USF Libraries
USF Digital Collections

Salinity (conductivity) sensor based on parallel plate capacitors

MISSING IMAGE

Material Information

Title:
Salinity (conductivity) sensor based on parallel plate capacitors
Physical Description:
Book
Language:
English
Creator:
Bhat, Shreyas
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Impedance
Oscillator
Dielectric
Polarization
Complex permittivity
Dissertations, Academic -- Electrical Engineering -- Masters -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This work is aimed at developing a high sensitivity salinity (conductivity) sensor for marine applications. The principle of sensing involves the use of parallel plate capacitors, which minimizes the proximity effects associated with inductive measurement techniques. The barrier properties of two different materials, AZ5214 and Honeywell's ACCUFLO T3027, were investigated for use as the insulation layer for the sensor. Impedance analysis performed on the two coatings using Agilent's 4924A Precision Impedance Analyzer served to prove that ACCUFLO was a better dielectric material for this application when compared to AZ5214.Two separate detection circuits have been proposed for the salinity sensor. In the Twin-T filter method, a variation in capacitance tends to shift the resonant frequency of a Twin-T oscillator, comprising the sensor. Simulations of the oscillator circuit were performed using Pspice. Experiments were performed on calibrated ocean water samples of 34.996 psu and a shift of 410 Hz/psu was obtained. To avoid the problems associated with the frequency drift in the oscillator, an alternate detection scheme is proposed which employs frequency-to-voltage converters. The sensitivity of this detection scheme was observed to be 10 mV/psu.
Thesis:
Thesis (M.S.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Shreyas Bhat.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 82 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001912789
oclc - 174112815
usfldc doi - E14-SFE0001381
usfldc handle - e14.1381
System ID:
SFS0025701:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001912789
003 fts
005 20071012090634.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 071012s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001381
040
FHM
c FHM
035
(OCoLC)174112815
049
FHMM
090
TK145 (ONLINE)
1 100
Bhat, Shreyas.
0 245
Salinity (conductivity) sensor based on parallel plate capacitors
h [electronic resource] /
by Shreyas Bhat.
260
[Tampa, Fla] :
b University of South Florida,
2005.
3 520
ABSTRACT: This work is aimed at developing a high sensitivity salinity (conductivity) sensor for marine applications. The principle of sensing involves the use of parallel plate capacitors, which minimizes the proximity effects associated with inductive measurement techniques. The barrier properties of two different materials, AZ5214 and Honeywell's ACCUFLO T3027, were investigated for use as the insulation layer for the sensor. Impedance analysis performed on the two coatings using Agilent's 4924A Precision Impedance Analyzer served to prove that ACCUFLO was a better dielectric material for this application when compared to AZ5214.Two separate detection circuits have been proposed for the salinity sensor. In the Twin-T filter method, a variation in capacitance tends to shift the resonant frequency of a Twin-T oscillator, comprising the sensor. Simulations of the oscillator circuit were performed using Pspice. Experiments were performed on calibrated ocean water samples of 34.996 psu and a shift of 410 Hz/psu was obtained. To avoid the problems associated with the frequency drift in the oscillator, an alternate detection scheme is proposed which employs frequency-to-voltage converters. The sensitivity of this detection scheme was observed to be 10 mV/psu.
502
Thesis (M.S.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 82 pages.
590
Adviser: Shekhar Bhansali, Ph.D.
653
Impedance.
Oscillator.
Dielectric.
Polarization.
Complex permittivity.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1381



PAGE 1

Salinity (Conductivity) Sensor Base d on Parallel Plate Capacitors by Shreyas Bhat A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Shekhar Bhansali, Ph.D. Larry Langebrake, P.E. Rudy Schlaf, Ph.D. Babu Joseph, Ph.D. Kendra Daly, Ph.D. Date of Approval: October 22, 2005 Keywords: Impedance, oscillator, dielect ric, polarization, complex permittivity Copyright 2005, Shreyas Bhat

PAGE 2

DEDICATION This dissertation is dedicated to my parent s, Gopal Krishna Bhat and Sulochana Bhat, my brother, Suraj and sister, Naina. Their love, support and patienc e have always been the inspiration of my life.

PAGE 3

ACKNOWLEDGEMENTS I wish to express sincere appr eciation to my major professo r Dr. Shekhar Bhansali for his valuable guidance and support during this thesis work. I also wish to extend my gratitude to Mr. Larry Langebrake for allowing me to be a part of this presti gious project. Thanks to Dr. Rudy Schlaf, Dr. Babu Joseph and Dr. Kendra Daly for agreeing to serve on my supervisory committee. I would like to thank Kevin, Shyam, Abdur and Saravana for assist ing me with various aspects of the project. I also appreciate the support of Dr. Paris Wiley of the Electrical Engineering Department for his help with circuit analysis. Special thanks are extended to the BioMEMS and Microsystems Research Group and my friends for their support and enc ouragement throughout this research.

PAGE 4

i TABLE OF CONTENTS LIST OF TABLES...............................................................................................................v LIST OF FIGURES...........................................................................................................vi ABSTRACT....................................................................................................................... ix CHAPTER 1: INTRODUCTION........................................................................................1 1.1 Definitions of Salinity..............................................................................................2 1.2 Composition of Seawater.........................................................................................3 1.3 Processes Influencing Ocean Salinity......................................................................4 1.3.1 Processes Decreasing Salinity.........................................................................4 1.3.2 Processes Increasing Salinity..........................................................................5 1.4 Need for Salinity Sensing........................................................................................5 1.4.1 Ocean Circulation...........................................................................................6 1.4.2 Marine Life.....................................................................................................7 1.4.3 Water Cycle....................................................................................................9 1.5 Motivation................................................................................................................9 1.6 Thesis Outline .......................................................................................................10 CHAPTER 2: LITERATURE REVIEW...........................................................................11 2.1 Conductivity...........................................................................................................11 2.2 Practical Salinity Scale (PSS) of 1978...................................................................12

PAGE 5

ii 2.3 Current Approaches at Conductivity Sensing........................................................13 2.3.1 Contact Type Sensors...................................................................................13 2.3.1.1 2-Probe Conductivity Sensors...........................................................13 2.3.1.2 4-Probe Conductivity Sensors...........................................................15 2.3.2 Non-Contact Type Sensors...........................................................................15 2.3.2.1 Single Transformer...........................................................................16 2.3.2.2 Double Transformer..........................................................................18 2.4 IDT Configuration in M easurement of Conductivity............................................22 CHAPTER 3: THEORY AND UNDERSTANDING.......................................................25 3.1 Fundamentals of a Capacitor.................................................................................25 3.2 Complex Permittivity and Dielectric Loss.............................................................26 3.3 Polarization of Materials........................................................................................27 3.4 Types of Polarization.............................................................................................29 3.4.1 Interface Polarization....................................................................................29 3.4.2 Electronic and Atomic Polarization..............................................................29 3.4.3 Ionic Polarization..........................................................................................30 3.4.4 Orientation (Di polar) Polarization................................................................30 3.5 Contribution of Polarization to Permittivity..........................................................31 3.6 Interfacial or Space Charge Polarization...............................................................32 3.7 Relaxation Time.....................................................................................................33 3.8 Debye Theory of Dielectric Behavior....................................................................34 3.9 Cole-Cole Diagram for Complex Permittivity.......................................................35 3.10 Dependence of Permittivity on the Concentration of Analyte...............................36

PAGE 6

iii 3.11 Temperature Dependence of Permittivity..............................................................37 3.12 Assessment of Barrier Properties of Polymeric Coatings on the Electrodes.........38 3.12.1 Impedance Spectroscopy............................................................................39 3.12.2 Double Charge Layer..................................................................................42 CHAPTER 4: DESIGN, OPTIMI ZATION AND CONSTRUCTION.............................44 4.1 Minimization of Fringe Fiel d Effects Using Guard Rings.....................................44 4.2 Detection Circuit Employing Twin-T Oscillator ..................................................49 4.3 Construction of the Sensor.....................................................................................53 4.4 Equivalent Circuit Model for the Sensor...............................................................56 4.5 Alternate Detection Circuit for Capacitive Salinity Sensor...................................57 CHAPTER 5: RESULTS...................................................................................................61 5.1 AZ5214 as the Insulation Layer.............................................................................61 5.1.1 Impedance Data for Va rying Salinities at 22 C...........................................62 5.1.2 Impedance Data for Varying Temp eratures at Constant Salinity.................65 5.1.3 Twin-T Oscillator Frequency Versus Salinity..............................................67 5.1.4 Twin-T Oscillator Frequency Versus Temperature......................................68 5.2 ACCUFLO Spin-on-Polymer as the Insulation Layer...........................................69 5.2.1 Impedance Data for Va rying Salinities at 22 C...........................................69 5.2.2 Impedance Data for Varying Temp eratures at Constant Salinity.................72 5.2.3 Twin-T Oscillator Response with Spin-on-Polymer as Insulation Layer.....73 5.2.4 Twin-T Oscillator Frequency Versus Temperature......................................74 5.3 Response of F-V Converter Circuit.......................................................................74 5.3.1 Output Voltage Versus Salinity at 22 C......................................................75

PAGE 7

iv 5.3.2 Output Voltage Versus Temp erature at Constant Salinity............................75 5.4 Comparative Study.................................................................................................76 CHAPTER 6: CONCLUSI ON AND FUTURE WORK...................................................77 6.1 Conclusion.............................................................................................................77 6.2 Future Work...........................................................................................................78 REFERENCES..................................................................................................................80

PAGE 8

v LIST OF TABLES Table 5.1: Parameters for the Randels Ce ll Obtained from the Nyquist Plot..................63 Table 5.2: Randels Cell Parameters for Varying Temperatures at Constant Salinity (34.821 psu) .......................................................................................66 Table 5.3: Comparison Chart for Materials Us ed and Detection Techniques Proposed .76

PAGE 9

vi LIST OF FIGURES Figure 1.1: Major Constituents of Seawater......................................................................3 Figure 1.2: Energy Transfer through the Conveyor Belt...................................................7 Figure 2.1: 2-Probe Conductivity Sensor........................................................................13 Figure 2.2: 4-Probe Conductivity Sensor........................................................................15 Figure 2.3: Single Transformer Conductivity Sensor Shown with Seawater as the Surrounding Media.................................................................................16 Figure 2.4: Equivalent Circuit for Single Transformer Design.......................................17 Figure 2.5: Schematic for a Double Transformer Conductivity Sensor..........................19 Figure 2.6: Capacitor Configurations: (a) Cylindrical Set-Up, (b) IDTs .......................22 Figure 2.7: IDT Conductivity Cell and Equivalent Circuit.............................................22 Figure 3.1: Parallel Plate Capacitor Configuration.........................................................25 Figure 3.2: Loss Tangent Vector Diagram......................................................................27 Figure 3.3: Polarization Phenome non in Dielectric Molecules.......................................28 Figure 3.4: (a) Dipole Rotation in Applied Fi eld, (b) Macroscopic Representation of Orientation Polarization Under the Influence of an Applied Electric Field..............................................................................................................30 Figure 3.5: Frequency Response of Various Polarization Mechanisms..........................31 Figure 3.6: Transition in Polarization with Sudden Field Drop......................................33 Figure 3.7: Debye Relaxation of Water at 30 C.............................................................34

PAGE 10

vii Figure 3.8: Cole-Cole Representation of the Debye Model of Water at 30 C...............35 Figure 3.9: Hydration Process in a Solute-Solvent Mixture............................................37 Figure 3.10: Typical Nyquist Plot of the Complex Impedance.........................................41 Figure 3.11: Randels Cell Circuit Model.........................................................................41 Figure 3.12: Gouy-Chapman for Meta l-Electrolyte Interaction........................................42 Figure 4.1: Capacitor Incorpor ating Kelvin Guard Ring.................................................44 Figure 4.2: 2-D Model of the Sensor Developed in FEMLAB.......................................46 Figure 4.3: 2-D Meshed Model of the Capacitor............................................................47 Figure 4.4: Potential and Electric Field Di stribution of the Sensor in Seawater.............47 Figure 4.5: Sensor with Guard Rings for Field Minimization.........................................48 Figure 4.6: Simulated Result fo r Capacitor with Guard Ring.........................................48 Figure 4.7: Twin-T Notch Filter Circuit..........................................................................49 Figure 4.8: Twin-T Filter Re sponse Simulated Using Pspice.........................................50 Figure 4.9: Schematic of Tw in-T Oscillator....................................................................51 Figure 4.10: Oscillator Response Simulated Using Pspice...............................................51 Figure 4.11: Set-Up for Twin-T Oscillator with Capacitive Sensor.................................52 Figure 4.12: Process Flow Steps for Fabrication of the Sensor.......................................53 Figure 4.13: Fabricated Sensors Prior to Alignment........................................................ 55 Figure 4.14: Top-View of the Completed Sensor.............................................................55 Figure 4.15: (a) Maxwell-Wagner Layered Capacitor Model, (b) Equivalent Circuit Model, (c) Reduced Mode l, (d) Equivalent Circuit for Reduced Model and e) Simplified Model.....................................................56 Figure 4.16: Block Diagram of the Circuit Employing Two F-V Converters and a Differential Amplifier Stage ...............................................................57

PAGE 11

viii Figure 4.17: Set-up for the Frequencyto-Voltage Converter Approach.........................58 Figure 4.18: Detailed Schematic of Sensor Detection Circuit with Calibration Adjustment................................................................................59 Figure 5.1: Plot of Complex Impedance for Different Salinity Con centrations (psu)....62 Figure 5.2: Plot of Capacitance Versus Frequency for Varying Salinities at 22 C........64 Figure 5.3: Plot of Capacitance Versus Salinity for 40 KHz at 22 C.............................65 Figure 5.4: Nyquist Plot for Varying Temperatures (Salinity = 34.821 psu)..................66 Figure 5.5: Plot of Capacitance Versus Frequency for Varying Temperatures at Constant Salinity...........................................................................................67 Figure 5.6: Twin-T Oscillator Fr equency Versus Salinity at 22 C................................68 Figure 5.7: Twin-T Oscillator Freque ncy Versus Temperature at 34.471 psu................68 Figure 5.8: (a) Plot of Complex Im pedance for Varying Salinities at 22 C, (b) Equivalent Circuit Model for Metal-Electrolyte System corresponding to the Plot in (a).....................................................................70 Figure 5.9: Plot of Capacitance Versus Frequency for Varying Salinities at 22 C........71 Figure 5.10: Plot of Capacitance Versus Salinity for a 2 KHz Signal at 22 C.................71 Figure 5.11: Nyquist Plot for Varying Temp eratures at Constant Salinity (34.856 psu)..72 Figure 5.12: Plot of Capacitance Versus Frequency for Varying Temperatures (34.856 psu)..................................................................................................73 Figure 5.13: Oscillator Fre quency Versus Salinity at 22 C.............................................73 Figure 5.14: Plot of Oscillator Frequency Versus Temperat ure at Constant Salinity of 32.891 psu.................................................................................................74 Figure 5.15: Output Voltage Versus Salinity at 22 C.......................................................75 Figure 5.16: Output Voltage Vers us Temperature at 34.65 psu........................................76

PAGE 12

ix SALINITY (CONDUCTIVITY) SENS OR BASED ON PARALLEL PLATE CAPACITORS Shreyas Bhat ABSTRACT This work is aimed at developing a hi gh sensitivity salinit y (conductivity) sensor for marine applications. The principle of sensing involves the use of parallel plate capacitors, which minimizes the proximity effects associated with inductive measurement techniques. The barrier prope rties of two different materials, AZ5214 and Honeywells ACCUFLO T3027, were investigated for use as the insulation layer for the sensor. Impedance analysis performed on the two co atings using Agilent s 4924A Precision Impedance Analyzer served to prove that ACCUFLO was a better di electric material for this application when compared to AZ5214. Two separate detection circuits have been proposed for the salinity sensor. In the Twin-T filter method, a variation in capacitance tends to shift the resonant frequency of a Twin-T oscillator, comprising the sensor. Simulations of the oscillator circuit were performed using Pspice. Experiments were performed on calibrated ocean water samples of 34.996 psu and a shift of 410 Hz/psu was obt ained. To avoid the problems associated with the frequency drif t in the oscillator, an alternate detection sc heme is proposed which employs frequency-to-voltage converters. The sensitivity of this detection scheme was observed to be 10 mV/psu.

PAGE 13

1 CHAPTER 1 INTRODUCTION Since archaic times, oceans have been used as a mode of transport and a source of food. However, it was not until late nineteenth century that oceans were perceived as information-rich reservoirs of scientific importance. The impetus for further oceanic studies began in the 1930s with the search fo r petroleum, continued with the emphasis for improved naval warfare and more recently e xpressed, to protect the ecosystem. Among the different facets of oceanography, physical oceanography relates to the study of physical properties and dynamics of the ocean, the primary interests being the oceanatmosphere interaction, the oceanic heat budge t and the coastal dynamics. No discussion on oceanography is complete without a men tion of parameters such as salinity, temperature, pressure and density. Extensiv e research has been done to understand the role of these parameters in regular oceanic pr ocesses, but there is still a lot of latent information that seems to remain elusive to the oceanographers worldwide. This work is primarily aimed at developing a salinity sensor for marine applications and hence, in the following sections, we shall focus on salin ity and its significance in oceanography.

PAGE 14

2 1.1 Definitions of Salinity Salinity (derived from the Latin equivalent salinus = salt) is commonly defined as the ratio between the weight of dissolved material in the sea water sample and the weight of the sample [1]. This ratio is generally expressed in parts per thousand. The dissolved material includes dissolved gases, because ev en gases dissolve in water. However, it excludes fine particles being he ld in suspension and other solids that are in contact with sea water. Salinity is also conveniently refe rred to as TDS or Total Dissolved Salts. It is nearly impossible to determine the to tal amount of dissolved salts in a sample by direct chemical analysis. In addition, it is impossible to obtain reproducible results by evaporating seawater to dryness and weighing the residue as some of the material present, mainly chloride, is lost during the final stages of drying (S verdrup, Johnson, and Fleming) [2]. Hence, it is logical to employ an alternate technique that yields reproducible results which albeit does not gi ve the exact amount of salts dissolved, yet represents a quantity of slig htly lower numerical value than TDS. The International Council for the Exploration of the Sea esta blished a technique in 1902, according to which salinity is defined as the total amount of solid material in grams contained in one kilogram of sea water when a ll of the carbonate has been co nverted to oxide, the bromine and iodine replaced by chlorine, and all orga nic matter completely oxidized. This method is accurate but difficult to use routinely [3]. William Dittmars analysis of 77 ocean water samples, collected during the HMS Challengers circumnavigation expedition around the globe, led to the establishment of The Principle of Constant Proportions This theory explains that the major dissolved constituents responsible for th e salinity of sea water occur nearly everywhere in the ocean

PAGE 15

in the exact same proportions, independent of its salinity. This constancy of composition was then exploited to deduce the salinity of sea water by measuring its single largest constituent, the chloride ion [1]. The weight of this ion in the water sample is termed as chlorinity. In 1966, a joint panel on Oceanographic tables and standards offered a more accurate definition of salinity based on its chlorinity, which is mathematically expressed by the relation, S = 1.80655 Cl (1.1) It is apparent that salinity is a ratio, and is hence, unitless. The Practical Salinity Scale (PSS) established in 1978, uses the term Practical Salinity Units (psu), as the standard unit for the salinity, and is essentially parts per thousand of sea water. The general representation for psu is %o. An advantage of expressing salinity in psu is that decimals are often avoided and values convert directly to grams of salt per kilogram of sea water. 1.2 Composition of Seawater Seawater is a solution of salts of nearly constant composition, dissolved in variable amounts of water. Potassium, 1.10%Calcium, 1.10%Sulfate, 7.70%Sodium, 30.60%Other, 0.80%Magnesium, 3.70%Chloride, 55% Figure 1.1: Major Constituents of Seawater 3

PAGE 16

4 The seawater is a mixture of more than 70 dissolved elements. However, only 6 of these constitute more than 99% of all the dissolved salts; all occurring as ions (electrically charged atoms or group of atoms) [1]. The major constituents of seawater are as shown in Figure 1.1. 1.3 Processes Influencing Ocean Salinity Since, the salinity of seawater is not the same throughout the ocean, it maybe of relevance to know what processe s actually influence salinity Apparently, salinity can be altered by varying either of the two compone nts the amount of dissolved salts or the amount of water. Nonetheless, the salts oc cur in almost the same proportions. An overview of the processes that tend to decrease/increase salinity is given in the following sections. 1.3.1 Processes Decreasing Salinity Precipitation, in the form of rain, snow, sl eet and hail, is one of the processes that contribute towards a decrease in seawater salinity. About three-quarters of all the precipitation worldwide falls di rectly into the ocean. Thus, it increases the amount of fresh water, decreasing salin ity. Runoffs from fresh water sources, like rivers and its tributaries also add fresh water. Icebergs are large chunks of glacial ice that have broken free from glaciers and flown into the ocean. Up on melting, they return fresh water to the oceans. Also, sea ice is formed at cold te mperatures in high latitude regions and is primarily composed of fresh water [1].

PAGE 17

5 1.3.2 Processes Increasing Salinity Formation of sea ice causes an increase in salinity. This is owing to the fact, that when sea water freezes into ice, only around 30% of the dissolved components are retained in the sea ice. This effectivel y causes fresh water removal from the ocean. Evaporation of water removes pure water fr om the oceans, thereby increasing salinity. Volcanic eruptions produce large amounts of gases, the major ones being sulfates and chlorides, that eventually reach the oceans. Submarine eruptions spew gases directly into the ocean, increasing its salt concentrati on. Weathering of rock s on the continents releases silica and elements like sodium, calcium, potassium and magnesium into the ocean increasing its salinity [1]. It is interesting to note how these processes affect the variation of salinity over depths. Since, most of these processes are surface phenomena, they lead to larger variations of salinity at the surface, and cause lit tle variation at greater depths. 1.4 Need for Salinity Sensing Salinity influences a lot of physical and biological processes in the ocean, both directly and indirectly. The following sectio ns will discuss in some detail, the vital information that can be inferred by determining seawater salinity. The key areas of salinity impact can be listed as: 1) Ocean Circulation. 2) Marine life. 3) Water Cycle.

PAGE 18

1.4.1 Ocean Circulation Before we proceed to appreciate ocean circulation and its impact on the global weather, it would be beneficial to comprehend the role of salinity in the determination of seawater density. Density of seawater is a function of its salinity, temperature and pressure and can be mathematically expressed as ),,(pST (1.1) Where, T = temperature, S = salinity and p = pressure [3]. An increase in salinity implies an addition of more dissolved material into the seawater, which translates to a corresponding increase in density. Similarly, an increase in pressure increases the density, because seawater is compressible to some extent. Conversely, temperature increase causes a decrease in the density, due to thermal expansion. Thermohaline graphs (plots of salinity and temperature) are used extensively to determine the density of seawater along the sea surface and at various depths. Density differences make water masses sink or float, and hence determine the vertical position and movement of water masses. The ocean like the earths interior is stratified (layered), with low density waters existing at the surface and high density waters below [1]. Evaporation and precipitation continually change the salinity of the surface waters and hence its density. Thus, there is constant circulation of water between the different layers in the sea. This circulation of water over the depth of the ocean plays a significant role in the circulation of dissolved oxygen and dissolved carbon between the various layers, which is vital for deep-sea aquatic life. Greater salinity, like lower temperature, causes higher density with a corresponding depression of the sea surface height. In warmer fresher waters, the density 6

PAGE 19

is lower resulting in an elevation of the sea surface. These height differences are related to the circulation of the ocean between the Equator and the Poles [4]. Figure 1.2: Energy Transfer through the Conveyor Belt [4] The changes in the density bring warm water from the Equator poleward to replace the dense sinking waters as illustrated in figure 1.2. Thus, the overall effects of surface current circulation and the deep-water thermohaline circulation lead to what is aptly called the conveyor-belt circulation model. This transport of heat, over the surface has a moderating effect on the climate of Europe. Current developments in mapping these parameters over broader areas, have indicated that changes in the thermohaline circulation can cause drastic climatic changes. As an instance, if surface waters stopped sinking, the surface water temperatures would be a lot higher, and consequently create hotter lands!! Salinity mapping can also give vital information about the buildup of greenhouse gases into the atmosphere, since these gases can alter the thermohaline circulations. 1.4.2 Marine Life The sensitivity response of marine animals to changes in the physical parameters of the ocean (like temperature, pressure, salinity, light transparency, buoyancy) varies 7

PAGE 20

8 between animals. Salinity in open ocean is le ss susceptible to large variations. Hence, evolution has led the deep-sea creatures to adap t to constant salinity and therefore, they can tolerate very small changes in the salinity levels. ( stenohaline animals). However, variation in salt concentrations in the estuarie s tends to be higher due to periodic rise and fall of tides. Hence, the aquatic diversity in the estuarine ecosystem has adapted itself to considerable variations in salinity levels. ( euryhaline animals) [1]. To better understand the role of salinity in marine animals, it might be beneficial to introduce the term osmosis, according to which, when two solutions of unequal concentrations are separated by a semi-permeab le membrane (animal skin, in this case), water molecules diffuse from the region of lo wer to the region of higher concentration. The pressure applied to the high concentra tion region to prevent water molecules from passing into it is called osmotic pressure. Marine organisms exchange nutrient molecules and waste molecules with their surrounding en vironment through osmosis. Unexpectedly large variations in salinity can lead to high osmotic imbalance, which may result in rupture of cells due to excessive pressure. Ther efore, it can be inferred that a variation in salinity brought about by natura l (precipitation etc.) or manmade factors (chemical spills etc.) can be consequential to the aquatic life. As an example, a study on lobsters performe d by Jury et. al. at the University of New Hampshire [5] reveals that drastic salinity variations can have fatal effects on their heart rates. Data like these aid in understand ing the vertical and hor izontal distribution of marine animals and their habitat preferences, and are widely used for purposes of fishing and conservation.

PAGE 21

9 1.4.3 Water Cycle As discussed in section 1.3, the salinity of seawater can be attributed to several physical, chemical and biological processe s. Sea Surface Salinity (SSS) is a key parameter in understanding how fresh water i nput and output affect ocean dynamics. It is possible to monitor variations in the water cycle due to processes like land run-off, sea ice melting and freezing and, evaporation and precipitation over oceans, by tracking the variations in SSS [6]. 1.5 Motivation Clearly, there is a need for salinity sensi ng. Salinity sensors are most often used in conjunction with pressure and temperature sensors for multi-parameter ocean monitoring. Integrated devices of this type are know n as CTD (conductivity, temperature and depth sensors) sensors. Most current CTD sensor s are sizeable and bulky devices, upto a meter in length [7]. In the oceanic environment, fo r reasons of corrosion protection, the smaller the surface areas the better. The sensor propos ed here is fabricated at the meso-scale (<2cm). Also, the motivation to minimize the proximity effects manifested in inductive sensors which are commonly used to measure salinity, led to the use of capacitive sensors for salinity sensing, more of which is expl ained in Chapter 2. The use of a Twin-T oscillator for the detection circuit stems from its ability to produce a low distortion output and relatively high sensitivity.

PAGE 22

10 1.6 Thesis Outline This thesis has been segregated into 6 chapters. The first chapter gave a description of the defi nitions of salinity, pa rameters influencing it, and the need for salinity sensing. Chapter 2 gives an overv iew of the various approaches used at measuring salinity (conductivity). Discussed in Chapter 3 are the theories and processes governing the operation of this sensor. Chapte r 4 talks about the design of the sensor, its optimization and construction details. The results are presented in Chapter 5. Chapter 6 is the conclusion to this work and also incl udes a brief discussion on the future work.

PAGE 23

CHAPTER 2 LITERATURE REVIEW The measurement of salinity has been done using several methods, both direct and indirect. The direct method for salinity measurement of a sample involves measuring its chlorinity, as discussed Section 1.1. The standard scheme adopted for estimating chlorinity is titration by the Mohrs method [8]. The principle of this method involves taking equal volumes of standard seawater and of the sample, and determining the volumes of given silver nitrate solution to precipitate completely the halogens. However, due to the level of difficulty associated with its experimentation and repeatability, oceanographers use salinitys dependence on conductivity, to their advantage [9]. An introduction to conductivity and the Practical Salinity Scale (PSS) precede the section on the current approaches at conductivity sensing. 2.1 Conductivity Conductivity is a unit measure of electrical conduction; a property by which matter conducts electricity. Electrical conductivity is the reciprocal of electrical resistivity, and is denoted by the symbol which is mathematically expressed by the equation ARL1 (2.1) 11

PAGE 24

Where, L = length of the material (in meters, m) R = resistance of the material to the flow of current. (in ohms, ) A = area of cross section of the material. (in meters, m) The unit for conductivity, is mhos / meter ( -1 m -1 ) or Siemens / meter (S.m -1 ). In an electrolyte, it is directly related to the total dissolved salts present in the solution. The more the dissolved salts, the higher is the conductivity of the electrolyte. Hence, conductivity is often measured to deduce the concentration of an analyte. The relation between conductivity and salinity for practical oceanographic purposes is given by the Practical Salinity Scale (PSS). 2.2 Practical Salinity Scale (PSS) of 1978 In 1978, UNESCO established the Practical Salinity Scale (PSS) which defines salinity in terms of a conductivity ratio, thereby, breaking its link with chlorinity. According to the PSS, salinity, SRRRRRSTTTT2/522/32/1157081.20261.70941.143851.251692.00080.0 )0,,(/)0,,(TKClCTSCRT )2.2(0144.0636.00375.00066.00056.0005.0))]15(0162.01(/)15[(2/522/32/1TTTTTRRRRRTTS For 422S Where C(S,T,0) is the conductivity of the seawater sample at temperature T and standard atmospheric pressure, and C(KCl,T,0) is the conductivity of the standard potassium chloride (KCl) solution at temperature T and standard atmospheric pressure [2]. 12

PAGE 25

2.3 Current Approaches at Conductivity Sensing Modern conductivity sensors are broadly classified under two categories: 1) Contact-type sensors 2) Non-contact type sensors 2.3.1 Contact Type Sensors As the name suggests, in these types of sensors there is a direct contact between the measurement probes and the surrounding media. There are 2 main types of contact-type conductivity sensors currently in use: 1) 2-probe conductivity sensor 2) 4-probe conductivity sensor 2.3.1.1 2-Probe Conductivity Sensors A schematic of such a sensor is as shown in figure 2.1. It consists of two conducting electrodes (metal or graphite in some cases), spaced a fixed distance d, apart from each other. 13 Fi gure 2 1 : 2Pro b e C o n d u ct i v i t y Sens or I x y d V

PAGE 26

14 The solution under test is allowed to flow between the electrodes. Upon application of an electric pot ential between the two electr odes, a current is generated between the electrodes and through the medium. The magnitude of this current is directly related to the electrical c onduction of the solution. Thus, the amount of current flowing through the circuit is a measure of the conducti vity of the solution. Transducers like these actually measure conductance of the solution. The probe/cell constant or K factor is what links the conductivity and conducta nce. K factor is determined by the geometry of the measuring cell and the distance between the el ectrodes. Mathematically, k is given by, K = (distance between the probes) / (c/s Area of the probe) For the sensor in figure 2.1, if x = y = d = 1c m, then the probe constant would be 1 / cm. The relationship between conductivity and conductance can be expressed as, Conductivity = Conductance K factor (2.3) K factor can vary from 0.01 to 100, but lies in the range 0.1 to 10 for most commercial probes [10]. The advantage of such systems is the si mplicity of design. However, they suffer from a few drawbacks, which can cause signif icant errors in measurement. Since, the probe is exposed to the solution, corrosion a nd fouling effects are an issue. Also, the problem of space-charge polarization is prono unced in such systems since, the same probes are used for applying the input voltage and measuring the resultant current. The problem of polarization has been minimized, if not eliminated, using a better design employing 4 probes.

PAGE 27

2.3.1.2 4-Probe Conductivity Sensor This design makes use of 4 conducting probes. The alternating voltage is applied across two probes while the current is measured across the other two. The 4-electrode cell uses a reference voltage to compensate for any polarization or fouling occurring at the voltage probes [11]. Figure 2.2: 4-Probe Conductivity Sensor V A Figure 2.2 illustrates the 4-pr obe conductivity cell. An altern ating voltage is applied to the two circular electro d es in the m i ddle. The current is m easured across the o u ter 2 electrodes. S e nsors of this kind can be used for measuri n g solut i ons with a wider range of conductiv ities than the two probe m e thod [11]. However, fouling is s t ill an is sue with such devices. 2.3.2 Non-Contact Type Conductivity Sensors This classification of conductiv ity sensors arises from the fact that there is no contact between the analyte and the sensing d e vice. Th is gives them a n apparent advantage over the contact type sensors, whose long term st ability is lim ited by polar ization and fouling. The m easuring principle for m o st non-contact conductivity sensors is inductive. The two commonly used types of inducti ve conductivity sensors are: 15

PAGE 28

1) Single transformer 2) Double transformer 2.3.2.1 Single Transformer This is the simplest form of inductive conductivity sensor. Secondar y loop formed b y Source Case Ring shaped core of the transforme r Primar y coil of the Figure 2.3: Single Transformer Conductivity Sensor Shown with Seawater as Surrounding Media [12] It consists of a magnetic induction device, a single coil which acts as the primary coil of the transformer as shown in figure 2.3. The secondary coil of the transformer is formed by the surrounding liquid whose conductivity is to be determined. An oscillating current of known frequency and amplitude is used to drive the primary coil. The current circulating in the coil produces a time-varying magnetic field, called the primary field, which induces an electric field, as defined by Faradays law of magnetism. When the coil is brought close to a conducting body, seawater in this case, the electric field generates eddy currents in that body. The density of these eddy currents is directly proportional to the conducting ability of the medium (given by Ohms law). These eddy currents generate a secondary magnetic field, pointing in a direction opposite to the primary field, according to Lenzs law. Again, by Faradays law, we have that a variation in the magnetic flux over time gives rise to an Electromotive force (emf) in the 16

PAGE 29

coil, whose phase is 180 relative to the driving current. The expression governing this relation is given by tEB 2.4 The conducting media is thus seen by the oscillator as a resistive load. Figure 2.4 shows the circuit representation of the single transformer. Knowing that the internal impedance of the oscillator is finite, a change in the load resistance brought about by a change in conductivity of the medium, causes the amplitude of oscillations to vary. Hence, a change in the oscillator amplitude is a measure of the change in conductivity of the analyte. Thus, the same coil acts as both the electromagnetic source and the receiver. K K 1 2 17 1 n 1 2 R l oad Fi gure 2 4 : E q ui val e nt C i rc ui t f o r Si n g l e Tr a n sf or mer Desi gn The equivalent circuit for the single tr ansform e r is shown in fig 2.4 where R eq is the equivalent resistance as seen by the source. The single transform e r conductiv ity sensor has n o t been applied to ocean ographic applications [12]. Howe ver, Lynn W. Hart et al. (1988) have described the use of such sensors for non-invasive electrom agnetic cond uctivity sensing in biom edical areas of research [13]. Their discussion involves the pos sible use of such sens ors to detect brain edem a induced by traum a to the skull. Brain edem a is known to propagate from the point(s) of traum a and hence de te ction is of prim e im portance in the tre a tm ent of injuries resulting f r o m traum a to the sku ll. A genera l ru le of thum b for such type of sensors is that R l oad L 11 n 1 n 2 = 1 n 1

PAGE 30

18 the depth of conductivity sensing is related to the radius of the sensing coil. The possible use of coils 2-5 cms in radius has been disc ussed. Initial experiments were carried out using HP 4275A LCR meter. The readings we re taken at oscillator frequency, f = 10 MHz for a coil radius, R = 10.27 cm, and number of turns, N = 2. The plot of resistance change to the conductivity variations in the sub-skull matter yielded a sensitivity of 76.4 m / (S/m). The Colpitts osci llator was chosen as an alternative to this expensive impedance meter, which gave a measurable dc output level change in response the change in conductivity of the analyte. A plot of the voltage change for variations in conductivity yielded a sensitivity of 77.2 mV / (S/m). The upper and lower bounds for conductivities for such detection lay in the range 0.10 S/m 0.5 S/m. However, experiments have not been performed with this set up for analytes with higher conductivities comparable to those of seawater. Robert Guardo et. al (1997) demonstrated the use of such contactless sensors in measuring thoracic conductivity [14]. This wa s an improved version of the contactless sensor developed earlier by Guar do et. al.(1995), with better li nearity achieved due to its ability to keep the voltage amplitude consta nt using a feedback loop. The change in conductance recorded for 0.05 S/m variation was 0.8 % of the measured output signal. 2.3.2.2 Double Transformer The double-transformer type of conductivity sensor wa s first introduced into physical oceanography by Relis (1951) and la ter on applied for Williams (1960) and Fatt (1962) [12].

PAGE 31

The transformer consists of two transformers, namely the driver transformer and the pick-up transformer. In this system, the surrounding media (sea water in this case) forms the secondary coil of the driver transformer and the primary coil of the pick-up transformer. Secondary loop formed by seawater 19 Case Source S e con d a r y co il of t he seco nd trans former Primary coil of the firs t trans former Ring sha ped core of the trans former Fi gure : 2. 5: Sc hem a t i c f o r a Do u b l e Tr a n sf orme r C o n d u c t i v i t y Sens or [ 1 2 ] A schem a tic of the design is as shown in the figure 2.5. The arrangem e nt of the two coils is im portant in that the o n ly couplin g between them is by the com m o n water loop. The voltage or current in the secondary coil of the pick-up transform e r is the sensor output signal. Here again, a high frequency curr ent excites the first coil that generates a current in the second coil. The amount of current generated depends on the am ount of m a gnetic flux, which is coupled through the so lution. Thus, the current in the second coil is proportional to the conduc tivity of the solution. Another contactless toroidal sensor em ploying 2 coils was i m plem ented by Natarajan et. al.[15]. The novelty w ith this approach however lay in the RF detection em ployed. Most conventional double transf orm e r sensors calibrate the m e dium s conductiv ity to the chan ge in am plitude of the current in the sensing coil. The detection

PAGE 32

20 strategy in this approach, invol ves measuring a phase shift that occurs in an alternating signal that couples between two closely held coils (drive and sense) placed in the conducting medium. The phase shift is dete cted using a phase detector circuit. Measurements taken at constant salinity of 36.15 ppt over varying temperatures show the device sensitivity to be 5mV for 1 S/m change in conductivity of the analyte. However, the device sensitivity for change in conduc tivity was recorded only for two different concentrations of the analyte. Conductivity sensors based on double transfor mers are the most widely used in oceanography. Contactless conductivity sensing technique serves to curb the limitations faced by the contact type electrodes, in that fouling and polarizat ion of the measuring electrodes is not of much concern. The co nstruction is robust an d preservation of geometry factor is easily achieved by cleaning with soaps, solvents and a brush. The large inherent hole in inductive sensor s permits free flushing of solution. However, inductive sensors suffer from a few drawbacks. The primary drawback of the inductive sensing approach is the pr oblem of external fields. Apparently, the measuring field in this type of sensing is exposed to the surrounding media. However, this entire field is not coupled to the s econdary coil of the transformer. Sea-bird Electronics Inc., in their comparative study on conductivity cells estimate that most of the inductive cells currently in use, have 11 to 20% of their field external [17]. The external fields lead to errors commonly known as proximity errors which can have major consequences on the data. Any material that has a conductivity value other than that of sea water can influence the cal ibration of the system. Cables sensor housings and marine growth close to the conductivity cell can contri bute to a shift in the systems calibration.

PAGE 33

21 It may be interesting to note that the antifouling coating on the cell is equally likely to cause errors. The simple reason for this is that any coating layer will alter the cell geometry. In the same study by Seabird, it is claimed that even a th in layer of coating (about 100 microns) will change the indicated sa linity by approximately 0.7 psu, which is significant in systems aiming at achieving high-a ccuracy standards. Even if the system is calibrated initially considering the thickness of the antifouling coating, subsequent leaching of the paint will lead to errors in calibration. The dependence of the sensor output on the coil inductance poses another problem in inductive sensors. In a study by Hinkelmann, it was shown that the permeability of the transformer core was susceptible to variability, caused by its temperature and pressure dependence [12]. However, precise electronic design employed in current inductive sensors can eliminate the undesired effects due to shift in the coils permeability, by calibrating the system using simplified assumptions. An alternate approach at minimizing the effect of permeab ility drift due to pressure variations has been the use of pressure-balance ports. Howeve r, they lead to significant escalation in costs. The problem of external fields associat ed with inductive sensing can be overcome by employing capacitive principles for conduction sensing. Parallel plate capacitors rely on trapping the electric field in the region between the plates. This provides an inherent confinement of the measuring field between the two plates, with little interference due to external fields/objects. Only a small portion of the field is exposed to the external media on account of fringing effects at the corners and edges of the plates. More discussion on the minimization of fringe fields in parallel plate capacitor follows in Chapter-4. Parallel

PAGE 34

plate arrangement is the simplest form of capacitor configuration. However, conductivity sensing using capacitors is not just limited to the parallel plate configuration. Various possible configurations have been discussed in [17] as shown in the figure 2.6. A B A C a) b) D B Fi gure 2 6 : C a paci t o r C o nf i g urat i ons ( a) C y l i ndri c al Set u p ( b) IDTs From a fabrication eng i neers pers pectiv e, the cylindrical configuration is not much preferred due to the le ve l of dif f i culty in volved in depositing a uniform layer of m e tal for the electrod e. The IDT structure is p r eferred due to its eas e o f fabrication as both the electrodes are co planar. 2.4 IDT Configuration in Measurem e n t of Conductiv ity A schem a tic and equivalent circuit of the m easuring cell is as shown in figure 2.7. Fi gure 2 7 : ID T C o nd uct i v i t y C e l l an d E q ui val e nt C i rc ui t [ 1 8 ] The stru ctu r e consists o f two electrode s, the impedance across which gives a m easure of t h e conductivity of the liquid between the com b s. The equivalent circuit for 22

PAGE 35

such a structure has been modeled as shown above. R sol is the solution resistance and is given by the equation, R sol = solcellK 2.5 Where, K cell is the cell constant of the conductivity cell and is dependant on its geometry. sol is the conductivity of the electrolyte. The direct capacitive coupling between the electrodes is represented by the cell capacitance C cell given by the relation, C cell = cellsolK 0 2.6 Where, sol is the relative permittivity of the electrolyte. C dl represents the double layer capacitance and is attributed to the formation of a layer of charge at the electrode, due to the electrode-electrolyte interaction. A plot of the complex impedance can be used to compute each of these parameters, more of which has been discussed in Chapter 3. However, the measurement of electrolyte of high concentrations has not been discussed in this study on IDTs. Fritz Wolter and Fritz Thom have demonstrated the use of a parallel plate capacitor to determine the complex permittivity of supercooled aqueous solutions in the 1 MHz range [19]. The electrode material used was copper, with a coating of platinum to avoid polarization effects. Studies were performed on absolute ethanol and aqua bidest for low conductivity ranges (< 0.02 S/m). The permittivities were obtained using the Schlumberger 120 impedance analyzer. The conductivity of the solutions was varied by varying its temperature. In the temperature range 10 C19 C, the change in capacitance observed for the sensor was 0.033 pF for every 1 C change in temperature (=0.85% 23

PAGE 36

24 change in capacitance for every 1 C). The response of the sensor was measured for different liquids. However, no data has been presented for change in conductivity for a particular solution of varying concentr ations, at a constant temperature. Hilland [21] presented a simple sensor sy stem to measure the dielectric properties of saline solutions. A co-axial probe was used to measure the dielectric properties of saline. The use of Sucoplate alloy has been de monstrated to avoid corrosion of the probes. Measurements were performed in the frequency range 500 MHz 40 GHz, at 20 C using the Network Analyzer HP8722C. The conductivity of the analyte was obtained as 0.126 S/m for every psu change in salinity.

PAGE 37

CHAPTER 3 THEORY AND UNDERSTANDING 3.1 Fundamentals of a Capacitor The most common form of capacitors is the parallel plate capacitors. It consists of two conducting plates, across which a potential difference is applied as shown in figure 3.1. The gap between the plates is filled with a material known as the dielectric. V DielectricConducting plates Figure 3.1: Parallel Plate Capacitor Configuration Capacitance for a parallel-plate capacitor is given by the expression, dAC (3.1) Where, A = overlapping area between the two conducting plates, d = distance between the plates. = dielectric constant (permittivity) of the material Permittivity is given by the relation, = 0 r (3.2) 25

PAGE 38

26 where, 0 = Permittivity of free space (vacuum). (= 8.854 F / m 2 ) r = Relative permittivity of material. The term dielectric is usually used to signify materials with low conductivity and high insulation characteristics. The electric fiel d distribution within the gap of a parallel plate capacitor depends on the type of dielectri c and its resistivity. Application of voltage across the plates leads to distribution of equal and opposite charges on the surface of the plates (from Gauss Law, it is known that charges cannot reside within the bulk of a conductor). 3.2 Complex Permittivity and Dielectric Loss It is desirable to have dielectric materials with zero loss, so that they confine more charge. However, in practical cases, no mate rial (except vacuum) behaves as a perfect dielectric and is associated w ith a finite loss. One of thes e losses is the conduction loss, representing the flow of actua l charge through the dielectric. The other is a dielectric loss and arises due to movement or rotation of mole cules in a time-varying electric field, more of which will be evident following a discussion on polarization in the next section. One way of describing dielectric losses is to c onsider the permittivity as a complex number, defined as = j = | | e -j (3.3) where, = ac capacitivity, = dielectric loss factor This treatment would necessitate the introdu ction of a complex capacitance, defined as C = C jC (3.4)

PAGE 39

Where, C and C are the real and imaginary parts of the complex capacitance respectively. This definition of complex capacitance permits us to use this complex value in any equation derived for real capacitance in a sinusoidal application, and get the correct phase shifts and power losses by applying the usual rules of circuit theory. Fig 3.2: Loss Tangent Vector Diagram When specifying lossy capacitors, the complex permittivity is generally expressed by and tan where tan = (3.5) Where, tan is called the loss tangent or simply, dissipation factor (DF), as shown in fig 3.2. A practical capacitor is more than pure capacitance. It is associated with a series and parallel resistance. Dissipation factor can also be expressed as, DF = XRcapacitorthebystoredEnergycapacitorthebylostEnergy (3.6) Where, R = resistive component of the capacitor X = capacitive reactance of the capacitor 3.3 Polarization of Materials The fundamental difference between a conductor and a dielectric is that a conductor has free electrons (loosely bound to the atom), whereas in a dielectric the electrons in the outermost shells are tightly bound to the nucleus. For any material in the 27

PAGE 40

absence of an electric field, the electrons form a symmetrical cloud around the nucleus, with the center of this cloud being at the center of the nucleus, as shown in figure 3.3(a). The electric field generated by the positively charged nucleus attracts the electron cloud around it, and the mutual repulsion of the electron clouds of adjacent atoms provides matter its form. Atom E ex ter n al E ex ter nal +q d -q el ectr on Posi ti v e l y char g ed nucl eus C enter o f el ec tr on cl ou d b) E ex ter nal 0 a) E ex ter nal = 0 c) Electric dipole Fig u r e 3.3: Pola riza tion Ph en om en on i n Dielectric Molecules [ 2 2 ] The electrons in a conductor behave differe ntly upon the applicat ion of an electric field. W h en subjected to external electric fields (dc), the loosely bound electrons in the conductor can jum p from one atom to the ot her, causing an electric current. However, this is not p o ssible in a diele c tr ic. A n extern al f i eld E external cannot cause m a ss m i gration of electrons in a d i electri c m e dium However, it can polarize the atom s of the dielectric. This process involves distorti ng the center of the electron cloud and the nucleus, so that the atom is no longer symmetric as evident fr om figure 3.3(b). This distortion causes the negativ ely charged portion of the atom to be displaced from the positive part, creating a dipole where the nucleus m a ybe represented as +Q and the center of m a ss of the electrons m a ybe represented by a charge Q. The for m ati on of dipole is illustrated in Figure 3.3(c). This kind of polarization is term ed as elec tronic p o lariza tion 28

PAGE 41

29 Thus, an externally applied field, caus es an induced electric field (due to the formation of dipoles), calle d the polarization field, E pol E pol is lesser in magnitude and opposite in direction to the E external Apparently, the induced field tends to reduce the magnitude of the total electric field on the mate rial. This reduction factor in the electric field is referred to as dielectric constant of the material. 3.4 Types of Polarization Mechanisms The interesting aspect of polarization in dielectrics is that th e contribution of the different polarization mechanisms to the overall permittivity is frequency dependant. Therere four basic types of polarization, of which, however, not all are relevant in this study of dielectric behavior in electrolytic samples. 3.4.1 Interface Polarization Surfaces, grain boundaries and interphase boundaries (which include the surface of precipitates) may be charged. In other wo rds, they contain dipoles which may become oriented to some degree in an external field and thus contribute to the polarization of the material. However, there is no simple method to determine the charge at the interfaces or their contribution to the total polarization of the material [25]. 3.4.2 Electronic and Atomic Polarization The polarization behavior referred to in figure 3.3 is also known as electronic polarization because it results from the shift of el ectronic charges within the atom, under

PAGE 42

the influence of an electric field. In atomic polarization, adjacent positive and negative ions stretch under the influence of an applied field. 3.4.3 Ionic Polarization This is generally seen in solids with some degree of ionic character. They have internal dipoles, which cancel each other out and are pretty much unable to move under the action of an electric field. The external field then induces net dipoles by slightly displacing the ions. 3.4.4 Orientation (Dipolar) Polarization This kind of polarization is generally associated with liquids and gases. A molecule is formed when one or more atoms combine to share their electrons, forming a single entity. A non-polar molecule is one in which the centers of positive and negative charges coincide. However, polar molecules have their positive and negative centers of charges displaced from each other, creating a permanent dipole moment. In the absence of an electric field, these dipoles are oriented in a random fashion so that no polarization exists. However, under the action of an electric field, the dipole will tend to rotate and align itself with the electric field, causing orientation polarization to occur as shown in figure 3.4a. 30 Torque E-field F o r ce Fo r c e Fi gure 3 4 : ( a ) Di p o l e R o t a t i on i n a n A p pl i e d Fi el d [ 2 4 ] ( b ) Macrosc o pi c Repre s ent a t i o n of t h e Orien t a tio n Pola riza tion Under th e In flu en ce o f a n Ap p lied Electric Field [2 2 ] a)

PAGE 43

31 Figure3.4 continued The friction accompanying the rotation of the dipole results in dielectric losses. Water, being a polar molecule, exhibits strong orientation polarization. 3.5 Contribution of Polarization to Permittivity Having discussed the various polarization mechanisms, we shall now try to explain their response to different frequencies of the applied field. Fig 3.5: Frequency Response of Various Polarization Mechanisms [24] Applied Electric Field Induced Electric Field dipole Dielectric Charged Conducting plates b)

PAGE 44

32 Electrons have much smalle r masses than ions. This allows them to respond more rapidly to a time-variant electric field. For el ectric fields of high fr equencies (like light), only electronic polariza tion can occur. At smaller fre quencies, relativ e displacement between positive and negative ions can occur. Orientation of permanent dipoles requires the rotation of bulky molecules. This can oc cur only if frequency of applied field is relatively low (i.e. in the MegaHertz range or lower). 3.6 Interfacial or Space Charge Polarization In addition to these polarization mechanisms, the existence of interfacial effects such as macroscopic discontinuities in the ma terial result in another form of polarization called the space-charge polarization [24]. Charge carriers also ex ist that can migrate over a distance through the material, upon the application of low frequency fields. Spacecharge polarization occurs when the motion of these migratin g charges is impeded. This can occur due to trapping of charges at the interfaces of materials, or when charges cannot be freely discharged or replaced at the electrodes. The field distortion consequential of the accumulation of these charges increases the overall capacitance of the material which appears as an increase in The behavior of this polarization to the frequency of the applied field is intuitional in that, at low frequencies the charges have sufficient time to accumulate at the borders of the conducting regions resulting in an increase in Conversely, at higher frequencie s the charges do not have time to

PAGE 45

accumulate and hence polarization effects are minimal. These low frequency effects are often referred to as the Maxwell-Wagner Effects. 3.7 Relaxation Time It maybe expected that as the applied electric field, E, is brought to zero, the polarization field, P, will follow instantaneously. However, it takes a certain finite amount of time for the dipoles to revert to their original state of randomness (such that P = 0). This time is known as the relaxation time (). EP tt Figure 3.6: Transition in Polarization with Sudden Field Drop [25] Figure 3.6 illustrates the relaxation time observed in dielectrics due to a drop in the applied field. All of the polarization mechanisms discussed so far can only operate up to a limiting frequency, after which a further frequency increase will result in their disappearance. Because of the spring-like nature of the forces involved, electronic and atomic polarization are accompanied by resonance type absorption. However, for orientation polarization, the disappearance accompanied by a broader peak in the loss factor, is more gradual, because the mechanism involved is of the relaxation type, and may involve a broad distribution of relaxation times. Indeed, the decline in may be so 33

PAGE 46

slow that may appear almost constant, and be correspondingly small, over a wide frequency range. 3.8 Debye Theory of Dielectric Behavior Figure 3.7 shows the dielectric behavior of water obtained for an input signal swept over a range of frequencies. Figure 3.7: Debye Relaxation of Water at 30C [24] The curves shown above follow the Debye theory of complex permittivity given by the following equations, 2)(1 s 3.7 2)(1)( s 3.8 Where, = 2f (f being the frequency of sweep) And, = 2f 0 (f 0 being the relaxation frequency). s = (at = 0) and = (at = ). 34

PAGE 47

From the Debye relation, the dependence of permittivity on the frequency of the applied field is evident. As frequency increases, decreases and this drop in is necessarily associated with a peak in Except for exceedingly high applied fields, is independent of the magnitude of the applied electric field for all dielectric materials used in practice, except ferroelectrics [24]. 3.9 Cole-Cole Diagram for Complex Permittivity Although, the Debye equations for permittivity are hailed as one of the most significant advancements in the realm of dielectric behavior, the Cole-Cole plots act as a better approximation in most cases. It is a plot of (on the X-axis) versus (on the Y-axis). The equation governing the Cole-Cole plot is given by, ])(1[)(10 js 3.9 Where, = j And, (/2) = angle between the real axis and the line drawn to the center of the of the circle from the high-frequency intercept. Figure 3.8: Cole-Cole Representation of the Debye Model of Water at 30C [24] 35

PAGE 48

36 The above plot is for a dielectric with single relaxation freque ncy and the maximum value for is as shown in the plot. 3.10 Dependence of Permittivity on the Concentration of Analyte Minute concentrations of el ectrolyte usually impart si gnificant condu ctivity to a liquid medium. Besides conductivity, addition of electrolytes can influence the dielectric behavior of the medium in two ways. Firstly, its ions may associate, generating an ionpair or similar solute species of appreciable dipole moment. Such species will then be expected to make their own contribution to the dielectric polarizat ion and dispersion. Secondly, the ions or their aggregates in the medium can, by the virtue of their strong localized electric fields, influence the solven ts molecular interactions. A corollary to the second aspect could be the explanation that solvent molecules maybe firmly bonded to the ions, and so imparted the new character of molecules of solvation. It is the second aspect that we are concerned with, during th is discussion on the effect of electrolyte concentration on the permittivity of the solu tion [26]. The electrostatic field in the immediate neighborhood of the sodium or chloride ion is such that th e interaction energy between the water-molecule dipole greatly exceeds that of th e typical hydrogen bond between the solvent species. This means that an appreciable number of water molecules will be frozen around each of the solute ion, and this produces a molecular pattern in thee liquid. This process is called hydration. Figure 3.9 shows the hydration process in a solvent upon introduction of salt ions. The hydrated salt ion is a bulkier enti ty than a nonhydrated ion. The hydrate sh eath around the dipoles preven ts them from orienting

PAGE 49

themselves with the changing electric field. This is the principal cause for the reduction in permittivity with addition of salt. Figure 3.9: Hydration Process in a Solute-Solvent Mixture [27] 3.11 Temperature Dependence of Permittivity The temperature dependence of polar materials is different than that of non-polar materials. While in non-polar materials, a change in temperature is associated with a fall in permittivity, due to the change in density of the material, the temperature dependence in polar materials is attributed to orientation polarization [26]. The dipoles in a polar solvent, are arranged randomly (such that P = 0) and are constantly moving, due to thermal motion. It may then be expected that as the temperature increases, the thermal energy of the individual molecule increases. Hence, increasing the thermal motion 37

PAGE 50

38 reduces the alignment of the permanent dipoles by the electric field, thereby reducing the orientation polarization. This causes a drop in permittivity of the sample. As may be apparent from the constructi on of the sensor, to be discussed in Chapter-4, there exists a thin layer of dielectric separating the sensor plates from the conducting solution. Having learnt the e ffects of frequency, temperature and concentration on the dielectric pr operties of materials, it is of relevance to understand the yardsticks in assessing the stability and barrier properties of the dielectric under consideration. 3.12 Assessment of Barrier Properties of Polymeric Coating on the Electrodes Polymers are high molecular weight mo lecules composed of repeating units (monomers), linked together to form long ch ains. Polymeric layers, coated on electrodes to prevent metal corrosion are thought to act as perfect barrie rs separating the metal from the surrounding media. However, the reliabilit y of these materials is limited largely by the fact that over time, diffusion of ions through the coating can establish contact between the media and the electr ode surface. This can have unde sired effects, particularly, in capacitive sensing applications where the ab sorption of water by the dielectric film can effect an alteration in its di electric properties. Absorption of water generally tends to increase the dielectric constant of the polym er. This can lead to erroneous results in calibration of the device. Another adverse effect of this diffu sion process is the deformation in the geometry of the film [2 8]. Water molecules fill the voids that are formed between polymeric chains and conse quently induce relaxati on or simply, swelling of the polymer. This effectively changes th e thickness of the diel ectric film. Hence, it

PAGE 51

39 becomes imperative to study the effects of the surrounding media on these barrier layers before making a final selection on the type of coating. One of the most trusted techniques for barrier layer analysis is impedance spectroscopy. 3.12.1 Impedance Spectroscopy Film breakdown results from ion transport through it and the re sulting failure is best understood in terms of the oxidation and reduction reactions at the metal-fluid interface. It is logical make an assessment of film properties usi ng electrochemical test methods. Although, ion diffusion and mass ab sorption (slow processes) have been identified as the causes for sw elling degradation, another cons ideration often neglected, is the surface morphology of the coating. Pore s in the polymer during its application process, can act as direct channels between the fluid and metal and can cause localized corrosion, ultimately leading to film failure (fast process) [29]. Of the several methods for th e assessment of barrier properties viz. film resistance measurement, film capacitance measurement, polarization resistance measurement, the electrochemical impedance spectroscopy, or AC impedance spectroscopy, is the most significant. It is clear that several processes take place at the metalsolution interface (like electrochemical reactions, diffusion, adsorption etc.). DC measurements average all these processes, from the slowest to the fastest. However, in AC measurements only those processes are taken, which have sufficient time to occur between the alternating fields of the applied signal. Hence, each of the pro cesses can be traced by sweeping an input signal over a range of frequencies.

PAGE 52

Electrochemical impedance is determined by applying a sinusoidal excitation of low amplitudes. This is done to keep the cells response pseudo-linear, where the resultant measured current signal follows the input signal in frequency but is shifted in phase relative to it [29]. Mathematically expressed, if the input voltage signal is represented by, E(t) = E 0 Cos (t) 3.10 Where, = 2f, (f being the frequency of the sinusoid), Then, the response current signal, can be expressed as I = I 0 Cos (t ) 3.11 Where, is the phase difference between the input and output signals. The impedance, Z, of the system can then be deduced from relation, Z = )()()()()()(000 tCostCosZtCosItCosEtItE 3.12 The impedance can be expressed in terms of real and imaginary terms as, Z = Z + jZ 3.13 Or in polar form as, Z = | Z | 3.14 A graphical method of representing complex impedance data is the Nyquist plot as shown in figure 3.10. A Nyquist plot is obtained by plotting Z on the X axis and Z on the Y axis. Such a plot is representative of the elements forming the equivalent circuit model of the electrode-medium interface. The length of the vector gives the magnitude of the impedance. 40

PAGE 53

41 Fi gure 3 1 0: Ty pi cal N y qui st Pl ot of t h e C o mpl ex I m pe d a n ce The graph represented above is typically the representation for a coated electrode where the coating prop erties have started to degrade due to absorption of water. The equivalent circuit m odel for su ch a represen tatio n is called the Randels cell and is shown in figure 3.11 [30]. Fig u r e 3.11: Ra nd els Cell Circu it Mod e l For a Randels c e ll eq uivalen t cir c uit, the individual circuit elem ents can be obtained from the Nyquist plot. The im pedance of su ch a circuit can be expressed using the following equation Re Z = 2 2 2 1 dl t t S C R R R = 0 |Z| R s R s + R t Re Z -Im Z

PAGE 54

Im Z = 22221tdltdlRCRC 3.15 The right hand corner of the semi-circle in the Nyquist plot for the Randel cell is indicative of low frequency, while the left indicates the response at high frequencies. The point of intersection of the semi-circle with the X-axis on the right corner yields the component R sol + R t while the point of intersection to the left yields the solution resistance R sol It may be observed at lower frequencies, the cumulative resistance is given by R sol + R t This indicates that at lower frequencies the polarization resistance is significant. C dl represents the double charge layer. More on this layer is explained in the next section. 3.12.2 Double Charge Layer The formation of this layer arises due to the interaction between the electrode and the electrolyte. The radius of the hydrated ions prevents direct contact between the ions in the electrolyte and the metal electrode and the diffusion of ions in the solution causes a non-linear charge distribution. The double charge layer can be best represented using the Gouy-Chapman model as shown in figure 3.12. Figure 3.12: Gouy-Chapman Model for Metal-Electrolyte Interaction [18] 42

PAGE 55

The following equation is used to compute the Randel cell parameters. dltCR21max 3.16 Where max is the frequency for which the peak of the curve occurs, R t is the charge transfer resistance or polarization resistance, C dl is the double layer capacitance. Double layer capacitance can be formulated as shown below: diffSterndlCCC111 3.17 Where, C Stern is the Stern capacitance, assumed to be a constant for most systems and is independent of the solution concentration, while C diff is the diffusion layer capacitance, which is given by the equation, TKNCezCAwaterdiff....10.20223 3.18 where, z is the valence of the ion, C is the ion concentration (mol/L), T is the absolute temperature, N A is the Avagadro number, 0 is the absolute dielectric constant of water, K is the Boltzmanns constant and T is the absolute temperature. It may be observed from the above equation that the diffusion capacitance increases as the ion concentration increases. 43

PAGE 56

CHAPTER 4 DESIGN, OPTIMIZATION AND CONSTRUCTION In this chapter, the design of the sensor, its optimization and fabrication have been described. Also, discussed are the detection strategies for the capacitive sensor. 4.1 Minimization of Fringe Field Effects Using Guard Rings As already discussed in Chapter 2, the only fields external to a parallel plate capacitor are the fringe fields caused due to electric field bending at the corners and edges. The capacitance of a parallel plate capacitor is given by equation 3.1. However, in practice, the measured capacitance of a capacitor is generally higher than the calculated capacitance. This increase is attributed to the fringing fields at the plate edges. It is difficult to calculate the fringe field analytically and hence, it is in our best interest to minimize the fringe field using a suitable strategy. Kelvin guard electrodes (ring) are usually employed to minimize these effects. It is sufficient to use this ring around just one of the electrodes. A typical arrangement for Kelvin guard ring is as shown. 44 V To p el e c tr o d e Figure 4.1: Capacitor Inc o rporating Kelvin Guard Ring Guard Rin g Bott om el ectro d e

PAGE 57

45 The use of Kelvin guard ring around the electrode serves two purposes: 1) It serves as a shield agains t external interference and, 2) It shifts the electric field out of the area of interest. A general rule of thumb, as explained by Heerens [31], is th at if the distance between the electrodes, is comparable to the smallest dimension of the electrode, then the fringe fields are significant. In our case, the distance between the electrodes is ~600 microns, and the smallest dimension is 1 cm which should make it less prone to fringe effects. However, a good design should incorpor ate strategies to eliminate the fringe fields. It should be noted that for excellent results in minimizing fringe effects, the following must be taken care of: 1) It must be ensured that the electrode and the surrounding guard ring should be totally in-plane with each other. Since, in this case, the electrode and guard ring are built on the same substrate, they can be assumed to be coplanar. 2) The distance between the guard ring an d the electrode should be minimal for maximum minimization. For the salinity sens or, the gap was chosen to be 400 microns, due to limitations of the PCB fabrication set-up in the laboratory. 3) Ideally, the larger the width of the guard ring, the better the shie lding. However, 5 mm was chosen as a convenient width for the gua rd ring for the capacitive salinity sensor. 4) The guard ring should be held at a potenti al equal to that of the opposite plate. An attempt was made to quantify the reduction in fringe field due to the incorporation of the Kelvin guard ring. This was done using the Finite Element Method tool FEMLAB. The sub-module Quasi-static of the Electromagnetic section was used for the simulation. The reason for having chosen the quasi-static mode is that the sensor

PAGE 58

is intended to be used at low frequencies. This follows from the discussion on polarization in section 3.5. The frequency of the signal was specified as 1 KHz. The first step in the simulation was developing the 2-D model which is as shown in figure 4.2. 46 Figure 4.2: 2-D Model of t h e Sens or Developed in FEMLAB. The Bou ndary Conditions f o r t h e Material s Used are as Indicated FR4 substrate; r = 4.5 Photoresist AZ5214; r = 6 Copper; (co nductivit y) = 5.7E7 Sea wat e r; = 5 S/m r = 80 Each shape in the g e ometry is as sig n ed a sub-d o m a in nam e This facilitates th e specification of m a terials used in th e cons truction of the sensor. FEML AB perm its the use of user-defined m a terials, allowing the user to specify the m a terial properties as indicated in figure 4.2. T o sim u late the s eawater m e dium around th e sensor, the capacitor was enclosed in a box with dielectric prope rties m a tching those of seawater. Next, the boundary conditions were specified, where th e voltag e for each of th e interfaces was assigned. T h is is the most im portant step pr io r to sim u lation, because the nature of the boundary dictates the electrom a gne tic equation used for calculati on of the electric fields in the system. The boundaries of the top m e tal plate were assigned a sinusoidal voltage, 5 Vp-p in m a gnitude, while those of the lower plate were assigned to ground. The junction of the liqu i d-dielectric interf ace was assign ed as a con tin uous interface and is h e nce, governed by the equation,

PAGE 59

n (J 1 J 2 ) = 0 (4.1) Where n is the normal vector and J 1 J 2 are the current density vectors of the two adjoining materials in consideration. In the next step, the mesh is developed with a predefined normal mesh size for the mesh elements. The meshed structure is shown in figure 4.3. Figure 4.3: 2-D Meshed Model of the Capacitor The sensor was analyzed in the time harmonic, small current mode and the solver used was UMFPACK. The simulated system results for the potential distribution and electric field are shown in figure 4.4. Figure 4.4: Potential and Electric Field Distribution of the Capacitive Sensor in Seawater Figure 4.4 shows the electric field distribution represented by the streamlines while the color gradient represents the potential distribution at various regions in the 47

PAGE 60

capacitor. The post-processing operation involves computation of the sensor capacitance. One of the post-processing results obtained from a quasi-static simulation in FEMLAB is the energy density, E. We know that the energy stored by a capacitor is given by the expression, E = 221CV Consequently, C = 22VE 4.2 Equation 4.2 computes the capacitance from the energy density and the simulated capacitance for the modeled capacitor was obtained as 4.608 pF. The next logical step was the addition of guard rings (held at ground potential) around the capacitor as shown in figure 4.5. Figure 4.5: Sensor with Guard Rings for Field Homogenization Area of inte rest Guard ring Fi gure 4 6 : Si mul a t e d Resul t f o r C a p a ci t o r w i t h G uar d Ri ng 48

PAGE 61

Capacitance for this model was computed using the expression 4.2 and was obtained as 4.1026 pF. As expected, the capacitance for this system was found lower than the sensor without guard ring. The reduction in the computed capacitance was 10.9 %. 4.2 Detection Circuit Employing Twin-T Oscillator The capacitive sensor forms a part of a Twin-T detection circuit. The Twin-T consists to two arms, the high pass arm and the low pass arm. The high pass arm consists of two capacitors and a resistor while the low pass arm, consists of two resistors and a capacitor. This circuit functions as a notch filter by eliminating a particular frequency from the incoming signal. This can be achieved by adjusting the values of the constituent elements of the circuit. The notch frequency is given according to the expression [33], CRF21 4.3 Where R1 = R2 = 2R = R, C1 = C2 = C3 / 2 = C 4.4 A schematic of this circuit is shown in figure 4.7. Figure 4.7: Twin-T Notch Filter Circuit A simulation of this circuit using Pspice, for an input frequency sweep from 100 Hz to 200 KHz, yields the following plot. 49

PAGE 62

Figure 4.8: Twin-T Filter Response Simulated Using Pspice Figure 4.8 shows the simulated result for a Twin T notch filter circuit. It is observed that at ~64 KHz (given by the circuit components), the phase of the output undergoes a near 180 shift. By introducing an additional 180 degrees of phase shift around a closed loop, it is possible to configure this filter as an oscillator, in accordance with Barkhausens criterion for sustained oscillations. The oscillator circuit is as shown in figure 4.9. The capacitors in the circuit were replaced with the capacitive sensors fabricated in house. Also, the resistors used have a tolerance of 2%, to minimize drift in the oscillator response. This kind of a circuit incorporating R and C components can be utilized in applications requiring low frequencies of operation. It can be observed from the expression 4.4 that for proper functioning of this circuit at the desired frequency, it is critical that the resistor and capacitor values are exactly matched. It is important to use resistors from the same manufacturer, as that will ensure closest matching. The resistor in the high pass arm (which is half the other two 50

PAGE 63

resistors) can be realized by running the same two such resistors in parallel. A schematic of the oscillator circuit is as shown in figure 4.9. R2100 C3100n 0 V 0 0 V110Vdc U7LF411 3 2 7 4 6 1 5+-V+V-OUTB1B2 0 C150n C250n V2-10Vdc R450 R1100 Figure 4.9: Schematic of Twin-T Oscillator Figure 4.10: Oscillator Response Simulated Using Pspice The test set-up for the Twin-T oscillator comprising the sensor is as shown in figure 4.11. Calibrated IAPSO Standard seawater solution obtained from Ocean Scientific International, with a salinity of 34.996 psu, was allowed to flow through the gap between the plates. 51

PAGE 64

Figure 4.11: Set-up for Twin-T Oscillator with Capacitive Sensor It is worth noting that other oscillator networks could be chosen to read the capacitive data from the sensor. The advantage of Twin-T oscillator over other single-capacitor circuits is a lower distortion sine wave output. This is one of the prime reasons that this oscillator is very popular in audio electronic circuits. Sensitivity is defined as the slope of the output versus input curve for a system, or mathematically represented as Sensitivity = output / input 4.6 The sensitivity for the frequency response is given by F/C. Differentiating expression 4.3 with respect to C, we obtain Sensitivity = dF / dC = 2**21RCpi 4.7 It is interesting to see that the sensitivity varies inversely as the square of the capacitance. Certain other R-C oscillators that provide the same extent of sensitivity involve more complex designs and components. 52

PAGE 65

4.3 Construction of the Sensor The sensor consists of two parallel conducting (copper) plates. One of the plates has a dimension of 2cm x 1cm and the other plate is built with slightly larger dimensions (2.1cm x 1.1cm) for purposes of alignment. The sensor plates are constructed on copper clad FR-4 substrate boards. The use of FR-4 boards incorporating sensors exposed to solutions of high salt concentrations was done by Natarajan et. al. This fabrication procedure is similar to the process of patterning copper clad boards for electronic circuits. The process flow steps for the fabrication are as depicted in figure 12. Figure4.12: Process Flow Steps for Fabrication of the Sensor 53

PAGE 66

54 The first step towards the fabrication of the sensor is the lithography step. The mask with the pattern was developed using AutoCAD 200 2. The exposure step in the lithography process was done using a UV exposure box. This was followed by the developing stage. The developer used for this photoresist was aq. KOH (9 grams of KOH per liter of Deionized water). This pr ocess was performed at 50 C in the presence of a magnetic stirrer for approximately 15 seconds. Since, positive photoresist is spun on the substrate, the areas exposed to the UV light were de veloped exposing the c opper underneath. After development, the exposed copper was etch ed using Ferric Chloride solution. Subsequently, the photoresist over the copper was dissolved using acetone and the board was now patterned with the plates of required dimensions. Next, a thin layer of a dielectric materi al was coated on the plates. Two sets of experiments were performed using two di fferent dielectric materials (AZ 5214 and Honeywells ACCUFLO) for the purpose of insulating the copper plates from the conducting media. The other purpos e served by this layer is th e charge storage. Seawater can be considered as a lossy dielectric, because of its ability to conduct. Since, the surrounding media (in this case seawater) is cond uctive, there is a need to provide a layer of dielectric for minimizing losse s through the conductive medium. The first material used was the positiv e resist AZ5214. The intended thickness of the coating was ~5 microns. AZ5214 was spin co ated on the substrate. The spin rate was set at 1000rpm and the spin time was 60sec. Once spun, the photoresist was baked on a hotplate at 140 C for 90 sec to dissolve the solv ents and harden the layer. Next stage, in the fabrication was to connect wires to th e contact pads. Tiny patches large enough to solder wires to the pads, were opened up on the dielectric layer, by careful application of

PAGE 67

acetone using cotton swabs. The wires were soldered to the exposed pads. Figure 4.13 shows the top view of the fabricated sensor. W i ndows opened to f acilitate soldering of wires Fi gure 4 1 3 : F abri c at ed Se ns ors P r i o r t o Al i g n m e n t Next step w a s to conceal the exposed c ontact areas again with photoresist. This was done by drop-coating the photor esist over the cont act areas. The plates intended to be arranged in parallel were fabricated on tw o separate substrates. The n e xt stag e was to align them face to face with a spacing of ~600 m i crons. The dim e nsion of the bottom plate was ch osen to b e a little larger than the top plate, as already m e ntioned, to allow for visual alignm ent through the FR 4 s ubstrate. The spacing between the two substrates was provided using two spacers, 600 m i crons thick, glued at each end of the substrate using m a rine glue. Figure 4.14 shows the com p leted sensor. Fi gure 4 1 4 : T o p Vi ew of t h e C o m p l e t e d Se n s or 55

PAGE 68

4.4 Equivalent Circuit Model for the Sensor As evident from the discussion on the construction, the sensor represents a complex dielectric system, comprising dielectrics of different permittivities and conductivities. A model of the sensor is shown in figure 4.15. Figure 4.15: (a) Maxwell-Wagner Layered Capacitor Model, (b) Equivalent Circuit Model, (c) Reduced Model, (d) Equivalent Circuit for Reduced Model, (e) Simplified Model Figure 4.15a shows the Maxwell-Wagner layered model for the sensor. According to the Maxwell-Wagner model, for a capacitor with alternate layers of dielectrics specified by their dielectric constants and conductivities, the complex dielectric can be simplified into 2 layers considering the aggregate thickness for both the materials. [32]. Assuming that the insulation coating has negligible conductivity, the equivalent model can then be represented as shown in figure 4.15c, where d 1 represents the total thickness of the insulation coating and d 2 the total thickness of the seawater. The equivalent circuit model for the reduced capacitor is shown in figure 4.15d, where the bulk conducting solution is represented by a capacitor and resistor in parallel and the insulation layer is represented by a resistor. 56

PAGE 69

4.5 Alternate Detection Circuit for Capacitive Salinity Sensor An alternative circuit for the detection of the capacitive salinity sensor, based on frequency to voltage converters (F-V) has also been proposed. The advantage of this approach over the Twin-T circuit is that it eliminates the issue of frequency stability. Also, since this design employs only one capacitor, it resolves the issue of matching the circuit components, as required by the Twin-T. A block diagram of the circuit is shown in figure 4.16. The principal components of this circuit are: 1) Signal generator (or) voltage controlled oscillator chip. 2) Frequency to Voltage converters (NJM 4151) 3) Buffer and Difference amplifier (LM 6142) 57 Figure 4.16: Block Diagram of Circuit Employing Two F-V Converters and a Differential Amplifier Stage The experimental set-up is as shown in figure 4.17. Function generator / Oscillato r IC 1 KHz sq. wave F-V converter NJM 4151 F-V converter NJM 4151 Buffer and Difference Am plifier Capac itiv e Sens or ~x nF (tested at 1 KHz with im pedan c e anal yser) Dc output Dc output Dc output Precision C a pacitor = x nF

PAGE 70

Figure 4.17: Set-up for the F-V Conversion Approach A detailed schematic of this circuit is shown in figure 4.18. In the schematic shown, the square wave required for the operation of the circuit was generated using a 555 timer. This was done to ensure that the entire circuit could be driven from a single voltage source of 10 V. However, during this experiment the timer was replaced by a function generator. The dc voltage developed on the output of both the F-V converters U1 and U2 is a function of the frequency of the input square-wave, the pulse-width of the one shot inside the F-V converter, the value of the constant current sources internal to the two F-V converters, and the value of the resistor connected from the output to ground. R18 is a 20 turn precision potentiometer, provided for calibration. The potentiometer adjusts the pulse-width of the one shot internal to the reference F-V converter U2 during the calibration of the circuit. The expression governing the output voltage of the F-V converters is given by equation 4.8. [34]. SBINoutRCRRFV486.0*00 4.8 58

PAGE 71

59

PAGE 72

11*486.0)1810(*5*RRRRFVINout 3*C 4.9 From equation 4.9 it may be observed that the sensitivity of the circuit is obtained as, Sensitivity = 111810486.0)(*RRRRFdCdVsinout = constant The sensitivity of this circuit is found to be independent of the capacitance. 60

PAGE 73

61 CHAPTER 5 RESULTS This chapter is broadly segregated into three main areas. The first section presents the data obtained upon the use of AZ 5214 as the photoresist. This in cludes an impedance analysis of the capacitor in seawater and the corresponding Twin-T oscillator response. The next section analyses the data obtai ned upon the use of Honeywells ACCUFLO Spin-on-polymer as the dielectric layer and the corresponding oscillator response. Also included are results obtained for the alternat ive detection circuit with ACCUFLO as the dielectric. 5.1 AZ 5214 as the Insulation Layer AZ 5214 is a thick photoresist with a di electric constant of 6. The barrier properties of this polymer were analyzed using the impedance an alyzer Agilent 4294A Precision Impedance Analyzer. Impedance data were collected for the sensor for varying salinities and temperatures of the solution. Calibration of the cables was performed prior to the experimentation stage. The amplitude of the applied voltage was 100 mV p-p.

PAGE 74

5.1.1 Impedance Data for Varying Salinities at 22 C Initial salinity of the solution was 34.996 psu. Subsequently, small amounts of water were added using a micropipette, to decrease the solution concentration. Figure 5.1 is the Nyquist plot for the sensor with the real part of the impedance plotted on the X axis and the imaginary on Y. The semi-circle shape observed clearly indicates that this circuit can be treated as a Randels cell as explained in Chapter 3. Nyquist plot1.10E+031.60E+032.10E+032.60E+033.10E+033.60E+034.10E+034.60E+035.10E+030.E+002.E+034.E+036.E+038.E+031.E+041.E+04Re(Z)-Im(Z) 34.966 34.961 34.926 34.891 34.856 34.821 Figure 5.1: Plot of Complex Impedance for Different Salinity Concentrations (psu) This shape of the Nyquist plot implies a low pore resistance of the coating on the electrode [36]. It may be observed that at lower frequencies, the cumulative resistance is given by R sol + R t This is indicative of the fact that, space charge polarization occurs at lower frequencies. It can be inferred that at higher frequencies the polarization effects subside and the only resistive component is the R sol It is further observed from the above plot that as the salinity decreases, on the right corner, the curves tend to shift towards the left. This implies an increase in the polarization resistance as the concentration increases. This suggests that chloride anion 62

PAGE 75

63 adsorption occurs on the electrode, favoring the blockage of activ e sites on the metal surface [36]. Also observed from the graph is the solution resistance at higher frequencies increases as the salinity levels decrease. This is logical, since, addition of ions decreases the solution resistance. From the Nyquist plot the follo wing parameters were obtained. Table 5.1: Parameters for the Randels Cell Obtained from the Nyquist Plot Salinity (psu) 34.996 34.961 34.926 34.891 34.856 34.821 Rs + Rt 8925.348 8685.2300 8550.3 440 8427.0640 8425.8980 8.42E+03 w_max 3.04E+03 3.29E+03 3.29E+0 3 3.29E+03 3.54E+03 3.29E+03 Rs 406.1793 407.6288 408.1606 408.5595 408.8128 4.09E+02 Rt 8519.17 8277.61 8142.18 8019.51 8017.07 8.02E+03 Cdl (=1 / (Rt w_max 1.32E-08 1.39096E08 1.4364E-08 1.47971E08 1.48036E08 1.48057E08 From the above table, it is seen that the double layer capacitance increases with decrease in concentration of the solution. This is attributed to the increase in diffusion capacitance as explained in Chapter 3. Figure 5.2 shows a plot of the capacitan ce versus frequency for varying salinity levels in seawater. The sweep frequency wa s 40 Hz to 100 KHz. It may be noted that there is a steep fall in the capacitance of the sensor with increasing frequency. This follows from the discussion on polarization in Chapter 3. At low frequencies, the effects of orientation polarization dominate and hence, the permittivity is high, leading to higher capacitance values. An interesting observation he re is the level of magnitude increase at low frequencies. While at higher frequencies, the capacitance was observed to be in the nF range; the capacitances for frequencies < 1 KHz are in the F range.

PAGE 76

Capacitance Vs. Frequency for varying salinities0.00E+005.00E-071.00E-061.50E-062.00E-062.50E-063.00E-063.50E-064.00E-064.50E-061.00E+011.00E+021.00E+031.00E+041.00E+05log Frequency (Hz)Capacitance (F) 34.996 psu 34.961 psu 34.926 psu 34.891 psu 34.856 psu 34.821 psu 34.786 psu Figure 5.2: Plot of Capacitance Versus Frequency for Varying Salinities at 22 C This is due to two reasons. Firstly, the effects of orientation polarization come into play at lower frequencies. Secondly, a path linking the electrolyte to the electrode through the barrier layer was formed, causing space charge polarization. This leads to a tremendous and anomalous increase in permittivity at low frequencies [32]. This link between the electrolyte and electrode causes film degradation due to diffusion of ions, water absorption and can result in flaking of the film [28]. These are also responsible for greater variations in capacitance (with increasing salinity levels) at lower frequencies. Shown in figure 5.3 is a plot of the capacitance versus salinity for a 40 KHz signal at 22 C. 64

PAGE 77

Capacitance Vs. Salinity for 40 KHz at 22 C4.92E-094.94E-094.96E-094.98E-095.00E-0934.7534.834.8534.934.953535.05Salinity (psu)Capacitance (F) Figure 5.3: Plot of Capacitance Versus Salinity for 40 KHz at 22 C This was plotted from the data used in figure 5.2. From this plot it is observed that with increasing salinity, the capacitance of the sensor decreases. This follows from the discussion in section 3.10, which states that as ions are added to water, the hydrate shell around them makes their rotation difficult. As a result, the permittivity of the solution decreases. The decrease in salinity from this plot was observed to be 0.02 nF (or 0.4%) for every 0.035 psu change in salinity. Hence, the sensitivity of the sensor was found to be 0.57 nF / psu. 5.1.2 Impedance Data for Varying Temperatures at Constant Salinity The temperature response of the sensor was obtained at the final concentration of 34.821 psu, after dilution with de-ionized water. Temperature was varied using a hotplate. The solution temperature was measured periodically using a mercury thermometer and a magnetic stirrer was used to maintain uniformity in the temperature profile throughout the solution. The Nyquist plot for this experiment is as shown in figure 5.4. The semi-circles are observed to shift to the left with increasing temperatures. This is indicative of the fact that as temperature increases the resistance of the solution and the film decreases. 65

PAGE 78

An explanation to this is that an increase in temperature increases the rate of diffusion of the electrolyte into the photoresist layer, thereby, reducing its resistance. Also, the rate of absorption of water in the film increases with increasing temperatures. This increases the effective dielectric constant of the film, increasing its capacitance. Nyquist plot for varying temperatures1100160021002600310036004100460051000.00E+002.00E+034.00E+036.00E+038.00E+031.00E+04Re(Z)-Im(Z) 20 25 30 35 40 Figure 5.4: Nyquist Plot for Varying Temperatures (Salinity = 34.821 psu) The Randels cell parameters for this plot and are as shown in table 5.2. Table 5.2: Randels Cell Parameters for Varying Temperatures at Constant Salinity (34.821 psu) Temperature (C) 20 25 30 35 40 Rs + Rt 8.55E+03 8.43E+03 8.32E+03 8.27E+03 8.26E+03 w_max 3.04E+03 3.29E+03 3.29E+03 3.29E+03 3.29E+03 Rt 8151 8033 7.93E+03 7941 7931 Rs 3.99E+02 3.97E+02 3.95E+02 3.92E+02 3.87E+02 Cdl (=1 / (Rt w_max 1.43E-08 1.47671E-08 1.51662E-08 1.52E-08 1.52648E-08 Figure 5.5 shows the variation of capacitance of the sensor versus frequency for varying temperatures at a constant salinity of 34.821 psu. From the graph it was observed that with an increase in temperature the capacitance of the sensor was found to decrease. This can be attributed to the reduction of orientation polarization as temperature increases. The 66

PAGE 79

change in capacitance at 40 KHz for a 5 C change in temperature was observed to be 0.002 nF. Therefore, the sensitivity of the sensor was found to be 0.0004 nF/ C. Capacitance Versus Frequency for varying temperatures at constant salinity (34.821 psu)0.00E+005.00E-071.00E-061.50E-062.00E-062.50E-063.00E-063.50E-064.00E-064.50E-065.00E-061.00E+011.00E+021.00E+031.00E+041.00E+05log frequency (Hz)Capacitance (F) 40C 35C 30C 25C 22C Figure 5.5: Plot of Capacitance Versus Frequency for Varying Temperatures at Constant Salinity 5.1.3 Twin-T Oscillator Frequency Versus Salinity Figure 5.6 shows the Twin-T oscillator response to the variation in salinity of seawater. The frequency drift in the output for constant salinity of the solution was ~400 Hz. The results obtained for the Twin-T oscillator with ACCUFLO as the dielectric will make clearer the fact, that this frequency drift can be attributed in part to the poor dielectric characteristics of the photoresist. 67

PAGE 80

oscillator frequency versus salinity4243444546474834.434.534.634.734.834.93535.1salinity (psu)oscillator frequency (KHz) Figure 5.6: Twin-T Oscillator Frequency Versus Salinity at 22 C The resistors used in the Twin-T network were 1K resistors (2% tolerance). The frequency of the oscillator was observed to increase with increasing salinity. The oscillator frequency is inversely proportional to the capacitance of the sensor and hence, as salinity increased, the capacitance decreased resulting in increased frequencies. The change in frequency was observed to be ~500 Hz for 0.0875 psu change in salinity. 5.1.4 Twin-T Oscillator Frequency Versus Temperature Oscillator Frequency Vs. Temperature4242.54343.54444.515202530354045Temperature (C)Oscillator Frequenc y (KHz) Figure 5.7: Twin-T Oscillator Frequency Versus Temperature at 34.471 psu 68

PAGE 81

69 From the plot in Figure 5.7, it is observed that as temperature increases the oscillator frequency decreases. This is becau se the sensors capacitance decreases with increasing temperature. Hence, frequency of the oscillator increases with temperature. The frequency was observed to shift by 0.12 KHz for every 5 C. The sensor was removed from the solution after 48 hours of immersion. The photoresist seemed to have almost leached out of the electrode surface. There are two possible explanations for this. Firstly, the de gradation of the film is caused due to water absorption and secondly, due to slow developing of the resist in the seawater. Photoresists are known to deve lop in alkaline solutions a nd seawater has a pH ~8 [1] making it slightly alkaline. This may have resu lted in accelerated l eaching of the resist layer. 5.2 ACCUFLO Spin-on-Polymer as the Insulation Layer Honeywells ACCUFLO Spin-on-Polymer (T3027), was used for the second set of experiments. Impedance analysis on the sens or was performed to test the strength of the barrier layer in seawater. 5.2.1 Impedance Data for Varying Salinities at 22 C The Nyquist plot for the sensor with this coating is as shown in figure 5.8a. From [35], we have that the equivale nt circuit model for this plot is as shown in figure 5.8b. According to GAMRY, this is the Nyquist plot for a good coating on the electrolyte as the impedance of the coating is found to be ex tremely high. This is because the film has not yet absorbed any water. Hence its permittivity is as expected and not anomalously

PAGE 82

high. The impedance of the layer was found to reduce with concentration increase. Thus, it can be inferred from the Nyquist plots that ACCUFLO is more suitable as an insulation layer than AZ5214, for this application. Nyquist plot for varying salinities at 22 C0200000004000000060000000800000001000000001200000001400000001600000001800000001.E+031.E+062.E+063.E+064.E+065.E+066.E+06Re z (Ohms)-Im Z (Ohms ) 34.996 psu 34.961 psu 34.926 psu 34.891 psu 34.856 psu a) b) Figure 5.8: (a) Plot of Complex Impedance for Varying Salinities at 22 C (b) Equivalent Circuit Model for Metal-Electrolyte System Corresponding to the Plot in (a) Figure 5.9 shows a plot of the capacitance versus frequency at varying salinities. Once again, capacitance is observed to fall as frequency increases. From this plot it can be seen that the capacitance drops as salinity increases. The reason for this drop is similar to that explained in the previous section. It may be noted that the variation of capacitance with frequency is not as pronounced as it was for the photoresist. Thus, it can be inferred that this coating exhibits better dielectric properties over wider frequencies than the photoresist. 70

PAGE 83

Capacitance vs. frequency for varying salinities (22 C)0.00E+005.00E-101.00E-091.50E-092.00E-092.50E-093.00E-093.50E-094.00E-091.00E+011.00E+021.00E+031.00E+041.00E+051.00E+06log frequency (Hz)Capacitance (F) 34.856 psu 34.891 psu 34.926 psu 34.961 psu 34.996 psu Figure 5.9: Plot of Capacitance Versus Frequency for Varying Salinities at 22 C Shown in figure 5.10 is a plot of the capacitance versus salinity for a 2 KHz signal at 20 C. This is from the data used for plotting figure B. For every 0.035 psu change in salinity, the change in capacitance observed was 0.01 nF. The sensitivity for this sensor was thus, obtained as 3.5 nF / psu. Capacitance versus salinity for 2 KHz at 22 C2.76E-092.77E-092.77E-092.78E-092.78E-092.79E-0934.834.8534.934.953535.05Salinity (psu)Capacitance (F) 71 Figure 5.10: Plot of Capacitance Versus Salinity for a 2 KHz Signal at 22 C

PAGE 84

5.2.2 Impedance Data for Varying Temperatures at Constant Salinity The Nyquist plot for varying temperatures of the solution is given in figure 5.11. The resistance of the film is found to decrease with increasing temperatures. The vertical lines obtained are indicative of the fact that the film has not yet absorbed any water from the medium. Nyquist plot 0.00E+002.00E+074.00E+076.00E+078.00E+071.00E+081.20E+081.40E+081.60E+081.80E+080.00E+001.00E+062.00E+063.00E+064.00E+065.00E+066.00E+06Re Z (ohms)-Im Z (ohms ) 22C 25C 30C 35C Figure 5.11: Nyquist Plot for Varying Temperatures at Constant Salinity (34.856 psu) Figure 5.12 shows a plot of the capacitance versus frequency for varying temperatures, at 34.856 psu. It is observed that for every 5 C increase in temperature, the capacitance was found to decrease by 0.14 nF. Thus, the sensitivity of the device for varying temperatures is 0.028 nF / C. 72

PAGE 85

Capacitance versus frequency for varying temperatures at 34.856 psu0.00E+001.00E-092.00E-093.00E-094.00E-095.00E-091.00E+011.00E+021.00E+031.00E+041.00E+051.00E+06frequency (Hz)Capacitance (F) 35 C 30 C 25 C 20 C Figure 5.12: Plot of Capacitance Versus Frequency for Varying Temperatures at 34.856 psu 5.2.3 Twin T Oscillator Response with Spin-on-polymer as Insulation Layer The Twin-T oscillators response was found for varying salinities at 22 C and is plotted in figure 5.13. The oscillator frequency was observed to increase by 0.058 KHz for every 0.14 psu change in salinity which implies 0.41 KHz / psu. Frequency Vs. Salinity at 22 C2.052.12.152.22.2534.234.334.434.534.634.734.834.93535.1Salinity (psu)Oscillator frequenc y (KHz) Figure 5.13: Oscillator Frequency Versus Salinity at 22 C Another point of interest with this material as the insulation layer was the good stability achieved in the output of the oscillator. The frequency was observed to drift only by ~10 Hz about a mean value at constant salinity and temperature. 73

PAGE 86

5.2.4 Twin-T Oscillator Frequency Versus Temperature The oscillator response for a constant salinity of 32.891 psu and varying temperatures is as shown in figure 5.14. Oscillator frequency versus temperature at constant salinity (32.891 psu)1.531.541.551.561.571.581.591.61.611.62202530354045Temperature (C)Frequency (KHz ) Figure 5.14: Plot of Oscillator Frequency Versus Temperature at Constant Salinity of 32.891 psu An interesting point to note is that for the same values of the R and C components in the Twin T network, the oscillation frequency can be made lower, by introducing a little positive feedback. This has a latching effect on the op-amp, and it tends to stay in the saturated state for a longer duration, thereby, reducing the output frequency. However, there is no expression that can relate the frequency to the amount of feedback applied. 5.3 Response of F-V Converter Circuit From the impedance analysis performed earlier, it was inferred that Honeywells ACCUFLO spin-on-polymer is more suitable as a dielectric layer for this application than photoresist AZ5214. Hence, this experiment was performed with ACCUFLO as the barrier layer. 74

PAGE 87

5.3.1 Output Voltage Versus Salinity at 22 C The output voltage of the schematic shown in figure 4.17 is shown in figure 5.15. The output voltage was found to decrease with increasing salinity. This is because the output voltage for this circuit, as seen in expression 4.9, is proportional to the sensor capacitance. Output Voltage Vs. Salinity4.2554.2564.2574.2584.2594.264.2614.26234.634.734.834.93535.1Salinity (psu)Output voltage (V ) Figure 5.15: Output Voltage Versus Salinity at 22 C For a change in salinity of 0.07 psu, the change in output voltage was obtained as 0.7 mV. Thus, the sensitivity of this detection circuit is 10 mV / psu. 5.3.2 Output Voltage Versus Temperature at Constant Salinity The effect of temperature on the output voltage was studied in this experiment, at constant salinity of 34.65 psu. The output voltage was observed to follow a decreasing trend with temperature due to the capacitance drop with increasing temperature. 75

PAGE 88

Output Voltage Vs. Temperature4.264.2614.2624.2634.2644.2654.26615202530354045Temperature (C)Output Voltage (V ) Figure 5.16: Output Voltage Versus Temperature at 34.65 psu Sensitivity of this approach to variations in temperature was found to be equal to 6.3 mV /C. 5.4 Comparative Study A comparison chart has been shown below, summarizing the results obtained for both the materials used as well as the detection techniques. Table 5.3: Comparison Chart for Materials Used and Detection Techniques Proposed Impedance Analysis using 4294 A Precision Impedance Analyzer Twin-T Oscillator Detection Frequency-Voltage Conversion Detection Barrier layer Salinity Variation (at 22 C) Varying temperatures Salinity variation Temperature Variation Salinity Variation Temperature Variation AZ 5214 Sensitivity 0.57 nF / psu (at 40 KHz) 0.0004 nF / C (@ 34.821 psu) 500 Hz / 0.035 psu 0.12 KHz / 5 C NA NA Honeywells ACCUFLO Sensitivity 3.5 nF/psu (at 2 KHz) 0.028 nF/ C 0.41 KHz /psu 0.04 KHz / 5 C 10 mV/psu 6.3 mV/ C 76

PAGE 89

77 CHAPTER 6 CONCLUSION AND FUTURE WORK 6.1 Conclusion A parallel-plate capacitive sensor was fabricated to detect varying salinity levels in the marine environment. Impedance an alysis was done on the capacitor using two separate materials as the dielectric layer be tween the electrode and seawater. The spinon-polymer ACCUFLO was fou nd to have better dielectric properties over the photoresist AZ5214, which was inferred base d on the Nyquist plot s obtained for both materials. Optimization of the capacitor wa s performed using FEMLAB, by addition of guard ring around one of the electrodes. The capacitance was observed to fall by 10% offering fringe field reducti on and minimized effects due to interference. Two detection strategies have been proposed for the sensor. The Twin-T oscillator provided a sensitivity of 500 Hz for every 0.0375 psu change in salinit y with photoresist as the dielectric, while the sensitivity of the oscillator with ACCU FLO as the dielectric layer was obtained as 410 Hz/ psu. However, frequency stability of the oscillator was enhanced upon the use of ACCUFLO. The frequency-to-voltage approach was suggested as an improvement since it eliminates the effects due to any drift in source frequency. The sensitivity of this detection circuit was obtained as 10 mV/psu.

PAGE 90

78 6.2 Future Work More robust materials need to be used as the dielectric layer for electrode protection. Parylene or Spinon-Teflon may be tried as alternatives for the polymers used in this research, which provide more effective protection from corrosive media. Also, materials which exhibit negligible dielectric constant variation with temperature need to be investigated. This will assi st to strengthen the argument th at the change in permittivity caused by a corresponding change in temperature, is purely due to the conducting media, with negligible or no contribution from the insu lation layer. Integrati on of the sensor and the electronics into one substrate will reduce the parasitic effects and capacitances due to the cables. Frequency stabilization circuits need to be incorporated to the Twin-T oscillator design to minimize the drift at constant sali nity levels. A study of the temperature response of the sensor needs to be performed using a constant temperature bath, and for temperatures lower than 20 C. This is because, in general, ocean temperatures can vary between 0 C 45 C. The fabrication techni que used here is the conventional PCB fabrication t echnique which is associated with significant undercuts. Although, the undercuts are advantageous fo r this sensor from the fringe field minimization point of view, an effort needs to be made to quantify this effect using FEMLAB. The effects of biofouling have not been studied in this work. This is an important consideration during deployment of the sensor in marine environments. Steps to minimize biofouling need to be incorporated into the system for accurate measurements. Reproducibility can be enhanced by ensuring that the design is coplanar. This is because coplanar arrangement can eliminate the errors due to mechanical placement of the spacers between the electrodes. This will require the use of IDT sensors

PAGE 91

79 over biplanar capacitors. Finally, study needs to be performed to observe the hysterisis properties of the sensor over increasing a nd decreasing cycles of both salinity and temperature.

PAGE 92

80 REFERENCES 1. Thurman H.P. and Trujillo A.P., Essentials of Oceanography Prentice Hall, 2002 2. Chapter 6: Temperature, Salinity and Density, from http://oceanworld.tamu.edu/resource s/ocng_textbook/chapter06/chapter06_01.ht m 3. Sverdrup H.U., Oceanography for Meteorologists George Allen and Unwin Ltd., 1952 4. The Blue Planet, from, http://www. globalchange.umich.edu/ globalchange1/ current/lectures/ samson/blue_planet/ 5. Jury S.H., Kinnison M.T., Howell W.H., Watson W.H., The effects of reduced salinity on lobster (Homarus Ameri canus Milne-Edwards) metabolism: implications for estuarine populations, J. Exp. Mar. Biol. Ecol. 176: 167-185, 1994 6. NASA Oceanography, Sea Surface Salinity, from http://science.hq.nasa.gov/oceans/physical/SSS.html 7. Seabird Electronics, FastCAT CTD Se nsor: SBE 49, from http://www.commtec.com/Prods/mfgs/SBE/brochur es_pdf/49brochurelores01-02.pdf 8. Barnes H., Apparatus and Methods of Oce anography Part One: Chemical George Allen and Unwin Ltd., 1959 9. Tucker M.J., Measurement in Oceanography, J. Phys. E: Sci. Instrum. 4 405413, 1971 10. A.P.C.S. Knowledge, Conductivity Measurement, from http://www.apcs.net.au/hlp/hlp0005/hlp0005.html 11. TOPAC, Conductivity Cells, from http://www.topac.com/conductivityprobes.html 12. Striggow K. and Dankert K., The Ex act Theory of Inductive Conductivity Sensors for Oceanograp hic Application, IEEE Journal of Oceanic Engineering Vol. OE-10, No.2, 1985

PAGE 93

81 13. Hart L.W., Ko H.W., Meyer J.H, Vas holz D.P., Joseph R.I., A Noninvasive Electromagnetic Conductivity Sensor for Biomedical Applications, IEEE Transactions on Biomedical Engineering Vol.35, No.12, December 1988 14. Guardo R., Charron G., Goussard Y. A nd Savard P., Contactless Measurement of Thoracic Conductivity Changes by Magnetic Induction, Proceedings 19th International Conference IEEE/EMBS 1997 15. Natarajan S.P., Huffman J., Weller T.M., Fries D.P., Contact-less Toroidal Fluid Conductivity Sensor Based on RF Detection, IEEE Sensors 2004 16. Sea-Bird Electronics Inc., Conductivity Sensors for Moored and Autonomous Operation, from http://www.seabird.co m/technical_references/condpaper.htm 17. Langereis G.R., An Integrated Sens or System for Monitoring Washing Purposes, vol. Thesis, University of Twente, the Netherlands, 1999 18. Timmer B.H., Sparreboom, Olthuis W., Be rgveld P., van den Berg A., Planar Interdigitated Conductivity Sensors for Low Electrolyte Concentrations, Proceedings Semiconductor Sens or and Actuator Technology pp. 878-883, 2001 19. Fritz Wolter and Fritz Thom, A Parallelplate Capacitor used to determine the complex permittivity of supercooled aque ous solutions in the 1 MHz range, Meas. Sci. Technol., 7 p969-975, 1996 20. Mamishev A.V., Sundara-Rajan K., Yang F., Du Y. and Zahn M., Interdigital Sensors and Transducers, Proceedings of the IEEE Vol. 92, No. 5, 2004 21. Hilland J., Simple Sensor System for measuring the Dielectric Properties of saline solutions, Meas. Sci. Technol. 8 p901-910, 1997 22. Hyldgard A., Hansen O., Thomsen E.V ., Fish & Chips: Single Chip MEMS CTDL Salinity, Temperature, Pressure and Light Sensor for use in Fisheries Research, MEMS 2005 23. Ulaby F.T., Fundamentals of Applied Electromagnetics, Prentice Hall, 2001 24. Agilent application note, Basics of M easuring the Dielectric Properties of Materials, from http://cp.literature. agilent.com/litweb/pdf/5989-2589EN.pdf 25. Mechanisms of Polarization from http://www.tf.unikiel.de/matwis/amat/elmat_en/kap_3/backbone/r3_2_1.html 26. Hill N.E., Vaughan W.E., Price A.H. and Davies M., Dielectric Properties and Molecular Behavior Van Nostrand Reinhold Company, London, 1969

PAGE 94

82 27. Kotz. C. John, Triechel P., Chemistry and Chemical Reactivity Saunders College Publishing, 1996 28. Taylor R.S., Assessing the Moisture Barri er Properties of Polymeric Coatings Using Electrical and Elec trochemical Methods, IEEE Transactions on Electrical Insulation Vol. 24, No.5, 1989 29. McDonough Laurie A., Microscopy and Spectroscopy of water uptake in polymer photoresists, Ph.D. Thesis, University of Colorado, 2004 30. Macdonald R., Impedance Spectroscopy: Emphasiz ing Solid Materials and Systems, Wiley Inter Science Publication, 1987 31. Heerens W. Chr., Multi-terminal capacitor sensors, in J. Phys.E: Sci. Instrum., Vol.15, 1982 32. Lu Z., Cho H., Manias E., Macdonald D. D., Dielectric Relaxation Spectroscopy Studies on Water-Saturated Nafion 117 Membranes, 204 th Meeting of the Electrochemical Society, 2003 33. R. Mancini, Op Amps for Everyone: Design Reference, Texas Instruments 34. Electronic Datasheet, NJM4151: V-F/FV Converters, New Japan Radio Co., Ltd. from http://www.njr.co.jp/pdf/be/be06016.pdf 35. Gamry Instruments, Application notes from http://www.gamry.com/App_Notes/Index.htm 36. Gonzalves R.S., Azambuja D.S., Lucho A.M.S., Reche M.P., Schmidt A.M., electrochemical Studies of Copper, Nickel and a Cu55/Ni45 Alloy in Aqueous Sodium Acetate, Mat. Res. vol.4 no.2, 2001