Evaluation of a hydride generation-atomic fluorescence system for the determination of arsenic using a dielectric barrier discharge atomizer

A new atomizer based on atmospheric pressure dielectric barrier discharge (DBD) plasma was specially designed for atomic fluorescence spectrometry (AFS) in order to be applied to the measurement of arsenic. The characteristics of the DBD atomizer and the effects of different parameters (power, discharge gas, gas flow rate, and KBH₄ concentration) were discussed in the paper. The DBD atomizer shows the following features: (1) low operation temperature (between 44 and 70 °C, depending on the operation conditions); (2) low power consumption; (3) operation at atmospheric pressure. The detection limit of As(III) using hydride generation (HG) with the proposed DBD-AFS was 0.04 μg L⁻¹. The analytical results obtained by the present method for total arsenic in reference materials, orchard leaves (SRM 1571) and water samples GBW(E) 080390, agree well with the certified values. The present HG-DBD-AFS is more sensitive and reliable for the determination of arsenic. It is a very promising technique allowing for field arsenic analysis based on atomic spectrometry.     

Metal furnace heated by flame as a hydride atomizer for atomic absorption spectrometry: Sb determination in environmental and pharmaceutical samples

The present work describes a metallic hydride atomizer for atomic absorption spectrometry, by evaluating the performance of the Inconel 600^® tube. For this purpose, stibine was used as the model volatile compound and antimony determination in river and lake sediments and in pharmaceutical samples was carried out to assess the metal furnace performance. Some parameters are evaluated such as those referring to the generation and transport of the hydride (such as KBH₄ and acid concentrations, carrier gas flow rate, injected volume, etc.), as well as those referring to the metal furnace (such as tube hole area, flame composition, long-term stability, etc.). The method presents linear Sb concentration from 2 to 80 μg L⁻¹ range (r > 0.998; n = 3) and the analytical frequency of ca. 140 h⁻¹. The limit of detection (LOD) is 0.23 μg L⁻¹ and the precision, expressed as R.S.D., is less than 5% (40 μg L⁻¹; n = 10). The accuracy is evaluated through the reference materials, and the results are similar at 95% confidence level according to the t-test.     

Determination of Se,Pb, and Sb by atomic fluorescence spectrometry using a new flameless,dielectric barrier discharge atomizer

A flameless atomizer for atomic fluorescence spectrometry (AFS), based on an atmospheric pressure dielectric barrier discharge, has been developed for the atomization of hydride-forming elements, such as Se, Sb and Pb. The atomizer (8 mm o.d, 35 mm length) was operated at a power less than 50 W. The discharge was sustained with argon at the flow rate of 0.85 L min^− 1 after optimization. The characteristics of the atomizer and the effects of different parameters (power, gas flow rate, and KBH₄ concentration) are investigated. The most attractive feature of this atomizer is its low operation temperature (~ 52 °C, detected at the outlet of the atomizer by a thermocouple), allowing both the radiation source and the detector to be placed in close proximity with the atomizer. The analytical performance of the atomizer has been evaluated, and detection limits for Se, Sb and Pb obtained with the present technique were 0.08, 0.11 and 0.27 μg L^− 1, respectively. The accuracy of the system was verified by the determination of Se, Sb and Pb in reference material of spinage GBW 10015. The concentrations of Se, Sb, and Pb determined by the present technique agreed well with the reference values (Se: 92 ± 24 mg kg^− 1, Sb: 43 ± 14 mg kg^− 1, Pb: 11.1 ± 0.9 mg kg^− 1). This detector is very promising for field elements detection with portable AFS.     

Feasibility of internal standardization in the direct and simultaneous determination of As, Cu and Pb in sugar-cane spirits by graphite furnace atomic absorption spectrometry

Bismuth and Sb were evaluated as internal standards (IS) to minimize matrix effects on the direct and simultaneous determination of As, Cu, and Pb in cachaça by graphite furnace atomic absorption spectrometry using W-coated platform plus Pd-Mg(NO₃)₂ as modifier. For 20 μL injected sample, calibration within the 0.5-10 μg L⁻¹ As, 100-1000 μg L⁻¹ Cu and 0.5-30 μg L⁻¹ Pb intervals were established using the ratios As absorbance to Sb absorbance, Cu absorbance to Bi absorbance and Pb absorbance to Bi absorbance versus analytes concentration, respectively. Typical linear correlations of 0.998, 0.999 and 0.999 were, respectively, obtained. The proposed method was applied for direct determination of As, Cu and Pb in 10 commercial cachaça samples and results were in agreement with those obtained by inductively coupled plasma mass spectrometry at 95% confidence level. The found characteristic masses were 30 pg As, 274 pg Cu and 39 pg Pb. The useful lifetime of the graphite tube was around 760 firings. Recoveries of As, Cu and Pb added to cachaça samples varied, respectively, from 98% to 109%, 97% to 108% and 98% to 104% with internal standards and from 48% to 54%, 53% to 92% and 62% to 97% without internal standards. The limits of detection were 0.13 μg L⁻¹ As, 22 μg L⁻¹ Cu and 0.05 μg L⁻¹ Pb. The relative standard deviations (n = 12) for a spiked sample containing 20 μg L⁻¹ As, Pb and 500 μg L⁻¹ Cu were 1.6%, 1.0%, and 1.8% with IS and 4.3%, 5.2%, and 5.5% without IS.     

A metallic furnace atomizer in hydride generation atomic absorption spectrometry: Determination of bismuth and selenium

A flow injection hydride generation system with a metal furnace atomizer (Inconel 600® alloy) was employed for Bi and Se determination. The presented methods have linear ranges up to 200 and 500 μg L^− 1 for Bi and Se, respectively, with good linearities (r² = 0.9997 and 0.9974, respectively). The limits of quantification obtained according to IUPAC recommendations were 2.3 μg L^− 1 for Bi and 6 μg L^− 1 for Se, and the relative standard deviations (N = 6) based on Bi and Se analytical responses from real samples were 2.7% and 10%, respectively. Accuracy evaluations were based on certified materials such as SRM 361, SRM 363, and SRM 364 (steel alloys) for Bi, Mess-3 (marine sediment), SRM 397 (human hair), and Bio-Rad2 — 69042 (urine) for Se. Good agreements between the results were obtained at the 95% confidence level, according to the t-test.     

An on-line preconcentration/separation system for the determination of bismuth in environmental samples by FAAS

An on-line preconcentration procedure for the determination of bismuth by flame atomic absorption spectrometry (FAAS) has been described. Lewatit TP-207 chelating resin, including iminodiacetate group, packed in a minicolumn was used as adsorbent material. Bi(III) was sorbed on the chelating resin, from which it could be eluted with 3 mol L⁻¹ HNO₃ and then introduced directly to the nebulizer-burner system of FAAS. Best preconcentration conditions were established by testing different resin quantities, acidity of sample, types of eluent, sample and eluent solution volumes, adsorption and elution flow rates, and effect of interfering ions. The detection limit of the method was 2.75 μg L⁻¹ while the relative standard deviation was 3.0% for 0.4 μg mL⁻¹ Bi(III) concentration. The developed method has been applied successfully to the determination of bismuth in pharmaceutical cream, standard reference materials and various natural water samples with satisfactory results.     

Hollow fiber based liquid-phase microextraction for the determination of mercury traces in water samples by electrothermal atomic absorption spectrometry

A three-phase liquid microextraction procedure for the determination of mercury at low concentrations is discussed. To the aqueous sample placed at pH 7 by means of a phosphate buffer, 0.002% (m/v) 1-(2-pyridylazo)-2-naphthol (PAN) is incorporated, and the mixture submitted to microextraction with a hollow-fiber impregnated with toluene and whose lumen contains a 0.05 mol L⁻¹ ammonium iodide solution. The final measurement of the extract is carried out by electrothermal atomic absorption spectrometry (300 °C and 1100 °C for the calcination and atomization temperatures, respectively). The pyrolytic graphite atomizer is coated electrolytically with palladium. An enrichment factor of 270, which results in a 0.06 μg L⁻¹ mercury for the detection limit is obtained. The relative standard deviation at the 1 μg L⁻¹ mercury level is 3.2% (n = 5). The reliability of the procedure is verified by analyzing waters as well as six certified reference materials.     

Sensitive determination of hydrazine in water by gas chromatography–mass spectrometry after derivatization with ortho-phthalaldehyde

A gas chromatography–mass spectrometric (GC–MS) method has been established for the determination of hydrazine in drinking water and surface water. This method is based on the derivatization of hydrazine with ortho-phthalaldehyde (OPA) in water. The following optimum reaction conditions were established: reagent dosage, 40 mg mL⁻¹ of OPA; pH 2; reaction for 20 min at 70 °C. The organic derivative was extracted with methylene chloride and then measured by GC–MS. Under the established condition, the detection and the quantification limits were 0.002 μg L⁻¹ and 0.007 μg L⁻¹ by using 5.0-mL of surface water or drinking water, respectively. The calibration curve showed good linearity with r² = 0.9991 (for working range of 0.05–100 μg L⁻¹) and the accuracy was in a range of 95–106%, and the precision of the assay was less than 13% in water. Hydrazine was detected in a concentration range of 0.05–0.14 μg L⁻¹ in 2 samples of 10 raw drinking water samples and in a concentration range of 0.09–0.55 μg L⁻¹ in 4 samples of 10 treated drinking water samples.     

Evaluation of polychlorotrifluoroethylene as sorbent material for on-line solid phase extraction systems: determination of copper and lead by flame atomic absorption spectrometry in water samples

Polychlorotrifluoroethylene (PCTFE) in the form of beads was applied, as packing material for flow injection on-line column preconcentration and separation systems coupled with flame atomic absorption spectrometry (FAAS). Its performance characteristics were evaluated for trace copper determination in environmental samples. The on-line formed complex of metal with diethyldithiophosphate (DDPA) was sorbed on the PCTFE surface. Isobutyl methyl ketone (IBMK) at a flow rate of 2.8 mL min⁻¹ was used to elute the analyte complex directly into the nebulizer-burner system of spectrophotometer. The proposed sorbent material reveal, excellent chemical and mechanical resistance, fast adsorption kinetics permitting the use of high sample flow rates up to 15 mL min⁻¹ without loss of retention efficiency. For copper determination, with 90 s preconcentration time the sample frequency was 30 h⁻¹, the enhancement factor was 250, which could be further improved by increasing the loading (preconcentration) time. The detection limit (3s) was c_L = 0.07 μg L⁻¹, and the precision (R.S.D.) was 1.8%, at the 2.0 μg L⁻¹ Cu(II) level. For lead determination, the detection limit was c_L = 2.7 μg L⁻¹, and the precision (R.S.D.) 2.2%, at the 40.0 μg L⁻¹ Pb(II) level. The accuracy of the developed method was evaluated by analyzing certified reference materials and by recovery measurements on spiked natural water samples.     

Determination of trace heavy metals in herbs by sequential injection analysis-anodic stripping voltammetry using screen-printed carbon nanotubes electrodes

A method for the simultaneous determination of Pb(II), Cd(II), and Zn(II) at low μg L⁻¹ concentration levels by sequential injection analysis-anodic stripping voltammetry (SIA-ASV) using screen-printed carbon nanotubes electrodes (SPCNTE) was developed. A bismuth film was prepared by in situ plating of bismuth on the screen-printed carbon nanotubes electrode. Operational parameters such as ratio of carbon nanotubes to carbon ink, bismuth concentration, deposition time and flow rate during preconcentration step were optimized. Under the optimal conditions, the linear ranges were found to be 2-100 μg L⁻¹ for Pb(II) and Cd(II), and 12-100 μg L⁻¹ for Zn(II). The limits of detection (S_bl/S = 3) were 0.2 μg L⁻¹ for Pb(II), 0.8 μg L⁻¹ for Cd(II) and 11 μg L⁻¹ for Zn(II). The measurement frequency was found to be 10-15 stripping cycle h⁻¹. The present method offers high sensitivity and high throughput for on-line monitoring of trace heavy metals. The practical utility of our method was also demonstrated with the determination of Pb(II), Cd(II), and Zn(II) by spiking procedure in herb samples. Our methodology produced results that were correlated with ICP-AES data. Therefore, we propose a method that can be used for the automatic and sensitive evaluation of heavy metals contaminated in herb items.     

Field measurement of nitrate in marine and estuarine waters with a flow analysis system utilizing on-line zinc reduction

A sensitive reagent-injection flow analysis method for the spectrophotometric determination of nitrate in marine, estuarine and fresh water samples is described. The method is based on the reduction of nitrate in a micro column containing zinc granules at pH 6.5. The nitrite formed is reacted with sulfanilamide and N-(1-naphthyl)ethylene diamine (Griess reagent), and the resulting azo compound is quantified spectrophotometrically at 520 nm. Water samples in the range of 3-700 μg L⁻¹ NO₃⁻-N can be processed with a throughput of up to 40 samples per hour, a detection limit of 1.3 μg L⁻¹ and reproducibility of 1.2% RSD (50 μg L⁻¹ NO₃⁻-N, n = 10). The proposed method was successfully applied for the determination of nitrate in estuarine waters and the reliability was assessed by the analyses of certified reference materials and recovery experiments. The method is suitable for waters with a wide range of salinities, and was successfully used for more than 3200 underway nitrate measurements aboard SV Pelican1 in the “Two Bays” cruise in January 2010.     

Adsorptive stripping voltammetry of nickel with 1-nitroso-2-napthol using a bismuth film electrode

A sensitive procedure is presented for the voltammetric determination of nickel. The procedure involves an adsorptive accumulation of nickel 1-nitroso-2-napthol (NN) complex on a bismuth film electrode prepared ex situ by electrodeposition. The most suitable operating conditions and parameters such as pH, ligand concentration (C_NN), adsorptive potential (E_ads), adsorptive time (t_ads), scan rate and others were selected and the determination of nickel in aqueous solutions using the standard addition method was possible. The adsorbed Ni-NN complex gives a well defined cathodic stripping peak current at −0.70 V, which was used for the determination of nickel in the concentration range of 10.0-70.0 μg L⁻¹ (pH 7.5; C_NN 6.5 μmol L⁻¹; E_ads −0.30 V; t_ads 60 s) with a detection limit of 0.1 μg L⁻¹. The relative standard deviation for a solution containing 10.0 μg L⁻¹ of Ni(II) was 3.5% (n = 4). The proposed method was validated determining Ni(II) in certified reference waste water (SPS-WW1) and Certified Reference Water for Trace Elements (TMDA 51.3) with satisfactory results. Then lake water samples were analyzed.     

Determination of hydrogen sulfide and volatile thiols in air samples by mercury probe derivatization coupled with liquid chromatography-atomic fluorescence spectrometry

A new procedure is proposed for the sampling and storage of hydrogen sulphide (H₂S) and volatile thiols (methanethiol or methyl mercaptan, ethanethiol and propanethiol) for their determination by liquid chromatography. The sampling procedure is based on the trapping/pre-concentration of the analytes in alkaline aqueous solution containing an organic mercurial probe p-hydroxymercurybenzoate, HO-Hg-C₆H₄-COO⁻ (PHMB), where they are derivatized to stable PHMB complexes based on mercury-sulfur covalent bonds. PHMB complexes are separated on a C₁₈ reverse phase column, allowing their determination by liquid chromatography coupled with sequential non-selective UV-vis (DAD) and mercury specific (chemical vapor generation atomic fluorescence spectrometry, CVGAFS) on-line detectors. PHMB complexes, S(PHMB)₂CH₃S-PHMB, C₂H₅S-PHMB and C₃H₇S-PHMB, are stable alt least for 12 h at room temperature and for 3 months if stored frozen (−20 °C).The best analytical figures of merits in the optimized conditions were obtained by CVGAFS detection, with detection limits (LODc) of 9.7 μg L⁻¹ for H₂S, 13.7 μg L⁻¹ for CH₃SH, 17.7 μg L⁻¹ for C₂H₅SH and 21.7 μg L⁻¹ for C₃H₇SH in the trapping solution in form of RS-PHMB complexes, the relative standard deviation (R.S.D.) ranging between 1.0 and 1.5%, and a linear dynamic range (LDR) between 10 and 9700 μg L⁻¹. Conventional UV absorbance detectors tuned at 254 nm can be employed as well with comparable R.S.D. and LDR, but with LODc one order of magnitude higher than AFS detector and lower specificity. The sampling procedure followed by LC-DAD-CVGAFS analysis has been validated, as example, for H₂S determination by a certified gas permeation tube as a source of 3.071 ± 0.154 μg min⁻¹ of H₂S, giving a recovery of 99.8 ± 7% and it has been applied to the determination of sulfur compounds in real gas samples (biogas and the air of a plant for fractional distillation of crude oil).     

Ion chromatography for rapid and sensitive determination of fluoride in milk after headspace single-drop microextraction with in situ generation of volatile hydrogen fluoride

In this article, we report a new method that involves headspace single-drop microextraction and ion chromatography for the preconcentration and determination of fluoride. The method lies in the in situ hydrogen fluoride generation and subsequent sequestration into an alkaline microdrop (15 μL) exposed to the headspace above the stirred aqueous sample. The NaF formed in the drop was then determined by ion chromatography. The influences of some crucial single-drop microextraction parameters such as the extraction temperature, extraction time, sample stirring speed, sulphuric acid concentration and ionic strength of the sample, on extraction efficiency were investigated. In the optimal condition, an enrichment factor of 97 was achieved in 15 min. The calibration working range was from 10 μg L⁻¹ to 2000 μg L⁻¹ (R² = 0.998), and the limit of detection (signal to noise ratio of 3) was 3.8 μg L⁻¹ of fluoride. Finally, the proposed method was successfully applied to the determination of fluoride in different milk samples. The recoveries of fluoride (at spiked concentrations of 200 μg L⁻¹ and 600 μg L⁻¹ into milk) in real samples ranged from 96.9% to 107.7%. Intra-day precision (N = 3) in terms of peak area, expressed as relative standard deviation, was found to be within the range of 0.24-1.02%.     

Determination and interference studies of bismuth by tungsten trap hydride generation atomic absorption spectrometry

The determination of bismuth requires sufficiently sensitive procedures for detection at the μg L⁻¹ level or lower. W-coil was used for on-line trapping of volatile bismuth species using HGAAS (hydride generation atomic absorption spectrometry); atom trapping using a W-coil consists of three steps. Initially BiH₃ gas is formed by hydride generation procedure. The analyte species in vapor form are transported through the W-coil trap held at 289 °C where trapping takes place. Following the preconcentration step, the W-coil is heated to 1348 °C; analyte species are released and transported to flame-heated quartz atom cell where the atomic signal is formed. In our study, interferences have been investigated in detail during Bi determination by hydride generation, both with and without trap in the same HGAAS system. Interferent/analyte (mass/mass) ratio was kept at 1, 10 and 100. Experiments were designed for carrier solutions having 1.0 M HNO₃. Interferents such as Fe, Mn, Zn, Ni, Cu, As, Se, Cd, Pb, Au, Na, Mg, Ca, chloride, sulfate and phosphate were examined. The calibration plot for an 8.0 mL sampling volume was linear between 0.10 μg L⁻¹ and 10.0 μg L⁻¹ of Bi. The detection limit (3 s/m) was 25 ng L⁻¹. The enhancement factor for the characteristic concentration (C_o) was found to be 21 when compared with the regular system without trap, by using peak height values. The validation of the procedure was performed by the analysis of the certified water reference material and the result was found to be in good agreement with the certified values at the 95% confidence level.     

In situ derivatization and hollow fiber membrane microextraction for gas chromatographic determination of haloacetic acids in water

An alternative method for gas chromatographic determination of haloacetic acids (HAAs) in water using direct derivatization followed by hollow fiber membrane liquid-phase microextraction (HF-LPME) has been developed. The method has improved the sample preparation step according to the conventional US EPA Method 552.2 by combining the derivatization and the extraction into one step prior to determination by gas chromatography electron captured detector (GC-ECD). The HAAs were derivatized with acidic methanol into their methyl esters and simultaneously extracted with supported liquid hollow fiber membrane in headspace mode. The derivatization was attempted directly in water sample without sample evaporation. The HF-LPME was performed using 1-octanol as the extracting solvent at 55 °C for 60 min with 20% Na₂SO₄. The linear calibration curves were observed for the concentrations ranging from 1 to 300 μg L⁻¹ with the correlation coefficients (R²) being greater than 0.99. The method detection limits of most analytes were below 1 μg L⁻¹ except DCAA and MCAA that were 2 and 18 μg L⁻¹, respectively. The recoveries from spiked concentration ranged from 97 to 109% with %R.S.D. less than 12%. The method was applied for determination of HAAs in drinking water and tap water samples. The method offers an easy one step high sample throughput sample preparation for gas chromatographic determination of haloacetic acids as well as other contaminants in water.     

Monoclonal antibody based electrochemical immunosensor for the determination of ochratoxin A in wheat

Competitive electrochemical enzyme-linked immunosorbent assays based on disposable screen-printed electrodes have been developed for quantitative determination of ochratoxin A (OTA). The assays were carried out using monoclonal antibodies in the direct and indirect format. OTA working range, I₅₀ and detection limits were 0.05-2.5 and 0.1-7.5 μg L⁻¹, 0.35 (±0.04) μg L⁻¹ and 0.9 (±0.1) μg L⁻¹, 60 and 100 μg L⁻¹ in the direct and indirect assay format, respectively. The immunosensor in the direct format was selected for the determination of OTA in wheat. Samples were extracted with aqueous acetonitrile and the extract analyzed directly by the assay without clean-up. The I₅₀ in real samples was 0.2 μg L⁻¹ corresponding to 1.6 μg/kg in the wheat sample with a detection limit of 0.4 μg/kg (calculated as blank signal −3σ). Within- and between-assay variability were less than 5 and 10%, respectively. A good correlation (r = 0.9992) was found by comparative analysis of naturally contaminated wheat samples using this assay and an HPLC/immunoaffinity clean-up method based on the AOAC Official Method 2000.03 for the determination of OTA in barley.     

Multi-residue method for the determination of organochlorine pesticides in fish feed based on a cleanup approach followed by gas chromatography–triple quadrupole tandem mass spectrometry

A multi-residue method for the determination of organochlorine pesticides in fish feed samples was developed and optimized. The method is based on a cleanup step of the extracted fat, carried out by liquid–liquid extraction on diatomaceous earth cartridge with n-hexane/acetonitrile (80/20, v/v) followed by solid phase extraction (SPE) with silica gel–SCX cartridge, before the identification and quantification of the residues by gas chromatography–triple quadrupole tandem spectrometry (GC–MS/MS). Performance characteristics, such as accuracy, precision, linear range, limits of detection (LOD) and quantification (LOQ), for each pesticide were determined. Instrumental LODs ranged from 0.01 to 0.11 μg L⁻¹, LOQs were in the range of 0.02–0.35 μg L⁻¹, and calibration curves were linear (r² > 0.999) in the whole range of explored concentrations (5–100 μg L⁻¹). Repeatability values were in the range of 3–15%, evaluated from the relative standard deviation of six samples spiked at 100 μg kg⁻¹ of fat, and in compliance with that derived by the Horwitz's equation. No matrix effects or interfering substances were observed in fish feed analyses. The proposed method allowed high recoveries (92–116%) of spiked extracted fat samples at 100 μg kg⁻¹, and very low LODs (between 0.02 and 0.63 μg kg⁻¹) and LOQs (between 0.05 and 2.09 μg kg⁻¹) determined in fish feed samples.     

Determination of cadmium in alcohol fuel using Moringa oleifera seeds as a biosorbent in an on-line system coupled to FAAS

In this study a new method for determination of cadmium in alcohol fuel using Moringa oleifera seeds as a biosorbent in an on-line preconcentration system coupled to flame atomic absorption spectrometry (FAAS) was developed. Flow and chemical variables of the proposed system were optimized through multivariate designs. The limit of detection for cadmium was 5.50 μg L⁻¹ and the precision was below 2.3% (35.0 μg L⁻¹, n = 9). The analytical curve was linear from 5 to 150 μg L⁻¹, with a correlation coefficient of 0.9993. The developed method was successfully applied to spiked alcohol fuel, and accuracy was assessed through recovery tests, with recovery ranging from 97.50 to 100%.     

The use of a polymer inclusion membrane for separation and preconcentration of orthophosphate in flow analysis

A highly sensitive flow analysis system has been developed for the trace determination of reactive phosphate in natural waters, which uses a polymer inclusion membrane (PIM) with Aliquat 336 as the carrier for on-line analyte separation and preconcentration. The system operates under flow injection (FI) and continuous flow (CF) conditions. Under optimal FI conditions the system is characterised by a linear concentration range between 0.5 and 1000 μg L⁻¹ P, a sampling rate of 10 h⁻¹, a limit of detection of 0.5 μg L⁻¹ P and RSDs of 3.2% (n = 10, 100 μg L⁻¹) and 7.7% (n = 10, 10 μg L⁻¹). Under CF conditions with 10 min stop-flow time and sample solution flow rate of 1.32 mL min⁻¹ the flow system offers a limit of detection of 0.04 μg L⁻¹ P, a sampling rate of 5 h⁻¹ and an RSD of 3.4% (n = 5, 2.0 μg L⁻¹). Interference studies revealed that anions commonly found in natural waters did not interfere when in excess of at least one order of magnitude. The flow system, operating under CF conditions, was successfully applied to the analysis of natural water samples containing concentrations of phosphate in the low μg L⁻¹ P range, using the multipoint standard addition method.