Determination of total antimony and inorganic antimony species by hydride generation in situ trapping flame atomic absorption spectrometry: a new way to (ultra)trace speciation analysis

Abstract

The analytical performance of non-chromatographic coupled hydride generation, integrated atom trap (HG-IAT) atomizer flame absorption spectrometry (FAAS) systems were evaluated for the speciation analysis of antimony in environmental samples. Antimony, using formation of stibine (SbH₃) vapors were atomized in an air–acetylene flame-heated IAT. A new design of HG-IAT-FAAS hyphenated technique that would exceed the operational capabilities of existing arrangements (a water-cooled single silica tube, double-slotted quartz tube or an “integrated trap”) was investigated. For the estimation of Sb(III) and Sb(V) concentrations in samples, the difference between the analytical sensitivities of the absorbance signals obtained for antimony hydride without and with previous treatment of samples with L-cysteine can be used. The concentration of Sb(V) was calculated by the difference between total Sb_tot and Sb(III). A dramatic improvement in detection limit was achieved compared with that obtained using either of the atom trapping techniques, presented above, separately. This novel approach decreases the detection limit down to low pg mL^–1 levels. The concentration detection limit, defined as 3 times the blank standard deviation (3sigma), was 0.2 ng mL^–1. For a 120 s in situ pre-concentration time (sample volume of 2 mL), sensitivity enhancement compared to flame AAS, was 550 fold for Sb, using hydride generation-atom trapping technique. The sensitivity can be further improved by increasing the collection time. The precision, expressed by RSD, was 8.0% (n = 6) for Sb. The designs studied include slotted tube, water-cooled single silica tube and integrated atom traps. Reference and real sample materials were analyzed. The accuracy of the method was verified by the use of certified reference materials (NIST SRM 2704 Buffalo River Sediment, SRM 2710 Montana Soil, SRM 1633a Coal Fly Ash, SRM 1575 Pine Needles, SRM 1643e Trace Elements in Water) and by aqueous standard calibration technique (solutions). The measured Sb content, in reference materials, were in satisfactory agreement with the certified values. The hyphenated technique was applied for antimony determinations in soil, sediment, coal fly ash, sewage and river water.

---

Introduction

Antimony is a toxic element, which exists mainly in two oxidation states as Sb(III) and Sb(V) in environmental, biological and geochemical samples,¹ and may form various inorganic and organic species having different physicochemical and toxic properties. Inorganic compounds of antimony are more toxic than its organic forms. Toxicity of Sb(III) has shown to be 10 times higher than that of Sb(V). Therefore, the selective determination of antimony(III) and antimony(V) species has become important in environmental, biological, clinical, metallurgical and other industrial samples studies because of the implications in human health. Critical reviews^1–3 have focused on the methodologies for the determination at trace levels of both total and the individual chemical species of antimony in various samples. Of all the methods for the determination of antimony in various samples, atomic absorption spectrometry, especially after a hydride generation step, is the most useful and sensitive technique.

The flame has been developed over the past decades into a powerful tool for elemental analysis by atomic absorption spectrometry (FAAS). Despite its widespread use, FAAS remains limited in routine applications by normal sample introduction methods which are insufficiently sensitive to permit direct quantitative determination of trace or ultratrace concentrations of some elements, such as hydride forming elements. Therefore, in most cases a pre-concentration stage, before the elements can be determined, is essential.

It is well established that significant improvements in the limits of detection of flame atomic absorption spectrometry (FAAS) and graphite furnace atomic absorption spectrometry (GF-AAS) may be achieved by chemical vapor generation, mostly hydride generation (HG) and in-atomizer (silica or graphite tube) trapping (pre-concentration) as reviewed^4,5 for the determination of As, Bi, Ge, In, Pb, Sb, Se, Sn, Te and Tl (as well as Hg).

Hydride generation combined with AAS detection (HG-AAS) is one of the most powerful analytical methods for the determination of total Sb. Initially, AAS using different atomization devices was the most employed technique.^6–10 Rondón et al.⁶ described the continuous flow hydride generation system AAS for the determination of Sb using a T-quartz cell tube heated by a conventional 10 cm path length air–acetylene flame. Krachler et al.⁷ suggested a flow injection method for the determination of trace amount of Sb by HG-AAS using a quartz tube atomizer. Petit de Penaet al.⁸ studied the performance of flow injection hydride generation system AAS in the determination of Sb in liver tissue and whole blood. Calixto de Campos et al.⁹ developed a method for the determination of Sb by HG-AAS using an electrically heated quartz cell. Marlon de Moraes Flores et al.¹⁰ described the generation of Sb(III) with atomization in a flame AAS using a quartz cell atom trap.

One way to determine antimony species in samples is to use a chromatographic speciation method. Hyphenated techniques using chromatographic separation coupled to atomic spectrometers have been widely utilized in the speciation of antimony in different matrices.¹¹ A possible non-chromatographic alternative method is HG-AAS; antimony speciation analysis is based on the slow kinetics of hydride formation from the pentavalent state.^12,13 In recent years, hydride generation has also been employed to selectively measure antimony species in natural waters, liver tissue, whole blood, samples of injectable drugs, etc.,^6,8,10,14 as different antimony species have different hydride generation activity, e.g., inorganic Sb(III) species may form hydrides while Sb(V) does not when reacting with NaBH₄.

Unfortunately, none of these studies reported a successful HG-AAS determination of Sb at the ultratrace levels in real samples. In addition, the reported limits of detection are not so low as those currently achievable by the use of an in situ atom trapping technique.

In situ trapping permits a significant enhancement in sensitivity for batch and continuous hydride generation approaches used in the ultra-trace determination of volatile hydride species. Due to its importance, in situ trapping, which allows the coupling of hydride generation to integrated atom trap (IAT),^15–19 was chosen for this study.

An analytical system was developed to trap and pre-concentrate Bi from the vapor phase stream. Bismuthine formed by sodium tetrahydroborate reduction was trapped on a tungsten coil previously heated to 270 °C. The analyte species were re-volatilized by increasing the coil temperature to 1200 °C and then transported to an externally heated silica T-tube by using a mixture of argon and hydrogen as the carrier gas.²⁰ Korkmaz et al.²¹ investigated the nature of re-volatilization from atom trap surfaces in flame by AAS. Analytes Au, Bi, Cd, Mn and Pb were trapped on a water-cooled, U-shaped silica trap or a slotted silica tube trap and re-volatilized by organic solvent aspiration. They concluded that, although heating was not necessarily associated with re-volatilization, direct contact between the flame and the active silica surface was required. Recently,²² the analytical performance of three trap systems (water-cooled U-shaped silica trap, water-cooled U-shaped silica trap combined with slotted silica tube and slotted tube trap) for flame AAS were evaluated for determination of Cd and Pb in waters. Guo and Guo²³ reported SeH₂ collection at gold wire heated to 200 °C situated in a quartz tube atomizer (atomic fluorescence spectrometry (AFS) or AAS detection) with a separate inlet for argon. A successful trapping of PbH₄ in a bare quartz tube was announced by Korkmaz et al.,²⁴ who also suggested that the same trap could be used also for other hydrides. Recently, a preliminary evaluation of a quartz tube trap for collection of SbH₃ and for volatilization of trapped analyte with subsequent atomization in a multiple microflame quartz tube atomizer for AAS was presented.²⁵ Kratzer and Dědina extended their investigation of stibine trapping in quartz tube traps²⁶ to stibine collection (and subsequent analyte atomization) in conventional quartz tube atomizers.²⁵ They employed the simplest possible experimental arrangement; just the commercially available externally heated quartz tube atomizer without any trap or additional heating device. Further, a modification of the externally heated quartz tube atomizer, making possible in situ trapping of bismuthine and subsequent analyte atomization for Bi was described.²⁷ Krejči et al.²⁸ investigated the collection of Sb and Bi on a molybdenum foil strip situated in a laboratory-made quartz T-tube, single-slot burner head, following hydride generation. Recently,²⁹ the analytical performance of a miniature quartz trap coupled with electrochemical hydride generator for antimony determination was described. A portion of the inlet arm of the conventional quartz tube atomizer was used as an integrated trap medium for on-line pre-concentration of electrochemically generated hydrides. Very recently,³⁰ a novel quartz device has been designed to trap arsine and selenium hydride and subsequently to volatilize the collected analyte and atomize it for AAS detection. The device is actually the multiple microflame quartz-tube atomizer with inlet arm modified to serve as the trap and to accommodate the oxygen-delivery capillary used to combust hydrogen during the trapping step.

This work has been mainly aimed at improving analytical performance of antimony conventional hydride generation flow system by combining this approach for in situ trapping technique with an integrated atom trap system for flame AAS that is applicable to determination of Sb in real and reference materials. In this work, the non-chromatographic speciation HG-IAT-FAAS method was developed to selectively measure the total and inorganic forms of antimony species exclusive of organic antimony (to distinguish between inorganic Sb(III)- and Sb(V)-species) in samples. The hydride generation technique brings the FAAS method closer to the detection limits of HG-GF-AAS (in situ trapping technique).^31–34 Finally, in order to check the accuracy of the proposed system, analyses of standard reference materials were performed.

Experimental

Spectrometer

A Carl Zeiss Jena (Jena, Germany) Model AAS3 flame atomic absorption spectrometer (components described in ref. 35) equipped with a 10 cm air–acetylene burner head assembly and an IBM-PC compatible computer was used throughout the study. The sampling rate for the PMT signal was 10 Hz. Signals were processed with in-house software (Turbo Pascal Version 7) to extract the transient peak heights, area and peak time. NARVA Sb hollow cathode lamp was used as the radiation source. No background correction was required in this mode of operation. Operating parameters of the AAS instrument are summarized in Table 1 after appropriate optimization.

Table 1 Instrumental operating conditions for determination of antimony by IAT-FAAS

Parameter	Setting
a Nebulizer uptake rate, 5 mL min^–1. b Air flow rate 475 L h^–1, acetylene flow rate 50 L h^–1 (fuel-lean flame); 10 cm slot burner.
Wavelength/nm	217.6
Spectralband pass/nm	0.2
Lamp current/mA	7
Flame type^a	Air–C2H2
Flame conditions^b	Lean
Silica tube obscuration (%)	ca. 30
Coolant water/L min^–1	ca. 2
Window of measurements/s	15
Read time/s	7
Smoothing	No
Signal measurement	AA (Peak area)

---

Hydride generation systems

Hydride generation was accomplished in the batch (discontinuous) mode and in a continuous mode using a manually controlled two channel peristaltic pump (Gilson, Model Minipuls 3, France). A mass flow controller with a precision pressure gauge (Models ERG 500 and ERG 2000, power supply Model ERG 2M, DHN, Warsaw, Poland) were used to regulate the purge and transfer gas flow rates accurately and reproducibly.

Even though the hydride generation cells served as gas–liquid separators, in an attempt to completely reduce water vapor, a quartz gas–liquid separator identical to that described earlier³⁶ was installed between the chemifolds and the nebulizer/spray chamber.

Batch hydride generation. Analyte hydrides were generated in a laboratory-made Pyrex cell (batch system, the total volume of the glass cell was about 80 mL) identical to the one described earlier,³⁷ into which NaBH₄ was introduced with a peristaltic pump. The PVC/PTFE pump tubing (3.9 mm i.d. × 40 cm) was fitted to the outlet of the batch system. The internal purge gas supply line to the nebulizer/burner was routed through into the side of the hydride cell and through the fine glass frit. The outlet of the glass cell and the nebulizer/burner system were connected, for practical reasons, by a piece of PVC/PTFE tubing of dimensions 1.5 mm i.d. × 2.0 mm o.d. having a length of 30 cm. The evolved hydrides were stripped from the solution and swept into the v nebulizer/burner–IAT system with an air purge gas. The cell assembly and the sequence of operations used to generate and trap the hydrides have been described in detail in a previous paper.³⁷

Continuous-flow hydride generation. Hydride generation was accomplished in the continuous-flow mode using a system similar to that described previously.³⁸ This system consists of a manually controlled peristaltic pump (Gilson, Minipuls-3), a gas handling network and a reaction cell (volume 40 mL). The PVC/PTFE pump tubing was fitted to the outlet of the continuous system. The internal purge gas supply line to the nebulizer/burner was routed through the inside part of the hydride cell. The outlet of the glass cell and the nebulizer/burner system were connected with a PVC/PTFE tubing of dimensions 1.5 mm i.d. × 2.0 mm o.d. having a length of 30 cm.

The operating conditions for batch and continuous-flow hydride generation atomic absorption spectrometry are summarized in Table 2. A schematic diagram of the batch and continuous-flow HG-IAT-FAAS system with in situ pre-concentration in the IAT unit is shown in Fig. 1.

Fig. 1 Schematic diagram of the HG-IAT-FAAS system.

Table 2 Optimized operating conditions for determination of Sb_tot and Sb(III) by HG-IAT-FAAS

Parameter	Batch hydride generation		Continuous-flow hydride generation
	Sbtot	Sb(III)	Sbtot	Sb(III)
a In 0.5% (m/v) NaOH solution. b Ref. 39
NaBH4 concentration^a (% (m/v))	1.0	1.0	1.0	2.0
NaBH4 solution flow rate/mL min^–1	2.5	2.5	1.0	1.0
HCl concentration/mol L^–1	2.0	—	3.0	—
Oxalic acid concentration (% (m/v))	—	5.0	—	6.0
L-cysteine concentration (% (m/v))	5.0	—	4.0	—
Reduction time (L-cysteine)/min	30	—	30	—
Sample flow rate/mL min^–1	—	—	1.0	1.0
Sample volume/mL	10	10	2	2
Carrier air flow rate/mL min^–1	75	75	100	100
Gas–liquid separator	As described in the literature.^34
PVC/PTFE peristaltic tubes		40 cm × 1.9 mm i.d. × 3.9 mm o.d.
Trapping time/s		120
Trapping temperature/°C		<100 (373 K)
Atomization temperature^b/°C		ca. 1330 (1600 K)

---

Atom trapping techniques

Three designs of atom trap were investigated. Since the atom trap system (trapping medium) was described in detail in previous papers^15,39 this will not be discussed again here, but briefly summarized only.

A double-slotted quartz tube atom trap (STAT) was installed on a standard 10 cm air–acetylene burner. The design permits changeover from analysis with the IAT¹⁵ to that for a conventional flame in a few seconds.

A water-cooled single silica tube atom trap (WCAT) was arranged, as previously described,¹⁵ and mounted on a 10 cm burner in such a manner in order to permit the system to be vertically and laterally adjusted to the flame. The tubes were of an o.d. 3 mm and an i.d. of 1 mm for water cooling.

An integrated atom trap (IAT) was designed and constructed in this laboratory,¹³ it consisted of a combination of a WCAT and a STAT (see Fig. 1 in ref. 15). The IAT system was mounted over an air–acetylene burner on a mounting bracket, which permitted calibrated movement both vertically and horizontally.

A modified cooling system was used for cooling the water¹⁶ (see Fig. 1 in ref. 16). The continuously flowing cooling water kept the surface of the silica tube at the temperature below 100 °C. This allowed the analyte atoms to condense on the surface of the tube.

The single silica and slotted quartz tubes were coated with lanthanum to prevent devitrification and to improve the silica surface properties (increasing the surface area) by continuous aspiration via the burner nebulizer of 1.0% lanthanum solution, for 15 min. The tubes were re-coated after approximately 100–200 runs.

Gases and reagents

Compressed air gas of N-50 purity (99.999%) obtained from BOC GAZY (Poznań, Poland) was employed as the carrier gas for the nebulizer/burner unit without further purification. Compressed medical purity acetylene (Cezal, Poznań, Poland) was used as the source of the air–acetylene flame.

Standard solutions of Sb(III) were prepared from a 1000 mg mL^–1Sb(III) atomic absorption standards (Titrisol grade, Merck, Darmstadt, Germany). Standard solutions of Sb(V) were prepared from potassium hexahydroxoantimonate(V) (Riedel-de-Haën). All working standard solutions of Sb(III) and Sb(V) were prepared daily to prevent any possible species change, by diluting appropriate aliquots of the stock solution in high-purity water.

Sodium tetrahydroborate(III) and potassium tetrahydroborate(III), used as reducing solutions, were prepared daily, or more frequently if required, by dissolving proper amounts of NaBH₄ and KBH₄ (pellets) (Alfa Inorganics, Ward Hill, USA) in high-purity water and stabilized with 0.1% (m/v) NaOH (Suprapur, Merck, Darmstadt, Germany) solution to decrease its rate of decomposition, and was used without filtration.

Thiourea was prepared by dissolving the powder (Fluka, Buchs, Switzerland) in water to yield a 1 mol L^–1 solution.

A 5% (m/v) potassium iodide (Merck) and a 5% (m/v) ascorbic acid (Fluka) solution were prepared in water.

Oxalic acid (Fluka) was prepared by dissolving the powder in water to yield a 10% (m/v) solution.

A 9% (m/v) L-cysteine (Merck) solution was prepared in water as a pre-reductor agent.

Citric and picolinic acids (Fluka) and sodium dithionite (Merck) were prepared fresh daily.

The lanthanum chloride solution, used to coat the quartz tubes (1.0% (m/v)) was prepared from a 10% (m/v) lanthanum chloride solution supplied by Alfa Inorganics (Ward Hill, MA, USA) as a releasing agent for the use in atomic absorption.

All mineral acids (HNO₃, HCl, HF) and hydrogen peroxide 30% (v/v) of the highest quality (Suprapur, Merck, Darmstadt, Germany) were used. High-purity water: deionized water (model DEMIWA 5 ROSA, Watek, Czech Republic), and doubly distilled water (quartz apparatus, Bi18, Heraeus, Hanau, Germany) were used throughout the experiments.

Reference materials and samples

Validation of the method described in this work was performed using five certified reference materials. The following materials were chosen: SRM 2704 (Buffalo River Sediment), SRM 2710 (Montana Soil), SRM 1633a (Coal Fly Ash), SRM 1575 (Pine Needles) and SRM 1643e (Trace Elements in Water) supplied by NIST (USA). The certified reference values are available for antimony for assessment of the method accuracy. All solid reference materials were used as bottled, without further grinding and sieving.

The following samples were used in this study: the untreated waste water and coal fly ash was sampled from Poznań Coal-fired Power Plant in Poland, soil from Silesia, sediment from Dębina lake in Poznań (Poland) and water from Warta river (Poznań, Poland).

To ensure homogeneity, it was necessary to grind the real, solid samples. This was achieved without difficulty in an agate pestle and mortar by manual grinding for coal fly ash and soil, and by a vibrational mixer mill Model S (Testchem, Pszów, Poland) equipped with 30 mL grinding chamber and rod (6 cm diameter), all made of tungsten carbide.

Microwave digestion system

A laboratory-built prototype of a high pressure–temperature focused microwave heated digestion system, equipped with a closed TFM-PTFE vessel (30 mL internal volume) based on a design outlined in detail by Matusiewicz⁴⁰ was employed for wet-pressure sample digestion.

Method development

The whole analytical procedure consists of various steps. It includes: (1) closed wet digestion of the samples, (2) generation of the SbH₃ vapors and its in situ trapping (collection) in an IAT system, (3) flame atomization of collected hydrides and (4) measurement by FAAS.

Microwave-assisted high pressure Teflon bomb digestion. Preparation of all standards and digestions of all samples were conducted under typical laboratory conditions. The microwave-assisted pressurized digestion technique used for biological and environmental samples has been described previously.⁴⁰

Approximately 300 mg of powdered inorganic reference materials and samples (Soil, River Sediment, Coal Fly Ash) were placed in the TFM-PTFE vessel (“bomb”) of the microwave digestion system and moistened by 1 mL of 30% H₂O₂; then 3 mL of concentrated HNO₃ and 1 mL of concentrated HF were added. The samples were heated for 15 min at 150 W. After dissolution, the clear digested solution was transferred into 10 mL calibrated flask and diluted to volume with water. When working with organic material (Pine Needles), approximately 300 mg of sample was first moistened by 1 mL of 30% H₂O₂, then 3 mL of concentrated HNO₃ was used. The sample was heated for 10 min at 100 W. Before further analysis these were appropriately diluted depending on the concentration level of the elements. In all cases, a corresponding blank was also prepared according to the above microwave-assisted digestion procedure.

Simplex optimization procedure. A simplex optimization approach was undertaken to establish, for antimony, the best conditions for volatile species generation, transport and atomization. The parameters optimized are listed in Table 3, along with the ranges over which optimization experiments were possible and conducted. In practice, the ranges were judiciously selected for each parameter in turn, taking into account the practical problems of maintaining a stable absorbance signal.

Table 3 Optimum operating conditions for HG-IAT-FAAS measurement of Sb obtained by simplex and univariate methods^a

Parameter (variable)	Boundary limits of parameters, range	Continuous-flow system		Batch mode
		Univariate method	Simplex method	Univariate method	Simplex method
a Response, peak area of the antimony absorption intensity of 217.6 nm.
HCl/mol L^–1	0.5–3.5	3.0	2.7	2.0	2.4
NaBH4 (% (m/v))	0.5–5.0	1.0	2.0	1.0	1.5
L-cysteine (% (m/v))	1.0–6.0	4.0	3.5	5.0	4.5
Oxalic acid (% (m/v))	0.5–7.0	6.0	6.4	5.0	5.2
Air carrier flow rate/mL min^–1	50–150	100	108	75	85
Sample and reductant uptake rate/mL min^–1	0.5–5.0	1.0	1.5	1.0	1.3

---

Simplex optimization experiments were performed using a software package obtained from the University of Plymouth. The optimization was carried out using aqueous standard solutions of element (antimony) determined. The net S/B ratio was taken as the criterion of merit. Some preliminary univariate experiments (searches) were performed prior to the simplex optimization in order to establish the boundaries of the values of each parameter. Three measurements for each variable were conducted at the factor of interest. Between each experiment, a blank corrective experiment was run to ensure stable and repeatable results.

The optimum conditions obtained from this procedure were then used to run standard antimony solutions and quantify the antimony present in the samples.

Hydride generation and procedure for in situ trapping. The procedures for hydride generation of samples from trapping to atomization are outlined below and were not conducted in a clean laboratory environment. Hydride generation was accomplished using two different generators, the continuous unit and the batch system (cell), the application of both generators has been described in our previous paper.³⁸ In brief, the hydrides were generated continuously or in batch mode and were introduced into the IAT system by the carrier gas (air) during the vapor-trapping step of the atomizer temperature program only; this step was 120 s in duration in all experiments.

Total antimony. After preparation of solid samples, all the antimony present in the sample will be obtained in solution in the +5 oxidation state (all species will be converted to Sb(V) as the main oxidation product). For the accurate determination of the total inorganic antimony content in a sample by the HG technique, it is necessary to reduce the Sb(V) to Sb(III). Determination of the total concentration of Sb was performed after a complete reduction of Sb(V) to Sb(III) by adding L-cysteine solution (1 mL) directly to the digested sample and the mixture was stored for 30 min before the hydride generation procedure was carried out. This allowed an ensuing determination of the total content of antimony in the sample [Sb(III) + Sb(V)] as Sb(III)viahydride generation (ensures complete reaction to Sb(III)).

Continuous-flow generation measurements of volatile antimony hydrides (SbH₃) were studied using the system shown schematically in Fig. 1. A 2 mL aliquot sample: 0.6 mL volume of 32% HCl solution, 0.8 mL volume of 9% (m/v) L-cysteine solution and 0.6 mL volume of water, were placed in a quartz vessel. The mixture was stored for 30 min before the hydride generation procedure was carried out. This allowed determination of the total content of antimony in the sample [Sb(III) + Sb(V)] as Sb(III)viahydride generation (ensures complete reaction to Sb(III)). The PVC/PTFE transfer line from the reaction cell was placed in the nebulizer/burner system. The Sb sample was being continuously introduced at a rate of 1 mL min^–1 to merge with a 1.0% (m/v) solution of NaBH₄ (flow rate 1.0 mL min^–1). The merging solution feeds the gas–liquid separator to the IAT system. Hydrogen antimony that evolved was transferred (carrier air flow rate, 100 mL min^–1) to the IAT system, where analyte species (SbH₃) were trapped and collected onto the quartz tube surface. A continuous flow of cooling-water permitted Sb volatile species to condense on the surface of the tube during hydrides collection. The liquid phase was being continuously removed to waste after neutralization with 0.1% NaOH solution.

Hydride generation was also accomplished using the batch system; the application of this generator has been described in a previous paper.³⁸ In brief, the hydrides were generated in batch mode and were introduced into the IAT system by the carrier gas (air) during the hydride-trapping step of the atomizer temperature program only; this step was 120 s in duration in all experiments.

Hydrides were generated from 10 mL volume of samples (batch mode). The NaBH₄ solution (1.0% (m/v)) was pumped for 60 s and the hydrides that evolved were transferred (carrier air flow rate, 75 L h^–1) to the IAT system, where they were collected onto the quartz tube. The merging solution feeds the gas–liquid separator to the IAT system. A continuous flow of cooling water permitted analyte atoms to condense on the surface of the tube during hydrides collection; a further 60 s air purge of the generator completed the transfer process.

Antimony(III) . By using the continuous and/or batch HG-IAT-FAAS system described above, the selective determination of Sb(III) was obtained by discriminating the presence of Sb(V) in mild oxalic acid. Stibine from Sb(III) was selectively generated from a sample solution aliquot mixed with the 10% oxalic acid and the 2% NaBH₄ solution. It is necessary to ensure the same reaction conditions for the determination of Sb(III) and total Sb. Thus, the sample must pass through the reduction reaction both in the determination of total Sb and in the determination of only Sb(III).

The amount of Sb(V) is determined from the arithmetic difference between the responses of total inorganic antimony obtained using the total antimony procedure and Sb(III) obtained using the antimony(III) procedure.

Flame atomization of collected hydrides. After collection, the vacuum water pump was then turned on to rapidly remove the cooling water from the quartz tube. The tube was rapidly heated in the flame, generating a transient atomic absorption signal as the analyte atoms were released from the surface. Finally, the analytes were atomized for 7 s at a feasible temperature of about 1330 °C.

HG-IAT-FAAS analysis. After completion of the hydride generation and collection stage, the analyte was vaporized and atomized for 7 s by heating the quartz tube up to 1330 °C. The integrated transient absorbance signals peak area were employed at the Sb line. Both peak height and peak area signals were recorded. Peak area absorbance signals were used for calculations. A simplex optimization approach was undertaken to establish the best conditions for hydrides generation, transport, in situ trapping and vaporization/atomization. Analytical blanks were also carried through the entire procedure outlined above, in order to correct possible contaminants in the reagents that were used for the sample preparation. The mean blank value, if necessary, was substracted from the sample value after all calculations. The system was manually operated during the experiments. Quantification of Sb was based on aqueous standard calibration curves (external calibration). All detection limits were calculated for raw unsmoothed data based on a 3σ criterion of the blank counts.

Results and discussion

The study included an investigation of the hydride generation with in situ trapping (pre-concentration, collection), the thermal vaporization–atomization into the FAAS, and its application to practical analysis. The optimization parameters affecting the efficiency of the hydride generation, collection, atomization and analysis technique will be discussed separately.

Continuous-flow system versus batch mode for hydride generation

First, the generation efficiency greatly depends on whether sodium or potassium tetrahydroborate is used. For this reason, the sodium salt was ultimately preferred because a better generation efficiency and detection limit were obtained. Use of NaBH₄ enhanced sensitivity by 1–2 fold over KBH₄ (Fig. 2). Therefore, NaBH₄ was selected for further experiments in order to ensure, with greater certainty, that all of the volatile species had been stripped from solution and transferred into the atom trap unit.

Fig. 2 Influence of (a) NaBH₄ and (b) KBH₄ concentration on the peak area signals of Sb_tot (30 ng mL^–1) and Sb(III) (30 ng mL^–1) in continuous-flow system. The experimental conditions employed are detailed in the Experimental.

The hydride vapor generation systems (two laboratory-made systems described in the Experimental) that we were using for these investigations allowed us to select from two basic modes of operation, a continuous-flow and a batch mode. Preliminary experiments were performed by FAAS using aqueous standard solutions of antimony. It was determined that the generation efficiency of the volatile Sb species did not greatly depend on the modes and the type of reactors used. However, the continuous-flow system appeared to offer some advantages, such as easy automation and, in combination with trapping of the generated antimony vapor in an IAT, the use of an almost unlimited sample volume simply by pumping sample solution for a correspondingly long period of time, increasing the sensitivity accordingly. In addition, continuous-flow hydride generation ensures rapid, efficient mixing of the sample and reductant, thereby reducing the required reaction time, so lengthy purge times are not necessary as the system is a closed one. Traditional “batch” type HG methods require large purge and/or reaction times to ensure the system is flushed of air and an efficient hydrogenation reaction.

Process blank

The determination of hydride forming elements is accompanied by the problem that there are many opportunities for introducing trace amounts of these elements in the system. Blanks using the continuous mode system were determined using the same treatment procedure as for samples (microwave-assisted sample digestion procedure). Although these experiments were conducted in an ordinary laboratory, the detectable source of the blank was determined experimentally to be the reductant solution and, in particular, the sodium hydroxide used to stabilize the tetrahydroborate. Even though the chemicals used were of the best quality available, trace amounts of the analytes in the reagent were pre-concentrated in the silica tube during the in situ trapping, resulting in blank signals. Using the continuous mode system and NaBH₄ as reductant, an absolute blank of 0.6 ng for Sb_tot, was achieved.

Simplex optimization of operational variables

The optimized hydride generation conditions are given in Tables 1 and 2.

The absorption characteristics of major lines for Sb (217.6; 261.4; 283.3 nm) were investigated by the HG-FAAS method. As a consequence, the Sb 217.6 nm line giving the largest S/B was used throughout as analytical line.

The stability of the flame is obviously controlled by the gas flow rate, it is therefore essential to keep the flame stable in the section of the atomizer. The effects of flame conditions on the trapping and release of the antimony were studied by varying the fuel flow rate. The influence of the flame condition on the signal intensity was investigated by fixing the air flow rate (475 L h^–1) and altering the acetylene flow rate. The result of our experiment showed that the best sensitivity was obtained by using a 50 L h^–1 of flow rate (lean flame) for acetylene. The absorbance of Sb was not very different for collecting or releasing in a lean, full-rich and stoichiometric flame. Therefore, the air flow rate of 475 L h^–1 and acetylene flow rate of 50 L h^–1 were used for further experiments during the collection and release cycles.

The water-cooled silica tube trap position was not optimized in our experiments, but was selected based upon previous experience.¹⁶ The optimum position of the trap tube (single silica tube and STAT) corresponded to the distance (gap) of 5 mm above the burner, and the position of the silica tube corresponded to obscuration of about one-fourth of the light beam by the upper part of the tube. No significant differences were found in the absorbances when a coolant water flow rate of 1–4 L min^–1 were used during the collection cycle of Sb hydrides.

The trapping time is one of the most important factors concerning the sensitivity of HG-IAT-FAAS. In a time-based pre-concentration system the sample loading time value indicates the pre-concentration time of the method and reflects the enrichment factor. It was demonstrated (although not shown) that, although the relationship is not in general linear, a longer trapping time increased the analytical signal. A reasonable trapping time per sample in a routine laboratory would be about 2 min, in order to ensure, with greater certainty, that all of the Sb hydrides had been stripped from the solution and transferred into the atomizer and as a compromise between medium sample consumption, high sensitivity and sufficient sampling frequency. This time was chosen to investigate the analytical performance of the HG-IAT-FAAS system with respect to linearity, sensitivity, precision and detection limit.

To optimize the sample and reductant flow rate for antimony determination, first the optimum flow for Sb was estimated in the range of 0.5–5 mL min^–1 (Fig. 3(a)). It was observed that when the flow rate was low (0.5–1 mL min^–1), the absorbance increased with the flow rate; when the flow rate went up to 1 mL min^–1, the absorbance reached maximum; however, with the further increase of the flow rate (above 2 mL min^–1), the absorbance decreased for antimony. Therefore, in this study a 1 mL min^–1 sample and reductant flow rate was chosen.

Fig. 3 Influence of (a) sample and reductant uptake rate; (b) L-cysteine concentration; (c) pre-reduction time; (d) HCl concentration and (e) flow rate of carrier air on the peak area signals of Sb_tot (30 ng mL^–1) and Sb(III) (30 ng mL^–1) in continuous flow system. The experimental conditions employed are detailed in the Experimental.

The continuous hydride generation system was optimized by varying the concentration of NaBH₄, HCl, L-cysteine, oxalic acid and the air carrier flow that regulate volatilization and transport. However, substantial optimization of the generation parameters for the antimony hydrides was not undertaken, as this information was readily available from a review paper⁴ (and references cited therein) pertaining to trace element detection by atom trapping and in situ pre-concentration for FAAS and on our own observations. All of the factors show a more or less significant effect.

A pre-reduction step is needed to overcome possible errors in total antimony determination, because the efficiency to produce SbH₃ depends on the state of oxidation of the element. In the present work, the use of L-cysteine, potassium iodide and thiourea was evaluated. A most efficient pre-reduction of Sb(V) was obtained when L-cysteine was used. Therefore, L-cysteine could be efficiently used as a pre-reducing agent for the reduction of Sb(V) to the lower oxidation state Sb(III), and has been proposed as effective pre-reducer of Sb(V). In addition, oxalic acid is one of the chelating agents that have been used to mask the interference from transition metals (i.e., Co²⁺, Cu²⁺, Fe³⁺, Ni²⁺); the effect is to prevent the reduction of the interfering metals. It should, therefore, be particularly well suited to the masking of transition metals. The results in Fig. 3(b) and (c) shows that the analytical signal increased with increasing L-cysteine concentration and contact time up to 3.5% (m/v) and 30 min, respectively, but above these values the signal reached a plateau. A concentration of 3.5% (m/v) of L-cysteine and a contact time of 30 min were selected for further experiments as convenient for the pre-reduction of all antimony species. At the same time, this concentration does not affect the determination of Sb(III). It is known that Sb(III) may be reduced to the hydride form in the presence of Sb(V) in some acid media. The efficiency of the generation of antimony hydrides in hydrochloric acid and nitric acid was investigated. When using HNO₃ the Sb signals were found to be about 20 to 30% lower than those obtained with HCl. Therefore, it was confirmed that HCl is the most appropriate acid to use. HCl was found to have a plateau of absorbance at a concentration range of 3–3.5 mol L^–1. Thus, a 3 mol L^–1 portion of HCl was chosen for the generation of Sb_tothydride vapors (Fig. 3(d)). Although, when 3 mol L^–1HCl was used very few Sb(III) species formed hydride products. The concentration of NaBH₄ has been recognized as one of the most critical variables in HG, so was also a parameter of importance for Sb hydride generation. Concentrations in the range 0.5–5.0% of NaBH₄ were assayed using 3 mol L^–1HCl (Fig. 2(a)). It was observed that the higher its concentration the higher the signal, so the faster should be the reaction and more active intermediates should be formed, but relative standard deviations also increased. On the other hand, higher NaBH₄ concentrations than 2% (m/v) would result in a violent reaction (more hydrogen gas was generated) in the gas–liquid separator (generator) and eventually lead to an unstable signal. High NaBH₄ concentrations had to be avoided and an optimum value of 2% was chosen for further experiments. The concentration of NaOH used for the stabilization of NaBH₄ was also found to have a significant effect on the analytical signal. The highest HG efficiency was obtained when the concentration of NaOH was 0.5% (m/v); therefore, that stabilizer was added in an effort to maintain, in addition, the reagent blank as low as possible.

The air carrier gas flow through the apparatus is one of the basic parameters influencing the transport of the stybine into the trap-flame system, and also the mixing effect of hydride-forming reaction solutions, and thus it can affect the determination of Sb(III) markedly. The hydrides were stripped from the hydride generator and were trapped in the silica tube (atomizer) at air flow rates in the range of 50–150 mL min^–1 with a 120 s collection time. It was evident that the use of low carrier gas (air) flow can successfully reduce high analyte losses caused by sorption on the inner surfaces of the apparatus (transport tubing), so such use leads to a slight improvement in the analytical signals. This may be connected with the more efficient separation of the hydrides from the reaction solution. On the other hand, higher carrier gas flow rates (>100 mL min^–1) can result in a slight decrease in trapping efficiency (higher flow rates reduces peak areas). The decrease observed is probably due to the diluting effect of the air flow. An air flow of 100 mL min^–1 was used throughout the experiments, independent of the Sb oxidation state, in order to transport the hydrides completely (Fig. 3(e)). In the atomization step, a signal with regular shape was observed for trapping temperature and was not significantly influenced by sample introduction time in the tested range between 1 and 3 min. It should be stressed that multiple absorption peaks were not observed. In order to be more illustrative, a recorded peak shape after pre-concentration is shown in Fig. 4 (the signals in Fig. 4 were obtained using a computer).

Fig. 4 Typical analytical signal shape (profile). 25 ng mL^–1Sb, 2 min collection.

The length of the transfer tubing (between the reaction vessel and the nebulizer/burner IAT system) was not optimized in our experiments, but was selected based upon previous experience.¹⁷ Therefore, for practical reasons, a transport tubing length of 30 cm was selected for this study.

Validation of the method by analysis of reference materials

Validation of the technique proposed included analysis of five standard reference materials (SRMs): sediment, soil, coal fly ash, pine needles and water. These reference materials were chosen as they were the closest available to biological samples and are certified for the analyte of interest to be determined in the environmental and biological samples. The matrix effects on the HG-IAT-FAAS analytical signals were evaluated by comparing the conventional calibration with the standard additions slopes. No significant differences were found between the slopes obtained by both calibration procedures when using HG-IAT-FAAS. Therefore, to determine total antimony in all samples by this analytical technique, the conventional calibration mode was used. Results obtained for the analysis of SRMs by HG-IAT-FAAS method using aqueous standard calibration technique are summarized in Table 4, together with the corresponding confidence interval and RSD. The short-term precision is expressed as the RSD of six replicate measurements of each sample. The results obtained by external calibration technique do agree with certified values for any reference material indicating that calibration against aqueous solution could produce accurate results. The concentrations reported for this element in reference materials are significantly higher than the detection limits that are attained for the measurement.

Table 4 Determination of inorganic antimony species (concentration in µg g^–1 ± SD) in reference materials using HG-IAT-FAAS technique (continuous-flow system)

Sample	Antimony species			Certified value
	Sb(III)	Sb(V)	Sbtot	Sb(III)	Sbtot
a Standard deviation for 6 parallel determinations. b Information (uncertified) value. c Concentration in µg L^–1.
NIST SRM 2704 (Buffalo River Sediment)	—	—	3.55 ± 0.26^a	—	3.79 ± 0.15
NIST SRM 2710 (Montana Soil)	—	—	36.30 ± 2.65^a	—	38.40 ± 3.01
NBS SRM 1633a (Coal Fly Ash)	—	—	6.71 ± 0.50^a	—	7^b
NBS SRM 1575 (Pine Needles)	—	—	0.27 ± 0.02^a	—	0.2^b
NIST SRM 1643e (Trace Elements in Water)	4.55 ± 0.30^c	51.85 ± 3.79^c	56.40 ± 4.12^c	—	58.30 ± 0.61^c

---

Determination of total Sb. Since the yield of SbH₃ formation differs from that of Sb(III) and Sb(V), a pre-reduction step is needed to overcome possible errors in total antimony determination. Antimony contents in a wide variety of SRMs and CRMs were determined by HG-IAT-FAAS following pre-treatment procedures and pre-reduction with L-cysteine, although Sb is certified only in its total content. Table 4 gives the total antimony concentration in samples subjected to microwave-assisted preparation procedures by HG-IAT-FAAS.

Sample preparation is without any doubt a critical stage in Sb determination in biological and environmental samples. Microwave-assisted acid digestion in closed pressurized PTFE vessels is a well-established tool for sample decomposition and/or dissolution prior to trace element determination. The Sb concentrations in the certified reference materials were determined using HNO₃ and H₂O₂ reagents (for organic samples) and HNO₃, H₂O₂ and HF reagents (for inorganic samples). These “classical” procedures have been considered well suited to Sb determination by HG-FAAS. Almost quantitative recoveries were obtained (in general, better than 96%) when certified reference materials were subject to a high pressure treatment in a TFM-PTFE vessel at high temperature. The results likely indicate that a high pressure and temperature must be employed to destroy completely the stable organic molecules (anionic and cationic species) contained in the samples: in particular, the temperature is the determining parameter; the higher the temperature, the better the quality of decomposition. The high pressure–temperature focused microwave heated digestion system⁴⁰ allows an effective digestion temperature up to 300 °C, essential to assuring the quantitative oxidation of antimony in samples.

Determination of Sb(III). Normally, speciation of a particular metal ion is carried out by determining the total concentration and the concentration of one oxidation state of that metal ion.

Antimony speciation is carried out by the determination of Sb(III) or Sb(V) and the total antimony, which is determined by the oxidation of Sb(III) to Sb(V) or the reduction of Sb(V) to Sb(III) depending on the method used. A method was developed for the speciation of antimony by reducing Sb(V) into Sb(III) by pre-reductant.

Dissolution procedure has been applied, with control of the reaction conditions, for the speciation of Sb based on direct generation of SbH₃ from solutions and the total reduction of Sb(V) to Sb(III) by addition of oxalic acid as a complexing agent. Data in Table 4 summarize the results found for the sample (Trace Elements in Water) under study. In the determination of total Sb, the mean obtained falls within the certified interval in this case.

Analytical figures of merit

A comparison of the detection limits of the present procedure are summarized in Table 5 for conventional FAAS, HG-GF-AAS (in situ trapping) and with four atom trap designs. While direct comparison of detection limits is often misleading owing to the use of different systems, operating conditions and modes, it is clear that the detection limits that can be achieved with in situ atom trapping FAAS are two to three orders of magnitude superior to those obtained with direct conventional flame AAS.

Table 5 Comparison of detection limits (DL)^a for antimony using various atom traps (ng mL^–1)

Hydride generation system	Element	Flame AAS		STAT		WCAT ^b		IAT ^b		HG-IAT^b		In situ trapping
												QAT-HG-AAS		HG-GF-AAS
		DL	X ^c	DL	X ^c	DL	X ^c	DL	X ^c	DL	X ^c	DL	X ^c	DL	X ^c
a Detection limit defined by 3 blank criterion (n = 6). b 2 min collection time. c Enhancement (improvement) factor. d Ref. 29. e Ref. 31. f Ref. 32. g Ref. 33. h Ref. 34.
Continuous	Sb(III)	110.0	1.0	35.0	3.1	16.0	6.9	8.0	13.8	0.2	550	0.009^d	12 222	0.005^e	22 000
	Sbtot	110.0	1.0	35.0	3.1	16.0	6.9	8.0	13.8	1.1	100	—	—	0.02^f	5500
Batch	Sb(III)	110.0	1.0	35.0	3.1	16.0	6.9	8.0	13.8	0.5	220	—	—	0.014^g	7860
	Sbtot	110.0	1.0	35.0	3.1	16.0	6.9	8.0	13.8	0.7	157	—	—	0.005^h	22 000

---

The analytical performance characteristics were evaluated for antimony. The limit of detection (LOD), calculated using the IUPAC recommendation (based on a 3σ_blank criterion), was obtained by use of optimized operating conditions. Since no detection limits obtained by identical techniques are available, the results are compared to the quartz atom trap and electrochemical hydride generation atomic absorption spectrometry (QAT-HG-AAS) (in situ trapping) and to the HG-GF-AAS (in situ trapping) technique (Table 5). From Table 5, it can be seen that the detection limits of the developed procedure are evidently better, as compared with other atom traps. Detection limits obtained with a 120 s collection time were in the ppt (pg mL^–1) range and are the best ones; they are low enough to suggest that this procedure provide a viable alternative to the use of QAT-HG-AAS and HG-GF-AAS (in situ trapping) methodology for the determination of this hydride forming trace element. The achieved limit of detection (LOD): 0.2 ng mL^–1 for Sb (sensitivity enhancement compared to FAAS is 550 fold), is at least ten times worse than the LOD achieved for Sbin situ trapping in a quartz surface reported by Menemenlioglu et al.,²⁹ and in commercial graphite furnace with subsequent detection reported by other authors.^31–34 Six replicate measurements of the total procedure (reagent) blank solution were carried out and the relative standard deviation (RSD) of the background values for the raw, unsmoothed data were calculated. Precision was in the range of 8.0% (Sb, evaluated as peak area); this reflects the cumulative imprecision of all of the sample handling, hydride generation, trapping, atomization and detection steps. The peak height precision was always slightly, but significantly, worse.

Inorganic antimony determination in selected real samples

Finally, in order to evaluate the usefulness of the proposed method in determining antimony contents in some real samples, the samples were analyzed using the experimental conditions previously optimized. The results for the samples analyzed using the evaluated method are given in Table 6. In all cases, the calibration was achieved using the aqueous standard calibration curves. The proposed method was validated by spiking the samples with known amount of Sb(III). The recoveries from spiked solutions were varied in the range 95–104%. The precision of replicate determinations is typically better than 10% RSD.

Table 6 Inorganic antimony species concentration in real samples (µg g^–1 or µg mL^–1)

Sample	HG-IAT-FAAS procedure
	Sb(III)	Sb(V)	Sbtot
a Standard deviation for 6 parallel determinations. b Below detection limit. c Concentration in µg mL^–1.
Coal fly ash	—	—	8.89 ± 0.72^a
Soil	—	—	0.42 ± 0.03^a
Sediment	—	—	3.15 ± 0.30^a
Sewage	0.09 ± 0.01^c	0.19 ± 0.02^c	0.28 ± 0.02^c
River water	<DL^b	—	0.0015 ± 0.0001^c

---

Conclusions

This work is a continuation of previous efforts to employ integrated silica tubes as a trap for species volatilized in a hydride generation system.^{15–19,38,39} The results presented in this work confirm the idea that the present hyphenated technique using a continuous mode hydride generation gas phase in situ trapping on an integrated silica tubes trap, followed by atomization in acetylene–air flame with simultaneous direct thermal heating of the atomizer, can be used for the determination of trace amounts of Sb(III) and total antimony (Sb_(tot)) in real samples and reference materials. This work demonstrates that coupling the selective stibine generation after pre-reduction of Sb(V) by L-cysteine, masking Sb(V) with oxalic acid in HCl medium, can be carried out in combination with hydride generation FAAS for speciation of inorganic Sb(III) and Sb(V). Speciation can be carried out in the same solution under the same conditions without pre-separation techniques. Following the trapping stage, the performance of the device and related problems are quite similar to the case of hydride generation-graphite furnace atomization (in situ trapping) HG-GF-AAS. The detection limit of this HG-IAT-FAAS system for Sb is considerably improved compared with those reported for measurements of Sb by any flame AAS approach. This very simple and cheap technique constitutes an attractive alternative to HG-in situ trapping-GF-AAS system at significantly lower cost. Although the achieved very low concentration detection limit for Sb, 0.2 ng mL^–1, is worse (ca. one order of magnitude) than those for the in situ trapping of Sb hydrides in a commercial graphite furnace: 0.005 ng mL^–1 (ref. 31); 0.02 ng mL^–1 (ref. 32); 0.014 ng mL^–1 (ref. 33); 0.005 ng mL^–1 (ref. 34), the determinations were equivalent within the precision of the methods. However, the precision for the water was rather poor for this method because the concentration of antimony was very low and close to the detection limit of the method.

The proposed experimental speciation approach offers an interesting perspective and good prospects for the determination of other hydride forming elements, in the range of pg mL^–1, e.g., Ge and Sn. This is the subject of on-going research. The use of the atom trapping technique is attractive for laboratories not equipped with any graphite furnace apparatus. Although time required for each measurement is longer than that required for the classic hydride generation technique.

Acknowledgements

Financial support by the State Committee for Scientific Research (KBN), Poland, Grant No. 3 T09A 140 27 is gratefully acknowledged.

References

1. P. Smichowski, Y. Madrid and C. Cámara, Fresenius’ J. Anal. Chem., 1998, 360, 623.
2. M. J. Nash, J. E. Maskall and S. J. Hill, J. Environ. Monit., 2000, 2, 97.
3. M. Krachler, H. Emons and J. Zheng, Trends Anal. Chem., 2001, 20, 79.
4. H. Matusiewicz, Spectrochim. Acta, Part B, 1997, 52, 1711.
5. H. Matusiewicz and R. E. Sturgeon, Spectrochim. Acta, Part B, 1996, B51, 377.
6. C. Rondón, J. L. Burguera, M. Burguera, M. R. Brunetto, M. Gallignani and Y. Petit de Peňa, Fresenius’ J. Anal. Chem., 1995, 353, 133.
7. M. Krachler, M. Burow and H. Emons, Analyst, 1999, 124, 777.
8. Y. Petit de Peňa, O. Vielma, J. L. Burguera, M. Burguera, C. Rondón and P. Carrero, Talanta, 2001, 55, 743.
9. R. Calixto de Campos, P. Grinberg, I. Takase and A. S. Luna, Spectrochim. Acta, Part B, 2002, 57, 463.
10. É. Marlon de Moraes Flores, E. Pereira dos Santos, J. S. Barin, R. Zanella, V. L. Dressler and C. F. Bittencourt, J. Anal. At. Spectrom., 2002, 17, 819.
11. I. Ali and H. Y. Aboul-Enein, Instrumental Methods in Metal Ion Speciation, CRC Taylor & Francis, Boca Raton, 2006.
12. A. R. Kumar and P. Riyzuddin, Anal. Sci., 2005, 21, 1401.
13. A. R. Kumar and P. Riyazuddin, Int. J. Environ. Anal. Chem., 2007, 87, 469.
14. S. C. Apte and A. G. Howard, J. Anal. At. Spectrom., 1986, 1, 221.
15. H. Matusiewicz and M. Kopras, J. Anal. At. Spectrom., 1997, 12, 1287.
16. H. Matusiewicz and M. Krawczyk, Anal. Sci., 2006, 22, 249.
17. H. Matusiewicz and M. Krawczyk, Microchem. J., 2006, 83, 17.
18. H. Matusiewicz and M. Krawczyk, J. Braz. Chem. Soc., 2007, 18, 304.
19. H. Matusiewicz and M. Krawczyk, Spectrochim. Acta, Part B, 2007, 62, 309.
20. O. Cankur, N. Ertaş and O. Y. Ataman, J. Anal. At. Spectrom., 2002, 17, 603.
21. D. Korkmaz, S. Kumser, N. Ertaş, M. Mahmut and O. Y. Ataman, J. Anal. At. Spectrom., 2002, 17, 1610.
22. N. Ertaş, D. K. Korkmaz, S. Kumser and O. Y. Ataman, J. Anal. At. Spectrom., 2002, 17, 1415.
23. X.-m. Guo and X.-w. Guo, J. Anal. At. Spectrom., 2001, 16, 1414.
24. D. K. Korkmaz, N. Ertaş and O. Y. Ataman, Spectrochim. Acta, Part B, 2002, 57, 571.
25. D. Korkmaz, J. Dědina and O. Y. Ataman, J. Anal. At. Spectrom., 2004, 18, 255.
26. J. Kratzer and J. Dědina, Spectrochim. Acta, Part B, 2005, 60, 859.
27. J. Kratzer and J. Dědina, J. Anal. At. Spectrom., 2006, 21, 208.
28. P. Krejči, B. Dočekal and Z. Hrušovská, Spectrochim. Acta, Part B, 2006, 61, 444.
29. I. Menemenlioglu, D. Korkmaz and O. Y. Ataman, Spectrochim. Acta, Part B, 2007, 62, 40.
30. J. Kratzer and J. Dědina, Anal. Bioanal. Chem., 2007, 388, 793.
31. H.-W. Sinemus, J. Kleiner, H.-H. Stabel and B. Radziuk, J. Anal. At. Spectrom., 1992, 7, 433.
32. W.-W. Ding and R. E. Sturgeon, J. Anal. At. Spectrom., 1996, 11, 225.
33. P. Niedzielski and M. Siepak, Anal. Lett., 2003, 36, 971.
34. J. Y. Cabon and C. L. Madec, Anal. Chim. Acta, 2004, 504, 209.
35. H. Matusiewicz, J. Anal. At. Spectrom., 1989, 4, 265.
36. M. Veber, K. Čujes and S. Gomišček, J. Anal. At. Spectrom., 1994, 9, 285.
37. R. E. Sturgeon, S. N. Willie and S. S. Berman, Anal. Chem., 1985, 57, 2311.
38. H. Matusiewicz, M. Kopras and R. E. Sturgeon, Analyst, 1997, 122, 331.
39. H. Matusiewicz, R. Sturgeon, V. Luong and K. Moffatt, Fresenius’ J. Anal. Chem., 1991, 340, 35 (and references cited therein).
40. H. Matusiewicz, Anal. Chem., 1994, 66, 751.

---