Determination of heavy metals by inductively coupled plasma mass spectrometry after on-line separation and preconcentration

A method for the determination of Cu, As, Se, Cd, In, Hg, Tl, Pb and Bi in waters and in biological materials by inductively coupled plasma mass spectrometry, after an on-line separation, is described. The matrix separation and analyte preconcentration is accomplished by retention of the analytes complexed with the ammonium salt of O,O-diethyl dithiophosphoric acid in a HNO₃ solution on C₁₈ immobilized on silica in a minicolumn. Methanol, as eluent, is introduced in the conventional pneumatic nebulizer of the instrument. In order to use the best compromise conditions, concerning the ligand and acid concentrations, the analytes were determined in two separate groups. The enrichment factors were in the range from 5 to 61, depending on the analyte. The limits of detection varied from 0.43 ng L⁻¹ for Bi to 33 ng L⁻¹ for Cu. The sample consumption is only 2.3 mL for each group and the sampling frequency is 21 h⁻¹. The accuracy was tested by analysing five certified reference materials: water, riverine water, urine, bovine muscle and bovine liver. The agreement between obtained and certified concentrations was very good, except for As. The relatively small volume of methanol, used as eluent, minimizes the problems produced by the introduction of organic solvent into the plasma.     

Determination of rare earth elements in human hair and wheat flour reference materials by inductively coupled plasma mass spectrometry with dry ashing and microwave digestion

A method was developed for the determination of all rare earth elements (REEs) at sub ng g⁻¹ levels in human hair (GBW 09101, SRM, Republic of China) and wheat flour (GBW 08503, SRM, Republic of China) by Inductively coupled plasma mass spectrometry (ICP-MS). The values obtained by dry ashing and microwave oven digestion procedures were compared with those obtained by traditional open vessel acid digestion method. The validity of the analytical procedure was examined by analyzing spiked samples and two vegetables (GBW 07603 and GBW 07605, SRMs, Republic of China). The results are satisfactory. The detection limits for 14 REEs ranged from 0.0039 to 0.0003 ng cm⁻³ in solution and the quantification limits ranged from 0.16 to 0.01 ng g⁻¹ in solid sample. The precision for most REEs were less than 10% RSD.     

In situ atom trapping of Bi on W-coated slotted quartz tube flame atomic absorption spectrometry and interference studies

Analytical performances of metal coated slotted quartz tube flame atomic absorption spectrometry (SQT-FAAS) and slotted quartz tube in situ atom trapping flame atomic absorption spectrometry (SQT-AT-FAAS) systems were evaluated for determination of Bi. Non-volatile elements such as Mo, Zr, W and Ta were tried as coating materials. It was observed that W-coated SQT gave the best sensitivity for the determination of Bi for SQT-FAAS and SQT-AT-FAAS. The parameters for W-coated SQT-FAAS and W-coated SQT-AT-FAAS were optimized. Sensitivity of FAAS for Bi was improved as 4.0 fold by W-coated SQT-FAAS while 613 fold enhancement in sensitivity was achieved by W-coated SQT-AT-FAAS using 5.0 min trapping with respect to conventional FAAS. MIBK was selected as organic solvent for the re-atomization of Bi from the trapping surface. Limit of detection values for W-coated SQT-FAAS and W-coated SQT-AT-FAAS was obtained as 0.14 μg mL^− 1 and 0.51 ng mL^− 1, respectively. Linear calibration plot was obtained in the range of 2.5–25.0 ng mL^− 1 for W-coated SQT-AT-FAAS. Accuracy of the W-coated SQT-AT-FAAS system was checked by analyzing a standard reference material, NIST 1643e.     

Determination of zinc and copper in human hair by slurry sampling employing sequential multi-element flame atomic absorption spectrometry

The present work proposes a direct method based on slurry sampling for the determination of zinc and copper in human hair samples by multi-element sequential flame atomic absorption spectrometry. The slurries were prepared by cryogenic grinding and sonication of the samples. The optimization step was performed using univariate methodology and the factors studied were: nature and concentration of the acid solution, amount sample/slurry volume, sonication time, and particle size. The established experimental conditions are the use of a sample mass of 50 mg, 2 mol L^− 1 nitric acid solution, sonication time of 20 min and slurry volume of 10 mL. Adopting the optimized conditions, this method allows the determination of zinc and copper with detection limits of 88.3 and 53.3 ng g^− 1, respectively, and precision expressed as relative standard deviation (RSD) of 1.7% and 1.6% (both, n = 10) for contents of zinc and copper of 100.0 and 33.3 μg g^− 1, respectively. The accuracy was checked and confirmed by analysis of two certified reference materials of human hair. The procedure was applied for the determination of zinc and copper in two human hair samples. The zinc and copper contents varied from 100.0 to 175.6 and from 3.2 to 32.8 μg g^− 1, respectively. These samples were also analyzed after complete digestion in a closed system and determination by FAAS. The statistical comparison by t-test (95% confidence level) showed no significant difference between these results.     

Direct determination of trace rare earth elements in ancient porcelain samples with slurry sampling electrothermal vaporization inductively coupled plasma mass spectrometry

A method for the direct determination of trace rare earth elements in ancient porcelain samples by slurry sampling fluorinating electrothermal vaporization inductively coupled plasma mass spectrometry was developed with the use of polytetrafluoroethylene as fluorinating reagent. It was found that Si, as a main matrix element in ancient porcelain sample, could be mostly removed at the ashing temperature of 1200 °C without considerable losses of the analytes. However, the chemical composition of ancient porcelain sample is very complicated, which makes the influences resulting from other matrix elements not be ignored. Therefore, the matrix effect of ancient porcelain sample was also investigated, and it was found that the matrix effect is obvious when the matrix concentration was larger than 0.8 g l^− 1. The study results of particle size effect indicated that when the sample particle size was less than 0.057 mm, the particle size effect is negligible. Under the optimized operation conditions, the detection limits for rare earth elements by fluorinating electrothermal vaporization inductively coupled plasma mass spectrometry were 0.7 ng g^− 1 (Eu)–33.3 ng g^− 1(Nd) with the precisions of 4.1% (Yb)–10% (La) (c = 1 μg l^− 1, n = 9). The proposed method was used to directly determine the trace rare earth elements in ancient porcelain samples produced in different dynasty (Sui, Ming and Qing), and the analytical results are satisfactory.     

A two-step atomizer system using a transversely heated furnace with Zeeman background correction: Design and first solid sampling applications

A two-step-atomizer consisting of a transversely heated graphite atomization tube and a movable vaporizer graphite cup is described. The atomizer is placed between the poles of an electromagnetic field providing longitudinal Zeeman-effect background correction capability. The tube and the cup are heated by independent power supplies enabling the performance of atomic absorption measurements at temporally and spatially isothermal conditions. The design of the vaporizer provides several advantageous features including direct introduction of solid and liquid samples with extremely low contamination risk and a sampling volume of up to 105 μl. The performance of this system was assessed by analysis of the bovine liver NIST SRM 1577b and of a well characterized titanium dioxide material. Calibration curves for quantification were recorded by using aqueous standards. In comparison of the results obtained by this method with the certified values and with the results of independent methods, excellent to reasonable agreement was achieved. For the elements Fe, K, Mg, Mn, Na and Zn in titanium dioxide, the achievable limits of detection were between 60 pg g^− 1 (Mg) and 0.7 ng g^− 1 (Fe).     

Laser ablation–inductively coupled plasma–dynamic reaction cell–mass spectrometry for the multi-element analysis of polymers

In this work, the potential of laser ablation–inductively coupled plasma–mass spectrometry for the fast analysis of polymers has been explored. Different real-life samples (polyethylene shopping bags, an acrylonitrile butadiene styrene material and various plastic bricks) as well as several reference materials (VDA 001 to 004, Cd in polyethylene) have been selected for the study. Two polyethylene reference materials (ERM-EC 680 and 681), for which a reference or indicative value for the most relevant metals is available, have proved their suitability as standards for calibration.Special attention has been paid to the difficulties expected for the determination of Cr at the μg g^− 1 level in this kind of materials, due to the interference of ArC⁺ ions on the most abundant isotopes of Cr. The use of ammonia as a reaction gas in a dynamic reaction cell is shown to alleviate this problem, resulting in a limit of detection of 0.15 μg g^− 1 for this element, while limiting only modestly the possibilities of the technique for simultaneous multi-element analysis. In this regard, As is the analyte most seriously affected by the use of ammonia, and its determination has to be carried out in vented mode, at the expense of measuring time.In all cases studied, accurate results could be obtained for elements ranging in content from the sub-μg g^− 1 level to tens of thousands of μg g^− 1. However, the use of an element of known concentration as internal standard may be needed for materials with a matrix significantly different from that of the standard (polyethylene in this work).Precision ranged between 5% and 10% RSD for elements found at the 10 μg g^− 1 level or higher, while this value could deteriorate to 20% for analytes found at the sub-μg g^− 1 level. Overall, the technique evaluated presents many advantages for the fast and accurate multi-element analysis of these materials, avoiding laborious digestion procedures and minimizing the risk of analyte losses due to the formation of volatile compounds.     

Study on the solid phase extraction and spectrophotometric determination of cobalt with 5-(2-benzothiazolylazo)-8-hydroxyquinolene

A highly sensitive, selective and rapid method for the determination of cobalt based on the rapid reaction of cobalt(II) with 5-(2-benzothiazolylazo)-8-hydroxyquinolene BTAHQ and the solid phase extraction of the Co(II)-BTAHQ complex with C18 membrane disks were developed. In the presence of pH = 6.4 buffer solution and cetylpyridenium chloride (CPC) medium, BTAHQ reacts with cobalt to form a deep violet complex with a molar ratio of 1:1 (cobalt to BTAHQ). This complex was enriched by the solid phase extraction with C18 membrane disks. An enrichment factor of 100 was obtained by elution of the complex from the disks with a minimal amount of isopentyl alcohol. In isopentyl alcohol medium, the molar absorptivity of the complex is 2.42 × 10⁵ L mol⁻¹ cm⁻¹ at 658 nm. Beer’s law is obeyed in the range of 0.01–0.38 μg mL⁻¹ in the measured solution. The relative standard deviation for 11 replicate samples of 0.20 μg mL⁻¹ level is 1.37%. The detection and quantification limits reach 3.1 and 9.7 ng mL⁻¹ in the original samples. This method was applied for the determination of cobalt in biological, water, soil and pharmaceutical preparation samples with good results.     

Investigation of immiscible alloy system of Al–Sn thin films as anodes for lithium ion batteries

The Al–Sn, which is immiscible alloy, film was prepared by e-beam deposition to explore the possibility as anode material for lithium ion batteries for the first time. The film has a complex structure with tiny Sn particles dispersed homogeneously in the Al active matrix. The diffusion coefficients of Li⁺ in these Al–Sn alloy films were determined to be 2.1–3.2 × 10⁻⁸ cm²/s by linear sweep voltammetry. The film electrode with high Al content (Al–33wt%Sn) delivered a high initial discharge capacity of 972.8 mA h g⁻¹, while the film electrode with high Sn content (Al–64wt%Sn) with an initial discharge capacity of 552 mA h g⁻¹ showed good cycle performance indicated by retaining a capacity of about 381 mA h g⁻¹ after 60 cycles. Our preliminary results demonstrate that Al–Sn immiscible alloy is a potential candidate for anodic material of lithium ion batteries.     

Study of the chemical composition of particulate matter from the Rio de Janeiro metropolitan region,Brazil, by inductively coupled plasma-mass spectrometry and optical emission spectrometry

Air quality in the metropolitan region of Rio de Janeiro was evaluated by analysis of particulate matter (PM) in industrial (Santa Cruz) and rural (Seropédica) areas. Total suspended particles (TSP) and fine particulate matter (PM_2.5) collected in filters over 24 h were quantified and their chemical composition determined. TSP exceeded Brazilian guidelines (80 μg m^− 3) in Santa Cruz, while PM_2.5 levels exceeded the World Health Organization guidelines (10 μg m^− 3) in both locations. Filters were extracted with water and/or HNO₃, and the concentrations of 20 elements, mostly metals, were determined by inductively coupled plasma mass spectrometry (ICP-MS) and optical emission spectrometry (ICP OES). Water soluble inorganic anions were determined by ion chromatography (IC). To estimate the proportion of these elements extracted, a certified reference material (NIST SRM 1648a, Urban Dust) was subjected to the same extraction process. Concordant results were obtained by ICP-MS and ICP OES for most elements. Some elements could not be quantified by both techniques; the most appropriate technique was chosen in each case. The urban dust was also analyzed by the United States Environmental Protection Agency (US EPA) method, which employs a combination of hydrochloric and nitric acids for the extraction, but higher extraction efficiency was obtained when only nitric acid was employed. The US EPA method gave better results only for Sb. In the PM samples, the elements found in the highest average concentrations by ICP were Zn and Al (3–6 μg m^− 3). The anions found in the highest average concentrations were SO₄^2 − in PM_2.5 (2–4 μg m^− 3) and Cl⁻ in TSP (2–6 μg m^− 3). Principal component analysis (PCA) in combination with enrichment factors (EF) indicated industrial sources in PM_2.5. Analysis of TSP suggested both anthropogenic and natural sources. In conclusion, this work contributes data on air quality, as well as a method for the analysis of PM samples by ICP-MS.     

Square-wave voltammetric determination of thallium using surface modified thick-film graphite electrode with Bi nanopowder

Trace analysis of thallium at surface modified thick-film graphite electrode with Bi nanopowder has been carried out using square-wave anodic stripping voltammetry (SWASV) technique. The Bi nanopowder electrode exhibited a well-defined response relating to the oxidation of Tl. From the linear relationship between Tl concentration and peak current, the sensitivity of the Bi nanopowder electrode was quantitatively estimated. The detection limit of Tl was determined to be 0.03 μg/L for 1.0 μg/L Tl solution under 10 min accumulation, which is lower than the reported values for a Bi film electrode. Furthermore, it is confirmed that EDTA addition effectively eliminates the Pb and Cd interferences in the course of Tl determination by forming complexes with Pb²⁺ and Cd²⁺.     

Highly active PtRuFe/C catalyst for methanol electro-oxidation

High methanol electro-oxidation activity was obtained on novel PtRuFe/C (2:1:1 at.%) catalyst. Mass and specific activities were 5.67 A  g⁻¹ catal. and 177 mA m⁻² for the PtRuFe/C catalyst while those of the commercial PtRu/C catalyst were 2.28 A g⁻¹ catal. and 87.7 mA m⁻², respectively. CO stripping results showed that on-set voltage for CO electro-oxidation was lowered by incorporation of Fe. XRD and XPS results revealed that Fe₂O₃ was formed instead of Fe(0), which resulted in large electron deficiency in Pt and easy CO electro-oxidation. The electron deficiency of Pt was proved by XPS results of Pt4f peaks, which moved to higher binding energies in PtRuFe/C than PtRu/C.     

Determination of atorvastatin in human serum by salting out assisted solvent extraction and reversed-phase high-performance liquid chromatography–UV detection

A simple and rapid technique based on salting out assisted solvent extraction was developed for extraction of atorvastatin from serum sample and high performance liquid chromatography–UV was used for its detection. In the present study, 1.0 mL serum was extracted by 0.5 mL of acetonitrile and some parameters that can affect extraction such as type and volume of extraction solvent, type of salt, and pH were optimized. Under optimized experimental conditions, the calibration curve was found to be linear in the range of 0.001–10 ng mL⁻¹ in human serum and the correlation coefficient (R²) and the limits of detection were >0.99 and 0.0005 ng mL⁻¹, respectively. The accuracy of the method in terms of average recovery of the compound in spiked serum and water samples was better than 90%.     

Determination of Cr(VI) in plants by electrothermal atomic absorption spectrometry after leaching with sodium carbonate

The determination of chromium (VI) compounds in plants by electrothermal atomic absorption spectrometry (ET AAS) is proposed based on their leaching with 0.1 M Na₂CO₃. Due to the presence of relatively high amounts of Na₂CO₃ in the resulting samples, the temperature and time of pyrolysis and atomization stages must be optimized to minimize the influence of the matrix. A limit of detection (LOD) for determination of Cr(VI) in plants by ET AAS was found to be 0.024 μg g⁻¹.The concentration of Cr(VI) and total chromium in plants collected in different geographical areas (South Africa and Russia), grown on soils high in chromium was determined. The concentration of Cr(VI) and total Cr in stems and leaves of plants was in the range of 0.04–0.7 μg g⁻¹ and 0.5–10 μg g⁻¹, respectively. The limited uptake of Cr(III) by plants, in comparison to its concentration in soil, can be explained by the very low solubility of natural Cr(III) compounds.Results for the determination of Cr(VI) were confirmed by the analysis of BCR CRM 545 (Cr(VI) in welding dust) with good agreement between certified (39.5 ± 1.3 μg mg⁻¹) and found (38.8 ± 1.2 μg mg⁻¹) values. The total concentration of Cr in plants has also been determined by ET AAS after dry ashing of samples at 650 °C. Results were confirmed by the analysis of BCR CRM 281 (Trace elements in Rye Grass) with good agreement between the found (2.12 ± 0.16 μg g⁻¹) and certified value (2.14 ± 0.12 μg g⁻¹).     

Extraction and preconcentration of trace levels of cobalt using functionalized magnetic nanoparticles in a sequential injection lab-on-valve system with detection by electrothermal atomic absorption spectrometry

A new approach to performing extraction and preconcentration employing functionalized magnetic nanoparticles for the determination of trace metals is presented. Alumina-coated iron oxide nanoparticles were synthesized and used as the solid support. The nanoparticles were functionalized with sodium dodecyl sulfate and used as adsorbents for solid phase extraction of the analyte. Extraction, elution, and detection procedures were performed sequentially in the sequential injection lab-on-valve (SI-LOV) system followed by electrothermal atomic absorption spectrometry (ETAAS). Mixtures of hydrophobic analytes were successfully extracted from solution using the synthesized magnetic adsorbents. The potential use of the established scheme was demonstrated by taking cobalt as a model analyte. Under the optimal conditions, the calibration curve showed an excellent linearity in the concentration range of 0.01–5 μg L^?1, and the relative standard deviation was 2.8% at the 0.5 μg L^?1 level (n = 11). The limit of detection was 6 ng L^?1 with a sampling frequency of 18 h^?1. The present method has been successfully applied to cobalt determination in water samples and two certified reference materials.     

Minimization of mass interferences in quadrupole inductively coupled plasma mass spectrometric (ICP-MS) determination of palladium using a flow injection on-line displacement solid-phase extraction protocol

A flow injection on-line displacement solid-phase extraction protocol was employed to minimize mass interferences with determination of palladium by inductively coupled plasma mass spectrometry (ICP-MS). The developed method involved in on-line complexing of Ag⁺ with pyrrolidine dithiocarbamate (PDC), presorption of the resultant Ag–PDC onto a microcolumn packed with the cigarette filter, displacement sorption of Pd²⁺ through loading the sample solution onto the microcolumn due to on-line displacement reaction between Pd²⁺ and the presorbed Ag–PDC, elution of the retained Pd²⁺ with 50 μL of ethanol for on-line ICP-MS detection. Interferences from co-existing heavy metal ions with lower stability of their PDC complexes relative to Ag–PDC were minimized/eliminated. No interferences from 5 mg L^− 1 Zn and 3 mg L^− 1 Pb for ¹⁰⁴Pd, 0.4 mg L^− 1 Cu for ¹⁰⁵Pd, 6 mg L^− 1 Zn and 2 mg L^− 1 Cd for ¹⁰⁶Pd, 6 mg L^− 1 Zn and 3 mg L^− 1 Cd for ¹⁰⁸Pd, and 2 mg L^− 1 Cd for ¹¹⁰Pd were observed for the determination of 100 ng L^− 1 Pd. The enhancement factors of 71–75, sample throughput of 23 samples h^− 1 and detection limits of 2.8–3.5 ng L^− 1 were achieved with the consumption of 3.0 mL of sample solution. The precision (RSD) for eleven replicate determinations of Pd at the 100 ng L^− 1 level was 1.8–2.7%. The developed method was applied to the determination of palladium in rock samples.     

Trace Cd,Co, and Pb elements distribution during Sulcis coal pyrolysis: GFAAS determination with slurry sampling technique

A simple and fast method based on graphite furnace atomic absorption spectrometry (GF AAS) and slurry sampling technique (SlS) was developed to determine trace Cd, Co and Pb in high-sulphur coal (Sulcis, Italy) and coal chars derived at 600, 750 and 950 °C under N₂ atmosphere for developing a clean coal for electricity production. The proposed method was then coupled to a four-step sequential chemical extraction method for assessment of metals distribution in coaled samples. The slurries were prepared by varying sample mass (1–50 mg), volume (1–3 mL) and kind of dispersing medium (1% v/v Triton X-100 and 2 N HNO₃), and sonication time (5–30 min). Pyrolysis/atomization temperatures as well carrier gas flow rate were optimised. Pd(NO₃)₂ and NH₄H₂PO₄ were employed to stabilize Cd and Pb, respectively, in the pyrolysis stage of furnace program. The use of HNO₃ as dispersing agent was found to be effective in lowering the high level of background absorption on the Cd analytical signal produced by raw coal matrix. Conversely, coal charred samples did not show significantly level of background interferences, so that Triton X-100 dispersing agent could be used for all analytes. Calibration curve against acid-matched standards was allowed for Cd, whereas the standard addition calibration was used for Co and Pb owing to chemical matrix interferences. The precision, expressed as relative standard deviation (% RSD, n = 5), was better than 5% for Cd, Co, and Pb at 1, 10, and 15 μg L^? 1 levels, respectively. The accuracy of the analytical method was checked by analyzing a BCR No. 182 steam coal certificated reference material and the results were in good agreement with certificated and informed values. The solid detection limits (3σ_blank) were as low as 0.001 Cd, 0.01 Co, and 0.01 Pb mg kg^? 1, using 30 mg sample mass and slurry concentration of 30 m v^? 1 for Cd, and 50 mg sample mass and 50 m v^? 1 slurry concentration for Co and Pb. The content of elements in Sulcis coal was found to be 0.33 Cd, 4.0 Co, and 3.8 Pb mg kg^? 1 and mostly associated to sulphates and di-sulphides as indicated by the leaching test. Under pyrolysis conditions Cd significantly volatilised (about 64%) at temperature higher than 600 °C, whereas residue chars at 950 °C are enriched in Co and Pb up to 20%. The proposed method is suitable for routine metals monitoring in coaled samples.     

Electrothermal vaporization coupled with inductively coupled plasma array–detector mass spectrometry for the multielement analysis of Al2O3 ceramic powders

An electrothermal vaporization (ETV) system useful for the analysis of solutions and slurries has been coupled with a sector-field inductively coupled plasma mass spectrometer (ICP–MS) equipped with an array detector. The ability of this instrument to record the transient signals produced for a number of analytes in ETV–ICP–MS is demonstrated. Detection limits for Mn, Fe, Co, Ni, Cu, Zn and Ga are in the range of 4–60 pg μL^− 1 for aqueous solutions and in the low μg g^− 1 range for the analysis of 10 mg mL^− 1 slurries of Al₂O₃ powders. The dynamic ranges measured for Fe, Cu and Ga spanned 3–5 orders of magnitude when the detector was operated in the low-gain mode and appear to be limited by the ETV system. Trace amounts of Fe, Cu and Ga could be directly determined in Al₂O₃ powders at the 2–270 μg g^− 1 level without the use of thermochemical reagents. The results well agree with literature values for Fe and Cu, whereas deviations of 50% at the 90 μg g^− 1 level for Ga were found.     

Application of modified multiwalled carbon nanotubes as solid sorbent for separation and preconcentration of trace amounts of manganese ions

In this work, the potential of modified multiwalled carbon nanotubes for separation and preconcentration of trace amounts of manganese ion is studied. Multiwalled carbon nanotubes were oxidized with concentrated HNO₃ and then modified with loading 1-(2-pyridylazo)-2-naphtol. Mn(II) ions could be quantitatively retained by modified multiwalled carbon nanotubes in the pH range of 8–9.5. Elution of the adsorbed manganese was carried out with 5.0 mL of 0.1 mol L^?1 HNO₃. Detection limit is 0.058 ng mL^?1 and analytical curve is linear in the range of 0.1 ng mL^?1–5.0 μg mL^?1 in the initial solution with a correlation coefficient 0.9977 and the preconcentration factor is 100. Relative standard deviation for eight replicate determination of 0.5 μg mL^?1 of manganese in the final solution is 0.41%. The effects of the experimental parameters, including the sample pH, flow rates of sample and eluent solution, eluent type, breakthrough volume and interference ions, were studied for preconcentration of Mn(II) ions in detail to optimize the conditions. The method was successfully applied for separation, preconcentration and determination of manganese in different samples.     

Photochemical vapor generation of carbonyl for ultrasensitive atomic fluorescence spectrometric determination of cobalt

UV photochemical vapor generation (photo-CVG) as sample introduction was first adapted for determination of ultratrace cobalt by atomic fluorescence spectrometry (AFS). Cobalt volatile species can be generated when the buffer system of formic acid and formate containing Co (II) is exposed to UV radiation. The generated gaseous products were separated from liquid phase within a gas–liquid separator and then transported to AFS for determination of cobalt. Factors affecting the efficiency of photo-CVG were investigated in detail, including type and concentration of low molecular weight (LMW) organic acid, buffer system, UV irradiation time, reaction temperature, carrier gas flow rate and hydrogen flow rate. With 4% (v/v) HCOOH and 0.4 mol L^− 1 HCOONa buffer solution, 150 s irradiation time and 15 W low pressure mercury lamp, a generation efficiency of 23–25% was achieved. A limit of detection (LOD) of 0.08 ng mL^− 1 without any pre-concentration procedure and a precision of 2.2% (RSD, n = 11) at 20 ng mL^− 1 were obtained under the optimized conditions. The proposed method was successfully applied in the analysis of several simple matrix real water samples.