Mercury is a naturally occurring element found in air, water and soil. It exists in three different forms, elemental or metallic mercury, inorganic mercury compounds, and organic mercury compounds. Coal-burning power plants are the largest source of mercury emissions to the air caused by humans in the United States, accounting for more than 40% of all domestic anthropogenic mercury emissions.1 Additional mercury emission sources include the burning of hazardous waste, chlorine production, mercury product breakage, mercury spillage, metal processing, and cement production.
Mercury travels in the air and is eventually deposited into water where microrganisms and abiotic reactions convert it to methyl mercury. Methyl mercury builds up in fish and shellfish, which constitute the main sources of human exposure to this highly toxic element. Exposure level depends on the fish consumption rate, the individual’s body weight, and the level of methyl mercury found in specific fish species consumed. Methyl mercury can reach extremely high levels in predatory fish such as swordfish, king mackerel, and shark.
Methyl mercury exposure at high levels can harm the brain, heart, kidneys, lungs, and immune system of people of all ages. Methyl mercury is acutely toxic to humans because of its ability to cross the blood/brain barrier. It has also been demonstrated that high levels of methyl mercury in the bloodstream of unborn babies and young children may harm their developing nervous systems, having a negative impact upon learning and cognitive abilities.
In response, U.S. regulatory bodies have enforced strict legislation to specify maximum allowable concentrations of methyl mercury in fish and to monitor mercury emissions to the air.
The Food and Drug Administration (FDA) has established action levels for poisonous or harmful substances in human food and animal feed.3 The regulation specifies a maximum level of 1 mg/kg methyl mercury in edible portions of fresh, frozen or processed fish, shellfish, crustaceans and other aquatic animals. This action level represents the limit at or above which the FDA will take legal action to prohibit imports and remove products from the market. The Environmental Protection Agency (EPA) has introduced a methyl mercury guideline that recommends a limit on mercury consumption based on body weight, more specifically 0.1 mg/kg body weight per day.
The EPA has also issued several stringent regulations to reduce mercury pollution. The final regulations for municipal waste combustors substantially reduced mercury emissions to the air by as much as 91% between 1990 and 2005. Information on compliance with the emission standards for medical waste incinerators (MWIs) indicates that mercury emissions from MWIs are 95% lower than in 1990. The final standards for mercury from chlor-alkali production, introduced in 2003, were projected to cut mercury emissions from point sources at these facilities by 74% and total mercury emissions by about 11% compared with 1999 levels.
In 2005, the EPA issued the Clean Air Mercury Rule to permanently cap and reduce mercury emissions from coal-fired power plants.5 This rule makes the United States the first country in the world to regulate mercury emissions from utilities. In its most recent regulation, the EPA aims to reduce mercury emissions from industrial boilers by 17% from 12 tons to 10 tons per year.
Due to these stringent regulations and the growing public food safety concerns, raw materials suppliers, food manufacturers and processors, government organizations and private food testing laboratories must analyze mercury in fish. The most established method for total mercury analysis is vapor generation atomic absorption (AA) spectrometry, which has emerged as the method of choice for routine, reliable total mercury screening. For samples contravening the legislative value of 1 mg/kg for total mercury, a speciated determination is likely to be required to determine the methyl mercury concentration.
Using powerful techniques such as high-performance liquid chromatography and inductively coupled mass spectrometry (HPLC-ICP-MS) or gas chromatography-ICP-MS can achieve this, but such determinations add a considerable degree of complexity and expense. The proportion of samples requiring further speciated analysis is relatively small and so vapor generation AA is therefore an ideal, cost-effective, first-line screening approach that, in many cases, eliminates the need for full-speciated analysis.
AA Spectrometry Capabilities
Most commonly, food analysis requires a fast, robust, flexible, and fully traceable technique capable of accommodating high sample throughput and covering a wide analytical range from low parts per billion (ppb) levels to percent levels. As a versatile method with flame, furnace, and vapor generation atomization options, AA spectrometry meets all of these criteria.
The technology caters to the upper concentration range using flame and autodilution and caters to lower levels with graphite furnace. Additionally, vapor generation provides an atomization method offering ppb-level analysis of hydride- and vapor-forming elements including arsenic, selenium, antimony, and mercury. Approximately 62 metallic elements are measurable using an AA spectrometer with a typical measurement precision of around 1% relative standard deviation (RSD) and good detection limits in the range between 0.0001 and 1 mg/L.
For laboratories interested in total mercury measurements, research has shown that vapor generation AA spectrometry provides fast and accurate sample analysis with detection limits of 0.07 ppb (mg/L) in solution. This equates to 0.014 mg/kg in a fish sample, based on a 0.5g in 100 mL preparative method. These levels are a suitable screening method for analyzing mercury in fish samples to highlight samples contravening legislative levels. We developed an experiment to demonstrate the capability of AA spectrometry to achieve precise, dependable analysis of low mercury levels in fish.
We used a Thermo Scientific iCE 3500 AA spectrometer combining high-precision optics, state-of-the-art design and user-friendly software for our analysis. The spectrometer was coupled to a VP100 vapor generation accessory using a continuous flow system to produce a steady-state signal, providing excellent analytical precision. The continuous flow of reagents ensured that the system was self-cleaning, reducing memory effects and increasing sample throughput.
The VP100 accessory was entirely controlled by the Thermo Scientific SOLAAR software, simplifying set up of the method and execution of the analysis. A mercury cell provided as standard with the VP100 was also used. This accessory provided an increased path-length compared to a normal vapor cell, offering exceptionally low detection limits.
The method was evaluated using both spiked fish samples and certified reference materials containing mercury levels relevant to current global legislation. Three different types of fish were evaluated. These were fresh salmon obtained from a local supermarket, canned sardine also obtained from a local supermarket, and dogfish muscle tissue certified reference material, DORM-2, supplied by the National Research Council of Canada, Institute for National Measurement Standards, Ottawa, Canada.
For sample preparation, we followed a four-step procedure including sample drying, sample preparation, sample digestion, and mercury reduction.
The sample drying phase may not be applicable to all situations, as it is only necessary in cases when the final mercury concentration is needed as a dry weight value, e.g., mg/kg dry weight. If dry weight measurements are needed, then fish samples should be ground in a mortar and pestle and dried in an oven at 80°C until they reach a constant weight. After drying, portions of approximately 0.5 g should be weighed out accurately for digestion.
Most countries and official regulatory bodies specify concentrations of mercury in a wet weight of sample. For wet weight measurements, fresh fish should be homogenized in a food processor and a portion of approximately 0.5 g should be accurately weighed and placed in a microwave digestion vessel. This provides a representative fish sample. For this particular experiment, 1 mL of 1,000 ppb Hg standard solution was added to half of the salmon and sardine samples. This spike gave a concentration of 10 ppb Hg in the final 100 mL sample. The other half of the samples did not have mercury added to them to allow for the calculation of spike recoveries.
The microwave digestion vessels containing the samples were placed in a fume extraction hood before adding 10 mL concentrated HNO3. The vessels were left for at least 30 minutes without their lids on to allow gases to escape. The vessels were subsequently digested using a microwave digestion system. Alternatively, digestion could have been performed using a hotblock.
Following digestion, the samples were transferred to a 100 mL graduated flask and 60 mL of 6% potassium permanganate solution was added. Sample vessels were left for at least two hours to ensure that all the mercury in the sample was reduced to Hg2+. It was very important to make sure that the vessels were not sealed at that stage, as gases produced could cause pressure to build up.
After the mercury was reduced, 15 mL of 20% hydroxylamine chloride solution was added to remove the excess potassium permanganate. It was essential to add the hydroxylamine chloride slowly during that stage and to gently mix the solution during the addition. The solution was then allowed to cool and deionized water was added to increase the volume up to 100 mL.
We prepared standards from a 1,000 parts per million (ppm; mg/L) mercury standard solution, which was first diluted to produce a 1,000 ppb (mg/L) stock solution to allow simple preparation of a range of standards. To demonstrate the linear range of AA spectrometry, we used a wide range of standards (1 to 100 ppb). The standards were matrix matched and prepared in the same order as the samples.
The VP100 requires both a reductant and an acid solution to perform the reactions that form the gaseous mercury. For this application, the reductant was a solution of 7.5% stannous chloride (SnCl2) stabilized in 10% HCl. The acid solution was 50% HCl.
The analysis was performed using the most sensitive absorption wavelength for mercury at 253.7 nm. Five replicate analyses were used, with each replicate taking four seconds, to assess thoroughly the short-term stability of the instrument during the development of this method. For normal use, three replicates would be adequate. Deuterium background correction was employed throughout the analysis. The VP100 and spectrometer parameters are shown.
The calibration curve showed excellent linearity up to 100 ppb, which is equivalent to 20 mg/kg in a fish sample weighing 0.5 g. The R2 value was 0.9989, demonstrating the superb performance of AA spectrometry over a wide concentration range. This calibration is equivalent to concentrations of 0 mg/kg to 20 mg/kg mercury in the original fish samples, assuming a sample mass of 0.5 g. The percent RSDs for each of the standards were less than 2.5%, demonstrating the excellent stability of both the spectrometer and the VP100 accessory.
We calculated the method detection limit (MDL) and characteristic concentration using the automated instrument performance wizard in the SOLAAR software. This feature is able to guide analysts through the steps necessary to quantify the performance of a method. It also automates all of the data processing, making the entire procedure quick and easy. The method was found to have a detection limit of 0.068 ppb (mg/L) in solution. This equates to a MDL of 0.014 mg/kg in the original fish sample, assuming a sample mass of 0.5 g. The MDL provides a measure of the noise and stability of the system. A lower detection limit allows for confident determination of lower mercury concentrations. The characteristic concentration is related to the sensitivity of the method. The characteristic concentration of this method was found to be 0.724 ppb in solution. This would be the equivalent of 0.145 mg/kg in the initial fish sample, assuming a sample weight of 0.5 g.
Salmon and sardine samples were spiked with 10 ppb mercury prior to digestion and compared with unspiked samples to calculate recoveries. These 10 ppb spikes would correspond to a concentration of 2 mg/kg in normal fish samples assuming a sample weight of 0.5 g and demonstrate the accuracy of the analysis at levels appropriate to current legislation. The agreement with expected results was excellent, with the recovered values all falling within 6% of the expected values. This demonstrated the repeatability and accuracy of both the sample digestion procedure and the vapor analysis using AA spectrometry.
To ensure the accuracy of sample preparation, digestion, and analysis, three separate samples of the DORM-2 standard reference material were also examined. The recoveries from these samples were excellent, with an accuracy of ±2% or better.
AA spectrometry coupled with a VP100 vapor generation accessory is the most comprehensive technology available for the analysis of trace levels of mercury in fish, combining speed, excellent linear range, stability, and accuracy. Analyzing samples in only 90 seconds, the method meets the high throughput requirements of laboratories while also offering superior sensitivity, full traceability and ease-of-use.
Dickson is an applications chemist at Thermo Fisher Scientific. For more information, reach her at hazel.dickson@ thermofisher.com or +44 (0) 1223 347 400.
- Environmental Protection Agency, Mercury, Basic Information, www.epa.gov/mercury/ about.htm.
- Environmental Protection Agency, Mercury, Health Effects, www.epa.gov/mercury/effects.htm.
- Food And Drug Administration, Industry Activities Staff Booklet, August 2000, Action Levels for Poisonous or Deleterious Substances in Human Food and Animal Feed, www.cfsan.fda.gov/~lrd/fdaact.html#merc.
- Public Affairs Television, Science and Health, The Mercury Story, www.pbs.org/now/science/mercuryinfish.html.
- US Environmental Protection Agency, Air and Radiation, Clean Air Rules of 2004, Clean Air Mercury Rule, www.epa.gov/camr/.