Combined Chemical and Mineralogical Evidence for Heavy Metal Binding in Mining- and Smelting-Affected Alluvial Soils

Pedosphere 18(4): 464–478, 2008

ISSN 1002-0160/CN 32-1315/P

c© 2008 Soil Science Society of China

Published by Elsevier Limited and Science Press

Combined Chemical and Mineralogical Evidence for Heavy

Metal Binding in Mining- and Smelting-Affected Alluvial Soils∗1

A. VANEK1, V. ETTLER2, T. GRYGAR3, L. BORUVKA1, O. SEBEK2 and O. DRABEK1

1Department of Soil Science and Soil Protection, Czech University of Life Sciences Prague, Kamycka 129, 165 21 Praha

6 (Czech Republic). E-mail: [email protected], [email protected] of Geochemistry, Mineralogy and Mineral Resources, Charles University, Albertov 6, 128 43 Praha 2 (Czech

Republic)3Institute of Inorganic Chemistry, Analytical Laboratory, Academy of Sciences of the Czech Republic, 250 68 Rez u Prahy

(Czech Republic)

(Received April 11, 2007; revised March 14, 2008)

ABSTRACTThe binding of metallic contaminants (Pb, Cd, and Zn) and As on soil constituents was studied on four highly con-

taminated alluvial soil profiles from the mining/smelting district of Prıbram (Czech Republic) using a combination of

mineralogical and chemical methods. Sequential extraction analysis (SEA) was supplemented by mineralogical investi-

gation of both bulk samples and heavy mineral fractions using X-ray diffraction analysis (XRD) and scanning electron

microscopy with an energy dispersive X-ray spectrometer (SEM/EDS). The mineralogy of Fe and Mn oxides was studied

by voltammetry of microparticles (VMP) and diffuse reflectance spectrometry (DRS). Zinc and Pb were predominantly

bound in the reducible fraction attributed to Fe oxides and Mn oxides (mainly birnessite, Na4Mn14O27·9H2O), which

were detected in soils by XRD and SEM/EDS. In contrast, Cd was the most mobile contaminant and was predominantly

present in the exchangeable fraction. Arsenic was bound to the residual and reducible fractions (corresponding to Fe

oxides or to unidentified Fe-Pb arsenates). SEM/EDS observations indicate the predominant affinity of Pb for Mn oxides,

and to a lesser extent, for Fe oxides. Thus, a more suitable SEA procedure should be used for these mining-affected soils

to distinguish between the contaminant fraction bound to Mn oxides and Fe oxides.

Key Words: alluvial soil, Fe and Mn oxides, heavy metals, mineralogy, mining

Citation: Vanek, A., Ettler, V., Grygar, T., Boruvka, L., Sebek, O. and Drabek, O. 2008. Combined chemical and

mineralogical evidence for heavy metal binding in mining- and smelting-affected alluvial soils. Pedosphere. 18(4): 464–

478.

INTRODUCTION

Heavy metal contamination of the fluvial environment is a very common phenomenon in mining andsmelting areas (Swennen et al., 1994; Hudson-Edwards et al., 1996; Cappuyns et al., 2006). Heavy metal-bearing particles derived from mining-related activities enter these systems primarily through mine orprocessing waste discharging (tailings from crushing, milling or dressing operations), or remobilizationof mining-contaminated alluvium and mine drainage (Hudson-Edwards, 2003). It is not uncommon formore than 90% of the total metal load in rivers to be transported in the solid phase, either adsorbedonto particle surfaces and coatings, or incorporated into mineral grains (Miller, 1997). Fluvial processesexert the greatest influence on deposition and redistribution of these heavy particles in the ecosystemssurrounding the stream.

The behaviour and further fate of heavy metals after their retention in alluvial systems during floodevents is controlled by the stability of the heavy metal-bearing phases and the ambient soil physicoche-

∗1Project supported by the Higher Education Development Fund (FRVS) of the Ministry of Education, Youth and Sports

of the Czech Republic (No. 217/2005), the Czech Science Foundation (No. GACR 205/04/1292), and the Ministry of

Education, Youth and Sports of the Czech Republic (Nos. MSM 6046070901 and MSM 0021620855).

HEAVY METAL BINDING IN ALLUVIAL SOILS 465

mical conditions. Metal ions bound to the solid phase can be mobilised into the solution phase by changesin the pH, temperature, redox potential, organic matter decomposition, leaching, and ion exchange, orby microbial activity (Kennedy et al., 1997). At the present time, the type of metal bonding in soils ismost frequently studied using the sequential extraction analysis (Kabala and Singh, 2001; Ettler et al.,2005; Tongtavee et al., 2005). The main disadvantage of this method is the lack of selectiveness causedby redistribution and readsorption of some metals during the extraction procedure and the possibleabsence of some phases, such as carbonates or oxides (Kheboian and Bauer, 1987).

This study was focused on the chemical and mineralogical investigation of the inorganic conta-minants in alluvial soils from the mining/smelting district of Prıbram, Czech Republic, and followedup several previous studies dealing with bulk metal contamination in this area (Boruvka et al., 1996,1997). The novelty of this study laid in the combination of sequential extraction results with furthermineralogical investigations using X-ray diffraction analysis (XRD), scanning electron microscopy withan energy dispersive X-ray spectrometer (SEM/EDS), and other methods, which enabled an accuratedetermination of heavy metal bonding and speciation in soils and thus assessed the risks for the localecosystem. The goals of this study were to determine the vertical distribution of Pb, Zn, Cd, and As inalluvial soil profiles located at different distances from the main pollution sources and to determine thespeciation of the studied contaminants in the individual soil horizons using a combination of chemicaland mineralogical methods.

MATERIALS AND METHODS

Study sites and soil samples

The Prıbram district, situated approximately 60 km southwest of Prague, the capital of the CzechRepublic, has a long history of Pb and Ag ore mining and processing. The most important polymetallicore deposits (Pb-Ag-Zn) are located at Brezove Hory (a district of the city of Prıbram) and Bohutın(Fig. 1a). These areas were contaminated between the end of the 18th century and 1979 by intensemining of Ag-bearing galena (PbS) and sphalerite (ZnS) (Bambas, 1990). The most significant pollutionsource of alluvial soils surrounding the Litavka River consisted in fine-grained tailings from the historicaltailing ponds from the Brezove Hory mining area (Fig. 1a). Its barriers were repeatedly damaged (inparticular in 1932 and 1952), leading to an extreme contamination of the river (Vurm, 2001). The heavy

Fig. 1 Location of the alluvial soil profiles and the main pollution sources in the Prıbram mining/smelting area (a) as

well as the flood peaks between 1931 and 2003 (according to the data of the Czech Hydrometeorological Institute, 2005)

(b). P1–10 are the peaks on May 31, 1932, July 9, 1954, August 23, 1977, June, 18, 1979, July 22, 1980, July 20, 1981,

August 11, 1983, May 30, 1986, June 26, 1995, and August 13, 2002, respectively.

466 A. VANEK et al.

metal load originating from this source was removed further downstream during historical extensiveflood events in the years 1932, 1954, 1977, 1979, 1980, 1981, 1983, 1986, 1995, and 2002 (Fig. 1b). Thisfinding was also supported by the Pb isotopic composition of the Litavka River sediments exhibiting thesignature of the local galena (PbS) mined at the Brezove Hory ore deposit and historically processed inthe smelter (Ettler et al., 2006).

Four profiles of alluvial soils were sampled in this study (Fig. 1a). The first profile was situated nearBohutın (49◦ 39.536′ N, 13◦ 56.990′ E) (Profile B) and the second was located 1.5 km downstream in theclose vicinity of Brezove Hory (49◦ 41.001′ N, 13◦ 59.014′ E), the centre of the historical mining (ProfileBH). Profile BH was sampled 0.1 km from the mining area because of the regulated river channel. Thelocation of the profiles was chosen so that they represent soils contaminated predominantly by mining-related activities in the area. Profile L close to Lhota (49◦ 42.672′ N, 13◦ 59.663′ E), where a Pb smelteris situated, and Profile TD close to Trhove Dusnıky (49◦ 43.55′ N, 14◦ 0.53′ E), sampled 3.8 and 5.5km downstream, respectively, may represent soils partly contaminated by smelting activities.

All the alluvial soils samples were collected from pits down to a depth of 90–120 cm. Three tosix soil horizons were distinguished in the profiles. The top surface composed of fresh grass cover wasremoved. Surface organomineral (humic) Ah horizons were collected on each site. Mineral horizonsincluded either sediment layers without apparent reductive changes (M1) and with slight reduction(Mg), or gleyic horizons with strong changes caused by different levels of reduction and reoxidation (Go,Gro, Gor, and Gr). The signatures and depths of the horizons including physicochemical parameters ineach particular profile are given in Table I. Samples were collected from the central part of each horizonto avoid heterogeneity in the transitional zones. Approximately 1 kg of each horizon was collected toget representative soil sample. The samples collected were stored in polyethylene (PE) bags and treatedimmediately on returning to the laboratory. The soils were classified as Gleyic Fluvisols (FAO, 2006).

TABLE I

Selected physicochemical propertiesa) of the alluvial soils studied

Profile Sample Horizon Depth Particle size distribution pH CEC TOC TS

No.Clay Silt Sand H2O KCl

cm % cmol kg−1 g kg−1 mg kg−1

Bohutın B1 Ah 0–25 4.5 30.3 65.2 4.8 4.3 17.3 28 232.4

(B) B2 M1 25–70 0.6 12.5 86.9 5.6 5.1 8.8 8 63.5

B3 Mg 70–90 0.0 7.3 92.7 5.7 5.3 5.8 6 50.0

B4 Gr > 90 4.5 14.4 81.1 4.6 4.3 16.9 34 2 399.0

Brezove BH1 Ahg 0–23 5.3 28.7 66.0 5.5 4.9 27.1 54 481.0

Hory BH2 Gro 23–40 8.7 36.4 54.9 5.6 4.9 14.6 2 105.1

(BH) BH3 Gor 40–70 4.9 25.1 70.0 5.7 5.4 9.1 7 129.0

BH4 Gr Not sampled

Lhota L1 Ahg 0–15 7.4 31.1 61.5 4.3 3.9 16.8 40 506.9

(L) L2 Go 15–25 9.2 22.3 68.5 4.7 4.4 9.6 12 194.7

L3 Gro1 25–35 8.4 32.7 58.9 4.8 4.2 10.7 12 102.7

L4 Mg 35–55 11.8 24.4 63.8 4.6 4.1 11.7 10 70.4

L5 Gro2 55–85 8.2 15.7 76.1 4.9 4.3 6.2 3 26.0

L6 Gr > 85 11.7 17.0 71.3 5.0 4.1 5.5 2 15.8

Trhove TD1 Ah 0–15 6.2 38.8 55.0 6.0 5.6 22.1 40 367.5

Dusnıky TD2 Go 15–25 0.0 30.8 69.2 5.5 5.1 7.5 9 97.6

(TD) TD3 Gro 25–35 6.4 51.0 42.6 5.3 4.8 14.5 18 183.4

TD4 Gor1 35–50 8.5 39.0 52.5 5.2 4.6 13.0 14 139.5

TD5 Gor2 > 50 2.5 37.8 59.7 5.5 4.7 13.4 8 67.4

a)CEC = cation exchange capacity; TOC = total organic C; TS = total S.

All the soil samples were air dried to constant weight, sieved to < 2 mm through a stainless steel

HEAVY METAL BINDING IN ALLUVIAL SOILS 467

sieve and homogenized. An aliquot part of 30 g obtained by quartering was finely ground in a planetaryagate mortar (Fritsch, Germany) and used for analytical procedures (total contents of metals, organic C,and total S) and mineralogical investigations. The < 2 mm fraction was used for sequential extractionprocedure to preserve in-site soil composition. Prior to microscopic investigation of the metal-bearingphases, the heavy mineral fraction was separated from the most contaminated soil samples (BH3, L1,and L2) by a 24-h specific-gravity separation in bromoform as a heavy liquid (0.2 g of soil in 10 mL ofheavy liquid).

Determination methods

Basic soil properties. The soil pH was measured in water with a 1:5 ratio of soil and wateror 1 mol L−1 KCl solution (ISO 10390:1994) using an inoLab Level 1 pH meter. Soil agitation for5 minutes and settling for 2 hours preceded the pH measurements. The contents of total organic C(TOC), total inorganic C (TIC), and total S (TS) were determined using an ELTRA Metalyt CS1000Selemental analyser. To determine the cation exchange capacity (CEC), the soil was saturated with Bacations using 0.1 mol L−1 BaCl2. Barium was subsequently released using MgSO4 (ISO 11260:1994).Excess Mg present in the solution was determined using a Varian SpectrAA 240 flame atomic absorptionspectrophotometer (FAAS). The particle size distribution was determined by the hydrometer method(Gee and Bauder, 1986).

Bulk metal concentrations. To determine the bulk metal concentrations, an amount of 0.2 g ofair-dried sample was dissolved using a mixture of hot acids (10 mL of HF and 0.5 mL of HClO4) andevaporated to dryness. This procedure was repeated with HF (5 mL) and HClO4 (0.5 mL) to totallydissolve all the silicates present. The residue obtained was dissolved in 100 mL of 20 mL L−1 HCl.During the analytical procedure, the MilliQ+ deionized water and double distilled HF, HClO4, andHCl acids (Merck, Germany) were used. The metal concentrations in solutions were determined usinga Varian SpectrAA 200 HT FAAS under standard analytical conditions.

Sequential extraction analysis (SEA). The sequential extraction method of Tessier et al. (1979)was used in this experiment. The detailed experimental scheme is given elsewhere (Tessier et al., 1979;Gleyzes et al., 2002). The interpretation of individual chemical fractions is described in Ettler etal. (2005). All the leachates, except the residual fraction, were separated from the solid phase usingcentrifugation. Each residue was washed with 2 mL of deionized water (MilliQ+) before the followingstep to completely wash out the previous reagent. The solution obtained was added to the leachate fromthe previous step before adding the acidifying agent (1 mL of double distilled HNO3) and dissolved in50 mL of deionized water. In the case of the silicate fraction, the residue after total dissolution wasdissolved in 50 mL of 20 mL L−1 HCl. The contents of Pb, Zn, Cd, As, Fe, and Mn in individual fractionswere determined using a Varian SpectrAA 280 FS (fast sequence) FAAS under standard analyticalconditions. The calibration was matrix-matched. A procedural blank was run for each extraction stepand the standard deviation of the duplicate analysis was < 10%. The sum of the individual fractionscorresponded well to the bulk metal and As concentrations obtained by total dissolution. The agreementwas very good for Zn and Pb (r2 = 0.994 and 0.992, respectively) and relatively good for As and Cd(r2 = 0.949 and 0.930, respectively). Although the Tessier SEA is not designed for anionic speciessuch as As (Gleyzes et al., 2002; Mihaljevic et al., 2003), the extracts were also analysed for As todetermine possible binding to the individual fractions and to correlate these data with the mineralogicalinvestigation of As binding to soil constituents. Mobility factor (MF), corresponding to the potentiallymobile amount of metallic contaminants, was calculated using the SEA results according to the equationof Kabala and Singh (2001): MF = [(A + B)/(A + B + C + D + E)] × 100 (%), where A to E are theindividual fractions of SEA. The exchangeable (A) and acid-extractable (B) fractions are considered tobe easily available.

Mineralogical investigations. Pulverised bulk soil samples and heavy mineral fractions of the mostcontaminated samples were analysed by X-ray diffraction analysis (XRD) on a PANalytical X’Pert Pro

468 A. VANEK et al.

instrument operating with CuKα radiation at 40 kV and 45 mA, scanning over the range 4◦–80◦ in2Θ, step size 0.02◦, counting time 150 s per step, and a X’Celerator multichannel detector. Qualitativeanalysis was performed with the X’Pert HighScore software Version 1.0 d (PANalytical, the Netherlands)and JCPDS PDF-2 database (ICDD, 2003). Voltammetry of microparticles (VMP) and ultraviolet-visible (UV-Vis) diffuse reflectance spectrometry (DRS) were employed to detect the presence of Feand Mn oxides. VMP was performed according to Grygar and van Oorschot (2002) using a µAutolabpotentiostat, a paraffin-impregnated graphite working electrode, 0.2 mol L−1 sodium acetate-aceticacid buffer (1:1), and a scan rate of 3 mV s−1. Goethite and hematite were distinguished after heatingsmall portions of samples at 320 ◦C. Interpretation of the voltammetric peak potentials was carriedout in accordance with Grygar et al. (2002). The voltammetric peaks were clearly discernible in thefirst scan without the need to subtract the second scan. DRS was performed according to the methoddescribed in Grygar et al. (2003). Semiquantitative estimates were obtained by comparison of theareas of the electron-pair transition bands of samples and a set of calibration mixtures of syntheticgoethite in a mineral matrix. The mineral grains were selected from the heavy mineral fractions undera binocular microscope and then analysed with a CamScan S4 scanning electron microscope equippedwith an Oxford Link ISIS energy dispersion spectrometer (SEM/EDS). For oxide and silicate phases,analytical conditions were: accelerating voltage 20 kV, beam current 1.5–3 nA, and counting time10 s. This instrument was used for semiquantitative chemical point analysis, for microphotographs inbackscattered and secondary electron modes, and for X-ray elemental mapping (the relative elementdistribution).

Statistical analysis. The statistical data treatment was performed using the NCSS software (Hintze,2001). Each sample was considered to be a vector with 33 variables (total metal, As, Fe, and Mn con-centrations, amounts of metals and As extracted by different methods, sample pH, CEC, TOC, TIC,and TS). As a result, a data matrix with dimensions of 18 × 33 was used for calculation of the Spearmanrank order correlations.

RESULTS AND DISCUSSION

Basic soil parameters

Selected physicochemical properties of the studied alluvial soil profiles are given in Table I. Thesurface horizons contained more organic C (up to 50 g kg−1) compared to the subsurface mineralhorizons (generally less than 20 g kg−1). Similarly, surface horizons showed higher total S (up to 500mg kg−1). The only example was the Gr horizon in the Profile B with a high S content (more than 2 000mg kg−1). The soil pH reached lower values (4.3–6.0 in water suspension), representing acidic soils. Thecation exchange capacity (CEC) varied from 5.5 to 27 cmol kg−1. The clay content in all the horizonsof the profiles was relatively low and accounted for 0–11.8% of the total particle size distribution. Thesoil texture of the studied soils was classified as sandy loam (FAO, 2006).

Chemical fractionation of metals and As

The bulk concentrations and results of SEA for Pb, Zn, Cd, and As are indicated in Table II andFig. 2. All the studied elements exhibit similar vertical distribution patterns in the individual alluvialsoil profiles. Profile B located upstream the main polymetallic mining site was the least contaminatedwith a uniform depth distribution of all the studied contaminants (Fig. 2). In contrast, Profile BH inthe close vicinity of the mines exhibited a significant increase in the metal and As concentrations,especially at the bottom of the profile (Fig. 2). This distribution can be caused by the late import ofless contaminated material from the upstream areas that settled during several recent flood events. Ageneral additional increase of metal and As concentrations was noticed in Profiles L and TD locatedfurther downstream (Fig. 2), showing that the contamination is migrating downstream in the fluvialsystem, especially during flood events.

HEAVY METAL BINDING IN ALLUVIAL SOILS 469

TABLE II

Total concentrations of metals and As in the studied alluvial soil profiles obtained by total digestion

Profile Sample No. Horizon Depth Fe Mn Pb Zn Cd As

cm mg kg−1

Bohutın (B) B1 Ah 0–25 30 200 2 057 1 410 257 6.5 80

B2 M1 25–70 29 450 1 616 1 615 174 5.5 110

B3 Mg 70–90 28 400 647 1 595 221 5.5 <DLa)

B4 Gr > 90 12 000 183 1 555 75 8.0 115

Brezove Hory (BH) BH1 Ahg 0–23 40 150 3 252 2 425 2 115 23.0 240

BH2 Gro 23–40 36 700 4 126 2 110 911 15.0 200

BH3 Gor 40–70 51 200 3 863 7 590 3 479 23.0 740

BH4 Gr Not sampled

Lhota (L) L1 Ahg 0–15 35 950 3 998 4 705 2 014 26.0 535

L2 Go 15–25 33 550 4 283 3 150 3 395 45.0 265

L3 Gro1 25–35 21 400 473 1 505 2 084 24.5 65

L4 Mg 35–55 18 050 773 55 1 382 19.0 20

L5 Gro2 55–85 19 700 261 15 501 3.0 70

L6 Gr > 85 11 000 124 20 333 3.0 25

Trhove Dusnıky (TD) TD1 Ah 0–15 45 344 6 340 4 500 3 687 31.0 103

TD2 Go 15–25 34 413 3 175 3 640 3 230 18.5 82

TD3 Gro 25–35 49 717 2 320 3 125 4 949 41.2 43

TD4 Gor1 35–50 20 759 1 707 875 3 552 40.0 <DL

TD5 Gor2 > 50 13 549 3 047 1 185 2 512 42.0 <DL

a)DL = detection limit.

The highest concentration of Pb was found in the bottom Gor horizon from Profile BH (7 590 mgkg−1) (Table II). Generally more than 50% of Pb was associated with the reducible fraction corre-sponding to Fe and Mn oxides (Figs. 2a–d). The relationship between Pb extracted within the reduciblefraction and Fe and Mn concentrations was manifested by statistically significant correlations: r = 0.83(P < 0.01) for Fe and 0.54 (P < 0.05) for Mn. A significant amount of Pb was bound in the residualfraction (up to 24% in the bottom horizon of Profile BH) and in the acid-extractable fraction (up to29%). The latter could be related to the presence of Pb carbonates that were observed in the Litavkastream sediments (Ettler et al., 2006) and could result from the mine tailing ponds whose barriers weredamaged several times, including the major flood events (Vurm, 2001; Vanek et al., 2005). In contrastto highly organic forest soils (Ettler et al., 2005), only up to 11% of the total Pb content was bound tothe organic (oxidisable) fraction. The exchangeable fraction of Pb (the fraction A) was correlated withTOC (r = 0.48, P < 0.05), which, in turn, was well correlated with CEC (r = 0.77, P < 0.001). Thisobservation shows that the most mobile Pb is bound to the organic matter. Mobility factor (MF) forPb ranged from 2% to 64%, indicating that the mobile Pb concentrations varied from 20 to 1 897 mgkg−1.

Although Zn concentrations were similar to those of Pb in Profiles B and BH, different patternsin the shallower subsurface horizons were observed in Profiles L and TD. This can be related either tothe deposition of material with different contaminant ratios (higher amounts of Zn in an earlier flood,i.e., deeper horizons), or to the post-depositional vertical migration of Zn in the alluvial soil profile.It is not possible to distinguish between these two possibilities in the absence of monitoring of floodsuspensions and sediments and their pollution effect in the past. The highest Zn concentration attained4 949 mg kg−1 (Table II). Zn speciation was similar to that of Pb with a predominant reducible fraction(Fe and Mn oxides), but with a slightly higher amount of the exchangeable fraction, indicating higherZn mobility (Figs. 2e–h). Zinc extracted in the reducible fraction was correlated with Fe obtained bythe same extraction (r = 0.51, P < 0.05), but not with Mn. Zinc extracted in the oxidisable fraction

470 A. VANEK et al.

Fig. 2 Results of Pb, Zn, Cd, and As fractionation obtained by the sequential extraction analysis (SEA) according to

Tessier et al. (1979) for the studied alluvial soil profiles: Profiles B (a, e , i, and m), BH (b, f, j, and n), L (c, g, k, and

o), and TD (d, h, l, and p).

was correlated with CEC and TOC (r = 0.49 and 0.52, respectively, P < 0.05). The MF values forZn in Profiles B and BH were relatively low, ranging from 6% to 25%, corresponding to the mobileconcentration of 7–222 mg kg−1. A significantly higher mobility was observed for Zn in Profiles L andTD with MF values of 19%–62%, corresponding to the mobile concentrations of 127–1 230 mg kg−1.This observation supports the hypothesis about the possibility of post-depositional vertical Zn migrationin the alluvial soil profiles, especially in the highly contaminated profiles downstream from the mainmining area.

HEAVY METAL BINDING IN ALLUVIAL SOILS 471

Cadmium showed similar concentration profiles to Zn probably because the principal source of Cdin the Prıbram mining district is Cd-bearing sphalerite (ZnS) (Bambas, 1990; Ettler et al., 2006). Themaximum concentration was 45 mg kg−1 in the horizon Go of Profile L (Table II). However, the Cdspeciation was completely different from that of Zn, as the former had more important exchangeableand acid-extractable fractions (Figs. 2i–l). This implies higher mobility of Cd as manifested by the MFvalues ranging from 51% to 94%, corresponding to the mobile concentrations of 2–46 mg kg−1. Nocorrelations between Cd extracted in the individual fractions and the other physicochemical parameterswere observed.

Interestingly, the majority of As was present in the residual fraction with a minor amount in thereducible fraction attributed to Fe and Mn oxides (Figs. 2m–p). Some previous studies showed, however,that the Tessier SEA designed for metallic cations does not correctly extract As bound to the targetphases. For example, Mihaljevic et al. (2003) demonstrated that Ca arsenate (if present in soil orsediment) may be dissolved not only in the acid-extractable fraction (as expected), but also in thereducible fraction. Similarly, if arsenopyrite (FeAsS) is present in the soil or sediment, the fractionD (organic matter and sulphides) is underestimated and the remaining arsenopyrite is subsequentlydissolved in the residual fraction (Mihaljevic et al., 2003). As a result, practically no As seems to bemobile in the soil profiles studied. This is also consistent with the fact that As is found in the Prıbrampolymetallic district mainly as arsenopyrite, stibarsen (AsSb), native As, or as admixtures in othersulphides or sulphosalts (Bambas, 1990). No correlations were observed between As extracted in theindividual fractions and the other physicochemical parameters.

Soil mineralogy

The XRD results showed that the bulk mineralogy of alluvial soils was dominated by quartz (SiO2)with minor amounts of feldspar (mainly albite, NaAlSi3O8), muscovite (KAl3Si3O10(OH)2), clinochlore((Mg,Al)6(Si,Al)4O10(OH)8), and kaolinite (Al2Si2O5(OH)4), and trace amounts of hornblende (mainlytremolite, Ca2Mg5Si8O22(OH)2). No metal-bearing phases were detected by XRD. The XRD analysiscarried out on the heavy mineral fraction of the most contaminated samples indicated the presence ofthe following phases: dominant siderite (FeCO3) derived from the ore gangue and also found in streamsediments from the Prıbram mining district (Bambas, 1990; Ettler et al., 2006), abundant hornblende,rutile/anatase (TiO2), and common Mn and Fe oxides such as birnessite (K4Mn14O27·9H2O), hematite(α-Fe2O3), or goethite (α-FeOOH).

Results of semiquantitative determination of Mn(III,IV) and Fe(III) oxides obtained by VMP andDRS (Table III) showed that the samples from Profile BH in the mining area were enriched in Mnoxides, especially in the most contaminated horizon Gor (Sample BH3). Less crystalline Fe(III) oxides(ferrihydrite and/or lepidocrocite) were mainly observed in the subsurface horizons of Profiles BH and L(Table III). In contrast, goethite (α-FeOOH) exhibited uniform concentrations in Profile BH (20 g kg−1)and a slight surface enrichment in Profile L (30 g kg−1). Nevertheless, similar concentrations of goethiteare common in natural soil samples. Good agreement was observed between the FeOOH concentrationobtained by DRS and that calculated from the amount of Fe in the reducible fraction. VMP revealed thepresence of free Pb(II) compounds in the majority of the samples (Table III), corresponding to readilysoluble Pb from chemical speciation, such as adsorbed and exchangeable forms. This observation mayalso be related to the occurrence of readily soluble Pb minerals (cerussite, PbCO3) or Pb bound toorganic matter.

The SEM/EDS observations showed that Fe and Mn oxides and smelter-derived particles werethe predominant heavy metal-bearing compounds in the studied alluvial soils. Reniform aggregates ofhematite (Fe2O3) originated from the fillings of hydrothermal veins in the Brezove Hory polymetallicmining district (Fig. 3a). The slight enrichment of Pb and Zn (Fig. 3b) in the hematite may be relatedto their precipitation from solutions percolating through the polymetallic (Pb-Ag-Zn) deposit. In thesoils from the mining area, Fe oxides precipitated as crusts on silicate grains and exhibited significantly

472 A. VANEK et al.

TABLE III

Semiquantitative determinationa) of Mn(III,IV) and Fe(III) oxides and Pb(II) compounds corresponding to easily solu-

ble Pb forms obtained by voltammetry of microparticles (VMP), diffuse reflectance spectroscopy (DRS) or sequential

extraction analysis (SEA) in the alluvial soil samples studied

Profile Sample Mn(III,IV) oxides Ferrihydrite or lepidocrocite Goethite Goethite Pb(II) compounds

No. by VMP by DRS by DRS by SEA by VMP

g kg−1

Brıezove BH1 +/– – 20 13 +

Hory (BH) BH2 + ++ 20 20 +

BH3 ++ – 20 13 ++

Lhota (L) L1 + +/– 30 18 +

L2 + + 30 19 +

L3 – – 10 9 +

L4 – – 10 8 +/–

L5 – – 30 9 –

L6 – – 10 4 –

a)++ = abundant, + = present, +/– = trace, – = below detection limit.

higher concentrations of Pb, Zn, and As (Figs. 3c, d). Grains approximately 5 µm in size, probably cor-responding to Pb-Fe arsenate, were also observed in the silicate matrix particles (Figs. 3c, d). Manganeseoxides formed flakes on silicate grains and exhibited high concentrations of Pb and Zn as revealed byEDS (Figs. 3e, f). These Mn oxides may correspond to birnessite, with composition Na4Mn14O27·9H2O,which has been identified as the most efficient Pb scavenger in contaminated environments (Hudson-Edwards, 2000; O’Reilly and Hochella, 2003). EDS analysis of this phase indicated the absence of theNa peak (and also of the K peak); however, it can be masked by Zn: the Kα line of Na at 1.041 keVoverlaps with the Lα line of Zn at 1.012 keV with resolution ±59 eV (Fig. 3f).

In contrast to Fe oxides from Profile BH, sphere-like aggregates of Fe(III) precipitates in Profile Lcontained only low concentrations of Zn as revealed by EDS (Figs. 4a, b). This observation was alsoconfirmed by X-ray elemental mapping performed on a zone with both Mn(III,IV) and Fe(III) oxides(Sample L1) (Fig. 5). The relative elemental distribution showed that all the Pb was bound in Mn oxides,and Fe oxides were Pb-free. The relative distribution of Zn also indicated locally higher concentrationswithin the Mn oxide zone. Arsenic was also inhomogeneously distributed, showing hotspots within theMn oxide zone. Smelter-derived particles consisted mainly of fragments of silicate slag glass enriched inPb (Figs. 4c, d). Partly altered metallic droplets (e.g., Cu sulphides) that were initially trapped withinthe slag glass were also observed (Figs. 4e, f). The chemical composition of the slag glass showing theelemental association of Si-Ca-Fe-Pb corresponds to that of the Pb-Zn pyrometallurgical slag glass fromthe Prıbram district (Ettler et al., 2001, 2002).

Contaminant binding and mobility

The concentrations of all the studied contaminants (Table 2, Fig. 2) significantly exceeded the ma-ximum tolerable levels defined by the Czech legislation (Czech Regulation 13/1994) (Ministry of En-vironment of the Czech Republic, 1994): Cd, 1 mg kg−1; Pb, 140 mg kg−1; Zn, 200 mg kg−1; andAs, 30 mg kg−1. The maximum contaminant concentrations found in the alluvial soils were up to 5times lower than those observed in the stream sediments in the same area (Ettler et al., 2006); thisobservation was in agreement with the data of Hudson-Edwards et al. (2001), who found similar resultsin the Rıo Pilcomayo mining-affected river system in Bolivia, where the metal and As concentrations inthe high-water channel sediments (alluvium) are lower than those in the low-water sediments (streamsediments).

The direct presence of ore minerals or slag particles in the stream sediment or alluvium can indicatethe provenance of the sediment (Hudson-Edwards, 2003). The secondary minerals can be derived directly

HEAVY METAL BINDING IN ALLUVIAL SOILS 473

Fig. 3 Scanning electron micrographs in secondary electrons or back-scattered electrons (BSE) and corresponding energy

dispersive X-ray spectra of heavy particles in the alluvial soil from the mining area (Profile BH): reniform aggregate of

primary hematite (Fe2O3) originating from the polymetallic ore gangue (a and b); silicate particles covered by newly

formed Fe(III) oxides and inlet image of particles in BSE with indication of heavy grains of Pb-Fe arsenate (c and d);

flakes of Mn(III,IV) oxides associated with silicate grains (e and f).

from the weathered mine tailing ponds or may form after deposition in the sediment (Hudson-Edwardset al., 1996; Hudson-Edwards, 2003). Generally, less than 10% of primary ore minerals (sulphides) ispresent in sediments and the metallic contaminants are mainly bound to the newly formed Fe and Mn(hydr)oxides found as coatings on lithic fragments (Hudson-Edwards, 2003).

Numerous studies on mining-affected river systems have shown that sorption and co-precipitationof metals and metalloids with Fe and Mn oxides are amongst the most important chemical processesof contaminant (Pb, Zn, Cd, and Cu) binding in stream sediments and alluvial soils (Hudson-Edwards,

474 A. VANEK et al.

Fig. 4 Scanning electron micrographs in secondary electrons or back-scattered electrons (BSE) and corresponding energy

dispersive X-ray spectra of heavy particles in the alluvial soil from the smelting area (Profile L): spherulite aggregates of

Fe(III) oxides (a and b); glassy particle of metallurgical slag (c and d); altered surface of slag glass with bubble texture

and etched Cu-sulphide droplet and Fe-Cr spinel crystal trapped within the glass (BSE zoomed image of zone indicated

in the previous photo) (e and f).

2003; Cappuyns et al., 2006). This fact was also found in alluvial soils of this study, in particular forPb and Zn, by both chemical and mineralogical methods (Figs. 2, 3, 4, and 5). In fact, the chemicalfractionation of metals was very similar to that of the stream sediments from the Litavka River (Ettleret al., 2006) with a predominant reducible fraction of Pb and Zn. X-ray elemental mapping and EDSanalyses showed that Pb was preferentially related to Mn oxides, while Zn was bonded to both Fe andMn oxides (Figs. 3, 4, and 5). It was observed that Mn oxides are generally more efficient sorbents ofPb than Fe oxides (O’Reilly and Hochella, 2003). This fact is strongly related to the negative surface

HEAVY METAL BINDING IN ALLUVIAL SOILS 475

Fig. 5 Scanning electron micrograph and X-ray distribution maps of Pb, Zn, As, Fe, Mn, Si, and Al of zone with newly

formed Fe and Mn oxides associated with primary quartz and clay minerals from the surface horizon of the alluvial soil

profile close to the Pb smelter (Sample L1).

charge in circumneutral conditions and to the efficient binding of Pb2+ and other divalent cations(Hudson-Edwards, 2000). The chemical fractionation study on floodplain sediments in northeast Eng-land showed that Mn oxides, representing important carriers for Pb and Zn, are more soluble than Feoxides, and high amounts of these metals can be readily released when metal-bearing Mn oxides aredissolved under reducing conditions (Hudson-Edwards, 2000). Cadmium was bound predominantly inexchangeable, and to a lesser extent, in acid-extractable and reducible fractions, as also observed inother studies (Hudson-Edwards, 2003; Ettler et al., 2006), confirming its higher mobility and availabi-lity. No Cd was detected by SEM/EDS in Mn and Fe oxides, probably due to concentrations below thedetection limit (DL). Similarly, no specific Cd-bearing phases were detected by XRD and SEM.

Complex speciation of As was indicated by SEA and SEM observations. With the exception of aslightly enriched zone detected by X-ray elemental mapping (Fig. 5), only extremely low As concentra-tions in several Fe oxides were observed by EDS analysis. Nevertheless, the presence of Fe-Pb arsenatewas observed (Figs. 3c, d), probably contributing to the recovery of As in the reducible fraction as dis-cussed above (see also Mihaljevic et al., 2003). Interestingly, we did not detect any primary As-bearingphases (e.g., arsenopyrite) to explain the predominant presence of the residual fraction obtained for Aschemical fractionation by SEA.

The present study indicated that the chemical fractionation study was properly supplemented by amulti-method mineralogical investigation of the contaminated soils. However, it also showed that theSEA of Tessier et al. (1979) was not capable of distinguishing sufficiently between metals bound to Mnand Fe oxides, but yielded only the total reducible fraction that included both these target phases.For example, the SEM/EDS investigation showed that Pb was preferentially bound to Mn oxides anddifferent SEA procedure with more extraction steps should be used to distinguish more soluble Mnoxides and less soluble Fe oxides (e.g., Leleyter and Probst, 1999; Hudson-Edwards, 2000).

Possible pathways of alluvium contamination

Boruvka et al. (1996) and Vanek et al. (2005) suggested that Pb smelter emissions may also con-tribute to the contamination of alluvial soils in the Prıbram district. However, the recent Pb isotopicinvestigation of stream sediments of the Litavka River (Ettler et al., 2006) and the alluvial soils from apresent study (data not yet published) showed that the isotopic ratio 206Pb/207Pb is close to the valueof 1.16 and corresponds perfectly to galena mined and historically processed in the Prıbram district.

476 A. VANEK et al.

Thus, the roles of smelter emissions and ore mining activities can not be distinguished on the basis of thePb isotopic compositions. In contrast to naturally developed soils polluted by smelter emissions (Ettleret al., 2005), we suggest that fluvial transport of contaminated suspension and water from the minedumps and tailing ponds at Brezove Hory would be the dominant source of the alluvium contamination.Only locally, the Pb smelting industry may be responsible for the increase in the metal and metalloidconcentrations in the stream sediments and alluvial soils: in the vicinity of slag dumps (Profile L), smallslag and matte particles can contribute to the alluvium contamination (Figs. 4c, e), especially owing tothe dissolution of metal-bearing slag glass or small metallic droplets (Ettler et al., 2002).

Flood events in the last century (Fig. 1b) were probably responsible for the most important trans-port and redeposition of contaminated materials. Failures of tailing ponds may be responsible for thedischarge of tailing particles into the stream (Hudson-Edwards, 2003). Most of the metal-bearingparticles are generally fine-grained (< 2 mm) and transported in the fluvial system as a suspension(Hudson-Edwards, 2003; Miller, 1997). In the periods of low to mean discharge, metal-bearing particlesconcentrate in the stream (bed and channel) sediments. However, during flood events, they are trans-ported downstream more efficiently and are redeposited in floodplains as vertically-accreted alluvium(for appropriate references, see the review paper by Hudson-Edwards, 2003). Accidents on the dams oftailing ponds (in particular in 1932 and 1952) at Brezove Hory caused leakage of contaminated waterand tailing particles into the Litavka River (Vurm, 2001). Although these old tailing ponds are currentlybeing reclaimed, they may still be assumed to be a source of discharge of contaminated materials intothe Litavka fluvial system.

The studied alluvial soil profiles near Lhota and Trhove Dusnıky had surface horizons that weregenerally enriched in Pb, whereas deeper horizons were also enriched in Zn and Cd. This phenomenonmay be caused either by historically different sources of contamination with higher Zn/Pb and Cd/Pbratios or by post-depositional vertical migration of Zn and Cd in the profile. With the exception ofProfiles B and BH, Zn bulk concentrations were often higher than those of Pb in the studied horizons.A similar situation was observed in the stream sediments of the Litavka River (Ettler et al., 2006). Thisinteresting phenomenon can be related to the fact that the initial sources of contamination (mine tail-ings) are probably enriched in Zn, as historically only Pb- and Ag-bearing ores were selectively processedin the district and Zn sulphides were partly abandoned within the gangue. It is necessary furthermoreto note that Zn was not smelted in the Prıbram area (Vurm, 2001; Bambas, 1990).

The contamination moved downstream from the principal point sources, as follows from the com-parison of the concentration trends in the profiles from Bohutın to Trhove Dusnıky (Figs. 1a, 2). Thiswas also documented by the development of horizons; i.e., the most contaminated horizon BH3 in theclose vicinity of the mining area was buried under the less contaminated horizons, whereas the profilesdownstream (L and TD) were greatly enriched at the surface, which can indicate the redeposition of themost contaminated materials coming from the mining area during the latest flood events. The alluviumof the Litavka River still represented an important pool of metals and metalloids that can be mobilized(in dissolved, but mainly in particulate forms) and transported further by the stream during high flooddischarges. At the present time, the movement of contaminated materials in the Litavka catchment area(suspension and dissolved species) is being studied by other research teams on the basis of measurementsmade last year.

CONCLUSIONS

The present study, focused on contaminated alluvial soils from the polymetallic mining/smeltingdistrict of Prıbram (Czech Republic), showed how detailed mineralogical investigation can properlyextend the chemical fractionation data. Such an approach constitutes a key contribution for correctinterpretations of chemical data on highly polluted soil and sediment systems. Zinc and Pb were mostlybound in the reducible fraction attributed to Fe oxides and Mn oxides (mainly birnessite), which weredetected in the soil by both XRD and SEM/EDS. In contrast, Cd was found to be the most mobile and

HEAVY METAL BINDING IN ALLUVIAL SOILS 477

was predominantly present in the exchangeable fraction. Arsenic was bound to the reducible fraction(bound to Fe oxides or unidentified Fe-Pb arsenates) and to the residual fraction. It was also foundthat Pb was mainly bound to Mn oxides and to a lesser extent to Fe oxides. Manganese(III,IV) oxidesare more soluble than Fe(III) oxides and the Tessier sequential extraction analysis is not capable ofdistinguishing between them. As a result, a sequential extraction procedure with specific extraction forthe Mn oxide fraction should be carried out for these samples polluted by mining activities.

ACKNOWLEDGEMENTS

A number of colleagues, Alexander Martaus (XRD) from the Institute of Chemical TechnologyPrague, Martin Mihaljevic (SEA), Maria Hojdova (statistical data treatment) and Radek Prochazka(SEM) from Charles University, assisted with the analytical procedures. Karel Zak from the Academyof Sciences of the Czech Republic assisted in the discussion of pollution pathways in the Prıbram district.The Czech Hydrometeorological Institute, especially Zdenka Vilhelmova, is thanked for the historicalflood data. The anonymous reviewer is thanked for improvement of the manuscript.

REFERENCES

Bambas, J. 1990. Ore District of Brezove Hory (in Czech). VZ Kamenna Publications, Prıbram. 200pp.

Boruvka, L., Huang-Wei, C., Kozak, J. and Kristoufkova, S. 1996. Heavy contamination of soils with cadmium, lead and

zinc in the alluvium of the Litavka River. Rostl. Vyr. 42: 543–550.

Boruvka, L., Kristoufkova, S., Kozak, J. and Huang-Wei, C. 1997. Speciation of cadmium, lead and zinc in heavily polluted

soils. Rostl. Vyr. 43: 187–192.

Cappuyns, V., Swennen, R., Vandamme, A. and Niclaes, M. 2006. Environmental impact of the former Pb-Zn mining and

smelting in East Belgium. J. Geochem. Explor. 88: 6–9.

Czech Hydrometeorological Institute. 2005. Hydrological Data of Maximum Flow Rates of Litavka River between the

Years of 1931–2003. Czech Hydrometeorological Institute, Prague.

Ministry of Environment of the Czech Republic. 1994. Czech Regulation 13/1994. Specifications for Protection of

Agricultural Soils. Ministry of Environment of the Czech Republic, Prague.

Ettler, V., Legendre, O., Bodenan, F. and Touray, J. C. 2001. Primary phases and natural weathering of old lead-zinc

pyrometallurgical slag from Prıbram, Czech Republic. Can. Mineral. 39: 873–888.

Ettler, V., Mihaljevic, M., Sebek, O., Molek, M., Grygar, T. and Zeman, J. 2006. Geochemical and Pb isotopic evidence for

sources and dispersal of metal contamination in stream sediments from the mining and smelting district of Prıbram,

Czech Republic. Environ. Pollut. 142: 409–417.

Ettler, V., Mihaljevic, M., Touray, J. C. and Piantone, P. 2002. Leaching of polished sections: an integrated approach for

studying the liberation of heavy metals from lead-zinc metallurgical slags. Bul. Soc. Geol. France. 173: 161–169.

Ettler, V., Vanek, A., Mihaljevic, M. and Bezdicka, P. 2005. Contrasting lead speciation in forest and tilled soils heavily

polluted by lead metallurgy. Chemosphere. 58: 1 449–1 459.

Food and Agriculture Organisation of the United Nations (FAO). 2006. World Reference Base for Soil Resources. FAO,

Rome.

Gee, G. W. and Bauder, J. W. 1986. Particle-size analysis. In Klute, A. (ed.) Methods of Soil Analysis. Part 1: Physical and

Mineralogical Methods. American Society of Agronomy-Soil Science Society of America, Madison, WI. pp. 383–411.

Gleyzes, C., Tellier, S. and Astruc, M. 2002. Fractionation studies of trace elements in contaminated soils and sediments:

a review of sequential extraction procedures. Trends in Anal. Chem. 21: 451–467.

Grygar, T., Bezdicka, P., Hradil, D., Domenech-Carbo, A., Marken, F., Pikna, L. and Cepria, G. 2002. Voltammetric

analysis of Fe oxide pigments. Analyst. 127: 1 100–1 107.

Grygar, T., Dedecek, J., Kruiver, P., Dekkers, M. J., Bezdicka, P. and Schneeweiss, O. 2003. Iron oxide mineralogy in

late Miocene red beds from La Glorie, Spain: rock-magnetic, voltammetric and Vis spectroscopy analyses. Catena.

53: 115–132.

Grygar, T. and van Oorschot, I. H. M. 2002. Voltammetric identification of pedogenic iron oxides in paleosol and loess.

Electroanalysis. 14: 339–344.

Hintze, J. 2001. NCSS and PASS: Number Cruncher Statistical System. Kaysville, Utah. Available online at http://www.

ncss.com.

Hudson-Edwards, K. A. 2000. Heavy metal-bearing Mn oxides in river channel and floodplain sediments. In Cotter-

Howells, J. D., Campbell, L. S., Valsami-Jones, E. and Batchelder, M. (eds.) Environmental Mineralogy: Microbial

Interactions, Anthropogenic Influences, Contaminated Land and Waste Management. The Mineralogical Society

Series Vol. 9. Turpin Distribution Services Ltd pp. 207–226.

Hudson-Edwards, K. A. 2003. Sources, mineralogy, chemistry and fate of heavy metal-bearing particles in mining-affected

river systems. Mineral. Mag. 67: 205–217.

478 A. VANEK et al.

Hudson-Edwards, K. A., Macklin, M. G., Curtis, C. D. and Vaughan, D. J. 1996. Processes of formation and distribution

of Pb-, Zn-, Cd-, and Cu-bearing minerals in the Tyne basin, northeast England: implications for metal-contaminated

river systems. Environ. Sci. Technol. 30: 72–80.

Hudson-Edwards, K. A., Macklin, M. G., Miller, J. R. and Lechler, P. J. 2001. Sources, distribution and storage of heavy

metals in the Rıo Pilcomayo, Bolivia. J. Geochem. Explor. 72: 229–250.

International Center for Diffraction Data (ICDD). 2003. JCPDS PDF-2 Database. ICDD, Newtown Square, USA.

Kabala, C. and Singh, B. R. 2001. Fractionation and mobility of copper, lead, and zinc in soil profiles in the vicinity of a

copper smelter. J. Environ. Qual. 30: 485–492.

Kennedy, V. H., Sanchez, A. L., Oughton, D. H. and Rowland, A. P 1997. Use of single and sequential extractants to

assess radionuclide and heavy metal availability from soils for root uptake. Analyst. 122: 89–100.

Kheboian, C. and Bauer, C. F. 1987. Accuracy of selective extraction procedures for metal speciation in model aquatic

sediments. Anal. Chem. 59: 1 417–1 423.

Leleyter, L. and Probst, J. L. 1999. A new sequential extraction procedure for the speciation of particulate trace elements

in river sediments. Int. J. Environ. Anal. Chem. 73: 109–128.

Mihaljevic, M., Ponavic, M., Ettler, V. and Sebek, O. 2003. A comparison of sequential extraction techniques for

determining arsenic fractionation in synthetic mineral mixtures. Anal. Bioanal. Chem. 377: 723–729.

Miller, J. R. 1997. The role of fluvial geomorphic processes in the dispersal of heavy metals from mine sites. J. Geochem.

Explor. 58: 101–118.

O’Reilly, S. E. and Hochella, M. F. 2003. Lead sorption efficiencies of natural and synthetic Mn and Fe-oxides. Geochim.

Cosmochim. Acta. 67: 4 471–4 487.

Swennen, R., van Keer, I. and de Vos, W. 1994. Heavy metal contamination in overbank sediments of the Guel River

(East Belgium): its relation to former Pb-Zn mining activities. Environ. Geol. 24: 12–21.

Tessier, A., Campbell, P. G. C. and Bisson, M. 1979. Sequential extraction procedure for the speciation of particulate

trace metals. Anal. Chem. 51: 844–850.

Tongtavee, N., Shiowatana, J. and McLaren, R. G. 2005. Fractionation of lead in soils affected by smelter activities using

a continuous-flow sequential extraction system. Int. J. Environ. Anal. Chem. 85: 567–583.

Vanek, A., Boruvka, L., Drabek, O., Mihaljevic, M. and Komarek, M. 2005. Mobility of lead, zinc and cadmium in alluvial

soils heavily polluted by smelting industry. Plant Soil Environ. 51: 316–321.

Vurm, K. 2001. The History of the Prıbram Smelter (in Czech). Kovohute Prıbram Editions.