Rapid Determination of Low-level Sulfite in Dry Vegetables and Fruits by LC-ICP-MS

Abstract

This paper describes a liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS) method for the determination of low-level sulfite in dry vegetables and fruits. Aqueous 0.1% solution of formaldehyde was chosen as extraction solution for releasing the bound sulfite and forming the more stable hydroxymethanesulfonate (HMS). Sulfite was separated using an anion exchange column with isocratic elution by a 50 mmol•L⁻¹ ammonium nitrate in 0.1% formaldehyde solution (pH 7). Sulfite was detected as ³²S¹⁶O⁺ by ICP-MS in DRC O₂ gas mode. In the optimum conditions, the linearity of the method is 0.05 – 5 mg•L⁻¹ for sulfite, and the detection limit is 0.02 mg•L⁻¹. The results demonstrated that the method achieved acceptable quantitative recoveries of 78.5% to 91.6% with relative standard deviations (RSDs) < 5%. Compared to the AOAC official method, the proposed method presented advantages of rapid and sensitive.

Introduction

Sulfite salts or sulfur dioxide are widely used as an additive in food and beverages to prevent oxidation, microbial spoiling, and to control enzymatic browning reaction during the food producing and storage process (Danilewicz, 2003; Lee and Shibamoto, 2002; Thorpe and Baynes, 2003). However, the excessive intake of sulfite is harmful to humans. Previous work demonstrated that daily consumption of sulfite exceeded 50 mg•kg⁻¹ body weight might cause many diseases, such as headaches, nausea and diarrhea (Suh et al., 2007). Due to potential health concerns, the use of sulfur dioxide and sulfite as additives in food is legally restricted in USA, EU and China. The Food and Drug Administration (FDA) of United States even requires sulfite warning on the label of foods if in the concentrations of sulfite exceed 10 mg•kg⁻¹.

Dry vegetables and fruits which are removed most of water from fresh product have less loss of nutrient substances. Compared to fresh product, dry product has much longer storage life. However, some producers may add a certain amount of sulfite to extend the dry product shelf period even in the poor storage environment. Those who claim to be the feature of “no preservatives” or “totally natural” in dry vegetables and fruits may consider to be exaggerated. So a reliable, sensitive method for sulfite analysis in dry vegetables and fruits is required. The method should be sensitive enough for covering the concentration range below 10 mg•kg⁻¹.

Various methods have been applied to determine sulfite in foods. Classical and official method of China National Standard (CNS, 2003) and Association of Official Analytical Chemists (AOAC, 1995) are based on the optimized Monier-Williams (OMW), which utilize the distillation of the samples under acidic condition and then analyzed by iodine or acid/base titration. However, these methods usually required the complicated sample preparation processes and cannot be used for fast or high-throughput analysis. Moreover, for dry vegetable and fruit samples, the OMW method tends to overestimate sulfite levels because of the presence of abundant volatile acidic compounds (Pizzoferrato et al., 1998; Zhong et al., 2012). In recent years, many new methods have been reported, including electrochemistry (Chen et al., 2012; Isaac et al., 2006; Karimi-Maleh et al., 2012), ion chromatography (Koch et al., 2010; Zhong et al., 2012), liquid chromatography (Pizzoferrato, 1998; Theisen et al., 2010; S. Wang et al., 2007; Zuo and Chen, 2003), optical chemical sensors or biosensors (Alamo et al., 2010; Dinckaya et al., 2007; Filik and Çetintaş, 2012), flow injection (Navarrro et al., 2010; Ruiz-Capillas and Jiménez-Colmenero, 2009; Tzanavaras et al., 2009). Above-mentioned methods indeed greatly shorten the analysis time. The limit of detection (LOD) was even lower than 0.1 mg•L⁻¹ in sample solution with some methods. However, by these methods, complicated operations and reagents, self-made equipments were required which prevent them from being applied widely in rapid determination of sulfite in food samples.

In recent years, liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS) technology has been well development for the determination of some metal and metalloid species due to its easy connecting to LC. Dynamic reaction cell (DRC) technique with O₂ makes the determination of total sulphur and sulphur species easier (Bandura et al., 2002; Lin and Jiang, 2009; Schaumlöffel et al., 2007; Stürup et al., 2006; M. Wang et al., 2007; Yang and Jiang, 2004), which eliminate the serious isobaric interferences from ¹⁶O¹⁶O⁺, ¹⁴N¹⁸O⁺ to ³²S⁺ (abundance: 95.018%). However, there still have some issues need to take into account when applying LC-ICP-MS for the determination of sulfite in food matrix sample.

Our work aimed to develop a simple, sensitive and reliable method for the determination of sulfite in dry vegetables and fruits. The separation was achieved on an anion-exchange column followed by ICP-MS detection. In the process, extraction conditions were optimized, and potential matrix interference was discussed.

Materials and Methods

Reagents and standards    All of the chemicals used were of analytical-reagent grade. All solutions were prepared with deionized water. Sodium sulfite, sodium sulphate, sodium thiosulfate, ammonium nitrate and formaldehyde were purchased from SCRC (Shanghai, China). Sulfur-containing amino acids was purchased from Wako Pure Chemical Industries, Ltd (Osaka, Japan). Deionized water was purified to 18.2 MΩ resistances using a Milli-Q system (Millipore, Bedford, MA, USA). A sulfite stock solution (1000 mg•L⁻¹) was prepared daily by dissolving 0.1575 g sodium sulfite in 100 mL. The solution was standardized by iodimetric titration. The working solution of sulfite was prepared by diluting the stock solution with 0.1% formaldehyde (v/v) (pH 11, adjusted by 20% ammonia).

Apparatus    SK5200H ultrasonic generator was purchased from Shanghai Kudos Instruments Inc (Shanghai, China). Chromatographic separations were performed using a Flexar Binary LC pump (PerkinElmer, Shelton, CT, USA) was used. Injections were performed using a 7725i Rheodyne valve. The anion exchange column Hamilton PRP-X100 (10 µm, 250 × 4.1 mm i.d.; Hamilton Company, Reno, NV, USA) was used and was connected to the quartz concentric nebulizer with PEEK tubing. A Perkin Elmer NexIon 300D ICP-MS (Shelton, CT, USA) instrument was used for elemental detection.

Sample preparation    Dry vegetable and fruit samples were purchased from local farmer's market and supermarket in Hangzhou, China. Samples were comminuted by stainless steel grinding mill. An amount of 1.00 g of the dry sample was weighed into a 50-mL polypropylene centrifuge tube. First, 20 mL 0.1% formaldehyde (v/v) (pH 7, adjusted by 20% ammonia) was added to the tube, then the mixture shaken well and extracted by ultrasonic for 15 min before it was filtered with a 0.22 µm tetrafluoroethylene membrane filter for further LC-ICP-MS analysis. All samples were prepared and determined in the same day.

Sulfite analysis by LC-ICP-MS    LC-ICP-MS was used for identification and determination of sulfite. The chromatographic separation was performed on a Hamilton PRP-X100 anion exchange column (250 × 4.1 mm i.d., 10 µm). A 100 µL aliquot of the sample was applied to the column, and an isocratic elution at a flow rate of 1.5 mL•min⁻¹ was used. The mobile phase was composed of 50 mmol•L⁻¹ ammonium nitrate (pH 7, adjusted by 20% ammonia) in 0.1% formaldehyde (v/v). The instrumental parameters of ICP-MS measurements were as follows: RF power was set to 1400 W, analog stage voltage, pulse stage voltage and lens voltage were 1670 V, 1055 V and 6.80 V, the plasma and auxiliary argon flow rates were 15 and 1.1 L•min⁻¹, respectively. The data acquisition parameters were: dwell time: 500 ms, sweeps per reading: 1, readings per replicate: 750. High purity oxygen (99.999%) was used for dynamic reaction cell (DRC) and the DRC was optimized to the following parameters: RPa: 0, RPq: 0.8, cell gas: 0.8 mL•min⁻¹, QRO:−6.5 V, CRO: −1.3 V, CPV: −16 V.

Safety considerations    In this study, 0.1% formaldehyde solution was used as extraction solution and mobile phase. Formaldehyde has been classified as a known human carcinogen (cancer-causing substance) by the International Agency for Research on Cancer and as a probable human carcinogen by the U.S. Environmental Protection Agency. Researchers wishing to conduct similar experiments are urged to exercise great care in sample handling and to ensure that sample lines, dilution apparatus, and pump exhausts have no leaks and are properly exhausted into appropriate waste air and water handling systems.

Results and Discussion

Stability of the extraction solution    To solve the substantial sulfites loss during extraction process, formaldehyde was used to convert sulfite into a more stable hydroxymethanesulfonate (HMS). Long time stability experiment has been done in our research. The results were shown in Fig.1. It showed that during the extraction process, lower pH was advantageous to improve the sensitive of the SO₃²⁻ determination, but it was unstable due to oxidation happening after longtime extraction. While higher pH would sacrifice the determination sensitivity of SO₃²⁻, it was advantageous to stability of the HMS. The concentration of formaldehyde less affected the stability of the HMS when the concentration >0.1% (v/v). So, considering formaldehyde to the environmental impact, 0.1% (v/v) formaldehyde in alkaline solution (pH 11) was selected in this experiment. Applying this solution to extract SO₃²⁻ from dry vegetable and fruit samples, it was found that after 15 minutes extraction, all experimental samples tend to be stable.

Fig. 1.

SO₃²⁻(0.5 mg•L⁻¹) in different pH formaldehyde solution with ultrasonic extraction

Optimization of the DRC instrument    It is well known that ¹⁶O¹⁶O⁺, ¹⁴N¹⁸O⁺ and ¹⁵O¹⁷O⁺ interfere greatly with sulfur isotope (³²S⁺). In this work, DRC system with oxygen gas (O₂) was employed to reduce the background interferences in ICP -MS analysis. In DRC system, sulfur reacted with O₂ to form ³²S¹⁶O⁺, which eliminating the interferences of isobars. The RPq and cell gas (O₂) flow parameters were of major importance to affect the operation of the dynamic reaction cell. When the RPq value was 0.45 and O₂ flow was 0.8 mL•min⁻¹, best S/N (the ratio of single to noise) was received. Since ³²S¹⁶O⁺ was chosen as the quantitative signal, the interference of ⁴⁸Ca⁺ should be taken into account. Dry vegetables and fruits contain a considerable number of water-solubility Ca ions. Researchers (Yang and Jiang, 2004) have shown that the significant interference of ⁴⁸Ca⁺ to ³²S¹⁶O⁺, when they had similar quantities in water solution. Fortunately, in our study, it had no signal peak after separating by LC system, even the concentration of Ca increased to 100 mg•L⁻¹ in water solution.

Interference by water-soluble sulfocompound    Except SO₃²⁻, some water-soluble sulfur compounds, such as sulfate (SO₄²⁻), thiosulfate (S₂O₃²⁻), sulfur-containing amino acids (taurine, methionine, cysteine and cystine) which could be found in dry vegetables and fruits may coelute with SO₃²⁻. Fifty mmol•L⁻¹ ammonium nitrate (NH₄NO₃) was chosen as the mobile phase. The pH value of mobile phase was irrelevant to the separation of the SO₃²⁻ from SO₄²⁻ and S2O₃²⁻. However, for sulfur-containing amino acids, when the pH value decreased to 5, the peaks of sulfur-containing amino acids overlapped considerably with that of SO₃²⁻. By increasing the pH to 7, the retention time became longer resulting in adequate separation of sulfur-containing amino acids and SO₃²⁻. In this pH condition, SO₃²⁻ was well separated from sulfate (SO₄²⁻), thiosulfate (S₂O₃²⁻) and sulfur-containing amino acids (Fig.2). The total runtime was approximately 5 minutes.

Fig. 2.

Chromatogram of (1) SO₃²⁻(0.5 mg L⁻¹) standard solution coexisting (2) SO₄²⁻, (3) S₂O₃²⁻ (dashed line) and (4 + 5) sulfur-containing amino acid (solid line) exuding by 50 mmol·L⁻¹ NH₄NO₃ (pH 7)

In this experiment, different types of dry vegetable and fruit samples with different matrix were tested in our experiment. All these samples didn't contain any sulfite (blank samples). The blank samples and blank samples added 0.5 mg•L⁻¹ sulfite were extracted by 0.1% formaldehyde, and tested by LC-ICP-MS. Typical chromatograms were shown in Fig.3 (a, b, c d). It showed that for different matrix samples, no interference peak appear at the retention time of SO₃²⁻, and other sulfur compounds were well separated from SO₃²⁻. In the meantime, the retention time of the sulfite in different matrix samples corresponded to that in the chromatogram of the sulfite standard solution (Figure 2). Extending the elution program time to 30 min, no other peak was found. Five minutes was long enough to separate SO₃²⁻ to other interference substance. This program was faster than reported ion chromatography methods (Koch et al., 2010; Zhong et al., 2012) and HPLC methods (Pizzoferrato et al., 1998; Theisen et al., 2010; S. Wang et al., 2007; Zuo and Chen, 2003).

Fig. 3.

Chromatograms of different blank matrix dry vegetable and fruit samples (solid line) with 0.5 mg L⁻¹ SO₃²⁻ standard solution (dashed line) (a: dry mushroom, b: dry garlic, c: dry mango and d: dry kiwi fruit)

Method validation    At the optimized conditions, the procedure was validated in terms of linearity, limits of detection and quantitation, precision, and accuracy.

The linear range of the established LC-ICP-MS method was 0.05 – 5 mg•L⁻¹ for sulfite, following the regression equation: Y = 2.2 × 10⁶X ⁺ 4.3 × 10³. The limit of detection (LOD) as three times the SD from the blank samples (n = 10) with the addition of the SO₃²⁻ concentration, which was close to the expected LOD. The limit of quantification (LOQ) was calculated as three times the LOD value.

The LOD and LOQ values of SO₃²⁻ were 0.02 mg•L⁻¹ and 0.06 mg•L⁻¹, respectively. It should be possible to determine the concentration of the sulfite under 10 mg•kg⁻¹ from this LOQ.

Accuracy and precision of the analytical method were calculated by the determination of six replicates of different samples at spiked with sulfite standard at 2, 10 and 50 mg•kg⁻¹. Accuracy and precision of the method were evaluated by means of recovery and relative standard deviation (RSD). The results are shown in Table 1 (the background contents of sulfite in dry mushroom, dry garlic, dry mango, and dry kiwi fruit were derived from sulfite added in the laboratory). The mean recoveries of the sulfite at three spiked levels were ranged from 78.5% to 91.6%, with RSD ranging from 2.5% to 4.8%.

Table 1. Recoveries for the determination of sulfite in selected dry vegetable and fruit samples (n = 6)

Sample	Backgroud (mg•kg−1)	Added (mg•kg−1)	Recovery (%)	RSD (%)
dry onion	ND a	2.00	78.5%	2.8
dry jujube	ND	2.00	86.5%	2.5
dry mushroom	10.1	10.00	87.6%	4.3
dry garlic	8.55	10.00	84.2%	3.4
dry mango	65.8	50.0	89.0%	4.8
dry kiwi fruit	50.1	50.0	91.6%	3.9

a.  not detected

Application to real samples and method comparison    The selected method was applied to analysis of sulfite in 4 laboratory added samples and 10 samples collected local from farmer's market and supermarket in Hangzhou. Sulfite was found in 2 collected samples which were dry raisin (8.19 mg•kg⁻¹) and dry bamboo shoot (68.8 mg•kg⁻¹) respectively. These 2 samples with another 4 laboratory added samples were also determined by AOAC Official Method 990.28 for method comparison. Student's t test was employed to estimate the significance of values. For high-level sulfite samples, at the 95% confidence level, |t_crit| is 2.120, and |t_calc| values were lower than |t_critis| (Table 2), there was no significant difference between the analytical values by the proposed and official methods. For low-level sulfite samples, |t_calc| values were higher than |t_critis|, which meant significant difference between the analytical values by the two methods. Because by official method, abundant volatile organic acids in vegetable and fruit were also titrated with standardized NaOH, which probably caused the results higher (Pizzoferrato et al., 1998). On the other hand, high LOD (10 mg•kg⁻¹) of official method may also lead to significant difference.

Table 2. Comparison of LC-ICP-MS method and AOAC official method on quantifying sulfite in selected dry vegetables and fruits samples (n = 9)

Sample	LC-ICP-MS method	AOAC official method	|tcalc| value
	Content (mg•kg−1)	Content (mg•kg−1)
dry raisin	8.19 ± 0.30	10.9 ± 1.4	4.166
dry bamboo shoot	68.8 ± 1.8	71.7 ± 1.8	2.014
dry mushroom (laboratory added)	10.1 ± 0.5	13.7 ± 2.4	6.301
dry garlic (laboratory added)	8.55 ± 0.23	11.0 ± 1.7	3.167
dry mango (laboratory added)	65.8 ± 1.6	67.6 ± 1.9	1.857
dry kiwi fruit (laboratory added)	50.1 ± 1.5	51.8 ± 1.2	1.978

Conclusion

In this study, a novel LC-ICP-MS has been developed for the determination of sulfite in dry vegetables and fruits. The proposal method was able to eliminate the interferences effectively, and showed highly selective. The proposed method show favourable recovery and RSD when determination of sulfite under 10 mg•kg⁻¹. Rapid separation with very low detection limit was the most advantage of the proposed method, and it was successfully applied for the determination of sulfite in dry vegetables and fruits.

Acknowledgements    This work was supported by the Special Funds from the Central Scientific Research Institute of Public Welfare (RISF2013011), Special Fund for Forestry Scientific Research in the Public Interest (201304705).

References

• AOAC International. Method 990.28. Official Methods of Analysis of AOAC International, 16th ed.;  Cunniff,  P., Ed.; The Association:Arlington, VA, 1995; Chapter 47, pp 27-29.
•  Alamo,  L. S.,  Tangkuaram,  T., and  Satienperakul,  S. (2010). Determination of sulfite by pervaporation-flow injection with amperometric detection using copper hexacyanoferrate-carbon nanotube modified carbon paste electrode. Talanta., 81, 1793-1799.
•  Bandura,  D. R.,  Baranov,  V. I., and  Tanner,  S. D. (2002). Detection of ultratrace phosphorus and sulfur by quadrupole ICPMS with dynamic reaction cell. Anal. Chem., 74, 1497-1502.
•  Chen,  P.-Y.,  Huang,  C.-C.,  Chen,  M.-C.,  Hsu,  J.-C., and  Shih,  Y. (2012). Determination of sulfite in hair waving products using oxygen-incorporated gold-modified screen-printed electrodes. Electroanaly., 24, 2267-2272.
•  Danilewicz,  J. C. (2003). Review of reaction mechanisms of oxygen and proposed intermediate reduction products in wine: Central role of iron and copper. Am. J. Enol. Vitic., 54, 73-85.
•  Dinckaya,  E.,  Sezginturk,  M.,  Akyilmaz,  E., and  Ertas,  F. (2007). Sulfite determination using sulfite oxidase biosensor based glassy carbon electrode coated with thin mercury film. Food. Chem., 101, 1540-1544.
•  Filik,  H. and  Çetintaş,  G. (2012). Determination of sulfite in water and dried fruit samples by dispersive liquid-liquid microextraction combined with UV-Vis fiber optic linear array spectrophotometry. Food. Anal. Method, 5, 1362-1367.
•  Isaac,  A.,  Davis,  J.,  Livingstone,  C.,  Wain,  A. J., and  Compton,  R. G. (2006). Electroanalytical methods for the determination of sulfite in food and beverages. TrAC Trend. Anal. Chem., 25, 589-598.
•  Karimi-Maleh,  H.,  Ensafi,  A. A.,  Beitollahi,  H.,  Nasiri,  V.,  Khalilzadeh,  M. A., and  Biparva,  P. (2012). Electrocatalytic determination of sulfite using a modified carbon nanotubes paste electrode: application for determination of sulfite in real samples. Ionics., 18, 687-694.
•  Koch,  M.,  Koppen,  R.,  Siegel,  D.,  Witt,  A., and  Nehls,  I. (2010).
•  Ruiz-Capillas,  C. and  Jiménez-Colmenero,  F. (2009). Application of flow injection analysis for determining sulphites in food and beverages: A review. Food. Chem., 112, 487-493.
•  Schaumlöffel,  D.,  Giusti,  P.,  Preud'Homme,  H.,  Szpunar,  J., and  Lobinski,  R. (2007). Precolumn isotope dilution analysis in nano HPLC-ICPMS for absolute quantification of sulfur-containing peptides. Anal. Chem., 79, 2859-2868.
•  Stürup,  S.,  Bendahl,  L., and  Gammelgaard,  B. (2006). Optimization of LC- DRC-ICP-MS for the speciation of selenotrisulfides with simultaneous detection of sulfur and selenium as oxides combined with determination of elemental and isotope ratios. J. Anal. At. Spectrom., 21, 201-203.
•  Suh,  H.-J.,  Cho,  Y.-H.,  Chung,  M.-S., and  Hee Kim,  B. (2007). Preliminary data on sulphite intake from the Korean diet. J. Food. Compos. Anal., 20, 212-219.
•  Theisen,  S.,  Hansch,  R.,  Kothe,  L.,  Leist,  U., and  Galensa,  R. (2010). A fast and sensitive HPLC method for sulfite analysis in food based on a plant sulfite oxidase biosensor. Biosens. Bioelectron., 26, 175-181.
•  Thorpe,  S. and  Baynes,  J. (2003). Maillard reaction products in tissue proteins: new products and new perspectives. Amino. Acids, 25, 275-281.
•  Tzanavaras,  P. D.,  Thiakouli,  E., and  Themelis,  D. G. (2009). Hybrid sequential injection-flow injection manifold for the spectrophotometric determination of total sulfite in wines using o-phthalaldehyde and gas-diffusion. Talanta., 77, 1614-1619.
•  Wang,  M.,  Feng,  W.,  Lu,  W.,  Li,  B.,  Wang,  B.,  Zhu,  M.,  Wang,  Y.,  Yuan,  H.,  Zhao,  Y., and  Chai,  Z. (2007). Quantitative analysis of proteins via sulfur determination by HPLC coupled to isotope dilution ICPMS with a hexapole collision cell. Anal. Chem., 79, 9128-9134.
•  Wang,  S.,  Cui,  X., and  Fang,  G. (2007). Rapid determination of formaldehyde and sulfur dioxide in food products and Chinese herbals. Food. Chem., 103, 1487-1493.
•  Yang,  C.-H. and  Jiang,  S.-J. (2004). Determination of B, Si, P and S in steels by inductively coupled plasma quadrupole mass spectrometry with dynamic reaction cell. Spectrochim. Acta. B., 59, 1389-1394.
•  Zhong,  Z.,  Li,  G.,  Zhu,  B.,  Luo,  Z.,  Huang,  L., and  Wu,  X. (2012). A rapid distillation method coupled with ion chromatography for the determination of total sulphur dioxide in foods. Food. Chem., 131, 1044-1050.
•  Zuo,  Y. and  Chen,  H. (2003). Simultaneous determination of sulfite, sulfate, and hydroxymethanesulfonate in atmospheric waters by ion-pair HPLC technique. Talanta., 59, 875-881.