The expansion/contraction of gold microparticles during voltammetrically induced amalgamation leads to mechanical instability

Article abstract of DOI:10.1039/b706810g

The mechanical stability of gold microparticles during anodic stripping voltammetric (ASV) detection over a large range of mercury concentrations was investigated. Mercury was detected at gold microparticles chemically deposited onto glassy carbon microspheres using ASV. Oxidation was observed at 0.5 and 0.8 V vs. SCE. Which peak was observed was dependent on the concentration of mercury and the deposition potential. The formation of the amalgam was of interest. As mercury was deposited for longer time intervals, scanning electron microscopy (SEM) analysis showed the microparticles increasing in size from 0.76 ± 0.03 μm (initial) to 1.51 ± 0.14 μm (Hg²⁺ deposited for 1980 s at 0.35 V) in diameter. In order to ascertain if multiple expansion and contraction cycles damaged the gold microparticles, cyclic voltammetry was used to monitor the amount of gold on the electrode as mercury was deposited and stripped repeatedly. It was seen that the area under the cathodic gold peak decreased with repetitive scans. SEM analysis revealed that the mechanical stress of repetitive deposition and stripping cycles of mercury caused the gold microparticles to fracture, appearing as irregular cuboid crystals rather than as the orderly polycrystallite formations seen initially. Energy dispersive X-ray (EDX) analysis indicated that the composition of the microparticles changed over the course of repetitive deposition and stripping cycles from gold to an Au-Hg amalgam, which may not be in electrical contact with the carbon support. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique.

Full text of DOI:10.1039/b706810g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
The expansion/contraction of gold microparticles during voltammetrically
induced amalgamation leads to mechanical instabilityw
Laura E. Barrosse-Antle,a Lei Xiao,a Gregory G. Wildgoose,a Ronan Baron,a
Chris J. Salter,b Alison Crossleyb and Richard G. Compton*a
Received (in Durham, UK) 8th May 2007, Accepted 13th July 2007
First published as an Advance Article on the web 7th August 2007
DOI: 10.1039/b706810g
The mechanical stability of gold microparticles during anodic stripping voltammetric (ASV)
detection over a large range of mercury concentrations was investigated. Mercury was detected at
gold microparticles chemically deposited onto glassy carbon microspheres using ASV. Oxidation
was observed at 0.5 and 0.8 V vs. SCE. Which peak was observed was dependent on the
concentration of mercury and the deposition potential. The formation of the amalgam was of
interest. As mercury was deposited for longer time intervals, scanning electron microscopy (SEM)
analysis showed the microparticles increasing in size from 0.76 Æ 0.03 mm (initial) to 1.51 Æ
0.14 mm (Hg2+ deposited for 1980 s at 0.35 V) in diameter. In order to ascertain if multiple
expansion and contraction cycles damaged the gold microparticles, cyclic voltammetry was used
to monitor the amount of gold on the electrode as mercury was deposited and stripped
repeatedly. It was seen that the area under the cathodic gold peak decreased with repetitive scans.
SEM analysis revealed that the mechanical stress of repetitive deposition and stripping cycles of
mercury caused the gold microparticles to fracture, appearing as irregular cuboid crystals rather
than as the orderly polycrystallite formations seen initially. Energy dispersive X-ray (EDX)
analysis indicated that the composition of the microparticles changed over the course of repetitive
deposition and stripping cycles from gold to an Au–Hg amalgam, which may not be in electrical
contact with the carbon support.
Thus, accurate and sensitive quantitative analysis is essential
1. Introduction
to monitor the mercury content in foodstuffs and ground-
Mercury is a toxic compound, which nevertheless is exten-
water.
sively employed over a range of applications. Organomercury
Currently, the World Health Organization recommends the
compounds are used in fungicides, antiseptics, preservatives
use of flame atomic absorption spectroscopy (AAS) or cold
vapour AAS for the determination of mercury.1 The wealth of
and pharmaceuticals. Elemental mercury is itself a crucial
component in certain types of instrumentation and industrial
literature on the spectrochemical methods used for mercury
processes.1 Inorganic mercury released into the environment
detection testifies to the methods’ utility.2–6 However, spectro-
by both natural and man-made processes—primarily volcanic
scopic techniques are far from ideal for use in ‘‘field’’ condi-
activity and coal combustion—is able to enter the food chain
tions. Electrochemical detection has come to the fore as a
via the action of methylating bacteria. Most mercury poison-
means of providing a cheap, rapid, facile, and portable method
ing that is not related to occupational hazards results from the
of mercury detection in the field. Methods exploiting these
consumption of carnivorous fish in which organomercury
strengths have been and are in the process of being developed.
compounds, in particular methyl mercury, have been concen-
Frequently these methods employ anodic stripping voltam-
trated due to biomagnification.1
metric techniques (ASV) involving
a preconcentration
Inorganic mercury primarily affects the kidneys, while
organic mercury attacks the nervous system, causing tremors,
muscle weakness, and impaired mental capacity, among other
symptoms.1 Toxic doses of either form result in shock, cardi-
ovascular collapse, renal failure, and gastrointestinal damage.1
or deposition step before detection, particularly for trace
analysis.7
Although many types of electrode substrates have been
investigated, gold electrodes are of particular interest due to
the high affinity of gold towards mercury and thus have been
extensively employed.8–11 The same affinity that enhances the
preconcentration of mercury at the electrode surface, however,
can also be a severe disadvantage. Under high mercury con-
centrations, preconcentration can cause the mercury to diffuse
far enough into the gold electrode that it cannot be completely
stripped away.12 Watson et al. suggest that bulk deposition of
mercury irreversibly changes the nature of the electrode,
forming a gold–mercury alloy electrode. Even repeated
a Physical and Theoretical Chemistry Laboratory, University of
Oxford, South Parks Road, Oxford, UK OX1 3QZ. E-mail:
richard.compton@chemistry.oxford.ac.uk; Fax: +44 (0) 1865 275
410; Tel: +44 (0) 1865 275 413
b Department of Materials, University of Oxford, Parks Road,
Oxford, UK OX1 3PH
w Electronic supplementary information (ESI) available: EDX spec-
trum of gold microparticles on glassy carbon microspheres. See DOI:
10.1039/b706810g
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 2071–2075 | 2071

View Article Online
stripping and cleaning is insufficient to restore the electrode to
its original condition.13
Copenhagen, Denmark) was used as the quasi-reference or
reference electrode, respectively. All potentials measured vs.
silver wire were calibrated against the SCE and are reported vs.
SCE. A scan rate of 50 mV sÀ1 was used unless otherwise
stated. All electrochemical measurements were performed in
0.1 M HNO3.
Therefore studies into the detection of mercury at gold
electrodes have begun to focus on renewable gold electrodes
such as gold films.10,11,14,15 These surfaces can be stripped
away as they become amalgamated and subsequently regen-
erated, avoiding any deleterious change in their electrochemi-
cal properties. Gold microarrays, for which the fouling of a
few individual elements does not significantly impair the
overall electrode function, have also been investigated.16 The
use of gold particles has only just begun to be fully explored
with regard to electrochemical mercury detection; Welch et al.
have described some of the characteristics and potential
applications of amalgamated gold nanoparticles,17 whilst
Keebaugh et al. have used the change in the resistance of gold
nanowires to monitor the concentration of mercury in both
the vapour and solution phase.18 Another group has used
SEM was carried out using a JEOL 6500F instrument. The
compositional analysis of the gold microparticles on the
electrode surface was done by employing wavelength disper-
sive X-ray spectrometry (WDS; Jeol JXA-8800 microprobe).20
2.2 Microparticle fabrication
The gold microparticles were formed on the glassy carbon
microspheres via electroless reduction of gold sodium thio-
sulfate using L-ascorbic acid as the reducing agent. Glassy
carbon microspheres (200 mg) were dispersed in 100 mL of
pure water and sonicated for one minute. The suspension was
then stirred during the addition of gold sodium thiosulfate
(150 mg). The flask was covered in foil, and L-ascorbic acid (90
mg) was added slowly. The solution was stirred for 18 hours
before being vacuum filtered and washed with deionised water.
The resulting Au-mp-GC powder was dried at 100 1C for 2
hours prior to use.
self-assembling gold nanoparticles as
a substrate for
urease, the inhibition of which by mercury(II) is sensed poten-
tiometrically.19
The present work is concerned with understanding the
processes occurring during the detection of mercury at gold
microparticles that are chemically deposited onto glassy car-
bon microspheres (Au-mp-GCs). Specifically, the mechanical
stability of the gold microparticles during the ASV detection
of a large range of mercury concentrations is investigated.
Repetitive expansion during deposition and contraction dur-
ing stripping is shown to be deleterious to the mechanical
stability of the gold microparticles, causing the microparticles
to fracture and become irregular in morphology as well as
changing the composition of the microparticles from bulk gold
to Au–Hg amalgam. An understanding of the amalgamation
and stripping processes at gold microparticle modified electro-
des may proffer opportunities for enhanced mercury detection
using these materials.
2.3 Immobilisation of Au-lp-GCs on bppg electrode
The Au-mp-GC powder was abrasively immobilised onto the
surface of a bppg electrode.21 This was done by polishing the
electrode on silicon carbide paper (P1000) and then removing
the upper layers of graphite using adhesive tape. The Au-mp-
GCs were then immobilised onto the bppg electrode by gently
rubbing the electrode surface on a fine qualitative filter paper
(Whatman) containing the Au-mp-GC powder.
3. Results and discussion
3.1 Characterisation of Au-lp-GCs
The presence of gold on the glassy carbon microspheres was
confirmed by cyclic voltammetry. The Au-mp-GC powder was
abrasively immobilised onto a basal-plane pyrolytic graphite
(bppg) electrode and scanned from 0 to 1.4 V vs. SCE.
Oxidation of bulk gold was observed as an anodic peak at
1.2 V that was partially obscured by the effect of solvent
breakdown. The corresponding reduction peak was clearly
resolved at 0.8 V. The gold fingerprint of the abrasively
modified bppg is overlaid on a cyclic voltammogram of the
unmodified bppg in Fig. 1. Scanning electron microscopy
(SEM) was used to confirm that the chemical deposition of
gold microparticles onto glassy carbon microspheres had been
successful, with gold microparticles with an average diameter
of 0.76 Æ 0.03 mm (n = 82) clearly evident on the surface of the
glassy carbon microspheres (Fig. 2).
2. Experimental
2.1 Reagents and instrumentation
Glassy carbon microspheres (10–20 mm) and gold sodium
thiosulfate (AuNa(S2O3)2 Á H2O, 99.9%) were purchased from
Alfa Aesar (Heysham, UK). L-ascorbic acid (C6H8O6, 99%),
mercury(II) nitrate monohydrate (Hg(NO3)2, 98%), nitric acid
(HNO3, 70%) and sodium hydroxide (97%) were purchased
from Aldrich (Gillingham, UK). Dimethylformamide (reagent
grade) was purchased from Fisher (Loughborough, UK).
Aqueous solutions were prepared using deionised water
(Vivendi, High Wycombe, UK) of resistivity not less than
18.2 MO cm at 25 1C.
Cyclic voltammetry and linear sweep voltammetry were
performed using a type II mAutolab (Eco Chemie, Utrecht,
Netherlands) which was interfaced with a PC using GPES
(version 4.9) software for Windows. Measurements were
performed using a three-electrode cell of 10 mL. The working
electrode consisted of Au-mp-GCs abrasively immobilised
onto a basal-plane pyrolytic graphite (bppg) electrode. A
platinum wire was used as the counter electrode, and a silver
wire or a saturated calomel electrode (SCE) (Radiometer,
3.2 Voltammetry of mercury detected at Au-lp-GCs
Initially the linear sweep anodic stripping voltammetry
(LSASV) detection of mercury(II) nitrate was performed over
a wide potential window, from 0.15 to 1.05 V vs. SCE, after
deposition at 0.15 V for 30 s. Two sharp, well-defined peaks
were observed at 0.43 and 0.50 V and a third broad peak at
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
2072 | New J. Chem., 2007, 31, 2071–2075

View Article Online
Fig. 1 Cyclic voltammetry confirming the presence of gold micro-
particles on glassy carbon microspheres (Au-mp-GCs) abrasively im-
mobilised on a basal-plane pyrolytic graphite electrode (bppg). Scan
performed in 0.1 M HNO3 at 50 mV sÀ1. The dashed line shows the
response of an unmodified bppg, and the solid line shows the response
of a bppg with gold present.
Fig. 3 Linear sweep voltammetry (LSV) of 50 mM additions of
mercury(II) nitrate to a 0.1 M HNO3 solution at 50 mV sÀ1. Mercury
was deposited for 60 s at 0.35 V. The inset shows linear response to
mercury over the concentration range of 0–300 mM.
peak, seen in Fig. 4, appeared at the base of the carbon peaks,
which reappeared at the less positive deposition potential.
0.85 V. Comparative scans using glassy carbon (GC), unmo-
dified bppg, and bulk gold electrodes revealed that, at the high
concentration of 120 mM Hg2+, the peak occurring at 0.85 V
can be attributed to mercury stripping from gold. The oxida-
tive stripping at lower potentials can be attributed to the
stripping of mercury deposited on the carbon components
(GC and bppg) of the abrasively modified electrode. To
selectively deposit mercury solely on the gold microparticles,
the deposition potential was altered to +0.35 V, with
the subsequent scans performed over the potential range
0.35–1.1 V.
3.3 Mechanical instability of Au-lp-GCs
The amalgamation process was of particular interest to the
authors. The literature concerning underpotential deposition
and stripping (UPD and UPS, respectively) of mercury on
gold is largely concerned with bulk gold electrodes. Thus the
stripping of mercury from any Au–Hg alloy formed does not
result in significant structural change to the bulk electrode,
though an increase in surface roughness has been observed.12
To better understand the ramifications of UPS with regard to
gold microparticles, a series of depositions were performed.
Mercury(II) (100 mM) was deposited at 0.35 V for 60 s and
scanned from 0.35 to 1.1 V. Then the electrode was removed
from the Hg2+ solution and placed into 0.1 M HNO3. A cyclic
According to Schadewald et al., deposition of mercury onto
the surface of a gold electrode in excess of a monolayer results
in diffusion of the mercury into the gold surface. This creates a
Au–Hg alloy, which is subsequently stripped at approximately
0.8 V vs. SCE. Bulk mercury that has not diffused into the
electrode is then stripped from the surface at 0.4 V.12 At the
high concentrations initially examined, only alloy stripping
was observed, as seen in Fig. 3. When the mercury concentra-
tion was reduced and the parameters were adjusted to deposit
at and scan from 0.2 V, a small peak attributed to bulk
stripping of mercury from gold was observed at 0.48 V. This
Fig. 4 LSV of 10 mM additions of mercury(II) nitrate to a 0.1 M
HNO3 solution at 50 mV sÀ1. Mercury was deposited for 30 s at 0.2 V.
Peaks at 0.25, 0.32, and 0.38 V show carbon response to mercury. The
peak at 0.48 V is the gold response to mercury. The inset shows linear
response to mercury over the concentration range of 0–50 mM.
Fig. 2 Scanning electron microscope (SEM) image of glassy carbon
microspheres after chemical deposition of gold microparticles.
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 2071–2075 | 2073

View Article Online
Fig. 5 Decrease in peak area associated with gold reduction is
correlated with the number of deposition and stripping steps of
100 mM mercury(II) nitrate on a bppg modified with Au-mp-GCs.
The inset shows cyclic voltammetry taken before and after five
deposition and stripping steps were performed.
Fig. 6 SEM images of gold microparticles on glassy carbon micro-
spheres after deposition of 400 mM mercury(II) nitrate at 0.35 V for (a)
990 s and (b) 1980 s.
voltammogram over the gold fingerprint range (0–1.4 V) was
taken and the area under the cathodic peak measured. A plot
of peak areas obtained in this fashion shows a linear decrease
in the amount of electroactive gold at the electrode with
repetitive scans (Fig. 5). One possible reason for the decrease
is a gradual detachment of the Au-mp-GCs from the electrode
surface. However, when the same procedure was performed in
the absence of mercury, the decrease in gold was less than that
observed when mercury is stripped from it, and therefore the
decrease in electroactive gold cannot be explained by the Au-
mp-GCs simply falling off in solution. Thus it was posited that
repetitive stripping of the Au–Hg amalgam damaged the
microparticles, causing some gold to be stripped away with
the mercury and resulting in a decrease in both the amount of
gold present on the electrode and the resulting electroanaly-
tical performance.
seven times. There is an obvious change in morphology from
the original particle to the repetitively stripped particle. En-
ergy dispersive X-ray (EDX) analysis showed that the micro-
particles are composed of gold, mercury (the ratio varying)
and a trace amount of chloride. These repetitively expanded
and contracted amalgam microparticles were larger than both
the original and half-saturated microparticles, with an average
diameter of 1.14 Æ 0.08 mm (n = 34). Comparison of the
morphology of damaged amalgam microparticles with that of
the original microparticles as seen using high definition SEM
(Fig. 7a) reveals distinct differences. Most notably, the original
cabbage-like arrangement of gold polycrystallites is no longer
In order to better understand the processes involved, three
samples of microparticles were imaged using SEM. Fig. 6a
shows microparticles after being partially saturated (400 mM
Hg2+ deposited for 990 s at 0.35 V). They are larger than the
original sample shown in Fig. 2, with an average diameter of
0.96 Æ 0.04 mm (n = 76). The volume of the spheres is
doubled, though no major morphological changes are ob-
served. When the microparticles are saturated (400 mM
Hg2+ deposited for 1980 s at 0.35 V), their size is found to
increase dramatically, enlarging to approximately nine times
their original volume. Their average diameter is 1.51 Æ 0.14
mm (n = 91), as shown in Fig. 6b, and the microparticles
appear flattened. In some cases, the microparticles expand into
each other to the point that they are barely distinguishable as
discrete units. In that situation, they appear to pull away from
close contact with the surface of the glassy carbon micro-
spheres, possibly leading to detachment of the Au–Hg alloy.
Fig. 7b shows a high resolution image of microparticles that
have undergone repetitive deposition and stripping cycles.
Mercury(II) (400 mM) was deposited at 0.35 V for 990 s and
then scanned from 0.35 to 1.1 V. This process was repeated
Fig. 7 High resolution SEM images of (a) gold microparticles prior
to exposure to mercury and (b) Au–Hg microparticles after seven
successive deposition and scanning steps in 400 mM mercury(II) nitrate
(mercury deposited for 990 s at 0.35 V).
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
2074 | New J. Chem., 2007, 31, 2071–2075

View Article Online
observed. Instead, the microparticles appear cuboid and irre-
gular. They appear to be individual crystals and are not
symmetrically grouped. It is clear that the morphology of
the gold microparticles has changed during the repetitive
electrochemical deposition and stripping of mercury.
College for a Junior Research Fellowship. L.X. thanks Wind-
sor Scientific and the EPSRC for an Industrial CASE award.
References
These observations, taken in context with the cyclic voltam-
metry showing the decreasing amount of gold on the surface,
suggest that the stripping events oxidise some combination of
mercury and gold, altering the composition of the remaining
microparticles until a stable amalgam is left on the carbon
surface. However, note that the amalgam may not be in
electrical contact with carbon. The mechanical stress of this
process fractures the ordered polycrystallites, resulting in the
irregular morphology of the remaining microparticles.
1 J. K. Fawell, J. Cortruvo, M. Giddings, P. Jackson, Y. Magara
and E. Ohanian, Mercury in Drinking-Water: Background Docu-
ment for Development of WHO Guidelines for Drinking-Water
Quality, World Health Organization, Geneva, Switzerland,
2005.
2 N. S. Bloom and E. A. Crecelius, Mar. Chem., 1983, 14, 49–59.
3 A. Iverfeldt, Water, Air, Soil Pollut., 1991, 56, 251–265.
4 P. Porvari and M. Verta, Environ. Pollut., 2003, 123, 181–191.
5 J. Schafer, G. Blanc, S. Audry, D. Cossa and C. Bossy, Appl.
¨
Geochem., 2006, 21, 515–527.
6 G. R. Wall, H. H. Ingleston and S. Litten, Water, Air, Soil Pollut.,
2005, 165, 233–248.
7 Method 7472: Mercury in Aqueous Samples and Extracts by Anodic
Stripping Voltammetry (ASV), ed. United States Environmental
Protection Agency, Washington, DC, 1996.
8 Y. Bonfil, M. Brand and E. Kirowa-Eisner, Anal. Chim. Acta,
2000, 424, 65–76.
9 H. G. Jayaratna, Curr. Sep., 1997, 16, 93.
10 P. Richter, M. I. Toral and B. Abbott, Electroanalysis, 2002, 14,
1288–1293.
11 E. A. Viltchinskaia, L. L. Zeigman and S. G. Morton, Electro-
analysis, 1995, 7, 264–269.
12 L. A. Schadewald, T. R. Lindstrom, W. Hussein, E. E.
Evenson and D. C. Johnson, J. Electrochem. Soc., 1984, 131,
1583–1587.
13 C. M. Watson, D. J. Dwyer, J. C. Andle, A. E. Bruce and M. R. M.
Bruce, Anal. Chem., 1999, 71, 3181–3186.
14 M. A. Augelli, R. A. A. Munoz, E. M. Richter, A. J. Gouveia and
L. Angnes, Electroanalysis, 2005, 17, 755–761.
15 N. Kariuki, L. B. Isreal, M. M. Maye and C.-J. Zhong, in 222nd
ACS National Meeting, American Chemical Society, Chicago, IL,
USA, 2001.
16 O. Ordeig, C. E. Banks, J. del Campo, F. X. Munoz and R. G.
Compton, Electroanalysis, 2006, 18, 573–578.
17 C. M. Welch, O. Nekrassova, X. Dai, M. E. Hyde and R. G.
Compton, ChemPhysChem, 2004, 5, 1405–1410.
18 S. Keebaugh, A. K. Kalkan, W. J. Nam and S. J. Fonash,
Electrochem. Solid-State Lett., 2006, 9, H88–H91.
19 Y. Yang, Z. Wang, M. Yang, M. Guo, Z. Wu and G. Shen, Sens.
Actuators, B, 2006, 114, 1.
Conclusion
The oxidation of mercury on gold microparticles chemically
deposited on glassy carbon microspheres follows patterns
observed in the literature. Anodic peaks at 0.5 V and 0.8 V
were seen and attributed to the stripping of bulk mercury and
the stripping of Au–Hg amalgam, respectively. Due to the size
of the microparticles, the processes that cause pitting in bulk
gold electrodes caused destruction of the original gold micro-
particles. SEM and EDX analyses, combined with voltamme-
try, show a decrease in the amount of gold on the surface,
indicating that repetitive deposition and stripping of mercury
from gold microparticles results in drastic morphological and
compositional changes due to the mechanical stresses induced
by the expansion and contraction cycles. The knowledge of the
mechanical instability of gold microparticles with repetitive
deposition and stripping cycles should guide their application
in electrochemical sensing of mercury(II). Practicality dictates
that easily renewable gold micro- and nanoparticle-based
sensors continue to be explored in order to achieve the
maximum reproducible range of mercury(II) detection on a
gold surface.
20 K. Jurkschat, S. J. Wilkins, C. J. Salter, H. C. Leventis, G. G.
Wildgoose, L. Jiang, T. G. J. Jones, A. Crossley and R. G.
Compton, Small, 2006, 2, 95–98.
21 G. G. Wildgoose, M. Pandurangappa, N. S. Lawrence, L. Jiang,
T. G. J. Jones and R. G. Compton, Talanta, 2003, 60, 887–893.
Acknowledgements
L.E.B.-A. thanks the Physical and Theoretical Chemistry
Laboratory for financial support. G.G.W. thanks St. John’s
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2007
New J. Chem., 2007, 31, 2071–2075 | 2075