Inductively

Mar 9, 2015 - We report an analytical methodology for the quantification of common arsenic species in rice and rice cereal using capillary electrophor...
8 downloads 0 Views 2MB Size
Article pubs.acs.org/JAFC

Arsenic Speciation in Rice by Capillary Electrophoresis/Inductively Coupled Plasma Mass Spectrometry: Enzyme-Assisted Water-Phase Microwave Digestion Haiou Qu, Thilak K. Mudalige,* and Sean W. Linder* Arkansas Regional Laboratory, Office of Regulatory Affairs, United States Food and Drug Administration (FDA), 3900 NCTR Road, Jefferson, Arkansas 72079, United States S Supporting Information *

ABSTRACT: We report an analytical methodology for the quantification of common arsenic species in rice and rice cereal using capillary electrophoresis coupled with inductively coupled plasma mass spectrometry (CE−ICPMS). An enzyme (i.e., αamylase)-assisted water-phase microwave extraction procedure was used to extract four common arsenic species, including dimethylarsinic acid (DMA), monomethylarsonic acid (MMA), arsenite [As(III)], and arsenate [As(V)] from the rice matrices. The addition of the enzyme α-amylase during the extraction process was necessary to reduce the sample viscosity, which subsequently increased the injection volume and enhanced the signal response. o-Arsanilic acid (o-ASA) was added to the sample solution as a mobility marker and internal standard. The obtained repeatability [i.e., relative standard deviation (RSD %)] of the four arsenic analytes of interest was less than 1.23% for elution time and 2.91% for peak area. The detection limits were determined to be 0.15−0.27 ng g−1. Rice standard reference materials SRM 1568b and CRM 7503-a were used to validate this method. The quantitative concentrations of each organic arsenic and summed inorganic arsenic were found within 5% difference of the certified values of the two reference materials. KEYWORDS: arsenic speciation, rice, capillary electrophoresis, inductively coupled plasma mass spectrometry, enzyme-assisted microwave digestion



INTRODUCTION Exposure to inorganic arsenic via drinking water has long been a concern of both public health agencies and scientists alike, but more recent studies have led to concerns about exposure to inorganic arsenic from food.1 Among all of the food commodities, rice and rice products are of particular interest because rice plants have higher accumulation rates of arsenic when compared to other crops.2,3 In 2014, the World Health Organization proposed a draft maximum level of 0.2 mg kg−1 for inorganic arsenic in polished rice.4 The United States Food and Drug Administration (FDA) has been monitoring the levels of arsenic in foods for decades. In 2012 and 2013, the FDA released findings from a large-scale survey of arsenic in rice.5,6 Approximately 1300 samples of rice and rice products were analyzed, and their arsenic levels ranged from below the detection limit to 1931 ng g−1.6 There are multiple forms of arsenic compounds in rice contributing to the total arsenic level. However, the toxicity and bioavailability of arsenic highly depends upon its chemical form. While inorganic arsenic compounds As(III) and As(V) are considered to be class I human carcinogens, a number of organic forms of arsenic, such as dimethylarsinic acid (DMA) and monomethylarsonic acid (MMA), are much less toxic.7 Additionally, there is evidence suggesting that some trivalent methylated arsenic species are also highly toxic. The measurement of total arsenic in rice, therefore, is unable to provide sufficient information for an accurate risk assessment. A speciation technique that can separate and quantify individual arsenic species is needed to evaluate the potential health hazard. This article not subject to U.S. Copyright. Published 2015 by the American Chemical Society

Many analytical methodologies have been established to speciate various arsenic compounds and measure their concentration in different matrices. Because of the intrinsic nature of low arsenic concentrations in rice and rice products, a hyphenated system, consisting of a high-resolution separation stage, followed by an ultrasensitive detection system for the quantification of each eluted compound, is the most applied strategy for arsenic speciation in rice products.8−10 For the separation stage, most reports focused on chromatographic methods, such as high-performance liquid chromatography (HPLC).8,11−13 One major drawback of these methods is column deterioration because of the residual matrix components, such as carbohydrates in the sample, which can lead to poor resolution and precision as the number of injections increases.11,14 Additionally, the presence of a few unknown arsenic species could interfere with the determination of As(III).15 A frequently used solution is to completely oxidize As(III) into As(V) and, subsequently, quantify only the amount of As(V) as an assessment of total inorganic arsenic.15 However, hydrogen peroxide residue left over from the oxidation treatment can also cause column deterioration. Capillary electrophoresis (CE) has been considered to be a versatile separation technique with excellent resolution for both small molecules and macromolecules.16−18 Because the Received: Revised: Accepted: Published: 3153

January 23, 2015 March 6, 2015 March 9, 2015 March 9, 2015 DOI: 10.1021/acs.jafc.5b00446 J. Agric. Food Chem. 2015, 63, 3153−3160

Journal of Agricultural and Food Chemistry

Article

As(III), and As(V) stock solutions with Milli-Q water. For the rice flour standard reference materials, SRM 1568b was purchased from NIST (Gaithersburg, MD) and CRM 7503-a [prepared by the National Metrology Institute of Japan (NMIJ)] was purchased from Waco Chemicals US (Richmond, VA). Rice (polished long grain rice) and rice cereal (single-grain infant cereal) were acquired from a local market. Instrumentation. CE was performed on a 7100 CE system (Agilent Technology, Santa Clara, CA). A 60 cm long coated fusedsilica capillary (Molex, Phoenix, AZ) with a 100 μm inner diameter and a 360 μm outer diameter was used for sample introduction. The separation of five arsenic species was carried out under basic condition through the use of Na2CO3 (8 mM, pH 11) buffer. A new capillary was initialized by flushing 1 M KOH at 950 mbar for 30 min and running buffer (8 mM Na2CO3, pH 11) for 30 min, followed by rinsing with water for 5 min. The capillary was conditioned with 0.1 M KOH and running buffer for 15 min prior the analysis. Before each run, the capillary was equilibrated with 0.1 M KOH for 3 min and running buffer for 3 min. After each run, the capillary was rinsed with water for 1 min. Samples were hydrodynamically injected at 15 mbar for 8 s, followed by an injection of running buffer at 15 mbar for 5 s. The temperature of the cartridge was set at 21 °C (ambient temperature), and the applied voltage for the separation was 20 kV. The coupling of CE with ICPMS (Agilent 7700x, Agilent Technology, Santa Clara, CA) was performed by directly connecting the outlet of the capillary to a Mira Mist CE nebulizer (Burgener Research, Inc., Mississauga, Ontario, Canada). A solution containing 1% HNO3 (v/v) and 10% 2propanol (v/v) was used as the makeup solution. Alcohol was added to the makeup solution to lower the liquid surface tension and facilitate the formation of an aerosol. Additionally, alcohol was used as a carbon source to enhance the signal intensity and overcome matrix effects.28 A final concentration of 10% 2-propanol was chosen for the analysis. Although a higher alcohol concentration can be used,29 carbon deposits on the sampling cone can occur and affect system stability. The peristaltic pump speed was set at 0.06 rpm using a 0.5 mm inner diameter tube. This configuration gives a flow rate of 15 μL min−1, with no visual droplets formulating inside the spay chamber, suggesting an approximate 100% sample delivery efficiency. The pH value of the running buffer was adjusted by adding 0.1 M NaOH or 0.1 M HNO3 solutions and was measured with an Orion Star A214 pH meter (Thermo Scientific, Waltham, MA). The ICPMS was tuned daily under reaction gas (H2) mode. The polyatomic interference of ArCl+ is eliminated by hydrogen gas through a proton-transfer process.30 The mass isotopes of m/z 75 and 77 were monitored for arsenic and the interference. Detailed operation conditions are listed in Table 1. The viscosity of the rice extract was measured on a VISCOlab

separation mechanism in CE is different from those in liquid chromatography, it is considered as an orthogonal analysis and provides complementary information.19 For online detection of species after separation, the detection method must possess sufficient sensitivity and be able to interface with the separation system. Among all of the techniques that have been applied for elemental analysis, atomic spectrometry has played a crucial role in the determination of the arsenic concentration in a variety of samples.8−10 Earlier development works involved the use of atomic emission spectrometry (AES) or atomic absorption spectrometry (AAS) coupled with liquid chromatography for arsenic speciation.20,21 In recent years, with the continuous improvement in sensitivity, inductively coupled plasma mass spectrometry (ICPMS) has become the most popular detector for elemental analysis. It offers several advantages over other atomic spectroscopy techniques, including improved detection limits, minimization of interferences with the application of reaction/collision cell gases, capability to easily interface with other instruments, and the simultaneous monitoring of multiple isotopes. It has been demonstrated to be an essential tool in the routine analysis of arsenic species.14,15,22,23 CE interfaced with ICPMS has been widely used in the speciation of elements in various matrices. Although there have been reports about speciation of selenium, arsenic, and other elemental species in wastewater, soil, and seafood matrices using CE,23−27 quantitation of common arsenic species in rice products using CE still remains relatively unexplored. In this study, we report the development of a hyphenated analytical technique consisting of CE and ICPMS for arsenic speciation in a rice matrix. An enzyme-assisted water-phase microwave digestion procedure was tested and used for the extraction. Multiple conditions that would affect the separation or detection of rice samples were evaluated. The optimized conditions were used for analysis of locally purchased rice samples. The precision and accuracy of this method was validated by analyzing rice standard reference materials and comparing the results to certified values.



EXPERIMENTAL SECTION

Materials and Chemicals. Methylarsonic acid monosodium salt (MMA, 98.5%) was purchased from Crescent Chemicals (Islandia, NY). Dimethylarsinic acid (DMA, 98%) was purchased from Strem Chemicals (Newburyport, MA), and o-arsanilic acid (o-ASA, 98%) was purchased from Pfaltz & Bauer (Waterbury, CT). National Institute of Standards and Technology (NIST)-traceable arsenite [As(III)] and arsenate [As(V)] stock solutions (1000 mg L−1) were purchased from Spex Certiprep (Metuchen, NJ). All chemicals that were used to prepare calibration standards were assayed using ICPMS to verify the arsenic concentration. The enzyme, α-amylase from Bacillus subtilis [165 000 bacterial amylase units (BAU)/g], was purchased from MP Biomedicals (Carlsbad, CA). TraceSELECT sodium carbonate (≥99.9999% metal basis), TraceSELECT sodium hydroxide monohydrate (≥99.9995% metal basis), and TraceSELECT 2-propanol (≥99.9%) were purchased from Sigma-Aldrich (St. Louis, MO). Potassium hydroxide pellets (>85%) and nitric acid (68−70%, OPTIMA ultrapure grade) were purchased from Fisher Scientific (Waltham, MA). All chemicals were used as received without further purification. Deionized water (>18 MΩ cm−1) from a Milli-Q reference system (Millipore, Billerica, MA) was used throughout the experiments. Individual stock solutions (1000 ng g−1) of DMA, MMA, o-ASA, As(III), and As(V) were prepared with Milli-Q water and stored at 4 °C in the dark. The As concentrations of the stock solutions were analyzed by ICPMS. Multi-analyte calibration standards were prepared fresh on the day of use by diluting DMA, MMA, o-ASA,

Table 1. CE−ICPMS Operating Parameters capillary running electrolyte voltage temperature sample injection RF power sample depth plasma gas carrier gas makeup gas collision/reaction gas monitored isotope (m/z) dwell time nebulizer 3154

CE Parameters polymer-coated fused-silica capillary, 100 μm inner diameter, 360 μm outer diameter, and 60 cm length 8 mM Na2CO3, pH 11.0 20 kV 21 °C 15 mbar, 8 s, hydrodynamic ICPMS Parameters 1500 W 8.0 mm 15.0 L min−1 0.9 L min−1 0.45 L min−1 H2, 2 mL/min 75 (As+), 77 (ArCl+) 0.4 s (75As), 0.1 s (77Se) Mira Mist CE DOI: 10.1021/acs.jafc.5b00446 J. Agric. Food Chem. 2015, 63, 3153−3160

Journal of Agricultural and Food Chemistry

Article

5000 viscometer with temperature control (Cambridge Viscosity, Medford, MA). The measurement of viscosity was performed in triplicate. Sample Preparation. Rice flour of NIST SRM 1568b and CRM 7503-a were received as fine powder and used directly for the extraction. Rice and rice cereal were first ground with an 8000 M mixer/mill (SPEX SamplePrep, Metuchen, NJ) and sieved using 0.1 mm mesh size. Rice powders were oven-dried at 90 °C for 2 h to determine the residual moisture content. An enzyme-assisted waterphase microwave extraction was used for sample extraction. Typically, in a 15 mL glass tube, 0.2 g of rice powder and 0.1 g of 0.5% (w/w) αamylase solution were diluted with water to a total weight of 2 g, which gave a final enzyme concentration of 0.05% (w/w). The tube was capped and put in a Discover SP microwave digestion system (CEM, Matthews, NC) at 80 °C for 1 h and 90 °C for 2 h with stirring. No pressure was generated inside the tube during the extraction. A method blank, which contains only 0.05% α-amylase, was also included in each extraction batch as a control to verify the absence of arsenic species. The extracted mixture was centrifuged, and any remaining solids were removed. An aliquot of 30 μL of o-ASA (50 ng g−1) was added to the 270 μL of rice extract for a final concentration of 5.0 ng g−1 of o-ASA. Each sample was prepared in triplicate. All of the standards and samples were prepared gravimetrically.



RESULTS AND DISCUSSION Optimization of the Separation Condition. The target arsenic species in this study are DMA, MMA, As(III), and As(V). In addition, o-ASA, which is not present in the rice, was also added for this analysis. It served as an internal standard and mobility marker for the other four analytes of interests. The five arsenic species were separated with the use of CE and then directly delivered to ICPMS for detection. Several factors that may affect the separation efficiency were evaluated. Separation conditions were first investigated at various buffer concentrations (i.e., 3−10 mM). The results suggested that a buffer solution with a concentration as low as 3 mM can be used to obtain acceptable resolution among the five tested arsenic species. From Figure 1a, baseline separations were achieved in all assayed conditions with enhanced peak resolution as the buffer concentration increased. Higher buffer concentrations led to stronger ionic strengths, which can suppress the electron double layer on the inner surface of the capillary, reducing electroosmotic flow (EOF) and resulting in longer migration times. Furthermore, a concentrated buffer solution largely increases the electric current in the capillary and results in larger variations because of Joule heating.19 After the resolution and analysis duration of each run were balanced, 8 mM Na2CO3 was selected as the running buffer for all of the measurements. Similar to the buffer concentration, the applied voltage had a strong effect on the EOF, which affected the migration speed of all analytes. However, it only exhibited a limited contribution to the peak separation. As shown in Figure 1b, voltages as low as 8 kV could resolve all of the species of interest. Because the magnitude of EOF generated from the low voltage is relatively small, the migration of the analytes is slow, leading to a longer elution time. Increasing the applied voltage was effective in shortening the analysis time without compromising the resolution. A voltage of 20 kV was found to be an optimum value. Although higher voltages can further reduce the analysis time, these voltages caused a significant loss in peak resolution. This is especially true for DMA and As(III). Furthermore, higher voltages also generate strong currents that can lead to an undesired heating effect, especially for capillaries with a large inner diameter. In other buffer systems, similar

Figure 1. (a) Separation of five arsenic species under different buffer concentrations (voltage, 15 kV), and (b) effect of the applied voltage on the migration of arsenic species (buffer concentration, 5 mM). Other conditions: buffer, Na2CO3, pH, 11.

effects of the buffer concentration and voltage on the migration of arsenic species have been observed.26,27 The effects of pH on the separation of arsenic species were also evaluated. A basic condition with positive polarity was selected for CE separation to take advantage of the strong EOF toward the cathode. From Table 2, the acid dissociation constants (pKa) of five As species ranged from 0.7 to 9.2. To maintain the magnitude of the EOF and avoid repeatability problems commonly encountered at weak basic conditions, moderate to strong basic conditions are favorable.31 Because As(V) is eluted last, its migration ultimately determines the duration of the analysis. It was found that higher pH conditions Table 2. Acid Dissociation Constants of As Species31,35

3155

As species

formula

pKa

As(V) As(III) DMA MMA o-ASA

OAs(OH)3 OAs(OH) (CH3)2AsO(OH) CH3AsO(OH)2 H2NC6H4AsO(OH)2

2.3, 6.8, 11.6 9.2 6.2 3.6, 8.2 0.7, 3.8, 8.9

DOI: 10.1021/acs.jafc.5b00446 J. Agric. Food Chem. 2015, 63, 3153−3160

Journal of Agricultural and Food Chemistry

Article

the two species were resolved at higher pH conditions. To verify the peak position of DMA and As(III), we analyzed a sample containing 10 ng g−1 of DMA and 5 ng g−1 of As(III) within the same pH range. It is clearly shown in Figure 2b that the elution time of As(III) and DMA continually changed as the pH increased. Both injection pressure and time are factors that determine the injection plug length, which directly affects the precision of the method. Different combinations of pressure and time were tested to find the optimal conditions (see Table S-1 of the Supporting Information). All of the measurements with injection pressures above 30 mbar resulted in relative standard deviations (RSD %) higher than 10%. It is very clear that high injection pressures combined with relatively short injection times resulted in poor repeatability. A high applied pressure usually requires a much longer injection time to equilibrate the vacuum level change inside the capillary.32 However, this would increase the plug length and enlarge the variation.19 When a low pressure with a moderate injection time was applied, the precision was significantly improved. With an injection pressure of 15 mbar, the RSD % was reduced to an average of 5%. After both injection volume and precision of the method were considered, an applied pressure of 15 mbar with 8 s of injection time was chosen for the analysis. Additionally, in our study, the application of an internal standard mixed with sample effectively negated the variance caused by inconsistent injection volumes. Analytical Performance. Figure 3 is a typical electropherogram of the separation of five arsenic compounds under

generated a faster EOF, increasing the migration speed of As(V), therefore reducing its overall retention time. Special attention was paid to DMA and As(III) for the accurate determination of their identities. In an earlier study, Liu and coworkers studied the migration behavior of multiple arsenic species in HBO3 and NaH2PO4 buffers at different pH conditions. They found that, as the pH values of the buffer solution increased from 8.5 to 9.8, the elution time of As(III) also increased, whereas the elution time of DMA was almost unaffected.26 We have observed a similar trend in our study. Furthermore, we discovered (Figure 2a) that, when the pH

Figure 3. Electropherogram of a standard solution containing 5 ng g−1 of five arsenic species. m/z 77 was selected to monitor ArCl+ (40Ar and 37 Cl) interference.

Figure 2. (a) Effect of the pH value on the migration of five arsenic species [1, DMA; 2, As(III); 3, o-ASA; 4, MMA; and 5, As(V)] and (b) electropherogram of a mixture containing 10 ng g−1 of DMA and 5 ng g−1 of As(III).

optimized conditions (as presented in Table 1) at a level of 5 ng g−1. Baseline separation was achieved for all species in less than 10 min. Hydrogen gas (H2) was used as the reaction gas to eliminate the polyatomic interference of ArCl+. A baseline was observed for m/z 77 (40Ar37Cl+) during the assay, demonstrating the absence of interference. A calibration range of 0.5−20 ng g−1 was analyzed, and a linear relationship between the concentration and peak area was plotted. From previous studies, most of the rice samples surveyed have less than 200 ng g−1 of each arsenic species.6,13,14 Because the extraction procedure usually involves a dilution factor of 10−20, our calibration range is sufficient to analyze most of the rice

value of the running buffer increased beyond 10, the elution order of the peaks for DMA and As(III) reversed. The separation between As(III) and DMA first became less resolved at a pH of approximately 10 and then recovered at higher pH conditions. Because As(III) has a pKa value of 9.2, it gradually acquired a negative charge when the pH was adjusted from 8.9 to 11. However, because DMA remained negatively charged throughout the entire range, its electrophoretic mobility was surpassed by that of As(III) at around pH 10 and the peaks of 3156

DOI: 10.1021/acs.jafc.5b00446 J. Agric. Food Chem. 2015, 63, 3153−3160

Journal of Agricultural and Food Chemistry

Article

extraction methods. Upon evaluation of the analytical results, it was concluded that the two different extraction methods provide equivalent results. We believe that the differences in signal response of samples prepared from two methods were primarily attributed to the viscosity of the extraction solution. In CE, the injection volume (Vinj) can be calculated by

products on the market. To further evaluate the method linearity, a standard spike (50 ng g−1) at approximate 2.5 times the highest calibrator was analyzed. The results (see Table S-2 of the Supporting Information) suggested that acceptable recoveries were reached for all analytes. As listed in Table 3, the Table 3. Analytical Performance of CE−ICPMS RSD %a

compound

linear range (ng g−1)

R2

migration time

peak area

DMA MMA As(III) As(V)

0.5−20 0.5−20 0.5−20 0.5−20

0.998 0.999 0.999 0.997

0.82 1.08 0.75 1.23

2.62 2.08 2.62 2.91

Vinj =

LOD LOQ (ng g−1)b (ng g−1) 0.27 0.16 0.15 0.24

ΔPd 4πt inj 128ηL

(1)

where ΔP is the pressure difference across the capillary, d is the capillary inner diameter, tinj is the injection time, η is the buffer viscosity, and L is the total capillary length. When the viscosity of the sample solution increases, it will result in a decrease in total injection volume and a lower signal response. We have measured the viscosity of sample solutions from the extraction procedures with and without the addition of α-amylase. The results show that the viscosity is 7.5 ± 0.1 mPa s (T = 23.3 °C) in samples (rice 1) extracted with only water. If α-amylase is introduced into the extraction procedure, the viscosity dropped to 1.3 ± 0.1 mPa s (T = 23.3 °C). Additionally, a similar signal loss was also observed for o-ASA in samples extracted without the enzyme, and in both cases, the signal ratios of o-ASA to analytes of interest were the same. Because o-ASA was added as an internal standard post-extraction to both samples, the results support the conclusion that the signal difference is from sample viscosity rather than the extraction efficiency. Under optimized conditions, common arsenic species in rice were successfully separated and detected by the CE−ICPMS methodology (Figure 4). Increased migration times were observed in the capillary column during analysis. We believe that this can be attributed to the dissolved sugar molecules left in the sample solution, which affected the movement of the analytes in the capillary. In addition to the four analytes of interest in this study, a trace amount of unknown arsenic species produced unidentified peaks. Rice and rice cereal (denoted as rice 1 and rice 2) acquired from a local market were analyzed. The rice powder was divided into two portions. One portion was digested with nitric acid for a total arsenic measurement. The second portion was first treated with the enzyme-assisted microwave digestion, and the supernatant was further analyzed by the CE−ICPMS method. From Table 4, the summation of the four arsenic species in the rice samples generated a quantitative value very close to the total arsenic amount. The recoveries were 96.2 and 92.2% for rice 1 and rice 2, respectively. Previous studies have suggested the possible existence of tetramethylarsonium and other trace amounts of arsenic species in rice grain.13 These species may contribute up to 10% of the total arsenic amount, but in our case, the intensity of unknown peaks were insignificant. To assess the reliability of this method, two rice standard reference materials (NIST SRM 1568b and NMIJ CRM 7503-a) were assayed. The analytical results were compared to certified values. As summarized in Table 4, the quantitative values of organic arsenic and the sum value of inorganic arsenic were within 5% of the certified values of the two reference materials. Small differences between the quantitative and certified values for As(III) and As(V) were observed in rice flour CRM 7503-a reference materials. We believed that these differences were due to a well-documented interspecies conversion that may occur during the extraction process.8,11

0.87 0.52 0.47 0.75

a Sample concentration = 5 ng g−1 (n = 7). bInstrument detection limit.

RSD % (n = 7) for migration time ranged from 0.75 to 1.23% for four arsenic species. After adjustment for the signal variation, the RSD % of the peak area was between 2.08 and 2.91%, demonstrating satisfactory repeatability for this method. The limit of detection (LOD, 3σ criteria) was determined to be 0.27, 0.16, 0.15, and 0.24 ng g−1 for DMA, MMA, As(III), and As(V), respectively (Table 3). Analysis of the Arsenic Concentration in Rice and Rice Cereal. The sample extraction procedure was performed prior to analysis to isolate arsenic species from the matrix. Previous studies have suggested several possible methods to extract common arsenic species from rice and rice products with acceptable confidence. Among these methods, nitric acidassisted extraction has received a considerable amount of attention and was used in HPLC−ICPMS for arsenic speciation by many groups.12,13,15 However, because the speciation in CE−ICPMS was carried out under basic pH conditions, injections of sample extracts from acid digestion negatively affect the results. Heat-assisted aqueous extraction is another effective alternative to fully extract arsenic from rice samples.11 A common problem in this methodology is that large amounts of starch remain in the solution, which cannot be separated by simple centrifugation or filtration. Loss of resolution, peak broadening, and column deterioration because of carbohydrates was observed in previous studies.11,14 Some methodologies have demonstrated that low-temperature extraction with enzymes or the addition of enzymes post-extraction can be effective in removing biological matrices, such as proteins and starch, in rice.14,33,34 To simplify the procedure, we focused our efforts on a direct enzyme-assisted microwave digestion for the extraction process. It is known that thermally stable α-amylase is able to catalyze the hydrolysis of starches and break them down to smaller sugar molecules. The addition of this type of enzyme during the extraction should reduce matrix effects. To compare the extraction efficiency, rice samples were microwaveextracted with and without the addition of α-amylase during preparation. The analysis (see Figure S-1 of the Supporting Information) exhibited significantly higher signal responses for all arsenic species when the enzyme was added during the extraction. The migration of arsenic species was faster for samples extracted with the enzyme, which can also be attributed to the lower viscosity. The extracts were also digested with nitric acid and analyzed using ICPMS for total arsenic concentration to compare the results from the two different 3157

DOI: 10.1021/acs.jafc.5b00446 J. Agric. Food Chem. 2015, 63, 3153−3160

Journal of Agricultural and Food Chemistry

Article

to compensate for signal drift and variance caused by the CE injection. Under a basic buffer condition of moderate concentration with a pH value around 11 and applied voltage of 20 kV, CE−ICPMS analysis can be completed within 10 min with excellent repeatability. The detection limits were determined to be lower than 0.3 ng g−1, with a linear range from 0.5 to 20 ng g−1. The linearity up to 50 ng g−1 was also demonstrated. The enzyme α-amylase was deemed necessary to achieve sufficient digestion and signal response. Rice samples purchased from a local market were successfully analyzed by this method, and the summed quantitative value of each arsenic species was consistent with the total arsenic quantitative concentration determined by standard ICPMS measurements. This methodology was further validated through the analysis of two rice standard reference materials. The quantitative results of DMA, MMA, and inorganic arsenic were well-matched with the certified values.



ASSOCIATED CONTENT

S Supporting Information *

Relative standard deviation (n = 5) of the peak area of arsenic species under different injection conditions (Table S-1), recovery values for As spikes at 2−3 times the highest calibration concentration (Table S-2), and electropherogram of the rice sample prepared by water-phase extraction and enzyme-assisted extraction (Figure S-1). This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Authors

*Telephone: +1-870-543-4665. Fax: +1-870-543-4041. E-mail: [email protected]. *Telephone: +1-870-543-4667. Fax: +1-870-543-4041. E-mail: [email protected]. Funding

This project was supported in part by an appointment to the Research Participation Program at the Office of Regulatory Affairs/Arkansas Regional Laboratory, FDA, administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the FDA. The views expressed in this document are those of the authors and should not be interpreted as the official opinion or policy of the FDA, Department of Health and Human Services, or any other agency or component of the U.S. government. The mention of trades names, commercial products, or organizations is for clarification of the methods

Figure 4. Electropherogram of arsenic standards and rice extracts with post extraction addition of o-AsA.

This study demonstrated a robust and reliable analytical method using CE−ICPMS, for the determination of common arsenic species, including DMA, MMA, As(III) and As(V), in rice and rice cereal by providing better separation efficiency over several liquid chromatography methods. An arseniccontaining compound o-ASA was used as an internal standard

Table 4. Analytical Results of the Arsenic Concentration (ng g−1) in Rice Samples and Rice Standard Reference Materials SRM 1568b and CRM 7503-a samplea SRM 1568b certified values measured values CRM 7503-a certified values measured values rice 1 rice 2

DMA

MMA

As(III)

As(V)

inAs sumb

sumc

totald

180 ± 12 186.49 ± 3.10

11.6 ± 3.5 11.62 ± 0.48

40.92 ± 6.78

92 ± 10 96.32 ± 5.71

284 294.45 ± 6.97

285 ± 14

55.42 ± 1.90

± ± ± ±

97 94.31 ± 1.05 221.63 ± 5.05 215.43 ± 5.59

98 ± 7

13.3 12.94 86.13 76.47

± ± ± ±

0.9 1.15 1.64 3.55