ADVANCEMENTS IN ELECTROCHEMISTRY FOR THE SENSING OF AQUEOUS

COMPOUNDS AND PERSISTENT COMPOUND DESTRUCTION

By

Mary Ensch

A DISSERTATION

Michigan State University

in partial fulﬁllment of the requirements

Submitted to

for the degree of

Chemical Engineering – Doctor of Philosophy

2020

ABSTRACT

ADVANCEMENTS IN ELECTROCHEMISTRY FOR THE SENSING OF AQUEOUS

COMPOUNDS AND PERSISTENT COMPOUND DESTRUCTION

By

Mary Ensch

A signiﬁcant research question to address is if existing electrochemical methods can be used to
advance the sensing of biological compounds and toxic heavy metals, and to destroy persistent
toxic pollutants to beneﬁt the planet and humankind. Using the methods of cyclic voltammetry
(CV), square wave voltammetry (SWV), square wave stripping voltammetry (SWSV), and electro-
chemical oxidation (EO), the sensing of isatin and manganese, and the destruction of perﬂuoroalkyl
compounds was investigated. Each task had the main focus of advancing our understanding of the
process and making signiﬁcant process improvements.

Isatin is an endogenous indole that has potential pharmaceutical applications based on its anti-
convulsant abilities through reaction pathways in the human body. Due to its low concentrations
in blood, serum, and tissues, it is typically detected using gas chromatography mass spectroscopy
(GC/MS) which requires a specialized individual and lengthy measurement times.
In order to
combat this, researchers have begun investigating electrochemical detection methods for this com-
pound. Here, the investigation of isatin detection on a 3-in-1 boron-doped diamond (BDD) sensor
was studied to simplify the electrochemical system requirements and set the ground work for a
potential plug and play sensor for isatin. Both sensor conﬁgurations tested showed applicability to
this application with the lowest limit of detection obtained to be 0.04 µM. Not only was isatin able to
be detected in a complex sample matrix in a short measurement time with this simpliﬁed technical
approach, it was done so with the elimination of extensive electrode pretreatment processes that
have been required for other electrode materials in the past. This study paves the way for a fully
developed BDD sensor that can expand the understanding of isatin reaction pathways and advance
its ability to be used in pharmaceuticals.

Manganese is an important nutrient but has toxic e ects in humans making it imperative to be
able to detect it in human blood, urine, and serum as well as drinking water. Indium tin oxide
(ITO) has been a typical electrode material for spectroelectrochemical measurements and metal
detection experiments on transparent substrates. This study goes through a few process parameters
for an ITO ﬁlm on a non-transparent silicon substrate. It was found that heating and annealing the
ﬁlm created the best ﬁlm structure for an electrochemical sensing application towards manganese
with a limit of detection of 0.1 parts per billion (ppb). This obtained lower detection limit allows
for improved accuracy in human serum measurements with no added membranes to the electrode’s
surface and uses an electrode material that already has the capabilities to be mass produced with
good repeatability. This makes ITO an economical sensing solution.

Per- and polyﬂuoroalkyl substances (PFAS) are toxic to human health and are wide spread in the
environment due to their numerous consumer and industrial applications. Most current remediation
techniques take advantage of ﬁltration or adsorption technologies to remove these compounds from
a water source. However, since they do not destroy the compounds, these remediation techniques
yield waste products. This study, utilizing electrochemical oxidation for the destruction of PFAS,
looks at tackling waste products from an ion exchange process. By utilizing BDD, a laboratory
study showed that PFAS completely mineralize in these complex solutions. Overall, it was found
that ammonia based solutions allow for the best mineralization of the most common PFAS using
electrochemical oxidation for an ion exchange regenerate solution. This study provides valuable
knowledge to ion exchange manufacturers who are looking to solve their residual PFAS waste
problem.

All three studies have shown that existing electrochemical techniques can be used to further
sensing and remediation knowledge in today’s world and provide safer, cleaner technology options
for the betterment of human health and the environment.

Copyright by
MARY ENSCH
2020

To Mom, Dad, Hannah, and Michaelangelo. The best support system I could have asked for.

v

ACKNOWLEDGEMENTS

First, I would like to thank Dr. Thomas Schuelke for providing ﬁnancial support through the
Fraunhofer USA, Inc. Center for Coatings and Diamond Technologies, which has allowed me to
complete my research as well as enable me to present the results at various scientiﬁc conferences.
I would also like to thank him for his ongoing support and guidance through my PhD studies.
Additionally I would like to thank both Michael Becker and Dr. Cory Rusinek for providing day-
to-day mentorship and support. I am also very grateful to Dr. Scott Calabrese-Barton, Dr. Greg
Swain, Dr. Elias Garratt, and Dr. Richard Lunt for serving as committee members and o ering
valuable insights and advice. Finally, I want to thank all of the dedicated researchers at Fraunhofer
USA as well as sta  members, interns, and MSU graduate and undergraduate students for all of
their help along the way.

vi

TABLE OF CONTENTS

.

.

.

.

.

.

.

x

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

xi

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
LIST OF TABLES .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
LIST OF FIGURES .
CHAPTER 1 MOTIVATION OF WORK . . . . . . . . . . . . . . . . . . . . . . . . . .
1
2
1.1 Overview .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
1.2 Electrochemical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
1.2.1 Cyclic Voltammetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
Square Wave Voltammetry . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.2
1.2.3
9
Square Wave Stripping Voltammetry . . . . . . . . . . . . . . . . . . . . .
1.2.4 Electrochemical Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.3 Background on Electrochemical Sensing . . . . . . . . . . . . . . . . . . . . . . . 11
Structure-property Connections
. . . . . . . . . . . . . . . . . . . . . . . 12
1.3.1.1 Boron-doped Diamond . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.1.2
. . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Prior State of the Art
1.3.3 Deﬁciencies of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . 17
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Indium Tin Oxide

1.3.1

.

.

.

.

.

.

.

.

BIBLIOGRAPHY .
CHAPTER 2

2.3 Results and Discussion .

Introduction .

2.1
.
2.2 Experimental Section .

.

.

.

.

.

.

.
.

ISATIN DETECTION USING A BORON-DOPED DIAMOND 3-IN-1
SENSING PLATFORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2.1 Chemicals and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Instrumentation .
2.2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3-in-1 Electrode Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2.3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.1 Electrode Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.2 BDD as a Quasi-Reference Electrode . . . . . . . . . . . . . . . . . . . . . 36
2.3.3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3.4 Urine Simulant Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3.5 Repeatability Study .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.3.6 Key Advancements Obtained from Study . . . . . . . . . . . . . . . . . . 47
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Isatin Detection .

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.

.

INDIUM TIN OXIDE FILM CHARACTERISTICS FOR CATHODIC
STRIPPING VOLTAMMETRY . . . . . . . . . . . . . . . . . . . . . . . . 56
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Introduction .

3.1
.
3.2 Experimental Section .

.
.

.

.

.

.

2.4 Summary .
.
BIBLIOGRAPHY .
CHAPTER 3

vii

.

.

.

.

.

.

Instrumentation .

3.3 Results and Discussion .

3.2.1 Chemicals and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.3 Electrode Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3.1
Physical and Optical Characterization: Atomic Force Microscopy . . . . . 61
3.3.2
Physical and Optical Characterization: XRD . . . . . . . . . . . . . . . . . 63
3.3.3
Physical and Optical Characterization: UV-Vis Spectroscopy . . . . . . . . 65
3.3.4 Electrical Properties
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.5 Electrochemical Characterization . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.6 CSV of Manganese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.3.7 Calibration and Figures of Merit
. . . . . . . . . . . . . . . . . . . . . . . 72
3.3.8
Possible Metals Interference . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.9 Key Advancements from Study . . . . . . . . . . . . . . . . . . . . . . . . 75
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

3.4 Summary .
.
BIBLIOGRAPHY .
CHAPTER 4 ELECTROCHEMICAL OXIDATION FOR THE MINERALIZATION

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

4.3.3

Introduction .

PFOS .
PFOA .

OF PER- AND POLYFLUOROALKYL SUBSTANCES . . . . . . . . . . . 84
4.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.2 Electrode Stack Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.3 Constant Current
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3.1 Methods .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
.
4.3.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.3.2.1
4.3.2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.3.2.3 Comparison of ISE Results to LC/MS Results . . . . . . . . . . . 96
Fluorination of BDD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.4 Combined Current Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.4.1 Comparison of ISE and LC/MS . . . . . . . . . . . . . . . . . . . . . . . 104
4.4.2 Remediation of Simulated Ion Exchange Regenerate Solutions . . . . . . . 105
4.4.2.1 Comparison of 50 mA/cm2 to 50&5 mA/cm2
. . . . . . . . . . 106
4.4.2.2 Comparison of four di erent regenerate solutions . . . . . . . . . 108
4.4.2.3 Byproduct Formation Investigation . . . . . . . . . . . . . . . . 114
4.4.3 Energy Usage and Cost Comparison . . . . . . . . . . . . . . . . . . . . . 115
4.5 Key Advancements from Bench Top PFAS Studies . . . . . . . . . . . . . . . . . . 118
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
.
.
4.6 Summary .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
.
.
BIBLIOGRAPHY .
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
CHAPTER 5 CONCLUSION .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
.
.
CHAPTER 6 FUTURE OUTLOOK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

6.1 Future Outlook .

5.1 Conclusion .

.
.
.
.

.
.

.
.

.
.

.
.

.

.
.

.

.
.

.
.

.
.

.

.

.

.

.

.

.

viii

Isatin Detection .
Indium Tin Oxide Characteristics

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
6.1.1
6.1.2
. . . . . . . . . . . . . . . . . . . . . . 129
6.1.3 Mineralization of Per- and Polyﬂuoroalkyl Substances . . . . . . . . . . . . 129

.

.

ix

LIST OF TABLES

Table 1.1: Summary table outlining the prior art for the electrochemical detection of isatin . 15

Table 1.2: Summary table outlining the prior art for the electrochemical detection of

manganese in simulated bu er solutions . . . . . . . . . . . . . . . . . . . . . . 16

Table 1.3: Summary table outlining the prior art for the simulated PFAS solutions

. . . . . 17

Table 1.4: Summary table outlining the prior art for the complex solutions containing PFAS 17

Table 2.1: Figures of merit for isatin detection using the 3-in-1 MAC and MEA sensors
compared to other reported literature studies on di erent detection methods.
Standard deviations calculated at lowest concentration detected for this study . . 43

Table 3.1: Comparison of each of the ﬁgures of merit for each of the four ITO ﬁlms that

were investigated. .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Table 3.2: Fitting parameters for the band gap determination from the Tauc plot of each ﬁlm. 67

Table 3.3: Redox potentials for other metals that are potentially in environmental and

biological samples

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Table 4.1: Comparison of remaining concentration of PFOA obtained from the ISE anal-

ysis and LC/MS analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Table 4.2: Rate constants and their linear ﬁts for the three tested combined current densities 104

Table 4.3: Comparison of 1 h PFOA degradation and 12 h PFOA degradation . . . . . . . . 111

Table 4.4: Comparison of 1 h PFOA degradation and 12 h PFOA degradation . . . . . . . . 113

Table 4.5: Energy requirements of each PFOA study to reach 75% deﬂuorination or at 1 hour116

x

LIST OF FIGURES

Figure 1.1:

(left) diagram of a beaker setup for electrochemical measurements and (right)
a clamp cell setup that can also be used alternatively. Each has (A) a working
electrode, (B) a reference electrode, and (C) a counter/ auxiliary electrode . . .

Figure 1.2: Setup used for an electrochemical oxidation experiment including an (A)

anode, (B) cathode, and (C) stir bar . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.3:

(left) Potential waveform for cyclic voltammetry and (right) the corresponding
current response .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

.

Figure 1.4:

(left) Potential waveform and (right) typical forward, reverse, and di erence
current response for square wave voltammetry . . . . . . . . . . . . . . . . . .

4

4

5

8

Figure 2.1: Molecular structure for isatin . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Figure 2.2: Optical micrograph of the MEA (left) and MAC (right) 3-in-1 sensors with

exposed BDD . .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Figure 2.3:

Figure 2.4:

(A) An SEM micrograph of the polycrystalline BDD surface that was used to
fabricate the two 3-in-1 sensors with no defects or large cracks in the ﬁlms
structure and (B) the corresponding Raman spectrum analysis for the BDD wafer 33
(A) Cyclic voltammograms of 1 mM Fe(CN)63- and 1 mM Ru(NH3)6Cl on
the MAC 3-in-1 BDD sensor oth dissolved in 1 M KCl and (B) of a 1:1 ratio
of Fe(CN)63-/4- and 1 mM of Ru(NH3)6Cl , both dissolved in KCl, on the
3-in-1 BDD MEA sensor. Scan rate: 0.1 V/s. . . . . . . . . . . . . . . . . . . . 34

Figure 2.5: The BDD’s reference potential in 0.1M phosphate bu ered saline at various

pH values measured against a Ag/AgCl reference electrode for 12 hours.

. . . . 37

Figure 2.6: The BDD’s reference potential for the extended 28 hours in 0.1M phosphate

bu ered saline at a pH of 7.4 measured against a Ag/AgCl electrode.

. . . . . . 38

Figure 2.7: Current versus voltage cyclic voltammogram curves for 350 µM isatin in 0.1
M PBS at pH 7.4 on the 3-in-1 (A) MAC sensor and (B) MEA sensor. Scan
rate of 0.1 V/s. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

.

.

.

Figure 2.8: Current versus voltage cyclic voltammogram curves of 350 µM isatin in 0.1 M
PBS at pH 7.4 (red) and after rinsing the 3-in-1 MAC electrode and rescanning
the same electrode in clean electrolyte solution (black). Scan rate of 0.1 V/s.

. . 40

xi

Figure 2.9: Squarewave voltammograms for various concentrations of isatin on the 3-in-1
(A) MAC and (B) MEA sensor conﬁgurations in 0.1 M phosphate bu ered
saline with pH of 7.4.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Figure 2.10: Calibration curves for (A) the MAC and (B) the MEA in 0.1 M phosphate

bu ered saline with a pH of 7.4 . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Figure 2.11: UV-vis absorbance spectra for isatin in 0.1 M of PBS pH 7.4 . . . . . . . . . . 44

Figure 2.12: Squarewave voltammograms for the 3-in-1 MAC sensor (A) and MEA sensor

(B) in Surine Negative Urine Control with no additional N2 bubbling . . . . . . 45

Figure 2.13: Calibration curves for the (A) MAC and (B) MEA for isatin in Surine Negative

Urine Control with no N2 bubbling . . . . . . . . . . . . . . . . . . . . . . . . 46

Figure 2.14: Repeatability data for the 3-in-1 (A) MAC and (B) MEA sensors with the
discussed electrochemical pretreatment and without any addition of a pre-
treatment

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

.

.

.

.

.

.

Figure 3.1: Tapping mode atomic force microscopy (AFM) images for the four analyzed

ITO ﬁlms (A) HA, (B) HNA, (C) NHA, and (D) NHNA.

. . . . . . . . . . . . 62

Figure 3.2: XRD scans for the (A) (222) peak and (B) (400) peak for each ITO ﬁlm

investigated.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Figure 3.3:

(left) ITO crystal quality (I(222)/I(400)) ratio obtained from XRD analysis, and
(right) the full width half max (FWHM) data for each ﬁlm from XRD analysis
pertaining to either the (222) or (400) peak . . . . . . . . . . . . . . . . . . . . 64

Figure 3.4:

ITO crystal size derived using the Scherrer equation and the (222) and (400)
peak of each ﬁlm .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

.

.

Figure 3.5: Transmittance spectra measured with UV/Vis for each of the four ﬁlms de-

posited on a transparent microscope slide . . . . . . . . . . . . . . . . . . . . . 66

Figure 3.6: Tauc graph used to determine the direct optical band gap for the four ITO

ﬁlms studied .

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Figure 3.7: Graph showing the current vs scan rate study of each of the ITO ﬁlms used

to determine their corresponding capacitance values . . . . . . . . . . . . . . . 69

Figure 3.8:

(A) CV curves of 0.1 mM ferrocyanide in 1.0M KCl with a scan rate of 0.1
V/s. (B) Potential windoe study performed in acetate bu er of pH 5.0. (C)
CSV od 25 ppb of Mn in acetate bu er of pH 5.0, deposition time of 3 minutes
and a deposition potential of +1.2 V.

. . . . . . . . . . . . . . . . . . . . . . . 69

xii

Figure 3.9: Plots for the Ohm’s resistance of each of the four ITO ﬁlms investigated . . . . 71

Figure 3.10: Mn stripping voltammograms for the full concentration range on HA (A1) and
the smaller range (A2), for the full range on HNA (B1) and the smaller range
(B2), and for the full range on NHA (C1) and smaller range (C2). Performed
in pH 5.0 acetate bu er with a deposition potential of 1.2 V and a deposition
time of 3 min . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

.

.

Figure 3.11: Calibration curves for Mn on HA (A), HNA (B), and NHA (C)

. . . . . . . . . 73

Figure 4.1: Molecular structure of (top) PFOA and (bottom) PFOS . . . . . . . . . . . . . 86

Figure 4.2: Scanning electron microscopy (SEM) image of the BDD surface . . . . . . . . 88

Figure 4.3: Raman spectroscopy of BDD electrodes

. . . . . . . . . . . . . . . . . . . . . 88

Figure 4.4: Current versus scan rate plots to determine the capacitance of BDD . . . . . . . 89

Figure 4.5: Cyclic voltammogram of both ferrocyanide and ruthenium hexamine on the

BDD surface .

. .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Figure 4.6: PFOS Temperature, pH, and voltage trends for the constant current density studies 91

Figure 4.7: PFOS reduction overtime with each current density tested . . . . . . . . . . . . 92

Figure 4.8: PFOS 1st order rate constants . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Figure 4.9: PFOA Temperature, pH, and voltage trends for the constant current density

studies .

.

.

.

. .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Figure 4.10: PFOA degradation over time at each current density value . . . . . . . . . . . . 95

Figure 4.11: 1st order rate constants for PFOA based on LC/MS measurements for each

current density study.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Figure 4.12: Comparison of the PFOA concentration obtained by the F- ISE and LC/MS . . . 97

Figure 4.13: 1mM ferrocyanide in 1 M KCl measured on both the used and unused portion
of the anode that was used for the constant current density PFOA degradation
studies .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

. .

.

.

.

.

.

Figure 4.14: CV in 1 M sulfuric acid for both the used and unused portion of the anode . . . 100

xiii

Figure 4.15: (A) CV comparison in1 mM ferrocyanide in 1 M KCl for the used anode,
unused anode, and used anode post hydrogen termination and (B) CV in 1 M
sulfuric acid for the used, unused, and used electrode post hydrogen termination 100

Figure 4.16: Temperature, pH, and voltage trends for the combined current density studies . . 102

Figure 4.17: PFOA reduction overtime with each combined current densities tested to-

gether as well at the standard 50 mA/cm2 for comparison . . . . . . . . . . . . 102

Figure 4.18: First order rate constant ﬁts for the ﬁrst 0.25 h, k1, and the remainder of each

test, k2. .

.

.

.

. .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Figure 4.19: Comparison of the PFOA concentration obtained by the F- ISE and LC/MS . . . 105

Figure 4.20: Temperature, pH, and voltage trends for the comparison of 50 and 50&5 in

IX simulated solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Figure 4.21: Comparison of the PFOA (left) and PFOS (right) degradation using 50

mA/cm2 and 50&5 mA/cm2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Figure 4.22: Temperature, pH, and voltage trends for the comparison of the four IX simu-

lated solutions . .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Figure 4.23: Comparison of the PFOA (left) and PFOS (right) degradation using the four

tested simulated IX solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Figure 4.24: 1st order rate constant kinetics for PFOA in each simulated solution . . . . . . . 110

Figure 4.25: Sequential 1st order rate constant kinetics for PFOS in each simulated solution . 112

Figure 4.26: Oxygen evolution reaction performed for all four solutions . . . . . . . . . . . . 113

Figure 4.27: Calibration curve for formaldehyde detection . . . . . . . . . . . . . . . . . . . 115

Figure 4.28: Extrapolation plots used to determine the time to 75% deﬂuorination for (A)

1 mA/cm2, (B) 5 mA/cm2, and (C) 10 mA/cm2

. . . . . . . . . . . . . . . . . 117

Figure 4.29: Bar graph comparing the conversion e ciency values for the simulated IX

regenerate solutions .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

xiv

CHAPTER 1

MOTIVATION OF WORK

1

1.1 Overview

An important question to consider is if electrochemical techniques can be advanced to improve
upon biological compound sensing, toxic compound sensing, and toxic compound destruction to
beneﬁt the planet and humankind. The ability to improve upon biological sensing is important to
advance pharmaceuticals based on how compounds react in the body and to understand the levels
at which they can be manipulated to better treat or prevent disease.
Isatin’s ﬁrst detection was
in 1991 with a gas chromatographic mass spectroscopy (GC/MS) method and a few years later
by high performance liquid chromatography measurements (HPLC). [1, 2] UV detection methods
were being coupled with HPLC to detect isatin, but only were successful at obtaining limits of
detection in the mg/L ranges. [3] These are not relevant to concentrations seen in the human body.
[4, 5] Application of electrochemical detection for HPLC was utilized to gain insight on how isatin
e ected the levels of neurotransmitters in rats that had Parkinson’s disease, and although it provided
promising results, the three electrode setup comprised of a chemically modiﬁed working electrode
material that required a multiple step preparation procedure. [6] Isatin’s detection would be greatly
advanced by removing the need of multiple sample pre-treatments, providing a measurement
technique that is accurate and portable without extensive electrode processing, and obtaining limits
of detection that allow understanding of isatin’s reactions in the human body.

Toxic heavy metals are persistent in the environment and have been shown to cause harm
to humans when above a required threshold.
If these metals go undetected in drinking and
other water sources, they can become a substantial health issue. Electrochemical heavy metal
analysis dates back to 1996 with the development of a microfabricated analysis system. [7] Since
this time, electrochemical sensors and biosensors have been developed with di erent materials,
conﬁgurations, and styles. For manganese detection speciﬁcally, electrode materials that have been
used include carbon, lithium doped or polymer coated indium tin oxide, bismuth, copper based
palladium, or platinum. [8–16] The main advancement that the ﬁeld is relying on is the ability to
mass produce reliable sensors, that require minimal pretreatment, at a low cost without hindering

2

on the obtained limits of detection.

Finally, the use of electrochemical methods to destroy organic compounds that are toxic to
humans are of utmost importance to mineralize the contaminants once and for all and prohibit
a contamination cycle from continuing. Using electrochemistry aimed at remediating organic
contaminants has been around since the early 1960s. These contaminated solutions include textile
dyes, petrochemicals, paper industry e uents, tannery e uents, and other industrial wastewater.
[17] As industry advances and develops new chemicals, the risks associated with them, and their
need for remediation, also increases. To improve upon this technique for modern times, a system
would need to be adapted to allow for the removal of ultra low concentrations of contaminants, in a
true environmental matrix, that are unable to be broken down by competing technologies in a way
that remains energy e cient.

These three important tasks can be accomplished using the electrochemical methods of cyclic
voltammetry, square wave voltammetry, square wave stripping voltammetry, and electrochemical
oxidation.
1.2 Electrochemical Methods

The main electrochemical methods used in this dissertation include those of cyclic voltammetry
(CV), square wave voltammetry (SWV), square wave stripping voltammetry (SWSV), and elec-
trochemical oxidation (EO). The next few sections outline the basic underlying principles that are
important in further chapters. Each of the voltammetry measurements utilizes a three-electrode
setup comprised of a working, reference and auxiliary electrode shown either of the conﬁgurations
in Figure 1.1. The three electrodes are connected to a potentiostat. Electrochemical oxidation
utilizes a two-electrode setup with the use of an anode and a cathode that are connected to a
powersupply. The setup for an electrochemical oxidation experiment is shown in Figure 1.2.

3

Figure 1.1: (left) diagram of a beaker setup for electrochemical measurements and (right) a clamp
cell setup that can also be used alternatively. Each has (A) a working electrode, (B) a reference
electrode, and (C) a counter/ auxiliary electrode

Figure 1.2: Setup used for an electrochemical oxidation experiment including an (A) anode, (B)
cathode, and (C) stir bar

1.2.1 Cyclic Voltammetry

CV is performed by scanning the potential between the working and reference electrodes in a
negative or positive direction and measuring the resulting current response. Typically, an experiment
would start at one potential, scan out in either direction to a further potential in a linear fashion,

4

and return to the starting point to complete one cycle. During this time, the current is measured.
A ﬁgure depicting the potential waveform used in CV and the corresponding current response
curve is shown in Figure 1.3. In order to understand repeatability and monitor change over time,
typically 4 to 6 cycles are performed. The rate at which the scan is performed is also an important
experimental parameter.

Figure 1.3: (left) Potential waveform for cyclic voltammetry and (right) the corresponding current
response

There is a relationship between the scan rate and the current that is measured on the working

electrode that is known as the capacitance, Equation 1.1. [18]

i = C ⇤

dV
dt

(1.1)

Where i is the current (A), V is the potential in V, t is the time in seconds, and C is the
capacitance (F). By selecting a potential and recording the current output at various scan rates, a
plot of the current versus the scan rate can be generated which provides the capacitance as the slope
(A*s/V or F). CV is also important to determine at which potentials chemical reactions take place
and how fast they occur on the working electrode’s surface. Since most electrolytes are aqueous,
the workable potential window of a speciﬁc electrode material can be determined by measuring the
potentials between the water oxidation and reduction reactions. This will also provide knowledge
on if the electrode can be oxidized or reduced and thus hinder its ability to detect other compounds
in solution.

5

How fast a reaction occurs on a given electrode can also be determined using CV. This allows
for the determination of the heterogenous rate constant for the material and allows characterization
and comparison of di erent electrode materials. Using a common redox couple, such as potassium
ferrocyanide, a CV experiment can be performed and the Randles-Sevcik equation can be used that
relates the peak heights measured, the area of the working electrode exposed in solution, the scan
rate used, and the di usion coe cient of the redox couple as seen in Equation 1.2.

ip = (2.69x105)⇤ n

3
2 ⇤ D

1
2 ⇤ C ⇤ A ⇤ v

1
2

(1.2)

Where ip is the peak current (A), n is the number of electrons in the reaction, D is the di usion
coe cient (cm2/s), C is the concentration of the redox compound (mol/cm3), A is the exposed area
of the electrode (cm2), and v is the scan rate (V/s). Once the di usion coe cient is determined
from this equation it can be used in two di erent methods to determine the rate constants for a
redox couple on an electrode material.

The ﬁrst method is known as the Nicholson Shain method and can be utilized when the separation
between the oxidation and reduction peaks is less than 212 mV. [19] The equation that describes
this method is provided as Equation 1.3.

ko = '(⇡Doxv nF
RT)
↵
( Dox
2
Dred)

1
2

(1.3)

Where ko is the heterogeneous rate constant (cm/s), ' is a kinetic parameter that is tabulated
based on peak potential separation at a set temperature for a one electron process and can be found
in a table in reference [19], Dred and Dox are the di usion coe cients for the reduced and oxidized
form of the redox couple, respectively (cm2/s), and ↵ is the transfer coe cient that usually has a
value of 0.5. The type of di usion that a redox couple has on a working electrode material can
be determined based on graphing the peak currents obtained from multiple scan rates against the
square root of the scan rate. If this relationship is linear, there is semi-inﬁnite linear di usion.

6

The second method used for rate constant determination is the Klinger and Kochi method that
utilizes fast scan CV for compounds that traditionally have peak separations larger than 212 mV or
those that show a steady state response. [20] Fast scan CV refers to using scan rates that are >100
V/s compared to traditional scans in the mV/s range. The equation for this method is provided as
Equation 1.4. [20]

ko = 2.18(D↵nv

F
RT )

1
2 e (↵2n F

RT )(Eox

p  Ered
p

)

(1.4)

Where Epox and Epred are the peak potentials for the oxidation and reduction peaks, respectively.
This method will also provide information on the di usion of the process and again shows semi-
inﬁnite linear di usion when the forward peak currents have a linear relationship with the square
root of the scan rate.

The method of CV provides the tools to characterize and compare di erent working electrode
materials. It is being utilized in the following chapters to advance knowledge on the properties that
boron-doped diamond has as a quasi-reference material, what factors change the electrochemical
responses seen with indium tin oxide, and if there are surface adsorption reactions occurring with
ﬂuoride in the use of BDD for PFAS oxidation. To further use these materials for sensing and
oxidation experiments, more advanced methods need to be utilized.

1.2.2 Square Wave Voltammetry

Square wave voltammetry (SWV) is similar to CV in that a potential range is scanned and an output
current is measured, however, this method utilizes a staircase waveform. This allows for background
noise suppression, high sensitivity, and the diagnostic ability of normal pulse voltammetry. [21]
The current is recorded at the ﬁrst pulse per cycle and at the end of the second pulse allowing for a
di erence current between the two points to be calculated. [21] A ﬁgure depicting this waveform
in provided in Figure 1.4 as well as the corresponding graph depicting a typical current response.
This pause after the potential step, before the current is measured, allows for non-faradaic current to
dissipate so that there is a lower background current and, subsequently, lower limits of detection for

7

the speciﬁed compound can be obtained. Non-faradaic current is the current that is not associated
with a the direct consumption of reactants to form products describes by Faraday’s law such as the
current from the formation of the double layer at the electrode’s surface. [18] Due to the sensitivity
of this technique care must be taken to ensure the cleanliness of the analysis equipment to avoid
interference from other compounds.

Figure 1.4: (left) Potential waveform and (right) typical forward, reverse, and di erence current
response for square wave voltammetry

This method can be utilized in combination with the standard addition method to obtain the
limit of detection (LOD) for a given compound when using a speciﬁc working electrode material.
In order to do this, a scan is performed with the desired working electrode in a pure electrolyte
solution. Next a set concentration of analyte is added to the solution and the scan is repeated. After
this, an additional concentration is added and the scan is performed again. The cycle of adding
analyte and scanning the potential range continues until there are enough scans that a graph can
be generated for the current obtained versus the added concentration. This graph should provide
a linear relationship between the concentration and the measured current. The slope of this line
provides the sensitivity (m) that the electrode has towards the speciﬁc compound detection.
In
order to ﬁnd the LOD, the scan for the lowest concentration of analyte is repeated 3 times. [22]
The standard deviation of the peak height (s) is then multiplied by 3 and divided by the obtained
sensitivity to ﬁnd the LOD, Equation 1.5. In order to get statistical data, the experiment can be
repeated multiple times in order to obtain a standard deviation. This method is utilized in the

8

following chapters to show that full BDD sensors, comprising of a working, auxiliary, and quasi
reference electrode, can be developed to reach lower detection limits for a biological compound is
a shorter amount of time than other techniques.

LOD = (3s)
m

(1.5)

This method is su cient for a variety of biological compounds, but in order to do heavy
metal analysis an additional step is needed. This leads to the method of square wave stripping
voltammetry.

1.2.3 Square Wave Stripping Voltammetry

Stripping voltammetry advances SWV by adding in a pre-concentration or deposition step. In this
pre-concentration/deposition step of anodic stripping voltammetry (ASV), the metal ion is reduced
from its ionic state to its elemental state onto the working electrode’s surface. In cathodic stripping
voltammetry (CVS), the metal ion is oxidized from its ionic state to a higher oxidation state and
plates onto the electrode’s surface as an oxide. After su cient pre-concentration, the analyte is
e ectively removed, or stripped, from the electrode surface by a potential sweep in the opposite
direction from the pre-concentration potential.
In general, these methods are known as square
wave stripping voltammetry (SWSV) when they use the same square wave waveform as previously
mentioned. The solution is stirred during the preconditioning step where a su cient potential
is applied to plate the analyte on the electrodes surface, the solution settles during a few second
rest, then the analyte is removed from the electrode’s surface through scanning the potential in
the opposite direction. The important parameters for this method include the deposition potential
and time. These must be carefully chosen to ensure adequate deposition of the metal onto the
electrode’s surface without sacriﬁcing on measurement time.

Not only can this method be used to determine the concentration of ions in solution, but it can
also provide information about the electrochemical activity of an electrode based on where the
ions are depositing onto the surface. For example, a metal can be plated onto the surface using

9

the pre-conditioning/ deposition step before removing the electrode from solutions and imaging
the surface to determine where the metal was actually deposited. When using anodic or cathodic
stripping voltammetry, the same approach used for SWV to determine the LOD of a compound can
be used. This method is utilized in the following chapters to advance the understanding of indium
tin oxide (ITO) as an electrode material for sensing applications in a way that lowers costs of sensor
fabrication with the ability to mass produce sensors from this material and allows for repeatable
and reliable results with minimal pretreatments.

1.2.4 Electrochemical Oxidation

Electrochemical oxidation (EO) has been used for wastewater treatment studies and commercial
applications throughout history.
[23–30] This method works by direct or indirect oxidation at
atmospheric conditions with no additional chemicals. The oxidation occurs with the transfer of one
electron from the species in solution to the anode or the generation of oxidizing agents at the anodes
surface that can then oxidize the contaminant in the bulk solution. There is no need for a reference
electrode in this experimental setup. An anode and a cathode are directly connected to a power
supply that controls the applied current or voltage. Oxidation occurs at the anode and involves
the transfer of an electron from the species of interest in the electrolyte solution to the electrode,
subsequently oxidizing the species in solution through a direct electron transfer process. In order
for this exchange to occur in a direct process, the species is required to be within an appropriate
distance from the electrode surface before di using away. Therefore, both mass transport and
electron transfer are important. This oxidation can also happen by having other compounds in
solution transform into oxidizing agents, such as hydroxyl radicals, that can oxidize the compound
of interest away from the electrode’s surface. This is known as indirect oxidation. This method is
utilized in the following chapters to advance the management of PFAS waste from ion exchange
processes in order to stop the contamination cycle from continuing to endanger the population,
while exploring energy conscious techniques.

10

1.3 Background on Electrochemical Sensing

The three classiﬁcations of electrochemical sensing are conductimetric, potentiometric, and
voltammetric. The work here focused on voltammetric sensing which is based on recording a
current-voltage curve over the course of time. Advantages of these electrochemical sensors include
the choice of electrode material having an inﬂuence on the ability to separate two similarly oxidizable
or reducible species, having high selectivity and speciﬁcity due to species undergoing a reaction
at a speciﬁc potential, having the ability to apply complex potential waveforms leading to sensors
with low detection limits, and the ability to miniaturize these sensors. [31] The ﬁrst instance of
voltammetric sensing came about after the invention of the dropping mercury electrode (DME)
in 1922. [31] This electrode allowed for the study of reducible species due to its wide negative
potential window. Shortly after in 1925, a polarograph was invented that allowed for a current-
voltage curve to be recorded photographically, but it was not until 1939 that the ﬁrst commercial
instrument became availible. [32] In order to advance sensing from the negative potential range to
cover both the negative and the positive potential range, more advanced solid electrode materials
began to be investigated. These included materials such as platinum, gold, or carbon based ﬁlms. In
the 1950s more advanced polarography appeared, including the use of square wave, when circuits
became available.
[31] Once the development of the instrumentation became widely available,
work in stripping analysis began to be investigated and compared to the standard voltammetric
analysis that was previously performed. [32]

Both linear scan voltammetry and cyclic voltammetry grew in popularity in the 1960s. [32]
Microelectrodes began to make an appearance after this time and the ﬁrst comprehensive review
of them was written in 1981. [32] It was found that microelectrodes have advantageous properties
such as immunity to solution resistance, an increase in current density, and a lower double-
layer capacitance. [32] Additionally, in the 1980s fast-scan linear sweep voltammetry and cyclic
voltammetry were added to high performance liquid chromatography (HPLC) to extend the range
of analytes that could be detected. [32] All these advancements require the understanding of the

11

material being used as the working electrode. The ability for a material to be applicable as a good
electrode material for voltammetric sensing can rely on its structure. Two newer electrode materials
are that of boron-doped diamond (BDD) and indium tin oxide (ITO). Discussed further in Section
1.3.1 are the structures of the electrode materials of BDD and ITO and how they play a role in the
properties that each material possesses.

1.3.1 Structure-property Connections

1.3.1.1 Boron-doped Diamond
The process parameters used when depositing a BDD ﬁlm can have an inﬂuence on its physio-
chemical properties. These parameters include the level of boron dopant. The boron doping
level of a diamond ﬁlm can have an inﬂuence on the surface and physical properties such as the
electrical conductivity, sp2 carbon content, surface termination, and surface area.
[33] Boron
content inﬂuences the grain size of the material with ﬁlms that have a higher boron concentration
have smaller grain sizes. [33, 34] The wetting ability of the electrode’s surface is also an inﬂuence
of the boron content of the ﬁlm. It has been shown that ﬁlm’s with a lower boron concentration
have a hydrophobic surface whereas those with a higher boron content show characteristics of
an oxygen terminated ﬁlm. [33, 34] A lower boron concentration allows for electrodes with an
increased electrochemical potential window due to their higher resistivity. [34, 35] In addition, a
higher boron content allows for a higher carrier mobility, a necessary parameter for electrochemical
sensing. [35]

One can also relate the boron content of a diamond ﬁlm to its ability to be used in certain
electrochemical applications. For example, electrochemical nitrate reduction has been shown to
have a larger peak potential and occur at a more positive potential for a BDD ﬁlm with a higher
boron content that has been attributed to a higher conductivity value. [33]. The highly doped ﬁlm
ultimately had a higher sensitivity and lower detection limit for nitrate compared to the other that
was not as heavily doped. The boron content has also been related to the detection of uric acid
on BDD ﬁlms. An increase in the linear detection range and a decrease in the detection limit is

12

obtained with an increase in boron concentration. [34] For the oxidation of sulphamerazine on
BDD, a higher concentration of boron to carbon was more e ective. [35] Similarly, the e ciency
of the electrochemical oxidation of azo dyes was inﬂuenced by the boron concentration of the ﬁlm.
An increase the doping level increased the e ciency of the process. [36]

Indium Tin Oxide

1.3.1.2
Similar to BDD, the deposition parameters used when depositing ITO through sputtering can
a ect the structural, electrical, and optical properties of ITO that can come into play when utilizing
the material as an electrode for a sensing application. The ﬁrst process dependent report on the
use of ITO for electrochemical applications was published in 1992. [37] More recently, a report
detailing how the microstructure and doping of the ﬁlm can have an inﬂuence on it’s electrochemical
properties was published in 2002. [38]

An electrochemical comparison between an as grown ITO ﬁlm and one with the application of
vacuum annealing has been made. It was found that more reversible behavior was obtained on the
ﬁlm that was vacuum annealed compared to the as-grown ﬁlm. [37] This is suggested to be due
to the fact that the resistivity decreases with annealing and increases the free carrier concentration.
[37] The vacuum annealed ﬁlms also allowed for a lower peak separation in peak potential than
those obtained on as-grown ﬁlms due to their increase in oxygen vacancies. [37] Although this study
had very limited ﬁlm characterization, it showed that the ﬁlm that went through post-deposition
annealing in a vacuum had a superior performance than the as-grown ﬁlm.

ITO can have an amorphous or crystalline microstructure. The shift to crystallinity comes from
the application of heat during deposition. In amorphous ﬁlms, the carrier concentration is low with
a high mobility due to the absence of crystal defects, but upon crystallization, the carrier density
increases and the mobility decreases. [38] It has also been shown that the tin in an amorphous
state is not electrically active in contrast to the crystalline material [38] The electron transfer rates
for three common redox couples were lower for the amorphous ﬁlm with lower peak currents
and larger peak separations. [38] This study provided evidence that the amorphous ITO ﬁlm was

13

not an adequate electrode for the oxidation of guanine. This study provides evidence that the
microstructure of the ﬁlm has an inﬂuence on its ability to be used in electrochemical applications
such as voltammetric sensing.

1.3.2 Prior State of the Art

Since this report will cover three distinct topics related to elecrochemistry, it is important to provide
the prior state of the art for each. When considering the electrochemical detection of isatin, there are
three primary works prior to this investigation. Two of which cover the electrochemical detection
of isatin after an initial system such as high-performance liquid chromatography (HPLC) or ﬂow
injection analysis (FIA), whereas, the third looks directly at the ﬁrst attempt to detect isatin solely
based on electrochemical methods. [6, 39, 40] Table 1.1 shows each study and the ﬁgures of merit
that were obtained. The di erential pulse study obtained similar limits of detections (LOD) to both
the electrodes used after FIA and the glassy carbon electrode used after HPLC. The lowest obtained
LOD was from using the chemically modiﬁed palladium hexacyanoferrate electrode after HPLC.
However, it is important to simplify the electrode requirements for this process to make it more
accessible.

Manganese detection using electrochemistry is not a new topic. To consolidate the prior
state of the art research, the detection of only manganese speciﬁcally using cathodic stripping
voltammetry in the last 25 years was researched. Working electrodes without any use of membranes
or additional techniques, such as ultrasonic or prior column methods, were focused on that had
no other interfering metals. The prior art scope was also limited to manganese detection in
simulated solutions so that it would be more applicable to the study performed here. Electrode
materials in this prior art research include modiﬁed carbon paste, glassy carbon, carbon ﬁlm,
graphite composite, boron-doped diamond, palladium, carbon nanotube, indium tin oxide, graphite,
and platinum.[41, 42, 8, 43–46, 14, 47–49] Table 1.2 compares the obtained limit of detection
and linear range for each study as well as the pretreatment the working electrode went through
It is important to consider an economical working
and the electrochemical parameters used.

14

Table 1.1: Summary table outlining the prior art for the electrochemical detection of isatin

electrode material for this application with minimal pretreatments to be able to advance this
sensing technology to a single use sensor for the medical or environmental ﬁeld.

The prior art research focused on perﬂuoroalkyl substance (PFAS) oxidation with electrochem-
istry focused on the studies that looked at using BDD as the anode material. There are ﬁve
categories included in this review: those involving industrial waste waters, those in natural water
matrix or ground water, simulated solutions with only perﬂuorooctanoic acid (PFOA), simulated
solutions starting with perﬂuorooctane sulfonate (PFOS) or perﬂuorobutane sulfonic acid (PFBS),
and studies that start with a mixture of PFAS. These categories are split into two tables, Table 1.3
that covers the simulated solutions studies and Table 1.4 that shows the PFAS studies performed in
industrial or ground water solutions.

Table 1.3 covers four studies with just PFOA, one study with just PFOS, one study with just

15

Table 1.2: Summary table outlining the prior art for the electrochemical detection of manganese
in simulated bu er solutions

PFBS, and one study with a mixture of a few chain lengths. [50–55, 27] Some of the removal
percentages were estimated based on the provided ﬁgures in the literature. Most studies achieved
greater than 50 % degradation for their perspective compound in under 10 hours.

Table 1.4 shows the studies that were performed in complex solutions. These include those
of industrial wastewater and contaminated groundwater. [56–61] All of those obtained removal
percentages of greater than 90 %. In terms of advancing the prior state of the art, it is important
to look at minimizing the treatment time needed for these solutions while also reducing the energy
consumption that is required.

16

Table 1.3: Summary table outlining the prior art for the simulated PFAS solutions

Table 1.4: Summary table outlining the prior art for the complex solutions containing PFAS

1.3.3 Deﬁciencies of Technology

Although electrochemical voltammetric sensing has advanced substantially over the years, there
are still a few deﬁciencies in the technology. Noble metals and advanced diamond materials can be
costly. In addition, in order to be more convenient, electrodes themselves need to be miniaturized

17

as well as the necessary equipment. For example the development of sensors that comprise of the
working, counter, and auxiliary electrodes all on a single platform. Both cost e ective materials
and miniaturization of the sensing technology will allow for more accessible sensors and more
single use sensors. In addition, the reduction in time and energy for PFAS remediation will play an
important role for it to be used in industry.

18

BIBLIOGRAPHY

19

BIBLIOGRAPHY

[1]

JM Halket, PJ Watkins, A Przyborowska, BL Goodwin, Angela Clow, Vivette Glover, and
M Sandler. Isatin (indole-2, 3-dione) in urine and tissues: Detection and determination by gas
chromatography—mass spectrometry. Journal of Chromatography B: Biomedical Sciences
and Applications, 562(1-2):279–287, 1991.

[2] Shigeo Manabe, Qiang Gao, Juan Yuan, Toshie Takahashi, and Akira Ueki. Determina-
tion of isatin in urine and plasma by high-performance liquid chromatography. Journal of
Chromatography B: Biomedical Sciences and Applications, 691(1):197–202, 1997.

[3] Zeid A Alothman, Masoom R Siddiqui, and Saikh M Wabaidur. Estimation of isatin in
spiked plasma samples using high performance liquid chromatography. Chemical Industry
and Chemical Engineering Quarterly, 19(4):605–614, 2013.

[4] Shigeo Manabe, Qiang Gao, Juan Yuan, Toshie Takahashi, and Akira Ueki. Determina-
tion of isatin in urine and plasma by high-performance liquid chromatography. Journal of
Chromatography B: Biomedical Sciences and Applications, 691(1):197–202, 1997.

[5] Alexei E Medvedev, Angela Clow, Merton Sandler, and Vivette Glover. Isatin: a link between

natriuretic peptides and monoamines? Biochemical pharmacology, 52(3):385–391, 1996.

[6] Haihong Xu, Dan Wang, Wen Zhang, Wei Zhu, Katsunobu Yamamoto, and Litong Jin. Deter-
mination of isatin and monoamine neurotransmitters in rat brain with liquid chromatography
using palladium hexacyanoferrate modiﬁed electrode. Analytica chimica acta, 577(2):207–
213, 2006.

[7] Richard J Reay, Anthony F Flannery, Christopher W Storment, Samuel P Kounaves, and Gre-
gory TA Kovacs. Microfabricated electrochemical analysis system for heavy metal detection.
Sensors and Actuators B: Chemical, 34(1-3):450–455, 1996.

[8] Christopher M.A. Brett Olga M.S. Filipe. Cathodic stripping voltammetry of trace mn(ii) at

carbon ﬁlm electrodes. Talanta, 61(5):643–650, 2003.
Junwei Di and Feng Zhang. Voltammetry determination of trace manganese with pretreatment
glassy carbon electrode by linear sweep voltammetry. Talanta, 60(1):31–36, 2003.

[9]

[10] Yeo May Ching, Tan Wee Tee, and Zulkarnain Zainal. Electrochemical studies of mn (ii)
mediated by li+ doped indium tin oxide (ito) electrode. Int. J. Electrochem. Sci, 6:5305–5313,
2011.

[11] Preetha Jothimuthu, Robert A Wilson, Josi Herren, Erin N Haynes, William R Heineman,
and Ian Papautsky. Lab-on-a-chip sensor for detection of highly electronegative heavy metals
by anodic stripping voltammetry. Biomedical microdevices, 13(4):695–703, 2011.

20

[12] He Xu, Qiaoli Zheng, Ping Yang, Jianshe Liu, and Litong Jin. Sensitive voltammetric detection
of trace heavy metals in real water using multi-wall carbon nanotubes/naﬁon composite ﬁlm
electrode. Chinese Journal of Chemistry, 29(4):805–812, 2011.

[13] Cory A Rusinek, Wenjing Kang, Keaton Nahan, Megan Hawkins, Cooper Quartermaine,
Angela Stastny, Adam Bange, Ian Papautsky, and William R Heineman. Determination
of manganese in whole blood by cathodic stripping voltammetry with indium tin oxide.
Electroanalysis, 29(8):1850–1853, 2017.

[14] Cory A Rusinek, Adam Bange, Mercedes Warren, Wenjing Kang, Keaton Nahan, Ian Pa-
pautsky, and William R Heineman. Bare and polymer-coated indium tin oxide as working
electrodes for manganese cathodic stripping voltammetry. Analytical chemistry, 88(8):4221–
4228, 2016.

[15] Sophia Karastogianni and Stella Girousi. Electrochemical behavior and voltammetric de-
termination of a manganese (ii) complex at a carbon paste electrode. AnAlyticAl chemistry
insights, 11:ACI–S32150, 2016.

[16] Mahmoud Roushani, Zahra Saedi, Fatemeh Hamdi, and Behruz Zare Dizajdizi. Preparation
an electrochemical sensor for detection of manganese (ii) ions using glassy carbon electrode
modiﬁed with multi walled carbon nanotube-chitosan-ionic liquid nanocomposite decorated
with ion imprinted polymer. Journal of Electroanalytical Chemistry, 804:1–6, 2017.

[17] Sergi Garcia-Segura, Joey D Ocon, and Meng Nan Chong. Electrochemical oxidation reme-
diation of real wastewater e uents—a review. Process Safety and Environmental Protection,
113:48–67, 2018.

[18] Thomas F Fuller and John N Harb. Electrochemical engineering. John Wiley & Sons, 2018.
[19] Richard S Nicholson. Theory and application of cyclic voltammetry for measurement of

electrode reaction kinetics. Analytical Chemistry, 37(11):1351–1355, 1965.

[20] RJ Klingler and JK Kochi. Electron-transfer kinetics from cyclic voltammetry. quantitative
description of electrochemical reversibility. The Journal of Physical Chemistry, 85(12):1731–
1741, 1981.

[21] Allan J. Bard and Larry R. Faulkner. Electrochemical Methods Fundamentals and Applica-

tions. John Wiley sons, Inc., United States of America, second edition, 2011.

[22] Daniel C. Harris. Quantitative Chemical Analysis. Clancy Marshall, New York, 8th edition,

2010.

[23] Angela Anglada, Ane Urtiaga, and Inmaculada Ortiz. Contributions of electrochemical
oxidation to waste-water treatment: fundamentals and review of applications. Journal of
chemical technology and biotechnology, 84(12):1747–1755, 2009.

[24] Minghua Zhou, Heikki Särkkä, and Mika Sillanpää. A comparative experimental study
on methyl orange degradation by electrochemical oxidation on bdd and mmo electrodes.
Separation and Puriﬁcation Technology, 78(3):290–297, 2011.

21

[25] Qiongfang Zhuo, Meiqing Luo, Qingwei Guo, Gang Yu, Shubo Deng, Zhencheng Xu,
Bo Yang, and Xiaoliang Liang. Electrochemical oxidation of environmentally persistent
perﬂuorooctane sulfonate by a novel lead dioxide anode. Electrochimica Acta, 213:358–367,
2016.

[26] Qiongfang Zhuo, Shubo Deng, Bo Yang, Jun Huang, and Gang Yu. E cient electrochemical
oxidation of perﬂuorooctanoate using a ti/sno2-sb-bi anode. Environmental science technol-
ogy, 45(7):2973–2979, 2011.

[27] Qiongfang Zhuo, Shubo Deng, Bo Yang, Jun Huang, Bin Wang, Tingting Zhang, and Gang
Yu. Degradation of perﬂuorinated compounds on a boron-doped diamond electrode. Elec-
trochimica Acta, 77:17–22, 2012.

[28] Marek Trojanowicz, Anna Bojanowska-Czajka, Iwona Bartosiewicz, and Krzysztof Kulisa.
Advanced oxidation/reduction processes treatment for aqueous perﬂuorooctanoate (pfoa) and
perﬂuorooctanesulfonate (pfos)–a review of recent advances. Chemical Engineering Journal,
336:170–199, 2018.

[29] BV Sarada, Tata N Rao, DA Tryk, and A Fujishima. Electrochemical oxidation of histamine
and serotonin at highly boron-doped diamond electrodes. Analytical chemistry, 72(7):1632–
1638, 2000.

[30] MA Rodrigo, P Cañizares, A Sánchez-Carretero, and C Sáez. Use of conductive-diamond
electrochemical oxidation for wastewater treatment. Catalysis Today, 151(1-2):173–177,
2010.

[31] Christopher MA Brett and Ana Maria Oliveira-Brett. Electrochemical sensing in solu-
tion—origins, applications and future perspectives. Journal of Solid State Electrochemistry,
15(7-8):1487–1494, 2011.

[32] Allen J Bard and Cynthia G Zoski. Voltammetry retrospective., 2000.
[33] JT Matsushima, WM Silva, AF Azevedo, MR Baldan, and NG Ferreira. The inﬂuence of
boron content on electroanalytical detection of nitrate using bdd electrodes. Applied surface
science, 256(3):757–762, 2009.

[34] Peng Yu, Jiawei Zhang, Tong Zheng, and Ting Wang. Inﬂuence of boron doped level on the
electrochemical behavior of boron doped diamond electrodes and uric acid detection. Colloids
and Surfaces A: Physicochemical and Engineering Aspects, 494:241–247, 2016.

[35] Aleksandra FabiaÒska, Robert Bogdanowicz, Patrycja Zi˛eba, Tadeusz Ossowski, Marcin
Gnyba, Jacek Ryl, Artur Zielinski, Sto el D Janssens, Ken Haenen, and Ewa M Siedlecka.
Electrochemical oxidation of sulphamerazine at boron-doped diamond electrodes: Inﬂuence
of boron concentration. physica status solidi (a), 210(10):2040–2047, 2013.

[36] Robert Bogdanowicz, Aleksandra FabiaÒska, L Golunski, Micha≥ Sobaszek, Marcin Gnyba,
Jacek Ryl, Kazimierz Darowicki, Tadeusz Ossowski, Sto el D Janssens, Ken Haenen, et al.
Inﬂuence of the boron doping level on the electrochemical oxidation of the azo dyes at si/bdd
thin ﬁlm electrodes. Diamond and related materials, 39:82–88, 2013.

22

[37] MA Martinez, J Herrero, and MT Gutierrez. Electrochemical stability of indium tin oxide

thin ﬁlms. Electrochimica acta, 37(14):2565–2571, 1992.

[38] Natasha D Popovich, Sze-Shun Wong, Brian KH Yen, Hyo-Young Yeom, and David C Paine.
Inﬂuence of microstructure on the electrochemical performance of tin-doped indium oxide
ﬁlm electrodes. Analytical chemistry, 74(13):3127–3133, 2002.

[39] Romana Jaroöová, Simon Sanchez, Lars Haubold, and Greg M Swain. Isatin analysis using
ﬂow injection analysis with amperometric detection–comparison of tetrahedral amorphous
carbon and diamond electrode performance. Electroanalysis.

[40] Victor C Diculescu, Sunita Kumbhat, and Ana Maria Oliveira-Brett. Electrochemical be-
haviour of isatin at a glassy carbon electrode. Analytica chimica acta, 575(2):190–197,
2006.

[41] Soo Beng Khoo, Ming Koon Soh, Qiantao Cai, Miriam R Khan, and Si Xuan Guo. Di erential
pulse cathodic stripping voltammetric determination of manganese (ii) and manganese (vii) at
the 1-(2-pyridylazo)-2-naphthol-modiﬁed carbon paste electrode. Electroanalysis, 9(1):45–
51, 1997.

[42] John Stanislaus Roitz and KW Bruland. Determination of dissolved manganese (ii) in coastal
and estuarine waters by di erential pulse cathodic stripping voltammetry. Analytica Chimica
Acta, 344(3):175–180, 1997.

[43] Ji-Ye Jin, Fan Xu, and Tomoo Miwa. Cathodic stripping voltammetry for determination of
trace manganese with graphite/styrene-acrylonitrile copolymer composite electrodes. Elec-
troanalysis: An International Journal Devoted to Fundamental and Practical Aspects of
Electroanalysis, 12(8):610–615, 2000.

[44] Craig E Banks, Jaanus Kruusma, Ryan R Moore, Peter Tom ík, Judith Peters, James Davis, äe-
bojka Komorsky-LovriÊ, and Richard G Compton. Manganese detection in marine sediments:
anodic vs. cathodic stripping voltammetry. Talanta, 65(2):423–429, 2005.

[45] Wenjing Kang, Xing Pei, Adam Bange, Erin N Haynes, William R Heineman, and Ian Pa-
pautsky. Copper-based electrochemical sensor with palladium electrode for cathodic stripping
voltammetry of manganese. Analytical chemistry, 86(24):12070–12077, 2014.

[46] Wei Yue, Adam Bange, Bill L Riehl, Bonnie D Riehl, Jay M Johnson, Ian Papautsky,
and William R Heineman. Manganese detection with a metal catalyst free carbon nanotube
electrode: anodic versus cathodic stripping voltammetry. Electroanalysis, 24(10):1909–1914,
2012.

[47] M Christine Welch, E Craig Banks, äebojka Komorsky-LovriÊ, and G Richard Compton.
Electroanalysis of trace manganese via cathodic stripping voltammetry: Exploration of edge
plane pyrolytic graphite electrodes for environmental analysis. 2006.

[48] Miroslav Rievaj, Peter Tomcik, Zuzana Janosikova, Du≥an Bustin, and Richard G Compton.
Determination of trace mn (ii) in pharmaceutical diet supplements by cathodic stripping
voltammetry on bare carbon paste electrode. Chem. Anal.(Warsaw), 53:153–161, 2008.

23

[49] Wenjing Kang, Cory Rusinek, Adam Bange, Erin Haynes, William R Heineman, and Ian Pa-
pautsky. Determination of manganese by cathodic stripping voltammetry on a microfabricated
platinum thin–ﬁlm electrode. Electroanalysis, 29(3):686–695, 2017.

[50] Tsuyoshi Ochiai, Yuichi Iizuka, Kazuya Nakata, Taketoshi Murakami, Donald A Tryk, Akira
Fujishima, Yoshihiro Koide, and Yuko Morito. E cient electrochemical decomposition of
perﬂuorocarboxylic acids by the use of a boron-doped diamond electrode. Diamond and
Related Materials, 20(2):64–67, 2011.

[51] Ane Urtiaga, Carolina Fernández-González, Sonia Gómez-Lavín, and Inmaculada Ortiz. Ki-
netics of the electrochemical mineralization of perﬂuorooctanoic acid on ultrananocrystalline
boron doped conductive diamond electrodes. Chemosphere, 129:20–26, 2015.

[52] Hanshuang Xiao, Baoying Lv, Guohua Zhao, Yujing Wang, Mingfang Li, and Dongming
Li. Hydrothermally enhanced electrochemical oxidation of high concentration refractory
perﬂuorooctanoic acid. The Journal of Physical Chemistry A, 115(47):13836–13841, 2011.
[53] NW Khun, E Liu, and XT Zeng. Corrosion behavior of nitrogen doped diamond-like carbon

thin ﬁlms in nacl solutions. Corrosion Science, 51(9):2158–2164, 2009.

[54] Kimberly E Carter and James Farrell. Oxidative destruction of perﬂuorooctane sulfonate using
boron-doped diamond ﬁlm electrodes. Environmental science technology, 42(16):6111–6115,
2008.

[55] Zhaohui Liao and James Farrell. Electrochemical oxidation of perﬂuorobutane sulfonate using
boron-doped diamond ﬁlm electrodes. Journal of applied electrochemistry, 39(10):1993–
1999, 2009.

[56] Beatriz Gomez-Ruiz, Sonia Gómez-Lavín, Nazely Diban, Virginie Boiteux, Adeline Colin,
Xavier Dauchy, and Ane Urtiaga. Boron doped diamond electrooxidation of 6: 2 ﬂuorotelom-
ers and perﬂuorocarboxylic acids. application to industrial wastewaters treatment. Journal of
Electroanalytical Chemistry, 2017.

[57] Beatriz Gomez-Ruiz, Sonia Gómez-Lavín, Nazely Diban, Virginie Boiteux, Adeline Colin,
Xavier Dauchy, and Ane Urtiaga. E cient electrochemical degradation of poly-and perﬂu-
oroalkyl substances (pfass) from the e uents of an industrial wastewater treatment plant.
Chemical Engineering Journal, 322:196–204, 2017.

[58] AM Trautmann, H Schell, KR Schmidt, K-M Mangold, and A Tiehm. Electrochemical
degradation of perﬂuoroalkyl and polyﬂuoroalkyl substances (pfass) in groundwater. Water
Science and Technology, 71(10):1569–1575, 2015.

[59] Álvaro Soriano, Daniel Gorri, and Ane Urtiaga. E cient treatment of perﬂuorohexanoic
acid by nanoﬁltration followed by electrochemical degradation of the nf concentrate. Water
Research, 112:147–156, 2017.

[60] Charles E Schaefer, Christina Andaya, Aniela Burant, Charles W Condee, Ana Urtiaga, Timo-
thy J Strathmann, and Christopher P Higgins. Electrochemical treatment of perﬂuorooctanoic

24

acid and perﬂuorooctane sulfonate: Insights into mechanisms and application to groundwater
treatment. Chemical Engineering Journal, 317:424–432, 2017.

[61] Charles E Schaefer, Sarah Choyke, P Lee Ferguson, Christina Andaya, Aniela Burant, Andrew
Maizel, Timothy J Strathmann, and Christopher P Higgins. Electrochemical transformations
of perﬂuoroalkyl acid (pfaa) precursors and pfaas in groundwater impacted with aqueous ﬁlm
forming foams. Environmental science & technology, 52(18):10689–10697, 2018.

25

CHAPTER 2

ISATIN DETECTION USING A BORON-DOPED DIAMOND 3-IN-1 SENSING

PLATFORM

26

2.1

Introduction

This work has been published in Analytical Chemistry1.

Indole-2,3-dione, or isatin, has
attractive properties as an endogenous indole that could be applied in the treatment of Parkinson’s
Disease, tremors, or epilepsy. [1–3] Its structure is shown in Figure 2.1. Isatin has a large amount
of biological activity and is found in the tissues and ﬂuids of mammals. [4–7] In 1988 scientists
discovered this compound using the technique of gas chromatography mass spectroscopy (GCMS).
[4] Concentrations of isatin range from above 1 µmol/L in blood to 0.1 to 10 µmol/l in tissues and
greater than 1 mmol/L in urine samples. [6, 8, 7] This compound has been of interest due to its
ability to bring on stress and anxiety, the inhibition of sodium in rat intestines, and the potential
link it may provide between a person’s mood, the natriuretic peptide system, and monoamine
neurotransmitter levels. [6, 9, 10] Moderate concentrations of isatin, 15-20 mg/kg of body weight,
have been found to be anxiogenic but a higher dose, 50 mg/kg of body weight, can induce sedation in
rats and larger mammals such as monkeys. [9] To gain a better understanding of all the interactions
isatin is capable of and how to take advantage of it as a pharmaceutical drug, a more complete
understanding of how it interacts in the human body and the mechanisms it goes through is essential.
To achieve this, precise measurement techniques for the detection of isatin are necessary.

Since the discovery of isatin using GCMS, liquid chromatography (LC) coupled with UV-vis
spectroscopy or an electrochemical voltammetric method have been used as detection methods.
[1, 8, 7, 11–14] A downside to GCMS is the long measurement times required. [14] A research
group using high performance liquid chromatography (HPLC) with UV-vis measured isatin’s
absorbance peak at 242 nm with a limit of detection (LOD) of 2 µmol/L. [14] Another group using
LC with electrochemical detection with a palladium hexacyanoferrate modiﬁed electrode measured
isatin in rat brains with a LOD of 0.03 µmol/L. [1] Diculescu et al. used UV-vis with a found LOD
of 0.8 µmol/L and began to study the detection possibilities of using electrochemical techniques
without the addition of a chromatography techniques. [11]

1Analytical Chemistry 2018, DOI:10.1021/acs.analchem.7b04045

27

Figure 2.1: Molecular structure for isatin

Isatin has both an oxidation and reduction reaction that are irreversible and dependent on pH
that can be detected via electrochemistry.
[11] When utilizing a glassy carbon (GC) electrode,
isatin displays an oxidation peak at +1.08 V and a reduction peak -0.48 V. [7, 11] One group used
a GC electrode and multiple voltammetry techniques to study isatin and were able to obtain a LOD
of 0.19 µM when using di erential pulse voltammetry. [11] Other groups have also studied isatin
on GC electrodes. [3, 7, 12]

Boron-doped diamond (BDD) is another electrode material known for its outstanding properties
when used for the electrochemical detection of biological compounds. [15–17] BDD has many
electrochemical advantages such as its wide potential window, low background current, chemical
inertness, and its good chemical and electrochemical stability. [18, 19] The wide potential window
that is characteristic of BDD o ers the ability to detect compounds that have higher positive or
negative redox potentials. BDD electrodes have excellent microstructure stability even at these
extreme potential values.
[18, 19] With minimal preconditioning steps, BDD displays quasi-
reversible redox systems with resistance to deactivation and fouling. [20, 21] Additionally, BDD
can be fabricated into di erent electrode microstructures such as micro electrode arrays (MEAs)

28

that can then be used as micro-sensors for free hydroxyl radical generation and neural activity
measurements. [21–29]

One factor that can produce unwanted results in experiments is the potential drop due to solution
resistance which can shift peak potential and increase peak separation. MEAs fabricated out of
BDD o er a lower electrical potential di erence, or IR drop, due to their lower electrode surface area
compared to typical macro (MAC) electrodes and they can provide steady state current responses
that MAC electrodes can not. [23, 24] After a potential step, MEAs require a short time to reach
equilibrium. [24] One group reported that an ultramicroelectrode array (UMEA) showed 50-100
times better sensitivity than a traditional MAC electrode with a superior reproducibility. [30] In
order to simplify the usability of MEAs and prevent breakage and loss, sensing platforms have been
fabricated that consist of the working, auxiliary, and reference electrodes. [30–40]

Previously, A 3-in-1 sensor was made from silver (Ag) and gold (Au) electrodes to detect nitrate
(NO3-) and another was fabricated with carbon nanotubes for the detection of glucose. [39, 40]
Many of these sensors use Ag/AgCl or Ag/Ag+ as a reference or quasi-reference electrode material.
[34, 37, 39, 40] One research group fabricated a device with a BDD quasi-reference material
but failed to mention how it behaved.
[35] However, two groups using sensors made with all
BDD components, demonstrated that BDD could behave as an acceptable quasi-reference material.
[38, 30] To evaluate BDD’s performance as a reference material, potential controlled experiments
were performed to determine its stability for 57 hours and showed that the potential only drifted by
+/- 8 mV. [38]

For the remainder of this chapeter, two fully BDD 3-in-1 sensors, one MAC and one MEA,
were fabricated and characterized. Both sensors were assessed for their ability to detect isatin.
The BDD quasi-reference material was characterized based on its potential stability over time
compared to a standard Ag/AgCl reference electrode in pH solutions of 4.0 to 7.4. The BDD MAC
electrode had a working area of 3.14 x 10-2 cm2. The MEA’s total working area amounted to 1.71
x 10-4 cm2. The BDD material was characterized with Raman spectroscopy and scanning electron
microscopy (SEM). The development of such a BDD sensing platform could be used in future

29

work to detect isatin in applications involving liquid chromatography coupled to electrochemical
detection methods while removing the need for three separate electrodes and simplifying the
required electrode pretreatment.
2.2 Experimental Section

2.2.1 Chemicals and Materials

Potassium ferrocyanide (K4Fe(CN)6) and potassium ferricyanide (K3Fe(CN)6) were purchased
from Mallinckrodt Specialty Chemicals and J.T. Baker, respectively.
Isatin and hexaamine-
ruthenium(III) chloride (Ru(NH3)6Cl) were purchased from Sigma Aldrich. A potassium chloride
(KCl) solution of 1.0 M was prepared with DI water and salt from ChemPure Brand Chemicals.
The KCl solution was used to prepare stock solutions of Fe(CN)63-/4- and Ru(NH3)63+/2+. A 10x
phosphate bu ered saline (PBS) solution (10.6 mM monobasic potassium phosphate, 1551 mM
sodium chloride, 29.7 mM dibasic sodium phosphate) was purchased from Gibco and diluted to
1x using DI water (1.1 mM monobasic potassium phosphate, 155 mM sodium chloride, 3.0 mM
dibasic sodium phosphate) to a ﬁnal pH of 7.4. The 10x PBS was also diluted and pH adjusted
with hydrochloric acid from Fisher Scientiﬁc to obtain pH values of 4, 5, 6, and 7. The 0.1 M PBS
was used to make a stock isatin solution. Surine Negative Urine Control was purchased from a
Simga-Aldrich Company, Ceriliant.

2.2.2

Instrumentation

A VCA Optima from AST Productions, Inc. was used to perform contact angle measurements
with a drop size of 1 µL of water. The systems software determined the contact angles of each
water drop on the sample’s surface. This measurement was repeated 3 times to obtain a mean
value and a standard deviation. A single compartment glass cell with 5 mL of solution was used
for cyclic and square wave voltammetry measurements. A clamp held each sensor in place once
immersed in solution. Each 3-in-1 sensor was electrically connected via a micro USB connector to
the leads from the potentiostat. Parameters for capacitance measurements using cyclic voltammetry

30

were: scan rates: 100-500 mV s-1, electrolyte: 1.0 M KCl. Square wave voltammetry parameters
were: frequency: 15 Hz; Amplitude, 0.05 V; increment E, 0.004 V. A CHI660C analyzer from
CH Instruments, Inc. was used. Nitrogen (N2) was bubbled for 5 minutes prior to the start
of the isatin experiments unless stated otherwise. Response curves for isatin were generated
using the standard addition method. the reproducibility of each electrode was determined from 4
scans at each concentration. The spectrophotometric analysis was performed using a UV-vis-NIR
spectrophotoeter (UV-3600) from Simadzu with quartz cuvettes that had a 10 mm path length.

2.2.3

3-in-1 Electrode Fabrication

Both the MAC and MEA were fabricated on the same conductive silicon (Si) wafer through multiple
processing steps. Microwave plasma-assisted chemical vapor deposition (MP-CVD) was ﬁrst used
to grow the BDD layer on top of the 3 inch diameter Si/ SiO2 wafer. This wafer was p-type with a
thickness o  350 µm and a resistivity between 1-5 Ohm cm. This was then coated with an insulating
SiO2 top layer that was 1 µm thick. The sensors were patterned using a photolithography mask that
allowed for the deposition of an insulating interlayer to be deposited on their surface using plasma-
enhanced chemical vapor deposition (PE-CVD). Through all of these steps, the exposed BDD,
which could be electrically connected at the top of each sensor, was the only electrochemically
active surface on the face of the sensor.

A second photoresist mask deﬁned the MAC and MEA conﬁgurations and allowed for the
three distinct electrodes on each to be patterned. The spacing between each electrode in the MEA
was carefully decided.
If the micro discs were too close together the advantages of the single
small electrodes would be lost as their depletion layers would overlap. [22, 23] Each 3-in-1 sensor
was released from the wafer by laser cutting. An earlier publication goes in to more detail on
the fabrication process. [22] Figure 2.2 shows optical micrographs of both the MEA (left) and
MAC (right) 3-in-1 sensors. The active surface area of the MAC was 3.14 x 10-2 cm2 and the 97
electrodes that comprised the hexagonal patterned MEA had a total active anode area of 1.71 x
10-4 cm2.

31

Figure 2.2: Optical micrograph of the MEA (left) and MAC (right) 3-in-1 sensors with exposed
BDD

2.3 Results and Discussion

2.3.1 Electrode Characterization

Boron is typically used as a dopant for diamond to charge carrier concentrations greater than 1020
cm-3 to obtain metal-like electrical conductivity.
[41–45] The surfaces of the MAC and MEA
were characterized using Raman spectroscopy and scanning electron microscopy (SEM). Figure
2.3 shows a Raman spectrum and SEM micrograph taken from the diamond-coated wafer that both
sensors were fabricated from.

The SEM micrograph, Figure 2.3 (A) shows grain sizes around 1-2 µm for the polycrystalline
diamond ﬁlm that was used for all sensors. The ﬁlm has no visual evidence of cracks or other
defects. The Raman spectrum has large peaks around 500 and 1200 cm-1 that are indicative of
highly doped BDD. [41, 46–48] These peaks grow in intensity with higher boron concentrations.
The boron concentration of the ﬁlm was determined by using a ﬁtting program in Matlab; providing
a concentration of 0.3 %.

Cyclic voltammetry was used to generate a current vs voltage curve to determine the capacitance

32

Figure 2.3: (A) An SEM micrograph of the polycrystalline BDD surface that was used to fabricate
the two 3-in-1 sensors with no defects or large cracks in the ﬁlms structure and (B) the
corresponding Raman spectrum analysis for the BDD wafer

values of the sensors. The MAC had a capacitance of 721 nF and the MEA had a capacitance
of 5.3 nF. The drastic di erence between the two comes from the di erent surface areas that are
exposed to solution. If the capacitances are normalized to each corresponding surface area they are
25 µF/cm2 and 31 µF/cm2 for the MAC and MEA, respectively. These capacitance values are in a
similar range to those reported for BDD in literature, 20-40 µF/cm. [49]

Cyclic voltammetry (CV) was used further to determine the electron-transfer kinetics for the
two redox couples of Fe(CN)63-/4- and Ru(NH3)63+/2+. Fe(CN)64- was used for the MAC electrode
conﬁguration whereas the MEA used a 1:1 ratio of Fe(CN)63- and Fe(CN)64-. The concentrations
of the redox couples were 1.0 mM in 1.0 M KCl. The peaks were analyzed after manual background
subtraction. [50, 51] The cyclic voltammograms for both redox couples on each 3-in-1 sensor are
shown in Figure 2.4.

Each electrode was ﬁrst conditioned in 1 M sulfuric acid (H2SO4) from -2.0 V to +2.8 V at 0.5
V/s for 30 scans to condition the electrode surface. This was decided based on the investigation

33

Figure 2.4: (A) Cyclic voltammograms of 1 mM Fe(CN)63- and 1 mM Ru(NH3)6Cl on the MAC
3-in-1 BDD sensor oth dissolved in 1 M KCl and (B) of a 1:1 ratio of Fe(CN)63-/4- and 1 mM of
Ru(NH3)6Cl , both dissolved in KCl, on the 3-in-1 BDD MEA sensor. Scan rate: 0.1 V/s.

into surface chemistry e ects as discussed next. For obtaining di erent surface terminations,
three di erent pretreatments were used: hydrogen termination (HT), oxygen termination (OT) and
electrochemical cycling (ET). The ET was done in 1.0 M H2SO4 using the procedure stated above.
For HT, the electrodes were placed in a microwave plasma system in ultra-pure hydrogen (H2) for
10 minutes. For the OT pretreatment, electrodes were placed in an oxygen (O2) atmosphere for
1 minute in a reactive ion etcher (RIE). Contact angle measurements were used to evaluate each
pretreatment process based on the obtained angle. Contact angles of 90 +/- 4.5°were measured
for HT whereas angles of 0°were measured for the OT pretreatments. These trends align with
those seen in literature for BDD surfaces. [52] The contact angles for the ET pretreatment were
predominantly oxygen terminated at 13 +/- 3.3°. However, these are slightly more hydrophobic
than the OT contact angles, suggesting di erent functional groups and surface coverage. The
ET process may be generating a slightly more hydrophilic surfaces due to a stepwise mechanism

34

involving bond breakage and carbon-oxygen bond formation. [53] Because of this, and the cleaning
of the surface impurities, the ET process was selected as the pretreatment for the remainder of this
study. It is also important to point out that this pretreatment procedure wold be accessible to the
users of the sensing technology. Not all users would have access to specialty equipment to perform
the hydrogen or oxygen termination procedures. Since the ET uses the same technology as the
method its-self, it is accessible to the common user. This pretreatment was only performed on the
electrodes before beginning a set of measurements. It was not performed between measurements
for the same calibration curve.

For the cyclic voltammograms of the redox couples, Figure 2.4, the MAC electrode provided
the expected semi-inﬁnite linear di usion response and the typical bell shaped curves. Peak
separations,   Ep, for Fe(CN)63-/4- and Ru(NH3)63+/2+ were 86 mV and 67 mV, respectively.
  Ep for Fe(CN)63-/4- on BDD electrodes is sensitive to the surface chemistry and cleanliness
whereas Ru(NH3)63+/2+ is insensitive to the surface chemistry, inner sphere or outer-sphere electron
transfer, respectively. [53–57] The heterogeneous rate constants for both of these redox couples
was determined using the Nicholson and Shain method. [58, 59] The rate constants on the MAC
for Fe(CN)63-/4- and Ru(NH3)62+/3+ were 1.74 +/- 0.10 x 10-3 cm/s and 1.17 +/- 0.09 x 10-2 cm/s,
respectively. Both redox couples indicated semi-linear di usion due to the peak currents on the
forward scans showing a linear relationship with the square root of the scan rate.

The CVs for the MEA was di erent from the MAC’s in that they displayed a sigmoidal shape
for both the oxidation and reduction of the redox couples. This shows that the reaction is limited
by di usion of the reactants to the electrodes surface due to the lack of overlap in the molecular
di usion towards each individual microelectrode causing the overall output current to be the sum
of all the elements in the array. [23, 24] Due to the small size of the microelectrodes, the reactant
transport is dominated by hemispherical di usion, improving the signal-to-noise ratio. [18, 21, 23]
This di usion proﬁle can shift back to semi-inﬁnite di usion behavior if the spacing between the
microelectrodes is too small. [22, 23]

The CV’s for the MEA are shown in Figure 2.4 (B). The rate constants for this sensor con-

35

ﬁguration were determined using fast-scan CV (FSCV) measurements and the Klinger and Kochi
method. [60] This was due to the fact that the peak separations were above the maximum limit for
the Nicholson and Shain method preciously used. The   Ep value at 100 V/s were 560 mV and
127 mV for Fe(CN)63-/4- and Ru(NH3)62+/3+, respectively. The rate constants for Fe(CN)63-/4-
and Ru(NH3)62+/3+ were 5.21 +/- 1.03 x 10-3 cm/s and 1.40 +/- 0.13 -1 cm/s, respectively. Both
redox couples displayed a linear relationship between the forward peak current and the square root
of the scan rate.

Surface morphology and cleanliness is a determining factor for Fe(CN)63-/4- reaction kinetics.
This is due to the fact that the reaction proceeds through surface sites which can become blocked
by adsorbed oxygen. [61, 47] Hydrogen terminated ﬁlms typically show rate heterogeneous rate
constants between 0.01 and 0.1 cm/s compared to oxygen terminated ﬁlms that can be 2 orders of
magnitude lower. [61, 47, 48] The resulting rate constants for both the MAC and MEA were in the
range typically displayed for oxygen terminated ﬁlms, being consistent with the studies into surface
termination after the ET pretreatment.

For Ru(NH3)62+/3+, the rate constant is inﬂuenced by the density of electronic states near its
formal potential. [57] Diamond ﬁlms have shown rate constants between 0.01 and 0.1 cm/s for
Ru(NH3)62+/3+.[47] Both the MAC and MEA fell within this range.

2.3.2 BDD as a Quasi-Reference Electrode

Since the three electrodes making up the two sensor conﬁgurations are made completely of BDD,
two questions needed to be addressed: (1) Does this BDD ﬁlm have a stable potential in solutions
of di ering pH values? (2) Is the potential stable over an extended period of measurement time?
These two questions were investigates using the MAC sensor by measuring its reference elec-
trode’s potential in solutions of various pH values compared to the potential of a Ag/AgCl electrode.
Figure 2.5 shows the potential-time curves at a pH of 4.0-7.4 over a 12 hour period of time. The
BDD’s steady-state potential shifts in a negative direction with increasing pH. This phenomenon
was reported in literature previously for diamond. [30] The potential stabilizes quickly in all the

36

tested electrolyte solutions regardless of the pH.

Figure 2.5: The BDD’s reference potential in 0.1M phosphate bu ered saline at various pH values
measured against a Ag/AgCl reference electrode for 12 hours.

Ponnuswamy attributed the change in BDD’s potential in di erent pH solutions to the reactions
between protons in the electrolyte solution and the surface.
[38] The results found here are
applicable to their suggestion that the change in proton concentration and interaction with the
diamond surface causes the shift in potential. At each pH value, the potential stabilized after the
ﬁrst hour and remained relatively unchanged for the full 12 hour period. These results indicate
that the quasi-reference material of BDD is suitable for a sensing application for a 3-in-1 device.
To further evaluate the quasi-reference material in this device, a 28 hour stability experiment was
performed. The average potential was -0.197 V with a standard deviation of 0.015 V as shown in
Figure 2.6. Although the response ﬂuctuated in the ﬁrst 4 hours, it reached a stable potential where
it remained for the remaining time. This shows that the BDD quasi reference can have a stable
response for an extended measurement time.

2.3.3

Isatin Detection

Due to the stability of the 0.1 M phosphate bu ered saline solution at pH 7.4, and its biological
relevance, it was chosen to be used for the isatin voltammetric studies. Before each measurement,
oxygen was removed from the solution by nitrogen (N2) bubbling. To further prevent oxygen from
re-entering the environment, a N2 blanket was used on top of the solution during a measurement.

37

Figure 2.6: The BDD’s reference potential for the extended 28 hours in 0.1M phosphate bu ered
saline at a pH of 7.4 measured against a Ag/AgCl electrode.

Figure 2.7 shows the CVs for 350 µM of isatin on both the (A) MAC and (B) MEA sensor
conﬁgurations at a scan rate of 0.1 V/s. Preconditioning of the sensors was done using the ET
pretreatment process as previously mentioned.

The CV for the MAC electrode, Figure 2.7 (A), shows the independent anodic and cathodic
peaks that have been reported previously. [11] The anodic peak is located at +0.832 V; -0.2 V from
the peak location reported in literature against a Ag/AgCl reference electrode. [11] This shift seen
on this electrode is due to the fact that the BDD quasi-reference displays a potential that is -0.2 V
when referenced against a Ag/AgCl reference electrode. This same phenomenon occurs with the
cathodic peak for isatin. The formal potential and half wave potential were determined as -0.034 V
and -0.686 V, respectively.

The MEA CV, Figure 2.7 (B), shows a sigmoidal response for the reactions occurring for
isatin. This sensor conﬁguration also displays a lower faradaic current and background current.
Since both sensor conﬁgurations clearly display the cathodic process and the cathodic process
avoids interferences with other compounds such as ascorbic acid and glucose, it was chosen for the
electrochemical detection comparison. [11]

Before performing the detection experiments the possible adsorption of isatin and its reduction
products were studied on the BDD surface using CV and the MAC sensor conﬁguration. The CV

38

Figure 2.7: Current versus voltage cyclic voltammogram curves for 350 µM isatin in 0.1 M PBS
at pH 7.4 on the 3-in-1 (A) MAC sensor and (B) MEA sensor. Scan rate of 0.1 V/s.

was measured for 350 µM of isatin on the sensor before removing it from solution and rinsing it
with DI water and immediately placing the sensor into a clean solution of 0.1 M PBS. N2 was
bubbled before measuring the original isatin CV and the following CV in the clean PBS bu er
for 5 minutes. Figure 2.8 shows the results from this study. It was determined that there was no
adsorption on the BDD’s surface from isatin or its reduction products even at these elevated levels.
This result coincides with that described by Diculescu et al. where they found the lack of adsorption
of isatin on a glassy carbon electrode.

Square wave voltammetry was used to study the reduction of isatin based on the standard
addition method with concentrations from 0.3 to 12 µM in 5 mL of 0.1 M PBS at a pH of 7.4. N2
bubbling was performed and blanketed on top the electrolyte solution. Both sensor conﬁgurations

39

Figure 2.8: Current versus voltage cyclic voltammogram curves of 350 µM isatin in 0.1 M PBS at
pH 7.4 (red) and after rinsing the 3-in-1 MAC electrode and rescanning the same electrode in
clean electrolyte solution (black). Scan rate of 0.1 V/s.

showed a linear range from 0 to 12 µM of isatin and a second linear range from 20 to 100 µM.
In order to determine the sensitivity, limit of detection (LOD), and limit of quantiﬁcation (LOQ)
of each sensor, three tests were performed. The LOD and LOQ were calculated using 3 / m
and 10  /m, respectively where   represents the standard deviation of the current for the lowest
concentration measured and m is the slope of the obtained calibration curve.   values for the MAC
and MEA were 7.3 nA and 0.01 nA, respectively.

After each standard addition of isatin, the solution was stirred for 10 seconds. Figure 2.9 shows
the voltammograms for 0.3 to 12 µM of isatin on the MAC (A) and MEA (B) sensors. The peak
heights at each concentration were measured from drawing a line at the background current level
and determining the distance from the background to the peak maximum. [62, 50, 51]

From Figure 2.9 two di erent peak shapes can be seen; one for the MEA and one for the MAC
electrode. The shape of the peak produced by the MEA sensor is much wider with a steep rise
on the leading edge of the peak and a drawn out decay peak. The MAC had a full width half
max (FWHM) of 0.23 V at 12 µM whereas the MEA had a FWHM of 0.26 V. The irreversibility
of the reduction of isatin is presumed to be the cause of these broad peak shapes for both sensor
conﬁgurations. SWV has been shown to result in wider peaks when analyzing irreversible systems.
[62] This phenomenon occurs due to the reverse pulse peak being more broad than the forward

40

Figure 2.9: Squarewave voltammograms for various concentrations of isatin on the 3-in-1 (A)
MAC and (B) MEA sensor conﬁgurations in 0.1 M phosphate bu ered saline with pH of 7.4.

pulse peak whereas in a reversible system they would be nearly mirror images of one another.
[63, 64, 52] The broader peak shape for these sensors can also be a result of the charge transfer
coe cient (↵) being below 0.2. ↵ values this low have been shown to distort peak shapes as seen
with those for the MAC conﬁguration. [50, 51]

The di erence in peak shapes could also be a result of the di erent di usive properties between
the two electrode conﬁgurations. On a literature reported MAC made of glassy carbon, isatin
reduction occurred in a two step reaction where the products reduced to isatide. [11] This reduction
appeared as a shoulder in their detection scans. In the study presented here, the MAC sensors have
a tail at the end of the stripping peak which could be related to the reduction of the products to
isatide on this BDD surface. The MEA does not display this same characteristic tail suggesting
the reduction products di use away faster on this sensor conﬁguration and do not become further

41

reduced. It has been shown that analytes di use away from MEA’s faster than they do when using
an MAC. [21–29]

Figure 2.10: Calibration curves for (A) the MAC and (B) the MEA in 0.1 M phosphate bu ered
saline with a pH of 7.4

Ultimately the MEA exhibited a lower sensitivity due to the standard deviation of the current
measurement being 2 orders of magnitude lower at 0.3 µM compared to the MAC. The calibration
curves for both sensor conﬁgurations are shown in Figure 2.10. The calculated LOD and LOQ were
both lower for the MEA due to this. The LOD of the MEA was 0.04 µM and that of the MAC was
0.22 µM. Both the MEA and MAC sensors displayed linearity with an excellent R2 value. These
values can be seen in Table 2.1.

Also shown in Table 2.1 is the comparison of the electrochemical results to those obtained with
UV-vis spectroscopy and other literature studies. The wavelength used for UV-Vis measurements

42

Table 2.1: Figures of merit for isatin detection using the 3-in-1 MAC and MEA sensors compared
to other reported literature studies on di erent detection methods. Standard deviations calculated
at lowest concentration detected for this study

was from 200 to 450 nm and the absorbance of isatin was recorded for concentrations of 0.3-12
µM. The absorbance spectra for these concentrations is provided as Figure 2.11. The LOD was
determined by generating a calibration curve from the peak absorbance at 242 nm. This value
was higher than that obtained for the electrochemical detection mainly due to the large standard
deviation of the signal between each test. In fact, this di erentiation is orders of magnitude higher
than that seen with both electrochemical sensors. Due to this, both the MEA and MAC have proven
to have lower detection limits than the UV-vis detection. Compared to the other electrochemical
sensing study that did not involve a pretreatment between concentration measurements, the study
performed here obtains a lower LOD and LOQ with a higher linear correlation coe cient and linear
range of concentrations. [11] The study here also advances this technology by removing the need
of separate counter and reference electrodes. Further, the use of BDD can be more reproducible
with the ET pretreatment compared to having to polish a glassy carbon electrode before each day
of measurements. Polishing drastically changes the surface of the electrode and can cause large
ﬂuctuations in data and a higher degree of uncertainty in the actual concentrations measured. BDD
is more suitable for understanding the concentrations in the human body and how the reactions
are occurring. This more advanced sensing technology will allow for a quicker turn around on
information related to the advancement of pharmaceuticals.

43

Figure 2.11: UV-vis absorbance spectra for isatin in 0.1 M of PBS pH 7.4

2.3.4 Urine Simulant Study

To further advance the understanding of this compound using electrochemical methods, calibration
experiments were performed in a urine simulant with no N2 bubbling prior to measurement for
evaluation in a more complex sample matrix. Isatin was studied in the range of 1.8 - 12 µM due to
the lower concentrations being hidden in the background of the new matrix. Due to the increased
complexity of this study, the sensitivity of the MAC decreased by 10x with an increase in the LOD
to 0.82 µM. The calibration curve still displayed linearity of 0.995. The MEA sensitivity only
decreased by a factor of 3 and the LOD increased to 0.80 µM but this sensor also displayed linearity
with an R2 value of 0.994. Figure 2.12 shows the voltammograms for both the MEA and MAC.

The calibration curves for both sensors are shown in Figure 2.13. The calculated LODs
are higher than those that were determined in the bu ered solutions, but most LC methods use
pretreatment methods as well as extraction procedures before determining isatin concentrations
that include diluting the complex matrix, pH adjustments, and using ethyl acetate for extraction.
[8, 14, 65] Even with these pretreatment procedures, the LOD when using LC-MS was well above
that obtained for the two electrochemical sensors at 2 µM. No studies of isatin had previously been
performed in synthetic urine solution. This study allows further understanding of the capability to
use electrochemistry in the measurement of real samples.

44

Figure 2.12: Squarewave voltammograms for the 3-in-1 MAC sensor (A) and MEA sensor (B) in
Surine Negative Urine Control with no additional N2 bubbling

45

Figure 2.13: Calibration curves for the (A) MAC and (B) MEA for isatin in Surine Negative Urine
Control with no N2 bubbling

46

2.3.5 Repeatability Study

A 5-day repeatability study was performed to analyze the repeatability of the BDD sensors. The
measured data is shown in Figure 2.14. The ET pretreatment procedure was used on the ﬁrst day
of the study for both sensors. Square wave voltammetry was then performed in a 3 µM solution of
isatin in PBS bu er of pH 7.4. Over the next 3 days, each sensor was used to rest a fresh solution
for 7 total trials without repeating the pretreatment method. Both sensors had a decrease in signal;
63% for the MEA and 73% for the MAC. The ET process was the repeated and the sensor were
retested. The MEA regained 64% of the original signal it had and the MAC regained 76%. Over
the course of the 7 trails, the MEA remained stable and regained full signal strength in trial 6.
The MAC showed a similar trend for those without the ET method applies but the signal did not
decrease as signiﬁcantly with the ET procedure compared to without. This study suggests that
there are substantial changes with the signal response over the course of a few days and additional
research is required to better understand how the surface regeneration works.

Figure 2.14: Repeatability data for the 3-in-1 (A) MAC and (B) MEA sensors with the discussed
electrochemical pretreatment and without any addition of a pretreatment

2.3.6 Key Advancements Obtained from Study

The advancements obtained from this study include:

47

• Simpliﬁed technical equipment for the electrochemical detection of isatin by removing the

need for three separate electrodes with a 3-in-1 sensor

• Time savings in the analysis of isatin samples by implementing one sample pretreatment
process compared to the previous complicated and multistep processes used on other electrode
materials

• The ability to directly detect isatin in a complex sample, such as a urine simulant, without

the necessary extraction methods typically used for analysis with gas chromatography.

• Using the MEA sensor conﬁguration obtained a sensitivity that was two orders of magnitude
lower than the MAC resulting in a lower LOD in simulated solution and as overall more
repeatable with the electrochemical pretreatment method compared to the MAC.

2.4 Summary

Two sensors were fabricated to have a reference, auxiliary, and working electrode, all comprised
of BDD, and were fully characterized. The reference electrode had a potential variation with pH, but
stabilized quickly and remained stable over the course of 30 hours. Both sensors, one comprising
of a MEA and one comprising of a macro electrode, were characterized with a surface sensitive and
surface insensitive redox couples and displayed comparable rate constant values. Adsorption of
isatin and its reduction products were evaluated and no adsorption was found for the BDD surface.
Comparison of the MEA and MAC lead to the MEA having a LOD of 0.04 µM whereas the MAC
had a LOD 0.22 µM; both below the LOD of UV-vis spectroscopy measurements. Both sensors
were further evaluated in a urine simulant study as well as repeatability tests. This work concludes
that the fully BDD devices can be feasible for the measurement of isatin. Future applications of
these sensors include the detection following a ﬂow injection analysis system.

48

BIBLIOGRAPHY

49

BIBLIOGRAPHY

[1] Haihong Xu, Dan Wang, Wen Zhang, Wei Zhu, Katsunobu Yamamoto, and Litong Jin. Deter-
mination of isatin and monoamine neurotransmitters in rat brain with liquid chromatography
using palladium hexacyanoferrate modiﬁed electrode. Analytica Chimica Acta, 577:207–213,
2006.

[2] Lin Hou, Chuanxia Ju, Jinyu Zhang, Jinlian Song, Yinlin Ge, and Wang Yue. Antitumor
e ects of isatin on human neuroblastoma cell line (sh-sy5y) and the related mechanism.
European journal of pharmacology, 589(1):27–31, 2008.

[3] Manjusha Verma, Surendra Nath Pandeya, Krishna Nand Singh, and James P Stables. Anti-
convulsant activity of schi  bases of isatin derivatives. ACTA PHARMACEUTICA-ZAGREB-,
54(1):49–56, 2004.

[4] Vivette Glover, JM Halket, PJ Watkins, Angela Clow, BL Goodwin, and M Sandier. Isatin:
identity with the puriﬁed endogenous monoamine oxidase inhibitor tribulin. Journal of
neurochemistry, 51(2):656–659, 1988.

[5] P Watkins, Angela Clow, Vivette Glover, J Halket, A Przyborowska, and M Sandler. Isatin,
regional distribution in rat brain and tissues. Neurochemistry international, 17(2):321–323,
1990.

[6] Alexei Medvedev, Natalia Igosheva, Michele Crumeyrolle-Arias, and Vivette Glover. Isatin:

Role in stress and anxiety: Review. Stress, 8(3):175–183, 2005.

[7] Severino Carlos B Oliveira, Isabel PG Fernandes, Bárbara V Silva, Angelo C Pinto, and
Ana Maria Oliveira-Brett. Isatin nitro-derivatives redox behaviour. Journal of Electroanalyt-
ical Chemistry, 689:207–215, 2013.

[8] Shigeo Manabe, Qiang Gao, Juan Yuan, Toshie Takahashi, and Akira Ueki. Determina-
tion of isatin in urine and plasma by high-performance liquid chromatography. Journal of
Chromatography B: Biomedical Sciences and Applications, 691(1):197–202, 1997.

[9] Alexei E Medvedev, Angela Clow, Merton Sandler, and Vivette Glover. Isatin: a link between

natriuretic peptides and monoamines? Biochemical pharmacology, 52(3):385–391, 1996.

[10] Surendra Nath Pandeya, Sivakumar Smitha, Mayank Jyoti, and SESHAIAH KRISHNAN

Sridhar. Biological activities of isatin and its derivatives. Acta Pharm, 55(1):27–46, 2005.

[11] Victor C Diculescu, Sunita Kumbhat, and Ana Maria Oliveira-Brett. Electrochemical be-
haviour of isatin at a glassy carbon electrode. Analytica chimica acta, 575(2):190–197,
2006.

[12] Alpana K Gupta and Rajendra S Sindal. A comparative study of electrochemical reduction of
isatin and its synthesized schi  bases at hmde. Journal of chemical sciences, 121(3):347–351,
2009.

50

[13] Cigdem Karaaslan and Sibel Suzen. Electrochemical behavior of biologically important

indole derivatives. International Journal of Electrochemistry, 2011, 2011.

[14] Naoya Hamaue, Noriko Yamazaki, Masaru Minami, Toru Endo, Masahiko Hirahuji, Yoshio
Monma, and Hiroko Togashi. Determination of isatin, an endogenous monoamine oxidase
inhibitor, in urine and tissues of rats by hplc. General Pharmacology: The Vascular System,
30(3):387–391, 1998.

[15] Yanli Zhou and Jinfang Zhi. The application of boron-doped diamond electrodes in ampero-

metric biosensors. Talanta, 79(5):1189–1196, 2009.

[16] Nicolae Spataru, Bulusu V Sarada, Donald A Tryk, and Akira Fujishima. Anodic voltammetry
of xanthine, theophylline, theobromine and ca eine at conductive diamond electrodes and its
analytical application. Electroanalysis, 14(11):721, 2002.

[17] Zhao Guo-Hua, Li Ming-Fang, and Li Ming-Li. Di erential pulse voltammetric determination
of dopamine with the coexistence of ascorbic acid on boron-doped diamond surface. Central
European Journal of Chemistry, 5(4):1114–1123, 2007.

[18] Christophe Provent, Werner Haenni, Eduardo Santoli, and Philippe Rychen. Boron-doped
diamond electrodes and microelectrode-arrays for the measurement of sulfate and perox-
odisulfate. Electrochimica acta, 49(22):3737–3744, 2004.

[19] Karolina Peckova, Jana Musilova, and Ji í Barek. Boron-doped diamond ﬁlm electrodes—new
tool for voltammetric determination of organic substances. Critical Reviews in Analytical
Chemistry, 39(3):148–172, 2009.

[20] Romana Jaroöová, Paula M De Sousa Bezerra, Catherine Munson, and Greg M Swain.
Assessment of heterogeneous electron-transfer rate constants for soluble redox analytes at
tetrahedral amorphous carbon, boron-doped diamond, and glassy carbon electrodes. physica
status solidi (a), 213(8):2087–2098, 2016.

[21] Ka L Soh, Weng P Kang, Jim L Davidson, S Basu, Yong M Wong, David E Cli el, AB Bonds,
and GM Swain. Diamond-derived microelectrodes array for electrochemical analysis. Dia-
mond and Related Materials, 13(11):2009–2015, 2004.

[22] Cory A Rusinek, Adam Bange, Mercedes Warren, Wenjing Kang, Keaton Nahan, Ian Pa-
pautsky, and William R Heineman. Bare and polymer-coated indium tin oxide as working
electrodes for manganese cathodic stripping voltammetry. Analytical chemistry, 88(8):4221–
4228, 2016.

[23] Kentaro Tsunozaki, Yasuaki Einaga, Tata N Rao, and Akira Fujishima. Fabrication and elec-
trochemical characterization of boron-doped diamond microdisc array electrodes. Chemistry
letters, 31(5):502–503, 2002.

[24] Nathan S Lawrence, Markus Pagels, Andrew Meredith, Timothy GJ Jones, Clive E Hall,
CS Jim Pickles, Herman P Godfried, Craig E Banks, Richard G Compton, and Li Jiang. Elec-
troanalytical applications of boron-doped diamond microelectrode arrays. Talanta, 69(4):829–
834, 2006.

51

[25] JL Davidson, DR Wur, WP Kang, DL Kinser, and DV Kerns. Polycrystalline diamond pressure

microsensor. Diamond and related Materials, 5(1):86–92, 1996.

[26] P Bergonzo, A Bongrain, E Scorsone, A Bendali, L Rousseau, G Lissorgues, P Mailley, Y Li,
T Kau mann, and F Goy. 3d shaped mechanically ﬂexible diamond microelectrode arrays for
eye implant applications: The medinas project. Irbm, 32(2):91–94, 2011.

[27] M Bonnauron, S Saada, L Rousseau, G Lissorgues, C Mer, and P Bergonzo. High aspect ratio
diamond microelectrode array for neuronal activity measurements. Diamond and Related
Materials, 17(7):1399–1404, 2008.

[28] Gaëlle Piret, Clément Hébert, Jean-Paul Mazellier, Lionel Rousseau, Emmanuel Scorsone,
Myline Cottance, Gaelle Lissorgues, Marc O Heuschkel, Serge Picaud, and Philippe Bergonzo.
3d-nanostructured boron-doped diamond for microelectrode array neural interfacing. Bioma-
terials, 53:173–183, 2015.

[29] Agnieszka Kapa≥ka, György Fóti, and Christos Comninellis. The importance of electrode
material in environmental electrochemistry: formation and reactivity of free hydroxyl radicals
on boron-doped diamond electrodes. Electrochimica Acta, 54(7):2018–2023, 2009.

[30] Waldemar Smirnov, Nianjun Yang, René Ho mann, Jakob Hees, Harald Obloh, Wolfgang
Muller-Sebert, and Christoph E Nebel. Integrated all-diamond ultramicroelectrode arrays:
optimization of faradaic and capacitive currents. Analytical chemistry, 83(19):7438–7443,
2011.

[31] Wenjing Kang, Xing Pei, Adam Bange, Erin N Haynes, William R Heineman, and Ian Pa-
pautsky. Copper-based electrochemical sensor with palladium electrode for cathodic stripping
voltammetry of manganese. Analytical chemistry, 86(24):12070–12077, 2014.

[32] Wenjing Kang, Xing Pei, Cory A Rusinek, Adam Bange, Erin N Haynes, William R Heine-
man, and Ian Papautsky. Determination of lead with a copper-based electrochemical sensor.
Analytical Chemistry, 89(6):3345–3352, 2017.

[33] Wenjing Kang, Cory Rusinek, Adam Bange, Erin Haynes, William R Heineman, and Ian Pa-
pautsky. Determination of manganese by cathodic stripping voltammetry on a microfabricated
platinum thin–ﬁlm electrode. Electroanalysis, 29(3):686–695, 2017.

[34] Gustav Liljegren, Jean Pettersson, Karin E Markides, and Leif Nyholm. Electrochemical
solid-phase microextraction of anions and cations using polypyrrole coatings and an integrated
three-electrode device. Analyst, 127(5):591–597, 2002.

[35] Markus Pagels, Clive E Hall, Nathan S Lawrence, Andrew Meredith, Timothy GJ Jones,
Herman P Godfried, CS James Pickles, Jonathan Wilman, Craig E Banks, and Richard G
Compton. All-diamond microelectrode array device. Analytical chemistry, 77(11):3705–
3708, 2005.

[36] Xing Pei, Wenjing Kang, Wei Yue, Adam Bange, William R Heineman, and Ian Papautsky.
Disposable copper-based electrochemical sensor for anodic stripping voltammetry. Analytical
chemistry, 86(10):4893–4900, 2014.

52

[37] Miroslav Pohanka and Petr Skládal. Electrochemical biosensors–principles and applications.

J. Appl. Biomed, 6(2):57–64, 2008.

[38] Thomas Ponnuswamy, Jin-Jian Chen, Fei Xu, and Oliver Chyan. Monitoring metal ion
contamination onset in hydroﬂuoric acid using silicon–diamond and dual silicon sensing
electrode assembly. Analyst, 126(6):877–880, 2001.

[39] Xinhao Wang, Manas Ranjan Gartia, Jing Jiang, Te-Wei Chang, Junle Qian, Yong Liu, Xian-
grong Liu, and Gang Logan Liu. Audio jack based miniaturized mobile phone electrochemical
sensing platform. Sensors and Actuators B: Chemical, 209:677–685, 2015.

[40] Youngsam Yoon, Gil S Lee, Koangki Yoo, and Jeong-Bong Lee. Fabrication of a micronee-
dle/cnt hierarchical micro/nano surface electrochemical sensor and its in-vitro glucose sensing
characterization. Sensors, 13(12):16672–16681, 2013.

[41] R Ramamurti, M Becker, T Schuelke, T Grotjohn, D Reinhard, G Swain, and J Asmussen.
Boron doped diamond deposited by microwave plasma-assisted cvd at low and high pressures.
Diamond and Related Materials, 17(4):481–485, 2008.

[42] E Bustarret, E Gheeraert, and K Watanabe. Optical and electronic properties of heavily

boron-doped homo-epitaxial diamond. physica status solidi (a), 199(1):9–18, 2003.

[43] HJ Xiang, Zhenyu Li, Jinlong Yang, JG Hou, and Qingshi Zhu. Electron-phonon coupling in

a boron-doped diamond superconductor. Physical Review B, 70(21):212504, 2004.

[44] T Teraji, H Wada, M Yamamoto, K Arima, and T Ito. Highly e cient doping of boron into
high-quality homoepitaxial diamond ﬁlms. Diamond and Related Materials, 15(4):602–606,
2006.

[45] Dario Narducci, C Richard Guarnieri, and Jerome J Cuomo. Defect clustering and boron elec-
trical deactivation in p-doped polycrystalline diamond ﬁlms. Journal of The Electrochemical
Society, 138(8):2446–2451, 1991.

[46] P Gonon, E Gheeraert, A Deneuville, F Fontaine, L Abello, and G Lucazeau. Characterization
of heavily b-doped polycrystalline diamond ﬁlms using raman spectroscopy and electron spin
resonance. Journal of applied physics, 78(12):7059–7062, 1995.

[47] Joel W Ager III, W Walukiewicz, Matthew McCluskey, Mary Anne Plano, and Maurice I
Landstrass. Fano interference of the raman phonon in heavily boron-doped diamond ﬁlms
grown by chemical vapor deposition. Applied Physics Letters, 66(5):616–618, 1995.

[48] E Gheeraert, P Gonon, A Deneuville, L Abello, and G Lucazeau. E ect of boron incorporation
on the “quality” of mpcvd diamond ﬁlms. Diamond and Related Materials, 2(5-7):742–745,
1993.

[49] NG Ferreira, AF Azevedo, AF Beloto, M Amaral, FA Almeida, FJ Oliveira, and RF Silva.
Nanodiamond ﬁlms growth on porous silicon substrates for electrochemical applications.
Diamond and related materials, 14(3):441–445, 2005.

53

[50] Bard Allan J. and Faulkner Larry R. Electrochemical Methods Fundamentals and Applica-

tions. John Wiley and sons, Inc., United States of America, second edition, 2011.

[51] Strobel Howard A. and Heineman William R. Chemical Instrumentation: A Systematic

Approach. John Wiley and Sons, Inc., United States of America, third edition, 1989.

[52] Julie V Macpherson. A practical guide to using boron doped diamond in electrochemical

research. Physical Chemistry Chemical Physics, 17(5):2935–2949, 2015.

[53] Michael C Granger, Malgorzata Witek, Jishou Xu, Jian Wang, Mateusz Hupert, Amy Hanks,
Miles D Koppang, James E Butler, Guy Lucazeau, and Michel Mermoux. Standard electro-
chemical behavior of high-quality, boron-doped polycrystalline diamond thin-ﬁlm electrodes.
Analytical Chemistry, 72(16):3793–3804, 2000.

[54] Anne E Fischer, Yoshiyuki Show, and Greg M Swain. Electrochemical performance of
diamond thin-ﬁlm electrodes from di erent commercial sources. Analytical chemistry,
76(9):2553–2560, 2004.

[55] Yoshiyuki Show, Ma≥gorzata A Witek, Prerna Sonthalia, and Greg M Swain. Characteriza-
tion and electrochemical responsiveness of boron-doped nanocrystalline diamond thin-ﬁlm
electrodes. Chemistry of materials, 15(4):879–888, 2003.

[56] Matt Hupert, Alexander Muck, Jian Wang, Jason Stotter, Zuzana Cvackova, Shannon Hay-
mond, Yoshiyuki Show, and Greg M Swain. Conductive diamond thin-ﬁlms in electrochem-
istry. Diamond and Related Materials, 12(10):1940–1949, 2003.

[57] Michael C Granger and Greg M Swain. The inﬂuence of surface interactions on the re-
versibility of ferri/ferrocyanide at boron-doped diamond thin-ﬁlm electrodes. Journal of the
Electrochemical Society, 146(12):4551–4558, 1999.

[58] Richard S Nicholson. Theory and application of cyclic voltammetry for measurement of

electrode reaction kinetics. Analytical Chemistry, 37(11):1351–1355, 1965.

[59] Joseph Wang. Analytical Electrochemistry. John Wiley and Sons, Inc, Online, 3 rd edition,

2006.

[60] RJ Klingler and JK Kochi. Electron-transfer kinetics from cyclic voltammetry. quantitative
description of electrochemical reversibility. The Journal of Physical Chemistry, 85(12):1731–
1741, 1981.

[61] Matthias Werner, Otto Dorsch, H-U Baerwind, Emst Obermeier, Colin Johnston, Paul R
Chalker, and Simon Romani. The e ect of metallization on the ohmic contact resistivity
to heavily b-doped polycrystalline diamond ﬁlms. IEEE Transactions on Electron Devices,
42(7):1344–1351, 1995.

[62] Kissinger Peter T. and Heineman William R. Laboratory Techniques in Electroanalytical

Chemistry. Marcel Dekker, Inc., New York, second, revised and expanded edition, 1996.

54

[63] Michael J Nuwer, John J O’Dea, and Janet Osteryoung. Analytical and kinetic investigations of
totally irreversible electron transfer reactions by square-wave voltammetry. Analytica chimica
acta, 251(1-2):13–25, 1991.

[64] John J O’Dea, Janet Osteryoung, and Robert A Osteryoung. Theory of square wave voltam-

metry for kinetic systems. Analytical Chemistry, 53(4):695–701, 1981.

[65] JM Halket, PJ Watkins, A Przyborowska, BL Goodwin, Angela Clow, Vivette Glover, and
M Sandler. Isatin (indole-2, 3-dione) in urine and tissues: Detection and determination by gas
chromatography—mass spectrometry. Journal of Chromatography B: Biomedical Sciences
and Applications, 562(1-2):279–287, 1991.

55

CHAPTER 3

INDIUM TIN OXIDE FILM CHARACTERISTICS FOR CATHODIC STRIPPING

VOLTAMMETRY

56

3.1

Introduction

This work has been published in ACS Applied Materials and Interfaces1. Indium tin oxide,
or ITO, is a common conductive and optically transparent ﬁlm that is used in applications such
as transparent electrodes for solar cells, surface heaters for windows, liquid-crystal displays, and
spectroelectrochemical applications. [1–10] Although sputtering is the most common deposition
method for ITO ﬁlms, it can also be performed using chemical vapor deposition, ion beam evapo-
ration, and spray pyrolysis. [1, 2, 4–7, 11–17] ITO ﬁlms can be tailored to having di erent optical
and electrical properties and can be deposited as amorphous or crystalline ﬁlms.

The wide range of characteristics that each ﬁlm can possess comes from the changes in the
ordered structure of the ﬁlm due to the generation of oxygen vacancies. [7] The tunable process
parameters that allow these changes include gas pressures, substrate temperatures, deposition time,
ﬁlm thickness, and post processing annealing steps in di erent gases and at di erent temperatures.
[2, 4, 5, 7, 11, 18, 19] Between the temperatures of 100 and 150  C, there is an amorphous to
crystalline transition for ITO. [19] This shift causes ﬁlms to di er in their presence of oxygen and
conducting carrier concentration. [1, 7, 19, 20] When the substrate temperature is above 100  C
the ﬁlm’s resistivity drops leading to an increase in carrier concentration and mobility. [6, 19] A
ﬁlm thickness of greater than 200 nm, deposited at room temperature, are typically crystalline, but
those below 100 nm are typically amorphous. [4] All of these process parameters can be tuned to
create unique ﬁlms for speciﬁc applications. [2, 4, 7, 11, 19, 21]

Post deposition process parameters can also have an e ect on the ITO ﬁlm’s characteristics.
dc-sputtered post deposition annealed ﬁlms show an increase in grain size, decrease in electrical
resistivity, and a decrease in intrinsic strain. [4, 5, 11, 20] The temperature needed for crystallization
during annealing is higher for ﬁlms that were deposited on non-heated substates compared to those
[4, 7] The optical properties of the ﬁlm are also a ected by
deposited on heated substrates.
annealing.
[4] Various

It can lead to an increased band gap and a steeper cuto  absorption.

1ACS Applied Materials and Interfaces 2019, DOI:10.1021/acsami.8b22157

57

studies have been carried out to study the changes in the ITO ﬁlm based on annealing atmosphere,
temperature, and if vacuum is used. [1, 7, 11, 20, 22]

The most common use of ITO in the ﬁeld of electrochemistry is as an optically transparent
ﬁlm for spectroelectrochemical measurements.
[9, 23, 24] It has been used to detect analytes
including Azure A, total alkaline phosphate in human blood, potassium ferricyanicde, and iron
using complexation with 2,2’-bipyridine. [2, 23, 25, 26] This application is primarily due to this
ITO ﬁlm possessing a large positive potential window beyond +1.5 V while remaining transparent
in the range of 390-700 nm. [23, 27, 28] Since the ﬁlm is already in its oxidized form, it does
not react in this positive potential range. [25] Several metals including cadmium (Cd), lead (Pb),
copper (Cu), cerium (Ce), mercury (Hg), and manganese (Mn), has also been detected with ITO,
by using the method of stripping voltammetry.
[27–30] Cd and Cu had their signal monitored
as they were plated onto the electrode surface using attenuated total internal reﬂectance stripping
voltammetry. [29] Ce additionally has been detected using cathodic stripping voltammetry (CSV)
and ITO. [27] Recently, bare ITO was used to detect Pb, Cu, and Hg simultaneously using this
method. [30]

Mn has been investigated on ITO also using CSV. [28] This was performed using an ITO coated
with polystyrene-block-poly(ethylene-ran-butylene)-block-polystyrene-sulfonate (SSEBS) as well
as compared to a bare ITO electrode. [28] This study was followed up with a paper reporting the
detection of Mn in digested bovine blood samples that obtained similar detection capabilities as
inductively coupled plasma mass spectrometry (ICP-MS) and graphite furnace atomic absorption
spectroscopy (GF-AAS). [31] The detection of Mn is important due to it having toxic e ects and
a maximum contaminant level (MCL) in drinking water of 50 parts per billion (ppb) set by the
United States Environmental Protection Agency (EPA). It is also signiﬁcant to test human blood,
urine, and serum levels when monitoring its e ects. [32, 33]

Palladium (Pd) on Cu, boron-doped diamond (BDD), and various other carbon materials have
also been used to detect Mn with CSV. [34–39] Compared to these other materials, the positive
potential window of ITO makes it more beneﬁcial for CSV. Cu-based Pd electrodes obtained a

58

LOD at 19 ppb which is below the MCL in drinking water but higher than the typically biological
concentrations.
[32, 33, 35] The BDD sensor was evaluated in sea water and required the use
of ultrasound, but only obtained a LOD of 5.5 ppb which is still not sensitive enough for serum
analysis. [36] Additionally BDD had a higher resistivity compared to ITO and can be expensive
to fabricate into sensor designs. [40] The bare carbon paste electrode previously studied obtained
the same LOD as the BDD sensor. [39] The most promising electrode material for Mn detecting
using CSV was edge plane pyrolytic graphite that obtained a LOD of 0.80 ppb. [37] However, the
previously studied ITO ﬁlms reached an even lower LOD compared to this material which may
be due to the metal oxide surface of ITO being beneﬁcial in the deposition of manganese dioxide
(MnO2) on the surface.

Sensors being used for metal detection ultimately do not require the use of optical transparency
when utilizing the method of CSV. ITO has been deposited on opaque substrates before, but never
has then been applied to CSV metal detection. [27, 41] Fabrication of such a sensor on silicon (Si)
is much safer and easier than the process used for sensor fabrication on transparent substrates due to
avoiding the use of hydroﬂuoric acid (HF). [42] Si can utilize dry etching processes for fabrication
allowing for faster and more precise etching without putting fabricators lives in jeopardy. ITO is
already mass produced for other applications on non-transparent substrates that would also allow
for an economical mass production of sensors. In addition to the unique electrochemical properties
that ITO possess, it is already widely used in mass production for solar cells which means it is
available at much lower costs, and is more reproducible than other electrode materials used in CSV.
This study reports on the optimization of ITO working electrodes on Si substrates for the
application of metal detection using CSV. At this time, there is no report in literature using
opaque substrates for this technique. Several processing parameters such as heated and non-heated
deposition and the e ects of post deposition annealing were investigated. Each corresponding ﬁlm
was thoroughly characterized and evaluated before using each to detect Mn by CSV.

59

3.2 Experimental Section

3.2.1 Chemicals and Materials

A stock solution of potassium ferrocyanide (K4Fe(CN)6) was prepared using powder purchased
from J.T Baker and deionized (DI) water. Manganese (Mn) was purchased from Acros Organics.
Deionized (DI) water and acetic acid (Fisher Scientiﬁc) were used to prepare a 0.2 M acetate
bu er with 0.2M sodium acetate purchased from Sigma-Aldrich. Ferrocene (Aldrich Chemical
Company Inc.) was prepared in acetonitrile (Sigma-Aldrich) using a 1mM solution with 0.1 M
tetrabutylammonium hexaﬂuorophosphate.

3.2.2

Instrumentation

An AFM5100 from Hitachi was used in tapping mode to complete the atomic force microscopy
(AFM) measurements on an area of 4 µm2 with a scan frequency of 0.84 Hz. The roughness of
each sample was determined using the open source software of Gwyddion using three evaluation
lines that were placed across the top, middle, and bottom of the scanned area. A cuto  length of
one-ﬁfth of the scan length was used in these calculations.

X-Ray di raction (XRD) measurements were performed with a Rigaku SmartLab 3 kW Cu
source generating a parallel beam with electron current of 44 mA and an acceleration voltage of
40 kV. Out of frame di raction peaks were collected at 4  /min using symmetric (2✓ - !) scanning
stepping at 0.02  over a range of 30  - 90 . Di racted X-rays were collected through collimating
slits at 0.5  pitch and aperture of 20 mm using a 0D SC-70 photomultiplier tube. An area of 10 x
12 mm was set to limit the X-ray beam spot on the sample.

A Shimadzu UV-vis-NIR spectrophotometer with a wavelength of 190-2500 nm was used for
all optical measurements. A 414513 semiconductor parameter analyzed from Hewlett-Packard was
used for the four-point probe measurements with a correction factor of 4.3. [43]

An Epsilon Eclipse from Bioanalytical Systems, Inc.

(BASi) was used in a three electrode
setup for all electrochemical measurements. Each ﬁlm was used as a working electrode with a 0.5

60

mm circle diameter exposed to the electrolyte solution. Inside the cell was a silver/silver chloride
(Ag/AgCl) reference electrode and either a platinum (Pt) or graphite auxiliary electrode. A silver
quasi-reference was used for the chronoamperometry (CA) experiments in ferrocene. Each CV
experiment utilized 0.1 mM Fe(CN)6 4- in 1.0 M KCl by scanning the potential from -0.1 to 0.6
V. Peak height measurements were performed with an extrapolated baseline method described in
literature. [44–46]

3.2.3 Electrode Fabrication

Pulsed dc magnetron sputtering was utilized for all fabrication runs. Each run was performed for
20 minutes using a Kurt J. Lesker PVD 775 system. One set was performed at room temperature
and one utilized a set substrate temperature of 190  C. An ITO target (In90/Sn10) with a 1.6%
oxygen (O2) to argon (Ar) ratio was used. The power density was 20 W/in. Two groups of
microscope slides and nonconductive Si substrates were formed: those annealed and not annealed.
The annealing was done at 400  C for 1 hour in ambient air. Four ﬁnal samples were evaluated;
heated and annealed (HA), heated and not annealed (HNA), not heated and annealed (NHA), and
not heated and not annealed (NHNA). All ﬁlm’s thicknesses were in the range of 230 and 250 nm.
3.3 Results and Discussion

3.3.1 Physical and Optical Characterization: Atomic Force Microscopy

Atomic force microscopy (AFM) was used to evaluate each electrode ﬁlm’s surface morphology.
Figure 3.1 shows the two dimensional topography views for HA, HNA, NHA, and NHNA on
nonconductive silicon (Si) substrates. From this Figure, visual similarities can be seen between the
two heated samples and the two non-heated samples.

NHA and NHNA, shown in (C) and (D) respectively, are comprised of smaller grain sizes that
appear to be more inhomogeneous than the two heated samples. Also between these two ﬁlms, the
annealing is noted to cause a signiﬁcant appearance of taller gains shown in the lighter color. HA
and HNA, shown in (A) and (B) respectively, have larger grain sizes compared to the non-heated

61

Figure 3.1: Tapping mode atomic force microscopy (AFM) images for the four analyzed ITO
ﬁlms (A) HA, (B) HNA, (C) NHA, and (D) NHNA.

samples. Annealing the heated sample (A) also brings about higher grain structures compared to
the not annealed heated sample.

The AFM data also provided the root-mean-square (rms) roughness values for each ﬁlm. NHNA
had the lowest rms value with HA having the highest value. All the ﬁlm’s rms values are provided
in Table 3.1. The roughness values are a ected more by the heating than the annealing, but all the
ﬁlms are smooth overall. There is a beneﬁt to having a greater roughness to electrode materials. A
higher roughness leads to a greater microscopic surface area which can lead to a greater sensitivity
and lower limit of detection (LOD) in CSV measurements.[44–46] From this physical and optical
characterization, HA and HNA are potentially more advantageous as a sensing material than the
other two ﬁlms.

62

Table 3.1: Comparison of each of the ﬁgures of merit for each of the four ITO ﬁlms that were
investigated.

3.3.2 Physical and Optical Characterization: XRD

All four ﬁlms were conﬁrmed as polycrystalline from the presence of multiple peaks in the di rac-
tion pattern performed with symmetric scans. The peaks were identiﬁed as Ia3 ITO using the
crystallography open data base. [47–50]

In Figure 3.2 (A) the (222) peak decreases as a function of heat treatment, regardless if performed
by deposition or annealing. The (400) peak, Figure 3.2 (B), shows an opposite trend by increasing
with heat treatment. Literature states that the (400) peak decreases when there is increasing oxygen
in the ITO ﬁlm suggesting that the heated deposition promotes the (400) orientation of the ﬁlm.
[50] The center of the (222) orientation shifts during both heating during deposition, and annealing
a non heated ﬁlm displaying a change in the strain of this phase. Using the NHNA ﬁlm as a zero
point, the NHA peak center is +0.16 °and the HNA ﬁlm’s is -0.2 °. HA’s center of the (222)
orientation peak is in roughly the same position as NHNA’s.

The ratio of the (222) and (400) peaks, I(222)/I(400), for NHNA is 23 times greater than HNA’s,
indicating a transition to a preferred orientation (Figure 3.3 (left)). The full width half-maximum
(fwhm) of the (222) and (400) peaks is also a function of heat treatment. In general, the fwhm
decreases from both heating and annealing as shown in Figure 3.3 (right). This trend shows that
either during heating or after annealing, the crystalline quality improves. Without the application
of heat during processing, the mobility of the tin and indium-oxygen clusters is restricted and limits

63

Figure 3.2: XRD scans for the (A) (222) peak and (B) (400) peak for each ITO ﬁlm investigated.

the likelihood of miss-orientations leading to bond distortions.

Figure 3.3: (left) ITO crystal quality (I(222)/I(400)) ratio obtained from XRD analysis, and (right)
the full width half max (FWHM) data for each ﬁlm from XRD analysis pertaining to either the
(222) or (400) peak

The FWHM for NHNA is caused by this process. The application of heat above 150 °C
during deposition enables the mobility of the tin and indium oxide clusters. [51] An exception
is the jump in FWHM from the (400) in NHNA and the (400) in NHA and the reverse FWHM
positions of (222) and (400) for HNA. This may be the result of having frozen clusters when using
room temperature deposition for NHA. Annealing has shown the ability to unfreeze these sites but
the di usion through the ﬁlm is limited and restricts their chance of occupying favorable lattice
positions. However, with heated deposition HNA allows the clusters a higher mobility and increases
the probability of stable bond formations and a lower FWHM.

64

Figure 3.4: ITO crystal size derived using the Scherrer equation and the (222) and (400) peak of
each ﬁlm

HA supports all the conclusions drawn when combining heating and annealing by having the
smallest FWHM, least amount of stress or strain, and the largest (400) crystallite size as seen in
Figure 3.4. This leads to the conclusion that HA could be the best ﬁlm for Mn detection due to
its larger grain size and increased conductivity. [19] This could also suggest that HNA may be
the optimum ﬁlm due to its mirrored properties for the FWHM and crystal size compared to HA.
Additionally, The I(222)/I(400) of HNA is lower than that of HA.

3.3.3 Physical and Optical Characterization: UV-Vis Spectroscopy

To gain insight into the charge carriers of the ﬁlms, microscope slides were included in each run
and analyzed using UV-Vis Spectroscopy. The Transmittance (T%) spectra of a blank microscope
slide, for reference, and each of the four ﬁlms is shown in Figure 3.5 The T% spectrums of HNA
and NHNA are similar to one another as well as HA and NHA, suggesting that the annealing step
changes the optical properties of the ﬁlms. Literature states that annealing of ITO ﬁlms shifts the
absorption edge to a lower wavelength, and this is the same trend seen with the ﬁlms in this study.
[4] HA and NHA are shifted to lower wavelengths and HNA is also seen shifted in this direction

65

which may be due to the heating of the substrate. Compared to NHNA, it is presumed that these
three ﬁlms will have a lower defect density and higher crystallinity due to this shift in their T%
spectra. [4]

Figure 3.5: Transmittance spectra measured with UV/Vis for each of the four ﬁlms deposited on a
transparent microscope slide

In the wavelength range of 1500-2500 nm, NHNA has the highest T% value with HA displaying
the lowest. Due to the fact that literature has shown that a higher carrier concentration displays a
lower T% in this wavelength, it is assumed that HA has the highest carrier concentration. [4, 52]
The direct optical band gap of each ﬁlm was also determined from the T% data using the reﬂectance
spectrum as well. The direct optical band gap provides information about the crystallinity of each
ﬁlm due to the relation between an increase in band gap with a decrease in localized state density.
[53, 54] In order to calculate the band gap, the energy gap (Eg) was ﬁrst determined using Equation
2.1. [55, 56]

↵ ⇤ h ⇤ v = (h ⇤ v   Eg)r

(3.1)

Where hv is the photon energy, ↵ is the absorption coe cient, and r is the transition coe cient
with a value of either 0.5 or 2 for a direct band gap or indirect band gap, respectively. ↵ is
determined using Equation 2.2.

66

T = (1   R)e(-↵d)

(3.2)

Where d is the ﬁlm’s thickness. By extrapolating the linear region of the Tauc Plot, Figure 3.6,
the direct optical band gap was determined for each ﬁlm. All of the ﬁtting parameters are displayed
in Table 3.2.

Figure 3.6: Tauc graph used to determine the direct optical band gap for the four ITO ﬁlms studied

Table 3.2: Fitting parameters for the band gap determination from the Tauc plot of each ﬁlm.

The lowest and highest optical band gap belonged to NHNA and HA, respectively. The band
gap was increased by both heating and annealing. According to the density of states model by
Mott and Davis, the process of annealing or heating removes the defects that were present in the
amorphous ﬁlm and forms saturated bonds that increase the band gap. [56] This trend agrees with
the FWHM trend from the XRD analysis.

67

Since HA has the highest band gap followed by NHA and HNA, these three ﬁlms are highly
likely to be applicable for a sensing application. All the optical band gaps determined for these
ﬁlms compare well with literature values for other reported ITO ﬁlms. [5, 7, 21]

3.3.4 Electrical Properties

In terms of having a good electrode material for electrochemical sensing applications, low resistance
is necessary. For each ﬁlm, the resistance and resistivity values are summarized in Table 3.1. NHNA
had the highest resistivity and sheet resistance and HA had the lowest. NHA and HNA had almost
the same sheet resistance and resistivity as one another. According to literature the resistivity of
an ITO ﬁlm decreases when it is heated or annealed to temperatures above 100  C. [4, 7, 11, 19]
Compared to other electrodes used for Mn detection, the ITO ﬁlms have a higher resistivity than
carbon, but lower than boron-doped diamond. [40, 38, 57] The results for the electrical properties
suggest that HA should be expected to perform as the best electrochemical sensor; agreeing with
the optical property analysis performed earlier.

3.3.5 Electrochemical Characterization

Cyclic voltammograms (CVs) and Ohm’s resistance plots were generated to investigate each ﬁlm’s
electrochemical behavior. The CVs allowed for the double layer capacitance of each ﬁlm to be
determined in 1 M potassium chloride (KCl) with a scan rate range of 0.1 to 0.4 V/s. Each current
value at 0.0V was recorded and was graphed against its corresponding scan rate, Figure 3.7. The
slope of this graph provided the capacitance of each ﬁlm that could then be divided by the exposed
area to obtain the typical units of F/cm2.

The inner sphere electron transfer redox couple of potassium ferrocyanide was used to evaluate
each ITO ﬁlm. This was chosen since it has a surface sensitive nature and can be used to compare
to other electrode materials although it is believed that the reaction of this redox on the surface
varies with its monolayer adsorption and does not vary with surface oxide coverage. [58] Figure
3.8(A) shows the CV for each ﬁlm obtained from a scan rate of 0.1 V/s.

68

Figure 3.7: Graph showing the current vs scan rate study of each of the ITO ﬁlms used to
determine their corresponding capacitance values

The peak separations determined at 0.1 V/s were 90 mV, 83 mV, 89 mV, and 95 mV for HA,
HNA, NHA, and NHNA, respectively compared to previously published at 160 mV. [59] Potassium
ferrocyanide is a common redox couple used to characterize boron doped diamond (BDD) and
glassy carbon. The peak separations on these materials are 70-85 mV for BDD and 61-65 mV
for glassy carbon.
[59–61] Depending on pretreatments, carbon electrode materials can have
drastically di erent peak separations including values of closer to 145 mV. [59, 61] The ITO ﬁlms
investigated here are similar to the peak separations for BDD.

Figure 3.8: (A) CV curves of 0.1 mM ferrocyanide in 1.0M KCl with a scan rate of 0.1 V/s. (B)
Potential windoe study performed in acetate bu er of pH 5.0. (C) CSV od 25 ppb of Mn in acetate
bu er of pH 5.0, deposition time of 3 minutes and a deposition potential of +1.2 V.

69

The rate constant of ferrocyanide on each ITO ﬁlm was determined using the Nicholson Shain
method for electrodes that display low peak separation with exact psi values that were interpolated
from the supplied table in the reference. [62] In terms of peak current, Figure 3.8(A), HA had
the highest followed by HNA and NHA with very similar responses. NHNA displayed the lowest
current value out of the four ﬁlms. This trend of peak current matches the trend for the conductivity
of each ﬁlm. The rate constants for each ﬁlm are shown in Table 3.1. HA and HNA had the fastest
rate constant and NHNA has the slowest. This result suggests, again, that HA or HNA would make
a great electrochemical sensor since a high current response and a fast rate constant are desired.

Each ITO ﬁlm’s electrochemical potential window was also studied; Figure 3.8 (B). The
potential began at 0.0V and was scanned anodically to +2.5 V before being reversed and scanned to
-1.75 V and then back to 0.0V. All four ﬁlms display an excellent anodic potential window, but have
several peaks in the cathodic scan. These peaks are believed to be from the reduction of oxygen
and ITO to metallic indium (In) and tin (Sn) as studied by Liu et al. [63] The peak at -0.3 V is the
largest on HA and is likely due to the reduction of oxygen in each ﬁlm. This would also support the
claim that the reduction of the (400) XRD peak is due to increasing oxygen in the ﬁlms. According
to all four scans, it appears that both heating and annealing ITO can cause an increase of oxygen
incorporation into the ﬁlms. The other peaks at around -1.2 V are also largest in HA and are related
to the reduction of the ITO ﬁlms. [63] This ﬁlm with the In- and Sn- terminated phases and larger
crystallite size has an important roll for the electrode performance.

Plotting the change in peak potential (Ep) versus the change in peak current (ip) for each ﬁlm
allowed for the generation of Ohm’s resistance plots. The slope of these plots is the resistance in
the electrochemical cell in Ohm. All four ﬁlm’s resistance plots are shown in Figure 3.9. These
plots were generated since the resistance values from the four-point-probe only provide information
of the resistance through the ﬁlm between the probes in air. Whereas these plots measure the
resistance when exposed to an electrolyte solution as the ﬁlms would be in a practical application.
Each of the four ﬁlm’s has a constant resistance when the temperature is constant and are ohmic
resistors. HA had the lowest resistance calculated and NHNA had the highest. NHA and HNA were

70

Figure 3.9: Plots for the Ohm’s resistance of each of the four ITO ﬁlms investigated

very similar to each other once again. All the values obtained from these plots is also displayed
in Table 3.1. The trends seen from the Ohm’s plots are expected based on the four-point-probe
measurements.

Chronoamperometry measurements were performed to determine the microscopic surface area
[44] All calculations
of each ﬁlm by using an applied potential step and the Cottrell equation.
were performed with a constant di usion coe cient of 2.24 x 10-5 cm/s. all four ﬁlms displayed
microscopic surface areas that were statistically similar at roughly 0.20 cm2. The geometric surface
area of 0.196 cm2 is almost identical to this value; probably due to the smoothness of each ﬁlm
provided in Table 3.1. From this, it was determined that the surface roughness may not be a critical
parameter in discriminating between the ﬁlms.

3.3.6 CSV of Manganese

Manganese was used to further test the applicability of each ITO ﬁlm by doing CSV with 25 ppb
of Mn with a deposition potential of +1.2V for 3 min, based on previous Mn experience when
using ITO. [28] The stripping voltammograms of each ﬁlm were performed a total of three times
and an average for each is plotted in Figure 3.8 (C). NHNA is the only ﬁlm that did not have a
response peak at this Mn concentration. HNA and HA displayed a comparable peak to one another

71

and NHA was visually comparable to them as well, but showed a smaller magnitude of current.
Table 3.1 displays the peak current responses for each ﬁlm. These results, and previous, support
the hypothesis that the heated deposition run makes the most appropriate ITO ﬁlm for Mn analysis,
but further investigation into HA and HNA still needs to be performed to decide between them.

3.3.7 Calibration and Figures of Merit

Figure 3.10: Mn stripping voltammograms for the full concentration range on HA (A1) and the
smaller range (A2), for the full range on HNA (B1) and the smaller range (B2), and for the full
range on NHA (C1) and smaller range (C2). Performed in pH 5.0 acetate bu er with a deposition
potential of 1.2 V and a deposition time of 3 min

Mn concentrations of 1, 5, 10, 25, and 50 ppb were used to generate calibration curves performed
in triplicate. Figure 3.10 shows the stripping voltammograms for the HA, NHA, and HNA ﬁlms.
Calibration curves pertaining to the full concentration range investigated are sown in Figure 3.11.
The uncertainty in the slope, y-intercept, and the uncertainty in the y and correlation coe cient
(R2) values were determined using a Linest function in Excel. The equation for propagation of
uncertainty in the best ﬁt line was used to determine the uncertainty in the LOD for each calibration
curve.
[64] NHNA was excluded from both Figure 3.10 and Figure 3.11 due to its inability to
perform for this application resulting from an R2 value of 0.663.

72

Figure 3.11: Calibration curves for Mn on HA (A), HNA (B), and NHA (C)

The overall average response of NHA was [I(A) = (-3.3 x 10-8 +/- 3.5 x 10-9)c (A/ppb) + (1.2
x 10-7 +/- 8.9 x 10-8) (A) R2 = 0.967] with a LOD of 0.8 ppb. The downsides to this ﬁlm were
that the background shifted slightly throughout the measurements and the lowest concentration of 1
ppb was hard to distinguish from the background, as seen in Figure 3.10(C2). HNA had a response
of [I(A) = (-3.0 x 10-8 +/- 3.4 x 10-9)c (A/ppb) +(1.4 x 10-7 +/- 8.6 x 10-8) (A) R2 = 0.964] and a
LOD of 0.3 ppb. Compared to the peaks obtained in NHA, the peaks on HNA are more symmetric
and the background level shifts less. The lowest concentration is still hard to distinguish from
the background in Figure 3.10(B2). HNA also exhibited an improved reproducibility and lower

73

standard deviation at 1 ppb compared to NHA. HA had an overall response with the lowest LOD,
highest sensitivity, and greatest correlation coe cient; [I(A) = (-3.4 x 10-8 +/- 1.2 x 10-9)c (A/ppb)
+ (8.2 x 10-8 +/- 3.1 x 10-8) (A) R2 =0.996]. The LOD on HA was 0.1 ppb which is 3x lower
than that of HNA. The R2 value for this ﬁlm was the only analytically acceptable value. Previously
recorded LODs for Mn on bare ITO are reported at 0.27 ppb, making the ﬁlm studied here superior.
[28] The lowest concentration of 1 ppb on this ﬁlm displays a clear and distinct peak compared
to HNA and NHA. The background also remains the most stable and appears the ﬂattest from the
other ﬁlms. HA was also the most reproducible ﬁlm overall.

The LODs obtained on HA, HNA, and NHA are lower than the MCL of 50 ppb for Mn in
drinking water, blood (4-15 ppb), urine (1-8 ppb), and serum (0.4-0.85 ppb). [32, 33] It is also
worth noting that the ﬁlms investigated here did not undergo a deposition time optimization which
can lead to even lower LOD values if necessary.

3.3.8 Possible Metals Interference

Since manganese is not the only metal that is present in environmental and biological samples, it
is important to consider if other metals may have an a ect on the sensors Mn detection response.
These metals can include cadmium, chromium, copper, iron, lead, and nickel. Table 3.3 shows each
of these metals reaction potentials. By using cathodic stripping voltammetry and the application
of a positive deposition potential, the interference from most metals plating onto the working
electrode’s surface is avoided due to them having a negative reduction potential. In other literature
it has been shows that cadmium chromium and zinc do not interfere with the manganese detection
when using a carbon ﬁlm electrode. [38] They also found that at elevated concentrations nickel,
copper and lead had an inﬂuence on the manganese signal. Since lead does have a positive redox
potential to form an oxide, this could also plate onto the electrode’s surface during the deposition
of manganese. Additionally iron has been shown to inﬂuence the ability to detect manganese on
both the carbon ﬁlm electrode and an ITO electrode. [38, 28] This interference is possibly due to
the fact that iron chemically strips the manganese oxide back o  the electrode before the potential

74

sweep can be performed. [28] However, a previous study showed that typically Fe2+ concentrations
in drinking water do not cause a concern for the ITO electrodes. Additionally anion interferences
can be reduced with a negatively charged polymer placed onto the ITO’s surface. [28] A literature
report also shows that using sodium citrate can help to eliminate this potential interference in waters
that have high iron concentrations. [65]

Table 3.3: Redox potentials for other metals that are potentially in environmental and biological
samples

3.3.9 Key Advancements from Study

Key advancements from this study include:

• Development of an economically viable and consistent ﬁlm material for metal sensing appli-

cations that is already mass produced for other applications

• Lower detection limits for manganese that allows for improved accuracy in human serum

measurements

• Understanding that heating and annealing of ITO ﬁlm’s create the best ﬁlm structure to be

used as an electrode for electrochemical sensing applications

75

3.4 Summary

The performance of four di erent indium tin dioxide ﬁlms were investigated to determine which
ﬁlm would be the optimum electrode material for sensing manganese (Mn) with cathodic stripping
voltammetry. Two samples were heated at 190  C during deposition and two samples were annealed
in air for 1 hour at 400  C; generating one heated and annealed (HA), one heated and not annealed
(HNA), one not heated and annealed (NHA), and one not heated and not annealed (NHNA). All
ﬁlms were thoroughly characterized.

The samples that were heated showed comparable trends using AFM and XRD analysis. Both
of these samples (HA and HNA) also showed higher roughness, crystallite sizes, and (400) XRD
peaks with lower fwhm values compared to the non-heated ﬁlms. The annealed samples had
higher optical band gap values and carrier concentrations than those not annealed. 4-point-probe
measurements displayed NHNA having the highest resistance and resistivity with HA having the
lowest. HA, HNA, and NHA all exhibited good electrochemical detection ability for Mn. Only
the HA ﬁlm displayed a calibration curve that was analytically in acceptable range for the linear
correlation coe cient of 0.996.
It can be concluded that heating during the deposition of ITO
and post deposition annealing creates an optimum ﬁlm for the CSV detection of Mn. The key
parameters of the ﬁlm may be its roughness, crystallite size, resistivity, and I(222)/I(400) obtained
from XRD analysis. This study expands previous understanding of ITO and lays the ground work
for further analyte detection on this material using CSV for an economical way to mass produce
sensors.

76

BIBLIOGRAPHY

77

BIBLIOGRAPHY

[1] Radhouane Bel Hadj Tahar, Takayuki Ban, Yutaka Ohya, and Yasutaka Takahashi. Tin doped
indium oxide thin ﬁlms: Electrical properties. Journal of Applied Physics, 83(5):2631–2645,
1998.

[2] Giampiero de Cesare, Domenico Caputo, and Mario Tucci. Electrical properties of
IEEE Electron Device

ito/crystalline-silicon contact at di erent deposition temperatures.
Letters, 33(3):327–329, 2012.

[3] F Kurdesau, G Khripunov, AF Da Cunha, Md Kaelin, and AN Tiwari. Comparative study
of ito layers deposited by dc and rf magnetron sputtering at room temperature. Journal of
non-crystalline solids, 352(9):1466–1470, 2006.

[4] C Guillén and J Herrero. Structure, optical, and electrical properties of indium tin oxide thin
ﬁlms prepared by sputtering at room temperature and annealed in air or nitrogen. Journal of
applied physics, 101(7):073514, 2007.
JH Park, C Buurma, S Sivananthan, R Kodama, W Gao, and TA Gessert. The e ect of post-
annealing on indium tin oxide thin ﬁlms by magnetron sputtering method. Applied Surface
Science, 307:388–392, 2014.

[5]

[6] Eunkyoung Nam, Young-Hun Kang, Dong-Jin Son, Donggeun Jung, Sung-Jei Hong, and
Young Sung Kim. Electrical and surface properties of indium tin oxide (ito) ﬁlms by pulsed
dc magnetron sputtering for organic light emitting diode as anode material. Surface and
Coatings Technology, 205:S129–S132, 2010.

[7] M Gulen, G Yildirim, S Bal, A Varilci, I Belenli, and M Oz. Role of annealing temperature
on microstructural and electro-optical properties of ito ﬁlms produced by sputtering. Journal
of Materials Science: Materials in Electronics, 24(2):467–474, 2013.

[8] Necati Kaval, Carl J Seliskar, and William R Heineman. Spectroelectrochemical sensing
based on multimode selectivity simultaneously achievable in a single device. 16. sensing by
ﬂuorescence. Analytical chemistry, 75(22):6334–6340, 2003.

[9] Yining Shi, Andrew F Slaterbeck, Carl J Seliskar, and William R Heineman. Spectroelectro-
chemical sensing based on multimode selectivity simultaneously achievable in a single device.
1. demonstration of concept with ferricyanide. Analytical chemistry, 69(18):3679–3686, 1997.
[10] Jennifer M DiVirgilio-Thomas, William R Heineman, and Carl J Seliskar. Spectroelectro-
chemical sensing based on multimode selectivity simultaneously achievable in a single device.
6. sensing with a mediator. Analytical chemistry, 72(15):3461–3467, 2000.

[11] L Kerkache, A Layadi, E Dogheche, and D Rémiens. Annealing e ect in dc and rf sputtered

ito thin ﬁlms. The European Physical Journal-Applied Physics, 39(1):1–5, 2007.

78

[12] Mika Yamaguchi, Ari Ide-Ektessabi, Hiroshi Nomura, and Nobuto Yasui. Characteristics
of indium tin oxide thin ﬁlms prepared using electron beam evaporation. Thin solid ﬁlms,
447:115–118, 2004.

[13] M Ait Aouaj, Rosy Diaz, Azzam Belayachi, F Rueda, and M Abd-Lefdil. Comparative study of
ito and fto thin ﬁlms grown by spray pyrolysis. Materials Research Bulletin, 44(7):1458–1461,
2009.

[14] Kunisuke Maki, Nobuo Komiya, and Asako Suzuki. Fabrication of thin ﬁlms of ito by aerosol

cvd. Thin Solid Films, 445(2):224–228, 2003.

[15] T Karasawa and Y Miyata. Electrical and optical properties of indium tin oxide thin ﬁlms
deposited on unheated substrates by dc reactive sputtering. Thin Solid Films, 223(1):135–139,
1993.

[16] C May and J Strümpfel. Ito coating by reactive magnetron sputtering–comparison of properties

from dc and mf processing. Thin solid ﬁlms, 351(1):48–52, 1999.

[17] M Rottmann and K-H Heckner. Electrical and structural properties of indium tin oxide ﬁlms
deposited by reactive dc sputtering. Journal of Physics D: Applied Physics, 28(7):1448, 1995.
[18] I Baia, B Fernandes, P Nunes, M Quintela, and R Martins. Inﬂuence of the process parameters
on structural and electrical properties of rf magnetron sputtering ito ﬁlms. Thin Solid Films,
383(1-2):244–247, 2001.

[19] Byung Hak Lee, Iee Gon Kim, Sung Woo Cho, and Si-Hyun Lee. E ect of process parameters
on the characteristics of indium tin oxide thin ﬁlm for ﬂat panel display application. Thin
Solid Films, 302(1-2):25–30, 1997.

[20] Eugenia Matveeva. Electrochemistry of the indium-tin oxide electrode in 1 m naoh electrolyte.

Journal of the electrochemical society, 152(9):H138–H145, 2005.

[21] Hamid Reza Fallah, Mohsen Ghasemi, Ali Hassanzadeh, and Hadi Steki. The e ect of
annealing on structural, electrical and optical properties of nanostructured ito ﬁlms prepared
by e-beam evaporation. Materials Research Bulletin, 42(3):487–496, 2007.

[22] K Sreenivas, T Sudersena Rao, Abhai Mansingh, and Subhash Chandra. Preparation and
characterization of rf sputtered indium tin oxide ﬁlms. Journal of Applied Physics, 57(2):384–
392, 1985.

[23] John N Richardson, Aubrey L Dyer, Michael L Stegemiller, Imants Zudans, Carl J Seliskar,
and William R Heineman. Spectroelectrochemical sensing based on multimode selectiv-
ity simultaneously achievable in a single device. 13. detection of aqueous iron by in situ
complexation with 2, 2 ‘-bipyridine. Analytical chemistry, 74(14):3330–3335, 2002.

[24] Shirmir D Branch, Amanda M Lines, John Lynch, Job M Bello, William R Heineman, and
Samuel A Bryan. Optically transparent thin-ﬁlm electrode chip for spectroelectrochemical
sensing. Analytical chemistry, 89(14):7324–7332, 2017.

79

[25] Hong-Jeong Kim and Juhyoun Kwak. Electrochemical determination of total alkaline phos-
phatase in human blood with a micropatterned ito ﬁlm. Journal of Electroanalytical Chem-
istry, 577(2):243–248, 2005.

[26] J Agrisuelas, D Giménez-Romero, JJ García-Jareño, and F Vicente. Vis/nir spectroelec-
trochemical analysis of poly-(azure a) on ito electrode. Electrochemistry communications,
8(4):549–553, 2006.

[27] Kolade Ojo, Daoli Zhao, Cory A Rusinek, Sarah K Pixley, and William R Heineman. Cathodic
stripping voltammetric determination of cerium using indium tin oxide (ito). Electroanalysis,
29(4):1124–1130, 2017.

[28] Cory A Rusinek, Adam Bange, Mercedes Warren, Wenjing Kang, Keaton Nahan, Ian Pa-
pautsky, and William R Heineman. Bare and polymer-coated indium tin oxide as working
electrodes for manganese cathodic stripping voltammetry. Analytical chemistry, 88(8):4221–
4228, 2016.

[29] Tanya Shtoyko, Sean Conklin, Anne T Maghasi, John N Richardson, Aigars Piruska, Carl J
Seliskar, and William R Heineman. Spectroelectrochemical sensing based on attenuated
total internal reﬂectance stripping voltammetry. 3. determination of cadmium and copper.
Analytical chemistry, 76(5):1466–1473, 2004.

[30] Shruti Peshoria and Anudeep Kumar Narula. Bare indium tin oxide electrode for electro-
chemical sensing of toxic metal ion. Journal of Materials Science: Materials in Electronics,
pages 1–6, 2018.

[31] Cory A Rusinek, Wenjing Kang, Keaton Nahan, Megan Hawkins, Cooper Quartermaine,
Angela Stastny, Adam Bange, Ian Papautsky, and William R Heineman. Determination
of manganese in whole blood by cathodic stripping voltammetry with indium tin oxide.
Electroanalysis, 29(8):1850–1853, 2017.

[32] US EPA, 2004.
[33] Agency for Toxic Substances Registry and Disease. Toxicological proﬁle for manganese.
[34] John Stanislaus Roitz and KW Bruland. Determination of dissolved manganese (ii) in coastal
and estuarine waters by di erential pulse cathodic stripping voltammetry. Analytica Chimica
Acta, 344(3):175–180, 1997.

[35] Wenjing Kang, Wei Yue, Xing Pei, Adam Bange, Erin Haynes, William R Heineman, and
Ian Papautsky. Palladium-based sensor for electrochemical detection of manganese in the
environment. pages 1146–1149. IEEE, 2013.

[36] Alexander Goodwin, Andrew L Lawrence, Craig E Banks, Frédéric Wantz, Dario OmanoviÊ,
äebojka Komorsky-LovriÊ, and Richard G Compton. On-site monitoring of trace levels of
free manganese in sea water via sonoelectroanalysis using a boron-doped diamond electrode.
Analytica chimica acta, 533(2):141–145, 2005.

80

[37] M Christine Welch, E Craig Banks, äebojka Komorsky-LovriÊ, and G Richard Compton.
Electroanalysis of trace manganese via cathodic stripping voltammetry: Exploration of edge
plane pyrolytic graphite electrodes for environmental analysis. 2006.

[38] Olga MS Filipe and Christopher MA Brett. Cathodic stripping voltammetry of trace mn (ii)

at carbon ﬁlm electrodes. Talanta, 61(5):643–650, 2003.

[39] Zuzana Janosikova Dusan Bustin Richard G. Compton Miroslav Rievaj, Peter Tomcik. Deter-
mination of trace mn(ii) in pharmaceutical diet supplements by cathodic stripping voltammetry
on bare carbon paste electrode. Chem. Anal. (Warsaw), 53(153), 2008.

[40] Jason Stotter, Yoshiyuki Show, Shihua Wang, and Greg Swain. Comparison of the electrical,
optical, and electrochemical properties of diamond and indium tin oxide thin-ﬁlm electrodes.
Chemistry of Materials, 17(19):4880–4888, 2005.

[41] Asma El Kasmi, Michael C Leopold, Ryan Galligan, Rebecca T Robertson, S Scott Saavedra,
Kacem El Kacemi, and Edmond F Bowden. Adsorptive immobilization of cytochrome c on
indium/tin oxide (ito): electrochemical evidence for electron transfer-induced conformational
changes. Electrochemistry Communications, 4(2):177–181, 2002.

[42] Ciprian Iliescu, Hayden Taylor, Marioara Avram, Jianmin Miao, and Sami Franssila. A practi-
cal guide for the fabrication of microﬂuidic devices using glass and silicon. Biomicroﬂuidics,
6(1):016505, 2012.

[43] SM Sze. Physics of semiconductor devices. New York, Wiley-Interscience, 1981. 878 p.,

1981.

[44] Bard Allan J. and Faulkner Larry R. Electrochemical Methods Fundamentals and Applica-

tions. John Wiley and sons, Inc., United States of America, second edition, 2011.

[45] Strobel Howard A. and Heineman William R. Chemical Instrumentation: A Systematic

Approach. John Wiley and Sons, Inc., United States of America, third edition, 1989.

[46] Kissinger Peter T. and Heineman William R. Laboratory Techniques in Electroanalytical

Chemistry. Marcel Dekker, Inc., New York, second, revised and expanded edition, 1996.

[47] Saulius Graûulis, Daniel Chateigner, Robert T Downs, AFT Yokochi, Miguel Quirós, Luca
Lutterotti, Elena Manakova, Justas Butkus, Peter Moeck, and Armel Le Bail. Crystallog-
raphy open database–an open-access collection of crystal structures. Journal of Applied
Crystallography, 42(4):726–729, 2009.

[48] Saulius Graûulis, Adriana Daökevi , Andrius Merkys, Daniel Chateigner, Luca Lutterotti,
Miguel Quiros, Nadezhda R Serebryanaya, Peter Moeck, Robert T Downs, and Armel Le Bail.
Crystallography open database (cod): an open-access collection of crystal structures and
platform for world-wide collaboration. Nucleic acids research, 40(D1):D420–D427, 2011.

[49] Andrius Merkys, Antanas Vaitkus, Justas Butkus, Mykolas Okuli -Kazarinas, Visvaldas
Kairys, and Saulius Graûulis. Cod:: Cif:: Parser: an error-correcting cif parser for the perl
language. Journal of applied crystallography, 49(1):292–301, 2016.

81

[50] Robert T Downs and Michelle Hall-Wallace. The american mineralogist crystal structure

database. American Mineralogist, 88(1):247–250, 2003.

[51] David C Paine, T Whitson, D Janiac, R Beresford, Cleva Ow Yang, and Brian Lewis. A
study of low temperature crystallization of amorphous thin ﬁlm indium–tin–oxide. Journal
of Applied Physics, 85(12):8445–8450, 1999.

[52] Ho-Chul Lee and O Ok Park. Electron scattering mechanisms in indium-tin-oxide thin ﬁlms:

grain boundary and ionized impurity scattering. Vacuum, 75(3):275–282, 2004.

[53] EA Davis and NFf Mott. Conduction in non-crystalline systems v. conductivity, optical
absorption and photoconductivity in amorphous semiconductors. Philosophical Magazine,
22(179):0903–0922, 1970.

[54] Nevill Francis Mott and Edward A Davis. Electronic processes in non-crystalline materials.

OUP Oxford, 2012.

[55] H Kim, CM Gilmore, A Pique, JS Horwitz, H Mattoussi, H Murata, ZH Kafaﬁ, and
DB Chrisey. Electrical, optical, and structural properties of indium–tin–oxide thin ﬁlms
for organic light-emitting devices. Journal of Applied Physics, 86(11):6451–6461, 1999.

[56] A El-Denglawey, M Dongol, and MM El-Nahass. Photoinduced absorption edge shift of as

20 se 60 tl 20 ﬁlms. Journal of Luminescence, 130(5):801–804, 2010.

[57] J Barjon, N Habka, C Mer, F Jomard, J Chevallier, and P Bergonzo. Resistivity of boron

doped diamond. physica status solidi (RRL)–Rapid Research Letters, 3(6):202–204, 2009.

[58] Peihong Chen and Richard L McCreery. Control of electron transfer kinetics at glassy carbon
electrodes by speciﬁc surface modiﬁcation. Analytical Chemistry, 68(22):3958–3965, 1996.
[59] Imants Zudans, Jean R Paddock, Hideki Kuramitz, Anne T Maghasi, Chamika M Wansapura,
Sean D Conklin, Necati Kaval, Tanya Shtoyko, David J Monk, and Samuel A Bryan. Electro-
chemical and optical evaluation of noble metal–and carbon–ito hybrid optically transparent
electrodes. Journal of Electroanalytical Chemistry, 565(2):311–320, 2004.

[60] Romana Jaroöová, Paula M De Sousa Bezerra, Catherine Munson, and Greg M Swain.
Assessment of heterogeneous electron-transfer rate constants for soluble redox analytes at
tetrahedral amorphous carbon, boron-doped diamond, and glassy carbon electrodes. physica
status solidi (a), 213(8):2087–2098, 2016.

[61] Melanie Poon, Richard L McCreery, and Royce Engstrom. Laser activation of carbon elec-
trodes. relationship between laser-induced surface e ects and electron transfer activation.
Analytical Chemistry, 60(17):1725–1730, 1988.

[62] Richard S Nicholson. Theory and application of cyclic voltammetry for measurement of

electrode reaction kinetics. Analytical Chemistry, 37(11):1351–1355, 1965.

[63] Liang Liu, Shai Yellinek, Ido Valdinger, Ariela Donval, and Daniel Mandler.

Important
implications of the electrochemical reduction of ito. Electrochimica Acta, 176:1374–1381,
2015.

82

[64] Daniel C. Harris. Quantitative Chemical Analysis. Clancy Marshall, New York, 8th edition,

2010.

[65] Kathleen E Berg, Jaclyn A Adkins, Sarah E Boyle, and Charles S Henry. Manganese detection
using stencil-printed carbon ink electrodes on transparency ﬁlm. Electroanalysis, 28(4):679–
684, 2016.

83

CHAPTER 4

ELECTROCHEMICAL OXIDATION FOR THE MINERALIZATION OF PER- AND

POLYFLUOROALKYL SUBSTANCES

84

4.1

Introduction

A portion of this work has been published in Water and Environment Journal1. Per- and
polyﬂuoroalkyl substances (PFAS) are man-made chemicals that began production in the early
1950’s. Since their discovery, they were used heavily in customer and industrial applications that
took advantage of their oil and water repelling properties. [1–9] The main property of this class
of chemicals that sets them apart from others are their fully ﬂuorinated carbon chains or "tails".
Under the broad term of PFAS, which now encompasses more than 4,000 compounds, there is a
main group of perﬂuoroalkyl acids (PFAA) which encompasses the perﬂuoroalkylcarboxylic Acids
(PFCA) and Perﬂuoroalkylsulfonates (PFSA) with each group comprising of compounds with four
to 18 carbons in their background. The di erences between the PFCA and PFSA are the head groups
that are attached to the carbon chain tails. The two most popular chains that are being studied in
recent literature are perﬂuorooctanoic acid (PFOA) and perﬂuorooctane sulfonate (PFOS) since
they are widely distributed and have potentially negative health e ects. The molecular structures
of PFOA and PFOS in shown in Figure 4.1. Due to their fully ﬂuorinated carbon chain, they are
stable in the environment and bioaccumulate rapidly. Both compounds have been phased out by
industry but wide spread contamination still remains an issue for human exposure.

Few standard water treatment methods have been successful at remediating these compounds.
Those that have been can be divided into two classes, that of adsorption and ﬁltration and that of
destruction. Methods for adsorption and ﬁltration include ion exchange columns, nanoﬁltration,
ultraﬁltration, and reverse osmosis. Ion exchange (IX) columns have the added ability of being
regenerated for reuse which saves on cost. Some of the resins researched for PFAS include Purolite
A600E, A5202E, A532E, PAD500, PAD428, MN102, Sorbix A3F, PFA300, IRA67, IRA958, and
IRA67 [10–15]. However, although they can remove PFAS from a certain source, they do not
destroy the compounds. Ultimately, the PFAS is just concentrated into a secondary waste product.
With improper disposal of this waste, the contamination cycle can continue. Due to this being the

1 Water and Environment Journal 2020, DOI:10.1111/wej.12616

85

Figure 4.1: Molecular structure of (top) PFOA and (bottom) PFOS

case will all ﬁltration and adsorption techniques, there is a need for a destructive treatment method
to destroy the PFAS once and for all.

Destructive methods include thermal, plasma based water treatment, and electrochemical oxi-
dation. Thermal destruction utilizes temperatures above 1000  C and may not completely combust
of all the PFAS chains. [4, 6, 3] In addition, the use of thermal techniques lead to the generation of
various undesirable green house gasses. [3] Plasma treatment is new for PFAS, but has obtained
promising results for PFCA utilizing a laminar jet bubbling reactor and argon gas. [16, 17] Elec-
trochemical oxidation (EO) does not require the use of additional gasses or chemicals. It operates
using applied current to perform direct and indirect oxidation of contaminant species. [18] Elec-
trode materials are limited for water based remediation of PFAS due to the need of high current and
voltage. Boron-doped diamond (BDD) has been a widely explored material for these applications.
[5, 19–31] It can withstand the high current and potential while promoting hydroxyl radicals which
can be of beneﬁt in indirect oxidation of co-contaminants. [18]

86

EO has been applied in literature studies to address PFAS contamination using current densities
from 0.04 - 50 mA/cm2. [5, 28, 22, 23, 27, 21, 25, 24, 26, 19, 20, 29, 31] Time frames for these
treatment experiments have ranged from minutes up to 120 hours.[21, 29, 26]. One of the daunting
concerns with EO can be the high cost of the electrode material and high energy consumption, but
it is important to remember that EO is one of the only remediation techniques that has proven to
provide complete mineralization of PFAS. [27] All possibilities to reduce operating costs should
be considered. The following sections discuss using EO independently and in conjunction with a
regenerable IX column. [3, 4, 6, 1] There is a focus on combining current densities to evaluate if
the overall energy associated with the process can be lowered while still obtaining the desired level
of compound destruction.
4.2 Electrode Stack Characterization

Before performing electrochemical oxidation experiments for the destruction of PFAS, the
electrode stack being used needed to be characterized. This stack was comprised of two double
side coated BDD on niobium (Nb) anode plates and three double side coated BDD on Nb cathode
plates. The two anodes were sandwiched in between the cathodes and an exposed anodic surface
area of 105 cm2 was used with 3 mm distance in between each plate. Characterization was done
using scanning electron microscopy (SEM), Raman spectroscopy, and electrochemical methods to
determine the capacitance and rate constants for two redox couples. The SEM image in Figure 4.2
shows a high surface area ﬁlm with no cracks on the surface.

Raman spectroscopy was used to determine the boron content of the ﬁlm. The spectrum is
shown in Figure 4.3 and the total boron content was 4.88 x 1020 cm-3 +/- 5.7 x 1019 cm-3. This
classiﬁes this ﬁlm as highly doped.

In order to ﬁnd the electrochemical capacitance of the ﬁlm, cyclic voltammetry (CV) was used.
This experiment was performed by exposing a small area of the electrode to a 1 M potassium
chloride (KCl) solution and measuring the current as the potential was scanned using a laboratory
potentiostat. After each scan, the scan rate was increased and the measurement was performed
again. The current was then recorded at 0 V and plotted against the scan rate at which is was

87

Figure 4.2: Scanning electron microscopy (SEM) image of the BDD surface

Figure 4.3: Raman spectroscopy of BDD electrodes

obtained, Figure 4.4. The scan rates of 0.1 - 0.5 V/s were used and three sets of each were
performed. The capacitance of the ﬁlm was 2.66 x10-04 F/cm2 +/- 4.88 x10-07 F/cm2, which was
obtained from the slope of the graph relating the current and scan rate.

The electrochemical characterization was completed by determining the rate constants of the
BDD for an inner and outer sphere electron redox couple. The inner sphere redox couple that was
chosen was potassium ferrocyanide (FeCN) and the outer sphere redox couple that was chosen was
ruthenium hexaammine (RuHX). Each redox couple was used in a 1 M KCl solution and the current

88

Figure 4.4: Current versus scan rate plots to determine the capacitance of BDD

was measured as the potential was scanned for scan rates of 0.05 to 1.0 V/s. The Nicholson and
Shain method was used to determine each rate constant. All together, each scan rate was performed
a total of three times so an average could be taken. Figure 4.5 shows a CV for both redox couples.
The rate constant for FeCN was 3.96 x10-04 +/- 4.47 x10-06 and the rate constant for RuHx was
9.00 x10-04 +/- 3.28 x10-05.

Figure 4.5: Cyclic voltammogram of both ferrocyanide and ruthenium hexamine on the BDD
surface

89

4.3 Constant Current

4.3.1 Methods

All electrochemical experiments were performed with a 180W multi range 60V/8A DC power
supply from BK Precision Instruments. All solution matrices were made with analytical grade
sodium sulfate (Na2SO4) from J.T. Baker Inc. at a concentration of 50 mM. Perﬂuorooctanoic
acid (PFOA) and perﬂuorooctane sulfonate (PFOS) were purchased from Sigma Aldrich. Nalgene
bottles were used for all solutions and testing volumes were 750 ml. Solutions were constantly
stirred during testing using a Hanna Instruments magnetic stirrer HI 190M on max. A pH probe,
temperature probe, and ﬂuoride ion selective electrode (ISE) were purchased from Thermoﬁsher
Scientiﬁc and a ﬂuoride standard that was used to calibrate the ISE was purchased from Sigma
Aldrich.

2ml of sample was removed and stored in the refrigerator in polyethylene vials at 4  C until
analysis was performed with the ﬂuoride ISE. A total ionic strength bu er (TISAB) was made
using glacial acetic acid (Fisher Chemical), reagent grade sodium chloride (NaCl) (Sigma Aldrich),
sodium hydroxide (NaOH) (Spectrum Quality Products Inc.) and DI water and used to dilute
each sample 1:1. The Michigan State University College of Natural Science Research Technology
Support Facility performed the liquid chromatography mass spectroscopy (LC/MS) measurements
for the samples spiked with only PFOA.

4.3.2 Results and Discussion

4.3.2.1 PFOS
PFOS studies were performed with the previously characterized electrode stack. Constant
current densities of 5, 10, 25, and 50 mA/cm2 were used. Each experiment was performed in
triplicate. The current densities of 5 and 10 mA/cm2 were performed for 1.5 hours but the current
densities of 25 and 50 mA/cm2 were only performed for 1 hour. The starting concentration of

90

PFOS was 10 parts per million (ppm) as calculated for the amount of powder that was added. With
increasing current density the initial temperature, pH, and potential of each system also increased.

Figure 4.6: PFOS Temperature, pH, and voltage trends for the constant current density studies

These ﬁgures are shown in Figure 4.6. pH for all tests began at 6 and increased over time. The
pH for 10, 25, and 50 mA/cm2 increased to 10.5 whereas the pH of the 5 mA/cm2 solution has an
upwards trend tending towards 10.5 but does not reach there in 1.5 hour. BDD in a known e cient
producer of hydroxyl radicals and previous studies using BDD have shown that the pH could
increase by 5 units depending on the applied current density. [32, 33] The temperature started at
room temperature for all tests and increased with time and potential for the low current densities of 5
and 10 mA/cm2 remained constant, whereas, the potential for both 25 and 50 mA/cm2 both slightly
decreased over time. PFOS degradation was monitored with the F- ISE and the concentration of
free F- was used to calculate the percentage of deﬂuorination. Figure 4.7 shows the percent of

91

deﬂuorination for the current densities studied.

Figure 4.7: PFOS reduction overtime with each current density tested

The three tests at each current density are taken into account and an average is graphed with
its corresponding error. It is clear that the PFOS degradation is independent of current density for
these studies. Each current density also had a standard deviation of 4 % at each time point. The 1st
order rate constants for each current density were also determined and graphed as an average value.
The PFOS concentrations were calculated based on the percentage of ﬂuoride released under the
assumption that no shorter chains were being formed due to lack of analysis equipment. The rate
constants for each were 0.39 +/- 0.02 h-1, 0.39 +/- 0.07 h-1, 0.45 +/- 0.10 h-1, and 0.45 +/- 0.10 h-1,
for 5 at 1.5 hours, 10 at 1.5 hours, 25 at 1 hour, and 50 mA/cm2 at 1 hour respectively. Each rate
constant had a linear correlation coe cient (R2) above 0.85. The average linear correlations were
0.9821, 0.958, 0.951, and 0.8628 for 5, 10, 25, and 50 mA/cm2, respectively. 5 and 10 mA/cm2
share the same rate constant as do 25 and 50 mA/cm2.

To ensure the validity of the PFOS results, a quality assurance check was performed with the
ISE. To do this, each of the 1 hour time points from the three 25 mA/cm2 runs were retested with
the ISE at the end of the complete study. The repeatability in the results obtained was 105 +/-
5.6 %, validating that the ISE measurements were accurate. To advance our understanding of how

92

Figure 4.8: PFOS 1st order rate constants

PFAS behave in electrochemical experiments, perﬂuorooctanoic acid (PFOA) was studied with the
addition of LC/MS analysis.

93

4.3.2.2 PFOA
The degradation of PFOA was studied at di erent applied current densities using the previously
characterized electrode stack. Current densities included 1, 5, 10, 25, and 50 mA/cm2. All tests
were performed for 1 hour. The stack was lowered into 750 ml of 50 mM sodium sulfate that the
PFOA was spiked into using a concentrate made from the dissolved powder. Cell voltages spanned
rom +4 V to +9 V at the lowest and highest current density, respectively. All solutions were started
at room temperature and experiences an increase to 40  C at 50 mA/cm2. All testing solutions
began at a pH of 5, one pH unit lower than the PFOS solutions, and increased over time to 10.5
except for the current density of 1 mA/cm2, where the pH remained constant. The current densities
chosen here suggest that the reactions a ecting the pH may be di erent between 1 mA/cm2 and the
rest of the current densities. Graphs showing the change in temperature, pH, and potential for each
current density are shown in Figure 4.9.

The temperature increased over time for all current densities studied with the highest having
the largest increase. The potentials were stable across the full testing time for all current densities
except for 50 mA/cm2 where a slight increase was observed. The starting concentration of PFOA
was 6.5 parts per million (ppm) as determined by liquid chromatography mass spectroscopy
(LC/MS) measurements. A graph representing the PFOA degradation over time, in terms of
percent deﬂuorination obtained by using the F- ISE, is shown in Figure 4.10. The release of ﬂuoride
increases over time with increasing current density. The ﬁve tests also show good repeatability
between the three sets that were performed.

Each current density study was also ﬁt to ﬁrst order kinetics using the PFOA degradation
over time from the LC/MS analysis. This way the direct rate constant relating to PFOA can be
determined instead of inferring from the ISE analysis as was done for PFOS. These plots are shown
in Figure 4.11. All current densities has a linear correlation coe cient of greater than 0.90 with
both 5 and 10 mA/cm2 being above 0.99. The rate is increasing as current density is increased. the
values for k1 are as follows: 0.34, 1.35, 2.04, 3.77, and 6.27 h-1, for 1, 5, 10, 25, and 50 mA/cm2,
respectively.

94

Figure 4.9: PFOA Temperature, pH, and voltage trends for the constant current density studies

Figure 4.10: PFOA degradation over time at each current density value

Another interesting takeaway from this constant current density study comes from taking a
closer look at Figure 4.10. This being that the slopes of the PFOA deﬂuorination lines only seem

95

Figure 4.11: 1st order rate constants for PFOA based on LC/MS measurements for each current
density study.

to di er in the ﬁrst 0.25 h of each test. For the remaining time of each, the slopes are relatively
the same. For instance, for the ﬁrst 0.25 hours of the test, 1 mA/cm2 had a slope of 0.28 %/h that
increased to 2.0 %/h at 50 mA/cm2. However, for 0.25h untill the end of 1 hour, the slope for 5
through 50 mA/cm2 were 0.54, 0.58, 0.62, and 0.49 %/h for 5, 10, 25, and 50 mA/cm2, respectively.
This raises the question as too if the higher current density is truly only beneﬁcial in the ﬁrst 0.25
h of each study and could then be lowered to save on energy consumption.

Fluoride quality analysis was also performed for PFOA using the 1 hour time sample from each
25 mA/cm2 test. The obtained concentration of ﬂuoride was 99.7 +/- 1.3 % of what the original
tested value was. This veriﬁed the F- ISE measurements were accurate over time for this study.

4.3.2.3 Comparison of ISE Results to LC/MS Results
At the time of analysis, the Michigan State University mass spectroscopy facility could only
analyze for the single compound of PFOA. Due to this limitation, shorter chain compounds could
not be directly identiﬁed, but by determining if there was a di erence in the concentration of PFOA
that remained from ISE analysis and what the LC/MS provided, it could be assumed that any
di erence was from the generation of shorter chain ﬂuorinated carboxylic acids. All ﬁve current

96

densities that were used were tested with LC/MS and the PFOA concentrations were graphed against
those calculated from the ISE.

Figure 4.12: Comparison of the PFOA concentration obtained by the F- ISE and LC/MS

Figure 4.12 shows the comparison of the concentrations obtained using the two di erent analysis
methods. The concentrations for both methods overlap for the lowest current density of 1 mA/cm2
at 4.3 +/- 0.2 ppm from the ISE and 4.4 from LC/MS after 1 hour, but as the current density is
increased, the two concentrations begin to di er. The di erences between the two methods are
shown in Table 4.1. Since there is a smaller amount of PFOA detected with LC/MS than there is
free ﬂuoride that would pertain to that amount of degradation, the remaining PFOA is assumed to
have converted to shorter chains through an unzipping mechanism.

97

Table 4.1: Comparison of remaining concentration of PFOA obtained from the ISE analysis and
LC/MS analysis

4.3.3 Fluorination of BDD

It has been shown in literature that perﬂuoroalkyl compounds can be used to ﬂuorinate a BDD
surface. [34, 35] As such, it is important to consider that some of the ﬂuoride released during
the decomposition of PFAS could adsorb onto the BDD surface. This was investigated in three
ways: (1) determining the contact angles between a unused and used BDD electrode for PFAS
remediation (2) measuring a potassium ferrocyanide (FeCN) CV on a used and unused electrode,
and (3) investigating the potential window changes between a used and unused electrode.

The contact angle measurements were done using a 0.5 µL drop of water on the surface of each
anode. The part of each electrode that was submerged into solution was labeled as "used" and the
part that was not submerged and remained above the solution volume was labeled as "unused". For
the anodes, the used electrodes displayed an average contact angle of 84  +/- 5.6  and the unused
portion displayed 75  +/- 4.9 .

The FeCN study was used due to its behavior as an inner sphere redox couple. If the ﬂuoride was
deposited onto the BDD surface, the surface sensitive reaction of ferrocyanide could be e ected.
Cyclic voltammetry was performed in 1 mM FeCN in 1M KCl at 0.1 V/s on the used and unused
portion of the BDD anode used in the previous PFOA degradation studies. Figure 4.13 shows the
voltammogram for both electrode portions. It is immediately apparent that the response from the
two di ers. The unused portion has a characteristic shape showing the oxidation and reduction
of ferrocyanide whereas the used portion has no visible peaks. This is indicative that there was a

98

change to the surface of the BDD anode during the remediation experiments with PFOS and PFOA.
To investigate this further, the potential windows of the used and unused portions were compared.

Figure 4.13: 1mM ferrocyanide in 1 M KCl measured on both the used and unused portion of the
anode that was used for the constant current density PFOA degradation studies

The potential window of each anode surface was investigated in 1 M sulfuric acid at a scan rate
of 0.1 V/s to determine if there was a shift in the hydrogen or oxygen evolution potentials. Four
scans were performed with the last two being graphed for each in Figure 4.14. It can be seen that
the used anode has a slightly larger oxygen evolution potential and a further negative hydrogen
evolution potential. According to literature, ﬂuorinated BDD electrodes have a larger window
compared to diamond electrodes and can show a potential window of 5 V [36–38]. There is a
larger di erence in the hydrogen evolution between a ﬂuorinated BDD and non ﬂuorinated BDD
compared to the oxygen evolution due to the fact that the ﬂuorinated surface is more hydrophobic,
seen with contact angles, which decreases the adsorption strength of water and atomic hydrogen
shifting the reaction more negative [37]. Between the contact angle measurements, investigation
into the redox couple of ferrocyanide, and the comparison of the potential window for the used
and unused BDD, it is conﬁrmed that there was ﬂuorination of the diamond surface during the
electrochemical oxidation experiments.

In an attempt to bring the BDD surface back to its original non-ﬂuorinated state, it underwent

99

Figure 4.14: CV in 1 M sulfuric acid for both the used and unused portion of the anode

a hydrogen termination process and re-evaluation of the FeCN CV and potential window study.
Hydrogen termination of the used surface was performed by placing it into a hydrogen atmosphere
in a hot ﬁlament chemical vapor deposition reactor for one minute. The resulting ferrocyanide
voltammogram and potential window scan are shown in Figures 4.15 in (A) and (B), respectively.

Figure 4.15: (A) CV comparison in1 mM ferrocyanide in 1 M KCl for the used anode, unused
anode, and used anode post hydrogen termination and (B) CV in 1 M sulfuric acid for the used,
unused, and used electrode post hydrogen termination

It can be seen that the FeCN peak showed a signiﬁcant current response and clear oxidation
and reduction peaks for the hydrogen terminated used sample. The current response after hydrogen
termination was greater than the response from the unused sample potentially due to now being

100

cleaner and fully hydrogen terminated whereas the unused portion may have become partially
oxygen terminated over time.
In addition, the potential window shifted closer to the unused
electrode by having the hydrogen evolution reaction shift to more positive values and the oxygen
evolution reaction shift to more negative values. Both of these tests veriﬁed that the ﬂuorination of
the BDD surface could be removed by placing the electrode in a hydrogen plasma after being used
for the degradation experiments.
4.4 Combined Current Density

To further evaluate on the ability of electrochemical oxidation to degrade PFOA over time and
to begin to investigate how to save on energy costs, a combined current density technique was
investigated. This technique employed the use of a high current density for the ﬁrst 15 minutes
of each test before lowering it down for the remaining 75 minutes. This decision came from the
fact that the most signiﬁcant beneﬁt for the high current density application seen with the constant
current densities was in the ﬁrst 15 min of each test. The experimental conditions used in the
previous constant current density section were employed again in this study. The high current
density chosen was 50 mA/cm2 as it showed the greatest initial change in PFOA concentration from
Section 4.3.2.2. Three lower current densities of 1, 5, and 10 mA/cm2 were investigated for the
remaining time frame of each test. Throughout each test, samples were removed at 0.25, 0.5, 0.75,
and 1 hours and stored in the refrigerator until analysis was performed.

Graphs showing the change in temperature, pH, and potential for each current density are shown
in Figure 4.16. The temperature from all three combined tests increases in the ﬁrst hour before
steadily decreasing with the reduction in applied current. The pH for all three tests also mirrored
one another with an increase before stabilizing out at 10.5. The potential slightly decreased
during the higher current density application and then remained stable during the lower current
density application for all three tests. Each PFOA sample was evaluated with the ﬂuoride ISE
as used previously. The starting concentration of PFOA was 6.5 ppm in 50 mM sodium sulfate
determined from LC/MS experiments. There was 750 ml of solution placed in a Nalgene bottle
and stirred continuously. The PFOA reduction in terms of deﬂuorination percentage over time for

101

Figure 4.16: Temperature, pH, and voltage trends for the combined current density studies

each combined current density compared to the constant current density of 50 mA/cm2 provided
as Figure 4.17.

Figure 4.17: PFOA reduction overtime with each combined current densities tested together as
well at the standard 50 mA/cm2 for comparison

102

It can be seen that the 50 mA/cm2 combined with 1 mA/cm2 (50&1), has the lowest evolution
of ﬂuoride between each run, but a low standard deviation (n=3) between each run. The 50&10
run also has a low standard deviation between sample sets. For the run using 50&5, the points line
up close to those of 50&10, and even overtake them at 1 h, but there is a much larger standard
deviation between the tests. As such, using 50&5 could outperform 50&10 as well as be very
similar to the degradation that is obtained with the constant current density run of 50 mA/cm2.
In order to evaluate the rate constants of each combined current density test, two sequential rate
constants were used to analyze the ﬂuoride ISE data. k1 for the rate constant from time 0 to 0.25
hours, and k2 from 0.25 hours until the end of the study, 1.5 hours.

Figure 4.18: First order rate constant ﬁts for the ﬁrst 0.25 h, k1, and the remainder of each test, k2.

Figure 4.18 shows the ﬁtted rate constants for the three combined current density studies. Table
4.2 shows the numerical values for each of the rate constants and their obtained linear ﬁts. It is
noted that k1 is only based o  of time 0 and 15 min. With only two points the linear coe cient will
always be equal to 1. k2 is based o  of 5 points allowing for a better understanding of if a linear
relationship exists.

k1 is the same for both 50&1 and 50&5 with 50&10 having a slightly lower value. Overall,
since this time point it taken based solely on the applied high current density of 50 mA/cm2, it
shows good repeatability. 50&5 had the largest standard deviation in the k1 value. For k2 all the

103

Table 4.2: Rate constants and their linear ﬁts for the three tested combined current densities

linear coe cients are all above 0.98 with 50&1 and 50&5 being greater than 0.99. 50&1 and
50&10 have a lower standard deviation in their values compared to 50&5. Due to this the k2 value
for 50&5 and 50&10 are similar. The k2 value for 50&1 is roughly half in comparison.

4.4.1 Comparison of ISE and LC/MS

As performed for the constant current density study, a LC/MS and F- ISE comparison was done
for the best combined current density study of 50&5. The di erence in PFOA concentrations
over time for each analysis technique are shown in Figure 4.19. The same trend seen before is
shown with this data. The concentration of PFOA obtained from the LC/MS data is lower than that
determined from the ISE analysis. Again, this suggests that the remaining ﬂuoride not detected
from the ISE is bound in other short chains that were formed as part of the degradation process,
or have ﬂuorinated the anode’s surface. 1.71 +/- 0.1 ppm of PFOA was estimated to be remaining
by the ISE analysis but LC/MS provided a ﬁnal concentration of 0.40 ppm showing the increased
ability in electrochemical oxidation to break apart PFOA. The rate constant was also determined
for 50&5 based on LC/MS measurements. For time 0 to 0.25h the rate constant was 4.20 h-1 and
from 0.25 to 1 h the rate constant was 2.85 h-1. These rate constants are much higher than those
corresponding to the ISE analysis. This again suggests that PFOA may be degrading quickly into
shorter chain PFAS.

104

Figure 4.19: Comparison of the PFOA concentration obtained by the F- ISE and LC/MS

4.4.2 Remediation of Simulated Ion Exchange Regenerate Solutions

To advance the application of the combined current density method, a treatment train approach that
can enhance an existing technology would be beneﬁcial. One of the most promising remediation
techniques is adsorption using ion exchange (IX). This has been shown successful in literature for
removing a wide array of PFAS. Some of these columns that remove PFAS in particular have the
ability to be regenerated and re-used. [10, 5, 13–15, 11, 12] This is a huge cost beneﬁt, but has
the downside of generating a concentrated waste product. The regenerate solution out of an IX
process can comprise of PFAS in the parts per million (ppm) range in a complex matrix of a brine
with an alcohol. [12] Due to di culties in disposal of this waste, and the potential to continue the
contamination cycle with improper disposal, a destruction method is needed to remove the PFAS
once and for all. However, due to the nature of the regenerate solution, i.e. high PFAS concentration
and high salt, it is a perfect electrolyte solution to apply electrochemical oxidation (EO). In order to
determine if EO is an appropriate treatment technology for these solutions, those currently existing
in literature to regenerate PFAS speciﬁc columns need to be tested.

The four main ion exchange (IX) regenerate solutions that were simulated in the laboratory
include: 6% ammonium hydroxide (NH4OH) with 2% methanol (MeOH), 6% sodium chloride
(NaCl) with 2% MeOH, 6% sodium hydroxide (NaOH) with 2% MeOH, and 0.5% NH4OH with

105

0.5% ammonium chloride (NH4Cl). The solutions with MeOH were done at 2% MeOH due to
the ability to distill o  the majority in a distillation process that can then be reused for future
regeneration.[12] NH4OH with MeOH was chosen due to its ability to remove PFOA from the IX
columns of PUROSORB PAD 500 and MACRONET MN102. [11] NaCl with MeOH was chosen
for its ability to remove PFOS from IX columns of PFA300, IRA958, and BAC. [14, 13, 15] NaOH
with methanol was chosen for its ability to regenerate PFA300 and IRA67 for PFOS removal.
[14, 13] The mixture of NH4OH and NH4Cl was chosen for its ability to regenerate Purolite A600E
and A520E with PFOA, PFBA, PFOS, and PFBS. [10, 11] Each solution was spiked with greater
than 10 ppm of PFOA and PFOS due to both being in the tens of ppm range in known ion exchange
solutions. [39] The high current density application time was increased from 15 minutes, from the
earlier studies, to 1 hour due to the high chemical oxygen demand (COD) present in most of the
simulated regenerate solutions. At 2% MeOH, the COD values for these solutions were around
30,000 ppm. The total testing time was also increased to 12 hours to account for the additional
competing COD content and higher initial PFAS concentrations.

4.4.2.1 Comparison of 50 mA/cm2 to 50&5 mA/cm2
The ﬁrst IX regenerate solution comparison performed utilized the solution of 6% NH4OH
with 2% MeOH was to compare the use of 50 mA/cm2 for an entire test to using a combination
of 50&5 mA/cm2. This was done by using the higher current density for an extended time of 1
hour before lowering the current density for the remainder of the test. Graphs showing the change
in temperature, pH, and potential for each current density are shown in Figure 4.20. The pH for
the 50&5 solution decreases much more rapidly in the ﬁrst 1.5 hours although both tests employed
50 mA/cm2 for the 1st hour. 50&5 then showed an increase in pH where it ended at about 10
compared to the constant current density that had a continued gradual decrease and ended closer to
8. The temperature for moth tests mirrored on another for the higher current density application.
after 50&5 was lowered to 5 mA/cm2, the temperature fell and stabled out. The 50 mA/cm2 run
only displayed a slight decrease in temperature over time. The potential for other tests started

106

relatively high and fell rapidly in the ﬁrst our before relatively stabling out. The 50&5 test had a
lower potential for the lower applied current.

Figure 4.20: Temperature, pH, and voltage trends for the comparison of 50 and 50&5 in IX
simulated solution

Figure 4.21: Comparison of the PFOA (left) and PFOS (right) degradation using 50 mA/cm2 and
50&5 mA/cm2

The change of PFOA and PFOS over time are shown in Figure 4.21 left and right, respectively.

107

When using the constant current density of 50 mA/cm2 for the entire test, the COD was reduced
by 72% and the ammonia was reduced by 88% compared to 43% and 47% for the combined
current density of 50&5, respectively. The PFOA concentration in the constant current density
test decreased by greater that 99% in 12 hours. Based on the dilution factor used when sending
samples for analysis, it is known that there is <66 parts per billion (ppb) remaining, but the exact
number is unknown. The PFOS degradation was also greater than 99% for this constant current
density study after 12 h. For the 50&5 study, the PFOA concentration degraded 90% and the PFOS
degraded 59% in the same total time frame.
It can be seen from the PFOS degradation graph
(Figure 4.21) that PFOS reached a higher degradation percent in 8 hours then at 12 hours using
the combined current density. This may be due to a number of factors including, generation of
PFOS from precursors or an outlier in the measurement technique. Generation from precursors is
unlikely since the solutions were spiked with purchased powders. However, these powders are not
100% pure. The human error in removing a sample and diluting x100 may have caused for the
discrepancy. Overall, 50&5 obtained good degradation results for PFOA and PFOS in the same
time frame as using 50 mA/cm2 for the entire time. Since this is only one potential IX regenerate
solution, it is of importance to test the other solutions researched in industry with a further analysis
into the degradation of PFAS.

4.4.2.2 Comparison of four di erent regenerate solutions
In order to compare the ability of using 50&5 further, each of the simulated solutions were
tested using 50&5 for the degradation of primarily PFOA and PFOS. Graphs showing the change
in temperature, pH, and potential for each current density are shown in Figure 4.22. The highest
pH belonged to the NaOH and MeOH solution at 14 compared to all of the others that ﬂuctuated
at levels below 10. The temperature increased for all solutions during the high current density
application and lowered once the current density was lowered. The highest temperature of 60  C
was reached with the NH4OH and MeOH solution. This was also the solution that had the highest
starting potential out of all those tested. The other three solutions had similar potentials at the high

108

and low current density applications.

Figure 4.22: Temperature, pH, and voltage trends for the comparison of the four IX simulated
solutions

Figure 4.23: Comparison of the PFOA (left) and PFOS (right) degradation using the four tested
simulated IX solutions

The change in PFOA and PFOS for each solution over the 12 hours is shown in Figure 4.23
with PFOA degradation on the left, and PFOS degradation on the right. PFOA degradation was
obtained in three out of the four tested solutions. Its degradation was 90% in the NH4OH and

109

MeOH solution, 52% in the NaCl and MeOH solution, and >99% in the NH4OH with NH4Cl
solution. It was not degraded in the NaOH based solution. This lack of degradation in NaOH
may be due to the signiﬁcantly higher pH that was characteristic of this solution compared to the
others (Figure 4.22). It has been shown in literature that the degradation of PFOA can be inhibited
in solutions with alkaline pH. [40, 29, 31] Each PFOA experiment was ﬁt to sequential 1st order
reaction kinetics, Figure 4.24.

Figure 4.24: 1st order rate constant kinetics for PFOA in each simulated solution

For the ﬁrst hour, the PFOA rate constants were 1.93, 0.06, 0.75, and 1.5 h-1 for NH4OH
and MeOH, NaOH and MeOH, NaCl and MeOH, and NH4OH with NH4Cl, respectively. After
one hour, the rate constant for every solution expect for NH4OH with NH4Cl was 0 h-1. That
of NH4OH with NH4Cl was 0.37 h-1. These results show that there is no additional degradation
in the simulated solutions for the majority of the tests after the current density is lowered. It is
interesting to note that out of the two solutions with the highest degradation for PFOA, the NH4OH
solution with MeOH obtains all of that degradation at the higher current density. Whereas, the
combination of NH4OH and NH4Cl does have slow continued degradation. In the ﬁrst hour, the
NH4OH with MeOH solution also has the lowest pH value during this time and the highest achieved
temperature which both may contribute to the increased degradation of PFOA. The other ammonia
based solution may have ultimately outperformed the solutions with MeOH due to the fact that

110

this solution does not have the competing COD that the other solutions contain. Table 4.3 lists the
PFOA concentrations, rate constants, and linear ﬁts obtained from the studied solutions at 1h and
12 h. This table shows that there is no added beneﬁt, in terms of PFOA degradation, to continue
to run the test past 1 h for all solutions except for that based on NH4OH and NH4Cl. It is also
noticeable that in 1h the NH4OH solution with MeOH obtained the most PFOA degradation with
the fastest rate constant despite having the added competition of COD although this rate constant
does not have a nice linear ﬁt.

Table 4.3: Comparison of 1 h PFOA degradation and 12 h PFOA degradation

For PFOS, only three out of the four solutions obtained degradation. However, the solution
that PFOS did not degrade in is di erent from the one that PFOA did not degrade in. PFOS did
not degrade in the NaCl with MeOH solution. In addition it obtained a much lower degradation in
the NaOH with MeOH compared to the two ammonia based solutions. The degradation was 40%
in the NaOH solution, 59% in the NH4OH with MeOH solution, and 91% in the NH4OH with
NH4Cl solution at the end of 12 hours. Sequential ﬁrst order rate constants were also determined
for PFOS for the ﬁrst hour followed by the remainder of the study. The two solutions with non zero
rate constants for the 1st hour of the study were the two NH4OH based solutions with 0.29 h-1 and
2.73 h-1 for that with MeOH and that with NH4Cl, respectively. They had a linear correlation of
0.089 and 0.59 for NH4OH with MeOH and NH4Cl, respectively. For the remaining portion of the
test, 1-12h, the rate constants were 0.05 hr-1 for both NH4OH and NaOH with MeOH and 0.02 for
NH4OH with NH4Cl. The ﬁttings for each rate constant can be shown in Figure 4.25. It is also

111

interesting to note that the NH4OH solution with NH4Cl now had the highest degradation in the
ﬁrst hour for PFOS compared to the NH4OH solution with MeOH; opposite of the PFOA trend. In
addition the mixed ammonia solution did not have continued degradation past this point whereas
the ammonia based MeOH solution did.

Figure 4.25: Sequential 1st order rate constant kinetics for PFOS in each simulated solution

A literature study on a di erent electrode materials also found that PFOS did not degrade in
their electrolyte of sodium perchlorate (NaClO4) when starting with 20 ppm and applying roughly
7 mA/cm2. [41] It has been proposed in the literature that PFOS prefers the gas-liquid interface
which may inhibit its degradation. [42] To investigate this electrochemically, the oxygen evolution
reaction for each solution on the BDD surface was studied. The oxygen evolution reaction was
investigated in each IX solution using a clamp cell placed on top of the used BDD anode. A counter
electrode of graphite and a reference electrode of Ag/AgCl was used. The exposed surface area
was 0.13 cm2. A graph of the current density compared to potential is shown in Figure 4.26

The two solutions with the lowest PFOS degradation have the earliest onset of the OER, the
NaCl and NaOH based electrolytes. The two solutions with the highest PFOS degradation had the
furthest potential where the OER began. This suggests that the oxygen evolution could play a role in
inhibiting the PFOS from reaching the anode surface to degrade. A report comparing two di erent
electrode materials in the same solution looked into the OER compared to the degradation % that

112

Figure 4.26: Oxygen evolution reaction performed for all four solutions

was achieved. The authors found that the electrode with the further onset of the OER had a greater
PFOS degradation. [43] Table 4.4 provides a summary of the PFOS concentrations, rate constants,
and linear ﬁts at 1h and 12h. In these complex solutions, the best electrolyte for treating both PFOA
and PFOS are ammonia based solutions. It has also been discovered that there is minimal beneﬁt
from running these tests passed the 1 h time point at the higher current density except to gain a
continued slow degradation of either PFOA or PFOS depending on the ammonia based solution
used.

Table 4.4: Comparison of 1 h PFOA degradation and 12 h PFOA degradation

113

4.4.2.3 Byproduct Formation Investigation
Since the majority of the IX solutions are comprised of a salt or a base and methanol, it is
important to look at the oxidation by-products from these constituents. For example, methanol
can be oxidized to for formaldehyde and chloride can be oxidized to perchlorate, both of which
are undesirable products. Formaldehyde is classiﬁed as a hazardous substance. The calorimetric
formaldehyde detection method used was performed by mixing formaldehyde with the Nash reagent.
[44] Formaldehyde reacts with the Nash reagent to form diacetyldihydrolutidine which is a yellow
compound whose absorbance can be detected at 410 nm using a UV/vis spectrophotometer. The
Nash reagent is a mixture of ammonium acetate, acetic acid, acetylacetone and DI water. Once the
solutions are mixed they are heated and kept in the dark too cool before analysis with the UV/vis.
Using this method solutions containing less than 8 µg of formaldehyde per ml can be analyzed.
[44] Those of higher concentrations can be diluted prior to analysis. A standard calibration curve
can be made for known concentrations of formaldehyde and then used to determine the value of an
unknown sample.

Methanol can be oxidized by hydroxyl radicals in the Equations 4.1-4.4 where formaldehyde is

generated in Equation 4.3.

CH3OH + OH· !· CH2OH + H2O
CH3OH + OH· ! CH3O· + H2O
·CH2OH + OH· ! HCHO + H2O

HCHO + CH3O· ! HCOOCH3 + H+ + e 

(4.1)

(4.2)

(4.3)

(4.4)

The constant current density study in NH4OH with MeOH and the three combined current
density studies in solutions containing MeOH were all investigated for their formaldehyde content
at the end of 12 hours. A calibration curve was generated for formaldehyde concentrations of 1 to
8 ppm, shown in Figure 4.27.

114

Figure 4.27: Calibration curve for formaldehyde detection

The formaldehyde concentrations were 0 ppm, 134 ppm, 104 ppm, 57 ppm, and 0 ppm for
NH4OH with MeOH at 50 mA/cm2, NH4OH with MeOH combined current density, NaCl with
MeOH, NaOH with MeOH, and NH4OH with NH4Cl, respectively.

The oxidation of chloride to perchlorate is another oxidation byproduct reaction that is un-
desirable. Perchlorate generation was analyzed using ion chromatography measurements for the
two solutions that contained chloride species, NaCl with MeOH and NH4OH with NH4Cl. Each
solution was diluted by 200x and 500x and an average perchlorate concentration was taken as an
estimated value for each 12 hour sample. Both solutions had generated 600-700 ppm of perchlorate
in 12 hours. This concentration is substantial and well above the water quality standard for per-
chlorate. However, there are existing remediation methods that could follow the electrochemical
treatment to remove this undesirable product. [45] If the tests in each solution were only ran for 1
h based on achieving the most degradation for both PFOA and PFOS in that time, the byproduct
formation may also be signiﬁcantly reduced and will need to be further evaluated.

4.4.3 Energy Usage and Cost Comparison

The overall energy for each PFOA constant current density study and combined current density
study of 50&5 was determined using Equation (4.5).

115

Q =

(4.5)

A ⇤ J

v π Edt

Where Q is the total energy in kW*h/m3, J is the applied current density in A/cm2, A is the
active anodic surface area of the anode in cm2, v is the volume of the solution in m3, and E is the
potential of the system in V. The energy usage for 50&5 was calculated as the sum of two integrals.
One pertaining to the ﬁrst time frame with the higher current density and one pertaining to the
time with the applied lower current density. The energy value for each simulated sodium sulfate
solution study to reach 75% deﬂuorination are provided in Table 4.5. In terms of comparing the
constant current densities to the combined current density, we see that 50&5 has a lower energy
consumption for 75% deﬂuorination and at 1 h compared to 50 mA/cm2. In addition to consuming
similar energy than the process at 25 mA/cm2, 50&5 also only required an additional 0.1 h time.
The time to reach 75% deﬂuorination was determined by extrapolating the percentage of PFOA
deﬂuorination over time graph for 1, 5, and 10 mA/cm2 on a log scale as they did not reach it in
under 1 hour. These graphs are included as Figure 4.28. It is also noted that these measurements
should be taken as a rough estimate.

Table 4.5: Energy requirements of each PFOA study to reach 75% deﬂuorination or at 1 hour

For the 12 hour simulated IX regenerate solutions, the energy was only determined for the two
ammonia based solutions as they degraded both PFOA and PFOS successfully. Since one solution
obtained the most degradation of PFOA in the ﬁrst hour and no additional degradation post 1h and
the other obtained the most degradation of PFOS in the ﬁrst hour and no additional decomposition

116

Figure 4.28: Extrapolation plots used to determine the time to 75% deﬂuorination for (A) 1
mA/cm2, (B) 5 mA/cm2, and (C) 10 mA/cm2

after, the 1 h marked was used as the energy comparison time. Since this is the length of the high
current density application, both solutions had the same current applied for the duration of this
time. However, the two solutions had di erent voltage values. In fact the voltage integral over time
for the NH4OH and MeOH was roughly double that of the NH4OH and NH4Cl solution leading to

117

an energy consumption of 106 kW*h/m3 compared to 53 kW*h/m3.

To further compare the four IX solutions, the electric energy per order has been determined.
The electric energy per order (EEO) is the total energy normalized per log removal of contaminant,
Equation 4.6. [46] The initial and ﬁnal contaminant concentrations are based on the total number
of PFAS chains in each solution, not just PFOA or PFOS.

EEO =

(4.6)

Q

log(Ci/C f )

Where Ci is the initial total PFAS concentration present in the solution and Cf is the ﬁnal total
PFAS concentration present in the solution. The EEO values at the 12 hour time point for the
NH4OH with MeOH, NaOH with MeOH, NaCl with MeOH, and NH4OH with NH4Cl were 362,
569, 2032, and 84 kW*h/m3, respectively. A bar graph for a visual comparison of these values is
provided as Figure 4.29. This conﬁrms that the mixed ammonia solution required the lowest energy
per log removal of total PFAS compared to the other three solutions. It is also interesting to note
that the NaCl and MeOH solution required the most energy per log removal of total PFAS.

Figure 4.29: Bar graph comparing the conversion e ciency values for the simulated IX
regenerate solutions

4.5 Key Advancements from Bench Top PFAS Studies

The key advancement from the bench top PFAS studies include the following:

• Step down current density techniques do save on energy without inhibiting PFOA degradation

118

in simulated solutions with the combination of 50&5 showing improved energy consumption
compared to 50 mA/cm2

• BDD can become ﬂuorinated from the treatment of PFAS, but this has yet to inhibit its ability

to remediate PFAS in these studies

• PFOA and PFOS can be broken down in di erent complex ion exchange regenerate solutions

• The optimum regeneration solutions to be treated with electrochemical oxidation are those

that are ammonia based

• Byproducts are generated from ion exchange regenerate solution treatments that will need to

be investigated further

4.6 Summary

Investigation into electrochemical oxidation of PFAS began at the bench scale in a beaker less
than 1 L. Initial studies that utilized constant current densities showed that there may be a way
to take advantage of the step down current density technique for these compounds even though
they do not ﬁt the traditional model. Further investigations in simulated solutions showed that
one step down current density method of 50&5 did indeed provide energy savings over a constant
current density method. Parameters were slightly adjusted when adapting this procedure to complex
simulated ion exchange regenerate solutions, showing that in the ﬁrst hour of high current density
application, most solutions reached a mass transport limit for PFOA or PFOS. Although there were
a few exceptions to this rule. Due to the generation of unwanted byproducts, further investigation
into limiting their development should be completed.

119

BIBLIOGRAPHY

120

BIBLIOGRAPHY

[1] Mohammad Feisal Rahman, Sigrid Peldszus, and William B Anderson. Behaviour and fate of
perﬂuoroalkyl and polyﬂuoroalkyl substances (pfass) in drinking water treatment: a review.
Water research, 50:318–340, 2014.

[2] Magali Houde, Jonathan W Martin, Robert J Letcher, Keith R Solomon, and Derek CG
Muir. Biological monitoring of polyﬂuoroalkyl substances: a review. Environmental science
technology, 40(11):3463–3473, 2006.

[3] Nancy Merino, Yan Qu, Rula A Deeb, Elisabeth L Hawley, Michael R Ho mann, and
Shaily Mahendra. Degradation and removal methods for perﬂuoroalkyl and polyﬂuoroalkyl
substances in water. Environmental Engineering Science, 33(9):615–649, 2016.

[4] Katarzyna H Kucharzyk, Ramona Darlington, Mark Benotti, Rula Deeb, and Elisabeth Haw-
ley. Novel treatment technologies for pfas compounds: A critical review. Journal of Environ-
mental Management, 2017.

[5] Kimberly E Carter and James Farrell. Oxidative destruction of perﬂuorooctane sulfonate using
boron-doped diamond ﬁlm electrodes. Environmental science technology, 42(16):6111–6115,
2008.

[6] C Fang, M Megharaj, and R Naidu. Electrochemical advanced oxidation processes (eaop)
to degrade per-and polyﬂuoroalkyl substances (pfass). Journal of Advanced Oxidation Tech-
nologies, 20(2), 2017.

[7] Laurel A Schaider, Simona A Balan, Arlene Blum, David Q Andrews, Mark J Strynar,
Margaret E Dickinson, David M Lunderberg, Johnsie R Lang, and Graham F Peaslee. Flu-
orinated compounds in us fast food packaging. Environmental Science Technology Letters,
4(3):105–111, 2017.

[8] Victor Andres Arias Espana, Megharaj Mallavarapu, and Ravi Naidu. Treatment technologies
for aqueous perﬂuorooctanesulfonate (pfos) and perﬂuorooctanoate (pfoa): A critical review
with an emphasis on ﬁeld testing. Environmental Technology Innovation, 4:168–181, 2015.
[9] Marek Trojanowicz, Anna Bojanowska-Czajka, Iwona Bartosiewicz, and Krzysztof Kulisa.
Advanced oxidation/reduction processes treatment for aqueous perﬂuorooctanoate (pfoa) and
perﬂuorooctanesulfonate (pfos)–a review of recent advances. Chemical Engineering Journal,
336:170–199, 2018.

[10] Alessandro Zaggia, Lino Conte, Luigi Falletti, Massimo Fant, and Andrea Chiorboli. Use
of strong anion exchange resins for the removal of perﬂuoroalkylated substances from con-
taminated drinking water in batch and continuous pilot plants. Water research, 91:137–146,
2016.

121

[11] Lino Contea, Luigi Fallettia, Alessandro Zaggiaa, and Marco Milanb. Polyﬂuorinated or-
ganic micropollutants removal from water by ion exchange and adsorption. CHEMICAL
ENGINEERING, 43, 2015.

[12] Steve Woodard, John Berry, and Brandon Newman. Ion exchange resin for pfas removal and

pilot test comparison to gac. Remediation Journal, 27(3):19–27, 2017.

[13] Pattarawan Chularueangaksorn, Shuhei Tanaka, Shigeo Fujii, and Chinagarn Kunacheva. Re-
generation and reusability of anion exchange resin used in perﬂuorooctane sulfonate removal
by batch experiments. Journal of Applied Polymer Science, 130(2):884–890, 2013.

[14] Shubo Deng, Qiang Yu, Jun Huang, and Gang Yu. Removal of perﬂuorooctane sulfonate
from wastewater by anion exchange resins: e ects of resin properties and solution chemistry.
Water Research, 44(18):5188–5195, 2010.

[15] Ziwen Du, Shubo Deng, Youguang Chen, Bin Wang, Jun Huang, Yujue Wang, and Gang Yu.
Removal of perﬂuorinated carboxylates from washing wastewater of perﬂuorooctanesulfonyl
ﬂuoride using activated carbons and resins. Journal of hazardous materials, 286:136–143,
2015.

[16] Gunnar R Stratton, Fei Dai, Christopher L Bellona, Thomas M Holsen, Eric RV Dickenson,
and Selma Mededovic Thagard. Plasma-based water treatment: e cient transformation of
perﬂuoroalkyl substances in prepared solutions and contaminated groundwater. Environmen-
tal science technology, 51(3):1643–1648, 2017.

[17] Raj Kamal Singh, Sujan Fernando, Sadjad Fakouri Baygi, Nicholas Multari, Selma Mede-
dovic Thagard, and Thomas M Holsen. Breakdown products from perﬂuorinated alkyl sub-
stances (pfas) degradation in a plasma-based water treatment process. Environmental science
technology, 2019.

[18] Ch Comninellis and C Pulgarin. Electrochemical oxidation of phenol for wastewater treatment

using sno 2, anodes. Journal of Applied Electrochemistry, 23(2):108–112, 1993.

[19] Beatriz Gomez-Ruiz, Sonia Gómez-Lavín, Nazely Diban, Virginie Boiteux, Adeline Colin,
Xavier Dauchy, and Ane Urtiaga. E cient electrochemical degradation of poly-and perﬂu-
oroalkyl substances (pfass) from the e uents of an industrial wastewater treatment plant.
Chemical Engineering Journal, 322:196–204, 2017.

[20] Beatriz Gomez-Ruiz, Nazely Diban, and Ane Urtiaga. Comparison of microcrystalline and
ultrananocrystalline boron doped diamond anodes: Inﬂuence on perﬂuorooctanoic acid elec-
trolysis. Separation and Puriﬁcation Technology, 2018.

[21] Zhaohui Liao and James Farrell. Electrochemical oxidation of perﬂuorobutane sulfonate using
boron-doped diamond ﬁlm electrodes. Journal of applied electrochemistry, 39(10):1993–
1999, 2009.

[22] Tsuyoshi Ochiai, Yuichi Iizuka, Kazuya Nakata, Taketoshi Murakami, Donald A Tryk, Akira
Fujishima, Yoshihiro Koide, and Yuko Morito. E cient electrochemical decomposition of

122

perﬂuorocarboxylic acids by the use of a boron-doped diamond electrode. Diamond and
Related Materials, 20(2):64–67, 2011.

[23] Tsuyoshi Ochiai, Hirofumi Moriyama, Kazuya Nakata, Taketoshi Murakami, Yoshihiro Koide,
and Akira Fujishima. Electrochemical and photocatalytic decomposition of perﬂuorooctanoic
acid with a hybrid reactor using a boron-doped diamond electrode and tio2 photocatalyst.
Chemistry Letters, 40(7):682–683, 2011.

[24] Charles E Schaefer, Christina Andaya, Aniela Burant, Charles W Condee, Ana Urtiaga, Timo-
thy J Strathmann, and Christopher P Higgins. Electrochemical treatment of perﬂuorooctanoic
acid and perﬂuorooctane sulfonate: Insights into mechanisms and application to groundwater
treatment. Chemical Engineering Journal, 317:424–432, 2017.

[25] Álvaro Soriano, Daniel Gorri, and Ane Urtiaga. E cient treatment of perﬂuorohexanoic
acid by nanoﬁltration followed by electrochemical degradation of the nf concentrate. Water
Research, 112:147–156, 2017.

[26] AM Trautmann, H Schell, KR Schmidt, K-M Mangold, and A Tiehm. Electrochemical
degradation of perﬂuoroalkyl and polyﬂuoroalkyl substances (pfass) in groundwater. Water
Science and Technology, 71(10):1569–1575, 2015.

[27] Ane Urtiaga, Carolina Fernández-González, Sonia Gómez-Lavín, and Inmaculada Ortiz. Ki-
netics of the electrochemical mineralization of perﬂuorooctanoic acid on ultrananocrystalline
boron doped conductive diamond electrodes. Chemosphere, 129:20–26, 2015.

[28] Hanshuang Xiao, Baoying Lv, Guohua Zhao, Yujing Wang, Mingfang Li, and Dongming
Li. Hydrothermally enhanced electrochemical oxidation of high concentration refractory
perﬂuorooctanoic acid. The Journal of Physical Chemistry A, 115(47):13836–13841, 2011.
[29] Qiongfang Zhuo, Shubo Deng, Bo Yang, Jun Huang, Bin Wang, Tingting Zhang, and Gang
Yu. Degradation of perﬂuorinated compounds on a boron-doped diamond electrode. Elec-
trochimica Acta, 77:17–22, 2012.

[30] Brian P Chaplin. Critical review of electrochemical advanced oxidation processes for water
treatment applications. Environmental Science: Processes Impacts, 16(6):1182–1203, 2014.
[31] Junfeng Niu, Yang Li, Enxiang Shang, Zesheng Xu, and Jinzi Liu. Electrochemical oxidation

of perﬂuorinated compounds in water. Chemosphere, 146:526–538, 2016.

[32] Juyuan Jiang, Ming Chang, and Peng Pan. Simultaneous hydrogen production and elec-
trochemical oxidation of organics using boron-doped diamond electrodes. Environmental
science technology, 42(8):3059–3063, 2008.

[33] Eleni Bitziou, Maxim B Joseph, Tania L Read, Nicola Palmer, Tim Mollart, Mark E Newton,
and Julie V Macpherson.
In situ optimization of ph for parts-per-billion electrochemical
detection of dissolved hydrogen sulﬁde using boron doped diamond ﬂow electrodes. Analytical
chemistry, 86(21):10834–10840, 2014.

123

[34] Bo Guan, Jinfang Zhi, Xintong Zhang, Taketoshi Murakami, and Akira Fujishima. Electro-
chemical route for ﬂuorinated modiﬁcation of boron-doped diamond surface with perﬂuo-
rooctanoic acid. Electrochemistry Communications, 9(12):2817–2821, 2007.

[35] Pralay Gayen and Brian P Chaplin. Fluorination of boron-doped diamond ﬁlm electrodes for

minimization of perchlorate formation. ACS Applied Materials Interfaces, 2017.

[36] Sergio Ferro and Achille De Battisti. The 5-v window of polarizability of ﬂuorinated diamond

electrodes in aqueous solutions. Analytical chemistry, 75(24):7040–7042, 2003.

[37] G Sine, L Ouattara, M Panizza, and Ch Comninellis. Electrochemical behavior of ﬂuorinated

boron-doped diamond. Electrochemical and solid-state letters, 6(9):D9–D11, 2003.

[38] Sergio Ferro and Achille De Battisti. Physicochemical properties of ﬂuorinated diamond

electrodes. The Journal of Physical Chemistry B, 107(31):7567–7573, 2003.

[39] Shangtao Liang, Randall" David" Pierce Jr, Hui Lin, Sheau-Yun Chiang, and Qingguo" Jack"
Huang. Electrochemical oxidation of pfoa and pfos in concentrated waste streams. Remedia-
tion Journal, 28(2):127–134, 2018.

[40] Hui Lin, Junfeng Niu, Shiyuan Ding, and Lilan Zhang. Electrochemical degradation of
perﬂuorooctanoic acid (pfoa) by ti/sno 2–sb, ti/sno 2–sb/pbo 2 and ti/sno 2–sb/mno 2 anodes.
Water research, 46(7):2281–2289, 2012.

[41] Qiongfang Zhuo, Xiang Li, Feng Yan, Bo Yang, Shubo Deng, Jun Huang, and Gang Yu.
Electrochemical oxidation of 1h, 1h, 2h, 2h-perﬂuorooctane sulfonic acid (6: 2 fts) on
dsa electrode: Operating parameters and mechanism. Journal of Environmental Sciences,
26(8):1733–1739, 2014.

[42] Pingping Meng, Shubo Deng, Xinyu Lu, Ziwen Du, Bin Wang, Jun Huang, Yujue Wang, Gang
Yu, and Baoshan Xing. Role of air bubbles overlooked in the adsorption of perﬂuorooctane-
sulfonate on hydrophobic carbonaceous adsorbents. Environmental science & technology,
48(23):13785–13792, 2014.

[43] Bo Yang, Jinbao Wang, Chaojin Jiang, Juying Li, Gang Yu, Shubo Deng, Shaoyou Lu, Peixin
Zhang, Caizhen Zhu, and Qiongfang Zhuo. Electrochemical mineralization of perﬂuorooctane
sulfonate by novel f and sb co-doped ti/sno2 electrode containing sn-sb interlayer. Chemical
Engineering Journal, 316:296–304, 2017.

[44] Ti any Nash. The colorimetric estimation of formaldehyde by means of the hantzsch reaction.

Biochemical Journal, 55(3):416, 1953.

[45] Asha Srinivasan and Thiruvenkatachari Viraraghavan. Perchlorate: health e ects and tech-
nologies for its removal from water resources. International Journal of Environmental Re-
search and Public Health, 6(4):1418–1442, 2009.

[46] Thi Xuan Huong Le, Holly Haﬂich, Amisha D Shah, and Brian P Chaplin. Energy-e cient
electrochemical oxidation of perﬂuoroalkyl substances using a ti4o7 reactive electrochemical
membrane anode. Environmental Science & Technology Letters, 6(8):504–510, 2019.

124

CHAPTER 5

CONCLUSION

125

5.1 Conclusion

This report outlines three important studies that advance the ﬁeld of electrochemistry. The
ﬁrst uses a boron-doped diamond (BDD) sensor that is comprised of a working, counter, and
auxiliary electrode. This sensor is used to demonstrate its ability to detect isatin while eliminating
pretreatment procedures that require additional time and additional equipment, simplifying the three
electrode set up into one plug and play sensor, and obtaining limits of detection (LOD) applicable
to biological ﬂuids. This study also demonstrated the additional beneﬁts of using a microelectrode
array (MEA) as the working electrode in the sensor due to its lower sensitivity and more repeatable
results. The second study focused on using a material that would allow for economical mass
production of sensors for heavy metal detection. Indium tin oxide (ITO) is already used in industry
for solar cells and does not require a transparent substrate for the detection of heavy metals, ideally
manganese (Mn), using cathodic stripping voltammetry (CSV). Investigating the use of heating
during deposition and post deposition annealing led to the discovery that a heated and annealed ﬁlm
had ideal properties to detect Mn in environmentally relevant concentrations due to its larger grain
sizes and increased conductivity. This led to the concluding that ITO can be an economic sensor
material for such applications. Finally, electrochemical oxidation was investigated to advance the
understanding of remediation for per- and polyﬂuoroalkyl substances (PFAS). Current remediation
techniques focus on adsorption or ﬁltration methods that produce a waste stream that can further
contribute to the contamination cycle. Using BDD and electrochemical oxidation, a combined
current density technique was evaluated using both simulated salt solutions and simulated ion
exchange regenerate solutions. Ultimately, the study showed that the two ideal regenerate solutions
currently in literature to be further treated with electrochemical oxidation are those based on
ammonia, ammonium hydroxide with methanol and the mixed solution of ammonium hydroxide
and ammonium chloride. Due to the byproducts produced from these processes, further studies
will need to be completed.

Overall, these three studies have shown that existing electrochemical methods can be advanced

126

to address problems in today’s world to beneﬁt the planet and humankind.

127

CHAPTER 6

FUTURE OUTLOOK

128

6.1 Future Outlook

6.1.1

Isatin Detection

The next steps for the detection of isatin with the 3-in-1 BDD sensors include the detection in a real
bodily ﬂuid. Since the sensors were only tested in a urine simulant it is important to study the in a
real solution to understand if there are other compounds that can inﬂuence the response of isatin.
Additionally, a process for renewing the surface of the working electrode should be investigated
to allow for improved reproducibilities from day to day studies. Ultimately, these sensors would
be used following a liquid chromatography system so a trial with a ﬂow through injection system
should also be completed.

6.1.2

Indium Tin Oxide Characteristics

Since other metals present in environmental samples have the potential to inﬂuence the response
for manganese on the HA ITO sensor, the next step in this study would be to run a set of calibration
curves with di ering levels of other metals present. Once a better understanding is obtained for how
the manganese response changes with the presence of other metals, a real environmental sample
will need to be studied. In addition, if this sensor is to be used to better understand the toxicology of
manganese in the human body, studies with a simulated biological ﬂuid will need to be performed
before allowing for the detection in a real biological solution.

6.1.3 Mineralization of Per- and Polyﬂuoroalkyl Substances

The next step in the PFAS remediation studies would be to obtain and characterize an industrial
ion exchange regenerate solution. These solutions could have co contaminants that can inﬂuence
the degradation rates of di erent PFAS chains. They could also be producing di erent byproducts
from those that are simulated. It is also important to begin to look into the byproducts that are
being generated in the o  gasses of the system. This will be critical to know when scaling this
technology for industrial use. It can also be of interest to reverse the combined current density

129

conﬁguration to look at the di erence in applying a lower current density before a higher one. This
may be beneﬁcial in precursor conversion in industrial solutions.

130