Highly selective flow injection spectrophotometric determination of gold based on its catalytic effect on the oxidation of variamine blue by potassium iodate in aqueous N,N-dimethylformamide medium

A highly selective and simple flow injection method is reported for the determination of Au(III) in jewel samples. The method is based on the catalytic effect of Au(III) on the oxidation of 4-amino-4′-methoxydiphenylamine hydrochloride (Variamine Blue B base, VB) by KIO₃. The colored reaction product was monitored spectrophotometrically at 546 nm. A volume fraction of 40% N,N-dimethylformamide (DMF) greatly enhances the selectivity of the method. The chemical (pH and concentrations of reagents) and instrumental variables (sample injection volume, reagents flow rates, reaction coil length) affecting the determination were studied and optimized. Under the selected values, the analyte could be determined in the range of 0.1-12.0 mg L⁻¹ (r = 0.9997), at a sampling rate of 120 h⁻¹. The proposed assay was precise (s_r = 0.8% at 5.0 mg L⁻¹ Au(III), n = 12) and adequately sensitive with a 3σ limit of detection of 0.03 mg L⁻¹. The method was successfully applied to the analysis of jewel samples. The obtained results were favorably compared to flame atomic absorption spectrometry (FAAS) used as a reference method.     

Selective stopped-flow sequential injection method for the spectrophotometric determination of titanium in dental implant and natural Moroccan phosphate rock

The present work reports the first sequential injection (SI) method for the spectrophotometric determination of Ti(IV). The method is based upon the reaction of Ti(IV) with chromotropic acid (CA) in acidic medium to form a water-soluble complex (λ_max = 420 nm). The chemical and instrumental variables of the system that affected the reaction were studied. Selectivity was greatly enhanced using ascorbic acid. A linear calibration graph was obtained in the range 0.2-10.0 mg l⁻¹ Ti(IV) at a sampling frequency of 24 h⁻¹. The precision was satisfactory (s_r = 1.5% at 5.0 mg l⁻¹ Ti(IV), n = 12) and the 3σ limit of detection, c_L, was 0.7 mg l⁻¹ (n = 10). The developed method proved to be adequately selective and was applied successfully to the analysis of real samples (dental implant and natural Moroccan phosphate rock) giving accurate results based on recovery studies (98-105%).     

Sequential injection method for the direct spectrophotometric determination of bismuth in pharmaceutical products

A spectrophotometric method is reported for the determination of bismuth in pharmaceutical products using sequential injection analysis. Methylthymol blue (MTB) was used as a color forming reagent and the absorbance of the Bi(III)-MTB complex was monitored at 548 nm. The various chemical and physical variables that affected the reaction were studied. A linear calibration graph was obtained in the range 0.0-75.0 mg l⁻¹ Bi(III) at a sampling frequency of 72 h⁻¹. The reagent consumption was considerably reduced compared to conventional flow injection systems, as only 150 μl of MTB were consumed per run. The precision was very satisfactory (s_r=0.5%, at 50.0 mg l⁻¹ Bi(III), n=12) and the limit of detection, c_L, was 0.250 mg l⁻¹. The developed method was applied successfully to the analysis of various pharmaceutical products containing Bi(III). The relative errors e_r, were <1.5% in all cases and were evaluated by comparison of the obtained results with those found using atomic absorption spectrometry as the reference method.     

Fluorimetric determination of phytic acid in urine based on replacement reaction

A sensitive fluorimetric method for determination of phytic acid in human urine samples was described. The method was based on a fluorimetric replacement reaction, in which the added phytic acid replaced the Cu²⁺ ion from Cu²⁺-gelatin complex, liberating the fluorescent gelatin molecule. The fluorescence of the solution was accordingly recovered proportionally to the amount of the foreign phytic acid. The excitation wavelength was 273.5 nm and the characteristic emission wavelength was 305.0 nm, respectively. The calibration graph was obtained by plotting the recovered fluorescent intensity at maximum 305.0 nm against the added standard phytic acid, and was divided into two sections. One section was linear over the range of 0.40-2.40 mg L⁻¹ with a linear regression equation of I_f = −0.895 + 15.146c (R² > 0.9993), and the other over the range of 2.40-9.20 mg L⁻¹ with a linear regression equation of I_f = −29.526 + 26.113c (R² > 0.9996), respectively. The relative standard deviation (R.S.D.) at 95% confidence degree for a 2.0 mg L⁻¹ of standard phytic acid within 1 month was less than 1.26% (n = 5), indicating the procedure is reproducible. The detection and the quantification limits of phytic acid were estimated to be 0.23 and 0.40 mg L⁻¹, respectively. The proposed method was applied to the determination of phytic acid in urine samples and the found concentrations of phytic acid in urine were in the range of 0.49-0.75 mg L⁻¹ with recoveries of 96.2-108.8%. Comparison of the obtained results with the reported HPLC was performed, indicating the proposed method was reliable.     

Sample cleanup-free determination of mycophenolic acid and its glucuronide in serum and plasma using the novel technology of ultra-performance liquid chromatography-electrospray ionization tandem mass spectrometry

Mycophenolic acid (MPA) is an immunosuppressant drug which powerfully inhibits lymphocyte proliferation. Since the early 1990s it has been used to prevent rejection in organ transplantation. The requirement of therapeutic drug monitoring shown in previous studies raises the necessity of acquiring accurate and sensitive methods to measure MPA and its major metabolite mycophenolic acid glucuronide (MPAG).The authors developed a sample cleanup-free, rapid, and highly specific method for simultaneous measurement of MPA and MPAG in human plasma and serum using the novel technology of ultra-performance liquid chromatography-electrospray ionization tandem mass spectrometry. MPA- and MPAG-determinations were performed during a 2.0-min run time. Multiple calibration curves for the analysis of MPA and MPAG exhibited consistent linearity and reproducibility in the range of 0.05-100 (r > 0.999) mg L⁻¹ and 4-4000 mg L⁻¹ (r > 0.999), respectively. Limits of Detection were 0.014 mg L⁻¹ for MPA and 1.85 mg L⁻¹ for MPAG. Lower Limits of Quantification were 0.05 mg L⁻¹ for MPA and 2.30 mg L⁻¹ for MPAG. Interassay imprecision was <10% for both substances. Mean recovery was 103.6% (range 78.1-129.7%) for MPA and 111.1% (range 73.0-139.6%) for MPAG. Agreement was good for MPA and MPAG between the presented method and a validated HPLC-MS/MS method. The Passing-Bablok regression line for MPA and MPAG was HPLC-MS/MS = 1.14 UPLC-MS/MS—0.14 [mg L⁻¹], r = 0.96, and HPLC-MS/MS = 0.77 UPLC-MS/MS + 0.50 [mg L⁻¹], r = 0.97, respectively. This sample cleanup-free and robust LC-MS/MS assay facilitates the rapid, accurate and simultaneous determination of MPA and MPAG in human body fluids.     

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.     

Portable microfluidic system for determination of urinary creatinine

A simple, low cost and portable microfluidic system based on a two-point alkaline picrate kinetic reaction has been developed for the determination of urinary creatinine. The creatinine reacts with picric acid under alkaline conditions, forming an orange-red colour, which is monitored on PDMS microchip using a portable miniature fibre optic spectrometer at 510 nm. A linear range was displayed from 0 to 40 mg L⁻¹ creatinine (r² = 0.997) with a detection limit of 3.3 mg L⁻¹ (S/N = 3). On-chip absorbance signals are reproducible, with relative standard deviations (RSDs) of 7.1%, when evaluated with 20 mg L⁻¹ creatinine (n = 10). The standard curves in which the intra-run CVs (4.7-6.8%) and inter-run CVs (7.9%) obtained were performed on three different days and exhibited good reproducibility. The method was highly correlated with the conventional spectrophotometric method when real urine samples were evaluated (r² = 0.948; n = 15).     

Artificial neural networks study of the catalytic reduction of resazurin: stopped-flow injection kinetic-spectrophotometric determination of Cu(II) and Ni(II)

An artificial neural network (ANN) procedure was used in the development of a catalytic spectrophotometric method for the determination of Cu(II) and Ni(II) employing a stopped-flow injection system. The method is based on the catalytic action of these ions on the reduction of resazurin by sulfide. ANNs trained by back-propagation of errors allowed us to model the systems in a concentration range of 0.5-6 and 1-15 mg l⁻¹ for Cu(II) and Ni(II), respectively, with a low relative error of prediction (REP) for each cation: REP_Cu(II) = 0.85% and REP_Ni(II) = 0.79%. The standard deviations of the repeatability (s_r) and of the within-laboratory reproducibility (s_w) were measured using standard solutions of Cu(II) and Ni(II) equal to 2.75 and 3.5 mg l⁻¹, respectively: s_r[Cu(II)] = 0.039 mg l⁻¹, s_r[Ni(II)] = 0.044 mg l⁻¹, s_w[Ni(II)] = 0.045 mg l⁻¹ and s_w[Ni(II)] = 0.050 mg l⁻¹. The ANNs-kinetic method has been applied to the determination of Cu(II) and Ni(II) in electroplating solutions and provided satisfactory results as compared with flame atomic absorption spectrophotometry method. The effect of resazurin, NaOH and Na₂S concentrations and the reaction temperature on the analytical sensitivity is discussed.     

Enzymatic chemiluminescent assay of glucose by sequential-injection analysis with soluble enzyme and on-line sample dilution

This work reports a sequential-injection analysis (SIA) method for the enzymatic assay of glucose with soluble glucose oxidase (GOD) and on-line sample dilution with chemiluminescence (CL) detection. A zone of sample was aspirated in the holding coil of the SIA manifold and, if necessary, was diluted on-line by means of an auxiliary dilution conduit. Then, a zone of GOD was aspirated adjacent to the sample zone and a stopped-flow period was applied to allow the enzymatic reaction to proceed with production of hydrogen peroxide. Then, zones of a catalyst (Co(II) solution) and alkaline luminol were aspirated into the holding coil. Finally, the flow was reversed and the stacked zones were sent to a flow-cell located in front of a photomultiplier tube (PMT) that monitored the CL intensity. The linear dynamic range was 1 × 10⁻⁵-1 × 10⁻³ mol L⁻¹ glucose, the coefficient of variation at 8 × 10⁻⁵ mol L⁻¹ of glucose was s_r = 3.1% (n = 8), the limit of detection at the 3σ level was c_L = 1 × 10⁻⁶ mol L⁻¹ and the sampling frequency was 28 h⁻¹. With on-line dilution by a factor of 1/200, the linear range could be extended up to 0.2 mol L⁻¹ glucose. The advantages of the proposed method are the simple manifold and instrumentation used, the scope for automated on-line dilution, the low consumption of sample and reagents and the elimination of enzyme immobilisation procedures. The method was applied to the analysis of commercial drinks and honey with percent relative errors in glucose determination in the range 100 ± 6.1%.     

Reversed flow injection spectrophotometric determination of chlorate

An interfacing has been developed to connect a spectrophotometer with a personal computer and used as a readout system for development of a simple, rapid and sensitive reversed flow injection (rFI) procedure for chlorate determination. The method is based on the oxidation of indigo carmine by chlorate ions in an acidic solution (dil. HCl) leading to the decrease in absorbance at 610 nm. The decrease in absorbance is directly related to the chlorate concentration present in the sample solutions. Optimum conditions for chlorate were examined. A linear calibration graph over the range of 0.1-0.5 mg L⁻¹ chlorate was established with the regression equation of Y = 104.5X + 1.0, r² = 0.9961 (n = 6). The detection limit (3σ) of 0.03 mg L⁻¹, the limit of quantitation (10σ) of 0.10 mg L⁻¹ and the RSD of 3.2% for 0.3 mg L⁻¹ chlorate (n = 11) together with a sample throughput of 92 h⁻¹ were obtained. The recovery of the added chlorate in spiked water samples was 98.5 ± 3.1%. Major interferences for chlorate determination were found to be BrO₃⁻, ClO₂⁻, ClO⁻ and IO₃⁻ which were overcome by using SO₃²⁻ (as Na₂SO₃) as masking agent. The method has been successfully applied for the determination of chlorate in spiked water samples with the minimum reagent consumption of 14.0 mL h⁻¹. Good agreement between the proposed rFIA and the reference methods was found verified by Student's t-test at 95% confidence level.     

A green analytical procedure for flow-injection determination of nitrate in natural waters

Nitrate determination in waters is generally carried out with cadmium filings and carcinogenic reagents or by reaction with phenolic compounds in highly concentrated sulfuric acid medium. In this work, it was developed a green analytical procedure for nitrate determination in natural waters based on direct spectrophotometric measurements in ultraviolet, using a flow-injection system with an anion-exchange column for separation of nitrate from interfering species. The proposed method employs only one reagent (HClO₄) in a minimum amount (equivalent to 18 μL concentrated acid per determination), and allowed nitrate determination within 0.50-25.0 mg L⁻¹, without interference of up to 200.0 mg L⁻¹ humic acid; 1.0 mg L⁻¹ NO₂⁻; 200.0 mg L⁻¹ PO₄³⁻; 75.0 mg L⁻¹ Cl⁻; 50.0 mg L⁻¹ SO₄²⁻ and 15.0 mg L⁻¹ Fe³⁺. The detection limit (99.7% confidence level) and the coefficient of variation (n = 20) were estimated as 0.1 mg L⁻¹ and 0.7%, respectively. The results obtained for natural water samples were in agreement with those achieved by the reference method based on nitrate reduction with copperized cadmium at the 95% confidence level.     

Selective determination of cyanides by gas diffusion-stopped flow-sequential injection analysis and an on-line standard addition approach

A highly selective sequential injection (SI) method for the automated determination of weak-acid-dissociable cyanides is reported. The analytical procedure is based on the on-line reaction of the analyte with ninhydrin in carbonate medium to form a coloured product (λ_max = 510 nm). Cyanides are removed from sample matrix by acidification through a gas-diffusion step incorporated in the SI manifold. The effect of instrumental and chemical variables was studied. By adopting an on-line standard addition protocol, the sensitivity of the proposed method was enhanced drastically, without affecting the determination range. The assay was validated in terms of linearity (up to 200 μg L⁻¹), limit of detection (c_L = 2.5 μg L⁻¹), limit of quantitation (c_Q = 7.5 μg L⁻¹), precision (s_r < 2.5% at 100 μg L⁻¹) and selectivity. High tolerance against critical species such as sulfides and thiocyanates was achieved. The applicability of the method was demonstrated by analyzing tap and mineral water samples at levels below the limits established by international E.U. and U.S. organizations. The percent recoveries were satisfactory in all cases, ranging between 94.2 and 103.6%.     

A sequential injection system for indirect spectrophotometric determination of lactic acid in yogurt and fermented mash samples

A suitable non-enzymatic method is presented as an alternative to the lactic acid determination in yogurt and fermented mash samples. The oxidative conversion of lactic acid by Ce⁴⁺ to CO₂ was performed in a sequential injection system with a heating coil set at 45 °C. A gas diffusion unit was coupled to the flow system for promoting the permeation of CO₂, which was collected into a bromothymol blue solution (pH 8.4), used as indicator solution for the spectrophotometric determination (619 nm). Simplicity in operation, low reagent consumption, low cost and ruggedness are some remarkable characteristics of the proposed system. Base line drift was < 0.005 h^− 1. A linear range from 20.0 to 100.0 mg L^− 1 lactic acid was obtained (r² = 0.998), and the detection and quantification limits were estimated as 0.158 mg L^− 1 and 1.6 mg L^− 1, respectively. The sampling rate was 22 h^− 1, with a consumption of 0.04 g Ce⁴⁺ per determination. Interferences of matrix components were not detected. Samples of yogurt and sugar cane fermented mash products were analyzed, and no significant difference at 95% confidence level was observed when comparing the proposed method with HPLC analysis.     

Silica gel modified with lumogallion for aluminum determination by spectroscopic methods

Simple, easy to use and selective method of Al(III) sorption-spectroscopic (SS) determination was proposed. For this purpose, silica modified with tridecyloctadecylammonium chloride(SGII) using adsorption technique and glass slide modified with thin silica-poly(dimethyldiallyl-ammonium chloride) (SGI) composite film obtained by sol-gel technique were worked out. It was shown that lumogallion (LG) easily absorbs on SGI and SGII. Obtained sorbents SGIII and SGIV, respectively, were used for aluminum(III) determination by diffuse reflectance and spectrophotometric methods. The ranges of determination were (mg L⁻¹): (0.08-0.54), s_r ≤ 0.13, n = 4 for SGIII and (0.05-2.0), s_r ≤ 0.11, n = 4 for SGIV. The detection limits (blank + 3σ) for aluminum were 70 and 30 μg L⁻¹ using SGIII and SGIV, respectively, where σ is the standard deviation of blank estimation. The accuracy of the developed spectrophotometric method was examined by the determination of standard addition of aluminum in alcohol-free beverages. The relative error did not exceed 9%. SGIII can be regenerated by 0.05 M EDTANa₂H₂ solution and reused. SGIV was shown to be perspective for determination of aluminum in solution in the range of 0.01-0.13 mg L⁻¹ by solid phase luminescent technique.     

Normal spectrophotometric and stopped-flow spectrofluorimetric sequential injection methods for the determination of alendronic acid, an anti-osteoporosis amino-bisphosphonate drug, in pharmaceuticals

A normal spectrophotometric and a stopped-flow (SF) spectrofluorimetric method have been developed and optimized for the determination of alendronic acid (ALD) in its pharmaceutical formulations. Both methods are automated using the sequential injection analysis (SIA) principle. The spectrophotometric assay is based on the reaction of the analyte with Cu(II) ions in acidic medium to form an UV-absorbing derivative (λ_max = 240 nm). The SF spectrofluorimetric method is based on the reaction of ALD with o-phthalaldehyde (OPA) in the presence of 2-mercaptoethanol at basic medium (λ_ex = 340 nm/λ_em = 455 nm). Linear calibration curves were obtained in the range 1.0-60.0 mg l⁻¹ ALD for the UV method, and in the range 0.13-10.0 mg l⁻¹ ALD for the SF spectrofluorimetric one. The sampling rates were 60 and 30 h⁻¹, respectively. The developed assays are critically compared and their advantages are discussed. Both methods were applied to the analysis of an ALD containing pharmaceutical formulation with satisfactory accuracy and precision.     

Development of a spectrophotometric Sequential Injection Analysis (SIA) procedure for determination of ammonium: A Response Surface Methodology (RSM) approach

This paper describes the optimization and use of a Sequential Injection Analysis (SIA) procedure for ammonium determination in waters. Response Surface Methodology (RSM) was used as a tool for optimization of a procedure based on the modified Berthelot reaction. The SIA system was designed to (i) prepare the reaction media by injecting an air-segmented zone containing the reagents in a mixing chamber, (ii) to aspirate the mixture back to the holding coil after homogenization, (iii) drive it to a thermostated reaction coil, where the flow is stopped for a previously established time, and (iv) to pump the mixture toward the detector flow cell for the spectrophotometric measurements. Using a 100 μmol L^− 1 ammonium solution, the following factors were considered for optimization: reaction temperature (25 - 45 °C), reaction time (30 - 90 s), hypochlorite concentration (20 - 40 mmol L^− 1), nitroprusside concentration (10 - 40 mmol L^− 1) and salicylate concentration (0.1 - 0.3 mol L^− 1). The proposed system fed the statistical program with absorbance data for fast construction of response surface plots. After optimization of the method, figures of merit were evaluated, as well as the ammonium concentration in some water samples. No evidence of statistical difference was observed in the results obtained by the proposed method in comparison to those obtained by a reference method based on the phenol reaction.     

Simple sequential injection analysis system for rapid determination of microalbuminuria

A simple, specific and sensitive sequential injection analysis (SIA) system based on non-immunoassay fluorescent detection has been developed for the determination of urinary albumin. The specific binding of the dye Albumin Blue 580 (AB 580) to albumin in urine generated high emission fluorescent signals. The excitation and emission wavelengths were set at 590 and 610 nm, respectively. The analytical range was obtained from 1 to 100 mg L⁻¹, with a detection limit of 0.3 mg L⁻¹ (S/N = 3). The SIA system gave high precision with relative standard deviations (R.S.D.s) of 0.9% and 1.4% when evaluated with 15 and 100 mg L⁻¹ albumin (n = 15), respectively. The method exhibited good reproducibility, as assessed by performing four analytical curves on different days, and intra-run CVs (2.3-3.3%) and inter-run CVs (3.8%) were obtained. Rapid operation was achieved with a sample throughput of 37 h⁻¹. This method was successfully applied to the determination of urinary albumin, and the method was highly correlated with the immunoturbidimetric method (r² = 0.965; n = 72).     

Assessment of cadmium and iron adsorption in sediment,employing a flow injection analysis system with on line filtration and detection by flame atomic absorption spectrometry and thermospray flame furnace atomic absorption spectrometry

This work presents an evaluation of iron and cadmium adsorption in sediment of the Furnas Hydroelectric Plant Reservatory located in Alfenas, Minas Gerais (Brazil). The metal determination was done employing a flow injection analysis (FIA) with an on-line filtering system. As detection techniques, flame atomic absorption spectrometry (FAAS) for iron and thermospray flame furnace atomic absorption spectrometry (TS-FF-AAS) for cadmium determinations were used. The developed methodology presented good limits of detection, being 190 μg L⁻¹ for iron and 1.36 μg L⁻¹ for cadmium, and high sampling frequency for both metals 144 and 60 readings h⁻¹ for iron and cadmium, respectively. Both metals obey the Langmuir model, with maximum adsorptive capacity of 0⋅169 mg g⁻¹ for iron and 7⋅991 mg g⁻¹ for cadmium. For iron, a pseudo-first-order kinetic model was obtained with a theoretical Q_e = 9⋅8355 mg g⁻¹ (experimental Q_e = 9⋅5432 mg  g⁻¹), while for cadmium, a pseudo-second-order kinetic model was obtained, with a theoretical Q_e = 0.3123 mg g⁻¹ (experimental Q_e = 0⋅3052 mg g⁻¹).     

Evaluation of alternate lines of Fe for sequential multi-element determination of Cu, Fe, Mn and Zn in soil extracts by high-resolution continuum source flame atomic absorption spectrometry

The usefulness of the secondary line at 252.744 nm and the approach of side pixel registration were evaluated for the development of a method for sequential multi-element determination of Cu, Fe, Mn and Zn in soil extracts by high-resolution continuum source flame atomic absorption spectrometry (HR-CS FAAS). The influence of side pixel registration on the sensitivity and linearity was investigated by measuring at wings (248.325, 248.323, 248.321, 248.329, and 248.332 nm) of the main line for Fe at 248.327 nm. For the secondary line at 252.744 nm or side pixel registration at 248.325 nm, main lines for Cu (324.754 nm), Mn (279.482 nm) and Zn (213.875 nm), sample flow-rate of 5.0 mL min⁻¹ and calibration by matrix matching, analytical curves in the 0.2-1.0 mg L⁻¹ Cu, 1.0-20.0 mg L⁻¹ Fe, 0.2-2.0 mg L⁻¹ Mn, 0.1-1.0 mg L⁻¹ Zn ranges were obtained with linear correlations better than 0.998. The proposed method was applied to seven soil samples and two soil reference materials (IAC 277; IAC 280). Results were in agreement at a 95% confidence level (paired t-test) with reference values. Recoveries of analytes added to soil extracts containing 0.15 and 0.30 mg L⁻¹ Cu, 7.0 and 14 mg L⁻¹ Fe, 0.60 and 1.20 mg L⁻¹ Mn, 0.07 and 0.15 mg L⁻¹ Zn, varied within the 94-99, 92-98, 93-101, and 93-103% intervals, respectively. The relative standard deviations (n = 12) were 2.7% (Cu), 1.4% (Fe - 252.744 nm), 5.7% (Fe - 248.325 nm), 3.2% (Mn) and 2.8% (Zn) for an extract containing 0.35 mg L⁻¹ Cu, 14 mg L⁻¹ Fe, 1.1 mg L⁻¹ Mn and 0.12 mg L⁻¹ Zn. Detection limits were 5.4 μg L⁻¹ Cu, 55 μg L⁻¹ Fe (252.744 nm), 147 μg L⁻¹ Fe (248.325 nm), 3.0 μg L⁻¹ Mn and 4.2 μg L⁻¹ Zn.