Article pubs.acs.org/est
Atmospheric Deposition of Indium in the Northeastern United States: Flux and Historical Trends Sarah Jane O. White,*,† Carrie Keach, and Harold F. Hemond Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States S Supporting Information *
ABSTRACT: The metal indium is an example of an increasingly important material used in electronics and new energy technologies, whose environmental behavior and toxicity are poorly understood despite increasing evidence of detrimental health impacts and human-induced releases to the environment. In the present work, the history of indium deposition from the atmosphere is reconstructed from its depositional record in an ombrotrophic bog in Massachusetts. A novel freeze-coring technique is used to overcome coring difficulties posed by woody roots and peat compressibility, enabling retrieval of relatively undisturbed peat cores dating back more than a century. Results indicate that long-range atmospheric transport is a significant pathway for the transport of indium, with peak concentrations of 69 ppb and peak fluxes of 1.9 ng/cm2/yr. Atmospheric deposition to the bog began increasing in the late 1800s/early 1900s, and peaked in the early 1970s. A comparison of deposition data with industrial production and emissions estimates suggests that both coal combustion and the smelting of lead, zinc, copper, and tin sulfides are sources of indium to the atmosphere in this region. Deposition appears to have decreased considerably since the 1970s, potentially a visible effect of particulate emissions controls instated in North America during that decade.
■
INTRODUCTION Humans have had a notable influence on the environmental cycling of metals, particularly over the last century (for example, see refs 1−4). The mining and smelting of metal ores, industrial applications of metals, use of consumer goods, disposal of waste, combustion of fossil fuels, and modification of hydrology, have altered natural metal cycling and in some cases resulted in increased human exposure to metals. Notable health issues have included neurocognitive deficits in children exposed to lead;5 elevated methylmercury concentrations in fish due to methylation of anthropogenically released mercury, and resulting in health advisories to limit fish consumption;6,7 and widespread arsenic poisoning in South Asia resulting from mobilization of subsurface arsenic to well water.8 More broadly, increased metal concentrations in the sedimentary record over the last century are one signature of the anthropocene, an era of anthropogenic dominance on the Earth and the environment.9 Rapid technological development in the past several decades, and a concomitant increase in the number of elements being employed in electronics and energy technologies, has resulted in the potential for increased environmental releases of these elements. Without adequate information about toxicity and environmental behavior, it is difficult to predict the potential consequences of such releases; timely availability of such knowledge may help predict whether a novel metal © XXXX American Chemical Society
contaminant will prove to be harmful to the environment or human health (for example, refs 2, 3, 10, and 11). One metal whose industrial use is rising rapidly is indium, an increasingly important metal in semiconductors and electronics. Primarily, indium is used as a light-transmissive conductive coating (in the form of ITO, indium tin oxide) for LCD displays, flat panel displays, and photovoltaic cells, or as a substrate in certain LEDs and photovoltaic cells.11 Despite the rapid increase in use of indium,11,12 very little is known about its environmental behavior, and concerns are emerging over its health impacts.11 Current releases of indium to the environment are now thought to be dominated by coal-fired power plants and the smelting of lead, zinc, copper, and tin sulfides,11,13,14 processes of which indium is a byproduct, but may be augmented in the future by releases from electronics manufacture or end-of-life disposal. There is significant uncertainty about these estimates, however, and few studies have characterized the sources of indium to the environment or how concentrations in the environment have changed over time. Received: June 30, 2015 Revised: October 1, 2015 Accepted: October 1, 2015
A
DOI: 10.1021/acs.est.5b03182 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology One approach to determining pollution histories is through the use of peat and sediment cores, which may contain a record of contaminant deposition. For example, an increase in lead emissions to the atmosphere in North America, associated with its industrial use beginning in the mid-1800s and accelerating in the 1920s, is well-documented in sediment cores, as is the dramatic decrease in lead emissions since the mid-1970s, which is widely attributed to the phaseout of leaded gasoline.15 Ombrotrophic bogs are particularly useful for tracking historical atmospheric deposition since they receive only atmospheric inputs, without confounding contributions from surface or groundwater inputs. In contrast to lead, far fewer sediment profiles have been examined for a record of indium deposition, and those that have been reported often suggest that indium deposition to sediments does not track indium’s industrial use. Instead, lake and sediment cores from North America, Australia, Europe, and Japan show increasing indium concentrations and/or fluxes beginning in the late 1800s or early 1900s, well before indium was widely used in industry.13,14,16−22 Additionally, most of these cores show a subsurface peak, suggesting that present-day atmospheric deposition fluxes are lower than such fluxes in recent decades. The inferred date of maximum deposition rate varies from as early as the late 1960s (Utah;20) to as late as 2000 (Japan;19), but falls generally during the 1970s and 1980s (Sweden:;16 Eastern Canada:;13,14 Australia:18). Several cores that are not dated or are of low resolution are consistent with this same general pattern.17,23 A few cores show more complicated histories with unclear trends for the last century.21,22,24 While the potential for postdepositional redistribution is unknown, indium has relatively low solubility and is thought to be heavily associated with particles.11 The authors are not aware, however, of any prior studies of atmospheric indium deposition in the northeastern U.S., whose airshed includes major industrial regions of North America and which historically has been a receptor of numerous contaminants, such as lead and acidic deposition, that are borne by long-range atmospheric transport. Therefore, the present study examines a nearly 150 year atmospheric deposition history of indium in the northeastern U.S. in order to determine its atmospheric sources and deposition over time. Retrieval of this record is enabled by the design and use of a novel freeze-corer that allowed the collection of a long and relatively undisturbed core from Thoreau’s Bog, an ombrotrophic bog in Massachusetts. Radioisotope dating, corroborated by cross-comparison with previous dated peat cores from this bog, enables interpretation of the core in terms of a record of historical atmospheric deposition of indium at this site.
Figure 1. Thoreau’s Bog is located within a larger wetland area in Concord, Massachusetts. There is no industry in close proximity; it is surrounded by residential, agricultural, and forested areas. Image from Google earth, ©2015 Google.
major industry in close proximity to the bog; the surrounding area is residential, agricultural, or forested. Boston is located 23 km downwind (southeast); the nearest upwind industrial centers are Worcester (43 km southwest) and Springfield (112 km southwest). Freeze-Corer. Sphagnum bogs can be difficult to core due to their compressibility and to the presence of woody plant roots; the corers presently available (for example, see refs 27−29) often have difficulty obtaining intact, uncompressed, and undisturbed cores at Thoreau’s Bog and many others like it.30,31 Past core studies at Thoreau’s Bog have generally been limited to the depths to which minimally disturbed cores can be cut with a sharp serrated knife or sampled by a sharpened, hammer-driven tube during winter freezing. To overcome this limitation we developed a corer that freezes a deep cylinder of peat, preserving its structure by providing mechanical strength and allowing it to be retrieved with minimal disturbance. It is based on a freeze-coring design previously used for lake sediments,32 with modifications to ensure minimal disturbance of the peat, and to allow the collection of a larger-diameter core. The corer is made up of two concentric tubes; the inner tube (the freeze-tube) used to freeze the peat, the other (the sawtube) used to extract the frozen core (Figure 2a). The freezetube, constructed of aluminum alloy 6061, is 1 m long, with a 2.5 cm outer diameter and 2.3 cm inner diameter. It is open on one end and capped at the other with a pointed, liquid-tight aluminum tip. This freeze-tube can be pushed into the peat
■
MATERIALS AND PROCEDURES Site Description. Thoreau’s Bog, known to Thoreau as Gowing’s Swamp, is an ombrotrophic, floating mat Sphagnum bog in Concord, Massachusetts (42°27′43″N, 71°19′42″W) (Figure 1). Henry David Thoreau first described its vegetation on February 17, 1854.25 Mainly ericaceous shrubs form a border around an open, floating Sphagnum mat, which is dominated by S. rubellum Wils. and S. magellanicum Brid., a species mix that has not changed significantly since Thoreau’s 1854 description. Its location on a water table divide results in the bog having no groundwater inputs; its water budget includes only precipitation, runoff, and evapotranspiration; there is no mineratrophic boundary.26 The water table is typically 10−20 cm below the moss surface.26 There is no B
DOI: 10.1021/acs.est.5b03182 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
did not contribute detectable indium to the peat, as discussed in the Supporting Information (SI). Several mm of peat were sliced away from the surfaces of the core that had been in contact with either the freeze-tube or the saw-tube, using stainless steel scissors that were cleaned with deionized water between samples. The scissors did not contribute indium to the peat samples, as confirmed in the same method as for the saw blade (SI). The dimensions of the core slices were measured in order to calculate volume, and the slices were weighed before and after drying. Although porewater could be extracted after slicing and before drying, indium was below detection in the few porewater samples analyzed, and therefore porewater was not systematically extracted. Each peat slice was dried in a 60 °C oven for 24− 48 h in a glass beaker covered by a glass watch glass. The dried peat was weighed, then homogenized in polystyrene vials with 8 mm methyl methacrylate balls in a SPEX CertiPrep 8000 Mixer/mill after manual removal of roots. All containers that contacted the peat were acid-washed, and blanks were tested at all steps of sample preparation and analysis, as described in more detail below and in Supporting Information. Radioisotope Dating. The peat core was dated using 210Pb and 137Cs chronostratigraphic techniques.33−35 After a 3-week equilibration period in polypropylene vials, samples were counted for 210Pb, 214Pb (as a proxy for supported 210Pb activity), and 137Cs on a γ-counter (Canberra Instruments, GL2020 detector with Series 40 multichannel analyzer). Model dates were assigned according to the Constant Rate of Supply (CRS) Model,33−35 which assumes a constant rate of input of 210 Pb on a year time scale (disintegrations per minute/(cm2yr)) (DPM/(cm2-yr)), but allows for variable sedimentation rates (g sediment/(cm2-yr)). This method therefore allows for any differences in net sedimentation rate that may occur among years, either due to changes in Sphagnum growth rate from year to year, or to compaction and decomposition that takes place as the peat ages. Of the 34 core samples analyzed, 25 were dated using this technique; intervening dates were linearly interpolated, and dates for the bottom several samples were linearly extrapolated for the purposes of plotting indium depth profiles. For complete details, see Supporting Information. Metal Analysis. Indium was analyzed using inductivelycoupled plasma−mass spectrometry (ICP-MS) after digestion of peat with nitric and perchloric acids. Homogenized peat samples (0.5−1 g dry weight) were digested based on a slightly modified EPA Method 3050B, taken to dryness, taken up in 2% nitric acid, then filtered with acid washed Whatman or VWR brand polypropylene 0.45 μm syringe filters. Acid blanks, reagent blanks, and filter blanks were carried through the procedures, and standard recoveries were calculated. See Supporting Information for complete digestion and quality control procedures. Indium analyses were performed using a Fisons PlasmaQuad 2+ Quadrupole Inductively-Coupled Plasma Mass Spectrometer (ICP-MS), using the conditions in Table 1. In order to account for matrix effects and drift of the instrument signal over time, the method of standard additions was used to quantify total indium. Each sample was split in two, one of the subsamples was spiked with 0.1 μg/L indium, and the two subsamples were then analyzed back-to-back on the ICP-MS. Isotope 115 (95.7% of naturally occurring indium) was monitored, and testing of potential polyatomic interferences at 115 showed that there were none present for this sample matrix. There is, however, an isobaric interferent, 115Sn (0.34%
Figure 2. A 1 m long, minimally disturbed core can be obtained with a freeze corer developed for hard-to-core peat bogs. (a) An inner tube is pushed into the peat, then filled with ethanol and dry ice to freeze approximately 13 cm diameter of peat around the tube. The undisturbed core can then be cut from the mat with the outer tube, which makes use of sharpened teeth to cut through plant roots. (b) A core retrieved in 2010 from Thoreau’s Bog, MA.
with minimal effort and minimal disturbance to the peat. The saw-tube is a 91 cm long, 12.7 cm outer diameter steel tube (Acker Drill Company, Inc., Scranton, PA), with saw-teeth cut into the bottom edge and holes at the top to fit a handle. In operation, pellets of dry ice are added to fill the freezetube, which is then topped up with 95% ethanol as a heat transfer fluid. A polypropylene funnel fitted with an O-ring seal is attached to the top of the freeze-tube to assist with filling. Because dry ice sublimates and ethanol evaporates quickly, both must be fed into the tube constantly until the desired thickness of peat freezes. The bottom of the peat tends to freeze more quickly than the surface peat, which is not cooled as consistently during freezing. In our experience, a 13 cm diameter core can be obtained in about 1 h. After the desired peat thickness is frozen, the funnel is removed, and the saw-tube is used to cut roots and free the frozen core from the surrounding peat. When the saw-tube has cut through the peat as far as possible, or if the frozen core begins to be pushed down by the saw-tube, pulling up on the saw-tube at a slight angle brings with it the frozen core and the freeze-tube. The saw-tube can then be removed from the core, and the core photographed, measured, wrapped in plastic, and kept frozen until processing. Sample Preparation. A 90 cm core was collected from Thoreau’s Bog using the above-described corer in May of 2010 (Figure 2b). Warm water was poured into the freeze-tube, loosening the tube and allowing it to be removed. The core was then solidly refrozen and sliced with a stainless steel butcher saw (Weston Supply 47-1601) that was cleaned between cuts by rinsing with deionized water and shaken dry. Accurately cut 1 cm slices were obtained by this means down to 15 cm depth, below which the core was cut into 2 cm slices. The saw blade C
DOI: 10.1021/acs.est.5b03182 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
ppb spike of indium added to each sample dilution. Total lead was quantified by analyzing counts at mass to charge ratios (m/ z) of 206, 207, and 208; the instrument used is a singlecollector instrument, therefore accurate isotopic ratios are not available. While measurement of total lead based only on one of these m/z ratios would be sufficient, the fact that total concentrations calculated using each individual m/z were within 3% of one another provided confidence that interferences were not present at any of the m/z analyzed.
Table 1. Analytical Conditions for ICP-MS conditions
Fisons PlasmaQuad 2+
mass-charge ratio carrier gas resolution CPS for 1 ppb In sample introduction nebulizer mode
In:115; Pb: 206, 207, 208 Ar low (unit) 200,000 free draw 1000 μL/min with frit to prevent clogging peak-jumping, pulse counting 200 sweeps/measurement In: standard additions; Pb: internal standard using In
quantification
■
RESULTS The freeze coring method developed was successful in retrieving a core that was 3× longer than has previously been accomplished for this bog and appeared to be minimally disturbed, as evidenced by cleanly cut roots, no visible evidence that any portions of the core had rotated during coring, a recovery ratio of ∼1, and similar or lower densities to cores previously collected.26,38 Plots of dry density and water content over the depth of the core are shown in Figures S1 and S2. Both density and water content are heavily influenced by the age and stage of the peat. Living moss at the surface contributes to low density and high relative water content. As the peat ages,
of total Sn), which is linear and predictable, and was corrected for by monitoring 117Sn or 118Sn and subtracting the corresponding counts that would be attributed to 115Sn.36 Lead was analyzed for comparison with published Pb flux profiles in the northeastern United States as validation of the dating and metal preservation in the present core. Lead was quantified using indium as an internal standard to account for matrix effects and drift of the instrument signal over time.37 Because the samples analyzed for lead were diluted by at least 200x, natural indium concentrations were much less than the 1
Figure 3. (a) Activity of unsupported 210Pb decreases exponentially with depth, but does not reach zero; sources of error are discussed in the main text. 214Pb is a proxy for supported 210Pb activity. (b) 137Cs has three major peaks, one at the surface believed to result from active uptake by Sphagnum mosses, one at 37 cm reflecting the 1986 Chernobyl accident, and one at 55 cm reflecting the 1963 peak of nuclear bomb testing. (c) Dating for the freeze-core from Thoreau’s Bog goes back to 1894. Dark circles are the model dates with the measured inventory. Empty circles are the model dates assuming that the actual inventory is 4% more than was collected. Diamonds are 137Cs dates. Error bars given represent the 1σ statistical uncertainty associated with each point. D
DOI: 10.1021/acs.est.5b03182 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
depth x, A0 is the inventory of the full core, λ is the decay constant for 210Pb of 0.0311 yr−1, and t is the time before present, in years (Equation S1)). Fitting an exponential curve to the bottom 40 cm of the 210Pb activity (Figure 3a) and integrating the function from 81 cm to infinity suggests that about 1.4 DPM/cm2, or 4% of the total inventory, may be missing. Figure 3c shows the date calculations with both the inventory given (solid circles), and with the assumption that the total inventory is 4% higher (40 DPM/cm2 instead of 38.4 DPM/cm2, open circles). Indium Deposition. Indium Concentration and Flux. Maximum indium concentrations are found between depths of 63−71 cm, corresponding to dates of ca. 1917−1938; the maximum is broad (Figure 4(a), Table S2). Concentration
it is compressed, leading to peaks in density and minimum water content at ∼40 cm; below this depth, density decreases again as the peat degrades further. Expansion during freezing is expected to be largely radial, and therefore does not impact the vertical stratigraphy significantly. Dating. 210Pb. 210Pb activity for this core decreases approximately exponentially with depth and shows no prominent evidence of mixing or disturbance (Figure 3a, Table S1). The deepest sample analyzed for 210Pb was dated to 1894, though uncertainty in the bottom 30 cm is high (Figure 3c). The upper 46 cm of the core, however, are dated to within ±9 yr, and the upper 60 cm are dated to within ±15 yr. Error in 210 Pb DPM/g (Figure 3a) arises from three sources: counting statistics, error in the efficiency of detection of disintegrations, and error in the mass of sample; error is higher for the shallower samples primarily because a smaller mass of sample was used for their analysis as compared to the deeper samples. Error in the modeled dates (Figure 3c) arises from error in 210 Pb DPM/g, as discussed above, as well as error in estimates of peat density and error in sample depth. The larger date errors at depth arise primarily from the smaller 210Pb inventories for the deeper samples. The “apparent” sedimentation rate of peat (Figure S3) is higher at the surface, where peat is actively growing and density is low, and lower at depth, where peat has been degraded and compressed. The total 210Pb inventory for the core is 38.4 DPM/cm2, and the deepest peat is estimated to have been deposited 116 years before present, leading to an average calculated 210Pb flux to the bog of 0.33 DPM/(cm2-yr). This flux is somewhat lower than other values reported for latitudes 40−50 N (0.5−1.4 DPM/(cm2-yr);39); this may be due to an overestimate of the deepest age, as discussed below. However, within the range of depth overlap, this core’s age profile and 210Pb activity profile compares favorably with other dated peat cores from Thoreau’s Bog.26,38 137 Cs. The profile of 137Cs shows three peaks, at 55 cm, 37 cm, and the surface (Figure 3b). The deepest peak is attributed to 1963, the peak year of nuclear bomb testing, and is consistent with the date calculated from the 210Pb dating model. The peak at 37 cm depth is attributed to 1986, the year of the nuclear reactor meltdown at Chernobyl, and is also consistent with the 210Pb dating model. The presence of a readily identifiable peak in 137Cs deposition arising from the Chernobyl accident has also been documented by Lima et al.40 in a nearby anoxic estuary. However, the peak of 137Cs activity at the surface implies some degree of mobility of this nuclide in peat, as has also been observed elsewhere.34,41,42 The surface peak is likely due to active biological uptake of Cs by Sphagnum mosses, a phenomenon also reported in this bog for potassium, another monovalent ion.26 The concentration of 137Cs in mosses has also been reported elsewhere.43 Further investigation of cesium mobility in Sphagnum peat and living Sphagnum mosses is warranted. Dating Error. In addition to experimental error, much of the error for the oldest dates arises from the lower count rates associated with peat that is of the order of 5 half-lives old, as well as from uncertainty in the total 210Pb inventory of the core: this latter value is underestimated if a core does not go deep enough to capture the horizon at which the unsupported 210Pb activity is zero,44 as is the case with the present core (Figure 3a). This error leads to the oldest dates calculated being older than their true age (arising mathematically from 1 t = − λ ln(A 0 /Ax ), where Ax is the inventory of 210Pb below
Figure 4. Indium concentrations in Thoreau’s Bog have been increasing since at least the early 1900s, peak broadly near 1940, and decrease to the present. Indium fluxes have been increasing since at least the early 1900s, peak in the late 1960s/early 1970s, and decrease to the present. Error bars are discussed in the main text; at a minimum, they represent a relative procedural error of 20% (1σ).
decreases from this maximum toward the surface. Error represented as central data points flanked by error bars is the minimum relative procedural error of 20% (1σ), based on two peat samples that were digested and analyzed in triplicate. In addition to this error, each sample digest was analyzed by ICPMS at least twice, and sometimes up to 5 times; where duplicate digest analyses produced relative error (1σ) greater than 20%, error is presented as a central line with a data point on each end. The flux profile of indium shows indium flux increasing from the late 1800s, peaking in the late 1960s/early 1970s, then decreasing to recent values lower than those for the year 1900 (Figure 4b, Table S2). The maximum indium flux of ca. 1.9 ng In/cm2-yr is calculated to occur at a more recent date than the E
DOI: 10.1021/acs.est.5b03182 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
■
Article
DISCUSSION The use of a freeze-coring methodology is shown to enable the retrieval of relatively long and minimally disturbed cores of peat from ombrotrophic bogs which, because of their unique hydrology and the sorptive characteristics of their peat, offer a unique opportunity to measure historical atmospheric deposition of many substances. Although to our knowledge this is the first published study of indium deposition in the northeastern United States, some similarity was expected between data from Thoreau’s Bog and data from eastern Canada, Sweden, Norway, and the western U.S. (Utah), particularly in regard to the decrease in atmospheric indium flux in recent decades. This expectation is borne out by the data,13,14,16,17,20 although the rate of decrease of deposition to Thoreau’s Bog appears to have been more rapid than elsewhere. The fact that the indium flux to Thoreau’s Bog and elsewhere is counter to indium’s exponential increase in consumption over the past 30 years (Figure 5a) suggests that neither electronics manufacturing nor disposal of electronic products is at present a dominant source of indium to the atmosphere. Many researchers have attributed the increase in metal fluxes as recorded in peat or sediment cores to increased industrial releases coinciding with the industrial revolution in the US and western Europe,13,15,16 and indeed the increase in indium flux to Thoreau’s Bog seen beginning in the late 1800s does track both U.S. coal consumption and metal production until ca. 1970 (Figure 5b and c). The indium content of coal is typically ca. 100 μg/g, and the indium content of zinc, lead, copper, and tin sulfides is typically 5−100 mg/kg.57,58 (Indium is not found naturally as its own mineral.) Several studies have shown that the smelting of sulfide ores is a significant source of indium to the atmosphere and surrounding surface environments,13,14,59,60 and our previous work suggests that both coal combustion and smelting are significant contributors of indium to the atmosphere.11 Because the closest upwind coal fired power plant is >40 km to the west/southwest of the bog and the closest upwind smelter is 325 km to the east/northeast of the bog, we conclude that long-range atmospheric transport is likely the dominant pathway for indium transport to Thoreau’s Bog. The large decrease in indium flux to the bog beginning in the 1970s appears to coincide with decreased PM2.5 emissions due to increased particulate control technologies for stationary sources (Figure 5b and c), regulated in the United States beginning in 1971 with the establishment of the National Ambient Air Quality Standards for Total Suspended Particles (a program of the Clean Air Act of 1970),61,62 and recommended in the early 1970s in Canada with the National Ambient Air Quality Objectives.63 Emissions controls did not specifically target emissions of indium, which has only recently been recognized as a contaminant of potential concern,11,64 but nevertheless appear to have had a significant impact on atmospheric concentrations of many particulate-associated metals. It is noted that many metals such as indium are byproducts of primary ores and are typically not subject to regulation or reporting requirements. At the same time, while this decrease in deposition of indium and other metals appears to offer evidence of the success of the Clean Air Act, decreases in long-range transport from large industrial facilities does not preclude the possible existence of severe local indium contamination due to local emissions, or emissions in countries
depth of highest indium concentration, resulting from the fact that peat is compressed and degraded as it ages. Flux Error. One flux point, corresponding to 77 cm depth, was omitted from the flux profile (Figure 4b). This was due to a net calculated peat deposition 4× higher than any other calculated rates, causing the calculated indium deposition flux to be 10x higher than earlier or later fluxes. This seemingly unreasonable value may have arisen because 210Pb activity was not measured at this depth, requiring the date to be estimated from interpolation; small changes in the date interval can cause significant changes in calculated sedimentation rates and subsequently on calculated fluxes. Because error on the model dates post-1940 are relatively small, the flux profile above this horizon is well-constrained. Metal Mobility. The dating of metal profiles in sediments depends on the assumption that those metals are strongly associated with the solid phase, causing them to be nearly immobile and to thus retain their vertical stratigraphy. Studies of indium mobility in peat are lacking. Our measurements of indium in porewater from several depths in this core are below analytical detection limits of about 5 ng/L. Taken with the sediment concentrations for those depths, a lower bound on a solid/aqueous partition coefficient, Kd, for indium in Thoreau’s Bog can be calculated to be 104.2 L/kg (where Kd = Cs/Caq; Caq is the concentration in the aqueous phase (g/L); and Cs is the concentration in the solid phase (g/kg)). Although Kds are best used on a site-specific basis, similar Kds of 104.5−105 L/kg have been reported for indium in lake sediments,16 and other studies have shown that indium migration in sediments is minimal, except in the presence of EDTA.13,45 Using our calculated lower bound Kd of 104.2 L/kg, In is calculated to travel at least 400 times more slowly than porewater travels in the bog (using a retardation factor, R = (mobile fraction + solid fraction)/ mobile fraction = 1 + Kdρbulk/η, where ρbulk is the dry density in g/cm3, and η is the porosity). Hemond26 reported precipitation of 1.45 m/y and evapotranspiration of 1.02 m/y averaged over a study period in excess of one year, yielding a maximum downward water velocity through the bog of 43 cm/yr (in fact, this is an upper bound, since water also leaves the bog via lateral flow). These values of Kd and vertical water flux translate into an upper bound on vertical indium migration of
