Total and Inorganic Arsenic Determination in Soil, Sediments, and Sludge by Hydride Generation High-Resolution Continuum Source Quartz Tube Atomic Absorption Spectrometry in Dilute Hydrochloric Acid Using Borohydride and L-Cysteine

Abstract A sensitive and interference-free method based on hydride generation high-resolution continuum source quartz tube atomic absorption spectrometry in dilute HCl was developed for total and inorganic As (tAs, iAs) determination in soil, sediments, and sludge. Sample preparation involved microwave-assisted digestion in aqua regia for tAs and extraction in 10 mol L−1 HCl for iAs. In both cases, the prereduction of As(V) to As(III) was carried out with L-cysteine, while derivatization to arsine with 0.6% NaBH4 was performed in the presence of 0.2% L-cysteine in 0.01 mol L−1 HCl (pH 2.00 ± 0.01). Inorganic arsenic determination was validated with 87 ± 7% extraction efficiency. The recovery of tAs in certified reference materials was 98 ± 23%, while the value for iAs in spiked samples was 95 ± 10%. The method was selective for iAs determination and did not require separation of iAs by liquid-liquid extraction in an organic solvent. The results were verified using the Tukeyʼs multiple comparison test for p > 0.05 (experimental values 0.063 to 0.999). The method is greatly simplified due to no requirement for the separation of iAs species, absence of spectral interference, similar derivatization conditions for tAs/iAs, and possibility of using external calibration. The limits of detection for tAs and iAs were 0.006 mg kg−1. Concentrations of 0.37–545 mg kg−1 tAs and 0.33–280 mg kg−1 iAs in real samples were determined with precisions from 3.7 to 12.2% and 6.8 to 15.0%.


Introduction
Water and agricultural soil contamination with heavy metals, pesticides, pharmaceutical residues, dyes, and arsenic compounds, resulting from industrial, municipal, agricultural and domestic practices has become an environmental issue of major concern and received high attention for chemists and specialists in health (Azam et al. 2022; mobility for the As(III) and As(V) species from soil in water, 1 mol L −1 HCl and 10 mol L −1 HCl dependent on As concentration.Chirita et al. (2023) speciated iAs in foodstuffs using only 0.01 mol L −1 HCl and 0.2% L-cysteine for sample extraction, prereduction of As(V) to As(III) and derivatization to arsine without separation by LLE in organic solvents (toluene).The authors reported the extraction in 0.01 mol L −1 HCl more effective than in 10 mol L −1 HCl for such samples.
The aim of this study was the development and validation of a sensitive and interference-free method for tAs and iAs determination by hydride generation high-resolution continuum source quartz tube atomic absorption spectrometry (HG-HR-CS-QTAAS) in environmental samples.The method is based on the derivatization to arsine with NaBH 4 and L-cysteine in dilute HCl.The method was investigated for the analysis of agricultural soil, water sediment, and sludge from wastewater treatment plants used as amendment.An analytical performance study was carried out to estimate the limit of detection (LOD), accuracy, and precision.The accuracy for tAs determination was checked by analyzing certified reference materials (CRMs).
Validation regarding the iAs fraction was more difficult, since CRMs with matrices matching environmental samples are still missing.Thus, the extraction of iAs in 10 mol L −1 HCl and determination using HG derivatization was validated directly in the extract without LLE separation, external and standard addition.The results were compared to those obtained using separation by LLE in toluene-diluted HCl and determination also by external and standard addition.The absence of non-spectral effects coming from the multimineral matrix was characterized using Tukey's statistical test (p > 0.05) (Tukey 1949) by comparing the results for iAs by the four approaches.The proposed procedure for the determination of iAs is easier than that developed by Chappell, Chiswell, and Olszowy (1995) based upon separation by LLE in the chloroform-water system and derivatization in 1 mol L −1 HCl.We determined the sum of As(III) and As(V) after derivatization in the presence of 0.2% L-cysteine and 0.01 mol L −1 HCl circumventing separation.Moreover, we extended the applicability of the method for iAs in soil without separation to the analysis of water sediment and sludge.
The iAs determination without separation was examined with the risk of traces of MMA and DMA in non-ionized form to cause mutual interferences during conversion of iAs(III) to arsine.Compared to the conventional separation of the iAs species by LLE extraction in the chloroform-water system, we used the extraction in toluene, as it was shown to be the most effective for the extraction of covalent species (Haghnazari et al. 2018).The elemental analysis of the matrix was carried out by HR-CS-FAAS using an acetylene-air flame.
A solution of 10 mol L −1 HCl as extraction reagent for iAs and 0.01 mol L −1 HCl as medium for arsine generation and extraction of As(III) from toluene were prepared.A solution of 40% NaOH was used for the pH adjustment to 2.00 ± 0.01 in the calibration standards and liquid samples before derivatization to arsine.A solution of 10% (m/v) L-cysteine in 0.01 mol L −1 HCl (pH 2.00 ± 0.01) was used as the prereductant of As(V) to As(III).Calibration standards of As(III) (0-10 lg L −1 ; n ¼ 8) in 0.01 mol L −1 HCl and 0.2% L-cysteine were prepared from As(V) and prereduction with 1 mL 10% L-cysteine (pH 2.00 ± 0.01) by heating at 90 ± 5 � C for 10 min on water bath, similar to Chirita et al. (2023).A solution of 0.2% L-cysteine in HCl (pH 2.00 ± 0.01) was used as the blank.For the accuracy evaluation of iAs determination by HG-HR-CS-QTAAS, the extracts in 10 mol L −1 were spiked with 5 mg L −1 As(III) or As(V) and subjected to prereduction and derivatization.A solution of 0.6% (m/v) NaBH 4 containing 0.05% (v/v) antifoam agent and stabilized in 0.01% (m/v) NaOH was prepared as derivatization reagent for As(III) to arsine.Ultrapure water (18 MX cm) was used throughout.A solution of 5% (v/v) HNO 3 was used to wash the glassware by soaking for 12 h.The cross contamination between sample measurements was avoided by cleaning the reaction cell of the HS55 hydride generator with ultrapure water.

Certified reference materials and test samples
The accuracy of the HG-HR-CS-QTAAS method for the determination of tAs was checked by analyzing CRMs matching the matrix of the test samples, namely: BCR − 280 R lake sediment (Institute for Reference Materials and Measurements -IRMM, Geel, Belgium), LGC6141 soil contaminated with clinker ash and RTC-SQC-001-30G metals in soil (Department of Trade and Industry, Teddington Midlesex, UK), Metranal-32 light sandy soil and Metranal-34 loam (Analytika Spol, Vysocany, Czech Republic) and CRM025-050 metals in soil (Resource Technology Corporation, Laramie, USA).
Test samples of agricultural soil, sludge from a wastewater treatment unit, river and estuarine sediment were analyzed to demonstrate the broad applicability of the HG-HR-CS-QTAAS method for tAs and iAs determination.In these samples, quantification of the toxic iAs fraction is of great concern because of the possible transfer of the bioavailable forms to groundwater and plant culture or sediments with high risk of exposure of humans and aquatic organisms.Also, sludge resulting from water treatment plants is still used as amendment for agricultural soils.Thus, such studies are essential, especially in areas with anthropogenic activities where wastes containing As are stored in tailings dumps, sometimes in neighborhoods of urban zones and rivers, as well as in areas with naturally high levels of arsenic in soil (Butaciu et al. 2017;Cordos et al. 2006;Levei et al. 2013).The soil samples were collected from areas that were in the past under the anthropogenic exposure from the non-ferrous metals industry, but also from areas where As has geogenic origin.The sediments were collected from a river that was also in the past under the anthropogenic action of non-ferrous metals mining.

Sample preparation and analysis
Sample preparation for tAs and iAs determination is schematically presented in Figure 1.

Sample preparation for tAs determination
The preliminary processing of samples and microwave-assisted digestion were carried out according to Mihaltan et al. (2013).0.1-0.5 g CRM or soil test, water sediment and sludge samples were digested in 12 mL aqua regia and diluted to 25 mL with ultrapure water.Prereduction of As(V) to As(III) was conducted on aliquot volumes of 5-20 mL under the conditions presented in Figure 1.Next, the pH was adjusted to 2.00 ± 0.01 by potentiometric titration with 40% (m/v) NaOH and diluted to 50 mL with HCl.The concentration of L-cysteine in the final solution was 0.2% (m/v).The tAs concentration was determined by the external calibration in 5 mL of sample after the addition of 3.5 mL of 0.6% NaBH 4 containing 0.05% antifoam agent using external calibration.

Sample preparation for iAs determination
The procedures used for iAs determination are schematically presented in Figure 1.The iAs fraction was extracted from 0.2-0.5 g CRM/test sample in 20 mL of 10 mol L −1 HCl solution by shaking for 20 min at room temperature, as previously described in our laboratory (Cordos et al. 2006).The extract was diluted to 25 mL with ultrapure water.The sample was subsequently processed for the determination of iAs with and without separation by LLE using the toluene-0.01mol L −1 HCl system.The prereduction of As(V) to As(III) was carried out with 10% L-cysteine under the same conditions as for tAs.For LLE, the inorganic As(III) species were separated by extraction in 2 � 10 mL toluene and back-extracted in 2 � 5 mL 0.01 mol L −1 HCl.The procedure was previously developed for iAs determination in foodstuffs by Chirita et al. (2023).The concentration of 0.2% L-cysteine was achieved by adding 1 mL of 10% L-cysteine and diluted to 50 mL with HCl solution (pH 2.00 ± 0.01).The concentration of iAs was determined by external/standard addition calibration under the conditions presented in Figure 1.
In the procedure without LLE separation, iAs was determined using external calibration in the acidic extract after the prereduction of As(V) to As(III) and dilution of the sample to 50 mL with HCl pH 2.00 ± 0.01, as for tAs.In the standard addition method, 5-20 mL extracts of iAs in 10 mol L −1 HCl were spiked with 5 mg L −1 As(V) for both procedures with and without separation by LLE.A rigorous control of pH 2.00 ± 0.01 is necessary as the arsine generation efficiency in diluted HCl in the presence of L-cysteine is strongly pH-dependent (Chirita et al. 2023).Samples were measured no later than 2 h after the prereduction of As(V) to As(III) because a decreasing of As signal was observed after several hours.

Instrumentation
The HG-HR-CS-QTAAS instrumentation (Analytik Jena AG, Jena, Germany) was previously described (Chirita et al. 2023).The experimental set-up is described in the supplemental online material (Section 1, Figure S1).The HG-HR-CS-QTAAS system consisted of the ContrAA 300 spectrometer coupled to HS55 hydride generator, a quartz tube atomizer (QT) (140 mm length, 15 mm i.d.) heated in an electric oven, and a reaction cell for mixing the aliquot sample with the NaBH 4 solution introduced via a single channel peristaltic pump.The absorption spectrum was recorded using a high-resolution double monochromator (2 pm full width at half maximum (FWHM)) equipped with a charge coupled device detector over a spectral range of ± 0.1 nm (200 pixels) around the analytical line As 193.696 nm.Five pixels (Central Pixel ± 2) in the middle of the spectral window were associated to the As 193.696 nm line while the others are used for background correction.Perma Pure MD-050-48 Nafion membrane tubing (120 cm length), Chromoservis (Prague, Czech Republic) was mounted between the reaction cell and the QT compared to the commercial ContrAA 300 system.According to the manufacturer 0 s specifications, the Perma Pure Nafion membrane removed more than 85% of the water vapor (Target Dew Point of at least −8 � C).The removal of water droplets entrained from the reaction cell was useful to overcome the spectral interference coming from the pseudo-continuum structure of background on each side of the As line.The jagged shape signal of the pseudo-continuum structure may prevent in the visualization of the analytical line and background correction (Kratzer et al. 2011;Lacko et al. 2022).
The microwave-assisted digestor Berghof MWS3þ system (Berghof, Germany) was used for samples preparation for tAs determination while the ContrAA 300 spectrometer HR-CS-FAAS with air-acetylene flame was employed for the multielemental analysis of the sample matrix.

Working conditions for the HG-HR-CS-QTAAS equipment
The operation of the HS55 batch system involved several sequences.A volume of 5 mL sample was introduced into the reaction cell and the oxygen dissolved in the liquid was removed by passing an Ar stream of 6 L h −1 for 60 s.Next 3.5 mL of NaBH 4 solution were introduced via the peristaltic pump and the generated arsine was drawn into the QT with Ar (6 L h −1 ) through the Nafion membrane tube.The removal of oxygen from the liquid sample before the NaBH 4 addition is necessary as the atomization of arsine in QT is only effective in an argon atmosphere containing traces of hydrogen and oxygen at a ratio of O 2 /H 2 < 0.1 (D'Ulivo and Sturgeon 2022; Dvo� r� ak et al. 2019; Kratzer et al. 2011Kratzer et al. , 2022;;Lacko et al. 2022).
Also, the quantity of hydrogen generated in the diluted HCl-L-cysteine system is lower than in the classical procedure using concentrated HCl in the absence of L-cysteine.No extra flow of hydrogen was used in this experiment.On the other hand, the fine structure of the background absorption of oxygen causes a spectral interference on As 193.696 nm line (Kratzer et al. 2011;Lacko et al. 2022).The fine structured absorption spectrum of O 2 below 200 nm is the result of B 3 P u ̅ -X 3 P g ̅ Schumann-Runge transition (Krupenie 1972), while the pseudo-continuum absorption spectrum of H 2 O between 190 and 260 nm is due to the water dimer ((H 2 O) 2 ) with 7.37 eV excitation energy (Harvey, Jung, and Gerber 1998;Makogon, Ponomarev, and Tikhomirov 2013) but is not yet completely understood.This interference is difficult to remove using background correction, such as least square background correction offered by the software of the instrument, since in batch operation the conditions in the QT change during the measurement.
D'Ulivo and Sturgeon (2022) and Lacko et al. (2022) showed that the atomization of arsine was favored by the hydrogen radicals developed in the reaction between hydrogen and oxygen.Thus, it was introduced a pre-washing step of the reaction cell with 6 L h −1 Ar for 60 s before adding the NaBH 4 over the 5 mL of sample in the reaction cell.Figure 2 presents the background spectrum and absorption signals of As 193.696 nm line for arsine generation with NaBH 4 in 0.01 mol L −1 HCl in the presence of 0.2% L-cysteine with/without pre-washing of the reaction cell and vapor drying.Also, the shape of the transient response is illustrated.The pre-washing of the system before the addition of NaBH 4 and vapor drying via the Nafion tubing membrane results in the complete removal of the pseudo-continuum background shape.Under these conditions, the correction of the continuum background is easy.Drying of the gas streams was The peak height measurement was selected to make the analysis faster, as the maximum of the transient response appeared from 10 to 15 s, while the measurement of the integrated absorbance required recording over a period of 70 s (Figure 2(b)).Although the sensitivity was better in the peak area measurement, it did not improve the limit of detection, since the signal-to-noise ratio remained almost the same.Figure 2(b) shows the stronger influence of noise upon the analytical signal for longer recording periods.
Some parameters are crucial for arsine generation in dilute HCl in the presence of Lcysteine to ensure a high sensitivity for the analysis by HG-HR-CS-QTAAS and must be strictly controlled.Thus, the optimum concentration of NaOH is 0.01% m/v and the pH of the sample should be carefully adjusted in the range of 2.00 ± 0.01 (Chirita et al. 2023).The low concentration of NaOH is consistent with the corresponding concentration of HCl in the sample, so that even under these conditions excess HCl necessary for the derivatization to arsine is ensured.Other details about the influence of L-cysteine on the derivatization of different As species were published by Shraim, Chiswell, and Olszowy (1999).
The working conditions for HG-HR-CS-QTAAS are summarized in the supplemental online material (Section 1, Table S1) from an earlier study aiming the efficient generation and atomization of arsine, absence of non-spectral interference from transitional metal ions and the pseudo-continuum background, good accuracy, precision and limit of detection for the analysis of foodstuffs (Chirita et al. (2023)).

Figures of merit of the HG-HR-CS-QTAAS method
The characteristics of the calibration curve were 0.9975 determination coefficient (R 2 ) and 0.018 L lg −1 slope, while the instrumental LOD was 0.070 mg L −1 with a relative standard deviation of 8% and 18% for 25 days, which shows good linearity and longterm reproducibility.The instrumental limit of detection was calculated using the 3r criterion (LOD ¼ 3 s b /m) where (s b ) is the standard deviation of the signal for the blank solution (n ¼ 11) and (m) is the slope of the calibration curve.
A 0.5 g sample digested to 25 mL and a dilution of 2.5 times resulted in a limit of detection in soil, sediments, and sludge of 0.006 mg kg −1 both for tAs and iAs, regardless of the sample preparation or calibration procedure.The limit of detection for tAs was approximately 800 and 1900 times lower than the guideline value in agricultural soil (5 mg kg −1 ) and sludge used as soil amendment (10 mg kg −1 ), regulated in the Romanian legislation (Order no. 756/19971997, Technical norm 2004).The limit of detection for iAs was 2000 times lower than the threshold of 12 mg kg −1 iAs according to the Canadian soil quality guidelines for the protection of environmental and human health (Canadian Council of Ministers of the Environment 2007).Thus, HG-HR-CS-QTAAS is very sensitive and suitable for the measurement of tAs and iAs fraction in soil, sediments, and sludge.Inorganic As and tAs may be quantified above 0.020 mg kg −1 which is considered to be the limit of quantification (3.3xLOD).
A comparison of the limits of detection using HG-HR-CS-QTAAS with those reported in literature for other methods in terms of sample processing and detection are presented in the supplemental online material (Section 2, Table S2).
The results in Table S2 show that the limit of detection for As by HG-HR-CS-QTAAS under the optimum conditions was similar or better than those obtained by Hesse, Ristau, and Einax (2015) using HG-HR-CS-GFAAS without preconcentration and Valdivia et al. (2018) using preconcentration by solid phase extraction (SPE), arsine generation, and in-situ trapping.
Also, the our limit of detection was better than reported by Mihucz et al. (2017b) by HR-CS-GFAAS after SPE preconcentration of As and direct liquid sampling, and that achieved by Irisov, Musin, and Zakharov (2021) using Pd-modifier without preconcentration and direct liquid sampling.The detection limit obtained by HG-HR-CS-QTAAS was better than reported in the HG-LR-LS-AAS method with a line source and lowresolution spectrometer (T� anase et al. 2010), and in GFAAS using ultrasound-assisted supramolecular solvent microextraction and direct liquid sampling (Kashanaki, Ebrahimzadeh, and Moradi 2017).The advantage of As derivatization to arsine in the presence of L-cysteine in diluted HCl and determination by HG-HR-CS-QTAAS compared to the direct liquid sampling methods coupled or not with preconcentration is obvious.On the other hand, the limit of detection in the reported method was poorer than using direct solid sampling and HR-CS-GFAAS of 0.04 mg kg −1 As in agricultural soil (Schneider et al. 2018).However, the percentage of iAs fraction cannot be evaluated by this method without a prior extraction from the solid matrix.

Validation of the HG-HR-CS-QTAAS method for tAs and iAs in environmental samples
The obtained results for tAs in CRMs of various matrices digested in aqua regia presented in Table 1 demonstrate that the differences between the measured and certified concentrations are lower than the associated expanded uncertainty (k ¼ 2), so that the results are not significantly different for 95% confidence level.The mean recovery was 98 ± 23%.The U lab was higher than provided for the certified concentrations (U), explained by the significant 98 ± 23 a U is the expanded uncertainty for the certified value (k ¼ 2; 95% confidence level).b U lab is the expanded uncertainty in laboratory for the found value (k ¼ 2, n ¼ 5 parallel measurements and 95% confidence level).c U mean is the mean of expanded uncertainties in laboratory (U mean ¼ ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi Þ, where U lab1 , U lab2 , … , U labm are expanded uncertainty for each CRM and (m) the number of CRMs (m ¼ 7).
contribution of uncertainty associated with aliquot analysis to the combined uncertainty (u lab ).The method for the determination of tAs is robust and free of matrix effects, so that the external calibration and the same derivatization conditions may be used regardless of the sample matrix, i.e., soil, sediment or sludge.The concentration of elements in the multimineral matrix is presented in the supplemental online material (Section 3, Table S3).The mineral matrix primarily contains Ca, Fe, Mg, K, Na and Al, characteristic for environmental samples.The elevated concentrations of Cu, Pb and Zn were in samples collected near a former Cu ore processing facility.
Earlier, our group validated the extraction of iAs from soil collected from areas located or not under the action of non-ferrous metal industry.The extraction degrees reported for iAs were 60 ± 21% in 1 mol L −1 HCl and 87 ± 7% in 10 mol L −1 HCl (Cordos et al. 2006).The increase in HCl concentration resulted in a greater improvement of the extraction of inorganic As(V) than inorganic As(III) species.The authors also reported that iAs extracted in water and 1 mol L −1 HCl increased with the As concentration in soil.Thus, for soils containing below 50 mg kg −1 As, such as unpolluted and low polluted soils, in which As species are bound to various bearing phases, such as ferric arsenate or scorodite (FeAsO 4 �2H 2 O) and arsenopyrite (FeS 2-x As x ), the extraction reagent of choice is 10 mol L −1 HCl for quantitative speciation.
Based on this previous study, we used 10 mol L −1 HCl as extraction reagent for iAs species prior the determination in environmental samples.The results obtained for the procedures with or without iAs separation by double LLE in the toluene-HCl (0.01 mol L −1 ) system and external/standard addition calibration are presented in Table 2.
Since in non-oxidizing media, such as HCl, the matrix is incompletely destroyed, the presence of non-spectral interferences was investigated by comparing the results obtained by external calibration and standard addition with and without LLE separation of iAs.The Tukeyʼs multiple comparison test indicated that the results for iAs obtained by the four approaches presented in Table 2 were not significantly different and the null hypothesis was retained for p > 0.05 (p-values in the range 0.063-0.999).Consequently, the concentrations of iAs in the last column of Table 2 represent the mean of the four values.
The precision of iAs measurements using external calibration without LLE separation in CRMs of soil and water sediment evaluated from the combined uncertainty was in the range from 4.8 to 14.9%.The percentage of iAs fraction was dominant in CRMs of soil and water sediment in the range of 75 ± 26% and 67 ± 6%, respectively.Since no CRMs of soil and water sediment for iAs are available, the accuracy of the method for iAs measurement was investigated by analyzing spiked extracts with As(III) and As(V) up to 5 lg L −1 with or without separation in the toluene-HCl system.The recovery was 95 ± 10% for a confidence level of 95%.Therefore, the separation of iAs fraction from soil and sediment by LLE was considered unnecessary so that a simple extraction in concentrated HCl and derivatization to arsine (NaBH 4 in 0.01 mol L −1 HCl, 0.2% Lcysteine) and external calibration were successfully used.The obvious advantage of this procedure consists in the direct determination of iAs fraction in the acidic extract, without using an organic solvent, which simplifies the protocol.As expected, the concentrations of concomitant were higher in the samples obtained without separation of iAs by LLE, but did not affect the measurements in HG-HR-CS-QTAAS, another relevant advantage.The concentration of elements in the sample matrix is presented in supplemental online material (Section 3, Table S4).

Analysis of real samples
The results obtained for test samples of soil, water sediment and sludge from waste water purification stations are presented in Table 3.
The Tukeyʼs multiple comparison test showed no significant differences (p > 0.05) among the results obtained for the iAs fraction with/without LLE and external calibration/standard addition (p-values 0.126-0.999).The precision evaluated from the combined uncertainty was 3.7-12.2%for tAs and 6.8-15.0%for iAs.The method fulfills the acceptance criterion of the Association of Official Agricultural Chemists (AOAC) guide to be better than 30% (Smith et al. 2015).The percentage of iAs fraction from tAs in soil, water sediments, and sludge from water treatment plants was 79 ± 13%, 83 ± 17% and 24 ± 3%, respectively.For the Soil 1 sample, the low percentage of iAs extractable in 10 mol L −1 HCl was attributed to the fact that this sample came from the vicinity of a waste deposit of non-ferrous ore flotation station.By flotation, the As species are concentrated in tailings in forms not readily leachable in HCl, which may be in the nearby agriculture soils (Cordos et al. 2006).Ko et al. (2003) showed that the As in the agricultural soil located in the vicinity of tailing dumps was in a residual detrital from as Fe-and Ca-associated fractions, soluble only in aqua regia.In the case of sludge, the lower percentage of iAs fraction may be attributed to the high concentration of organic matter, and presence of organic As species non-reducible to arsine, but further studies are needed.However, the results highlight the 6.5-11.6 7.2-12.04.8-14.97.2-12.67.5-11.9a U lab is the expanded uncertainty in laboratory (k ¼ 2, n ¼ 5 parallel measurements and 95% confidence level).b RSD is the relative standard deviation evaluated from combined uncertainty (n ¼ 5 parallel measurements and 95% confidence level).
importance of knowing the level of the iAs fraction with high toxicity in these environmental samples.

Conclusions
HG-HR-CS-QTAAS was demonstrated to be sensitive and interference-free for the multimineral matrix and thus suitable for the determination of tAs and iAs in soil, water sediments, and sludge.The same derivatization conditions to arsine in the presence of L-cysteine in dilute HCl were used for all samples.Also, the use of external calibration with As(III) standards and avoidance of iAs separation by LLE simplified the protocol.
The results in this study may be the basis for the development of standard methods for As determination and iAs speciation in environmental samples.

Disclosure statement
No potential conflicts of interest was reported by the authors..7-12.2 4.4-20.2 5.7-15.2 6.8-15.0 a U lab is the expanded uncertainty in laboratory (k ¼ 2, n ¼ 5 parallel measurements and 95% confidence level).b RSD is the relative standard deviation evaluated from combined uncertainty (n ¼ 5 parallel measurements and 95% confidence level).

Figure 1 .
Figure 1.Analytical scheme of the sample preparation for tAs and iAs determination in environmental samples.

Figure 2 .
Figure 2. (a) Background and arsenic 193.696 nm absorption signals with and without pre-washing of the reaction cell and vapor drying obtained by arsine generation with NaBH 4 in 0.01 mol L −1 HCl in the presence of 0.2% L-cysteine and (b) the transient signal.

Table 1 .
Concentrations of tAs (mean ± expanded uncertainty) and recovery (mean ± expanded uncertainty) obtained by HG-HR-CS-QTAAS method using external calibration in CRMs of environmental samples after aqua regia digestion and derivatization to arsine in optimum conditions.

Table 2 .
Concentrations of iAs (mean ± expanded uncertainty (U lab ) a ) in CRMs of soil and river sediment after extraction in 10 mol L −1 HCl obtained by HG-HR-CS-QTAAS method with/without liquid-liquid extraction (LLE) in toluene-0.01mol L −1 HCl system and external calibration/standard addition and derivatization to arsine in 0.01 mol L −1 HCl and 0.2% L-cysteine.

Table 3 .
Concentrations (mean ± expanded uncertainty (U lab ) a ) in the analysed soil, water sediment and sludge samples for tAs determination after microwave-assisted digestion in aqua regia and iAs using extraction in 10 mol L −1 HCl with/without separation by liquid-liquid extraction (LLE) in toluene-0.01mol L −1 , prereduction and derivatization in 0.01 mol L −1 HCl and 0.2% L-cysteine.