environmentally
beneficial as it also a greenhouse gas. Liquid manure from ponds, lagoons and
3
holding tanks also promotes methane production as the manure is produced from
decomposition of the organic matter in the livestock and poultry manure. A
flooded rice field is an ideal environment for methane production as it contains
high levels of organic substrates, oxygen-depleted conditions and moisture for
anaerobic decompositions. The level of emissions varies with soil conditions and
production practices. In countries like Indonesia, biomass is burned as part of
their agricultural system as well as for fuel. A small but significant amount of
methane is produced - 95% is carbon dioxide and carbon monoxide. Waste water
treatment can only produce methane if the organic matter in the waste water is
treated anaerobically. If methane is produced, it is directly released to the air.
1.1.1.2 Natural methane sources
The known natural sources of methane are wetlands, fossil, termites and
freshwaters. Natural wetlands are a rich anaerobic environment and abundant in
organic matter. As such, it is a conducive habitat for methanogenic bacteria
(methane producing bacteria) and enhances the decomposition of the organic
matter, thus producing methane. Methane was created in the geologic past and
found in the earth’s crust in the form of gas hydrates and permafrost. Hydrates
are solids comprising water molecules that contain methane molecules which are
found in both the polar regions and ocean sediments. Permafrost methane
originates from biological processes and is trapped in shallow permafrost ice and
soil before it reaches the atmosphere. Today, the amount of permafrost is
decreasing and more methane is being released to the environment. Cicerone and
Oremland (1988) reported the emissions of methane from termites depended on
the termite population, amount of organic matter consumed, type of species and
the methane-oxidizing bacteria activity. The freshwater environment is an ideal
place for the decomposition of wetlands plants which then emits methane.
Emissions from these natural sources are dependent on the temperature and
rainfall. For example, temperature changes can promote microbial activity, thus
enhancing methane production.
4
1.1.2
Ceramics
The term ceramics is defined as the art and science of making and using
solid articles which have their essential component, and are composed in large
part of, inorganic non-metallic materials (Kingery et al., 1976). This definition is
not limited to just pottery, porcelain, refractories, structural clay products,
abrasives, porcelain enamels, cements and glass but it also applies to non-metallic
magnetic materials, ferroelectrics, single crystals and glass-ceramics. Barsoum
(1997) defined ceramics as solid compounds that are formed by the application of
heat and sometimes heat and pressure, comprising at least one metal and a nonmetallic elemental solid or non-metal, a combination of at least two non-metallic
elemental solids, or a combination of at least two non-metallic elemental solids
and a non-metal.
Today, ceramics are, in short “solid inorganic non-metallic materials made
by firing” (Murata, 2000). The term ceramics is now classified as traditional and
fine ceramics. These ceramics have common features; resistance to rust, heat
resistance, non-flammability, extreme hardness and ease of forming. These
features are meaningful because neither plastics nor metals have all these features.
Today’s fine ceramics are a new breed or a new kind of material. To distinguish
between fine and traditional ceramics, the latter are made of natural materials
whilst the former are produced by putting the atomic compositions of various,
refined elements together through scientific forming and sintering processes. In
other words, fine ceramics are made by scientifically controlling chemical
compositions and this brings the realisation of new materials customised to the
unlimited amount of purpose they serve. Therefore, fine ceramics can be grouped
as functional materials such as electronic ceramics, optical ceramics and catalyst
or structural materials such as bio-ceramics, heat resistance structure and artificial
jewellery.
5
1.1.2.1 Electronic ceramics
The secret of the characteristics of fine ceramic is in its microstructure.
To a layman it is like an artificially created small piece of stone as shown in
Figure 1.1
grain
grain boundary
FIGURE 1.1: Microstructure of a fine ceramic showing grain and grain
boundary of a typical ceramic.
Technically, they are finely aggregated grains and traditional ceramics are
comparatively far more porous and more irregular. The grains and grain
boundaries are all scientifically controlled and show specific electrical responses
to electrical potential or environmental changes. These specific reactions are
utilised for specific purposes. For example, titanium dioxide (TiO2) ceramics or
barium titanate (BaTiO3) ceramics are polarised when voltage is applied to them.
On the other hand, other types of ceramics containing different additives, though
mainly composed of the same BaTiO3, serve as unique semiconductors which turn
an electric flow on and off under a given condition. This is therefore an
application of the electrical changes in their grain boundaries. Another example is
when the inclusion of a catalyst in semiconductor ceramics such as TiO2 or SnO2
affects the conductance of the material which can serve as sensing element in a
gas detector. Thus the function of electronic ceramics varies according to their
internal microstructure (Saito, 1988).
6
Most natural stones are insulators. In contrast, fine ceramics can be
designed with different conductivities by adjusting their composition; some are
conductors and others are insulators. This variation is one of the greatest
advantages of electronic ceramics. Electronic ceramics can further sub-divided
into magnetic ceramics, transparent ceramics, pyroelectric ceramics,
semiconductive ceramics, piezoelectric ceramics, insulating ceramics and
dielectric ceramics.
1.2
Justification for Research
In the past, gas sensors were used to control industrial processes and to
warn of poisonous gas leakages (Carotta et al., 1991). In Europe, controlling air
quality was mandatory by 2001 as stipulated by Council Directives such 96/61/EC
and 96/62/EC (Saul Garcia and Fernandez, 1999 and O’Malley 1999). For
example, the National Air Quality Standard for CO adopted by the UK
government in January 2000 is currently 10 ppm for a running 8 hours mean
(Stewart, 2000). It was therefore necessary to focus research on sensors capable
of monitoring pollutant gases and controlling combustion processes both at home
and in industry (Ruiz et al., 2002). The demands for more accurate and dedicated
sensors to monitor and control environmental pollution have led to the
development of new sensing materials to improve sensitivity, selectivity and
stability of sensors (Sharma et al., 2001).
With reference to an environmental issue, gases like CO2, methane, water
vapour, ozone, nitrous oxide and halocarbons play a significant role in enhancing
the greenhouse effect. The greenhouse effect is primarily a function of the
concentration of water vapour, carbon dioxide and other trace gases in the
atmosphere that absorb the terrestrial radiation leaving the earth surface. The
changes in the atmospheric concentration of these greenhouse gases will alter the
7
balance of energy transfers between the atmosphere, space, land and the ocean
which will give rise to global warming (Houghton et al., 1996).
1.2.1
Methane gas and global warming
Methane gas is amongst the greenhouse gases and the atmospheric
concentrations of methane have doubled over the last 200 years and continue to
rise, although the rate of increase is slowing (Dlugokencky et al., 1998). The
natural methane emissions to the atmosphere are 30% from wetlands, oceans,
termites while the remaining 70% is anthropogenic, from human activities such as
agriculture, usage of fossil fuel and waste disposal (Fung et al., 1991). When
methane enters the atmosphere, it reacts with molecules of oxygen and hydrogen
known as OH radicals. The OH radicals combine with methane and they
decompose into carbon dioxide and water vapour. Increasing emissions of
methane will reduce the concentration of OH radicals, a feedback which may
increase methane’s atmospheric lifetime. While most greenhouse gas studies
focus on CO2, methane is 20 times more potent as a heat trapping gas in the
atmosphere (Houghton, 2001). Thus, methane gas is also a leading contributor to
global warming after carbon dioxide.
1.2.2
Methane gas explosions
Methane is the major component (95 %) of natural gas, thus it can be used
to produce energy. The lower explosion limit (LEL) is 5% methane in air and the
upper explosion limit is 15% methane in air. However, gas explosions are
frequently reported in homes, pipelines and coalmines world wide. For example,
an explosion at one of the gas pipeline owned by Brunei Shell company in Seria
destroyed two homes and hundreds of residents in the vicinity were forced to
abandon their homes (Othman, 2000). The cause of the explosion was due to a
8
corroded pipeline which leaked out methane gas (Teo, 2000). Another incident, at
a Terengganu gas processing plant in Kertih, owned by Petronas Malaysia caught
fire and killed three workers (Alias and Hamidah, 2002). Frequent gas explosions
occurred in the 1960s in Japan largely associated with the popular usage of bottled
liquid petroleum gas for domestic purposes (some 23 million households use them
for cooking requirements and another 18 million used piped gas). In countries
like the United Kingdom a similar problem was reported in the Ronan Point
disaster. The disaster at Ronan Point in London, England in May 1968 was a gas
explosion which led to the collapse of one whole corner of a high rise building
and the death of three people. In Malaysia, closed packed condominiums,
apartments and flats which utilize bottled or channelled gas pipeline are also
prone to disasters like those in Japan.
1.2.3
The importance of methane sensing in Malaysia
There are two main reasons for researching into methane sensing in
Malaysia. Malaysia has substantial resources of natural gas from offshore fields
such as those based in Terengganu and Sarawak. Its reserves of natural gas
ranked 12th in the world (Wu, 2000a). The gas reserves here are dedicated to the
Peninsular Gas Utilization Project which provided 37% of the main sources of
primary commercial energy for the period 1996 to 2000. By 2005, the
contribution is expected to rise to 39.9%. Under the 8th Malaysia Plan (20012005), the Government will continue to promote gas usage. Malaysia has the
largest natural gas reserves among the Southeast Asian economies and is the third
largest amongst the Asia Pacific economies. At the turn of the century, the
recoverable gas reserves were 84.4 trillion standard cubic feet, 43% offshore east
coast of West Malaysia, 48% located offshore Sarawak and 9% offshore Sabah.
This introduces 1753 km of gas pipelines into the network of both domestic and
industrial sectors (Balce, 2002). Ambitious constructions of natural gas pipeline
in Malaysia like the one from Kuala Terengganu to Segamat and its branches and
the 220 miles of gas pipeline from Gulf of Thailand to the northern state of Kedah
9
which will provide natural gas for industries and home would certainly need gas
sensors to detect leaks and seepage of gases. It is foreseen that detection and
measurement of natural gas leaks such as methane is required on a day-to day
basis especially in the natural gas industry such as Petronas, Gas Malaysia and
from gas appliances, gas piping inside buildings or buried gas piping.
In modern, high rise flats, apartments and condominiums in Malaysia,
methane gas is supplied via such a network of pipelines. The gas is normally used
for domestic cooking or drying clothes. Such areas are enclosed due to the usage
of air-conditioning. Therefore, methane leakages in the concealed pipeline
network will accumulate in high concentrations in a very short time. This
inevitably needs methane monitoring for both public safety and environmental
issues. The early warning of methane presence would led to necessary steps that
could save lives and preserve the environment.
Malaysia also has large resources of tin and its tin reserves ranked as the
world’s third largest (Carlin, 2001). Tin is mined by various methods; gravel
pump (53.5%), open cast (20.2%), retreatment or Amang plant (12.6%), panning
and underground (8.3%) and dredging (5.4%). Malaysia exported 20 614 tons of
refined tin and the domestic demand was 5639 tonnes in 2000 (Wu, 2000b). The
local consumption are from the Malaysian solder industry (56.3%), pewter
industry (14.8%), tin plating industry (10.6%) and other end users (18%) reported
by the Wu (2000b).
An oxide of tin known as stannic oxide (SnO2) is easily obtained from
pure tin or tin derivatives in the form of thin, thick film or pellet (ceramic) form.
Stannic oxide is a well-known material used for CO and CH4 gas sensing.
Nevertheless, the state of methane gas sensing using stannic oxide needs further
investigation as it has not attained its projected capability (Clifford, 1981).
10
The proposed methane sensing project will contribute to future R&D in
methane sensors for Malaysian natural gas pipelines, domestic actuators and an
environmental monitoring system to reduce global warming. The usage of tin in
the form of stannic oxide will introduce an alternative use of the local tin and
promote a diversified Malaysian economy.
1.3
Scope of study
The knowledge of gas sensors has led to high-volume applications which
are publicised via periodic international sensor conferences which are devoted to
fundamental research, for example Transducers/Eurosensors, Semiconductor Gas
Sensors (SGS), Pittsburgh Conference (PITTCON), Electrochemical Gas Sensors
and also a book series covering the state of sensor chemistry, physics and
technology (Gopel et al., 1990). Much work has been done in the field of
methane sensing, for example thick film based on SnO2 (Carotta et al.,1991) in
relating the sensitivity to methane. The performance of the sensor material
depends strongly on its composition and preparation conditions (de Angelis and
Roberto, 1995). It was intended that the material chosen was SnO2 and takes the
form of a sintered pellet. The sensor element that comprised SnO2 only has
limited sensitivity to CH4 (Williams, 1987). The incorporation of noble metals is
to enhance the sensitivity (Yamazoe et al., 1983).The use of additives like Pt was
also reported to enhance sensitivity and selectivity and to reduce the response
time and operating temperature of the sensing material in the form of thin film
(Wu et al., 1993; Schierbaum et al., 1991; Zakrzewska et al., 1997 and Atashbar
et al., 1998). Gélin and Primet (2002) reported that Pt as a catalyst has
advantageous over Pd with respect to methane sensing. Firstly, Pd appeared to be
more sensitive to poisoning than Pt and secondly Pt could fully and easily
restored while the deactivation of Pd was irreversible. Therefore, a great interest
in Pt-SnO2 ceramic to achieve a considerable degree of performance for methane
sensing in air was anticipated mainly due to the catalytic behaviour of Pt.
11
The electrical properties of the ceramic semiconductor Pt-SnO2 was
studied using impedance spectrometer and Sensor Element Characterization
System (SECS) which are home-made at the Warwick University, United
Kingdom and Universiti Teknologi Malaysia respectively. The measurements
made were electrical resistance, conductance and sensitivity of the sensor
element. The microstructure of the sensor element was studied using High
Resolution X-Ray Diffraction (HRXRD), Energy Dispersive Analysis using XRay (EDAX), Scanning Electron Microscopy (SEM), Transmission Electron
Microscopy (TEM), Atomic Force Microscopy (AFM), X-Ray Photoemission
Spectroscopy (XPS), Nuclear Magnetic Resonance spectroscopy (NMR), Fourier
Transform Infra-red spectroscopy (FTIR), Raman-Shift spectroscopy and
Mössbauer spectroscopy. The physical properties of the ceramic was gauged via
Vickers hardness, bulk density, porosity and elastic modulus measurements.
Brunauer-Emnett-Teller (BET) method was employed to determine the specific
surface area and the particle size of the ceramic whilst the Barrett-Joyner-Halenda
(BJH) method is for calculating pore size distributions The highlight of the
research was to relate where ever possible the sensitivity and the microstructure
properties. It is reasonable to expect that if the same microstructure is present in
the fabrication of a similar ceramic, then it will show the similar corresponding
sensitivity.
1.4
Statements of hypotheses
The hypotheses made are as follows;
1. The usage of Pt-SnO2 would form a stable and sensitive material for
methane sensing in air via the dry pressing method,
2. The amount of Pt in SnO2 sintered will be minimized in obtaining the
optimum composition of the ceramic and the optimum operating
temperature of the gas sensor.
12
1.5
Objectives of the study
The objectives of this study are;
1. To construct and improve a new sensing probe in the acquisition of
electrical measurements,
2. To determine the optimum composition and optimum operating
temperature of the methane sensing,
3. To determine the sensor resistance, conductance and sensitivity with
varying ceramic compositions, operating temperatures of the methane
sensor and flow rates of the methane gas,
4. To determine the physical and microstructural properties of the
Pt-SnO2 ceramics.
1.6
Thesis plan
This thesis comprises of nine chapters. In the introduction, the state of gas
sensing is briefly mentioned with an emphasis on methane sensing with respect to
global warming and perils of gas explosion. The notions of methane sensing as a
research project with regards to Malaysian resources are highlighted. The
research tools and expectations are stated.
The second chapter deals with methane sensor viewed through the work
by researchers in the last four decades, since the birth of the gas sensor in the
1960s. Various methane sensors are mentioned including the stannic oxide based
sensors. The introduction of various dopants and their effects on the performance
of the gas sensor are also mentioned. The problems that arise from the sensors are
also pointed out.
13
The third chapter mentions the theory of sensing mechanisms known up to
the time this thesis was written. These include the well known Spillover and
Energy Barrier Model. The role of Pt in SnO2 in the sensing property is also
highlighted.
The fourth chapter states the experimental and measurement techniques
which include sample preparation and the apparatus used for both electrical and
microstructure analysis. The parameters and physical measurements are defined.
The fifth chapter presents the results of the electrical analysis which are
basically the measurement of sensor element resistance, conductance and
sensitivity. The variables in the experiments are sintering temperature and the Pt
loadings in the SnO2 ceramics. The effects of methane gas concentration and of
methane gas flow rate are also reported. The stability of the methane is viewed
via its long term performance.
The sixth chapter highlights two important parameters which will gauge
the performance of the methane sensor, namely optimum composition of the
ceramics and optimum operating temperature of the sensor with respect to
methane sensing in air. These two parameters are then related to the mean
crystallite size of the ceramics.
The seventh chapter looks at the response and recovery times of the
methane sensor which are influenced by methane gas concentration, operating
temperature of the sensor, doping level and the flow rate of the methane gas.
The eighth chapter deals with microstructure analysis of the ceramics.
The ceramics analysed using HRXRD, EDAX, SEM, TEM, XPS, NMR,
Mössbauer, Raman-Shift and FTIR will lead to the understanding of the
microstructure observed. The physical properties in relation to its stability are
gauged via DTA, TGA analysis and measurements of density, porosity, BET and
elastic modulus.
14
The ninth or final chapter summarizes the findings and comments on the
Pt-SnO2 ceramics in relation to methane sensing in air. Recommendations for
further work are also mentioned.
15
CHAPTER 2
GAS SENSORS REVIEW
2.1
OVERVIEW
The earliest commercial gas sensor comprised of a hot platinum wire
(Arai et al., 1986). Basically, the platinum filament is maintained at several
hundred degrees centigrade and via catalysis it will detect any combustible gas in
air. Such a phenomenon will cause a rise in temperature in the platinum wire and
a measurable increase in resistance. For example, 1000 ppm of isobutene will
show an output signal of a few millivolts via a resistance bridge circuit. The
sensor was improved by introducing a catalyst such as palladium as done by
Baker (1959). The output signal increased to a magnitude of 15 –20 mV.
However, the setback was that the catalyst deteriorated with time but modern
versions of the sensor known as pellistors and are still being used in coal mines
until today (Ihokura and Watson, 1994).
One remarkable development was to utilize the resistance change in a
metallic oxide semiconductor which results from gas adsorption (Tarama,1958).
This was realised in thin film zinc oxide gas sensors and the phenomenon is
known as the semiconductor catalysis mechanism (Seiyama, 1962). Seiyama et
al. (1962) reported on a gas detector which is based on the adsorption and
desorption of gases which cause a change in electrical conductivity of
semiconductors. The inclusion of a noble metal in the base material can improve
16
the sensitivity. The result is a patented stannic oxide gas sensor and the world’s
first commercial device led by Taguchi (1962).
In the 1970s and 1980s the search was on for the materials as a priority
and the effects of additives for enhancing the gas sensor sensitivity. The
selectivity of a gas sensor was also improved by changing the sensor temperature
between limits (Murakami, 1982). A gas identification system was introduced as
a result of combined sensors having different characteristics (Ikegama, 1985).
Another method to approach selectivity involved surface modification of the base
metal oxide with hydrophobic groups, calcium oxide, super-corpuscles of gold,
zinc oxide, sulphur or lanthanum oxide (Kanefusa et al., 1985; Fukui, 1985;
Nakahara, 1987; Haruta et al., 1989; Maekawa et al., 1990 and Matsushima et al.,
1989). The approach operates different from valence control (as in the case of
additives such as antimony or vanadium) or catalysis (as in the case of additives
of noble metal). This led to further refinement in the quality of the gas sensor
where efforts were made to detect gases (such as H2S, NOx, mercaptans, air
pollutants) at low concentrations (Satakae et al., 1989). Exotic applications of gas
sensors are for the sensing of trimethylamine (Takao, 1989) for testing the
freshness of fish, cooking gases (Shiratori, 1985) in the control of microwave
ovens, leaking freon and other refrigerants (Nomura, 1990) in evaluating the
ozone destruction in the upper atmosphere.
From the 1960s until 2001, the groundwork was followed to produce a
series of commercial gas sensors covering a very wide variety of applications
(Table 2.1). Table 2.2 shows the development from late 1990s until 2001
focussing on the stannic oxide based gas sensor.
17
Table 2.1: Development in the research field of gas sensors, 1960s – 1980s.
(Ihokura and Watson, 1994).
No.
1
2
3
4
5
6
7
Material
ZnO (thin film)
SnO2
SnO2+ Pd, Pt, Ag
ZnO, SnO2
SnO2+Al2O3
WO3 + Pt
In2O3 + Pt
8
9
10
11
12
13
14
15
SnO2 + SiO2
ZnO + Pt + Ga2O3
SnO2 + Pd
La1-x Srx CoO3
V2O5 + Ag (thin film)
ZnO + Ga2O3 + Pd, Pt
TiO2
CoO
16
17
18
19
20
21
22
23
Pd/CdS
SnO2 + Pd + ThO2
γ-Fe2O3
Co3O4
Ag2O
Pd/TiO2
Metal-phthalocyanine
Anthracene
24
25
26
27
28
29
30
ZnO (thin film)
SnO2-ultra fine particle
α-Fe2O3
ZnO + V2O5 + MoO3
ZnO (single crystal)
SnO2 (thin film)
SnO2 (thin film)
Detecting gas
H2, alcohol
combustible gas
combustible gas
reducing gas
combustible gas
H2, N2H4, WH3,
H2S
H2, hydrocarbon
combustible gas
CH4, NH3
propane
alcohol
NO2
H2, CO,
hydrocarbon
O2
O2
H2
CO
propane
CO
mercaptan
H2
NO2
amine, carboxylic
acid
alcohol
combustible gas
CH4, H2
freon
CO
combustible gas
combustible gas
Researcher/s
Seiyama, et. al.(1962)
Taguchi (1962)
Taguchi (1963)
Seiyama, et. al. (1966)
Taguchi (1966)
Shaver (1967)
Loh (1967)
Taguchi (1969)
Bott, et al. (1971)
Seiyama, et al. (1972)
Sakurai, et al. (1975)
Nakagawa, et al. (1975)
Ichinose, et al. (1975)
Tien, et al. (1975)
Logothetis, et al. (1975)
Steele, et al. (1976)
Nitta, et al. (1977)
Matsuoka, et al. (1978)
Stetter (1978)
Subomura, et al. (1978)
Subomura, et al. (1978)
Sadaoka, et al. (1978)
Suzuki, et al. (1978)
Heiland, et al. (1979)
Abe, et al. (1980)
Nakatani, et al. (1981)
Shiratori, et al. (1981)
Jones, et al. (1982)
Chang (1982)
Sotomura, et al. (1982)
18
Table 2.2: Development in the research field of stannic oxide based gas sensorslate 1990s until 2001.
No.
1
Topics/Acitvity
Effects of gas diffusivity and reactivity on
sensing properties of SnO2
Researcher/s
Shimizu et al. (1998)
2
Computational approaches to the chemical
sensitivity of semiconducting tin dioxide
Rantala, et al. (1998)
3
New design of an SnO2 gas sensor on low
temperature cofiring ceramics
Teterycz, et al. (1998)
4
Transient response of thick-film tin oxide
gas sensors multicomponent gas mixtures
Llobet, et al. (1998)
5
SnO2-based thick-film-resistive sensor for
H2S detection
Malyshev & Pislyakov
(1998)
6
SnO2-TiO2 solid solutions for gas sensors
Radecka, et al. (1998)
7
Electrical and spectroscopic
characterization of SnO2 and Pd-SnO2
thick films studied as CO gas sensors
Chiorino, et al. (1998)
8
In-situ EXAFS analysis of the local
environment of Pt particles incorporated in
thin films of SnO2 semiconductor oxide
used as gas sensor
Gaidi, et al. (1998)
9
The influence of molybdenum on the
properties of SnO2 ceramic sensor
Ivanovskaya, et al. (1998)
10
Pd-doped SnO2 thin films deposited by
assisted ultrasonic spraying CVD for gas
sensing: selectivity and effect of annealing
Briand, et al. (1998)
11
H2 selective gas sensor based on SnO2
Katsuki & Fukui (1998)
12
Electrical and CO gas sensing properties of Haeng Yu & Man Choi
ZnO-SnO2 composites
(1998)
13
Micromachined thin film SnO2 gas sensors
in temperature pulsed operation mode
Jaegle, et al.(1999)
19
No.
14
Topics/Acitvity
Improvement of gas-sensing properties of
SnO2 by surface chemical modification
with diethoxydimethylsilane
Researcher/s
Wada & Egashira (1998)
15
High ethanol sensitivity in sol-gel derived
SnO2 thin films
Varghese, et al. (1999)
16
Humidity insensitive thick film methane
sensor based on SnO2/Pt
Licznerski, et al. (1999)
17
Selectivity improvement of SnO2 films by
superficial metallic films
Sauvan & Pijolat (1999)
18
Selectivity enhancement of SnO2 gas
sensors: simultaneous monitoring of
resistances and temperatures
Heilig, et al. (1999)
19
Preparation and characterization of SnO2
and MoOx-SnO2 nano-sized powders for
thick film gas sensors
Chiorino, et al. (1999)
20
New method to obtain stable small-sized
SnO2 powders for gas sensors
Cirera, et al. 1999)
21
Ageing and p-type conduction in SnO2 gas
sensors
Ionescu, R. (1999 )
22
A SnO2 microsensor device for SUB-PPM
NO2 detection
Cobianu, et al. (1999)
23
MCM-41 modified SnO2 gas sensors:
sensitivity and selectivity properties
Li & Kawi (1999)
24
Surface chemistry of SnO2 and MoOxSnO2 nano-sized powders and electrical
responses of the related thick films
Chiorino & Ghiotti (1999)
25
Electrical properties under polluting gas of
Pt and Pd doped polycrystalline SnO2 thin
film
Matko, et al. (1999)
26
Integrated solid-state gas sensors based on
SnO2 (Pd) for CO detection
Renault, et al. (1999)
27
Relationships between sensitivity, catalytic Li, et al. (1999)
activity and surface areas of SnO2 gas
sensors
20
No.
28
Topics/Acitvity
Role of water vapour in the interaction of
SnO2 gas sensor with CO and CH4
Researcher/s
Ionescu, et al. (1999)
29
Influence of oxygen concentration in the
carrier gas on the response of tin dioxide
sensor under hydrogen and methane
Toumier & Pijolat (1999)
30
Electrical and CO gas sensing properties of Yu & Choi (1999)
ZnO/SnO2 hetero-contact
31
CO2-sensing characteristics of SnO2 thick
film by coating lanthanum oxide
Kim, et al. (2000)
32
Determination of the Pd content in Pddoped SnO2 films
Laureyn, et al. (2000)
33
Light enhanced gas sensing properties of
indium oxide and tin dioxide sensors
Comini, et al. (2000)
34
Improvement of humidity dependence in
gas sensor based on SnO2
Fukui & Katsuki (2000)
35
Computational studies for the
interpretation of gas response of SnO2
(110) surface
Rantala, et al. (2000)
36
Gas chromatographic study on adsorption
selectivity of tin dioxide gas sensor to
organic vapours
Wang, et al. (2000)
37
The effect of dopants on the electronic
structure of SnO2 thin film
Liu, et al. (2000)
38
The temperature change in SnO2 based gas
sensors during sensing reaction
Liu, et al. (2000)
39
Gas-sensitive properties of nanometerseized SnO2
Pan, et al. (2000)
40
Investigation of a new catalytic
combustion type CH4 gas sensor with low
power consumption
Sun, et al. (2000)
41
Comparison study of SnO2 thin and thick
film gas sensors
Lee, et al. (2000)
42
Sprayed SnO2 thin films for NO2 sensors
Leo & Rella (2000)
21
No.
43
Topics/Activity
Effects of surface modification with
platinum and ruthenium on temperature
and humidity dependence of SnO2-based
CO gas sensors
Researcher/s
Morimitsu, et al. (2000)
44
Detection and differentiation of CH4
hydrocarbon isomers over the Pd-SnO2
compressed powder sensor
Bulpitt & Tsang (2000)
45
Effect of particle size and dopant on
properties of SnO2-based gas sensor
Zhang & Liu (2000)
46
Analysis of the noble metal catalytic
additives introduced by impregnation of as
obtained SnO2 sol-gel nanocrystals for gas
sensors
Cabot, et al. (2000)
47
Material properties and the influence of
metallic catalysts at the surface of a highly
dense SnO2 films
Wollenstein, et al. (2000)
48
Reactivity of SnO2-CuO nanocrystalline
materials with H2S
Pagnier, et al. (2000)
49
PECVD prepared SnO2 thin films for
ethanol sensors
Hellegouarch, et al.
(2000)
50
CH4 sensing characteristics of K-, Ca-, Mg
impregnated SnO2 sensors
Choi & Lee (2001)
51
Comparison of CO-gas sensing
characteristics between mono and multilayer Pt/SnO2 thin films
Kim, et al. (2001)
52
Pd and Pt-SiC Schottky diodes for
detection of H2S and CH4 at high
temperature
Kim, et al. (2001)
53
A study on thin film gas sensor based on
SnO2 prepared by pulsed laser deposition
method
Kim, et al. (2001)
54
SnO2 sol-gel derived thin films for
integrated gas sensors
Cobianu, et al. (2001)
55
Chemical diffusion of oxygen in tin
dioxide
Kamp, et al. (2001)
.
22
No.
56
Topics/Activity
Surface state trapping models for SnO2based microhotplate sensors
Researcher/s
Ding, et al. (2001)
57
The properties of strongly pressed tinbased gas sensors
Kocemba, et al. (2001)
58
The influence of Rh surface doping on
anomalous properties of thick-film SnO2
gas sensors
Licznerski, et al. (2001)
59
Conductivity of SnO2 thin films in the
presence of surface adsorbed species
Kissine, et al. (2001)
60
Thin film sensors SnO2-CuO-SnO2
sandwich structure to H2S
Yuanda, et al. (2001)
61
Effect of alumina addition on methane
sensitivity of tin dioxide thick films
Saha, et al .(2001)
62
Selective CO gas detection of SnO2Zn2SnO4 composite gas sensor
Moon, et al. (2001)
63
Cerium oxide/SnO2 based semiconductor
gas sensors with improved sensitivity to
CO
Khodadadi, et al.
(2001)
2.2
Methane gas sensing
The sensor element is the active part of a gas sensor. A complete
instrument using such a sensor element can take the form of a gas monitor, gas
detector, gas alarm or gas actuator. Table 2.1 on page 17 shows a variety of
sensor element for sensing different gases. For methane sensing, the sensor
element is usually a metal oxide such as SnO2 (Carotta et al., 1991), Fe2O3
(Malyshev et al., 1994), Ga2O3 (Fleischer and Meixner, 1995), BaSnO3 (Ostrick
et al., 1996) or ZnO (Gruber et al., 2003). Amongst these the most reported
metal-oxide methane sensor is the SnO2 based. The behaviour of this type of
sensor element is characteristic of many sintered, thin or thick film metal oxide
semiconductors (Hagen et al., 1983). There are other methods of detecting
23
methane reported such as pellistors (Jones, 1987), semistor (Williams and Coles,
1999), planar heat-wire (Quan et al., 2000), Pd-SiC or Pt-SiC Schottky diodes
(Kim et al., 2001) and Thermal Wave Pyroelectrics Film(TWPFS) (Dorojkine,
2003). There are also commercial methane sensors such as Figaro sensors such as
TGS 813 and TGS 842 studied by (Pascale et al., 1995) for methane detection in
domestic premises.
2.2.1
Pellistors
This is also known as a catalytic gas sensor and this is the earliest form of
methane sensor developed in the 1960s. It comprised of thin Pt wire coils
surrounded by a catalyst which is supported on alumina. The flammable gas such
as methane is oxidised at the catalytic surface at an elevated temperature and this
shows as an increase in the resistance of the coil. The operating temperature of
this type of sensor is typically 700-1300 K (Jones, 1988).
2.2.2
Semistors
This is similar to that of the pellistor but is semiconductor-based in which
the alumina is replaced by a polycrystalline gas-sensitive ceramic such as stannic
oxide as reported by Komatsu et al. (1984), Orlik et al. (1993) and Fukui (1991).
However, the sensor-operating mechanism was not clearly defined (Williams and
Coles, 1999).
24
2.2.3
Planar heat wire
This was developed by Kokuen et al. (1990) to detect i-C4H10. The planar
heat wire was incorporated on an Al2O3 substrate by RF sputtering. The resistance
of the wire at a given temperature will be different in air and methane.
2.2.4
Pd-SiC or Pt-SiC Schottky diodes
Lundstrom et al. (1975), Arbab et al. (1993) and Kang et al. (1994) used
Schottky diodes in a gas sensitive device using silicon technology. Methane
dissociates on the catalytic surface and hydrogen diffuses through the metal to the
oxide interface thus forming an electric double layer on the internal surface of the
electrode. This in turn causes a change in the electrical properties of the device
when methane is being adsorped. In the case of Pt-SiC, the methane can
dissociate on Pt metal at high temperature (Arbab et al., 1993).
2.2.5
Thermal Wave Pyroelectrics Film (TWPF)
The sensor substrate is an Al2O3/SiO2 ceramic or sitall. The pyroelectric
detectors which are organic polycyclic molecular microcrystals and belong to the
P1 group are deposited onto the substrate. This will give rise to both
pyroelectricity and piezoelectricity (Nye, 1957). When an electric current flows
through the heating strip, the substrate will be heated steadily and the thermal
wave will propagate into the bulk of the substrate. This material generates
electrical current when its temperature changes. The sensor is capable of
measuring the relative thermal conductivity (which is temperature dependent) of
methane with respect to air and the TWPF studied by Dorojkine (2003) was able
to measure rather low methane concentrations.
25
2.2.6
Metal oxides
The metal oxide gas sensors for methane sensing can take the form of thin
or thick film or sintered pellet. As already mentioned, the metal oxide alone has
limited sensing ability. The metal oxides are usually semiconductors and
problems like selectivity, sensitivity and stability still persist. Nevertheless, efforts
are being made to overcome the problems. For example, in the case of Ga2O3, the
sensor was operated at high temperatures between 600-900oC (Flingelli et al.,
1998). At high temperature, the carrier mobility is not influenced by the grain
boundaries as other types of conduction prevail (Fleischer, 1993). Also, at such
elevated temperature, most organic vapours burn more readily than methane
(Fleischer and Meixner, 1995; Reti et al., 1995). The novel doping of SnO2 in
Ga2O3 enhanced the sensitivity and conductivity as reported by Frank et al.
(1998). This is due to the influence of Sn4+ as a donor which fits the available
lattice sites leading to improved gas-sensitive electrical properties.
2.3
Stannic oxide as a sensing element
This is the most widely used metal oxide for methane sensing and can take
the form of thin or thick film or sintered pellet. A typical method of producing
stannic oxide is by reacting aqueous SnCl4.5H2O with an aqueous hydrazine
monohydrate, N2H4.H2O (Popescu-Amalric, 1999). The reaction proceeds as
follows:
SnCl4.5H2O + 2N2H4 Æ SnO2 + 4HCl(g) +2N2 +4H2 + 3H2O.
Another typical method is by pouring aqueous solution of HNO3 on highpurity metallic Sn powder (Donaldson and Fuller, 1968). The reaction produces
NO2 gas which is reddish-brown and stannic oxide which is a white precipitate.
26
A new way of producing stannic oxide is microwave processing (Cirera et
al., 2000). In this method microwave radiation is applied to a saturated solution
of tin chloride in methanol. Rheotaxial growth thermal oxidation (RGTO)
involves two steps: first, by means of dc sputtering, several metallic tin layers are
deposited on a substrate which is maintained at a temperature higher than the
melting point of tin. These layers are then thermally oxidized (Dieguez et al.,
2000). Other methods include rf magnetron sputtering (Lee et al., 2001), atomic
layer epitaxy deposition (Utriainen, 2000), sol-gel (Morazzoni et al., 2001),
mechanochemical processing (Cukrov et al., 2001), PECVD (Hellegouarc’h,
2001) and atomic-layer chemical vapour deposition (Rosental, 2001). The stannic
oxide produced by the methods mentioned above range from microns to submicrons to nano-sized particles. Commercialized stannic oxide is produced by
Fluka-Chemika, Riedel-de Haën and Aldrich Chem.
The performance of SnO2 alone as a gas senor element is rather limited.
An effective method to improve its performance can be achieved by doping. The
doping species can be either catalytic (for example Pd and Pt), electroactive (In,
Sb, Cu, Ni and Mn) or stabilizer (Nb, Al and Si) (Kohl, 1990; Yamazoe, 1991; Xu
et al., 1991 and Fliegel et al., 1994). Catalytic dopants increase the rate of
reaction on the surface of SnO2 grains due to spill-over effect or modification of
surface energy states. An electroactive doping metal, which acts as an acceptor
contribute to the formation of an energy barrier at the grain boundary. It does so
by changing the electron concentration in the bulk of the tin dioxide due to a
compensation mechanism. Stabilizing dopants inhibit grain growth especially in a
nanocrystalline material. Other reported metal dopant used in SnO2 are Ru, Ag,
Bi, Ca, Fe, Au and Cd (Egdell et al., 1996; Kim et al., 1997; Zhou et al., 1997;
Nomura et al., 1997; Chaudhary et al., 1998; Tianshu et al., 1999 and Gaidi et al.,
2000). Metal oxide dopants like Bi2O3, Sb2O3, CeO2, CuO, Ag2O, TiO2 and
MoO3 are also reported to be incorporated in SnO2 (Devi et al., 1998; Teterycz et
al., 1998; Malyshev and Pislyakov, 1998; Radecka and Zakrzewska, 1998 and
Ivanovskaya et al., 1998) for gas sensing.
27
The use of dopant in SnO2 is to make the sensing element sensitive to a
particular gas and it is selective against interfering gases. For such tasks in
particular methane sensing, a high Pt content in SnO2 enhances the sensor
response to CH4 rather than CO (Khodadadi et al., 2001).
2.4
Problems and improvements with SnO2-based methane sensor
Although the SnO2 based methane sensor is one of the earliest methane
sensor and is widely used the sensitivity and selectivity need to be further
improved (Watson et al., 1993; Gopel and Reinhardt, 1996 and Zhou et al., 1997).
As mentioned in Section 2.3, an effective way to overcome this is by metal or
metal-oxide doping. Operating the gas sensor at higher temperature makes it
more selective to methane (Khodadadi et al., 2001). This is because CH4 gas is
thermodynamically the most stable alkane, more stable than most other reducing
gases and relatively difficult to sense at temperatures lower than 350-400oC. The
SnO2-based gas sensor is more sensitive to interferent gas such as ethanol vapour
in a domestic ambience and has caused false alarms. In such a case, the methane
selectivity and sensitivity can be enhanced by using a filter (SiO2, Al2O3)
containing noble metal catalysts (Dutronc et al., 1993; Papadopoulos et al., 1996
and de Angelis and Riva, 1995). The gas filter hinders the interfering gas from
reaching the sensitive layer and transforms the gas mixture to be detected into a
mixture which favours the detection of the gas to be detected or target gas (Madou
and Morisson, 1989). Cross-sensitivity in stannic oxide based sensor can be
decreased using thick-film (Corcoran et al., 1993; Fraigi et al., 1994 and Ménil et
al., 1983) or differential conductivity sensors (Dutronc et al., 1993 and Logothetis
et al., 1986). The common cross sensitivity is between CO and CH4 due to their
same reducing character. Research using signal post processing to distinguish
between these two gases use arrays or signal treatments algorithms (Ortega et al.,
2000; Hahn et al., 2000 and Capone et al., 2000). Niranjan et al. (2001) reported
that amorphous thin-film oxide has an enhanced selectivity and sensitivity as
compared to pure tin oxide. The amorphous thin film is rough which provides a
28
good adsorption site and exhibit a highly porous structure. Chaudhary et al.
(1999) improved the selectivity and sensitivity of tin oxide pellets by surface
modification using Ru. In this method, the physical and chemical properties of tin
oxide altered due to misfits created at the grain boundaries. This appears as large
number of grains and is due to the presence of Ru species. Methane sensitivity in
a stannic oxide based sensor can also be achieved by using alumina in admixture
with SnO2 and without doping in SnO2 lattice. This is due to the electronic
interaction between the semiconducting tin oxide grains and the Lewis acid sites
of the alumina grains in close contact (Saha et al., 2001).
One of the main disadvantages of a chemical gas sensor based on stannic
oxide is its strong dependence on humidity (Fleischer, 1996). Bârsan et al.
(1999a) stated that the presence of water increases the sensitivity towards CO and
decreases the sensitivity towards hydrocarbons. It is assumed that tin oxide
sensors work through the oxidation of the analyte, in this case CO. The resulting
loss of ionsorbed oxygen on the sensor surface results in an increase in
conductivity. This will be observed or read out electrically. This problem is not
totally eradicated by using dopants (Sayago et al., 1995; Debeda et al., 1996 and
Bauer et al., 1997). Solutions are sensor arrays with signal processing methods as
reported by Huyberechts et al. (1997) and the usage of a new cerment
composition based on stannic oxide and platinum black (Pt-black) (Licznerski et
al., 1999).
Choi and Lee (2001) reported that Ca in stannic oxide particle plays an
outstanding role of crystallite growth stabilization especially above 350oC. This
gives rise to high sensitivity when the crystallite size is twice the depth of space
charge which is approximately 6 nm (Ogawa et al., 1982; Xu et al., 1991 and
Yamazoe, 1991). This led to the usage of nano-sized SnO2 and dopants like the
one reported by Fau et al. (2001).
In the case of the catalytic methane sensor, poisoning of platinum surfaces
by hexamethyldisiloxane (HMDS) was reported by Ehrhardt et al. (1997). The
29
HMDS is a molecule-(CH3)3-Si-O-Si(CH3)3 and inhibited by silicon-containing
compounds (Cullis and Willatt, 1984; Gentry and Jones, 1984 and Sommer et al.,
1992). Exposure to minute amount of HMDS results in total poisoning of
methane oxidation and the decomposition of HMDS is strongly catalyzed by
transistion metals as reported by Cullis and Willatt (1984). As this type of sensor
works in the high temperature range of 700-1300 K, it is widely accepted that the
catalytic oxidation obeys the Langmuir-Hinshelwood mechanism involving
reactions between adsorbed species on the surface (Trimm and Lam, 1980; Lintz
et al., 1962 and Cullis et al., 1970). At 873 K, Gentry et al. (1978) claimed the
sensors are irreversibly poisoned towards catalytic oxidation of methane. Also as
methane oxidation uses high energy adsorption sites, it is strongly influenced by
the interaction with HMDS and this will give rise to surface polyorganosiloxanes
or silica. Cullis et al.(1984) and Gentry and Jones (1986) also studied poisoning
of catalytic methane sensors giving rise to chemical products with molecules
containing sulphur, halogen and silicone. The poisoning effect is also a factor
which cause ageing in the sensor. In normal practice, this needs scheduled
replacement which makes it a relatively high cost of ownership.
2.5
Current research trend on SnO2-based methane sensor
Barsan et al. (1999b) commented that metal oxide in particular SnO2 have
attracted the attention of many users and scientists who are interested in gas
sensing under atmospheric conditions. However, the Research and Development
(R&D) of such sensors still leaves researcher with very poor basic understanding
of practically useful gas sensors. This is because the understanding is based on
different models due to different approaches. The first approach is chosen by the
users of gas sensors. These users tested the phenomenological parameters of
existing sensors, and envisaged that there exist minimum parameter to describe
the sensor selectivity, sensitivity and stability. The second approach is chosen by
the developers who empirically optimized the sensor material preparations, test
structures, ageing, etc., for different applications and thus optimizing the sensor
30
technology. The third approach is chosen by basic research scientists, who
attempted to identify the atomistic processes of gas sensing by applying
spectroscopies in addition to phenomenological techniques of sensor
characterization. Eventually all the different approaches lead to the well-known
structural and pressure gaps between the ideal and the real world of surface
science.
The same trend of work done to improve selectivity, sensitivity and
stability exist and are foreseen to be ongoing, with an emphasis on thick films
with controlled nanocrystalline sizes and thin films with monolayers of
nanocrystals. However, the nanotechnology has its drawbacks; high cost for
material production and miniaturization of sensor, low yield and unstable
structure at temperatures > 400oC.
The basic understanding of SnO2-based sensors can be aided using
spectroscopies which are in-situ measurements such as infrared (IR), electron
paramagnetic resonance (EPR) and Raman (Willett, 1991 and Lenaerts et al.,
1995). Spectroscopies can also be done under ideal conditions such as
thermodesorption, X-ray photoemissions (XPS) and ultraviolet photoemissions
(Szuber, 1998 and Schweizer-Berberich, 1998). Atomic resolution may be
obtained from Scanning Tunnelling Microscope (STM) or Atomic Force
Microscope (AFM). This leads to better understanding of the molecule-SnO2
interaction. For example, Thermal Programming Desorption (TPD) and infrared
(IR) studies show that at temperatures above 200oC molecular water is no longer
present at the surface of SnO2 (Barsan, 1999b). Another example, IR studies can
identify CO-related species on SnO2 (Barsan, 1999b).
31
CHAPTER 3
SENSING MECHANISMS AND MODELS
3.1
SENSING PARAMETERS
It is customary that electrical measurements such as resistance and
conductance are reported in most gas sensing procedures. The electrical
conductance, G is defined as the reciprocal of the electrical resistance, R. Thus,
G = 1/R and its unit in S.I. is ohms-1(Ω-1) or Siemens (S). The sensitivity of a gas
is defined as the ratio of the change in conductance to the conductance in air.
Hence, the sensitivity S/ = (G –Go)/Go, where G is the conductance in the test gas
and Go is the conductance in air. Yamazoe et al. (1983) observed a distinct
maximum in sensitivity at a particular temperature which is dependent on both
dopant and the sensing gas. Another parameter is selectivity which is defined as
the ratio of the sensitivity of two gases, Š = S/gas1/S/gas2.
3.2
SENSING MECHANISMS
The sensing mechanisms in section 3.2.1 to section 3.2.3 are based on
conductance, section 3.2.4 – section 3.2.5 are based on sensitivity and sections
3.2.6. are based on selectivity. These gas sensing mechanisms are for
temperatures < 500oC.
32
3.2.1 DEFECT FORMATION
SnO2 is an n-type semiconductor with band-gap approximately 3.6 eV
(Kohl, 1989) and slightly non-stoichiometric, whereby the bridging oxygen in the
SnO2 is removed. The removal of bridging oxygen atom does not provide
additional donors and the non-stoichiometric configuration results in a stable
surface as the Sn ion has two stable oxidation states, Sn4+ and Sn2+. In ambient
atmosphere, oxygen vacancies are formed where the loss of bulk lattice oxygen
causes the generation of oxygen lattice defects and electron donor states (McAleer
et al., 1987). For sintered porous SnO2, the observable conductivity change is
very substantial, even in minute concentrations of combustible gas like methane
in a large excess of oxygen. Thus, it is assumed that the observable conductance
change is a surface controlled process, not in equilibrium with the bulk (McAleer
et al., 1987).
3.2.2 DEPLETION LAYER FORMATION
Windischmann and Mark (1979) suggested a two step process for n-type
semiconductor-electrical response to reactive gases. The first step is where
oxygen from the ambient adsorbs onto the surface of the sensing material and
extracts electrons from it as shown in the equation (Morrison, 1987a) :
O2 + e- ↔ O2-ads
(3.1)
O2 +2e- ↔ 2O-ads
(3.2)
with O- dominating at temperatures > 160oC. At this stage the sensor surface is of
high resistance. This initial high resistance in the sensor as observed by
Windischmann and Mark (1979) is due to the adsorbed oxygen and is essential for
a sensor response. On exposure to methane, the adsorbed methane reacts with the
oxygen species and electrons are injected back into the material, thus increases
the conductivity.
33
or
CH4 + O2- ↔ CH4O2 + e-
(3.3)
CH4 + O-ads ↔ CH4O + e-
(3.4)
In the case of an n-type semiconductor, the oxygen acts as an acceptor on
it and extracts electrons from the valence band. Thus, a positively charged double
layer or space charge layer appears just below the very negatively charged surface
(Figure 3.1). As the negative surface charge increases, the surface barrier
potential increases too, according to the following equation;
Vs = eNs2/2εoεrNi
(3.5)
where Vs is the surface barrier potential, e is the electronic charge, Ns is the
negative surface charge, εo is the permittivity of free space, εr is the relative
permittivity of the medium and Ni is the net density of ions in the space charge
region.
ES = eVS
EC
Ef
surface
EV
FIGURE 3.1: Formation of surface depletion layer in SnO2 . Ec is the conduction
band energy, Ef is the Fermi energy, Ev is the valence band energy and ES is the
surface barrier energy (Yannopoulos, 1987).
34
The surface state energy approaches the Fermi energy with increasing
band bending of the depletion layer. If there is a further increase in the amount of
adsorbate, the increment in Vs would raise the surface levels above the Fermi level
and the ion would lose their electrons. However, this cannot appreciably occur
and the Weisz limitation of the surface coverage is about 1012-1013 cm-2. The
corresponding band bending is about 1 eV and a Schottky depletion depth of 10-610-5 cm (Kohl, 1989).
3.2.3 POLYCRYSTALLINE MATERIALS
Sintered porous pellets or thick films are the typical fabrication for a gas
sensor. In this form the electrical resistance of the material depends very strongly
on gas adsorption (Moseley, 1992 and Morrison, 1987b). Figure 3.2 shows a
model of a typical compressed powder.
FIGURE 3.2:
Model of a typical compressed powder and Schottky barrier
formation (Morrison, 1987b).
35
In ambient atmosphere, oxygen is adsorbed on the exposed grains and
electrons are extracted from the near-surface region, thus forming a depletion
region. Electrons from the bulk must pass from grain to grain and cross a high
energy barrier at the intergranular contacts for conductance to occur. Thus, the
electrical resistance at this stage is fairly high. When methane is introduced,
electrons are released and injected back into the conduction band and the barrier
is lowered. This results in an increase in conductance. Lantto et al. (1988)
expressed the conductance, G = Goexp(-eVs/kBT), where G is the conductance in
methane, Go is the conductance in air and eVs is the barrier energy. Therefore, for
a porous powder sample, the Schottky barriers limit the conductivity at the
intergranular contacts.
For a well sintered material, the conductivity is determined by a fully
open and undepleted neck regions between the grains (Figure 3.3). Kiselev and
Krylov (1987) then expected that fully sintered polycrystalline or single crystal
will exhibit lower sensitivity than porous compressed powder as the conductance
of the former material depends on the reduction of the Schottky barriers.
FIGURE 3.3: Well-sintered polycrystalline-formation of undepleted neck
regions.
36
3.2.4
GRAIN SIZE EFFECTS
The sensitivity increases as the grain size decreases and approaches the
Debye length or the thickness of the depletion layer for stannic oxide (Xu et al.,
1991 and Yamazoe, 1991). In this case, the space charge region encompasses a
large fraction of the grain bulk, hence the electrical resistance is sensitive to the
adsorbed gas (Ippommatsu et al., 1991).
3.2.5
HIGH TEMPERATURE SENSING
For gas sensing at temperatures < 500oC, the characteristics of SnO2 based
gas sensors agree with the models already mentioned. However, it may be
possible that at high temperatures (> 500oC), gas sensing occurs via different
mechanisms. Sukharev (1993) reported that the reducing gas interacts directly
with the semiconductor which does not involve adsorbed oxygen at all.
3.2.6
CATALYST DOPING
The use of a catalyst in n-type semiconductor gas sensors is fundamentally
to accelerate the oxidation rate of the reducing gas, thus improving the response
time and enhancing the selectivity of the gas sensors (Morrison, 1987b;
Ambrazeviciene et al., 1993). To enhance the effect of catalytic doping on the
material, it is essential that the doping have a great influence on the electrical
resistance of the material, for example, it is accessible to the intergranular contact
regions between the grains. There are two processes in which it may be achieved
namely Fermi energy control and spillover.
37
3.2.6.1 FERMI ENERGY CONTROL
In this process, the adsorbing gas exchanges electrons with the supported
catalyst, which then exchanges electrons with the support (Yamazoe, 1991) as
shown in Figure 3.4. Yamazoe et al. (1983) reported for Ag-SnO2 or Pd-SnO2,
that the electronic sensing is characterized by a change of oxidation state of the
noble metal which acts as an electron donor or acceptor. The catalyst will control
the barrier if the concentration of surface states is greater than 1012 cm-2. Initially,
the Fermi level in the semiconductor is higher than that of the catalyst and
electrons are transferred from the semiconductor to the catalyst (Figure 3.5).
Thus, the Fermi energy is pinned and defines the surface barrier. The surface
depletion layer which is determined by the Fermi energy of the catalyst will
influence the intergranular contact resistance. If the catalyst is adequately
dispersed, the depletion regions will overlap and extend across the intergranular
contacts, thus dominating the resistance (Figure 3.6).
FIGURE 3.4: Effect of noble metal in SnO2– Fermi energy control (Yamazoe,
1991).
38
ECS
Ef
EV
FIGURE 3.5: Band model for catalyst and n-type semiconductor. Ecs is the
surface conduction band edge of the n-type semiconductor. EV is the valance
band and Ef is the Fermi energy.
FIGURE 3.6. Depletion region formed by a well-dispersed catalyst.
For Ag-SnO2, the AgO was reduced to Ag metal upon exposure to H2
(Yamazoe et al., 1983). The Fermi energy level shifted and an increase in surface
conductivity was observed. Thus, an additive replaces adsorbed oxygen in
controlling the Schottky barrier height and hence influencing the electrical
resistance of the gas sensor.
39
3.2.6.2 SPILLOVER MECHANISM
In this case, the reactive gas adsorbs itself to the catalyst, becomes
activated and spillsover onto the semiconductor surface (McAleer et al., 1988) as
shown in Figure 3.7. Sermon and Bond (1973) reported that Pt or Pd are the
most efficient at initiating spillover. McAleer et al. (1988) reported Pt or Pd
catalyst greatly modify the effects due to moisture in SnO2 gas sensors. In pure
SnO2, the hydroxyl (OH-) group was reported to dominate the surface when
exposed to moisture. Nevertheless, the ionosorbed oxygen will still dominate as
an acceptor due to increasing oxygen adsorption from spillover effects.
FIGURE 3.7: Spillover of oxygen and hydrogen on a catalyst doped
semiconductor.
For Pt-SnO2, Yamazoe (1991) reported the chemical sensing involved in
the change of the oxidation state of SnO2 and the role of Pt is the activation and
spillover of the sensing gas as shown schematically in Figure 3.8.
40
FIGURE 3.8: Chemical model of spillover mechanism of SnO2 doped noble
metal. (Yamazoe, 1991).
3.3 CURRENT STATUS ON SENSING MECHANISMS
Study on gas sensing still focuses on sensors operating at low temperature
(<500oC). Reasonable explanations and models for selectivity enhancement using
metal and metal oxide are still lacking in the published literature. High
temperature gas sensing has not been successful and the existing models fail.
3.4 METHANE DETECTION
Methane gas detection was extensively done by workers such as Heiland and
Kohl (1985) and Yamazoe and Miura (1992) using SnO2 thick film doped with Pt
and Pd and pure SnO2 thick film respectively. The doped samples were able to
detect methane at optimum temperatures between 300-500oC (Kohl, 2001). The
sol-gel route for obtaining SnO2 which was utilized for methane sensing was
reported to detect methane at optimum temperatures between 400-600oC (Feng et
al., 1993).
41
Activation of methane means excitation of the C-H bond as a consequence
of the formation of radicals. Thus, methane can be activated on the SnO2 surface
and forms CH3 radicals (Harber, 1984). The surface morphology does influence
the activation process and in the case of hydrogen, it is strongly and nearly
irreversibly absorbed on the surface of lattice oxygen of SnO2 as is a typical for
semiconducting oxides (Bond, 1987). In the case of pure SnO2 which is exactly
stoichiometric (Figure 3.9), it cannot chemisorbs oxygen.
FIGURE 3.9: Stoichiometric SnO2 (110) surface with bridging oxygen atoms on
the top. (Barsan et al., 1999b).
However, in most cases, pure SnO2 is non-stoichiometric (Figure 3.10) and
oxygen deficient and it can chemisorbs oxygen until the charge averaged over the
depth of the depletion layer is balanced. For n-type semiconductor like SnO2 the
concentration of the surface oxygen ions remains some order of magnitude lower
than the concentration of lattice oxygen in the uppermost surface layer. The
adsorbed superoxide O2- and O- species and the exposed oxygen atoms at the steps
of terraced surfaces are electrophilic reactants which preferentially attack the C=C
double bond of the adsorbates and abstract electrons. Hydrogen-containing gases
such as methane produces oxygen vacancies which are known to be donor.
42
FIGURE 3.10. Non-stoichiometric (reduced) SnO2 (110) surface with in-plane
oxygen only. (Barsan, et al., 1999b).
However, de Fresart et al. (1982) deduced from conductivity measurements, that
surface defects on SnO2 do not act as donor states. This fact is also supported by
ultraviolet photoemissions spectroscopy (UPS) (Sinner-Hettenbach, 2001) and
band structure calculations by Munnix and Schmeits (1987). However, surface
oxygen vacancies may diffuse towards the bulk and this can activate the vacancies
as donors. Kohl (1992), based on Lannto calculations, which show that the
vacancies have to cross only two lattice planes, found that the onset for activation
can be achieved at temperatures as low as 550 K.
The decomposition of hydrocarbon molecules on a catalytically active
metal is used in typical semiconductor-type sensors with metal cluster deposits.
The sensitivity is enhanced only by partial oxidation of the target gas on the metal
deposit and subsequent spillover of hydrogen to the semiconductor. The activity
of a catalyst depends greatly on its oxidation state. Prolonged runs at the
operation temperature and in the absence of the target gas will cause the
semiconductor gas sensors (for combustibles gas like methane) to `fall into sleep’
(Kohl, 1996).
43
The role of lattice oxygen and ionosorbed oxygen is demonstrated via the
reaction of methane in Ultra-High-Vacuum (UHV). In the thermodesorption
spectra, CH4 show a maximum value at 500 K (Thoren, 1985) and methane
dissociates to methyl and hydrogen;
CH4(g) ↔ CH3ads + Hads
(3.6)
It should be noted that even in the presence of gaseous oxygen, total oxidation
does not occur due to Weisz limitation of oxygen ion density (Weisz, 1953).
Methyl groups are then formed on top of the O ion and the methoxy group CH3Ocan be converted to a formate group. The formate group then decays to CO, CO2
and H, hence consuming the ionosorbed oxygen (Figure 3.11).
A
B
FIGURE 3.11: Reaction scheme: methane with oxygen on a sputtered SnO2 film
at 774 K. (Kohl, 2001) A; decay of methane on a surface by adsorbed oxygen.B;
decay of methane on a surface without adsorbed oxygen.
44
At temperatures well below 700 K, continuous exposure to methane will block
further methane adsorption and reduce conductance which are caused by stable
acetate-like species (Egashira et al., 1987).
The methane sensing mechanism of noble metal doped SnO2 based thick
film devices for hydrocarbons have been proposed by Morrison (1987a); Heiland
and Kohl (1988); Lee and Chung (1993) and Lee (1994). The mechanism
involves methane oxidation which yields CO2 and H2O via CHn or CHnO
intermediates (0 < n < 4) and the reaction is promoted by noble metals such as Rh,
Pt and Pd as reported by Pitchai and Klier (1986); Hicks et al. (1990); Ribeiro et
al. (1994) and Burch et al. (1995). The noble metal catalyzes the methaneabsorbed oxygen reaction on the semiconducting oxides, leaving oxygen
vacancies or conduction electrons behind which is supposed to explain the high
sensitivity observed in the noble metal doped semiconductors sensors. The
promoting effects of the noble metals are prominent and clear but the mechanisms
are not. Researchers like Dutronc et al. (1993), Semencik and Fryberger (1990),
Matsushima et al. (1989b) and Xu et al. (1990) which work on these sensors have
failed to provide direct evidence for the mechanisms involved. Kim et al. (1997)
suggested that the role of the noble metals in the sensor should be based on the
dispersion states of the noble metals and the interactions between the noble metals
and the support.
45
CHAPTER 4
EXPERIMENTS AND MEASUREMENT TECHNIQUES
4.1
INTRODUCTION
The experiments were basically categorized into electrical and
microstructure characterization. The main aims of the electrical characterization
were to determine the optimum operating temperature and optimum doping of Pt
in SnO2. The electrical properties studied include resistance, conductance and
sensitivity. The response and recovery times under the influence of various
factors like methane in air concentration and flow rate were also determined. The
microstructure characterization was emphasized on samples which exhibited the
optimum operating temperature and composition. The experiments used
techniques like HRXRD, SEM, TEM, BET, BJH, EDAX, FTIR, Raman Shift,
Mössbauer, XPS and NMR. The physical properties of the ceramic which include
density, porosity, Vickers hardness and elastic modulus were also investigated.
4.2
CERAMICS PREPARATIONS
The raw material tin oxide powder (purity 99.995%, product number;
20471-4) and platinum powder (purity 99.99%, product number; 748303)
obtained from Aldrich and Metalor (PUPT 518) respectively (Appendix A and
Appendix B). The SnO2 powder was sieved using a sieve (Endecotts) with an
46
aperture of 50 µm, mixed mechanically with Pt powder (with size ≤ 13µm) in a
beaker using a glass rod to give a nominal composition of SnO2 (100-x) Pt (x)
with x in the range 0 ≤ x ≤ 5 wt.%. The density of the pure SnO2 is 6.9 g/cm3 and
the density of the Pt is 21.4 g/cm3 (JCPDS, 1997a and 1977b). To ensure the
homogeneity of the mixture, the mixing of the powders was done for 15 minutes
until the mixture appears to have the same colouration throughout. The total
weight of a sample was (10.000±0.001) g. The weighing was done using an
electronic balance (Mettler AE 163). The powdered sample was pressed between
sheets of plastic in a dry pressing machine (Herzog) with a pressure of 40 MPa for
a duration of 5 minutes at room temperature. The dry-pressed sample was a
circular disc with a diameter of about 40 mm and a thickness of about 2 mm was
then transferred onto a ceramic plate and heated in an electric furnace to the
required temperature (100-1000oC) at a heating rate of 20oC per minute for a
sintering time of one hour. The sample was cooled down at the same rate and to
avoid contamination, only one sample was sintered at any one time. The
preparation of the ceramics is schematically shown in Figure 4.1. The sample can
be easily cut into any shape using a sharp knife. For example, for SEM analysis
the sample was cut into a disc of about 12 mm in diameter and thickness of about
2 mm and for electrical measurements the sample was cut into a 10X10X2 mm3
cuboids. The pure SnO2 is also referred to as the undoped sample and the doped
or modified sample have Pt incorporated in the SnO2. The differences between
the sintered and the green-body was that the former was much whiter and stronger
than the latter (buff-white and flaky). The green body was grey in colour due to
Pt but became white after sintering.
47
SnO2 powder sieved
using a 50 µm sieve.
SnO2 and Pt powder were weighted in nominal
composition of SnO2 (100-x) Pt (x) with x in wt % in
the range 0 ≤ x ≤ 5.
Mixed mechanically
Dry pressing at room temperature
in cylindrical mould ( ~ 40 mm diameter)
Pressure: 40 MPa
Time : 5 minutes
Sintering
Temperature : (100-1000)oC
Time : 1 hour
Ceramic sample
(Shape: Disc)
Diameter: ~ 40 mm
Thickness:~ 2 mm
Colour: White
FIGURE 4.1: Preparation of the Pt-SnO2 ceramics.
48
4.3
BULK DENSITY AND POROSITY
The bulk density of the ceramics was determined based on Archimedes
principle using a Precisa Model XT 220A. The sample was cut into
10X10X2 mm3 cuboids, weighed in air (W1) and then weighed in toluene (W2)
(after immersing in toluene for two weeks). Toluene was used as it does not react
chemically with the sample and a period of two weeks was long enough for the
toluene to penetrate the pores in the ceramic. The bulk density (Jones and
Bernard, 1991) was calculated using the formula;
ρ=
W1 ρ toluene
W1 − W2
(4.1)
where ρtoluene is the density of the toluene (0.8647 g/cm3).
The porosity (Humaidi, 2000) was calculated using the formula;
p/ =
4.4
W1 − W2
x 100 %
W2
( 4.2)
Sensor Element Characterization System (SECS)
This is a home-made equipment for measuring the electrical resistance of a
sensor element at operating temperatures between 30-450oC (Figure 4.2). The
SECS has already been used and described (John, 2001) as in Appendix C . The
changes made to the system are;
(i)
construction of a new probe which is much slimmer, with springloaded holder to accommodate the sample,
(ii)
data acquisition in the form of Microsoft EXCEL plotted (plot of
load resistance against time as in Appendix D),
49
The sample was cut into 10X10X2 mm3 and fitted snugly between the
spring loaded electrodes (Figure 4.3). Typical load resistances were between 100999 Ω with supply voltage between 5-20 V. The temperature controller
maintained temperatures between 30-450oC and mass-flow controller maintained
flow rates of synthetic air or methane between 100-500 standard cubic centimetre
per minute (sccm).
Electrical
measurement parts
gas chamber
methane supply
FIGURE 4.2: Sensor Element Characterization System (SECS).
electrodes
sample
2 mm
power supply
10 mm
Variable
Resistance
FIGURE 4.3: Configuration of sample between electrodes.
50
The sensor resistance, Rs (John, 2001) can be calculated using the formula;
R S = R L ( VC V L − 1 )
(4.3)
where RL is the load resistance, VL is the voltage across the load resistor and VC is
the voltage supply.
The conductance G is simply the reciprocal of Rs in equation (4.3), i.e.
G = 1/Rs
(4.4)
The sensitivity S/ is defined as;
S/ =
∆G
G0
x100% =
G methane − G air
x100%
G air
(4.5)
where ∆G is the change in conductance, Go is the conductance in air and Gmethane
is the conductance in methane.
4.4.1
Response and recovery time
The performance of the gas sensor is also gauged on its ability to respond
to the presence of methane. The response time is defined as the time taken for the
signal from its initial value to reach 90% of its maximum or saturated value. The
onset of the counting commences when methane is immediately introduced. The
signal can take the form of load voltage, sensor resistance, conductance or
sensitivity.
The recovery time is defined as the time from the maximum or saturation
value to return to its initial value. The counting starts immediately when the
51
methane was cut-off. Figure 4.4 shows schematically the determination of the
response and recovery time.
2.5E-04
recovery time
2.0E-04
response
time
1.5E-04
90% of
max.
value
1.0E-04
5.0E-05
0.0E+00
0
100
200
300
t400
(s)
500
600
700
800
FIGURE 4.4 : Schematic diagram for the determination of response and recovery
times. The quantity in the vertical axis can be VL,, RS, G or S/.
4.5
Impedance spectroscopy
The impedance spectrometer setup is shown in Figure 4.5a. It comprises
of an impedance analyzer which is capable of measuring the impedance of the
ceramic sample via a PC monitor. The ceramic sample was cut into a circular
shape of diameter of about 5 mm and a thickness of about 2 mm (Figure 4.5b)
which was placed between slabs of electrodes connected to the impedance
analyzer (Hewlett Packard 4291A RF) at working frequencies from 50 kHz up to
800 kHz. The ceramic sample was heated up in an electric oven which houses the
sample from room temperature to 450oC at a rate of 20oC per minute. The
temperature of the sample can be monitored via the pc display. The data obtained
from the run was saved in the form of an ASCII file. The resistance of the
52
sample, R was calculated using the relationship, R = Z cos θ/, where Z is the
impedance and θ/ is the phase angle.
To pc
(a) Setup for the impedance spectrometer
electrode
sample
(b) sample and oven
Figure 4.5: Impedance spectrometer at the Physics Department, University of
Warwick, England.
4.6 ELASTIC MODULUS
The Ultrasonic System for Mechanical Characterization of Material
(UMC) is a home-made system for measuring elastic modulus such as Young,
Bulk and Shear modulus and other related measurements include Poisson ratio,
compressibility via the velocities of both longitudinal and transverse wave
propagated in the sample.
53
4.6.1
The Ultrasonic Mechanical Characterization System (UMC)
Figure 4.6(a) shows the setup of a simplified pulse ultrasonic system of the
UMC setup.
Figure 4.6: Ultrasonic Mechanical Characterization System (UMC). (a) simple
pulse ultrasonic system, (b) envelope of pulse echo train, (c) echo as seen on
oscilloscope.
54
The quartz transducer which is piezoelectric, converts an electrical energy
into mechanical energy, and vice versa at ultrasonic frequency. When a radio
frequency signal from ultrasonic equipment arrives at the transducer, it causes the
quartz to vibrate at its fundamental resonance frequency ~10 MHz and this will
allow an ultrasonic plane wave to propagate in the sample. There are two types of
cuts, namely X-cut and Y-cut. The X-cut generates and receives longitudinal
waves and the Y-cut generates and receives shear or transverse waves. The
difference between these two, the Y-cut transducer is marked by a flat edge
perpendicular to the polarization direction. The acoustic waves transmitted by the
quartz transducer can only be launched successfully into the sample by physically
bonding the transducer to one face of the sample with Nonag stopcock grease.
The ceramic sample was immersed in castor oil prior to the run. The
samples as prepared have typical values of non-parallelism and misorientation of
less than 1x10-4 radians and 8.7x10-3 radians respectively. The pulse echo overlap
system verifies the sample parallelism, by ensuring that the echo train is
exponential as shown in Figure 4.6(b) and Figure 4.6(c). As an electromagnetic
acoustic transducer (EMATS) only generate on electrically conductive or
magnetic materials, a thin layer of Al foil (< 1 mm) was used to transmit the
energy from the EMAT into the ceramics. The foil was bonded with very thin
layer of Nonag stopcock grease.
The pulse echo overlap system used in this work extends the simple pulse
echo single ended system by means of overlapping any two selected visually
intensified echoes on the oscilloscope. The block diagram of the pulse echo
overlap system is as shown in Figure 4.7. A pulse of 10 MHz radio frequency is
generated by the pulse modulator and receiver unit as shown in Figure 4.7a. A
high resolution oscillator set itself generates a synchronizer square wave which is
used to power the decade divider and dual delay set as in Figure 4.7b. The
frequency of this square wave is similar to that of the sine wave produced by a
pulse modulator and receiver unit. This frequency is then reduced in the decade
divider and dual delay set by dividing it with a factor of 10 in order to trigger the
55
pulse modulator and receiver (Figure 4.7c). Each triggered pulse creates a radio
frequency burst which is in turn transformed by the quartz transducer into a
mechanical vibration within the crystal (Figure 4.7d). Longitudinal and shear
wave velocities (VL and VT, X-cut and Y-cut transducers plated with chrome gold
were used to convert the rf input pulses to longitudinal and shear ultrasonic
pulses, respectively) were determined from transit time of the ultrasonic waves
using the pulse-echo-overlap technique.
Figure 4.7: Pulse echo overlap waveforms
The velocity of the ultrasonic wave, v = 2lf , where l is the sample thickness and f
is the frequency of the superimposed echo ( Figure 4.8).
56
Transducer
Bonding Material
Sample
l
FIGURE 4.8: Acoustic wave propagation in sample.
4.6.2 Young Modulus
If the shear and longitudinal wave velocities are known, the ultrasonic
wave propagation in a solid can be used to determine the elastic constants of the
polycrystalline ceramics. It is assumed that plane and nearly plane longitudinal
and shear waves can exist with velocities,
v L2 = ( λ + 2 µ ) / ρ
(4.6)
vS 2 = ( µ / ρ )
(4.7)
where vL = longitudinal velocity
vS = shear velocity
ρ = density of the ceramics
µ = shear modulus
λ + 2µ = longitudinal modulus
In order to express the relationships amongst the strain or elongation force
and area, consider a block of initial length, L and cross sectional area A. If a force,
F pulls at the end of this block, the elongation is ∆L and the fractional elongation
is ∆L /L. This fractional elongation is directly proportional to the force and
inversely proportionally to the area A, ∆L /L = 1/Y . F/A; thus
57
Y=
(
)
F/A
2
2
= ρ 3v L − 4 v S / 3
∆L / L
(4.8)
where Y is the Young’s Modulus.
4.6.3
Bulk Modulus
The proportionality of strain and stress is also valid for compressional
deformation defined as the ratio of decrease ∆ V of the volume, V to the initial
volume and this fractional deformation is proportional to the force F pressing on
each face of the block and inversely proportional to the area A of the face thus,
∆ V/V = 1/B. F/A;
B=
(
)(
F/A
2
2
2
2
= ρv S 3 v L − 4 v S / v L − v S
∆A / A
)
(4.9)
where B is the bulk modulus.
4.7
Vickers Hardness
This tests the hardness which is a measure of resistance against damage or
permanent deformation. The resistance in this case is the resistance which the
sample offers to indentation by an indenter which is a pyramid-shaped diamond
under an applied load. The microhardness (when the load is in the order of
grammes) is the ratio of load to the surface area of the pyramid-shaped permanent
indentation (produced by the load). The Vickers hardness, HV can be calculated
using the following formula;
HV =
1854 xP
dv
2
(4.10)
where P is the load and dv is the average length of the Vickers diagonals in µm as
shown in Figure 4.9. The unit of HV is GPa in the S.I. system or sometimes
expressed in kg/mm2. The Vickers hardness test was performed using M12a
58
micro-hardness equipment (Vickers Instrument Limited). The load used was
9.807 N. As the ceramic was off-white in colour, it was blackened using a black
marker. When the load penetrates into the sample, it leaves a white indentation
which is easily resolved for d1 and d2 measurements.
FIGURE 4.9 : The average length of the Vickers diagonals.
59
4.8
Brunauer-Emnett-Teller (BET) and Barrett-Joyner-Halenda (BJH)
analysis
Autosorb Micromeritic (Model ASAP 2010) as shown in Figure 4.10 was
used to perform the BET and BJH analysis.
FIGURE 4.10: Autosorb Micromeritic (Model ASAP 2010)
The BET analysis calculates the specific surface area, average pore diameter and
total pore volume whilst the BJH analysis calculates the pore size distributions in
the ceramic (Barrett et al., 1951). About 100 mg of the powdered form of the
ceramics was used. Initially, the chamber was fed with helium gas at a pressure
of 30 mmHg for 30 minutes. The absorption was allowed to take place for 4
hours at the same pressure and at a temperature of 120oC. The absorption that
took place may be classified into six types referred to as the BET classification
(Gregg and Sing, 1982) and shown in Figure 4.11.
60
FIGURE 4.11: BET classification of absorption isotherms (Gregg & Sing, 1982).
Isotherms Type IV and Type V show hysterisis loop, with the lower
branch is the result of progressive addition of nitrogen to the system whilst the
upper branch is the result of progressive withdrawal of the gas. Isotherms that are
likely to yield a true value of the specific surface area are Type II and Type IV.
On the other hand isotherms Type I, Type II and Type VI are unnameable to the
BET analysis. From the isotherm, the amount of nitrogen can be calculated. It is
assumed that the absorption of a monolayer of liquid nitrogen onto the surface of
a mass of particles in the sample took place. Thus, the amount of nitrogen
calculated is the vapourized monolayer of liquid nitrogen. Based on this quantity,
the surface area can then be calculated.
The particle size (Li et al., 1997 ) can be calculated using the following
equation:
R/ =
6
ρS
(4.11)
61
where ρ is the density of the ceramic (kg/m3) and S is the specific surface area of
the ceramic (m2/kg) determined by the BET experiment.
The Barrett-Joyner-Halenda (BJH) method for calculating pore size
distributions is based on a model of the absorbent as a collection of cylindrical
pores. The theory accounts for capillary condensation in the pores using the
classical Kelvin equation, which assumes a hemispherical liquid vapour meniscus
and a well-defined surface tension. The ratio of core radius of the fluid to pore
radius of the sample is assumed to be constant. The BET isotherm was used as an
input to the BJH calculation.
4.9
High Resolution X-Ray Diffraction (HRXRD)
For solid samples, they were mounted onto the centre of a circular disc by
means of blue-tack and if the samples were powder, it was filled onto a similar
structure with a disc volume of approximately 57 mm3. The circular disc was
then mounted snugly into a jig which was placed on the diffraction circle. The
high resolution X-ray diffractometer used was a Bruker AXS D5005. The X-rays
produced were from a copper target with a wavelength, λ = 1.54312 Å. The
machine was operated at 30 mA, 40 kV and the jig holding the specimen and spun
at 15 rpm for uniform scanning. The scanning was between 2θ values of 25o to
56o with a step time of 5 seconds per step and a resolution of 0.01o. The settings
of these parameters were chosen using EVA software programme in a pc which
was interfaced with the diffractometer. Each run lasted for approximately 4 ½
hours and the data was saved as an EVA file. The EVA file in its raw form was
then converted as an ASCII file which was then analysed using Excel or Origin
software.
62
The mean crystallite size of the samples can be calculated using Scherrer's
equation (Cullity, 1978) ,
RX = 0.9 λ/β cos θ
(4.12)
where λ is the wavelength of the X-ray radiation, θ is the diffraction angle and
β2 = βm2 - βi2 , where βm is the measured Full Width Half Maximum (FWHM)
and βi is the FWHM from a Si standard and β is the corrected FWHM (Appendix
E). The FWHM is the full width (2θ) at half maximum value of the intensity (I).
The Bragg equation proposed by Yu et al. (1997) was used to calculate
lattice parameter a and c;
4 sin 2 θ
λ2
=
h2 + k 2 l2
+ 2
a2
c
(4.13)
where θ is the diffraction angle, λ is the wavelength of the X-ray radiation and
h,k,l are the Miller indices. The distortions in the lattice parameter were
calculated according to the equations;
δa = a220 - a110
(4.14)
δc = c101-110 – c101-220
(4.15)
and
The mean strain (e) was calculated using the following equation;
d relaxed − d T
< e2 >1/ 2 = 100 hkl relaxed hkl
d hkl
(4.16)
T
represents the position of the plane with Miller indices (hkl) of sample
where d hkl
relaxed
sintered at temperature T and d hkl
refers to d values indicated in the JCPDS
(1997a).
63
4.10
Scanning Electron Microscopy (SEM)
The SEM enabled the examination of the microstructure of the sample.
The ceramic sample was cut into a circular shape with a diameter of 10 mm and
mounted on a stud by means of a carbon tape. The sample was carbon coated
using a carbon sputter. The scanning electron microscope used was JEOL (Model
JSM-6100) as in Figure 4.12. The typical voltage used was 25 kV and the
magnification was between 15 000-25 000.
FIGURE 4.12: Scanning Electron Microscope (JEOL, model JSM-6100) at the
Physics Department, University of Warwick, England.
4.11
Transmission Electron Microscopy (TEM)
The JEOL Electron Microscope (JEM), model 2000FX transmission
electron microscope used for a higher magnification on the sample as compared to
SEM (Figure 4.13). The sample was in a powder form which was emulsified in
methanol. The suspended powder was pipetted onto a Cu grid and dried under the
heat of a filament light bulb. The images were recorded on a film and then
64
developed for viewing. An EDAX was also incorporated so as to distinguished
between SnO2 and Pt particles in the sample.
FIGURE 4.13: Transmission Electron Microscope (JEOL Electron Microscope
(JEM), model 2000FX) at the Physics Department, University of Warwick,
England.
4.12
Atomic Force Microscopy (AFM)
AFM operates by measuring the attractive or repulsive forces between the
tip and the sample. It could operate either in repulsive contact mode or the noncontact mode. In the former case, the instrument lightly touches a tip at the end of
the cantilever to the sample whilst in the latter mode, yield topographic images
from the attractive force measurements and the tip does not touch the sample at
all. However, in the contact mode, the AFM measures the repulsive forces
between the tip and the sample. As a raster scan drags the tip over the sample
surface, a sensor measures the vertical deflection of the cantilever which registers
65
as the height of the sample. High resolution of ~ 10 picometer (pm) and can
image sample in air as well as in liquids makes AFM differs from conventional
electron microscopes (Binnig et al., 1993; Albrecht et al., 1993 and Baselt, 1993).
The sample was cut into a shape of a disc with a diameter of about 10 mm
and a thickness of about 2 mm and mounted horizontally on the viewing platform
of the atomic force microscope (ARIS, model 3300). The atomic force
microscope used is shown in Figure 4.14.
To pc
FIGURE 4.14: Atomic Force Microscope (ARIS, model 3300), at the Physics
Department, University of Warwick, England.
The sample was illuminated by a He-Ne laser. The optical lever operates
by reflecting a laser beam off the cantilever. The angular deflection of the
cantilever doubles the angular deflection of the laser beam. The reflected laser
beam strikes a position sensor which comprised of two photodiodes, side by side.
The difference in signals between the two photodiodes is the position of the laser
spot on the detector and thus the angular deflection of the cantilever (Putman, et
al., 1992). The equipment was coupled to a personal computer (pc) in which the
scanning can be viewed via a monitor. In this manner, data from the topography
scans can be accessed and the scan can be controlled from the pc. The images
captured were stored and processed using Photoshop 7.0 software.
66
4.13
Fourier Transform Infra-Red (FTIR) Spectroscopy and FT-Raman
Shift Spectroscopy
FTIR spectroscopy has been widely used for both organic and inorganic
materials such as ceramics for material characterization. Specifically, when
applied to a semiconductor material, FTIR spectroscopy gives information such as
electrical conductivity variations. FTIR spectroscopy is based on infrared
absorption. The absorbance A/ of a sample at a given wavelength λ is related to
the intensity transmitted, IT by the sample according to the equation;
A/(λ) = log [Io(λ)/IT(λ)]
(4.17)
where Io is the intensity of the incident beam. According to the Beer Lambert law,
the absorbance is directly related to the absorption coefficient K:
A/(λ) = K(λ) z
(4.18)
where z is the thickness of the sample.
However, for a semiconducting material, part of the infrared absorption is
due to the free carriers. The Drude-Zener theory relates the absorption coefficient
to the electrical conductivity σ as in the following equation (Gibson, 1958);
K (λ) = σ(λ)/εoc/n
(4.19)
where σ(λ) is the electrical conductivity which is wavelength dependent, εo is the
permittivity in free-space, c/ is the velocity of light in a vacuum and n is the
refractive index of the material. Hence, the absorbance of a semiconductor at a
given wavelength is directly related to its electrical conductivity. Also part of the
absorption due to the free carriers is directly proportional to the free carrier
density and is a linear function of λ2 ( Harrick, 1962 and Baraton et al., 1998).
67
In the customary setup of a FTIR, an infrared light source is sent through a
transparent sample, the components brought into this cell absorb the infrared light
at its specific wavelength. The absorption is proportional to the concentration.
The equipment used in the experiments was a Spectrum GX Fourier Transform
Infra-Red (Perkin Elmer). The sample was in the form of pellet embedded in
KBr. Initially, KBr and the sample in the form of powder in the ratio of KBr to
sample which is typically 100:1 were mixed (over a period of 15 minutes)
homogeneously, then pressed in an evacuable die until a transparent disc was
obtained. It was then mounted onto the spectrometer and scanned at wave
numbers between 400 – 4000 cm-1.
The same equipment was used for FT-Raman shift experiments covering
the same range. The samples needed no further preparations as they can be
mounted on a jig. Raman spectroscopy is a complementary technique to infrared
spectroscopy and provides information on the vibrational energy levels of the
molecules (Herzberg, 1945; Colthrup et. al., 1975; Long, 1997 and Durig, 1981).
Figure 4.15 shows the equipment used for FTIR and FT-Raman spectroscopy.
FIGURE 4.15: Equipment used for FTIR and FT-Raman spectroscopy at Institute
of Ibnu Sina, Universiti Teknologi Malaysia.
68
4.14
X-Ray Photoemission Spectroscopy (XPS)
This is a technique for surface analysis in which monoenergetic soft Xrays are irradiated on a solid in ultra-high vacuum and the emitted electrons
analysed by energy. The typical spectrum is a plot of the number of detected
electrons per energy interval against their binding energy (X-ray photon energy –
measured electron kinetic energy). XPS is a unique surface-sensitive technique
for chemical analysis because the mean free path of electrons in solids is very
small, thus the detected electrons originate from the first few atomic layers (from
the irradiated surface). Data can be obtained from peak heights or peak areas and
chemical state identification from exact measurement of peak positions and
separations and spectral features.
XPS data was acquired using a Scienta ESCA 300 X-ray photoelectron
spectrometer at the RUSTI facility at the Daresbury Laboratory, Warrington,
United Kingdom. About 0.5 g of the powdered samples were used and filled in a
cavity about the same level as the stub. A monochromated Al Kα source
(hυ = 1486.6 eV) was used in conjunction with a 300 mm radius concentric
hemispherical analyser and the acquisition parameters were typically 150 eV pass
energy, 0.5 mm entrance slit size and 0.05 eV incremental step size. The absolute
binding energies of the photoelectron spectra were determined by referencing to
the C 1s transition at 284.6 eV which resulted from background hydrocarbons
from the ultra-high-vacuum environment. The O 1s, Pt 4f and Sn 3d were taken
before and after all region scans and the shape and intensity were consistent to
approximately 1% and the binding energies were reproducible to within ± 0.05
eV.
69
4.15
Mössbauer Spectroscopy
The Mössbauer effect is the recoil-free emission of γ radiation from a solid
radioactive material. Such recoil-free gamma emission can be resonantly
absorbed by atoms in a solid. The nuclear transistions are very sensitive to the
local environment of the atom and such spectroscopy is a sensitive probe of the
different environments an atom occupies in a solid material. This is because the
line widths of the recoilless transistions are very small compared to their energies.
The resolution of the Mössbauer effect is typically in the order of one part in 1012.
The Mössbauer spectroscopy were conducted in collaboration with Professor C.E.
Johnson of the Physics Department, Liverpool University, England.
The spectrometer is shown schematically in Figure 4.8. The gamma
source was mechanically vibrated back and forth to Doppler shift of the energy of
the emitted gamma radiation. The Doppler shifts occurred in the energy of the
photons as there was a relative movement between the emitter (source) and the
absorber (sample). By varying the relative velocity of the emitter, it is possible to
sweep photon energies over a range of values. As the energy of the gamma
radiation was scanned by Doppler shifting, the frequencies of the gamma radiation
that was absorbed by the sample was recorded by the dectector. In a typical
Mössbauer spectroscopy, the absorption rate is measured as a function of source
velocity.
70
FIGURE 4.16: Mössbauer spectrometer (schematic diagram).
The detailed account of the instrumentation and experimental procedures
are well documented by Benezer-Koller and Herber (1968), Spijkerman (1971)
and Kalvius and Kankeleit (1972). The samples were powdered and made to their
`optimum thickness’ as described in ‘The ideal Mossbauer Effect Absorber
Thickness’ by Long et al. (1983). Calibration was done using an iron source
(57CoRh) and the corresponding standard absorber (α-Fe metal) at room
temperature. All measurements of the velocity were relative to the centroid of
this 6 line reference spectrum. Both Pt-SnO2 and pure SnO2 were run at room
temperature. The data was collected via a pc.
4.16
Nuclear Magnetic Resonance (NMR)
In nuclear magnetic resonance (NMR) spectroscopy, the signal frequency
detected is proportional to the magnetic field applied to the nucleus. However,
the response of the atomic electrons to the externally applied magnetic field is to
produce a small in magnitude magnetic field at the nucleus which are in
71
opposition to the externally applied field. This changes the effective magnetic
field on the nuclear spin causes the NMR signal frequency to shift. This is known
as the chemical shift of the NMR which depends on the type of nucleus and the
electron motion in the neighbouring atoms and molecules. The NMR is able to
measure precisely the chemical shift and from this, chemical bonds and structure
of molecules can be ascertained. Bloch et al.(1946) and Purcell (1946) were
amongst the pioneers who observed the NMR effect in protons (1H). Knight
(1949) was the first to measure the chemical shifts in metals.
0.5 g of the sample was ground to fine powder for MAS-NMR analysis
using CMX 11-B. The 119Sn and 195Pt static-NMR spectra were obtained using a
204 MHz spectrometer of frequency 71.54 MHz. The chemical shift of the peaks
were referred to SnCl2(aq) as it is an interest to look at the doner states formed by
oxygen vacancies (Sn2+ ions) with activation energies of 0.03 eV up to 0.15 eV
(Ihokura and Watson, 1994).
4.17
Thermal Gravimetric Analyzer (TGA)
In this thermal analysis, the change in weight of the sample is recorded as
it is heated, cooled or held at constant temperature. The equipment used was
(Perkin Elmer) Thermal Gravimetric Analyzer, Series 7. The sample was in a
powdered form and the amount used was < 15 mg. The data collected was in
form of a plot of weight loss against sample temperature between room
temperature up to 1000oC.
72
CHAPTER 5
SENSOR ELEMENT CHARACTERIZATION SYSTEM (SECS) OPTIMUM OPERATING TEMPERATURE AND OPTIMUM
COMPOSITION
5.1
Measurements from SECS
The SECS was setup to investigate the electrical characterization of the
sensor material, Pt-SnO2 in the ceramic form. In this section, the improvements
made on the SECS is reported; how the sensor resistance, conductance and
sensitivity were derived from the load voltage measurement; the dependent of
sensor resistance (Rs) and load resistance (RL) in air and methane environment
were investigated and the reproducibility feature is shown. In section 5.2, the
optimum composition and operating temperature were determined via the
sensitivity measurements.
5.1.1
Improvements on the SECS
The SECS was initially known as the Gas Sensor Characterization System
(GSCS) and was used by John (2001) for sensing methane in air using Pd-SnO2
ceramic (Appendix C). A new probe was designed and constructed as shown in
Figure 5.1 with the following features;
73
1.
It is much slimmer and fits into the test chamber without problem of
touching the heating element,
2.
Most of the electrical connections are concealed using ceramics to avoid
excess heating,
3.
The new probe can be adjusted to measure either the bulk resistance or
the
surface resistance of the sample.
Another new feature of the SECS is that it is capable of storing data for
running time > 1000 s. This was not possible in the GSCS due to the software
limitations.
LEFT: OLD SENSOR PROBE
RIGHT: NEW SENSOR PROBE
FIGURE 5.1: The sensor probe used in the GSCS (left) and SECS (right).
74
5.1.2
Resistance (R), Conductance (G) and Sensitivity (S/) determination
The electrical circuit for measuring resistance of the ceramic sample is
shown in Figure 5.2.
IS
RS
VS
VC
RL
VL
FIGURE 5.2: Electrical circuit for sensor resistance measurement.
The data collected from SECS showed as the variation of the voltage across the
load, VL with time, t and was also tabulated in an EXCEL file. A typical data set
collected is shown in Figure 5.3. Between t = 0 s and t = 100 s, the sample was in
dry synthetic air, showing an almost constant value of VL ~ 0.25 V. The flow rate
of the synthetic air was maintained at 400 standard cubic centimetre per minute
(sccm) cm3/min using a flow controller (Cole Parmer). At t = 100 s, 25 000 parts
per million (ppm) methane in air was introduced but cut-off at t = 300 s. The
choice of using 25 000 ppm CH4 and a flow rate of 400 sccm was because that
was the only methane concentration available and the flow rate was easily
monitored. The load voltage increased sharply from ~ 0.25 V to ~ 2.70 V. The
spectrum in Figure 5.3 showing VL values for t > 200 s does not show much
increase and almost saturates. As such any value in such a range is considered as
a typical value of VL. With this value of VL , VC = 20 V and RL = 700 Ω and using
equation 4.3, the sensor resistance R S = R L ( VC V L − 1 ) can be calculated (Figure
5.4).
75
3.0
CH4 out
2.5
VL (V)
2.0
1.5
1.0
0.5
CH4 in
0.0
0
100
200
300
400
500
600
700
800
t (s)
FIGURE 5.3: Typical data collection from SECS, graph of load voltage (VL)
versus time (t). Methane gas was introduced at t = 100 s and cut-off at t = 300 s.
The sample used was 0.5 wt.% Pt-SnO2 and operating at 400oC. VC = 20 V and
RL = 700 Ω.
76
70.0
CH4
in
3
RS x10 (Ω)
60.0
50.0
40.0
CH4
out
30.0
20.0
10.0
0.0
0
100
200
300
400
500
600
700
800
t (s)
FIGURE 5.4: Corresponding sensor resistance (Rs) versus time (t) graph.
Methane gas was introduced at t = 100 s and cut-off at t = 300 s.
The conductance, G is the reciprocal of RS as in equation 4.4 and the sensitivity S/
can be evaluated using equation 4.5. Thus, the corresponding graph of G versus t
and S/ versus t graphs are shown in Figure 5.5 and Figure 5.6 respectively.
CH4 out
2.5
G x10-4 (S)
2.0
1.5
1.0
0.5
CH4 in
0.0
0
100
200
300
400
500
600
700
t (s)
FIGURE 5.5: Corresponding conductance (G) versus time (t) graph.
800
77
CH4 out
1400
1200
1000
S (%)
800
/
600
400
CH4 in
200
0
0
100
200
300
400
500
600
700
800
t (s)
FIGURE 5.6: Corresponding sensitivity (S/) versus time (t) graph.
5.1.3
Effect of RL on Rs in dry synthetic air
In this experiment the load resistor RL was varied with voltage supply,
Vc = 20 V. In this case the sensor was in dry synthetic air. The results are
tabulated in Table 5.1 and plotted in Figure 5.7. From Table 5.1 and graph in
Figure 5.7, the sensor resistance Rs increased linearly between RL = (100-800) Ω.
After RL = 800 Ω, there was an abrupt increase in RS with an increase in RL. As
such, a suitable range of RL would be between 100-800 Ω for the experiment. At
RL >800 Ω, the signal VL was infested with noise. The choice in most experiments
followed reported uses RL = 700 Ω (unless otherwise stated) as the noise was
tolerable.
78
TABLE 5.1: Load resistance and sensor resistance with voltage supply Vc = 20 V
in dry synthetic air.
Sensor
Resistance
RS ± 10 (Ω)
25501
29059
27629
26908
29619
30121
30467
31970
50303
55333
Load Resistance
RL ± 5 (Ω)
100
200
300
400
500
600
700
800
900
990
60000
50000
RS (Ω)
40000
30000
20000
10000
0
0
100 200 300 400 500 600 700 800 900 1000
R L (Ω)
FIGURE 5.7: Graph of sensor resistance (RS) against load resistance (RL) in dry
synthetic air.
79
5.1.4
Effect of RL on Rs in 25 000 ppm methane (in air)
In this case the sensor was in 25 000 ppm methane in air. The results are
tabulated in Table 5.2 and plotted in Figure 5.8. From Table 5.2 and Figure 5.8.
There are two distinct regions of linearity between RL and RS. For the values of
100 Ω ≤ RL ≤ 400 Ω, the slope is 3.64 whilst for 500 Ω ≤ RL ≤ 1000 Ω, the slope is
6.06. Thus, the relationships can be written as follows;
and
RS = 3.64RL + 1500
100 Ω ≤ RL ≤ 400 Ω
RS = 6.06RL + 1000
500 Ω ≤ RL ≤ 1000 Ω
TABLE 5.2: Load resistance and sensor resistance with voltage supply Vc =20 V
in 25 000 CH4 in air.
Load Resistance
RL ± 5 (Ω)
Sensor Resistance
RS ± 10 (Ω)
100
1908
200
2268
300
2654
400
3071
500
3839
600
4436
700
4989
800
5454
900
6136
990
6633
The sensor resistance were much lower in methane than in the dry
synthetic air. The ratio of RS in air to RS in methane ~ 12 ± 2 for RL between (100990) Ω. The RS versus time spectrum showed higher noise at RL values > 800 Ω.
80
7000
RS = 6.06RL + 1000
6000
RS (Ω)
5000
RS (Ω)
4000
3000
RS = 3.64RL + 1500
2000
1000
0
0
100
200
300
400
500
600
700
800
900 1000
RRLL (Ω)
(Ω)
FIGURE 5.8: Graph of RS against RL in 25 000 ppm CH4.
5.1.5
Reproducibility
The methane gas was introduced at time t = 100 s and cut-off at t = 350 s
as shown in Figure 5.9. The sample showed a response with a load voltage of
1.25 V. Then, when the sample was exposed to methane again at t = 900 s, the
load voltage reached a value of 1.24 V. This indicates that the data obtained were
almost reproducible. In a run, it was a common practice that such reproducible
feature occurred. The good signal was attributed to the ability to maintain a
constant temperature, air and methane flow rate in the SECS. The improved
system makes it possible to perform experiments and record data for longer time
(t > 1000 s).
81
CH4 out
1.60
1.40
VL (V)
1.20
1.00
0.80
0.60
0.40
0.20
CH4 in
0.00
0
500
1000
1500
2000
t (s)
FIGURE 5.9: Graph of load voltage (VL) against time (t) showing the
reproducibility feature.
5.2 Optimum operating temperature and composition
For this series of experimental runs, the following parameters were kept
constant; methane gas in air concentration (25 000 ppm), flow rate (400 sccm),
applied voltage (20 V), operating temperature (250 – 400oC), load resistance
(RL = 700 Ω). The sensor element were cut into 10x10x2 mm3 pieces and placed
between the slabs of electrodes in the SECS. The value of VL which varies with
time was recorded using a ADC signal processor which can be monitored on a PC
screen. A plot of load voltage (VL) against time (t) was obtained using Microsoft
EXCEL program.The methane gas in air was introduced at t = 100 s for a duration
of 300 s, in which the value of VL almost saturates. The value of VL at t = 400 s
was taken for A plot of sensitivity versus Pt loading in stannic oxide was then
82
obtained so as to determine the optimum composition. The result are tabulated in
Table 5.3 and plotted out as shown in Figure 5.10.
TABLE 5.3: Sensitivity of various compositions of Pt-SnO2 ceramics at
operating temperatures 250 – 430oC.
Sensitivity of samples (%): in 25000 ppm
Operating
Temperature
± 1 (oC)
250
300
350
380
400
430
0.1 wt.%
0
0
66.99
167
200
100
0.3 wt.%
0
0
668.84
1012.39
1123.76
802.58
0.5 wt.%
0
0
767.39
1351.28
1577.85
1083.28
CH4
0.7 wt.%
0
0
200
401.96
403.94
200.58
1800
1600
1400
0.1wt.%
S/ (%)
1200
0.3wt.%
1000
0.5wt.%
800
0.7wt.%
600
1.0wt.%
400
200
0
250
300
350
400
450
T (oC)
Figure 5.10: Sensitivity curves (S/) against operating temperature (T).
Methane in air could not be detected at temperatures below 200oC. This is
because the electrical resistances of the samples were too high and the change in
electrical resistance in synthetic air and methane (in air) was very small to show
1 wt.%
0
0
55.00
200.58
353.10
301.17
83
any sensitivity readings. However, there was a marked increase in sensitivity for
all the samples when the operating temperature approached 350- 380oC. The
sensitivity reached its maximum value in the vicinity of ~ 400oC and fell
thereafter for all the samples. Thus, the optimum operating temperature was
~ 400oC. The maximum value of sensitivity at ~ 400oC is attributed to the
dissociation of methane into carbon dioxide and water at this temperature. The
sample which shows the highest magnitude of sensitivity was 0.5 wt.% Pt in
SnO2. Hence, this is an indication that the sample is the optimum composition for
methane sensing in air. The increase in sensitivity up to ~ 0.5 wt.% Pt is probably
due to the uniform distribution of Pt in the ceramics which in turn increase the
surface density. The effective density in terms of particles per unit area
consequently decrease after ~ 0.5wt.% Pt loading, hence the decrease in
sensitivity (Ihokura and Watson, 1994). Another probable reason could be the
rate of oxidation on the particles reaches its maximum activity at ~ 0.5wt.% Pt
loading at the optimum operating temperature of ~ 400oC.
The sensitivity of the Pt-doped stannic oxide for the various composition
at an operating temperature of 400oC (extracted from graph in Figure 5.10) is
shown in Table 5.4 and plotted as in Figure 5.11. The sensitivity increases for Pt
loading (0.1 - 0.3 wt.%) from ~ 200% to ~ 1124%. The sensitivity reaches its
maximum value (~ 1578%) at ~ 0.5 wt.% Pt and falls gradually thereafter. This
confirms that the optimum composition was 0.5wt.% Pt doped in SnO2 for
methane sensing in air. As the maximum signal (sensitivity) with the above
composition is maximum at an operating temperature of ~ 400oC, then this is the
optimum operating temperature for methane sensing in air. These results have the
same value of optimum operating temperature but lower value of doping level
when compared to the work done by John (2001) on palladium doped stannic
oxide ceramics. Such findings will economize the use of Pt in the making of the
Pt-SnO2 ceramics for methane sensing in air. The working temperature is easily
achieved using technologies available nowadays and the Pt-SnO2 ceramic is a
good candidate for an active element for methane sensing in air.
84
TABLE 5.4: Sensitivity of 0.1-1.0 wt.% Pt in SnO2 at operating temperature of
400oC extracted from Figure 5.10.
Wt.% Pt in stannic oxide
Sensitivity (%)
0.1
200.00
0.3
1123.76
0.5
1577.85
0.7
403.94
1.0
353.10
2000
S/ (%)
1500
1000
500
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
W* (wt.%)
FIGURE 5.11: Graph of Sensitivity (S/) against of Pt loading (W*) at operating
temperature of 400oC.
85
5.3 Conclusion
The SECS was capable of measuring the voltage across the load, hence the
voltage across the sensor which is the base signal for the methane sensing. The
sensor resistance, conductance and sensitivity were then easily derived. The
sensor resistance and load resistance varied almost linearly for load resistance
< 800 Ω with low noise, both in dry synthetic air and in methane (in air)
environment. However, for almost all experiments conducted reported in Chapter
5 to Chapter 7, RL used was 700 Ω. Another noble feature of the SECS is the
reproducibility which makes the data obtained reliable.
The important results obtained from the experiments using SECS were the
optimum composition of the Pt doped in SnO2 and optimum operating
temperature. The optimum composition was found to be 0.5 wt.%. This result
reflects the economy when incorporating Pt in SnO2. The addition of Pt in SnO2
was easily achieved as compared to methods like spin-coating and sol-gel route.
The optimum operating temperature was ~ 400oC which was also achieved by
other workers in methane sensing (Khodadadi et al., 2000.). This is a favourable
operating temperature as at this temperature methane combustion occur which
enhances the methane sensing activity. Furthermore, this temperature can be
easily achieved by today’s technology like micro-hotplate (Cavicchi et. al., 1996)
and micromachining (Isolde et. al, 2001). To sense methane at lower temperature
for example between 200-300oC, would only cause cross-sensitivity problem with
other gases like CO and NOx.
86
CHAPTER 6
RESISTANCE, CONDUCTANCE AND SENSITIVITY MEASUREMENTS
6.1
Introduction
The following work focussed on the resistance, conductance and
sensitivity measurements operating at a temperature of 400oC which is the
optimum temperature for methane sensing in air as determined in Chapter 5. Both
the SECS (temperature range: 200-440oC) and impedance analyzer (temperature
range: 30-440oC, frequency: 50-800 kHz) were used for this purpose. Section 6.2
is an account of the resistance in air of the doped samples (0.1, 0.5 and 1.0) wt.%
Pt in SnO2. As a comparison, the resistance of the pure SnO2 was also studied. It
is expected that the resistance in air would differ greatly than in methane. It is the
difference in resistance that is essential for methane detection. The resistance of
the 0.5 wt.% Pt-SnO2 in air can then be compared to its resistance in methane in
section 6.3. In section 6.3, the resistance of the 0.5 wt.% Pt –SnO2 in methane (25
000 ppm) was studied. Section 6.4 is a study of the conductance in air for the
pure and doped samples described in section 6.2 and also the comparison of
conductance in air and methane (25 000 ppm) for the sample with the optimum
composition (0.5 wt. % Pt-SnO2). In section 6.5, the conductance-time dependent
study was made on the 0.5 wt. % Pt-SnO2. The conductance power law was
studied on the 0.5 wt. % Pt-SnO2 sample in section 6.6. The effect of methane
concentration on conductance was investigated in section 6.7 using the 0.5 wt. %
Pt-SnO2 sample. In the last section (6.8), sensitivity studies were made on the 0.5
wt. % Pt-SnO2. The reported resistance in all the sections is the bulk resistance.
87
6.2
Resistance measurements in air
Samples of pure SnO2 and Pt doped SnO2 (0.1, 0.5, and 1.0 wt.%)
were studied using an impedance analyzer. The samples were heated from
room temperature up to 450oC. The resistance in air for the samples in the
temperature range are plotted out in Figure 6.1.
250000
200000
R (Ω)
150000
pure stannic oxide
100000
0.1wt.% Pt
1.0wt.%Pt
0.5wt.% Pt
50000
0
0
50 100 150 200 250 300 350 400 450
T (oC)
FIGURE 6.1: Graph of resistance (R) against temperature (T) in air.
All samples show a decrease in resistance with an increase in temperature
(for T > 150oC). The pure SnO2 shows the lowest resistance over the temperature
range of 50 – 450oC. The doping of Pt increases the maximum resistance up to
200 kΩ for the 1 wt.% Pt-SnO2 sample. All the doped samples show a prominent
maximum value which shifted to a lower temperature when the Pt loadings
increased. However, for comparison the pure stannic oxide does not show any
88
maximum. An Arrhenius plot was also plotted as shown in Figure 6.2. The
activation energies were calculated and are tabulated in Table 6.1 and plotted in
Figure 6.3. There are two distinct regions with different activation energy. The
activation energy changes in value around 200oC. The activation energy at
temperature T < 200oC is lower than the activation energy well above 200oC.
7
6
ln (R/T)
5
4
pure stannic
oxide
0.1wt.% Pt
3
2
0.5wt.% Pt
1
1wt.% Pt
0
1.2
1.6
2.0
2.4
2.8
3.2
1000/T (K-1)
FIGURE 6.2: Arrhenius plot for samples at temperatures between 50-450oC.
TABLE 6.1: Activation energy for various Pt loadings in SnO2 in intrinsic and
extrinsic regions.
Pt loading
(wt.%)
0
0.1
0.5
1.0
Activation energy
(±0.01 eV)
T > 200oC
0.40
0.29
0.36
0.43
Activation energy
(±0.01 eV)
T < 200oC
0.11
0.08
0.11
0.18
89
0.50
0.45
0.40
>200oC
0.35
EA (eV)
0.30
0.25
0.20
0.15
< 200oC
0.10
0.05
0.00
0.0
0.2
0.4
0.6
0.8
1.0
W* (wt. %)
FIGURE 6.3: Graph of Activation energy (EA) against Pt loadings (W*) for
temperatures between 21-200oC and 200-450oC.
6.3
Resistance measurements in methane
Only the 0.5 wt.%Pt-SnO2 was investigated using the SECS for
temperatures between 200-450oC. The variation of the sample resistance in
methane is tabulated as in Table 6.2 and plotted as in Figure 6.4. The resistance
of the sample in 25 000 ppm methane in air decreases with temperature from
~ 54 kΩ at 200oC to ~ 4.6 kΩ when sample was heated up to 430oC. The value of
resistance is much lower compared with the same sample heated in air from 200450oC. The Arrhenuis plot is shown in Figure 6.5.
90
TABLE 6.2: Resistance of 0.5 wt. % Pt-SnO2 in 25 000 ppm CH4.
T ± 1(oC)
440
430
420
410
400
380
350
300
280
250
200
RCH4 ± 10 (Ω)
4650
4690
4814
4989
4989
6063
8042
16368
21701
31884
59442
70000
60000
RS (Ω)
50000
40000
30000
20000
10000
0
200
250
300
350
400
450
T (oC)
FIGURE 6.4: Plot of resistance (R) in 25 000 ppm methane against temperature
(T) for 0.5 wt.% Pt in SnO2.
91
7
6
ln(R/T)
5
4
3
2
1
0
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
1000/T
FIGURE 6.5: Arrhenius plot for 0.5 wt.% Pt in SnO2 in 25 000 ppm methane.
From the Arrhenius plot, there are two distinct regions with different
activation energies. The transistion occurred at T ~ 400oC. For T between 200oC
and 400oC, the activation energy was 0.48 eV and for T > 400oC up to 440oC was
0.21 eV. For the case of SnO2 thin film, de Angelis and Minnaja (1991) reported
two different activation energies: at low temperature (< 427oC), its activation
energy was 0.42 eV and higher temperature ( > 427oC K up to 527oC), its
activation energy was 0.15 eV. Mandayo et al., (2003) who worked on thin film
Pt-SnO2 reported two activation energies: 0.35 eV at temperatures
(300< T< 400)oC and 0.19 eV at temperatures (> 400oC up to 500oC). The
activation energy at temperatures T > 400oC, seemed to be in agreement with de
Angelis and Minnaja (1991) whilst the activation energy at temperatures < 400oC
has a value near to the value obtained by Mandayo et al.(2003). The difference
in activation energies is associated with the formation of different chemisorbed
oxygen species (O2-, O- and O2-) on the surface of the sample at different
temperature range which is discussed in section 6.9. No value of activation
energy was available at temperatures < 200oC as methane could not be detected
by the SECS.
92
6.4
Conductance in air and methane
The electrical conductance is the reciprocal of the resistance obtained in
sections 6.1–6.3. The conductance in air for the samples in the temperature range
investigated are plotted out in Figure 6.6 and Figure 6.7.
14.0
12.0
pure stannic
oxide
G x10-4(S)
10.0
0.5wt.% Pt
8.0
1.0wt.%Pt
6.0
0.1wt.% Pt
4.0
2.0
0.0
0
100
200
300
400
500
T (oC)
FIGURE 6.6: Graph of conductance against temperature in air.
93
3.0
2.5
0.5wt.% Pt
G x10-4(S)
2.0
1.0wt.%Pt
1.5
0.1wt.% Pt
1.0
0.5
0.0
0
100
200
300
400
500
T (oC)
FIGURE 6.7: Graph of conductance (G) against temperature (T) in air (for doped
samples only).
In air, all the samples show an increase in conductance with temperature. All
samples show a slow increase in conductance up to a temperature of 250oC.
Above 250oC, all samples show an exponential increase in conductance with
temperature. The magnitude of conductance was large for pure SnO2 compared to
the doped samples for any given temperature. For the doped samples, increasing
the Pt loading lowers the conductance value in air. The conductance of 0.5 and
1.0 wt.% Pt loadings was almost constant below 230oC. The conductance of 0.5
wt.% Pt loading in methane is shown in Figure 6.8.
94
2.50
2.00
in methane
G x10-4 (S)
in air
1.50
1.00
0.50
0.00
200
250
300
350
400
450
T (oC)
FIGURE 6.8: Graph of conductance (G) against temperature (T) in methane for
0.5 wt.% Pt-SnO2.
The conductance for the 0.5 wt.% Pt-SnO2 in methane is much greater than in air.
The difference in conductance becomes more prominent at temperatures from
300oC up to 450oC. When the conductance is plotted against 1000/T, as in Figure
6.9, the behaviour is closer to the thick-film behaviour reported by Lantto et al.
(1988), Figure 6.10. It also shows that curve obtained experimentally differs from
the plot of G against 1000/T for a Taguchi Gas Sensor (TGS) which has a
pronounced minimum and maximum.
95
2.50
2.00
G x10-4(S)
in methane
in air
1.50
1.00
0.50
0.00
1.2
1.4
1.6
1.8
2.0
2.2
2.4
1000/T (K-1)
FIGURE 6.9: Graph of conductance (G) against 1000/T in methane for 0.5 wt.%
Pt-SnO2.
96
FIGURE 6.10: Graph of conductance (G) against inverse temperature (1000/T) of
a thick film sensor (Lantto et al., 1988).
97
6.5 Conductance –time dependent
The SECS provides resistance-time curves, hence the conductance-time
curve was derived from such curves by taking the reciprocal of the resistance.
Figure 6.11 shows the conductance-time curve in air and methane.
x10-4
Conductance (Siemens)
2.5E-04
G (S)
o
400 C
2.0
2.0E-04
380 C
1.5E-04
1.5
350 C
o
o
pure SnO2, 400oC
o
1.0
1.0E-04
330 C
o
0.5
5.0E-05
300 C
a
0.0E+00
0
100
200
300
400
500
600
700
800
Time
t (s)(s)
FIGURE 6.11: Graph of conductance (G) against time (t) in methane for
0.5 wt.% Pt-SnO2 at temperatures of (300, 330, 350, 380 and 400)oC.
The conductance-time curves show (i) the conductance of sample in air between
time t = 0 s and t = 100 s, (ii) the conductance in methane between t = 100 s and
t = 300 s and (iii) the conductance in air for t > 300 s. At a given constant
temperature, (300-400)oC, the conductance in air initially was fairly constant and
this is assigned as Go , then rose steeply as soon as the methane was introduced
and finally decayed exponentially when methane was cut-off and the sample was
once again in air. The high conductance in methane arises from bulk material
transformation as a result of lattice oxygen consumption (Safonova et al., 2000).
At high temperatures T > 300oC, the rate of diffusion is high as the rate of oxygen
migration from the bulk to the surface predominates. Safonova et al. (2000)
observed two successive stages of conductance increase. The first stage is a small
rise in conductance which is small in magnitude which is attributed to the surface
reactions of gas molecules with chemisorbed oxygen, typical of CO detection.
98
Since such a rise of conductance was absent in this study, it can be concluded that
the gas sensed using the SECS was mainly methane. The second rise in
conductance is similar to the one obtained in the experiment. It can also be
asserted here that XPS could distinguish whether the rise of conductance observed
from the experiment could be due to a deviation of SnO2 composition from
stoichiometry or a change of the chemical state of doping metals from oxide to
metallic. Such modifications could be brought about by the formation of doping
metals at the grain boundaries. The pure SnO2 sample operated at 400oC shows
similar characterization but exhibits a high Go value compared to the doped
sample but shows a negligible change of conductance in methane.
6.6 Conductance –power law
The observed dependence of sample conductance, G on the gas partial
pressure, p or simply G(p) is frequently approximated by the power law (Kissine
et al., 2001):
G ~ pm
(6.1)
where m is an integer. The flow rate, χ was approximated as the gas partial
pressure (p) (Kissine et al., 2001) . The slope of the graph of ln G versus ln χ will
yield m. The data are shown in Table 6.3 and the graph is plotted as in Figure
6.12. From the graph m ~ - 0.5, hence, G ~ p-0.5 which is in agreement with
Kupriyanov (1996). The relationship established from the experiment can be
written as ln G = -0.5002ln χ – 6.3333. According to Kupriyanov (1996), the
concentration of electrons in the doped material is varied with the oxygen
chemisorption according to the reaction: O2 + e- Æ O2- or O2 + 2e- Æ 2Owhere the type of chemisorbed ions depends mainly on temperature. The graph
also shows that the conductance decreases with the gas partial pressure which is
an indicative that the Pt-SnO2 ceramic is an n-type semiconductor. Thus, the
decrease in the sample conductance is related to the formation of chemisorbed
oxygen at the sample surface according to the equations above.
99
TABLE 6.3: Conductance and gas flow rate
χ (sccm)
G (S)
ln χ
ln G
100
1.56x10-4
4.605170
-8.76565
200
1.43x10-4
5.298317
-8.85267
300
1.18x10-4
5.703782
-9.04483
400
9.52x10-5
5.991465
-9.25953
500
6.41x10-5
6.214608
-9.65507
ln
ln χ
χ
-6
-7
ln G
-8
ln G
3
4
5
6
7
ln G = -0.5002ln
χ – 6.3333
y = -0.5002x
- 6.3333
-9
-10
-11
-12
FIGURE 6.12: Graph of ln G against ln χ.
100
6.7 Conductance –methane gas concentrations
The conductance of the doped sample (0.5 wt.% Pt-SnO2) changes with the
methane concentration as shown in Figure 6.13. The graph shows a sharp
increase in conductance at low methane in air concentration (< 2500 ppm CH4 in
air) but becomes monotonic at high methane concentrations. The conductance
(G)-CH4 concentration (c) graph takes the form G = kcn, where k is a constant. A
plot of ln G against ln c is shown in Figure 6.14. The graph shows a linear
correlation between ln G and ln c. The slope of the graph yield a value of ~ 0.35,
thus G = kc0.35. This value is 2.5 times larger than the value obtained by Tournier
and Pijolat (1999) using a tin dioxide sensor developed by Coreci Company in
Lyons, France. The difference could be due to the production of the ceramics
which was undoped, cold pressed (4 tons/cm2) powder supplied from Prolabo
(France) and sintered at 800oC. Interestingly the value obtained from experiment
is similar to the value (slope ~ 0.38) when Tournier and Pijolat (1999) plot of ln G
against ln O2 concentration.
Graph of Conductance against CH4 concentrations
x10-4
Conductance (Siemens)
2.5E-04
0.35
G =ln(x)
kc - 0.0001
y = 3E-05
2.0E-04
G (S) 1.5E-04
1.0E-04
0.5
5.0E-05
0.0E+00
0
5000
10000
15000
20000
25000
30000
(ppm)
CH4 cconc.
(ppm)
FIGURE 6.13: Graph of conductance (G) in methane against CH4 concentration
(c).
101
ln CH4 concentration
ln c
0
ln Conductance
-2 0
2
4
6
8
10
12
-4
ln G
-6
ln G = 0.35ln c – 11.93
y = 0.3466x - 11.933
-8
-10
-12
-14
FIGURE 6.14: Graph of ln G against ln c. (G is conductance in methane and c is
methane concentration).
6.8 Sensitivity-time dependent
Samples of 0.5 wt.% Pt were analysed using the SECS. The sensitivity
was calculated using equation (4.5) and the time dependent form is as shown;
S / (t) =
G(t)methane − G(t)air
x100%
G(t)air
(6.1)
where t is time in seconds. G(t)air is the conductance in air at any given time and
this is constant for a given set of experimental conditions. G(t)air is also Go value
of the sample in air and thus the sensitivity of the sample in air is zero. G(t)methane
is the conductance of the sample in methane. Methane was introduced into the
SECS at t = 100 s and cut-off at t = 300 s when there is not much change in the
conductance.
Graphs in Figure 6.15 shows the sensitivity curves for the 0.5 wt.%PtSnO2 in 25 000 ppm CH4 at operating temperatures 350 – 440oC. For all curves,
between t = 0 s and t = 100 s, the sensitivity is almost constant and zero.
102
Graph of Sensitivity against time
1400
o
Sensitivity (%)
1200
S/ (%)
400 C
1000
o
420 C
800
o
440 C
600
400
o
380 C
200
o
350 C
0
0
100 200 300
400
500 600 700 800
900 1000
timet (s)
(s)
FIGURE 6.15: Graphs of sensitivity (S/) against time (t) of 0.5 wt.% Pt-SnO2 in
air and in 25 000 ppm CH4 at operating temperatures 350-440oC at flow rate of
400 sccm.
At 100 ≤ t ≤ 300 s, the sensitivity increases sharply at first but plateaus-off.
When the methane is cut-off at t = 300 s, the sensitivity falls off exponentially.
Notice that the sensitivity increases for operating temperatures 350-400oC but
decreases thereafter. This is attributed to the oxidation of methane which occurs
in the temperature range (300oC up to 400oC). In the experiment, it was
envisaged that the methane oxidation increased from 300oC, became maximum at
400oC and decreased thereafter. Therefore, the sensitivity of the Pt-SnO2 ceramic
to methane is influenced by the methane oxidation activity in such a manner.
103
6.8.1 Relative conductance and Sensitivity
The term
G( t )methane − Go
is called the relative conductance change.
Go
Kohl et al. (2000) states that the relationship between the amount of gas detected
and the relative conductance change is:
G (t )methane − Go
= K c , where c is the
Go
concentration of the gas detected. A plot of relative conductance change versus
square root of the concentration is shown in Figure 6.16. From the slope of the
graph, K = 0.0006 (ppm)-1/2 for pure SnO2 and K = 0.0802 (ppm)-1/2 for 0.5 wt.%
Pt-SnO2. Thus, doping with 0.5wt.% Pt in SnO2 increases the sensitivity of the
base material by a factor of 133.
0.06
15.0
∆ G/G = 0.0802c
13.5
1/2
+ 1.2048
12.0
0.02
10.5
-0.02
7.5
∆ G/G = 0.0006c 1/2 - 0.0727
6.0
-0.04
∆G/G
0.00
9.0
∆G/G
0.04
-0.06
4.5
doped
undoped
3.0
1.5
-0.08
-0.10
-0.12
0.0
0
20
40
60
80
100 120 140 160 180
(c) 1/2 (ppm-1/2)
FIGURE 6.16: Graph of relative conductance change (∆G/G) against square root
of methane concentration (c1/2). Samples used were pure SnO2 and 0.5 wt.% PtSnO2 sintered at 1000oC.
104
This is a desirable feature for methane sensing in air as the signal are large in
magnitude and easy to detect electrically. However, the conductance change in
the pure stannic oxide which was sintered at 1000oC for 1 hour shows very small
change in conductance when in air and in methane. Kohl et al. (2000) reported
the introduction of Pt in WO3 produces a sensitivity increase of about a factor
about 34 in sensing volatile namely 2,4-nonadienal.
6.8.2 Effects of flow rates
The effect of flow rate of CH4 on the sensitivity was investigated on the
optimum composition sample (0.5 wt.% Pt-SnO2) operating at 400oC. The results
are depicted in Figure 6.17. 25 000 ppm methane was introduced at t = 100 s and
cut-off at t = 300 s. The sensitivity was calculated as in previous sections. The
maximum sensitivity for each flow rate is shown in Table 6.4. For the flow rates
100-500 sccm, the sensitivity increases with the flow rate as shown in Table 6.4.
This increase could be attributed to the efficient replenishment of oxygen species
(O-, O2-), adsorption and desorption process on the sample. Thus, the flow rate of
the gas can control the sensitivity of the sample towards methane.
105
500 sccm
400 sccm
/
Sensitivity (%)
1200
1000
300 sccm
800
200 sccm
S (%) 600
100 sccm
400
200
0
0
250
500
750 1000 1250 1500 1750
time
(s)
t (s)
FIGURE 6.17: Sensitivity curves (S/) against time at flow rates (χ).
Table 6.4: Flow rate and maximum sensitivity
Flow rate (sccm)
100
200
300
400
500
Max. Sensitivity (%)
559
751
921
962
1088
6.9 Discussion and conclusion
As expected, the electrical characteristics (R, G and S/ measurements) are
different from that of the pure SnO2 ceramics. This could arise from the
modification of the surface and bulk due to the presence of Pt. Cabot et al. (2000)
suggested that such different electrical characteristics is due to the variations of
the grain barriers. The resistance in air for the doped sample showed a higher
resistance than the pure stannic oxide sample. The resistance of the doped
106
samples also increased with an increase in Pt doping in SnO2. Similar result was
obatained by Ambrazeviciene et al. (1993) but with chemically deposited SnOx
films. The reason that causes the increase in resistance could be due to the Pt2+
states which are directly related to the density of the surface states of the sample.
The presence of this band generates a marked increase in the resistance, of the
order of 5 for the case of the 1 wt.%Pt-SnO2. Cabot et al. (2001) reported that
from XPS analysis at the oxygen region in a Pt-doped stannic oxide the surface
states facilitate the adsorption of oxygen and hence increase the negative surface
charge. A higher grain barrier is needed to maintain the charge neutrality and
this resulted in a higher resistance value compared to the undoped samples. This
will give rise to an increase in electrical charge at the grain surface due to the
variation of surface states density. Another explanation for the resistance increase
in the modified samples is the tin oxide conduction-band depletion near Pt(0) as
explained by de Angelis and Minnnaja (1991). The XPS technique shows the
presence of Pt(0) in the modified sample as in Figure 6.18.
Pt(0)
39000
38000
Pt(2)
I (arb. unit)
37000
36000
35000
34000
33000
82
80
78
76
74
72
70
68
66
64
EB (eV)
FIGURE 6.18: Graph of intensity (I) against binding energy (EB). XPS spectrum
of modified sample showing Pt(0) and Pt(2) states.
107
De Angelis and Minnnaja (1991) also pointed out that introducing Pt into SnO2
will reduce the unstable behaviour of pure SnO2 because the shallow levels are
affected by external atmosphere, thus undoped SnO2 is not a suitable candidate for
a reliable sensing element.
The high resistance in the doped sample corresponds to a higher grain
barrier (Cabot et al., 2000) when compared to the undoped sample. The high
grain barrier may be attributed to more oxygen absorption sites as seen in the
SEM analysis introduced by the presence of Pt. Yamazoe et al.(1983) points out
that the effects of noble metal such as Pt on the atomic (O-, O2-) and molecular
(O2-) species is not the same and Cabot et al. (2001) deduced that at the regime
where the resistance increase was measured, only the adsorption of O-, O2appeared to be enhanced by the noble metal introduced. Another reason could be
directly related to the presence of Pt metal clusters more localised at the surface
sample (Kang et al., 1994). In comparing the resistance values between the
modified samples and the undoped samples, it can be said that the Pt states have
an important contribution to the resistance values. This influence depends greatly
on the Pt loadings in SnO2. The introduction of CH4 modifies the resistance of
RS x103(Ω)
the doped samples at higher temperature (T>300oC) (Figure 6.19).
140
in air
120
in methane
100
80
60
40
20
0
200
250
300
350
400
450
T (oC)
FIGURE 6.19: Comparison of sensor resistance (RS) for 0.5 wt.% Pt-SnO2.
108
The resistance change in air for the pure SnO2 is related to the presence of two
donor levels, 30 and 150 meV below the bottom of the conduction band and
oxygen adsorption on the sample surface (Jarzebski, 1972). In air, the resistance
of the doped sample decreases with temperature up to ~ 280oC. This could be
due to an activation of 30 meV shallow donor state as most shallow donor levels
are fully activated near 250oC (Shim et al., 2002). Thus, an increase in
temperature up to 450oC will not result in an increase in resistance. However, the
observed increase in resistance for temperature above 280oC is due to the change
of chemical state of oxygen which is adsorbed on the sample surface as suggested
by Madou and Morrison (1989). It is suggested that at temperature between 200280oC, oxygen molecule (O2) are chemically adsorbed in the form of O2- and
result in the following reaction; O2- + e Æ 2O-. Hence, O- species evolve at the
sample surface. The next step is the transformation of O2- to 2O-. According to
Chang (1983) and McAleer et al. (1987) the transition O2- Æ O- occurs at T >
150oC, O2- desorption occur at T > 150oC, O- desorption at T > 520oC from SnO2.
The oxygen adsorbed on the sample surface led to the formation of depletion
layer near to the surface of the sample. This resulted in the increase of resistance
of the sample. The apparent result is that the sample resistance decreases up to
280oC and increases thereafter (up to 450oC).
However, in methane, the resistance of the sample decreases with
increasing temperature. The resistance is 2 order of magnitude less than the
resistance in air for temperatures between 200-300oC. Between 300-450oC, the
difference in resistance becomes as large as 10 order of magnitude. This is
attributed to the onset of the methane oxidation which occurs at these temperature
and in the case of the doped sample, the rate of oxidation is greatly promoted by
the catalyst Pt. The activation energy in air for doped samples differs slightly
from the pure SnO2 sample but increases slightly with an increase in Pt loadings.
In air, there are two distinct values associated with temperatures below and above
200oC. The activation energy below 200oC has a lower magnitude than the
activation energy above 200oC for both pure and doped samples. The mere
difference of activation energy in the same sample with respect to the temperature
(200oC) indicates that there is a change in the conduction mechanism for both
109
pure and doped samples as observed by Mandayo et al. (2003). Korotchenkov et
al. (1999) concured these activation energies are related to O2-, O- and O2- which
are surface chemisorbed species and that the activation energy is approximately
equal to the energy of oxygen desorption from the SnO2 surface. The high values
of the activation energy are associated with O- desorption and the low values of
the activation energy are associated with O2- desorption, according to Williams
(1987). The values obtained from the experiments are in agreement with other
workers (Mandayo et al., 2003). In the present study, the difference in activation
energy merely show that the sensing mechanism favours the chemical spillover
model over the electronic energy barrier model as a mechanism for methane
sensing in air. This is in agreement with researchers like Bond et al. (1975) and
Yamazoe (1991). This is evident from the activation energy calculations which
show that the activation energy in methane was higher than in air, making it
impossible for enhanced electrical conduction via grain to grain and proceed via
spillover.
110
CHAPTER 7
RESPONSE AND RECOVERY TIME
7.1
Response and recovery time
The performance of a gas sensor is also discriminated via its response and
recovery time. The response and recovery time can be in seconds or minutes. As
a rule of thumb the shorter response and recovery times mean that the sensor is of
a good quality. Many factors influence the response and recovery time, amongst
them are flow rate of gas, level of doping in the base material, gas concentrations,
operational temperature. The work reported here is based on the sensitivity
curves in determination of the response and recovery times.
7.2
Effect of flow rate
The effect of flow rate of CH4 on the sensitivity was investigated for the
optimum composition sample (0.5 wt.% Pt-SnO2) operating at 400oC. The
sensitivity curves are shown in Figure 7.1.
111
500 sccm
400 sccm
/
Sensitivity (%)
1200
S (%)
1000
800
600
300 sccm
200 sccm
100 sccm
400
200
0
0
250
500
750 1000 1250 1500 1750
time
t (s)(s)
FIGURE 7.1: Sensitivity (S/) curves against time (t).
Methane (25 000 ppm in air) was introduced at t = 100 s and cut-off at t = 300 s.
The response and recovery times are shown in Table 7.1 and plotted as in Figure
7.2. For the flow rates 100-500 sccm, the response time decreases with flow rates.
The shortest response time was 60 s and the longest response time was
approximately 121 s. At flow rates ≥ 200 sccm, both the response and recovery
time vary linearly with the flow rate. The values of the slope are -0.466 s/sccm for
the recovery time and -0.083 s/sccm for the response time. This linear portion
could be utilized to control the response and recovery time via the flow rates. The
recovery time also show the same trend but the recovery times are much longer
than the response times. This was probably due to the poor expulsion of excess
methane at the top vent of the chamber. This vent was latter slightly enlarge and
this reduces the recovery time appreciably. Flow rates above 500 sccm was not
possible to be monitored and controlled in the setup.
112
Table 7.1: Response and recovery time for various flow rates.
Flow rate (sccm)
Response time (± 2s)
Recovery time (± 2s)
100
130
1317
200
85
771
300
69
701
400
62
647
500
60
631
1400
1200
1000
Time (s)
t (s)
slope =-0.466
/
800
recovery time
600
t = -0.4664χ + 864.2
400
slope = -0.083 s/sccm
200
response time
t = -0.083χ + 101.6
0
0
100
200
300 400
(sccm)
Flowχ rate
(sccm)
500
600
Figure 7.2: Response and recovery time (t) at various flow rates (χ).
113
7.3
Effect of methane concentrations
The pure SnO2 was doped with 0.5 wt.% Pt. The experimental
conditions were; operational temperature = 400oC, flow rate = 500 sccm,
methane concentration = 500-25 000 ppm, VS = 20 V, RL = 700 Ω. The sensitivity
curves are plotted as in Figure 7.3. The gas was introduced into the SECS at
t = 100 s and cut-off at t = 300 s.
1600
1400
25 000 ppm
1200
Sensitivity (%)
S/ (%) 1000
800
5000 ppm
600
2500 ppm
400
200
500 ppm
0
0
100
200
300
400
500
600
700
800
time
t (s)(s)
Figure 7.3: Sensitivity (S/) against time (t) at various CH4 concentrations.
The response and recovery times are shown in Table 7.2 and plotted in Figure 7.4.
The response time decreases with CH4 concentrations and the recovery time
increases with CH4 concentration. At low CH4 concentrations (≤ 500 ppm), the
sensitivity continues to plateau-off for 10-20 s before decreasing exponentially.
This characteristic is not shown at higher concentration. The sensitivity was still
114
high after the CH4 was cut-off because of the remains of CH4 in the chamber due
to poor expulsion in the SECS.
Table 7.2: Response and recovery time for various CH4 concentration.
CH4 conc. (ppm)
500
2500
5000
25000
Response time (± 2 s)
154
100
70
56
Recovery time (± 2 s)
440
550
600
1000
1200
t = 50c -124000
1000
recovery time
time (s)
800
t (s)
slope = 50 s/ppm
600
400
slope = -0.024 s/ppm
200
response time
t = -0.024c + 166
0
0
5000
10000
15000
20000
25000
30000
CH4 cconc.
(ppm)
(ppm)
Figure 7.4: Graph of response and recovery time (t) against CH4 concentrations
(c).
For the response time curve, a linear relationship exists (the slope is ~ -0.024
s/ppm) from 50 ppm up to 5000 ppm CH4. For the recovery time curve, there
exist a linear relationship (the slope is ~ 50 s/ppm at and above 2500 ppm CH4.
This too shows that both the response and recovery time could be controlled via
the CH4 concentrations.
115
7.4
Effect of operational temperature
The experimental conditions were; operational temperature = 280-400oC,
flow rate = 500 sccm, methane (in air) concentration = 25 000 ppm, VS = 20 V,
RL = 700 Ω. The sample used was the 0.5 wt.% Pt in SnO2. The gas was
introduced into the SECS at t = 100 s and cut-off at t = 300 s. The sensitivity
curves are plotted in Figure 7.5.
Graph of Sensitivity against time
1000
900
o
800
400 C
Sensitivity (%)
700
/
S (%)
600
o
350 C
500
400
o
300 C
300
o
280 C
200
100
o
250 C
0
0
100
200
300
400
500
600
700
800
900 1000
Time
t (s)(s)
Figure 7.5: Sensitivity (S/)-time (t) curves at operating temperatures 250-400oC.
The response and recovery times derived from the data are shown in Table 7.3
and plotted in Figure 7.6.
116
Table 7.3: Response and recovery time at various operating temperature.
Operating temp. (oC)
Response time (± 2 s)
Recovery time (± 2 s)
280
134
600
300
124
> 700
350
100
> 700
380
100
> 700
400
87
> 700
t (s)
Response time (s)
Graph of response time against operating temperature
160
140
120
100
80
60
40
20
0
slope = -0.3917 s/oC
250 270 290 310 330
350 370 390 410 430 450
o
Operating T
temperature
( C)
(oC)
Figure 7.6: Response time (t) against operating temperature (T).
The response time decreases with operating temperature in a linearly fashion. The
slope of the graph t against T yield a value of ~ -0.4 s/oC. The recovery time for
operating temperature 280oC was 600 s. For operating temperature > 280oC, the
recovery times were > 700 s. The long response time could be due to the poor
expulsion of methane in the SECS system and the oxidation of methane is
favourable at temperatures 300-400oC.
117
7.5 Effect of Pt loadings
The pure SnO2 was doped with 0.1-1.0 wt.% Pt. The experimental
conditions were; operational temperature = 400oC, flow rate = 500 sccm,
methane concentration = 25 000 ppm, VS = 20 V, RL = 700 Ω. The gas was
introduced into the SECS at t = 100 s and cut-off at t = 300 s. The sensitivity
curves are plotted in Figure 7.7.
1200
0.5 wt.%Pt
1100
1.0 wt.% Pt
1000
900
Sensitivity (%)
S/ (%)
800
0.75 wt.% Pt
700
0.3 wt.% Pt
600
0.1 wt.% Pt
500
400
300
200
100
0
0
100 200 300 400 500 600 700 800 900 1000
t (s)
time
(s)
Figure 7.7: Graph of sensitivity (S/) against time (t) for various Pt loadings.
The response and recovery times derived from the data are shown in Table 7.4
and plotted in Figure 7.8.
118
Table 7.4: Response and recovery time for various Pt loadings in SnO2.
Doping level (wt.% Pt)
0.10
0.30
0.50
0.75
1.00
Response time (± 2s)
39
50
42
33
34
Recovery time ( ± 2 s)
600
600
600
600
600
700
600
Recovery time
500
Time (s)
t (s) 400
300
200
100
Resopnse time
0
0
0.1 0.2
0.3 0.4
0.5 0.6
0.7 0.8
0.9
1
1.1
W* (wt.
%) Pt)
Pt loadings
(wt.%
Figure 7.8: Response and recovery time (t) against Pt loadings (W*).
The response times decrease with Pt loadings but increase at 0.3 wt.% Pt then
decrease thereafter. Note that 0.5 wt. Pt is the optimum loading in SnO2. The
recovery times are almost constant with a value in the vicinity of 600 s. Thus, the
recovery time is almost independent of the Pt loadings.
119
7.6 Conclusion
The response and recovery time in the experiments were based on the
parameters namely; flow rate of gas, gas concentrations, operating temperature
and doping level. Each parameter has a unique variation with both the response
and recovery time. The linear relationship between the parameters and the
response/recovery time could be utilized to control the methane sensor operation.
As the response time was adequately fast, the operation of the methane sensor
using the Pt-SnO2 ceramics was satisfactory. It was found that the response and
recovery time were sensitive to high flow rate of gas (500 sccm), optimum
operational temperature (400oC) and high CH4 concentration (25 000 ppm) but
less sensitive to Pt loadings in SnO2 (up to 1 wt.%).
120
CHAPTER 8
MICROSTRUCTURE AND PHYSICAL PROPERTIES ANALYSIS
8.1
Introduction
This part of the experimental work was focussed on the microstructure
and the physical properties of the active element Pt-SnO2 with an emphasis on the
optimum composition (0.5 wt.% Pt-SnO2). As shown in Chapters 5-7, this
material has an optimum operational temperature at 400oC and its electrical
properties can be utilized to sense methane in air. This material which is a
candidate for methane sensing in air is in the form of a ceramic which was
sintered at 1000oC for a duration of 1 hour. The analysis are as follows; Section
8.2: (TGA); Section 8.3: Density; Section 8.4: True porosity; Section 8.5: BET
and BJH analysis; Section 8.6: Vickers hardness; Section 8.7: Elastic modulus;
Section 8.8: High Resolution XRD; Section 8.9: Raman Shift spectroscopy;
Section 8.10: FTIR spectroscopy; Section 8.11: Surface analysis-EDAX, SEM,
AFM, XPS, Mössbauer spectroscopy and NMR spectroscopy.
8.2
Thermal Gravimetric Analysis (TGA)
The TGA was performed on the powder of the unsintered 0.5 wt.% Pt-SnO2 from
room temperature to 1000oC. About 13.888 mg of the mentioned sample was
used. The TGA result is shown in Figure 8.1.
121
W (%)
T (oC)
Figure 8.1: TGA of unsintered 0.5 wt.% Pt - SnO2 powder. Graph of weight loss
(W) against temperature (T).
The sample shows weight loss ~ 0.5% between (100 – 700)oC and almost
no weight loss at temperatures > 700oC. Between 37oC and 100oC, the weight
loss was mainly due to water molecules. The weight loss above 100oC up to
about 700oC is probably due to hydroxyl ions. Heiland and Kohl (1988) pointed
out that two types of OH- are generated by moisture which existed on the SnO2
through a reaction with O- and O2-. As such, the choice of sintering at
temperatures > 700oC and up to 1000oC is a favourable choice as the powder
shows no weight loss.
122
8.3
Density
The densities of the ceramics were determined for pure SnO2 and
(0.1-1.0)wt.%Pt-SnO2. The graph of density of pure SnO2 against temperature is
shown in Figure 8.2.
7.00
3
ρ (g/cm )
6.95
6.90
6.85
6.80
6.75
6.70
300
400
500
600
700
800
900 1000 1100
o
T ( C)
Figure 8.2: Graph of density (ρ) of pure SnO2 against temperature (T).
From the graph, the density increases with temperature. The density of the
ceramic sintered at 300oC is 6.870 g/cm3 and the density sintered at 1000oC is
6.920 g/cm3. Therefore, the fabricated SnO2 ceramics did not reach its full
density. The theoretical (full) density of SnO2 is 6.994 g/cm3 and the melting
point of pure SnO2 is 1127oC (Kumar Das, 1992).
The graph in Figure 8.3 shows the variation of the density of the doped
samples at various Pt loadings which were all sintered at 1000oC for 1 hour. The
density of the doped samples increases with Pt loadings. The density at 0.1 wt.%
Pt loading was ~ 7 g/cm3 and does not change much at Pt loading of 1 wt.%. The
inclusion of Pt in SnO2 has inevitably caused the bulk density to exceed the bulk
123
density of the pure SnO2. This feature adds strength to the modified ceramic
which was also shown by the elastic moduli calculated in Section 8.7.
7.10
7.08
7.06
3
ρ (g/cm )
7.04
7.02
7.00
6.98
6.96
6.94
6.92
6.90
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
W* (wt.%)
Figure 8.3: Graph of density (ρ) of doped SnO2 against Pt loading (W*) sintered
at 1000oC.
8.4
True Porosity
The corresponding true porosities for the samples mentioned in section 8.3
are shown in Figure 8.4 and Figure 8.5. The true porosity of the ceramics
decreases very little with both sintering temperature and Pt loadings. The
decrease in true porosity with sintering temperature was expected as at high
temperatures less pores are present in the ceramics. This is evident from the SEM
micrographs shown in Figure 8.6.
124
14.5
p/ (%)
14.4
14.3
14.2
14.1
300
400
500
600
700
800
900
1000
T (oC)
Figure 8.4: Graph of true porosity (p/) of pure SnO2 against sintering temperature
/
p (%)
(T).
15.0
14.8
14.6
14.4
14.2
14.0
13.8
13.6
13.4
13.2
13.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
W* (wt. %)
Figure 8.5: Graph of true porosity (p/) of doped SnO2 against Pt loadings (W*)
sintered at 1000oC.
125
A. Pure SnO2 sintered at 500oC
B. Pure SnO2 sintered at 700oC
Figure 8.6: SEM micrographs. A:Pure SnO2 sintered at 500oC and B: Pure SnO2
sintered at 700oC.
Notice that the ceramic grains that were sintered at 500oC are smaller than the one
sintered at 700oC and the pores are less visible in B compared to A. From the
density measurements, the ceramic samples were more dense if sintered at high
temperatures. Nevertheless, the magnitude of the true porosity change can be
regarded as negligible.
126
The variations of both true porosity and density with Pt loading is shown
in Figure 8.7. It shows the similar variation as in the case of sintering at 1000oC.
The true porosity decreases with Pt loadings but the density increases with Pt
loadings. This is evident from SEM micrographs as shown in Figure 8.8. The
increase in bulk density is a desired feature for a gas sensing active element in a
gas sensor. The negligible decrease in true porosity nevertheless still yielded a
good sensitivity in the case of the 0.5 wt.% Pt-SnO2 ceramic sintered at 1000oC
and operated at 400oC. The corresponding bulk density is ~ 7.04 g/cm3 and true
porosity of 14.05%. The true porosity showed < 0.1% change when the Pt
loading increased from 0.1-1.0 wt.%. From the SEM micrographs (Figure 8.8) it
is also reasonable to assume that methane can easily pass through the pores and
interact with the sample surface as to initiate the adsorption and desorption
processes.
7.05
14.15
14.10
7.00
3
ρ (g/cm )
6.95
14.00
/
p (%)
14.05
13.95
6.90
13.90
ρ
p
13.85
6.85
/
13.80
6.80
0.1
0.3
0.5
0.7
1
W* (wt. %)
Figure 8.7: Graphs of true porosity (p/) and bulk density (ρ) against Pt loadings
(W*) sintered at 1000oC.
127
A. 0.1 wt % Pt-SnO2, sintered at 1000oC
B. 0.5 wt % Pt-SnO2, sintered at 1000oC
Figure 8.8: SEM micrographs. A:0.1 wt.% Pt sintered at 1000oC and B: 0.5 wt.%
Pt sintered at 1000oC.
128
8.5
Brunauer-Emnett-Teller (BET) analysis
The BET was performed on the undoped powdered sample and the effects
of sintering temperature on the particle size is shown in Figure 8.9. The particle
10
200
9
180
8
160
7
140
6
120
5
100
4
80
3
R/ (nm)
S (m2/g)
size was calculated using equation 4.11.
60
S
2
40
R/
1
20
0
0
25
500
700
1000
T (oC)
Figure 8.9: Graph of BET specific surface area (S) and particle size (R/) versus
sintering temperature (T).
The BET specific surface area decreases with sintering temperature. This
give rise to bigger particle size ceramic at high sintering temperature. The
starting material has an average particle size of ~ 92 nm and when it was sintered
at 1000oC the particle size grew to ~ 174 nm. The effects of Pt loading in SnO2 is
shown in Figure 8.10. The ceramics were all sintered at 1000oC for 1 hour. The
doped ceramics also show a decrease in BET specific surface area and increasing
particle size with increasing Pt loading. The average particle size with 0.1 wt.%
Pt loading was ~ 221 nm and Pt loading of 1 wt. % in SnO2 gave rise to an
average particle size of ~ 358 nm.
129
4.5
350
4
300
3.5
S (m2/g)
2.5
200
2
150
1.5
1
0.5
S
100
R/
50
0
R/ (nm)
250
3
0
0.1
0.3
0.5
0.8
1
W* (Wt. %)
Figure 8.10: Graph of BET specific surface area (S) and particle size (R/) versus
Pt loadings (W*) sintered at 1000oC.
The 0.5 wt% Pt-SnO2 which was found to be the optimum composition
has an average particle size of 225 nm.The adsorption and desorption curves also
indicate the presence of micropores with micropore volume of ~ 0.001847 cm3/g.
Observations from TEM images (Figure 8.12) revealed that the starting powders
have smaller particle size compared to the particle size obtained using BET
analysis. This is attributed to the powder tends to be agglomerated and this
contributes to the discrepancy. The typical adsorption and desorption curves
obtained from the BET analysis was typically type III ( Figure 8.11) which
according to Gregg (1982) will be unlikely to yield a true value of the surface
area.
130
Figure 8.11: Adsorption/Desorption curve from BET analysis.
131
A
B
Figure 8.12: TEM: A fresh Pt powder; B fresh SnO2 powder.
Typical pore size distributions for the powders were derived from the
isotherms using the BJH method as shown in Figure 8.13. The pores are
distributed in two distinct regions; (i) between 25 and 32 nm, with a distribution
centred at ~ 28 nm, and (ii) between 10 and 100 nm, with a distribution centred at
~ 50 nm.
132
Figure 8.13: Pore size distributions by the BJH method.
133
8.6
VICKERS HARDNESS
The Vickers hardness measurements were performed on the doped
0.5 wt.% Pt-SnO2 ceramic samples by the microindentation technique with a
1.961 N load. The influence of temperature on the ceramic samples is shown in
Figure 8.14.
15
14
HV (GPa)
13
12
11
10
9
8
500
550
600
650
700
750
800
850
900
950 1000
o
T ( C)
Figure 8.14: Graph of Vickers hardness (HV) against temperature (T).
The Vickers hardness of the doped samples decreases with temperature. The
Vickers hardness for this doped material is similar to the value for TiO2 rutile
ceramics (Hv = 7 -11 GPa) reported by Lackey et al. (1987). The decrease in
Vickers hardness is attributed to the increase in density of the sample with
temperature. This leads to a decrease in porosity and hence a decrease in the
Vickers hardness. Wang and Hon (1999) suggest that the decrease in hardness, H
as temperature increases can be fitted by a functional relationship of the form
H = Hoe-aT, where Ho and α are constant and T is the test temperature (oC). As
such, the dependence of hardness on density shows that hardness may
demonstrate energy density characteristics. As temperature rises, the energy
density and therefore the hardness decreases due to thermal expansion.
134
Figure 8.15 shows the influence of Pt loadings on the ceramic samples
which were sintered at 1000oC. The Vickers hardness of the doped samples
increases with Pt loading. The density of these doped ceramics increases with Pt
doping but there was a negligible change in the true porosity of the ceramics.
As such it is expected that the Vickers hardness to increase with Pt doping. The
Pt which was incorporated in the SnO2 adds hardness to the modified ceramic.
Thus this is also a feature of stability apart from the mechanical strength shown
by the material which is discussed in the following section. As mentioned earlier,
hardness is the resistance against force which causes damage. The doped ceramic
7.10
20
18
16
14
12
10
8
6
4
2
0
7.08
7.06
7.04
7.02
ρ
7.00
H
6.98
0.5
1
2
2.5
W* (Wt. %)
Figure 8.15: Graph of Vickers hardness (HV) and bulk density (ρ) versus Pt
loading (W*) sintered at 1000oC.
3
ρ (g/cm )
HV (GPa)
is therefore durable.
135
8.7
ELASTIC MODULUS
The elastic moduli were determined for the ceramic samples by ultrasonic
non-destructive testing. The results are shown in Figures 8.16 and 8.17.
Both the transverse and longitudinal wave velocities increase with Pt loading.
However, the magnitude of the transverse wave was much greater than the
longitudinal wave. The elastic moduli show a sharp increase in magnitude for Pt
loading < 0.3 wt.% and a gradual increase thereafter up to 1 wt.% Pt. However,
the greatest magnitude was the bulk modulus followed by Young modulus and
shear modulus. The magnitude obtained reflects the stability of the doped
ceramic and for Pt loadings ≥ 0.5 wt.%, resulted in a slight increase only in
magnitude of the bulk, Young and shear modulus.
7
transverse
wave
6
v x103(m s-1)
5
4
3
longitudinal
wave
2
1
0
0
0.2
0.4
0.6
0.8
W* (wt. %)
Figure 8.16: Graph of velocity (v) of transverse/longitudinal wave versus Pt
loadings (W*) sintered at 1000oC.
1
136
90
80
Bulk Modulus
70
E (GPa)
60
Young
Modulus
50
40
30
Shear
Modulus
20
10
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
W* (wt.%)
Figure 8.17: Graph of elastic modulus (E)-bulk, Young and shear modulus
against Pt loadings (W*) sintered at 1000oC.
8.8
HIGH RESOLUTION X-ray DIFFRACTION (HRXRD)
8.8.1 STARTING POWDERS
X-ray diffraction was performed on both doped and undoped samples.
Figure 8.18 shows the XRD diffraction pattern of the starting powder. The
starting powder shows the main peaks at 2θ ~ 26.8o, 34o, 38o, 52o and 55o which
correspond to the (110), (101), (200), (211) and (220) respectively. Thus, this
shows that it is a typical rutile tetragonal cassiterite SnO2 with peaks as reported
in the literature (JCPDS, 1997a). The mean crystallite size calculated using
equation 4.11 was ~ 14 nm. TEM images of the starting SnO2 powder are shown
in Figure 8.22. The mean crystallite size calculated was approximately (± 5%) the
particle size of the powder shown in the TEM images.
137
2500
(110)
(
(101)
(211)
I (arb. unit)
2000
1500
1000
(200
500
(220)
0
25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55
2θ ( ο)
Figure 8.18: XRD pattern of fresh pure SnO2 powder. Plot of intensity (I) against
Bragg angle (2θ).
Figure 8.19: Typical TEM images of fresh pure SnO2 powder.
138
The XRD pattern of the platinum powder is shown in Figure 8.20. The
peaks at 2θ ~ 40o and 46.3o are typical Pt metal peaks (JCPDS, 1997b). These
two peaks are sufficient for the purpose of the experiment as they are no peak
appearing at these 2θ values for the pure SnO2.
18000
16000
(111)
14000
I (c.p.s)
12000
10000
8000
(200)
6000
4000
2000
0
25 27 29 31 33 35 37 39 41 43 45 47 49
2θ ( Ο )
Figure 8.20: XRD pattern of fresh Pt powder.
The calculated mean crystallite size was ~ 18 nm and the particle size of
the platinum powder as observed from TEM images (Figure 8.21) was ~ 14 nm.
As such, the observed particle size and mean crystallite size differs slightly but
reliable as compared to BET particle size calculations. The starting powders of
both SnO2 and Pt showed approximately the same magnitude of particle sizes.
139
.
A
B
Figure 8.21: Typical TEM images of Pt powder: A; scale bar 50 nm, B; scale bar
20 nm.
140
8.8.2
Pt–SnO2 CERAMICS
8.8.2.1 SINTERING EFFECTS ON PURE SnO2
The influence of sintering in pure SnO2 is shown in Figure 8.22.
16000
14000
o
1000 C
12000
I (arb. unit)
I (arb.unit) 10000
o
900 C
8000
o
800 C
6000
o
700 C
4000
o
600 C
2000
o
500 C
0
25
30
35
40
45
50
55
οo
2θ2θ( ( ) )
Figure 8.22: XRD pattern of pure SnO2 sintered at temperatures 500-1000oC.
Diffractograms show the plot of intensity (I) against Bragg angle (2θ).
All diffractograms show the main peaks (110), (101) and (211) of the tetragonal
cassiterite SnO2. An interesting feature of the XRD pattern is that the FWHM of
the intensity decreases slightly with sintering temperature. Hence, the mean
crystallite size of the SnO2 increased with sintering temperature. Figure 8.23
shows the calculated mean crystallite size which varies with sintering
temperature.
141
60
55
(110)
Rx (nm)
50
45
(101)
40
35
(211)
30
25
(220)
20
500
550
600
650
700
750
800
850
900
950 1000
T (oC)
Figure 8.23: Calculated mean crystallite size (RX) of pure SnO2 against sintering
temperature (T).
The mean crystallite size was the largest in the (110) direction and the smallest
was in the (220) direction. In both cases, a marked increase in the mean
crystallite size was observed at sintering temperatures of 800oC and above, whilst
the mean crystallite size in both (101) and (211) showed almost monotonic
growth between sintering temperatures 500-1000oC.
8.8.2.2 EFFECTS OF Pt LOADINGS ON PURE SnO2
The XRD pattern of Pt-SnO2 ceramics sintered at 1000oC are shown in
Figure 8.24. The modified ceramics show XRD reflections almost similar to that
of the pure SnO2. Apart from the main peaks (110), (101) and (211) of the
tetragonal cassiterite SnO2, at Pt loadings of 2 wt.% and above, prominent peaks
of Pt (111) and Pt (220) metal evolved at 2θ ~ 39.9o and 46.3o respectively. It
was noticed that there were no PtOx formed as they should show up as strong
142
diffraction peaks at 2θ = (50 – 55)o or (15-30)o (Morazzoni et al., 2001). This
shows that the Pt have a remarkable thermal stability up to 1000oC in the SnO2
+
+
+
+
3000
fresh SnO2 powder
+
I (arb. unit)
2500
+
* Pt peak
+ Sn peak
o
SnO2 sintered 1000 C
2000
0.5 wt.% Pt
1500
1 wt.% Pt
1000
*
*
*
*
*
500
0
25
30
35
2 wt.% Pt
3 wt.% Pt
*
40
45
5 wt.% Pt
50
55
60
ο
2θ ( )
Figure 8.24: XRD pattern of Pt-SnO2 at Pt loadings 0.5-5 wt % sintered at
1000oC. Diffractograms show the plot of intensity (I) against Bragg angle (2θ).
matrix. The mean crystallite size of the Pt (111) was ~ 33 nm while the mean
crystallite size of the Pt(111) in the fresh Pt powder was ~ 12 nm. Such large Pt
crystallites like the one found in the modified sample are said to be resistant
against oxidation (Balint et al., 2002) so it could be the reason for the absence of
PtOx. The absence of the Pt peaks at lower doping could be either due to
evaporation at high temperatures or buried well in the recess of the ceramic or too
small to be diffracted by the X-rays. The mean crystallite size of the Pt-SnO2
ceramics at various Pt loadings is shown in Figure 8.25.
143
60.0
55.0
(110)
RX (nm)
50.0
45.0
(101)
40.0
35.0
(211)
30.0
(220)
25.0
20.0
0
1
2
3
4
5
W* (wt. %)
Figure 8.25: Mean crystallite size (RX) of Pt- SnO2 at Pt loadings (W*) sintered at
1000oC.
Again, the mean crystallite size was the largest in the (110) direction and
the least growth was in the (220) direction and the mean crystallite size increased
with Pt loadings.
The mean crystallite size of the 0.5wt.%Pt-SnO2 at different sintering
temperatures are shown in Figure 8.26. The mean crystallite size of the
0.5wt.%Pt- SnO2 increased by small amount (2-5 nm) with sintering temperature.
Such increment is also small when compared to the case of sintering pure SnO2.
Also, the inclusion of 0.5 wt Pt in SnO2 produces smaller mean crystallite size.
Thus, the Pt inhibits the grain growth of the modified sample. It is obvious that
the growth of the crystallites in this case is different from the mechanisms such as
coalescence (Family and Meakin, 1988), Ostwald ripening (Lin et al., 1994) and
nucleation (Yu, 1997). The XRD pattern of the doped ceramics has a higher
intensity when compared to that of the pure SnO2 as shown Figure 8.27.
144
50
(110)
45
RX (nm)
40
(101)
35
30
(211)
25
(220)
20
500
600
700
800
900
1000
T (oC)
Figure 8.26: Mean crystallite size (RX) of 0.5 wt.% Pt- SnO2 against sintering
Intensity (arb. unit)
temperatures (T).
I
(c.p.s)
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
doped
undoped
33.5 33.6 33.7 33.8 33.9 34.0 34.1 34.2 34.3 34.4 34.5
2θ( ο )
2θ
Figure 8.27: XRD of doped (3 wt. % Pt) and undoped sample sintered at 1000oC
in the (101) direction. Diffractogram shows a plot of intensity (I) against Bragg
angle (2θ).
145
Morazzoni et al. (2001) suggested that Pt diffuses into the SnO2 structure,
modifying the structure factors and therefore the reflection intensity of the phase.
SEM micrographs of the doped ceramics (Figure 8.28) showed that Pt metal
clusters on the surface of the modified ceramics (Pt-SnO2) which is an indication
of the Pt diffusion. However, the difference in the intensity could also be
attributed to density and grain size as observed from earlier experiments. There is
also a broad shoulder on the high angle side of the Pt (111) peak (Figure 8.29)
which is the Kα2. A closer look at both the doped and undoped XRD reflection in
the (101) direction (Figure 8.30) revealed that the peak of the doped sample
(3 wt.% Pt-SnO2) shifted ~ 0.01o to lower angle with respect to the peak of the
undoped sample. This shift has a magnitude which is in the same order of the
error in 2θ and this dismiss the probable of the shift in the d values.
Figure 8.28: Dispersion of 0.5 wt.% Pt-SnO2. The white speckles are Pt clusters
against the dark background of SnO2.
146
400
Intensity (arb. unit)
350
I
(c.p.s)
300
250
200
broad-shoulder
150
100
50
0
39.5 39.6 39.7 39.8 39.9 40.0 40.1 40.2 40.3 40.4 40.5
2θ2θ( ο )
Figure 8.29: Broad shoulder formation of Pt (111) at high angle side of 3 wt.%
Pt-SnO2 sintered at 1000oC.
12000
o
0.01
Intensity
(arb. unit)
10000
I
(c.p.s)
doped
8000
6000
undoped
4000
2000
0
33.80 33.85 33.90 33.95 34.00 34.05 34.10 34.15 34.20
22θθ( ο )
Figure 8.30: Peak shifts of 0.01o to lower angle side of doped (3 wt.% Pt-SnO2)
sintered at 1000oC with respect to the peak of the undoped SnO2 sintered at
1000oC in the (101) direction.
147
8.8.2.3 INTENSITY RATIO (I211/ I220)
The peak intensity ratio for the SnO2 system (Cirera et al., 2000) can be
expressed as,
/
71.36 + 4.48α
I 211 F211
= 2 ≈
/
I 220 F220
69.79 + 9.98α
2
2
2
(8.1)
where F/hkl is the structure factor and α is the occupation factor which keeps unit
value when every oxygen position is occupied by an oxygen atom. For highenergy electron diffraction, the (220) reflection has a higher contribution from the
oxygen. In the tin oxide rutile structure, the (220) plane intercept high number of
oxygen atoms as compared to the (211) plane. It is predicted from the equation
above that the intensities ratio decreases when the occupancy grows (relaxation of
vacancies).
The plot of such intensities from the experiment are shown in Figure (8.31
- 8.33). From graphs in Figures 8.31 and 8.33, the intensity ratio decreases with
sintering temperature. The effect of sintering was to reduce the intensity ratio.
Thus, the oxygen vacancies vanish (but not completely) with sintering
temperature. However, this is limited by the amount of Pt in the SnO2. At doping
levels > 2 wt.%, the relative intensity ratio rises as shown in Figure 8.32.
148
5.0
4.9
4.8
I 211/ I 220
4.7
4.6
4.5
4.4
4.3
4.2
4.1
4.0
500
600
700
800
900
1000
T (oC)
Figure 8.31: Graph of intensity ratio (I 211/I 220) against sintering temperature (T)
for pure SnO2.
1.05
1.04
1.03
I 211/ I 220
1.02
1.01
1.00
0.99
0.98
0.97
0.96
0.95
0
1
2
3
4
5
W* (wt. %)
Figure 8.32: Graph of intensity ratio (I 211/I 220) against Pt loadings (W*) sintered
at 1000oC.
149
1.04
1.03
1.02
I 211/I 220
1.01
1.00
0.99
0.98
0.97
0.96
700
750
800
850
900
950
1000
o
T ( C)
Figure 8.33: Graph of intensity ratio (I211/I220) against sintering temperature (T)
for 0.5 wt.% Pt-SnO2.
8.8.2.4 INTENSITY RATIO (Ihkl/I110)
Matsuhata et al. (1994), considered the numerical difference between the
(220) and (110) reflections in the effects of occupancy growth. For high energy
electron diffraction, the (110) reflections depends strongly on the ionicity of tin.
Figures 8.34-8.36 shows the intensity ratio for undoped and doped samples.
The intensity ratio in all cases shows a low magnitude and does not change much
with either sintering temperature or Pt loadings. The intensity ratio was highest in
the (211) and lowest in the (220) directions. Thus, this suggest that the
incorporation of Pt in SnO2 has not much effect on the ionicity of tin.
150
1.6
(211)
1.4
1.2
(101)
I hkl/ I 110
1.0
0.8
0.6
0.4
(220)
0.2
0.0
500
550
600
650
700
750
800
850
900
950
1000
T (oC)
Figure 8.34: Graph of intensity ratio (Ihkl/I110) against sintering temperature (T)
for pure SnO2.
1.6
1.4
(211)
1.2
(101)
I hkl/ I 110
1.0
0.8
0.6
(220)
0.4
0.2
0.0
700
750
800
850
900
950
T (oC)
Figure 8.35: Graph of intensity ratio (Ihkl/I110) against sintering temperature (T)
for 0.5 wt.% Pt-SnO2.
1000
151
1.6
(211)
1.4
1.2
(101)
I hkl / I 110
1.0
0.8
0.6
0.4
(220)
0.2
0.0
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
W* (wt.%)
Figure 8.36: Graph of intensity ratio (Ihkl/I110) against Pt loading (W*) sintered at
1000oC.
8.8.2.5 INDUCED MEAN STRAIN CALCULATIONS
d relaxed − d T
The mean strain, < e2 >1/ 2 = 100 hkl relaxed hkl was calculated and the
d hkl
results are shown in Figures 8.37-8.39. The undoped sample shows almost
compressive strain except at sintering temperatures between 700 and 850oC. The
most strain was observed in the (110) direction as the bond between metal and
oxygen decomposes mainly in the (110) direction for TiO2 in the same rutile
lattice (Mattossi, 1951). In the doped sample, the strain in the (220) direction was
predominantly tensile for T > 700oC. According to Cirera et al. (2000), when an
oxygen atom is removed, the resulting Sn-O bond relaxes, thus increasing the
<e2>1/2
152
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1500
-0.2
(110)
(220)
(211)
(101)
550
600
650
700
750
800
T (oC)
850
900
950 1000
-0.3
-0.4
-0.5
-0.6
Figure 8.37: Graph of mean strain (<e2>1/2) of the atoms in plane (hkl) in the
normal direction to the plane against sintering temperature (T) for undoped SnO2.
T (oC)
0
500
-0.1
550
600
650
700
<e2>1/2
-0.2
((
-0.3
-0.4
750
(
(220)
800
850
900
950 1000
(211)
(101)
(110)
-0.5
-0.6
Figure 8.38: Graph of mean strain (<e2>1/2) of the atoms in plane (hkl) in the
normal direction to this plane against sintering temperature (T) for 0.5wt.%PtSnO2.
153
0.8
0.7
(110)
0.6
<e2>1/2
0.5
0.4
(101)
0.3
0.2
(220)
0.1
(211)
0
0
1
2
3
4
5
W* (Wt. %)
Figure 8.39: Graph of mean strain (<e2>1/2) of the atoms in plane (hkl) in the
normal direction to this plane against Pt loadings (W*) sintered at 1000oC.
distances between atoms and moving away from the (220) atomic planes. Note
that the O-Sn-O bonds are orthogonal to the (220) atomic plane. It is possible that
a deformation potential that mainly decomposes in the (220) direction was
induced by a high density of oxygen vacancies in the lattice. The Pt loading at 1
wt. % shows the most strain but relaxes thereafter. The sintering at high
temperatures or any level of Pt loadings does not reduce the strain in either the
doped or the undoped samples. This is also shown (Table 8.1) in the
discrepancies in calculating dhlk, the distance between planes with (hkl) Miller
indexes by means of Bragg formula. Thus, the distortions are directly associated
with the observed strain. The distortions were calculated using the equations
(4.14) and (4.15) (Chapter 4) which consider the differences between planes (110)
and (220). The results are shown in Figure 8.40-8.42.
6
154
Table 8.1: Distance dhkl calculated using Bragg formula for 0.5 wt% Pt-SnO2
sintered between 700-1000oC.
T /dhkl
d220 (Å)
d110 (Å)
d101 (Å)
d211 (Å)
700oC
1.678
3.354
2.648
1.767
800oC
1.675
3.340
2.640
1.764
900oC
1.675
3.341
2.640
1.764
1000oC
1.672
3.327
2.632
1.761
Yu (97)
1.672
3.351
2.637
1.762
JCPDS(97a)
1.675
3.347
2.643
1.764
155
1.13
1.1295
δc (Ǻ )
1.129
1.1285
1.128
1.1275
1.127
1.1265
500 550 600 650 700 750 800 850 900 950 1000
T (oC)
T (oC)
0.004
δa (Ǻ)
0.002
0
500
-0.002
550
600
650
700
750
800
850
900
950 1000
-0.004
-0.006
-0.008
Figure 8.40: Distortions δc and δa against sintering temperature (T) for pure
SnO2.
156
1.128
1.128
δc (Ǻ)
1.127
1.127
1.126
1.126
1.125
1.125
1.124
700
750
800
850
900
950
1000
900
950
1000
T (oC)
T (oC)
δa (Ǻ)
0.002
0
700
-0.002
750
800
850
-0.004
-0.006
-0.008
Figure 8.41: Distortions δc and δa against sintering temperature (T) for
0.5 wt.%Pt-SnO2.
157
1.1258
1.1256
δc (Ǻ)
1.1254
1.1252
1.1250
1.1248
1.1246
1.1244
1.1242
0
1
2
3
4
5
4
5
W* (wt. %)
δa (Ǻ)
W* (wt.%)
0.0000
-0.0010 0
-0.0020
-0.0030
-0.0040
-0.0050
-0.0060
-0.0070
-0.0080
-0.0090
-0.0100
1
2
3
Figure 8.42: Distortions δc and δa versus Pt loadings (W*) sintered at 1000oC.
158
The effects of sintering in both doped and undoped sample was reduction in the
distortions at high temperatures. For the doped sample, the evolution of the
distortion with sintering temperature suggests that there was a strong reduction of
the vacancies at temperatures between 700oC and 800oC. The distortions were
greatly reduced when the sample was sintered above 800oC. This suggests that
sintering at temperatures at and above 800oC is a way to minimize distortions
w.r.t. the (110) and (220) planes. The reduction in distortions mentioned means
the oxygen vacancies decompose. The magnitude of δa < δc and the
incorporation of Pt in SnO2 does not reduce the distortions especially after 1 wt.%
Pt loadings.
The hypothesis of oxygen vacancy reduction with sintering temperature is
inevitably likely. In the rutile structure of SnO2, the (110) and (220) planes
intercept more oxygen atoms than the (101) and (211) planes. Thus, the changes
in the morphology (distortions) of the mentioned planes can be related with the
evolutions of the oxygen vacancies. These oxygen vacancies result in mechanical
effects induced on the lattice which give rise to the induced strain. Consequently,
the presence of the oxygen vacancies on the lattice parameters induced
distortions. The calculated distortions calculations δa and δc were based on the
formula proposed by Yu et. al (1997) in the form
4 sin 2 θ
λ2
h2 + k 2 l2
=
+ 2 and
a2
c
Cirera (2000) computed δa = a220 - a110 and δc = c101-110 – c101-220. As such, the
doped and undoped common characteristic is that they originate from oxygen
defective tin oxide in which the oxygen vacancies diminish with sintering
temperature. Other hypotheses that could also produce the variations in the
structural parameter to be considered are interstitial tin in the tin oxide lattice
(Agashe et al., 1991) and the presence of hydroxyl group. The former hypothesis
showed that the defective tin oxide produced an anomalous increase in intensity in
the X-ray diffractogram and the that high magnitude stress was observed due to
O-H groups being substituted in the oxygen sites. However, these two hypotheses
can be disputed from the samples investigated as the magnitude of stress is small,
there was no evidences for the presence of O-H group from XPS, Raman, FTIR
and no anomalous XRD peak intensity was observed.
159
8.9
RAMAN SHIFT SPECTROSCOPY
The Raman-shift modes obtained for pure SnO2 sintered at temperatures
100-1000oC are shown in Figure 8.43. The main cited modes of vibrations are at
473, 633 and 773 cm-1. These active modes were also observed by Katiyar et al.
(1971).
I (arb. unit)
δ (cm-1)
Figure 8.43: Raman Shift spectra; plot of intensity (I) against Raman shifts (δ) of
the undoped SnO2 dry-pressed ceramics sintered at temperatures 100-1000oC.
An interesting feature in the Raman –shift spectrum is the appearance of an active
mode at ~ 555 cm-1. This is identified as the A2g mode which was supposed to be
inactive Raman mode in a perfect rutile (Yu et al., 1997). The appearance of this
mode in the sample indicates that there is a high density of oxygen vacancies and
with this the composition rule ω 2 ( A2 g ) = ω 2 ( B1 g ) + ω 2 ( B 2 g ) − ω 2 ( A1 g ) that
can be verified. However, this mode vanishes at T > 300oC. Another feature of
the spectrum is that the peaks become sharper as the sintering temperature
increases. This indicates that the grain size of the SnO2 increases with sintering
160
temperature which was also shown by HRXRD. This is also observed by Yu et
al. (1997). The Eg mode at ~ 473 cm-1, A1g mode at ~ 633 cm-1 and B2g mode at ~
773 cm-1 become prominent with sintering temperature too. The Raman-shift
obtained for doped SnO2 sintered at a temperature of 1000oC is shown in Figure
8.44.
I (arb. unit)
δ (cm-1)
Figure 8.44: Raman Shift spectra; plot of intensity (I) against Raman shift (δ) of
the Pt- SnO2 dry-pressed ceramics sintered at 1000oC.
The main cited mode of vibrations is at ~ 633 cm-1. This A1g mode appears in all
the doped samples but the intensity decrease and the FWHM increase slightly
with Pt loading. The other two modes at 774 and 472 cm-1 almost diminishes at 1
wt.% Pt loading. The diminution of this effect is attributed to growth in grain size
(due to Pt incorporation and sintering at 1000oC), thus surface/volume
contribution to Raman spectra diminishes. Cabot et. al (2000) expected similar
results for using Pt or Au and usage of such noble metals alters the tin oxide
crystallography which was deduced by analysing the Raman spectra. The slight
increase in FWHM of the A1g mode with Pt loading seemed to indicate the
161
inhibitions of grain growth up to 1 wt.% Pt loadings which is also shown in the
HRXRD analysis. Such change is also considered as a distortion due to the
addition of Pt in the SnO2 matrix which is again shown in the HRXRD distortion
calculations. This distortion manifests not only on the surface but extends to the
bulk as SEM micrographs show clustering of Pt, both on the surface and also in
the bulk.
8.10
FTIR SPECTROSCOPY
The FTIR on the fresh SnO2 powder showed very small concentrations of
H2O and OH as shown in Figure 8.45.
Sn-O
T (%)
cm-1
Figure 8.45: FTIR spectra of the fresh pure SnO2 powder.
162
If hydroxyl groups were present, they will appear as a broad band between 3200
and 3600 cm-1 as these are assigned to O-H stretching and deformation vibrations
of weak bound water (Li and Kawi, 1998). The wet method of obtaining SnO2
based materials is reported to show prominent peaks of water and surface
hydroxyls even at temperatures 400-440oC (Korotchenkov, 1999; McAleer,
1987). The broad bands between 400-800 cm-1 are attributed to the framework
vibrations of SnO2 (Sala and Trifiro, 1974) and the intense band at 1518 is due to
Sn-O stretching (Chaudhary, 2001). The FTIR spectrum of the pure Pt is shown in
Figure 8.46.
T (%)
cm-1
Figure 8.46: FTIR spectra of fresh Pt.
163
The spectrum shows no contaminations and negligible amount of H2O. The
evolution of the FTIR spectra with Pt loading is shown in Figure 8.47.
632
T
(arb.unit)
cm-1
Figure 8.47: FTIR spectrum of Pt-SnO2 ceramic sintered at 1000oC at various Pt
loadings.
The increasing amount of Pt led to the evolution of the prominent minimum at
632 cm-1 found in the pure SnO2 sample. This was also observed for the RamanShift spectrum in which this corresponded to the A1g mode. Also, the decrease in
transmissions was due to the growth in particle size with Pt doping as the
scattered light intensity is proportional to d3f4, where d is particle diameter and
f is the frequency of the incident light (Sergent, 2002). A strong decrease of
transmission is proposed to be induced by free electrons in the conduction band as
Pt concentrations increase. This suggests the presence of oxygen vacancies at the
surface of the doped SnO2 crystallites.
Only the bands at 630 and 550 cm-1 were observed which correspond well
to those given in the literature for SnO2 single crystal (Hofmeister et al., 1990) as
stannic oxide belongs to the space group D4h for which four fundamental lattice
164
modes are expected to be IR active as predicted from the group-theoretical
analysis of the rutile structure. However, the bands typical of SnO2 skeletal
vibrations became progressively less defined as a result of sintering (Scarlat et al.,
2003) and the incorporation of Pt in SnO2. Large and broad adsorption bands
with low value of transmission as observed in the doped samples can be attributed
to sintering too.
However, there was an overall an increase in the absorption of the FTIR
spectrum after exposure to 25 000 ppm of CH4 at an operating temperature of
400oC with a maximum shown by the 0.5 wt.% Pt-SnO2 as shown in Figure 8.48.
The electrical conductivity of an n-type semiconductor decreases when the
adsorption of oxygen on a semiconductor surface occurs. This is caused by a
decrease in the electron density in the conduction band due to the formation of
ionosorbed oxygen species such as O- or O2- (and electrons become localised).
When CH4 gas is adsorbed, electrons are injected into the conduction band and
the electrical conductivity increases. Thus, the free carrier density is modulated
by CH4 gas adsorbing on the doped SnO2 sample surface. By observing the
variation of the background infrared absorption under oxidizing and reducing
gases, it will enable one to determine the type of its semiconductivity (Baraton et
al., 2001). The results of these adsorptions on the FTIR of the doped SnO2
showed an increase in the overall absorption under the CH4 gas. This is attributed
to the desorption and oxidation of methane on the surface of the sample which
produces CO2 and H2O molecules which contributes to a higher absorption
intensity in the FTIR spectrum as compared to air. This absorption becomes more
pronounced in the Pt-SnO2 samples in which the activity is catalysed by Pt, hence
increasing the absorption intensity. Such increase in absorption is observed in the
samples and this indicates that the Pt-SnO2 is an n-type semiconductor.
165
A/
(arb. unit)
cm-1
Figure 8.48: FTIR absorption spectrum of Pt-SnO2 ceramic sintered at 1000oC at
various Pt loadings after exposure to 25 000 ppm CH4 at 400oC.
166
8.11
SURFACE ANALYSIS
The surface analysis was based on EDAX, SEM, TEM, AFM, XPS,
Mössbauer spectroscopy and NMR spectroscopy.The evolutions of the surface
profiles due to sintering and Pt loadings are described.
8.11.1 EDAX
The doped samples contained very low amounts of Pt and the EDAX
analysis showed almost no Pt on the surface. Figure 8.49 shows the EDAX result
for pure SnO2 and Figure 8.50 shows the EDAX result for the 0.5 wt.%Pt-SnO2.
There were no prominent peaks that show the existence of Pt metal or PtOx.
Thus, from the EDAX spectrum no Pt species were detected and the concentration
of Pt species were too low to be detected. To ascertain this, the SEM technique
was employed. Figure 8.51 shows the SEM micrographs of the surface and crosssection of the 0.5 wt.% Pt-SnO2. The micrographs clearly show the existence of
Pt metal clusters (shown as white specks) against the dark background of SnO2.
The distribution of these clusters is fairly uniform which is not only found on the
surface of the SnO2 but also in the bulk. Further more, there were more Pt
clusters found on the surface as compared to the bulk (beneath the surface). This
is probably due to the difference in density (the density of pure SnO2 starting
powder was 6.95 g/cm3 whilst the density of the Pt powder was 21.5 g/cm3). Thus
the Pt metal exists but is below the detectable level. The presence of Pt clusters at
the grain surface favours the Fermi energy control during chemical sensing
(which requires agglomeration of dopants at the surface).
167
Figure 8.49: EDAX spectrum of undoped (pure SnO2) sintered at 1000oC.
Figure 8.50: EDAX spectrum of doped 0.5 wt.%Pt- SnO2 sintered at 1000oC.
168
Surface of 0.5wt.%Pt-SnO2
( scale-bar 20µm)
Cross-section of 0.5wt.%Pt-SnO2
( scale-bar 500µm)
Figure 8.51: SEM micrographs of doped 0.5 wt.% Pt-SnO2 sintered at 1000oC.
The white speckles are Pt metal clusters and the dark background is SnO2.
169
8.11.2 SEM and AFM
The effect of sintering and Pt loadings in SnO2 are shown in Figure 8.52.
Figure 8.52 on the left shows the SEM micrographs (1-3) of pure SnO2 sintered at
500oC, 700oC and 1000oC. The SEM micrographs show that the samples have
polycrystalline cassiterite structures with grain sizes of the order less than 1 µm.
The sintering at 500-1000oC shows microstructure changes from a non-uniform
agglomeration to grain coarsening without densification. This is clearly shown as
rough surface with protrusions. The grain coarsening is mainly attributed to
sintering as observed by Varela et. al (1999). However, in this case the dry
pressed samples undergo grain coarsening without densification as also observed
by Botter et al.(1994) who worked on SnO2 ceramic formed by pressure filtration.
The grain size of SnO2 sintered at 1000oC is clearly larger than that sintered at
500oC. The average grain size of SnO2 is roughly 0.3 µm and 0.8 µm when
sintered at 500oC and 1000oC respectively. Hence, an increase in sintering
temperature will increase the grain size. This is generally true and in agreement
with other researchers Choi and Lee (2001) and Lee et al.(2000) who worked on
thin and thick film of SnO2.
The pure stannic oxide was also doped with 0.5 , 1 and 2 wt% Pt and
sintered at 1000oC. The SEM micrographs are as shown on the right (4-6) of
Figure 8.52. For these samples of platinum doped stannic oxide with 0.5-2 wt.%
Pt, the grain size increased and porosity decreased with the platinum loading.
These are in agreement with the results mentioned in earlier sections of this
chapter. It can be seen that at 0.5 wt.% Pt loading into SnO2, small size grains (~
0.1 - 0.2 µm) are more abundant than the larger grains (~ 0.7 - 0.8 µm). However,
the small grains grew comparable to the larger grain as the doping level increased
from 1 to 2 wt.% Pt in SnO2. Hence, the increase in the amount of Pt loading in
SnO2 resulted in the increase of the grain size.
170
1. SnO2 ceramic sintered at 500oC.
4. 0.5 wt.% Pt-SnO2 sintered 1000oC
2. SnO2 ceramic sintered at 700oC.
5. 1.0 wt. % Pt-SnO2 sintered 1000oC
3. SnO2 ceramic sintered at 1000oC.
6. 2.0 wt.% Pt -SnO2 sintered 1000oC
Figure 8.52: SEM micrographs 1-3 of pure SnO2 sintered at various sintering
temperature and SEM micrographs 4-6 of pure SnO2 doped at various Pt loadings
and sintered at 1000oC.
171
As the grain size grew due to sintering and Pt loadings, this led to the
rough surfaces morphology (with scanned heights of 0.00-2.55 µm) as shown in
Figure 8.53 of an AFM topography of the 0.5 wt.% Pt-SnO2 which is sintered at
1000oC.
y
z
x
Figure 8.53: AFM topography of 0.5wt.% Pt-SnO2 sintered at 1000oC. Scanned
area 14x14 (µm)2 in the x-y direction with maximum height-z direction, 2.55 µm.
The rough surface are ideal adsorption sites for oxygen in ambient air (Lee et al.,
2000). The faceted grains shown in Fig. 8.55 (plates 3-6) are an indication that it
is a stabilised material due to high sintering temperature Cirera et al. (2001). The
grain size and types of contacts that developed between the grains depend on the
method of preparation and sintering temperature (Barsan et al., 1999; Tunstall and
Patou, 1999).
From all the figures shown, the pure SnO2 and Pt-SnO2 showed that the
samples exhibit a rutile structure as most of the particles were round in shape.
HRXRD analysis confirmed the above observation.
172
8.11.3 X-ray photoemission spectroscopy (XPS)
The XPS survey spectrum of the 3 wt.% Pt-SnO2 sintered at 1000oC is
shown in Figure 8.54.
Sn 3d5/2
Sn 3d3/2
O 1s
Sn MNN
Sn 3p3/2
Sn 3p1/2
O KLL
Sn 3s
1400
1200
1000
800
600
Sn 4d
Sn 4p
Fermi
C 1s Sn 4s
level
400
200
0
EB (eV)
Figure 8.54: XPS spectra of 3 wt.% Pt-SnO2 dry-pressed ceramic sintered at
1000oC. Plot of intensity (I) against binding energy (EB).
The spectra have all been referenced to a C 1s binding energy value of 284.6 eV.
The XPS spectrum shows the Sn 4d, Sn 4p, Sn 4s, C 1s,Sn 3d5/2, Sn 3d3/2, O 1s, Sn
3p1/2, Sn 3s. The high resolution scan of O 1s in Figure 8.55 revealed that the
binding energy was 530.2 eV which is lower than the expected value 530.6 eV.
This lower value indicates that there were oxygen vacancies in the doped material.
The FWHM of the doped (3 wt.%Pt-SnO2) XPS spectrum is broader than the
173
doped
50000
40000
530.2
undoped
I (c.p.s)
30000
20000
10000
0
535
530
525
EB (eV)
Figure 8.55: O 1s of doped (3 wt.% Pt-SnO2) and undoped (SnO2) sintered at
1000oC.
undoped case, which shows that there is a greater range of environments and
hence binding energies. The asymmetry in the O 1s also shows that adsorbed
oxygen exists on the surface of the sample in ambient atmosphere (Yoo et al.,
1995). The Sn 3d5/2 and Sn 3d3/2 XPS high resolution spectrum is plotted in
Figure 8.56. The Sn 3d5/2 has a binding energy of 486.2 eV and the binding
energy of the Sn 3d3/2 was 494.2 eV which is in agreement with other data
(Perkin-Elmer, 1992). Thus the separation between the Sn 3d5/2 and Sn 3d3/2 is
8.0 eV which is smaller than the value reported (8.5 eV) by Yoo et al.(1995) who
worked on thin film SnO2-x. The larger value reported is probably attributed to
the non-stoichiometry in the SnO2-x.
174
I (c.p.s)
3d 5/2
5x10
4
4x10
4
3x10
4
2x10
4
1x10
4
8.0 eV
3d 3/2
0
478 480 482 484 486 488 490 492 494 496 498 500
E B (eV)
Figure 8.56: 3d5/2 and 3d3/2 of 3 wt.% Pt-SnO2 dry-pressed ceramic sintered at
1000oC. I is the intensity and EB is the binding energy.
The Sn 3d5/2 is symmetric and has a small FWHM indicating that the compound
has one component only. The atomic ratio of oxygen and tin (ratio of O 1s and
Sn 3d5/2) is ~ 1.30 :1, a deviation of stoichiometry which was caused by oxygen
deficiency on the surface region. Notably, the XPS is based on the photoelectric
effect and the X-ray penetration is fairly low which gives information up to a
depth of ~ 20 Å from the surface of the sample. The binding energies of both the
Sn 3d5/2 and Sn 3d3/2 shifted by 0.01 eV with respect to the pure SnO2 XPS
spectrum sintered at the same temperature (1000oC) and this is an indication that
the chemical environment was changing (Wagner et al., 1986) due to the
incorporation of Pt in SnO2. The Pt 4 f XPS spectrum is shown in Figure 8.57.
The Pt(0) or Pt 4f 7/2 has peaks at ~ 71.17 and 74.63 eV and the Pt(2) or 4f5/2 has
peaks at ~ 72.38 and 75.82 eV.
175
Pt(0)
39000
38000
Pt(2)
I (c.p.s)
37000
36000
35000
34000
33000
82
80
78
76
74
72
70
68
66
64
EB (eV)
Figure 8.57: Graph of intensity (I) against binding energy (EB). Pt 4f 7/2 and
4f5/2 of 3 wt.% Pt-SnO2 dry-pressed ceramic sintered at 1000oC.
This intensity ratio Pt 4f 7/2 : 4f5/2 was found to 1.29 and 1.38 (See Figure 8.58).
This result is in good agreement with Fildman and Mayer (1986) and Lee and
Chung (1989). This photoemission from the f electronic states produces a spinorbit doublet which is the Pt 4f 7/2 - 4f5/2 lines. High resolution XRD only shows
the Pt(0) states but not Pt(2). The Pt(2) which shows from the XPS spectrum was
probably an oxide layer on the Pt metal or possibly dissolved in the SnO2. The
latter case is supported by the observation of solid solutions as reported in the
HRXRD analysis. However, the XPS analysis above showed that the Pt 4 f looks
like mainly Pt(0) or Pt metal. This is further supported by the fact that Pt
oxidation is known to be highly passivating (McCabe et al., 1988) and that
platinum oxide, PtO2 which is a relatively stable oxide of platinum (Remy, 1956)
was confined to particle sizes < 0.75 nm and temperatures below 875 K (Lee and
Chung, 1989). None of the reported TEM work show such sizes even at room
temperature nor in the ceramics that were sintered at 1000oC.
176
I
(c.p.s)
EB (eV)
Figure 8.58: Peak ratio Pt 4f 7/2 and 4f5/2 of 3 wt.% Pt-SnO2 dry-pressed ceramic
sintered at 1000oC.
The XPS valance band of pure and doped SnO2 is shown in Figure 8.59. The
introduction of 3 wt. % Pt has modified the valance band top edge. These band of
surface states just above the valance band edge are associated with the Pt surface
states. Altman and Gorte (1989) and Henshaw et al. (1996) reported similar
bands in additive-modified oxide semiconductors in their XPS and UPS analysis.
In their case, the band of surface states is directly associated to the additive
presence whilst in another case it was an indirect association with the additive
presence (Sanjines et al., 1994). In the case of SnO2, surface band states was
observed in sputtered tin oxide surfaces. The bridging oxygen ions have been
177
Intensity
230
I (arb. unit)
180
130
3 wt.%Pt-SnO2
80
SnO2
30
12
11
10
9
8
7
6
5
(eV) (eV)
BindingEBenergy
4
3
2
1
0 -20
EEFF
Figure 8.59: XPS valance band of the pure and doped SnO2 sintered at 1000oC.
The introduction of Pt introduce a higher density of states between the Fermi
level EF and the valence band. EB is the binding energy and EF is the Fermi
energy.
removed which give rise to a local SnO-like configuration which is the source of
the new valence states (Themlin et al, 1990; Manassidis et al., 1995). When SnO2
was exposed to CH4, Kawabe et al. (2000) found that the Sn2+ associated bands
increased. Thus, the evolution of the total density of the mentioned surface states
was identical to that of the Pt chemical state concentrations. The band could arise
from Pt +2 ions at the outermost layers of the Pt localised in the surface of the
modified tin oxide crystallite which was observed and mentioned earlier.
8.11.4 Mössbauer spectroscopy analysis
The Mössbauer spectrum at room temperature for pure SnO2 and doped
SnO2 samples sintered at 1000oC are shown in Figures 8.60 and 8.61 respectively.
178
The doped sample showed higher intensity than the pure sample by an order of a
magnitude. This indicates that the doped sample has more chemical environment
due to the incorporation of Pt in SnO2. Figure 8.62 showed the combined results.
I (c.p.s)
v (mm/s)
Figure 8.60: Mössbauer spectrum of pure SnO2 sintered at 1000oC.
I (c.p.s)
v (mm/s)
Figure 8.61: Mössbauer spectrum of 3 wt.% Pt-SnO2 sintered at 1000oC.
179
doped sample
undoped sample
v (mm/s)
Figure 8.62: Combined Mossbauer spectrum of undoped (pure SnO2) and doped
(3wt.% Pt-SnO2) sintered at 1000oC.
undoped sample, a velocity of 1.2 mm/s was adequate to drop the intensity to
half its peak value. For the doped sample, the FWHM was 0.9 mm/s. It is also
noted that the doped sample shows distribution of velocities lie closer to the
fitted line. The isomer shifts in pure SnO2 was - 0.21 mm/s whilst the isomer
shifts for the doped SnO2 was - 0.18 mm/s. The strongly negative values of
isomer shifts supports the view that tin ionicity in the modified ceramic matrix are
greater than those in the pure SnO2. As the isomer shift values of the pure SnO2
was in the vincinity of – 0.21 mm/s fourfold coordination is the most probable for
Sn4+ ions residing in the pure SnO2. The quadraople splitting for the pure sample
was 0.22 mm/s and the quadrapole splitting for the doped sample was 0.61 mm/s.
The large isomer shift and quadraople splitting values in the doped sample
suggests that it is unfeasible for the sample to be assign to tetrahedrally
coordinated Sn4+ ions.
180
8.11.5 NMR analysis
The NMR chemical shifts (δ) of both pure SnO2 and doped SnO2 samples are
referred with respect to SnCl2 as shown in Figure 8.63. The standard reference
for 119Sn for NMR is tin tetramethyl but this is a flammable toxic liquid and
therefore a secondary reference is often used whose chemical shift relative to tin
tetramethyl is accurately known. Hence, SnCl2 is chosen. The oxidation state of
the compound used as reference does not matter as long as the chemical shift is
not too far removed from that of the SnO2. SnCl2 liquid gives a much narrower
line than a solid (even when spun). Furthermore, a saturated solution of SnCl2
can be easily made whereas it is difficult to produce aqueous solutions of Sn (IV)
compound. Typical NMR chemical shifts components of 119Sn for δ11, δ22, δ33 and
δiso for the magnetic field applied were observed for both the pure and doped
samples. The marked difference that was observed in the NMR chemical shifts
was the high intensity for the case of the doped sample, again reflecting a much
richer chemical environment as observed in both XPS and Mössbauer results.
12000
11000
10000
9000
8000
I (c.p.s.)
7000
undoped
6000
5000
doped
4000
3000
2000
1000
0
400
350
300
250
200
150
100
50
0
-50
-100
-150
-200
δ (ppm)
Figure 8.63: Chemical shift (δ) with respect to SnCl2 solution for pure SnO2 and
3wt.% Pt-SnO2 sintered at 1000oC.
181
8.12
Summary and Conclusion
The microstructure and physical properties of the Pt-SnO2 ceramics are
summarized in Table 8.2.
Table 8.2 Summary of microstructure and physical properties of the Pt-SnO2.
Analysis
(TGA)
Comments
sintering powders at temperatures
>700oC is favourable as powder show
no weight loss.
Bulk Density
ceramics sintered at 1000oC attained
bulk denity exceeding full density of
pure SnO2 ceramics. Bulk density
increases with Pt loadings.
Porosity
porosity decreases with Pt loadings.
BET
BET specific surface area decreases
with Pt loadings. Particle size
increases with Pt loadings.
Vickers Hardness
Vicker hardness decreases with
sintering temperature but increases
with Pt loadings (bulk density).
Elastic Modulus
Bulk, Young and Shear Modulus
increase with Pt loadings.
(XRD)
Pt in doped ceramic exist as Pt metal,
resistance to oxidation up to 1000oC,
mean crystallite size grew with Pt
loadings.
FTIR spectroscopy
shows purity of starting powders,
attaining vibration at 633 cm-1.
182
Analysis
EDAX
Comments
could not detect Pt due to low
concentration.
SEM
growth of grains with Pt loadings.
TEM
starting powders SnO2 and Pt were
nanosized.
AFM
rough surface profile.
XPS
existence of Pt 2+ as a crust on Pt
metal, existence of additional band of
states.
Mössbauer spectroscopy
existence of richer chemical
environment compared to pure sample.
NMR spectroscopy
chemical shifts shows richer chemical
environment compared to pure sample.
The microanalysis carried out showed that the Pt-SnO2 ceramics were
physically strong and stable as an active element for the methane sensor. The
rough surface of the doped ceramic is an additional feature which enhances the
adsorption of methane during sensing. The existence of additional band of states
or richer chemical environment is suggested to be the contributing factor for the
ease of electrical conduction and producing good signal for methane sensing
activity. Samples with 3 wt.%Pt-SnO2 were used for XPS, Mössbauer and NMR
as the equipment was unable to detect any change (from the pure SnO2) at Pt
loading < 3 wt.%. The observation made on the 3 wt.%Pt-SnO2 is assumed to be
equally observed in the lower Pt loading ceramics.
183
CHAPTER 9
CONCLUSIONS
9.1
Summary of findings
The findings reported in this thesis are solely based on the Pt-SnO2
ceramics which were produced by the dry pressing method. The incorporation of
Pt in SnO2 was done by mechanically, homogeneously mixing Pt powder and
SnO2 powder followed by sintering. This ceramic form of Pt-SnO2 serves as an
active element for methane sensing in air. Electrical and microstructure
characterization as well as its physical properties were investigated.
The notable features of the home-made SECS are the reproducibility
which makes the data acquisition reliable, long running time with temperature
control and stabilization and easy switching from synthetic air to methane and
vice versa. The SECS was capable of detecting methane in air via the voltage
drop across the load resistance which was in series with the sensor element. This
known voltage was converted to sensor resistance, conductance and sensitivity.
There are two distinct regions of linearity between RL and RS when operating at
400oC in 25 000 ppm of methane. For 100 Ω ≤ RL ≤ 400 Ω, the slope is 3.64
whilst for 500 Ω ≤ RL ≤ 1000 Ω, the slope is 6.06. It was found that the optimum
operating temperature for methane sensing in air was 400oC and the optimum
composition was 0.5 wt.% Pt in SnO2. This optimum operating temperature is in
agreement with other researchers and none has reported lower values that work
very well. This is probably due to methane combustion at 400oC. The optimum
184
composition which utilizes only 0.5 wt.% Pt is a low value reported which
economizes the usage of Pt. The inclusion of such a small amount of Pt in SnO2
has inevitably showed a large signal compared to pure SnO2. The choice of
sintering the ceramic Pt-SnO2 at 1000oC was mainly for stability.
In air, the resistance of the doped samples (0.1-1.0)wt.% Pt-SnO2 showed
a maximum between temperatures of 100-200oC and decrease thereafter while the
pure SnO2 did not show any maximum in the same temperature range 50-450oC.
The corresponding activation energy of the samples showed higher value
(0.29-0.43 eV) at temperatures > 200oC than at temperatures < 200oC
(0.08-0.18 eV). In methane, the resistance of the 0.5wt.%Pt-SnO2 decreased
between temperatures of 200-450oC and the activation energy was 0.21 eV for
T ≥ 400oC up to 440oC and 0.48 eV for 200 ≤ T ≤ 400oC. The conductance in air
for the doped samples increased with temperature (100-450)oC but its magnitude
was less compared to the pure SnO2. However, the 0.5wt.%Pt-SnO2 ceramic in
methane showed larger conductance (than in air) between temperatures (200450oC). The sample also obeyed the power law in the form G ~ p-0.5 and as the
conductance decreased with the gas partial pressure, it indicates that the sample is
an n-type semiconductor. The decrease in conductance was associated with the
formation of chemisorbed oxygen at the sample surface. The conductance (G) methane gas concentration (c) formed the relationship G = kc0.35.
The response and recovery times were greatly influenced by operating
temperature, material composition, gas concentration and gas flow rate. The
linear relationships between the response/recovery time the parameters could be
utilized to further enhance the sensitivity of the doped material to methane in air.
The TGA result suggested that the ideal sintering temperature for the PtSnO2 powder was at temperatures between 700-1000oC as there were no thermal
change and mass loss in this temperature range. The density of the doped samples
increased with temperature and the inclusion of Pt in SnO2 and sintered at 1000oC
yielded ceramics with bulk density greater than the full density of the pure SnO2.
The porosity of the doped ceramics decreased very slightly (negligible change)
185
with temperature. The BET analysis showed that the specific surface area
decreased with sintering temperature and the particle size increased with sintering
temperature.
The Vickers hardness decreased with temperature. The Vickers hardness
increased with Pt loadings for ceramics which were sintered at 1000oC. The
inclusion of Pt in the SnO2 showed little change in the magnitudes of the Young,
Bulk and Shear modulus. The magnitude of the moduli indicated that the doped
Pt-SnO2 ceramics were relatively strong for gas sensing purposes.
The XRD results showed that the doped ceramics are tetragonal cassiterite
polycrystalline in nature. The doped ceramics have a higher intensity than the
pure SnO2 ceramics probably due to the modification of the structure factor. The
relative intensities calculations showed that high sintering temperature decreased
the oxygen vacancies in the doped material to a certain extend. The oxygen
vacancies was also greatly reduced at Pt loadings < 2 wt.%. The strain
calculations showed that both the sintering temperature and Pt loadings could
reduce the strain but not totally eradicating it. The distortions calculation also
showed that the distortions observed are directly associated with the observed
strain.
The Raman-shift spectra of the doped ceramics showed that the inclusions
of Pt in SnO2 resulted in the diminishing the modes at ~ 773 and 473 cm-1 but the
mode A1g at ~ 633 cm-1 remained dominant. The FTIR spectroscopy results
showed that there was a small amount H2O and OH radicals in both the starting
materials. However, the inclusions of Pt in SnO2 led to almost the reduction of
the prominent minimum at ~ 632 cm-1. The decrease in the FTIR transmissions
spectrum in the doped samples was due to particle growth with Pt loadings and
there was an overall increase in the absorption of the FTIR spectrum after
exposure to 25 000 ppm CH4 at an operating temperature of 400oC.
186
The surface analysis showed that the Pt in the doped ceramics were mostly
found as clusters on the surface. SEM micrographs showed that the doped
samples grew in grain size as Pt loadings were increased. Faceted feature on the
doped sample was an indication that the material was stable. The rough surface as
shown by AFM topography provides an ideal adsorption sites for oxygen in
ambient air. As indicated by HRXRD, the material has a rutile structure and this
is shown as rounded grains in the SEM micrographs of the doped material. XPS
data indicated that the specific surface area and absorbability of the doped
material were larger than the undoped material. The state of Pt in the doped
material was dominantly Pt metal. The inclusion of Pt in SnO2 gave rise to
additional band of states which were supported by the NMR and Mössbauer
analysis.
In the light of the results from both electrical and microstructure analysis,
the sensitised mechanism involved is of the chemical type. In this mechanism, the
Pt acts as a chemical catalyst. The role of Pt is to increase the rate of reaction of
the methane gas, which adsorbed on the metallic Pt clusters and latter moved or
fashionly termed as `spillover’ the SnO2 surface. In other words, the process is
called spillover. The evidences for this type of mechanism was clearly shown by
the activation energy calculations in air and in methane and SEM micrographs
which show Pt clusters both on the surface and in the bulk of the modified
ceramics. This is in agreement to researchers like Bond et. al (1975) and
Yamazoe (1991).
In conclusion, the Pt-SnO2 ceramics have showed a remarkable electrical
property which has satisfactorily sensed methane in air with high sensitivity.
187
9.2 Recommendations
It is recommended that the following studies be attempted;
9 perform detection on desired methane concentration using gas
blenders,
9 in situ microstructure analysis using built-in heater/temperature
controller for temperatures 100-500oC.
188
REFERENCES
Abe, J. (1980). Gas-sensitive sensor with ultrafine particle film deposited
on monolithic IC chip. National Technical Report 1980 (Tech. J. Matshushita
Elec. Co. Ltd.). 26 : 457.
Agashe, C., Takwale, M.G., Bhide, V.G., Mahamuni, S., Kulkarni, S.K. (1991).
Effect of Sn incorporation on the growth mechanism of sprayed SnO2 films.
J. Appl. Phys. 70 12: 7382-7386.
Alias, M. and Hamidah, A. (2002). Petronas gas processing plant fire toll rises to
three. New Straits Times. May 4.
Albrecht, T.R., Akamine, S., Carver, T.E. and Quate, C.F. (1990).
Microfabrication of cantilever styli for the atomic force microscope. J. Vac.
Sci. Technol. A. 8(4): 3386-3396.
Altman, E.I., and Gorte, R.J. (1989). A photoelectron spectroscopy study of Pt
particles supported on α-Al2O3 {0001}. Surf. Sci. 216: 386-394.
Ambrazeviciene, V., Galdikas, A., Grebinskij, S., Mironas, A. and Tvardauskas,
H. (1993). Gas-sensing properties of chemically deposited SnOx films doped
with Pt and Sb. Sens. and Act. B. 17: 27-33.
Arai, H.,T. Yamada, Eguchi, K. and Seiyama, T. (1986). Catalytic combustion of
methane over various perovskite-type oxides. Applied Catalysis. 26: 265276.
189
Arbab, A., Spetz, A., and Lundstrom, I. (1993). Gas sensors for high
temperature based on metal-oxide silicon carbide (MOSiC) devices. Sens.
and Act. B. 15/16: 19-23.
Atashbar, M.Z., Sun, H.T., Gong, B., Wlodarski, W. and Lamb, R. (1998). XPS
study of Nb-doped oxygen sensing TiO2 thin films prepared by sol-gel
method. Thin Solid Films. 326: 238-244.
Baker, A.R. (1959). Catalytic combustion sensor. (Brit. Pat. 892, 530.)
Balce, G. (2002). Natural Gas Pipeline Development. Asia Pacific Energy
Research (APERC). Tokyo.; 2.
Balint, I., Miyazaki, A., and Aika, K.I. (2002). NO reduction over wellstructured Pt nanocrystals supported on γ-Al2O3. App. Catal. B.:
Environmental. 37: 217-229.
Baraton, M.I., Merhari, L., Chancel, F., and Tribout, J. (1998).
Nanostructured
Materials: Science and Technology. NATO-ASI Series, ed. Dordrecht:
Kluwer Academic Pub.
Baraton, M.I. and Merhari, L. (2001). Determination of the gas sensing
potentiality of nanosized powders by FTIR spectroscopy. Scripta Mater.
44:1643-1648.
Barrett, E.P., Joyner, L.G. and Halenda, P.P. (1951). The determination of pore
volume and area distributions in porous substances, I. Computations from
nitrogen isotherms. J. Am. Chem. Soc. 73: 373-380.
190
Barsan, N., Heilig, A., Kappler, J., Weimar, U., and Gopel, W. (1999a). COwater interaction with Pd-doped SnO2 gas sensors: simultaneous monitoring
of resistances and work functions. Proceedings of the EURO-SENSORS XIII
Conference. 12-15 September 1999. The Hague, The Netherlands: Euro
Sensors, 183-184.
Barsan, N., Schweizer-Berberich, M., and Gopel, W. (1999b). Fundamental and
practical aspects in the design of nanoscaled SnO2 gas sensors: A Status
Report. J. Anal. Chem. 356: 287-304.
Baselt, D. (1993). The tip-sample interaction in atomic force microscopy and its
implications for biological applications. California Institute of Technology:
Ph.D. Thesis.
Barsoum, M. (1997). Fundamentals of Ceramics. New York: McGraw Hill.
Bauer, M., Barsan, N., Ingrisch, K., Zeppenfeld, A., Denk., J., Schumann, B.,
Weimar, U. and Gopel, W. (1997). Influence of measuring voltage and
geometry of the sensing layer on the characteristics of thick film SnO2 gas
sensors. Proc. 11th European Microelectronics Conference. Venice, Italy.
May 1997. 37-44.
Benezer-Koller, N., and Herber, R.H. (1968). Experimental Methods. In:
Goldanskii, V.I. and Herber, R.H. ed. Chemical Applications of Mossbauer
spectroscopy. New York: Academic Press.
Berry, F.J., and Maddock, A.G. (1977). Mossbauer investigation of thermal
decomposition of tin (IV) hydroxide. Radiochim. Acta. 24: 32-33.
Binning, G., Quate, C.F. and Gerber, Ch. (1986). Atomic force microscope.
Phys. Rev. Lett. 56(9): 930-933.
191
Bloch, F., Hansen, W.W. and Packard, M.Z. Nuclear Induction. (1946). Phys.
Rev. 69: 127.
Bond, G.C., Molloy, L.R. and Fuller, M.J. (1975). Oxidation of carbon monoxide
over palladium-Tin(IV) oxide catalysts: an example of spillover catalysis.
J.C.S. Chem. Commun. 796-797.
Bond, G.C. (1987). Heterogeneous Catalysis: Principles and ApplicationsOxford Chemistry Series. Oxford: Clarendon.
Bott. B. (1971). Methane and ammonia gas sensors. (British Pat. 1 374 575.)
Botter, R., Aste, T. and Beruto, D. (1994). Influence of microstructures on the
functional properties of tin oxide-based gas sensors. Sens. and Act.B 22: 2735.
Briand, D., Labeau, M., Currie, J.F. and Delabouglise,G. (1998). Pd-doped SnO2
thin films deposited by assisted ultrasonic spraying CVD for gas sensing:
selectivity and effect of annealing. Sens. and Act.B 48: 395-402.
Brailsford, A. D., and Logothetis, E.M. (1998). Selected aspects of gas sensing.
Sens and. Act. B. 52: 195-203.
Bulpitt, C. and Tsang, S.C. (2000). Detection and differentiation of CH4
hydrocarbon isomers over the Pd-SnO2 compressed powder sensor. Sens. and
Act. B. 69: 144-152.
Burch, R., Urbano, F.J., and Loader, P.K. (1995). Methane combustion over
palladium catalyst: the effect of carbon dioxide and water on activity. Appl.
Catal. 123: 173-184.
192
Cabot, A., Arbiol, J., Morante, J.R., Weimar, U., Bârsan, N. and Göpel, G. (2000).
Analysis of the noble metal catalytic additives introduced by impregnation of
obtained SnO2 sol-gel nanocrystals for gas sensors. Sens. and Act. B. 70: 87100.
Cabot, A., Dieguez, A., Romano, R., Morante, J.R., and Barsan, N. (2001).
Influence of the catalytic introduction procedure on the nano-SnO2 gas sensor
performances-Where and how stay the catalytic atoms. Sens. and Act. B. 79:
98-106.
Capone, S., Siciliano, P., Barsan, N., Weimer, U., and Vasanelli, L.(2000).
Analysis of CO and CH4 gas mixtures by using a micromachined sensor
array. Sens. and Act. B. 78: 40-48.
Carlin, J.F. Jr., (2001). Tin. U.S. Geological Survey Mineral Commodity
Summaries 2001. 172-173.
Carotta, M. C., Dallara, C., Martinelli, G., Passari, L. and Camanzi, A. (1991).
CH4 thick-film gas sensors: Characterization method and theoretical
explaination. Sens. and Act. B. 3: 191-196.
Cavicchi, R. E., Suehle, J.S, Kreider, K.G., Gaitan, M. and Chaparala, P. (1996)
Optimized temperature-pulse sequences for the enhancement of chemically
specific response patterns from micro-hotplate gas sensors. Sens. and Act B
33: 142-146.
Chang, S.C. (1983). Sensing mechanism of thin film tin oxide. Proceedings of
the International Meeting on Chemical Sensors. Fukuoka, Japan: Elseiver,
78-83.
Chaudhary, V.A., Mulla, I.S., and Vijayamohanan, K. (1998). Comparative
studies of doped and surface modified tin oxide towards hydrogen sensing:
Synergistic effects of Pd and Ru. Sens. and Act. B. 50: 45-51.
193
Chaudhary, V.A., Mulla, I.S., and Vijayamohanan. (1999). Selective gas sensing
properties of surface ruthenated tin oxide. J. Mater. Res. 14: 185-188.
Chaudhary, V.A., Mulla, I.S., and Vijayamohanan, K. (2001). Hydrocarbon
sensing mechanism of surface ruthenated tin oxide: An in situ IR, ESR, and
adsorption kinetics study. J. Phys. Chem. B. 105: 2565-1571.
Chiorino, A., Ghiotti, G., Carotta, M.C. and Martinelli, G. (1998). Electrical and
spectroscopic characterization of SnO2 and Pd-SnO2 thick films studied as CO
gas sensors. Sens. and Act. B. 47: 335-338.
Chiorino, A., Ghiotti, G., Prinetto, F., Carotta, M.C., Gnani, D. and Martinelli, G.
(1999). Preparation and characterization of SnO2 and MoOx-SnO2 nano-sized
powders for thick film gas sensors. Sens. and Act. B. 58: 338-349.
Choi, S.D., and Lee, D.D. (2001). CH4 sensing characteristics of K-, Ca-, Mg
impregnated SnO2 sensors. Sens. and Act. B. 47: 205-212.
Cicerone, R.J. and Oremland, R.S. (1988). Biochemical aspects of atmospheric
methane. Global Biochemical Cycles 2: 299-327.
Cirera, A., Dieguez, A., Diaz, R., Cornet, A., and Morante, J.R. (1999). New
method to obtain stable small-sized SnO2 powders for gas sensors. Sens. and
Act. B. 58: 360-364.
Cirera, A., Vila, A., Dieguez, A., Cabot, A., Cornet, A., and Morante, J.R.
(2000). Microwave processing for low cost, mass production of undoped and
in situ catalytic doped nanosized SnO2 gas sensor powders. Sens. and Act. B.
64: 65-69.
Clifford, P. K. (1981). Mechanisms of Gas Detection by Metal Oxide Surfaces.
Carnegie Mellon University: Ph.D. Thesis.
194
Cobianu, C., Savaniu, C., Arnautu, A., Iorgulescu, R., Dascalu, D., Leo, G.,
Mazzer, M., Rella, R., Siciliano, P., Capone, S. and Vasanelli, L. (1999). A
SnO2 microsensor device for SUB-PPM NO2 detection. Sens. and Act. B. 58:
552-555.
Cobianu, C., Savaniu, C., Siciliano, P., Cupono, S., Utriainen, M. and Niinisto, L.
(2001). SnO2 sol-gel derived thin films for integrated gas sensors. Sens. and
Act. B. 77: 534-542.
Colthrup, N.B., Daly, L.H., and Wiberly, E. (1975). Introduction to Infrared and
Raman Spectroscopy. New York: Academic Press.
Comini, E., Cristalli, A., Faglia, G. and Sberveglieri, G. (2000). Light enhanced
gas sensing properties of indium oxide and tin dioxide sensors. Sens. and Act.
B. 65: 260-263.
Corcoran, P., Shurmer, H.V., and Gardner, J.W. (1993). Integrated tin oxide
sensors of low power consumption for use in gas and odour sensing. Sens.
and Act. B. 15: 32-37.
Cukrov, L.M., McCormick, P.G., Galatsis, K., and Wlodarski, W. (2001). Gas
sensing properties of nanosized tin oxide synthesised by mechanochemical
processing. Sens. and Act. B. 77: 419-495.
Cullis, C.F., Keene, D.E., and Trimm, D.L. (1970). Studies of the partial
oxidation of methane on heterogeneous catalysts. J. Catal. 19: 378-385.
Cullis, C.F. and Willatt, B.M. (1984). The inhibition of hydrocarbon oxidation
over supported precious metal catalysts. J. Catal. 86: 187-200.
Cullity, B.D. (1978). Elements of X-ray diffraction. 2nd ed. Massachusetts:
Addison-Wesley Pub.
195
de Angelis, L. and Minnaja, N. (1991). Sensitivity and selectivity of a thin-film
tin oxide gas sensor. Sens. and Act. B. 3: 197-204.
de Angelis, L. and Roberto, R. (1995). Selectivity and stability of tin dioxide
sensor for methane. Sens. and Act. B. 28: 25-29.
de Fresart, E., Darville, J. and Gilles, J.M. (1982). Influence of the surface
reconstruction on the work function and surface conductance of (110)SnO2.
Appl. Surface Sci. 11/12: 637-649.
Debeda, H., Dulan, L., Dandon, F., Menil, F., Lucat, C., and Massok, P.
(1996). Development of an intelligent methane sensor. Proc. Eurosensors X.
Sept. 8-11. Leuven, Belgium. 59-62.
Devi, G.S., Manorama, S.V., and Rao, V.J. (1998). SnO2/Bi2O3: a suitable
system for selective carbon monoxide detection. J. Electrochem. Soc. 145:
1039-1044.
Dieguez, A., Romano-Rodriguez, A., Morante, J.R., Sangaletti, L., Depero,
L.E., Comini, E., Faglia, G., and Sberveglieri, G. (2000). Influenece of the
completion of oxidation on the long-term response of RGTO SnO2 gas
sensors. Sens. and Act. B. 66: 40-42.
Ding, J., McAvoy, T.J., Cavicchi, R.E. and Semancik, S. (2001). Surface state
trapping models for SnO2-based microhotplate sensors. Sens. and Act. B. 77:
597-613.
Dlugokencky, E.J., Masarie, K.A., Lang, P.M., and Tans., P.P. (1998).
Continuing Decline in the Growth Rate of the Atmospheric Methane Burden.
Nature. 393: 447-450.
196
Donaldson, J.D., and Fuller, M.J. (1968). Ion exchange properties of tin (IV)
materials-I Hydrous tin (IV) oxide and its cation exchange properties. J.
Inorg. Nucl. Chem. 30: 1083-1092.
Dorojkine, L.M. (2003). The non-catalytic thermal wave-based chemical gas
sensor for methane and natural gas. Sens. and Act. B. 89: 76-85.
Durig, J.R. (1981). Vibrational Spectra and Structure. New York: Elsevier.
Dutronc, P., Lucat, C., Menil, F., Loesch, M., Horrillo, M.C., Sayago, I.,
Gutierrez, J., and de Agapito, J.A. (1993). A potentially selective methane
sensor based on the differential conductivity responses of Pd- and Pt-doped tin
oxide thick layers. Sens. and Act. B. 15: 384-389.
Dutronc, P., Lucat, C., Menil, F., and Loesch, M. (1993). A new approach to
selectivity in methane sensing. Sens. and Act. B. 15/16: 24-31.
Egashira, M., Matsumoto, T., Shimizu,T., and Iwanaga, H. (1987). Ar+
sputtering effect on gas sensing characteristics of tin dioxide whiskers. Proc.
Transducers `87. Tokyo, Japan: IEEE, 622-625.
Egdell, R.G., Rastomjee, C.S., Dale, R.S., Scheffer, R.J., Jones, F.H.,
Georgiadis, G.C., Lee, M.J., Tate, T.J., and CaO, L.L. (1996). Investigation
of doping of SnO2 by ion implantation and application of ion-implanted films
as gas sensors. Thin Solid Films. 279: 98-105.
Ehrhardt, J.J., Colin, L., and Jamois, D. (1997). Poisoning of platinum surfaces
by hexamethyldisiloxane (HMDS): application to catalytic methane sensors.
Sens. and Act B. 40: 117-124.
Family, F., and Meakin, P. (1988). Scaling of the droplet-size distribution in
vapour-deposited thin films. Phys. Rev. Lett. 61 4: 428-431.
197
Fau, P., Sauvan, M., Trautweiler, S., Nayral, C., Erades, L., Maisonnat, A., and
Chaudret, L. (2001). Nanosized tin oxide sensitive layer on a silicon
platform for domestic gas applications. Sens. and Act. B. 3955: 1-6.
Fayon, F., Capron, M., King, I., Le Calve, S., Alonso, B., Durand, J.O.,
Bujoli,B., Gan, Z., and Hoatson, G. (2000). Magnetic Resonance in
Chemistry.
20: 70.
Feng, C.D., Shimizu, Y. and Egashira, M. (1993). Gas sensing properties of
layer built SnO2 sensors fabricated by sol-gel processing. Proc. Symp. on
Chemical Sensors II. In: Butler, M., Ricco, A., and Yamazoe, N. Sensor
Group Proc. Vol. 93-97. Pennington, NJ, USA: The Electrochemical Society
Inc. 538-542.
Fildman, L.C., and Mayer, J.W. (1986). Fundamentals of Surface and Thin Film
Analysis. New York: Elsevier Science.
Fliegel, W., Behr, G., Werner, J., and Krabbs, G. (1994). Preparation,
development of microstructure, electrical and gas-sensitive properties of pure
and doped SnO2 powders. Sens. and Act. B. 18-19: 474-477.
Flingelli, G.K., Fleischer, M.M., and Meixner, H. (1998). Selective detection of
methane in domestic environments using a catalyst sensor system based on
Ga2O3. Sens. and Act. B. 48: 258-262.
Fleischer, M., and Meixner, H. (1993). Electron mobility in single and
polycrystalline Ga2O3. J. Appl. Phys. 74/1: 300-305.
Fleischer, M., and Meixner, H. (1995). Sensitive, selective and stable CH4
detection using semiconducting Ga2O3 thin films. Sens. and Act. B. 26-27:
81-84.
198
Fleisher, M. (1996). Fast gas sensors based on metal oxides which are stable at
high temperatures. Proc. Eurosensors X. September 8-11. Leuven, Belgium.
25-33.
Fukui, K. (1985). Alcohol selective gas sensor, Digest of 4th Chemical Sensor
Symposium. Electrochem. Soc. Japan; Japan Assoc. Chem. Sens. Tokyo. 51.
Fukui, K. and Katsuki, A. (2000). Improvement of humidity dependence in gas
sensor based on SnO2. Sens. and Act. B. 65: 316-318.
Fukui, V. (1991). Detection and measurements of odour by sintered tin oxide.
Sens. and Act. B. 5: 27-32.
Fraigi, L. Lupi, D., and Malatto, L. (1994). A thick-film pressure transducer for
cars propelled by natural gas. Sens. and Act. A. 42: 439-441.
Frank, J., Fleischer, M., Meixner, H., and Feltz. A. (1998). Enhancement of
sensitivity and conductivity of semiconducting Ga2O3 gas sensors by doping
with SnO2. Sens. and Act. B. 49: 110-114.
Fung, I., John, J., Lerner, E., Matthews, E., Prather, M., Steele, L.P., and Fraser,
R (1991). Three-dimensional model synthesis of global methane cycle.
Journal of Geophysical Research. 96: 13,033-13,065.
Gaidi, M., Labeau, M., Chenevier, B. and Hazemann, J.L. (1998). In-situ
EXAFS analysis of the local environment of Pt particles incorporated in thin
films of SnO2 semi-conductor oxide used as gas-sensors Sens. and Act. B. 48:
227-284.
Gaidi, M, Chenevier, B., and Labeau, M. (2000). Electrical properties evolution
under reducing gaseous mixtures (H2, H2S, CO) of SnO2 thin films doped with
Pd/Pt aggregates and used as polluting gas sensors. Sens. and Act. B. 60: 4348.
199
Gélin, P. and Primet, M. (2002). Complete oxidation of methane at low
temperature over noble metal based catalyst: a review. Appl. Catal. B:
Environmental. 1023: 1-37.
Gentry, S.J., and Jones, A. (1984). Poison-resistant catalytic flammable gas
sensing elements. Sens. and Act.. 5: 141.
Gentry, S.J., and Jones, T.A. (1986). The role of catalysis in solid-state sensors.
Sens. and Act. 10: 141.
Gibson, A.F. (1958). Infrared and microwave modulation using free carriers in
semiconductor. J. Sci. Instrum. 35: 273.
Gopel, W. ,Hesse, J., and Zemel, J.N. (1990). Sensors, a Comprehensive
Survey. Weinheim: VCH.
Gopel, W., and Reinhardt, G. (1996). Metal oxide sensor: new devices. In:
David, S.N. Sensors Update. New York: Wiley. 49-120.
Grasselli, S.G. (1991). Analytical Raman Spectroscopy. Grasselli, S.G. and
Bulkin, B.J. ed. New York: Wiley and Sons Inc.
Gregg, S.J. and Sing, S.W. (1982). Absorption, Surface Area and Porosity.
London: Academic Press.
Gruber, D., Kraus, F., and Muller, J. (2003). A novel gas sensor design based on
CH4/H2/H2O plasma etched ZnO thin films. Sens. and Act. B. 92: 81-89.
Haeng Yu, J. and Man Choi, G. (1998). Electrical and CO gas sensing properties
of ZnO–SnO2 composites. Sens. and Act. B. 52: 251-256.
Hagen, W., Lambrich, R. E., and Lagois, J. (1983). Semiconducting Gas Sensors.
Adv. Solid State Physics. 23: 259-274.
200
Hahn, S., Barsan, N. and Weimer, U. (2000). Impact of different feature
extraction methods on signal evaluation of sensors coated with differently
doped SnO2. Eurosensors XIV. 27-30 August. Copenhagen, Denmark.
Technical Digest.
Harber, J. (1984). Catalysis and surface chemistry of oxides. Proc. 8th Int.
Congress on Catalysis. Weinheim: Verlag Chemie, 85-107.
Harrick, N.J. (1962). Optical Spectrum of the Semiconductor Surface States
from Frustrated Total Internal Reflections. Phys. Rev. 125: 11651170.
Haruta, M., Yamada, N., Kobayashi, T., and Iijima, S. (1989). Gold catalysts
prepared by co-precipitation for low-temperature oxidation of hydrogen and
carbon monoxide. J. Catal. 115: 301-309.
Heiland, G. and Kohl, D. (1985). Problems and possibilities of oxidic and
organic semiconductor gas sensors. Sens. and Act. 8: 227-239.
Heiland, G. and Kohl, D. (1988). Physical and chemical aspects of oxidic
semiconductor gas sensors. Chem. Sens. Tecnol. 1: 15-38.
Heilig, A., Barsan, N., Weimar, U. and Göpel, W. (1999). Selectivity
improvement of SnO2 films by superficial metallic films. Sens. and Act. 58:
302-319.
Hellegouarc’h, F., Arefi-Khonsari, F., Planade, R., and Amouroux, J. (2001).
PECVD prepared SnO2 thin films for ethanol sensors. Sens. and Act. B. 73:
27-34.
Henshaw, G.S., Ridley, R., and Williams, D.E. (1996). Room-temperature
response of platinised tin dioxide gas-sensitive resistors. J. Chem. Soc.,
Faraday Trans. 92: 3411-3417.
201
Herzberg, G. (1945). Infrared and Raman Spectra. New York: Van Nostrand
Reinhold Company.
Hicks, R.F., Qi, H., Young, M.L., and Lee, R.G. (1990). Structure sensitivity of
methane oxidation over platinum and palladium. J. Catal. 122: 280-294.
Hofmeister, A.M., Xu, J., and Akimoto, S. (1990). Infrared Spectroscopy of
Synthetic and Natural Stishovite. Am. Mineral. 75: 951-955.
Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, Dai, X.,
Johnson, C.A., Maskell, K. (1996). Climate change 1995: The Science of
Climate Change. Intergovernmental Panel on Climate Change (IPPC).
Cambridge, U.K.: Cambridge University Press.
Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, Dai, X.,
Johnson, C.A., Maskell, K. (2001). Climate change 2001: A Scientific
Basis. Intergovernmental Panel on Climate Change (IPPC). Cambridge,
U.K.: Cambridge University Press.
Humaidi, S. (2000). The sensitivity of ceramic gas sensor based on SnO2-Nb2O5CuO system. Universiti Teknologi Malaysia: M.Sc. Thesis.
Huyberechts, G., Szecowka, P.M., Roggen, J., and Licznerski, B.W. (1997).
Simultaneous quantification of carbon monoxide and methane in humid air
using sensor array and an artificial neural network. Sens. and Act. B. 45: 123130.
Ichinose, N. (1975). High performance gas sensor. Material for Study Group
on Elec. Devices, Inst. Elec. Eng. Japan. 75-53.
Ihokura, K. and Watson, J. (1994). The Stannic Oxide Gas Sensor: Principles and
Applications. Florida: CRC Press Inc.
202
Ikegama, A. (1985). Olfactory detection using integrated sensor. 3rd Int. Conf.
on Solid-State Sensors & Actuators. 11-14 June. Philadelphia:
Proc.Transducers`85 ,136.
Ionescu, R. (1999). Ageing and p-type conduction in SnO2 gas sensors. Sens.
and Act. B. 58: 375-379.
Ionescu, R., Vancu, A., Moise, C. and Tomescu, A. (1999). Role of water
vapour in the interaction of SnO2 gas sensor with CO and CH4. Sens. and Act.
B. 61: 39-42.
Ippommatsu, M., Ohnishi, H., Sasaki, H., and Matsumoto, T. (1991). Study on
the sensing mechanism of tin oxide flammable gas sensors using the Hall
effect. J. Appl. Phys. 69: 8368-8374.
Ivanovskaya, M., Lutynskaya, E., and Bogdanov, P. (1998). The influence of
molybdenum on the properties of SnO2 ceramic sensors. Sens. and Act. B.
48: 387-391.
Jaegle, M., Wöllenstein, J., Meisinger, T., Bötter, H., Müller, G., Becker, T. and
Bosch-v. Braunmühl, C. (1999). Micromachined thin film SnO2 gas sensors
in temperature pulsed operation mode. Sens. and Act. B. 57: 130-134.
Jarzebski, Z.M. (1972). Oxide Semiconductors. Oxford: Pergamon Press.
JCPDS. (1997a). JCPDS XRD card 41-1445 (SnO2 cassiterite synthetic).
International Centre for Diffraction Data. Newton Square, PA.
JCPDS. (1997b). JCPDS XRD card 4-0802 (Pt metal). International Centre for
Diffraction Data. Newton Square, PA.
John, O.D. (2001). The effect of Pd on electrical properties of SnO2 in CH4
detection. Universiti Teknologi Malaysia. Ph.D. Thesis.
203
Jones, E., and Walsh, P.T. (1988). Flammable gas detection. Platinum Metals
Rev. 32: 50-60.
Jones, E. (1987). The pellistor catalytic gas sensor. In: Mosley, P.T. and Tofield.
Solid State Gas Sensors. B.C. (Eds.). Bristol: IOP Publ. 17-31.
Jones, J.T., and Bernard, M.F. (1991). Ceramics-Industrial Processing and
Testing. Ames, Iowa: Iowa State University Press.
Jones, T.A. (1982). “ Solid-state sensors: zinc oxide and pthalocyanine.”
Abs. Int. Seminar on Solid-state Gas Sensors (23rd WEH Seminar), Bad
Honnef, Germany, 7.
Jones, T.A. (1988). Flammable gas detection. Platinum Metals Rev. 32: 50-60.
Kalvius, G.M. and Kankeleit, E. (1972). Recent Improvements in
Instrumentation and Methods of Mossbauer Spectroscopy. In: Mössbauer
Spectroscopy and its Application. Vienna: IAEA.
Kamp, B., Merkle, R. and Maier, J. (2001). Chemical diffusion of oxygen in tin
dioxide. Sens. and Act. B. 77: 534-542.
Kang, W.P., Gurbuz, Y., Davidson, J.L., and Kerns, D.V.
(1994). A new
hydrogen sensor using a polycrystalline diamond-based Schottky diode. J.
Electrochem. Soc. 141: 2231-2234.
Kanefusa, S., Nitta, M., and Haradome, M. (1985). High sensitivity H2S gas
sensors. J. Electrochem. Soc. 132: 1770.
Katiyar, R.S., Dawson, P., Hargreave, M.M., and Wilkinson, G.R. (1971).
Dynamics of the rutile structure III. Lattice dynamics, infrared and Raman
spectra of SnO2. J. Phys. C. 4: 2421-2431.
204
Katsuki, A. and Fukui, K. (1998). H2 selective gas sensor based on SnO2. Sens.
and Act. B. 52: 30-37.
Kawabe, T., Shimomura, S., Karasuda, T., Tabata, K., Suzuki, E., and
Yamaguchi, Y. (2000). Photoemission study of dissociatively adsorbed
methane on a pre-oxidised SnO2 thin film. Surf. Sci. 448: 101-107.
Khodadadi, A., Mohajerzadeh, S.S., Mortazavi, Y., and Miri, A.M. (2001).
Cerium oxide/SnO2-based semiconductor gas sensors with improved
sensitivity to CO. Sens. and Act. B. 80: 267-271.
Kim, C.K., Choi, S.M., Noh, I.H., Lee, J.H., Hong, C., Choe, H.B., Jang, G.E. and
Park, H.D. (2001). A study on thin film gas sensor based on SnO2 prepared
by pulsed laser deposition method. Sens. and Act. B. 77: 427-431.
Kim, C.K., Lee, J.H., Choi, S.M., Noh, I.H., Kim, H.R., Cho, N.I., Hong, C.,
and Jang, G.E. (2001). Pd- and Pt-SiC Schottky diodes for detection of H2S
and CH4 at high temperature. Sens. and Act. B. 77: 455-462.
Kim, D.H., Yoon, J.Y., Park, H.C and Kim, K.H. (2000). CO2-sensing
characteristics of SnO2 thick film by coating lanthanum oxide. Sens. and Act.
B 62: 61-66.
Kim, J.C., Jun, H.K., Huh, J.S. and Lee, D.D. (1997). “Tin oxide-based
methane gas sensor promoted by alumina-supported Pd catalyst.” Sens. and
Act. B 45: 271-277.
Kingery, W.D., Bowen, H.K., Uhlmann, D.R. (1976). Introduction to Ceramics.
2nd Editon. New York: John Wiley and Sons.
Kiselev, V.F., and Krylov, O.V. (1987). Electronic phenomena in Adsorption
and Catalysis. Gomer, R. ed. Berlin, Heidelberg: Springer-Verlag.
205
Kissine, V.V., Sysoev, V.V., and Voroshilov, S.A. (2001). Conductivity of SnO2
thin films in the presence of surface adsorbed species. Sens. and Act. B. 79:
63-170.
Knight, W.D. (1949). Nuclear Magnetic Resonance Shift in Metals. Phys. Rev.
76: 1259.
Kocemba, I., Szafran, S., Rynkowski, J. and Paryjczak, T. (2001). The properties
of strongly pressed tin-based gas sensors. Sens. and Act. B. 79: 28-32.
Kohl, D. (1989). Surface processes in the detection of reducing gases with SnO2based devices. IEEE Trans-Electron Device. 18: 71-113.
Kohl, D. (1990). The role of noble metals in the chemistry of solid-state gas
sensors. Sens. and Act. B. 1: 158.
Kohl, D. (1992). Fundamentals and Recent Developments of Homogeneous
Semiconducting Sensors. In: Gardner, J.W. and Bartlett, P.N. Sensors and
Sensory Systems for an Electronic nose. Dordrecht: Kluwer. 53-76.
Kohl, D. (1996). Semiconductor and calorimetric sensor devices and array. In:
Kress-Rogers, E ed. Handbook of Biosensors and Electronic Noses:
Medicine, Food and the Electronic noses. Boca Raton, FL: CRC Press.
Section 41.
Kohl, D., Heinert, L., Bock, J., Hofmann, Th., and Schieberle, P. (2000).
Systematic studies on responses of metal-oxides sensor surfaces to straight
chain alkanes, alcohols, aldehydes, ketones, acids and esters using the
SOMMSA approach. Sens. and Act. B. 70: 43-50.
Kohl, D. (2001). Function and applications of gas sensors. J. Phys. D. 34: 125129.
206
Koller, A. (1994). Structure and Properties of Ceramics. Amsterdam: Elseiver
Science Publishers.
Kohl, A. (2001). Topical Review; Function and applications of gas sensors.
J. Phys. D: Appl. Phys. 34: 125-149.
Kokuen, A., Takahashi, K.H., and Nakanouchi, Y.K. (1990). Hot wire type gas
sensor of thin film tin oxide. Chem. Sens. 6 (Suppl. A): 45-48.
Komatsu, K., Sakai, S. and Fukui, K. (1984). Gas Sensor, European patent.
EPA-0115953.
Korotchenkov, G., Brynazri, V., and Drntriev, S. (1999). SnO2 films for thin film
gas sensor design. Mater. Sci. Eng. B. 56: 195-204.
Kumar Das, V.G., Ng, S.W. and Gielen, M. (1992). Chemistry and Technology
of Silicon and Tin. Oxford: Oxford University Press.
Kupriyanov, L. (1996). Semiconductor sensors in Physico-Chemical Studies. In:
Middelhoek, S. ed. Handbook of Sensors and Actuators.
Amsterdam:
Elsevier. 4.
Kvenvolden, K.A. (1991). A review of arctic gas hydrates as a source of
methane in global change. In: Weller, G., Wilson, C. and Severin, B. eds.
Proceedings of the International Conference of the Role of the Polar Regions
in Global Change. 11-15 June 1990. Fairbanks AK.: NSIDC, 696-701.
Lackey, W.J., Stinton, D.P., cerny, G.A., Schaffhauser, A.C. and Fehrenbacher,
L.L. (1987). Ceramic Coatings for Advanced Heat Engines- A Review and
Projection. Adv. Ceram Mater. 2: 24-30.
207
Lantto, V., Romppainen, P., and Leppavuori, S. (1988). A study of the
temperature dependence of the barrier energy in porous tin oxide. Sens. and
Act. B. 14: 149-163.
Laureyn, W., Delabie, L., Huyberechts, G., Maes, G., Roggen, J., Stevens, G. and
Vinckier, C. (2000). Determination of the Pd content in Pd-doped SnO2
films. Sens. and Act. B. 65: 193-194.
Lee, D.D. and Chung, W.Y. Gas-sensing characteristics of SnO2—x thin film with
added Pt fabricated by the dipping method. (1989). Sens. and Act. 20: 301305.
Lee, D.D. and Chung, W.Y. (1993). High sensitivity and selectivity methane gas
sensors doped with Rh as a catalyst. Sens. and Act. B. 13/14: 252-255.
Lee, D.D. (1994). Hydrocarbon gas sensors. Chem. Sens. Technol. 5: 79-99.
Lee, D.S., Rue, G.H., Huh, J.S., Choi, S.D., and Lee, D.D. (2001). Sensing
characteristics of epitaxially-grown tin oxide gas sensor on sapphire. Sens.
and Act. B. 77: 90-94.
Lee, S.W., Tsai, P.P., and Chen, H. (2000). Comparison study of SnO2 thin and
thick film gas sensors. Sens. and Act. B. 67: 122-127.
Leer, S.F. (1992). Global methane and the coal industry. Coal Industry Advisory
Board(CIAB). New York: IEA.
Lenaerts, S., Roggen, J., and Maes, G. (1995). FT-IR characterization of tin
dioxide gas sensor materials under working conditions. Spectrochimica Acta
Part A-Molecular Spectroscopy. 51: 883.
208
Leo, G., Rella, R., Siciliano, P., Capone, S., Alonso, J.C., Pankov, V. and Ortiz,
A. (2000). Sprayed SnO2 thin films for NO2 sensors. Sens. and Act. B. 58:
370-371.
Li, Q., Yuan, X., Zeng, G., and Xi, S. (1997). Study on the microstructure and
properties of nanosized stannic oxide powders. Mater. Chem. and Phys. 47:
239-245.
Li, G.J., and Kawi, S. (1998). Synthesis, characterization and sensing application
of novel semiconductor oxides. Talanta. 45: 759-766.
Li, G.S. and Kawi, S. (1999). MCM-41 modified SnO2 gas sensors: sensitivity
and selectivity properties. Sens. and Act. B. 59: 1-8. Sens. and Act. B. 59: 18.
Li, G.J., Zhang, X.H. and Kawi, S. (1999). Relationships between sensitivity,
catalytic activity and surface areas of SnO2 gas sensors. Sens. and Act. B. 60:
64-70.
Licznerski, B.W., Nitsch, K., Teteryez, H, Szecowka, P.M., and Wisniewski, K.
(1999). Humidity insensitive thick film methane sensor based on SnO2/Pt.
Sens. and Act. B. 79: 157-162.
Licznerski, B.W., Nitsch, K., Teteryez, H, Szecowka, P.M., and Wisniewski, K.
(2001). The influence of Rh surface doping on anomalous properties of thickfilm SnO2 gas sensors. Sens. and Act. B. 57: 192-196.
Lin, X.W., Washburn, J., Lliental-Weber, Z., Bernas, H. Coarsening and phase
transition of FeSi2 precipitates in Si. J. Appl. Phys. 75 9: 4686-4693.
Lintz, H.G., Le Goff, P., and Pentenero, A. (1962). Etude par spectrometrie de
mass de la combustion heterogene des vapeurs organiques sur platine. I Cas
du methane avec exces d’oxygene. J. Chim. Phys. 59: 933.
209
Liu, W., Cao, X., Zhu, Y. and Cao, L. (2000). The effect of dopants on the
electronic structure of SnO2 thin film. Sens. and Act. B. 66: 219-221.
Liu, W., Ma, W., Li, L. and Li, J. (2000). The temperature change in SnO2based gas sensors during sensing reaction Sens. and Act. B. 66: 222-224.
Llobet, E., Vilanova, X., Brezmes, J., Sueiras, J.E. and Correig, X. (1998).
Transient response of thick-film tin oxide gas-sensors to multicomponent gas
mixtures. Sens. and Act. B. 47: 104 -112.
Logothesis, E.M. (1975). Oxygen sensors using CoO ceramics. Appl. Phys.
Letters. 26: 209.
Logothesis, E.M., Hurley, M.D., Kaiser, W.J., and Ao, Y.C. (1986). Selective
methane sensors. Proceedings of the 2nd International Meeting on Chemical
Sensors. 8-12 July. Bordeaux, France: Chemical Sensors, 1-17.
Long, D.A. (1997). Raman Spectroscopy. New York: McGraw-Hill.
Long, G.J., Cranshaw, T.E. and Longworth, G. (1983). Mössbauer Effect Ref.
Data J. 6: 42.
Loh, J.C. (1967). Hydrogen and hydrocarbon sensor. (French pat. 1545292.)
Lundstrom, I. and Shivaraman, M.S., Svensson, C., and Lundkvist, L. (1975)
Hydrogen sensitive MOS field-effect transistor. Appl. Phys. Lett. 26: 55-57.
Madou, M., and Morisson, S.R. (1989). Chemical Sensing with Solid State
Devices. San Diego: Academic Press.
Maekawa, T., Tamaki, J., Miura, N., and Yamazoe, N. (1990). H2S detection
using CuO-SnO2 sensor. Chemical Sensors, (J. Jpn. Assoc. Chem. Sens.).
6(suppl. B. Tokyo): 21.
210
Malyshev, V.V., Eryshkin, A.V., Koltypin, E.A., Varfolomeev, A.E., and
Vasiliev, A.A. (1994). Gas sensitivity of semiconductor Fe2O3-based thickfilm sensors to CH4, H2, and NH3. Sens. and Act. B. 19: 434-436.
Malyshev, V.V., and Pislyakov, A.V. (1998). SnO2-based thick film resistive
sensor for H2S. Sens. and Act. B. 47: 181-188.
Manassidis, I., Goniakowski, J., Kantorrovich, L.N.,and Gillan, M.J. (1995). The
structure of the stoiciometric and reduced SnO2 (110) surface. Surf. Sci. 339:
258- 271.
Mangamma, G., Jayaraman, V., Gnanasekaran, T. and Periaswami, G. (1998).
Effects of silica additions on H2S sensing properties of CuO–SnO2 sensors.
Sens. and Act. B. 53: 133-139.
Mandayo, G.G., Castano, E., Gracia, F.J., Cirera, A., Cornet, A. and Morante, J.R.
(2003). Strategeries to enhance the carbon monoxide sensitivity of tin oxide
thin films. Sens. and Act. B. 95: 90-96.
Matko, I., Gaidi, M., Hazemann, J.L., Chenevier, B. and Labeau, M. (1999).
Electrical properties under polluting gas of Pt and Pd doped polycrystalline
SnO2 thin film. Sens. and Act. B. 59: 210-215.
Matsuhata, H., Ejenne, J., and Tafte, J. (1994). A study of the structure factors in
rutil-type SnO2 by high-energy electron diffraction. Acta Cryst. A. 50: 115123.
Matsuoka, M. (1978). Gas sensitivity characteristics of γ-Fe2O3 ceramic. Material
for Study Group on Elec. Devices. Inst. Elec. Eng. Japan. EDD:78-22.
Matsushima, S., Maekawa, T., Muira, N., and Yamazoe, N. (1989a). Role of
additives on alcohol sensing by semiconductor gas sensor. Chem. Soc. Jpn.
845.
211
Matsushima, S., Tamaki, J., Miura, N. and Yamazoe, N. (1989b). TEM
observation of the dispersion state of Pd on SnO2. Chem. Lett. 1651-1654.
Mattosi, F. (1951). The vibration spectrum of rutile. J. Chem. Phys. 19: 15431546.
McAleer, J.F., Moseley, P.T., Norris, J.O.W., and Williams, D.E. (1987). Tin
dioxide gas sensors Part 1 – aspects of the surface chemistry revealed by
electrical conductance variations. J. Chem. Soc., Fraraday Trans. 1. 83:
1323-1346.
McAleer, J.F., Moseley, P.T., Norris, J.O.W., Williams, D.E. and Tofield, B.C.
(1988). Tin dioxide gas sensors Part 2 – the role of surface additives. J.
Chem. Soc., Fraraday Trans. 1. 84: 441-457.
McCabe, R.W., Wong, C. and Woo, H.S. (1988). The passivating oxidation of
platinum. J. Catal. 114: 354-367.
Ménil, F., Portier, J., Lucat, C., Miquel, C., Salardenne, J., and Destriau, M.
(1983). Influence of the glass concentration in thick layers SrFeO3-x for
methane detection. Proceedings of the 9th International Meeting on Chemical
Sensors. 21 -24 September. Fukuoka, Japan: Chemical Sensors, 193-197.
Morazzoni, F., Canevali, C., Chiodini, N., Mari, C., Ruffo, R., Scotti,
Armelao, L., Tondello, E., Depero, L.E., and Bontempi, E. (2001).
Nanostructured Pt-doped tin oxide films: Sol-gel preparation, spectroscopic
and electrical characterization. Chem. Mater. 13: 4355-4361.
Morimitsu, M., Ozaki, Y., Suzuki, S.and Matsunaga, M. (2000). Effects of
surface modification with platinum and ruthenium on temperature and
humidity dependence of SnO2-based CO gas sensors. Sens. and Act. B. 67:
184-188.
212
Morrison, S.R. (1974). Electron exchange processes in olefin oxidation.
J. Catal. 34: 462-478.
Morrison, S.R. (1987a). Mechanism of semiconductor gas sensor operation.
Sens. and Act. 11: 283-287.
Morrison, S.R. (1987b). Selectivity in semiconductor gas sensors. Sens. and Act.
12: 425-440.
Moseley, P.T., (1992). Materials selection for semiconductor gas sensors. Sens.
and Act. B. 2: 329-323.
Moon, W.J., Yu, J.H., and Choi, G.M. (2001). Selective CO gas detection of
SnO2-Zn2SnO4 composite gas sensor. Sens. and Act. B. 80: 21-27.
Muller, J., and Weiβenrieder, S. (1994). ZnO-thin film chemical sensors. J.
Anal. Chem. 349: 380-384.
Munnix, S., and Schmeits, M. (1987). Band structure calculation of oxygen
vacancies on SnO2 (110) faces. J. Vac. Sci. Technol. A. 5: 910-922.
Murakami, N. (1982). The selective detection of carbon monoxide with stannic
oxide gas sensor using sensor temp cycle. Digest of 5th Chemical Sensor
Symposium, Electrochem. Soc. Japan: Japan Assoc.Chem. Sens., Tokyo. 53.
Murata, MFG (2000). Electronic Ceramics Guides. Murata Electronic:
Yokohama-shi, Kanagawa, Japan.
Nakagawa, M. (1975). Adsorption field-effect transistor. Material for Study
Group on Elec. Devices. Inst.Elec. Eng. Japan. EDD: 75-50.
213
Nakahara, T. (1987). High sensitive SnO2 gas sensor I. Detection of volatile
sulphides. Proc. Symp. Chem. Sens., 1987 Joint Congress, Electrochem.
Socs. Japan & USA., Honolulu. 55.
Nakatani, K. (1981). Gas sensors for halogenated hydrocarbons. Tech Digest,
1st Sensor Symposium 1981., Inst. Elec. Eng. Japan, Tsukuba. 25.
Niranjan, R.S., Chaudhary, V.A., Sainkar, S.R., Patil, K.R., Vijayamohanan, K.,
and Mulla, I.S. (2001). Surface ruthenated tin oxide thin-film as a
hydrocarbon sensor. Sens. and Act. B. 79: 132-136.
Nitta, M. (1978). Oscillation phenomenon in ThO2-doped SnO2 exposed CO gas.
Appl. Phys. lett. 32: 590.
Nomura, K., Shiozawa, H., Takada, T., Reuther, H., and Richter, E. (1997).
Gas-sensor properties of SnO2 films implanted with gold and iron ions. J.
Mater. Sci.: Mater. Electron. 8: 301-306.
Nomura, T. (1990). Development of semiconductor Freon gas sensor. Chemical
Sensors, (J. Jpn. Assoc. Chem. Sens.). 6(suppl. B.): 13.
Nye, J.F. (1957). Properties of Crystals. Oxford. Oxford University Press.
Ogawa, H., Nishikawa, M., and Abe, A. (1982). Hall measurement studies and
electrical conductive model of tin oxide ultra-fine particle films. J. Appl.
Phys. 53: 4448-4454.
Ogita, M., Higo, K., Nakanishi, Y., and Hatanaka, Y. (2001). Ga2O3 thin films for
oxygen sensors at high temperature. Appl. Surf. Sci. 175-176: 721-725.
O’Malley, V. (1999). The integrated pollution prevention and control (IPPC)
directive and its implications for the environment and industrial activities
in Europe. Sens. and Act.B. 59: 78-82.
214
Orlik, D.R., Ivanovskaya, G.A., Bratinsky, G.A., and Bogdanov, P. (1993).
Monoelectrode gas sensors based on SnO2 semiconductor ceramics. Sens.
and Act. B. 13-14: 155-158.
Ortega, A., Marco, S., Perera, A., Sundic,T., Pardo, A., and Samitier, J. (2000).
An intelligent detector feauturing real-time signal processing of temperature
modulated gas sensor signals. Eurosensors XIV. 27-30 August. Copenhagen,
Denmark: Techical Digest.
Ostrick, B., Fleischer, M., and Meixner, H. (1996). Preparation of
stoichiometric barium stannate thin films: Hall measurements and gas
sensitivities, In: Proceedings Eurosensors X, Leuven, Belgium.
Othman, M.H. (2000). Brunei Gas Blast in Seria. Borneo Bulletin. 27 April.
Papadopoulos, C.A., Vlachos, D.S., and Avaritsiotis, J.N. (1996). Comparative
study of various metal-oxide-based gas-sensor architectures. Sens. and Act. B.
32: 61-69.
Pagnier, T., Boulova, M., Galerie, A., Gaskov, A. and Lucazeau, G. (2000).
Reactivity of SnO2-CuO nanocrystalline materials with H2S. Sens. and Act. B.
71: 134-139.
Pan, Q., Xu, J., Dong, X. and Zhang, J. (2000). Gas-sensitive properties of
nanometer-sized SnO2. Sens. and Act. B. 66: 237-239.
Pascale, M., Muriel, L., and Daniel, B. (1995). Comparison between two Figaro
sensors (TGS 813 and TGS 842) for the detection of methane, in terms of
selectivity and long-term stability. Sens. and Act. B. 24-25: 525-528.
Perkin-Elmer (1992). Handbook of X-ray Photoelectron Spectroscopy. Edca
Prairie, MN. USA: Perkin Elmer.
215
Pitchai, R., and Klier, K. (1986). Partial oxidation of methane. Catal. Rev. Sci.
Eng. 28(1): 13 -88.
Popescu-Amalric, D. (1999). Oxydes d'étain et catalyseurs Pd/SnO2: préparation,
caractérisation et activité catalytique en DeNOx . Paris VI University: Ph.D.
Thesis.
Purcell, E.M., Torrey, H.C. and Pound, R.V. (1946). Phys. Rev. 69: 37.
Putman, C.A.J., De Grooth, B.G., Van Hulst, N.F. and Greve, J. (1992). A
detailed analysis of the optical beam deflection technique for use in atomic
force microscopy. J. App. Phys. 72: 6-12.
Quan, B., Sun, C., Wang, Z., Quan, H., Zhou, L., and Li, Y. (2000). Planar
heat-wire type CH4 gas sensors. Sens. and Act. B. 66: 243-245.
Radecka, M., and Zakrzewska, K. (1998). SnO2-TiO2 solid solutions for gas
sensors. Sens. and Act. B. 47: 194-204.
Rantala, T., Lantto, V. and Rantala, T. (1998). Computational approaches to the
chemical sensitivity of semiconducting tin dioxide. Sens. and Act. B. 47: 5964.
Rantala, T.S., Rantala, T.T. and Lantto, V. (2000). Computational studies for the
interpretation of gas response of SnO2(110) surface. Sens. and Act. B. 65:
375-378.
Remy, H. (1956). Treatise on Inorganic Chemistry, Vol.II. Amsterdam:
Elsevier.
Renault, O., Tadeev, A.V., Delabouglise, G. and Labeau, M. (1999). Integrated
solid-state gas sensors based on SnO2 (Pd) for CO detection. Sens. and Act.
B. 59: 260-264.
216
Reti, F., Fleischer, M., Meixner, H., and Giber, J. (1995). Detection of
reducing gases in air by β-Ga2O3 thin films using self heated and externally
operation modes. Proc. Eurosensors’95, Stockholm, Sweden, 456-D10 ,854857.
Ribeiro, F.H., Chow, M., and Betta, R.A.D. (1994). Kinetics of the oxidation of
methane over supported palladium catalyst. J. Catal. 146: 537-544.
Rosental, A., Tarre, A., Gerst, A., Uustare, T., and Sammelselg, V. (2001).
Atomic –layer deposition of SnO2 for gas sensing applications. Sens. and Act.
B. 77: 297-300.
Ruiz, A., Arbiol, J., Cirera, A., Cornet, A., and Morante, J.R. (2002). Surface
activation by Pt-nanoclusters on titania for gas sensing applications.
Materials Science and Engineering C. 19: 105-109.
Sadaoka, Y.(1980). The NO2 detecting ability of the phthalocyanine gas sensor.
Denki Kagaku. (J. Chem. Soc. Jpn.). 48: 486.
Safonova, O.V., Rumyantseva, M.N., Kozlov, R.I., Labeau, M., Delabouglise, G.,
Ryabova, L.I. and Gaskov, A.M. (2000). Two successive effects in the
interaction of nanocrystalline SnO2 thin films with reducing gases.
Materials Science and Engineering B. 77: 159-166
Saha, M., Banerjee, A., Halder, A.K., Mondal, J., Sen, A., and Maiti, H.S.
(2001). Effect of alumina addition on methane sensitivity of tin dioxide thick
films. Sens. and Act. B. 79: 192-195.
Saito, S. (1988). Fine Ceramics. New York: Elseiver.
Sakurai, H. (1975). A sensor made of rare transistion metal oxide in
perovskite structure. Material for Study Group on Elec. Devices. Inst.
Elec. Eng. Japan. EDD: 75-53.
217
Sala, F., and Trifiro, F. (1974). Oxidation catalysts based on tin-antimony
oxides. J. Catal. 34: 68-78.
Sanjines, R., Rosenfeld, D., Gozzo, F., Almerras, Ph., Perez, L., Levy, F.,
Margaritondo, G.,and Schreiner, W.H. (1994). ESCA investigation of SnOx
films used as gas sensors. Surf. Interf. Anal. 22: 372-375.
Satake, K. et al. (1989). NOx sensors for exhaust monitoring. Digest of 9th
Chemical Sensor Symposium, Electrochem. Soc. Japan: Japan. Assoc.
Chem. Sens. Tokyo. 97.
Saul Garcia, D.S.A.J., and Fernandez Patier, R. (1999). The environmental
control of atmospheric pollution. The framework directive and its
development. The new European approach. Sens. and Act. B. 59: 69-74.
Sauvan, M. and Pijolat, C. (1999). Humidity insensitive thick film methane
sensor based on SnO2/Pt. Sens. and Act. B. 58: 295-301.
Sayago, J., Gutierrez, J., Ares, L., Robla, J.I., Horrillo, M.C., Getino, J., Rino,
J., and Agapito, J.A. (1995). The effect of additives in tin oxide on the
sensitivity and selectivity of NOx and CO. Sens. and Act. B. 26-27: 19-23.
Scarlat, O., Mihaiu, S., Aldica, Gh., Zaharescu, M., and Groza, R. (2003).
Enhanced Properties of Tin (IV) Oxide based materials by Field-Activated
Sintering. J. Am. Ceram. Soc. 86[6]: 893-897.
Schierbaum, K.D., Kirner, U.K., Geiger, J.F. and Gopel, W. (1991). Schottkybarrier and conductivity gas sensors based upon Pd/SnO2 and Pt/TiO2. Sens.
and Act. B. 4:87-94.
Shimizu, Y., Maekawa, T., Nakamura, Y. and Egashira, M. (1998). Effects of
gas diffusivity and reactivity on sensing properties of thick film SnO2-based
sensors. Sens. and Act. B. 46: 163-168.
218
Schultz, V.M. (1979). Messung von fremdgasen in der luft mit halbleitersensoren. (Measurement of extraneous gases in air by means of
semiconducting sensors), Technisches Messen tm (Germany). 11: 405.
Schweizer-Berberich, M. (1998). Fundamental and practical aspects in the
design of nanoscaled SnO2 sensor. University of Tübingen. Ph.D Thesis.
Seiyama, T. (1962). Study on a detector for gaseous components using a
semiconductive thin film. Anal. Chem. 38: 1069.
Seiyama, T. (1972). Gas detection by activated semiconducting gas sensor,
Denka Kagaku (J. Electrochem. Soc. Jpn.). 40: 244.
Seiyama, T., Kato, A., Fujiishi, K., and Nagatani, M. (1962). A new detector for
gaseous components using semiconductive thin films. Analytical Chemistry.
34: 1502-1503.
Semencik, S., and Fryberger, T.B. (1990). Model studies of SnO2-based gas
sensors: vacancy defects and Pd additive effects. Sens.and Act. B. 1: 97102.
Sergent, N., Gelin, P., Perier-Camby, L., Praliaud, H., and Thomas, G. (2002).
Preparation and characterization of high surface area stannic oxides:
structural, textural and semiconducting properties. Sens. and Act. B. 84: 176188.
Sermon, P.A. and Bond, G.C. (1973). Hydrogen spillover. Catal. Rev. 8: 211239.
Sharma, R.K., Chan, P.C., Tang, Z.N., Yan, I., Hsing, M., Sin, J.K. (2001).
Investigation of stability and reliability of tin oxide thin-film for integrated
micro-machined gas sensor devices. Sens. and Act. B. 81 (1): 9-16.
219
Shaver, P.J. (1967). Activated tungsten gas detectors. Appl. Phys. Lett. 11: 255.
Shim, C.H., Lee, D.S., Hwang, S.I., Lee, M.B., Huh, J.S. and Lee, D.D. (2002).
Gas sensing characteristics of SnO2 thin film fabricated by thermal oxidation
of a Sn/Pt double layer. Sens. and Act. B. 81: 176-181.
Simon, I., Barsan, Bauer, M. and Weimar, U. (2001). Micromachined metal
oxides gas sensors: opportunities to improve sensor performance. Sens. and
Act. B. 73: 1-26.
Sinner-Hettenbach, M. (2001). SnO2 (110) and nano-SnO2: Characterization by
surface analytical techniques. Tübingen University: Ph.D. Thesis.
Shin-Chia Chang (1982). Thin film tin oxide gas sensor. Abst. Inst. Seminar on
Solid-State Gas Sensors (23rd WEH Seminar), Bad Honnef, Germany. 28.
Shiratori, M. (1985). A thin film gas sensor for ventilation. Digest of 4th
Chemical Sensor Symposium: Electrochemical. Soc. Japan: Japan Assoc.:
Chem. Sens. 6. (supp. B. Tokyo): 57.
Shiratori, M. (1981). Gas sensors for halogenated hydrocarbon.
1st Sensor Symposium. Eng. Japan, Tsukuba: Tech Digest, 25.
Sommer, V., Rongen, R., Tobias, P. and Kohl, D. (1992). Detection of
methane/butane mixtures in air by a set of two microcalorimetric sensors.
Sens. and Act. B. 6:262-265.
Sotomura, S. (1980). Metal-semiconductor junctions for the detection of
reducing gases and the mechanism of the electrical responses. Nippon
Kagaku Kaishi (J. Chem. Soc. Jpn.). 1585.
Sotomura, T. (1982). Gas sensitivity of SnO2 sputtered film. Electrochem. Soc.
Japan: Assoc. Chem. Sens., Tokyo. 28.
220
Spijkerman, J.J. (1971). Instrumentation. In: Leopold, M. ed. An Introduction to
Mossbauer Spectroscopy. London: Adam Hilger.
Steele, M.C. (1976). Palladium/cadmium sulfide Schottky diodes for hydrogen
detection. Appl. Phys. Lett. 28: 687.
Stetter, J.R. (1978). A surface chemical view of gas detection. J. Colloid
Interface Sci. 65: 432.
Stewart, H. (2000). The State of Environment of England and Wales: the
atmosphere. London: The Air Quality Strategy(AQS).
Sukharev, V.Y. (1993). Percolation model of adsorption-induced response of the
electrical characteristics of polycrystalline semiconductor adsorbents. J.
Chem. Soc. Faraday Trans. 89: 559-572.
Sun, L., Qiu, F. and Quan, B. (2000). Comparison study of SnO2 thin- and thickfilm gas sensors. Sens. and Act. B. 66: 237-239.
Suzuki, K. (1978). Detection of gases using an organic semiconductor.
anthracine, unseki Kagaku (J. Jpn. Soc. Anal. Chem.). 27: 472.
Szuber, J. (1988). Photoemission Study of the Fermi Level Position on SnO2
Surface and Thin Films. Proceedings of the International Seminar on
Semiconductor Gas Sensors. 22-25 September. Ustron, Poland:Elsevier,
invited paper.
Taguchi, N. (1962). Japanese Pat. S45-38200.
Taguchi, N. (1966). Japanese Pat. S47-38840.
Taguchi, N.(1966). Japanese pat. S50-30480.
221
Taguchi, N. (1969). Japanese pat. S50-23317.
Takao, Y. (1989). The detection of trimethylamine by Ru-doped semiconductor
gas sensor. Digest of 6th Chemical Sensor Symposium: Electrochem.
Soc.Japan: Japan Assoc.: Chem. Sens.: J. Electrochem. Soc. Japan. Tokyo.
23.
Tarama, K. (1958). On the catalysis of transistion metal oxide. Yukigosi Kagaku
J.Synth. Org. Chem. Japan). 16: 433.
Teo, I. (2000). Brunei: BSP Reveals Cause of Pipe Blaze. Borneo Bulletin. June
14.
Teterycz, H., Kita, J., Bauer, R., Golonka, L.J., Licznerski, B.W., Nitsch, K. and
Wisniewski, K. (1998a). New design of an SnO2 gas sensor on low
temperature cofiring ceramics. sensitive composition. Sens. and Act. B. 47:
100-103.
Teterycz, H., Licznerski, B.W., Nitsch, K., Wisniewski, K., and Golonka, L.J.
(1998b). Anomalous behaviour of new thick film gas sensitive composition.
Sens. and Act. B. 47: 153-157.
Themlin, J.M., Sporken, R., Darville, J., Caudano, R., Gilles, J.M., and Johnson,
R.L. (1990). Resonant-photoemissions study of SnO2: cationic origin of the
defect band-gap states.” Phys. Rev. B. 42: 11914-11925.
Thoren, W. (1985). Function and applications of gas sensors. Aachen
University: Ph.D. Thesis.
Tianshu, Z., Hing, P., Li, Y., and Li, Z. (1999). Selective detection of ethanol
vapour and hydrogen using Cd-doped SnO2-based sensors. Sens. and Act. B.
60: 208-215.
222
Tien, T.Y. (1975). TiO2 as an air-fuel ratio sensor for automobile exhausts.
Amer. Ceram. Soc. Bull. 54: 280.
Tournier, G., Pijolat, C., Lalauze, R., and Patissier, B. (1995). Selective
detection of CO and CH4 with gas sensors using SnO2 doped with palladium.
Sens. and Act. B. 26-27: 24-28.
Tournier, G., and Pijolat, C. (1999). Influence of oxygen concentration in the
carrier gas on the response of tin dioxide sensor under hydrogen and methane.
Sens. and Act. B. 61: 43-50.
Tunstall, D.P. and Patou, S. (1999). Size effects in the NMR of SnO2 powders.
Materials Research Bulletin. 14: 1513-1520.
Trimm, D.L., and Lam, C.W. (1980). The combustion of methane on platinum
fibre catalyst. I. Kinetics and mechanism. Chem. Eng. Sci. 35: 1405-1413.
Utriainen, M., Lattu, H., Virola, H., Niinisto, L., Resch, R. and Friedbacher, G.
(2000). Atomic Force Microscopy studies of SnO2 thin film microstructures
deposited by atomic layer epitaxy. Mikrochim. Acta. 133: 119-123.
Varela, J.A., Cerri, J.A., Leite, E.R., Longo, E., Shamsuzzoha, M., and Braadt,
R.C. (1999). Microstructure revolution during sintering of CeO doped SnO2
ceramics. Ceram. Int. 25: 253-256.
Varghese, O.K., Malhotra, L.K. and Sharma, G.L. (1999). High ethanol
sensitivity in sol–gel derived SnO2 thin films. Sens. and Act. B. 55: 161-165.
Wada, K. and Egashira, M. (1998). Improvement of gas-sensing properties of
SnO2 by surface chemical modification with diethoxydimethylsilane. Sens.
and Act. B. 53: 147-154.
223
Wagner, C.D., Riggs, W.M., Davis, L.E., Moulder, J.F., and Muilenberg, G.E.
(1986). Handbook of X-ray Photoelectron Spectroscopy. MN, USA: Perkin
Elmer.
Wang, D., Ma, L., Zhao, G., Tang, Z., Chan, C.H., Sin, K.O.and Sheng, L.Y.
(2000). Gas chromatographic study on adsorption selectivity of tin dioxide
gas sensor to organic vapours. Sens. and Act. B. 66: 156-158.
Wang, H.L., and Hon, M.H. (1999). Temperature dependence of ceramics
hardness. Ceram. Int. 25: 267-271.
Watson, J., Ihokura, K., and Coles, G.S. (1993). The tin dioxide gas sensors.
Meas. Sci. Technol. 4: 711-719.
Weisz, P.B. (1953). Weisz limitation in semiconductor gas sensors. J. Chem.
Phys. 21: 1531-1538.
Williams, D. E. (1987). Conduction and Gas Response of Semiconductor Gas
Sensors. In: Moseley, P. T. and Tofield, B. G. (eds.). Solid State Gas Sensors.
Bristol: Hilger. 71-123.
Williams, G., and Coles, G.S.V. (1999). The semistor: a new concept in selective
methane detection. Sens. and Act. B. 57: 108-104.
Willett, M.J. (1991). Techniques and Mechanisms in Gas Sensing. In: Moseley,
P.T., Norris, J., and Williams, D.E. (eds). Solid State Gas Sensors. Bristol:
Adam Hilger.
Windischmann, H., and Mark, P. (1979). A model for the operation of a thin film
SnO2 conductance-modulation carbon monoxide sensor. J. Electrochem. Soc.
126: 627-633.
224
Wollenstein, J., Böttner, H., Jaegle, M., Becker, W.J. and Wagner, E. (2000).
Material properties and the influence of metallic catalysts at the surface of a
highly dense SnO2 films. Sens. and Act. B. 70: 196-202.
Wu, J.C. (2000a). Worldwide look at reserves and production. Oil and Gas
Journal. 98: 122-123
Wu, J.C. (2000b). The mineral industry of Malaysia. U.S. geological survey
minerals yearbook-2000. 16.
Wu, J.C. (2001). KLTM tin prices, production by mining method and domestic
tin consumption. Malaysian Tin Bulletin. 12: 3.
Wu, M.T., Yao, X., Yuan, Z.H., Sun, H.T., Wu, W.C., Chen, Q.H. and Xu, G.Y.
(1993). Effect of noble metal catalyst on titania exhaust gas oxygen sensor.
Sens. and Act. B. 13-14: 491.
Xu, C., Tamaki, J., Miura, N., and Yamazoe, N. (1990). Relationship between
gas sensitivity and microstructure of porous SnO2. J. Electrochem. Soc. Jpn.
58: 1143-1148.
Xu, C., Tamaki, J., Miura, N., and Yamazoe, N. (1991). Grain size effects on
gas sensitivity of porous SnO2-based elements. Sens. and Act. B. 3: 147-155.
Yamamoto, N. (1981). Influence of chemisorption of 2-propanethiol on the
electrical properties of silver oxide compaction. Bull. Chem. Soc. Jpn. 54:
696.
Yamazoe, N., Kurokawa, Y., and Seiyama, T. (1983). Effects of additives on
semiconductor gas sensors. Sens. and Act. 4: 283-289.
Yamazoe, N. (1991). New approaches for improving semiconductor gas sensors.
Sens. and Act. B. 5: 7-19.
225
Yamazoe, N., and Miura, N. (1992). Some basic aspects of semiconductor gas
sensors-Chemical Sensor Technology Vol. IV. Seiyama, T. ed. Tokyo:
Kodansha in cooperation with Elsevier, Amsterdam.
Yannopoulos, L.N. (1987). A p-type semiconductor thick film gas sensor. Sens.
and Act. 12: 263-273.
Yoo, K.S., Cho, N.W., Song, H.S., and Jung, H.J. (1995). Surface morphology
and gas sensing characteristics of SnO2-x thin films oxidised from Sn films.
Sens. and Act. B. 24-25: 474-477.
Yu, K.N., Xiong, Y., Liu, Y., and Xiong, C. (1997). Mirostructural change of
nano-SnO2 grain assemblages with the annealing temperature. Phys. Rev. B.
55: 2666-2671.
Yu, J.H. and Choi, G.M. (1999). Electrical and CO gas sensing properties of
ZnO/SnO2 hetero-contact. Sens. and Act. B. 61: 59-67.
Yuanda, W., Maosong, T., Xiuli, H., Yushu, Z. and Guorui, D. (2001). Thin film
sensors SnO2-CuO-SnO2 sandwich structure to H2S. Sens. and Act. B. 79:
163-170.
Zakrzewska, K., Radecka, M.and Rekas, M. (1997). Effect of Nb, Cr, Sn
additions on gas sensing properties of TiO2 thin films. Thin Solid Films. 310:
161.
Zhang, G. and Liu, M. (2000). Effects of particle size and dopant on properties
of SnO2-based gas sensors. Sens. and Act. B. 69: 144-152.
Zhou, X., Xu, Y., CaO, Q., and Niu, S. (1997). Metal-semiconductor ohmic
contact of SnO2-based ceramic gas sensors. Sens. and Act. B. 41: 163-167.
226
APPENDIX A
Tin oxide powder specification:
227
APPENDIX B
Platinum powder specification:
228
APPENDIX C
The SECS (Sensor Element Characterization System) or GSCS (Gas Sensor
Characterization System) –John (2000). It was formally known as Gas Sensor
Characterization System.
Gas Sensor Characterization System (GSCS)
The experimental arrangement of the GSCS setup is depicted in Figure A1. It
comprised a gas sensor test chamber, a temperature controller (Cole Parmer), two
D.C. power supplies, a resistance box, a flow controller (Cole Parmer), a rotary valve
(Swagelock), a gas injection port and various gas cylinders. The temperature
controller regulates, via a thermocouple, the operating temperature of the sample in
the Gas Sensor Test Chamber by controlling a D. C. power supply to the heating
element inside the chamber. The other D. C. power supply provides the sensor
voltage. The flow controller is used to regulate the flow rate of gas up to 500
Standard Cubic Centimeter per Minute (sccm) into the sensor test chamber with a
high precision (maximum error of ± 1 sccm) where 1sccm is equivalent to 1
milliliters per minute (ml/min). The rotary valve is used to switch from synthetic air
to the various concentrations of CH4 in synthetic air and vise versa when measuring
the response characteristics, while the gas injection port is used to inject concentrated
CH4 into the flow of synthetic air for other types of experiments. The various gas
cylinders contain synthetic air and various concentrations of CH4 in synthetic air from
50 ppm to 25000 ppm.
The Gas Sensor Test Chamber (GSTC)
The Gas Sensor Test Chamber (GSTC) is the most important and
indispensable part of the GSCS. All of the parameters investigated in the study are
varied or controlled in this chamber.
229
Figure A2 shows the basic design of the GSTC. The main structure consists of two
aluminum blocks and a borosilicate glass tube of inner diameter (ID) 6.3 cm, outer
diameter (OD) of 7.0 cm and a height of 13.0 cm, which implies a volume of about
405 cm3, held together in a vertical position. The aluminum blocks are secured to the
glass interface by means of gaskets applied to all mating surfaces, while screws via
four stainless steel rods tighten these blocks together. The chamber is airtight and on
one side of the bottom and top aluminum blocks, gas inlet and outlet ports have been
drilled. These are fitted with Swagelok screws for securing a teflon tube of 3 mm
inner diameter for conveying gas into the chamber and over the sample.
The Reaction Unit
Figure A3 shows details of the reaction unit in the GSTC where the sensor
elements interact with gas flowing over it at a particular flow rate. Temperature
control in the reaction unit is provided by a type K thermocouple that is connected to
the temperature controller which regulates a D. C. power supply to a heater coil
mounted and secured by screws to the bottom of the chamber. The heater coil consists
of a high-wattage molybdenum resistor wire (Aldrich) of about 1.5 meter length and a
diameter of 0.5 mm coiled into a compact form that fits inside an alumina tube
forming the reaction unit. Synthetic air and premixed concentrations of CH4/air can
be introduced into the chamber via a gas inlet port bored into the aluminum block at
the bottom of the chamber and out through a similar gas outlet at the top of the
chamber. This vertical configuration of the gas sensor test chamber is justified
because of the fact that CH4 is lighter than air and, therefore, floats upwards and
reduces the possibility of its accumulation at the bottom of the chamber when the
experiment is being run.
230
Figure A1: Schematic diagram illustrating the Gas Sensor Characterization System
(GSCS).
231
Figure A2: Schematic diagram of the Gas Sensor Test Chamber (GSTC).
Reaction
unit
Gas
Figure A3: The Reaction Unit and Sensor Probe.
232
APPENDIX D
The Data Acquisition System
233
The Displayed Data
234
Data Obtained (Extracts from Microsoft Excel Sheet)
The raw data are in columns 1 and 2. Column 1 displays the time and column 2 are
the corresponding load voltage. Column 3, 4 and 5 are calculated values of the sensor
resistance Rs, the corresponding conductance and sensitivity respectively.
second
3.129375
4.169375
5.209375
6.259375
7.299375
8.339375
9.389375
10.42937
11.47937
12.51937
13.55937
14.60938
15.64937
16.68937
17.73937
18.77938
19.81937
20.86938
21.90937
22.94938
23.99937
25.03938
26.08937
27.12938
28.16937
29.21938
VL(volts)
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.33201
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
0.33201
0.33201
0.33201
0.33201
Rs
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
41467.4
41467.4
41467.4
41467.4
G
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.41E-05
2.41E-05
2.41E-05
2.41E-05
S
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5.649337
5.649337
5.649337
5.649337
1692.579
1693.619
1694.669
1695.709
1696.749
1697.799
0.35154
0.35154
0.35154
0.35154
0.35154
0.35154
39124.77
39124.77
39124.77
39124.77
39124.77
39124.77
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
2.56E-05
0
0
0
0
0
0
Vs = 20V
RL=700
ptp5,1000
100 sccm
.
.
.
The time for running this experiment was 1698 s or 28 minutes and 3 seconds.
235
APPENDIX E
Calculation of FWHM, β
The mean crystallite size of the samples can be calculated using Scherrer's
equation (Cullity, 1978) ,
RX = 0.9 λ/β cos θ
(4.12)
β2 = βm2 - βi2
and
where λ is the wavelength of the x-ray radiation, θ is the diffraction angle. βm is the
measured FWHM and βi is the FWHM from a Si standard and β is the corrected
FWHM. The value of βi is due to the instrument broadening which is just the FWHM
value of the Si standard. A plot of FWHM against 2θ is as shown below.
FWHM (radians)
Graph of FWHM against 2θ
SILICON STANDARD
0.160
0.150
0.140
0.130
0.120
0.110
0.100
0.090
0.080
FWHM = 5x10-5(2θ)x2 - 0.0022(2θ)+ 0.1204
0
10
20
30
40
50
2θ (degrees)
Thus βi=5x10-5(2θ)2 – 0.0022(2θ) + 0.1204. With known values of βm and βi, the
value of β can then be calculated.
60
236
PUBLICATIONS
1.
Zuhairi Ibrahim, John Ojur Dennis, Mohd Mustamam Abd Karim and Md
Rahim Sahar (2001). Development of a sensor characterization system
(SCS) for testing semiconductor ceramic material as a gas sensor element.
Solid State Science and Technology (2001). 9: 1-8.
2.
Zuhairi Ibrahim, John Ojur Dennis, Mohd Mustamam Abd Karim and Md
Rahim Sahar (2001). Sensitivity, response and recovery time of Pd-SnO2
semiconductor ceramic material for methane sensing. Jurnal Fizik UTM
(2001). 8: 23-30.
3.
Zuhairi Ibrahim, John Ojur Dennis, Mohd Mustamam Abd Karim and Md
Rahim Sahar. (2001). Characterization of SnO2-Pd as an active
semiconductor ceramic material for methane sensing. Proceedings of
National Seminar on Advanced Materials Development in Malaysia-2001.
May 15-16. Johor Bahru, Malaysia: FME (UTM), 60-64.
4.
Zuhairi Ibrahim, John Ojur Dennis, Mohd Mustamam Abd Karim and Md
Rahim Sahar. (2002). XRD analysis on pure and platinum doped stannic
oxide ceramics produced by dry pressing. Proceedings of Malaysian
Science and Technology Congress(MSTC) 2002(Symposium A Physical
Sciences, Engineering and Technology). September 19-21. Kuala
Lumpur, Malaysia: MSTC, 15-18.
237
5.
John Ojur Dennis, Mohd Mustamam Abd Karim, Zuhairi Ibrahim, Mohd
Khairi Saidin (2002). Hysterisis behaviour of sensitivity in CH4
detection in air using SnO2 with Pd as sensitizing additive. Jurnal
Teknologi (2002). 36(C): 55-60.
6.
Zuhairi Ibrahim, John Ojur Dennis, Mohd Mustamam Abd Karim (2002).
SEM analysis on undoped stannic oxide and platinum doped stannic oxide
as an active element for gas sensing in air. Proceedings of the 11th
Scientific Conference Electron Microscopy Society of Malaysia (2002).
December 17 & 18. Johor Bahru, Malaysia: EMSM&IMM, 145-148.
7.
Zuhairi Ibrahim, Zulkafli Othman and Mohd Mustamam Abd Karim (2003).
Morphological examination of undoped stannic oxide and platinum doped
stannic oxide for gas sensing in air. Journal of the Institute of Materials
Malaysia(JIMM) 2003. 4(2):107-116.
8.
Zuhairi Ibrahim, Zulkafli Othman, Mohd Mustamam Abd Karim and Mohd
Khairi Saidin, Rosli Husin (2003). On optimum composition of platinumdoped stannic oxide as an active element for methane sensing in air via
sensitivity measurements. Solid State Science and Technology Letter
(2003). 10(2):15-20.
9.
Zuhairi Ibrahim, Zulkafli Othman, Mohd Mustamam Abd Karim (2004).
The effect of Pt in SnO2 ceramics for methane sensing in air. To be
published in Solid State Science and Technology Letters 2004.