ORIGINAL PAPER

Automated method for the determination of total arsenicand selenium in natural and drinking water by HG-AAS

Mariela Piston • Javier Silva •

Ramiro Perez-Zambra •

Isabel Dol • Moises Knochen

Received: 10 February 2010 / Accepted: 11 April 2011

� Springer Science+Business Media B.V. 2011

Abstract A multicommutated flow system was

designed and evaluated for the determination of total

arsenic and selenium by Hydride Generation Atomic

Absorption Spectrometry (HG-AAS). It was applied to

the determination of arsenic and selenium in samples

of natural and drinking water. Detection limits were

0.46 and 0.08 lg l-1 for arsenic and selenium, respec-

tively; sampling frequency was 120 samples h-1 for

arsenic and 160 samples h-1 for selenium. Linear

ranges found were 1.54–10 lg l-1 (R = 0.999) for

arsenic and 0.27–27 lg l-1 (R = 0.999) for selenium.

Accuracy was evaluated by spiking various water

samples and using a reference material. Recoveries

were in the range 95–116%. Analytical precision

(sr (%), n = 10) was 6% for both elements. Compared

with the Standard Methods, APHA, 3114B manual

method, the system consumes at least 10 times less

sample per determination, and the quantities of acid

and reducing agent used are significantly lower with

a reduction in the generation of pollutants and waste.

As an additional advantage, the system is very fast,

efficient and environmentally friendly for monitoring

total arsenic and selenium levels in waters.

Keywords Flow analysis � Selenium � Arsenic �Water � Trace element monitoring

Introduction

Arsenic and selenium are two widely distributed

semimetals which for different reasons have attracted

interest in connection with public health issues.

Arsenic has been known for centuries as a toxic

element: in popular imagination, it has become a

synonymous of poison. Besides the obvious behavior

of some of its compounds in acute exposure, there is an

increasing concern about the risk associated with long-

term exposure. There exists considerable evidence

suggesting that consumption of arsenic-containing

water can cause skin, bladder and lung cancer (Goyer

1996).

One of the main problems in some geographical

areas arises from groundwater contamination of either

mineral or anthropogenic origin. A number of chronic

intoxication episodes have occurred that have drawn

attention to the risk to which the population of certain

areas is exposed due to consumption of water from

contaminated wells. Thus, it has been classified as a

high-priority substance for screening in drinking water

sources. Maximum acceptable levels of total arsenic in

drinking water have been the object of some contro-

versy, although there is a trend toward the establish-

ment of lower values. WHO, in its Guidelines for

M. Piston (&) � J. Silva � R. Perez-Zambra �I. Dol � M. Knochen

Facultad de Quımica, Universidad de la Republica,

Catedra de Quımica Analıtica. Av. Gral. Flores 2124,

P.O. Box 1157, 11800 Montevideo, Uruguay

e-mail: mapiston@gmail.com

123

Environ Geochem Health

DOI 10.1007/s10653-011-9436-9

Drinking-water Quality (World Health Organization

2004), sets up the recommendation for a provisional

guideline value of 10 lg l-1. This value coincides with

other standards such as that set up by EPA (Environ-

mental Protection Agency 2001) and several regula-

tory bodies throughout the world. However, values as

high as 20 lg l-1 (Instituto Uruguayo de Normas

Tecnicas 2008) can still be found in some countries.

For drinking water sources, such as river or ground-

water, the maximum values set up are often lower

(\10 lg l-1).

Selenium on the other hand is now known as

necessary for life. It occurs in all environments

unevenly distributed throughout the planet; in some

places, selenium concentrations in soil are so low that

consumption of vegetables by animals or humans

cannot guarantee the recommended daily intake of

selenium.

In other areas, it can be found at high levels, for

example, in soils in very dry zones of USA where there

are plants with high concentrations of selenium

(Fordyce 2005).

In the 1950s, the essentiality of selenium in animals

and humans was revealed with the discovery of its

ability to prevent the Keshan disease, first described in

1935 in the Chinese province of that name. This

population showed symptoms attributed to selenium

deficiency: different degrees of heart disease or

cardiac deficiencies; myocardial histology showed

degeneration and necrosis of ultrastructure of muscle

fibers. Selenium is an essential component of enzymes

such as glutathione peroxidase that is found in human

tissues (Flynn and Cashman 1997).

Subsequently, its important role as part of the

body’s antioxidant mechanisms and the role in

protecting the body against heavy metals as well as

its importance in the immune system were discovered

(Tiglea and Mello De Capitani 2003; Fordyce 2005).

Related to the latter, the patients with AIDS show

evidence of selenium deficiency, such as a marked

decrease in the activity of glutathione peroxidase.

RDA (Recommended Dietary Allowance) accepted

value in the USA ranges from 20 lg/day for children

under 3 years of age up to 55 lg/day for adults. These

values can be increased to 60 lg/day for pregnant

women and 70 lg/day during lactation (Flynn and

Cashman 1997). These nutritional requirements should

be satisfied by food and water ingested by the

individual.

Selenium has one of the narrowest ranges between

dietary deficiency and toxic levels, so control in the

intake of selenium by humans and other animals is

imperative (Fordyce 2005). The WHO guideline value

for selenium in drinking water is 10 lg l-1 (World

Health Organization 2004).

In Uruguay, the government promotes the moni-

toring of some trace elements in waters through

different programs involving surface and groundwater

as well as drinking water.

The increasing interest in these trace elements has

expanded the need for reliable data, which puts

considerable pressure on environmental analytical

laboratories. This has triggered research for new and

improved automated analytical methods for the deter-

mination of ultra-trace levels of arsenic and selenium

in water samples. These methods should not only be

reliable but also fast and environmentally friendly.

Total arsenic and selenium at the trace and ultra-

trace levels are determined by atomic absorption

spectrometry (AAS) and atomic fluorescence spec-

trometry (AFS) (Standard Methods, APHA 1998;

Welz and Sperling 1999). For AAS determinations,

electrothermal atomization can be used (Aleixo and

Nobrega 2003; Paiva Oliveira et al. 2005; Hung et al.

2004; Michon et al. 2007). However, the use of

hydride generation (HG) (Dedina and Tsalev 1995) is

preferable since it provides separation of the analyte

from the matrix which in turn reduces the effects of a

number of interferences. Hydride generation can be

coupled to atomic absorption (HG-AAS) or to atomic

fluorescence (HG-AFS) spectrometric detection. The

latter provides lower detection limits at the cost of

more expensive instrumentation.

Hydride generation can be automated employing

different flow techniques such as flow injection

analysis (FIA) (Zhang and Adeloju 2008; Li et al.

2009) or sequential injection analysis (SIA) (Seme-

nova et al. 2000). Multicommutated flow analysis

(MCFA) is an emerging flow technique based on flow

networks built around solenoid valves which can be

commutated independently under computer control

in order to perform specific tasks (Reis et al. 1994;

Rocha et al. 2002; Catala Icardo et al. 2002; Feres et al.

2008). This technique has proven to be very flexible

since several modifications can be carried out simply

by changing parameters in the control software.

For instance, the sample volume can be easily changed

by modifying the time a given solenoid valve is

Environ Geochem Health

123

energized, thus allowing for different concentration

ranges. Since MCFA is a microanalytical technique,

methods based on it usually present low reactive

consumption and generate low volumes of chemical

residues. The literature reflects the usefulness of this

technique that has been applied to the determination of

different analytes in a wide range of matrixes (Kno-

chen et al. 2005; Piston et al. 2006; Sixto and Knochen

2009; Piston et al. 2009).

The goal of this work was to develop a fast, efficient

and environmentally friendly method for the determi-

nation of total arsenic and selenium in water samples.

For this purpose, an MCFA system with HG-AAS

detection was designed. The analytical method was

validated and applied to the determination of arsenic

and selenium in samples of natural and drinking water.

Materials and methods

All reagents used in sample treatment were of

analytical reagent grade or better.

Sodium tetrahydroborate (Hydride-Generation

grade) was obtained from Fluka (Buchs, Switzerland).

0.5% (w/v) and 1.2% (w/v) solutions were prepared

daily by dissolving the solid in 0.05% (w/v) sodium

hydroxide.

Purified water (ASTM Type I) was obtained from a

Millipore (Sao Paulo, Brazil) Simplicity 185 purifier

fed with glass-distilled water.

A selenium standard solution (1,000 mg l-1) was

prepared from selenium metal (Aldrich, St. Louis,

MO, USA, 99.99%) dissolved in nitric acid and made

up to volume with 10% (v/v) hydrochloric acid.

An intermediate standard solution (0.8 mg l-1) was

prepared daily by stepwise dilution with 1.5% (v/v)

hydrochloric acid.

An arsenic standard solution (1,000 mg l-1) was

prepared from arsenic trioxide (Sigma–Aldrich, Pri-

mary standard grade 99.95%) dissolved in 0.4% (w/v)

sodium hydroxide solution. An intermediate standard

solution (0.6 mg l-1) was prepared by dilution with

1.5% (v/v) hydrochloric acid.

All glassware were soaked overnight in 10% (v/v)

nitric acid and then rinsed exhaustively with purified

water.

Samples of surface water, groundwater and drinking

water were obtained from different locations of

Uruguay. For arsenic, the calibration curve in the

range 2.0–15 lg l-1 was prepared by accurately mea-

suring 0.05-, 0.1-, 0.2- and 0.3-ml of the 0.6-mg l-1

intermediate standard solution, followed by the addition

of 10 ml of water, 0.2 ml of 7% (v/v) sulfuric acid and

0.4 ml of 5% (w/v) potassium persulfate. The mixture

was heated on a hot plate for 1 h at gentle boiling and

then cooled down at room temperature. Afterward, 1 ml

of 7% (w/v) sodium iodide and 1 ml of concentrated

hydrochloric acid were added to carry out the pre-

reduction of As(V) to As(III), followed by dilution to

15 ml with water.

For selenium, calibration solutions in the range

1.0–10.0 lg l-1 were prepared by accurately measur-

ing 0.05-, 0.1-, 0.2-, 0.3- and 0.4-ml aliquots of the

0.8 mg l-1 intermediate standard solution, to which

20 ml of water and 10 ml of concentrated hydrochlo-

ric acid were added. The mixture was heated on a hot

plate for 1 h at gentle boiling to carry out the pre-

reduction of Se(VI) to Se(IV) and then cooled down to

room temperature and diluted to 30.0 ml with water.

Water samples (10 ml) were processed similarly to

the standards according to the procedures described

above for the respective elements.

MCFA system

The flow system was described elsewhere (Piston et al.

2009). It was based upon a Gilson (Villiers-le-Bel,

France) Minipuls 2 multichannel peristaltic pump and

two 3-way 12 V solenoid valves (NResearch, West

Caldwell, NJ, USA, model 161T031) which were used

for fluid control. In order to avoid unnecessary

consumption, sample and NaBH4 solution were recy-

cled to the bottle when not used.

Nitrogen (dried and purified by a combined

Drierite/molecular sieve trap) was used as carrier gas.

All measurements were taken on a Perkin-Elmer

(Norwalk, CT, USA) model 5000 atomic absorption

spectrometer, fitted with a 10-cm burner (air-acetylene

flame) and a T-shaped quartz atomization cell (Preci-

sion Glassblowing, Centennial, CO, USA) and oper-

ated at 193.7 nm (As) or 196.0 nm (Se). The slit width

was 0.7 nm for both determinations. Superlamp�

intensified emission hollow cathode lamps (Photron,

Narre Warren, Australia) were used as light sources

and operated as recommended by the manufacturer.

For the operation of the system, a program was

compiled in Visual Basic 6.0 (Microsoft) using the

Environ Geochem Health

123

Software 3.1 (Measurement Computing) graphical

programming environment. The program controlled

the timing and the activation of the solenoid valves

using the parameters set up before the beginning of the

analysis and also handled the data acquisition, visual

display and storage on hard disk.

Results and discussion

The influence of the most significant variables (con-

centration of HCl and NaBH4, carrier gas flow rate,

sample volume and sample and reagents flow rates)

was studied by means of multivariate experiments

based on central composite design. Optimal conditions

found were as follows: for arsenic, 1.2% (w/v) NaBH4

and 1.0 ml sample volume; for selenium, 0.5 (w/v) and

0.6 ml, respectively. For both elements, 5% (v/v) HCl

was used, while optimum flow rates were 0.32 l min-1

(nitrogen), 7.2 ml min-1 (sample), 3.2 ml min-1

(HCl solution) and 1.7 ml min-1 (NaBH4 solution).

Under the final operating conditions, sampling

frequencies were 120 samples h-1 (arsenic) and 160

samples h-1 (selenium).

Precision, accuracy, linear range as well as detec-

tion and quantification limits were determined

(Table 1). Precision (repeatability) was estimated by

the analysis of a given sample (n = 10) and expressed

as relative standard deviation [sr(%)].

Accuracy was evaluated as recovery from analysis

of a reference material provided by Mexico’s Centro

Nacional de Metrologıa (CENAM) for proficiency

testing and coded ‘‘620-Q004-0127-PA MERCOSUR

y Chile/MA2008’’, as well as from analysis of spiked

samples.

Linear range was estimated by the analysis of

regression coefficients and visual inspection of regres-

sion plots. Detection (3 s) and quantification (10 s)

limits were also estimated.

In our laboratory, we have gathered data for arsenic

and selenium content in natural and drinking water

from different areas in Uruguay for a period of 4 years.

Over 340 samples were analyzed for arsenic and 250

for selenium using the APHA 3114B HG-AAS manual

standard method (Standard Methods, APHA 1998).

The values found were always in the range up to

84 lg l-1 (arsenic) and 4 lg l-1 (selenium). A map

representing the different levels of selenium found in

Uruguay is presented in Fig. 1. The proposed system is

suitable for handling these concentration ranges, and

the methods present sampling frequencies of 120 and

160 samples h-1 compared with just 15 samples h-1 of

the APHA standard manual methods. Other analytical

methods widely used for these elements, such as

Table 1 Validation of arsenic and selenium determination by the proposed methods (MCFA) compared with the Standard Method

APHA 3114B

Figures of merit As Se

APHA 3114 B MCFA APHA 3114 B MCFA

LD (lg l-1), n = 5 0.40 0.46 0.3 0.08

LQ (lg l-1), n = 5 1.50 1.54 1.0 0.27

Precision, sr (%), n = 10 \6 \3 \6 \4

Recovery (%) 90–109* 96–99* 98–106** 95–116**

Linear range (lg l-1) 1.50–5

(R = 0.999)

1.54–15 (R = 0.999) 1.0–15 (R = 0.998) 0.27–27 (R = 0.999)

Sample consumption per determination (ml) 10 1.0 10 0.6

HCl consumption per determination (ml) 2.5 0.08 2.5 0.05

NaBH4 consumption per determination (mg) 320 20 320 5

Sampling frequency (samples h-1) 15 120 15 160

LD, LQ detection (3 s) and quantification (10 s) limits, sr (%) precision (relative standard deviation) determined by repeated analysis

of samples

* Accuracy evaluated by recovery on reference material provided by Centro Nacional de Metrologıa, CENAM (Mexico) for

participation in proficiency testing and coded ‘‘620-Q004-0127-PA MERCOSUR y Chile/MA2-2008’’

** Accuracy evaluated by spiking various water samples

Environ Geochem Health

123

electrothermal atomic absorption spectrometry, are

also very slow.

Thus, the proposed methods seem suitable for rapid

monitoring of these elements in an environmentally

friendly way, as less than 2 ml of chemical residues

are generated per determination, while the manual

standard methods generate at least 10 times more.

Conclusion

Multicommutated flow analysis was successfully used

for the generation of arsenic and selenium hydride and

applied to the determination of these trace elements by

HG-AAS in the samples of natural and drinking water.

Methods based on the proposed technique were

validated and resulted sufficiently accurate and precise

for the determination of total arsenic and selenium

contents in that type of samples.

The use of multicommutation provided several

advantages such as low consumption of reagents and

sample with less generation of chemical residues, and

a fast and flexible system which can be easily modified

by software to allow for different operating conditions.

The system was applied to the determination of total

arsenic and selenium in samples of drinking water from

Uruguay and could be used for monitoring total arsenic

and selenium in natural waters in a routine work.

Acknowledgments The authors thank UdelaR-CSIC for a

research grant and PEDECIBA-Quımica for partial support.

References

Aleixo, P. C., & Nobrega, J. A. (2003). Direct determination

of iron and selenium in bovine milk by graphite furnace

atomic absorption spectrometry. Food Chemistry, 83(3),

457–462.

Catala Icardo, M., Mateo Garcıa, J. V., & Martinez Calatayud, J.

(2002). Multicommutation as a powerful new analytical

tool. Trends in Analytical Chemistry, 21(5), 366–378.

Dedina, J., & Tsalev, D. L. (1995). Hydride generation atomicabsorption spectrometry. Chichester: Wiley.

Environmental Protection Agency. (2001). National primarydrinking water regulations; arsenic and clarificationsto compliance and new source contaminants monitoring.http://www.epa.gov/fedrgstr/EPA-WATER/2001/January/

Day-22/w1668.htm. Accessed 24 March 2010.

Feres, M. A., Fortes, P. R., Zagatto, A. G., Santos, J. L. M., &

Lima, J. L. F. C. (2008). Multi-commutation in flow

analysis: Recent developments and applications. AnalyticaChimica Acta, 618(1), 1–17.

Flynn, A., & Cashman, K. (1997). Nutritional aspects of min-

erals in bovine and human milks. In P. F. Fox (Ed.),

Advanced dairy chemistry volume 3. Lactose, water, saltsand vitamins (2nd ed., pp. 257–289). Great Britain:

Chapman & Hall.

Fordyce, F. (2005). Selenium deficiency and toxicity in the

environment. In O. Selinus, B. Alloway, J. A. Centeno,

R. B. Finkelman, R. Fuge, U. Lindh, & P. Smedley (Eds.),

Essentials of medical geology. Impacts of the naturalenvironment on public health (pp. 373–415). MA: Elsevier

Academic Press.

Goyer, R. A. (1996). Toxic effects of metals. In C. D. Klaassen

(Ed.), Casarett and Doull’s toxicology. The basic scienceof poisons (5th ed., pp. 696–698). USA: Mc Graw-Hill.

Hung, D. Q., Nekrassova, O., & Compton, R. G. (2004). Ana-

lytical methods for inorganic arsenic in water: A review.

Talanta, 64(2), 269–277.

Instituto Uruguayo de Normas Tecnicas. (2008). Agua pota-

ble—Requisitos. Norma UNIT, 833, 2008.

Knochen, M., Piston, M., Salvarrey, L., & Dol, I. (2005). Mul-

ticommuted flow system for the determination of dextrose

in parenteral and hemodialysis concentrate solutions.

Journal of Pharmaceutical and Biomedical Analysis, 37(4),

823–828.

Li, N., Fang, G., Zhu, H., Gao, Z., & Wang, S. (2009). Deter-

mination of As(III) and As(V) in water samples by flow

injection online sorption preconcentration coupled to

hydride generation atomic fluorescence spectrometry.

Microchimica Acta, 165, 135–141.

Michon, J., Deluchat, V., Al Shukry, R., Dagot, C., & Bollinger,

J. C. (2007). Optimization of a GFAAS method for deter-

mination of total inorganic arsenic in drinking water.

Talanta, 71(1), 479–485.

Paiva Oliveira, A., Gomes Neto, J. A., Araujo Nobrega, J., Mir-

anda Correira, P. R., & Oliveira, P. V. (2005). Determination

of selenium in nutritionally relevant foods by graphite fur-

nace atomic absorption spectrometry using arsenic as inter-

nal standard. Food Chemistry, 93(2), 355–360.

Piston, M., Dol, I., Knochen, M. (2006). Multiparametric flow

system for the automated determination of sodium,

Fig. 1 Map of selenium concentrations (mg l-1) in different

regions of Uruguay. Average concentrations (coded in gray-scale) have been calculated for the results obtained in a period of

4 years

Environ Geochem Health

123

potassium, calcium, and magnesium in large-volume par-

enteral solutions and concentrated hemodialysis solutions.

Journal of Automated Methods and Management inChemistry, 2006, 1–6.

Piston, M., Silva, J., Perez-Zambra, R., & Knochen, M. (2009).

Determination of total selenium by multicommutated-flow

hydride generation atomic absorption spectrometry.

Application to cow’s milk and infant formulae. AnalyticalMethods, 1, 139–143.

Reis, B. F., Gine, M. F., Zagatto, E. A. G., Lima, J. L. F. C., &

Lapa, R. A. (1994). Multicommutation in flow analysis.

Part 1. Binary sampling: concepts, instrumentation and

spectrophotometric determination of iron in plant digests.

Analytica Chimica Acta, 293(1–2), 129–138.

Rocha, F. R. P., Reis, B. F., Zagatto, E. A. G., Lima, J. L. F. C.,

Lapa, R. A., & Santos, J. L. M. (2002). Multicommutation

in flow analysis: concepts, applications and trends. Anal-ytica Chimica Acta, 468(1), 119–131.

Semenova, N. V., Bauza de Mirabo, F. M., Corteza, R., & Cerda,

V. (2000). Sequential injection analysis system for total

inorganic arsenic determination by hydride generation-

atomic fluorescence spectrometry. Analytica ChimicaActa, 412(1–2), 169–175.

Sixto, A., & Knochen, M. (2009). Multicommutated flow

system for the determination of glucose in honey with

immobilized glucose oxidase reactor and spectrophoto-

metric detection. Talanta, 77(4), 1534–1538.

APHA. (1998). Standard methods for the examination of waterand wastewater, 3114 B (20th ed.). Washington DC:

American Public Health Association.

Tiglea, P., Mello De Capitani, E. (2003). Selenio. In F. A. En De

Azeredo, & A. A. Da Matta Chasin (Eds.), Metais,geranciamento da toxicidade (pp. 239–262). San Pablo:

Editora Atheneu.

Welz, B., & Sperling, M. (1999). Atomic absorption spec-trometry (3rd ed.). Weinheim: Wiley-VCH.

World Health Organization. (2004). Guidelines for drinking-

water quality, 3rd edn, Vol. 1. http://www.who.int/water_

sanitation_health/dwq/gdwq3rev/en/. Accessed 24 March

2010.

Zhang, Y., & Adeloju, S. B. (2008). Flow injection–hydride

generation atomic absorption spectrometric determination

of selenium, arsenic and bismuth. Talanta, 76(4), 724–730.

Environ Geochem Health

123