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

 

DATE DUE DATE DUE DATE DUE

 

New :1 Emma

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DatoDue.p65-p. 15

APPLICATION OF DIAMOND THIN-FILM ELECTRODES FOR
THE ELECTROCHEMICAL DETECTION OF TRACE METAL
IONS
By

Prerna Virendra Sonthalia

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2004

ABSTRACT
APPLICATION OF DIAMOND THIN-FILM ELECTRODES FOR THE
ELECTROCHEMICAL DETECTION OF TRACE METAL IONS
By
Prerna Virendra Sonthalia

Determination of trace metal ion contamination using anodic stripping
voltammetry (ASV) is an established method. The use of boron-doped diamond (BDD)
thin-ﬁlm electrodes as alternates to Hg based electrodes for use in ASV was investigated.
This is mainly because of the toxicity of Hg. High-quality conductive diamond is
attractive for the detection of trace metal ions due to the material’s (i) wide working
potential window, (ii) stable surface chemistry and microstructure, (iii) chemically inert
surface, and (iv) good activity for metal deposition and stripping reactions. The
application of boron-doped microcrystalline and nanocrystalline diamond thin-ﬁlm
electrodes for the analysis of Ag(I), Cu(II), Pb(II), Cd(II), and Zn(II) yielded analytical
detection ﬁgures of merit comparable and better than those obtained for Hg-coated GC
and bare GC, respectively. Limits of quantitation in the low ppb range, good sensitivity,
excellent response precision (< 4 °/o) and long-term stability (> 96 %) were obtained.

Nanocrystalline diamond has been found to perfomI slightly better than the more
conventional microcrystalline form of diamond in this application. It was used in
conjunction with differential pulse voltammetry (DPASV) to detect these metal ions in
lake water, well water, tap water, wastewater treatment sludge, and soil. The

electrochemical results were comparable (< 10% difference in most cases) with data from

inductively coupled plasma mass spectrometric (ICP-MS) and or atomic absorption
spectrometric (AAS) measurements of the same samples.

A research goal was to determine how factors such as the diamond surface
morphology and chemistry affected the nucleation and growth of metal phases. Scanning
electron micrographs of the metal deposits are presented. In general, the fraction of the
surface area of the diamond covered by metal deposits was larger for nanocrystalline
diamond. This indicates that a larger amount of metal phase formation occurs on
nanocrystalline diamond during the preconcentration step, yielding greater sensitivity and
lower limits of quantitation in the DPASV measurements.

The quality of the DPASV measurement could be improved if the relatively large
peak widths seen in most of our presented data were reduced. We demonstrate the use of
pulsed galvanostatic deposition to produce Cu deposits that are nominally smaller in size
and more uniformly dispersed over the diamond surface than what is achieved using
constant potential or constant current. This in turn leads to relatively sharp and narrow
stripping peaks at 40% of the effective deposition time. We also found a possibility of
interrnetallic formation between Cu and Zn during their simultaneous measurement on
diamond. Initial results from the investigation of pulsed galvanostatic deposition as a
means to reduce the interrnetallic complex formation of the Cu-Zn compound proved -
promising. The hypothesis was that if the probability of a physical/chemical interaction
of the surface ad-atoms were reduced, it would help reduce interrnetallic complex

formation.

ACKNOWLEDGEMENTS

I would like to begin by thanking a man whose inﬂuence on my life is just
beginning. He is my adviser Dr. Greg Swain. His guidance, patience and support have
seen me through to this day. I deeply admire his personality, character, love for science
and value for family life. There was never a time in these four years that I have known
him when I could not approach him with an idea or a problem. His ability for hard work,
quest for excellence, vast knowledge of analytical chemistry, patience, attention to detail
and congenial nature amazes me. Thank you Greg!

I would like to thank my committee members — Dr. Simon Garrett, Dr. Vicki
Mcgufﬁn, Dr. Merlin Bruening and Dr. Dave Long. Their input and support has been
valuable to my development as a scientist. I am very grateful to Dr. Kathy Severin, she
has been my mentor and through her own example inspired me to make contributions to
science outreach. Appreciation is also due to Ewa Danielewicz (SEM facility) and Dr.
Lina Patino (ICP-MS facility) who helped a great deal in my research efforts.

The Swain group has been like an extended family. Dr. Jian Wang, Dr.
Malgorzata Witek, Dr. Mateuz Hupert and Dr. Shannon Haymond welcomed me into the
group, I got to learn a great deal from them. Visiting post-docs and scholars Dr. Josef
Cvaéka, Zuzana Cvackové, Veronika Quaiserové, Karolina Peckova, Dr. Jesse
Seegmiller and Dr. Yoshiyuki Show added a lot to my understanding of science. I thank
them all for their friendships and their contribution. Dr. Yoshiyuki Show’s help with all
the diamond ﬁlm growth and his advice is deeply appreciated. Jason Stotter and Grace
Muna, who joined the group the same year as me shared my pains and joys. Graces “go

girl” always helped me to carry on. Sanjeev Dalavoy, Jinwoo Park, Anne Fischer “Fish”,

iv

Gloria Pimienta, Doug Knigge, Jason Bennett “JB”, Audrey Martin, Shihua Wang, Yang
Song, Pushwinder Kaur and Elizabeth McGaw have been very supportive. Thanks guys!
Elizabeth McGaw has collected some of the data presented in this thesis. I thank her for
all her help and wish her lots of success with the project. JB developed the pulse
deposition setup, thanks JB! JB and Fish also helped with the rehydrogenation of the
diamond thin-ﬁlms. Jesse Seegmiller and Shihua Wang assisted in the SECM work. As I
ﬂy the “nest” I cannot help but feel a sense of sorrow at leaving the company of all these
wonderful people.

My family has been a constant source of encouragement, their sacriﬁces and love
cannot be emphasized enough. My grandmother who literally let go in her words her
“walking stick,” when she so selﬂessly supported my decision to undertake higher
education. My mother who silently suffered great hardships to see her children succeed.
My father who is a scientist at heart, although not one of his “inventions” has seen the
light of day! I would like to thank my brother Aayush and bhabhi Rakhee for their
unconditional love and support. I also thank mamaji and mamiji my “home away from
home” here in Lansing. I thank my ﬁ’iend Amit Goradia for his encouragement. I also
thank Lavanya Parthasarathy, Kiran Misra, Freddy Kharoliwala, Bhaskar Patham, Priya
Heerwani and all the others for their support.

Funding from the US. Department of Agriculture (2001-35102-10045) and
NASA (NA69-1481; Advanced Environmental Monitoring and Control) is gratefully

acknowledged.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. ix
LIST OF FIGURES ............................................................................. xi
CHAPTER 1

INTRODUCTION ............................................................................... 1
1.1 The Need for Monitoring Metal Ions ................................................... 3
1.2 The Differential Pulse Anodic Shipping Voltammetry (DPASV) Technique. . ....6
1.3 “Alternate” Electrodes and Their Limitations ......................................... 12
1.4 Diamond Thin-Films as an “Alternate” Electrode ...................................... 15
1.5 Outline of the Dissertation ................................................................ 17
1.6 References .................................................................................. 19
CHAPTER 2

THE PHYSICAL CHARACTERIZATION AND ELECTROCHEMICAL
PROPERTIES OF BORON-DOPED DIAMOND THIN-FILM ELECTRODES

2.1 Introduction ................................................................................. 23
2.2 Experimental ............................................................................... 24
2.3 Results and Discussion .................................................................... 28
2.3.1 Structural Characterization ......................................................... 28
2.3.2 Electrochemical Characterization .................................................. 32
2.3.3 Electrochemical Activity as studied by the Scanning
Electrochemical Microscope ........................................................ 36
2.4 Conclusions .................................................................................. 43
2.5 References ................................................................................... 45
CHAPTER 3

ELECTROCHEMICAL AND ATOMIC FORCE MICROSCOPY STUDIES OF
METAL PHASE NUCLEATION AND GROWTH ON BORON-DOPED
DIAMOND THIN-FILM ELECTRODES

3.1 Introduction .................................................................................. 48
3.2 Theory of Metal Nucleation and Growth ................................................ 48
3.3 Experimental ................................................................................ 53
3.4 Results and Discussion .................................................................... 55
3.4.1 Metal Ion Voltammetry ............................................................. 55
3.4.2 Chronocoulometn'c Measurements ................................................ 59
3.4.3 Metal Ion Chronoamperometric Measurements .................................. 61
3.4.4 EC-AFM Study of Copper Deposition ........................................... 66
3.5 Conclusions ................................................................................. 70
3.6 References .................................................................................... 71
CHAPTER 4

ANODIC STRIPPING VOLTAMMETRIC DETECTION OF TRACE METAL
IONS: A PERFORMANCE COMPARISON OF BORON-DOPED DIAMOND

vi

THIN-FILM ELECTRODES AND SOME COMMONLY USED ELECTRODE
MATERIALS

4.1 Introduction ................................................................................. 74
4.2 Experimental ................................................................................ 75
4.3 Results and Discussion .................................................................... 77
4.3.1 Comparison of Diamond and Glassy Carbon for the LSV Detection of
Metal Ions ........................................................................... 77
4.3.2 Comparison of Diamond and Hg-Film Electrodes for the DPASV Detection
of Metal Ions ......................................................................... 80
4.3.3 Linear Sweep and Differential Pulse Voltammetry of Individual Metal
Ions ...................................................................................... 85
4.3.4 Differential Pulse Voltammetry of Mixtures of Metal
Ions ..................................................................................... 89
4.3.5 Scanning Electron Microscopy (SEM) Studies of Metal Phase
Formation .............................................................................. 94
4.4 Conclusions .................................................................................. 98
4.5 References ................................................................................... 100
CHAPTER 5

TRACE METAL ANALYSIS IN CONTAMINATED WATER SAMPLES USING
AN ODIC STRIPPING VOLTAMMETRY AND A NANOCRYSTALLINE
DIAMOND THIN-FILM ELECTRODE

5.1 Introduction ................................................................................. 102
5.2 Experimental .............................................................................. 103
5.3 Results and Discussion ................................................................... 109
5.3.1 Analysis of a NIST Standard Solution ............................................ 109
5.3.2 Drinking Water Analysis .......................................................... 112
5.3.3 Wastewater Treatment Sludge Analysis ......................................... 115
5.3.4 Contaminated Soil Analysis ...................................................... 118
5.4 Conclusions ................................................................................ 120
5.5 References ................................................................................. 122
CHAPTER 6

CHALLENGES AND SOLUTIONS: TOWARDS GAINING BETTER
SELECTIVITY AND RESOLUTION DURING THE AN ODIC STRIPPING
VOLTAMMETRY OF METAL IONS USING BORON-DOPED DIAMOND THIN-

FILM ELECTRODES

6.1 Introduction ................................................................................ 123

6.2 Experimental .............................................................................. 125

6.3 Results and Discussion ................................................................... 130
6.3.1 Intermetallic Compound Formation ............................................. 130
6.3.2 Characterization of the Cu-Zn metals deposited on diamond.................137
6.3.3 Pulsed Galvanostatic Deposition ................................................ 141
6.3.4 Observation of Voltammetric i-E Curves of Narrow FWHM. . . . . . . . ........152

6.4 Conclusions ................................................................................ 155

vii

6.5 References .................................................................................. 157

CHAPTER 7
SUMMARY .................................................................................... l 59

viii

LIST OF TABLES

Table 1.1 Factors affecting the toxicity of metals .............................. 4

Table 1.2 Characteristics and attributes of Hg and diamond electrodes for
ASV ........................................................................ 13

Table 2.1 Instrumental techniques used for characterization .................. 24

Table 3.1 Summary of the cyclic voltammetric peak currents, peak
potentials and peak potential separations for the Cu(H)/Cu redox
system at various scan rates for a microcrystalline diamond thin-
ﬁlm electrode .............................................................. 56

Table 3.2 Chronocoulometric charge data for Ag(I) in 0.1 M acetate
buffer, pH 4.5. Potential steps from 0.6 to 0 to 0.6 V were
applied for 5 3 each to diamond electrodes of two different
morphologies .............................................................. 61

Table 3.3 Current-time transient data for copper electrodeposition on
microcrystalline diamond ................................................ 66

Table 4.1 LSV analytical detection ﬁgures of merit for the detection of
several metal ions at boron-doped microcrystalline diamond and
glassy carbon electrodes ................................................. 80

Table 4.2 Analytical detection ﬁgures of merit for the differential pulse
anodic stripping voltammetric (DPASV) detection of the Cd(II)
at a nanocrystalline diamond and Hg-coated GC electrode. (Data
obtained by EA. McGaw) ............................................... 84

Table 4.3 Summary of the differential pulse anodic stripping voltammetric
(DPASV) detection ﬁgures of merit for the measurement of

individual metal ions at microcrystalline (M) and nanocrystalline
(N) diamond .............................................................. 89

Table 4.4 Summary of the differential pulse anodic shipping voltammetric
detection ﬁgures of merit for mixtures of metal ions at
microcrystalline (M) and nanocrystalline (N) diamond ............. 93

Table 4.5 Comparison of the nominal particle size, density and fractional
surface area of the diamond ﬁlm covered by the metal phase, as

determined from scanning electron microscope (SEM) images... 96

Table 5.1 Concentration and percent error data for Ag(I), Cu(II) and Pb(II)
in the SRM 1640 water samples ........................................ 111

ix

Table 5.2

Table 5.3

Table 5.4

Table 6.1

Concentration and percent difference data for Cu(II) in two
different tap water samples from the MSU campus as determined
by DPASV and ICP-MS ................................................. 114

Concentration and percentage difference data for Cd(II) and
Pb(II) in a wastewater treatment sludge extract as determined by
DPASV and AAS ........................................................ 117

Concentration and percent difference data for Pb(II) and Cu(II)
in contaminated soil extracts as determined by DPASV and AAS 120

Work functions, kinds of interactions in binary systems and
types of anodic voltammetric i-E curves of electrochemically
formed deposits. (Adapted from Ref 20) ........................... 136

Figure 1.1

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

LIST OF FIGURES

The differential pulse anodic stripping voltammetric (DPASV)
potential waveform and resulting current response. (1) The
preconcentration step involves a cathodic potential-step of some
magnitude and duration and the (2) the oxidation step involves a
potential scan in the positive direction to selectively oxidize
each metal phase formed. An enlarged view of the applied pulse
program depicting the various parameters is shown to the right.
The resulting voltammetric i-E curve is constructed by plotting
the current difference for each pulse as a function of the
potential ...................................................................

5x5 um2 AF M images (height mode) of boron-doped (A)
microcrystalline and (B) nanocrystalline diamond ﬁlms .........

Visible macro-Raman spectra of (A) microcrystalline and (B)
nanocrystalline diamond ﬁlms .......................................

XRD patterns of (A) microcrystalline and (B) nanocrystalline
diamond ﬁlms ...........................................................

Background cyclic voltammetric i-E curves in 1 M KCl for (A)
microcrystalline diamond and (B) nanocrystalline diamond thin-
ﬁlm electrodes ............................................................

Cyclic voltammetric i-E curves for (A) Fe(CN)6'3M, (B)
Ru(NH3)6+2/+3, (C) 1:064”, and (D) methyl viologen (Mv+ ’*)
at increasing scan rates of 50, 100, 200, 500, 1000 and 2000
mV/s ........................................................................

Tip-approach curve at a polycarbonate substrate. The data were
acquired in a 1 mM solution of Ru(NH3)63+ in 1 M KCl. The tip
is a 10 um-radius Pt disk. The tip was biased at -0.3 V

Tip approach-curves for the different electrode materials:
platinum, microcrystalline and nanocrystalline diamond. The
simulated current-distance data for a tip with an RG ratio 6.5 is
also shown. Experimental curves were acquired in a 1 mM
solution of Ru(NH3)53+ in 1 M KCl. The tip is a 2 um-radius Pt
disk of RG ratio between 6 and 7. The tip was biased at -0.3 V
and the substrate at 0.05 V ............................................

Schematic illustration of the SECM feedback mode with a ﬁnite

substrate (Adapted from reference 26). The heterogeneous redox
reactions occur at a diffusion-controlled rate at the ultra-

xi

29

31

32

33

34

37

40

Figure 2.9

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

microelectrode and conductive portion of the
tip ........................................................................

Simulated i/in.o vs. L curves for ﬁnite disk-shaped substrates:
(ﬁlled symbols) multidimensional integral equations
simulations; (open symbols) ﬁnite difference method
simulation. The h values are indicated. (Figure adapted from
reference 26) ............................................................

The stages in the formation of a stable, growing nucleus of a
metal on a substrate of a different material such as carbon.
(Adapted from reference 10) ..........................................

Deposition of a thick metal layer of metal on an inert cathode
via a progressive nucleation and growth mechanism. (Adapted
from reference 10) ......................................................

Schematic representation of different growth modes of metal
deposition on foreign substrate (A) 3-D metal island formation
(Volmer-Weber grth mode) ‘I’Mads — S << ‘I’Mads — M (B)
Metal layer by layer formation, ‘PMads — S >> ‘I’Mad, — M and
ratio do)“ - do,s/ dos z 0, (C) 3—D metal island formation on top
of predeposited 2D metal overlayers on substrate, ‘I’Mads — S >>
‘I’Mads — M and ratio doM - do,s/ dos > 0 (positive misﬁt) or doM
- dos / dos < 0 (negative misﬁt). (Adapted from reference
12) ........................................................................

Cyclic voltammetric i-E curves for the reduction of 1 mM
Cu(II) and subsequent oxidation at a microcrystalline diamond
thin-ﬁlm electrode. Scan rate = 100 mV/s. The redox reaction
for the Cu/Cu(Il) system is shown ...................................

Cyclic voltammetric i-E curves for 10 um solutions of Pb(II),
Cu(II), Cd(II) and Zn(II) at a microcrystalline thin-ﬁlm diamond
thin-ﬁlm electrode. The metals deposit and oxidize off near their
standard reduction potentials. Scan rate = 100 mV/s, supporting
electrolyte = 0.1 M acetate buffer, pH 4.5 ...........................

(A) Schematic diagram of the time-dependent grth of
diffusion zones around forming metal nuclei on an electrode
surface and their eventual overlap. The arrows indicate the
directions of the diffusion ﬁelds during the grth of nuclei. (B)
Expected current transient; K is constant given by
nFADo‘lzCoi/nm, the terms have their usual meanings. . .. . . . .

Comparison of experimental current-time transients with the

xii

41

42

49

50

52

56

59

63

Figure 3.8

Figure 3.9

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

theoretical transients corresponding to instantaneous and
progressive nucleation at different overpotentials .................

ECAFM images (height mode, 5 pm X 5 um) collected during
Cu deposition on microcrystalline diamond from a 50 ppm
CuSOII/OJ M acetate buffer solution at different cathodic
overpotentials. Arrows indicate the formation of new nuclei. . ..

EC-AFM images (height mode, 5 mm X 5 pm) collected during
Cu deposition on microcrystalline diamond from a 50 ppm
CuSO4 /0.1 M acetate buffer solution at various times at a ﬁxed
overpotential. Arrows indicate the formation of new
nuclei ......................................................................

Linear sweep voltammetric i-E curves for Ag(I) concentrations
of 10 uM, 100 uM and 1 mM. Inset shows the i-E curves for the
1 and 10 uM Ag(I) concentrations. The solutions were prepared
in 0.1 M acetate buffer, pH 4.5. The log-log current-
concentration plots for the assays on BDD and GC are also
presented ..................................................................

Electrochemically deposited Hg on the surface of a
microcrystalline diamond thin-ﬁlm electrode. The Hg tended to
form spherical droplets on the substrate. (Adapted from
reference 8) ...............................................................

Differential p ulse anodic s tripping voltammetric (DPASV) i-E
curves for standard solutions of Cd(H) ranging in concentration
of 10 to 1000 ppb. The working electrodes were (A)
nanocrystalline diamond and (B) Hg-coated GC. The supporting
electrolyte was 0.1 M acetate buffer, pH 5.2. Deposition
potential = -900 mV vs. Ag/AgCl. Deposition time = 3.5 min.
The other DPASV conditions are described in experimental
section. Geometric area = 0.222 cmz. The peak areas were
determined by integration between —700 mV and — 900 mV.
(Data obtained by EA. McGaw) ......................................

Linear sweep stripping voltammetric i-E curves for 0.1 mM
Cd(II) in 0.1 M acetate buffer, pH 4.5, at a microcrystalline
diamond electrode, depicting the dependence of ip"x on
deposition time. Deposition time was varied from 30 to 180 s
(30 5 increments). Deposition potential —1000
mV ........................................................................

xiii

65

68

69

78

81

84

86

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Figure 5.1

Figure 5.2

Figure 5.3

Representative differential pulse anodic shipping voltammetric
(DPASV) i-E curves for a 10 uM concentration of each metal
ion at both microcrystalline and nanocrystalline diamond thin-
ﬁlm electrodes. Supporting electrolyte = 0.1 M acetate buffer,
pH 4.5 .....................................................................
Differential p ulse anodic s nipping v oltammetric ( DPASV) i-E
curves for Ag, Pb and Cd on boron doped nanocrystalline
diamond in acetate buffer, pH 4.5.The preconcentration was
performed for 3.5 min. in quiescent solution at -1000 mV. The
concentration of the metal ions in each solution was (a) 5 (b) 1
(c) 0.5 (d) 0.1 uM .......................................................

Differential pulse stripping voltammetric (DPASV) i-E curves
for the detection of 10 uM concentrations of Ag, Cu, Pb, Cd and
Zn for nanocrystalline diamond ﬁlm in acetate buffer, pH
4.5.The preconcentration was carried out for 3.5 min in
quiescent solution at -1200 mV ......................................

Scanning electron micrographs of metal phases deposited on
microcrystalline (top) and nanocrystalline (bottom) diamond
thin-ﬁlm electrodes. The deposition, in each case, was for 3 min
at —1.2 V (no stirring). Arrows indicate a few representative
metal deposits ............................................................

Differential pulse anodic stripping voltammetric (DPASV) i-E
curves for a standard solution (SRM 1640) and known solutions
of Pb, Cu and Ag using a boron-doped nanocrystalline diamond
electrode. Deposition potential = - 900 mV. The DPASV
conditions are described in the Experimental section. Geometric
area = 0.2 cm2. Peak area integration range = 200 to 370 mV for
Ag(I), -200 to 100 mV for Cu(II), and -550 to -380 mV for
Pb(II) .............................................................

Standard addition plot of charge vs. concentration for the
addition of solutions containing Ag(I) in increasing
concentrations ranging from 10 ppb to 90 ppb to a solution of
the SRM diluted in acetate buffer (1: 9). All concentrations
reported have been corrected for dilution ............................

Differential pulse anodic shipping voltammehic (DPASV) i-E
curves for Cu in tap-water samples #1 and #2 using a boron-
doped nanocrystalline diamond electrode. Deposition potential =
- 200 mV. The DPASV conditions are described in the
Experimental section. Geometric area of electrode = 0.2 cm2.
Peak area integration range = -100 to lOOmV for Cu(II) ..........

xiv

88

90

92

97

110

111

114

Figure 5.4

Figure 5.5

Figure 6.1

Figure 6.2

Figure 6.3

Figure 6.4

Differential p ulse anodic shipping voltammetric (DPASV) i-E
curves for Cd and Pb in acid-digested waste sludge using a
boron-deped nanocrystalline diamond thin-ﬁlm electrode.
Deposition p otential = - 1000 m V. The D PASV conditions are
described in the Experimental section. Geometric area = 0.2
cm2. Peak area integration range = -600 to -450 mV for Pb(II),
and —850 to -650 mV for Cd(II) .......................................

Differential p ulse anodic s tripping v oltammetric (DPASV) i-E
curves for Pb and Cu in MGP contaminated soil extracts using a
10 ppm boron-doped nanocrystalline diamond thin-ﬁlm
electrode. Deposition potential = - 1000 mV. The DPASV
conditions are described in the Experimental section. Geometric
area = 0.2 cm2. Peak area integration range = -50 to 80 mV for
Cu(II), and -525 to -425 mV for Pb(II) ...............................

The electrical set-up used for the galvanostatic pulse
electrodeposition. (Adapted from Reference 15). CE -— counter
electrode; WE — working electrode and Ref - reference
electrode ..................................................................

Differential p ulse anodic stripping v oltammetric (DPASV) i-E
curves for Cu and Zn. Standard additions of solutions containing
Cu(II) were made in increments of 20 ppb to a solution of 100
ppb Zn(II) in pH 5.2 acetate buffer. Deposition potential = -
1300 mV. The DPASV conditions are as described in the
Experimental section. Geomehic area of the electrode = 0.163
cm .................................................................

Differential pulse anodic shipping v oltammetric ( DPASV) i-E
curves for Cu(II) and Zn(II). Standard additions of solutions
containing Zn(II) were made in increments of 20 ppb to a
solution of 100 ppb Cu(II) in pH 5.2 acetate buffer. Deposition
potential = - 1300 mV. The DPASV conditions are as described
in the Experimental section. Geometric area of elech'ode = 0.163

cm2 .................................................................

SEM images (secondary electron) of Cu and or Zn deposited on
microcrystalline diamond electrodes formed at -1.3 V for 3 min
from varying ratios of the metal ions in solution: (A) 100 ppm
Cu(II); (B) 100 ppm Cu(II) + 10 ppm Zn(II); (C) 100 ppm
Zn(II); and (D) 100 ppm Zn(II) + 10 ppm
Cu(II) .....................................................................

XV

117

119

129

133

133

139

Figure 6.5

Figure 6.6

Figure 6.7

Figure 6.8

Figure 6.9

Figure 6.10

Energy dispersive x-ray analysis spectrum for Cu and or Zn
deposited on microcrystalline diamond electrodes at -1300 mV
for 3 min. at metal ion concentrations of: (A) 100 ppm Cu(II) +
10 ppm Zn(II) and (B) 100 ppm Zn(II) + 10 ppm Cu(II) ..........

Differential pulse stripping voltammetric i-E curves for Cu on
nanocrystalline diamond deposited at (A) constant current
deposition at 15 mA/cm2 for 903, (B) pulsed current deposition
at 15 mA/cmz, 0.5 s pulses at a 50% duty cycle for a total time
of 180 s, (C) constant potential step deposition at — 400 mV for
903, and (D) pulsed potential deposition at - 400 mV with 0.5 s
pulses at a 50% duty cycle for a total time of 180 s ................

Differential pulse anodic stripping voltammetric i-E curves for
Cu from a nanocrystalline diamond surface using pulsed
galvanostatic deposition at (A) 2.7, (B) 5.5, (C) 12, (D) 15, and
(E) 21 mA/cmz. The solution was 100 ppb Cu(II) in 0.1 M
acetate buffer, pH 5.2. Total deposition time = 180 5. Pulse
width = l s. Duty cycle = 50%. Arrow indicates the optimum
current density ............................................................

Differential pulse anodic shipping voltammetric i-E curves for
Cu from a nanocrystalline diamond surface using pulsed
galvanostatic deposition at effective deposition times of (A)
3603 (25 % duty cycle, 90, l-s on pulses), (B) 1803 (50 % duty,
90, l-s on pulses), and (C) 1203 (75 % duty, 90, l-s on pulses).
The solution was 100 ppb Cu(II) in 0.1 M acetate buffer, pH
5.2. Current density = 15 mA/cm2 ....................................

Differential pulse anodic shipping voltammetric i-E curves for
Cu from a nanocrystalline diamond surface using pulsed
galvanostatic deposition at pulse widths of (A) 1.5, (B) 3, (C)
0.5, (D) 1.0, (E) 0.1, and (F) 0.25 s. The solution was 100 ppb
Cu(II) in 0.1 M acetate buffer, pH 5.2. Current density = 15
mA/cmz. Duty cycle = 50%. Total deposition time = 180
s .............................................................................

Scanning electron micrographs (back scattered images) of Cu
deposited on nanocrystalline diamond at 15 mA/cm2 in the form
of a (A) constant current applied for 90 s and (B) a pulsed
current applied for 180 s using 0.5 3 pulse width. Other
conditions have been provided in the experimental
section ....................................................................

xvi

140

141

145

146

147

148

Figure 6.11

Figure 6.12

Figure 6.13

(A) Background-subtracted differential pulse shipping
voltammetric i-E curves for metal deposits formed from 50, 100,
250, 500 and 1000 ppb of Cu(II) solution in pH 5.2 acetate
buffer. The inset shows the shipping voltammetric i-E curves
for the 1 ppb (dashed line) and 10 ppb Cu(II) solutions.(B) Plot
of charge vs. concentration for the data in 6.11.A. The
integration range was 100 mV (potential limits for integration
varied with concentration) .............................................

Comparison of representative background subtracted differential
pulse anodic voltammetric i-E curves for 100 ppb Cu(II) and
100 ppb Zn(II) in pH 5.2 acetate buffer. Preconcentration was
performed by (A) pulsed current deposition at 15 mA/cm2 with
a 0.5 8 pulse width and a 50% duty cycle and a total time of
deposition of 180 s, and (B) constant potential deposition at -
1200 mV for 90 s .........................................................

Background-subtracted stripping voltammehic i-E curves
recorded for (A) 100 ppb Ag(I) and (B) 100 ppb Cd(II) in 0.1 M
acetate buffer, pH 5.2 (with added 30 ppm Ca(II) and 5 ppm
Mg(II) salts). The solid lines depict the ﬁrst recorded i-E curves
and the dashed lines depict subsequent i-E curves (no purging
with N2 undertaken) ......................................................

xvii

150

151

153

CHAPTER 1

INTRODUCTION

There is an increasing demand for the development of sensitive, reproducible,
low-cost, and ﬁeld-deployable methods for the detection of trace metal ions in water
supplies. Determination of trace metal ion contamination in the environment using anodic
stripping voltammetry (ASV) has all the above listed advantages. ASV involves a two
step measurement: preconcentration or metal deposition followed by an anodic potential
scan to selectively oxidize each deposited metal phase. Historically, Hg drop (HMDE)
and Hg-ﬁlm—coated (MFE) electrodes have been used in ASV; however, there is a need
to develop Hg-free “alternate” electrodes. This is mainly because of the toxicity of Hg.
The use of boron-doped diamond (BDD) thin-ﬁlm electrode as an alternate to Hg for this

application is the focus of this work.

Diamond thin-ﬁlm electrode shows promise for this electroanalytical application,
as the material possesses many of the same attributes as Hg. For example, high—quality
diamond exhibits (i) large overpotentials for hydrogen evolution and oxygen reduction,
(ii) a large overpotential for oxygen evolution (large positive window), (iii) low
background current, (iv) resistance to electrode fouling, (v) rapid electrode reaction
kinetics for metal deposition and shipping reactions, and (vi) no chemical interaction
with metal deposits. Diamond consists of an sp3-bonded carbon network. Its structural
robustness and chemical inertness result from high atomic density and strong, directional

covalent bonding.l

The unique properties of boron-doped diamond make it ideally suited for ASV, as
our group has demonstrated for Ag(I), Cu(II), Pb(II), Cd(II) and Zn(II).“4 Work in our
laboratory has helped to establish the feasibility of electrodepositing metal phases on
diamond surfaces.5'6 Diamond electrodes provide improvements over other alternate
electrodes, in terms of several electroanalytical features of merit: linear dynamic range,

sensitivity, detection limit, response variability and response stability.

This research sought to answer several key questions.

1. Is diamond a viable electrode material for trace metal ion analysis?

2. How does the performance of microcrystalline and nanocrystalline diamond thin-ﬁlms
compare, given their differences in surface morphology and microstructure, for the
detection of several priority metal ions? What type of analytical detection ﬁgures of merit

are possible for real and simulated contaminated water samples?

3. How do factors, such as the diamond surface morphology and chemistry, affect the
nucleation and growth of metal phases? Where do the metal phases form on the surface

and what particle size and distribution do they possess?

4. Is intermetallic compound formation on diamond a major issue and, if so, can
formation be minimized through the use of pulsed galvanostatic deposition rather than

conventional constant potential or current deposition?

1.1 The Need for Monitoring Metal Ions

Metals differ from other toxic substances in that they are neither created nor
desh'oyed by humans. Their utilization has ramiﬁcations for human health in at least two
major ways: anthropogenic contributions to air, water, soil and food, and altering the

speciation or biochemical form of the element.7

Accurate determination of metal ion concentrations in water supplies is critical
because of their potential for causing adverse health effects in humans.8 An interference
of the metal contaminant with an essential metal may occur when the metabolism of the
two are similar.9 For instance, Pb(II) competes with Ca(II) in several ways: (i) inhibiting
the release of neurotransmitters and interfering with the regulation of cell metabolism by
binding to second-messenger calcium receptors, (ii) blocking ion transport by calcium
channels and calcium-sodium ATP pumps, and (iii) by competing for calcium-binding
protein sites and uptake by mitochondria. Dietary deﬁciencies of Ca(II), Fe(II)/(III) and
Zn(II) enhance the effects of Pb(II) on cognitive and behavioral development. Iron
deﬁciency increases the gastrointestinal absorption of cadmium, and cadmium competes
with zinc for binding sites on metallothionein, which is important in the storage and
transport of zinc during development. The EPA regulates the maximum allowed levels of
these metals in potable water (National Primary Drinking Water Standards: Cd — 5 ppb,
Cu - 1.3 ppm, Pb- 15 ppb; Secondary Maximum Contaminant Levels: Ag — 100 ppb, Zn

—5 ppm). Therefore, they are analytes of importance.10

The toxicity of a particular metal depends on several factors with the important
ones listed in Table 1.1.11 A critical determinant of metal retention is its biological half-
life, that is, the time it takes for the body or organ to excrete half of an accumulated
amount. The half-life varies according to the metal as well as the organ or tissue. For
example, the biological half-lives of Cd(II) in the kidney and Pb(II) in bone are 20 to 30
years, in contrast to As(III, IV) or Li(I) which are only a few hours or days.11 This means

that some metal ions have a tendency to bio-accumulate in speciﬁc organs of the human

body.

Table 1.1 Factors affecting the toxicity of metals

 

Interactions with essential metals

 

Formation of metal-protein complexes

 

Age and stage of development of the person

 

Lifestyle factors

 

Chemical form or speciation of the metal ion

 

Immune status of the host

 

 

 

A brief review of the sources, degree of toxicity and exposure routes of these ﬁve priority

ions (Ag(I), Cu(II), Cd(II), Pb(II) and Zn(II)) is given below.”

Ag (Ag(I)): The principal use of Ag is as a halide in the manufacture of photographic
plates. Other uses are for jewelry, coins and eating utensils. Ag salts are known to have

medicinal uses as antiseptics and astringents. Ag compounds can be absorbed orally, by

inhalation, and through damaged skin. The major effect of excessive absorption of Ag(I)
is local or generalized impregnation of the tissues, where it remains as silver sulﬁde. This
forms an insoluble complex in elastic ﬁbers, resulting in agyria. Chronic bronchitis has

also been reported to result from medicinal use of colloidal Ag.

Cu (Cu(II)): Cu is widely distributed in nature and is a nutritionally essential element.
Ambient air levels are generally low in the United States. For the general population,
food, beverages, and drinking water are potential sources of excess exposure.
Experimental studies in humans suggested that ingestion of drinking water greater than 3
ppm of Cu produced gastrointestinal symptoms including nausea, vomiting and diarrhea.
Ingestion of large amounts of Cu salts, most frequently CuSO4, could result in hepatic

necrosis and death.

Cd (Cd(II)): A modern toxic metal. It was discovered as an element only in 1817,
however, it has many current applications. Because of its non-corrosive properties, its
main use is in electroplating or galvanizing. It is also used as a color pigment for paints
and batteries and as a cathode material for nickel-cadmium batteries. Cd is a by—product
of Zn and Pb mining and smelting, which are important sources of environmental
pollution. For the general population, the major source of Cd is food. Gastrointestinal
absorption of Cd(II) is about 5 to 8 percent. A major non-occupational source of
respirable Cd(II) is cigarettes, although workplace exposure to Cd(II) has been known to

be particularly hazardous.

Pb (Pb(II)): Pb is toxic to most living things at high exposure levels and there is no
demonstrated biological need for it. The major risk is toxicity to the nervous system and
the most susceptible populations are children. The principle route of exposure for people
in the general population is food. Other sources include Pb-based indoor paint, Pb(II) in
air from combustion of lead containing industrial emissions and Pb(II) in contaminated

water.

Zinc (Zn(II)): Zn is a nutritionally essential metal, and a deﬁciency results in severe
health consequences. More than 200 metalloenzymes require Zn as a cofactor. Adverse
health effects from excessive exposure to Zn(II) are relatively uncommon, but
gastrointestinal distress and diarrhea have been reported following ingestion of beverages
standing in galvanized cans or from the use of galvanized utensils. With regard to
industrial exposure, metal fume fever resulting from inhalation of freshly formed fumes

of Zn(II) presents the most signiﬁcant effect.

1.2 The Differential Pulse Anodic Stripping Voltammetry
(DPASV) Technique

Several analytical techniques are routinely used for the detection of heavy metal
ions including atomic absorption, plasma emission, x-ray ﬂuorescence spectroscopy,
anodic stripping voltammetry (ASV), and neutron activation.12 However, none of these
methods offer all the advantages of ASV. Voltammetry refers to a class of
electroanalytical techniques in which the current at a working electrode is measured as a
function of a potential waveform applied to the it.12 The working electrode is the
electrode at which the reaction of interest occurs. ASV is a type of voltammetric

technique that utilizes a bulk electrolytic step to preconcentrate the analyte from the

sample solution into or onto the working electrode. The preconcentration step, which can
be thought of as an electrochemical extraction step, yields very favorable signal-to-noise
ratios of the recorded voltammetric i-E curves. In fact, the detection limits are lowered by
2 to 3 orders compared to solution-phase voltammetric measurements.” In summary,
ASV is a sensitive detection method for many metal ions and possesses multielement
analysis capability, low instrumental and maintenance costs, ﬁeld deployability and low

power demand with no need for cooling and or ventilation.12

As depicted pictorially in Figure 1.1, ASV involves a two-step measurement
sequence: (i) electroreduction of the metal ions at a negative potential to form metal
phases on the electrode surface (the preconcentrating step), and (ii) electrooxidation of
each metal during a slow potential scan toward positive potentials (the stripping step). In
theory, each metal is selectively oxidized near its standard reduction potential, E°', and
the oxidation current or charge is recorded and used for quantitative purposes. One
voltammetric method commonly used is differential pulse voltammetry. In this method, a
series of pulses of equal amplitude are superimposed on a potential ramp. As shown in
Figure 1.1, each pulse has parameters of amplitude, width, period and step height that can
be controlled to increase the magnitude of the current. The currents are sampled twice,
once just prior to the pulse application and again at the end of the pulse period. The ﬁrst
measured current is subtracted from the second one and the current difference is plotted
versus potential. The technique is useful for discriminating the faradaic current from the
background current. When the potential pulse is applied, the total current in the system

increases because of the resultant increase in both the faradaic and double layer charging

currents. The charging current decays rapidly in comparison to the diffusion-controlled

-t/Rqu

faradaic current (ic ~ e versus if ~ t'w, where t is the time (s), Ru the uncompensated

solution resistance (0) and Cd the double layer capacitance (F)).

Some researchers are also of the opinion that an overall improvement in the
signal-to-background ratio occurs because some of the metal ions that are stripped off
from the electrode surface during the application of the pulse are replated onto the
electrode during the waiting time between pulses.'2 The use of the differential pulse
voltammetry with a Hg thin-ﬁlm deposited on glassy carbon (GC) gave detection limits
3-5 times lower than those achieved with linear scan voltammetry.M The same authors
noted that the differential pulse peak currents are more susceptible to interference by
surface-active substances than linear scan measurements, with undetected absorption
effects causing greater errors.'4 This can be explained by the fact that potential step-based
techniques tend to depend heavily on the electron-transfer rate constant of the reaction.15
If the system has slow charge-transfer kinetics (high activation energy) for the reaction of
interest, contaminants in solution would manifest themselves in the voltammetric i-E

CUI’VCS.

Two small-amplitude controlled-potential techniques, differential pulse
voltammetry and square-wave voltammetry have been tested for use in the stripping step.
Square—wave voltammetry is a differential technique in which a waveform composed of a
symmetrical square-wave is superimposed on a base staircase potential. Although both

techniques yielded very favorable signal-to—background ratios, the peak widths of the

voltammetric i-E curves in the differential pulse method were narrower. An advantage of
the differential pulse method is the improved resolution between two species with similar
redox potentials due to its peak-shaped response. In various situations, peaks separated by

only 50 mV could be measured.13

 

 

 

   

 

  

 

 

 

 

 

 

—) é-Pulse Width
Step 5‘1 Sample Period
(1) Preconcentration step E - —
' ‘ Pulse
A (2) Stripping step Amphtude , sample Perm
A K—Pulse Period
;>
v I
'8 1
g I $uiet
' me
8 I
a? :
i t
(1) :
I

 

 

 

 

Time (s)

---—-——p()-—---——

-—------- #1

 

Current Difference
(PA)

_----—-

P---—-

 

Potential (V)

Figure 1.1 The differential pulse anodic stripping voltammetric (DPASV) potential
waveform and resulting current response. (1) The preconcentration step involves a
cathodic potential-step of some magnitude and duration and (2) the oxidation step
involves a potential scan in the positive direction to selectively oxidize each metal phase
formed. An enlarged view of the applied pulse program depicting the various parameters
is shown to the right. The resulting voltammetric i-E curve is constructed by plotting the
current difference for each pulse as a function of the potential.

Several researchers such as R.L. Birke, R.A. Osteryoung, J. Osteryoung, K.W.
Hanck and KB. Oldham developed the theory of differential pulse voltammetry in the
period from the mid-19605 to late 19703.1("22 However, the concept of pulse polarography
was originally developed in England by GC. Barker as an out—grth of his work on

23, 24

square wave polarography. The equation representing the maximum current

difference (Aimax) obtained is given by:21

. _ nzFZAEADomCo
Almax - 4RTIZ’l/2tl/2 (1.1)

 

Ai is the difference in current (A) recorded at potentials E1 and E2, Co is the concentration
of the reactant (mol/cm3), t is the step width (5), Do is the diffusion coefﬁcient (cmZ/s), A
is the area of the electrode (cmz), F is the Faraday constant (C/mol e"), R is the gas
constant (Jmol'lK’l) and n is the stoichiometric number of electrons involved in the
electrode reaction (mol e'). The above expression however applies when the pulse
amplitude AE (V), AB = E2 - E1, is smaller than RT/nF. When a large pulse is used, the

expression for the maximum current becomes:

nFACoDo”2(0'—1)
ﬂl/Ztl/2(O,+1)

 

A1max =

(1.2)

E2_El
2

 

nF
and, 0 : €Xp( —El/2)ﬁ (13)

When AE/2 becomes very large with respect to RT/nF, equation 1.3 is simply the Cottrell
equation. The above expressions should apply for the case of planar solid electrodes
because the effect of electrode sphericity or capillary parameters in case of the HMDE
has not been taken into consideration. The expressions hold for mass-transfer limited

conditions.

10

To determine the effect of pulse amplitude on peak height, the term (0-1)/( 0+1)
can be evaluated at various pulse amplitudes and for various values of n. This term is
unity at maximum height and the values obtained represent the fraction of maximum
peak height under the conditions listed. Although, increasing the pulse amplitude would
give increased sensitivity of the measurements, it would also result in decreased
resolution.

Barker also gives an expression for peak half width, W1 ,2

W1/2=3.52 *R—T- (1.4)
nF

Again, this expression is valid only for small pulse amplitudes. The equation gives peak
half widths of 90.4, 45.2 and 30.1 mV, for 1, 2 and 3 electrons, respectively.
To derive the expression fOr half width, which is valid for any pulse amplitude, the terms

P and A are derived:

(A3+3A2 —3A—1)

 

 

P2+P +1=0 1.5
(A-AZ) ( )
AE nF
Where,A=ex — — 1.6
P 2 (RT) ( )
E +E nF
Also,P=exp( ‘2 2 -E”2)R_T_ (1.7)

From P, two values of E are obtained and the difference between these values gives the

peak half width.

In summary, the selection of the pulse amplitude and potential scan rate usually

involves a trade-off among sensitivity, resolution, and speed.13 Our initial efforts with the

11

project involved the optimization of the differential pulse parameters to yield the

maximum analytical signal.

1.3 “Alternate” Electrodes and Their Limitations

Historically, Hg drop (HMDE) and Hg ﬁlm-coated (MFE) electrodes have been
used in ASV. However, there is much interest in ﬁnding non-toxic, “alternate” electrodes
that function as well as Hg does.25’ 26 Hg can exist in three forms: elemental, inorganic
and organic. All three forms are toxic, however, their biological uptake mechanisms are
different. Hg in forms such as methyl mercury is extremely toxic and mainly affects the
central nervous system (CNS). Elemental Hg readily passes across the blood-brain barrier
into the central nervous system and also into the fetus. Mercuric chloride salts (HgClz and
ngClz) are known to cause corrosive ulceration, bleeding and necrosis of the

gastrointestinal tract accompanied by shock and circulatory collapse.

There are also other reasons for seeking alternate electrodes. Hg cannot be used
for the detection of metal ions having more positive standard reduction potentials than it
(e. g., Ag(I) and Au(II)) or which are insoluble in it (e.g., As(III)).26 Hg is a ﬁne electrode
material for ASV because the liquid environment provides a medium into which the
deposited metals can dissolve and form amalgams. The oxidation of a metal ion from its
amalgam is a much better understood and easily controlled process than the metal
nucleation and growth and subsequent oxidation from solid electrode surfaces. Relatively
intense and narrow peak currents are normally seen for the oxidation of metal ions from

amalgams. There seems to be minimal interference between co-deposited metals whereas

12

it is still an open challenge to get the same/better response from solid electrodes. A
comparison of the properties of Hg-based electrodes and diamond, pertinent to their use
in ASV detection of trace metal ions, is given in Table 1.2. Clearly, diamond possesses

many of the same athibutes as Hg.

Table 1.2 Characteristics and attributes of Hg and diamond electrodes for ASV.

 

 

 

Hg-based electrodes Diamond electrodes
Potential window limited by self- . . .
. . . . . Large working potential wrndow
oxrdation In the posrtrve range (~ _1 2 V to 1 5 V)
(~-l.2Vto 0.2 V) ' '
Surface can be renewed by depositing a Surface is stable with no need for
fresh ﬁlm of Hg renewal

 

Surface of heterogeneous conductivity
and microstructure, leads to broader
peaks

Uniform surface, shipping peaks are
narrow

 

Cannot be used for the stripping analysis
of metals with relatively positive E°
values (e. g., Au, Hg) or for metals
insoluble in Hg (e. g., As)

Can be used for the assay of metal
ions having relatively positive E°
values.

 

Highly toxic, volatile and unstable Non-toxic and chemically inert

 

Foreign metals deposit as a solid
phase, dependent on electrode surface
morphology and microstructure

Foreign metals deposit within the Hg
volume forming their amalgams

 

Higher background currents Very low background currents

 

Formation of interrnetallic compounds is Methods such as pulsed deposition are
reduced due to low surface area to being researched to yield low surface
volume ratio of the Hg drops area to volume ratio of metal deposits

 

 

 

 

Proper choice of the working electrode is crucial for the success of any ASV
measurement. The ideal electrode must meet the requirements of high electrical

conductivity, electrochemical inertness over the potential region of interest and chemical

13

inertness. Alternate electrodes being investigated are 11'26, Bi”, Auzs, Ag29 and graphite30 .
However, the overall response of most of these alternates has not approached that of Hg.
Some of the reasons for this include the heterogeneous surface structure of solid
electrodes and the higher background current compared to Hg electrodes.“7 Alternate
electrodes like “noble” metals31 and glassy carbon, can be intrinsically limited in terms of

sensitivity because of surface oxidation and reduction processes.

The practical utility of any electrode is determined by its effectiveness at
determining metal ions in real water samples. Kounaves et al. reported on the use of
microfabricated Au ultra-microelectrode arrays for the on-site analysis of As(III) in
groundwater.28 Limits of detection were investigated, and 0.05 ppb As(III) could be
measured while maintaining a S/N of 3/ 1. The same group reported on the use of a
microfabricated array of Ir micro-disks for Cu(II) analysis in tap and spring water
samples, and good agreement was obtained with conventional methods.26 An unexplained
effect of CT on the Ir surface was noted. Further investigation by atomic force
microscopy revealed that changes on the surface occurred but could be eliminated when
Cl' leakage from the reference electrode was minimized. Wang and coworkers reported
the use of Bi-coated, screen-printed carbon electrodes for ASV.32 These electrodes were
successfully used for the determination of Pb(II) in drinking water samples}2 Of all the
possible “alternate” electrodes, Bi appears to exhibit performance closest to that of Hg,
and it is environmentally friendly. However, the potential window of Bi is between -l.2
and -0.2 V vs. Ag/AgCl, and this limits its usefulness to the determination of metal ions

with standard potentials more negative than — 0.2 V.

14

Carbon electrodes have also been investigated for ASV. Recently, the
simultaneous determination of Cd(II) and Pb(II) at a glassy carbon electrode modiﬁed
with multi-wall carbon nanotubes was reported.33 Compared with bare glassy carbon, the
modiﬁed electrode exhibited improved sensitivity for these metal ions. A reason for this
observation could be the fact that multi-walled carbon nanotubes are chemically stable
and possess excellent electrical conductivity.34 The linear dynamic range for the detection
of Cd(II) and Pb(II) was reported to be from 2.5 x 10'8 to 1 x 10'5 M (2.8 ppb to 1.1
ppm) and from 2 X 10'8 to 1 X 10'5 M (4.1 ppb to 2.1 ppm), respectively, for deposition or

preconcentration times of 5 minutes.

1.4 Diamond Thin-Film as an “Alternate” Electrode

As will be shown in this dissertation, electrically conducting diamond is an
electrode material that shows great promise for application in ASV. Diamond is normally
a superb elechical insulator, but when doped with boron, the material possesses semi-
metallic electronic properties.1 For example, a doping level of ~1020 cm'3 results in an in-
plane resistivity of ca. 0.01 Q-cm, or less. Such a low resistivity is sufﬁciently adequate

to allow the material to be used in electroanalysis.35

In this relatively new area of research, it has been shown that diamond is useful
for the detection of Hg(II), Pb(II), Mn(II), Cd(II), Cu(II) and Ag(I).”42 The properties of
diamond are compared with those of Hg in Table 1.2. Clearly, diamond possesses many
of the same attributes as Hg. Boron-doped diamond electrodes have been employed for

the detection of Pb(II) in river sediments.” 43 It has also been successfully employed for

15

the detection of Mn(II) in tea samples.44 Prado et al. reported the simultaneous
electrochemical detection and determination of Pb(II) and Cu(II) at boron-doped diamond
electrodes.45 Independent detection of Cu(II) and Pb(II) via conventional standard
addition procedures was accomplished with diamond. The use of nitrogen-doped,
diamond-like carbon ﬁlm electrodes for the analysis of Pb (II), Cu (II), and Cd(II) was
also recently described.46 A linear dependence of the Pb(II) stripping peak current with
concenh'ation from 5 X 10‘7 to 2 X 10’6 M ( 100 to 400 ppb) was observed for a deposition

time of 2 minutes (with stirring).

Very recently, Kruusma and co-workers reported the electroanalytical
determination of Pb(II) by ASV using a Bi-modiﬁed, boron-doped diamond electrode.
Detection limits of 1.1 x 10'8 M (0.2 ppb) and 9.6 x 10'8 M (2.3 ppb) were obtained after
5 and 1 min deposition times, respectively.47 Babyak and Smart described the detection
of Cd(II) and Pb(II) at boron-doped diamond electrodes. The lowest concentrations
detected within the linear dynamic range (5 min deposition time) for Cd(II) were 10 ppb

and 50 ppb in KCl and KNO3, respectively, and 10 ppb for Pb(II) in KN 03.48

In summary, a few groups have recently reported on the use of conducting
diamond in ASV and shown that it is a viable electrode material for the detection of metal
ions. We recently reported the detection and quantiﬁcation of Ag (I), Cu (11), Pb (11), Cd
(II), and Zn (II) in several contaminated water samples using DPASV and a
nanocrystalline diamond thin-ﬁlm electrode.4 Diamond was shown to function well in

this electroanalytical application, providing a wide linear dynamic range, limits of

16

quantitation in the low ppb range, excellent response precision, and good response
accuracy.3 For the analysis of Pb(H), bare diamond provided a response nearly identical

to that obtained with a H g-coated GC electrode.

1.5 Outline of the Dissertation

Overall, the research effort sought to answer the key questions enumerated in the
ﬁrst section. The dissertation is organized as follows. In Chapter 2, the experimental
aspects of this work are described. Information about the diamond ﬁlm synthesis and
subsequent characterization by atomic force microscopy, visible-Raman spectroscopy,
cyclic voltammetry and scanning electrochemical microscopy is given. In Chapter 3, the
theory of metal phase formation (nucleation and growth) on solid electrodes is described.
The characteristic voltammetric responses for the metal ions studied are presented.
Results from the in-situ atomic force microscopic and chronoamperometric studies of Cu

deposition on diamond thin-ﬁlms are also described.

Boron-doped diamond thin ﬁlms of different morphologies (microcrystalline and
nanocrystalline) were used to assay for Ag(I), Pb(II), Cd(II), Cu(II), and Zn(II),
individually and in mixtures, using DPASV. A performance comparison of the two

diamond ﬁlm types with bare and Hg-coated GC is given in Chapter 4.

Chapter 5 reports on the detailed analysis of Ag(I), Cu(II), Pb(II), Cd(II), and
Zn(II) in contaminated water samples by DPASV using a boron-doped nanocrystalline

diamond thin-ﬁlm electrode. Data for the detection of these metal ions in four sample

17

types are presented: lake water, tap water, wastewater treatment sludge, and soil. The

data are validated through comparison measurements using ICP-MS and or AAS.

Intermetallic compound formation can be a complicating factor in ASV. A series
of anodic stripping voltammetric i-E curves, obtained by the oxidation of metal deposits
formed from binary solutions of Cu(II) and Zn(II) in varying ratios, were studied to learn
how much of a factor intermetallic compound formation is on diamond. Results and
discussion of these experiments is presented in Chapter 6. Our initial efforts using pulsed
galvanostatic deposition instead of constant potential deposition, to minimize
interrnetallic compound formation are also reported in chapter 6. The co-deposition of Cu

and Zn was used to evaluate the effectiveness of the pulsed current approach.

18

1.6
(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

References

Swain, G. M.; Anderson, A. B.; Angus, J. C. MRS Bulletin 1998, 23, 56-60.

Granger, M. C.; Xu, J .; Strojek, J. W.; Swain, G. M. Analytica Chimica Acta
1999, 397, 145-161.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chemistry of Materials
2003, 15, 879-888.

Sonthalia, P.; McGaw, E.; Show, Y.; Swain, G. M. Analytica Chimica Acta 2004,
522, 35-44.

Awada, M.; Strojek, J. W.; Swain, G. M. Journal of the Electrochemical Society
1995, 142, L42-L45.

Wang, J .; Swain, G. M.; Tachibana, T.; Kobashi, K. Journal of New Materials for
Electrochemical Systems 2000, 3, 75-82.

Friberg, L.; Nordberg, G. F.; Vouk, V. 3.; Editors, Eds. Handbook on the
Toxicology of Metals, 1979

Anon, Ed. Principles of biochemical toxicology, edited by John T imbrell, 2000
Goyer, R. A. American journal of clinical nutrition, 61, 6468-6508.
Online hlﬂjf/Al’lt’it‘. epa. gov/safewater/Incl. [11ml . (Accessed June 2004)

Klaassen, C. D.; Editor, Eds. Casarett and Doull's Toxicology: The Basic Science
of Poisons. 6th Ed; McGraw-Hill, 2001

Wang, J ., Ed. Stripping Analysis: Principles, Instrumentation, and Applications;
VCH Publishers, Inc., 1985

Wang, J ., Ed. Analytical Electrochemistry; John Wiley and Sons, Inc.: New York,
2000

19

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

Florence, T. M. Analytica Chimica Acta 1980, 119, 217-223.

Kissinger, P. T.; Heineman, W. R.; Editors, Eds. Monographs in Electroanalytical
Chemistry and Electrochemistry, Vol. 5: Laboratory Techniques in
Electroanalytical Chemistry, 1984

Birke, R. L. Analytical Chemistry 1978, 50, 1489-1496.
Dillard, J. W.; Hanck, K. W. Analytical Chemistry 1976, 48, 218-222.

Dillard, J. W.; Turner, J. A.; Osteryoung, R. A. Analytical Chemistry 1977, 49,
1246-1250.

Keller, H. E.; Osteryoung, R. A. Analytical Chemistry 1971, 43, 342-348.
Oldham, K. B. Analytical Chemistry 1968, 40, 1024-1031.

Osteryoung, R. A.; Parry, E. P. Journal of Electroanalytical Chemistry (1959-
1966) 1965, 9, 299-304.

Osteryoung, R. A.; Christie, J. H. Analytical Chemistry 1974, 46, 351-355.

Barker, G. C.; Gardner, A. W. At. Energy Research Estab. (Gt. Brit.) Rept. 1961,
AERE C/R-2297, 12 pp.

Barker, G. C.; Gardner, A. W. Zeitschriftfuer Analytische Chemie 1960, 1 73, 79-
83.

Wang, J.; Tian, B. Analytical chemistry, 65, 1529-1532.
Nolan, M. A.; Kounaves, S. P. Analytical Chemistry 1999, 71 , 3567-3573.

Hutton, E. A.; Ogorevc, B.; Hocevar, S. B.; Weldon, F.; Smyth, M. R.; Wang, J.
Electrochemistry Communications 2001, 3, 707-711.

Feeney, R.; Kounaves, S. P. Analytical Chemistry 2000, 72, 2222-2228.

20

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

Bonﬁl, Y.; Kirowa-Eisner, E. Analytica Chimica Acta 2002, 45 7, 285-296.

Brainina, K. Z.; Malakhova, N. A.; Ivanova, A. V. NA T 0 AS] Series, Series 2:
Environment 1997, 38, 255-269.

Laitinen, H. A.; Watkins, N. H. Analytical Chemistry 1975, 47, 1352-1358.

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

Wu, K.; Hu, 8.; Fei, J.; Bai, W. Analytica Chimica Acta 2003, 489, 215-221.

Kwok, K.; Chiu, W. K. S. Carbon 2005, 43, 437-446.

Xu, J .; Granger, M. C.; Chen, Q.; Strojek, J. W.; Lister, T. E.; Swain, G. M.
Analytical Chemistry 1997, 69, 59lA-597A.

Saterlay, A. J .; Tibbetts, D. F.; Compton, R. G. Analytical Sciences 2000, 16,
1055-1060.

Manivannan, A.; Seehra, M. S.; Tryk, D. A.; Fujishima, A. Analytical Letters
2002, 35, 355-368.

Manivannan, A.; Tryk, D. A.; Fujishima, A. Electrochemical and Solid-State
Letters 1999, 2, 455-456.

Saterlay, A. J .; Agra-Gutierrez, C.; Taylor, M. P.; Marken, F.; Compton, R. G.
Electroanalysis 1999, 11 , 1083-1088.

Ramesham, R. Journal of Materials Science 1999, 34, 1439-1445.

Rychen, P.; Perret, A.; Gobet, J .; Madore, C.; Haenni, W. Proceedings -
Electrochemical Society 2001, 2001-23, 97-107.

Peilin, Z.; J ianzhong, Z.; Shenzhong, Y.; Xikang, Z.; Guoxiong, Z. F resenius'
Journal of Analytical Chemistry 1995, 353, 171-173.

21

(43)

(44)

(45)

(46)

(47)

(48)

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

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

Prado, C.; Wilkins, S. J .; Marken, F .; Compton, R. G. Electroanalysis 2002, 14,
262-272.

Zeng, A.; Liu, E.; Tan, S. N.; Zhang, S.; Gao, J. Electroanalysis 2002, 14, 1294-
1298.

Kruusma, J.; Banks, C. E.; Compton, R. G. Analytical and Bioanalytical
Chemistry 2004, 3 79, 700-706.

Babyak, C.; Smart, R. B. Electroanalysis 2004, 16, 175-182.

22

CHAPTER 2

THE PHYSICAL CHARACTERIZATION AND
ELECTROCHEMICAL PROPERTIES OF BORON-DOPED
DIAMOND THIN-FILM ELECTRODES

2.1 Introduction

The structural and electrochemical characterization data for diamond thin-ﬁlm
electrodes deposited using microwave assisted plasma chemical vapor deposition (CVD)
are reported. Two types of conducting diamond with different surface morphologies and
microstructures were investigated and compared, microcrystalline and nanocrystalline
ﬁlms. The motivation for studying these two types of diamond stems from the fact that
the different morphology and microsh'ucture of each could signiﬁcantly inﬂuence the
nucleation site density, particle density and particle size of the metal deposits. This could,

in turn, affect the electrode response for the metal ions.

Table 2.1 lists the instrumental techniques used in our laboratory for
characterizing of the physical, chemical and electronic properties of diamond thin-ﬁlms
along with the information gained from each.“ 2 Results from the characterization by
atomic force microscopy, visible Raman spectroscopy, x-ray diffraction and cyclic
voltammetry, using Fe(CN)6'3/4, Ru(NH3)6+3/+2, IrCl(,'2/'3 and methyl viologen as redox
test probes, are presented in this Chapter. Scanning electrochemical microscopy (SECM)
was also used to characterize the spatial electrochemical activity of the diamond

electrodes using probe approach curves in the positive feedback mode.3’ 4

23

Table 2.1 Instrumental Techniques used for characterization.

 

 

 

 

 

 

 

Technique In ormation
SEM Morphology
AFM Morphology
Raman Microstructure
XRD Crystallinity
4-pt probe Resistivity
SECM Local Electrochemical Activity

 

 

 

 

2.2 Experimental

Film Growth and Characterization. The diamond thin ﬁlms were deposited on p-

type Si(100) substrates ( ~10'3 Q-cm, Virginia Semiconductor Inc., Fredricksburg, VA),
using a commercial microwave-assisted chemical vapor deposition (CVD) system (1.5
kW, Astex, Inc., Lowell, MA). The Si surface was prepared for diamond growth by
mechanically scratching the surface on a felt pad with 1 pm diameter diamond powder
(GE Superabrasives, Worthington, OH). The scratched substrate was then sequentially
washed with ultrapure water, isopropyl alcohol (IPA), acetone, IPA, and ultrapure water.
Finally, it was ultrasonicated for 5 min each in acetone and ultrapure water. The cleaned

substrate was placed in the CVD reactor.

24

The microcrystalline ﬁlms were grown using a CH4/H2/B2H6 source gas mixture
at a CH4/H2 volumetric ratio of 0.50%, a total gas ﬂow of 200 seem, a microwave power
of 1000 W, a growth pressure of 45 torr, an estimated substrate temperature of 800°C, and
a growth time of ca. 10 h. Ultrahigh purity (99.999%) CHx, at a ﬂow rate of 1 seem, and
H2 at a ﬂow rate of 197 sccm were used. B2H6, diluted in H2 (0.1%) was used for boron
doping at a ﬂow rate of 2 seem. The concentration of B2H6 in the source gas mixture was
10 ppm (1x10'3%). Aﬁer growth, the CH4 and B2H6 ﬂows were stopped and the ﬁlms
remained exposed to the hydrogen plasma (atomic hydrogen) for an additional 10 min.
Next, the microwave power was gradually reduced over a 15 min period down to 400 W
in order to cool the samples in the presence of atomic hydrogen. Post growth annealing in
atomic hydrogen is necessary for full hydrogen surface termination and ensuring the

removal of adventitious non-diamond sp2 carbon impurity.

For the nanocrystalline ﬁlms, ultrahigh purity CH4, Ar, and H2 (99.999 %) were
used as the source gas mixture. The gas ﬂow rates were 1, 94, and 5 seem, respectively,
for CH4 ( 1%), Ar (94%), and H2 (5%). B2H6 was used for doping at a ﬂow rate of 0.001
sccm (10 ppm). The microwave power and deposition pressure were maintained at 800 W
and 140 torr, respectively. The substrate temperature was estimated to be about 800 °C
using an optical pyrometer. The deposition time was 2 h. The plasma was ignited and
deposition initiated with all the gases ﬂowing into the reactor. At the end of the
deposition period, the CH4 and B2H6 ﬂows were stopped, and the Ar and H2 ﬂows

continued. The ﬁlm was exposed to H2/Ar plasma for 10 min. The substrate was then

25

cooled in the presence of hydrogen to an estimated temperature of less than 300 ° C by

slowly reducing the power and pressure over a 4 min period.

The resulting diamond thin ﬁlms were approximately 4-6 pm thick, as estimated
from cross sectional SEM images. The nominal resistivity was found to be 0.01 Q-cm, or

less, for the microcrystalline and 0.1 Q-cm, or less, for the nanocrystalline ﬁlms using the

four-point probe method.

The surface morphology of the boron-doped diamond thin-ﬁlms was investigated
by the atomic force microscopy (N anoscope IIIa, Digital instruments, Santa Barbara, CA).
The crystalline structure was characterized by Raman spectroscopy.5 The turn key
instrument consisted of a diode-pumped, frequency-doubled CW Nd: yttrium aluminium
garnet (YAG) excitation laser of 523 nm and objective lens of x 20 (Raman 2000,
Chromex, Inc Albuquerque, NM). Spectra were collected with an incident power density
of ~ 500 kW/cmz, a 10 5 integration time and spectral resolution of about 2 cm”. A white
light spectrum collected under the same conditions was used for background correction.
X-ray diffraction (XRD) was employed to characterize the bulk crystallinity of the
diamond thin ﬁlms. A Rigaku Rotaﬂex diffractometer equipped with a rotating anode and
a Cu K-oz radiation source was used to obtain the XRD patterns. The thin ﬁlm rotates by

an angle, 0, with respect to a ﬁxed x-ray source, and the detector rotates by 20 to measure

the reﬂected light. The measurements were made from 20 of 30° to 100°.

26

Electrochemistry. A single-compartment glass cell was used for all the

electrochemical measurements, along with a computer-controlled potentiostat (Model
650a CH Instruments Inc., Austin, TX).6 The diamond thin-ﬁlm working electrode was
pressed against a Viton® o-ring and clamped to the bottom of the cell. The backside of
the Si substrate was scratched with a carbon rod to ensure good ohmic contact with the
current-collector plate. A graphite rod was used as the counter electrode, and a
commercial Ag/AgCl electrode (saturated KCl) served as the reference (EOAg/Agc. = -0.045
V vs. SCE). The geometric area of the working electrode was ca. 0.163 cmz. All
measurements were made in solutions deoxygenated with N2 for at least 5 min initially.
The solutions were blanketed with the gas during all measurements. The reference
electrode was placed inside a cracked-glass capillary (double junction) that was ﬁlled
with 1 M KCl. All measurements were performed at room temperature, ~ 25 OC. The
supporting electrolyte was 1 M KCl. The concentrations of Fe(CN)6'3M, Ru(NH3)6 ””3,

IrClr,‘2/'3 and methyl viologen were 1 mM.

SECM experiments were conducted using a CHI 900 SECM (CH Instruments,
Austin, TX). The tip could be moved in the x,y and 2 directions using an inch-worrn
controller. The x,y plane is along the substrate surface and the 2 plane is perpendicular to
this plane. For approach curves, the tip was moved towards the substrate surface in 0.2
urn increments (along the —z axis). The approach was terminated when the tip current

reached 115% (positive feedback) and 75% (negative feedback) of its initial value. The
initial value of the current, i190, is deﬁned as the current when the tip was “inﬁnitely” far

from the surface.

27

Pt probe tips were obtained from CH Instruments (Austin, TX). They were 2 um
diam. Pt wire sealed into glass. The SECM cell was made of .Teﬂon. Pt wire was used as
a counter electrode and an Ag/AgCl electrode served as the reference for all
measurements. Prior to use the tips were cleaned by potential cycling in 0.5 M H2SO47
between -0.3 V and 0.9 V at 100 mV/s fori112.5 cycles. The resulting cyclic voltammetric
i-E curves showed some/all the features for Pt, Pt oxide formation, oxygen evolution, Pt
oxide reduction, hydrogen adsorption and desorption and hydrogen evolution. The

+3/+2

voltammehic i-E curves for Ru(NH3)(, , at the ultra-microelectrode tips, showed

/ . . . .
+3 +2 was chosen as a mediator because its kinetics

classical sigmoidal shapes. Ru(NH3)(,
for electron transfer are mainly dependent on the density of electronic states and would
indicate the charge carrier density at the formal potential of the material being examined.
1 mM Ru(NH3)6 ””3 in 1 M KCl was used as the redox mediator for all the probe-
approach measurements. The solution was degassed using N2 for 5 minutes prior to

introduction into the cell. For all the probe approach measurements, the tip was poised at

a potential of -0.3 V and the substrate at a potential of 0.05 V.

2.3 Results and Discussion

2.3.1 Structural Characterization.

Atomic force microscopy (AFM) provides information about the ﬁlm
morphology, while Raman spectroscopy is useful for evaluating the ﬁlm microstructure
and for detecting the presence of any localized non-diamond spz-bonded carbon impurity.
Figure 2.1. shows 5x5 um2 AFM images (height mode, ex-situ) of microcrystalline and

nanocrystalline thin-ﬁlm electrodes. The microcrystalline ﬁlm is polycrystalline and

28

well-faceted with randomly-oriented crystallites having lateral dimensions of 0.5 to 4 pm
at the base. The nanocrystalline ﬁlm is also polycrystalline but much smoother with
feature sizes of ca. 100 nm in lateral dimension. These features are actually clusters of
randomly-oriented diamond grains that are approximately 15 nm in diameter, as revealed
by transmission electron microscopy (TEM).8 As a consequence of the much smaller
grain size, the nanocrystalline ﬁlm contains a much higher fraction of grain boundary.

The abrupt grain boundaries consist of 1: - bonded carbon atoms.9

 

Figure 2.1 5x5 um2 AF M images (height mode) of boron-doped (A) microcrystalline and
(B) nanocrystalline diamond ﬁlms.

Figures 2.2.A and 2.2.B show macro—Raman spectra for the microcrystalline and
nanocrystalline diamond thin-ﬁlm electrodes, respectively. Raman spectroscopy is a
nondestructive technique often used to determine the nanostructure of carbon thin-ﬁlms
due to its ability to discriminate between the sp2 and sp3-bonded carbon as well as the
local bonding environment of carbon.10 The ﬁrst-order diamond phonon line is observed

at 1332 cm'I for the microcrystalline ﬁlm, which originates from the sp3-bonded carbon

29

microstructure.9 For the microcrystalline diamond ﬁlm, weak asymmetry and broadening
appears in the one phonon band and also an appearance of a new band centered at around
1200 cm'1 is observed. This has been attributed to the high boron doping effect.11 As
diamond thin-ﬁlms undergo a. transition from a microcrystalline to nanocrystalline
morphology, the Raman spectra become dominated by scattering from the spZ-bonded
carbon located at the grain boundaries.12 Due to resonance effects, the Rarnan cross-
section scattering coefﬁcient (visible excitation) is larger (50X) than that for the sp3-
bonded diamond.l3 At the nanocrystalline diamond, there are peaks observed at 1332 cm‘
1 and near 1500 cm". These peaks are the D-band and G-band routinely observed in spz-
bonded carbon.12 The diamond peak at 1332 cm'1 is not visible above the signal
generated by the spz-bonded carbon. The 1550 cm" peak is down shifted from the
expected 1580 cm'1 position for graphite. This shift results because the spZ-bonded
carbon is mixed with sp3-bonded carbon. Further there is also a peak observed at 1140
cm’l in the nanocrystalline ﬁlm. It was widely believed that this peak originates from the
presence of conﬁned phonon modes in diamond, due to the reduced crystallite size of
nanocrystalline diamond.14 However, Ferrari et al. suggested that these features may have
their origin in spz-bonded carbon.” More recently this peak has been reported to be

speciﬁcally caused by carbon-hydrogen bonds in the grain boundaries.12

30

3.00 25.0

 

 

 

= 1333 I

250* A l * B frimcml

- - , 20.0- 1 2550mm
2,» 2.00 — l m»
g _ , 15.0- 1140?"
3 1.50 100.
E. 1.00 ' 7

“TM w
0.50 5.0_
i l r
000

 

 

 

 

 

. 1 I 1 l r l 1 l L 1 l g- . L l 1 l 1 l 1 l 1 l 1 l
700 900 1100 1300 1500 1700 1900 0 900 900 1100 1300 1500 1700 1900
Raman Shift (crrr‘)

Figure 2.2 Visible macro-Raman spectra of (A) microcrystalline and (B) nanocrystalline
diamond ﬁlms

X-ray diffraction (XRD) was employed to characterize the bulk crystallinity of
the diamond thin-ﬁlms.8 Figure 2.3 shows the XRD spectra for a (A) microcrystalline and
(B) nanocrystalline diamond thin ﬁlm. For both ﬁlms, three reﬂections are observed at
20 values of 44°, 76°, and 92°, and are assigned to the (111), (220), and (311) planes of
cubic diamond, respectively.“ ‘6 Well-deﬁned and sharp (111) peaks indicate high quality
of the diamond ﬁlms. The peaks for the nanocrystalline ﬁlm are broader than those for
the microcrystalline ﬁlm due to the smaller grain size. This result conﬁrms that both ﬁlm

types possess a crystalline diamond structure.

31

 

 

 

 

 

 

 

 

 

 

. 2 r
E 1200’ V i
g . 1000-
o 1&1)» '
E“. ’ 80°"
H 800- .
ea ’ ax),-

C. 6(1)-
8 ’ r
4m».
5 400- ’
b 8 2
20°" :3 .2: 20°"
0....1 -1: 0.....11 . 1..
40 so 80 100 40 so so 100

 

29 ((1615166)

Figure 2.3 XRD patterns of (A) microcrystalline and (B) nanocrystalline diamond ﬁlms

2.3.2 Electrochemical Characterization

Figure 2.4 shows background cyclic voltammetric i-E curves for boron-doped (A)
microcrystalline and (B) nanocrystalline diamond thin ﬁlms at 100 mV/s in 1.0 M KCl.
The curves shown are for the tenth scan, and are ﬂat and featureless over much of the
potential range. The working potential window is about 2.7 V (i 100 11A) with a
featureless response between the working potential limits. The anodic current ﬂowing at
1150 mV is due to the oxidation of chloride to chlorine. The cathodic current ﬂowing at —

1400 mV is due to the reduction of water to hydrogen.

At 500 mV, the anodic current for the nanocrystalline ﬁlm is 1.4 [AA (7 uA/cmz)

and the anodic current for the microcrystalline ﬁlm is 1.5 uA (8 uA/cmz). The current for

32

both diamond electrodes, however, is less than that for freshly polished glassy carbon
(GC), 7-10 uA (35—50 uA/cmz). A low background current and wide working potential
window (3 to 4 V) are characteristic features of high quality, boron-doped diamond."
This leads to improved S/N ratios in electroanalytical measurements.‘8 The lower
potential window seen for the two diamond ﬁlms is due to the fact that KC] was used as
the supporting electrolyte rather than acid (HClO4) and chlorine evolution occurs on

diamond with relatively low overpotential.

20

 

15 -

10 —

 

 

Current (11A)

 

 

 

l 4

-2 l A I A I A l g ‘ A l

0 r l .
-150&1200-900-600-30 0 300 600 900 1200
Potential (mV vs. Ag/AgCl)

Figure 2.4 Background cyclic voltammetric i-E curves in l M KCl for (A)

 

microcrystalline diamond ( ) and (B) nanocrystalline diamond ( ------- ) thin-ﬁlm

electrodes.

The electrochemical responsiveness of the boron-doped microcrystalline and

nanocrystalline diamond thin ﬁlms towards four different redox systems was investigated

using cyclic voltammetry. The physical, chemical and electronic properties of the

33

diamond inﬂuence the electrode kinetics and reaction mechanism for these systems." 8’ '9’
20 Figure 2.5 shows representative cyclic voltammetric i-E curves for (A) Fe(CN)6'3M,
(B) Ru(NH3)6+2/+3, (C) IrC16'2/'3, (D) methyl viologen (MVW 1) at a microcrystalline
diamond thin-ﬁlm electrode. The electrolyte was 1 M KCl and the potential sweep rate
ranged from 50 to 2000 mV/s. Ey, for these redox systems ranges from approximately

800 to -1100 mV, so they are useful for probing the ﬁlm’s electronic properties over a

wide potential (i.e., energy) range.

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

2
3.
v
H
t:
at): .
4o-
:3 , 40r
C.) 20_ -
20-
O_ l-
o.
-20.. '
. ~20-
-40_ h
. 40-
m’ r
r 1 . 1 . r . l . .60 l A l A 1 . 1 .
500 600 700 800 900 1000 4200 4000 -800 ~600 400

Potential (mV vs Ag/AgCl)

Figure 2.5 Cyclic voltammetric i-E curves for (A) Fe(CN)6'3"‘, (B) Ru(NH3)6*2’+3, (C)
IrC16'2/'3, and (D) methyl viologen (MV+2/+) at increasing scan rates of 50, 100, 200, 500,
1000 and 2000 mV/s.

34

A nearly reversible response is observed for Fe(CN)6’3M with a AEp of 69 mV at
100 mV/s and an ipoxlipred of 1.0. Fe(CN)6'3/4 is a surface-sensitive redox system at both
glassy carbon and boron-doped microcrystalline diamond.20 The electrode reaction
kinetics for this couple is strongly inﬂuenced by the amount of exposed edge plane on
spZ—bonded carbon as well as the surface cleanliness.21 Granger and co-workers showed
that surface carbon-oxygen functionalities on microcrystalline diamond signiﬁcantly
inﬂuence AEp with increasing oxygen content causing a proportional increase in AEP. The
small AEp seen for the boron-doped microcrystalline diamond ﬁlm is indicative of a high

level of surface cleanliness and a low surface oxide coverage.22

A nearly reversible response is also seen for Ru(NH3)(,+2/+3 with a AEp of 71 mV
at 100 mV/s and an ip""/i,,"ed of 1.0. A nearly reversible response is also seen for IrC16'2/'3
at 100 mV/s with a AEp of 69 mV and an i,,°"/ipred of 1.0. The electrode reaction kinetics
for both of these redox systems are not as sensitive to the physicochemical properties of
both spz-bonded carbon and diamond electrodes as is Fe(CN)6'3M.9‘ 2° The kinetics are
mainly inﬂuenced by the density of electronic states at the potentials where the
electrolysis occurs (i.e., near the E°’). In other words, the diamond electrode is
sufﬁciently conductive at both —200 mV, a potential negative of the ﬂatband potential
(ca. 0.5 V vs. Ag/AgCl) for p-type semiconducting diamond, and 800 mV, a potential

positive of the ﬂatband potential.17

A reversible response is seen for MVHMl with a AEp of 62 mV and an him/1,,”d of

1.0. Like Ru(NH3)6+2/+3 and IrC16'2/'3, the electrode reaction kinetics for MVWH are

35

relatively insensitive to the physicochemical properties of both spz-bonded carbon and
diamond electrodes.20 The kinetics are mainly inﬂuenced by the density of electronic
states at the potential where the electrolysis occurs. MV undergoes two redox processes.
Good responsiveness was also observed for the other redox couple, MVH/O. The cathodic

peak at ~99O mV is associated with the reduction of MV” to MVO. The peak shape is

. . . . . . . . / /
consrstent wrth a drffusron-lrmrted reactron. Plots of 1p” versus scan rate12 (V'2

) were
constructed for the four redox couples in question at the various scan rates studied. It was
observed that the oxidation peak current for each redox system varied linearly with the
vm, with a near-zero y-axis intercept, indicative of reactions limited by semi-inﬁnite
linear diffusion of reactants to the electrode surface. Similar cyclic voltammetric results
for the four couples were observed on the nanocrystalline diamond. The quasi-reversible
voltammetry for all the couples indicates that both the kinds of diamond have a sufﬁcient

charge carrier density over the wide potential range to support relatively rapid electron

transfer.

2.3.3 Electrochemical Activity as Studied by the Scanning
Electrochemical Microscopy (SECM)

Probe Approach Curves. SECM can be used to study the spatial electrochemical

activity of an electrode and distinguish between conducting and insulating regions on a
substrate through probe-approach curves. In the feedback mode, an ultra-microlectrode
(UME) tip is perturbed from its value in bulk solution by a much larger substrate in close
proximity.23 For example, if an oxidizing species “0” in the bulk solution is reduced at
the UME tip, and the substrate is conductive and is held at a sufﬁciently positive

potential, then the tip-generated reduced species “R” is oxidized back to “O” at the

36

substrate and the tip current becomes larger than its magnitude when it is “inﬁnitely” far

from the substrate, as denoted by iTm. This is called the positive feedback mode. At

insulating substrates, no oxidation of “R” occurs and the tip current decreases from iT,00

due to blockage of the diffusional ﬂux of “R” to the tip (negative feedback). For all the
probe approach measurements, the tip was poised at a potential of -0.3 V and the

substrate at a potential of 0.05 V. In Figure 2.6, an example of negative feedback is

shown. As the tip, set at a reducing potential, approaches the substrate i100 decreases

because the ﬂux of reactant to the tip is inhibited when the tip is in close proximity to the

substrate. All tip current values are normalized to ir,oo- L is the normalized tip-substrate

distance (d/a), where d is the distance traveled by the tip and a is the tip electrode radius.

 

1.00 -

0.95 -

1011,00

0.90 -

0.85 -

 

 

0.80

 

 

Figure 2.6 Tip-approach curve at a polycarbonate substrate. The data were acquired in a
1 mM solution of Ru(NH3)63+ in 1 M KCl. The tip is a 10 rim-radius Pt disk. The tip was
biased at -0.3 V.

37

Figure 2.7 shows representative positive feedback current-distance approach
curves obtained using boron-doped microcrystalline and nanocrystalline diamond and a
Pt ﬁlm as substrates and a 2 um Pt tip as the probe. Five to six spots on the diamond
surface were probed. Positive feedback was observed at all the locations and the pattern
of the curves and magnitude of the current enhancement was similar. To facilitate
comparison between approach curves, the tip currents were normalized to the steady-state
values obtained at inﬁnite distance. The theoretical curve for a uniformly active substrate
has also been shown in the Figure. The data for the diamond ﬁlms indicate that these
electrodes are electrically conducting with fast heterogeneous kinetics for Ru(NH3)(,+3/+2
at this potential. The activity is comparable to Pt. Furthermore, probe-approach curves
recorded for the microcrystalline diamond were reproducible for multiple spots whereas
those on nanocrystalline diamond varied slightly, possibly indicating that this surface was
more heterogeneous in conductivity. Microcrystalline diamond electrodes have fewer
grain boundaries separating the individual crystallites. The grain boundaries are known to
be rich in spz-bonded carbon while the grains predominantly contain sp3-bonded carbon.
The microcrystalline diamond is thus expected to be microstructurally homogeneous
compared to the nanocrystalline variety. However the spZ-bonded carbon grain
boundaries are known to be more electrically conductive. The behavior of the
nanocrystalline diamond was observed to be closer to the theoretical curve indicating that

this diamond type has a higher overall surface conductivity than the microcrystalline

variety.

38

According to Amphlett and Denault, an equation for a current-distance curve
controlled by a positive feedback is given below.24 They showed how the four
coefﬁcients, k1, k2, k; and k4, change with RG (ratio of the insulating glass sheath radius

to that of the electroactive disk) values,
itip/itipm = k] + (kg/L) + k3 exp (kt/L) (2.1)

The theoretical curve depicted in Figure 2.7 has been constructed using the above
equation and values of the constants interpolated to an RG ratio of 6.5. The values of the

four coefﬁcients k1, k2, k3 and k4 were 0.7220, 0.7555, 0.2647 and 1.5640, respectively.
The current values were plotted upto a limit of 115% of the i100, to facilitate comparison

to the experimental curves. The theoretical curve has a current increase further away from
zero than the experimental curves, either due to a different true RG value or due to the
fact that the experimental substrates may not be uniformly reactive. Various RG values
ranging from 1.5 to 7.5 were tested, however, the experimental and theoretical curves did
not coincide. To explain this data, the theory of ﬁnite heterogeneous electron-transfer

kinetics and arbitrary substrate size has been discussed next.

39

 

 

Pt

 
  
   

 

 

 

1.14 _ - ‘
------ microcrystalline
. ----- nanocrystalline
1.12 - 1' theoretical
1.10 - i .
. i
1.08 - '=
8 i-
€~ 1.06 -
.._r* r
1.04 -
1.02 -
1.00 - "
0 5 10 15 20
L

Figure 2.7 Tip approach-curves for the different electrode materials: platinum,
microcrystalline and nanocrystalline diamond. The simulated current-distance data for a
tip with an RG ratio 6.5 is also shown. Experimental curves were acquired in a 1 mM
solution of Ru(NH3)63+ in 1 M KC1. The tip is a 2 um-radius Pt disk of RG ratio between
6 and 7. The tip was biased at -0.3 V and the substrate at 0.05 V.

The theoretical curve in the previous section was calculated based on the
assumption of a macroscopic (i.e., inﬁnitely large compared to the tip), uniformly active
substrate. This assumption may not be accurate when the substrate contains three-
dimensional features, or consists of regions of different elechical conductivity or
electrochemical activity. A schematic model showing SECM in the feedback mode with
substrate of ﬁnite dimensions is shown in Figure 2.8. Only a portion of the substrate is
active in terms of the regeneration of 0 from R. Very recently, the elechical conductivity
and electrochemical activity at the surfaces of hydrogen-terminated boron-doped

diamond electrodes, with different boron doping levels, was measured using conductive-

probe atomic force microscopy (CP-AFM) and scanning electrochemical microscopy

4O

(SECM).25 SECM approach curves obtained using Ru(NH3)(,3+’2+ as the redox mediator
were characteristic of those obtained at partially blocked electrodes of non-uniform
activity or a microelectrode array. The active area of the electrode was related to the
boron-doping level, however, even the heavily doped (10 ppm source gas B2H6) diamond

electrode showed the behavior of partially blocked electrodes.

Ultra-microelectrode tip

0 R diffusion

 

insulator i, ‘ . insulator

conductive zone

 

 

 

 

Figure 2.8 Schematic illustration of the SECM feedback mode with a ﬁnite substrate
(Adapted from reference 26). The heterogeneous redox reactions occur at a diffusion-
controlled rate at the ultra-microelectrode and conductive portion of the tip.

Further, Bard et a1. studied the effect of ﬁnite substrate size by simulating current-
distance curves over various sizes of microdisk substrates using two methods,
multidimensional integral equations and the ﬁnite difference method?" The working
curves, for various substrate sizes, representing the dependence of the steady-state
diffusion-controlled current on the dimensionless distance between the tip and substrate
are presented in Figure 2.9. Parameter, h, has been deﬁned as:

h = as/a where: as is the active substrate radius and a is the tip electrode radius.

41

 

 

 

 

5.0 - A . h
E 3.% --- 00
4'0 T1 1 a o 1.0
i i‘ 1'3 A A 0.7
8' 1 \i 6 I i 0.5
i 3'0 :1'2‘ iii-1‘ ' v 0.3
I" 1‘ Rik. like 0 0.1
2.0 : "1‘ I...“ .‘jr‘~‘
(3‘ i ‘1‘» f "1“‘u-c.‘
R km ..." 'G... ------ _ ________ -..,
-~ u . _ - J
10 ' v’ v - ‘. !-_ ‘ i '
V ":= 7.3;. z 1. . .7 _.
6 Q —- Q A
0 0 0 5 1.0 1 5 20

Figure 2.9 Simulated i/ipm vs. L curves for ﬁnite disk-shaped substrates: (ﬁlled symbols)
multidimensional integral equations simulations; (open symbols) ﬁnite difference method
simulation. The h values are indicated. (Figure adapted from reference 26)

Bard et al. found that with decreasing h, the current—distance curves have a
shallow slope and a sharp current rise close to L = 0. Comparing the two ﬁgures 2.7 and
2.9, it seems that the microcrystalline diamond has a slightly smaller “h” parameter (i.e.,
smaller conducting substrate features) than nanocrystalline diamond. This is again
explained by the fact that the nanocrystalline diamond has a higher fraction of exposed
grain boundary compared to the microcrystalline variety. As mentioned, the grain
boundaries are more conductive than the grains.8 The resolution of the instrument is a
limiting factor. It is only possible to identify particles 10-20 times smaller than the tip by
SECM.26 With a tip 2 pm in diameter, this corresponds to 100-200 nm in size. The
inﬂuence of the grains of the nanocrystalline diamond which are only 10-15 nm on the
probe-approach curve may not have been resolvable. The curve for the Pt substrate also

did not trace the theoretical curve, probably due to non-uniform conductivity caused by

42

the presence of Pt surface oxides. Overall, it can concluded that since the current-distance
curves obtained for diamond electrodes of both morphologies are similar to the one
obtained for Pt macroelectrode, it seems that at this doping level, the diamond substrate

can be treated as sufﬁciently conductive.

2.4 Conclusions

The physical characterization and electrochemical properties of diamond thin-ﬁlm
electrodes was reported on. Two types of conducting diamond with different surface
morphologies and microstructures were investigated and compared, microcrystalline and
nanocrystalline thin-ﬁlms. Based on the electrochemical results, it is clear that both
diamond types have a sufﬁcient charge carrier density over a wide potential range to
support relatively rapid electron transfer for several analytes. This evidence suggests that
both types of diamond ﬁlms are electrically conductive and electrochemically active
electrodes. From the SECM approach curve measurements, it was demonstrated that
diamond electrodes of both morphologies showed a positive feedback response that was
comparable to a Pt macro-electrode. Probe-approach curves recorded for the
microcrystalline diamond were reproducible for multiple spots whereas those on
nanocrystalline diamond varied slightly, possibly indicating that this surface was more
electrically heterogeneous. However, since the nanocrystalline diamond has a higher
fraction of grain boundary than does microcrystalline diamond, a larger fraction of the
former surface was determined to be elechically conducting (i.e., electrochemically
active). Sufﬁcient understanding of how the grth conditions of the diamond ﬁlms

affect the efﬁciency of stripping will enable development of even superior electrodes in

43

the ﬁrture. For instance, previous studies in our group, have shown that the copper
nucleation mechanism on boron doped diamond ﬁlm strongly depends on boron doping

level of diamond ﬁrm.27

44

2.5 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Xu, J .; Granger, M. C.; Chen, Q.; Strojek, J. W.; Lister, T. E.; Swain, G. M.
Analytical Chemistry 1997, 69, 591A-597A.

Hupert, M.; Muck, A.; Wang, J .; Stotter, J .; Cvackova, Z.; Hayrnond, S.; Show,
Y.; Swain, G. M. Diamond and Related Materials 2003, 12, 1940-1949.

Bard, A. J .; Mirkin, M. V.; Editors, Eds. Scanning Electrochemical Microscopy;
Marcel Dekker, Inc., 2001

Bard, A. J .; Fan, F.-R.; Mirkin, M. Physical Electrochemistry 1995, 209-242.

Dennison, J. R.; Holtz, M.; Swain, G. Spectroscopy (Eugene, Oregon) 1996, I 1,
38-46.

Xu, J .; Chen, Q.; Swain, G. M. Analytical Chemistry 1998, 70, 3146-3154.

Seegmiller, J. C.; Buttry, D. A. Journal of the Electrochemical Society 2003, 150,
B413-B418.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chemistry of Materials
2003, 15, 879-888.

Granger, M. C.; Xu, J .; Strojek, J. W.; Swain, G. M. Analytica Chimica Acta
1999, 397, 145-161.

Solin, S. A.; Ramdas, A. K. Physical Review B: Solid State 1970, [3] 1, 1687-
1698.

Ushizawa, K.; Watanabe, K.; Ando, T.; Sakaguchi, 1.; Nishitani-Gamo, M.; Sato,
Y.; Kanda, H. Diamond and Related Materials 1998, 7, 1719-1722.

Birrell, J .; Gerbi, J. E.; Auciello, 0.; Gibson, J. M.; Johnson, J .; Carlisle, J. A.
Diamond and Related Materials 2005, 14, 86-92.

45

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

Knight, D. S.; White, W. B. Journal of Materials Research 1989, 4, 385-393.

Nemanich, R. J .; Glass, J. T.; Lucovsky, G.; Shroder, R. B. Journal of Vacuum
Science & Technology, A: Vacuum, Surfaces, and Films 1988, 6, 1783-1787.

Ferrari, A. C .; Robertson, J. Physical Review B: Condensed Matter and Materials
Physics 2001, 63, 1214051-1214054.

Sharda, T.; Misra, D. S.; Seibt, E. W.; Selvam, P. Journal of Vacuum Science &
Technology, A: Vacuum, Surfaces, and Films 1998, 16, 413-418.

Swain, G. M.; Anderson, A. B.; Angus, J. C. MRS Bulletin 1998, 23, 56-60.

Strojek, J. W.; Granger, M. C.; Swain, G. M.; Dallas, T.; Holtz, M. W. Analytical
Chemistry 1996, 68, 2031-2037.

Witek, M.; Wang, J .; Stotter, J .; Hupert, M.; Haymond, S.; Sonthalia, P.; Swain,
G. M.; Zak, J. K.; Chen, Q.; Gruen, D. M.; Butler, J. E.; Kobashi, K.; Tachibana,
T. Journal of Wide Bandgap Materials 2002, 8, 171-188.

Granger, M. C.; Witek, M.; Xu, J .; Wang, J .; Hupert, M.; Hanks, A.; Koppang, M.
D.; Butler, J. E.; Lucazeau, G.; Mermoux, M.; Strojek, J. W.; Swain, G. M.
Analytical Chemistry 2000, 72, 3793-3804.

Kneten, K. R.; McCreery, R. L. Analytical Chemistry 1992, 64, 2518-2524.

Granger, M. C.; Swain, G. M. Journal of the Electrochemical Society 1999, 146,
4551-4558.

Kwak, J .; Bard, A. J. Analytical Chemistry 1989, 61, 1794-1799.

Amphlett, J. L.; Denuault, G. Journal of Physical Chemistry B 1998, 102, 9946-
9951.

Holt, K. B.; Bard, A. J .; Show, Y.; Swain, G. M. Journal of Physical Chemistry B
2004, 108,15117-15127.

46

(26) Bard, A. J .; Mirkin, M. V.; Unwin, P. R.; Wipf, D. 0. Journal of Physical
Chemistry 1992, 96, 1861-1868.

(27) Wang, J .; Swain, G.M; Unpublished research 1999.

47

CHAPTER 3

ELECTROCHEMICAL AND ATOMIC FORCE MICROSCOPY
STUDIES OF METAL PHASE NUCLEATION AND GROWTH ON
BORON-DOPED DIAMOND THIN-FILM ELECTRODES

3.1 Introduction

Electrochemical metal phase formation is a surface sensitive redox process which

generally follows a nucleation-and-growth mechanism.l

The shape, coverage and
location of the initial metal deposits are inﬂuenced by the inherent rate of electron
transfer and also by the physical, chemical and electronic properties of the electrode
material.“ Previous work in our laboratory has established the feasibility of
electrodepositing metals on the diarrrondf"7 In this Chapter, a review of some of the

theoretical aspects of metal nucleation and grth is given along with some preliminary

studies of metal phase formation on boron-doped diamond thin-ﬁlm electrodes.

3.2. Theory of Metal Nucleation and Growth

Electrodeposition occurs by a process of nucleation and growth. Nuclei form at
active sites on the electrode according to a nucleation rate law, and then grow via the
incorporation of additional ions from the surrounding solution.8 Metal electrodeposition
takes place at electrode/electrolyte interfaces under the inﬂuence of an electric ﬁeld and
involves solid phase formation.9 Derek Pletcher in his book, “A First Course in Electrode
Processes,” explains the steps involved in the deposition of a metal onto an inert
cathode.IO These are depicted in Figure 3.1 (adapted from his book) and are: (1) transport
of solvated metal ions through the solution to the electrode surface; (ii) electron transfer;

(iii) partial or complete loss of the solvation sheath to form an adatom; (iv) surface

48

diffusion of adatoms; (v) clustering of the adatoms to form a nucleus of sufﬁcient size for
it to be stable; and (vi) growth of the nucleus by incorporation of adatoms at favorable

sites in the lattice structure of the metal.

).@ ion

bulk diffusion

 

 
  

solution

electrode I ad— atom growth center
c.

Figure 3.1 The stages in the formation of a stable, growing nucleus of a metal on a
substrate of a different material such as carbon. (Adapted from reference 10)

Pletcher also explains the various stages in the deposition of a thick metal layer on an
inert cathode. These have been pictorially described in Figure 3.2, for the case of
progressive nucleation and three-dimensional growth. In the “progressive” nucleation
model, new nuclei are formed with time at a given overpotential. Growth from these
nuclei occurs and eventually the metal particles coalesce to form a continuous ﬁlm.”
As depicted in Figure 3.2, the deposition of a thick layer of metal is a multistep
process and each of the processes could be the rate-determining step. An
understanding of the metal nucleation and growth phenomenon on diamond is

important for determination of the optimum experimental parameters for

49

preconcentration in the ASV method, such as the overpotential and deposition time.
Optimizing these parameters is important for improving the selectivity and sensitivity

of the ASV method.

 

Clean, smooth surface
_I

Formation of ﬁrst stable nuclei

—.——.—.—— Growth of ﬁrst nucleus and formation of others

_._I_”_I. Further nucleation and growth
“4 Commencement of overlap
jun-g overlap complete

_-_ Thickening of complete layer

Figure 3.2 Deposition of a thick metal layer of metal on an inert cathode via a
progressive nucleation and growth mechanism. (Adapted from reference 10)

 

 

Budevski, Staikov and Lorenz in their book, “Electrochemical Phase Formation and
Growth,” give a very nice introduction to the processes governing the initial stages of
metal deposition.12 Important factors which determine the mechanism of metal phase
formation on a foreign substrate are the metal-substrate binding energy and the
crystallographic misﬁt between the substrate and the 3D metal bulk deposit. They

present two different growth models, which are illustrated in Figure 3.3. In the ﬁrst

case (Fig 3.3.A), the binding energy of the metal on a foreign substrate (\Irmds-s) is

lower than that of the metal on the native metal substrate (‘I’Mads-M). In this case the

overpotential deposition of 3D metal bulk phase takes place on an unmodiﬁed

50

substrate surface according to the “Volmer-Weber” or 3D island growth mode,

independent of crystallographic metal-substrate misﬁt.

In the second case, the binding energy of the metal on the foreign substrate
(‘I’Madsg is higher than that of the metal on the native metal substrate (‘I’MadsM). In

this case, 2D metal phases can form in the underpotential range (i.e., potentials more

positive of the E°') and the surface concentration of the adsorbed metal at EM/M2+ can
reach one or more monolayers depending upon ‘I’Mad,-s. They describe two different

sub-cases: (i) substances with negligibly small crystallographic misﬁt; 2D metal
(underpotential deposition) and 3D metal crystallites are epitaxially oriented
following a “layer-by—layer” growth mode (crystallographic misﬁt is characterized by
the interatomic distances doMe and dos of 3D metal and substrate bulk phases,
respectively, as shown in the Figure); and (ii) systems with signiﬁcant
crystallographic metal-substrate misﬁt; the formation and growth of unstrained 3D
metal crystallites (islands) on top of strained 2D metal overlayers is energetically

favored.

51

/\PMads - M do,M

 

Figure 3.3 Schematic representation of different growth modes of metal deposition on
foreign substrate (A) 3-D metal island formation (Volmer—Weber growth mode) ‘I’Mads —
S << LPM“, — M (B) Metal layer by layer formation, ‘I’Mad, — S >> ‘I’Mads — M and ratio
do)“ - do,s/ do,S zO, (C) 3-D metal island formation on top of predeposited 2D metal
overlayers on substrate, ‘I’Mad, — S >> ‘I’MadS — M and ratio do)“ - do,s/ dos > 0 (positive
misﬁt) or doM - dos / dos < 0 (negative misﬁt). (Adapted from reference 12)

Apparent from experimental data presented in the remainder of the Chapter, the

deposition of metal phases on boron-doped diamond electrodes occurs by the Volmer-

Weber growth mode. The binding energy of the metal on diamond (‘I’Mads-S) seems to be

lower than that of the metal on the native metal substrate (‘I’Mads-M). No underpotential

deposition of any metal has been seen for diamond surfaces, consistent with weak deposit
substrate interaction.l3 Similar observations were made by other researchers. For
instance, Pitter et a1. observed that the interaction of Ag with single-crystal boron-doped

diamond was rather weak.14 In fact, a procedure to entrap the metal particles into the

52

surface microstructure is often needed to stabilize electrodeposited metal catalyst

particles on diamond.‘5

3.3 Experimental

Chronoamperometric and Voltammetric Measurements. Most of the

electrochemical measurements were made using a CHI 650a computerized potentiostat
(CH Instruments Inc., Austin, TX) and a single compartment, three-electrode glass cell
placed inside a Faraday cage for shielding. The electrolyte solution was 0.1 M acetate
buffer, pH 4.5, unless otherwise speciﬁed. The other electrochemical parameters were as

described in Chapter 2.

Chemicals. All solutions were prepared with ultrapure water (>17 MQ-distilled,

deionized and passed over activated carbon) from an E-pure water puriﬁcation system
(Bamstead). Silver nitrate (Fisher), cadmium nitrate (Aldrich), zinc acetate (Aldrich),
cupric nitrate (Aldrich), and lead nitrate (Aldrich) were all reagent grade quality used
without additional puriﬁcation. 0.1 M acetate buffer was prepared by mixing appropriate
amounts of 99% sodium acetate (Aldrich) and acetic acid (Aldrich) solutions. All
glassware was cleaned by a three-step procedure: ethanol/ KOH bath, alconox/ ultrapure
water solution, and ultrapure water rinse. The CuSO4-5H2O and H2SO4 used in the
ECAFM measurements were ultrahigh purity grade (99.999%) (Aldrich). The
electrochemical ﬂuid cell was cleaned by a two-step sonication procedure that involved

soaking in dilute alconox and in ultrapure water.

53

Electrochemical Atomic Force Microscopy (E CAFM). Experiments were

conducted with a Nanoscope IIIa atomic force microscope (Digital Instruments Inc. and
Veeco Metrology Group, Santa Barbara, CA) equipped with a 16 um scanning head. The
measurements were made in the contact mode. A standard AFM ﬂuid cell equipped with
a gold-coated spring clip and a pyramidal Si3N4 tip formed on a gold-coated 100 um v-
shaped silicon nitride cantilever (0.38 N/m spring constant) were used. The tip was held
in an optically transparent holder in the ﬂuid cell. The ﬂuid cell was sealed with a special
o-ring, obtained from the manufacturer, designed to minimize leaks due to its special
concave shape. The exposed area of the electrode was 0.38 cmz. The cell had tapered
channels allowing for the injection of the solution and also placement of auxiliary and
reference electrodes. Pt wire was used as the counter electrode and a piece of Ag wire
was used as the quasi-reference electrode (QRE). All the images were obtained with 512
x 512 pixels. ECAFM images were obtained in either the constant force mode in which
the tip height was adjusted via a feedback mechanism so as to maintain the force constant
and or the constant height mode in which the tip height is constant so as to measure the
force. A potentiostat was used to control the electrode potential during potential sweep or
potential step measurements. The potential difference between the Ag QRE and an SCE
reference electrode was — 27 mV in 1 M KCl. For convenience, all potentials are reported
after being converted to the Ag/AgCl reference scale. A complete AFM image took

approximately ca. 60 s to record. All measurements were made at room temperature, ~ 25

°C.

54

3.4 Results and Discussion

Cyclic voltammetry and chronoamperometry are useful for studying the
nucleation and growth of metal phases. Potentiostatic transients at different potentials
give experimental “current-time” relationships which can form the basis of evaluation of

nucleation rate constants and particle number.

3.4.1 Metal Ion Voltammetry. Figure 3.4 shows a cyclic voltammetric i-E curve

for 1 mM CuSO4 in 50 mM H2SO4 at 100 mV/s. The curve for the last scan (total of ﬁve
scans) is shown. Curves were recorded from potential sweep rates of 50 to 500 mV/s. A
summary of the peak currents, peak potentials and the peak separations at these scan rates
is given in Table 3.1. The difference in current magnitude between the current maximum

and a baseline have been denoted by 1p , peak currents. A baseline was manually drawn

by extending the part of the curve before the current increase/decrease.

In Figure 3.4, the potential scan was begun at 400 mV and scanned in the negative
direction, in this case at 100 mV/s. The recorded voltammetric i-E curve shows a
cathodic current ﬂow due to the reduction of the metal ion to the metal, beginning at — 96
mV. The cathodic current gradually increases due to the formation and growth of metal
nuclei, as depicted in Figure 1.1, and ﬁnally reaches a maximum before decaying at more
negative potentials due to a reduced ﬂux of reactant species diffusing to the electrode
surface. After reversal at — 500 mV, cathodic current continues to ﬂow with a current

crossover at -110 mV. This crossover is indicative of an overpotential for nucleation and

55

grth on the bare diamond surface. In other words, it is energetically more favorable to
form metal phases on existing metal than on the bare diamond surface. In summary, the
voltammetric i-E curve exhibits the characteristic features of metal phase nucleation and
growth, i.e., the large negative shift of the reduction peak potential and the current

crossover during the reverse scan.

 

100 * Cu(II) + 26' m CU

A
s
v
4—r
5 o
1:
3 1
U

-50-+

 

 

 

 

I

-400 ' -200 i o I 200 ' 400
Potential (mV vs. Ag/AgCl)

Figure 3.4 Cyclic voltammetric i-E curves for the reduction of 1 mM Cu(II) and
subsequent oxidation at a microcrystalline diamond thin-ﬁlm electrode. Scan rate = 100
mV/s. The redox reaction for the Ctr/Cu(II) system is shown.

Table 3.1 Summary of the cyclic voltammetric peak currents, peak potentials and peak
potential separations for the Cu(II)/Cu redox system at various scan rates for a
microcrystalline diamond thin-ﬁlm electrode.

 

 

 

 

 

 

 

Onset reduction 1,,red 1pox “’d °" A
potential (mV) (pA) (pA) (mV) (mV) (mV)
50 mV/s -80 40 120 -193 178 413
100 mV/s - 96 44 126 -229 203 432
200 mV/s -110 47 135 -276 206 473
400 mV/s -1 19 56 150 -296 214 510
500 mV/s -121 58 158 -310 229 S39

 

 

 

 

 

 

 

 

56

The overall reaction for the Cu(II) system has been cited as;'6

Cu(II) + 2e‘ <:> Cu, E° = 53 mV vs. Ag/AgCl
The reduction peak current for each electrode increased linearly with the square root of
scan rate (Levich plot) over the range, 1'2 = 0.991, suggestive of a diffusion controlled
process. In contrast, the oxidation peak current increased rather linearly with scan rate
indicative of a surface conﬁned process (r 2 = 0.994). The onset reduction potential
shifted negative with increasing scan rate indicating that the electron-transfer rate for

metal ion reduction was less than the rate of the potential sweep.

Figure 3.5 shows the cyclic voltammetric i-E curves for 10 um solutions of Pb(II),
Zn(II), Cu(II) and Cd(II) in 0.1 M acetate buffer, pH 4.5, at 100 mV/s at a
microcrystalline diamond thin-ﬁlm electrode. The voltammograms exhibit features
characteristic of metal nucleation and growth. Listed near each i-E curve is the
experimental observed oxidation peak potential, Ep°", and the standard potential, E°', for
the reaction. The standard potentials listed have been corrected for concentration (10 um
solutions) and are reported with respect to the Ag/AgCl reference electrode. Little
overpotential is seen for the oxidation of the metal phases. This means that the oxidation
reaction rate for each of the metal species is relatively fast. On the other hand, the onset
potentials for metal phase formation are shifted more negative of the E“. The
overpotential for the reduction reaction was, in general, larger than the overpotential for
the oxidation reaction. In reality however, the oxidation potentials are observed to be
within 100 - 150 mV of the theoretical potential. Diamond consists of certain defect sites

and grain boundaries where the nucleation of the metal is favored. Further as shown in

57

Chapter 2, the surface of diamond consists of regions of different electrical conductivity
or electrochemical activity. The non uniform conductivity of the surface and the difﬁcult
nucleation can both contribute to the initial deposition overpotentials.l3 Microscopy
revealed the metal phases form in isolated islands of 3-D growth. Overall, the fact that
good activity for metal deposition and oxidation (stripping) reactions is observed with
modest overpotentials (ca. 100 mV) for diamond electrodes demonstrates its potential for

use in metal analysis.

Berzins and Delahay derived equations for the voltammograrrrs in the case of
reversible metal deposition on a solid electrode.17 The peak current was given as:
ipo" = 367 n3/2AD01/201/2C0* (3.1)
where, n is the number of electrons transferred in the reaction, A is the electrode surface

area, Do is the diffusion coefﬁcient of the ion, 0 is the scan rate and C0* is the bulk

solution concentration of the metal ion. The experimental correlations of the values of ipox
with the above equation tended to be relatively poor (> 20% error) for Ag(I) and Cd(II).
This could be due to a disparity in the value of D (literature values were used),
uncertainty about the area term in the expression and or surface phenomenon at the

electrode.

An example of a surface phenomenon could be the incomplete stripping of
deposited metal from the surface of the electrode. Vinokur et al. and also Hyde et al.
observed the incomplete stripping of Ag from the surface of boron-doped diamond

electrodesu’ '8 One possible reason cited by the Vinokur group was the model of

58

“undercutting”, initially proposed by Adzic and co-workers.19 “Undercutting” is the

phenomenon of faster dissolution of metal grains closer to the electrode surface, cutting

off electrical contact with the rest of the deposited Ag, which would therefore not

contribute to the oxidation current.

 

 

 

 

 

 

 

 

  

 

 

 

 

   

  
   

30.0 - .
2,0} Pb(II) Epox = - 429 mV '50: Cd(II)
200* 100}
150; . . . — Ep°*=-691 mV
' i Indicative ofNucleaho 5.0-
10.0: and Growth 1
5.0- \ 0.0-
0.0- 3" /' ,
A 5 0. —-* I. -501— E0' = _ 746 mV
<1: W E°=~469mV A , ,
1-100 1111111 111111111111... 400111111113411111~1111111«111111
v -900 -800 -700 4500 .500 400 -300 61800 -950 -900 -850 -800 -750 -700 -650 4100
4—3 . . _.
1
5 ioi-Zn(II) LCu(II) f
T 4.0
g i ) Ep°x = + 6 mV
0'»—
0 r 20” E°'=-4mv

 
  
  

Ep0" = - 986 mV
E°' - 1106 mV

1 1 1 1
-1350

11111111

-30 I 1 ' 1
-l450 -1250 -950

 

 

1 -20 -

1 4.9100

    

 

 

-500

Potential (mV vs. Ag/AgCl)

Figure 3.5 Cyclic voltammetric i-E curves for 10 um solutions of Pb(II), Cu(II), Cd(II)
and Zn(II) at a microcrystalline thin-ﬁlm diamond thin-ﬁhn electrode. The metals deposit
and oxidize off near their standard reduction potentials. Scan rate = 100 mV/s, supporting

electrolyte = 0.1 M acetate buffer, pH 4.5.

3.4.2 Chronocoulometric

Measurements.

Double potential-step

chronocoulometric measurements were conducted to determine if complete oxidation

(stripping) of deposited Ag occurred ﬁom the electrode surface. 50, 250 and 1000 ppb of

Ag(I) solutions in 0.1 M acetate buffer were studied. The potential at the working

59

electrode was held for 5 5 each, at 0 V to deposit the Ag from solution and then at 0.6 V
to oxidize the metal deposit. Summary of the results for both types of diamond have been
presented in Table 3.2. The ratio of background corrected charge for oxidation (onx) and

red) are presented. It is observed that in all cases OP" is less than Qpred with

reduction (Qp
an average ratio of 0.76 i 0.05. This indicates incomplete stripping of Ag from the
surface of the diamond. Vinokur et a1. and also Hyde et al. have observed the incomplete
stripping of Ag from the surface of boron-doped diamond electrodes.” 18 One possible
reason attributed for this observation by the Vinokur group was the model of
“undercutting” as discussed in Section 3.4.1. Another reason for incomplete stripping is
speculated to be the non-uniformity of the electrode surface, due to which metal nuclei
preferentially localize to active sites such as grain boundaries.20 During the positive
potential step, dissolution from these sites may be occurring at lower applied electrical
ﬁeld.

A clear trend of dependence of shipping efﬁciency on the concentration of the
metal ions in solution was not observed, although the shipping efﬁciency at the
nanocrystalline diamond was slightly higher at the lower concentration. Further, the
efﬁciency of oxidation of the metal from the electrode surface was slightly higher on the

microcrystalline diamond. This could have resulted from the greater microstructural

homogeneity (as discussed in Chapter 2) of the microcrystalline diamond.

60

Table 3.2 Chronocoulometric charge data for the deposition and shipping of Ag from a
Ag(I) solution in 0.1 M acetate buffer, pH 4.5. Potential steps from 0.6 to 0 to 0.6 V were
applied for 5 5 each to diamond electrodes of two different morphologies.

 

Ratio: Qp°x/ Qpred

 

Ag(I) Concentration Microcrystalline diamond Nanocrystalline diamond

 

 

 

50 ppb 0.81 :011 0.77 :010
250 ppb 0.76 i 0.09 0.70 i 0.08
1000 ppb 0.80 i 0.09 0.70 1 0.05

 

 

 

 

 

3.4.3 Metal Ion Chronoamperometric Measurements. As discussed above,

the early stages of electrochemical phase transformations often involve a 3-D nucleation
process on boron-doped diamond electrodes. The rate of formation and the number of
nuclei formed are dependent on the overpotential applied. Due to the fast charge-transfer
step, the rates of growth of mature nuclei can be subsequently described in terms of
control by mass-transfer of reactant metal to the growth centers. Three-dimensional
nucleation/growth processes have been described in different theoretical models?" 22 The
nucleation kinetic parameters, A (nucleation rate constant per site) and No (number
density of active sites over the substrate surface), can be estimated. The growth is best
described in terms of localized diffusion to small spheres of (assumed to be so) nuclei. At
early times in the nucleation event, hemispherical diffusion zones develop around and

extend outward from each nuclei. The thickness of the diffusion zone is given by:

23,-(1131)“? (3.1)

61

Where 5t (cm) is the diffusion zone thickness; k is a constant; and D is the diffusion

coefﬁcient (cmz/s) of the reacting species. As the radius of the growth centers increase,
the hemispherical diffusion zones for the nuclei begin to overlap. Therefore, the ﬂux of
reactant decreases and eventually the primary direction of reactant ﬂux is normal to the
surface. The corresponding diffusion path can then be expressed in terms of semi-inﬁnite
linear diffusion toward a fraction of the electrode surface. These relationships have been
pictorially depicted in Figure 3.6. Such three-dimensional nucleation and growth has
been described by different theoretical models as proposed by Hills et al., and Schariﬂcer

and Mostany.

Hills et al., classify nucleation as either “instantaneous” in which rapid nucleation
takes place on a ﬁx number of active sites, or “progressive” in which the number of
nuclei increases with time and the number of active sites available for nucleation are
considered to be virtually inﬁnite. It should be noted that these are limiting cases. In the

present work, these two limiting models were used to analyze the experimental results.

62

 

long t

short t

 

 

 

 

 

 

Current Density A\ CIIl'E—r

Time —>

 

Figure 3.6 (A) Schematic diagram of the time-dependent growth of diffusion zones
around forming metal nuclei on an electrode surface and their eventual overlap. The
arrows indicate the directions of the diffusion ﬁelds during the owth of nuclei. (B)
Expected current transient; K is constant given by 11F ADol/ZCoi/n” , the terms have their
usual meanings.

Single-step chronoamperometric measurements were performed to investigate the
metal nucleation growth process on diamond electrode surfaces.23 Figure 3.7 presents a
set of current-time transients recorded during the formation of Cu deposits on a
microcrystalline diamond thin-ﬁlm electrode from a solution of 1 mM CuSO4 in 50 mM
H2804. The measurements were performed by stepping the potential from 500 mV,
where Cu(II) is stable, to different ﬁnal potentials ( -400, -500, -600 and -700 mV) past
the E°'. These potentials were sufﬁciently cathodic to induce nucleation and subsequent
crystallite growth. The working electrode was kept at a positive potential of 0.6 V for 3
min after every potential step to completely oxidize the copper deposits. Transients

obtained by stepping the potential to — 400 and —700 mV have been shown in the Figure,

63

along with the theoretical curves (the cases for instantaneous and progressive nucleation).
The plots are presented in dimensionless units, (i/iM)2 vs. t/tM, facilitating comparison

with the behavior predicted for each of these limiting nucleation mechanisms.

The current initially increases as a result of the formation of stable nuclei and
growth of the metal deposits. The rising current eventually terminates in a broad
maximum (iM) as the individual hemispherical diffusion zones of the growing crystallites
coalesce and give way to linear diffusion to a planar surface. After iM has been reached,
the current decays at a t'“2 rate predicted from the Cottrell equation, due to the formation
of a depletion layer in the solution phase adjacent to the electrode that grows in thickness

as a function of time.

As mentioned, the theoretical curves were constructed by plotting (i/im)2 vs. t/tm.
At ﬁrst the nucleation process was classiﬁed as “instantaneous” or “progressive”. The
expressions presented in non-dimensional form by plotting (i/iM)2 vs. t/tM, for

11151311141190“ (3-2) and Progressive (3.3) nucleation are 8’ 9

(i/iM)2=1.9542(ti(4/t){l-exp[-l.2564(t/tM)]}2 (3.2)

(i/iM)2=1.2254(tM/t){1-exp[-2.3367(t/t),4)2]}2 (3.3)

The experimental data did not exactly ﬁt the theoretical current transients, however, the
nucleation of Cu on microcrystalline diamond appears to more closely track, the transient
for progressive nucleation. The reason for the discrepancy in the experimental and
theoretical curves maybe the fact that the diagnostic criteria for distinguishing between

extreme cases of progressive and instantaneous nucleation they do not contain parameters

64

related to physical or chemical properties of the system, such as molar volume,

concentratron, drffusron coefﬁcrent or kinetic constants.23 2

 

 

32 11 = 455 mV 3.2 n = 755 mV

 

 

 

 

 

 

 

 

Figure 3.7 Comparison of experimental current-time transients with the theoretical
transients corresponding to instantaneous and progressive nucleation at different
overpotentials.

Once established that the nucleation is progressive at high supersaturation, the
product ANo that deﬁnes the steady-state nucleation rate over the entire surface, was
calculated from the current maxima, by means of the following equations,

1... = ( 4.6733/ ANonkD)‘/2 and im = 0.461 5.7.I=D3"‘c(A1~I,1<)"4 (3.4)
where D is the diffusion coefﬁcient, zF the molar charge of the electrodepositing species
and k is a numerical constant given by,

k = (811cM/p)“2 (3.5)
where M and p are the molecular weight and the density of deposited material,

respectively. The values obtained appear in Table 3.3.

65

Table 3.3 Current-time transient data for copper electrodeposition on microcrystalline

 

 

 

 

 

diamond.
E38122; im (mA.cm'2) 1,, (s) No(cm'2) A (8")
455 1.17 0.099 1.45x107 23.6
555 1.61 0.06 2.38x107 38.6
655 1.91 0.04 3.57x107 58.2
755 2.33 0.02 7.15x107 117

 

 

 

 

 

 

 

D = 2.5 x 10' cmz/s. E07 = 55 mV.

In all cases, the reaction rate, reﬂected by the current maximum, im, both of which
increased with applied overpotential, whereas tm decreased. In other words, the time
required for the coalescence of the hemispherical diffusion zone for each mature nuclei,
decreased with overpotential. This implies an increase in the number density of growth
centers with overpotentials. The observation is consistent with the fact that the number
density of sites, N0 and the rate constant, A, increases with overpotential. The
electrodeposition of copper on boron-doped microcrystalline diamond ﬁlms at these

conditions is potential dependent and progressive in nature on microcrystalline diamond.

3.4.4 EC-AF M Study of Copper Deposition. Electrochemical-atomic force

microscopy (EC-AFM) was used to monitor copper electrodeposition on boron-doped,
polycrystalline diamond thin-ﬁlm electrodes with respect to time and overpotential. EC-
AFM studies of copper stripping/deposition reactions on boron-doped diamond thin-ﬁlm
electrodes under potentiostatic control in solutions of different ability complexing ability

such as sulfates and phosphates, have been previously reported.25 The method proved

66

useful for revealing the nucleation state and the early stages of complex formation when

cations were released from the deposited metal.

Figures 3.8 and 3.9 show height mode images (5 X 5 pm) collected during c0pper
deposition on microcrystalline diamond from a 50 ppm CuSO4 in 0.1 M acetate buffer,
The images were collected at the various overpotentials (vs. Ag/AgCl) and times as
indicated. Images in Figure 3.8 were collected by stepping the potential from + 600 mV
to the particular cathodic value indicated for ~ 60 s and then back to + 600 mV. The
potential was then held at + 600 mV for at least 3 min. Multiple images collected at + 600
mV veriﬁed that the metal deposits were completely oxidized from the electrode surface.
From Figure 3.8, new nuclei are observed to form with increasing overpotentials. Metal
particles fornr at both the crystal facets and crystalline grains boundaries and defects. The
greatest densities of particles seems to be present at the grain boundaries at lower
overpotentials. The particles formed on the crystalline facets increase in number at the
higher overpotentials. Notice the particle formed on the grain facet indicated by the arrow
at the overpotential of — 866 mV. The grth on the facet surfaces initiates at defect sites.
Another important factor with diamond in terms of location of the metal deposits is the

local ﬁlm conductivity.

In related work, Enea et al. electrochemically deposited Pt particles onto boron-
doped (1 ppm of trimethyl boron in the source gas) diamond thin-ﬁlms. Their study of the
morphology and dispersion of Pt deposits by AFM (air, tapping mode) proved the
existence of a small number of preferential diamond sites on which the Pt particles

initially formed.26 They reasoned that the behavior resulted from the higher electrical

67

conductivity in some regions relative to others, presumably due to differences in boron-
doping levels. From a manual count of the metal particles in multiple images, the particle
density is estimated to be in the range of 6.0X107 to 1.0X108 cm'z. The number of
particles formed increased with the overpotential. These observations are also consistent
with a progressive nucleation and growth mechanism. The particles broadly range in size

from ~ 100 to 250 nm.

Bare electrode

   
 

211m

Figure 3.8 ECAFM images (height mode, 5 pm X 5 pm) collected during Cu deposition
on microcrystalline diamond from a 50 ppm CuSO4/0.l M acetate buffer solution at
different cathodic overpotentials. Arrows indicate the formation of new nuclei.

68

Figure 3.9 shows the ECAFM images (height mode, 5 11m X 5 pm) collected
during copper deposition on microcrystalline diamond at one overpotential for various
times. From a manual count, the particles density is in the range of 8.0X107 to 1.0X108
cm'z. The particles broadly range in size from ~ 100 to 320 nm. The nucleation looks to
be largely progressive as new nuclei are formed with time. The time dependent formation

of nuclei is consistent with a progressive nucleation and grth mechanism.

Bare electrode

      

- 466 mV, 608

 

2 11m - 466 mV, 180s - 466 mV, 120s
Figure 3.9 EC-AFM images (height mode, 5 pm X 5 pm) collected during Cu deposition

on microcrystalline diamond from a 50 ppm CuSO4 /0.1 M acetate buffer solution at
various times at a ﬁxed overpotential. Arrows indicate the formation of new nuclei.

69

3.5 Conclusions

The Chapter describes the theoretical aspects of the metal nucleation and grth
on solid electrodes. Results from cyclic voltammetric, chronoarnperomehic and EC-AFM
measurements demonstrated how these methods are useful for both the quantitative and
qualitative study of the mechanisms governing metal phase formation on diamond. The
experimental data reveal that the deposition of metal phases on boron-doped diamond
electrodes occurs by the Volmer-Weber growth mode. No underpotential deposition of
metal was observed for diamond, indicating a weak metal phase-substrate interaction.
The analysis of current transients, under the framework proposed by Schariﬂ<er and
Montany, was particularly useful since it allowed the determination of the values for the
nucleation rate, A, and number density of nuclei, No, for Cu deposition on
microcrystalline diamond. Results from cyclic voltammehic, chronoamperometric and
EC-AFM measurements all are consistent with a progressive nucleation and growth

mechanism.

Understanding metal phase formation on diamond is important for application of
this new electrode in ASV. For instance, one ideally desires a high dishibution of metal
particles on diamond with a small and uniform particle size for ASV. This should lead to
more narrow and intense oxidation and shipping peaks. Therefore, an instantaneous
nucleation and growth mechanism would be ideal. As will be discussed in Chapters 5 and
6, the stripping peaks for Cu and the other metals are rather broad. This is presumably,
due, in part to the fact that, at least for Cu, metal phase formation on diamond occurs via

a progressive nucleation and growth mechanism.

70

3.6 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Henry, C. R. Crystal Research and Technology 1998, 33, 1119-1140.

Gunawardena, G.; Hills, G.; Montenegro, 1.; Schariﬂcer, B. Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry 1982, 138, 225-239.

Gunawardena, G.; Hills, G.; Montenegro, 1. Journal of Electroanalytical
Chemistry and Interfacial Electrochemistry 1982, 138, 241-254.

Gunawardena, G.; Hills, G.; Montenegro, 1.; Scharifker, B. Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry 1982, 138, 255-271.

Wang, J .; Swain, G. M.; Tachibana, T.; Kobashi, K. Journal of New Materials for
Electrochemical Systems 2000, 3, 75-82.

Swain, G. M.; Wang, J. In US. Pat. Appl. Publ.; (Board of Trustees of Michigan
State University, USA). Us, 2004, pp 13 pp.

Awada, M.; Strojek, J. W.; Swain, G. M. Journal of the Electrochemical Society
1995, I42, L42-L45.

Hyde, M. E.; Compton, R. G. Journal of Electroanalytical Chemistry 2003, 549,
1-12.

Budevski, E.; Staikov, G.; Lorenz, W. J. Electrochimica Acta 2000, 45, 2559-
2574.

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

Abyaneh, M. Y. Journal of the Electrochemical Society 2004, 151, C737-C742.

Budevski, E.; Staikov, G.; Lorenz, W. J .; Editors, Eds. Electrochemical Phase
Formation and Growth: An Introduction to the Initial Stages of Metal Deposition,
1996

71

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

Vinokur, N.; Miller, B.; Avyigal, Y.; Kalish, R. Journal of the Electrochemical
Society 1999, 146, 125-130.

Pitter, M.; Hugenschmidt, M. B.; Behm, R. J. Applied Physics Letters 1996, 69,
4035-4037.

Wang, J .; Swain, G. M. Journal of the Electrochemical Society 2003, 150, E24-
E32.

A.J.Bard; L.R.Faulkner, Eds. Electrochemical Methods, Fundamentals and
Applications,; John Wiley and sons, Inc: New York, 2001

Berzins, T.; Delahay, P. Journal of the American Chemical Society 1953, 75, 555-
559.

Hyde, M. E.; Jacobs, R. M. J .; Compton, R. G. Journal of Electroanalytical
Chemistry 2004, 562, 61-72.

Adzic, G. D.; Drazic, D. M.; Despic, A. R. Journal of Electroanalytical
Chemistry and Interfacial Electrochemistry 1988, 239, 107-111.

Holt, K. B.; Bard, A. J .; Show, Y.; Swain, G. M. Journal of Physical Chemistry B
2004, 108,15117-15127.

Pitter, M.; Hugenschmidt, M. B.; Behm, R. J. Applied Physics Letters 1996, 68,
2508-2510.

Schariﬂter, B.; Hill, G. Electrochimica Acta 1983, 28, 879-889.

Wang, J. Dissertation, Michigan State University Fabrication, characterization,
and electrocatalyticactivity of metal/diamond composite electrodes,2002.

Palomar-Pardave, M.; Ramirez, M. T.; Gonzalez, 1.; Serruya, A.; Schariﬂcer, B. R.
Journal of the Electrochemical Society 1996, 143, 1551-1558.

Zak, J .; KOlOdziej-Sadlok, M. Electrochimica Acta 2000, 45, 2803-2813.

72

(26) Enea, O.; Riedo, B.; Dietler, G. Nano Letters 2002, 2, 241-244.

73

CHAPTER 4

ANODIC STRIPPING VOLTAMMETRIC DETECTION OF TRACE
METAL IONS: A PERFORMANCE COMPARISON OF BORON-
DOPED DIAMOND THIN-FILM ELECTRODES AND SOME
COMMONLY USED ELECTRODE MATERIALS

4.1 Introduction

The use of two different types of boron-doped diamond thin-ﬁlm electrodes for
the anodic stripping voltammetric detection of trace metal ions is described in this
Chapter. Both microcrystalline and nanocrystalline morphologies were studied.
Comparison of the two diamond electrode types revealed that both perform well for the
detection of the metal ions and exhibit superb detection ﬁgures of merit. In general, the
limit of quantitation was lower and the sensitivity was greater for nanocrystalline
diamond. The nanocrystalline diamond morphology appears to aid in the formation of
smaller metal particle sizes and higher particle densities during the deposition step. This
leads to the improved ﬁgures of merit. Multielement analysis was also performed and the
detection ﬁgures of merit are reported. There are some limitations with the use of
diamond, speciﬁcally, interferences due to alloy and intermetallic compound formation
between co-deposited metals. The response of diamond with bare and Hg-coated glassy
carbon (GC) electrodes was compared. Scanning electron micrographs of the metal
deposits, useful for the speculation about the nucleation and growth mechanism of each

metal is presented.

74

4.2 Experimental

Diamond Film Deposition and Characterization of Metal Deposits. Boron-

doped diamond thin ﬁlms were deposited on p-type Si(100) substrates ( ~10’3 Q-cm),
using a commercial microwave-assisted chemical vapor deposition (CVD) system.
Details regarding the growth conditions, as well as the follow-up material
characterization, have been provided elsewhere and in Chapter 2." 2 B2H6 was used for
doping at a ﬂow rate of 0.001 sccm (10 ppm). Scanning electron microscopy (JEOL-

JSM-6400V) was used for imaging the metal deposits on the diamond surface.

Electrochemistry. Most of the electrochemical measurements were made using a CHI

650a or 900 - computerized potentiostat (CH Instruments Inc., Austin, TX) and a single
compartment, three-electrode glass cell placed inside a Faraday cage for shielding. The
diamond working electrode was pressed against a Viton® o-ring and clamped to the
bottom of the cell. Either a bead of In/Ga alloy was placed on the backside of the
scratched and cleaned substrate for ohmic contact or it was rubbed with a carbon rod. A
graphite rod was used as the counter electrode and a Ag/AgCl electrode (3.5 M KCl)
served as the reference (EOAg/Agm = -0.045 V vs. SCE). The geometric area of the working
electrode was ca. 0.163 cmz. All measurements were made in solutions deoxygenated
with nitrogen (99.99%) for at least 5 min. initially and then for 2 min after each anodic
stripping or oxidation cycle. The solutions remained blanketed with the inert gas during a
measurement. The electrolyte solution was 0.1 M acetate buffer, pH 4.5, unless otherwise
stated. The reference electrode was placed inside a second cracked glass capillary (ﬁlled

with the acetate buffer), which served as a double junction. All measurements were made

75

in an unstirred solution. Typical differential pulse voltammetric settings were a 3 min
deposition time, a 0.5 min “quiet” time at the deposition potential prior to initiation of the
anodic sweep, a 100 mV pulse height, a 2 mV step height, a 100 msec pulse width (unless
otherwise speciﬁed), a 100 msec cycle period, a 35 msec or less sampling time, and a
constant potential of 600 mV for 120 s after completion of the anodic sweep to ﬁrlly

oxidize all metal deposits.

Prior to the deposition of the metal adlayer, the glassy carbon (Tokai, GC-30) was
polished under very clean conditions using a series of alumina slurries (1.0, 0.3, and 0.05
pm). The electrode was rinsed and ultrasonicated with ultrapure water between each
polishing step to remove polishing debris and clean the surface. For the comparison of
diamond and Hg—coated GC, the concentration of Hg(CH3CO2)2 in the Hg-containing
solutions was 2><10‘5 M. All the experiments comparing diamond and Hg-coated GC were
conducted in acetate buffer of pH 5.2 with 30 ppm of Ca(II) and 5 ppm Mg(II) added.
The geometric area of the working electrode was 0.222 cmz. The other deposition and

stripping conditions were same as above.

Chemicals. All solutions were prepared with ultrapure water (>17 MQ-distilled,
deionized and passed over activated carbon) from an E-pure water puriﬁcation system
(Bamstead). Silver nitrate (Fisher), cadmium nitrate (Aldrich), zinc acetate (Aldrich),
cupric nitrate (Aldrich), lead nitrate (Aldrich) and mercuric acetate (Aldrich) were all
reagent grade quality used without additional puriﬁcation. 0.1 M acetate buffer was

prepared by mixing appropriate amounts of 99% sodium acetate (Aldrich) and acetic acid

76

(Aldrich). All glassware was cleaned by a three-step procedure: ethanol/ KOH bath,
alconox/ ultrapure water solution, and ultrapure water rinse. The glassware for the
comparison of diamond and Hg-coated GC were cleaned by a tlrree-step procedure
consisting of: alconox/ ultrapure water solution, soaking for 10 min each in l M

hydrochloric acid, and l M nitric acid, followed by ultrapure water rinse.

4.3 Results and Discussion

4.3.1 Comparison of Diamond and Glassy Carbon for the LSV

Detection of Metal Ions. Comparison studies of the performance of microcrystalline

diamond and GC were initially performed. A complete set of detection analytical ﬁgures
of merit were obtained with both electrodes for three metal ions using linear sweep
voltammetry (LSV): Ag(I), Cd(II) and Hg(II). Figure 4.1 shows the linear sweep
voltammograms for the oxidation of Ag at microcrystalline diamond. The metal phase
was formed at 0 V with a deposition time of 3 min. The peak current and charge
increased with the solution concentration of Ag(I). The current rises sharply on the
forward edge of the peak as the metal phase oxidizes. The peak then goes through a
maximum followed by a slow decay on the trailing edge. This reﬂects diffusion of the
oxidized metal ion away from the surface. As the concentration was increased from 0.001
to 1 mM, it was observed that Ep°x for Ag shifted from 360 to 450 mV, for Cd from —700
to —610 mV and for Hg from 460 to 500 mV. The peak potential shifts positively because
the surface coverage increases with increasing solution concentration and it takes longer

to oxidize the metal deposits.3 The limit of quantitation was found to be 107, 110 and 200

77

ppb for Ag(I),Cd(II) and Hg(II) respectively (S/N>3). Figure 4.1 also shows the log—log
plots of ip°x vs. the Ag(I) concentration at diamond. The plots were fairly linear, (BDD, r2
= 0.985, 993, 0.939 and GC, r2 = 0.998,0.9902, 0.977) for Ag(I), Cd(II) and Hg(II)
respectively. It was found that plotting the peak charge (me‘) vs. the concentration
generally gave better correlation coefﬁcients and was used more frequently from this

point on.

 

 

 

 

 

 

é GC

8

1: BDD
3

U

 

 

 

 

 

 

 

--7-6-5-4-3-2-101
Potential (mV vs. Ag/AgCl) Log 01119311113110“ (111M)

Figure 4.1 Linear sweep stripping voltammetric i-E curves for Ag(I) concentrations of 10
pM, 100 pM and 1 mM. Inset shows the i-E curves for the l and 10 11M Ag(I)
concentrations. The solutions were prepared in 0.1 M acetate buffer, pH 4.5. The log-log
current-concentration plots for the assays on BDD and GC are also presented.

Table 4.1 summarizes the analytical detection ﬁgures of merit for Ag(I), Cd(II)
and Hg(II) for both electrodes. Microcrystalline diamond outperformed GC in terms of

linear dynamic range, limit of quantitation, response precision and stability. The response

precision was measured over 10 consecutive measurements of a 10 11M metal ion

78

concentration. The response stability was measured over a three-day period with a total of

6 measurements of the solution of same concentration.

A reason for the superior performance of diamond could be the fact that these
ﬁlms exhibit voltammetric background currents and double-layer capacitances up to an
order of magnitude lower than for glassy carbon.‘1 The diamond surfaces are non-polar
and relatively oxygen-free, with the surface carbon atoms terminated by hydrogen.
Conversely, the GC surfaces are polar, with the exposed defect sites terminated by

carbon-oxygen functionalities.5

The surface carbon-oxygen functionalities promote
strong dipole-dipole and ion-dipole interactions with polar and ionic molecules and the
speciﬁc adsorption of electrolyte ions could occur. Another contributing factor may be a
lower density of surface electronic states near the fermi level caused by the semimetal-
semiconductor nature of boron-doped diamond. A lower surface charge carrier density at
a given potential would lead to a reduced accumulation of counter-balancing ions and
water dipoles on the solution side of the interface, thereby lowering the background
current and capacitance.6 Recently it has been proved that the diamond surface has
"electrochemically active" sites separated by more insulating regions.7 The reduced

electrochemically active area of diamond would further account for low background

currents observed.

79

Table 4.1 LSV analytical detection ﬁgures of merit for the detection of several metal ions

at boron-doped microcrystalline diamond and glassy carbon electrodes.

 

 

 

 

 

 

 

 

Microcrystalline Diamond Glassy Carbon
Parameter Ag(I) Cd(II) Hg(II) Ag(I) Cd(II) Hg(II)
Lma’ Dynam‘c 0.001-1 0.001—1 0.001-1 0.01-1 0.001-1 0.01-1
Range, (mM)
Limit of Quantitation
(S/N>3), (ppb) 107 1 12 200 1000 1 12 2000
Sensitivity, (pA/mM) 53 516 39 40 339 266
PreCiSion’ RSD’ (%) 1 8 0 81 3 3 6 7 4 5 12
(n = 10) . . . . .
Stability, (%)
(3 day test) 3.6 8.1 7.9 3.9 8.0 8.4

 

 

 

 

 

 

 

 

4.3.2 Comparison of Diamond and Hg—Film Electrodes for the DPASV
Detection of Metal Ions

Awada et al. deposited Hg on polycrystalline diamond thin-ﬁlms.8 The Hg
was deposited potentiostatically at -1.0 V vs. Ag/AgCl in a thin-layer ﬂow cell during
which time 5 injections (20 11L) of a 1 mM Hg(CH3CO2)2 in 0.1 M sodium acetate buffer
solution at pH 4.5 were made. The injections were carried out at a ﬂow rate of 1.4
mUmin. A scanning electron micrograph of the Hg-coated diamond thin ﬁlm (from their
work) is presented in Figure 4.2. They observed that the Hg-ﬁlm consists of individual
droplets and tended to form spheres ranging from 0.5 to 3 pm on the surface of the
electrode. Hg deposits on this diamond were found to be highly dispersed and existed

almost exclusively at the intercrystalline grain boundaries between the larger diamond

microcrystallites.

8O

 

7.5 pm

Figure 4.2 Electrochemically deposited Hg on the surface of a microcrystalline diamond
thin-ﬁlm electrode. The Hg tended to form spherical droplets on the substrate. (Adapted
from reference 8)

The deposition and stripping process of metals from Hg electrodes is different
from the metal nucleation and growth on a solid electrode. Foreign metals deposit within
the Hg volume forming amalgams. On the other hand, foreign metals deposit as a
metallic phase on solid electrodes. Figure 4.3 presents differential pulse anodic stripping
voltammetric i-E curves for Cd at (A) nanocrystalline diamond and (B) Hg-deposited
coated GC. The oxidation peak curents vary linearly from 10 to 1000 ppb for both
electrodes. (r2 = 0.997 and 0.994 for nanocrystalline diamond and Hg coated GC,
respectively). The background noise is higher for the Hg—coated GC than for the
nanocrystalline diamond. Typical values were 5.5 and 0.53 nA at E = - 750 mV,
respectively. The increased noise is presumably due, in part, to the greater capacitance

and more reactive surface of GC (electroinactive surface carbon-oxygen functionalities).

81

Further, the sensitivity is higher and the limit of quantitation (LOQ) slightly lower
for the detection of Cd(II) at Hg-coated GC compared to the planar nanocrystalline
diamond electrode. This can be explained due to the fact that the deposited metal is
amalgamated within the individual drops of Hg. In considering the shipping of the metal
from stationary Hg drop and ﬁlm electrodes, the metal concentration in the Hg becomes
important. A spherical correction term has to be applied to the Randles-Sevcik equation
(used to describe the forward peak current for a reversible system using linear sweep
voltammetry).9 There is also a divergence in the diffusion ﬂux in metal stripping from a
drop electrode due to its spherical nature. The peak current is not proportional to the drop
surface area as in the original Randles-Sevcik equation, but is proportional to the drOp
surface radius.10 In general, the peak current increases with decreasing values of the drop

radius and is proportional to the volume of Hg in the drop.

Peak potentials for the oxidation of Cd from the nanocrystalline ﬁlm are observed
to be more negative than those from Hg-coated GC. For instance at 100 ppb Cd(II), Epox
is observed to be -789 mV and -732 mV for nanocrystalline diamond and Hg-coated GC,
respectively. A reason for this could be the fact that in the case of the amalgamated metal,
the metal has to undergo diffusion to the surface of the Hg drop in order for it to be
oxidized. This takes more time than in the case of the solid electrode where the deposited
metal is readily available on the electrode surface. Also the recorded oxidation charge for
metal oxidation from the diamond electrode is almost three times lower than the

oxidation from Hg-coated GC, indicating a greater amount of total metal deposited inside

82

the Hg drOplet. The activity of the metal inside Hg is different from a pure metal phase.
Therefore, the peak current for the oxidation of metal from the Hg-coated electrode
appears at more positive potentials. The FWHM values of the peaks at the various
concentrations for the Hg-coated GC were slightly lower (< 6 mV in most cases) than
widths observed for nanocrystalline diamond. For instance at the 1000 ppb concentration,
the FWHM for Hg-coated GC was 52 mV and for nanocrystalline diamond was 58 mV.
However, the peaks were sharper at the Hg-coated GC surface, probably due to the more
homogeneous nature of the surface (i.e., the Hg environment from which the metal was

oxidizing).

In summary, the response of nanocrystalline diamond for Cd(II) was determined
to be comparable to that of Hg-coated GC. Both electrode types yielded an analytically
useful response for the metal ion. The sensitivity was higher for the Hg-coated GC and
the limit of quantitation was slightly lower. On the other hand, the background current

and noise were greater for Hg-coated GC.

83

-—l

5...

.5
o
r

 

 

Current (11A)
0'

 

 

O
4

 

 

 

 

 

 

 

-850 800 i -750 ‘ 430° '750 '700

Potential (mV vs. Ag/AgCl)

Figure 4.3 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for
standard solutions of Cd(II) ranging in concentration of 10 to 1000 ppb. The working
electrodes were (A) nanocrystalline diamond and (B) Hg-coated GC. The supporting
electrolyte was 0.1 M acetate buffer, pH 5.2. Deposition potential = -900 mV vs.
Ag/AgCl. Deposition time = 3.5 min. The other DPASV conditions are described in
experimental section. Geometric area = 0.222 cmz. The peak areas were determined by
integration between —700 mV and — 900 mV. (Data obtained by EA. McGaw)

Table 4.2 Analytical detection ﬁgures of merit for the differential pulse anodic shipping
voltammetric (DPASV) detection of the Cd(II) at a nanocrystalline diamond and Hg-
coated GC electrodes. (Data obtained by EA. McGaw)

 

 

 

 

 

 

 

Cd(II)
Parameter Nanocrystalline Hg-Coat ed GC
Diamond
Background Signal (pA) 0.53 5.5
Limit of Quantitation (ppb) 5 0 1 0
(S/N > 3) ' '
Sensitivity (nA/ppb) 15 51
Linear Correlation
Coefﬁcient, r2 0'997 09%

 

 

 

84

 

4.3.3 Linear Sweep and Differential Pulse Voltammetry of the

Individual Metals. Figure 4.4 shows linear sweep stripping voltammetric i-E curves

for 0.1 mM Cd in 0.1 M acetate buffer, pH 4.5 at a microcrystalline diamond electrode.
The deposition time was varied from 30 to 180 s in 30 s increments. A linear dependence
of the peak current with the time of deposition is observed (r2 = 1.00). A similar trend was
observed for both Ag and Hg (r2 = 0.997 and 0.972, respectively). These results prove
that the peak current increases in a proportional manner with the deposition time. A
similar relationship was observed for the peak charge, Qp°", and the time of deposition for
all the three metal ions. The deposition time conh'ols the amount of metal phase formed
and therefore the oxidation response sensitivity. A lowering of the limit of quantitation

can be achieved with longer deposition times.

In case of metallic ﬁlms deposited on an inert solid electrode, the number of

moles of metal, M, accumulated on the surface is given by:H

't
M: [Fri (4.1)

where, iL is the limiting current for the deposition of the metal, F the Faraday, td the
deposition time and n is the number of electrons transferred in the reaction. From the
above relation the total amount of metal deposited on the solid electrode is a function of

the total time of deposition, which in turn is related to the experimental peak current

(equation 1.1).

85

 

 

35.0F‘vviltf‘lvvvlvvilrti 400f T ‘ I l T I 1

    
  

30-01' Deposition Time

25.01-

 

Current( 11A)

 

 

5’11?"
r2 = 1.000

 

 

 

_5.orarInrAL1417lrnrloorlL00 00w 1 l 1 I 1 I l
4000 -900 -800 - 00 -6 -5 - 0 50 100 150 200

Potential (mV VS. Ag/AgCI) Tlmc(s)

 

Figure 4.4 Linear sweep stripping voltarrrmetric i-E curves for 0.1 mM Cd in 0.1 M
acetate buffer, pH 4.5, at a microcrystalline diamond electrode, depicting the dependence
of ip°" on deposition time. Deposition time was varied from 30 to 180 s (30 5 increments).
Deposition potential -1000 mV.

Figure 4.5 shows representative shipping voltarnmehic i-E curves for a 10 1.1M
concentration of each metal ion at both microcrystalline and nanocrystalline diamond in
0.1 M acetate buffer, pH 4.5. All deposition and shipping conditions were maintained
constant for the comparison. The curves reveal the complexity of metal deposition and
stripping at diamond, as the stripping peak potential, current, charge and FWHM all very
much depend on both the metal and the diamond ﬁlm type. Metal deposition and
stripping at diamond is more complex than the same processes at Hg-coated GC. For all
the metals except Zn, the shipping peak width, FWHM, was slightly smaller for
nanocrystalline diamond. This is consistent with a smaller nominal deposit size and

narrower particle size dishibution, as will be shown in scanning electron micrographs

presented in Section 4.3.5. The ip°x and Qp°" values tended to be greater for

86

nanocrystalline than for microcrystalline diamond. The same trends were observed at
other metal ion concentrations of 100 and 1 pM. The greater peak current and charge for
nanocrystalline diamond is consistent with more metal phase formation

(preconcentration). There is also some variance in the Epm‘ values.

A summary of the analytical detection ﬁgures of merit for the ﬁve individual
metal ions at both diamond types is presented in Table 4.3. Both electrodes exhibit a
linear dynamic range of 4 to 5 orders of magnitude with a limit of quantitation in the low
ppb range for all ﬁve metal ions. The limits of quantitation are generally lower for
nanocrystalline than for microcrystalline diamond. Due to the morphological inﬂuences
of the smaller grain size and higher fraction of grain boundary, a higher nucleation
density and higher particle density is achieved on this surface. We are presently
investigating the relative importance of the grain versus the grain boundaries in the metal
nucleation and growth process. Except for Cu(H), the sensitivity for all the metal ions is
slightly higher for nanocrystalline diamond. The coefﬁcients of variation for all the metal
ions are similar for both electrode types (1-4 %). The response precision was evaluated
from 10 successive measurements of a 10 11M solution of each metal ion. The response
stability was measured over a three-day period with a total of 6 runs of the same solution
concentration. The sensitivity values were calculated from the slope of the charge vs.

concentration plots and a minimum of ﬁve points was used in the statistical analysis.

87

 

 

 

 

Ag(I) 180CU(II) 90Pb(II)

 

 

 

 

 

 

 

 

 

 

 

Current (11A)
8

 

 

 

 

 

 

 

. 1 1 0 1 1 . r 1
~91) 31) -7(1) 411) 41(1) -1(II) 9(1)

Potential (mV vs. Ag/AgCl) - f 111111111 Nanocrystalline

—- Microcrystalline

Figure 4.5 Representative differential pulse anodic stripping voltammetric (DPASV) i-E
curves for a 10 pM concentration of each metal ion at both microcrystalline and

nanocrystalline diamond thin-ﬁlm electrodes. Supporting electrolyte = 0.1 M acetate
buffer, pH 4.5.

The limit of quantitation is highest for Zn, and this is due, at least to some extent,
to the hydrogen-evolution reaction interfering with the coulometric efﬁciency of the

metal deposition. The stripping peak potential for each metal shifted positive with

88

deposition of a metal ﬁom solutions of increasing metal ion concentration, consistent
with an increased surface coverage. It, therefore, takes more time during the scan to
completely oxidize the metal phase and the peak appears at more positive potentials. The
response precision and stability of the measurement were generally < 5% for all the ﬁve

metal ions tested at both diamond types.

Table 4.3 Summary of the differential pulse anodic stripping voltammetric (DPASV)
detection ﬁgures of merit for the measurement of individual metal ions at
microcrystalline (M) and nanocrystalline (N) diamond.

 

 

Linear Dynamic Limit of S . . . Response Response
Parameter Range(ppb) Quantitation (erghvgy (‘5)ICI81clt0) (stfggy
11 pp 0 n = o y
(r2 > 0.99) (ppb)(S/N > 3) test)
Metal ‘0“ M N M N M N M N M N

 

Ag(I) 0.1-1000 0.1-10,000 0.11 0.11 0.098 0.111 1.1 0.91 2.6 3.7

 

Cu(II) 0.6-6400 0.6-6400 0.64 0.64 0.085 0.045 1.0 1.2 1.9 4.6

 

 

Pb(II) 20-20,000 2-20,000 20.1 2.07 0.045 0.050 1.6 0.41 2.9 1.7

 

Cd(II) 11-1100 1.1-11,000 11.2 1.12 0.072 0.099 0.7 1.2 4.6 3.9

 

Zn(II) 65-6500 6.5-6500 65.3 6.53 0.028 0.035 3.5 4.5 6.5 4.5

 

 

 

 

 

 

 

 

 

 

 

 

4.3.4 Differential Pulse Voltammetry of Mixtures of Metals Ions. The

simultaneous analysis of three metal ions, Ag(I), Pb(II) and Cd(II) from a solution
containing them all was investigated using nanocrystalline diamond.2 Figure 4.6 shows
the differential pulse anodic stripping voltammehic i-E curves for Ag, Pb and Cd for a
boron-doped nanocrystalline diamond electrode in acetate buffer, pH 4.5. The

concentration of the metal ions in each solution was (a) 5 (b) l (c) 0.5 (d) 0.1 uM. The

89

deposition was performed at -1000 mV for 3 min. The peaks were identiﬁed by peak
potential matching of the response for individual metal ions. The shipping peaks for the
Cd, Pb, and Ag, deposited from 10 uM solution concentration, appear at -781 (106), -

432(228), and 332 (116) mV. The FWHM values for each peak in mV are given in

 

 

 

 

 

 

 

parentheses.

80

Pb(II)
A I

A 60 __ 8()
<1:
3
E
g 40 ...
6 b

20 _

Cd(II)
0 _. “A , . - ~-
-1000 -600 -200

 

600
Potential (mV vs. Ag/AgCl)

Figure 4.6 Differential pulse anodic shipping voltammehic (DPASV) i-E curves for Ag,
Pb and Cd on boron-doped nanocrystalline diamond in acetate buffer, pH 4.5.The
preconcentration was performed for 3.5 min. in quiescent solution at -1000 mV. The
concentration of the metal ions in each solution was (a) 5 (b) l (c) 0.5 (d) 0.1 uM.

The stripping peaks all increased with the metal ion concentration in solution. The
linear dynamic range for both electrodes was 4 to 5 orders of magnitude (r2 > 0.99). The
limit of quantitation for all three metal ions was 0.01 11M, or 1, 2 and 1 ppb for Cd(II),

Pb(II) and Ag(I), respectively. The S/N ratio for Cd peak was 3 and for Ag and Pb was

90

 

ca. 30 at this concentration. It can be seen that the Pb stripping peak formed a doublet
when both Cd and Ag are co-deposited. In another series of experiments, it was observed
that the standard additions of either Cu(II) or Ag(I) to a test solution containing just
Pb(II) produced a Pb stripping peak doublet. Co-deposition of Cd or Zn did not inﬂuence
the Pb peak. It was also observed that the Cd peak was suppressed in the mixture
compared to the peak magnitude in the individual metal ion measurement. Experiments
were conducted to ascertain the cause for the suppression. The addition of either Ag(I),
Cu(II) or Pb(II) to the solution caused the suppression of the Cd peak. However, the
effect was greatest with the addition of Ag(I). It seems from experiments involving
standard additions of Cd(II) to the solution mixture that the ratio of Cd(II) to the rest of
the metals ions is critical. At a ratio of 2:1, the ratio of the Cd shipping peak charge in the
mixture to the peak charge for the individually stripped metal is approximately 1. When
co-deposited with Cd, the stripping charge for Ag is enhanced suggesting that some of

the deposited Cd is stripping at the Ag oxidation potential.12

Further, Figure 4.7 shows a example of a differential pulse anodic shipping
voltammetric i-E curve for a mixture of Zn, Cd, Pb, Cu, and Ag at nanocrystalline
diamond. The metal ion concentrations were all 10 uM. Quantiﬁable peaks are seen for
Cd, Pb, Cu, and Ag. The peak for Zn is suppressed and is not distinguishable due to the
intermetallic compound formation with Cu.13 '15 The peak for Cd is also suppressed, as
observed in the three metal ion measurement. Further, at concentrations as low as 0.01
uM, the Ag peak also tended to form a doublet. The limit of quantitation of Ag(I)

(S/N=7), Cu(II) (SfN=8), Pb(II) (S/N=3.2) and Cd(II) (S/N=3.1) in the mixture is 0.01

91

11M or 1, 0.6, 2 and 1 ppb, respectively. These data along with the other ﬁgures of merit
for similar measurements using microcrystalline diamond are compiled in Table 4.4. The
coefﬁcients of variation for both kinds of diamond were generally less than 5 % for the
mixture of these 5 metals. The stripping peaks for nanocrystalline diamond were

observed at potentials 30-50 mV positive of those seen for microcrystalline diamond.

 

80
Pb(II)

M Ag(I)
Cu(II)
60 p /

1 l /

20 a.

.___.._.-
M!

M"
4 MAM—
...._._A —~—‘—-

Current (uA)

Cd(II) /
Zn(II)

 

 

0 l l l l l l l l
-1200 -1000 -800 -600 -400 -200 0 200 400 600

 

Potential (mV vs. Ag/AgCl)

Figure 4.7 Differential pulse stripping voltammetric (DPASV) i-E curves for the
detection of 10 uM solution concentrations of Ag, Cu, Pb, Cd and Zn for nanocrystalline

diamond ﬁlm in acetate buffer, pH 4.5.The preconcentration was carried out for 3.5 min
in quiescent solution at -1200 mV.

92

Table 4.4 Summary of the differential pulse anodic stripping voltammetric detection
ﬁgures of merit for mixtures of metal ions at microcrystalline (M) and nanocrystalline
(N) diamond.

 

Response

 

Linear Dynamic Limit of . . . . . Response

Range(ppb) Quantitation 851157113), Pig/13011 Stability(%)

Metal ( r2 > 0.98) (ppb)(S/N > 3) 1‘ pp in = 10) ( 3 day test)
M N M N M N M N M N

 

Ag(I) 1-10000 1-10000 1.1 1.1 0.42 0.53 3.2 2.9 3.7 1.1

 

Cu(II) 0.6-6400 0.6-6400 0.6 0.6 0.22 0.47 3.5 1.2 4.5 2.3
Pb(II) 2-20000 2-20000 2.1 2.1 0.19 0.37 3.0 2.5 1.2 1.8
Cd(II) l-11000 1-1100 1.1 1.1 0.01 0.02 4.2 3.1 5.3 3.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

As shown earlier, the simultaneous deposition and stripping of the metals leads to
complex voltammograms. The Pb stripping peak splits into two components at higher
concentrations when either Ag or Cd is co-deposited. One possible explanation for this is
that the metal is depositing and stripping from two energetically distinct sites on the
electrode surface. For example, the Pb may be oxidizing from a homogeneous Pb deposit
on the diamond surface and from heterogeneous sites where multiple metals have co-
deposited. Another possibility is the formation of an intermetallic compound. It is
supposed that the dual Pb peaks observed presently are caused by the formation of Ag-Pb

and Cu-Pb interrnetallic compounds.16

The suppression of the Cd(II) peak may be partly attributed to the formation of an
Ag-Cd complex. Voltammetric data showing the deposition of Ag-Cd alloys from

mixtures of Cd(II) and Ag(I) have been reported.16 The alloy was of a solid solution type

93

at low concentration of Cd(II) and a mixture of phases at higher concentration. In alkaline
media, several researchers have reported two distinct peaks for the transitions of silver to
argentous oxide and argentous oxide to argentic oxide.l7 It is still uncertain whether silver
is undergoing the same kind of transition on diamond at very low concentrations of metal

in this relatively acidic media.

Further, the sensitivity readings for the five metal detection were greater than
individual metal detection by a factor of approximately 4-10 for Ag(I), Cu(II), and Pb(II),
however they were less by a factor of 2-5 for Cd(II). Zn(II) was not detected in the
mixture. This could indicate that Cd(II) and Zn(II) are interacting with the other metals
and getting stripped off as interrnetallic compounds at the peak potentials of the other
metals forming binary/ ternary compounds. See Chapter 6 for a detailed study of the Cu-

Zn intermetallic compound formation.

4.3.5 Scanning Electron Microscopy (SEM) Studies of Metal Phase

Formation. As described in Chapter 2, metal deposition on diamond surfaces is a

complicated process and is not completely understood. The metal under study must
nucleate and grow on the surface. The sites at which this occurs, as well as the nucleation
and growth mechanism for many metals are not known. Cu deposits equally well on both
the facets and the grain boundaries of highly boron-doped microcrystalline diamond.2
The metal deposits by an instantaneous nucleation and growth mechanism at low
overpotentials and progressive mechanism at high overpotentials. Ag deposits on

microcrystalline diamond have also been studied.‘8 At low overpotentials, an

94

 

instantaneous nucleation and growth mechanism was observed, while at high
overpotentials, a progressive mechanism was found. Zn deposits on microcrystalline

diamond by a progressive mechanism at both low and high overpotentials.‘9

SEM was used to image the metal deposits formed on the diamond surface.
Representative images of films containing different deposited metals are shown in Figure
4.8. The deposition, in each case, was for 3 min at —l.2 V (no stirring). The Cu deposits
are small (mean diam. of 40 i 13 nm), uniform across the surface, numerous, and well
dispersed on the microcrystalline film. All the facets are decorated with deposits. The Ag
deposits are also relatively small (mean diam. of 142 i 73 nm) and well dispersed, but
they have a wide range of particle diameters. The Cd deposits are fewer in number and
larger (mean diam. 245 i- 211 nm). They appear to have distinct prismoidal geometries.
The Pb deposits are fewer in number, small in diameter and scattered (mean diam. ~ 156
i 21 nm). The Zn deposits (mean diam. ~ 138 i 76 nm), do not appear to be well

distributed and have a clustered appearance.

A similar trend is observed for the deposition of metals on the nanocrystalline
diamond. Deposits of Ag are fairly spherical and well dispersed but differed in size (
mean diam. ~ 82 :t 48 nm). The Cu deposits have a greater uniformity in size and are
numerous and well distributed (mean diam. ~ 32 i 11 nm). The Cd deposits are few and
unevenly distributed, although did appear to have a somewhat geometrical shape (mean

diam. ~ 115 i 33 nm). Pb deposits were globular, and are not well dispersed over the

95

surface. They possess a very large size (mean diam. ~ 481 i 19 nm). The Zn deposits are
also few in number and randomly dispersed (mean diam. of 137 i 21 nm). The surface
area of a ﬁlm, which was covered by metal, was determined using software assisted
image analysis. Table 4.5 shows a comparison of nominal particle size, particle density
and fractional surface area of the metal deposits for both diamond types. Except for Pb,
the nominal particle size was less and the particle density greater for nanocrystalline
diamond. The fraction of the surface area of the diamond covered by metal deposits was
larger in all cases (except on Cd where it was nearly equal) for nanocrystalline diamond.
This means that more metal phase formation occurs on diamond during the
preconcentration step and this translates into greater sensitivity and lower limits of

quantitation in the DPASV measurements.

Table 4.5 Comparison of the nominal particle size, density and fractional surface area of
the diamond ﬁlm covered by the metal phase, as determined from scanning electron
microscope (SEM) images.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Microcrystalline Diamond Nanocrystalline Diamond
Nominal Particle Fractional Nominal Particle Fractional
Metal Particle density Surface Particle density Surface
Diam. (nm) (cm-2) Area Diam. (nm) (cm-2) Area
Ag 142:73 3,0x107 0.019 823:48 11 x 108 0.173
Cu 40.2 :1: 13 111x108 0.040 32 i 11 2,1 x 108 0.109
Cd 245 i 211 4,0 x106 0.037 115 i 33 2.1 x 107 0.023
Pb 156 i 21 7,0 x106 0.040 481119 10 x 106 0.146
Zn 138 :t 76 7.0 x106 0.027 137 i 21 7,0 x 106 0.019

 

96

 

 

1000 nm Microcrystalline Diamond (top)
200 nm Nanocrystalline Diamond (bottom)

 

Figure 4.8 Scanning electron micrographs of metal phases deposited on microcrystalline
(top) and nanocrystalline (bottom) diamond thin-ﬁlm electrodes. The deposition, in each
case, was for 3 min at —1.2 V (no stirring). Arrows indicate a few representative metal
deposits.

4.4 Conclusions

The use of two types of diamond thin ﬁlms for the detection of trace metal ions,
individually and in mixtures, was demonstrated. The detection limits were lower or
comparable to those mandated by EPA for the maximum allowable levels of these metal
ions in drinking water (given in Chapter 1). The analytical detection ﬁgures of merit
obtained with diamond compared well with those obtained using Hg-coated GC (the state
of the art electrode material for this measurement).

Both diamond types functioned well for the detection of the metal ions and
yielded compelling detection ﬁgures of merit. However, the limit of quantitation was
slightly lower and the sensitivity greater for the detection of the metal ions, both
individually and in mixtures, with nanocrystalline diamond. As an example, the linear
dynamic range for the detection of Pb (11) at nanocrystalline diamond was four orders of
magnitude from 2 to 20,000 ppb, the limit of quantitation was 2 ppb (S/N > 3), and the
response precision was 0.4% (n=10). The linear dynamic range at microcrystalline
diamond was from 20 to 20,000 ppb, the limit of quantitation was 20 ppb (S/N > 3), and
the response precision was 1.6 % (n=10). The greater sensitivity and lower limit of
quantitation for nanocrystalline diamond is due to the increased particle density and
smaller particle diameter (i.e., increased metal phase formation) during the
preconcentration or deposition step. It appears as though the surface morphology of the
nanocrystalline diamond possesses a higher nucleation density for metal phase formation
then does microcrystalline diamond. This may be caused by the higher fraction of grain

boundary on the nanocrystalline ﬁlm and or a higher fraction of electrically conducting

98

regions. We conclude from this work that nanocrystalline diamond performs in a superior

fashion to microcrystalline diamond for this measurement.

In comparison with LSV, the use of the DPASV technique enabled improved

analytical detection ﬁgures of merit for the analysis of Ag(I), Cd(II) and Hg(II) (From

Tables 4.3 and 4.1, respectively).

99

4.5 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Granger, M. C.; Xu, J .; Strojek, J. W.; Swain, G. M. Analytica Chimica Acta
1999, 397, 145-161.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chemistry of Materials
2003, 15, 879-888.

Berzins, T.; Delahay, P. Journal of the American Chemical Society 1953, 75, 555-
559.

Xu, J.; Granger, M. C.; Chen, Q.; Strojek, J. W.; Lister, T. B.; Swain, G. M.
Analytical Chemistry 1997, 69, 591A-597A.

Xu, J .; Chen, Q.; Swain, G. M. Analytical Chemistry 1998, 70, 3146-3154.

Alehashem, 8.; Chambers, F.; Strojek, J. W.; Swain, G. M.; Ramesham, R.
Analytical Chemistry 1995, 67, 2812-2821.

Holt, K. B.; Bard, A. J .; Show, Y.; Swain, G. M. Journal of Physical Chemistry B
2004,108,15117-15127.

Awada, M.; Strojek, J. W.; Swain, G. M. Journal of the Electrochemical Society
1995, 142, L42-L45.

Reinmuth, W. H. Analytical Chemistry 1961, 33, 1793-1794.

Vydra, F.; et al., Eds. Electrochemical Stripping Analysis; Ellis Horwood Ltd.,
1976

Wang, J ., Ed. Stripping Analysis: Principles, Instrumentation, and Applications;
VCH Publishers, Inc., 1985

Temmerrnan, E.; Verbeek, F. Analytica Chimica Acta 1972, 58, 263-272.

100

(13)

(14)

(15)

(16)

(17)

(18)

(19)

Copeland, T. R.; Osteryoung, R. A.; Skogerboe, R. K. Analytical Chemistry 1974,
46, 2093-2097.

Stromberg, A. G.; Gorodovykh, V. E. Zhurnal Neorganicheskoi Khimii 1963, 8,
2355-2359.

Marques, A. L. B.; Chierice, G. O. Talanta 1991, 38, 735-739.

Bonﬁl, Y.; Kirowa-Eisner, E. Analytica Chimica Acta 2002, 457, 285-296.

Clarke, T. G.; Hampson, N. A.; Lee, J. B.; Morley, J. R.; Scanlon, B. Berichte der
Bunsen-Gesellschaft 1969, 73, 279-283.

Vinokur, N.; Miller, B.; Avyigal, Y.; Kalish, R. Journal of the Electrochemical
Society 1999, 146, 125-130.

Wang, J .; Swain, G. M. unpublished results 2002.

101

CHAPTER 5

TRACE METAL ANALYSIS IN CONTAMINATED WATER
SAMPLES USING ANODIC STRIPPING VOLTAMMETRY AND A
NANOCRYSTALLINE DIAMOND THIN-FILM ELECTRODE

5.1 Introduction

There is emphasis for the use of diamond electrodes in stripping voltammetric
detection of metal ions due to their favorable properties for this application."3 Recently,
cathodic stripping voltammetry at a bare boron-doped diamond electrode was shown to
be a reliable and selective method to evaluate the Mn content of marine sediments with a
limit of detection of 7.4 X 10'7 M (41 ppb) with a deposition time of 60 s.4 The
electroanalytical determination of Pb(II) by ASV at in-situ-formed, bismuth-ﬁlm-
modiﬁed, boron-doped diamond electrodes was also demonstrated. Detection limits in
0.1 M nitric acid of 9.6 x 10'8 M (0.2 ppb) and 1.1 x 10’8 M (2.3 ppb) were obtained aﬁer

60 and 300 s deposition times, respectively. 5

In this Chapter, results from analysis of Ag(I), Cu(II), Pb(II), Cd(II), and Zn(II) in

“real world” water samples by DPASV are presented.6 Chapter 4 showed that
nanocrystalline diamond typically outperformed microcrystalline diamond in DPASV of
metal ions. Therefore, nanocrystalline diamond ﬁlms were used to conduct the
measurements described herein. Good analytical detection ﬁgures of merit were observed
With limits of quantitation in the low ppb range, good sensitivity, and excellent response
Precision and stability (~ 4 %). Data for the detection of these metal ions in four sample
tYPes are presented: lake water, tap water, wastewater treatment sludge, and soil. The

%0 als for this work were (i) to investigate the feasibility of detecting trace metal ions in

102

“real-world” water samples using diamond electrodes, (ii) to obtain detection ﬁgures of
merit for metal ions in several types of contaminated water samples, and (iii) to validate
the method through comparison measurements with inductively coupled plasma mass-

spectrometry (ICP-MS) and or atomic absorption spectrometry (AAS).

5.2 Experimental

Diamond F iim Deposition. The boron-doped nanocrystalline diamond thin ﬁlm was

deposited on a p-type Si(100) substrate ( ~10'3 Q-cm), using a commercial microwave-
assisted chemical vapor deposition (CVD) system. Details of the growth conditions, as
well as the follow-up material characterization, were given in Chapter 2. Ultrahigh-purity
CH4, Ar, and H2 (99.999%) were used as the source gases. The gas ﬂow rates were 1, 94,
and 5 sccm, respectively, for CH4, Ar, and H2. B2H6 was used for doping at a ﬂow rate of
0.001 sccm (10 ppm). The microwave power and deposition pressure were maintained at
800 W and 140 torr, respectively. The substrate temperature was estimated to be around
800 °C using an optical pyrometer. The deposition time was 2 h. The diamond thin ﬁlm
was approximately 4 pm thick, as estimated from the weight change after growth, the
density of diamond (3.52 g/cm3), and the geometric area of the Si substrate. The apparent

resistivity was 0.1 Q-cm, or less, as determined using the four-point probe method.

Inductively Coupled Plasma Mass Spectrometry (I CP-MS Analysis). An

inductively coupled plasma mass spectrometer (Micromass Platform ICP-MS) ﬁtted with
a hexapole collision cell was used. A Meinhard concentric nebulizer was used for sample

introduction. Before data acquisition, the performance of the ICP-MS was optimized

103

using a
and U
the m:
C011C€1
stand;
HNO;
conce
soluti
cone:
my
10 e;

£01113 1

01).
Pb <1
Were
We.

using a standard solution containing 10 ug/L of Be (II), Co (11), In (111), Ce (111), Bi (III),
and U (11). Parameters, such as the torch position and gas ﬂow rates, were adjusted until
the maximum and most stable signal was observed for each element. The metal-ion
concentrations in the unknown samples were determined from the calibration curves for
standard solutions. 1 ppm commercial stock solutions of the individual metal ions in 2%
HNO3 were diluted with 2% HNO3 to prepare six standard solutions over a wide
concentration range. For multiple metal-ion analysis, the individual metal ion stock
solutions were mixed together to prepare standard mixed-metal solutions of different
concentration. For analysis, 1 mL of the standard solution was pipetted into a separate
sample tube. 1 mL of an internal standard solution of In (111) and Bi (III) was then added
to each tube. The standard solutions were analyzed ﬁrst, in order of increasing
concentration, and this was followed by analysis of the unknowns. The response for each
sample was acquired over a one-minute period. The system was thoroughly rinsed with
the blank solution (2% HNO3) after each measurement. The peak intensities of the

resultant spectra recorded for the standard solutions were used to construct calibration

curves.

1 ppm stock solutions of Cu (11) (CPI, Santarosa, CA), Zn (II) (CPI, Santarosa,
CA), Cd (11) (Spex Certiprep, Metuchen, NJ), Ag (1) (Spex Certiprep, Metuchen, NJ), and
Pb (11) (Spex Certiprep, Metuchen, NJ) in 2% HNO3 were used. Nalgene sample bottles
were used for solution storage and were ﬁrst cleaned in a clean room, using the following
procedure: (i) degreasing with VersaClean soap, (ii) washing with acetone, (iii) soaking

in conc. HCl, (iv) soaking in conc. HNO3, (v) soaking in warm (40-50°C) 2 M HNO3,

104

(vi) soaking in 0.5% HNO3, and (vii) rinsing with Milli-Q water (Steps 5 and 6 done in

Class 100 lab).

Atomic Absorption Spectrometry (AAS) Analysis. Wastewater samples were
analyzed for Cu(II), Cd(II), Zn(II), Ag(I), and Pb(H), using a Hitachi Z-9000 Atomic
Absorption Spectrophotometer with a graphite tube fumace attachment. The instrument
possessed a diffraction grating monochromator, which could be set to the desired
wavelength for monitoring the absorbance by a particular metal ion. Cd, Cu, Pb, Ag, and
Zn hollow cathode lamps were used with excitation wavelengths of 228.8, 324.8, 283.3,
328.1, and 213.9 nm, respectively. The lamp currents were set at 7.5 mA. The sheath gas
used was argon, and its ﬂow was maintained at 200 mL/min. A 20 uL sample volume
was introduced into the furnace via an autoinjector. The sample was dried and ashed at
pre-speciﬁed temperatures and times. Standard solutions containing known

concentrations of Cd(II), Cu(II), Pb(II), Ag(I), and Zn(II) were prepared and response

curves constructed for each.

Differential Pulse Anodic Stripping Voltammetry (DPAS l0 Analysis. A

single-compartment glass cell was used for all the electrochemical measurements, along
with a computer-controlled potentiostat (Model 650a CH Instruments Inc., Austin,
TX)[9]. The electrochemical cell was placed inside a Faraday cage for shielding. The
diamond thin-ﬁlm working electrode was pressed against a Viton® o-ring and clamped
to the bottom of the cell. A bead of an In/Ga alloy was placed on the backside of the

scratched and cleaned Si substrate to ensure good ohmic contact with a copper current-

105

collector plate. A graphite rod was used as the counter electrode, and a commercial
Ag/AgCl electrode (saturated KCl) served as the reference (E°Ag/Aga = -0.045 V vs.
SCE). The geometric area of the working electrode was ca. 0.2 cm2. All measurements
were made in solutions deoxygenated with N2 for at least 10 min initially, and then for 2
min after each anodic stripping or oxidation cycle. The solutions were blanketed with the
gas during all measurements. The electrolyte solution was 0.1 M acetate buffer, pH 4.5.
The reference electrode was placed inside a cracked-glass capillary (double junction)
that was ﬁlled with the acetate buffer. All measurements were made in an unstirred
solution. The typical differential pulse voltammetric settings were (i) a 3 min deposition
time, (ii) a 0.5 min. “quiet” time prior to initiation of the anodic sweep, (iii) a 50 mV
pulse height, 2 mV step height, 50 msec pulse width, 100 msec cycle period, and 35
msec sampling time during the stripping step, and (iv) a constant potential of 600 mV for
120 s after completion of the anodic sweep to fully oxidize all metal deposits prior to the

HBXI measurement.

NIS T Standard Solution. A standard reference material® (1640, henceforth referred

to as SRM 1640) was procured for testing from the National Institute of Standards and
Technology (N IST). The aqueous sample contained certiﬁed amounts of Ag(I), Cu(II),
Pb(II), Cd(II), and Zn(II), in addition to at least 24 other dissolved elements. The
certiﬁed values of Cd(II), Pb(II), and Ag(I) were 22.79 :1: 0.96, 27.89 :b 0.14, and 7.62 :h
0.25 ppb, respectively. The reference values of Cu(II) and Zn(II) were given as 85.2 i
1.2 and 53.2 a: 1.1 ppb, respectively. According to NIST, the sample was prepared in the

following manner. About 3500 L of natural (fresh) water was obtained by US.

106

Geological Survey personnel at Clear Creek, CO. It was ﬁltered through a 0.1 pm ultra
ﬁlter and acidiﬁed with nitric acid. Analysis of the water by ICP-MS, before and after
the stabilization process, showed that arsenic, beryllium, cobalt, selenium and zinc
decreased in concentration during the stabilization process. These elements were
adjusted to their original concentrations by the addition of the appropriate salts. The
stabilized solution was then pumped through an ultra ﬁlter, past a UV light source (for
sterilization purposes), and then to a bottling station. At the bottling station, the bottles

were ﬁrst rinsed with the sample and then ﬁlled.

Protocols for Field Sampling. The river, lake, and well water samples were

obtained using a homemade (~ 1 meter) sampling rod with an attached 250 mL ultrapure
Nalgene bottle. The bottle was ﬁlled and rinsed three times with the sample water before
collecting a sample for analysis. The bottle was then capped, labeled, and stored in an

acid-washed bag for return to the laboratory. Prior to analysis, the sample was ﬁltered to
remove solid particulates using a syringe ﬁlter (ACRODISCS Syringe Filter, 0.2 pm,

Tuffryn Membrane, low protein binding, non-pyrogenic, hydrophilic and bi-directional),
and transferred into a 125 mL Nalgene bottle. 20 11L of acid (70% Seastar HNO3) was
then added. The sample was thoroughly mixed and stored in the reﬁ'igerator until use.

The acidiﬁed water samples were analyzed by DPASV, AAS, and or ICP-MS.

Wastewater and Soil Samples. A heavy wastewater sludge sample was obtained

from the East Lansing Water Treatment Facility. The heavy sludge solution (50 mL) was

acid digested for several hours after adding 100 mL of ultrapure, concentrated HNO3.

107

 

The solution was then decanted and centrifuged for several minutes at ca. 3200 rpm. The

supernatant was decanted and then analyzed by DPASV and AAS.

Three contaminated soil samples were procured from a former manufactured-gas
plant (MGP) facility located near downtown Grand Ledge, Michigan. The soil samples
were obtained using geo-probes that were sent down 10-11 (sample 1), 4-6 (sample 2),
and 8-10 (sample 3) feet into the ground. The collected soil samples were acid digested
following EPA Method 3050A. This involved the digestion of 1g of soil (wet weight) in
25 mL of 1M nitric acid and 10 mL of 30% hydrogen peroxide. The resultant digestate
was reduced in volume by solvent evaporation and then diluted to a ﬁnal volume of 100
mL using ultrapure water. Particulates in the digestate were removed by ﬁltration
(Whatman No. 41 ﬁlter paper). The resultant soil extract was tested for Zn(II), Pb(II) and

Cu(II) using DPASV and AAS.

Chemicals (AAS and DPAS V) All solutions were prepared with ultrapure water

(>17 MQ-cm) from an E-pure puriﬁcation system (Bamstead). Silver nitrate (Aldrich),
Cadmium nitrate (Aldrich), zinc acetate (Aldrich), cupric nitrate (Aldrich), and lead
nitrate (Aldrich) were all reagent- grade quality and used without additional puriﬁcation.
0.1 M acetate buffer was prepared by mixing appropriate amounts of sodium acetate
(99%, Aldrich) and acetic acid (99%, Aldrich) solutions. All solutions were prepared
fresh daily from their stock solutions and purged with N2 (99.99%) for 10 min prior to
any electrochemical measurement. All glassware was cleaned by a three-step procedure:

ethanol/KOH bath, alconox/ultrapure water solution, and ultrapure water rinse.

108

.L' 3‘2”

 

5.3 Results and Discussion

5.3.1 Analysis of a NIST Standard Solution

Figure 5.1 shows a differential pulse stripping voltammetric i-E curve for the
NIST sample, as well as the curves for standard solutions of Pb, Cu, and Ag at equal
concentrations ranging from 10 to 90 ppb. The curves are all background corrected. The
background was an differential pulse stripping voltammtric i-E curve of the electrolyte
solution under the same experimental conditions. The NIST sample has peaks for Pb, Cu,
and Ag at potentials of ca. —450, ~80 and 280 mV. The peak for Pb in all the solutions is
nearly symmetric with a base width of 150 mV. The peak for Cu, particularly for the
NIST sample, is asymmetric with a secondary stripping current about 100 mV positive of
the main peak. As a result, the base width of the peak is about 300 mV for this sample.
The peak width for the standard solution is ca. 150 mV. The peak for Ag for the NIST
sample, as well as the standard solutions, is symmetric with evidence for two current
maxima. The base width is about 150 mV. In all cases, the peak current and charge
increase proportionally (r2 > 0.99) with the solution concentration of each metal ion.
Table 5.1 presents DPASV and AAS and ICP-MS determined concentration data for all
ﬁve metal ions. The concentrations were determined by the standard addition method. As
an example, a plot of peak charge versus the concentration of Ag(I) added is shown in
Figure 5.2. The concentration of the unknown is determined from the intercept on the
negative x-axis, which in this case is 7.3 ppb. There is good agreement between the AAS

and DPASV data for all the metal ions, except Cd(II). The response precisions for the

109

 

 

measurements with diamond are very good as the coefﬁcients of variation range from 3-4
%. The percent differences between the values for the two methods range from a low of
2.7 % for Zn(II) to high of 11.1 % for Cd(II). The percent differences for some of the
elements (e.g., Cd and Cu) are higher than desired. It appears that the large differences
seen for Cd(II) and Cu(II) are due, at least in part, to the complexity of the NIST solution
as such large differences for these two metal ions were not observed for other real world
samples, as is reported below. The data indicate that diamond electrodes provide an
analytically useful response for these metal ions that is comparable to other established

methods.

 

#- ii 90 ppb C1101)

 
 
   
   

 

Pb 11) 10 ppb

SRM 1640

Current ( 11A)

 

 

-600 A -400 -200 A 0 i 200 400
Potential (mV vs. Ag/AgCl)

Figure 5.1 Differential pulse anodic shipping voltammetric (DPASV) i-E curves for a
standard solution (SRM 1640) and known solutions of Pb, Cu and Ag using a boron-
doped nanocrystalline diamond electrode. Deposition potential = - 900 mV. The DPASV
conditions are described in the Experimental section. Geometric area = 0.2 cm2. Peak

area integration range = 200 to 370 mV for Ag(I), -200 to 100 mV for Cu(II), and -550 to
-380 mV for Pb(II).

110

 

 

14W
124 /'
8 101
:1. y"
T; 8‘ /
g3 6‘. /
4. /
23
O ' I l ' I I I I
-20 O 20 40 6O 80 100
Concentration (ppb)

 

Figure 5.2 Standard addition plot of charge vs. concentration for the addition of solutions
containing Ag(I) in increasing concentrations ranging from 10 ppb to 90 ppb to a solution
of the SRM diluted in acetate buffer (1:9). All concentrations reported have been
corrected for dilution.

Table 5.1 Concentration and percent error data for Ag(I), Cu(II), Cd(II), Pb(II) and
Zn(II) in the SRM 1640 water samples.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Results Ag(I) Pb(II) Cd(II)
(329353? ICP AAS ASV ICP AAS ASV ICP AAS ASV
8.04 7.32 25.1 30.2 26.5 25.3
SRM 1640 7‘01 :073 :056 28'3 :1.6 :08 23 '7 :1.2 :15
Error % 8.0 5.0 4.2 1.4 10.0 8.4 4.03 16.3 11.1
Results Cu(II) Zn(II)
0?: Egg”): ICP AAS ASV ICP AAS ASV
89.3 76.5 57.0 51.8
SRM 1640 80'9 :06 :31 57'6 : 3.7 :28
Error % 5.1 4.8 10.2 4.4 7.1 2.7

 

111

 

 

5.3.2 Drinking Water Analysis

Our ﬁrst attempt to employ diamond electrodes for metal ion detection in “real
world” samples was in the analysis of tap water. Two different samples from the MSU
campus were analyzed by DPASV and ICP-MS. Cu(II) was the only metal ion found at
detectable concentrations in either water sample. Figure 5.3 shows differential pulse
stripping voltammetric i-E curves for the two tap-water samples, and for standard
solutions of Cu ranging in concentration from 30 to 130 ppb. The curves shown are all
background corrected. Unlike in the voltammetric i-E curve for the NIST sample (see
Fig. 5.1), the stripping peak for Cu in these solutions is nearly symmetric, even though
the electrode was the same in both cases. The more asymmetric peak shape for the NIST
sample is probably caused by the complex nature of the solution from which the Cu(II) is
being analyzed (25+ elements present in addition to Cu(II)). The base peak widths in
Figure 5.3 range from 150 to 300 mV, with the width widening with increasing
concentration of Cu(II) (i.e., increasing surface coverage). The peak current and charge
both increase proportionally with the solution concentration of Cu. Based on the peak
charge, tap-water sample #2 has a lower Cu(II) concentration than does sample #1. The
peak charge was used for quantitation via the standard addition method. The standard
addition curve of the peak charge versus the added Cu(II) solution concentration was
linear over two orders of magnitude (1-130 ppb) with a least squares correlation
coefﬁcient of >0.990. This does not represent the full linear dynamic range for the assay,
but rather is only the range of concentrations used in the measurement.7 Table 5.2
presents the concentration of Cu(II) in the two samples, as determined by DPASV and

ICP-MS. The Cu(II) concentration in samples #1 and #2 was determined by DPASV to

112

A I. V!"

 

 

‘17-! m: «ﬂu-'5. bl!

be 91.2 i 3.5 and 29.1 i 1.3 ppb, respectively. At least three measurements were made to
evaluate the precision of the DPASV technique. The coefﬁcients of variance were 3.8
and 4.5% (n23), respectively, meaning that the measurements with diamond are very
reproducible. Only one ICP-MS measurement was made on each sample so the precision
of the measurement cannot be assessed. The ICP-MS measurements revealed a Cu(II)
concentration of 87.3 and 26.1 ppb, respectively, for the two samples. Table 5.2 also
presents percent difference data between the two methods. For both samples, the DPASV
value is larger than the ICP-MS one, with a difference of 4.5% in the values for sample
#1 and 11.4% for sample #2. These numbers indicate a good agreement between the
values determined by DPASV and ICP-MS, especially for sample #1. The difference in
the values for sample #2 is larger than one would like, but since we do not know the
variance of the ICP-MS measurement, we cannot be certain that this difference is really
this large. The results reveal that diamond provides an analytically-useful response for
Cu(II) in these tap water samples, in terms of the sensitivity, response reproducibility and

accuracy, at least to the extent that we can evaluate the latter.

113

14
13 L Cu(II)
12
ll
10

 

  
 

‘ﬁ—F‘TW’VIV

Tap water 1

fj’ '7 I V

Tap water 2

l'T'1

Current (11A)

 

OHNw-ﬁ-MGQ

 

l . _l_

-100 0 100 200

. :—
N
8

r-

Potential (mV vs. Ag/AgCl)

Figure 5.3 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for Cu
in tap-water samples #1 and #2 using a boron-doped nanocrystalline diamond electrode.
Deposition potential = -200 mV. The DPASV conditions are described in the
experimental section. Geometric area of electrode = 0.2 cm2. Peak area integration range
= -100 to lOOmV for Cu(II).

Table 5.2 Concentration and percent difference data for Cu(II) in two different tap water
samples from the MSU campus as determined by DPASV and ICP-MS.

 

 

 

 

 

 

 

 

Cu(II)
Concentrations in ppb, r2 > 0.98 ICP-MS ASV Difference%
Tap Water 1 (Spartan Village
sample, MSU campus) 87.3 91.3 i 3.5 4.5
Tap Water 2 (Unrvers1ty V1llage 26.1 29.1 i 1.3 11.4
sample, MSU campus)

 

 

Lake, Well and River Water Analysis. The nanocrystalline diamond electrode was

also used to assay for metal ions in lake (Lake Lansing), well, and river water (Red
Cedar) samples. The lake, well and river water samples were from the East Lansing, MI

area. Surprisingly, none of the metal ions probed for (Ag(I), Cu(II), Pb(II), Cd(II), and

114

 

Zn(II)) were detected in the river water sample. On the other hand, Zn(II) and Cu(II)
were found in both the well and lake water samples. The samples were analyzed by both
DPASV and AAS, for comparison. The stripping peaks for Zn(II) and Cu(II) (data not
presented) were symmetric and easily quantiﬁable. Quantitation was accomplished from
the peak charge via the standard addition method. The concentrations of Cu(II) and Zn(II)
in the well water sample were determined to be 28.4 i 1.4 and 64.1 i 0.4 ppb,
respectively. At least three measurements were made on the sample, with a resulting
coefﬁcient of variance of 4.8 and 0.68% for Cu(II) and Zn(II), respectively. The accuracy
of the electrochemical method was assessed by analyzing the same sample using AAS.
The concentrations of Cu(II) and Zn(II) were determined to be 26.7 i 2.9 ppb and 63.0 i
0.3 ppb), respectively. The response coefﬁcient of variance was found to be 11% for
Cu(II) and 0.54 % for Zn(II) (n23). The data obtained by DPASV are in good agreement
with the AAS results, reflecting the accuracy of the method. In the lake-water sample,
only Cu (11) was detected by DPASV at a concentration of 6.30 i 2.8 ppb. This value is

in good agreement with the AAS value of 7.29 i 2.8 ppb.

5.3.3 Wastewater Treatment Sludge Analysis

Figure 5.4 shows differential pulse stripping voltammetric i-E curves for a
wastewater treatment sludge sample, as well as the curves for standard solutions of Cd(II)
and Pb(II) ranging in concentration from 50 to 1000 ppb. The curves are all background
corrected. Sludge is the solid residue generated during the treatment of wastewater.
Heavy metal ions tend to collect in this medium along with poorly biodegradable organic

compounds and potentially pathogenic organisms (viruses, bacteria, etc). There is great

115

interest in analyzing wastewater and sewage sludge because of the potential for

contamination of surface and groundwater supplies, if inadequately processed.8

Pb(II) and Cd(II) are the two metal ions detected in the sludge extract. The sample
preparation is described in the Experimental section. Stripping peaks for Cd (ca. —750
mV) and Pb (ca. —525 mV) are present. The peaks for both elements in the standard
solutions are symmetric with base peak widths of approximately 150 mV each; However,
the peaks in the curve for the sludge extract are broader, particularly for Pb. The peak
current and charge for both elements increase proportionally with the solution
concentration (r2 > 0.99). The standard addition curves of peak charge versus the
concentration of added Pb(II) and Cd(II) were linear over at least two orders of
magnitude (10-1000 ppb) with a correlation coefﬁcient 20.990. Again, this does not
represent the full linear dynamic range of the assay but rather is only the range of
concentrations used for the calibration. The concentrations of Cd(II) and Pb(II) were
determined by DPASV to be 145 i 3.2 and 124 i 4.9 ppb, respectively, with coefﬁcients
of variance of 2.2 and 3.9 % (n23). The solution was also analyzed by AAS, and
concentrations of 156 i 5.3 and 134 i 1.6 ppb, respectively, were determined for Cd(II)
and Pb(II). The DPASV concentrations are lower by 7-8 % than those determined by
AAS. This relatively small percent difference reﬂects the accuracy of the DPASV

method. The results reveal that diamond provides a sensitive, reproducible and accurate

response for Cd(II) and Pb(II) in this somewhat complex sample.

116

 

I" .‘I an"? . I ..

 

 

 

 

100 r- Pb(II)
1000 ppb
80 ~
r-x
<2
:1. 60 -
\_/
q—a
E; 40 l d
r Waste 5 u e
t 8
3 r-
0 20 -
O F-

 

 

 

-1000 -800 -600 -400 -200

Potential (mV vs. Ag/AgCl)

Figure 5.4 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for Cd
and Pb in acid-digested waste sludge using a boron-doped nanocrystalline diamond thin-
ﬁlm electrode. Deposition potential = -1000 mV. The DPASV conditions are described in
the experimental section. Geometric area = 0.2 cm2. Peak area integration range = -600 to
-450 mV for Pb, and -850 to -650 mV for Cd.

Table 5.3 Concentration and percentage difference data for Cd(II) and Pb(II) in a
wastewater treatment sludge extract as determined by DPASV and AAS.

 

 

 

 

Cd(II) Pb(II)
(ppb) (29b)
DPASV 145 i 3 124 :1: 5
AAS 156 :t 5 134 i 2
Difference % 7.5 8.1

 

 

 

 

 

117

5.3.4 Contaminated Soil Analysis

Figure 5.5 shows differential pulse stripping voltammetric i-E curves for an
acidiﬁed extract from a contaminated soil sample, as well as the curves for standard
solutions of Cd(II) and Pb(II) ranging in concentration from 5 to 120 ppb. Again this does
not represent the full dynamic range of the assay but rather only the concentration range
used for calibration. The curves are background corrected. Details regarding the soil
sample preparation are reported in the experimental section. The results reveal that the
soil is contaminated with Cu(II) (ca. 46 mV) and Pb(II) (ca. -482 mV). The peaks for Cu
are reasonably symmetric at all solution concentrations with base widths of ca. 100-200
mV, depending on the solution concentration (i.e., surface coverage of Cu). The single
peak for Pb at ca. -550 mV is symmetric at low solution concentrations. However, a
second peak forms at ca. -410 mV at higher solution concentrations. Both Pb stripping
peaks grow larger in amplitude and area with increasing solution concentration. If these
peaks are associated with oxidation from two different sites, then these sites do not
saturate in coverage during the deposition step. One possibility for the development of
the more positive stripping peak is that some Pb is oxidizing from a site that is a little bit
less active kinetically. This double stripping peak is something we have seen for some
contaminated samples but not all. The stripping peak current and charge for both
elements increase proportionally with the solution concentration with a correlation
coefﬁcient 20.99. Table 5.4 presents the concentration of each element in the two
different soil samples, as determined by DPASV and AAS. The concentrations of Pb(II)
and Cu(II) in soil sample 1 were determined by DPASV to be 97.6 d: 3.1 and 74.5 i 2.7

ppb, respectively. The precision of the measurement was good with coefﬁcients of

118

 

variance of 3.1 and 3.6 % (n23), respectively. The concentrations of Pb(II) and Cu(II) in
soil sample 2 were determined by DPASV to be 272.5 :1: 5.6 and 995 :t 14 ppb,
respectively. Again the precision was good with coefﬁcients of variance of 2.1 and 1.4%
(n23), respectively. There was good agreement between the AAS and DPASV results as

the percent differences range from 2 to 7 %.

 

1

18
16 -
142
12:
10;

V

120 ppb Pb(II)

Sample 1 Cu(II)
5 ppb

 

Current ( 11A)

r

 

    

 

 

7 I ‘7 T V

 

 

ON-hO)

_l_ _

.l.
I
‘5~-.-.--"’

- _ -- 7-1/I
—800 -600 -400 -200

 

 

20 400
Potential (mV vs. Ag/AgCl)

Figure 5.5 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for
Pb and Cu in MGP contaminated soil extracts using a 10 ppm boron-doped
nanocrystalline diamond thin-ﬁlm electrode. Deposition potential = - 1000 mV. The
DPASV conditions are described in the experimental section. Geometric area = 0.2
cm2. Peak area integration range = -50 to 80 mV for Cu(II), and -525 to -425 mV for
Pb(II).

119

 

Table 5.4 Concentration and percent difference data for Pb(II) and Cu(II) in
contaminated soil extracts as determined by DPASV and AAS.

 

 

 

 

 

DPASV AAS Difference %
Sample 1 97.6 : 3.1 92.9 d: 4.5 5.1
Pb(11)(ppb)
Sample 2 272 :t 6 259 :1: 12 5.0
Sample 1 74.5 :1: 2.7 69.5 :1: 1.5 7.2
0101) (Ppb)
Sample 2 995 i 14 972 3: 21 2.4

 

 

 

 

 

 

 

5.4 Conclusions

Boron-doped diamond thin ﬁlm is a viable alternate electrode material for anodic
stripping voltammetry. The material possesses many of the same properties as Hg but is
non-toxic. For Ag(I), Cu(II), Pb(II), Cd(II) and Zn(II), the electrode provides good
sensitivity, low limits of detection, good response precision, and response stability for
several “real-world” sample types: lake and well water, tap water, wastewater treatment

sludge, and contaminated soil.

Even with the favorable detection ﬁgures of merit, research remains to be
conducted in order to further optimize the electrode performance. First, the stripping
peaks are rather broad (ca. 100-150 mV). This is due, at least in part, to the wide range of
metal particle sizes on diamond when using ﬁxed potential deposition. The heterogeneity
of the electrode surface also plays a role. Our initial efforts for the investigation of pulsed
galvanostatic deposition as a means to reduce the nominal particle size are presented in

Chapter 6. Second, although not a signiﬁcant problem in the present analyses,

120

 

intennetallic compound formation remains an issue to be addressed when using non-Hg
alternate electrodes. Several researchers have observed reduced intermetallic compound
formation with electrodeposited Ga.9 Third, the issue of electrode fouling needs to be
further explored. Speciﬁcally, the roles of the diamond surface chemistry, the deposited
metal, and the type of organic matter on the electrode response stability needs to be
further elucidated. In one study reported, the presence of cationic, anionic, and neutral
surfactants and humic substances affected the peak heights of Cu, Pb, and Cd in synthetic
seawater analyzed by differential pulse anodic-stripping voltammetry.10 At surfactant
concentrations of less than 0.1 ppm, the effect was insigniﬁcant, but at higher
concentrations the peak heights usually decreased, although for Cu(II) an increase in the

peak height was also observed.

121

 

5.5 References

(1) Zhang, Y.; Yoshihara, S. Journal of Electroanalytical Chemistry 2004, 573, 327-
331.

(2) Manivannan, A.; Kawasaki, R.; Tryk, D. A.; Fujishima, A. Electrochimica Acta
2004, 49, 3313-3318.

rs
(3) Chatterjee, A.; Wiltshire, R.; Holt, K. B.; Compton, R. G.; Foord, J. S.; Marken,

F. Diamond and Related Materials 2002, 11, 646-650.
(4) Banks, C. E.; Kruusma, J .; Moore, R. R.; Tomcik, P.; Peters, J .; Davis, J .;

Komorsky—Lovric, S.; Compton, R. G. T alanta 2005, 65, 423-429. .3, .

 

(5) Kruusma, J.; Banks, C. E.; Compton, R. G. Analytical and Bioanalytical
Chemistry 2004, 379, 700-706.

(6) Sonthalia, P.; McGaw, B.; Show, Y.; Swain, G. M. Analytica Chimica Acta 2004,
522, 35-44.

(7) Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chemistry of Materials
2003, 15, 879-888.

(8) Farre, M.; Barcelo, D. TrA C, Trends in Analytical Chemistry 2003, 22, 299-310.

(9) Copeland, T. R.; Osteryoung, R. A.; Skogerboe, R. K. Analytical Chemistry 1974,
46, 2093-2097.

(10) Sagberg, P.; Lund, W. T alanta 1982, 29, 457-460.

122

CHAPTER 6

CHALLENGES AND SOLUTIONS: TOWARDS GAINING BETTER
SELECTIVITY AND RESOLUTION DURING THE ANODIC
STRIPPING VOLTAMMETRY OF METAL IONS USING BORON-
DOPED DIAMOND THIN-FILM ELECTRODES

6.1 Introduction

Intermetallic Compound Formation. An intermetallic phase is deﬁned as a

“compound or intermediate solid that contains two or more metals, which has
characteristic properties and a crystal structure different from those of the pure metals or

”1

the terminal solid solution. Intermetallic compound formation can be a complicating
factor in ASV. When an intermetallic compound forms, the stripping peaks for the
constituent metals can be attenuated, shifted or enhanced.2 Intermetallic compound
formation is known to cause peak suppression and multiple stripping peaks due to the

formation of a new phase that may not oxidize at the same potentials as the individual

metals. Some of the more common interrnetallics are Ag-Cd, Ag-Zn, Cu-Cd, and Cu-

,4
Zn.3

A large amount of recent research on the formation of interrnetallic compounds in
ASV has focused on the Cu-Zn complex.5'9 This is because of the co-existence of these
metals in a diverse number of environmental samples and the fact that the response of
each signiﬁcantly affects that of the other. The simultaneous deposition of Cu and Zn
results in the formation of three compounds, CuZn, CuZn2 and CuZn3, at least with an
HMDE. These compounds oxidize near the E°' for Cu, resulting in an enhanced Cu peak
and a diminished Zn peak.10 When a Hg—ﬁlm electrode and a Cu(II)/Zn(II) ratio of 2.0

were used, the Zn peak completely diminished.ll The formation of CuZn (insoluble) and

123

 

CuZn2 (soluble) compounds were also reported for the simultaneous deposition of Cu and
Zn at a Hg thin-ﬁlm electrode. The electrodeposition potential of - 0.85 V (vs. SCE) was
determined as the potential for the reduction of Cu in presence of Zn.8 In this Chapter, we
have used the Cu-Zn intermetallic complex to illustrate the challenges and proposed
technical solutions for the anodic stripping voltammetric detection of these metal ions

using diamond thin-ﬁlm electrodes.

Increasing the Peak Currents and Reducing the Peak Widths in the

Voltammograms. The activity of metal deposit depends not only on the surface

coverage but also on the interaction of the metal with the electrode material. The quality
of the DPASV measurement could be improved if the relatively large peak widths seen in
most of our presented data were reduced. There are several causes of the broad peaks.
First, if the metal deposits are not of a small nominal diameter with a narrow size
distribution, then the stripping peaks will be broadened due to the wide range of times
required for the metal phase to oxidize and the metal ions to diffuse away. Second, if
there are activity differences across the diamond surface, speciﬁcally variations in the
electrode conductivity, then broadened peaks will result due to variable rates of reaction.
From results presented in Chapter 2, we have observed that such heterogeneity, although
present, is minimized at the high doping level of the ﬁlms being used in this work.
Further, it is speculated that if the metal deposits are well dispersed over the electrode
surface and of small nominal diameter, then the probability of a physical/chemical
interaction of the surface ad—atoms is reduced, thereby reducing intermetallic complex

formation.

124

 

 

In this Chapter, we demonstrate the use of pulsed galvanostatic deposition to
produce metal deposits that are nominally smaller in size and more unifome dispersed
over the diamond surface than what is achieved using either constant potential or constant
current deposition.12 Pulsed galvanostatic deposition produces highly dispersed metal
deposits of a small nominal diameter, which in turn leads to relatively sharp and narrow
stripping peaks. An advantage of the galvanostatic method is that the metal nucleation
and growth can be forced to take place at a pre-determined rate.13 This Chapter includes:
(a) a comparison of differential pulse stripping voltammetric i-E curves for Cu and Zn
with the metal deposition accomplished by pulsed current, constant current and constant
potential deposition, (b) scanning electron micrographs showing a comparison of the
metal deposit size and dispersity for the pulsed and constant current deposition methods,
(c) analytical detection ﬁgures of merit for Cu(II) when using pulsed galvanostatic
deposition to preconcentrate the metal, and (d) a study of how interrnetallic compound
formation between Cu and Zn is affected by the use of constant potential and pulsed

current deposition.

6.2. Experimental

Diamond Film Deposition. Boron-doped diamond thin ﬁlms were deposited on p-

type Si(100) substrates ( ~10'3 Q-cm) using a commercial microwave-assisted chemical
vapor deposition (CVD) system.” '5 Details regarding the grth conditions, as well as
the follow-up material characterization, have been provided in Chapter 2. B2H6 was used
for doping at a ﬂow rate of 0.001 sccm (10 ppm). In order to provide a reproducible

hydrophobic surface, the diamond thin-ﬁlms used in this work were all freshly acid-

125

washed and rehydrogenated. Acid washing involved soaking the ﬁlms in warm aqua-
regia and then in warm 30% hydrogen peroxide for 30 min each. The ﬁlms were

rehydrogenated by exposing them to a hydrogen plasma for 30 min.

Electrochemistry. Most of the electrochemical measurements were made using either

a CHI 650a or a CHI 900 computerized potentiostat (CH Instruments Inc., Austin, TX)
and a single compartment, three-electrode glass cell placed inside a Faraday cage for
shielding. The diamond working electrode was pressed against a Viton® o-ring and
clamped to the bottom of the cell. The backside of the cleaned substrate was scratched
with a carbon rod for ohmic contact with a metal plate current collector. A graphite rod
Was used as the counter electrode and a home-made Ag/AgCl electrode (4 M KCl) served
as the reference (EOAg/Aga = -0.045 V vs. SCE). The geometric area of the working
electrode was 0.163 cm2 (interrnetallic study) and 0.222 cm2 (pulse deposition study). All
measurements, unless otherwise noted, were made in solutions deoxygenated with
nitrogen (99.99%) for at least 2 min initially and then for 3 min after each anodic
Shipping or oxidation cycle. The solutions were blanketed with the inert gas during a
measurement. The electrolyte solution was 0.1 M acetate buffer, pH 5.2. 30 ppm of
calcium acetate and 5 ppm of magnesium acetate were added to the electrolyte solution.
The reference electrode was placed inside a cracked glass capillary, ﬁlled with the acetate
buffer, which served as a double junction. All measurements were made in an unstirred
SOlution. The electrode was left mounted in the electrochemical cell, and covered with
Water, for the duration of the pulse deposition experiments reported. This allowed for the

Same area of the electrode to be used in all the experiments. Typical differential pulse

126

 

voltammetric settings were a (i) 3 min deposition time without stirring, (ii) a 0.5 min
“quiet” time at the deposition potential prior to initiation of the anodic sweep, (iii) a 50
mV pulse height, a 2 mV step height, a 50 msec pulse width, a 100 msec cycle period, a
35 msec sampling time, and (iv) a constant potential of 600 mV for at least 120 s and up

to 300 s after completion of the anodic sweep to oxidize all metal deposits.

Chemicals. All solutions were prepared with ultra-pure water (>17 MQ-cm, deionized

and passed over activated carbon) from an E-pure water puriﬁcation system (Bamstead).
Silver nitrate (Fisher), cadmium nitrate (Aldrich), zinc acetate (Aldrich), cupric nitrate
(Aldrich) and lead nitrate (Aldrich) were reagent grade quality used without additional
Puriﬁcation. 0.1 M acetate buffer was prepared by mixing appropriate amounts of sodium
acetate (99%, Aldrich) and acetic acid (Aldrich) in water. All glassware and nalgene
Storage bottles were cleaned by a three-step procedure: thorough rinsing with an alconox
and ultra-pure water solution, soaking in 1 M hydrochloric and nitric acid for at least 10

min. each, and a ﬁnal ultrapure water rinse. The cleaned item was then dried in an oven at

\55° C.

Physical Characterization of the Metal Deposits. The metal deposits on

diamond were characterized using a JSM-64OOV scanning electron microscope (JEOL
LTD., Tokyo, Japan). Micrographs were recorded using both secondary and
backseattered electrons generated with an accelerating voltage of 20 kV. The presence of
Cu and Zn was veriﬁed by energy dispersive x-ray (EDX) analysis (Noran Instruments

Inc., Middleton, WI) in conjunction with the JSM-6400V. The usual parameters were an

127

integration time of 100 s and a take off angle of 30°. Particle analysis was performed with

the AnalySIS imaging software (Software Imaging System Corp., Lakewood, CO).

Scanning electron micrographs and energy dispersive spectra were recorded for
Cu and or Zn deposited onto microcrystalline diamond for 3 min at a deposition potential
of -1.3 V. The applied potential was cathodic of both E°' values so that the reduction
reactions occurred at a mass-transfer limited rate. The solutions were deoxygenated for 3
min in the electrochemical cell prior to the application of the deposition potential. Metal

deposition was accomplished using solutions of different Cu(II) and Zn(II) concentration.

Scanning electron micrographs of Cu deposited from a 10 ppm solution of Cu(II)
in pH 5.2 acetate buffer on nanocrystalline diamond ﬁlms were also acquired after
galvanostatic deposition at 15 mA/cm2 using the constant current and pulsed current
deposition conditions. The solutions were degassed in the electrochemical cell for 3 min

prior to the deposition.

Pulsed Galvanostatic Deposition. The hardware for the pulsed current deposition

consisted of a homemade galvanostat, conﬁgured from an OMNI 90 potentiostat
(Cypress Systems, Inc., Lawrence, KS), that was interfaced with a PC1-5401E controller
and data acquisition board (National Instruments, Austin, TX). The pulse generation was
computer-controlled using National Instruments LabVIEW software (version 5.1 or 6.0).
The scheme of the electrical set-up used has been described by Countanceau et al. and it

is shown in Figure 6.1.16 A resistor was placed in between the working and the reference

128

 

output leads from the potentiostat in order to generate the desired current. The current
was generated as a function of the applied potential. Cu was galvanostatically deposited
from a 100 ppb Cu(II) solution by varying the pulse conditions (current density, pulse
width, pulse number and duty cycle). The metal was then oxidized potentiostatically by
applying a DPASV potential scan, in most cases, from -300 to 300 mV. An Ag/AgCl
reference electrode kept in a cracked capillary junction containing the electrolyte was
introduced into the cell during the potentiostatic oxidation step. The reference electrode
was removed from the cell during the galvanostatic step to minimize short-term potential

ﬂuctuations and instabilities.

 

-—~ Computer

 

 

 

 

ontrol Input
SoCE TWE l?”

Resistor

BR Diamond Electrode

Pulse
Generator

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6.1 The electrical set-up used for the galvanostatic pulse electrodeposition.
(Adapted from Reference 15). CE — counter electrode, WE — working electrode and Ref —
reference electrode.

The current density, pulse width, pulse number and duty cycle were

systematically investigated for their effect on minimizing the stripping peak widths and

129

 

maximizing the peak amplitude. The goal was to obtain intense and narrow peak shapes
and to maximize the ratio ip°"/FWHM . After some study, the optimized pulse conditions
were found to be a 0.5 5 width with a 0.5 s delay time between each pulse, and a current
density of 15 mA/cmz. The deposition period was 180 s with a total of 180 pulses. The

effective deposition time was therefore 90 s.

6.3 Results and Discussion

6.3.] Intermetallic Compound Formation

DPAS V Interference Studies of Binary Metal Ion Solutions. These

experiments revealed at what solution concentrations of the constituent metal ions
interrnetallic compound formation is most problematic at diamond electrodes, if at all.
The possibility of the formation of Cu-Zn complexes was investigated by recording the
stripping voltammetric i-E curves for a solution containing a ﬁxed concentration of

Cu(II) and as a function of the concentration of added Zn(II), and vice-versa.

Zn-Cu binary solutions. Figure 6.2 shows the background-subtracted differential pulse
stripping voltammetric i-E curves for Zn and Cu at a boron-doped nanocrystalline
diamond thin-ﬁlm electrode. Cu(II) was added in increments of 20 ppb to a solution
containing 100 ppb Zn(II). The supporting electrolyte was 0.1 M acetate buffer at pH 5.2.
The Zn stripping peak initially appears as a broad curve with a maximum around — 915
mV (bold solid line). Notice that the peaks have shifted below the zero line. This effect is
occasionally observed at low concentrations of Zn(II). At these potentials there is a

substantial amount of charge passed for the hydrogen evolution due to the reduction of

130

 

 

water. This means that the background current associated with hydrogen evolution, is
larger in the absence of Zn. In other words, the deposition of Zn makes the surface less
active for hydrogen evolution. Further, this particular voltammetric i-E curve for the 100
ppb Zn(II) solution is less well deﬁned than was normally seen for the same solution
concentration of Zn(II) in previous work. It is possible that there was some Cu(II)
contamination in the system. Suppression of the Zn stripping peak current by co-

deposited Cu is widely known.3

With the addition of Cu(II), there is an increase in the area under the Zn stripping
peak and it transitions to a well deﬁned peak shape. Although enhancement in the peak
current and charge occur with each added aliquot of Cu(II), the maximum enhancement
occurs after the addition of the ﬁrst aliquot. There are two possible explanations for the
observation. Some of the deposited Cu may be oxidizing off at the potentials of Zn. This
could indicate the formation of an electroactive interrnetallic phase between Cu and Zn
with E°’ close to that of Zn, at these experimental conditions. However, it is also possible
that Cu is deposited preferentially as opposed to Zn and the deposition of Zn takes place
on the already formed Cu deposits. The deposition of Zn on Cu may have a larger metal
loading than on the bare diamond surface yielding the enhanced peaks for Zn. From the
literature, thermodynamic considerations and from scanning electron microscopy data,

this is the more plausible explanation for enhancement of the Zn peak.4’ ‘7

The Cu stripping peak between -200 to 100 mV was also studied as a function of

the added Zn(II). With only the Zn(II) initially present in the solution, the i-E curve for

131

 

Cu is ﬂat and featureless (no Cu phase formed). After addition of 20 ppb Cu(II) to the
100 ppb Zn(II) solution, two Cu stripping peaks initially appear at -5 mV and - 105 mV.
With increasing amounts of Cu(II), the peak at — 105 mV grows in intensity and the peak
at —5 mV somewhat merges into and broadens this peak. A possible reason for the
observation of the dual peaks could be the oxidation of Cu from two energetically
different sites on the diamond surface. Another possibility is that the peak at -5 mV is a
result of the oxidation of a Cu-Zn complex that was formed between the two metals. This
theory is strengthened by the fact that there is a limited amount of Zn(II) initially present
in the solution and this peak does not track the solution concentration of Cu(II). When

only Cu(II) is present in solution, it was uncommon to observe the dual/multiple peaks.

Cu-Zn binary solutions. Figure 6.3 shows the background-subtracted differential pulse
stripping voltammetric i-E curves for Zn and Cu recorded to test the effect of making
standard additions of Zn(II) to a solution containing Cu(II). The area under the i-E curve
with no Zn(II) present represents the charge generated due to hydrogen evolution. Cu is
known to catalyze this reaction and shift the hydrogen evolution positive.18 On the
addition of aliquots of Zn(II) to the solution, the total area under the Zn peak somewhat
increased. However, as observed from the Figure, there is no clear trend of increase of
current magnitude for the Zn peak with Zn(II) concentration. This is probably due to

some of the Zn interacting with the Cu and oxidizing at the potentials close to that for Cu.

132

 

 

.0
...-L

 

 

  
 
   
     

9
0

Current (11A)

 

 

 

 

 

 

Potential (mV vs. Ag/AgCl)

Figure 6.2 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for Cu
and Zn. Standard additions of solutions containing Cu(II) were made in increments of 20
ppb to a solution of 100 ppb Zn(II) in pH 5.2 acetate buffer. Deposition potential = - 1300

mV. The DPASV conditions are as described in the experimental section. Electrode
geometric area = 0.163 cmz.

 

 

       

0.10 0.3 E T
. Zn(II) Cu(H) 1:1; . _\:.\‘ }T 20 to 60 ppb Zn(II)
a '|
l- b ' 3 no Zn(II)
A 015 100 ppb Zn(ll) \\' 100 ppb Zn(II)
<5; 02 V 4/‘/
v
... 0.00
t:
E
0%
U 0.1
-0.10 -
0.0 - 1 I I J
-0.15 4

 

 

 

 

 

 

Potential (mV vs. Ag/AgCl)

Figure 6.3 Differential pulse anodic stripping voltammetric (DPASV) i-E curves for Cu
and Zn. Standard additions of solutions containing Zn(II) were made in increments of 20
ppb to a solution of 100 ppb Cu(H) in pH 5.2 acetate buffer. Deposition potential = -

1300 mV. The DPASV conditions are as described in the Experimental section. Electrode
geometric area = 0.163 cmz.

133

 

The intensity of the Cu peak at about — 100 mV is slightly enhanced with the addition of
Zn(II); however, it decreases at concentrations of 80 ppb Zn(II) and higher. A sister peak
at about — 31 mV gets prominent at higher concentrations of Zn(II). It is possible that this

peak is formed due to the oxidation of a product of the Cu-Zn interaction.lo

Overall, the co-deposition of Cu and Zn signiﬁcantly affected the response of the
other. The addition of a metal ion to the solution in the presence of the other resulted in
an increase in peak charge at or near the E°' of the added metal ion, in both cases.
Presence of Cu(II) shifted the hydrogen evolution positive, making the analysis of Zn(II)
challenging. Evidence of Cu and Zn forming a new phase that that undergoes oxidation
near the E°' for Cu was observed. A generalized standard addition procedure could be
useful for the determination of the metal ion of interest nullifying matrix effects for the

other. I 9

The oxidation of a metal system depends on various factors, information on which
is available in phase diagrams. The electronic work function, (1), can serve as an important

1.20 Brainina and

quantitative characteristic of the electrochemical behavior of a meta
Neyman have compared the electronic work function of metals and the interaction type
(phase diagrams) of binary systems.21 Table 6.1 (adapted from their book) shows an
arrangement of selected metals in descending order of the electronic work ﬁrnction (1)
horizontally and general data on the character of the phase diagrams in each box that

corresponds conventionally to the system E2 — E1. The table speciﬁes groups of elements

forming continuous solid solutions and intermediate phases and chemical compounds as

134

 

 

 

well as the number and kinds of voltammetric peaks expected. According to the phase
diagrams, Cu is most prone to the formation of interrnetallic compounds, and this is the
case both from our experimental results and also from reported literature. As predicted
from the table, peaks of Zn(II) are experimentally observed to be suppressed in the
presence of Cu(II). In contrast, Pb-Zn, Pb-Cd, Cd-Zn and Zn-Ag are systems without
mutual interaction, both from the table as well as experimental results. Overall, it may be
inferred that for some binary metal systems the character of interaction and the properties
of the phases being formed are determined by the values of the electronic work functions

of the corresponding metals.

There are several ways to minimize interrnetallic compound formation. One
approach is the preferential formation of another interrnetallic compound.4 This approach
is based on the addition of a “thir ” element that forms a more stable interrnetallic
compound with one component of the binary system. For example, adding an excess of
Ga(III) to the sample solution usually solves the Cu-Zn interrnetallic problem, as the Cu-
Ga complex has a larger formation constant than Cu-Zn and the Zn(II) solution
concentration can then be determined without any interference.22 Another approach to
solving this problem was proposed by Heineman and coworkers."1 Using a twin electrode
thin layer cell, they selectively plated Cu at one electrode and measured Zn(II) by ASV at
the other electrode. In another report, the addition of CN' and Triton X-100 was shown to
suppress the formation of the Cu-Zn intermetallic compound in ASV making it possible

to determine trace levels of Zn(II) in the presence of an excess of Cu(II).7

135

 

 

Table 6.1 Work functions, Kinds of Interactions in Binary Systems and Types of Anodic
Voltammetric i-E curves of Electrochemically Formed Deposits. (Adapted from Ref 21)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(b 4.55 4.30 4.30 4.24 3.97
E1 Cu Ag Zn Pb Cd
2 3 2 3 Cu
lilillllllﬂmns. solid solution ’
1 A
9
intermetalllc compound '
1 1 Zn
system without interaction 1
r Pb
Kinds of anodic voltammograms:
l - peaks ofEl and E2 Cd
2 — one additional peak
3 - peak of E1 disappears while E2 is in excess
4 - several additional peaks E 2

 

 

 

Our approach, at present, is to limit the surface concentration of the metals by
maintaining nominally small and highly dispersed deposits over the electrode surface.
This is discussed below. Wise et al. studied conditions to avoid intermetallic compound
formation between Cu and Zn at Hg electrodes. The use of shortened preconcentration
times helped reduce errors for Cu(II) and Zn(II) concentrations in the ppb range.3 This
may be attributed to reduced interaction between the metals as decreased amounts of

metal would get deposited at shorter preconcentration times.

136

 

 

6.3.2 Characterization of the Cu-Zn metals deposited on diamond

Figure 6.4 and 6.5 show scanning electron micrographs and energy dispersive x-
ray spectra recorded for Cu/Zn metal deposits on a microcrystalline diamond thin-ﬁlm
electrode. The deposit was formed for 3 min at -l300 mV. Deposition was performed
from solutions of Cu(II) and or Zn(II) in the following proportions: (A) 100 ppm Cu(II);
(B) 100 ppm Cu(II) + 10 ppm Zn(II); (C) 100 ppm Zn(II); and (D) 100 ppm Zn(II) + 10
ppm Cu(II). The purposes for conducting these experiments at varying concentration
ratios of the two metal ions were: (i) to ascertain if there were any differences in the
shape, size, density and location of the metal deposited on the diamond surface,
individually and from their mixtures; and (ii) to determine if one or both the metals were

being deposited on the diamond surface at various experimental conditions.

As can be seen from the micrographs in Figure 6.4, the metal particles are well
dispersed and present on both the grains and the boundaries in all four cases. The Cu-only
deposits are large in number, uniform in size and well dispersed. The average particle
size is ~85 nm and the particle density is ~ 2.3 x 108 cm'z. The deposits formed from the
10:1 Cu(II):Zn(II) ratio are aggregated and several new particles appear to form on
existing deposits. The average particle size is ~125 nm and the particle density was found
to be ~ 3.4 x 108 cm'z. It is possible that Cu deposits ﬁrst because its E°' is ahnost 1000
mV positive of the applied potential, and the Zn deposits on the Cu more easily than on
the bare diamond. The deposits of only Zn are uniform in size, evenly distributed and
possess a relatively small nominal diameter of about 100 nm and the particle density is ~

1.6 x 107 cm'z. In comparison, particles deposited from the 10:1 Zn(II):Cu(II) ratio are

137

 

greater in number, evenly distributed and of nominal diameter similar to the deposits of
pure Zn (~ 100 nm). The particle density was found to be ~ 5.8 x 107 cm'z. There is

evidence of new metal formation on existing metal phases.

It appears that Cu deposits form in large number over the entire diamond surface
whereas Zn deposits are fewer in number and more isolated. A greater number of

particles are expected to form at a faster rate for Cu than for Zn, due to the differences in

 

applied overpotential. When low concentrations of Zn(II) are present in a Cu(II) solution,

it is likely that a new metallic phase is getting formed because of the larger particle i-

 

density and clustered nature of the resultant particles. Further, Zn appears to be L

preferentially deposited on Cu deposits.

 

138

 

Figure 6.4 SEM images (secondary electron) of Cu and or Zn metal deposits on a
microcrystalline diamond electrode formed at -1.3 V for 3 min from varying ratios of the
metal ions in solution: (A) 100 ppm Cu(II); (B) 100 ppm Cu(II) + 10 ppm Zn(H); (C) 100
ppm Zn(II); and (D) 100 ppm Zn(II) + 10 ppm Cu(II).

Figures 6.5 A and B show the corresponding energy dispersive spectra for the
metal deposits in images B and D of the previous ﬁgure. All experimental conditions
(e.g., image magnitude, accelerating voltage etc.) were the same. The C Koz (0.275 KeV),
0 K0: (0.531 KeV), Si Ka(1.747 KeV), Cu 1111 (0.933 KeV), Cu kozl (8.037 KeV) and Zn
al and (12 around 1.0117 KeV, are observed in both spectra.20 The relative intensities,

however, are much greater for the solution with larger proportion of Cu(II). This is

139

 

 

consistent with more metal coverage and with the SEM images of surfaces with metal

deposited from Cu(II):Zn(II) containing solutions with a 10:1 ratio (6.4.B). The EDX

results prove that both metals were deposited on diamond from their binary metal ion

 

 

 

 

 

 

solutions.
6000'
A
5000'
4000'
3000< L
‘ Si
2000 L Cu, II
Zn ll
1000. o v. ., Zn
JHbMTW«TTMWW%WHWWWW_Zn %
0.0 1.0 2.0 6.0 40 5.0 6.0 710 8.0
I , I-
1100iII f 1 I3
9001 l '
7001 1 lo I I
l I Z
L ' Cu,_Zn I...) I; n
500 J II I.II ”It. ﬁt.
. l .i d‘iI—IJ Viv" II
300. 1} :1 n.1,, .
I ' II I 'Ir 1" Anﬁhbfl “Atlantis-Lg a,“
100 I. . """’~'\"'r~»--~r-«---»-.s~......... . .. -9”
L T ‘7.—‘ fit . .__. . ___..._ __-- --- .__..w M :
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Figure 6.5 Energy dispersive x-ray analysis spectrum for Cu and or Zn deposited on
microcrystalline diamond electrodes at -1300 mV for 3 min at metal ion concentrations
of: (A) 100 ppm Cu(II) + 10 ppm Zn(II) and (B) 100 ppm Zn(II) + 10 ppm Cu(II).

140

 

 

6.3.3 Pulsed Galvanostatic Deposition

Figure 6.6 shows a comparison of the differential pulse voltammetric i-E curves
for the stripping of metal phases deposits formed from 100 ppb Cu(II) in pH 5.2 acetate
buffer using four different methods: (a) constant current deposition at 15 mA/cm2 for 905,
(b) pulsed current deposition at 15 mA/cm2 with 0.5 s pulses at a 50 % duty cycle for a
total of 1805, (c) constant potential deposition at - 400 mV for 90s, and (d) pulsed
potential deposition at - 400 mV with 0.5 s pulses at 50 % duty cycle for a total of 180 s.
The charge passed during the galvanostatic deposition was kept constant at 0.3 C. The
pulsed deposition parameters used in this comparison were found to be optimum as

shown in the next section.

1.5

 

o...
0..
.0
0..
.......

Current (uA)

.
“".
.’

~
~
.. .—
‘o O. ’-
O-oo-oo-O.’.. .v'
.

 

 

 

 

'-200 A -100 A o A 100
Potential (mV vs. Ag/AgCl)

Figure 6. 6 Differential pulse stripping voltammetric 1- -E curves for Cu on nanocrystalline
diamond deposited at (A) constant current deposition at 15 mA/cm2 for 905, (B) pulsed
current deposition at 15 mA/cmz, 0. 5 s pulses at a 50% duty cycle for a total time of 180
s, (C) constant potential step deposition at - 400 mV for 90S, and (D) pulsed potential
deposition at - 400 mV with 0.5 s pulses at a 50% duty cycle for a total time of 180 s.

141

 

The two largest currents are seen for the metal deposits formed under
galvanostatic conditions. This results ﬁom the fact that more charge is passed during the
galvanostatic than the potentiostatic deposition. The most well deﬁned and narrow peak
is seen for the pulsed galvanostatic method. This is attributed to the formation of well
dispersed metal deposits during the pulse sequence due to avoidance of inter-particle
coupling.23 This occurs because of the replenishment of the individual metal ion
depletion layers around each of the growing metal particles during the pulse “off” stage.
This prevents the metal centers from coalescing and forming large area deposit clusters.
Pulsed electrodeposition approaches are known to offer advantages in terms of controlled
particle size, stronger adhesion, more uniform electrodeposition etc.24 Some of the
particles are more closely spaced than the diffusion layer thickness, as evidenced by SEM
given below. Therefore, hemispherical diffusion is maintained at some particles while
semi-inﬁnite linear diffusion occurs at other closely spaced particles and clusters. In a
separate set of potential step experiments, the resultant charge passed during a potential
step to -400 mV for 180 s was measured for 100 ppb Cu(H). The magnitude of the charge
was 7.5x10’5 C. This is much less than the charge passed during the galvanostatic
deposition and is the reason for the low signal intensities seen in Figures 6.6.C and D.
The above results suggest that the use of pulsed current deposition is favorable for the

DPASV detection of metals with diamond.

142

 

Optimization of Various Pulse Parameters. There are several parameters that can

be varied to optimize the metal deposit: pulse current, pulse time, duty cycle and pulse
ntunber.

a) Optimization of the current density. Choi et al. found that the pulsed current
deposition of Pt on carbon electrodes to be optimum at 25 mA/cmz. At higher current
densities, the performance dropped sharply because of the growth of dendritic crystals
and the displacement of deposited metal particles due to hydrogen generation.24
However, at lower current densities the nucleation and grth mechanism can be charge
controlled where the grth of the already formed nuclei is preferred whereas at higher
current densities the mechanism is likely mass transport controlled. It seems that an '
optimum applied current density would be a trade—off between obtaining a large metal
loading on the electrode versus the other competing reactions like hydrogen evolution
becoming dominant. Current densities ranging from 2 to 21 mA/cm2 were used with a

deposition time of 180 s and l-s pulse length at a 50% duty cycle.

Natter and Hempelmann deposited Au from a commercial sulﬁte bath onto
stainless steel and titanium electrodes, and tested current densities in the range from 50 to
500 mA/cmz. They observed that the smallest crystallite size of 12 nm was formed at 50
mA/cmz.25 At higher current densities, powder formation occurred on the electrode
surface. Figure 6.7 shows a few representative differential pulse stripping voltammetric i-
E curves for Cu deposited by pulsed galvanostatic deposition at different current

densities. The peak intensity initially increases with applied current density as expected,

until the highest current density is reached 15 mA/cmz, at which point the peak current

143

 

decreases some. Analyses of the peak charge reveals that, the decrease in peak current
may be due to the dislodging of metal deposits by hydrogen evolution reaction as gas
bubbles were observed at the electrode surface at current densities higher than 15
mA/cmz. The resultant electrode potential measured at the working electrode, at these
current densities, was between —2.7 to —3.2 V versus an Ag/AgCl reference electrode.
Dual peaks at ~ -100 and 100 mV are observed. These may be due to the oxidation of
metal deposited at two different sites on the electrode surface. The current density of 15
mA/cm2 (total charge passed: 0.3 C) was accepted as optimum based on the peak shape

and intensity.

b) Optimization of duty cycle. The duty cycle is the percentage of the time the pulse is
in the “on” state (ton) relative to the total time that also includes a dead time. The duty
cycle can be an important factor because the properties of metal deposits can be
inﬂuenced by both the on-time, during which nuclei formation and grth occur, and the
off—time or dead time, during which replenishment of the depletion layer around each
particle(s) occurs.26 Figure 6.8.A shows representative differential pulse stripping
voltammetric i-E curves for the stripping of Cu from deposits formed as a function of the
duty cycle. The deposition charge passed was kept constant at 0.3 C. The total time of
deposition was 120 s (75 % duty cycle, 90 l-s pulses), 180 s (50 % duty cycle, 90 1-s
pulses) and 360 s (25 % duty cycle, 90 l-s on pulses). As can be observed from Figures
6.8.A and 6.8.B, the peak parameter (current/FWHM) increases with decreasing duty

cycle. The optimum duty cycle from the few tests was found to be 25 %; however 50 %

144

 

 

was used because the total analysis time for a 25 % duty cycle is 360 s and this is too

long for practice.

 

 

 

 

c ..‘—_
D.’ '
1.5 " o .
o . E
r ,' C
A ' /"\
< O .0 ‘.
3; 1.0 — ,' .’ B ,
H 0 ’. ‘\
a ' o. a
E P .IOOJ'f I ‘0 ...,
0 ~ o’,o’.. .. :o” ‘\‘
o ’ \
0.5 '- 0"," A \“ .............. 0.0.:
u
a II
D. ."
I"
0.0 l 4 l 1 l
-200 0 200

Potential (mV vs. Ag/AgCl)

Figure 6.7 Differential pulse anodic stripping voltammetric i-E curves for Cu from a
nanocrystalline diamond surface using pulsed galvanostatic deposition at (A) 2.7, (B) 5.5,
(C) 12, (D) 15, and (E) 21 mA/cmz. The solution was 100 ppb Cu(II) in 0.1 M acetate
buffer, pH 5.2. Total deposition time = 180 s. Pulse width = 1 s. Duty cycle = 50%.
Arrow indicates the optimum current density.

According to Natter and Hempelmann, for a constant ton-time and current density,
the crystallite size increases for long tofrtimes due to Ostwald ripening.25 Ostwald
ripening is a spontaneous process that is said to be occurring when initially formed

smaller crystals in a system merge to form larger crystals. This happens because larger

crystals are more energetically favored than smaller crystals.

145

 

Current (11A)

This is contrary to what we infer from the voltammetric curves. They were found
to be the broadest for lowest ton- (75 %) duty cycle, consistent with large particle sizes. A
reason for this disagreement could be that Natter and Hempelmann tested the variance of
toﬂ‘ in the milli-second regime. The variation of the electrode potential during the current
pulse leads to a non-faradaic current due to the charging of the double layer, so that only
a part of the overall applied current is involved in the electrochemical reaction.16 The
charging of the double layer is typically in the milli-second regime. It is possible that the
duration of pulses in their experiments were too short for the bulk of the applied current

to be used for metal deposition.

 

 

 

 

 

 

 

A B
0010
9 . ° A
'25 0.009~
1 .
v 0033
2 .
§ 00071 ,B
E .
0.(XB«
.é‘ .
3 0006
. 0004- '
0.0 ... ‘ 1 1 '— r ' I ' I P l ' I ' l
-200 0 200 20 30 40 50 60 70 80

Potential mV vs. Ag/AgCl Duty CYCIC (%)

Figure 6.8 Differential pulse anodic stripping voltammetric i-E curves for Cu from a
nanocrystalline diamond surface using pulsed galvanostatic deposition at effective
deposition times of (A) 360$ (25 % duty cycle, 90, l-s on pulses), (B) 1808 (50 % duty,
90, l-s on pulses), and (C) 1205 (75 % duty, 90, l-s on pulses). The solution was 100 ppb
Cu(II) in 0.1 M acetate buffer, pH 5.2. Current density = 15 mA/cmz.

146

 

u-. I.“

 

c) Optimization of pulse width. Figure 6.9 shows representative differential pulse
stripping voltammetric i-E curves for Cu from a nanocrystalline diamond surface as a
function of the pulse width (ton). The total galvanostatic charge passed was kept constant
at 0.3 C. The pulse widths tested were: (A) 3.0, (B) 1.5, (C) 1.0, (D) 0.5, (E) 0.25, and (F)
0.1 s. The plot of peak current/FWHM (uA/mV) vs. the pulse width shows that at 0.5 3
pulse width, this ratio is a maximum. The peak current/FWHM (uA/mV) parameter was
chosen as an indicator of peak shape. Intense narrow voltammetric peaks are desired as
this could enable (i) quantitation using peak currents rather than peak charge, and (ii) the

simultaneous assay of metal ions having relatively close E°' values.

 

 

 

 

 

 

 

90mm 0
A E
ii i -
v V B °
E 2 0.006- A
5 .- E .C
‘3 F
5 0004- '
E ' -
U I ' EU " T ‘ I ' I ' I ' I ﬂ
1 . 4 . - 0.0 0.5 1.0 1.5 2.0 2.5 3.0
-1m 0 1(1)
Pulse width (8)

Potential (mV vs. Ag/AgCl)

Figure 6.9 Differential pulse anodic stripping voltammetric i-E curves for Cu from a
nanocrystalline diamond surface using pulsed galvanostatic deposition at pulse widths of
(A) 1.5, (B) 3, (C) 0.5, (D) 1.0, (E) 0.1, and (F) 0.25 s. The solution was 100 ppb Cu(II)
in 0.1 M acetate buffer, pH 5.2. Current density = 15 mA/cmz. Duty cycle = 50%. Total
deposition time = 180 s.

147

 

Scanning Electron Micrographs of Cu Deposits on Nanocrystalline
Diamond. Scanning electron micrographs of Cu deposited from a 10 ppm solution of
Cu(II) in 0.1 M acetate buffer, pH 5.2, at 15 mA/cm2 using both the constant and pulsed
conditions are presented. The deposition charge passed was constant at 0.3 C. The ﬁlms
used were acid-washed and rehydrogenated boron-doped nanocrystalline diamond thin-
ﬁlms. Figure 6.10 shows the back-scattered electron images for (A) a 90 s constant
current deposition and (B) pulsed current deposition using 0.5 s pulses for 180 s (180
pulses). Analysis of the Cu deposits on at least two nanocrystalline ﬁlms at four different
spots each ﬁlm for each set of deposition conditions revealed mean particle diameters of
109 i 87 nm and 34 i- 60 nm, respectively, for the constant and pulsed current deposition.

The particle density was 1.5><109 and 1.1X 10Io cm‘z, respectively, for the constant current

and pulsed deposition conditions.

   

Figure 6.10 Scanning electron micrographs (back scattered images) of Cu deposited on
nanocrystalline diamond at 15 mA/cm2 in the form of a (A) constant current applied for
90 s and (B) a pulsed current applied for 180 s using 0.5 s pulse width. Other conditions
have been provided in the experimental section.

148

The fractional surface coverage is 0.1386 and 0.1058 respectively, for the constant
current and pulsed deposition conditions. These data show that in comparison to the
constant current deposition, the pulsed deposition yields well dispersed particles of
smaller nominal diameter and lower metal surface coverage. Although, the particle size
deviation was large using both the deposition protocols, it was larger for the constant

current approach.

Analytical Detection Studies. Studies were performed to determine if there was a

concentration dependence on the response for Cu(II) using the pulse galvanostatic
deposition method. Figure 6.11.A depicts background-subtracted differential pulse
stripping voltammetric i-E curves for deposits from Cu(II) solutions of increasing
concentrations ranging from 1 to 1000 ppb. The peak current and charge increase with
increasing solution concentration. The Ep0x shifts positive with increasing solution
concentration due to the greater surface coverage. The baseline peak widths are ca. 150
mV and the FWHM values are ca. 75 mV. Figure 6.11.B shows a plot of peak charge vs.

the Cu(II) concentration. The S/N for the 1 ppb run is 10/1.

149

 

 

 

 

 

 

 

 

 

 

 

 

02
8- , _ 50t01000ppb
0 0/‘—’/ (Insetl and 10 ppb) 18:
i I 16- .II
5 14‘ r2=0.992
{0‘2 -200 0 a
A
< 612:
3' 1104
E; 4* T; 8‘
no i
g i / a 6-i _.
O 2* I/8 6 4: II
,2.’, d ‘ \. -\I 2: ,r'.
ifon I- _ 0‘ r. 1 , I f I f I I I
0'- ‘ - :' O 200 4m w) 800 101]
L A * Concentration (ppb)
’1w .5 0 75

Potential (mV vs. Ag/AgCl)

Figure 6.11 (A) Background-subtracted differential pulse stripping voltammetric i-E
curves for metal deposits formed from 50, 100, 250, 500 and 1000 ppb of Cu(II) solution
in 0.1 M acetate buffer, pH 5.2. The inset shows the stripping voltammetric i-E curves for
the 1 ppb (dashed line) and 10 ppb Cu(II) solutions.(B) Plot of charge vs. concentration
for the data in 6.11.A. The integration range was 100 mV (potential limits for integration
varied with concentration).

Comparison of Cu-Zn Voltammetric i-E Curves for Pulsed Constant

Current and Constant Potential Deposition Methods. Figure 6.12 shows a

comparison of representative background-subtracted differential pulse anodic stripping
voltammetric i-E curves for 100 ppb Cu(II) and 100 ppb Zn(II) in pH 5.2 acetate buffer
recorded from -1200 mV to 200 mV. Curve 6.12.A. was recorded for metal phase
formation using pulsed current deposition of both Cu(II) and Zn(II) at 15 mA/cm2 with a
0.5 3 pulse width and a 50 % duty cycle for a total time of 180 s, while curve 6.12.B was

recorded for metal phase formation at constant potential at - 1200 mV for 90 s.

150

 

 

 

Current (11A)

 

 

 

n 1

-400 -200 0

m

 

-10001--800 ‘ -600
Potential (mV vs. Ag/AgCl)

Figure 6.12 Comparison of representative background subtracted differential pulse
anodic voltammetric i-E curves for 100 ppb Cu(II) and 100 ppb Zn(II) in 0.1 M acetate
buffer, H 5.2. Preconcentration was performed by (A) pulsed current deposition at 15
mA/cm with a 0.5 5 pulse width and a 50% duty cycle and a total time of deposition of
180 s, and (B) constant potential deposition at - 1200 mV for 90 s.

Discemible peaks for Cu (-90 mV) and Zn (-1000 mV) are observed for the pulsed
current deposition. In contrast, only a weak Zn peak is seen and the curve is shifted below
the zero line for the constant potential deposition. This could be attributed to the
difference in charge for the hydrogen evolution reaction on the bare electrode versus the
electrode containing metal deposits. During the potential scan in regions cathodic of ~ -1
V, a signiﬁcant amount of charge is passed due to hydrogen evolution reaction. These
results indicate that small amounts of metal were deposited even at 90 5 effective time of

application of constant potential in pH 5.2 acetate buffer, however the amount of metal

deposited by galvanostatic deposition for the same effective times was much higher.

151

 

 

6.3.4 Observation of Voltammetric i-E curves of Narrow FWHM

The DPASV procedure used a “cleaning step” in between runs. This step typically
involved poising the working electrode at + 0.6 V for 3 min with the N2 line placed in
close proximity. The N2 purge aid in the solution mixing of metal ions from the vicinity
of the electrode. An unusual observation was occasionally made when solutions were not
purged with N2 in between runs during the “cleaning” step. Figure 6.13 shows
voltammetric i-E curves recorded for a (A) 100 ppb solutions of Ag(I) and (B) 100 ppb
solution of Cd(H), both in 0.1 M acetate buffer, pH 4.5 (with added 30 ppm Ca(II) and 5
ppm Mg(II) salts). In all cases, an N2 purge was n_ot performed during any of the

procedural steps.

The voltammetric i-E curves (solid lines) were ﬁrst recorded after deposition of
the metal ions for 3.5 min at 0 and — 1 V for Ag and Cd respectively. The cell potential
was then poised at 0.6 V for 3 min and the deposition and stripping steps repeated and the
voltammetric i-E curves (dashed lines) recorded. It was observed that the Ag peak shifted
negative by about 20 mV and became narrow (FWHM ~ 20 mV) and relatively intense.
A similar observation was made for Cd, as the peak shifted negative by about 150 mV
and also became narrow (FWHM ~ 45 mV) and relatively intense. The intense narrow
peaks persisted in subsequent scans. The total area was greater in both cases for the

second run (dashed lines).

152

 

 

 

 

 

 

 

0.30 1-41
4 "° . i ‘
025 I..I A 1.2+ B
, .0;
Q 020 z: ‘
1 I II 0.8‘
v .
0.15« 3‘- i
E . I. 0.6
a) :2 ‘
E 0.104 3 g 04‘ .
0.06M, -. . Lea/.7 ..... ¥~
IIIIIIII .. 0.0«I ..........
. . r . r r 1 0.2 ' I ' I I I . T Y 1
100 200 300 400 -1000 -900 4300 -700 60° “500

Figure 6.13 Background-subtracted stripping voltammetric i-E curves recorded for (A)
100 ppb Ag(I) and (B) 100 ppb Cd(II) in 0.1 M acetate buffer, pH 5.2 (with added 30
ppm Ca(II) and 5 ppm Mg(II) salts). The solid lines depict the ﬁrst recorded i-E curves
and the dashed lines depict subsequent i-E curves (no purging with N2 undertaken).

These results suggest that incomplete removal of the remnant metal deposits takes
place when the solution is not purged with a gas in between runs. The fact that the peaks
moved negative and became intense could be explained by the fact that when there were
favorable sites in the form of remnant metal already present on the diamond surface, the
new metal phase formation would take place on these sites. The oxidation of metal from
these sites would require less driving force or overpotential and would also take place
with a more uniform rate than oxidation from the relatively heterogeneous diamond
surface, making the peaks appear intense. It is possible that more metal gets deposited on
the remnant metal centers due to faster charge transfer and or radial ﬂux of depositing
species. This would explain the greater charge under the voltammetric i-E curves (dashed
lines). From chronocoulometric measurements (Chapter 3), the typical ratio of the

Qp‘”‘/Qpred for the metal oxidation and deposition was found to be in the range of 0.65-

0.75 for Ag(I) on diamond. This indicated that all the deposited metal is not oxidized.

153

 

 

 

The above theory was strengthened by the fact that previous electrode response
(denoted by solid lines in Figure 6.13) could be regained by bubbling with N2 and that a
background voltammetric i-E curve recorded after the “cleaning” step without the N2
purge showed an oxidation peak for the metal ion. The same effect was observed when
purging with O2 instead of N2, proving that the observed response may not necessarily be
due to the nature of the gas. Recent results from our laboratory have also shown that

dissolved 02 does not affect the stripping voltammetric response for Ag(I).

Another approach to explaining the observation could be to model the metal phase
formation chemistry on diamond surfaces. The Nemst equation dictates the position of
the peak potential, Ep. Changes in surface chemistry of diamond (e.g., surface oxygen
content) can decrease the concentration of the adsorbed ad-atom. This would shift the
observed Ep negative. Although these results are preliminary, they could open the door
for a new method of obtaining narrower peaks in ASV. Experimental conditions for
reproducibly obtaining the intense narrow ASV peaks observed above need further study.
If the ﬁrst explanation is valid, metals with positive E°' such as Pt, Pd or Au could be
plated onto diamond prior to its use for assaying for metal ions for obtaining narrower

stripping peaks.

154

 

 

6.4 Conclusions

The electroreduction of the metal ions via potential step deposition produces
considerable nucleation and grth of metal deposits. This approach, however, led to
peak broadening, multiple peaks, suppression of peaks etc. during the electrooxidation
step because the deposition process was not controlled. The experiments conducted by
co-depositing the metal ions provided valuable insight into the complexity of this
problem and the possibility of interference by interrnetallic complex formation between
them. It was observed that Cu(II) and Zn(II) signiﬁcantly affected the voltammetric
response for the other. It is possible that Cu-Zn interrnetallic complexes are formed on

diamond.

Scanning electron micrographs of the Cu and Zn deposits demonstrated that the
morphology of the deposits formed when the metals were simultaneously deposited is
different from the morphology of the individually deposited metals. The energy
dispersive x-ray spectra of the deposits showed that both Cu and Zn are deposited from
their mixtures. The difference in morphology could be due to the formation of a complex

between the metals or deposits of Zn forming on the already formed deposits of Cu.

Pulsed galvanostatic deposition lead to highly dispersed Cu deposits of a small
nominal diameter. It is felt that with the further optimization of experimental conditions
this approach could be useful for producing sharp and narrow stripping voltammetric
peaks. Results presented in this Chapter indicated that constant current pulsed deposition

of metal ions was useful in increasing the S/N ratio and improving peak shape, even at

155

 

 

 

effective deposition times of only 90 s. Discemible peaks for 100 ppb Cu(II) and Zn(II)
- were observed after their pulsed current deposition for 90 3, whereas potentiostatic

deposition did not yield such peaks after 90 s deposition time.

A shifting of the voltammetric peaks to negative potentials and improved peak

shapes was observed occasionally, speculated to be caused by remnant metal on the

surface of the diamond. It might be worthwhile to explore this phenomemon ﬁlrther.

156

6.5 References

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

Eckerlin, P.; Kandler, H., Eds. Structure Data of Elements and Intermetallic
Phases (Landolt-Boernstein Numerical Data and Functional Relations in Science
and Technology, Group 111: Crystal and Solid-State Physics, Vol. 6), 1971

Wang, J., Ed. Stripping Analysis: Principles, Instrumentation, and Applications;
VCH Publishers, Inc., 1985

Wise, J. A.; Roston, D. A.; Heineman, W. R. Analytica Chimica Acta 1983, 154,
95-104.

Copeland, T. R.; Osteryoung, R. A.; Skogerboe, R. K. Analytical Chemistry 1974,
46, 2093-2097.

Lastres, E.; De Armas, G.; Catasus, M.; Alpizar, J .; Garcia, L.; Cerda, V.
Electroanalysis 1997, 9, 251-254.

Lau, O.-W.; Cheng, O.-M. Analytica Chimica Acta 1998, 376, 197-207.
Marques, A. L. B.; Chierice, G. O. T alanta 1991, 38, 735-739.
Piccardi, G.; Udisti, R. Analytica Chimica Acta 1987, 202, 151-157.

Scollary, G. R.; Cardwell, T. J .; Cattrall, R. W.; Nan, C. G.; Vicente-Beckett, V.
A.; Hamilton, 1. C.; Roden, S. Electroanalysis 1993, 5, 685-689.

Stromberg, A. G.; Gorodovykh, V. E. Zhurnal Neorganicheskoi Khimii 1963, 8,
2355-2359.

Roston, D. A.; Brooks, E. E.; Heineman, W. R. Analytical Chemistry 1979, 51,
1 728-1732.

Bennett, J. A.; Show, Y.; Swain, G. M. In preparation, Journal of the
Electrochemical Society 2004.

157

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

Gunawardena, G.; Hills, G.; Montenegro, 1.; Schariﬂ<er, B. Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry 1982, 138, 225-239.

Granger, M. C.; Xu, J .; Strojek, J. W.; Swain, G. M. Analytica Chimica Acta
1999, 397, 145-161.

Show, Y.; Witek, M. A.; Sonthalia, P.; Swain, G. M. Chemistry of Materials
2003, 15, 879-888.

Coutanceau, C.; Rakotondrainibe, A. F.; Lima, A.; Gamier, B.; Pronier, S.; Leger,
J. M.; Lamy, C. Journal of Applied Electrochemistry 2004, 34, 61-66.

Shuman, M. S.; Woodward, G. P., Jr. Analytical Chemistry 1976, 48, 1979-1983.

Behzadian, B.; Piron, D. L.; Fan, C.; Lessard, J. International Journal of
Hydrogen Energy 1991, 16, 791-796.

Gerlach, R. W.; Kowalski, B. R.; Lab. Chemometrics,Univ.
Washington,Seattle,WA,USA. :, 1981, pp 23 pp.

Kolb, D. M.; Przasnyski, M.; Gerischer, H. Journal of Electroanalytical
Chemistry and Interfacial Electrochemistry 1974, 54, 25-38.

Brainina, K.; Neyman, E., Eds. Electroanalytical stripping methods; John Wiley
and Sons, 1993

Brett, C. M. A.; Beatriz Quinaz Garcia, M.; Lima, J. L. F. C. Analytica Chimica
Acta 1997, 339, 167-172.

Liu, H.; Penner, R. M. Journal of Physical Chemistry B 2000, 104, 9131-9139.
Choi, K. H.; Kim, H. S.; Lee, T. H. Journal of Power Sources 1998, 75, 230-235.

Natter, H.; Hempelmann, R. Electrochimica Acta 2003, 49, 51-61.

Ibl, N. Surface Technology 1980, 10, 81-104.

158

 

 

CHAPTER 7

SUMMARY

The answers to several research questions were sought at the outset of this effort.
The ﬁrst answer sought was the viability of using boron-doped diamond as an electrode
material for trace metal ion analysis using DPASV. As mentioned, this new electrode
material has very favorable attributes for this electroanalytical application, that include a
large working potential window, a chemically inert surface and elechical conductivity.
Data presented in this dissertation reveals that diamond electrodes are useful for the trace
analysis of dissolved metal ions. Diamond electrodes yielded favorable analytical
detection ﬁgures of merit for ﬁve priority metal ions with very low limits of detection,
and a stable and accurate response. The electrode surface could be regained by a simple
oxidation procedure and the electrodes were used for months at a time with no response
degradation. The analytical performance of diamond compared well with the historically-
used Hg thin-ﬁlm electrode and was better than that for the commonly-used GC

electrode.

The performance of two morphologically and microstructurally different diamond
electrodes was compared. Both types of diamond showed very favorable analytical
detection ﬁgures of merit, although the nanocrystalline ﬁlm had a slight edge in terms of
sensitivity and limit of detection for some metal ions. This observation is due to the
increased particle density (i.e., increased metal phase formation) and smaller particle
diameter during the preconcentration or deposition step. This could be both due to the

smaller feature size, and the fact that a larger fraction of the surface, compared to the

159

 

 

microcrystalline diamond, was electrically conducting (i.e., electrochemically active).
The results obtained with the nanocrystalline diamond and DPASV compared favorably
with results from other more established methods, such as AAS and or ICP-MS. The

method was applied to the analysis of several types of contaminated samples.

An understanding of the metal nucleation and grth phenomenon on diamond is
important for determining of the optimum experimental parameters for preconcentration
in the DPASV method. For instance, when using solid electrodes for ASV, like diamond,
one ideally desires a high distribution of metal particles with a small and uniform particle
size. This should lead to more narrow and intense oxidation or stripping peaks. Therefore,
conditions under which an instantaneous nucleation and growth of the metal deposits
occur should be selected. As discussed in Chapters 5 and 6, the stripping peaks for Cu,
and the other metals, were observed to be rather broad. This is presumably due to the fact
that, at least for Cu, metal phase formation on diamond occurs via a progressive
nucleation and growth mechanism. Results from cyclic voltammetric,
chronoamperometric and EC-AFM measurements presented in Chapter 3 demonstrated
the utility of these methods for both the quantitative and qualitative study of the
mechanisms governing metal phase formation on diamond. Further, the experimental data
revealed that the deposition of metal phases on boron-doped diamond electrodes occurs
by the Volmer-Weber grth mode. No underpotential deposition of metal was observed

indicating a weak metal phase-substrate interaction.

The quality of the DPASV measurement could be improved if the relatively large
peak widths seen in most of our presented data were reduced. In Chapter 6 we

demonstrated the use of pulsed galvanostatic deposition to produce Cu deposits that were

160

 

 

 

nominally smaller in size and more uniformly dispersed over the diamond surface than
was achieved using constant potential or constant current deposition. Resultant stripping
voltammetric i-E curves indicated that the constant current pulsed deposition of metal
ions was useful in increasing the S/N ratio and improving peak shape, even at effective

deposition times of only 90 s.

Intermetallic compound formation can be a complicating factor in ASV. When an
intermetallic compound forms, the stripping peaks for the constituent metals can be
attenuated, shifted or enhanced. Some of the more common interrnetallics are Ag-Cd,
Ag-Zn, Cu-Cd, and Cu-Zn. From results presented in Chapter 6, intermetallic compound
formation on diamond was found to be problematic in the simultaneous assay of Cu(II)
and Zn(II). It was further shown that the selectivity, and the peak shape and intensity
could be improved through the use of pulsed galvanostatic deposition rather than the

conventional constant potential deposition for this measurement.

161

 

iiiiiiiiiiiii