New strategies to improve selectivity and sensitivity with ultramicroelectrodes

MISSING IMAGE

Material Information

Title:
New strategies to improve selectivity and sensitivity with ultramicroelectrodes
Physical Description:
xiii, 195 leaves : ill. ; 29 cm.
Language:
English
Creator:
Hsueh, Chen-Chan, 1963-
Publication Date:

Subjects

Subjects / Keywords:
Electrochemical analysis   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 187-194)
Statement of Responsibility:
by Chen-Chan Hsueh.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002051812
oclc - 33436668
notis - AKN9771
System ID:
AA00009028:00001


This item is only available as the following downloads:


Full Text










NEW STRATEGIES TO IMPROVE SELECTIVITY AND SENSITIVITY WITH
ULTRAMICROELECTRODES
















By

CHEN-CHAN HSUEH


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


1995














ACKNOWLEDGMENTS


I would like to express my appreciation to my research advisor, Dr. Anna Brajter-

Toth, for her guidance and assistance, especially for her encouragement and her great help

in writing scientific papers. I wish to thank a former graduate student, Dr. Michael S. Freund,

for his help in designing an on-line iR compensation circuit and writing a semi-integral

analysis program.

My appreciation to my brothers, Chen-Nan and Chen-Sen, and my mother; without

their support and love, I would never have been able to finish this work. Thanks to all the

members of the Toth group (Maurice Thompson, Quan Cheng, Lisa Spurlock, and Merle

Regino) for their assistance and friendship. The assistance of electronics shop's staff

(especially Steven Miles) in designing the fast-scan instrument is also acknowledged.

I would like to thank Jau Yoh in the Pesticide Research Laboratory for her financial

help, and the coworkers in PRL for their friendship. The financial support of this work

through an Electrochemical Society Summer Fellowship is also gratefully acknowledged.

Last but certainly not least, I would like to thank my dear wife, YarJing Yang, for her

support and love.















TABLE OF CONTENTS



ACKNOWLEDGMENTS ............................................. ii

LIST OF FIGURES .................................................... vi

ABSTRACT ............................... ......... ........ ....... xi

CHAPTERS

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

Types ofUltramicroelectrodes ................................... 1
Properties of UMEs and Their Analytical Applications ............. 4
Small Physical Dimensions ............................. 5
In-vivo detection and single cell measurements ....... 5
Microsensors and microdetectors ................... 6
Scanning electrochemical microscopy ............... 6
Low iR Drop and Small Cell Time Constant ................ 7
Electrochemistry in highly resistive media .......... 18
Fast scan voltammetry ....................... 19
Importance of iR compensation for accurate fast
kinetic measurements ..................... 26
High Efficiency of Mass Transport ......................... 31
Circuit Design for UMEs Experiments ......................... 38
Potentiostat for UMEs experiments at low scan rates ........ 38
Potentiostat for UMEs experiments at high scan rates ....... 39
Potentiostat for high-scan-rate measurements at UMEs with
on-line iR compensation ........................ 40
Surface Modification of UMEs ............................... 41
Polymer-film-modified electrodes ....................... 41
Surface-oxide-modified electrodes ...................... 43
Purpose of Work ........................................ 46

2. EXPERIMENTAL SECTION ............................... 48









Reagents and Solutions ..................................... 48
Electrodes .............................................. 48
Reference and Auxiliary Electrodes ..................... 48
W working Electrodes .................................. 49
Fabrication of UM Es ................................. 49
Carbon fiber electrodes sealed in heat shrinkable
Teflon tubes ............................ 49
Carbon fiber electrodes sealed in epoxy ............ 50
Platinum wires sealed in soft glass ................ 53
Electrode Pretreatment ................................ 54
Instrumentation .......................................... 54
IR Compensation Circuit and Its Operational Principle ...... 54
The Current Transducer and the Potentiostat for the UME
Experiments .................................. 62
Instrumental Setup for Low and High Scan Rate
Experim ents .................................. 66
Voltammetry at low scan rates .................... 67
Voltammetry at high scan rates ................... 72
Fundamentals of Electrochemical Methods ...................... 78
Cyclic Voltammetry .................................. 78
Kinetic Measurements with Cyclic Voltammetry ........... 83
Semi-Integration Analysis ............................. 85
Measurement of Diffusion Coefficients with Rotating Disk
Electrodes ............................... .... 91
Chronocoulometry .................................. 92

3. FAST KINETIC MEASUREMENTS IN AQUEOUS SOLUTIONS
WITH ON-LINE IR COMPENSATION ....................... 93

Background .............................................. 93
Signal Averaging and its Effect on Surface Properties ............. 95
Surface Properties of Carbon Fiber and Their Effect on Response of
Ru(NH3)63'/2), Fe(CN)63-/'), and Uric Acid ............... 96
Kinetics of Ru(NH3)6(3+/2) on Freshly Cut Carbon Fiber ........... 104
Kinetics of Fe(CN)6(3-4-) on Freshly Cut Carbon Fiber ............ 108
Redox Reaction of Uric Acid .............................. 109
Conclusions .......................................... 113

4. ELECTROCHEMICAL PREPARATION OF ULTRATHIN
OVEROXIDIZED POLYPYRROLE FILMS AND THEIR
ANALYTICAL APPLICATIONS .......................... 115

Background ........................................ 115








Procedure ofUltrathin OPPy Film Formation by Polymerization and
Overoxidation ..................................... 121
Procedure for Coating Ultrathin OPPy Films on Pt ............... 122
Preparation of Ultrathin OPPy Films .......................... 122
Cation Permselectivity of Ultrathin OPPy at GC Electrodes ........ 126
Permselectivity of Ultrathin OPPy Films at Carbon Fiber
Electrodes ......................................... 130
Applications of OPPy-Coated Ultramicroelectrodes .............. 133
Permselectivity of Ultrathin OPPy Films at Pt
Ultramicroelectrodes ................................ 134
Stability of the OPPy Films ................................. 140
Coating Polymer Films of PPy-2-COOH on Pt Electrodes ......... 140
Conclusions .......................................... 146

5. NEW STRATEGIES FOR IMPROVING SENSITIVITY AND
SELECTIVITY WITH FAST SCAN VOLTAMMETRY ......... 148

Background .......................................... 148
Effects of Surface Oxides on Sensitivity and Selectivity .......... 151
Detection of DA Without the Interference of AA ................ 159
Fast Scan Voltammetry with Signal Averaging .................. 167
Limitations of Fast Scan Voltammetry ........................ 168
Conclusions.............................................. 171

6. CONCLUSIONS AND FUTURE WORK ..................... 173

APPENDICES

A. DATA ACQUISITION, DISPLAY AND PROCESSING
PROGRAM SOURCE CODE ............................. 176

B. SEMI-INTEGRAL ANALYSIS PROGRAM SOURCE CODE ...... 185

REFERENCE LIST ................................................... 187

BIOGRAPHICAL SKETCH .......................................... 195














LIST OF FIGURES


Figure page
1.1. Different types of ultramicroelectrodes. ................... ....... 3

1.2. Illustration of the basic operation principle of scanning electrochemical
microscopy (SECM). As the diffusion of electroactive species is
blocked by the protruding part of the substrate, the tip current
decreases. By plotting the fluctuations of the tip current vs. the
position of the tip, the topography of the substrate is revealed. ....... 9

1.3. Illustration of an electrochemical cell and its equivalent circuit. The
electrochemical cell contains a working electrode and a reference
electrode. Under common experimental conditions, the capacitance
of the reference electrode is much higher than the double layer
capacitance of the working electrode, thus the capacitance of the
reference electrode can be neglected. The electrochemical cell can be
simplified into a combination of cell resistance and the working
electrode double layer capacitance. .......................... 11

1.4. A graphic illustration of the double layer capacitance. As the electrode
surface becomes more positive, the negative ions in the solution begin
to accumulate on the electrode. These negative ions then attract
positive ions. Therefore, a double layer forms. ................. 13

1.5. The origin of the faradaic current. An electroactive molecule (in this
case, the oxidant) diffuses to the electrode surface and exchanges
electrons with the electrode. The oxidant is reduced to the reductant
and diffuses away. The electron exchange between the electroactive
molecule and the electrode produces the faradic current. The current
is a function of redox kinetics (rate of electron transfer between the
electroactive species and the electrode) and the diffusion process (rate
of diffusion of the electroactive species). ........................ 15

1.6. Illustrations of fast cyclic voltammetry and background subtraction: (a)
applied potential waveform as a function of time; (b) current as a
function of time; the solid line is the background current and the








dashed line is the current in the presence of electroactive
species; (c) A background subtracted voltammogram of (b);
dashed line in (b) subtracted from solid line in (b) and the
result plotted as a function of potential. ........................ 25

1.7. Simulation of the effects of the cell resistance and double layer
capacitance on the effective scan rates (the shape of the applied
potential); measured (effective) potential (solid line), applied
potential (dashed line). Top: the effects of cell resistance only.
Bottom: the effects of resistance and capacitance. ................ 30

1.8. Growth of a diffusion layer as a function of time (t, < t2 < t < t4 where the lines represent a cross section of the interface between the
diffusion layer and the solution bulk. At time t,, diffusion is linear.
As time scale increases, diffusion becomes a mix of linear and radial
diffusion. Eventually, diffusion behavior transforms to a radial
diffusion. ............................................... 34

1.9. The effects of electrode size on diffusion behavior at common
experimental time scales (several seconds or less) ................ 37

1.10. A possible structure of surface oxide groups on a carbon electrode
(reference 78). ......................................... 45

2.1. Schematics ofUMEs used in this work. (A). Carbon fiber (71pm) sealed
with Teflon. (B). Carbon fiber sealed with epoxy (C). Pt wire sealed
(5pm) in glass. ......................................... 52

2.2. Schematic of a potentiostat built for fast measurements with iR
compensation. ....................... ............... ... 56

2.3. Schematic of the current transducer. ............................ 64

2.4. Instrumental setup for UME experiments at low scan rates ........... 69

2.5. Instrumental setup for UME experiments at high scan rates. ........... 71

2.6. Cyclic voltammogram of 0.1 mM Fe(CN), in 70 mM pH 7.0
phosphate buffer obtained from the current transducer connected to
BAS. The scan rate is 10 mV/s. The gain of the current transducer
is 10000 and the time constant of the first order filter is 100 ps. The
working electrode is a carbon electrode (7 rm). .................. 74








2.7. Voltammogram of 0.18 mM dopamine at 2,000 V/s in 70 mM pH 7.0
phosphate buffer with a carbon electrode (7 pm). The voltammogram
is background subtracted and signal averaged 1000 times .......... 77

2.8. A typical cyclic voltammogram. .............................. 81

2.9. A plot ofAEp vs kinetic parameter i (reference 88 and 89)............ 87

2.10. Division of experimental i(t) vs. t [or vs. E(t)] curve for semi-
integration. ................... ........................... 90

3.1. Semiintegrated cyclic voltammetric current of 10 mM Ru(NH3)6( '*3) in
1 M KCl solution at a freshly cut carbon fiber electrode (7 gm). Scan
rate is 5000 V/s .......................................... 98

3.2. Cyclic voltammograms of 1 mM uric acid in 1 M pH 7.0 phosphate
buffer at a carbon fiber electrode (7 gim). Scan rate 400 V/s. Solid line
freshly cut; dashed line electrochemically pretreated. (For
pretreatment procedure see the Experimental Section)............. 101

3.3. Semiintegrated cyclic voltammetric currents of uric acid in Figure 3.2.
Solid line freshly cut; dashed line electrochemically pretreated. .. 103

3.4. Cyclic voltammograms of 10 mM Ru(NH3)623) in 1 M KCI solution
at a freshly cut carbon fiber electrode (7 gm). Scan rate is 11,000 V/s.
The solid line is the current response with iR compensation
(Compensated resistance Ri=6.9 KO). The dashed line is the current
response without iR compensation .............. .......... 107

3.5. Redox and follow-up reactions of uric acid in aqueous solutions. ..... 111

4.1. Chemical structures of Py PPy, and OPPy ....................... 117

4.2. A cartoon representation of the proposed carbonyl group in OPPy films
hindering anion diffusion. .................................. 119

4.3. A cartoon representation of pin holes gradually filled with OPPy by a
repeated coating procedure ..................................... 125

4.4. Cyclic voltammograms: (a) 10 mM Fe(CN)63 in 0.5 M phosphate buffer
(pH 7.0) at OPPy modified glassy carbon electrode. Curve 1 is the
response of the bare electrode. Curves 2-7 are the responses after first
time (curve 2) to sixth time (curve 7) of repeated coating; (b)10 mM








Ru(NH3)613 at the same electrode and buffer as in (a). Solid
line is the response before coating. Dashed line is the
response after coating six times. Scan rate is 20 mV/s.
Electrode area is 0.067 cm ................................. 129

4.5. Cyclic voltammograms: (a) 10 mM Fe(CN)63 in 0.5 M phosphate buffer
(pH 7.0) at OPPy modified carbon fiber ultramicroelectrode. Curve
1 is the response of the bare electrode. Curves 2-5 are the responses
after first (curve 2) to four times (curve 5) of repeated coating; (b) 10
mM Ru(NH3)63 at the same electrode and buffer as in (a). Solid line
is the response before coating. Dashed line is the response after
coating four times. Scan rate is 20 mV/s. Electrode diameter is 7
um.r ...... ......... .......................... .... 132

4.6. Cyclic voltammograms of 10 mM dopamine (solid line) and ascorbic
acid (dashed line) in 0.5 M phosphate buffer (pH 7.0) at carbon fiber
electrode modified four times with the same film as in Figure 4.5.
Scan rate is 20 mV/s. Electrode diameter is 7 im. ................ 136

4.7. Cyclic voltammograms: (a) 10 mM Fe(CN)63- obtained at Pt
ultramicroelectrode in 0.5 M phosphate buffer (pH 7.0). Solid line is
the response at the bare electrode. Dashed line is the response at an
electrode modified with adsorbed/polymerized monolayer of OPPy
film; (b) 10 mM Ru(NH3)63 at the same electrode and buffer as in (a).
Solid line is the response at the bare electrode. Dashed line
corresponds to the conditions for dashed line in (a). Scan rate is 20
mV/s. Electrode diameter is 5 pm. .................. ..... 139

4.8. (a) Cyclic voltammograms of three consecutive scans at Pt
ultramicroelectrode in 50 mM Py-2-COOH and 0.1 M TBAP in
MeCN. Scan rate is 50 mV/s. The reference electrode is a Ag wire.
Electrode diameter is 5 pim. The peak potential is ca. +1.5 V vs. Ag
wire; (b) Cyclic voltammograms of three consecutive scans at a
glassy carbon electrode in 50 mM Py-2-COOH and 0.1 M TBAP in
MeCN. Scan rate is 100 mV/s. Electrode area is 0.067 cm2. The peak
potential is ca. +1.65 V vs. Ag wire. ......................... 143

4.9. Cyclic voltammograms: (a) 10 mM Ru(NH3)63 and (b) 10 mM
Fe(CN)63- in 0.5 M phosphate buffer (pH 7.0) at a Pt
ultramicroelectrode before and after coating with PPy-2-COOH in
one CV scan. Solid lines are the voltammograms at the bare
electrodes. Dashed lines are the voltammograms at the modified
electrodes. Scan rate is 20 mV/s. Ultramicroelectrode diameter is 5










5.1. Background currents after repetition of fast cycle scans. Cyclic
voltammogram obtained: (line 1) with a freshly polished carbon fiber
electrode (7 pm diameter); (line 2) after 30 minutes of repeated
cycling, at a scan rate of 100 V/s, in the potential window of -0.8 to
+1.2 V vs SCE; ( line 3) after one hour of cycling; (line 4) after one
and a half hour of cycling. ................... .............. 154

5.2. Plots of log of 0.1 mM DA peak current, log i, vs log scan rate. Scan
range is from 50 to 10,000 V/s. Working electrode is a carbon fiber
electrode (7 ptm diameter) and buffer solution is 70 mM pH 7.4
phosphate buffer. ......................................... 158

5.3a. Cyclic voltammograms of 104 M DA (solid line) and 10.2 M AA
(dashed line) at a scan rate of 100 V/s at a 7 pm carbon fiber
electrode. The buffer used in the experiment is a 70 mM phosphate
buffer. Electrode cycled for ca. 30 minutes in the buffer solution
before use (see Experimental). Time constant of the potentiostat filter
is 1 s. Gain of the current transducer of the potentiostat is
IV/pA. ............................................ 161

5.3b. Cyclic voltammograms of 104 M DA (solid line) and 10-2 M AA
(dashed line) at a scan rate of 10,000 V/s. Same electrode and
solution conditions as in Figure 5.3A .......................... 163

5.4. Cyclic voltammogram of 5 pM DA in the presence of 1.2mM AA in 70
mM pH 7.4 phosphate buffer. Scan rate is 2,000 V/s, electrode
diameter 7 pm The voltammogram was averaged 1000 times. .... 166

5.5. Signal averaged cyclic voltammograms of 50 pM DA in 70 mM
phosphate buffer, scan rate 2,000 V/s. Only the oxidation peaks are
shown. Lines 1, 2, 3, and 4 averaged 1, 10, 100, and 1000 times,
respectively. ............................................ 170














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

NEW STRATEGIES TO IMPROVE SELECTIVITY AND SENSITIVITY WITH
ULTRAMICROELECTRODES

By

Chen-Chan Hsueh

May 1995

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

The objective of this work is to develop new analytical strategies for analytical

measurements with ultramicroelectrodes (UMEs), especially in biosensing. We focus most

of our efforts on the development of fast scan technique using UMEs in characterizing fast

redox reactions of biological species, and improving signal-to-noise ratio (S/N), selectivity,

and detection limit. Instrumention which has been built to perform these experiments with

UMEs is described.

There are three major parts in the dissertation. The first part describes fast scan

voltammetry in the investigation of rapid redox reactions in aqueous solutions with UMEs.

Fast scan voltammetry has been used in organic solvents and most of the electrodes used

were metal electrodes. This part of the work tries to establish the use of fast scan

voltammetry in the investigation of fast redox reactions in aqueous solutions with carbon








electrodes. An on-line iR-drop-compensation circuit was designed for this work. This circuit

can perform fast scan voltammetry up to 200,000 V/s without iR-drop distortion. The surface

characteristics of carbon fiber ultramicroelectrodes have been investigated under fast scan

voltammetry conditions. The standard reaction rate constants of a biological compound (uric

acid) and two well-know inorganic probes (Ru(NH-3)6d3) and Fe(CN)3'"*) were measured.

The influence of surface oxides on the measured rate constants is discussed.

The second part focuses on the improvement of selectivity of UMEs with polymer

film coatings. The complexity of a biological matrix makes selectivity an important

consideration in biosensing and bioanalysis. The work aimed to improve the selectivity and

stability in electrochemical measurements by modifying electrode surface properties with

polymer films. A method of making ultrathin overoxidized polypyrrole films was designed

to modify electrode surfaces without sacrificing sensitivity.

The third part of this research focuses on the potential analytical advantages of fast

scan voltammetry. A specific consideration was the method design for improving selectivity

and sensitivity of analytical measurements in complex environments. To reach this goal, a

circuit with low instrumental noise was designed to perform fast scan voltammetry. The new

method was tested in dopamine detection at low concentrations in the presence of a large

excess of ascorbic acid. By pushing the scan rate up to 10,000 V/s, it is demonstrated that

a substantial improvement in selectivity can be achieved even though the reaction potentials

of dopamine and ascorbic acid are similar. At high scan rates, it is possible to detect

dopamine in the presence of 1000-fold excess of ascorbic acid because of the kinetic

differences in the electrochemical responses of the two probes. Other advantages of ultrafast

xii








scan rate measurements in aqueous solutions are also investigated in this work. These

include improved temporal resolution and higher signal-to-noise ratios resulting from the fact

that more voltammetric scans can be acquired and averaged in a short period of time.














CHAPTER 1
INTRODUCTION


Types of Ultramicroelectrodes


In the 1970s a number of research groups exploited the advantages of

ultramicroelectrodes (UMEs), normally defined as electrodes with characteristic dimensions

smaller than 20 Ilm.12'3 A variety of materials, shapes, and sizes of UMEs have been

reported since. Carbon, gold, and platinum are the most commonly used materials, with a

few publications describing copper and mercury. Carbon fiber is the most popular material

used in bioanalysis due to its rich chemical properties and compatibility with bio-tissue. The

geometries of several different types of UMEs are illustrated in Figure 1.1. The cylindrical

UMEs are usually prepared by allowing a portion of the wire or fiber to protrude from the

insulator. The cylindrical geometry has the advantage of larger currents, but these electrodes

tend to be mechanically fragile and cannot be easily polished.

At present, the most widely used UMEs are probably ultramicrodisk electrodes with

a radius of less than 10 pm. They are constructed relatively easily by encasing a metal wire

or a carbon fiber in glass or epoxy; the flat surface of the end of the insulated wire serves as

the active electrode surface. Because disk-like UMEs have smaller electrode areas and good

spatial resolution, they are commonly used in fast measurements and in in-vivo detection.

Recently, the use of band and ring electrodes is being reported more often.4.',6 Electrodes



























Figure 1.1. Different types of ultramicroelectrodes.













Typical types of ultramicroelectrodes


drop


band


cylinder


interdigitated array


disk


ring


array


'41 P, 00 1 a
q mh0
0 04








4

with a band or ring geometry are attractive because they can be fabricated on a nanometer

scale in one dimension. One approach to their fabrication is to sandwich thin metal film

between glass or epoxy insulators. Bands with a smallest dimension of 20 A have been

prepared in this way.7 The surface areas of band and ring UMEs can be enlarged by

changing the length of the band or the circumference of the ring without losing the desired

properties of UMEs.

It should be noticed that the area of a UME can be as large as that of a conventional

electrode. As long as one of the dimensions of the electrode is less than 20 pm, it can be

called an ultramicroelectrode. An enlarged surface area ensures large faradaic currents which

can be detected without difficulty by conventional potentiostats. The same objective can be

achieved with ultramicroelectrode arrays. One of the most interesting designs of UME

arrays is the interdigitated UME.8 The individual band UMEs are arranged close to each

other, yet different potentials can be applied to these individual UMEs. This can make redox

species diffuse back and forth between two adjacent electrodes, thus amplifying the signal.

A common approach to the construction of an ultramicroelectrode array is to use

lithographic techniques to prepare arrays of band electrodes on insulating substrates.9-10



Properties of UMEs and Their Analytical Applications



Because many obstacles in electrochemistry can be reduced or eliminated with

UMEs, their use has grown rapidly in the past ten years.1112,13,14 Compared to the

conventional macroelectrodes, UMEs have many unique properties including small physical








5

dimensions, low iR drop, and high mass transport. These unique properties of UMEs have

opened totally new possibilities for electrochemists. In the following discussion, these

unique properties and their analytical applications will be introduced.



Small Physical Dimensions



One of the most obvious advantages of UMEs is related to their small physical

dimensions. Since the electrode is typically very small, a microliter volume of solution can

be used in the experiments with UMEs. Because the electrochemical measurements are

concentration sensitive rather than mass sensitive, the mass detection limit can be extremely

low ifa small volume of a sample is used. A 10-21 mole detection limit of insulin has been

reported in the literature."1



In-vivo detection and single cell measurements

The most common applications based on the small dimensions of UMEs are in-vivo

detection and single cell measurements. 16.17.18.1920.21 The small dimensions of the electrode

allow in-vivo detection with minimal tissue damage and with a high spatial resolution. The

most intensively investigated area of in-vivo measurements is in neuroscience where UMEs

allow pinpointing the area of interest in the brain without damaging the brain tissue. Recent

development of even smaller electrodes allows electrochemical detection inside a single

cell.22










Microsensors and microdetectors

In the separation science, the miniaturization of instrumentation is the current trend.

Miniaturization reduces the amount of required sample, the waste produced, and in many

cases the time required to complete the analysis. One of the biggest challenges in

miniaturization is to maintain the performance of the detector as the size of the detector

decreases. Most detectors require a large detection volume in order to perform well, which

makes miniaturization difficult. However, the small dimensions of UMEs and the inherent

concentration sensitive properties of the electrochemical detection make UMEs very

compatible with detector miniaturization. Moreover, the small diffusion layer at UMEs

makes them less dependent upon the flow rate. Many reports have shown the power of

UMEs combined with capillary zone electrophoresis (CZE) and micro-column high

performance liquid chromatography (HPLC).232425



Scanning electrochemical microscopy

Another interesting application based on the small dimensions of UMEs is scanning

electrochemical microscopy (SECM). This technique resembles scanning tunneling

microscopy (STM) as far as the movement of the electrode across the substrate surface is

concerned, but the principles of the measurement are different. In SECM, an UME with a

tip radius on the order of 10 pm or less is moved in close proximity to a substrate of interest,

in contact with a solution containing an electroactive species. The electrochemical reactions

at the tip give rise to a tip current that is affected by substrate topography. Generally, the tip

current is controlled by diffusion of electroactive species at the electrode tip. As the tip








7

moves across the substrate, the mountain part of the substrate blocks the diffusion of

electroactive species (lower faradaic current), and the valley part of the substrate allows

efficient diffusion (higher faradaic current). Consequently, the fluctuation of the tip current

reflects the topography of the substrate as the tip moves across the substrate (Figure 1.2).



Low iR Drop and Small Cell Time Constant



Two main obstacles of conventional large electrodes are large iR drop and large time

constant. An electrochemical cell can be considered as a circuit consisting of a resistor (cell

resistance) and a capacitor (double-layer capacitance). Figure 1.3 illustrates an

electrochemical cell and its equivalent circuit. The double-layer capacitance originates at a

charged electrode surface in contact with an electrolyte solution which attracts ions of

opposite charge and repels ions of like charge. A graphic illustration of the double-layer

capacitance is shown on Figure 1.4. When a change of voltage occurs at a working

electrode, the charge of the electrode surface changes accordingly, thus the charged ions need

to move into or move out from the double-layer. The movement of the charged ions

produces a current which is called the double-layer charging current. Another source of

current is from the redox reaction of electroactive species in the cell. When the applied

voltage is sufficient to provide energy for the redox reaction, faradaic current flows (see

Figure 1.5).

As the current is passing through the cell, the applied voltage is consumed by the cell

resistance, contributing to the iR loss or the iR drop, as the following equation shows,



























Figure 1.2. Illustration of the basic operation principle of scanning electrochemical
microscopy (SECM). As the diffusion of electroactive species is blocked by the protruding
part of the substrate, the tip current decreases. By plotting the fluctuations of the tip current
vs. the position of the tip, the topography of the substrate is revealed.















tip current


diffusion blocked


diffusion unblocked


electrode


subtrate


ILsl ~---------



























Figure 1.3. Illustration of an electrochemical cell and its equivalent circuit. The
electrochemical cell contains a working electrode and a reference electrode. Under common
experimental conditions, the capacitance of the reference electrode is much higher than the
double layer capacitance of the working electrode, thus the capacitance of the reference
electrode can be neglected. The electrochemical cell can be simplified into a combination
of cell resistance and the working electrode double layer capacitance.











(A)


working electrode


Solution


reference electrode


C ref


cell
a
-------- ------- MM/ -------


Cd

a


Rcell
VVAAAAAAA
vWVVv


R e



























Figure 1.4. A graphic illustration of the double layer. As the electrode surface becomes
more positive, the negative ions in the solution begin to accumulate on the electrode. These
negative ions then attract positive ions. Therefore, a double layer capacitance forms. For
the sake of simplicity, the solvent molecules are omitted in the figure.













ions


.--_ /


+

+

+ I


6>


+

+


electrode surface


+


-I-


0


double layer
structure



























Figure 1.5. The origin of the faradaic current. An electroactive molecule (in this case, the
oxidant) diffuses to the electrode surface and exchanges electrons with the electrode. The
oxidant is reduced to the reductant and diffuses away. The electron exchange between the
electroactive molecule and the electrode produces the faradic current. The current is a
function of redox kinetics (rate of electron transfer between the electroactive species and the
electrode) and the diffusion process (rate of diffusion of the electroactive species ).








faradaic current







red








e ox


electrode surface











E =E .iR (1.1)



where E, (V) is the potential applied to the cell, E, (V) is the true potential of the working

electrode, i (pA) is the current flowing through the working electrode, and R (0) is the cell

resistance.

The cell resistance R can be expressed as follows,


R = P 1 (1.2)
4r r


where r is the radius of the disk electrode (cm) and p is the specific resistivity of the solution

(acm). According to equation 1.2, the cell resistance increases as the size of the electrode

decreases. However, the charging current from the double-layer capacitance, and the

faradaic current from the redox reactions of the electroactive species in the electrochemical

cell, are proportional to the electrode area because both of them originate at the electrode

surface. The total cell current can be described as follows,

i = (i.i9) A r2 (1.3)





where i (pA) is the cell current, i, (pA) is the charging current, if (pA) is the faradaic current,

A (cm2) is the electrode area, and r (cm) is the radius of the disk electrode. According to

equation 1.3, the magnitude of the current decreases with the square of the electrode radius,

which is a significant decrease. Combining equations 1.2 and 1.3, a relationship between the

iR drop and the electrode size is follows.










iR r (1.4)



Equation 1.4 clearly shows that the iR drop will be reduced as the size of the electrode

decreases. This ensures the superiority of UMEs over conventional macroelectrodes in terms

of reducing the iR drop.

Another problem associated with the conventional macroelectrodes is the large cell-

time constant. When a sudden change of voltage is applied to the working electrode, time

is required to charge the capacitor (double-layer capacitance) to reflect the change in the

voltage. This time is called the cell-time constant and can be expressed as follows,

S- C, R (1.5)



where r (sec) is the cell-time constant, Cd is the double-layer capacitance (F), and R is the

cell resistance (0). The doubl- layer capacitance is a function of the electrode area,

Cd. CA = Cr' r2 r' (1.6)



where C is the double-layer capacitance per unit area (F/cm2). Substitution of equations 1.2

and 1.6 into equation 1.5 gives the following,

S-= C R r (1.7)



According to equation 1.7, the cell-time constant will be very small when a UME is used

because the time constant is proportional to the electrode radius. Experimentally, with

UMEs of 5 unm radius cell-time constants are less than a microsecond.26










Electrochemistry in highly resistive media

One important application exploiting the low iR drop at UMEs is in the

electrochemical measurements in highly resistive media which are typically inaccessible with

conventional size electrodes due to the enormous iR drop. The reduced iR drop allows

measurements to be made in novel systems such as nonpolar solvents with low

concentrations of supporting electrolytes,27 polar solvents in the absence of supporting

electrolytes,28 and gas phase.29-30,31 The ability to perform electrochemistry in these novel

media makes UMEs a very valuable tool.

Nonpolar solvents were incompatible with electrochemical measurements due to their

extremely high resistance. Lines and Parker were the first to show that voltammetry at

UMEs is possible in benzene (a nonpolar solvent).32 Other groups, such as Bond's,33"-'

Fleischmann's,36 and Wightman's37,8'39 had reported voltammetric measurements with UMEs

in benzene, toluene, and even hexane.

In conventional electrochemistry, a supporting electrolyte is required to ensure that

the medium conducts properly. Supporting electrolytes often limit the maximum useful

potential range because the electrolyte itself may become electroactive. The ability to

perform the measurements in polar solvents in the absence of supporting electrolyte enables

the expansion of the useful potential range. Pons and Fleischmann have reported that

alkanes can be oxidized in electrolyte-free acetonitrile at potentials above + 3.5 V.4 Bard's

group has reported that alkali metal ions can be successfully oxidized in liquid SO2 at

potential of+5.0 V.4'

Applications of UMEs to electrochemistry in the gas phase are another interesting








19

field. Pons and Fleischmann have published several reports of such experiments.29'30-'I The

UMEs coated with an ionic conducting membrane were used to determine the quantity of

organic gases. The cell conductivity was maintained by proton transfer in the membrane

while the gases could diffuse into the membrane to react at the electrodes or affect the

conductivity of the cell.30



Fast scan voltammetry

At the conventional macroelectrodes, it is impossible to perform fast scan

voltammetry (voltammetry at scan rates higher than 100 V/s) because the cell time constant

and the iR drop are too large. In voltammetric experiments, the dependence of the charging

current (ic) on the cell time constant (r) and the scan rate (v) can be expressed by the

following equation,


i, [-exp(--).l]vCd (1.8)


where t is time (s) and v is scan rate (V/s). The first term indicates the time (t) needed to

charge the double-layer as the electrode potential changes. After the double-layer is fully

charged, the charging current will remain as indicated by the second term (Cdv). This

charging current is independent of time. According to equation 1.8, the first term approaches

zero and the second term dominates if t/T is much larger than one. In other words, if the

experimental time scale is much larger than the cell-time constant, the effect of time constant

on the charging current can be neglected. In fast scan voltammetry, the time scale (t) is

extremely small. For fast scan voltammetry to be feasible, a small cell-time constant (r) is

needed to ensure that t/r is large enough.








20

The second requirement for fast scan voltammetry is that the iR drop must be small.

According to equation 1.8, the first term will be close to zero when the cell time constant

(defined by equation 1.7) is small. Equation 1.8 can be simplified to the following.




i = vCd (1.9)





therefore, we have the following equation.

iCR vCd O v (1.10)



Equation 1.10 shows that the iR drop is proportional to scan rate. To minimize the iR drop

at high scan rates, CdR needs to be as small as possible. In summary, a small cell-time

constant and a small iR drop are the two essential requirements for fast scan voltammetry.

As equations 1.4 and 1.5 demonstrate, the cell-time constant and the iR drop are

greatly reduced at a UME. Thus, the scan rate can be increased greatly. Experimentally,

scan rates of 1,000,000 V/s have been achieved with a 51pm gold electrode in acetonitrile as

a solvent (cell resistance = 16 kQ and the double-layer capacitance = 5.5 pF).42,43,44 This is

an improvement of five orders of magnitude compared to scan rate of 10 V/s at

macroelectrodes.

Fast scan voltammetry has been applied to measurements of fast redox reactions and

fast follow-up chemical reactions (coupled chemical reactions). Since the introduction of

UMEs fast heterogeneous electron-transfer kinetics have been studied with fast scan








21

voltammetry. For example, redox reactions of anthracene, anthraquinone, naphthoquinone,

and benzoquinone in organic solvents have been characterized.45'4647 The scan rates used in

the experiments ranged from 1000 V/s to 1,000,000 V/s and the standard rate constants

obtained were as large as 3.8 cm/s.44'47

Since 1980, many voltammetric investigations have been carried out on electron

transfer reactions coupled with follow-up chemical reactions such as addition, isomerization,

dimerization, and homolytic cleavage.48,49,50 All of these reactions involved fast-decaying

redox intermediates. Fast scan voltammetry can be used to measure the life time of redox

intermediates. Fast scan voltammetry also prevents the intermediate from decaying and

producing by-products which may react with the original reactants and complicate the

reaction mechanism. In some applications, the follow-up products of the intermediates are

electroactive and have close redox potentials to the parent molecules. By using fast scan

voltammetry, the redox potentials and the redox kinetics of such follow-up products can be

revealed.51'52

In addition to the application to the study of fast redox kinetics, fast scan

voltammetry can be a powerful analytical tool. Recently, research has shown that fast scan

voltammetry can be advantageous in trace analysis and in bioanalysis.354 For example, fast

scan voltammetry has been applied in quantitative measurements of biological species, such

as NADH and dopamine.54'"

Inherently, fast scan voltammetry is not suitable for quantitative analysis due to the

interference of large background currents. At solid electrodes, such as graphite, the

background current consists of two components, namely pure capacitive charging current,








22

and the current due to the surface redox processes. Charging current is proportional to scan

rate (equation 1.9), and the current due to the surface redox processes is also proportional to

scan rate. For surface bound species, peak current for an irreversible redox reaction is



SnanF'AvI (1.11)
$ 2.718RT


where C' is the reactant concentration in solution bulk in mole/cm3, n is the number of

electrons involved in the reaction, a is the electron transfer coefficient, n. is number of

electrons involved in the rate determining step, and r,* is the surface excess (mole/cm2).

Equation 1.11 shows that the faradaic background current is proportional to the scan

rate. However, faradic peak current for a diffusion-controlled irreversible process is

proportional to a square root of the scan rate.


i (2.99x 10)n(ancz)AC D 'nv' (1.12)


where D is the diffusion coefficient of the reactant in cm2/s. Thus the ratio of the faradic

diffusional current to background current can be expressed as follows,



i, Vr 1 (1.13)
i' v VV


where if and ib are the faradic diffusional current and the background current, respectively.

Equation 1.13 indicates that as the scan rate increases, the current if(signal) decreases

relative to the background current, ib. Consequently, at high scan rates, this current is buried








23

in the enormous background current. Thus, faradaic currents can not be measured directly

from such voltammograms. Furthermore, the instrumental noise becomes noticeable as the

background current increases. However, with the aid of background subtraction and noise-

filtering techniques, fast scan voltammetry has been demonstrated to be useful for

quantitative analysis.52,' In fact, submicromolar detection limits have been reported.52'56

Figure 1.6 shows the effect of background subtraction on extraction of the faradaic current

out of the background current.

One of the great advantages of fast scan voltammetry is its high temporal resolution,

which allows observation of rapid concentration changes in the micromolar range. This

advantage is particularly useful in monitoring rapid changes of concentration of biological

species in vivo which may require subsecond or better temporal resolution.15 Another

significant advantage of fast scan voltammetry is its ability to differentiate different species

by their kinetic differences as well as their redox potentials.52"56 Species with fast electron

transfer rates show reversible voltammetry at high scan rates, while species with slow

electron transfer rates react at larger overpotentials at high scan rates and consequently, can

be easily distinguished from the more reversible species. This strategy of improving

selectivity has been very effective in reducing interference in the measurement of dopamine

in vivo.56

The fact that the electrode stability can be improved with fast scan methods is of

particular importance in bioanalysis. Many redox reactions of biological species (such as

dopamine, uric acid, and NADH) have follow-up chemical reactions which produce side

products. These products can adsorb and foul the electrode surface. In most experiments


























Figure 1.6. Illustrations of fast cyclic voltammetry and background subtraction: (a) applied
potential waveform as a function of time; (b) current as a function of time; the solid line is
the background current and the dashed line is the current in the presence of electroactive
species; (c) A background subtracted voltammogram of (b); dashed line in (b) subtracted
from solid line in (b) and the result plotted as a function of potential.'1













Ox
(b) C

200 nA /O

(c) O



1 \\20 nA\
Red
I I

10 ms

(a) ..


1 -
1V


0.5 V


*








26

electrode activity decreases after several scans. In fast scan voltammetry, because voltages

are swept back and forth in a short time, the reductive or oxidative products are converted

back to the original analyte before they have an opportunity to undergo follow-up reactions.

Thus, surface fouling due to adsorption of products can be prevented. Consequently, the

stability of electrodes can be improved.

Compared with chronoamperometry, fast scan voltammetry can provide more

information about the analyte, which can help identify analytes of interest. Because different

redox reactions have unique redox potentials, the information can add more confidence that

the analyte of interest is being detected in a complicated environment. This advantage may

be undermined during in vivo measurements if the diffusion layer of the analyte is blocked

by tissue or clogged by blood. However, fast scan rates can alleviate this problem because

the diffusion layer is smaller at higher scan rates.



Importance of iR compensation for accurate fast kinetic measurements

Even though iR loss and cell time constant are small at a UME, they are not

negligible at scan rates above 1000 V/s if an accurate kinetic measurement is desired,

especially at electrodes such as carbon. Carbon electrodes have a large and irregular

background current in aqueous solutions which comes from large double-layer capacitance

and from redox reactions of surface groups. Since a variety of surface-bound functional

groups with different formal potential (E) values may be present, the faradaic background

current may be potential-dependent. This background current may affect the accuracy of

kinetic measurements because background current is proportional to scan rate.








27

The large and irregular background current on carbon electrodes limits the upper limit

of useful scan rates to around several hundred V/s.52'56 Due to the large iR loss and the

irregular shape of the background current at high scan rates (above 1000 V/s), the elimination

of the iR drop and background subtraction become necessary if significant kinetic

information is to be extracted from the voltammogram.

The relationship between the ohmic drop (iR) and the effective scan rate (vel) on a

working electrode can be expressed by the following equations,

dE d(E .iR) dEe di di
v d ) R(&)=vR() (1.14)
0 dt dt dt dt dt


where E, (V) is the potential applied to the cell, E, (V) is the true potential of the working

electrode, i (pA) is the current flowing through the working electrode, R (Q) is the cell

resistance, ve (V/s) is the effective scan rate at the working electrode, and v (V/s) is the

scan rate applied to the cell which is constant during the experiment. The effective scan rate

at the working electrode is not a constant when R is not zero and a faradaic current (

background faradaic current due to the surface reactions or faradaic current of analytical

species in solution) is present (equation 1.12), and the voltammograms cannot be treated

according to theories developed for cyclic voltammetry which assume constant scan rate at

the working electrode during the experiment.

Figure 1.7 demonstrates the effects of cell resistance and the double-layer capacitance

on the applied potential. Figure 1.7 (top) is a simulation of the effect of the cell resistance

on the potential at the working electrode. The deviations of the effective potential (potential

on the working electrode) from the applied potential (triangular waveform) are caused by the








28

faradic current (from the electroactive species in solution) and the cell resistance (i.e., iR

drop). Furthermore, when the double-layer capacitance is considered ( Figure 1.7 bottom),

the combined effects of the double-layer charging and the cell resistance severely disfigure

the shape of potential sweep.5s

Two approaches are commonly used to compensate the iR drop. The first approach

is to simulate and fit the distorted experimental voltammograms by incorporating the ohmic

and the capacitive factors.5""596 However, it is difficult and complicated to compensate the

iR drop with the simulation method because the effective scan rate (vf) and the total current

( i ) are mutually dependent on each other. To make the situation worse, the charging current

is affected by the presence of the faradaic current, since,

di di
i=Cd C=(v.R(d)] = Cv.C )( d) (1.15)


where ic is the charging current, Cd is the capacitance of the electrode, i is the total current

(faradaic and charging) passed at the working electrode. According to equation 1.14, v,f is

a function of the faradaic current and the charging current.

On the other hand, because the charging current affects the effective scan rate

(equation 1.14), which affects the faradaic current, the faradaic current in the presence of the

charging current is not the same as the pure faradaic current. These factors coupled together

complicate the simulation approach.

Another approach to solve the problem of the iR drop is to use a potentiostat with an

on-line iR compensation circuit.61,62,63 This approach is simple and straightforward. A

potentiostat with an on-line iR compensation circuit can be used to compensate cell































Figure 1.7. Simulation of the effects of cell resistance and double layer capacitance on the
effective scan rates (the shape of the applied potential); measured (effective) potential (solid
line), applied potential (dashed line)." Top: the effect of the cell resistance only. Bottom:
the effects of the resistance and the capacitance.
















-0.1-


E/0.0-


0.1-


0.2-


0.3 r
0.3- I I I I I
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
tls


-0.3- /
S
:
*
-0.2-


-0.1-*

V0.0
EN 1


0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
SUs








31

resistance (R) to negligible values. This makes the second term (R(di/dt)) in equation 1.14

close to zero, thus ensuring that the scan rate at the working electrode is constant and the iR

loss is eliminated. An iR corrected voltammogram can be acquired on-line and the kinetic

information can be obtained directly, which is a simple and easy approach compared to the

off-line mathematical simulation.



High Efficiency of Mass Transport



As the redox reaction occurs, the reactant is being depleted at the electrode surface.

The depletion of the reactant at the surface causes the reactants in solution bulk to diffuse to

the electrode due to the concentration gradient between the solution and the surface. If the

rate of the redox reaction at the electrode surface is faster than the rate of mass transport

(diffusion), a diffusion layer will begin to grow with time. The thickness of the diffusion

layer, 6 (cm), can be described by the following equation,

6 vTn (1.16)



where D is the diffusion coefficient of the diffusing species (cm2/s) and t is the experimental

time (s). Figure 1.8 illustrates the growth of the diffusion layer as a function of time at a disk

electrode. As the Figure 1.8 shows, the diffusion layer thickness at a short time (t1) is very

small compared to the electrode radius. When the reaction time becomes longer (t3-t5), the

diffusion layer becomes thicker. At the UMEs, the diffusion profile is transformed from

linear diffusion to radial diffusion, and the mass transport is greatly enhanced due to the








32

diffusion from the edges. This enhancement of mass transport is called an edge effect or an

effect of radial diffusion. Under radial diffusion conditions, the mass transport rate is given

by the mass transport coefficient m (cm/s). The dependence of m on the electrode radius is

described as follows,"




m = D (1.17)
r


where m is the mass transport coefficient (cm/s), D is the diffusion coefficient of the

diffusing species (cm2/s), and r is the radius of the electrode. Consequently, for a very small

electrode the mass transport coefficient can be very high.

As Figure 1.8 illustrates, the diffusion type (linear or radial) is determined by the ratio

of the diffusion layer thickness (6) to the radius of the electrode (r). If 6/r is much less than

one, diffusion is linear. If 6/r is significantly greater than one, diffusion becomes radial.

Because 6 is a function of time (as equation 1.12 describes), radial diffusion can be achieved

at an electrode of any size (r) as long as the experimental time is sufficiently long. However,

as the radius of the electrode decreases, the value of 6/r increases, which ensures radial

diffusion at shorter time scales. A more quantitative illustration of the relationship of the

diffusion layer thickness, electrode size, and the experimental time scale can be obtained

from the following equation,"




S. (1.18)






























Figure 1.8. Growth of a diffusion layer as a function of time (t < t2 < t3 < t4 lines represent a cross section of the interface between the diffusion layer and the solution
bulk. At time t,, diffusion is linear. As time scale increases, diffusion becomes a mix of
linear and radial diffusion. Eventually, diffusion behavior transforms to a radial diffusion.















radial diffusion


\


electrode








35

where 1n is a unitless quantity, D is the diffusion coefficient (cm2/s), t is time (s), and r is the

radius of the electrode (cm). For values of t greater than 6, diffusion is considered as radial.

For values of r less than 1, diffusion is linear.

The radius of an UME is small enough to establish radial diffusion at time scales of

1 s or less. For a 1 and a 10 tm UME, the time needed to reach radial diffusion is about 0.01

and 1.3 s, respectively. Thus, at common experimental time scales the enhancement in mass

transport can be taken advantage of at UMEs. Figure 1.9 illustrates the effect of electrode

size on diffusion behavior of time scales of seconds.

For a chronoamperometric experiment, the effect of enhanced mass transport on

current density can be calculated,'3




1* (1.19)



where i" (pA/cm2) is the current density under radial diffusion conditions, iin (pA/cm2)

is the current density under linear diffusion conditions, D (cm2/s) is the diffusion coefficient

(cm2/s), t is time (s), and r is the electrode radius (cm). Based on equation 1.19, after 1 s the

current density at an UME with a diameter of 1 gm is about 100 times greater than at a

conventional size electrode (1 cm). This higher current density is particularly useful in

analytical measurements. Since the noise or background in the electrochemical

measurements is proportional to the electrode area, the high current density (faradic current

per electrode area) can allow higher sensitivity and lower detection limits.



























Figure 1.9. The effects of electrode size on diffusion behavior at common experimental time
scales (several seconds or less).

















Edge effect of UME








r r=0.5 cm
macro electrode





ultramicroelectrode r = 5 um










Circuit Design for UMEs Experiments



Although current densities of UMEs are much higher than for conventional size

electrodes, the absolute current magnitudes at UMEs are extremely small (nA to pA) due to

the small electrode radius which determines the electrode current. Thus, current

measurements at UMEs are difficult and can not be done with conventional potentiostats.

A part of this project was to design and build a potentiostat for measurements at UMEs. The

design and construction of such instrumentation played an important role in this work

because it allowed us to investigate and exploit unique properties of UMEs under different

experimental conditions. The instrumentation design took advantage of very small current

magnitudes at UMEs. In a conventional three-electrode configuration potentiostat, an

auxiliary electrode is needed to bypass the large current from the reference electrode,

otherwise the reference electrode will be polarized by the large current which will cause the

shift of the reference electrode potential. Unlike conventional potentiostats with a three-

electrode configuration, the potentiostat for UME measurements can have a simple two-

electrode configuration which is possible due to the small magnitude of the currents passed

at UMEs.6465 The following sections introduce the basic design of such instrumentation.

Details of the instrumental design will be discussed in Chapter 2.



Potentiostat for UMEs experiments at low scan rates (less than 10 V/s)


Because conventional potentiostats can not measure currents lower than pA, a current








39

amplifier is necessary to amplify currents for a conventional potentiostat. The current

amplifier built in our laboratory has different gains (102 to 10') and time constants ( 1 10,

and 100 gis) to fit different experimental requirements. By using this current amplifier with

a conventional potentiostat, pico ampere currents can be measured.



Potentiostat for UMEs experiments at high scan rates (higher than 10 V/s)



The conventional potentiostats were designed to perform cyclic voltammetry at scan

rates below 10 V/s, thus, the time constant of the noise filter in these potentiostats is set

above 100 ts to eliminate high frequency noise. In addition to the time constant of the noise

filter, the instrumental time constant is also limited by the speed of the analog-to-digital

converter (ADC) for digital potentiostats or by the speed of the plotter for analog

potentiostats. Because the large cell time constant of macroelectrodes sets the limit on a

maximum scan rate in cyclic voltammetry, before the instrumental time constant can affect

the voltammograms, there has been no need for a short instrumental time constants or fast

ADC in conventional potentiostats. However, at UMEs, the cell time constant is small (ca.

1 is) and the highest scan rate reported is as high as 200,000 V/s.26 Because of the large

instrumental time constants of the conventional potentiostats, voltammograms begin being

distorted at scan rates above 10 V/s even though the cell time constant at UMEs permits scan

rates up to 200,000 V/s. Thus, a new design of a potentiostat is needed to perform fast scan

voltammetry. Since the small current at UMEs allows a two-electrode configuration

potentiostat to be used, we modified the current amplifier described above to make a two-








40

electrode configuration potentiostat for fast scan measurements. The potentiostat which was

constructed is able to perform fast scan voltammetry up to 4,000 V/s without distortion from

the time constant of the potentiostat. For the purpose of detecting low concentrations of

analytes at high scan rates, the noise level of this homemade potentiostat was reduced by a

first order filter and low noise operational amplifiers.



Potentiostat for high-scan-rate measurements at UMEs with on-line iR compensation



This potentiostat has a two-electrode configuration with an on-line positive current

feedback iR compensation circuit. The operational amplifiers used in the circuit are current

feedback operational amplifiers, having very wide bandwidths (above 100 MHz) and high

gains. The main consideration in the circuit design was the speed (or bandwidth) of the

operational amplifiers. In order to compensate on-line for the iR drop, the current passing

through the working electrode needs to be fedback to the reference electrode as fast as

possible to makeup the iR loss in the applied voltage. The circuit is very noisy due the

wide bandwidth of the operational amplifiers. To extract the signal out of the noise,

concentrations of analytes above ImM are needed in order to produce sufficiently large

signals. Therefore, this circuit is well suited for fast kinetic measurements but not for trace

analysis.










Surface Modification of UMEs



Polymer-film-modified electrodes



Modified electrodes have received tremendous attention in the past 10 years.66'67 The

main goal is to improve the selectivity and the sensitivity of the electrodes with the

modifiers. One of the best known modifiers is Nafion (a perfluorosulfonated ionomer) which

is anion selective and repels anions with its negatively charged sites SO3". As mentioned

above, UMEs have many advantages due to their small size. However, selectivity of UMEs

is not improved as the size of the electrode decreases. An useful strategy to improve

selectivity is to modify the electrode surface with permselective films. For example, a

common application of UMEs is in in-vivo measurements of neurotransmitters such as

dopamine (DA) in the central nervous system. However, voltammetric responses of the

neurotransmitters usually suffer from the interference of ascorbic acid (AA) which coexists

in-vivo at concentrations 103 to 10 times higher than the concentration of the

neurotransmitter. UMEs coated with Nation films are usually used to resolve this problem,

since Nafion can repel anions such as AA. Nevertheless, the use of Nafion has some

disadvantages. Electrodes modified with Nafion films usually suffer from slow response

times due to low diffusion coefficients of analytes in the film, memory effects due to the

strong binding between the cation analytes and Nafion and saturation of negative sites when

cations are present at high concentration.6'6970. Besides these problems with the performance

of Nafion, thickness and quality of Nafion films is difficult to control and reproduce.








42

Recently, a new class of cation permselective films has been investigated by our

group.71'72"73 It was shown that overoxidized polypyrrole (OPPy) films can be made quite

permeable, and have ionic conductivity, although the electrical conductivity of polypyrrole

(PPy) is lost after overoxidation. During overoxidation, carbonyl groups are introduced into

the polymer backbone forming overoxidized polypyrrole (OPPy).74 The high electron

density of the carbonyl groups acts as a barrier to anion diffusion in the film. In our previous

investigation, the OPPy films (thickness of 0.1 pm) were shown to have excellent cation

permselectivity. Since OPPy repels anions with neutral carbonyl groups instead of the

negative charge sites, electrostatic binding does not occur and the OPPy has a fast response

time, fewer memory effects, and is free from the problem of binding site saturation.

A thinner ion-selective film is always desirable due to faster response time and higher

sensitivity. The permeability of films is expressed by the following equation,

aD D
P.,- -D (1.20)
b

where Pm (cm/s) is ilm permeability, a is membrane partition coefficient, Dm (cm2/s) is the

diffusion coefficient within the film, D, (cm2/s) is the apparent diffusion coefficient in the

film, and ,. (cm) is the film thickness. According to equation 1.20, the permeability of the

film increases as 6, decreases. High permeability will allow analyte to reach the surface

easily, hence a fast response time and a higher sensitivity can be achieved.

Besides the higher sensitivity gained from an ultrathin film, ultrathin organic films

may be used as a precoated material to assist growth of compact conducting polymers. This

subject has been addressed by several authors recently. Gottesfeld and coworkers have








43

reported that preassembling an organic monolayer on electrodes can assist in the formation

of a more compact conducting polymer with better adhesion characteristics." Matsue and

Uchida's group have reported that hydrophobic pretreatment of a substrate promotes lateral

growth of PPy.76 In general, growth of the conducting polymers can be manipulated at

precoated organic films because of the hydrophobic interactions between the precoated

organic film and the monomers. The OPPy ultrathin film may serve as a precoating material

to control the growth of the conducting polymers. Since the film is compact but yet

permeable to monomers, it may help grow more compact and adhesive conducting polymers.

Similar work has been reported by B61anger's group who used an ionically conducting

polymer (Nafion) to promote lateral growth of PPy.77



Surface-oxide-modified electrodes



Electrochemical pretreatment (ECP) has been found to be quite effective in

improving electrode selectivity and sensitivity.78 It is particularly useful to pretreat carbon

electrodes due to the ability to produce rich oxide surfaces.78 The properties of the oxidized

surface are different because of the high charge density and the ability of the surface

functional groups to catalyze redox reactions.7 Figure 1.10 illustrates various surface

functional groups on carbon electrodes, such as phenol, quinone, carbonyl, and carboxylic

acid, which can form as a result of surface treatment.7" The negative charge of the carboxylic

group at neutral pH and the abundance of electron density on oxygen on treated electrode

surfaces can prevent anions from approaching the electrode surface. Consequently, redox




























Figure 1.10. A possible structure of surface oxide groups on a carbon electrode (reference
78).











Phenol

Carbonyl



Lactone




Carboxylic Acid


o-quinone


p-quinone








46

reactions of anions can be deterred.79 On the other hand, cations have been known to be

concentrated on the pretreated electrode surfaces presumably due to the electrostatic

attraction by the opposite charge of the surface.78 The ability of the surface functional groups

to preconcentrate and catalyze reactions of cations improves the sensitivity and the

electrochemical kinetics increase at treated electrodes. ECP may also improve sensitivity

by cleaning debris and impurities on electrode surfaces. ECP usually increases background

currents as the surface becomes rougher and the amount of the surface functional groups is

increased. Resistance of the electrode also increases after the electrochemical pretreatment

which may be attributed to the insulating nature of the surface groups, especially oxides,

which can form on the treated electrodes. We observed that the resistance of a freshly cut

fiber carbon electrode is about 6 kQ and it can go up to 50 kQ after pretreatment.


Purpose of Work



The purpose of this work was to develop new strategies and fast scan techniques for

analytical measurements with UMEs in aqueous solutions. We focus most of our efforts on

the exploitation of fast scan technique with UMEs in the characterization of fast redox

reactions of biological species, improvement of S/N, selectivity and detection limits. Fast

redox reactions investigated with fast scan voltammetry in aqueous solutions at carbon fiber

UMEs helped us understand the properties of the electrode surface and its effect on the redox

reactions at fast scan rates. To improve selectivity of UMEs in biosensing, UMEs were

modified with ultrathin polymer films. The ultrathin films provide selectivity by deterring








47

anion diffusion but allowing cations to reach the electrode surface efficiently which results

in higher sensitivity for cationic analytes. The properties of the ultrathin films were

characterized and their applications in bioanalysis were demonstrated in this work.

New strategies for using fast scan voltammetry in detection of trace biological species

in complex environments have been proposed and developed. These strategies involve signal

averaging, improved temporal resolution, improved instrumental design, and selectivity

enhancement through enhancement of the kinetic differences between the analyte and the

interferant.













CHAPTER 2
EXPERIMENTAL SECTION


Reagents and Solutions



Acetonitrile (MeCN) was obtained from Fisher. Pyrrole and tetrabutylammonium

perchlorate (TBAP) were from Kodak. Potassium ferricyanide (Fe(CN)63) and ascorbic acid

(AA) were obtained from Mallinckrodt. Hexaamineruthenium (III) chloride (Ru(NH3)63+)

was purchased from Alfa Products. Pyrrole-2-carboxylic acid (Py-2-COOH) was purchased

from Aldrich. Dopamine (DA) was purchased from Sigma. All chemicals were used as

received. Uric acid (Sigma) was prepared in 1 M phosphate buffer of pH 7.0.

Hexaamineruthenium (III) chloride and potassium ferricyanide were prepared in 1 M

aqueous KC1 solution or in 0.5 M phosphate buffer at pH 7.4. DA and AA were prepared

in 70 mM phosphate buffer of pH 7.4.



Electrodes



Reference and Auxiliary Electrodes



A saturated calomel electrode (SCE) was used as the reference electrode. When








49

MeCN was used as the solvent, a quasi-reference electrode (a 5 cm long Ag wire) was used

as a reference to avoid water contamination of the organic solvent and to prevent

establishment of liquid-liquid junction potentials.80 If a three-electrode potentiostat was

used, a 1 cm2 platinum foil was used as an auxiliary electrode.



Working Electrodes



Glassy carbon electrodes (ca. 0.67 cm2 Electrosynthesis Co.), carbon fiber

ultramicroelectrodes (7 pm diameter, Textron Specialty Materials, Inc.) and platinum

ultramicroelectrode (5 pm Pt wire sealed in glass, Goodfellow Co.) were used as working

electrodes. The experiments using rotating disk electrode (RDE) were conducted with an

IBM RDE Controller and Rotator. The screw-on electrode tip was made from Teflon, with

a 3 mm diameter glassy carbon rod sealed into the end by heat pressing. Prior to each

measurement the working electrodes were polished on an Alpha A polishing cloth (Mark V

laboratory) with gamma alumina suspensions of 0.1 Pm particle size (Gamal, Fisher

Scientific Co.) and were ultra-sonicated afterwards in doubly distilled water for 5 minutes.



Fabrication of UMEs



Carbon fiber electrodes sealed in heat shrinkable Teflon tubes

Carbon fiber ultramicroelectrodes were prepared from 7 pm diameter carbon fibers

(Textron specialty materials 7440-44-0) and were sealed in a heat shrinkable Teflon tube








50

(Zeus ZDS-S-036). This Teflon tube had two layers and when it was heated, the outer layer

of the tube shrunk and the inner layer melted. The inner diameter of the tube could shrink

to virtually zero. The details of making the UMEs are given below.

A piece of carbon fiber (ca. 3 cm long) was glued to a copper wire (ca. 6 cm long)

with silver epoxy ( EPO-TEK 410 E, Epoxy Technology Inc.). The epoxy was allowed to

dry. The silver epoxy serves as a conductive connector which connects the carbon fiber and

the copper wire. The carbon fiber connected to the copper wire was then placed into a

Teflon tube and the tube was wrapped with an aluminum foil. A lighter or gas torch was

used to heat the foil for about half a minute until the carbon fiber was well sealed in the tube.

The heating process needs to be carefully controlled to avoid overheating. The tip of the

carbon fiber sealed with Teflon was then cut with a scalpel to expose the disk-like UME. By

following this procedure a well sealed ultramicroelectrode can be made easily and quickly.

To avoid introducing contaminants, the electrodes were used directly without polishing. An

illustration of this UME is shown in Figure 2.1-A.



Carbon fiber electrodes sealed in epoxy

Carbon fiber was first connected to a copper wire with silver epoxy. After the epoxy

dried, the copper wire with attached carbon fiber was inserted into a capillary glass tube or

a plastic micropipet tip (ca. ImM radius). The copper wire was then glued with epoxy to the

capillary glass tube for easy handling. Another epoxy mixture was made from the epoxy

resin and the hardener. The mixture was heated on a hot plate until the epoxy became

transparent and liquid-like. The epoxy liquid was carefully stirred to get rid of air bubbles


























Figure 2.1 Schematics of UMEs used in this work. (a). Carbon fiber (7pm) sealed with
Teflon. (b). Carbon fiber sealed with epoxy (c). Pt wire sealed (5gm) in glass.







































Copper wire







Soder

insulator

Soft glass


Pt wir


(B) (C)








53

trapped in the liquid. The epoxy liquid was then pipetted into the capillary glass tube. The

whole tube was then placed in an oven and heated for one hour at ca. 150 OC. After the

epoxy hardened, the extra epoxy at the tip was removed with sandpaper to expose a disk-like

UME. This working electrode was polished prior to each measurement on an Alpha A

polishing cloth (Mark V laboratory) with gamma alumina suspension of particle size of 0.1

pm (Gamal, Fisher Scientific Co.). An illustration of this UME is shown in Figure 2.1-B.



Platinum wires sealed in soft glass

Metallic electrodes can be to be sealed with glass. A piece of Pt wire (r=2.5 pm) was

inserted into a soft glass tube (melting point = 500 oC, Kimble Inc). A gas torch was used

to melt the tip of the glass, which caused the melting glass to seal the Pt wire. The melting

process has to be carefully controlled to minimize the formation of air bubbles in the melting

glass which will cause imperfections in the seal. After the glass tube cooled to room

temperature, the glass tip was removed with sandpaper until the end of the Pt wire was

exposed. A disk-like Pt UME was formed. The electrical contact was made by inserting a

Cu wire into the other end of the glass tube. Pieces of solder were put into the glass tube and

the glass tube was gently heated until the solder melted and embedded the Pt and the Cu

wire. The connection between the Pt wire and the Cu wire was formed after the glass tube

cooled down to room temperature and the melted solder became solid. Figure 2.1-C shows

a cartoon illustration of this type of electrodes.










Electrode Pretreatment



The electrochemical pretreatment of carbon fiber electrodes was carried out in 1 M

phosphate buffer in the experiments with uric acid as analyte, and in 1 M KCI solutions in

kinetic measurements with Ru(NH3),(3+/2+) and Fe(CN)6/4-'). The parameters for the

treatment were as follows: square-wave potential, lower potential limit -1.0 V, upper

potential limit +2.0 V, frequency 70 Hz; duration 5 min. The conditions for the pretreatment

were chosen because a previous report had shown that such pretreatment can improve the

reaction kinetics of uric acid."

When analysis with fast scan voltammetry at carbon fiber electrodes, the electrodes

were pretreated to stabilize the background. The electrodes were treated by a cycling for 30

minutes in the potential window from -0.8 to +1.2 V in 70mM phosphate buffer.



Instrumentation



IR Compensation Circuit and Its Operational Principle



All iR compensation experiments were carried out with a homemade instrument (see

Figure 2.2) of a design which was adapted from that described by Saveant's group.8 The

design was modified to suit our experimental purposes. An operational amplifier (OP27)

was added to the circuit to further amplify signal and eliminate noise at low scan rates (less

than 1000 V/s). The potentiostat had a two-electrode configuration and an on-line positive



























Figure 2.2 Schematic of a potentiostat built for fast measurements with iR compensation.






56

N





5J
:3,t




aU





a,
0: -e
U0 (0




'IL





B E






I 3+
-rJ








57

current feedback iR compensation circuit. The potentiostat was based on two types of current

feedback operational amplifiers (CLC 401 and CLC 400) obtained from the Comlinear

Corporation.

Voltage at the positive input of the amplifier U3 equals to the current to be measured

(i) times the sampling resistor (R8). Voltage iR8 is applied to the positive input of the

amplifier U3 and is then amplified by a factor of 1 + R7/R5 (R7 and R5 are the feedback and

the gain resistors of the amplifier U3, respectively). Thus, the output voltage at output 1, U,,

to be measured by the digital oscilloscope, is:



U=R( I.R/R)i (2.1)



where U, is the voltage output (V), R are the resistances (0) of the resistors in the circuit, and

i is the current measured at the working electrode, W. The output voltage was monitored and

recorded on the LeCroy 9310 digital oscilloscope.

The operational amplifiers used in the circuit, CLC 400 and CLC 401, are current

feedback operational amplifiers. Unlike the conventional voltage feedback operational

amplifiers, the bandwidth of the current feedback operational amplifiers depends only on the

magnitude of the feedback resistance such as R7 (see Fig 2.2). Therefore, the current

feedback operational amplifiers have a larger bandwidth (at least of a factor of 5) which is

very important for high speed experiments, such as fast scan voltammetry. The optimum

value of R7 is 1500 0 for the CLC401 operational amplifier used in the current

measurements.








58

The current to be measured crosses the sampling resistance R, (332 0), the value of

which is chosen so as to be negligible compared to the input impedance of the amplifier U3

(ca. 200 ka). The output voltage, Us, is fed back to the operational amplifier U2 through a

variable resistor, R&. The magnitude of the feedback is controlled by the fraction of the

variable resistance which is applied (i.e. p; p= (R6-x)/R6; OsxsR,). The summation of the

positive feedback voltage, pR,(l+R7/R5)i, with the voltage delivered by the function

generator is carried out at the amplifier U2. Because the positive feedback is injected into

the positive input of the amplifier U2, and the amplifier has an unit gain, the potential at the

output of the amplifier U2 is thus:

U 2pR,(l. R, /R,)i = U R i (2.2)



where U (V) is the voltage from the waveform generator (i.e. the original voltage on the

reference electrode), p is the fraction of the positive feedback applied to the amplifier U2,

i (nA) is the current to be measured, and R, = 2pR,(I+R7/R5) is the "compensation"

resistance. Re can be change by adjusting p. A major concern in the design of an iR

compensation circuit is to obtain an exact compensation, not an undercompensation nor an

overcompensation. In other words, we need to ensure that the compensation resistance (Re)

is exactly equal to the uncompensated cell resistance (R). By doing so, the potential

affected by the iR loss, iR., will be exactly compensated by the iR,.

With conventional macroelectrodes, it is possible to design an instrument such that

sustained oscillation of the output signal occur when the compensation and the cell

resistances are practically equal. A simple determination of R is ensured, leading to a precise








59

knowledge of the remaining uncompensated resistance in each experiment. This is so

because the time constant of the instrument is much shorter than the cell time constant of a

macroelectrode, RkCd. This is no longer true at a UME, because cell time constants with

UMEs as working electrodes are comparable with the instrumental time constants. The

following discussion shows how to compensate the iR drop with UMEs as working

electrodes with on-line iR compensation.

The on-line iR compensation circuit can be analyzed in terms of a second order

approximation as shown by the type of oscillations obtained upon increasing p (i. e. the

fraction of the compensation resistance R&) when it is connected to a resistor-capacitor

dummy cell representing the solution resistance and the double layer charging in the absence

of the faradaic current. The operational expression for the output voltage is thus:83-"


R,(l. R I RCy,v
U (2.3)
(R,CF/)p 2+[R,(1/Cd)-Rj]Cd p+.



where p is the Carson-Laplace variable, Cd is the double layer capacitance (F), v is the scan

rate (V/s), o is the bandpass pulsation of the instrument and R, is the total resistance to be

compensated, i.e., R4 + R +R8. A sustained oscillatory behavior is obtained when:

R, (I/Cd a) R, 0 (2.4)



i.e., when:


R, R R- = 0


(2.5)








60

The Rk is the "overcompensation resistance" which equals to 1/Cdo. Thus, if we know R,

exact compensation can be obtained by first increasing R, until sustained oscillatory behavior

is reached and then decreasing it by an amount equal to R1. R, can be derived from the value

of the double layer capacitance (Cd) and from w. Cd is obtainable from the height of the

double layer charging current. The parameter o can be obtained as follows. Dummy cells

with known resistances (R1) and capacitances (Cd) were used to simulate the electrochemical

cell. The compensation resistances (Re) were obtained by increasing p until sustained

oscillations appeared. Since R1, Cd, and R, were known, the parameter o can be obtained

from equation 2.4. In our experiments, the value of o was determined as 3.3 x 107 rad/s.

During the experiments, the ohmic drop compensation was performed by increasing

the positive feedback (p) until sustained oscillations appeared on the cyclic voltammetric

trace. The feedback was then decreased by l/Cdw before recording the current-potential

curve. The feedback should be carefully tuned to reach sustained oscillations but total

oscillations should be avoided. Sustained oscillations are apparent when a noticeable

oscillation appears in the beginning of the voltammogram as the compensation resistance

increases. If the compensation resistance continues to increase, the magnitude of the

oscillation will suddenly change from several mV to 7 V (same as the power voltage of the

operational amplifiers) and will destroy the voltammogram. This type of oscillation is called

total oscillation. The reason for avoiding total oscillations is that the total oscillations

(magnitude +7 to -7V) feed back to the electrochemical cell and will cause pretreatment of

the carbon fiber electrode. In our experiments, after total oscillations occurred a change in

the AEp of electrochemical probes, and the background current was observed.








61

Experiments were performed using iR compensation when the scan rates exceeded

1000 V/s. At scan rates up to 1000 V/s, iR compensation was not necessary, since the iR

loss was insignificant. Because the current decreases as the scan rate decreases, at scan rates

below 1000 V/s, the noise in the feedback circuit severely interferes with the small faradaic

current. Thus, the circuit was modified to suit the experimental needs. An amplifier ( OP27,

8 MHz bandwidth) was added to further amplify the current and filter out the high frequency

noise when the scan rates were lower than 1000 V/s. The OP27 amplifier has a much

narrower bandwidth (8 MHz) compared to the other two amplifiers (150 MHz for CLC 401

and CLC 400). The narrow bandwidth of OP 27 can function as a filter to eliminate the high

frequency noise. Rechargeable Ni-Cd batteries ( Radio Shack) were used as a power supply

to avoid 60 Hz power line noise. An offset adjustment was used to adjust the offset of the

output voltage due to the voltage drift of the operational amplifiers. All connections were

made on printed circuit boards to minimize stray capacitances and crosstalk between leads.

For the same reason, the leads connecting the digital oscilloscope, the working electrode, and

the reference electrode were shortened as much as possible. At lower scan rates, majority

of signals were acquired repeatedly 50 times, then averaged to improve signal-to-noise ratio.

Data acquisition, signal averaging, background substraction and the measurements of the

peak current and the peak potential were all performed on the LeCroy 9310 digital

oscilloscope.

The cyclic voltammograms of the analytical species in fast measurements were

repeatedly acquired, averaged and stored in the digital oscilloscope. Then the

voltammograms of a blank solution were acquired, averaged and subtracted from the








62

voltammograms of the analytical species. The data acquired by the LeCroy oscilloscope were

transferred to a Northgate 386 PC for the display and the semi-integration. The data transfer

is through an RS-232 interface on the oscilloscope and the communication port number two

on the computer (Northgate 386 PC). Two QBasic programs were locally written to

communicate with the digital oscilloscope and to perform the semi-integration (see Appendix

A and B). A function generator (Model 182A, Wavetek) and an universal programmer

(model 175 EG & G) were used to generate triangular waveforms for cyclic voltammetry.

The function generator was used as a trigger source to simultaneously trigger the

oscilloscope and the universal programmer.



The Current Transducer and the Potentiostat for the UME Experiments



The schematic of the current transducer is shown in Figure 2.3. The design of the

current transducer was based on Faulkner's design."' We modified the circuit to make it

more flexible in the instrumental time constant, and added an output (bypassing the OP27)

for fast measurements. The power supply for this instrument was made from rechargeable

batteries to minimize 60 Hz power line interference. Two 7.2 V Ni-Cd Turbo racing batteries

(purchased from Radio Shack) were used because of their high current output (1200 mAh).

The input operational amplifier (AD515A, Analog Devices) was a monolithic,

precision, low power, FET-input operational amplifier. It served as a current-to-voltage

converter which amplifies and converts the input currents to voltages. This operational

amplifier has a fast slew rate (3.0 V/ts), ultralow bias current (less than 50 fA with 5V




























Figure 2.3 Schematic of the current transducer.






































I pF

I Mn
0.1 pF

10 M
10 pF

I MD
--pF

10 MQ
100 pF Rf

I MO
10 pF

10 Mo
-I I- ipF

100 Ma


--- -----
0.1 nP



-c- 4---
I nF


10 onF


R2

--10 Kn


To oscilloscope
For high scan rate caperiments


To BAS potentiostat


\ For low scan rate ex
Ref 2

SADSI5
S10 3 OP27 AWN

+ 010 IK


periments


.... ........... r .......








65
power supplies), low input noise voltage (50nV/Hz), and very low input noise current

(0.01 pA rms). The bandwidth of the operational amplifier (AD515A) is 1 MHz. The low bias

current and the low noise make this operational amplifier suitable for low-level current

measurements, yet the fast slew rate and the wide bandwidth make it adequate for fast scan

voltammetry at scan rates of kV/s. Various resistors (1, 10 and 100 MQ ) and capacitors

(0.1, 1, 10 and 100 pF) in the feedback loop control the gains (100, 1000 and 10000) and RC

time constants (1, 10 and 100 ps). Since the current-to-voltage converter (AD 515A) inverts

the input signal phase, a second operational amplifier (OP27) with a unit gain is used as an

inverter to invert the phase of the output signal back to normal. The capacitors and the

resistor (R2) on OP27 function as a first order filter, and the RC time constant should be

adjusted not to exceed the RC time constant of AD 515A

The commercial potentiostat used in the experiments is Bioanalytical Systems

(BAS100) Electrochemical Analyzer. When used with the BAS the output of OP 27 has a

10 kO (Ro) resistor connected to the working electrode lead of the BAS. In most commercial

potentiostats such as the BAS the working electrode lead feeds a virtual-ground summing

junction of a current to voltage converter. The current measured with the BAS in

conjunction with the current transducer is Rfx(RI/R2)x(RBAs/R)xi, and is RBASXi without the

current transducer. Because R, equals to R2, the net gain of the current transducer will be the

ratio of R/R (i.e. the ratio of the feedback resistance of OP AD515A to the output resistance

of OP 27). For example, for a 10 kO value of Ro the gain is 100, 1000, and 10000 when Rf

is 1 10 and 100 MO, respectively.

For stand alone use in fast scan measurements, the output of OP AD515A bypassed








66

the OP27 and was directly connected to a digital oscilloscope (LeCroy 9310). The digital

oscilloscope measures the output voltage (the transduced current). The ratio of the transduced

current (output voltage) to the input current (i.e. the current flowing at the UME) was

determined alone by the R, on OP AD515A. The conversion factors for the potentiostat were

1, 10 and 100 V/pA when the resistors (Rf) in the feedback loop of the current transducer

were 1, 10 and 100 MO respectively. Thus, for a conversion factor of 10V/IA a 10 mV of

voltage measured by the oscilloscope corresponds to a 1 nA of the analytical current. The RC

time constant of the current transducer was controlled by the feedback resistance (Rf) and the

capacitance on OP AD515A. For example, a 1 MO of Rfand a 1 pF of a capacitor will give

an 1 ps of the RC time constant.



Instrumental Setup for Low and High Scan Rate Experiments



At low scan rates, the current amplifier is connected to the BAS. The ground of the

instrument should be connected to the ground wire of the BAS to insure proper grounding

(see Figure 2.4). The grounding is important for the output voltage of the circuit to be

measured correctly by the BAS. A faraday cage was used to shield the circuit from noise and

the faraday cage was also connected to the ground. At high scan rates the potential

waveform is generated from the function generator and the waveform is applied to the SCE

reference electrode. The current at the working electrode is transduced to voltage and

amplified by OP AD515A. The transduced current bypasses OP27 and is directly measured

with the digital oscilloscope (Figure 2.3). Simultaneously, the waveform applied to the








67

reference electrode is measured at a separate channel of the digital oscilloscope (Figure 2.5).

The stored waveform can be later transferred to a computer for plotting the cyclic

voltammograms. Since the phase of the potential at the working electrode is reversed

relative to the waveform potential at the reference electrode the phase of the potential

waveform measured with the oscilloscope needs to be inverted. In addition, the phase of the

current passed at the working electrode is inverted by the inverting input of OP AD515A.

Thus, to obtain a correct plot of current vs potential (i.e. voltammogram), the phase of the

potential waveform and the current stored in the digital oscilloscope are inverted by a

computer program (Origin) before plotting the cyclic voltammograms on the computer.



Voltammetry at low scan rates

At low scan rates, the circuit shown in Figure 2.3 can be used as a current amplifier

which allows pico ampere currents to be measured with the BAS. The minimum current

range of the BAS is 0.1 pA. The gain of the current amplifier ranges from 100 to 10,000.

Combing the BAS with the current amplifier, the minimum current range can be low to 100

pico amperes. To minimize noise, a first order filter is used in the circuit (Figure 2.3, OP

AD515A ). The time constant of the potentiostat is controlled by the time constant of the

first order filter in the circuit (Figure 2.3).

Ideally, the time constant should be as large as possible if the experimental conditions

allow it. Traditionally, for performing cyclic voltammetry a suitable time constant of a

potentiostat is decided by the following equation,"687



























Figure 2.4 Instrumental setup for UME experiments at low scan rates.




























Reference electrode


SAuxiliary electrode


Cell






Working electrode


Input


Current Output
transducer
/amplifier



Ground


BAS


Ground


i



























Figure 2.5 Instrumental setup for UME experiments at high scan rates.
































Function generator


Digital O'Scope


Reference electrode
Working electrode



Potentiostat



Cell Syrnge


RS 232


Personal Computer











RCnv 4mV (2.6)



where the R is the resistance (0) of the first order filter (i. e. the feedback resistance Rf for

OP AD515A), C is the capacitance (F) of the first order filter (i.e. the capacitance of the

feedback capacitor for OP AD515A), n is of electrons number in the redox reaction, and u

is the scan rate (V/sec). At low scan rates (less than 40 V/s), the time constant of the system

is set at 100 ps which allows use of scan rates up to 40V/s for one electron reactions (or 20

V/s for two electron reactions) with negligible distortion in the AEp. Figure 2.6 shows the

measured current (pico amperes) obtained from the BAS in conjunction with the amplifier.



Voltammetry at high scan rates

One of the advantages of using UMEs is that the current is small and thus there is no

need to use an auxiliary electrode. Therefore, the circuit of the potentiostat can be simplified

into a two-electrode configuration. The current transducer used for low scan rates can be

used as a potentiostat for fast scan measurements in conjunction with a waveform generator

and a digital oscilloscope. The potentiostat was designed to have a two-electrode

configuration (one reference and one working electrode). The potential waveform generated

from the function generator is applied to the reference electrode. Current passing through

the UME working electrode is converted and is amplified to voltage through the current

transducer. The output of the current transducer (voltage) and the potential waveform at the

reference electrode are then simultaneously measured with the digital oscilloscope. The

transducer is designed with flexibility and low noise in mind. Different gains (106 to 108

V/A) are obtained with different resisters (1 to 100 MO).



























Figure 2.6 Cyclic voltammogram of 0.1 mM Fe(CN) '- in 70 mM pH 7.0 phosphate buffer
obtained from the current transducer connected to the BAS. The scan rate is 10 mV/s. The
gain of the current transducer is 10000 and the time constant of the first order filter is 100
ps. The working electrode is a carbon electrode (7 pm).



















80 I


70 L


60 -


50 h


40 k


//
Ji
//

//



r /
--!


30 H


20 H


10 H


0I I I I 1 I
400 300 200 100 0


E(mV)


0
<
*M








75

As was described above, a first order filter with different time constants (1 to 100 ps)

is used with OP AD515A (Figure 2.3) to minimize noise. Again, equation 1 can be used as

a guide for setting the time constant of the potentiostat. A 10 is time constant, for example,

will allow a scan rate of 400 V/s to be used for one electron reactions with negligible

distortion in the voltammogram. With 1 us time constant, scan rate of 4,000 V/s will be the

maximum scan rate allowed for obtaining undistorted voltammograms. Figure 2.7 shows the

voltammogram of dopamine at a scan rate of 2,000 V/s. Since the large charging current can

obscure the faradic current at high scan rates, the voltammogram needs to be background

subtracted. The background voltammogram was obtained in a buffer solution and stored in

the digital oscilloscope or in the computer. A buffer solution with analyte was then added

and the voltammogram was acquired. The voltammogram was background subtracted to

obtain a clear voltammogram.

The potentiostat design is not limited to fast scan voltammetry. It can be used in all

electrochemical applications which need a fast instrumental time, such as fast potential step

amperometry. The digital oscilloscope is the most expensive part of the potentiostat, because

the digital oscilloscope has a very high bandwidth (150 MHZ) as well as other extra

functions. However, a bandwidth of 10 MHZ is sufficient for all fast electrochemical

measurements described above. Thus, the digital oscilloscope may be replaced with an

analog-to-digital converter which has a narrower bandwidth but is much less expensive.



























Figure 2.7 Voltammogram of 0.18 mM dopamine at 2,000 V/s in 70 mM pH 7.0 phosphate
buffer with a carbon electrode (7 jpm). The voltammogram is background subtracted and
signal averaged 1000 times.

















































SI I I I


0.5

E (V)


00 -0.5 -1.0


1 1


U










Fundamentals of Electrochemical Methods



Cyclic Voltammetry



In cyclic voltammetry (CV) experiments, the potential is scanned from an initial

value where no reaction occurs to a final value where the reaction rate is limited by diffusion

current. The potential is then swept back to the initial potential. The time scale of the

experiment is expressed as scan rate, v (V/s), which is the rate of potential sweep. The

potential range between initial and final potential is called the potential window. In a cyclic

voltammogram which is recorded, the current from the redox reaction gradually rises after

the potential is swept past the potential where the electrochemical reaction occurs. The

current then reaches a peak and then gradually decreases. The potential where the current

reaches a maximum is called the peak potential, Ep. The potential difference between the

reduction, E, and the oxidation, Ep,, peak potentials is expressed as a AEp =Ep-E,. Figure

2.8 illustrates the details of a voltammogram and the relationship of time scale potential,

and current.

The value of AE, is an indication of a reversibility of a redox reaction. For an one

electron reversible reaction, the AE, value is around 58 mV." For a quasi-reversible redox

reaction, the AEp is between 60 and 212 mV.88 If the value of AE, is above 212 mV, the

reaction is considered irreversible." The theoretical peak current for a diffusion controlled

reversible reaction can be expressed as follows:88












S. (2.69x 10)n nAD, 2vl 'C (2.7)


and for a diffusion controlled irreversible reaction, the peak current is:




i (2.99x1IO)n(an.)'nAD, "v'/2C (2.8)


where n is the number of electrons per mole oxidized or reduced, A is the electrode area

(cm2), Do is the diffusion coefficient (cm2/s), v is the scan rate (V/s), Co' is the bulk

concentration of the redox species (mol/cm3), a is the transfer coefficient, and n, is the

number of electrons involved in the rate determining step.

Equations 2.7 and 2.8 define voltammetric peak currents for diffusing species. For

adsorbed species, the peak current for a reversible reaction is:

n'F'Avr
n .iF2 (2.9)
4RT

and for an irreversible reaction is:




nan F2Avr
S nan.F Avl (2.10)
ip 2.718RT



where n is the number of electrons per mole oxidized or reduced, A is the electrode area

(cm2), v is the scan rate (V/s), r, is the surface excess of the redox species (mol/cm2), a is

the transfer coefficient, n. is the number of electrons involved in the rate determining step,




























Figure 2.8 A typical cyclic voltammogram.





































0
L.
UJ


V
0.
w

Lu
II

LU






Lu


__cc __---








82

R is the gas constant (J mol' K-'), T is the temperature (K), and F is the Faraday constant

(C).

Voltammograms at conventional electrodes (ca. mm in diameter) are peak-shaped and

the peak currents depend on scan rate (equations 2.7 and 2.8). At UMEs, the

voltammograms are sigmoidal in shape and the limiting current is independent of scan rate

at scans up to ca. 1 V/s which results from an edge effect (radial diffusion). This effect has

been described in Chapter One. Since the radius of UMEs is small enough compared to the

thickness of the diffusion layer under common experimental conditions (i.e. scan rates of up

to 1 V/s), radial diffusion can be established at a UME. Because radial diffusion is so

efficient, steady state mass transport can be obtained. In the steady state, the current is time-

independent and 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 is expressed by the following

equation:3

i = 4nFDC'r (2.11)



where n is the number of electrons per mole oxidized or reduced, r is the electrode radius

(cm), D is the diffusion coefficient (cm2/s), C' is the bulk concentration of the redox species

(mol/cm3) n is the number of electrons involved in the reaction, and F is the Faraday

constant (C).

It should be noted that the unique voltammetric response of UMEs exists only at low

scan rates (less than 1 V/s) where the diffusion layer is thick compared to the electrode

radius. As the scan rate increases (higher than 1V/s for a 5mrn UME), the sigmoidal shape








83

of the voltammograms is gradually transformed into a peak-shaped voltammogram.

Eventually, at sufficiently high scan rates (above 100 V/s), the shape of a voltammogram

obtained with a UME at high scan rates is the same as that with a conventional electrode. In

fact, the diffusion behavior at a UME at high scan rates is the semi-infinite diffusion

behavior as described for conventional size electrodes. Therefore, all theories for

conventional size electrodes can be applied to analyze the response of UMEs at high scan

rates.



Kinetic Measurements with Cyclic Voltammetry



Voltammetric AE (Fig. 2.8) is an indicator of the reversibility of a redox reaction.

For a one electron reaction with no adsorption, AE between 58 and 212 mV is considered

as a quasi-reversible behavior and from the AEp values for a quasi-reversible reaction, kinetic

information can be obtained for the reaction."

The electrochemical redox reaction rate constants depend on the potential applied to

the electrode surface. Therefore, a redox reaction with a low inherent reactivity may have a

large reaction rate constant if a large overpotential is applied to the electrode. The

electrochemical (heterogeneous) reaction rate constant can be expressed as follows:88


k ke -e'AE-E)


(2.12)













k,,d kOe (-'*E-E") (2.13)


where k,, (cm/s) is the heterogeneous oxidation rate constant at a potential E, k, (cm/s) is

the heterogeneous reduction rate constant at a potential E, ko (cm/s) is the standard reaction

rate constant measured at E', a is the transfer coefficient, f is equal to F/RT which is a

constant (38.92V-') at a temperature of 25 C, n is the number of electrons involved in the

redox reaction, E' (V) is the formal potential for the redox reaction, and the term (E-EO) is

the overpotential.

According to equations 2.12 and 2.13, kox and k) are a function of the applied

potential E (or the overpotential). At the potential of E=Eo, both ko, and k, are equal to k.

The ko is independent of the applied potential but depends on the inherent reactivity of the

redox system. Thus, we can use standard reaction rate constant (k0) as a reference point (an

indicator) for the reactivity of a redox system without confusion caused by differences in the

applied potential.

A method developed by Nicholson can be used to calculate standard reaction rate

constants from cyclic voltammograms.9 A kinetic parameter is determined from the AE,

values in cyclic voltammograms of quasi-reversible systems. The kO is then determined from

the value r and the scan rate. Their relationship is expressed by the following equation:8"89



ko
9 [D (2.14)
[Dn vy(nFIRT)]'








85

where is a function of AE,. The larger the AEp, the smaller the *. k (cm/s) is the standard

reaction rate constant. The value of r can only be obtained by numerical analysis. Figure

2.9 is a plot of AE vs ,. It should be noticed that the coordinate of is logarithmic, thus,

a small error in AEp will cause a large deviation in (or in calculated k).

For redox reactions of adsorbed species, Nicholson's method is not valid. However,

a method developed by Laviron can be used to calculate the standard reaction rate constant.

According to Laviron's theory, the standard rate constant of an adsorbed species can be

expressed by a following equation:"




RT a(1-a)nFAE
logk0 alog(1-a).(l-a)loga-log( --)- -. (2.15)
nFv 2.3RT


The equation is valid for nAEp > 200 mV. In the above equation, k is the standard

reaction rate constant (s'), a is the electron transfer coefficient, n is the number of electrons

involved in the reaction, R is the gas constant (J mol' K'), T is the temperature (K), F is the

Faraday constant (C), v is the scan rate (V/s), and AE, is the peak separation measured from

a cyclic voltammogram.



Semi-Integration Analysis



Voltammetric current (i vs E(t)) can be integrated as a function of square root of time

to produce a steady state response and the method is referred to as semi-integration analysis.

Semi-integration analysis is useful for obtaining kinetic information and in verification of


























Figure 2.9 A plot of AE, vs kinetic parameter i (reference 88 and 89).











































IO
\



S\50 \





100











1 10
i : \




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EMQDJE2DK_5XFKPU INGEST_TIME 2012-02-29T17:09:56Z PACKAGE AA00009028_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES