Sampling Procedures
Currently, there are no generally available methodologies for measuring low levels of mercury in the ambient air. Many of the ambient measurement efforts for mercury that have been made to date have occurred as part of research studies and provide data of limited usefulness in assessing broader questions relating to geographical sources and deposition of this metal. The USEPA plans to issue a Compendium of Inorganic Ambient Monitoring methods in the near future which contains a recommended method (IO-4) for measuring mercury vapor, particulate mercury and mercury in precipitation. The forthcoming Compendium method will most likely contain components of the procedures discussed below.
Vapor phase mercury can be collected using adsorbent media normally used for collecting volatile organic compounds. Activated carbon has been found to efficiently collect mercury vapor under a variety of conditions. Gold coated sand adsorbent traps and gold annual denuders are available for collecting vapor phase mercury as an amalgam (Xiao et. al.,., 1991). Denuders have been proposed to avoid the unwanted collection of particulate mercury which would be possible using packed adsorbent media traps. Some doubts have been raised regarding the ability of the gold based adsorbent to collect mercury effectively during humid conditions. Adsorbents can also be used to collect organic mercury.
Particulate mercury can be collected on filter media (Teflon or Glass Fiber) using generally available particulate sampling equipment (e.g. Code of Federal Regulations, CFR 40 Ch.( 1; 7-1-93) Part 50; Reference Method for the Determination of Lead in Suspended Particulate Matter Collected from Ambient Air). Size selective equipment, such as dichotomous samplers, have been used to future characterize the size range of particulates which contain mercury. Glass fiber filters must be baked before use to drive off any mercury which is originally contained in the "clean" filter. Because of the low expected ambient level values, strong efforts need to be made to minimize "blank" levels which may be contained in the sampling medium.
A variety of wet/dry deposition collection devices exists. Vermont uses a precipitation collector which is specially designed to collect samples for trace metal analysis.
Analytical Procedures
Several trace element analytical techniques are available to analyze samples for elemental and particulate mercury. These include cold vapor atomic fluorescence spectroscopy (CVAFS), direct current plasma atomic emission spectroscopy (DC-AES), cold vapor atomic absorption (AA) and instrumental neutron activation analysis (INAA).
Neutron activation is a nondestructive technique, which is capable of analyzing solid matrix samples without prior digestion. This method is capable of achieving a minimum detection limit of 5 pg/M3 on particulate filters which have collected 21.6 cubic meters of air. Neutron activation analyzers require a nuclear reactor. In Eastern Massachusetts both Massachusetts Institute of Technology (MIT) and University of Massachusetts-Lowell have research reactors and neutron activation capabilities. MIT has done significant research and development work in monitoring mercury using neutron activation.
The University of Michigan has done significant development work in the gold coated sand and glass fiber filter sampling techniques and CVAFS analysis both in association with the Great Lakes and Lake Champlain Basin studies. The minimum detection level for these techniques was reported to be 9 pg/M3 at 43 cubic meters of collected air. Filters analyzed this way must be extracted with an acid solution prior to analysis.
Organic mercury can be detected using DC-AES. It also has been experimental shown to be detectable using standard gas chromatograph - mass spectrometry (GC-MS) such as that used for volatile organic compounds (VOC) analysis, although it is not a standard target analyte. GC-MS analysis for organic mercury compounds could require a special GC separation phase and/or program for optimal detection.
Quality Control
Standard quality control procedures for trace atmospheric levels of mercury will be included with the upcoming USEPA compendium method. However, these quality control procedures will conform with those usually applied to monitoring of this type.
Sampler precision is determined by statistically comparing the concentration results from the analysis of two discrete samples taken simultaneously at the same location. Analytical precision is determined by the analysis of two separate aliquots of sample extract. Statistical precision ranges by percentage can be expected to be larger for trace mercury measurement than for other measurements, because of the low expected absolute concentrations.
Method accuracy for metals measurement is usually evaluated through the analysis of blind matrix spiked samples. The preparation of accurate reproducible spiked samples would be a challenge, again because of the low expected levels.
The maintenance of the integrity and cleanliness of sample media would be a major part of a mercury monitoring QC program. Blank filter (particulates) and adsorbent trap (vapor) samples must be analyzed with field samples to evaluate collection media contamination levels. All surfaces used to collect deposition samples for mercury must be kept scrupulously clean prior to exposure. Adsorbent trap media must undergo special QC procedures which include an initial empirical collection/desorption efficiency evaluation and periodic in series backup samples to detect breakthrough.
Siting and Meteorological Conditions
The selection of monitoring locations depends on the scope and objective of the study. The Great Lakes, Upstate New York and Vermont studies have focused on the impacts on rural locations from mercury sources far upwind and fluctuations with time (and seasons) of those concentrations (Hoyer et. al., 1992; Scherbatshoy, 1993).
A smaller scale sampling network would be appropriate to evaluate the impact from local or area-wide sources of mercury. This network would include a regional upwind or background site to evaluate area concentrations at a location which is not thought to be significantly impacted by sources of mercury. One or more "pristine" statewide locations should be included to evaluate fluctuations in long-range transport of airborne mercury and to help provide background data to truly evaluate the impact of industrial sources. Very little ambient mercury concentration data currently exists for comparison.
A strategy for evaluating the impact and mechanism of airborne mercury on a local environment would probably require a long-term sampling program with vapor, particulate and precipitation deposition components. An area-wide program may indicate the placement of samplers at locations near receptors which are sensitive to airborne mercury or are under multimedia study for the element. Locations in the maximum downwind impact area of known mercury sources as indicated by dispersion modeling results may be appropriate. As previously indicated, upwind non-impacted locations should also be included.
Local placement of samplers would likely follow general Federal Register ambient criteria, although extra attention would be given to avoiding close nearby potential sources of mercury including rooftop fossil fuel combustion chimneys and re-entrained soil.
Costs
Because trace atmospheric mercury monitoring is still in the development stage, good cost estimates are not available. However, the cost of analyzing each sample for trace levels of mercury is in the $200 to $400 in comparison with $50 to $100 per sample for other elements. A conservative estimate or the operation of a mercury monitoring site for one year where air samples are taken every sixth day (including vapor and particulate) and deposition samples are collected every three to four weeks is in the range $75 - $100,000. This estimated cost includes initial sampling equipment, analysis, sample media, quality control and labor. These costs can be reduced through in-house analysis, method refinements and reducing sample frequency. The initial sampling equipment costs (about $25,000) would be reduced during subsequent years of operation.
Conclusions
Monitoring for trace levels of mercury in the ambient air (or by deposition) is not now commonly performed nor is it currently a generally available technology, although the techniques are feasible for research oriented operations. If trace mercury monitoring is under consideration, several options would be available. The first would be to wait for the USEPA recommended method to become available and follow this method using in-house and/or contractor resources. The second option is to carefully scope a research project which includes participation of both agency and academic personnel and assets. There are several good reasons to take the second approach. These include the need to establish reliable ambient background concentrations, the need to apply and document meticulous and intricate measurement and quality control procedures and the value of integrating air monitoring with research being conducted with other media.
Stack Emission Testing
In the September 20, 1994 Federal Register (Federal Register dated September 20, 1994, pages 48259-48261), EPA proposed the addition of Method 29 to Appendix A of 40 CFR Part 60 and Amendments to Method 101A of Appendix B of 40 CFR Part 61.
Both methods can be used to measure mercury however Method 29 can be used also for particulate and other metals determination. This makes Method 29 more practical to use because more than just mercury analysis will be required of testing programs for incinerators.
Method 29 requires a stack sample to "withdrawn isokinectically from the source, particulate emissions are collected in the probe and on a heated filter, and gaseous emissions are then collected in an aqueous acidic solution of hydrogen peroxide (analyzed for all metals including mercury) and an aqueous acidic of potassium permanganate (analyzed only for mercury). The recovered samples are digested, and appropriate fractions are analyzed for mercury by cold vapor atomic absorption spectroscopy (CVAAS) and for Sb, As, Ba, Be, Cd, Cr, Co, Cu, Pb, Mn, Ni, P, Se, Ag, Tl, and Zn by inductively coupled argon plasma emission spectroscopy (ICAP) or atomic absorption spectroscopy (AAS)."
Three separate test runs are required in emission test programs. The duration of each test is at least one hour but may be longer to assure that all contaminants being measured will be present at detectable levels of the analytical methods employed.