1i: Comparable measurement results for mercury analysis and speciation

Mercury is a highly volatile element and a potentially useful tracer for cosmochemical studies, but the controls on Hg concentration and speciation in meteorites are poorly understood. Previous studies reported surprisingly large variations in Hg concentrations (~101–104 ppb) that did not follow the general volatile depletion trend of other elements. Some studies proposed terrestrial contamination as the origin for the high Hg concentrations, while others argued against this explanation. A key factor to solve this conundrum is the form, in which Hg occurs in meteorites. In addition, mass-dependent (MDF) and mass-independent fractionation (MIF) of stable Hg isotopes can provide additional constraints on the origin and modification of Hg in meteorites.

Here, we present data from three different meteorite groups (ordinary and carbonaceous chondrites, eucrites). Concentrations were determined by combustion AAS, isotope ratios by CV-MC-ICP-MS on acid digests and/or after pre-enrichment using a dual-stage combustion oven, and Hg species by pyrolysis--thermo-desorption-AAS (PTD), in which the Hg release was continuously recorded up to 700°C under N2 gas flow. A wide range of Hg concentrations from 2.8 ppb (NWA 2714) to 88420 ppb (Orgueil) was found without systematic trends with meteorite class. MDF (δ202Hg) relative to NIST-3133 varied from -3.55‰ (Orgueil, CI) to +0.14‰ (Bouvante, eucrite), but no systematic trends with Hg concentration or meteorite class were apparent. Small but resolvable MIF (Δ199Hg) was observed from -0.07‰ (Bouvante) to +0.15‰ (Orgueil) with a negative correlation between δ202Hg and Δ199Hg. This is a typical feature for MIF caused by the nuclear volume effect.

All TD Hg release curves of ordinary chondrites showed a uniform single peak maximum temperature between 180 and 200°C. Based on comparison with Hg(0) incubated minerals, this may indicate the release of metallic Hg trapped in pores of the mineral matrix. TD measurements of carbonaceous chondrites present a more nuanced picture: Orgueil shows two distinct release peaks at ~200 and ~280°C representing 40% and 60% of total Hg release, respectively. In contrast, Allende and Vigerano yielded a main peak with an extinction maximum at 180-190°C and a small peak shoulder at ~230°C. The first Orgueil peak falls between standards of Hg(0) adsorbed on Fe(OH)3 and β-HgS (metacinnabar). The second Orgueil peak appears slightly earlier than our α-HgS (cinnabar) standard, which could indicate the presence of microcrystalline α-HgS. Our data on Hg concentrations, species and isotopic variations in meteorites argue against terrestrial contamination and provide new constraints for cosmochemical studies.

Significant levels of mercury (Hg) are present in the underground storage tanks at the US Department of Energy Savannah River Site, and removal of the Hg from the radioactive liquid waste stored in these tanks is a major concern and challenge. The species of Hg in the tank wastes are suspected to have changed over time, resulting in changes in the stabilization efficiency of Hg in the solid waste products (e.g., Saltstone). It was originally believed that the Hg in the tank wastes was present primarily as HgO and Hg(II). Increasing Hg concentrations from leachate characterization testing (TCLP) of the saltstone triggered an investigation. Initial testing of the saltstone waste feed showed a significant amount of organomercury species. As individual tanks were further tested, poor mass balance was often observed between the total Hg concentration and the sum of the Hg species that were measured, resulting in ambiguous data interpretation. Since the efficacy of the treatment process is dependent on the molecular forms of Hg present in the waste, the site contractors investigated whether alternative analytical techniques would yield a better mass balance.

Diluted tank waste samples were analyzed for total Hg, dissolved Hg, inorganic Hg, elemental Hg, methylmercury, ethylmercury, and dimethylmercury by Brooks Applied Labs (BAL) using methods based on CV-AFS technology, which have been in use for many years. BAL also developed alternative analytical techniques for the simultaneous analysis of inorganic Hg, methylmercury, ethylmercury and phenylmercury using ion pairing chromatography with cold vapor generation and inductively coupled plasma mass spectrometry detection (IP-CV-ICP-MS) and dimethylmercury by reverse phase chromatography (RP-ICP-MS). Additionally, methylethylmercury and diethylmercury reference materials were prepared as quality control standards. Results from the two methods are compared and discussed.

Accurate quantification of atmospheric mercury (Hg) concentrations are critical for understanding the Hg biogeochemical cycle and for the development of effective policy and management. However, a number of recent studies have revealed uncertainties in the accurate measurement of gaseous oxidised mercury (GOM), where biased low measurement has shown to occur with the use of KCl-coated annular quartz denuders traditionally used for atmospheric GOM measurement. Additionally, there is also evidence to suggest that environmental parameters affect the accurate quantification of GOM, including relative humidity and tropospheric ozone (O3) concentration. In this study, we present a method comparison between KCl-coated annular quartz denuder based methods and a novel method which utilises cation exchange membranes and nylon membranes to measure total GOM and differentiate GOM species. We additionally explore the dependence of GOM on environmental variables, and investigate the variables which control GOM events. This study presents the first of its kind within the Southern Hemisphere, and is undertaken at an urban site in Sydney, Australia, from the period April 2016 onwards. Mean background concentrations at this location as determined by the UNRARM system are 20.2 pg m-3, and 5.4. The findings from this study aim to assist in the development of more accurate GOM sampling methodology, as well as providing a greater understanding of the ways in which environmental parameters influence GOM measurement.

Within the EMRP (European Metrology Research Programme) project “MeTra” a mercury vapour generator is being developed to establish SI traceability for mercury vapour measurement results in the range from 5 µg Hg/m3 up to 55 µg Hg/m3.

The novel calibration facility generates a tuneable mercury vapour directly traceable to primary mass standards. The facility can be used for the calibration of mercury monitors as well as for the loading of sorbent traps.

The facility consists of a modified type of diffusion tube, a new measurement method to weigh the loss in (mercury) mass of these diffusion tubes during use (with a minimum 6-8 μg mass difference between successive weighings), and a new housing for the diffusion tubes to optimize flow characteristics and to minimize temperature variations and adsorption effects.

The specifics of the newly introduced capabilities, to calibrate mercury monitors at levels of 5 ng Hg/m3 and higher, will be highlighted. This new calibration service is especially important for measurement services testing ambient air monitoring (1 – 2 ng Hg/m3), indoor and work place related mercury content levels according to health standards (from 50 ng Hg/m3) as well as mercury vapour concentrations relevant to stationary source emissions (upwards of 1 µg Hg/m3).

Measurements of MeHg in sediment and soil samples represent a challenge. This is due to the nature of the matrix, high ratio between inorganic mercury (Hg) and MMHg and risks of artefact MMHg formation during extraction and measurement procedures. The latter is of particular concern due to the high inorganic Hg presence in sediments. Simple extraction method for quantitative separation of MMHg from inorganic mercury has been developed.

The method consists of alkaline leaching with 1N KOH in ethanol from a sediment (usually 0.2-0.5 g of wet weight and 0.2 g of dry sample) followed by addition of NH2OHHCl and EDTA-4Na. Sample is then slightly acidified with 1N HCl and purged with nitrogen gas to remove possible S2+ ions as H2S. The next step involves derivatisation by dithizone and quantitative extraction of both MMHg and inorganic Hg dithizone complexes into toluene. Hg(II) and MMHg dithizonates are then separated on silica gel column and selectively collected into separate test tube. The collected sample is evaporated to dryness and digested with mixture of acids (HNO3/HClO4 and H2SO4) and measured by CV AAS. The limit of detection and limit of quantification are 0.2 ng/g and 0.67 ng/g, respectively.

Separation of inorganic and organic mercury is quantitative, independent on inorganic mercury content and absence of artefact MMHg formation. The method was validated using certified reference materials (CRMs).

Compared to other methods this method is simple and inexpensive from the point of view that the instrumentation which is used for total mercury measurements can be used to measure organic mercury. The detailed analytical methods as well as the comparison of MMHg in various sediment samples between conventional methods and newly developed method will be presented.

Changes in climate, global and local mercury emissions into the environment may influence the cycling of methylmercury (MeHg) and its bioaccumulation in marine organisms. Following the decrease of atmospheric Hg emissions or discharge into rivers a decline in biota MeHg concentrations is expected. Analysis of historic biota samples for their MeHg concentrations could offer a pathway to unravel temporal patterns. Because copepods are a strong link between primary producers and upper trophic levels it is important to better understand their MeHg bioaccumulation. There is a vast depository of biological samples preserved in formalin, but it is unclear if these samples can be used for MeHg measurements. We tested for temporal changes of MeHg concentrations in formalin-preserved copepods. Because the life span of many copepod species is on the order of a weeks to months, and MeHg that is highly assimilated (>90%) is lost very slowly, copepod MeHg represents recent cumulative exposure. We tested the null hypothesis that formalin preservation does not influence MeHg concentration in copepods over time. We collected two copepod species Acartia tonsa and Temora longicornis from a local embayment in Connecticut, US. These species were selected to represent smaller and larger size classes of copepods found in coastal embayments. Copepods (n=3 replicates of 50 individuals) were picked prior to formalin preservation and stored frozen on a polycarbonate filter. The remaining copepods were preserved with formalin for time intervals e.g. 1 week, 1 month, 6 months and up to a year. At the end of each preservation period, copepods were filtered via polycarbonate filter and kept frozen prior to analysis. Filters with copepods were acid-digested and analyzed for MeHg by direct ethylation of digest aliquot diluted in deionized water. Initial values were 0.20 pg per an individual for the small copepod, A. tonsa (average prosome length or PL: 0.9 mm) and 0.30 ng per copepod for the larger T. longicornis (average PL: 1.2 mm). MeHg concentrations dropped to ~ 25% of initial values within two weeks (T. longicornis) to a month (A. tonsa), and subsequently increased by a factor of ~7 within 6-12 months for A. tonsa and 1.5 for T. longicornis. The mechanism for these dramatic increases has not been tested yet. Additional research on abiotic MeHg formation during formalin incubation is needed.

Stable Hg isotope analysis is now widely used in environmental sciences, essentially for discriminating Hg sources and better understanding its bio-geochemical cycle. Up to now, total Hg isotopic composition has been mainly studied. Further, the compound specific Hg isotopic analysis (CSIA) is an emerging tool providing the isotopic signature of each different Hg species such as inorganic Hg (InHg) and methyl Hg (MeHg). However, one of the main challenges remains the low Hg concentration in most environmental compartments. In this work, we propose to overcome this limitation using an on line Hg pre-concentration strategy combined with gas chromatography (GC) hyphenated to MC-ICP-MS.

The GC is here fitted with a programmable temperature volatilization (PTV) injector and a liner filled with a specific solid phase. This allows the injection of a large volume of sample (up to 100 times larger than conventional injector) and a pre-concentration of the Hg species in the liner before their separation and detection. The pre-concentration and separation parameters such as the inlet and GC temperature during the injection phase, the gas flow and duration of the venting period, were first optimized using a regular quadrupole ICP-MS instrument. When transferring the method to MC-ICP-MS, the main challenge was to deal with very short transient signals (down to 11 s) and this aspect was solved using a suitable data treatment strategy. Contrary to previous GC/MC-ICP-MS methods, we also demonstrated that our method required a compound specific standard bracketing procedure due to the nature of the pre-concentration involved.

The main advantage of the proposed methods is the high pre-concentration capability which allows the compound specific isotopic analysis with InHg or MeHg concentrations down to 50 ng.L-1 in liquid matrices and 200 ng.g-1 for solid samples (after extraction). At this level, the method gives a good accuracy and a maximum uncertainty of 0.30‰ (as 2SD) for both δ and Δ.

We will discuss the main advantages and limitations of the proposed method, such as the potential pre-concentration capabilities and associated limits in term of Hg concentration. We’ll also present the critical steps of technical validations including the analysis of environmental CRMs and comparison to the actual reference technique.

Rice is a staple food for many countries around the globe, but it is also known to accumulate MeHg the most toxic mercury species to humans. Therefore, accurate determination of this species in rice ismandatory.However, while a certified Reference material (CRM) exists for total Hg in rice,there is noCRMfor MeHg in a rice matrix. In earlier work, we could show that the % of MeHg in rice can vary substantially. Therefore, we aimed tocreate alaboratory reference material(LRM) for our purposes,in order toassess theperformance ofroutine measurements.Onekg of long grain white rice was purchased from a local supermarket and ground into a fine powder and mixed well. It was thenanalysedrepeatedly,using a TMAH/HCl digestion followed bypre-concentration CV-AFS.MeHg and Hg2+were separated by RP-LC and on-line chemical oxidation and reduction with acidified Br-/BrO3-and SnCl2was used prior to AFS detection. This method hasbeen proven to be accurate and precise by cross-calibration using species specificisotope dilution mass spectrometry.

For the determination of precision and accuracy, the rice powder was split into several sub-sets andanalysed repeatedlyto assess variation between the sub-sets. The measurement of the samples gave a concentration range between 3.05mg kg-1and 3.68mg kg-1, averaging at 3.36mg kg-1.Further analysis will include the testing of the materials shelf-life.

Coal is apt to concentrate mercury and other toxic elements whose content is governed by geological and geochemical specificity of the coal basins. The mercury concentration in coal varies in a wide range from < 1 ppb to 300 ppm. Metamorphism and subsequent hydrothermal processes generate different mercury species in coal and in coal-bearing strata. The mercury species having different matrix binding energy can be determined by thermospeciation technique based on the study of mercury release from a sample during its gradual heating. Total mercury concentration and mercury thermospeciation in coal were determined using RA-915M Zeeman atomic absorption spectrometer. We studied features of the mercury distribution in samples of brown and hard coals with different metamorphism degree of the organic matrix.

Total mercury concentration in the studied coal samples lies in a wide range of less than 2 to 7600 ppb. The thermoscanning reveals various mercury thermospecies in coal. The low-temperature peaks, below 200 OC, can be attributed to mercury bound by physical sorption, and the mid- and high- temperature peaks to mercury bound to organic coal matrix, sulfides (mainly pyrite), and silicates. Low-temperature forms of mercury evaporating at 240-350 OC prevail in brown coal. The most high-temperature species releasing at 640 900 OC were found in shungite.

The scanning electron microscopy and electron probe microanalysis reveal peculiarities of mercury and other chemical elements distribution in organic matrix and inorganic impurities. The mercury accumulation in coals was found in organic matrix, sulfides, iron oxides and hydroxides, clay minerals and phosphates. The scanning electron microscopy helped us to identify the following terrigenous minerals within coal matrix: quartz, potassium and sodium feldspars, hydromica, kaolinite, chlorite, rutile, apatite, and sphene. Correlation analysis has shown that the mercury concentration correlates with the elements of the halophile group: Fe, Mo, Ag, Ni, Zn, Sb, Cu. Correlation coefficients vary from 0.74 to 0.94. Mercury has significant correlation with V, Ge, K, and Al, which can be explained by the effective mercury Hg(0) sorption by clay minerals; this finding is supported by the thermoscanning data.

This research is supported by the National Research Foundation of South Africa and the Russian Foundation for Basic Research, project № 14-05-93961.

We recently developed a passive air sampler (PAS) for gaseous mercury, and determined a sampling rate (SR; defined as the volume of air stripped of analyte per unit time) of 0.121 ± 0.005 m3 day−1 in a year-long, outdoor calibration in a temperate environment. A model based on Fick’s First Law of Diffusion predicted a SR within 10% of this calibrated value. Here, we present results from a series of controlled laboratory studies that tested the potential impacts of variable wind speed, relative humidity, and temperature on the SR of the PAS. We find a logarithmic relationship between wind speed and SR, with the greatest impacts at wind speeds less than 1 m s-1, which are not common for outdoor deployments. When we consider only data from wind speeds between 1 and 6 m s-1 we find a significant positive linear relationship (p < 0.001), indicating a 0.003 m3 day-1 increase in SR (corresponding to 2.3% of the previously calibrated SR) for every m s-1 increase in wind speed. Similarly, SR increases linearly with temperature in the range -15 to +35 °C (p < 0.001). This temperature dependence can be fully explained by the effect of temperature on the molecular diffusivity of gaseous mercury in air, which is calculated to lead to a 0.008 m3 day-1 increase in SR (6.9% of the calibrated SR) per 10 K increase in temperature. SR is not significant affected by relative humidity, tested between 44 to 80% at 20 °C (p = 0.08). While differences in SR caused by wind speed and temperature are relatively small, the accuracy of passively sampling gaseous Hg in air can be improved by adjusting SRs using the slope of the relationships described here, if measured or estimated temperature and wind speed data at or near sampling sites are available.

In response to the increasingly severe issue of mercury pollution, the EU funded Global Mercury Observation System (GMOS) project (FP7) established a global monitoring network, that in the last five years provided harmonized and detailed information on mercury ambient measurements through a dedicated database platforms (www.gmos.eu). Among others, the atmospheric mercury data obtained from the GMOS network has become an essential data source for many current and forthcoming studies as well as for the formulation and implementation of the Minamata regulatory policies. In respect to data management, within this project an ad-hoc centralized system, namely GMOS-Data Quality Management (G-DQM), was developed and used to release high-quality atmospheric mercury data collected with the Tekran speciation instrument. Based on a web application, the system is able to fulfil the demands of processing monitoring data in near real-time, verifying if instruments adhere to standard procedures, and rapidly identifying non-representative measurements. The use of automated common checks represents an improvement because it ensures consistency and reduces human bias thus avoiding misinterpretation and inappropriate data use. Despite the wide use of the Tekran speciation system, recent studies increasingly acknowledge that the atmospheric reactive mercury species, Gaseous Oxidized Mercury (GOM) and Particle-Bound Mercury (PBM), have large measurement uncertainties. Besides that, in our study we want to elucidate how significant are the quality control issues and identify how much they-self can additionally affect the resulting validated datasets. Referring to 2013 and 2014 years, during which we had a good coverage of speciation data for 7 GMOS Master stations, at the end of the whole validation process we were able to discriminate between valid and invalid speciated data. For both the years, we had an average percentage data loss of about 30%. The incidence of rejected data, in terms of discrepancy from the final quality-validated concentrations was, on annual basis, in the range from 0.02 to 4.89 pg.m-3 for PBM, and from -0.79 to 3.11 pg.m-3 for GOM, respectively. In terms of monthly means, we obtained a larger gap ranging from -0.46 to 40.23 pg.m-3 for PBM and from -0.79 to16.40 pg.m-3 for GOM, respectively. Based on GMOS monitoring data, this work thus intend to highlightthat quality control issues could significantly affect the resulting atmospheric mercury data. Furthermore, it suggests that researchers using data from any source be expected to manage data with due caution depending on the fact that they are or not quality controlled.

Dried Blood Spots (DBS) are drops of whole blood collected on standardized paper from a finger or heel prick. They have been used since the 1960s for newborn screening programs, pharmacokinetics studies, and demographic health surveys. The use of DBS for the analysis of mercury exposure in newborns has been studied previously in two studies, although both suffered from poor detection limits and challenges with accuracy and precision. Challenges in measuring mercury in DBS (as supported by the European Bioanalysis Forum) include: 1) the impact of hematocrit levels on analyte and spot formation; 2) differences between measures taken from venous and capillary blood; and 3) matrix effects. Therefore the objective of this study is to evaluate whether total mercury and methylmercury content in DBS differs between blood sources (venous or capillary), and hematocrit levels. Whole blood and DBS from venous and capillary sources were obtained from 49 healthy individuals associated with McGill University (Canada). Total mercury in the blood and DBS will be determined by acid digestion and by Dual-Stage Gold Preconcentration, and methylmercury by tetramethylammonium digestion and Gas Chromatography (GC) - Cold Vapor Atomic Fluorescence Spectrometry (CVAFS). Initial blood analysis showed no significant difference in hematocrit levels between venous and capillary blood, although a range of other clinical blood parameters are under investigation. We are currently in the midst of analyzing all samples for mercury. The results are expected to increase our understanding of DBS as a potentially novel means to assess mercury exposure. This could have immediate applications in established programs such as newborn screening and demographic health surveys.

Monitoring fish mercury spatiotemporal concentration trends is of great importance in terms of scientific research and regulatory design. A robust and reliable analysis of trends of fish mercury is not only evidence for identifying sites with elevated mercury but also crucial reference for proposing mercury mitigation strategies and developing fish consumption advisories. Unfortunately, effective analysis based on fish monitoring data is normally hindered by difficulties in distinguishing the true spatial and temporal trends from total variances in fish-mercury concentrations complexed due to heterogeneity in fish samples of diverse species and trophic levels. This problem occurs since collecting fish samples with consistent characteristics (e.g., fish species, size, and sex) over large regions or over long time is difficult. Furthermore, budgetary constraints lead to a shortage of well-designed continuous monitoring programs, which force scientists studying spatiotemporal trends of fish mercury concentrations to use composite monitoring database. Composite database commonly integrates data from different monitoring program with distinct research objectives and protocols (e.g., aiming at specific species, regions and using different fish-cut for analysis). Using a composite database would thus complicate the interpretation of long-term trends embedded within fish mercury monitoring data. Surprisingly, there has been minimal efforts to evaluate the relative performance of these methods systematically. This research aims to review and to evaluate the performance of several published and representative statistical standardization models/methods on fish mercury: ANCOVA, multivariate regression, mixed-effects regression, polynomial regression, partial least squares regression and biotic mercury partitioning index (BMPI). We employed several statistical techniques (e.g., Principle Component Analysis (PCA), MANOVA, etc.) and the concept of machine learning (i.e., cross validation) to test the accuracy and robustness of the targeted standardization methods. This research is important since any further analysis (e.g., economics valuations, policy analysis) based on monitoring data may reach spurious results and misleading conclusions if the true spatial and temporal trends of fish mercury are not properly elucidated. This study is striving to serve as a clear guide on how to choose the right model for further research involving fish mercury trend analysis.

Measurement of low methylmercury concentrations (MeHg) in seawater represents a major analytical problem due to low natural concentrations. Reference analytical method (aqueous phase ethylation coupled with gas chromatography and cold vapor atomic fluorescence spectrometry, Eth-GC-CVAFS) is widely used for MeHg determinations. Derivatisation of mercury species by hydride generation (HG), coupled with cryogenic trapping (CT), gas chromatography (GC) and cold vapor atomic fluorescence spectrometry has an advantage due to lower detection limits. Following optimization of HG-CT-GC-CVAFS for MeHg determination in seawater, we calculated measurement uncertainty budget using GUM/Eurachem guidelines for both methods. Expanded relative combined standard uncertainty (Uex,r) at the 95% confidence interval is obtained by applying a coverage factor k=2. At the highest MeHg concentration level (>80 pg/L), middle (2030 pg/L) and the lowest (<10 pg/L), Uex,r has a value of 11.1%, 15.0% and 21.3% for hydride generation, and 15.8%, 18.2% and 19.3% for ethylation, respectively. The highest relative contribution (index) to the expanded relative standard uncertainty in HG is from sample repeatability which accounts for 46.6% of Uex,r at the lower concentration level and slightly lower value at the highest MeHg level (39.9%). Similar trend is also observed in ethylation method, but the contribution indexes are lower (29.936.4%). Contribution indexes of MeHg standard repeatability are increasing with increased MeHg concentration in both methods; however, this trend is directly proportional to Uex,r decrease. Relative standard uncertainties for recovery are about 1.4% and 5% at all MeHg concentration levels for hydride generation and ethylation, respectively. Higher values in ethylation method are due to unreproducible and lower recoveries in ethylation (average 79%) caused by several procedure steps and incomplete sample extraction. At the highest MeHg concentration level (>80 pg/L), recovery has the highest uncertainty index for ethylation method (37%). Sample areas have a big contribution to Uex,r at lower MeHg concentrations for hydride generation (24.3%) due to readability of small peaks which is a consequence of 35% lower sensitivity of this method. Sample volume has three-fold higher uncertainty contribution index in ethylation than in hydride generation. This uncertainty is caused by sample losses during phase separation, back-extraction of sample, incomplete sample transfer to measuring vial and to instrument. Other uncertainty sources do not represent a significant contribution to expanded relative standard uncertainty of both methods.

Mercury is a global pollutant released into the environment from both natural and anthropogenic sources. Various mercury species differ greatly in their bio-physico-chemical properties such as toxicity and rate of bioaccumulation by organisms. Therefore the analysis of samples only for total mercury is no longer completely acceptable and provides only partial information about their impact on human health and the environment. As a consequence, considerable progress has been made in the development of techniques, which are capable of quantifying various mercury species. Future regulations on MeHg will require the standardized procedures for quantitative determination of alkylmercury species, therefore there is a strong need for development of robust analytical procedures providing reliable data on both, total mercury and its chemical species in the complex marine matrices.

Among its multiple tasks, IAEA Environment Laboratories (IAEA EL) in Monaco acts as the analytical support center for IAEA Member States laboratories and is the pillar of the Quality Assurance program for determination of contaminants in the marine environment. Several recommended methods for mercury and MeHg, based on the application of different analytical techniques are set up. Development and validation of reliable and sensitive method based on gas chromatographic separation and pyrolysis-atomic fluorescence spectroscopy (GC-py-AFS) in a variety of marine matrices is described in the the present study.

Since reliable results are still mostly dependent on sample preparation, this study aims to evaluate the efficiency of different procedures, namely microwave-assisted extraction (MAE) with HCl, TMAH KOH, alkaline digestion with KOH and enzymatic hydrolysis with protease, for the extraction of MeHg in marine samples. Determination was accomplished by hyphenated GC-py-AFS. The full validation process following international guidelines is presented and described.

The estimation of the total uncertainty associated to each measurement result was fundamental tool for sorting the main sources of measurement biases. Preliminary forecast of the uncertainty budgets was used as a strategy to ensure that determination of methyl mercury in biota samples could be achieved with demonstrated traceability to a stated system of reference within less than 3 % expanded uncertainty (k=2).

Overall, the results obtained with the developed analytical procedures are reliable, with stated expanded uncertainty, demonstrated traceability and fully validated methods. Hence, the methods presented could be recommended for its implementation in IAEA Member States Laboratories.

With the eventual ratification of the Minamata Convention on Mercury, United Nations/IAEA Member States will be required to establish or strengthen environmental mercury monitoring efforts. Among its many tasks, the IAEA Marine Environment Studies Laboratories (IAEA-MESL) in Monaco acts as the analytical support center for IAEA Member States laboratories and is the pillar of the IAEA Quality Assurance program for determination of contaminants in the marine environment. As such, IAEA-MESL has developed and validated a method for ultra-low level analysis of total mercury (THg) in seawater in order to better assist Member States laboratories with mercury analysis and mercury data quality assurance.

An analytical procedure using cold-vapor atomic fluorescence spectrophotometry (CVAFS) along with improved cleaning and sample processing methods was optimized and validated according to the recommendations of ISO-17025 standard and Eurachem guidelines. Importantly, a limit of detection and limit of quantification, after extensive cleaning and careful sample preparation, was possible at ultra-low levels (LOD = 0.01 ng kg-1; LOQ = 0.04 ng kg-1) using a relatively small sample volume (25mL). However, these values were variable over time and highly dependent on clean analytical technique. Additional parameters such as recovery, working range, repeatability, intermediate precision and combined uncertainty (28%, k=2; ~ 0.2 ng kg-1 concentration levels) were also carefully estimated.

All sample preparation steps were carried out in a clean laboratory environment under clean hoods (Class 100).

The traceability of obtained measurement results was also demonstrated. Particular attention was paid to the correct definition of the measurand in this validation study. The mass fraction of mercury dissolved in seawater (~0.20 ng kg-1) and mercury associated with solid particles (~0.22 ng kg-1) was estimated for a coastal seawater sample (~0.43 ng kg-1) from the Mediterranean Sea and the respective mass balance provided.

As part of the method validation, sample preservation methods, sample preparation methods and stability of samples were also investigated. As was found in previous research, preservation of seawater samples with 0.1% (v/v) nitric acid was acceptable for short-duration (up to 1 month) holding times. In this instance, nitric acid proved effective for sample stability for both unfiltered and filtered sample processing at concentrations around 0.47 ng kg-1 and 0.20 ng kg-1, respectively.

Additional validation of the proposed analytical procedure was achieved by participation in an Inter-Laboratory Comparison (ILC) study. The results provided were in excellent agreement with the assigned values of the ILC.

The simple portable device for calibration of mercury analyzers is offered. The device uses a known method introduction of microvolumes of saturated mercury vapors in analytical cell. The basis of the method - known concentration of mercury vapor in unit volume at a specific temperature and pressure. In the proposed device the glass conical flask with mercury installed in thermal insulation box with melting ice, and microvolumes of saturated vapors sampled at 0°C. The concentration of mercury vapor at 0°C is 2.205 ng/ml. The dosage of saturated mercury vapors was performed by conventional insulin syringes with volume of 1 ml. The minimum volume for these syringes is 0.1 ml, i.e. at 0°C and 760 mm Hg. 220 pg of Hg. Using gas-tight chromatographic syringes with volume of 0.05 ml and below it is possible to dose amounts of mercury since 10 pg. The closed 75 ml flask is completely saturated with mercury for 4 minutes. The effect of reduction of mercury concentration in flask at selection intervals more than 30 s is absent. For stable results it is necessary to carry out preliminary training of syringes and introduce amendments to the atmospheric pressure. The assessment of possible diffusive losses of mercury from a syringe needle during transfer of vapors to analyzer was made. Reproducibility of various volumes dosage does not exceed 2%. Requirements to carrying out of calibration in pg range are developed. Filled with ice device keeps the temperature not less than 8 hours. This simple and cheap method provides precise, accurate and reproduced calibration. The device can be used in field conditions.

Mercury is widely used in compact fluorescent lamps. The demand for these lamps is rising due to CFLs energy efficiency. Increased use of CFLs raised concerns due to potential exposure to toxic mercury vapour when broken indoor. Moreover, inappropriate storage, handling and recycling causes environmental concerns. Fully traceable data on total mass of Hg released after CFL breakage and validated measurement methods is currently lacking.

Mercury was trapped (preconcentrated) under constant flow of Hg free gas (i.e. nitrogen, argon, air) by several absorbers including gold traps and activated carbon traps. The most frequently used method for pre-concentration of Hg is based on amalgamation of elemental Hg using gold (Au) traps, which is followed by thermal desorption of Hg at 600 C and detection by CV AAS or CV AFS. The method is suitable for lower masses of mercury (up to 1000 ng). The precision of measurements is less than 15 %. Memory effects and passivation of Au traps can lead to loss of mercury or reduction in the amalgamation efficiency. Au coated silica sand is mostly used as absorbing media, however other noble metals as Au-Pt can be used to pre-concentrate Hg. Instead of silica Al2O3 can be used as Al2O3 is more stable at high temperatures. Au and Au-Pt coated Al2O3 sand were also used during experiment, however further research is still needed.

For higher Hg levels (in range of ng and g) traps with activated carbon are widely used. Mercury adsorbed on carbon from traps is further analysed by pyrolysis and CV AAS, CV AFS, and mass spectrometry (MS). Neutron activation analysis (NAA) is also used as a reference method. Activated carbon traps impregnated by sulphur or chlorine were tested for trapping elemental mercury in ng and g levels.

The main goal of this work was to develop an easy, on-line, sensitive and traceable method for the determination of released Hg from broken CFL bulb. A procedure involving a gas tight Plexiglas box with bulb crushing system, heated gold amalgamation trap coupled to a CV AAS mercury analyser with Zeeman background correction was developed. Calibrations of the CVAAS detection were conducted using diluted NIST 3133 Hg standard solution where the mass fraction of mercury is certified and traceable. Certified reference materials were used to check performance of analytical methods used.