Tin Oxide Sensors for Insect Infestation Detection 
 
by 
 
Junpin Teng 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial Fulfillment of the 
requirements for the Degree of 
Master of Science 
 
 
 
 
Auburn, Alabama 
December 18, 2009 
 
 
Copyright 2009 by Junpin Teng 
 
 
Approved by 
 
Bryan A. Chin, Chair, Professor of Materials Engineering 
Zhongyang Cheng, Associate Professor of Materials Engineering 
Jong Wook Hong, Associate Professor of Materials Engineering 
Arnold Vainrub, Associate Professor of Anatomy, Physiology and Pharmacology 
 
Abstract 
 
      Agriculture suffers a significant loss every year due to pests. Early detection 
of insect infestation is important to reduce that kind of loss. Plants emit volatile 
organic compounds whose composition will change after the infestation of pests. 
Detection of such change can provide the critical information for pest control. In this 
research thick film hotplate type tin oxide sensors are used to detect Methyl 
Jasmonate, an important plants? defense relating volatile chemical. The tests are 
performed by a gas flowing system. For the concentration dependence study of the 
sensors different partial flow rate (0.02L/min, 0.04L/min, 0.06L/min, 0.08L/min, 
0.10L/min, 0.12L/min, 0.16L/min, 0.20L/min) of Methyl Jasmonate at 25?C in dry air 
flow is tested. Gas chromatography is used to determine the exact concentration of 
Methyl Jasmonate at different partial flow rate. The sensors? relative sensitivity is 
found to be from 2 to 5 and follow the power law relationship with the concentration 
of Methyl Jasmonate. For the temperature dependence study of the sensors the tests 
are performed in dry air flow at different temperature (-22?C, -12?C, -2?C, 7?C, 17?C, 
22?C, 27?C, 32?C, 37?C). The sensor?s relative sensitivity is found to be relatively 
low about 4 around 20?C and increases with decreasing ambient temperature below 
10?C. For humidity dependence study the tests are performed at 25?C with different 
relative humidity (20%, 64%, 70%, 82%, 94%). The relative sensitivity of the sensor 
is decreasing with increasing relative humidity, but still above 3 in high relative 
humidity. According to the performance of the sensors to different concentration of
 
ii
Methyl Jasmonate at different temperature and relative humidity the capability of tin 
oxide sensors in detecting insect infestation is discussed. 
 
iii
Acknowledgements 
 
The author would like to thank Dr. Bryan Chin for advisement and thank 
research group members (Suiqiong Li, Shichu Huang, Wen Shen, John Shu, Shin 
Horikawa, Kanchana Weerakoon) for help during the research. Thanks are also to my 
family members especially my wife Jun Lu for the support during the course of 
investigation. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
iv
 
v
Table of Contents 
 
Abstract????????????..????????..???????..??ii 
Acknowledgements?????????????????.???..?????iv 
List of Tables??.?????????????????.???..?????vii 
List of Figures??.???...???.??????????..???..????viii 
Chapter 1 Introduction?????????...?????...?????????.1 
Chapter 2 Literature review????...??????????????????.5 
2.1 Review on current technologies for gas detection???????..????..5 
2.1.1 Optical gas sensor???.???????????????????.5 
2.1.2 Thermal gas sensor???.??????????????????...8 
2.1.3 Electrochemical gas sensor?????.???????...?????...9 
2.1.4 Gravimetric gas sensor???????????????..????..12 
2.2 Fundamentals of tin oxide semiconductor gas sensor??????????14 
2.2.1 Gas-solid surface adsorption mechanism???.???????...??14 
2.2.2 Theory of adsorption and tin oxide sensor sensing mechanism?..???19 
Chapter 3 Objectives of the research???????????????????28 
Chapter 4 Experimental procedure?????????????????..??.29 
4.1 Concentration dependence test?????..???????????.??31 
4.2 Temperature dependence test????????????...?????.?.32 
4.3 Humidity dependence test??????????????????....?..33 
Chapter 5 Results and discussion????????????????????.34 
5.1 Concentration dependence test?????..??????????.???34 
5.2 Temperature dependence test????????.?????...?????.42 
5.3 Humidity dependence test??????????.......................??.??..42 
Chapter 6 Conclusions???..??????????????...??????47 
References?...???????????????????????????.54 
 
vi
List of Tables 
 
Table 1:  Comparison of different types of gas sensors???????????..14 
Table 2:  Different formulas describing response of semiconductor gas 
sensor???????????????????????????????27 
Table 3:  Calculated concentrations of different gas samples?????????40 
 
vii
 
viii
List of Figures 
 
Figure 1:  Plants? time dependent defensive responses????????????1 
Figure 2:  Different types of chemical vapor sensors????????????...5 
Figure 3:  Schematic of a fluorescence gas sensor?????????????...6 
Figure 4:  Schematic of a SPR based gas sensor??????????????..7 
Figure 5:  Schematic of an infrared absorption optical gas sensor???????...8 
Figure 6:  Schematic of conductometric gas sensor?????????????.9 
Figure 7:  Schematic of conventional potentiometric gas sensor??????.?..10 
Figure 8:  Schematic of an amperometric gas sensor??????????.??11 
Figure 9:  Schematic of a surface acoustic wave gas sensor?????????..13 
Figure 10:  Schematic of the remote monitoring system for plants? 
health???................................................................................................................14 
Figure 11:  The energy of the system as a function of adsorbate/adsorbent 
distance.........................................................................................................................15 
Figure 12:  Typical temperature dependence of adsorption 
coverage??????????????????????????...???18 
Figure 13:  Band bending in the near surface region ?????????...??.20 
Figure 14:  Depletion layer formed at the surface of tin oxide??????...?...21 
Figure 15:  Functions of physical properties???????????????..23 
Figure 16:  Open neck????????????????????????25 
Figure 17:  Closed neck???????????????????????..25 
Figure 18:  Schottky contact?????????????????????...26 
Figure 19:  Schematic of gas flow circuit for concentration dependence and 
temperature dependence tests???????.???????????...??...29 
Figure 20:  Schematic of TGS 2620 tin oxide sensor????????????.30 
Figure 21:  Schematic of gas flow circuit for humidity dependence 
tests?????????????????????????????....?...33 
Figure 22a:  Sensors? responses to MeJa with partial flow rate 
0.02L/Min??........................................................................................................?.34 
Figure 22b:  Sensors? responses to MeJa with partial flow rate 
0.04L/Min?????????????????????????????.35 
Figure 22c:  Sensors? responses to MeJa with partial flow rate 
0.06L/Min?????????????????????????????.35 
Figure 22d:  Sensors? responses to MeJa with partial flow rate 
0.08L/Min?????????????????????????????.36 
Figure 22e:  Sensors? responses to MeJa with partial flow rate 
0.10L/Min?????????????????????????????.36 
Figure 22f:  Sensors? responses to MeJa with partial flow rate 
0.12L/Min?????????????????????????????.37 
Figure 22g:  Sensors? responses to MeJa with partial flow rate 
0.16L/Min?????????????????????????????.37 
Figure 22h:  Sensors? responses to MeJa with partial flow rate 
0.20L/Min?????????????????????????????.38 
Figure 23:  GC calibration curve????????????????????39 
Figure 24:  Sensors? responses to MeJa at different concentration?????...?.40 
Figure 25:  Relative sensor sensitivity S to MeJa at different 
concentration????????????????????????????41 
 
ix
 
x
Figure 26a:  Responses of sensor 1 to MeJa at different temperatures?????..42 
Figure 26b:  Responses of sensor 2 to MeJa at different temperatures?????..42 
Figure 27:  Sensors? resistances at different temperature?????????...?43 
Figure 28:  Relative sensor sensitivity S to MeJa at different ambient 
temperature????????????????????????????...44 
Figure 29:  Ambient temperature effect on sensors? working temperature as a linear 
superposition????...?????????????????????.??45 
Figure 30:  Survey of oxygen species detected at different temperatures at SnO2 
surfaces??????????????????????????..???...46 
Figure 31:  Sensors? logarithmic resistances in dry air at different 
temperature????????????????????????????...46 
Figure 32a:  Sensors? responses at relative humidity 20%.........................................47 
Figure 32b:  Sensors? responses at relative humidity 64%.........................................48 
Figure 32c:  Sensors? responses at relative humidity 70%.........................................48 
Figure 32d:  Sensors? responses at relative humidity 82%.........................................49 
Figure 32e:  Sensors? responses at relative humidity 94%.........................................49 
Figure 33:  Relative sensor sensitivity at different relative humidity?????....50 
Figure 34:  Sensors? initial resistances at different relative humidity........................51 
Figure 35:  Sensors? resistances in the presence of MeJa at different relative 
humidity???????????????????...??????????51 
CHAPTER 1:  INTRODUCTION 
 
The World Food and Agriculture Organization of the United Nations (FAO) 
estimates that the annual loss of edible food plants due to pests is about 14%.  This is 
equivalent to more than $200 billion dollars in economic loss every year.  In the 
USA, one third of all agriculture products are damaged by pests which leads to an 
average annual economic loss of over $4 billion [1].  Hence early detection of insect 
pests could significantly reduce current economic losses. 
Time dependent defensive responses
0 102030405060708
Vm Changes
[Ca2+] Changes
ROS Produced
VOCs Change
Gene expression
Metabolites Change
Ev
en
t
s
Time
SEC Min Hours Days
0
 
Figure 1:  Plants? time dependent defensive responses 
 
Plants have a natural phytochemical defense system that is activated by invading 
herbivores. After being attacked by herbivores, plants have a series of responses at
 
1
different time stages (Fig 1) [2]. The chain responses start from the plasma membrane 
of plant cells which directly contact the environment. Seconds after herbivores? attack 
plants the plasma transmembrane potential (Vm) changes [3]. Vm is directly 
influenced by ion influxes among which Ca
2+
 is predominant in cell signaling and 
recognized as an important second messenger [4]. Following the Ca
2+
 concentration 
change, reactive oxygen species (ROS) such as H
2
O
2
 are increasingly produced. At 
low concentration ROS also acts as secondary messengers [5-6]. As the level of ROS 
increases when the insects? attack persist ROS especially H
2
O
2
 can defend a plant?s 
wounded area from further invasion of bacteria, fungi, or viruses [7]. After ROS 
changes volatile organic compounds (VOC) emitted by the plants change. VOC have 
a significant role in the plants? defense system which will be explicated later. 
Regulated by the network of phytohormones, the process of plants? biosynthesis 
changed. As a result the composition of plants? metabolites changes. Many secondary 
metabolites which usually act as a toxin to herbivores are produced to directly defend 
plants from the herbivores? invasion [8-10]. 
To communicate with the environment, plants emit hundreds of volatile organic 
compounds. At present more than 1700 kinds of VOC are identified from various 
plant families [11]. These chemical substances can not only regulate processes of 
growth and reproduction but also defend plants from herbivore?s attack or deal with 
physical stress [12]. When plants are fed on by herbivores they will emit a different 
mixture of VOC blend. The difference of the mixture from normal conditions can be 
quantitative or qualitative. Quantitative difference is the relative composition ratio 
change of VOC mixture emitted by plants after and before a herbivore?s infestation. 
Qualitative difference is due to new kinds of VOC that are produced by plants after an 
insects? attack. 
 
2
The VOC emitted by infested plants can defend against herbivores in three major 
ways. First, some VOC can directly repel insects [13-14]. Second, some VOC can 
attract natural predators of herbivores or egg parasitoids [15-19]. Therefore, the 
population of herbivores can be reduced to prevent plants from being overtaken. Third, 
some VOC can warn neighboring plants to activate their defense system [20-25]. 
Those VOC can trigger the biosynthesis of plant secondary metabolites which usually 
act as a toxin or inhibitors of insect digestive enzymes [8-10, 26-28]. Thus, 
neighboring plants can avoid or deter the future feeding of insects. 
Among the many VOC, one family named jasmonates which is composed of 
Jasmonic acid (JA) and its derivatives should be noticed. These compounds will 
change quantitatively or qualitatively at the early stage of VOC changes that are 
observed minutes after plants being damaged by herbivores [29]. And they have 
intimate relationship with the plants? defense system. On one hand, they are important 
components in systemic wound signaling. Farmer et al discovered that local treatment 
of jasmonates can activate the systemic expression of wound response genes. Based 
on this finding jasmonates are proposed as an important signaling VOC [30]. Later 
this hypothesis was substantiated by famous grafting experiments conducted by Li et 
al [31]. On the other hand, jasmonates play an integral role in crosstalk between 
signaling networks [32]. The jasmonate?s signaling pathway interacts with many other 
signaling pathways such as ethylene, light, auxin, salicylic acid, and abscisic acid 
[33-37]. MeJa is a volatile counterpart of JA which is converted from JA by JA 
carboxyl methyltransferase (JMT) [38]. MeJa is vital substance in plants and is 
proposed as a strong candidate for an important role in systemic signaling [39]. It can 
diffuse to distal parts of the plant in vapor phase and activate formation of defense 
proteins, stress protective proteins, and thicker cell wall [40, 41]. When the plants are 
 
3
 
4
infested the emission of Methyl Jasmonate the volatile counterpart of JA increases 40 
fold to the concentration of several ppm [42]. Thus, it can be used to identify the early 
stages of an insect infestation. Once detected, appropriate treatment can be carried out 
in time to control the insects and prevent further loss. 
The metal oxide sensor is one kind of conductometric sensor. Conductometric 
sensors have a significant portion of the commercial gas sensor market. SnO
2
 is one 
of the most commonly used materials in metal oxide sensors.  Metal oxide sensors 
have the merits of high sensitivity, quick response and low cost.  Metal oxide sensors 
have been used to detect a wide range of gas species such as CO, H
2
, CH
4
, H
2
S, NO
X
, 
CO
2
, and O
2
 [43-73]. However, there has been no investigation involving detection of 
Methyl jasmonate and other gases emitted by a plants? phytochemical defense system. 
Hence, there is a demand to identify the capability of tin oxide sensors in detecting 
volatile organic compounds for the application of insect infestation detection. This 
thesis will investigate the performance of the tin oxide sensor in the detection of 
insect infestation.
CHAPTER 2:  LITERATURE REVIEW 
 
2.1.  Review of current technologies for gas detection 
There are many sensing technologies that may be used to detect Methyl 
Jasmonate the volatile counterpart of Jasmonic acid. Different principles that may be 
used for detection are summarized in Figure 2. The following are the discussions on 
different types of gas sensors. 
Chemical Vapor Sensors
Optical Thermal Electrochemical Gravimetric
Fluorescence
Surface 
Plasmon 
Resonance
Absorption
Pellistor Conductometric
Potentiometric
Amperiometric
Surface 
Acoustic 
Wave
 
Figure 2:  Different types of chemical vapor sensors. 
 
2.1.1.  Optical gas sensor 
There are various types of optical gas sensors based on one or more different
 
5
optical phenomena (1) Fluorescence (2) Surface Plasmon Resonance (SPR) (3) 
Absorption. Fluorescence is the emission of photons by molecules after being 
electronicly excited by absorbing light. Figure 3 illustrates a typical gas sensor based 
on fluorescence. After the fluorescence material absorbs the excitation light it will 
emit a light with longer wavelength which is the fluorescence signal. In the presence 
of target gas the fluorescence intensity emitted by fluorescence material will change 
due to interaction between the gas and the sensing material. Hence, the gas 
concentration can be related to the fluorescence intensity F as: 
?
?
?
?
?
?
?
?
?
?
=
FF
FF
KC
max
min
    (1) 
where  is concentration of the target gas,  is fluorescence intensity without 
presence of target gas,  is fluorescence intensity when the fluorescence material 
is saturated by target gas. This kind of sensor has high sensitivity and wide detection 
range. The major problem with this type of sensor is durability due to the degrading of 
the fluorescence material which also reduces the sensitivity and accuracy of the sensor 
[75]. 
C
min
F
max
F
 
Figure 3:  Schematic of a fluorescence gas sensor [74] 
 
Surface Plasmon resonance (SPR) sensors are a second type of optical sensor 
 
6
that can be used for gas detection. Surface plasmon is a charge-density oscillation 
which may occur at the boundary of two media with opposite sign dielectric constants. 
It can be considered as an electromagnetic wave propagating along the surface in 
transverse magnetic mode. Figure 4 is a typical gas sensor based on SPR. When the 
component of a light?s wave vector which is parallel to the interface matches the 
propagation constant of the surface plasmon, SPR occurs. The energy will transfer 
from the incident light to surface plasmon, so the intensity of reflected light decreases. 
At a certain angle 
SPR
?  the intensity of reflected light is reduced to minimum. After 
target gas interacting with the binding layer the surface condition of the metal layer 
changes. As a result 
SPR
?  changes which can be used for detection. This type of gas 
sensor is easy to use and extremely sensitive. However, it is bulky, expensive and 
mainly used in laboratories [76]. 
 
Figure 4:  Schematic of a SPR based gas sensor [76]. 
 
Figure 5 is a typical gas sensor based on the light absorption by the gas. When 
white light passes through gases it will be absorbed at different extent at different 
wavelengths; each specific gas has its own light absorption spectrum which can be 
 
7
used to discriminate the gases. Based on the Beer?Lambert law, the light intensity of 
the sensor can be related to target gas concentration as, 
cl
e
I
I
???
=
?
0
    (2) 
where I  is the transmitted intensity of the light after absorption,  is the incident 
intensity of the light before absorption, 
0
I
?  is the absorptivity of the gas species,  is 
the optical path length and  is the concentration of the gas species. The merits of 
this kind of sensor are high sensitivity and selectivity. The disadvantages are 
relatively long response time and high cost. Though optical fibers can make the 
system more cost effective, it limits the working wavelength range; consequently, 
limiting the gas species that can be detected [74]. 
l
c
 
Figure 5:  Schematic of an infrared absorption optical gas sensor [74]. 
 
2.1.2.  Thermal gas sensor 
Pellistor is the major kind of thermal gas sensor. It consists of a platinum wire 
coil in an alumina bead covered with a thick film catalyst, which is normally 
palladium [77]. Gas is combusted by electrically heating the catalyst material to the 
required temperature (about 500 ?C) with the platinum wire. As a result the 
temperature increases due to oxidation of the combustible gas and is measured in 
 
8
terms of the platinum resistance relative to a reference resistance. The resistance 
change is related to the concentration of the gas. This type of gas sensor is not widely 
used due to high power consumption and poisoning of catalyst material by sulfur 
compounds [78]. 
 
2.1.3.  Electrochemical gas sensor 
There are three major kinds of electrochemical gas sensors (1) Conductometric; 
(2) Potentiometric and (3) Amperometric. Conductometric gas sensors are usually 
fabricated as a chemiresistor as shown in Figure 6. The sensing element of the 
chemiresistor is conducting polymer or semiconductor. When the sensing element is 
exposed to the target gas the conductance (inverse resistance) of the sensor changes 
from that measured in normal air due to the adsorption of target gas. Hence the 
conductance (inverse resistance) change can be related to the gas concentration. This 
type of gas sensor has high sensitivity, quick response and good stability. The main 
drawback is poor selectivity [79-80]. 
 
Figure 6:  Schematic of conductometric gas sensor. 
 
Figure 7 illustrates the conventional potentiometric gas sensor. Due to the gas 
concentration difference between the sensing (working) electrode side and reference 
 
9
electrode side there is an electrochemical potential difference that develops between 
the electrodes (emf). Based on the Nernst law the open circle voltage measured 
between the electrodes can be related to the gas concentration as, 
0
ln
p
p
kE =     (3) 
where E  is the potential measured, k  is constant, p  is the partial pressure of 
target gas which needs to be detected,  is the partial pressure of gas concentration 
which is known. This type of gas sensor is highly sensitive and selective, but lacks 
long term stability. Moreover it is restricted to detect the gas which corresponds to the 
mobile ionic species in the electrolyte. Auxiliary electrodes can be added to detect 
other gases. However the complexity and cost of the system is increased [81-82]. 
0
p
 
Figure 7:  Schematic of conventional potentiometric gas sensor with (a) oxygen 
conductor (b) protonic conductor [81]. 
 
 
10
Figure 8 is a schematic of an amperometric gas sensor. After introduction of 
analyte gas the membrane will allow the target gas to diffuse into the membrane and 
effectively prevent the diffusion of other unwanted gases. The reference electrode is 
used to maintain the sensing electrode at a known electrochemical potential. By 
applying a constant potential between the working electrode and the reference 
electrode, the electroactive species of electrolyte will participate in electrochemical 
reaction. A reduction reaction occurs at the cathode while an oxidation reaction 
occurs at the anode. Thus, a current flow is established. Based on Faraday?s law the 
relationship between the current flow of the circuit and gas concentration can be 
expressed as, 
kpi =     (4) 
where  is constant, k p  is the gas concentration. This kind of gas sensor has good 
sensitivity and selectivity but has relatively long response time and is poor in long 
term stability [83]. 
 
1. Gas inlet 2. Gas filter 3. Gas outlet 4. Gas membrane 5. Working electrode 6. Reference 
electrode 7. Counter electrode 8. Electrolyte 
Figure 8:  Schematic of an amperometric gas sensor [83]. 
 
11
2.1.4.  Gravimetric gas sensor 
The surface acoustic wave (SAW) sensor is the dominate gravimetric sensor used 
in gas detection. As shown in Figure 9 a SAW system is usually constructed on a 
piezoelectric substrate. The interdigitated transducer is used to generate a SAW. The 
SAW, first discovered by Lord Rayleigh, can couple with any media contacting the 
surface. Hence, when target gas is adsorbed or absorbed on the selective coating at the 
surface of the substrate, the amplitude and velocity of the SAW will change. Such a 
change can be interpreted by the receiving oscillator in terms of frequency shift which 
can be expressed as: 
mKff ??=?
2
0
    (5) 
where  is the frequency shift, f? K  is a constant,  is the resonant frequency, 
and  is the mass change. This kind of sensor has the merits of high sensitivity, 
rapid response, low cost and small size. The main challenge comes from the sensing 
film which is coated on to the surface. The sensor response is dependent on film 
thickness, so the deviation of film thickness significantly affects the reproducibility of 
the sensor. Also the interaction between specific molecules and the sensing film 
changes the viscoelasticity which affects the sensor response [85]. 
0
f
m?
 
12
 
Figure 9:  Schematic of a surface acoustic wave gas sensor [84]. 
 
Among the many different gas sensors discussed above, four types, (1) infrared 
absorption (optical); (2) electrochemical sensors (potentiometric or amperometric); (3) 
conducting polymer (conductometric) and (4) semiconductors (conductometric) are 
most attractive in gas detecting [86]. In practical use many features of gas sensors 
should be considered to fulfill the specific requirements. The final goal of this 
research project is to establish a remote monitoring system for plant health which is 
illustrated in Figure 4. The in field sensing system will detect the VOC emitted by 
plants and send the data to a host computer wirelessly. By analyzing the data, the 
in-situ plants? health condition will be monitored. The semiconductor gas sensor has 
merits of good sensitivity, quick response and high durability. Moreover, it is easy to 
combine the sensing element, signal converter, control electronics and other required 
electronics in one device [87]. Therefore, according to the comparison of the four 
popular gas sensors [88] shown in table 1 the semiconductor gas sensor, which will be 
described in detail later, was selected for this research. For this study tin oxide sensor 
 
13
model TGS2620, manufactured by Figaro Inc., was investigated. 
 
Figure 10:  Schematic of the remote monitoring system for plants? health. 
 
Table 1:  Comparison of different types of gas sensors. 
Parameter
Semiconductor Conducting Polymer Electrochemical Infrared Absorption
Sensitivity e e g e
Accuracy g g g e
Selectivity p p g e
Response time e e p p
Stability g g b g
Durability g b p e
Maintenance e g g p
Cost e e g b
Type of gas sensors
 
e: excellent; g: good; p: poor; b: bad 
 
2.2.  Fundamentals of tin oxide semiconductor gas sensor 
2.2.1.  Gas-solid surface adsorption mechanism 
When gas atoms or molecules approach a solid surface they can be attracted and 
maintained on the surface in two modes: (1) physisorption and (2) chemisorption. 
 
14
Physisorption or physical adsorption is a weak bonding due to the induced dipole 
moment of a nonpolar adsorbate interacting with its own image charge in the 
polarizable solid, i.e. Van der Waals force. During this process the electronic state of 
the atom or molecule barely changes. Chemisorption or chemical adsorption is a 
strong bonding between a gas adsorbate and solid surface, i.e. ionic bonding or 
covalent bonding. This process involves the electronic structure change of the gas 
atom or molecule. 
 
Figure 11:  The energy of the system as a function of adsorbate/adsorbent distance 
(a) physical adsorption and (b) chemical adsorption [89]. 
 
According to the Lennard-Jones model, the energy of the system is illustrated as 
a function of distance between gas atom or molecule and solid surface as shown in 
Figure 11 [89]. When the distance between the molecule and surface is infinite, the 
energy of the system is defined as zero. As the molecule approaches the surface it can 
polarize and induce an equivalent dipole in the solid. Due to the dipole-dipole 
 
15
interaction, the energy of the system decreases. When the distance between the 
molecule and the surface decreases to , the
 
Van der Waals force will be balanced 
by the repulsive force between the atom and substrate. By further approaching 
repulsion will dominate which will lead to an increase of the system energy. 
Therefore, at equilibrium distance r
p
 the system has the lowest energy. The 
physisorbed molecule is stable there and has the binding energy  with the 
surface which is typically 10~100 meV. From the distance dependent energy curve 
for physisorption (curve a) it should be noticed that no activation energy is needed for 
such a process which means that physisorption is temperature independent. By 
contrast for desorption an activation energy 
p
r
PHY
H?
PHY
H?  is needed. Hence, the 
physisorption rate can be expressed as, 
() ?
?
?
?
?
? ?
??????=
?
RT
H
kPk
dt
d
PHY
d
O
a
PHY
exp       (6) 
Where,  is the density of surface adsorption sites, 
O
? ? is the density of surface 
sites covered by adsorbate, P  is gas pressure,  is adsorption rate constant.,  
is desorption rate constant, 
a
k
d
k
R  is gas constant, T  is temperature. At steady state, the 
adsorption rate equals desorption rate i.e. physisorption rate equals zero. By 
defining
?
?
=?
O
, the fractional coverage, equation 1 yields, 
?
?
?
?
?
? ?
?+
=
RT
H
k
k
P
P
PHY
a
d
exp
?      (7) 
At low gas pressure, equation 7 can be reduced to, 
P
RT
H
k
k
PHY
d
a
?
?
?
?
?
??
? exp?     (8) 
This equation indicates that fractional coverage has a linear relationship with gas 
 
16
pressure at low gas concentration and decreases with a temperature increase which 
means that at high temperature physisorption is negligible. 
Besides physisorption, the gas molecules can bond to the solid surface in another 
mode, chemisorption. The process (curve b) starts from the dissociation of molecules 
into atoms which needs a dissociation energy
D
E? . As the atom approaches the 
surface to certain distance, , a strong chemical bond will form by electron transfer 
or sharing. In this case a binding energy or heat of chemisorption, , is large, 
typically 1-10 eV. Due to a large dissociation energy, chemisorption generally does 
not happen directly. In practice, it usually follows physisorption which needs a much 
smaller amount of activation energy
c
r
CHEM
H?
a
E? . Hence, the chemisorption rate can be 
expressed as:  
() ?
?
?
?
?
? ?+?
????
?
?
?
?
? ?
????=
?
RT
EH
k
RT
E
Pk
dt
d
aCHEM
d
aO
a
CHEMY
expexp    (9) 
Therefore, at steady state the equilibrium fractional coverage can be written as, 
?
?
?
?
?
??
=
RT
H
k
k
CHEM
d
a
exp?     (10) 
It?s easy to understand from equation 9 that both adsorption and desorption process 
require activation energy. So chemisorption occurs at higher temperatures than 
physisorption. 
 
17
 
Figure 12:  Typical temperature dependence of adsorption coverage [89]. 
 
The typical temperature dependence of adsorption coverage is sketched as a 
dotted curve in Figure 12. At low temperature (region I), the thermal energy is not 
enough to overcome the activation energy, 
a
E? , so only physisorption occurs. The 
equilibrium coverage follows the physisorption curve which can be expressed as 
equation 3. At high temperatures (region III), enough thermal energy is provided to 
activate both adsorption and desorption processes of chemisorption. Therefore, 
equilibrium coverage can be reached and the chemisorption curve is followed which 
is given by equation 5. In the intermediate temperature range (region II), on one hand 
thermal energy is enough to activate the adsorption process of chemisorption. So 
chemisorption sets in and transition from physisorption to chemisorption starts. On 
the other hand, the thermal energy is not enough to activate the desorption process of 
chemisorption. Therefore, equilibrium coverage of chemisorption can?t be reached. 
 
18
Then, a question arises: how can the adsorption be balanced otherwise it never stops? 
Actually, it can be explained by the activation energy
a
E? . In practice,  is not 
constant, it increases with the coverage. The solid surface is heterogeneous. So 
adsorbate will first fill the site with lowest energy i.e. highest
a
E?
CHEM
H? . As a result, 
 decreases with increasing coverage. In the Lennard-Jones model, this is 
expressed as the upward shift of curve b in Figure 4. As curve a is not affected, 
CHEM
H?
a
E?  
will increase as a result. When the thermal energy is not enough to overcome the 
increased  chemisorption stops. Hence, if the temperature increases again, the 
adsorption will continue which explains the increasing coverage with temperature in 
region II. It should be noticed that such chemisorption is irreversible due to the lack of 
chemical desorption. So only lowering the temperature or pressure will result in 
residual coverage. Therefore, metal oxide semiconductor sensors operated in 
intermediate temperature range require high temperature surface cleaning. Otherwise, 
the sensor will inevitably become less sensitive due to the reduction of active sensing 
sites. 
a
E?
 
2.2.2.  Theory of adsorption and tin oxide sensor sensing mechanism 
It is easy to understand that the electron energy level of an atom in vacuum is 
different from the one in a material. At the surface which can be recognized as a 
defect of materials, electronic band structure will change from the bulk material to the 
vacuum. Thus, new electronic surface states are formed at atom layers which are 
closest to the surface. For ionic materials, the unoccupied orbital on the cation acts as 
an acceptor-like surface state (Lewis acid site) and the occupied orbital on the anion 
acts as a donor-like state (Lewis base site). For a covalent material, atoms at the 
 
19
surface form dangling bonds. Besides the intrinsic surface state which is only due to 
termination of lattice periodicity another surface state, so called extrinsic surface state 
originates from the adsorption of gas impurities. When a foreign atom or molecule is 
adsorbed on the surface, electrons can either transfer from bulk material to foreign 
species if the induced surface state is lower than the Fermi level of the solid (average 
energy level of electrons), or from foreign species to bulk material if the induced 
surface state is higher than the Fermi level of the solid. Thus, foreign species are 
chemisorbed and cause an electric field due to the surface charge layer formed. This 
electrostatic field will bend the energy band of the solid. A negative surface charge 
layer will bend the bands upward while a positive surface charge layer will bend the 
bands downward as shown in Figure 13. As the Fermi level is pushed back to the 
band gap, the charge carrier concentration of semiconductors is reduced and a 
depletion region is formed which compensates for the surface charge layer. 
 
Figure 13:  Band bending in the near surface region of (a) an n-type semiconductor 
 (b) a p-type semiconductor; ES, surface state energy level 
 
20
 
Figure 14:  Depletion layer formed at the surface of tin oxide 
 
Taking tin oxide solid and oxygen gas as the example which is illustrated in 
Figure 14, electrons flow due to the electron energy level difference between the 
oxygen induced surface state and the tin oxide bulk material. Oxygen captures 
electrons and forms a negative surface charge layer on the surface. A depletion layer 
within  from the surface is formed and the electronic bands bend upward. The 
charge distribution can be quantitatively derived from the one dimensional Poisson?s 
equation [90]: 
0
x
 
21
( )
r
x
dx
Vd
??
?
0
2
2
?=     (11) 
where  is the potential, V x  is the distance from surface, 
0
?  is the permittivity of 
vacuum, 
r
?  is the relative permittivity of the semiconductor, ?  is the net charge 
density which consists of electrons ( ) , holes (n p ), ionized donor ( ) and ionized 
acceptor ( ). In the depletion region, the concentration of electrons and holes is 
assumed to be zero. Thus, the net charge density within the distance can be 
expressed as: 
+
D
N
?
A
N
0
x
( )
?+
?=
AD
NNe?     (12) 
Substituting equation (12) into equation (11) and integrating once, 
()
( )
.
0
consx
NNe
dx
dV
xE
r
AD
+
?
?==
?+
??
    (13) 
Assuming the potential in the depletion region smoothly changes into the potential of 
the bulk material  one can get the boundary condition that , 
b
V
b
VV = 0=
dx
dV
, at 
. Thus, equation (13) becomes, 
0
xx =
()
( )
(
0
0
xx
NNe
dx
dV
xE
r
AD
?
?
?==
?+
??
)    (14) 
Integrating equation (14) and given the boundary one can obtain, 
()
( )
()
b
r
AD
Vxx
NNe
xV +?
?
?=
?+
2
0
0
2 ??
    (15) 
According to the calculation above, functions of the space charge, electric field, and 
potential energy can be illustrated as in Figure 15. 
 
22
 
Figure 15:  Functions of physical properties (a) the space charge, (b) the electric 
field, and (c) the potential [90]. 
 
By defining 0=
b
V , surface barrier height, , at 
S
V 0=x  becomes, 
( )
2
0
0
2
x
NNe
V
r
AD
S
??
?+
?
=     (16) 
Thus, the energy barrier, , which electrons must overcome to move to surface 
states becomes, 
S
eV
( )
2
0
0
2
2
x
NNe
eV
r
AD
S
??
?+
?
=     (17) 
Based on charge neutrality, surface charge should equal space charge i.e. 
 
23
( )
0
xNNN
ADS
?+
?=     (18) 
where  is the number of occupied surface states per unit area. Substituting 
equation (18) into equation (17) and eliminating , one can obtain, 
S
N
0
x
()
?+
?
=
ADr
S
S
NN
Ne
eV
??
0
22
2
    (19) 
It is easy to see that the activation barrier is proportional to the square of surface 
coverage ( ) and increases with increasing coverage. For an n-type 
semiconductor , so the surface coverage can be expressed as, 
S
N
N
?+
>>
AD
N
2
1
0
2
?
?
?
?
?
?
?
?
=
+
e
VN
N
SDr
S
??
    (20) 
Based on equation (20) and assuming a reasonable impurity concentration (10
25
/m
3
), a 
fairly high barrier potential (1V) and a total number of surface sites (10
19
/m
2
), Weisz 
[91] calculated the fractional coverage ?  and obtained the value 0.9%. Given the 
high sensitivity of such kind of sensor, the surface barrier at intergranular contact is 
believed to play a key role. 
By comparing the neck width and Debye length of electron L
D
 which can be 
expressed as equation (21), the intergranular contacts can be classified into three 
kinds: (1) Open neck (2) Closed neck and (3) Schottky contact. 
2
1
2
0
?
?
?
?
?
?
?
?
=
b
r
D
ne
kT
L
??
    (21) 
where  is the Boltzmann constant,  is the electron density of bulk material. k
b
n
 
24
 
Figure 16:  Open neck 
 
When the neck width is larger than  the contact is called open neck as 
shown in Figure 16. The highly resistive depletion layer is parallel connected with the 
lower resistive bulk grain at the neck. Hence, the conductance is governed by the 
undepleted layer width  and can be expressed as: 
D
L2
w
?
?
?
?
?
?
?
??
kT
E
wG
b
exp     (22) 
where  is the activation energy of electron from donor level to conduction band 
in bulk material. This case is often found in well sintered ceramics. 
b
E?
 
Figure 17:  Closed neck 
 
When the neck width is smaller than , the contact is called closed neck as 
shown in Figure 17. The depletion layers at opposite surfaces at the neck overlap and 
D
L2
 
25
form a more resistive contact. The conductance is dominated by the activation of 
electron from surface state to conduction band at the surface and can be written as: 
?
?
?
?
?
?
?
??
kT
E
G
S
exp     (23) 
where  is activation energy of electron from surface state to conduction band at 
the neck. This case usually can be found in less sintered ceramics. 
S
E?
 
Figure 18:  Schottky contact 
 
When no neck exists and two particles with diameter greater than 2 L
D 
are only 
?pressed? together, the contact is called Schottky contact as shown in Figure 18. 
Electrons must overcome the energy barrier  which appears in equation (19). 
Thus, the conductance can be expressed as: 
S
eV
?
?
?
?
?
?
??
kT
eV
G
S
exp     (24) 
In practice a sensor usually consists of all three types of contact. At high 
 
26
sintering temperature Schottky contact can convert to neck type contact. Thus, 
thermal treatment of the sensing element will affect the composition of different 
contacts which determines the response characteristics of a sensor. To describe the 
response of the tin oxide sensor various empirical formulas have been proposed as 
shown in table 2. 
 
Table 2:  Different formulas describing response of semiconductor gas sensor. 
Reference Formula
[92]
[93]
[94]
[95]
[96]
[97]
[98]
( )
?
KCRR += 1
0
?
KCRR
0
=
?? += CR log
?
KCG =
?
GCGG =?
0
( )
?
KCGG exp
0
=
?
KCGG +=
0
 
G: conductance in gas; G
0
: conductance in air; R: resistance in gas; R
0
: resistance in 
air; C: gas concentration; K,? ,? : constants. 
 
27
CHAPTER 3:  OBJECTIVES OF RESEARCH 
 
The final goal of the research project is to establish a remote monitoring system 
for plant health. The objective of this research is to identify the capability of tin oxide 
semiconductor sensor to detect Methyl Jasmonate, an important chemical vapor 
related to a plants? natural defense system. By studying the concentration dependence, 
temperature dependence, humidity dependence, recovery time, stability of the sensor, 
the suitability of a semiconductor type sensor for detection of insect infestation will 
be determined. 
 
28
CHAPTER 4:  EXPERIMENTAL PROCEDURE 
 
 
Figure 19:  Schematic of gas flow circuit for concentration dependence and 
temperature dependence tests. 
 
The gas flow circuit for concentration dependence and temperature dependence 
tests is presented in Figure 19. A compressed air tank furnished by Airgas, Inc was 
used as the carrier gas supply. A water vapor trap was installed to remove any 
moisture in the gas supply. Gas flow rate was monitored by a Mass flow meter 
(FMA1700/1800) made by OMEGA Engineering, Inc. Two TGS 2620 tin oxide 
sensors manufactured by Figaro, Inc were put in the flask B for detection tests. In
 
29
addition, a thermocouple and a battery-powered remote temperature and humidity 
logger (USB 502) are located in flask B to monitor the ambient temperature and the 
humidity. The flask B is kept in an environmental chamber (ECT-3) manufactured by 
ESPEC North America, Inc to control the temperature. TGS 2620 (Figure 20) is a 
thick film type semiconductor sensor with an alumina substrate. It is highly sensitive 
to the vapors of organic solvents as well as other volatile vapors. The sensing element 
is integrated with a heater to raise the temperature to about 400?C. Pins 1 and 4 are 
connected to a DC power supply (Agilent E3611A) for heating. Pins 2 and 3 are 
connected to a data acquisition system (Agilent 34970A) for resistance response 
measurements. 
 
Figure 20 Schematic of TGS 2620 tin oxide sensor. 
 
 
 
30
4.1.  Concentration Dependence Test 
For the concentration dependence test the total gas flow rate was fixed at 
0.2L/min. Before the tests, the sensors were kept in air in an environmental chamber 
for 7 days for conditioning. To establish the base line of the signal, regulator 1 was 
open; gates 1 and 2 were closed. When the resistances of the sensors were stable 
regulator 1 and gates 1 and 2 were opened. Thus, carrier air gas will flow into the 
liquid Methyl Jasmonate which was kept in flask A and take the chemical vapor into 
the testing chamber (flask B). The sensors were exposed to the chemical vapor for 30 
minutes. Then gates 1 and 2 were closed for recovery test. When the resistances of the 
sensors recover and became stable again another test cycle was started. To change the 
concentration of the target gas vapor, the gas flow rate through flask A and through 
the bypass (regulator 1) were regulated by gas flow regulator 1. The partial flow rate 
through flask A was set at 8 different points (0.02 L/min, 0.04 L/min, 0.06 L/min, 
0.08 L/min, 0.10 L/min, 0.12 L/min, 0.16 L/min, 0.20 L/min) but the total flow rate 
was kept a constant at 0.2L/min. The temperature is maintained at 25?C by the 
environmental chamber and monitored by thermocouple. 
To determine the actual concentration of Methyl Jasmonate at different flow 
rates Gas Chromatography (GC) was used. A Varian CP-3380 GC machine with 
DB-5 column and flame ionization detector (FID) was used for this study. To make 
standard solutions different amounts of Methyl Jasmonate (0.1?L , 0.5?L, 1?L, 5?L, 
10?L, 100?L) was dissolved in 5mL Dichloromethane. Gas samples at different flow 
rates were collected using a volatile collection trap (VCT - 1/4 - 3 - HSQ-P), which 
was fabricated by ARS, Inc. The volatile collection trap was connected to the gas flow 
tube for three hours. Then 2mL Dichloromethane was injected into the trap to extract 
the Methyl Jasmonate adsorbed. Every time 0.1?L of standard solutions or gas sample 
 
31
solutions was injected into the GC for test using a micro syringe. During the 
experiment the injector of the GC was set at 275?C and the FID was set at 320?C. The 
column was initially set at 45?C for 5 minutes and then heated up to 300?C at the rate 
of 15?C/min. The hydrogen pressure for the column was fixed at 40 psi, the air 
pressure was 60 psi and the carrier gas helium was 80 psi. 
 
4.2.  Temperature Dependence Test 
For the temperature dependence test the gas flow rate was kept constant at 
0.2L/min. Before the tests the sensors were kept in air in the environmental chamber 
for 7 days for conditioning. At the beginning regulator 1 was opened; gates 1 and 2 
were closed to establish a baseline. When the resistances of the sensors were stable 
regulator 1 was closed and gates 1 and 2 were opened in order to expose the sensors 
to chemical vapor for 30 minutes. Then regulator 1 was opened, gates 1 and 2 were 
closed for the recovery test. When the resistances of the sensors recovered, the 
temperature was changed by the environmental chamber. Consequently, the 
resistances of the sensors changed. When the sensor outputs were stable again, 
another response and recovery test cycle was performed. Test cycles were performed 
at 9 different ambient temperatures (-22?C, -12?C, -2?C, 7?C, 17?C, 22?C, 27?C, 
32?C, 37?C). During the test the temperature was monitored by thermocouple. 
 
 
 
 
 
 
 
32
4.3.  Humidity Dependence Test 
 
Figure 21:  Schematic of gas flow circuit for humidity dependence tests. 
 
Figure 21 illustrates the gas flow circuit for humidity dependence tests. Flask C 
with saturated salt solution was added to provide constant moisture. For humidity 
dependence tests the total gas flow rate was fixed at 0.2L/min. Before the tests the 
sensors were kept in air in the environmental chamber for 7 days for conditioning. To 
establish the base line of the signal, regulator 1 was open; gates 1 and 2 were closed. 
When the resistances of the sensors were stable regulator 1 was closed and gates 1 
and 2 were opened. The sensors are exposed to the chemical vapor for 30 minutes. 
Then regulator 1 was opened and gates 1 and 2 were closed for the recovery test. 
When the resistances of the sensors recovered, the humidity was changed by 
alternating the kind of saturated salt solution in Flask C. Consequently, the resistances 
of the sensors changed. When they were stable again, another response and recovery 
test cycle was performed. The humidity was monitored by the humidity sensor.
 
33
CHAPTER 5:  RESULTS AND DISCUSSION 
 
5.1.  Concentration dependence study 
Response to  MeJa with flow rate 0.02L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(a)
Gas
2
Air
 
Figure 22a:  Sensors? responses to MeJa with partial flow rate 0.02L/Min. 
 
34
Response to  MeJa with flow rate 0.04L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(b)
Gas
2
Air
 
Figure 22b:  Sensors? responses to MeJa with partial flow rate 0.04L/Min. 
 
Response to  MeJa with flow rate 0.06L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(c)
Gas
2
Air
 
Figure 22c:  Sensors? responses to MeJa with partial flow rate 0.06L/Min. 
 
35
Response to  MeJa with flow rate 0.08L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(d)
Gas
2
Air
 
Figure 22d:  Sensors? responses to MeJa with partial flow rate 0.08L/Min. 
 
Response to  MeJa with flow rate 0.10L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(e)
Gas
2
Air
 
Figure 22e:  Sensors? responses to MeJa with partial flow rate 0.10L/Min. 
 
36
Response to  MeJa with flow rate 0.12L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(f)
Gas
2
Air
 
Figure 22f:  Sensors? responses to MeJa with partial flow rate 0.12L/Min. 
 
Response to  MeJa with flow rate 0.16L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(g)
Gas
2
Air
 
Figure 22g:  Sensors? responses to MeJa with partial flow rate 0.16L/Min. 
 
37
Response to  MeJa with flow rate 0.20L/Min
0
10000
20000
30000
40000
50000
60000
01234567891011
Time (hr)
R
e
sistan
ce (Oh
m
)
(h)
Gas
2
Air
 
Figure 22h:  Sensors? responses to MeJa with partial flow rate 0.20L/Min. 
 
Figure 22a to Figure 22h illustrate the resistances change of the sensors to MeJa 
at different partial flow rates. When MeJa is introduced, the resistance of the sensor 
decreases sharply and reaches the lowest value in 4 minutes. Then the resistance 
gradually increased and became stable again. After being exposed to the chemical 
vapor for 30 minutes the sensors are exposed to air flow again. The resistance can 
recover to 80 percent of initial resistance in about 40 minutes and fully recover in 
about 8 hours. Therefore, the sensors have a good reusability. The flat base lines 
indicate a good stability of the sensors. 
To determine actual concentration of MeJa at different partial flow rates, GC was 
used. Six standard solutions were made for the tests with different amounts of Methyl 
Jasmonate (0.1?L, 0.5?L, 1?L, 5?L, 10?L, 100?L), which were dissolved in 5mL 
Dichloromethane. The results are shown in Figure 23. 
 
38
GC Calibration Curve
y = 6.5076x + 0.139
R
2
 = 0.9998
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Volume pencentage (%)
A
r
ea p
en
cen
tag
e 
(%)
 
Figure 23:  GC calibration curve. 
 
The peak area is proportional to the amount chemicals injected into the GC. The 
volume percentage of a chemical can be expressed as, 
?
=
ii
ii
i
Af
Af
V %     (25) 
%
i
V  is the volume percentage of a chemical,  is the response factor of GC,  is 
the peak area of a chemical. In certain ranges the response factor can be assumed to be 
the same for similar chemicals. Hence, the volume percentage becomes proportional 
to the peak area percentage. The high relevance coefficient indicates that the 
assumption here is safe. Using the calibration curve the concentrations of gas samples 
were calculated and are shown in table 3. 
i
f
i
A
 
 
 
 
39
Table 3:  Calculated concentrations of different gas samples 
Flow rates of gas samples (L/Min) Calculated concentration (ppm)
0.02 3.32
0.04 6.58
0.06 9.21
0.08 11.04
0.10 12.84
0.12 14.10
0.16 15.59
0.20 17.46
 
 
Sensors' responses to MeJa at different concentration
y = -0.5647x + 4.6947
R
2
 = 0.991
y = -0.5682x + 4.6441
R
2
 = 0.9824
3.8
3.9
4
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
LogC
L
ogR
In Test Gas
In Air
 
Figure 24:  Sensors? responses to MeJa at different concentration. 
 
Using the concentration obtained by GC, the logarithmic resistances of sensors is 
plotted verses logarithmic concentration as shown in Figure 24. The resistances of the 
sensors in air  are almost constant which indicates a good reusability of the 
sensors. The resistances of the sensors in test gas  decrease with increasing 
concentration of MeJa and show a linear relationship with concentration on a 
a
R
g
R
 
40
logarithmic scale. Therefore, the resistance of the sensor in gas is found to follow the 
expression (26) which is the well known power law [99]. 
?? += LogCLogR
g
 or  (26) 
n
g
KCR =
Where  is the concentration of the gas, C ? ,? , K ,  are constants. n
By defining the relative sensor sensitivity as, S
g
a
R
R
S =     (27) 
it is found that the relative sensor sensitivity increases from about 2 to 5 with 
increasing concentration of MeJa and follows the power law (Figure 25). 
Relative sensor sensitivity S to MeJa at different concentration
y = 0.5654x + 0.0396
R
2
 = 0.9846
y = 0.56x + 0.0226
R
2
 = 0.992
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
LogC
Lo
g
S
 
Figure 25:  Relative sensor sensitivity S to MeJa at different concentration 
 
 
 
 
 
 
41
5.2.  Temperature dependence study 
Responses of sensor 1 to MeJa at different temperatures
0
10000
20000
30000
40000
50000
60000
70000
80000
2 4 6 8 10121416182
Time (hr)
Re
si
st
a
n
c
e
 (
O
h
m
)
Relative Humidity=0=Constant
0
Temperature= ?2?C
Temperature= ?22?C
Temperature= ?12?C
Temperature=17?C
Temperature=7?C
Temperature=37?C
Temperature=32?C
Temperature=27?C
Temperature=22?C
 
Figure 26a:  Responses of sensor 1 to MeJa at different temperatures 
 
Responses of sensor 2 to MeJa at different temperatures
0
10000
20000
30000
40000
50000
60000
70000
80000
2 4 6 8 10121416182
Time (hr)
Re
si
st
a
n
c
e
 (
O
h
m
)
Relative Humidity=0=Constant
0
Temperature= ?2?C
Temperature= ?22?C
Temperature= ?12?C
Temperature=17?C
Temperature=7?C
Temperature=37?C
Temperature=32?C
Temperature=27?C
Temperature=22?C
 
Figure 26b:  Responses of sensor 2 to MeJa at different temperatures 
 
 
42
Figure 26a and Figure 26b show the responses of two sensors at different 
temperatures. The flat initial resistances of the sensors and the ability to recover to the 
initial resistances indicate good stability and reusability at different temperatures. 
With increasing ambient temperature, the sensor resistances in air solely decrease, and 
the sensor resistances in test gas first increase then decrease as shown in Figure 27. 
Sensors' resistances at different ambient temperatures
0
10000
20000
30000
40000
50000
60000
70000
80000
-30 -20 -10 0 10 20 30 40
Ambient Temperature (?C)
R
e
sistan
ce (Oh
m
)
Resistance in air
Resistance in test gas
 
Figure 27:  Sensors? resistances at different ambient temperatures. 
 
Hence, the sensors have relatively low relative sensitivity S at around 20?C about 4 to 
5, and S decreases with increasing temperature (Figure 28). 
 
43
Relative sensor sensitivity S at different abmient temperature
0
1
2
3
4
5
6
7
8
9
-30 -20 -10 0 10 20 30 40
Ambient Temperature (?C)
S
 
Figure 28:  Relative sensor sensitivity S at different ambient temperatures. 
 
According to the report of Figaro Inc, the TGS 2620 sensor is heated by an 
integrated heater whose working voltage and current is 5V and 42mA respectively. 
Thus, the working temperature of the sensing element is elevated to about 400?C 
when the sensor is working at the ambient temperature of 20?C. For practical in field 
use, the sensor probably will work in the temperature range from -30?C to 40?C. 
Hence, the sensor is tested at different temperatures within that range. Jian-wei et al 
reports that the ambient temperature has a linear superposition effect on sensor?s 
working temperature as shown in Figure 29 [100]. Hence, during the temperature 
dependence test, the temperature of the sensing element is estimated to be in the range 
from 350?C to 420?C. 
 
44
 
Figure 29:  Ambient temperature effect on sensors? working temperature as a linear 
superposition [100] 
 
According to the band bending theory mentioned in Chapter 2 the resistance of 
the tin oxide sensor can be expressed as [98], 
)exp(
kT
eV
RR
S
O
=     (28) 
where  is the flat band resistance. The schottky potential barrier, , in air is 
mainly controlled by the oxygen partial pressure and types of oxygen species 
adsorbed on the sensor surface [101]. According to the survey by Barsan and Weimar 
as shown in Figure 30 [102] in the testing temperature range  is dominant and has 
no significant change in types of oxygen species adsorbed on the sensor surface. 
Hence, there should not be a large change in Schottky potential barrier, , as 
oxygen partial pressure can be regarded as constant at different temperatures during 
O
R
S
V
?
O
S
V
 
45
the test. 
 
Figure 30:  Survey of oxygen species detected at different temperatures at SnO2 
surfaces with IR (infrared analysis), TPD (temperature programmed desorption), EPR 
(electron paramagnetic resonance) [102]. 
y = 2921x + 6.5997
R
2
 = 0.9977
y = 2471.4x + 7.2068
R
2
 = 0.9962
10.75
10.8
10.85
10.9
10.95
11
11.05
11.1
11.15
11.2
11.25
11.3
0.00144 0.00146 0.00148 0.0015 0.00152 0.00154 0.00156 0.00158 0.0016
1/T
Ln
R
a
 
Figure 31:  Sensors? logarithmic resistances in dry air at different temperatures. 
 
46
By plotting the logarithmic resistances of the sensors in air verses inverse temperature 
of the sensor surface as shown in Figure 31, a linear relationship is obtained which is 
consistent with the theory. 
The Schottky potential barrier, , in reducing gases is very complex which is 
affected by adsorption and desorption rates of oxygen species and/or reducing gases 
or even reaction products, reaction rates between adsorbed oxygen species and 
reducing gases, charge carrier concentration and Debye length in the semiconductor 
[101]. The sensors? resistances are found to increase below 10?C, be almost constant 
between 10?C and 30?C and decrease above 30?C (Figure 27). It can be inferred that 
there are different mechanisms controlling the sensor resistance behavior in the 
testing temperature range. 
S
V
 
5.3.  Humidity dependence study 
Sensor response at 20%RH
0
5000
10000
15000
20000
25000
30000
35000
0 5 10 15 20 25
Time (hr)
R
e
sistan
ce (Oh
m
)
10
15
20
25
30
35
40
Relativ
e h
u
m
id
ity
 (
%
)
(a)
Gas
Humid Air
Temperature=25?C=Constant
 
Figure 32a:  Sensors? responses at relative humidity 20%. 
 
 
47
Sensor response at 64%RH
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1020304050
Time (hr)
R
e
sistan
ce (Oh
m
)
50
55
60
65
70
75
80
85
90
Relat
iv
e
 Hu
m
id
ity
 (
%
)
(b)
Gas
Humid Air
Temperature=25?C=Constant
 
Figure 32b:  Sensors? responses at relative humidity 64%. 
 
Sensor response at 70%RH
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
0 5 10 15 20 25 30 35 40 45 50
Time (hr)
R
e
sistan
ce (Oh
m
)
50
55
60
65
70
75
80
85
90
Relat
iv
e
 Hu
m
id
ity
 (
%
)
(c)
Gas
Humid Air
Temperature=25?C=Constant
 
Figure 32c:  Sensors? responses at relative humidity 70%. 
 
 
48
Response at 82%RH
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
0 5 10 15 20 25 30 35 40 45
Time (hr)
R
e
sistan
ce (Oh
m
)
60
65
70
75
80
85
90
95
100
105
110
Relat
iv
e
 Hu
m
id
ity
 (
%
)
(d)
Humid Air
Gas
Temperature=25?C=Constant
 
Figure 32d:  Sensors? responses at relative humidity 82%. 
 
Sensor response at 94%RH
0
2000
4000
6000
8000
10000
12000
14000
16000
0 5 10 15 20 25 30 35 40 45
Time (hr)
R
e
sistan
ce (Oh
m
)
80
85
90
95
100
105
110
115
120
Relativ
e h
u
m
id
ity
 (
%
)
(e)
Gas
Humid Air
Temperature=25?C=Constant
 
Figure 32e:  Sensors? responses at relative humidity 94%. 
 
    Figure 32a to Figure 32e illustrate the sensors? responses at different relative 
 
49
humidities. The thick line in the lower part of the figure is the relative humidity 
detected by the humidity sensor. The two thin lines are the resistances of two sensors. 
The flat initial resistances of the sensors and the ability to recover to initial resistances 
show good stability and reusability of the sensor at different relative humidities. The 
water vapor reduced the resistance of the sensor compared with in dry air. The sensors 
still respond to MeJa quickly and reach the lowest value in 4 minutes however they 
recover much slower than in dry air. They need more than 30 hours to recover to 
initial resistances in high relative humidity. The relative sensor response S is found to 
be decreasing with increasing relative humidity and decrease faster in high relative 
humidity (Figure 33). 
Relative sensor response S at different relative humidities
3
3.2
3.4
3.6
3.8
4
4.2
0 10203040506070809010
Relative Humidity (%)
S
 
Figure 33:  Relative sensor sensitivity S at different relative humidities. 
 
By plotting the resistances of the sensors verses water vapor concentration in air 
on a logarithmic scale it was found that the sensors? resistances in humid air also 
follow the power law (Figure 34). 
 
50
Resistances at different water vapor concentration
y = -0.4887x + 6.3404
R
2
 = 0.9992
y = -0.481x + 6.2933
R
2
 = 0.9991
4.1
4.15
4.2
4.25
4.3
4.35
4.4
4.45
4.5
4.55
3.73.83.9 4 4.14.24.34.44
LogC
L
ogR
.5
 
Figure 34:  Sensors? initial resistances at different relative humidity. 
 
Sensors' resistances in the presence of MeJa at different relative humidities
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0 10203040506070809010
Relatvie Humidity (%)
R
e
sistan
ce (Oh
m
)
 
Figure 35:  Sensors? resistances in the presence of MeJa at different relative 
humidities. 
 
 
51
 
52
In the presence of MeJa the sensors? resistances decrease with increasing relative 
humidity but don?t show the power law behavior (Figure 35). This means that water 
vapor contribution and the MeJa contribution to the reduction of the sensors? 
resistances are not independent of each other. Otherwise, the sensors? resistances at 
different relative humidity in the presence of MeJa should also follow the power law 
as the concentration of MeJa is fixed. However, it is not supported by the data. The 
reason for this is that the water vapor is competing with the reducing gases for 
preadsorbed oxygen species under different adsorption?desorption equilibrium 
conditions [103]. This also explains why the relative sensor response is decreasing 
with increasing relative humidity. 
When the sensors are used in the field they can be distributed in different area. 
First, the baselines of the sensors are established at certain ambient temperature and 
humidity. When the ambient temperature and humidity changes the resistances of the 
sensors will gradually changes to corresponding status in a similar shape. If insects 
attack plants in some area, the volatile compounds which are emitted by plants change. 
The concentration of MeJa?an important defense system related chemical increases 
significantly in a short time. Hence, the resistances of the sensors in corresponding 
area should decrease quickly and have different resistance change shape from the 
sensors in other area. Thus, it can be used as a qualitative alarm for the insect 
infestation. The interfering effect of other volatile compounds is not tested this time 
and will be performed in future study. 
CHAPTER 6:  CONCLUSIONS 
 
Thick film hotplate type tin oxide sensors can successfully detect Methyl 
Jasmonate and show good stability and reversibility at different ambient temperature 
and relative humidity. The sensors have a quick response to MeJa but the recover time 
of the sensor is very long, about 8 hours in dry air and even longer?more than 30 
hours in high relative humidity. In the presence of MeJa from about 3ppm to 18ppm 
the relative sensor sensitivity S increases from 2 to 5 in dry air at ambient temperature 
25?C and follow the power law. Sensors? resistances in dry air flow at different 
temperatures are proportional to 1/T on a logarithmic scale. Sensors? resistances in the 
presence of MeJa at different temperature in dry air flow are controlled by different 
dominant mechanisms in different testing temperature ranges. The relative sensor 
sensitivity in dry air flow is relatively low about 4 around ambient temperature 20?C 
and increases with decreasing ambient temperature below 10?C. Sensors? resistances 
in humid air flow at 25?C decrease with increasing water vapor concentration and also 
follow the power law. The relative sensor sensitivity decreases with increasing 
relative humidity and is still over 3 in high humidity. 
By distributing the sensors in different area of the field, concentration change of 
MeJa can be qualitatively detected and discriminated from temperature and humidity 
changes. Hence, thick film type hotplate tin oxide sensor has great potential to be used 
as an in-situ qualitative alarm to the infestation of insects in site. 
 
53
REFERENCES 
[1] Chatterjee, Samprit, Biometrics 29 (1973), 727-734 
[2] Massimo E. Maffei, Axel Mithofer, Wilhelm Boland, TRENDS in Plant Science 
12 (2007) 7 310-316 
[3] Ebel, J., Mithofer, Planta 206 (1998) 335?348 
[4] White, P.J., BBA-Biomembranes 1465 (2000) 171?189 
[5] Foyer, C.H., Noctor, G., Plant Cell Environ. 28 (2005) 1056?1071 
[6] Asada, K., Plant Physiol, 141 (2006) 391?396 
[7] Orozco-Ca?rdenas, M., Ryan, C.A., Proc. Natl. Acad. Sci. USA 96 (1999) 
6553?6557 
[8] Bennett, R.N. and Wallsgrove, R.M., New Physiologist 127 (1994)617-633 
[9] Theis, N., Lerdau, M., International Journal of Plant Science 164 (2003) S93-S102 
[10] Mao, J., Burt, A.J., Ramputh, A.-L., Simmons, J., Cass, L., Hubbard, K., Miller, 
S., Altosaar, I., Arnason, J.T., Journal of Agriculture and Food Chemistry 55 (2007) 
2582-2589 
[11] Jette T. Knudsen, Roger Eriksson, Jonathan Gershenzon, Bertil St?hl, The 
Botanical Review 72(1) (2006) 1-120 
[12] Natalia Dudareva, Florence Negre, Dinesh A. Nagegowda and Irina Orlova, 
Critical Reviews in Plant Sciences 25 (2006) 417?440 
[13] Bernasconi, M. L., Turlings, T. C. J., Ambrosetti, L., Bassetti, P., Dorn S., 
Rhopalosiphum maidis. Entomol. Exp. Appl. 87 (1998) 133?142 
[14] De Moraes, C. M., Mescheer, M. C., Tumlinson, J. H., Nature 410 (2001) 577?580
 
54
[15] Price, P. W., C. E. Bouton, P. Gross, B. A. McPheron, J. N. Thompson, A. E. 
Weis, Annual Review of Ecology and Systematics 11 (1980) 41?65 
[16] Takabayashi, J., Dicke, M., Trends Plant Sci. 1 (1996) 109?113 
[17] Dicke, M., Entomologia Experimentalis et Applicata 92 (1999b) 131?142 
[18] Anderson, P., Alborn, H. Entomol. Exp. Appl. 92 (1999) 45?51 
[19] Hilker, M., Meiners, T. Entomol. Exp. Appl. 104 (2002) 181?192 
[20] Baldwin, I. T. & J. C. Schultz, Science 221 (1983) 277?279 
[21] Rhoades, D. F., Recent Advances in Phytochemistry 19 (1985)195?218. 
[22] Dicke, M., Abelis, M. W., Takabayashi, J., Bruin, J., Posthumus, M. A., J. Chem. 
Ecol. 16 (1990) 3091?3117. 
[23] Arimura, G., Ozawa, R., Nishioka, T., Boland, W., Koch, T., Kuhnemann, F., 
Takabayashi, J., Plant J. 29 (2002) 87?98. 
[24]Arimura, G., Ozawa, R., Kugimiya, S., Takabayashi, J., Bohlmann, J., Plant 
Physiol. 135 (2004b) 1976?1983 
[25] Ruther, J., and Kleier, S., J. Chem. Ecol. 31 (2005) 2217?2222 
[26] Green, T.R., Ryan, C.A. Science 175 (1972) 776-777 
[27] Ussuf, K.K., Laxmi, N.H. and Mitra, R., Current Science, 80 (2001) 847-853 
[28] Rawlings, N.D., Tolle, D.P. and Barrett, A.J., Biochemical Journal 378 (2005) 
705-716 
[29] Birgit Schulze, Ryan Lauchli, Mesmin Mekem Sonwa, Annika Schmidt, 
Wilhelm Boland, Analytical Biochemistry 348 (2006) 269?283 
[30] Farmer EE, Johnson RR, Ryan CA. Plant Physiol 98 (1992) 995?1002 
[31] Li L, Li C, Lee GI, Howe GA. 2002a. Proc Natl Acad Sci USA 99 (2002a) 
6416?6421 
[32] Fujita M, Fujita Y, Noutoshi Y, Takahashi F, Narusaka Y, Yamaguchi, Shinozaki 
 
55
K, Shinozaki K, Curr Opin Plant Biol 9 (2006) 436-442 
[33] Rojo E, Solano R, Sanchez-Serrano JJ, Plant Growth Regul 22 (2003) 82-98 
[34] Tiryaki I, Staswick PE, Plant Physiol 130 (2002) 887-894 
[35] Spoel SH, Koornneef A, Ctaessens SMC, Korzelus JP, Van Pelt JA, Mueller M J, 
Buchala A J, Metraux JP, Brown R, Kazzan K, Van Loon LC, Dong XN, Pieterse 
CMJ.. Plant Cell 15 (2003) 760-770 
[36] Li J, Brader G, Palva ET., Plant Cell 16 (2004) 319-331 
[37] Anderson JP, Badruzsaufari E, Schenk PM, Manners M, Desmond OJ, Ehlert C, 
Maclean D], Ebert PR, Kazan K, Pant Cell 16 (2004) 3460-3479 
[38] Seo HS, Song JT, Cheong JJ, Lee Y-H, Lee YW, Hwang I, Lee JS, Choi YD 
Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 4788?4793 
[39] Jong-Joo Cheong and Yang Do Choi, TRENDS in Genetics 19 (2003) 409-413 
[40] R. Karban, I.T. Baldwin, K.J. Baxter, G. Laue, G.W. Felton, Oecologia 125 
(2000) 66?71 
[41] Wasternack C, Hause B, Progr Nucl Acid Res Mol Biol 72 (2002) 165-221 
[42] Riaan Meyer, George F. Rautenbach and Ian A. Dubery, Phytochemical. 
Analysis 14 (2003) 155?159 
[43] N. Butt& L. Cinquegrani, E. Mugno, A. Tagliente and S. Pizzini, sensors and 
Actuators B, 6 (1992) 253-256 
[44] Geraint Williams and Gary S. V. Coles, Sensors and Actuators 8, 15-16 (1993) 
349-353 
[45] Williams, E.W. Lawlor, C.M. Keeling, A.G. Gould, R.D., International Journal 
of Electronics, Vol. 76 Issue 5, May (1994) 815 
[46] Ulrich Hoefer, Gerd K?hner, Werner Schweizer, Gerd Sulz, Klaus Steiner, 
Sensors and Actuators B 22 (1994) 115-119 
 
56
[47] M. Di Giulio, G. Micocci, R. Rella, p. Siciliano, A. Tepore, Sensors and 
Actuators B 23 (1995) 193-195 
[48] Ping Ping Tsai, I-Cherng Chen, Ming Hann Tzeng, Sensors and Actuators B 
24-25 (1995) 537-539 
[49] Klaus Steiner, Ulrich Hoefer, Gerd Kiihner, Gerd Sulz, Elmar Wagner, Sensors 
and Actuators B 24-25 (1995) 529-531 
[50] Kiyoshi Fukui, Masanori Nakane, Sensors and Actuators B 24-25 (1995) 
486-490 
[51] Tadashi Mochida, Kei Kikuchi, Takehiko Kondo, Hironobu Ueno, Yoshinobu 
Matsuura, Sensors and Actuators B 24-25 (1995) 433-437 
[52] G. Sarala Devi, S. Manorama, V.J. Rao, Sensors and Actuators B 28 (1995) 
31-37 
[53] W.K. Choi, S.K. Song, J.S. Cho, Y.S. Yoon, D. Choi, H., J. Jung, S.K. Koh, 
Sensors and Actuators B 40 (1997) 21-27 
[54] Sang woo Lee, Ping ping Tsai, Haydn Chen, Sensors and Actuators B 41 (1997) 
55-61 
[55] V.A. Chaudhary, I.S. Mulla, S.R. Sainkar, A.A. Belhekar, K. Vijayamohanan, 
Sensors and Actuators A 65 (1998) 197-202 
[56] Jae Chang Kim, Hee Kwon Jun, Jeung-Soo Huh, Duk Dong Lee, Sensors and 
Actuators B 45 (1997) 271?277 
[57] V.N. Mishra, R.P. Agarwal, Microelectronics Journal 29 (1998) 861-874 
[58] Zhihong Jin, Huan-Jun Zhou, Zhang-Li Jin, Robert F. Savinell, Chung-Chiun Liu, 
Sensors and Actuators B 52 (1998) 188?194 
[59] A. Katsuki, K. Fukui, Sensors and Actuators B 52 (1998) 30?37 
[60] A. Chaturvedi, V.N. Mishra, R. Dwivedi, S.K. Srivastava, Microelectronics 
 
57
Journal 30 (1999) 259?264 
[61] V.A. Chaudhary, I.S. Mulla, K. Vijayamohanan, Sensors and Actuators B 55 
(1999) 154?160 
[62] M.C. Horrillo, I. Sayago, L. Are?s, J. Rodrigo, J. Gutie?rrez, A.Go?tz, I. Gra`cia, 
Fonseca, C. Cane, E. Lora-Tamayo, Sensors and Actuators B 58 (1999) 325?329 
[63] Go Sakai, Nam Seok Baik, Norio Miura, Noboru Yamazoe, Sensors and 
Actuators B 77 (2001) 116-121 
[64] Yu-De Wang, Chun-Lai Ma, Xing-Hui Wu, Xiao-Dan Sun, Heng-De Li, Sensors 
and Actuators B: Chemical 85 (2002)270-276 
[65] K. Anothainart, M. Burgmair, A. Karthigeyan, M. Zimmer, I. Eisele, Sensors and 
Actuators B 93 (2003) 580?584 
[66] S. Shukla, S. Patil, S.C. Kuiry, Z. Rahman, T. Du, L. Ludwig, C. Parish, S. Seal, 
Sensors and Actuators B 96 (2003) 343?353 
[67] Behzad Esfandyarpour, Shams Mohajerzadeh, Abbas Ali Khodadadi, and 
Michael D. Robertson, IEEE SENSORS JOURNAL, 4 (2004) 449-454 
[68] S.M.A. Durrani, E.E. Khawaja, M.F. Al-Kuhaili, Talanta 65 (2005) 1162?1167 
[69] U lo Kersen, Lauri Holappa, Analytica Chimica Acta 562 (2006) 110?114 
[70] Satyajit Shukla, Peng Zhang, Hyoung J. Cho, Sudipta Seal, Lawrence Ludwig, 
Sensors and Actuators B 120 (2007) 573?583 
[71] Yun-Hyuk Choi, Seong-Hyeon Hong, Sensors and Actuators B 125 (2007) 
504?509 
[72] Chang Sup Moon, Hae-Ryong Kim, Graeme Auchterlonie, John Drennan, 
Jong-Heun Lee, Sensors and Actuators B 131 (2008) 556?564 
[73] Jun Zhang, ShurongWang, YanmeiWang, YanWang, Baolin Zhu, Huijuan Xia, 
Xianzhi Guo, Shoumin Zhang,Weiping Huang, ShihuaWu, Sensors and Actuators B 
 
58
135 (2009) 610?617 
[74] Cesar Elosua, Ignacio R. Matias, Candido Bariain and Francisco J. Arregui, 
Sensors, 6 (2006)1440-1465 
[75] Alexander P. Demchenko, Introduction to Fluorescence Sensing, Springer (2009) 
[76] Jiri Homola, Sinclair S. Yee a, Gunter Gauglitz, Sensors and Actuators B 54 
(1999) 3?15 
[77] P. T. Moseley and B. C. Tofield, Solid State Gas Sensors, (1987) Adam Hilger, 
Bristol 
[78] Deng Y, Nevell T G, Ewen R J, Honeybourne C L, Appl Catal A-Gen, 101 (1993) 
51 
[79] L. A. Obvintseva,, Russian Journal of General Chemistry, 78, (2008) 2545?2555 
[80] Hua Bai and Gaoquan Shi, Sensors, 7 (2007) 267-307 
[81] P. Pasierb, M. Rekas, J Solid State Electrochem, 3 (2009) 3?25 
[82] C. O. Park, J. W. Fergus , N. Miura, Jinsu Park, Angi Choi, Ionics, 15 (2009) 
261?284 
[83] Stetter, J. R.; Chang, S. C.; Cha, C. S. Talanta, 40, (1993) 461-477 
[84] David James, Simon M. Scott, Zulfiqur Ali, William T. O?Hare, Microchim. 
Acta 149, (2005) 1?17 
[85] Ralf Lucklum . Peter Hauptmann, Anal Bioanal Chem, 384 (2006) 667?682 
[86] P.T. Moseley, Meas. Sci. Technol. 8 (1997) 223?237 
[87] S. Capone, A. Forleo, L. Francioso, R. Rella, P. Siciliano, J. Spadavecchia, D.S. 
Presicce, A.M. Taurino, J. Optoelectr. Adv. Mater. 5 (5) (2003) 1335?1348 
[88] G. Korotcenkov, Materials Science and Engineering B 139 (2007) 1?23 
[89] C. O. Park, S. A. Akbar, Journal of Materials Science 38 (2003) 4611 ? 4637 
[90] Matthias Batzill, Ulrike Diebold, Progress in Surface Science 79 (2005) 47?154 
 
59
 
60
[91] P . B. WEISZ, J. Chem. Phys. 21 (1953) 1531 
[92] P. K. Clifford, D. T. Tuma, Sensors and Actuators, 3 (1983) 233-254. 
[93] J Watson, Sensors and Actuators, 5 (1984) 29 ? 42 
[94] Shigeki Hirobayashi, Mohammed Afrose Kadir, Toshio Yoshizawa, Tatsuo 
Yamabuchi, Sensors and Actuators B 92 (2003) 269?278 
[95] Williams D E 1987 Conduction and gas response of semiconductor gas sensors 
in Solid State Gas Sensors ed P T Moseley and B C Tofield (Bristol: Adam Hilger) 71 
[96] G. Sberveglieri, S. Groppelli, P. nelli?Sensors and Actuators B, 4 (1991) 457-461 
[97] Rajeev K. Srivastava, P. Lal, R. Dwivedi, SK. Srivastava, Sensors and Actuators 
B 21 (1994) 213-218 
[98] G. Martinelli, M.C. Carotta, Sensors and Actuators B 23 (1995) 157-161 
[99] Noboru Yamazoe, Kengo Shimanoe, Sensors and Actuators B 128 (2008) 
566?573 
[100] Jian-Wei Gong, Quan-Fang Chen, Ming-Ren Lian, Nen-Chin Liu, Claude 
Daoust, IEEE SENSORS JOURNAL, 6 (2006) 139-145 
[101] Andrew P. Lee, Brian J. Reedy, Sensors and Actuators B 60 (1999) 35?42 
[102] Nicolae Barsan, Udo Weimar, Journal of Electroceramics, 7 (2001) 143?167 
[103] Dorota Koziej, Nicolae Ba?rsan, Udo Weimar, Jacek Szuber, Kengo Shimanoe, 
Noboru Yamazoe, Chemical Physics Letters 410 (2005) 321?323