Using citric acid stabilizing reagent to improve selective hydride generation-ICP-MS method for determination of Sb species in drinking water

Abstract

An improved selective hydride generation-inductively coupled plasma-mass spectrometry (SHG-ICP-MS) method using citric acid as the stabilizing reagent was established and evaluated for the speciation of inorganic Sb in drinking water. The use of citric acid as the stabilizing reagent improved the sensitivity of the Sb species over 2-fold, when compared to the conventional pH-dependent SHG strategy. The SHG reaction in the presence of citric acid was studied using high-resolution quadrupole-Orbitrap mass spectrometry (HR-MS). Almost 100% of the Sb(V) and Sb(III) reacted with citric acid to produce [Sb(OH)₃(C₆O₇H₅)]⁻ and [Sb(C₆O₇H₆)₂]⁻, respectively. The resultant Sb(V)-complex ([Sb(OH)₃(C₆O₇H₅)]⁻) cannot be reduced by NaBH₄, while the resultant Sb(III)-complex ([Sb(C₆O₇H₆)₂]⁻) originating from Sb(III) can be reduced thoroughly by NaBH₄. The ICP-MS signal intensity for Sb was initially low due to its high first ionization energy of 8.61 eV; however, an increase in sensitivity was observed with an addition of 2.0 mL min⁻¹ CH₄ to the ICP. Three elements (Ge, Sn, and Te) were evaluated as potential internal standards (ISs) to improve the stability of the HG signal. After careful evaluation, ¹¹⁸Sn was selected as the IS, and the relative standard deviation (RSD) for the of Sb(III) signal (1.0 μg L⁻¹) was improved from 5 to 2%. Under optimized conditions, excellent detection limits (4.0 ng L⁻¹ for Sb(III) and 6.0 ng L⁻¹ for total Sb) and satisfactory recoveries for Sb(III) (from 93.1 to 108%) and total Sb (from 98.5 to 103%) were obtained. The proposed method was employed for analysis of three types of drinking water: bottled water, tap water, and well water. The average Sb(III) and total Sb contents were 0.032 and 0.320 μg L⁻¹, respectively, in bottled water (N = 10), 0.197 and 1.05 μg L⁻¹, respectively, in tap water (N = 10), and 0.211 and 1.93 μg L⁻¹, respectively, in well water (N = 40). Thus, this high throughput method has great potential for the determination of inorganic Sb species at ultra-trace levels in drinking water.

---

Introduction

Sb is a toxic element that is present in the environment originating from natural sources as well as anthropogenic activity.¹ Sb compounds are listed as precursor pollutants by the United States Environmental Protection Agency (USEPA) and the European Commission (EC), with strictly concentration limits for drinking water (6.0 μg L⁻¹ for the USEPA, 5.0 μg L⁻¹ for the EC).² Inorganic Sb species are more toxic than organic species, with the toxicity of Sb(III) being ten times that of Sb(V).³ Therefore, speciation data for inorganic Sb in drinking water is required in order to fully assess its risks to human health.

Over the last three decades, the most commonly used for determining Sb species have been those based on chromatographic separation (e.g. high-performance liquid chromatography (HPLC),⁴ solid-phase extraction (SPE),⁵ liquid–liquid extraction⁶ or cloud-point extraction⁷) followed by element-specific detection (e.g. atomic absorption spectrometry (AAS),⁸ atomic fluorescence spectrometry (AFS),⁹ and inductively coupled plasma-mass spectrometry (ICP-MS)^10,11). The primary advantage of these approaches is unequivocal Sb species separation and specific online detection.¹² However, most of these techniques have disadvantages, including the risk of species inter-conversion and contamination, increased process complexity, co-elution of species of the same element, long retention times, and severe peak-tailing for Sb(III), as well as analytical problems associated with detector instability and interferences caused by the use of organic solvents.^13–17 Additionally, these types of instruments needed for analysis are not standard equipment in typical water-quality monitoring laboratories. Fortunately, non-chromatographic selective hydride generation (SHG) atomic spectrometric techniques have received increasing attention for Sb speciation analysis owing to their sensitivity, simplicity, and speed.^18,19 Selective hydride formation mainly relies on the pH-dependency of SbH₃ formation from Sb(III) and Sb(V) species. Sb(III) forms SbH₃ over a wide range of pH values (2–8), total Sb can be determined under strongly acidic conditions (pH < 1) or with a pre-reduction procedure, and Sb(V) content is calculated from the difference between total Sb and Sb(III).²⁰ Methods based on these strategies, (e.g. HG-AAS, HG-AFS, and HG-ICP-MS) are widely used for direct determination of inorganic Sb species in water due to their high throughput, low cost, and ease of handling, which are especially important for routine analysis.^21–24 Unfortunately, these methods are not highly selective and the interference of Sb(V) on the determination of Sb(III) starts to become significant at a ratio of 4 : 1.^24,25 Because most natural waters have Sb(V)/Sb(III) ratios on the order of 100 or more,²⁵ a series of complicated equations must be used in order to differentiate between Sb(III) and Sb(V) in real water samples. Mohammad et al.²⁴ have reported that citric acid forms a stable complex with Sb(V), suppressing its hydride generation. Based on this strategy, Santos et al.²⁶ established an SHG-inductively coupled plasma-optical emission spectroscopy (SHG-ICP-OES) method for determination of Sb species in bottled water, and achieved a limit of detection (LOD) of 50 ng L⁻¹ for Sb(III) and 110 ng L⁻¹ for total Sb. Unfortunately, this method suffers from poor reproducibility (17% RSD for Sb(III) and 18% (RSD) for total Sb(V)) due to poor signal stability and limited sensitivity.²⁶ In addition, the exact mechanism of selective hydride generation with citric acid is still unclear.²⁶ The analysis of natural drinking water samples by ICP-OES and ICP-MS is severely limited due to the high first ionization potential of Sb (8.61 eV) leading to poor Sb sensitivity.²⁷ More than 80% of natural samples cannot be analysed without pre-concentration²⁶ unless a more reproducible technique producing higher Sb sensitivity is developed.

This work focuses on the development of a reliable method for the determination of inorganic Sb species at ultra-trace levels in drinking water by SHG-ICP-MS using citric acid as a stabilizing reagent. The selectivity of the HG reaction in the presence of citric acid is discussed, and details on the improvement of the stability of the SbH₃ signal by selecting a suitable internal standard (IS), as well as increasing the poor ICP-MS signal of Sb, are presented. The optimization of the technique, its analytical performance, and its application to the direct determination of ng L⁻¹ levels of inorganic Sb species in different types of drinking water, are discussed in detail.

Experimental

Instrumentation

An Agilent 7700x ICP-MS (Agilent Technologies, USA) equipped with an aerosol dilution system was used as the element detector.^28,29 A flow of methane was merged with the sample aerosol between the gas–liquid separator and the torch using a T piece.³⁰ The optimized instrument operating parameters are listed in Table 1. Quantification was carried out by an external calibration approach using ¹¹⁸Sn as an IS to compensate for signal drift. A continuous flow system hydride generator (PerkinElmer, USA) with slight modification was used (Fig. 1). The acidified sample and IS were continuously pumped and mixed with a pumped stream of NaBH₄ to produce SbH₃ within the continuous flow system. Two mixing T junctions (Chemifold Assembly) were used to combine the sample and reductant streams and to add the stripping argon. A Q Exactive hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific, Germany) was used to investigate the complexation between the citric acid and the Sb compounds. The instrument operation conditions were similar to previous work and described in Yang et al.³¹

Table 1 SHG-ICP-MS operating parameters

ICP-MS instrument	Agilent 7700x
RF power, W	1550
Plasma gas flow, L min^−1	15
Auxiliary gas flow, L min^−1	1.0
Sampling depth, mm	7
Dwell time, ms	30
Sweeps	100
Readings	1
Replicates	3
Isotopes monitored	^121Sb
Internal standard	50 μg L^−1 Sn
CH4 flow, mL min^−1	2

---

HG system	PerkinElmer
0.5% NaBH4 uptake rate, mL min^−1	1.5
Sample uptake rate, mL min^−1	1.5
Carrier gas flow, L min^−1	1.15
Concentration of HCl, %	2.0

---

Fig. 1 Schematic of the hydride generator used for Sb speciation.

Reagents and materials

High-purity water (18.2 MΩ cm) used for the preparation of all blanks, standards, and samples was obtained from a Millipore water purification system (Millipore, France). Ultra-pure hydrochloric acid (HCl) was prepared from analytical grade reagents (Alfa Aesar Ltd., Tianjin) using a sub-boiling distillation system. Standard solutions (1000 μg mL⁻¹) were prepared by dissolving SbCl₃ (99.99%, Sigma Aldrich, USA) in 20% (v/v) HCl and potassium hexahydroxoantimonate(V) (99.0%, Sigma Aldrich, USA) in hot water. The reductant solution was prepared from NaBH₄ (99.99%, Sigma Aldrich, USA) in 0.1% (v/v) NaOH (99.99%, Sigma Aldrich, USA). Single element solutions of Ge, Sn, and Te (1000 μg mL⁻¹) were obtained from the National Centre for Analysis and Testing of Steel Materials. Citric acid (99.5%) and KI were obtained from Sigma Aldrich (USA).

Sample preparation

After filtering through a 0.45 μm membrane filter, 18 mL water samples were acidified with 2% HCl and divided into two 9.0 mL portions. For Sb(III) determination, 1.0 mL of 10% citric acid was added to the first portion for a final solution containing 0.5% citric acid. One mL of 30% KI was added to the second portion to pre-reduce the Sb(V) for total inorganic Sb determination in a final solution contained 6% KI.

Results and discussion

SHG for Sb speciation

SHG provides a faster, less expensive and more widely applicable method for the determination of inorganic Sb species. Generally, it can be used to directly discriminate Sb(III) and Sb(V) based on pH.²⁵ To check the performance of this conventional pH-dependent method, the effects of solution pH values on the separation efficiencies of Sb(III) and Sb(V) were tested in our initial experiment. The acidified sample (50 ppb Sb(III) or Sb(V) standard in 2% HCl matrix) was adjusted from pH 1 to 8 by addition of HCl or NaOH solution, before it was run on the SHG-ICP-MS system (Fig. 1). Fig. 2 shows the effect of pH on the reduction yields of 50 μg L⁻¹ of Sb(III) and Sb(V). Both Sb(III) and Sb(V) are reduced to SbH₃ in highly acidic solutions. However, Sb(V) does not generate the hydride above pH 5.0, while Sb(III) does. Thus, Sb(III) can be determined under mild pH conditions (i.e., 5–8), while total inorganic Sb is determined after a pre-reduction procedure. However, direct determination of trace Sb species in drinking water remains challenging because the sensitivity for Sb(III) is decreased ∼50% at pH = 5–6 (Fig. 2). Additionally, there is a possibility of the transformation between Sb(III) and Sb(V) in this situation.

Fig. 2 Relationship between the pH values and the reduction yield of Sb species. The solution pH values of 50 μg L⁻¹ Sb(III) or Sb(V) standard in 2% HCl matrix was adjusted from pH 1 to 8 by addition of a NaOH or HCl solution.

The use of citric acid to generate stable Sb-complexes could be a viable strategy to improve the pH-dependent SHG-ICP-MS method. In order to evaluate this possibility, different concentrations of citric acid were added to the acidified water sample (50 ppb Sb(III) or Sb(V) standard in 2% HCl matrix) before it was run on the HG-ICP-MS system (Fig. 1). No additional NaOH or HCl was added to the samples during these experiments. Fig. 3 shows the effects of citric acid concentration on the intensities of Sb(III) and Sb(V) at 10 μg L⁻¹ each. As shown in Fig. 3, the clearest discrimination of Sb(III) and Sb(V) is observed when the concentration of citric acid is more than 0.5%. At this concentration (0.5% citric acid), the Sb(V) signal approaches the LOD and the Sb(III) signal remains at maximum intensity. Compared with the conventional pH-dependent method (Fig. 2), this improved SHG method achieves excellent sensitivity, approximately 2-fold that of the conventional pH-dependent HG method (Fig. 3).

Fig. 3 Effect of the citric acid concentration on the intensity of Sb species. 50 μg L⁻¹ Sb(III) or Sb(V) in 2% HCl matrix was contained different concentrations of citric acid.

Some researchers have investigated these Sb-complexes by HPLC-ICP-MS. For example, Ulrich et al.³² reported the complexation of Sb(III) with citric acid based on chromatographic retention time changes, and observed no Sb(V) complexes. In contrast, Guy et al.³³ demonstrated complexation between Sb(V) and citric acid, while no Sb(III) complex was observed. Recently, by electrospray (ES)-MS/MS and HPLC-ICP-MS, Hansen and Pergantis³⁴ found that the stable product of the reaction of Sb(V) with citric acid was a 1 : 1 Sb(V)–citrate complex. In the current work, we examined the complicated interactions of Sb species with citric acid and the subsequent reactions of their products with reductant NaBH₄ by a quadrupole-Orbitrap high-resolution electron spray ionisation (EIS)-MS. Solutions of Sb(III) and Sb(V) (10 μg mL⁻¹ each) were mixed with 0.5% citric acid, and then divided into two portions, one with and one without 2% NaBH₄ to simulate the SHG procedure. These solutions were diluted with 99.9% MeOH (1 + 1) prior to ESI-MS measurement. As shown in Fig. 4, the obtained ESI-MS spectrum is characterized by the base peak at m/z 191.019, which is attributed to the excess deprotonated citrate. Fig. 4a and b show the mass spectra of the reaction products without NaBH₄. Peaks resulting from Sb species occurred at m/z 500.928 and 502.928 (Fig. 4a) for the Sb(III)/citric acid solution (attributed to [Sb(C₆O₇H₆)₂]⁻), and at m/z 360.917 and 362.917 (Fig. 4b) for the Sb(V)/citric acid solution (attributed to [Sb(OH)₃(C₆O₇H₅)]⁻). The signal intensity of the Sb(III)-complex is almost the same as that of the Sb(V) complex, which indicates that the complexation efficiency of Sb(III) is similar with Sb(V) (Fig. 4a and b). To further explore the selectivity of the SHG procedure, three different concentrations (1, 10, and 100 μg mL⁻¹) of Sb(III) and Sb(V) in 0.5% citric acid were monitored by ESI-MS with and without the reductant 2% NaBH₄. As shown in Table 2, there is an obvious linearity between the concentration of Sb(V) and the signal intensity for the Sb(V)-complex ([Sb(OH)₃(C₆O₇H₅)]⁻), irrespective of the presence of 2% NaBH₄. In contrast, the signal intensities for the Sb(III) complex ([Sb(C₆O₇H₆)₂]⁻) are almost ignored (<10⁴) after addition of the NaBH₄, although there is a linearity between the content of Sb(III) and the Sb(III) complex ([Sb(C₆O₇H₆)₂]⁻) in the absence of 2% NaBH₄. It can be verified that the significant difference in the Sb(III)-complex intensities with or without NaBH₄ may be due to the complete reduction of the Sb(III) complex by NaBH₄ (Table 2 and Fig. 4c and d). In conclusion, the selectivity of the SHG-ICP-MS with citric acid can be rationalized by the fact that nearly all of the Sb(V) or Sb(III) reacts with citric acid, and the resultant Sb(V)-complex ([Sb(OH)₃(C₆O₇H₅)]⁻) cannot be reduced by NaBH₄, while the resultant Sb(III)-complex ([Sb(C₆O₇H₆)₂]⁻) originating from Sb(III) can be reduced thoroughly by NaBH₄.

Fig. 4 ESI-MS spectra of the Sb/citrate complex in negative mode: (a) Sb(III)-complex without NaBH₄; (b) Sb(V)-complex without NaBH₄; (c) Sb(III)-complex with NaBH₄; (d) Sb(V)-complex with NaBH₄.

Table 2 Gradient test: intensity of the complexes (N = 3)

Identified targets by ESI-MS		Sb(III)				Sb(V)
		1 μg mL^−1	10 μg mL^−1	100 μg mL^−1	1 μg mL^−1	10 μg mL^−1	100 μg mL^−1
a Sb(III)-complex ([Sb(C6O7H6)2]^−) was monitored at m/z 500.928; Sb(V)-complex ([Sb(OH)3(C6O7H5)]^−) was monitored at m/z 360.917.b Citric acid was monitored at the m/z of 191.019.
Sb-complex^a	Without NaBH4	2.30 ± 0.05 × 10^6	17.0 ± 0.4 × 10^6	150 ± 2 × 10^6	3.0 ± 0.1 × 10^6	25.0 ± 0.5 × 10^6	220 ± 6 × 10^6
	With NaBH4	<10^4	<10^4	<10^4	3.0 ± 0.2 × 10^6	23.0 ± 0.4 × 10^6	200 ± 7 × 10^6
Citric acid^b	Without NaBH4	1.7 ± 0.1 × 10^9	1.7 ± 0.2 × 10^9	1.4 ± 0.1 × 10^9	1.8 ± 0.1 × 10^9	1.6 ± 0.2 × 10^9	1.4 ± 0.1 × 10^9
	With NaBH4	0.9 ± 0.2 × 10^9	0.9 ± 0.1 × 10^9	0.8 ± 0.2 × 10^9	0.9 ± 0.1 × 10^9	0.9 ± 0.2 × 10^9	0.8 ± 0.1 × 10^9

---

Improvements in the measurement precision by an IS method

Santos's group²⁶ reported poor signal stability (RSDs 16–17%) when using the SHG-ICP-OES method for Sb species determination. Musil et al.³⁵ also observed a 40% signal difference for arsenic within three hours in SHG-ICP-MS analysis. They claimed the relative instability of the SHG method may be due to the dry plasma only introducing AsH₃.³⁵ In this study, we attempted to use an IS method to improve the measurement precision in SHG-ICP-MS. We evaluated three potential elements (Ge, Sn, and Te), all of which have similar efficiencies of hydride generation, transportation, and atomization to those of Sb. The IS element was added online into the acidified sample by the same pump to ensure a good homogeneity. As shown in Fig. 5, the RSD of the Sb signal is improved from 4.7 to 2.8% when Sn is used as the IS. Therefore, Sn was selected as the IS for our SHG-ICP-MS method.

Fig. 5 Signal recoveries corrected by different internal standards.

Increasing the Sb signal by addition of CH₄ to ICP plasma

Because of its high first ionization energy (8.61 eV), only some of the Sb atoms can be ionized by ICP. Hu et al. successfully used 3% (v/v) ethanol to improve the Sb signal in ICP-MS analysis.²⁷ In our study, CH₄ gas was used as the carbon source by directly introducing it to the spray chamber. Signal intensity for ¹²¹Sb⁺ (after blank correction) as a function of CH₄ gas flow rate is illustrated in Fig. 6. The results show that 2.0 mL min⁻¹ CH₄ leads to a 2.1-fold improvement in the maximum sensitivity for ¹²¹Sb⁺ compared to that without CH₄. This signal enhancement phenomenon can be explained by a charge transfer reaction occurring between the positively charged carbon species and the high ionization energy element in the central channel of the plasma.^36–38

Fig. 6 Effect of CH₄ gas flow rate on the signal intensity of Sb. The concentration of Sb(III) was 1.0 μg L⁻¹.

Optimization of SHG-ICP-MS operation conditions

Some important operational parameters of the SHG-ICP-MS method need to be optimized in order to obtain the best analytical performance. These parameters include the concentration of NaBH₄ and HCl for HG, the Ar carrier gas flow rate, the radiofrequency (RF) power of the ICP-MS, and the concentration of KI in the pre-reduction of Sb(V).

An optimum concentration of NaBH₄ is necessary to obtain rapid and quantitative conversion of Sb(III) to SbH₃. As shown in Fig. 7a, the analytical signal is enhanced by NaBH₄ (w/w) concentrations in the range of 0.1–0.5%, and the optimum concentration is 0.5%. Higher concentrations of NaBH₄ may produce two much hydrogen, diluting the SbH₃ and decreasing the stability of the plasma, resulting in lower sensitivity and poor reproducibility. The formation of SbH₃ by the reaction of NaBH₄ requires HCl medium. As shown in Fig. 7b, an HCl concentration of 2% (v/v) provides the highest signal intensity, and thus this concentration was selected for all subsequent experiments.

Fig. 7 Optimization of SHG-ICP-MS operation parameters: (a) effect of NaBH₄ concentration on Sb signal; (b) effect of HCl concentration on Sb signal; (c) effect of carrier gas flow rate on Sb signal; (d) effect of RF power on Sb signal. The concentration of Sb(III) was 1.0 μg L⁻¹.

The Ar carrier gas flow rate and RF power are two important parameters for the ICP because they significantly affect the Sb signal intensity. When the optimum values (1.15 L min⁻¹ for Ar flow rate and 1550 W for RF) are used, the best transport efficiency for the SbH₃ gas and ionization of Sb⁺ in plasma are obtained (Fig. 7c and d). The concentration of Sb(V) is obtained by subtracting the Sb(III) concentration from the total inorganic Sb. KI was used as the pre-reduction reagent to quantitatively convert Sb(V) to Sb(III) for the determination of total inorganic Sb. As shown in Fig. 8, the conversion efficiency for Sb(V) to Sb(III) reaches 98% using 6% KI as the reduction medium in the pre-reduction procedure. The optimized conditions for this work are summarized in Table 1.

Fig. 8 Effect of KI concentration on the efficiency of Sb(V) reduction. The concentration of Sb(V) was 50 μg L⁻¹.

Potential interferences that could affect Sb(III) determination when using the hydride generation method include transition metals (i.e., Cu⁺, Ni²⁺ and Fe³⁺) and Sb(V).^24,25 The effects of the presence of these four mixed ions on Sb(III) recovery in the proposed SHG-ICP-MS (with 0.5% citric acid) system were examined here. The results indicated the selected ions (i.e. 10 000-fold of Cu⁺, Ni²⁺, Fe³⁺ and Sb(V)) did not interfere with the determination of 1 μg L⁻¹ Sb(III) (recovery > 97%). The high interference tolerance of the proposed method is an advantage of the proposed SHG-ICP-MS method for the determination of trace Sb(III) or total Sb analysis in natural drinking waters.

Analytical performance and drinking water analysis

Six standard solutions (0, 0.01, 0.1, 0.5, 5.0 and 25 μg L⁻¹ Sb(III) or total Sb(Sb(III) : Sb(V) = 1 : 1)) were analyzed to construct a calibration curve with a correlation coefficient of 0.9995 for Sb(III) and 0.9996 for total Sb. The LODs for our method are 4.0 ng L⁻¹ for Sb(III) and 6.0 ng L⁻¹ for total Sb, with RSDs for 1.0 μg L⁻¹ of Sb(III) and 1.0 μg L⁻¹ of total Sb of 2.2 and 2.0%, respectively. A comparison of the analytical performance obtained by this method with those of other chromatographic or non-chromatographic SHG approaches^4,26,39–42 for speciation of inorganic Sb reported in the literature is shown in Table 3. The superior LOD, stability, and throughput all indicate that this simple method has great potential for direct determination of ng L⁻¹ levels of inorganic Sb species in drinking water. Table 4 shows the recovery results from spiked well water samples. The recoveries for Sb(III) and total Sb vary from 93.1 to 108% and from 98.5 to 103%, respectively.

Table 3 Recovery of Sb species in a spiked well water (n = 3)

Sample	Sb(III)		Sb(T)^a
	Value, μg L^−1	Rec. (%)	Value, μg L^−1	Rec. (%)
a Sb(T): total inorganic Sb.
Water	1.46 ± 0.08		3.95 ± 0.21
Water + 1 ppb Sb(III) + 1 ppb Sb(V)	2.39 ± 0.22	93.1	5.92 ± 0.16	98.5
Water + 10 ppb Sb(III) + 10 ppb Sb(V)	12.1 ± 0.9	106	23.8 ± 0.9	99.2
Water + 1 ppb Sb(III) + 10 ppb Sb(V)	2.54 ± 0.18	108	15.3 ± 0.7	103
Water + 10 ppb Sb(III) + 1 ppb Sb(V)	11.2 ± 0.7	97.4	15.1 ± 0.6	101

---

Table 4 Comparison of the reported method with procedures in the literature for determination of inorganic Sb species

Method^a	Sb(III)		Sb(V) or Sb(T)^b
	LOD, ng L^−1	RSD, %	LOD, ng L^−1	RSD, %
a Abbreviations for method: HPLC-HG-AFS (high performance liquid chromatography-hydride generation-atomic fluorescence spectrometry); HPLC-ICP-MS (high performance liquid chromatography-inductively coupled plasma-mass spectrometry); HPLC-HG-ICP-MS (high performance liquid chromatography-hydride generation-inductively coupled plasma-mass spectrometry); SHG-ICP-OES (selective hydride generation-inductively coupled plasma-optical emission spectroscopy); SHG-AFS (selective hydride generation-atomic fluorescence spectrometry) and SHG-ICP-MS (selective hydride generation-inductively coupled plasma-mass spectrometry).b Sb(T): total inorganic Sb.c No available.
HPLC-HG-AFS^39	190	6.1	180	5.8
HPLC-ICP-MS^4	83	2.4	38	2.1
HPLC-HG-ICP-MS^41	40	7.0	8	6.0
SHG-ICP-OES^26	50	17	110	18
SHG-ICP-OES^42	1200	—^c	4500	—
SHG-AFS^40	10	5.2	13	5.0
SHG-ICP-MS (this method)	4	2.2	6	2.0

---

The proposed method was applied to the determination of Sb species in ten bottled waters, ten tap waters, and forty well waters. As shown in Fig. 9, the maximum Sb(III) and total Sb contents are 1.32 and 3.85 μg L⁻¹, respectively, for all the samples. The average Sb(III) and total Sb contents in bottled water (N = 10) are 0.032 and 0.320 μg L⁻¹, respectively. The average Sb(III) and total Sb contents in tap water (N = 10) are 0.197 and 1.05 μg L⁻¹, respectively, and the average Sb(III) and total Sb contents in well water (N = 40) are 0.211 and 1.93 μg L⁻¹, respectively. Most of the natural drinking water samples (>90%) contain low ratios (<30%) of Sb(III)/[Sb(III) + Sb(V)]. Our results for total Sb (average 1.48 μg L⁻¹) and Sb(III) (average 0.175 μg L⁻¹) are within the range of results previously published by other researchers for drinking water from Europe, Canada, Japan, Hong Kong, and Brazil (1.22–3.14 μg L⁻¹ for total Sb, 0–0.31 μg L⁻¹ for Sb(III)).^43,44

Fig. 9 Results of Sb(III) and total inorganic Sb contents for 60 drinking water samples.

Conclusions

A reliable method has been developed for direct determination of inorganic Sb species at ultra-trace levels in drinking water by SHG-ICP-MS using citric acid as the stabilizing reagent. ESI-MS analysis was used to rationalise the selectivity of the HG with citric acid. An IS technique using Sn and the introduction of CH₄ to the ICP improved the precision and sensitivity of the method. The results of this study strongly indicate that this high throughput, validated method can be used for routine in-lab determinations of ng L⁻¹ levels of inorganic Sb species in drinking water.

Acknowledgements

This work was supported by the China Scholarship Council, the National Nature Science Foundation of China (No. 41521001 and No. 21207120), the Ministry of Science and Technology of China (No. 2014DFA20720), the Fundamental Research Funds for the Central Universities, China University of Geosciences (Wuhan) (No. CUGL140411), and the Research Program of State Key Laboratory of Biogeology and Environmental Geology of China (No. GBL11505).

References

1. M. Filella, N. Belzile and M.-C. Lett, Earth-Sci. Rev., 2007, 80, 195–217.
2. M. Filella, N. Belzile and Y.-W. Chen, Earth-Sci. Rev., 2002, 57, 125–176.
3. M. M. Silva Jr, D. J. Leao, I. T. Moreira, O. M. de Oliveira, A. F. de Souza Queiroz and S. L. Ferreira, Environ. Sci. Pollut. Res., 2015, 22, 8386–8391.
4. M. Marcinkowska, I. Komorowicz and D. Baralkiewicz, Anal. Chim. Acta, 2016, 920, 102–111.
5. X. W. Wang, X. K. Li, X. Zhang and S. H. Qian, J. Anal. At. Spectrom., 2014, 29, 1944–1948.
6. L. Fang, L. Wang, J. Li and Q. Gao, Anal. Lett., 2015, 48, 1954–1964.
7. N. Altunay and R. Gurkan, Food Chem., 2015, 175, 507–515.
8. I. López-García, R. E. Rivas and M. Hernández-Córdoba, Talanta, 2011, 86, 52–57.
9. W. Quiroz, H. Arias, M. Bravo, M. Pinto, M. G. Lobos and M. Cortes, Microchem. J., 2011, 97, 78–84.
10. H. R. Hansen and S. A. Pergantis, Anal. Chem., 2007, 79, 5304–5311.
11. J. Zheng, A. Iijima and N. Furuta, J. Anal. At. Spectrom., 2001, 16, 812–818.
12. E. A. Gonzalvez, S. Armenta, M. L. Cervera and M. de la Guardia, TrAC, Trends Anal. Chem., 2010, 29, 260–268.
13. L. A. Portugal, L. Ferrer, A. M. Serra, D. Gonçalves da Silva, S. L. C. Ferreira and V. Cerdà, J. Anal. At. Spectrom., 2015, 30, 1133–1141.
14. S. L. C. Ferreira, W. N. L. dos Santos, I. F. dos Santos, M. M. S. Junior, L. O. B. Silva, U. A. Barbosa, F. A. de santana and A. F. de S. Queiroz, Microchem. J., 2014, 114, 22–31.
15. J. Lu, J. Tian, W. Zhao and Z. Yang, Spectrochim. Acta, Part B, 2011, 66, 74–80.
16. A. Ramesh Kumar and P. Riyazuddin, Int. J. Environ. Anal. Chem., 2007, 87, 469–500.
17. S. L. Anderson and S. A. Pergantis, Talanta, 2003, 60, 821–830.
18. E. M. de M. Flores, E. P. dos Santos, J. S. Barin, R. Zanella, V. L. Dressler and C. F. Bittencourt, J. Anal. At. Spectrom., 2002, 17, 819–823.
19. M. A. Vieira, P. Grinberg, C. R. R. Bobeda, M. N. M. Reyes and R. C. Campos, Spectrochim. Acta, Part B, 2009, 64, 459–476.
20. M. B. de la Calle-Guntiñas, Y. Madrid and C. Cámara, Fresenius' J. Anal. Chem., 1992, 343, 597–599.
21. H. S. Ferreira, S. L. C. Ferreira, M. Luisa Cervera and M. de la Guardia, Spectrochim. Acta, Part B, 2009, 64, 597–600.
22. A. Gonzalvez, M. L. Cervera, S. Armenta and M. de la Guardia, Anal. Chim. Acta, 2009, 636, 129–157.
23. P. Cava-Montesinos, A. de la Guardia, C. Teutsch, M. L. Cervera and M. de la Guardia, Anal. Chim. Acta, 2003, 493, 195–203.
24. B. Mohammad, A. M. Ure, J. Reglinski and D. Littlejohn, Chem. Speciation Bioavailability, 1990, 3, 117–123.
25. M. O. Andreae, J. F. Asmode, P. Foster and L. Van't dack, Anal. Chem., 1981, 53, 1766–1771.
26. V. C. G. dos Santos, M. T. Grassi and G. Abate, Anal. Lett., 2015, 48, 2921–2935.
27. Z. C. Hu, S. H. Hu, S. Gao, Y. H. Liu and S. L. Lin, Spectrochim. Acta, Part B, 2004, 59, 1463–1470.
28. Q. Xu, W. Guo, L. L. Jin, Q. H. Guo and S. H. Hu, J. Anal. At. Spectrom., 2015, 30, 2010–2016.
29. Y. T. Li, W. Guo, Z. W. Wu, L. L. Jin, Y. Q. Ke, Q. H. Guo and S. H. Hu, Microchem. J., 2016, 126, 194–199.
30. W. Guo, Y. X. Wang, J. X. Li, Y. E. Peng, L. L. Jin, Q. H. Guo and S. H. Hu, At. Spectrosc., 2016, 37, 7–12.
31. Y. Yang, Y. Peng, Q. Chang, C. Dan, W. Guo and Y. Wang, Anal. Chem., 2016, 88, 1275–1280.
32. N. Ulrich, P. Schaked and D. Zilberstein, Fresenius' J. Anal. Chem., 2000, 368, 62–67.
33. A. Guy, P. Jones and S. L. Hill, Analyst, 1998, 123, 1513–1519.
34. H. R. Hansen and S. A. Pergantis, J. Anal. At. Spectrom., 2006, 21, 1240–1248.
35. S. Musil, A. H. Petursdottir, A. Raab, H. Gunnlaugsdottir, E. Krupp and J. Feldmann, Anal. Chem., 2014, 86, 993–999.
36. P. Allain, L. Jaunault, Y. Mauras, J. M. Mermet and T. Delaporte, Anal. Chem., 1991, 63, 1497–1498.
37. E. H. Larsen and S. Stürup, J. Anal. At. Spectrom., 1994, 9, 1099–1105.
38. W. Guo, S. H. Hu, J. Y. Zhang, L. L. Jin, X. J. Wang, Z. L. Zhu and H. F. Zhang, J. Anal. At. Spectrom., 2011, 26, 2076–2080.
39. W. Quiroz, H. Arias, M. Bravo, M. Pinto, M. G. Lobos and M. Cortés, Microchem. J., 2011, 97, 78–84.
40. X. Wang, X. Li, X. Zhang and S. Qian, J. Anal. At. Spectrom., 2014, 29, 1944–1948.
41. P. Smichowski, Y. Madrid, M. B. de La Calle Guntinas and C. Cámara, J. Anal. At. Spectrom., 1995, 10, 815–821.
42. Y. Li. Feng, H. Narasaki, H. Y. Chen and L. C. Tian, Anal. Chim. Acta, 1999, 386, 297–304.
43. R. Dabeka, H. Conacher, J. Lawrence, W. Newsome, A. McKenzie, H. Wagner, R. Chadha and K. Pepper, Food Addit. Contam., 2002, 19, 721–732.
44. H.-B. Hou and H. Narasaki, Anal. Sci., 1999, 15, 911–914.

---