a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/aca

Square wave adsorptive cathodic stripping voltammetry

automated by sequential injection analysis
Potentialities and limitations exemplified by the
determination of methyl parathion in water samples

Luciana B.O. dos Santos, Jorge C. Masini ∗

Instituto de Quı́mica, Universidade de São Paulo, C.P. 26077, 05513-970, São Paulo, SP, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: This paper describes the development and evaluation of a sequential injection method to
Received 31 August 2007 automate the determination of methyl parathion by square wave adsorptive cathodic strip-
Received in revised form ping voltammetry exploiting the concept of monosegmented flow analysis to perform in-line
1 November 2007 sample conditioning and standard addition. Accumulation and stripping steps are made
Accepted 6 November 2007 in the sample medium conditioned with 40 mmol L−1 Britton–Robinson buffer (pH 10) in
Published on line 13 November 2007 0.25 mol L−1 NaNO3 . The homogenized mixture is injected at a flow rate of 10 ␮L s−1 toward
the flow cell, which is adapted to the capillary of a hanging drop mercury electrode. After
Keywords: a suitable deposition time, the flow is stopped and the potential is scanned from −0.3 to
Methyl parathion −1.0 V versus Ag/AgCl at frequency of 250 Hz and pulse height of 25 mV. The linear dynamic
Adsorptive voltammetry range is observed for methyl parathion concentrations between 0.010 and 0.50 mg L−1 , with
Sequential injection detection and quantification limits of 2 and 7 ␮g L−1 , respectively. The sampling throughput
Waters is 25 h−1 if the in line standard addition and sample conditioning protocols are followed,
but this frequency can be increased up to 61 h−1 if the sample is conditioned off-line and
quantified using an external calibration curve. The method was applied for determination
of methyl parathion in spiked water samples and the accuracy was evaluated either by com-
parison to high performance liquid chromatography with UV detection, or by the recovery
percentages. Although no evidences of statistically significant differences were observed
between the expected and obtained concentrations, because of the susceptibility of the
method to interference by other pesticides (e.g., parathion, dichlorvos) and natural organic
matter (e.g., fulvic and humic acids), isolation of the analyte may be required when more
complex sample matrices are encountered.
© 2007 Elsevier B.V. All rights reserved.

1. Introduction important crops such as cotton, soybeans, cereals, fruits, sugar

cane, etc. It is used to control aphids, boll weevils, and mites
Methyl parathion (O,O-dimethyl O-p-nitrophenyl phospho- on cotton, soybeans, wheat, alfalfa, rice, lettuce, onion, sug-
rothioate) is a restricted-use organophosphorus insecticide arbeets, and artichokes. Because of its high toxicity, methyl
used throughout the world on a number of economically parathion is not approved for nonprofessional use. Poisoning

∗
Corresponding author. Tel.: +55 11 3091 3837x233; fax: +55 11 3815 5579.
E-mail address: jcmasini@iq.usp.br (J.C. Masini).
0003-2670/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2007.11.009
210 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216

with methyl parathion leads to cholinergic overstimulation, tion system were exploited to perform sample conditioning,
with signs of toxicity including sweating, dizziness, vomiting, sample accumulation and concentration, as well as automatic
diarrhea, convulsions, cardiac arrest, respiratory arrest, and, calibration, facilitating the implementation of on-line auto-
in extreme cases, death [1]. Methyl parathion enters the envi- mated monitoring.
ronment primarily by spraying on farm crops, but it breaks
down quickly to other chemicals such as methyl paraoxon,
p-nitro phenol, dimethyl phosphoric acid, methyl phospho- 2. Experimental
rothioic acid and aminomethyl parathion as a consequence of
biodegradation, hydrolysis, interactions with natural organic 2.1. Apparatus and reagents
matter and photolysis [1–3]. The drinking water equivalent
level (DWEL) for methyl parathion is 7 ␮g L−1 , according to the Voltammetric measurements were carried out using an EG&G
2006 Edition of the Drinking Water Standards and Health Advi- PAR model 263A potentiostat. An EG&G PAR model 303A
sories of the United States Environmental Protection Agency static mercury drop electrode (SMDE) was used in all exper-
[4]. iments. Purified and doubly distilled mercury was used in
The determination of residues of methyl parathion in the working electrode. The flow cell (estimated internal vol-
waters is usually made by high performance liquid chromatog- ume of 15 ␮L) adapted to the Hg capillary has already been
raphy with ultra violet [2,5,6], fluorescence [7], electrochemical described [23]. The Hg drop radius used was 0.46 mm. The
[8] and mass spectrometry [3,9] detection modes. Gas electrochemical cell was completed with an Ag/AgCl reference
chromatography has also been used associated to mass electrode (KCl saturated) and a platinum auxiliary electrode.
spectrometry or flame thermoionic detector [10–14]. These Ultrapure N2 (O2 < 2 ppm) was used to remove dissolved O2
methods require expensive instruments and analyte extrac- from the solutions and to provide an inert atmosphere inside
tion and concentration before analysis. Electroanalytical the cell. Prior to the analysis, the carrier solution was stored
methods based on differential pulse and square wave voltam- in a 250 mL Schott flask (Schott North America, Inc., Elms-
metry meet the requirements of rapid analyses and low cost ford, NY, USA) and de-oxygenated for at least 30 min with N2 .
instrumentation. Sample treatment is minimal or even unnec- De-oxygenation was maintained during all the experiment.
essary. De Souza and Machado [15] described the application Standard and sample solutions were stored in 15 mL glass
of gold and carbon fiber microelectrodes for determination of tubes and de-oxygenated for 5 min prior to the analysis. In
methyl parathion and other organophosphorus insecticides case of serial analysis, several samples were simultaneously
and bipyridilium herbicides by square wave voltammetry, bubbled with N2 , so that the sampling frequency was not lim-
finding detection limits smaller than 15 ␮g L−1 in pure water ited by the de-oxygentaion time. The covers of the carrier
samples. Liu and Lin [16] proposed an electrochemical strip- flask and sample/standard tubes contained three holes, one
ping analysis of organophosphosphate pesticides and nerve to introduce the sampling tube, another to introduce the N2
agents based on adsorption of the analytes on a carbon paste- flow, and the last one to allow the N2 exit.
working electrode, followed by square wave voltammetry, A FIAlab 3500 (FIAlab Instruments, Bellevue, WA, USA)
finding a detection limit of 13 ␮g L−1 at an accumulation time instrument was used in all experiments in the sequential
of 600 s. Ni et al. [17] proposed the simultaneous determi- injection mode according to Fig. 1. Solutions were driven by
nation of three organophosphorus pesticides by differential a 5.00 mL syringe pump and an eight port rotary valve, RV
pulse stripping voltammetry and chemometrics using the (Valco Instrument Co., Houston, TX). The holding coil, HC, was
hanging mercury drop electrode, reporting a detection limit made of 3 m Teflon (polytetrafluoroethylene, PTFE) tubing with
of 4.8 ␮g L−1 for methyl parathion. 0.8 mm i.d. The tubing connecting RV to the flow cell (ECFC)
Flow injection systems have been proposed for determi- was 27 cm long, made of 0.5 mm i.d. PTFE tubing. All other
nation of methyl parathion in water samples by enzyme tubing connections were made of 0.5 mm i.d. PTFE tubing and
linked immunosorbent assay [18], amperometric detection PTFE nuts and ferrules (Upchurch, Oak Harbor, WA). Port 5 of
using biosensors [19] and luminol based chemiluminescence RV was connected to a 4 mm i.d. glass tube used to homog-
[20]. Flow injection adsorptive stripping voltammetry was enize the monosegment. An auxiliary peristaltic pump (not
first proposed by Romanus et al. [21] for determination of shown in Fig. 1) was used to continuously draw off the excess
trace metals in brine, but few attention has been given to of solution inside the glass three-electrode cell, as described
this approach for determination of pesticides in waters. Flow previously [23,24].
injection system improves efficiency of sample concentra- An LC 9A Shimadzu high performance liquid chromato-
tion during the accumulation step, requires minimum sample graph (HPLC), equipped with a SPD 6 AV UV detector, and
treatment, facilitates the medium exchange between the the LC Workstation Class-LC 10 software was used in all
accumulation and stripping steps and has ready adaptability experiments for quantification of methyl parathion. A SB C-18
for automation and on-line monitoring. The later two char- Zorbax—HP column (3.5 ␮m, 150 mm × 4.6 mm) connected to
acteristics are especially facilitated by sequential injection a C-18 Phenomenex guard column was used. Sample injection
analysis because of its mechanical configuration and robust- was made with a rotary Rheodyne valve using a 20 ␮L sample
ness [22]. The present paper reports the automation of square loop.
wave adsorptive cathodic stripping voltammetry (SI-ACSV) All reagents used in this work were of analytical grade and
by sequential injection analysis for determination of methyl all working solutions were prepared in deionized water (Sim-
parathion in surface and drinking waters using the hanging plicity 185 system from Millipore coupled to an UV lamp). A
mercury drop electrode. The properties of the sequential injec- 1000 ␮g mL−1 stock methyl parathion solution was prepared by
 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216 211

Fig. 1 – Sequential injection manifold to perform SI-ACSV. C = carrier (40 mmol L−1 BR buffer at pH 10 in 0.25 mol L−1 NaNO3 );
SV = syringe valve; SP = syringe pump; HC = holding coil (made of 3 m of PTFE tubing of 0.8 mm internal diameter); RV = eight
port rotary selection valve; S = sample reservoir; B = 80 mmol L−1 BR buffer and 0.50 mol L−1 NaNO3 ; STD = 0.50 mg L−1 methyl
parathion standard solution (in the case of calibration curves this standard is prepared in the same medium as C, but in
case of standard addition it is prepared in medium similar as B); DC = dilution/homogenization device made of 0.3 m of glass
tube of 4.0 mm internal diameter; ECFC = electrochemical flow cell (connection between ECFC and RV is made of 0.27 m of
PTFE tubing of 0.5 mm i.d.); W = waste. More details about the ECFC are given in ref. [21].

dissolving the solid standard (Riedel-de Haën, purity 99.8%, at 100 ␮L s−1 , washing the system for the next analysis. The
molar mass of 263.23 g mol−1 ) in ethanol. Working solutions total analysis time is 59 s.
were prepared by diluting this stock solution in distilled deion-
ized water. The voltammetric experiments were performed
2.2.2. Construction of analytical curve by in-line dilution
in medium of 40 mmol L−1 BR buffer at pH 10 in 0.25 mol L−1
of a single standard
NaNO3 .
The HC, the electrochemical cell and the tubing connecting
the port 1 of RV are filled with C. The tubing connecting the
2.2. Sequential injection procedures port 2 is filled with the previously de-oxygenated 0.50 ␮g L−1
methyl parathion standard solution (STD) prepared in the
2.2.1. Construction of analytical curve with off-line same medium as the carrier. As in the preceding protocol,
prepared standards the programs controlling both the potentiostat and SIA sys-
The first step of the SIA system is to fill HC and the elec- tems start simultaneously, but in the present case the delay
trochemical cell connected to port 3 of the rotary selection time for applying the deposition potential was programmed at
valve (Fig. 1) with the de-oxygentated carrier solution (C). 87 s. The syringe valve (SV) is set at position ‘OUT’, connect-
The tubing connecting the port 2 of RV is filled with de- ing the syringe with the carrier reservoir and, at a flow rate
oxygenated standard or sample solutions (STD or S). To of 500 ␮L s−1 , 600 ␮L of carrier are aspirated inside the syringe.
perform this calibration, methyl parathion standard solutions SV switches to position ‘IN’, connecting the syringe with RV
with concentrations between 10 and 500 ␮g L−1 were prepared through HC. After that, RV switches to port 7, and SP aspi-
in volumetric flasks, in presence of 40 mmol L−1 BR buffer at rates 100 ␮L of air inside HC using a flow rate of 50 ␮L s−1 . In
pH 10 in 0.25 mol L−1 NaNO3 . All other ports are filled with C, the sequence, a monosegment with a total volume of 800 ␮L
with exception of port 7, which is used to introduce air bubbles. is formed into HC. This total volume is composed by the fol-
The potentiostat and the SIA programs start simultaneously. lowing standard/carrier volumes (in ␮L): 0/800, 16/784, 40/760,
A delay time of 24 s is programmed in the potentiostat. While 80/720, 160/640, 400/400, and 800/0 generating, respectively,
this delay time runs, the SIA system, with SV at position ‘OUT’ methyl parathion concentrations of 0, 10, 25, 50, 100, 250 and
(Fig. 1), aspirates 600 ␮L of C inside the syringe at a flow rate 500 ␮g L−1 inside the monosegment, which is completed by
of 500 ␮L s−1 . Next, with SV at position ‘IN’, 100 ␮L of air and aspirating 100 ␮L of air into HC. The volumes of standard (S)
450 ␮L of standard or sample solutions are sequentially aspi- and carrier (C) are fractionated four times, resulting a stack
rated to HC at 50 ␮L s−1 from ports 7 and 2 of RV, respectively. of S/C/S/C/S/C/S/C zones between the two air bubbles, with
As the delay time ends, the potentiostat applies −0.3 V to the each one of the S/C zones having a total volume of 200 ␮L
electrochemical cell during 30 s. Simultaneously, RV of the SIA [25,26]. The rotary valve switches to port 5 and a total volume
system switches to port 3 and SP dispenses 400 ␮L of stan- of 1000 ␮L (monosegment plus the two air bubbles) is injected
dard or sample solutions toward the flow cell at 10 ␮L s−1 . The into DC at 50 ␮L s−1 homogenizing the mixture zone. Next, at a
rotary valve (RV) switches to port 6 and SP dispenses 250 ␮L flow rate of 50 ␮L s−1 , 850 ␮L of the DC are aspirated back to HC
at 200 ␮L s−1 , discarding the air bubble and the excess of sam- completing the homogenization and leaving the front air bub-
ple. Under stopped flow conditions inside the flow cell, after ble inside DC. The rotary selection valve switches to port 3 and
5 s of equilibrium time, the potential is scanned from −0.3 to SP injects 400 ␮L through the flow cell at 10 ␮L s−1 . While the
−1.0 V using the frequency of 250 Hz and pulse height of 25 mV. homogenized zone of the monosegment is traveling through
Finally, RV switches back to port 3 and SP empties the syringe the flow cell, the potentiostat applies −0.3 V to the three elec-
212 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216

Fig. 2 – Reduction of methyl parathion.

trodes electrochemical cell for 30 s. Then, RV switches to port Paulo city (Brazil). These reservoirs are used as water sup-
6 and SP dispenses 650 ␮L at 200 ␮L s−1 , discarding the air bub- plies for São Paulo city. Other samples were collected in the
ble and the excess standard or sample. While these steps are Praia dos Namorados (PN) and Salto Grande (SG) reservoirs
being made by the SIA system the potentiotat sweeps the of Atibaia river, near to the municipality of Americana (São
potential from −0.3 to −1.0 V versus Ag/AgCl using a square Paulo state), in an agricultural area dominated by sugar cane
wave frequency of 250 Hz and pulse height of 25 mV. Finally, cultivation. Water samples were filtered through a 0.45 ␮m cel-
RV switches back to port 3 and SP empties the syringe at a flow lulose acetate membrane and stored in glass bottles at 4 ◦ C.
rate of 100 ␮L s−1 washing the flow cell. For in-line analyses Analyses were performed by the standard addition approach
and conditioning of unknown samples, the buffering solution described in Section 2.2.3. Recovery experiments were per-
connected to port 1 is changed by a 80 mmol L−1 BR buffer (pH formed by standard addition, after spiking the samples with
10) in 0.50 mol L−1 NaNO3 . Equal volumes of buffer and sam- 0.10 mg L−1 methyl parathion. A delay time of 24 h was adopted
ple are aspirated into HC following the fractionation pattern between the spiking and the analyses, as was previously made
described above, so that the sample and buffering solutions for atrazine [24] and picloram [27].
are diluted at a 1:1 ratio. Three drinking water samples consumed in São Paulo city
were also studied. These samples were spiked with methyl
2.2.3. In-line standard addition parathion at the concentration level of 10 ␮g L−1 and analyzed
The procedure was similar to the one described for analytical by the protocol described in Section 2.2.1.
curve with in-line standard dilution, but with the following
modifications. The delay time for potential scanning was 106 s.
The volume of the monosegment is 800 ␮L, but formed by 3. Results and discussion
combining the total volumes described in Table 1. During the
aspiration step to HC, the sample volume is fractionated five Accumulation of methyl parathion on the electrode is a con-
times and properly intercalated by four zones of buffer and sequence of the interaction of the sulfur atom with the Hg
standard zones [24,27]. To obtain the monosegment solution surface [28,29] at a suitable potential. After the accumula-
composed by 40 mmol L−1 BR buffer at pH 10 in 0.25 mol L−1 tion step, the cell potential is scanned toward cathodic region,
NaNO3 , the standard methyl parathion (STD) and the diluting reducing the adsorbed methyl parathion in an irreversible and
(B) solutions were both prepared in 80 mmol L−1 BR buffer and pH-dependent process that occurs at the nitro group, produc-
0.50 mol L−1 NaNO3 . ing a hydroxylamine according to the reaction shown in Fig. 2
[30].
2.3. HPLC analysis The hydroxylamine can be reversibly oxidized to a nitrous
group compound, but this reaction occurs at potentials more
Experiments were made in isocratic elution mode with a positive than that one needed for oxidizing Hg [15], so that only
mobile phase constituted by 50% (v/v) acetonitrile:water mix- the reduction reaction was explored for the method develop-
ture. Both solutions constituting the mobile phase were ment. Despite of exploring an irreversible reaction, the use of
previously filtered through 0.45 ␮m PTFE membranes. Helium square wave voltammetry is still attractive because of the high
was used as degassing gas in all experiments. The analyses speed at which the measurements are made, being suitable for
were performed under a flow rate of 1.2 mL min−1 . The UV application in flow systems.
detector monitored the absorbance at 273 nm. Fig. 3 shows the square wave voltammograms obtained at a
frequency of 100 Hz and pulse height of 25 mV in 40 mmol L−1
2.4. Water samples BR buffers adjusted to pH 4.0, 6.0, 8.0 and 10, all of them pre-
pared in 0.25 mol L−1 NaNO3 . The peak potential shifts towards
The water samples were collected in the Billings (B) and Guara- more cathodic values as the pH increases, a behavior that
piranga (G) reservoirs located in the Metropolitan Area of São has also been described for other nitrated organophosphate

Table 1 – Combination of volumes of sample (S), carrier solution (CS) and standard methyl parathion solution (SS)
aspirated into the holding coil (HC) to perform in line standard addition
Concentration (␮g L−1 ) Sample volume (␮L) Standard volume (␮L) Carrier volume (␮L)

0 400 0 400
25 400 40 360
50 400 80 320
75 400 120 280
 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216 213

Fig. 3 – Influence of pH on peak current and peak potential

on the adsorptive cathodic stripping square wave
voltammmograms of a 0.50 mg L−1 methyl parathion
solution in 40 mmol L−1 BR buffers in 0.25 mol L−1 NaNO3 .
The voltammograms were obtained at a square wave
frequency of 100 Hz and pulse height of 25 mV under
stopped flow conditions and deposition time of 30 s
performed at flow rate of 10 ␮L s−1 . The sequential injection
protocol is described in Section 2.2.1.

pesticides such as fenitrothion and parathion [17,29,31]. Peak

potentials, plotted as a function of pH, fitted a linear equation
with slope of 52 ± 3 mV pH−1 , which is in reasonable agree-
ment with the involvement of one proton per electron in the
reduction process. In agreement with Ni et al. [17], peak cur-
rents increased with the pH, so that pH 10 was chosen for the
proposed method. Higher pH might improve the sensitivity,
but at the cost of more frequent need for silanization of the
Hg capillary electrode and less stability of the Hg drop. Addi-
tionally, hydrolysis of methyl parathion may occur at pH > 8
[32,33], but this reaction is slow in comparison to the time
scale of the proposed method, in which buffering and sample
conditioning are performed in-line, immediately before the
measurement step.

3.1. Optimization of instrumental parameters

The first parameter to be optimized for determination of

methyl parathion employing SI-SWV was the synchronization
between the SIA system and the potential scanning on the
working electrode. The steps for preparing the SIA system,
Fig. 4 – Influence of deposition potential (a), deposition
as well as sampling, sample conditioning and standard addi-
time (b) and square wave frequency (c) on the peak current
tion protocols demand a time that have to be accounted for in
measurements. The experiments were made with a
the program of the potentiostat to allow the potential scan to
0.50 mg L−1 methyl parathion solution in medium of
coincide with presence of sample solution inside the flow cell,
40 mmol L−1 BR buffer (pH 10) in 0.25 mol L−1 NaNO3 . Error
in contact with the working electrode surface [24]. The time
bars correspond to the standard deviations of three
needed to synchronize SIA steps with potential scan depends
experiments. The sequential injection protocol is described
on the procedure used. For off-line conditioned samples this
in Section 2.2.1.
time was 59 s, leading to a sampling frequency of 61 h−1 . In line
buffering required 87 s for solution handling, plus 35 s for accu-
mulation, equilibrium and potential sweeping steps, leading
to sampling frequency of 29 h−1 . For in-line standard addition
214 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216

protocol, the sampling frequency was 25 h−1 , but in this case, 10 to 90 s. Fig. 4b shows a sharp increase in peak currents up
additionally to the sample analysis, three standard additions to 30 s. From 30 to 60 s, only a small increase in the signal
are made, and the total time for each sample increases to 564 s, magnitude is observed, so that the deposition time of 30 s was
implying that six different samples can be processed per hour. chosen for the method as a compromise between sensitivity
The experimental conditions needed to maximize the and sampling frequency.
adsorption of the analyte on the Hg electrode were studied The cathodic stripping voltammetry performed by sweep-
with a 0.50 mg L−1 methyl parathion solution in the previ- ing the potential from −0.30 to −1.0 V under continuous flow
ously selected pH 10 BR buffer. The square wave frequency was conditions through the cell caused a significant decrease of
100 Hz, with a pulse height of 25 mV. The deposition potential the peak currents, so that this step was made in stopped
was studied in the range from −0.50 to 0.0 V versus Ag/AgCl at flow conditions. This fact is consistent with weak interactions
a flow rate of 10 ␮L s−1 (Fig. 4a). The maximum peak current in between the P = S group of methyl parathion with the Hg sur-
the stripping step was observed for the deposition potential of face, as has been evidenced by El-Shahawi and Kamal [34].
−0.30 V, in agreement with Bourque et al. [28] and Ni et al. [16], Weak interaction requires low flow rate to avoid desorption
who found −0,20 and −0,30 V versus Ag/AgCl, respectively. from the electrode surface. Sensitivity is also greatly reduced
This result is also consistent with the findings of El-Shahawi after medium exchange, being a major drawback of the pro-
and Kamal [34], who found a broad peak at −0.29 V versus posed method. Medium exchange is an approach to perform
Ag/AgCl for reduction of the complex Hg(II)-clorpyrifos. Other matrix exchange, so that the stripping step is performed in
compounds containing the phosphorothioate group (P = S) are absence of the sample matrix, eliminating potential interfer-
likely to suffer similar processes because of the interaction of ences that do not accumulate on the electrode, but that are
the sulfur atom with the surface of the mercury electrode [34], electroactive, producing cathodic currents during the poten-
so that the proposed methodology should be considered as not tial scan.
selective toward methyl parathion, but to this related class of The influence of square wave frequency was studied
pesticides. The deposition potential of −0.30 V was used in the between 10 and 500 Hz at a pulse height of 25 mV using the
methodology and in the other optimization experiments. SI-SWV method and a 0.50 mg L−1 methyl parathion solution
Optimization of the deposition time was made under flow in 40 mmmol L−1 BR buffer (pH 10) in 0.25 mol L−1 NaNO3 . The
rate of 10 ␮L s−1 during the passage of solution through the flow signal to noise ratio increased significantly up to the frequency
cell. Sample volumes ranging from 100 to 900 ␮L were injected of 250 Hz (Fig. 4c), which was adopted as the frequency to be
toward the working electrode to provide deposition times from used in the proposed method.

3.2. Figures of merit

The equation for the analytical curve obtained by in-

line dilution of a 0.50 mg L−1 methyl parathion solution
was Ip = (−9.7 ± 0.1)Cm-parathion + (0.06 ± 0.03), with Ip being
expressed in ␮A and the methyl parathion concentration
(Cm-parathion ) in mg L−1 . The linear relation between peak
current and methyl parathion concentrations was observed
in the range between 0.010 and 0.50 mg L−1 , with r2 > 0.999.
Homogenization of the monosegment is a key factor to allow
in-line and off-line statistical parameters of the analytical
curves to coincide. To perform this task, fractioning of the
zones in small volumes [25] followed by three flow rever-
sals [35], from the holding coil to the homogenization tube,
from the homogenization tube back to the holding coil and
finally from the holding coil to the detection flow cell. Addi-
tionally, homogenization is also achieved by the turbulence
promoted by changing the internal diameter of holding coil
Fig. 5 – Square wave voltammograms obtained by in line and homogenization tube. The previously proposed SI-SWV
dilution of a buffered 0.50 mg L−1 methyl parathion at a methodologies have used polyethylene homogenization coils
frequency of 250 Hz and pulse height of 25 mV, under [24,27], but in the present case, systematic low recoveries
stopped flow conditions, after concentrating the analyte at were observed in comparison with off-line prepared stan-
a deposition potential of −0.3 V for 30 s at a flow rate of dards. After changing the polyethylene homogenization coil
10 ␮L s−1 . Methyl parathion concentrations generated by a glass tube, indistinguishable statistical parameters were
inside the monosegmented were (a) 0; (b) 0.010; (c) 0.025; (d) obtained for the analytical calibration curves prepared either
0.050; (e) 0.10; (f) 0.25 and (g) 0.50 mg L−1 . The inset shows in-line or off-line. A possible explanation for this fact is
the analytical curve obtained by plotting the peak currents adsorption of methyl parathion onto the polyethylene surface.
read at −540 mV (against Ag/AgCl, KCl saturated reference Repeatability of the method was evaluated at two con-
electrode) vs. methyl parathion concentrations. Error bars centration levels, 0.050 and 0.50 mg L−1 , resulting in mean
correspond to the standard deviations of three peak current values of −0.52 ± 0.02 and −5.0 ± 0.2 ␮A, lead-
experiments. ing to variation coefficients of 4.5 and 3.7%, respectively.
 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216 215

insectide, but not containing the P = S group, decreased the

Table 2 – Effect of some possible interference on the peak
current of a 0.20 mg L−1 methyl parathion solution in methyl parathion signal by a factor of 33%, probably because
40 mmol L−1 BR buffer (pH 10) and 0.25 mol L−1 NaNO3 , adsorption and passivation of the Hg surface. Because the
according the procedure described in Section 2.2.1 sensitivity decreases if the medium is exchanged, the strip-
Interference Interference Peak current ping step is performed in presence of sample solution. As a
compound concentration (mg L−1 ) variation (%) consequence, potential interferences are expected for nitro
compounds such as methyl paraoxon and p-nitro phenol,
Parathion 0.2 +109
which are products of methyl parathion breakdown in the
Dichlorvos 0.2 −33
Picloram 0.2 −6
environment. These compounds do not accumulate on the Hg
Endosulfan 0.2 −5 electrode, but remain in the diffusion layer during the cathodic
Gliphosate 0.2 −3 scan, being reduced to hydroxylamine and causing positive
2,4-D 0.2 −2 interference.
Paraquat 0.2 −5 Interference by organochlorine pesticides (picloram, 2,4-D,
Atrazine 0.2 +5
and endosulfan) was not statistically significant since the peak
Humic Acid 5.0 −100
current deviations caused by these compounds did not exceed
Fulvic Acid 5.0 −67
Cu2+ 0.050 −2 the variation coefficient of the measurements at the 95% con-
Pb2+ 0.050 −3 fidence level. Similarly, no interference was observed for the
Cd2+ 0.050 −5 herbicides glyphosate, atrazine and paraquat. Humic and ful-
Zn2+ 0.050 −6 vic acid at concentration of 5 mg L−1 depleted the methyl
Mn2+ 0.050 +3 parathion signal, probably as a consequence of electrode pas-
Fe3+ 0.100 +2
sivation. Interference of humic substances is a typical case of
matrix interference that can be corrected by the standard addi-
Fig. 5 shows the voltammograms obtained by in-line dilu- tion approach. Metal cations were investigated since they can
tion of the 0.50 mg L−1 methyl parathion stock solution. be either reduced at the Hg mercury, especially Pb2+ and Cd2+ ,
The limit of detection (LOD) was obtained from the equa- at potentials close to the one observed for methyl parathion,
tion LOD = 3 × (Sb/m), and the quantification limit (LOQ) was or interact with the S atom of the P = S group, interfering in
obtained from LOQ = 10 × (Sb/m), where Sb is the standard the accumulation step. As one can see from Table 2, none of
deviation of current measurements of the blank solution at the metal cations caused statistically significant interference
the potential range between −542 and −590 mV (Fig. 4), and on the analytical signal of methyl parathion.
m is the slope of the analytical curve. In deionized water, LOD
and LOQ were 0.002 and 0.007 mg L−1 , respectively, so that the
detectability of the proposed method is suitable to monitor the 3.4. Methyl parathion recovery from spiked water
Drinking Water Equivalent Level (DWEL) for methyl parathion, samples
which is 7 ␮g L−1 [4].
Four samples of natural surface waters were analyzed, but
3.3. Selectivity none of them contained detectable amounts of the insecticide,
so that spiking was performed providing methyl parathion
The selectivity of the proposed method was evaluated by concentrations of 0.10 mg L−1 . The determinations were per-
performing the measurements with a 0.20 mg L−1 methyl formed by the proposed SI-SWV standard addition and in-line
parathion solution in presence of several other pesticides buffering protocols. Table 3 shows that the recoveries were
(Table 2). Ethyl parathion, another organophosphorus insec- between 82 and 97%, denoting reasonable accuracy for the pro-
ticide, interferes as a consequence of its chemical and posed method, a fact that was also confirmed by the results
electrochemical similarity to methyl parathion. This fact obtained by HPLC, for which the recoveries were between 89
can be extended to other compounds such as fenitrothion and 98%. Because the slope of the standard addition curves did
and isocarbophos. In these cases the proposed method is not differ significantly from the slope of the in-line prepared
not able to distinguish among the different compounds, so calibration curve, the samples could be directly analyzed by
that chromatographic or chemometric approaches have to external calibration, leading to higher sampling throughput.
be considered [17,36]. Dichlorvos, another organophosphorus Results reported in Table 3 were obtained from triplicate deter-

Table 3 – Recovery of methyl parathion by the proposed SI-ACSV and HPLC for 0.10 mg L−1 spiked water samples
Samples SI-ACSV HPLC |t|

Found (mg L−1 ) Recovery (%) Found (mg L−1 ) Recovery (%)

B 0.097 ± 0.002 97 0.092 ± 0.001 92 2.45

G 0.089 ± 0.006 89 0.097 ± 0.001 97 1.38
PN 0.0820 ± 0.0002 82 0.089 ± 0.009 89 1.16
SG 0.097 ± 0.002 97 0.098 ± 0.004 98 0.42

B: Billings; G: Guarapiranga; PN: Praia dos Namorados; SG: Salto Grande. Critical t-value = 2.78 (P = 0.05 and four degrees of freedom).
216 a n a l y t i c a c h i m i c a a c t a 6 0 6 ( 2 0 0 8 ) 209–216

minations by SI-SWV and by the HPLC method, resulting references

a critical t-value of 2.78 (P = 0.05) for 4 degrees of freedom.
As all the experimental t-values (Table 3) were smaller than
2.78, there is no evidence of systematic errors between the [1] S.J. Garcia, A.W. Abu-Qare, W.A. Meeker-O’Connell, A.J.
SI-SWV and HPLC results. These results suggest that the Borton, M.B. Abou-Donia, J. Toxicol, Environ. Health, Part B 6
proposed method has potential application in automatic sam- (2003) 185.
pling stations, being able to detect spills or run-off events of [2] X.F. Guo, U. Jans, Environ. Sci. Technol. 40 (2006) 900.
[3] J. Liu, L. Wang, L. Zheng, X.R. Wang, F.S.C. Lee, J. Chromatogr.
organophosphorus pesticides to aquatic environments.
A 1137 (2006) 180.
Because in-line conditioning and in-line standard addition [4] http://www.epa.gov/waterscience/criteria/drinking/
approaches require a 1:1 sample dilution to obtain suitable dwstandards.pdf, accessed in May, 2007.
homogenization, the minimum original sample concentra- [5] S. Lacorte, D. Barcelo, Anal. Chim. Acta 296 (1994) 223.
tion that can be analyzed (within the linear dynamic range) is [6] R.C. Martinez, E.R. Gonzalo, M.J.A. Moran, J.H. Mendez, J.
20 ␮g L−1 . To evaluate the performance of the method at con- Chromatogr. 607 (1992) 37.
[7] T. Perez-Ruiz, C. Martinez-Lozano, V. Tomas, J. Martin, Anal.
centrations closer to the quantification limit, three samples
Chim. Acta 540 (2005) 383.
of drinking water were spiked with 10 ␮g L−1 . These sam-
[8] T. Galeano-Diaz, A. Guiberteau-Cabanillas, N. Mora-Diez, P.
ples were off-line buffered, at a 9:1 volumetric ratio, with a Parrilla-Vazquez, F. Salinas-Pópes, J. Agric. Food Chem. 48
400 mmol L−1 BR buffer (pH10) in 2.5 mol L−1 NaNO3 , and ana- (2000) 4508.
lyzed by the external calibration curve. Recoveries between 96 [9] S. Lacorte, D. Barcelo, Anal. Chem. 68 (1996) 2464.
and 115% were observed, indicating the good accuracy of the [10] M. Schellin, B. Hauser, P. Popp, J. Chromatogr. A 1040 (2004)
proposed method. 251.
[11] D.A. Lambropolou, V.A. Sakkas, T.A. Albanis, Anal. Bioanal.
Chem. 374 (2002) 932.
4. Conclusion [12] D.A. Lambropolou, T.A. Albanis, J. Chromatogr. A 922 (2001)
243.
[13] D.A. Lambropolou, T. Sakellarides, T.A. Albanis, Fresenius J.
This paper demonstrates the feasibility of sequential injec-
Anal. Chem. 368 (2000) 616.
tion analysis to automate adsorptive cathodic square wave
[14] M.R. Driss, M.L. Bouguerra, Intern. J. Environ. Anal. Chem. 65
stripping voltammetry for determination of methyl parathion (1996) 1.
in water samples. The proposed methods consume only [15] D. de Souza, S.A.S. Machado, Electroanalysis 18 (2006) 862.
400–450 ␮L of sample per analysis. This small sample vol- [16] G. Liu, Y. Lin, Electrochem. Commun. 7 (2005) 339.
ume allows one to obtain high enrichment factors if solid [17] Y. Ni, P. Qiu, S. Kokot, Anal. Chim. Acta 516 (2004) 7.
phase extraction protocols are used to concentrate the ana- [18] M.A. Kumar, R.S. Chouhan, M.S. Thakur, B.E.A. Rani, B.
Mattiasson, N.G. Karanth, Anal Chim. Acta 560 (2006) 30.
lyte, as well as to eliminate interferences. Sampling frequency
[19] P. Mulchandani, W. Chen, A. Mulchandani, Environ. Sci.
of 61 h−1 is possible for off-line buffered samples that do not Technol. 35 (2001) 2562.
exhibit significant matrix effects. The method is favored by [20] Z.M. Rao, J.N. Wang, L.D. Li, X.R. Zhang, Chin. J. Anal. Chem.
the advantages of the hanging mercury drop electrode, which 29 (2001) 373.
allows one to use a fresh drop in each injection, overcoming [21] A. Romanus, H. Muller, D. Kirsch, Fresenius J. Anal. Chem.
limitation of memory effects that could occur as a conse- 340 (1991) 371.
[22] W.W. Kubiak, R.M. Latonen, A. Ivaska, Talanta 53 (2001)
quence of methyl parathion adsorption during the deposition
1211.
step. Other advantages of the proposed method are the low
[23] G. Abate, J. Lichtig, J.C. Masini, Talanta 58 (2002) 433.
consumption of reagents and minimal generation of wastes. [24] L.B.O. dos Santos, M.S.P. Silva, J.C. Masini, Anal. Chim. Acta
The used Hg may be easily recovered and reused in the elec- 258 (2005) 21.
trode capillary. The run time of an HPLC determination is [25] B.F. Reis, M.F. Giné, E.A.G. Zagatto, J.L.F.C. Lima, R. Lapa,
about 3.0 min, at a flow rate of 1.2 mL min of 50% (v:v) ace- Anal. Chim. Acta 293 (1994) 129.
tonitrile:water mobile phase, implying in a consumption of [26] J.A. Vieira, I.M. Raimundo Jr., B.F. Reis, E.A.G. Zagatto, J.L.F.C.
Lima, Anal. Chim. Acta 366 (1998) 257.
1.8 mL of acetonitrile, in addition to the column conditioning
[27] L.B.O. dos Santos, J.C. Masini, Talanta 72 (2007) 1023.
time (∼30 min) before starting the serial analyses. On the other
[28] T.M. Florence, J. Electroanal. Chem. 97 (1979) 219.
hand, the proposed method has poor selectivity, being liable [29] C.L. Bourque, M.M. Duguay, Z.M. Gautreau, Intern. J. Environ.
to interference caused by other organophosphorus pesticide Anal. Chem. 37 (1989) 187.
and nitro compounds, so that isolation of the analyte may be [30] M.R. Smyth, J.G. Osteryoung, Anal. Chim. Acta 96 (1978)
required for analysis of complex sample matrices. 335.
[31] L.Z. Zelayaran Melgan, S.A.S. Machado, J. Braz. Chem. Soc 16
(2005) 743.
Acknowledgements [32] J.A. Manzanilla-Cano, E. Reyes-Salas, M.H. Barceló-Quintal,
Intern. J. Environ. Anal. Chem. 75 (1999) 387.
[33] G.M. Castanho, C.M.P. Vaz, S.A.S. Machado, J. Braz. Chem.
Authors are grateful to Fundação de Amparo à Pesquisa
Soc. 14 (2003) 594.
do Estado de São Paulo (FAPESP) and to Conselho Nacional
[34] M.S. El-Shahawi, M.M. Kamal, Fresenius J. Anal. Chem. 362
de Desenvolvimento Cientı́fico e Tecnológico (CNPq) for the (1998) 344.
financial support and fellowships. The Ph.D. students Allan [35] T. Gubeli, G.D. Christian, J. Ruzicka, Anal. Chem. 63 (1991)
Cezar Vieira dos Santos and Carlos Martı́n Infante Córdova 2407.
are acknowledged for their helpful assistance during the water [36] P. Qiu, Y.N. Ni, S. Kokot, Anal. Let. 39 (2006) 1967.
sampling.