DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO.

3, 2011 947

RESIDUES AND TRACE ELEMENTS

Validation of a Method for Arsenic Speciation in Food by Ion

Chromatography-Inductively Coupled Plasma/Mass
Spectrometry After Ultrasonic-Assisted Enzymatic Extraction
VINCENT DUFAILLY, MARINE NICOLAS, JANIQUE RICHOZ-PAYOT, and ERIC POITEVIN1
Nestlé Research Center, Vers-Chez-Les-Blanc, CH-1000 Lausanne 26, Switzerland

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

A fully validated and rapid quantitative method is food (3, 4), followed by methylated compounds that are
presented for determination of inorganic arsenic significantly less toxic, such as methylarsonic acid (MA) and
[arsenite, As(III) and arsenate, As(V)] and organic cacodylic acid (dimethyarsinic acid, DMA; 5). Finally,
arsenic species (methylarsonic acid, dimethylarsinic organic species, such as arsenobetaine (AB), arsenocholine
acid, and arsenobetaine) by ion chromatography (AC), and arsenosugars are considered to be essentially
paired with inductively coupled plasma/MS after nontoxic (6). Chronic exposure to AsI may give rise to several
ultrasonic-assisted enzymatic extraction (UAEE) in health effects, including effects on the gastrointestinal tract,
rice- and seafood-based raw materials and finished respiratory tract, skin, liver, cardiovascular system,
products. This method gives toxicological meaning hematopoietic system, and nervous system (1). The World
to arsenic analysis, since the sum of the toxic Health Organization (WHO) has set a guideline of 10 mg/L as
chemical forms As(III) and As(V) can be determined. the drinking water standard (6).
In contrast to classical water–methanol extraction, Many methods have been developed in the last decades to
UAEE enables drastic acceleration of sample extract, separate, and quantify these As species in various
extraction (5 min instead of several hours), while matrixes, including food samples (7). Nevertheless, most food
total arsenic extraction efficiency is improved legislations and official methods [atomic absorption
without species conversion. Validation was spectrometry (8–11) or inductively coupled plasma
performed to evaluate the method for selectivity, (ICP)/MS; 12] focus on total As value, whereas only the AsI
linearity, LOD/LOQ (0.007–0.020 mg/kg), trueness, amount has a toxicological meaning.
precision (HorRat values, 0.2–0.6), recovery European Union and United States food regulations on AsI
(93–122%), and uncertainty. The method was also are nonexistent for the moment, but the European Food Safety
satisfactorily tested using two proficiency tests. Authority (EFSA) pointed out the need to produce speciation
Performance characteristics are reported for four data for different food commodities to support dietary
certified reference materials, standard reference exposure assessment and dose-response data for the possible
material (SRM) 1568a (rice flour), Institute for health effects (13). AOAC INTERNATIONAL, Chemical
Reference Materials and Measurements 804 (rice Contaminants and Residues in Food Community, Metals
flour), SRM 2976 (mussel tissue), certified Subgroup, listed in the “2008 call for method needs” (14) an
reference material-627 (tuna fish), and several As speciation method to determine the level of the toxic
commercial food samples populating five AOAC chemical forms of As [As(III), As(V), MA, and DMA] in food
triangle food sectors. The results indicated that (primarily seafood and dietary supplements).
this speciation method is cost-efficient, Considering total As amount, seafood and rice are two of
time-saving, and accurate, as well as the most important sources of As in food (15–17). However,
fit-for-purpose, according to International seafood has the ability to bioaccumulate the nontoxic AB
Organization for Standardization/International form, whereas rice has a tendency to accumulate toxic
Electrotechnical Commission 17025:2005 standard, AsI (17–19). For example, Meharg et al. (15) showed that
and could be used for routine analysis. 35% of the United Kingdom baby rice samples analyzed had
an amount of AsI higher than the Chinese regulatory limit
(0.15 mg/kg). Thus, this information pointed out the fact that
t is well known that the toxicity of arsenic (As) depends on the food industry needs suitable and robust methods for

I its chemical forms (1, 2). Indeed, inorganic As (AsI)

species, arsenite As(III) and arsenate As(V), are
considered the most toxic of the common forms found in
determination of AsI in food (mainly rice- and seafood-based
raw materials and finished products) in order to be ready to
answer existing and forthcoming legislations. The first and
most important step for a speciation method is to ensure an
Received August 12, 2010. Accepted by AK November 24, 2010. efficient method of extraction of the analyte without species
1
Corresponding author’s e-mail: [email protected] conversion. Many methods of extraction were reported and
948 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

Table 1. ICP/MS instrumental settings and column allows a satisfactory separation of both inorganic and
chromatographic conditions organic As species in a single chromatographic run. It has
ICP/MS Elan DRC II (PerkinElmer) been applied to matrixes such as water (30), mushrooms (23),
plants (24), and fish products (23, 25–29, 31).
Several years ago, our laboratory developed a method for
RF power 1000–1600 W
As speciation by ion chromatography (IC)-ICP/MS after
Nebulizer gas flow 0.8–1.0 L/min
water–methanol extraction (32), but this method was time-
Autolens ON, calibrated consuming and gave low recovery for rice samples. The
Detector mode Pulse or dual purpose of this work was to develop a fast and efficient
Dwell time on each mass 500 ms method in order to provide a suitable, sensitive, and
Sweeps/reading 656 quantitative method for routine As speciation in seafood- and
75 rice-based raw materials and finished products, and to validate
Analytical mass As, 77ArCl, 115In
this method in order to answer the needs of official food

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Total acquisition time 6.0 min
organizations and existing and forthcoming legislations. The
Sampling 1.9 points/s validation of this method in accordance with Standard
HPLC Series 200 (PerkinElmer) Procedure International Organization for Standardization/
International Electrotechnical Commission (ISO/IEC)
No. of injections 1/sample
17025:2005 (33) involved single-laboratory validation
Injection volume 100 mL (SLV); verification of the SLV by a second laboratory (i.e.,
Flush Milli-Q water checking that the target performance characteristics,
Flush volume 2500 mL previously established during the SLV, were reached); and
Guard column Dionex IonPac AG 7 two external QCs by participating in proficiency tests.
(10 mm, 4 ´ 50 mm)
Experimental
Column Dionex IonPac AS 7
(10 mm, 4 ´ 250 mm) Apparatus
Flow rate 1.3–1.5 mL/min
Mobile phase A 0.7 mM acetate buffer solution
(a) Ultrasonic device.—Sonics Vibracell (ACIL SARL,
Chatou, France) ultrasonic processor (130 W, 20 kHz)
Mobile phase B 0.5 mM HNO3
equipped with a 3 mm titanium probe.
Mobile phase C 50 mM HNO3 (b) Digestion system.—HPA-S high-pressure Asher
Gradient program Before injection: 4 min system (PerkinElmer, Waltham, MA), maximum pressure
(1.3 mL/min; 100% A) 150 bar, maximum temperature 320°C, used with 15 mL
After injection: 1 min quartz vessels and 21-position heating blocks.
(1.3 mL/min; 100% A)
(c) Centrifuge.—Heraeus Minifuge T (Thermo Fisher
1.5 min (1.3 mL/min; 75% B; 25% C) Scientific, Waltham, MA) centrifuge equipped with suitable
3 min (1.5 mL/min; 100% C) rotor.
(d) Centrifuge filters.—Amicon Ultra-15 (Millipore,
Billerica, MA).
(e) Chromatographic system.—Series 200 (PerkinElmer)
compared in the literature (7, 20). Generally, these methods
HPLC system equipped with a Series 200 quaternary pump,
are time-consuming and show poor recovery for As in rice.
autosampler, vacuum degasser, injection valve with a 100 mL
Ultrasonic-assisted enzymatic extraction (UAEE) showed
polyether ether ketone (PEEK) sample loop, and data
efficient extraction of As species in several different food
software (Chromera, PerkinElmer). An IonPac AG7 guard
samples compared to classical solid–liquid extraction (21, 22). column (50 ´ 4 mm, 10 mm particles) and an IonPac AS7
By focusing the energy on a localized sample zone anion exchange column (250 ´ 4 mm, 10 mm particles; both
ultrasonication induces extreme temperatures and pressures from Dionex, Sunnyvale, CA) were used for the
generated in the liquid at a micro level, as well as formation of chromatographic separation of compounds. The analytical
highly reactive radical species (21). All these physical and column was connected to the nebulizer of the ICP/MS
chemical effects enhance the reactivity of targeted enzymes instrument using 50 cm long PEEK tubing (0.17 mm id). The
and accelerate drastically extraction of As species in the operating conditions are summarized in Table 1.
sample (i.e., a few minutes instead of hours for classical (f) Spectrometer.—Elan DRC II (PerkinElmer) ICP/MS
extractions). instrument equipped with mini cyclonic spray chamber
Most speciation methods mentioned in the literature (Cinabar) fitted with a micromist nebulizer, 0.2 mL/min
coupled LC with ICP/MS to separate and quantify As (Glass Expansion, Melbourne, Australia). For total As
species (7). Several studies based on the work of analysis, the sample solutions were pumped by a peristaltic
Londesborough et al. (23) used a Dionex IonPac AS7 pump from tubes arranged on an ASX 500 Model 510
anion-exchange column with a nitric acid eluent (23–31). This autosampler (CETAC, Omaha, NE) and aspirated into the
 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011 949

Table 2. Samples used for optimization of extraction method, SLV, and verification

Material type Sample reference AOAC food triangle sector Origin Test codea

Polished rice Polished rice A 5 Italy a

Polished rice Polished rice B 5 Italy a
Polished rice Polished rice C 5 Italy a
Polished rice Polished rice D 5 Italy a
Rice-based baby food Baby food A 5 Poland a
Rice-based baby food Baby food B 5 Poland a
Rice-based baby food Baby food C 5 France a
Rice-based baby food Baby food D 5 Belgium a

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Rice flour SRM 1568a 5 NISTb, U.S. a, b, and c
Mussel tissue SRM 2976 9 NIST, U.S. a and c
Nondefatted lobster hepatopancreas LUTS-1 4 NRCCb, Canada a
Mussel GBW 08571 9 NRCC, Canada a
Milk powder T0766 6 FAPASb, UK a
Soya flour T0756 5 FAPAS, UK a
Seafood- and rice-based baby food — 5 Commercial product, b
Switzerland
Tuna fish CRM 627 9 BCRb, Belgium b and c
Cereal-and fish-based baby food — 5 Commercial product, c
Singapore
Seaweed flakes — 8 Commercial product, c
Singapore
Rice flour IRMM 804 5 JRCb, Belgium c

a
a = Optimization of extraction method, b = SLV, and c = verification.
b
NIST = National Institute of Standards and Technology; NRCC = National Research Council Canada; FAPAS = Food Analysis Performance
Assessment Scheme; and JRC = Joint Research Center.

argon plasma. To avoid introduction of the internal standard (h) Sodium arsenate dibasic heptahydrate.—Purity
in the chromatographic column, online automatic addition ³99.995% (Sigma-Aldrich).
[two-way in, one-way out/T-piece Socochim (Lausanne, (i) Disodium methyl arsenate.—Purity ³99.0%
Switzerland), polypropylene, 1.5 mm id] of the internal (Sigma-Aldrich).
standard was used. Further details of the instrumental settings (j) DMA.—Purity ³99.0% (Sigma-Aldrich).
are given in Table 1. (k) AB standard solution, BCR 626.—Purity = 97%
(Sigma-Aldrich).
Chemicals and Reference Solutions
(l) AB.—Purity ³95% (Sigma-Aldrich).
(a) High grade water, H2O (18.2 MW).—Milli-Q™ PLUS (m) Indium standard.—1000 mg/mL (VWR).
system (Millipore). (n) Protease from Streptomyces griseus type XIV.—
(b) Nitric acid (HNO3).—65% (w/v), Suprapure (Merck, ³3.5 units/mg solid (Sigma-Aldrich).
Darmstadt, Germany). (o) a-Amylase from Bacillus subtilis.—Approximately
(c) Nitric acid (HNO3).—65% (w/v), analytical grade 380 units/mg (Sigma-Aldrich).
(Merck). (p) Chloride standard solution.—1000 mg/mL, NaCl in
(d) Multistandard solution.—CPI International (Santa H2O, solution ready for use (Merck).
Rosa, CA). Composition in mg/L: aluminium = 5; As = 5;
Purification of Mixed Enzyme Solution
cadmium = 1; lead = 2; selenium = 1; chromium = 1;
molybdenum = 2; cobalt = 2; and tin = 10. Enzymes (protease type XIV and a-amylase) used for the
(e) Acetate buffer solution.—pH 4.65, sodium extraction of As species contain a non-negligible amount
acetate–acetic acid solution, ready for use (Sigma-Aldrich, St. of metals (e.g., up to 0.2% for As(V); 22, 34). The mixture
Louis, MO). of enzymes needs to be purified before use. Quantities of
(f) 2-Isopropanol.—p.a. (VWR, West Chester, PA). 1200 mg of protease type XIV and 360 mg a-amylase were
(g) Sodium (meta)arsenite.—Purity ³99.0% (Sigma-Aldrich). dissolved in 15 mL Milli-Q water. A volume of 2.5 mL
950 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Figure 1. Chromatogram observed for 1 mg/L (as As) standard solution of each As species using the optimized
chromatographic conditions (Table 1).

nonpurified mixed enzyme solution was pipetted on the top (f) Mobile phase C.—50 mM HNO3 (analytical grade).
part of each of six centrifuge filters (Amicon Ultra-15). The (g) Internal standard solution.—50 mg/L indium solution
filters were centrifuged at 6500 rpm (4000 ´ g) for 1 h. The from 1000 mg/mL indium standard solution was prepared in
residue of each filter was dissolved in 1 mL ultrapure water 4% of isopropanol. Isopropanol was used to reduce the carbon
and centrifuged again at 6500 rpm (4000 ´ g) for 1 h. This effect on As sensitivity (35).
procedure was repeated twice. Residues of the filters were
dissolved in 120 mL Milli-Q water and stored in a plastic Samples
bottle for 1 week at room temperature. The efficiency of the The choice of matrixes was based on the AOAC food
purification was checked by analyzing at m/z 75 a blank triangle approach, with the aim to cover five sectors (4, 5, 6, 8,
sample containing 30% purified enzymatic solution (same and 9) of the scope of application of the method (rice and
proportion as in an extracted sample). seafood) according to their relative proportion of As, their
relative proportion of fat, protein, and carbohydrate
Preparation of Reagents and Reference Solutions
content (36), and their total amount of As. All samples used
All solutions were prepared in plastic volumetric flasks are listed in Table 2.
with analytical reagent grade chemicals and ultrapure water
UAEE Procedure
(18.2 MW) generated by purifying distilled water with a
Milli-QTM PLUS system. The procedure described was based on studies of Sanz et
(a) 1000 mg/L (as As) standard stock solutions of As(III), al. (21, 22) and optimized according to the samples tested and
As(V), MA, DMA, and 200 mg/L (as As) of AB.—Prepared chromatographic conditions used.
from sodium (meta)arsenite, sodium arsenate dibasic One randomly selected vessel was filled with the reagents
heptahydrate, disodium methyl arsenate, DMA, and AB, and taken through the entire procedure as a reagent blank.
respectively. AB standard solution BCR 626 was used About 200 mg dry sample (or 400 mg wet sample) was
initially, but it was replaced by the AB salt (purity ³95%) after weighed in a 15 mL polypropylene calibrated tube (Falcon,
it was not available for purchase anymore. The stability of Becton Dickinson, Franklin Lakes, NJ). A 3 mL volume of
these standards in terms of total As content and purity of purified enzymatic solution was added. An ultrasonic probe
the species was checked by ICP/MS and IC-ICP/MS, was immersed in the middle of the solution. The mixture was
respectively. Standard stock solutions were prepared in 1% sonicated at a power of 60% (about 80 W) for 5 min. The
HNO3 (Suprapure) to facilitate dissolution of As salts and probe tip was rinsed with Milli-Q water to recover deposited
stabilization of As species. Stock solutions were stored at 4°C sample remaining on the probe. The solution was centrifuged
in the dark. for 10 min at 3500 rpm (1000 ´ g), and the supernatant was
(b) Calibration standards.—Each stock solution was filtered through a 0.22 mm syringe filter into a new 15 mL
further diluted to 1 and 0.1 mg/L (as As), out of which Falcon tube (Tube 1). The solution was diluted to the 8 mL
multicompounds solution were prepared daily. mark of the Falcon tube with Milli-Q water. The accuracy of
(c) 0.2, 0.5, 1, 3, 5, and 10 mg/L (as As) working standard the 8 mL mark of the Falcon tube had been checked before. A
solutions of As(III), As(V), MA, DMA, and AB.—Prepared volume of 4 mL diluted solution was pipetted into a new
from 1 and 0.1 mg/L (as As) diluted stock solutions. Each 15 mL Falcon tube (Tube 2).
solution was prepared in 5% isopropanol and 30% purified In Tube 1, 750 mL Milli-Q water and 250 mL isopropanol
mixed enzyme solution. were added. Isopropanol was added to reduce the activity of
(d) Mobile phase A.—0.7 mM acetate buffer solution enzymes in samples and to buffer the carbon effect on As
pH 4.65, sodium acetate–acetic acid. sensitivity. Tube 1 was stored at 4°C in the dark until As
(e) Mobile phase B.—0.5 mM HNO3 (analytical grade). speciation analysis by IC-ICP/MS. In case of additional
 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011 951

Table 3. Total As recoveries observed by ICP/MS after optimized UAEE of several samples

Matrix Reference value, mg/kga Observed value, mg/kgb Extraction rate, %

Polished rice A 0.082a 0.078 95

a
Polished rice B 0.093 0.092 98
Polished rice C 0.136a 0.136 100
Polished rice D 0.113a 0.116 102
Rice flour (SRM 1568a) 0.290 0.280 97
a
Baby food A 0.101 0.091 90
Baby food B 0.249a 0.230 92
a
Baby food C 0.170 0.166 98

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Baby food D 0.069a 0.078 114
Tuna fish (CRM 627) 4.80 4.03 84
Mussel tissue (SRM 2976) 13.3 12.0 90
Mussel (GBW 08571) 6.10 5.01 82
Lobster (LUTS-1) 2.83 3.23 114
Milk powder (FAPAS T0766) 0.056 0.051 91
Soya flour (FAPAS T0756) 0.349 0.382 109

a
Reference values presented in the table were obtained by ICP/MS after HPA digestion (refs. 35, 37).
b
Total extracted As obtained by ICP/MS after UAEE.

dilution of sample, the same proportion of enzymes and enzymatic solution or a real sample induced the split of
isopropanol was maintained in the final solution. MA (25, 26, 31). Even if the MA1 area represented less than
In Tube 2, 750 mL Milli-Q water and 250 mL nitric acid 10% of total MA, the sum of the two MA peak areas was taken
(analytical grade) were added to precipitate enzymes. The into account for calculation. A 6-point external calibration
solution was centrifuged for 10 min at 3500 rpm (1000 ´ g), [0.2–10 mg/L (as As)] was carried out with the respective
and the supernatant filtered through a 0.22 mm syringe filter standard compounds that were prepared in a reagent blank
into a new 15 mL Falcon tube (Tube 3). Tube 3 was stored at (purified mixed enzyme solution and isopropanol). The peak
4°C in the dark until total As analysis by ICP/MS. identification and quantification (peak area at m/z = 75) were
performed using Chromera software.
Digestion Procedure for Total As Determination
Total As Determination Procedure
Prior to use, quartz vessels were decontaminated and then
rinsed with Milli-Q water and dried in an oven. One randomly The total As concentration in extracted and digested
selected vessel was filled with the reagents and taken through samples was determined by accredited ICP/MS methods
the entire procedure as a reagent blank. Samples (0.25 to 1 g described previously (35, 37).
wet mass) were weighed into closed quartz digestion vessels
SLV
and wet digested with 2 mL nitric acid (65%). The sealed
containers were placed in a high-pressure asher and heated (a) Linearity.—Linearity of the method was evaluated by
according to the digestion program described previously analyzing six working standard solutions [0.2, 0.5, 1, 3, 5, and
(35, 37). After digestion, sample solutions were cooled to 10 mg/L (as As)] in triplicate for each As species [As(III),
room temperature, then transferred quantitatively into a As(V), MA, DMA, and AB]. These concentration levels
15 mL Falcon tube and diluted to 10 mL with Milli-Q water. covered the working range of the method for all As species
In case of additional sample dilution, the same proportion of studied.
nitric acid was maintained in the final solution. (b) LOD and LOQ.—LOD and LOQ were estimated by
the S/N procedure. A blank sample (3 mL purified mixed
As Speciation Determination Procedure
enzyme solution and 0.5 mL isopropanol diluted to 10 mL
As species in extracted samples were analyzed by with Milli-Q water) was spiked at three different low
IC-ICP/MS. An acetate buffer–HNO3 gradient was used to concentration levels (0.1, 0.2, and 0.3 mg/L As) of each
separate compounds in a single chromatographic run. Table 1 species studied [As(III), As(V), MA, DMA, and AB], and
details the chromatographic conditions of separation, and analyzed in triplicate. The S/N for each replicate of the spiked
Figure 1 shows the chromatogram observed in these samples was calculated. The slope and intercept of the
conditions. MA divided in two different peaks, at 1.8 (MA1) regression line of the S/N versus spiked concentration were
and 3.7 (MA2) min. Indeed, the presence of salts in an determined for each species. LOD was calculated as:
952 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

Table 4. Results of the evaluation of linearity for each As species

Analyte Concn range, mg/L Slope Intercept Intercept = 0? (Y/N) R2 SD of residuals

As(III) 0.2–10 14656 642 N 1.000 484

As(V) 0.2–10 11580 3258 N 1.000 891
DMA 0.2–10 12541 –20 Y 0.999 1473
MA 0.2–10 10655 153 Y 1.000 310
AB 0.2–10 10545 106 Y 1.000 737

3 - Intercept fraction) of the analyte in question (valid for concentration

LOD =

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Slope ratio above 10–9). The HorRat value is a simple performance
parameter that reflects the acceptability of a chemical method
and LOQ as: of analysis with respect to precision (38). A HorRat value of
1.0 is judged satisfactory with limits of acceptability of
10 - Intercept 0.3–1.5. Consistent deviations from the ratio on the low side
LOQ =
Slope (values <0.3) may indicate unreported averaging or excellent
training and experience. Consistent deviations on the high
(c) Precision.—Three matrixes, National Institute of side (values >1.5) may indicate inhomogeneity of the test
Standards and Technology (NIST) 1568a (rice flour), a rice- samples, need for further method optimization or training,
and seafood-based baby food, and certified reference material operating below the LOQ, or an unsatisfactory method.
(CRM) 627 (tuna fish), were spiked at two concentration It is now one of the acceptability criteria for many of
levels (1 and 5 mg/L As) for each species [As(III), As(V), MA, the recently adopted chemical methods of analysis of AOAC
DMA, and AB]. Each spiked sample was analyzed in INTERNATIONAL, the European Union, and other
duplicate (k = 2) on 6 different days (n = 6) by three different European organizations dealing with food analysis (38).
analysts. Nonspiked sample was also considered when the (d) Trueness and recovery.—The same spiked matrixes
native content was higher than the LOQ. Total As extracted in analyzed to evaluate precision of the method were also used
nonspiked samples was also quantified by ICP/MS for each for evaluation of the trueness. Depending on the sample used,
matrix. Both repeatability SD (SDr) and intermediate the recovery and associated RSD were calculated as follows:
reproducibility SD (SDiR) were determined.
The Horwitz ratio (HorRat value) was calculated for SLV ctotal × measured - cnative × measured
Spiked sample: Recovery
according to the equation: cspike

CV(iR), % ctotal × measured

HorRat Value = CRM: Recovery =
2 ´ C-0.1505 , % creference

where CV(iR) is the relative intermediate reproducibility SD The CODEX procedural manual, 19th Ed., set acceptance
of each species in each of the three tested samples and C is the criteria for recovery according to concentration range:
relative concentration (expressed as a dimensionless mass 60–115% for 0.010 and 80–110% for 0.1–10 mg/kg (39).

Table 5. Estimated LOD and LOQ for dry sample and wet sample for each species

Estimated LOD Estimated LOQ

Dry sample, Wet sample, Dry sample, Wet sample,

Analyte Intercept Slope mg/La mg/kga mg/kga mg/La mg/kga mg/kga

As(III) 0.026 43.4 0.068 0.003 0.002 0.230 0.011 0.006

As(V) –7.31 93.5 0.110 0.005 0.003 0.185 0.009 0.005
AsIb 0.178 0.008 0.005 0.415 0.020 0.011
DMA 1.59 24.4 0.057 0.003 0.001 0.344 0.017 0.009
MA 0.917 32.9 0.063 0.003 0.002 0.276 0.014 0.007
AB 2.412 22.2 0.026 0.001 0.001 0.341 0.017 0.009

a
LOD and LOQ were expressed to three decimal places.
b
LOD and LOQ of AsI were calculated as the sum of LOD and LOQ of As(III) and As(V).
 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011 953

Table 6. Results of the evaluation of precision for each As species

SRM 1568a (rice flour) Baby food CRM 627 (tuna fish)

Meanb, CV(iR), Meanb, CV(iR), Meanb, CV(iR),

Analyte Spikea mg/kgc CV(r), % % HorRatd mg/kgc CV(r), % % HorRatd mg/kgc CV(r), % % HorRatd

AsI a 0.086 6.6 6.7 0.3 0.010 7.3 20.0 0.6 — — — —

b 0.186 2.9 9.9 0.5 0.058 3.3 10.9 0.4 0.094 8.3 11.8 0.5
c 0.580 2.2 4.7 0.3 0.248 7.8 8.3 0.4 0.481 1.2 7.3 0.4
DMA a 0.131 4.2 13.4 0.6 — — — — — — — —
d 0.184 4.1 7.0 0.3 0.025 3.3 6.2 0.2 0.062 4.5 5.2 0.2
e 0.374 5.0 5.9 0.3 0.128 4.0 4.1 0.2 0.250 1.8 6.1 0.3

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

MA d 0.072 6.0 7.4 0.3 0.028 4.3 5.4 0.2 0.059 1.1 5.7 0.2
e 0.285 6.0 11.4 0.6 0.139 6.4 11.3 0.5 0.257 4.5 5.7 0.3
AB a — — — — 0.020 8.3 7.8 0.3 0.365e 4.7 6.1 0.3
e
d 0.052 5.0 12.7 0.5 0.046 5.0 10.0 0.4 0.422 4.1 3.5 0.2
e 0.253 3.4 6.7 0.3 0.149 3.0 3.7 0.2 0.628e 2.7 7.1 0.4
f f f
Total As a 0.274 9.6 9.1 0.5 0.044 4.8 9.5 0.4 3.46 4.9 4.9 0.4

a
a = Nonspiked (only considered when native content was higher than LOQ), b = spiked at 2 mg/L [sum of 1 mg/L additions of As(III) and
As(V)], c = spiked at 10 mg/L [sum of 5 mg/L additions of As(III) and As(V)], d = spiked at 1 mg/L, and e = spiked at 5 mg/L.
b
Mean observed for duplicate (k = 2) on 6 different days (n = 6) by three different analysts.
c
Volume of dilution = 10 mL, sample weight = 0.2 or 0.4 g for dry or wet sample, respectively.
d
HorRat = Ratio using CV(iR) and Horwitz denominator.
e
Values observed after 10 times dilution of test sample solution.
f
Total extracted As obtained by ICP/MS after UAEE (refs. 35, 37).

(e) Measurement uncertainty (MU).—MU is estimated (a) Repeatability, intermediate reproducibility, recovery,
using the simplified approach based on existing validation and MU.—SRM 1568a (rice flour) was spiked at 1 mg/L As
data proposed by Barwick and Ellison (40), mainly precision for each As species [As(III), As(V), MA, DMA, and AB].
and trueness studies which, if properly planned to cover as Each spiked and nonspiked sample was analyzed in duplicate
many of the uncertainty sources previously identified as (k = 2) on 3 different days by three different analysts. Total As
possible, provide the necessary data required to calculate MU. extracted in spiked and nonspiked samples was also
Precision and trueness contributions are combined as follows quantified for each matrix in order to check extraction
to obtain the overall uncertainty: recovery according to total As reference value. The HorRat
value was calculated the same way as for the SLV.
Standard uncertainty: u = Mean ´ CV(iR) 2 + RSD(Rec) 2corrected (b) Repeatability of several samples.—SRM 1568a (rice
flour), Institute for Reference Materials and Measurements
Expanded uncertainty: U = 2 ´ u (IRMM) 804 (rice flour), CRM 627 (tuna fish), SRM 2976
(mussel tissue), cereal- and fish-based baby food, and
(which gives a level of confidence of approximately 95%).
seaweed flakes were analyzed in duplicate (k = 2) on the same
Verification day and by the same analyst. Total As extracted in spiked and
nonspiked samples was also quantified for each sample in
Before using an official method or an in-house validated
order to check extraction recovery according to the total As
method (without any modification in the procedure and in the
reference value. Total As reference values for seaweed flakes
scope of application), each laboratory has to demonstrate that
and cereal- and fish-based baby food were determined by
it is able to perform the method correctly. This step is called
ICP/MS after High Pressure Asher digestion.
verification and means checking that the target performance
characteristics, previously established during the SLV, are External QC
reached. Method verification was performed in two steps:
checking repeatability, intermediate reproducibility, recovery, The method was complemented with information on
and MU on one reference sample; and checking repeatability accuracy demonstrated with regular participation in
on several samples representative of those that will be proficiency schemes. The method was tested on an
routinely analyzed in the laboratory. The verification was interlaboratory comparison (ILC) of the European
performed in another Nestlé laboratory (Asia) on the same Union-Reference Laboratory for Heavy Metals in Feed and
type of instruments as those used for the SLV. Food, and an external proficiency testing (p-test) scheme of
954 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

Table 7. Results of the evaluation of trueness for each As species

SRM 1568a (rice flour) Baby food CRM 627 (tuna fish)

Ref. Ref. Ref.

value, Mean,c Recovery, RSD value, Mean,c Recovery, RSD value, Mean,c Recovery, RSD
Analyte Spikea mg/kgb mg/kgb % (Rec.), % mg/kgb mg/kgb % (Rec.), % mg/kgb mg/kgb % (Rec.), %

AsI b 0.100 0.100 100 7.5 0.050 0.048 95 5.3 0.100 0.093 93 4.2
c 0.500 0.494 99 2.1 0.250 0.238 95 2.6 0.500 0.480 96 3.0
DMA d 0.050 0.051 102 9.4 0.025 0.025 100 2.4 0.050 0.048 97 2.2
e 0.250 0.243 97 3.0 0.125 0.128 102 1.2 0.250 0.236 94 2.6
MA d 0.050 0.061 122 9.5 0.025 0.028 111 5.1 0.050 0.057 114 6.5

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

e 0.250 0.274 110 4.5 0.125 0.139 111 4.2 0.250 0.255 102 1.9
AB d 0.050 0.052 103 5.0 0.025 0.026 105 6.6 0.050 0.057 114 5.7
e 0.250 0.253 101 2.6 0.125 0.129 103 1.4 0.250 0.263 105 6.7
Total As a 0.290 0.274d 94 2.3 0.045e 0.044d 97 4.5 4.80 3.46d 72 19.5

a
a = Nonspiked, b = spiked at 2 mg/L [sum of 1 mg/L additions of As(III) and As(V)], c = spiked at 10 mg/L [sum of 5 mg/L additions of As(III) and
As(V)], d = spiked at 1 mg/L, and e = spiked at 5 mg/L.
b
Volume of dilution = 10 mL, sample weight = 0.2 or 0.4 g for dry or wet sample, respectively.
c
Ctotal·measured – Cnative·measured [observed for duplicate (k = 2) on 6 different days (n = 6) by three different analysts].
d
Total extracted As obtained by ICP/MS after UAEE.
e
Reference value was obtained by ICP/MS after HPA digestion (refs. 35, 37).

the Central Science Laboratory–Food Analysis Performance time (5 min); sonication power (60%); number of sonication
Assessment Scheme (CSL–FAPAS), which focused on the treatments, 1; volume of enzymatic solution (3 mL); and
determination of both total and inorganic As in rice. The test concentration of protease type XIV (10 mg/mL).
item used in the ILC (IMEP-107) was a rice sample purchased
SLV
in a local supermarket and provided by the University of
Aberdeen. All details of this ILC are presented in the de la Linearity.—Table 4 shows the results of the evaluation of
Calle et al. report (41). The second p-test sample (FAPAS the linearity for each species studied. The calibration curves
round 07134) was powdered rice. Participant’s z-scores for constructed by plotting species concentration versus peak area
both p-tests were calculated as: response showed good linearity. The intercept was different
than zero for As(III) and As(V), since there remained a small
( x - X$ ) amount of AsI (<LOQ) in purified enzymes. In routine
z=
sp
Table 8. Results of the evaluation of expanded
where x = the participant’s report results, X$ = the assigned uncertainty U (in %) for each As species
value, and sp = the target SD. The z-score can be interpreted SRM 1568a CRM 627
as:|z| £2, satisfactory result; 2 < |z| £ 3, questionable result; and Analyte Spikea (rice flour) Baby food (tuna fish)
|z| > 3, unsatisfactory result.
AsI b 24.8 24.2 25.1
Results and Discussion
c 10.3 17.4 15.8
Optimization of the Extraction Method DMA d 23.4 13.3 11.3
In preliminary work, the effect of several parameters on e 13.2 8.5 13.3
total As recoveries was tested in rice samples to determine the MA d 24.1 14.9 17.3
best conditions of extraction including sonication time and e 24.5 24.1 12.0
power, number of sonication treatments, volume of enzymatic AB d 27.3 24.0 13.4
solution, and concentration of protease type XIV (data not
e 14.4 7.9 19.5
shown). Other parameters were fixed, as described in Sanz et
al. (21, 22). Total As a 18.8 20.8 40.2

Table 3 shows total As recoveries observed by ICP/MS a

a = Nonspiked, b = spiked at 2 mg/L [sum of 1 mg/L additions of
after optimized UAEE of several samples. Total As recoveries As(III) and As(V)], c = spiked at 10 mg/L [sum of 5 mg/L additions of
(between 82 and 114%) were satisfactory for all samples. As(III) and As(V)], d = spiked at 1 mg/L, and e = spiked at 5 mg/L.
Optimal conditions for the extraction method were: sonication
 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011 955

Table 9. Results of the verification of the method for SRM 1568a

Mean,b Ref. value, Mean,e Recovery,

Analyte Spikea mg/kgc CV(r), % CV(iR), % HorRatd mg/kg mg/kgc % RSD(Rec.), % U, %

AsI a 0.101 2.8 12.2 0.5

b 0.191 3.0 9.4 0.5 0.100 0.090 90 5.4 21.7
DMA a 0.134 5.3 6.7 0.3
c 0.177 4.0 12.0 0.6 0.050 0.043 86 11.0 32.6
MA a 0.014 3.7 13.0 0.4
c 0.064 4.7 11.6 0.5 0.050 0.050 100 8.4 28.6
AB c 0.047 2.8 9.0 0.4 0.050 0.046 91 3.7 19.5

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Total As a 0.262f 2.8 8.9 0.3 0.290 0.262f 91 4.9 20.3

a
a = Nonspiked, b = spiked at 2 mg/L [sum of 1 mg/L additions of As(III) and As(V)], and c = spiked at 1 mg/L.
b
Mean observed for duplicate (k = 2) on 6 different days (n = 6) by three different analysts.
c
Volume of dilution = 10 mL, sample weight = 0.2 or 0.4 g for dry or wet sample, respectively.
d
HorRat = Ratio using CV(iR) and Horwitz denominator.
e
Ctotal·measured – Cnative·measured.
f
Total extracted As obtained by ICP/MS after UAEE (refs. 35, 37).

conditions (single calibration at the beginning of the batch), interconversion between As(III) and As(V) could be observed
linear regression coefficients for all elements were with the conditions of extraction used. Values of CV(r) for all
satisfactory (R2 > 0.999). A second aliquot of a midrange species and total As extracted in SRM 1568a, baby food, and
standard [1 mg/L (as As)] was measured every six samples and CRM 627 were 2–10, 3–8, and 1–8%, respectively.
at the end of the batch to detect variations in response. Corresponding values of the reproducibility CV(iR) were
LOD and LOQ.—Table 5 shows the LOD and LOQ 5–13, 4–11, and 3–12%, respectively. However, the value for
estimated for dry and wet sample (volume of dilution = AsI in nonspiked baby food was 20% due to levels of native
10 mL, dry and wet sample weight = 0.2 and 0.4 g, concentration close to the LOQ. HorRat values (0.2–0.6)
respectively). LOQ values (0.005–0.011 and 0.009–0.020 mg/kg, never exceeded 1.5. Several values below 0.3 could be
for wet and dry sample, respectively) determined in this work explained by experience, training of analysts, a good
were the same order of magnitude, or slightly lower, than knowledge of the tested matrixes, and careful application.
those described in the literature, ranging from 0.001 to Trueness and recovery.—Table 7 shows the results of the
0.200 mg/kg (21, 22, 25, 27, 42–44). evaluation of trueness and recovery for each species and each
Selectivity.—A blank sample was analyzed and matrix. RSDs of recovery RSD(Rec.) regarding values for all
satisfactory enzyme purification was indicated since no species and total As extracted in SRM 1568a, baby food, and
significant signal was observed at the retention times of the As CRM 627 were 2–10, 1–7, and 2–7%, respectively.
species studied. Therefore, reagents in the blank (enzymes, Recoveries found for all species and total As extracted in
isopropanol) did not induce significant interferences in
chromatograms. Table 10. Results of the evaluation of repeatability for
The presence of a high amount of chloride (Cl) in the SRM 1568a, IRMM 804, SRM 2976, CRM 627, cereal- and
matrixes could induce an overestimation of As with ICP/MS fish-based baby food, and seaweed flakes for each
detection. Indeed, the polyatomic interference 40Ar35Cl has species
the same m/z (75) as As (75As). A blank sample (3 mL purified
Concn range, mg/kg Repeatability
mixed enzyme solution and 0.5 mL isopropanol diluted to
10 mL with Milli-Q water) spiked at 10 mg/L with Cl standard Analyte Minimum Maximum [CV(r)], %
solution (1000 mg/L) was analyzed to check the potential
effect of this interference on quantification of the As species AsI 0.011 0.203 3.9
studied. No significant signal was observed at the retention a
DMA 0.129 0.142
times of the As species studied. The selectivity of the method
MA 0.129 8.53 0.8
regarding 40Ar35Cl interference for the As species studied was
verified. AB 0.010 6.74 1.8
Precision.—Table 6 shows results of the evaluation of Total As 0.021 16.87 1.7
repeatability and intermediate reproducibility for each species a
Not enough values higher than the LOQ for this species to apply
and for each matrix. AsI values [sum of As(III) and As(V) calculation.
values] are presented in this and following tables, since
956 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

Table 11. As species values (mg/kg) observed in CRM 627 and SRM 1568a

Sample AsI DMA MA AB Total As Ref,

CRM 627 0.036 0.141 <0.014 3.9 4.1 This work

0.150 3.9 4.8 Certified
0.076 0.157 <0.016 4.1 4.4 25
0.007 0.154 0.010 4.1 4.2 51
NDa 0.139 ND 3.8 4.1 48
0.015 56
0.140 3.7 4.1 55
0.163 <0.003 4.1 4.9 58

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

ND ND 5.3 5.8 43
0.010 0.140 ND 3.6 3.8 21
SRM 1568a 0.098 0.129 0.016 <0.017 0.262 This work
0.290 Certified
0.089 0.135 0.008 ND 0.286 21
0.088 0.135 0.008 ND 0.286 22
0.092 0.174 0.008 0.274 49
0.278 15
0.096 0.173 0.012 0.281 53
0.102 0.155 0.009 0.267 50
0.106 0.158 0.013 0.277 28
0.109 0.165 0.015 0.288 47
0.0101 0.148 0.011 0.260 46
0.087 0.168 0.012 0.267 46
0.058 0.116 0.010 0.184 52
0.080 0.160 0.002 0.240 59
0.110 0.180 0.290 57
0.082 0.268 54

a
ND = Not detected.

SRM 1568a, baby food, and CRM 627 (94–122, 95–111, and tested matrixes. The low recoveries (70–80%) observed for
93–114%, respectively) were satisfactory compared to CRM 627 gave RSD (Rec.) and U slightly higher than target
acceptance criteria set by CODEX (39). However, recovery values.
for MA in SRM 1568a (rice flour) spiked at 1 mg/L (122%)
Verification
was slightly higher than the acceptance criteria (60–115%).
This slight overestimation could not be explained, but further Repeatability, intermediate reproducibility, recovery, and
investigations are planned to clarify this point. Recovery MU.—Table 9 shows the results of the evaluation of
(72%) for total As extracted in CRM 627 (tuna fish) was repeatability and intermediate reproducibility, trueness,
significantly different from 100% with an RSD (Re.c) value recovery, and MU for each species in SRM 1568a. Values of
of 20%. Sanz et al. (21) also observed lower total As recovery CV(r) for all species and total As extracted in SRM 1568a
for this sample. They supposed that the nonextracted As ranged between 3 and 5%. Corresponding values of the
fraction is strongly covalently bound to this matrix and cannot intermediate reproducibility CV(iR) ranged between 7 and
be measured without more aggressive conditions of extraction 13%. Recoveries (86–100%) for all species and total As
inducing risk of species transformation (21). This As fraction extracted were satisfactory for SRM 1568a. RSD (Rec.) and U
could be arsenolipid species found as natural constituents of for all species and total As extracted ranged between 4 and
tuna by Taleshi et al. (45) that are only extractable by specific 11% and 20 and 33%, respectively. These values were similar
lipid phase extraction. to those observed during the SLV. HorRat values (0.3–0.6)
MU.—Table 8 shows the results of the evaluation of MU never exceeded 1.5 and were higher than 0.3.
for each species and each matrix. Calculated derived relative Repeatability of several samples.—Table 10 shows the
U values range from 8 to 40% for all species in the three results of the evaluation of repeatability for several samples
 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011 957

Table 12. Results (mg/kg) of AsI and total As quantification in IMEP-107 and FAPAS 07134

Sample Analyte Mean Ref. value z-Score AsI, %a

IMEP 107 (Rice) AsI 0.096 ± 0.015b 0.107 ± 0.014 –0.7 57.8
Total As 0.166 ± 0.009 0.172 ± 0.018 –0.2
FAPAS 07134 (Powdered rice) AsI 1.410 ± 0.352 1.330 ± 0.204 0.4 84
Total As 1.679 ± 0.235 1.516 ± 0.228 0.7

a
Percentage of AsI found in the sample.
b
The uncertainty reported is based on the measurement of the three replicates (i.e., precision) since the uncertainty estimated in the SLV was
not yet available at deadline for submission of results.

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

for each species. Values of CV(r) for all species and total As References
extracted in several samples ranged between 1 and 4%. These
values were similar to CV(r) values obtained during the SLV. (1) Mandal, B.K., & Suzuki, K.T. (2002) Talanta 58, 201–235
Table 11 shows values of As species and total As extracted (2) Sharma, V.K., & Sohn, M. (2009) Environ. Int. 35, 743–759
observed in SRM 1568a (rice flour) and CRM 627 (tuna fish) (3) Bissen, M., & Frimmel, F.H. (2003) Acta Hydrochim.
compared to those reported in the literature (17, 21, 22, 25, 28, Hydrobiol. 31, 9–18
43, 46–59). Values obtained in this work were generally in (4) Tseng, W.P. (1977) Environ. Health Perspect. 19, 109–119
good agreement with those reported in previous studies. (5) Gebel, T.W. (2002) Int. J. Hyg. Environ. Health 205,
505–508
External QC.—Table 12 shows the results of the
(6) Arsenic and Arsenic Compounds (2001) Environmental
quantification of AsI and total As in IMEP-107 and FAPAS Health Criteria 224, 2nd Ed., WHO, Geneva, Switzerland
round 07134 by the validated method described in this paper.
(7) Francesconi, K.A., & Kuehnelt, D. (2004) Analyst 129,
Results of AsI and total As obtained were satisfactory since 373–395
z-scores for both tests were lower than ç2ç. (8) Determination of Arsenic in Seafood by Graphite Furnace
Atomic Absorption Spectrometry (GFAAS) After Microwave
Conclusions Digestion (2004) Brussels, Belgium, European Standard EN
14332
(9) Determination of Arsenic by Hydride Generation Atomic
An As speciation method using IC-ICP/MS after UAEE was Absorption Spectrometry (HGAAS) After Dry Ashing (2005)
developed. The extraction method is faster and simpler than Brussels, Belgium, European Standard EN 14546
previous classical water–methanol extraction methods (32). (10) Determination of Arsenic and Selenium by Hydride
Optimized chromatographic conditions allow the Generation Atomic Absorption Spectrometry (HGAAS) After
satisfactory and rapid separation and detection of the five Pressure Digestion (2005) Brussels, Belgium, European
major As species [As(III), As(V), DMA, MA, and AB]. The Standard EN 14627
method was validated according to ISO/IEC 17025:2005 (11) Official Methods of Analysis (2000) AOAC
standard (33) in rice, seafood, and rice/seafood-based baby INTERNATIONAL, Gaithersburg, MD, Method 986.15
food. Recoveries for total As extracted and all individual As (12) Determination of Arsenic, Cadmium, Mercury, and Lead in
Foodstuffs by Inductively Coupled Plasma/Mass
species in rice, baby food, and seafood were satisfactory.
Spectrometry (ICP/MS) After Pressure Digestion (2009)
Verification results obtained for precision, trueness, and European Standard, Brussels, Belgium, EN 15763
uncertainty were in good agreement with those obtained
(13) EFSA (2009) EFSA J. 7, 1351–1550
during in-house validation. The method was also tested in two
(14) Chemical Contaminants and Residues in Food Community
interlaboratory comparisons with satisfactory z-scores. The Metals Subgroup (2008) Metals Subgroup Issues Call for
method, providing confident quantification of AsI species in Methods Needs, AOAC INTERNATIONAL, Gaithersburg,
rice and seafood raw materials and finished products, will MD
answer increasing regulatory requirements and ensure (15) Meharg, A.A., Sun, G., Williams, P.N., Adomako, E.,
As-containing product safety for the consumer. Deacon, C., Zhu, Y.G., Feldmann, J., & Raab, A. (2008)
Environ. Pollut. 152, 746–749
Acknowledgments (16) Meharg, A.A., Lombi, E., Williams, P.N., Scheckel, K.G.,
Feldmann, J., Raab, A., Zhu, Y., & Islam, R. (2008) Environ.
Sci. Technol. 42, 1051–1057
We wish to acknowledge the following collaborators for (17) Pizarro, I., Gomez, M., Palacios, M.A., & Camara, C. (2003)
their skillful participation in verification of the method: Anal. Bioanal. Chem. 376, 102–109
Kwok-Wah Mak, Bee-Leng Teo, and Yin-Leng Choy from (18) Pizarro, I., Gomez, M., Camara, C., & Palacios, M.A. (2003)
Nestlé Quality Assurance Center Singapore, Singapore. Anal. Chim. Acta 495, 85–98
958 DUFAILLY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 3, 2011

(19) Schoof, R.A., Yost, L.J., Eickhoff, J., Crecelius, E.A., (40) Barwick, V.J., & Ellison, S.L.R. (2000) Accredit. Qual.
Cragin, D.W., Meacher, D.M., & Menzel, D.B. (1999) Food Assur. 5, 47–53
Chem. Toxicol. 37, 839–846 (41) de la Calle, M.B., Linsinger, T., Emteborg, H., Charaud-Got,
(20) Gong, Z., Lu, X., Ma, M., Watt, C., & Le, X.C. (2002) J., & Verbist, I. (2010) Report of the Seventh Interlaboratory
Talanta 58, 77–96 Comparison Organized by the European Union-Reference
(21) Sanz, E., Munoz-Olivas, R., & Camara, C. (2005) J. Laboratory for Heavy Metals in Feed and Food, JRC-IRMM,
Chromatogr. A 1097, 1–8 Geel, Belgium
(22) Sanz, E., Munoz-Olivas, R., & Camara, C. (2005) (42) Hirata, S., Toshimitsu, H., & Aihara, M. (2006) Anal. Sci. 22,
Anal.Chim. Acta 535, 227–235 39–43
(23) Londesborough, S., Mattusch, J., & Wennrich, R. (1999) (43) Mato-Fernandez, M.J., Otero-Rey, J.R., Moreda-Pineiro, J.,
Fresenius’ J. Anal. Chem. 363, 577–581 Alonso-Rodriguez, E., Lopez-Mahia, P.,
(24) Daus, B., Wennrich, R., Morgenstern, P., Weiss, H., Muniategui-Lorenzo, S., & Prada-Rodriguez, D. (2007)
Leonhardt Palmieri, H.E., Nalini, H.A., Leonel, L.V., Guedes Talanta 71, 515–520
(44) Wang, R.Y., Hsu, Y.L., Chang, L.F., & Jiang, S.J. (2007)

Downloaded from https://academic.oup.com/jaoac/article-abstract/94/3/947/5655420 by guest on 15 April 2020

Monteiro, R.P., & Moreira, R.M. (2005) Microchim. Acta
151, 175–180 Anal. Chim. Acta 590, 239–244
(25) Dufailly, V., Noel, L., Frémy, J.M., Beauchemin, D., & (45) Taleshi, M.S., Edmonds, J.S., Goessler, W., Ruiz-Chancho,
Guérin, T. (2007) J. Anal. Atom. Spectrom. 22, 1168–1173 M.J., Raber, G., Jensen, K.B., & Francesconi, K.A. (2010)
(26) Dufailly, V., Guérin, T., Noel, L., Frémy, J.M., & Environ. Sci. Technol. 44, 1478–1483
Beauchemin, D. (2008) J. Anal. Atom. Spectrom. 23, (46) Ackerman, A.H., Creed, P.A., Parks, A.N., Fricke, M.W.,
1263–1268 Schwegel, C.A., Creed, J.T., Heitkemper, D.T., & Vela, N.P.
(27) Kohlmeyer, U., Kuballa, J., & Jantzen, E. (2002) Rapid (2005) Environ. Sci. Technol. 39, 5241–5246
Commun. Mass Spectrom. 16, 965–974 (47) D’Amato, M., Forte, G., & Caroli, S. (2004) J. AOAC Int. 87,
(28) Kohlmeyer, U., Jantzen, E., Kuballa, J., & Jakubik, S. (2003) 238–243
Anal. Bioanal. Chem. 377, 6–13 (48) Gomez Ariza, J.L., Morales, E., Sanchez-Rodas, D., &
(29) Kohlmeyer, U., Jakubik, S., Kuballa, J., & Jantzen, E. (2005) Giraldez, I. (2000) TrAC Trends Anal. Chem. 19, 200–209
Microchim. Acta 151, 249–255 (49) Heitkemper, D.T., Vela, N.P., Stewart, K.R., & Westphal,
(30) Ronkart, S.N., Laurent, V., Carbonnelle, P., Mabon, N., C.S. (2001) J. Anal. Atom. Spectrom. 16, 299–306
Copin, A., & Barthélemy, J.P. (2007) Chemosphere 66, (50) Matos Reyes, M.N., Cervera, M.L., Campos, R.C., & de la
738–745 Guardia, M. (2007) Spectrochim. Acta B 62, 1078–1082
(31) Simon, S., Tran, H., Pannier, F., & Potin-Gautier, M. (2004) (51) Nakazato, T., Taniguchi, T., Tao, H., Tominaga, M., &
J. Chromatogr. A 1024, 105–113 Miyazaki, A. (2000) J. Anal. Atom. Spectrom. 15, 1546–1552
(32) Zbinden, P., Andrey, D., & Blake, C. (2000) Atom. (52) Nam, S.H., Cheng, J., Mindak, W.R., & Capar, S.G. (2006)
Spectrosc. 21, 205–215 Bull. Korean Chem. Soc. 27, 903–908
(33) General Requirements for the Competence of Testing and (53) Narukawa, T., Inagaki, K., Kuroiwa, T., & Chiba, K. (2008)
Calibration Laboratories (1999) 1st Ed., International Talanta 77, 427–432
Organization for Standardization/International (54) Signes, A., Mitra, K., Burlo, F., & Carbonell-Barrachina,
Electrotechnical Commission 17025, Geneva, Switzerland A.A. (2008) Food Addit. Contam. A 25, 41–50
(34) Cuderman, P., & Stibilj, V. (2009) Anal. Bioanal. Chem. (55) Sloth, J.J., Larsen, E.H., & Julshamn, K. (2003) J. Anal.
393, 1007–1013 Atom. Spectrom. 18, 452–459
(35) Zbinden, P., & Andrey, D. (1998) Atom. Spectrosc. 19, (56) Sloth, J.J., Larsen, E.H., & Julshamn, K. (2005) J. Agric.
214–219 Food Chem. 53, 6011–6018
(36) Wolf, W.R., & Andrews, K.W. (1995) Fresenius’ J. Anal. (57) Sun, G.X., Williams, P.N., Zhu, Y.G., Deacon, C., Carey,
Chem. 352, 73–76 A.M., Raab, A., Feldmann, J., & Meharg, A.A. (2009)
(37) Hammer, D., Nicolas, M., & Andrey, D. (2005) Atom. Environ. Int. 35, 473–475
Spectrosc. 26, 203–208 (58) Suner, M.A., Devesa, V., Munoz, O., Vélez, D., & Montoro,
(38) Horwitz, W., & Albert, R. (2006) J. AOAC Int. 89, R. (2001) J. Anal. Atom. Spectrom. 16, 390–397
1095–1109 (59) Williams, P.N., Price, A.H., Raab, A., Hossain, S.A.,
(39) Codex Procedural Manual (2010) 19th Ed., Codex Feldmann, J., & Meharg, A.A. (2005) Environ. Sci. Technol.
Alimentarius Commission, Rome, Italy 39, 5531–5540