Grantee Research Project Results

Year 2 of the project has focused on continuing experimental development work started in Year 1, collection of experimental data, and initial interpretation of results. Such work has been progressing in the following four areas, as defined in the initial proposal:

 

1) Studies of mercury cycling to plants, soils, and other environmental surfaces at the UW-Madison Biotron controlled environment facility, using on-line mercury instruments and mercury isotope spiking studies.

2) Smog chamber studies of mercury oxidation during controlled ozone and SOA formation studies using expertise at the University of New Hampshire 

3) Laboratory studies of the chemical transformations of mercury with cloud and fog water collected using ultra-clean sampling methods along with parallel studies using artificial cloud and fog waters

4) Regional chemical transport modeling to study atmospheric mercury deposition sensitivity to temperature, precipitation, and atmospheric circulation patterns associated with climate change

 

Atmospheric deposition is the primary pathway by which mercury enters aquatic environments in which bacterial methylation and subsequent accumulation in the food chain can occur. The purpose of this module is to comprehensively determine the climate sensitivity of dry atmospheric deposition velocities for gaseous elemental mercury (GEM) and reactive mercury (RM) to a range of environmental surfaces. This is being done by determining the functional dependencies of deposition on temperature, light irradiance, and relative humidity, which will be employed in the atmospheric modeling portion of the project described below. Various plants, soils, and environmental surfaces have been exposed to gaseous elemental mercury enriched in stable isotope 198 (GEM-198), in a controlled environment room at the UW-Biotron facility. Plots of two types of locally collected soil, peat, sand, concrete, asphalt, white ash trees, and white spruce trees were placed in the room alongside a water surface sampler^1,2, and deposition coupons made from quartz fiber filters, some with absorbent coatings to collect RGM and GEM. GEM-198 was introduced into the room augmenting the background concentration of ~5 ng m^-3 over the course of 7 days at two ambient temperatures (10^oC and 30^oC) under a summer light condition. The first design apparatus for GEM-198 dosing yielded large initial input of isotope into the room, which tapered off as the experiment progressed through the 7 days (Figures 2-4). This design has been revised and is in the process of being optimized to provide continuous, stable, and sufficiently large inputs of isotope into the room.

 

We have been successful in getting our experimental design to yield results. In short we are seeing significant time dependent enrichment in white ash, white spruce, grass, and at least one of the soil types. Preliminary data are presented in Figures 1-7.

 

GEM mercury uptakes to white ash was observed in both the light and dark experiments at all the temperatures tested (Figures 1-3). In many of the experiments uptake initially followed the concentration spike of the GEM at the beginning of the experiment, and then settled out at some lower, but temporally constant enrichment value. This initial experiment demonstrates two things. Firstly, uptake of GEM to the white ash was relatively quick, on the order of less than 1 day. Secondly, the GEM-198 concentrations decreased to a much lower concentration than was seen in the relative enrichment of Hg isotope 198 in the leaves. This trend implied that uptake was made up of two components; a portion of mercury that was exchangeably transferred to the leaves and a portion that was non-reversibly retained. This will be further explored with experiments in which concentrations of GEM-198 will be held steady for a period of time at the start of the experiment, and then the source of GEM-198 will be removed from the room. After the source is removed from the room the experiment will continue to see if the enrichment of GEM -198 in the leaves decreases or is conserved. The initial period during which the source is placed in the room will be varied in length to see if GEM-198 retention after removal varies as a function of contact time of the plant with GEM-198. During the dark experiments, enrichment, which for the moment will be assumed to be linearly related to mass uptake, was greatest at 20^oC (Figure 2), while during the light experiments we observed the greatest enrichment at 10^oC (Figure 3). A similar enrichment at 30^oC as was seen under dark and light conditions. We are working to understand the relationship between optimum uptake and conditions, and will verify these findings with duplicate experiments.

 

Enrichments of GEM-198 by white spruce at 0% light conditions (Figure 4) followed trends seen in the white ash data reasonably closely, implying that the enrichment trends observed in the pine needles were dictated either by analogous physiological and physicochemicial processes in the trees, or by the GEM-198 trends seen in each experiment. This demonstrates the importance of optimizing the performance of the GEM-198 source so that it can be operated reproducibly, which we are addressing.

 

GEM-198 uptakes were measured on soil collected at Arlington Agricultural Research Station (Figure 5). An enrichment in Hg stable isotope 198 was observed in the soil, although not as large of a relative uptake as observed in some of the tree leaves. Statistically significant uptakes of GEM-198 to peat (Sphagnum Moss peat) were also measured. We are currently submitting more samples from soil experiments to determine the temporal trend of GEM-198 uptake to the soil substrates. Furthermore, we are in the process of determining how much mercury mass was taken up per gram of each of the different substrates, so that the relative importance of GEM uptake to trees and soils can be assessed.

 

We measured the uptake of GEM-198 to grass (Kentucky Blue Grass; Figure 7) and observed uptakes of GEM-198 in most experiments. Uptake in the dark at 30^oC exhibited similar behaviour as the white ash, and white spruce. As was the case in all but the dark 30^oC white ash data, no initial peak in uptake was seen to correspond with the initially elevated concentrations in GEM-198.

 

In addition to the aforementioned work we intend to conduct to better measure and understand GEM-198 uptake, we are currently evaluating how best to parameterize the uptake of mercury by the different substrates in order to enter them into CMAQ. Lastly we will invert the data shown here to provide absolute measures of mercury flux to the different substrates to enable intercomparisons to determine which surfaces are most important to the atmospheric deposition of GEM.

 

Figure 1. Enrichments of 198 Hg stable isotope in the leaves of white ash trees as a function of light and temperature. Relative humidity was always set to 70%. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.).

Figure 2 Enrichments of 198 Hg stable isotopes in white ash at 0% light conditions and at temperatures of 10^oC, 20^oC, and 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope (GEM-198) collected on gold traps and measured using a High Resolution ICPMS. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.). Some measurements of GEM-198 are yet to be analyzed and added to the figures below.

 

 

Figure 3 Enrichments of 198 Hg stable isotopes in white ash at 100% light conditions and at temperatures of 10^oC, 20^oC, and 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope collected on gold traps and measured using a High Resolution ICPMS. No data is yet available for the 20^oC experiment. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.). Some measurements of GEM-198 are yet to be analyzed and added to the figures below.

 

Figure 4 Enrichments of 198 Hg stable isotopes in white spruce at 0% light conditions and at temperatures of 10^oC, 20^oC, and 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope collected on gold traps and measured using a High Resolution ICPMS. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.). Some measurements of GEM-198 are yet to be analyzed and added to the figures below.

 

Figure 5 Enrichments of 198 Hg stable isotopes in white spruce at 0% light conditions and at temperatures of 10^oC, 20^oC, and 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope collected on gold traps and measured using a High Resolution ICPMS. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.).

Figure 6 Enrichments of 198 Hg stable isotopes in peat at 100% light conditions and at 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope collected on gold traps and measured using a High Resolution ICPMS. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.).

Figure 7 Enrichments of 198 Hg stable isotopes in grass at 0% light conditions and 30^oC. Relative humidity was always set to 70%. Also pictured in figures are the total gaseous elemental mercury concentrations (measured with the Tekran 2537B), and the gaseous concentration of 198 stable isotope collected on gold traps and measured using a High Resolution ICPMS. Uncertainties were determined from the standard error of duplicate samples across all collected substrates (i.e., leaves, soils, etc.).

 

 

 

GEM is most effectively transferred from the atmosphere to aquatic ecosystems by wet and dry deposition after oxidation to reactive mercury. The oxidation of GEM by oxidants such as ozone, OH, and various halogen species have previously only been studied in homogeneous reactions systems^3-15, which do not effectively represent the heterogeneous aerosol reaction systems that are present in the environment. This module aims to evaluate heterogeneous reaction rates of GEM of a range of oxidation reactions. This is being achieved by observing the effect of a complex smog reaction system that leads to the formation of secondary organic aerosol on reaction rate kinetics for oxidants such as ozone, OH, Cl and Br. The effect of the updated reaction kinetics data on dry and wet deposition rates will be studied in the climate sensitivity modeling portion of the project described below. During the summer of 2008, we conducted 6 weeks of experiments in the smog chamber at the University of New Hampshire. GEM-198 was added to ongoing oxidations of propene, alpha-pinene, isoprene, and toluene, and 2-butanol (OH scavenger) in the dark and under ultraviolet (350BL) fluorescent lamps. Concentrations of GEM were monitored real-time using a Tekran 2537A GEM analyzer, and GEM oxidized to reactive mercury (RM) was collected on specially prepared filter substrates¹⁶ as described above in the dry deposition experiments.

 

Initial experiments in the smog chamber were performed to evaluate wall losses of GEM to the walls, which were determined to be less than 1% during the first 4 hours of the experiments at starting concentrations between 70 and 100 ng m^-3 (Figure 8). These losses were then compared to losses measured for GEM when ozone (nominally 250 ppb) was added at the start of the experiment. The oxidative losses of GEM upon reaction with ozone were much larger than the losses to the walls, and could be approximated with a pseudo first order reaction rate model. In such a model, which has been applied by previous authors^{6, 9}, the gradient of a ln[GEM] vs t plot is equal to the rate coefficient multiplied by the ozone concentration, which is held constant during these initial reactions (Figure 9). The rate coefficient calculated using the measurements made during this study was -4 x 10^-19 cm³ molec^-1 s^-1, which compared well with previously published measurements (Table 1) giving us confidence that the reaction coefficient measurements made with the smog chamber were in the previously observed range of values. The oxidation of GEM in the dark was much faster than wall losses allowing an initial determination of the GEM + ozone reaction rate coefficient. We have assumed the system to be bi-molecular, but that it acted as a pseudo first order system due to ozone remaining constant so that our data could be compared with previous publications^{6, 9} (Figure 9; Table 1). Our preliminary rate estimates agree very well with Pal and Ariya, 2004, but are higher than that reported by Hall et al., 1995. We think that we have successfully demonstrated that our experimental method refers well to previously conducted experiments providing a reference point to GEM oxidation experiments conducted in SOA systems, which are presented in Figures 10 and 11. We tested the validity of using only an oxidation of GEM by ozone in the atmospheric mercury chemical transport models by performing oxidation experiments of GEM with ozone and SOA precursors (Figure 10). This was done to determine what effect the presence of reactive organic oxidants, and surfaces with dicarboxylic acid groups would have in the presence of UV¹⁷. The decrease in losses of elemental mercury to oxidation by ozone and secondary OH implied that a competing reductive reaction on the surfaces of the SOA particles involving dicarboxylic acid groups could be occurring. Another explanation was that some reaction mechanism was inhibiting the oxidation of GEM by ozone and secondary OH from occurring. This unknown process accounted for between approximately 50 to 100% of the GEM conversion observed in the GEM + ozone experiments. Losses of GEM in the SOA system with and OH scavenger present demonstrated some differences between the ozone only and ozone + secondary OH oxidations of GEM. However, numerical modeling and the inclusion of more experimental data from the completed experiments will be necessary to be able to robustly rationalize those differences, as none is obvious at present.

 

Our results demonstrate very clearly that net oxidation rates of GEM in SOA systems, with and without an OH scavenger (2-butanol), are significantly slower than those measured for the simple GEM + ozone bimolecular system. We are currently beginning a numerical modeling project to understand the reaction chemistry occurring in these experiments, in order to determine robust, mechanistically based reaction rate coefficients. The implications of these results are considerable. While the fundamental reaction work performed by previous authors^6,9 is hugely valuable in understanding complex heterogeneous atmospheric oxidations of GEM, our data complements these publications by providing a significant step forward in our ability to accurately model atmospheric conversion of GEM to the depositable, reactive mercury species. As current Chemistry Transport Models (i.e. CMAQ¹⁸) rely upon only the GEM + ozone reaction to estimate photochemical production of reactive mercury, these models would cause a significant overprediction of reactive mercury concentrations in the atmosphere. This has indeed been seen to be the case in the CMAQ verification study performed in Part 4 of this project. This means that the current paradigm for predicting GEM oxidation in the atmosphere is overemphasizing the importance of reactive mercury production from the background GEM pool, and thereby underestimating the importance of local point source impacts^19,20. This behaviour has been confirmed in the comparisons of CMAQ Hg with ambient measurements conducted in Wisconsin¹⁹. Clearly these findings could have important policy implications.

 

Figure 8. Smog chamber experiments to measure the wall losses of gaseous elemental mercury to the chamber walls, and the oxidative losses of gaseous elemental mercury under reaction with ozone in the dark.