Determination of arsenic and mercury in sunflower oil by ...

Food Additives and Contaminants, 2002, Vol. 19, No. 10, 948 ±953

Determination of arsenic and mercury in sun¯ower oil by electrothermal atomic absorption I. Karadjova* and T. Venelinov

Introduction

Faculty of Chemistry, University of So®a, J. Bourchier 1, 1126 So®a, Bulgaria

(Received 10 January 2001; revised 1 October 2001; accepted 4 November 2001)

A simple and fast method for the determination of arsenic (As) in sun¯ower oil by electrothermal atomic absorption (ETAAS) is described. The optimal instrumental parameters for ETAAS measurement of As species in stable and homogeneous soap emulsions prepared from oil samples with tetramethylammonium hydroxide (TMAH) have been established. The limit of ¡ determination is 5 ng g 1 total As in sun¯ower oil. A second approach involving extraction of As and mercury from sun¯ower oil and consequent ETAAS measurement is also described. Simultaneous quantitative extraction of As(III), As(V), monomethylarso nate (MMA) and dimethylarsinat e (DMA) as well as Hg(II), monomethylmercury(II)chlorid e (MMC), dimethylmercury (DMM), diethylmercury (DEM) and diphenylmercury (DPM) is achieved by using an extraction mixture comprising of 0.1 m NH3 /0.01 m EDTA. Pre-reduced palladium is applied as an e ective modi®er for the next ETAAS measurement of all extracted species. The method of standard addition is employed for calibration. The accuracy and reproducibly of the methods was established by spike and recovery experiments and parallel analysis of sun¯ower oils from the marketplace. Limits of determination of ¡ ¡ 2 ng g 1 for As and 3 ng g 1 for mercury were obtained. Keywords : sun¯ower oil, arsenic, mercury, ETAAS, extraction

* To whom correspondence should be addressed. e-mail: Karadjova@ chem.uni-so®a.bg

Two main approaches can be applied for the determination of trace elements in edible oils: direct spectrometric determination, e.g. ETAAS (Olejko 1976, Hendrikse et al. 1988, Lacoste et al. 1993, Dalen and Galan 1994, Martin-Polvillo et al. 1994, Karadjova et al. 1998), ICP-AES (Ibrahim 1991), ICP-MS (Castillo et al. 1999) or sample pretreatment followed by spectrometric measurement. Several procedures have been used for sample pretreatment: dry or acid sample digestion (Saleh et al. 1988) or extraction (Jacob and Klevay 1975, Ooms and van Pee 1983) of trace elements. There are few papers dealing speci®cally with the analysis of arsenic (As) and mercury (Hg) in edible oils and the most commonly used procedure consists of acid sample digestion followed by ETAAS or HGAAS measurement (Elson and Ackman 1978, Puig-Deu and Baxaderas 1988). The aim of the study was to assess the possibility of direct ETAAS for the determination of As and Hg in sun¯ower oil after sample solubilization in tetramethylammonium hydroxide (TMAH). The optimal TMAH concentration, optimal temperature programme for ETAAS measurement and a suitable modi®er have been established. As an alternative, a simple and rapid method consisting of an extraction and preconcentration step followed by ETAAS measurement of As and Hg was developed. Several extraction systems (HCl/H2 O2 , HCl/EDTA, NH3 / EDTA and Na2 S2 O3 ) were investigated to achieve simultaneous, quantitative extraction of di erent species of As and Hg. The thermal behaviour of organic and inorganic species of As and Hg in the presence of the most suitable extracting agents was studied. Based on the results obtained, the optimal temperature programmes and e ective modi®ers are proposed. The accuracy and reproducibilit y of the developed analytical method was established by spike and recovery experiments and by parallel analysis of sun¯ower oils from the marketplace by using an

Food Additives and Contaminant s ISSN 0265±203X print/ISSN 1464±5122 online # 2002 Taylor & Francis Ltd http://www.tandf.co.uk/journals DOI: 10.1080/0265203021015538 7

Arsenic and mercury in sun¯ower oil

analytical procedure based on wet digestion of the sample in an autoclave and measurement of As and Hg by hydride generation (Elson and Ackman 1978).

Table 2. Temperature programme for ETAAS determination of arsenic and mercury species in sun¯ower oil after extraction of 0.1 M NH3 /0.01 M EDTA.

Step

Materials and methods Instrumentation Electrothermal atomic absorption measurements were carried out using a Perkin Elmer Zeeman 3030 (Uberlingen, Germany) spectrometer coupled with an HGA-600 atomizer. The light sources used were electrodeless discharged lamps for As and Hg. The spectral band-pass and wavelengths used were as recommended by Perkin-Elmer. Pyrolytic graphiteTable 1. Temperature programme for ETAAS determination of As in sun¯ower oil after solubilization with TMAH. As (III), As (V), MMA, DMA Step Drying Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min † Pre-treatment 1 Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Pre-treatment 2 Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Atomization Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Cleaning Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min )

Wall atomization

Platform atomization

120 15 10 300

130 15 15 300

700 15 20 300

800 15 20 300

1100 10 25 300

1200 15 30 300

2000 0 3 0

2300 0 3 0

2500 2 4 300

2500 4 4 300

Sample injection volume, 20 ml; modi®er, 10 ml; tetrachloroammonium ¡1 palladate, 300 mg ml .

949

Drying Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Pre-treatment Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Atomization Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min ) Cleaning Temperature (°C) Ramp time (s) Hold time (s) ¡1 Ar ¯ow rate (ml min )

As(III), As(V), MMA, DMA

Hg(II), MMC, DMM, DEM, DPM

120 15 10 300

120 15 15 300

1000 20 20 300

250 15 15 300

2100 0 3 0

1000 1 10 0

2500 2 4 300

1800 2 4 300

Sample injection volume, 20 ml; modi®er, prereduced Pd, wall atomization.

coated graphite tubes and uncoated graphite tubes with pyrolytic platforms were used as atomizers. Solutions (20 ml) were introduced into the graphite atomizer using an AS-60 autosampler. The graphite furnace operating parameters are presented in tables 1 and 2. Atomic absorption signals were recorded on an Anadex printer. Only peak areas were used for quanti®cation.

Reagents All reagents and solvents were of analytical-reagen t grade (E. Merck, Darmstadt, Germany) unless speci®ed otherwise. Hydrochloric acid (p.a. Merck) was additionally puri®ed by sub-boiling distillation. Hydrogen peroxide was additionally puri®ed by ion exchange. Triton-X 100 (Ferak, Berlin, Germany) and TMAH as a 25% aqueous solution (Fluka Chemie AG, Switzerland) were used as emulsi®ers. Aqueous standard solutions for As(III), As(V) and

950

I. Karadjova and T. Venelinov

Hg(II) were prepared by appropriate dilution of ¡ 1000 mg ml 1 stock solutions, which were made from Titrisol concentrates (Merck). Individual stock solu¡ tions (1000 mg ml 1 as As or Hg) for organic species were prepared by dissolving monomethylarsonate (MMA) disodium salt, dimethylarsinate (DMA) sodium salt in water and methylmercury(II) chloride (MMC), dimethylmercury (DMM), diethylmercury (DEM) and diphenylmercury (DPM) in isopropanol. Doubly distilled water was used throughout.

Procedures for sample preparation Sample preparation for direct ETAAS determination of As and Hg. The oily matrix was modi®ed with an aqueous solution of TMAH. An accurately weighed sun¯ower oil sample of around 1 g was mixed with 2.0 ml 5% (v/v) aqueous solution of TMAH and 0.1 ml 1% (v/v) aqueous solution of Triton X-100. The mixture was treated for 10 min at 60°C on the hot plate and than treated in an ultrasound bath for 30 min. A soap emulsion, which was stable for about 20 min, was obtained and used for ETAAS measurements. Sample preparation for extraction ETAAS determination of As and Hg. An accurately weighed sample of sun¯ower oil of around 10 g was diluted with 5.0 ml CHCl3 . Extraction was performed for 30 min with 2 ml of various extraction mixtures at (1) room temperature, (2) 70°C under re¯ux and (3) using ultrasound treatment. After centrifugation, 20 ml of the extracts obtained were used for ETAAS measurements. Recovery test. The recovery of the inorganic and organic species of As and Hg was evaluated by extracting spiked oil samples with various extraction mixtures and comparing the results with those obtained for standard solutions for the same concentrations. Individual spiked oil samples were prepared by mixing sequentially 10 g oil with 0.05 ml aqueous standard solution of Hg(II), As(III), As(V), MMA and DMA or with 0.05 ml isopropanol solution of MMC, DMM, DEM and DPM with appropriate concentrations in order to obtain ®nal ¡ spikes for As in sun¯ower oil (5, 10 and 100¡ ng g 1 ) and for Hg in sun¯ower oil (5 and 10 ng g 1 ). The solutions were shaken for 5 min and then used for the recovery tests.

Results and discussion Direct ETAAS determination Di erent TMAH concentrations in various solvents (water, methanol, 1,4-dioxan and di erent concentrations of Triton X-100 surfactant) were investigated for the preparation of soap emulsions. It was found that stable and homogeneous emulsions were obtained using a 1:1 dilution of oil sample with a 5% (v/v) aqueous solution of TMAH and 0.1 ml 1% (v/v) Triton X-100 as surfactant. Pretreatment and atomization curves were constructed for spike samples in order to de®ne optimal temperature programmes for ETAAS measurement of di erent species of As and Hg as well as to choose the most suitable modi®er for each of them. Unfortunately, all attempts to ®nd possibilities for the determination of Hg in emulsions were unsuccessful. A maximum pretreatment temperature achieved in the presence of palladium as the modi®er (used as an aqueous solution or as a permanent modi®er) was 400°C, which was not high enough to ensure matrix removal. Very high background absorptions owing to almost simultaneous atomization of Hg and oil matrix mineralization make ETAAS measurement of any kind of Hg species impossible. Hg determination would be possible by using higher oil dilution, but this is useless for analytical practice and routine oil analysis. Results obtained for di erent inorganic and organic species of As are shown in ®gure 1. As can be seen, the di erences observed in the thermal behaviour of As(III), As(V), MMA and DMA in the presence of lead as modi®er were insigni®cant. Thermal stabilization of inorganic and organic species of As in the soap emulsion was investigated with an aqueous solution of palladium as tetrachloroammonium palladate(II) ¡ (300 mg ml 1 ) and with pre-reduced palladium ob¡ tained by injection of 20 ml 1000 mg ml 1 palladium, charred at 1000°C. For both modi®ers, thermal stabilization of all As species studied up to 1100°C was achieved. It should also be noted that equal sensitivity was obtained for all inorganic and organic As species by using the optimal furnace programme shown in table 1. Matrix interferences were evaluated through the ratio between the slopes of calibration curves in the presence of soap emulsion and aqueous standard solutions. A signi®cant degree of suppression, ratio of 0.90 for As(III), of 0.91 for A(V) and of 0.89 for MMA and DMA, was found. As can be expected,


951 (a)

MMA

Pd as tetrach loroammonium palladate (II), 300

As(V)

m g ml

-1

0,14 0,12

0,12

As(III)

Integrated absorbance


0,14

0,10

DMA 0,08

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

0,10

Hg(II)

0,08 0,06

CH3HgCl

0,04

(CH3)2)Hg

0,02 0,00

(C 6H5)2Hg

(C 2H5)2Hg

o

Temperature, C

100

Figure 1. Pretreatment and atomization curves for ¡ ¡ ¡ As(III), 7 ng g 1 ; As(V), 9 ng g 1 ; MMA, 10 ng g 1 ¡1 and DMA, 5 ng g with modi®er 20 ml Pd as aqueous ¡ tetrachloroammoniu m palladate (II), 300 mg ml 1 .

150

200

250

300

350

400

o

Temperature, C (b) prereducedPd as mod i fier 0,14

platform atomization eliminates to some extent matrix interferences and increases sensitivity (around 1.3-fold). Experiments carried out showed that in both cases wall or platform atomization by the method of standard addition should be used for calibration, but since equal sensitivity was obtained for all As species, standard additions prepared from only one calibrant are possible. Note that the real advantage of this approach for total As determination in sun¯ower oil is the possibility of using aqueous standard solution of As (III) for calibration.

Extraction ETAAS determination of As and Hg Preliminary experiments showed that in the presence of nitric acid, considerable losses of Hg were observed, which was why hydrochloric was chosen as the extraction reagent. Taking into account the chemical properties of As and Hg, the extraction mixtures investigated were 0.1 m HCl/0.1 m H2 O2 , 0.1 m HCl/0.01 m EDTA(disodium salt), 0.1 m NH3 /0.01 m EDTA (acid) and 0.01 m Na2 S2 O3 . As a ®rst step, careful optimization of the temperature programmes for ETAAS measurement of di erent Hg and As species in the extraction mixtures studied was performed through the pretreatment and atomization curves. The modi®ers used were palladium as an aqueous solution of tetrachloroammonium ¡ palladate(II) (300 mg ml 1 ) and pre-reduced palladium. As can be expected, insigni®cant di erences were observed in the thermal behaviour of As species


0,12

Hg(II)

0,10 0,08

(C 6H 5)2Hg

0,06 0,04

(CH3)2)Hg

0,02

CH 3HgCl

(C 2H5)2Hg

0,00 100

150

200

250

300

350

400

o

Temperature, C

¡

Figure 2. Pretreatment curves for Hg (II), 70 ng g 1 ; ¡ monomethylmercur y chloride, 40 ng g 1 ; dimethylmer¡1 ¡ cury, 30 ng g ; diethylmercury, 20 ng g 1 ; and diphenyl¡1 mercury, 50 ng g , with modi®er (a), 20 ml Pd as aqueous ¡ tetrachloroammoniu m palladate (II), 300 mg ml 1 , and (b), pre-reduced Pd. Di erent concentrations of Hg species studied were used to obtain diverse curves.

and the optimal instrumental parameters are summarized in table 2. The results shown in ®gure 2a demonstrate signi®cant losses of organic Hg species even in the presence of palladium as modi®er. Obviously, these losses are during the drying step and thermal stabilization is only possible by chemical reaction with a modi®er or by chemical sorption on the active surface of already reduced modi®er. It might be assumed that this is the explanation for the much better e ciency of pre-reduced palladium in comparison with the aqueous solution of tetrachloroammonium palladate(II) for the thermal stabilization of organic Hg (®gure 2b). As described by Karadjova et al. (1995) a ramp time of 1 s in the

I. Karadjova and T. Venelinov

952

Table 3. Degree of extraction (R%) of inorganic and organic arsenic species obtained according to the recovery tests. R(%) Extraction system HCl/H2 O2 , 20°C HCl/EDTA, 20°C NH3 /EDTA, 20°C HCl/H2 O2 , 70°C HCl/EDTA, 70°C NH3 /EDTA, 70°C HCl/H2 O2 , ultrasound HCl/EDTA, ultrasound NH3 /EDTA, ultrasound As ˆ 10 ng g

¡1

As (III)

As (V)

MMA

DMA

>99 >99 >99 >99 >99 >99 >99 >99 >99

>99 >99 >99 >99 >99 >99 >99 >99 >99

79 § 6 83 § 7 93 § 5 81 § 4 65 § 8 96 § 5 90 § 6 85 § 7 95 § 3

83 § 6 81 § 7 94 § 4 80 § 6 59 § 8 96 § 4 89 § 7 78 § 7 95 § 4

Accuracy of the proposed methods

as the spike concentration.

atomization step ensures higher a sensitivity for all Hg species in comparison with maximum power. The optimal instrumental parameters for Hg species determination are shown in table 2. The degree of extraction of each chemical species of As or Hg from sun¯ower oil in extraction mixtures under investigation was obtained sequentially in the separate experiments according to the described recovery tests. The results achieved are summarized in table 3 for As and table 4 for Hg. It is evident from table 4 that 0.01 m Na2 S2 O3 is a very suitable extraction reagent for Hg species. However, for As species, low recoveries were obtained. The simultaneous, quantitative extraction of all studied Table 4.

species of As and Hg is achieved by using a mixture of 0.1 m NH3 /0.01 m EDTA. As recoveries were in the range 93±95%, the method of standard addition should be used for calibration. As a conclusion, the correct determination of total As and Hg content in sun¯ower oil is possible after extraction with 0.1 MNH3 /0.01 m EDTA followed by ETAAS measurement under the instrumental parameters shown in table 2 by using the method of standard addition for calibration and aqueous standard solutions of As(III) and Hg(II) as calibrants.

To check the accuracy of the proposed methodology, spike and recovery experiments were carried out at 5, ¡ ¡ 10 and 100 ng g 1 content for As and 5 and 10 ng g 1 for Hg. For all spiked samples, the recoveries were in the range 90±105% with RSD for these concentrations of around 10% for Hg and of 8% for As. Since no standard reference materials with a certi®ed As and Hg content in sun¯ower oil were available in Bulgaria, the accuracy of the method was checked by comparison with an independent method, i.e. sample acid digestion and hydride generation atomic absorption spectrometry (HGAAS) measurement of As and Hg. Values obtained for sun¯ower oils from Bulgarian market are shown in table 5 and are in good agreement, thus proving the validity of the analytical methods proposed.

Degree of extraction (R%) of inorganic and organic mercury species obtained according to the recovery tests. R(%)

Extraction system

Hg(II)

CH3 HgCl

(CH3 )2 Hg

(C2 H5 )2 Hg

(C6 H5 )2 Hg

>99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99

51 § 8 62 § 7 92 § 4 95 § 4 57 § 9 68 § 7 93 § 4 96 § 5 95 § 4 69 § 4 95 § 3 96 § 3

45 § 9 46 § 10 91 § 5 93 § 4 51 § 3 55 § 6 95 § 5 95 § 4 52 § 9 55 § 10 93 § 3 94 § 3

43 § 10 41 § 9 91 § 4 93 § 4 46 § 8 49 § 9 93 § 4 94 § 3 49 § 9 48 § 8 94 § 4 94 § 3

41 § 10 42 § 9 91 § 4 92 § 4 45 § 9 48 § 9 92 § 3 93 § 3 48 § 9 47 § 8 94 § 4 93 § 3

HCl/H2 O2 , 20°C HCl/EDTA, 20°C NH3 /EDTA, 20°C Na2 S2 O3 , 20°C HCl/H2 O2 , 70°C HCl/EDTA, 70°C NH3 /EDTA, 70°C Na2 S2 O3 , 70°C HCl/H2 O2 , ultrasound HCl/EDTA, ultrasound NH3 /EDTA, ultrasound Na2 S2 O3 , ultrasound Hg ˆ 5 ng g

¡1

as the spike concentration.


Table 5.

953

Determination of total As and Hg in sun¯ower oils from the market in Bulgaria. ¡1

¡1

As (ng g ) Sample Slantse Kaliakra ¡1 Kaliakra spiked with As and Hg at 5 ng g Pampas Biser Vidona ¡1 Vidona spiked with As and Hg at 5 ng g ¡1

Hg (ng g )

Direct ETAAS

ETAAS after extraction

HGAAS after acid digestion

ETAAS after extraction

HGAAS after acid digestion

4.4 § 0.3