silica- and sulfate-bearing rock coatings in smelter areas ... · 2009; durocher and schindler,...

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 90 (2012) 221–241

Silica- and sulfate-bearing rock coatings in smelter areas:Part II. Forensic tools for atmospheric metal(loid)-

and sulfur-isotope compositions

Nathalie M. Mantha a, Michael Schindler a,⇑, T. Kurtis Kyser b

a Department of Earth Sciences, Laurentian University, Sudbury, ON, Canada P3E 2C6b Department of Geological Sciences and Geological Engineering, Queen’s University, Kingston, ON, Canada K7L 3N6

Received 5 March 2012; accepted in revised form 3 May 2012; available online 16 May 2012

Abstract

Black silica- and sulfate-bearing rock coatings in the Greater Sudbury area, Canada provide a record of atmospheric pro-cesses and emitted particulate matter associated with historical smelting operations in this area. Coating samples collectedover the Greater Sudbury region are characterized with scanning electron microscopy, electron microprobe analysis, laserablation inductively coupled mass-spectrometry, S-stable isotope measurements and micro X-ray fluorescence. On themicrometer scale, Cu, Pb, As, Se and S occur in close association within metal–sulfate rich layers composed of Fe- andCu-sulfates. The concentrations of these and other elements do not represent their chemical proportions in the smelter plumesdue to dissolution–precipitation processes, element substitutions and the stability of various phases involved in the coatingformation. On the regional scale, the atomic ratios of Pb:Ni, As:Ni and Se:Ni decrease in the coatings with increasing distancefrom the smelting centers. This observation is consistent with higher wet deposition rates of small diameter Pb, As and Se-bearing primary sulfate aerosols (<2.5 lm), compared to larger diameter (>2.5 lm) Ni-bearing particulate matter. The mixingof primary (higher d34S values) and secondary (lower d34S values) sulfates explains the d34S values of sulfates within the coat-ings close to smelting centers and the decrease in these values is attributed to the decrease in the ratio of primary to secondarysulfates with distance from the smelting centers. The information preserved in mineral surface-coatings together with anunderstanding of stoichiometry, geochemical processes and former environmental conditions provide a valuable record ofatmospheric compositions, mixing, scavenging, deposition rates and oxidation processes, and the nature and source of anthro-pogenic releases to the atmosphere.� 2012 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Surface coatings on rocks and building stones are usefulforensic tools for monitoring recent and past air pollutantsand tracing their sources. Coatings formed in deserts, inproximity to smelters and in highly polluted urban environ-ments may capture common pollutants such as metal-bear-ing particulate matter, sulfate aerosols and radionuclidesduring their formation (Nord et al., 1994; Hodge et al.,

0016-7037/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.gca.2012.05.013

⇑ Corresponding author.E-mail address: [email protected] (M. Schindler).

2005; Wayne et al., 2006; Smith and Prikryl, 2007; Hoar etal., 2010; Mantha et al., 2012). The forensic examinationof rock coatings is also valuable in providing an understand-ing of past terrestrial or planetary environments (i.e. onMars, McLennan, 2003; Milliken et al., 2008; Squyres etal., 2008), in which coatings may have preserved particlesand mineral phases containing information about past cli-matic conditions, volcanic events and features related tobiological activity.

An important fingerprinting tool for identifying sourcesof pollutants once present in the atmosphere is the identifi-cation of isotopic compositions of anthropogenic SOx.Studies have been undertaken to examine preserved traces


mailto:[email protected]

http://dx.doi.org/016/j.gca.2012.05.013

222 N.M. Mantha et al. / Geochimica et Cosmochimica Acta 90 (2012) 221–241

of atmospheric pollutants in urban areas (e.g. Normanet al., 2006; Sinha et al., 2008), around coal-fired powerplants (e.g. Forrest and Newman, 1977a) and smelters(Nriagu and Coker, 1978) as well as to interpret the relatedformation of weathering crusts on building stones (Buzekand Sramek, 1985). Complications may arise from applyingthis tool in environmental studies including the fraction-ation or absence of fractionation of S-bearing atmosphericspecies during different oxidation processes, the input fromdifferent S-sources, seasonal changes in oxidant concentra-tions as well as differences in deposition mechanisms.

The use of the chemical and sulfur-isotopic signatures ofcoatings as forensic tools for identifying environmentalconditions can be tested quite effectively using silica coat-ings formed in the vicinity of smelters, such as those formedin Greater Sudbury, Canada (Fig. 1a). These coatings wereformed in close proximity to a major point source of SO2

and metals, active for over 70 years and for which historicalrecords of local emissions and their chemical and isotopiccomposition are available. In this case, we can assume thatthe coating components originate predominantly from thelocal point source or sources whereas rock coatings formedin proximity to volcanoes (e.g. Schiffman et al.,, 2006), indeserts (e.g. Dorn, 1998; Perry et al., 2006; Schindler etal., 2010) and on Antarctica (Giorgetti and Baroni, 2007)have been affected by unknown inputs and processes thatmay range from local to global scales. Furthermore, anadvantage with respect to coatings formed of encrustingsulfates such as gypsum-bearing coatings formed in urbanareas (Nord et al., 1994; Smith and Prikryl, 2007), is theirhigher resistance to weathering and thus their ability tomore effectively preserve isotopic and chemical signaturesof embedded materials.

S

H2S

SO2

(a)

(c)RCI

ACI

Fig. 1. (a) The black rock-coating of the Sudbury area, Ontario, Canadatmosphere-coating interface (ACI) and rock-coating-interface (RCI): (Iweathered exposed rock surfaces and formed a silica-gel type alteration labearing particulates and aerosols; (II) dissolution of the latter componentthe formation of a metal–sulfate-rich layer on the ACI; (III) corrosionmetal–sulfate-rich layer on the RCI; (c) occurrence of metal–sulfate-rich lasmelter (Mantha et al., 2012).

In this paper, we will compare the chemical, mineralog-ical and isotopic composition of silica coatings formed inthe vicinity of base-metal smelters with historical data onmetal emissions, metal speciation in smelter plumes andthe isotopic composition of the smelted ore, sulfate aerosolsand smelter-derived SO2 gases. We will show that the chem-ical and isotopic signatures of the coatings provide recordsof the speciation of trace metals from the smelter plume aswell as changes in the proportion of secondary and primaryaerosols, and their sulfur isotope signature with distancefrom the smelters.

1.1. Environmental impact of smelters

Former and current smelting activities contribute tolocal and regional anomalies of atmospheric pollution inthe form of metals, metalloids and sulfur dioxide. Metal(loid)-bearing particles emitted by smelter stacks influencethe quality of soil and water systems and have an impacton human health through the inhalation of nanometer-sizeairborne particulate matter (Kalkstein and Greene, 1997).Sulfur dioxide emissions from smelters add to global sul-fur concentrations in the atmosphere and contribute toacidic precipitation, climate forcing and global pollution(Brock et al., 1995; Hamill et al., 1997; Bao and Reheis,2003).

Smelters are considered point sources of metal(loid) pol-lution where the concentration of emitted elements com-monly decreases in surficial environments with distancefrom the source (Davies, 1983). The distribution pattern ofelements emitted from point sources has been mapped insoils (e.g. Freedman and Hutchinson, 1980; Hogan andWotton, 1984; Wren, 2012), peat (e.g. Zoltai, 1988), lake

Sulfate-rich

Atmosphere

Rocki, K, Al, Ca, Na, Fe

Cu, Ni, Fe, Pb, As particulate

O4 Pb, Cu, As sulfate aerosols

mosp

Sulfate-rich

Amorphous silica

(b)

a; (b) model for the formation of metal–sulfate-rich layers on the) precipitations and fumigations of sulfuric acid and sulfur dioxideyer, which promoted the incorporation of atmospheric metal(loid)-s and the leaching of elements from the underlying rock resulted inof the underlying rock by metal–sulfate-rich solutions produces ayers on the ACI and RCI in a coating sampled around the Coniston

N.M. Mantha et al. / Geochimica et Cosmochimica Acta 90 (2012) 221–241 223

sediments (e.g. Jackson, 1978), snow (Telmer et al., 2004)and in humus and till sediments (e.g. Henderson et al., 1998).

1.2. Silica and metal–sulfate-rich coatings of the Sudbury

area

The release of 11.2 million tons of SO2 from 1890 to 1930by roast yards (Laroche et al., 1979) and ca. 60 million tons ofSO2 between 1935 and 1975 by smelters (estimated on the ba-sis of ca. 1.5 million tons per year; Wren, 2012) resulted inacidic fumigations and rain, which enhanced the chemicalweathering rate of exposed rocks in the Sudbury area (Man-tha et al., 2012, Fig. 1b). Non-stoichiometric dissolution ofsilicate minerals under these acidic conditions resulted inthe formation of a silica-gel type coating (Schindler et al.,2009; Durocher and Schindler, 2011), which promoted theuptake and preservation of detrital grains from adjacentrocks and soils (feldspar, quartz, hematite, chlorite, montmo-rillonite) and smelter-derived nano- to micro-size particulatematter (metal–silicates, metal–oxides, C-spheres, Manthaet al., 2012). Dissolution of the underlying rock and metal-bearing particles by sulfuric acid resulted in the in situ forma-tion of metal–sulfate-rich layers on the atmosphere-coating(ACI) and rock-coating interface (RCI) (Fig. 1b and c).These metal–sulfate-rich layers are composed of nanometeraggregates of Fe-sulfate-hydroxide and Fe–Cu-sulfate minerals(Mantha et al., 2012). Scanning electron microscopy (SEM)studies on selected samples showed that metal–sulfate-richlayers on the ACI are common, whereas layers on the RCI onlyoccur on mafic rocks and coatings from roast yards.

TEM, SEM and XRD studies showed that the metal–sulfate-rich layers contain Fe-sulfate-hydroxide, Fe3+

4SO4

(OH)10, mereiterite, K2Fe2+(SO4)2(H2O)4, goldichite,KFe3+(SO4)2(H2O)4, butlerite, Fe3+(SO4)(OH)(H2O)2, guil-dite, CuFe3+(SO4)2(OH)(H2O)4 and antlerite Cu3(SO4)(OH)6, whereas smelter-derived particles are mainly com-posed of magnetite, Fe3O4, fayalite, Fe2SiO4 and unidenti-fied Cu-oxides (Mantha et al., 2012).

Although the release of SO2 and metal-bearing particleshas been drastically reduced in the last 40 years, blackmicrometer-thick metal–sulfate-rich coatings still occur onexposed rocks in the Greater Sudbury area (Fig. 1a). X-rayphotoelectron spectroscopy (XPS) indicated the occurrenceof a metal-depleted silica-rich layer on the surface of the coat-ings and suggested the current leaching of metal(loid)s bymeteoric water on a nanometer scale (Mantha et al., 2012).

1.3. Coatings on building stones formed through atmospheric

pollution

Black coatings are also common features on building-stone surfaces in polluted, usually urban, environments(Nord et al., 1994; Smith and Prikryl, 2007). Similar to theblack coatings found in the Sudbury area, the coatings formthrough interaction between minerals of the underlyingbuilding stone with S-bearing atmospheric components(SO2, H2SO4, HSO3

�). Here, dissolution of Ca-bearing min-erals (mainly calcite and Ca-silicates) by sulfuric acid resultsin the formation of Ca–H2SO4-rich water films, which uponevaporation form gypsum-rich precipitates. Incorporated

particulate matter commonly originates from anthropo-genic sources and contains metal salts, nitrogen and sul-fur-oxide compounds, radionuclides and carbon-bearingparticles. Distribution and thickness of the gypsum coatingsdepends on the geographic location, air pollution level, rain-fall infiltration and evaporation and the Ca-content of theunderlying rock (e.g. Buzek and Sramek, 1985; Pye andSchiavon, 1989; Nord et al., 1994; Prikryl et al., 2004; McAl-ister et al., 2006; Smith and Prikryl, 2007 and referencestherein).

1.4. Environmental relevance of coatings in riverine, soils and

tailings settings

The surface properties of minerals and amorphousphases change through their interaction with surface-,ground-, and pore waters. The properties and processesthat occur at the mineral–water interface are critical forunderstanding the uptake and release of toxic elements incontaminated sediments, soils, rocks, aquifers and acid-mine drainage (AMD) systems. Abiotic and biotic leachingor dissolution–reprecipitation processes drive non-stoichi-ometric weathering of primary minerals forming secondaryminerals (e.g. Fein et al., 1999). These processes promotethe formation of clay-sized layers of silicate phases, as wellas various Fe, Si, Al and Mn oxides and hydroxides, whichform discontinuous thin layers that coat mineral surfaces(e.g. Chorover et al., 2007).

Hydrous oxides of Mn and Fe have long been recog-nized as important scavengers of metals in stream sediment,soil and in the ocean (e.g. Carpenter et al., 1975; Lion et al.,1982; Robinson, 1981, 1982; Deng et al., 2000). Studieshave shown that films, cements and coatings of these oxidescontain anomalously high concentrations of metals nearoxidizing sulfide deposits (Filipek et al., 1981; Whitney,1981). The “scavenging” of metals by Mn–Fe oxides hasbeen widely investigated and is believed to include bothco-precipitation and adsorption processes (Chao and Theo-bald, 1976). Robinson (1981) showed that adsorption maybe more important than co-precipitation in producinganomalous levels of metals in these coatings. These coatingsrepresent a complex multiphase system consisting of vari-ous Mn and Fe oxides and hydroxides (with different de-grees of crystallinity) plus a mixture of organiccomponents and silicates, such as clay minerals (Potterand Rossman, 1979; Filipek et al., 1981).

Iron-hydroxide and Fe-sulfate coatings are common inoxidized zones of sulfide mineral assemblages (e.g. McGre-gor et al., 1998; Durocher and Schindler, 2011) and maycontrol the fate of metals and metalloids within many minetailings facilities. The high specific surface area of the Fe-hydroxides allows them to act as powerful sorbents for dis-solved species, particularly heavy metal contaminants,(PO4)3� and (AsO4)3� species (Cornell and Schwertmann,1996). Iron-sulfate minerals of the jarosite group are ableto accommodate significant proportions of cations of differ-ent valence and size within their structures (Dutrizac andJambor, 2000).

Micrometer-thick hydrous silica-gel coatings have beenfound on exposed bedrock at tailings ponds located within


the Copper Cliff mine tailings disposal area in Sudbury,Ontario, Canada (Schindler et al., 2009; Durocher andSchindler, 2011). These silica-rich coatings were found toform a more or less continuous layer below Fe-hydroxidecoatings with variable silica concentrations. Schindleret al. (2009) and Durocher and Schindler (2011) noted high-er or similar trace–metal concentrations in the silica-rich toFe-hydroxide coatings. Similar to the observations on theblack rock-coatings of the Sudbury area (Mantha et al.,2012), metals in the silica-rich coatings occurred eitherfinely distributed or were either found as discrete particlesof oxy-hydroxide and K–Fe-sulfate particles (most likelyjarosite).

1.5. Objectives

The black coatings of the Sudbury area formed during aperiod of high SO2 emissions between 1888 and 1972, whichcorrespond to the start of mining activities in the Sudburyarea and the construction of a 381 m Super Stack respectively(a recent review on the mining history of Sudbury can befound in Mantha et al. (2012)). The latter stack is one ofthe tallest and largest point sources in North America, whichinitially emitted ca. 20% of all the SO2 emitted annually inCanada (Nriagu and Coker, 1978). The corresponding smel-ter plume has been a desirable subject for chemical and isoto-pic studies by virtue of its high SO2 concentration and itscoherent form, at times extending to hundreds of kilometers(Lusis and Wiebe, 1976; Forrest and Newman, 1977b).

Chemical and isotopic studies were not carried out onsmelter plumes emitted from the smaller stacks during theformation of the coatings. However ore smelted at the Cop-per Cliff centers did not change significantly over time(Wren, 2012) and therefore, emitted plumes may have hadsimilar ratios between major-, minor and trace-elementsas well as ratios between sulfur isotopes. In this way, chem-ical and isotopic composition of the black coatings arecompared with those of the Copper Cliff Super Stack plumein order to derive a mixing model for the various S-bearingatmospheric species around the Copper Cliff smelter area.The objectives of this paper are in detail:

1. to describe the occurrence and speciation of metal(loid)s)within the coating;

2. to illustrate their distribution and concentration in coat-ings over the Greater Sudbury area;

3. to depict the variation of the d34S values in the coatingsover the Greater Sudbury area;

4. to examine if these variations are related either to frac-tionation of sulfur isotopes during oxidation of SO2 orto mixing of primary metal–sulfate aerosols with second-ary sulfate species formed during oxidation of SO2.

2. EXPERIMENTAL

2.1. Sampling and sample preparation

Coatings were sampled from 31 different locations with-in the Greater Sudbury area (Fig. 2). A sampling procedure

was designed to procure samples in proximity to the threesmelting centers (Copper Cliff, Falconbridge and Coniston)and the surrounding areas. Coatings are not prolific in thenorth-eastern area of Greater Sudbury because outcroppingbedrock is rare due to dominant glacio-fluvial sediments.As a result, coated rock samples in the area of Falconbridgewere collected mainly west of the Falconbridge smelter(Fig. 2). The collected samples were washed with distilledwater and cross-sectioned samples with thick coatings(>10 lm) were polished in epoxy discs using silicon carbide.

2.2. Scanning electron microscopy (SEM)

Scanning electron microscopy (SEM) was conductedwith a JEOL 6405 at 20 kV, equipped with both backscat-tered (BSE) and secondary (SE) electron detectors and anEnergy Dispersive X-ray Spectrometer (EDS). The averagecomposition of the coatings was determined via EDS anal-yses of selected areas in the distribution maps for Si and Fe.

2.3. Laser ablation inductively coupled mass-spectrometry

(LA ICP-MS)

One coating sample from each location (except for theroast yards) was analyzed in situ with laser ablation induc-tively coupled mass-spectrometry (LA ICP-MS) using aThermo-Fisher X II Series inductively coupled plasmaquadrupole mass spectrometer with an ultra-violet laserbeam (New Wave Nd:YAG 213 nm solid state laser). Abla-tion was carried out in a He atmosphere to which Ar wasadded for efficient transport of the ablated aerosol. Thecomposition of the coating was probed via depth profiling,where material from each sample was ablated from the topof the coating down to the underlying rock. Five depth pro-files were completed for each sample using beam diametersof 55, 40 and most commonly 30 lm. All analyses were car-ried out at a repetition rate of 5 Hz and the energy densitymaintained at 11 J cm�2. Each spot was measured for 30–45 s following 30 s of background acquisition. The syn-thetic glass standard NIST612 which contains trace-elementabundances of about 30–100 mg kg�1 was used as the exter-nal standard. At the beginning of each analytical run andafter each sample, the standard was ablated under the sameconditions. All data were blank corrected and instrumentdrift was corrected using a linear interpolation betweenstandard blocks (Ulrich et al., 2009).

2.3.1. Depth profiling versus line scans on coating surfaces

and cross-sections

Depth profiling of all 30 samples was preferred to sur-face line-scans and surface rastering (e.g. Wayne et al.,2006), because the latter scans remove material from a cer-tain depth (i.e. 30 lm) that can exceed the thickness of thecoating (on average from <10 lm to >100 lm, Manthaet al., 2012). In such cases, material from the underlyingrock would have ablated and their contributions wouldhave resulted in erroneous count rates and thus in incorrectelement concentrations for the coatings.

The depth profiling was also preferred to line scan anal-yses of cross-sectional samples because the former method

Fig. 2. Map of the Sudbury region with the location of the 30 sampling sites; the smelter areas of Copper Cliff, Coniston and Falconbridge areindicated with a sketched chimney; an index in the upper left corner shows the location of Sudbury within Ontario, Canada; an arrow in thelower left corner indicates the prevailing SW!NE wind direction.


allows the use of larger laser spot-sizes without any loss inresolution (with respect to contributions of coating androck). The use of large laser spot-sizes (i.e. 50 lm) has theadvantage of ablating a larger volume of material, yieldingbetter count statistics for elements of low concentration(e.g. Cd, Se and Co).

2.3.2. Selection of integration areas in the LA-ICP-MS depth

scans

Fig. 3a shows stacked depth-scan patterns for Fe and Siand selected elements through the black coating of sampleLA-1 (Fig. 2). These scan patterns depict a prominent peakfor all elements once material is ablated from the coating’ssurface. The peaks extend for ca. five seconds, which corre-sponds to a depth of ca. 25 lm. At greater depths, only thepattern for Si shows significant CPS, indicating the ablationof material from the underlying siliceous rock. Contrary toline scan traverses in cross sectional samples (Mantha et al.,2012), metal–sulfate- and silica-rich layers could not be re-solved in the patterns of the depth profiles (Fig. 3a). Hence,the depth scan-pattern was quantified using two integrationareas: one for the bulk coating and one for the underlyingweathered rock. The integration area for the bulk coatingwas selected in the depth profile through inspection of theCPS of Pb, As, Cu and S (representing the metal–sulfate-rich part of the coating) as well as the CPS of Mg, Na,K, Ca, Al and Si (representing the underlying rock andmost likely portions of the silica-rich layer of the coating).The integration area of the coatings was taken from the

inflection point at the rising limb to the inflection point atthe falling limb of the Pb, As, Cu and S-peaks (Fig. 3a).Hence, the selected integration boundary between a coatingwith a metal–sulfate-rich layer on the ACI and the underly-ing rock occurred within the silica-rich layer, whereas theintegration boundary between a coating with a metal–sul-fate-rich layer on the RCI and the underlying rock occurredmost likely along the RCI. Consequently, the quantifiedLA-ICP-MS data of the coatings should represent the me-tal–sulfate-rich parts of the coatings rather than an averageof the metal–sulfate and silica-rich layers. This can be ver-ified when comparing concentrations obtained from LA-ICP-MS depth scans and electron microprobe (EMP).For example, EMP spot analyses on the ACI of sampleF1 (sampled around the Falconbridge smelter, Fig. 2) showthat the average concentrations of Cu, As and Pb (majormetal(loid)s) in the coatings are 0.5, 0.6 and 3.1 wt.% andthat their concentrations in the metal–sulfate-rich layervary in the range of 0.7–1.0, 0.8–1.4 and 3.9–8.2 wt.%,respectively (Mantha et al., 2012). The latter concentrationsare more consistent with the corresponding LA-ICP-MSdata of sample F1 (Cu = 0.9–1.1, As = 1.3–2.0; Pb = 5.8–8.6 wt.%, Supplementary data A), in agreement with theargumentation above.

2.3.3. Quantification of the LA-ICP-MS data

An internal standard was required to quantify the LA-ICP-MS data because the ablation yield between the stan-dard and the sample was not constant. SEM, EMP and

y = 2.62ln(x) + 27.8R² = 0.9420

22

24

26

28

30

0.0 0.2 0.4 0.6 0.8 1.0

Si w

t% (

SEM

)

Si : Fe raw counts (LA ICP-MS)

(b)

Si

Time (sec)35 40 45 50 55 60

S

Fe

As

K

Cu

Ni

Pb

Se

Coating Rock[C

PS]

(a)

No ablation

Fig. 3. (a) Depth-scan patterns for Fe and Si and selected elementsthrough the coating of sample LA-1; for clarity reasons, the scansare stacked on top of each other and integration areas for thecoatings and underlying weathered rock are indicated with shadedvertical bars; (b) plot with the average ratios between the rawcounts of Si and Fe (measured on six coatings with LA-ICP-MSdepth scans) versus the corresponding Si-concentrations (measuredwith an EDS-SEM system); the uncertainty of the EDS measure-ment is approximately represented by the size of the symbols andthe standard deviations of the average ratios are indicated witherror bars.


LA-ICP-MS indicated that Si is the most common elementin the coating (Mantha et al., 2012) and therefore, the con-centration of Si determined with an energy dispersive spec-trometer (EDS) was used as an internal standard for thequantification of the LA-ICP-MS data. However, average

concentrations of Si in the coatings could not be obtainedfor each sample because in most cases, coatings were toothin to be resolved adequately in the SEM images. Hence,a correlation between the concentrations of Si and the cor-responding LA-ICP-MS data had to be identified.

Ratios between raw counts measured with LA ICP-MScan be used to identify associations between elementsregardless of differences in ablation (Ulrich et al., 2009;Durocher and Schindler, 2011). Concentrations of Si wereplotted versus the ratios of the Si:Fe raw counts from thecorresponding LA ICP-MS data (Fig. 3b). A regressionanalysis (R2) showed a strong exponential relationship(f(x) = 2.62ln(x) + 27.8, R2 = 0.94), which was used to cal-culate wt.% of Si for all other samples based on theirrespective Si:Fe raw counts. The resulting Si concentrationswere then incorporated as an internal standard in additionto the external NIST 612 standard values in the Plasma Labsoftware to quantify the LA ICP-MS data of the coatings.

2.3.4. Error estimation and detection limits

The counts for S and K from the NIST standard werenot above background and thus the measured concentra-tions of S and K were not considered in the data analysis.The ratios between the raw counts of Si and Fe for onesample (five depth profiles) varied on average in the rangeof ±0.05 (Fig. 3b) and resulted in an uncertainty ofca. ±1 wt.% for the corresponding Si-concentration. Onthe basis of this value, calculated concentrations for ele-ments with >100 mg kg�1 varied in the range of ±10%and concentrations of elements with <100 mg kg �1 in therange of ±7%.

Element-specific detection limits depend on the experi-mental setting of the laser scan (i.e. spot-size, repetitionrate). Detection limits for elements measured in this studyare provided in Supplementary data B.

2.4. Micro X-ray fluorescence using Synchrotron radiation

Polished samples embedded in epoxy were analyzed onthe Very Sensitive Elemental and Structural Probe Employ-ing Radiation (VESPER) Synchrotron beamline at theCanadian Light Source, Saskatchewan. Micro X-ray fluo-rescence (l-XRF) spectra were collected over raster areasto produce 2D element maps of two samples known to havemultiple coating layers. Beam penetration is on the order of100–250 lm, therefore only coating areas greater than100 lm thick were selected to minimize the interference offluorescence radiation emitted from mineral phases of theunderlying rock. The samples were scanned using a beamsize of 7 lm, a step size of 0.01 (dwell time 1.5 s) to 0.005(dwell time 1 s) and analyzed for the elements Mn, Fe,Ni, Cu, Cr, Zn, As and Pb employing a Si solid state detec-tor (Vortex-ME4).

2.5. Stable isotope measurements of sulfur

A total of 26 samples were prepared for sulfur isotopemeasurements by collecting scrapings of the coatings. Thescrapings were dissolved in HNO3 and S was precipitatedas BaSO4 under Ar atmosphere. The precipitates were


centrifuged, washed and subsequently treated with HNO3

to remove any carbonates. Sulfur isotopes were analyzedusing a MAT 252 Stable Isotope Ratio Mass Spectrometercoupled with a Carlo Erba NCS 2500 Elemental Analyzerat the Queen’s Facility for Isotope Research. The d34S val-ues for the coating material were calculated by normalizingthe 34S:32S ratios in the sample to the same ratio in theVienna Canyon Diablo Troilite (VCDT) standard:

d34S ¼d34Sd32S

� �sample

d34Sd32S

� �VCDT

� 1

0@

1A� 103 ð1Þ

and are expressed as d values in units of permil. The uncer-tainty of the d34S values is ±0.1 and is based on measuredvariations of d34S values for the VCDT standard.

The scrapings of the coatings likely contained materialfrom the underlying rock; however optical microscopyand SEM examination of the underlying rock did not indi-cate the presence of S-bearing minerals.

3. RESULTS

Black rock coatings are generally limited to exposedrocks within the Greater Sudbury area. Their distributionis consistent with the smelter impact zones in regards toatmospheric SO2 pollution, metals in surface soils, low soilpH, and vegetation damage which were controlled by dom-inant SW! NE winds. The distribution of metals from a

f(x) = 0.53(3)x

0

0.2

0.4

0.6

0.8

0 0.5 1 1.5

As [a

t%]

Pb [at%]

[mm]6.1 6.2 6.3 6.4 6

[mm

]

-4.3

-4.2

-4.1

-4.0

-3.9

[mm]6.1 6.2 6.3 6.4 6.5 6.6

[mm

]

-4.3

-4.2

-4.1

-4.0

-3.9

(d)

RY1

Silica-rich

Metal-sulfate-rich

Fig. 4. Micro X-ray fluorescence element distribution maps for (a) Pb andistribution map; (d) plot with the atomic percent (at.%) for Pb and As alayers in the coatings of F1 and RY-1 (Mantha et al., 2012) the size of the(e) plot with the concentration of Pb versus As (in mols per 100 g coatindepth scan analyses); the uncertainty for the LA-ICP-MS data (±10%) istheir equations are indicated in (d) and (e).

point source and subsequent environmental damage is con-sistently shown to be controlled in part by prevailing winds(Knight and Henderson, 2006).

3.1. Metal(loid)s in the coatings: distributions and

associations on a micrometer, millimeter- and regional scale

In the following sections, the distribution and associa-tion of metal(loid)s in the coatings are discussed on a mi-cro- and millimeter scale (i.e. within the coatings and theweathered surface of the rock) and on a regional scale(i.e. Greater Sudbury area). The results are presented inindividual sections that discuss elements closely associatedwith each other (Pb and As), main elements of locallysmelted ore (Cu and Ni) and elements that are toxic in smallconcentrations (>15 mg kg �1 for Se). The occurrence anddistribution of S will be addressed when discussing the vari-ations of d34S values within the coatings of the Greater Sud-bury area.

3.1.1. Lead and Arsenic

Fig. 4a and b show l-XRF distribution maps for Pb andAs in the cross section of sample RY-1. The large penetra-tion depth of the beam (100–250 lm) and its inclination an-gle of 45� created three-dimensional distribution maps, inwhich not every feature can be related to those observedin an optical or SEM image. However, the location of thecoating/weathered portion of the rock and the epoxy can

R2 =0.9654

.5 6.6

Pb [cps]

0.00 0.02 0.04

As [c

ps]

0.00

0.02

0.04

0.06 (c)

(e)

Pb [mol]0.00 0.04 0.08 0.12

As [m

ol]

0.00

0.02

0.04

0.06

f(x) = 0.45(2)x

d (b) As; (c) a plot between the CPS for Pb and As in the l-XRFs determined by EMP analyses for the metal–sulfate- and silica-richsymbol represents approximately the uncertainty of each data point;g) for the 30 examined coating samples (a total of 152 LA-ICP-MS

indicated for one data point; calculated linear regression lines and


be clearly distinguished and are labeled in both distributionmaps. Fig. 4c shows a linear plot of the counts per seconds(cps) for Pb versus As from the distribution maps (Fig. 4aand b). Distribution maps and linear plot indicate the com-mon occurrence of both elements and their strong associa-tion in the coatings and the weathered portion of theunderlying rock.

The concentrations of all major elements in the metal–sulfate- and silica-rich layers of the samples F1 and RY-1

were characterized with 21 EMP spot analyses (Manthaet al., 2012). A correlation between the concentrations(at.%) of Pb versus As records their common occurrencein the metal–sulfate-rich layer and the corresponding linearregression curve depicts an average atomic As:Pb ratio0.53(3) (Fig. 4d).

The chemical compositions from the 152 LA ICP-MSdepth-scan analyses (Fig. 4e) indicate an average As:Pb ra-tio of 0.45(2), which is quite similar to the observed ratio inthe metal–sulfate-rich layers of sample F1 and RY-1

(Fig. 4d). The apparent correlations indicate the close asso-ciation of Pb and As in all examined coatings, independentof the geographic location or the type of underlying rock.

Coating samples were grouped by geographical locationbased on their proximity to former or current smelters(Copper Cliff, Coniston and Falconbridge) and their rela-tive location in regards to the Greater City of Sudbury(Sudbury Center and Sudbury South, Fig. 2). The highestconcentrations of Pb and As occur around the formerand current smelters and decrease with distance from thesepoint sources (Fig. 5a and b). They further show that thehighest average concentrations of Pb are found in coatingsnear the Coniston smelter and that the highest average con-centrations of As occur in the Copper Cliff and Coniston re-gions (Table 1 and Fig. 5a and b).

(b)

(a)F

Con

Sudbury South

Sudbury Center

CopperCliff

Fig. 5. Maps of the Sudbury area with the average concentrations of (a)Sudbury area; the concentrations are expressed in ranges of weight %interpretation of the references to color in this figure legend, the reader

3.1.2. Copper and Nickel

Copper and Ni occur in different chemical environmentswithin the black coatings. SEM, EMP and transmissionelectron microscopy (TEM) examinations of the coatingsshow that Cu is one of the major elements in the metal–sul-fate-rich layer of sample F1 (Fig. 6a, Mantha et al., 2012).Nickel is typically found within spherical Fe-oxide smelterparticles (Fig. 6b), which are embedded in the silica-richmatrix of the black coatings (Mantha et al., 2012). EMPspot analyses of multiple particles and regions within thecoatings confirm these observations (Fig. 7a).

The different occurrences of Cu and Ni are also apparentin the l-XRF distribution maps (Fig. 7b and c) and the plotof the cps of Ni versus Cu (Fig. 7d), wherein there are atleast three different trends between Cu and Ni. Similarly,the concentrations of Ni and Cu do not correlate in all coat-ings of the Greater Sudbury area (Fig. 7e), which can beattributed to proportions of spherical particles and metal–sulfate-rich layers varying with the type of underlying rockand the geographic location of the coating (Mantha et al.,2012).

On the regional scale, higher concentrations of Ni andCu occur at Copper Cliff, Falconbridge and Coniston. Con-centrations of Ni are much higher at Copper Cliff than atConiston and Falconbridge whereas the concentrations ofCu are similar at Copper Cliff and Coniston (Table 1 andFig. 8a and b).

3.1.3. Selenium

The distribution of Se within the coatings was recordedwith depth-profiles and LA-ICP-MS line scans, the latter ofwhich traverse the coatings in cross-sectional samples(Mantha et al., 2012). Stacked LA-ICP-MS line scans ofSi, S, Se and Cu through the metal–sulfate- and silica-rich

5 km

As wt% <0.1 0.1-0.50.5-1 >1.0

Pb wt %< 5 5-7 7-9 >9

5 km

alconbridge

iston

Pb and (b) As in the coatings from 30 locations within the Greaterand indicated with solid spheres of different size and color. (Foris referred to the web version of this article.)

Fig. 6. (a) BSE image of the cross section of sample F1 (top) and the corresponding distribution map for Cu (bottom); the coating-rockinterface visible in the BSE image has been re-drawn in the distribution map of the element; (b) Back-scattering-electron (BSE) image of aspherical smelter-derived particle in a coating from Copper Cliff (top) and the corresponding chemical distribution map for Ni (bottom).

Ni [cps]0.0 0.1 0.2

Cu

[cps

]

0.0

0.1

0.2

0.3

0.4

[mm]6.1 6.2 6.3 6.4 6.5 6.6

[mm

]

-4.3

-4.2

-4.1

-4.0

-3.9

[mm]6.1 6.2 6.3 6.4 6.5 6.6

[mm

]

-4.3

-4.2

-4.1

-4.0

-3.9

Ni [wt%]0.0 0.4 0.8 1.2

Cu

[wt%

]

0

1

2

3(d) (e)

0

0.2

0.4

0.6

0.8

1

1.2

0 0.1 0.2

Cu

[wt%

]

Ni [wt%]

(a)

spheres

Metal-sulfate-richlayer

Fig. 7. (a) Plot with the concentrations of Ni (in weight %) versus Cu determined from 21 EMP spot analyses of the metal–sulfate-rich layers,spherical particles and silica-rich matrix; the uncertainty of a data point corresponds approximately to the size of its symbol (b) and (c) l-X-ray fluorescence element distribution maps for (b) Cu and (c) Ni; (d) plot of the CPS for Ni versus Cu in the l-XRF distribution map; (e) plotwith the weight percent of Ni versus Cu for the 30 examined coating samples, the uncertainty of the LA-ICP-MS data is indicated for one datapoint.


layer in the coatings of sample F1 indicate that the peak/background ratio for the line scan of Se is much smallerthan for the scans of Cu, S and Si. Similarities betweenthe line-scan patterns of Se, Cu and S are apparent and sug-

gest their common association in the coatings (Fig. 9b). Thehighest concentrations of Se occur in the coatings aroundCopper Cliff and decrease with distance from the CopperCliff smelter (Fig. 9c and Table 1).

Table 1Element concentrations (mg kg �1 and weight %) for coatings from five different regions; n denotes the number of samples per region (Fig. 2),each sample was analyzed with five to six LA-ICP-MS depth scans.

Region Statistic Se (mg kg �1) Ni (mg kg �1) Fe (wt.%) Cu (wt.%) As (wt.%) Pb (wt.%)

Coniston Min 67 610 2.9 0.3 0.2 0.7n = 6 Max 165 1459 27.3 1.5 1.8 10.3

Mean 110 1215 13.9 0.8 0.9 5.0Copper Cliff Min 93 1189 0.5 0.1 0.1 0.2n = 9 Max 1030 9400 21.5 2.2 1.7 8.0

Mean 482 3035 11.7 0.9 0.9 4.8Falconbridge Min 7 580 1.4 0.1 0.1 0.2n = 4 Max 331 2109 22.1 1.1 1.5 6.7

Mean 136 1149 10.6 0.5 0.6 2.5Sudbury Center Min 45 159 4.1 0.2 0.1 0.8n = 7 Max 471 2471 11.6 0.8 0.9 6.7

Mean 189 1199 8.6 0.4 0.4 2.6Sudbury South Min 10 632 3.5 0.1 0.1 0.4n = 4 Max 128 1690 13.6 0.4 0.3 1.4

Mean 52 1137 8.2 0.2 0.2 0.8

Cu wt% < 0.1 0.1-0.5 0.5-1.0 1.0-1.5 >1.5

Ni wt% < 0.08 0.08-0.20 0.20 - 0.30 > 0.30

5 km

(a)

5 km

(b)

Falconbridge

Coniston

Sudbury South

Sudbury Center

Copper Cliff

Fig. 8. Maps of the Sudbury area with the average concentrations of (a) Ni and (b) Cu in the coatings from 30 locations within the GreaterSudbury area; the concentrations are expressed in ranges of weight % and indicated with solid spheres of different size and color. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


Cu [wt%]

0 1 2 3

Se [w

t%]

0.00

0.04

0.08

0.12

0.16

Time (sec)60 65 70 75 80 85 90 95 100

S

Cu

Se

Si

(c)

(a) (b)

Se wt%

< 0.01

0.01-0.04

0.04-0.080

>0.080

MS Si altered rock

Fig. 9. (a) Stacked plots of LA-ICP-MS line scans for Cu, Se, S and Si; areas associated with the metal–sulfate-rich layer (MS), silica-richlayer (S) and weathered part of the rock are indicated with shaded vertical bars. (b) Plot of the concentrations (weight %) of Cu versus Se forthe 30 examined coating samples; the uncertainty for the LA-ICP-MS data is given for one data point. (c) Map of the Sudbury area with theaverage concentrations of Se in the coatings; the concentrations are expressed in ranges of weight % and indicated with solid spheres ofdifferent size and color. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of thisarticle.)


3.1.4. Sulfur and d34S

Mantha et al. (2012) showed that the concentrations of Sstrongly correlate with those of Fe measured by EMP(WDS) in the coatings of F1 and RY-1. Similarly, the con-centration of all major metal(loid)s in the metal–sulfate-richlayers of RY-1 and F1 (Pb, As and Cu) correlate with thoseof S. Correlations of high confidence (R2 > 0.70) are rarebetween the concentrations of S and metal(loid)s for the30 coatings examined using LA ICP-MS, most likely a re-sult of the bulk analysis of the coatings and absence of me-tal–sulfate-rich layers in some coatings (Mantha et al.,2012). The correlation with the highest confidence(R2 = 0.71) occurs between the cps of S and Pb(Fig. 10a), suggesting that S is more closely associated withPb than with other metal(loid)s.

The sulfur isotopic composition (d34S) of the coatingsvaries between +2.6& and +5.7&. The higher values are

associated with samples in closer proximity to the smeltingcenters, notably Copper Cliff and Coniston (Fig. 10b). Thed34S values in Copper Cliff showed the largest spread(+3.3& to +5.7&) and included the single highest d34S va-lue recorded (+5.7&). The samples from Coniston and Fal-conbridge exhibit a smaller range of d34S values with+4.1& to +5.3& and +3.5& to +4.4&, respectively. Thesamples from Falconbridge display the smallest d34S valuesof all samples collected around former and recent smelters.Sudbury Center and Sudbury South samples exhibit thelowest d34S values, +3.0& to +4.1& and +2.6& to+3.7&, respectively.

A correlation between the d34S values of the coatingsand their distance to the smelter centers can be estab-lished if one considers the prevailing SW! NE winddirection and the emission data from the various smelt-ers. The smelters at the Copper Cliff area emitted more

S [cps]0 4000 8000 12000

Pb [c

ps]

0

4e+7

8e+7

1e+8 (a)(b)

Sulfates inCoa�ngs

Sulfate aerosols insmelter plume

Ni-Cu sulfides

(d)SO2 instackandplume

Sulfates in rain

0 1 2 3 4 5 6 7 8

Distance from smelter [km]0 2 4 6 8 10 12 14 16 18 20

2.02.53.03.54.04.55.05.56.0

(c)

Fig. 10. (a) Plot of the cps of S versus Pb for all 152 LA-ICP-MS depth scans; (b) map of the Greater Sudbury area with the d34S values forcoatings from 26 locations; (c) correlation between the d34S values of the coatings and their distance to the smelters at Copper Cliff (circles),Falconbridge (triangles) and Coniston (squares); the size of the symbol represents approximately the uncertainty of a data point; the averaged34S value for the sulfate aerosols is indicated with a star (see text for details); (d) sketch indicating the range in d34S values for the Sudburyore, SO2 gas in the smelter, (SO4)2�-bearing smelter plume and the coatings.


sulfur dioxide and metals than the smelters at Conistonand Falconbridge. For example, the former smelters emit-ted 18-times more primary sulfates during the period1973–1981 than the smelter at Falconbridge and eight-times more particular matter during the period 1930–1960 than the smelter at Coniston (Ozvacic, 1982; Wren,2012). Emission data and prevailing wind direction sug-gest therefore that all coatings sampled west of Conistonand Falconbridge were predominantly affected by theemission of the Copper Cliff smelters and that only coat-ings sampled around Coniston and Falconbridge weresignificantly affected by the emissions of the stacks inthese areas. In this way, distances between sample loca-tions and the Copper Cliff smelter center (Table 2) werecalculated for all areas west of Coniston and Falcon-bridge (labeled with stars, squares and crosses for the dif-ferent areas indicated in Fig. 2 and listed in Table 1),whereas distances to the Coniston and Falconbridgestacks (Table 2) were only calculated for sampling loca-tions in their vicinity (triangles and circles). Fig. 10c indi-cates a good correlation between the d34S values of thecoatings and their distances to the various smelters(R2 = 0.52 and 0.62 for a linear and exponential regres-sion, respectively).

4. DISCUSSION

The regional distribution of the black coatings indicatesthat its formation is primarily related to the occurrence of ex-posed bedrock and secondarily to the distance of the smelt-ers. Anthropogenic exposures of bedrock surfaces as wellas natural geomorphic processes provide a first-order controlfor the occurrence of the rock coatings according to Dorn’shierarchal model of coating development (Dorn, 1998). Inthe Sudbury area, these first-order processes include exten-sive logging prior to and during early smelting activities(ore roasting in roast yards), and the occurrence of (sulfuric)acid rain and metal emissions, which both contributed to abarren landscape lacking soil and vegetative cover. Thenear-absence of black rock coatings in the northern andnortheastern portion of Sudbury is related to a lack of ex-posed bedrock surfaces due to the presence of glacio-fluvialsediments, which were deposited during glacial events.

Transport and availability of raw ingredients for the for-mation of coatings, i.e. a positive mass balance at the rock-atmosphere interface is ranked as a fourth-order process inDorn’s hierarchal model (Dorn, 1998). The absence of rockcoatings in areas west of Sudbury is a product of such afourth-order process, where the prevailing northeasterly

Table 2Sampling sitesa around the three smelter areas, their distance to the smelter stacks and the average d34S, As:Ni, Se:Ni and Pb:Ni ratios in theircoatings.b

Location km d34S As:Ni Se:Ni Pb:Ni

Data for all locations around the Copper Cliff smelter area

17-1 8.6 2.6 1.8(2) 0.13(2) 9(1)17-2 12.8 3.2 2.2(3) 0.04(1) 11(3)69N-1 9.6 3.0 1.5(2) 0.11(2) 13(2)69S-1 13.0 3.7 0.38(5) 0.012(2) 2.6(4)69S-2 17.5 2.8 1.0(1) 0.015(2) 8(1)A1 5.2 3.3 4.8(9) 0.18(3) 22(4)A3 6.4 5.7 7.0(7) 0.21(3) 30(3)AD1 8.0 3.5 1.2(2) 0.04(1) 7(1)B1 8.3 3.3 2.1(3) 0.09(1) 16(2)DE2 1.7 5.6 6.5(9) 0.23(3) 32(5)FR1 5.4 3.7 2.1(3) 0.14(2) 12(1)K1 2.9 5.2 1.7(2) 0.26(3) 19(3)LA1 12.2 3.4 2.9(4) 0.12(2) 24(3)M1 11.3 3.4 2.3(3) 0.09(1) 14(2)CC1 1.6 5.5 12(1) 0.33(2) 61(8)S1 7.4 4.1 2.4(2) 0.019(3) 18.5(9)L1 8.4 nd 2.5(2) 0.22(2) 19(3)CC3 1.5 4.6 Nd Nd Nd

Data for locations around the Falconbridge smelter

F1 0.8 4.0 7(1) 0.09(1) 32(4)SK1 4.5 3.4 0.17(2) 0.011(2) 2.7(4)SK2 7.4 4.4 1.2(4) 0.021(2) 6.2(7)G1 4.0 3.8 9(1) 0.38(5) 29(4)

Data for locations around the Conistion smelter

C1 1.9 5.3 3.0(4) 0.05(1) 17(1)W1 4.6 4.6 8(1) 0 67(9)W2 4.3 4.1 6.5(9) 0.094(2) 19(3)CH 1.8 nd 5.5(5) 0.12(2) 38(4)W3 4.1 nd 1.9(3) 0.06(1) 6.0(5)CE 1.8 nd 29(3) 0.00 169(13)CH 1.8 nd 5.5(4) 0.12(2) 38(4)

a The samples RY-1, TB and CC4 were not considered due to their locations within former Roast Yards and tailings disposal areas.b Uncertainties are given in parentheses.


winds resulted in lower deposition rates of sulfuric acid,particular matter and metal–sulfate aerosols west of thecity.

4.1. Emission data and metal(loid) concentrations in the

coatings

The distribution of Pb, As, Cu, Ni, Se and S in the coat-ings on the micrometer, millimeter and regional scaleshowed that

1. Pb and As occur in close association in all coatings overthe Greater Sudbury area.

2. Ni and Cu occur often in different areas on the micro-and millimeter scale and their concentrations in the coat-ings do not correlate on the regional scale.

3. Se and S are in closer association with Cu and Pb,respectively, than with other metal(loid)s in thecoatings.

We will explain first the occurrences and associations ofthese metal(loid)s using the emission data by the smelters

and the minerals present in the coatings. The regional dis-tribution of the metal(loid)s in the coatings can be, how-ever, only addressed when one considers their speciationin the plume and the deposition and mixing of the differentS-bearing atmospheric species. The latter information canbe gained by (a) comparing the d34S values of the coatingswith those of the smelter plume and precipitations and (b)by evaluating the possibility of S fractionation during oxi-dation of atmospheric SO2.

4.1.1. Lead and Arsenic: emission data and potential hosts

On the basis of the historical data from Hutchinson andWhitby (1974), Ozvacic (1982) and Wren (2012), an esti-mated 200 tonnes year�1 of Pb and 120 tonnes year�1 ofAs were emitted by the Copper Cliff smelter complex be-tween 1925 and 1975. In comparison, the smelters at Fal-conbridge and Coniston emitted together approximatelyone-fourth of the Pb and As emissions emitted by the Cop-per Cliff smelters during this time period (Ozvacic, 1982;Wren, 2012). On the basis of these emission data, the As:Pbatomic ratio within the smelter plume was approximately1.5:1. This ratio differs from the average As:Pb atomic ratio


of 1:2 in the coatings of the Greater Sudbury area (Fig. 4dand e) and indicates that the incorporation of As and Pbwas mainly controlled by the stoichiometry of the phasesin the metal–sulfate-rich layers and not by the proportionsof the elements in the smelter plume.

The common occurrence of K, Fe, S, Pb and As in themetal-rich layers and the correlation between Pb and S(Fig. 10a) suggests that Pb and As must be associated withthe K–Fe sulfates mereiterite and goldichite. The potentialsubstitution of K+ by Pb2+ in the interstitial complexes ofmereiterite, and goldichite, is, however, only possible bymeans of a coupled substitution to maintain charge balance(Schindler et al., 2006).

The coupled substitution :

Pb2þ þ ðAs5þO4Þ�3 $ Kþ þ ðSO4Þ�2 ðR1Þ

is well documented for jarosite (Scott, 1987; Rattray et al.,1996; Rocca et al., 1999; Dutrizac and Jambor, 2000) andcould occur in both goldichite, K[Fe3+(SO4)2(H2O)4] andmereiterite, K2[Fe2+(SO4)2(H2O)4]. The As:Pb ratio of 1:2indicates that for every second Pb2+ that substitutes forK+ in the K–Fe-sulfates, one As5+ substitutes for oneS6+. This requires an additional coupled substitution in or-der to balance the incorporation of each additional Pb2+.There are two possibilities: (1) In the structure of theFe3+-sulfate goldichite, Fe2+ can substitute for Fe3+, main-taining the charge balance of the mineral structure:

KFe3þðSO4Þ2ðH2OÞ4 $ Pb2½Fe3þ;Fe2þðSO4Þ3ðAsO4Þ�ðH2OÞ4ðR2Þ

(2) the deprotonation of H2O groups, which would berequired for mereiterite, K2[Fe2+(SO4)2(H2O)4], a mineralwithout significant amounts of Fe3+. The deprotonationwould involve (H2O) groups of the Fe-octahedra, leavingbehind underbonded Fe2+-OH terminations along the[Fe(SO4)(H2O)4] clusters (Giester and Rieck, 1995). Thistermination would most likely bond directly to Pb2+ atoms,in order to minimize the number of underbonded O-atomswithin the structure of the mineral:

K2½Fe2þðSO4Þ2ðH2OÞ4� $ Pb2½Fe2þðSO4ÞðAsO4Þ� ðOHÞðH2OÞ3� ðR3Þ

4.1.2. Copper and Nickel: emission data and hosts

Historical records of metal emissions from the Sudburysmelting facilities indicate that over 1000 tonnes of Cu andNi were annually released between 1935 and 1975 (Hutch-inson and Whitby, 1974; Ozvacic, 1982; Wren, 2012). Theatomic ratios between these elements in the smelter plumevaried annually and from stack to stack but were on aver-age close to 1:1. Note that the smelters at Coniston andFalconbridge emitted only one fourth of the Cu andNi-emissions by the Copper Cliff smelters (Wren, 2012).

The coatings contain up to eight times more Cu than Ni(Table 1), which must be a result of their dominant occur-rences in the metal–sulfate-rich layer and spherical parti-cles, respectively. Concentrations of Ni in the metal–sulfate-rich layers are much lower relative to Cu (Fig. 7a).This observation may be explained by

1. The presence of Cu-sulfates and the absence of Ni-sul-fates in the metal–sulfate-rich layers; most likely a resultof the higher solubility of common acidic Ni-sulfates,Ni(SO4)�n(H2O) (n = 1–7) with respect to basic Cu-sul-fates such as guildite [Cu2+Fe3+(SO4)2(OH)2�4(H2O)]and antlerite [Cu3(SO4)(OH)6] (Jambor et al., 2000).

2. The preferred incorporation of Cu in K–Fe-sulfates withrespect to Ni (Dutrizac et al., 1996).

3. The higher affinity of Cu relative to Ni to sorb on silia-nol groups within the silica-matrix of the coatings (Dug-ger et al., 1964; Stumm, 1992).

4.1.3. Selenium: emission data and potential hosts

The Sudbury ore contains on average 40 mg kg�1 Se andSudbury smelters emitted ca. 50 tonnes year�1 Se in theearly 1980s (Nriagu and Wong, 1983).

High Se concentrations can be commonly found in Feand Cu-sulfides, where the selenide ion replaces the sulfideion (Hawley and Nichol, 1959). Similarly, the selenate ionreplaces the sulfate ion in Fe- and Cu-sulfate minerals.For example, Dutrizac et al. (1981) successfully synthesizedjarosite-group minerals in which (SeO4)2� substituted for(SO4) 2� in mol ratios controlled by the composition ofthe mother solution.

The correlation between the concentrations of Se and Cu(Fig. 9b) suggests the presence of Se-rich Cu-sulfates in thecoatings. A similar correlation does not exist for Se and Fe,which raises the question whether the correlation betweenCu and Se is a result of the preferential incorporation ofSe into Cu-sulfates, the occurrence of Se-rich Cu-sulfidesin the Sudbury ore or the lower ratio between oxides (i.e.

non-selenate bearing phases) and sulfates of Cu versus Fe.Firstly, there are no crystal-chemical indications that

(SeO4)2� would preferentially replace (SO4)2� in a Cu-sul-fate rather than a Fe-sulfate. Selenium-rich Cu-sulfidesare quite common in sulfide ore deposits in Canada. Forexample, Hawley and Nichol (1959) showed that Cu-sul-fides in the magmatic ore deposit at Rouyn-Noranda, Que-bec contain up to 1000 mg kg �1 Se. Although the authorsshowed that Se is equally distributed between Fe-sulfides(pyrohotite) and Cu–Fe-sulfides (chalcopyrite) in the Sud-bury ore, we cannot rule out that Se-rich Cu-sulfides wereand are smelted at the Copper Cliff facilities. Finally,SEM and TEM studies showed that Fe-oxides are indeedmore common in the coatings than Cu-oxides (Manthaet al., 2012), explaining the poorer correlation betweenthe concentrations of Fe and Se relative to Cu and Se.

4.2. Sulfur isotope composition of the ore, smelter plume,

precipitations, lakes and coatings in the surroundings of the

Greater Sudbury area

The isotopic composition of sulfur-bearing aerosols iscommonly used to trace their source and distribution in theatmosphere (e.g. Norman et al., 2006; Sinha et al., 2008).The isotopic signature of atmospheric sulfur may vary be-tween �5& and +25& depending on the various contribut-ing sources of sulfur (Krouse and Mayer, 2000). Sulfur ofvolcanic origin displays d34S values of approximately +5&


(Lein, 1991) whereas sulfur-bearing aerosols from salt sprayrange from +15& to +21& (Krouse and Mayer, 2000). Itshould be noted that there is no significant fractionation ofsulfur during the transition from fluid (ocean) to vapor(atmospheric aerosol) (Mayer, 1998). Anthropogenic sulfurmeasured from precipitation (rain and snowfall) typicallyvaries between �3& and +9& (Mayer, 1998).

Smelters in Sudbury have been documented to releaseboth SO2 gas and sulfate (SO4)2�-bearing particulate mat-ter (Forrest and Newman, 1977b; Wren, 2012). Stable iso-tope studies of the main smelter plume (the Super Stack)showed that SO2 in the plume displayed low positive d34Svalues of +0.9& to +1.2& (Nriagu and Coker, 1978),which is consistent with the isotopic composition of themajority of the smelted Ni–Cu ores (+0.2& to 3.5&) in thisregion (Fig. 10d, Thode et al., 1962; Schwarcz, 1973).

Sulfates in the smelter plume mainly occur in the form ofH2SO4 with minor amounts of metal–sulfate aerosols (seebelow). The proportion of these primary sulfates in theplume varies between 0.4% and 3% of the total SO2 emis-sions (Chan et al., 1982a,b, 1983) and their d34S values arein the range of +4.2& to 7.6& (Fig. 10d, Lusis and Wiebe,1976). Nriagu and Coker (1978) and Nriagu and Harvey(1978) examined in detail the sulfur-isotope compositionof wet precipitation and lake water around Sudbury (up to100 km from the Super Stack). They found seasonal varia-tions of the d34S values in wet precipitation with lower(+2.99& to 4.66&) and higher (+5.64& to 6.19&) valuesin the summer and winter, respectively. The d34S valuesfor lake water varied over a larger range with values between2.66& and 8.43&. In both studies, the authors noted thatthe average d34S values in precipitation and lakes were sim-ilar to the values for the sulfate aerosols in the smelter plume(Fig. 10d) and that the individual values did not show anysystematic variations with the distance to the smelter.

4.2.1. S- and O-fractionation during oxidation of SO2

The fractionation of S isotopes in the atmosphere can bemediated by a gas-phase or an aqueous-phase oxidationprocess. The gas-phase oxidation (also known as homoge-neous oxidation) occurs via OH radicals in the troposphereand stratosphere and produces sulfuric acid (H2SO4) whichcondenses, forming aqueous (SO4)2� and (SO4)2�-bearingaerosols (Saltzman et al., 1983; Tanaka et al., 1994):

SO2 þOHþMetal! HSO3 þMetal ðR4ÞHSO3 þO2 ! HO2 þ SO3 ðR5ÞSO3 þ 2H2O! H2SO4ðgÞ H2O ðR6Þ

The measured fractionation factor ahom (Harris et al.,2012) for the equations (R4)–(R6) is

a ¼ 1:0089� 4� 10� 5T ð�CÞ ð2Þ

The aqueous-phase oxidation (also known as heteroge-neous oxidation) occurs in the aqueous component or onparticle surfaces in the atmosphere. The major oxidantsare H2O2, O3 and O2, the latter being catalyzed by Fe3+

and other transition metal ions (Hermann et al., 2000). Dis-solution and oxidation of SO2 in the aqueous phase can beformulated as follows: (Eriksen, 1972a–c)

SO2ðgÞ $ SO2ðaqÞ ðR7ÞSO2ðaqÞ þH2O$ HSO3

� þHþ ðR8Þ

HSO3� þ ðH2O2=O3=particulatesÞ $ SO4

2�ðaqÞ þHþ

ðR9Þ

Measured fractionation factors are for the equilibrium reac-tion (R7) (Eriksen, 1972b),

a ¼ 1:0029 at 10 �C ð3Þ

for the equilibrium reactions (R7) and (R8) (Eriksen,1972a),

a ¼ 1:01033 at 25 �C ð4Þ

for the equilibrium reactions (R7)–(R9) with O3 and H2O2

as oxidants (Harris et al., 2012):

a ¼ 1:0167� 8:7� 10� 5T ð�CÞ ð5Þ

and for the equilibrium reactions (R7)–(R9) with Fe and O2

as oxidants (Harris et al., 2012):

a ¼ 0:9894ðat 19 �CÞ ð6Þ

The mechanism for SO2 oxidation can also be inferredfrom its O-isotopes. Holt et al. (1982), Holt and Kumar(1984) and Holt (1991) showed that d18O in primary sul-fates (sulfates formed before their release to the atmo-sphere) are isotopically heavier than secondary sulfates(formed through oxidation of SO2 in the atmosphere).For example, Norman et al. (2004a,b, 2006) used S- andO-isotope measurements to study the atmospheric sulfurbudget for two urban environments with different degreesof humidity (Vancouver and Calgary). Their studiesshowed that primary and secondary sulfates emitted fromdifferent sources can be distinguished on the basis of theirS and O isotope signature and that S fractionation didnot occur during oxidation of SO2 in the atmosphere.

It should be noted here that d18O measurements on theblack coatings would have been possible with, for example,Secondary Ion Mass Spectroscopy (SIMS), but would haveresulted in the measurement of d18O values from embeddedsulfates, smelter-particles, Fe-oxides and clays as well asfrom the surrounding silica matrix.

4.2.2. Rate of SO2 oxidation and proposed S fractionation in

the Super Stack plume

Lusis and Wiebe (1976) and Forrest and Newman(1977b) showed that the oxidation rate of SO2 in theSuper Stack plume is on average 1% per hour and sug-gested that heterogeneous oxidation is the predominantprocess in the conversion of SO2 into H2SO4. Forrestand Newman (1977b) argued further that the relativelylow oxidation rate in the Sudbury plume (in comparisonto plumes emitted by coal-fired power plants) was the re-sult of a low particulate loading (i.e. low amounts of me-tal–surface catalysts) and a low pH value, which sloweddown the conversion of SO2(aq) and HSO3

� in the equi-librium reactions (R7)–(R9).

On the basis of these small oxidation rates, Nriagu andCoker (1978) explained the differences in the d34S values be-tween sulfates and SO2 in the smelter plume with the frac-tionation of S during the conversions of SO2(gas) into


SO2(aq) and SO2(aq) into HSO3� (reactions (R7) and (R8)).

They explained furthermore the seasonal variations ofd34S in precipitation around Sudbury with higher washoutrates of (SO2)gas in the summer (lower positive d34S values)than during the winter months (higher positive d34S values).

The interpretations by Nriagu and Coker (1978) did notaccount for the occurrence of primary sulfates in the smel-ter plume (0.4–3%), which occur in at least equal propor-tions to secondary sulfates (formed through the oxidationof SO2) within the first or second hour the smelter plumeis emitted (considering an oxidation rate of 1%). The stud-ies on the smelter plume by Lusis and Wiebe (1976), Forrestand Newman (1977b) and Nriagu and Coker (1978) showedthat the plume traveled on average 40 km h�1. The samplelocations of the black coatings were all within 40 km of thesmelter centers, suggesting that their isotopic signature wassignificantly influenced by primary sulfates and their mixingwith secondary sulfates.

4.2.3. Decrease of d34S with the distance from the smelter, a

result of S fractionation?

The small variations of d34S for SO2 in the Super Stackplume (+0.9& to +1.2&) suggest that SO2 fumigations andplumes prior to the Super Stack had similar isotopic com-positions. If this was indeed the case, the isotopic composi-tion of secondary sulfates (d34SH2SO4

) formed duringRayleigh distillation can be calculated using the followingequation:

d34SH2SO4¼ aðd34S0

SO2 þ 103Þf a�1 � 103 ð7Þ

where d34S0SO2

is the initial d34S value for SO2 (average valueof SO2 in the plume = 1.11& for f = 1) and f the fraction ofSO2(gas) remaining (Faure, 1998). For the equilibriumreactions (R7) and (R8), ðR7Þ–ðR9ÞH2O2=O3

and (R7)–(R9)Fe-surface at T = 19 �C, the fractionation factors a wouldbe 1.01033, 1.01505 and 0.9894, respectively.

If the remaining fraction of (SO2) gas decreases in stepsof 0.01 (corresponding to an oxidation rate of 1% h�1), sec-ondary sulfates would have either large positive or negativevalues depending whether the oxidation was catalyzed byO3/H2O2 or metal–surface catalysts. For example two hoursafter emission of the plume with f = 0.98, the d34SH2SO4

would be in the range of 11&, 16& and �9& for the equi-librium reactions (R7) and (R8), ðR7Þ–ðR9ÞH2O2=O3

and

(R7)–(R9)Fe-surface, respectively. These calculated d34SH2SO4

values for secondary sulfates formed within the greater Sud-bury area (assuming plume velocities of 40 km h�1) are notwithin the range of d34S values observed in the black coat-ings. Even if one assumes higher oxidation rates of SO2(gas)

on the surface of exposed bedrocks (i.e. 3% h�1), the corre-

sponding d34SH2SO4would be still much larger (+7.8& to

+9.5&) or smaller (�6.5&) than any observed value inthe coatings (note that an oxidation rate of 3% h�1 corre-sponds to the largest oxidation rate observed in a plume inNorth America; Newman, 1981). Hence, mixing of similarproportions of primary sulfates (+4.2& to +7.6&) with sec-ondary sulfates containing fractionated sulfur (with valueshigher than +7.8& or less than �6.5&) would result in

d34SH2SO4values which are not consistent with those mea-

sured in the coatings (+2.6& to +5.7&). As a result, S-frac-

tionation during the equilibrium reactions (R7) and (R8)(R7)–(R9) ðR7Þ–ðR9ÞH2SO4

and (R7)–(R9)Fe-surface is disre-

garded as the cause of the decrease of d34S with distancefrom the smelters.

4.2.4. Decrease of d34S with distance from the smelter, a

result of mixing?

Studies by Chan et al. (1982c–e) and Chan and Lusis(1985) showed that sulfates and trace metals from theplume are efficiently scavenged during wet precipitationevents. For example, typical scavenging coefficients for sul-fates were found to be about 30–40% h�1 and for trace met-als such as Pb, Cu, Se and As in the 60–80% h�1 range (seebelow). The same studies also showed that the scavengingcoefficient for SO2(gas) was insignificantly small relativeto those of sulfates, indicating that mainly primary and sec-ondary sulfates scavenged via wet precipitation likely inter-acted with exposed rocks in the Sudbury area.

The amount of primary sulfates in the plume was mea-sured to be between 0.4% and 3%. Hence, an oxidation rateof 1% h�1 for SO2(gas) would have resulted in approximatelyequal amounts of primary and secondary sulfates within thefirst two hours after emission of the plume. The averaged34S value for the primary sulfates (5.4&) is similar to thed34S values observed in the coatings proximal to the smelt-ers (Fig. 10c). If the equilibrium oxidation reactions (R7)–(R9) did not result in the fractionation of S, as observedin the urban areas of Calgary and Vancouver (Normanet al., 2004a,b), the d34S value for the secondary sulfateswould have been similar to those for SO2(gas) (averagevalue = 1.11&) from which they formed. Hence, anincreasing ratio between secondary and primary sulfatesthrough scavenging of primary sulfates by wet precipitation(see below) and oxidation of SO2(gas) would have yielded adecrease in the d34S-value for atmospheric sulfates (andthus for the coatings) with increasing distance from thesmelters. This mixing model can be verified through inspec-tion of the regional distribution of trace metals in the coat-ings, which were found to be closely associated withprimary sulfate aerosols in the smelter plume.

4.2.5. Deposition rates of sulfate aerosols and larger

particular matter

Studies of airborne particulate matter in the CopperCliff Super Stack smelter plume found that Pb, As and Seoccurred in smaller particles with average diameters<2.5 lm, compared to Cu and Ni which were found mainlyin larger particles (>2.5 lm; Chan et al., 1982a,b). The firstobservation is in agreement with studies of other smelterplumes, which in addition, showed that Pb commonly oc-curred as metallic Pb or Pb-sulfates in plumes (Sobanskaet al., 1999; Choel et al., 2006). The close association ofPb with sulfate aerosols in the Sudbury plumes may beexemplified by the correlation between the raw counts forPb and S from the LA-ICP-MS study (Fig. 10a), whichindicates that almost all Pb in the coatings is associatedwith S-bearing minerals.

Studies on the wet and dry deposition rate of metalsfrom the Super Stack plume showed that deposition ratesfor the larger Fe–Ni particulate matter were similar during


Se :

Ni r

atio

0.0

0.1

0.2

0.3

0.4


As :

Ni r

atio

02468

101214


Pb :

Ni r

atio

010203040506070

(a)

(b)

(c)

Fig. 11. Correlations between the distance of the coatings to theCopper Cliff smelter facility and their concentration ratios for (a)Pb:Ni, (b) As:Ni and (c) Se:Ni; the uncertainty of a data point isgiven in Table 2; data from coatings in the Falconbridge andConiston area were not considered because different ores andsmelter and sintering techniques in comparison to the Copper Cliffsmelters resulted in different emission data, especially with respectto the ratios Pb:Ni, As:Ni and Se:Ni (Wren, 2012); data fromcoatings in former and present Roast yards and tailings disposalareas were also not considered because their composition wasaffected by particles emitted from the roast piles and mine tailings;it is also important to note that these metal(loid): metal ratios arenot affected by wind directions, because variations in winddirections do not cause any changes in the ratios betweendeposition rates of larger and smaller particles.


wet and dry conditions whereas the rates for the smallermetal–sulfate aerosols dramatically increased during wetdeposition (Chan and Lusis, 1985). These results suggestthat precipitations and consequently coatings closer to thesmelters became more enriched in Pb-, As- and Se-bearingmetal–sulfates relative to Fe–Ni-bearing particulate matterthan precipitations and coatings farther away from thesmelters. These enrichments may be visualized in plots withthe ratios between the concentrations of Pb and Ni, As andNi and Se and Ni in the coatings versus their distance to theCopper Cliff smelter (for more details see caption of

Fig. 11a–c). The inverse correlations between these metal(loid):metal ratios and the distance to the smelter indicateindeed that higher deposition rates of primary metal–sul-fates occurred closer to the smelter. They also suggest thatthe ratio between primary Pb–As–Se bearing sulfates andsecondary sulfates decreased with distance to the smelter,in agreement with our interpretation on the decrease ind34S with the distance to the smelter (see above).

4.2.6. Comparison with the study by Nriagu and Coker

(1978)

Nriagu and Coker (1978) collected rain samples fromfive to eight locations over ten different time intervals be-tween January 1975 and October 1976 and found no varia-tions in the d34S values with the distance to the Copper Cliffsmelter. On the contrary, the d34S values of the coatingsrepresent the isotopic composition of almost 70 years ofwet and dry deposition events. The apparent statistical mis-match between both studies and their different results indi-cate that any systematic change in the ratio betweenatmospheric primary and secondary sulfates with distancecan be only detected when monitoring the isotopic compo-sition of dry and wet depositions over a longer period oftime.

4.2.7. Comparison with S-isotope studies on gypsum coatings

of building stones

The atmospheric origin of S in gypsum coatings onbuilding stones was verified by many sulfur-isotope studies(e.g. Pye and Schiavon, 1989; Prikryl et al., 2004). The ob-served decrease in d34S between gypsum coatings in the citycenter of Prague (4&) and its surroundings (2.5&) were ex-plained with fractionation during the isotopic exchangereactions (R7) and (R8) and a decrease in particulates cat-alyzing the heterogeneous oxidation reaction (R9) (Buzekand Sramek, 1985). Considering the d34S values for SO2(gas)

(2.3&), aerosols (3.7&) and sulfates scavenged in wet pre-cipitation (5.7&) for the city of Prague, the observed de-crease of d34S in the gypsum coatings could be alsoexplained with a decrease in the ratio between primaryand secondary atmospheric sulfate species.

4.3. Conclusions and outlook on the use of silica coatings as

forensic tools for compositional atmospheric studies

The black coatings of the Sudbury area preserved tracemetals from atmospheric constituents in the form of smelterparticles (mainly Ni and minor Cu) and precipitated sul-fates (S, Pb, As, Se and Cu). Good to excellent correlationsbetween concentrations for Cu versus Se, Pb versus As andPb versus S on the micrometer, millimeter and on the regio-nal scale indicate the common occurrence of these elementsin minerals within coatings as well as their common occur-rence in atmospheric particulate matter and aerosols. Onthe regional scale, the coatings preserved in addition:

(a) A decrease in metal concentrations with the distancefrom the smelter point sources; an expected resultthat has been, however, never before described forrock coatings.


(b) The sulfur isotope composition of primary and sec-ondary atmospheric sulfates incorporated into coat-ings, and an increasing ratio between secondaryand primary sulfates with the distance to the smelter.

(c) Higher scavenging and thus higher deposition ratesof primary sulfates containing Pb, As and Se bywet precipitations.

(d) Similar deposition rates of larger Ni-bearing particlesunder wet and dry conditions.

(e) A decreasing ratio between primary metal–sulfateaerosols and larger Ni-bearing particulate matterwith the distance to the smelter.

Although the latter observations may have been ex-pected from the monitoring of wet and dry deposition ratesof aerosols and particular matter, they have not been ob-served for soils, lakes and coatings in the vicinity of smelt-ers prior to this study.

The comparison between emission data and atomic ra-tios for Pb and As and Cu and Ni in the coatings as wellas the inverse correlations between the ratios for Pb:Ni,As:Ni and Se:Ni and the distance to the Copper Cliff smel-ter showed that the coatings did not preserve the atomic ra-tios of elements in the smelter plume. This study as well asthe study by Mantha et al. (2012) showed that the atomicratios of elements in the coatings is mainly controlled bythe occurrence of minerals in the sulfate-rich layers as wellas their ability to incorporate trace metals with respect totheir size, coordination and charge.

Similar to the black coatings of the Greater Sudburyarea, silica-rich coatings on Mars are believed to be prod-ucts of the weathering of the underlying rock and containsulfates and amorphous silica as their main constituents(Tosca et al., 2005; Hurowitz et al., 2006; Tosca andMcLennan, 2006). This study showed that silica and Fe-sul-fate coatings preserve the isotopic signature of sulfur in-volved in their formation as well as deposition rates foratmospheric sulfate species. In this way, silica coatings onMars may contain similar information with respect toatmospheric sources and processes.

ACKNOWLEDGMENTS

This work was supported by a NSERC Discovery grant toM.S., and an Ontario Graduate Scholarship to N.M. K.K.acknowledges financial support from NSERC, CFI and OCE.We would like to thank T. Ulrich for technical assistance at theLA-ICP-MS, K. Klassen for her help at the Queen’s UniversityStable Isotope Lab, B. Kamber and D. Pearson for comments onan earlier version of the Manuscript and J. Durocher and K. Sker-ies for their assistance in sampling, sample preparation and samplemeasurements. We also like to thank four anonymous reviewersand Associate Editor Edward Ripley for their comments on an ear-lier version of the manuscript.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.gca.2012.05.013.

REFERENCES

Bao H. M. and Reheis M. C. (2003) Multiple oxygen and sulfurisotopic analyses on water-soluble sulfate in bulk atmosphericdeposition from the southwestern United States. J. Geophys.

Res. 108, 4430–4438.

Brock C. A., Hamill P., Wilson J. C., Jonsson H. H. and Chan K.R. (1995) Particle formation in the upper tropical troposphere –a source of nuclei for the stratospheric aerosol. Science 270,

1650–1653.

Buzek F. and Sramek J. (1985) Sulfur isotopes in the study of stonemonument conservation. Stud. Conserv. 30, 171–176.

Carpenter R. H., Pope T. A. and Smith R. L. (1975) Fe–Mn oxidecoatings in stream sediment geochemical surveys. J. Geochem.

Explor. 4, 349–363.

Chao T. T. and Theobald, Jr., P. K. (1976) The significance ofsecondary iron and manganese oxides in geochemical explora-tion. Econ. Geol. 71, 1560–1569.

Chan W. H., Vet R. J., Skelton G. B. and Lusis M. A. (1982a) Size

distribution and emission rate measurements of particulates in the

93m Falconbridge smelter stack plume, 1979. Report ARB-TDA-57-80, Air Resources Branch, Ontario Ministry of theEnvironment.

Chan W. H., Vet R. J., Lusis M. A. and Skelton B. G. (1982b) Size

distribution and emission rate measurements of particulates in the

Inco 381 m chimney and iron ore recovery plant stack fumes,

1979–80. Report ARB-TDA-62-80, Air Resources Branch,Ontario Ministry of the Environment.

Chan W. H., Vet R. J., Ro C. U., Tang A. J. S. and Lusis M. A.(1982c) Precipitation quality and wet deposition in the Sudbury

basin: Sudbury environmental cumulative precipitation network

results. Report ARB-04-82-ARSP, Air Resources Branch,Ontario Ministry of the Environment.

Chan W. H., Vet R. J., Ro C. U., Tang A. J. S. and Lusis M. A.(1982d) An analysis of the impact of smelter emissions on

precipitation quality and wet deposition in the Sudbury area:

Sudbury environmental study event precipitation network results.Report ARB-05-82-ARSP, Air Resources Branch, OntarioMinistry of the Environment.

Chan W. H., Vet R. J., Ro C. U., Tang A. J. S. and Lusis M. A.(1982e) An analysis of the impact of smelter emissions on

atmospheric dry deposition in the Sudbury area: Sudbury

environmental study event precipitation network results. ReportARB-12-81-ARSP, Air Resources Branch, Ontario Ministry ofthe Environment.

Chan W. H., Vet R. J., Lusis M. A. and Skelton G. B. (1983)Airborne particulate size distribution measurements in nickelsmelter plumes. Atmos. Environ. 17, 1173–1181.

Chan W. H. and Lusis M. A. (1985) Post-superstack smelteremissions and their fate in the atmosphere: an overview of theSudbury environment study results. Water Air Soil Pollut. 26,

43–58.

Choel M., Deboudt K., Flament P., lecornet G., Perdix E. andSobanska S. (2006) Fast evolution of tropospheric Pb- and Zn-rich particles in the vicinity of a lead smelter. Atmos. Environ.

40, 4439–4449.

Chorover J., Kretschmar R., Garcia-Pichel F. and Sparks D. L.(2007) Soil biogeochemical processes within the Critical Zone.Elements 3, 321–326.

Cornell R. M. and Schwertmann U. (1996) The Iron Oxides.

Structure, Properties, Reactions, Occurrences and Uses. VCHVerlagsgesellschaft, New York, USA.

Davies B. E. (1983) Heavy metal contamination from base metalmining and smelting: implications for man and his environ-ment. In Applied Environmental Geochemistry (ed. I. Thornton).Academic Press London, pp. 425–462.




Deng D., Nelson Y. M., Lion L. W., Shuler M. L. and Ghiorse W.C. (2000) Adsorption of Pb and Cd onto metal oxides andorganic material in natural surface coatings as determined byselective extractions: new evidence for the importance of Mnand Fe oxides. Water Res. 34, 427–436.

Dorn R. (1998) Rock Coatings. Elsevier, Amsterdam, 429p.Dugger D. L., Stanton J. H., Irby B. N., McConnell B. L.,

Cummings W. W. and Maatman R. W. (1964) The exchange oftwenty metal ions with the weakly acidic silanol group of silicagel. J. Phys. Chem. 68, 757–760.

Durocher J. and Schindler M. (2011) Iron-hydroxide, iron-sulfateand hydrous-silica coatings in acid-mine tailings facilities: acomparative study of their trace-element composition. Appl.

Geochem. 26, 1337–1352.

Dutrizac J. E. and Jambor J. L. (2000) Jarosites and theirapplication in hydrometallurgy. Rev. Mineral. Geochem. 40,

405–452.

Dutrizac J. E., Dinardo O. and Kaiman S. (1981) Selenateanalogues of jarosite-type compounds. Hydrometallurgy 6,

327–337.

Dutrizac J. E., Hardy D. J. and Chen T. T. (1996) The behavior ofcadmium during jarosite precipitation. Hydrometallurgy 41,

269–285.

Eriksen T. E. (1972a) Sulfur isotope-effects I. Isotopic-exchangecoefficient for sulfur isotopes 34S–32S� in the system SO2g–HSO3

� aq at 25, 35, and 45 degrees C. Acta Chem. Scand. 26,

573–580.

Eriksen T. E. (1972b) Sulfur isotope-effects II. The isotopicexchange coefficient for the sulfur isotopes 34S–32S in thesystem SO2g-aqueous solutions of SO2. Acta Chem. Scand. 26,

581–584.

Eriksen T. E. (1972c) Sulfur isotope-effects III. Enrichment of 34S�by chemical exchange between SO2g and aqueous-solutions ofSO2. Acta Chem. Scand. 26, 575–579.

Faure G. (1998) Principles and Applications of Geochemistry,second ed. Prentice-Hall Inc., New Jersey, 600p.

Fein J. B., Brady P. V., Jain J. C., Dorn R. I. and Lee J. (1999)Bacterial effects on the mobilization of cations from a weath-ered Pb-contaminated andesite. Chem. Geol. 158, 189–202.

Filipek L. H., Chao T. T. and Carpenter R. H. (1981) Factorsaffecting the portioning of Cu, Zn, and Pb in boulder coatingsand stream sediments in the vicinity of a polymetallic sulfidedeposits. Chem. Geol. 33, 45–65.

Forrest J. and Newman L. (1977a) Further studies on the oxidationof sulfur dioxide in coal-fired power plant plumes. Atmos.

Environ. 11, 465–474.

Forrest J. and Newman L. (1977b) Oxidation of sulfur dioxide inthe Sudbury smelter plume. Atmos. Environ. 11, 517–520.

Freedman B. and Hutchinson T. C. (1980) Pollutant inputs fromthe atmosphere and accumulations in soils and vegetation neara nickel–copper smelter at Sudbury, Ontario, Canada. Can. J.

Bot. 58, 108–132.

Giester G. and Rieck B. (1995) Mereiterite, K2Fe[SO4]2 4H2O, anew leonite-type mineral from the Lavrion mining district,Greece. Eur. J. Mineral. 7, 559–566.

Giorgetti G. and Baroni C. (2007) High-resolution analysis of silicaand sulphate-rich rock varnishes from Victoria Land (Antarc-tica). Eur. J. Mineral. 19, 381–389.

Hamill P., Jensen E. J., Russell P. B. and Bauman J. J. (1997) Thelife cycle of stratospheric aerosol particles. Bull. Am. Meteorol.

Soc. 78, 1395–1410.

Harris E., Sinha B., Hoppe P., Crowley J. N., Ono S. and Foley S.(2012) Sulfur isotope fractionation during oxidation of sulfurdioxide: gas-phase oxidation by OH radicals and aqueousoxidation by H2O2, O3 and iron catalysis. Atmos. Chem. Phys.

12, 407–424.

Hawley J. E. and Nichol I. (1959) Selenium in some Canadiansulfides. Econ. Geol. 54, 608–628.

Henderson P. J., McMartin L., Hall G. E., Percival J. B. andWalker D. A. (1998) The chemical and physical characteristicsof heavy metals in humus and till in the vicinity of the basemetal smelter at Flin Flon, Manitoba, Canada. Environ. Geol.

34, 39–58.

Hermann H., Ervens B., Jacobi H. W., Wolke R., Nowacki P. andZellner R. (2000) CAPRAM2.3: a chemical aqueous phaseradical mechanism for tropospheric chemistry. J. Atmos. Chem.

36, 231–284.

Hoar K., Nowinski P., Hodge V. and Cizdziel J. (2010) Rockvarnish: a passive forensic tool for monitoring recent airpollution and source identification. Nucl. Technol. 175, 351–

359.

Hodge V. F., Farmer D. E., Diaz T. A. and Orndorff R. L. (2005)Prompt detection of alpha particles from 210Po: another clue tothe origin of rock varnish? J. Environ. Radioact. 78, 331–342.

Hogan G. D. and Wotton D. L. (1984) Pollutant distribution andeffects on forests adjacent to smelters. J. Environ. Qual. 13, 377–

382.

Holt B. D. (1991) The oxygen isotope analysis of sulphur andoxygen isotopes: oxygen isotopes. In Stable Isotopes: Natural

and Anthropogenic Sulphur in the Environment (eds. H. R.Krouse and V. A. Grinenko). Wiley, New York, pp. 55–64,

Scope 43.

Holt B. D. and Kumar R. (1984) Oxygen-18 study of high-temperature air oxidation of SO2. Atmos. Environ. 15, 2089–

2094.

Holt B. D., Kumar R. and Cunningham P. T. (1982) Primarysulfates in atmospheric sulfates: estimation by oxygen isotoperatio measurements. Science 217, 51–53.

Hurowitz J. A., McLennan S. M., Tosca N. J., Arvidson R. E.,Michalski J. R., Ming D. W., Schroder C. and Squyres S. W.(2006) In situ and experimental evidence for acidic weatheringor rocks and soils on Mars. J. Geophys. Res. 111, E02S19.

Hutchinson T. C. and Whitby L. M. (1974) Heavy-metal pollutionin the Sudbury mining and smelting region of Canada. I. Soiland vegetation contamination by nickel, copper, and othermetals. Environ. Conserv. 1, 123–131.

Jackson T. A. (1978) The biogeochemistry of heavy metals inpolluted lakes and streams at Flin Flon, Canada, and aproposed method for limiting heavy-metal pollution of naturalwaters. Environ. Geol. 2, 173–189.

Jambor J. L., Nordstrom D. K. and Alpers C. N. (2000) Metal–sulfate salts from sulfide mineral oxidation. Rev. Mineral.

Geochem. 40, 305–350.

Kalkstein L. S. and Greene J. S. (1997) An evaluation ofclimate/mortality relationships in large U.S. and the possibleimpacts of a climate change. Environ. Health Perspect. 105, 84–

93.

Knight R. D. and Henderson P. J. (2006) Smelter dust in humusaround Rouyn-Noranda, Quebec. Geochem. Exploit. Environ.

Anal. 6, 203–214.

Krouse H. R. and Mayer B. (2000) Sulphur and oxygen isotopes insulfate. In Environmental Tracers in Subsurface Hydrology (eds.P. G. Cook and A. L. Herczeg). Kluwer Academic Publishers,

Boston, pp. 195–231.

Laroche C., Sirois G. and McIlveen W. D. (1979) Early Roasting

and Smelting Operations in the Sudbury Area – An Historic

Outline. Ontario Ministry of Environment and Energy, Sud-bury, Ontario.

Lein A. Y. (1991) Flux of volcanogenic sulphur to the atmosphereand isotopic composition of total sulphur. In Stable Isotopes:

Natural and Anthropogenic Sulphur in the Environment. Wiley,

Chichester, pp. 116–125.


Lion L. W., Altmann R. S. and Leckie J. O. (1982) Trace metaladsorption characteristics of estuine particulate matter: evalu-ation of contributions of Fe/Mn oxide and organic surfacecoatings. Environ. Sci. Technol. 16, 660–666.

Lusis M. A. and Wiebe H. A. (1976) The rate of oxidation of sulfurdioxide m the plume of a nickel smelter stack. Atmos. Environ.

10, 793–798.

Mantha N. M., Schindler M., Murayama M. and Hochella M. F.(2012) Silica- and sulfate-bearing rock coatings in smelter areas:products of chemical weathering and atmospheric pollution: I.Formation and mineralogical composition. Geochim. Cosmo-

chim. Acta 82, 254–274.

Mayer B. (1998) Potential and limitations of using sulphur isotopeabundance ratios as an indicator for natural and anthropogenicenvironmental change. In Isotope Techniques in the Study of

Environmental Change. Proceedings of an International Sympo-

sium on Isotope Techniques in the Study of Past and Current

Environmental Changes in the Hydrosphere and the Atmosphere;

Proceedings series (International Atomic Energy Agency), pp.423–435.

McAlister J. J., Smith B. J. and Torok A. (2006) Elementpartitioning and potential mobility within surface dusts onbuilding in a polluted urban environment, Budapest. Atmos.

Environ. 40, 6780–6790.

McGregor R. G., Blowes D. W., Jambor J. L. and Robertson W.D. (1998) The solid-phase controls on the mobility of heavymetals at the Copper Cliff tailings area, Sudbury, Ontario,Canada. J. Contamin. Hydrol. 33, 247–271.

McLennan S. (2003) Sedimentary silica on Mars. Geology 31, 315–

318.

Milliken R. E., Swayze G. A., Arvidson R. E., Bishop J. L., ClarkR. N., Ehlmann B. L., Green R. O., Grotzinger J. P., Morris R.V., Murchie S. L., Mustard J. F. and Weitz C. (2008) Opalinesilica in young deposits on Mars. Geology 36, 847–850.

Newman L. (1981) Atmospheric oxidation of sulfur dioxide: areview as viewed from power plants and smelter plume studies.Atmos. Environ. 15, 2231–2399.

Nord A. G., Svardh A. and Tronner K. (1994) Air pollution levelsreflected in deposits on building stone. Atmos. Environ. 28,

2615–2622.

Norman A. L., Belzer W. and Barrie L. A. (2004a) Insights into thebiogenic contribution to aerosols total sulphate and precipita-tion in the Fraser Valley afforded by isotopes of sulfur andoxygen. J. Geophys. Res. D 109, D05311.

Norman A. L., Krouse H. R. and MacLeod J. (2004b) Appor-tionment of pollutant S in urban airshed: Calgary, Canada. Acase study. In Air Pollution Modeling and its Application (eds. C.Borrego and S. Incecik). Kluwer, New York, pp. 107–125.

Norman A. L., Anlauf K., Hayden K., Thompson B., Brook J. R.,Li S. M. and Bottenheim J. (2006) Aerosol sulphate and itsoxidation on the Pacific NW coast: S and O isotopes in PM2.5.Atmos. Environ. 40, 2676–2689.

Nriagu J. O. and Coker R. D. (1978) Isotopic composition ofsulphur in atmospheric precipitation around Sudbury, Ontario.Nature 274, 883–885.

Nriagu J. O. and Harvey H. H. (1978) Isotopic variation as anindex of sulphur pollution in lakes around Sudbury, Ontario.Nature 273, 223–224.

Nriagu J. O. and Wong H. K. (1983) Selenium pollution of lakesnear the smelters at Sudbury, Ontario. Nature 301, 55–57.

Ozvacic V. (1982) Emissions of sulphur oxides, particulates and trace

elements in the Sudbury basin. Ontario Ministry of EnvironmentReport SES 008/82, pp. 82.

Perry R. S., Lynne B., Sephton M., Kolb V., Perry C. and Staley J.(2006) Baking black opal in the desert sun: the importance ofsilica in desert varnish. Geology 34, 537–540.

Potter R. M. and Rossman G. R. (1979) Mineralogy of manganesedendrites and coatings. Am. Mineral. 64, 1219–1226.

Prikryl R., Svobodova J., Zak K. and Hradil D. (2004) Anthro-pogenic origin of salt crusts on sandstone sculptures of Prague’sCharles Bridge (Czech republic): evidence of mineralogy andstable isotope geochemistry. Eur. J. Mineral. 16, 609–618.

Pye K. and Schiavon N. (1989) Cause of sulphate attack onconcrete, render and stone indicated by sulphur isotope ratios.Nature 342, 663–664.

Rattray K. J., Taylor M. R., Bevan D. J. M. and Pring A. (1996)Compositional segregation and solid solution in the lead-dominant alunite-type minerals from Broken Hill, N.S.W..Mineral. Mag. 60, 779–785.

Robinson G. D. (1981) Adsorption of Cu, Zn and Pb near sulfidedeposits by hydrous manganese-iron oxide coatings on streamalluvium. Chem. Geol. 22, 65–79.

Robinson G. D. (1982) Trace metal adsorption potential of phasescomprising black coatings on stream pebbles. J. Geochem.

Explor. 17, 205–219.

Rocca A., Vinals J., Arranz M. and Calero J. (1999) Character-ization and alkaline decomposition/cyanidation of beudantite–jarosite materials from Rio Tinto gossan ores. Can. Metall. Q.

38, 93–103.

Saltzman E. S., Brass G. and Price D. (1983) The mechanism ofsulfates aerosol formation: chemical and sulfur isotopic evi-dence. Geophys. Res. Lett. 10, 513–516.

Schiffman P., Zierenberg R., Marks N., Bishop J. and Dyar M.(2006) Acid-fog deposition at Kilauea volcano: a possiblemechanism for the formation of siliceous–sulfate rock coatingson Mars. Geology 34, 921–924.

Schindler M., Huminicki D. M. C. and Hawthorne F. C. (2006)Sulfate minerals: I. Bond topology and chemical composition.Can. Mineral. 44, 1403–1430.

Schindler M., Durocher J. L., Abdu Y. and Hawthorne F. C.(2009) Hydrous silica coatings: occurrence, speciation of metals,and environmental significance. Environ. Sci. Technol. 43, 8775–

8780.

Schindler M., Fayek M. and Hawthorne F. C. (2010) Uranium-richopal from the Nopal I uranium deposit, Pena Blanca, Mexico:evidence for the uptake and retardation of radionuclides.Geochim. Cosmochim. Acta 74, 187–202.

Schwarcz H. P. (1973) Sulfur isotope analyses of some Sudbury,Ontario, ores. Can. J. Earth Sci. 10, 1444–1459.

Scott K. M. (1987) Solid solution in, and classification, of gossan-derived members of the alunite–jarosite family, northwestQueensland, Australia. Am. Mineral. 72, 178–187.

Sinha B. W., Hoppel P., Huth J., Foley S. and Andreae M. O.(2008) Atmospheric chemistry and physics sulfur isotopeanalyses of individual aerosol particles in the urban aerosol ata central European site (Mainz, Germany). Atmos. Chem. Phys.

8, 7217–7238.

Smith B. J. and Prikryl R. (2007) Diagnosis decay: the value ofmedical analogy in understanding the weathering of buildingstones. In Building Stone Decay: From Diagnosis to Conserva-

tion, vol. 271 (eds. R. Prikryl and B. J. Smith). Geological Soc.,

London, pp. 1–8.

Sobanska S., Ricq N., Laboudigue A., Guillermo R., Bremard C.,Laureyns J., Merlin J. C. and Wignacourt J. P. (1999)Microchemical investigation of dust emitted by a lead smelter.Environ. Sci. Technol. 33, 1334–1339.

Squyres S. W., Arvidson R. E., Ruff S., Gellert R., Morris R. V.,Ming D. W., Crumpler L., Farmer J. D., Des Marais D. J., YenA., McLennan S. M., Calvin W., Bell, III, J. F., Clark B. C.,Wang A., McCoy T. J., Schmidt M. E. and de Souza, Jr., P. A.(2008) Detection of silica-rich deposits on Mars. Science 320,

1063–1067.


Stumm W. (1992) Chemistry of the Solid–Water Interface.

Processes at the Mineral–Water and Particle–Water Inter-

face in Natural Systems. John Wiley and Sons, New York,428p.

Tanaka N., Rye D. M., Xiao Y. and Lasaga A. C. (1994) Use ofstable sulfur isotope systematics for evaluating oxidationreaction pathways and in-cloud scavenging of sulfur-dioxidein the atmosphere. Geophys. Res. Lett. 21, 1519–1522.

Telmer K., Bonham-Carter G. F., Kliza D. A. and Hall G. A.(2004) The atmospheric transport and deposition of smelteremissions: Evidence from the multi-element geochemistry ofsnow, Quebec, Canada. Geochim. Cosmochim. 68, 2691–2980.

Thode H. G., Dunford H. B. and Shima M. (1962) Sulfur isotopeabundances in rocks of the Sudbury district and their geologicalsignificance. Econ. Geol. 57, 565–578.

Tosca N. J. and McLennan S. M. (2006) Chemical divides andevaporate assemblages on Mars. Earth Planet. Sci. Lett. 241,

21–31.

Tosca N. J., McLennan S. M., Clark B. C., Grotzinger J. P.,Hurowitz J. A., Knoll A. H., Schroder C. and Squyres S. W.(2005) Geochemical modeling of evaporation processes on

Mars: insight from the sedimentary record at meridian Planum.Earth Planet. Sci. Lett. 240, 122–148.

Ulrich T., Kamber B. S., Jugo P. J. and Tinkham D. K. (2009)Imaging element-distribution patterns in minerals by laserablation – inductively coupled plasma – mass spectrometry(LA–ICP–MS). Can. Mineral. 47, 1001–1012.

Wayne D., Diaz T., Fairhurst R., Orndorff R. and Pete D. (2006)Direct major- and trace-element analyses of rock varnish byhigh resolution laser ablation inductively-coupled plasma massspectrometry (LA-ICP MS). Appl. Geochem. 21, 1410–1431.

Wren C. (2012) Risk assessment and environmental management: acase study in Sudbury, Ontario, Canada. Prog. Environ. Sci.

Technol. Manage. 1, 1–450.

Whitney P. R. (1981) Heavy metals and manganese oxides in theGenese Watershed, New York State: effects of geology and landuse. J. Geochem. Explor. 14, 95–118.

Zoltai S. (1988) Distribution of base metals in peat near a smelterat Flin Flon, Manitoba. Water Air Soil Pollut. 37, 217–228.

Associate editor: Edward M. Ripley

silica- and sulfate-bearing rock coatings in smelter areas ... · 2009; durocher and schindler,...

Documents