USF Libraries
USF Digital Collections

Theoretical modeling of cortisol sensor

MISSING IMAGE

Material Information

Title:
Theoretical modeling of cortisol sensor
Physical Description:
Book
Language:
English
Creator:
Gordic, Milorad
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Dielectrophoresis
Square Wave Voltammetry
BioMEMS
Electrochemistry
Thiol chemistry
Dissertations, Academic -- Chemical Engineering -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This thesis describes the theoretical modeling of a response of an electrochemical BioMEMS sensor for detecting small amounts of cortisol hormone. The electrochemical sensor utilizes a catalyst enzyme (3α-HSD) to convert cortisone to cortisol and the Square Wave Voltammetry (SWV) as a preferred method to measure the forward and reverse current of the system. The parameters and equations necessary to estimate the Square Wave Voltammetry (SWV) theoretical response are determined and outlined. The response is modeled and the results are compared to the experimental data. Further, the design of the sensor is analyzed and suggestions are made on how to improve the repeatability of the sensor's response. The diffusion coefficients for cortisone and cortisol hormone are calculated to be 2.87*10⁻¹⁰ and 2.84*10⁻¹⁰ square meters per second respectively with 10 percent tolerance. The dimensionless peak current (ψ) for the system is approximately 10 percent lower than the one theoretically postulated by Bard et al. 3. The surface area of the working electrode of the sensor varies with and is directly proportional to the concentration of the analyte. Theoretical current peaks are hypothesized to be within 10 percent tolerance limits (mainly due to the reason that the surface area of the working electrode is itself a variable).
Thesis:
Thesis (M.S.)--University of South Florida, 2008.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Milorad Gordic.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 71 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 - 002046387
oclc - 496014722
usfldc doi - E14-SFE0002689
usfldc handle - e14.2689
System ID:
SFS0027006:00001


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 2200397Ka 4500
controlfield tag 001 002046387
005 20100106141107.0
007 cr bnu|||uuuuu
008 100106s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002689
035
(OCoLC)496014722
040
FHM
c FHM
049
FHMM
090
TP145 (Online)
1 100
Gordic, Milorad.
0 245
Theoretical modeling of cortisol sensor
h [electronic resource] /
by Milorad Gordic.
260
[Tampa, Fla] :
b University of South Florida,
2008.
500
Title from PDF of title page.
Document formatted into pages; contains 71 pages.
502
Thesis (M.S.)--University of South Florida, 2008.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Shekhar Bhansali, Ph.D.
3 520
ABSTRACT: This thesis describes the theoretical modeling of a response of an electrochemical BioMEMS sensor for detecting small amounts of cortisol hormone. The electrochemical sensor utilizes a catalyst enzyme (3-HSD) to convert cortisone to cortisol and the Square Wave Voltammetry (SWV) as a preferred method to measure the forward and reverse current of the system. The parameters and equations necessary to estimate the Square Wave Voltammetry (SWV) theoretical response are determined and outlined. The response is modeled and the results are compared to the experimental data. Further, the design of the sensor is analyzed and suggestions are made on how to improve the repeatability of the sensor's response. The diffusion coefficients for cortisone and cortisol hormone are calculated to be 2.87*10 and 2.84*10 square meters per second respectively with 10 percent tolerance. The dimensionless peak current () for the system is approximately 10 percent lower than the one theoretically postulated by Bard et al. [3]. The surface area of the working electrode of the sensor varies with and is directly proportional to the concentration of the analyte. Theoretical current peaks are hypothesized to be within 10 percent tolerance limits (mainly due to the reason that the surface area of the working electrode is itself a variable).
653
Dielectrophoresis
Square Wave Voltammetry
BioMEMS
Electrochemistry
Thiol chemistry
690
Dissertations, Academic
z USF
x Chemical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2689



PAGE 1

Theoretical Modeling of Cortisol Sensor By Milorad Gordic A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Bi omedical Engineering Department of Chemical a nd Biomedical Engineering College of Engineering University of South Florida Major Professor: Shekhar Bhansali, Ph.D. Alberto Sagues, Ph.D. Vinay Gupta, Ph.D. Arun Kumar, Ph.D. Shyam Aravamudhan, Ph.D. Date of Approval: October 27, 2008 Keywords: Dielectrophoresis, Square Wave Voltammetry, BioMEMS, Electrochemistry, Thiol Chemistry Copyright 2008, Milorad Gordic

PAGE 2

DEDICATION To all students who are curre ntly pursuing thei r education while being a full-time employee There were countless moments when I thought of quitting while working on this thesis, but somehow found a reason not to To my en tire family who supported me and always encouraged me, never to give up To all who prayed for me every night for enlightenment and patience to finish this thes is (especially my wife Daniella, my mother Anica, my sister Ruzica, and my now deceas ed grandmother Pera and nanny Manda) And to all whom I may be forgetting... Tha nk you from the bottom of my heart! This dissertation is dedicat ed to all of you.

PAGE 3

ACKNOWLEDGEMENTS I would like to exte nd special thanks to Dr. Sh ekhar Bhansali and Dr. Arun Kumar for inviting me to be a part of the Cortis ol Sensor project. I al so wish to thank Dr. Alberto Sagues for his help and guidance w ith electrochemistry. Thanks to Dr. Shyam Aravamudhan, Dr. Niranjan Ramgir, Ke, Do rielle and Brian for assisting me in fabrication and analysis of the sensor itself. I would like to extend thanks to all of the committee members for their input to the project And last but not least, thank you, Lord, for all your blessings.

PAGE 4

i TABLE OF CONTENTS LIST OF TABLES .............................................................................................................iiiLIST OF FIGURES ...........................................................................................................ivABSTRACT .......................................................................................................................viCHAPTER 1: HISTORY OF CORTISOL RESEARCH ..................................................11.1Introduction to Cortisol (Hydrocortisone) ..............................................................21.2Synthesis of Cortisol Hormone ...............................................................................31.3Secretion of Cortisol Hormone...............................................................................61.4Need for Cortisol Sensing .......................................................................................71.4.1Need for Cortisol Sens or in Medical Field .........................................................71.4.2Need for Cortisol Sensor in Other Fields ............................................................81.5Collection of Sample ...............................................................................................91.6Motivation .............................................................................................................10CHAPTER 2: CURRENT METHODS OF CORTISOL ANALYSIS ............................112.1Radioimmunoassay (RIA) ....................................................................................112.2Competitive Protein Binding (CPB) .....................................................................13

PAGE 5

ii 2.3Fluorometric Method ............................................................................................152.4High Performance Liquid Chromatography (HPLC) ...........................................172.4.1Reverse-phase HPLC (RP HPLC) ....................................................................172.5Electrochemical BioMEMS Cortisol Sensor ........................................................18CHAPTER 3: THEORY..................................................................................................243.1Dielectrophoresis ..................................................................................................243.2Electrochemistry ...................................................................................................273.2.1Electrode Function Breakdown .........................................................................283.2.2Fundamentals of Electrochemistry ...................................................................283.2.3Square Wave Voltammetry (SWV) ..................................................................323.2.4Construction of Electrochemical Cell ...............................................................39CHAPTER 4: RESULTS .................................................................................................424.1Calculation of Diffusion Coefficients ...................................................................424.2 Calculation of Peak Dimensionless Current .........................................................474.3Calculation of Surface Ar ea of Working Electrode ..............................................524.4Calculation of Total Peak Current for an Electrochemical Response ...................55CHAPTER 5: DISCUSSION AND FUTURE WORK ...................................................615.1Discussion .............................................................................................................615.2Future Work ..........................................................................................................67REFERENCES .................................................................................................................69

PAGE 6

iii LIST OF TABLES Table 3.1: Dimensionless Peak Current vs. SWV Operating Parameters...................37 Table 4.1: Calculation of Surface Ar ea Using Dimensionless Peak Current From Figure 4.3.........................................................................................53 Table 4.2: Calculation of Surface Ar ea Using Dimensionless Peak Current From Table 3.1...........................................................................................53

PAGE 7

iv LIST OF FIGURES Figure 1.1: Synthesis of Cortisol....................................................................................4 Figure 1.2: Struct ure of Cortisol (C21H30O5)..................................................................5 Figure 1.3: Cortisone Cortisol Conversion...............................................................5 Figure 1.4: Secretion of Cortisol....................................................................................6 Figure 2.1: Ra dioimmunoassay....................................................................................12 Figure 2.2: Competitiv e Protein Binding.....................................................................14 Figure 2.3: Fluorometric Measurement........................................................................15 Figure 2.4: High Performa nce Liquid Chromatograph................................................18 Figure 2.5: Electrochem ical BioMEMS Sensor...........................................................19 Figure 2.6: Cortisol Detection Scheme........................................................................20 Figure 2.7a: SWV at Different C oncentrations of Cortisol (10-50 ).........................21 Figure 2.7b: SWV at Different C oncentrations of Cortisol (60-80 ).........................22 Figure 2.8: Calibration Curve for Cortisol...................................................................23 Figure 3.1: Forward and Revers e Current in El-Chem System....................................32 Figure 3.2: Time-Potential Profile for SWV................................................................33 Figure 3.3: Square Wave Voltammogram....................................................................34 Figure 3.4: Dimensionless Current...............................................................................36 Figure 3.5: Graphical Repres entation of El-Chem Cell...............................................40

PAGE 8

v Figure 3.6: Block Diagram of Apparatus for El-Chem Analysis.................................41 Figure 4.1a: Views of Cortisone Molecule....................................................................44 Figure 4.1b: Rectangular Prism Unit Cell......................................................................45 Figure 4.1c: Structural Info of One Molecule of Cortisone...........................................45 Figure 4.2: Simulation of SW Excitation on Working Electrode.................................49 Figure 4.3: Simulation of Dimensionless Current........................................................50 Figure 4.4: Simulation of Dime nsionless Current Over Time.....................................51 Figure 4.5: Graphical Repres entation of Whatman Disc..............................................54 Figure 4.6: Theoretically Calculat ed SW Voltammogram for Different Concentrations of Cortisone/Cortisol........................................................58 Figure 4.7: Theoretically Calcul ated Current Based on Different Concentrations of Cortisone/Cor tisol Using Dimensionless Peak Current From Figure 4.3............................................................................59 Figure 4.8: Theoretically Calcul ated Current Based on Different Concentrations of Cortisone/Cor tisol Using Dimensionless Peak Current From Table 3.1..............................................................................60 Figure 5.1: Random Alignment of Nanowires.............................................................64 Figure 5.2: Ag/AgCl Reference Electrode...................................................................66 Figure 5.3: Graphical Re presentation of Design Improvement for BioMEMS El-Chem Sensor.........................................................................................68

PAGE 9

vi Theoretical Modeling of Cortisol Sensor Milorad Gordic ABSTRACT This thesis describes the th eoretical modeling of a res ponse of an electrochemical BioMEMS sensor for detecting small amounts of cortisol hormone. The electrochemical sensor utilizes a catalyst enzyme (3 -HSD) to convert cortisone to cortisol and the Square Wave Voltammetry (SWV) as a preferred method to measure the forward and reverse current of the system. The parameters and equations necessary to estimate the Square Wave Voltammetry (SWV) theoretical res ponse are determined and outlined. The response is modeled and the results are compared to the experimental data. Further, the design of the sensor is analyzed and s uggestions are made on how to improve the repeatability of the sensors response. The diffusion coefficients for cortisone and cortisol hormone are calculated to be 2.87*10-10 and 2.84*10-10 m2/s respectively with 10% tolerance. The dimensionless peak current ( ) for the system is ~10% lower than the one theoretically postulated by Bard et al. [3]. The surface area of the worki ng electrode of the sensor varies with and is directly proportional to the concentration of the analyte. Theoretical current peaks are hypothesized to be within 10% tolerance limits (mainly due to the reason that the surface area of the working electrode is itself a variable).

PAGE 10

1 CHAPTER 1 HISTORY OF CORTISOL RESEARCH Just like many other hormones, corticos teroids (a family of steroid hormones produced by the adrenal gland) were discovered through their absence from a system. In 1849 British scientist Thomas A ddison observed a fatal form of anemia, as he described it at the time, which was reflected by diseased supra-renal capsules (glands located above the kidneys). In a later description of the condition, which came to be known as Addisons disease, he emphasized the weakness of the body and the heart, anemia, irritability of the stomach, and discoloration of the skin. However, it was not until 70 years later that the distinction between th e adrenal hormones was first seen through the administration of the extracted substance from the adrenal gland that extended life. The functions of these substances were further divided into those that involve carbohydrate metabolism (glucocorticoids), and those related to elec trolyte and water balance (mineralocorticoids). However, it was f ound that with the admi nistration of these substances, certain side effects occurred whic h reflected themselves by causing patients to gain weight, develop moon face and buffalo hump (from the excess fat-tissue stored in these areas), osteoporosis of the spine, insulin resistant diabetes, etc. More than 80 years after Addisons disease was named (primarily characterized by insufficient secretion of adrenal gland), Cushings disease was describe d (primarily characterized by

PAGE 11

2 an over-active adrenal gland and excess amount s of glucocorticoids). These conditions further motivated research into glucocortico ids. Then in 1949, Hench, Kendall, Slocumb, and Polley published a paper that turned out to be a major scientific breakthrough of its time. In it, they discussed the anti-inflamma tory effects of the ad renal glucocorticoid hormone. The discovery benefited millions of people; however, dangerous side effects were still present since the effect of the new drugs did not include those of mineralocorticoids [10]. The above-mentione d scientists are considered to be the discoverers of the cortisol hormone, as we know it today. Ever since its discovery in 1949, cortisol has been thoroughly researched by scientists worldwide. Extensive research has been done to understand multiple roles cortisol has on a system [12, 19, 24, 32], but one can still say that its detection (using sensors) is in its infancy. This work is aimed at developing a cortis ol sensor for medical and defensive purposes, and the following sections focus on its si gnificance in both fields. 1.1 Introduction to Cortisol (Hydrocortisone) Cortisol is a member of glucocorticoids a family of steroid hormones produced by the adrenal gland. It has multiple roles a nd functions in the human system, including maintaining normal blood glucose levels, bl ood pressure regulation, and regulating the homeostatic balance of cardiovascular system immune system, kidneys, skeletal muscle, nervous system, and endocrine system [10, 18]. It is not possible to make a quick esti mate of how much cortisol is in an individuals bloodstream at a particular instance, partly due to the numerous factors affecting its levels. Levels of cortisol va ry throughout the day, being highest in the

PAGE 12

3 morning, and lowest right before bed-time, a nd after we fall asleep. Some controllable factors that can affect levels of cortisol are eating patterns an d levels of activity throughout a day. Although, there are many ch anges in the system that can cause disruption of levels of cortisol hormone in the blood, none has been more effective than physical and/or emotional stress. As a matter of fact, stress is so effective in disrupting levels of cortisol that the hormone r eceived a pet name stress-hormone. Side effects of abnormal levels of cortis ol range from case to case, and in some instances can even be beneficial Most of the time, however, abnormal levels of cortisol are associated with negative side effects, often resulting in a se rious condition after prolonged, untreated exposure. Addisons and Cushings disease are, unfortunately, only two examples that can result in such an inst ance. Cortisol is t hought to be a possible precursor to some other conditions such as ep ilepsy [9]. Developi ng a fast and reliable method of cortisol detection would be very bene ficial; later in this chapter we will discuss the motivation behind the work described in this thesis. 1.2 Synthesis of Cortisol Hormone Synthesis of most steroid hormones starts with cholesterol. Contrary to popular belief, cholesterol is essential for the homeos tatic balance in the system. The organism produces most of it, but a good portion comes from food intake through daily consumption. Since the adrenal gland has no visible i nnervation, it can be inferred that the Adrenocorticotropic hormone (ACTH) is a sole stimulant to production of cortisol. The adrenal gland does not store the cortisol hormone, but it does store significant amount of

PAGE 13

its precursors. Once the ACTH stimulates the gland, cholesterol concentration drops within a very short time and the level of cortisol increases [18]. Although there is no documented research to support this, one c ould surmise that the relationship between adrenal cholesterol and cortisol is somewhat inversely proportional. Figure 1.1: Synthe sis of Cortisol In order to fully understand the synthesis of cortisol, it is necessary to first understand that the entire process will undergo a few reactions involving enzymes critical for the creation of each hormone at a particular stage. First, the cholesterol will undergo a catalyzed oxidation of carbons 20 and 22. This reaction will result in pregnenolone and aldehyde, which then oxidizes into an isocapro ic acid. In this reaction a small amount of 17 -OH-pregnenolone may also be formed; however, this portion will be converted to 17 -OH-progesterone by the 17 -hydroxylase enzyme. A lthough tiny portions of pregnenolone escape from the adrenal gl and, most stays and undergoes further processing. The remaining pregnenolone is then converted to progesterone by 3 hydroxysteroid dehydrogenase and involves two microsomal enzymes. Progesterone is 4

PAGE 14

converted to 17 -OH-progesterone by the 17 -hydroxylase enzyme. Finally, this hormone goes through two more hydroxylation phases before it is finally converted to cortisol (see figur e 1.1 and 1.2) [19]. Figure 1.2: Structure of Cortisol (C21H30O5). Adapted from Citizendium. Online encyclopedia. 10/2008 Each enzyme from figure 1.1 is crucial for that particular step. Having an enzyme deficiency will not result in cr eation of cortisol, and exposure to such conditions over a prolonged period of time will result in a serious homeostatic imbalance. Figure 1.3: Cortisone Cortisol Conversion Further, inter-conversion between active cortisol to inactiv e cortisone and vice versa may be possible by introduction of another enzyme called 11 -hydroxysteroid dehydrogenase type 1 (11 -HSD-1) and 11 -hydroxysteroid dehydr ogenase type 2 (11 -HSD-2) (see figure 1.3) [19]. 5

PAGE 15

1.3 Secretion of Cortisol Hormone Figure 1.4: Secretion of Cortisol The pathway to release of cortisol is relatively simple to explain. During a stressful event, for example, the hypothalamu s starts releasing Corticotropin-releasing hormone (CRH). The Corticotropin-releasing hormone (CRH) triggers the release of Adrenocorticotropic hormone (ACTH) by the pituita ry gland. Finally, it is the release of the Adrenocorticotropic hormone (ACTH) that stimulates the releas e of glucocorticoids (mainly cortisol) by the adrenal gland. Risi ng levels of cortisol will then provide feedback to the hypothalamus and pituitary gl and and signal to stop the release of the Corticotropin-releasing hormone (CRH) a nd Adrenocorticotropic hormone (ACTH) (See figure 1.4) [18]. Transcortin, a protein serum, is responsible for transporting cortisol throughout the system [32]. As such, most cortisol hormones are bound to some kind of compound in the bloodstream. Only 4-10 % of the cortisol is thought to be free-moving in the system [12]. 6

PAGE 16

7 As mentioned earlier, levels of cortisol hormone will constantly vary throughout the day, being highest in the morning and lowest at night, right before the bed-time. In normal adults, providing they are not under stre ss, normal secretion of cortisol hormone in a whole day is 10-20mg. Morning levels vary anywhere between 140-180ng/ml, while afternoon and night levels significantly drop to anywhere between 20-40ng/ml [10]. The rate of increase/decrease of cortisol at any pa rticular instance (i.e. epileptic seizure attack) is not known. This area of medicine is st ill open for further study and research. 1.4 Need for Cortisol Sensing Detection of cortisol pote ntially has a wide variety of uses in the medical field, but these could also be applie d to any other industry. A pot ential new market for a highly reliable cortisol sensor might be customs screening or ai rport screening for passengers carrying illicit materials (assuming that such passengers stress levels would be unusually abnormal when compared to the norm). Another possible use for a highly sensitive cortisol sensor could be a fo rm of polygraph screening to improve the accuracy of the test. The possibilitie s are endless. 1.4.1 Need for Cortisol Sensor in Medical Field As mentioned, the cortisol sensor would have numer ous applications in the medical field. Cortisol is one of the major metabolic regulators in the body, and as such, abnormal levels could indicate some other, more serious condition. Reliably detecting precise levels of cortisol might also give better idea about the level of cortisone in the system and its ratio to cortisol (two hor mones that can be inter-converted using 11 -

PAGE 17

8 hydroxysteroid dehydrogenase type 1 and 2 enzy mes, as already shown). For example, children with hypoadrenalism have a lower ra nge of cortisol as opposed to cortisone, while children with adrenal cancer have been found to have increased levels of cortisol as opposed to cortisone [21]. Another important recent discovery ties co rtisol with an epileptic seizure attack [9]. Having a cortisol sensor would help further research into exactly what enables canines to predict an epileptic seizure. Sin ce cortisol is a steroid and stress hormone, the possibility whether it is this hormone that dogs sense could further be researched. In other words, can canines smell an ep ileptic seizure as well as fear? Also, the dangerous side effects of artificial cortisol used as a drug to treat certain conditions, such as, rheumatoid arthritis, cannot be forgotten. Overdosing a patient will most definitely result in Cush ings syndrome and later in Cu shings disease. Having a sensor to monitor the natural levels of cor tisol, and then feeding back the automatic drugdelivery system could potentially minimize the negative side effects that such patients can experience. Finally, another possible market for a reliab le sensor is in the field of sports medicine. Well over the last forty years, athletes have been illegally using glucocorticoids to enhance their performance. Drug screening for illegal use of artificial steroids is yet another possible us e of this kind of sensor [23]. 1.4.2 Need for Cortisol Sensor in Other Fields There are many possible industrial applica tions for a cortisol sensor, but none have the appeal like the possi ble screening tool for people im porting illicit materials into

PAGE 18

9 a country. While a terrorist threat is at an all time high and lives of innocent people are in danger all around the globe, having a sensor to monitor the stress level of passengers going through customs or airport security woul d be a great advantage. The heart rate, blood pressure, and other conditions connected to sympathetic nervous system, become abnormal in most people with any kind of confrontation [16]. Having a good cortisol monitor could segregate highly stressed travelers as opposed to ones experiencing low or no stress at all when confronted by customs agents. Another application where a cortisol sensor might pot entially find a use is the polygraph machine. Along with monitoring for respiratory rate, sweatiness of fingertips, blood pressure, and heart rate, the machine co uld monitor cortisol secretion and improve in accuracy. Also, biological research facili ties that study the effect of man-made objects on flora and fauna mainly test the cortisol levels in the bl ood of a subject. For example, one particular study focused on the effect a fluctuating hydro-power plant had on oneyear old trout [8]. The re searchers in this study mainly focused on the level of blood cortisol and the nature of its secretion. 1.5 Collection of Sample To monitor cortisol levels requires co llecting two specimens of blood (collected between 6 and 8am and 6 to 11pm) or cont inuous collection of urine over a 24-hour period. For blood, two 5ml samples are collected and mixed with anticoagulant to keep the sample in liquid state. Upon receiving a sufficient amount of sample, the sample is then transferred in a refrigerated environmen t to the laboratory where it is analyzed using a preferred method.

PAGE 19

10 1.6 Motivation The basic motivation of this thesis is to investigate the possibility of an electrochemical sensor and its se nsitivity to increasing levels of cortisol. While the most common methods for cortisol level diagnosis are still blood and urine tests, other potential methods for cortisol anal ysis are saliva and sweat tests. Since cortisol levels in the human b ody differ throughout the day, blood and urine tests require repeatable sample collections th roughout the day. Collecting all samples for analysis (especially blood) can result in a traumatic experience for the patient and it takes considerable time to implement another dow n-side to the whole process. Having a highly selective sensor integrated within an instrument that could monitor either saliva and/or sweat throughout the day could not only reduce the trauma of the sample collection but also may aid in giving an insigh t into an epileptic seizure attack and serve as an aid to customs agents for separa ting passengers who may have malicious. Clearly, there is a need for a cortisol sensor. There are also many ways to produce it, but finding an effective one is wher e the puzzle starts. In this thesis, the possibility of using the electrochemical sensi ng of the cortisol-biotin -streptavidin bond to Au nanowires will be examined.

PAGE 20

11 CHAPTER 2 CURRENT METHODS OF CORTISOL ANALYSIS The following chapter discusses a few met hods utilized to test the levels of cortisol in a sample: Radioimmunoassay (RIA), Competitive Protein Binding Assay (CPB), Fluorometric method, Reverse-phase High Performance Liquid Chromatography, and Electrochemical BioMEMS Cortisol Se nsor. The first three are stand-alone applications; they are used to find higher concentrations, and are not very sensitive to minute variations of the hormone in the te st sample. Reverse-phase High Performance Liquid Chromatography is described mainly becau se it aids in the accuracy of the utilized sensor. Finally, electrochemical sensor (d escribed in section 2.5) is a novel way proposed to monitor cortisol hormone le vels. Design, fabrication, and method of measurement of this sensor are briefly de scribed. Experimental results from the measurements are outlined and will be used in theoretical analysis in the chapters that follow. 2.1 Radioimmunoassay (RIA) Developed in 1959, the RIA method can be somewhat dangerous and caution is practiced because it utilizes a radioactive antigen. This proc edure utilizes the fact that proteins and antibodies are adsorbed by some plastics (see figure 2.1).

PAGE 21

Figure 2.1: Radioimmunoassay Serum containing the fixed amount of antibodies is added to plasti c tubes, and left to incubate for a few minutes for adsorption to take place. Next, the serum is removed and the tube is rinsed with a saline solution, leaving a coat of anti bodies on the surface. Solutions containing a fixed number of radi oactive cortisol molecules (procedure not described) and test sample are then simultaneous ly added to the tube, and left to incubate. After incubation, the solutions are removed, a nd tubes are rinsed. The radioactivity of the tube is then measured depending on the ratio of radio-labeled cortisol and regular cortisol, and the level of normal cortisol in the sample is determined. The radioactivity of 12

PAGE 22

13 the tube depends on the concentration of cortisol in the test sample since the two attach to a limited number of antibodies in a specific ratio [4, 11]. Even though RIA offers more specificity, it still can lead to er roneous results. A major problem with the RIA cortisol test is th at various steroids c ontained in the sample (along with cortisol) can still react with cortisol antibodies contained in the assay. This can cause false elevations of the cortisol le vel in the measurement. The measurement is dependent on the antibody used. To improve the results of the test, it is possible to purify the sample prior to performing an RIA analysis. The process is called chromatography, and it involves extracting specific elements out of the samp le for further processing. However, the purification procedure is not always the same and RIA kits are not always the same, either. These variables, combined with resear ch showing that reference levels of cortisol vary between individuals and genders, c ould make RIA an undesirable method for diagnosing cortisol levels in the system [4, 11]. 2.2 Competitive Protein Binding (CPB) In theory, the CPB and RIA methods are very similar. The major difference between these two methods is the amount of sa mple needed to perform an analysis, with CPB requiring considerably a smaller amount to perform the analysis. Like RIA, CPB also uses radioactive atoms in order to carry out the measurement and caution is emphasized. The CPB method consists of immobilizing a fixed number of antibodies over a surface to which the cortis ol will bind (see figure 2.2).

PAGE 23

Figure 2.2: Competitive Protein Binding A fixed amount of cortisolhormone (labeled with a radioactive atom) is added to bond with antibodies (at this point the radioact ivity will reach the peak because all of the radioactive cortisol is present in the sample). The test sample is then added to the assay (already containing bonded radioactive cortisol and antibody). Cortisol from the sample will then proportionally bind to antibodies depe nding on the ratio to radioactive cortisol. For example, if there are twelve molecule s of antibodies and the same number of radioactive cortisol, and we add six molecules of normal cortisol, only four molecules of normal cortisol and eight molecules of radio active cortisol will bond to the antibodies. Two molecules of normal cortis ol and four molecules of ra dioactive cortis ol do not bond and can be separated from the sample by centrifugation or some other method. The radioactivity measurement will then reveal th e amount of normal cortisol present since there is an interdependent rela tion between the two, just like w ith RIA. Since the cortisol 14

PAGE 24

molecules (normal and radioactive) compete for the binding sites with antibodies, the assay is called Competitive Protein Binding Assay [4, 11]. In practice, the measurements are not so simple and they depend on many other factors. For instance, the sp ecificity of the antibody used may not be very good and other molecules may bind to them. Cortisol-binding protein called transcortin is used as an immobilizing, binding agent. But other hormones such as cortisone, 11-deoxycortisol, and progesterone also have affinity towa rds transcortin. For this reason, it is recommended to perform chromatography of the test sample before assaying. Further, too small amount of radioactivity may produce erroneous results, thus invalidating the measurement [4, 11]. 2.3 Fluorometric Method Figure 2.3: Fluorometric Measurement 15

PAGE 25

16 Fluorescence is a phenomenon that describes the property of certain materials/matter to adsorb light energy a nd then reemit this light energy in longer wavelengths (less energy). Exciting light usually has to be within a certain range in order to maximize the effect of fluorescence of a certain matter. Likewise, to maximize this effect, fluorometers generally have two filters that separate different light frequencies, only allowing light with a speci fied wavelength to reach the te st sample. The first filter only allows shorter wavelength exciting light to pass through to excite the sample. The second filter serves to filter out scattered exciting light, and on ly allows filtered light to pass through to the photo cell (see figure 2.3). In order to make use of this assay for cortisol measurements, the hormone must first be extracted and marked with fluorescent reagents. The test sample is first mixed with ethylene dichloride, and shaken rigorous ly for 20 minutes, in order to enable the attachment of fluorescent reagent to the cortis ol hormone. At this point, the sample is centrifuged to separate it from ethylene dichlo ride. An aliquot of this extract is then placed in a test tube. Fluorescent reag ent (absolute ethyl alcohol mixed with concentrated sulfuric acid) is then adde d to the aliquot extract every 1 minute and vigorously shaken for 20 seconds for binding to take place. After a fixed amount of time, the sample is tested under the fluorometer utilizing a system of two filters described above. For ethyl alcohol mixed with concentr ated sulfuric acid marker, the first filter should have a cut-off at 450nm, while the se cond filter should not pass light lower than 520nm in wavelength. Although not very specific, the fluorometr ic assay method can still be used for some clinical tests to determine the level of co rtisol. However, like with the previous two

PAGE 26

17 methods, there are factors that can skew the results and thus invalidate the te st. It has been noted that large consumption of alcohol and tobacco products can greatly skew the tests. Further, fluorometric instruments can greatly differ between each other. To avoid errors with data comparison due to this reason, it is necessary to establish a reference level (baseline) of the known good sample be fore measuring the fluorometric response of blood sample or urine sample [4, 11]. 2.4 High Performance Liquid Chromatography (HPLC) The HPLC method was established in late 1960 s. This apparatus utilizes six basic components for its operation. These are the liquid mobile phase (also known as carrier liquid), sample injector, mechanical pump (w hich maintains pressure of the system), column, detector, and data recorder. Dependi ng on the properties of the material being investigated, detection can be perfor med using refractive index, conductivity, electrochemistry, absorbance or fluorescence. There are five separation techniques th at pertain to liquid chromatography: adsorption, partition, ion-exchange, affinit y, and size exclusion chromatography. Of these five, partition is the most popular one and most often us ed for cortisol analysis. Further, partition is divided into Normal -phase Liquid Chromatography and Reversephase Liquid Chromatography (with the latter bei ng used for cortisol separation) [11, 17]. 2.4.1 Reverse-phase HPLC (RP HPLC) Reverse-phase liquid chromatography uses a non-polar statio nary phase. This means that as the liquid mobile phase passes through the column carry ing the test sample,

PAGE 27

undesired elements from the sample are filtered out by the stationary phase. For cortisol, the stationary phase (or the filter) is made up of C18 chains that filter small molecules and peptides out of the sample. Finally, when the sample has been properly filtered, it is further passed through a detector of choice, a nd data graphs are plo tted (see figure 2.4). Figure 2.4: High Performance Liquid Chromatograph Of the four methods for cort isol testing, HPLC seems to offer the best specificity. It seems to be the best option for precise measurements with as little interference from other hormones as possible. Unlike the other three methods, it offers an option of a fully automated process which eliminates the human variable from anal ysis, resulting in a more precise measurement [11, 17]. 2.5 Electrochemical BioMEMS Cortisol Sensor The electrochemical cortisol sensor is desi gned and fabricated at the University of South Florida [15] (see figure 2.5). Micr oelectrodes are first fabricated on the Si wafer. 18

PAGE 28

Figure 2.5: Electrochemical BioMEMS Se nsor. Reprinted from Biosensors and Bioelectronics, 22, Kumar et al., Ultrasensi tive detection of cortisol with enzyme fragment complementation technology using functionalized nanowires, 2138-2144, Copyright (2007), with pe rmission from Elsevier The working electrode and counter elect rode are fabricated by evaporating Ti and Pt onto the wafer. Afterwards, the Si wafer is exposed to a few more lithography steps to fabricate a 60 m tall microfluidic chamber using an SU-8 Further, Au nanowires are fabricated using a Whatman anodisc template by electroplating in Techni Gold 25 ES solution for one hour. The Au nanowires are aligned between the Pt electrodes on the surface of Si wafer using dielectrophoresis. Cortis ol antibodies are then attached to Au nanowires utilizing the biotin -streptavidin link added by activ ation with thioctic acid. 19

PAGE 29

20 Finally, the electrochemical measurements are performed by mixing 5 l of 0.1M catalyst enzyme 3 -HSD with 25 l of dissolved cortisone at different concentrations. The system is left to react for 5 seconds, and electroc hemical measurements are performed (see figure 2.6). Figure 2.6: Cortisol Detection Scheme. Re printed from Biosensors and Bioelectronics, 22, Kumar et al., Ultrasensitive detecti on of cortisol with enzyme fragment complementation technology using f unctionalized nanowires, 2138-2144, Copyright (2007), with permission from Elsevier Several scans are performed before the ra nge of the sensor is determined. The increase in current peaks is found to be pr oportional with the increase in cortisone concentrations. The graph is broken down in to two separate entities for clarity (see figure 2.7a and 2.7b). The peak value of curre nt (observed to be at 40mV potential) is afterwards used to calibrate a curve where current is reflected as a function of the concentration of cortisol (see figure 2.8, obtained direc tly from the authors).

PAGE 30

Figure 2.7a: SWV at Different Concentrations of Cortisol (10-50 ). Reprinted from Biosensors and Bioelectronics, 22, Kumar et al., Ultrasensitive detection of cortisol with enzyme fr agment complementation technology using functionalized nanowires, 2138-2144, Copyright (2007), w ith permission from Elsevier 21

PAGE 31

Figure 2.7b: SWV at Different Concentrations of Cortisol (60-80 ). Reprinted from Biosensors and Bioelectronics, 22, Kumar et al., Ultrasensitive detection of cortisol with enzyme fr agment complementation technology using functionalized nanowires, 2138-2144, Copyright (2007), w ith permission from Elsevier 22

PAGE 32

23 8.01E-06 1.68E-05 2.87E-05 4.41E-05 6.45E-05 9.35E-05 1.27E-04 1.55E-04 y = 2E-08x2 + 1E-07x + 5E-06 R2 = 0.9986 0.00E+00 2.00E-05 4.00E-05 6.00E-05 8.00E-05 1.00E-04 1.20E-04 1.40E-04 1.60E-04 1.80E-04 0102030405060708090 concentration (M)current (A) Figure 2.8: Calibration Curve for Cortisol. Obtained from Arun Kumar, Shyam Aravamudhan, and Shekhar Bhansali. University of South Florida. 2008

PAGE 33

24 CHAPTER 3 THEORY Chapter 3 introduces the concepts necessary to understand theore tical analysis of the sensor. The first one, dielectrophoresis, is a method used to control the assembly of nanowires as a working electrode of the sensor. The paramete rs that can be varied to maximize the dielectrophoretic force for ultimate assembly will be discussed. The assembly itself does not seem important at fi rst; however, later it will be shown that a surface area of the working elec trode may be an important factor when estimating the total current. The second part of this chapter (introduction to electrochemical measurements) primarily discusses Square Wave Voltammetry (SWV). This helps theoretically confirm and expl ain the experimental results obtained in the lab. 3.1 Dielectrophoresis Dielectrophoresis is a scient ific method used to align pa rticles in a certain manner using a uniform and/or non-uniform electric fiel d. The subject matter does not need to be charged for the electric field to alter its position. The factors th at affect the final result of the alignment of the matter are strength of the electric field, frequency, shape of matter, size of matter, electrical properties of matter, and the medium in which dielectrophoresis

PAGE 34

is taking place. The following simplified e quation is commonly used to estimate the dielectrophoretic force: ))()(( tEtmFDEP 3.1 where E is the electric field, m is the dipole moment, and is the del vector (actual equation includes higher order terms as well as Maxwell stress tensor ) [5, 25]. As the non-uniform electric field passes fr om one electrode to the other, it will create a torque on the particle thus causing it to move. Cylindrically shaped object should therefore align in the same direction of the electric field. The dipole moment induced on the particle can be expr essed using the following equation: )( )( tEKVtmpm 3.2 where m is the absolute permittivity of the medium, Vp is the volume of the particle (for cylindrical object ), E is the electric field, and K is the complex polarization factor. Since the nanowires used for this project are cylindrical in shape, polarizability along the length is more pronounced than either radial orientation. Therefore, complex polarization factor is expressed: lrVp 2 m mpK 3.3 where p and m are complex permittivity of particle and medium respectively. The word complex indicates the presence of both real and imaginary factors, with the latter containing both conductivity and angular frequency Therefore, the new equation for the complex permittivity is: i 3.4 25

PAGE 35

Due to the fact that the E-field and nanowire polarization are in-phase, the dielectrophoretic force can be expressed by looking at the real part of equation 3.1, and can therefore be re-written: } Re{ 2 1 Em FDEP 3.5 Combining equations 3.2 and 3.5 yields a new expression for dielectr ophoretic force, and is written: 2 2||}Re{ 2 1rms m DEPEKlrF 3.6 Root mean square value of the electric field Erms as well as complex polarization factor K can further be broken down as following: dx dV dx dV Erms x rms rms 2|| 3.7 2 2 2 2 2 2}Re{m m mpm mpmK 3.8 Finally, equation 3.6 can be combined with equations 3.7 and 3.8 to yield the final equation of the dielectrophoretic force: dx dV dx dV lr Frms x rms m m mpm mpm m DEP2 2 2 2 2 2 22 1 3.9 Upon the alignment of each nanowire betw een electrodes, the magnitude of the electric field between electrodes is reduced Depending on the application, this can create a serious issue because a majority of nanowires can end up being scattered, randomly melting and welding to the electrode If nanowires are goi ng to be used as a 26

PAGE 36

27 working electrode for electrochemical measur ements (like in this project), it becomes more complicated to estimate the true surface area of the working electrode. The exact value of the surface area in this case is necessary for final data analysis and estimation. At frequencies below 1kHz nanowires may have a tendency to evaporate, further complicating the above problem. Evaporation do es not seem to be the problem at higher frequencies. At 10kHz or highe r, the strength of the dielect rophoretic force seems to be weaker [25]. On the other hand, the same strength of the dielectrophoretic force is exponentially proportional to the peak voltage applied across the electrodes. In short, to manipulate/maximize the strength of the fo rce, voltage and frequency are the two variables most commonly altered to tweak the force as desired. The user may not have much choice with the permittivity due to the fact that different liquids/buffers may force an unpredicted reaction in a sensor itself, thus skewing results. Further, it is thought that di electrophoresis is applicable to structures between 1 and 1000 m. Gravity is suspected to be a major interference for structures larger than 1000 m. For small structures under 1 m in size, Brownian motion (random movement of particles suspended in liquid) overwhelms the DEP force [5, 25]. 3.2 Electrochemistry By definition, electrochemistry is the st udy of the interchange of chemical(s) and electric current. There are di fferent types of electrochemical measurements, but for this project Square Wave Voltammetry (SWV) was used, and therefore will be discussed. This method utilizes a system of three elec trodes to provide an excitation voltage, and

PAGE 37

28 measure the current from the electrochemical cell: working electrode (WE), reference electrode (RE), and counter electr ode (CE) [2, 3, 13, 20, 30, 35]. 3.2.1 Electrode Function Breakdown Working electrode (WE) is where all the el ectrochemical changes occur. In other words, it is a point where the apparatus pr ovides electrical exc itation and performs measurements with respect to the referen ce electrode and the counter electrode. Reference electrode (RE) is the electrode of a fixed, consta nt potential with respect to the electrolyte and is kept at equi librium. The potential of the RE serves as a reference when measuring the potential of the WE. Because of the need to keep the RE at equilibrium, it is not recommended to pass charge through it. Most reference electrodes contain the chemical element Cl and current flow will cause the concentration to fluctuate. Finally, the count er electrode (CE) is included to facilitate the passage of current at the WE [2, 3, 13, 20, 30, 35]. 3.2.2 Fundamentals of El ectrochemistry To reflect on some fundamentals and terminology in electrochemistry, it is imperative to start with the term of equilibrium. In electrochemistry, equilibrium means zero current during potentiometric measurement. In other words, it is common to say that a system in equilibrium has an absent pow er source driving the reaction (effectively equivalent to an open circuit electronic wi se). However, even though the current does not exist between the electrodes, the electrodes still have some potential. This potential is equilibrium electrode potential and very impor tant when determini ng the overpotential of

PAGE 38

an electrode, which is the difference between new electrode potentia l and its equilibrium equivalent while forcing charge. = Eelectrode Eequilibrium 3.10 When forcing the charge on the electrode, the equilibrium of the system will be disturbed, thus resulting in va rying the electrode potential. At this point, the flow of charge is accompanied with electron uptake (r eduction) or electron lo ss (oxidation) of the electroactive analyte in the electrolyte by the working electrode, thus changing the concentration and activities of the analyte in the electrolyte. Activity and concentration are closely related. Activity is defined as concentration perceived by electrode, and is expressed by: a = c 3.11 (where a is activity, c is concentration, and is activity coefficient). However, at low concentrations, this is considered equal to unity, and the new form of equation is: a = c 3.12 Altering this ratio of activities at the elec trode/solution interface and conversion of the material from reduced to oxidized form and back will in turn result in production/consumption of charge. This phenomenon is somewhat described by a form of the Nernst equation: ) (@ Re ) (@ 0lnsurface electrode d surface electrode Ox electrode workinga a nF RT E E 3.13 (E0 is standard electrode potential, R is universal gas constant (8.314510 J K-1 mol-1 ), T is temperature (in Kelvin), n is number of electrons involved in re action, F is Faraday constant (9.64853094 C mol-1 ), and aOx and aRed are chemical act ivities for the 29

PAGE 39

oxidized and reduced species respectively). The true Nernst Equati on is only valid when the system is in equilibrium, and is: ) (@ Re ) (@ 0lnsurface electrode d surface electrode Ox mequilibriua a nF RT E E 3.13a Under this condition net current density is equal to either forward or reverse current of the reaction: Inet = Iforward = Ireverse 3.14 Forward current (also known as oxidative or an odic current) is obtained when a species of interest loses electrons. Likewise, revers e current (also known as reductive or cathodic current) is obtained when a species of intere st gains electrons. It can be mathematically shown that the Nernst equation is therefore a derivative of the above equation (and vice versa). However, equation 3.13 is a form of the Nernst equation because the system is not in equilibrium. When the system is not in equilibrium, forward and reverse currents are not the same, and net current is the addition of the two. I net = I forward I reverse 3.15 In practice, the currents flow opposite to one another, so to account for that, a negative sign is used in the formula. Following fo rmulas are adopted for forward and reverse currents: RT EEnF Ox forwardmequilibriueaII) ()1( 0 3.16 RT EEnF d reversemequilibriueaII) ( Re0 3.17 30

PAGE 40

( is a transfer coefficient, and in mo st cases has a value of 0.5). The term I0 is also known as exchange current. It is calculated by multiplying the exch ange current density i0 [A/m2] with total surface area of a working electrode A[m2]. This is the rate constant of electron transfer while the system as a whole is in equilibrium. Combining the equations 3.15, 3.16, and 3.17, the final equation for estimating the net current is obtained: RT EEnF d RT EEnF Ox netmequilibriu mequilibriuea eaII) ( Re ) ()1( 0 3.18 The above equation is the Butler-Volmer equation and it may be used to estimate the total current at any particular instance. Finally, to conclude this part of the chapter, it is important to emphasize that the above theory is not always useful when predicting forward and reverse currents. The graphica l representation of the above currents has a form as in figure 3.1 below, and currents do not peak [2, 3, 13, 20, 30, 35]. The method of measurement used fo r this experiment, Square Wave Voltammetry, significantly differs from figure 3.1 in output. Square Wave Voltammetry is characterized by forward and reverse current graphs that reach a peak at one point, and converge to a fixed value on eith er end. This method is briefly described in the following section. 31

PAGE 41

Figure 3.1: Forward and Reverse Current in El-Chem System 3.2.3 Square Wave Voltammetry (SWV) Electrochemical voltammetry is a method where voltage is varied over a time period, while at the same instance, current is measured in a reaction. The square wave excitation signal applied to the working electrode (WE) is sh own in figure 3.2. Voltage pulses are applied at a user directed rate (Vs-1) and period (ms). Ep is usually 50mV, Es is commonly about 5mV, and tp is usually anywhere between 20-40ms. 32

PAGE 42

Figure 3.2: Time-Potential Profile for SWV. Electrochemical Methods Fundamentals and Applications. Allen J. Bard & Larry R. Faulkner. David Harris, Elizabeth Swain & Eugene Aiello. Copyright (2001) John Wiley & Sons, Inc. Reprinted with permission of John Wiley & Sons, Inc. While the voltage on the working elect rode (WE) is applied, the current measurements are performed by the apparatus first in the forward, then in the reverse direction. The net current is the addition (or difference) be tween these two currents, and the net current is the one refl ected on the monitor of the inst rument. An example of one such reading is illustrated in figure 3.3. 33

PAGE 43

Figure 3.3: Square Wave Voltammogram. Reprinted with permission from (Anal. Chem., Vol: 49, No.: 13, 1904-1908). Copy right (1977) American Chemical Society. A few equations are utilized to describe the SWV waveform. First, depending on the selected parameters, potential waveform is applied to the working electrode, and is described as: )1()1(1 2 1 mforE E m IntEEp m s im 3.19 (where m denotes a series of half cycles from the first forward pulse (m = 1), and Int[(m=1)/2] denotes truncation of the ratio to the highest in teger). The balance of the concentrations of oxidative a nd reductive species at the surf ace of the working electrode (also known as Nernstian balance) is then described as: )(exp ),0( ),0(2/1EE RT nF tC tCm R O m 3.20 34

PAGE 44

(where n is the number of electrons involved). The above equation provides an input for calculating another parameter required for estimation of total current. 1)0( 10 Qi Qi i i 3.21 50 R OD D 3.22 (where DO is the diffusion coefficien t of the O-species, and DR is the diffusion coefficient of the R-species). In Square Wave Voltammetry (SWV), cu rrent is expressed as a dimensionless unit Since in every cycle, there is one forw ard current sample and one reverse current sample obtained, the total forward and reverse current will be the addition of all the previous (preceding) half cycles, and the present one. Therefore, the dimensionless current equation is expressed as: m i ii mim QQ1 50 1)1( 3.23 where odd values of m correspond to forward current, while even values correspond to reverse current samples. Finally, the dimens ionless current difference (analogous to net current from equation 3.18) is written as: 1 mmm 3.24 where m covers only odd values, and odd m is taken first. The dimensionless current Voltammogram is illustrated in figure 3.4 below (not to be confused with figure 3.3 that 35

PAGE 45

represents the response of a particular sensor measured in amperes. E1/2 is very close to E0, and is expressed by equation O RD D nF RT EE ln'0 2/1). Figure 3.4: Dimensionless Curr ent. Reprinted with permission from (Anal. Chem., Vol: 53, No.: 4, 695-701). Copyright ( 1981) American Chemical Society. Determining the peak reflects the peak dimensionless current p from figure 3.4, which enables estimation of the peak of the current according to equation 3.25: p p OO pt CnFAD i 50 50 *50 3.25 36

PAGE 46

37 (where CO is the bulk concentration of O-species). An important fact to note is that difference peak current p happens near the half wave potential E1/2, which has to be taken into account when graphing the fina l current waveform ve rsus potential. Current equals 0 when the forced potential is much larger than E1/2. Since the forced potential is far away from the electr olysis point, the forwar d and reverse current do not flow. When the forced potential starts approaching E1/2, electrolysis starts to occur and forward and reverse currents start to flow. Finally, when the forced potential becomes much smaller than E1/2, electrolysis happens at the diffusion controlled rate independent of the applied pot ential and both forward and reverse current become similar [1-3, 13, 20, 22, 26-28, 30, 33-35]. The peak of dimensionless current is also dependent on square wave parameters. From table 3.1 it is evident that peak is dependent and increases with Ep and Es. The value of p can then be used to estimate othe r unknown parameters from equation 3.25. Table 3.1: Dimensionless Peak Curre nt vs. SWV Operating Parameters. Electrochemical Methods F undamentals and Applications. Allen J. Bard & Larry R. Faulkner. David Harris, Elizabeth Swain & Eugene Aiello. Copyright (2001) John Wiley & Sons, Inc. Reprinted with permission of John Wiley & Sons, Inc. n ES /mV n Ep /mV 1 5 10 20 10 0.2376 0.2549 0.2726 0.2998 20 0.4531 0.4686 0.4845 0.5077 50 0.9098 0.9186 0.9281 0.9432 100 1.1619 1.1643 1.1675 1.1745

PAGE 47

Table 3.1 is helpful in estimating a single p eak point of the current but equation 3.23 has to be incorporated in order to see how the current really behaves over a larger range of voltage. So, the current for the mth half-cycle is: m i ii p OO mim QQ t CnFAD i1 50 1 50 50 *50)1( 3.26 Again, an important thing to note about this equation is that odd values of m correspond to forward current, and even correspond to reve rse current. Net current is the difference between the forward and reverse samples shown in: 1 mmmiii 3.27 (where m covers only odd values, and odd m is taken first). Further, the diffusion coefficient is described in terms of the radius of the moving partic les and molecules in the solvent by equation: rN TR D6 3.28 (where v is the viscosity of the solvent, N is Avogadros number, and r represents the radius of a molecule/particle) [1]. The radius is then described by Stokes radius (which is the equivalent of the radius of a hard sphere that diffuses at the same rate as an actual molecule) in equation 3.29: 34 3V r 3.29 (where V is a cell volume of a particle/molecule). Square Wave Voltammetry (SWV) features a number of advantages when compared to other methods. From figure 3.3 it can be seen that since the peak value is 38

PAGE 48

39 the addition of two currents and therefore larger, the value of the current is more easily estimated (hence, increasing the accuracy). Further, Square Wave Voltammetry (SWV) minimizes the capacitive contributions to the overall current, resulting in dramatic increase of scan rate (capacitive current is always present as long as there is AC excitation on the electrode. Keeping the ex citation signal constant greatly reduces the capacitive effect. For this reason, current m easurements are performed right before the potential changes in order to minimize the capacitive effect). Finally, the height of the peak of the net current is directly proportiona l to the concentration of analyte, making the interpretation of the results easier [1-3, 13, 20, 22, 26-28, 30, 33-35]. 3.2.4 Construction of Electrochemical Cell The next step is the construction of the cell where the measurements will take place. This is a simple yet complicated pro cess, due to so many rules to follow, and the greatest challenge is trying to satisfy all of the requirements. Only a few important rules are mentioned and implemented for the sake of this experiment, and will later be discussed. The very first rule states that the tip of the RE should be positioned as close as possible to the surface area of the WE. This minimizes the ohmic drop and polarization of the WE should be uniform. Second, the CE should be positioned downstream from the WE, and it is recommended that the CE be excluded from the solution bulk or significantly away from it, but still in contact with main body of the cell. The reason for this is because the CE reaction with the el ectrolyte may produce negative effects on the

PAGE 49

measurement. Finally, one last rule of utmost importance is to thermally isolate the cell because the reactions are temperatur e dependent [2, 3, 13, 20, 30, 35]. Failure to satisfy these few rules can greatly skew the results. But again, these are just a few requirements to keep in mind. The design could get expensive and laborintensive if all requirements were to be sa tisfied, especially if the cell had to be dimensionally small. One possible scheme for constructing the electrochemical cell is illustrated in figure 3.5 below. Figure 3.5: Graphical Repres entation of El-Chem Cell Upon completing the cell, the electrodes are connected to the apparatus and electrochemical measurements of choice are pe rformed. Figure 3.6 below depicts a block diagram of such an apparatus [2, 3, 13, 20, 30, 35]. 40

PAGE 50

Figure 3.6: Block Diagram of A pparatus for El-Chem Analysis. Adapted from class notes. A. Sagues. El ectrochemical Diagnostic Techniques. 2006 41

PAGE 51

CHAPTER 4 RESULTS This chapter attempts to give an explana tion to the experiment al results shown in section 2.6. It is necessary to refer back to chapter 3 equations (3.19 3.29) in order to calculate the peak current for a given concen tration of cortisone/co rtisol using the SWV measurement method. First, looking at e quation 3.26, a few unknown quantities are left to be determined. m i ii p OO mim QQ t CnFAD i1 50 1 50 50 *50)1( 3.26 The oxidative and reductive species diffu sion coefficient, dimensionless current (summation term), and the surface area of th e working electrode are for now unknown. The following section explains the approach ta ken to estimate the diffusion coefficient of the oxidative and reductive specie s cortisone and cortisol. 4.1 Calculation of Diffusion Coefficients Diffusion coefficients were estimated us ing equation 3.28. Since the reaction takes place in phosphate buffer saline (PBS), which has properties similar to water, viscosity of water 0.001 2m Ns is therefore used for this equation. 42

PAGE 52

rN TR D6 3.28 The value of the radius for equation 3.28 is obtained from equation 3.29. The unit cell volume for equation 3.29 is obtained from the Cambridge Structural Database [6, 29] for both cortisone and cortisol. 34 3V r 3.29 The software patch for cortisone molecule is downloaded from the Cambridge Structural Database. This software patch is uploaded in to Mercury software (also obtained from the Cambridge Structural Database), which enable s viewing of all the st ructural and physical information of the cortisone molecule (see figures 4.1a, 4.1b, and 4.1c below). An important thing to stress is that cell volume is expressed as a volume of a rectangular prism that a single molecule occupies. Since molecules are not spherical in shape, the radius of the molecule is then the radius of a perfect sphere that diffuses at the same rate as the actual molecule, also known as Stokes radius. Therefore, the volume of a rectangular prism is equated to the volume of a perfect sphere. The rect angular prism volume (see figure 4.1c) is incorporated into equation 3.29, upon which the final value for th e Stokes radius of the cortisone molecule is obtained. ][10*60.7 *4 22.1838*310 3m r 43

PAGE 53

Figure 4.1a: Views of Cortisone Molecule. Mercury Software from Cambridge Structural Database. [http://www.ccdc.cam.ac.uk/support/product_references/] 44

PAGE 54

Figure 4.1b: Rectangular Prism Unit Ce ll. Mercury Software from Cambridge Structural Database. [http://www.ccdc.cam.ac.uk/support/product_references/] Figure 4.1c: Structural Info of One Mo lecule of Cortisone. Mercury Software from Cambridge Structural Database. [http://www.ccdc.cam.ac.uk/support/product_references/] 45

PAGE 55

The radius of a single cortisone molecule is estimated to be approximately 7.6 Incorporating this value into equation 3.28, the diffusion coefficient for cortisone (oxidative species) is calculated to be: s m m m sN mol K molK J rN TR D2 10 10 2 231087.2 106.7 001.0 1 10022.66 15.298 314.8 6 The dimensions of a unit cell for cortisol (a = 6.435 b = 15.626 c = 18.912 ) yield a cell volume of 1901.67 3 [6]. Using the equivalent calculation procedure described above yields the radius of a singl e cortisol molecule (re ductive species) to be approximately 7.69 The diffusion coefficient of co rtisol is calculated to be 2.84*10-10 m2/s. Of course, the calculations above represen t rough estimates of what the true value of both diffusion coefficients are. The molecu les are too small to precisely calculate their volume. Further, the value of the di ffusion coefficient greatly depends on the environment in which the measurements are performed. The viscosity of PBS, for example, is not available from a manufactur ers specification sheet The viscosity of water is used instead, even though the two me diums may slightly differ in this aspect. Combination of all of these factors may introduce error into further calculation of total current. For this reason, it is hypothesized that a deviation of 10% will account for the rough estimate of the above-calculated values. So, for cortisone, the diffusion coefficient range is 2.58*10-10 3.16*10-10 m2/s. For cortisol, the same range is 2.56*10-10 3.12*1010 m2/s. 46

PAGE 56

Finally, estimating the diffusion coefficients eliminates two variables from the list of unknowns. The following section explains the calculation of dimensionless current (summation term of equation 3.26). 4.2 Calculation of Peak Dimensionless Current The dimensionless current is calculated using the system of equations 3.19 through 3.24 from chapter 3. First, Square Wave Voltammetry (SWV) excitation voltage is calculated using equation 3.19 over so many half cycles. )1()1(1 2 1 mforE E m IntEEp m s im 3.19 All the parameters from this equation are known (Ei = 0.08V, Es = 0.004V, Ep = 0.02V, f = 8Hz) [15], so modeling the waveform presents a simple task (see figure 4.2). Next, the value of Em (from equation 3.19) is used in conjunction with E1/2 (which is calculated to be approximately 0.039V, since E0 was found to be 0.04V [14]) to determine the balance between concentrations of oxidative and reduc tive species at the working electrode by equation 3.20. )(exp ),0( ),0(2/1EE RT nF tC tCm R O m 3.20 This balance changes every half cycle a nd is used further down in the estimation of dimensionless peak current. Equations 3. 20 and 3.22 are then combined into equation 3.21 to find the next term in a sequence, Q 50 R OD D 3.22 47

PAGE 57

48 1)0( 10 Qi Qi i i 3.21 Finally, equation 3.23 yields the dimensionless current at any particular instance, while equation 3.24 represents the total net cu rrent (or the difference between the forward and reverse dimensionless currents) analogous to equation 3.15. The peak dimensionless current is obtaine d from equation 3.24, and is the highest point in the curve p (see figure 4.3). It is determined to equal 0.4247. m i ii mim QQ1 50 1)1( 3.23 1 mmm 3.24 Using the frequency of 8Hz for the square wave excitation [15], it is calculated that the period of the square wave is 0.125s (an inverse of frequency). It then follows that the half cycle of the squa re wave lasts 0.0625s. Using th e logic that current is probed every 0.0625s (first in forward then in reverse direction), it is hypothesized that the whole measurement can be performed in the amount of time it takes the square wave excitation to reach a certain level lower than that of E1/2. In other words, forward and reverse currents become similar and electrolysis starts to occur at the diffusion controlled rate. In this case, it is hypothesized th at the whole measurement could be performed in 5 seconds (see figure 4.4). It is assumed that estima ting the diffusion coefficient to 10% does not affect the results since the e rror is canceled by the equation 3.22. In other words, the term will not change if the diffusion coefficient is 10% larger (or smaller), since the ratio remains the same.

PAGE 58

Square Wave Voltage Waveform0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 00.20.40.60.811.21.41.61.82 time [s]E [V] SW Excitation Figure 4.2: Simulation of SW Excitation on Working Electrode 49

PAGE 59

Dimensionless Current, PSI-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 -0.45 -0.4 -0.35 -0.3-0.25-0.2 -0.15 -0.1 -0.05 0 0.05 0.1n(E-E1/2) [V]PSI PSI Forward PSI Reverse PSI Total Figure 4.3: Simulation of Dimensionless Current 50

PAGE 60

51 PSI vs. t-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 00.511.522.533.544.555.56time [s]PSI PSI vs. t Figure 4.4: Simulation of Di mensionless Current Over Time

PAGE 61

4.3 Calculation of Surface Area of Working Electrode After estimating the dimensionless curre nt, the surface area of the working electrode remains the last unknow n in equation 3.26. However, in order to estimate the surface area of the electr ode, it is first necessary to estimate the peak currents obtained in the lab experimentally (from figures 2.7a a nd 2.7b). These peak currents are necessary because the area is estimated from equation 4.1, which is a modified form of equation 3.25. *50 50 50OOp ppCnFD it A 4.1 Dimensionless peak current value dete rmined in the previous section ( p = 0.4247) is used in the above equation. The value extrapolated from table 3.1 ( p = 0.4647) can also be used; however, square wave voltammogram c ould not be obtained since this value strictly represents a single value at the peak. In other words, it would not yield a continuous line which illu strates the potential at which peaks occur. But to verify the information, both methods will be perf ormed and compared for consistency. Finally, looking at the figu res 2.7a and 2.7b, eight different current peaks are estimated for each concentration of hormone ad ded to the cell. An important thing to note is that the concentrati ons must be converted to appropriate units. A good check would be to write out the un it equation using equation 4.1, and to determine whether the final result is m2, such as performed in section 4.1. ][ ** ** *2 3 2 3 2 *50 50 50m m mol s m mol C s C s m mol s m mol C As CnFD itOOp pp 52

PAGE 62

53 Table 4.1: Calculation of Surface Area Using Dimensionless Peak Current From Figure 4.3 Concentration [mol/m3] Peak Current [A] (fig. 2.7a and 2.7b) Diffusion Coefficient [m2/s] Net PSI (fig. 4.3) Surface Area [m2] 0.01 8.01E-06 2.87E-10 0.4247 5.11E-04 0.02 1.68E-05 2.87E-10 0.4247 5.36E-04 0.03 2.87E-05 2.87E-10 0.4247 6.10E-04 0.04 4.41E-05 2.87E-10 0.4247 7.04E-04 0.05 6.45E-05 2.87E-10 0.4247 8.23E-04 0.06 9.35E-05 2.87E-10 0.4247 9.94E-04 0.07 1.27E-05 2.87E-10 0.4247 1.16E-04 0.08 1.55E-05 2.87E-10 0.4247 1.24E-04 After estimating the areas by using the p eak value from the dimensionless graph, a second method of estimating the area is pe rformed. The dimensionless peak current value ( p = 0.4647) from table 3.1 is used ba sed on the square wave excitation parameters specified. This is a faster, si mpler method since it does not involve software analysis. Table 4.2: Calculation of Surface Area Using Dimensionless Peak Current From Table 3.1 Concentration [mol/m3] Peak Current [A] (fig. 2.7a and 2.7b) Diffusion Coefficient [m2/s] Net PSI (Table 3.1) Surface Area [m2] 0.01 8.01E-06 2.87E-10 0.4647 4.67E-04 0.02 1.68E-05 2.87E-10 0.4647 4.90E-04 0.03 2.87E-05 2.87E-10 0.4647 5.58E-04 0.04 4.41E-05 2.87E-10 0.4647 6.43E-04 0.05 6.45E-05 2.87E-10 0.4647 7.52E-04 0.06 9.35E-05 2.87E-10 0.4647 9.09E-04 0.07 1.27E-05 2.87E-10 0.4647 1.05E-04 0.08 1.55E-05 2.87E-10 0.4647 1.13E-04

PAGE 63

It is evident from the above calculation s that the surface areas using different values of the dimensionless peak current vary roughly up to 10%. During the final current estimate, it is important to keep th e above calculations separate. They may be describing the same entities, but the methods utili zed to calculate that entity are different. Also, looking at the above values for th e surface area of the nanowires, it might seem surprising that such small particles co uld create such a la rge surface area of the working electrode. In order to understand this better, it is imperative to refer back and analyze the original method utilized to fabricate the Au nanowires [15]. Understanding the amount of surface area that the entire set of nanowires from the Whatman disc can form determines whether the above surface area calculations make sense. The nanowires were fabricated by electr oplating in Techni Gold 25 ES solution for 1 hour [15]. Graphical representation of the Whatman disc clarifies physical dimensions of the Au nanowires (see figure 4.5). Figure 4.5: Graphical Repres entation of Whatman Disc Electroplating for 1 hour create s a nanowire that is roughly 20 m in length. However, after releasing the nanowires in KOH and then methanol, it is hypothesized 54

PAGE 64

55 that they break in half since they are thought to be extremely brittle. Also, for the sake of calculation, it is assumed that the nanowires are perfectly cy lindrical. Taking this into account (perfect cylinder; 10 m in length; 200nm base diamet er) the surface area of a single nanowire is ca lculated to be 6.35*10-12 m2. The Whatman disc comes in two sizes, 13mm and 26mm in diameter. Also, the pore density ranges between 1013-1014 holes per meter squared [31, 36]. Depending on which size of the disc is selected, the total number of holes per disc is 2.65*109 (for the 13mm diameter disc) and 1.06*1010 (for 26mm diameter disc ). Multiplying these numbers by the surface area of a single na nowire yields a tota l surface area of 1.68*10-2 m2 (168cm2) for a 13mm diameter disc and 6.74*10-2 m2 (674cm2) for a 26mm diameter disc. Finally, the surface areas from all the nanowires in the disc are compared with the surface areas of the working electrode thought to be present in the cell. It is obvious that the total surface area of the nanowires thought to be in the cell makes only a small fraction of the total surface area produced by all the nanowires that are in either size Whatman disc. Therefore the nanowires do ind eed have the potential for creating a total surface area comparable to that determined independently from the electrochemical calculations shown earlier in tables 4.1 and 4.2. 4.4 Calculation of Total Peak Current for an Electrochemical Response After all the unknown parameters have been estimated, the tota l peak current of the electrochemical system can finally be calculated and graphed using equations 3.25 and 3.26

PAGE 65

p p OO pt CnFAD i 50 50 *50 3.25 m i ii p OO mim QQ t CnFAD i1 50 1 50 50 *50)1( 3.26 Equation 3.25 utilizes the sec ond part of area calculations (table 4.2) illustrated in the previous section. This equation yields single point current p eaks at a particular concentration. These values are compared against figure 2.8 from experimental analysis. Equation 3.26 illustrates the entire square wave voltammogram and highlights the half potential of the cell (since that is the point where peak occurs). The first part of the equation ( 50 50 *50 p OOt CnFAD) represents a constant that is combined with the graph of the dimensionless net current ( m i iiim QQ1 50 1)1( ) from figure 4.3. The areas from table 4.1 (section 4.3) are incorporated in to this equation. The values are also compared against the ones obtained in figure 2.8 fr om experimental analysis. Further, like in previous sections, it is recommende d to write out the equation with units to make sure e rrors do not carry over (the final result is in Ampere). In order to do this, all the parameters and their respective values and units are listed below. 1. n Number of electrons involved in react ion = 1 (when cortisone is converted to cortisol it releases 1 electron) 2. F Faraday constant 96,500 C mol-1 56 3. A surface area of working electrode see section 4.3 (steps 1-8)

PAGE 66

57 4. DO Diffusion coefficient for O-species 2.87*10-10 m2/s 5. CO Bulk concentration of O-species = 10-80 (figures 2.7a and 2.7b) or 0.010.08 mol/m3 6. tp half-cycle of the square wave = 0.0625 seconds 7. p estimated from figure 4.3 0.4247 A s C ss C s m mol s m m mol C t CnFADp OO ***3 2 2 50 50 *50 After performing a unit check, theoretical current peaks are finally graphed and calculated using Microsoft Excel. The maximum value of each peak is then determined and graphed against each concentration. Th e theoretical current peaks are compared against experimental current peaks from chapter 2 (see figures 4.6 and 4.7). Error bars are set to 10% to account for hypothesized pe rcent error when estimating the diffusion coefficient. Equation 3.25 is used in order to see whet her the results would match if the value of dimensionless peak current from table 3.1 we re used. So the only two parameters that changed from the list above are the surface area calculations (step 3) and dimensionless peak current (step 7). In this instan ce, surface area equation 4.1 would utilize calculations from table 4.2 (s ection 4.3), while dimensionless peak current would equal 0.4647. Likewise, error bars are set to 10%. The results are discussed in chapter 5.

PAGE 67

i vs. E0.00E+00 1.50E-05 3.00E-05 4.50E-05 6.00E-05 7.50E-05 9.00E-05 1.05E-04 1.20E-04 1.35E-04 1.50E-04 1.65E-04 1.80E-04-0.05 -0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07E [V]i [A] 10uM 20uM 30uM 40uM 50uM 60uM 70uM 80uM Figure 4.6: Theoretically Calculated SW Voltammogram fo r Different Concentrations of Cortisone/Cortisol 58

PAGE 68

Theoretical vs. Experimental currenty = 2E-08x2 + 1E-07x + 5E-06 R2 = 0.9984 y = 2E-08x2 + 1E-07x + 5E-06 R2 = 0.99840.00E+00 2.00E-05 4.00E-05 6.00E-05 8.00E-05 1.00E-04 1.20E-04 1.40E-04 1.60E-04 1.80E-04 0102030405060708090Concentration [uM]Current [A] I_theoretical [A] I_experimental [A] Poly. (I_theoretical [A]) Poly. (I_experimental [A]) 59 Figure 4.7: Theoretically Calculated Current Based on Different Concentrations of Cortisone/Cortisol Using Dimensionless Peak Current From Figure 4.3

PAGE 69

60 Figure 4.8: Theoretically Calculated Current Based on Different Concentrations of Cortisone/Cortisol Using Dimensionless Peak Current From Table 3.1 Peak Current vs. Concentrationy = 2E-08x2 + 1E-07x + 5E-06 R2 = 0.9984 y = 2E-08x2 + 1E-07x + 5E-06 R2 = 0.9984 0.00E+00 2.00E-05 4.00E-05 6.00E-05 8.00E-05 1.00E-04 1.20E-04 1.40E-04 1.60E-04 1.80E-04 02 04 06 08 01 0 0Concentration [uM]Current [A] I vs. C theoretical I vs. C experimental Poly. (I vs. C experimental) Poly. (I vs. C theoretical)

PAGE 70

61 CHAPTER 5 DISCUSSION AND FUTURE WORK In this research, a theoretical model has b een developed for a sensor to detect the cortisol hormone. An in-depth look into Square Wave Voltammetry resulted in the information needed to optimize the design of the sensor itself. The following conclusions are primarily based on numerical analysis of the data. At the end of this chapter, suggestions to optimize the design of the sensor are given. 5.1 Discussion The greatest challenge of this analysis was to estimate the surface area of the working electrode to aid in the approximation of the current during the electrochemical measurement. The theoretically estimated values from figures 4.7 and 4.8 are mirror images of the experimental values shown in figure 2.8 (unsurprisingly, because figures 4.7 and 4.8 are after all revers e-engineered by using the va lues from figure 2.8). However, one fundamental question arises from the entire analysis that deals with the accuracy and reliability of the surface area of the working electrode (estimated in section 4.3) that is used to estimate curves in fi gures 4.7 and 4.8. Are the values calculated in table 4.1 (section 4.3) more precise than values calculated in table 4.2 (section 4.3), even though they fall within 10% tolerance of one another? To better answer this question, it

PAGE 71

62 is necessary to refer to the original process utilized to align the na nowires as the working electrode of the sensor. Alignment of Au nanowires was completed utilizing the dielectrophoresis technique discussed in chapter 3. First, th e resistance was checked between electrodes to ensure complete isolation between the pos itive and negative micr oelectrode using a simple Digital multi-meter. 200 l of nanowire-methanol so lution was then dispersed over the microelectrodes using a syringe. Agilent 33250 waveform generator was connected to microelectrodes and programmed to output 10 VRMS at 1kHz. This signal was applied over the period of 15 seconds. Af ter this initial peri od, the excess nanowire solution was removed, and the working area was washed with pure methanol. Potential and frequency were then varied gradually over time up to 50VRMS and 10kHz for an additional 30 seconds. The assembly was observed by monitoring the voltage drop over the series 1k resistor and was later verified with a scanning electron microscope (SEM). The set-up was left at room temperature over night to dry off. Finally, resistance was checked again between microelectrodes to ensure the presence of nanowires which indicated alignment. Resistance should decrease during the sec ond check, suggesting a short path between the electrodes. The problem with the above method is that the amount of nanowires dispersed with 200 l of methanol cannot be controlled. If the process is required to be repeated multiple times, the sole number of nanowires becomes unpredictable and starts to vary, making the surface area parameter from equation 3.26 a variable. Further manipulation of the working area of the se nsor (like washing the area with piranha solution, which is done to activate the Au nanowires and attach cortisol antibodies) may remove the loose

PAGE 72

63 nanowires off the sensor completely. This is a factor that only a dds variability to the design. But even if the above-stated problems we re not an issue, the dielectrophoretic alignment of the nanowires presents another challenge. The force of the electric field during the alignment will not align the nanowir es to perfectly bridge the gap between electrodes. The final outcome is more similar to a random, unc ontrolled alignment as shown in figure 5.1. Some nanowires may ge t welded to the microelectrodes under the influence of electric field; others may just stay in contact with the microelectrodes without forming a firm joint, while some nanowires may not be in contact with the microelectrodes at all. The latter presents a problem because the excitation voltage never reaches this portion of the working el ectrode, thus skewing the result. The above factors result in the surface area paramete r being a variable, which makes it (and the sensor) extremely difficult to study. In order to overcome this problem, it is necessary to design a sensor with a fi xed electrode surface area. For example, evaporating a fixed amount of Au onto a Si chip and creating an electrochemical cell around it. On the other hand, utilizing the nanowire s brings some advantages. The surface area provided by a small fraction of the nanowir es is comparable to the surface area of evaporating a larger amount of Au onto a wafer. For a larg e scale operation, evaporating the Au onto a Si wafer uses more Au ; thus raising the cost. Creating a larger surface area on multiple wafers by using nanowires from a single Whatman disc can lower the cost substantially. Also, because of their size, nanowires can potentially achieve higher sensitivity, even though the area f ootprint is significantly reduced.

PAGE 73

Figure 5.1: Random Alignment of Nanowires. Reprinted from Nanotechnology, 16, Boote et al., Dielectrophoretic manipulation and electrical c haracterization of gold nanowires, 1500-1505, Copyright (2005), with permissions from IOP Publishing Ltd. and S.D. Evans. 64

PAGE 74

65 The reference electrode used for this se nsor is fabricated externally and is suspended above the cell [14] making it another potential ca ndidate for a variable since the electrode is inserted arbi trarily into the cell. This may introduce variations if the process is required to be repeated multiple ti mes, since the location of the electrode is different every time. Another problem with the external referen ce electrode is that there is a risk of shorting the surface of the me tal between the reference electrode (which is freely suspended over the sensor in electrolyte) and the working and counter electrodes. It was observed in a separate unrelated experiment that shorting the reference electrode to the working and counter electrodes raises the temperature of PBS to a boiling point; thus making the test sample useless. Further MEMS analysis is necessary to come up with a design where all three electrode s (primarily the reference el ectrode) would be fixed in a cell. An example of such an electrode is illustrated in figure 5.2. First, 300 of Ti and Ni are evaporated (or sputtered) on the surface of the Si wafer. This step provides a good adhesion layer for the material that follows. Then, 5000 of Ag is evaporated to create the first layer of the electrode. The epoxy is applied to the wafer to form the groove for bonding and the AgCl paste. Finally, the AgCl paste is added and the reference electrode characterization can be performed.

PAGE 75

Figure 5.2: Ag/AgCl Reference Electrode. Reprinted from Sensors and Actuators B, 97, Kim et al., Enhancement of physical and chemical properties of thin film Ag/AgCl reference electrode using a Ni buffer layer 348-354, Copyright (2004), with permission from Elsevier 66

PAGE 76

67 Finally, having an open, unsealed cell wher e electrochemical measurements are performed leaves room for environmental inte rference with the set-up. Primarily, there is a human factor. Depending on th e time of day the tests are performed (cortisol secretion is most active in the morning, least active at ni ght) it may be possible to skew the results if the person performing the expe riment disturbs the cell by bei ng in close contact with it. More people in a surrounding area only increase the chance for an error. For this reason, there is a definite need for a sealed electro chemical cell as well as protective gear worn by the operator(s) performing the experiment. 5.2 Future Work Based on the above observations, a reco mmendation is made for an improved design of an electrochemical sensor. Fi rst, for the time being, the nanowires are completely eliminated from the sensor si nce they introduce a problem when estimating the surface area of the electrode. Instead, Au is evaporated directly on the Si wafer to form the layer of the working electrode of a fixed surface area. Afte rwards, the reference electrode is fabricated via the procedure de scribed in the previous section; and the counter electrode is fabricat ed by evaporating a layer of Pt onto the surface of Si wafer. Finally, after activation and characterization of Au the assay procedure can be performed (see figure 5.3). The measurement is recommen ded to be performed later in the day by an operator wearing all necessa ry protection to minimize inte raction with the cell. The mathematical model developed in this thesis can be used to model the response of the sensor. After developing a re peatable model of the sensor research to find a better alignment method of the nanowires is recomm ended to repeat the above experiment.

PAGE 77

Figure 5.3: Graphical Representation of Desi gn Improvement for BioMEMS El-Chem Sensor 68

PAGE 78

69 REFERENCES [1] Atkins P.W. Physical Chemistry. Oxford University Press. 1978. [2] Bagotzky V.S. Fundamentals of Electrochemistry . Plenum Press. 1993. [3] Bard A.J. & Faulkner L.R. Electrochemical methods Fundamentals and Applications . John Wiley & Sons, Inc. 2001. 2nd Edition. [4] Bauer J.D., Ackerman P.G., & Toro G. Clinical Laboratory Methods . The C.V. Mosby Company. 1974. 8th Edition. [5] Boote J.J. & Evans S.D. 2005. Diel ectrophoretic manipulation and electrical characterization of gold nanowires. Journal of Nanotechnology Vol: 16: 15001505. [6] Castellano E.E., Main P., & Westbrook E. 1980. The Disordered Structure of Cortisol (11 ,17 ,21-Trihydroxy-4-pregnene-3,20-dione) and Iodocortisol (11 l 7 -Dihydroxy-2 l-iodo-4-pregnene-3,20-dione). Acta Crystallographica Section B. 36. Vol: 12: 3063-3067. [7] Citizendium. Online encyclopedia. 10/09/2008 [ http://en.citizendium.org/wiki/Cortisol ] [8] Flodmark L.E.W., Urke H.A., Halleraker J.H., Arnekleiv J.V., Volestad L.A., & Poleo A.B.S. 2005. Cortisol and glucose responses in juvenile brown trout subjected to a fluctuating flow regime in an artificial stream. Journal of Fish Biology. Vol: 60(1): 238. [9] Galimberti C.A., Magri F., Copello F., Arbasino C., Cravello L., Casu M., Patrone V., & Murialdo G. 2005. S eizure Frequency and Cortisol and Dehydroepiandrosterone Sulfate (DHEAS ) Levels in Women with Epilepsy Receiving Antiepileptic Drug Treatment. Epilepsia. Vol: 46(4): 517 [10] Goulding N.J. & Fl ower R.J. (Editors). Glucocorticoids (Milestones in Drug Therapy). Birkhauser 2001. 1st Edition.

PAGE 79

70 [11] Henry J.B. Clinical Diagnosis and Management by Laboratory Methods . W.B. Saunders Company. 1997. 19th Edition. [12] Jones C. The Adrenal Cortex . Cambridge University Press, 1957. [13] Kissinger P.T. & Heneman W.R. Laboratory Techniques in Electroanalytical Chemistry . Marcel Dekker, Inc. 1996. 2nd Edition. [14] Kumar A., Ph.D. Pers onal Interview. University of South Florida. 04/08/2008. [15] Kumar A., Aravamudhan S., Gordic M., Bhansali S., & Mohapatra S.S. 2007. Ultrasensitive detection of cortisol with enzyme fragment complementation technology using functionalized nanowires. Biosensors and Bioelectronics Vol: 22: 2138-2144. [16] Larkin K.T., Semenchuk E.M., Frazer N.L., Suchday S., & Taylor R.L. 1998. Cardiovascular and behavioral response to social confrontation: measuring reallife stress in the laboratory. Annals of Behavioral Medicine. Vol: 20(4): 294301. [17] Lin C.L., Wu T.J., Machacek D.A., Ji ang N.S., & Kao P.C. 1997. Urinary Free Cortisol and Cortisone Determined by High Performance Liquid Chromatography in the Diagnosis of Cushings Syndrome. Journal of Clinical Endocrinology and Metabolism. Vol: 82(1): 151-155 [18] Marieb E.N. Human Anatomy and Physiology . Pearson Benjamin Cummings. 2004. 2nd Edition. [19] Martin C.R. Endocrine Physiology . Oxford University Press. 1985. [20] Monk P.M.S. Fundamentals of Electroanalytical Chemistry. John Wiley & Sons, Inc. 2001. [21] Nomura S., Fujitaka M., Jinno K ., Sakura N., & Ueda K. 1996. Clinical significance of cortisone and cortisone/cor tisol ratio in evaluating children with adrenal diseases. Clinica Chimica Acta. Vol: 256(1): 1. [22] ODea J.J., Osteryoung J., & Ostery oung R.A. 1981. Theory of Square Wave Voltammetry for Kinetic Systems. Analytical Chemistry. Vol: 53(4): 695-701. [23] Petkus M.M., McLauchlin M., V uppu A.K., Rios L., Garcia A.A., & Hayes M.A. 2006. Detection of FITC-cortisol via Modulated Supraparticle Lighthouses. Analytical Chemistry. Vol: 78: 1405.

PAGE 80

71 [24] Pfaff D.W., Phillips I.M., & Rubin R.T. Principles of Hormone/Behavior Relations Elsevier Academic Press. 2004. [25] Pohl H.A. Dielectrophoresis . Cambridge University Press. 1978 [26] Ramaley L. & Krause M.S. Jr. 1969. Theory of Square Wave Voltammetry. Analytical Chemistry. Vol: 41(11): 1362-1365. [27] Ramaley L. & Krause M.S. Jr. 1969. Analytical Application of Square Wave Voltammetry. Analytical Chemistry. Vol: 41(11): 1365-1369. [28] Rifkin S.C. & Evans D.H. 1976. G eneral Equation for Voltammetry with StepFunctional Potential Change s Applied to Differential Pulse Voltammetry. Analytical Chemistry. Vol: 48(11): 1616-1618. [29] Shikii K., Sakamoto S., Seki H., Utsumi H., & Yamaguchi K. 2004. Narcissistic aggregation of steroid co mpounds in diluted solution elucidated by CSI-MS, PFG NMR and X-ray analysis. Tetrahedron. Vol: 60: 3487-3492. Retrieved on 08-01-2008 from the Cambridge Structural Database. [30] Stulik K. & Pacakova V. Electroanalytical measurem ents in flowing liquids . Ellis Horwood Limited. 1987. [31] Sun K., Ph.D. Candidate. Personal In terview. University of South Florida. 09/05/2008. [32] Tepperman J. Metabolic and Endocrine Physiology . Yearbook Medical Publishers. 1968. 2nd Edition. [33] Turner J.A., Christie J.H., & Osteryoung R.A. 1977. Square Wave Voltammetry at the Dropping Mercury Electrode: Theory. Analytical Chemistry. Vol: 49(13): 1899-1903. [34] Turner J.A., Christie J.H., Vukovic M., & Osteryoung R.A. 1977. Square Wave Voltammetry at the Dropping Mercur y Electrode: Experimental. Analytical Chemistry. Vol: 49(13): 1904-1908. [35] Vanysek P. Modern Techniques in Electroanalysis . John Wiley & Sons. 1996. [36] Whatman Inc. Technical Data Anopore Inorganic Membranes. 09/05/2008. [ http://www.whatman.com/PRODAnoporeInorganicMembranes.aspx ]