Wuuflfi
x

22‘3"“

x541

Wilmififif‘

N100

.. :2...
0‘ .4 1..

u .
.. III

.36

 

 

 

 

 

 

- Luv;
mama:
,

. lflrbvhsn.

3

95‘
iii.

a-Ei‘i‘f’éffiff

» 1. uiindfltn!
; ,. its .1
f It :11 fin!

“H".
.11. 1%? l
«@8511. .. . . :flmu
”.121... rue
‘ ”1:11;!

.&1

 

 

 

L007

I

 

LIBRARY
Michigan State
University

This is to certify that the
dissertation entitled

APPLICATION OF BORON—DOPED DIAMOND
ELECTRODES FOR METAL ION ANALYSIS BY ANODIC
STRIPPING VOLTAMMETRY

presented by

ELIZABETH ANN MCGAW

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemistry

 

 

 

C/zfirfl/M

flajor Professor’s Signature

7%; /Z 250}

Date

MSU is an affirmative-action, equal-opportunity employer

--—.-.----—n- .- -uu.-_--

--o-.--I-.-.--.—---—-.- - --.----u-u--n------

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATEDUE

DATEDUE

DAIEDUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p/ClRC/DaleDue.indd-p 1

 

APPLICATION OF BORON-DOPED DIAMOND ELECTRODES FOR METAL
ION ANALYSIS BY ANODIC STRIPPING VOLTAMMETRY

By

Elizabeth Ann McGaw

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2007

ABSTRACT

APPLICATION OF BORON-DOPED DIAMOND ELECTRODES FOR METAL
ION ANALYSIS BY ANODIC STRIPPING VOLTAMMETRY

By
Elizabeth Ann McGaw

Anodic stripping voltammetry (ASV) is an established method for
monitoring heavy metal ions, and the overall aim of this work was to fully
evaluate and understand the properties of boron-doped diamond (BDD) as they
relate to its use in ASV. In order to fully characterize the properties of BDD,
several inquiries were conducted: (i) comparison of the electrochemical
properties and response with the traditional Hg electrode, (ii) exploration of the
utility of the material in microgravity, (iii) determination of the usefulness of BDD
for more complex matrices (e.g., urine, blood), and (iv) investigation of ways of
preventing electrode fouling (e.g., coatings, complexing agents) in solutions
containing high protein concentrations (e.g., whole blood).

In addition to having many of the desirable features of Hg, BDD has some
inherent advantages over Hg electrodes; it is non-volatile and non-toxic. BDD
has a wider anodic potential limit and a tower background current density than
Hg. Both BDD and Hg coated glassy carbon (Hg-GC) provided detection limits
for all the metal ions (Zn2+, Cd2+, Pb”, Cu”, Ag“) in the mid to low ppb range
(S/N 2 3). Although the magnitude of the stripping peak current and charge were
greater for Hg-GC; this was offset by a comparable reduction in the background

noise for BDD resulting comparable or lower limit of detection values.

The utility of BDD for ASV analysis of heavy metal ions (e.g., Ag“) in the
water supply on-board the lntemational Space Station (ISS) was also
investigated. A concentrated buffer solution allows adjustment of the solution
conditions (e.g., pH, electrolyte concentration) and a prototype cell for
microgravity was developed. Because elimination of the N2 purge step was
necessary for this application, dissolved 02 was determined to have some effect
on Pb2+ and Cu2+ determinations. The use of an oxygen scavenging molecule
(ascorbic acid) to chemically remove the 02 mitigated this problem to a large
extent. BDD was found to exhibit good long-term response stability for Ag“
(<10% relative standard deviation in response over 90 days). Analysis of
simulated water samples in the microgravity cell using BDD yielded accurate Ag"
concentrations (< 5% error).

BDD was also used for ASV analysis of metal ions in biological matrices.
A linear calibration curve (10 - 250 ppb) was obtained in a urine simulant and the
response was reproducible electrode-to-electrode (R80 6%). Stability tests in a
rat urine sample showed no fouling effects over a 6-hour period. Although BDD
is resistant to fouling, it cannot be used directly in whole blood. However, Nafion-
coated BDD is successfully used for sz‘” analysis with a linear calibration
response (r2 > 0.98). A solution-based fouling prevention method using sodium
dodecyl sulfate (SDS) to aggregate with the blood protein was also promising.
When the electrolyte contained SDS, the signal for sz+ obtained in the presence

of albumin was comparable to the signal obtained in a standard solution.

Copyright by
ELIZABETH ANN MCGAW
2007

ACKNOWLEDGMENTS

There have been many people that have helped me during my time here
at Michigan State. First, I would like to thank my advisor, Dr. Greg Swain, for
giving me a place to learn and grow, and for all of the guidance and
encouragement. I would also like to thank my committee members, Dr. Medin
Bruening, Dr. Gavin Reid, and Dr. Keith Lookingland. All of their input and
support has been very valuable to me throughout this process.

To Dr. Kathy Severin and all of the fellow members of Women in
Chemistry, so fondly known as WiC - thanks for all the fun we had and all of the
things we accomplished. It was great to have an opportunity to participate in so
many outreach opportunities, and a good excuse to get out of the lab once and
awhile!

I need to thank all of my labmates, past (Prerna, Stotter, Grace, Jesse,
Gloria, Anne, Bennett, Yang, Audrey, Luther, Veronika, Jason T, and Bhavik) and
present (Wang, Hua, Yingrui, Matt Swope, Ayten, lsao, Martin, Dasa, Michael,
Lisa, Matt Smith, and Jia) for all of their support. For all of you who have left
before me I miss all of you and the guidance, discussions, and fun we had. For
the current members I wish you the best of luck as you go on to pursue your
dreams. The lab is a constantly changing environment as different people come
and go and all of you have added special memories to my time in graduate

school and I will not forget you.

And finally, I need to thank my friends and family for their encouragement.
For Mel, you’ve been here with me since the beginning and it has been wonderful
to share everything with such a great friend. To everyone else, Amber, Audrey,
Luther, Anne, Carl and JB, thanks for listening when I had silly questions and for
keeping me sane while I was trying to finish two degrees at once. I will miss all
of the times we shared and hope that we can continue to stay in touch since we
are now spread all over the country. To my parents for all of their love and
support and for teaching me a strong work ethic that has gotten me through
graduate school. To Scott, thank you for all of your love throughout all of this and
being willing to tolerate me coming home late day after day, I could not have

done this without you by my side.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................. x
LIST OF FIGURES .............................................................................................. xii
CHAPTER 1
INTRODUCTION .................................................................................................. 1
Background ..................................................................................................... 1
Anodic Stripping Voltammetry ......................................................................... 2
Working Electrodes ......................................................................................... 5
Diamond Electrodes ........................................................................................ 6
Dissertation Overview ................................................................................... 11
References .................................................................................................... 15
CHAPTER 2
EXPERIMENTAL METHODS ............................................................................. 18
Diamond Thin Film Deposition ...................................................................... 18
Material Characterization .............................................................................. 20
Electrochemical Characterization ............................................................ 20
Raman Spectroscopy ............................................................................... 22
Resistivity Measurements ........................................................................ 23
Differential Pulse Anodic Stripping Voltammetry ........................................... 23
Comparison to Hg Electrode ......................................................................... 24
Hg-Coated Electrode ............................................................................... 24
Electrochemical Measurements ............................................................... 24
Microgravity Modification Experiments .......................................................... 25
Electrochemical Measurements ............................................................... 25
X-ray Diffraction Spectroscopy ................................................................ 25
Potable and Technical Water Simulants .................................................. 25
Microgravity Cell ...................................................................................... 26
Urine Analysis ............................................................................................... 27
Electrochemical Measurements ............................................................... 27
Simulated Urine and Rat Urine Samples ................................................. 27
Blood Analysis ............................................................................................... 29
Electrochemical Measurements ............................................................... 29
Blood Sample .......................................................................................... 29
Nafion Coating ......................................................................................... 30
Atomic Force Microscopy ......................................................................... 30
Chemicals ..................................................................................................... 30
References .................................................................................................... 32

vii

CHAPTER 3
COMPARISON OF BORON DOPED DIAMOND AND Hg-GC WORKING

ELECTRODES ................................................................................................... 33
Introduction ................................................................................................... 33
Potential Window .......................................................................................... 34
Differential Pulse Anodic Stripping Voltammetry ........................................... 34
Detection Figures of Merit ............................................................................. 42
Comparison of Metal Deposition and Stripping Charge for BDD ................... 46
Conclusions ................................................................................................... 52
References .................................................................................................... 54

CHAPTER 4

METHOD ADJUSTMENTS FOR MICROGRAVITY APPLICATION ................... 56
Introduction ................................................................................................... 56
Solution Parameters ...................................................................................... 59

Effect of pH Adjustment ........................................................................... 59
Effect of Mineralization ............................................................................. 61
Effect of Dissolved Oxygen ...................................................................... 61
Development of a Sampling Procedure ......................................................... 67
Addition of the Supporting Electrolyte ...................................................... 67
Cell Design .............................................................................................. 69
Analysis of Simulated ISS Water ................................................................... 73
Long-Term Response Stability ...................................................................... 74
Conclusions ................................................................................................... 77
References .................................................................................................... 78

CHAPTER 5

ANALYSIS OF HEAW METAL IONS IN URINE ................................................ 79
Introduction ................................................................................................... 79
Deposition Parameters .................................................................................. 81
Detection Figures of Merit ............................................................................. 83
Standard Addition .......................................................................................... 85
Cadmium Complexation ................................................................................ 85
Real Sample Analysis ................................................................................... 89
Conclusions ................................................................................................... 91
References .................................................................................................... 92

CHAPTER 6

ANALYSIS OF HEAW METAL IONS IN BLOOD .............................................. 94
Introduction ................................................................................................... 94
Sample Preparation ...................................................................................... 95
Prevention of Fouling .................................................................................... 97

Nafion Coating ......................................................................................... 98
Sodium Dodecyl Sulfate ......................................................................... 105
Conclusions ................................................................................................. 107
References .................................................................................................. 109

viii

CHAPTER 7
CONCLUSIONS ............................................................................................... 112

Table 1.1.

Table 1.2.

Table 1.3.

Table 2.1.

Table 3.1.

Table 3.2.

Table 3.3.

Table 3.4.

Table 3.5.

Table 4.1 .

Table 4.2.

LIST OF TABLES

Comparison of the electrochemical properties of BDD and Hg
electrodes for use in anodic stripping voltammetry. ........................ 7

Analysis of NIST SRM 1640. Comparison of concentrations
determined by anodic stripping voltammetry (ASV) and atomic
absorption spectroscopy (AAS). The error reported for each
analytical method is compared to NIST certified values. ................. 9

Concentrations of Cd2+ and Pb2+ in the acid digest of the waste
treatment sludge as determined by ASV and AAS.31 The error is
the difference between the two techniques. .................................. 10

Components and concentrations in simulated urine ...................... 28

Stripping peak potentials and peak widths (FWHM) for BDD and
Hg-GC electrodes in the presence of individual metal ions at a 100
ppb concentration are displayed. Mean and standard deviation
values (n=3) are presented. .......................................................... 36

Detection figures of merit for metal ions measured using BDD and
Hg-GC electrodes. ........................................................................ 45

Deposition and stripping charges for individual metal ions at a
concentration of 1 ppm in 0.1 M acetate buffer, pH 5.2 using a BDD
electrode. ...................................................................................... 48

Deposition and stripping charges for 1 ppm Cu2+ in 0.1 M acetate
buffer, pH 5.2 using a BDD electrode, varying deposition
potential. ....................................................................................... 50

Deposition and stripping charges for 1 ppm Cu2+ in 0.1 M acetate
buffer, pH 5.2 using a BDD electrode, varying deposition time. 51

Comparison of the minimum concentrations of several priority metal
ions detectable with a diamond thin-film electrode at pH 4.5 and pH
5.2. ................................................................................................ 60

Comparison of the stripping response (peak area) for each priority
metal ion before and after the addition of Ca2+ and Mg2+ salts ...... 61

Table 4.3.

Table 4.4.

Table 4.5.

Table 5.1.

Table 5.2.

Table 5.3.

Table 6.1.

Table 6.2.

Powder X-ray diffraction data for reference materials for Pb,
Pb01,44, and PbOz. d is the lattice spacing, intensity indicates the
relative intensity of the peaks, and hkl is the crystallographic
orientation. The entries that are bolded and italicized are closest to
the peaks seen experimentally. ..................................................... 65

Comparison of X-ray counts for peaks from XRD spectra of four
diamond electrodes, bare, with Pb, Pb stripped - N2 purge, and Pb
stripped - no N2 purge. The - indicates no peak seen above the
background. N/A indicates this region was not scanned .............. 66

Results from the analysis of Ag+ in NASA simulated potable and
technical water samples. Each water sample was measured in
both the open cell as well as the microgravity cell. The error is the
standard deviation of concentration calculated by 3 runs. The %
error is calculated with respect to the specified value by NASA,
which was determined by ICP-MS. ............................................... 74

Reproducibility of Cd2+ signal on 3 BDD electrodes. 75 ppb Col2+ in
simulated urine (pH 8.8). ............................................................... 85

Formation constants for Cd2+ complexes.26 Listed as cumulative
formation constants where Kn is equal to k1k2....kn if k is the
stepwise formation constants. ....................................................... 86

Calculated concentration of free Cd2+ in the presence of complexing
agents: ammonia, citric acid, and lactic acid. ................................ 87

Minimum concentration of sz+ in blood when different health
effects are observed ...................................................................... 94

Comparison of three electrodes with Nafion coating. Each
electrode was used to analyze blood with 445 ppb Pb”. Two trials
were performed on each film with a short cleaning step in 0.3 M
HCI between. The exposure time was 15 minutes after which a
deposition time of 180 s was used prior to ASV analysis. ........... 101

xi

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 2.1.

Figure 2.2.

Figure 3.1.

Figure 3.2.

LIST OF FIGURES

Differential pulse anodic stripping voltammetry (DPASV) waveform
(top), and voltammetric i-E curve (bottom) for a solution containing
two metals. The deposition step (1) followed by the stripping step
(2). The current is sampled at two points noted by A, the current
difference (i2 - i1) is plotted versus potential. .................................. 4

Anodic stripping voltammetric i-E curves for a standard reference
solution obtained from NIST (SRM - NIST 1640). Voltammetric
curves for standard solutions, in acetate buffer, pH 4.5, of equal
concentrations of containing Pb”, Cu”, and Ag+ are also shown.
Deposition was for 180 s at -0.9 V vs. AglAgCI. Figure reproduced
with permission from Elsevier .......................................................... 8

Anodic stripping voltammetric i-E curve for sludge, overlaid with
responses for standard solutions, in acetate buffer, pH 4.5,
containing Cd” and Pb” in equal concentrations ranging from 50
to 1000 ppb. The deposition time was 180 s at -1.0 V vs. AglAgCI.
Image reprinted with pemlission from Elsevier .............................. 10

Schematic of microwave assisted chemical vapor deposition
reactor used for diamond electrode growth. .................................. 19

Diagram of electrochemical cell. ................................................... 21

Overlaid individual stripping voltammetric i-E curves for 100 ppb
solutions of each metal ion in 0.1 M acetate buffer, pH 5.2, 30 ppm
Ca2+, and 5 ppm Mg”. The tap set of voltammetric i-E curves were
obtained with Hg-GC and the bottom set with BDD. The curves are
offset for clarity. All curves are uncorrected for the background
current except the one for Cu on Hg-GC. Figure reprinted with
permission from Elsevier. .............................................................. 35

Differential pulse stripping voltammetric i-E curves for 100 ppb Cu2+
in 0.1 M acetate buffer, pH 5.2. 30 ppm Ca2+ and 5 ppm Mg2+ were
added. The background voltammetric i-E curves are also shown for
comparison. The deposition potential was -400 mV vs. AglAgCl
and the deposition time was 210 s. The working electrodes were
BDD (A) and Hg-GC (B). Note the difference in current scales.
Figure reprinted with permission from Elsevier. ............................ 38

xii

Figure 3.3.

Figure 3.4.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

Figure 4.6.

Figure 4.7.

Differential pulse stripping voltammetric i—E curves for 100 ppb Cd2+
in 0.1 M acetate buffer, pH 5.2, 30 ppm Ca2+, and 5 ppm Mg”. The
deposition potential was -400 mV vs. AglAgCl and the deposition
time was 210 s. The working electrodes were BDD (A) and Hg-GC
(B). Note the difference in current scales. .................................... 41

Differential pulse stripping voltammetric i-E curves for standard
solutions of Cd2+ in 0.1 M acetate buffer, pH 5.2, 30 ppm Ca2+, and
5 ppm Mg”. The deposition potential was -1ooo mV vs. Ag/AgCl
and the deposition time was 210 s. The working electrodes were
BDD (A) and Hg-GC (B). Note the difference in current scales.
Figure reprinted with permission from Elsevier. ............................ 43

Stripping voltammetric i-E curves for 100 ppb Ag+ (A) and 100 ppb
Pb2+ (B) in acetate buffer, pH 5.2, with and without a N2 purge.
Image adapted from McGaw, et. al. .............................................. 63

Powder XRD spectra taken of a bare diamond electrode (gray) and
a diamond electrode with Pb electrochemically deposited on the
surface (black). Peaks due to the diamond are marked with BDD
and peaks due to Pb are also marked accordingly. The BDD signal
is multiplied by 5 and spectra are offset slightly for clarity ............. 64

Anodic stripping voltammetric i-E curves for Pb were recorded in a
solution containing, 100 ppb Pb2+ in acetate buffer, pH 5.2, with 30
ppm Ca2+ and 5 ppm Mg”. (A) no added ascorbic acid and (B) 0.1
M ascorbic acid. ............................................................................ 69

A diagram of the sealed electrochemical cell designed for use in
microgravity. The body is made of Teflon. Three Iuer fittings with
valves serve as inlets and outlets. The planar working electrode is
held in place by a copper cylinder inside a Teflon screw. The
working electrode area is defined by a Viton o-ring (0.32 cm2). 71

Anodic stripping voltammetric i-E curves for Ag in the microgravity
prototype cell. Measurements were made using Ag" solutions
ranging in concentration from 25 - 500 ppb. Buffer concentrate
was used to add electrolyte, the concentration in the final solution
was 0.1 M acetate, pH of 5.2. The deposition time was 3 min. 72

Stability test of a 400 ppb Ag+ solution using a single diamond
electrode. The electrode was stored wet, under deionized water,
between measurements. ............................................................... 75

Stability test of a 400 ppb Ag+ solution using a single diamond
electrode. The electrode stored dry between measurements. ..... 76

xiii

Figure 5.1.

Figure 5.2.

Figure 5.3.

Figure 5.4.

Figure 5.5.

Figure 6.1.

Figure 6.2.

Figure 6.3.

Peak current as a function of deposition time for 250 ppb Cd” in
urine simulant (pH 8.8). Deposition time varied between 2 s and 5
minutes. A) no convection method used, B) N2 bubbling used for
convection. Each point is an average of 3 measurements. .......... 82

Anodic stripping curves (left) and calibration curve (right) for urine
simulant (pH 8.8) with Cd” concentrations in the range of 10 to 250
ppb. Deposition occurred at -1.0 V vs. AglAgCl for 3003 with
convection (N2 bubbling) Error bars on calibration curve represent
standard deviation. ........................................................................ 84

Anodic stripping voltammetric i-E curves for Cd in an acetate buffer,
pH 8.8, containing 0.1M ammonia, 3mM citric acid, or 4mM lactic
acid. 30 ppb Cd” deposited for 180 s with convection (N2
bubbling). ...................................................................................... 88

Anodic stripping voltammetric i-E curves in rat urine. Bottom curve
is rat urine alone and aliquots of Cd” were added to increase the
Cd” concentration to the range of 90 -300 ppb. Curves are offset
for clarity. ...................................................................................... 90

Stability test performed in 300 ppb Cd” in real rat urine.
Measurements are sequential runs taken over a time period of
approximately 6 hours. Deposition conditions, 300 s with
convection (N2 bubbling). .............................................................. 90

Anodic stripping voltammetric i-E curves for BDD in 0.1 M acetate
buffer, pH 5.2, and 0.1 M NaCl, with 100 ppb Pb” (top) and without
Pb” (bottom). The deposition time was 180 s. ............................ 96

Anodic stripping voltammetric i-E curves of 450 ppb Pb” in HCI (1
part Pb” solution to 5 parts 0.3 M HCI), before exposure to bovine
blood (a), immediately after exposure to bovine blood (b), after a 24
h soak in distilled water (c), and after an additional 20 min soak in
isopropanol (d). The deposition time was 180 s. .......................... 98

Pb stripping current for Pb” in bovine blood as a function of the
preconcentration time. The blood was mixed 1:5 with 0.3 M HCI.
The BDD electrode was coated with a Nafion film. The time axis
indicates the amount of time blood solution was in contact with
Nafion coated electrode before deposition began. The deposition
time was 180 s. ........................................................................... 100

xiv

Figure 6.4.

Figure 6.5.

Figure 6.6.

Stripping voltammetric i-E curves of Pb” in blood at varying
concentrations. The bovine blood was mixed 1:5 with 0.3 M HCI.
The electrode exposure time was 30 min. The deposition time was
180 s followed by ASV analysis. Curves are offset for clarity ..... 103

AFM contour plots of three BDD electrodes, taken in contact mode.
The electrodes were coated in Nafion, which was removed with
acetone (A), ethanol (B), and acid washing followed by
rehydrogenation (C). ................................................................... 105

Stripping voltammetric i-E curves of 450 ppb Pb” solution mixed
1:5 with 0.3 M HCI. The solutions used before and after BSA
exposure (black and grey solid lines) only contain Pb” and HCI.
The other sample (dashed line) also contained 50 mglL BSA and 3
mM SDS. The deposition time was 180 s. ................................. 107

XV

CHAPTER 1
INTRODUCTION
1.1. Background

Monitoring heavy metal ion levels is essential for human health and safety.
There are numerous health problems associated with exposure to high levels of
metal ions (e.g., Cd”, Pb”, Hg”, As”’5+) because of their tendency to
accumulate in the body and their low rate of clearance. For instance, the
biological half-life of Cd is 10 to 30 years and the half-life of Pb in bone is more
than 20 years.1 There are many routes for contaminant exposure: water, food
supply, air, and environment/occupation exposure (e.g., steel production). The
Environmental Protection Agency (EPA) estimates that nearly twenty percent of
human exposure to Pb is through contaminated drinking water.2 It is, therefore,
critical for humans to experience minimal exposure to these contaminants and
this can be accomplished through effective water quality monitoring.

Regulatory agencies, such as the EPA, have established maximum
allowable contaminant levels in drinking water to ensure public health. These
standards are meaningless, though, unless analytical methods exist that can
monitor these contaminants with adequate sensitivity, selectivity, and
reproducibility. The present EPA recommended methods for metal ion analysis
in water are atomic absorption spectroscopy (AAS), inductively couple plasma
mass spectrometry (ICP-MS), and anodic stripping voltammetry (ASV). ASV is
commonly employed because of its wide linear dynamic range, low detection limit

(ppb), and multielement analysis capability. An additional advantage of ASV,

over AAS or lCP-MS, is the simplicity of the instrumentation, which is relatively
inexpensive, small in size, requires low electrical power, and is portable enabling
its field deployment.

In addition to environmental monitoring there is also need for analysis of
clinical samples (e.g., blood, urine) that may contain heavy metals. Since the
metals are toxic to humans the concentration of metal determined in these fluids
can help diagnose clinical symptoms. In some cases, such as symptomatic
patients with high levels of Pb”, chelation therapy can be effective.1 Chelation
therapy involves administering a chelating agent, with ethylenediamine
tetraacetic acid (EDTA) being the most common. These agents complex the
toxic metal ion and aid in its removal from the body. Analysis of the metal ions in
body fluids can also be useful for monitoring the effectiveness of this treatment.

Currently, the most common method for analysis of metal ions in body
fluids is AAS, typically using a graphite furnace to atomize the samples for
detection.”6 The use of ASV for clinical samples also has some advantages,
particularly the portability and low cost of the instrumentation that would make it
practical for use in a doctor’s office or at bedside. In fact, there is a commercially
available Pb” analysis system marketed by BAS, Inc. (LeadCare) that is based
on ASV.

1.2. Anodic Stripping Voltammetry

ASV is an electroanalytical technique in which the current at the working

electrode is measured as a function of a potential waveform applied to the

working electrode. The working electrode is the electrode at which the

electrochemical reaction of interest takes place. The applied potential controls
the reducing or oxidizing strength of the electrode.

Stripping voltammetry gains advantages in sensitivity and detection limit
because of an initial preconcentration step. The deposition potential is chosen to
be more negative than the standard reduction potential (E0) of the analyte in
order to reduce and deposit the metal ion on the surface. The deposition can
vary from seconds to minutes.

Once the analyte has been preconcentrated/deposited, it is then stripped
off by applying a positive-going (anodic) potential. The current is recorded during
this step, and can provide both qualitative and quantitative information. The
oxidation of the metal will occur near the reversible Nemstain potential (E0) for
the metal, so the potential at which the highest current is obtained will serve to
identify the metal present. The magnitude of this current can be used to plot
calibration curves and to determine the metal ion concentration in the bulk
solution. An example of the voltammetric i-E curve obtained for a solution
containing two metal ions is shown in Figure 1.1; the dashed lines demonstrate
the correlation of the potential waveform (top) and the voltammetric i-E curve
(bottom).

There are several variations on the simple case of linear scan stripping
voltammetry that change the way the positive potential sweep is carried out in
order to reduce the contribution of the background charging current. In linear
sweep voltammetry the total current is recorded and is a combination of

background current and faradaic current. However, only the faradaic current

reflects the oxidation of the metal. The charging current is created by changing
the potential across the electrode-solution interface, which acts as a capacitor.
When a potential is applied to the working electrode the charge must be
balanced on the solution side with an array of ions and oriented dipoles. This
separation of charge at the electrode-solution interface creates a capacitor;
potential changes create a current (charging current) that flows to charge and

discharge this capacitor.

(1)

 

I Potential (V) 4;

 

 

 

 

  

 

:2
EV
03A
go<
3c:
mv
0&3
PE
0

 

Potential (V)

Figure 1.1. Differential pulse anodic stripping voltammetry (DPASV) waveform
(top), and voltammetric i-E curve (bottom) for a solution containing two metals.
The deposition step (1) followed by the stripping step (2). The current is sampled
at two points noted by A, the current difference (i2 — i1) is plotted versus
potential.

Differential pulse anodic stripping voltammetry (DPASV) is a common
technique used to minimize the contribution of the charging current. The
potential waveform is shown in Figure 1.1. In this technique, pulses of equal
amplitude are superimposed upon the linear scan. The current is sampled at two

points, one just before the pulse is applied and one just before the pulse is

finished, the former is subtracted from the latter and this difference is plotted
versus the potential. This takes advantage of the different decay rates of
faradaic and charging currents, since charging currents decay exponentially with
time and faradaic currents only decay with t'1’2. The pulse width is such that
when the current is sampled it will be mostly faradaic. It is assumed that the
charging current at both sampling points will be approximately the same so any
differences can be attributed to faradaic current.
1.3. Working Electrodes

The choice of working electrode is critical for high quality ASV
measurements. Many different factors can influence the performance of a
working electrode. The first ASV studies were performed using Pt; however, it
was quickly realized that Hg had superior properties. These pr0perties include a
large negative potential window, reproducible response, and good sensitivity.7
Hg is used in several forms. Initially, an Hg-coated Pt electrode was used, which
then evolved into the Hg pool electrode, and then to a hanging drop.”13 In 1970,
Florence proposed co-deposition of Hg and the analyte metals onto a glassy
carbon surface in the form of a thin film.14 This thin layer Hg electrode exhibited
increased peak resolution as compared to the hanging-drop or even thicker films
of Hg. This is still one of the more commonly used Hg electrode preparation
methods for ASV because it is simple and leads to reproducible results.

Hg is unique when compared to other electrodes because it forms an
amalgam with the analyte metals. This reduces surface interactions and

interrnetallic compound formation, both of which can cause distortion in the shape

of the stripping voltammetric i-E curves. Even with these attributes, however,
alternate electrodes are desired because of Hg’s toxicity and volatility.

There are some specific interactions of electrodes with analytes that can
enhance the signal. For example, an interaction between As” and Au makes it
the best choice for analysis of As”.15 The other consideration is the matrix
components and the effect they may have on the working electrode. For
instance, CI' interacts with a Hg surface forming calomel (HgCl) which will affect
the electrochemical response. Therefore, Hg is not a good choice for use in a
matrix containing CI'.

Several alternate electrodes to Hg have been investigated, such as lr,16'17
Au,18'22 Ag,” 24 and boron-doped diamond (BDD).”’36 One alternative that has
shown promising behavior for stripping analysis is Bi-modified electrodes, as

37“" and others.42 Bismuth can be co-deposited with

studied by the Wang group
the contaminant metals, similar to the preparation of a Hg-film electrode, and this
electrode provides detection figures of merit that are comparable to those for Hg.
A limiting factor with Bi, however, is the low anodic potential limit.
1.4. Diamond Electrodes

Diamond is also an attractive alternative with properties similar to those of
Hg. The comparison of these properties is presented in Table 1.1. BDD has a
wider potential window than Hg, and has been found to yield good detection
figures of merit for many metal ions. One of the very first metal ions investigated

with BDD using ASV was Pb”. Fujishima and co-workers demonstrated a low

ppb limit of detection on a bare diamond using a preconcentration time of 15

min.26 The more positive potential limit of BDD has been exploited in detection

of Hg” with detection limits in the low ppb range for short preconcentration times
(30 s), and in the mid-ppt range for longer deposition times (20 min)” 32 The
Compton group has investigated the enhancing effects of ultrasound on the
stripping current, and showed detection limits with BDD in the high ppt range for
Ag‘ and Cd”.”° 3‘ The stripping current can also be enhanced with microwaves
focused in the vicinity of the BDD surface. This lowered the detection limit for
Pb” by an order of magnitude, likely due to enhanced convective mass transport
from the temperature gradients created.28

Table 1.1. Comparison of the electrochemical properties of BDD and Hg
electrodes for use in anodic stripping voltammetry.

 

 

 

 

 

 

 

 

 

Boron Doped Diamond (BDD) Hg
Wide cathodic and anodic potential limits Wide cathodic limit
Lower background current Higher background current
Good sensitivity Good sensitivity
Chemically inert Interaction with Cl'
Reusable surface — no pretreatment required Easily refreshed surface
Non-toxic Toxic
Non-volatile Volatile

 

 

 

Early work from our group showed the usefulness of BDD for real sample
analysis, such as river, and tap waters, as well as digests of waste treatment
sludge, and soils.31 The accuracy of the method using this electrode was
demonstrated using a standard reference material (SRM) obtained from the
National Institute for Standards and Technology (NIST). The anodic stripping
voltammetric i-E curve recorded for the SRM sample is shown in Figure 1.2.

Comparison curves for solutions (acetate buffer, pH 4.5) containing Pb”, Cu”,

and Ag+ in equal concentrations ranging from 10 — 90 ppb are also shown. The
metal ion concentrations in the SRM were determined using the standard addition
method and the resulting values are shown in Table 1.2. The solution was also
analyzed by flame atomic absorption spectroscopy (AAS), and the results are

also shown in the table.

 

10 - — Standard Solution
- - SRM 1640 (NIST)

 

 

 

8 q CU”

Current (pA)

 

 

 

 

-0.6 -0.4 -0.2 0 0.2 0.4 0.6
Potential vs. AglAgCl (V)

Figure 1.2. Anodic stripping voltammetric i-E curves for a standard reference
solution obtained from NIST (SRM - NIST 1640). Voltammetric curves for
standard solutions, in acetate buffer, pH 4.5, of equal concentrations of
containing Pb”, Cu”, and Ag+ are also shown. Deposition was for 180 s at -O.9
V vs. AglAgCI. Figure reproduced with permission from Elsevier.31

The values determine by ASV compare favorably to both the NIST certified
values (error s 10%) as well as to AAS values, the latter is a well established
technique for metal ion analysis. This result demonstrates the capability of

multimetal ion analysis with BDD using ASV.

Table 1.2. Analysis of NIST SRM 1640. Comparison of concentrations
determined by anodic stripping voltammetry (ASV) and atomic absorption
spectroscopy (AAS). The error reported for each analytical method is compared
to NIST certified values.31

 

 

 

 

 

 

 

 

 

 

 

Pb” Cu” Ag“
NIST: 27.89 :I: 0.14 NIST: 85.2 i 1.2 NIST: 7.62 :l: 0.25
Concentration Error Concentration Error Concentration Error
(ppb) (ppb) (ppb)
ASV 30.2 1 0.8 8% 77 i 3 10% 7.3 t 0.6 4%
AAS 25 i 2 10% 89.3 t 0.6 5% 8.0 :I: 0.7 5%

 

 

In addition to the aqueous samples some additional solid samples were
tested: soil and sludge. The sludge was obtained from a wastewater treatment
facility, and heavy metals and other unwanted materials (e.g., organic
components not easily biodegradable) collect in this material. In order to analyze
the sludge, and similarly the soil, an acid digestion (concentrated HNO3) was
used to solubilize the metals. Two metals were detected in the sludge, Cd and
Pb, and were quantified using standard addition. The stripping voltammetric i-E
curve obtained for the sludge digest is displayed in Figure 1.3 and is compared to
standard solutions of Cd” and Pb”.

Concentrations of metal in the acid digest of the sludge sample were
determined by AAS well, and the results obtained by ASV are compared to the
AAS results in Table 1.3. Again the results from ASV on BDD compare favorably

with the established technique, AAS.

120-

 

— Standards
-- Sludge

1 00 T Pb2+

 

 

 

80 . 1000 ppb

I

Current (uA)

 

 

 

 

 

-1 -0.8 0.6 -0.4 -0.2
Potential vs. Ag/AgCl (V)

Figure 1.3. Anodic stripping voltammetric i-E curve for sludge, overlaid with
responses for standard solutions, in acetate buffer, pH 4.5, containing Cd” and
Pb * in equal concentrations ranging from 50 to 1000 ppb. The deposition time

was 180 s at -1.0 V vs. AglAgCI. Image reprinted with permission from
Elsevier.31

Table 1.3. Concentrations of Cd” and Pb” in the acid digest of the waste
treatment sludge as determined by ASV and AAS,:31 The error is the difference
between the two techniques.

 

 

 

 

 

ASV (ppb) AAS (ppm Error
Cd” 156 :l: 5 145 :l: 3 7.5%
Pb” 124:5 134:2 8.1%

 

 

 

 

Other groups have also verified results from ASV analysis on BDD with
more traditional techniques (e.g., ICP-AES, AAS) with favorable results.“ 33 For
example, Babyak and Smart compared concentrations of Cd” determined for a

sample of spiked river water using ASV on BDD and lCP-AES (atomic emission

1O

spectroscopy) and found there was an error of 8.7% between the two techniques
(59.8 ppb and 55 ppb determined with ASV and lCP-AES, respectively).33
Overall BDD has been successfully used for ASV analysis and is a potential
alternative to the Hg-based electrodes.

1.5. Dissertatlon Overview

Earlier work has shown that BDD is a suitable electrode for analysis of a
variety of samples including natural waters, soils, and sludge using ASV. The
focus of this project was to further explore the properties of BDD for ASV. While
BDD has been shown to be useful for ASV analysis there has not been any direct
comparison of the BDD and the Hg-based electrodes traditionally used. This
study serves to investigate the properties of BDD including analytical figures of
merit, manner of metal deposition, stripping peak potentials, and shapes as
compared to a Hg-coated glassy carbon electrode. BDD has advantages over
Hg, particularly the non-toxic, non-volatile nature of the electrode. Since BDD
has already been used for ASV analysis the goal was to determine the similarities
and differences in the electrode performance with respect to Hg and the utility as
an alternate electrode.

One particular application of BDD that was of interest was to develop a
method for heavy metal analysis in microgravity. On-board the lntemational
Space Station (ISS) Ag+ is used as a biocide in the water supply, and a method
was needed to analyze the metal ion content (both Ag“ as well as potential
contaminants) continuously or semi-continuously to maintain a safe water

supply.” “3‘45 ASV was the method chosen for analysis because of its portability

11

and low power requirements, but because of the toxicity and volatility of Hg the
use of a Hg electrode in microgravity is not possible. The properties of BDD
(e.g., non-toxic, non-volatile) and the comparable electrochemical response to Hg
electrodes make BDD a good alternate electrode for this application. In
developing a method for on-board use several issues had to be addressed
including addition of electrolyte, constraints of microgravity, and long-term
stability of the method. Because the water samples of interest were potable and
technical (used for toilet flushing, hygiene, etc.) waters, they contained minimal
amounts of electrolyte and the pH could vary. There were also many challenges
encountered to modify the method for microgravity: all of the liquid had to be
enclosed, the pH needed to be between 5.5 and 9, and deoxygenation using N2
sparging was not possible. Addressing the needs of an assay in microgravity and
determining the performance of the BDD electrode was the main focus of this
portion of the work.

BDD has been successfully applied to environmental and water samples.
The hydrophobic nature of the electrode surface tends to make this electrode
fouling resistant which could be very beneficial to use in a more complex matrix
than a water sample or a soil digest sample. Biological samples (e.g., blood,
urine) are particularly troublesome for electrochemical methods because of the
number of components (e.g., proteins, biomolecules) that can foul the electrode
surface; therefore, these samples were chosen to probe the fouling resistance of

BDD.

12

A simulated urine sample was used initially to explore the stability and
anti-fouling nature of the BDD surface. In urine samples, the concentration of
metal ions is low (high ppt to low ppb), and therefore, the deposition conditions
were optimized in order to obtain the lowest detection limits. The analytical
figures of merit were measured to gauge the performance of the BDD electrode.
A sample of rat urine was also tested on the BDD surface and a series of
measurements were made over the course of several hours in order to determine
the stability of the BDD response in the urine matrix.

Blood was the other biological fluid that was tested on the BDD electrodes.
Blood is a much more complex matrix than urine, containing a much higher
concentration of proteins as compared to urine. In metal analysis the blood
needs to be analyzed whole (i.e., not plasma) because the metal ions are bound
to the red blood cells. Even though BDD has a fouling resistant surface it is still
not able to be used directly in a blood solution. Therefore, two options were
explored to protect the electrode from fouling. A polymer coating (Nafion) was
used to physically block the proteins from reaching the electrode surface, and a
solution based complexing agent (sodium dodecyl sulfate) for the proteins was
used to prevent fouling. The effectiveness of each of these methods was
explored.

The aims of this project were to further explore the properties of BDD for
ASV analysis, specifically to: (i) compare the electrochemical properties and
response to a traditional Hg electrode, (ii) explore the utility in microgravity

applications, (iii) determine the usefulness of BDD for more complex matrices

13

(e.g., urine, blood), and (iv) to explore potential fouling prevention methods (e.g.,
coatings, complexing agents) to use BDD in solutions containing high protein

concentrations (e.g., whole blood).

14

1 .6. References

(1)

(2)

(3)
(4)

(5)
(6)

(7)
(8)
(9)
(10)
(1 1)

(12)
(13)
(14)

(15)
(15)
(17)

(13)
(19)

Goyer, R. A. In Casarett & Doull's Toxicology: The Basic Science of
Poisons, 5th ed.; Klaassen, c. 0., Ed.; McGraw-Hill: New York, 1996, pp
691-736.

Drinking Water Standards. http://www.epa.gov/safewater/standards.html
(accessed May 2007)

Sharma, R. P.; McKenzie, J. M.; T., K. J. Anal. Tax. 1982, 6, 135-138.

Plotkin, E.; McAleer, J. F.; Cordeiro, M. L.; Ackland, M. R.; Sheehan, T.
M.; Braithwaite, R. A. Clin. Chem. 1997, 43, 2187-2189.

Yassin, A. S.; Martonik, J. F. J. Occup. Environ. Hyg. 2004, 1, 324-333.
Castro Maciel, C. J.; Miranda, G. M.; Oliveria, D. P.; Siqueria, M. E. P. B.;
Silveria, J. N.; Leite, E. M. A.; Silva, J. B. B. Anal. Chim. Acta 2003, 491,
231-237.

Lord, S. S.; O'Neill, R. 0.; Rogers, L. B. Anal. Chem. 1952, 24, 209-213.
Gardiner, K. W.; Rogers, L. B. Anal. Chem. 1953, 25, 1393-1397.

Porter II, J. T.; Cooke, W. D. J. Am. Chem. Soc. 1955, 77, 1481-1486.
Ross, J. W.; DeMars, R. D.; Shain, l. Anal. Chem. 1956, 28, 1768-1771.

Mamantov, G.; Papoff, P.; Delahay, P. J. Am. Chem. Soc. 1957, 79, 4034-
4040.

Nikelly, J. G.; Cooke, W. D. Anal. Chem. 1957, 29, 933-939.
DeMars, R. D.; Shain, I. Anal. Chem. 1957, 29, 1825-1827.

Florence, T. M. J. Electroanal. Chem. Interfacial Electrochem. 1970, 27,
272-281 .

Song, Y.; Swain, G. M. Anal. Chem. 2007, 79, 2412-2420.
Nolan, M. A. Anal. Chem. 1999, 71, 3567-3573.

Herdan, J.; Feeney, R.; Kounaves, S. P.; Flannery, A. F.; Storment, C. W.;
Kovacs, G. T. A.; Darling, R. B. Environ. Sci. Technol. 1998, 32, 131-136.

Andrews, R. W.; Johnson, D. C. Anal. Chem. 1975, 47, 294-299.
Wang, J.; Tian, B. Anal. Chem. 1993, 65, 1529-1532.

15

(20)

(21)

(22)
(23)

(24)
(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)
(34)

(35)
(36)

(37)

Vlltchinskaia, E. A.; Zeigman, L. L.; Morton, S. G. Electroanal. 1995, 7,
264-269.

Bonfil, Y.; Brand, M.; Kirowa-Eisner, E. Anal. Chim. Acta 1999, 387, 85-
95.

Feeney, R.; Kounaves, S. P. Anal. Chem. 2000, 72, 2222-2228.

Kirowa-Eisner, E.; Brand, M.; Tzur, D. Anal. Chim. Acta 1999, 385, 325-
335.

Bonfil, Y.; Kirowa-Eisner, E. Anal. Chim. Acta 2002, 457, 285-296.

Saterlay, A. J.; Foord, J. S.; Compton, R. G. Analyst 1999, 124, 1791-
1796.

Manivannan, A.; Tyrk, D. A.; Fujishima, A. Electrochem. Solid-State Left.
1999, 2, 455-456.

Saterlay, A. J.; Marken, F.; Foord, J. S.; Compton, R. G. Ta/anta 2000, 53,
403-415.

Tsai, Y.-C.; Coles, B. A.; Holt, K.; Foord, J. S.; Marken, F.; Compton, R. G.
Electroanal. 2001, 13, 831-835.

Manivannan, A.; Seehera, M. 8.; Tryk, D. A.; Fujishima, A. Anal. Lett.
2002, 35, 355-368.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chem. Mater. 2003,
15, 879.

Sonthalia, P.; McGaw, E.; Swain, G. M. Anal. Chim. Acta 2004, 522, 35-
44.

Manivannan, A.; Seehra, M. S.; Fujishima, A. Fuel Process. Technol.
2004, 85, 513-519.

Babyak, 0.; Smart, R. B. Electroanal. 2004, 16, 175-182.

Banks, C. E.; Hyde, M. E.; Tomcik, P.; Jacobs, R.; Compton, R. G.
Talanta 2004, 62, 279-286.

Wantz, F.; Banks, C. E.; Compton, R. G. Electroanal. 2005, 17, 655-661.

McGaw, E.; Sonthalia, P.; Swain, G. M. SAE Technical Paper 2005-01-
2888 35th lntemational Conference on Environmental Systems 2005.

Wang, J.; Lu, J.; Hocevar, S. B.; Farias, P. A. M.; Ogorevc, B. Anal.
Chem. 2000, 72, 3218-3222.

16

(38)
(39)

(40)

(41)

(42)

(43)

(44)

(45)

Wang, J.; Lu, J.; Hocevar, S. B.; Ogorevc, B. Electroanal. 2001, 13, 13-16.

Wang, J.; Lu, J.; Kirgoz, U. A.; Hocevar, S. B.; Ogorevc, B. Anal. Chim.
Acta 2001, 434, 29-34.

Flechsig, G.-U.; Korbout, O.; Hocevar, S. B.; Thongngamdee, S.;
Ogorevc, B.; Grundler, P.; Wang, J. Electroanal. 2002, 14, 192-196.

Pauliukaite, R.; Hocevar, S. B.; Ogorevc, B.; Wang, J. Electroanal. 2004,
16, 719-723.

Kruusma, J.; Banks, C. E.; Compton, R. G. Anal. Bioanal. Chem. 2004,
379, 700-706.

Plumlee, D. K.; Mudgett, P. D.; Schultz, J. R. SAE Technical Paper 2002-
01-2537 32’” lntemational Conference on Environmental Systems 2002.

Mudgett, P. D.; Benoit, M. J.; Schultz, J. R. SAE Technical Paper 2002-
01-2532 32’“ lntemational Conference on Environmental Systems 2002.

Plumlee, D. K.; Mudgett, P. D.; Schultz, J. R. SAE Technical Paper 2003-
01-2401 33’” lntemational Conference on Environmental Systems 2003.

17

CHAPTER 2
EXPERIMENTAL METHODS
2.1. Diamond Thin Film Deposition

Nanocrystalline diamond thin-film electrodes were used in all of the
studies detailed in this work. The nanocrystalline diamond was deposited on
highly conductive p-type Si (~10'3 n-cm) (Virginia Semiconductor, Inc.) by
microwave-assisted chemical vapor deposition (CVD) (1.5 kW, 2.54 GHz, Model
AX3120, Astex, lnc., Lowell, MA).1 A schematic of this reactor is shown in Figure
2.1 . The Si substrate was first prepared by scratching the surface with a 0-2 pm
diameter diamond powder (GE Superabrasives, Worthington, OH) in an ultrapure
water slurry. After scratching the substrates were rinsed with ultrapure water and
isopropanol. The surface was further cleaned by sonication (5 min each) in
acetone, isopropanol, and water, sequentially. The clean substrates were dried
and examined under a polarized light microscope (Olympus BX60M, Olympus
America, Inc.) to check for surface cleanliness. If polishing debris remained on
the surface the substrates were resonicated until it was removed. The
preparation by scratching the surface was necessary to create defects in the Si
surface and embed small diamond particles to serve as nucleation sites for
diamond growth. It was important that a uniform density of scratches was
present on the surface to obtain uniform diamond film growth.

Once the substrates were clean they were placed on a sample stage
made up of layers starting with a stainless steel plate, Mo, W, quartz, and finally

the Si substrate. The sample stage was loaded into the reactor chamber,

18

consisting of a quartz bell jar that could be evacuated of atmospheric gases using
a rotary pump. The chamber was allowed to pump down for at least 2 hours and
a pressure of < 20 Torr was obtained. Following evacuation of atmospheric

gasses the source gases were introduced into the chamber.

|mlcrowave

generator

 

 

 

 

 

 

 

 

quartz
chamber

 

.... substrate

 

 

 

 

 

rotary pump

Figure 2.1. Schematic of microwave assisted chemical vapor deposition reactor
used for diamond electrode growth.

The source gas mixture, containing ultrahigh-purity methane, hydrogen,
and argon, entered the reactor at flow rates of 1, 5, and 94 sccm, respectively.
All gases were ultrahigh purity (99.999%) grade (methane and hydrogen from
AGA Specialty Gas, Cleveland, OH; argon from BOC Group, Inc., Murray Hill,
NJ). Deposition was accomplished using a power of 800 W, a system pressure

of 140 Torr, a substrate temperature estimated (optical pyrometry) to be between

19

700 and 800 ‘C, and a deposition time of 2 h. Each film was boron-doped during
deposition by adding 10 ppm diborane (0.1% B2H6 diluted in H2) to the source
gas mixture. At the end of the growth period, the film remained exposed to
argon-hydrogen plasma for 10 min followed by a gradual reduction in power and
pressure to slowly cool the film in the presence of atomic hydrogen. Post-growth
annealing in atomic hydrogen is essential for removal of any adventitious non-
diamond carbon and assurance of full hydrogen termination. After deposition, the
film was cleaned by chemical oxidation, exposing the film to warm aqua-regia
then to warm hydrogen peroxide (30%), each for 30 min. The film was then
rehydrogenated in a hydrogen plasma (200 sccm H2, 35 Torr, 1000 W) for 30
min, and cooled in the same manner as described for the growth procedure.2
The resulting film thickness was approximately 5 pm, as determined by cross
sectional scanning electron microscopy imaging. Hall effect measurements of
other films grown in a similar manner revealed a carrier concentration (holes) in
the low 1020 cm'3 range and a carrier mobility of 0.1 to 1 cmzN-s. The relatively
low mobility (single crystal diamond ~3000 cmzN-s) results from electron

scattering by the high fraction of grain boundary.

2.2. Material Characterization
2.2.1. Electrochemical Characterization

A single-compartment glass cell (~5 mL volume), shown in Figure 2.2, was
used for the electrochemical measurements along with a computer-controlled
potentiostat (Model 6508, 850, or 900 CH Instruments Inc., Austin, TX).2 The

electrochemical cell was housed inside a grounded Faraday cage for electrical

20

shielding. The working electrode was clamped against a Viton o-ring at the
bottom of the cell to confine the solution. Prior to use, a small amount of carbon
from a graphite rod was rubbed on the backside of the scratched and cleaned Si
substrate to ensure good ohmic contact with the copper current-collector plate.
The geometric area of the electrode exposed to the solution was 0.2 cm2, unless
otherwise noted. A large area carbon rod served as the counter electrode, and
was placed normal to the working electrode. A Ag/AgCl reference electrode
(saturated KCI, E0 = -45 mV vs. SCE) was placed inside a cracked-glass capillary

(double junction) and filled with the supporting electrolyte.

Auxiliary (C rod)

   
  
  

Reference N2 inlet

Viton o-ring (0.2 cm?)

Current Collector\ Working Electrode

(copper)

Figure 2.2. Diagram of electrochemical cell.

Prior to their use for stripping voltammetry experiments initial
characterization experiments were performed on all new diamond electrodes to
ensure they had similar electrochemical properties. Background cyclic

voltammograms were measured in 0.1 M perchloric acid solution. The potential

21

was swept between approximately -1.5 V and 1.5 V vs. AglAgCl to determine the
potential window. The potential window was defined as the region between :I: 50
uA. The electrodes used had potential windows that spanned approximately 3 V.
Cyclic voltammograms were also measured in 0.1 M ferrocyanide and 0.1 M
ruthenium hexamine each in 1 M potassium chloride. The peak splitting was
measured for scans taken at a 100 mV/s scan rate. The ferrocyanide is used to
probe the surface cleanliness; a clean surface has a peak splitting close to 59
mV. Both the ferrocyanide and ruthenium hexamine probe the conductivity of the
electrode, and a peak splitting of 59 mV indicates good conductivity. The
diamond electrodes used had a peak splitting of approximately 70 mV for both
analytes.
2.2.2. Raman Spectroscopy

Raman spectroscopy was used to study the microstructure of the diamond
film. Spectra were obtained at room temperature using a Chromex RAMAN 2000
spectrograph (Chromex, |nc., Albuquerque, NM). The instrument consisted of a
diode-pumped, frequency-doubled continuous wave Nd:YAG (yttrium aluminum
garnet) laser (500 mW at 532 nm, COHERENT), a Chromex 500is spectrometer
(f/4, 600 grooves/mm holographic grating, 50 um slit width), and a charge-
coupled device (CCD) detector (ANDOR Tech., Ltd.), 1024 x 256 element,
thermoelectrically cooled. Spectra were collected with an incident power density
of ~500 kW/cmz, a 10 5 integration time, and an approximate spectral resolution

of 4 cm". A white-light spectrum was collected and used as the background for

22

spectral correction. The Raman shift was calibrated using a spectrum of
acetaminophen.
2. 2.3. Resistivity Measurements

Resistivity of the diamond film is measured using a tungsten four point
probe connected to an HP 3748A multimeter (Hewlett-Packard, Palo Alto, CA).
The films used in this study showed semi metallic behavior and had resistivity
values of < 0.05 Q-cm.
2.3. Differential Pulse Anodic Stripping Voltammetry

Differential pulse anodic stripping voltammetry (DPASV) was used for
metal analysis. The measurement involves two steps, first, a negative potential is
applied for a fixed period of time to reduce the metal ions and form metal
deposits on the electrode surface (i.e., preconcentration). Second, the potential is
subsequently scanned toward positive values to selectively oxidize or “strip” the
individual metals. In the present work, metal deposition was performed for 180 -
210 s (as noted in each chapter) at a potential sufficiently negative of the
standard reduction potential. The deposition potentials for each metal ion were
-1.3 , -1.0, -0.8, -0.4, and 0 V for Zn”, Cd”, Pb”, Cu”, and Ag“, respectively.
Differential pulse voltammetry was used for the positive (anodic) scan with the
following parameters: a pulse height of 50 mV, a step height of 2 mV, 3 pulse
width of 50 ms, and a cycle period of 100 ms. A constant potential of 0.600 V
was applied for 180 s after completion of an anodic sweep to ensure full oxidation
of all metal deposits. All measurements were made in an unstirred solution and

at room temperature (~ 22-25 °C).

23

2.4. Comparison to Hg Electrode
2.4.1. Hg-coated Electrode

The Hg-coated electrode was formed by electrodepositing the metal on
glassy carbon (Hg-GC). Prior to use, the glassy carbon (GO-20 Tokai, Ltd.) was
polished under clean conditions using a series of alumina slurries (1.0, 0.3, and
0.05 pm) (Buehler, Lake Bluff, IL) in water. The polishing debris was removed by
ultrasonic cleaning for 15 min in ultrapure water after each polishing step.
Finally, the polished electrode was soaked in distilled isopropanol for 15 min prior
to use.3 In a measurement, the Hg was co-deposited along with metal analyte
from a 20 mM Hg” solution containing some concentration of the metal ion of
interest.4 The deposition potential used depended on the E°' for the M"*/M redox
couple but was sufficiently negative to co-deposit both Hg and the analyte metal.
2.4.2. Electrochemical Measurements

A 0.1 M acetate buffer, pH 5.2, with additional Ca” (30 ppm) and Mg” (5
ppm) ions served as the supporting electrolyte. Ca” and Mg” acetate salts were
added to meet NASA requirements for potable water on-board ISS. It was
determined that their presence did not affect the DPASV signal for any of the
metal ion contaminants.5 A cracked glass capillary was used to house the
reference electrode (Ag/AgCl) to provide a barrier against CI' ion transport into
the analyte solution; such leakage would cause complications with the
measurement due to calomel formation. All solutions were purged with N2 for at
least 3 min prior to analysis in order to remove dissolved 02. All measurements

were made in an unstirred solution at room temperature (~22-25 ‘C). Analysis

24

for the dissolved metal ions was carried out using DPASV as described
previously with a deposition time of 210 s.
2.5. Microgravity Modification Experiments
2.5.1. Electrochemical Measurements

A glass cell would be unsafe for on-orbit use, so a specially-designed
Teflon cell was used in this work. In the Teflon cell the working electrode was
mounted from the bottom using the same size Viton o-rings as the glass cell
described previously to define the working electrode area (~0.2 cmz). Long-teml
potential stability of the reference electrode is an important consideration for on-
board use of an electrochemical method. To this end, a “no-leak” AglAgCl
reference electrode (EE-009, Cypress Systems, Inc., Lawrence, KS) has been
incorporated which has similar stability to a standard AglAgCl reference
electrode.
2.5.2. X-ray Diffraction Spectroscopy

Powder X-ray diffraction spectroscopy (XRD) was used to investigate the
presences of metal deposits on the electrode surface following stripping. A
Rikagu Rotaflex RTP300 RC (Tokyo, Japan) instrument was used. The anode
used to generate the X-rays was a rotating Cu plate and the K10 X-ray line was
used (1.540 A). The scans were performed in the 28/8 mode at a scan rate of 1
degree/second.
2.4.3. Potable and Technical Water Simulants

Simulated Shuttle transfer water, potable (25-0203-02) and technical (25-

0203-03) types, were provided by the Wyle Laboratory (Johnson Space Center).

25

Each of these waters contained Ag” in a concentration between 300 and 500
ppb. The potable water also contained Ca” and Mg” at concentrations of 30
ppm and 5 ppm, respectively. The analysis results for the Ag+ concentrations
determined by the Wyle laboratory using inductively-coupled plasma mass
spectrometry (lCP-MS) were provided for each sample.

2.4.4. Microgravity Cell

A completely sealed electrochemical cell was designed for use in
microgravity to fit the requirements that all liquids be contained. A diagram of this
cell can be seen in Chapter 4, Figure 4.4. The body of the cell is made from
Teflon, and the top is sealed with a polymethylmethacrylate (Plexiglas) so the
inside of the cell could be visualized. There are 3 inlet and outlets on 3 of the
sides that are fitted with Iuer fittings and valves to be compatible with the Iuer
fittings on the current sampling bags used by NASA. The inner chamber is cubic
with side length 18 mm, and a volume of 5—6 mL.

There is a swagelok fitting on the top of the Plexiglas cover that is
intended for a Ag/AgCI “no-leak” electrode (EE-009, Cypress Systems, Inc.,
Lawrence, KS). The counter electrode consisted of a coiled Pt wire inserted
through a small hole in the Teflon body, and sealed with silicon sealant (DAP,
Inc., Baltimore, MD). The working electrode was a planar diamond thin-film that
was sealed by a Viton o-ring (i.d. 1/4 inch, ad. 5/16 inch) and had an exposed
area of 0.32 cm2, slightly larger than the conventional electrochemical cell.

The microgravity cell was used in conjunction with a portable potentiostat,

PG580 (Princeton Applied Research, Oak Ridge, TN). This particular model is

26

about the size of a calculator, and can be used in a standalone mode or with a
laptop computer. It can also be run on intemal rechargeable batteries. In these
studies the P6580 was run on battery power but was connected to a laptop
computer for data acquisition.

2.6. Urine Analysis

2.6.1. Electrochemical Measurements

No additional supporting electrolyte was added to urine simulant and urine
samples. For control samples 0.1 M acetate, pH adjusted to 8.8 (to match urine
simulant) with sodium hydroxide, was used as the supporting electrolyte. All
solutions were used within 1 week of preparation and stored in a refrigerator to
prevent bacterial growth in the medium.

Before analysis, each solution was deoxygenated in the electrochemical
cell by N2 purging for at least 3 min. The N2 purge was continued through the
deposition and stripping steps as a source of convection. A flow meter was used
to regulate the flow of N2 day-to-day. A negative potential of -1 V was applied for
300 s (unless otherwise noted) to deposit Cd on the electrode surface. Following
deposition, the metal was subsequently oxidized using DPASV as described
previously. Following analysis, a cleaning step was performed by holding a
positive potential (2 0 V) for 180 5 while purging with N2 purge, to ensure that any
Cd remaining on the surface is oxidized leaving a fresh BDD surface.

2.6.2. Simulated Urine and Rat Urine Samples
The components and concentrations used for the urine simulant were

obtained from a NASA recipe for simulated early planetary base mission

27

wastewater.6 The components are listed in Table 2.1 in order from highest to

lowest concentration.

Table 2.1. Components and concentrations in simulated urine.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Component Concentration
Ammonium Bicarbonate 250
Ammonium Hydroxide 100
Sodium Chloride 100
Potassium Bicarbonate 30
Creatinine 20
Potassium Dihydrogen 10
Hippuric Acid 7
Potassium Bisulfate 6
Lactic Acid 4
Citric Acid 3
DL-Tyrosine 3
D-Glucuronic Acid 2

 

 

 

After the components were mixed, the solution was sonicated for at least 15
minutes to ensure homogeneity.

Rat urine samples were obtained from the Galligan group from the
Department of Pharmacology & Toxicology at Michigan State University. Several
samples (2-3 mL) were obtained frozen (—20‘C) in plastic vials. A larger volume
was desired to allow for replicate measurements, therefore, 5 smaller samples
(2-3 mL) were combined to generate a single homogenous sample that was.
approximately 10 mL. The combined sample was then centrifuged for 15 min,
and the supematant filtered with a syringe filter (PTFE, 0.1 pm). The sample was

stored frozen (-20‘C) until use.

28

2.7. Blood Analysis
2. 7. 1. Electrochemical Measurements

The analyte of interest for this study was Pb”. The Pb” is associated
with the red blood cells, and digestion was necessary to release the Pb” for
electrochemical analysis. Hydrochloric acid (0.3 M) was used and was mixed
with the blood in a 5:1 (acidzblood) ratio. The sample was mixed directly in the
electrochemical cell, first adding the acid followed by the blood. A Teflon cell,
described above in the microgravity work, was used for the analysis of the blood
samples. Glass is typically not used as it can cause the blood to coagulate.
2. 7.2. Blood Sample

The blood obtained was bovine blood spiked with 47.2 ug/dL (~445 ppb)
of Pb” and was stabilized with EDTA to prevent coagulation. The blood had also
undergone two freeze/thaw cycles to lyse the red blood cells. These blood
samples were obtained from ESA, Inc., and the Pb” was added by the scientists
at ESA using the following procedure. The blood (Lampire Biological
Laboratories, Pipersville, PA) was weighed and corrected for density (1.05 g/mL),
then spiked with Pb” from a Pb” in DI water solution. The solution is rocked
gently at room temperature for 4 hours to allow the blood to uptake the added Pb.
Once the Pb” is added and allowed to mix the blood is analyzed by graphite
furnace atomic absorption spectroscopy and/or isotope diluted mass
spectrometry to confirm the concentration of Pb in the sample. The blood is then

divided into small plastic vials (~2 mL) and is frozen for long term storage or

29

refrigerated for short term storage. Blood samples can be used for to 2-5 years if
stored frozen.
2. 7. 3. Naflon Coating

A 3% Nafion solution was used to coat the diamond electrodes. The 3%
solution was made by diluting a commercial 5% Nafion in ethanol solution
(ElectroChem, Inc., Woburn, MA). 60 (IL of the 3% solution was placed on the
diamond film (~1 cmz) using an Eppendorf pipet (Eppendorf, Hamburg). A single
drop was placed in each comer of the square electrode and finally the center so
the solution would be evenly distributed over the electrode surface. The
electrodes were then allowed to air-dry on a flat surface. The electrodes were
covered by a beaker during the drying process to prevent dust particles from
settling into the polymer.
2. 7.4. Atomic Force Microscopy

Atomic force microscopy (AFM) images were taken to determine if the
Nafion coating was completely removed from the diamond surface using different
cleaning methods. A Nanoscope llI instrument (Veeco Metrology Group, Inc.,
Santa Barbara, CA) was used in the contact mode. Pyramidal-shaped Si3N4 tips,
mounted on gold cantilevers (100 mm legs, 0.38 N/m spring constant), were used
in air to acquire the topographical images. The scan rate ranged from 4 — 12 Hz

depending on the scan size (0.5 — 3 pm).

2.8. Chemicals
All solutions were prepared with ultrapure water from an E-pure water

purification system (Bamstead) having a resistivity greater than 17 MQ-cm.

30

Cadmium nitrate, 99.999% (Sigma-Aldrich), zinc acetate, 99.999% (Sigma-
Aldrich), cupric nitrate, 99.999% (Sigma-Aldrich), lead nitrate, 99.999% (Sigma-
Aldrich), silver nitrate 99.999% (Sigma-Aldrich), and mercuric acetate 99.999%
(Sigma-Aldrich) were all high purity and used without additional purification. 0.1
M acetate buffer was prepared by mixing the appropriate amounts of sodium
acetate, 99.995% (Sigma-Aldrich), acetic acid, 99.7% (Sigma-Aldrich), calcium
acetate, 99.999% (Sigma-Aldrich), and magnesium acetate, 99.999% (Sigma-
Aldrich).

The urine simulant contained: ammonium bicarbonate, reagent grade
(Columbus Chemical Industries, Columbus, WI), ammonium hydroxide, reagent
grade (Columbus Chemical Industries), sodium chloride, 99.7% (Jade Scientific),
potassium bicarbonate, 99.7% (Spectrum Chemical, Gardena, CA), creatinine,
98% (Sigma-Aldrich), potassium dihydrogen phosphate, 99% (Sigma-Aldrich),
hippuric acid, 99% (Sigma-Aldrich), potassium bisulfate (Sigma-Aldrich), lactic
acid, 85%, (Sigma-Aldrich), citric acid, ACS grade (Columbus Chemical
Industries), DL-tyrosine, 98% (Sigma-Aldrich), and D-glucuronic acid (Sigma-
Aldrich). All chemicals were used without additional purification.

All glassware and nalgene storage bottles were cleaned by a three-step
procedure consisting of a thorough washing in an alconox solution, soaking in 1
M hydrochloric acid and 1 M nitric acid for at least 10 min each, and rinsing with

ultrapure water. The cleaned glassware was then dried in an oven at ~ 55° C.

31

2.9. References

(1) Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chem. Mater. 2003,
15, 879.

(2) Granger, M. C.; Witek, M.; Xu, J.; Wang, J.; Hupert, M.; Hanks, A.;
Koppang, M. 0.; Butler, J. E.; Lucazeau, G.; Memnoux, M.; Strojek, J. W.;
Swain, G. M. Anal. Chem. 2000, 72, 3793-3804.

(3) Ranganathan, S.; Kuo, T.-C.; McCreery, R. L. Anal. Chem. 1999, 71,
3574-3580.

(4) Florence, T. M. J. Electroanal. Chem. Interfacial Electrochem. 1970, 27,
272-281. _

(5) McGaw, E.; Sonthalia, P.; Swain, G. M. SAE Technical Paper 2005-01-
2888 35th lntemational Conference on Environmental Systems 2005.

(6) NASA. Simulated Early Planetary Base Mission Wastewater, 2002.

32

CHAPTER 3

COMPARISION OF BORON-DOPED DIAMOND AND Hg-GC WORKING
ELECTRODES

3.1. Introduction

Hg based electrodes, drop or thin-film, have been the electrode of choice
for ASV for many decades. Hg has many desirable features for this
measurement including a large negative potential window, reproducible
response, and good sensitivity. However, the toxic and volatile nature of this
material are not desirable, and therefore, there is a need for alternative electrode
materials with comparable properties. Boron doped diamond (BDD) electrodes
have such properties including, a large potential window (negative and positive),
a chemically inert surface, a low background current, and a non-toxic and non-
volatile nature. These qualities make it a viable alternative to Hg electrodes.

A direct comparison of the performance of BDD and Hg-coated glassy
carbon (Hg-GC) electrodes for the detection of Zn”, Cd”, Pb”, Cu”, and Ag+ in
standard solutions by ASV was conducted.1 The basic electrochemical
properties of BDD, as they compare with Hg, are described herein. As well as
describing the fundamentally different nature of metal deposition and stripping on
the two electrode materials, the stripping peak potentials and shapes, and the
detection figures of merit for the analytes are presented. The results confirm that
BDD is a suitable alternate electrode for ASV that provides comparable or

superior detection figures of merit.

33

3.2. Potential Window

The working potential window is an important characteristic of an
electrode used for ASV because it determines which metal ions can be analyzed.
For carbon electrodes, the anodic limit in aqueous media is determined by the
potential at which oxygen evolution and or electrolyte electrolysis occurs. The
cathodic limit is determined by the potential at which hydrogen evolution
commences. The kinetic overpotential for these reactions varies depending on
the electrode material (i.e., its microstructure and electronic properties). The
wide cathodic potential limit for Hg is one of its desirable properties that is difficult
to find in alternates. The cathodic limit for Hg extends to -1.3 V (vs. Ag/AgCI) in a
pH 4.5 acetate bufferz. Only BDD3 and Bi4 (another alternate ASV electrode)
have cathodic limits this negative. The anodic potential limit, thus the potential
window, is also important for multimetal ion analysis. Hg has a working potential
window of 1.6 V but, a very low anodic limit.2 Most other alternate electrodes
have more narrow potential windows (e.g., 0.5 V for Au,5 1 V for Bi,4 0.7 V for
A96) except for BDD, which has a 3 V window.3 The cathodic limit for BDD is just
as negative as that for Hg but the anodic limit of BDD is nearly 2 V more positive.
This means that in addition to all of the metal ions Hg can be used to analyze,
BDD can also be used for metal ions with E“ values more positive than that for
Hg”/Hg (e.g., Hg” and Ag‘”).
3.3. Differential Pulse Anodic Stripping Voltammetry

Standard solutions of Zn”, Cd”, Pb”, Cu”, and Ag+ were used to

compare the ASV responses of BDD and Hg-GC. The same conditions were

used for deposition and stripping so that direct comparison of the detection
figures of merit for both electrodes could be made. Figure 3.1 shows overlaid

stripping voltammetric i-E curves for solutions of each individual metal ion.

Cd

2 “A I Cu
Zn fl Hg-GC

K/LyKD
\xL—JLJL JL/Aim

-1 300 -1 000 -700 -400 -100 200 500
Potential vs. AglAgCI (mV)

 

 

 

 

 

Figure 31. Overlaid individual stripping voltammetric i-E curves for 2100 ppb
solutions of each metal ion in 0.1 M acetate buffer, pH 5. 2, 30 ppm Ca” ,and 5
ppm Mg”. The top set of voltammetric i-E curves were obtained with Hg-GC and
the bottom set with BDD. The curves are offset for clarity. All curves are
uncorrected for the background current except the one for Cu on Hg-GC. Figure
reprinted with permission from Elsevier.

The oxidation potential for Ag is positive of that for Hg so this metal ion
cannot be analyzed using Hg-GC. Well defined stripping peaks are seen for all of
the metals using both electrodes, although the peak current and charge (area) for

diamond are lower than Hg-GC. A summary of the stripping peak potentials

(Ep°") for both electrodes is presented in Table 3.1, which also includes the

35

equilibrium potential, Eeq for the M""/M couple, as calculated from the standard
reduction potential using the Nemst equation, corrected for the reference
electrode (Ag/AgCI), and metal ion activity. The activity coefficients were
calculated with the extended Debye-Hiickel equation.7 The activity of the
reduced metal phase on BDD is unity because the metal is a pure substance. In
the case of the Hg electrode the activity of the reduced metal phase was
approximated as unity for these calculations. The reproducibility of E,“ for both
electrodes was comparable. For all the metals, Ep°" for both electrodes was
approximately the same, within 70 mV, even for the more electronegative metals.
This means that the BDD surface was just as active for the metal phase oxidation
as was Hg.

Given that the E°’ values for Zn”/Zn, Cd”/Cd, and Pb”le are well
negative of the flat band potential for hydrogen-terminated boron-doped
(semiconducting) diamond, the voltammetric data indicate that BDD is sufficiently
doped to be degenerate electronicallyf" 9 In other words, the electrode exhibits
metallic-like electrochemical behavior over a wide potential range.

Table 3.1. Stripping peak potentials and peak widths (FWHM) for BDD and Hg-
GC electrodes in the presence of individual metal ions at a 100 ppb

concentration are displayed. Mean and standard deviation values (n=3) are
presented.

 

 

 

Metal Ion E: (mV) FWHM (mV)
BDD H9-GC E.q BDD Hg-GC
_AQ+ 279 :l: 4 - 239 59 :l: 3 _.

 

Cu” 4713 434:0 -40 60:3 32:1
Pb” -559 a 1 -554 1; 0 -521 49 :l: 2 41 :l: 1
Cd” -789 i 1 -732 1: 2 -790 50 :l: 2 48 :I: 1
_z_n2* 407714 -114710 -1143 60:1 49:1

 

 

 

 

 

 

 

 

 

 

 

36

The Cd stripping peak for BDD occurs about 50 mV negative of the peak
for Hg—GC; however, the standard reduction potential for the Cd”/Cd couple is
approximately 50 mV more negative of that for the Cd”/Cd(Hg) reaction.10 This
accounts for the difference in Ep°" for the two electrodes. In the case of Zn, Ep°"
is near the predicted value for Hg-GC, however, for BDD the value is 65 mV more
positive. This suggests that there is a slight overpotential for this reaction on
BDD. There is also a difference in Ep°" for Cu for the two electrodes. Ep°" for Hg-
GC is approximately 45 mV more negative than the calculated Eeq. There is a
small difference of 5 mV between the standard reduction potential for the Cu”/Cu
and Cu”/Cu(Hg), so this does not account for the difference.10 In fact, Hg-GC
displayed very unusual behavior in the potential region of Cu stripping, which
likely affects Ep°". A very broad peak centered about 0 mV vs. AglAgCI was
observed in the background. An overlay of the background and total
voltammetric stripping current for a phase formed from a 100 ppb Cu” solution at
both Hg-GC and BDD is presented in Figure 3.2.

The shape, position, and magnitude of this background peak on Hg-GC
did not vary with the Hg” solution concentration, and it was present at the same
magnitude even when no Hg” was present. This confirms that the oxidation
peak is not caused by Hg or a contaminant in the Hg” salt. Further investigation
by cyclic voltammetry revealed that the peak height varied linearly (r2 = 0.997)
with scan rate between 20 and 300 mV/s. This trend is consistent with an
oxidation reaction involving a surface confined molecule. The peak potential also

shifted by -54 mVIpH (r2 = 0.999) as the pH of the buffer was varied from 3.0 to

37

5.2. This is very close to the theoretical -59 mV/pH slope predicted by the Nemst
equation for a reaction involving equal numbers of H+ and e’. Based on these
data, the background peak was ascribed to hydroquinone functionalities on the
60 surface that are oxidized to the corresponding quinone at this potential. As
expected, this peak is not seen in the voltammetric i-E curves for BDD because
such functionalities do not exist on the hydrogen-terminated surface. The
presence of the peak in the Hg-GC background complicated the analysis for Cu”
making background subtraction necessary for quantitation. It is supposed that
the oxidation of the underlying carbon simultaneously with the oxidation of the

solid Cu phase causes the apparent shift in Ep°".

 

 

 

 

2.1
W A Mi B
1.8' 12
1.5- 10
$12 i
is: ' 2,2; 8'
50-9‘ 5 6-
O 0
0.6- 4.
/
0.3'( 2‘
0.0. U I I I j 0 fl r ' ' '

 

 

-300 -200 -100 0 100 200 -300 -200 -100 0 100 200
Potential vs. AglAgCI (mV) Potential vs. AglAgCI (mV)

Figure 3.2. Differential pulse stripping voltammetric i—E curves for 100 ppb Cu”
in 0.1 M acetate buffer, pH 5.2. 30 ppm Ca” and 5 ppm Mg” were added. The
background voltammetric i-E curves are also shown for comparison. The
deposition potential was -400 mV vs. AglAgCI and the deposition time was 210 s.
The working electrodes were BDD (A) and Hg-GC (B). Note the difference in
current scales. Figure reprinted with permission from Elsevier.1

38

Another important parameter is the stripping peak width, which is
assessed by the width at half height (FWHM). Analytically, narrow peaks are
desired to improve the resolution of the measurement, thereby enabling the
detection of metal ions with closely spaced formal potentials. In Table 3.1, a
summary of the FWHM values for each metal at both electrodes is presented.
Both electrodes show similar FWHM values for all the metals, with the values for
BDD being slightly higher.

The broader peaks for BDD are attributed to the heterogeneous nature of
the electrical and electrochemical properties across the surface.11 The Hg
environment is simply more chemically and electrically homogeneous than is the
polycrystalline diamond surface. In other words, metal phase oxidation from Hg
occurs with one apparent heterogeneous electron-transfer rate constant, while
the oxidation from BDD likely occurs with a variable rate constant, depending on
the particular location on the surface. Moreover, the peak widths for BDD are
also influenced by the morphology of the metal deposit. For BDD or any
alternate electrode, a unifome dispersed metal phase of small nominal particle
size and narrow size distribution is desired.12 The dc deposition condition used in
the present work produced metal deposits of varying size (volume) on BDD, far
from optimum?" ‘2 This coupled with the heterogeneous chemical and electrical
properties inherently lead to broader stripping peaks than are seen for Hg-GC.
Even with this heterogeneity, the peak widths for BDD are not excessively broad

such that they preclude the analysis of multiple metal ions.

39

In addition to the broader peaks, there is a difference in peak shapes for
the two electrodes (Figure 3.3). The i-E curves for Hg—GC are more symmetric in
shape than are those for BDD, the latter being asymmetric with peak fronting.
The shape differences likely occur because of the manner in which the metal
phase is formed and subsequently oxidized at each electrode. In the case of Hg,
foreign metals deposit within a volume of Hg to form an amalgam. Thus, the
analyte metal resides in a very homogeneous environment with the concentration
of metal being proportioned to the volume of the Hg present. Because of this
chemical and electrical homogeneity, the oxidation reaction kinetics are uniform
for all metal atoms giving rise to a symmetric peak shape. However, metal
deposition on BDD occurs in a completely different manner. The metal deposits
form on the surface as particles with some metal atoms having only metal-metal
interactions and others having metal-diamond interactions. The polycrystalline
nature of BDD, in terms of the site heterogeneity and the non-uniform electrical
conductivity, makes for a complex surface from which metal oxidation occurs.11
It is expected that at slower scan rates the effect of the surface heterogeneity
would be minimized and the peaks would be more symmetric. In fact symmetric
peaks were observed on BDD for scan rates < 5 mV/s, which is consistent with

the complex surface leading to the asymmetric peak shape.

40

   

 

 

 

2.0 - 10.0 -
A e

1.5 - 8.0 -
a 1.2 - : a 5.0 - :
‘5 I ‘5 I
'5 0.8 - 5 g 4.0 - :

0.4 - E 2.0 - E

0.0 .' . . 0.0 4 L. .

 

 

 

 

-0.9 -0.8 -0.7 -0.6 -0.9 -0.8 -0.7 -0.6
Potential vs. AglAgCI (V) Potential vs. AglAgCI (V)

Figure 3.3. Differential pulse stripping voltammetric i-E curves for 100 ppb Cd”
in 0.1 M acetate buffer, pH 5.2, 30 ppm Ca”, and 5 ppm Mg”. The deposition
potential was -400 mV vs. AglAgCI and the deposition time was 210 s. The
working electrodes were BDD (A) and Hg-GC (B). Note the difference in current

scales.

The rising portion of the i-E curves for Cd, Pb, and Cu on Hg-GC can be fit
to an exponential trend (r2 2 0.997), and the peak occurs near the theoretically
predicted potential. Such shape is consistent with fast oxidation reaction kinetics,
and an apparent rate constant that is the same across the surface. Oxidation of
the metal phase is an exhaustive electrolysis process (pseudo thin-layer
behavior) so that the peaks tend to be relatively narrow with no diffusional tail
past the maximum. As long as the Hg volumes from which the metal phase is
being oxidized are relatively constant across the surface, the peak widths will be
narrow, approaching the theoretical value 45.3 mV (n=2).13 On the other hand,
the rising portion of the i-E curve on BDD increased more slowly with potential,

and could not be accurately described by an exponential function. As discussed

41

above, there are many different surface sites (chemically and electrically) from
which metal phases can be oxidized. This site variability leads to drawn out i-E
curves. Even though the BDD peaks have a different shape, overall both
electrodes yielded similar responses in terms of peak position and width. The
similarity of these parameters indicates the metal oxidation reactions are nearly
as favorable for BDD as they are for Hg.
3.4. Detection Figures of Merit

Despite the lower stripping peak current and charge (area) for all the metal
ions, the limit of detection and linear dynamic range for BDD are practically the
same as for Hg-GC. Figure 3.4 presents a series of stripping voltammetric i-E
curves for standard solutions of Cd” at BDD and Hg—GC. For both electrodes,
the current and charge increase proportionally with the solution concentration.
However, the slopes of the response curves for Hg-GC are about a factor of four
higher than those for BDD. A summary of the metal ion analysis data is
presented in Table 3.2. The peak current and charge data presented were
background corrected. The background current values reported were measured
at the stripping peak potential for each metal. In all cases, the background
current for BDD was lower than that for Hg-GC by a factor of three to ten. This is
especially apparent for the more electronegative metals, Cd” and Zn”. A
characteristic property of BDD is a low and featureless background current that is
stable with potential and time.3' ‘4 The background current for Hg-GC depends
on the condition of the GC surface as well as the amount of Hg present but, in

general, is a factor of 5-10 greater than for BDD of the same geometric area.

42

The lower background current leads to enhanced signal-to-background ratios in

electroanalytical measurements. Additionally the peak-to-peak fluctuation (noise)

in the background current is lower for BDD.

 

 

 

 

 

 

 

 

  

15- -
A II 1000 ppb 3 1000 ppb
. 4f 55‘ AL
11 -
A 3?
3 335-
E 7‘ t‘: .
D 3
O 0
3+ 15‘ A
'A10ub
-1 I I I I -5 I I I I

 

 

-0.9 -0.8 -0.7 -0.6 -0.9 -0.8 -0.7 -0.6
Potential vs. AglAgCI (V) Potential vs. AglAgCI (V)

Figure 3.4. Differential pulse stripping voltammetric i-E curves for standard
solutions of Cd” in 0.1 M acetate buffer, pH 5.2, 30 ppm Ca”, and 5 ppm Mg”.
The deposition potential was -1000 mV vs. AglAgCI and the deposition time was
210 s. The working electrodes were BDD (A) and Hg-GC (B). Note the
difference in current scales. Figure reprinted with permission from Elsevier.1

The minimum concentrations detectable at signal-to—noise ratio (SIN) 2 3
are also presented in Table 3.2. For Cd”, Pb”, Cu”, and Ag", concentrations
from 1 — 10 ppb were detectable with BDD. These values are as good or
superior to the values for Hg-GC. Ag+ was detectable down to a level of 1 ppb
but only with BDD. The minimum concentration of Zn” detectable with BDD was

a factor of five higher than with Hg-GC. The exact reason for this is unknown but

may have to do with the electronic properties of the electrode at the stripping

43

potential. There is also a discrepancy between the minimum concentrations of
Cu” detectable with the two electrodes. The lowest concentration detectable
with BDD is a factor of five lower than the value for Hg-GC. As discussed
previously, the large background signal from the redox-active carbon-oxygen
functionalities on GC complicate the analysis of Cu”. The magnitude of this
background peak is in the (1A range, making it difficult to separate from the Cu”
signals, and impossible to obtain a low detection limit.

The sensitivity is an important parameter for low detection limits. Hg-GC
exhibits a response sensitivity that is three to five times greater than BDD for
each of the metal ions. Typically, a higher sensitivity value will result in a lower

limit of detection, which is seen in equation 3.1

3N
CLoo = __ (3.1)

m
CLOD is the minimum concentration detectable (ppb), N is the amplitude of the
noise (A) and m is the sensitivity of the measurement (A/ppb). Even though the
sensitivity was lower for BDD, the lower noise/background current enabled as
good or superior CLOD values to be achieved. While a direct comparison of the

sensitivity is often useful for determining how well two electrodes perform

analytically, it is not a good comparative of these two electrodes.

:E: L000: 05 22000000: Ho: 2 0:0 005000;. 5000:0800 $00.0. 0009.00 4

 

 

 

 

 

 

 

 

 

 

 

 

- - - - - 00-0.... m<
$4.0 .009 - F 0F.0 H 9.0 09 no.0 H 40.0 com +
$N.0 .000? - 00 0.0 H 4.0 00 0 H NV 00-“: :0
$5 .009 - 0_. F40 H 4_..n or 5.0 H 00.0 new a
$0.0 .000? - 0 4; H 0.40 0.0 N0 H 09 00.“: an.
$0.0 .000? - 0 4.0 H 0.0? 0.0 00.0 H 00.0 com a
$0N .009 - F 00 H 0.00 0.? No H he 00.0.: no
$ v.0 .009 - v 0.0 H NE. 0._. 00.0 H 04.0 com A
$0.0 1009 - or 04 H 0.04 or 0.0 H 0.0 00.0.... :N
$0.4 000.. - 00 00.0 H 04.0 00' F0 H No com a
a": see a N 25 as

an 00 .0 am”: 00:0”. 3002.5 305 00000000 50:000.". x000
a 0 332200 coznhc0ocoo .

.£__e_oa_eoaom

 

 

0_Emc>o 50:5

 

 

“0026..

 

5 230.3000

 

 

 

0002520 00.0... new. 000 9:0: 005000E 0:2 .90E .60 E0E ho 00.50: 800000 0.0 030%

45

Another important analytical parameter to compare is the linear dynamic
range, which is assessed from response curves. The response curves were
linear for all metal ions at both electrodes from the low to the high (1000) ppb
range with linear regression correlation coefficients of 0.99, or greater. Previous
studies have shown that BDD provides a linear dynamic range of four orders of
magnitude for various metal ions.3 This range is comparable with that for Hg-GC,
and is more than adequate for most analytical applications. Lastly, BDD must
have the same reproducibility of results as a Hg-based electrode. For all of the
metals tested, BDD showed a run-to-run variability of less than 5 % RSD (n=3),
and this is comparable to values for Hg-GC. However, if BDD is to be a viable
alternate electrode, reproducible results must also be obtained from electrode-to-
electrode. This was assessed by analyzing a 100 ppb Ag+ solution with four
different films using a minimum of three measurements for each. It was found
that the electrode-to-electrode reproducibility was excellent with a variability of
only 3.8% RSD.

3.5. Comparison of Metal Deposition and Stripping Charge for BDD

In past studies with BDD, issues with both incomplete metal oxidation and
metal phase detachment have been reported, both of which limit the response
sensitivity in ASV.1‘“7 No evidence for metal phase detachment during past work
with Pt electrodeposition on BDD was observed.”20 Since the ASV analysis
involves different metal phases, a few experiments were conducted to probe for

metal loss by either incomplete oxidation or detachment.

46

The first set of experiments was designed to look for incomplete metal
oxidation. Double potential step chronoamperometry was used to record the
current passed as a function of time during the deposition and stripping steps.
The i-t curves were integrated to determine the charge passed. The first step
was to apply 0.6 V for 3 min to oxidize any remnant metal particles from previous
measurements. Following this, the metal particles were formed via
electrodeposition at a potential for which the reaction is mass transfer limited.
This potential was different for each metal, depending on its E°'. The deposition
time in each case was 210 s. The potential was then switched to 0.6 V for the
same period of time to oxidize the metal phase. The apparent faradaic charge for
the metal deposition and oxidation were determined by subtracting the
background charged recorded in the absence of the analyte metal ion. The
resulting data are presented in Table 3.3.

Theoretically, the stripping/deposition charge ratio should be unity if both
reactions occur with 100% efficiency. In the case of the electropositive Ag‘lAg
couple, the ratio was very near unity. This confirms that incomplete metal phase
oxidation did not occur to any appreciable extent on these films, at least for this
metal, as has been reported previously.” ‘7 The ratio decreased progressively
for the more electronegative metals ranging from 93% for Ag‘lAg to only 8% for
Zn”/Zn. This was because the deposition charge progressively increased with
the deposition potential (e.g., Zn and Cd), and it reflects all reactions occurring at

the electrode. The more negative the potential is, the more reactions that

47

0.00.2... .0 500300 .8550 E00 0000.00.00 0.0 00.0.. 05:0

05000.0 0.050 m 5.3 50.0. .005 0.:0E05000E

00.00.50 05200.05 05

0.00 .04 00.000 0500.00 0:0 50.00000 0.0 5050 00 00.0.0005 0.03 00.090 0535.50 0500.00
0:0 50.00000 0:... .> 0.0 .0 305.00 0500.00 0 0:0 0.00. 05 5 00.0.. 00 0.00550 50.00000
.0 0pm .6 0.053 00.0 050: 0050000 003 0:05.098 00.0E050E5505 00.0 0.05.00 03:00 <

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4.00... 4.0.. 400 4.0.0 4.4.0 00m
. H . . H . . H . . H . . H . .553 on. e. E0
woo moo Bo Bo So 0:. So So 080 Boo can: .920
4.0.0 4.: 4.0.... 4.0.. 4.0.. :0:
. . . . . . . . . . 83
830:. 2.3.2 8032.. :30: .33.: 5920055500
4.8.0 4.8.0 40.0 4.0.» 4.0.0 00m
. . - . . - . . - . . - .0...
83.8.. No.08. 3403 00400.1 2.32 8.9.0 cosmoaoal
.> a... u new. .> 4.0. u 9.0. .> 0.... u 9.0. .> e. F. u .450. .> a... u .50. .222
0< :0 ed 00 an

 

 

05000.0 000 m 050: «.0 I0 .505 0.0.000 .2 ...0
5 E00 .. .0 5000:0050 m 00 05. .90.: .5520... 5.. 000.000 0500.00 0:0 50.00000 .00 0.00...

48

are possible. For these particular metals, hydrogen evolution is a key parasitic
reaction. While the reaction kinetics are not known for each of the metals it is
reasonable to suppose that the hydrogen evolution reaction kinetics are more
rapid for all of the metal phases than for diamond. This would account for a
greater cathodic charge passed during the deposition step for the more
electronegative metals (i.e., the decreasing stripping/deposition charge ratio).

To confirm that the stripping/deposition charge ratio decrease was a
function of the deposition potential rather than the property of the specific metals
a similar experiment was performed using only Cu”. The same deposition
potentials were used as in the previous experiment, and the results are presented
in Table 3.4. From these data the same trend is observed again, a decreasing
stripping/deposition charge ratio with decreasing deposition potential, suggesting
that this phenomenon is related to parasitic reactions.

A second set of experiments was performed to look for metal phase
detachment. Double-potential step chronoamperometry was used to record the
current passed as a function of time during deposition and stripping steps. The
measurements were made as a function of time with the hypothesis that longer
deposition times would lead to more metal phase formation, and consequently,
greater detachment. Cu was chosen to probe this effect because detachment of
this metal has been previously reported, and the results are summarized in Table
3.5.16

As expected, the charge passed increased proportionally with deposition

time. The charge ratio basically remained constant (~70%) for all deposition

49

0.0 0.00 =<

000000.0 0.050 0 5.3 50.0. Anus
0.:0E05000E 0.00.08 .0 :000.>00 0505.0 .50 00.0.0200 0.0 00.0.. 05.5 00.00000 05905000

00.000 0500.00 0:0 50.00000 0... 50.00 0. 5.0.00.5 0.03 00.0000 0600:0000

0500.00 0:0 50.00000 0...... .> 0.0 .0 .00:0.00 0500.00 0 0:0 0.00. 0:. 5 00.0.. 00 0.00:0.00 50.00000
.0 0.0 .0 00.0.3 00.0 050: 0050.50 003 .:0E..00x0 00.050.000.00550 00.0 .00:0.00 0.0000 .0.

 

 

 

 

 

 

 

 

 

 

 

 

 

4.00.0 4...... 4.00.. 4.00.0 _ 00m
- :0 mcnw mph—mm..— ‘
00.0 H 00.0 000.0 H 00.0 0000.0 H 00.0.0 00000.0 H 00000.0 0 as... ”“0300. .0
4.00.0 4.0.0 4.0... 4.00.. 000
. . . . . . . . 6:.
00300 SH. .0 00H000 0 .H000 00.20 00.00.00
4.00.0 4.00.. 4. .00 4.0... 000
a
.0 H .00- 0 H .0.- 00 H 000.- 000 H 00000- 00.000.mo_.._0o._00
0.0. 0.0. 0..- 0..- .>. .30

 

 

 

._0..:0.00 50.00000 05b0> .000000_0 com
0 050: «.0 I0 000:0 0.0.000 s. _..0 5 000 E00 _. 5. 000.000 0500.00 0:0 50.00000 .00 0.00...

50

. 00050.0 0.0:.0
0 5.3 50.0. Amucv 0.:0E0.0000E 0.0...0E .0 :000.>00 0.00:0.0 05 E00 00.0.0200 0.0 00.0.. 0500
00.00.50 0:00.00000 0.03 0.00 :0. 00.000 0500.00 0:0 50.00000 05 50.00 0. 00.0.00.:. 0.03
00.5.0 0E....:0..00 00... .> 0.0 003 .00:0.00 0500.00 05 0:0 > 0.0- 003 .00:0.00 50.00000 0:...
.05. 50.00000 0:.>.0> 5.3 00E.0..00 003 .:0E..00x0 0...0E0.00E005.:0 00.0 .00:0.00 0.0000 <

 

$0.. 0000.0 00.0.. $30.0 $05.0 — 0mm

 

500.303.300.00.
0000 00.000

  

5.0 H 00.0 30.0 H and 3.0 H and mood H nmd mood H mmd

 

 

 

 

 

4.... 4.00.0 4.00.0 4.00.0 4.00.0 00m
0
0.0 H .0. 00.0 H 00.0 00.0 H 0.0 .00 H 00.0 .00 H 00.. 00 .0..%W:..00. 00
4.00.0 4.00.0 4.0.. 4.00 4.000 000
80.

N6 H Qmwu —..o H 0N..- mNd H mad- 00.0 H 00.0. 00.0 H N50- 09.000 005.0092—

 

0 0.0 0 00 0 00 0 0. 0 0 0E... 000.02.8'

 

 

 

 

 

 

 

.05. 50.00000 0:.>.0> 00950.0 000
0 0500 N0 I0 ..0..00 0.0.000 s. ..o 5 000 E00 . .0. 000.000 0500.00 0:0 50.00000 0.0 0.00-.-

51

 

times except the shortest, for which there was a significant decrease. The fact
that the stable ratio is less than 100% is attributed to the parasitic evolution of
hydrogen on the depositing metal phase. The relatively constant ratio is
evidence for minimal metal phase detachment. The trend in the
stripping/deposition charge ratio with deposition time is consistent with a metal
formation and growth model.21 In this model, metal phase formation involves
nucleation of a metal phase and growth of the deposit from the initially formed
nuclei. The nuclei must reach some critical size before they become stable on
the surface and then function as sites for growth. Sufficiently long deposition
times (i.e., induction time) are needed to achieve the critical size. At the short
deposition times (e.g., 5 s and 10 s), the efficiency of the deposition is not as high
as at long times because few of the nuclei reach critical size (i.e. are stable).
Experimentally, this same trend was observed by Hyde et al. in a study of Ag
electrodeposition on BDD. They noted a 15 s “induction period” during which few
nuclei formed on the surface.22
3.6. Conclusions

It was shown through direct comparison measurements that BDD is a
suitable alternative electrode to Hg for the ASV determination of common metal
ion contaminants. BDD exhibits many of the same electrode properties as does
Hg, but it is not toxic or volatile. BDD also has a wider anodic potential limit, and
a lower background current density than Hg. Due to the hydrophobic nature of
the hydrogen-terminated surface, BDD is quite resistive to deactivation such that

conventional time-consuming pretreatment is normally not needed. This is an

52

attractive feature for real sample analysis. Overall, the detection figures of merit
for BDD are as good or superior to those for Hg-GC. Both BDD and Hg-GC
provided detection limits for all the metal ions in the mid to low ppb range (SIN 2
3). Only the minimum detectable concentration for Zn2+ was lower for Hg-GC.
The response reproducibility was < 5% (RSD) for BDD from run-to-run and from
electrode-to-electrode. The biggest difference between the two electrodes is the
magnitude of the stripping peak current and charge, which are greater for Hg-GC.
The larger electrochemical response on Hg-GC resulted in a higher sensitivity,
but this was offset by a similar reduction in the background noise for BDD which
resulted in comparable or lower CLOD values. Electrochemical measurements
revealed that reduced sensitivity due to incomplete metal oxidation or metal

phase detachment was not an issue for these BDD films.

53

3.7. References

(1) McGaw, E.; Swain, G. M. Anal. Chim. Acta 2006, 575, 180-189.

(2) Wang, J.; Lu, J.; Hocevar, S. B.; Farias, P. A. M.; Ogorevc, B. Anal.
Chem. 2000, 72, 3218-3222.

(3) Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chem. Mater. 2003,
15, 879.

(4) Chatzitheodorou, E.; Economou, A.; Voulgaropoulos, A. Electroanal.
2004, 16, 1745-1754.

(5) Bonfil, Y.; Brand, M.; Kirowa-Eisner, E. Anal. Chim. Acta 1999, 387, 85-
95.

(6) Kirowa-Eisner, E.; Brand, M.; Tzur, D. Anal. Chim. Acta 1999, 385, 325-
335.

(7) Debye, P.; Huckel, E. Physik. Z. 1923, 24.
(8) Swain, G. M.; Ramesham, R. Anal. Chem. 1993, 65, 345-351.

(9) Alehashem, 8.; Chambers, F.; Strojek, J. W.; Swain, G. M.; Ramesham,
R. Anal. Chem. 1995, 67, 2812-2821.

(10) Bard, A. J.; Faulkner, L. R. Electrochemical Methods Fundamentals and
Applications, 2nd ed.; John Wiley & Sons, Inc.: New York, NY, 2001.

(11) Holt, K. B.; Bard, A. J.; Show, Y.; Swain, G. M. J. Phys. Chem. B 2004,
108, 15117-15127.

(12) Sonthalia, P.; McGaw, E.; Swain, G. M. Anal. Chim. Acta 2004, 522, 35-
44.

(13) Ramaley, L.; Krause Jr., M. S. Anal. Chem. 1969, 41, 1362-1365.

(14) Granger, M. C.; Witek, M.; Xu, J.; Wang, J.; Hupert, M.; Hanks, A.;
Koppang, M. 0.; Butler, J. E.; Lucazeau, G.; Mermoux, M.; Strojek, J. W.;
Swain, G. M. Anal. Chem. 2000, 72, 3793-3804.

(15) Vinokur, M; Miller, B.; Avyigal, Y.; Kalish, R. J. Electrochem. Soc. 1999,
146, 125-130.

(16) Nakabayashi, S.; Tryk, D. A.; Fujishima, A.; Ohta, N. Chem. Phys. Lett.
1999, 300, 409-413.

(17) Awada, M.; Strojek, J. W.; Swain, G. M. J. Electrochem. Soc. 1995, 142,
L42-L45.

(18)
(19)
(20)

(21)

(22)

Wang, J.; Swain, G. M. Electrochem. Solid-State Lett. 2002, 5, E4-E7.
Wang, J.; Swain, G. M. J. Electrochem. Soc. 2003, 150, E24-E32.

Bennett, J. A.; Show, Y.; Swain, G. M. J. Electrochem. Soc. 2005, 152,
E184-E192.

Pletcher, D. A First Course in Electrode Processes; Alresford Press Ltd.:
Alresford, Great Britain, 1991.

Hyde, M. E.; Jacobs, R. M. J.; Compton, R. G. J. Electroanal. Chem.
2004, 562, 61-72.

55

CHAPTER 4
METHOD ADJUSTMENTS FOR MICROGRAVITY APPLICATION
4.1. Introduction

Water on-board the lntemational Space Station (ISS) presently comes
from three sources: recovered humidity condensate, fuel cell transfer water from
the Shuttle, and preparations on earth."3 A large portion of the crew’s drinking
water is supplied from the condensate (recovered water), which is treated in and
dispensed from the SRV-K system. The resulting deionized water is then passed
through a conditioning bed that adds 300-500 ppb of Ag+ for biocidal purposes
and minerals (calcium, magnesium, and fluoride ions) for taste. The water is
then pumped to a galley where it is pasteurized and dispensed. The water from
the SRV-K system is not prepared for extended storage but rather is supplied on
demand.

The remaining supply comes from the stored water, which is mainly the
Shuttle fuel cell transfer water. Iodine is added to this stored water at a
concentration of 2-4 ppm for biocidal purposes. This water is prepared for
transfer to I88 by first deiodinating it, and then storing it in contingency water
containers (CWCs, 44 L). As water is being collected in the CWC, Ag+ is added
(300-500 ppb) along with minerals for taste. These chemicals are added during
the collection process from ground-loaded syringes to ensure homogeneous
mixing.2 This stored water, referred to as potable, is transferred to I88 in the
CWC, and is used for drinking, food rehydration, and hygiene. It can be

dispensed directly from a CWC using the SVO-ZV system. Some of the fuel cell

56

water Is also prepared for transfer by adding Ag+ without any mineralization. This
water is referred to as technical, and is used mainly for toilet flushing and
electrolytic oxygen generation. In the event that the volume of recovered water
available does not meet the crew demand, the potable water can be added to the
SRV-K system - so-called make-up water.

The water quality is a critical issue for crew health and safety. There is
presently minimal monitoring capability of the recovered and stored water on-
board ISS. Therefore, there are no means for rapidly responding to a
contamination event. Monitoring of the ISS water is currently performed by
periodically collecting (~1 month intervals) both SRV-K and SVO-ZV water in
special Teflon storage bags."3 These samples are then returned to earth during
resupply expeditions, and comprehensively analyzed for turbidity, pH, organic
and inorganic chemical content, and biological agents. Ag”, of course, is
important to regularly monitor in both the recovered and stored water in order to
assess biocidal activity. Additionally, other metal ion contaminants have been
detected in water samples returned from previous Shuttle expeditions." 3 For
example, Pb2+ levels near or above the Environmental Protection Agency (EPA)
action limit of 15 ppb and Cd” levels above NASA’s Medical Operations
Requirement Document (MORD) limit of 5 ppb have been found. The sources of
these contaminants have been identified, and system corrections made to
eliminate them. However, the unexpected appearance of these toxic inorganic

contaminants points to the need for on-board monitoring and control.

57

The requirements for an analytical technique suitable for space
applications are different from those needed on earth. First of all, the
instrumentation must be Iightvveight, compact, robust, and energy efficient. The
technique must have long-term stability, as replenishing missions are several
months apart, and require little user interaction. Furthermore, there are
limitations because of the microgravity environment. All liquids must be enclosed
and, for safety reasons, the pH range is limited to between 5.5 and 9.

Electrochemical methods are attractive for on-board monitoring because
they fulfill many of these requirements. Electrochemical methods require minimal
maintenance and user interaction especially, if a stable and sensitive electrode
material is employed. The goal of this work was to develop an electrochemical
method for water quality monitoring using a boron-doped diamond thin film (BDD)
electrode and anodic stripping voltammetry (ASV).4 The focus was on heavy
metal ions (e.g., Ag+ biocide).

Due to the stringent requirements on-board, work was conducted to adapt
a previously developed ground-based method. Efforts to evaluate the ASV
method (i) at increased pH from 4.5 to 5.2, (ii) after adding Ca2+ and Mg” salts
for mineralization, (iii) after eliminating the nitrogen purge step (dissolved oxygen
removal), and (iv) during long-term electrode response stability, are reported.
Additionally, the development of a prototype microgravity cell, and results from
the analysis of Ag“ in simulated potable and technical water samples, provided

by the Wyle Laboratory (Johnson Space Center) are discussed.

58

4.2. Solution Parameters
4.2.1. Effect of pH Adjustment

Ground-based ASV methods are usually performed in an acetate buffer
supporting electrolyte solution at pH 4.5.5' 6 The pH of this solution is outside the
range allowable for use on-board ISS (5.5 to 9). Tests were therefore conducted
to determine the detection figures of merit for Ag‘, Cuz“, sz“, Cd”, and Zn2+ in
an acetate buffer supporting electrolyte solution at pH 5.2.

Increasing the solution pH can have two harmful effects on metal ion
analysis by ASV. First, the activity of a metal ion can be reduced due the
formation of metal hydroxide species (soluble and insoluble). Second, the
increase in pH shifts the onset for hydrogen evolution toward more positive
potentials, and this can interfere with the detection of metal ions having very
negative standard reduction potentials (e.g., Zn”). Table 4.1 shows the
minimum concentration of each metal ion detected along with the SIN ratio at pH
5.2, and comparison data at pH 4.5 are also presented. It should be noted that
the measurements at pH 4.5 were made using a pulse height of 100 mV while
the measurements at pH 5.2 were made with a pulse height of 50 mV. The
decrease in pulse height may slightly decrease the response sensitivity by
reducing the stripping peak current. However, the choice was made to use the
more standard 50 mV pulse height because no significant improvement was
seen using larger pulse heights.7' 8

Even with the different voltammetric parameters, and the solution pH

increase, the minimum detectable concentrations for Cd”, Pb”, and Cu”,

59

measured or estimated for a S/N of 3, are in the low ppb range. The exceptions
are Zn” and Ag*. The higher minimum detection for Zn” at the higher pH is
expected do to the greater interference by hydrogen evolution on Zn metal
formation during the deposition step. The data for Ag+ reveals that the minimum
detectable concentration is a factor of 10 higher at pH 5.2. This was an
unexpected result but it may be caused by the formation of Ag+ acetate
complexes, which would reduce the Ag” activity in solution. There are two known
Ag acetate (OAc) complexes: AgOAc and Ag(OAc)2'. Using complexation
equilibrium equations and mass balances in the pH range of 4 to 6 there is a 2%
decrease in the amount of free Ag” when OAc is present. Even at pH 5.2, the
detection limit for Ag+ is still in the low ppb range, well below the MORD required
limit of 50 ppb.

Table 4.1. Comparison of the minimum concentrations of several priority metal
ions detectable with a diamond thin-film electrode at pH 4.5 and pH 5.2.

 

 

 

 

 

 

 

pH 4.5 pH 5.2
Metal Ion ppb S IN ppb S/N
Zn2+ 6.5 3 50 20
Cd?+ 1 .1 5 10 30
Pb2+ 2.1 3 5.0 3
Cu” 0.64 3 10 20
Ag“ 0.1 1 4 1.0 4

 

 

 

 

 

 

 

In terms of the response sensitivity, the effect of increasing the pH
appears minimal for all the metal ions except Zn”. For example, the response

sensitivity for Ag+ was 110 nC/ppb at pH 4.5 and 130 nC/ppb at 5.2.

60

4. 2. 2. Effect of Mineralization

Ca” and Mg” forrnate salts are added to the SRV-K water for taste at
concentrations of 30 and 5 ppm, respectively. Measurements were conducted to
determine if the presence of the salts affected the electrochemical response of
standard solutions of the metal ions. In this study, acetate salts of Ca” and Mg”
were used instead of the forrnate salts used in NASA preparations of potable
water.

Table 4.2. Comparison of the stripping response (peak area) for each priority
metal ion before and after the addition of Ca” and Mg” salts.

 

 

 

 

 

 

 

Metal Ion 2:: ”5: (Spa? % Difference
100 ppb Zn” 0.66 0.69 -3.8
50 ppb Cd” 0.86 0.85 0.92
100 ppb Pb” 1.9 1.9 -0.24
500 ppb Ag+ 3.9 3.8 3.0

 

 

 

 

 

The addition of these cations has little effect on the ASV response for any
of the priority metal ions. Less than a 4% change was observed.
4.2.3. Effect of Dissolved Oxygen

Solutions are normally purged of dissolved oxygen prior to making an ASV
measurement. The reduction of oxygen both at the BDD surface and the
deposited metal surface compete with metal deposition during the
preconcentration step. This may reduce the amount of metal formation resulting
in decreased stripping current. On earth the removal of oxygen is usually
accomplished by sparging the solution with an inert gas (e.g., N2, Ar). Oxygen
removal with this method is not possible on-board ISS so the effect of dissolved

oxygen on the diamond thin-film electrode response for each of the priority metal

61

ions was investigated. Standard solutions containing the metal ions (Cd”, Pb”,
Cu”, and Ag“) were tested individually, both with and without a 3 min N2 purge.
All measurements were made in acetate buffer, pH 5.2.

The presence of dissolved oxygen did not significantly affect the ASV
response for Ag+ or Cd” at BDD, as demonstrated for Ag+ in Figure 4.1 A. The
stripping peak potential, current, and charge (area) were unchanged in the
presence of dissolved 02. The only difference between the Ag+ and Cd”
responses was the higher background current without an N2 purge at potentials
where Cd is detected. However, the faradaic current for Cd” stripping was not
affected. This higher background was due to the 02 reduction reaction that
occurs in the potential range where the Cd is oxidized (stripped).

On the other hand, there was a significant effect on the stripping peak for
Pb” as shown in Figure 4.1 B. It is clear that when dissolved oxygen is present,
the stripping peak current and charge are reduced. There is also a small
reduction in the signal from Cu” (~15%) when dissolved oxygen is present. The
effect is not related to the E0 of the metals because Cd was not affected, and it
has a more negative E° than either Pb or Cu. It is thought that the reason only
these two metal ions were affected by the dissolved 02 has to do with their
reactivity with oxygen. There are two possible causes for this effect. First, the
deposited Pb (solid) could react with oxygen to make PbO2, passivating the
surface and inhibiting the oxidation. Second, reactive oxygen species (e.g.,
H202, 0') could chemically oxidize the deposited Pb from the surface, removing it

prior to electrochemical oxidation.

62

 

 

 

 

 

10- 207
A \ e — N2 purge
- - N0 N2 purge
0.9 - 1.6 q
A 0.8 - A
g g 1.2 -
E 0.7 - ‘5
E $0.,
0 0.6 - O
/ \
0.54 0°44...—— t--.
0.4 . . a 0.0 . r a
100 200 300 400 -675 -575 -475
Potential vs. AglAgCI (mV) Potential vs. AglAgCI (mV)

Figure 4.1. Stripping voltammetric i-E curves for 100 ppb Ag+ (A) and 100 ppb

Pb” (B) in acetate buffer, pH 5.2, with and without a N2 purge. Image adapted
from McGaw, et. al.9

In order to determine which of these possibilities was true, a series of
experiments was performed to measure the residual Pb on the diamond
electrode. In the case where PbO2 is formed there would be Pb remaining on the
surface after stripping, but in the case where Pb is chemically oxidized there
would not. Powder X-ray diffraction (XRD) was used to determine the presence
of the Pb. Initially, a spectrum was taken of the nanocrystalline diamond
electrode to determine which peaks were due to the diamond (Figure 4.2), and a

second spectrum was taken of the electrode after electrochemical deposition of

Pb (Figure 4.2).

63

— Pb deposited on BDD

 

 

 

 

600 - 3.12 A (Pb)
— BDD (x 5)

0 500 - 1.05 A (Pb)
‘65
3
8 400 -
>4
S
f, 300 4
O
9
§ 200 -
0% 1.57 A (Pb) 1.07 A (BDD)

100 ' 2.06 A (BDD) L 1.26 A (BDD)

.. ,,
0 I’M—W
20 4O 60 80 100

26 (‘)
Figure 4.2. Powder XRD spectra taken of a bare diamond electrode (gray) and a
diamond electrode with Pb electrochemically deposited on the surface (black).
Peaks due to the diamond are marked with BDD and peaks due to Pb are also
marked accordingly. The BDD signal is multiplied by 5 and spectra are offset
slightly for clarity.

There are three peaks from the diamond, 2.06 A, 1.26 A, and 1.07A,
which correspond to the 111, 220, and 311 crystal faces, respectively.5 There
are also three peaks that are due to the Pb. The lattice spacings were compared
with lattice spacings of reference materials: Pb, PbO1_44, and PbO2, see Table
4.3.“M2 The peaks obtained in the experimental sample (i.e., Pb on BDD) do not
directly correspond with any of the reference materials; the closest correlation is
with PbO1,44. There are peaks that are missing (e.g., 2.6 — 2.8 A range), and the

intensity ratios do not match the reference material. Notice that the different

lattice spacings correspond to different crystallographic orientations of the

material (e.g., 111). The reference spectra were taken of powdered materials
where the sample in this case was deposited on a surface. When the Pb
deposits on the surface certain orientations may be preferential which would
explain why all of the peaks from the reference material are not present or are at
different ratios in the experimental results. In summary, it appears the Pb
deposited on the BDD surface is a mixture of Pb and PbO2. The electrode was
removed from the electrochemical cell after electrodeposition and the Pb
deposits were exposed to air so it is likely that some amount of Pb oxidation
occurred prior to XRD analysis.

Table 4.3. Powder X-ray diffraction data for reference materials for Pb, PbO1,44,
and PbO2. d is the lattice spacing, intensity indicates the relative intensity of the

peaks, and hkl is the crystallographic orientation. The entries that are bolded
and italicized are closest to the peaks seen experimentally.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1:11110 PbO1,4412 PbO2"

d(A) Intensity hkl d(A) Intensity hkl d(A) Intensity hkl
2.86 100 111 3.16 100 111 3.09 100 111
2.48 50 200 2.73 80 200 2.67 so 200
1.75 31 220 1.93 65 220 1.89 80 220
1.49 32 311 1.65 70 311 1.61 so 311
1.14 10 331 1.53 11 222 1.54 30 222
1.26 14 331 1.34 30 400

1.22 17 420 1.23 60 331

1.05 11 511 1.20 80 420

1.09 60 422

 

 

Two more samples were then generated, first, was a diamond electrode
that had Pb electrochemically deposited and stripped (ASV) in a solution that

was purged with N2, the second was created the same way but the solution was

65

not purged with N2. These two samples were analyzed only in the regions where

Pb or BDD peaks were expected. The results are displayed in Table 4.4.

Table 4.4. Comparison of X-ray counts for peaks from XRD spectra of four
diamond electrodes, bare, with Pb, Pb stripped - N2 purge, and Pb stripped - no
N2 purge. The - indicates no peak seen above the background. N/A indicates
this region was not scanned

 

 

 

 

 

 

 

 

 

X-ray counts
d (A) Bare With Pb Psztripped - Pb stripped —
2 purge no N2 purge
3.12 (Pb) 20 262000 133 -
2.06 (BDD) 43 110 N/A 43
1.57 (Pb) - 4940 - N/A
1 .26 (B DD) 130 170 N/A N/A
1.07 (BDD) 12 - N/A 13
1.05 (Pb) - 179000 63 -

 

 

 

 

 

 

The electrode that was prepared by depositing and stripping the Pb in a
N2 purged solution showed a small amount of Pb. The peaks, corresponding to
d-spacing of 3.12 and 1.05 A, were less than 1% of the size of the peak when Pb
was deposited on the diamond (see Table 4.4). While XRD is only semi-
quantitative this still suggests that there is only a small amount of residual Pb.
For the electrode prepared by depositing and stripping the Pb in a solution that
was not N2 purged all peaks due to Pb were below the limit of detection. BDD
peaks were still visible on this electrode, and were comparable to the signal seen
for the bare BDD, indicating that the instrument was functioning properly, and the
sample was mounted correctly. Therefore, there was no Pb detected in the case

where no N2 purge was used. This suggests that the lower current observed

66

when no N2 purge was used results from chemical removal (oxidation) of the Pb
from the surface of the electrode prior to electrochemical stripping.
4.3. Development of a Sampling Procedure

As Indicated above, to make the technique practical for on-board
monitoring, several adjustments were necessary. First, a method for introducing
the buffer, adjusting the pH, and possibly remediating the effect of the dissolved
02 present. Also a cell needed to be designed for use in microgravity
environments.
4.3.1. Addition of the Supporting Electrolyte

An acetate buffer concentrate was developed that could be used to adjust
the solution pH and add supporting electrolyte to the water sample. The buffer
concentrate consisted of a high concentration of acetate salt (8000 ppm) in a 5%
acetic acid solution. The composition of the solution was determined semi-
empirically. The goal was to add 100 uL buffer concentrate to a 5 mL water
sample, since this would only cause a 2% change in overall volume. The
Henderson-Hasselbalch equation was used to calculate the amount of acetate
needed to create a solution of pH 5.2 with an overall electrolyte concentration of
0.1 M. The solution was prepared, and the amount added was adjusted
experimentally until the proper pH was reached. The final amount was 147 uL
added to a 5 mL sample.

The second solution issue to address was the effect of the dissolved 02
on the stripping voltammetric current, particularly for Pb. While a linear

calibration curve was observed for the detection of Pb” even without an N2

67

purge, the minimum detectable concentration was 50 ppb (S/N 2 3). This
concentration is too high to be useful to assess toxicity of potable waters, as the
limit of detection needs to be at or below the EPA drinking water standard. An
approach that was explored involved the addition of an oxygen scavenging
molecule to the solution to chemically remove the 02 before analysis. Ascorbic
acid was selected because it is a good 02 scavenger that has been used
previously for anodic stripping analysis.13 Also, it is a non-toxic molecule; in fact,
it is a common health supplement. Therefore, its addition to the water supply
should pose no health problems. Based on previous use of ascorbic acid for
stripping voltammetry a concentration of 0.1 M was chosen.13 The ascorbic acid
was added to a solution containing 100 ppb Pb” in the standard acetate buffer
solution (pH 5.2). Anodic stripping voltammetric i-E curves were recorded in this
solution, both with and without a N2 purge. These voltammetric i-E curves were
then compared to ones recorded without ascorbic acid (Figure 4.3).

From the figure, it is easy to see that ascorbic acid does have a positive
effect on the Pb stripping current when dissolved O2 is present. However, the
overall signal with an N2 purge is not as high as the case where no ascorbic acid
is added. While the cause of this is unknown, it is possible that the high
concentration of ascorbic acid (0.1M) is affecting the deposition of the Pb” either
by complexing the Pb” in the solution phase or by adsorbing to the deposited Pb

and preventing further deposition at that site.

68

4.0 -

 

 

 

 

 

 

4'0 T A B — N2 purge
3.5 . n 3.5 . - - N0 N2 purge
3.0 - 3,0 4
5? 2-5 - 2‘ 2.5 «
EL- 3
E 2.0 ' E 2 0 c1
2 E I
1.0 - 1.0 - I “
\ /
0.5 1 I / (l 0.5 - \
0.0 . r — — l 0.0 . . — .
-0.7 06 -0.5 04 -0.7 -0.6 -0.5 -0.4
Potential vs. AglAgCI (V) Potential vs. AglAgCI (V)

Figure 4.3. Anodic stripping voltammetric i-E curves for Pb were recorded in a
solution containing, 100 ppb Pb” in acetate buffer, pH 5.2, with 30 ppm Ca” and
5 ppm Mg”. (A) no added ascorbic acid and (B) 0.1 M ascorbic acid.

4.3.2. Cell Design

An electrochemical cell was designed as a prototype for use in
microgravity (Figure 4.4). When assembled, the cell was fully sealed, and had 3
Iuer ports with valves that served as inlets and outlets for the solution. The
working electrode was secured between a viton o-ring and a copper cylinder for
backside contact. A sleeve of Teflon was placed around the copper cylinder and
was threaded to screw the assembly to the body of the cell. The reference
electrode port was a swagelok fitting in the Plexiglas cover that was intended to
be used with a “no-leak” Ag/AgCl reference electrode (EE-009, Cypress
Systems, Inc., Lawrence, KS). The assembly was a cube with a side length of

46 mm; the inner chamber was also cubic with a side length of 18 mm. The inner

69

volume was 5 - 6 mL. The auxiliary electrode was a coil of Pt wire sealed in the
Teflon side wall.

Three inlet and outlet ports were placed in the cell to enable easy loading,
replacement, and rinsing of fluids. The Iuer fittings were used for ease of
interfacing with NASA’s current water system and sample bags. The cell could
be used either with syringes to manually fill and empty the cell, or by connecting
the system to a pump for automated flow.

Proper mixing of the water sample with the supporting electrolyte is a
critical step with this cell. This was accomplished in the following manner. Two
syringes were used; one contained 300 uL of buffer concentrate and the second
contained 10 mL of water sample. The two syringes were connected using a
male-to-male Iuer connection. The buffer concentrate was injected into the larger
syringe containing the water sample, allowing the two solutions to mix. The
syringe containing the combined solution was attached to an open inlet valve on
the cell. A second valve was opened to allow the intemal air to be displaced by
the sample solution. The water solution was introduced into the cell slowly to
minimize the entrapment of air. Once the cell was filled any remaining air was
dislodged from the walls by tapping. The collected air was then displaced with
additional sample introduction. Once the air was completely removed the values

were closed, and the cell was ready for analysis.

70

Swage fitting
for reference

 

Plexuglas wmdow Teflon screw

   

Copper current

 

   
  

, collector
O-ring gasket \ “I" Auxiliary (Pt)
Luer fitting (inlet/outlet)
Luer valve
Top view
Copper current
collector

Teflon screw

l Teflon body

Working electrode

O-ring
( ‘/4 in. inner dia.)

Cut away view of worklng electrode

Figure 4.4. A diagram of the sealed electrochemical cell designed for use in
microgravity. The body is made of Teflon. Three Iuer fittings with valves serve
as inlets and outlets. The planar working electrode is held in place by a copper
cylinder inside a Teflon screw. The working electrode area is defined by a Viton
o-ring (0.32 cmz).

71

Current (uA)

 

 

 

 

0 0.2 0.4 0.6
Potential vs. AglAgCI (V)

Figure 4.5. Anodic stripping voltammetric i-E curves for Ag in the microgravity
prototype cell. Measurements were made using Ag+ solutions ranging in
concentration from 25 — 500 ppb. Buffer concentrate was used to add
electrolyte, the concentration in the final solution was 0.1 M acetate, pH of 5.2.
The deposition time was 3 min.

A calibration curve was constructed for Ag+ using this procedure to
determine the sensitivity of the method. The calibration curve was linear (r2 >
0.99) for both current and charge (area) with sensitivities of 0.8 nA/ppb and 47
nC/ppb, respectively. The stripping voltammetric i-E curves used for the
calibration curve are presented in Figure 4.5. In addition to the microgravity, cell
a battery operated potentiostat (PG580, Princeton Applied Research, Oak Ridge,
TN) was used to demonstrate the usefulness of the technique with equipment

that could be used on space station. Because this is a smaller model and is

battery operated it does not contain the filtering capability of a bench model

72

potentiostat, and this contributed to the noise seen in the voltammetric i-E
curves.
4.4. Analysis of Simulated ISS Water

Simulated potable and technical water samples were provided by the Wyle
Laboratory (Johnson Space Center) for analysis. The potable sample contained
both Ag+ and the Ca” and Mg” forrnate salts, and the technical sample
contained just Ag+. These simulants were evaluated using two cells: the
traditional single compartment 3-electrode cell with the electrode mounted from
the bottom and the microgravity prototype cell. Calibration curves were
generated, using standard solutions of Ag“, ranging in concentration from 350 to
550 ppb. All solutions were prepared using deionized water. The acetate buffer
concentrate was used as the supporting electrolyte and for pH adjustment. In the
open cell, the acetate buffer concentrate was added directly to the cell containing
the water sample (147 pL concentrate per 5 mL water). The simulant samples
were analyzed the same way as the calibration standards, and the
concentrations were determined from the standard calibration curves. The
results for each of the cells are displayed in Table 4.5.

Using both cell types, the Ag“ concentration determined experimentally
was close to the NASA specified concentration with an error of less than 5%.
The exception to this is the result for the technical water in the standard cell.
These results demonstrate two things. First the ASV method using a buffer
concentrate provides accurate values for the Ag” concentration. Second, the

microgravity cell functions well for this measurement. The standard deviation for

73

the concentrations obtained with the microgravity cell are a bit larger, however,
the portable potentiostat was used with this cell, and the slightly higher
background noise of this instrument and could have contributed to the larger

standard deviations.

Table 4.5. Results from the analysis of Ag” in NASA simulated potable and
technical water samples. Each water sample was measured in both the open
cell as well as the microgravity cell. The error is the standard deviation of
concentration calculated by 3 runs. The % error is calculated with respect to the
specified value by NASA, which was determined by lCP-MS.

 

 

 

 

 

 

 

 

 

 

NASA (ICP-MS) Open Cell % Error Microgravity Cell % Error
(ppb) (ppb) (9pr
Simulated Potable Water
422 I 432 i 4 l 2 % I 432 i 13 l 2%
Simulated Technical Water
480 T 530 :l: 5 | 10% I 506 :1: 15 | 5%

 

 

4.5. Long-Term Response Stability

The last performance criterion evaluated was the long-term stability of the
electrode response. Since replenishing missions are months apart, it is
important that the electrode maintain a steady signal over this time period. To
evaluate this, a standard water sample containing 400 ppb Ag+ was prepared in
deionized water, and analyzed three times per week. A single electrode was
used for all tests, and the electrode remained mounted in the electrochemical cell
during this period to ensure that the same area was used in each measurement.
The cell used was a plastic version of the open cell made of high density
polyethylene (HDPE). The water sample (5 mL) was added to the cell and mixed
with a small amount (147 1.1L) of acetate buffer concentrate (which also contained

Ca” and Mg” salts). At the end of each run a cleaning step was performed,

74

holding the electrode at a potential of 0.6 V with convection (N2 bubbling) for 5
min to ensure a clean surface for the subsequent analyses. The electrode was
stored in one of two ways, wet or dry. For the wet storage, deionized water was
added to the cell at the end of the analysis. Dry storage involved removing all of
the liquid from the cell and storing the electrode exposed to the atmosphere. The

results from each of these stability tests are shown in Figures 4.6 and 4.7.

3.00

 

 

 

 

 

 

 

 

5.00
0 I Current

2,50 :1 D 0 0 Charge -. 4.00
2 c] C] 6
32.00 ~ 0 0 U D 3
.5 0 1:1 00”“ 60 00° 0‘36 ~— 3.00 4!»,
c€1.50JIII 0 ‘3 ”on” U” ”U D 2
5 D ' " + 2 oo 0
x . I -I ' x
81'00 al.." In a .‘I- I g
0- .d - .-. I O.

0.50 - ~~ 1.00

0.00 . . 1 r 0.00

0 20 40 60 80 100
Time (days)

Figure 4.6. Stability test of a 400 ppb Ag+ solution using a single diamond

electrode. The electrode was stored wet, under deionized water, between
measurements.

75

 

 

 

 

 

 

5.0 f 12.0
I Current

4.0 -0 c1 Chargei 4 10,0
f; 1:] 8.0 3:;
g 3.0 {0.0 DU 1:: 2..”
t c- 13 I c1 1:) -- 50 0
5 I I. II . 5

2.0 ~' an
if: 0 C, U [,3 1:] a 4 4.0 T3
52 ' 0000 0 D 00 c1- 20 0 33

1'0— .2..-I-Il"“:- .- 'I I 1» 2.0

I
0.0 . . r 1 D 0.0
0 2O 40 60 80 100

Time (days)
Figure 4.7. Stability test of a 400 ppb Ag+ solution using a single diamond

electrode. The electrode stored dry between measurements.

In the wet storage stability test (Figure 4.6) the response was very stable
over the entire 90 day period with relative standard deviations of 9.3% and 9.1%
for peak current and charge respectively. The dry storage had some fluctuations
in the data in the initial 35 days; however, once the response stabilized the
relative standard deviations were comparable to what was obtained with the wet
storage method: 11.7% and 10.4% for peak current and charge respectively.
Both the wet and dry storage had the same average value for peak current (1 (1A)
and peak charge (3 1.10). It appears that the wet storage is a better method
because it avoids the initial stabilization time, and the relative standard deviations

are a little lower.

76

4.6. Conclusions

A ground-based ASV method was modified for possible on-orbit use. In
general, BDD working electrodes provide low detection limits, excellent response
precision, and good response stability for several priority metal ions, including
Ag“. The ability of the method to provide quantitative information on the Ag“,
Cu”, Pb”, and Cd” levels in water was not compromised by increasing the
supporting electrolyte pH, mineralizing the water with Ca” and Mg” salts, or
eliminating the N2 purge step to remove dissolved O2. The exceptions for the
effect of dissolved O2 were Pb”, and Cu”, however, the use of ascorbic acid
offered a promising mitigation to this problem. The electrode exhibited good
long-term response stability, observed for Ag“ (<10% relative standard deviation
in response over 90 days). lmportantly, preliminary measurements of simulated
transfer water samples revealed that the diamond thin-film electrode provides an

accurate measurement of Ag+ in the microgravity prototype cell.

77

4.7. References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(3)
(9)

(10)

(11)

(12)

(13)

Plumlee, D. K.; Mudgett, P. D.; Schultz, J. R. SAE Technical Paper 2002-
01-2537 32"“ lntemational Conference on Environmental Systems 2002.

Mudgett, P. D.; Benoit, M. J.; Schultz, J. R. SAE Technical Paper 2002-
01-2532 32"" lntemational Conference on Environmental Systems 2002.

Plumlee, D. K.; Mudgett, P. D.; Schultz, J. R. SAE Technical Paper 2003-
01-2401 33" lntemational Conference on Environmental Systems 2003.

Swain, G. M. In Electroanal. Chem; Bard, A. J., Rubinstein, l., Eds.;
Marcel Dekker, Inc., 2004; Vol. 22, pp 181-277.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chem. Mater. 2003,
15, 879.

Sonthalia, P.; McGaw, E.; Swain, G. M. Anal. Chim. Acta 2004, 522, 35-
44.

Wang, J. Stripping Analysis Principals, Instmmentation, and Applications;
VCH Publishers: Deerfield Beach, Florida, 1985.

Lund, W.; Onshus, D. Anal. Chim. Acta 1976, 86, 109-122.

McGaw, E.; Sonthalia, P.; Swain, G. M. SAE Technical Paper 2005-01-
2888 35th lntemational Conference on Environmental Systems 2005.

In Powder Diffraction File; McClune, W. F., Ed.; lntemational Centre for
Diffraction Data: Swarthmore, PA, 1986, pp 4-686.

In Powder Diffraction File; McClune, W. F., Ed.; lntemational Centre for
Diffraction Data: Swarthmore, PA, 1986, pp 22-389.

In Powder Diffraction File; McClune, W. F., Ed.; lntemational Centre for
Diffraction Data: Swarthmore, PA, 1986, pp 27-1201.

Florence, T. M.; Farrar, Y. J. Electroanalytical Chemistry and Interfacial
Electrochemistry 1973, 41, 127-1 33.

78

CHAPTER 5
ANALYSIS OF HEAW METAL IONS IN URINE
5.1. Introduction

Analysis of heavy metals in biological fluids can be an important tool for
diagnostics as well as monitoring of occupational exposure. High levels of toxic
metals in humans can indicate a single or a prolonged exposure, and it is
important that there are clinical tools for the analysis of these analytes in humans.
Depending on the properties and interactions of the metal ions in the body, they
will be excreted in different ways, and will be present at different concentrations
in different body fluids. For example, Cd” is more likely to appear in the urine,
while Pb” is more likely to appear in the blood.1 There are also essential metals
(Zn”, Cu”) that can be found in the urine.1 It is important to monitor these
metals as well, since increased or decreased levels can indicate a problem. For
instance, elevated Cu” levels in the urine can be an indicator of Wilson’s
disease.

Anodic stripping voltammetry is a common method for determination of
metal ion concentration. Traditionally this type of analysis has been done using a
Hg based electrode (e.g., hanging drop, thin-film); however, there have been
recent attempts to replace the Hg based electrodes with a more environmentally
friendly electrode material. Boron-doped diamond (BDD) is a relatively new
electrode material that is being developed for stripping analysis by many groups.
As discussed in Chapter 3, BDD is comparable to a more traditional Hg-coated

glassy carbon electrode in a simple aqueous media.2

79

Atomic adsorption spectroscopy”5 and anodic stripping voltammetry
(ASV) are commonly used methods for analysis of urine samples. For ASV it is
typical that a digestion step is used because of the soluble proteins present in the

d6—9

urine. Many digestion procedures, including hot aci acid and ultrafiltration,10

""3 acid digestion bombs,“ '5 dry ashing.16 and freeze

microwave digestion,
drying,17 have been used successfully. Some have even used a polymer coating
on the electrode to prevent interferences from compounds in the urine.” ‘9
Direct determination of metal ion concentration in urine has been done using both

”'22 and glassy carbon electrodes.”

Hg-based electrodes

In this study, rat urine and urine simulant were analyzed by ASV without
digestion or acidification. There were several factors in this decision: ease of
analysis, the naturally high electrolyte concentration of the urine, and the fouling-
resistant properties of the BDD, which make it a suitable candidate for the direct
analysis of the urine. Also in preliminary studies the urine was acidified,
however, this resulted in an interfering peak in the potential range of Cd oxidation.
The identity of this peak is not known, however, it is has also been observed by
Pauliukaite and co-workers.9

Cd” was the metal ion chosen as the focus of this study because the
urine concentration of Cd” is the most important clinical indicator of excessive

cadmium exposure, whereas other metals are typically analyzed in the blood

(e.g., Pb”).1

80

5.2. Deposition Parameters

Toxic levels of Cd” in urine are in the low ppb to high ppt range. For
example a study by Yassin and Martonik reported levels ranging from 0.01 ppb
(normal) to 15.57 ppb (toxic).5 Therefore, low detection limits are necessary for
this analysis. The ASV detection limit of Cd” in an acetate buffer solution with a
210 s deposition was found to be 1.0 ppb using BDD as a working electrode (see
Chapter 3). A lower detection limit was desired for this analysis; to lower the
detection limit the deposition parameters (e.g., deposition time and ion flux) were
studied.

Initially, a single solution of 250 ppb Cd” in simulated urine (pH 8.8) was
used, and the deposition time was varied between 2 and 360 seconds (no
convection). The results of this study are shown in Figure 5.1 A. The points are
an average of 3 measurements and the error (standard deviation) is within the
size of the markers.

It is clear that increasing the deposition time caused the peak current to
increase, up to a point. However, the current begins to level off around 200 s
with little enhancement occurring with additional time. The deposition conditions
used in previous studies were 180-210 s, and increasing the deposition time
beyond this does not appear to have a significant benefit.

The curve (Figure 5.1 A) begins to level off because the mass transport of
the Cd” to the electrode surface is limited. By adding convection during
deposition the flux of Cd” to the electrode will be enhanced to give higher

stripping peak currents. Several methods of convection were explored: stirring

81

 

 

with a stir bar, stirring with a rotator with a variety of “fins” attached, and bubbling
with N2 gas. The N2 gas was the most successful method yielding the largest
enhancement of peak current. To regulate the N2 flow, a gas flow meter was
used to ensure the same flow rate was used each day. In addition the tubing that

was used was marked to ensure that the gas line was placed at the same height

with respect to the electrode.

 

 

 

 

0.6 - 5.0 -
A o 0 B -
I
I U D
0 5 D 4.0 -
A A I
g 0.4 - 0 g
g D E 3.0 -
g 0.3 - U 5 I
x x 2.0 - .
‘° 0 2 - a“;
6’. ' o. I With convection
U 0 Without convection 1 0 j . 0 Without convection
0.1 - '
J' n D D D D D D D
0.0 . . . . a 0.0 U . . . .
0 100 200 300 400 0 100 200 300 400
Deposition Time (s) Deposition Time (s)

Figure 5.1. Peak current as a function of deposition time for 250 ppb Cd” in
urine simulant (pH 8.8). Deposition time varied between 2 s and 5 minutes. A)

no convection method used, B) N2 bubbling used for convection. Each point is
an average of 3 measurements.

Using N2 as a convection method the deposition time was again varied
(Figure 5.1 B). The first observation is that the current is much higher than the
case without convection (plotted on same the graph for comparison). For
example, at 60 s the peak current was 0.26 pA without convection versus 1.0 uA
with convection. Based on these experiments, a deposition time of 300 s (with

convection) was selected for use in analysis.

82

 

 

5.3. Detection Figures of Merit

To characterize the BDD electrode for use directly in a urine sample, the
detection figures of merit in the urine simulant were determined. A calibration
curve was generated for concentrations in the 10 to 250 ppb range (Figure 5.2).
The calibration curve is linear (r2 > 0.99) for both peak current and peak charge
(area) with sensitivities of 27.3 nA/ppb and 99.7 nC/ppb, respectively. As
expected the sensitivity in the urine simulant is higher than the sensitivity
reported in standard acetate buffer at pH 5.2 (14.7 nA/ppb) due to the longer
deposition time and added convection.2 The linear dynamic range was not
specifically investigated, however, linearity in the response was observed over 2
orders of magnitude in the range tested (10 — 250 ppb).

The lowest concentration of Cd” detected in the urine simulant with a
signal/noise (SIN) > 3 was 10 ppb (S/N=5). This is higher than the 1.0 ppb
reported for the standard solutions in acetate buffer pH 5.2.2 While the sensitivity
of this measurement is higher than in the acetate buffer, the noise is also larger
in the urine simulant, 6 nA vs. 2 nA. The limit of detection is a function of both
the noise as well as the sensitivity. During the stripping step in the urine
simulant, the N2 was still bubbling and this added to the background noise. The
noise could be lowered by removing the N2 line before the stripping step but
because this gas line was place in the cell manually it was difficult to reproducibly
remove it just prior to the stripping step. Since the convection from the bubbling
enhances the deposition, if it was not moved at exactly the same time it lead to

problems with the reproducibility of the stripping curves. Therefore, it was

83

ET)...

 

 

decided that the increased noise level was more tolerable than the poor
reproducibility. However, if an automated convection method was used it could
be reproducibly shut off before stripping, and it is expected that the noise would

decrease also leading to a decrease in the limit of detection.

   
   

 

 

Current (uA)
N (.0 H 01
Peak Current (11A)
.5

 

 

 

 

0 I I r fl 0‘ I r I
-1 -O.9 -O.8 -0.7 O 100 200 300

Potential vs AglAgCI (V) Concentration (ppb)

 

Figure 5.2. Anodic stripping curves (left) and calibration curve (right) for urine
simulant (pH 8.8) with Cd * concentrations in the range of 10 to 250 ppb.
Deposition occurred at -1.0 V vs. Ag/AgCl for 300s with convection (N2 bubbling)
Error bars on calibration curve represent standard deviation.

For the method to be useful, it is important to have run-to-run I

reproducibility as well as electrode-to-electrode reproducibility. This was tested

 

on 3 nanocrystalline diamond electrodes, 3 runs each, using 75 ppb Cd” in the L
urine simulant. The average peak currents and standard deviations are shown in
Table 5.1 for each of the electrodes as well as overall. The run-to-run variability

is described by the relative standard deviation for each electrode, and is less

than 5% for each. The electrode-to-electrode variability is also excellent at only

6%.

Table 5.1. Reproducibility of Cd” signal on 3 BDD electrodes. 75 ppb Cd” in
simulated urine (pH 8.8).

 

 

 

 

 

 

 

 

 

Average Peak Relative Standard
Current (uA) Deviation
Electrode 1 (n=3) 1.08 t 0.03 2.72%
Electrode 2 (n=3) 0.95 t 0.01 0.65%
Electrode 3 (n=3) 1.01 :t 0.05 4.44%
3 Electrode Avergagejn=9) 1.01 :I: 0.06 6.15%

 

5.4. Standard Addition

Real urine samples posses a complex matrix and variable pH so the Cd”
concentrations are best determined by the standard addition method. For the
standard addition, 10 ppb Cd” in urine simulant was used as the initial solution,
and two additions to create 20 and 30 ppb Cd” were made for each sample. A
solution with a high concentration of Cd” was added to cause a minimal change
in overall volume. The change in volume was accounted for in plotting the added
Cd” concentration. The concentration of the initial urine (10 ppb Cd”) was
determined by this procedure to be 1.7 ppb, an error of 83%. One possible
source for this error is the complexation of Cd” by other constituents in the
sample that would reduce its activity.
5.5. Cadmium Complexation

If components of the urine simulant complex with the Cd”, then the
stripping voltammetric current will be lower due to less free Cd” available for

deposition. ASV will only yield a response, at the thermodynamic potential, for

85

 

 

free metal ions or metal ions in labile complexes”: 25 Components of the urine
simulant that were tested as possible interferants were lactic acid, citric acid, and
ammonia.

Calculations were performed initially using known formation constants to
determine the amount of free Cd” available when each complexing agent is
present at the concentration used in the urine simulant. To simplify the math,
each component was considered individually. The formation constants used
were obtained from Lange’s Handbook and are shown in Table 5.2.26
Table 5.2. Formation constants for Cd” complexes.26 Listed as cumulative

formation constants where K" is equal to k1k2....kn if k is the stepwise formation
constants.

 

log K1 log K2 log K, log K, lgq K5 M K2,
Ammonia 2.65 4.75 6.19 7.12 6.80 5.14
Citric Acid 11.3 - - - - -

Lactic Acid 1.7 - - - - -

 

 

 

 

 

 

 

 

 

 

 

 

The estimated free Cd” concentration was calculated for several Cd”
solutions ranging from 25 to 100 ppb Cd”, and the results are displayed in Table
5.3 as the percentage of the total Cd2+ dissolved in the solution that is free Cd”.
It is clear that both the ammonia and citric acid have a major impact on
decreasing the amount of free Cd” in solution.

These calculations were followed up by additional experiments that tested
the effect of each complexing agent separately in an acetate solution at pH 8.8;
the same pH as the urine simulant. Each solution contained 30 ppb Cd” The
resulting voltammetric i-E curves are displayed in Figure 5.3. The data show

there is a significant effect from each of the complexing agents, with all three

86

 

 

causing a decrease in the peak current and charge as compared to the control.
Ammonia and citric acid caused the stripping peak to shift slightly negative. The
relative effect (ammonia > citric acid > lactic acid) is the same trend suggested
by the calculations (Table 5.3).

Table 5.3. Calculated concentration of free Cd” in the presence of complexing
agents: ammonia, citric acid, and lactic acid.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ammonia (0.1 M a
[Cdlmllzpbl 4001000) 100101100... "
25 0.02 0.07%
50 0.04 0.07%
75 0.06 0.07% '
100 0.07 0.07% "
Citric Acid (3 mM)
[Cdlioeltlapm [(3th (ppb) 10d” lfreeIICdItotal
25 0.24 0.95%
50 0.48 0.95%
75 0.71 0.95%
100 0.95 0.95%
Lactic Acid (4 mM
[Cdltol_al(PPb) [Cdz free (ppb) [Cd2+]free/[Cd]total
25 20.8 83.3%
50 41.7 83.3%
75 62.5 83.3%
100 83.3 83.3%

 

 

 

 

 

The effect of increasing the pH (8.8 in urine simulant compared to 4-5 in

 

acetate buffer) was also considered as a cause for the inaccurate determination f

of Cd” concentration in the urine simulant. However, based on the Pourbaix

 

i,
diagram, increasing the pH from 4 to 9 at the deposition potential of -1.0 V vs. '
AglAgCI, has no effect on the Cd” concentration (i.e., Cd(OH)2 formation).27
This was confirmed by making anodic stripping voltammetric measurements in
acetate buffer solutions ranging from pH 5 to 9. There was no change in the

Cd” stripping peak potential, current, or charge.

87

 

 

" —Buffer
0 7 — Lactic Acid
- ‘ - - Citric Acid
0.6 . - - Ammonia
2?
3 0.5 - ’ ..
g 0.4 ' ’ I O . ‘~ ~
0 - y . , , ‘ .
0.3 '.’ ' 0’ “ . ~ '''''
I . . ' ' ‘ ~ ~
0.2 . ‘ ‘ - - - .
0.1 . . . .
-1 -0.9 -0.8 -0.7

Potential vs. AglAgCI (V)
Figure 5.3. Anodic stripping voltammetric i-E curves for Cd in an acetate buffer,

pH 8.8, containing 0.1M ammonia, 3mM citric acid, or 4mM lactic acid. 30 ppb
Cd2+ deposited for 180 s with convection (N2 bubbling).

Because complexation causes such a major decrease in the Cd stripping
response, it is necessary to remove these interferences before analysis. This
can be accomplished with a digestion. Several methods have been successfully
used including hot acid,M acid and ultratiltration,10 microwave digestion,"'13 or

acid digestion bombs,“ ‘5 as mentioned previously. Even though the BDD

electrode may not be specifically affected by these components, they affect the

electrochemical signal for Cd”.

88

 

 

 

5.6. Real Sample Analysis

To test the utility of the method on a real sample, rat urine was obtained
for testing. The filtered (PTFE syn‘nge filter, 0.1 pm) rat urine was directly
analyzed with the diamond electrode. This curve is shown in Figure 5.4. No Cd
stripping response was observed in the urine alone. Measurements were then
made on urine samples spiked with higher concentrations of Cd”. Curves for
concentrations between 90 and 300 ppb are also shown in Figure 5.4. Around a
potential of -0.8 V an increase in current is seen indicating that the electrode is
responding to the increasing Cd2+ concentration. The sensitivity of the
measurement is poor as only a small response is seen at a high concentration of
300 ppb.

Because the rat urine contains other components (e.g., proteins) that are
not present in the urine simulant, it was also important to evaluate the stability of
the electrode response to check for any fouling effects. To accomplish this rat
urine containing 300 ppb Cd2+ was added to the cell and several sequential runs
were performed over a 6 hour time period (Figure 5.5).

There was no evidence of any electrode fouling over the course of the
measurements. After the initial 20 measurements the response stabilized, and
the relative standard deviation (RSDs) for runs 20 - 35 was 3.1% for current and
3.2% for area. These values indicate that the diamond electrode is very stable

even in a complex matrix.

89

 

)7-

 

300 ppb

MM v

90 ppb
:l:0.2 uA
r , ' urine '
-0.95 -0.85 -0.75 -0.65

Potential vs. AglAgCI (V)

Figure 5.4. Anodic stripping voltammetric i-E curves in rat urine. Bottom curve is
rat urine alone and aliquots of Cd2+ were added to increase the Cd2+
concentration to the range of 90 -300 ppb. Curves are offset for clarity.

 

 

 

 

0.10 - T 0.60
. I. -
_ . I. II" II." 0.50
0.08 .-... .-l I.
A ' ' ' -- 0.40 A
3 0_06 4 I.. I DUDE-DDDDUDDDDDDDDDQ (:31
8 000000.30 __ D D -- 0.30 8
S 0.04 - a D D 2
0 C10 " 0.20
I Current
0.02 0 Area .- 0_10
0.00 . . . 0.00
O 10 20 30
Run Number

Figure 5.5. Stability test performed in 300 ppb Cd” in real rat urine.
Measurements are sequential runs taken over a time period of approximately 6
hours. Deposition conditions, 300 s with convection (N2 bubbling).

90

 

 

 

5.7. Conclusions

The Cd stripping signal can be enhanced by increasing the deposition
time as well as adding a convection step to enhance the Cd2+ flux to the surface.
The best deposition time was determined to be a 300 s with N2 bubbling for
convection. In the urine simulant a linear calibration curve was obtained over a
range of 2 orders of magnitude (10—250 ppb), and the response was
reproducible electrode-to-electrode (RSD 6%). A digestions step, or other
purification step, to remove complexing agents, is imperative to obtain accurate
Cd2+ concentrations. For the urine simulant, it was found that at least three
components, ammonia, citric acid, and lactic acid, have a negative effect on the
electrochemical signal by reducing the Cd2+ activity. This causes a higher limit of
detection and inaccurate determination of the Cd2+ concentration by standard
addition. A possible solution to this would be to digest the urine simulant prior to
analysis, either using a hot acid digest or a microwave digest. The digestion
would destroy the organic and inorganic molecules that were complexing the
Cd2+ and lead to a more accurate determination of Cd”.

The diamond electrode performed adequately in a rat urine sample. An
increased signal was observed for increasing concentrations of Cd”. The
electrode showed no fouling effects from proteins or biological molecules in the
real sample over a 6 hour period. Attempts were also made to determine the
Cd2+ concentration in rat urine after centrifugation and filtration. Even though the
electrode response was stable and reproducible, minimal sensitivity was

observed for the metal ion.

91

 

 

 

5.8. References

(1) Goyer, R. A. In Casarett & Doull's Toxicology: The Basic Science of
Poisons, 5th ed.; Klaassen, C. D., Ed.; McGraw-Hill: New York, 1996, pp
691-736.

(2) McGaw, E.; Swain, G. M. Anal. Chim. Acta 2006, 575, 180-189.

(3) Sharma, R. P.; McKenzie, J. M.; T., K. J. Anal. Tax. 1982, 6, 135-138.

(4) Castro Maciel, C. J.; Miranda, G. M.; Oliveria, D. P.; Siqueria, M. E. P. B.;

 

Silveria, J. N.; Leite, E. M. A.; Silva, J. B. B. Anal. Chim. Acta 2003, 491, 3“
231-237. i
(5) Yassin, A. S.; Martonik, J. F. J. Occup. Environ. Hyg. 2004, 1, 324-333.
(6) Searle, 8.; Chan, W.; Davidow, B. Clin. Chem. 1973, 19, 76-80.
(7) Lund, w.; Eriksen, R. Anal. Chim.Acta1979, 107, 37-46. gel

(8) Jeong, E.-D.; Won, M.-S.; Shim, Y.-B. Electroanal. 1994, 6, 887-893.

(9) Pauliukaite, R.; Metelka, R.; Svancara, I.; Krolicka, A.; Bobrowski, A.;
Vytras, K.; Norkus, E.; Kalcher, K. Anal. Bioanal. Chem. 2002, 374, 1 15-
1 158.

(10) Yantasee, W.; Timchalk, 0.; Lin, Y. Anal. Bioanal. Chem. 2007, 387, 335-
341.

(11) Pan, T. C.; Homg, C.-J.; Lin, 8. R.; Lin, T. H.; Huang, C. W. Biol. Trace
Elem. Res. 1993, 38, 233-241.

(12) Homg, C.-J. Analyst1996, 121, 1511-1514.

(13) Homg, C.-J.; Homg, P.-H.; Hsu, J. W.; Tsai, J.-L. Archives of
Environmental Health 2003, 58, 104-110.

(14) Kinard, J. T. Anal. Lett. 1977, 10, 1147-1161.

 

(15) Onar, A. N.; Temizer, A. Analyst 1987, 112, 227-229.
(16) Newberry, C. L.; Christian, G. D. J. Electroanal. Chem. 1965, 9, 468-472.

(17) Golimowski, J.; Valenta, P.; Stoeppler, M.; Numberg, H. W. Talanta 1979,
26, 649-656.

(18) Hoyer, 8.; Florence, T. M.; Batley, G. E. Anal. Chem. 1987, 59, 1608-
1614.

92

(19)

(20)
(21 )

(22)

(23)
(24)

(25)

(26)

(27)

Kefala, G.; Economou, A.; Voulgaropoulos, A. Analyst 2004, 129, 1082-
1 090.

Kemula, W.; Kublik, 2. Nature (London) 1961, 189, 57-58.

Copeland, T. R.; Christie, J. H.; Osteryoung, R. A.; Skogerboe, R. K. Anal.
Chem. 1973, 45, 2171-2174.

Herbeuval, X.; Maso, J. L.; Baudot, P.; Hutin, M. F.; Burnel, D. Pathol.
Biol. 1975, 379-386.

Ramadan, A. A.; Mandil, H.; Asaad, W. Asian J. Chem. 2006, 18, 50-56.

Vydra, F.; Stulik, K.; Julakova, E. Electrochemical Stripping Analysis; Ellis
Horwood Limited: Chichester, 1976.

Wang, J. Stripping Analysis Principals, Instrumentation, and Applications;
VCH Publishers: Deerfield Beach, Florida, 1985.

Dean, J. A., Ed. Lange's Handbook of Chemistry, 15th ed.; McGraw—Hill,
1999.

Deltombe, E.; Pourbaix, M.; Zoubov, N. In Atlas of Electrochemical

Equilibria; Pourbaix, M., Ed.; Pergamon Press Ltd.: Oxford, 1966, pp 414-
420.

93

 

 

CHAPTER 6
ANALYSIS OF HEAW METAL IONS IN BLOOD
6.1. Introduction

The analysis of biological fluids for heavy metals is important for clinical
diagnosis and treatment of poisoning. As stated previously, in the discussion of
analysis in urine, different metals will be present in different body fluids
depending on the physiological activity of each. One metal ion that is important
clinically, and that is found in the blood is Pb”. The Pb2+ is biologically similar to
Ca”, and can replace Caz“, or possibly an“, thus affecting synaptic neuro-
transmission.1 Pb2+ can also replace Ca2+ in the bone, weakening it, and the half
life of Pb” in the bone is 20 years.1 The adverse health effects from Pb2+ are
mainly neurological and hematological, but it can also affect reproduction. The
Pb2+ concentrations at which some of these adverse health effects manifest
themselves are displayed in Table 6.1.

Table 6.1. Minimum concentration of sz+ in blood when different health effects
are observed.1

 

Pb” Concentration in Blood mph)

 

 

 

 

 

 

 

Effect Children Adults
Low IQ 100 - 150
Hearing loss 200
Encephalopathy 800 — 1000 1000 - 1200
Nephropathy 400
Reproductive (e.g., sterility) 400
Anemia 800 - 1000 800 — 1000

 

 

 

 

 

Elevated Pb+ levels are particularly dangerous for children because they are

growing and developing. According to the Centers for Disease Control and

94

 

 

 

 

Prevention (CDC), over 300,000 children in the United States have elevated
blood Pb2+ levels (> 100 ppb).2 They are committed to eliminating elevated blood
sz+ levels in children by 2010. The CDC recommends that children at risk of
high Pb2+ levels (e.g., live in a home with lead based paint) be tested yearly.
Because of the large number of children who are tested, there is a need for a
simple and quick measurement that could even be done in a doctor’s office.

Graphite furnace atomic absorption spectroscopy is the most commonly
used analytical technique for Pb2+ analysis in blood.3 However, other methods
may be better suited to high throughput analysis or screening in a doctor’s office.
Electrochemical methods are ideal for this type of analysis because they are
portable, inexpensive, and easy to operate. Anodic stripping voltammetric
analysis of sz“ in blood is becoming more routine, and there is even a
commercial point-of—care testing system, LeadCare, developed by ESA, Inc.
using this technique.“'6

A variety of working electrodes have been used for the analysis of Pb in

7‘9 Bi,‘°' 1‘ and sp2 carbon materials.12 Boron-doped diamond

blood including Hg,
(BDD) has many qualities that make it attractive for this type of analysis including
resistance to fouling and a need for minimal activation pretreatment. The intent
of this work was to examine the utility of BDD for blood analysis.
6.2. Sample Preparation

Anodic stripping voltammetry (ASV) is sensitive to the activity of free metal

ions in solution. The metal ions must not be bound to other solution species, and

must be able to be electrochemically reduced and deposited on the electrode

95

 

 

 

 

surface. More than 90% of the Pb2+ in blood exists in the red blood cells, and is
associated with the cell membrane and the hemoglobin.1 Therefore, a digestion
method is routinely used to lyse the red blood cells, thereby releasing the sz“.
The ESA, Inc. procedure for this uses 0.3 M hydrochloric acid (HCI) mixed with
blood in a 5-to-1 ratio. This mixed solution is then directly analyzed on their
carbon electrode, and the HCI serves as the electrolyte for the electrochemical

analysis. A similar preparation was used in these preliminary studies.

0.9 -
0.8 -
0.7 -
0.6 4

Current (uA)

9.0.0.0
NOD->01
llll

.0
_L
l

 

0.0 . . .
-0.75 -0.65 -0.55 -0.45
Potential vs. AglAgCI (V)

 

Figure 6.1. Anodic stripping voltammetric i-E curves for BDD in 0.1 M acetate
buffer, pH 5.2, and 0.1 M NaCI, with 100 ppb sz+ (top) and without Pb2+
(bottom). The deposition time was 180 s.

In preliminary studies using HCI as an electrolyte, a second stripping

peak, slightly more positive than the Pb peak, was observed. In acetate buffer

alone the background is flat; no second peak is observed in the region of the Pb

96

 

 

_ .1.’v_”

 

peak. Therefore, it was supposed that the peak was related to the Cl'. When CI'
was added to the acetate buffer (0.3 M NaCl) the voltammetric i-E curve did have
a second peak (Figure 6.1). For reference 100 ppb sz+ was added to the same
buffer containing Cl‘, and the peak due to the Cl' is observed adjacent to the Pb2+
peak. This is similar to what is observed when HCI is used as the electrolyte
confirming that this peak is due to CI”.
6.3. Prevention of Fouling

Proteins adsorb on many surfaces; a process that can passivated an
electrode making it unusable for measurement. Measurements in blood are
challenging because of the large number of proteins present that can foul the
electrode. Polar molecules weakly adsorb on BDD because of the relatively non-
polar hydrogen surface termination and the absence of an extended 1T electron
system.“°"16 BDD has been shown to have resistance to fouling from surfactants
and proteins17 making it potentially useful for analysis in whole blood. Even
though the surface of BDD is typically resistant to fouling, preliminary analysis of
l=>b2+ in bovine blood (mixed 1 part blood to 5 parts 0.3 M HCI) yielded a small
response that decreased with each subsequent run. The Pb stripping response
could not be fully regained after exposure to the blood even after a 24 h soak in
distilled water and an additional 20 min soak in isopropanol (Figure 6.2).

It was, therefore, necessary to develop a means to protect the electrode
surface from molecular adsorption. There have been numerous polymeric
coatings used for fouling prevention including dialysis membrane“,

polylysine/polystyrene,19 cellulose acetate,2024 and Nafion.7' 9' 2533 Ultrasound

97

 

 

has been used to lesson the effects of fouling by acting to clean the surface in
situ through cavitational activity of bubbles formed.1o Nafion is a cation exchange
ionomer that has been used as an electrode modification in numerous
electroanalytical applications. It has been used with ASV for the determination of
heavy metal ions in biological matrices.7' 9' 25' 26 Because of the extensive

literature based, Nafion was selected for use in this work.
3.0 -

2.5 -

N
o
I

Current (uA)
;
a.

_L
O
n

0.5- , b

 

0.0 r . .
-0.6 -0.5 -0.4 -0.3
Potential vs. AglAgCI (V)

 

Figure 6.2. Anodic stripping voltammetric i-E curves of 450 ppb Pb2+ in HCI (1
part sz+ solution to 5 parts 0.3 M HCI), before exposure to bovine blood (3),
immediately after exposure to bovine blood (b), after a 24 h soak in distilled water

(c), and after an additional 20 min soak in isopropanol (d). The deposition time
was 180 5.

6.3.1. Nafion Coating

There was no reported literature on Nafion-coated BDD electrodes, so the

first task was to investigate the physicochemiwl properties of the film deposited

on BDD. The electrode was modified with Nafion using a simple drop-coating

98

 

._ ........a

 

method. A 3% Nafion solution was prepared in ethanol, and 60 |JL of this solution
was dropped onto the BDD (~1 cm2). This volume was sufficient to fully cover
the surface. The solution was then allowed to dry for 20 min in a covered beaker
to prevent contamination from particles in the air.23 Care was taken to place the
electrodes on a flat and level surface to ensure that the film dried uniformly.

The thickness of the Nafion film was estimated by weighing the electrode
before and after coating. The Nafion weight (dry) was 0.0015 1 0.0002 g. The
density of hydrated Nafion is between 1.4 and 2.0 g/cm3 (depending on the
source of the Nafion and curing method)?4 Using this range of densities,
knowing the film weight and area (1 cm2), the film thickness was calculated to
range from 7.7 to 11 pm. The uniformity of the film thickness was not specifically
investigated; however, there were minimal refraction patterns on the surface.
The absence of refraction patterns was consistent with a relatively uniform film
thickness.

The negatively charged (highly acidic) sulfonate groups in the interior of
the polymer electrostatically preconcentrate sz“ near the electrode surface.
Because of this preconcentration the amount of metal phase formed during the
deposition step is larger than that in the absence of the polymer. The increased
metal phase loading leads to increased stripping signal (i.e., increased
signal/background). The Pb stripping peak current in blood is plotted as a
function of the preconcentration time for three different Pb2+ solution

concentrations using a Nafion-coated BDD electrode (Figure 6.3).

99

 

 

_x
C
j

 

 

 

 

 

 

‘ 445 ppb Pb2+
I 645 ppb Pb2+
0.8 d A 845 ppb Pb2+ A
2 A
a A
*g' 0.6 . .
l: A I
8
x 0.4 " .
8 A
0. I
0.2 - I .
o o ’ ° .
0 10 20 30
Time (min)

Figure 6.3. Pb stripping current for Pb” in bovine blood as a function of the
preconcentration time. The blood was mixed 1:5 with 0.3 M HCI. The BDD
electrode was coated with a Nafion film. The time axis indicates the amount of
time blood solution was in contact with Nafion coated electrode before deposition
began. The deposition time was 180 s.

The results reveal that longer preconcentration time, prior to deposition,
for a given Pb” concentration, results in a larger stripping peak current. A single
preconcentration time, the stripping current increased with the solution
concentration. The peak current is influenced by the amount of Pb”
preconcentrated in the Nafion film, which increases with time, until the polymer
becomes saturated. The amount of Pb” loaded in the film depends on the flux of
the Pb” into the polymer as well as the amount of time the polymer is in contact
with the Pb” solution. Even though the time is increased the flux will decrease

over time. The flux is related to electrostatic attraction and the concentration

gradient, which will decrease as the concentration of Pb” increases. Therefore,

100

 

 

 

 

there is not a simple linear relationship between electrode exposure time and the
amount of Pb” preconcentrated (i.e., electrochemical response).

It is also important that the Nafion film can be reproducibly deposited on
the electrode surface. The drop-coating method was tested using three
electrodes all cast and cured the same way. Each of these electrodes was then
used to analyze for Pb” (445 ppb) in bovine blood. The resulting stripping peak
currents and charges (areas) are presented in Table 6.2. The relative standard
deviation (RSD) of the response is quite high electrode-to-electrode indicating
that the reproducibility of the coatings is not sufficient. The Nafion-loading needs
to be carefully controlled to produce a reproducible surface. The variation in the
weight of Nafion on the surface had an RSD of 13% (0.0015 :1: 0.0002 g), and is
likely the source of variation. A couple of factors that may need to be better
controlled are the homogeneity of the Nafion solution used for coating, and the
volume applied to the electrode surface.

Table 6.2. Comparison of three electrodes with Nafion coating. Each electrode
was used to analyze blood with 445 ppb Pb”. Two trials were performed on
each film with a short cleaning step in 0.3 M HCI between. The exposure time

was 15 minutes after which a deposition time of 180 s was used prior to ASV
analysis.

 

 

 

 

 

 

 

 

 

 

 

 

Electrode Peak Current (pA) Peak Charge (pC)
1 0.218 0.505
0.161 0.369
2 0.244 0.693
0.177 0.419
3 0.248 0.478
0.310 0.673
Average 0.23 :l: 0.05 0.5 :l: 0.1
Relative Standard Deviation 24% 25%

 

 

101

 

 

 

 

 

Another important point with Nafion-coated electrodes is how well the
cationic analyte can be removed from the polymer. In order to obtain a clean
background (i.e., no Pb signal) in 0.3 M HCI following Pb” exposure a series of
cleaning steps were necessary. The cell was rinsed with 0.3 M HCI three times
then filled with clean HCI, and 0.6 V was applied to the electrode for 3 min to
oxidize any remaining Pb on the electrode. The cell was then rinsed again and a
stripping voltammetric i-E curve was recorded, if a Pb signal was observed this
procedure was repeated. When the electrode had been exposed to Pb” for a
short amount of time (< 30 min) a clean background could be obtained after
repeating the process 2-4 times. However, when the electrode was soaked in
bovine blood containing Pb” for more than 12 h a clean background was not
obtained even after repeating the cleaning procedure 10 times. Therefore, as
long as the Nafion film is not in contact with the Pb” for extended periods of time
a clean background can be obtained.

Another parameter investigated was the electrode sensitivity for Pb”. To
study this, the blood was spiked with increasing concentrations of Pb” and a
stripping voltammetric i-E curve was recorded for each concentration after 30
minutes of exposure to the Nafion film (Figure 6.4). With increasing Pb”
concentration in solution, the stripping peak current increased at -0.55 V. A
second oxidation peak, due to the Cl‘, was present in the voltammograms at a
slightly more positive potential. The blood samples were run one after another
using the same electrode with a cleaning step between runs. The increase in

peak height is nearly linear (r = 0.98) with a sensitivity of 1.6 anpb.

102

 

[0.5 pA
M

 

645 ppb Pb”
445 ppb Pb2+
-0.65 -0.55 -0.45

Potential vs. AglAgCI (V)
Figure 6.4. Stripping voltammetric i-E curves of Pb” in blood at varying
concentrations. The bovine blood was mixed 1:5 with 0.3 M HCI. The electrode
exposure time was 30 min. The deposition time was 180 s followed by ASV
analysis. Curves are offset for clarity.

The cell setup in these experiments has the BDD electrodes mounted
from the bottom with a Viton o-ring that is used to define the area exposed.
When the electrode was removed from the cell, some of the Nafion film was also
removed because it remained attached to the o-ring. This makes re-use of these
electrodes impossible because of polymer film damage, making it necessary to
fully remove Nafion for re-coating.

The solvent used for the colloidal Nafion was ethanol so this was the first
choice for Nafion removal from the electrode surface. The electrode was wetted

with ethanol then wiped with a disposable laboratory wipe to remove the bulk of

the Nafion from the surface. Next, the electrode was sonicated in ethanol for 15

103

 

 

 

 

 

min. Because the diamond surface is rough and consists of many grains and
grain boundaries it is possible that the Nafion may remain in these grain
boundaries. To examine if there was any material in the grain boundaries atomic
force microscopy (AFM) was used. Displayed in Figure 6.5 are three contour plot
images of the BDD surface using different methods for Nafion removal: acetone
(A), ethanol (B), acid washing and rehydrogenation (C). Comparing the images,
the most defined grain structure is seen on the rehydrogenated surface (C).
While the surface cleaned with ethanol (B) shows more definition in the grain
boundaries than the acetone cleaned substrate (A), but is still not as clean as the
rehydrogenated surface. The Nafion is not very soluble in the acetone, therefore,
this was the least effective method, the bulk of the Nafion was removed from the
surface by physical wiping, however, a significant amount remains in the grain
boundaries. It seems the best way to fully remove any Nafion from the diamond
surface is to use ethanol to remove most of the Nafion first followed by a
standard acid washing and rehydrogenation procedure.

In summary, it appears that Nafion prevents fouling of the BDD electrode
surface by components in blood. Using a 30 min exposure time the response
was close to linear with a sensitivity of 1.6 nA/ppb. However, a more
reproducible coating procedure needs to be adopted before this will be useful
analytically. The disadvantage of an ionomer coating is that the response

depends on the amount of time the electrode is in a solution prior to ASV.

104

 

 

 

Figure 6.5. AFM contour plots of three BDD electrodes, taken in contact mode.
The electrodes were coated in Nafion, which was removed with acetone (A),
ethanol (B), and acid washing followed by rehydrogenation (C).
6.3.2. Sodium Dodecyl Sulfate

An alternative to protecting the surface with Nafion was to reduce the
affinity of the proteins for the electrode surface. Hoyer and Jensen suggested
that sodium dodecyl sulfate (SDS) could be used for this purpose to suppress
electrode fouling.3H7 They found that the 808 forms soluble aggregates with
proteins in solution, and prevents their interaction with the surface. This was
successfully applied to the ASV determination of Cd”, Pb”, and Cu” in fruit
juice, beer, milk powder, and waste water using a rotating glassy carbon
electrode.36 They also found that 303 had little or no effect on the ASV signal for

the metal ions at this electrode.” 36

105

 

 

Hoyer used a concentration of SDS that was approximately 3 mM (1000
ppm) which is above the critical micelle concentration (CMC) for a high ionic
strength solution.”39 The CMC of SDS in a pure aqueous solution is around 8
mM.40 However, the CMC is strongly influenced by the ionic strength because
the repulsion between head groups is the limiting factor in micelle formation,
when the ionic strength is higher there is less repulsion, therefore, the micelle
size is increased, and the CMC is decreased.39 The effect of SDS on ASV
measurements depends on many experimental parameters, including electrolyte
and working electrode material, so it was necessary to determine the effect of the
SDS using a BDD electrode in an HCI solution.35

To determine the effectiveness of SDS at protecting the diamond
electrode surface in blood a 3 mM SDS in 0.3 M HCI solution was mixed with a
Pb” sample. Rather than blood, 50 ppm bovine serum albumin (BSA) was
added to a solution of 450 ppb Pb”, to simulate the blood matrix. The electrode
response was first tested using a solution of Pb” in HCI. Next the Pb” solution
containing BSA was tested in the HCI with SDS. Finally the electrode response
was re-tested in Pb” in HCI. The electrode responses are presented in Figure
6.6. There was no significant difference in the Pb stripping response before or
after the exposure to BSA, and the response is not suppressed in the presence of
BSA. This indicates that BDD is not fouled by the BSA when introduced in a

solution also containing SDS.

106

 

 

 

 

 

 

 

 

0.40 '-
— before BSA
- — with BSA & SDS
0.36 . — after BSA
g 0.32 M“, out
*5 \
d)
S
0 0.28 -
0.24 -
0.20 t r

 

 

-0.65 -0.6 -0.55 -0.5
Potential vs. AglAgCI (V)

Figure 6.6. Stripping voltammetric i-E curves of 450 ppb Pb” solution mixed 1:5
with 0.3 M HCI. The solutions used before and after BSA exposure (black and
grey solid lines) only contain Pb” and HCI. The other sample (dashed line) also
contained 50 mg/L BSA and 3 mM SDS. The deposition time was 180 s.

In contrast to the initial measurements in blood (Figure 6.2), the addition of
the SDS appears to prevent the proteins from fouling the bare BDD surface. This
is important to allow measurement of Pb” on an uncoated surface with the
addition of SDS to the blood prior to exposure to the BDD surface. However, the
concentration of BSA in the test was only 50 mg/L, and in the blood the protein
concentration is likely to be in the range of g/L. Therefore, it may be necessary to
use much higher SDS concentrations for use in blood because of the higher
concentration of protein.

6.4. Conclusions

While BDD cannot be used for blood analysis directly, there are a number

of options available to prevent fouling of the surface, making it useful for this

107

 

 

 

measurement. A Nafion-coated electrode was able to be used for Pb” analysis,
and had an approximately linear response with solution concentration. The
advantage and disadvantage of the Nafion-coated electrodes is the
preconcentration of Pb” in the polymer. While this can yield higher signals and
sensitivity it also makes the measurement sensitive to the electrode exposure
time. An alternative option to a coated electrode is a solution based aggregation
of the proteins using an anionic surfactant such as SDS. Employing SDS in this
sampling method is promising because even in the presence of albumin 3 very
similar response was obtained for Pb” with the addition of SDS. The SDS

prevented the protein from fouling the bare BDD.

108

 

 

6.5. References

(1)

(2)

(3)

(4)

(5)

(5)

(7)

(8)

(10)
(11)

(12)

(13)

(14)
(15)

Goyer, R. A. In Casarett & Doull's Toxicology: The Basic Science of
Poisons, 5th ed.; Klaassen, C. 0., Ed.; McGraw-Hill: New York, 1996, pp
691-736.

General Lead lnforrnation: Questions and Answers.
http://www.cdc.gov/nceh/lead/faq/about.htm (accessed 2007)

Plotkin, E.; McAleer, J. F.; Cordeiro, M. L.; Ackland, M. R.; Sheehan, T.
M.; Braithwaite, R. A. Clin. Chem. 1997, 43, 2187-2189.

Matson, W. R. Apparatus for stripping voltammetry. US. Patent
4,201,646, May 6, 1980

Matson, W. R. Sampling system and analysis cell for stripping
voltammetry. US. Patent 5,284,567, February 8, 1994

Matson, W. R. Sampling system and analysis cell for stripping
voltammetry. US. Patent 5,290,420, March 1, 1994

Hoyer, 8.; Florence, T. M.; Batley, G. E. Anal. Chem. 1987, 59, 1608-
1614.

lbanez, J. C. V.; Lorente, M. M. F.; Suarez, J. R. C. Quim. Anal.
(Barcelona) 1998, 1 7, 95-103.

Jaenicke, S.; Sabarathinam, R. M.; Fleet, B.; Gunasingham, H. Talanta
1998, 45, 703-711.

Banks, C. E.; Compton, R. G. Analyst 2004, 129, 678-683.

Kruusma, J.; Banks, C. E.; Compton, R. G. Anal. Bioanal. Chem. 2004,
379, 700-706.

Feldman, B.; D'Alessandro, A.; Osterloh, J. D.; Hata, B. H. Clin. Chem.
1995, 41, 557-563.

Xu, J.; Granger, M. C.; Chen, 0.; Strojek, J. W.; Lister, T. E.; Swain, G. M.
Anal. Chem. 1997, 59, 591A-597A.

Xu, J.; Chen, 0.; Swain, G. M. Anal. Chem. 1998, 70, 3146-3154.
Granger, M. C.; Witek, M.; Xu, J.; Wang, J.; Hupert, M.; Hanks, A.;

Koppang, M. D.; Butler, J. E.; Lucazeau, G.; Mermoux, M.; Strojek, J. W.;
Swain, G. M. Anal. Chem. 2000, 72, 3793-3804.

109

 

 

 

 

(15)

(17)

(18)

(23)
(24)
(25)

(26)

(27)
(28)
(29)

(30)

(31)
(32)
(33)
(34)
(35)
(36)

Hupert, M.; Muck, A.; Wang, J.; Stotter, J.; Cvackova, Z.; Haymond, 8.;
Show, Y.; Swain, G. M. Diamond Relat. Mater. 2003, 12, 1940-1949.

Shin, D.; Tryk, D. A.; Fujishima, A.; Merkoci, A.; Wang, J. Electroanal.
2005, 17, 305-311.

Smart, R. 8.; Stewart, E. E. Environ. Sci. Technol. 1985, 19, 137-140.

Kamei, K.-i.; Mie, M.; Yanagida, Y.; Aizawa, M.; Kobatake, E. Sens.
Actuators, B 2004, 99, 106-112.

Wang, J.; Hutchins, L. D. Anal. Chem. 1985, 57, 1536-1541.

Wang, J.; Hutchins-Kumar, L. D. Anal. Chem. 1986, 58, 402-407.
Kuhn, L. S.; Weber, S. G.; Ismail, K. 2. Anal. Chem. 1989, 61, 303-309.
Hoyer, B.; Jensen, N. Talanta 1995, 42, 767-773.

 

Marinesco, S.; Carew, T. J. J. Neurosci. Methods 2002, 117, 87-97.

Kristensen, E. W.; Kuhr, W. G.; Wightman, R. M. Anal. Chem. 1987, 59,
1752-1757.

lbanex, J. C. V.; Lorente, M. M. F.; Suarez, J. R. C. Quimica Analytica
1998, 17, 95—103.

Liu, K.-z.; Wu, Q.-g.; Liu, H.-l. Analyst 1990, 115, 835-837.
Dalangin, R. R.; Gunasingham, H. Anal. Chim. Acta 1994, 291, 81-87.

Dam, M. E. R.; Thomsen, K. N.; Pickup, P. G.; Schroder, K. H.
Electroanal. 1995, 7, 70-78.

Matysik, F.-M.; Matysik, S.; Vrett, A. M. 0.; Brett, C. M. A. Anal. Chem. E
1997, 69, 1651-1656. 5

Hurst, M. P.; Bruland, K. W. Anal. Chim. Acta 2005, 546, 68-78.
Kefala, G.; Economou, A. Anal. Chim. Acta 2006, 576, 283-289.

 

Legeai, S. Anal. Chim. Acta 2006, 560, 184-190.

Zook, L. A.; Leddy, J. Anal. Chem. 1996, 68, 3793-3796.

Hoyer, B.; Jensen, N. Electrochem. Commun. 2003, 5, 257-261.
Hoyer, B.; Jensen, N. Analyst 2004, 128, 751-754.

110

(37)
(38)
(39)

(40)

Hoyer, B.; Jensen, N. Electroanal. 2005, 17, 2037-2042.
Emerson, M.; Holtzer, A. J. Phys. Chem. 1967, 71, 1898-1907.

Helenius, A.; McCaslin, D. R.; Fries, E.; Tanford, C. Methods Enzymol.
1979, 56, 734.

Bales, B. L.; Messina, L.; Vidal, A.; Peric, M.; Nascimento, O. R. Journal of
Physical Chemistry B 2998, 102, 10347-10358.

111

 

 

CHAPTER 7
CONCLUSIONS

The overall aim of this work was to fully evaluate and understand the
properties of boron-doped diamond (BDD) as they related to its use in anodic
stripping voltammetric (ASV) measurements. Initial studies in this area showed
that BDD was a suitable electrode with ASV for metal ion analysis in river water,
tap water, soil, and sludge. In order to more fully evaluate BDD as an alternative
electrode to Hg for ASV several studies were conducted: (i) comparison of the
electrochemical properties and response with the traditional Hg electrode, (ii)
exploration of the utility of the material in microgravity applications, (iii)
determination of the usefulness of BDD for more complex matrices (e.g., urine,
blood), and (iv) investigation of ways of preventing electrode fouling (e.g.,
coatings, complexing agents) in solutions containing high protein concentrations
(e.g., whole blood).

In addition to having many of the desirable features of Hg, BDD has some
inherent advantages; it is non-volatile and non-toxic. This makes it an
advantageous electrode choice for clinical and environmental analysis. The
response of BDD was compared to a Hg-coated glassy carbon (Hg-GC)
electrode in terms of the analytical figures of merit, nature of metal deposition,
and stripping voltammetric peak potentials and shapes. BDD has a wider anodic
potential limit and a lower background current density than Hg. Overall, the
detection figures of merit for BDD are as good or superior to those for Hg-GC.

Both BDD and Hg-GC provided detection limits for all the metal ions (Zn”, Cd”,

112

 

 

Pb”, Cu”, and Ag’) in the mid to low ppb range (S/N 2 3). The biggest
difference between the two electrodes was the magnitude of the stripping peak
current and charge, which are greater for Hg-GC. This is because of the large
volume of Hg into which the foreign metal can deposit compared to the surface
area of BDD available for deposition. The greater amount of metal deposited on
Hg-GC resulted in a higher sensitivity, but this was offset by a comparable
reduction in the background noise for BDD, which resulted in comparable or
lower limit of detection values. Electrochemical measurements revealed that
reduced sensitivity due to incomplete metal oxidation or metal phase detachment
was not an issue for these BDD films.

A significant effort was spent developing BDD for use in water supply
monitoring on-board the lntemational Space Station (ISS). The properties of
BDD (e.g., non-toxic, non-volatile) made it attractive for this application. The
previously developed ground-based ASV method was modified for microgravity
use. A concentrated buffer solution was developed to adjust the solution
conditions (e.g., pH, electrolyte concentration), and a prototype cell for
microgravity was developed. The prototype cell was a fully sealed
electrochemical cell that allowed introduction of sample through Iuer fittings
compatible with the current NASA water sampling system. The method provided
accurate concentration determinations for Ag”, Cu”, Pb”, and Cd”, and the
response was unaffected by the increasedng (5.2), or the minerals added for
taste (Ca” and Mg” salts). Elimination of the N2 purge step to remove dissolved

02 was not a problem for Ag" and Cd” analysis; however, the presence of

113

 

 

dissolved 02 did affect the Pb” and Cu” analysis. The use of an oxygen
scavenging molecule (ascorbic acid) to chemically remove the 02 mitigated this
problem to a large extent. BDD exhibited good long-term response stability for
Ag+ (<10% relative standard deviation in response over 90 days). Analysis of
simulated water samples in the microgravity cell using BDD yielded accurate Ag+
concentrations (< 5% error).

Once BDD had been successfully applied to environmental and water
samples, the next type investigated was biological samples (e.g., blood, urine) to
probe the fouling resistance of the surface in more complex matrices. A urine
simulant containing Cd” was used and a linear calibration curve was obtained
over a range of 2 orders of magnitude (10—250 ppb), and the response was
reproducible from electrode-to-electrode (RSD 6%). Urine contains compounds
that can complex the Cd” (e.g., ammonia, lactic acid, citric acid), which reduced
the activity of free Cd” thereby reducing the electrochemical response. To
obtain the highest electrochemical signal and accurate concentration
determination for Cd”, a digestion step or other purification step is necessary.
While the simulant contains many of the major components of urine, it did not
contain any protein, therefore, real urine (rat) was tested as well. BDD
performed adequately, and an increased signal was observed for increasing
concentrations of Cd”. The electrode showed no fouling effects over a 6-hour
penod.

One of the more complex biological matrices is blood, and for metal ion

analysis, it is necessary to use the whole blood rather than just a plasma sample.

114

 

 

Blood contains a large amount of proteins that can passivate the electrode
surface, and even though BDD is resistant to fouling, it cannot be used directly in
whole blood. Two different fouling prevention strategies were investigated: a
polymer coating (Nafion) and an added surfactant to aggregate with the proteins
in solution (sodium dodecyl sulfate). BDD was successfully coated with Nafion,
and was able to be used for Pb” analysis with a linear calibration response
(r2>0.98). However, Nafion coating does have some disadvantages. First, the
reproducibility of the coating can affect the reproducibility of the measurement
from run-to-run. Second, the analysis is time dependent because Nafion is a
cation-exchange polymer, and the longer it remains in contact with the Pb”
solution the more preconcentration occurs increasing the electrochemical signal.
The solution-based aggregation of protein with sodium dodecyl sulfate (SDS) is a
promising option, and even in the presence of albumin the BDD showed very
similar electrochemical response to the analysis in standard solutions. The
benefit of the solution based method is that it is much easier to add a
reproducible amount of SDS to a solution than to have a reproducible polymer
coafing.

Overall, BDD performed well for ASV analysis, and is definitely a suitable
alternate to Hg. The inert nature of the electrode material and the ability to reuse
the surface for multiple replicate analyses are definite advantages. In this work,
the utility of BDD was demonstrated for many different sample types; however,
one thing that was not specifically addressed was potential interferences from

interrnetallic compound formation.

115

 

 

lnterrnetallics can form during the deposition from a sample that contains
different metal ions. This tends to be more of a problem on a solid electrode,
such as BDD. In a Hg electrode the metal will form an amalgam, which helps to
reduce formation of intermetallic species by physically separating the metals. On
a solid electrode such as BDD the different metals can co-deposit at the surface
and form an intermetallic. The intermetallic formation can affect ASV analysis by
shifting the potential at which a metal is detected because the E0 value for a pure
metal will be different from an intermetallic.

In the previous studies using BDD to detect multiple metal ions in a
standard reference material there was some evidence for the formation of
intermetallics on BDD; a secondary peak associated with the Cu peak was
observed. However, the concentrations of the metals in the standard reference
material were still able to be accurately determined. While the formation of
intermetallics did not pose a problem in the concentration determination in this
particular sample it is worth investigating further to determine the utility of BDD in

multimetal solutions.

116

 

 

   

lllllllllllllllli