Use and Legacy of Mercury in the Andes - American Chemical Society

Apr 18, 2013 - Department of Anthropology, Yale University, New Haven, Connecticut 06520, United States. ∥. Department of Chemistry, Trent Universit...
1 downloads 5 Views 3MB Size
Article pubs.acs.org/est

Use and Legacy of Mercury in the Andes Colin A. Cooke,*,†,‡ Holger Hintelmann,∥ Jay J. Ague,† Richard Burger,§ Harald Biester,⊥ Julian P. Sachs,# and Daniel R. Engstrom∇ †

Department of Geology and Geophysics, Yale University, New Haven, Connecticut 06520, United States Department of Anthropology, Yale University, New Haven, Connecticut 06520, United States ∥ Department of Chemistry, Trent University, Peterborough, Ontario K9J 7B8, Canada ⊥ Institute for Geoecology, Technical University-Braunschweig, Langer Kamp 19c, 38106 Braunschweig, Germany # School of Oceanography, University of Washington, Seattle, Washington 98195, United States ∇ St . Croix Watershed Research Station, Science Museum of Minnesota, Marine-on-St. Croix, Minnesota 55047, United States §

S Supporting Information *

ABSTRACT: Both cinnabar (HgS) and metallic mercury (Hg0) were important resources throughout Andean prehistory. Cinnabar was used for millennia to make vermillion, a red pigment that was highly valued in pre-Hispanic Peru; metallic Hg0 has been used since the mid-16th century to conduct mercury amalgamation, an efficient process of extracting precious metals from ores. However, little is known about which cinnabar deposits were exploited by pre-Hispanic cultures, and the environmental consequences of Hg mining and amalgamation remain enigmatic. Here we use Hg isotopes to source archeological cinnabar and to fingerprint Hg pollution preserved in lake sediment cores from Peru and the Galápagos Islands. Both pre-Inca (pre-1400 AD) and Colonial (1532−1821 AD) archeological artifacts contain cinnabar that matches isotopically with cinnabar ores from Huancavelica, Peru, the largest cinnabar-bearing district in Central and South America. In contrast, the Inca (1400−1532 AD) artifacts sampled are characterized by a unique Hg isotopic composition. In addition, preindustrial (i.e., pre-1900 AD) Hg pollution preserved in lake sediments matches closely the isotopic composition of cinnabar from the Peruvian Andes. Industrial-era Hg pollution, in contrast, is distinct isotopically from preindustrial emissions, suggesting that pre- and postindustrial Hg emissions may be distinguished isotopically in lake sediment cores.



INTRODUCTION Cinnabar (HgS) forms a bright red pigment (vermillion) when powdered. In the South American Andes, vermillion is found in graves of high-status individuals and as a paint covering funerary masks and adorning ceremonial artifacts (Figure 1A). Vermillion has been recovered in association with a range of archeological artifacts spanning one of the first (Chavı ́n) to the last (Inca) pre-Hispanic Andean civilizations. Deposits of cinnabar are known from a range of hydrothermal ore deposits located across Central and South America,1 the largest of which is the Huancavelica quicksilver district in central Peru (Figure 1B).2,3 It has been suggested that Huancavelica cinnabar was mined and traded for in pre-Hispanic times,4 but clear confirmation is lacking, as is information regarding other possible cinnabar sources. A renewed interest in cinnabar, this time as a source of metallic mercury (Hg0), occurred after Spanish conquest of the Andes in 1532 AD. By ∼1570 AD, metallic Hg0 was relied upon across Central and South America to conduct mercury amalgamation, a technological development that allowed for the extraction of silver and gold from even low-grade ores.5,6 © XXXX American Chemical Society

The rapid adoption of mercury amalgamation across the Americas stimulated cinnabar mining on an unprecedented scale, and the cinnabar mines within the Huancavelica quicksilver district, the largest of which was the Santa Barbara mine,2,3 supplied much of the metallic Hg 0 used for amalgamation. Mercury amalgamation dominated silver production globally until ∼1900 AD,7 and it is estimated to have emitted >100 Gg of gaseous Hg0 to the global atmosphere.6,8 Resolving the geographic scope and biogeochemical impact of preindustrial mercury mining and emissions is important because mercury can be recycled repeatedly between various earth-surface compartments9 and may persist for centuries in biogeochemically active pools before being sequestered in soils or sediments.10,11 High-precision measurements of Hg isotopes have afforded new insight into source apportionment and the biogeochemical Received: November 26, 2012 Revised: April 1, 2013 Accepted: April 3, 2013

A

dx.doi.org/10.1021/es3048027 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

accordance with published methodologies.17,18 Additional details are provided in the Supporting Information. Analytical uncertainty was evaluated using replicate analyses of both the UM-Almadén standard and the certified standard reference materials (CRM) MESS-3 (marine sediment) and NIST 1944 (New York/New Jersey waterway sediment); the CRM were processed and analyzed in the same manner as our samples. The results for these standards (Table S1) were indistinguishable (within uncertainty ranges) from published values.15,17,19,20 We estimate a typical analytical uncertainty of a given isotope ratio as 2 SD of the measurement of the ratio in procedural standards (e.g., δ202Hg uncertainty = 0.09‰). Digestion duplicates yielded δ202Hg values with 2 SD uncertainty ranges spanning ±0.02‰ to ±0.56‰ (δ202Hg); ±0.02‰ to ±0.09‰ (Δ199Hg) and 0.01‰ to ±0.09‰ (Δ201Hg) (Table S1). While we cannot readily explain the relatively large δ202 Hg heterogeneity observed in sample H5, it seems likely that this reflects heterogeneity within the original deposit (discussed below). Cinnabar Ores and Archeological Artifacts. To constrain the Hg isotopic signature of cinnabar sources across the South American Andes, we obtained 12 samples of cinnabar ore from eight deposits (Table 1; Figure 1B). Our samples,

Figure 1. Artifact photos and sample locations. (A) Photographs of some of the artifacts included in this study; the red pigment on each item is cinnabar. Details about the provenance, cultural affiliation, and age of each artifact are provided in Table 2; photographs of the artifacts not shown here (samples A7−8, A10, A12−13, and A15−16) are included in the Supporting Information. (B) Map showing the locations of cinnabar samples (C#; white squares), archeological artifacts (A#; gray squares), and lake sediment cores (red stars). Also indicated is the approximate extent of the Inca Empire at the time of Spanish conquest (1532 AD), which extended across much of modern-day Peru and Ecuador, and parts of Colombia, Bolivia, Chile, and Argentina. Sample labels in both panels correspond to Tables 1 and 2.

Table 1. List of Cinnabar Ores, with Sample IDs Corresponding to Labels in Figures 1 and 2

cycling of Hg.12−14 There are seven stable isotopes of Hg (196−204 amu), and mass-dependent fractionation (MDF) of Hg isotopes is expressed as δ202Hg, which is derived using standard delta notation: δxHg (‰) = ([(xHg/198Hg)unknown/ (xHg/198Hg)NIST SRM 3133] − 1) × 1000, where (xHg/198Hg)NIST SRM 3133 is the average Hg isotope ratio of bracketing standards.15 MDF of Hg is known to occur during reduction−oxidation transformations, biological cycling, and volatilization of Hg. Hg isotopes can also undergo massindependent fractionation (MIF) in which the even- and oddmass-number isotopes fractionate from each other.12 MIF of Hg isotopes is reported as the deviation of a measured δ202Hg value from that theoretically predicted on the basis of MDF. MIF is thus reported as Δ199Hg and Δ201Hg in per mil (‰), and for variations ≤5‰ can be calculated using the following: Δ199Hg = δ199Hg − (δ202Hg × 0.252) and Δ201Hg = δ201Hg − (δ202Hg × 0.752);15 we use Δ199Hg as the default value to report MIF. MIF of Hg is largely produced by photochemical reactions and remains unchanged by most dark and biological transformations.12,16 Here we present the results of an interdisciplinary study using Hg isotopes to link resource exploitation, the archeological record, and environmental pollution. Our new Hg isotope data offer unprecedented insight into artisanal cinnabar mining and exchange and the environmental legacy of mercury mining in the Andes.

sample ID

location

H1−5 C1 C2 C3 C4 C5 C6 C7

Huancavelica, Peru Jalaca, Honduras Antioquia, Colombia Quindio, Colombia Chonta, Peru Cerro Colorado, Bolivia Mina de Pedernal, Bolivia Algarrobo Mine, Chile

which were obtained from museum mineral collections, are by no means a complete sampling of Andean cinnabar occurrences. But they offer an initial evaluation of Hg isotope heterogeneity across the Andes and include the only two deposits (Huancavelica and Chonta, Peru) mined historically (i.e., since 1532 AD).1 We hypothesized that Hg isotopes might be used to provenance archeological cinnabar. To test this hypothesis, we obtained samples of cinnabar found either as an offering (samples A1−7) or as a pigment covering ceramic (sample A8), metal (sample A9), or wooden (samples A10−17) artifacts (Table 2; Figure 1). A complete list of the objects sampled, including information about the archeological site, cultural affiliation, and age, is provided in Table 2; additional details about each item are provided in the Supporting Information. The artifacts sampled include a number of pre-Inca (i.e., pre1400 AD), Inca (ca. 1400−1532 AD), and Colonial (1532− 1821 AD) objects. Lake Sediment Core Mercury Content. To reconstruct the history of Hg deposition, we collected sediment cores from two lakes, one located in the Peruvian Andes (Laguna Negrilla: 13° 09′ S, 72° 58′ W; 4125 masl) and one located on San Cristóbal Island, Galápagos archipelago, Ecuador (El Junco Lake: 0° 53′ S, 89° 28′ W; 660 masl) (Figure 1B). Details about the recovery, chronology, and Hg measurements of the Laguna Negrilla sediment core have been published pre-



MATERIALS AND METHODS Mercury Stable Isotopes. Hg isotopic compositions were measured using Thermo-Finnigan Neptune continuous-flow cold vapor generation MC-ICP/MS at Trent University and in B

dx.doi.org/10.1021/es3048027 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Table 2. List of Archaeological Cinnabar Samples: Sample IDs Correspond to Labels in Figures 1 and 2a sample id Pre-Inca Artifacts Al−2 A3−6 A7 A8 A9 A10−11

object

cinnabar in grave funerary offering cinnabar in ceremonial offering bottle rim gold funerary mask wooden mummy bundle masks

Colonial Artifacts A12−14 wooden drinking vessels (qeros) Inca Artifacts A15−17 wooden digging boards

archaeological site or region

cultural affiliation

age estimate

Kuntur Wasi Chongoyape Las Huacas

Chavı ́n Chavı ́n Cupisnique

Early Horiozn (800−300 BC) Early Horizon (800−300 BC) Initial Period (1200−800 BC)

Cerro Blanco Northern Peruvian coast Central Peruvian coast

Chavı ́n Lambayeque (Sican) unknown

Early Horiozn (800−300 BC) Late Intermediate Period (1000−1250 AD)

Cusco

Colonial

A12: 1670−1890 ADA13: 1670−1950 AD A14: 1520−1790 AD

Southern Peruvian coast

Inca

A15: 1400−1440 AD A16: 1460−1630 AD

A10: 1020−1150 AD All: 890−1020 AD

a Archaeological samples occur either as pure cinnabar entombed within a burial or offering (A1−7) or as a decorative pigment (A8−17). Samples A10−17 were radiocarbon dated, and the calibrated 2σ age ranges are also provided. More details about each artifact and the radiocarbon results are provided in the Supporting Information.

viously.21 El Junco Lake occupies an extinct volcanic crater on the southwest summit San Cristóbal Island in the Galápagos Archipelago. The lake, which is roughly circular with a diameter of ∼220 m, lies within a ∼320-m diameter crater. Stratocumulus cloudswhich are an efficient scavenger of atmospheric Hg2+ (ref 22)usually conceal the lake in a dense fog (locally called garúa), and strong surface winds channel air year round from the southeast. El Junco Lake is thus ideally situated to record the long-range atmospheric transport of Hg emitted from the South American mainland. The sediment cores from both lakes were dated using both 210 Pb (Table S2) and 14C (Table S3) radioisotopes, and agedepth models for the sediment cores (Figure S2) were generated using the CLAM software package.23 To quantify the total sediment Hg content, we freeze-dried and homogenized ∼200 mg of sediment from 1-cm intervals and measured the Hg content using a DMA-80 direct mercury analyzer. Duplicate analyses and the CRM MESS-3 were measured after every 10th sample and were always within 5% of each other (for duplicates) and certified values (for CRM). We also measured a suite of other sediment parameters including organic carbon and nitrogen contents (%C and %N), and the concentration of aluminum (Al), titanium (Ti), sodium (Na), and manganese (Mg) (Figure S3).

Figure 2. Three-isotope plot of δ202Hg and Δ199Hg values for cinnabar. The sample numbers correspond to the sample ID numbers in Table 1 (for the cinnabar ore) and Table 2 (for the artifacts) but without the capital letters. A significant and linear (r2 = 0.99, p = 0.001) relationship is noted within cinnabar samples from Huancavelica (black line). The majority of the pre-Inca (pre-1400 AD) and Colonial (1532−1900 AD) archeological cinnabar samples (circled) plot within the 2σ uncertainty ranges for the Huancavelica cinnabar ores (gray shading), suggesting Huancavelica was the source of cinnabar on these objects. In contrast, cinnabar from Inca-era (ca. 1400−1532 AD) artifacts plot away from the Huancavelica, pre-Inca, and Colonial cinnabar samples.



RESULTS AND DISCUSSION Ancient Cinnabar Mining and Exchange. Many of the cinnabar ores we analyzed exhibit both MDF and MIF of Hg (Figure 2; Table S1). Cinnabar δ202Hg values span ∼3.5‰, and Δ199Hg and Δ201Hg values span ∼0.4‰. There is, furthermore, greater variability in δ202Hg values within the Huancavelica ores than among the different deposits, reflecting the heterogeneity of this district. Our cinnabar samples were obtained from museum collections that are largely devoid of information concerning their geological provenance beyond their general location of collection. Formation conditions are, therefore, unknown for the majority of the samples we analyzed, making it difficult to explain the wide range of isotopic compositions observed (Figure 2). However, MDF of Hg has been observed previously

within a variety of hydrothermal ores, including cinnabar,24−28 and is thought to result from a combination of processes associated with the emplacement of Hg-bearing minerals.25,26 In contrast, significant MIF of Hg has not been observed previously in cinnabar, and we observe a linear relationship between MDF and MIF within Huancavelica ore (r2 = 0.99, p = 0.001). MIF of Hg is thought to be initiated by either the nuclear volume effect (NVE) or the magnetic isotope effect (MIE). The NVE, which generates smaller MIF (≤0.4‰) and a Δ199/201Hg ratio of 1.60−1.65, has been observed during C

dx.doi.org/10.1021/es3048027 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

liquid−vapor Hg0 evaporation, Hg2+ reduction in the absence of light, and Hg2+−thiol complexation.29−31 In contrast, the MIE produces larger MIF with a Δ199/201Hg ratio between 1.0 and 1.3, and is initiated during photoreduction of Hg2+ or methyl-Hg (ref 12). Our cinnabar samples are characterized by a Δ199/201Hg ratio of 1.23 ± 0.09 (Figure S1), suggesting that the MIE drives MIF in these samples. While photochemically initiated reduction of oxidized Hg2+ (to gaseous Hg0) is thought to be the main driver of MIF of Hg in natural samples,13,32 it cannot explain the MIF observed within our cinnabar samples. Instead, the MIF anomalies preserved within Central and South American cinnabar may indicate isotopic inheritance from interactions with sedimentary source-rocks, which, for example, are found in close association with the cinnabar deposits at Huancavelica.2 Coal, peat, and marine sediments all exhibit MIF of Hg generated by the MIE,13,32 and a similar suggestion was made by Sonke et al. to explain significant MIF anomalies of trace Hg found within sphalerite (ZnS).20 We also note that, given our small sample size, we cannot rule out the existence of other isotopic end-members. Regardless of the exact cause of MIF within cinnabar, the relatively high degree of isotopic variability among cinnabar deposits raises the possibility that Hg isotopes can be used to source cinnabar preserved as part of the archeological record. The archeological artifacts we analyzed can be broadly grouped according to their antiquity and Hg isotopic signatures. The first group (samples A1−11) includes all of the artifacts predating the Inca Empire (i.e., pre-1400 AD). Most of these pre-Inca artifacts do not reveal significant MIF of Hg, and, with the lone exception of sample A2, plot within the 2σ uncertainty range of the Huancavelica data (gray band) (Figure 2). The second group of artifacts (samples A15−17), which contains all of the Colonial era (1532−1821 AD) samples, is characterized by δ202Hg and Δ199Hg values that overlap closely with Huancavelica cinnabar. This is consistent with historical records that indicate Huancavelica was the only significant source of cinnabar exploited by the Spanish.2 The final group of artifacts all date to the Inca era (ca. 1400−1532 AD). The cinnabar adorning these items is characterized by Hg isotopic compositions that are significantly different from both the Huancavelica cinnabar ore and the other artifacts. The isotopic overlap between the Huancavelica ores and the majority of the artifacts is consistent with previous suggestions4 that the Huancavelica region was an important prehistoric source of cinnabar. However, we cannot eliminate the possibility that some of the archeological cinnabar might have come from deposits at Chonta in the Department of Huánuco, Peru (sample C4), Cerro Colorado in the Nor Chicas Province, Bolivia (sample C5), or other deposits that we did not analyze. We are also unable to constrain the Hg isotopic heterogeneity that may be present in these other cinnabar deposits. Historically, Chonta was exploited only briefly (