Sensors and Actuators B 123 (2007) 82–88
Detection of arsenic in groundwater using a surface
plasmon resonance sensor
Erica S. Forzani a , Kyle Foley a , Paul Westerhoff b,∗∗ , Nongjian Tao a,∗
a
Department of Electrical Engineering & Center for Solid State Electronics Research, Arizona State University, Tempe, AZ 85287, United States
b Department of Civil and Environmental Engineering, Arizona State University, Tempe, AZ 85287, United States
Received 28 March 2006; received in revised form 25 July 2006; accepted 27 July 2006
Available online 11 September 2006
Abstract
We have built a highly sensitive surface plasmon resonance (SPR) sensor to detect arsenic (As) in groundwater. Using several thiol-containing
organic compounds as sensor probes, we have been able to discriminate arsenic levels below and above the US EPA maximum containment
level (0.010 mg/L =10 ppb, according to 2006 regulation), in drinking water. The SPR sensor is simple comparing to conventional spectroscopic
techniques currently employed for As detection, and can be potentially used for As level screening in groundwater.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Arsenic; Arsenate; Arsenite; Sensor; Groundwater; Surface plasmon resonance
1. Introduction
The threat of arsenic pollution in drinking water is a serious
environmental [1–3] and health concern [1,4] because of the toxicity of arsenic on human being and on other living organisms.
In drinking water supplies, arsenic is usually naturally occurring although some pesticides and preservatives also contain
arsenic. The US EPA recently lowered the maximum containment level (MCL) for arsenic to 0.010 mg/L in drinking water.
This new mcl has created an urgent need of arsenic sensors to
detect arsenic in water supplies and manage treatment system
performance. To date, arsenic in the environment is usually measured with spectroscopic techniques, including atomic fluorescence spectroscopy (AFS), graphite furnace atomic absorption
(GFAA), hydride generation atomic absorption spectroscopy
(HGAAS), inductively coupled plasma–atomic emission spectrometry (ICP–AES), and inductively coupled plasma–mass
spectrometry (ICP–MS) [5,6]. These techniques are well established but require samples to be collected and transported to a
centralized laboratory for analysis because the instrumentation
∗
Corresponding author. Tel.: +1 480 965 3708; fax: +1 480 965 8118.
Corresponding author. Tel.: +1 480 965 2885; fax: +1 480 965 0557.
E-mail addresses: [email protected] (P. Westerhoff),
[email protected] (N. Tao).
∗∗
0925-4005/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.snb.2006.07.033
is bulky, expensive and requires significant maintenance and
operator expertise. They also involve sample preservations and
chemical reduction steps, which may introduce sample contamination [5] and prolong turnout time. In situ measurements are
thus highly desirable because they could provide early detections
of problems while minimizing errors, labor and cost associated
with the spectroscopic methods.
One of the newest promising in situ arsenic sensors is based
on anodic stripping voltammetry [7–13]. To achieve detection
limits of trace elements in the microgram to nanogram per
liter range, a preconcentration or separation technique is normally used. Arsenate, the most abundant redox specie, must
be chemically reduced to arsenite before electrochemical deposition as As(0). Following the deposition step, the electrode
potential is scanned to a more positive value to oxidize and
strip the deposited As(0) from the electrode. The oxidation
current reaches a peak at a potential characteristic of arsenic,
which provides the identity of the oxoanion. Other approaches
include bacterial fluorescent biosensors [14], electrochemical
detection with enzymatic inhibition [15], electrophoresis techniques, laser induced breakdown spectroscopy and colorimetric
assays [5,6,16]. While each method has its own advantages and
disadvantages, a simple and inexpensive yet reliable and sensitive sensor requires further effort [5,6].
Here, we present a differential SPR sensor for total arsenic
(As) screening in groundwater. SPR-based technology is an
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
accepted and established tool for investigation of biological
processes, adsorption phenomena, characterization of binding
properties of antibodies and other biological molecules, gene
analysis and drug discovery [17–24]. Recently, we have shown
that a high resolution SPR can be used to detect and quantify
heavy metal ions in drinking water [25]. In the present work,
we have developed a method to functionalize the SPR sensing surfaces for selectively and sensitively detecting arsenic in
groundwater, and demonstrating a potentially simple and costeffective device for on-site screening test of arsenic in surface
and groundwater.
A key towards arsenic detection using the SPR is to modify a ∼50 nm-thick gold film into two areas, a sensing area
coated with arsenic recognition element, and reference area that
does not adsorb arsenic. When a laser beam is reflected from
the gold film (through a prism), the absorbance of the incident light by the surface plasmons in the gold film results in
83
two dark lines, associated with the sensing and reference areas
(Fig. 1). When arsenic binds to the sensing element, the dark line
associated with the sensing element shifts with respect to that
associated with the reference area, which are sensitively monitored with a quadrant cell photodetector (Fig. 1A). We have
modified the sensing element using different arsenic recognition molecules, glutathione (GSH), dithiothreitol (DTT) and N(dithiocarboxy)-N-methyl-d-glucamine (dTGluc) (Fig. 2). GSH
was used because of its capability to metabolize As in plants
[26,27]. However, we selected dithiothreitol, a commercially
available GSH parent molecule, as a better recognition element
for As detection in water samples. This molecule can not only
reduce arsenate to arsenite and complex arsenite in situ, but also
modify the gold films more easily than GSH. We tested the SPR
sensor using groundwater and the results were compared with
those obtained with Graphite Furnace Atomic Absorption Spectroscopy (GFAAS). Our SPR arsenic sensor can discriminate
Fig. 1. Schematic representation of the differential SPR set-up. (A) A diode laser is focused with a cylindrical lens through a prism onto the gold film supported on
the prism (not shown for clarity). The gold film is divided into a sensing area and a reference area. The reflected beams from the two areas have dark lines (SPR
dips), corresponding to resonance of the surface plasmon, when the incident angle is appropriately adjusted. A quadrant cell photodetector simultaneously measures
the SPR dips from the reference (C and D) and sensing (A and B) areas. (B) Prior to each measurement, the quadrant photodetector is adjusted to balance A, B, C
and D, so that [(A − B)/(A + B)] − [(C − D)/(C + D)] ∼ 0. When the analyte is injected into the cell, the specific adsorption onto the sensing area causes a shift in the
SPR dip position, which is detected by the differential signal [(A − B)/(A + B)] − [(C − D)/(C + D)].
84
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
Fig. 2. Schematic representation of the different routes to modify sensing and reference area (see Section 2 for details).
samples with As concentrations above and below 10 ppb, the
highest permitted As level according to the EPA regulation set
in 2006 [1–4].
subtracted out, thus providing a simple and accurate method to
detect the oxoanion.
2.2. Gold film modification
2. Experimental
2.1. Instrumentation
The SPR set-up was described elsewhere [25,28]. Briefly,
it is based on the Kretschmann configuration, in which a ppolarized laser beam (λ = 635 nm) is focused through a prism
onto a metal film placed on the prism. At the so-called resonance
angle, the incident light is absorbed by the surface plasmon and
the reflection drops to a minimum, which results in a dark thin
line known as the SPR dip (Fig. 1B). We use a gold film that is
divided into a sensing area and a reference area. The dark lines
from the two areas are simultaneously monitored by a quadrant
cell photodetector, containing four nearly identical photocells
(A–D). The resonance angles from the sensing and reference
areas are detected by differential signals (A − B)/(A + B) and
(C − D)/(C + D), respectively which are recorded with a digital
oscilloscope (Yokogawa, DL708). Prior to each measurement,
the prism is rotated to bring the dark lines to the center of the
reflected beam spot, and the quadrant photodetector is adjusted
to balance not only A and B for the sensing signal but also
C and D for the reference signal (Fig. 1A). When the analyte
is injected into the cell, the index of refraction of the solution changes which causes a shift in the SPR resonance angle.
However, since the sensing area is modified with arsenic recognizing molecules, specific binding of As onto the sensing area
causes an additional resonance angle shift. The differential signal [(A − B)/(A + B)] − [(C − D)/(C + D)] eliminates the effect
due to the solution refractive index change, which allows us to
detect the specific As ion binding with an accuracy of ∼10−5◦
(Fig. 1B). Because the four photocells (on a single chip) are
nearly identical, thermal drift and mechanical noises are also
The gold film (∼50 nm thick) was evaporated on BK7 glass
slides by an ion beam coater (Model 681, Gatan Inc.). The
film was separated in two parts with a gap of 100–200 ␮m.
One part was modified as reference area with a self-assembled
monolayer of 1-dodecanethiol (DDT) or mercaptoethanol/
TRISamidepropanethiol and the second part was designated as
sensing area and modified with glutathione (GSH), dithiothreitol
(DTT) or N-(dithiocarboxy)-N-methyl-d-glucamine (dTGluc)
(Fig. 2). Two modification procedures were employed. The simpler one consisted of stamping a DDT monolayer by using
an impregnated DDT-polydimethylsiloxane (PDMS) stamp for
5 min followed by overnight or 3 h adsorption step of 15 mM
dithiothreitol in water (step 1) or 2 mM N-(dithiocarboxy)-Nmethyl-d-glucamine in 20/80 (v/v) methanol/water (step 2),
respectively. In this case, the hydrophobicity generated after
DDT modification precluded the adsorption of hydrophilic
species in the reference area during the second modification
step.
The second method of modification was developed for
modifying the sensing area with GSH. The gold film was
completely modified with 14 mM 3,3 -dithiodipropionic acid
di(N-succinimidyl ester) (NHS-thiol) dissolved in dimethylformamide (DMF) for 15–20 min. After thoroughly rinsing with
DMF, a drop of 10–20 mM glutathione +1 mM mercaptoethanol
in 10 mM phosphate buffer pH 7.5 (step 3) was set on the sensing
area and distributed uniformly by using an unmodified PDMS
stamp for 30 min. After that, the sample was dipped in 1 mM
mercaptoethanol +0.05 mM EDTA +10 mM Tris buffer pH 7.5
for at least 1 h (step 4) for modification of reference area. Mercaptoethanol was used as a surface-blocking agent, Tris as an
NHS-group quencher and EDTA as a metal ion scavenger to
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
85
protect GSH. In this way, both sensing and reference areas were
hydrophilic. The modification of sensing area with DTT and
GSH leave exposed thiol groups (Fig. 2) for quantification of
arsenate + arsenite (see below) while dTGluc allows the detection of arsenate [29].
2.3. Arsenite and arsenate detection
The response of the SPR sensors were tested in 10 mM
Tris buffer and the pH condition was optimized for stable As
detection (see below). Successive injections of different concentrations of arsenate solutions prepared in the same Tris buffer
were added after two to four additions of Tris buffer. The injections of Tris buffer allowed quantification of the differential
SPR signal drift for a baseline correction and accurate evaluation of the total SPR differential signal change in presence
of the analyte molecule. All the solutions were prepared in
Eppendorf ultracentrifuge plastic tubes thoroughly washed with
ultrapure water. The arsenate mother solution was calibrated
with GFAAS. Dilutions of this solution were prepared using
calibrated micropipettes. The SPR cell was a Teflon® cylinder of 400 ␮L capacity washed with piranha solution (98%
H2 SO4 :30% H2 O2 = 3:1, v/v) and then sonicated in 18 M cm
water three times before use (caution: piranha solution reacts
violently with most organic materials and must be handled with
extreme care). Five to ten microliters groundwater sample or
arsenate solution was injected into 310–320 ␮L of Tris buffer.
2.4. Arsenic detection in groundwater samples
Groundwater samples were from Arizona and Nebraska.
Samples of Nebraska groundwater include ambient groundwater and groundwater after two levels of arsenic treatment. Since
Nebraska’s samples contained very low As concentration they
were spiked with known As concentration and used for evaluation of SPR sensor accuracy. The samples were analyzed
with GFAAS and delivered to the Electrical Engineering Lab as
unknown As concentration samples for analysis with SPR sensors. The water samples were filtered with 0.1 ␮m nylon syringe
filter (Whatman Inc., NJ), and then treated with cation-exchange
resin filter (Dionex On Guard® , II-H) at flux rate of 10 ␮L s−1
and neutralized with 0.1 M NaOH solution before SPR analysis. Internal standard addition method was used for unknown As
level determination. Following the groundwater sample addition, the standard solutions were injected into the SPR sample
Fig. 3. (A − B) Time course of the differential SPR signal [(A − B)/A + B)] −
[(C − D)/(C + D)] upon arsenate injections. The sensing area is modified
with a monolayer of dTGluc (A) and DTT (B) and the reference area
is covered with a self-assembled monolayer of DDT (A and B). The
arsenate concentrations added to the solution cell in each injection are
indicated. The spikes observed after each injection are a consequence of
initial inhomogeneities in the solution (inhomogeneous solution refractive
index) arisen from initial injection of concentrated arsenate solutions to buffer
solution in contact with sensing (positive spikes) or reference (negative spikes)
area and also from later mechanical stirring of the solution. The time course of the
experiments is partially shown for sake of clarity. (C) Response of phosphate on
SPR sensor modified with dTGluc. (D) Schematic representation of the proposed
reaction for arsenate detection on SPR sensor modified with DTT or GSH.
Despite the fact that a H2 AsO4 − /HAsO4 2− mixture (pKa = 6.76) is dominant at
pH 6.5, only H2 AsO4 − was considered for schematic representation. D inset:
cyclic voltammogram obtained in an arsenate free solution of 2 M HCl on gold
electrode with a DTT monolayer previously exposed to arsenate for 15 min from
−300 to 400 mV vs. Ag at 100 mV s−1 .
86
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
cell. The SPR differential signal as a function of As concentration was analyzed by fitting the data with the Langmuir equation
(see below).
3. Results and discussion
3.1. Arsenate detection using different probes
Fig. 3 shows the time course of the differential signal
[(A − B)/(A + B)] − [(C − D)/(C + D)] upon introduction of Tris
buffer and arsenate into the sample cell. The sensing area was
modified with dTGluc (A) and DTT (B). The SPR differential
signals in the two cases are shown in Fig. 3A and B, respectively,
showing that both sensors are able to detect arsenate by different
sensing mechanisms.
It has been reported that dTGluc can provide selective complexation of arsenate through a combination of stereochemical
placement of –OH groups from sorbitol moieties and electrostatic interaction between the oxoanion and the quaternary amine
group of glucamine (Fig. 2) [29]. We tested the selectivity by
exposing the dTGluc-coated sensing area to high concentrations
of phosphate, the closest parent oxoanion to arsenate (Fig. 3C).
We found no significant changes in SPR signal were up to relatively high phosphate concentrations (9 ppm). In contrast, we
observed an increase in the SPR resonance angle when the sensing area was exposed to even a very dilute (∼3 ppt) arsenate
(Fig. 3A). This test confirms the selectivity of dTGluc to bind
to arsenate.
DTT-, as well as GSH-coated sensing areas are also sensitive
to arsenate injections (Fig. 3B), the SPR angle decreases, which
is opposite from the case of dTGluc-coated sensing surfaces.
Typically, a binding event increases the index of refraction and
therefore causes an increase in the SPR angle [25]. However,
a binding event can cause a decrease in the effective thickness
of the sensing molecules due to conformational changes in the
sensing molecules, which will decrease the SPR angle. The net
change in the SPR angle depends on the balance of the refractive
index and thickness effects. Since the changes we are able to
detect for chemical binding events of arsenate are smaller than 4
millidegrees, there is no significant change of width upon arsenic
saturation of the sensor surface.
An distinctive feature of DTT is the capability for detecting
As(V). This is because the reduction of As(V) to As(III) mediated by thiol groups [30], which is followed by the complexation
of As(III) to the thiol groups [31] (Fig. 3D). The proposed reaction is given as follows at pH 6.5 [26,32]:
H2 AsO4 − + 5GSH + H+
→ As–(SG)3 (III) + GS–SG + 4H2 O
(1)
HAsO4 2− + 5GSH + 2H+
→ As–(SG)3 (III) + GS–SG + 4H2 O
(2)
We believe that the release of water, the formation of GS–SG
and As(III)–glutathione complex would be responsible for the
thickness decrease of the sensing layer mentioned above. In
Fig. 4. (A) Absolute differential SPR signal change as a function of arsenate
concentration on sensors modified with GSH, DTT, dTGluc. 10 mM Tris pH
6.5. Full lines show the corresponding fitting to Eq. (3). (B) Effect of pH on the
response of GSH modified SPR sensor. The dominant arsenate–thiol complex
is indicated according to the pH value.
order to confirm the presence of As(III) on the SPR sensing
surface, we performed cyclic voltammetry on DTT-coated gold
electrodes previously incubated in arsenate solution for 15 min
in deoxygenated arsenate free HCl (2 M). The cyclic voltammogram reveals a peak corresponding to the oxidation of As(0)
to As(III) [7], which is expected from the above proposed
reaction.Fig. 4A shows SPR signals as a function of arsenate
concentration at pH 6.5 using gold films modified with GSH,
DTT and dTGluc. By fitting the data using Langmuir equation
[25] given by
θ =
θmax C
Kd + C
(3)
where θ is the SPR angle shift; θ max the SPR angle shift at
the saturation; C the arsenate concentration; Kd is the dissociation constant, we have determined Kd (Table 1 and full lines
Fig. 4A). Due to that the SPR signal changes were probably
not taken under real equilibrium situations and, in the cases of
DTT and GSH, the changes are a consequence of coupled chemical reactions, the extracted dissociation constants are apparent
and useful only for the purpose of comparing results obtained
under similar experimental conditions. The apparent Kd values
were rather small (few ppb to tens of nM) for all the sensing
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
87
Table 1
Comparative analytical performance and “apparent” langmuirian parameters for
SPR sensorsa
Sensing area probe GSH
Kd (ppb)
SPR shiftmax (◦ )
Detection limit
a
DTT
dtGluc
2.0 ± 0.8
0.25 ± 0.10
2.0 ± 10
(9.1 ± 0.8) × 10−4 (2.1 ± 0.4) × 10−3 (1.3 ± 02) × 10−3
1 ppb
3 ppt
1.5 ppt
See text for details of SPR shiftmax and Kd calculus.
molecules, however Kd for DTT is much smaller than the Kd
for GSH and dTGluc, indicating a stronger apparent affinity of
DTT surfaces for arsenate. However, since DTT-modified sensors have better maximum response (Table 1) and can quantify
both As(V) and As(III), it was chosen for the detection of total
As in groundwater samples.
We have determined optimal working pH conditions for As
detections. Fig. 4B shows as example the response of arsenate
on a GSH modified sensor at pH 6.5 and 9.0. Although the
SPR differential signal change for 10 ppt obtained at pH 9.0 is
higher than the one at pH 6.5, the total sensitivity defined at
higher concentration at pH 9.0 resulted smaller than the one at
pH 6.5. This is because the dominant As–thiol complex at pH 9.0,
As(OH)2 –GSH, is not stable for long period of time (half-life
time of about 7 min) [33]. As consequence, pH 6.5 was chosen
as working pH for groundwater analysis.
3.2. Arsenic analysis in groundwater samples
The chemical composition of groundwater can be very complex and may vary over space and time since it is influenced
by temperature, human activities, presence of microorganisms,
reactions with soil and rocks from aquifers, etc. [34]. Organic
compounds, bacteria, viruses, colloidal particles, inorganic ions
can be present in groundwater. Thus, pretreatment of groundwater samples results essential in order to perform quantitative
analysis of small molecular weight species. In our case, removal
of potential microorganisms and large suspended particles was
achieved by filtration with 0.1 ␮m nylon syringe filters of the
groundwater samples. Since our probes modifying sensing area
(thiolated molecules) were not only sensitive to arsenate but
also to heavy metal ions, we also needed to remove potential
interferents cations as Pb2+ , Cd2+ , Cu2+ , Zn2+ , Ni2+ , Fe2+/3+ ,
Hg2+ [35]. We found that the sample pretreatment with cation
exchange resins (Dionex On Guard® , II-H) was crucial to minimize the heavy metal ion interference and correlate SPR sensor
results with GFAAS values.
The test of arsenic concentration was carried out using the
following procedure: 5 ␮L of cation-exchange treated groundwater was first injected into 310–320 ␮L of 10 mM Tris buffer
(pH 6.5) which was followed by successive injections of arsenate
standard solutions until the sensing surface is saturated with As
(Fig. 5). The data were then fit using the Langmuir equation to
extract unknown As concentration and the dissociation constant,
which is 0.8–1.0 ppb in all groundwater samples for DDT-coated
sensing surface. Table 2 summarizes the average results obtained
in four groundwater samples (A from Arizona and B–D from
Fig. 5. Calibration plot for As internal standard addition analysis of a groundwater using a SPR sensor built with DTT modified sensing area. Full-line fitting
to Eq. (3).
Table 2
Arsenic analysis of groundwater samples performed SPR sensors and GFAAS
Groundwater
samples
GFAAS
(ppb)
SPRa (ppb)
SPR diagnosis to EPA’s
As permitted level
A
B
C
D
32
19
49
4
25 ± 15
14 ± 9
78 ± 42
Not detectable
+
+
+
−
Injection of 5 ␮L of pretreated sample to 310–320 ␮L of 10 mM Tris buffer
pH 6.5.
a
Nebraska). The average represents the results from two to four
tests carried out on each sample. The standard deviation for each
sample analysis is relatively high, which is a consequence of the
fitting procedure involving only four to five data points in the
curve. This limitation comes from the high apparent affinity constant involved in the recognition reaction and the relatively low
dynamic range of the technique. However, the As concentration
determined for each sample from SPR sensors is in agreement
with the value obtained with GFAAS.We shown earlier that the
SPR technique can detect arsenate in pure buffer at the ppt level.
This detection limit could not be reached for the groundwater
samples, which is likely due to sample matrix effect arisen from
non-specific adsorption on sensing and reference area of other
non-filtrated groundwater compounds as organic compounds or
inorganic oxoanions. No further characterization of these nonspecific interferences was performed. However, the SPR sensor
was able to discriminate between samples with As concentration
lower and higher than the maximum containment level established according the new EPA’s regulations.
4. Conclusions
In summary, we have demonstrated a method to analyze
total As in groundwater by combining a high performance
differential SPR sensor with proper modifications of the SPR
sensing surfaces. In comparison to the widely used techniques,
such as Atomic Absorption Spectrometry, the SPR method is
88
E.S. Forzani et al. / Sensors and Actuators B 123 (2007) 82–88
simpler, faster and less expensive. The use of DTT allowed in
situ reduction of arsenate [32] and avoid the chemical pretreatment stages required by most of arsenic detection methods.
In comparison with Anodic Stripping Voltammetry, the SPR
method is faster and does not require sample preconcentration.
Acknowledgements
We thank A. Baumgardner and Troy Benn for assisting us
to collect groundwater samples and perform atomic absorption spectroscopy and Salt River Project (Phoenix, AZ), EPA
(R82962301) and NSF(CHE-0243423) for funding. K.F. thanks
support from IGERT (Integrative Graduate Education and
Traineeship) program.
References
[1] NRC, NRC Subcommitte on Arsenic in Drinking Water Committee on
Toxicology, Arsenic in Drinking Water, NRC, Washington, DC, 1999.
[2] NRC, Subcommittee on Arsenic in Drinking Water, Committee on Toxicology, Arsenic in Drinking Water, National Research Council, Washington,
DC, 1999.
[3] NRC, Subcommittee to Update the 1999 Arsenic in Drinking Water, Committee on Toxicology, Arsenic in Drinking Water, National Research Council, Washington, DC, 2001.
[4] NRC, NRC Subcommittee to Update the 1999 Arsenic in Drinking Water,
Committee on Toxicology, NRC, Washington, DC, 2002.
[5] D. Melamed, Anal. Chim. Acta 532 (2005) 1–13.
[6] D.Q. Hung, O. Nekrassova, R.G. Compton, Talanta 64 (2004) 269–277.
[7] X. Dai, O. Nekrassova, M.E. Hyde, R.G. Compton, Anal. Chem. 76 (2004)
5924–5929.
[8] G. Hignett, et al., Electroanalysis 16 (2004) 897–903.
[9] A. Salimi, M.E. Hyde, C.E. Banks, R.G. Compton, Analyst 129 (2004)
9–14.
[10] R. Feeney, S.P. Kounaves, Anal. Chem. 72 (2000) 2222–2228.
[11] H.L. Huang, P.K. Dasgupta, Anal. Chim. Acta 380 (1999) 27–37.
[12] J. Wang, M. Ariel, J. Electroanal. Chem. 128 (1981) 147–153.
[13] R. Feeney, S.P. Kounaves, Talanta 58 (2002) 23–31.
[14] J. Stocker, et al., Environ. Sci. Technol. 37 (2003) 4743–4750.
[15] M. Stoytcheva, V. Sharkova, M. Panayotova, Anal. Chim. Acta 364 (1998)
195–201.
[16] P.K. Dasgupta, H.L. Huang, G.F. Zhang, G.P. Cobb, Anal. Chim. Acta 58
(2002) 153–164.
[17] T.M. Battaglia, et al., Anal. Chem. 77 (2005) 7016–7023.
[18] X.D. Su, Y.J. Wu, W. Knoll, Biosens. Bioelectr. 21 (2005) 719–726.
[19] H.J. Lee, A.W. Wark, Y. Li, R.M. Corn, Anal. Chem. 77 (2005) 7832–7837.
[20] Z.Z. Wang, T. Wilkop, Q. Cheng, Langmuir 21 (2005) 10292–10296.
[21] S. Fang, H.J. Lee, A.W. Wark, H.M. Kim, R.M. Corn, Anal. Chem. 77
(2005) 6528–6534.
[22] H.J. Lee, Y. Li, A.W. Wark, R.M. Corn, Anal. Chem. 77 (2005) 5096–5100.
[23] T. Wilkop, Z.Z. Wang, Q. Cheng, Langmuir 20 (2004) 11141–11148.
[24] M. Piliarik, H. Vaisocherova, J. Homola, Biosens. Bioelectr. 20 (2005)
2104–2110.
[25] E.S. Forzani, H.Q. Zhang, W. Chen, N.J. Tao, Environ. Sci. Technol. 39
(2005) 1257–1262.
[26] N. Scott, K.M. Hatlelid, N.E. Mackenzie, D.E. Carter, Chem. Rev. Tox. 6
(1993) 102–106.
[27] C. Rensing, Personal Communication, Seminar at Arizona State University,
2005.
[28] H.Q. Zhang, S. Boussaad, N.J. Tao, Rev. Sci. Instrum. 74 (2003) 150–
153.
[29] L. Dambies, R. Salinaro, S.D. Alexandratos, Environ. Sci. Technol. 38
(2004) 6139–6146.
[30] M. Delnomdedieu, M.M. Basti, J.D. Otvos, D.J. Thomas, Chem. Biol.
Interact. 90 (1994) 139–155.
[31] N.A. Rey, O.W. Howarth, E.C. Pereira-Maia, J. Inorg. Biochem. 98 (2004)
1151–1159.
[32] http://ehp.niehs.nih.gov/members/1995/Suppl-1/carter-full.html.
[33] A. Raab, A.A. Meharg, M. Jaspars, D.R. Genney, J. Feldmann, J. Anal. At.
Spectrom. 19 (2004) 183–190.
[34] http://www.umwelt-schweiz.ch/buwal/eng/fachgebiete/gewaesserschutz/
grundwasser/beschaffenheit/index.html.
[35] http://www.cluin.org/download/toolkit/metals.pdf.
Biographies
Erica S. Forzani is an Assistant Research Professor in the Department of Electrical Engineering at ASU in Tempe, AZ. She joined ASU in 2003 as Research
Associate in NJ Tao’s group after receiving her PhD in Chemistry in 1999 from
Cordoba National University, Argentina and a postdoctoral degree in University of Buenos Aires, Argentina (2000–2003). Her current research interest is the
development of nanosensors for chemical and biochemical detection of environmental and health care analytes as well as the optimization of different detection
methods based on optical, electrical and acoustic detection.
Kyle Foley received his BSc degree in electrical engineering from the Milwaukee
School of Engineering in 2003 and is currently pursuing the PhD degree from
Arizona State University in Tempe, AZ. His main research interests include
surface plasmon resonance sensors and biomedical applications.
Paul Westerhoff, PhD, is an Associate Professor in the Department of Civil
and Environmental Engineering at Arizona State University (ASU) in Tempe,
AZ. He joined ASU in 1995 after obtaining his PhD from the University of
Colorado at Boulder. His research focuses on water quality and drinking water
treatment, and more specifically on reactions and fate of oxo-anions (bromate,
nitrate, arsenate), characterization of dissolved organic matter in watersheds and
treatment processes and reactions with oxidations, formation of disinfection byproducts, and removal of emerging organic contaminants.
Nongjian Tao joined the ASU Faculty as a Professor of Electrical Engineering
and an Affiliated Professor of Chemistry and Biochemistry in August 2001.
Before that, he worked as an Assistant and Associate Professor at Florida
International University. His research interests focus on molecular electronics,
nanostructured materials and devices, chemical and biological sensors, interfaces between biological molecules and solid materials, and electrochemical
nanofabrications.