Widespread Atmospheric Tellurium Contamination in Industrial and

Article Cite This: Environ. Sci. Technol. XXXX, XXX, XXX−XXX

pubs.acs.org/est

Widespread Atmospheric Tellurium Contamination in Industrial and Remote Regions of Canada Johan A. Wiklund,*,† Jane L. Kirk,*,† Derek C. G. Muir,† Jacques Carrier,‡ Amber Gleason,† Fan Yang,† Marlene Evans,§ and Jonathan Keating§ †

Aquatic Contaminants Research Division and ‡National Laboratory of Environmental Testing, Environment Canada, Burlington, Ontario, Canada L7R 4A6 § Aquatic Contaminants Research Division, Environment Canada, Saskatoon, Saskatchewan, Canada S7N 3H5 S Supporting Information *

ABSTRACT: High tech applications, primarily photovoltaics, have greatly increased demand for the rare and versatile but toxic element tellurium (Te). Here we examine dated lake sediment Te concentration profiles collected near potential point sources (metal smelters, coal mining/combustion facilities, oil sands operations) and from rural regions and remote natural areas of Canada. Te contamination was most prevalent near a Cu/Zn smelter where observed deposition infers 21 g Te released per metric ton (t) of Cu processed. Globally, 9,500 t is predicted to have been atmospherically deposited near Cu smelters post-1900. In a remote area of central Canada (Experimental Lakes Area; ELA), preindustrial Te deposition rates were equivalent to the estimated average global mass flux supplied from natural sources; however more surprisingly, modern Te deposition rates were 6-fold higher and comparable with Te measurements in precipitation. We therefore suggest that sediment cores reliably record atmospheric Te deposition and that anthropogenic activities have significantly augmented atmospheric Te levels, making it an emerging contaminant of potential concern. Lake water residence time was found to influence lake sediment Te inventories among lakes within a region. The apparent settling rate for Te was comparable to macronutrients (C, N, P), likely indicative of significant biological processing of Te.

■

enriched element (EF = 107) in fumarolic emissions from Erta Ale volcano Ethiopia relative to local basalts.14 Sen and Peucker-Ehrenbrink15 estimated the natural and anthropogenic global mass flows of Te to be 780 ± 270 and 790 ± 50 t year−1, respectively. Volcanoes contribute ∼35% of the natural Te mass flow, which is higher than seen for Hg (∼10%), while soil erosion, riverine transport, and aeolian dust comprise the remaining significant natural sources. While no estimates exist, biotic releases may also contribute significantly to the global mass flow of Te. Anthropogenic mass flows are estimated to be dominated by coal burning (700 ± 100 t year−1 and mining activities (125 t year−1).15 The commercial production of Te has been used to estimate the mass flow due to the mining sector.15 Due to its scarcity, Te is not mined on its own but recovered primarily as a byproduct from the processing of Cu, with Canada, Japan, Peru, Russia, Sweden, and USA being major producers.2−4,16,17 However, due to the low efficiency (2−4.5%)2,18 of Te extraction from Cu ore, the estimated anthropogenic mass

INTRODUCTION

Tellurium (Te) is one of rarest elements on earth with crustal abundance of 1−5 μg kg−1, which is similar to gold and platinum.1−3 Te is used in alloy production, rubber vulcanization, and increasingly in the electronics sector, particularly in cadmium-telluride photovoltaic panels and thermoelectric devices.2,3 The increased demand for photovoltaic panels has increased Te demand, and world production has risen from ∼100 t (t; t = 1 Mg) yr−1 in 2000 to ∼500 t yr−1 in 2010.2,3 Concerns have thus been raised about potential environmental and human health issues as some forms of Te are highly toxic.4−11 Little is known about the natural cycle and mass balance of Te through the atmosphere and bio- and lithospheres.3,12 The distribution of Te in the environment appears to be one of extremes, very low in most bulk components of Earth’s system but highly enriched in a few specific compartments. Like that of Earth’s crust, Te content in oceanic waters3 is exceedingly low (0.05−40 ng L−1). The major geogenic sink for Te in the world’s oceans is incorporation into deep ocean Fe−Mn nodules13 in which Te is the most enriched element (Enrichment Factor (EF) = 5 × 104). Volcanic gases are also exceedingly enriched in Te. For example, Te was the most Published XXXX by the American Chemical Society

Received: December 4, 2017 Revised: March 26, 2018 Accepted: April 4, 2018

A

DOI: 10.1021/acs.est.7b06242 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 1. Tellurium concentration profiles from lake sediment cores collected across Canada (see also Figure S1). Note that the Flin Flon smelter is a significant local point source for Te emissions. While the ELA and Dorset Ontario (ON) are remote from industry, both show major Te enrichment indicating large point source(s) must exist within the Laurentian Great Lakes basin. Sudbury ON (SB) is one major source as were the former Cu smelters of the Keweenaw Peninsula (KP) MI, USA, although others (smelters + coal plants) exist. Method detection limits are represented by the blue vertical dashed blue lines in each plot. Map modified from Natural Resources Canada 2001, Atlas of Canada (https://open. canada.ca/en/open-government-licence-canada).

■

flow due to mining15 is low by a factor of ≥20. While the majority (∼88%) of the Te lost during copper mining is to tailings, the second highest loss is to aerosol and gaseous products during smelting/refining.2 If not intercepted by emission abatement systems, this waste stream emits Te into the atmosphere at a rate similar to that of refined Te produced for sale. Here we examine Te concentration profiles in dated lake sediment cores from across Canada located near the following: base metal smelting operations (Flin Flon and Thompson MB), coal mining and burning facilities (Estevan SK), oil sands mining and upgrading (Northern AB), rural regions (central AB), and natural areas remote from human disturbance (ELA ON, Dorset ON, Kejimkujik NS). The five lakes of the Experimental Lakes Area (ELA ON) are examined in more detail to reconstruct the history of anthropogenic sourced Te deposition from 1860 to 2010. Calculated modern and natural Te flux rates are compared to literature values for Te concentration in modern precipitation and the estimated rate of Te supplied by natural sources. Catchment effects and hydrologic control on lake retention of Te are also examined.

MATERIALS AND METHODS

Lake sediment cores were collected from several locations in the Canadian Provinces of Alberta (AB), Saskatchewan (SK), Manitoba (MB), Ontario (ON), and Nova Scotia (NS) as part of the mercury trends research component of the Clean Air Regulatory Agenda/Climate Change and Air Pollutant (CARA/ CCAP) and Joint Oil Sands Monitoring (JOSM) programs (see Figure 1, Tables S1−S2). Sediment cores were obtained from the deepest point in each lake using a Uwitec gravity corer designed to collect undisturbed sediment profiles. The deepest point in each lake was located by sonar or by using bathymetric maps, avoiding highly sloped areas that may be subject to slumping. All cores were sliced into 0.5 cm sections for the top 20 cm and then at 1 cm sections for the bottom 20−40 cm. Sediments were stored in Whirl-Pak polyethylene bags or in wide-mouth polypropylene jars and were frozen immediately after collection. Sediment Core Dating. Measurement of 210Pb activity and subsequent dating of sediment cores were carried out as previously described for sediment records from the ELA ON and Flin Flon MB19 or for Plastic Lake ON.20 The additional cores were similarly dated at the Canada Centre for Inland Waters, Burlington using alpha and/or gamma ray spectrometry B



Article

was maintained at 240 °C for 30 min. The vessel was removed from the microwave oven, and the cooled digest was brought to a 50 mL volume with deionized water. The following day, the solution was diluted by a factor of 12.5 and analyzed by CRCICP-MS as described above. Calculation of Anthropogenic Te Fluxes and Inventories. The background or natural Te concentration (Tebackground) was estimated from the average Te/Al ratio21,22 observed in each of the ELA lake sediment records for pre-1860 sediment intervals as that date preceded significant human industrial influence within the NW-Laurentian Great Lakes region.19,23 For the Flin Flon MB lakes, the pre-1915 sediments were used as the presmelter (est. 1930) baseline when calculating smelter sourced anthropogenic inventories. Postdevelopment sediment intervals were then compared to the “natural baseline” or background Te content, and anthropogenic enrichment was estimated for each core as

(see SI Methods). Resulting 210Pb, 226Ra, and 137Cs activity profiles and age depth model results are shown in Figure S2. Analytical Methods. All analysis was conducted at the National Laboratory for Environmental Testing (NLET) in the Canada Centre for Inland Waters (CCIW), Burlington Ontario, Canada, a premier facility for analysis of water, sediment, and biological matrices for environmental contaminants. The analytical methods used here for Recoverable Metals in Sediment is comparable to EPA Method # 3051 and to the aqua-regia extraction method HMARSOIL-E3062A used by the Ontario Ministry of the Environment. Metals extracted by this procedure are taken to represent the “environmentally available” metals. With the exception of the sediment cores from Thompson MB, all samples were digested using the HotBlock Method (NLET 2404), while those of Thompson MB used the Microwave-Digest Method (NLET 2403). See Tables S3s and S4 for Method Detection Limits, % recovery, and % precision of elemental analysis of Standard Reference Materials. Hot-Block Digest Method. 500 mg of homogeneous freeze-dried sediment was weighed into a 60 mL polypropylene digestion tube, and 1.0 mL of nitric acid was added. The tube was shaken, the reaction was allowed to settle, and 3.0 mL of HCl was added. The tube was shaken again, and the solution was allowed to settle overnight. The following day, the sample cap was loosened, and the sample tube was placed on a hot block digestion system, held at 60 °C for 1 h, then slowly ramped up to 100 °C at a rate of 10 °C per 30 min, and then held at 100 °C for 1 h. The combination of temperature and the strong oxidizing conditions ensures destruction of the organic and the majority of inorganic components of the substrate. With the exception of strongly siliceous bound elements, the aqua-regia digestion provides effective extraction of analytes of interest. After the digests were cooled down, they were diluted to 50 mL with reagent water. The following day, the solution was then diluted by a factor of 25 and analyzed by an inductively coupled argon plasma-collision/reaction cell mass spectrometer (CRC-ICP-MS: Agilent 7700x ICP-MS) using discrete sampling pneumatic nebulization. Each respective element was measured at a specific mass to charge ratio, with the m/z value expressed in amu (atomic mass units). Elements were analyzed in reaction gas {H2} mode: [Se(78)] only; in collision gas {He} mode: [Al(27), As(75), B(11), Co(59), Cr(52), Cu(63), Fe(56), K(39), Mg(26), Mn(55), Na(23), Ni (60), P(31), Sc(45), Sr(88), Ti(47), V(51), Zn(66), Zr(90)]); in collision gas {Xe} mode: [S(32)] only; and in normal no gas mode: [Ag(107), Ba(137), Be(9), Bi(209), Ca(43), Cd(111), Ce(140), Cs(133), Ga (71), Gd(157), La(139), Li(7), Mo(98), Nb(93), Pb(208), Pd(108), Pt(195), Rb(85), Rh(103), Sb (121), Sn(118), Tl(205), Te (128), U(238), W(184), Y(89)]. The results were reported on a dry weight basis. An examination of possible mass-interferences (CdO, SnO, and SrAr) was conducted using blanks and prepared Cd (10 μg L−1), Sn (5 μg L−1), Sr (100 μg L−1), and Te (10 μg L−1) solutions under He gas and no collision gas modes. No appreciable interference was found, and Te data is free of bias at the method detection limit of 0.02 mg kg−1 sediment. Microwave Digest Method. 250 mg of freeze-dried sediment was weighed into a TFM Teflon vessel, and a mixture of nitric acid and hydrochloric acid (4.5:2 mL) was added. The vessel was assembled, sealed, and fitted onto a microwave rotor which was placed in a high-pressure Ultrawave microwave oven. The sample was digested, and the temperature

EFi =

Teconci /Alconci Te background /Albackground

(1)

where EFi = the enrichment factor of interval i (EF = 1 is equivalent to background conditions), Te and Alconci = the Te and Al concentrations of core interval i, and Te background /Al background = mean the Te to Al ratio for pre-1860 intervals from a sediment core. The flux of excess or anthropogenic sourced Te (Te FluxAnthro) for a given sediment core interval (i) was then determined by ⎛ EF − Te FluxAnthroi = Teconci⎜ i ⎝ EFi

210 1 ⎞⎛ F Pbu,local ⎞ ⎟(ri) ⎟⎜⎜ 210 ⎠⎝ F Pbu,regional ⎟⎠

(2)

where ri = the dry mass sedimentation rate for interval i (determined from 210Pb dating), F210Pbu,local = the flux of unsupported 210Pb measured in the sediment core (equal to the inventory of unsupported 210Pb multiplied by the 210Pb decay constant), and F210Pbu,regional is the expected flux of unsupported 210 Pb from direct atmospheric fallout for the region as predicted based on latitude20 or previous local published value,24 with the ratio of these giving the focus factor (FF) correction. As not every sediment core interval was analyzed for elemental composition, linear interpolation of anthropogenic Te deposition was performed between analyzed horizons. The net deposition or inventory of anthropogenic Te was then integrated for each sediment record. The total (anthropogenic + natural) Te deposition and subsequently the total Te inventory was similarly determined by excluding the second term from eq 2. The “natural” or preindustrial annual Te deposition rate for a given lake sediment core is equal the post-1860 Total Te inventory minus the anthropogenic Te inventory all divided by the time elapsed in years (2010−1860). ⎛ Te InventoryTotal − Te InventoryAnthro ⎞ Te FluxNatural = ⎜ ⎟ 2010 − 1860 ⎝ ⎠ (3)

Calculation of the Anthropogenic Landscape Inventory of Te near Flin Flon. The total anthropogenic inventory of Te deposited within a 50 km radius of the Flin Flon Cu−Zn smelter was estimated in the manner previously described for Hg19 (see eqs 6−8) integrating the relation between anthropogenic Te inventory (weighted by mean wind source C



Article

concentrations (Figure 1d, Figure S3). This is not unexpected as Te is often associated with the gold content of volcanogenic massive sulfide deposits mined near Flin Flon.31 Moreover, world Te production is mainly a byproduct from copper refinery anode sludges,2,3 with Flin Flon being one of the early producers of Te starting in 1935.32 Using the method previously used for Hg,19 we estimate the inventory of anthropogenically sourced Te deposited within a 50 km radius of the Flin Flon smelter at 72.2 t (see Figure S3) over its operational history (1930−2010). Other major copper refining centers in the world likely show similarly enhanced Te deposition surrounding them. Twenty-five Flin Flon area mines have contributed ore containing 3.4 × 106 t of copper33 to the smelter, yielding an emission factor of 21 g of Te atmospherically deposited near Flin Flon per t of Cu processed (72.2 t Te/3.4 × 106 t Cu = 21 g Te/t Cu). As net 1900−2010 global Cu production34 (minus production from recycling) is 451 × 106 t, we estimate that 9,500 t of Te has been deposited near Cu smelters globally. As net global refined Te production2 (1940−2010) is estimated at 11,000 t, Te emissions to air from Cu smelters is both a large source of Te contamination and a very large loss in potential Te production. This assumes the Flin Flon smelter process and the trace element composition of the Volcanogenic-Massive Sulfide (VMS) deposits exploited at Flin Flon are comparable to other 20th century Cu producers. This appears reasonable considering current information, which while limited indicates porphyry Cu deposits (dominant global Cu and Te source) have an equivalent Te content to VMS Cu deposits.35 In Thompson MB (Figure 1e), Te in lakes located 13−48 km from the smelter shows some increase after nickel mining and refining began (1958/1961),36 but it is only highly apparent (20× increase) in the lake closest to the smelter (1A). Large changes in sedimentation rates are seen in all the Thompson cores (Figure S2d-g), masking some of the trends in Te deposition (Figure S4). Still, Ni operations in Thompson MB contributed significant Te emissions to the atmosphere but to a lesser extent than Flin Flon. Sediment core Te concentration profiles from two Ontario sites, ELA and Dorset (Figure 1f-g) which are relatively remote (240 and 160 km, respectively) from major industry and population centers, demonstrated large (10−20×) increases in sediment Te concentrations during the 20th century, suggesting regional Te emission sources and long-range transport within the Laurentian Great Lake region. Although there have been few reports of atmospheric Te pollution to date, one definitive study comes from a Ni smelter in Clydach, UK37 that imported ore from Sudbury, ON, which is located within the Laurentian Great Lake region. Soil Te concentrations were elevated with increasing proximity to the Clydach smelter (maximum concentrations of 11 mg kg−1) and were comparable to those observed in lake sediments from near the Flin Flon smelter (7.6 mg kg−1; Douglas Lake) and near Sudbury, (29.5 and 24.3 mg kg−1 from McFarlane and Clearwater lakes, respectively38). Sudbury is a world class Ni−Cu and platinum group metal mining and processing region, with smelting beginning in 1888 followed by major expansions in the early and mid-20th century. In 1972, a 380 m super stack began operation, reducing local pollution but greatly increasing long-range dispersion of emissions. Thus, the large increase in Te concentration seen at Dorset after ∼1970 (Figure 1g, Figure S4) likely reflects implementation of new technology which enhanced dispersion from Sudbury. In fact,

direction) and distance from the smelter stack over a 50 km radius from the stack. ELA Lake Sediment Te Inventories and Relations with Catchment Area and Hydrology. ELA lake water renewal times (τw) and areal discharge estimates (Table S2) were calculated in the manner described by Curtis and Schindler,25 using lake area and volume data. The relation between sediment inventories of anthropogenic Te for low order lakes vs either lake water renewal times (Figure 4a) or areal discharge rate (m yr−1; annual discharge divided by lake area) (Figure 4b) was then used to estimate the potential maximum Te inventory of the ELA for a theoretical lake of infinite water renewal time (1.042 ± 0.006(1SD) mg m−2) or a lake of zero discharge (1.053 ± 0.024(1SD) mg m−2). Using either of these equivalent estimates of maximal Te inventory allowed calculation of the retention (R) of Te in the low order ELA lakes where R = observed ELA lake Te inventory/maximum Te inventory (4)

As retention can also be expressed as R = v /(v + qs)

(5)

where v = apparent settling velocity, and qs = areal water discharge rate (= annual total discharge/lake area). Plotting the relation 1/R versus qs, we solved for the apparent settling velocity (m yr−1) as the slope (m) of the relation (Figure 4c) is equal to 1/v.26,27 Graphical and regression analysis was performed using Sigma Plot 13.0. Unless stated otherwise, all uncertainty estimates graphical or within text are standard deviations (1 SD).

■

RESULTS AND DISCUSSION Local and Regional Sources of Atmospheric Te Contamination. Te concentrations in lake sediment were generally steady and low (100 years) been a major Au and Ag producer.41 As Te often accompanies Au and Ag in Ontario as gold and silver tellurides,42−44 refining of precious metals may have contributed historical Te deposition to the ELA and Dorset ON. The many coal-fired stations, smelters, and industrial activities in adjacent regions of Quebec and USA also likely contributed Te deposition to the Great Lakes region. In another remote site (∼140 km from industry and population centers) in Eastern Canada (Kejimkuijk National Park, Nova Scotia), lake sediment cores show low present-day levels of Te (≤0.02 mg kg−1) but a modest rise during the early 20th century (Figure 1h). During the late 19th and early 20th century, Kejimkujik was one of several Nova Scotia Au producing districts.45 As Te is associated with the Au-bearing sediment hosted vein deposits of Nova Scotia,46 processing of Au may have contributed to Te emissions to air. Alternately, Kejimkuijk is downwind of industrial centers of the American eastern seaboard. A sediment core from Chesapeake Bay located along the American eastern seaboard found Te to be one of the most enriched metal(oid)s, which was attributed to copper processing in Baltimore, Maryland.47 While Baltimore is ∼1000 km from Kejimkuijk, our results suggest that significant anthropogenic releases of Te were occurring in North America and elsewhere in the world well before Te was of economic value or interest. Given that Te concentration profiles correspond with changes in local and regional emission sources at several locations across Canada and that Te also has a high particulate/ solid phase affinity in aquatic systems under a wide range of pH-redox conditions (in mildly oxic to anoxic conditions under normal pH (such as lake sediment), Te should be predominantly in the solid Te0 form),10,48 we suggest that lake sediment cores are promising for reconstructing Te depositional fluxes. Quantitative Tellurium Deposition History. Most of the sediment records examined here either do not extend well into preindustrial times or find Te below detection limits during such times. However, the lakes of the ELA possessing both low dry mass accumulation rates and low bulk density of the accumulating sediments produce a high-resolution record of anthropogenically enhanced atmospheric metal(oid) deposition for the last 150 years (Figure 1f). Sediment Te concentrations have increased ∼6-fold during the 20th century at the ELA (mean 1900−2010 vs mean pre-1860) only exceeded by Sb and Pb (9-fold and 8-fold respectively) and greater than seen for Sn, As, Bi, Cd, Hg, and Zn (∼6 to 1-fold relative increase; Figure 2).

Figure 2. Mean Enrichment Factor (EF) averaged (±1 SD) over the post-1900 period relative to pre-1860 baseline (using Al as the normalizing cofactor excepting Hg19) for each of the sediment cores collected from the ELA in descending order of enrichment (for elements showing an EF > 1.2). Note that for comparison this includes such elements as Mn and Fe whose up-core enrichment is due to sediment redox processes and not to anthropogenic atmospheric metals deposition.

The anthropogenic sourced Te deposition history for each of the five ELA lake sediment records (Figure 3) reveals a quantitatively similar trend among the five different lakes over time. Briefly, anthropogenic Te deposition to the ELA increased markedly after 1880 (not long after the establishment of major copper smelting in Michigan’s Northern Peninsula and small-scale gold mining near the ELA,19 with a further sharp rise after 1920, peak deposition occurring during the 1940s −1950s declining in the 1960s and modestly thereafter (Figure 2)). The general agreement between ELA lake sediment based reconstructions suggests the sediment records of each lake are responding to an external supply of Te in at least a semiconservative manner. Modern and Preindustrial ELA Te Deposition Compared with Precipitation. The few Te measurements of precipitation in the literature3 range between 2 and 25 ng L−1, consistent with the deposition of Te recorded in ELA sediment records. Assuming the anthropogenic deposition to the ELA (Figure 2) was entirely in the form of wet deposition (ELA precipitation ∼700 L m−2 yr−1),25 a mean precipitation concentration of ∼10 ng L−1 Te of anthropogenic origin would suffice to explain the deposition recorded in lake sediments for recent times while ∼15 ng L−1 Te would account for fluxes during the period of peak deposition (1940s−1950s). We estimate natural or approximately preindustrial Te deposition to be 1.75 ± 0.42 μg m−2 yr−1, which is highly compatible with the estimated natural global Te mass flow,15 (2012) being equivalent to 1.53 ± 0.64 μg m−2 yr−1. Assuming the preindustrial flux of Te to the ELA arrived primarily by wet deposition, a preindustrial Te precipitation concentration of 2.5 ± 0.6 ng L−1 is inferred, also compatible with the lower values previously measured in precipitation.3 Catchment and Hydrologic Controls on Lake Retention of Tellurium. Atmospheric deposition of heavy metal(oid)s to lake catchments has the potential to increase the apparent deposition recorded in lake sediment cores as a portion of the deposition intercepted by the terrestrial catchment may be exported to the lake, as is observed for Hg.19,49−51 The degree to which catchments export atmospheric sourced deposition to lakes is often assessed via the relation between lake sediment metal inventories and the ratio E



Article

Figure 3. Anthropogenic Te deposition at the ELA reconstructed from lake sediment cores. Sediment derived anthropogenic Te fluxes corrected for lake specific Te background (mean pre-1860 Te/Al), sediment focusing (FF), and changing sedimentation are shown in a). Average anthropogenic Te fluxes for the ELA lakes are shown in b) with uncertainty bars (±95% Confidence Intervals).

of catchment to lake area (CA/LA) of low order lakes within a background region.19,49,51 The slope of a positive relation between these two measures infers catchment loading and can be used to infer the proportion of catchment deposition exported to lakes in a region. We find that Te is unique in being the only metal(oid) where a strong negative linear relationship was observed for inventory versus CA/LA (see Figure S6). This has two implications: 1) catchment export of Te is not a factor (or is strongly overshadowed by another process) and 2) there exists some mechanism of Te loss from lakes which is strongly related to lake-catchment morphology. To our knowledge this behavior has not previously been reported for heavy metal(oid)s but is like that of dissolved and total organic carbon (DOC, TOC) retention in lakes25,52 due to a negative relation between catchment to lake area ratios and lake water residence time in low order lakes.25,26,52 The retention of both DOC and TOC within lakes is positively related to lake water residence time with sediment burial and CO2 evasion predominating over outflow losses when lake water residence is >2 years.25,52 Like DOC,25 lake sediment inventory of anthropogenic Te versus lake water renewal times (Tw) follows a hyperbolic function (Figure 4a) signifying that Te delivery to lake sediments is less complete in lakes with higher water throughflow. Hyperbolic regression predicts the maximum anthropogenic Te inventory (1.042 ± 0.006 mg m−2) at infinite water renewal time and that sediment retention will exceed outflow losses when water residence time ≥0.84 ± 0.04 years (Tw for 1/2 max). Expressing the lake water renewal times instead as the areal discharge rate (m yr −1 ) versus anthropogenic Te inventory (Figure 4b) again shows that sediment Te inventory decreases with higher lake water turnover (higher qs), with the theoretical anthropogenic Te inventory of a ELA lake with no water outflow being 1.053 ± 0.024 mg m−2. Using this estimate of maximal anthropogenic Te inventory, we calculated the apparent settling velocity of Te in the water to be 7.03 ± 1.23 m yr−1 (Figure 4c) or approximately twice that previously calculated for DOC and total nitrogen for the central Ontario lakes26,53 and about half of that previously calculated for phosphorus in southern Ontario lakes.54,55 As the behavior of Te contrasts with other metal(loid)s but appears akin to C, N, and P, Te may also undergo considerable biological uptake and processing within the water column before entering the sedimentary environment.56

Figure 4. ) Low order ELA lakes show strong relation between anthropogenic Te inventory and a) water renewal times and b) areal water discharge (discharge/lake area). The theoretical maximal anthropogenic Te inventory for the ELA (1.05 mg m−2) calculated in b) allows determination of Te retention (R) in ELA lakes. From the inverse of lake Te retention vs the areal discharge shown in c) we calculated the apparent settling rate of Te (7.03 m yr−1 ± 1.23 (1 SD)).

F



■

ACKNOWLEDGMENTS We thank Dr. Michael Paterson (Chief Research Scientist IISDExperimental Lakes Area) who provided the ELA catchment and hydrology data. We would also like to thank Igor Lehnherr, Colin Cooke, Jing Ma, and CCIW Technical Operations for field support as well as Amy Sett, Greg Lawson, Xiaowa Wang, Carolyn Tunks, Jana Coty, and Alyah Thawer for laboratory support, the staff of the National Laboratory for Environmental Testing (NLET), and Dr. Roland Hall (University of Waterloo) who provided facilities for additional QA-QC for sediment core dating. Funding for this research was provided by the Government of Canada under the Clean Air Regulatory Agenda/Climate Change and Air Pollutant (CARA/CCAP), Joint Oil Sands Monitoring (JOSM) programs.

Unfortunately, studies on Te burdens in natural populations are few. Soil invertebrates reportedly show high Te enrichment relative to environmental abundances,54 and one study on the Te content of fish also reported high values (0.1−1.2 mg Te kg−1 fresh weight).57 In a fungal culture study >80% of media Te(IV) was bioaccumulated (bioaccumulation of Bi(III) > Te(IV) > Hg(II) > Se(IV) > Te(VI) > Sb(III)).58 These observations suggest community bioaccumulation and biomagnification of Te may occur. Anthropogenic releases of Te to the atmosphere have elevated Te deposition to the landscape both near and far from major metallurgical centers in central Canada and likely elsewhere in the world. Reconstructions of atmospheric Te deposition history using dated lake sediment cores are in agreement with limited independent data sources and are promising for further work, although low environmental Te abundances and associated analytical issues create some challenges, particularly for defining preindustrial Te levels. With results from the ELA, Dorset and Kejimkujik indicate that long-range atmospheric transport and deposition of Te are significant, with likely contribution from multiple distant sources. As monitoring data is absent, lake sediment core based reconstructions of fluxes and inventories of Te and other high-tech elements are crucial to understand both past and present anthropogenic loadings. The low apparent settling velocity for Te (similar to macronutrients; C, N, and P) despite its high particulate matter affinity10,48 implies that some process(s) are acting within the aquatic environment slowing its apparent descent, possibly significant biological Te uptake and reprocessing. While Te is normally rare in the environment, it is highly toxic for most bacteria, with effects seen at concentrations 100× lower than required to produce toxic effects for more common elements of concern (Se, Cr, Hg, and Cu).7 As Te utilization and potential human and environmental exposure has greatly increased in the past decade and is likely to increase further, it would be prudent to acquire a better understanding of Te interactions within the environment.

■

■

REFERENCES

(1) Wedepohl, K. H. The composition of the continental crust. Geochim. Cosmochim. Acta 1995, 59, 1217−1232. (2) Kavlak, G.; Graedel, T. E. Global anthropogenic tellurium cycles for 1940−2010. Resources, Conservation and Recycling 2013, 76, 21−26. (3) Belzile, N.; Chen, Y.-W. Tellurium in the environment: A critical review focused on natural waters, soils, sediments and airborne particles. Appl. Geochem. 2015, 63, 83−92. (4) Shie, M. D.; Deeds, F. E. The importance of Tellurium as a health hazard in industry. A preliminary report. Public. Health Rep. 1920, 35 (16), 939−954. (5) Taylor, A. Biochemistry of Tellurium. Biol. Trace Elem. Res. 1996, 55, 231−239. (6) Widy-Tyszkiewicz, E.; Piechal, A.; Gajkowska, B.; Smialek, M. Tellurium induced cognitive deficits in rates are related to neuropathological changes in the central nervous system. Toxicol. Lett. 2002, 131, 203−214. (7) Chasteen, T. G.; Fuentes, D. E.; Tantaleán, J. C.; Vásquez, C. C. Tellurite: history, oxidative stress and molecular mechanisms of resistance. FEMS Microbiol. Rev. 2009, 33 (4), 820−832. (8) Roy, S.; Hardej, D. Tellurium tetrachloride and diphenyl ditelluride cause cytotoxicity in rat hippocampal astrocytes. Food Chem. Toxicol. 2011, 49, 2564−2574. (9) Vij, P.; Hardej, D. Evaluation of tellurium toxicity in transformed and non-transformed human colon cells. Environ. Toxicol. Pharmacol. 2012, 34, 768−782. (10) Wu, X.; Jinming, S.; Xuegang, L. Occurrence and distribution of dissolved tellurium in Changjiang River estuary. Chin. J. Oceanol. Limnol. 2014, 32 (2), 444−454. (11) MAK 2012. Tellurium and its inorganic compounds [MAK Value Documentation, 2006]. The MAK Collection for Occupational Health and Safety; 2012; pp 82−307. DOI: 10.1002/ 3527600418.mb1349480vere0022. (12) Cunha, R. L.; Gouvea, I. E.; Juliano, L. A glimpse on biological activities of tellurium compounds. An. Acad. Bras. Cienc. 2009, 81, 393−407. (13) Hein, J. R.; Koschinsky, A.; Halliday, A. N. Global occurrence of tellurium-rich ferromanganese crusts and a model for the enrichment of tellurium. Geochim. Cosmochim. Acta 2003, 67, 1117−1127. (14) Zelenski, M. E.; Fischer, T. P.; Maarten der Moor, J.; Marty, B.; Zimmermann, L.; Ayalew, D.; Nekrasov, A. N.; Karandashev, V. K. Trace elements in the gas emissions of Erta Vale volcano, Afar, Ethiopia. Chem. Geol. 2013, 357, 95−116. (15) Sen, I. S.; Peucker-Ehrenbrink, B. Anthropogenic Disturbance of Element Cycles at the Earth’s Surface. Environ. Sci. Technol. 2012, 46, 8601−8609. (16) USGS 2010 minerals Yearbook. Selenium and Tellurium [Advance release]; 2012. (17) USGS 2015 minerals Yearbook. Selenium and Tellurium [Advance release]; 2017. (18) Hayes, S. M.; Spry, P. G.; Spaleta, K. J.; Skidmore, A. E.; Knight, N.; Kinght, D.; Milke, K. Mineralogy and chemistry of tellurium in the

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.7b06242. Additional methods, figures (site locations, geochronologies, Flin Flon Te inventories vs distance, anthropogenic Te flux history for other locations, flux history of Plastic Lake ON for major metal products of Sudbury metallurgical district, ELA Te inventories vs CA/LA), and tables (site locations and characteristics, method detection limits and method recoveries of SRMs) (PDF)

■

Article

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Johan A. Wiklund: 0000-0002-2327-5651 Jane L. Kirk: 0000-0001-6888-0727 Notes

The authors declare no competing financial interest. G



Article

hypogene and supergene environments. 2005; USGS Award MRERP Number# USGS/G12AP20054. (19) Wiklund, J. A.; Kirk, J. L.; Muir, D. C. G.; Evans, M.; Fan, F.; Keating, J.; Parsons, M. T. Anthropogenic mercury deposition in Flin Flon Manitoba and the Experimental Lakes Area Ontario (Canada): A multi-lake sediment core reconstruction. Sci. Total Environ. 2017, 586, 685−695. (20) Muir, D. C. G.; et al. Spatial trends and historical deposition of mercury in Eastern and Northern Canada inferred from lake sediment cores. Environ. Sci. Technol. 2009, 43, 4802−4809. (21) Loring, D. H. Normalization of heavy-metal data from estuarine and coastal sediments. ICES J. Mar. Sci. 1991, 48, 101−115. (22) Kersten, M.; Smedes, F. 2002. Normalization procedures for sediment contaminants in spatial and temporal trend monitoring. J. Environ. Monit. 2002, 4, 109−115. (23) Pirrone, N.; Allegrini, I.; Keeler, G. J.; Nriagu, J. O.; Rossmann, R.; Robbins, J. Historical atmospheric mercury emissions and depositions in North America compared to mercury accumulations in sedimentary records. Atmos. Environ. 1998, 32 (5), 929−940. (24) Lamborg, C. H.; Fitzgerald, W. F.; Damman, A. W. H.; Benoit, J. M.; Balcom, P. H.; Engstrom, D. R. Modern and historic atmospheric mercury fluxes in both hemispheres: global and regional mercury cycling implications. Global Biogeochem. Cycles 2002, 16 (4), 51-1−5111. (25) Curtis, P. J.; Schindler, D. W. Hydrologic control of dissolved organic matter in low-order Precambrian shield lakes. Biogeochemistry 1997, 36, 125−138. (26) Dillon, P. J.; Molot, L. A. Dissolved organic and inorganic carbon mass balances in central Ontario lakes. Biogeochemistry 1997, 36, 29−42. (27) Evans, H. E.; Dillon, P. J.; Molot, L. A. The use of mass balance investigations in the study of the biogeochemical cycle of sulfur. Hydrol. Processes 1997, 11, 765−782. (28) SaskPower 2017-Online information: http://www.saskpower. com/our-power-future/our-electricity-system/ (accessed April 27, 2017). (29) Chiou, K. Y.; Manuel, O. K. Tellurium and selenium in aerosols. Environ. Sci. Technol. 1986, 20 (10), 987−991. (30) Wittrock, V. Climatic extremes and energy sectors vulnerability: Now and in the future − focus on the Canadian portion of the Souris River watershed: A literature review. Prepared for Environmental Systems Assessment Canada Ltd. as part of the Natural Resources Canada Adaptation Platform Energy Working Group. Saskatchewan Research Council (SRC), Environment Division: 2016; SRC Publication No. 13757-1E16, 78pp. (31) McClenaghan, M. B.; Peter, J. M. Till geochemical signatures of volcanogenic massive sulphide deposits in glaciated terrain: summary of Canadian examples; Geological Survey of Canada: 2013; Open File 7354, 36pp, DOI: 10.4095/292803. (32) CDTC (Canadian Department of Trade and Commerce Dominion Bureau of Statistics Mining, Metallurgical and Chemical Branch). Chronological record of Canadian mining events from 1604− 1943 and historical tables of the mineral production of Canada; Edmond Cloutier Printer to the King’s most excellent Majesty: Ottawa, 1945. (33) Gibson, H.; Lafrance, B.; Pehrsson, S.; DeWolfe, M.; Gilmore, K.; Simard, R.-L. The Volcanological and Structural Evolution of the Paleoproterozoic Flin Flon Mining District, Manitoba: The Anatomy of a Giant VMS System; Geological Association of Canada− Mineralogical Association of Canada Joint Annual Meeting, Field Trip Guidebook FT-C3; Manitoba Innovation, Energy and Mines, Manitoba Geological Survey, 2013; Open File OF2013-6, 47. (34) USGS Copper statistics U.S. Geological Survey, World Copper Production History. Available online: https://minerals.usgs.gov/ minerals/pubs/historical-statistics/#copper (accessed Aug 2017). (35) Goldfarb, R. J.; Berger, B. R.; George, M. W.; Seal, R. R. Tellurium, Chapter R of. Critical mineral resources of the United StatesEconomic and environmental geology and prospects for future supply; Schulz, K. J., DeYoung, J. H., Seal, R. R., Bradely, D. C., Eds.; 2017; USGS Professional Paper 1802, pp R1−R27, DOI: 10.3133/pp1802R.

(36) M-STEM (Manitoba Science, Technology, Energy and Mines). Mineral Inventory File #573, Thompson Mine, 1983. (37) Perkins, W. T. Extreme selenium and tellurium contamination in soils  An eighty year-old industrial legacy surrounding a Ni refinery in the Swansea Valley. Sci. Total Environ. 2011, 412−413, 162−169. (38) Alzahrani, A. Determination and speciation of Tellurium in environmental samples using hydride generation atomic fluorescence spectrometry. MSc Thesis, Laurentian University, Sudbury, Ontario, Canada, 2013. (39) SARA (Sudbury Area Risk Assessment). Chapter 3, Historical review of air emissions from the smelting Operations (see table 3.6). In Sudbury Soils Study, Volume I: Background, Study Organization and 2001 Soil Survey; Prepared by the SARA Group c/o AECOM 2009 http://www.sudburysoilsstudy.com/EN/media/volume_I.asp. (40) Bornhorst, T. J.; Barron, R. J. Copper deposits of the western Upper Peninsula of Michigan. In Geological Society of America Field Guide 24; Miller, J. D., Hudak, G. J., Wittkop, C., McLaughlin, P. I., Eds.; 2011; pp 83−99, DOI: 10.1130/2011.0024(05). (41) Robert, F.; Poulsen, K. H. World-class Archaean gold deposits in Canada: An overview. Aust. J. Earth Sci. 1997, 44, 329−351. (42) Giddings, S. D.; Perkins, D. Gold Telluride Mineralization at the Goldlund Mine, Northwestern Ontario. I. Ore Mineral Log. Can. Mineral 1987, 25, 659−665. (43) Ispolatov, V.; LaFrance, B.; Dube, B.; Creaser, R.; Hamilton, M. Geologic and Structural Setting of Gold Mineralization in the Kirkland Lake-Larder Lake Gold Belt, Ontario. Econ. Geol. Bull. Soc. Econ. Geol. 2008, 103, 1309−1340. (44) Mériaud, N.; Jébrak, M. Review: From intrusion-related to orogenic mineralization: The Wasamacdeposit, Abitibi Greenstone Belt, Canada. Ore Geol. Rev. 2017, 84, 289−308. (45) Bates, J. L. E. Gold in Nova Scotia; Nova Scotia Department of Mines and Energy: Halifax, Nova Scotia, 1987; C87-099804-8. (46) Kontak, D. J.; Smith, P. K. A metaturbidite-hosted lode gold deposit: The Beaver Dam Deposit, Nova Scotia. I. vein paragenesis and mineral chemistry. Can. Mineral. 1993, 31, 471−522. (47) Dolor, M. K.; Helz, G. R.; McDonough, W. F. 2009. Sediment profiles of less commonly determined elements measured by laser Ablation ICP-MS. Mar. Pollut. Bull. 2009, 59, 182−192. (48) Harada, T.; Takahashi, Y. Origin of the difference in the distribution behavior of tellurium and selenium in a soil-water system. Geochim. Cosmochim. Acta 2008, 72, 1281−1294. (49) Sorensen, J. A.; Glass, G. E.; Schmidt, K. W.; Huber, J. K.; Rapp, G. R., Jr Airborne mercury deposition and watershed characteristics in relation to mercury concentrations in water, sediments, plankton, and fish of eighty northern Minnesota Lakes. Environ. Sci. Technol. 1990, 24 (11), 1716−1727. (50) Harris, R. C.; et al. Whole-ecosystem study shows rapid fishmercury response to changes in mercury deposition. Proc. Natl. Acad. Sci. U. S. A. 2007, 104 (42), 16586−16591. (51) Blais, J. M.; Kalff, J. Atmospheric loading of Zn, Cu, Ni, Cr, and Pb to lake sediments: The role of catchment, lake morphometry, and physico-chemical properties of the elements. Biogeochemistry 1993, 23, 1−22. (52) Algesten, G.; Sobek, S.; Bergström, A.-K.; Ǻ gren, A.; Tranvik, L. J.; Jansson, M. Role of lakes for organic carbon cycling in the boreal zone. Global Change Biol. 2004, 10, 141−147. (53) Molot, L. A.; Dillon, P. J. Nitrogen mass balances and denitrification rates in central Ontario Lakes. Biogeochemistry 1993, 20, 195−212. (54) Kirchner, W. B.; Dillon, P. J. An empirical method of estimating the retention of phosphorus in lakes. Water Resour. Res. 1975, 11 (1), 182−183. (55) Chapra, S. C. Comment on ’An Empirical Method of Estimating the Retention of Phosphorus in Lakes’ by W. B. Kirchner and P. J. Dillon. Water Resour. Res. 1975, 11 (6), 1033−1034. (56) Gongal’skii, K. B.; Krivolutskii, D. A. Rare and scattered elements in soil invertebrates. Dokl. Biol. Sci. 2003, 389, 134−137. H



Article

(57) Mert, R.; Alaş, A.; Bulut, S.; Ö zcan, M. M. Determination of heavy metal contents in some freshwater fish. Environ. Monit. Assess. 2014, 186, 8017−8022. (58) Boriová, K.; Ceranasky, S.; Matus, P.; Bujdos, M.; Simonivcova, A. Bioaccumulation and biovolatilization of various elements using filamentous fungus Scopulariopsis brevicaulis. Lett. Appl. Microbiol. 2014, 59, 217−223.

I


Widespread Atmospheric Tellurium Contamination in Industrial and

Recommend Documents