The inorganic studies complement the radionuclide assessments because they provide information about the types of aerosols in the atmosphere and how their concentrations vary with time. In addition, some of the trace elements being studied (As, Ba, Cd, Cr, Pb, Hg, Se and Ag) are listed as components of the Permitted TRU Mixed Wastes in the WIPP hazardous waste permit (Waste Acceptance Criteria for the Waste Isolation Pilot Plant, DOE/WIPP-069, November 8, 1999). Since no mixed waste has yet been delivered to the site, the nonradiological data presented here can be considered part of a continuing effort to characterize background conditions.

Methods

Gravimetric determinations of the aerosol mass collected on the filters were conducted at CEMRC, using standard operating procedures detailed in the 1998 CEMRC Report. These determinations were only made on the Teflo® (IC) filters because static problems with the Metricel® filters caused their weights to vary erratically. The total mass that accumulated during a sampling period was divided by the total air volume drawn through the filter to calculate aerosol mass concentrations.

All of the aerosol data presented in this report were produced by CEMRC. Standard operating procedures have been developed for the chemical analyses, and where possible these are based on applicable standard U.S. Environmental Protection Agency (EPA) procedures. A summary of the analytical procedures and quality control aspects of the chemical analyses of the aerosols and other WIPP EM samples are presented elsewhere in this report (Appendix K).

For the IC analyses, individual Teflo® filters were extracted with de-ionized water, using an ultrasonic water bath to facilitate the process. Aliquots of the same aqueous extracts of the aerosol samples were used for both anion and cation analyses.

Aerosol filters were prepared for elemental analyses using a microwave digestion system with HNO₃, HCl, HF, H₂O₂ and H₂O. The concentrations of major and trace elements were determined in the aerosol samples by AAS, ICP-ES and ICP-MS. The ICP-MS was used in place of ICP-ES for analyses after January 1999.

The paired t-tests showed that the mass concentrations at the On Site station were significantly different from those at either Near Field or Cactus Flats at p < 0.0001. The mean difference between paired samples from On Site vs. Near Field was 6.17 µg m^-3; thus the mean TSP mass loading at the On Site station was ~30% higher than that at Near Field (Fig. 13). The difference in mean mass concentrations between On Site and Cactus Flats was slightly smaller at 5.16 µg m^-3, a difference of more than 25%.

The next step in the analysis was to compare the TSP mass concentrations at Near Field and Cactus Flats, again using a paired t-test. Results of this test showed that the aerosol mass concentrations at the two sites were not significantly different (p >0.05), indicating that the elevated TSP mass concentrations at the On Site station (presumably associated with WIPP site activities) did not affect Near Field.

The PM₁₀ and PM_2.5 mass concentrations for the 1997-1999 samples were quite comparable to a much smaller data set obtained with high volume samplers and summarized in the 1998 CEMRC Report and in Lee et. al. (1999, J. Radioanalytical and Nuclear Chem. 234, 267). As a follow-up on those studies, mass concentrations were compared among samplers for the much larger data set from the low-volume samples. For these tests a two-way analysis of variance (ANOVA) was used, with sampling site and particulate size sampler type (i.e., TSP, PM₁₀ or PM_2.5) as variable classes. Data for the On Site station could not be included in this test because only a TSP sampler was deployed at that station.

Results of the ANOVA for aerosol mass confirmed that the differences in mass concentrations between Near Field and Cactus Flats were not significant (p > 0.05), and this test expanded the results to include the PM_2.5 and PM₁₀ data. Furthermore, the effect (if any) was small because the differences between the mean PM₁₀ and PM_2.5 mass concentrations at Near Field and Cactus Flats were small (< 1% and ~5%, respectively).

Despite the elevated salt concentrations in the FAS samples, the Na concentrations in TSP samples from the On Site station were not significantly different from those at Near Field or Cactus Flats (paired t-tests, p > 0.05, see Fig. 14). It is worth noting that the Na determined through the IC analyses is water soluble Na, and this presumably includes little if any Na released from the matrices of insoluble mineral aerosol particles.

Further analyses showed that the Na-to-mass ratios did not differ among sites either. Therefore, salt did not make up as large a percentage of the aerosol mass at the On Site station as one might expect if salt were a disproportionately large contributor to the aerosol mass there. In fact, aerosol Na amounted to only 1-2% of the total mass at the sites (Table 4). If one assumes that the composition of the salt deposits is similar to that of sea water, one would multiply the Na concentration by 3.26 to estimate the mass of salt in the atmosphere, using the seawater composition of Millero (1996, Chemical Oceanography 2^nd edition, CRC Press, Boca Raton, FL.). Based on this presumptive source material, the salt particles in the atmosphere would amount to < 10% of the aerosol mass, which is clearly not sufficient to account for the observed differences in mass concentrations among sites.

Aerosol nitrate and sulfate are other important atmospheric constituents, in large measure as a result of their roles in the biogeochemical cycling of nitrogen and sulfur. Aerosols composed of these ions also are of considerable interest to atmospheric scientists because of their light scattering properties and associated effects on the fluxes of solar radiation. Aerosol nitrate mainly originates from fossil fuel combustion (Singh, H. B., 1987, Environ. Sci. and Tech. 21, 320), including emissions from motor vehicles (Dignon, J., 1992, Atmos. Environ. 26A, 1157) while anthropogenic sulfate is mainly produced through the combustion of coal and residual fuel oil (Cullis, C. F. and M. M. Hirschler, 1980, Atmos. Environ. 14, 1263). Thus the concentrations of these major ions could be elevated at the WIPP by the normal operations of the physical plant (i.e. emissions from motor vehicles, generators and other combustion sources).

Somewhat surprisingly, the amount of aerosol sulfate at each of the sites was greater than the concentration of atmospheric salt aerosols (see Fig. 14). The mass of aerosol sulfate also was greater than that of nitrate, with sulfate amounting to 10-15% of the total mass loading, and nitrate < 5% of the mass. However, neither sulfate nor nitrate differed significantly between the On Site and Near Field or Cactus Flats TSP samples (paired t-tests, p > 0.05). Therefore, like salt, the combination of these two anions cannot account for the differences in aerosol mass concentrations among sites.

The actual contributions of the sulfate- and nitrate-containing aerosols to the mass concentrations cannot be determined more precisely because the cations (Na⁺, Ca²⁺, Mg²⁺ NH4⁺, etc.) that are bound to the anions, have not been established. In fact, the cations would increase their contributions of these types of aerosols considerably. For example, if the sulfate aerosol were predominantly in the form of the mineral gypsum (CaSO₄·2H₂0), the increase in mass contribution of the aerosol would be nearly 80% relative to sulfate alone, but if the sulfate aerosol were in the form of anhydrite (CaSO₄), the increase in mass contribution would be ~40%.

Based on the IC data obtained to date, the combination of salt and the major anions account for roughly 25% to 30% of the total TSP aerosol mass at On Site, assuming a 50% contribution of the cations associated with nitrate and sulfate. Again, an important point of these analyses with respect to the WIPP EM is that none of these substances differed significantly among sites, and their contributions to the aerosol mass concentrations cannot explain the observed higher TSP masses at On Site.

The mass differences between sites were further investigated using the cation data generated by IC analyses of the aqueous extracts of the Teflo® filters. These comparisons showed that the differences between sites in Ca and Mg concentrations were highly significant (ANOVA, p < 0.001), and K concentrations were marginally significant (0.05 > p > 0.01). Moreover, the concentrations of all of these cations were highest in the samples from the On Site station (Fig. 15), and all of these elements are strongly enriched with respect to the cation/Na ratios expected from the previously referenced sea salt composition of Millero (1996) (Table 5).

While Ca and Mg contributed to the higher mass concentrations at the On Site station, they cannot, even in combination, account for more than ~40% of the observed differences in mass. For example, if one assumes that Ca exists as a 50%/50% mixture of (CaSO₄·2H₂0)/CaCO₃, which would not be unreasonable given the mass concentrations of sulfate, then the corresponding difference in masses between On Site and Near Field attributable to these Ca-bearing minerals would be 2.6 µg m^-3. This is about 35% of the observed mass difference between the two sites (i.e. 22.3 minus 15.2 µg m^-3, Fig. 13). The corresponding value for Mg, based on the same assumed anion mixture, would be much smaller, about 3%. These comparisons indicate that Ca- and Mg-bearing mineral aerosols contribute to the higher mass concentrations at the On Site station, but these minerals alone cannot explain the observed differences.

One might speculate that construction or maintenance activities, including any operations involving cement or road construction at the WIPP may have mobilized some Ca-containing particles, but we have no way to investigate this further with the existing data.

The sources for atmospheric Ba are largely unstudied, and this element was not considered in either of the two seminal references on the sources of trace elements in the atmosphere (Pacyna, J. M., 1986. In Nriagu, J. O. and C. I. Davidson (eds.), Toxic Metals in the Atmosphere, John Wiley, NY; 2; Nriagu, J. O., 1989, Nature 338, 47). Ba is known to be a component of some lubricants, however. On the other end of the scale, the biogeochemical cycling of pollutant Pb has been extensively studied, largely as a result of the global contamination caused by the use of leaded gasoline and the related health effects (Patterson, C. C., and D. M. Settle, 1987, Marine Chem. 22, 137).

Gravimetric analyses were not performed on the cellulose ester filters used for the TE sampling because of problems obtaining stable weights, and therefore it is not possible to unequivocally determine which of the elements may have contributed to the higher mass concentrations at the On Site station relative to the other sites. However, a simple examination of the data shows that Ca has the highest concentration per unit volume air of all the elements determined, with an arithmetic mean of 0.85 µg m^-3. This is somewhat lower than the average concentration determined by IC, but some of this difference may be due to the fact that the time periods encompassed by the two sets of measurements are slightly different as a result of the TE/IC sequence in sampling. Also, different types of filters were used for the IC and elemental analyses, and it is possible that the filters differ somewhat in their particle collection efficiencies. Except for a few extreme instances in which the IC Ca concentrations were higher, the data produced by the two methods were in good agreement.

After Ca, the element exhibiting the next highest aerosol concentration is Al. The average Al concentrations were similar at all sites, ranging from 0.405 µg m^-3 at Cactus Flats to 0.477 µg m^-3 at Near Field, with On Site intermediate at 0.442 µg m^-3. Al is commonly used as an indicator of mineral dust (e.g. Rahn, R. A., 1976, The Chemical Composition of the Atmospheric Aerosol, Tech. Report, Grad. Sch. of Oceanogr., Univ. Rhode Island, Kingston.), with average crustal material containing ~8% Al (Taylor, S. R. and S. M. McLennan, 1995, Rev. of Geophys. 33, 241), and therefore mineral matter is one of the dominant components of aerosol, amounting to ~25% of the total mass concentration On Site (based on the IC mass data) and ~30-40% at the other two aerosol sampling stations.

Differences in mineral aerosol concentrations between sites cannot explain the higher mass concentrations at the On Site station. The differences in mineral dust concentrations among sites were small, and in the case of On Site and Cactus Flats they were the reverse of the patterns for mass concentrations.

The final analysis of the elemental data was to compare the concentrations On Site to those at Near Field and Cactus Flats. A simple indicator was derived by dividing the mean concentration for each element for the On Site station by the corresponding means at the other two sites. This was done only for those elements with ≥ 50% of the samples above detection limits (Fig. 17). This index provides a simple measure of whether the concentrations were higher at On Site than the other sites, although the index does not lend itself to rigorous statistics.

Corroborating the previous interpretations from the IC data, the ICP-MS Ca and Mg concentrations at the On Site station were 30% to 70% higher than those at the other two sites. It is important to emphasize that these results are from a fully independent analysis on a separate set of samples, and therefore this parallel method provides a powerful means for validating the results and conclusions from the IC analyses.

Sr exhibited the third highest concentration difference between On Site and the other sites. Even though the concentration of Sr was not high enough to account for the differences in mass concentrations, these elemental differences in aerosol composition do provide some information on what types of aerosols are most abundant at each site. Fig. 17 also shows that the On Site:Cactus Flats elemental ratios tended to be higher than the On Site:Near Field ratios, which is further evidence for compositional differences between Cactus Flats and Near Field.

The concentrations of the elements exhibiting high levels at On Site (relative to the other stations) were divided by the matching Al concentrations to provide a means for determining whether the elements of concern are mainly associated with mineral dust or are enriched relative to that source. For all of the elements that had high concentrations On Site, including Ca and Mg, the ratios to Al were higher than in the average crustal material composition of Taylor and McLennan (1995) (Table 6).

Another important use of the elemental data in the future will be for comparisons to radionuclide analyses. These comparisons will be particularly useful for U and Th. The ICP-MS analyses have generated a set of data for these two elements in aerosols, and these data will not only provide background information but also a basis for comparisons against other media.

The interpretation of the elemental data produced to date has provided some insight into the apparent perturbation of mass concentrations at the WIPP site, but much of the mass difference cannot be accounted for, suggesting a contribution by organic substances. Follow-up studies on elemental and organic carbon in aerosols are being considered for the upcoming year.

Tables presenting the aerosol data summarized herein are available on the CEMRC web site at http://www.cemrc.org.