Speciation without Chromatography Using Selective Hydride Generation

iAs forms volatile arsine species with high efficiency when treated with .... Stephen E. Long , Michael Ellisor , Frances Nilsen , Laura Wood , Karen ...
7 downloads 0 Views 346KB Size
Download PDF
Technical Note pubs.acs.org/ac

Speciation without Chromatography Using Selective Hydride Generation: Inorganic Arsenic in Rice and Samples of Marine Origin Stanislav Musil,†,‡,∥ Á sta H. Pétursdóttir,†,§,∥ Andrea Raab,† Helga Gunnlaugsdóttir,§ Eva Krupp,† and Jörg Feldmann*,† †

TESLA-Trace Element Speciation Laboratory, Department of Chemistry, University of Aberdeen, Aberdeen, AB24 3UE, Scotland, U.K. ‡ Institute of Analytical Chemistry of the ASCR, v. v. i., Veveří 97, 602 00 Brno, Czech Republic § Matis, Environment and Genetics Department, Vinlandsleid 12, 113 Reykjavik, Iceland S Supporting Information *

ABSTRACT: Because of the toxicity of inorganic arsenic (iAs), only iAs needs to be monitored in food and feedstuff. This demands the development of easy and quick analytical methods to screen large number of samples. This work focuses on hydride generation (HG) coupled with an ICPMS as an arsenic detector where the HG is added as a selective step to determine iAs in the gaseous phase while organically bound As remains in the solution. iAs forms volatile arsine species with high efficiency when treated with NaBH4 at acidic conditions, whereas most other organoarsenic compounds do not form any or only less volatile arsines. Additionally, using high concentrations of HCl further reduces the production of the less volatile arsines and iAs is almost exclusively formed, therefore enabling to measure iAs without a prior step of species separation using chromatography. Here, we coupled a commercially available HG system to an ICPMS and optimized for determination of iAs in rice and samples of marine origin using different acid concentrations, wet and dry plasma conditions, and different reaction gas modes. Comparing this method to conventional HPLC−ICPMS, no statistical difference in iAs concentration was found and comparable limits of detections were achieved using less than half the instrument time.

T

Among them hydride generation (HG)9 is one of the most straightforward approaches with minimum sample pretreatment. It usually uses sodium borohydride (NaBH4) and hydrochloric acid (HCl) to convert species in aqueous solutions into volatile hydrides.10 It is most often employed to increase sensitivity, eliminate matrix interferences, and/or improve the selectivity by adding an extra step of postcolumn generation.6,11−14 Its direct benefit for speciation of arsenic is the selectivity of the hydride formation when only a few arsenic species form hydrides and at different efficiencies. The individual hydrides were generated either selectively at different conditions in terms of pH, type of acid or buffer, borohydride concentration, or additives15−20 or together with other species in different media. Since the sensitivities of all species in the different media were different, systems of linear equations was used to calculate their concentrations21,22 The motivation for this study is to assess the potential of whether determination of iAs in foodstuff without chromatog-

he toxicity of arsenic depends on its chemical form and varies from the most toxic inorganic arsenic species (iAs = As(III) + As(V)) to organic arsenic species such as innocuous arsenobetaine (AB) and other organic arsenicals that toxicological assessment has declared potentially toxic.1−3 Because of these different toxicities, research over the last decades has focused on differentiating between these arsenic species. The most commonly applied method of speciation today is using HPLC for separation of arsenic species, often hyphenated with atomic absorption spectroscopy (AAS), atomic fluorescence spectroscopy (AFS), or most often in recent years to inductively coupled plasma mass spectrometry (ICPMS), as element specific detector.4−6 There has been increasing interest in robust and reliable methods to determine the iAs concentration in a range of food items with growing awareness of the iAs in food, where EU legislation on iAs content in food items might become a reality in the near future. This need for methods was highlighted in the scientific opinion paper on arsenic in food by the European Food Safety Authority (EFSA, 2009).5 There are many nonchromatographic strategies that are less time-consuming and more cost-effective than high-pressure liquid chromatography (HPLC) for speciation of arsenic.7,8 © 2013 American Chemical Society

Received: October 23, 2013 Accepted: December 19, 2013 Published: December 19, 2013 993

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

Figure 1. (a) Standard setup by Agilent, dry plasma conditions. (b) Modified setup to introduce IS through a nebulized solution in wet plasma conditions. Schematic adapted from Agilent manual by courtesy of Agilent Technologies.

Sample Preparation. For determination of total As concentration, approximately 0.1−0.2 g was microwave digested in 1 mL of concentrated HNO3 and 2 mL of 30% w/w H2O2 using open vessel digestion in a CEM Mars microwave system. The samples were diluted to the final volume of 25 mL with deionized water. Each sample was prepared in triplicate. H2O2 is commonly used during the extraction of iAs and has been observed to be an important factor;25 therefore, two of the extraction methods used H2O2. However, since H2O2 can compromise the HG reaction, the effect of H2O2 was investigated. Microwave extraction was employed for all samples. Rice samples (0.1 g) were extracted in 10 mL of 1% HNO3 and 1% H2O2 (5 min 50 °C, 5 min 75 °C, 30 min 95 °C). The seaweed samples (0.2 g) were extracted with three different extraction solvents (10 mL) abbreviated as E1−E3: E1, H2O; E2, 2% HNO3 + 3% H2O2; E3, 0.07 M HCl + 3% H2O2 (5 min 50 °C, 5 min 75 °C, 10 min 85 °C). For E1, H2O2 was added before analysis to make up 3% H2O2 concentration in the samples in order to oxidize As(III) to As(V). All samples were centrifuged at 13 000 rpm for 10 min prior to analysis with HG-ICPMS or HPLC−ICPMS. During optimization of HG, As(V) water standard was also measured as As(III) after prereduction in 5% KI and 1% ascorbic acid in 4 M HCl (at least 1 h). Before measurement, the prereduced solution was diluted 5 times by water (final: 1% KI, 0.2% ascorbic acid in 0.8 M HCl). The prereduced As(V) will be referred to as As(III) hereafter for simplicity. Hydride Generation Setup. The Agilent Hydride Generation (HG) Accessory for ICPMS was used. The sample was injected via an ASX-500 autosampler and transported to the hydride generator by the peristaltic pump (PP1) in continuous flow mode (CF). The sample was mixed with HCl and NaBH4 in a mixing coil before entering the gas liquid separator (GLS), a cyclonic spray chamber with a nebulizer. The reacting mixture was immediately removed from GLS to the waste by the second peristaltic pump (PP2). The gaseous sample was transported to the ICPMS with an argon gas flow, where a filter prevented droplets from the reaction solution reaching the plasma, Figure 1. Then either an additional dry argon flow was mixed with this gas flow (Figure 1a) or an argon flow carrying a nebulized solution of the IS using the peristaltic pump of the ICPMS (Figure 1b) creating dry and wet plasma conditions respectively. Optimized flow rates and conditions are shown in Table 2. ICPMS Instrumentation. The Agilent triple quadrupole ICPMS 8800 (ICP-QQQ) was used for arsenic detection.

raphy is a feasible option. This work will use the selectivity that HG can offer by using high concentrations of HCl to volatilize almost exclusively iAs leaving all organoarsenic species in solution or forming far less volatile compounds.12,15−18 The aim is to have a method available for quick screening of samples for iAs, tailored for high-throughput of large batches of samples for feed and food quality control. ICPMS as an arsenic detector was chosen since it offers sensitivity suitable for measurements of foodstuff extracts and a broad linear range.

■

EXPERIMENTAL SECTION Chemicals and Standards. Ultrapure water (>18 MΩ cm) was used for all analytical purposes. For calibration of total As and measurements with HG, a 1 002 mg As L−1 certified As stock solution (as H3AsO4 in 0.5 M HNO3) was supplied by Merck (U.K.). Quantification for speciation using HPLC− ICPMS was performed with sodium dimethylarsinic acid (DMA, 98%; ChemService). Rhodium (Specpure, Alfa Aesar, Germany), 1 000 mg L−1 solution, was diluted to 1 or 25 μg L−1 and used as an internal standard for HG measurement or total arsenic/speciation, respectively. Tellurium (BDH, U.K.), 1 000 mg L−1 solution, was diluted to 80 μg L−1 and used as internal standard for HG measurements. Disodium methyl arsonate (MA) was obtained from Chem Service. Nitric acid (HNO3, 69%) was supplied by Fluka (U.K.). Ammonium nitrate (98+ %) was obtained from Sigma-Aldrich (U.K.). Ammonium solution (28%) and ammonium carbonate were obtained from BDH (U.K.). Hydrogen peroxide (H2O2, >30% w/v), ammonium phosphate (NH4H2PO4), sodium hydroxide (NaOH, laboratory reagent grade (LR)), KI, ascorbic acid, and hydrochloric acid (HCl, 32%, LR grade, used for the hydride generation reaction), were obtained from Fisher Scientific (U.K.). Sodium borohydride (NaBH4, 99%) was from Acros Organics (U.K.). Antifoam B emulsion (aqueous− silicone emulsion) was purchased from Sigma-Aldrich. All chemicals used were at least of analytical grade unless otherwise stated. Samples. Two white rice samples used in a proficiency test reported elsewhere23 were analyzed, and rice CRM NIST 1568a is certified for total arsenic concentration and the iAs concentration has been reported extensively in the literature.24 Three edible seaweed samples, dulse (Palmaria palmata), were measured; the samples were collected around Iceland in 2011− 2012. Lobster hepatopancreas CRM TORT-2, certified for total arsenic concentration, was used as a biological matrix reference material. 994

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

Figure 2. Signal stability under (a) dry plasma, no gas and oxygen, (b) wet plasma, no gas and oxygen conditions over time; HG conditions summarized in Table 2.

and introduce an IS to the spray chamber of the ICPMS instrument creating wet plasma conditions, Figure 1b. Since the IS is introduced independently from the HG, the influence of various HG parameters on the IS signal is minimized; therefore, the IS signal monitors only the plasma stability. Both setups were compared, and the signal stability was monitored. The signal in no gas mode at dry plasma conditions was not as stable as for wet plasma conditions (using 5 M HCl for HG) as can be seen in Figure 2 where the individual signal of the blanks was related to the average signal of the blanks normalized to 100%. It is highly probable that the signal increase using dry plasma was a result of 40Ar35Cl interference caused by fumes of HCl evaporating from the reacting solution and slowly being transported to the spray chamber and subsequently to the plasma, Figure 2a. The theory is that the 40Ar35Cl interference should be eliminated by using the reaction/collision cell followed by the second quadrupole, and it was noted that the signal is more stable using oxygen (6% RSD) as a reaction gas than in no gas mode (15%), Figure 2a. When applying wet plasma conditions, the fumes may condense in the spray chamber and be washed out with the liquid waste resulting in far more stable conditions (1.4% RSD in no gas mode and 1.6% RSD in O2 mode) illustrated in Figure 2b. These results indicate that in O2 mode at dry plasma conditions most 40 35 Ar Cl interference are removed; however, considerably more stable conditions are obtained at wet plasma conditions, independent of gas mode. Therefore, wet plasma conditions were used for all further experiments. Optimization of Parameters. Gas Flows. Flow rate for the makeup gas introduced into the GLS was found to be optimal at 0.3 L min−1. The maximum sensitivity of arsenic for the carrier gas was typically achieved in the region between 0.85 and 0.95 L min−1. This was also the optimal flow rate for the highest signal of the nebulized IS. The carrier gas flow rate was manually tuned every day before analysis. HCl Concentration. The influence of HCl on generation of arsine, mono- and dimethylarsine, was investigated for HCl concentrations ranging from 1 to 10.2 M. The main interest was to keep the ratio of DMA/iAs as low as possible with little emphasis on the MA/iAs ratio since MA is generally not present, or only in trace amounts, in most samples including rice and seafood.5,26−28 Further, DMA is almost exclusively the only breakdown product of other more complicated organoarsenic species such as arsenosugars in seaweed.

Measurements were carried out in four gas modes (no gas, He, H2, and O2) in the reaction/collision cell. In O2 mode, arsenic was measured indirectly as 75As16O+ on m/z 91. After optimization of the HG setup, real samples were measured only in no gas mode and O2 mode due to substantial loss of sensitivity in both He and H2 modes. Chromatographic Conditions for HPLC−(HG)-ICPMS Method. Speciation was carried out on an Agilent 1100 HPLC system connected directly, or via postcolumn HG system, to the ICPMS. PRPX-100 Hamilton anion exchange column (10 μm, 4.6 mm × 250 mm) with a flow rate of 1 mL min−1 was used. Mobile phase was 25 mM ammonium carbonate (pH 8.5) for HPLC−ICPMS and a phosphate buffer (6.2 mM ammonium nitrate and 6.5 mM phosphoric acid adjusted to a pH of 6.0 with ammonia) for measurements with HPLC−HGICPMS.13

■

RESULTS AND DISCUSSION The ICP-QQQ was used because it allows for easy monitoring and better control of possible interferences compared to normal single quadrupole instruments. This was thought to be of critical importance since high concentrations of HCl were necessary for selective HG of arsine increasing the chance of chlorine containing gas attributing to the 40Ar35Cl interferences on 75 m/z of As. Since the ICP-QQQ selectively introduces m/ z 75 into the reaction cell, other chloride clusters would be excluded limiting the possibility of interferences. Although this instrumentation is of higher cost than conventional single quadrupole instruments, its use was beneficial for optimization of selective HG, including the optimal way of connecting the generator to the plasma and for verifying the accuracy of the analysis of real samples. Dry versus Wet Plasma. The standard setup suggested by Agilent is shown in Figure 1a. This dry plasma experimental setup uses carrier gas for the HG and makeup gas mixed with the gaseous phase from the GLS. Liquid internal standard (IS) cannot be used under these conditions, and to introduce IS in the plasma the IS must form hydrides together with arsine. Rare hydride forming elements were tested (Ge, In, Te, and Tl); however, only Ge and Te gave a reasonable signal under the given conditions of high concentration of HCl (8 M). Since the change in generation efficiency of IS does not have to correlate with that of arsine introduction of the IS to the plasma independent of the HG process, it was tested during the optimization of HG. The makeup gas was used to transfer the gaseous sample from the GLS and the carrier gas to nebulize 995

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

Figure 3. Influence on the signal of arsenic species dependent on (a) HCl concentration, (b) NaBH4 concentration, standards in water and 3% H2O2; HG conditions as in Table 2 except tested parameters. All results given in no gas mode corrected to 103Rh.

ratio of DMA/iAs at high concentration of HCl where the ratio of DMA/iAs increased with higher concentration of NaBH4. The signal of DMA related to As(V) was 2.7 ± 0.5%, 3.4 ± 0.7%, 4.5 ± 1.0%, and 7.0 ± 0.8% for 0.5, 1, 2 and 3% NaBH4, respectively (using 5 M HCl). Similar dependencies were obtained when HG was performed with 7 M HCl. In order to have low DMA/iAs ratios, using low concentrations of NaBH4 would be optimal; however, since lower concentrations of NaBH4 result in lower sensitivity that must be taken into consideration. The choice of optimal NaBH4 concentration depends on the form in which iAs should be determined. iAs could be measured as As(V) after oxidation by H2O2, often used to facilitate the extraction of iAs or as As(III) after prereduction (information on prereduction in the Supporting Information). Since H2O2 is an oxidant, it influences the NaBH4 decomposition; therefore, the optimization of concentration of NaBH4 was also carried out with the standards prepared in 3% H2O2, as a typical condition for sample extracts (Figure 3b). It can be seen in Table 1 that the H2O2 concentration has an advantageous effect on the DMA/iAs ratio since 4.4% was

The signal of iAs was constant in the whole range of HCl concentration, Figure 3a. However a gradual decrease in MA and DMA signals was observed which corresponds to the data in literature.12,15−19 The species are protonated in strong acidic conditions and thus not prone to hydride formation and remain in solution. The signal of DMA was 4.6 ± 0.3%, 3.9 ± 0.7%, and 3.4 ± 0.4% relative to the signal of iAs for 5, 7, and 8 M HCl, respectively. These signals were not caused by iAs impurity in the DMA standard since the purity of DMA was found by means of HPLC−ICPMS to be >99%. At the highest concentration of HCl, a small increase was found, Figure 3a. This could be due to additional formation of volatile chlorides, AsCl3.29 The same trends were found in He, H2, and O2 modes. The 5 M HCl was chosen for further experiments since it showed low concentrations of DMA and practically it was thought best to keep the HCl concentration as low as possible for safety reasons, the lifetime of tubing, and to keep the cost of consumables at a minimum. Volume of the Reaction Coil. The influence of the volume of the reaction coil/mixer (Figure 1) was investigated since it has been shown to have a significant influence on arsine formation, especially from As(V).15 Four different lengths of PTFE (0.8 mm i.d.) reaction coils were tested (0.23 mL, 0.5 mL, 1 mL, and 1.5 mL). The shortest one was the standard reaction coil supplied by Agilent. The dependence was measured at two concentrations of NaBH4, 1% and 2% (w/ v). Generally 2% NaBH4 gave higher signal for arsenic than 1% NaBH4, Supporting Information, Figure S1. There was a linear increase of the signals of As(V), MA, and DMA and steady signal for As(III) with increased volume of the reaction coil. More important than the overall sensitivity of iAs is the signal ratio of DMA to As(V) where the ratio between DMA and As(V) remained constant independent of the volume of the reaction coil. This was true for both concentrations of sodium borohydride. These results indicate that the sensitivity of As(V) determination can be improved by increasing the reaction coil volume but not the ratio between DMA and iAs, adding to the robustness of the method. For convenience, the standard reaction coil from Agilent (0.23 mL) was therefore chosen for further measurements. Influence of NaBH4 Concentration. Six different concentrations of NaBH4 (0.5 to 3% NaBH4 in 0.1 M NaOH) were tested. The signals for As(V), MA, and DMA increased with increasing concentration of NaBH4, Figure 3b. The concentration of NaBH4 was an important parameter influencing the

Table 1. Contribution of DMA and MA to the Signal of iAs Assessed by Comparison of Sensitivities (0, 0.2, 1, 4, and 15 μg L−1of As in Standards) In No Gas Mode, HG Conditions as in Table 2, Results Corrected for 103Rh MA/iAsa DMA/iAsa

1% H2O2

3% H2O2

4.4 ± 0.03%

42.9 ± 0.4% 2.0 ± 0.04%

a

Ratios expressed as %, errors as combined uncertainty (from slope errors).

found for 1% H2O2 addition whereas only 2.0% for 3% H2O2. For further experiments, a concentration of 2% NaBH4 was chosen. At this concentration, the signals of As(V) standards prepared in 3% H2O2 related to water standards were 86 ± 3%. This highlights the need for calibration standards to be prepared with the same addition of H2O2 as in the sample extracts. Optimal Conditions. The optimized parameters are summarized in Table 2. The excess of HCl is a very important factor for low ratios of DMA/iAs; this is reflected in the high flow rate of HCl compared to the considerably lower flow rate of NaBH4, Table 2. At the optimal conditions, the DMA/iAs ratio was found to range from 2 to 4%, Table 1, depending on 996

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

Table 2. Optimal Parameters of HG Generator: sample flow rate HCl flow rate NaBH4 flow rate reaction coil volume Ar flow rate (for HG) Ar flow rate (for nebulization of IS) Reagents:

0.5 mL min−1 2.5 mL min−1 0.5 mL min−1 0.23 mL 0.3 L min−1 0.85−0.95 L min−1

concentration of HCl concentration of NaBH4

5M 2% (w/v)

H2O2 concentration. Further, it was noted that the DMA/iAs ratio stayed the same over the linear range (0.2−15 μg L−1). The LODs were calculated using 3 × SD of blanks with LOD 7 ng L−1 in no gas mode and 6 ng L−1 in O2 mode. LODs in sample matrix (using an average dilution factor of rice samples) were 1.1 μg kg−1 and 0.9 μg kg−1 in no gas mode and O2 mode, respectively. The LOD values are strongly influenced by iAs contamination from HCl. This blank value has to be subtracted from the standards and sample signals or intercept of the calibration has to be taken into account. The LODs could be improved by using acid of a higher purity but for most of samples these LODs are suitable. Samples. Antifoaming Agent. During HG of samples, substantial foaming could be seen in the GLS, which caused insufficient waste removal. Arsine release from the solution is also influenced and can be retained within the bubbles of the foaming reacting mixture. The foaming extent is unpredictable and depends on the individual sample. Some memory effect and carryover effect were observed since the foam was hard to remove from the GLS. This was remedied with the addition of an antifoaming agent (350 mg L−1, resulting in a 50 mg L−1 concentration in the GLS) to the NaBH4 solution, improving the method performance. No significant changes in the signals of As(V) and DMA were found when the standards or rice extract were measured with and without the addition of Antifoam B emulsion. Quantification. iAs in all extracts was measured as As(V) after oxidation in H2O2; an external calibration was used with matrix matched calibration standards. To validate using external calibration, a recovery test was applied to a seaweed sample by spiking iAs (3 and 6 μg L−1). No significant influence of the matrix was found using HG giving recovery 100.6 ± 2.0% for no gas mode and 95.8 ± 1.5% for O2 mode validating the use of an external calibration for quantification. Rice Samples. Two rice samples and a CRM (NIST 1568a) were analyzed for iAs. Results are presented in Figure 4 for no gas mode and O2 mode. Two different IS were used for correction of the sensitivity drift, 103Rh (ionization potential (IP) 7.46 eV), and 125Te (IP 9.00 eV), Figure 4. Te has an IP closer to that of arsenic, and both IS were included to test whether this difference in IP would have an effect on the quantification. Figure 4 and Table 3 show the robustness of the HG method where there was not found to be a difference between the two modes tested, nor which IS was used, neither was the speciation method a relevant influence, Table 3. No statistical difference was found between using different gas modes and different IS (ANOVA on ranks using Sigmastat) nor was a statistical difference found when using different means of speciation of iAs, HPLC vs HG and additionally verified by combining both

Figure 4. Measured content of iAs in rice samples and CRM (n = 6) in μg kg−1. Error is given as 1 SD. HG conditions as in Table 2

methods and the concentration measured with HPLC-HG. Further, the iAs concentration determined for CRM NIST 1568a is in agreement with values reported in the literature (103 ± 7 mg kg−1 with HG-ICPMS compared to 94 ± 12 mg kg−1)24 as well as for the rice samples used in a previous proficiency testing (95 ± 9 and 107 ± 20 mg kg−1 compared to 107 ± 14 mg kg−1).23 The sum of all species also correlates reasonably with the total amount of arsenic in the samples, Table 3. Seafood Samples. Three dulse samples, SW1−SW3, were measured after three different types of extraction: E1, water extracts; E2, 2% HNO3 + 3% H2O2; E3, 0.07 M HCl + 3% H2O2. Further CRM TORT-2 was measured with E2 and E3. Figure 5 depicts that there is generally good agreement between the concentrations found with HG and HPLC. Using HG-ICPMS does not seem to overestimate the iAs in the seaweed indicating that the As-sugars are not volatilized under these conditions, which might have contributed as As-sugars have been shown to form hydrides in low quantities.30,31 This also gives better confidence that there is no coelution of other organoarsenic species with the iAs when using HPLC−ICPMS given the close agreement between the measurements. Further, it is noted that the water extraction (Figure 5a) gave lower values for all dulse samples than extractions with the dilute acids (Figure 5b), indicating that some iAs might have been bound as thio-arsenic species.32 Quantification of other arsenic species present in the samples is given in Table S1 in the Supporting Information. TORT-2 is not certified for iAs; however, the values found with HG-ICPMS and HPLC−ICPMS agree well with the average of values reported in the literature (0.59 ± 0.01 and 0.57 ± 0.02 with HG, 0.72 ± 0.02 mg kg−1 with HPLC compared to 0.59 ± 0.23 mg kg−1).33 The iAs concentration found with the HPLC−ICPMS was a little higher than found with HG-ICPMS, which could be due to day-to-day differences of instrument and different calibrations. In general the results for both seaweed and TORT-2 agreed for HPLC and HG, in spite of the fact that seafood matrixes are far more complex than the rice matrix.

■

CONCLUSIONS HG-ICPMS was optimized to selectively measure iAs, with only a minor contribution of DMA. MA would contribute significantly to the iAs concentration quantified; however, 997

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

Table 3. iAs in Rice Samples Measured with Different Methods of Speciation in μg kg−1 Presented with SD, HG Conditions as in Table 2, Measured in No Gas Mode Corrected for IS 103Rh HPLC−ICPMS

JRC 203 JRC 165 NIST 1568a

HG-ICPMSa

HPLC−HG-ICPMS

ICPMS

DMA

iAs

iAs

iAs

totAs

(n = 3)

(n = 3)

(n = 6)

(n = 3)

(n = 3)

± ± ± ± ±

95 ± 9 107 ± 20

94 ± 7 95 ± 4

103 ± 7

99 ± 4

57 ± 1 60 ± 1 175 ± 2

105 105 [107 98 [94

4 3 14]a 1 12]b

130 142 [172 258 [290

± ± ± ± ±

12 2 18]a 5 30]c

a Values from proficiency testing.23 bValue from a literature compilation.24 cCertified value. LODs HG-ICPMS, 1.1 μg kg−1; HPLC−HG-ICPMS, 0.6 μg kg−1; HPLC−ICPMS, 0.5 μg kg−1.

Figure 5. Comparison of three extraction methods and HG (no gas mode) vs HPLC (O2 mode) for speciation of iAs (a) water extraction E1, (b) acid extractions E2 and E3, additionally for sample SW2 a measurement with HPLC−HG is included.

■

MA is generally not present in rice and seafood or only in trace amounts. When running HG-ICPMS, the signal stability was much more stable using wet plasma than dry plasma conditions. Further no difference was found between O2 mode and no gas mode, indicating no significant chloride interferences when using wet plasma conditions, despite high concentrations of HCl used for HG. Therefore a single quadrupole instrument can be used as an arsenic detector with no risk of compromising the results by 40Ar35Cl interference. The combination of HG-ICPMS has been shown to give comparable results to HPLC−ICPMS for the determination of iAs when applied to rice and seaweed samples. This method can be used as a quick screening method for large sample sets as each sample takes just over 4 min, measuring 5 replicates, in comparison speciation with HPLC most commonly takes about 10 min for each sample giving only one replica. Further, when measuring samples of seafood origin, the arsenic profile is more complex and therefore to prevent coelution of the iAs with other arsenic species longer HPLC run times are often needed. In addition using HG allows for easier data treatment as no integration of sample peaks is needed. Using HG-ICPMS is a realistic and simple alternative to the conventional HPLC− ICPMS approach for large scale screening of relevant samples, e.g., where a limit of iAs will be set for rice.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Author Contributions ∥

S.M. and Á .H.P. contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Agilent Technology for the loan of the 8800 ICPMS. Á . Pétursdóttir thanks the Icelandic research fund and the Icelandic research fund for graduate students for the financial support. S. Musil is grateful to Institute of Analytical Chemistry AS CR, v. v. i. (Institutional Research Plan RVO:68081715), to the Academy of Sciences of the Czech Republic (Project No. M200311271), and Journal Grant for International Authors programme by the Royal Society of Chemistry that sponsored his visit in the TESLA lab. Finnbogi Gudmundson is acknowledged for the dulse samples, and Cornelius Brombach is kindly thanked for the table of content/ abstract graphic.

■

REFERENCES

(1) Feldmann, J.; Krupp, E. M. Anal. Bioanal. Chem. 2011, 399, 1735−1741. (2) Edmonds, J. S.; Francesconi, K. A. Mar. Pollut. Bull. 1993, 26, 665−674. (3) ATSDR. Toxicological Profile for Arsenic; United States Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry: Atlanta, GA, 2007. (4) Francesconi, K. A.; Kuehnelt, D. Analyst 2004, 129, 373−395. (5) European Food Safety Authority. EFSA J. 2009, 7 (10), 1351.

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. 998

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999

Analytical Chemistry

Technical Note

(6) Maher, W.; Krikowa, F.; Ellwood, M.; Foster, S.; Jagtap, R.; Raber, G. Microchem J. 2012, 105, 15−31. (7) Rasmussen, R. R.; Qian, Y. T.; Sloth, J. J. Anal. Bioanal. Chem. 2013, 405, 7851−7857. (8) Vieira, M. A.; Grinberg, P.; Bobeda, C. R. R.; Reyes, M. N. M.; Campos, R. C. Spectrochim. Acta, Part B: At. Spectrosc. 2009, 64, 459− 476. (9) J. Dědina and Tsalev, D. L. Hydride Generation Atomic Absorption Spectrometry; John Wiley & Sons, Inc.: Chichester, U.K., 1995. (10) Howard, A. G. J. Anal. At. Spectrom. 1997, 12, 267−272. (11) Hwang, C. J.; Jiang, S. J. Anal. Chim. Acta 1994, 289, 205−213. (12) Magnuson, M. L.; Creed, J. T.; Brockhoff, C. A. J. Anal. At. Spectrom. 1996, 11, 893−898. (13) Pétursdóttir, Á .; Gunnlaugsdóttir, H.; Jö rundsdóttir, H.; Mestrot, A.; Krupp, E.; Feldmann, J. Anal. Bioanal. Chem. 2012, 1−7. (14) Nakazato, T.; Tao, H. Anal. Chem. 2006, 78, 1665−1672. (15) Anderson, R. K.; Thompson, M.; Culbard, E. Analyst 1986, 111, 1143−1152. (16) Shraim, A.; Chiswell, B.; Olszowy, H. Talanta 1999, 50, 1109− 1127. (17) Tasev, K.; Karadjova, I.; Stafilov, T. Microchimica Acta 2005, 149, 55−60. (18) Carrero, P.; Malave, A.; Burguera, J. L.; Burguera, M.; Rondon, C. Anal. Chim. Acta 2001, 438, 195−204. (19) Karadjova, I. B.; Lampugnani, L.; Onor, M.; D’Ulivo, A.; Tsalev, D. L. Spectrochim. Acta, Part B: At. Spectrosc. 2005, 60, 816−823. (20) Pitzalis, E.; Ajala, D.; Onor, M.; Zamboni, R.; D’Ulivo, A. Anal. Chem. 2007, 79, 6324−6333. (21) Cava-Montesinos, P.; Nilles, K.; Cervera, M. L.; de la Guardia, M. Talanta 2005, 66, 895−901. (22) Reyes, M. N. M.; Cervera, M. L.; Campos, R. C.; de la Guardia, M. Spectrochim. Acta, Part B: At. Spectrosc. 2007, 62, 1078−1082. (23) de la Calle, M. B.; Emteborg, H.; Linsinger, T. P. J.; Montoro, R.; Sloth, J. J.; Rubio, R.; Baxter, M. J.; Feldmann, J.; Vermaercke, P.; Raber, G. TrAC, Trends Anal. Chem. 2011, 30, 641−651. (24) Tyson, J. ISRN Anal. Chem. 2013, 2013, 24. (25) de la Calle, M. B.; Baer, I.; Robouch, P.; Cordeiro, F.; Emteborg, H.; Baxter, M. J.; Brereton, N.; Raber, G.; Velez, D.; Devesa, V.; Rubio, R.; Llorente-Mirandes, T.; Raab, A.; Feldmann, J.; Sloth, J. J.; Rasmussen, R. R.; D’Amato, M.; Cubadda, F. Anal. Bioanal. Chem. 2012, 404, 2475−2488. (26) Maher, W.; Foster, S.; Krikowa, F.; Donner, E.; Lombi, E. Environ. Sci. Technol. 2013, 47, 5821−5827. (27) Meharg, A. A.; Williams, P. N.; Adomako, E.; Lawgali, Y. Y.; Deacon, C.; Villada, A.; Cambell, R. C. J.; Sun, G.; Zhu, Y. G.; Feldmann, J.; Raab, A.; Zhao, F. J.; Islam, R.; Hossain, S.; Yanai, J. Environ. Sci. Technol. 2009, 43, 1612−1617. (28) Ruttens, A.; Blanpain, A. C.; De Temmerman, L.; Waegeneers, N. J. Geochem. Explor. 2012, 121, 55−61. (29) Mester, Z.; Sturgeon, R. E. J. Anal. At. Spectrom. 2001, 16, 470− 474. (30) Regmi, R.; Milne, B. F.; Feldmann, J. Anal. Bioanal. Chem. 2007, 388, 775−782. (31) Schmeisser, E.; Goessler, W.; Kienzl, N.; Francesconi, K. A. Anal. Chem. 2004, 76, 418−423. (32) Larsen, E. H.; Engman, J.; Sloth, J. J.; Hansen, M.; Jorhem, L. Anal. Bioanal. Chem. 2005, 381, 339−346. (33) Pétursdóttir, Á . H.; Gunnlaugsdóttir, H.; Krupp, E. M.; Feldmann, J. Food Chem. 2014, 150, 353−359.

999

dx.doi.org/10.1021/ac403438c | Anal. Chem. 2014, 86, 993−999