Electrochemical Characterization of Metal Catalyst Free Carbon Nanotube

Electrode and Its Application on Heavy Metal Detection

A dissertation submitted to the

Graduate School of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in the Department of Chemistry

of the College of Arts and Sciences

Wei Yue

M.S. University of Cincinnati

October 2014

Committee Chair: W. R. Heineman, Ph.D.

Abstract of dissertation

Heavy metals are claimed to be essential to health. On one hand, heavy metal pollution, such as lead and cadmium is one of the serious pollution problems in nature due to the stability of metals at contaminated sites and high toxicity to the biosphere. While some other heavy metals, such as Mn2+ and Zn2+ are required elements in human body and play important roles.

Carbon nanotubes (CNTs) have attracted scientists as a novel electrode material due to their excellent electrochemical features: large potential window, fast electron transfer rate and large surface area. The metal catalyst free carbon nanotubes (MCFCNTs) is synthesized via Carbo

Thermal Carbide Conversion method which leads to residual transition metal free in the CNTs structure. The new material shows very good results in detecting heavy metal ions, such as Pb2+,

Cd2+, Zn2+ and Mn2+.

The work described in this dissertation includes two parts: the first part includes electrochemical characterization of metal catalyst free carbon nanotubes (MCFCNTs) and heavy metal detection with stripping voltammetry; the second part extends the use of metal catalyst free carbon nanotube whiskers as electrode modifier for trace zinc detection in bovine serum matrix with the sample being treated with a double extraction procedure using dithizone in chloroform as a zinc chelating ligand.

A composite film of Nafion and MCFCNT whiskers was applied to a gold electrode surface to form a mechanically stable sensor. The sensor was then used for zinc detection in both acetate buffer solution and extracted bovine serum solution. The sensor has reproducible behavior for the anodic stripping voltammetry (ASV) measurement of zinc. A limit of detection of 53 nM was

achieved for a 120 s deposition time. The modified electrode was found to be both reliable and sensitive for zinc measurements in both matrices.

iii

Acknowledgements

First, I would like to thank my advisor Dr. William R. Heineman for all his professional guidance, patience and financial support during the past five years. You have been a great advisor as well as a role model to me both as a researcher and more importantly, as a human being. You have consciously and unconsciously influenced me in many details which make me become a better person. It is truly my honor to finish my Ph.D in Dr. Heineman’s research lab and I could not have asked for a better environment to learn in.

I would also like to acknowledge my committee members: Dr. Ayres, Dr. Ridgway for their guidance and influence professionally, as well as our collaborators Dr. Ian Papautsky, Wenjing

Kang, Xing Pei and Dr. Bange for their many insightful discussions on my research. I would also thanks my group members and Dr. Necati for their help on my research.

Finally, I would thank my family members who have supported me through this long journey.

I need to thank my wife Yan Zhang for her endless support, encouragement, and patience. We have shared ups and downs during these five years and worked hard to make it to this point. Last but not the lease, special thanks to my parents and grandparents for all the support and love they have given me.

Table of contents

ABSTRACT OF DISSERTATION…………………………………………………………….i

ACKKNOLEDGEMENTS……………………………………………………………………..iii

TABLE OF CONTENTS……………………………………………………………………….iv

LIST OF FIGURES………………………………………………………………………….viii

CHAPTER 1 | INTRODUCTION……………………………………………………………..1

1.1 INTRODUCTION…………………………………………………..………………………2 1.2 STRIPPING VOLTAMMETRY ANALYSIS……………………………………………2 1.3 CARBON NANOTUBES and ELECTROACTIVITY of CARBON NANOTUBES…..3

1.4 FABRICATION of CNT ELECTRODES………………………………………………….4

1.4.1 Direct growth of CNT electrodes ………………………………………………….………4

1.4.2 CNT modified electrodes…………………………………………………………….…….7

1.4.3 Chemically Modified CNTs electrodes……………………………………………………...8

1.5 PROJECT OBJECTIVES………………………………………………………………….9

1.6 REFERENCE…………………………………………………………………………….…11

CHAPTER 2 | ANODIC STRIPPING VOLTAMMETRY OF HEAVY METALS ON A METAL CATALYST FREE CARBON NANOTUBES ELECTRODE………………….15

2.1 INTRODUCTION………………………………………………………………………..16

2.2 EXPERIMENTAL……………………………………………………………………..17

2.2.1 Chemicals and instrumentation……………………………………………………17

2.2.2 Procedure ………………………………………………………………………..18

2.3 RESULT AND DISCUSSION……………………………………………………………..19

2.3.1 Optimization of parameters …………………………………………………………….19

2.3.1.1 Deposition potential………………………………………………………………..19

2.3.1.2 Deposition time……………………………………………………………………..20

2.3.1.3 Conditioning potential and time………………………………………………….21

2.3.2 Metal-catalyst-free carbon nanotube electrode…………………………………………22

2.3.3 Calibration data………………………..………………………………………….…..24

2.3.4 Analysis of tap water………………..………………………………………………..33

2.4 CONCLUSION…………………………………………………………………………..35

2.5 ACKNOWLEDGEMENT………………………………………………………………..36

2.6 REFERENCE……………….………………………………………………………..36

CHAPTER 3 | MANGANESE DETECTION WITH A METAL CATALYST FREE CARBON NANOTUBE ELECTRODE: ANODIC VERSUS CATHODIC STRIPPING VOLVAMMETRY……………………………………………………………………………39

3.1 INTRODUCTION……………………………………………………………………….40

3.2 EXPERIMENTAL…………………………………………………………………….42

3.3 RESULT AND DISCUSSION……………………………………………………………..42

3.3.1 Anodic Stripping Votammetry Study of Mn……………………………………………42

3.3.1.1 Cyclic Voltammetry of Mn in NH4Cl………………………………………………..43

3.3.1.2 pH and Deposition Potential Optimization………………………………………..44

3.3.1.3 Reproducibility and Cleaning Step Study ………………………………………….45

3.3.1.4 Calibration Data………………………………………………………………….46

3.3.2 CSV Study of Mn…………………………………………………………………….48

3.3.2.1 Cyclic Voltammetry of Mn in Borate Buffer………..………………………………48

3.3.2.2 pH Optimization of Buffer Solution ……………….………………………………49

3.3.2.3 Deposition Potential Optimization of CSV of MnO2…………………………………50

3.3.2.4 Reproducibility and Cleaning Step Study …………………………………………51

3.3.2.5 Calibration Data………………………………………………………………….51

3.4 NATUAL MATRIX of CSV STUDY of Mn……………………………………………..54

3.5 CONCLUSION…………………………………………………………………………….56

3.6 ACKNOWLEDGMENT…………………………………………………………………..56

3.7 REFERENCE……………………………………………………………………………….57

CHAPTER 4 | THE APPLICATION OF NAFION METAL CATALYST FREE CARBON NANOTUBE MODIFIED GOLD ELECTRODE: VOLTAMMETRIC ZINC DETECTION IN BOVINE SERUM…………………………………………………………59

4.1 INTRODUCTION………………………………………………………………………..60

4.2 EXPERIMENTAL………………………………………………………………………62

4.2.1 Reagents ……………………………………………………………………………..62

4.2.2 Apparatus…………………………………………………………………………………….63

4.2.3 Preparation of Nafion/whiskers-Au electrode………………………………………….63

4.2.4 Extraction procedures………………………………………………………………………64

4.3 RESULT AND DISCUSSION ……………………………………………………………68

4.3.1 MCFCNT whiskers……………………………………………………………..68

4.3.2 Comparison of bare gold, Nafion coated gold (Nafion-Au) and Nafion/whiskers-Au electrodes ………………………………………………………………………………..68

4.3.3 MCFCNT whiskers amount optimization……………………………………………….73

4.3.4 pH, deposition potential and deposition time optimization …………………………….75

4.3.5 Calibration data ……………………………………..…………………………………78

4.3.6 Sample preparation and standard addition of extraction from bovine serum…….……..82

4.4 CONCLUSION ………………………………………………………….………………..85

4.5 ACKNOWLEDGMENT……………………………………………………………………………85

4.6 REFERENCE………………………………………………………………………………………86

vii

CHAPTER 5 | HEAVY METAL DETECTION ON LAB-ON-A-CHIP SENSOR……….89

5.1 INTRODUCTION………………………………………………………………………90

5.2 EXPERIMENTAL……………………………………………………………………91

5.2.1 Reagents and instrumentation……………………………………………………………….91

5.3 Results and discussion…………………………………………………………………..92

5.3.1 Electron transfer rate study of the MCFCNT electrode……………………………………92

5.3.2 Cyclic voltammetry of lead, cadmium and zinc……………………………………………..93

5.3.3 ASV study of lead, cadmium and zinc……………………………………………………….96

5.3.4 Comparison of ASV results between two sets of parameters on zinc……………………105

5.4 Conclusion………………………………………………………………………………108

5.5 Acknowledgement……………………………………………………………………….108

5.6 Reference…………………………………………………………………………...……109

CHAPTER 6 | CONCLUSIONS…………………………………………………………110

viii

List of Figures

Figure 2.1 The effect of deposition potential on the stripping current of Pb2+, Cd2+ and Zn2+, concentrations were 3 µM Pb2+, 3 µM Cd2+ and 5 µM Zn2+ in 0.1M acetate buffer (pH = 4.65). Deposition time: 20 s…………………………………………………………………………20

Figure 2.2 The effect of deposition time on the stripping current of Pb 2+, Cd2+ and Zn2+, concentrations were 3 µM for Pb2+, 3 µM for Cd2+ and 5 µM for Zn2+ in 0.1M acetate buffer (pH = 4.65). Deposition potential: -1.35 V, cell volume: 15 ml. ………………………………21

Figure 2.3 Cyclic voltammograms of 0.1 M pH 7 phosphate buffer and 5 mM hydrogen peroxide.

……………………………………………………………………………………………………23 Figure 2.4A) OSWSV and calibration curve for Pb2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.3, 0.5, 1.0, 2.0, 3.0, 5.0 and 8.0 µM of Pb2+; Deposition time: 150 s, deposition potential: -1.35 V………………………………………………………………………………25

Figure 2.4B) OSWSV and calibration curve for Cd2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.5, 0.7, 1.0, 2.0, 3.0, 6.0 and 8.0 µM; Deposition time: 150 s, deposition potential: -1.35 V………………………………………………………………………………………26

Figure 2.4C) OSWSV and calibration curve for Zn2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.5, 1.0, 2.0, 3.0, 4.0, 6.0 and 8.0 µM. Deposition time: 150 s, deposition potential: -1.35 V……………………………………………………………………………………….27

Fig. 5 OSWSV of 0.1 nM Pb2+, 8 nM Cd2+ and 40 nM Zn2+ in 0.1 M acetate buffer (pH=4.65). Deposition time: 600 s, deposition potential:-1.35 V. ………………………………………29

Figure 2.6 A) OSWSV of simultaneous detection of Pb2+, Cd2+ and Zn2+, concentration 0.5, 1.0, 1.5, 1.7 and 2.5 µM. ……………………………………………………………………………30

Figure 2.6 B) Calibration curve for Cd2+ and Zn2+, concentration 0.5, 1.0, 1.5, 1.7 and 2.5 µM. Deposition time:150 s, deposition potential: -1.35 V. ………………………………………31

Figure 2.7 Trend curve of simultaneous detection of Pb2+, Cd2+ and Zn2+, concentration 0.5, 1.0, 1.5, 1.7, 2.5, 3.0 and 3.5 µM. ……………………………………………………………….32

Figure 2.8 Anodic stripping voltammograms for analysis of tap water by method of standard addition, sample was diluted 1:1 with 0.1 M acetate buffer pH 4.65. Dashed line is voltammogram of tap water; solid lines are voltammograms after spiking sample with concentrated Zn2+ to give the concentrations of 0.1 µM, 0.2 µM, 0.3 µM, 0.45 µM, 0.6 µM. Deposition time: 900 s; Deposition potential: -1.35 V………………………………………34

2+ Figure 3.1 Cyclic voltammetry of 1.5 mM Mn in 0.05 M NH4Cl on MCFCNTs electrode, scan rate 100 mV/s………………………………………………………………………………44

2+ Figure 3.2 Deposition potential optimization of 100 µM Mn in 0.05 M NH4Cl………………45

2+ Figure 3.3 A) Calibration plot for ASV of Mn in 0.05 M NH4Cl……………………………47

2+ Figure 3.3 B) Anodic stripping voltammograms of Mn in 0.05 M NH4Cl in the concentration range of 3.5 µM to 13.5 µM……………………………………………………………………48

Figure 3.4 Cyclic voltammotram of 1.5 mM Mn2+ in 0.1 M pH 8.5 borate buffer, scan rate 100 mV/s…………………………………………………………………………………………49

Figure 3.5 CSV pH optimization of Mn2+ from pH 7.5-9.5, deposition potential: 600 mV……50

Figure 3.6 CSV deposition potential optimization of Mn2+ from 600 mV to 800 mV, pH 8.5; Mn2+ concentration for A and B: 1 µM, deposition time for A and B: 60 s……………………51

Figure 3.7 A) Dynamic range of CSV measurements of Mn in the range of 0.12 µM - 12 µM………………………………………………………………..……………..52

Figure 3.7 B) shows CSV stripping voltammograms of Mn at 350 mV with increasing Mn concentrations in the linear concentration range of 0.6 - 6.7 µM. Each voltammogram represents three replicas of measurements which behave similarly ………………………………………53

Figure 3.8 Accumulation time effect on peak current from 1–20 min; deposition potential: 600 mV; pH: 8.5; Mn concentration: 0.12 µM …………………………………………………….54

Figure 3.9 Voltammograms of CSV measurements of Mn in spiked pond water sample; deposition time: 60 s; deposition potential: 600 mV; pH: 8.5…………………………………55

Figure 4.1 A) 5 mM Dithizone in chloroform; B) Deprotonated dithizone in pH 9, 1 M ammonia/0.5 M ammonium buffer solution; C) 10 ml deprotonated dithizone with solution containing Zn2+ (upper layer) + 10 ml chloroform (bottom layer); D) Mixture in C after 5 min sonication (zinc dithizone complex in the bottome layer); E) Mixture of bottom layer from D and 10 mL 1 M sulfuric acid after 5 min sonication (Zn2+ in the upper aqueous layer and dithizone in bottom chloroform layer) ………………………………………………………………………67

Figure 4.2 TEM images of MCFCNT whiskers: A) 4000X magnification; B) 200,000X magnification……………………………………………………………………………………69

Figure 4.3 ASV voltammograms of 4 µM zinc in acetate buffer solution (pH 6) on different electrodes (bare gold electrode, Nafion-Au gold electrode and Nafion/whiskers-Au electrode). Deposition time 120 s; deposition potential -1400 mV………………………………………71

Figure 4.4 SEM images of Nafion/whiskers-Au electrode: A) surface topology of Nafion/whiskers-Au electrode with a gas secondary electron detector; B) surface composition of Nafion/whiskers-Au electrode with a backscattered electron detector………………………72

Figure 4.5 EDX images of Nafion/whiskers-Au electrode: A) The bright area (horizontal arrow) in Figure 4.4 B; B) The dark area (vertical arrow) in Figure 4.4 B……………………………73

Figure 4.6 Effect of CNT concentration on peak current for zinc stripping voltammetry. Coating: 10 µL of MCFCNTs whisker dispersion in 1 wt. % Nafion ethanolic solution with different CNT concentrations. Deposition time 120s, deposition potential -1400 mV; zinc concentration 5 µM…………………………………………………………………………………….……75

Figure 4.7 A) Cyclic voltammograms of Nafion/whiskers-Au electrode in acetate buffer with different pH values; scan rate: 10 mV / s……………………………………………………77

Figure 4.7 B) pH optimization of Zn2+ on Nafion/whiskers-Au electrode in acetate buffer with optimized coating parameters from pH 4.5 to 6. Zinc concentration 5 µM; deposition potential - 1400 mV; deposition time 120 s………………………………………………………………78

Figure 4.8 Deposition potential optimization with optimized coating parameters: deposition time 120 s; zinc concentration 5 µM; pH 6.0………………………………………………………79

Figure 4.9 ASV voltammograms of Zn2+ with Nafion/whiskers-Au electrode in 0.1 M pH 6 acetate buffer in the concentration range of 0.5 µM to 7.0 µM. Deposition time 120 s; deposition potential -1400 mV. ……………………………………………………………………….80

Figure 4.10 Dynamic ranges of zinc ASV measurements in the range of 0.5 µM–11 µM on Nafion-Au electrode and Nafion/whiskers-Au electrode. Deposition time 120 s; deposition potential -1400 mV……………………………………………………………………………81

Figure 4.11A) Voltammograms of standard addition of bovine serum extract for ASV zinc detection with Nafion/whiskers-Au electrode. ……………………………………………..84

Figure 4.11B) Standard addition plot of bovine serum extract for ASV zinc detection with Nafion/whiskers-Au electrode. Deposition time 300 s; deposition potential -1400 mV; pH 6. ………………………………………………………………………………………………….85

Figure 5.1 Cyclic voltammograms of 5 mM K4Fe(CN)6 in 1 M KNO3 solution on MCFCNT electrode and glassy carbon electrode, scan rate 100 mV/s……………………………………94

Figure 5.2 A) Cyclic voltammograms of Pb2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1 mM…………………………………………………….95

xii

Figure 5.2 B) Cyclic voltammograms of Cd2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1 mM……………………………………………………..96

Figure 5.2 C) Cyclic voltammograms of Zn2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1 mM……………………………………………………97

Figure 5.3 A) Anodic stripping voltammograms of Pb2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s; Deposition potential: -1350 mV……………………………………………………………….99

Figure 5.3 B) Anodic stripping voltammograms of Cd2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s; Deposition potential: -1350 mV……………………………………………………………….100

Figure 5.3 C) Anodic stripping voltammograms of Zn2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s; Deposition potential: -1350 mV……………………………………………………………….101

Figure 5.4 A) Cyclic voltammograms of Zn2+ in 0.1 M acetate buffer at MFCNT electrode, Concentration: 1 mM, scan rate 25 mV/s………………………………………………………102

Figure 5.4 B) Anodic stripping voltammograms of Zn2+ on MFCNT electrode in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s; Deposition potential: -1350 mV…………………………………………………..103

Figure 5.5 A) Reverse currents for anodic stripping voltammograms at a CNT electrode for 10 µM Zn in 0.1 M acetate buffer (pH 4.5), deposition time: 150 s, amplitudes from 25 mV to 200 mV………………………………………………………………………………………………104

Figure 5.5 B) Reverse currents of anodic stripping voltammograms for 10 µM Zn in 0.1 M acetate buffer (pH 4.5), deposition time: 150 s; amplitude: 200 mV; step size: 5 mV; frequency from 10 Hz to 50 Hz………………………………………………………………………….105

xiii

Figure 5.5 C) Reverse currents for anodic stripping voltammograms at CNT electrode for 10 µM Zn in 0.1 M acetate buffer (pH 4.5), deposition time: 150 s; frequency: 25 mV; amplitude: 200 mV; step size: 5 mV, 8 mV, 10 mV, 12 mV, 15 mV………………………………………….106

Figure 5.6 A) Voltammograms of ASV of zinc in the range from 2 µM to 7 µM in 0.1 M acetate buffer (pH 4.5) with default ASV parameters………………………………………………….107

Figure 5.6B Voltammograms of ASV of zinc in the range from 2 µM to 7 µM in 0.1 M acetate buffer (pH 4.5) with optimized ASV parameters; deposition time: 60 s………………………108

xiv

Chapter 1| Introduction

1.1 Introduction

There are many adverse health effects associated with exposure to heavy metals and abnormal concentrations of metals in the body can be linked to various diseases. A wide range of processes, including smelting, heavy industry, and consumer electronic waste, are responsible for the release of dangerous metals into the environment, and the metals are transported through the biosphere and accumulate in various matrices. Heavy metal pollution is different from pollution by many organic species in that metals do not decay into harmless compounds with time. Some of the metals that pose health risks are necessary for humans at trace levels (zinc, copper, manganese, nickel) while others are toxic (cadmium, mercury, arsenic, lead). In addition, some toxic semi-metals are also considered to be as toxic as heavy metals1. Due to the extensive industrial and agricultural use of heavy metals and the hazardous effect to both human and other organisms, the development of sensitive monitoring methods has become increasingly crucial 2, 3.

1.2 Stripping Voltammetry Analysis

Stripping voltammetry is a versatile analytical technique that has been used for a wide variety of applications. Multiple analytes can be measured simultaneously in various matrices down to the nanomolar level with stripping volotammetry. The stripping voltammetry technique consists of two major steps: accumulation and stripping. During the accumulation step, the analyte is reduced (Mn+ to M0) with an applied negative potential and deposited at the working electrode surface. This allows for analyte pre-concentration and provides the ability to obtain very low limits of detection. In most cases only a small fraction of the total analyte in the sample is deposited because complete deposition of all the analyte would be time-consuming and is usually unnecessary to reach the required limit of detection. Since the deposition is not exhaustive, it is

important to control the amount of analyte deposited. To do this the electrode surface area, deposition time, and rate of convective mass transport must be carefully optimized and controlled throughout the experiment. Deposition time can vary widely depending on the analyte concentration, the type of electrode, and the stripping technique used. The stripping signal is proportional to the total amount of analyte deposited, so less concentrated solutions require longer deposition times in order to give adequate stripping peaks. Following the deposition step, the electroplated metal is oxidized (M0 to Mn+) as the potential of the working electrode is scanned in the positive direction and the anodic current is measured.

1.3 Carbon Nanotubes and Electroactivity of Carbon Nanotubes

Historically, mercury-based electrodes have been used more than any other type for stripping voltammetry. The typical electrodes used have traditionally been the hanging mercury drop electrode (HMDE) and mercury film electrode (MFE). Mercury-based electrodes have been used for years due to their reproducible electrode surface, good negative potential window, and overall excellent performance with stripping voltammetry4. Due to their toxicity, however, mercury-based electrodes are no longer a suitable choice for some applications. Many alternatives are being explored in the search for a replacement5-8. Since they were first discovered9, carbon nanotubes (CNTs) have been attracting increased attention due to their excellent electrochemical properties, including wide potential window, fast electron transfer rate and large surface area10-12. Applications with CNTs have a wide range of use, including in biomedicine, nanoelectronics, environmental engineering and electrochemical sensing1. In this chapter, I critically focus on a review of applications of different forms of CNTs for heavy metal detection by stripping voltammetry from the past five years (2008-present).

For more than 20 years, CNTs have attracted attention worldwide as a novel electrode material9. Generally, CNTs have two classes: single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs). SWCNTs consist of a simple geometry with a rolled single graphene sheet which has a diameter in the range of 0.4-2 nm and up to 20 cm in length13.

MWCNTs consist of several concentric tubes fitted into each other with a diameter up to 100 nm14. It has been confirmed that, in most of the cases, the electroactive sites on CNTs are at the open ends of CNTs, which are edge-plane like sites, and defects on the side walls of CNTs15, 16.

In some cases, oxygenated species and residual metal oxide can result in electrocatalytical behavior on CNTs17, 18.

1.4 Fabrication of CNT electrodes

1.4.1 Direct growth of CNT electrodes

In general, CNTs have two sub-categories, SWCNTs and MWCNTs. SWCNTs are formed by rolling a single layer of graphene, whereas MWCNTs consist of several graphene tubes fitted one inside of the other19. Depending on the desired properties of the CNTs, different methods have been proposed and developed to synthesize CNTs. There are 3 main methods for the synthesis of CNTs: arc discharge, laser ablation, and chemical vapor deposition (CVD)20. Arc discharge and laser ablation belong to the high temperature (above 3000 K) and short time (μs- ms) synthesis techniques20. Prasek et al. has provided a thorough discussion of the synthesis of

MWCNTs and SWCNTs using arc discharge and laser ablation19. Although arc discharge and laser ablation produce high quality nanotubes with superior straightness and crystallinity, the major disadvantage of using those two techniques is the requirement of large amounts of energy and solid carbon/graphite as substrate, which limits their application in large scale production20,

21. As compared with this higher temperature synthesis technique, lower temperature synthesis methods, such as CVD provide better control of CNT orientation, nanotube length, diameter, alignment, purity and yield, making CVD the most prevalent technique for CNT growth nowadays19. Also, CNTs can be synthesized using various substrates and the reaction rate of

CNT formation and site density can be controlled with the application of electron-beam lithography22. In the CVD process, CNTs are formed by the decomposition of hydrocarbon precursors with continuous flow of gas over patterned metal catalyst sites23. Most commonly used CNT precursors are methane, ethylene, benzene, xylene and carbon monoxide. Popularly used metal catalysts are transition metals (Fe, Co, Ni) and some less commonly used metals (Cu,

Pt, Pd, Mn, Mo, Cr, Sn, Au, Mg, Al) have also been reported for the CVD process. Two previous reviews have comprehensively discussed the proper selection of precursors and catalysts for

CVD22, 23. CNTs have been grown on iron patterned silicon substrate with ethylene with controllable length by Fan24 and have been grown on nickel patterned glass substrate with acetylene by Ren25. During the growth, neighboring CNTs interact with each other via Van der

Waals forces to form high density CNT bundles with high density and rigidity. In general, there are two steps for catalytic CNT formation. First, a meta-stable carbide is formed at the catalyst surface, resulting in the formation of an amorphous carbon rod in a quick step. This is followed by graphitization of the carbon rod which allows the slow formation of the hollow tubes from rods and the termination of growth26.

High density CNT arrays and forests have fast electron transfer rate as well as electrocatalytic features. However, they also show tremendous background current which are caused by the CNT sidewalls since the CNTs do not act as individual electrodes, but instead as clusters with the diffusion layers of single CNTs overlaping17, 27. To make each CNT work as an

individual electrode and to overcome the overlapping of diffusion layers from neighboring CNTs, the spacing needs to be larger than the CNT diameter28.

The pursuit of improvements in the field of CNT fabrication led to the discovery of the non- metal catalyzed system. Liu et al. has reported that a non-metallic substrate (SiO2 particles) acts as a catalyst for CNT growth using CVD of ethanol without catalysts. The annealing of SiO2 in hydrogen at high temperature leads to the formation of defects at SiO2 surfaces, which provide nucleation sites for the CNT growth29.

Another process used to produce aligned CNTs in a non-metallic catalytic system is the sublimation and decomposition of SiC at high temperature in a vacuum, which was first reported in 1997 by Kusunok30. Recently, CNT derivative solid carbon nanorods (SCNRs) fabricated by

Carbo-Thermal Carbide Conversion (CTCC) has been reported31. CTCC is a chemical process for producing carbon nanostructures from solid phase carbide source materials. The commonly used substrate is carbide ceramics, such as SiC in the states of single crystal, polycrystalline or amorphous. In the CTCC process, the metal or metalloid impurities from a carbide material react with reducing gases (H2O, CxOy, air or a mixture of these gases) and form gaseous compounds at a high processing temperature. As a result, the metal/metalloid species can be selectively removed leaving only carbon species. In the CTCC process, the metal or metalloid species from a crystalline carbide material is selectively removed with only carbon species remaining. CTCC- grown materials are also referred to as Solid Carbon Nanorods (SCNRs) since they do not have the traditional hollow core but instead have a solid core structure inside. SCNRs are similar in appearance, physical and chemical properties to CVD grown CNTs, with concentric spacing on the order of interplanar spacing of graphite, approximately 0.4 nm32. The spectroscopic and electrochemical properties of single-walled carbon nanotube material produced by the CTCC

process have been characterized33. Metal catalyst free CNTs (MCFCNTs) synthesized via a solid-phase growth mechanism have also been reported, and have shown excellent performance in voltammetric heavy metal detection34, 35. Also, a recent application of MWCNT fabricated by the CTCC process was reported to construct a europium sensor36.

1.4.2 CNT modified electrodes

The main challenge for developing CNT modified electrodes is that CNTs do not dissolve or disperse well in most of the common organic and inorganic solvents since the high Van der

Waals force causes CNTs to aggregate into bundles37, 38. Several approaches such as direct dispersion of CNTs and polymer-assisted and surfactant-assisted dispersion have been explored to fabricate CNT based sensors. Among all of the modification techniques, drop casting of a

CNT dispersion is favored and widely used for electrode surface modification because of its simplicity and residual solvent-free feature. In general, the CNT dispersion is evenly cast on the surface of the electrode and followed up by evaporation of solvent.

Surfactants can successfully solubilize CNTs in aqueous solution by sheathing the CNT surface with hemimicelles39, 40. The alkyl groups of surfactants lie on the CNT surface parallel to the tube axis. In addition, surfactants with benzene rings are proved to have better dispersing ability due to the π-π stacking interaction between surfactant and CNT, which can increase the binding surface significantly39. CNT dissolution in water has great significance for electroanalytical applications. Dodecyl benzene sulfonate (NaDBS) and dihexadecyl phosphate have been used for CNTs dispersion and heavy metal detection41, 42. Another approach to increase CNT dispersion in water is to use high molecular weight polymer to wrap around the surface of CNTs and therefore to solve CNT aggregation by disrupting the Van der Waals force

without disrupting π bonds in CNTs38. For instance, due to its advantages of thermal stability, chemical inertness, mechanical strength and resistance to fouling, Nafion has been widely used to disperse CNTs and modify electrodes in electrochemistry43, 44. Another simple approach that has been widely used is to mix CNTs with binders, such as Teflon or mineral oil, to form composites for CNT based sensor fabrication45, 46.

1.4.3 Chemically Modified CNTs electrodes

In some cases, CNTs are chemically modified prior to the sensor fabrication process. Since the edge plane-like end of CNTs is more reactive than the inert sidewalls, most chemical modification happens on the open ends of CNTs. Soluble SWCNTs have been obtained by derivatizing SWCNTs with thionychloride and octadecylamine which resulted in substantially improved solubilityof SWCNTs in organic solvents47. With respect to heavy metal detection with chemically modified CNT electrodes, in general, CNTs are first washed with concentrated strong acid to generate carboxyl group at the open ends48, 49. Then, the open ends with the carboxyl groups are chemically modified with certain classes of small molecules that have strong affinities to heavy metals. Cysteine49, polyaniline (PANI)50, hydroxyapatite (HAP,

51 52 Ca10(PO4)6(OH)2) and chitosan have been immobilized on CNTs through interaction with carboxyl groups at the open end for heavy metal detections.

In addition, not only the open ends but the sidewalls of CNTs can also be chemically modified for heavy metal sensing purpose. Plasma treatment can strongly interact with CNTs by breaking the sp2 hybridized C=C bonds in the CNT lattice and generate functional groups during the plasma treatment53, 54. Amino groups can be directly introduced onto the CNT surface by

NH3/Ar plasma treatment without destroying the CNT structure. The amino group serves as a

heavy metal trap through acid-base pairing interaction between electron-rich amino ligands and

electron-deficient heavy metal ions55, 56.

1.5 Project Objectives

This dissertation focuses on two main objectives: 1) characterize metal catalyst free carbon

nanotube electrodes with electrochemical methods for heavy metal detection and 2) develop a

method for zinc detection with a lab-on-a-chip sensor using anodic stripping voltammetry in

biological matrices such as bovine serum.

Table 1. Representatives of CNT/CNT modified electrodes for heavy metal detection with

stripping voltammetry

Electrode composition Modifier Target Detection LOD Linear range Refs

analytes mode

MWCNT-GC 5-Br-PADAP Pb2+ ASV 0.1 ug/L 0.9-114.6 ug/L 57

MWCNT(Pt, Fe nano-particles)-GC Nafion As3+ ASV 10 nM Up to 0.3 uM 58

CNPE CTS-ECH, Cu2+ ASV 10 nM 79 nM-16uM 59

mineral oil

MWCNT-CPE EHPO Ag+ DPASV 0.08 ng/ml 0.5-235ng/ml 60

MWCNT-CPE AV Mo5+ ASV 0.1 nM 0.4-100 nM; 61

0.2-8 uM

SWCNT-Au PhSH Hg2+ SWASV 3 nM 5-90 nM 62

MWCNT-GC PAN Cd2+ DPASV 0.1 ug/L 0.8-220.4 ug/L 63

MWCNT/Bi film-GC ABTS Cd2+ DPASV 0.2 ug/L; 0.1 0.5 -35 ug/L; 64 ug/L Pb2+ 0.2 -50 ug/L

MWCNT - Pb2+ OSWSV 13 nM, 32 nM, 0.3 -8 uM; 35 50 nM Cd2+ 0.5 -5 uM;

Zn2+ 0.5 –7 uM

2+ 57 MWCNT(-NH3)-GC - Cd ASV 0.0272 nM, 2.5 -25 nM 0.314 nM, 0.226 Zn2+ 0.2 – 2.8 uM nM 0.144 nM

Cu2+ 0.2 – 2.8 uM

Hg2+ 0.02 – 0.6 uM

DPASV: Differential Pulse Anodic Stripping Voltammetry; SWASV: Square Wave Anodic Stripping

Voltammetry; OSWSV: Osteryoung Square Wave Stripping Voltammetry

1.6 Reference

1. A. K. Wanekaya, Analyst 2011, 136, 4383-4391.

2. Pipat Chooto, P. Wararatananurak, C. Innuphat, Science Asia 2010, 36, 150-156.

3. M. Rajabi, A. Asghari, H. Zavvar Mousavi, J. Anal. Chem. 2009, 65, 511-517.

4. J. Wang, Practical Considerations. Analytical Electochemistry, John Wiley & Sons, Inc: 2002; p

115-163.

5. J. Wang, B. Tian, Anal. Chem. 1993, 65, 1529-1532.

6. E. P. Achterberg, C. Braungardt, Anal. Chim. Acta 1999, 400, 381-397.

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

8. J. J. Hatch, A. A. Gewirth, J. Electrochem. Soc. 2009, 156, D497-D502.

9. S. Iijima, Nature 1991, 354, 56-58.

10. R. H. Baughman, Science 2000, 290, 1310-1311.

11. G. Dukovic, M. Balaz, P. Doak, N. D. Berova, M. Zheng, R. S. McLean, L. E. Brus, J. Am. Chem.

Soc. 2006, 128, 9004-9005.

12. B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, L. G. Bachas, Science 2004, 303, 62-

65.

13. H. W. Zhu, C. L. Xu, D. H. Wu, B. Q. Wei, R. Vajtai, P. M. Ajayan, Science 2002, 296, 884-886.

14. M.-F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, R. S. Ruoff, Science 2000, 287, 637-640.

15. K. Jurkschat, S. J. Wilkins, C. J. Salter, H. C. Leventis, G. G. Wildgoose, L. Jiang, T. G. J. Jones,

A. Crossley, R. G. Compton, Small 2006, 2, 95-98.

16. C. E. Banks, R. R. Moore, T. J. Davies, R. G. Compton, Chemical Communications 2004, 1804-

1805.

17. J. J. Gooding, R. Wibowo, Liu, W. Yang, D. Losic, S. Orbons, F. J. Mearns, J. G. Shapter, D. B.

Hibbert, Journal of the American Chemical Society 2003, 125, 9006-9007.

18. M. Pumera, H. Iwai, Chemistry – An Asian Journal 2009, 4, 554-560.

19. J. Prasek, J. Drbohlavova, J. Chomoucka, J. Hubalek, O. Jasek, V. Adam, R. Kizek, Journal of

Materials Chemistry 2011, 21, 15872-15884.

20. K. B. Teo, C. Singh, M. Chhowalla, W. I. Milne, Encyclopedia of nanoscience and nanotechnology 2003, 10.

21. M. M. A. Rafique, J. Iqbal, Journal of Encapsulation and Adsorption Sciences 2011, 1, 29.

22. M. Kumar, Y. Ando, Journal of Nanoscience and Nanotechnology 2010, 10, 3739-3758.

23. A. Szabo, C. Perri, A. Csato, G. Giordano, D. Vuono, J. B. Nagy, Materials 2010, 3, 3092-3140.

24. S. Fan, M. G. Chapline, N. R. Franklin, T. W. Tombler, A. M. Cassell, H. Dai, Science 1999, 283,

512-514.

25. Z. F. Ren, Z. P. Huang, J. W. Xu, J. H. Wang, P. Bush, M. P. Siegal, P. N. Provencio, Science

1998, 282, 1105-1107.

26. A. Yasuda, N. Kawase, F. Banhart, W. Mizutani, T. Shimizu, H. Tokumoto, The Journal of

Physical Chemistry B 2002, 106, 1849-1852.

27. X. Yu, D. Chattopadhyay, I. Galeska, F. Papadimitrakopoulos, J. F. Rusling, Electrochemistry

Communications 2003, 5, 408-411.

28. Y. Tu, Y. Lin, Z. F. Ren, Nano Letters 2002, 3, 107-109.

29. H. Liu, D. Takagi, S. Chiashi, Y. Homma, Carbon 2010, 48, 114-122.

30. M. Kusunoki, M. Rokkaku, T. Suzuki, Applied Physics Letters 1997, 71, 2620-2622.

31. C. P. Jones, K. Jurkschat, A. Crossley, R. G. Compton, B. L. Riehl, C. E. Banks, Langmuir 2007,

23, 9501-9504.

32. Z. K. Tang, L. Y. Zhang, N. Wang, X. X. Zhang, J. N. Wang, G. D. Li, Z. M. Li, G. H. Wen, C. T.

Chan, P. Sheng, Synthetic Metals 2003, 133–134, 689-693.

33. T. Wang, H. D. Manamperi, W. Yue, B. L. Riehl, B. D. Riehl, J. M. Johnson, W. R. Heineman,

Electroanalysis 2013, 25, 983-990.

34. W. Yue, A. Bange, B. L. Riehl, B. D. Riehl, J. M. Johnson, I. Papautsky, W. R. Heineman,

Electroanalysis 2012, 24, 1909-1914.

35. W. Yue, B. L. Riehl, N. Pantelic, K. T. Schlueter, J. M. Johnson, R. A. Wilson, X. Guo, E. E.

King, W. R. Heineman, Electroanalysis 2012, 24, 1039-1046.

36. T. Wang, D. Zhao, X. Guo, J. Correa, B. L. Riehl, W. R. Heineman, Analytical Chemistry 2014,

86, 4354-4361.

37. L. A. Girifalco, M. Hodak, R. S. Lee, Physical Review B 2000, 62, 13104-13110.

38. A. Star, J. F. Stoddart, D. Steuerman, M. Diehl, A. Boukai, E. W. Wong, X. Yang, S.-W. Chung,

H. Choi, J. R. Heath, Angewandte Chemie International Edition 2001, 40, 1721-1725.

39. M. F. Islam, E. Rojas, D. M. Bergey, A. T. Johnson, A. G. Yodh, Nano Letters 2003, 3, 269-273.

40. J. F. Liu, W. A. Ducker, Langmuir 2000, 16, 3467-3473.

41. Y. Q. Tian, N. B. Li, H. Q. Luo, Electroanalysis 2009, 21, 2584-2589.

42. R. M. Snider, M. Ciobanu, A. E. Rue, D. E. Cliffel, Analytica Chimica Acta 2008, 609, 44-52.

43. W. S. Huang, C. H. Yang, S. H. Zhang, Analytical and Bioanalytical Chemistry 2003, 375, 703-

707.

44. J. Wang, M. Musameh, Y. H. Lin, Journal of the American Chemical Society 2003, 125, 2408-

2409.

45. J. Wang, M. Musameh, Analytical Chemistry 2003, 75, 2075-2079.

46. J. Wang, G. Chen, M. Wang, M. P. Chatrathi, Analyst 2004, 129, 512-515.

47. J. Chen, M. A. Hamon, H. Hu, Y. Chen, A. M. Rao, P. C. Eklund, R. C. Haddon, Science 1998,

282, 95-98.

48. J. E. Riggs, Z. Guo, D. L. Carroll, Y.-P. Sun, Journal of the American Chemical Society 2000,

122, 5879-5880.

49. J. Morton, N. Havens, A. Mugweru, A. K. Wanekaya, Electroanal. 2009, 21, 1597-1603.

50. Z. Wang, E. Liu, D. Gu, Y. Wang, Thin Solid Films 2011, 519, 5280-5284.

51. D. Pan, Y. Wang, Z. Chen, T. Yin, W. Qin, Electroanalysis 2009, 21, 944-952.

52. B. C. Janegitz, L. H. Marcolino-Junior, S. P. Campana-Filho, R. C. Faria, O. Fatibello-Filho,

Sensors and Actuators B-Chemical 2009, 142, 260-266.

53. Y. J. Lee, H. H. Kim, H. Hatori, Carbon 2004, 42, 1053-1056.

54. T. I. T. Okpalugo, P. Papakonstantinou, H. Murphy, J. McLaughlin, N. M. D. Brown, Carbon

2005, 43, 2951-2959.

55. C. Chen, B. Liang, D. Lu, A. Ogino, X. Wang, M. Nagatsu, Carbon 2010, 48, 939-948.

56. Y. Wei, R. Yang, X. Chen, L. Wang, J.-H. Liu, X.-J. Huang, Analytica Chimica Acta 2012, 755,

54-61.

57. A. Salmanipour, M. A. Taher, J Solid State Electrochem 2011, 15, 2695-2702.

58. S. H. Shin, H. G. Hong, Bull Korean Chem Soc 2010, 31, 3077-3083.

59. B. C. Janegitz, L. H. Marcolino-Junior, S. P. Campana-Filho, R. C. Faria, O. Fatibello-Filho,

Sensors and Actuators B: Chemical 2009, 142, 260-266.

60. S. Jahandari, M. A. Taher, H. Fazelirad, I. Sheikhshoai, Microchimica Acta 2013, 180, 347-354.

61. P. H. Deng, J. J. Fei, Y. L. Feng, Sensors and Actuators B: Chemical 2010, 148, 214-220.

62. J. Wei, D. Yang, H. Chen, Y. Gao, H. Li, Sensors and Actuators B: Chemical 2014, 190, 968-974.

63. A. Mohadesi, H. Beitollahi, M. A. Karimi, Chinese Chemical Letters 2011, 22, 1469-1472.

64. W. Deng, Y. Tan, Z. Fang, Q. Xie, Y. Li, X. Liang, S. Yao, Electroanalysis 2009, 21, 2477-2485.

Chapter 2| Anodic Stripping Voltammetry of Heavy Metals on a Metal

Catalyst Free Carbon Nanotube Electrode

2.1 Introduction

Heavy metal pollution is one of the serious pollution problems in nature due to the stability of metals at contaminated sites and high toxicity to the biosphere. Lead, cadmium and zinc are major heavy metal contaminants from industry and are being continuously added to the biosphere. Because of the extensive use of these metals and the hazardous effect to both humans and other organisms, the development of sensitive monitoring methods becomes increasingly crucial 1, 2. Existing effective methods such as atomic absorption spectroscopy (AAS), inductively coupled plasma atomic emission spectrometry (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS) and anodic stripping voltammetry (ASV) can all give satisfactory results in detecting heavy metals. These techniques, except for ASV, although highly sensitive and reliable, require complicated operation, relatively large volume of sample and expensive instrumentation to maintain and operate. On the other hand, ASV has the advantages over the above measurements of low cost and simple experimental apparatus 3.

ASV is well known as a powerful technique for detecting diverse trace heavy metal ions.

Analytes are deposited on the electrode during the accumulation step which may last as long as the measurement requires4, 5. In this way, sensitivity may be improved by increasing accumulation time. Previously, mercury based electrodes have been used for their excellent performance with stripping voltammetry. The typical one is the hanging mercury drop electrode

(HMDE) which has been used for years due to its amenable electrode surface and good performance with stripping voltammetry6. However, because of its toxicity, mercury-based electrodes are no longer favored. Many substitutes are being explored in the search for electrodes to replace mercury-based electrodes 7, 8. Bismuth film coated electrodes have emerged as a good substitute for mercury based electrode due to its lower toxicity and high sensitivity. However,

bismuth film coated electrodes have a narrower potential window compared with carbon-fiber and gold electrodes 9. Gold electrodes are also significantly affected by chloride environments and oxidizing species such as chromate 10, 11. In recent years, carbon nanotubes (CNTs) have attracted scientists as a novel electrode material due to their excellent electrochemical features: large potential window, fast electron transfer rate and large surface area 12-14. Having a large portion of edge plane character makes CNTs become a superb electrochemical material in stripping voltammetry analysis, which surpasses other carbon forms such as: glassy carbon and boron doped diamond. Because the edge plane sites have faster electron transfer rate which allows low detection limit, high sensitivity, improved signal to noise ratio and low overpotentials15-18. Currently, the most common ways to synthesize CNTs are arc discharge, laser ablation and chemical vapor deposition 19, 20. These methods inevitably use metal catalysts in the fabrication process and result in metal impurities within the CNTs. These features can hinder the application of CNTs in producing sensors because metal particles trapped in CNTs are difficult to remove and in many instances can degrade the experimental result 21, 22.

Previously, Banks’s group has reported metal catalyst free CNTs (MCFCNTs) synthesized via a solid-phase growth mechanism 23. In this paper, we report how the MCFCNTs behave as a new electrode material in ASV for heavy metal detection. Good sensitivity was obtained with the novel CNTs electrode. Osteryoung square-wave mode was used as the detection method24.

Experimental variables, such as deposition time, deposition potential and conditioning potential were optimized.

2.2 Experimental

2.2.1 Chemicals and instrumentation

0.1 M Acetate buffer (pH 4.65) was purchased from Aldrich; Pb(NO3)2，Cd(NO3)2·4H2O and hydrogen peroxide were purchased from Fisher Scientific, zinc acetate was purchased from

J.T. Baker Chemical Co. MCFCNTs electrodes were supplied by SCNTE LLC (Beavercreek ,

OH). The CNTs electrodes were produced via carbo-thermal carbide conversion (CTCC), a metal catalyst free process. Thus the electrodes are guaranteed to be metal catalyst free 23. All the chemicals and electrodes were used without further purification and treatment. Stock solutions were prepared by dissolving solute with deionized water (resistivity 18.2 MΩ from Milli-Q

System, Barnstead) and working solutions were diluted with deionized water to various concentrations.

Measurements were carried out in a 20 ml conventional three-electrode cell consisting of

MCFCNTs electrode as working electrode, Ag/AgCl as reference electrode (filled with 3M KCl solution), Pt wire as auxiliary electrode. The solution volume is 15 ml for anodic stripping voltammetry experiments. A BASi 100B Electrochemical Analyzer from BASi was used as the potentiostat. Basic set-up parameters for Osteryoung ASV were S.W. Amplitude = 25 mV, step potential = 5 mV and frequency = 15 Hz 25. A Perkin Elmer AAnalyst 300 was used for AAS measurements. AA standard zinc stock solution was used for AAS measurements. Experiments were performed under room temperature and acetylene and oxygen mixture was used as fuel for nebulizing samples. Voltammetric response was determined by peak height and averaging 3 replicas.

2.2.2 Procedure

ASV measurements were performed in acetate buffer solution. Pb2+, Cd2+, and Zn2+ were reduced at –1.35 V with a deposition time of 150 s (or 600 s for longer deposition time

experiments) in a stirred solution, the stirring rate was held at a constant rate by using a magnetic stir bar in the cell. After a quiet time of 20 s voltammograms were recorded. Before each voltammogram, the electrode was refreshed by applying 0.8 V for 1-2 min to remove residual metals.

2.3 Results and discussion

2.3.1 Optimization of parameters

2.3.1.1 Deposition potential

The effect of deposition potential on peak current was explored for Pb2+, Cd2+ and Zn2+ to find the optimum potential for each analyte. Because Pb2+, Cd2+ and Zn2+ had the maximum signals (Figure 2.1) at a deposition potential of -1.35 V vs. Ag/AgCl (3 M KCl), this potential was selected as the optimized deposition potential in the subsequent measurements. The decreasing trend for peak current at more negative potentials is probably due to the reduction of some other species (e.g., water or H+ to form hydrogen gas), which could influence the availability of electrochemically active surface of the electrode by competing for electron transfer sites or interfere with mass transfer during deposition.

18 Pb 16 Cd

14 Zn

) 12

 ( 10

Peak current Peak 4

0 -1000 -1100 -1200 -1300 -1400 -1500 Deposition potential (mV) vs. Ag/AgCl

Figure 2.1 The effect of deposition potential on the stripping current of Pb2+, Cd2+ and Zn2+, concentrations were 3 µM Pb2+, 3 µM Cd2+ and 5 µM Zn2+ in 0.1M acetate buffer (pH = 4.65).

Deposition time: 20 s.

2.3.1.2 Deposition time

Next the effect of deposition time on peak current was studied by varying deposition time from 50 s to 400 s. Peak signals increased significantly (as shown in Figure 2.2) indicating an enhancement of metal deposition on the electrode as deposition time increased. Also, the slope of the curve decreases as time increases which means a decrease of sensitivity. To better understand the reason for this, another experiment was carried out in a 50 ml solution with the same concentration of lead nitrate in acetate buffer using all the same experimental set-up and parameters. A linear response was observed from 50 s to 600 s. Based on this, we can conclude

that the decreasing trend of slope was because of the depletion of the analyte in the 15 ml solution. Thus, a deposition time of 150 s was selected for subsequent measurements as a compromise between sensitivity, required sample volume and a relatively short experiment time.

70 Pb

60 Cd Zn

)

A  ( 40

20 Peak current Peak

0 0 50 100 150 200 250 300 350 400 450 Deposition time (s)

Figure 2.2 The effect of deposition time on the stripping current of Pb2+, Cd2+ and Zn2+, concentrations were 3 µM for Pb2+, 3 µM for Cd2+ and 5 µM for Zn2+ in 0.1M acetate buffer

(pH = 4.65). Deposition potential: -1.35 V, cell volume: 15 ml.

2.3.1.3 Conditioning potential and time

After each measurement, especially the measurements with higher concentrations than 5 µM or longer deposition times, it is likely to leave some residual metal on the electrode surface which can interfere with the peak current for the subsequent measurements. In order to minimize this memory effect of residual metal from the previous measurement, a conditioning potential

was applied to the CNTs electrode after each experiment. In this process, timebase mode was selected to apply a constant potential on the electrode. The effects of conditioning potential and time were studied and good reproducibility was obtained at a potential of 0.8 V for 60 s. When the trace on the current vs. time graph levels off, the cleaning step is finished. Also, after being used for 200-300 measurements, the electrodes became less reproducible and background noise increased. The peak current can vary 10 % - 20 % and several tiny peaks show up on the background. In this situation, an even higher voltage is required to refresh the electrode, typically

1.5 V vs. Ag/AgCl for 1-2 min 26. Because of the unique array structure of MCFCNTs, it allows the applied high voltage to refresh the electrode surface by oxidizing the outer CNTs edges and exposing new CNTs edge plane character surfaces 27. This extends the useful electrode life-time to several months of MCFCNTs when being used in stripping voltammetry analysis. Stripping voltammograms were taken after electrochemical activation of 1.5 V with a deposition time of

30 s to verify the conditioning effect. Voltammograms comparable to those obtained on fresh electrodes were obtained showing that the electrode was refreshed after the conditioning step.

2.3.2 Metal-catalyst-free carbon nanotube electrode

Traditional CNTs synthesis methods usually utilize transition metals to catalyze the formation process 28. The residual metal oxide cannot be fully removed by posttreatment has been reported that in the case which metal oxide impurities are partially trapped in the CNTs, the material has a very significant electrochemical activity such as reduction of the metal oxide from CNTs 29. In addition, the batch-to-batch variation of CNTs in terms of metallic impurity content can also hinder the application of CNTs in electrochemistry. Residual metal catalyst can decrease the negative potential window, which can interfere with the deposition of more electronegative metals.

And it can also increase the background current in square-wave voltammetry, which degrades the

limit of detection 23, 30. The CNTs used here are produced via a metal catalyst free process. The

MCFCNTs are synthesized via Carbo Thermal Carbide Conversion method which uses silicon carbide as a deletional matirx to grow carbon nanotubes, thus no transition metal is involved in the synthesis process. Figure 2.3 shows the cyclic voltammograms of the MCFCNTs electrode in both phosphate buffer and hydrogen peroxide. There is almost no difference between the two voltammograms indicating the electrode is free of metallic impurities because there will be a significant reduction peak of hydrogen peroxide if there are metal oxide impurities in carbon nanotubes 31.

)

A 

( 0

-10 Current

-20 5 mM H O in phosphate buffer pH 7 2 2 phosphate buffer pH 7

-30 -1000 -800 -600 -400 -200 0 200 400 600 Potential (mV) vs. Ag/AgCl

Figure 2.3 Cyclic voltammograms of 0.1 M pH 7 phosphate buffer and 5 mM hydrogen peroxide

2.3.3 Calibration data

The OSWSV determination of Pb2+, Cd2+, and Zn2+, was performed under optimized parameters described above. Figure 2.4A shows sharp peaks for Pb at -0.45 V and a linear response was obtained in the concentration range of 3*10-7~ 8*10-6 M. The correlation equation was I(µA)=10.42[Pb(µmol)]-2.867 (R2=0.9990 for 7 concentrations within the range). The limit of detection was calculated to be 13 nM (based on 3σ). Figure 2.4B shows sharp peaks for Cd at

-0.7 V and a linear response was obtained in the concentration range of 5*10-7~5*10-6 M. The correlation equation was I(µA)=7.81[Cd(µmol)]-3.959 (R2=0.998 for 7 concentrations within the range). The limit of detection was calculated to be 32 nmol. Figure 2.4C shows broader peaks for

Zn at -0.95 V and a linear response was obtained in the concentration range of 5*10-7~7*10-6 M.

The correlation equation was I(µA)=2.3787[Zn(µmol)]-0.6928 (R2=0.998 for 7 concentrations within the range). The limit of detection was calculated to be 50 nM. Error bars show extremely reproducible property of the electrode for stripping voltammetry (Figure 2.4).

0 Pb -10 A -20

-30

) 90 A -40 Pb

 80 (

70 )

-50 A 60

 ( 50

-60 40

Current 30 Peak current Peak -70 20

10 R2=0.999 -80 0 0 1 2 3 4 5 6 7 8 9 Concentration (mol/L) -90

200 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV) vs.Ag/AgCl

Figure 2.4A) OSWSV and calibration curve for Pb2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.3, 0.5, 1.0, 2.0, 3.0, 5.0 and 8.0 µM of Pb2+; Deposition time: 150 s, deposition potential: -1.35 V

0 Cd

-10 B

-20 90 Cd

80 ) 70

A -30 

( 60 )

A 50  -40 ( 40

30 -50 Current Current 20 2 R =0.9983 10

-60 0 0.5 1.0 1.5 2.0 2.5 3.0 Concentration (mol/L) -70

200 0 -200 -400 -600 -800 -1000 -1200 Potential (mV)

Figure 2.4B) OSWSV and calibration curve for Cd2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.5, 0.7, 1.0, 2.0, 3.0, 6.0 and 8.0 µM; Deposition time: 150 s, deposition potential:

-1.35 V

0 Zn

-5 C -10

) 14 Zn A -15

 12

( )

A 10

 ( -20 8

6 Current

-25 current Peak 4

2 R2=0.9981 0 -30 0 1 2 3 4 5 6 7 Concentration (mol/L)

-35 -200 -400 -600 -800 -1000 -1200 Potential (mV) vs.Ag/AgCl

Figure 2.4C) OSWSV and calibration curve for Zn2+ in 0.1M acetate buffer (pH=4.65), concentration: 0.5, 1.0, 2.0, 3.0, 4.0, 6.0 and 8.0 µM. Deposition time: 150 s, deposition potential:

-1.35 V

The descending trend of calibration slope in the sequence of Pb2+, Cd2+ and Zn2+ indicates the electrode has the most sensitive response to Pb2+ followed by Cd2+ and Zn2+. In Cd2+ and

Zn2+ voltammograms, extra peaks were seen at potentials more positive than the normal dissolution potential of the metal film. This phenomenon has been suggested for copper, cadmium and zinc as the oxidation of the first monolayer of metal on the electrode which requires much more energy 32, 33. In the case of solid electrodes, such as graphite and CNTs electrodes the first monolayer deposited on another solid substrate can have a considerably larger bonding energy than the bonding energy of metal to itself 32-35 .

In order to explore how low a detectable concentration level can be reached, 10 min was chosen as a longer deposition time. An extremely low concentration 0.1 nM was the lowest that could be detected for Pb2+ and also 8 nM and 40 nM were reached for Cd2+ and Zn2+, respectively (Figure 2.5).

0 Pb Cd -5 -4

-10

)



( ( -15

-6 Current

Current -20

-25

-8 -30 -200 -400 -600 -800 -1000 0 -200 -400 -600 -800 -1000 Potential (mV) Potential (mV)

0 Zn -5

-10

) -15

 ( -20

-25

Current -30

-35 -40 0 -200 -400 -600 -800 -1000 -1200 Potential (mV)

Fig. 5 OSWSV of 0.1 nM Pb2+, 8 nM Cd2+ and 40 nM Zn2+ in 0.1 M acetate buffer (pH=4.65).

Deposition time: 600 s, deposition potential:-1.35 V.

Finally, we explored simultaneous detection of Pb2+, Cd2+ and Zn2+ which showed a very well-defined and highly resolved votammogram (Figure 2.6A). Stripping peaks of metals in the mixture remained at the same potential compared with voltammograms of single metals. The voltammogram shows very reproducible linear response for all of the three analytes in the range of 0.5-2.5 µM with slopes of 19.74 µA*L*µmol-1 for Pb2+, 5.392 µA*L*mol-1 for Cd2+ and 2.467

µA*L*µmol-1 for Zn2+ (Figure 2.6B).

0 A -5 -10 Pb,Cd,Zn

-15

) -20

A 

( -25

-30 -35

Current -40

-45 -50

-55

200 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)vs. Ag/AgCl

Figure 2.6 A) OSWSV of simultaneous detection of Pb2+, Cd2+ and Zn2+, concentration 0.5, 1.0,

1.5, 1.7 and 2.5 µM.

45 Pb Cd 40 Zn

35 B )

A 30

 ( 25

10 Peak current Peak

0.5 1.0 1.5 2.0 2.5 Concentration (mol/L)

Figure 2.6 B) Calibration curve for Cd2+ and Zn2+, concentration 0.5, 1.0, 1.5, 1.7 and 2.5 µM.

Deposition time:150 s, deposition potential: -1.35 V.

The comparable sensitivities between single and simultaneous measurements imply that simultaneous detection is quite practicable. The relative standard deviation for three consecutive measurements was 1.0% for Pb2+, 1.1% for Cd2+ and 0.38% for Zn2+ in the sample containing 1.5

µM of each of the three analytes. When the concentration goes higher, the trend starts to level off

(Figure 2.7). Also, 50 ml simultaneous detection was done to compare with the 15 ml results.

Linear response was observed for Pb2+and Zn2+ in the range from 1µM to 7 µM, but two linear responses were observed for Cd2+.

Pb 50 Cd Zn

)

A  ( 30

10 Peak current Peak

0.5 1.0 1.5 2.0 2.5 3.0 3.5 Concentration (mol/L)

Figure 2.7 Trend curve of simultaneous detection of Pb2+, Cd2+ and Zn2+, concentration 0.5, 1.0,

1.5, 1.7, 2.5, 3.0 and 3.5 µM.

Previously, CNTs were used mostly as substrates for coated electrodes on stripping analysis, such as bismuth and nafion film coated electrodes 1, 2, 36-38. As shown from Table 2.1, there are usually one or more than one modifiers used in modified CNTs electrode in stripping analysis， such as Nafion modified electrode and ABTS/Bi modified electrodes 39, 40. Bi film coated CNTs electrodes showed excellent stripping behavior among the modified electrodes 3, 39-41. The

MCFCNTs electrode showed excellent reproducibility and low limit of detection as well.

Detection limits of heavy metals in this work are comparable with the lowest ones reported previously (shown in Table 1). However, considering the absence of film coating such as

mercury and bismuth, the MCFCNTs electrode has the advantage of ease of operation which is

especially important when being applied in nature water quality monitoring.

Electrode Analytical Depositio Lowest detectable References

Modifier technique n time(s) detection limit (nM) Electrode

Pb2+ Cd2+ Zn2+

MCFCNTs _ ASV 600 0.1 8 40 This work electrode

MWCNTs/Nafion Nafion LSV 360 5 _ _ 39

Bi-coated carbon Bi ASV 600 1.4 _ _ 42 electrode

Bi/ABTS- Bi/ABTS DPV 300 0.5 1.8 _ 40

MWCNTs/GC

Cysteine-CNTs Cysteine ASV 600 4.8 _ _ 3

Table 1. Detection limits of heavy metals with ASV, LSV and DPV on different electrodes

2.3.4 Analysis of tap water

The MCFCNTs electrode then was applied to the determination of Zn2+ in tap water. Tap

water was collected from University of Cincinnati and diluted in a ratio of 1:1 with 0.1 M acetate

buffer (pH = 4.65), and the standard addition method was used. Preliminary experiment showed

that a deposition time of 900 s was adequate for detection of zinc. The stripping voltammetric

response to additions of standard Zn2+ is shown in Figure 2.8. The result was then compared with AAS as shown in Table 2. In AAS analysis, a calibration curve was built up which incorporate the concentration of the sample in the middle of the range. Then the Zn2+ concentration in tap water was calculated. Since the result from stripping voltammetry agrees

93.6% with AAS result, we can conclude that the MCFCNTs electrode can be applied to natural water sample detection.

-5 )

A -10

 (

-15 Current

-20

-25 -600 -800 -1000 -1200 -1400 Potential (mV)

Deposition time: 900 s; Deposition potential: -1.35 V

Determination method Zn2+ concentration in tap water (µmol/L)

ASV 1.91

AAS 2.04

Table 2. Zn2+ concentration in tap water detected by ASV and AAS

2.4 Conclusion

In conclusion, we have demonstrated the suitability of the metal catalyst free carbon nanotube electrode for heavy metals detection. The novel electrode provides excellent performance on both single and simultaneous detection of Pb2+, Cd2+ and Zn2+. The deposition time, deposition potential, conditioning time and conditioning potenial were all optimized. Tap water sample detection also confirms that MCFCNTs electrode is reliable in heavy metal detection in natural water samples. The MCFCNTs electrode shows comparable limit of detection with other mercury free electrodes, such as: CNTs modified electrodes and bismuth film electrode. Because other than the MCFCNTs there are no other additives added to the solution, this can be considered as an environmental friendly material for ASV.

Another attractive respect of the MCFCNTs is due to its unique structure, the electrode can be electrochemically rejuvenated after being used for hundreds of cycles. This property is very meaningful for the application of natural water detection. Preliminary results show that the

MCFCNTs electrode is capable of detecting Cu2+, Hg2+ and Mn2+. Thus, for future prospects, we will study the behavior of MCFCNTs electrode in other major heavy metals. Acknowledgement

2.5 Acknowledgement

The authors wish to acknowledge support provided by the Ohio Department of Development,

Ohio Third Frontier Sensors Program (Grant TECH 10-072). We also thank the Environmental

Chemistry Lab at the University of Cincinnati for the analysis by AAS.

2.6 References

1. P. Chooto, P. Wararatananurak, C. Innuphat, Science Asia 2010, 36, 150-156.

2. M. Rajabi, A. Asghari, H. Zavvar Mousavi, Journal of Analytical Chemistry 2009, 65, 511-517.

3. J. Morton, N. Havens, A. Mugweru, A. Wanekaya, Electroanalysis 2009, 21, 1597-1603.

4. G. Tölg, Angewandte Chemie International Edition in English 1986, 25, 485-485.

5. J. Wang, Controlled-Potential Techniques. Analytical Electochemistry, John Wiley & Sons, Inc:

2006; p 67-114.

6. J. Wang, Practical Considerations. Analytical Electochemistry, John Wiley & Sons, Inc: 2002; p

115-163.

7. J. Wang, B. Tian, Analytical Chemistry 1993, 65, 1529-1532.

8. E. Achterberg, C. Braungardt, Analytica Chimica Acta 1999, 400, 381-397.

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

10. D. Mandler, A. Bard, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry

1991, 307, 217-228.

11. J. Hatch, A. Gewirth, Journal of The Electrochemical Society 2009, 156, D497-D502.

12. R. Baughman, Science 2000, 290, 1310-1311.

13. G. Dukovic, M. Balaz, P. Doak, N. Berova, M. Zheng, R. McLean, L. Brus, J. Am. Chem. Soc.

2006, 128, 9004-9005.

14. B. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, L. Bachas, Science 2004, 303, 62-65.

15. C. Banks, R. Compton, Anal. Sci. 2005, 21, 1263-1268.

16. X. Ji, R. O. Kadara, J. Krussma, Q. Chen, C. Banks, Electroanalysis 2010, 22, 7-19.

17. K. Wu, S. Hu, J. Fei, W. Bai, Anal. Chim. Acta 2003, 489, 215-221.

18. T. Marek, TrAC, Trends Anal. Chem. 2006, 25, 480-489.

19. S. Huang, L. Dai, A. Mau, J. Phys. Chem. B 1999, 103, 4223-4227.

20. R. B. Mathur, S. Seth, C. Lal, R. Rao, B. Singh, T. Dhami, A. Rao, Carbon 2007, 45, 132-140.

21. U. Michael, Journal of Nanomaterials 2008, 2008, 1.

22. G. Tibbetts, G. Meisner, C. Olk, Carbon 2001, 39, 2291-2301.

23. C. Jones, K. Jurkschat, A. Crossley, R. Compton, B. Riehl, C. Banks, Langmuir 2007, 23, 9501-

9504.

24. R. Osteryoung, Analytical chemistry (Washington) 1985, 57, A101-&.

25. A. J. Bard, Electroanalytical chemistry 1986, 14, 209-308.

26. C. Zoski, Handbook of Electrochemistry 2007.

27. M. Pumera, T. Sasaki, H. Iwai, Chem. Asian J. 2008, 3, 2046-2055.

28. S. Fan, M. Chapline, N. Franklin, T. Tombler, A. Cassell, H. Dai, Science 1999, 283, 512-514.

29. X. Ji, R. Kadara, J. Krussma, Q. Chen, C. E. Banks, Electroanalysis 2010, 22, 7-19.

30. C. E. Banks, A. Crossley, C. Salter, S. Wilkins, R. Compton, Angew. Chem. Int. Ed. 2006, 45,

2533-2537.

31. M. Pumera, Langmuir 2007, 23, 6453-6458.

32. B. Vassos, H. Mark Jr, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry

1967, 13, 1-9.

33. K. Gardiner, L. Rogers, Analytical Chemistry 1953, 25, 1393-1397.

34. M. Nicholson, Journal of the American Chemical Society 1957, 79, 7-12.

35. S. Perone, Analytical Chemistry 1963, 35, 2091-2094.

36. J. He, X. Lin, J. Pan, Electroanalysis 2005, 17, 1681-1686.

37. N. Lawrence, R. Deo, J. Wang, Electroanalysis 2005, 17, 65-72.

38. C. Tarley, V. Santos, B. Baeta, A. Pereira, L. Kubota, Journal of Hazardous Materials 2009, 169,

256-262.

39. D. Sun, Z. Sun, Journal of Applied Electrochemistry 2008, 38, 1223-1227.

40. W. Deng, Y. Tan, Z. Fang, Q. Xie, Y. Li, X. Liang, S. Yao, Electroanalysis 2009, 21, 2477-2485.

41. G. Liu, Y. Lin, Y. Tu, Z. Ren, Analyst 2005, 130, 1098-1101.

42. J. Wang, J. Lu, S. Hocevar, P. Farias, B. Ogorevc, Analytical Chemistry 2000, 72, 3218-3222.

Chapter 3| Manganese Detection with a Metal Catalyst Free Carbon

Nanotube Electrode: Anodic versus Cathodic Stripping Voltammetry

3.1 Introduction

Manganese (Mn) is commonly found throughout most aquatic environments. While Mn is a required trace metal, elevated concentrations of Mn are associated with a host of health issues, including neurotoxicity and development of Parkinson’s disease symptoms [1, 2]. For this reason,

Mn in drinking water is regulated at a low concentration of 50 ppb, and thus the determination of

Mn in aqueous samples is of direct practical importance [3].

Both spectroscopic and electroanalytical methods are commonly used for the determination of trace metal contaminants such as Mn; both methods offer advantages for different applications.

Typically, electroanalytical methods are less expensive, more portable, and require fewer separation steps than spectroscopic techniques. The electroanalytical technique most commonly used for the determination of metal analytes at the trace level is stripping voltammetry.

Stripping voltammetry broadly describes a variety of electroanalytical techniques that are often used for many analytical applications where a high level of sensitivity is required.

Stripping methods are generally more sensitive than other voltammetric techniques because of a preconcentration step which accumulates the desired analyte on the surface of the electrode. The analyte may be preconcentrated by either electrodeposition or by physical adsorption, depending on the analyte and the stripping method being used. Once sufficient preconcentration is achieved, the potential of the working electrode is swept as to strip the analyte off of the electrode surface, with the associated faradaic current being measured to quantitatively determine the concentration of analyte present [4].

Anodic stripping voltammetry (ASV) is the most commonly used form of stripping voltammetry. In this technique the analyte, typically a metal ion is preconcentrated on the electrode surface by reductive electrodeposition. The electrode potential is then swept in the positive direction, and metal ions are oxidatively liberated from the electrode surface at their

oxidation potentials [5, 6]. The determination of numerous metals has been reported using ASV, on solid electrode surfaces as well as on liquid films and drops of mercury [5, 7]. While most commonly applied to metals such as copper, cadmium, lead, and zinc, ASV has also been used for the determination of Mn [8]. Mn is more difficult to determine by ASV in some applications because of the negative potential of the Mn2+ + 2e-  Mn redox couple. This reduction potential is beyond the working range of most common solid electrode materials, and most reported ASV methods use a Hg surface [4, 6, 8]. Due to toxicity concerns, the use of Hg is often limited, making these analyses unsuitable for many applications [9-12].

A second form of stripping voltammetry that has been used for Mn determination is adsorptive stripping voltammetry (AdSV) [13, 14]. In AdSV, the deposition of the analyte is not accomplished by means of electrolysis, but rather by a physical or chemical interaction with the electrode surface. Many of the AdSV techniques for Mn that have been reported in the literature use Hg, making these techniques unsuitable for applications requiring mercury-free analysis [14].

CSV is a third stripping technique used for Mn determination [15-17]. CSV is the reverse of

ASV in that the analyte is accumulated as an oxidized species, and is stripped by a potential sweep in the negative direction. This technique offers several advantages for Mn determination, including insensitivity to oxygen and intermetallic interferences, Hg free analysis, and a redox potential within the working range of many common electrode materials. CSV methods have been reported using glassy carbon, platinum, and boron doped diamond electrodes [15, 18, 19].

Recently, carbon nanotubes (CNTs) have emerged as a novel electrode material due to their useful electrochemical properties, such as large potential window, fast electron transfer rate and large surface area [20-22]. CNT electrodes have been created in a number of ways, including arc discharge, laser ablation and chemical vapor deposition [23, 24]. A hindrance of these methods

when producing electrodes for trace stripping analysis is that they generally use metal catalysts

that can contaminate the CNTs and reduce the analytical effectiveness for sensor applications

[25-27]. Previously, metal catalyst free CNTs (MCFCNTs) synthesized via a solid-phase growth

mechanism has been reported by the Banks research group [28]. CNTs are grown on a silicon

carbide matrix which does not have a residue of transition metal catalyst in the CNT structure

[28, 29]. Because of its unique array structure, this novel material is very robust and the CNTs

can be “rejuvenated” by refreshing the electrode with a high voltage treatment [30]. In this paper,

the use of MCFCNT electrodes for trace determination of Mn using ASV and CSV is explored.

3.2 Experimental

All chemicals were purchased without further purification: Mn AAS standard solution with 2%

HNO3 from Fisher Scientific, 20X borate buffer from Thermo Scientific. MCFCNT electrodes

were supplied by SCNTE LLC (Beavercreek, OH) and used without any pretreatment. All

solutions were prepared with deionized water (18.2 MΩ from Milli-Q System, Barnstead, MA).

ASV measurements were carried out in a 20 mL conventional three-electrode cell consisting of

MCFCNT electrode as working electrode, Ag/AgCl as reference electrode (filled with 3 M KCl

solution), Pt wire as auxiliary electrode. The solution volume was 15 mL for both ASV and CSV

experiments. A BASi 100B Electrochemical Analyzer from BASi (West Lafayette, IN) was used

as the potentiostat. Basic set-up parameters for Osteryoung square wave voltammetry were S.W.

amplitude = 25 mV, step potential = 5 mV and frequency = 25 Hz [31].

3.3 Results and Discussion

3.3.1 Anodic Stripping Votammetry Study of Mn

3.3.1.1 Cyclic Voltammetry of Mn in NH4Cl

ASV of Mn in NH4Cl has been done and shown to give repeatable and reliable results on rotating solid silver amalgam electrode in 0.05 M NH4Cl solution from Mikkelsen’s lab [32]. For

2+ this reason, aqueous solutions consisting of 0.05 M NH4Cl (pH 4.5) with 1.5 mM Mn were prepared and studied using cyclic voltammetry on MCFCNT electrodes to characterize the redox behavior of Mn. A flat background was seen for 0.05 M NH4Cl in the range of 1400 mV to -

1500 mV and hydrolysis started from -1500 mV. Then, the potential was cycled with 1.5 mM

2+ Mn in 0.05 M NH4Cl both from 0 mV to 1500 mV and from 1500 mV to -2000 mV to observe multiple relevant redox couples, as seen in Figure 3.1. As illustrated in the cyclic voltammogram, the reduction of Mn2+ starts at approximately -1000 mV with a peak width of 600 mV and a peak potential of -1300 mV and the oxidation peak of deposited Mn is at -1250 mV. This peak location is compatible with ASV in the potential window of the MCFCNT electrode, so experiments were done to optimize the experimental conditions for ASV of Mn2+.

2+ Figure 3.1 Cyclic voltammetry of 1.5 mM Mn in 0.05 M NH4Cl on MCFCNTs electrode, scan rate 100 mV/s

3.3.1.2 pH and Deposition Potential Optimization

Osteryoung square wave mode was used for ASV measurements, and the first variable that was investigated was pH. Decreasing pH from 4.5 causes increasing peak current of Mn and for pH values lower than 3, evolution of hydrogen gas starts to significantly interfere with Mn electrodeposition on the electrode surface. The ASV analysis at pH 3 was shown to have the sharpest and highest peak current. Thus, pH 3 was selected for further optimization of ASV parameters for the measurement of Mn. The next parameter explored was deposition potential.

This was done using 100 µM Mn in 0.05 M NH4Cl (pH 3). As shown in Figure 3.2, -1800 mV produces the sharpest peaks and largest peak current. At each potential, the measurements were done three times and the standard deviation is shown in Figure 3.2. The decrease in peak current for more negative deposition potentials is attributed to hydrogen gas evolution from the reduction of water at the electrode surface which interferes with the mass transport of manganese analyte to the electrode surface for preconcentration. Thus, -1800 mV was selected as the most suitable deposition potential for further measurements.

)

A  ( 35

30 PeakCurrent

-1650 -1700 -1750 -1800 -1850 -1900 Deposition Potential (mV)

2+ Figure 3.2 Deposition potential optimization of 100 µM Mn in 0.05 M NH4Cl

3.3.1.3 Reproducibility and Cleaning Step Study

Between measurements, a constant potential of 600 mV was applied to thoroughly oxidize residual manganese on the electrode surface. After this step, the electrode showed reproducible behavior for CSV manganese detection.

3.3.1.4 Calibration Data

With the deposition time of 60 s, the peak height of stripping peaks shows a linear relationship in the concentration range from 3.5 µM to 13.5 µM with the correlation equation I

(µA) = (0.66±0.0089) * [Mn (µmol)] – [(1.89±0.052) (µA)] and another linear range from 13.5

µM to 36.5 µM with the correlation equation I (µA) = (0.18±0.0097) * [Mn (µmol)] +

[(4.47±0.20) (µA)] as shown in Figure 3.3. Severe hydrogen evolution occurs in the manganese oxidation potential region, which leads to a distorted background for Mn voltammograms. The steep baseline causes the poorly-defined Mn stripping peaks since the manganese oxidation is superimposed on the hydrogen reduction wave. The narrow linear range is due to saturation of the working electrode surface with a layer of Mn. The second linear range follows the mechanism of deposition of Mn on top of the base Mn layer which has a different slope as reflected in the calibration curve. Based on the first linear range, the limit of detection by ASV was calculated (3σ/slope) to be 120 nM. Also, under the same experimental conditions, a 10 min deposition was explored and longer deposition time does not change either the two linear range patterns or increase the sensitivity. Because of the relatively narrow linear range of detection and poorly defined peak shape within the detection range of ASV, CSV was investigated as an alternative technique for Mn detection.

12 A

10 )

A 8

 (

4 Peak current Peak

0 0 5 10 15 20 25 30 35 40 Concentration (M)

2+ Figure 3.3 A) Calibration plot for ASV of Mn in 0.05 M NH4Cl;

-20 )

A -40

 (

-60 Current

-80

-100 0 -400 -800 -1200 -1600 Potential (mV)

2+ Figure 3.3 B) Anodic stripping voltammograms of Mn in 0.05 M NH4Cl in the concentration range of 3.5 µM to 13.5 µM

3.3.2 CSV Study of Mn

3.3.2.1 Cyclic Voltammetry of Mn in Borate Buffer

Cyclic voltammetry was first performed with Mn2+ to study the redox behavior of Mn2+ in a

2+ wide potential window. The accumulation of Mn is the oxidation of Mn to insoluble MnO2:

2+ + - Mn (H2O) x (aq)  MnO2 (H2O) x-y(s) +4H +(y-2) H2O+2e

Since a basic medium helps the formation of MnO2, borate buffer solution was selected as a basic buffer [15]. The voltammogram was first scanned from 0 to 1000 mV and then scanned from 1000 mV to -1000 mV and back to 0. Figure 3.4 shows two major reactions of Mn2+ on the

2+ CNT electrode in pH 8.5 borate buffer solution: oxidation of Mn to MnO2 at ca. 500 mV and

2+ 2+ reduction of Mn to Mn at ca. -1400 mV. The reduction peak of Mn is superimposed on the water reduction wave which, as discussed earlier, significantly interferes with ASV measurement

of Mn2+. On the other hand, the oxidation peak of Mn2+ is in a more central region of the potential window which has much less interference from water electrolysis compared to the ASV method.

600

500

400

) 300

 ( 200

100 Current 0

-100

-200 1000 500 0 -500 -1000 -1500 -2000 Potential (mV)

Figure 3.4 Cyclic voltammotram of 1.5 mM Mn2+ in 0.1 M pH 8.5 borate buffer, scan rate 100 mV/s

3.3.2.2 pH Optimization of Buffer Solution

Since the solubility of MnO2 is quite dependent on solution pH, optimization of buffer pH is critical for reproducible and sensitive experimental results. Figure 3.5 shows how pH affects the

CSV peak height of Mn2+. It is observed that with a deposition potential of 600 mV at pH 8.5 in borate buffer, the CSV peak has a larger peak height and better looking peak shape than the ASV peak. Therefore, pH 8.5 was selected as the buffer pH for CSV measurements. Stripping peaks are observed to shift to a more negative potential as the pH becomes more basic. This is because

the more acidic solution makes the reduction of MnO2 easier by pushing the equilibrium in the following electrode reaction to the right.

+ - 2+ MnO2 (H2O) x-y(s) +4H +(y-2) H2O+2e Mn (H2O) x (aq)

45 pH 7.5 A pH 8.0 pH 8.5 44 pH 9.0

pH 9.5 )

A 43

 (

42 Current

40 600 500 400 300 200 100 0 -100 Potential (mV)

Figure 3.5 CSV pH optimization of Mn2+ from pH 7.5-9.5, deposition potential: 600 mV;

3.3.2.3 Deposition Potential Optimization of CSV of MnO2

A deposition potential study was performed in pH 8.5 borate buffer with 1 µM Mn2+ solution.

Based on the CV of Mn2+ (Figure 3.4), 600 mV was chosen as the minimum deposition potential that would preconcentrate MnO2 onto the CNT electrode, and the deposition potential was investigated using increases of 50 mV up to 800 mV, as illustrated in Figure 3.6. Increasing the potential did not change the size or shape of MnO2 reduction, so 600 mV was chosen as the deposition potential for CSV.

50 600 mV B 49 650 mV 700 mV 48 750 mV 800 mV

47 )

A 46

 ( 45

Current 43

40 800 600 400 200 0 -200 Potential (mV)

Figure 3.6 CSV deposition potential optimization of Mn2+ from 600 mV to 800 mV, pH 8.5;

Mn2+ concentration for A and B: 1 µM, deposition time for A and B: 60 s.

3.3.2.4 Reproducibility and Cleaning Step Study

Similar to ASV measurements, a reproducibility test was carried out to establish a protocol for reliable CSV manganese measurements. Between measurements, a constant potential of -

1500 mV was applied to the electrode for 2 min to thoroughly reduce manganese oxide. After this step, the electrode showed reproducible behavior for CSV manganese detection.

3.3.2.5 Calibration Data

CSV measurements of a series of Mn standards were performed under optimized conditions as described above in 0.1 M pH 8.5 borate buffer solution. Figure 3.7A shows the dynamic range of CSV measurements of Mn in the concentration range from 0.12 µM to 12 µM. Each

measurement was done three times and the standard deviation was shown in Figure 3.7A. The linear range was from 0.6 to 6.7 µM and the correlation equation was I (µA) = (1.45±0.0029) *

[Mn (µmol)] – [(0.65±0.12) (µA)] (R2=0.99 for 11 concentrations within the range). The limit of detection was calculated to be 93 nM (based on 3σ/slope).

14 A

12 )

A 10

 ( 8

4 Peakcurrent

0 0 2 4 6 8 10 12 Concentration (M)

Figure 3.7 A) Dynamic range of CSV measurements of Mn in the range of 0.12 µM - 12 µM;

48 0.6 M B 1.2 M 1.8 M 46 2.4 M 3.0 M 3.6 M 44 4.2 M 4.8 M 5.4 M

) 42 6.0 M

6.6 M

 ( 40

38 Current 36

600 500 400 300 200 100 0 -100 Potential (mV)

A longer deposition time of 15 min under the same experimental condition gave a detectable concentration of 30 nM. A 0.12 µM Mn solution was used to study how accumulation time affects Mn CSV behavior and the result is shown in Figure 3.8 shows how accumulation time affects the peak current. It is seen that for an accumulation time of less than 13 min, peak current follows a linear trend vs. accumulation time; as time increases, peak current starts to level off due to depletion of the analyte from the sample.

4.5

4.0

3.5

3.0

) A

 2.5 (

2.0

1.5 Current 1.0

0.5

0.0 0 3 6 9 12 15 18 21 Accumulation time (min)

Figure 3.8 Accumulation time effect on peak current from 1–20 min; deposition potential: 600 mV; pH: 8.5; Mn concentration: 0.12 µM

3.4 Natural Matrix of CSV Study of Mn

In general, CSV has several advantages over ASV for Mn measurements on MCFCNT electrodes, such as lower limit of detection, better sensitivity and better reproducibility. To demonstrate that this method shows promise for Mn2+ detection in more complex matrices, a sample of natural water was run to evaluate this method in the presence of potential interferences.

Burnet Woods (Cincinnati, OH) pond water was collected from shore using a plastic sampling bottle on November 11th 2011. There was no further treatment to the pond water and it was used immediately after collection as a natural water matrix into which Mn2+ was spiked. Atomic absorption spectroscopy showed no detectable Mn2+ in the unspiked sample. For CSV analysis

the sample was adjusted to pH 8.5 using borate buffer. A 0.090 µM artificial Mn2+ sample was prepared by diluting AAS standard Mn solution (1000 ppm) into the buffered pH 8.5 pond water matrix and tested with the standard addition method using CSV. The peak current shows a linear response vs. additions of Mn stock solution: I (µA) = (0.83±0.037) * [Mn (µmol)] +

[(0.074±0.011) (µA)] (R2=0.99). The concentration of Mn in the spiked sample was calculated to be 0.09±0.01 µM which is in close agreement with 0.090 µM.

Sample 3.4 0.12 M addition 0.24 M addition 3.2 0.36 M addition

0.48 M addition )

A 3.0

 (

2.8 Current 2.6

2.4

600 500 400 300 200 100 0 -100 -200 Potential (mV)

Figure 3.9 Voltammograms of CSV measurements of Mn in spiked pond water sample; deposition time: 60 s; deposition potential: 600 mV; pH: 8.5.

3.5 Conclusions

We have evaluated two stripping voltammetry methods for detecting Mn2+ in aqueous

solutions using metal catalyst free carbon nanotube electrodes. The ASV method we examined

was shown to have a narrow linear range of detection and poor stripping peak shape. The CSV

method that was explored as an alternative shows a lower limit of detection and better sensitivity

(93 nM vs. 120 nM). The CSV method was then shown to be a very reproducible and reliable

technique for the detection of Mn2+, and robust enough to operate in a sample such as pond water.

The MCFCNT electrode for CSV detection of Mn has a low limit of detection, wide linear

dynamic range and good reproducibility. Previously, our lab has designed a bismuth film chip

which was used for ASV detection of Mn in blood samples [33]. However, due to the high

electronegativity of Mn, the limit of detection of Mn was only 5 µM. Since MCFCNT has shown

a promising CSV result for Mn detection with a limit of detection of only 93nM, it can

potentially be a coating material on our chip to provide a significantly better limit of detection

than was obtained on a bismuth electrode.

3.6 Acknowledgement

The author gratefully acknowledges support provided by NIEHS R21ES019255 and SCNTE

LTD.

3.7 References

[1] R. Witholt, R. H. Gwiazda and D. R. Smith, Neurotoxicology and Teratology 2000, 22, 851.

[2] N. V. Hue, S. Vega and J. A. Silva, Soil Sci. Soc. Am. J. 2001, 65, 153.

[3] M. P. Colombini and R. Fuoco, Talanta 1983, 30, 901.

[4] P. T. Kissinger and W. R. Heineman, eds., Laboratory Techniques in Electroanalytical Chemistry,

Second Edition, Marcel Dekker, INC, New York, ch24, 1996.

[5] G. Tölg, Angew. Chem. Int. Ed. 1986, 25, 485.

[6] J. Wang, Controlled-Potential Techniques, in Analytical Electrochemistry, Third Edition, John

Wiley & Sons, Inc., Hoboken, NJ, USA.ch3, 2006.

[7] C. R. T. Tarley, V. S. Santos, B. E. L. Baêta, A. C. Pereira and L. T. Kubota, J. Hazard. Mater.

2009, 169, 256.

[8] O. H. Roger J, Anal. Chim. Acta 1982, 140, 51.

[9] J. Barek and J. Zima, Electroanalysis 2003, 15, 467.

[10] E. P. Achterberg and C. Braungardt, Anal. Chim. Acta 1999, 400, 381.

[11] S. P. K. M. A. Nolan, Anal. Chem 1999, 71, 3567.

[12] J. Wang, J. Lu, S. B. Hocevar, P. A. M. Farias and B. Ogorevc, Anal. Chem. 2000, 72, 3218.

[13] J. Wang and J. Lu, Talanta 1995, 42, 331.

[14] N. Abo El-Maali and D. Abd El-Hady, Anal. Chim. Acta 1998, 370, 239.

[15] E. Hrabanko, J. Dolezal and V. Masin, J. Electroanal. Chem. 1969, 22, 195.

[16] J. Labuda, M. Vaníčková and E. Beinrohr, Microchim. Acta 1989, 1, 113.

[17] J. S. Roitz and K. W. Bruland, Anal. Chim. Acta 1997, 344, 175.

[18] A. J. Saterlay, J. S. Foord and R. G. Compton, Analyst 1999, 124, 1791.

[19] J. W. Di and F. Zhang, Talanta 2003, 60, 31.

[20] R. H. Baughman, Science 2000, 290, 1310.

[21] B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas and L. G. Bachas, Science 2004, 303,

62.

[22] G. Dukovic, M. Balaz, P. Doak, N. D. Berova, M. Zheng, R. S. McLean and L. E. Brus, J. Am.

Chem. Soc. 2006, 128, 9004.

[23] S. Huang, L. Dai and A. W. H. Mau, J. Phys. Chem. B 1999, 103, 4223.

[24] R. B. Mathur, S. Seth, C. Lal, R. Rao, B. P. Singh, T. L. Dhami and A. M. Rao, Carbon 2007, 45,

132.

[25] G. G. Tibbetts, G. P. Meisner and C. H. Olk, Carbon 2001, 39, 2291.

[26] M. Pumera, T. Sasaki and H. Iwai, Chem. Asian J. 2008, 3, 2046.

[27] M. U. Niemann, S. S. Srinivasan, A. R. Phani, A. Kumar, D. Y. Goswami and E. K. Stefanakos, J.

Nanomater. 2008, 2008, 950967.

[28] C. P. Jones, K. Jurkschat, A. Crossley, R. G. Compton, B. L. Riehl and C. E. Banks, Langmuir

2007, 23, 9501.

[29] C. E. Banks, A. Crossley, C. Salter, S. J. Wilkins and R. G. Compton, Angew. Chem. Int. Ed.

2006, 45, 2533.

[30] W. Yue, B. L. Riehl, N. Pantelic, K. T. Schlueter, J. M. Johnson, R. A. Wilson, X. Guo, E. E.

King and W. R. Heineman, Electroanalysis 2012, 24, 1039.

[31] J. G. Osteryoung and R. A. Osteryoung, Anal. Chem. 1985, 57, 101A.

[32] L. Lesven, S. M. Skogvold, Ø. Mikkelsen and G. Billon, Electroanalysis 2009, 21, 274.

[33] P. Jothimuthu, R. A.Wilson, J. Herren, E. N. Haynes, W. R.Heineman and I. Papautsky, Biomed.

Microdevices 2011, 13, 695.

Chapter 4| The Application of Nafion Metal Catalyst Free Carbon Nanotube

Modified Gold Electrode: Voltammetric Zinc Detection in Serum

4.1 Introduction

Zinc is an essential trace element that is required in the human body for catalyzing enzyme activity, transcription and protein functions1. Maintaining controlled levels of zinc is critical for multiple physiological functions, including the immune system, insulin action, and anti-oxidant systems2, 3. Lack of zinc can cause immune dysfunction, growth and reproduction problems; also, inflammation and infection are associated with low zinc level in blood2. It has been widely accepted that normal zinc level is necessary to maintain a healthy immune system4, 5. As there are health concerns associated with zinc deficiencies, it is often necessary to analyze complex biological matrices such as blood, serum, tissue, and urine. A variety of methods has been reported in the literature, including inductively coupled plasma – mass spectrometry (ICP-MS) and other spectroscopic and electrochemical methods6-8. Of these, electrochemical methods have advantages associated with high sensitivity, low detection volumes, compact instrumentation, and low cost9, 10.

Carbon nanotubes (CNTs) are an increasingly important group of nanomaterials possessing chemical, physical, and mechanical properties that make them well suited for electrochemical sensing applications11, 12. Nanotubes have high electrical conductivity, are relatively inert chemically, can be used over a wide potential range, and have a very high microscopic surface area due to their surface structure13, 14. CNTs can be prepared in a variety of ways, including arc discharge, laser ablation and chemical vapor deposition15, 16. A hindrance shared by these methods is the general use of metal catalysts in their preparation that can contaminate the CNTs and reduce their effectiveness for stripping voltammetry applications17, 18. Previously, metal

catalyst free carbon nanotubes (MCFCNTs) fabricated via a solid-phase growth mechanism have been reported as a good electrode material with fast electron transfer rate and clear cyclic voltammetry background as well as for trace metal detection by stripping voltammetry with low limit of detection and high sensitivity19, 20. The MCFCNT whisker used in this paper is a powder material with a whisker shape, which is different from the solid CNT electrode that we used previously19, 20. In this work, we continue to explore the potential of MCFCNTs for anodic stripping voltammetry (ASV) by using this whisker material as an electrode modifier for zinc detection in biological samples.

A potential limitation of stripping analyses in complex matrices such as blood is the presence of organic and inorganic substances that may display redox activity near the potential of the analyte of interest, interfering with the detection signal. In addition, organic biological substances that can adsorb on the electrode surface can decrease its effective surface area and passivate (foul) the electrode surface. Also, a large fraction of the zinc in blood or serum is coordinated with binding sites, making it unavailable for detection by ASV21, 22. Thus, a robust analytical procedure must overcome these challenges in order to reliably determine the analyte concentration. A variety of techniques can be used to overcome the matrix interferences, such as

UV irradiation decomposition, ultrasound-assisted extraction and microwave-digestion7, 23, 24. Of these techniques, analyte extraction which physically separates the targeted analyte from the complicated matrix is much simpler since it does not require an external energy supply. A double extraction method using dithizone as a chelating reagent has been reported as an effective method for extracting metal ions from different matrices7, 25-27.

Another approach that applies to electroanalytical techniques is the use of a chemically ion selective electrode coating28. Because faradaic electron transfer occurs at the interface of the

sample with the electrode surface, a material that limits which chemical species are able to reach the electrode can also be an effective means of removing interferences. Nafion is a perfluorinated sulfonated cation-exchanger with the advantages of thermal stability, chemical inertness, mechanical strength and resistance to fouling and has been widely used to modify electrodes in electrochemistry29, 30. Nafion ethanolic solution can also serve as an excellent CNT dispersion solvent that facilitates the electrode modification procedure significantly29.

The aim of this preliminary work is to develop a simple metal based sensor that allows accurate and rapid serum zinc detection using square-wave stripping voltammetry. Previously, our group has developed a lab-on-a-chip sensor with bismuth film working electrode on gold substrate for zinc detection by ASV31. Here we evaluate MCFCNT whiskers as an electrode material by immobilizing them in a Nafion film coated on a gold substrate. This approach has improved the limit of detection of the sensor for zinc detection due to the Nafion/MCFCNT whiskers coating and, more importantly, the new platform worked very well for detection of zinc in a bovine serum sample.

4.2 Experimental

4.2.1 Reagents

All chemicals were purchased without further purification: Nafion perfluorinated resin solution 5 wt. %, acetate buffer solution (pH 4.65) and 1000 ppm Zn standard solution for atomic absorption spectroscopy (AAS) were from Sigma Aldrich; bovine serum was purchased from

Fisher Scientific; MCFCNT whiskers were supplied by SCNTE LLC (Beavercreek, OH) and used without any pretreatment. All other chemicals used in this work were ACS certified reagent

grade and all solutions were prepared with deionized water (18.2 MΩ from Milli-Q System,

Barnstead, MA).

4.2.2 Apparatus

ASV measurements were carried out in a 20 mL conventional three-electrode cell containing

15 mL of sample solution and consisting of a MCFCNT modified gold (Nafion/whiskers-Au) electrode as working electrode, Ag/AgCl as reference electrode (filled with 3 M KCl solution) and Pt wire as auxiliary electrode. A BASi 100B Electrochemical Analyzer from BASi (West

Lafayette, IN) was used as the potentiostat. Basic set-up parameters for Osteryoung square wave voltammetry were square wave amplitude = 25 mV, step potential = 5 mV and frequency = 25

Hz. The sonicator used in this work was FS20D from Fisher Scientific; the centrifuge was

Marathon 6K from Fisher Scientific.

Transmission electron microscopy (TEM) images were taken on a Tecnai F20 at 200 kV high angle annular dark field (HAADF) in scanning TEM mode. Scanning electron microscopy (SEM) images were taken on a FEI XL 20 ESEM in environmental SEM mode. Energy-dispersive X- ray spectroscopy (EDX) images were taken on an EDAX detector from EDAX Inc.

AAS measurements were done using a Varian AA240FS atomic absorption spectrometer.

The extract samples of bovine serum were diluted in 10% HNO3 and analyzed using the parameters specified by the instrument. For serum samples, the same procedure was followed except that we spiked the serum with zinc standard and calculated the recovery.

4.2.3 Preparation of Nafion/whiskers-Au electrode

MCFCNT whiskers powder was dispersed in 1 wt. % Nafion ethanol solution assisted by sonication for 10 min to yield a homogeneous dispersion. A gold electrode was first sonicated for

1 min in distilled water to clean the electrode surface. Then, 10 µL of the MCFCNT whiskers

Nafion dispersion was applied to the dry gold surface. The coating was left to let ethanol evaporate for 30 min and then used directly for ASV measurements.

4.2.4 Extraction procedures

Dithizone (H2Dz) has been used as a highly efficient chelating ligand to extract metals such as silver, zinc and lead from different matrices7, 26, 27. Also, it has been reported that the extraction process can be expedited by using potassium thiocyanate to form a complex with zinc

25, 27 and applying sonication during the extraction . Herein, we use H2Dz as the extraction reagent and thiocyanate ion as auxiliary ligand to extract and quantify zinc in bovine serum with the

ASV method on the Nafion/whiskers-Au electrode. The mechanism is described by the following equations:

- - H2Dz (org) + OH (aq)  H2O + HDz (aq) (1)

- 2+ 2HDz (aq) + Zn (aq)  H2Dz (org) (2)

- - HDz (aq) + Zn(SCN)2 (aq)  Zn(HDz)2 SCN (org) (3)

+ 2+ Zn(HDz)2 (org) + 2H (aq)  Zn (aq) + H2Dz (org) (4)

First, 10 mL (5 mM) dithizone was dissolved in chloroform which gives a dark blue solution as shown in Figure 4.1A. Then it was deprotonated by mixing with 10 mL (pH 9) 1 M ammonia/0.5 M ammonium buffer solution.

- - H2Dz (org) + OH (aq)  HDz (aq) + H2O

(H2Dz)

Then the aqueous layer of deprotonated form of dithizone (Dark orange as shown in Figure

4.1B) was mixed with the solution containing Zn (II) and 0.5 mL of 0.05 M potassium thiocyanate in ethanol and sonicated with 10 ml chloroform for 5 min.

- 2+ 2HDz (aq) + Zn (aq)  Zn(HDz)2 (org)

- - HDz (aq) + Zn(SCN)2 (aq)  Zn(HDz)2 SCN (org)

H N S C N N N Zn H N N N N C S

(Zn (HDz)2)

After sonication, the solution was transferred to a 50 mL plastic tube and centrifuged for 10 min at 4000 rpm to separate the two phases. Zinc dithizone compound is formed and dissolved in the organic chloroform phase which was collected and sonicated with 10 mL 1 M sulfuric acid for another 5 min. In this step, dithizone is protonated and dissolved in chloroform again with

2+ dark blue color and Zn is released from Zn(HDz)2 into the aqueous phase.

+ 2+ Zn(HDz)2 (org) + 2H (aq)  Zn (aq) + 2H2Dz (org)

Then the upper clear aqueous phase was collected and mixed with 0.1 M acetate buffer and the pH was adjusted to 6 with 5 M NaOH solution. The resulting solution was used for ASV measurements with Osteryoung square wave mode.

B A C D E

10 mL 1 M sulfuric acid after 5 min sonication (Zn2+ in the upper aqueous layer and dithizone in bottom chloroform layer)

4.3 Results and discussion

4.3.1 MCFCNT whiskers

Traditionally, transition metals are used as catalyst to synthesize CNTs32. The residual metal oxide cannot be fully removed even by harsh acid wash, which can result in an electrochemical interference such as reduction of the metal oxide in the CNTs33, 34. In addition, batch to batch variation of CNTs in terms of metallic impurity can introduce variation for the application of

CNTs in electrochemistry13, 35. The MCFCNT whiskers used here were synthesized via the carbo thermal carbide conversion method, which grows carbon nanotubes on a silicon carbide matrix35,

36. Since no metal catalyst is used in the synthesis process, batch to batch variability is eliminated in terms of impurities and no post treatment is required. High purity material obtained with this process has been reported as a good material for electrodes19. The TEM images of the MCFCNT whiskers in Figure 4.2 clearly show the whisker shape of the MCFCNT.

Figure 4.2 TEM images of MCFCNT whiskers: A) 4000X magnification; B) 200,000X magnification

4.3.2 Comparison of bare gold, Nafion coated gold (Nafion-Au) and Nafion/whiskers-Au electrodes

In order to separate the effect of the Nafion coating from the MCFCNT whiskers, control experiments for zinc detection were conducted on a bare gold electrode and a Nafion-Au electrode without MCFCNT whiskers. Figure 4.3 shows anodic stripping voltammograms of 4

µM zinc on these different electrodes. A stripping peak for Zn on bare gold electrode occurs at

480 mV. When comparing the bare gold electrode to the Nafion-Au electrode, it can be seen that the Nafion coating causes a small negative shift in the peak potential to 520 mV. Although the coating does not increase the zinc stripping peak current, it improves the electrochemical response in terms of smoother background and better defined peak shape. On both bare gold electrode and Nafion-Au electrode, the zinc stripping peak is at ca. -500 mV, which is shifted to a more positive potential compared to the Nernst potential of 0.98 V (vs. Ag/AgCl) due to the

underpotential deposition (UPD) of depositing zinc on gold substrate37. For the Nafion/whiskers-

Au electrode, a double peak pattern is observed: a major stripping peak at ca. -1050 mV, which is zinc stripping from MCFCNT whiskers, followed by a minor stripping peak at ca. -500 mV, which is zinc stripping from gold. We attribute this to incomplete coverage of the gold surface by MCFCNT whiskers. The height of the “Zn stripping from gold” peak is diminished compared to this peak on bare gold and the Nafion-Au electrodes because most of the gold is now covered by MCFCNT whiskers. The “Zn stripping from MCFCNT whiskers” peak is substantially larger than the “Zn stripping from gold” peak, indicating that it is now the primary surface area for depositing Zn during the deposition step.

0 Bare gold electrode Gold electrode with 1% Nafion -5 Gold electrode with 1% Nafion and MCFCNT whiskers -10

) -15

A 

( -20

-25

-30 Current

-35

-40

-45 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 4.3 ASV voltammograms of 4 µM zinc in acetate buffer solution (pH 6) on different electrodes (bare gold electrode, Nafion-Au gold electrode and Nafion/whiskers-Au electrode).

Deposition time 120 s; deposition potential -1400 mV.

The surface coverage of gold by MCFCNT whiskers can be seen in the scanning electron micrographs in Figure 4.4. Figures 4.4A and 4.4B were obtained at the same surface with different detectors. Figure 4.4A was taken with a gaseous secondary electron detector that shows the surface topology of the Nafion/whiskers-Au electrode where the MCFCNT whiskers coating can be identified by the rough layers. Figure 4.4B, which was taken with a backscattered electron detector, shows the surface composition of the Nafion/whiskers-Au electrode. The darker colored part in Figure 4.4B shows the film of Nafion/whiskers coating which correlates to the ridged part in Figure 4.4A.

Figure 4.4 SEM images of Nafion/whiskers-Au electrode: A) surface topology of

Nafion/whiskers-Au electrode with a gas secondary electron detector; B) surface composition of

Nafion/whiskers-Au electrode with a backscattered electron detector.

Figure 4.5 shows the EDX results for the bright and dark areas in Figure 4.4B as indicated with arrows in Figure 4.4B. Clearly, Figure 4.5A shows a strong Au signal which indicates the exposed gold surface on the Nafion/whiskers-Au electrode (horizontal arrow in Figure 4.4B). On the other hand, the signature gold peak has disappeared in Figure 4.5B, indicating that the gold surface was fully covered by the Nafion/whiskers coating for the dark area (vertical arrow in

Figure 4.4B).

Figure 4.5 EDX images of Nafion/whiskers-Au electrode: A) The bright area (horizontal arrow) in Figure 4.4 B; B) The dark area (vertical arrow) in Figure 4.4 B.

4.3.3 MCFCNT whiskers amount optimization

The effect of amount of MCFCNT whiskers in the Nafion film coated on the gold electrode was investigated. As shown in Figure 4.6, the peak current for zinc stripping at -1050 mV from the MCFCNT whiskers increases almost linearly with increasing amount of the MFCNT whiskers dispersion up to about 3 mg/mL above which it levels off. Also, as more CNT is coated on the gold surface, the zinc peak at -500 mV gets smaller as more of the gold is covered. As shown by the error bars, reproducibility was much poorer at concentrations above 3 mg/mL. The thicker whiskers films are less mechanically stable and can be disrupted by hydrogen bubbles formed by electrolysis of water during the Zn deposition step, which causes larger variation38.

Overall, 3 mg/mL gives a good compromise for zinc ASV detection with respect to both peak height and reproducibility. Therefore, the optimized film composition of 10 µL of 3 mg/mL

MCFCNT whiskers in 1 wt. % Nafion ethanol solution was used for all subsequent experiments.

)

 ( 8

4 Peakcurrent

0 0 1 2 3 4 5 6 CNT concentration (mg/mL)

Figure 4.6 Effect of CNT concentration on peak current for zinc stripping voltammetry. Coating:

10 µL of MCFCNTs whisker dispersion in 1 wt. % Nafion ethanolic solution with different CNT concentrations. Deposition time 120s, deposition potential -1400 mV; zinc concentration 5 µM.

Between measurements, the electrode was cleaned by applying a constant potential of 500 mV for 1 min to remove any zinc residue from the electrode surface. A single electrode can give reproducible results for 50 - 80 measurements. The relative standard deviation of zinc detection

(4 µM in acetate buffer) with different Nafion/whiskers-Au electrodes (N=5) was found to be

5.1 %.

4.3.4 pH, deposition potential and deposition time optimization

Electrolysis of water is an important factor to consider for ASV detection of zinc since hydrogen evolution can interfere with zinc deposition on the electrode surface because of the very negative potential required. Hydrogen evolution is quite sensitive to solution pH, deposition potential and electrode material. Figure 4.7A shows cyclic voltammograms of the

Nafion/whiskers-Au electrode in acetate buffer with different pH values. Electrolysis of water starts at ca. -1000 mV at pH values lower than 6.0, which will interfere with zinc deposition for

ASV measurements at ca. -1300 mV.

Figure 4.7B shows how buffer pH affects zinc ASV detection at the Nafion/whiskers-Au electrode. The ASV peak currents are lower and with larger variations at lower pH values compared with pH 6.0. Overall, ASV at pH 6 buffer has the most moderate background and the most reproducible peak current since significant hydrogen evolution occurs at lower pH. Also,

ASV has the sharpest and highest stripping peak at a deposition potential of -1400 mV. The peak current drops at more negative deposition potentials due to hydrogen gas evolution at the electrode surface, which interferes with the mass transport of zinc analyte to the electrode surface for deposition and damages the Nafion/whiskers film.

1200 A 1000 pH 4.5 800 pH 5.0 pH 5.5 pH 6.0

) 600

A  ( 400

200

Current 0

-200

-400

1500 1000 500 0 -500 -1000 -1500 Potential (mV)

Figure 4.7 A) Cyclic voltammograms of Nafion/whiskers-Au electrode in acetate buffer with different pH values; scan rate: 10 mV / s;

17.5 B

17.0

) 16.5

 ( 16.0

15.5

15.0 Peakcurrent

14.5

4.5 5.0 5.5 6.0 pH

Figure 4.7 B) pH optimization of Zn2+ on Nafion/whiskers-Au electrode in acetate buffer with optimized coating parameters from pH 4.5 to 6. Zinc concentration 5 µM; deposition potential -

1400 mV; deposition time 120 s.

Figure 4.8 shows the deposition potential optimization of ASV zinc measurement at the optimized pH value. A potential of -1400 mV was chosen as optimized deposition potential due to the highest peak height and acceptable reproducibility for the Zn stripping peak at -1050 mV.

Thus, pH 6 and a deposition potential of -1400 mV were used for subsequent experiments. The optimization of deposition time was also explored with optimized CNT concentration, buffer PH and depositon potential. It was found that the peak current increases linearly with deposition time up to 120 s after which the curve slope starts to decrease as coverage of the CNTs on the

electrode surface increases with longer deposition time. So, 120 s was selected as optimized deposition time.

)

A  ( 6

4 Peakcurrent

-1250 -1300 -1350 -1400 -1450 -1500 Deposition potential (mV)

Figure 4.8 Deposition potential optimization with optimized coating parameters: deposition time

120 s; zinc concentration 5 µM; pH 6.0

4.3.5 Calibration data

ASV determination of Zn2+ in a series of standard solutions was carried out in 0.1 M pH 6 acetate buffer solution using optimized experimental parameters. Figure 4.9 shows ASV voltammograms in the linear range of zinc detection with increasing concentrations from 0.5 to 7

µM. The double peak pattern is also seen in the voltammogram and the major zinc peak at -1050 mV was used for zinc stripping peak measurements. Each voltammogram represents three replicates of zinc measurements for the same concentration.

-10

-20

)

A 

( -30 Background 0.5 M -40 1.0 M 1.5 M Current 2.0 M 3.0 M -50 4.0 M 5.0 M 6.0 M 7.0 M -60 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

A plot of peak current versus concentration is shown in Figure 4.10. The correlation equation for the linear range is I (µA) = (3.77±0.03) * [Zn (µM)] - (1.05±0.06) (µA) (R2=0.99 for 9 concentrations in the range of 0.5- 7 µM and the standard deviations are shown for the

slope and intercept). The limit of detection was calculated to be 53 nM (based on 3σ/slope) which is close to the 50 nM that we reported previously for a MCFCNT electrode for zinc detection19. Measurements of higher concentrations fit into another linear range which is attributed to stripping of zinc that was deposited on zinc rather than the zinc deposited directly on MCFCNT whiskers.

35 1% Nafion and MCFCNT whiskers coated gold electrode 1% Nafion coated gold electrode 30

) 25

 ( 20

10 Peakcurrent 5

0 0 1 2 3 4 5 6 7 8 9 10 11 Concentration (M)

Figure 4.10 Dynamic ranges of zinc ASV measurements in the range of 0.5 µM–11 µM on

Nafion-Au electrode and Nafion/whiskers-Au electrode. Deposition time 120 s; deposition potential -1400 mV.

The bare gold electrode has a dissatisfactory behavior for zinc detection in terms of poor reproducibility, non-linear calibration and distorted background. After coating with 10 µL 1 wt. %

Nafion solution, the Nafion-Au electrode showed excellent reproducibility for zinc detection and a linear range of 0.5 - 4 µM was obtained as shown in Figure 4.10. For concentrations higher than 4 µM, the deposition follows a zinc-on-zinc deposition mechanism which leads to a smaller slope value. The limit of detection (based on 3σ) calculated for the Nafion-Au electrode within the first linear range was 80 nM. Notably, the Nafion-Au electrode has a smoother background and a better defined zinc peak than bare gold electrode at the same current scale. The

Nafion/whiskers-Au electrode exhibits 2.4 times increase in peak current compare to the Nafion-

Au electrode which is attributed to the increased electrode surface area provided by the

MCFCNT whiskers and the fast electron transfer rate at MCFCNT whiskers. Compared with the

Nafion-Au electrode, the Nafion/whiskers-Au electrode has a much wider linear range, larger calibration curve slope and bigger peak current (shown in Figure 4.9). Also, the limit of detection of zinc for the Nafion/whiskers-Au electrode is lower than for the Nafion-Au electrode. In this comparison, all of the experimental parameters were the same optimized values described above.

Comparison of zinc detection with different CNT-based electrodes is shown in Table 4.1. As shown in Table 4.1, the combination of CNT, Nafion film and bismuth film has the lowest limit of detection and largest linear range. The Hg film CNT electrode has a similar linear range to the

MCFCNT electrode with a higher limit of detection. One obvious advantage of this work is that there is no extra film needed to be co-deposited (e.g., Hg and Bi film39, 40) on the electrode which allows an easier fabrication process.

Sensor Deposition time (s) Limit of detection Linear range (µM)

(µM)

MWCNTs/Hg film40 180 0.43 0.89 - 9.9

MWCNTs/NA/Bi/SPE39 120 0.0046 0.0077 - 1.5

This work 120 0.053 0.5 - 7

Table 4.1 Comparison of different CNT based electrodes for ASV Zn2+ detection

4.3.6 Sample preparation and standard addition of extraction from bovine serum

The extraction efficiency was first examined by extracting zinc from aqueous zinc AAS standard solution. Table 4.2 shows the extraction efficiency of various ligand to metal ratios in the range 10:1, 50:1 and 100:1. A ligand to metal ratio of 100:1 is needed for high extraction yields. Bovine serum samples were treated with the procedure described in the Experimental

Section. The resulting solution was adjusted to pH 6 and quantified by ASV with a

Nafion/whiskers-Au electrode using the standard addition method. The low concentration of zinc caused us to use a longer deposition time of 5 min to obtain a better defined ASV peak for zinc detection in the serum extracts. Figure 4.11A shows the stripping voltammograms for a bovine serum extract and standard additions from 1 - 4 µM. The zinc stripping peak at ca. -500 mV is not changed linearly by the addition os zinc due to the limited exposed Au surface area. Also, shoulder peaks are present in Figure 4.11A at ca. -750 mV which were not present in Figure 4.9.

These shoulder peaks are possibly due to the presence of some residual organic species which can alter the ASV behavior of heavy metals41.

[L]:[Zn2+] ratio Extraction efficiency (%)

10:1 68±1

50:1 83±2

100:1 97±1

Table 4.2 Efficiency of extraction from zinc stock solution with different [L]: [Zn2+] ratios

-10 A

-20

) -30

 ( -40 Serum extraction Serum extraction+1 M zinc addition

Current -50 Serum extraction+2 M zinc addition Serum extraction+3 M zinc addition Serum extraction+4 M zinc addition -60

0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 4.11A) Voltammograms of standard addition of bovine serum extract for ASV zinc detection with Nafion/whiskers-Au electrode.

) 12

 (

6 Peakcurrent 3

0 0 1 2 3 4 5 Concentration (M)

Figure 4.11B) Standard addition plot of bovine serum extract for ASV zinc detection with

Nafion/whiskers-Au electrode. Deposition time 300 s; deposition potential -1400 mV; pH 6.

Figure 4.11B shows the plot [I (µA) = (3.24±0.15) * [Zn (µM)] + (2.54±0.33) (µA)]

(standard deviations are shown for the slope and intercept) for standard addition. The plot slope is higher than the 2.38 µA/µM that we reported before for zinc detection with a MCFCNT electrode19. The sensitivity increase is due to the Nafion film with MCFCNT whiskers coating.

The zinc concentration in bovine serum was tested to be 45.1±0.6 µM which is in close agreement with an independent AAS result of 48.2 ±0.5 µM.

4.4 Conclusions

For the first time, MCFCNT whiskers are used as an electrode coating material confined in a

Nafion film coated on a gold electrode. This novel material has advantages over the traditional carbon nanotube material such as low cost, high purity and good consistency from batch to batch.

MCFCNT whiskers are homogeneously dissolved in 1 wt. % Nafion ethanolic solution and coated onto the gold electrode surface. The Nafion/whiskers-Au electrode shows better reproducibility and lower limit of detection for zinc detection in acetate buffer than both Au and

Nafion-Au electrodes. A double peak pattern was observed for zinc detection with

Nafion/whiskers-Au electrode due to UPD on exposed gold. Previously, multi-walled carbon nanotubes (MWCNTs) have been used as an electrode modifier for trace level voltammetric zinc detection. The double extraction method used in this work to extract zinc from bovine serum was optimized to give a high yield of zinc extraction. The Nafion/whiskers-Au electrode works very well for zinc detection in bovine serum extracts, and the results are in good agreement with independent AAS measurements. This work shows the good potential of using MCFCNT materials together with a gold substrate for a sensing system that measures zinc in blood.

4.5 Acknowledgement

The authors gratefully acknowledge NIEHS R21ES019255 for financial support, SCNTE

LTD for providing the MCFCNT whiskers, and Dr. Necati Kaval for technical support on the

SEM and EDX.

4.6 References

1. D. K. Heyland, N. Jones, N. Z. Cvijanovich, H. Wong, J. Parenter. Enteral. Nutr. 2008, 32, 509-

519.

2. B. Y. Besecker, M. C. Exline, J. Hollyfield, G. Phillips, R. A. DiSilvestro, M. D. Wewers, D. L.

Knoell, Am. J. Clin. Nutr. 2011, 93, 1356-1364.

3. K. H. Ibs, L. Rink, J. Nutr. 2003, 133, 1452S-1456S.

4. H. R. Wong, T. P. Shanley, B. Sakthivel, N. Cvijanovich, R. Lin, G. L. Allen, N. J. Thomas, A.

Doctor, M. Kalyanaraman, N. M. Tofil, S. Penfil, M. Monaco, M. A. Tagavilla, K. Odoms, K. Dunsmore,

M. Barnes, B. J. Aronow, S. S. S. Genomics Pediat, Physiol. Genomics 2007, 30, 146-155.

5. T. P. Shanley, N. Cvijanovich, R. Lin, G. L. Allen, N. J. Thomas, A. Doctor, M. Kalyanaraman,

N. M. Tofil, S. Penfil, M. Monaco, K. Odoms, M. Barnes, B. Sakthivel, B. J. Aronow, H. R. Wong, Mol.

Med. 2007, 13, 495-508.

6. C. S. Muniz, J. M. Larchante-Gayon, J. I. G. Alonso, A. Sanz-Medel, J. Anal. At. Spectrom. 1999,

14, 193-198.

7. J. Kruusma, C. E. Banks, L. Nei, R. G. Compton, Anal. Chim. Acta 2004, 510, 85-90.

8. M. Ghaedi, A. Shokrollahi, K. Niknam, E. Niknam, S. Derki, M. Soylak, J. AOAC Int. 2009, 92,

907-913.

9. J. Wang, Practical Considerations. Analytical Electochemistry, John Wiley & Sons, Inc: 2002; p

115-163.

10. G. Tölg, Angew. Chem. Int. Ed. 1986, 25, 485-485.

11. R. H. Baughman, Science 2000, 290, 1310-1311.

12. B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, L. G. Bachas, Science 2004, 303, 62-

65.

13. X. Ji, R. O. Kadara, J. Krussma, Q. Chen, C. E. Banks, Electroanalysis 2010, 22, 7-19.

14. K. Wu, S. Hu, J. Fei, W. Bai, Anal. Chim. Acta 2003, 489, 215-221.

15. S. Huang, L. Dai, A. W. H. Mau, J. Phys. Chem. B 1999, 103, 4223-4227.

16. R. B. Mathur, S. Seth, C. Lal, R. Rao, B. P. Singh, T. L. Dhami, A. M. Rao, Carbon 2007, 45,

132-140.

17. G. G. Tibbetts, G. P. Meisner, C. H. Olk, Carbon 2001, 39, 2291-2301.

18. M. U. Niemann, S. S. Srinivasan, A. R. Phani, A. Kumar, D. Y. Goswami, E. K. Stefanakos, J.

Nanomater. 2008, 2008, 950967.

19. W. Yue, B. L. Riehl, N. Pantelic, K. T. Schlueter, J. M. Johnson, R. A. Wilson, X. Guo, E. E.

King, W. R. Heineman, Electroanalysis 2012, 24, 1039-1046.

20. W. Yue, A. Bange, B. L. Riehl, B. D. Riehl, J. M. Johnson, I. Papautsky, W. R. Heineman,

Electroanalysis 2012, 24, 1909-1914.

21. J. C. Ebert, R. B. Altman, Protein Sci. 2008, 17, 54-65.

22. I. L. Alberts, K. Nadassy, S. J. Wodak, Protein Sci. 1998, 7, 1700-1716.

23. B. Liu, L. Lu, M. Wang, Y. Zi, Electroanalysis 2008, 20, 2363-2369.

24. E. Šucman, M. Mahrová, J. Páč, M. Vávrová, P. Kettisch, Electroanalysis 2008, 20, 386-389.

25. Y. S. Kim, G. In, J. M. Choi, C. W. Lee, Bull. Korean Chem. Soc. 2000, 21, 855-859.

26. J. L. Manzoori, G. Karim-Nezhad, Anal. Chim. Acta 2003, 484, 155-161.

27. J. Kruusma, P. Tomčík, C. E. Banks, R. G. Compton, Electroanalysis 2004, 16, 852-859.

28. O. Niwa, M. Morita, H. Tabei, Electroanalysis 1994, 6, 237-243.

29. W. S. Huang, C. H. Yang, S. H. Zhang, Analytical and Bioanalytical Chemistry 2003, 375, 703-

707.

30. J. Wang, M. Musameh, Y. H. Lin, Journal of the American Chemical Society 2003, 125, 2408-

2409.

31. P. Jothimuthu, R. Wilson, J. Herren, E. Haynes, W. Heineman, I. Papautsky, Biomed.

Microdevices 2011, 13, 695-703.

32. S. Fan, M. G. Chapline, N. R. Franklin, T. W. Tombler, A. M. Cassell, H. Dai, Science 1999, 283,

512-514.

33. X. Ji, R. O. Kadara, J. Krussma, Q. Y. Chen, C. E. Banks, Electroanalysis 2010, 22, 7-19.

34. M. Pumera, Langmuir 2007, 23, 6453-6458.

35. C. E. Banks, A. Crossley, C. Salter, S. J. Wilkins, R. G. Compton, Angew. Chem. Int. Ed. 2006,

45, 2533-2537.

36. C. P. Jones, K. Jurkschat, A. Crossley, R. G. Compton, B. L. Riehl, C. E. Banks, Langmuir 2007,

23, 9501-9504.

37. V. Sudha, M. V. Sangaranarayanan, J. Chem. Sci. 2005, 117, 207-218.

38. A. Krolicka, A. Bobrowski, A. Kowal, Electroanalysis 2006, 18, 1649-1657.

39. L. Fu, X. Li, J. Yu, J. Ye, Electroanalysis 2013, 25, 567-572.

40. J. Wang, C. Bian, J. Tong, J. Sun, S. Xia, Electroanalysis 2012, 24, 1783-1790.

41. M. Boussemart, L. Menargues, J. Y. Benaïm, Electroanalysis 1993, 5, 125-133.

Chapter 5| Kinetic Study of Voltammetric Behavior of Heavy Metals by

Osteryoung Squarewave Voltammetry on a Metal Catalyst Free Carbon

Nanotube Electrode

5.1 Introduction

Anodic stripping voltammetry (ASV) has been well known for many years as a powerful technique for detecting diverse trace heavy metal ions1, 2. In the first decades, mercury-based electrodes have been used for their excellent performance with stripping voltammetry. The hanging mercury drop electrode (HMDE) was one of the widely used mercury-based electrodes due to its amenable electrode surface and good performance with stripping voltammetry3. The main advantages of the HMDE are the low residual current, the excellent negative potential range and the fresh electrode surface formed for each measurement. And using mercury as the working electrode for ASV is that many metals dissolve in mercury to form amalgams. However, the fatal point of mercury-based electrodes is their toxicity and therefore, they are no longer favored and many substitutes are used to replace mercury-based electrodes4-6. Since then, various carbon-based materials, such as glassy carbon, edge plane pyrolytic graphite, carbon paste, and carbon nanotube, have been widely used as electrode materials in electrochemical applications, especially for ASV7, 8. Since its discovery9, the carbon nanotube has emerged as a good candidate for electrochemistry among these carbon materials due to its wide potential window, fast electron transfer rate and large surface area10-12.

Lead, cadmium and zinc are three major heavy metals being detected by ASV due to their extensive use and their hazardous effects to both humans and other organisms13. The detections of the above metals are satisfactory by ASV with low limit of detection and good sensitivity.

However, there is one experimental phenomenon always associated with the stripping voltammograms for the detection of these three metals: for equimolar concentrations of these

analytes, it always follows the same pattern that lead has the largest ASV peak, followed by cadmium, and zinc has the lowest response. So far, the reason for this phenomenon has not been addressed by researchers. In this paper, we explore the explanation for this commonly observed, yet poorly understood phenomenon by electrochemistry methods.

5.2 Experimental

5.2.1 Reagents and instrumentation

Acetate buffer was purchased from Aldrich; 1000 ppm lead stock solution, 1000 ppm cadmium stock solution and 1000 ppm zinc stock solution were purchased from Fisher Scientific.

Metal catalyst free carbon nanotube (MCFCNT) electrodes (geometric surface area 0.066 cm2) were supplied and fabricated with carbo thermo carbide conversion method by SCNTE LLC

(Beavercreek, OH). Glassy carbon electrode (surface area 0.066 cm2) was purchased from BASi and polished with alumna before each measurement. All the chemicals were ACS grade and electrodes were used without further purification. Lead, cadmium and zinc working standards were prepared by diluting the stock solution from Fisher Scientific with deionized water

(resistivity 18.2 MΩ from Milli-Q Integral Water Purification System). All measurements were conducted on solutions in a 20 ml three electrode cell using a BASi 100B Electrochemical

Analyzer to obtain the stripping voltammograms. A three-electrode system was employed that consisted of either MCFCNT electrode or glassy carbon electrode as working electrode,

Ag/AgCl as reference electrode from BASi (filled with 3M KCl filling solution), Pt wire as auxiliary electrode.

5.3 Results and discussion

5.3.1 Electron transfer rate study of the MCFCNT electrode

5 mM of Potassium ferrocyanide (K3Fe(CN)6) was used as a benchmark because of its quasi- reversibility on carbon electrodes to explore the electron transfer rate on the MCFCNT electrode surface. Cyclic voltammetry (CV) measurements were performed on a glassy carbon, which is a commonly used working electrode material, to make a comparison with the MCFCNT electrode under the same experimental conditions, and 0.5 M KCl was used as supporting electrolyte. The theoretical peak-to-peak separation (ΔEp) for ferro-ferricyanide redox reaction is 59 mV. The closer the ΔEp is to 59 mV, the faster the electron transfer kinetics on the electrode. As shown in

Figure 1, the peak-to-peak separation (ΔEp) for the cyclic voltammogram on the MCFCNT electrode is 60 mV compared to 200 mV on the glassy carbon electrode, which indicates faster electron transfer rate at the MCFCNT electrode surface than at glassy carbon for ferricyanide.

Due to the high microscopic surface area on CNT electrode, both anodic and cathodic peak density are higher at CNT electrode than at GC electrode surface14.

150 MCFCNT electrode Glassy carbon electrode

100

) 50

 (

0 Current -50

-100

700 600 500 400 300 200 100 0 -100 -200 Potential (mV)

Figure 5.1 Cyclic voltammograms of 5 mM K4Fe(CN)6 in 1 M KNO3 solution on MCFCNT electrode and glassy carbon electrode, scan rate 100 mV/s.

5.3.2 Cyclic voltammetry of lead, cadmium and zinc

Lead, cadmium and zinc are three major heavy metals that have been widely detected by

ASV. 1 mM solutions of each metal analyte were made separately with 0.1 M pH 4.5 acetate buffer solution as supporting electrolyte and characterized by CV on MCFCNT electrode. Figure

2 shows cyclic voltammograms of the three analytes with various scan rates. The sweeping potential range is between 0 to -1350 mV, since -1350 mV was reported as optimized potential for metal deposition on the CNT electrode15. As shown in Figure 2A and 2B, the magnitude of the peak currents increased with increasing scan rate and the shapes of the CV waves were unchanged. However, as for Zn (shown in Figure 2C), the reduction peak becomes less

discernible as the scan rate increases and a low scan rate (less than 25 mV/s) is required to obtain a discernible Zn reduction peak. In addition, with respect to ΔEp, the ΔEp values are 115 mV, 130 mV and 180 mV for lead, cadmium and zinc at a scan rate of 25 mV/s, respectively. The much larger ΔEp value for Zn indicates the poor reversibility of Zn on the MCFCNT electrode.

100 A

)

A  ( -50

-100 Current

-150 25 mV/s 50 mV/s 75 mV/s -200 100 mV/s 150 mV/s -250 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.2 A) Cyclic voltammograms of Pb2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1 mM.

250 B 200

150

100

) A

 50 (

-50 Current -100 25 mV/s 50 mV/s -150 75 mV/s 100 mV/s -200 150 mV/s

0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.2 B) Cyclic voltammograms of Cd2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1

220 C 200 180 5 mV/s 160 10 mV/s 140 25 mV/s 50 mV/s 120 ) 75 mV/s

A 100  ( 80 60 40

Current 20 0 -20 -40 -60 -80 -600 -700 -800 -900 -1000 -1100 -1200 -1300 -1400 Potential (mV)

mM.

Figure 5.2 C) Cyclic voltammograms of Zn2+ in 0.1 M acetate buffer at MCFCNT electrode with different scan rates, Concentration: 1 mM.

5.3.3 ASV study of lead, cadmium and zinc

Osteryoung square wave (OSW) mode was used for ASV which results the net current by subtracting reverse current from forward current16. During the deposition step, the solution was stirred so the total amount of analyte being deposited on the electrode surface is determined by the diffusion coefficient of metal ions, deposition time and the kinetics between the analyte and the electrode. The diffusion coefficients for lead, cadmium and zinc are listed in Table 1.

Table 5.1. Diffusion coefficient at infinite dilution for lead, cadmium and zinc 17

Ion Pb2+ Cd2+ Zn2+

Diffusion coefficient (10-5cm2s-1) 0.945 0.719 0.703

Further inspection of the forward and reverse currents is necessary to get better understanding of the mechanism. Figure 3A-3C are ASV voltammograms of lead, cadmium and zinc measured under the same experimental conditions measured by Osteryoung squarewave mode (step size 5 mV, frequency 15 Hz and amplitude 25 mV) and shown with forward current, reverse current and net current.

100 Pb 75 Net current Forward current 50 Reverse current

) 25

 ( 0

-25 Current -50

-75

-100 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.3 A) Anodic stripping voltammograms of Pb2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s;

Deposition potential: -1350 mV

200 Cd

150 Net current Forward current Reverse current

100

)

A  ( 50

0 Current

-50

-100 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.3 B) Anodic stripping voltammograms of Cd2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s;

Deposition potential: -1350 mV

100 Zn

80 Net current Forward current

60 Reverse current

) A

 40 (

Current 0

-20

-40 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.3 C) Anodic stripping voltammograms of Zn2+ in 0.1 M acetate buffer with forward current, reverse current and net current; Concentration: 10 µM; Deposition time: 150 s;

Deposition potential: -1350 mV

The result of equal molar heavy metals shows the same trend as previous reports15: lead has the largest stripping peak current followed by cadmium, and zinc has the smallest peak current.

Obviously, OSW mode detection of lead and cadmium has reverse current with opposite sign from the forward current which increases the net current. On the other hand, the reverse current for zinc does not show up at the same potential which as a result cannot help to increase the net current with small amplitude value.

100

A mercury film carbon nanotube (MFCNT) electrode was utilized to run control experiments to study the electron transfer kinetics. The mercury film was coated onto a CNT electrode by reducing 1 µM Hg(Cl)2 solution in 0.1 M acetate buffer for 1 min at -1 V with solution stirred .

Then, the same CV and ASV experiments on zinc were done at the mercury film electrode.

Figure 4A shows closer ΔEp between the reduction and the oxidation peaks for zinc on the

MFCNT electrode compared with bare the CNT electrode zinc CV in Figure 2C, which implies

Zn(Hg)/Zn2+ is a more electrochemically reversible couple on mercury electrode than on the bare

CNT electrode. The ASV of zinc on the mercury film electrode is shown in Figure 4B in which reverse current has an opposite sign as opposed to the forward current and to increase net current signal.

200 A

150

100

)

A 

( 50

Current -50

-100

-150

-600 -700 -800 -900 -1000 -1100 -1200 -1300 Potential (mV)

Figure 5.4 A) Cyclic voltammograms of Zn2+ in 0.1 M acetate buffer at MFCNT electrode,

Concentration: 1 mM, scan rate 25 mV/s

101

400 B

200

) 0

 ( -200

-400 Current Net current Reverse current -600 Forward current

-800

-200 -400 -600 -800 -1000 -1200 -1400 Potential(mV)

The Osteryoung square wave mode employs a pulsing amplitude which enables the reverse pulses to redeposit metal onto the electrode surface that enhances the signal. From the cyclic voltammogram of zinc, because the oxidation peak and reduction peak are separated, this small pulse amplitude would not be able to reduce metal ion in the reverse pulse when the potential is swept positively through the oxidation potential of zinc. To understand this better, an amplitude study of the reverse current of Osteryoung square wave ASV for zinc was performed. Figure 5A shows the result of this study at a CNT electrode. For amplitudes larger than 150 mV, reverse currents start to have an opposite sign with forward currents which agrees with the peak-to-peak

102

separation in CV. Also, step size and frequency were also studied as two important variables that affect the reverse current. Figure 5B and 5C show that with amplitude of 150 mV, the peak current in the reverse scan has an opposite sign with the forward peak current with various scan frequency and step size values.

400 25 mV A 50 mV 350 100 mV 150 mV 300 200 mV

) 250

 ( 200

150 Current

100

0 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.5 A) Reverse currents for anodic stripping voltammograms at a CNT electrode for 10

µM Zn in 0.1 M acetate buffer (pH 4.5), deposition time: 150 s, amplitudes from 25 mV to 200 mV

103

400 10 Hz B 15 Hz 350 25 Hz 30 Hz 300 35 Hz 40 Hz ) 250

A 50 Hz

 ( 200

150 Current

100

0 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

104

450

C 400

350

) 300

 ( 250

200 Current 150

100

0 -200 -400 -600 -800 -1000 -1200 -1400 Potential (mV)

Figure 5.5 C) Reverse currents for anodic stripping voltammograms at CNT electrode for 10 µM

Zn in 0.1 M acetate buffer (pH 4.5), deposition time: 150 s; frequency: 25 mV; amplitude: 200 mV; step size: 5 mV, 8 mV, 10 mV, 12 mV, 15 mV

5.3.4 Comparison of ASV results between two sets of parameters on zinc

In order to lower the limit of detection for zinc on MCFCNT electrode, different sets of parameters (listed in Table 2) were used based on the research discussed above. Figure 6A and

6B show two series of stripping voltammograms with two sets of different parameters. Clearly, the bigger pulse size not only increases the reverse current, but also increases the net current.

This inevitably introduces much variation in terms of peak current and background current. Thus,

105

the limit of detection of zinc is not much improved due to the large current variation when calculating the limit of detection base on the 3σ method.

-26 A Background 2 M -28 3 M 4 M 5 M

) -30

6 M

A 

( 7 M -32

-34 Current -36

-38

-40 -600 -700 -800 -900 -1000 -1100 -1200 -1300 Potential (mV)

Figure 5.6 A) Voltammograms of ASV of zinc in the range from 2 µM to 7 µM in 0.1 M acetate buffer (pH 4.5) with default ASV parameters;

106

-200 B

-225

) -250

 (

-275 Background 2 M 3 M

Current -300 4 M 5 M 6 M -325 7 M

-350 -600 -700 -800 -900 -1000 -1100 -1200 -1300 Potential (mV)

Figure 5.6B Voltammograms of ASV of zinc in the range from 2 µM to 7 µM in 0.1 M acetate buffer (pH 4.5) with optimized ASV parameters; deposition time: 60 s.

Table 5.2. Comparison between the two sets of parameters

Step size (mV) Frequency (Hz) Amplitude (mV) Limit of detection (µM)

Default 5 15 25 0.4

Optimized 5 25 150 0.3

107

5.4 Conclusion

Cyclic voltammetry for Pb2+, Cd2+ and Zn2+ under the the same experimental conditions shows an electron transfer rate difference for the three analytes. This difference in kinetics is the key factor leading to the difference in the magnitude of the peak current obtained by anodic stripping voltammetry behavior. The Osteryoung square wave mode was used to study the kinetic behavior of anodic stripping voltammetry of these metals. The forward and reverse current study confirms that the anodic stripping voltammetry behavior is determined mostly by the intrinsic kinetic property of the analyte. The large ΔEp value of zinc in cyclic voltammetry shows the inability to redeposite zinc during the reverse pulse and hnce decreases the the reverse current. By increasing the pulsing amplitude, a reduction peak for Zn2+ in the reverse current can be achieved. Although this increases the currents, it does not improve the limit of detection for zinc when calculated by the 3σ method due to the large current variation.

5.5 Acknowledgement

The author gratefully acknowledges NIEHS R21ES019255 for financial support and SCNTE

LTD for providing the MCFCNT electrode.

108

5.6 Reference

1. G. Tölg, Angew. Chem. Int. Ed. 1986, 25, 485-485.

2. J. Wang, Controlled-Potential Techniques, in Analytical Electrochemistry, Third Edition. Third ed.; John Wiley & Sons, Inc., Hoboken, NJ, USA.ch3: 2006; Vol.

3. J. Wang, Practical Considerations. Analytical Electochemistry, John Wiley & Sons, Inc: 2002; p

115-163.

4. J. Wang, B. Tian, Anal. Chem. 1993, 65, 1529-1532.

5. E. P. Achterberg, C. Braungardt, Anal. Chim. Acta 1999, 400, 381-397.

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

7. X. Ji, R. O. Kadara, J. Krussma, Q. Chen, C. E. Banks, Electroanal. 2010, 22, 7-19.

8. C. E. Banks, R. G. Compton, Anal. Sci. 2005, 21, 1263-1268.

9. S. Iijima, Nature 1991, 354, 56-58.

10. R. H. Baughman, Science 2000, 290, 1310-1311.

11. G. Dukovic, M. Balaz, P. Doak, N. D. Berova, M. Zheng, R. S. McLean, L. E. Brus, J. Am. Chem.

Soc. 2006, 128, 9004-9005.

12. B. J. Hinds, N. Chopra, T. Rantell, R. Andrews, V. Gavalas, L. G. Bachas, Science 2004, 303, 62-

65.

13. M. Rajabi, A. Asghari, H. Zavvar Mousavi, J. Anal. Chem. 2009, 65, 511-517.

14. T. Wang, H. D. Manamperi, W. Yue, B. L. Riehl, B. D. Riehl, J. M. Johnson, W. R. Heineman,

Electroanalysis 2013, 25, 983-990.

15. W. Yue, B. L. Riehl, N. Pantelic, K. T. Schlueter, J. M. Johnson, R. A. Wilson, X. Guo, E. E.

King, W. R. Heineman, Electroanalysis 2012, 24, 1039-1046.

16. J. G. Osteryoung, R. A. Osteryoung, Analytical Chemistry 1985, 57, 101A-110A.

17. P. Vanysek, CRC Handbook of Chemistry and Physics 1999, 83.

109

Chapter 6| Conclusions

110

Conclusions

This dissertation has demonstrated several applications in the electrochemical characterization of metal catalyst free carbon nanotubes. Several heavy metals in trace level have been tested using stripping voltammetry method. The novel electrode provides excellent performance on both single and simultaneous detection of Pb2+, Cd2+ and Zn2+. The deposition time, deposition potential, conditioning time and conditioning potenial were all optimized. Tap water sample detection also confirms that MCFCNTs electrode is reliable in heavy metal detection in natural water samples. The MCFCNTs electrode shows comparable limit of detection with other mercury free electrodes, such as: CNTs modified electrodes and bismuth film electrode.

Also, this dissertation evaluated two stripping voltammetry methods for detecting Mn2+ in aqueous solutions using MCFCNTs electrodes. The ASV method was shown to have a narrow linear range of detection and poor stripping peak shape. The CSV method that was explored as an alternative shows a lower limit of detection and better sensitivity. The CSV method was then shown to be a very reproducible and reliable technique for the detection of Mn2+, and robust enough to operate in a sample such as pond water. The MCFCNT electrode for CSV detection of

Mn has a low limit of detection, wide linear dynamic range and good reproducibility. Previously, our lab has designed a bismuth film chip which was used for ASV detection of Mn in blood samples. However, due to the high electronegativity of Mn, the limit of detection of Mn was only

5 µM. Since MCFCNT has shown a promising CSV result for Mn detection with a limit of detection of only 93nM, it can potentially be a coating material on our chip to provide a significantly better limit of detection than was obtained on a bismuth electrode.

111

For the first time, MCFCNT whiskers are used as an electrode coating material confined in a

Nafion film coated on a gold electrode. This novel material has advantages over the traditional carbon nanotube material such as low cost, high purity and good consistency from batch to batch.

Nafion-Au electrodes. A double peak pattern was observed for zinc detection with

112

be achieved. Although this increases the currents, it does not improve the limit of detection for zinc when calculated by the 3σ method due to the large current variation.

Because other than the MCFCNTs there are no other additives added to the solution, this can be considered as an environmental friendly material for stripping voltammetry. Another attractive respect of the MCFCNTs is due to its unique structure, the electrode can be electrochemically rejuvenated after being used for hundreds of cycles. This property is very meaningful for the application of natural water detection.

113