Citation
Purine templated overoxidized polypyrrole modified graphite for detection of adenosine

Material Information

Title:
Purine templated overoxidized polypyrrole modified graphite for detection of adenosine
Creator:
Spurlock, Lisa Denise, 1969-
Publication Date:
Language:
English
Physical Description:
xiv, 163 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Carbon fibers ( jstor )
Electric current ( jstor )
Electrodes ( jstor )
Film thickness ( jstor )
Oxidation ( jstor )
pH ( jstor )
Polymers ( jstor )
Polypyrroles ( jstor )
Potassium phosphates ( jstor )
Voltammetry ( jstor )
Chemistry thesis, Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 150-162).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Lisa Denise Spurlock.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
023272502 ( ALEPH )
35122408 ( OCLC )

Downloads

This item has the following downloads:


Full Text










PJRINE TEMPLATED OVEROXIDIZED POLYPYRROLE MODIFIED
GRAPHITE FOR DETECTION OF ADENOSINE












By

LISA DENISE SPURLOCK


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1996
























This work is dedicated to my parents and family who have always encouraged me to pursue a higher education and never settle for anything but the best.














ACKNOWLEDGMENTS


I would like to thank my parents for their constant love and total support. They have always encouraged me to get an education and to enjoy life.

I would also like to express my appreciation to my research advisor Anna Brajter-Toth for her guidance and helpful hints on surviving graduate school. I would like to especially thank my group mates who graduated before me for their words of wisdom, teaching and lasting friendship. They are Ana Marino, Maurice Thompson, Quan Cheng and Charlie Hsueh. I would also like to acknowledge the friendship and assistance of my other group members, Merle Regino and Roberto Bravo. Several others who were undergraduates working with me also deserve recognition for their hard work and dedication; they are Jackie Lewis, Andrew Praserthdam, Constantine Panakos and Ngoc Chou. Also, I would like to give my sincere appreciation to Alonso Jaramillo for his friendship, encouragement and devotion these past few years.

Special thanks go to the faculty and staff who have helped me with instrumentation and experiments. They are Vanecia Young for ESCA analysis of my electrode surfaces, Kathryn Williams for the use of her instruments, Russ Pierce for all of his help with fixing our instruments and Steve Miles for his electronic expertise.

I would also like to thank Jeff Brouwer for hanging in with me and my best friends in Gainesville, Susan Rasmussen and Andrea Pless, for their support and friendship.













TABLE OF CONTENTS




ACKNOW LEDGMENTS .............................................. iii

LIST OF TABLES ................................................... vii

LIST OF FIGU RE S ................................................... ix

AB STR A C T ....................................................... xiii

CHAPTERS

1 INTRODUCTION ......................................... 1

Biological Significance of Adenosine ........................... 1
Metabolism and Catabolism of Adenosine in Cardiomyocytes ......... 3 Advantages of Using Cardiomyoctes in Bioanalysis of Adenosine ...... 4 Electrochemical Properties of Adenosine and Its Metabolites ......... 5 Previous Strategies for Detection of Adenosine ................... 7
Amperometric Based Detection .......................... 7
Potentiometric Based Detection ......................... 8
IPLC for Detection of Adenosine in Bioanalysis ............. 9
Capillary Zone Electrophoresis for Detection of Adenosine .... 12 Properties of Carbon Electrodes .............................. 12
Polypyrrole and Overoxidized Polypyrrole in Sensor Design ......... 15 M olecular Templating ...................................... 21
Fast Scan Voltammetry ..................................... 24
Characterization of Electrode Surfaces by X-ray Photoelectron
Spectroscopy ...................................... 27
Purpose of W ork ......................................... 30

2 EXPERIM ENTAL ........................................ 33

Reagents and Solutions ..................................... 33
E lectrodes .............................................. 34









Reference and Auxiliary Electrodes ...................... 34
W orking Electrodes ................................. 34
Instrum entation .......................................... 38
Electrochemical Experiments .......................... 38
UV Absorption and ESCA Experiments .................. 41
Fundamentals of Electrochemical Measurements .................. 41
Cyclic Voltammetry ................................. 41
Chronocoulometry .................................. 44
Rotating Disk Electrode Experiments .................... 45

3 PREPARATION AND CHARACTERIZATION OF OPPy
AND TEMPLATED OPPy ELECTRODES ............... 47

Procedure for Preparing Ultrathin OPPy Films by Polymerization and
O veroxidation ...................................... 47
Procedure for Preparing Templated OPPy Films on GC by Polymerization
and Overoxidation ................................... 50
Procedure for Preparing Ultrathin OPPy and Templated OPPy Films
at U M E s .......................................... 54
Procedure for Template Release During Overoxidation of Templated
Polypyrrole ........................................ 55
ESCA Analysis of Bare GC, OPPy and OPPy/ATP Films on GC ..... 58

4 ANALYSIS OF THE ELECTROCHEMICAL RESPONSE AT
GLASSY CARBON, GLASSY CARBON COATED WITH
OPPy AND TEMPLATED OPPy ...................... 76

Sensitivity Data at Glassy Carbon and Glassy Carbon Coated with
Ultrathin Films of OPPy and Templated OPPy .............. 76
Selectivity at Bare, OPPy and OPPy Templated Electrodes .......... 87 Adenosine Detection ...................................... 88
Characterization of Film Permeability by Electrochemical Methods .... 91
Determination of Apparent Membrane Diffusion Coefficients .. 91 Effect of Diffusion vs Surface Interactions on Response ...... 93
C onclusions ............................................. 95

5 ANALYSIS OF SURFACE INTERACTIONS AT GC AND
GC TEMPLATED OPPy ELECTRODES ...................... 98

Saturation Binding of Ultrathin OPPy and Templated OPPy Electrodes 98 Analysis of Membrane Interactions ........................... 110
Scatchard Plot Analysis .............................. 110
Langmuir Isotherm Analysis .......................... 115









OPPy and OPPy Templated Surface Stability ................... 118
C onclusions ............................................ 120

6 CHARACTERISTICS OF PURINE TEMPLATED OPPy FILMS ON
CARBON FIBER ULTRAMICROELECTRODES .............. 123

Effect of Film Thickness on the Response of Macroelectrodes ...... 125 Sensitivity of Bare, OPPy and OPPy/ATP Modified Carbon Fiber .... 129 Analysis of Film Interactions at OPPy Modified UMEs ............ 133
Fast Scan Voltammetry of Uric Acid ......................... 133
C onclusions ............................................ 146

7 CONCLUSIONS AND FUTURE WORK ..................... 147

LIST OF REFERENCES ............................................. 150

BIOGRAPHICAL SKETCH ........................................... 163













LIST OF TABLES


Table page

3-1 Comparison of Various Atom Area Ratios obtained from ESCA spectra
for Bare and OPPy and OPPy/ATP modified GC electrodes ............... 72

4-1 Sensitivity of Ru(NH3)63+, Uric Acid and Adenine at Different Surfaces ...... 82 4-2 Non (0,0) fits of Ru(NH3)63+, Uric Acid and Adenine Calibration Plots at
D ifferent Surfaces .............................................. 83

4-3 Sensitivity of OPPy Films Prepared from Different Polymerization Solvents ... 86 4-4 Sensitivity of Adenosine at OPPy Modified GC Electrodes ................ 90

4-5 Apparent Diffusion Coefficients (cm2/s) for OPPy, OPPy/ado and OPPy/ATP . 92 4-6 Slopes of Log Peak Current vs Log Scan Rate Plots at GC and OPPy Film
E lectrodes .................................................... 94

5-1 Scatchard Plot Analysis for Ru(NH3)63' at bare and OPPy Modified GC
E lectrodes ................................................... 113

5-2 Langmuir Isotherm Analysis for Ru(NH3)63+ and Uric Acid at Bare and OPPy
M odified Electrodes ............................................ 117

5-3 Limits of Detection at Bare GC and OPPy Modified Electrodes for Ru(NH3)63+,
Uric Acid and Adenine .......................................... 119

6-1 Sensitivity at OPPy and OPPy/ATP Modified Macro GC Electrodes with
Varying Degrees of Film Thickness ................................ 126

6-2 Langmuir Isotherm Analysis for Ru(NH3)6" and Uric Acid Calibration data
at OPPy and OPPy/ATP ........................................ 128









6-3 Sensitivity of Ru(NH3)63+, Uric Acid and Adenine at Ultramicroelectrode
Surfaces ..................................................... 130

6-4 Langmuir Isotherm Analysis for Ru(NH3)63" and Uric Acid Sensitivity data at
Bare and OPPy Modified Carbon Fiber Ultramicroelectrodes ............. 134

6-5 Limits of Detection at Bare and OPPy Modified Carbon
Ultramicroelectrodes ........................................... 135

6-6 Background Subtracted Fast Scan Voltammetry Results ................ 141













LIST OF FIGURES


Figure pge

1-1 Structures for adenosine, adenine and uric acid .......................... 2

2-1 Diagram of ESCA sample holder with GC electrode and Teflon insulation .... 37 3-1 Suppression of 10 mM Fe(CN)63 response in 0.5 M pH 7.0 potassium
phosphate buffer after repeated coatings of OPPy at GC; scan rate 0.100 V/s,
electrode area 0.07 cm2, deposition charge per coating ca. 35jVC/cm2, film
thickness after 7th coating ca. 28A .................................. 49

3-2 Cartoon representation of template, adenosine or ATP, incorporation into
polypyrrole ................................................... 51

3-3 Cartoon representation of the templated OPPy structure after overoxidation of
polypyrrole (Beck et al., 1987) ..................................... 53

3-4 Suppression of 10 mM Fe(CN)63 response in 0.5 M pH 7.0 potassium
phosphate buffer after repeated coatings of OPPy at UME; radius 7 gm,
scan rate 0.100 V/s, deposition charge 3.17 x 10-2 C/cm2 per coating, film
thickness after 4th coating ca. 0.53gm .............................. 56

3-5 ESCA valence band spectra of bare and OPPy modified GC electrodes
film thickness ca. 12A ........................................... 60

3-6 Wide scan spectrum of OPPy film on GC (ca. 12A thick) showing
the presence of oxygen, carbon and nitrogen .......................... 61

3-7 Structure of overoxidized polypyrrole as proposed by Beck et al.(1987) ...... 63 3-8 Structure of OPPy as proposed by Ge et al. (1994) ..................... 64

3-9 ESCA spectrum of chloride and phosphorous from OPPy film (ca. 12A)
on G C ....... .... .. ............ .... ... ..... .... .. ... . ... .... . 65









3-10 ESCA spectrum of phosphorous, aluminum and silicon of OPPy film
(ca. 12 A thick) on GC ........................................... 66

3-11 Mechanism of polypyrrole overoxidation in water (Beck et al., 1987) ........ 68

3-12 ESCA spectrum of aluminum and silicon in OPPy/ATP films (ca. 16A thick)
on G C ................... . ... .... ... ...... .. .......... . ...... 70

3-13 ESCA valence band spectrum of OPPy/ATP film (ca. 16A thick) on GC ..... 71

3-14 ESCA spectrum of chloride and phosphorous in OPPy/ATP film on GC
(ca. 16A thick) ................................................. 74

3-15 ESCA spectrum of calcium in OPPy/ATP film on GC (ca. 16A thick) ........ 75

4-1 Cyclic voltammograms of 0.4 mM Ru(NH3)63" in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm2, film thickness ca. 16A, scan rate 0.020 V/s ................ 78

4-2 Cyclic voltammograms of 0.2 mM uric acid in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm2, film thickness ca. 16A, scan rate 0.020 V/s ................ 79

4-3 Cyclic voltammograms of 0.3 mM adenine in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm2, film thickness ca. 16A, scan rate 0.020 V/s ................ 80

4-4 Cyclic voltammogram of 0.005 M adenosine (solid line) in 0.5 M pH 7
potassium phosphate buffer at OPPy/ado film (ca. 16 A) GC electrode
(0.07 cm2 area), scan rate 0.050 V/s ................................. 89

5-1 Current vs concentration for Ru(NH3)631 in 0.5 M pH 7.0 potassium phosphate
buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s ........................... 99

5-2 Current vs concentration for Ru(NH3)63" in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 100

5-3 Current vs concentration for Ru(NH3)63 in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ATP. Electrode area 0.07 cm2, film thickness ca. 16 A,
scan rate 0.020 V /s ............................................ 101









5-4 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate
buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 103

5-5 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 104

5-6 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ATP coated electrode. Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 105

5-7 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate
buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 107

5-8 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 108

5-9 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate
buffer at OPPy/ATP coated electrode. Electrode area 0.07 cm2,
film thickness ca. 16 A, scan rate 0.020 V/s .......................... 109

6-1 Background subtracted cyclic voltammogram of 10 IM Ru(NH)63 in
0.070 M pH 7.4 sodium phsophate buffer at electrochemically pretreated carbon
fiber UME (see text for details), electrode radius 7pm, scan rate 1000 V/s,
250 scans signal averaged ........................................ 138

6-2 Background subtracted cyclic voltammogram of 10 pM Ru(NH3)631 in
0.070 M pH 7.4 sodium phsophate buffer at OPPy coated UME,
electrode radius 7pm, film thickness ca. 0.53 jim, scan rate 500 V/s,
500 scans signal averaged ........................................ 139

6-3 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in
0.070 M pH 7.4 sodium phsophate buffer at pretreated UME (see text).
Electrode radius 7pm, scan rate 500 V/s, 250 scans signal averaged ........ 142

6-4 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in
0.070 M pH 7.4 sodium phsophate buffer at pretreated UME (see text).
Electrode radius 7ptm, scan rate 1000 V/s, 250 scans signal averaged ....... 143









6-5 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in
0.070 M pH 7.4 sodium phsophate buffer at OPPy coated UME.
Electrode radius 7pm, film thickness 0.53 pim, scan rate 500 V/s,
50 scans signal averaged ........................................ 144










Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy


PURINE TEMPLATED OVEROXIDIZED POLYPYRROLE MODIFIED GRAPHITE FOR DETECTION OF ADENOSINE

By

Lisa Denise Spurlock

May 1996

Chairperson: Anna Brajter-Toth
Major Department: Department of Chemistry

The objective of this work was to design an amperometric sensor for adenosine by modifying carbon surfaces with ultrathin films of templated overoxidized polypyrrole (OPPy) (ca. 16A thickness) to improve the detection selectivity and sensitivity. Adenosine detection is of interest as a marker of oxygen deficiency to the heart.

The response of adenosine at the film electrodes was greatly enhanced considering that no response was apparent at the bare surface. Despite this, the film electrodes were not sensitive enough for routine detection. The films were characterized with adenine as a model compound because of the greater detection sensitivity to adenine in order to identify interactions favorable for adenosine detection. The effect of molecular templates on film response was characterized additionally with ruthenium hexaamine, a positively charged hydrophilic probe, and a purine, uric acid, a negatively charged relatively hydrophobic probe, and neutral adenine. The results obtained from the electrochemical film characterization









illustrated an increase in film compactness and small changes in film interactions with templating and was verified by Scatchard and Langmuir analysis. Overall, templating with purines decreased film sensitivity with a small improvement in selectivity to adenine. The films also suppressed large background currents (noise) found to interfere with adenosine detection.

The structure of the films and the consequences of templating were additionally characterized by ultraviolet and x-ray photoelectron spectroscopy. Film transport properties were determined from rotating disk electrode measurements.

Carbon fiber ultramicroelectrodes were also modified with templated OPPy films to further characterize the films and to design an in vivo sensor for adenosine. The compact, templated OPPy films were more hydrophilic based on the exclusion of uric acid and good sensitivity for the other probes. A new type of selectivity for uric acid based on counterion exclusion was discovered. These films showed good sensitivity to uric acid in KC1 while suppressing its response in phosphate buffer.

In order to press the limits of detection for uric acid, fast scan voltammetry was used. The results showed that the increased scan rate at the OPPy films improved the sensitivity for uric acid without the need for lengthy surface pretreatment.














CHAPTER 1
INTRODUCTION


Biological Significance of Adenosine


Cardiovascular disease is the number one killer in America and affects one out of every five persons. Because one person out of six will die of a heart attack before reaching age sixty-five, the American Heart Association has spent $93 million in research of cardiovascular disease in the past year (American Heart Association, 1993). Through efforts of intensive research, the heart attack death rate has decreased 32.4% in the last ten years, and will continue to decrease as new methods are developed to prevent cardiovascular disease and to detect heart attacks. In order to detect early warning signs of heart attacks, this research has focussed on developing methods to detect adenosine.

In 1929 Drury and Szent-Gy6rgyi discovered that adenosine modified physiological processes by injecting it directly into mammals. Adenosine as shown in Figure 1-1 is an endogenous cardiac nucleoside produced from the degradation of adenosine triphosphate (ATP). The limited energy reserves from ATP and creatine phosphate in the myocardium, and the high metabolic demand require a constant supply of oxygen and substrates to myocardial cells, in order to preserve normal metabolic and contractile function (Neely and Morgan, 1974; Reimer and Jennings, 1986). If this balance is disturbed, e.g. myocardial ischaemia or heart attack, then rapid depletion of the high energy phosphate occurs, in which







NH2

N " N
tN

HO 10


OH OH

NH2
NN

NH


H
N


H


Uric Acid


Figure 1-1 Structures of adenosine, adenine and uric acid


Adenosine








Adenine









3
80/o of creatine phosphate is lost within minutes of regional ischaemia, and more than half the ATP in fifteen minutes (Jennings et al., 1981). Continual usage of ATP, along with the inability of mitochondria to rephosphorylate adenosine diphosphate (ADP), results in the accumulation of ADP and adenosine monophosphate (AMP) (Jennings et al., 1981). AMP is metabolized via 5-nucleotidase from the cytosolic fraction of the cell to adenosine (Schrader, 1983). Adenosine is then transported from ischaemic myocardial cells, in large quantities, within seconds after onset of oxygen deficiency (Belardinelli et al., 1989; Berne, 1963; Gerlach et al., 1963). Within thirty seconds, the adenosine concentration increases a hundred fold in coronary venous effluent since neutral adenosine can rapidly penetrate the cell membrane (Schrader et al., 1977). If one could monitor this influx of adenosine, then it would be possible to predict a heart attack.



Metabolism and Catabolism of Adenosine in Cardiomyocytes


After adenosine diffuses out of the myocardial cells arteriolar dilation occurs, thus increasing coronary blood flow and increasing myocardial oxygen tension. This in turn reduces the rate of adenine nucleotide degradation and hence adenosine concentration (Berne, 1963). Excess adenosine can be rapidly degraded to inosine by adenosine deaminase which can be further degraded to uric acid (Baer et al., 1966; Belle, 1969), taken into the myocardium and transformed into AMP (Wiedmeier et al., 1972; deJong, 1972; Namm, 1973), or washed out by coronary circulation.

The specific roles of adenosine are still controversial, but it is generally accepted that









4
adenosine relaxes smooth muscle. On the surface of the sarcolemma or muscle membrane, receptors for adenosine and catecholamines exist to control cardiac contractility. Catecholamines control the inward flux of Ca2 which causes muscles to contract. Adenosine counteracts this by binding to its receptor site and acting as negative feedback inhibitor of the catecholamines (Schrader, 1981).



Advantages of Using Cardiomyoctes in Bioanalysis of Adenosine


The high levels of adenosine released during periods of oxygen deficiency make adenosine the ideal molecule to detect for predicting heart attacks. For experimental purposes isolated myocytes are ideal for studying adenosine release. Most importantly, myocytes are large and sediment rapidly without centrifugation, allowing sample medium to be tested or changed without trauma to the cells (Dow, 1989). Myocytes are stable in the presence of Ca2", and they do not degrade nucleosides to free bases. The use of myocytes also allows one to distinguish myocyte specific processes from those occurring in interstitial cells of the myocardium and to eliminate hormonal or neural influences. From an analytical viewpoint, digestion of the heart to form myocytes is ideal since multiple samples from a homogeneous population are produced.

In order to design an adenosine sensor several factors must be taken into consideration. The sensor must have a fast response time because once released, adenosine is rapidly degraded to inosine or taken up by the cell. Also, the sensor must have a small area especially for in vivo and in vitro measurements so as not to disturb the cell metabolism. For









5
these reasons amperometric detection is well suited to adenosine detection, but inherent difficulties have to be considered first.



Electrochemical Properties of Adenosine and Its Metabolites


The structures of adenosine, adenine and uric acid are shown in Figure 1-1. Adenosine of 0.25x103 M has a peak potential (Er) of 1.778-0.087pH V vs saturated calomel electrode (SCE) at a pyrolytic graphite (PG) electrode at 0.005 V/s (Dryhurst, 1972). Under the same conditions, Ep for adenine is 1.338-0.063pH V and a single well formed anodic voltammetric peak can be found at PG between pH 0 and 11. Both peak potentials are concentration dependent and will shift to higher oxidation potentials with increasing concentration (Dryhurst, 1972). The absence of reduction peaks for both adenosine and adenine indicates that the oxidation process is irreversible (Dryhurst and Elving, 1968). Dryhurst also found that both of these compounds adsorb strongly to the PG electrode.

The exact oxidation mechanism for adenosine has not been postulated possibly because the large EP, which is close to the background discharge potential, makes the process difficult to study by electrochemical methods.

Adenine, or 6-aminopurine, undergoes two sequential two electron-two proton oxidations to give first 2-oxyadenine and then 2,8-dioxyadenine. Finally another two electron oxidation at the C,4-C,5 double bond produces an unstable dicarbonium ion intermediate which can react further (Dryhurst, 1972). In vivo adenine was found to be oxidized to 2,8dioxyadenine directly (Bendich et al., 1960). Dryhurst discovered that equal numbers of









6
electrons and protons were transferred during the rate determining step since in cyclic voltammetry the peak potential shifted 51 mV/pH unit to more negative potentials at 0.06 V/s and 56 mV/pH at 0.6 V/s. Analysis at both pH 2.3 and 4.7 using coulometry verified that ca. six electrons are involved in the overall oxidation, but since the electrolysis of 10-3 M adenine takes two days to complete, a very slow intermediate step, an insoluble reaction product, film formation on the electrode, or any combination of these may occur. Dryhurst concluded that adsorption and a slow chemical intermediate step is present in the overall process (Dryhurst and Elving, 1968).

Uric acid, or 2,6,8-trioxypurine, at pH 4.7 shows a well formed two electron oxidation peak at ca. 0.47 V at scan rates below 0.3 V/s at PG (Struck and Elving, 1965). At faster scan rates a well formed cathodic peak appears at ca. 0.42 V and with increasing scan rate becomes relative height to the anodic peak (Dryhurst, 1969). The appearance of this reduction peak at higher scan rates indicates an unstable product which can undergo a rapid chemical follow up reaction (Dryhurst et al., 1983). At a scan rate of 0.005 V/s between pH l and 12, EP is 0.76-0.069pH at rough PG and 0.685-0.055pH V at PG between pH 0 andi 1.5 (Brajter-Toth et al., 1981). The primary oxidation process involves a two electrontwo proton oxidation at the C4-C5 double bond to give a diimine which is very unstable and can be rapidly hydrated to an imine-alcohol (Owens et al., 1978). In high concentrations of phosphate buffer, H2P04" attacks the diimine, and in low concentrations of phosphate buffer H20 attacks the diimine (Goyal et al., 1982) The reduction peak due to the diimine can be observed at rough PG at 0.2 V/s at pH 8.0 provided that the surface has a large area because the diimine strongly adsorbs to the surface, and this adsorbed species is more stable than in









7
solution (Owens et al., 1978). Because the oxidation potential is lower than adenine and adenosine, uric acid is easier to detect and could also be useful in predicting cardiac arrest since it is a metabolite of adenosine (Brown, 1991)



Previous Strategies for Detection of Adenosine Amperometric Based Detection


Dryhurst determined adenine in the presence of large concentrations of adenosine by using polarographic reduction at a dropping mercury electrode (DME) and coupling these results with voltammetric oxidation at a PG electrode (Dryhurst, 1972). In the DME studies, the polarographic limiting current was directly proportional to the sum of the concentrations of adenine and adenosine. In order to quantitate the concentrations of each, oxidation current of the mixture in pH 4.7 acetate buffer was measured at PG electrode, and the fact that both compounds adsorb on the surface was used in quantitation. The concentration of adenine was found to be dependent on the concentration of adenosine present, but at large concentrations of adenosine, adenine is desorbed from the electrode and its oxidation becomes a diffusion controlled process. So, the total amount of adenine and adenosine can be determined from DME studies; the amount of adenine can be determined from the voltammetric results and then respective concentrations for each can be found. Even though this process is relatively straightforward, extensive pretreatment of the PG electrode must be done in order to ensure reproducible results.

The PG electrode was resurfaced with 600-grade silicon carbide paper and then placed









8
into supporting electrolyte where a potential of 0.OV was applied for ten seconds, and then the potential was scanned to 1.4 or 1.5 V three times. Next, the electrode was removed without drying the tip and placed in the background solution, and the process was repeated. The first scan was typically not recorded because it was inconsistent with the rest due to the reduction of surface oxides. Once the background was recorded, the same procedure was repeated in the analyte solution. This method has use for determining both adenine and adenosine in a mixture at pH below 7, but it could not be used successfully for in vivo analysis due to the methods involved as well as the required treatment of the electrode.

Yao et al. used a similar pretreatment procedure for determining purine bases and nucleosides at a glassy carbon (GC) electrode in Britton-Robinson's buffer at pH 2.0-11 (Yao et al., 1977). They also found that these compounds strongly adsorb to GC, and the oxidation peaks of adenine and adenosine are masked in the presence of chloride ions. Adenine and adenosine were best determined at pH 2-4; so this method could not be used for physiological measurements either.

Potentiometric Based Detection


Potentiometric sensors are popular because ammonia, which can be detected by a gas sensing membrane, is a product in the enzymatic conversion of adenosine to inosine by adenosine deaminase (ADA). Deng and Enke have expoited this reaction to detect adenosine by immobilizing ADA onto an ammonia gas permeable membrane. This sensor has a limit of detection (LOD) at pH 9.0 in the micromolar range and is stable for a month. The problems








9
with this sensor are dependence on pH, which limits its use in some biological samples, and a response time often minutes.

Xiuli et al. have also designed a sensor based on an ammonia sensing membrane, but they have immobilized rabbit thymus tissue as the enzyme source (1992). This sensor also has a LOD of ca. 10' M with a response time of seven minutes. This sensor is useful for body fluid samples, but is not sensitive enough for use in blood. However, adenosine can be detected in blood with no interferences or sample pretreatment if it is added directly to the sample. Another sensor based on this method was developed by Bradley and Rechnitz in 1984 who used mouse small intestinal mucosal cells as the enzyme source.

In order to overcome the long response times of the previous sensors, Liu et al. designed an asymmetric polyurethane membrane consisting of a very thin hydrophilic polyurethane membrane, with a high density of polylysine groups to which ADA is attatched, coated to a hydrophobic plasticized polyurethane/poly(vinylchloride/vinyl acetate/vinyl alcohol) membrane, which adheres well to silicon dioxide surfaces and contributes to the good stability (Liu et al., 1993). This electrode has an LOD ca. 105M and takes sixty seconds to reach steady state. However, it has only been tested in buffer solutions of adenosine.

Potentiometric sensors are advantageous because of the specificity of ADA to adenosine, but these sensors suffer from problems such as response time and difficulty in miniaturization that limits their routine use in bioanalysis of adenosine. HPLC for Detection of Adenosine in Bioanalysis


High performance liquid chromatography (HPLC) is the most widely used method for








10
detection of adenosine. HPLC with electrochemical (EC) based detection is preferred because EC detection is more sensitive than ultraviolet absorption (UV), and because it is a more direct method than fluorescence, as adenosine and its metabolites have to be tagged for use in fluorescence detection. Henderson and Griffin have separated adenine and adenosine as well as other purine compounds with a reversed phase column using a glassy carbon electrochemical cell set at 1.5 V vs Ag/AgC1 reference (Henderson and Griffin, 1984). Their procedure can be used for biomaterials with some sample clean up, but this method suffers from fouling of the EC cell due to adsorption of purine molecules to GC and contaminants which can be oxidized due to the large detection potential applied.

Berne, et al. also have used EC detection, but the adenosine in the sample is first enzymatically converted to uric acid (Beme et al., 1986). This method was used to determine the interstitial fluid content of adenosine in different tissues and had a LOD of 5 finol. The samples were collected from a small chamber placed on the epicardial surface of a dog's heart. After ADA, nucleoside phosphorylase and xanthine oxidase were added to convert adenosine to uric acid. The reaction was stopped by the addition of HC1O4, and the precipitate was removed by centrifugation and stored overnight at 4�C. Before analysis, the sample was centrifuged once again. This technique can measure adenosine below basal levels so that any increase in concentration, as expected during cardiac arrest, can be measured, but a lengthy sample pretreament has to be done before analysis.

Ontyd and Schrader have determined adenosine, inosine and hypoxanthine in human plasma by reversed phase HPLC using UV detection at 254 nm. The inaccurracy of the determination of adenosine levels in plasma stems from the sample collection since adenosine








11
can be rapidly degraded. Ontyd and Schrader have overcome this by designing a syringe which stops any degradation once the blood is drawn from the patient by immediately mixing the blood with dipyridamole in Locke solution. After collection the sample was run once; then ADA and nucleoside phosphorylase were added and another chromatogramn was taken. The enzymatic peak shift was used to confirm the presence of the analytes, since a number of unidentified peaks are present in the first chromatogram. The only drawback of this approach was that the chromatograms were not consistent between individuals mainly because the nutritional states and pathological conditions varied.

HPLC with fluorescence detection has also been developed for detection of adenosine and inosine. Gardiner monitored the formation of dichlorofluorescein which is the oxidation product from the reaction of dichlorofluorescin with H202 in the presence of horseradish peroxidase (Gardiner, 1979). H202 is a product in the conversion of hypoxanthine by xanthine oxidase to uric acid. The amount of adenosine is quantitated by measuring the conversion of inosine plus adenosine to hypoxanthine and subtracting the conversion of inosine to hypoxanthine. This method has a linear dynamic range of 6.2 x 10-9to 9.3 x 10-7M and has been tested on dogs during a five minute reversible occlusion of the left anterior descending coronary artery. The concentration changes of adenosine in blood could be detected, but at least ten minutes were needed for the enzymatic conversion processes.

Another HPLC with fluorescence detection method was developed by Preston for use in determining nucleoside concentration in marine phytoplankton (Preston, 1983). The adenosine samples had to be incubated in acetate buffer and chloroacetaldehyde for thirty








12
minutes. As a result, 1,N6-etheno derivatives which fluoresce are the detected products. The LOD is 2 x 10.9 M and no interferences have been discovered. Capillary Zone Electrophoresis for Detection of Adenosine


Kuhr and Yeung have used capillary zone electrophoresis (CZE) with laser induced indirect fluorescence detection for the determination of nucleosides and proteins (Kuhr and Yeung, 1986). This method has good sensitivity with a linear dynamic range (LDR) of 50 to 100 amol and has been tested for detection of lysozymes, but has not been used in real samples.

Separations as well as amperometric and potentiometric sensors have offered insight into the metabolism and chemistry of adenosine, but still no sensor exists which can be used directly in vivo for long periods of time. This research has set out to develop a sensor surface which does not foul due to adsorption of the analyte and products, which has real time response and requires no sample pretreatment.



Properties of Carbon Electrodes


Carbon is the most widely used and extensively studied electrode material for analytical purposes (McCreery, 1991). Carbon has high conductivity, is readily available, is cheap, has low chemical reactivity and has a wide potential window at biological pH (Kinoshita, 1988). The most widely used forms of macroelectrode materials are PG and GC (Dryhurst and McAllister, 1984), and carbon fiber has also become popular for use in ultramicroelectrodes.







13

Glassy carbon (GC) or vitreous carbon is made from polymeric resins such as polyacrylonitrile or phenol/formaldehyde polymers heat treated at 1000-3000'C under pressure to release H, N, or 0 atoms to leave an extensively conjugated sp2 carbon structure. The polymeric backbone remains intact forming a complex structure of interwoven graphitic ribbons which accounts for its mechanical hardness. GC is impermeable to gases, resistant to chemical attack, electrically conductive and can be obtained in high purity, making this material very advantageous for analysis (Kinoshita, 1988). However, this surface is very inactive despite its small degree of hydrophilicity.

Pyrolytic graphite (PG) is prepared by high temperature decomposition, above 1200�C but below 3800�C, of gaseous hydrocarbons such as methane and propane onto a hot surface. PG consists of highly conjugated hexagonal rings or carbon atoms arranged in planes held together by weak van der Waals forces (Kinoshita, 1988). The conductivity along the planes (edge plane) is metallic, but conductivities perpendicular to the planes (basal) are semiconducting (Kinoshita, 1988). PG is a more active surface than GC due to these properties. The exact structural and physical properties of PG depend on the surface of deposition, temperature of curing and further treatment after manufacturing. For our purposes PG was roughened to form rough pyrolytic graphite (RPG) using 600-grit silicon carbide paper.

Because of the highly desirable properties of carbon and the need for miniaturization for in vivo analysis carbon fiber has become popular. Carbon fibers generally range in diameter from 5 to 25 p.m and are made from heat treatment of polymeric precursors or catalytic chemical vapor deposition (CCVD) (McCreery, 1991). Heat treatment procedures








14

are similar to those used in the preparation of GC, but the molten polymer is spun or extruded to form fibers. CCVD involves a catalytic dehydrogenation of hydrocarbons like benzene onto small particles of Fe, Ni or Co. Carbon atoms diffuse to the surface or through the bulk of the metal particle after decomposition of the reagent gas and precipitate to form graphite on the lower cooler side of the metal particle. The carbon fibers formed from either method can have different structures depending on the specific procedure, but the Raman spectra of carbon fibers prepared at high temperatures are similar to those of highly ordered pyrolytic graphite, with preferred graphitic plane orientations (McCreery, 1991). Carbon fiber surfaces have not been as extensively characterized as GC and PG have been.

One of the disadvantages of graphite as an electrode material in bioanalysis is the poor reproducibility of the electrode surface. The irreproducibility is related to adsorption as well as to the oxidation of the surface, which is especially apparent at high positive potentials (McCreery, 1991). To improve the reproducibility the electrodes must be pretreated before measurements. Additionally, large background currents at high positive potentials, from electrode charging and from the faradaic surface reactions, involving surface groups such as quinones, have to be controlled (McCreery, 1991).

For analysis at graphite several surface pretreatment procedures, including polishing (Kaman, 1988), electrochemical activation (Engstrom, 1982; Engstrom and Strasser, 1985), laser pretreatment (Poon and McCreery, 1986, 1987; Poon et al., 1988) and heat treatment (Fagan et al., 1985) have been developed. Other promising methods include dynamic modification of graphite with surfactants (Marino and Brajter-Toth, 1993) and with membranes (Murray, 1992) such as Nafion (Waller, 1986), a perfluorosulfonated ionomer.








15

Nafion membranes preconcentrate positively charged analytes while excluding interfering anions, and the membranes limit adsorption, protecting the electrode surface. In vivo measurements of neurotransmitters are a successful example of Nafion applications, where favorable partitioning of neurotransmitters into the membrane improves detection sensitivity while the interfering negatively charged ascorbic acid is excluded (Gerhardt et al., 1984; Nagy et al., 1985), with the membrane protecting the surface from direct contact with the complex biological sample. Nevertheless, Nafion coated electrodes suffer from several problems such as slow response time due to low diffusion coefficients in the film, memory effects due to strong binding with the SO3 groups in the film, and saturation (Guadalupe and Abrujia, 1985; Whiteley and Martin, 1988; Wang and Tuzhi, 1986). Also, Nafion coatings are often nonuniform in thickness and reproducibility of film formation is poor (Gao et al., 1993). Gelimmobilized enzymes have recently been developed for bioanalysis (Wang and Heller, 1993) although these require multicomponent films to facilitate the enzymatic reaction.



Polypyrrole and Overoxidized Polypyrrole in Sensor Design


Ultrathin overoxidized polypyrrole (OPPy) films are another type of membrane that has recently been developed for analytical applications (Witkowski et al., 1991; Witkowski and Brajter-Toth, 1992; Hsueh and Brajter-Toth, 1994). It has been shown that OPPy films can be easily prepared by electropolymerization so that the response is not limited by slow infilm diffusion and irreversible interactions (Hsueh and Brajter-Toth, 1994; Witkowski and Brajter-Toth, 1992). Ultrathin OPPy films have been shown to be cation-selective and have








16

been shown to eliminate the response of Fe(CN)63" (Hsueh and Brajter-Toth, 1994) presumably because of the high electron density of the film carbonyl groups (Palmisano et al., 1995; Beck et al., 1987; Christensen and Hamnett, 1991); similar interactions can suppress response of uric acid (Witkowski and Brajter-Toth, 1992) which is a common biointerferent. Another desirable feature of OPPy films is that the membrane sensitivity and selectivity can be changed depending upon the conditions in which the films were made. This work has sought to alter the properties of ultrathin OPPy films by templating these films with purine molecules in order to design a sensor for in vitro use for adenosine and its metabolites.

In 1980, Diaz and Castillo reported the growth of thin polymer films of polypyrrole, grown on Pt electrodes by electropolymerization (Diaz and Castillo, 1980). Polypyrrole prepared in this manner is deposited as a cation and consists primarily of a,a' linkages, but n-coupling can also occur (Salmon et al., 1982). Polypyrrole has variable conductivity and can change reversibly from a conducting to a nonconducting polymer within a given potential window (Diaz and Castillo, 1980; Feldman et al., 1985). Because of these properties, polypyrrole films have been used for a wide variety of applications.

Several factors influence the behavior of polypyrrole films such as the applied potentials during polymerization (Asavapiriyanont et al., 1984), the solvent employed (Diaz and Castillo, 1980; Asavapiriyanont et al., 1984; Ferreira, 1990) and the temperature (Ogasawara et al., 1986). However, the most important factor is the counterion incorporated during polymerization (Imisides et al., 1991). The morphology, conductivity (Yamaura et al., 1988), adhesion and mechanical strength are all affected by this choice (Skotheim, 1986; Freund et al., 1991). Polypyrrole can be readily synthesized from a range of solvent media,







17
including aqueous and nonaqueous solvents, so the choice of the counterions is endless (Imisides et al., 1991). In general, polypyrrole films grown from nonnucleophilic solvents and electrolytes, i.e. acetonitrile with 1% H20, have good conductivity, and aqueous solvents produce less conductive films (Imisides et al., 1991). Aqueous solvents are advantageous for most applications, especially industrial, since a larger variety of dopants can be used as the counterion, but the potential of the anion must be higher than pyrrole to allow pyrrole to polymerize without competition from the electrolyte (Takakubo, 1987).

Typical dopant levels are one counter ion for every three to four pyrrole units (Mitchell et al., 1988). These ions are incorporated as counterions during the oxidation of pyrrole, in which every third repeat unit has a positive charge (Diaz and Castillo, 1980). The exact chemical configuration of the polypyrrole chains is ambiguous, but it is accepted that the molecular organization is highly disordered. Mitchell, et al. have discovered that more anisotropic counterions lead to a higher level of preferred orientation in the film, but anions with more than three aromatic units do not make a good film because these larger counterions may separate the polymer chains or separate some sections (Mitchell et al., 1988). Aromatic conterions are preferred since these will aid in retention of the polymer to the electrode surface since polypyrrole is planar (Mitchell et al., 1988). Some example of counterions used are surfactants, such as dodecylbenzenesulfonate ion (Lyons et al., 1993), ferricyanide (Chen et al., 1993) and cobalt porphyrin (Armengaud et al., 1990).

Polypyrrole has been polymerized on various substrates such as Pt (Diaz and Castillo, 1980), Au, Fe, indium-tin oxide coated glass (Street et al., 1983), mercury (Bradner and Shapiro, 1988) and plastic films coated with gold, tin oxide or silver (Barisci and Wallace,








18
unpublished). Titanium, aluminum, mild steel and brass (Cheung et al., 1988) have been also been used. Glassy carbon is the best choice because of its large potential window and polypyrrole films have been shown to strongly adhere (Imisides et al., 1991); polypyrrole cannot be removed except by mechanical grinding or treatment with chromic acid. Nishizawa et al. have shown that polypyrrole adheres well to hydrophobic substrates, so glassy carbon is a good choice since glassy carbon is mostly hydrophobic (Nishizawa et al., 1991).

If the oxidation potential is increased beyond 0.8 V, polypyrrole becomes nonconducting (Asavapiriyanont et al., 1984). Beck et al. have shown that overoxidation in aqueous solutions occurs via a nucleophilic attack of the hydroxide or water on the pyrrole unit, followed by oxidation of the hydroxy group to a carbonyl group ( Beck et al., 1987). Wernet and Wegner have also reported that polypyrrole can be overoxidized by cycling in very basic solutions like NaOH (Wernet and Wegner, 1987). This treatment may produce a mixture of carbonyl groups as well as negatively charged hydroxyl groups (Gao et al., 1994). Overoxidation results in oxidation and final scission of the polypyrrole chains at a small number of sites, such that smaller chains are entangled at the surface. As a result, the conjugation of the polymer is disrupted, but no significant material is lost (Christensen and Hamnett, 1991) even though total release of the counterions may occur during the course of overoxidation (Beck et al., 1987). This infers that the structure of polypyrrole is not greatly changed by overoxidation; so incorporation of a counterion to change the structure will not be futile even though the counterion could be lost during overoxidation.

Several groups have used overoxidized polypyrrole (OPPy) films for sensing purposes because of its ability to exclude anions (Freund et al., 1991). Gao and Ivaska polymerized








19
pyrrole on glassy carbon in the presence of sodium dodecyl sulfate, and then overoxidized polypyrrole in NaOH (Gao and Ivaska, 1993). They observed high selectivity for dopamine in the presence of ascorbic acid with a two minute preconcentration. The exclusion of ascorbic acid was presumably due to the presence of the hydroxyl and carbonyl groups within the OPPy structure which also can contribute to the increase in sensitivity (LOD 40 nM) of dopamine which is positively charged. The enhanced selectivity to dopamine is accounted for by the use of sodium docecyl sulfate during the polymerization process because once the film is overoxidized the surfactant is removed leaving a more porous structure than a smaller ion. The presence of the film also protects the electrode surface from fouling by the oxidation product of dopamine. However, the drawbacks with this sensor are long preconcentration times as well as renewal of the film which has to be done before each measurement, since the film thickness is ca. 1 pim thick.

Gao et al. have used indigo carmine as the counterion during polymerization of pyrrole to develop a sensor for dopamine in the presence of ascorbic acid (Gao et al., 1994) for use in in vivo measurements. The same procedure as described above was used. They found that with ascorbic acid concentrations lower than 0.2 mM, which is the normal concentration level in mammalian brain, the selectivity with 0.25 to 1.0 gim thick film is optimal for detection of dopamine between 0.1 to 10 [M. Thinner films produce better detection limits, but large oxidation currents stemming from the oxidation of ascorbic acid cause problems with detection of dopamine. The problems of this sensor are the difficulty in controlling film thickness, and hence film reproducibility, and the small lifetime of two hours.

Besides changing the conditions during polymerization, Gao et al. have also varied the








20
overoxidation conditions to change the permeability of the films (Gao et al., 1994). A wide variety of solvents, e.g. CH3OH, NaOH, HCI and CH3CN, were used to overoxidize polypyrrole after polymerizing with sodium dodoceyl sulfate. Their results showed that the most sensitive and well defined voltammograms for dopamine were obtained from the polypyrrole film overoxidized in NaOH solution, due to the formation of both carbonyl groups and hydroxyl groups. The existence of these functional groups was confirmed by Fourier transform IR spectroscopy. The porosity of the films was controlled by the doping ion, but the permeability was increased by increasing the pH of the electrolyte used to overoxidize. For practical purposes this sensor is limited to dopamine concentrations lower than 1 X 10" M due to the saturation and renewal problems. These films only need a sixty second preconcentration time, but renewal by soaking the electrode in phosphate buffer takes up to three hours for complete removal of analyte. Despite this, these films exhibited excellent antifouling properties in albumin solutions with doparnine where a rapid loss in the oxidation current at bare GC occurred.

Centonze et al. have manipulated the permselective and antifouling properties of OPPy for use as a glucose sensor. This group has immobilized glucose oxidase in 0.67 Atm, or thicker, OPPy films on GC overoxidized in phosphate buffer for use in a flow injection apparatus. This sensor has a shelf life of ten days and shows a 75% decrease in sensitivity after six days of continuous use. In addition this sensor is interference free as the currents for several common interferents (ascorbic acid, uric acid, cysteine, acetaminophen) are suppressed in comparison to the glucose response. The linear dynamic range of the sensor is 1.0 X 10-2 to 5.0 X 10.2 M, but can be extended to lower concentrations by use of a film








21

twice as thick. This sensor also shows the feasibility for use in real samples; a pooled serum sample was tested by the sensor in a flow injection apparatus as well as by a routine enzymatic colorimetric methods. The results were not significantly different, indicating that this sensor can be used in real samples.

Based on these results, the feasibility for tailoring overoxidized polypyrrole by use of specific counterions during the polymerization process as well as by the optimization conditions for overoxidation can produce a sensor with desirable characterics for detection of adenosine and its metabolites. Others have designed polymers and membranes for structurally similar molecules to adenosine in a similar manner by molecular templating.



Molecular Templating


Molecular templating or molecular imprinting involves the preparation of polymers that are selective for a particular compound (Edelman and Wang, 1992). The compound of interest acts as a template in which monomers are prearranged and complementary interactions occur. Next, the monomers are polymerized about the template and in the final step, the template is removed by an extraction or other method to leave a specific binding site for the template or a structurally related compound. In general, two approaches have been taken: (1) the template has been covalently but reversibly bound (Wulff, 1986), or (2) the initial interactions between monomers and the print molecule have been non-covalent (Arshady and Mosbach, 1981). An example of reversible covalent binding is the polymerization of vinylphenylboronic acid with phenyl-a-D-mannopyranoside as the template







22

to design polymers that could resolve racemic mixtures of carbohydrates (Wulff, 1986). Boronate groups have also been used in this manner for detection of nucleotides (Norrlow et al., 1986). This method is hampered by a limited number of compounds with suitable binding groups and useful reversible interactions that do not disrupt the polymer matrix (Ekberg and Mosbach, 1989).

Polymers designed by non-covalent interactions are more versatile because more monomer-print molecule interactions including ionic, hydrogen bonding, hydrophobic, or charge transfer can be used (Ekberg and Mosbach, 1989). Early work with this approach used dyes as the template in acrylic monomers (Arshady and Mosbach, 1981) for selectivity to the dye itself Mosbach's group has used this approach for separating amino acid derivatives by interacting the amino group of the print molecule with the carboxyl group of a monomer to separate amino acid derivatives on the basis of substrate and enantioselectivity (Ekberg and Mosbach, 1989).

In order to design a selective polymer, several factors have to be taken into consideration. In general, templates that can form multiple interactions with the monomer form polymers with the best resolution or higher selectivity (Wulff and Lohmar, 1979; Wulff and Gimpel, 1982; Wulff et al., 1980, 1984). Reduced specificity can also occur if the template molecule is polymerized very close to the surface in that an incomplete imprint is formed (Wulff, 1986). The polymerization conditions are crucial since polymerization at extreme conditions such as thermal decomposition destabilizes the complex between the template and monomers and limits the types of templates that can be used. In order to maintain the polymer structure as dictated by the template, crosslinkers are employed and








23

generally, less crosslinking should produce a less defined matrix with poor resolving ability (Wolff, 1986). A compromise in rigidity has to be made because the polymer has to be flexible enough to allow fast binding and diffusion from the cavity (Wulff, 1986).

Molecular imprinting has been used to design a wide variety of sensors especially for biomolecules since the theory behind this process relies on the mechanism for molecular recognition as found in nature. Network polymers have been prepared for detection of adenine derivatives (Shea, et al., 1993), antibody mimics (Vlatakis et al., 1993) and for HPLC column packing (Sellergren et al., 1985; Andersson et al., 1990; Kempe and Mosbach, 1994). Others have used the concept of molecular imprinting to develop anion sensors (Ikariyama and Heineman, 1986) and potentiometric sensors for Cl- (Dong et al., 1988) and NO3 (Hutchins and Bachas, 1995) but in the potentiometric methods, the template molecule was left in the polymer matrix for additional interactions with the anions and was not removed to leave the molecular binding site.

Based on the principles of molecular templating and the permselectivity of OPPy, this work has sought to design templated OPPy structures for use in detection of adenosine and its metabolites. The templates used in templating polypyrrole first and then in forming templated OPPy were structurally related to adenosine and were expected to be incorporated and to weakly interact with the forming polypyrrole through hydrogen bonding, hydrophobic and electrostatic interactions, provided the template was anionic. Templating should alter polypyrrole morphology based on the changes reported in polypyrrole morphology with electrolyte, solvent and the polymerization conditions (Gao et al., 1994). The template was expected to be released during overoxidation of polypyrrole in phosphate buffer typically







24

carried out at potentials slightly lower than the oxidation potentials of the templates. The molecular sites from templating were expected to remain in OPPy since the carbonyl groups introduced in the formation of OPPy do not significantly alter polymer morphology (Christensen and Hamnett, 1991). Characterization and utility of these structures were done by UV absorption, x-ray photoelectron spectroscopy and electrochemical analysis including fast scan voltammetry.



Fast Scan Voltammetry


In order to further enhance the sensitivity and selectivity at cabon fiber electrodes, fast scan voltammetry can be used since the current is proportional to the square root of scan rate for diffusion controlled systems and proportional to the scan rate for an adsorption controlled process (Bard and Faulkner, 1980). At scan rates above 100 V/s adsorption can dominate even for weakly adsorbing species since the current increases faster with scan rate, thus allowing higher peak current, lower detection limits and an increase in signal to noise (Hsueh and Brajter-Toth, 1993; Freund and Brajter-Toth, 1992; Wiedemann et al., 1991). Fast scan rates can only be used at ultramicroelectrodes since they have a small time constant and a small iR drop. Fast scan voltammetry was first introduced by Millar's group (Millar et al., 1981) and was later improved by Wightman's group (Kuhr and Wightman, 1986). One of the advantages of fast scan voltammetry is that the differences between the electrochemical kinetics of the analyte and the interferant can be used to resolve the two signals (Baur et al., 1988). An example of this is the measurement of dopamine in the presence of ascorbic acid.







25

The electrochemical kinetics of ascorbic acid are slow, so fast scan voltammetry enhances this affect so that the oxidation peak potential of ascorbic acid shifts away from the oxidation peak potential of dopamine as the scan rate increases (Hsueh and Brajter-Toth, submitted).

Another advantage of fast scan voltanmetry is the preservation of the working electrodes. Many redox reactions of biomolecules are followed by chemical reactions which produce side products. These can adsorb on the electrode surface and eventually cause fouling after several scans. But because of the short time span of fast scan voltammetry, the products can undergo electrochemical reactions and be converted back to the original analyte before chemical reactions can occur in the solution that would foul the surface.

Fast scan voltanmetry can also provide qualitative information about the analyte which may be used to confirm the identity of the analyte of interest in complex samples. At high scan rates, thickness of the diffusion layer of biomolecules like doparnine will be greatly reduced, so the chance of blocked diffusion by an in vivo environment is reduced. Hence, fast scan voltammetry can guarantee that voltammograms obtained in vivo will be similar to those obtained in buffer solution.

Further enhancement can be achieved by signal averaging which produces high temporal resolution since the time required to complete a voltammogram is short. Wightman's group has investigated this technique to improve detection limits or increase signal to noise (Wiedemann et al., 1991). Signal averaging along with analog and digital filtering allowed the detection of 1.0 x 10-7 M of dopamine in vivo. They averaged forty voltammograms at a scan rate of 300 V/s to improve the signal to noise. The number of voltammograms signal averaged is limited by the scan rate and the experimental time scale.








26

Shorter cycling time allows more cyclic voltammograms to be acquired in a specific time. The enhancement in signal to noise is expected to be proportional to the square root of the number of scans (Hsueh and Brajter-Toth, submitted). For example, for 3125 averaged scans, the signal to noise is improved fifty-six times. High frequency noise is rapidly reduced by signal averaging as the number of scans increases, but the low frequency noise remains the same even after a thousand scans (Hsueh, 1995). The low frequency noise may be harmonic with or close to the frequency of cycling.

Fast scan voltammetry can improve the selectivity and sensitivity of bare carbon fiber as well as coated carbon fiber electrodes. Wightman's group has used fast scan voltammetry to enhance the kinetic differences of dopamine and ascorbic acid at Nation coated electrodes, since ascorbic acid shows slow kinetics at Nation and dopamine partitions favorably into the film (Baur et al., 1988). The direct result of coupling fast scan with Nafion is an increase in sensitivity when sufficient time is allowed for the diffusion layer to relax following a scan. The presence of the Nation film decreases adsorption at the surface, but still increases the sensitivity for dopamine. However, the voltammogram is more irreversible than at the bare electrode. Despite this, the Nation film keeps the electrode surface in uniform condition so that the large background current is kept to a minimum allowing a clear observation of the faradaic events (Kristensen et al., 1987).

This large background current is the major limitation to fast scan voltammetry. The charging current increases proportionally to the scan rate (Wipf et al., 1988); so most of the resolution of the digital oscilloscope is consumed by the higher charging current which accompanies all electroanalytical signals at high scan rates. The left over resolution does not








27

accurately reflect the shape of the analyte signal because of the noise introduced by the lack of resolution of the digitized noise. This noise will prevent low concentrations of analyte from being detected even if the noise attributed to the cell and the current measuring circuit could be eliminated.

The high frequency noise also becomes more difficult to eliminate as the scan rate increases, but this can be reduced by adequate circuit design, digital filtering or smoothing of the data.



Characterization of Electrode Surfaces by X-ray Photoelectron Spectroscopy


In X-ray photoelectron spectroscopy (XPS) or electron spectroscopy for chemical analysis (ESCA) photons from a monochromatic X-ray beam of known energy can displace electrons from atomic orbitals of atoms, ions or molecules or from bands of solids. The kinetic energy of discharged electrons (or the power of the electron beam) is plotted as a function of the energy (or the frequency or wavelength) of the emitted electrons. Hence, ESCA provides a means of qualitative identification of the elements present on the surface of solids since every element in the periodic table has one or more energy levels that will result in the appearance of peaks from 0 to 1250 eV binding energy in a low-resolution, wide-scan spectrum. Usually, the peaks are well resolved and lead to unambiguous identification if the element is present in concentrations greater than 0.1%. If one of the peaks is further analyzed, using a higher energy resolution, the surface environment can be characterized because the position of the maximum depends upon the chemical environment








28
of the atom responsible for the peak. The variations in the number of valence electrons and the type of bonds they form influence the binding energies of core electrons. In general, binding energies increase as the oxidation state becomes more positive, because when an electron is removed, the effective charge on the core electron increases thus increasing the binding energy.

In addition to giving qualitative information about the types of atoms present in a sample, ESCA can provide information on the relative number of each atom type as well as their oxidation state. For this reason, ESCA is widely used to characterize surfaces. Normally, only the top layer of the surface is analyzed, but with sputtering the elemental composition of the surface bulk can be analyzed. Sputtering allows the depth profile of the surface to be studied as it is being etched away by a beam of argon ions. This has been very useful in applications to corrosion chemistry, catalyst behavior and properties of semiconductors.

In 1971 to 1981, Clark and Harrison investigated core-level binding energy shifts for atoms in polymers, but since then no major additions to this work have been made, largely due to tedious curve fitting. Since then computerized curve fitting has made ESCA a powerful analytical tool. As mentioned earlier, ESCA is favored for the characterization of surfaces, so with the development of a wide range of sensors based on polymer coatings, ESCA has become a widely accepted method of surface characterization.

ESCA has been especially useful for the characterization of polypyrrole and overoxidized polypyrrole because polypyrrole films are difficult to characterize since they are insoluble and less crystalline (Street et al., 1982; Street et al., 1983) than other polymers








29

(Pfluger and Street, 1984). The lack of structural data has hindered accurate band structure calculations and has complicated the interpretation of available data. Pfluger and Street have studied conducting polypyrrole grown on Pt substrates with thicknesses of 1-5 Pm (Pfluger and Street, 1984). The pyrrole P carbons have a binding energy centered at 283.6 eV, and the a carbons have a binding energy at 284.5 eV. The main pyrrole peak for NIs appears at 399.6 eV, but the spectrum shows the presence of three inequivalent nitrogen sites. This indicates that the charge is fairly localized at the N sites. Also, the presence of counterions was detected and for polypyrrole/perchlorate polymer, the anion to ring ratio was 1:3, confirming that every third repeat unit has a negative charge.

Ge et al. have studied OPPy grown on gold foil and overoxidized in perchlorate solution by ESCA (Ge et al., 1994). In comparison of the spectra of polypyrrole and OPPy, the Cl 2p signal at 207 eV was lost, the amount of high binding energy nitrogen was reduced significantly as the low binding energy component increased, and a new peak at 287.4 eV appeared on the high binding energy tail of C ls electrons at 284.6 eV with overoxidation. The composition of polypyrrole was C5.3Ni.000.37(C104) 0.24, and the composition of OPPy was found to be C6.5N1.001.58 (C104) 0.o5. This shows that the amount of 0 groups greatly increased as postulated by the formation of carbonyl groups during overoxidation (Beck et al., 1987) while the number of counteranions decreased significantly due to their expulsion from the film as the positive charge is lost. These results are consistent with previously proposed mechanisms of overoxidation (Beck et al., 1987).

Palmisano et al. also characterized polypyrrole and OPPy structues by ESCA. Their membranes were grown on a Pt disk and were overoxidized in phosphate buffer (Palmisano








30

et al., 1995), and the findings are in general agreement with the spectra obtained by Ge et al. The carbon to nitrogen ratio remained nearly constant in going from polypyrrole to OPPy and the oxygen content as measured by the oxygen to nitrogen ratio increased. The CI 2p signal also decreased with overoxidation. The spectra also showed a small P2p signal possibly belonging to the buffer which suggested that positive charges were present in the OPPy structure. However, the residual positive N was not fully balanced by phosphate species considering the atom ratios, so partial formation and ionization of COOH groups might have occurred. Both a and P3 carbons are involved in these COOH groups,so this implied that the loss in conductivity during overoxidation was due to the breakage of polymer chains. This breakage of polymer chains suggests different kinds of regions in the film. The removal of positive charges on the polymer decreasing N' and increasing N=C creates hydrophobic regions, and the introduction of carbonyl and carboxylic groups makes the film more hydrophilic. Palmisano and coworkers developed a model based on this theory where the bulk polymer is hydrophobic with hydrophilic micropores which favor neutral species over anionic species. This was previously hypothesized by Witkowski and Brajter-Toth (Witkowski and Brajter-Toth, 1992). This theory was tested by measuring the permeability by rotating disk electrode experiments and was confirmed.



Purpose of Work


The focus of this work was to design an amperometric sensor for adenosine by modifying carbon surfaces with templated ultrathin OPPy films. Adenosine is inherently








31

difficult to dectect due to its large oxidation potential where the oxidation current can be obscured by large oxidation currents stemming from the background. Adenine, inosine and adenosine triphosphate, all structurally similar to adenosine, were used as template molecules and were incorporated into polypyrrole polymerziation to enhance the sensitivity and selectivity of the polymer film to adenosine. The modified electrodes enhanced the response of adenosine considering that no response was apparent at the bare surface. Despite this, the electrodes were not sensitive or selective enough to use as routine sensors for adenosine, so the majority of the work focussed on characterization of the electrode surfaces in order to understand and to control the response of the templated OPPy films.

Scatchard and Langmuir isotherm analysis of the calibration data was performed to provide a model for the surface environment. The physical microstructure of the films was characterized by UV spectroscopy and x-ray photoelectron spectroscopy. Electrochemical studies were done to characterize film permeability. These included rotating disk electrode experiments to determine apparent diffusion coefficients of the probes in the films and log peak current vs log scan rate studies, to determine if the electrochemical processes were diffusion or adsorption controlled. Based on the measurements, a simple surface model was proposed.

Carbon fiber electrodes were also modified with OPPy films to enhance the sensitivity to adenine and uric acid and to further characterize the films. The design of the ultramicroelectrode sensors were important for in vivo or cellular use, since the films are necessary to prevent fouling of the surface during the measurements.

Finally, the OPPy films were used to provide a stable background and to suppress the









32

large background currents present in fast scan voltammetry. Fast scan voltammetry was attempted in order to press the limits of detection for uric acid.














CHAPTER 2
EXPERIMENTAL


Reagents and Solutions


HPLC grade acetonitrile (MeCN), certified ACS methanol (MeOH), tetrabutyl ammonium perchlorate (TBAP) and potassium phosphate monobasic were obtained from Fisher. Pyrrole (Py), dopamine, adenine, uric acid, adenosine, adenosine 5'-triphosphate (ATP) and inosine were from Sigma. Potassium ferricyanide (K2Fe(CN)6"), potassium phosphate dibasic and ascorbic acid were from Mallinckrodt. Sodium perchlorate was from Aldrich. Ruthenium hexaamine (Ru(NH3)63") was obtained from Johnson Matthey. All solutions were made with doubly distilled water, and all chemicals were used as received. Pyrrole was purified to obtain pure monomer by passing the monomer solution, which consists of dimers, trimers, etc., over activated alumina, but films formed from the purified pyrrole showed similar sensitivity to those formed from unpurified pyrrole, so pyrrole was used as received. Fresh or purified pyrrole solutions showed shorter polymerization times, but this did not seem to alter the characteristics of the ultrathin films tested here.

Ru(NH3),3+, uric acid and adenine in 0.5 M pH 7.0 potassium phosphate buffer and in 0.5 M KCI were used as probes to characterize the films. Ru(NH3)63+, an electrochemically fast system, undergoes a one electron transfer at ca. E�'=-0.290 V, and the analytical currents were measured at -0.350 V, the cathodic peak, vs SCE (Witkowski et al., 1991). Uric acid









34

(Goyal et al., 1982) and adenine (Dryhust and Elving, 1968) are both slower electron/proton transfer systems. For uric acid, the currents were measured at the oxidation peak potential of ca. 0.3 50 V, and for adenine the currents were measured ca. 1.1 V vs SCE respectively. Except for at the bare electrode, no peaks in the cyclic voltammetric response were seen for adenine, so the currents were measured on the rising portion of the curve where the background was minimal.



Electrodes


Reference and Auxiliary Electrodes


A saturated KC1 calomel electrode (SCE) or 5 cm long Ag wire was used as the reference electrode. Ag was used as a quasi reference electrode when MeCN was the solvent to avoid water contamination and to prevent liquid-liquid junction potentials (Sawyer and Roberts, 1974) and also during electrooxidation of PPy/ATP in the preparation of ATP solutions for analysis by UV to avoid contamination. When a conventional three electrode setup was necessary a 1 cm2 or larger, depending on the the area of the working electrode, platinum foil electrode was used as the auxiliary electrode.



Working Electrodes


Glassy carbon electrodes (GC) were constructed from 3mm and 5mm diameter GC rods obtained from Electrosynthesis. The GC rod was cut into 1 cm in length pieces and each








35

piece was sealed, cut side facing out, into the end of a glass rod using EpoxiPatch epoxy (Dexter Corporation). After drying overnight, the excess epoxy was sanded off the GC surface using 600 grit silicon carbide paper (Fisher). The GC disk was then polished to a mirror finish with Gamal y-alumina/water slurry on a microcloth with an Ecomet 1 polishing wheel (Beuhler), and sonicated in deionized water for one minute. Electrical contact was made to the unpolished side of the electrode using mercury and a piece of copper wire. Finally, the open end of the electrode was sealed with Teflon tape. Before modification, the GC electrodes were polished with the alumina slurry, and then sonicated for one minute. Electrode areas were determined by chronocoulometry by stepping the potential from 0.4 to

-0.1 V with 3 x 10.- M K3Fe(CN)6 in 0.1 MKCl (Do=7.63 x 10' cm2/s; Stackelberg et al., 1953). Typical GC electrode areas were 0.07 cm2.

For rotating disk electrode (RDE) experiments, the RDE GC electrode tips were made by heat pressing a GC rod into a Teflon cylinder. Electrical contact to the GC was made with a Pt wire with silver epoxy (Type 410E, Epoxy Technology, Inc.). The GC RDE areas used were 0.06 and 0.23 cm2. The RDE electrode was polished as described for GC.

A 2 cm2 rough pyrolytic graphite (RPG) (Electrosynthesis) electrode was used for the determination of ATP release from PPy/ATP by UV spectroscopy. RPG was chosen, as opposed to GC, since RPG can be machined easier and can be polished in the same manner as GC to produce a similar surface to GC. The RPG was sealed into a nylon block using EpoxiPatch epoxy and electrical contact was made by a copper wire attatched to the carbon via silver epoxy (EPO-TEK, 40E, Epoxy Technology).

In ESCA experiments a 0.5 mm thick GC disk of 3mm diameter was glued to the









36

ESCA sample holder using silver epoxy. To electrochemically modify GC attatched to the ESCA sample holder, electrical contact to the GC was made using a copper wire coated with silver epoxy, and a water-tight Teflon casing was designed to fit around the sample holder for electrical insulation of the ESCA sample holder as shown in Figure 2-1. Before modification this GC electrode was polished as described above.

Carbon fiber UMEs were made from ca. 7jim radius carbon fiber obtained from Textron specialty materials (Hsueh and Brajter-Toth, 1994). Fibers were glued to a copper wire using silver epoxy and were inserted into a micropipet tip or glass capillary. The copper wire was attatched to the side of the micropipet tip using EpoxiPatch. After letting the epoxy dry overnight, the micropipet tip was back filled with liquid epoxy (Epoxy-Shell EPON Resin 828, hardner-Metaphenylenediamine, both from Miller Stephenson Chemical Co.) and placed in an oven at 150 �C for one hour to cure. To make the epoxy, both resin and hardner were slowly heated until transparent and viscous like water. After curing, the tip was sanded off using 600 grit silicon carbide paper and then Gamal y-alumina/water slurry on a polishing cloth to produce a carbon fiber disk. Prior to electrochemical analysis and modification with OPPy films, the carbon disk was polished with alumina on a polishing cloth like GC and was ultrasonicated for one minute. Electrode radius was verified from the limiting currents obtained using cyclic voltammetry.

























C; 1 1.5mm


9mm


Glassy carbon disk (0.5 mm thick) glued to ESCA sample holder




1.8cm


ESCA sample holder (stainless steel)
6.5 mm


- Teflon insulation


2.6 cm


Figure 2-1 Diagram of ESCA sample holder with GC electrode and Teflon insulation











INSTRUMENTATION



Electrochemical Experiments


A Bioanalytical Systems electrochemical analyzer (BAS-100) was used in all electrochemical experiments except in fast scan measurements at scan rates from 0.005V/s to 0.500 V/s. The electrochemical data were downloaded to an IBM PS/2 Model 50 computer and analyzed using Grapher or Origins commercial programs. For use with UMEs a homemade current amplifier based on Faulkner's design (Huang et al., 1986), which allowed picoampere currents to be measured, was connected to the BAS and properly grounded (Hsueh and Brajter-Toth, in press).

The input operational amplifier (AD515 A, Analog Devices) of the current transducer was a monolithic precision, low power, FET-input operational amplifier. The AD515A functioned as a current-to-voltage converter which amplified and converted the input currents to voltages. Various resistors (1, 10 and 100 M(Q) and capacitors (1, 10 and 100 pF) in the feedback loop controlled the gains (100, 1000 and 10000) and RC time constants (1, 10 and 100 lis). Because the current-to-voltage converter inverted the signal phases, a second operational amplifier (OP27) with a unit gain was used as an inverter to invert the phase of the signal back to normal. Capacitors and a resistor on OP27 functioned as a first order filter to minimize noise in the circuit.

The minimum current measurable by the BAS was 0.1 ptA, but the gain of the current amplifier ranged from 100 to 10,000, so the BAS with the preamplifier could measure









39

currents as low as 100 picoamperes. The time constant of the potentiostat was controlled by the time constant of the first order filter, which was used to minimize electrical noise in the preamplifier circuit. At low scan rates, i.e. less than 40 V/s, the time constant of the system was set at 100 gas which allowed use of scan rates up to 40 V/s for one electron reactions, 20 V/s for two electron reaction, etc., with negligible distortion in the separation of peaks (Howell et al., 1986; Wipfet al., 1988).

For fast scan experiments with scan rates from 100 V/s to 10000 V/s, a potential waveform from a function generator (EG&G Parc Model 175 Universal Programmer) was applied to a SCE reference electrode of a two electrode cell configuration, and the waveform was recorded at one channel of a digital oscilloscope (LeCroy 9310). A two electrode configuration was used since the current measured at the working electrode was small. Current at the working electrode was transduced to voltage by the current transducer of the preamplifier, amplified by OP AD515A and measured directly with the oscilloscope. The output of OP AD515A bypassed the OP27 and was directly connected to the oscilloscope. The oscilloscope measured the output voltage (the transduced current). The ratio of the transduced current to the input current of the working UME was determined by the feedback resistance of OP AD515A. The conversion factors for the potentiostat were 1, 10 and 100 V/tA when the resistors in the amplifier were 1, 10 and 100 MD respectively. The RC time constant of the current transducer was controlled by the feedback resistance and the capacitance on OP AD515A. A 10 ps time constant would allow a scan rate of 400 V/s for one electron reactions with little distortion in the voltammogram, and a 1 gs time constant would permit a scan rate of 4000 V/s (Hsueh and Brajter-Toth, in press).








40

Stored waveforms were transferred to a computer via a simple BASIC program (Hsueh and Brajter-Toth, submitted) for plotting cyclic voltammograms using Origins. The phase of the potential at the working electrode was reversed relative to the waveform potential at the reference electrode by the potentiostat (Hsueh and Brajter-Toth, in press), so the phase of the potential waveform measured with the oscilloscope had to be inverted. Also, the phase of the current at the working electrode was inverted by the inverting input of OP AD515A. To obtain a standard cyclic voltammogram, the phase of the potential waveform and the current stored in the digital oscilloscope were inverted using Origins before being plotted.

Since the large charging current could obscure the faradic current at high scan rates, background subtraction had to be performed. For background subtraction, the background was measured in the absence of analyte in the supporting electrolyte alone, by cycling the electrode, and stored, and then the stored background was subtracted from the analyte current after completion of the analytical measurements. Signal averaging of 250 scans was used in both the background and analyte measurements before background subtraction. Generally, 250 scans for both the background and the analyte response was sufficient to obtain a good cyclic voltammogram. Other combinations such as a greater number of scans for the background were attempted, but these did not produce better results. Larger numbers of scans produced poorer resolution since more memory was taken up in the oscilloscope. In order to obtain accurate measurements, solutions were injected with a syringe into a microliter electrochemical cell, which allowed solutions to be pumped into the cell without moving the electrodes (Hsueh and Brajter-Toth, 1993).











UV Absorption and ESCA Experiments


UV absorption measurements were made with a Hewlett-Packard 8450A UV/Vis spectrophotometer. Matched quartz cells with a 1 cm path length were used.

ESCA experiments were performed with a Kratos Xsam 800 spectrometer using AR, excitation. Spectra were recorded for a survey scan with an energy window of 1100 eV, various core levels with energy window of 20 eV or 40 eV, and the valence level with the energy window of 50 eV, using low magnification and high resolution, high magnification and low resolution, and high magnification and high resolution. The sample analyzer chamber pressure was kept at less than 1 x 10.' Torr.



Fundamentals of Electrochemical Measurements Cyclic Voltammetry


In cyclic voltammetry (CV) experiments, the potential is scanned linearly from an initial potential, where typically no Faradaic reaction of the analyte occurs, to a final potential where the reaction rate is limited by diffusion. The potential is then scanned linearly back to the starting potential. The rate of potential change is the scan rate, v (V/s), and the potential range between the initial and the final potential is the potential window, which depends upon the electrochemical properties of the analyte, the electrode and the solvent/electrolyte. Cyclic voltammograms are plotted as current vs potential. The current stemming from the redox reaction gradually increases past the potential where the reaction starts and reaches a








42
maximum, called the peak potential (Ep). After reaching this potential, the current gradually decreases as the analyte diffusion controls the response of the electrode. The potential difference between the reduction and oxidation peaks (E, and E. respectively) is known as AEp.

The value of AE, can be used as an indicator of the reversibility of the electrode reaction. A one electron reaction at 25 'C is considered reversible if AEp is ca. 0.059 V; quasi-reversible if AEp is between 0.060 to 0.212 V and irreversible if AEp is greater than

0.212 V (Bard and Faulkner, 1980).

The theoretical peak current,i (A), for a diffusion controlled, reversible reaction can be written as follows (Bard and Faulkner, 1980): i =(2.69x105)n 312AD /2v112C (2.1)


and for an irreversible reaction:


1, =(2.99x10 5)n(on)12AD /2v "2C (2.2)


where n is the number of electrons transferred per mole, A is the electrode area (cm2), D. is the diffusion coefficient (cmE/s), v is the scan rate (V/s), Co" is the bulk concentration of the analyte (mol/cm3), a is the transfer coefficient and n. is the number of electrons in the rate determining step. For adsorption controlled reactions, the peak current for a reversible reaction can be written as:


n 2F2AvP (2.3) 4RT










and for an irreversible reaction:


i n -n FAvro (2.4) P 2.718RT


where r. is the surface excess of the analyte (mol/cm2), R is the gas constant (J mol-'K'), T is the temperature (K), F is Faraday's constant (C), and the remaining variables are the same as for the difflusional controlled processes.

Cyclic voltammograms at macro electrodes (mm diameter) are typically peak-shaped, and the peak currents are proportional to the square root of the scan rate for a diffusion controlled system, and proportional to the scan rate for adsorption controlled reactions (equations 2.1 and 2.2). For UMEs, the cyclic voltammograms are sigmoidal in shape and the limiting current measured at the plateau is independent of scan rate for scan rates up to I V/s due to radial diffusion. Under these conditions, the radius of an UME is small compared to the thickness of the diffusion layer. Because of the radial diffusion, steady state mass transport is attained at the electrode surface, and hence the current is time independent. As a result, the scan rate does not affect the shape and the size of the voltammetric wave. For a disk UME, the limiting current at steady state can be expressed by (Heinze, 1993): i =4nFDC * r (2.5)


where n is the number of electrons transferred per mole, F is Faraday's constant (C), D is the diffusion coefficient (cm2/s), C is the bulk concentration of the analyte (mol/cm3) and r is the electrode radius (cm).









44

The plateau potential, which is the potential where the limiting current is reached, and the E,,2, the half-wave potential value, can be used to judge the difficulty or ease of electron transfer. For an oxidation process, a plateau and E1V2 potential significantly more positive of E�1 would indicate difficulty in electron transfer. As a note, in a reaction with very slow kinetics a well defined plateau of current is not observed.

Using membrane coated UMEs, information about membrane structure can also be obtained since diffusion in solution should not the limit the response at time scales where diffusion layer thickness is much greater than the electrode radius. Hence, the response is determined by transport and other processes in the film (Cheng and Brajter-Toth, 1992) and can give information about the film microstructure/microenvironment.

As mentioned above, the signoidal shape current-potential curves for UMEs only exist at scan rates less than 1 V/s for ca. 5 pm radius (Heinze, 1993). At higher scan rates, this shape changes to peaks as with the macro electrodes, and at scan rates above 100 V/s, the shape is the same as with macro electrodes. Thus, the peak current equations for the macroelectrodes can be applied to the UMEs at high scan rates.



Chronocoulometry


In chronocoulometry (CC) the potential is stepped from a potential where the rate of the redox reaction is negligible to a potential where the reaction rate is diffusion limited. The charge passed is monitored as a function of time (Q(t)) and is expressed as (Bard and Faulkner, 1980):












2nFAD 2C *t (2.6) 1E 1/2



where n is the number of electrons per mole transferred, F is Faraday's constant, A is the electrode area (cm2), D is the diffusion coefficient (cm2/s), C" is the bulk concentration of the analyte (mol/cm3), and t is the pulse width (s).

According to equation 2.6, a plot of Q(t) vs t'2 should give a straight line with a slope of 2nFAD/ 'CY/"2. The value of the slope is used to determine electrode areas or diffusion coefficients. In this work, CC was used to find electrode areas of the macroelectrodes and to control the amount of pyrrole polymerized on the electrode surface by regulating the charge during polymerization.



Rotating Disk Electrode Experiments


Rotating disk electrodes (RDE) can be used to calculate the apparent diffusion coefficients of probes in films. The diflusion limiting current (id) is proportional to the square root of the rotation rate, 6) (sl) (Gough and Leypoldt, 1979) and a Koutecky-Levich plot of id"1 vs w -12 gives an intercept of id' at infinite rotation rate (Gough and Leypoldt, 1979). At infinite rotation rate, the diffusion of the probe in solution becomes negligible and id depends only on the diffusion of the probe through the film (Gough and Leypoldt, 1979, 1980; Leddy et al., 1985). For a membrane covered RDE, the membrane current, id, can be written as:











D
Id=nFAC *P,=nFAC * app (2.7)




where n is the number of electrons, F is Faraday's constant, A is the electrode area (cm2), C" is the bulk concentration of the probe (M), Pm is the permeability of the film (cm/s), DP is the apparent diffusion coefficient of the probe in the film and 6. is the thickness of the film

(cm). The Dapp values can be obtained from this equation since id can be found from a Koutecky-Levich plot of the data, and the rest of the variables are typically known. This method can only be used to determine DP values for probes at macroelectrodes since RDE UME electrodes are difficult to fabricate, and the hydrodynamic processes at the small electrodes are difficult to control.














CHAPTER 3
PREPARATION AND CHARACTERIZATION OF OPPy AND TEMPLATED OPPy ELECTRODES


Procedure for Preparing Ultrathin OPPy Films by Polymerization and Overoxidation of Polypyrrole on GC


Preparation procedure for ultrathin OPPy films was based on a technique developed by Hsueh (Hsueh and Brajter-Toth, 1994). In this work, 0.020 M pyrrole was polymerized on GC from MeCN with 0.1 M TBAP at 0.950 V vs Ag wire. Hsueh's procedure was modified in order to control the deposition charge at 35 gC/cm2 which corresponds to a monolayer surface coverage (monolayer surface coverage is defined as 0.15 nmol/cm2 to 0.3 nmol/cm2 (Murray, 1992)), based upon 2.25 electrons involved in the polymerization process (Diaz and Castillo, 1980).

Polymerization was initiated by chronocoulometry, with a potential step from 0.650 to 0.900 V vs Ag wire (Bull et al., 1982). Typical polymerization times were ca. 50 ms. Polypyrrole was then overoxidized from 0.5 M potassium phosphate buffer of pH 7.0 at 0.950 V vs SCE as described by Hsueh (Hsueh and Brajter-Toth, 1994). Typical time for overoxidation was ca. five minutes. Formation of pinhole-free ultrathin OPPy films was performed using the procedure developed by Hsueh (Hsueh and Brajter-Toth, 1994). To form a pinhole-free ultrathin film, polypyrrole was first polymerized and then overoxidized. Since the overoxidized film was nonelectroactive, this process could be repeated, and only








48

the pinholes and gaps in the film would be filled in by subsequent polymer deposition steps. This process was repeated until the response of 0.010 M Fe(CN)63- in 0.1 M KCI or in pH 7.0 0.5 M potassium phosphate buffer was suppressed to background level after four to seven coatings as shown in Figure 3-1. The choice of the electrolyte for Fe(CN)63" detection did not influence the response of the OPPy film electrodes.

The thickness of the films which suppressed Fe(CN)63- response and which were formed by repeating the polymerization/overoxidation procedure was roughly calculated from the total charge required to form the pinhole-free film, on average four times 35 A C/cm2, and from the diameter of pyrrole (ca. 4A*), estimated from the bond lengths and the bond angles (Hsueh and Brajter-Toth, 1994). Since ca. four multilayers of polypyrrole (each layer 4A thick) were deposited, the films that were used in the majority of the experiments with the GC electrodes were ca. 16A thick.

As mentioned in Chapter 2, pyrrole was used as received without any further purification. Pyrrole was purified to obtain only monomer solution by passing the solution over activated alumina. This purified pyrrole was then used to make an OPPy film electrode in order to compare the effect of purified and unpurified pyrrole on the film characteristics. The sensitivities of OPPy film electrodes made from unpurified pyrrole were similar to those formed from purified pyrrole. However, the time required to deposit charge when purified pyrrole was used was considerably shorter, especially when thick (> 16,A) films were being prepared. Otherwise, no difference was observed in the investigated properties of the ultrathin films with the use of the purified or unpurified pyrrole.


















60 bare GC
/ ' 1stcoating /coating / N 3rd coating
II
40
4th coating
I/
5th coating
S20 / / 6th coaling


7th coating

S0




-20/
N / N I



-4 I I I
400 300 200 100 0

potential (mV)



Figure 3-1 Suppression of 10 mM Fe(CN)63" response in 0.5 M pH 7.0 potassium
phosphate buffer after repeated coatings of OPPy; scan rate 0.100 V/s,
electrode area 0.07 cm2, deposition charge per coating ca. 35 A C/cm2, film
thickness after 7th coating ca. 28 A.








50

Procedure for Preparing Templated OPPy Films on GC by Polymerization and Overoxidation


OPPy was templated with adenosine (OPPy/ado), inosine (OPPy/ino) and ATP (OPPy/ATP) by polymerizing 3 x 10' M polypyrrole (PPy) at GC from 80% MeOH, 0.1 M TBAP with 3 x 10' M adenosine or inosine to form first PPy/ado and PPy/ino respectively. To obtain PPy/ATP, 3 x I0 M pyrrole was electropolymerized from water with 0.0 10 M ATP. MeOHlwater solutions and water were used in the formation of the templated polymers to accomodate the solubility of the templates. The scheme proposed for incorporation of the templates is illustrated in Figure 3-2. Note that the exact orientation of the pyrrole monomers is unknown, but Nishizawa et al. found that lateral growth was promoted at hydrophobic substrates (Nishizawa et al., 1991). As shown in Figure 3-2, adenosine can be incorporated during the polymerization of pyrrole by weak interactions with the forming polypyrrole, such as hydrogen bonding or hydrophobic interactions. In contrast, ATP can undergo stronger interactions such as electrostatic as well as hydrogen bonding and hydrophobic interactions with the forming polymer. Typical dopant levels for polypyrrole are one counter ion for every three to four pyrrole units (Mitchell et al., 1988). Larger concentrations than the typical dopant levels were used to ensure incorporation of adenosine and inosine because of their neutrality, and to also use ATP as the supporting electrolyte.

A polymerization potential of 0.400 V vs SCE as suggested by Ko et al. for aqueous solutions (Ko et al., 1990) was applied to deliver a charge of 35 tC/cmz in the formation of all templated polypyrrole structures in all mixed/aqueous solvents. Polypyrrole was overoxidized to form OPPy/ado, OPPy/ino and OPPy/ATP at 0.950 V vs SCE, as were the



























0 0 0
II II II P-P-P
I - I - I 0 0 0


'N
I
ribose


Adenosine


Figure 3-2 Cartoon representation of template, adenosine or ATP, incorporation into
polypyrrole


H, ,H
N


A TP
/








52

pyrrole films without the templates, with ca. five minutes required for the current to reach steady state. These films overoxidized in approximately the same time as those formed without the templates. During overoxidation, the template molecules were expected to be expelled from the oxidized polypyrrole due to the removal of the net positive charge of the polymer, requiring expulsion of the charge balancing ions, and due to the resulting unfavorable interactions with the carbonyl groups in the newly forming OPPy structure. The newly formed OPPy structure without the templates is depicted in Figure 3-3. The polymerization/overoxidation of the templated polymer was typically repeated four to seven times until the response of 0.010 M Fe(CN)63" in 0.5 M pH 7.0 potassium phosphate buffer was suppressed to a background level. No apparent difference in the suppression of Fe(CN)63, in comparison to the films formed without the templates, was noted. However, in some instances with the templated films, the first coating seemed to suppress Fe(CN)63" response more. In addition, polypyrrole was polymerized/overoxidized at GC without the templates but in the same solvents as used during the formation of the templated polypyrrole to assess the effect of the solvent on film formation and properties. In 80% MeOH, 3 x 10-3M pyrrole was polymerized with 0.1 M TBAP; in water 3 x 10- M pyrrole and 0.010 M NaCIO4 was used. These OPPy films were prepared using the polymerization and overoxidation conditions described above for the templated films.

In order to analyze the effect of film thickness on response and interpret the data from macroelectrodes compared to the data obtained at carbon fiber UMEs, films with thicknesses of 32, 44 and 48 A were prepared at GC to investigate the effect of film thickness on sensitivity. For a 32 A film a charge of 84gjC/cm2was applied per coating, and for the 44 and





















































Figure 3-3 Cartoon representation of the templated OPPy structure after overoxidation of
polypyrrole (Beck et al., 1987)









54

and 48A films charges of 124 and 133 j.tC/cm2 were applied per coating respectively. Polymerization times generally required ca. 15 to 20s which was much longer than the millisecond time required for the ultrathin films. The thicker polypyrrole films were overoxidized in 0.5 M pH 7.0 potassium phosphate buffer at 0.950 V vs SCE until the current reached a steady state, ca. ten minutes. These thicker films required approximately twice longer overoxidation time than the ca. 16A films. Polymerization and overoxidation steps were repeated as described previously to fill in the pinholes in the films. Typically three coatings were needed to suppress the response of 10 mM Fe(CN)63- in 0.5 M pH 7.0 potassium phosphate buffer.



Procedure for Preparing Ultrathin OPPy and Templated OPPy films at UMEs


Preparation of OPPy films at carbon fiber UMEs was based on the work of Hsueh (Hsueh and Brajter-Toth, 1994). The same conditions as used for GC electrodes were used for the UMEs, except the charge applied was 31.7 mC/cm2 per coating (three coatings on average) which corresponds to a thickness of ca. 0.53 gim, based on a charge of 24 mC/cm2 resulting in a 0.1 tm thick film (Diaz and Castillo, 1980). Hsueh's films were ca. 32 A thick, but this thickness was difficult to duplicate because very short times were required for the deposition due to the high current density at the UMEs from edge effect. Polymerization generally required three milliseconds for each coating, and overoxidation required less than one minute, as compared to ca. fifty milliseconds for polymerization and ca. five minutes for overoxidation at GC. Suppression of the response of 0.010 M Fe(CN)63- in 0.5 M pH 7.0








55

potassium phosphate buffer was typically achieved after three to four coatings as shown in Figure 3-4. This was shorter compared to the films formed at the macoelectrodes since up to seven coatings could have been used, possibly due to the thicker films at the UMEs.



Verification of Template Release During Overoxidation of Templated Polypyrrole


A calibration plot of absorbance vs concentration was prepared in the concentration range 5.0 x 10' to 1.0 x 10- M ATP to measure the ATP release from polypyrrole grown on RPG, into 0.1M pH 7.0 potassium phosphate buffer. Potassium phosphate buffer concentration of 0.1 M was used because the use of 0.5 M pH 7.0 potassium phosphate buffer produced a high background in the UV absorption spectra. The molar absorptivity of 1.41 x 10 M cm"1 of ATP was determined at 1., =260 nm in agreement with the literature value of E=1.54 x 104 Mcnfm" (Pyo et al., 1994). To measure the amount of ATP released from the PPy templated with ATP, a 2.0 cm2 square RPG electrode encased in nylon and sealed with Epoxipatch epoxy was used as the substrate for the deposition of the ATP templated polypyrrole film. RPG was used because a large electrode area was needed so that a sufficient amount of ATP could be incorporated into polypyrrole in a reasonable amount of time, and RPG, if not roughened, has a similar structure to GC. The RPG electrode was polished with alumina on a polishing cloth and sonicated for one minute prior to the polymerization as decribed for GC in Chapter 2. A 0.020 M ATP solution with 0.007 M pyrrole in water was used as the polymerization solution. The surface coverage of polypyrrole was 5 x 10.' mol/cm2 as determined from the deposition charge (0.108 C/cm2), which



















bare carbon fiber


2 A


coating 1
coating 2

coating 3
coating 4




S I I I 1I
400 300 200 100 0

potential (m V)



Figure 3-4 Suppression of 0.010 M Fe(CN)63" in 0.5 M pH 7.0 potassium phosphate
buffer after repeated coatings of OPPy at carbon fiber LME; radius 7 gm, scan rate 0. 100 V/s, deposition charge 3.17 x 10-2 C/cm2 per coating, film
thickness after 4th coating ca. 0.53 i








57

corresponded to a film thickness of ca. 0.45 ptm based on 24 mC/cm2 resulting in a 0.1 prm thick film (Diaz and Castillo, 1980). This thickness required a polymerization time of ca. ten minutes which was achieved by bulk electrolysis at 0.900 V vs SCE. To determine the amount of ATP incorporated during the polymerization of polypyrrole and released during the overoxidation of the film, polypyrrole was overoxidized in 0.1 M pH 7.0 potassium phosphate buffer at 0.950 V vs SCE until the current decayed to a steady state value, ca. thirty minutes.

To verify template incorporation into polypyrrole and template release (Li and Dong, 1992) during polypyrrole overoxidation, ATP template concentration was monitored at X.=260 nm in the overoxidation solution. A thicker than normal PPy/ATP film (0.45 jgm) was formed to allow detection of ATP above the micromolar limit of detection for ATP set by the UV method. The amount of ATP that was detected in the overoxidation solution corresponded to a molar ratio of 7.7:1 of pyrrole to ATP, with the expected ratio of 9:1, based on +0.33 charge/pyrrole unit in polypyrrole (Beck et al., 1987) and a -3 charge on ATP. The 7.7:1 molar ratio confirmed ATP incorporation into polypyrrole and the release of the template during the overoxidation of polypyrrole to OPPy in nearly a stoichiometric amount. Since the polymerization solution consisting of ATP, pyrrole and water had pH 3.0, and ATP has pIK of 4.1 (H2ATP2" H+ + HATP3") and 6.95 (HATP3" - H+ + ATP4-), then at pH 3 ca. 9 % ATP should have a charge of-3 (Zubay, 1988). Therefore differences in the experimental and the theoretical value of the ratio of pyrrole to ATP may be due to the greater presence of H2ATP2", which would theoretically give a molar ratio of 6:1.












ESCA Analysis of Bare GC. OPPy and OPPy/ATP Films on GC


Initial ESCA spectra were taken of the GC surface alone in order to obtain a background level of carbon and oxygen atoms and their ratios present before polymerization. GC was polished with alumina on a polishing cloth and sonicated for one minute before the spectra were run. Spectra were recorded to obtain a wide scan spectrum or a survey scan (binding energy window 0 to 1100 eV) as illustrated in Figure 3-6, various core energy levels for different atoms (binding energy window 20 eV or 40 eV) as shown in Figures 3-9, 3-10, 3-12 and 3-13 and a spectrum of valence levels (binding energy window 0 to 50 eV) as pictured in Figure 3-5 using low magnification and high resolution, high magnification and low resolution, and high magnification and high resolution. The spectra of the bare GC indicated the presence of C1,, O,, Ag3d, and Si,., 2P. The silicon present was due to silica; silica has been observed at other types of carbon surfaces, such as carbon foil and carbon molecular seives (Goodfellow, 1990). Silicon most likely segregated to the surface from heating or ion bombardment during ESCA analysis, or normal aging of the material. From the survey scan of the bare GC surface, the C/O area ratio was 1.5. After the electrode was etched with Ar' (4 keV, 1 pA) for two hours, the C/O area ratio was 1.7. Therefore, the etching removed some of the oxygenated carbon from the surface even though the ratios did not change significantly. This also showed that GC could possibly be used as a substrate in ESCA analysis of OPPy films since the C/O area ratios remained fairly constant at the carbon surface with etching.









59

Next, the GC electrode was modified with ultrathin OPPy by first polishing with alumina on a polishing cloth, sonicating for one minute, and then polymerizing 0.020M pyrrole from MeCN with 0.1 M TBAP at 0.950 vs Ag wire. The polypyrrole electrode was then overoxidized from 0.5 M pH 7.0 potassium phosphate buffer at 0.950V vs SCE. This process was repeated until the response of 0.010 M Fe(CN)63" in 0.5 M potassium phosphate buffer pH 7.0 was not suppressed any further. For the electrode studied by ESCA, three polymerization/overoxidation steps were needed. Fe(CN)63 response was not suppressed to the background level, but the fourth coating did not suppress the response any further than the third coating did, so the process was not repeated. The behavior was similar to that observed in other experiments with GC electrodes. After coating with the OPPy film, the electrode was rinsed with deionized water and stored in vacuum until ESCA analysis was performed. The film thickness was ca. 12 A as calculated from the deposition charge of 35 pC/cm2 per coating (3 coatings applied) which corresponds to a monolayer coverage, 2.25 as the number of electrons involved in the process and 4 A as the diameter of pyrrole.

As shown in Figure 3-5, the ESCA valence band spectra for bare GC and OPPy modified GC are significantly different, indicating variations in the chemistry of the two surfaces. A wide scan ESCA spectrum taken on GC covered with ultrathin OPPy (film thickness ca. 12 A) as shown in Figure 3-6, clearly illustrates the presence of oxygen, carbon and nitrogen. Carbon and oxygen were expected since they are present at the GC surface, but the presence of nitrogen indicated the presence of pyrrole or TBAP on the surface. By comparing the C/Si area ratios for bare GC and OPPy, it was concluded that ca. 13 % of the electrode was not covered with the film. This was reasonable since more hydrophilic regions

























OPPy modified GC electrode


40 30 20 10 0 Binding Energy (eV)


Figure 3-5 ESCA valence band seectra of bare and OPPy modified GC electrodes
film thickness ca. 12 A












61







































80000 01
Si CII

a
60000




a 40000 NIs C0



20000




0
1000 800 B00 400 200 0 Binding Energy (eV)


























Figure 3-6 Wide scan spectrum of OPPy film on GC (ca. 12A thick) showing the presence of oxygen, carbon and nitrogen









62

on GC may not support the growth of polypyrrole, and in constructing the polymer only three repeated processes of polymerization and overoxidation were performed which did not completely suppress Fe(CN)6". In general, the film coated regions charged more, as expected for a non-conductor, than the bare carbon which was the region that charged the least.

For identification of the film on the electrode surface, the types of nitrogen on the surface were of interest. The nitrogen spectra showed charge corrected peaks at 397.8 eV and 400.7 eV. According to Pfluger et al. and Ge et al., who studied OPPy films, the peak at 397.8 eV was due to =N- from the OPPy structure as shown in Figure 3-3 (Pfluger et al., 1983; Ge et al., 1994). Although no polymer structures were proposed by either group, the presence of this nitrogen was consistent with the structures proposed by Beck as shown in Figure 3-7 (Beck et al., 1987).The peak at 400.7 eV could be a nitrogen with a partial positive charge as suggested by Pfluger et al. (Pfluger et al., 1983; Pfluger and Street, 1984). Ge et al. attributed this peak to -NH on polypyrrole as shown in Figure 3-8 (Ge et al., 1994).

Other peaks in the ESCA spectra were from carbon and oxygen as shown in Figure 3-6, phosphorous, chloride, silicon, aluminum ( Figures 3-9 and 3-10), and potassium. The presence of silica from the low charging region was attributed to exposed carbon from defects in the film, and aluminum was identified as a contaminate from the polishing of GC with alumina. The presence of phosphorous and potassium was likely due to H2PO4" and HP042 from the potassium phosphate buffer used to overoxidize polypyrrole. The presence of HP042 was speculated to produce the low binding energy phosphorous component, and the presence of H2P04",which was present in a larger quantity, was believed to be the high binding energy phosphorous component. Based on the C/P area ratios of the 2s (area ratio
































H0 0 HH

N A&


Figure 3-7 Structure of overoxidized polypyrrole as proposed by Beck et al. (1987)





















OH


Figure 3-8 Structure of OPPy as proposed by Ge et al. (1994)













































25000,

C1


20000


P

0 15000.
0




1 30000





5000.




0

220 210 200 190 180
Binding Energy (eV)

























Figure 3-9 ESCA spectrum of chloride and phosphorous from OPPy film (ca. 12 A)

on GC













66


































P ZS 15000

M Si Zp




10000



a


- 5000






0

150 140 130 120 110 tOo
inding Energy (eV)


























Figure 3-10 ESCA spectrum of phosphorous, aluminum and silicon of OPPy film

(ca. 12 A thick) on GC









67

is 2.8) and 2p (area ratio is 2.3) spectra of phosphorous and the Ka value of H2P04- (6.32 x 10"1; Harris, 1991), the pH in the film was calculated to be ca. 6.8 as compared to the measured overoxidation solution pH of 7.0, so a reasonable amount of the ions from the phosphate buffer must have been trapped in the film. During overoxidation of pyrrole there is a movement of counterions due to the loss of charge on the polymer and the formation of carbonyl groups. Counterions like H2PO4" or BPO42- could be trapped in the film. The movement of ions during overoxidation is illustrated in Figure 3-11. Figure 3-11 shows the oxidation mechanism of polypyrrole to overoxidized polypyrrole in aqueous solution as proposed by Beck et al.

Since -NH' could be present in the film as postulated by Ge et al. (Ge et al., 1994), the detected presence of the phosphate ions may be for charge balancing purposes. Cations in the films could also exist to balance the charge of any anions in the film. Chloride found in the films could be from the TBAP. The C/K area ratio was 83, and the C/Cl area ratio was 34, indicating more chloride was present in the film. Cl- may have served to balance charge in the film from the presence of -NH' groups. Hence, these results confirmed that nitrogen containing films were present on the GC surface from the nitrogen peaks and a new C/O ratio of 4.5 compared to 1.5 at bare GC. If an OPPy film was present at the surface, the film must be porous due to the presence of several types of ions. Finally the polymerization/overoxidation process produced a well covered surface nearly free of pinholes, pointing to a polymer film rather than a TBAP covered surface.

In order to investigate the effect of templating on the surface microenvironment, GC was modified with ATP templated OPPy using the procedure described previously in this











+2A
(1) +A- +

H H conducting polypyrrole


OH
+ 2A- H20
(2)+
N -2A- N H H



OH 0

(3) 14 AIZ N N H H





-2e
(4) -2H N / N H overoxidized polypyrrole

Figure 3-11 Mechanism of polypyrrole overoxidation in water (Beck et al., 1987)









69

chapter. Polymerization from a 0.003 M pyrrole/0.010 M ATP solution was performed at 0.400 V vs SCE. Polypyrrole was then overoxidized at 0.950 V vs SCE in 0.5 M pH 7.0 potassium phosphate buffer. Four coatings were needed to suppress the response of 0.010 M Fe(CN)63" in 0.500 M pH 7.0 potassium phosphate buffer. The OPPy/ATP film was ca. 16 A thick. The electrode was rinsed with deionized water and dried before ESCA analysis.

The OPPy/ATP film also showed differential charging (different binding energies for one atom) similar to the OPPy film, as indicated by two sets of silver 3d peaks. No silver was observed in the least charging region. Ag was likely from the silver epoxy used to attatch the GC to the holder before film modification. The OPPy film had three different types of regions indicated from three sets of silver 3d peaks. The area ratio for C/Si (Si spectrum shown in Figure 3-12) indicated that ca. 10% of the electrode was not covered with the film since the C/Si area ratio for film covered and bare GC were 1.2 and 9.4 respectively. The percentages of uncovered GC was similar to that for the nontemplated OPPy film, indicating that the ATP template or somewhat thicker film did not significantly affect coverage.

The ESCA valence band spectrum for OPPy/ATP illustrated in Figure 3-13 showed differences and similarities to the valence band spectra for the OPPy film shown in Figure 3-5. Table 3-1 summarizes the area ratios of carbon to selected atoms for OPPy and OPPy/ATP electrodes from obtained spectra. The differences in the area ratios indicated that the surface microenvironments were significantly different. The C/N area ratio for OPPy/ATP was 190 in comparison to 33 for OPPy, which indicated fewer nitrogens on the surface. Hence, polymerization with the template altered the film structure as observed from ESCA analysis.






















































-; 10000 r ao Si Zp







5 5000








01

160 140 120 100 Bindi S Energy (eV)


























Figure 3-12 ESCA spectrum of aluminum and silicon in OPPy/ATP films (ca. 16A

thick) on GC























































a 30000.

5
a


20000





10000





0

40 30 20 to 0 Binding Energy (eV)



























Figure 3-13 ESCA valence band spectrum of OPPy/ATP film (ca. 16A thick) on GC












Table 3-1 Comparison of Various Atom Area Ratios obtained from ESCA spectra for Bare
and OPPy and OPPy/ATP modified GC electrodes



area ratio observed binding energy bare GC OPPy/GC OPPy/ATP/GC
C/X (eV) electrode electrode electrode
b
C/N 405.9 + 408.8 33 190
b
C/P 195.0 101 79
(HPO42)
b
C/P 200.0 36 101
(H2P04)
b b
C/K 301.0 83
b b
C/Ca 356.9 147
b
C/Cl 207.4 34 49
b
C/Al 76.0 41 41

C/Si 111.2 9.4 56 57

C/O 539+541 1.5 4.5 5.5

Note: area normalized for time of scan and size of window,
C/X = (area C/o)/(area X/o), where a is the Scofield cross section (Scofield,
1976)

'electrode area 0.07 cm2, film thickness ca. 12 k at OPPy and ca. 16 A at OPPy/ATP


bundetected









73

Phosphorous was present in the OPPy/ATP film as shown in Figure 3-14, but more HP042" (area C/P=-79) than H2PO4" (area C/P=101) was present than in the OPPy film, where significantly more H2P04" (area C/P=36) was found. Potassium was not detected in the OPPy/ATP film, but small amounts of calcium (area C/Ca= 147), shown in Figure 3-15, was present. Calcium impurity originated from the ATP since it was not present in the OPPy film, and more than likely, calcium replaced the K' because Ca2+ can more effectively balance the charge on HP042. Since Ca2+ was from the ATP, it could be postulated that ATP was present initially in the film. Carbon, oxygen and nitrogen ratios did not indicate that ATP was present after oxidation. The OPPy/ATP film also contained chloride, probably from the SCE, but the OPPy/ATP film had less Cl" than OPPy which was expected since CI" was not present during the polymerization. Based on the results, the OPPy/ATP film was concluded to have a different microenvironment than the OPPy film. The microstructure of the OPPy/ATP film may be more compact since fewer ions were found at the surface. However, this could also be a result of less polypyrrole on the surface due to the initial presence of ATP ions in spite of initially greater thickness. ATP ions are much larger than pyrrole and could act as spacers in the structure during the polymerization. Since there was less polypyrrole, as attributed from lower amounts of N atoms, ATP must have altered the structure of the film at the GC electrode surface as expected when incorporated into the ultrathin film.













74


































C
25000 " 20000

a

15000 '~10000 5000





220 210 200 190 180 Binding Energy (eV)


























Figure 3-14 ESCA spectrum of chloride and phosphorous in OPPy/ATP film on C

(ca. 16A thick)












































































370 365 360 355 Biading Energy (eV)


Figure 3-15 ESCA spectrum of calcium in OPPy/ATP film on GC (ca. 16A thick)

















CHAPTER 4
ANALYSIS OF THE ELECTROCHEMICAL RESPONSE AT GLASSY CARBON, GLASSY CARBON COATED WITH OPPy AND TEMPLATED OPPy


Sensitivity Data at Glassy Carbon and Glassy Carbon Coated with Ultrathin Films of OPPy and Templated OPPy


This chapter will describe the properties of ultrathin OPPy membranes templated with adenosine, inosine and ATP. Ru(NH3),3, uric acid and adenine were used to characterize the voltammetric response of the films (ca. 16A) and were chosen based on their electrochemical reactivity, charge and ability to undergo hydrophilic and hydrophobic interactions with GC and the membrane. Based on the results of Witkowski et al., thick OPPy films grown on GC were expected to be relatively compact (Witkowski et al., 1991). Ultrathin OPPy films were prepared here to detect changes in the film interactions from templating with good sensitivity.

Good sensitivity was expected for Ru(NH3)63 at OPPy coated GC due to the formation of carbonyl groups in the OPPy structure during polypyrrole overoxidation. These carbonyl groups have a high electron density which should favor their interactions with cations. In contrast, uric acid should have lower sensitivity at OPPy membrane electrodes because of the negative charge density of the carbonyl groups, especially at pH 7.0 when uric acid is negatively charged (pK,=5.4, 11.3) (Brown, 1991). The high electron density from the carbonyl groups on OPPy has been shown to exclude anions as with the response of









77

Fe(CN)6' shown in Figure 3-1. Uric acid may not be excluded as well as Fe(CN)63- since the charge on uric acid is smaller and delocalized about the structure, and uric acid is capable of hydrophobic interactions with the OPPy backbone. With -NW groups in the film, as possibly indicated from ESCA analysis, uric acid may show higher sensitivity than expected if the OPPy structure consisted only of repelling carbonyl groups. Reasonable sensitivity at OPPy membrane electrodes was expected for neutral adenine (pKa=2.0, 4.1, 9.8) (Brown, 1991) which can interact hydrophobically with the OPPy backbone.

The cylic voltammetric reponse of each probe at GC, OPPy and templated OPPy film coated electrodes (ca. 16A thickness) is shown in Figures 4-1 through 4-3. The cyclic voltammograms (CVs) for 0.4 x 10' M Ru(NH3)63+ (Figure 4-1) were peak shaped, and the kinetics were quasireversible with AEp values of 75 for GC, 75 for OPPy, 99 for OPPy/ado, 101 for OPPy/ino and 91 for OPPy/ATP electrodes. The cathodic peak potentials of -312 mV for OPPy, -333 mV for OPPy/ado and OPPy/ino and -328 mV for OPPyATP were not significantly shifted for the OPPy and templated OPPy film electrodes compared to the cathodic peak potential at the bare GC electrode of -326mV which indicated that the ultrathin films did not significantly alter the response of Ru(NH3)6+.

In contrast, the CVs for 0.2 x 10" M uric acid (Figure 4-2) and 0.3 x 10-3 M adenine (Figure 4-3) were irreversible as evidenced by no return peak on the reduction and a AEp value greater than 30 mV in the case of uric acid at the coated GC. In general, the CVs for uric acid were peak shaped with the exception of OPPy/ino and OPPy/ATP where a plateau was observed while the CVs of adenine did not show peaks, except at GC.







78






5



4



3
GC
_ OPPy
...OPPyado
i2 OPPy"ino
--------..OPPy/ATP 4, /

. 2 ~ ~.......OP. /in....�
o -/o











o -ioo -200 -300 -400 -500 potential (m V)


Figure 4-1 Cyclic voltammograms of 0.4 mM Ru(NH3 )6 31 in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm, film thickness ca. 16A, scan rate 0.020 V/s.

































I' I.
I.
I, I, I,
I


-GC
OPPy
....... OPPy/ado
------ OPPy/ino
-.----- OPPy/ATP


- I


. I * I * I * I


400


300


200


100


potential (mV)


Figure 4-2 Cyclic voltammograms of 0.2 mM uric acid in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm2, film thickness ca. 16A, scan rate 0.020 V/s.


0.5 0.0



-0.5


-1.0



-1.5


c U)
L.
0


-2.0



-2.5



-3.0


500


I








80






6


4 -"/


2 -I




0~~~~~1 --- ~-



4-4


SGC
---OPPy
.OPPy/ado
-8 -----OPPy/ino
-------- OPPy/ATP

-10


-12

1200 1000 800 600 4W0 200 0 potential (mV)



Figure 4-3 Cyclic voltammograins for 0.3 mM1 adenine in 0.5 M pH 7.0 potassium
phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode
area 0.07 cm', film thickness ca. 16A, scan rate 0.020 V/s.









81

The current for each probe was measured at the peak, or in the case of adenine, at a potential where the response was greater than the background response, but where the background currents were not too high to sacrifice sensitivity.

The potentials used in the measurements were ca. -0.30V vs SCE for Ru(NH3)63, 0.35V vs SCE for uric acid and 1.OV vs SCE for adenine. All measurements of analyte current were background subtracted by first measuring the background currents (at least three measurements were taken), averaging the background currents and then subtracting them from each analyte signal. The analyte currents used to calculate sensitivities were based on averaged results of at least three analyte measurements. Sensitivities at bare GC and OPPy modified GC electrodes were obtained as the slope of the current vs concentration plots for each respective electrode. The slope was calculated by regressing the data through (0,0) because in theory since the background currents were subtracted from the analyte currents, the current should be close to zero at zero concentration of analyte. The sensitivity results, based on the fits of the data through the origin, for the three probes are summarized in Table 4-1. Regression analysis of the data not through the origin was also calculated for comparison, and the results are fisted in Table 4-2. Typical intercept values of less than �1jLA indicated that fitting the data through the origin produced a small error in most cases. In addition, theoretical sensitivities calculated from equation 2.1 for Ru(NH3)6 for a reversible system, and from equation 2.2 for uric acid for an irreversible system (Bard and Faulkner, 1980), are listed. For adenine, the sensitivity was estimated for a two electron irreversible system, since the exact oxidation mechanism at pH 7.0 is still unknown, and in acidic solution up to six electrons can be transferred during the oxidation (Dryhurst and Elving, 1968).











Table 4-1 Sensitivity of Ru(NI3)63 , Uric Acid and Adenine at Different Surfaces



Ru(NH3)63+ Uric Acid Adenine


peak potential (mV) - -300 - +350 - +1100
pKb 5.4, 11.3 < 2.0, 4.1, 9.8
theoretical sensitivity' 2.7 4.2 4.2
bare GC 6.0 � 0.5 14.8 0.9 15 � 1 OPPyd 10 � 40 6.7 0.5 15 � 2
OPPy/adenosine ,d 6.6 � 0.5 8.0 - 0.3 11.9 � 0.1
OPPy/inosined 6.8 � 0.1 8.6 � 0.1 9 � 2 OPPy/ATPd 4.7 � 0.2 5.1 � 0.3 9�1


Note: sensitivity (iiA/mM) as a result of linear regression of the data through the
origin
aelectrode area ca. 0.07 cm2, scan rate 0.020 V/s, analyte concentrations
0.1-1 x 10"3M in 0.5 M pH 7.0 potassium phosphate buffer, all potentials vs SCE

b(Brown, 1991)

ctheoretical current for Ru(NH3)63",equation 2.1, for uric acid and adenine,
equation 2.2, D. estimated at 10' cm2/s for each probe
'film thickness ca. 16 A


'obtained from non (0,0) regression analysis of the data











Table 4-2 Non (0,0) fits of Ru(NH3)63", Uric Acid and Adenine Calibration Plots at
Different Surfaces


electrode Ru(NH3)6 3+ b Uric Acid b

bare GC y=3.9(�0.2)x+ 1.2(�0.1) y= 6.7(�-0.8)x -0.2(+0.1) OPPy, y=10(�4)x+2(�3) y=7.6(�0.6)x-0.I (�0. 1)

OPPy/ado' y-4.6(�0.5)x+1.2(�0.3) y=8.6(�0.4)x-0.08(�0.04) OPPy/ino' y=6.3(�0.2)x+0.3(�0.1) y=8.6(+0.1)x-0.08(�0.07) OPPy/ATP y=4.0(�0.5)x+0.4(+0.3) y=4.7(0.3)x+0.09(-0.05)


Note: slope (iiA/mM) and intercept (itA) electrode area 0.07 cm2, scan rate 0.020 V/s, all potentials vs SCE banalyte concentrations 0.1-1 x 10" M in 0.5 M pH 7.0 potassium phosphate buffer


'film thickness ca. 16A


electrode2 adenine b bare GC y=18(�2)x-1.0(�0.6) OPPy� y=20(+8)x-1(�2) OPPy/ado' y= 12.2(�0.6)x-0.1 (�O.1) OPPy/ino' y=3.01(�l)x+l(�) OPPy/ATPC y=2.7(-0.5)x+ 1.4(�O. 1)









84

For all the probes the sensitivity at bare GC, in slow scan voltammetry, was higher than the theoretical sensitivity. Ru(NH3)631 can interact with the surface oxides (Kovach et al., 1986), and both uric acid (Dryhurst and De, 1972) and adenine (Dryhurst, 1972) have been reported to adsorb on GC. The adsorption interactions can contribute to the increased sensitivity because of the resulting preconcentration which gives a higher concentration of analyte at the surface. The higher than theoretical sensitivity for adenine may be due in part to a larger number of electrons in the oxidation (Dryhurst and Elving, 1968) than used in the calculation.

At ca. 16 A thick OPPy film electrodes Ru(NH3)631 sensitivity increased (-37%) compared to the sensitivity at the bare GC. The sensitivity of uric acid was lower (-55%) than at the bare GC but remained higher (-60%) than the theoretical sensitivity. The result for uric acid showed that at OPPy film electrodes the response of uric acid was not as efficiently suppressed as the response of Fe(CN)63" probably as a result of a smaller negative charge on uric acid and possible hydrophobic interactions of uric acid with the film and electrode. If -NH' groups were present in the OPPy film as indicated by ESCA analysis discussed in Chapter 3, then the reponse of uric acid could also be due to favorable interactions with these groups. Since Fe(CN)6' should also interact favorably with any -NH' groups present, exclusion of Fe(CN)6' indicated that either multiple interactions between the probe and the film, e.g. hydrophobic and hydrophilic, contributed to the response or the density of the -NH' groups was low. Adenine sensitivity did not change at OPPy coated GC compared to the sensitivity at the bare GC indicating that the surface processes important in adenosine detection were not significantly changed by the presence of the ultrathin OPPy film.









85

The high sensitivity in slow scan voltammetry at ultrathin film electrodes showed (Table 4-1) that at the highly permeable ultrathin films, slow in-film transport, which has been reported at thicker OPPy electrodes (Witkowski and Brajter-Toth, 1992), did not limit the response. For example, the high sensitivity of Ru(NH3)63" was not influenced by slow diffusion observed in thicker films (D3 =2.8 x 108 cm2/s, Witkowski et al., 1991), indicating that the high sensitivity must be influenced by the favorable interactions of Ru(NH3)631 with the film. Similarly, favorable interactions, rather than slow transport, must control the response of uric acid and adenine, which showed a relatively sensitive response at ultrathin OPPy films.

Templating ultrathin polypyrrole (ca. 16A thickness) with adenosine, inosine and ATP produced a small decrease in sensitivity for all probes as compared to OPPy electrodes with some exceptions for uric acid, which showed an increase in sensitivity at OPPy/ado and OPPy/ino electrodes (Table 4-1).

The sensitivity of the nontemplated OPPy membranes prepared from MeCN was higher than the sensitivity of the templated membranes prepared from an aqueous solvent, again with some exceptions for uric acid which showed an increase in sensitivity for OPPy films prepared in 80%/a MeOH and a decrease in sensitivity for OPPy films polymerized from water and NaClO4 electrolyte.

The sensitivity data listed in Table 4-3 were calculated using linear regression through (0,0) since the calibration data were background subtracted. Theoretically, at zero concentration there should be zero current if the background currents are subtracted.











Table 4-3 Sensitivity of OPPy Films Prepared from Different Polymerization Solvents


Ru(NH3)63+ b Uric Acid b Adenine b OPPy'(MeCN) 8.2 � 0.8 6.7 0.5 15 � 2 OPPyZ(80% MeOH) 6.0 � 0.2 10.5 0.1 4-2 OPPy (-20, NaCIO4) 6.4 �0.1 3.8 - 0.2 2.8 � 0.7


Note: sensitivity (jiA/mM)
aelectrode area 0.07 cm2, scan rate 0.020 V/s, film thickness ca. 16,A


banalyte concentrations 0.1-1 x 10-3M in 0.5 M pH 7.0 potassium phosphate buffer




Full Text

PAGE 1

PURINE TEMPLATED OVEROXIDIZED POLYPYRROLE MODIFIED GRAPHITE FOR DETECTION OF ADENOSINE By LISA DENISE SPURLOCK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1996

PAGE 2

This work is dedicated to my parents and family who have always encouraged me to pursue a higher education and never settle for anything but the best.

PAGE 3

ACKNOWLEDGMENTS I would like to thank my parents for their constant love and total support. They have always encouraged me to get an education and to enjoy life. I would also like to express my appreciation to my research advisor Anna Brajter-Toth for her guidance and helpful hints on surviving graduate school. I would like to especially thank my group mates who graduated before me for their words of wisdom, teaching and lasting friendship. They are Ana Marino, Maurice Thompson, Quan Cheng and Charlie Hsueh. I would also like to acknowledge the friendship and assistance of my other group members. Merle Regino and Roberto Bravo. Several others who were undergraduates working with me also deserve recognition for their hard work and dedication; they are Jackie Lewis, Andrew Praserthdam, Constantine Panakos and Ngoc Chou. Also, I would like to give my sincere appreciation to Alonso Jaramillo for his friendship, encouragement and devotion these past few years. Special thanks go to the faculty and staff who have helped me with instrumentation and experiments. They are Vanecia Young for ESCA analysis of my electrode surfaces, Kathryn Williams for the use of her instruments, Russ Pierce for all of his help with fixing our instruments and Steve Miles for his electronic expertise. I would also like to thank Jeff Brouwer for hanging in with me and my best friends in Gainesville, Susan Rasmussen and Andrea Pless, for their support and friendship.

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGMENTS iii LIST OF TABLES vii LIST OF FIGURES ix ABSTRACT xiii CHAPTERS 1 INTRODUCTION 1 Biological Significance of Adenosine 1 Metabolism and Catabolism of Adenosine in Cardiomyocytes 3 Advantages of Using Cardiomyoctes in Bioanalysis of Adenosine 4 Electrochemical Properties of Adenosine and Its Metabolites 5 Previous Strategies for Detection of Adenosine 7 Amperometric Based Detection 7 Potentiometric Based Detection 8 HPLC for Detection of Adenosine in Bioanalysis 9 Capillary Zone Electrophoresis for Detection of Adenosine .... 12 Properties of Carbon Electrodes 12 Polypyrrole and Overoxidized Polypyrrole in Sensor Design 15 Molecular Templating 21 Fast Scan Voltammetry 24 Characterization of Electrode Surfaces by X-ray Photoelectron Spectroscopy 27 Purpose of Work 30 2 EXPERIMENTAL 33 Reagents and Solutions 33 Electrodes 34 IV

PAGE 5

Reference and Auxiliary Electrodes 34 Working Electrodes 34 Instrumentation 38 Electrochemical Experiments 38 UV Absorption and ESCA Experiments 41 Fundamentals of Electrochemical Measurements 41 Cyclic Voltammetry 41 Chronocoulometry 44 Rotating Disk Electrode Experiments 45 3 PREPARATION AND CHARACTERIZATION OF OPPy AND TEMPLATED OPPy ELECTRODES 47 Procedure for Preparing Ultrathin OPPy Films by Polymerization and Overoxidation 47 Procedure for Preparing Templated OPPy Films on GC by Polymerization and Overoxidation 50 Procedure for Preparing Ultrathin OPPy and Templated OPPy Films at UMEs 54 Procedure for Template Release During Overoxidation of Templated Polypyrrole 55 ESCA Analysis of Bare GC, OPPy and OPPy/ ATP Films on GC 58 4 ANALYSIS OF THE ELECTROCHEMICAL RESPONSE AT GLASSY CARBON, GLASSY CARBON COATED WITH OPPy AND TEMPLATED OPPy 76 Sensitivity Data at Glassy Carbon and Glassy Carbon Coated with Ultrathin Films of OPPy and Templated OPPy 76 Selectivity at Bare, OPPy and OPPy Templated Electrodes 87 Adenosine Detection 88 Characterization of Film Permeability by Electrochemical Methods .... 91 Determination of Apparent Membrane Diffusion Coefficients .91 Effect of Diffusion vs Surface Interactions on Response 93 Conclusions 95 5 ANALYSIS OF SURFACE INTERACTIONS AT GC AND GC TEMPLATED OPPy ELECTRODES 98 Saturation Binding of Ultrathin OPPy and Templated OPPy Electrodes 98 Analysis of Membrane Interactions 110 Scatchard Plot Analysis 110 Langmuir Isotherm Analysis 115 v

PAGE 6

OPPy and OPPy Templated Surface Stability 118 Conclusions 120 6 CHARACTERISTICS OF PURINE TEMPLATED OPPy FILMS ON CARBON FIBER ULTRAMICROELECTRODES 123 Effect of Film Thickness on the Response of Macroelectrodes 125 Sensitivity of Bare, OPPy and OPPy/ ATP Modified Carbon Fiber .... 129 Analysis of Film Interactions at OPPy Modified UMEs 133 Fast Scan Voltammetry of Uric Acid 133 Conclusions 146 7 CONCLUSIONS AND FUTURE WORK 147 LIST OF REFERENCES 150 BIOGRAPHICAL SKETCH 163 vi

PAGE 7

LIST OF TABLES Table page 31 Comparison of Various Atom Area Ratios obtained from ESC A spectra for Bare and OPPy and OPPy/ATP modified GC electrodes 72 41 Sensitivity of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine at Different Surfaces 82 4-2 Non (0,0) fits of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine Calibration Plots at Different Surfaces 83 4-3 Sensitivity of OPPy Films Prepared from Different Polymerization Solvents ... 86 4-4 Sensitivity of Adenosine at OPPy Modified GC Electrodes 90 4-5 Apparent Diffusion Coefficients (cm 2 /s) for OPPy, OPPy/ado and OPPy/ATP . 92 46 Slopes of Log Peak Current vs Log Scan Rate Plots at GC and OPPy Film Electrodes 94 51 Scatchard Plot Analysis for Ru(NH 3 ) 6 3+ at bare and OPPy Modified GC Electrodes 113 5-2 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid at Bare and OPPy Modified Electrodes 117 53 Limits of Detection at Bare GC and OPPy Modified Electrodes for Ru(NH 3 ) 6 3+ , Uric Acid and Adenine 119 61 Sensitivity at OPPy and OPPy/ATP Modified Macro GC Electrodes with Varying Degrees of Film Thickness 126 6-2 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid Calibration data at OPPy and OPPy/ATP 128 Vll

PAGE 8

6-3 Sensitivity of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine at Ultramicroelectrode Surfaces 130 6-4 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid Sensitivity data at Bare and OPPy Modified Carbon Fiber Ultramicroelectrodes 134 6-5 Limits of Detection at Bare and OPPy Modified Carbon Ultramicroelectrodes 135 6-6 Background Subtracted Fast Scan Voltammetry Results 141 Vlll

PAGE 9

LIST OF FIGURES Figure page 11 Structures for adenosine, adenine and uric acid 2 21 Diagram of ESC A sample holder with GC electrode and Teflon insulation .... 37 31 Suppression of 10 mM Fe(CN) 6 3 ' response in 0.5 M pH 7.0 potassium phosphate buffer after repeated coatings of OPPy at GC; scan rate 0. 100 V/s, electrode area 0.07 cm 2 , deposition charge per coating ca. 3 5 pC/cm 2 , film thickness after 7th coating ca. 28A 49 3-2 Cartoon representation of template, adenosine or ATP, incorporation into polypyrrole 51 3-3 Cartoon representation of the templated OPPy structure after overoxidation of polypyrrole (Beck et al., 1987) 53 3-4 Suppression of 10 mM Fe(CN) 6 3 ' response in 0.5 M pH 7.0 potassium phosphate buffer after repeated coatings of OPPy at UME; radius 7 pm, scan rate 0.100 V/s, deposition charge 3.17 x 10‘ 2 C/cm 2 per coating, film thickness after 4th coating ca. 0.53 pm 56 3-5 ESCA valence band spectra of bare and OPPy modified GC electrodes film thickness ca. 12A 60 3-6 Wide scan spectrum of OPPy film on GC (ca. 12A thick) showing the presence of oxygen, carbon and nitrogen 61 3-7 Structure of overoxidized polypyrrole as proposed by Beck et al.(1987) 63 3-8 Structure of OPPy as proposed by Ge et al. (1994) 64 3-9 ESCA spectrum of chloride and phosphorous from OPPy film (ca. 12A) on GC 65 IX

PAGE 10

3-10 ESC A spectrum of phosphorous, aluminum and silicon of OPPy film (ca. 12 A thick) on GC 66 3-1 1 Mechanism of polypyrrole overoxidation in water (Beck et al., 1987) 68 3-12 ESCA spectrum of aluminum and silicon in OPPy/ ATP films (ca. 16A thick) on GC 70 3-13 ESCA valence band spectrum of OPPy/ ATP film (ca. 16A thick) on GC 71 3-14 ESCA spectrum of chloride and phosphorous in OPPy/ ATP film on GC (ca. 16A thick) 74 315 ESCA spectrum of calcium in OPPy/ ATP film on GC (ca. 16A thick) 75 41 Cyclic voltammograms of 0.4 mM Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s 78 4-2 Cyclic voltammograms of 0.2 mM uric acid in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s 79 4-3 Cyclic voltammograms of 0.3 mM adenine in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s 80 44 Cyclic voltammogram of 0.005 M adenosine (solid line^ in 0.5 M pH 7 potassium phosphate buffer at OPPy/ado film (ca. 1 6 A) GC electrode (0.07 cm 2 area), scan rate 0.050 V/s 89 51 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 99 5-2 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 100 5-3 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ ATP. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 101 x

PAGE 11

5-4 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 103 5-5 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 104 5-6 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ ATP coated electrode. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 105 5-7 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 107 5-8 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 108 59 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ ATP coated electrode. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s 109 61 Background subtracted cyclic voltammogram of 10 pM Ru(NH 3 ) 6 3+ in 0.070 M pH 7.4 sodium phsophate buffer at electrochemically pretreated carbon fiber UME (see text for details), electrode radius 7 pm, scan rate 1000 V/s, 250 scans signal averaged 138 6-2 Background subtracted cyclic voltammogram of 10 pM Ru(NH 3 ) 6 3+ in 0.070 M pH 7.4 sodium phsophate buffer at OPPy coated UME, electrode radius 7pm, film thickness ca. 0.53 pm, scan rate 500 V/s, 500 scans signal averaged 139 6-3 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phsophate buffer at pretreated UME (see text). Electrode radius 7pm, scan rate 500 V/s, 250 scans signal averaged 142 6-4 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phsophate buffer at pretreated UME (see text). Electrode radius 7pm, scan rate 1000 V/s, 250 scans signal averaged 143 xi

PAGE 12

6-5 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phsophate buffer at OPPy coated UME. Electrode radius 7pm, film thickness 0.53 pm, scan rate 500 V/s, 50 scans signal averaged 144 Xll

PAGE 13

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PURINE TEMPLATED OVEROXIDIZED POLYPYRROLE MODIFIED GRAPHITE FOR DETECTION OF ADENOSINE By Lisa Denise Spurlock May 1996 Chairperson: Anna Brajter-Toth Major Department: Department of Chemistry The objective of this work was to design an amperometric sensor for adenosine by modifying carbon surfaces with ultrathin films of templated overoxidized polypyrrole (OPPy) o (ca. 16A thickness) to improve the detection selectivity and sensitivity. Adenosine detection is of interest as a marker of oxygen deficiency to the heart. The response of adenosine at the film electrodes was greatly enhanced considering that no response was apparent at the bare surface. Despite this, the film electrodes were not sensitive enough for routine detection. The films were characterized with adenine as a model compound because of the greater detection sensitivity to adenine in order to identify interactions favorable for adenosine detection. The effect of molecular templates on film response was characterized additionally with ruthenium hexaamine, a positively charged hydrophilic probe, and a purine, uric acid, a negatively charged relatively hydrophobic probe, and neutral adenine. The results obtained from the electrochemical film characterization xm

PAGE 14

illustrated an increase in film compactness and small changes in film interactions with templating and was verified by Scatchard and Langmuir analysis. Overall, templating with purines decreased film sensitivity with a small improvement in selectivity to adenine. The films also suppressed large background currents (noise) found to interfere with adenosine detection. The structure of the films and the consequences of templating were additionally characterized by ultraviolet and x-ray photoelectron spectroscopy. Film transport properties were determined from rotating disk electrode measurements. Carbon fiber ultramicroelectrodes were also modified with templated OPPy films to further characterize the films and to design an in vivo sensor for adenosine. The compact, templated OPPy films were more hydrophilic based on the exclusion of uric acid and good sensitivity for the other probes. A new type of selectivity for uric acid based on counterion exclusion was discovered. These films showed good sensitivity to uric acid in KC1 while suppressing its response in phosphate buffer. In order to press the limits of detection for uric acid, fast scan voltammetry was used. The results showed that the increased scan rate at the OPPy films improved the sensitivity for uric acid without the need for lengthy surface pretreatment. xiv

PAGE 15

CHAPTER 1 INTRODUCTION Biological Significance of Adenosine Cardiovascular disease is the number one killer in America and affects one out of every five persons. Because one person out of six will die of a heart attack before reaching age sixty-five, the American Heart Association has spent $93 million in research of cardiovascular disease in the past year (American Heart Association, 1993). Through efforts of intensive research, the heart attack death rate has decreased 32.4% in the last ten years, and will continue to decrease as new methods are developed to prevent cardiovascular disease and to detect heart attacks. In order to detect early warning signs of heart attacks, this research has focussed on developing methods to detect adenosine. In 1929 Drury and Szent-Gyorgyi discovered that adenosine modified physiological processes by injecting it directly into mammals. Adenosine as shown in Figure 1-1 is an endogenous cardiac nucleoside produced from the degradation of adenosine triphosphate (ATP). The limited energy reserves from ATP and creatine phosphate in the myocardium, and the high metabolic demand require a constant supply of oxygen and substrates to myocardial cells, in order to preserve normal metabolic and contractile function (Neely and Morgan, 1974; Reimer and Jennings, 1986). If this balance is disturbed, e.g. myocardial ischaemia or heart attack, then rapid depletion of the high energy phosphate occurs, in which 1

PAGE 16

Adenine Uric Acid Figure 1-1 Structures of adenosine, adenine and uric acid

PAGE 17

3 80% of creatine phosphate is lost within minutes of regional ischaemia, and more than half the ATP in fifteen minutes (Jennings et al., 1981). Continual usage of ATP, along with the inability of mitochondria to rephosphorylate adenosine diphosphate (ADP), results in the accumulation of ADP and adenosine monophosphate (AMP) (Jennings et al., 1981). AMP is metabolized via 5-nucleotidase from the cytosolic fraction of the cell to adenosine (Schrader, 1983). Adenosine is then transported from ischaemic myocardial cells, in large quantities, within seconds after onset of oxygen deficiency (Belardinelli et al., 1989; Berne, 1963; Gerlach et al., 1963). Within thirty seconds, the adenosine concentration increases a hundred fold in coronary venous effluent since neutral adenosine can rapidly penetrate the cell membrane (Schrader et al., 1977). If one could monitor this influx of adenosine, then it would be possible to predict a heart attack. Metabolism and Catabolism of Adenosine in Cardiomvocvtes After adenosine diffuses out of the myocardial cells arteriolar dilation occurs, thus increasing coronary blood flow and increasing myocardial oxygen tension. This in turn reduces the rate of adenine nucleotide degradation and hence adenosine concentration (Berne, 1963). Excess adenosine can be rapidly degraded to inosine by adenosine deaminase which can be further degraded to uric acid (Baer et al., 1966; Belle, 1969), taken into the myocardium and transformed into AMP (Wiedmeier et al., 1972; deJong, 1972; Namm, 1973), or washed out by coronary circulation. The specific roles of adenosine are still controversial, but it is generally accepted that

PAGE 18

4 adenosine relaxes smooth muscle. On the surface of the sarcolemma or muscle membrane, receptors for adenosine and catecholamines exist to control cardiac contractility. Catecholamines control the inward flux of Ca 2+ which causes muscles to contract. Adenosine counteracts this by binding to its receptor site and acting as negative feedback inhibitor of the catecholamines (Schrader, 1981). Advantages of Using Cardiomvoctes in Bioanalvsis of Adenosine The high levels of adenosine released during periods of oxygen deficiency make adenosine the ideal molecule to detect for predicting heart attacks. For experimental purposes isolated myocytes are ideal for studying adenosine release. Most importantly, myocytes are large and sediment rapidly without centrifugation, allowing sample medium to be tested or changed without trauma to the cells (Dow, 1989). Myocytes are stable in the presence of Ca~ + , and they do not degrade nucleosides to free bases. The use of myocytes also allows one to distinguish myocyte specific processes from those occurring in interstitial cells of the myocardium and to eliminate hormonal or neural influences. From an analytical viewpoint, digestion of the heart to form myocytes is ideal since multiple samples from a homogeneous population are produced. In order to design an adenosine sensor several factors must be taken into consideration. The sensor must have a fast response time because once released, adenosine is rapidly degraded to inosine or taken up by the cell. Also, the sensor must have a small area especially for in vivo and in vitro measurements so as not to disturb the cell metabolism. For

PAGE 19

5 these reasons amperometric detection is well suited to adenosine detection, but inherent difficulties have to be considered first. Electrochemical Properties of Adenosine and Its Metabolites The structures of adenosine, adenine and uric acid are shown in Figure 1-1. Adenosine of 0.25xl0' 3 M has a peak potential (Ep) of 1 . 778-0. 087pH V vs saturated calomel electrode (SCE) at a pyrolytic graphite (PG) electrode at 0.005 V/s (Dryhurst, 1972). Under the same conditions, E p for adenine is 1.338-0. 063 pH V and a single well formed anodic voltammetric peak can be found at PG between pH 0 and 1 1 . Both peak potentials are concentration dependent and will shift to higher oxidation potentials with increasing concentration (Dryhurst, 1972). The absence of reduction peaks for both adenosine and adenine indicates that the oxidation process is irreversible (Dryhurst and Elving, 1968). Dryhurst also found that both of these compounds adsorb strongly to the PG electrode. The exact oxidation mechanism for adenosine has not been postulated possibly because the large E p which is close to the background discharge potential, makes the process difficult to study by electrochemical methods. Adenine, or 6-aminopurine, undergoes two sequential two electron-two proton oxidations to give first 2-oxyadenine and then 2,8-dioxyadenine. Finally another two electron oxidation at the C,4-C,5 double bond produces an unstable dicarbonium ion intermediate which can react further (Dryhurst, 1972). In vivo adenine was found to be oxidized to 2,8dioxyadenine directly (Bendich et al., 1960). Dryhurst discovered that equal numbers of

PAGE 20

6 electrons and protons were transferred during the rate determining step since in cyclic voltammetry the peak potential shifted 51 mV/pH unit to more negative potentials at 0.06 V/s and 56 mV/pH at 0.6 V/s. Analysis at both pH 2.3 and 4.7 using coulometry verified that ca. six electrons are involved in the overall oxidation, but since the electrolysis of 10' 3 M adenine takes two days to complete, a very slow intermediate step, an insoluble reaction product, film formation on the electrode, or any combination of these may occur. Dryhurst concluded that adsorption and a slow chemical intermediate step is present in the overall process (Dryhurst andElving, 1968). Uric acid, or 2,6,8-trioxypurine, at pH 4.7 shows a well formed two electron oxidation peak at ca. 0.47 V at scan rates below 0.3 V/s at PG (Struck and Elving, 1965). At faster scan rates a well formed cathodic peak appears at ca. 0.42 V and with increasing scan rate becomes relative height to the anodic peak (Dryhurst, 1969). The appearance of this reduction peak at higher scan rates indicates an unstable product which can undergo a rapid chemical follow up reaction (Dryhurst et al., 1983). At a scan rate of 0.005 V/s between pH 1 and 12, E p is 0.76-0.069pH at rough PG and 0.685-0.055pH V at PG between pH 0 andl 1.5 (Brajter-Toth et al., 1981). The primary oxidation process involves a two electrontwo proton oxidation at the C4-C5 double bond to give a diimine which is very unstable and can be rapidly hydrated to an imine-alcohol (Owens et al., 1978). In high concentrations of phosphate buffer, H 2 P0 4 attacks the diimine, and in low concentrations of phosphate buffer H 2 0 attacks the diimine (Goyal et al., 1982) The reduction peak due to the diimine can be observed at rough PG at 0.2 V/s at pH 8.0 provided that the surface has a large area because the diimine strongly adsorbs to the surface, and this adsorbed species is more stable than in

PAGE 21

7 solution (Owens et al., 1978). Because the oxidation potential is lower than adenine and adenosine, uric acid is easier to detect and could also be useful in predicting cardiac arrest since it is a metabolite of adenosine (Brown, 1991) Previous Strategies for Detection of Adenosine Amperometric Based Detection Dryhurst determined adenine in the presence of large concentrations of adenosine by using polarographic reduction at a dropping mercury electrode (DME) and coupling these results with voltammetric oxidation at a PG electrode (Dryhurst, 1972). In the DME studies, the polarographic limiting current was directly proportional to the sum of the concentrations of adenine and adenosine. In order to quantitate the concentrations of each, oxidation current of the mixture in pH 4.7 acetate buffer was measured at PG electrode, and the fact that both compounds adsorb on the surface was used in quantitation. The concentration of adenine was found to be dependent on the concentration of adenosine present, but at large concentrations of adenosine, adenine is desorbed from the electrode and its oxidation becomes a diffusion controlled process. So, the total amount of adenine and adenosine can be determined from DME studies; the amount of adenine can be determined from the voltammetric results and then respective concentrations for each can be found. Even though this process is relatively straightforward, extensive pretreatment of the PG electrode must be done in order to ensure reproducible results. The PG electrode was resurfaced with 600-grade silicon carbide paper and then placed

PAGE 22

8 into supporting electrolyte where a potential of 0.0V was applied for ten seconds, and then the potential was scanned to 1.4 or 1.5 V three times. Next, the electrode was removed without drying the tip and placed in the background solution, and the process was repeated. The first scan was typically not recorded because it was inconsistent with the rest due to the reduction of surface oxides. Once the background was recorded, the same procedure was repeated in the analyte solution. This method has use for determining both adenine and adenosine in a mixture at pH below 7, but it could not be used successfully for in vivo analysis due to the methods involved as well as the required treatment of the electrode. Yao et al. used a similar pretreatment procedure for determining purine bases and nucleosides at a glassy carbon (GC) electrode in Britton-RobinsonÂ’s buffer at pH 2.0-1 1 (Yao et al., 1977). They also found that these compounds strongly adsorb to GC, and the oxidation peaks of adenine and adenosine are masked in the presence of chloride ions. Adenine and adenosine were best determined at pH 2-4; so this method could not be used for physiological measurements either. Potentiometric Based Detection Potentiometric sensors are popular because ammonia, which can be detected by a gas sensing membrane, is a product in the enzymatic conversion of adenosine to inosine by adenosine deaminase (ADA). Deng and Enke have expoited this reaction to detect adenosine by immobilizing ADA onto an ammonia gas permeable membrane. This sensor has a limit of detection (LOD) at pH 9.0 in the micromolar range and is stable for a month. The problems

PAGE 23

9 with this sensor are dependence on pH, which limits its use in some biological samples, and a response time of ten minutes. Xiuli et al. have also designed a sensor based on an ammonia sensing membrane, but they have immobilized rabbit thymus tissue as the enzyme source (1992). This sensor also has a LOD of ca. 10' 5 M with a response time of seven minutes. This sensor is useful for body fluid samples, but is not sensitive enough for use in blood. However, adenosine can be detected in blood with no interferences or sample pretreatment if it is added directly to the sample. Another sensor based on this method was developed by Bradley and Rechnitz in 1984 who used mouse small intestinal mucosal cells as the enzyme source. In order to overcome the long response times of the previous sensors, Liu et al. designed an asymmetric polyurethane membrane consisting of a very thin hydrophilic polyurethane membrane, with a high density of polylysine groups to which ADA is attatched, coated to a hydrophobic plasticized polyurethane/poly(vinylchloride/vinyl acetate/vinyl alcohol) membrane, which adheres well to silicon dioxide surfaces and contributes to the good stability (Liu et al., 1993). This electrode has an LOD ca. 10" 5 M and takes sixty seconds to reach steady state. However, it has only been tested in buffer solutions of adenosine. Potentiometric sensors are advantageous because of the specificity of ADA to adenosine, but these sensors suffer from problems such as response time and difficulty in miniaturization that limits their routine use in bioanalysis of adenosine. HPLC for Detection of Adenosine in Bioanalvsis High performance liquid chromatography (HPLC) is the most widely used method for

PAGE 24

10 detection of adenosine. HPLC with electrochemical (EC) based detection is preferred because EC detection is more sensitive than ultraviolet absorption (UV), and because it is a more direct method than fluorescence, as adenosine and its metabolites have to be tagged for use in fluorescence detection. Henderson and Griffin have separated adenine and adenosine as well as other purine compounds with a reversed phase column using a glassy carbon electrochemical cell set at 1.5 V vs Ag/AgCl reference (Henderson and Griffin, 1984). Their procedure can be used for biomaterials with some sample clean up, but this method suffers from fouling of the EC cell due to adsorption of purine molecules to GC and contaminants which can be oxidized due to the large detection potential applied. Berne, et al. also have used EC detection, but the adenosine in the sample is first enzymatically converted to uric acid (Berne et al., 1986). This method was used to determine the interstitial fluid content of adenosine in different tissues and had a LOD of 5 fmol. The samples were collected from a small chamber placed on the epicardial surface of a dog’s heart. After ADA, nucleoside phosphorylase and xanthine oxidase were added to convert adenosine to uric acid. The reaction was stopped by the addition of HC10 4 , and the precipitate was removed by centrifugation and stored overnight at 4°C. Before analysis, the sample was centrifuged once again. This technique can measure adenosine below basal levels so that any increase in concentration, as expected during cardiac arrest, can be measured, but a lengthy sample pretreament has to be done before analysis. Ontyd and Schrader have determined adenosine, inosine and hypoxanthine in human plasma by reversed phase HPLC using UV detection at 254 nm. The inaccurracy of the determination of adenosine levels in plasma stems from the sample collection since adenosine

PAGE 25

11 can be rapidly degraded. Ontyd and Schrader have overcome this by designing a syringe which stops any degradation once the blood is drawn from the patient by immediately mixing the blood with dipyridamole in Locke solution. After collection the sample was run once; then ADA and nucleoside phosphorylase were added and another chromatogram was taken. The enzymatic peak shift was used to confirm the presence of the analytes, since a number of unidentified peaks are present in the first chromatogram. The only drawback of this approach was that the chromatograms were not consistent between individuals mainly because the nutritional states and pathological conditions varied. HPLC with fluorescence detection has also been developed for detection of adenosine and inosine. Gardiner monitored the formation of dichlorofluorescein which is the oxidation product from the reaction of dichlorofluorescin with H 2 0 2 in the presence of horseradish peroxidase (Gardiner, 1979). H 2 0 2 is a product in the conversion of hypoxanthine by xanthine oxidase to uric acid. The amount of adenosine is quantitated by measuring the conversion of inosine plus adenosine to hypoxanthine and subtracting the conversion of inosine to hypoxanthine. This method has a linear dynamic range of 6.2 x 10' 9 to 9.3 x 10' 7 M and has been tested on dogs during a five minute reversible occlusion of the left anterior descending coronary artery. The concentration changes of adenosine in blood could be detected, but at least ten minutes were needed for the enzymatic conversion processes. Another HPLC with fluorescence detection method was developed by Preston for use in determining nucleoside concentration in marine phytoplankton (Preston, 1983). The adenosine samples had to be incubated in acetate buffer and chloroacetaldehyde for thirty

PAGE 26

12 minutes. As a result, CbTÂ’-etheno derivatives which fluoresce are the detected products. The LOD is 2 x 10' 9 M and no interferences have been discovered. Capillary Zone Electrophoresis for Detection of Adenosine Kuhr and Yeung have used capillary zone electrophoresis (CZE) with laser induced indirect fluorescence detection for the determination of nucleosides and proteins (Kuhr and Yeung, 1986). This method has good sensitivity with a linear dynamic range (LDR) of 50 to 1 00 amol and has been tested for detection of lysozymes, but has not been used in real samples. Separations as well as amperometric and potentiometric sensors have offered insight into the metabolism and chemistry of adenosine, but still no sensor exists which can be used directly in vivo for long periods of time. This research has set out to develop a sensor surface which does not foul due to adsorption of the analyte and products, which has real time response and requires no sample pretreatment. Properties of Carbon Electrodes Carbon is the most widely used and extensively studied electrode material for analytical purposes (McCreery, 1991). Carbon has high conductivity, is readily available, is cheap, has low chemical reactivity and has a wide potential window at biological pH (Kinoshita, 1988). The most widely used forms of macroelectrode materials are PG and GC (Dryhurst and McAllister, 1984), and carbon fiber has also become popular for use in ultramicroelectrodes.

PAGE 27

13 Glassy carbon (GC) or vitreous carbon is made from polymeric resins such as polyacrylonitrile or phenol/formaldehyde polymers heat treated at 1000-3 000 °C under pressure to release H, N, or O atoms to leave an extensively conjugated sp 2 carbon structure. The polymeric backbone remains intact forming a complex structure of interwoven graphitic ribbons which accounts for its mechanical hardness. GC is impermeable to gases, resistant to chemical attack, electrically conductive and can be obtained in high purity, making this material very advantageous for analysis (Kinoshita, 1988). However, this surface is very inactive despite its small degree of hydrophilicity. Pyrolytic graphite (PG) is prepared by high temperature decomposition, above 1200°C but below 3800°C, of gaseous hydrocarbons such as methane and propane onto a hot surface. PG consists of highly conjugated hexagonal rings or carbon atoms arranged in planes held together by weak van der Waals forces (Kinoshita, 1988). The conductivity along the planes (edge plane) is metallic, but conductivities perpendicular to the planes (basal) are semiconducting (Kinoshita, 1988). PG is a more active surface than GC due to these properties. The exact structural and physical properties of PG depend on the surface of deposition, temperature of curing and further treatment after manufacturing. For our purposes PG was roughened to form rough pyrolytic graphite (RPG) using 600-grit silicon carbide paper. Because of the highly desirable properties of carbon and the need for miniaturization for in vivo analysis carbon fiber has become popular. Carbon fibers generally range in diameter from 5 to 25 pm and are made from heat treatment of polymeric precursors or catalytic chemical vapor deposition (CCVD) (McCreery, 1991). Heat treatment procedures

PAGE 28

14 are similar to those used in the preparation of GC, but the molten polymer is spun or extruded to form fibers. CCVD involves a catalytic dehydrogenation of hydrocarbons like benzene onto small particles of Fe, Ni or Co. Carbon atoms diffuse to the surface or through the bulk of the metal particle after decomposition of the reagent gas and precipitate to form graphite on the lower cooler side of the metal particle. The carbon fibers formed from either method can have different structures depending on the specific procedure, but the Raman spectra of carbon fibers prepared at high temperatures are similar to those of highly ordered pyrolytic graphite, with preferred graphitic plane orientations (McCreery, 1991). Carbon fiber surfaces have not been as extensively characterized as GC and PG have been. One of the disadvantages of graphite as an electrode material in bioanalysis is the poor reproducibility of the electrode surface. The irreproducibility is related to adsorption as well as to the oxidation of the surface, which is especially apparent at high positive potentials (McCreery, 1991). To improve the reproducibility the electrodes must be pretreated before measurements. Additionally, large background currents at high positive potentials, from electrode charging and from the faradaic surface reactions, involving surface groups such as quinones, have to be controlled (McCreery, 1991). For analysis at graphite several surface pretreatment procedures, including polishing (Kaman, 1988), electrochemical activation (Engstrom, 1982; Engstrom and Strasser, 1985), laser pretreatment (Poon and McCreery, 1986, 1987, Poon et al., 1988) and heat treatment (Fagan et al., 1985) have been developed. Other promising methods include dynamic modification of graphite with surfactants (Marino and Brajter-Toth, 1993) and with membranes (Murray, 1992) such as Nafion (Waller, 1986), a perfluorosulfonated ionomer.

PAGE 29

15 Nafion membranes preconcentrate positively charged analytes while excluding interfering anions, and the membranes limit adsorption, protecting the electrode surface. In vivo measurements of neurotransmitters are a successful example of Nafion applications, where favorable partitioning of neurotransmitters into the membrane improves detection sensitivity while the interfering negatively charged ascorbic acid is excluded (Gerhardt et al., 1984; Nagy et al., 1985), with the membrane protecting the surface from direct contact with the complex biological sample. Nevertheless, Nafion coated electrodes suffer from several problems such as slow response time due to low diffusion coefficients in the film, memory effects due to strong binding with the S0 3 ' groups in the film, and saturation (Guadalupe and Abruna, 1985; Whiteley and Martin, 1988; Wang and Tuzhi, 1986). Also, Nafion coatings are often nonuniform in thickness and reproducibility of film formation is poor (Gao et al., 1993). Gelimmobilized enzymes have recently been developed for bioanalysis (Wang and Heller, 1993) although these require multicomponent films to facilitate the enzymatic reaction. Polypyrrole and Overoxidized Polvpvrrole in Sensor Design Ultrathin overoxidized polypyrrole (OPPy) films are another type of membrane that has recently been developed for analytical applications (Witkowski et al., 1991; Witkowski and Brajter-Toth, 1992; Hsueh and Brajter-Toth, 1994). It has been shown that OPPy films can be easily prepared by electropolymerization so that the response is not limited by slow infilm diffusion and irreversible interactions (Hsueh and Brajter-Toth, 1994; Witkowski and Brajter-Toth, 1992). Ultrathin OPPy films have been shown to be cation-selective and have

PAGE 30

16 been shown to eliminate the response of Fe(CN) 6 3 ' (Hsueh and Brajter-Toth, 1994) presumably because of the high electron density of the film carbonyl groups (Palmisano et al., 1995; Beck et al., 1987; Christensen and Hamnett, 1991); similar interactions can suppress response of uric acid (Witkowski and Brajter-Toth, 1992) which is a common biointerferent. Another desirable feature of OPPy films is that the membrane sensitivity and selectivity can be changed depending upon the conditions in which the films were made. This work has sought to alter the properties of ultrathin OPPy films by templating these films with purine molecules in order to design a sensor for in vitro use for adenosine and its metabolites. In 1980, Diaz and Castillo reported the growth of thin polymer films of polypyrrole, grown on Pt electrodes by electropolymerization (Diaz and Castillo, 1980). Polypyrrole prepared in this manner is deposited as a cation and consists primarily of a,a' linkages, but P-coupling can also occur (Salmon et al., 1982). Polypyrrole has variable conductivity and can change reversibly from a conducting to a nonconducting polymer within a given potential window (Diaz and Castillo, 1980; Feldman et al., 1985). Because of these properties, polypyrrole films have been used for a wide variety of applications. Several factors influence the behavior of polypyrrole films such as the applied potentials during polymerization (Asavapiriyanont et al., 1984), the solvent employed (Diaz and Castillo, 1980; Asavapiriyanont et al., 1984; Ferreira, 1990) and the temperature (Ogasawara et al., 1986). However, the most important factor is the counterion incorporated during polymerization (Imisides et al., 1991). The morphology, conductivity (Yamaura et al., 1988), adhesion and mechanical strength are all affected by this choice (Skotheim, 1986; Freund et al., 1991). Polypyrrole can be readily synthesized from a range of solvent media,

PAGE 31

17 including aqueous and nonaqueous solvents, so the choice of the counterions is endless (Imisides et al., 1991). In general, polypyrrole films grown from nonnucleophilic solvents and electrolytes, i.e. acetonitrile with 1% H 2 0, have good conductivity, and aqueous solvents produce less conductive films (Imisides et al., 1991). Aqueous solvents are advantageous for most applications, especially industrial, since a larger variety of dopants can be used as the counterion, but the potential of the anion must be higher than pyrrole to allow pyrrole to polymerize without competition from the electrolyte (Takakubo, 1987). Typical dopant levels are one counter ion for every three to four pyrrole units (Mitchell et al., 1988). These ions are incorporated as counterions during the oxidation of pyrrole, in which every third repeat unit has a positive charge (Diaz and Castillo, 1980). The exact chemical configuration of the polypyrrole chains is ambiguous, but it is accepted that the molecular organization is highly disordered. Mitchell, et al. have discovered that more anisotropic counterions lead to a higher level of preferred orientation in the film, but anions with more than three aromatic units do not make a good film because these larger counterions may separate the polymer chains or separate some sections (Mitchell et al., 1988). Aromatic conterions are preferred since these will aid in retention of the polymer to the electrode surface since polypyrrole is planar (Mitchell et al., 1988). Some example of counterions used are surfactants, such as dodecylbenzenesulfonate ion (Lyons et al., 1993), ferricyanide (Chen et al., 1993) and cobalt porphyrin (Armengaud et al., 1990). Polypyrrole has been polymerized on various substrates such as Pt (Diaz and Castillo, 1980), Au, Fe, indium-tin oxide coated glass (Street et al., 1983), mercury (Bradner and Shapiro, 1988) and plastic films coated with gold, tin oxide or silver (Barisci and Wallace,

PAGE 32

18 unpublished). Titanium, aluminum, mild steel and brass (Cheung et al., 1988) have been also been used. Glassy carbon is the best choice because of its large potential window and polypyrrole films have been shown to strongly adhere (Imisides et al., 1991); polypyrrole cannot be removed except by mechanical grinding or treatment with chromic acid. Nishizawa et al. have shown that polypyrrole adheres well to hydrophobic substrates, so glassy carbon is a good choice since glassy carbon is mostly hydrophobic (Nishizawa et al., 1991). If the oxidation potential is increased beyond 0.8 V, polypyrrole becomes nonconducting (Asavapiriyanont et al., 1984). Beck et al. have shown that overoxidation in aqueous solutions occurs via a nucleophilic attack of the hydroxide or water on the pyrrole unit, followed by oxidation of the hydroxy group to a carbonyl group ( Beck et al., 1987). Wernet and Wegner have also reported that polypyrrole can be overoxidized by cycling in very basic solutions like NaOH (Wernet and Wegner, 1987). This treatment may produce a mixture of carbonyl groups as well as negatively charged hydroxyl groups (Gao et al., 1994). Overoxidation results in oxidation and final scission of the polypyrrole chains at a small number of sites, such that smaller chains are entangled at the surface. As a result, the conjugation of the polymer is disrupted, but no significant material is lost (Christensen and Hamnett, 1991) even though total release of the counterions may occur during the course of overoxidation (Beck et al., 1987). This infers that the structure of polypyrrole is not greatly changed by overoxidation; so incorporation of a counterion to change the structure will not be futile even though the counterion could be lost during overoxidation. Several groups have used overoxidized polypyrrole (OPPy) films for sensing purposes because of its ability to exclude anions (Freund et al., 1991). Gao and Ivaska polymerized

PAGE 33

19 pyrrole on glassy carbon in the presence of sodium dodecyl sulfate, and then overoxidized polypyrrole in NaOH (Gao and Ivaska, 1993). They observed high selectivity for dopamine in the presence of ascorbic acid with a two minute preconcentration. The exclusion of ascorbic acid was presumably due to the presence of the hydroxyl and carbonyl groups within the OPPy structure which also can contribute to the increase in sensitivity (LOD 40 nM) of dopamine which is positively charged. The enhanced selectivity to dopamine is accounted for by the use of sodium docecyl sulfate during the polymerization process because once the film is overoxidized the surfactant is removed leaving a more porous structure than a smaller ion. The presence of the film also protects the electrode surface from fouling by the oxidation product of dopamine. However, the drawbacks with this sensor are long preconcentration times as well as renewal of the film which has to be done before each measurement, since the film thickness is ca. 1 pm thick. Gao et al. have used indigo carmine as the counterion during polymerization of pyrrole to develop a sensor for dopamine in the presence of ascorbic acid (Gao et al., 1994) for use in in vivo measurements. The same procedure as described above was used. They found that with ascorbic acid concentrations lower than 0.2 mM, which is the normal concentration level in mammalian brain, the selectivity with 0.25 to 1.0 pm thick film is optimal for detection of dopamine between 0.1 to 10 pM. Thinner films produce better detection limits, but large oxidation currents stemming from the oxidation of ascorbic acid cause problems with detection of dopamine. The problems of this sensor are the difficulty in controlling film thickness, and hence film reproducibility, and the small lifetime of two hours. Besides changing the conditions during polymerization, Gao et al. have also varied the

PAGE 34

20 overoxidation conditions to change the permeability of the films (Gao et al., 1994). A wide variety of solvents, e g. CH 3 OH, NaOH, HC1 and CH 3 CN, were used to overoxidize polypyrrole after polymerizing with sodium dodoceyl sulfate. Their results showed that the most sensitive and well defined voltammograms for dopamine were obtained from the polypyrrole film overoxidized in NaOH solution, due to the formation of both carbonyl groups and hydroxyl groups. The existence of these functional groups was confirmed by Fourier transform IR spectroscopy. The porosity of the films was controlled by the doping ion, but the permeability was increased by increasing the pH of the electrolyte used to overoxidize. For practical purposes this sensor is limited to dopamine concentrations lower than 1 X 10' 3 M due to the saturation and renewal problems. These films only need a sixty second preconcentration time, but renewal by soaking the electrode in phosphate buffer takes up to three hours for complete removal of analyte. Despite this, these films exhibited excellent antifouling properties in albumin solutions with dopamine where a rapid loss in the oxidation current at bare GC occurred. Centonze et al. have manipulated the permselective and antifouling properties of OPPy for use as a glucose sensor. This group has immobilized glucose oxidase in 0.67 pm, or thicker, OPPy films on GC overoxidized in phosphate buffer for use in a flow injection apparatus. This sensor has a shelf life of ten days and shows a 75% decrease in sensitivity after six days of continuous use. In addition this sensor is interference free as the currents for several common interferents (ascorbic acid, uric acid, cysteine, acetaminophen) are suppressed in comparison to the glucose response. The linear dynamic range of the sensor is 1.0 X 1 O' 2 to 5.0 X 10' 2 M, but can be extended to lower concentrations by use of a film

PAGE 35

21 twice as thick. This sensor also shows the feasibility for use in real samples; a pooled serum sample was tested by the sensor in a flow injection apparatus as well as by a routine enzymatic colorimetric methods. The results were not significantly different, indicating that this sensor can be used in real samples. Based on these results, the feasibility for tailoring overoxidized polypyrrole by use of specific counterions during the polymerization process as well as by the optimization conditions for overoxidation can produce a sensor with desirable characterics for detection of adenosine and its metabolites. Others have designed polymers and membranes for structurally similar molecules to adenosine in a similar manner by molecular templating. Molecular Templating Molecular templating or molecular imprinting involves the preparation of polymers that are selective for a particular compound (Edelman and Wang, 1992). The compound of interest acts as a template in which monomers are prearranged and complementary interactions occur. Next, the monomers are polymerized about the template and in the final step, the template is removed by an extraction or other method to leave a specific binding site for the template or a structurally related compound. In general, two approaches have been taken: (1) the template has been covalently but reversibly bound (Wulff, 1986), or (2) the initial interactions between monomers and the print molecule have been non-covalent (Arshady and Mosbach, 1981). An example of reversible covalent binding is the polymerization of vinylphenylboronic acid with phenyl-a-D-mannopyranoside as the template

PAGE 36

22 to design polymers that could resolve racemic mixtures of carbohydrates (Wulff, 1986). Boronate groups have also been used in this manner for detection of nucleotides (Norrlow et al., 1986). This method is hampered by a limited number of compounds with suitable binding groups and useful reversible interactions that do not disrupt the polymer matrix (Ekberg and Mosbach, 1989). Polymers designed by non-covalent interactions are more versatile because more monomer-print molecule interactions including ionic, hydrogen bonding, hydrophobic, or charge transfer can be used (Ekberg and Mosbach, 1989). Early work with this approach used dyes as the template in acrylic monomers (Arshady and Mosbach, 1981) for selectivity to the dye itself. MosbachÂ’s group has used this approach for separating amino acid derivatives by interacting the amino group of the print molecule with the carboxyl group of a monomer to separate amino acid derivatives on the basis of substrate and enantioselectivity (Ekberg and Mosbach, 1989). In order to design a selective polymer, several factors have to be taken into consideration. In general, templates that can form multiple interactions with the monomer form polymers with the best resolution or higher selectivity (Wulff and Lohmar, 1979; Wulff and Gimpel, 1982; Wulff et al., 1980, 1984). Reduced specificity can also occur if the template molecule is polymerized very close to the surface in that an incomplete imprint is formed (Wulff, 1986). The polymerization conditions are crucial since polymerization at extreme conditions such as thermal decomposition destabilizes the complex between the template and monomers and limits the types of templates that can be used. In order to maintain the polymer structure as dictated by the template, crosslinkers are employed and

PAGE 37

23 generally, less crosslinking should produce a less defined matrix with poor resolving ability (Wolff, 1986). A compromise in rigidity has to be made because the polymer has to be flexible enough to allow fast binding and diffusion from the cavity (Wulflf, 1986). Molecular imprinting has been used to design a wide variety of sensors especially for biomolecules since the theory behind this process relies on the mechanism for molecular recognition as found in nature. Network polymers have been prepared for detection of adenine derivatives (Shea, et al., 1993), antibody mimics (Vlatakis et al., 1993) and for HPLC column packing (Sellergren et al., 1985; Andersson et al., 1990; Kempe and Mosbach, 1994). Others have used the concept of molecular imprinting to develop anion sensors (Ikariyama and Heineman, 1986) and potentiometric sensors for C1‘ (Dong et al., 1988) and N0 3 * (Hutchins and Bachas, 1995) but in the potentiometric methods, the template molecule was left in the polymer matrix for additional interactions with the anions and was not removed to leave the molecular binding site. Based on the principles of molecular templating and the permselectivity of OPPy, this work has sought to design templated OPPy structures for use in detection of adenosine and its metabolites. The templates used in templating polypyrrole first and then in forming templated OPPy were structurally related to adenosine and were expected to be incorporated and to weakly interact with the forming polypyrrole through hydrogen bonding, hydrophobic and electrostatic interactions, provided the template was anionic. Templating should alter polypyrrole morphology based on the changes reported in polypyrrole morphology with electrolyte, solvent and the polymerization conditions (Gao et al., 1994). The template was expected to be released during overoxidation of polypyrrole in phosphate buffer typically

PAGE 38

24 carried out at potentials slightly lower than the oxidation potentials of the templates. The molecular sites from templating were expected to remain in OPPy since the carbonyl groups introduced in the formation of OPPy do not significantly alter polymer morphology (Christensen and Hamnett, 1991). Characterization and utility of these structures were done by UV absorption, x-ray photoelectron spectroscopy and electrochemical analysis including fast scan voltammetry. Fast Scan Voltammetry In order to further enhance the sensitivity and selectivity at cabon fiber electrodes, fast scan voltammetry can be used since the current is proportional to the square root of scan rate for diffusion controlled systems and proportional to the scan rate for an adsorption controlled process (Bard and Faulkner, 1980). At scan rates above 100 V/s adsorption can dominate even for weakly adsorbing species since the current increases faster with scan rate, thus allowing higher peak current, lower detection limits and an increase in signal to noise (Hsueh and Brajter-Toth, 1993; Freund and Brajter-Toth, 1992; Wiedemann et al., 1991). Fast scan rates can only be used at ultramicroelectrodes since they have a small time constant and a small iR drop. Fast scan voltammetry was first introduced by MillarÂ’s group (Millar et al., 1981) and was later improved by WightmanÂ’s group (Kuhr and Wightman, 1986). One of the advantages of fast scan voltammetry is that the differences between the electrochemical kinetics of the analyte and the interferant can be used to resolve the two signals (Baur et al., 1988). An example of this is the measurement of dopamine in the presence of ascorbic acid.

PAGE 39

25 The electrochemical kinetics of ascorbic acid are slow, so fast scan voltammetry enhances this affect so that the oxidation peak potential of ascorbic acid shifts away from the oxidation peak potential of dopamine as the scan rate increases (Hsueh and Brajter-Toth, submitted). Another advantage of fast scan voltammetry is the preservation of the working electrodes. Many redox reactions of biomolecules are followed by chemical reactions which produce side products. These can adsorb on the electrode surface and eventually cause fouling after several scans. But because of the short time span of fast scan voltammetry, the products can undergo electrochemical reactions and be converted back to the original analyte before chemical reactions can occur in the solution that would foul the surface. Fast scan voltammetry can also provide qualitative information about the analyte which may be used to confirm the identity of the analyte of interest in complex samples. At high scan rates, thickness of the diffusion layer of biomolecules like dopamine will be greatly reduced, so the chance of blocked diffusion by an in vivo environment is reduced. Hence, fast scan voltammetry can guarantee that voltammograms obtained in vivo will be similar to those obtained in buffer solution. Further enhancement can be achieved by signal averaging which produces high temporal resolution since the time required to complete a voltammogram is short. WightmanÂ’s group has investigated this technique to improve detection limits or increase signal to noise (Wiedemann et al., 1991). Signal averaging along with analog and digital filtering allowed the detection of 1.0 x 10" 7 M of dopamine in vivo. They averaged forty voltammograms at a scan rate of 300 V/s to improve the signal to noise. The number of voltammograms signal averaged is limited by the scan rate and the experimental time scale.

PAGE 40

26 Shorter cycling time allows more cyclic voltammograms to be acquired in a specific time. The enhancement in signal to noise is expected to be proportional to the square root of the number of scans (Hsueh and Brajter-Toth, submitted). For example, for 3125 averaged scans, the signal to noise is improved fifty-six times. High frequency noise is rapidly reduced by signal averaging as the number of scans increases, but the low frequency noise remains the same even after a thousand scans (Hsueh, 1995). The low frequency noise may be harmonic with or close to the frequency of cycling. Fast scan voltammetry can improve the selectivity and sensitivity of bare carbon fiber as well as coated carbon fiber electrodes. WightmanÂ’s group has used fast scan voltammetry to enhance the kinetic differences of dopamine and ascorbic acid at Nafion coated electrodes, since ascorbic acid shows slow kinetics at Nafion and dopamine partitions favorably into the film (Baur et al., 1988). The direct result of coupling fast scan with Nafion is an increase in sensitivity when sufficient time is allowed for the diffusion layer to relax following a scan. The presence of the Nafion film decreases adsorption at the surface, but still increases the sensitivity for dopamine. However, the voltammogram is more irreversible than at the bare electrode. Despite this, the Nafion film keeps the electrode surface in uniform condition so that the large background current is kept to a minimum allowing a clear observation of the faradaic events (Kristensen et al., 1987). This large background current is the major limitation to fast scan voltammetry. The charging current increases proportionally to the scan rate (Wipf et al., 1988); so most of the resolution of the digital oscilloscope is consumed by the higher charging current which accompanies all electroanalytical signals at high scan rates. The left over resolution does not

PAGE 41

27 accurately reflect the shape of the analyte signal because of the noise introduced by the lack of resolution of the digitized noise. This noise will prevent low concentrations of analyte from being detected even if the noise attributed to the cell and the current measuring circuit could be eliminated. The high frequency noise also becomes more difficult to eliminate as the scan rate increases, but this can be reduced by adequate circuit design, digital filtering or smoothing of the data. Characterization of Electrode Surfaces by X-rav Photoelectron Spectroscopy In X-ray photoelectron spectroscopy (XPS) or electron spectroscopy for chemical analysis (ESC A) photons from a monochromatic X-ray beam of known energy can displace electrons from atomic orbitals of atoms, ions or molecules or from bands of solids. The kinetic energy of discharged electrons (or the power of the electron beam) is plotted as a function of the energy (or the frequency or wavelength) of the emitted electrons. Hence, ESCA provides a means of qualitative identification of the elements present on the surface of solids since every element in the periodic table has one or more energy levels that will result in the appearance of peaks from 0 to 1250 eV binding energy in a low-resolution, wide-scan spectrum. Usually, the peaks are well resolved and lead to unambiguous identification if the element is present in concentrations greater than 0.1%. If one of the peaks is further analyzed, using a higher energy resolution, the surface environment can be characterized because the position of the maximum depends upon the chemical environment

PAGE 42

28 of the atom responsible for the peak. The variations in the number of valence electrons and the type of bonds they form influence the binding energies of core electrons. In general, binding energies increase as the oxidation state becomes more positive, because when an electron is removed, the effective charge on the core electron increases thus increasing the binding energy. In addition to giving qualitative information about the types of atoms present in a sample, ESCA can provide information on the relative number of each atom type as well as their oxidation state. For this reason, ESCA is widely used to characterize surfaces. Normally, only the top layer of the surface is analyzed, but with sputtering the elemental composition of the surface bulk can be analyzed. Sputtering allows the depth profile of the surface to be studied as it is being etched away by a beam of argon ions. This has been very useful in applications to corrosion chemistry, catalyst behavior and properties of semiconductors. In 1971 to 1981, Clark and Harrison investigated core-level binding energy shifts for atoms in polymers, but since then no major additions to this work have been made, largely due to tedious curve fitting. Since then computerized curve fitting has made ESCA a powerful analytical tool. As mentioned earlier, ESCA is favored for the characterization of surfaces, so with the development of a wide range of sensors based on polymer coatings, ESCA has become a widely accepted method of surface characterization. ESCA has been especially useful for the characterization of polypyrrole and overoxidized polypyrrole because polypyrrole films are difficult to characterize since they are insoluble and less crystalline (Street et al., 1982; Street et al., 1983) than other polymers

PAGE 43

29 (Pfluger and Street, 1984). The lack of structural data has hindered accurate band structure calculations and has complicated the interpretation of available data. Pfluger and Street have studied conducting polypyrrole grown on Pt substrates with thicknesses of 1-5 pm (Pfluger and Street, 1984). The pyrrole P carbons have a binding energy centered at 283.6 eV, and the a carbons have a binding energy at 284.5 eV. The main pyrrole peak for Nls appears at 399.6 eV, but the spectrum shows the presence of three inequivalent nitrogen sites. This indicates that the charge is fairly localized at the N sites. Also, the presence of counterions was detected and for polypyrrole/perchlorate polymer, the anion to ring ratio was 1:3, confirming that every third repeat unit has a negative charge. Ge et al. have studied OPPy grown on gold foil and overoxidized in perchlorate solution by ESCA (Ge et al., 1994). In comparison of the spectra of polypyrrole and OPPy, the Cl 2p signal at 207 eV was lost, the amount of high binding energy nitrogen was reduced significantly as the low binding energy component increased, and a new peak at 287.4 eV appeared on the high binding energy tail of Cls electrons at 284.6 eV with overoxidation. The composition of polypyrrole was C 5 3 Nj 0 O 037 (ClO 4 ) 024 , and the composition of OPPy was found to be C^N) qOj 58 (C10 4 ) 005 . This shows that the amount of O groups greatly increased as postulated by the formation of carbonyl groups during overoxidation (Beck et al., 1987) while the number of counteranions decreased significantly due to their expulsion from the film as the positive charge is lost. These results are consistent with previously proposed mechanisms of overoxidation (Beck et al., 1987). Palmisano et al. also characterized polypyrrole and OPPy structues by ESCA. Their membranes were grown on a Pt disk and were overoxidized in phosphate buffer (Palmisano

PAGE 44

30 et al., 1995), and the findings are in general agreement with the spectra obtained by Ge et al. The carbon to nitrogen ratio remained nearly constant in going from polypyrrole to OPPy and the oxygen content as measured by the oxygen to nitrogen ratio increased. The Cl 2p signal also decreased with overoxidation. The spectra also showed a small P2p signal possibly belonging to the buffer which suggested that positive charges were present in the OPPy structure. However, the residual positive N was not fully balanced by phosphate species considering the atom ratios, so partial formation and ionization of COOH groups might have occurred. Both a and P carbons are involved in these COOH groups, so this implied that the loss in conductivity during overoxidation was due to the breakage of polymer chains. This breakage of polymer chains suggests different kinds of regions in the film. The removal of positive charges on the polymer decreasing N + and increasing N=C creates hydrophobic regions, and the introduction of carbonyl and carboxylic groups makes the film more hydrophilic. Palmisano and coworkers developed a model based on this theory where the bulk polymer is hydrophobic with hydrophilic micropores which favor neutral species over anionic species. This was previously hypothesized by Witkowski and Brajter-Toth (Witkowski and Brajter-Toth, 1992). This theory was tested by measuring the permeability by rotating disk electrode experiments and was confirmed. Purpose of Work The focus of this work was to design an amperometric sensor for adenosine by modifiying carbon surfaces with templated ultrathin OPPy films. Adenosine is inherently

PAGE 45

31 difficult to dectect due to its large oxidation potential where the oxidation current can be obscured by large oxidation currents stemming from the background. Adenine, inosine and adenosine triphosphate, all structurally similar to adenosine, were used as template molecules and were incorporated into polypyrrole polymerziation to enhance the sensitivity and selectivity of the polymer film to adenosine. The modified electrodes enhanced the response of adenosine considering that no response was apparent at the bare surface. Despite this, the electrodes were not sensitive or selective enough to use as routine sensors for adenosine, so the majority of the work focussed on characterization of the electrode surfaces in order to understand and to control the response of the templated OPPy films. Scatchard and Langmuir isotherm analysis of the calibration data was performed to provide a model for the surface environment. The physical microstructure of the films was characterized by UV spectroscopy and x-ray photoelectron spectroscopy. Electrochemical studies were done to characterize film permeability. These included rotating disk electrode experiments to determine apparent diffusion coefficients of the probes in the films and log peak current vs log scan rate studies, to determine if the electrochemical processes were diffusion or adsorption controlled. Based on the measurements, a simple surface model was proposed. Carbon fiber electrodes were also modified with OPPy films to enhance the sensitivity to adenine and uric acid and to further characterize the films. The design of the ultramicroelectrode sensors were important for in vivo or cellular use, since the films are necessary to prevent fouling of the surface during the measurements. Finally, the OPPy films were used to provide a stable background and to suppress the

PAGE 46

32 large background currents present in fast scan voltammetry. Fast scan voltammetry was attempted in order to press the limits of detection for uric acid.

PAGE 47

CHAPTER 2 EXPERIMENTAL Reagents and Solutions HPLC grade acetonitrile (MeCN), certified ACS methanol (MeOH), tetrabutyl ammonium perchlorate (TBAP) and potassium phosphate monobasic were obtained from Fisher. Pyrrole (Py), dopamine, adenine, uric acid, adenosine, adenosine 5'-triphosphate (ATP) and inosine were from Sigma. Potassium ferricyanide (K 2 Fe(CN) 6 3 ’), potassium phosphate dibasic and ascorbic acid were from Mallinckrodt. Sodium perchlorate was from Aldrich. Ruthenium hexaamine (Ru(NH 3 ) 6 3+ ) was obtained from Johnson Matthey. All solutions were made with doubly distilled water, and all chemicals were used as received. Pyrrole was purified to obtain pure monomer by passing the monomer solution, which consists of dimers, trimers, etc., over activated alumina, but films formed from the purified pyrrole showed similar sensitivity to those formed from unpurified pyrrole, so pyrrole was used as received. Fresh or purified pyrrole solutions showed shorter polymerization times, but this did not seem to alter the characteristics of the ultrathin films tested here. Ru(NH 3 ) 6 3 , uric acid and adenine in 0.5 M pH 7.0 potassium phosphate buffer and in 0.5 M KC1 were used as probes to characterize the films. Ru(NH 3 ) 6 3+ , an electrochemically fast system, undergoes a one electron transfer at ca. E° =-0.290 V, and the analytical currents were measured at -0.350 V, the cathodic peak, vs SCE (Witkowski et al., 1991). Uric acid 33

PAGE 48

34 (Goyal et al., 1982) and adenine (Dryhust and Elving, 1968) are both slower electron/proton transfer systems. For uric acid, the currents were measured at the oxidation peak potential of ca. 0.350 V, and for adenine the currents were measured ca. 1.1 V vs SCE respectively. Except for at the bare electrode, no peaks in the cyclic voltammetric response were seen for adenine, so the currents were measured on the rising portion of the curve where the background was minimal. Electrodes Reference and Auxiliary Electrodes A saturated KC1 calomel electrode (SCE) or 5 cm long Ag wire was used as the reference electrode. Ag was used as a quasi reference electrode when MeCN was the solvent to avoid water contamination and to prevent liquid-liquid junction potentials (Sawyer and Roberts, 1974) and also during electrooxidation of PPy/ATP in the preparation of ATP solutions for analysis by UV to avoid contamination. When a conventional three electrode setup was necessary a 1 cm 2 or larger, depending on the the area of the working electrode, platinum foil electrode was used as the auxiliary electrode. Working Electrodes Glassy carbon electrodes (GC) were constructed from 3 mm and 5 mm diameter GC rods obtained from Electrosynthesis. The GC rod was cut into 1 cm in length pieces and each

PAGE 49

35 piece was sealed, cut side facing out, into the end of a glass rod using EpoxiPatch epoxy (Dexter Corporation). After drying overnight, the excess epoxy was sanded off the GC surface using 600 grit silicon carbide paper (Fisher). The GC disk was then polished to a mirror finish with Gamal y-alumina/water slurry on a microcloth with an Ecomet 1 polishing wheel (Beuhler), and sonicated in deionized water for one minute. Electrical contact was made to the unpolished side of the electrode using mercury and a piece of copper wire. Finally, the open end of the electrode was sealed with Teflon tape. Before modification, the GC electrodes were polished with the alumina slurry, and then sonicated for one minute. Electrode areas were determined by chronocoulometry by stepping the potential from 0.4 to -0.1 V with 3 x 103 M K 3 Fe(CN) 6 in 0.1 M KC1 (D 0 =7.63 x 10 * cm 2 /s; Stackelberg et al„ 1953). Typical GC electrode areas were 0.07 cm 2 . For rotating disk electrode (RDE) experiments, the RDE GC electrode tips were made by heat pressing a GC rod into a Teflon cylinder. Electrical contact to the GC was made with a Pt wire with silver epoxy (Type 410E, Epoxy Technology, Inc.). The GC RDE areas used were 0.06 and 0.23 cm 2 . The RDE electrode was polished as described for GC. A 2 cm rough pyrolytic graphite (RPG) (Electrosynthesis) electrode was used for the determination of ATP release from PPy/ATP by UV spectroscopy. RPG was chosen, as opposed to GC, since RPG can be machined easier and can be polished in the same manner as GC to produce a similar surface to GC. The RPG was sealed into a nylon block using EpoxiPatch epoxy and electrical contact was made by a copper wire attatched to the carbon via silver epoxy (EPO-TEK, 40E, Epoxy Technology). In ESCA experiments a 0.5 mm thick GC disk of 3mm diameter was glued to the

PAGE 50

36 ESCA sample holder using silver epoxy. To electrochemically modify GC attatched to the ESCA sample holder, electrical contact to the GC was made using a copper wire coated with silver epoxy, and a water-tight Teflon casing was designed to fit around the sample holder for electrical insulation of the ESCA sample holder as shown in Figure 2-1 . Before modification this GC electrode was polished as described above. Carbon fiber UMEs were made from ca. 7pm radius carbon fiber obtained from Textron specialty materials (Hsueh and Brajter-Toth, 1994). Fibers were glued to a copper wire using silver epoxy and were inserted into a micropipet tip or glass capillary. The copper wire was attatched to the side of the micropipet tip using EpoxiPatch. After letting the epoxy dry overnight, the micropipet tip was back filled with liquid epoxy (Epoxy-Shell EPON Resin 828, hardner-Metaphenylenediamine, both from Miller Stephenson Chemical Co.) and placed in an oven at 1 50 C for one hour to cure. To make the epoxy, both resin and hardner were slowly heated until transparent and viscous like water. After curing, the tip was sanded off using 600 grit silicon carbide paper and then Gamal y-alumina/water slurry on a polishing cloth to produce a carbon fiber disk. Prior to electrochemical analysis and modification with OPPy films, the carbon disk was polished with alumina on a polishing cloth like GC and was ultrasonicated for one minute. Electrode radius was verified from the limiting currents obtained using cyclic voltammetry.

PAGE 51

37 1.5 mm 9 mm Glassy carbon disk t 2.6 cm Figure 2-1 Diagram of ESC A sample holder with GC electrode and Teflon insulation

PAGE 52

38 INSTRUMENTATION Electrochemical Experiments A Bioanalytical Systems electrochemical analyzer (BAS100) was used in all electrochemical experiments except in fast scan measurements at scan rates from 0.005V/s to 0.500 V/s. The electrochemical data were downloaded to an IBM PS/2 Model 50 computer and analyzed using Grapher or Origins commercial programs. For use with UMEs a homemade current amplifier based on FaulknerÂ’s design (Huang et al., 1986), which allowed picoampere currents to be measured, was connected to the BAS and properly grounded (Hsueh and Brajter-Toth, in press). The input operational amplifier (AD5 1 5 A, Analog Devices) of the current transducer was a monolithic precision, low power, FET-input operational amplifier. The AD515A functioned as a current-to-voltage converter which amplified and converted the input currents to voltages. Various resistors (1, 10 and 100 MQ) and capacitors (1, 10 and 100 pF) in the feedback loop controlled the gains (100, 1000 and 10000) and RC time constants (1, 10 and 100 ps). Because the current-to-voltage converter inverted the signal phases, a second operational amplifier (OP27) with a unit gain was used as an inverter to invert the phase of the signal back to normal. Capacitors and a resistor on OP27 functioned as a first order filter to minimize noise in the circuit. The minimum current measurable by the BAS was 0. 1 pA, but the gain of the current amplifier ranged from 100 to 10,000, so the BAS with the preamplifier could measure

PAGE 53

39 currents as low as 1 00 picoamperes. The time constant of the potentiostat was controlled by the time constant of the first order filter, which was used to minimize electrical noise in the preamplifier circuit . At low scan rates, i.e. less than 40 V/s, the time constant of the system was set at 100 ps which allowed use of scan rates up to 40 V/s for one electron reactions, 20 V/s for two electron reaction, etc., with negligible distortion in the separation of peaks (Howell et al., 1986; Wipf et al., 1988). For fast scan experiments with scan rates from 100 V/s to 10000 V/s, a potential waveform from a function generator (EG&G Parc Model 175 Universal Programmer) was applied to a SCE reference electrode of a two electrode cell configuration, and the waveform was recorded at one channel of a digital oscilloscope (LeCroy 9310). A two electrode configuration was used since the current measured at the working electrode was small. Current at the working electrode was transduced to voltage by the current transducer of the preamplifier, amplified by OP AD515A and measured directly with the oscilloscope. The output of OP AD515A bypassed the OP27 and was directly connected to the oscilloscope. The oscilloscope measured the output voltage (the transduced current). The ratio of the transduced current to the input current of the working UME was determined by the feedback resistance of OP AD515A. The conversion factors for the potentiostat were 1,10 and 100 V/pA when the resistors in the amplifier were 1, 10 and 100 MO respectively. The RC time constant of the current transducer was controlled by the feedback resistance and the capacitance on OP AD515A. A 10 ps time constant would allow a scan rate of 400 V/s for one electron reactions with little distortion in the voltammogram, and alps time constant would permit a scan rate of 4000 V/s (Hsueh and Brajter-Toth, in press).

PAGE 54

40 Stored waveforms were transferred to a computer via a simple BASIC program (Hsueh and Brajter-Toth, submitted) for plotting cyclic voltammograms using Origins. The phase of the potential at the working electrode was reversed relative to the waveform potential at the reference electrode by the potentiostat (Hsueh and Brajter-Toth, in press), so the phase of the potential waveform measured with the oscilloscope had to be inverted. Also, the phase of the current at the working electrode was inverted by the inverting input of OP AD515A. To obtain a standard cyclic voltammogram, the phase of the potential waveform and the current stored in the digital oscilloscope were inverted using Origins before being plotted. Since the large charging current could obscure the faradic current at high scan rates, background subtraction had to be performed. For background subtraction, the background was measured in the absence of analyte in the supporting electrolyte alone, by cycling the electrode, and stored, and then the stored background was subtracted from the analyte current after completion of the analytical measurements. Signal averaging of 250 scans was used in both the background and analyte measurements before background subtraction. Generally, 250 scans for both the background and the analyte response was sufficient to obtain a good cyclic voltammogram. Other combinations such as a greater number of scans for the background were attempted, but these did not produce better results. Larger numbers of scans produced poorer resolution since more memory was taken up in the oscilloscope. In order to obtain accurate measurements, solutions were injected with a syringe into a microliter electrochemical cell, which allowed solutions to be pumped into the cell without moving the electrodes (Hsueh and Brajter-Toth, 1993).

PAGE 55

41 UV Absorption and ESCA Experiments UV absorption measurements were made with a Hewlett-Packard 8450A UV/Vis spectrophotometer. Matched quartz cells with a 1 cm path length were used. ESCA experiments were performed with a Kratos Xsam 800 spectrometer using AlK a excitation. Spectra were recorded for a survey scan with an energy window of 1 100 eV, various core levels with energy window of 20 eV or 40 eV, and the valence level with the energy window of 50 eV, using low magnification and high resolution, high magnification and low resolution, and high magnification and high resolution. The sample analyzer chamber pressure was kept at less than 1 x 10' 9 Torr. Fundamentals of Electrochemical Measurements Cyclic Voltammetry In cyclic voltammetry (CV) experiments, the potential is scanned linearly from an initial potential, where typically no Faradaic reaction of the analyte occurs, to a final potential where the reaction rate is limited by diffusion. The potential is then scanned linearly back to the starting potential. The rate of potential change is the scan rate, v (V/s), and the potential range between the initial and the final potential is the potential window, which depends upon the electrochemical properties of the analyte, the electrode and the solvent/electrolyte. Cyclic voltammograms are plotted as current vs potential. The current stemming from the redox reaction gradually increases past the potential where the reaction starts and reaches a

PAGE 56

42 maximum, called the peak potential (E p ). After reaching this potential, the current gradually decreases as the analyte diffusion controls the response of the electrode. The potential difference between the reduction and oxidation peaks (E^ and E pa respectively) is known as AE, The value of AE p can be used as an indicator of the reversibility of the electrode reaction. A one electron reaction at 25 °C is considered reversible if AE p is ca. 0.059 V; quasi-reversible if AE p is between 0.060 to 0.212 V and irreversible if AE p is greater than 0.212 V (Bard and Faulkner, 1980). The theoretical peak current, ip (A), for a diffusion controlled, reversible reaction can be written as follows (Bard and Faulkner, 1980): i p =(2.69x]0 5 )n m AD" 2 v m C‘ (2.1) and for an irreversible reaction: i p =(2.99xl 0 5 )«(a« a ) 1/2 ^D 0 1/2 v 1/2 C 0 * (2.2) where n is the number of electrons transferred per mole, A is the electrode area (cm 2 ), D 0 is the diffusion coefficient (cm 2 /s), v is the scan rate (V/s), C 0 * is the bulk concentration of the analyte (mol/cm 3 ), a is the transfer coefficient and n a is the number of electrons in the rate determining step. For adsorption controlled reactions, the peak current for a reversible reaction can be written as: n 2 F 2 A\T : _ O ART (2.3)

PAGE 57

43 and for an irreversible reaction: nan F 2 A\T j a o p 2.718 RT (2.4) where r o is the surface excess of the analyte (mol/cm 2 ), R is the gas constant (J moUKÂ’ 1 ), T is the temperature (K), F is FaradayÂ’s constant (C), and the remaining variables are the same as for the diffusional controlled processes. Cyclic voltammograms at macro electrodes (mm diameter) are typically peak-shaped, and the peak currents are proportional to the square root of the scan rate for a diffusion controlled system, and proportional to the scan rate for adsorption controlled reactions (equations 2. 1 and 2.2). For UMEs, the cyclic voltammograms are sigmoidal in shape and the limiting current measured at the plateau is independent of scan rate for scan rates up to 1 V/s due to radial diffusion. Under these conditions, the radius of an UME is small compared to the thickness of the diffusion layer. Because of the radial diffusion, steady state mass transport is attained at the electrode surface, and hence the current is time independent. As a result, the scan rate does not affect the shape and the size of the voltammetric wave. For a disk UME, the limiting current at steady state can be expressed by (Heinze, 1993): i p =4nFDC V ( 2 . 5 ) where n is the number of electrons transferred per mole, F is FaradayÂ’s constant (C), D is the diffusion coefficient (cm 2 /s), C is the bulk concentration of the analyte (mol/cm 3 ) and r is the electrode radius (cm).

PAGE 58

44 The plateau potential, which is the potential where the limiting current is reached, and the E 1/2 , the half-wave potential value, can be used to judge the difficulty or ease of electron transfer. For an oxidation process, a plateau and E 1/2 potential significantly more positive of E would indicate difficulty in electron transfer. As a note, in a reaction with very slow kinetics a well defined plateau of current is not observed. Using membrane coated UMEs, information about membrane structure can also be obtained since diffusion in solution should not the limit the response at time scales where diffusion layer thickness is much greater than the electrode radius. Hence, the response is determined by transport and other processes in the film (Cheng and Brajter-Toth, 1 992) and can give information about the film microstructure/microenvironment. As mentioned above, the sigmoidal shape current-potential curves for UMEs only exist at scan rates less than 1 V/s for ca. 5 pm radius (Heinze, 1993). At higher scan rates, this shape changes to peaks as with the macro electrodes, and at scan rates above 100 V/s, the shape is the same as with macro electrodes. Thus, the peak current equations for the macroelectrodes can be applied to the UMEs at high scan rates. Chronocoulometrv In chronocoulometry (CC) the potential is stepped from a potential where the rate of the redox reaction is negligible to a potential where the reaction rate is diffusion limited. The charge passed is monitored as a function of time (Q(t)) and is expressed as (Bard and Faulkner, 1980):

PAGE 59

45 Q(t> InFAD m C *t 1/2 (2.6) where n is the number of electrons per mole transferred, F is FaradayÂ’s constant, A is the electrode area (cm 2 ), D is the diffusion coefficient (cnr/s), C* is the bulk concentration of the analyte (mol/cm 3 ), and t is the pulse width (s). According to equation 2.6, a plot of Q(t) vs t 1/2 should give a straight line with a slope of 2nFAD 1/2 C /n l The value of the slope is used to determine electrode areas or diffusion coefficients. In this work, CC was used to find electrode areas of the macroelectrodes and to control the amount of pyrrole polymerized on the electrode surface by regulating the charge during polymerization. Rotating Disk Electrode Experiments Rotating disk electrodes (RDE) can be used to calculate the apparent diffusion coefficients of probes in films. The diffusion limiting current (ij is proportional to the square root of the rotation rate, co (s *) (Gough and Leypoldt, 1979) and a Koutecky-Levich plot of id 1 vs w ' 12 gives an intercept of i d _1 at infinite rotation rate (Gough and Leypoldt, 1979). At infinite rotation rate, the diffusion of the probe in solution becomes negligible and i d depends only on the difiiision of the probe through the film (Gough and Leypoldt, 1979, 1980; Leddy et al., 1985). For a membrane covered RDE, the membrane current, i d , can be written as.

PAGE 60

46 i,=nFAC *P =nFAC * c* m ( 2 . 7 ) where n is the number of electrons, F is FaradayÂ’s constant, A is the electrode area (cm 2 ), C* is the bulk concentration of the probe (M), P m is the permeability of the film (cm/s), D app is the apparent diffusion coefficient of the probe in the film and 5 m is the thickness of the film (cm). The D app values can be obtained from this equation since i d can be found from a Koutecky-Levich plot of the data, and the rest of the variables are typically known. This method can only be used to determine values for probes at macroelectrodes since RDE UME electrodes are difficult to fabricate, and the hydrodynamic processes at the small electrodes are difficult to control.

PAGE 61

CHAPTER 3 PREPARATION AND CHARACTERIZATION OF OPPy AND TEMPLATED OPPy ELECTRODES Procedure for Preparing Ultrat hin OPPy Films by Polymerization and Overoxidation of Polypyrrole on GC Preparation procedure for ultrathin OPPy films was based on a technique developed by Hsueh (Hsueh and Brajter-Toth, 1994). In this work, 0.020 M pyrrole was polymerized on GC from MeCN with 0.1 M TBAP at 0.950 V vs Ag wire. HsuehÂ’s procedure was modified in order to control the deposition charge at 35 pC/cm 2 which corresponds to a monolayer surface coverage (monolayer surface coverage is defined as 0. 15 nmol/cm 2 to 0.3 nmol/cm 2 (Murray, 1992)), based upon 2.25 electrons involved in the polymerization process (Diaz and Castillo, 1980). Polymerization was initiated by chronocoulometry, with a potential step from 0.650 to 0.900 V vs Ag wire (Bull et al., 1982). Typical polymerization times were ca. 50 ms. Polypyrrole was then overoxidized from 0.5 M potassium phosphate buffer of pH 7.0 at 0.950 V vs SCE as described by Hsueh (Hsueh and Brajter-Toth, 1994). Typical time for overoxidation was ca. five minutes. Formation of pinhole-free ultrathin OPPy films was performed using the procedure developed by Hsueh (Hsueh and Brajter-Toth, 1994). To form a pinhole-free ultrathin film, polypyrrole was first polymerized and then overoxidized. Since the overoxidized film was nonelectroactive, this process could be repeated, and only 47

PAGE 62

48 the pinholes and gaps in the film would be filled in by subsequent polymer deposition steps. This process was repeated until the response of 0.010 M Fe(CN) 6 3 ‘ in 0. 1 M KC1 or in pH 7.0 0.5 M potassium phosphate buffer was suppressed to background level after four to seven coatings as shown in Figure 3-1 . The choice of the electrolyte for Fe(CN) 6 3 ' detection did not influence the response of the OPPy film electrodes. The thickness of the films which suppressed Fe(CN) 6 3 ’ response and which were formed by repeating the polymerization/overoxidation procedure was roughly calculated from the total charge required to form the pinhole-free film, on average four times 3 5 pC/cm 2 , and from the diameter of pyrrole (ca. 4A), estimated from the bond lengths and the bond angles (Hsueh and Brajter-Toth, 1994). Since ca. four multilayers of polypyrrole (each layer 4A thick) were deposited, the films that were used in the majority of the experiments with the GC electrodes were ca. 16A thick. As mentioned in Chapter 2, pyrrole was used as received without any further purification. Pyrrole was purified to obtain only monomer solution by passing the solution over activated alumina. This purified pyrrole was then used to make an OPPy film electrode in order to compare the effect of purified and unpurified pyrrole on the film characteristics. The sensitivities of OPPy film electrodes made from unpurified pyrrole were similar to those formed from purified pyrrole. However, the time required to deposit charge when purified pyrrole was used was considerably shorter, especially when thick (>16A) films were being prepared. Otherwise, no difference was observed in the investigated properties of the ultrathin films with the use of the purified or unpurified pyrrole.

PAGE 63

current (^A) 49 60 bare GC 40 20 -20r -40 L -J 1 » L 400 300 200 100 potential (mV) _i L 0 Figure 3-1 Suppression of 10 mM Fe(CN) 6 3 ' response in 0.5 M pH 7.0 potassium phosphate buffer after repeated coatings of OPPy; scan rate 0.100 V/s, electrode area 0.07 cm 2 , deposition charge per coating ca. 35 pC/cm 2 , film thickness after 7th coating ca. 28 A.

PAGE 64

50 P rocedure for Preparing Templated O P Py Films on GC bv Polymerization and Overoxidation OPPy was templated with adenosine (OPPy/ado), inosine (OPPy/ino) and ATP (OPPy/ATP) by polymerizing 3 x 10Â’ 3 M polypyrrole (PPy) at GC from 80% MeOH, 0. 1 M TBAP with 3 x 1 0 3 M adenosine or inosine to form first PPy/ado and PPy/ino respectively. To obtain PPy/ ATP, 3 x 10 3 M pyrrole was electropolymerized from water with 0.010 M ATP. MeOH/water solutions and water were used in the formation of the templated polymers to accomodate the solubility of the templates. The scheme proposed for incorporation of the templates is illustrated in Figure 3-2. Note that the exact orientation of the pyrrole monomers is unknown, but Nishizawa et al. found that lateral growth was promoted at hydrophobic substrates (Nishizawa et al., 1991). As shown in Figure 3-2, adenosine can be incorporated during the polymerization of pyrrole by weak interactions with the forming polypyrrole, such as hydrogen bonding or hydrophobic interactions. In contrast, ATP can undergo stronger interactions such as electrostatic as well as hydrogen bonding and hydrophobic interactions with the forming polymer. Typical dopant levels for polypyrrole are one counter ion for every three to four pyrrole units (Mitchell et al., 1988). Larger concentrations than the typical dopant levels were used to ensure incorporation of adenosine and inosine because of their neutrality, and to also use ATP as the supporting electrolyte. A polymerization potential of 0.400 V vs SCE as suggested by Ko et al. for aqueous solutions (Ko et al., 1990) was applied to deliver a charge of 35 pC/cm 2 in the formation of all templated polypyrrole structures in all mixed/aqueous solvents. Polypyrrole was overoxidized to form OPPy/ado, OPPy/ino and OPPy/ATP at 0.950 V vs SCE, as were the

PAGE 65

^ a: 51 A denosine Figure 3-2 Cartoon representation of template, adenosine or ATP, incorporation into polypyrrole

PAGE 66

52 pyrrole films without the templates, with ca. five minutes required for the current to reach steady state. These films overoxidized in approximately the same time as those formed without the templates. During overoxidation, the template molecules were expected to be expelled from the oxidized polypyrrole due to the removal of the net positive charge of the polymer, requiring expulsion of the charge balancing ions, and due to the resulting unfavorable interactions with the carbonyl groups in the newly forming OPPy structure. The newly formed OPPy structure without the templates is depicted in Figure 3-3. The polymerization/overoxidation of the templated polymer was typically repeated four to seven times until the response of 0.010 M Fe(CN) 6 3 ' in 0.5 M pH 7.0 potassium phosphate buffer was suppressed to a background level. No apparent difference in the suppression of Fe(CN) 6 3 , in comparison to the films formed without the templates, was noted. However, in some instances with the templated films, the first coating seemed to suppress Fe(CN) 6 3 ' response more. In addition, polypyrrole was polymerized/overoxidized at GC without the templates but in the same solvents as used during the formation of the templated polypyrrole to assess the effect of the solvent on film formation and properties. In 80% MeOH, 3 x 1 0' 3 M pyrrole was polymerized with 0. 1 M TBAP; in water 3 x 10‘ 3 M pyrrole and 0.010 M NaC10 4 was used. These OPPy films were prepared using the polymerization and overoxidation conditions described above for the templated films. In order to analyze the effect of film thickness on response and interpret the data from macroelectrodes compared to the data obtained at carbon fiber UMEs, films with thicknesses of 32, 44 and 48 A were prepared at GC to investigate the effect of film thickness on sensitivity. For a 32 A film a charge of 84pC/cm 2 was applied per coating, and for the 44 and

PAGE 67

53 Figure 3-3 Cartoon representation of the templated OPPy structure after overoxidation of polypyrrole (Beck et al., 1987)

PAGE 68

54 o and 48 A films charges of 124 and 133 pC/cm 2 were applied per coating respectively. Polymerization times generally required ca. 15 to 20s which was much longer than the millisecond time required for the ultrathin films. The thicker polypyrrole films were overoxidized in 0.5 M pH 7.0 potassium phosphate buffer at 0.950 V vs SCE until the current reached a steady state, ca. ten minutes. These thicker films required approximately twice longer overoxidation time than the ca. 16A films. Polymerization and overoxidation steps were repeated as described previously to fill in the pinholes in the films. Typically three coatings were needed to suppress the response of 10 mM Fe(CN) 6 3 ‘ in 0.5 M pH 7.0 potassium phosphate buffer. Procedure for Pre paring Ultrathin OPPv and Templated OPPv films at UMEs Preparation of OPPy films at carbon fiber UMEs was based on the work of Hsueh (Hsueh and Brajter-Toth, 1994). The same conditions as used for GC electrodes were used for the UMEs, except the charge applied was 31.7 mC/cm 2 per coating (three coatings on average) which corresponds to a thickness of ca. 0.53 pm, based on a charge of 24 mC/cm 2 resulting in a 0. 1 pm thick film (Diaz and Castillo, 1980). Hsueh’s films were ca. 32 A thick, but this thickness was difficult to duplicate because very short times were required for the deposition due to the high current density at the UMEs from edge effect. Polymerization generally required three milliseconds for each coating, and overoxidation required less than one minute, as compared to ca. fifty milliseconds for polymerization and ca. five minutes for overoxidation at GC. Suppression of the response of 0.010 M Fe(CN) 6 3 ' in 0.5 M pH 7.0

PAGE 69

55 potassium phosphate buffer was typically achieved after three to four coatings as shown in Figure 3-4. This was shorter compared to the films formed at the macoelectrodes since up to seven coatings could have been used, possibly due to the thicker films at the UMEs. Verification of Template Release During Overoxidation of Templated Polvpvrrole A calibration plot of absorbance vs concentration was prepared in the concentration range 5. Ox 10 to 1. 0x1 O' 6 M ATP to measure the ATP release from polypyrrole grown on RPG, into 0.1M pH 7.0 potassium phosphate buffer. Potassium phosphate buffer concentration of 0. 1 M was used because the use of 0.5 M pH 7.0 potassium phosphate buffer produced a high background in the UV absorption spectra. The molar absorptivity of 1 .41 x 10 M cm ‘of ATP was determined at X mix =260 nm in agreement with the literature value of e=1.54 x 10 4 M ‘cm 1 (Pyo et al., 1994). To measure the amount of ATP released from the PPy templated with ATP, a 2.0 cm 2 square RPG electrode encased in nylon and sealed with Epoxipatch epoxy was used as the substrate for the deposition of the ATP templated polypyrrole film. RPG was used because a large electrode area was needed so that a sufficient amount of ATP could be incorporated into polypyrrole in a reasonable amount of time, and RPG, if not roughened, has a similar structure to GC. The RPG electrode was polished with alumina on a polishing cloth and sonicated for one minute prior to the polymerization as decribed for GC in Chapter 2. A 0.020 M ATP solution with 0.007 M pyrrole in water was used as the polymerization solution. The surface coverage of polypyrrole was 5 x 10' 7 mol/cm 2 as determined from the deposition charge (0.108 C/cm 2 ), which

PAGE 70

56 — 1 — 1 1 i 1 i I , L 400 300 200 100 0 potential (mV) Figure 3-4 Suppression of 0.010 M Fe(CN) 6 3 ' in 0.5 M pH 7.0 potassium phosphate buffer after repeated coatings of OPPy at carbon fiber UME; radius 7 pm, scan rate 0. 100 V/s, deposition charge 3. 17 x 10' 2 C/cm 2 per coating, film thickness after 4th coating ca. 0.53 pm

PAGE 71

57 corresponded to a film thickness of ca. 0.45 pm based on 24 mC/cm 2 resulting in a 0. 1 pm thick film (Diaz and Castillo, 1980). This thickness required a polymerization time of ca. ten minutes which was achieved by bulk electrolysis at 0.900 V vs SCE. To determine the amount of ATP incorporated during the polymerization of polypyrrole and released during the overoxidation of the film, polypyrrole was overoxidized in 0.1 M pH 7.0 potassium phosphate buffer at 0.950 V vs SCE until the current decayed to a steady state value, ca. thirty minutes. To verify template incorporation into polypyrrole and template release (Li and Dong, 1992) during polypyrrole overoxidation, ATP template concentration was monitored at ^m®t = 260 nm in the overoxidation solution. A thicker than normal PPy/ATP film (0.45 pm) was formed to allow detection of ATP above the micromolar limit of detection for ATP set by the UV method. The amount of ATP that was detected in the overoxidation solution corresponded to a molar ratio of 7.7:1 of pyrrole to ATP, with the expected ratio of 9:1, based on +0.33 charge/pyrrole unit in polypyrrole (Beck et al., 1987) and a -3 charge on ATP. The 7.7:1 molar ratio confirmed ATP incorporation into polypyrrole and the release of the template during the overoxidation of polypyrrole to OPPy in nearly a stoichiometric amount. Since the polymerization solution consisting of ATP, pyrrole and water had pH 3.0, and ATP has pK a of 4. 1 (H 2 ATP 2 ' H + + HATP 3 ') and 6.95 (HATP 3 ' * H + + ATP 4 '), then at pH 3 ca. 9 % ATP should have a charge of -3 (Zubay, 1988). Therefore differences in the experimental and the theoretical value of the ratio of pyrrole to ATP may be due to the greater presence of H 2 ATP 2 ', which would theoretically give a molar ratio of 6: 1.

PAGE 72

58 ESC A Analysis of Bare GC. OPPv and OPPv/ATP Films on GC Initial ESCA spectra were taken of the GC surface alone in order to obtain a background level of carbon and oxygen atoms and their ratios present before polymerization. GC was polished with alumina on a polishing cloth and sonicated for one minute before the spectra were run. Spectra were recorded to obtain a wide scan spectrum or a survey scan (binding energy window 0 to 1 100 eV) as illustrated in Figure 3-6, various core energy levels for different atoms (binding energy window 20 eV or 40 eV) as shown in Figures 3-9, 3-10, 3-12 and 3-13 and a spectrum of valence levels (binding energy window 0 to 50 eV) as pictured in Figure 3-5 using low magnification and high resolution, high magnification and low resolution, and high magnification and high resolution. The spectra of the bare GC indicated the presence of C u , O l5 , Ag 3d , and Si^^^ The silicon present was due to silica; silica has been observed at other types of carbon surfaces, such as carbon foil and carbon molecular seives (Goodfellow, 1990). Silicon most likely segregated to the surface from heating or ion bombardment during ESCA analysis, or normal aging of the material. From the survey scan of the bare GC surface, the C/O area ratio was 1.5. After the electrode was etched with Ar (4 keV, IgA) for two hours, the C/O area ratio was 1.7. Therefore, the etching removed some of the oxygenated carbon from the surface even though the ratios did not change significantly. This also showed that GC could possibly be used as a substrate in ESCA analysis of OPPy films since the C/O area ratios remained fairly constant at the carbon surface with etching.

PAGE 73

59 Next, the GC electrode was modified with ultrathin OPPy by first polishing with alumina on a polishing cloth, sonicating for one minute, and then polymerizing 0.020M pyrrole from MeCN with 0. 1 M TBAP at 0.950 vs Ag wire. The polypyrrole electrode was then overoxidized from 0.5 M pH 7.0 potassium phosphate buffer at 0.950V vs SCE. This process was repeated until the response of 0.010 M Fe(CN) 6 3 ' in 0.5 M potassium phosphate buffer pH 7.0 was not suppressed any further. For the electrode studied by ESC A, three polymerization/overoxidation steps were needed. Fe(CN) 6 3 ' response was not suppressed to the background level, but the fourth coating did not suppress the response any further than the third coating did, so the process was not repeated. The behavior was similar to that observed in other experiments with GC electrodes. After coating with the OPPy film, the electrode was rinsed with deionized water and stored in vacuum until ESCA analysis was performed. The film thickness was ca. 12 A as calculated from the deposition charge of 35 pC/cm per coating (3 coatings applied) which corresponds to a monolayer coverage, 2.25 as the number of electrons involved in the process and 4 A as the diameter of pyrrole. As shown in Figure 3-5, the ESCA valence band spectra for bare GC and OPPy modified GC are significantly different, indicating variations in the chemistry of the two surfaces. A wide scan ESCA spectrum taken on GC covered with ultrathin OPPy (film thickness ca. 12 A) as shown in Figure 3-6, clearly illustrates the presence of oxygen, carbon and nitrogen. Carbon and oxygen were expected since they are present at the GC surface, but the presence of nitrogen indicated the presence of pyrrole or TBAP on the surface. By comparing the C/Si area ratios for bare GC and OPPy, it was concluded that ca. 13 % of the electrode was not covered with the film. This was reasonable since more hydrophilic regions

PAGE 74

60 OPPy modified GC electrode Binding Energy (eV) Figure 3-5 ESC A valence band spectra of bare and OPPy modified GC electrodes film thickness ca. 12 A

PAGE 75

Figure 3-6 Wide scan spectrum of OPPy film on GC (ca. 12A thick) showing the presence of oxygen, carbon and nitrogen

PAGE 76

62 on GC may not support the growth of polypyrrole, and in constructing the polymer only three repeated processes of polymerization and overoxidation were performed which did not completely suppress Fe(CN) 6 3 '. In general, the film coated regions charged more, as expected for a non-conductor, than the bare carbon which was the region that charged the least. For identification of the film on the electrode surface, the types of nitrogen on the surface were of interest. The nitrogen spectra showed charge corrected peaks at 397.8 eV and 400.7 eV. According to Pfluger et al. and Ge et al., who studied OPPy films, the peak at 397.8 eV was due to =Nfrom the OPPy structure as shown in Figure 3-3 (Pfluger et al., 1983; Ge et al., 1994). Although no polymer structures were proposed by either group, the presence of this nitrogen was consistent with the structures proposed by Beck as shown in Figure 3-7 (Beck et al., 1987). The peak at 400.7 eV could be a nitrogen with a partial positive charge as suggested by Pfluger et al. (Pfluger et al., 1983; Pfluger and Street, 1984). Ge et al. attributed this peak to -NH + on polypyrrole as shown in Figure 3-8 (Ge et al., 1994). Other peaks in the ESCA spectra were from carbon and oxygen as shown in Figure 3-6, phosphorous, chloride, silicon, aluminum ( Figures 3-9 and 3-10), and potassium. The presence of silica from the low charging region was attributed to exposed carbon from defects in the film, and aluminum was identified as a contaminate from the polishing of GC with alumina. The presence of phosphorous and potassium was likely due to H 2 P0 4 ‘ and HP0 4 2 ' from the potassium phosphate buffer used to overoxidize polypyrrole. The presence of HP0 4 was speculated to produce the low binding energy phosphorous component, and the presence of H 2 P0 4 .which was present in a larger quantity, was believed to be the high binding energy phosphorous component. Based on the C/P area ratios of the 2s (area ratio

PAGE 77

63 Figure 3-7 Structure of overoxidized polypyrrole as proposed by Beck et al. (1987)

PAGE 78

64 t OH ' t Figure 3-8 Structure of OPPy as proposed by Ge et al. (1994)

PAGE 79

65 Figure 3-9 ESCA spectrum of chloride and phosphorous from OPPy film (ca 12 A) on GC

PAGE 80

66 Figure 3-10 ESC A spectrum of phosphorous, aluminum and silicon of OPPy film (ca. 12 A thick) on GC

PAGE 81

67 is 2.8) and 2p (area ratio is 2.3) spectra of phosphorous and the K a value of H 2 P0 4 '(6.32 x 1 O' 8 ; Harris, 1991), the pH in the film was calculated to be ca. 6.8 as compared to the measured overoxidation solution pH of 7.0, so a reasonable amount of the ions from the phosphate buffer must have been trapped in the film. During overoxidation of pyrrole there is a movement of counterions due to the loss of charge on the polymer and the formation of carbonyl groups. Counterions like H 2 P0 4 ' or HP0 4 2 ' could be trapped in the film. The movement of ions during overoxidation is illustrated in Figure 3-1 1 . Figure 3-11 shows the oxidation mechanism of polypyrrole to overoxidized polypyrrole in aqueous solution as proposed by Beck et al. Since -NH + could be present in the film as postulated by Ge et al. (Ge et al., 1994), the detected presence of the phosphate ions may be for charge balancing purposes. Cations in the films could also exist to balance the charge of any anions in the film. Chloride found in the films could be from the TBAP. The C/K area ratio was 83, and the C/Cl area ratio was 34, indicating more chloride was present in the film. Cl' may have served to balance charge in the film from the presence of -NH + groups. Hence, these results confirmed that nitrogen containing films were present on the GC surface from the nitrogen peaks and a new C/O ratio of 4.5 compared to 1.5 at bare GC. If an OPPy film was present at the surface, the film must be porous due to the presence of several types of ions. Finally the polymerization/overoxidation process produced a well covered surface nearly free of pinholes, pointing to a polymer film rather than a TBAP covered surface. In order to investigate the effect of templating on the surface microenvironment, GC was modified with ATP templated OPPy using the procedure described previously in this

PAGE 82

68 Figure 3-11 Mechanism of polypyrrole overoxidation in water (Beck et al., 1987)

PAGE 83

69 chapter. Polymerization from a 0.003 M pyrrole/0.010 M ATP solution was performed at 0.400 V vs SCE. Polypyrrole was then overoxidized at 0.950 V vs SCE in 0.5 M pH 7.0 potassium phosphate buffer. Four coatings were needed to suppress the response of 0.010 M Fe(CN) 6 3 ' in 0.500 M pH 7.0 potassium phosphate buffer. The OPPy/ATP film was ca. 16 A thick. The electrode was rinsed with deionized water and dried before ESCA analysis. The OPPy/ATP film also showed differential charging (different binding energies for one atom) similar to the OPPy film, as indicated by two sets of silver 3d peaks. No silver was observed in the least charging region. Ag was likely from the silver epoxy used to attatch the GC to the holder before film modification. The OPPy film had three different types of regions indicated from three sets of silver 3d peaks. The area ratio for C/Si (Si spectrum shown in Figure 3-12) indicated that ca. 10% of the electrode was not covered with the film since the C/Si area ratio for film covered and bare GC were 1.2 and 9.4 respectively. The percentages of uncovered GC was similar to that for the nontemplated OPPy film, indicating that the ATP template or somewhat thicker film did not significantly affect coverage. The ESCA valence band spectrum for OPPy/ATP illustrated in Figure 3-13 showed differences and similarities to the valence band spectra for the OPPy film shown in Figure 3-5. Table 3-1 summarizes the area ratios of carbon to selected atoms for OPPy and OPPy/ATP electrodes from obtained spectra. The differences in the area ratios indicated that the surface microenvironments were significantly different. The C/N area ratio for OPPy/ATP was 190 in comparison to 33 for OPPy, which indicated fewer nitrogens on the surface. Hence, polymerization with the template altered the film structure as observed from ESCA analysis.

PAGE 84

70 Figure 3-12 ESCA spectrum of aluminum and silicon in OPPy/ATP films (ca. 16A thick) on GC

PAGE 85

71 Figure 3-13 ESC A valence band spectrum of OPPy/ATP film (ca. 16A thick) on GC

PAGE 86

72 Table 3-1 Comparison of Various Atom Area Ratios obtained from ESCA spectra for Bare and OPPy and OPPy/ATP modified GC electrodes area ratio C IX observed binding energy (eV) bare GC electrodeÂ’ OPPy/GC electrodeÂ’ OPPy/ATP/GC electrodeÂ’ C/N 405.9 + 408.8 b 33 190 C/P (HP0 4 2 -) 195.0 b 101 79 C/P (H 2 P0 4 ) 200.0 b 36 101 C/K 301.0 b 83 b C/Ca 356.9 b b 147 C/Cl 207.4 b 34 49 C/Al 76.0 b 41 41 C/Si 111.2 9.4 56 57 C/0 539 + 541 1.5 4.5 5.5 Note: area normalized for time of scan and size of window, C IX = (area C/oJ^area X/oJ, where o is the Scofield cross section (Scofield, 1976) Â’electrode area 0.07 cm 2 , film thickness ca. 12A at OPPy and ca. 16 A at OPPy/ATP b undetected

PAGE 87

73 Phosphorous was present in the OPPy/ATP film as shown in Figure 3-14, but more HP0 4 2 ’ (area C/P=79) than H^O; (area C/P=101) was present than in the OPPy film, where significantly more H 2 P0 4 " (area C/P=36) was found. Potassium was not detected in the OPPy/ATP film, but small amounts of calcium (area C/Ca=147), shown in Figure 3-15, was present. Calcium impurity originated from the ATP since it was not present in the OPPy film, and more than likely, calcium replaced the K + because Ca 2+ can more effectively balance the charge on HP0 4 2 ‘. Since Ca 2+ was from the ATP, it could be postulated that ATP was present initially in the film. Carbon, oxygen and nitrogen ratios did not indicate that ATP was present after oxidation. The OPPy/ATP film also contained chloride, probably from the SCE, but the OPPy/ATP film had less CT than OPPy which was expected since C1‘ was not present during the polymerization. Based on the results, the OPPy/ATP film was concluded to have a different microenvironment than the OPPy film. The microstructure of the OPPy/ ATP film may be more compact since fewer ions were found at the surface. However, this could also be a result of less polypyrrole on the surface due to the initial presence of ATP ions in spite of initially greater thickness. ATP ions are much larger than pyrrole and could act as spacers in the structure during the polymerization. Since there was less polypyrrole, as attributed from lower amounts of N atoms, ATP must have altered the structure of the film at the GC electrode surface as expected when incorporated into the ultrathin film.

PAGE 88

74 Figure 3-14 ESCA spectrum of chloride and phosphorous in OPPy/ATP film on GC (ca. 16A thick)

PAGE 89

75 370 365 360 355 Binding Energy ( eV) Figure 3-15 ESCA spectrum of calcium in OPPy/ATP film on GC (ca. 16A thick)

PAGE 90

CHAPTER 4 ANALYSIS OF THE ELECTROCHEMICAL RESPONSE AT GLASSY CARBON, GLASSY CARBON COATED WITH OPPy AND TEMPLATED OPPy Sensitivity Data at Glassy Carbon and Glassy Carbon Coated with Ultrathin Films of OPPy and Templated OPPy This chapter will describe the properties of ultrathin OPPy membranes templated with adenosine, inosine and ATP. Ru(NH 3 ) 6 3+ , uric acid and adenine were used to characterize the voltammetric response of the films (ca. 16A) and were chosen based on their electrochemical reactivity, charge and ability to undergo hydrophilic and hydrophobic interactions with GC and the membrane. Based on the results of Witkowski et al., thick OPPy films grown on GC were expected to be relatively compact (Witkowski et al., 1991). Ultrathin OPPy films were prepared here to detect changes in the film interactions from templating with good sensitivity. Good sensitivity was expected for Ru(NH 3 ) 6 3+ at OPPy coated GC due to the formation of carbonyl groups in the OPPy structure during polypyrrole overoxidation. These carbonyl groups have a high electron density which should favor their interactions with cations. In contrast, uric acid should have lower sensitivity at OPPy membrane electrodes because of the negative charge density of the carbonyl groups, especially at pH 7.0 when uric acid is negatively charged (pK a =5.4, 1 1.3) (Brown, 1991). The high electron density from the carbonyl groups on OPPy has been shown to exclude anions as with the response of 76

PAGE 91

77 Fe(CN) 6 3 ' shown in Figure 3-1. Uric acid may not be excluded as well as Fe(CN) 6 3 ' since the charge on uric acid is smaller and delocalized about the structure, and uric acid is capable of hydrophobic interactions with the OPPy backbone. With -NH + groups in the film, as possibly indicated from ESCA analysis, uric acid may show higher sensitivity than expected if the OPPy structure consisted only of repelling carbonyl groups. Reasonable sensitivity at OPPy membrane electrodes was expected for neutral adenine (pK a =2.0, 4.1, 9.8) (Brown, 1991) which can interact hydrophobically with the OPPy backbone. The cylic voltammetric reponse of each probe at GC, OPPy and templated OPPy film coated electrodes (ca. 16A thickness) is shown in Figures 4-1 through 4-3. The cyclic voltammograms (CVs) for 0.4 x 10' 3 M Ru(NH 3 ) 6 3+ (Figure 4-1) were peak shaped, and the kinetics were quasireversible with AE p values of 75 for GC, 75 for OPPy, 99 for OPPy/ado, 101 for OPPy/ino and 91 for OPPy/ATP electrodes. The cathodic peak potentials of -3 12 mV for OPPy, -333 mV for OPPy/ado and OPPy/ino and -328 mV for OPPyATP were not significantly shifted for the OPPy and templated OPPy film electrodes compared to the cathodic peak potential at the bare GC electrode of -326mV which indicated that the ultrathin films did not significantly alter the response of Ru(NH 3 ) 6 3+ . In contrast, the CVs for 0.2 x 10' 3 M uric acid (Figure 4-2) and 0.3 x 10‘ 3 M adenine (Figure 4-3) were irreversible as evidenced by no return peak on the reduction and a AE p value greater than 30 mV in the case of uric acid at the coated GC. In general, the CVs for uric acid were peak shaped with the exception of OPPy/ino and OPPy/ATP where a plateau was observed while the CVs of adenine did not show peaks, except at GC.

PAGE 92

78 potential (mV) Figure 4-1 Cyclic voltammograms of 0.4 mM Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s.

PAGE 93

current (pA) 79 0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 500 400 300 200 potential (mV) _l i L 100 0 Figure 4-2 Cyclic voltammograms of 0.2 mM uric acid in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s.

PAGE 94

80 =3 O -12 I I I i I i 1 1 1 1 L 1200 1000 800 600 400 200 0 potential (mV) Figure 4-3 Cyclic voltammograms for 0.3 mM adenine in 0.5 M pH 7.0 potassium phosphate buffer at GC, OPPy and templated OPPy electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16A, scan rate 0.020 V/s.

PAGE 95

81 The current for each probe was measured at the peak, or in the case of adenine, at a potential where the response was greater than the background response, but where the background currents were not too high to sacrifice sensitivity. The potentials used in the measurements were ca. -0.30V vs SCE for Ru(NH 3 ) 6 3+ , 0.35V vs SCE for uric acid and 1.0V vs SCE for adenine. All measurements of analyte current were background subtracted by first measuring the background currents (at least three measurements were taken), averaging the background currents and then subtracting them from each analyte signal. The analyte currents used to calculate sensitivities were based on averaged results of at least three analyte measurements. Sensitivities at bare GC and OPPy modified GC electrodes were obtained as the slope of the current vs concentration plots for each respective electrode. The slope was calculated by regressing the data through (0,0) because in theory since the background currents were subtracted from the analyte currents, the current should be close to zero at zero concentration of analyte. The sensitivity results, based on the fits of the data through the origin, for the three probes are summarized in Table 4-1. Regression analysis of the data not through the origin was also calculated for comparison, and the results are listed in Table 4-2. Typical intercept values of less than ±l|iA indicated that fitting the data through the origin produced a small error in most cases. In addition, theoretical sensitivities calculated from equation 2. 1 for Ru(NH 3 ) 6 3+ for a reversible system, and from equation 2.2 for uric acid for an irreversible system (Bard and Faulkner, 1980), are listed. For adenine, the sensitivity was estimated for a two electron irreversible system, since the exact oxidation mechanism at pH 7.0 is still unknown, and in acidic solution up to six electrons can be transferred during the oxidation (Dryhurst and Elving, 1968).

PAGE 96

82 Table 4-1 Sensitivity of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine at Different Surfaces Ru(NH 3 ) 6 3+ Uric Acid Adenine peak potential (mV) -300 ~ ±350 ~ ±1100 pK„ b 5.4, 11.3 <2.0, 4.1, 9.8 theoretical sensitivity' 2.7 4.2 4.2 bare GC a 6.0 ±0.5 14.8 ±0.9 15 ± 1 OPPy*d 10 ± 4‘ 6.7 ±0.5 15 ± 2 OPPy/adenosine a,d 6.6 ±0.5 8.0 ±0.3 11.9 ±0.1 OPPy/inosine a,d 6.8 ±0.1 8.6 ±0.1 9 ± 2 OPPy/ATP a,d 4.7 ±0.2 5.1 ±0.3 9 ± 1 Note: sensitivity (gA/mM) as a result of linear regression of the data through the origin ‘electrode area ca. 0.07 cm 2 , scan rate 0.020 V/s, analyte concentrations 0.1-1 x 10' 3 M in 0.5 M pH 7.0 potassium phosphate buffer, all potentials vs SCE b (Brown, 1991) ‘theoretical current for Ru(NH 3 ) 6 3+ , equation 2.1, for uric acid and adenine, equation 2.2, D 0 estimated at lO" 6 cm 2 /s for each probe d film thickness ca. 16 A ‘obtained from non (0,0) regression analysis of the data

PAGE 97

83 Table 4-2 Non (0,0) fits of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine Calibration Plots at Different Surfaces electrode 3 Ru(NH 3 ) 6 3+b Uric Acid b bare GC y=3.9(±0.2)x+1.2(±0.1) y=16.7(±0.8)x -0.2(±0. 1) OPPy c y r =10(±4)x+2(±3) y=7 . 6(±0 . 6)x-0 . 1 (±0 . 1 ) OPPy/ado c y=4.6(±0.5)x+1.2(±0.3) y=8 . 6(±0 . 4)x-0 . 0 8 (±0 .04) OPPy/ino c y=6 . 3 (±0 . 2)x+0 . 3 (±0 . 1 ) y=8.6(±0.1)x-0.08(±0.07) OPPy/ ATP c y=4 . 0(±0 . 5 )x+0 . 4 (±0 . 3 ) y=4.7(±0.3)x+0.09(±0.05) electrode a adenine b bare GC y= 1 8 (±2)x1 . 0(±0 . 6) OPPy c y=20(±8)x-l(±2) OPPy/ado c y=12.2(±0.6)x-0.1(±0.1) OPPy/ino c y=3.01(±l)x+l(±l) OPPy/ATP c y=2.7(±0.5)x+l ,4(±0. 1) Note: slope (pA/mM) and intercept (pA) a electrode area 0.07 cm 2 , scan rate 0.020 V/s, all potentials vs SCE b analyte concentrations 0.1-1 x 10" 3 Min0.5M pH 7.0 potassium phosphate buffer o c film thickness ca. 16A

PAGE 98

84 For all the probes the sensitivity at bare GC, in slow scan voltammetry, was higher than the theoretical sensitivity. Ru(NH 3 ) 6 3+ can interact with the surface oxides (Kovach et al., 1986), and both uric acid (Dryhurst and De, 1972) and adenine (Dryhurst, 1972) have been reported to adsorb on GC. The adsorption interactions can contribute to the increased sensitivity because of the resulting preconcentration which gives a higher concentration of analyte at the surface. The higher than theoretical sensitivity for adenine may be due in part to a larger number of electrons in the oxidation (Dryhurst and Elving, 1968) than used in the calculation. At ca. 16 A thick OPPy film electrodes Ru(NH 3 ) 6 3+ sensitivity increased (-37%) compared to the sensitivity at the bare GC. The sensitivity of uric acid was lower (-55%) than at the bare GC but remained higher (-60%) than the theoretical sensitivity. The result for uric acid showed that at OPPy film electrodes the response of uric acid was not as efficiently suppressed as the response of Fe(CN) 6 3 " probably as a result of a smaller negative charge on uric acid and possible hydrophobic interactions of uric acid with the film and electrode. If -NH + groups were present in the OPPy film as indicated by ESCA analysis discussed in Chapter 3, then the reponse of uric acid could also be due to favorable interactions with these groups. Since Fe(CN) 6 3 ' should also interact favorably with any -NH + groups present, exclusion of Fe(CN) 6 3 ‘ indicated that either multiple interactions between the probe and the film, e.g. hydrophobic and hydrophilic, contributed to the response or the density of the -NH + groups was low. Adenine sensitivity did not change at OPPy coated GC compared to the sensitivity at the bare GC indicating that the surface processes important in adenosine detection were not significantly changed by the presence of the ultrathin OPPy film.

PAGE 99

85 The high sensitivity in slow scan voltammetry at ultrathin film electrodes showed (Table 4-1) that at the highly permeable ultrathin films, slow in-film transport, which has been reported at thicker OPPy electrodes (Witkowski and Brajter-Toth, 1992), did not limit the response. For example, the high sensitivity of Ru(NH 3 ) 6 3+ was not influenced by slow diffusion observed in thicker films (D app =2.8 x 10' 8 cm 2 /s, Witkowski et al., 1991), indicating that the high sensitivity must be influenced by the favorable interactions of Ru(NH 3 ) 6 3+ with the film. Similarly, favorable interactions, rather than slow transport, must control the response of uric acid and adenine, which showed a relatively sensitive response at ultrathin OPPy films. Templating ultrathin polypyrrole (ca. 16A thickness) with adenosine, inosine and ATP produced a small decrease in sensitivity for all probes as compared to OPPy electrodes with some exceptions for uric acid, which showed an increase in sensitivity at OPPy/ado and OPPy/ino electrodes (Table 4-1). The sensitivity of the nontemplated OPPy membranes prepared from MeCN was higher than the sensitivity of the templated membranes prepared from an aqueous solvent, again with some exceptions for uric acid which showed an increase in sensitivity for OPPy films prepared in 80% MeOH and a decrease in sensitivity for OPPy films polymerized from water and NaC10 4 electrolyte. The sensitivity data listed in Table 4-3 were calculated using linear regression through (0,0) since the calibration data were background subtracted. Theoretically, at zero concentration there should be zero current if the background currents are subtracted.

PAGE 100

86 Table 4-3 Sensitivity of OPPy Films Prepared from Different Polymerization Solvents Ru(NH 3 ) 6 3+b Uric Acid b Adenine b OPPy 3 (MeCN) 8.2 ±0.8 6.7 ±0.5 15 ± 2 OPPy 3 (80% MeOH) 6.0 ±0.2 10.5 ±0.1 4 ± 2 OPPy 3 (H ? 0, NaCIO,) 6.4 ±0.1 3.8 ±0.2 2.8 ±0.7 Note: sensitivity (pA/mM) a electrode area 0.07 cm 2 , scan rate 0.020 V/s, film thickness ca. 16A b analyte concentrations 0.1-1 x 10‘ 3 M in 0.5 M pH 7.0 potassium phosphate buffer

PAGE 101

87 As shown in Table 4-2 the sensitivity of Ru(NH 3 ) 6 3+ , but not of uric acid and adenine, was similar at ultrathin OPPy film electrodes (ca. 16A thickness) prepared from the same solvent as the templated OPPy (MeOH/water) films but without the templates. This indicated that Ru(NH 3 ) 6 3+ was less sensitive to changes in OPPy morphology with templating and polymerization solvent than uric acid and adenine. The template, ATP, shown in Figure 3-2, can interact electrostatically with the positively charged polypyrrole during the formation of the polymer. The sensitivity of all the probes at the OPPy/ ATP film electrodes was low even though the sensitivity remained higher than the theoretical value. For Ru(NH 3 ) 6 3+ the sensitivity was ca. 74% higher than the theoretical sensitivity indicating some favorable interactions of Ru(NH 3 ) 6 3+ with the film. The highest sensitivity was observed for adenine showing film selectivity to adenine. The response of all the test probes at the OPPy film electrodes prepared in the absence of ATP, but from the same solvent, was significantly different (Table 4-2) confirming the effect of ATP on film morphology and the effect of film morphology on probe response. Selectivity at Bare. OPPy and OPPv Templated Electrodes GC electrodes were not selective, and hence, showed similar sensitivity to uric acid and adenine presumably because of similar interactions with the electrode and reasonably fast kinetics of the probes at the GC surface. OPPy films were more selective to adenine, compared to uric acid and Ru(NH 3 ) 6 3+ , and OPPy/ ATP films had similar higher selectivity for adenine. Overall, after templating OPPy films with purines less sensitive films were produced.

PAGE 102

88 The possible reasons may be that the templated films were more hydrophobic/compact films allowing fewer favorable interactions with the probes. Adenosine Detection Adenosine cannot be detected at pH 7.0 at bare GC above the high background at the oxidation potential of ca. 1.2 V vs SCE (Dryhurst, 1972). CVs of adenosine did not show a peak and were similar to those of adenine (Figure 4-3). Adenosine currents were measured at the potential where the background currents were small compared to adenosine response. To illustrate this point, the cyclic voltammetric response of 0.005 M adenosine in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado film electrode is shown in Figure 4-4. In this CV the current for adenosine was measured at ca. 1090mV. The calibration results obtained using this approach are summarized in Table 4-4. The results showed that adenosine could be detected at ultrathin (ca. 1 6A thickness) OPPy membrane electrodes probably because the OPPy films suppressed GC and bulk solvent oxidation at the detection potential. The results showed a possible route for optimization of adenosine detection by coating GC with a membrane which can suppress the oxidation currents arising from the bulk background oxidation, yet still allow favorable interactions for adenosine response. With the exception of OPPy/ino, templating of polypyrrole decreased the sensitivity of OPPy to adenosine which indicated that templating does not significantly improve interactions of adenosine with OPPy films.

PAGE 103

89 _J i I i I i I i I i 1 1 1 1 1200 1000 800 600 400 200 0 potential (mV) Figure 4-4 Cyclic voltammogram of 0.005 M adenosine (solid line) in 0.5 M pH 7 potassium phosphate buffer (phosphate buffer response alone dashed line) at OPPy/ado film (ca. 16 A) GC electrode (0.07cm 2 area), scan rate 0.050 V/s

PAGE 104

90 Table 4-4 Sensitivity of Adenosine at OPPy Modified GC Electrodes electrode sensitivity (pA/mM) bare GC no response OPPy 0.87 ±0.03 OPPy/ado 0.62 ± 0.03 OPPy/ino 1.71 ±0.05 OPPy/ATP 0.50 ±0.20 Note: 0.1-1 x 10' 3 M in 0.5 M pH 7.0 potassium phosphate buffer, pK a =3.5, 12.5 (Brown, 1991), electrode area 0.07 cm 2 , scan rate 0.020 V/s, current measured at ca. 1.1 V vs SCE., film thickness ca. 16A

PAGE 105

91 Characterization of Film Permeability bv Electrochemical Methods Determination of Apparent Membrane Diffusion Coefficients RDE measurements can be used to determine the apparent diffusion coefficients (D app =aD m (cm 2 /s)) through films on electrodes, where a is the membrane partition coefficient and D m the diffusion coefficient in the film as described in Chapter 2. The values of D app can provide information about film morphology (Witkowski et al., 1991). The RDE measurements were used here to shed light on the morphology of the OPPy films. Djpp values are generally lower for compact films because of the possible interactions between the probe and the film slowing in-film transport (Witkowski et al., 1991). For example, the low D^ of Ru(NH 3 ) 6 3+ determined at the ultrathin OPPy film electrode as shown in Table 4-5 was expected from the compact structure of the pinhole-free film (Hsueh and Brajter-Toth, 1994; Witkowski et al., 1991) with high local density of the carbonyl groups slowing Ru(NH 3 ) 6 3+ transport through the film. D app values were higher at thicker OPPy films on GC electrodes prepared by one step polymerization/overoxidation (D app =2.8 x 10~ 8 cm 2 /s at ca. 0.1 pm thick OPPy film; Witkowski and Brajter-Toth, 1992) indicating a less compact structure for these films (Hsueh and Brajter-Toth, 1994). For uric acid at ultrathin OPPy films (ca. 1 6 A) the D app values were lower than D app values for Ru(NH 3 ) 6 3+ (Table 4-5). This indicated that the compact film structure resulted in less efficient transport/partitioning of uric acid.

PAGE 106

92 Table 4-5 Apparent Diffusion Coefficients (cm 2 /s) for OPPy, OPPy/ado and OPPy/ATP Film Ru(NH 3 ) 6 3+ a Uric Acid a OPPy c 7.2 x 10' 9b 1.0 x 10' 9 OPPy/ado d 1.5 x 10' 9 3.8 x 10' 11 OPPy/ATP 6 1.7 x 10‘ 9 1.2 x 10' 9f Note: For RDE measurements electrode area 0.23 cm 2 , scan rate for linear sweep voltammetry 0.010 V/s, rotation rate (w) = 0 to 1200 rpm “Probe concentration 0.001 M in 0.5 M pH 7.0 potassium phosphate buffer, potential window for Ru(NH 3 ) 6 3+ 0 to -0.5 V and for uric acid 0. 1 to 0.6 V. b (Hsuehand Brajter-Toth, 1994) ‘film thickness ca. 16A d film thickness ca. 4A ‘film thickness for Ru(NH 3 ) 6 3+ ca. 4A and film thickness for uric acid ca. 16A f electrode area 0.064 cm 2

PAGE 107

93 D app values for Ru(NH 3 ) 6 3+ decreased when the ultrathin films were templated (templated films had a thickness of ca. 4 A except for the OPPy/ATP film electrode with uric acid which had thickness ca. 16A) indicating a more compact film morphology. A further increase in the local density of the carbonyl groups in the resulting templated OPPy films was unlikely based on the thickness of these films. The observed decrease was likely the result of a more compact, hydrophobic film morphology after templating. In general, changes observed in D w on the templated films (Table 4-4) for Ru(NH 3 ) 6 3 confirmed small changes in film morphology and interactions with templating. The transport results were consistent with the results of ESCA analysis which showed differences in the microenvironment of the templated films from the multiple peaks for several types of atoms in the ESCA spectra, consistent with hydrophilic as well as hydrophobic regions in the films as proposed by Witkowski et al. (1991). Effect of Diffusion vs Surface Interactions on Response The dependence of the voltammetric peak current, ip, on scan rate, v, can be used to assess if the electrochemical process is diffusion or adsorption-controlled. On film coated electrodes this information can provide insight into film structure. As shown in Table 4-6, the slopes of log ip vs log v plots on all the investigated electrodes were typically less than 0.5 for Ru(NH 3 ) 6 3+ . The slopes were obtained by linear regression of the data, and the fits were not obtained through the origin in accordance with equations 2.1 to 2.4. A slope of 0.5 was expected for a diffusion controlled process described in Chapter 2 (Bard and Faulkner, 1980).

PAGE 108

94 Table 4-6 Slopes of Log Peak Current vs Log Scan Rate Plots at GC and OPPy Film Electrodes Ru(NFLL 3+ Uric Acid Adenine bare GC 0.40 ±0.03 0.60 ± 0.08 0.80 ±0.10 OPPy 0.44 ±0.02 0.14 ± 0.10 0.24 ± 0.07 OPPy/ado 0.46 ±0.01 0.35 ±0.02 0.28 ±0.05 OPPy/ino 0.13 ± 0.10 0.34 ±0.00 0.36 ±0.03 OPPy/ATP a 0.002 ± 0.00 0.55 ±0.05 Note: electrode area ca. 0.07 cm 2 , scan rate 0.020-0.200 V/s, analyte concentrations 0.25 x 10' 3 M in 0.5 M pH 7.0 potassium phosphate buffer, all potentials vs SCE, film o thickness ca. 16A “slopes could not be obtained since the kinetics were quasi-reversible, for Ru(NH 3 ) 6 3+ all values used in the calculations were obtained from data that had reversible kinetics

PAGE 109

95 The lower than 0.5 slopes observed for Ru(NH 3 ) 6 3+ may reflect partial surface blockage (Zhang and Bard, 1989) in the Ru(NH 3 ) 6 3+ reaction at bare and coated electrodes indicating transport problems at the film electrodes. For uric acid and adenine, the slopes were greater than 0.5 at bare GC indicating mixed diffusion-adsorption control of the electrode response (Freund and Brajter-Toth, 1992) and were in agreement with the sensitivity data where higher than theoretical sensitivity was measured at bare GC (Table 4-1) reflecting enhanced interactions of these probes with the GC surface and indicating few complications with transport. At OPPy film electrodes (ca. 16A thickness) the slopes for uric acid and adenine were less than 0.5 indicating a partially blocked surface (Zhang and Bard, 1989) in reactions of the two probes. For adenine at OPPy/ ATP film electrodes (ca. 16A thickness), the slope approached 0.5 as expected for a diffusion controlled process. The results obtained on OPPy film electrodes indicated, in general, a partially blocked surface in agreement with the model of a relatively compact film on the surface complicating transport consistent with the RDE results and the D app values that were obtained. Conclusions At templated OPPy film electrodes the sensitivity observed in slow scan voltammetry generally decreased, but some selectivity was shown towards adenine. This indicated that the favorable interactions with the OPPy films, present before templating, were not enhanced by templating possibly because the interactions involved in forming the templated polymer and

PAGE 110

96 the new template-analyte interactions were quite weak. Nevertheless, templating did change the interactions of the probes with the OPPy film, and the selectivity of the templated OPPy films. Small changes in sensitivity have been observed in other practical applications of imprinted polymers (Vlatakis et al., 1993; Palmisano et al., 1995). The exception was the recent report of significant changes in potentiometric selectivity of imprinted polymers (Dong et al., 1988; Hutchins and Bachas, 1995). Potentiometric measurements, unlike amperometric measurements, are equilibrium measurements and generally require film equilibration with the analytical sample. Additionally, in the reported potentiometric measurements the templates remained in the film after the films were imprinted. In spite of the weak templating interactions imprinting of OPPy films changed the sensitivity and selectivity of the films. However, since the interactions and film compactness present before the templating were relatively significant, they appeared to continue to control the response of the templated OPPy films and no dramatic changes in the electrode response were observed when OPPy and templated OPPy film response was compared. To improve sensitivity of the templated films, thicker, more permeable, and initially less strongly interacting polymer films may be required for templating purposes. Thin films were deliberately used here to insure high film permeability in order to limit sensitivity problems common at thicker films where in-film transport is slow. In the final outcome the use of the ultrathin films was advantageous because it produced high sensitivity and allowed detection of the small morphological changes from templating. The RDE measurments allowed the determination of the D app values for the films. From the low values and the high sensitivity in slow scan voltammetry measurements the

PAGE 111

97 high permeability of the ultrathin films was confirmed. The RDE results confirmed that the templated OPPy films were a little more compact consistent with the fewer amount of ions present in the templated OPPy films as evidenced by ESCA analysis. The scan rate results confirmed surface blocking by the films with the exception of the detection of adenine at OPPy/ ATP films. Finally, adenosine detection was shown to be more sensitive at the modified surfaces.

PAGE 112

CHAPTER 5 ANALYSIS OF SURFACE INTERACTIONS AT GC AND GC TEMPLATED OPPy ELECTRODES Analysis of surface interactions at GC and GC coated with templated OPPy films is presented in this chapter. The analysis aimed to provide quantitative information about the interactions controlling the selectivity and sensitivity of the films. In the determination of Ru(NH 3 ) 6 3+ , uric acid and adenine, OPPy films appear to limit adsorption at graphite and suppress high background currents at large positive potentials, allowing oxidative detection of adenosine at potentials greater than one volt. This chapter also describes the long term stability of OPPy membranes. Saturation Binding of Ultrathin OPPy and Templated OPPy Electrodes The Ru(NH 3 ) 6 3+ calibration curves are shown in Figures 5-1 to 5-3. Regression of the calibration data gives y=3.93x + 1.22 (R=0.99, all points fitted) at bare GC,y=9.6x + 2.2 (R=0.87, last 3 points fitted) at OPPy, y=4.60x +1.19 (R=0.98, all points fitted) at OPPy/ado, y=6.30x + 0.32 (R=0.99, all points fitted) at OPPy/ino, y=4.03x + 0.53 (R=0.97, all points fitted) at OPPy/ ATP electrodes with slope in pA/mM and the intercept in pA. The linear portions of the plot were used in the regression analysis, and at least three measurements were averaged for each point. A complete list of the data is given in Table 4-2. Inspection of the 98

PAGE 113

99 Figure 5-1 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 114

100 5.04.54.03.53.0§ 2.5\ fc § 2 .^ 1.5i.o0.5r o.o — 0.0 T 0.1 0.2 “i 1 1 1 i 1 i 1 i 0.3 0.4 0.5 0.6 0.7 concentration (mM) 0.8 76 c S 3 O 5432 1x 5 Oi 1 1 1 1 1 1 i ' i 0.0 0.2 0.4 0.6 0.8 1.0 concentration (mM) Figure 5-2 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 115

101 Figure 5-3 Current vs concentration for Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ATP. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 116

102 calibration data indicated saturation binding for Ru(NH 3 ) 6 3+ at lower concentrations where Figure 5-1 for the OPPy film plateaued and at higher concentrations where the OPPy/ATP film began to plateau in Figure 5-3. The plateau at lower concentrations of Ru(NH 3 ) 6 3+ indicated strong binding of Ru(NH 3 ) 6 3+ with the film. The results for the modified electrodes showed slopes greater than 3.93 pA/mM, the sensitivity at the bare electrode, so the OPPy film electrodes were more sensitive than the bare GC electrode, and hence had more favorable interactions with Ru(NH 3 ) 6 3+ . With templating there was a decrease in OPPy film sensitivity, shown by slopes lower than the 9.6 pA/mM of the OPPy film electrode. The results showed that coating GC with an ultrathin OPPy film provided a simple method for improving the sensitivity and reproducibility of the electrode. The reproducibility was verified by low standard deviation of multiple measurements at the same membrane electrode (represented by error bars in Figures 5-1 to 5-9) and by good intralaboratory precision of measurements at different electrodes. OPPy and templated OPPy film electrodes were used for at least sixty measurements from scan rates of 0.020 to 0.200 V/s at potential windows as large as -0.50 V to 1.2 V, without showing signs of film loss, as monitored by 0.010 M Fe(CN) 6 3 ' in 0.50 M KC1. The electrodes were stored dry at room temperature. For uric acid the calibration curves in Figures 5-4 to 5-6 showed a plateau at concentrations above 0.2 mM for bare GC, and OPPy and OPPy/ado film electrodes indicating saturation binding and confirming surface interactions of uric acid (Ahmad and Brajter-Toth, 1992). The calibration curves at OPPy/ino and OPPy/ATP film electrodes (Figures 5-5 and 5-6) showed no plateau, possibly due to weaker interactions of uric acid

PAGE 117

103 Figure 5-4 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated GC (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 118

104 0.0 0.2 0.4 0.6 0.8 concentration (mM) 1.0 Figure 5-5 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 119

current (p A) 105 concentration (mM) Figure 5-6 Current vs concentration for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ATP coated electrode. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 120

106 with the film. For uric acid y=16.66x-0.26 (R=0.99, first 5 points fitted) at bare, y=7.58x0.13 (R=0.98, first 4 points fitted) at OPPy, y=8.57x-0.08 (R=0.99, first 4 points fitted) at OPPy/ado, y=8.57x-0.09 (R=0.99, all points fitted) at OPPy/ino, and y=4.71x+0.09 (R=0.98, all points fitted) at OPPy/ ATP. The plateaued portions were not used in the regression analysis, and at least three measurements were averaged for each point. Uric acid sensitivity at bare GC (16.66pA/mM) was significantly higher than the sensitivity of Ru(NH 3 ) 6 3+ (3.93 pA/mM) and higher than expected from the two fold increase in the number of electrons in the electron transfer process, indicating significantly higher surface concentration or more favorable interactions for uric acid with GC. Unlike for Ru(NH 3 ) 6 3+ where the sensitivity generally increased at modified electrodes (Table 4-2), uric acid sensitivity decreased 54% at OPPy film electrode in comparison to bare GC. At templated OPPy film electrodes a 49% decrease in sensitivity for OPPy/ado and OPPy/ino films and a 72% decrease in sensitivity for OPPy/ ATP films were observed in comparison to bare GC. Templating of OPPy films showed a 11% improvement in sensitivity over OPPy with the exception of OPPy/ ATP films which showed a 38% decrease in sensitivity in relation to OPPy. Figures 5-7 to 5-9 showed plateaus in adenine calibration curves indicating adenine surface interactions and apparent saturation binding. For adenine at OPPy/ino and OPPy/ ATP electrodes where the calibration data were very similar [y=3.01x+1.34 (R=l, first 2 points fitted) and y=2.71x+1.39 (R=0.95, first 4 points fitted) respectively], the plateaus shifted to lower concentrations which coincided with a significant decrease in sensitivity. The calibration data were 17.94x-1.01 (R=0.95, first 6 points fitted) at bare, y=19.93x-1.44 (R=0.77, first 4 points fitted) at OPPy and y=12.15x+0.06 (R=0.99, first 3 points fitted) at

PAGE 121

107 12-i 10o I 2 o OH — 0.0 0.2 0.4 0.6 0.8 concentration (mM) 1.0 7 6 5 c a> 3 2 10 — 0.0 0.2 I — i ' 1 ' 1 — 0.4 0.6 0.8 concentration (mM) T 1 1.0 Figure 5-7 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at bare (top) and OPPy coated (bottom) GC electrodes. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 122

108 2 1 . 0 3 O 0.5H O.CH 1 1 1 1 1 1 ' r1 r— 0.0 0.2 0.4 0.6 0.8 1.0 concentration (mM) Figure 5-8 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ado (top) and OPPy/ino (bottom). Electrode area 0.07 cm 2 , film thickness ca.16 A, scan rate 0.020 V/s.

PAGE 123

109 Figure 5-9 Current vs concentration for adenine in 0.5 M pH 7.0 potassium phosphate buffer at OPPy/ATP coated electrode. Electrode area 0.07 cm 2 , film thickness ca. 16 A, scan rate 0.020 V/s.

PAGE 124

110 OPPy/ado electrodes, showing relatively high sensitivity at these electrodes for adenine. A significant sensitivity decrease at OPPy/ino (3.01 gA/mM) and OPPy/ATP (2.71 pA/mM) was seen. Apparent saturation binding observed from the plateaus in the calibration curves of uric acid (Figures 5-4, 5-5), adenine (Figures 5-7 to 5-9), and Ru(NH) 3 3+ at low and high concentrations (Figure 5-1) indicated surface interactions. The apparent interactions occurred at both GC and OPPy film electrodes. In spite of the interactions, reproducibility of measurements was significantly enhanced at film electrodes since up to sixty reproducible measurements could be made on the same film whereas at GC reproducible measurements could not be made without resurfacing prior to each measurement. The film electrodes were rinsed with deionized water before each measurement and stored dry at room temperature. Although probe interactions with the film were apparent from the calibration data, it was difficult to determine their strength or assign the interactions from sensitivity data. Analysis of Membrane Interactions Scatchard Plot Analysis Validity of Scatchard analysis (Scatchard, 1949) in the quantitative analysis of surface interactions at film coated and bare GC electrodes was evaluated. Scatchard analysis describes reversible, noncooperative interactions such as observed for small molecules with DNA (McGhee and von Hippel, 1974) and can be used to obtain binding constants and the maximum number of binding sites. The analysis applies to analytes in a single form interacting

PAGE 125

Ill with a discrete population of sites, which have a single affinity for the analyte. A Scatchard plot of [bound species]/[ffee species] vs [bound species] is linear with a slope equal to -1/K, where K is the binding constant equal to [free species]/[bound species] and the x intercept is equal to the maximum number of binding sites, B,^. In order to quantitate the surface interactions of an electroactive analyte using this approach several assumptions have to be fulfilled (Millan and Mikkelson, 1993): 1) the measured analyte current must be due to the diffusing analyte while the bound analyte is electroinactive; 2) equilibration of the bound and the diffusing analyte occurs rapidly on the electrochemical time scale; 3) the diffusion coefficient of the analyte in the film must be the same as in the solution bulk, and 4) preconcentration of the analyte in the film on the electrode cannot significantly alter the analyte concentration in the solution bulk. For OPPy films, assumption 1 should be fulfilled because the concentration of the bound analyte is expected to be low and because, with the exception of Ru(NH 3 ) 6 3+ , the selfexchange rate constants for the bound analyte are expected to be small. Membrane porosity, which for the ultrathin membranes was assumed to be high (Hsueh and Brajter-Toth, 1994), and the presumed reversibility of the membrane interactions, which should be weak (Witkowski and Brajter-Toth, 1992; Witkowski et al., 1992), is expected to satisfy assumption 2. Finally, the solution concentration of the analyte should not be altered by the interactions as expected by assumption 4. Assumption 3, however, is not fulfilled directly because the in-film diffusion coefficients are low (Hsueh and Brajter-Toth, 1994; Witkowski et al., 1992); high film permeability should satisfy this assumption in practice. Scatchard analysis is often limited by the assumptions which are not fulfilled for complex systems where

PAGE 126

112 different interactions or combinations of interactions can lead to curved or non-linear plots, as in oligopeptide binding with DNA (Browne et al., 1993). In contrast, small molecule binding with DNA often fulfills the conditions of Scatchard analysis (Hogan et al., 1979). In the Scatchard analysis carried out on the regression data, the concentration of the bound species was obtained from the maximum current measured for each concentration divided by the sensitivity. This assumed analyte response only as a result of the surface interactions. The sensitivity used to determine the concentration was obtained from the slope of the regression data taken through the origin except for Ru(NH 3 ) 6 3+ at OPPy film electrode. Scatchard analysis was also done using the sensitivity data from the slope of the regression line not through the origin. The two sensitivities produced very similar K and values indicating that the non-zero intercepts were small and had a small effect on the sensitivity. In determining the concentration of the bound analyte, the bare electrode current was not subtracted from the membrane current because the response of the bare and the OPPy film coated GC electrodes was considered to be a result of different processes, including different surface interactions. In Scatchard analysis the initial concentration of the analyte was taken as the concentration of the free species in solution. Scatchard analysis results are summarized in Table 5-1, and the binding constants, defined as [free]/[bound] analyte or the concentration of the analyte at which the binding sites are half occupied, are listed. The binding contants in M" 1 are listed for comparison. For Ru(NH 3 ) 6 3+ , where the binding interactions were verified

PAGE 127

113 Table 5-1 Scatchard Plot Analysis for Ru(NH 3 ) 6 3+ at bare and OPPy Modified GC Electrodes sensitivity 3 (pA/mM) K (M) 1C 1 (M1 ) B max (M) bare GC 6.0 6.9 x 10* 4 1.4 x 10 3 1.3 x 10' 3 OPPy 9.6 1.9 x 10' 5 5.3 x 10 4 3.4 x 104 OPPy/ado 6.6 N/A b N/A b N/A b OPPy/ino 6.8 3.3 x 10' 3 3.0 x 10 2 4.1 x 10' 3 OPPy/ ATP 4.7 2.6 x 10' 3 3.8 x 10 2 3.3 x 10' 3 Note: analyte concentrations 0. 1-1 .0 x 10' 3 M in 0.5M potassium phosphate buffer pH 7.0, cyclic voltammetric currents measured at -0.350 V at GC with 0.07 cm 2 area, scan rate 0.020 V/s, film thickness ca. 16A “sensitivity data (see text for fitted points) calculated from a linear fit regression analysis through the origin except for OPPy (see text) ‘’Nonlinear Scatchard plot

PAGE 128

114 mainly by higher than theoretical sensitivity (Figures 5-1 to 5-3), the plots were reasonably linear, with an exception of the curved plot for OPPy/ado film electrode. The Scatchard analysis of Ru(NH 3 ) 6 3+ data showed strongest binding at OPPy film electrodes with weaker binding at the bare GC and templated OPPy film electrodes. From K and B mix the interactions at templated OPPy/ino and OPPy/ ATP film electrodes with Ru(NH 3 ) 6 3+ were similar. If, as previously proposed, templating produced more compact films, the observed decrease in binding strength with templating was reasonable if the new films became more hydrophobic, reducing local high density of carbonyl groups presumably needed for the favorable interactions of Ru(NH 3 ) 6 3+ (Witkowski and Brajter-Toth, 1992). The lower sensitivity pointed to this model of a compact, hydrophobic film with templating since the binding strength, as indicated by lower K (M' 1 ) values, decreased after templating of the OPPy films, pointing to less efficient local interactions even though the density of the binding sites increased (Table 5-1). Overall, the binding constant of Ru(NH 3 ) 6 3+ measured at OPPy was an order of a magnitude higher than at the bare GC indicating relatively strong interactions. For instance, K values indicating strong binding are 10" 6 M for epinephrine bound to P-adrenoreceptors (Voet and Voet, 1990) and 1.74 x 10 3 M' 1 for Co(phen) 3 3+ bound to DNA (Millan and Mikkelson, 1993). For association of Ru(NH 3 ) 6 3+ with negatively charged polyacrylates, Jiang and Anson calculated binding constants of ca. 10 3 tolO 4 M' 1 for 0.2 mM Ru(NH) 3 3+ in 0.05 M pH 9.2 Na 2 B 4 0 7 , which was similar to the values obtained here at GC and OPPy electrodes where it has been postulated that similar groups may be involved in binding. The

PAGE 129

115 similar values to Jiang and Anson showed that the assumptions made for quantitating surface interactions by Scatchard analysis were valid. The Scatchard plot analysis of Ru(NH 3 ) 6 3+ data at OPPy/ado electrode produced a plot which was concave upward indicating negative cooperativity or several types of noninteracting binding sites with unchanging and different affinity for the ligand. The upward curvature could also indicate several types of interactions with the receptor sites or more than one kind of binding site (Limbird, 1986). No reasonable Scatchard plots could be obtained for uric acid and adenine, with the exception of the results for uric acid at OPPy/ATP which showed negative cooperativity. This indicated that the surface interactions with uric acid and adenine did not fulfill the simple model under assumptions 1-4. Langmuir Isotherm Analysis Analysis of the interactions based on the Langmuir isotherm model was also tested. This analysis has been used to quantitate the effect of a polymer microenvironment on polymer ligand interactions (Alexandratos et al., 1992). Langmuir isotherm describes conditions of surface saturation at high analyte concentrations (Tohda et al., 1995). The analysis assumes that the surface is a lattice of noninteracting sites and a monolayer is formed when the number of adsorbed molecules is at a maximum (Castro et al., 1991). According to the theory, a plot of the inverse of the concentration of the bound analyte, l/[bound], vs the inverse of the analyte concentration in solution, l/[free], should be linear, with a slope equal tol/ (N^^K), where is the maximum number of binding sites and K is the binding

PAGE 130

116 constant (M' 1 ) defined as K=[bound]/[free], The value for is found from the inverse of the y-intercept. As in Scatchard analysis, peak current or the maximum current for a given analytical concentration divided by the sensitivity, from the regression of the calibration data in Figures 5-1 to 5-6 through the origin except in the case of Ru(NH 3 ) 6 3+ at OPPy electrode described earlier, was used to obtain the [bound] analyte at the surface. The concentration of the bound analyte was also calculated using the sensitivity data obtained from fits not through the origin, but no significant difference was apparent in the results of the Langmuir isotherm analysis. The results of the Langmuir analysis are summarized in Table 5-2. For Ru(NH 3 ) 6 3+ the plots were linear (R^O.99) except for OPPy where R=0.41 and for OPPy/ado where R=0.91. The binding constants and the number of binding sites that were obtained were nearly equal to those found from Scatchard analysis. As in Scatchard analysis there was reasonable agreement between the number of binding sites for Ru(NH 3 ) 6 3+ at GC, OPPy/ino and OPPy/ ATP electrodes. For Ru(NH 3 ) 6 3+ , similar K and values at OPPy/ino and OPPy/ATP electrodes, which were somewhat lower and higher respectively to the values at GC, suggested similar surface interactions of Ru(NH 3 ) 6 3+ at these surfaces. Since Ru(NH 3 ) 6 3+ can only interact electrostatically, the microenvironment at GC and the membranes must be weakly electrostatic or predominantly hydrophobic, consistent with previous models of GC having a predominantly hydrophobic surface (Nishizawa et al., 1991). Based on the Scatchard analysis, since Langmuir analysis was not valid for that case, a more strongly interacting environment for Ru(NH 3 ) 6 3+ was present at OPPy films, which could be a more hydrophilic environment.

PAGE 131

117 Table 5-2 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid at Bare and OPPy Modified Electrodes Ru(NH 3 ) e 3+a Uric Acid 3 Nm(M) K (M' 1 ) (M) K (M' 1 ) GC 1.2 x 10' 3 1.5 x 10 3 8.0 x 10‘ 5 5.0 x 10 3 OPPy N/A b N/A b 6.0 x 10‘ 5 5.7 x 10 3 OPPy/ado N/A b N/A b 5.5 x 104 1.3 x 10 3 OPPy/ino 3.2 x 10‘ 3 4.0 x 10 2 3.8 x 10‘ 3 2.3 x 10 2 OPPy/ ATP 3.7 x 10' 3 3.3 x 10 2 5.5 x 104 2.9 x 10 2 a analyte concentrations 0. 1-1.0 x 10' 3 M in 0.5M pH 7.0 potassium phosphate buffer, cyclic voltammetric currents measured at GC with 0.07 cm 2 area, scan rate 0.020 V/s, OPPy film thickness ca. 16 A. Maximum currents for Ru(NH 3 ) 6 3+ measured at ca. -0.350 V and for uric acid measured at ca. 0.350 V ‘’Langmuir isotherm plots were nonlinear.

PAGE 132

118 Langmuir analysis of uric acid data gave linear plots; at OPPy the plot approached linearity with R=0.98. The binding constants (Table 5-2) were similar at GC and OPPy coated GC. In contrast, according to Scatchard analysis, Ru(NH 3 ) 6 3+ binding to OPPy was significantly stronger than to GC. Templating caused a decrease in the strength of the surface interactions for uric acid. The binding constants obtained for uric acid were reasonably high compared to the polymer supported dimethylamine ligand substituted benzoic acids interactions, where K was 15-126 M' 1 (Alexandratos et al., 1992). Adenine Langmuir isotherm plots were nonlinear (R=0.87 for bare, R=0.71 for OPPy and much lower R values for the templated films) possibly because, in spite of the appearance of saturation binding, other processes may be responsible for the response which do not fit the simple Langmuir model of surface interactions. OPPy and OPPy Templated Surface Stability OPPy films could be used for at least sixty measurements and were stable throughout a wide potential window (-0.50 to 1.20 V). In general the limits of detection (calculated from 3 a/m, where a is the standard deviation of the blank and m is the sensitivity) for Ru(NH 3 ) 6 3+ and adenine were lower at OPPy and templated OPPy film electrodes compared to bare GC (Table 5-3), with the exception of adenine at the OPPy/ ATP film electrode. The standard deviation of the blank was calculated from at least three measurements in the electrolyte solution. For uric acid the limits of detection were higher at OPPy coated GC than at bare GC. In general for adenine the LODs were better at OPPy coated electrodes. Despite partial

PAGE 133

119 Table 5-3 Limits of Detection at Bare GC and OPPy Modified Electrodes for Ru(NH 3 ) 6 3+ , Uric Acid and Adenine Ru(NH 3 ) 6 3+b Uric Acid b Adenine b GC a 4.7 x 104 8.2 x 10‘ 7 2.9 x 10' 5 OPPy“ 1.5 x lO* 4 4.7 x 106 7.0 x 10" 6 OPPy/ado a 4.0 x Iff 5 2.4 x 10' 7 7.0 x 106 OPPy/ino a 4.0 x 10" 6 2.7 x 10" 6 C OPPy/ATP a 2.0 x 10" 6 2.4 x Iff 6 7.0 x 10‘ 5 Note: LOD (M) obtained by 3o/m (a standard deviation of the blank, m sensitivity) “sensitivity obtained from cyclic voltammetry, electrode area 0.07 cm 2 , film thickness ca. 16 A scan rate 0.020 V/s b analyte concentrations 0.1-1 x 1 O' 3 M in 0.5 M pH 7.0 potassium phosphate buffer, sensitivity data in Table 4-1 c not enough background data were available for calculation

PAGE 134

120 exclusion of uric acid by the OPPy films, the LODs were in the micromolar range. The limits of detection for adenine were typically lower than Ru(NH 3 ) 6 3+ and higher than uric acid which confirmed that OPPy and templated OPPy films contributed to the improved LODs. With the exception of adenine, the OPPy/ino film electrode produced the largest linear dynamic range (LDR). For Ru(NH 3 ) 6 3+ , all the calibration plots in Figures 5-1 to 5-3 except for the OPPy electrode illustrated a wide LDR and a LOD below that at GC electrode (ca. KT 4 M). For uric acid, the LDRs extended up to 0.2 mM with the exception at OPPy/ino and OPPy/ ATP electrodes which were linear throughout the range studied. The LDRs for adenine were much smaller than for the other two probes, but despite this, all the LODs for adenine were comparable to Ru(NH 3 ) 6 3+ and uric acid. Conclusions According to the Scatchard and Langmuir analysis, the strength of interactions of structurally different probes, Ru(NH 3 ) 6 3+ , uric acid and adenine at GC and OPPy membrane electrodes was quite high especially for Ru(NH 3 ) 6 3+ at OPPy. The magnitude of the interactions was estimated from Scatchard and Langmuir isotherm analysis of the calibration data. For Ru(NH 3 ) 6 3+ the reasonably strong interactions with OPPy films (K=5.3 x 10 4 M' 1 ) decreased after templating of OPPy films (K ca. 10 2 ), confirming changes in OPPy film morphology with templating, and supporting a model of more compact, hydrophobic templated OPPy films with a less favorable distribution/density of the hydrophilic binding groups. For uric acid, the strength of interactions with GC (K=5.0 x 10 3 M' 1 ) did not

PAGE 135

121 significantly change after GC was coated with an ultrathin OPPy film (K =10 3 M' 1 ), possibly because of the continued interactions with the GC substrate at the ultrathin OPPy film (ca. 16 A) electrode. The lower sensitivity for uric acid measured at OPPy electrodes could account for this result if the effective GC area decreased. The apparent decrease in the strength of interactions for uric acid at templated OPPy films was consistent with the idea of a compact, templated film decreasing the strength of interactions of uric acid with GC. The goal of the Scatchard and Langmuir analysis was to obtain quantitative information about the surface interactions in order to rationalize the sensitivity data and to verify previous conclusions (mostly from the sensitivity and transport data) about the effect of templating on the interactions. The results confirmed that for Ru(NH 3 ) 6 3+ detection sensitivity and limits of detection were better at OPPy films compared to those at bare GC consistent with a cation selective, interacting membrane model (Hsueh and Brajter-Toth, 1994). After templating, the more compact OPPy films appeared to be less interactive with Ru(NH 3 ) 6 3+ . This could be reasonable if the interacting groups were redistributed over the surface, in agreement with the observed higher density of the interacting groups on the templated films. For uric acid the sensitivity was better at bare GC, in agreement with the anion excluding properties of the OPPy films and the higher hydrophobicity of GC. The interactions of uric acid with GC appeared to dominate the reponse of OPPy film electrodes, with the strength of these interactions decreasing at templated OPPy films, consistent with their increased compactness. OPPy and templated OPPy film electrodes were sensitive to adenine and adenosine.

PAGE 136

122 The analysis indicated that surface interactions contributing to the observed adenine response, which were apparent from the calibration data of adenine, did not fit the surface interactions model used here to quantitatively analyze the interactions. This was not surprising in view of the likely complexity of the oxidation mechanisms of adenine which may involve additional steps such as water oxidation in addition to adenine surface interaction (Dryhurst, 1972). The Scatchard and Langmuir analysis supported previous conclusions that the sensitivity optimization in adenosine detection would require less control of the response by surface (Chen et al., 1994) or polymer backbone interactions.

PAGE 137

CHAPTER 6 CHARACTERISTICS OF PURINE TEMPLATED OPPy FILMS ON CARBON FIBER ULTRAMICROELECTRODES In order to design a practical in vivo sensor for adenosine, both spatial and temporal resoltuion is required because of the rapid turnover rate of adenosine in vivo and a possible disruption of cellular metabolism by a large sensor (Fredholm, 1987; Soderback et al., 1987). Ultramicroelectrodes (UMEs), because of their small size, small iR drop, and a small time constant, can have submicrosecond temporal resolution and are well suited for sensor design (Fleischmann et al., 1987; Tariaka and Kashiwagi, 1989). Bare carbon fiber UMEs have been tested in adenosine detection; in the detection bare carbon fiber had low sensitivity, unless activated, and the surface apparently fouled after extensive use (Chen et al., 1994). To obtain reproducible results Chen et al. activated carbon fiber before each measurement by applying an anodic potential of 2.2V for 125 ms followed by a cathodic potential of -1.0 V for 75 ms (Chen et al., 1994). Without the activation adenosine concentration could not be measured due to apparent surface fouling and passivation of the electrode presumably by strong adsorption of adenosine and its metabolites (Chen et al., 1994) Another approach to electrode activation used in this work was to coat the UME with a selective, preconcentrating film. A wide range of films (Turner et al., 1987; Bailey et al., 1991) including enzyme (Kuhr et al., 1993), conductive polymer (Bailey et al., 1991), ionicpolymer conductors e.g. Nafion (efficient in bioanalysis when combined with fast scan) (Millar 123

PAGE 138

124 et al., 1981; Kuhr and Wightman, 1986; Bauer et al., 1988) and OPPy films (Hsueh and Brajter-Toth, 1994) have been used to enhance the selectivity and sensitivity of UMEs. In Chapter 4 OPPy/ ATP films on GC electrodes were shown to enhance selectivity in adenine detection and to provide excellent electrode stability. The OPPy films allowed direct measurement of adenosine without additional surface activation. In this chapter, characteristics of carbon fiber UMEs modified with OPPy and OPPy/ATP films will be discussed as reusable surfaces for sensors for adenine and uric acid. The structure of OPPy films grown on the carbon fiber UMEs was expected to be similar to the structure of OPPy films grown on GC because of the similarities in the surface structure of GC and carbon fiber (McCreery, 1991). The graphite surface micro structure has been shown to control the structure of polypyrrole grown on graphite (Witkowski and Brajter-Toth, 1992). Marino et al. have shown that long carbon chain cationic surfactant assemblies at carbon fiber and at GC were structurally similar which was attributed to the similarity in the surface structure of the two forms of graphite (Marino, 1994). In addition to the electrochemical characterization of OPPy films on UMEs analysis of membrane interactions by Langmuir isotherm analysis, as described in Chapter 5, was performed. Information about the membrane structure was easier to obtain at the film coated UMEs because of the rapid mass transport to the electrode, conditions under which electrode response is determined by the film properties(Cheng and Brajter-Toth, 1992). Also, the possibility of sensitivity improvement of the UMEs by fast scan voltammetry has been investigated and is described in Chapter 1 .

PAGE 139

125 Effect of Film Thickness on the Response of Macroelectrodes Films of a thickness greater than 16A were first characterized on GC to determine the effect of film thickness on sensitivity of Ru(NH 3 ) 6 3+ and of uric acid in 0.5 M pH 7.0 potassium phosphate buffer and in 0.5 M KC1 since the thinnest OPPy films on UMEs that could be prepared were ca. 32 A (Hsueh and Brajter-Toth, 1994). The results in Table 6-1 for GC show that as the film thickness on GC increased the sensitivity of Ru(NH 3 ) 6 3+ in 0.5 M pH 7.0 potassium phosphate buffer remained reasonably constant until the thickness was increased to 48 A, when the response decreased ca. 41%. The sensitivity in Table 6-1 is the slope of the linear regression of the calibration data through the origin. For both types of OPPy films the increased sensitivity at films thicker than 16 A was due mainly to the increased scan rate; however, at 48 A OPPy/ ATP films no Ru(NH 3 ) 6 3+ response was detected above background. Table 6-1 also shows (at 44 A films) similar sensitivity for Ru(NH 3 ) 6 3+ in 0.5 M KC1 and in 0.5 M pH 7.0 potassium phosphate buffer. For uric acid, the response of the OPPy film electrodes in 0.5 M pH 7.0 potassium o phosphate buffer was not detected above the background for films thicker than 16A. However in 0.5 M KC1 a response was observed at 32 and 44 A thick films (Table 6-1). At films thicker than 44 A the response of uric acid was not detected. The sensitivity of uric acid was lower than the sensitivity of Ru(NH 3 ) 6 3+ . The results showed that the electrolyte strongly influenced the sensitivity of the films toward uric acid.

PAGE 140

126 Table 6-1 Sensitivity at OPPy and OPPy/ATP Modified Macro GC Electrodes with Varying Degrees of Film Thickness Ru(NH^ 3 " film thickness (A) OPPy OPPy/ATP 16 b 10 ± 4 4.7 ±0.2 32 10.1 ±0.1 6.6 ±0.3 44 9.2 ±0.5 7.6 ±0.1° 8.9 ±0.1 7.2 ±0.2° 48 4.6 ±0.2 d Uric Acid 3 film thickness (A) OPPy OPPy/ATP 16 b 6.7 ±0.5 5.1 ±0.3 32 no response 6.0 ± 1.0° no response 3.2 ±0.6° 44 no response 1.8 ±0.4 C no response 0.80 ±0.09 c 48 d d Note: (iA/mM, sensitivity determined from linear regression of the data through the origin “electrode area 0.07 cm 2 , scan rate 0.1 V/s, analyte concentration 0.25-1.0 mM in 0.5 M potassium phosphate buffer pH 7.0, all potentials vs SCE b from Table 4-1, data obtained at 0.02 V/s c analyte concentration 0.25-1.0 mM in 0.5 M KC1 d analyte signal was below background

PAGE 141

127 Langmuir isotherm analysis of the calibration data (Table 6-1) was used to identity the interactions with the films of different thickness. As described in Chapter 5, this method can be used when the calibration plot shows a plateau which indicates surface-saturating interactions as observed at the thicker films. Table 6-2 lists the values obtained for the binding constant, K (M' 1 ), and the maximum number of binding sites, (M), for the calibration data in Table 6-1. The linear fits were based on R>0.98 and n=4 for all the plots. The K and values of Ru(NH 3 ) 6 3+ at all the thick films (>16A) were similar (K ca. 10 2 ). The analysis showed that the film thickness had no significant effect on the binding constant and on the density of the binding sites, at films thicker than 16 A . This indicated that the Ru(NH 3 ) 6 3+ film interactions were thickness independent and that the film structure did not significantly change with thickness. Therefore, the significant decline in sensitivity at 48 A films must not have reflected a change in probe-film interactions but rather was a consequence of transport limitations in thicker films (Witkowski and Brajter-Toth, 1992). The absence of response for uric acid in 0.5 M pH 7.0 potassium phosphate buffer at films thicker than 16 A (Table 6-1) probably reflected a combined effect of unfavorable interactions and inefficient transport of phosphate ions. The newly discovered counter ion based selectivity came at the expense of sensitivity (T able 61 ) where the response of uric acid was observed in unbuffered KC1 but not in phosphate buffer. The morphological changes in the OPPy film structure from templating as shown by similar binding constants (ca. 10 2 ) at templated and non-templated films were less apparent at the thicker films with Ru(NH 3 ) 6 3+ as a probe. The results also showed that electrolyte did not significantly influence the sensitivity of Ru(NH 3 ) 6 3+ (Table 6-1).

PAGE 142

128 Table 6-2 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid Calibration data at OPPy and OPPy/ATP Modified Macro GC Electrodes Ru(NH 3 ) 6 3+i Njmx (M) K (M’ 1 ) 16 A OPPyh 4.1 x KT 4 5.3 x 10 4 44 A OPPy 1.39 x 10' 2 7.94 x 10‘ 3 c 6.6 x 10 1 1.4 x 10 2 c 48 A OPPy 3.45 x 10' 3 3.8 x 10 2 16 A OPPy/ATP 3.71 x 10' 3 3.3 x 10 2 32 A OPPy/ATP 7.19 x 10' 3 1.3 x 10 2 44 A OPPy/ATP 3.02 x 1 O’ 3 2.22 x 10' 2 c 3.6 x 10 2 4.2 x 10 1 c Uric Acid* 16 A OPPy“ 6.0 x 10' 5 5.7 x 10 3 16 A OPPy/ATP d 5.50 x 104 2.90 x 10 2 32 A OPPy/ATP c 6.06 x 104 5.26 x 10 3 Note: Rs:0.98, n=4 ‘calibration data listed in Table 6-1 b from Scatchard plot analysis (Table 5-1) Langmuir isotherm analysis not valid see text c in 0.5 M KC1 “Table 5-2

PAGE 143

129 o o For uric acid Langmuir analysis the interactions at 32 and 44 A OPPy films and of 44A OPPy/ATP films could not be performed since the regression data had correlation coefficients less than 0.98. At 32 A OPPy/ATP film in 0.5 M KC1, was similar to that at the 16 A OPPy film in 0.5 M pH 7.0 potassium phosphate buffer. The K value was an order of a magnitude greater at the OPPy/ATP film in KC1 indicating stronger interactions of uric acid with the templated film in KC1. Sensitivity of Bare. OPPy and OPPv/ATP Modified Carbon Fiber Table 6-3 summarizes the calibration data at the carbon fiber UMEs fitted using linear regression through the origin. The results showed that the senstivity ratio of the bare fiber vs the theoretical sensitivity was greater than two for Ru(NH 3 ) 6 3+ . At GC, the sensitivity ratio was 6. 0/2. 7 (Table 4-1) and was due to Ru(NH 3 ) 6 3+ adsorption (Kovach et al., 1986). The high ratio at carbon fiber confirmed interactions of Ru(NH 3 ) 6 3+ with the fiber surface. After coating the carbon fiber with OPPy films Ru(NH 3 ) 6 3+ sensitivity decreased 49%. The decrease was consistent with that observed at thick film coated GC and likely resulted from transport limitations. The films at carbon fiber were thicker than at GC, ca. 0.53 pm. At OPPy/ATP modified UME the sensitivity was similar to that at the OPPy film as expected from the results for thicker film GC electrodes. The sensitivity was electrolyte independent for Ru(NH 3 ) 6 3+ . The sensitivity ratio for uric acid at carbon fiber (Table 6-3) compared to the theoretical sensitivity was approximately two, somewhat smaller than the ratio at GC (ca. 3.5)

PAGE 144

130 Table 6-3 Sensitivity of Ru(NH 3 ) 6 3+ , Uric Acid and Adenine at Ultramicroelectrode Surfaces Ru(NH,)« 3+b Uric Acid b Adenine b theoretical sensitivity 0 0.44 0.88 0.88 bare carbon fiber 2 0.91 ±0.01 1.95 ±0.05 0.37 ±0.07 OPPy 2 0.46 ± 0.06 1.19 ± 0.03 0.29 ±0.07 OPPy/ATP 2 0.3 ±0.1 d 0.52 ±0.02 OPPy/ ATP 20 0.4 ±0.1 2.3 ±0.6 0.5 ±0.1 Note: nA/mM, obtained from linear regression of the calibration data through the origin (plateaued portions were not fitted) 2 electrode radius 7pm, film thickness ca. 0.53 pm, scan rate 0.100 V/s all potentials vs SCE b analyte concentration 0. 1-1.0 mM in 0.5 M potassium phosphate buffer pH 7.0 ‘Equation 2.5, D=10' 6 cm 2 /s d no response over background e in 0.5 M KC1

PAGE 145

131 (Hsueh and Brajter-Toth, 1993), confirming adsorption of uric acid at the fiber. At OPPy modified carbon fibers uric acid could be detected although the sensitivity decreased 39% compared to the sensitivity at the bare UME. The decrease in sensitivity was similar to that observed for Ru(NH 3 ) 6 3+ . At OPPy/ ATP electrodes no response above background was apparent in the phosphate buffer electrolyte. The difference in response of templated and nontemplated films indicated morphological changes with templating of OPPy to which uric acid but not Ru(NH 3 ) 6 3+ were sensitive. Detection of uric acid at the thick OPPy coated UMEs indicated the possibility of a different film structure than on GC. In fact, uric acid sensitivity at the bare and coated UMEs was significantly higher than the sensitivity of Ru(NH 3 ) 6 3+ . At GC, the sensitivity of uric acid in comparison to the sensitivity of Ru(NH 3 ) 6 3+ was much greater at bare GC and lower at OPPy films. The high sensitivity of uric acid at OPPy coated UMEs indicated that uric acid interactions/transport were not limited at these films, possibly as a result of more efficient partitioning and transport into the film. The behavior was different than at GC where after coating with thicker OPPy films uric acid response was suppressed. For adenine, the response at the bare carbon fiber (Table 6-3) was below the theoretical sensitivity which was determined assuming a direct two electron oxidation of adenine (Dryhurst and Elving, 1968). The behavior was different than at GC where the sensitivity was significantly higher than theoretical. The reason might be that the complicated oxidation mechanism (Dryhurst and Elving, 1968) becomes more apparent with efficient mass transport to the UME. At OPPy coated UMEs, adenine response decreased 22% which was a somewhat smaller decrease than observed for the other probes at the OPPy coated fibers.

PAGE 146

132 The response of adenine at the OPPy/ATP films on carbon fiber in comparison to the sensitivity at OPPy films increased in contrast to the behavior of the other two probes. This result showed that templating OPPy films with ATP on carbon fiber produced more selective films to adenine. The selectivity was also observed at ultrathin OPPy/ATP films on GC described in Chapter 4. The films at the UMEs were selective to adenine while excluding uric acid and showing lower sensitivity to Ru(NH 3 ) 6 3+ (Table 6-3). At OPPy/ATP coated GC selectivity to adenine described in Chapter 4 was accompanied by an overall decrease in sensitivity; at UME greater selectivity was accompanied by increased sensitivity. The effect of electrolyte was explored to further investigate the origins of the selectivity of OPPy/ATP films on carbon fiber. At the film electrodes the response of Ru(NH 3 ) 6 3+ and adenine was not electrolyte dependent, as shown by the results in 0.5 M KC1 and 0.5 M pH 7.0 potassium phosphate buffer in Table 6-3. However, the response of uric acid was strongly dependent on the electrolyte. The significant result was a major increase in response of uric acid in 0.5 M KC1 while in the 0.5 M pH 7.0 potassium phosphate buffer no current was observed above the background. The increase in response indicated that the film selectivity could be modified by the electrolyte and that electrolyte transport/partitioning plays an important role in uric acid response as observed at OPPy coated GC. The absence of the same effect in the responses of Ru(NH 3 ) 6 3+ and adenine may be related to the fact that uric acid is the only anionic probe that was investigated while Fe(CN) 6 3 was excluded.

PAGE 147

133 Analysis of Film Interactions at OPPv modified UMEs Langmuir isotherm analysis of the calibration data is summarized in Table 6-4 for results where R^0.98 and n=6. All the calibration plots showed a plateau or curvature indicating validity of Langmuir analysis. The binding constants for Ru(NH 3 ) 6 3+ were generally simil ar to the values obtained at the thicker films on GC (K ca. 10 2 M' 1 , Table 6-2) indicating similar Ru(NH 3 ) 6 3+ interactions with the films. For uric acid the K values were much smaller at OPPy films on carbon fiber (K ca. 10 1 ) than at ultrathin films on GC (K ca. 10 3 M *). The smaller K values confirmed weaker interactions of uric acid with the OPPy modified carbon fiber indicating different surface environment than at GC. At the OPPy/ ATP films the values obtained for uric acid in 0.5 M KC1 agreed with the previously reported values (K ca. 10“M L ) at ultrathin films on GC in 0.5 M pH 7.0 potassium phosphate buffer indicating similar film interactions. Since uric acid response at OPPy/ ATP coated carbon fiber was not apparent in phosphate buffer, the micro structure of OPPy/ ATP at carbon fiber must be similar to OPPy/ ATP at GC, but film response must be dependent on the electrolyte. Fast Scan Voltammetry of Uric Acid The LOD values that were obtained at the UMEs are summarized in Table 6-5. The LOD values suggested that uric acid sensitivity could be very high at bare and OPPy modified carbon fiber. Fast scan voltammetry was used to determine if the LOD could be improved further in order to design a useful sensor for determining both basal (nanomolar) and peak

PAGE 148

134 Table 6-4 Langmuir Isotherm Analysis for Ru(NH 3 ) 6 3+ and Uric Acid Sensitivity data at Bare and OPPy Modified Carbon Fiber Ultramicroelectrodes Ru(NH 3 ) 6 3+ N.(M) K (M' 1 ) bare carbon fiber 1.65 x 10' 3 4.6 x 10 2 OPPy a a OPPy/ ATP 1.53 x lO' 3 1.89 x 10 3 OPPy/ATP b a a Uric Acid N^M) K (M' 1 ) bare carbon fiber a a OPPy 1.64 x 10' 2 7.4 x 10 1 OPPy/ATP b 2.15 x 10’ 3 5.9 x 10 2 Note: analyte concentration 0.05-1.0 mM in 0.5 M pH 7.0 potassium phosphate buffer, electrode radius 7pm, scan rate 0. 1 V/s. All potentials vs SCE, film thickness ca. 0.53 pm “Langmuir isotherm plots were nonlinear b in 0.5 M KC1

PAGE 149

135 Table 6-5 Limits of Detection at Bare and OPPy Modified Carbon Ultramicroelectrodes Ru(NH 3 ) 6 3+b Uric Acid b Adenine b bare carbon fiber 3 6.6 x 10' 5 1.2 x 10' 5 1.6 x Iff 4 OPPy 3 2.6 x 104 2.5 x 10' 5 2.1 x 10' 4 OPPy/ATP 3 3.0 x 104 N/A 5.8 x 10’ 5 OPPy/ATP 3 ’ 0 7.5 x 10' 5 2.6 x Iff 5 4.8 x Iff 4 Note: LOD (M) obtained by 3o/m (o is the standard deviation of the blank, m is the sensitivity) ‘sensitivity obtained from cyclic voltammetry (Table 6-3), electrode radius 7 pm, film thickness 0.53 (am, scan rate 0.1 V/s b analyte concentrations 0.05-1.0 mM in 0.5M pH 7.0 potassium phosphate buffer % 0.5MKC1

PAGE 150

136 (micromolar) levels of uric acid in vivo. One drawback of fast scan voltammetry is the high background current obtained at carbon electrodes. The functional groups on the carbon surface, i.e. quinones and the oxide groups, contribute to the large Faradaic background (Fagan et al., 1985) along with the large contribution from the physical charging of the electrical double layer. The surface functional groups can react during voltammetric experiments and the amount of surface oxygen can change during the experiment both causing large changes in the background current (Hsueh and Brajter-Toth, 1993). These changes in the background can cause problems in background subtraction since a stable background cannot be obtained. Background subtraction is essential in fast scan measurements to obtain an analytical signal. Hsueh and Brajter-Toth have established a carbon fiber pretreatment procedure to obtain a stable background (Hsueh and Brajter-Toth, submitted). They found that after thirty minutes of repeated cycling from -0.8 to 1.2 V vs SCE in 0.07 M pH 7.4 sodium phosphate buffer at 100 V/s, the background current at the carbon fiber electrode stabilizes (Hsueh and Brajter-Toth, submitted). Another problem associated with background subtraction is that the large background consumes most of the digital resolution of the oscilloscope during digitization of the data which leaves little resolution for the analyte signal. This lack of resolution will introduce “digitization noise” after background subtraction (Hsueh and Brajter-Toth, submitted). The goals of this work was to develop a less time consuming pretreatment procedure that would produce stable background currents similar to those obtained from the cycling procedure. The goal was to use OPPy films to improve the sensitivity and reusability and as a means to reduce the background currents associated with high scan rates and in turn reduce

PAGE 151

137 the "digitization noise”. Since the background currents seemed to be suppressed at GC macroelectrodes with OPPy films as discussed in Chapter 4, OPPy films on carbon fiber were considered for background suppression. The analysis was performed using OPPy coated carbon fiber because this surface was most stable compared to the OPPy/ ATP coated carbon fiber at slow scan rates and showed the best sensitivity to uric acid (Table 6-3). Electrode stability was verified by monitoring the response of 0.01 M Fe(CN) 6 3 ' in 0.5 M pH 7.0 potassium phosphate buffer. A response over background indicated that the film had degraded. OPPy films on UMEs had a lifetime of ca. fifteen measurements when used in a potential window of 0.0 to 1.2 V at a scan rate of 0. 1 V/s. The electrode was stored dry at room temperature between experiments. OPPy/ ATP films on UMEs at a scan rate of 0. 1 V/s were less reliable since some films only lasted for five measurements. The reliability of the OPPy and OPPy/ ATP films on UMEs in slow scan voltammetry can be represented by the standard deviations of the sensitivity values given in Table 6-3. The OPPy/ ATP films have the largest standard deviations as well as the smaller lifetime. The OPPy coated fiber could be used for at least ten studies at fast scan rates (100 to 1500 V/s) without any further pretreatment other than rinsing. A study could consist of up to 500 signal averaged scans for both background and analyte signal, but smaller numbers of averaged signals helped to preserve the films longer. In contrast to the OPPy coated UME, it was discovered that the bare carbon fiber had to be pretreated by cycling before each use in fast scan measurements to obtain a good response. The cyclic voltammogram for 10 pM Ru(NH 3 ) 6 3 * in 0.07 M pH 7.4 sodium phosphate buffer at pretreated carbon fiber at a scan rate of 1000 V/s is shown in Figure 6-1 . The

PAGE 152

138 Z3 O -5 -10 _-|5 i i l i l i 1 1 1 1 1 1 1 1 1 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.8 potential (V) Figure 6-1 Background subtracted cyclic voltammogram of 10 pM Ru(NH 3 ) 6 3+ in 0.070 M pH 7.4 sodium phosphate buffer at electrochemically pretreated carbon fiber UME (see text for details), electrode radius 7 pm, scan rate lOOOV/s, 250 scans signal averaged.

PAGE 153

139 potential (V) Figure 6-2 Background subtracted cyclic voltammogram of 10 gM Ru(NH 3 ) 6 3+ in 0.070 M pH 7.4 sodium phosphate buffer at OPPy coated UME, electrode radius 7 gm, film thickness ca. 0.53 gm, scan rate 500 V/s, 500 scans signal averaged

PAGE 154

140 electrode was first pretreated by HsuehÂ’s method of cycling described earlier before background signals were signal averaged for background subtraction. Well defined anodic and cathodic peaks at ca. -0.3 V and -0.49 V respectively are illustrated. Figure 6-2 which shows 10 pM Ru(NH 3 ) 6 3+ at OPPy coated carbon fiber at a scan rate of 500 V/s illustrates that similar results can be obtained at OPPy modified electrodes as at carbon fiber electrodes pretreated by cycling. At OPPy films it was found that the currents decreased at a higher scan rate (1000 V/s) which could be due to transport problems. The results for 10 uM Ru(NH 3 ) 6 3+ at pretreated carbon and OPPy coated carbon fiber are summarized in Table 6-6 for scan rates of 500 and 1000 V/s. The data without error bars represent preliminary measurements, so only general conclusions can be made. The measurements were signal averaged before performing background subtraction in order to improve the response. It was found that 250 averaged scans each for the background and the analyte current at bare carbon fiber produced a good response. Larger number of scans for the background or larger number of scans for the analyte were attempted, but more scans were time consuming and showed little improvement in S/N, and fewer scans produced poor S/N since not enough scans were available to obtain a good averaged signal. For the OPPy coated carbon fiber electrodes fewer scans (ca. 10) were used in signal averaging to preserve the film. More than 100 scans caused the film to degrade rapidly from the surface as verified by large currents for 0.010 M Fe(CN) 6 3 ' in 0.5 M pH 7.0 potassium phosphate buffer. Fewer scans (less than 100) still showed good S/N for the OPPy coated electrodes. Cyclic voltammetric responses for 250 averaged scans at 500 and 1000 V/s are shown for 0.3 mM uric acid in 0.07 M pH 7.4 sodium phosphate buffer at pretreated carbon fiber in

PAGE 155

141 Table 6-6 Background Subtracted Fast Scan Voltammetry Results 10 pM Ru(NH 3 ) 6 3+ in 0.07 M pH 7.4 sodium phosphate buffer scan rate (V/s) AE p b ip(nA)‘ carbon fiber 2 500 200 ± 100 3.4 ±0.6 1000 190 ± 18 7 ± 2 OPPy 500 152 4.8 1000 234 1.3 0.3 mM Uric Acid in 0.07 M pH 7.4 sodium phosphate buffer scan rate (V/s) AE p b ip(nA)‘ carbon fiber 2 500 580 ± 10 21 ±3 1000 632 17.2 OPPy 500 439 2.85 1000 616 4.84 ±0.02 Note: electrode radius 7 pm, film thickness ca. 0.53 pm, all potentials vs SCE a pretreated by cycling from -0.8 to 1.2 V vs. SCE at 100 V/s in 0.070 M pH 7.4 sodium phosphate buffer for 30 minutes b difference between the anodic and cathodic peak potentials c peak current measured at cathodic peak

PAGE 156

142 potential (V) Figure 6-3 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phosphate buffer at pretreated UME (see text). Electrode radius 7pm, scan rate 500 V/s, 250 scans signal averaged.

PAGE 157

143 potential (V) Figure 6-4 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phosphate buffer at pretreated UME (see text). Electrode radius 7pm, scan rate 1000 V/s, 250 scans signal averaged.

PAGE 158

144 potential (V) Figure 6-5 Background subtracted cyclic voltammogram of 0.3 mM Uric Acid in 0.070 M pH 7.4 sodium phosphate buffer at OPPy coated UME. Electrode radius 7 pm, film thickness ca. 0.53 pm, scan rate 500 V/s, 50 scans signal averaged.

PAGE 159

145 Figures 6-3 and 6-4 respectively. In this case, there was a small sensitivity decrease (Table 6-6) with increasing scan rate possibly reflecting kinetic effects since at faster scan rates the peaks were also broader. With the OPPy modified electrodes, although the sensitivity for uric acid decreased, the cyclic voltammogram at 500 V/s (Figure 6-5) shows well defined peaks with only 50 scans used in the signal averaging. The kinetics of uric acid appear faster at the OPPy film shown by smaller AE p values. Although the sensitivity of uric acid at OPPy decreased compared to bare UME with a decrease in the number of signals averaged (250 to 50), the response for uric acid can be obtained at OPPy coated UME without lengthy electrode pretreatment procedures such as cycling. The response of the OPPy film at 1 OOOV/s obtained from 200 signal averaged scans was not as well defined more than likely as a result of poor background subtraction. At 500 V/s AE p values were smaller at OPPy films for both probes compared to pretreated carbon fiber indicating that better kinetics could result at OPPy films than at pretreated carbon fiber possibly because of the greater activity of the less pretreated carbon surface which was used to coat OPPy films. For Ru(NH 3 ) 6 3 the sensitivity at OPPy films was 41% larger than at extensively pretreated bare carbon fiber, showing that at 1000 V/s the OPPy film improved the sensitivity for Ru(NH 3 ) 6 3+ . For uric acid, the sensitivity decreased 72% at OPPy in comparison to the bare UME at 1000 V/s indicating that the OPPy film did not favor interactions with uric acid. The preliminary results with OPPy films for fast scan measurements predicted that improvement in sensitivity over the results of slow scan measurements could be achieved by using a combination of fast scan rates and favorably

PAGE 160

146 interacting films which limit the need to pretreat the electrode surface before each measurement and limit the number of scans used in signal averaging. Conclusions Modification of carbon fiber electrodes with templated OPPy/ATP films produced a film more selective to adenine in phosphate buffer than the OPPy/ATP films on GC since the response of uric acid was suppressed. Despite this, the sensitivity of uric acid was much greater at bare and OPPy coated carbon fiber electrodes than for Ru(NH 3 ) 6 3 or adenine. Even though the response for uric acid was suppressed in phosphate buffer at OPPy/ATP electrodes, excellent sensitivity was apparent in unbuffered KC1. This good sensitivity of uric acid was exploited for use in fast scan measurements. Preliminary results showed that the OPPy film coating could possibly replace the lengthy cycling procedure necessary to obtain a stable background before each set of measurements at carbon fiber electrodes. Stabilization of the background could be presumably from the suppression of background currents seen at the OPPy coated GC electrodes.

PAGE 161

CHAPTER 7 CONCLUSIONS AND FUTURE WORK The main objective of this work was to design an amperometric biosensor for adenosine. In order to achieve this goal ultrathin films of overoxidized polypyrrole grown on graphite surfaces were templated with purine molecules in order to enhance the selectivity and sensitivity of the carbon surface. While the surfaces developed were not sensitive enough for routine use as sensors, characterization of these films provided useful information about surface interactions with OPPy films and purine molecules like uric acid and adenine and furthered development of new methods for electrode pretreatment for use in designing sensors and controlling adenosine response. In general OPPy films without templates at GC showed cation permselectivity and good sensitivity to adenine. Templating OPPy films on GC produced new film structures with slightly increased selectivity to adenine. Both templated and nontemplated OPPy films suppressed large background currents allowing detection of adenosine and improved reproducibility of measurements at GC considering that the film electrodes did not have to be renewed after each measurement. UV analysis confirmed incorporation of ATP into polypyrrole during the polymerization process and verified ATP release during overoxidation of polypyrrole. ESCA analysis showed that HsuehÂ’s polymerization/overoxidation procedure produced a good 147

PAGE 162

148 surface coverage by the film and that templating OPPy films produced a different microenvironment, possibly more compact than for OPPy films without the templates. ESCA analysis should also be done on OPPy/ado or OPPy/ino films to determine if these microstructures are significantly different than OPPy/ ATP films. A compact film structure with templating was also confirmed by low apparent diffusion coefficients determined from rotating disk electrode measurements. Langmuir and Scatchard analysis of the films at GC indicated reasonably strong interactions of the probes at bare and OPPy film electrodes. Templating decreased the strength of these interactions which points to a different possibly more compact, hydrophobic structrue for the templated films. Modification of carbon fiber electrodes with OPPy/ ATP films produced a film more selective to adenine in phosphate buffer than the OPPy/ ATP films on GC. A new type of selectivity to uric acid based on counterion exclusion was discovered. The response of uric acid was suppressed at OPPy/ ATP in phosphate buffer, but excellent sensitivity was apparent in KC1. The sensitivity for uric acid at the bare and OPPy modified carbon fiber electrodes was much greater than for Ru(NH 3 ) 6 3+ or adenine. This good sensitivity along with the fact that OPPy films were shown to suppress large background currents was exploited for use in fast scan measurements. Preliminary results showed that the OPPy film coating could possibly replace the lengthy cycling procedure necessary to obtain a stable background before each set of measurements. Further work should include a more thorough investigation of film coatings for stabilization of background currents in fast scan voltammetry. It should also be determined

PAGE 163

149 if further selectivity can be achieved with the use of templated films at fast scan rates. Next, this methodology should be applied for determining uric acid concentrations in cardiac perfusates since fast scan rates will discriminate against interferants such as ascorbic acid and allow micromolar detection limits. Finally, the concept of templating should be applied to other types of polymers that can be electropolymerized at the surface and which more closely resemble the binding site environment of adenosine to enhance selectivity to this molecule and to continue the original goals of this work.

PAGE 164

LIST OF REFERENCES Ahmad, K. and Brajter-Toth, A., Effect of ultramicroelectrode array surface structure on analytical figures of merit of structurally different compounds, Anal. Chim. Acta 1992 , 259, 139-144. Alexandratos, S.; Grady, C.; Merda, P.; Kaiser, P., Binding constants as a probe of the microenvironment in polymer-supported reagents, J. Am. Chem. Soc. 1992 , 114, 3078-3080. American Heart Association. Research Facts-Update 1994 Pamphlet, AMA: Dallas, 1993. Andersson, L.; OÂ’Shannessy, D.; Mosbach, K., Molecular recognition in synthetic polymers: preparation of chiral stationary phases by molecular imprinting of amino acid amides, J. Chromatogr. 1990 , 513, 167-179. Armengaud, C.; Moisy, P.; Bedioui, F.; Devynck, J., Electrochemistry of conducting polypyrrole films containing cobalt porphyrin, J. Electroanal. Chem. 1990 , 277, 197211 . Arshady, R. and Mosbach, K., Synthesis of substrate-selective polymers by host-guest polymerization, Makromol. Chem. 1981 , 182, 687-692. Asavapiriyanont, S.; Chandler, G. Gunawardena, G.; Pletcher, D., The electrodeposition of pyrrole films from aqueous solutions, J. Electroanal. Chem. 1984 , 177, 229-244. Baer, H.; Drummond, G.; Duncan, E., Formation and deamination of adenosine by cardiac muscle enzymes. Mol. Pharmacol. 1966 , 2, 67-76. Bailey, F.; Malinski, T.; Kiechle, F., Carbon fiber ultramicroelectrodes modified with conductive polymeric tetrakis (3-methoxy-4-hydroxyphenyl) porphyrin. Anal. Chem. 1991 , 63, 395-398. Bard, A. and Faulkner, L., Electrochemical Methods ; John Wiley & Sons: New York, 1980, chapter 6. Barisci, N. and Wallace, G., unpublished work. 150

PAGE 165

151 Baur, I; Kristensen, E.; May, L.; Wiedemann, D.; Wightman, R., Fast-scan voltammetry of biogenic amines. Anal. Chem. 1988 , 60, 1268-1272. Beck, F.; Braun, P.; Oberst, M., Organic electrochemistry in the solid state-overoxidation of polypyrrole, Ber. Bunges. Phys. Chem. 1987 , 91, 967-974. Belardinelli, L.; Linden, J. Berne, R. The cardiac effects of adenosine, Prog. Cardiovasc. Dis. 1989 , 32, 73-97. Belle, H., Uptake and deamination of adenosine by blood, species differences, affect of pH, ions, temperature and metabolic inhibitors, Biochim. Biophys. Acta 1969 , 192, 124132. Bendich, A.; Brown, G.; Philips, F.; Thiersch, J., The direct oxidation of adenine in vivo, J. Biol. Chem. 1950 , 183, 267-271. Berne, R, Cardiac nucleosides in hypoxia: possible regulation in coronary blood flow, Am. J. Physiol. 1963 , 204, 317-322. Berne, R; Cumish, R.; Gidday, J.; Rubio, R., Measurement of Femtomolar Concentrations of Adenosine, J. Liq. Chrom. 1986 , 9, 113-119. Bradley, C. and Rechnitz, G., Kinetic analysis of enzyme electrode response. Anal. Chem. 1984 , 56, 664-667. Bradner, F.; Shapiro, J., Improvement in the quality of polypyrrole films prepared electrochemically on a mercury anode, Synth. Met. 1988, 26, 69-77 . Brajter-Toth, A.; Goyal, R.; Wrona, R. Lacava, T.; Nguyen, N.; Dryhurst, G., Electrochemical and enzymatic oxidation of biological purines, Bioelectrochem. Bioenerg. 1981 , 8, 413-435. Brown, T., Ed , Molecular Biology, Bios Scientific Publishing: Oxford, 1991. Browne, K.; He, G.; Bruice, T., Microgonotropens and their interactions with DNA, J. Am. Chem. Soc. 1993 , 115, 7072-7079. Bull, R.; Fan, F.; Bard, A., Polymer films on electrodes, J. Electrochem. Soc. 1982 , 129, 1009-1015.

PAGE 166

152 Castro, A.; Bhattacharyya, K.; Eisenthal, K., Energetics of adsorption of neutral and charged molecules at the air/water interface by second harmonic generation: hydrophobic and solvation effects, J. Chew. Phys. 1991 , 95, 1310-1315. Centonze, D.; Guerrieri, A.; Malitesta, C.; Palmisano, F.; Zambonin, P., Interference-free glucose sensor based on glucose-oxidase immobilized in an overoxidized nonconducting polypyrrole film, Fres. J. Anal. Chem. 1992 , 342, 729-733. Chen, C.; Wei, C.; Rajeshwar, K., Flow assay of ions at chemically modified electrodes: the polypyrrole/Fe(CN)6 4 Â’ model system, Anal. Chem. 1993 , 65, 2437-2442. Chen, T.; Strein, T.; Abe, T.; Ewing, A., Voltammetry of adenosine after electrochemical treatment of carbon-fiber electrodes, Electroanal. 1994, 6, 746-751. Cheng, Q. and Brajter-Toth, A., Selectivity and sensitivity of self-assembled thioctic acid electrodes, Anal. Chem. 1992 , 64, 1998-2000. Cheung, K.; Bloor, D.; Stevens, G., Characterization of polypyrrole electropolymerized on different electrodes, Polymer 1988 , 29, 1709-1717. Christensen, P. and Hamnett, A., In situ spectroscopic investigations of the growth, electrochemical cycling and overoxidation of polypyrrole in aqueous solution, Electrochim. Acta 1991 , 36, 1263-1286. Clark, D. and Harrison, A., ESCA applied to polymers XXXI. A theoretical investigation of molecular core binding and relaxation energies in a series of prototype systems for nitrogen and oxygen functionalities in polymers, J. Polym. Sci., Polym. Chem. Ed. 1981 , 19, 1945-1955. deJong, J., Phosphorylation and deamination of adenosine by the isolated, perfused rat heart, Biochim. Biophys. Acta 1972 , 286, 252-259. Deng, I. and Enke, C. Adenosine-selective electrode, Anal. Chem. 1980, 52, 1937-1940. Diaz, A. and Castillo, J., A polymer electrode with variable conductivity, polypyrrole, J.C.S. Chem. Comm., 1980 , 397-398. Dong, S.; Sun, Z.; Lu, Z., Chloride chemical sensor based on an organic conducting polypyrrole polymer, Analyst 1988 , 113, 1525-1528. Dow, J. CRC Isolated Adult Cadiomyocytes Structure and Metabolism, Metabolism of Purines by Adult Cardiomyocytes\ CRC Press: Boca Raton, 1989.

PAGE 167

153 Drury, A. and Szent-Gyorgyi, A. The physiological activity of adenine compounds with special reference to their action upon the mammalian heart, J. Physiol. 1929, 68, 213237. Dryhurst, G., Dicarbonium ions as products of electrochemical oxidation of biologically important purines at the pyrolytic graphite electrode, J. Electrochem. Soc. 1969, 116, 1411-1412. Dryhurst, G. Primary products of electrochemical oxidation of uric acid in aqueous and methanolic solution, J. Electrochem. Soc. 1971, 118, 699-701. Dryhurst, G., Electrochemical determination of adenine and adenosine, Talanta 1972, 19, 769-787. Dryhurst, G. and De, P., Direct electrochemical methods for determination of allopurinol and uric acid mixtures. Adsorption of uric acid at pyrolytic graphite electrodes. Anal. Chim. Acta 1972, 58, 183-191. Dryhurst, G. and Elving, P., Electrochemical oxidation of adenine: reaction products and mechanisms, J. Electrochem. Soc. 1968, 115, 1014-1020. Dryhurst, G.; Nguyen, N.; Wrona, M.; Goyal, R.; Brajter-Toth, A.; Owens, J; Marsh, H., Elucidation of the biological redox chemistry of purines using electrochemical techniques, J. Chem. Edu. 1983, 60, 315-319. Edelman, P. and Wang, J., in Biosensors and Chemical Sensors: Optimizing Performance Through Polymeric Materials', ACS: Washington, DC., 1992. Ekberg, B. and Mosbach, K., Molecular imprinting: a technique for producing specific separation materials. Trends in Biotechnol. 1989, 7, 92-96. Engstrom, R., Electrochemical pretreatment of glassy carbon electrodes. Anal. Chem. 1982, 54, 2310-2314. Engstrom, R. and Strasser, V., Characterization of electrochemically pretreated carbon electrodes, Anal. Chem. 1984, 56, 136-141. Fagan, D.; Hu, I.; Kuwana, T., Vacuum heat treatment for activation of glassy carbon electrodes^! nal. Chem. 1985, 57, 2759-2763. Feldman, B.; Burgmayer, P.; Murray, R., Potential dependence of electrical conductivity and chemical charge storage of polypyrrole films, J. Am. Chem. Soc. 1985, 107, 872-878.

PAGE 168

154 Ferreira, C.; Aeiyach, S.; Delamar, M.; Lacaze, P., Electropolymerization of pyrrole on iron electrodes,/. Electroanal. Chem. 1990, 284, 351-369. Fleischmann, M.; Pons, S.; Rolison, D.; Schmidt, P., Ultramicroelectrodes\ Datatech System, Inc.: Morganton, N.C., 1987. Fredholm, B., Analysis of purines, Life Sci. 1987, 41, 837-840. Freund, M.; Bodalbhai, L.; Brajter-Toth, A., Anion-excluding polypyrrole films, Talanta 1991,55, 95-99. Freund, M. and Brajter-Toth, A., Semiintegral analysis in cyclic voltammetry: determination of surface excess and concentration in the presence of weak adsorption and thin films, J. Phys. Chem. 1992, 96, 9400-9406. Gao, Z.; Chen, B.; Zi, M., Voltammetric determination of dopamine in the presence of ascorbic acid at over-oxidized polypyrrole-indigo carmine film-coated electrodes, Analyst, 1994, 119, 459-464. Gao, Z. and Ivaska, A., Electrochemical behavior of dopamine and ascorbic acid at overoxidized polypyrrole(dodecyl sulphate) film-coated electrodes, Anal. Chim. Acta 1993, 284, 393-404. Gao, Z.; Zi, M.; Chen, B., Permeability controllable overoxidised polypyrrole film modified glassy carbon electrodes, Anal. Chim. Acta, 1994, 286, 213-218. Gao, Z.; Zi, M.; Chen, B., The influence of overoxidation treatment on the permeability of polypyrrole films, J. Electroanal. Chem. 1994, 373, 141-148. Gardiner, D., A rapid and sensitive fluorimetric assay for adenosine, inosine, and hypoxanthine, Anal. Biochem. 1979, 95, 377-382. Ge, H.; Qi, G.; Kang, E.; Neoh, K., Study of overoxidized polypyrrole using X-ray photoelectron spectroscopy, Polymer 1994, 35, 504-508. Gerlach, E.; Deuticke, B.; Dreisbach, R., Nucleotide cleavage in heart muscle by O deficiency and its possible importance for coronary blood circulation, Naturwissenschaften 1963, 50, 228-229. Gerhardt, G.; Oke, A_; Nagy, G.; Moghaddam, B.; Adams, R., Nafion-coated electrodes with high selectivity for CNS electrochemistry, Brain Res. 1984, 290, 390-395. Goodfellow, Metals arid Materials Catalog, Malvern, 1990.

PAGE 169

155 Goyal, R.; Brajter-Toth, A.; Dryhurst, G., Further insights into electrochemical oxidation of uric acid, J. Electroanal. Chem. 1982 , 131, 181-202. Gough, D. and Leypoldt, J., Membrane covered, rotated disk electrode. Anal. Chem. 1979, 51, 439-444. Gough, D. and Leypoldt, J., Rotated, membrane-covered oxygen electrode, Anal. Chem. 1980 , 52, 1126-1130. Guadalupe, A. and Abruna, H., Electroanalysis with chemically modified electrodes, Anal. Chem. 1985 , 57, 142-149. Harris, D., Quantitative Chemical Analysis, W. H. Freeman and Company: New York, 1991, 3rd ed. Heinze, J., Ultramicroelectrodes in electrochemistry, Angew. Chem. Int. Ed. Engl. 1993, 32, 1268-1288. Henderson, R. and Griffin, C., Electrochemical detection of adenosine and other purine metabolites during high-performance liquid chromatographic analysis, J. Chrom. 1984 , 298, 231-242. Hogan, M.; Dattagupta, N.; Crothers, D., Transmission of allosteric effects in DNA, Nature 1979 , 278, 521-524. Howell, J.; Kuhr, W.; Ensman, R.; Wightman, R., Background subtraction for rapid scan voltammetry, J. Electroanal. Chem. Interfacial Electrochem. 1986 , 209, 77-90. Hsueh, C. and Brajter-Toth, A., Fast scan voltammetry in aqueous solutions at carbon fiber ultramicroelectrodes with on-line iR compensation. Anal. Chem. 1993, 65, 15701575. Hsueh, C. and Brajter-Toth, A., Electrochemical preparation and analytical applications of ultrathin overoxidized polypyrrole films, Anal. Chem. 1994, 66, 2458-2464. Hsueh, C. and Brajter-Toth, A., Enhancement of selectivity and sensitivity in analysis with fast scan voltammetry at scan rates above 1000 V/s, Anal. Chem. submitted. Hsueh, C. and Brajter-Toth, A., A simple current transducer for UME measurements at a wide range of time scales, Anal. Chim. Acta in press. Huang, H.; He, P.; Faulkner, L., Current multiplier for use with ultramicroelectrodes, Anal. Chem. 1986 , 58, 2889-2891.

PAGE 170

156 Hutchins, R. And Bachas, L., Nitrate selective electrode developed by electrochemically mediated imprinting/doping of polypyrrole. Anal. Chem. 1995, 67, 1654-1660. Ikariyama, Y. and Heineman, W., Polypyrrole electrode as a detector for electroinactive anions by flow injection analysis. Anal. Chem. 1986 , 58, 1803-1806. Imisides, M.; John, R.; Riley, P.; Wallace, G., The use of electropolymerization to produce new sensing surfaces: a review emphasizing electrodeposition of herteroaromatic compounds, Electroanal. 1991 , 3, 879-889. Jennings, R.; Reimer, K.; Hill, M.; Mayer, S. Total ischemia in dog hearts, in vitro. 1. Comparison of high energy phosphate production, utilization, and depletion, and of adenine nucleotide catabolism in total ischemia in vitro vs severe ischemia in vivo. Circ. Res. 1981, 49, 892-900. Jiang, R. and Anson, F., Association of electroactive counterions with polyelectrolytes. 5. Electrostatic binding of Ru(NH 3 ) 6 3+ and Co(NH 3 ) 6 3+ to polyacrylates, J. Phys. Chem. 1992 , 96, 10565-10571. Kamau, G., Surface preparation of glassy carbon electrodes, Anal. Chim. Acta 1988 , 207, 116. Kempe, M. and Mosbach, K., Chiral recognition of N-alpha-protected amino-acids and derivatives in noncovalently molecularly imprinted polymers, Int. J. Peptide Protein Res. 1994 , 44, 603-606. Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties', Wiley & SonÂ’s: New York, 1988. Ko, J.; Rhee, H.; Park, S.; Kim, C., Morphology and electrochemical properties of polypyrrole films prepared in aqueous and nonaqueous solvents, J. Electrochem. Soc. 1990 , 137, 905-909. Kovach, P.; Deakin, M.; Wightman, R., Electrochemistry at partially blocked carbon-fiber microcylinder electrodes, J. Phys. Chem. 1986 , 90, 4612-4617. Kuhr, W.; Barrett, V.; Gagnon, M.; Hopper, P.; Pantano, P., Dehydrogenase-modified carbon-fiber microelectrodes for the measurement of neurotransmitter dynamics, Anal. Chem. 1993 , 65, 617-622. Kuhr, W. and Wightman, R., Real-time measurement of dopamine release in rat brain, Brain Res. 1986 , 381, 168-171.

PAGE 171

157 Kuhr, W. and Yeung, E., Optimization of sensitivity and separation in capillary zone electrophoresis with indirect fluorescence detection, Anal. Chem. 1986 , 60, 26422646. Leddy, J.; Bard, A.; Maloy, J.; Saveant, J., Kinetics of film-coated electrodes. Effect of a finite mass transfer rate of substrate across the film-solution interface at steady state, J. Electroanal. Chem. 1985 , 1 87, 205-227 . Li, Y. and Dong, S., Electrochemically controlled release of adenosine 5 '-triphosphate from polypyrrole film, J. Chem. Soc. Chem. Commun. 1992, 827-828. Limbird, L. Cell Surface Receptors: A Short Course on Theory and Methods', Martinus Nijhoff Publishing: Boston, 1986. Liu, D.; Meyerhoff, M.; Goldberg, H.; Brown, R., Potentiometric ionand bioselective electrodes based on asymmetric polyurethane membranes. Anal. Chim. Acta 1993 , 274, 37-46. Lyons, M.; Lyons, C.; Fitzgerald, C.; Bannon, T., Polypyrrole based amperometric sensors: theory and characterization. Analyst 1993, 118, 361-369. Marino, A., Dynamic Modification of Graphite Surfaces with Surfactants for Electrochemical Detection of Biological Molecules, Ph.D. Dissertation, University of Florida, Gainesville, 1994. Marino, A. and Brajter-Toth, A., Ionic surfactants as molecular spacers at graphite electrodes. Anal. Chem. 1993 , 65, 370-374. McCreery, R. In Electroanalytical Chemistry', Bard, A., Ed.; Marcel Dekker: New York, 1991;Vol. 17; pp. 221-364. McGhee, J. and von Hippel, P., Theoretical aspects of DNA-protein interactions: cooperative and non-co-operative binding of large ligands to a one-dimensional homogeneous lattice, J. Mol. Biol. 1974 , 86, 469-489. Millan K. and Mikkelsen, S., Sequence-selective biosensor for DNA based on electroactive hybridization indicators, Anal. Chem. 1993 , 65, 2317-2323. Millar, J.; Armstrong-James, M.; Kruk, Z., Polarographic assay of iontophoretically applied dopamine and low-noise unit recording using a multibarrel carbon fiber microelectrode, Res. 1981 , 205, 419-424.

PAGE 172

158 Mitchell, G.; Davis, F.; Legge, C., The effect of dopant molecules on the molecular order of electrically-conducting films of polypyrrole, Synth. Met. 1988, 26, 247-257. Murray, R., Ed. Molecular Design of Electrode Surfaces, Wiley-Interscience Publication, 1992. Nagy, G.; Gerhardt, G.; Oke, A.; Rice, M.; Adams, R.; Moore, R.; Szentirmay, M.; Martin, C., Ion exchange and transport of neurotransmitters in Nafion films on conventional and microelectrode surfaces, J. Electroanal. Chem. 1985, 188, 85-94. Namm, D., Myocardial nucleotide synthesis from purine bases and nucleosides. Circulation Res. 1973, 33, 686-695. Neely, J. Morgan, H. Relationship between carbohydrate and lipid metabolism and energy balance of the heart, Annu. Rev. Physiol. 1974, 36, 413-459. Nishizawa, M.; Shibuya, M.; Sawaguchi, T.; Matsue, T.; Uchida, I., Electrochemical preparation of ultrathin polypyrrole film at microarray electrodes, J. Phys. Chem. 1991, 95, 9042-9044. Norrlow, O.; Mansson, M.; Mosbach, K., Swedish patent application no. 8605372-5, 1986. Ogasawara, M.; Funahashi, K.; Demura, T.; Hagiwara, T.; Iwata, K., Enhancement of electrical conductivity of polypyrrole by stretching, Synth. Met. 1986, 14, 61-69. Ontyd, J. and Schrader, J., Measurement of adenosine, inosine, and hypoxanthine in human plasma, J. Chrom. 1984, 307, 404-409. Owens, J.; Henry, A.; Dryhurst, G., An investigation by cyclic voltammetry, double potential step chronoamperometry and thin-layer spectroelectrochemistry, J. Electroanal. Chem. 1978,97, 231-247. Palmisano, F.; Malitesta, C.; Centonze, D.; Zambonin, P., Correlation between permselectivity and chemical structure of overoxidized polypyrrole membranes used in electroproduced enzyme biosensors, Anal. Chem. 1995, 67, 2207-2211. Pfluger, P.; Krounbi, M.; Street, G.; Weiser, G., The chemical and physical properties of pyrrole-based conducting polymers: the oxidation of neutral polypyrrole, J. Chem. Phys. 1983, 78, 3212-3218. Pfluger, P. and Street, G., Chemical, electronic, and structural properties of conducting heterocyclic polymers: a view by XPS, J. Chem. Phys. 1984, 80, 544-553.

PAGE 173

159 Poon, M. and McCreery, R., In situ laser activation of glassy carbon electrodes, Anal. Chem. 1986 , 58, 2745-2750. Poon, M. and McCreery, R, Repetitive in situ renewal and activation of carbon and platinum electrodes: applications to pulse voltammetry. Anal. Chem. 1987 , 59 , 1615-1620. Poon, M.; McCreery, R; Engstrom, R., Laser activation of carbon electrodes. Relationship between laser-induced surface effects and electron transfer activation, Anal. Chem. 1988 , 60, 1725-1730. Preston, M., Determination of adenine, adenosine and related nucleotides at the low picomole level by reversed-phase high-performance liquid chromatography with fluorescence detection, J. Chrom. 1983 , 275, 178-182. Pyo, M.; Maider, G.; Kennedy, R.; Reynolds, J., Controlled release of biological molecules from conducting polymer modified electrodes. The potential dependent release of adenosine 5 '-triphosphate from poly(pyrrole adenosine 5'-triphosphate) films, J. Electroanal. Chem. 1994 , 368, 329-332. Reimer, K.; Jennings, R. Myocardial ischemia, hypoxia and infarction. In The Heart and Cariovascular System, Fozzard, H., Ed.; Raven Press: New York, 1986; 1133-1201. Salmon, M.; Kanazawa, D.; Diaz, A.; Krounbi, M., A chemical route to pyrrole polymer films, J. Polym. Sci., Polym. Lett. Ed. 1982 , 20, 187-193. Sawyer, D. and Roberts, J. Experimental Electrochemistry for Chemists, John Wiley & Sons: New York, 1974. Scatchard, G., Molecular interactions in protein solutions, Ann. N.Y. Acad. Sci. 1949 , 51, 660-681. Schrader, J. Sites of action and production of adenosine in the heart. In Purinergic Receptors', Bumstock, G., Ed.; Chapman and Hall: London, 1981; 120-162. Schrader, J. Metabolism of adenosine and sites of production in the heart. In Regulatory Function of Adenosine', Berne, R.; Rail, R.; Rubio, R., Eds.; Martinus/Nijhoff Publishers: The Hague, 1983; 133-156. Schrader, J.; Baumann, G.; Gerlach, E., Adenosine as inhibitor of myocardial effects of catecholamines, Pfliigers Arch. 1977 , 372, 29-35. Scofield, J., Theoretical photoionization cross sections from 1 to 1500 eV, J. Electron Spectrosc.Relat. Phenom. 1976 , 8, 129-137.

PAGE 174

160 Sellergren, B.; Ekberg, B.; Mosbach, K., Molecular imprinting of amino acid derivatives in macroporous polymers, demonstration of substrateand enantio-selectivity by chromatographic resolution of racemic mixtures of amino acid derivatives, J. Chromatogr. 1985 , 347, 1-10. Shea, K.; Spivak, D.; Sellergren, B., Polymer complements to nucleotide bases. Selective binding of adenine derivatives to imprinted polymers, J. Am. Chem. Soc. 1993 , 115, 3368-3369. Skotheim, T., Ed. Handbook of Conducting Polymers, Marcel Dekker: New York, 1986; Vol. 1. Soderback, U.; Sollevi, A.; Fredholm, B., The disappearance of adenosine from blood and platelet suspension in relation to the platelet cyclic AMP content, Acta Physiol. Scand. 1987 , 129, 189-194. Stackelberg, M.; Pilgram, M.; Toome, V., Determination of diffusion coefficients of ions in aqueous solution in the presence of foreign electrolytes, Z. Electrochem. 1953 , 57, 342-347. Street, G.; Clarke, T.; Geiss, R.; Lee, V.; Nazzal, A.; Pfluger, P.; Scott, J., Characterization of polypyrrole, J. Phys. Colloq. 1983 , 44, 599-606. Street, G.; Clarke, T.; Krounbi, M.; Kanazawa, K.; Lee, V.; Pfluger, P ; Scott, I; Weiser, G., Preparation and characterization of neutral and oxidized polypyrrole films, Mol. Cryst. Liq. Cryst. 1982 , 83, 1285-1296. Struck, W.; Elving, P., Electrolytic oxidation of uric acid: products and mechanism , Biochem. 1965 , 4, 1343-1353. Takakubo, M., Electrochemical polymerization of pyrrole in aqueous solutions, Synth. Met. 1987 , 18, 53-58. Tanaka, K.and Kashiwagi, N., In vivo voltammetry with an ultramicroelectrode, J. Electroanal.Chem. Bioelectrochem. Bioenerg. 1989 , 275, 95-98. Tohda, K.; Umezawa, Y.; Yoshiyagawa, S.; Hashimoto, S.; Kawasaki, M., Cation permselectivity at the phase boundary of ionophore-incorporated solvent polymeric membranes as studied by optical second harmonic generation, Anal. Chem. 1995 , 67, 570-577. Turner, A.; Karube, I.; Wilson, G., Biosensors Fundamentals and Applications, Oxford University Press: Oxford, 1987.

PAGE 175

161 Vlatakis, G.; Andersson, L.; Muller, R.; Mosbach, K., Drug assay using antibody mimics made by molecular imprinting, Nature 1993 , 361, 645-647. Voet, D. and Voet, J. Biochemistry, John Wiley & Sons: New York, 1990, chap. 34. Wagner, C.; Riggs, W.; Davis, L.; Moulder, J., Handbook of X-Ray Photoelectron spectroscopy, Perkin-Elmer Corporation: Eden Prairie, MN, 1979. Waller, F. In Polymeric Reagents and Catalysts ; Ford, W., Ed.; American Chemical Society: Washington, D C., 1986. Wang, D. And Heller, A., Miniaturized flexible amperometric lactate probe, Anal. Chem. 1993 , 65, 1069-1073. Wang, J. and Tuzhi, P., Selectivity and sensitivity improvements at perfluorinated ionomer/cellulose acetate bilayer electrodes, Arwl. Chem. 1986 , 58, 3257-3261. Wemet, W. and Wegner, G. Electrochemistry of thin polypyrrole films, Makromol. Chem. 1987 , 188, 1465-1475. Whiteley, L. and Martin, C., Ionomer film-coated electrodes as electrochemical sensors, Mol. Cryst. Liq. Cryst. 1988 , 160, 359-376. Wiedemann, D.; Kawagoe, K.; Kennedy, R.; Ciolkowski, E.; Wightman, R., Strategies for low detection limit measurements with cyclic voltammetry, Anal. Chem. 1991 , 63, 2965-2970. Wiedmeier, V.; Rubio, R.; Berne, R., Incorporation and turnover of adenosine-U14 C in perfused guinea pig myocardium. Am. J. Physiol. 1972, 223, 51-54. Wipf, D.; Kristensen, E.; Deakin, M.; Wightman, R., Fast-scan cyclic voltammetry as a method to measure rapid, heterogeneous electron-transfer kinetics, Anal. Chem. 1988 , 60, 306-310. Witkowski, A. and Brajter-Toth, A., Overoxidized polypyrrole films: a model for the design of permselective electrodes. Anal. Chem. 1992 , 64, 635-641. Witkowski, A.; Freund, M.; Brajter-Toth, A., Effect of electrode substrate on the morphology and selectivity of overoxidized polypyrrole films, Anal. Chem. 1991 , 63, 622-626. Wulff, G., Molecular recognition in polymers prepared by imprinting with templates, Am. Chem. Soc. Symp. Ser. 1986 , 308, 186-230.

PAGE 176

162 Wulff, G.; Best, W.; Akelah, A., Enzyme-analog built polymers, 17. Investigations on the racemic resolution of amino acids, Reactive Polymers 1984 , 2, 167-174. Wulff, G. And Gimpel, J., On polymers with enzyme-analogous structure, 16. On the influence of the binding group flexibility on the ability for racemic resolution, Makromol. Chem. 1982 , 183, 2469-2477. Wulff, G. and Lohmar, E., Enzyme-analog built polymers. XII. Specific binding effects in chiral microcavities of crosslinked polymers, Isr. J. Chem. 1979 , 18, 279-284. Wulff, G.; Schulze, I.; Zabrocki, K.; Vesper, W., Enzyme-analogue built polymers, 11. Binding sites in polymers with differect numbers of binding groups, Makromol. Chem. 1980 , 181, 531-544. Xiuli, G.; Yuqi, L.; Guanghua, Y., Study and application of a new adenosine electrode with thymus tissue, Biosen. Bioelectron. 1992, 7, 21-26. Yamaura, M.; Hugiwara, T.; Iwata, K., Enhancement of electrical conductivity of polypyrrole film by stretching: counterion effect, Synth. Met. 1988, 26, 209-224. Yao, T.; Wasa, T.; Soichiro, M., Linear-sweep voltammetry and the simultaneous determinations of purine bases and their nucleosides in the glassy carbon electrode. Bull. Chem. Soc. Jpn. 1977 , 50, 2917-2920. Zhang, X. And Bard, A., In-trough cyclic voltammetric studies of langmuir-blodgett monolayers of a surfactant derivative of Ru(bpy) 3 2+ at an indium-tin oxide electrode, J. Am. Chem. Soc. 1989 , 111, 8098-8105. Zubay, G., Ed. Biochemistry, Macmillan Publishing Co.: New York, 1988; 2nd edition.

PAGE 177

BIOGRAPHICAL SKETCH Lisa Spurlock was bom in Seoul, Korea, on December 17, 1969. A week later, she was adopted by Ken and Clara Spurlock. She graduated valedictorian from Maynard Evans High School in Orlando, Florida, in 1987 and attended Rollins College in Winter Park, Florida, where she received a B.A. in chemistry in 1991. In August of that year Lisa began her graduate work in analytical chemistry at the University of Florida under the supervision of Anna Brajter-Toth. In the latter part of 1995 Lisa accepted a position with Eli Lilly Animal Sciences Research in Greenfield, Indiana, as Senior Analytical Chemist to begin in March of 1996. While finishing her degree, Lisa married her best friend Jeff Brouwer on January 20, 1996. 163

PAGE 178

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 4 (m-sKs -Totn, Chair ~ — Anna Brajter Associate Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. W2^^r_ ier imes D. Wineforc ^Graduate Research Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Robert T. Kennedy W Assistant Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and isiully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy Randolph S. Duran Associate Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. RoberCJ. Cohen Associate Professor of Biochemistry and Molecular Biology

PAGE 179

This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May, 1996 Dean, Graduate School