Determination of beryllium and selenium in human urine and of selenium in human serum by graphite-furnace atomic absorption spectrophotometry.

For human urine beryllium (Be), each sample (500 microl) was diluted (1+1) with Nash reagent (containing 0.2% (v/v) acetylacetone and 2.0 M ammonium acetate buffer at pH 6.0) and then a 20-microl volume of Triton X-100 (0.4%, v/v) aqueous solution was added. An aliquot (10 microl) of the diluted urine mixture was introduced into a graphite cuvette and was atomized according to a temperature program. The method detection limit (MDL, 3sigma) for Be was 0.37 microg/l in the undiluted urine sample and the calibration graph was linear up to 65.0 microg/l. Calibration graphs were prepared by the standard addition method. Accuracies of 98.6-102% were obtained when testing standard reference material (SRM 2670) freeze dried human urine samples. Precision (relative standard deviation, RSD) for urine Be was < or = 2.3% (withinrun, n = 5) and was < or = 3.0% (between-run, n = 3). For human urine and serum selenium (Se), samples (100 microl) were diluted with HNO3 (0.2%, v/v) to make a (1+1) dilution for urine analysis or a (1+4) dilution for serum analysis. An additional aliquot (10 microl) of Triton X-100 (0.1%, v/v) was added to each 200 microl of (1+1) diluted urine (or 20 microl of the Triton X-100 was added to each 500 microl of (1+4) diluted serum) sample. After the diluted sample mixture (10 microl) was introduced into a graphite cuvette, the corresponding chemical modifier (10 microl, containing Ni2+ + Pd + NH4NO3 in HNO3 (0.2%, v/v)) was added to it and the mixture was atomized. The MDL (3sigma) for Se in urine and in serum was 4.4 and 21.4 microg/l in undiluted sample, respectively, and the calibration graphs were linear up to 150 and 400 microg/l. Accuracies of urine Se were 98.9 - 99.4% by testing SRM 2670 (NIST) urine standards with RSD (between-run, n = 3) within 2.9%; and that of serum Se was 97.2% when testing a certified second-generation human serum (No. 29, #664) with RSD (between-run, n = 3) of 1.4%. The proposed method can be applied easily, directly, and accurately to the measurement of Be and Se in real samples (including six urine Se and four serum Se from patients of Blackfoot Disease in Taiwan).     

Determination of pesticides in vegetables using large-volume injection column liquid chromatography-electrospray tandem mass spectrometry

Direct injection of a large volume (900 microl) of a sample extract onto a liquid chromatographic (LC) column, LC separation and electrospray tandem mass spectrometric detection were used for the quantitative analysis of a wide polarity range of pesticides in carrots and potatoes. Rapid sample preparation involved extraction of a small amount of sample (2 g) with a small volume of organic solvent (3 ml), clean-up over a filter and dilution of the organic extract with the aqueous LC eluent. The extraction efficiency for the selected pesticides was studied using methanol, acetone and acetonitrile as solvents. Evaluation of the performance of the overall method, using extraction with acetonitrile and detection in the selected-reaction-monitoring mode, showed excellent linearity in the range of 2-100 microg/kg with limits of detection of 0.5-2 microg/kg for both types of vegetable. With relative standard deviations of the MS peak area measurements of less than 6.5% (n=8) the repeatability of the method was fully satisfactory.     

Trace enrichment and measurement of platinum by flow injection inductively coupled plasma mass spectrometry

A flow injection system incorporating an alumina microcolumn has been coupled to inductively coupled plasma mass spectrometry (ICP-MS) for on-line preconcentration and determination of platinum (IV) in natural waters. Depending on the nature of the sample, a nominal preconcentration factor of up to 600 can be achieved by eluting with 50microl of 2 mol/l NH(4)OH. The limit of detection after a 5 min preconcentration time was 4 ngl(-1), with a relative standard deviation of 4% (100 ngl(-1) working solution). The proposed method was assessed for the determination of platinum (IV) in natural waters, motor car exhaust and some common analytical reagents.     

Speciation of metal-EDTA complexes by flow injection analysis with electrospray ionization mass spectrometry and ion chromatography with inductively coupled plasma mass spectrometry

Flow injection analysis (FIA) with ESI-MS and ion chromatography (IC) with inductively coupled plasma-MS (ICP-MS) as the complementary technique have been explored for the determination of metal ions as their metal-EDTA complexes. ESI-MS enabled the identification of metal-EDTA complexes such as [Mn(EDTA)](2-), [Co(EDTA)](2-), [Ni(EDTA)](2-), [Cu(EDTA)](2-), [Zn(EDTA)](2-), [Pb(EDTA)](2-), and [Fe(EDTA)](1-) and their MS spectral showed that these metal-EDTA complexes were present in solution. Based on the ESI-MS, ion chromatographic separation and ICP-MS detection of these complexes are possible because IC-ICP-MS requires stable metal-EDTA complex during the chromatographic separation. The separation of these metal-EDTA complexes was achieved on an anion-exchange column with a mobile phase containing 30 mM NH(4)(HPO(4))(2) at pH 7.5 within 7 min with ICP-MS providing element specific detection. The ICP-MS LODs for the metal-EDTA were in the range of 0.1-0.5 microg/L with the exception of Fe (15 microg/L). The proposed method was a simple procedure for sample processing, using direct injection of sample without removal of sample matrix and was successfully applied to the determination of metal-EDTA complexes in real samples.     

Ion chromatography determination of trace level bromate by large volume injection with conductivity and spectrophotometric detection after post column derivatisation

Bromate is a well known by-product produced by the ozonisation of drinking water; the allowed concentration for human consumption has to be regulated to the low microg l(-1) range. A direct injection, ion chromatographic method was developed using a tetraborate eluent with serially connected conductivity and spectrophotometric detection. Bromate was detected after post-column reaction with fuchsin at 520 nm. Sample capacity was investigated by injecting large volumes (up to 6 ml) using a high total hardness and chloride tap water. Linear correlation of bromate response with volumes from 1 ml to 6 ml was demonstrated, the main limitation being the overlapping of the chloride peak with bromate. Up to 1.5 ml sample can be injected without any pre-treatment. With more than 1.5 ml injection volume, a sample pre-treatment with a cartridge in Ag and H form, followed by a 10 min degassing in an ultrasonic bath, was needed. This method was validated by analysing secondary reference materials and real samples from a drinking water treatment plant. The method was linear from the limit of quantification to 20 microg l(-1). Reproducibilities in tap water were 18% (5 microg l(-1), n=12) and 21% (1 microg l(-1), n=4) respectively for 1.5 and 6 ml injection volumes with conductivity detection, and 17% at 0.5 microg l(-1) (n=9) with spectrophotometric detection. Calculated detection limits were 0.5 microg l(-1) (6 ml) ahd 2 microg l(-1) (1.5 ml) for conductivity detection and 0.3 microg l(-1) (1.5 ml) for spectrophotometric detection.     

Preconcentration and determination of trace elements with 2,6-diacetylpyridine functionalized Amberlite XAD-4 by flow injection and atomic spectroscopy

An on-line flow injection method for the direct determination of trace elements in environmental samples is described. A mini-column packed with 2,6-diacetylpyridine functionalized Amberlite XAD-4 was used to preconcentrate and separate 8 trace metals (Cd, Co, Cu, Mn, Ni, Pb, U and Zn) from water and extracts from solid samples. The metals were eluted with 0.1 M HNO(3) directly to the detection system (either inductively coupled plasma-mass spectrometry (ICP-MS) or flame atomic absorption spectrometry (FAAS)). As well as demonstrating that the resin could be used to preconcentrate ultra-trace analytes from natural waters, it was also shown to work well at a pH of 5.5. Therefore, after treatment of sample digests with sodium fluoride, samples that contain extremely large concentrations of iron may be analysed for trace analytes without the excess iron overloading the capacity of the resin. To this end, the analytes Cd, Co, Cu and Ni were preconcentrated from acid extracts of certified soil/sediment samples and then eluted with nitric acid to be determined on-line. Limits of detection (3sigma) of Cd = 0.33 microg l(-1), Co = 0.094 microg l(-1), Cu = 0.34 microg l(-1), Mn = 0.32 microg l(-1), Ni = 0.30 microg l(-1), Pb = 0.43 microg l(-1), U = 0.067 microg l(-1) and Zn = 0.20 microg l(-1) for the FI-ICP-MS system and Cd = 22 microg l(-1), Co = 60 microg l(-1), Cu = 10 microg l(-1) and Ni = 4.8 microg l(-1) for the FI-FAAS system were obtained. Analysis of certified reference materials showed good agreement with the certified values using the two methods.     

Determination of 0.1 to 1000 μ g/ml cadmium in a hydrometallurgical zinc refining process stream by a flow injection technique with computer controlled injection method

A computer-controlled flow injection system was developed for the determination of cadmium in a hydrometallurgical zinc refining process stream. An anion-exchange method in acidic potassium iodide medium was used for the on-line separation of cadmium from the matrix zinc. 1-(4-Nitrophenyl)-3-(4-phenylazophenyl)triazene (Cadion) was used as the chromogenic reagent for the spectrophotometric detection of cadmium. In order to expand the dynamic range of the flow injection - spectrophotometry, a computer-aided time-based variable-volume injection method has been employed for the introduction of the sample into the flow injection system. Samples ranging from 0.56 to 350 microl can be delivered by controlling the time period of the sample introduction valve and the flow rate of the carrier solution. The system permits a throughput of 5 samples per hour. The reproducibility has been proven to be satisfactory with a relative standard deviation of less than 6.2% (sample injected: 0.56 microl of 850 microg Cd/ml; n=100) and 5.0% (350 microl of 0.14 microg Cd/ml; n=5). The determination limit was 20 microg Cd/ml with 0.56 microl sample injection and 0.05 microg Cd/ml with 350 microl sample injection (the absolute amount of cadmium injected into the system was 11 ng and 17.5 ng, respectively).     

Determination of chlorophenoxy acid herbicides in water by in situ esterification followed by in-vial liquid-liquid extraction combined with large-volume on-column injection and gas chromatography-mass spectrometry

A new approach for rapidly analysing chlorophenoxy acid herbicides in water is presented. The chlorinated acids are derivatised with dimethyl sulphate in the water sample itself (800 microl) and, next, the methyl esters are extracted with 800 microl of n-hexane. A 200-microl volume of the extract is injected into the GC-MS system. The miniaturisation of both the methylation and extraction steps could be implemented because of the use of large-volume on-column injection and mass spectrometric detection. The optimisation of the methylation reaction for the simultaneous determination of (3,6-dichloro-2-methoxy)benzoic acid, (2-methyl-4-chlorophenoxy)- and (2,4-dichlorophenoxy)acetic acids, (+/-)-2-(4-chloro-2-methylphenoxy)- and 2-(2,4-dichlorophenoxy)propanoic acids and 4-(4-chloro-2-methylphenoxy)- and 4-(2,4-dichlorophenoxy)butyric acids showed that tetrabutylammonium salts act as catalysts. Addition of sodium hydroxide was required to obtain quantitative reaction yields for 4-(4-chloro-2-methylphenoxy)- and 4-(2,4-dichlorophenoxy)butyric acids. The methylation-cum-extraction procedure takes only 3 min per sample for a batch of seven samples. Linear calibration plots were obtained for the complete procedure and the limits of detection were of 10-60 ng/l with a signal-to-noise ratio (S/N) of 6. Relative standard deviations ranged from 8 to 15% (n=7) for analyte concentrations of 0.5 microg/l in surface water.     

Liquid-phase microextraction-gas chromatography-mass spectrometry for the determination of bromate, iodate, bromide and iodide in high-chloride matrix

In the determination of bromate and iodate, any free bromide and iodide present was quantitatively removed by anion exchange with silver chloride exploiting the differences in silver salts solubility product, being AgCl, 1.8 x 10(-10), AgBr, 5.0 x 10(-13), AgI, 8.3 x 10(-17), AgBrO(3), 5.5 x 10(-5) and AgIO(3), 3.1 x 10(-8). The oxyhalides were reduced with ascorbic acid to halides and converted to 4-bromo-2,6-dimethylaniline and 4-iodo-2,6-dimethylaniline by their reaction with 2-iodosobenzoate in the presence of 2,6-dimethylaniline at pH 6.4 and 2-3, respectively. Single drop microextraction (SDME) of the haloanilines in 2 microl of toluene and injection of the whole extract into GC-MS, or liquid-phase microextraction (LPME) into 50 microl of toluene and injection of 2 microl of extract, resulted in a sensitive method for bromate and iodate. The latter method of extraction has been found more robust, sensitive and to give better extraction in shorter period than SDME. Total bromine/iodine was determined without any treatment with silver chloride. High concentration of chloride in the matrix did not interfere. A rectilinear calibration graph was obtained for 0.05 microg-25 mg l(-1) of bromate/bromide and iodate/iodide, the limit of detection were 20 ng l(-1) of bromate, 15 ng l(-1) of iodate, 20 ng l(-1) of bromide and 10 ng l(-1) of iodide (by LPME in 50 microl of toluene). The method has been applied to seawater and table salt. From the pooled data, the average recovery of spiked oxyhalide/halide to real samples was in range 96.7-105.7% with RSD in range 1.6-6.5%.     

Gas chromatographic determination of benzene, toluene, ethylbenzene and xylenes using flame ionization detector in water samples with direct aqueous injection up to 250 microl

A simple method of solventless extraction of volatile organic compounds (benzene, toluene, ethylbenzene and xylenes) from aqueous samples was developed. This method allows direct injection of large volume of water sample into a gas chromatograph using the sorption capacity of the sorbent Chromosorb P NAW applied directly in the injection port of gas chromatograph. The system prevent water penetration into a column, keep it adsorbed on its surface until the analytes are stripped into a column, and the residual water is purging using split flow. The limit of detection ranging from 0.6 for benzene to 1.1 microg l(-1) for o-xylene and limit of quantification ranging 2.0-3.6 microg l(-1) are lower that those reached by gas chromatography with flame ionization detection and direct aqueous injection before.     

Fiber optic multi-channel protein detector for use in preparative continuous annular chromatography

On-line coupling continuous-flow liquid membrane extraction (CFLME) with HPLC, a novel automatic system was developed for the determination of sulfonylurea herbicides in water. After an automatic trace-enrichment process by CFLME, which is the combination of continuous flow liquid-liquid extraction and support liquid membrane (SLM) extraction, the target analytes were concentrated in 50 microl of 0.2 M Na2CO3-NaHCO3 (pH 10.0) buffer. The concentrated sample solutions were injected directly onto a C18 analytical column with a valve, and detected at 240 nm with a diode array detector. Metsulfuron methyl (MSM), and DPX-A 7881 were baseline separated with a mobile phase consisting of methanol and 67 mM KH2PO4-Na2HPO4 (pH 5.91) buffer (45+55, v+v) at a flow-rate of 1.0 ml min(-1). With an enrichment time of 10 min and enrichment sample volume of 20 ml, the enrichment factors and detection limits are 100 and 0.05 microg l(-1) for MSM, and 96 and 0.1 microg l(-1) for DPX-A 7881, respectively. The linear range and precision (RSD) are 0.1-50 microg l(-1) and 7.0% for MSM, and 0.2-50 microg l(-1) and 9.2% for DPX-A 7881, respectively. This proposed method was applied to determine MSM and DPX-A 7881 in seawater, tap water, and bottled mineral water with spiked recoveries in the range of 83-95% for MSM and 88-100% for DPX-A 7881, respectively.     

A DNA assay protocol in a lab-on-valve meso-fluidic system with detection by laser-induced fluorescence

An automatic protocol for in-situ assay of dsDNA is presented by employing a micro-sequential injection lab-on-valve meso-fluidic system, which facilitates precise fluidic handling at the 0.1-10 microl level. Sub-nano-liter to a few micro-liters of DNA sample and ethidium bromide (EB) solutions were introduced into the meso-fluidic system, where EB binding onto DNA takes place and an intercalated DNA-EB adduct was formed, which was afterwards excited in the flow cell of the LOV by a 473 nm laser beam, and the emitted fluorescence was monitored in-situvia optical fibers. The experimental variables, i.e., pH of the buffer solution, the concentration and volume of EB solution, the reaction time and the fluid flow rates, were investigated. By loading 600 nl sample and 1.0 microl EB solution, a linear calibration graph was obtained within 0.03-3.0 microg ml(-1)(dsDNA), and a detection limit (3sigma) of 0.009 microg ml(-1) was achieved, along with a sampling frequency of 60 h(-1) and a precision of 1.9% at the 1.0 microg ml(-1) level. The detection limit was further improved to 0.006 microg ml(-1) by increasing the sample volume to 2.0 microl. Plasmid DNA in E. Coli extraction and lambda-DNA/Hind III in four synthetic samples were assayed by using this procedure. For the plasmid DNA, a good agreement with the documented UV method was obtained, while spiking recoveries for the synthetic samples were 95.6-103.4%.     

Determination of tetracycline by flow injection with chemiluminescence detection

Tetracycline hydrochloride (4 × 10^?5?1 × 10^?3M) in a 40-μl aqueous sample is determined in a flow system by measurement of the chemiluminescence emitted on reaction with bromine (9.3 × 10^?3 M) at pH 10.4 (carbonate buffer). The limit of detection is 1.6 nmol per 40-μl injection.     

Determination of organophosphorus pesticides in water samples by single drop microextraction and gas chromatography-flame photometric detector

In this paper we have developed single drop microextraction (SDME) with modified 1.00 microl microsyringe, followed by gas chromatography with flame photometric detector (GC-FPD) for determination of 13 organophosphorus pesticides (OPPs) in water samples. By using a 1.00 microl microsyringe the repeatability of drop volume and injection were improved, because of using maximum volume of microsyringe and no dead volume. On the other hand, the modification of needle tip caused increasing cross section of needle tip and increasing adhesion force between needle tip and drop, thereby increasing drop stability and achieving a higher stirrer speed (up to 1700 rpm). The method used 0.9 microl of carbon tetrachloride as extractant solvent, 40 min extraction time, stirring at 1300 rpm and no salt addition. The enrichment factor of this method ranged from 540 to 830. The linear ranges were 0.01-100 microg/l (four orders of magnitude) and limits of detection were 0.001-0.005 microg/l for most of analyte. The relative standard deviation (RSD%) for 2 microg/l of OPPs in water by using internal standard was in the range 1.1-8.6% (n = 5). The recoveries of OPPs from farm water at spiking level of 1.0 microg/l were 91-104%.     

Simultaneous derivatization and extraction of anilines in waste water with dispersive liquid-liquid microextraction followed by gas chromatography-mass spectrometric detection

The one-step derivatization and extraction technique for the determination of anilines in river water by dispersive liquid-liquid microextraction (DLLME) is presented. In this method the anilines are extracted by DLLME and derivatized with pentafluorobenzaldehyde (PFBAY) in aqueous solution simultaneously. In this derivatization/extraction method, 0.5 ml acetone (disperser solvent) containing 10 microl chlorobenzene (extraction solvent) and 30 g/l pentafluorobenzaldehyde (PFBAY) dissolved in methanol was rapidly injected by syringe into 5 ml aqueous sample (pH 4.6). Within 20 min the analytes extracted and derivatized were almost finished. After centrifugation, 2 microl sedimented phase containing enriched analytes was determined by GC-MS. The effects of extraction and disperser solvent type and their volume, pH value of sample solution, derivatization and extraction time, derivatization and extraction temperature were investigated. Linearity in this developed method was ranging from 0.25 to 70 microg/l, and the correlation coefficients (R2) were between 0.9955 and 0.9989, and reasonable reproducibility ranging from 5.8 to 11.8% (n=5). Method detection limits (MDLs) ranged from 0.04 to 0.09 microg/l (n=5).     

Determination of Ba, Cd, Cu, Pb and Zn in saliva by isotope dilution direct injection inductively coupled plasma mass spectrometry

Trace elements in small sample volumes of saliva were determined by coupling a high efficiency direct injection nebulizer to inductively coupled plasma mass spectrometry and employing quantification by isotope dilution. Aliquots of 0.4 ml of human saliva were mixed with 0.1 ml of concentrated nitric acid and diluted to 2 ml with water. Sample solutions were spiked with an isotopic solution enriched in 135Ba, 112Cd, 65Cu, 206Pb and 66Zn. The amount of each isotope added to the samples and the measurement procedure were adjusted to attain precise analytical results calculated from the isotope ratios 135Ba/138Ba, 112Cd/114Cd, 65Cu/63Cu, 206Pb/208Pb and 66Zn/68Zn. Data acquisition for Ba, Cu and Zn isotopes was performed for a single sample injection of 50 microl and in another sample injection the Cd and Pb isotopes were measured. Concentrations ranging from 5.0 to 16 microg l(-1) for Ba, from 0.50 to 1.1 microg l(-1) for Cd, from 6.0 to 50 microg l(-1) for Cu, from 0.8 to 18.8 microg l(-1) for Pb and from 46.0 to 230 microg l(-1) for Zn were found in saliva samples. Detection limits of 0.11, 0.03, 0.40, 0.05 and 0.59 microg l(-1) were determined for Ba, Cd, Cu, Pb and Zn, respectively. The concentrations found by isotope dilution were in agreement with those of the completely digested samples quantified by external calibration. The direct analysis of 30 samples per hour was attained with the proposed procedure, avoiding time-consuming digestion steps, contamination risks and matrix effects.     

Characterization of the interactions in polymer-filler systems by inverse gas chromatography

Chromium is a primary drinking water contaminant in the USA with hexavalent chromium, Cr(VI), being the most toxic form of the metal. As a required step in developing a revised state drinking water standard for chromium, the California Department of Health Services recently issued a new Public Health Goal (PHG) of 2.5 microg/l for total chromium and 0.2 microg/l for Cr(VI). Hexavalent chromium can be determined (as chromate) by ion chromatography, as described in US Evironmental Protection Agency Method 218.6; however, the method as originally published does not allow sufficient sensitivity for analysis at the California PHG level of 0.2 microg/l. Modification of the conditions described in Method 218.6, including the use of a lower eluent flow-rate, larger reaction coil, and larger injection volume, significantly increases the method sensitivity. The modified method, which uses IonPac NG1 and AS7 guard and analytical columns, an eluent of 250 mM ammonium sulfate-100 mM ammonium hydroxide operated at 1.0 ml/min, a 1000 microl injection volume, and postcolumn reaction with 2 mM diphenylcarbazide-10% methanol-0.5 M sulfuric acid (using a 750 microl reaction coil) followed by UV-Vis detection at 530 nm, permits a method detection limit for chromate of 0.02 microg/l. This results in a quantitation limit of 0.06 microg/l, which is more than sufficient for analysis at the California PHG level. Calibration is linear over the range of 0.1-10 microg/l and quantitative recoveries (>80%) are obtained for chromate spiked at 0.2 microg/l in drinking water. The modified method provides acceptable performance, in terms of chromate peak shape and recovery, in the presence of up to 1000 mg/l chloride or 2000 mg/l sulfate.     

A new fully automated on-line digestion system for ultra trace analysis of mercury in natural waters by means of FI-CV-AFS

A fully automated flow injection (FI) system utilizing the extraordinary oxidation power of bromine monochloride (BrCl) for the transformation of dissolved mercury species to Hg(2+) and oxidation of dissolved organic carbon (DOC) has been developed and coupled to cold vapor (CV) atomic fluorescence spectrometry (AFS) for highly sensitive mercury detection. The system can be applied to natural waters, sea water as well as freshwater and provides a detection limit as low as 16 pg Hg l(-1) from a sample volume of 7 ml. The relative standard deviation is about 4-10%. A 3-fold measurement of one sample is completely processed within 15 min. Dissolved organic carbon, chloride and iodide ions are tolerated in concentrations of 15 mg DOC l(-1), >21 g Cl(-)l(-1), and 10 mg I(-)l(-1). Validation of the proposed method yielded a good recovery of total mercury in a moorland water sample and in the certified reference material ORMS-3, river water. Investigation of eight real water samples with mercury concentrations in the range of 0.3-1.4 ng l(-1) also confirmed the suitability of the proposed method.     

Flow injection electrochemical hydride generation of hydrogen selenide on lead cathode: critical study of the influence of experimental parameters.

A flow injection system for the determination of selenium by electrochemical hydride generation and quartz tube atomic absorption spectrometry is described. The generator consists of an electrolytic flow-through cell with a concentric arrangement and a packed cathode made of particulated lead. The influences of sample flow rate, carrier gas flow rate and electrolysis current on the hydrogen selenide generation have been critically studied. Both sample flow rate and electrolysis current play important roles in the efficiency of the hydride generation process. A characteristic mass of 2.4 ng and a concentration detection limit of 17 microg l(-1) were obtained for a sample volume of 420 microl.     

Determination of cadmium and bismuth in high-purity zinc metal by inductively coupled plasma mass spectrometry with on-line matrix separation

Traces of cadmium and bismuth in high-purity zinc metal were determined by inductively coupled plasma mass spectrometry (ICP-MS) in combination with flow injection (FI) on-line matrix separation (FI-ICP-MS). The anion-exchange separation method of the potassium iodide (KI) system was applied to the separation of the analytes from the matrix zinc. The analytes, cadmium and bismuth, were adsorbed on the anion-exchange (BIO. RAD AG1-X8) mini-column (1.0 mm i.d.x 100 mm bed length), while the matrix zinc can be completely removed from the anion-exchange resin. The analytes were eluted by 2 mol/l HNO(3) and directly introduced into the ICP-MS. The detection limits (D.L.) obtained by using a single injection (350 microl) were 0.81 and 0.075 ng g(-1) for cadmium and bismuth, respectively. In the case of multi-injection concentration onto the anion-exchange mini-column (five injections 350 microl each), the detection limits could be improved to 0.16 and 0.014 ng g(-1) for cadmium and bismuth, respectively. The reproducibilities of the single injection and the multi-injection method were satisfactory with a relative standard deviation of less than 5% (at the 10 and 1 ng ml(-1) level for the single injection and the multi-injection method, respectively). The method was successfully applied to the determination of trace impurities in four samples of high-purity zinc metal (7 nines grade) and three standard reference materials of high-purity unalloyed zinc samples (from NIST).