NOTE TO USERS
This reproduction is the best copy available.
®
UMI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Study o f Gas Specificity in M 0 O 3 /W O 3 Thin Film Sensors and their Arrays
A Dissertation Presented by Arun Kapaleeswaran Prasad
to The Graduate School in Partial fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Materials Science and Engineering
Stony Brook University May 2005
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3179613
Copyright 2005 by Prasad, Arun Kapaleeswaran
All rights reserved.
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
®
UMI UMI Microform 3179613 Copyright 2005 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Copyright by Arun Kapaleeswaran Prasad 2005
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Stony Brook University The Graduate School
Anrn Kapaleeswaran Prasad______
We, the dissertation committee for the above candidate for the
Doctor of Philosophy
degree,
hereby recommend acceptance of this dissertation.
Prof. Pelagia I. Gouma, Dissertation Advisor
kjuU
o ' Prof. Richard J. Gajauftno, SUNY Stony Brook
9
Prof. GaTy P. Halada, SUN1 SUNY Stony Brook
0
it ■
Dr. David J. Kubinski, Ford Research Laboratory This dissertation is accepted by the Graduate School
Dean of the Graduate School
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract of the Dissertation Study of Gas Specificity in M 0 O3 /WO3 Thin Film Sensors and their Arrays by Arun Kapaleeswaran Prasad Doctor of Philosophy in Materials Science and Engineering Stony Brook University May 2005 Chemical sensors that monitor gas concentrations through changes in the electrical resistance of their sensing elements are called resistive type detectors. Gas specificity is a highly desirable property in gas sensing, and is defined as the preferred response to a particular gas in the presence of other interfering compounds. The lack of specificity and limited selectivity of existing chemosensors often results in false alarms, thus reducing the reliability of these devices. This work focuses on specific gas-oxide interactions in transition metal oxides that exhibit structure sensitivity in catalytic processes. The hypothesis is that gas specificity depends on the oxide’s polymorph phase used in sensing. M 0 O3 and WO 3 thin films have been chosen for this study as the model gas sensing elements. Ammonia, nitrogen oxides and hydrocarbons are the target gases selected.
Gas sensing films were fabricated using ion beam deposition and sol-gel techniques.
The effects of processing parameters, film thickness, stabilization heat
treatment, and sensing temperature on the films’ microstructures have been studied. The
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
microstructures of the films are characterized using transmission electron microscopy, Xray diffraction, and scanning electron microscopy. Differential scanning calorimetry experiments are performed to establish the phase stability fields for the various polymorphs of the model systems under study. Gas sensing tests were carried out using the orthorhombic and monoclinic phases of M 0 O3 and WO3 .
It was found that it is possible to control the microstructure and operating temperature of a single semiconducting metal oxide film (viz, M 0 O3 or WO3 ) so as to produce polymorphs that are sensitive to a particular gas. A suitable sensor array is proposed consisting of un-doped metal oxide which can detect different target gas species selectively.
IV
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dedicated to
Amma, Pappa, Atha, Papamma and my beloved Late Tatha
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CONTENTS List of Figures..........................................................................................................................x List of T ables........................................................................................................................ xv Acknowledgements.............................................................................................................xvi Vita...............................................................................
xviii
Publications.......................................................................................................................... xix
1. Introduction.......................................................................................................................1 1.1 Metal Oxide Resistive Gas Sensors...................................................................... 1 1.2 Detection of Ammonia and NOx ......................................................................... 2 1.3 M 0 O3 and WO3 as chemiresistivesensors............................................................4 1.3.1 M 0 O3 as Gas Sensor....................................................................................4 1.3.2 WO3 as Gas Sensor......................................................................................7 1.4 Statement of Problem........................................................................................... 18 References................................................................................................................................20
2. Experimental Techniques............................................................................................... 27 2.1 Dual Ion Beam Deposition.................................................................................. 27 2.2 Sol-gel Deposition............................................................................................... 30 2.3 TEM Characterization........................................................................................31 2.3.1 IBD Sam ples...................................................................................... 31 vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3.2 Sol-gel Sam ples...................................................................................32 2.4 SEM Characterization..........................................................................................32 2.5 XRD Characterization.........................................................................................32 2.6 Testing of M 0 O3 and WO3 F ilm s....................................................................... 33 References...............................................................................................................................36
3. M 0 O3 Thin Films as a Gas Sensor..................................................................................37 3.1 Effect of Thickness.............................................................................................. 37 3.1.1 SEM Characterization......................................................................... 38 3.1.2 XRD Characterization.......................................................................... 40 3.1.3 Sensing Tests........................................................................................42 3.2 Effect of Polymorphism.......................................................................................44 3.2.1 Films Heat Stabilizedat 350°C.......................................................... 44 3.2.2 Films Heat Stabilizedat 400°C.......................................................... 45 3.2.3 Films Heat Stabilizedat 450°C.......................................................... 47 3.2.4 Films Heat Stabilizedat 500°C.......................................................... 49 3.3 Effect of Operating Temperature........................................................................ 51 References................................................................................................................................ 55
4. Novel WO 3 Thin Films as a Gas Sensor....................................................................... 57 4.1 Thin Film Preparation.......................................................................................... 57 4.2 Characterization of WO 3 thin films....................................................................58 4.2.1 SEM Characterization.........................................................................59
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2.2 XRD Characterization........................................................................61 4.3 Sensing Tests........................................................................................................63 4.3.1 W2 Films............................................................................................... 64 4.3.1.1 Sensing Temperature 200°C................................................64 4.3.1.2 Sensing Temperature 300°C................................................65 4.3.1.3 Sensing Temperature 400°C................................................67 4.3.1.4 Sensing Temperature 500°C................................................ 6 8 4.3.2 W3 Films............................................................................................... 71 4.3.2.1 Sensing Temperature 200°C................................................71 4.3.2.2 Sensing Temperature 300°C................................................72 4.3.2.3 Sensing Temperature 400°C................................................74 4.3.2.4 Sensing Temperature 500°C................................................75 4.4 Conclusions...........................................................................................................78 References................................................................................................................................ 79
5. Sensor Arrays for Detection of Isoprene, Methanol and Carbon dioxide............ 82 5.1 Methanol Detection.............................................................................................. 82 5.2 Isoprene Detection............................................................................................... 83 5.3 Experimental.........................................................................................................83 5.4 Results................................................................................................................... 85 5.5 Conclusions..........................................................................................................93 References
.....................................................................................................................94
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
. Discussion..........................................................................................................................96 6.1 Processing Issues - Comparing responses from different processing routes..96 6.1.1 M 0 O3 thin films..........................................................................................96 6.1.2 WO3 thin films............................................................................................ 97 6.2
PolymorphicPhase
Selection -
Key
to
achieving
specificity
in
sensing............................................................................................................... 99 6.3 Building Optimum Sensors - Identifying optimum temperature regimes and crystal structure for specificity...........................................................................105 References.............................................................................................................................. 107
7. Conclusions and Future W ork ..................................................................................... 109 7.1 Summary of Conclusions........................................................................109 7.2 Future Work............................................................................................. I l l
A ppendix.............................................................................................................................. 113
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Figures Figure 2.1. Dual Ion Beam Deposition System.................................................................. 29 Figure 2.2. Electrical circuit for resistance measurement.................................................34 Figure 2.3. Schematic Diagram of the experimental setup of the Sensing System
35
Figure 3.1. SEM images of films deposited for 4 min (a) before sensing (b) after sensing...................................................................................................................................... 38 Figure 3.2. SEM images of films deposited for 40min(a) before sensing deposited on alumina (b) after sensing - on alumina region (c) before sensing deposited on gold electrodes............................................................... ................................................................39 Figure 3.3. SEM images of films deposited for 400 min (a) before sensing (b) after sensing..................................................................................................................................... 40 Figure 3.4. XRD of films (a) before sensing (b)aftersensing.......................................... 41 Figure 3.5. Plot of resistance of films (with and without ammonia gas) against sensing temperature for films of different thickness........................................................................42 Figure 3.6 Sensitivity Versus Thickness of IBD M0 O3 film s............................................42 Figure 3.7 SEM image of sol-gel M0 O3 films heat treated at 350°C for
8
h o u rs............44
Figure 3.8 XRD spectrum of films heat treated at 350°C ..................................................45 Figure 3.9 SEM image of M0 O3 film heat treated at 400°C for
8
hours...........................46
Figure 3.10 TEM image showing grains of Monoclinic phase, inset —SADpattern of one of the grains showing [011] zone axis of monoclinic phase (Space Group :P21/c (14)).. .46 Figure 3.11 XRD spectrum of sol-gel M0 O3 films heat treated at 400°C.........................47 Figure 3.12 SEM image of films heat treated at 450°C for
8
h ours................................. 48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.13 TEM image of M 0 O3 film after heat treatment at 450°C; inset - SAD pattern of one of the grains showing [010] zone axis of monoclinic phase (Space Group :P21/c ( 1 4 ) ) ........................................................................................................................................ 48 Figure 3.14 XRD spectrum of films heat treated at 450°C ................................................49 Figure 3.15 SEM image of films heat treated at 500°C...................................................... 50 Figure 3.16 TEM image of films heat treated at 500°C inset - SAD pattern of one of the grains showing [100] zone axis of orthorhombic phase (Space Group :Pbnm (62) ) ........50 Figure 3.17 XRD Spectra of M 0 O3 thin film heat stabilized at 500°C ............................. 51 Figure 3.18 Response of ion beam deposited M 0 O3 thin films to (a) 400 ppm ammonia (b) 100 ppm am m onia............................................................................................................ 52 Figure 3.19 Comparison of Sensitivities of sol-gel M 0 O3 films to various concentrations of ammonia at different operating temperatures.................................................................. 53 Figure 3.20 Comparison of Sensitivity of sol-gel M 0 O3 to NO 2 at different operating temperatures.............................................................................................................................54
Figure 4.1 SEM image of as-received W2 film a) at 20kX b) at 1 kX .............................. 59 Figure 4.2 SEM image of W2 film after heat treatment at 500°C for
8
hoursa) at20kX
b) at 1 kX................................................................................................................................. 59 Figure 4.3 SEM image of W3 film in as received state a) at 20 kX b) at 1 kX ............... 60 Figure 4.4 SEM image of W3 film after heat treatment at 500°C for
8
hours a) at 20kX
and b) at 1 k X .........................................................................................................................60 Figure 4.5 Comparison of XRD spectrum of W2 and W3 films in as-received sta te
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Figure 4.6 Comparison of XRD spectrum of W2 and W3 films heat treated at 500°C for 8
hours...................................................................................................................................... 62
Figure 4.7 DSC/TGA graphs of dried sol-gel WO 3 ........................................................... 63 Figure 4.8 Response of W2 films to NO 2 at 200°C (50 ppm to 400 ppm)........................64 Figure 4.9 Response of W2 films to NH 3 at 200°C (50 ppm to 500 ppm)........................65 Figure 4.10 Response of W2 films to NO 2 at 300°C (50 ppm to 500 ppm)..................... 6 6 Figure 4.11 Response of W2 films to NH 3 at 300°C (50 ppm to 500 ppm)..................... 6 6 Figure 4.12 Response of W2 films to NO 2 at 400°C (50 ppm to 400 ppm)..................... 67 Figure 4.13 Response of W2 films to NH 3 at 400°C (50 ppm to 400 ppm)......................6 8 Figure 4.14 Response of W2 films to NO 2 at 500°C (50 ppm to 400 ppm)......................69 Figure 4.15 Response of W2 films to NH 3 at 500°C (50 ppm to 400 ppm)......................69 Figure 4.16 Comparison of Sensitivities of W2 films towards NO 2 at various sensing temperatures.............................................................................................................................70 Figure 4.17 Comparison of Sensitivities of W2 films towards NH 3 at various sensing temperatures.............................................................................................................................70 Figure 4.18 Response of W3 films to NO 2 at 200°C (50 ppm to 400 ppm)......................71 Figure 4.19 Response of W3 films to NH 3 at 200°C (50 ppm to 500 ppm)......................72 Figure 4.20 Response of W3 films to NO 2 at 300°C (50 ppm to 500 ppm)..................... 73 Figure 4.21 Response of W3 films to NH 3 at 300°C (50 ppm to 500 ppm)..................... 73 Figure 4.22 Response of W3 films to NO 2 at 400°C (50 ppm to 200 ppm)..................... 74 Figure 4.23 Response of W3 films to NH 3 at 400°C (50 ppm to 400 ppm)......................75 Figure 4.24 Response of W3 films to NO 2 at 500°C (50 ppm to 400 ppm)......................76 Figure 4.25 Response of W3 films to NH 3 at 500°C (50 ppm to 500 ppm)......................76
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.26 Comparison of Sensitivities of W3 films towards NO 2 at various sensing temperatures............................................................................................................................. 7 7 Figure 4.27 Comparison of Sensitivities of W3 films towards NH 3 at various sensing temperatures............................................................................................................................. 7 7
Figure 5.1 (a) SEM image of inter-digitated electrodes (b) SEM image of heaters
84
Figure 5.2 Multisensor gas sensing setup............................................................................ 85 Figure 5.3 Sensing responses of two sets of M 0 O3 thin films towards ammonia (15 ppm), methanol (400 ppm) and CO (15 ppm).................................................................................8 6 Figure 5.4 Comparison of sensitivities of M 0 O3 thin films towards various gases/gas mixtures at 400°C................................................................................................................... 87 Figure 5.5 Sensing responses of two sets of M 0 O3 thin films towards ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm)......................................................... 87 Figure 5.6 Comparison of sensitivities of M 0 O3 thin films towards various gases/gas mixtures at 420°C................................................................................................................... 89 Figure 5.7 Sensing responses of two sets of M 0 O3 thin films towards ammonia (15 ppm), NO 2 (500 ppm), isoprene
(1 0
ppm) and CO2 (5 ppm)......................................................... 90
Figure 5.8 Comparison of sensitivities of M 0 O3 thin films towards various gases/gas mixtures at 450°C...................................................................................................................91 Figure 5.9 Comparison of XRD spectra of films heat treated at 420°C and 450°C.........92
Figure 6.1 Comparison of SEM images of films obtained by (a) Sol-gel (b) Ion Beam Deposition and (c) Acidic Precipitation................................................................................ 98
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6.2 Phase transformation from (3 to a phase in M 0 O3 ........................................... 101 Figure 6.3 XPS spectra obtained from the same M 0 O3 film after exposure to different gas environments at ~465°C. (a) The Mo 3d spectra after exposure to 1000 ppm NH 3 in 10% O2 . (b) After exposure to 10% O2 only, (c) After exposure to 1000 ppm NH 3 in 0.5% C>2 .(d) After 1000 ppm C 3 H6 in 10% O2 ............................................................................. 102 Figure 6.4 Structure of monoclinic WO3 ........................................................................... 103
Figure A.1 Comparison of Sensitivities of sol-gel M 0 O3 to NH 3 at different operating temperatures.......................................................................................................................... 114 Figure A.2 Comparison of sensitivities of sol-gel M 0 O3 to NO2 at different operating temperatures.......................................................................................................................... 114 Figure A.3 Comparison of sensitivities of sol-gel M 0 O3 films toward NH 3 and NO 2 at 4 7 5 °C .................................................................................................................................... 115 Figure A.4 Comparison of Sensitivities of W2 films towards NO 2 at various sensing temperatures.........................................................................................................................116 Figure A.5 Comparison of sensitivities of W2 films towards NH 3 at various sensing temperatures.......................................................................................................................... 116 Figure A . 6 Comparison of sensitivities of W3 films towards NO 2 at various sensing temperatures..........................................................................................................................117 Figure A.7 Comparison of Sensitivities of W3 films towards NH 3 at various sensing temperatures..........................................................................................................................117 Figure A . 8 Comparison of sensitivities of W2 and W3 films towards NO 2 and NH 3 at 200°C..................................................................................................................................... 118
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
List of Tables Table 1.1 Resistance responses for p-type and n-type m aterials.........................................2
Table 5.1 Response of sensor arrays of M 0 O3 to square pulse of methanol (400ppm), carbon monoxide (15 ppm) and ammonia (15 ppm) at 400°C............................................ 8 6 Table 5.2 Response of sensor arrays of M 0 O3 to square pulse of ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 420°C.................................................. 8 8 Table 5.3 Response of sensor arrays of M 0 O3 to square pulse of ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 450°C.................................................. 90 Table 5.4 Percentage distribution of each phase in films heat
treated at 420°C and
450°C....................................................................................................................................... 92 Table 5.5 Selective Sensor Array prototype......................................................................... 93
Table 6.1 Comparison of Sensitivities of M 0 O3 thin film prepared by sol-gel and ionbeam deposition.......................................................................................................................96 Table 6.2 Comparison of Sensitivities of WO3 thin
films prepared by various
methods.................................................................................................................................... 97 Table 6.3 Classification of oxides, gases they detect and
sensing
mechanism
involved.................................................................................................................................. 106 Table 6.4 Optimum Sensor Arrays- Temperature and
Phase domainsand Gases
detected.................................................................................................................................. 106
xv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Acknowledgements I would like to extend my sincere thanks to my advisor Prof. Perena Gouma, for her guidance, motivation and support throughout the course of the research. Her valuable advice, criticism and encouragement have greatly helped in the materialization of this dissertation.
Sincere thanks are also due to Prof. Richard Gambino, for his guidance during the initial stages of the research and the use of his lab facilities. His comments and ideas have helped me focus the research in a systematic manner.
I would like to thank Prof. Gary Halada for serving on my defense committee and for his valuable suggestions and inputs during my research.
I am deeply grateful to Dr. David Kubinski for walking me through the various aspects of sensor technologies including assistance with the gas sensing setup. I express my thanks to him for several sensing experiments done in his lab, XPS measurements and support for the project.
I would like to thank Dr. James Quinn for always being there in times of need and emergencies related to laboratory equipments. Without his help and suggestions, this dissertation would not have been possible.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I would like to thank Dr. Csaba Balazsi (Hungary) for providing me with powders for preparation of sensor films. I wish to acknowledge the sensing tests performed in laboratory of Prof. Sbverglieri ( Italy).
Thanks are also due to Ms. Laurie Dudik for supplying the sensor substrates for use throughout the course of the research. Special thanks to Kitty, Debby and Gertha for providing me all the academic and research-related help throughout my study.
Finally, thanks to my parents, all my relatives and friends for their moral support throughout the course of this work.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Vita Arun K. Prasad Born: July 6, 1979, Bangalore, India.
Education
M.S., (Aug 2000-July 2002), Materials Science and Engineering, SUNY at Stony Brook, NY (Thesis Title: Processing and Microstructural Effects on the Gas Sensing Properties of M 0 O3 and WO3 Thin films)
B.Tech, (July 1996 - May 2000), Chemical and Electrochemical Engineering, Central Electrochemical Research Institute, Karaikudi, India
Work Experience
Research Assistant
(August 2001-till date), Dept, of Materials Science and
Engineering, SUNY at Stony Brook, NY
Summer Intern (June 2001-August 2001), Lucent Technologies (Agere Systems) Murray Hill, NJ
Teaching Assistant (August 2000-July 2002), Dept, of Materials Science and Engineering, SUNY at Stony Brook, NY
xviii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Publications (Journal and Conference Proceedings) This Work
1. A.K. Prasad, D. Kubinski, and P.I. Gouma, Processing effects on MoOj-based gas sensors fo r selective ammonia detection. Sensors and Actuators B-Chemical, 93 (2003) 25-30. 2. A.K. Prasad, P.I. Gouma, D.J. Kubinski, J.H.Visser, R.E.Soltis, and P.J.Schmitz, Reactively Sputtered M 0 O3 films fo r ammonia sensing. Thin Solid Films. 436 (2003)46-51 3. A.K. Prasad, P. I. Gouma, M 0 O3 and WO3 based thin film conductimetric sensors fo r automotive applications. Journal of Materials Science 38 (2003) 4347 -4 3 5 2 4. A.K. Prasad, P.I. Gouma, Effect o f Thickness o f Ion Beam Deposited Molybdenum Trioxide Thin films on Gas Sensing Properties, 2004 MRS Fall Meetings Proceedings, Vol 828, 2004 5. A.K. Prasad, P. I. Gouma, Nano-structured Sensor Materials fo r Selective BioChemical Detection, Ceramic Transactions, Vol 159 p 257-264 (2004) 6. K.M. Sawicka, A.K. Prasad, P. I. Gouma, Processing and Characterization o f Nanostructured Metal Oxides and Nanocomposites for use in Chemical Sensing Applications, Proceedings of the 2nd AIST International Workshop on Chemical Sensors, p 41-49, Nagoya Japan (2004) 7. K.M. Sawicka, A.K. Prasad, P.I. Gouma, Metal Oxide Nanowires fo r use in Chemical Sensing Applications, Sensor Letters 3 (1) pp 31-35 (2005)
xix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Other Work
8
. P. Giridhar, A. K. Prasad, N. Kalaiselvi, K. Gopalakrishnan, M. Ganesan, and A. Veluchamy, A Review on Lithium - Ion polymer electrolyte batteries. Bulletin of Electrochemistrv. 15 (9-10): p. 414-418, (1999).
9. A.K. Prasad, P. Giridhar, V. Ravindran, and V.S. Muralidharan, Zinc-cobalt alloy:
Electrodeposition
and
Characterization.
Journal
of
Solid
Electrochemistry, 6(1): p. 63-68, (2001).
xx
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
State
Chapter 1 Introduction This work focuses on understanding the relationship between the microstructure and gas sensing properties of M 0 O3 and WO3 based thin film gas sensors. Gas specificity is a highly desirable property in gas sensing and the lack of it often results in false alarms, which is detrimental to their reliability. The reason for the specific response of these semiconducting oxides towards a particular gas is identified and the mechanism of response is proposed. Firstly, an introduction is given to metal oxide resistive gas sensors to understand the basic principle underlying a chemiresistive sensor. Then, the need to detect the gases used in this study is explained. Next a brief introduction to the materials used for fabricating the sensor is discussed which leads to the statement of the problem of this work.
1.1 Metal Oxide Resistive Gas Sensors A resistive gas sensor is a kind of chemical sensor in which the changes in gas concentration are manifested as a change in resistance of the sensor material. The idea of chemical sensors started after the invention of pH electrode in 1922, which is considered as the first chemical sensor to detect chemical changes (concentration). Other sensing technologies based on oxidation-reduction reactions at electrodes were extensively pursued in the 1940’s and 1950’s.
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Taguchi then devised the first commercial sensor in 1968. A Taguchi sensor operates on the principle that a chemical reaction between the gas species to be detected and a semiconducting metal oxide sensing materials produces a change in the film’s resistance. Materials can be classified as either n-type or p-type based on whether they show a decrease or an increase of resistance when they are exposed to either a reducing or an oxidizing gas in an atmosphere of fixed oxygen partial pressure (as in air). Reducing gases inject electrons into the semiconductor by releasing the adsorbed oxygen thereby decreasing the resistance in an n-type semiconductor. Oxidizing gases accept electrons from the semiconductor thereby increasing the resistance of an n-type semiconductor. For a p-type material, the absorption or release of electrons is manifested as release or adsorption of holes respectively. Hence the response is opposite to that of ntype materials. Based on this, the resistance changes expected for reducing and oxidizing gases on n-type and p-type semiconducting oxides can be summarized as shown below. Table 1.1 Resistance responses for p-type and n-type materials M aterial
1 R edu cing gas
O xid izin g G as
n -t\p e
i Resistance decreases
Resistance increases
p-lype
! Resistance increases
Resistance decreases
1.2 Detection of Ammonia and NOx Ammonia is a poisonous and colorless gas with a pungent smell. It is a common feedstock in many chemical processes involving nitrogen compounds [1], Gas odors,
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
which emanate from such processes and livestock facilities, can pose serious health problems and hence should be eliminated from our working and living environments. Long term exposure limit
(8
hour reference period) of ammonia gas for an average
human being is 25 ppm while short-term exposure limit (10 min reference period) is 35 ppm. But the lower limit of human perception of ammonia is 53 ppm [2]. These levels thus emphasize the need for a sensor to operate in such low ammonia concentrations. Ammonia sensors are being used for diverse applications such as in food technology, chemical plants, medical diagnosis and for environmental protection [3], In food technology, ammonia from cold storage chillers needs to be monitored to ensure safety from bacterial attack. The use in chemical plants and for environmental protection has already been described above. In medical diagnosis, ammonia sensors are used to track the activity of urease-producing bacteria [4]. Ammonia sensors also find uses in selective catalytic reduction (SCR) systems [5, 6
]. These SCR systems are employed in the exhaust system of commercial vehicles,
combustion systems in power plants and in industrial boilers to monitor the NOx emissions (NO2 and NO). Nitrogen dioxide is a harmful gas causing potential hazard to humans. Hence the detection of NO 2 is critical and needs to be controlled from emissions occurring from power plants, industrial boilers and fossil fuel combustion [6 , 7]. In a SCR converter, ammonia serves as a reducing agent for NOx converting it into environmentally safe nitrogen and water vapor.
An ammonia sensor is required to
calculate the amount of unreacted and excess ammonia, which is fed into the inlet stream. This quantity of ammonia has to be kept at a minimum. In other words, the ammonia in the inlet should be just sufficient to convert all the NOx and the ammonia liberated, if
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
any, should be minimal. By proper monitoring, this excess ammonia can be re-injected
into the inlet stream for continuous operation. Thus, there is a need for both NOx as well as an ammonia sensor.
NOx (NO and NO 2 ) gases are known to be toxic gases that can cause diseases of the respiratory system and are also harmful for the environment as a source of acid rain and photochemical smog. NOx gases are thus critical factors in formation of ozone in the troposphere. But excessive ground level ozone is more injurious to health and is responsible for choking; coughing and can cause lung damage and other respiratory disorders. Currently, about one-half of all NOx emissions into the environment are due to power plants and industrial boilers [6 ].
1.3 M 0 O3 and W 0 3 as chemiresistive sensors 1.3.1 M 0 O 3 as Gas Sensor Molybdenum trioxide has been studied as a sensor material for almost a decade. The sensing behavior of sputtered M 0 O3 thin films was first investigated by Mutschall et al [8 ]. Thin films of M 0 O3 were prepared by rf sputtering and films of different thicknesses were prepared. Various gases were tested and the films of thickness 600nm to lp m were found to be highly sensitive to ammonia in the temperature range of 400450°C with a response time of around 30 seconds. Ferroni et al [9] reported the response of similarly prepared films to CO. The sputtered films were heat stabilized at 500°C as in the earlier work by Mutschall et al. The films were found to be sensitive to as low as 10 ppm of CO in the temperature range
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of 300-500°C with a response time of 1-2 min. The phase obtained was the stable orthorhombic phase. The effect of the processing techniques on the sensing behavior of M 0 O3 has been studied by Comini et al [10]. The films obtained from RF-sputtering had a welldeveloped nanoparticle microstructure and a larger surface area as compared to sol-gel films. The effect of stabilization heat treatment temperature was studied for sol-gel films. Transformation from equiaxed grains at 400°C to long needle-like structure at 500°C to continuous planar amorphous structure at 600°C and above is observed. The response of such films is studied towards CO. The selectivity towards ammonia was enhanced by addition of Ti over layers [11]. It is supposed that partial substitution of Ti atoms into the M 0 O3 lattice affect the sensing properties of the M 0 O3 films by inducing lattice distortions. The films in this case were prepared by RF sputtering and were about 450 nm thick with a 50nm layer of Ti deposited over the M 0 O3 layer. Sensing tests were carried out with various gases such as ammonia, CO, H2 and SO2 in the temperature range of 100°C to 350°C. The un-doped M 0 O3 showed maximum sensitivity toward ammonia (200 ppm) at 250°C. The Ti-doped M 0 O3 showed maximum sensitivity toward ammonia (200 ppm) at 200°C with a response time of around 2 0 seconds. Galatsis et al [12] have reported the gas sensing properties of Mo 0 3 -Ti0 2 mixed oxide towards O2 , CO and NO 2 . The films were prepared through sol-gel technique. The films were more sensitive towards oxygen (1000 ppm) at 400°C with a response time of around 2-3 min. Further studies from the same group with M 0 O 3 -WO3 showed better
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sensitivity towards oxygen than their T i0 2 counterpart. The optimal temperature was around 420°C. The gas sensing properties of r.f. sputtered V2O5-M0O3 for H2 detection were studied by Imawan et al [13]. Low cross-sensitivity was observed towards N 0 2, NH3, CO, CH4 and S 0 2 gases. The response time was around 20 sec and the sensor was selective to hydrogen at a low temperature of 150°C. The crystallization of (0 k 0) textured orthorhombic phase of M0O3 is observed at stabilization heat treatment temperatures above 350°C. They have also discussed the increased sensitivity attributing it to the promoting effect of V2O5 in facilitating the release of lattice oxygen through interaction with the parent oxide (M0O3). Li et al [14] have developed Mo0 3 -Ti0
2
mixed oxide sensors through sol-gel
technique and discussed their suitability as oxygen sensors. They have identified a metastable hexagonal phase which seemed to be more dominant phase over the stable orthorhombic phase at temperatures above 450°C.
They have also observed the
volatilization of the M0O3 films at temperatures above 550°C leaving large holes within the film. Nanosized Ti-doping on M0O3 thin films was used as novel gas sensing materials in a study by Guidi et al [15]. The Ti doping was performed to enhance the conductance of the M0O3 films. The sensor showed best response to CO and N 0 2 at 3 0 0 °C with a response time of 1 min. Imawan et al [16] illustrate how a multi-layered M0O3 film obtained through repetitive sputtering could improve sensitivity, selectivity, response time and long term stability of M0O3 sensors. The multilayered films also possessed denser and smoother films with smaller crystallite size than their single layer counterpart.
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Single layer films with high porosity showed bad stability which was overcome in multilayered structure. The mechanism of H2 sensing is also explained by a reduction in Mo+ 6 to lower Mo+ 5 and Mo+ 4 states. The existence of inter-grain contact resistance is the dominant effect in single-layered film which controls the resistance of the film.
1.3.2 WO 3 as Gas Sensor Tungsten oxide has been used to detect various gases due to its excellent semiconducting properties. Shaver reported the first WO3 gas sensor for the detection of hydrogen [17]. Since then there has been numerous reports on WO3 sensors for detecting NO 2 and NH 3 and other gases such as ozone, CO, CH4 , and H2 S. The research on WO 3 based gas sensors for NOx detection has increased tremendously during the last 5 years. Akiyama et al [18] established the suitability of WO3 based semiconductor sensor towards NO and NO 2 . They have prepared films of WO3 using the pyrolysis technique. The films were sensitive to NO 2 and NO at 300°C with a response time of 10 sec and 20 sec respectively. They have also demonstrated the selectivity of the films to NO and NO 2 in the presence of other exhaust gases from combustion facilities such as CO, H 2 , CH4 and Isobutane. According to their theory, adsorption was supposed to be the underlying mechanism for the response to NO and NO 2 . In a later study by the same group, the effect of noble metal additions is discussed. Pt electrodes are found to be more suitable for NO detection but Au and Ru are found to be better dopants than Pt. Nelli et al [19] developed sub-ppm NO 2 sensors based on nanosized thin films of Ti-W oxides. The films were prepared through r.f. sputtering followed by thermal oxidation. The films showed response to concentrations as low as 0.5 ppm to 20 ppm.
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Annealing at 500°C yielded monoclinic phase of WO3 . The Ti atoms were segregated onto the surface. The sensor showed selective response to NO 2 in the temperature range of 200°C to 400°C in presence of other gases such as ethyl alcohol, CO and CH4 . A novel semiconductor NO gas sensor which could operate at room temperature was devised by Zhang et al [20], A 300 nm thin film of WO3 was the active ingredient, and was deposited on the p-Si/Al electrode through sputtering process. The response was measured as a function of vertical current across the diode structure (Pt/Pd- WO 3 - p-Si Al structure). The films were heat stabilized at 500°C to transform the amorphous phase into crystalline phase. The concentration of gas used was in the range of 50-250 ppm. The room temperature response is attributed to the interfacial effect of p-Si/WCb or PtPd/W 03. Penza et al [21] have studied the NOx gas sensing characteristics of WO3 thin films activated by noble metals. The W 0 3 films were prepared by r.f sputtering. The noble metal layers were then evaporated on them and Al layers were used as top electrodes. The sensors showed excellent sensitivity 0-10 ppm of NO 2 and 0-440 ppm of NO at an optimum temperature range of 150°C to 200°C. The noble metals have a promotional effect on the speed of response and it also helped to enhance the selectivity in the presence of various other gases such as CO, CH4 , H2 , SO2 , H2 S and NH 3 . The microstructure of the films is reported in a later work by the Penza and other coworkers [22], The tetragonal phase of W 0 3 was identified for the sputtered films through XRD characterization. The average grain size was in the range of 25-30 nm. The films grown at lower growth rate underwent no change in phase after exposure to gases, whereas the
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
films grown at higher growth rate showed that the tetragonal phase was reduced in intensity. The response and recovery times were 1 min and 15 min respectively. Thick films of WO3 and their gas sensing properties dependent of processing conditions was investigated by Chung et al [23]. The parameters varied to obtain different conditions were firing temperature and annealing temperature. These resulted in differences in crystallite size, specific surface area and microstructure. The effect of the oxygen deficiency in WO3 films proved to be an important aspect affecting the sensing behavior. The films were prepared through screen printing and fired at 600°C. The sensing experiments were carried out between 100°C and 250°C. The kink observed in sensing response is attributed to the formation of nitrite type adsorbates which dissociates into nitrosyl type adsorbates. The sensitivity of the gas sensor is also found to be dependent on its resistance and oxygen deficiency. Oxygen content of about 40-50% in the gas flow stream during annealing was found to be suitable towards the detection of N 0 2. Lee et al [24] have developed micro-gas sensors based on WO 3 for nitrogen oxide gas detection. WO 3 films were thermally evaporated onto micro-hotplates prepared by finite element simulation. Annealing was performed at various temperatures ranging from 500°C to 1000°C. The films annealed at 500-800°C were found to be triclinic and those annealed at 800°C were found to be tetragonal. The grain sizes ranged from 0.1pm to 2 pm. The sensor showed optimal response at 300°C towards 10 ppm of N 0 2. The response time was between
1 -2
minutes.
Ehrmann et al [25] devised a novel cheap instrument for breath odor analysis. The sensor array consisted of 40 monolithic sensor elements made of Sn0 2 and WO3 . The
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
sensors were developed mainly to detect NH 3 , H2 S, CH3 SH, indole and putrescine with little cross-sensitivity towards ethanol and humidity. Ammonia and hydrogen sulphide were the main gases responsible for bad breath odor whereas ethanol was an interfering gas. WO3 showed better selectivity than SnC>2 in the presence of ethanol. Xu et al [26] have reported their studies of selective detection of NH3 over NO in combustion exhausts by using Au and M 0 O3 doping on WO3 elements. The oxide materials were prepared by thermal evaporation and Au was introduced by impregnation from a colloidal dispersion. The gas sensing experiments were performed in the temperature range of 200°C to 550°C. They have proposed that M 0 O3 occupies the sites available for NO adsorption thereby suppressing the NO response. The effect of dopants on W 0 3 based films was extensively studied by Wang et al [27]. The sensitivity of the WO 3 film was compared with films doped with 1% metal oxide at 350°C. A wide range of dopant oxides were studied and an equivalent point sensor consisting of a sensor material and a catalyst was designed. These NO/NH 3 equivalent point sensor could be used for real time monitoring and controlling reduction of NO by NH3. The substrate dependence of W 0 3 thin film gas sensors was investigated by Lee et al [28], The sensing behavior on unpolished alumina, polished alumina and silicon substrates were compared. The films on unpolished alumina substrates exhibited the highest sensitivity to
10 ppm NO 2 (S = 70) at an operating temperature of 300°C and
those on Si showed lower sensitivity of S= 3.
The films differed in their surface
roughness and had variable sensitivies. The film thickness was approximately 1.5 pm was deposited through thermal evaporation. The films were annealed at 600°C for 2 hours in
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
air. The sensing tests were carried out within a temperature range of 200°C to 400°C. Rougher films enhanced the absorption of NOx in voids and cracks which increased the sensitivity. The films on different substrates however showed the same phase (triclinic) after annealing. The trimethlylamine (TMA) sensing properties of WO3 prepared by sol-gel technique was investigated by Tong et al [29]. The films were annealed at three different temperatures at 150°C, 300°C and 500°C and it was found that only the films annealed at 500°C showed crystallinity. The TMA response was highest at 75°C in the concentration range of 100-1000°C. The films were also selective to TMA in the presence of NH 3 , gasoline, ethanol, CH4 and CO in the same temperature and concentration range showing that WO 3 films could be used as a good TMA sensor at low temperatures. The response time was 2-6 sec and recovery time was 24-30 sec. In a study by Wang et al [30] the effects of calcining temperature and operating temperature on the gas sensing behavior of WO 3 thin films was investigated. The gases under study were ethanol, petrol, butane and methane. The films calcined at 500°C showed monoclinic and triclinic phases. With increasing the calcinations temperature, higher degree of crystallinity was observed. According to their study, the electron concentration of WO 3 semiconductor is determined mainly by the concentration of stoichiometric defects such as oxygen vacancy. From 100°C to 300°C, the non-linear dependence of resistance with operating temperature was attributed to the change in the charge state of the chemisorbed oxygen-related species such as 0~2 ads, 0 'adS, O H 'adS, and 0
2'ads. Above 300°C, their intrinsic defects such as oxygen vacancies are responsible for
the conductance of the sensor. Maximum sensitivity is obtained at 500°C calcination
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature towards ethanol gas. The gas sensing mechanism is attributed to change in conductance of WO 3 . The reducing gas reacted with oxygen adsorbed on the surface of the sensor leading to e' release from WO3 which leads to decrease in resistance. Fruhberger et al [31] have shown the suitability of WO3 based chemiresistive microsensor for detection and quantification of NO in human breath detection which is an important odor correlated with inflammatory response in asthma. The principle of operation of the sensors was based on oxidizing NO to NO 2 by using an oxidizing agent such as permanganate and a Pt/AfCf catalyst. Cross-sensitivity towards CO2 is less and isoprene interference was minimized by the incorporation of a silicalite filter. Oxygen gas sensing properties of M 0 O3 -WO3 sol-gel based thin films were studied by Galatsis et al [32]. Films of various ratios of the oxides are deposited on Si substrates by spin coating technique. The M 0 O3 -WO3 film showed reproducible response to O2 in the concentration range of 10-10000 ppm at 420°C. The films were annealed at 500°C. Single metal oxide M 0 O3 showed orthorhombic symmetry and the presence of this phase reduced when WO 3 dominates the film composition. The unstable response of single oxides was greatly improved in the case of mixed oxide over wide range of concentrations
(10
ppm -
10000
ppm) however the response signal amplitude was lower
than that of pure M 0 O3 . Extensive stoichiometry and microstructure characterization of tungsten oxide chemiresistive films was performed by Moulzolf et al [33], The films were prepared by rf sputtering in various oxygen/argon mixtures. The oxygen concentration in the plasma had a major effect in the level of oxygen deficiencies in the film. Though bulk WO 3 had well defined phase regions, surface and interface energies in the thin film regime can
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
significantly affect the equilibrium stability. Hence a high temperature phase such as tetragonal phase (stabilized at >720°C) could be stabilized at a much lower temperature depending on the substrate and processing conditions. Films deposited at on a epitaxial lattice match with the (012) 01-AI2 O3 substrate lattice is hence sufficient to induce stability of the thin film tetragonal phase at lower temperatures. The H2 S sensing response of the tetragonal phase and monoclinic phase showed marked differences (tetragonal phase being several orders of magnitude higher) suggesting that oxygen vacancy and charge transport mechanism are dependent on film microstructure. Solis et al [34] have reported the room temperature response of nanocrystalline tungsten oxide thick films to H2 S. The films were prepared by evaporation and a mixture of monoclinic and tetragonal phases were obtained after annealing. The films showed excellent response to low concentrations of H2 S even at room temperature. At heat treatment temperature of about 300°C, the conductance increased by a factor of 104 whereas at heat treatment temperatures above 600°C the response completely vanished. This can be explained by the disappearance of tetragonal phase at those temperatures. This suggests that a specific crystal structure could be responsible for the unique gas sensing properties of nanocrystalline WO3. The films were not sensitive to other gases such as 500ppm of H 2 , 100 ppm of SO2 and 10 ppm of NO 2 at room temperature. The lowest limit of H2 S detectable in their study was
1
ppm.
A semiconducting metal oxide sensor array for the detection of NOx and NH 3 was fabricated by Marquis et al [6 ], The purpose of their study was to engineer a small robust sensitive and selective sensor array capable of detecting NOx and NH 3 from flue gas stream where NH 3 is injected to convert the harmful NOx to benign compounds such as
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
nitrogen and water vapor. WO 3 thin films were used as sensor materials and its properties were altered by modifying substrate material, film thickness, doping, deposition temperature and operating temperature. The films were prepared by sputtering and annealed at 400°C in air for 5 h. The operating temperatures was varied from 200°C to 400°C. The film deposition temperature also varied from 200°C to 400°C. Films were also deposited on various substrates such as sapphire, alumina with varying thicknesses and dopants (Au or Ru). Thus a possible 70 different states of sensors were fabricated and tested. The undoped film was more sensitive to NOx and the film doped with Au was found sensitive to NH 3 . The incorporation of these sensors in the arrays facilitated in selective and sensitive detection of the gases by employing principal component analyses technique. Solis et al [35] in a later study, developed and compared tungsten oxide based semiconductor gas sensors prepared by two different methods. Advanced reactive gas evaporation was used to directly deposit films and to produce ultra fine WO3 powder which was later used for screen printing of thick films. The deposited films were heat treated at 480°C and the screen-printed films were heat treated at 500°C. They consisted of a mixture of monoclinic and tetragonal phases and were both sensitive to H 2 S at room temperature. The directly deposited films were more porous of grain size 15 nm whereas the screen printed consisted of larger grains. Both the films showed similar response to 10 ppm ofH 2 S. Ozone sensing properties of reactively sputtered WO 3 films was investigated by Aguir et al [36]. The films were deposited by rf sputtering and annealed at 400°C. The sensing response is correlated to thin film deposition parameters such as substrate
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature, oxygen partial pressure and annealing conditions.
The films showed 3
orders change in magnitude of conductance upon introduction of 54 ppb of ozone. The response time was 15 sec. The conduction mechanisms depend on oxygen concentration which is in turn affected by operating temperatures. The effect of morphology/microstructure of WO3 crystals prepared by two kinds of gels was studied by Choi et al [37]. The sols were subjected to centrifugal treatment for lh and 10 h at room temperature. Heat treatment was performed in the temperature range of 200-700°C for 2 h. The preferred orientation of (010) was observed in case of room temperature films which disappeared as the heat treatment temperature was increased to 500°C. The sensors have capability to detect NO 2 at sub ppm levels. The films heat treated for
10
h gives better NO 2 sensing properties than those heat treated at
lh. The sensitivity tends to increase with increasing calcination temperature while a reverse tendency is exhibited by the rate of response. The response time at lower heat treatment temperatures such as 300°C was around 15 sec as against 4.3 min for 500°C. It is supposed that the lamellar structures observed in the case of the films heat treated for 10 h provide for most of the adsorption sites for NO 2 . The thickness of the lamella also plays an important effect in sensitivity of NO 2 . NO 2 gas sensing properties of WO3 has also been studied by He et al [38]. The sensor was fabricated using MEMS technology and the sensing tests were performed at around 250°C. The films were deposited onto the membrane by magnetron sputtering. Annealing was done at 600°C for 4 h and films were found to be monoclinic. Films deposited at higher temperatures showed better sensitivity than those deposited at lower temperatures. But the operating temperatures had reverse effect on sensitivity.
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Pitcher et al [39] have reported the current-voltage characteristics of WO 3 thin film sensor to determine the theory of electron transport mechanism in the WO 3 films. The sensor response to ethylene gas was tested to study the effect of thermionic emission and tunneling processes occurring in the films. The difference between amorphous and crystalline films is modeled using dc equivalent circuit composed of contact and interfacial resistances. The oxygen vacancy concentration can be determined through I/V and R/V studies. Tomchenko et al[40] designed and fabricated semiconducting metal oxide sensor array for the selective detection of combustion gases. Drop coating technique was adopted for metal oxide deposition followed by annealing at 600°C. The various oxides used in the array include WO3, ZnO, SnC>2 , CuO and In2 0 3 . PCA was used to identify the various combustion gases such as nitrogen oxides, ammonia, sulfur dioxide and other gaseous pollutants. Choi et al [41] prepared thick films of WO 3 through wet-process sol-gel technique for NO 2 sensing. The wet prepared films were calcined at 300°C leading to the formation of lamellar structures. The films could detect NO 2 down to sub-ppm levels. Higher calcination temperature lead to increased sensor response but was offset by longer response and recovery times. This is believed to be due to formation of micropores in the lamellas at higher temperatures. Tamaki et al [42] have studied the effect of thickness on NO 2 sensing properties of WO 3 films prepared through suspension drop coating. They have observed that thinner films exhibited higher sensitivity than thicker films. In the case of thinner films, NO 2 molecules can easily reach the gaps between two electrodes and are more effectively
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adsorbed leading to higher resistance thus inducing higher sensitivity. For the thicker films, the upper part of the film traps the NO 2 molecules and disturbs the diffusion into the gap between the electrodes inducing lower sensitivity. The films showed excellent response to NO 2 in the concentration range of 0 .0 1 - lppm. Teoh et al [43] have developed a novel mesoporous WO3 based gas sensor capable of detecting NO 2 even at room temperature. The films were prepared by sol-gel encapsulated in a block copolymer. After calcination at 250°C, the copolymer was washed away. After calcination, the films were found to be monoclinic in structure. The grains were very of fine (approx. 3.8 nm) size. These small grains contribute to the excellent gas sensitivity since the diameter is comparable with or less than the space charge region of the grain. The sensor showed response to 3 ppm of NO 2 with a response time of 1-2 min in the temperature range of 35-100°C. It is proposed that the mechanism underlying the NO 2 sensing is due to the adsorption of the unpaired electron in NO 2 which is known as a strong oxidizer. Upon adsorption, charge transfer is likely to occur from mesoporous WO3 to NO2 because of the electron-withdrawing power of NO2 molecules. The NO 2 ions adsorbed at low temperatures on oxide semiconductor surfaces are nitrite type (ONO‘) and dissociate into nitrosyl type adsorbates (NO+, NO'). There have been numerous other reports on M 0 O3 and WO3 in catalysis based on their ability to selectively catalyze certain chemical reactions [44-46]. However, there is limited literature available on gas specificity and mechanisms of gas interactions with various structure sensitive metal oxides. This is the motivation of study for this work.
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1.4 Statement of Problem M 0 O 3 and WO3 are metal oxides which are used in gas sensing, catalysis, photochromic and electrochromic devices due to its excellent semiconducting properties. Most of their physical and chemical properties are same though there are some marked microstructural dissimilarities as well. M 0 O3 has been used as gas sensing material since 1996 and there are few reports indicating its suitability to sensing different gases such as NH 3 , H2 , CO and NO 2 . M 0 O3 has a high resistivity which has limited its applications in gas sensing due to electrical circuitry difficulties. The high resistivity is due to wide band gap of 3.2 eV. M 0 O3 exhibits good response characteristics, but it is very unstable and irreversible which degrades as the operating temperature increases. This is due to its low melting point of 795°C. These two drawbacks have led to very little research on M 0 O3 in gas sensing applications. However, in the lower temperature range (350°C-500°C), it shows excellent sensing response and high selectivity in certain cases. This is the origin of motivation in pursuing this study on M 0 O3 . The sensing response can be controlled by processing conditions and altering the microstructure. Sputtering and sol-gel techniques are investigated due to the ease in varying the control parameters to achieve the desired microstructure. The phase transformation from monoclinic phase to orthorhombic phase usually occurs in the temperature range of 350°C for bulk single crystals. This transition is exploited and suitably tailored by selecting the sol properties. WO3 on the other hand, has been extensively studied during the last 5 years. A
clear understanding of the polymorph selection specific to a particular gas is still lacking.
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Many processing techniques have been chosen to process the thin films; however, we propose to use a novel WO3 thin film preparation route for use as gas sensor material. The aim is to establish the hypothesis that gas specificity depends on the oxide’s polymorph phase used in sensing and is irrespective of the processing route followed. The effect of the polymorph on the gas sensing behavior will aid in the microstructure selection for specific gas detection. Suitable sensor arrays consisting of un-doped metal oxide will be constructed based on the above results.
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References: 1.
N. Guernion, R.J. Ewen, K. Pihlainen, N.M. Ratcliffe, and G.C. Teare, The fabrication and characterisation o f a highly sensitive polypyrrole sensor and its electrical responses to amines o f differing basicity at high humidities. Synthetic Metals, 2002. 126(2-3): p. 301-310.
2.
A.A. Vaughan, M.G. Baron, and R. Narayanaswamy, Optical ammonia sensing film s based on an immobilized metalloporphyrin. Analytical Communications, 1996. 33(11): p. 393-396.
3.
H. Nanto, T. Minami, and S. Takata, Zinc-Oxide Thin-Film Ammonia Gas Sensors with High-Sensitivity and Excellent Selectivity. Journal of Applied Physics, 1986. 60(2): p. 482-484.
4.
L. Deng, H.W. Tan, Y.J. Xu, L.H. Nie, and S.Z. Yao, On-line rapid detection o f urease-producing bacteria with a novel bulk acoustic wave ammonia sensor. Enzyme and Microbial Technology, 1997. 21(4): p. 258-264.
5.
R. Moos, R. Muller, C. Plog, A. Knezevic, H. Leye, E. Irion, T. Braun, K. Marquardt, and K. Binder, Selective Ammonia Exhaust Gas Sensor for Automotive Applications. Sensors and Actuators B-Chemical, 2002. 83: p. 181189.
6
.
B.T. Marquis and J.F. Vetelino, A semiconducting metal oxide sensor array fo r the detection ofN O x and NH3. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 100-110.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7.
J. Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura, and N. Yamazoe, Grain-Size Effects in Tungsten Oxide-Based Sensor fo r NitrogenOxides. Journal of the Electrochemical Society, 1994.141(8): p. 2207-2210.
8
.
D. Mutschall, K. Holzner, and E. Obermeier, Sputtered molybdenum oxide thin films fo r NH 3 detection. Sensors and Actuators B-Chemical, 1996. 36(1-3): p. 320-324.
9.
M. Ferroni, V. Guidi, G. Martinelli, P. Nelli, M. Sacerdoti, and G. Sberveglieri, Characterization o f a molybdenum oxide sputtered thin film as a gas sensor. Thin Solid Films, 1997. 307(1-2): p. 148-151.
10.
E. Comini, G. Faglia, G. Sberveglieri, C. Cantalini, M. Passacantando, S. Santucci, Y. Li, W. Wlodarski, and W. Qu, Carbon monoxide response o f molybdenum oxide thin films deposited by different techniques. Sensors and Actuators B-Chemical, 2000. 68(1-3): p. 168-174.
11.
C. Imawan, F. Solzbacher, H. Steffes, and E. Obermeier, Gas-sensing characteristics o f modified-MoO 3 thin films using Ti-overlayers fo r NH 3 gas sensors. Sensors and Actuators B-Chemical, 2000. 64(1-3): p. 193-197.
12.
K. Galatsis, Y.X. Li, W. Wlodarski, E. Comini, G. Faglia, and G. Sberveglieri, Semiconductor M0O3-HO2 thin film gas sensors. Sensors and Actuators BChemical, 2001. 77(1-2): p. 472-477.
13.
C. Imawan, H. Steffes, F. Solzbacher, and E. Obermeier, Structural and gassensing properties o f V2 O5 -M0 O3 thin films for H 2 detection. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 346-351.
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14.
Y.X. Li, K. Galatsis, W. Wlodarski, M. Ghantasala, S. Russo, J. Gorman, S. Santucci, and M. Passacantando, Microstructure characterization o f sol-gel prepared M oO j-T itf thin films fo r oxygen gas sensors. Journal of Vacuum Science & Technology a-Vacuum Surfaces and Films, 2001.19(3): p. 904-909.
15.
V. Guidi, D. Boscarino, L. Casarotto, E. Comini, M. Ferroni, G. Martinelli, and G. Sberveglieri, Nanosized Ti-doped M 0 O3 thin films for gas-sensing application. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 555-560.
16.
C. Imawan, H. Steffes, F. Solzbacher, and E. Obermeier, A new preparation method fo r sputtered M 0 O3 multilayers fo r the application in gas sensors. Sensors and Actuators B-Chemical, 2001. 78(1-3): p. 119-125.
17.
P J. Shaver, Activated Tungsten Oxide Gas Detectors. Applied Physics Letters, 1967.11(8): p. 255-&.
18.
M. Akiyama, J. Tamaki, N. Miura, and N. Yamazoe, Tungsten Oxide-Based Semiconductor Sensor Highly Sensitive to NO and N 0 2 - Chemistry Letters, 1991(9): p. 1611-1614.
19.
P. Nelli, L.E. Depero, M. Ferroni, S. Groppelli, V. Guidi, F. Ronconi, L. Sangaletti, and G. Sberveglieri, Sub-ppm NO 2 sensors based on nanosized thin films o f titanium-tungsten oxides. Sensors and Actuators B-Chemical, 1996. 31(12): p. 89-92.
20.
W.Y. Zhang, H. Uchida, T. Katsube, T. Nakatsubo, and Y. Nishioka, A novel semiconductor NO gas sensor operating at room temperature. Sensors and Actuators B-Chemical, 1998. 49(1-2): p. 58-62.
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21.
M. Penza, C. Martucci, and G. Cassano, NOx gas sensing characteristics o f WO3 thin film s activated by noble metals (Pd, Pt, Au) layers. Sensors and Actuators BChemical, 1998. 50(1): p. 52-59.
22.
M. Penza, M.A. Tagliente, L. Mirenghi, C. Gerardi, C. Martucci, and G. Cassano, Tungsten trioxide (WO3) sputtered thin films fo r a NOx gas sensor. Sensors and Actuators B-Chemical, 1998. 50(1): p. 9-18.
23.
Y.K. Chung, M.H. Kim, W.S. Urn, H.S. Lee, J.K. Song, S.C. Choi, K.M. Yi, M.J. Lee, and K.W. Chung, Gas sensing properties o f WO3 thick film fo r NO 2 gas dependent on process condition. Sensors and Actuators B-Chemical, 1999. 60(1): p. 49-56.
24.
D.S. Lee, J.W. Lim, S.M. Lee, J.S. Huh, and D.D. Lee, Fabrication and characterization o f micro-gas sensor fo r nitrogen oxides gas detection. Sensors and Actuators B-Chemical, 2000. 64(1-3): p. 31-36.
25.
S. Ehrmann, J. Jungst, J. Goschnick, and D. Everhard, Application o f a gas sensor microarray to human breath analysis. Sensors and Actuators B-Chemical, 2000. 65(1-3): p. 247-249.
26.
C.N. Xu, N. Miura, Y. Ishida, K. Matsuda, and N. Yamazoe, Selective detection o f NH 3 over NO in combustion exhausts by using Au and M 0 O3 doubly promoted WO3 element. Sensors and Actuators B-Chemical, 2000. 65(1-3): p. 163-165.
27.
X.S. Wang, N. Miura, and N. Yamazoe, Study o f W0 3 -based sensing materials fo r NH 3 and NO detection. Sensors and Actuators B-Chemical, 2000. 66(1-3): p. 74-76.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28.
D.S. Lee, K.H. Nam, and D.D. Lee, Effect o f substrate on N 0 2 -sensing properties o f WO3 thin film gas sensors. Thin Solid Films, 2000. 375(1-2): p. 142-146.
29.
M.S. Tong, G.R. Dai, and D.S. Gao, WO3 thin film sensor prepared by sol-gel technique and its low-temperature sensing properties to trimethylamine. Materials Chemistry and Physics, 2001. 69(1-3): p. 176-179.
30.
Y.D. Wang, Z.X. Chen, Y.F. Li, Z.L. Zhou, and X.H. Wu, Electrical and gassensing properties o f WO3 semiconductor material. Solid-State Electronics, 2001. 45(5): p. 639-644.
31.
B. Fruhberger, N. Stirling, F.G. Grillo, S. Ma, D. Ruthven, R.J. Lad, and B.G. Frederick, Detection and quantification o f nitric oxide in human breath using a semiconducting oxide based chemiresistive microsensor. Sensors and Actuators B-Chemical, 2001. 76(1-3): p. 226-234.
32.
K. Galatsis, Y.X. Li, W. Wlodarski, and K. Kalantar-zadeh, Sol-gel prepared M 0 O3 -WO 3 thin-films fo r O2 gas sensing. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 478-483.
33.
S.C. Moulzolf, S.A. Ding, and R.J. Lad, Stoichiometry and microstructure effects on tungsten oxide chemiresistive films. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 375-382.
34.
J.L. Solis, S. Saukko, L.B. Kish, C.G. Granqvist, and V. Lantto, Nanocrystalline tungsten oxide thick-films with high sensitivity to H 2S at room temperature. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 316-321.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35.
J.L. Solis, S. Saukko, L. Kish, C.G. Granqvist, and V. Lantto, Semiconductor gas sensors based on nanostructured tungsten oxide. Thin Solid Films, 2001. 391(2): p. 255-260.
36.
K. Aguir, C. Lemire, and D.B.B. Lollman, Electrical properties o f reactively sputtered WO3 thin film s as ozone gas sensor. Sensors and Actuators B-Chemical, 2002. 84(1): p. 1-5.
37.
Y.G. Choi, G. Sakai, K. Shimanoe, Y. Teraoka, N. Miura, and N. Yamazoe, Preparation o f size and habit-controlled nano crystallites o f tungsten oxide. Sensors and Actuators B-Chemical, 2003. 93(1-3): p. 486-494.
38.
X.L. He, J.P. Li, X.G. Gao, and L. Wang, NO 2 sensing characteristics o f WO3 thin film microgas sensor. Sensors and Actuators B-Chemical, 2003. 93(1-3): p. 463-467.
39.
S. Pitcher, J.A.
Thiele, H.L. Ren, and J.F. Yetelino,
Current/voltage
characteristics o f a semiconductor metal oxide gas sensor. Sensors and Actuators B-Chemical, 2003. 93(1-3): p. 454-462. 40.
A.A. Tomchenko, G.P. Harmer, B.T. Marquis, and J.W. Allen, Semiconducting metal oxide sensor array fo r the selective detection o f combustion gases. Sensors and Actuators B-Chemical, 2003. 93(1-3): p. 126-134.
41.
Y.G. Choi, G. Sakai, K. Shimanoe, N. Miura, and N. Yamazoe, Wet processprepared thick film s o f WO3 for NO 2 sensing. Sensors and Actuators B-Chemical, 2003. 95(1-3): p. 258-265.
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42.
J. Tamaki, A. Hayashi, Y. Yamamoto, and M. Matsuoka, Detection o f dilute nitrogen dioxide and thickness effect o f tungsten oxide thin film sensors. Sensors and Actuators B-Chemical, 2003. 95(1-3): p. 111-115.
43.
L.G. Teoh, Y.M. Hon, J. Shieh, W.H. Lai, and M.H. Hon, Sensitivity properties o f a novel N 0 2 gas sensor based on mesoporous WO3 thin film. Sensors and Actuators B-Chemical, 2003. 96(1-2): p. 219-225.
44.
G. Mestl, C. Linsmeier, R. Gottschall, M. Dieterle, J. Find, D. Herein, J. Jager, Y. Uchida, and R. Schlogl, Molybdenum oxide based partial oxidation catalyst: 1. Thermally induced oxygen deficiency, elemental and structural heterogeneity and the relation to catalytic performance. Journal of Molecular Catalysis A: Chemical, 2000.162: p. 463-492.
45.
J. Haber, J. Janas, M. Schiavello, and R.J.D. Tilley, Tungsten-Oxides as Catalysts in Selective Oxidation. Journal of Catalysis, 1983. 82(2): p. 395-403.
46.
T. Ressler, J. Wienold, R.E. Jentoft, and T. Neisius, Bulk structural investigation o f the reduction o f M 0 O3 with propene and the oxidation o f M 0 O2 with oxygen. Journal of Catalysis, 2002. 210(1): p. 67-83.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 2 Experimental Techniques The films used for gas sensing and microstructure characterization in this study are mainly prepared by two techniques viz. dual ion beam deposition and sol-gel technique. These two techniques were chosen since they offer better control over deposition than other techniques. Since the sensing behavior is affected greatly by deposition conditions (both pre deposition and post-deposition), care is taken while designing the experiment for processing. The film properties such as thickness, grain sizes, polymorphs, vacancy concentration, porosity etc which contribute to differences in sensing behavior are characterized using SEM, TEM, XRD and DSC techniques. The sensing setup is calibrated with the existing setup at Ford SRL through use of a commercial NOx sensor.
2.1 Dual Ion Beam Deposition The dual ion beam deposition (IBD) system used for thin film deposition is shown in *7
Figure 2.1. The base pressure of this system is lx 10' Torr and the process pressure is 1.6 x 10'4 Torr. The deposition system consists of a filamentless Radio-Frequency Inductively Coupled Plasma (RFICP) primary source and a RFICP assist source. The target is watercooled and is mounted at 45 degrees with respect to the primary source. The sample stage can be rotated continuously (2 rpm) or tilted from 0-75 degrees with respect to the target normal and is also water-cooled. The distance from the primary source to the target, the assist
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
source to the sample stage and the target to the sample are 356 mm, 508 mm and 381 mm respectively. Argon is the working gas used inside the chamber, which serves as the gas in primary source channel, plasma bridge neutralizer and one of the two channels in the assist source. The other gas channel in the assist source can be used to inject reactive gases such as oxygen for reactive sputtering applications. The gas flow rates are controlled by MKS mass flow controllers of range 0-10 seem (standard cubic centimeters per minute). A plasma bridge neutralizer with tungsten filament was used to neutralize the excess positive ions and to provide a neutral ion beam. For detailed account of the ion beam deposition process consult reference [1-3]. M 0 O 3 films (-140 nm thickness) were prepared by reactive ion beam deposition from molybdenum and tungsten targets with oxygen in secondary plasma. The ratio of oxygen to argon in the secondary plasma was maintained at 5:5 (seem). An acceleration voltage o f 700V and an extraction voltage of -180V from the deposition source were maintained to provide a continuous ion beam from the primary source. The substrate was kept rotating during deposition to attain better uniformity.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Process Chamber: Base Pressure~ 1*10'7 Torr Process Pressure ~1.6 *10'4 Torr
Gas inlet RIM-210 Assist Source
Turbomolecular and Mechanical Pumps
Working gas: Ar, O 2 , Ne
Cleaning, Etching
Gas inlet (Ar) for PBN
508 mm Water-coole Target
RF Helical Coil Collimated Ar Beam — Gas inlet
356 mm
3 Grid Ion Optics
381 mm
Cryogenic Pump
J
RIM-210 Deposition Source
Rotating Substrate Holder
Figure 2.1. Dual Ion Beam Deposition System (Ref.[3])
29
; Plasma Bridge Neutralizer (PBN)
2.2 Sol-gel Deposition Sol-gel processing is the process of preparing a sol, gelation of sol, and removal of solvent [4]. A sol is a colloidal suspension of solid particles in a liquid. Usually, the particle size of the dispersed phase is ~l-1000nm. A gel is a substance that contains a continuous solid phase enclosing a continuous liquid phase. The continuity of the solid phase gives the elasticity to the gel. The sol may be produced from inorganic or organic precursors. A precursor is a starting compound for the preparation of a colloid. Alkoxides are the most common precursors used. Metal alkoxides have an organic ligand attached to the metal atom or metalloid atom. These alkoxides are used mainly because they react readily with water [5], M (OR) z + H20 --------- ►HO - M(OR)z-i + ROH
(2.1)
Where R is the alkyl group, z is dependant on valency of metal M
In the preparation of transition metal oxides, an alkoxide reaction with an alcohol is often used to adjust the rate of gelation due to alcohol interchange [4], This method is adopted for processing thin films of M 0 O 3 and WO3 . Precursors for molybdenum and tungsten oxide films are prepared from molybdenum isopropoxide and tungsten isopropoxide respectively. These alkoxides were mixed with n-butanol to obtain 0.1M solutions. Since these alkoxides were reactive to atmosphere, the mixing was done inside a glove box under nitrogen atmosphere. After mixing, the sols were mechanically agitated for 5 minutes inside the glove box and then sealed airtight. Ultrasonic agitation was then performed for 2 hours and the sols were
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
allowed to age and settle. In the case of molybdenum oxide sol, a black opaque liquid was obtained after 24 hours of aging. A transparent brownish yellow sol was obtained in the case of tungsten oxide sol after the same aging period. Thin film deposition was done by spin coating method. The sols were dropped on the sensor substrates and spun at 2500 rpm for 30 sec in a spin coater (Chemat Technology- KW-4A). For obtaining thicker films of comparable thickness to that obtained by ion beam deposition, the spinning was repeated 10 times with baking at 75°C between spins for 15 minutes. After 5 spins, the films were left overnight for gelation and condensation and then heat treated at 500°C for 1 hour in air before the final 5 spins. The final heat treatment step was repeated after
10
spins.
2.3 TEM Characterization In order to determine the grain sizes and the phases present in the films deposited by ion beam deposition and sol-gel processes, TEM characterization was performed. The transmission electron microscope used for this purpose is the Philips CM 12 with LaE$6 cathode. The incident energy of electrons under which this was carried out is 120 keV.
2.3.1 IBD Samples Samples for TEM characterization from ion beam sputtered samples were prepared by direct deposition of the films on standard TEM grids made of Cu with formvar coating. The TEM samples were heat treated at 500°C for
8
hours to stabilize the
phase. After stabilization heat treatment, the samples were observed under the TEM. This method was employed for both M 0 O3 and WO3 samples.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.3.2 Sol-gel Samples Samples for TEM characterization from sol-gel method were prepared by dropping the sol over the TEM grid with formvar coating and allowing it to gelate and condense overnight. After condensation, the grids were heat treated at 500°C for 1 hour and
8
hours. The same method was employed for both M0O3 and WO3 samples.
2.4 SEM Characterization In order to understand the sensing behavior of the sensor due to different processing conditions, scanning electron microscopy (SEM) investigations were carried out on a LEO-1550 Field Emission Gun SEM. The response of the sensor is explained in terms of morphology and porosity. Secondary electrons were used for imaging and images were obtained at various accelerating voltages and working distances. All the sensors were observed directly (without any special specimen preparation techniques) before and after they were tested.
2.5 XRD Characterization Philips X-Ray Diffractometer PW 1729 was used to characterize the films deposited on sensor substrates and to confirm the phases present in the stabilized sensors. The diffractometer operates at 40kV and 30mA. X-rays are irradiated over an area of 2x1.5 cm2 region and has a 20 range from 20° to 80°. The Philips computer generates a file of 20 versus I from which d values can be calculated from the relation d = X / (2 sin 0)
(2.2)
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
where X is the wavelength of Cu Ka radiation which is 1.54184 A. The d values are then compared with standard JCPDS powder diffraction data for the material under test and possible phases are identified.
2.6 Testing of M o03 and W 03 Films The heat-treated films were used for gas sensing in the setup described below. The gases used in the sensing setup were UHP nitrogen (Praxair), UHP oxygen (Praxair), lOOOppm ammonia in nitrogen (BOC gases) and 1000 ppm NO 2 in nitrogen (BOC gases). Concentration of ammonia and nitrogen dioxide was varied by varying their flow rates in conjunction with nitrogen flow rates. The gases were controlled through 1479 MKS Mass flow controllers whose channels were connected to a Type 247-MKS 4-channel readout which is calibrated to read the flow rate of the gases directly in seem. The combined flow rate of the gases was maintained at
1000
seem.
The gas mixture is then passed through a tube furnace (Lindberg/Blue), which can be heated at a programmed rate. The sensor is placed inside the tube furnace with quartz tube and is electrically connected to outside leads using gold wires (Alfa Aesar, 0.25mm dia, 99.998%). Connection at the hot end was made by spot welding using Mo-clad Cu electrodes on a Unitek Miyachi Unibond II system (Model
86
) and at the cold end the
gold wires were soldered to the Cu-Ni alloy wires fused to the quartz. The change in resistance of the sensor is measured in terms of voltage change across it using an Agilent 34401A digital multimeter and by calculating the resistance from monitoring the voltage across a standard resistor using another digital multimeter (Fluke 29592). In order to measure the resistance of the sample, the following electrical circuitry was devised.
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
+
-
Figure 2.2. Electrical circuit for resistance measurement
Rs
= Standard 10 or 1 Mf2 resistor
Vs
= Voltage across Rs
R
= Sample resistance (variable)
V
= Voltage across R
I = Vs/Rs
(2.3)
R = V/I Substituting for I from equation (2.3), the resistance of the sample R is R = (VRS)/VS
(2-4)
The resistance variation with concentration of the gases is plotted against time for both M 0 O3 and WO 3 samples.
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The schematic diagram of the experimental setup of the gas sensing system shown in Figure 2.3. Exhaust
Mass Flow IControllers
n
2
Furnace
o2 nh3
nnnn
Computer
Sensor
Multimeter
Figure 2.3. Schematic Diagram of the experimental setup of the Sensing System
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References: 1.
J.L. Vossen and W. Kern, Thin Film Processes. 1978, New York: Academic Press.
2.
K. Williams, A Study o f Reactive Ion Beam Assisted Deposition o f Ni,NiO, and Ni/NiO Cermets and their Magnetic Properties, Ph.D. Dissertation in Material Science and Engineering. 2000, SUNY at Stony Brook: Stony Brook, p. 31-37.
3.
K.W. Lin, Structural and Magnetic Characterization o f Ion-Beam Deposited NiFe/NixFei.xO Composite Films, Ph.D. Dissertation in Material Science and Engineering. 2002, SUNY at Stony Brook: Stony Brook, p. 17-19.
4.
M.A. Aegerter, J.M. Jafelicci, D.F. Souza, and E.D. Zanotto, Sol-Gel: Science and Technology. 1989, Singapore: World Scientific.
5.
C.J. Brinker and G.W. Scherer, Sol-Gel Science: The Physics and Chemistry o f Sol-Gel Processing. 1990, San Diego: Academic Press, Inc.
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 3 M 0 O3 Thin Films as a Gas Sensor This chapter focuses on the study of M 0 O3 thin films as a gas sensor material. Thin films of M 0 O3 were prepared by ion beam deposition and sol-gel techniques. Three sets of experiments were carried out - (a) The effect of thickness on the gas sensing behavior was studied in an effort to determine whether the sensing process is surface processes-controlled or bulk-diffusion controlled; (b) Stabilization heat treatments over a range of temperatures to determine phase transformation regions and (c) Sensing temperature to determine suitable operating temperatures for gas detection.
3. 1 Effect of Thickness Three sets of samples of different thicknesses were prepared by ion beam deposition. The film thicknesses were controlled by the time of deposition. A 4 min deposition time yielded 15 nm thick films; 40 min deposition yielded 150 nm and 400 min deposition yielding films of 1500 nm thickness. Following deposition, the samples were annealed at 500°C for
8
hours to stabilize the phases. The samples were analyzed
using SEM to assess the morphology of the films before and after sensing. XRD tests were used to identify the phases obtained before and after sensing and the results are compared.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.1 SEM characterization Figure 3.1 (a) and 3.1 (b) show the morphology of the films deposited for 4 min before and after sensing respectively. In Figure 3.1 (a) the observed fine features (as indicated by arrow marks) on the relief pattern (substrate) represent the M 0 O3 grain aggregates. These are approximately 200-250 nm in size.
S T O N Y SSSMA.t«u™. 3 3 0 ,w
IMag*iCBOKX
T « w :t* 1 ;3 6
B R f l l S K £H T « 15.00 VV T W ».ta-3840
Figure 3.1. SEM images of films deposited for 4 min (a) before sensing (b) after sensing In Figure 3.1 (b) the white particles dispersed on the larger grains are M 0 O3 particle structures. These particles are 150-200 nm in size. There is considerable reduction in the number of grains following the sensing tests but their morphology remains the same.
BRi#K
[W D » 8m m l« a g » 1000 K X
ts««
2jitn j f
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
c)
■
■ If
wmm
Figure 3.2. SEM images of films deposited for 40 min (a) before sensing deposited on alumina (b) after sensing - on alumina region (c) before sensing deposited on gold electrodes Figure 3.2 (a) and 3.2 (b) show the SEM image of the films deposited for 40 min before and after sensing respectively. As seen in Figure 3.2(a), the films are denser than in case of the films deposited for 4 min. Moreover, the morphology of the M 0 O3 appears to take the form of the substrate beneath it. The films deposited on the alumina portion of the substrate are shown above. The part of the film grown on the gold electrodes is smoother and consists of faceted structures as shown in Figure 3.2 (c). In figure 3.2 b), the morphology of the film remains almost the same as before sensing except that the grain clusters are more “broken”. This is possibly due to grain coarsening. This is confirmed by the intensity of the XRD peaks obtained after heat treatment. This is discussed in later section (Section 3.1.2) in this chapter. The grains are 1.2 pm in length and 400-450 nm thick.
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 3.3 (a) and 3.3 (b) show the SEM image of the films deposited for 400 min before and after sensing respectively.
W 0»' ftwm
**»$»taoexx
|
«e Na nc s *
ST#NY
O«4»c1«Nw»J04
BW NMCeMf*w-00kv Tkwa&s**4
Figure 3.3 SEM images of films deposited for 400 min (a) before sensing (b) after sensing As seen in Figure 3.3 (a), the film is similar in density as in the case of that deposited at 40 min. The deposition is more uniform on both the alumina-part and goldelectrode part o f the substrate. The grains are larger than the ones deposited for 40 min, but the morphology of the grains is similar to those obtained after 40 min deposition. As observed in the above case, the films after sensing (Figure 3.3 b) are more broken than the films before sensing due to possible grain coarsening. The grains are 1.3-1.5 pm in length and 300-400 nm in thickness.
3.1.2 XRD characterization Figure 3.4 a) and b) show the XRD plots of the films before and after sensing. As observed there isn’t any phase difference in films before and after sensing confirming the reversible nature of sensing mechanism. The XRD peaks from the films are marked (F)
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and they are almost o f same relative intensity. JCPDS file matching indicates that the
sensing phase is orthorhombic M 0 O3 . Before Sensing
a)
< £
400 min 6000
- 4 0 min
-4 min
s
i
is
I idfaiwi
i - J i — ih-------- :— L - l .
? N i o O ) i o i n s i - N i n n r t t i o s ® « N ’- © » - ( O N « ) t ,o
s
^
N
O
K
N
i f
i ' t
n
N
' -
o
c
V^ ( 0 < * ) ( i n ( MNCMNNNCNN
i ' i
a
s
N
i D
i o
i
f
i i
n
' f
^
T
IL N »o nr N »
t n
d (Angstroms)
After Sensing
b)
J
3 2500
iw
I 2000
rs. o ^
- 4 min
l
<0 <0 01 0
t o o> t o i-» ^ CO CO
- 40 min 400 min
.. N .. co c o o > < o m i« t i n < o o o T N I O ^ M N ' - O O I O I C O N N p CO (N
NCiciNNNNrr
N
M
N
(
00
) i n N
» m
d (Angstroms)
Figure 3.4. XRD of films (a) before sensing (b) after sensing (S- Substrate peaks, FFilm peaks)
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.3 Sensing tests Figure 3.5 shows the comparison of resistances of the films of different thicknesses and the change in resistance upon introduction of 100 ppm and 400 ppm NH 3 gas. The resistance is plotted against sensing temperature. 1% 0 . 1e+11
— 40 min - A — 400 min ♦ 4 min
1e+10 1e+9
1e+8 1e+7
1e+6 1e+5
^
1e+4
lOOppmNH. 400ppm^lH.
1e+3
- —
A
1e+2 360 380 400 420 440 460 480 500 520
T(°C) Figure 3.5 Plot of resistance o f films (with and without ammonia gas) against sensing temperature for films of different thickness
15
150
1500
T h ick n ess (nm)
Figure 3.6 Sensitivity Versus Thickness of IBD M 0 O3 films
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The resistance of films deposited for 4 min is higher than the resistance of the films deposited for 40 min and 400 min due to non-uniform distribution of the film over the substrate as is confirmed by the SEM results. As is seen from the plot, the resistance drop of the sensor upon introduction to a reducing gas such as ammonia, is lowest for the film deposited for 4 min (sensitivity is less than one order of magnitude) and is somewhat similar for the films deposited at 40 min and 400 min (2 orders change in magnitude at 100 ppm ammonia, and 3 orders change in magnitude upon introduction of 400 ppm of ammonia). This shows that the sensitivity is more dependent on the film density rather than just on film thickness. The films deposited for 40 min and 400 min are more or less equal in density and hence the sensitivity which results from transfer of charge across grain boundaries is more or less the same for these films. The films deposited for 4 min consists of M 0 O3 grains which are disconnected from each other thereby increasing its resistance. Hence, the gas adsorption rate on the surface on these films (deposited for 4 min) is not a rate limiting step in affecting the sensitivity since the high resistance due to disconnected grains controls the sensitivity. In Figure 3.6, the sensitivity of IBD M 0 O3 films to ammonia gas at two different concentrations is plotted against the thickness. It is observed that the films deposited for 4 min have a lower sensitivity to ammonia than the films deposited for 40 min and 400 min. Hence, it can be inferred that the gas sensitivity is independent of thickness once a monolayer of film has been formed over the substrate sufficient enough to aid the transfer of charges across the electrodes. This shows that the film thickness is not a major factor in sensitivity and the surface reactions are more critical than the oxygen diffusion rate through the bulk.
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2 Effect of Polymorphism In order the study the effect of stabilization heat treatment on sensing behavior, the following sets of samples were prepared. Four sets of samples were prepared from sol-gel synthesis. The methodology followed in depositing the samples were according the procedure explained in Chapter 2. The first set of samples were heat treated at 350°C for
8
hours, the second set at 400°C for
final set at 500°C for
8
8
hours, third set at 450°C for
8
hours and the
hours. All the heat treatments were done in static air.
3.2.1 Films Heat Stabilized at 350°C Figure 3.7 shows the SEM image of the films heat treated at 350°C for
WO - 9 m m Mag = 5 .0 0 K X
1 0 Mm Hie Name * 072904-056.tif |----------------------1
ST @ N Y B R IM H C
8
hours.
S ig n a lA = R B S O Date :29 Jul 2 0 0 4 EHT= 15.00 kV Tim e :4:06:57
Figure 3.7 SEM image of sol-gel M 0 O3 films heat treated at 350°C for
8
hours
As seen from the figure above, the films are amorphous. Selected area diffraction of the films on the TEM showed diffuse rings confirming the presence of amorphous film. Figure 3.8 shows the XRD spectrum of the above films. The peaks observed are
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
those of the substrate with gold interdigitated electrodes. The broad monoclinic phase (indicated by arrow mark) of the film indicates the onset of crystallization. 35 0 d eg C
3500
3000 -
2500
6 2000
1500
1000
500
in CO
co vo • »o
■«t C O n.
CO CO
CO
^
n CM
in
CM
n CM
d (Angstroms)
Figure 3.8 XRD spectrum of films heat treated at 350°C (S- Substrate peaks, F- Film peaks) Sensing tests performed on the above films showed no response towards any gas (viz. ammonia or NO 2 ). This shows that amorphous M 0 O3 is not a sensing material for ammonia or NO 2 sensors.
3.2.2 Films Heat Stabilized at 400°C Figure 3.9 shows the SEM image of the M 0 O3 films heat treated at 400°C. The formation of the plate-like structures which corresponds to monoclinic phase of (Oil) orientation is observed. These structures are formed on discrete islands of the M 0 O3 film over the substrates. Figure 3.10 shows the TEM image of the grains of the M 0 O3 film
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
deposited over the substrates. The TEM samples were prepared by removing material from sensor substrate and dispersing them in ethanol and dropping over TEM grids. The grains are around 150-200 nm.
File N am e
=0 7 2 S tm J 4 1 .tif
— 1
STONY
Signal A * R B SD D ate :29 Jul 2 0 0 4
B R # # K
E H T =1500kv
Figure 3.9 SEM image of M 0 O3 film heat treated at 400°C for
8
T im e :3:36:08
hours.
Figure 3.10 TEM image showing grains of Monoclinic phase, inset - SAD pattern of one of the grains showing [011] zone axis of monoclinic phase (Space Group:P21/c (14)) The inset shows the SAD pattern of grains showing [011] zone axis of monoclinic phase.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
400 deg C
400 Inset 3500
350 3000
-
1
• •
250
2500
£ 200 a
2000 150
100
1500
1000
500
d (Angstroms)
Figure 3.11 XRD spectrum of sol-gel M 0 O3 films heat treated at 400°C (S-Substrate peaks, F - Film peaks, Inset- monoclinic peak around 3.74 A) Figure 3.11 shows the XRD spectrum of the films heat treated at 400°C. There is a reduction in amorphous phase as seen by sharpening of peaks around d=3.74 A. At this temperature, there is also onset of transformation of P-monoclinic phase to aorthorhombic phase as indicated by the peak around d=1.82 A corresponding to (0 0 2 ) phase of a-phase.
3.2.3 Films Heat Stabilized at 450°C Figure 3.12 shows the SEM image of the sol-gel films heat treated at 450°C. The plate-like structure observed after heat treatment at 400°C disappears and is transformed into orthorhombic phase. Small grainy structure starts to appear as is seen in TEM image shown in Figure 3.13. 47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
WD ~ Mag -
10m m 5 .0 0 K X
10Mm Rle Name - 072904-032.bf |----------------------1
ST®NY S ig n a lA = R B S D B R # # K
EHT = 1 5 .0 0 k V
Figure 3.12 SEM image of films heat treated at 450°C for
8
Date :29 Jul 2 0 0 4 Time :3:14:09
hours
15nm
Figure 3.13 TEM image of M 0 O3 film after heat treatment at 450°C; inset - SAD pattern of one of the grains showing [010] zone axis of monoclinic phase (Space Group :P21/c (14)) The grains are much smaller than the monoclinic grains formed at 400°C. This is due to formation of new grains upon transformation. The grains are approximately 15-30 nm in size. The inset shows the selected area diffraction image of one of the grains with
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
monoclinic phase of [010] zone axis. This state consists of mixture of monoclinic and orthorhombic phases as is shown by XRD spectrum in Figure 3.14. 4 5 0 d eg C
450 Inset 3500
GOO 3000 -
-r 550 D < 500 — ———
2500 -
3 2000
-
400 350
1500
d (Angstroms) 1000
500 -
d (Angstroms)
Figure 3.14 XRD spectrum of films heat treated at 450°C (S-substrate peaks, F- Film peaks, Inset - Mixed monoclinic and orthorhombic peaks) The peaks around d=3.74A, d=3.69A and d=3.24A represent a mixed-phase of orthorhombic and monoclinic phase of M 0 O3 . The sharp peak around d=1.82A shows the formation of strongly (0 0 2 ) oriented orthorhombic phase.
3.2.4 Films Heat Stabilized at 500°C Figure 3.15 shows the SEM image of the films heat treated at 500°C. As is seen from the image, the film has gone complete transformation into uniform aggregates. Grain growth has occurred from previous stage as is confirmed by the grain size from TEM image in Figure 3.16.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
WD = 10 mm Mag - 5.00 KX
10(im~
File Name = 072904-016.tif
|----------- 1
ST#NY BR®#K
Signal A = RBSD Date :29 Jul 2004 EHT = 15.00kV Time :2:35:13
Figure 3.15 SEM image of films heat treated at 500°C.
Figure 3.16 TEM image of films heat treated at 500°C inset - SAD pattern of one of the grains showing [100] zone axis of orthorhombic phase (Space Group :Pbnm (62) ) The grain size is of the order of 100-120 nm and the selected area diffraction shows orthorhombic structure of [100] zone axis. This is confirmed by the XRD results (Figure 3.17) which shows complete transformation to orthorhombic phase.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M ( N T r { D o > c M t D r ' - i n ( O i n * - i ^ r ) ^ r c o c y3 r ) i ' - t n c ' j t^ ■oc os T ^j - rN~ ^ or ^ c-
t
ri
PI
ri
PI
N
d (Angstroms)
Figure 3.17 XRD Spectra of M 0 O3 thin film heat stabilized at 500°C (S- Substrate peaks, F- Film peaks, Inset - orthorhombic peaks) The peaks around d=3.69A represent of orthorhombic phase of M 0 O3 (110). The sharp peak around d=1.82A shows the formation of strongly (0 0 2 ) oriented orthorhombic phase. Phase transformation from P to a phase of M 0 O3 occurs in the region of 400°C to 500°C. This transformation is gradual as is seen from the presence of small monoclinic peaks even at 450°C. However, above 450°C, the transformation occurs rapidly and the transformation is complete at 500°C.
3.3 Effect of Operating Temperature In order to identify the optimum operating temperature for specific sensing, sensing experiments were carried out in the range of 385°C to 500°C. The thin films
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
obtained from ion beam deposited samples and sol-gel samples were used for this study. The experiments on ion beam deposited samples were carried out in Ford Scientific Research Labs.
a)
b) 1
%o. = 1V
1% 0 2 Vapplied = 1V
applied
2.5e-4
4e-5 400ppm NH3
T1finnsor = 493 C "
2.0e-4 100ppm NH3
3e-5 1.5e-4 2e-5 -
2* 1.0e-4 5.0e-5
1e-5 -
0.0
8
10
12
14
16
18
20
22
.I.... I.
24
38
time (min)
40
42
, I.... I.... I.... I. 44
46
48
50
52
54
56
time (min)
Figure 3.18 Response of ion beam deposited M 0 O3 thin films to (a) 400 ppm ammonia (b)
100
ppm ammonia Figure 3.18 shows the response of the thin films to ammonia at two different
concentrations (a) at 400 ppm (b) at 100 ppm. The sensing tests were performed at three different temperatures as shown - 385°C, 438°C and 493°C. As is observed from the graphs, the sensitivity is low till 438°C (less than one order change in magnitude) above which there is more than an order change in magnitude. This might be due to transformation in structure above 450°C to orthorhombic phase which favors ammonia sensitivity. This effect is also observed with sol-gel films shown below. Figure 3.19 shows the comparison of sensitivities to ammonia at various concentrations at 4 different operating temperatures - 400°C, 425°C, 450°C and 475°C.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
NH3 Sensitivity Vs Concentration
0.7
£
>
-♦ -4 0 0 - a - 425 — 450 -x -4 7 5
0 -6
1 0.5 c
»
0.4
C C C C
0.3 0.2
50
100
300
200
400
500
Concentration (ppm)
Figure 3.19 Comparison of Sensitivities of sol-gel M0 O3 films to various concentrations of ammonia at different operating temperatures From the graph shown above, it is observed that the gas sensitivities increase above 425°C. While sensitivity of the films towards ammonia increases in the temperature range from 400°C to 500°C, the sensitivity towards nitrogen dioxide decreases. This is expected behavior based on the fact that at 400°C, the film is mainly monoclinic (P-M0 O3 ) and this phase is iso-structural with Re0 3 -type structure of WO3 [1] which is known to be sensitive to N 0 2 [2-9]. Above 450°C however, the monoclinic phase is absent which leads to reduced N 0 2 sensitivity. Figure 3.20 shows the sensitivity of sol-gel M0 O3 at various operating temperatures to N 0 2.
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
NO; Sensitivity Vs Concentration 0.5 0.45 0.4 0.35
*>
°-3
400 deg C 450 deg C - A - 475 degC
1 0.25 c « to 0.2
0.15
0.05
50
100
300
200
400
500
Concentration (ppm)
Figure 3.20 Comparison of Sensitivity of sol-gel M 0 O3 to NO 2 at different operating temperatures.
From the above results, it is concluded that the temperature range of 450°C 500°C is a suitable range for specific ammonia sensing by 01-M0 O3 . Irrespective of the processing method chosen (through either sol-gel or ion beam deposition), through suitable polymorph selection, specificity is thus achieved.
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References: 1.
E.M. Mccarron, \S-M0 O3 : A Metastable Analogue o f WO3 . Journal of Chemical Society, Chemistry Communications, 1986: p. 336-338.
2.
M. Akiyama, J. Tamaki, N. Miura, and N. Yamazoe, Tungsten Oxide-Based Semiconductor Sensor Highly Sensitive to NO and NO 2 . Chemistry Letters, 1991(9): p. 1611-1614.
3.
M. Penza, C. Martucci, and G. Cassano, NOx gas sensing characteristics o f WO3 thin film s activated by noble metals (Pd, Pt, Au) layers. Sensors and Actuators 13Chemical, 1998. 50(1): p. 52-59.
4.
M. Penza, M.A. Tagliente, L. Mirenghi, C. Gerardi, C. Martucci, and G. Cassano, Tungsten trioxide (WO3) sputtered thin films fo r a NOx gas sensor. Sensors and Actuators B-Chemical, 1998. 50(1): p. 9-18.
5.
Y.K. Chung, M.H. Kim, W.S. Urn, H.S. Lee, J.K. Song, S.C. Choi, K.M. Yi, M.J. Lee, and K.W. Chung, Gas sensing properties o f WO3 thick film fo r NO 2 gas dependent on process condition. Sensors and Actuators B-Chemical, 1999. 60(1): p. 49-56.
6
.
D.S. Lee, K.H. Nam, and D.D. Lee, Effect o f substrate on N 0 2 -sensing properties o f WO3 thin film gas sensors. Thin Solid Films, 2000. 375(1-2): p. 142-146.
7.
Y.G. Choi, G. Sakai, K. Shimanoe, N. Miura, and N. Yamazoe, Wet processprepared thick film s o f WO3 for NO2 sensing. Sensors and Actuators B-Chemical, 2003. 95(1-3): p. 258-265.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
.
J. Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura, and N. Yamazoe, Grain-Size Effects in Tungsten Oxide-Based Sensor fo r NitrogenOxides. Journal of the Electrochemical Society, 1994. 141(8): p. 2207-2210.
9.
J. Tamaki, A. Hayashi, Y. Yamamoto, and M. Matsuoka, Detection o f dilute nitrogen dioxide and thickness effect o f tungsten oxide thin film sensors. Sensors and Actuators B-Chemical, 2003. 95(1-3): p. 111-115.
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 4 Novel WO3 Thin Films as a Gas Sensor WO 3
thin
films
prepared
by
various
processing
techniques
vacuum
evaporation[l], pyrolysis [2], sputtering [3-11], sol-gel [12-16] have been investigated as suitable materials for gas sensors. This chapter focuses on development of WO 3 thin films prepared by acidic precipitation [17-19] as suitable materials for gas sensors. A comparison of the sensing response to NO 2 and NH 3 among the novel material and earlier work with sol-gel and ion beam deposited materials [20] is provided in Chapter 6 .
4.1 Thin Film Preparation Tungsten oxide thin films were prepared from WO3 powders obtained through acidic precipitation. The powders were obtained from Dr. Csaba Balazsi *. The detailed preparation method is described elsewhere [17-19]. A brief description is given here. Hydrated tungstic acid is prepared by acidification of sodium tugstate solution. Sodium tungstate was dissolved in deionized water and the solution was cooled to 5°C. Hydrochloric acid acid cooled to the same temperature was added to the solution in several doses. The mixture was stirred for 1.5 h in an ice bath and for half an hour at room temperature. The precipitate was separated by centrifuging, and the top layer, rich in sodium is removed. Washing of the precipitate was carried out by addition of water to * Csaba Balazsi, Senior Scientist, Research for Technical Physics and Materials Science, Hungarian Academy of Sciences, H-1525, Budapest, Hungary.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
precipitate at room temperature followed by stirring and keeping the solid particles dispersed in the liquid [18]. The washing process is repeated several times to obtain high purity precipitates low in sodium content. After washing, precipitate is dried and calcined. The two different phases in this study, viz, the orthorhombic and monoclinic phases, were obtained by varying calcination temperatures. The calcined and dried powders were mixed with butanol (1 gm in 5 ml) to obtain solutions of tungsten oxide. The solutions were then spin coated onto sensor substrates using spin coater. The spinning was done at 1500 rpm for 30 sec and was repeated 5 times and then solution was dropped on the spun coated layers to obtain films of 150-200 nm thickness.
The films were then dried and observed under SEM in as-received room temperature state, and after heat treatment at 500°C. XRD patterns were recorded at the two conditions using PW1729 Philips X-ray diffractometer (description in Chapter 2). Sensing experiments were carried out in the temperature range of 200°C to 500°C and the sensitivities are plotted. The sensitivity is defined as S = A R / R air
(4.1)
Where AR is the difference between R ajr and R gas
4.2 Characterization of the W 03 films The monoclinic and orthorhombic phases of WO3 powders and the thin films prepared from them by the above technique were labeled W2 and W3 respectively throughout this chapter and later references.
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2.1 SEM Characterization Figures 4.1a&b show the SEM image of the as-received W2 film before heat treatment.
%f«siA«RBS0 Oai» 35 Eeb 20&5 ENT* IS.SOfcV Tra*
[wo* sarara
m
* G2240S-042.tf
jMag* 1.80KX
Oste 33 F#b 2006
BKINWC ^T*?S.S0kVT1ra**«4;<«
Figure 4.1 SEM image o f as-received W2 film a) at 20kX b) at 1 kX As seen from Figure 4.1a&b) the film is comprised of fibrous mass. Some individual fibers are of 200-300 nm in lateral size.
EMT-* 15.00 feV Tms 2:27:18
£HT*15.00feV TiflW.-2:21:31
Figure 4.2 SEM image of W2 film after heat treatment at 500°C for
8
hours a) at 20kX
b) at 1 kX
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.2a&b show the film after heat treatment at 500°C for
8
hours. The films
are similar, consisting of 250-350 nm lateral sized fibers (calculated from Figure 4.2a). This shows that the film morphology doesn’t change considerably upon heat treatment.
£™'ki ' .j* Hie t a n * « Q & m 023.tif
___ . ' ■. ijkl3K -r*t5t50WTwmS;2852 ST0N Y
S«nalA»fc6S0 U « ^ r « j 2 0 S 5 EHfMS.COW
S«T4fA»RBSO
Figure 4.3 SEM image of W3 film in as received state a) at 20 kX b) at 1 kX Figure 4.3a&b show the W3 film in the as-received state. The film consists of small grains as compared to the fibrous the monoclinic films. The average grain size is calculated to be around 150 nm and the agglomerates are of 80pm average size.
WO* 11r
Mag- 1.00KX|20pmJ
WO* 11m » Km Msg* 20.00 XX f—H
ffte n*m » 031205-029,W
Figure 4.4 SEM image of W3 film after heat treatment at 500°C for
5Sj»nalA*8B5D 0«*;12:Mar2eos ENT* 15,00 kV Tim* :3:C4:41
8
hours a) at 20 kX
and b) at 1 kX
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.4a&b show the W3 film after heat treatment at 500°C for
8
hours. Film
show similar morphology in single grains but there is disappearance of film in agglomerates which might be due to evaporation of film over prolonged heat treatment. The grain sizes are in the range of 150-170 nm indicating that there is grain growth though the average agglomerate size remains constant around 80pm.
4.2.2 XRD Characterization Figure 4.5 shows the comparison of XRD spectrum of W2 and W3 films in the asreceived state. The most prominent higher order peaks are indexed (JCPDS 88-0545 for monoclinic and JCPDS 20-1324 for orthorhombic). As Received Film 600 -i
OO 500 -
400 -
W2 * 300 ■
W3
200-
100
-
Figure 4.5 Comparison o f XRD spectrum of W2 and W3 films in as-received state As is observed from Figure 4.6, the W2 film undergoes phase transformation from monoclinic to orthorhombic phase upon heat treatment at 500°C. From the height of the
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
peaks o f W3 film, it is also confirmed that grain growth occurs during prolonged heat treatment. HT Films 600
OO
500
400 -
-----W2
£ 300 •
200
-
100
•
-----W3
Figure 4.6 Comparison of XRD spectrum of W2 and W3 films heat treated at 500°C for 8
hours Monoclinic phase transforms to orthorhombic from in the temperature range of
350-450°C depending upon the grain growth and the processing route. This transformation is also observed by Csaba et al [17] and in studies in our group using tungsten isoproproxide sol-gel route. DSC studies performed with the sol-gel precursor in the temperature range of 25°C to 700°C range confirmed the formation of monoclinic and orthorhombic phases around 370°C and 490°C. The transformation is gradual in the range of 390°C to 490°C and is complete at around 500°C. The above results suggest that the monoclinic phase, once stabilized, is stable up to the temperature of around 400°C above which phase transformation from monoclinic
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to orthorhombic phase occurs which leads to disturbances in properties of the film including changes in sensing response. The orthorhombic phase, once stabilized, is stable well up to 500°C and beyond till 600°C after which the tetragonal phase starts to appear
DSC/TGA data for sol-gel W03 M ass r»l 105
DSC
TGA 0.60
■■
0.40
- • \
100
0 .2 0 - • -•95
0.X -•90
- 0.20 -
-0.40 -
-■85 - 0.60 -■
-
1.00
-
- 1.20
149
449
209
509
569
629
Figure 4.7 DSC/TGA graphs of dried sol-gel WO 3 (M- formation of Monoclinic phase, O- Formation of orthorhombic phase)
4.3 Sensing Tests Sensing tests were performed with NO 2 and NH 3 from 100°C up to 500°C. However at 100°C, the sensor took infinitely long to recover to the initial resistance due to high activation energy for the adsorbed NO 2 to be released back and the recovery times
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were long. Considering fast recovery times as an essential criterion while selecting materials for sensors, the results from 200°C onwards is reported here.
4.3.1 W2 Films Gas flow is switched on for 100 seconds manually, and allowed to recover to original resistance after the flow is shut off. The gas concentration is varied from 50 ppm to 500 ppm.
4.3.1.1 Sensing Temperature 200°C The response of W2 films to NO 2 and NH 3 at 200°C is shown in Figure 4.8 and Figure 4.9. The response time is 15 seconds and the recovery time is less than one minute for NO 2 . Maximum sensitivity of 4704 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.771. Hence, the response to ammonia is negligible in comparison to nitrogen dioxide. The resistance increase is more than three orders in magnitude for NO 2 as against less than l/3rd towards NH 3 . Hence by stabilizing a phase at higher temperature, sensing response is achieved even at low temperatures. W2-20G-NO* 140.00 400ppm
/
100.00
300ppm
/ } /
200ppm 40.00 A SOpnrn
.
.
.
A
100» m
.
.
n 10
1
12
.. 14
, 16
L 18
_ 20
V ... 22
24
L 26
28
80
.... 32
34
36
Time (min)
Figure 4.8 Response of W2 films to NO 2 at 200°C (50 ppm to 400 ppm)
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W2-200-NH3
0.18 0.18 0.14
1
^ 0.12 010
/
1"
/!
I't / i /
50 ppm 0.08
V
0.06
/
/
/ .............
V
” ”
004
/
f ........... 1
1
/
/
/
/
I
/
I
/
\
1 I / 40f^ Brft
\ V 300 ppm
200 ppm
/
/ /
X
/
j s, 500 ppm \
0.02 0.00 6
8
10
12
14
16
18
20
22
24
26
28
80
32
34
36
38
40
42
T im e (m in)
Figure 4.9 Response of W2 films to NH 3 at 200°C (50 ppm to 500 ppm)
4.3.1.2 Sensing Temperature 300°C The response of W2 films to NO2 and NH3 at 300°C is shown in Figure 4.10 and Figure 4.11. The response time is 15 seconds and the recovery time is less than one minute for NO 2 . Maximum sensitivity of 2026 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.626. Hence, the response to ammonia is negligible in comparison to nitrogen dioxide and has reduced from the response at 200°C. The resistance increase is more than three orders in magnitude for NO 2 as against less than l/4th towards NH3. At this temperature, the response to NO2 is still considerably large and the ammonia response is suppressed. This represents a suitable condition for selective NO2 detection in the presence of NH3.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W2*30G*NOi .00
500 ppm .00
70.00 60.00 50.00 ppm ,2 40.00
30.00 300 ppm 20.00
200 ppm 10.00
50 ppm
0.00
0
4
2
6
8
10
12
14
16 18 Time fmln)
20
22
24
26
28
30
32
Figure 4.10 Response of W2 films to NO 2 at 300°C (50 ppm to 500 ppm)
W 2-300-N H , 0,30
0.25
0.20
i0 ppm
g 0,15
100 ppm 200 ppm
0.10
400 ppm
300 ppm
500 ppm 0.05
0,00 0
2
4
8
8
10
12
14
16
18
20
22
24
26
28
30
32
34
T im e (min)
Figure 4.11 Response of W2 films to NH 3 at 300°C (50 ppm to 500 ppm)
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3.1.3 Sensing Temperature 400°C The response of W2 films to NO 2 and NH 3 at 400°C is shown in Figure 4.12 and Figure 4.13. The response time is 20-25 seconds and the recovery time is slightly longer than the previous two cases and is around two to three minutes for NO 2 . Maximum sensitivity of 518 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.93. The decrease in NO 2 sensitivity may be attributed to the phase transformation occurring at this temperature from monoclinic to orthorhombic phase. Though the response to ammonia is negligible in comparison to nitrogen dioxide it is higher than at 300°C. The resistance increase is more than three orders in magnitude for NO 2 as against less than l/4th towards NH 3 . At this temperature, the response to NO 2 is still considerably large but the recovery times are high. Hence this does not represent a suitable condition for selective NO 2 detection in the presence of NH 3 . W 2-400-N 02
120.00
400 ppm
100.00
300 ppm
0
2
4
6
8
10
12
14
16
18 20 22 24 Time (min)
26
28 30
32
34
36
38 40
42
Figure 4.12 Response of W2 films to NO 2 at 400°C (50 ppm to 400 ppm)
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W240CMM, 0.20
018
0.14 0.12
0.06 0.04
50 ppm 100 ppm
200 ppm
:300 ppm
0.00 0
2
4
6
8
10
12
14
16 18 Time {min)
20
22
24
26
28
30
32
34
Figure 4.13 Response of W2 films to NH 3 at 400°C (50 ppm to 400 ppm)
4.3.1.4 Sensing Temperature 500°C The response of W2 films to NO 2 and NH 3 at 500°C is shown in Figure 4.14 and Figure 4.15. The response time is 20-25 seconds and the recovery time is around three to four minutes for N 0 2. Maximum sensitivity of 6.13 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.947. As is observed, the sensitivity for the N 0 2 is the lowest at this temperature as compared to other sensing temperatures investigated above. The decrease in N 0 2 sensitivity may be attributed to the complete phase transformation occurring at this temperature from monoclinic to orthorhombic phase. Hence, formation of new grains of orthorhombic phase may be the rate limiting step. The response to ammonia is negligible in comparison to nitrogen dioxide however it is the maximum of all the sensing temperatures discussed above. The
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
recovery times are also considerably large both towards NO 2 as well for NH 3 . Hence this does not represent a suitable condition for selective NO 2 detection or for NH 3 . WZ-SOO-NOj
5.00
200 ppm
300 ppm
400 ppm
4.50 100 ppm 3.5.0
S
300
50 ppm
1 2.50
1.50
0.50 000 0
2
6
4
8
10
12
14
16
!8
20
22
24
26
28
30
Tim# (min)
Figure 4.14 Response of W2 films to NO 2 at 500°C (50 ppm to 400 ppm) W2"600-NHj
0.045 0.040 0.035 0.030 0.025 5 0020 0.015 0,010
50 ppm 0.005
O >00 ppm
100 ppm 0
2
4
6
8
10
12
14
16
400 ppm
300 ppm
0.000 18
20
22
24
26
28
30
32
34
Time (min)
Figure 4.15 Response of W2 films to NH 3 at 500°C (50 ppm to 400 ppm)
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W2 Sensing Temperature Dependence on N02 Sensitivity 5000 4500
3500 <
3000
i
2500
—♦ —200 deg C —SB—300 deg C -- * -4 0 0 deg C ~*r- 500 deg C
« 2000 1500
Ml
1000 500
50
100
200
300
400
500
Concentration (ppm)
Figure 4.16 Comparison of Sensitivities of W2 films towards NO 2 at various sensing temperatures W2 Sensing Temperature Dependence on NHj Sensitivity
0.9
0.7
-* -2 0 0
deg C
•^r
deg C
0 0 10
d eg C
i *
0 0
1
•
0.4
degC
0 0 <0
0.5
0.3
0.2
50
100
300
200
400
500
Concentration (ppm)
Figure 4.17 Comparison of Sensitivities of W2 films towards NH 3 at various sensing temperatures
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figures 4.16 and 4.17 show the comparison of sensitivities of W2 films towards NO 2 and NH 3 respectively for various sensing temperatures (200°C, 300°C, 400°C and 500°C). From the graphs, the optimum conditions for specific NO 2 detection is concluded to be best in the range 200°C to 300°C where the response to ammonia is lowest and the response to NO 2 is fast with good recovery times. Monoclinic phase of WO 3 acts as a suitable material for low temperature NO 2 sensing.
4.3.2 W3 Films 4.3.2.1 Sensing Temperature 200°C The response of W3 films to NO 2 and NH 3 at 200°C is shown in Figure 4.18 and Figure 4.19. The response time is 20 seconds and the recovery time two to three minutes for NO 2 . Maximum sensitivity of 1328 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.883. The resistance increase is more than three orders in magnitude for NO 2 as against less than 1/10th towards NH 3 . The maximum sensitivities at this temperature are lower than that obtained with W2 films. W3-200-N0*
400 ppm 300 ppm
j
/ 200 ppm h T /
............................................... i
100 ppm /
50 ppm
(I. 0
4
8
12
11. L,. I L 16
20
24
26
32
39
40
44
48
52
56
60
T im s (min)
Figure 4.18 Response of W3 films to NO 2 at 200°C (50 ppm to 400 ppm)
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In addition to this, the recovery times are slower than the W2 films indicating that orthorhombic films desorb NO 2 slower than monoclinic films. W3-200-NH,
0.06
0.05
0.04
/
S 0.03
0.02
/ »Y
\ J 100 ppm
f
/
/
50 ppm 0.01
/
/
/ u
200 ppm
/
300 ppm
/
A
/
j
1
/ /
u
^ ~ 's0 Q p p m
400 ppm
0.00
6
8
10
12
14 16 18 Time (min)
20
22
24
26
28
30
Figure 4.19 Response of W3 films to NH 3 at 200°C (50 ppm to 500 ppm)
4.3.2.2 Sensing Temperature 300°C The response of W3 films to NO 2 and NH 3 at 300°C is shown in Figure 4.20 and Figure 4.21. The response time is 15 seconds and the recovery time is less than one minute for NO 2 . Maximum sensitivity of 588 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.74. The resistance increase is more than three orders in magnitude for NO 2 as against less than l/3rd towards NH 3 . At this temperature, the response to NO 2 is still considerably reduced and the ammonia response is higher than that obtained with monoclinic films. But the recovery times are faster than at 200°C.
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W3-300-NQ, 70-00
500 ppm 60.00
50-f 400 ppm £ . 40,00 ppm
« 30.00
200 ppm 100 ppm 10.00
0.00
50 ppm -
0
2
4
6
8
10
$
14
12
18
1
20
Time (min)
Figure 4.20 Response of W3 films to NO 2 at 300°C (50 ppm to 500 ppm)
W3-30Q-NHj
0,05 0.04 0.04 0 03 f - 0.03
I
50 ppm
0-02 S ~J
100 ppm 0.02
200 ppm 0.01
300 ppm
400 ppm
500 ppm
0.01 0.00 0
2
4
6
8
12
10
14
18
18
20
Time (min)
Figure 4.21 Response of W3 films to NH 3 at 300°C (50 ppm to 500 ppm)
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3.2.3 Sensing Temperature 400°C The response of W3 films to NO 2 and NH 3 at 400°C is shown in Figure 4.22 and Figure 4.23. The response time is 30-35 seconds and the recovery time is significantly long for both NO 2 as well as NH 3 . The recovery time at 200 ppm NO 2 was close to 20 minutes. Maximum sensitivity of 1427 (arbitrary units) is obtained at 200 ppm. For ammonia, the maximum sensitivity is around 0.911. Though the sensitivity is high, the high recovery time makes it practically difficult for use as a sensor. Flence this does not represent a suitable condition for selective NO 2 detection in the presence of NH 3 . The irregular shape and spikes in the sensing plots are due to artifacts created by instrument range switching and flow controllers. W3 4 0 0 -NO2 2u0ppm
100 ppm
50 ppm
/
Tim* (min)
Figure 4.22 Response of W3 films to NO 2 at 400°C (50 ppm to 200 ppm)
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W3-400-NH, 0018
0.014
0.002 0,000 -I 0
1------- 1------- 1------- 1------- !------- 1------- ■-------1------- 1------- 1------- 1------- 1-------1------- 1------- 1------- r4 8 12 16 20 24 28 32 Tims (min)
Figure 4.23 Response of W3 films to NH 3 at 400°C (50 ppm to 400 ppm)
4.3.2.4 Sensing Temperature 500°C The response of W3 films to NO 2 and NH 3 at 500°C is shown in Figure 4.24 and Figure 4.25. The response time is 20-25 seconds and the recovery time is around 30-60 minutes for NO 2 . Maximum sensitivity of 164 (arbitrary units) is obtained at 500 ppm. For ammonia, the maximum sensitivity is around 0.941. As is observed, the sensitivity for the NO 2 is the lowest at this temperature as compared to other sensing temperatures investigated above. The response to ammonia highest of all the sensing temperatures discussed above. The recovery times are also considerably large both towards NO 2 as well for NH 3 . Flence this does not represent a suitable condition for selective NO 2 detection or for NH3. Again, the spikes are caused by artifacts arising from instrument switching range.
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W3-50Q.NO,
ppm
400 ppm 2.50 300 p p m
300 »Pm
200 ppm 2,00
O c
a
1
oo
100 ppn
spin 0.50
000
90 100 110 120 130 140
160 1T0 180 190 200 210
Time (min)
Figure 4.24 Response of W3 films to NO 2 at 500°C (50 ppm to 500 ppm)
W3-500.no , 3.00
S0O ppm 400 ppm
2.S0 200 ppm 2.00
o c w & 2w
lOOppri 1.00
50 spm 0.50
0.00 0
10
20
30
40
50
80
70
80
90 100 110 120 130 140 150 160 170 180 190 200 210 Time (min)
Figure 4.25 Response of W3 films to NH 3 at 500°C (50 ppm to 500 ppm)
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W3 S e n sin g T e m p e ra tu re D e p e n d e n c e o n N 0 2 Sensitivity
1600
1400
1200
- e - 200 deg C - a - 300 deg c
800
400 deg C ~*r- 500 deg c
600
400
.4B200
50
100
300
200
400
500
Concentration (ppm)
Figure 4.26 Comparison of Sensitivities of W3 films towards NO 2 at various sensing temperatures W3 Sensing Temperature Dependence on NH} Sensitivity
0.9
0.7
—* - 200 deg C 300 deg C - * - 4 0 0 d egC
0.5
- x - 500 deg C
« 0.4 0.3 0.2
50
100
200
300
400
500
Concentration (ppm)
Figure 4.27 Comparison of Sensitivities of W3 films towards NH3 at various sensing temperatures
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figures 4.26 and 4.27 show the comparison of sensitivities of W3 films towards N 0 2 and NH 3 respectively for various sensing temperatures (200°C, 300°C, 400°C and 500°C). From the graphs, the optimum conditions for specific N 0 2 detection is concluded to be best near 300°C where the response to ammonia is lowest and the response to N 0 2 is fast with good recovery times. However the high recovery times for all sensing temperatures make the orthorhombic phase of WO3 phase an unsuitable material for low temperature specific N 0 2 sensing.
4.4 Conclusions Novel WO3 thin films have been prepared and investigated as gas sensors. Both films (monoclinic and orthorhombic phases) showed selective N 0 2 detection over NH 3 . Monoclinic phase is suitable for low temperature (200°C-300°C) N 0 2 detection due to higher sensitivity and faster response and recovery times whereas films of orthorhombic phase had longer recovery times. The mechanism of sensing and comparison with other techniques is discussed in Chapter 6 .
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
References: 1.
P.J. Shaver, Activated Tungsten Oxide Gas Detectors. Applied Physics Letters, 1967. 11(8): p. 255-&.
2.
M. Akiyama, J. Tamaki, N. Miura, and N. Yamazoe, Tungsten Oxide-Based Semiconductor Sensor Highly Sensitive to NO and NO 2 . Chemistry Letters, 1991(9): p. 1611-1614.
3.
P. Nelli, L.E. Depero, M. Ferroni, S. Groppelli, V. Guidi, F. Ronconi, L. Sangaletti, and G. Sberveglieri, Sub-ppm NO 2 sensors based on nanosized thin film s o f titanium-tungsten oxides. Sensors and Actuators B-Chemical, 1996. 31(12): p. 89-92.
4.
M. Penza, M.A. Tagliente, L. Mirenghi, C. Gerardi, C. Martucci, and G. Cassano, Tungsten trioxide (WO3) sputtered thin films fo r a NOx gas sensor. Sensors and Actuators B-Chemical, 1998. 50(1): p. 9-18.
5.
M. Penza, C. Martucci, and G. Cassano, NOx gas sensing characteristics o f WO3 thin film s activated by noble metals (Pd, Pt, Au) layers. Sensors and Actuators BChemical, 1998. 50(1): p. 52-59.
6
.
W.Y. Zhang, H. Uchida, T. Katsube, T. Nakatsubo, and Y. Nishioka, A novel semiconductor NO gas sensor operating at room temperature. Sensors and Actuators B-Chemical, 1998. 49(1-2): p. 58-62.
7.
S.C. Moulzolf, S.A. Ding, and R.J. Lad, Stoichiometry and microstructure effects on tungsten oxide chemiresistive films. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 375-382.
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
.
B.T. Marquis and J.F. Vetelino, A semiconducting metal oxide sensor array for the detection ofN O x and NH 3 . Sensors and Actuators B-Chemical, 2001. 77(1-2): p.
9.
100
-1 1 0 .
K. Aguir, C. Lemire, and D.B.B. Lollman, Electrical properties o f reactively sputtered WO3 thin film s as ozone gas sensor. Sensors and Actuators B-Chemical, 2002. 84(1): p. 1-5.
10.
X.L. He, J.P. Li, X.G. Gao, and L. Wang, N 0 2 sensing characteristics o f WO3 thin film microgas sensor. Sensors and Actuators B-Chemical, 2003. 93(1-3): p. 463-467.
11.
A.K. Prasad and P.I. Gouma, M 0 O3 and WO3 based thin film conductimetric sensors for automotive applications. Journal of Materials Science, 2003. 38(21): p. 4347-4352.
12.
C.I. Reneker DH, Nanometre diameter fibres o f polymer, produced by electrospinning. Nanotechnology, 1996. 7(3): p. 216-223.
13.
L.G. Teoh, Y.M. Hon, J. Shieh, W.H. Lai, and M.H. Hon, Sensitivity properties o f a novel NO 2 gas sensor based on mesoporous WO3 thin film. Sensors and Actuators B-Chemical, 2003. 96(1-2): p. 219-225.
14.
M.S. Tong, G.R. Dai, and D.S. Gao, WO3 thin film sensor prepared by sol-gel technique and its low-temperature sensing properties to trimethylamine. Materials Chemistry and Physics, 2001. 69(1-3): p. 176-179.
15.
K. Galatsis, Y.X. Li, W. Wlodarski, and K. Kalantar-zadeh, Sol-gel prepared M 0 O3 -WO 3 thin-films fo r O2 gas sensing. Sensors and Actuators B-Chemical, 2001. 77(1-2): p. 478-483.
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16.
Y.G. Choi, G. Sakai, K. Shimanoe, N. Miura, and N. Yamazoe, Wet processprepared thick film s o f WO3 for NO 2 sensing. Sensors and Actuators B-Chemical, 2003. 95(1-3): p. 258-265.
17.
C. Balazsi, M. Farkas-Jahnlce, I. Kotsis, L. Petras, and J. Pfeifer, The observation o f cubic tungsten trioxide at high-temperature dehydration o f tungstic acid hydrate. Solid State Ionics, 2001.141: p. 411-416.
18.
C. Balazsi and J. Pfeifer, Structure and morphology changes caused by wash treatment o f tungstic acid precipitates. Solid State Ionics, 1999. 124(1-2): p. 7381.
19.
C. Balazsi and J. Pfeifer, Development o f tungsten oxide hydrate phases during precipitation, room temperature ripening and hydrothermal treatment. Solid State Ionics, 2002. 151(1-4): p. 353-358.
20.
A.K. Prasad, Processing and Microstructural Effects on the Gas Sensing Properties o f M 0 O3 and WO3 Thin Films, M.S. Thesis in Materials Science and Engineering. 2002, SUNY at Stony Brook: Stony Brook, p. 82.
21.
E. Salje and K. Viswanathan, Physical-Proper ties and Phase-Transitions in WO3 . Acta Crystallographica Section A, 1975. A 31(MAY1): p. 356-359.
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 5 Sensor Arrays for Detection of Isoprene, Methanol and Carbon Dioxide This section focuses on the development of sensor arrays based on M0 O3 thin films. The various gases detected are methanol, isoprene and carbon dioxide. Conditions suitable for selective detection of each of the gases are discussed. A brief introduction of the gases to be detected is also given below.
5.1 Methanol Detection Methanol (CH3 OH) is a colorless liquid with B.P. 65°C. It is a potent nerve poison and is also shown to be a carcinogenic agent [1]. Breath analysis is employed to test the variations in metabolic pathways of alcohol breakdown. Detection of methanol in breath may give information in patient’s pulmonary health. Research on methanol sensors using semiconducting oxides is very limited [2 - 1 0 ] hence the reaction mechanism of methanol with semiconducting oxides is not known clearly. An attempt is made to propose a suitable methanol sensing mechanism based on catalysis literature. The work presented in this chapter will reveal the conditions under which selective methanol detection is achieved. A possible reaction mechanism for methanol detection by M0 O3 thin films is given in Chapter 6 .
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5.2 Isoprene Detection Isoprene (2-methyl - 1,3-butadiene) is a reactive aliphatic hydrocarbon. It is a colorless liquid, B.P. 35°C (101.325 kPa). It is synthesized by nearly all animals and is present among the hydrocarbon metabolites in human breath. Isoprene originates from the decomposition of di-methylallyldiphosphate, a member of cholesterol and isoprenoids synthetic pathway [11]. In other words, the amount of isoprene in expired breath is an indirect measure o f amount of cholesterol synthesized and its concentration corresponds to the activity of the enzyme producing cholesterol in our body. Cholesterol is synthesized mainly at night due to high demands on energy supply. Hence isoprene concentration measurement should aid in monitoring the patients suffering from hypercholesterolemia (disorder in cholesterol mechanism), which is a risk factor for atherosclerosis development. This necessitates the need for an isoprene detector which can be used to monitor breath isoprene concentrations. The work reported here presents the conditions suitable for selective isoprene detection in presence of other interfering gases such as carbon monoxide and ammonia.
5.3 Experimental Thin films of M 0 O3 were prepared by sol-gel technique as explained in Chapter 2. These films were desposited on 3 x 3 mm alumina substrates with platinum interdigitated electrodes on one side and platinum heaters on the reverse side. The SEM image of the substrates is as shown in Figure 5.1.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 5.1 (a) SEM image of inter-digitated electrodes (b) SEM image of heaters The width of the electrodes as well as the heaters is approximately 200 pm. The contacts were spot-welded onto the electrodes. The thickness of the films deposited on the substrates was controlled by the amount of sol-gel dropped. Two sets of films of different thicknesses (denoted in the graphs by 1L (1 layer-150 nm) and 2L (2 layers-250 nm)) were prepared and allowed to dry before heat treating them in air at 500°C for 10 hours. The heat treatment was done in multisensor testing gas flow bench setup (shown in Figure 5.2) prior to starting the sensing experiments. The gases used for sensing include methanol, isoprene, ammonia, nitrogen dioxide, carbon monoxide and carbon dioxide. These gases were flown individually and in mixtures and the flow rates were computer controlled. The change in current upon exposure to gas is measured over a constant applied voltage of 1 V. The resistance is then calculated and plotted over time.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CH
Figure 5.2 Multisensor gas sensing setup (P rof Sberveglieri 's Lab, Italy used in these experiments)
5.4 Results The sensing experiments were carried out three different temperatures 400°C, 420°C and 450°C. Figure 5.3 shows the sensor output plotted as resistance and concentration as a function of time. Gases are introduced in 10 min square pulses. The response of sensors of two different thicknesses (MO-1 - 250 nm, MO-2 - 150 nm) towards ammonia, methanol and CO are shown. As is seen from the sensitivities, the M 0 O3 film is highly sensitive to methanol at 400°C. The effect of thickness is negligible with increasing thickness barely affecting the sensitivity. As seen from Figure 5.4, the effect of ammonia in combination with methanol is almost the same as the response to methanol alone indicating that methanol sensitivity is the dominating factor. It is concluded that M 0 O3 thin film behaves as a good methanol detector at 400°C.
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MO (1 L.2L) 450
1200 T
400 1000
350
800 ••
-3 0 0
"&s' a 250 T
V
£.
1 200
|
L> 400
- M02-R -M01-R -Ammonia -Methanol 0050
150
••100 200
-
50
500
600
700
Time (min)
Figure 5.3 Sensing responses of two sets of M 0 O3 thin films towards ammonia (15 ppm), methanol (400 ppm) and CO (15 ppm) at 400°C.
Table 5.1 Response of sensor arrays of M 0 O3 to square pulse of methanol (400 ppm), carbon monoxide (15 ppm) and ammonia (15 ppm) at 400°C. Response time
Recovery time
(min)
(min)
MOl
0.61
1 0 .1
0.75
M 02
0.67
12
0.67
Carbon
MOl
0.27
20
0 .1 2
Monoxide
M 02
0.3
19
0.25
MOl
0.35
14
0.13
M 02
0 .6
17
0.25
Gas
Sensor
Sensitivity (S)
Methanol
Ammonia
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M o03 1L vs 2L 400 d e g C
SM0O3-2L Mo03-1L
J J Figure 5.4 Comparison o f sensitivities of M0 O3 thin films towards various gases/gas mixtures at 400°C MoOj (1L vs 2L) at 420 deg C 30 0
500
250 -•4 0 0
200 -•3 0 0
a s
£
/
M02-R M01-R
200 S
Isoprene
100
100 6 0 ••
700
800
90 0
1000
Time (mini
Figure 5.5 Sensing responses o f two sets of M0 O3 thin films towards ammonia (15 ppm), NO2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 420°C.
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 5.2 Response of sensor arrays of M 0 O3 to square pulse of ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 420°C. Response time Gas
Recovery time
Sensor
Sensitivity (S) (min)
(min)
MOl
0.9
7.6
0.36
M 02
0.9
8
0.35
MOl
0.27
20
0.25
M 02
0.28
19
0 .2 2
MOl
0.4
14
0.3
M 02
0 .6
15
0.26
Isoprene
Carbon Dioxide
Ammonia
Figure 5.5 shows the sensing response at 420°C towards various gases/gas mixtures namely isoprene, carbon dioxide, ammonia and nitrogen dioxide. The reason for irregular response observed in case of ammonia and CO2 is not known and is probably due to instrumental artifacts. The sensitivities, response times and recovery times are tabulated in Table 5.2. As is observed from the results, the response of the films towards isoprene is higher than that towards ammonia and CO2 . The response towards methanol is negligible, hence is not shown. As the film is heated to 420°C, the monoclinic phase of M 0 O 3 begins to transform to orthorhombic phase, though the transformation is still not complete. As reported in Chapter 3, the orthorhombic form favors the response towards ammonia which is seen by the increase in response at 420°C than 400°C wherein the film is predominantly monoclinic phase.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The film is not sensitive to NO 2 and the presence o f NO 2 along with NH 3 in the
gas stream tends to reduce the sensitivity of the film. This is as expected since NO 2 reacts with NH 3 at high temperatures, thereby reducing the concentration of NH 3 which leads to a reduced response from the sensor. Mo031LVs 2L 420 deg C 0.4 -1---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Isoprene
NH3-C02
N02-co2
co2
no2
no2-nh3
nh3
Figure 5.6 Comparison of sensitivities of M 0 O3 thin films towards various gases/gas mixtures at 420°C
The sensitivity towards isoprene is highest at 420°C wherein the phase distribution in the film consists of higher monoclinic phase than orthorhombic phase. Hence it can be concluded that at this temperature, the film acts as a good isoprene sensor.
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
MO (1L vs2L) at 450 degC 100
400
■•300
sa -•200 1
-•
100
& 200
240
280
320
380
400
440
480
520
580
T im e {min)
Figure 5.7 Sensing responses of two sets of M 0 O3 thin films towards ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 450°C. Figure 5.7 shows the response of the sensor at 450°C. The response is similar to that at 420°C except for a reverse in specificity towards CO2 than isoprene. Table 5.3 Response of sensor arrays of M 0 O3 to square pulse of ammonia (15 ppm), NO 2 (500 ppm), isoprene (10 ppm) and CO2 (5 ppm) at 450°C.
Gas
Response time
Recovery time
(min)
(min)
MOl
1.9
7.6
0 .2 0
M 02
0.92
5.6
0 .1 2
MOl
0.27
19
0.375
M 02
0.25
17
0.35
MOl
1.5
16.4
0 .1
M 02
1.4
15
0.114
Sensor
Sensitivity (S)
Isoprene
Carbon Dioxide
Ammonia
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
M o03-1L Vs 2L 450 deg C 0.4
Iso p re n e
N H 3 -C 0 2
NO2-C02
co2
no2
no2-nh3
nh3
Figure 5.8 Comparison of sensitivities of M0 O3 thin films towards various gases/gas mixtures at 450°C At 450°C, the response of the sensor is highest to CO2 and the response to isoprene and ammonia is low (as seen from sensitivities in Table 5.3). The response to NO 2 is negligible. At this temperature, there is still a mixture of monoclinic and orthorhombic phases existing in the sensing film. The presence of monoclinic phase hinders the response to ammonia. The response to methanol is also negligible (not shown) indicating that only pure monoclinic phase favors the response towards methanol. It can be concluded that at 450°C wherein the phase distribution of the film contains higher percentage of orthorhombic phase than monoclinic phase, the sensing film behaves as a good carbon dioxide sensor.
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The phase distribution of monoclinic and orthorhombic phases in films heat treated at 420°C and 450°C is obtained by quantitative XRD.
420 Vs 450 1000 v
900
800
700
600
£
500
400
300
200
100
U) CD ^ CN CN
cn
on
is .
CN CN PI 2
is .
(I (Angstroms)
Figure 5.9 Comparison of XRD spectra of films heat treated at 420°C and 450°C (SSubstrate peaks) Figure 5.9 shows the comparison of the XRD spectra of the films heat treated at 420°C and 450°C. Prominent substrate peaks are marked “S”. The (002) peaks of monoclinic and orthorhombic phases are marked through dotted lines.
Table 5.4 Percentage distribution of each phase in films heat treated at 420°C and 450°C Phases
420°C
450°C
Monoclinic (%)
72.15
23.2
Orthorhombic (%)
27.85
76.8
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
By measuring the intensity of the peaks in each film, the percentage composition of each phase in two sets of films is calculated. The result is tabulated as above. (Table 5.4)
5.5 Conclusions By altering the phase composition of the single oxide film, namely M 0 O3 , through stabilization heat treatments differing sensing responses are obtained. The phase stability domains are identified and suitable sensor arrays for selective detection of target gases can be constructed. Table 5.5 shows the prototype of a selective sensor array based on the above results. Table 5.5 Selective Sensor Array prototype Selective Sensor Arrays Sensor
Methanol
Isoprene
Carbon Ammonia Sensor
Sensor
Sensor
Dioxide Sensor
Monoclinic M 0 O3
75% Monoclinic 25% Orthorhombic M 0 O3
25% Monoclinic 75% Orthorhombic M 0 O3
Orthorhombic M 0 O3
400°C
420°C
450°C
500°C
Phase
Operating Temperature
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reference: 1.
P. Lirk, F. Bodrogi, and J. Rieder, Medical Applications o f proton transfer reaction-mass spectrometry: ambient air monitoring and breath analysis. International Journal of Mass Spectrometry, 2004. 239: p. 221-226.
2.
X.L. Cheng, H. Zhao, L.H. Huo, S. Gao, and J.G. Zhao, ZnO nanoparticulate thin film: preparation, characterization and gas-sensing property. Sensors and Actuators B-Chemical, 2004.102: p. 248-252.
3.
A.K. Srivastava, Detection o f volatile organic compounds (VOCsj using SnC>2 gas-sensor array and artificial neural network. Sensors and Actuators BChemical, 2003. 96: p. 24-37.
4.
N.G. Patel, P.D. Patel, and V.S. Vaishnav, Indium tin oxide (ITO) thin film gas sensor fo r detection o f methanol at room temperature. Sensors and Actuators BChemical, 2003. 96: p. 180-189.
5.
A. Salehi, Preparation and Characterization o f proton implanted indium tin oxide selective gas sensors. Sensors and Actuators B-Chemical, 2003. 94: p. 184-188.
6
.
C. Garzella, E. Bontempi, L.E. Depero, A. Vomiero, G. Della Mea, and G. Sberveglieri, Novel selective ethanol sensors:W/Ti0 2 thin films by sol-gel spincoating. Sensors and Actuators B-Chemical, 2003. 93: p. 495-502.
7.
R.E.
Cavicchi,
R.M.
Walton,
M.
Aquino-Class,
J.D.
Allen,
and
B.
Panchapakesan, Spin-on nanoparticle tin oxide for microhotplate gas sensors. Sensors and Actuators B-Chemical, 2001. 77: p. 145-154. 8
.
M. Mabrook and P. Hawkins, A rapidly-responding sensor fo r benzene, methanol and ethanol vapours based on films o f titanium dioxide dispersed in a polymer
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
operating at room temperature. Sensors and Actuators B-Chemical, 2001. 75: p. 197-202. 9.
C. Garzella, E. Comini, E. Tempesti, C. Frigeri, and G. Sberveglieri, HO 2 thin film s by a novel sol-gel processing fo r gas sensor applications. Sensors and Actuators B-Chemical, 2000.
10.
6 8
: p. 189-196.
A. Teeramongkonrasmee and M. Sriyudthsak, Methanol and ammonia sensing characteristic o f sol-gel derived thin film gas sensor. Sensors and Actuators BChemical, 2000.
11.
6 6
: p. 256-259.
R. Hyspler, S. Crhova, J. Gasparic, Z. Zadak, M. Cizkova, and V. Balasova, Determination o f isoprene in human expired breath using solid-phase microextraction
and gas chromatography-mass spectrometry.
Journal
of
Chromatography B-Analytical Technologies in the Biomedical and Life Sciences, 2000. 739(1): p. 183-190.
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 6 Discussion 6.1 Processing Issues - Comparing responses from different processing routes The various processing techniques used for comparison include ion beam deposition, sol-gel, and acidic precipitation. The films were heat treated at 500°C for
8
hours and the sensing temperature was maintained at 450°C. The sensitivities of the films are compared for ammonia and nitrogen dioxide.
6.1.1 M 0 O 3 thin films M 0 O3 thin films prepared by ion beam deposition and sol-gel methods are compared based on the sensitivity towards ammonia and NO 2 . It has been shown that M 0 O 3 thin films operated in the temperature range of 450°C-500°C is specific to ammonia [ 1 , 2 ]. Table 6.1 Comparison of Sensitivities of M 0 O3 thin film prepared by sol-gel and ionbeam deposition to gases at 450°C Gas
Sol-gel
Ion Beam Deposition [1]
Ammonia
0.706
0.989
Nitrogen dioxide
0.024
0.35
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Processing route affects the sensing behavior due to the differences in morphology obtained through each method. The sensitivity is higher in the case of ion beam deposited thin films due to continuous film morphology; however, both processing routes yield films which are specific to ammonia. This shows that specificity can be obtained irrespective of the processing method chosen.
6
.1.2 WO 3 thin films Tungsten oxide thin films prepared by sol-gel, IBD and acidic precipitation are
compared based on the sensitivity towards ammonia (500 ppm) and nitrogen dioxide (500 ppm). Table 6.2 Comparison of Sensitivities of WO3 thin films prepared by various methods Sol-gel [3]
IBD [4]
Acidic Precipitation*
Nitrogen
0.124
3.61
4704
0 .0 1
0.065
0.947
Dioxide Ammonia
* The concentration o f NO 2 and NH 3 was 400 ppm for this study As observed from the above table, WO 3 thin films prepared by all the three methods are more specific to NO 2 than NH 3 though the sensitivity levels are vastly different. This is due to difference in resistance of the films prepared by each technique. In sol-gel technique (reported in our article [3]), the sensitivity towards NO 2 was an order of magnitude greater than towards ammonia. The sensitivity of films obtained through IBD technique towards NO 2 was 2 orders of magnitude more than that towards ammonia.
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The sensitivity of films obtained through acidic precipitation method towards NO 2 was more than 3 orders of magnitude than that towards NH 3 . This difference is due to the fact that in acidic precipitation method, the film was prepared directly from heat-stabilized pure oxide phase whereas in sol-gel and ion beam deposition the films were either made from precursor (sol-gel) or oxidized from pure metal (IBD). The sensitivity is directly proportional to the surface area of the film in contact with the reacting gas [5, 6 ]. Though the films prepared by sol-gel and ion beam deposition contain nano-sized grains [3, 4] (30 nm and 50 nm grains respectively), the films prepared from acidic precipitation (150 nm) were more continuous (Figure 6.1 a,b,c).
1 ^ "
* e s w o j m t# ST© NY s^ a * sk 2 — j_________ _______________ £HT« kV Tires:l
2Qm:
o a » : u May 2002
I
?fefW*«bSM^TCUif
______
STflNY
BKfMftK.
A»t«l»rre bate -14m ; 2002
£Ht» 6.00ay ran*-33:14,-sc
(c)
........
vfina A R ?..0 Set* 12fta>2 £H1 • ISIJSW T*me:2;27:tf
Figure 6.1 Comparison of SEM images of films obtained by (a) Sol-gel (b) Ion Beam Deposition and (c) Acidic Precipitation
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6.1 (a) shows the discrete islands in sol-gel thin films. Figure 6.1 (b) shows porous ion beam deposited films. Figure 6.1 (c) shows denser and more continuous films prepared from powders obtained from acidic precipitation. Flence the difference in sensitivity between films prepared from acid precipitation and those from the sol-gel and ion beam deposited films is higher (3 orders of magnitude as against less than 2 orders magnitude change).
6.2 Polymorphic Phase Selection - Key to achieving specificity in sensing The detection process of oxidizing/reducing gases by semiconducting metal oxides involves the change in conductivity of the oxide in the presence of the gas due to catalytic oxidation-reduction reactions occurring at the metal oxide- gas interface [7]. Moseley et al [7] proposed different models based on which conductivity changes occur and a generalization is obtained as is tabulated in Chapter 1 (Table 1.1). Though this generalization works with most semiconducting oxides, this model is not sufficient to explain the phenomenon of selectivity observed with certain metal oxides viz. M 0 O3 and W 0 3. The differences in the sensing response of M 0 O3 and WO3 as studied in this work suggest that crystallographic modifications of the single oxide phase are the key to achieving selectivity. Based on the above results, a simplified model of polymorphic phase selection process for specific gas sensing is explained as below. Semiconducting metal oxides such as M 0 O3 and WO 3 show structure sensitivity [8 ] which favors their use as catalysts in selective oxidation/reduction processes. M 0 O3
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for example undergoes reduction reaction in presence of hydrocarbons through simultaneous formation of oxygen vacancies and crystallographic shear (CS) planes[9]. The formation of oxygen vacancies might be represented by the following quasi-chemical reaction (6 . 1 ) as reported in literature [2 , 0 0* ~
>/2
10
]
0 2 (g) + V02+ + 2 e'
(6.1)
Where 0 0* represents an unstable oxygen atom in an oxygen site, V02+ represents an oxygen vacancy with double positive charge. When oxygen is incorporated into these vacancies, a reversible reaction (6.2) occurs as shown below [V02+] +
'/2
0 2 <-> 0 0 + 2h+
(6.2)
Reaction as represented by Equation (6.1) occurs due to increased oxygen mobility at elevated temperatures [11] or the presence of reducing atmospheres. The slightly reduced metal oxides thus formed may either undergo reoxidation through reaction represented by Equation (6.2) by gaseous oxygen or other oxidizing gases such as N 0 2 which is the mechanism for adsorption based sensing. On the other hand, the oxygen vacancies may rearrange to form CS structures which gives rise to reaction based sensing mechanism. CS planes in these oxides are part of crystals containing blocks of four edgeshared octahedral which occurs due to removal of oxygen from lattice during interaction with target gas species. The formation of CS planes in M 0 O3 and WO3 is discussed below. Molybdenum oxide is found to exist in 3 different phases -
a M 0 O3
(orthorhombic phase), P-M0 O 3 (monoclinic phase-metastable) and MoO ’ 3 (hexagonal-
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
metastable) [12].
-M 0 O3 crystallizes in a unique 2D layered structure (Figure 6.2 (b)).
01
They are built up of double chains of edge-sharing [MoOg] octahedral connected through vertices. The (3 phase and MoO ’ 3 phase are stable upto 400°C beyond which a phase begins to stabilize until 550°C. P phase (Figure 6.2 (a)) is more commonly observed than MoO ’ 3 phase. The P-phase is ReC>3 -type structure and does not contain the Van der Waals gap as in a-phase. It has been observed that M 0 O3 phase transformation from a to p phase is topotactic meaning that the transformation results in film displaying a strong preferred orientation from Okk p phase to OkO a phase [13]. The transformation is believed to be due to shearing of every other row of (0 1 1 ) octahedral of the p structure half a unit along the [Oil] monoclinic crystal direction. A simulation adapted from [13] is shown below, (a)
P-phase
(b)
a-phase
[010]
[O il]
Figure 6.2 Phase transformation from p to a phase in M 0 O3
Upon reduction, the a-MoC>3 forms the M 0 1 8 O 52 structure instead of the Re0 3 type M 0 8 O2 3 shear structure [9]. As observed from results in Chapter 3, orthorhombic M 0 O 3 (a-phase) is selective to ammonia and the sensing mechanism is consistent with
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the reduction of M 0 O3 and formation of M0 1 8 O52 shear structures. This suggests that ammonia detection with 01-M 0 O 3 is through reaction based mechanism. This is confirmed by XPS results[l] obtained with the films after sensing with ammonia and is shown in Figure 6.3.
These results and curve fitting were obtained from work done by P.J.
Schmitz, Ford SRL. It is evident from Fig 6.3 (a), for the case of lOOOppm NH 3 in 10% O2 , the appearance of a second set of peaks at binding energies 231.6eV (Mo 3 ds/2 ) and 234.7eV (Mo 3d3/2) that the surface of Mo is reduced to lower oxidation states of Mo+ 5 and Mo+4. This reduction of Mo is more pronounced in case of Figure 6.3 (c) wherein the accompanying oxygen level is reduced to 0.5%. Exposure to C3 H 6 caused negligible change in the Mo oxidation state.
1
binding energy (eV)
Figure 6.3 XPS spectra obtained from the same M 0 O3 film after exposure to different gas environments at ~465°C. (a) The Mo 3d spectra after exposure to 1000 ppm NH 3 in 10% O2 . (b) The Mo 3d spectra after exposure to 10% O2 only, (c) After exposure to 1000 ppm NH 3 in 0.5% 0 2. (d) After 1000 ppm C3 H 6 in 10% 0 2.
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The crystal structure of WO3 is a distortion of rhenium oxide cubic structure in which tungsten atoms are located in cube corners and the oxygen atoms are located on the cube edges[14]. Each tungsten atom is surrounded by six oxygen atoms forming an octahedron. The slight rotation of these octahedral with respect to each other, as well as unequal bond lengths in the octahedral coordination, causes lattice distortion and reduces the symmetry. The distorted structure is stable in several forms giving rise to different phases depending on the temperature. WO 3 exists in various polymorphic forms. There are at least 7 known polymorphic transformations between the temperature range of 0 to 1220K [12]. Triclinic phase is stable below 17°C. Monoclinic phase is stable between 17-320°C and orthorhombic is stable from 320-720°C above which tetragonal phase is the dominant phase.
Figure 6.4 Structure of monoclinic WO 3 (adapted from [14])
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 6.4 shows the crystal structure of the monoclinic WO3. Oxygen atoms are shown in black with the comer sharing W06 octahedra shaded. All of the polymorphs of WO3 are distorted Re0 3 type structured, meaning all are nearly of the above shape
(Figure 6.4). Stoichiometric WO3 readily loses oxygen to form WO 3 .X. When oxygen is removed from the lattice in sufficient amounts, some of the octahedral change their character from corner-sharing to edge-sharing. In slightly reduced WO3, shear planes are introduced in many directions at irregular intervals. But this shear planes doesn’t seem to affect the electrical properties or the phase transformation in WC>3[ 14 ]. Hence, NO2 sensitivity occurs by re-oxidation of oxygen vacancy and is dominated by adsorption based sensing mechanism which doesn’t affect the bonds in the metal oxide surface. The behavior of monoclinic and orthorhombic WO3 towards NO2 and NH3 is similar because they are iso-stmctural. However, WO3 which possesses distorted Re 0 3 structure favors adsorption of oxidizing gases such as NO 2 . Ammonia sensing is through reaction of lattice oxygen with gas and reduction of off-stoichiometric compounds through formation of CS structures in layered oxide structures such as those in (X-M0 O3 . It has been shown in Chapter 5 that through polymorphic phase modifications, variations in sensing response can be achieved. Let us consider the case of methanol, isoprene and carbon dioxide detection. As mentioned before, literature on the interaction of semiconducting oxides with methanol for use in sensors is limited. However, it is known methanol is an active reducing agent. This is indicated by the drop in resistance of the M 0 O3 films upon exposure to methanol. Private communications with co-workers suggests an adsorption
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
based mechanism for methanol sensing. Methanol abstracts oxygen atom from M 0 O3 lattice and adsorbs on the surface reducing the, resistance of the oxide. Formation of oxygen vacancies have been mentioned earlier for the case of both layered structure (aM 0 O3 ) and ReC>3 type structures (P-M0 O 3 and WO 3 ). The mechanism of detection of isoprene at 420°C, wherein the film consists of 25% orthorhombic phase-75% monoclinic phase is still unclear. It can however be proposed that the existence of two double bonds (2-methyl 1,3 buta-diene) provide for easy sharing of electrons with monoclinic M 0 O3 , and an adsorption based mechanism can be used to explain the isoprene sensing behavior. The mechanism of CO2 response can also be explained in a similar fashion to NO 2 adsorption, however, detailed studies are required for confirming these hypotheses. As the percentage of orthorhombic phase increases, the CO 2 response decreases, since the film become more prone towards reaction-based mechanism for detection of reducing gases. This also explains the increased ammonia response above 450°C.
6.3 Building Optimum Sensors - Identifying optimum temperature regimes and crystal structure for specificity Based on above discussion, the sensing behavior of two metal oxides under research, M 0 O3 and WO 3 can be classified according to their crystallographic characteristics [15] and their target species which they detect.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 6.3 Classification of oxides, gases they detectand sensing mechanism involved Oxide type
Examples
of Gas Detected
Sensing Mechanism
oxides 2d-Layered
(X-M0 O3
structure Re 0 3 -type
Ammonia,
Reaction based mechanism
amines P-M0 O3 , W 0 3
structure
N 0 2,
NO,
0 2, Adsorption based mechanism
CO2 , Isoprene
The optimum operating temperature range is chosen carefully such that it does not interfere with phase transformation domains and still exhibiting maximum sensitivity and specificity. A list of different operating temperatures, polymorphic phase selected and target gas is tabulated in Table 6.4.
Temperature Range
Sensor Material
Phase
Gas
200°C - 300°C
W 03
Monoclinic
N 02
400°C
M 0 O3
Monoclinic ((3)
Methanol
420°C
M 0 O3
75% p - 25% a
Isoprene
450°C
M 0 O3
25% a - 75% p
O O to
Table 6.4 Optimum Sensor Arrays- Temperature and Phase domains and Gases detected
>450°C - 500°C
M 0 O3
Orthorhombic (a)
Ammonia
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reference: 1.
A.K. Prasad, P.I. Gouma, D.J. Kubinski, J.H. Visser, R.E. Soltis, and P.J. Schmitz, Reactively sputtered M 0 O3 films fo r ammonia sensing. Thin Solid Films, 2003. 436(1): p. 46-51.
2.
A.K. Prasad, D. Kubinski, and P.I. Gouma, Comparison o f sol-gel and ion beam deposited M 0 O3 thin film gas sensors fo r selective ammonia detection. Sensors and Actuators B-Chemical, 2003. 93: p. 25-30.
3.
K.M. Sawicka, A.K. Prasad, and P.I. Gouma, Metal Oxide Nanowires fo r Use in Chemical Sensing Applications. Sensor Letters, 2005. 3(1): p. 31-35.
4.
A.K. Prasad and P.I. Gouma, M 0 O3 and WO3 based thin film conductimetric sensors fo r automotive applications. Journal of Materials Science, 2003. 38(21): p. 4347-4352.
5.
L.G. Teoh, Y.M. Hon, J. Shieh, W.H. Lai, and M.H. Hon, Sensitivity properties o f a novel NO2 gas sensor based on mesoporous WO3 thin film. Sensors and Actuators B-Chemical, 2003. 96(1-2): p. 219-225.
6
.
J. Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura, and N. Yamazoe, Grain-Size Effects in Tungsten Oxide-Based Sensor fo r NitrogenOxides. Journal of the Electrochemical Society, 1994. 141(8): p. 2207-2210.
7.
P.T. Moseley and A.J. Crocker, Sensor Materials. 1996, Bristol: Institute of Physics.
8
.
J. Haber, J. Janas, M. Schiavello, and R.J.D. Tilley, Tungsten-Oxides as Catalysts in Selective Oxidation. Journal of Catalysis, 1983. 82(2): p. 395-403.
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9.
T. Ressler, J. Wienold, R.E. Jentoft, and T. Neisius, Bulk structural investigation o f the reduction o f M 0 O3 with propene and the oxidation o f M 0 O2 with oxygen. Journal of Catalysis, 2002. 210(1): p. 67-83.
10.
H. Yamada and G.R. Miller, Point-Defects in Reduced Strontium-Titanate. Journal of Solid State Chemistry, 1973. 6(1): p. 169-177.
11.
M.A. Khilla, Z.M. Hanafi, B.S. Farag, and A. Abuelsaud, Transport-Properties o f Molybdenum Trioxide and Its Suboxides. Thermochimica Acta, 1982. 54(1-2): p. 35-45.
12.
E. Haro-Poniatowski, M. Jouanne, J.F. Morhange, C. Julien, R. Diamant, M. Fernandez-Guasti, G.A. Fuentes, and J.C. Alonso, Micro-Raman characterization o f WO3 and M 0 O3 thin films obtained by pulsed laser irradiation. Applied Surface Science, 1998.129: p. 674-678.
13.
P.F. Carcia and E.M. Mccarron, Synthesis and Properties o f Thin-Film Polymorphs o f Molybdenum Trioxide. Thin Solid Films, 1987.155(1): p. 53-63.
14.
L.J. LeGore, R.J. Lad, S.C. Moulzolf, J.F. Vetelino, B.G. Frederick, and E.A. Kenik, Defects and morphology o f tungsten trioxide thin films. Thin Solid Films, 2002. 406(1-2): p. 79-86.
15.
P.I. Gouma, Nanostructured Polymorphic Oxides for Advanced Chemosensors. Reviews on Advanced Materials Science, 2003. 5: p. 123-138.
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Chapter 7 Conclusions and Future Work 7.1 Summary of Conclusions This work has established the suitability of M 0 O3 and WO3 thin films as suitable materials for specific gas sensor arrays. M 0 O3 thin films fabricated through sol-gel and ion beam deposition possessed excellent gas sensing properties. The effect of processing, thickness, stabilization heat treatment temperature, and operating temperature on the gas sensing properties has been studied and established. It is found that thickness does not play an important role in gas sensitivity once a monolayer film has been formed. It is further concluded that the gas interactions with M 0 O3 are surface controlled rather than bulk controlled. Stabilization heat treatments carried out on the films revealed the phase transition regions. Thin film microstructure characterization through SEM, TEM and XRD revealed the microstructure at different stabilization regions. Studies on operating temperature effects on sensing behavior have helped in identifying the temperature range for specific ammonia sensing. Different processing routes (sol-gel and ion beam deposition) resulted in similar temperature range for optimum specific sensing for ammonia (450°C-500°C).
This shows that polymorph
selection is the key to specific gas detection.
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
WO3 thin films processed by a novel technique are investigated in this study. The acid precipitation route is adopted for producing pure single phase metal oxide. Spin coating combined with drop coating is employed to apply the film on to the sensor substrates. WO 3 thin films prepared by this route are tested as sensing material for the first time. Two different phases of the material (monoclinic and orthorhombic) are used for study. XRD studies are used to identify the phases before and after heat stabilization. Phase transformation from monoclinic to orthorhombic is found to occur in the temperature range of 350°C - 400°C. The morphology of the films are investigated by SEM technique. Both the films were consisted of nano-crystalline particles. Sensing tests were performed in the temperature range of 200°C - 500°C to identify the optimum temperature for specific sensing. It is found that the films prepared by both routes were specific to NO 2 in the presence of ammonia however the monoclinic phase was found to operate well in lower temperature range of 200°C-300°C possessing better sensitivity, specificity, and response and recovery times. The response is compared with results obtained from other techniques. The nature of response (specific to NO 2 ) is similar for all processing routes. M0 O3 thin films prepared by sol-gel technique are chosen for use in gas sensor arrays. Harmful hydrocarbons such as methanol, isoprene and emission gas CO2 is used as target gases. Specific temperature range is identified for specific detection of each of the gases. The polymorphic phase distribution is calculated for each case and a suitable sensor array is constructed. Study of crystal structure of M 0 O3 and WO 3 and their interactions with various gases obtained from catalysis literature gave insight to the sensing mechanism of the thin
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
film sensors. Semiconducting oxides are characterized based on the crystal structure and the gases they detect. Re 0 3 -type structures favor adsorption based sensing mechanism and are good sensors for detecting oxidizing gases such as NO 2 , CO2 , O2 , NO and hydrocarbons like methanol and isoprene. The unique 2D-layered structure such as that of stable (X-M0 O3 is favorable for the detection of reducing gases such as ammonia and amines. The sensing mechanism involved in this type of oxides is through reaction based mechanism. It is shown in this study that by controlling the microstructure to obtain different polymorphs, specificity in gas sensing is achieved. A specific sensor array is proposed which consists of sensing elements of un-doped metal oxide.
7.2 Future Work The polymorphs investigated in this study are monoclinic and orthorhombic phases of M0O3 and WO3. However, there are still other polymorphs such as hexagonal, triclinic, cubic, tetragonal etc which are to be explored. Though there is scattered literature available on each of the phase as suitable material for gas sensor, a comprehensive survey of each phase including phase stabilization, identification of specificity towards a particular target gas species, phase transformation mechanism and sensing mechanism of these phases is still lacking. The effect of microstructure of these phases on gas sensing properties will provide better choice of materials for use in specific sensor arrays. In-situ XRD experiments need to be carried out to determine the phases formed in-situ at high temperatures. This would provide a better picture of exact phase
111
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
transformation regions. The phase distribution can be better determined if XRD spectra is taken in-situ at smaller temperature intervals. In-situ TEM experiments should also be carried out to see the grain size changes during phase transformation. Extensive XPS study need to be carried out in case of every sensing test carried out. A detailed XPS study before and after sensing of all gas analytes would confirm the reaction mechanism with those gases. An in-situ XPS with gas atmosphere control would give better idea as to what happens to the electronic structure of the metal oxide during high temperature gas interaction similar to that attained during sensing experiments. New techniques of processing such as through acidic precipitation which yields very high sensitivity and selectivity needs to be explored and standardized for obtaining other oxides as well. Such novel processes which would improve sensor characteristics such as sensitivity, selectivity, response time and recovery times needs to be developed in the search for better sensor materials for sensor arrays and electronic noses.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX
Definition of Sensitivity The sensitivity of a gas sensor is usually defined as
S = | AR | / R 0
(A.l)
Where S = Sensitivity AR = Rg - Ro; difference in resistance with and without the presence of gas Ro = Resistance in air Rg = Resistance in presence of gas This definition has been used throughout the dissertation. However, various other definitions of the term exist in literature. An alternative method of defining sensitivity for oxidizing gas and reducing gas is given as
^reducing Ro/Rs
(A.2)
^ o x id iz in g
(A.3)
R g /R o
This method of defining sensitivity gives a more normalized measure of the sensitivity towards both oxidizing and reducing gas. Based on the above formula, the sensitivities for some of the gas sensors reported are recalculated.
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Results Figure 3.19 in Chapter 3, which shows the sensitivity of sol-gel M 0 O3 to NH 3 at different temperatures can be re-plotted using the new definition of sensitivity as shown in
Figure A.* 1* ®
NHj Sensitivity Vs Concentration
-4 0 0 C -4 2 5 C -4 5 0 C -4 7 5 C
| -------
50
100
200
300
400
500
Concentration (ppm)
Figure A.1 Comparison of Sensitivities of sol-gel M 0 O3 to NH 3 at different operating temperatures. N02 Sensitivity Vs Concentration 1 .5 1 .4 5 1 .4 1 ,3 5 1 .3 -♦ -4 0 0 C -B ~ 4 5 0 C
1 .2 5
- A -475 C
I
12 1 .1 5
1.1 1 .0 5
1 50
100
200
300
400
500
Concentration (ppm)
Figure A.2 Comparison of sensitivities of sol-gel M 0 O3 to NO 2 at different operating temperatures.
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
As is seen, the trend observed is similar as observed with earlier calculation with the only difference being the sensitivity values. The sensitivities towards NO2 at different operating temperatures is also re-plotted as shown Figure A.2 and similar plot is obtained. Figure A. 3 shows the comparison of sensitivities between NH3 and NO2 at 475°C for the sol-gel films. This plot shows the specific response towards NH3 .
Comparison of NH, Vs NO; Sensitivity of Sol-gel MoO, Films at 475 deg C 8 7
6
3 5 < # 4 ■£
e>
c«
w 3
2 1
0 50
100
200
300
400
500
Concentration (ppm)
Figure A.3 Comparison of sensitivities of sol-gel M0 O3 films toward NH 3 and NO2 at 475°C
The comparison of sensitivities of W2 films towards NO2 and NH3 in Chapter 4, can be re-plotted as shown in Figure A.4 and Figure A.5. As observed in the above case, the definition of sensitivity affects only its magnitude and does not affect the nature of the curve.
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W2 S e n s in g T e m p e ra tu re D e p e n d e n c e o n NO* S en sitiv ity
5000 4500 4000 3500 =
3000
-2 0 0 C - 300 C
2500
-4 0 0 C - 500 C
» 2000
1500 1000 500
50
100
200
300
400
500
Concentration (ppm)
Figure A.4 Comparison of Sensitivities of W2 films towards NO 2 at various sensing temperatures W2 Sensing Temperature Dependence on NH3 Sensitivity 25
20
<
15
-200C
- 300 C -4 0 0 C - 500 C
s> 10
5 -HS
0 50
100
200
300
400
500
Concentration (ppm)
Figure A.5 Comparison of sensitivities of W2 films towards NH 3 at various sensing temperatures
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
W3 Sensing Temperature Dependence on N02 Sensitivity 1600
1400 1200
800 ~ T lr ~ 4 0 0 C
500 C 600
400
200 0 50
100
200
300
400
500
Concentration (ppm)
Figure A . 6 Comparison of sensitivities of W3 films towards NO 2 at various sensing temperatures W3 Sensing Temperature Dependence on NH* Sensitivity 18 .-x
16 14
12
D < 10
-2 0 0 C
- 300 C - 400 C - 500 C
>
£
8 6 4 2 0 50
100
200
300
400
500
Concentration (ppm)
Figure A.7 Comparison of Sensitivities of W3 films towards NH3 at various sensing temperatures
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A comparison of sensitivities of W3 films is also re-plotted as shown in Figure A . 6 and Figure A.7. The NO2 sensitivities remain almost the same as calculated before, whereas NH 3 sensitivities have a higher scale of magnitude though the nature of the graphs is same as before. Figure A . 8 shows the comparison of the two films W2, W3 towards NO2 and NH3 at 200°C on a logarithmic scale. It is evident that the sensitivity towards NO2 is two orders of magnitude higher than that towards NH3 . Sensitivities C o m p ariso n N 0 2 Vs NH3
10000
1000
=> < £v>
- 200 C- W2-N02 - 200 C - W2-NH3 • 200 C - W3-N02 -200 C - W3-NH3
100
c
50
100
200
300
400
500
C oncen tratio n (ppm)
Figure A.8 Comparison of sensitivities of W2 and W3 films towards NO2 and NH3 at 200°C.
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Note: Definition of sensitivity only affects the magnitude and not the nature of the graphs as is shown by the plots. Sensitivity plots have been re-plotted for both M 0 O3 solgel films as well as for WO3 films reported in Chapters 3 and 4 respectively. The new definition helps in standardizing the definition for oxidizing gases and reducing gases and provides better resolution in sensitivity differences on a normalized scale.
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
