Determination of As, Cd and Hg in emulsified vegetable oil by flow injection chemical vapour generation inductively coupled plasma mass spectrometry

Abstract

An inductively coupled plasma mass spectrometry (ICP-MS) method has been developed for the determination of As, Cd and Hg in vegetable oils using flow injection vapour generation (VG). An oil emulsion containing 10% m/v vegetable oil, 2% v/v Triton X-100 and 1.2% v/v HCl was injected into a VG-ICP-MS system for the determination of As, Cd and Hg. The quantifications have been carried out using standard addition and isotope dilution methods. The influences of vapour generation conditions and emulsion preparation on the ion signals were reported. This method has been applied for the determination of As, Cd and Hg in soybean oil and peanut oil samples obtained from a local market. The analytical results of various vegetable oil samples determined by standard addition and isotope dilution methods were in good agreement with those of digested samples analysed by pneumatic nebulization ICP-MS. Under the optimum operating conditions, the detection limit obtained from the standard addition curve was 0.01, 0.04 and 0.04 ng g^–1 for As, Cd and Hg, respectively, in the original oil samples.

---

Introduction

Inductively coupled plasma mass spectrometry (ICP-MS) is a versatile technique for elemental and isotopic analysis. The most common sample introduction method for ICP-MS is the pneumatic nebulization of solution. Although it is simple and offers good stability, sample introduction is inefficient. Hence, efforts have been made to couple alternative sample introduction systems to ICP-MS in order to extend its application range. Vapour generation (VG) is one of the sample introduction techniques that was reported to improve the sensitivity and detection limits of vapour-forming elements when they are introduced into the plasma as vapours.^1,2 Additionally, it was also shown that vapour generation can remove or separate analyte from problematic matrix species which would otherwise cause spectral and non-spectral interferences.^3,4 Vapour introduction in flow injection mode is a simple and rapid technique that has been coupled to ICP-MS previously.^5,6

The determination of trace elements in vegetable oils is one of the criteria for the assessment of quality regarding freshness and the storable period. Traces of heavy metals in vegetable oils are known to affect the rate of oxidation. Moreover, some of the metals are the subject of food legislation.⁷ Hence, determination of trace metals in vegetable oils is important.

The articles published dealing with the determination of trace elements in oils at sub-ppm levels using atomic absorption,^8–10 atomic emission,^7,10–12 and ICP-MS,^13,14 require sample pre-treatment. The other methods that are used to destroy the organic matter in vegetable oil include extraction, solubilization, dilution and microwave digestion.^15,16 In most of the cases these pre-treatment methods are tedious and time consuming with the consequent risk of sample contamination and analyte loss. One of the relatively straightforward and rapid pre-treatment systems for oils is dilution with a xylene-type or methyl isobutyl ketone-type organic solvent.¹⁶ However, increase in the organic content of the sample matrix is not compatible with ICP-MS analysis.¹⁷ An alternative to simple dilution of the oil with an organic solvent which would overcome the previously mentioned problems is the formation of an emulsion with the oil and emulsifying agent.^7,18 This procedure has already been used for metal analysis in vegetable oils with ICP-OES and ICP-MS detection systems.⁷

The aim of the present work is to develop an accurate FI-ICP-MS method with a vapour generation sample introduction device for the determination of As, Cd and Hg in vegetable oils. The vegetable oil samples were diluted in 2% v/v of Triton X-100 to form the emulsified solutions which were then determined by VG-ICP-MS. Optimization studies for the vapour generation of As, Cd and Hg are also reported. This method has been applied for the determination of As, Cd and Hg in soybean oil and peanut oil obtained locally.

Experimental

Apparatus and conditions

An ELAN 6100 DRC II ICP-MS instrument (PE-SCIEX, Concord, ON, Canada) and a simple, laboratory-built vapour generation system were used for this study. Vapours generated from samples were introduced into the ICP torch through Teflon tubing. The ICP-MS operating conditions used in this work were summarized in Table 1. A Star 2 focused microwave digester (CEM, Matthews, NC, USA) was used to digest the vegetable oil samples using the parameters listed in Table 2.

Table 1 ICP-MS equipment and operating conditions

ICP-MS instrument	PerkinElmer Sciex ELAN 6100 DRC II
Plasma conditions
rf power/W	1300
Plasma gas flow/L min^–1	15.0
Intermediate gas flow/L min^–1	1.13
Carrier gas flow/L min^–1	0.95
Mass spectrometer settings
Resolution	0.7 amu at 10% peak maximum
Isotopes monitored	^75As, ^111Cd, ^113Cd, ^201Hg, ^202Hg
Dwell time/ms	50
Sweeps per reading	5
Reading per replicate	250
Peak signal	Integrated
Autolens	On
Vapour generation system
Sample volume/µL	200
Carrier solution	1.2% v/v HCl, 0.5% m/v thiourea, 0.5 µg mL^–1Co(II)
Carrier flow rate/mL min^–1	1.5
Reductant solution	6% m/v NaBH4 in 0.1 mol L^–1NaOH
Reductant flow rate/mL min^–1	1.5

---

Table 2 Microwave digestion method^a

Stage	Ramp time	Temperature/°C	Time at parameter	Reagent	Aliquot size/mL	Added at start
a Volume of sample: 3 mL. Initial reagent: 5 mL HNO3, 10 mL H2SO4.
1	7 min	150	0	None	0	No
2	1.5 min	175	0	1.5 mL HNO3	1.5 mL	Yes
3	1.5 min	200	0	1.5 mL HNO3	1.5 mL	Yes
4	0	200	4 min	3 mL HNO3	1 mL	Yes
5	2 min	250	2 min	3 mL HNO3	1.5 mL	Yes
6	0	200	5 min	10 mL H2O2	2 mL	Yes

---

In this study, a continuous-flow vapour generation system was coupled with ICP-MS for the determination of As, Cd and Hg using FI analysis. It was assembled from a six-port injection valve (Rheodyne Type 50) with a 200 µL sample loop. A detailed description of the working principle of this vapour generator was given in a previous paper.¹⁹ The operating conditions for the vapour generation were optimized by the flow injection (FI) method.

Reagents

All reagents were of analytical-reagent grade, and de-ionized water was used throughout. Trace metal grade HNO₃ (70% m/m), HCl (35% m/m) and As(III) elemental standard were obtained from Fisher (Fair Lawn, NJ, USA). NaBH₄, NaOH, H₂SO₄ and thiourea were obtained from Merck (Darmstadt, Germany). NaBH₄ solution containing 0.1 mol L^–1NaOH was freshly prepared just prior to analysis. H₂O₂ was obtained from Riedel-deHaën (Seelze, Germany). Triton X-100 was obtained from Sigma Chemicals (St. Louis, MO, USA). Element standard solutions were obtained from SPEX CertiPrep (Metuchen, NJ, USA). Isotope enriched ¹¹¹CdO (96.4%) was purchased from the Oak Ridge National Laboratory (Oak Ridge, TN, USA) and ²⁰¹HgO (82.3%) was obtained from Cambridge Isotope Laboratories (Andover, MA, USA). The concentrations of the spike solutions were verified by reversed spike isotope dilution ICP-MS.

Preparation of oil emulsions

A 1-g portion of the vegetable oil was transferred into a 10 mL flask with the use of a pipette. Suitable amounts of Triton X-100 and HCl were added to make the final solution containing 10% m/v vegetable oil, 2% v/v Triton X-100 and 1.2% v/v HCl. After addition of suitable amounts of enriched isotopes or various amounts of As, Cd and Hg element standard solutions, this emulsified solution was diluted to the mark with pure water. The emulsion was then sonicated for 5 min in an ultrasonic bath. The prepared oil emulsion was stable for at least 20 min without separation into two layers. A blank solution was also prepared as outlined above. These solutions were analysed for As, Cd and Hg using ICP-MS by injecting 200 µL of the emulsified oil into the VG system. The concentrations of analytes in the sample were calculated by the equation described in the previous paper²⁰ and/or from the standard addition calibration curves. Owing to the mass bias effect, the sensitivity of the instrument at different m/z might be different. The isotopic compositions of Cd and Hg in both natural elements and enriched isotopes were obtained by determining the intensities of all isotopes by ICP-MS with solution nebulization. The intensities of each isotope obtained during this measurement were used for the isotope ratios and atomic weights calculation of the elements studied. Since the mass bias effect could be eliminated during isotope dilution calculation, in this study the measured isotope ratio was not corrected for mass bias effect. The effect of vapour generation conditions on the ion signals was studied using the oil emulsion prepared by the procedure described above and spiked with 1 ng mL^–1 of As, Cd and Hg.

Results and discussion

Oil emulsion preparation

In this study Triton X-100 was selected as the surfactant since it has a medium hydrophilic–lipophilic balance value.⁷ In order to balance sample homogeneity, analyte signals and complete vaporization of introduced sample, an oil emulsion containing 10% m/v oil and 2% v/v Triton X-100 was selected in the following experiments after many tryouts.

Selection of hydride generation conditions

Guo and Guo suggested that thiourea and cobalt act as catalysts for the cadmium vapour formation reaction with NaBH₄.²¹ In the present work, a solution containing thiourea, Co(II) and HCl was used as the carrier. The concentrations of NaBH₄ and HCl were optimized to obtain better signals of As, Cd and Hg. At constant HCl concentration, the signal of Cd increased with the increase in NaBH₄ concentration when it was less than 6% m/v. However, the concentration of NaBH₄ did not affect the signal of As and Hg when the concentration was greater than 0.5%. To obtain a better signal for all the elements studied, a NaBH₄ concentration of 6% m/v in 0.1 mol L^–1NaOH was employed for the subsequent experiments. Similarly, at constant NaBH₄ concentration, the signal of Cd increased significantly with the concentration of HCl as long as it was less than 1.2% v/v, whereas the signals of As and Hg increased slightly with the HCl concentration. To achieve better signal to background ratio for the elements studied, in the following experiments 1.2% v/v HCl was added to the carrier solution and the injected sample solution.

The effects of the concentrations of thiourea and cobalt in the carrier on As, Cd and Hg vapour formation were also studied. The peak areas of the elements studied increased gradually with the increase in thiourea concentration. The cadmium signal did not change significantly when the thiourea concentration was greater than 0.5% m/v. Though the signals of As and Hg increased with the increase of thiourea concentration, the backgrounds of As and Hg were also increased. To achieve better signal to noise ratio for the elements studied, 0.5% m/v thiourea was selected. The ion signal of cadmium increased rapidly with increase in cobalt concentration and reached a maximum when the cobalt concentration was about 0.5 µg mL^–1 whereas the signals of As and Hg increased gradually. However, the cadmium signal decreased significantly when the cobalt concentration was greater than 0.5 µg mL^–1. This could have been due to the interference caused in the solution phase of the vapour generation process.⁴ In the subsequent experiments, 0.5 µg mL^–1Co(II) was used as the catalyst for the vapour formation reaction.

Studies on the effect of flow rate of NaBH₄ and carrier solution on the peak height and peak area of the flow injection signal revealed that the peak height increased with the increase of flow rate with a slight decrease in the peak area. It could be due to the incomplete reaction when the flow rate was increased. As a compromise between detection limit and the repeatability of quantification, in the following experiments a solution flow rate of 1.5 mL min^–1 was selected. A summary of operating conditions of the FI vapour generation system is given in Table 1.

Since HCl was used in the vapour generation, ⁴⁰Ar³⁵Cl⁺ might have been formed in the process and interfered in the determination of ⁷⁵As⁺. In this study, H₂ was tested as the reaction gas in the DRC system to alleviate this interference. From the experimental results, it was found that the detection limits did not improve significantly. To simplify the experiments, the ICP-MS was operated in the standard mode for the real sample analysis.

The repeatability of the peak area and peak height was determined by seven consecutive injections of 0.2 ng mL^–1 As, Cd and Hg in 10% m/v emulsified oil solution. The relative standard deviation of the peak area and peak height determination was better than 5.6% for seven consecutive injections.

Determination of As, Cd and Hg in vegetable oil by FI-VG-ICP-MS

As As(III) and As(V) show different sensitivities in the vapour generation process, it is preferred to reduce arsenic to the lower oxidation state to increase the hydride generation efficiency. In order to discover the difference in the hydride generation efficiency of As(III) and As(V) in the oil emulsion, calibration curves obtained by standard addition method using 10% m/v emulsified oil solution and As(III) and/or As(V) standard solution have been plotted. From the experiments, it was found that the sensitivity was similar when the emulsified oil solution was spiked with As(III) and/or As(V). The mixture of concentrated HCl and thiourea used for vapour generation is believed to be acting as an effective reducing agent for the reduction of As(V) to the lower oxidation state. These results indicated that As in the emulsified oil could be readily quantified by FI-VG-ICP-MS using As(III) as the calibration standard.

The FI-VG-ICP-MS method has been applied to the determination of As, Cd and Hg in four vegetable oil samples. In order to evaluate the possibility of using an external calibration method, calibration curves obtained by standard addition method of 10% m/v oil emulsion solution and external calibration of aqueous standard were compared. The sensitivities of the elements studied were slightly different. This could be due to the difference in the vapour generation efficiency between oil emulsion matrix and aqueous solution. Therefore, the external calibration method could not be used for the quantification of these elements in the samples. Hence, standard addition and isotope dilution methods were used for the determination of As, Cd and Hg in vegetable oil samples. Aliquots of 200 µL of the emulsified oil were injected for the determination of As, Cd and Hg using the FI vapour generation system. Typical element-selective flow injection signals (ICP-MS detection) for a solution containing 10% m/v emulsified peanut oil #1 solution are shown in Fig. 1. The concentration of As, Cd and Hg in the injected solution was about 0.27, 0.10 and 0.12 ng mL^–1, respectively. The peak area of the flow injection signals was used for quantification. Since another isotope of the same element represents the ideal internal standard for that element, isotope dilution results are expected to be highly accurate even when the sample contains high concentrations of concomitant elements and/or losses occur during sample preparation or during sample introduction into the ICP. In this work, the concentrations of Cd and Hg were also determined by the isotope dilution method. Analytical results are shown in Table 3. As shown, the analytical results of various vegetable oil samples determined by the standard addition and isotope dilution methods are in good agreement with those of digested samples using microwaves and analysed by pneumatic nebulization ICP-MS. The relative standard deviation obtained by the reported procedure was 2–9.7% (26% for Cd) for all determinations. The higher value obtained for Cd in soya bean oil is due to a greatly lower concentration. Furthermore, poorer precision in quantification is expected when transient signals are measured. It is interesting to see that the concentrations of As and Cd were elevated in the used (fried) soybean oil. In contrast, the concentration of Hg was reduced. This could be due to the evaporation of Hg during the cooking or, on the other hand, As and Cd might have been leached from the foods during cooking and increased the concentration. According to the regulations of the Taiwan government, the maximum allowable concentrations of As and Hg in edible oil are 0.1 and 0.05 µg g^–1, respectively. The concentration of As and Hg in the analysed oil samples are far below the maximum allowable concentration.

Fig. 1 Typical element-selective flow-injection peaks of 10% m/v emulsified peanut oil #1 solution. The concentrations of As, Cd and Hg in the injected solution are about 0.27, 0.10 and 0.12 ng mL^–1, respectively. Operating conditions of FI vapour generation are given in Table 1.

Table 3 Determination of As, Cd and Hg in vegetable oil by FI-VG-ICP-MS^a (n = 3)

Sample	Method^b	Concentration/µg g^–1
		As	Cd	Hg
a Values are means of three measurements ± standard deviation. b Method 1: standard addition method. Method 2: isotope dilution method. Method 3: obtained by pneumatic nebulization ICP-MS. Cd and Hg were determined by isotope dilution method while the As was determined by external calibration method after dissolution. The total dilution factor was 50 times. c Method detection limit.
Soybean oil (before frying)	Method 1	0.13 ± 0.01	0.23 ± 0.06	3.09 ± 0.18
	Method 2		0.21 ± 0.02	2.97 ± 0.05
	Method 3	<0.99^c	0.25 ± 0.05	2.72 ± 0.30
Soybean oil (after frying)	Method 1	1.98 ± 0.04	0.78 ± 0.03	0.72 ± 0.02
	Method 2		0.74 ± 0.01	0.69 ± 0.06
	Method 3	2.02 ± 0.16	0.78 ± 0.09	0.78 ± 0.11
Peanut oil #1	Method 1	2.80 ± 0.22	0.97 ± 0.07	1.24 ± 0.05
	Method 2		0.99 ± 0.10	1.21 ± 0.04
	Method 3	2.94 ± 0.29	1.08 ± 0.07	1.20 ± 0.06
Peanut oil #2	Method 1	3.51 ± 0.31	1.99 ± 0.13	2.91 ± 0.19
	Method 2		2.06 ± 0.20	2.69 ± 0.23
	Method 3	3.56 ± 0.57	2.07 ± 0.15	2.88 ± 0.48

---

The detection limits were estimated from the standard additions curves based on the concentration necessary to yield a net peak height equal to three times the standard deviation of the blank. The detection limits of As, Cd and Hg were 0.001, 0.004 and 0.004 ng mL^–1, respectively, in prepared emulsified solution, which corresponded to 0.01, 0.04 and 0.04 ng mL^–1 for As, Cd and Hg, respectively, in the original oil sample. The limits of detection (LODs) obtained by the present procedure are superior to the LODs of recently reported procedures based on ultrasonic extraction (ICP-OES),¹⁵ complete digestion (ICP-OES)²² and emulsions (ETAAS, ETV-ICP-MS and ICP-MS).^14,18,23

Conclusion

The use of flow injection vapour generation ICP-MS provides a simple, rapid and accurate procedure to determine As, Cd and Hg in vegetable oils without complicated sample pre-treatment. The detection limits obtained for As, Cd, and Hg with this system are low enough to determine them in vegetable oils.

References

1. M. J. Powell, D. W. Boomer and R. J. McVicars, Anal. Chem., 1986, 58, 2864.
2. T. Dagnac, A. Padro, R. Rubio and O. Rauret, Talanta, 1999, 48, 763.
3. T.-J. Hwang and S.-J. Jiang, J. Anal. At. Spectrom., 1996, 11, 353.
4. Z.-C. Chen and S.-J. Jiang, J. Anal. At. Spectrom., 2006, 21, 566.
5. C.-S. Chen and S.-J. Jiang, Spectrochim. Acta, Part B, 1996, 51, 1813.
6. M.-F. Huang, S.-J. Jiang and C.-J. Hwang, J. Anal. At. Spectrom., 1995, 10, 31.
7. M. Murillo, Z. Benzo, E. Marcano, C. Gomez, A. Garaboto and C. Marin, J. Anal. At. Spectrom., 1999, 14, 815.
8. C. M. Canário and D. A. Katskov, J. Anal. At. Spectrom., 2005, 20, 1386.
9. A. M. F. Fernández, A. M. Piñeiro and P. B. Barrera, J. Anal. At. Spectrom., 2007, 22, 573.
10. L. B. Allen, P. H. Siitonen and H. C. Thompson, J. Am. Oil Chem. Soc., 1998, 75, 477.
11. M. Murillo, Z. Benzo, E. Marcano, C. Gomez, A. Garaboto and C. Marin, J. Anal. At. Spectrom., 1999, 14, 815.
12. A. N. Anthemidis, V. Arvanitidis and J. A. Stratis, Anal. Chim. Acta, 2005, 537, 271.
13. J. R. Castillo, M. S. Jimenez and L. Ebdon, J. Anal. At. Spectrom., 1999, 14, 1515.
14. S. J. Huang and S. J. Jiang, J. Anal. At. Spectrom., 2001, 16, 664.
15. M. Martin-Polvillo, T. Albi and A. Guinda, J. Am. Oil Chem. Soc., 1994, 71, 347.
16. R. M. de Souza, A. L. Saraceno, C. L. P. da Silveira and R. Q. Aucelio, J. Anal. At. Spectrom., 2006, 21, 1345.
17. E. H. Evans and L. Ebdon, J. Anal. At. Spectrom., 1990, 5, 425.
18. W. N. L. dos Santos, F. de S. Dias, M. V. Reboucas, M. de G. Pereira, V. A. Lemos and L. S. G. Teixeira, J. Anal. At. Spectrom., 2006, 21, 1327.
19. H.-J. Yang and S.-J. Jiang, J. Anal. At. Spectrom., 1995, 10, 963.
20. K.-H. Lee, S.-J. Jiang and H.-W. Liu, J. Anal. At. Spectrom., 1998, 13, 1227.
21. X.-W. Guo and X.-M. Guo, J. Anal. At. Spectrom., 1995, 10, 987.
22. I. Juranovic, P. Breinhoelder and I. Steffan, J. Anal. At. Spectrom., 2003, 18, 54.
23. M. S. Jimenez, R. Velarte and J. R. Castillo, J. Anal. At. Spectrom., 2003, 18, 1154.

---