877 ANALYTICAL SCIENCES JULY 2008, VOL. 24 877 2008 © The Japan Society for Analytical Chemistry Introduction Arsenic exists in several chemical… [600242]

877 ANALYTICAL SCIENCES JULY 2008, VOL. 24 877
2008 © The Japan Society for Analytical Chemistry
Introduction
Arsenic exists in several chemical forms in the nature, both
organic and inorganic. Its toxicity depends on the chemical
form, inorganic arsenic(III) being the most toxic of the water-
soluble species. Arsenic is a ubiquitous trace element
originating from natural and anthropogenic sources. Hence it is
important to develop a rapid, selective and sensitive method for
the arsenic determination and speciation.
A number of analytical methods are used for the arsenic
determination and speciation in environmental samples. The
speciation of arsenic is mostly based on the determination of total
arsenic after As(V) prereduction and determination of As(III) alone.
Some of the methods used are flow injection hydride generation
atomic absorption spectrometry (FI-HG-AAS),1–3 HG-AAS,4–6
sequential injection hydride generation atomic fluorescence
spectrometry (SIA-HG-AFS),7 flow injection inductively coupled
mass spectrometry (FI-ICP-MS),8,9 electrochemical methods,10–12
pervaporation-flow injection determination with photometric
detection,13 spectrophotometrical,14–16 colorimetric17,18 and FIA with
chemiluminiscence detection.19 Most methods are based on a
hydride generation in an acidic media; according to the detection
methods, they vary in the detection limits, as well as solution
and sample consumption and the expense of the analysis itself.
Farrell et al. have reported the determination of arsenic by
hydride generation gas diffusion flow injection analysis with
electrochemical detection.20 Their electrochemical detector was
equipped with the gold electrode, their acceptor solution was
0.01 mol/dm3 H2SO4 and their reagent was 0.5% NaBH4. Their
limit of detection of was 10 mg/dm.3
An indirect gas-diffusion FIA method has been developed for
the selective and sensitive determination of tetrahydroborate.21 The precision of the method was better than a relative standard
deviation of 2.1% at 60 mmol/dm3 levels and better than 0.5% at
0.1 mmol/dm3, with a throughput of 60 samples per hour.
In this research a gas-diffusion FIA method was developed
and applied for arsenic determination. The method is based on
arsine generation by NaBH4. The formed arsine diffuses through
the PTFE (polytetrafluoroethylene) membrane and is quantified
amperometrically. Our amperometrical detector was equipped
with a platinum working electrode; by using iodine/iodide
acceptor solution we achieved low limit of detection (5 mg/dm3).
FIA experiments were confirmed by cyclic voltammetry. This
optimized system was applied for arsenic determination in some
power plant waste water samples and the results were compared
with those obtained by the HG-AAS technique.
Experimental
Reagents and chemicals
All the chemicals used were of analytical reagent grade.
Deionized water, distilled over potasium permanganate, was
used throughout.
A stock arsenic solution was prepared weekly. Sodium
arsenite (Carlo Erba, Italy) was dissolved in 6 mol/dm3 HCl and
then diluted, yielding appropriate concentration. Each stock
solution concentration was tested bromatometrically. Sodium
borhydride was used as the reduction agent (Riedel de-Haen,
Germany). A stock solution was prepared daily in 0.1% sodium
hydroxide (Lachema, Czech Republic). Stock iodine solution
was prepared weekly by dissolving iodine (Zorka, Serbia),
refined by sublimation in potassium iodide (Carlo Erba).
Arsenic(V) solution (Carlo Erba), Sb(III), Sn(II) (Fluka,
Germany) and Se(IV) (Schering Kahlbaum, Germany) were
prepared in 6 mol/dm3 hydrochloric acid, and were diluted to the
appropriate concentrations.
Appropriate volumes of waste water samples were acidified Optimization of a Flow Injection System with Amperometric
Detection for Arsenic Determination
Aleksandar LOLIC,† Snezana NIKOLIC, and Jelena MUTIC
Faculty of Chemistry, University of Belgrade, Studentski trg 16, P. O. Box 158, 11000 Belgrade, Serbia
A selective and sensitive analytical procedure for rapid arsenic determination by gas-diffusion flow injection analysis with
amperometric detection was developed. The method is based on the arsenite reduction by NaBH4. Derived arsine diffuses
through a PTF membrane into the acceptor flow stream and is amperometrically determined on a platinum working
electrode. The limit of detection (3s) at room temperature was 5 mg/dm3 of As(III). The relative standard deviation for a
1 mg/dm3 As(III) standard was 1.96% for six repetitive injections. Arsenic(V) was determined after its prereduction with
potassium iodide. Arsenic determination was not interferred with by 1 mg/dm3 Sb(III), 5 mg/dm3 Sn(II), 10 mg/dm3
Se(IV), 1 mg/dm3 As(V), 1 mg/dm3 hydrasine, 1 mg/dm3 Fe(II) or 0.5 mg/dm3 Fe(III) solution. The throughput of this
method was 60 analyses per hour. This method was successfully applied to arsenic determination in some power plant
waste water samples.
(Received October 25, 2007; Accepted January 28, 2008; Published July 10, 2008)

† To whom correspondence should be addressed.
E-mail: lolix@chem.bg.ac.yu

878 ANALYTICAL SCIENCES JULY 2008, VOL. 24
with hydrochloric acid (0.1 mol/dm3); the samples were set on
an ultrasonic bath for 30 min and then injected into the flow
system.
Instruments and apparatus
Figure 1 presents the FIA manifold used for the determination
of As(III). Three peristaltic pumps were used. One pump was a
Model HPB 5400 (Iskra, Slovenia), the second one was a Model
MS Reglo (Ismatec, Switzerland) with flow rate control unit and
the third pump was a Model Mini S-840 (Ismatec). The
injection valve was a Model 5020 (Rheodyne, USA), equipped
with a 0.2-cm3 sample loop. The gas-diffusion unit was
manufactured after the model provided by Shenyang Film
Projector Factory, China. All connections were made with 0.5
mm i.d. tubing. Due to oxide film formation on its surface, the
platinum electrode was occasionally cleaned by polishing with
fine alumina pasta. Working potential was regulated by
polarogaph MA 5450, Iskra, Slovenia. The obtained FIA signals
were recorded on the Servograph Model 61 (Radiometer,
Denmark) writer. Temperature was regulated with a thermostat,
Model Messgerate-Werk Lauda, Germany.
Cyclic voltammetric experiments were registered on a Metrohm
797 VA Computance instrument (Herisau, Switzerland). The
triple electrode system consisted of: working electrode, rotating
platinum disk (RDE); reference electrode, Ag/AgCl in potassium
chloride (3 mol/dm3); and auxiliary platinum electrode. The PC
software provided controls for the measurement, recording the
measuring data and evaluating it.
Results and Discussion
In order to achieve optimal conditions for the arsenic
determination we investigated the effects of a few parameters.
The first effect to be investigated was the optimal potential.
Arsenic standard was injected at several potential values in the
range –0.10 to +0.20 V versus Ag/AgCl reference electrode. The
concentration of the standard was 1 mg/dm3 in 0.1 mol/dm3 HCl.
Reagents were 0.2% NaBH4 in 0.1% NaOH, the carrier was 0.1
mol/dm3 HCl, the acceptor solution was 1 mg/dm3 I2 in 0.05
mol/dm3 KI. From the obtained hydrodynamic voltammogram
(Fig. 2) the highest peak was reached at +0.10 V. Figure 3
presents obtained cyclic voltammograms that gave us another
proof for choosing the potential for further experiments.
The next effect to be investigated was the carrier solution.
Several solutions were chosen as carriers: sulfuric, perchloric and hydrochloric acid. Sulfuric acid was dismissed as it gave a
signal for blank injections, probably because of its impurities.
When perchloric acid was used, a signal split occurred. Since
the highest and the proper signals were obtained when
hydrochloric acid was injected, it was chosen to be the carrier.
Further experiments with hydrochloric acid as the carrier
showed that 0.1 mol/dm3 HCl gave the best results (Fig. 4).
Examination of NaBH4 effect as a reagent showed that 0.1%
NaBH4 in 0.1% NaOH gave the highest signals (Fig. 5). Using
less concentrated reagent we avoided H2 production, which may
cross the membrane with arsine and thus cause greater
dispersion in the streams, which leads to decrease of the peak
height signal.
One of the important effects was the mixing coil volume.
When arsenic standard was injected with different coils (30, 60
and 130 cm in length) and a tube (10 cm in length), peak
currents obtained were 1.22, 0.89, 0.74 and 0.54 mA,
respectively. Hence for further experiments, the shortest coil
(30 cm) was used.
The acceptor stream solution components were investigated
by injecting arsenic standard solution with several acceptor
solutions. The examined solutions were sulfuric acid, potassium
hexacyanoferrate(III) and iodine dissolved in potassium iodide.
When sulfuric acid was used as the acceptor, there was no signal
for the arsenic standard solution. When a ferro salt was used a
signal for the blank injection was noticed. The proper signals
were obtained for iodine in potassium iodide solution (0.5 and 1
mg/dm3 I2 in 0.05 mol/dm3 KI). To prove why iodine/iodide
solution was used as the acceptor we recorded cyclic
voltammograms (Fig. 3). The first voltammogram was recorded
for the acceptor solution alone (1); when 1 mg/dm3 As(III)
standard was injected, voltammograms 2 – 5 were recorded. The
obtained voltammograms clearly show why the iodine/iodide
solution was chosen as the acceptor solution (Fig. 3a) compared
to 1 mg/dm3 potassium hexacyanoferrate(III) in 0.01 mol/dm3
NaCl (Fig. 3b) and 0.01 mol/dm3 H2SO4 (Fig. 3c). At a potential
of +0.100 V, there is a signal for iodine/iodide reduction (Fig.
3a) which decreases when As(III) is added.
Effects of stream directions and rates were also investigated.
In an optimized system, the opposite direction gave a peak
current of 0.21 mA, whereas for the parallel direction the peak
current was 0.18 mA for 1 mg/dm3 As(III). With the flow rate
increase from 0.5 to 0.9 cm3/min, identical peak currents were
obtained for the opposite direction. In this case compensation
of two effects takes place: the slower donor rate increases the
Fig.1Schematic presentation of the FIA apparatus used for
amperometric determination of arsenic content. (C) Carrier; (R)
reagent; (A) acceptor flow stream; (P) peristaltic pump; (V) injection
valve; (MC) mixing coil; (GDU) gas-diffusion unit; (FC)
electrochemical flow through cell; (PO) potentiostat; (RE) writer;
(W) waste. Numbers present flow rates in cm3/min.
Fig.2Hydrodynamic voltammogram for a 0.200-cm3 injection of a
1 mg/dm3 arsenic(III) standard.

879 ANALYTICAL SCIENCES JULY 2008, VOL. 24
AsH3 amount that passes through PTFE membrane and at the
same time the higher acceptor rate increases the base current
caused by iodine reduction, causing significant peak current
change with As(III) injection. For further experiments, a higher
acceptor flow rate was used (0.9 cm3/min) because it enables
better throughput (60 samples per hour). The donor flow rate
was kept constant at 1.1 cm3/min.
Temperature effects were studied by injecting the same As(III)
standard (1 mg/dm3), while varying the temperature in the
interval 20 – 50˚C. It was observed that there was no signal
change with temperature increase, which can be explained as
due to the high As(III) reduction rate, its volatility and its
diffusion through the PTFE membrane.
Possible interferants of this determination were Se, Sn and Sb,
since they form volatile hydrides as well as some other wastewater components such as hydrasine, Fe(II) and Fe(III).
We injected solutions containing various ratios of interferants
and As(III), whose concentration was kept constant (1 mg/dm3).
Results showed that 10 mg/dm3 Se(IV) and 5 mg/dm3 Sn(II) did
not interfere with the determination, neither did 1 mg/dm3 Sb(III)
and 1 mg/dm3 As(V). Hydrasine of the same concentration as
As(III) did not interfere, with the As(III) signal, nor did Fe(II),
whereas Fe(III) did increase the signal for 50%; but when we
injected 0.5 mg/dm3 Fe(III) solution, no effect was noticed.
Our proposed method could be used for speciation of As(III)
and As(V). A set of experiments was performed. Standard
solutions containing only As(III) or As(V) or their mixtures
were injected; each solution was treated with potassium iodide
for As(V) prereduction. A 1-cm3 volume of 0.02 mol/dm3
potassium iodide was added to 100.00 cm3 of working solution
containing As(III) 30 min prior to its injection. The results
obtained are presented in Table 1. It is important to highlight
that there was no significant interferring from potassium iodide
and iodine excess.
The limit of detection of this flow-injection method was 5
mg/dm3 of arsenic (3s), which corresponds to 100 pg of As(III)
(the sample loop volume was 0.2 cm3).
The reproducibility of the analytical system was investigated
by six repetitive injections of the arsenic standard. The relative
standard deviation for a 1 mg/dm3 As(III) standard was 1.96%.
Linearity was investigated by triple injection of standard As(III)
5 ¥ 10–6A5 ¥ 10–6A5 ¥ 10–6A
Fig.3Cyclic voltammograms obtained at 1 mm Pt electrode at u =
100 mV/s. a) Cyclic voltammograms of 1 mg/dm3 I2 in 0.05 mol/dm3
KI, 0.1% NaBH4 in 0.1% NaOH, 0.1 mol/dm3 HCl (1) and with
1 mg/dm3 As(III) injection (2 – 5); b) 1 mg/dm3 potassium
hexacyanoferrate(III) in 0.01 mol/dm3 NaCl, 0.1% NaBH4 in 0.1%
NaOH, 0.1 mol/dm3 HCl (1), and with 1 mg/dm3 As(III) injection (2 –
5); c) 0.01 mol/dm3 H2SO4, 0.1% NaBH4 in 0.1% NaOH, 0.1 mol/dm3
HCl (1), and with 1 mg/dm3 As(III) injection (2 – 5).
Fig.4Effect of hydrochloric acid concentration as a carrier.
Fig.5Effect of tetrahydroborate as a reagent.

880 ANALYTICAL SCIENCES JULY 2008, VOL. 24
solutions; it spans over two ranges, 0.1 – 1.0 mg/dm3 and 1.0 –
10.0 mg/dm3; the corresponding correlation coefficients were
r = 0.9944 and 0.9996.
Our optimized system was applied to arsenic determination in
the waste water samples from the power plant “Kolubara”,
Lazarevac, Serbia. Table 2 represents the obtained results,
which were compared to those obtained by the reference
technique HG-AAS. The reference results were obtained on a
atomic absorption spectrometer, Perkin-Elmer 2380, equipped
with hydride generator unit MHS-20. The determinations were
carried out at 193.7 nm, slit 0.7 and lamp current of 10 mA.
Conclusion
A rapid, indirect gas-diffusion flow injection method with
amperometric detector has been developed for selective arsenic
determination. Several physical and chemical parameters were
investigated. Their optimal conditions were: the potential of the
working Pt electrode (+0.10 V vs. Ag/AgCl); the hydrochloric
acid as the carrier (0.1 mol/dm3); the acceptor solution 0.5 and 1
mg/dm3 I2 in 0.05 mol/dm3 KI; the reagent solution 0.1% NaBH4
in 0.1% NaOH; both parallel and opposite direction of acceptor
and carrier streams; and room temperature.The optimized system had good reproducibility (RSD for a 1
mg/dm3 As(III) standard was 1.96% for six repetitive injections),
was sensitive (limit of detection (3s) was 5 mg/dm3, which
corresponds to 100 pg of As(III), the sample loop volume was
0.2 cm3) and was fast with a throughput of 60 samples per hour.
The method was succesfully applied for the arsenic
determination in the waste water samples from the power plant
“Kolubara” near Lazarevac, Serbia; the accuracy was confirmed
by HG-AAS technique.
References
1. N. M. M. Coelho, A. C. da Silva, and C. M. da Silva, Anal.
Chim. Acta, 2002, 460, 227.
2. A. Shraim, B. Chiswell, and H. Olszowy, Talanta, 1999, 50,
1109.
3. J. Y. Cabon and N. Cabon, Anal. Chim. Acta, 2000, 418, 19.
4. C. J. Hsieh, C. H. Yen, and M. S. Kuo, Anal. Sci., 1999, 15,
669.
5. P. K. Petrov, I. Serafimovski, T. Stafilov, and D. L. Tsalev,
Talanta, 2006, 69, 1112.
6. L. Zhang, Y. Morita, K. Yoshikawa, and A. Isozaki, Anal.
Sci., 2007, 23, 365.
7. N. V. Semenova, F. M. Bauzá de Mirabó, R. Forteza, and V.
Cerdá, Anal. Chim. Acta, 2000, 412, 169.
8. L. F. R. Machado, A. O. Jacintho, A. A. Menegário, E. A. G.
Zagatto, and M. F. Giné, J. Anal. At. Spectrom., 1998, 13,
1343.
9. X. P. Yan, R. Kerrich, and M. J. Hendry, Anal. Chem., 1998,
70, 4736.
10. P. Salaun, B. Planner-Friedrich, and C. M. G. van den Berg,
Anal. Chim. Acta, 2007, 585, 312.
11. Y. Song and G. M. Swain, Anal. Chim. Acta, 2007, 593, 7.
12. M. S. S. Pereira, E. Winter, J. R. Guimarães, S. Rath, and
A. H. Fostier, Environ. Chem. Lett., 2007, 5, 137.
13. T. Rupasinghe, T. J. Cardwell, R. W. Cattrall, M. D. Luque de
Castro, and S. D. Kolev, Anal. Chim. Acta, 2001, 445, 229.
14. H. D. Revanasiddappa, B. P. Dayananda, and T. N. K.
Kumar, Environ. Chem. Lett., 2007, 5, 151.
15. A. Baghel, B. Singh, P. Pandey, and K. Sehkar, Anal. Sci.,
2007, 23, 135.
16. M. Hashemi and P. Modasser, Talanta, 2007, 73, 166.
17. K. Morita and E. Kaneko, Anal. Sci., 2006, 22, 1085.
18. M. Rahman, Y. Seike, and M. Okumura, Anal. Sci., 2006,
22, 475.
19. S. Satienperakul, T. J. Cardwell, S. D. Kolev, C. E. Lenehan,
and N. W. Barnett, Anal. Chim. Acta, 2007, 554, 25.
20. J. R. Farrell, P. J. Iles, and Y. J. Yuan, Anal. Chim. Acta,
1996, 334, 193.
21. S. D. Nikolic, E. B. Milosavljevic, J. L. Hendrix, and J. H.
Nelson, Talanta, 1993, 40, 1283.Table1Determinatio n of As(III) and As(V) in standar d
solutions
As(III) added/
mg dm–3As(V) added/
mg dm–3Found/
mg dm–3 ± RSDaReco very,
%
10 0 9.8 ± 0.2 98
5 5 9.9 ± 0.2 99
0 10 9.8 ± 0.2 98
a. Results given are obtained as averages of three measurements ±
RSD.
Table 2 Results of arsenic determination in waste water
samples
Waste water sampl eArsenic(III) found/mg dm–3
FIAaHG-AASa
1 0.191 ± 0.004 0.195 ± 0.008
2 0.136 ± 0.003 0.130 ± 0.006
3 0.058 ± 0.001 0.060 ± 0.003
4 0.287 ± 0.006 0.280 ± 0.012
5 0.119 ± 0.002 0.120 ± 0.005
a. Results given are obtained as averages of three measurements ±
RSD.