Available online at www.sciencedirect.

com

J. Biochem. Biophys. Methods 70 (2008) 1287 – 1291

www.elsevier.com/locate/jbbm

Evaluation of the ET-AAS and HG-AAS methods of selenium determination

in vegetables
Ondrej Hegedűs a,⁎, Alžbeta Hegedűsová b , Silvia Šimková b , Vladimír Pavlík a , Klaudia Jomová b
a
Regional Authority of Public Health Nitra, Štefánikova 58, 949 63 Nitra, Slovak Republic
b
Department of Chemistry, Faculty of Natural Sciences, Constantine the Philosopher University, Tr. A. Hlinku 1, 949 01 Nitra, Slovak Republic
Received 2 July 2007; received in revised form 8 July 2007; accepted 9 January 2008

Abstract

ET-AAS and HG-AAS methods of selenium determination were compared and evaluated. The ET-AAS method has been followed with
deuterium background correction and Zeeman background correction respectively. The following validation parameters were determined:
accuracy (under repeatability conditions), trueness, calibration curve and linearity, limit of detection, limit of determination and combined standard
uncertainty of the method. The HG-AAS method and the ET-AAS method with Zeeman correction were more suitable for the determination of
low selenium concentrations in vegetables. The results were calculated by the standard addition method because of the strong matrix effect.
© 2008 Elsevier B.V. All rights reserved.

Keywords: Selenium determination; ET-AAS; HG-AAS; Validation

1. Introduction Recommended Daily Allowances in the United States, the

average daily intake should be 50–200 μg/day. [10] Optimal
Biological value of agricultural products depends on the daily intake considers 1 μg/kg of weight per day [9]. On the other
quality of growing medium — soils. Biological elements hand a maximum safe daily dietary intake has been estimated at
presented in soils are transferred into plant and thereby into the 400 µg [10].
food chain. Selenium is an essential micronutrient, required by Research in Se determination and speciation is still an
animals and humans in small amounts for basic vital functions. attractive area. Determination of selenium in low levels in
It is an integral component of glutathione peroxidases, which biological material is described in several methods.
are important antioxidant enzymes that catabolize hydroper-
oxides [1]. Its level in human organism is dependent on its - spectrometric (ET-AAS, HG-AAS, fluorescent spectrometry,
content in the soil–plant system [2–4]. ICP-MS),
Vegetables and their positive material composition play an - electrochemical (cathodic dissolving voltamperometry),
important role in human nutrition. Selenium content in - radiochemical (neutron activation analysis, X-ray fluores-
vegetables achieve levels ranging from 0.001 to 0.034 mg/kg cence analysis),
in fresh matter. Vegetables rich in starch and sugar are generally - separative (gas chromatography, liquid chromatography).
poor in Se. Mustard, cabbage, broccoli, garlic, and mushrooms
supply Se to the food chain [5–7]. Tomatoes (0.034 mg/kg) and Lower detection limits and speciation of selenium com-
carrots (0.020 mg/kg are rich in selenium [8,9]. According to the pounds can be achieved with recent powerful techniques, such
as ICP-MS or LC–MS–MS [11]. However aforementioned
⁎ Corresponding author. methods are not readily accessible to many laboratories. On the
E-mail addresses: hegedus@ruvznr.sk (O. Hegedűs), ahegedusova@ukf.sk other hand the hyphenation of common techniques such as
(A. Hegedűsová). chromatography and atomic absorption or fluorescence
0165-022X/$ - see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jprot.2008.01.002
1288 O. Hegedűs et al. / J. Biochem. Biophys. Methods 70 (2008) 1287–1291

spectrometry (AAS or AFS) as a substitute of mass spectro- Germany) and hydrogen peroxide (30%, suprapur, Merck,
metry (MS) techniques and ET-AAS, HG-AAS/AFS and HG- Germany).
ET-AAS techniques for total Se quantification are considered to For the HG-AAS technique was Se-hydride formed using the
be current procedures of great interest [12]. reducer 0.6% NaBH4 (p.a., Merck, Germany) with 0.5% NaOH
Atomic absorption spectrometry (AAS) is often used for total (water solution). The stock Se (IV) solution was prepared from
selenium determination in plant and animal samples. Matek et high purity standard stock solution (1000 ± 2 mg Se/l as
al. [13] suggested that biological material containing selenium H2SeO3, Titrisol, Merck, Germany) in 1.7 mol/l HCl. Work
can be successfully measured by the ET-AAS method with Pd solutions HCl 1.7 mol/l, 7 mol/l and 10 mol/l were prepared
as a matrix modifier. HG-AAS more easily measures samples from hydrochloric acid (30% suprapur, Merck) with deionized
with lower Se concentrations, because of higher sample water.
volumes which can be taken for hydride formation. Analysis The stock Se (IV) solution required for ET-AAS technique
of biological matrix requires complicated decomposition of was prepared from the same standard Se solution in 1% (v/v)
sizable sample amounts. Relatively higher levels of selenium HNO3 (suprapur, Merck) diluted with deionized water.
are determined often by AAS method after electrothermic
atomization (ET-AAS) with deuterium or with Zeeman back- 2.3. Analyzed material
ground correction.
The hydride generation technique (HG-AAS) is mainly Compared samples were prepared from fresh tomatoes,
used within special analysis. This method is based on collected from the research fields in Research Institute of
hydride generation from analyte which is first reduced to Vegetables in Nové Zámky and for two standard reference
hydride in the liquid phase, and consequently converted into materials 1515 (Trace Elements in Apple Leaves, USA) and N
the vapour phase which is further atomized in atomic 1570a (Trace Elements in Spinach Leaves, USA) were used as
absorption spectrometer. This technique uses analyte separa- external controls.
tion from the matrix to increase analyte concentration in
absorption environment compared with standard AAS 2.4. Measuring conditions
methods. Hydride atomization eliminates the problem of
spectral interferences. Non-spectral interferences are shown HG-AAS technique required a selenium hollow cathode
in liquid phase during analyte reduction to hydride, then lamp to operate at a wavelength of 196.0 nm with a slit width set
during hydride release out of the solution and in vapour to 1.0 nm and an electrodeless discharge lamp set at 10 mA
phase in atomizator [14]. These non-spectral interferences current without background corrector. Aspiration time was 45 s,
are non-additive and they cannot be corrected by the measuring time 5 s, the height signal evaluation, argon flow
addition technique. 60 ml/min. Atomizing environment was silicous cell heated to
The HG-AAS and ET-AAS techniques are widely used 900 °C, floated solution HCl (10 mol/l) and reducers were 0,6%
therefore the purpose of this study was a comparison of three NaBH4 and 0,5% NaOH.
AAS techniques of selenium determination and consequent For the both ET-AAS techniques (with deuterium and
statistic evaluation of the results proposing the method suitable Zeeman background correction) was used a selenium hollow
for low level determination of selenium in vegetables. cathode lamp to operate at the same conditions as HG-AAS
technique. Injected sample volume was 10 μl. Palladium matrix
2. Materials and methods modifier Pb(NO3)2 with concentration 0.1 mol/l digested in
0.1% HNO3 and ascorbic acid (1%) were used as modifiers.
2.1. Apparatus Evaluation of the results was done with the method of
calibration curve.
Atomic absorption spectrometer SpectrAA200 Varian (Mul-
grave Virginia, Australia) equipped with models VGA-77 for 2.5. Methods validation
hydride technique and GTA-100 for electrothermal atomization
(ET-AAS) with deuterium background correction was used for Three techniques of selenium determination, ET-AAS (with
selenium determination respectively. Atomic absorption spec- deuterium and Zeeman background correction) and HG-AAS,
trometer AA240Z Varian (Mulgrave Virginia, Australia) with were submitted to validation.
Zeeman background correction was used as the second ET-AAS Accuracy (repeatability) was characterized by selected
technique. Wet digestion of the plant samples was done in variation coefficient sR, calculated from standard deviation s
mineralization autoclaves ZA-1 (JZD Pokrok Zahnašovice, and arithmetic average from a series of measurements under
Czech Republic). repeatability conditions [15]. Gained sR is evaluated according
to pre-established requirements.
2.2. Chemicals and reagents Two parameters of the calibration curve are calculated (slope
of a straight line m, shift of the regress straight line b and its
Deionized water was used to prepare all solutions. Standards statistic testing, coefficient of determination R2 and reliability
were prepared daily by appropriate dilution of Merck standards. of the calibration relation presented by standard deviation sy, or
Samples were digested with nitric acid (65%, suprapur, Merck, relative residual deviation syrel). If statistically significant
 O. Hegedűs et al. / J. Biochem. Biophys. Methods 70 (2008) 1287–1291 1289

difference in the shift of regression straight line from zero is Digestion procedure: Fresh matter and dry matter (reference
observed, value b is involved otherwise b is unkept. material) samples were weighed directly into mineralization
Linearity of the method was evaluated as an ability to autoclaves (type ZA-1) at 10 g and 1 g respectively.
provide results proportional to concentration within the defined Samples were digested in digestion mixture 5 ml conc.
interval [16]. A maximum allowing deviation from average HNO3 and 2 ml H2O2, and dry matter samples had an extra
relative response was assigned, which was graphically repre- addition of 1 ml deionized water to moisturize. Samples were
sented as a lower and upper limit of acceptable value of relative closed in iron mineralization autoclaves and digested in hot-air
response. Linearity of the method was provided to the point sterilizers at 140 °C for 120 min.
with relative average response intervening between lower and Sample modification for HG-AAS determination: after
upper limit of acceptance value of relative response. cooling the samples, mineralizates were quantitative transferred
For the purposes of result evaluation using the standard to 25 ml volumetric flasks, 4 ml HCl (7 mol/l) was added and
addition method, limit of detection as a triple and limit of than heated 30 min at 80 °C. After cooling the samples
determination as a decuple of standard deviation from the volumetric flasks were filled by HCl (1.7 mol/l) to the match
blanks are calculated. For the purposes of result evaluation mark and measured immediately.
using the calibration curve method, limit of detection and limit Sample modification for ET-AAS determination: after cool-
of determination are calculated from the upper limit approach ing the samples, mineralizates were quantitatively transferred to
(ULA) [17,18]. The condition is equidistant distribution of 25 ml volumetric flasks, and filled with deionized water to the
concentrations. Relation for detection limit calculation match mark.
k :s
LODucD ¼ Dm y , kD is table value for (n-2; 0.01) in calibration
relation y = b + mc, or for (n-1; 0.01) in calibration relation 3. Results
y = mc, sy is residual standard deviation. Relation for determina-
tion limit calculation LOQ = 3 · LOD. Comparison of the possibility to determinate Se was
All mentioned validation characteristics were calculated in performed from the results gained from Se determination in
Excel program using computing tables. Created computing tables blanks and in real samples. Measurements were performed with
allow calculations of validation characteristics from measured the series of blank and tomato samples analysis of by the HG-
data. The results are submitted to numeral and verbal evaluation AAS and both ET-AAS technique respectively. Absorbance of the
according to pre-established requirements for this purpose. blanks from AAS measurements was compared to absorbance of
All partial results of evaluation are summarized in one table the tomato samples. Results of measurements and statistical
as the validation protocol. comparison of three techniques is described in Table 1.
Standard uncertainty of these methods is calculated as an Results show clear distinction of the absorbances of the
extended combined uncertainty according to Eurachem docu- blank and vegetable samples performed by ET-AAS technique
ments in the PC program Metro2003, version 3.02 [19]. with Zeeman background correction and technique HG-AAS.
Means of absorbance (± sd) of ET-AAS technique with Zeeman
2.6. Sample and sample preparations background correction as well as the technique HG-AAS are
significantly different, whereas the ET-AAS technique with
Samples were washed with demineralized water to remove deuterium background correction is not significantly different at
all the salts produced by soil, sweat, etc. Clean tomatoes were 5% level by Mann–Whitney's U-test.
homogenized. Homogenized fresh matter was submitted to Validation of the three techniques listed above was realized
mineralization. by the main validation characteristics determination in water

Table 1
Statistical evaluation of absorbance differences between the tomato samples and blanks
Tomatoes ET-AAS deuterium b.c. a blank ET-AAS Zeeman b.c. a blank HG-AAS blank
Absorbance Mean sd Absorbance Mean sd Absorbance Mean sd Absorbance Mean sd
0.024 0.024 0.0099 0.0025
0.019 0.026 0.0037 0.0024
0.014 0.014 0.0080 0.0019
0.027 0.011 0.0019 0.0047
0.028 0.023 0.0041 0.020 0.019 0.0052 0.0001 0.0049 0.0032 0.0115 0.0057 0.0032
0.021 0.021 0.0079 0.0085
0.026 0.018 0.0033 0.0091
0.025 0.017 0.0030 0.0067
0.023 0.024 0.0072 0.0051
0.024 0.013 0.0037 0.0046
Mann–Whitney's U-test
P 0.054 0.00016 0.00016
a
Background correction.
1290 O. Hegedűs et al. / J. Biochem. Biophys. Methods 70 (2008) 1287–1291

Table 2 Table 4
Evaluation of repeatability of selenium determination methods — water Detection limits of selenium determination methods — water solutions model
solutions model
Calculated ET-AAS HG-
Calculated parameter ET-AAS HG-AAS parameter AAS
Deuterium b.c. a Zeeman b.c. a
Deuterium Zeeman
Evaluation of intercept in used calibration model b
b.c. a b.c. a
tcal 2.901 1.192 2.645
Average, mg/l P 0.0148 0.0179 0.0095 tcrit 2.160 2.145 2.160
x
m,1/mg 9.367 14.541 13.082
B 0.0093 0.000 − 0.0039
Standard deviation, mg/l Sd 0.0011 0.0011 0.0010
Coefficient of variation, % 100:s 7.5 6.2 11.0 Evaluation of linear regression c
sR ¼ P
x sy 0.0053 0.0024
kD (n-2) 2.953 2.953 2.953
kD (n-1) 2.624 2.624 2.624
Requirement sR b 15% Satisfied Satisfied Satisfied
LOD, mg/l 0.0017 0.0011 0.00049
a
Background correction. LOQ, mg/l 0.0050 0.0033 0.0017
Requirement Satisfied Satisfied Satisfied
a
Background correction.
solutions. Validation results were calculated in Excel tables and b
Absorbance = f (c), c is concentration [mg/l], according to evaluation of
are shown in Tables 2, 3 and 4. calibration model for ET-AAS with deuterium b.c. and HG-AAS is valid
Calculation of the standard combined uncertainty of the calibration model y = mx + b, for ET-AAS with Zeeman b.c. y = mx.
method requires uncovering all uncertainty sources throughout
c
Requirement for limit of quantification (LOQ) ≤ 0.005.
whole analytical process. The result of this calculation is
standard combined uncertainty (Table 5). digestion apparatus recommend weight of dry material in the
Trueness of the methods was verified by the analysis of range 0.3–0.4 g. Mineralization of the quantities may cause that
matrix reference material (RM). RM was dried apple leaves and determined Se values are below the limit of determination,
dried spinach leaves. The results of the analysis show Table 6. according to low Se content in many foodstuff. Some
techniques of samples preparation such as weight increase of
4. Discussion the shot or preconcentration of the samples can increase
selenium levels. However, they bring other technological
Low selenium levels in vegetable require sensitive determi- problems or influence the stability of analyzed matter [18].
nation method, whereas problems with matrix effect in plant Whereas very low amounts of selenium was expected, high
matter often occur. Digestion of plant material is the key factor sensitivity method should be used.
in the sample preparation procedure. The technical manuals of The results of blanks and tomato samples analysis measure-
ment showed that low contents of selenium in analyzed
vegetables cannot be distinguished reliably by the ET-AAS
technique with deuterium background correction, which means
Table 3
Evaluation of calibration model and linear regression of selenium determination
that the sensitivity of this method under shown conditions is
methods — water solutions model insufficient for Se determination in vegetable samples with low
Calculated ET-AAS HG-AAS
Se content (Table 1). Selenium data obtained by HG-AAS were
parameter considered the true values because it is an accepted reference
Deuterium b.c. a Zeeman b. c. a
method of the AOAC, Method Number 996.17 [20].
Evaluation of intercept in used calibration model b According to results those two techniques may be success-
tcal 2.901 1.192 2.645
fully used for determination of the low selenium amounts in
tcrit 2.160 2.145 2.160
m,1/mg 9.367 14.541 13.082 vegetables. Even though the both ET-AAS techniques are
B 0.0090 0.000 −0.004 characterized by lower value of the repeatability than the HG-
AAS technique (6.2% for ET-AAS with Zeeman background
Evaluation of linear regression c correction, 7.5% with deuterium background correction and
sy 0.0053 0.0091 0.0024
sy rel 3.35 6.03 2.14
Requirement Satisfied Satisfied Satisfied
Table 5
Evaluation of linearity d Calculation of uncertainty with program Metro2003
Requirement Satisfied Satisfied Satisfied
Parameters ET-AAS with Zeeman HG-AAS
a
Background correction. b.c. a
b
Absorbance = f (c), c is concentration [mg/l], according to evaluation
Concentration of Se, mg/kg 0.046 0.021
of calibration model for ET-AAS with deuterium background correction and
Combined standard uncertainty, mg/kg ±0.0022 ±0.0058
HG-AAS is valid calibration model y = mx + b, for ET-AAS with Zeeman
Coverage factor 1.96 1.96
background correction y = mx.
c Probability 95% 95%
Requirement: sy rel b 10.0.
d a
Requirement: all values of relative response. Background correction.
 O. Hegedűs et al. / J. Biochem. Biophys. Methods 70 (2008) 1287–1291 1291

Table 6 grown in relation to soil hygiene on extremely acid soil. Chem Listy
Verification of trueness of methods 2003;97:801.
[4] Terry N, Zayed AM, de Souza MP, Tarun AS. Selenium in higher plants.
Used reference material Value of Se[mg/kg] Recovery
Annu Rev Plant Physiol Plant Mol Biol 2000;51:401–22 ProQuest
[%]
Certified Measured Agriculture Journals.
HG-AAS [5] Finley JW. Selenium accumulation in plant foods. Nutr Rev 2005;63:196–202.
SRM 1515 — apple leaves 0.050 0.045 91.0 [6] Sager M. Selenium in agriculture, food, and nutrition. Pure Appl Chem
2006;78(1):111–33.
ET-AAS Zeeman background correction [7] Irion CW. Growing alliums and brassicas in selenium-enriched
SRM 1570a — spinach leaves 0.117 0.113 96.6 soils increases their anticarcinogenic potentials. Med Hypotheses
1999;53:232–5.
[8] Maďarič A, Kadrabová J. Selén v potravinách a možnosť jeho
11.0% for HG-AAS respectively). Obtained data of repeat- suplementácie. Bull Potrav Výsk 1998;1:11–7.
ability are comparable to other references [21,13]. [9] Mosnáčková J, Kováčiková E, Pastorová J, Košická M, Vojtaššáková A,
Holčíková K, Simonová E. Selén v potravinách. Bratislava: VÚP80-
The linearity of the measurements was provided and 89088-22-8; 2003.
confirmed for all three techniques. Lower detection limit was [10] Selenium. Trace elements in human nutrition and health. Geneva: World
calculated in the ET-AAS technique with deuterium background Health Organisation; 1996. p. 105–22.
correction and HG-AAS (Table 4), that is necessary for the [11] Nogueira CW, Zeni G, Rocha JBT. Organoselenium and organotellurium
selenium determination in vegetables and determines successful compounds: toxicology and pharmacology. Chem Rev 2004;104:6255–86.
[12] Capelo JL, Fernandez C, Pedras B, Santos P, Gonzalez P, Vaz C. Trends in
application of the method. selenium determination/speciation by hyphenated techniques based on
AAS or AFS. Talanta 2006;68:1442–7.
5. Conclusion [13] Matek M, Blanuša M, Grgič J. Comparison of two methods using atomic
absorption spectrometry for determination of selenium in food. Arh Hig
Comparison of the following methods for selenium determi- Rada Toksikol 1999;50:283–8.
[14] Dědina J, Fara M, Kolihová D, Korečková J, Musil J, Plško E, Sychra V.
nation (ET-AAS with deuterium background correction, Zee- Vybrané metody analytické atomové spektrometrie. Československá
man background correction and HG-AAS) showed that method spektroskopická společnost. Sekce optické atomové sprektroskopie,
ET-AAS with deuterium background correction was not able to Praha; 1987.
distinguish the signal of blanks and vegetable samples with low [15] STN ISO 3534-1. Štatistika. Slovník a značky. Časť 1: Pravdepodobnosť a
Se content. That means: sensitivity of this method is insufficient všeobecné štatistické termíny; júl 1999.
[16] Holík M. Validace analytických metod (Postup při práci a příprava
for selenium determination in vegetable samples with low Se protokolu se zaměřením na HPLC a TLC). Brno: Lachema; 1990.
content. HG-AAS method and ET-AAS method with Zeeman [17] Mocák J, Bond AM, Mitchell S, Scollary G. A statistical overview of
background correction are able to distinguish signal of blank standard (IUPAC and ACS) and new procedures for determining limits of
and vegetable samples correctly. Comparison of values of detection and quantification — application to voltammetric and strip-
ping techniques (IUPAC Technical Report). Pure and Appl Chem
detection and determination limits showed that HG-AAS
1997;69:297–328.
method produces lower values. This is a very important [18] Hegedűs O, Hegedűsová A, Gašparík J, Ivičičová A. Evaluation of the
conclusion for selenium determination in the plant samples ETA-AAS and HG-AAS methods of selenium determination in vegetables.
with low selenium content. Chem Listy 2005;99:518–24.
[19] Baričič P, Mackov M. Metro2003 (Verzia: 2.30). Laboratórny softvér pre
References Windows. Bratislava: Chemmea s.r.o.; 2003.
[20] Miksa IR, Buckley CL, Carpenter NP, Poppenga RH. Comparison of
selenium determination in liver samples by atomic absorption spectro-
[1] Burk RF, Levander OA. Selenium. In: Shils ME, Olson JA, Shike M, et al, scopy and inductively coupled plasma-mass spectrometry. J Vet Diagn
editors. Modern nutrition in health and disease. Baltimore, MD: Williams
Invest 2005;17:331–40.
and Wilkins; 1999. p. 265–76. [21] Diaz-Alarcón JP, Navarro-Alarcón M, López-García de la Serrana H,
[2] Bajčan D, Žemberyová M, Klimek J, Rúriková D. Stanovenie López-Martínez MC. Determination of selenium in cereals, legumes and
obsahu bioprístupného selénu v pôdach metódou AAS. Chem Listy dry fruits from southeastern Spain for calculation of daily dietary intake.
2001;95:638–41.
Sci Total Environ 1996;184:183–9.
[3] Vollmannová A, Tóth T, Tomáš J, Jomová K. The affection of intake of
some content of heavy metals in soils micronutrients by grain of bean