Further research on the photo-SPME of triclosan

In this study the photoinduced degradation of triclosan has been investigated by photo-solid-phase microextraction (photo-SPME). In photo-SPME, photodegradation is carried out on the SPME fibre containing the target compound. Triclosan was extracted from aqueous solutions by use of polydimethylsiloxane SPME fibres and these were subsequently exposed to UV irradiation (power 8 W, wavelength 254 nm) for different times (from 2 to 60 min). The photodegradation kinetics of triclosan were investigated, the photoproducts generated were tentatively identified, and the photochemical behaviour of these products was studied by use of this on-fibre approach followed by gas chromatographic–mass spectrometric analysis. Eight photoproducts were tentatively identified, including chlorinated phenols, chlorohydroxydiphenyl ethers, 2,8-dichlorodibenzo-p-dioxin, and a possible dichlorodibenzodioxin isomer or dichlorohydroxydibenzofuran. The main photodegradation mechanisms were postulated and photodegradation pathways proposed. The effect of pH on triclosan degradation and on triclosan-to-dioxin conversion was also investigated. Triclosan degradation occurred, and generation of 2,8-dichlorodibenzo-p-dioxin was confirmed, throughout the pH range studied (from 3 to 9).      

In situ photoproduction of dichlorodibenzo-p-dioxin from non-ionic triclosan isolated in solid argon

The infrared spectrum of triclosan [or 5-chloro-2-(2,4-dichlorophenoxy)phenol] isolated in a low temperature (～15 K) argon matrix has been recorded and assigned with help of DFT claculations undertaken with the B3LYP functional and the 6-311++G(d,p) basis set. The obtained spectrum doubtlessly exhibits the characteristic vibrational signature of the neutral (phenol) form of the compound, which exists in two different conformations (forms I and II) in the matrix, in a I: II population ratio of ca. 0.75. Upon broadband UV irradiation of the matrix-isolated triclosan with unfiltered light provided by a xenon arc lamp, formation of 2,8-dichlorodibenzo-p-dioxin (2,8-DCDD) was observed, together with HCl. The reaction seems to occur through initial photoproduction of the triclosan phenol radical derivative, and involve in the initial step participation of dissociative (πσ¹) excited states along the OH stretching coordinate, as observed previously for other phenol derivatives. The photochemically detached hydrogen atom derived from triclosan may then react with the closest located chlorine atom in the triclosan molecule to yield HCl and a biradical species, which can subsequently undergo a ring-closure reaction by intramolecular recombination, leading to the observed 2,8-DCDD.     

Stir-bar-sorptive extraction and ultra-high-performance liquid chromatography–tandem mass spectrometry for simultaneous analysis of UV filters and antimicrobial agents in water samples

Stir-bar-sorptive extraction (SBSE) with liquid desorption (LD) and ultra-high-performance liquid chromatography–electrospray ionization triple-quadrupole tandem mass spectrometry (UHPLC–(ESI)MS–MS) were used for analysis of six personal care products in environmental water: four UV filters (2,2-dihydroxy-4-methoxybenzophenone, benzophenone-3, octocrylene, and octyldimethyl-p-aminobenzoic acid) and two antimicrobial agents (triclocarban and triclosan). Experimental conditions that affect SBSE-LD sorption efficiency (extraction time and temperature, sample pH, and ionic strength) and desorption efficiency (solvent, temperature, and time) were optimized. The method proved to be sensitive—a 50-mL sample was used to determine these compounds in environmental waters at trace levels. The detection limits of the analytical method were 2.5 ng L⁻¹ for river water and 5–10 ng L⁻¹ for effluent and influent sewage water. In river waters, benzophenone-3 was found at levels from 6 ng L⁻¹ to 28 ng L⁻¹ and triclosan at levels <LOQ. Benzophenone-3 was found between 75 and 127 ng L⁻¹ in influent sewage, whereas concentrations of benzophenone-3 and triclosan were commonly below 25 ng L⁻¹ in effluent sewage.     

Cationic Surfactant-Selective Electrode Based on a Hydrophobic Cation Exchanger

 A cationic surfactant-selective electrode for sensitive analysis of cationic surfactants has been developed by using a plasticized poly (vinyl chloride) membrane based on a hydrophobic cation exchanger, sodium tetrakis (3,5-bis(trifluoromethyl)phenyl) borate. The electrode shows a Nernstian response to dodecyltrimethylammonium (DTA) ion in the concentration range from 8 × 10⁻⁷ M to 10⁻² M with a slope of 55.3 ± 2.0 mV/decade. The electrode was used over a wide pH range of pH 2–12. The electrode is excellently selective for the DTA ion over inorganic anions, but interferences of other cationic surfactants such as cetylpyridinium ion and tetradecyldimethylbenzylammonium ion (zephiramine) are great. The present electrode was applied to determine total cationic surfactants in commercial disinfectants. Received February 27, 2002; accepted June 14, 2002     

Sorption of Paraquat and 2,4-D by an Oscillatoria Sp.-Dominated Cyanobacterial Mat

The present study characterises sorption of two pesticides, namely, paraquat (PQ) and 2,4-dichlorophenoxyacetic acid (2,4-D) by an Oscillatoria sp.-dominated cyanobacterial mat. Sorption of PQ onto the test mat was not significantly affected by the pH of the solution within the pH range 2–7. However, 2,4-D sorption was strongly influenced by the solution pH and was maximum at pH 2. Whereas PQ sorption increased with increase in temperature, 2,4-D sorption showed an opposite trend. The sorption of PQ and 2,4-D achieved equilibrium within 1 h of incubation, independent of concentration of pesticide and mat biomass in the solution. The pseudo-second-order kinetic model better defined PQ sorption than the pseudo-first-order model, whereas 2,4-D sorption was well defined by both the models. Sorption isotherms of both the pesticides showed L-type curve. Freundlich model more precisely defined PQ sorption than Langmuir model, thereby suggesting heterogeneous distribution of PQ binding sites onto the biomass surface. However, the Langmuir model more correctly defined 2,4-D sorption, thus, indicating homogeneous distribution of 2,4-D binding sites onto the biomass surface. The test biomass is a good sorbent for the removal of PQ because it could, independent of pH of the solution, sorb substantial amount of PQ (q _max = 0.13 mmol g⁻¹).     

Spectrophotometric determination of triclosan in personal care products

A spectrophotometric method for the determination of triclosan in personal care products was proposed. It was based on the reaction of sodium nitrite with p-sulfanilic acid in an acidic medium to form diazonium ion, with which triclosan further formed an azo compound in an alkaline medium. The resulting yellow colored product has a maximum absorption at 452 nm. A good linear relationship (r = 0.9999) was obtained in the range of 0–30 mg L⁻¹ triclosan. A detection limit of 0.079 g L⁻¹ was achieved and the relative standard deviation was 0.24% (n = 11) at 14 mg L⁻¹ triclosan. The proposed method has been applied to the analyses of triclosan in several personal care products and the results were in good agreement with those obtained by high-performance liquid chromatography.     

Visual spectroscopy detection of triclosan

The azo coupling reaction with 2-aminonaphthalene-4,8-disulfonic acid (I) was used to develop a new cheap and rapid method of triclosan (II) determination in hygiene products. The calibration graph was linear in the range of 2.0−100 × 10⁻⁶ mol L⁻¹. The detection limit was 2.0 μmol L⁻¹.     

Determination of UV filters and antimicrobial agents in environmental water samples

Although there is increasing concern about residues from personal care products entering the aquatic environment and their potential to accumulate to levels that pose a health threat to humans and wildlife, we still know little about the extent and magnitude of their presence in the aquatic environment. In this study we describe a procedure for isolation, and subsequent determination, of compounds commonly added to personal care products. The compounds of interest include UV filters with the commercial name Eusolex (homosalate, 4-methylbenzylidenecamphor, benzophenone-3, octocrylene, butylmethoxydibenzoylmethane, ethylhexyl methoxycinnamate) and two common anti-microbial agents, clorophene and triclosan. Water samples were filtered, acidified, and extracted by use of solid-phase extraction. Extracted compounds were then derivatised before analysis by gas chromatography–mass spectroscopy. By use of our method we obtained limits of detection of 13–266 ng L⁻¹ for UV filters, and 10–186 ng L⁻¹ for triclosan and clorophene. Recoveries were 82–98% for deionised water and 50–98% for natural water (seawater, pool water, lake water, and river water). Samples collected in Slovenia included seventeen recreational waters (seawater, pool water, lake water, and river water; August 2004) and four wastewaters (January 2005). The most abundant UV filter was benzophenone-3 (11–400 ng L⁻¹). Of the two anti-microbial agents studied, trace amounts, only, of triclosan were present in the river Kolpa (68 ng L⁻¹) and in an hospital effluent (122 ng L⁻¹).     

Online phototransformation–flow injection chemiluminescence determination of triclosan

A highly selective and sensitive chemiluminescence method for the determination of triclosan is proposed. The method is based on the phototransformation of triclosan to a light-emitting precursor in the presence of fluorescein in alkaline medium and the chemiluminescence reaction is then triggered by strong base or oxidants such as N-bromosuccinimide. Based on this reaction an online phototransformation–flow injection manifold was developed, in which the photoreactor comprises a 150-cm-long × 0.8-mm-i.d. piece of PTFE tubing coiled around a 25-W fluorescent lamp, and the phototransformed products were then injected into a carrier stream of borate buffer. After mixing with the oxidant stream the produced light was detected by a photomultiplier. A wide calibration range from 8.0 × 10⁻⁸ to 1.0 × 10⁻⁴ mol L⁻¹ was obtained under the optimized conditions, and the detection limit was as low as 5.0 × 10⁻⁸ mol L⁻¹. The whole process of analysis, including the online phototransformation and subsequent chemiluminescence detection, could be completed in 6 min. Most of the foreign substances tested showed high tolerance levels, and the proposed method was directly applied to the determination of triclosan in toothpaste samples without any pre-separation procedure. Figure Schematic representation of the phototransformation of triclosan and subsequent chemiluminescence reaction     

Electrochemical oxidation behavior of 2,4-dinitrophenol at hydroxylapatite film-modified glassy carbon electrode and its determination in water samples

Hydroxylapatite (HAP)-modified glassy carbon electrode (GCE) was fabricated and used to investigate the electrochemical oxidation behavior of 2,4-dinitrophenol (2,4-DNP) by cyclic voltammetry, differential pulse voltammetry, and chronocoulometry. The oxidation peak current of 2,4-DNP at the modified electrode was obviously increased compared with the bare GCE, indicating that HAP exhibits a remarkable enhancement effect on the electrochemical oxidation of 2,4-DNP. Based on this, a sensitive and simple electrochemical method was proposed for the determination of 2,4-DNP. The effects of HAP concentration, accumulation time, accumulation potential, pH, and scan rate were examined. Under optimal conditions, the oxidation peak current of 2,4-DNP was proportional to its concentration in the range from 2.0 × 10⁻⁶ to 6.0 × 10⁻⁴ M with a correlation coefficient of 0.9987. The detection limit was 7.5 × 10⁻⁷ M (S/N = 3). The proposed method was further applied to determine 2,4-DNP in water samples with recoveries from 96.75% to 106.50%.     

Polar herbicides,pharmaceutical products,perfluorooctanesulfonate (PFOS), perfluorooctanoate (PFOA), and nonylphenol and its carboxylates and ethoxylates in surface and tap waters around Lake Maggiore in Northern Italy

A survey of contamination of surface and drinking waters around Lake Maggiore in Northern Italy with polar anthropogenic environmental pollutants has been conducted. The target analytes were polar herbicides, pharmaceuticals (including antibiotics), steroid estrogens, perfluorooctanesulfonate (PFOS), perfluoroalkyl carboxylates (including perfluorooctanoate PFOA), nonylphenol and its carboxylates and ethoxylates (NPEO surfactants), and triclosan, a bactericide used in personal-care products. Analysis of water samples was performed by solid-phase extraction (SPE) then liquid chromatography–triple-quadrupole (tandem) mass spectrometry (LC–MS–MS). By extraction of 1-L water samples and concentration of the extract to 100 μL, method detection limits (MDLs) as low as 0.05–0.1 ng L⁻¹ were achieved for most compounds. Lake-water samples from seven different locations in the Southern part of Lake Maggiore and eleven samples from different tributary rivers and creeks were investigated. Rain water was also analyzed to investigate atmospheric input of the contaminants. Compounds regularly detected at very low concentrations in the lake water included: caffeine (max. concentration 124 ng L⁻¹), the herbicides terbutylazine (7 ng L⁻¹), atrazine (5 ng L⁻¹), simazine (16 ng L⁻¹), diuron (11 ng L⁻¹), and atrazine-desethyl (11 ng L⁻¹), the pharmaceuticals carbamazepine (9 ng L⁻¹), sulfamethoxazole (10 ng L⁻¹), gemfibrozil (1.7 ng L⁻¹), and benzafibrate (1.2 ng L⁻¹), the surfactant metabolite nonylphenol (15 ng L⁻¹), its carboxylates (NPE₁C 120 ng L⁻¹, NPE₂C 7 ng L⁻¹, NPE₃C 15 ng L⁻¹) and ethoxylates (NPE _n Os, n = 3-17; 300 ng L⁻¹), perfluorinated surfactants (PFOS 9 ng L⁻¹, PFOA 3 ng L⁻¹), and estrone (0.4 ng L⁻¹). Levels of these compounds in drinking water produced from Lake Maggiore were almost identical with those found in the lake itself, revealing the poor performance of sand filtration and chlorination applied by the local waterworks.     

Evaluation of the Counter Ions in Doped Porous Polyaniline Coatings on Stainless Steel Wires for Solid Phase Microextraction of Chlorophenolics in Water Combined with High-Performance Liquid Chromatography

Porous polyaniline coatings doped with different counter ions were electrodeposited on the surface of stainless steel wires using controlled potentiostatic coulometry. Prior to electropolymerization, the stainless steel wires were chemically etched to improve subsequent immobilization of the polyaniline coatings on the substrate and to increase the effective surface area. Porous polyaniline coatings doped with sulfate, nitrate, and perchlorate counter ions were employed for solid-phase microextraction (SPME) of 2-chlorophenol, 2,4-dichlorophenol, triclosan, and 2,4-dichlorophenoxyacetic acid coupled to high-performance liquid chromatography. The results demonstrated that the perchlorate doped polyaniline coating exhibited the highest extraction efficiency for 2-chlorophenol, 2,4-dichlorophenol, and triclosan at pH 5.0, indicating that the extraction capability was modified by introducing different counter ions into the coatings. As a result, the perchlorate doped polyaniline coated fiber was further used for the optimization of extraction condition s . The method provided linear dynamic ranges over 2 to 4 orders of magnitude. The limits of detection were 0.006 µg · L^?1, 0.005 µg · L^?1, and 0.040 µg · L^?1 for 2-chlorophenol, 2,4-dichlorophenol, and triclosan, respectively. The precision expressed as the relative standard deviation ranged from 2.20% to 5.04% for spiked water at 10 µg · L ^?1 (n = 5) and the fiber to fiber reproducibility was between 3.27% and 5.91% (n = 5). The method was successfully applied to the determination of chlorophenolics in real water samples. The recoveries of chlorophenolics in spiked water at 5.0 µg · L^?1 were between 99.60% and 108.7% with relative standard deviations between 3.24% and 5.47%.     

Interaction of partially ionized poly(acrylic acid) with a surfactant counterion at constant pH values

Binding of a cationic surfactant ion, dodecylpyridinium ion, to poly(acrylic acids) of low charge densities was examined by potentiometry using surfactant-selective electrodes in the solutions, where the pH was kept constant by employing a pH buffering system. The binding of the surfactant counterions was thus able to be studied at a constant pH during the binding process. The binding took place in two steps, the first cooperative binding step and the second gradual binding step. The critical association concentration decreased as the pH increased, indicating the predominant role of the electric interaction in the binding. The binding isotherms obtained at different but constant pH values were analyzed by the matrix method, taking into account the nearest-neighbor interactions among three different kinds of sites on the polymer: ionized, protonated, and surfactant-bound. The theoretical analysis could describe only the first step but could not explain the second step. A relatively large cooperativity parameter, u, was found for the first step and it can be between 3 × 10³ and 1 × 10⁴. When the ionic strength was decreased tenfold, the cooperativity of the binding decreased (u∼1 × 10³). The binding constants of the isolated site were 5.5–6.0 × 10⁴ kg mol⁻¹ and slightly increased to 6.5 × 10⁴ kg mol⁻¹ as the ionic strength decreased. The deviation of the second step from the theoretical analysis was supposed to arise from a change of proton dissociation constant in the nonpolar space formed by the bound surfactants. Received: 29 November 2000/Accepted: 24 January 2001     

Trace determination of triclosan and triclocarban in environmental water samples with ionic liquid dispersive liquid-phase microextraction prior to HPLC–ESI-MS–MS

A novel and environmentally friendly microextraction method, termed ionic liquid dispersive liquid-phase microextraction (IL-DLPME), has been developed for rapid enrichment of triclosan and triclocarban before analysis by high-performance liquid phase chromatography–electrospray tandem mass spectrometry (HPLC–ESI-MS–MS). Instead of using toxic organic solvents, an ionic liquid was used as a green extraction solvent. This also avoided the instability of the suspending drop in single-drop liquid-phase microextraction, and the heating and cooling step in temperature-controlled ionic liquid dispersive liquid phase microextraction. Factors that may affect the enrichment efficiency, for example volume of ionic liquid, type and volume of dispersive solvent, pH, extraction time, and NaCl content were investigated in detail and optimized. Under optimum conditions, linearity of the method was observed over the range 0.2–12 μg L⁻¹ for triclocarban and 1–60 μg L⁻¹ for triclosan with correlation coefficients ranging from 0.9980 to 0.9990, respectively. The sensitivity of the proposed method was found to be excellent, with limits of detection in the range 0.040–0.58 μg L⁻¹ and precision in the range 7.0–8.8% (RSD, n = 5). This method has been successfully used to analyze real environmental water samples and satisfactory results were achieved. Average recoveries of spiked compounds were in the range 70.0–103.5%. All these results indicated that the developed method would be a green method for rapid determination of triclosan and triclocarban at trace levels in environmental water samples.     

A study of photodegradation of quetiapine by the use of LC-MS/MS method

Photodegradation of quetiapine under UVC irradiation in methanol solution was investigated and structural elucidation of its photodegradation products was performed with the use of the reversed phase UHPLC system coupled with accurate mass hybrid ESI-Q-TOF mass spectrometer. During one run all essential data for the determination of photodegradation kinetics and for the structural elucidation of the products was collected with the use of auto MS/MS mode. Five degradation products were found and their masses and formulas were obtained with high accuracy (0.26–5.02 ppm). For all the analyzed compounds, MS/MS fragmentation spectra were also obtained allowing structural elucidation of the unknown degradation products and indicating photodegradation pathways of quetiapine. The main photodegradation product was identified as 2-[2-[4-(5-oxidodibenzo[b,f][1,4]thiazepin-11-yl)-1-piperazinyl]ethoxy]-ethanol and the photodegradation reaction yields the first-order kinetics with the rate constant k = 0.1094 h⁻¹.     

Assessment of the acute toxicity of triclosan and methyl triclosan in wastewater based on the bioluminescence inhibition of Vibrio fischeri

In this work, the contributions of triclosan and its metabolite methyl triclosan to the overall acute toxicity of wastewater were studied using Vibrio fischeri. The protocol used in this paper involved various steps. First, the aquatic toxicities of triclosan and methyl triclosan were determined for standard substances, and the 50% effective concentrations (EC₅₀) were determined for these compounds. Second, the toxic responses to different mixtures of triclosan, methyl triclosan, and surfactants were studied in different water matrices, i.e., Milli-Q water, groundwater and wastewater, in order to evaluate (i) the antagonistic or synergistic effects, and (ii) the influence of the water matrices. Finally, chemical analysis was used in conjunction with the toxicity results in order to assess the aquatic toxicities of triclosan and its derivative in wastewaters. In this study, the toxicities of 45 real samples corresponding to the influents and effluents from eight wastewater treatment works (WWTW) were analyzed. Thirty-one samples were from a wastewater treatment plant (WWTP) equipped with two pilot-scale membrane bioreactors (MBR), and the influent and the effluent samples after various treatments were characterized via different chromatographic approaches, including solid-phase extraction (SPE), liquid chromatography coupled to tandem mass spectrometry (LC–MS/MS), and SPE coupled to gas chromatography–mass spectrometry (GC–MS). The toxicity was determined by measuring the bioluminescence inhibition of Vibrio fischeri. In order to complete the study and to extrapolate the results to different WWTPs, the toxicity to V. fischeri of samples from seven more plants was analyzed, as were their triclosan and methyl triclosan concentrations. Good agreement was established between the overall toxicity values and concentrations of the biocides, indicating that triclosan is one of the major toxic organic pollutants currently found in domestic wastewaters.     

高效液相色谱-电喷雾质谱联用研究三氯生的光聚合产物

三氯生在光解时可能会产生二噁英类物质,因而它在日用化学品中的安全性受到学术界的广泛关注.本文采用高效液相色谱-电喷雾质谱研究了三氯生在紫外光照射下的光聚合反应,通过对产物的准分子离子分析,初步确定了7种光聚合产物.聚合物的分子离子分别为505、540、792,且有同分异构体存在,其异构体个数分别为2个、3个和2个.这些物质的产生主要由三氯生分子间脱除氯化氢、分子上脱除氯离子等过程来聚合完成.在光解溶液中也测到2,8-二氯代二苯并-对-二噁英,该物质则是通过分子自身脱除氯化氢而环化产生.同时应用液相色谱定性分析了不同光源对光聚物产生及含量变化的影响.结果表明,三氯生的光聚合反应受光源的影响显著,UVB比UVA更易产生聚合反应.     

Molecularly imprinted solid-phase extraction and flow-injection chemiluminescence for trace analysis of 2,4-dichlorophenol in water samples

Highly sensitive flow-injection chemiluminescence (CL) combined with molecularly imprinted solid-phase extraction (MISPE) has been used for determination of 2,4-dichlorophenol (2,4-DCP) in water samples. The molecularly imprinted polymer (MIP) for 2,4-DCP was prepared by non-covalent molecular imprinting methods, using 4-vinylpyridine (4-VP) and ethylene glycol dimethacrylate (EGDMA) as the monomer and cross-linker, respectively. 2,4-DCP could be selectively adsorbed by the MIP and the adsorbed 2,4-DCP was determined by its enhancing effect on the weak chemiluminescence reaction between potassium permanganate and luminol. The enhanced CL intensity was linear in the range from 1 × 10⁻⁷ to 2 × 10⁻⁵g mL⁻¹. The LOD (S/N = 3) was 1.8 × 10⁻⁸g mL⁻¹, and the relative standard deviation (RSD) was 3.0% (n = 11) for 1.4 × 10⁻⁶g mL⁻¹. The proposed method had been successfully applied to the determination of 2,4-DCP in river water. Figure Effect of 4-VP content on the ultraviolet spectrum of 2,4-DCP in chloroform     

Photocatalytic mineralization of chlorinated organic pollutants in water by polyoxometallates. Determination of intermediates and final degradation products

Homogeneous aqueous solutions of organochlorine pesticides and chlorophenols, namely, lindane, hexachlorobenzene and 2,4-dichlorophenol (2,4DCP), undergo effective photodegradation upon photolysis with UV and near visible light in the presence of a characteristic polyoxometallate catalyst PW₁₂O₄₀ ³⁻. These substrates remained, practically, intact (lindane, HCB) or underwent minor degradation under similar conditions in absence of catalyst. The main oxidant appears to be OH radicals formed by the reaction of the excited polyoxometallate with H₂O. The system compares with the widely published TiO₂. Chlorohydroquinone (ClHQ), hydroquinone (HQ), chlorobenzoquinone (ClBQ), benzoquinone (BQ), 3,5-dichlorocatechol (3,5DCC) and 4-chlorocatechol (4CC), among others, were identified as the main aromatic intermediates in the photodegradation of 2,4DCP. Acetic acid was detected as ring cleavage product. In all cases the final photodegradation leads to complete mineralization of substrates to CO₂ and HCl.     

Nonequilibrium hollow-fiber liquid-phase microextraction with in situ derivatization for the measurement of triclosan in aqueous samples by gas chromatography–mass spectrometry

Hollow-fiber liquid-phase microextraction (HF-LPME), a relatively new sample preparation technique, has attracted much interest in the field of environmental analysis. In the current study, a novel method based on hollow-fiber liquid-phase microextraction with in situ derivatization and gas chromatography–mass spectrometry for the measurement of triclosan in aqueous samples is described. Hollow-fiber liquid-phase microextraction conditions such as the type of extraction solvent, the stirring rate, the volume of derivatizing reagent, and the extraction time were investigated. When the conditions had been optimized, the linear range was found to be 0.05–100 μg l⁻¹ for triclosan, and the limit of detection to be 0.02 μg l⁻¹. Tap water and surface water samples collected from our laboratory and Wohushan reservoir, respectively, were successfully analyzed using the proposed method. The recoveries from the spiked water samples were 83.6 and 114.1%, respectively; and the relative standard deviation (RSD) at the 1.0 μg l⁻¹ level was 6.9%.