Environmental Research 133 (2014) 77–89

Contents lists available at ScienceDirect

Environmental Research journal homepage: www.elsevier.com/locate/envres

Distribution of chemical elements in soils and stream sediments in the area of abandoned Sb–As–Tl Allchar mine, Republic of Macedonia Katerina Bačeva a, Trajče Stafilov a,n, Robert Šajn b, Claudiu Tănăselia c, Petre Makreski a a

Institute of Chemistry, Faculty of Natural Sciences and Mathematics, Ss. Cyril and Methodius University, POB 162, 1000 Skopje, Macedonia Geological Survey of Slovenia, Ljubljana, Slovenia c INCDO-INOE 2000 Research Institute for Analytical Instrumentation (ICIA), Cluj-Napoca, Romania b

art ic l e i nf o

a b s t r a c t

Article history: Received 19 December 2013 Received in revised form 29 March 2014 Accepted 31 March 2014

The aim of this study was to investigate the distribution of some toxic elements in topsoil and subsoil, focusing on the identification of natural and anthropogenic element sources in the small region of rare As–Sb–Tl mineralization outcrop and abandoned mine Allchar known for the highest natural concentration of Tl in soil worldwide. The samples of soil and sediments after total digestion were analyzed by inductively coupled plasma–mass spectrometry (ICP–MS) and inductively coupled plasma–atomic emission spectrometry (ICP–AES). Factor analysis (FA) was used to identify and characterize element associations. Six associations of elements were determined by the method of multivariate statistics: Rb– Ta–K–Nb–Ga–Sn–Ba–Bi–Li–Be–(La–Eu)–Hf–Zr–Zn–In–Pd–Ag–Pt–Mg; Tl–As–Sb–Hg; Te–S–Ag–Pt–Al–Sc– (Gd–Lu)–Y; Fe–Cu–V–Ge–Co–In; Pd–Zr–Hf–W–Be and Ni–Mn–Co–Cr–Mg. The purpose of the assessment was to determine the nature and extent of potential contamination as well as to broadly assess possible impacts to human health and the environment. The results from the analysis of the collected samples in the vicinity of the mine revealed that As and Tl elements have the highest median values. Higher median values for Sb are obviously as a result of the past mining activities and as a result of area surface phenomena in the past. & 2014 Elsevier Inc. All rights reserved.

Keywords: Soil Sediments As–Sb–Tl Allchar Toxic elements distribution

1. Introduction In many regions natural mineral deposits contain particularly large quantities of heavy metals, although the anthropogenic activities, such as mining and smelting of metal ores, have increased distribution of trace elements that appear to be a main reason for environmental pollution. Specifically, mine, opencast mining activities and mine tailings have a serious environmental impact on soils and water streams. In addition, these areas present severe erosion problems caused by wind and water run-off, where soil and mine spoil texture, landscape topography and regional and local meteorological conditions play an important role (Navarro et al., 2008). The abandoned mines and mining areas are important issue because can be a major source of environmental pollution. In many areas worldwide, present and historical mining and smelting activities are causing a variety of environmental problems such as elevated metal concentrations in soils/sediments, dispersion of toxic elements in soil and water and ecological damage caused by

n

Corresponding author. Fax: þ 38 92 322 6865. E-mail address: [email protected] (T. Stafilov).

http://dx.doi.org/10.1016/j.envres.2014.03.045 0013-9351/& 2014 Elsevier Inc. All rights reserved.

extensive metal pollution (Alloway, 1995). Because ore is only a small fraction of the total volume of mined material, ore extraction, beneficiation processes and further processing of ores produce large amounts of waste that can contain metals (Siegel, 2002). Although the contaminants have a long residence time in soils and waste materials at historical mining sites, research indicates that the mobility and bioavailability of metals in many of these environments is still high; the transfer of contaminants to the food chain and exposure of the local population still occur; significant quantities of contaminants are still being transported off the site (Alvarenga et al., 2004; Concas et al., 2004; Lee et al., 2001; Merrington and Alloway, 1994). The objective of this investigation is to present the results from the soil survey in the As–Sb–Tl Allchar locality, Republic of Macedonia, abounded mine in the last 100 years (Boev and Jelenković, 2012; Volkov et al., 2006) and the river transport of sediments enriched with As, Sb, and Tl. These sediments are transported from Majdanska River to Crna River, and deposited in its alluvial sediment, on area of intensive agriculture activities. The content of Tl in the Allchar locality represents the highest established natural concentration in soil worldwide (Rieck, 1993). Special attention was given to the behavior of As, Sb and Tl and

78

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

other trace elements following the main As–Sb–Tl mineralization. Therefore, the goals of this research were: (a) to investigate the distribution of various elements (Ag, Al, As, Ba, Be, Bi, Br, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hf, Hg, Ho, I, In, K, La, Li, Lu, Mg, Mn, Mo, Na, Nb, Nd, Ni, P, Pb, Pd, Pr, Pt, Rb, Rh, S, Sb, Sc, Sm, Sn, Sr, Ta, Tb, Te, Th, Tl, Tm, V, W, Y, Yb, Zn and Zr) analyzed by inductively coupled plasma–mass spectrometry (ICP–MS) and inductively coupled plasma–atomic emission spectrometry (ICP– AES), in topsoil, subsoil and the river sediments; (b) to define by statistical methods the main geochemical association and their spatial distribution in soil and river transport; (c) to identify the distribution of the elements in the survey area as either geogenic or anthropogenic.

2. Materials and methods 2.1. Study area

Fig. 1. The location of the study area in the Republic of Macedonia.

The Allchar locality is rare antimony–arsenic–thallium mineralization outcrop, located on the northwestern part of Kožuf Mt., Republic of Macedonia (Fig. 1). The locality of Allchar is unique in its mineral composition, and excluding very intriguing mineral lorandite (TlAsS2), there are 45 other minerals, some of them very rare minerals. It is worldwide known locality as the richest deposit with thallium minerals and the largest number (12) different thallium minerals, four of them nowadays known as type-locality species: jankovicite (Tl5Sb9(AsSb)4S22), picotpaulite (TlFe2S3), rebulite (Tl5Sb5As8S22), simonite (TlHgAs3S6), (Boev et al.,

Fig. 2. Geological map of the studied area after Boev and Jelenković (2012).

K. Bačeva et al. / Environmental Research 133 (2014) 77–89 2001–2002; Frantz et al., 1994; Rieck, 1993; Jovanovski et al., 2012). In addition to economic Sb and As grades, the ore is substantially enriched in Tl. Allchar is the first Calin-type gold deposit discovered in the Balkan Peninsula in the late 1980s (Percival et al., 1990). Arsenic reserves at the deposit amount were approximately to 15,000 t (Boev et al., 2001–2002). The Sb mineralization was studied after World War II from 1958 to 1974; the estimated Sb reserves exceed 20,000 t with an average Sb grade of 0.5% (Boev et al., 2001–2002). The Allchar deposit is classed as large in terms of thallium reserves (over 500 t) and as medium according to gold and antimony reserves (Volkov et al., 2006). The locality is hydrothermal volcanogenic deposit close to the border between the Republic of Macedonia and Greece. The location of this volcanic complex in the Kožuf–Kilkis transverse zone and the intersection with the Vardar zone indicates a central type volcanism, activated on the tectonic intersection formed by the reactivated regional fault structures of Vardar strike (NW–SE to N–S) and the

79

Kožuf–Kilkis (E–W) fault structure formed during the neotectonic period. This type of volcanism is characterized by ring–radial structures (Boev and Jelenković, 2012).

2.2. Geological and mineralogical characteristics of the Allchar deposit Geological and mineralogical description are summarized by Jelenković and Boev (2011); Jelenković et al. (2011) and Boev and Jelenković (2012). The geological description of the investigated area is presented in Fig. 2. The Allchar Sb–As–Tl–Au volcanogenic hydrothermal deposit is situated at the northwestern margins of Kožuf Mts. From the geotectonic point of view, ore mineralization is related to a Pliocene volcano-intrusive complex located between the rigid Pellagonian block in the west, and the labile Vardar zone in the east. From the metallogenic point of view, the Allchar deposit belongs to the Kožuf ore district as part

Fig. 3. Shaded relief map with defined areas and sampling locations of soil and river sediment.

80

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

of the Serbo–Macedonian metallogenetic region. Deposition of sandstone and claystone, followed by bedded and massive carbonate rocks (limestone, dolomite, marble) took place in the Middle and Upper Triassic. These rocks are the base of the Allchar deposit. The quartz–sericite–feldspar schists are developed along the eastern flank of the deposit, while the central part is built of dolomite, marble, and sporadically limestone. The dolomite series underlies marbel. The Mesozoic rocks are unconformable overlain by Pliocene cover and glacial till. The unit of Pliocene felsic tuffs covers a large portion of the Allchar deposit. This volcanic sequence includes ash, crystal tuffs, tuff breccia and lacustrine tuffaceous sediments. The Allchar deposit is a NNW–SSE oriented antiform. It comprises several ore bodies within a zone 2 km long and around 300–500 m wide. The mineralization is characterized by increased porosity and permeability, typically related to fractures and fractured zones in the vicinity of subvolcanic intrusive bodies. The Allchar polychronous and polygenetic formation results of complex physico-chemical processes occurring in (i) heterogeneous geological environment, (ii) the interaction of multi-stage hydrothermal fluids with the products of polyphase magmatic activity and (iii) surrounding sedimentary and metamorphic rocks. The major elemental components of the Allchar deposit are Sb, As, Tl, Fe and Au, accompanied by minor Hg and Ba, and traces of Pb, Zn, Cu. Enrichment of Tl in the Allchar deposit is closely associated with increased concentrations of volatiles, such as As, Sb, Hg. The elevated Tl concentration is related to parts of the deposit where As, Sb and Hg are also high (Janković, 1993). It is the only deposit in Macedonia that contains economic grades of Tl (0.1–0.5%), Sb (up to 2.5%), As (1.5%), and Au ( 41 g/t) (Volkov et al., 2006). The Crven Dol ore body is situated in the northwestern part of the Allchar area. Nowadays, it is an abandoned undergrounds mine. The deposit of Crven Dol belongs to the second zone of the Allchar ore region. The basic rocks of andesites and dolomites consist mostly of sulfide minerals of arsenic (realgar, As4S4, and orpiment As2S3) and thallium (vrbaite, Hg3Tl4As8Sb2S20, and lorandite, etc.). This second zone contains less antimony, iron and other minerals (marcasite, FeS2, pyrite, FeS2, melnikovite var. marcasite, FeS2, etc.). Probable ore reserves are estimated at 4000 t with 6% As and 0.3% Tl. Thallium occurs primarily as the mineral lorandite, but also in association with other Tl–minerals, realgar is the most present arsenic mineral while orpiment is less abundant (Janković and Jelenković, 1994). 2.3. Sampling and sample preparation The entire investigated region (c. 13 km2) was covered by a basic sampling grid of 500  500 m2. The sampling density was increased especially in the central part of the studied area in mineralization outcrop and at abandoned Allchar mine. Altogether, 67 locations were defined (Fig. 3). With regard to the basic lithological units, 13 sampling sites were located on the Quaternary moraines; 4 were on the Pliocene–Quaternary tuff; 6 on Pliocene– Quaternary quartz–latite breccias; 5 on Pliocene andesitic, 6 on Jurassic serpentines and diabase; 11 on Triassic carbonates; 13 on Triassic clastites and 9 on altered Triassic Carbonates. Due to the expected extreme concentrations especially with As, Sb and Tl, the following zones (groups) were determined: Area of hydrothermally altered rocks (24 sampling locations); Above the As–Tl–Sb ore body Crven Dol area (5 location: A-36, A-37, A-41, A-116 and A-117); Group of samples, taken at tailings (2 locations: A-24 and A-29) and Area of non-altered rocks–background (43 locations). Samples of soil were collected according to European guidelines for soil contamination studies and also according to our own experience (Theocharopoulos et al., 2001; Šajn, 2005, 2006; Stafilov et al., 2008, 2010b, 2010a). The samples of topsoil (0–5 cm) and subsoil (20–30 cm) were collected at each location. One soil sample of about 1 kg weight represents the composite material collected from the four points within a radius of 10 m around the central point (Šajn, 2005, 2006). The eventually present organic horizon was excluded. To control a river transport of chemical elements, the samples of stream (river) sediments have been collected above the hydrothermally altered rocks (background), below the hydrothermally altered rocks and 10–18 km downstream of the hydrothermally altered rocks. One stream sediment sample represents the composite of 5 or more sub-samples collected along the river bank. The soil samples were air-dried indoors, cleaned of extraneous material and sifted through a plastic sieve with a 2 mm mesh. For chemical analysis, this fraction was ground in an agate mill and pulverized to fine powder. The size fraction of stream sediments smaller than 0.125 mm was prepared for the chemical analyses by sieving (Salminen et al., 2005). Soil and sediment samples (0.25 g) were placed in a Teflon digestion vessel and were digested on a hot plate. In the first step, HNO3 was added to remove all organic compounds, and then a mixture of HF and HClO4 was added, followed by a total digestion method where HCl and water were added to dissolve the residue. The solution was transferred quantitatively in 25 ml volumetric flask. 2.4. Instrumentation The investigated elements (Ag, Al, As, Ba, Be, Bi, Br, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hf, Hg, Ho, I, In, K, La, Li, Lu, Mg, Mn, Mo, Na, Nb, Nd, Ni, P, Pb, Pd, Pr, Pt, Rb, Rh, S, Sb, Sc, Sm, Sn, Sr, Ta, Tb, Te, Th, Tl, Tm, V, W, Y, Yb, Zn and Zr) were analyzed by the application of inductively coupled plasma with mass spectrometry (ICP–MS). All samples were analyzed also by ICP-AES (Varian, model

715-ES) for the elements with high contents (Al, Ca, Fe, K, Mg and Na). For ICP–AES, operating and instrumental conditions were presented in the paper of Balabanova et al. (2010). For ICP–MS measurements a SCIEX Perkin Elmer Elan DRC II (Canada) inductively coupled plasma mass spectrometer (with quadruple and single detector setup) was used. Optimal instrumental and operating conditions were presented in our previously published papers (Bačeva et al., 2012, 2011). The X-ray powder diffraction (XRPD) measurements were conducted on a Rigaku Ultima IV X-ray powder diffractometer. Each studied sample, previously powdered to dimensions bellow 100 μm, was manually positioned over a silicon sample plate and the data were collected on an ultra-fast D/tex detector in the 2θ range from 51 to 601 (scan rate 2θ 1/min). CuKα radiation was obtained from a generator set at 40 kV and a current of a 40 mA was applied. The debyegrams were analyzed by PDXL-2 software (Rigaku Corp.) and ICDD-PDF-2 database (release 2012) for phase identification and peak indexing was employed. Certified reference materials were used to validate the method for all considered elements and the difference between measured and certified values was satisfied ranging within 15%. Standard soil reference material (JSAC 0401) also ensured quality control. The measured concentrations were in good agreement with the recommended values.

2.5. Data processing and statistical analyses Analytical data and measurements were included to the data matrix. Each observation is described with several variables such as sample identification number, locality, geographic coordinates, sample type and concentration level for 62 elements. Geostatistical data interpretation and visualization (mapping) have been performed at Geological Survey of Slovenia by using following software's: Statistica (Stat Soft, Inc.), Autodesk MAP 3D (Autodesk, Inc.), ArcINFO (ESRI, Inc.) and Surfer (Golden Software, Inc.). Parametric and nonparametric statistical methods were used (Zhang et al., 1998), and normality tests of data distributions were performed. The degree of association of chemical elements was assessed using the linear coefficient of correlation—Peaerson r (Cohen, 1988). Multivariate R-mode factor analyses (FA) was used to reveal associations of chemical elements (Garson, 2000; Šajn, 2005, 2006). The factor analysis was performed on variables standardized to zero mean and unit standard deviation. Varimax method has been used for orthogonal rotation. The factor analysis (FA) from accurate number of variables provides smaller number of new variables, so called factors that present association of statistical significant variables. Universal kriging with linear variogram interpolation method was used for the construction of maps showing spatial distribution of factor scores, as well as maps displaying the distribution of trace elements (Davis, 1986). The basic grid cell size for interpolation was 25  25 m. For class limits, seven classes of the percentile values of distribution of interpolated values were chosen (0–10, 10–25, 25–40, 40–60, 60–75, 75–90 and 90–100).

3. Results and discussion 3.1. Statistical analysis The descriptive statistics and results of statistical tests of the 62 analyzed elements for 67 values of average concentrations of topsoil and subsoil samples in the same location are briefly summarized in the Tables 1a and 1b. On the basis of the normality tests and compared with histograms of distribution for the content of all analyzed elements in soil samples, the normality was assumed for naturally values of Al, Eu, Ga, I, In, K, Mg, Na, Nd, Pr, Rb and Th. For the rest of the elements, the distribution was assumed on the bases of the logarithms of their contents. ANOVA test was used to compare significance of the difference between the soil horizons, and the distribution of individual geologic units (soil). The differences between soils horizons are not significant except for the elements Cd, Pb and Zn in the soil. The significant variations for majority of the elements have been found in samples collected on different geological units. However, the variations between the soil horizons could not be determined due to the presence of mineralization. Decomposed mineralized fragments in soil cause a high local variance, which is the reason for large deviation among the element contents in soil horizons. Therefore, the average value of both topsoil and subsoil horizons has been used for further statistical analysis (Tables 1a and 1b).

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

81

Table 1a Descriptive statistics for values for the content of chemical elements (Ag–Mg) in the soil samples (0–30 cm), n ¼134.

Ag Al As Ba Be Bi Br Ca Cd Ce Co Cr Cs Cu Dy Er Eu Fe Ga Gd Ge Hf Hg Ho I In K La Li Lu Mg

Dis

Unit

DL

X

Xg

Md

Min

Max

S

Sx

CV

F (Soil)

F (Geo)

Log N Log Log Log Log Log Log Log Log Log Log Log Log Log Log N Log N Log Log Log Log Log N N N Log Log Log N

mg kg  1 % mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 % mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 % mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 % mg kg  1 mg kg  1 mg kg  1 %

0.01 0.00009 0.1 0.01 0.01 0.01 0.01 0.001 0.01 0.05 0.001 0.001 0.01 0.01 0.01 0.01 0.01 0.000001 0.01 0.01 0.01 0.01 10 0.01 10 10 0.0002 0.01 0.01 10 0.00005

0.51 4.2 410 540 2.6 0.16 2.2 3.3 0.39 28 13 83 6.1 24 1.5 0.77 0.55 2.0 10 2.1 0.18 1.0 550 0.28 95 22 2.2 17 21 130 1.2

0.46 4.3 53 400 2.2 0.14 1.7 1.2 0.27 24 12 65 4.4 18 1.4 0.73 0.53 1.9 10 2.0 0.17 0.77 110 0.27 88 22 2.0 14 18 130 1.1

0.48 3.8 73 380 2.2 0.14 0.64 1.5 0.27 22 11 61 4.1 19 1.3 0.67 0.48 1.9 9.3 1.9 0.17 0.78 110 0.24 75 20 1.9 13 18 110 1.0

0.21 0.84 10 32 0.29 0.028 o 0.01 0.12 0.049 1.9 2.8 11 0.36 3.2 0.29 0.13 0.085 0.63 1.5 0.47 0.046 0.080 o 10 0.049 o 10 o 10 0.27 0.85 3.1 28 0.26

1.3 7.8 9200 1900 10 0.67 16 26 2.7 120 37 500 37 270 5.3 2.8 1.3 5.0 17 6.5 0.39 3.9 11000 1.0 310 58 4.5 60 49 350 3.0

0.20 1.7 1400 450 1.8 0.096 2.4 5.0 0.38 18 7.0 77 6.2 29 0.78 0.41 0.25 0.75 4.0 1.1 0.066 0.83 1600 0.15 53 9.7 0.98 12 11 55 0.65

0.017 0.15 120 39 0.15 0.0083 0.21 0.43 0.033 1.6 0.60 6.6 0.53 2.5 0.067 0.036 0.022 0.064 0.35 0.092 0.0057 0.072 140 0.013 4.6 0.84 0.085 0.99 0.91 4.8 0.056

39 41 338 84 67 60 112 150 98 67 53 92 101 120 52 54 46 37 39 50 37 80 292 54 57 43 45 68 51 44 54

1.05NS 0.14NS 0.11NS 0.00NS 0.00NS 0.55NS 0.01NS 0.01NS 9.17NS 0.01NS 0.04NS 0.08NS 0.01NS 0.01NS 0.02NS 0.03NS 0.00NS 0.00NS 0.07NS 0.01NS 0.04NS 0.01NS 0.00NS 0.03NS 0.01NS 0.38NS 0.48NS 0.12NS 0.05NS 0.16NS 0.01NS

8.36n 8.35n 31.00n 14.96n 9.04n 11.14n 10.61n 22.25n 4.95n 9.01n 5.93n 18.70n 11.25n 2.90n 8.14n 9.47n 7.25n 1.19NS 8.55n 8.03n 1.37NS 13.48n 13.93n 9.05n 0.83NS 2.23n 15.53n 10.84n 6.67n 5.77n 18.02n

Dis.—distribution (N—normal, Log—lognormal;); DL—detection limit; X—mean; XG—geometric mean; Md—median; Min—minimum; Max—maximum; S—standard deviation; SX—standard error of mean ;CV—coefficient of variation; F (soil)—F ratio between soil horizon 0–5 cm vs 20–30 cm (ANOVA); F ratio between geological units (ANOVA); NS—no significant; n—significance at po 0.05; nn—significance at p o 0.01. ).; Data rounded at two digits.

Table 1b Descriptive statistics for values for the content of chemical elements (Mn–Zr) in the soil samples (0–30 cm), n¼ 134. Dis Mn Mo Na Nb Nd Ni P Pb Pd Pr Pt Rb Rh S Sb Sc Sm Sn Sr Ta Tb Te Th Tl Tm V W Y Yb Zn Zr

Log Log N Log N Log Log Log Log N Log N Log Log Log Log Log Log Log Log Log Log N Log Log Log Log Log Log Log Log

Unit

DL 1

mg kg mg kg  1 % mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1 mg kg  1

0.01 0.01 0.00005 0.1 0.01 0.01 10 0.01 0.1 0.01 0.01 0.1 10 10 0.01 0.1 0.1 0.01 0.01 0.1 0.01 10 0.1 0.1 10 0.01 0.01 0.01 0.01 0.01 0.01

X 570 1.1 0.62 7.4 11 58 820 150 0.39 2.9 0.19 55 12 1200 76 8.6 2.1 1.4 130 0.51 0.27 82 3.3 660 110 75 3.5 6.7 0.75 62 45

Xg 510 0.64 0.67 6.4 11 34 700 78 0.35 2.8 0.18 52 5.0 490 2.1 6.9 2.0 1.3 99 0.47 0.27 26 3.4 2.5 110 68 2.9 6.0 0.72 55 31

Md 480 0.35 0.49 6.3 9.4 39 720 87 0.32 2.4 0.19 47 9.2 570 3.4 6.6 1.8 1.2 93 0.43 0.24 30 0.83 5.2 97 69 2.7 5.8 0.66 55 32

Min 89 o 0.01 0.020 0.74 1.2 4.9 190 7.8 o 0.1 0.26 0.098 4.3 o 10 150 0.24 0.94 0.31 0.33 13 o 0.1 0.061 o 10 o 0.1 0.11 20 18 0.41 1.2 0.13 13 4.1

Max 1800 17 1.4 20 29 330 1800 1900 1.1 8.4 0.46 130 40 51000 3200 31 6.1 4.8 560 1.8 0.92 2400 9.6 20000 380 190 18 28 2.4 170 190

S 360 1.8 0.33 4.2 5.9 59 390 270 0.23 1.6 0.064 29 9.5 4600 400 6.5 1.1 0.66 120 0.28 0.14 240 3.2 2900 57 32 2.7 3.9 0.37 30 40

Sx 31 0.15 0.029 0.36 0.51 5.1 34 24 0.020 0.13 0.0055 2.5 0.82 400 34 0.56 0.093 0.057 10 0.024 0.012 21 0.27 250 4.9 2.8 0.23 0.33 0.032 2.6 3.4

CV 62 156 53 56 53 102 48 177 59 54 33 53 80 372 520 75 52 49 86 54 51 292 95 436 51 43 78 58 50 48 88

F (Soil) NS

0.91 1.17NS 0.00NS 0.01NS 0.00NS 0.05NS 0.72NS 4.49n 0.06NS 0.01NS 1.31NS 0.00NS 0.07NS 1.07NS 0.24NS 0.21NS 0.01NS 0.89NS 0.06NS 0.01NS 0.01NS 0.48NS 1.03NS 0.25NS 0.05NS 0.05NS 0.01NS 0.03NS 0.05NS 4.06n 0.01NS

F (Geo) 1.48NS 5.28n 7.20n 15.06n 9.92n 21.89n 8.99n 1.23NS 6.71n 10.18n 5.03n 21.29n 4.78n 9.37n 9.40n 7.82n 8.59n 5.91n 7.26n 13.86n 7.72n 7.57n 3.71n 33.61n 8.95n 2.86n 10.89n 9.87n 8.86n 2.37n 14.01n

Dis.—distribution; N—normal, Log—lognormal; DL—detection limit; X—mean; XG—geometric mean; Md–median; Min – minimum; Max–maximum; S—standard deviation; SX—standard error of mean;CV—coefficient of variation; F (soil)—F ratio between soil horizon 0–5 cm vs 20–30 cm (ANOVA); F ratio between geological units (ANOVA); NS—no significant; �—significance at p o 0.05; —significance at p o 0.01. ).; Data rounded at two digits.

82

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Fig. 4. Enrichment ratio Allchar area averages vs. European averages (Salminen et al., 2005) – 54 selected chemical elements.

Enrichment ratios between concentrations of chemical elements of studied area (Allchar) vs. European average (Salminen et al., 2005) are shown in Fig. 4 where one can conclude that concentrations of the most analyzed elements are within the range of European values. The highest enrichment ratios are indicative for Hg, As, Sb and particularly for Tl. 3.2. Association of chemical elements Statistical analysis was undertaken on the basis of the matrix of correlation coefficients and factor multivariate. Factor analysis (Table 2) was used to identify and characterize element associations. Twenty three variables (Br, Ca, Cd, Ce, Cs, Dy, Er, Ho, I, Mo, Na, Nb, Nd, P, Pb, Pr, Rh, Sm, Sr, Tb, Th, Tm and Yb) out of 62 analyzed variables (analyzed elements) were eliminated from further analysis either due to their low factor loading or low tendency to form independent factor (absence of reasonable correlation with other chemical elements). Elements with low share of communality or tendency to form independent factors were also excluded. Due to the high correlation coefficients between the group of lanthanides (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu), their standardized values have been joined into two new variables: light lanthanides (La, Ce, Pr, Nd, Sm and Eu) and heavy lanthanides (Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu) on the basis that high correlation coefficients of lanthanides cause a tightness of data and, consequently, disable calculation of factor analysis. The matrix of rotated factor loadings are presented and six factors were identified (Table 2): F1 (Rb, Ta, K, Nb, Ga, Sn, Ba, Bi, Li, Be, La-Eu, Hf, Zr, Zn, In, Pd, Ag, Pt, Mg); F2 (Tl, As, Sb, Hg); F3 (Te, S, Ag, Pt, Al, Sc, Gd-Lu, Y); F4 (Fe, Cu, V, Ge, Co, In); F5 (Pd, Zr, Hf, W, Be) and F6 (Ni, Mn, Co, Cr, Mg) associations, interpreted as Factors (F1–F6), which account for 82.7% of the total variability of treated elements (Figs. 5 and 6). The first group consists of the following elements Rb, Ta, K, Nb, Ga, Sn, Ba, Bi, Li, Be, La-Eu, Hf, Zr, Zn, In, Pd, Ag, Pt and Mg which are slightly affected by anthropogenic activities. The existence of the group is confirmed by factor analysis results (Table 2). Factor 1 is strongest factor, contains high values of mentioned chemical elements loadings, explaining c. 30% of total variability within the data. Such association of chemical elements is probably naturally distributed. The content of these elements mostly depends on the basic geological structure being reflected from the natural processes. Characteristic example is Ba that occurs as a natural marker in piroclastites and its tuffs (Bačeva et al., 2011; Stafilov et al., 2010b).

Table 2 Matrix of rotated factor loadings (n¼ 67).

Rb Ta K Nb Ga Sn Ba Bi Li Be La  Eu Hf Zr Zn In Sb Hg Tl As Ag Pt Te Gd  Lu Y Al S Sc Ge Fe V Cu Pd W Mn Cr Co Ni Mg Var

F1

F2

F3

F4

F5

F6

Comm

0.94 0.92 0.89 0.87 0.84 0.83 0.77 0.77 0.76 0.75 0.70 0.68 0.66 0.65 0.57 0.13  0.04  0.06  0.22 0.53 0.51 0.45 0.30 0.13 0.00  0.08  0.22 0.07 0.06 0.05  0.28 0.54 0.48 0.07  0.19  0.23  0.28  0.53 29.8

0.10  0.03  0.11  0.17  0.15 0.09 0.08  0.01  0.15 0.09 0.36 0.01  0.02 0.15  0.18 0.86 0.78 0.92 0.89  0.08  0.13 0.05 0.35 0.39 0.05 0.46 0.01  0.12  0.06  0.47  0.28 0.12 0.46 0.12  0.30  0.15  0.19 0.18 12.0

0.10 0.07 0.04 0.26  0.13 0.00 0.34 0.23  0.29 0.03  0.33 0.21 0.31  0.12  0.36 0.14  0.39  0.02 0.00 0.62 0.54 0.68  0.70  0.65  0.94 0.65  0.82 0.06  0.20 0.11 0.06  0.03 0.06 0.01  0.39  0.20  0.30  0.29 13.9

 0.12  0.20  0.14  0.08 0.43 0.08  0.02  0.11 0.20 0.05  0.18 0.01  0.02 0.40 0.54  0.20  0.17  0.20  0.09 0.21 0.15 0.04 0.00  0.03 0.04  0.06 0.28 0.67 0.84 0.76 0.79 0.04  0.23 0.43 0.17 0.56 0.03 0.00 10.6

0.13 0.16 0.03 0.22  0.08 0.09 0.14 0.21  0.09 0.53 0.37 0.59 0.60 0.07  0.16  0.10  0.01 0.09 0.23 0.23 0.16 0.03 0.35 0.37  0.02 0.20  0.14  0.02  0.22 0.12 0.17 0.76 0.56 0.14  0.34  0.12  0.11  0.10 7.8

 0.12  0.06  0.30  0.09  0.15 0.03  0.25  0.13 0.13 0.03 0.01  0.27  0.24 0.20 0.10 0.06  0.13  0.02  0.03  0.14  0.15 0.12 0.31 0.39 0.11  0.18 0.27 0.03 0.14 0.19 0.14  0.05  0.07 0.71 0.66 0.67 0.83 0.56 8.8

95.4 93.0 91.5 92.0 96.2 72.0 80.3 71.7 75.0 85.6 89.8 93.0 94.1 66.7 81.6 82.6 81.1 89.1 90.6 79.0 64.1 68.4 92.0 88.3 89.2 71.6 89.1 46.8 82.5 85.4 83.4 88.2 81.7 73.4 86.3 88.7 90.5 72.2 82.7

F1, F2 … F6—factor loading; Var—explorated variance (%); Comm—communality (%) Numbers in bold present factor loadings higher than 0.5.

The geographical distribution of Factor 1 association is shown in Fig. 5A. High values are typical for Triassic clastites rocks, PLQ quartz Latite Breccia (N and E of the study area), the low values for hydrothermally altered Triassic carbonates (central area), and the

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

83

Fig. 5. Spatial distribution of factor scores and its levels according to basic lithological units. A—Factor 1 scores (Rb, Ta, K, Nb, Ga, Sn, Ba, Bi, Li, Be, La - Eu, Hf, Zr, Zn, In, Pd, Ag, Pt and Mg); B—Factor 2 scores (Tl, As, Sb and Hg); C—Factor 3 scores (Te, S, Ag, Pt, Al, Sc, Gd–Lu and Y).

Jurassic diabase and serpentines (SW part). In addition, high values clustered around reefs and low values of Q alluvium indicate that Factor 1 is the distribution of clay fractions. The second geochemical association is indicated by the Factor 2 (Table 2), which associates high contents of Tl, As, Sb and Hg (c. 12% of total variability within the data). Association of Tl, As, Sb and Hg indicates mainly natural enrichment and also anthropogenic influence in the study area. This geochemical association is the most important for our research being well separated. Distribution map of the Factor 2 association (Fig. 5B) clearly points out the higher content of these elements deposited either on the area of hydrothermally altered stones or at close proximity of the mine. The observed association of elements was correlated to the former mining activities and appears as a result of past surface phenomena in the studied area (Frantz et al., 1994; Janković, 1993).

The highest value are related to hydrothermally altered Triassic carbonates (central area), the lower values Triassic carbonates and clastite (N part of the study area) and the central part of the Pliocene andesites. Very low values of F2 association are Triassic clastites outcrops (part E) and Jurassic diabase and serpentines (SW part). Topologically high values are clustered in the valleys and the slopes of the Q alluvium which implies that F2 distribution is somewhat coarse fraction. Third geochemical association is indicated by the second strongest Factor 3 (Table 2), which associates high contents of Te, S, Ag, Pt, Al, Sc, Gd–Lu and Y (c. 14% of total variability within the data). Distribution of F3 association is very interesting because it represents a “golden factor”. The content of Au was under the DL in most of the samples, despite the fact that the southern part of the deposit is characterized by dominance of gold mineraliza-

84

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Fig. 6. Spatial distribution of factor scores and its levels according to basic lithological units. A—Factor 4 scores (Fe, Cu, V, Ge, Co and In); B—Factor 5 scores (Pd, Zr, Hf, W and Be); C—Factor 6 scores (Ni, Mn, Co, Cr and Mg).

tion accompanied by variable amounts of antimony (Janković and Jelenković, 1994). It should be noted that the gold is not distributed uniformly in the ore. Its maximum contents (3–4 g t  1 on average and up to 20 g t  1 or higher in particular samples) were detected in the southern part of the deposit (Volkov et al., 2006). High values are related primarily to the Plio-Quaternary latite breccias (N and E of the study area) and lower at hydrothermally altered Triassic carbonates (central area) which indicate a small part of the distribution related to its hydrothermal activity (Fig. 5C). The forth group is comprised of Fe, Cu, V, Ge, Co and In. The association correlated typical “heavy metals” and the distribution of these elements is neither related to the geologic units in any topology nor due to the presence of high content of the heavy metals in the soil (Boev and Jelenković, 2012; Frantz et al., 1994). The only explanation for their distribution lies in the fact that some concentrated elements

are part of the marginal boundary of hydrothermal activity as similar to porphyric deposits (eg. Bor, Serbia). A grid of samples was sufficiently dense to encompass these changes (Fig. 6A). Association of Pd, Zr, Hf, W, Be (Factor 5—Table 2) comprises the geochemical group related to the area of Triassic hydrothermally altered carbonate outcrop (W part). The highest values were clustered on a ridge below the alluvium of the Majdanska River pointing out towards distribution of fine particles (Fig. 6B). Association of Ni, Mn, Co, Cr and Mg (Factor 6 – Table 2) is typical for a flysch (serpentine) outcrop. Similar relationships have been observed in soils formed on outcrops of certain formations in Slovenia and Croatia (Šajn, 2005), Bosnia and Herzegovina Alijagić and Šajn (2010) and Macedonia (Stafilov et al., 2008). Mostly, accumulations appear as moraine taking into account the glacier transport material from the higher areas of Jurassic serpentines.

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

85

Fig. 7. Spatial distribution of As, Tl and Sb in topsoil (0–30 cm) and concentration levels of As, Tl in Sb in topsoil and stream sediment according to determined zones.

Thus, the relatively high values of factor scores are found in Triassic carbonates (W part) most likely appear from the remains of moraines distribution of the material (Fig. 6C). The network was established on the basis of expected results and previous experience (Boev and Jelenković, 2012). The obtained results vary due to influence of the mineralization areas. The distribution was stabilized averaging the results from topsoil and subsoil samples. Certainly, the results were sufficiently stable providing statistical significant differences between the geological units and the defined zones. 3.3. Natural and anthropogenic behavior of As, Sb and Tl in investigated area The main interest of the work was focused on the behavior of As, Sb and Tl (Fig. 7) and on the investigation of the natural

enrichment and anthropogenic influence in the studied area due to (i) impact of the volcanic phenomena of the past, (ii) mineralization present in the mine, (iii) pedogenetic processes and (iv) former mining activities. However, if the data for the toxic elements in soil samples collected from the Allchar (AR) mining area (area of hydrothermally altered rocks, n ¼24) are compared with the data for the samples collected from the rest of the wider Allchar vicinity (Rest) (area of unchanged rocks, n ¼43), we can distinguish the differences in the content especially evident for the As, Sb and Tl (Fig. 3, Table 3). The content of As in the soil around the Allchar mine (AR) was a significantly high, which is 17 times higher compared with the Rest of the Allchar area, and 89 times higher compared to the European average (Table 3) (Salminen et al., 2005). Therefore, the median value of As in the soil around the Allchar mine is 3.8 times higher compared to values for the samples from the rest of the

86

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Allchar area. The average amount of 5 mg kg  1 of As in the world's soils (Rosler and Lange, 1972) is significantly lower than the corresponding value determined in soils of the Allchar locality. The differences between the average of As in soil samples around the Allchar mine was more than 100 times higher compared with the As

Table 3 Basic statistics and enrichment ratios of As, Tl and Sb regarding to sampling materials and selected groups of samples. Material

Zone

N

X

Min

Max

ER (EU)

ER (B)

As As As As As Tl Tl Tl Tl Tl Sb Sb Sb Sb Sb

Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil Topsoil

Europe Background Altered rocks Crven Dol Tailings Europe Background Altered rocks Crven Dol Tailings Europe Background Altered rocks Crven Dol Tailings

 43 24 5 2  43 24 5 2  43 24 5 2

12 62 1000 4100 1100 0.82 7.1 1800 8000 1800 1 5.2 21 46 2200

 13 20 250 810  0.2 1.4 350 1500  0.38 0.51 3.1 2000

 330 8300 8300 1500  190 17000 17000 2200  79 2400 99 2400

 5.4 89 350 99  8.6 2200 9700 2200  5 20 44 2100

  17 65 18   260 1100 260   4 8.9 430

As As As As Tl Tl Tl Tl Sb Sb Sb Sb

R. R. R. R. R. R. R. R. R. R. R. R.

Europe Zone 1 Zone 2 Zone 3 Europe Zone 1 Zone 2 Zone 3 Europe Zone 1 Zone 2 Zone 3

 3 3 3  3 3 3  3 3 3

10 15 330 45 0.48 0.78 17 2.3 1.1 0.53 83 10

 14 180 41  0.62 11 1.4  0.41 46 7.9

 15 420 51  0.87 25 3.3   100 12

 1.4 33 4.4  1.6 36 4.8  0.5 77 9.5

  23 3.1   22 2.9   160 19

Sediment Sediment Sediment Sediment Sediment Sediment Sediment Sediment Sediment Sediment Sediment Sediment

Europe—European average Salminen et al. (2005) for each sampling material; Background—Area of non-altered rocks; Altered rocks—Area of hydrothermally altered rocks; Crven Dol—Area above the As–Tl–Sb ore body Crven Dol; Tailings)— Group of samples, taken at tailings; Zone 1—River sediment collected above of the hydrothermally altered rocks (background); Zone 2—River sediment collected below the hydrothermally altered rocks; Zone 3—River sediment collected 10–18 km downstream of the hydrothermally altered rocks; ER (EU)—enticement ratios: European average vs. selected groups of samples for each sample materials; ER (B)—enticement ratios: local background vs. selected groups of samples for each sample materials. All data given in mg kg  1; Data rounded at two digits

soil content obtained from the study of “Geoinstitut” for 280 soil samples from Macedonia (average of 9.6 mg kg  1 with a range of 5– 120 mg kg  1, year 1999). High content of As determined in the most of the studied area of the Allchar locality, exceeded the New Dutch List (http://www.esdat.net/Environmental%20Standards/Dutch/ annexS_I2000Dutch%20Environmental%20Standards.pdf) optimum level of 29 mg kg  1 and action level of 55 mg kg  1 in soil. In addition, the spatial distribution of the aforementioned element in both soil horizons was highly dependent on lithology. As an example, the distribution of As with regard to basic lithological units as well as spatial distribution of As is provided in Fig. 7A. The highest concentrations were found at the Crven Dol area on Triassic altered carbonates, whereas the lowest concentrations are determined on Jurassic serpentines and Plio–Quaternary tuffs. The average amount of Sb in the soil for the entire study area is 76 mg kg  1, ranging from 0.24 to 3200 mg kg  1 (Table 1a,b). The average amount of Sb in the soil around the Allchar mine (AR) was significantly high—4 times higher compared with the (Rest) of the Allchar area, and 20 times above the European average (Table 3). The average amount of Sb in the world's and European topsoils is 0.2 mg kg  1 (Bowen, 1979) and 1.0 mg kg  1, respectively (Salminen et al., 2005). The obtained results for the study indicate that the entire investigated Allchar area has 380 times higher content of Sb compared with the world's average and 76 times higher from the European average (Salminen et al., 2005). The content for Sb in the soil from the Allchar area (tailings – present from the former mining activities) shows 430 times higher values compared with the (Rest) of the Allchar area. The spatial distribution of Sb is strongly dependent on lithology and is very similar to arsenic spatial distribution. The distribution of Sb is highest in Triassic altered carbonates and lowest in Plio–Quaternary tuffs (Fig. 7C). The comparison made with data (Table 3) for Tl in the soil around the Allchar area obtained from hydrothermally altered rocks with the (Rest) of the Allchar area has shown significantly high enrichment ratio, exceeding 260. The enrichment ratio for Tl in the soil at the Crven Dol area was even 1100 times higher compared with the (Rest) of the Allchar area (Table 3) and more than 9700 times higher compared with the European average (Salminen et al., 2005). Similarly to the distribution of the As and Sb, the higher contents of Tl were found in Crven Dol area on Triassic altered carbonates, whereas lower content was found in Quaternary moraines, Pl Andesite and also in Triassic clastites. The lowest content of Tl was observed on Plio-Quaternary latite breccias and Plio-Quaternary tuffs (Fig. 7B).

Fig. 8. Polynomial regression As vs. Tl (left) and As vs. Sb (right) in topsoil (0–30 cm).

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Natural deposits rich with As, Sb and Tl are scarce and, therefore, it seemed important to carry out Polynomial regression between the content of elements, with special concern on the ratio of (As/Tl) and (As/Sb). The spatial distribution of As, Sb and Tl (Fig. 7) was very similar compared to each other element. The highest determined average concentrations of As (0.41%) and Tl (0.8%), was related to the area of outcrop of the ore body Crven Dol (Table 3). These concentrations are probably the highest established natural concentration in soil worldwide and thousands times exceed the European average (Table 3) (Ivanov, 1986; Janković, 1988). The average concentrations of As (0.10%) and Tl (0.18%) were also relatively high in the area of hydrothermally altered rocks and rapidly decreased with the distance. In the area of unaltered rock (background) the average values of As (62 mg kg  1) and Tl (7.1 mg kg  1) are 5 to 8 times higher than the European average. Polynomial regression of As–Tl contents was significantly high (r ¼ 0.92), which leads to the idea that their concentration level is affected by the unique process (single phase) of hidrotermal action (Fig. 8). The exeptions are areas outside hidrotermal action (background) (As/Tl ¼26.0) and where the concentrations of Tl is much higher than As. This effects confirm the theory of Boev and Jelenković (Boev and Jelenković, 2012; Jelenković and Boev, 2011; Jelenković et al., 2011), the research of separated zones and enriched As, Sb and Tl ore deposits. Distribution of Sb in the soil varies from As and Tl distribution. Enrichment of Sb was primarily related to the area along the Majdanska River where the abandoned Sb mine dating from the Ottoman Empire was exploited (Fig. 7). The average content of Sb in two selected samples from this area were extremely high (0.21%) and exceeded the European average for more than 400 times (Table 3). The area with the higher values was limited and the average ratio between (As/Sb) was very low, meaning that the concentration of Sb excess the concentration of As. Concentrations of Sb did not follow the areas of extremely high concentrations of As and Tl (Crven Dol area and the area of hydrothermally altered rocks). Thus, the amount of 46 mg kg  1 was found in the Crven Dol area and 21 mg kg  1 at the area of hydrothermally altered rocks. The average value of Sb in the (background) area of unaltered rock is found to be similar to the As and Tl (5.2 mg kg  1), about 5 times surpassing the European average (Table 3). The polynomial regression between As and Sb was very interesting (Fig. 8). Two sample groups were singled out: the first group with the As/Sb 410 (r ¼ 0.86) and the other group with the content ratio As/Sb o10; (r ¼ 0.96). The first samples group represents the area where the content of Sb correlates the distribution of As and Tl. The second group comprises samples collected at wider area or natural enrichment of Sb ore-shoots, which can not be distinguished from anthropogenic influences. It confirms that Sb distribution affected several stages of hydrotemal action as such conclusion was previously postulated (Boev and Jelenković, 2012; Jelenković and Boev, 2011; Jelenković et al., 2011). 3.4. X-ray powder diffraction (XRPD) analysis X-ray powder diffraction (XRPD) analysis was conducted on three selected samples as an attempt to detect the Sb, As and Tlminerals in to correlate the results with the AES-ICP measurements. The samples denoted A-24 (20–30), A-29 (0–5) and A-36 (20–30), in which the content of As, Sb and Tl was determined by AES-ICP (As—7551, 2109 and 23270 mg kg-1, respectively, Sb—1926, 7491 and 262 mg kg-1, respectively, Tl—1431, 390 and 4306 mg kg-1, respectively) were analyzed. As seen (Fig. 9), the XRPD enabled identification of mineral arsenolite (As2O3) in A-24 sample, where the content of As, expressed as As2O3, reaches 1%. The presence of the latter mineral (Table S1) as minor constituent in the sample was

87

Fig. 9. X-ray powder diffraction patterns of the analyzed samples. To avoid complexity, only the sole or combination peaks arising from the detected As(arsenolite), Sb- (stibnite, manganostibnite, valentinite) and Tl-minerals (lorandite) are denoted. Abbreviations: ars—arsenolite, dol—dolomite, lor—lorandite, mns— manganostibnite, mus—muscovite, qua—quartz, stb—stibnite, val—valentinite.

manifested by the appearance of four peaks observed at very similar 2theta positions (strongest peaks in arsenolite occur at 2theta: 13.82671 (44), 27.86331 (100), 35.2752 1 (34), 42.3431 (12)).The automatic peak indexing pointed out that major constituents in the sample are dolomite and gypsum (widely distributed around Allchar) as well as quartz, SiO2 and muscovite-3T, KAl2(AlSi3O10)(F,OH)2 (Fig. 9, top plot, and Table S1). Despite the presence of dolomite and quartz as major components in sample A-29, muscovite, valentinite (Sb2O3) and manganostibnite (Mn7SbAsO12) were also characterized (Fig. 9, middle plot). The appearance of latter two minerals nicely complements the obtained Sb content in the sample (0.75%). Having in mind relatively low content of arsenic and thallium in the sample, no corresponding mineral phases were found in the diffraction pattern. The strongest peaks of the obtained Sb-containing minerals (manganostibnite—2θ: 17.8005 (31), 27.8439 (6), 29.4445 (18), 33.7099 (100), 35.1738 (19), 39.4904 (12) and valentinite: 2θ: 19.4536 (17), 28.4299 (100), 58.8827 (10)) coincide very well with the obtained results (Table S2). In the sample denoted A-36 where highest content of arsenic and thallium was found (see above determined contents), XRPD indicated presence of corresponding arsenolite and lorandite

88

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Fig. 10. Distribution of As, Tl in Sb in stream sediment according to determined zones.

mineral (Fig. 9, bottom plot, Table S3). In addition, stibnite (Sb2S3), as one of the most abundant minerals in Allchar, was also registered. 3.5. Transport of As, Tl and Sb from investigated area and possible environmental impact Due to the hazardous nature of these minerals to public health, particular attention was dedicated to their transport and subsequent accumulation in the important agricultural lower basin of the Crna River. River sediments were divided into three groups: Zone 1—above of the hydrothermally altered rocks (background); Zone 2—below the hydrothermally altered rocks and Zone 3—downstream 10–18 km of the hydrothermally altered rocks (Figs. 7 and 10). As reported in the previous section, here, a special attention to the ratio between the elements As/Tl and As/Sb in river sediments has also been devoted. The average content of As, Sb and Tl in Zone 1 was comparable to the European average in river sediments (Table 3), whereas in the zone 2 (hydrothermally altered rocks impact), As and Tl were enriched more than 20 times in the lower part of the river (Zone 3) and about 3 times compared with the Zone 1 (background). The behavior of Sb was found to be somewhat different. Namely, Sb was enriched 160 times in Zone 2, at downstream about 20 times (Table 3), and generally one can establish Sb distribution with significant anthropogenic impact. Transport of river sediments is very important because, volcanic activity with intermittent eruptions of tuffs was followed by erosion of material lying above the Allchar ore deposit (Kolios et al., 1980) and at lower parts of the Crna and Vardar Rivers basins, a huge amount of material rich with As, Sb and Tl was transferred (Stafilov et al., 2013). In this small study area (c. 13 km2), except the village Majdan that is mostly uninhabited, the extremely high As, Sb and Tl concentrations are of no particular impact. The settlement Majdan is on the edge of hydrothermal activity and although the obtained concentrations in this area were also high (74 mg kg  1 As, 3.3 mg kg  1 Tl and 1.9 mg kg  1 Sb—see Location A-28) and exceed norms (New Dutch List) for soil, the area lies on high mountain and lacks agriculture activities. The bigger problem poses cattle grazing in the Crven Dol (Table 3), which is covered with grasslands. Anthropogenic influence in the studied Allchar area was limited and almost has no impact on the distribution of As, Sb and Tl. The distribution of these elements is related almost exclusively to decomposition of mineralized rocks and later pedogenetic processes. The mining was limited and linked to sole exploitation of Sb showing slight tendency of anthropogenic

influence (tailings), which later was reflected in the Sb distribution in river sediments of Majdanska River. The most important problem lies in the river transport of sediments enriched with As, Sb, and Tl. Those sediments are transported from the Majdanska River to the Crna River, and deposited in its alluvial sediment, on area of intensive agriculture activities. In our previous work (Stafilov et al., 2008, 2010a) it has been determined that the samples collected on Holocene alluvium of the Crna River contain high concentrations of the As, Sb, and Tl. Their average enrichment ratios exceed the average of the total investigated area by 4 to 4.5 times. Both in the topsoil and subsoil, a clear anomaly in the area of Holocene alluvium of the rivers Crna and Vardar was observed. Namely, the highest content of elements are found on Holocene alluvium of the Crna Reka (topsoil: 32 mg kg  1 As, 4.8 mg kg  1 Sb, 1.4 mg kg  1 Tl; subsoil: 30 mg kg  1 As, 4.2 mg kg  1 Sb, 1.4 mg kg  1 Tl).

4. Conclusion The activities carried out in As–Sb–Tl Allchar mine bring over natural increased content of certain toxic elements into environment, which was determined through the results from the soil survey and river sediments from the Allchar locality, Republic of Macedonia. Enforcement of statistical multivariate analysis, concretely principal component factor analysis, revealed that the geochemical association of the arsenic, antimony and thallium indicates mainly natural enrichment, and several high-values for Sb were explained as a result of the former Ottoman Empire mining activity in the study area. Anthropogenic distribution of Sb is reflected in the river sediments as well. The content of Tl determined in the Allchar locality probably reflects the highest established natural concentration in soil worldwide. In zone 2, As and Tl are enriched with a more than 20 times whereas their contents decrease in the lower part of the river (Zone 3) about 3 times compared to the Zone 1 (background). It is important to point out that Sb was enriched 160 times in Zone 2 (downstream about 20 times) and, therefore, Sb distribution involves significant hazardous anthropogenic impact.

Acknowledgments This study was funded jointly by the Institute of Chemistry, Faculty of Science, Skopje, Macedonia, Slovenian Research Agency (ARRS)–Program Groundwater and Geochemistry and Research Institute for Analytical Instrumentation, Cluj-Napoca, Romania.

K. Bačeva et al. / Environmental Research 133 (2014) 77–89

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.envres.2014.03.045. References Alloway, B.J., 1995. Heavy metals in soils. Blackie Academic, London. Alvarenga, P.M., Araújo, M.F., Silva, J.A.L., 2004. Elemental uptake and root-leaves transfer in Cistus ladanifer L. growing in a contaminated pyrite mining area (Aljustrel–Portugal). Water. Air. Soil. Pollut. 152, 81–96. Alijagić, J., Šajn, R., 2010. Distribution of chemical elements in an old metallurgical area, Zenica (Bosnia and Herzegovina). Geoderma 162, 71–85. Bačeva, K., Stafilov, T., Šajn, R., Tǎnǎaselia, C., 2012. Moss biomonitoring of air pollution with heavy metals in the vicinity of a ferronickel smelter plant. J. Environ. Sci. Health Part A 47, 645–656. Bačeva, K., Stafilov, T., Šajn, R., Tănăselia, C., Ilić Popov, S., 2011. Distribution of chemical elements in attic dust in the vicinity of ferronickel smelter plant. Fres. Environ. Bull. 20, 2306–2314. Balabanova, B., Stafilov, T., Bačeva, K., Šajn, R., 2010. Biomonitoring of atmospheric pollution with heavy metals in the copper mine vicinity located near Radoviš, Republic of Macedonia. J. Environ. Sci. Health Part A 45, 1504–1518. Boev, B., Bermanec, V., Serafimovski, T., Lepitkova, S., Mikulcic, S., Soufek, M., Jovanovski, G., Stafilov, T., Najdoski, M., 2001-2002. Allchar mineral assemblage. Geol. Maced. 15/16, 1–23. Boev, B., Jelenković, R., 2012. Allchar deposit in Republic of Macedonia—petrology and age determination. In: Juboury, A.I. (Ed.), Petrology—New Perspectives and Applications. InTech, Croati, Rijeka, pp. 131–168. Bowen, H.J. M., 1979. Environmental Chemistry of the Elements. Academic Press, NewYork. Cohen, J., 1988. Statistical power analysis for the behavioural sciences. Lawrence Erlbaum Associates Publishers, New York. Concas, A., Patteri, C., Cincotti, A., Cao, G., 2004. Metal contamination from abandoned mining sites: experimental investigation on possible remediation techniques. Land. Contam. Reclam. 12, 9–20. Davis, J.C., 1986. Statistic and Data Analysis in Geology. Willey in Sons, New York. Frantz, E., Palme, H., Todt, W., El Goresy, A., Pavićević, M.K., 1994. Geochemistry of Tl–As minerals and host rocks at Allchar (Macedonia).Solar neutrino detection with TI-205 the “LOREX” Project; geology, mineralogy and geochemistry of the Allchar deposit locality Crven Dol. Neues Jahrb. Miner. Abh 167, 359–399. Garson, D., 2000. Factor Analysis. North Carolina State University, Raleigh. Ivanov, T., 1986. Allchar the richest ore deposit of Tl in the world. Feasibility Sol. Neutr. Detect. GSI-86-9, Darmstad, 6. Janković, S., 1988. The Allchar Tl–As–Sb deposit, Yugoslavia and its specific metallogenic features. Nucl. Instrum Methods Phys. Res. A 271, 286. Janković, S., 1993. Metallogenic features of the Alshar epithermal Sb-As-Tl deposit (the Serbo-Macedonian metallogenic province). Neues Jahrb. Miner. Abh 166, 25–41. Janković, S., Jelenković, R., 1994. Thallium mineralozation in the Allchar Sb-As–Tl– Au depoit. Neues Jahrb. Miner. Abh 167, 283–297. Jelenković, R., Boev, B., 2011. Vertical mineralization interval and forecast of the position of an ore-body in the Alšar Sb–As–Tl deposit, Macedonia. Geol. An. Balkan. Pol. 72, 119–129. Jelenković, R., Boev, B., Pavićević, M., 2011. Određivanje primarne dubine rudnih tela Sb–As–Tl ležišta Alšar (Republika Makedonija). Tehnika 66, 399–405. Jovanovski, G., Boev, B., Makreski, P., 2012. Minerals from the Republic of Macedonia with an introduction to mineralogy. Macedonian Academy of Sciences and Arts, Skopje. Kolios, N., Innocenti, F., Maneti, P., Peccerillo, A., Guliano, O., 1980. The pliocene vulcanisam of the Voras Mt. (Central Macedonia, Greece). Bull. Volc. 43, 3.

89

Lee, C.G., Chon, H.T., Jung, M.C., 2001. Heavy metal contamination in the vicinity of the Daduk Au–Ag–Pb–Zn mine in Korea. Appl. Geochem., 16. Merrington, G., Alloway, B.J., 1994. The transfer and fate of Cd, Cu, Pb and Zn from two historic metalliferous mine sites in the UK. Appl. Geochem. 9, 67–77. Navarro, M.C., Pérez-Sirvent, C., Martínez-Sánchez, M.J., Vidal, J., Tovar, P.J., Bech, J., 2008. Abandoned mine sites as a source of contamination by heavy metals: a case study in a semi-arid zone. J. Geochem. Explor. 96, 183–193. Percival, T.J., Radtke, A.S., Jancović, S.R., et al., 1990. Gold mineralization of the Calin type in the Alsar district, Macedonia, Yugoslavia. Eight Quadren. IAGOD, Ottawa. Rieck, B., 1993. Allchar, Macedonia. Miner. Rec. 24, 437–448. Rosler, H.J., Lange, H., 1972. Geochemical Tables. Elsevier, Amsterdam. Theocharopoulos, S.P., Wagner, G., Sprengart, J., Mohr, M.E., Desalues, A., Muntau, H., Christou, M., Quevauviller, P., 2001. European soil sampling guidelines for soil pollution studies. Sci. Total Environ. 264, 51–62. Šajn, R., 2005. Using attic dust and soil for the separation of anthropogenic and geogenic elemental distributions in an old metallurgic area (Celje, Slovenia). Geochem. Explor. Environ. A 5, 59–67. Šajn, R., 2006. Factor analysis of soil and attic-dust to separate mining and metallurgy influence, Meža Valley, Slovenia. Mat. Geol. 38, 735–747. Salminen, R., Batista, M.J., Bidovec, M., Demetriades, A., De Vivo, B., De Vos, W., Duris, M., Gilucis, A., Gregorauskiene, V., Halamic, J., Heitzmann, P., Jordan, G., Klaver, G., Klein, P., Lis, J., Locutura, J., Marsina, K., Mazreku, A., O'Connor, P.J., Olsson, S.Å., Ottesen, R.T., Petersell, V., Plant, J.A., Reeder, S., Salpeteur, I., Sandströ m, H., Siewers, U., Steenfelt, A., Tarvainen, T., 2005. Geochemical Atlas of Europe. Geolog. Surv. Finland. Siegel, F.R., 2002. Environmental Geochemistry of Potentialy Toxic Metals. Springer, Verlag Berlin, Heidelberg. Stafilov, T., Šajn, R., Alijagić, J., 2013. Distribution of Arsenic, antimony and thallium in soil in Kavadarci and the environs. Republic of Macedonia. Soil. Sediment. Contam. 22, 105–118. Stafilov, T., Šajn, R., Boev, B., Cvetković, J., Mukaetov, D., Andreevski, M., 2008. Geochemical Atlas of Kavadarci and the Environs. Faculty of Science, Sts Cyril and Methodius University, Skopje. Stafilov, T., Šajn, R., Boev, B., Cvetković, J., Mukaetov, D., Andreevski, M., Lepitkova, S., 2010b. Distribution of some elements in surface soil over the Kavadarci Region. Republic of Macedonia. Environ. Earth Sci. 61, 1515–1530. Stafilov, T., Šajn, R., Pančevski, Z., Boev, B., Frontasyeva, M.V., Strelkova, L.P., 2010a. Heavy metal contamination of surface soils around a lead and zinc smelter in the Republic of Macedonia. J. Hazard. Mater. 175, 896–914. Volkov, A.V., Serafimovski, T., Koehneva, N.T., Tomson, I.N., Tasev, G., 2006. The Allshar epithermal Au–As–Sb–Tl deposit Southern Macedonia. Geol. Ore. Deposit 48, 205–224. Zhang, C., Selinus, O., Schedin, J., 1998. Statistical analyses for heavy metal contents in till and root samples in an area of southern Sweden. Sci. Total Environ. 212, 217–232.

Web references Autodesk, Inc. 2012. Autodesk MAP 3D, Version 2012 - Software. Autodesk, Inc. 〈http://www.autodesk.com〉. ESRI, Inc. 2004. ArcINFO ver 9, Software. Environmental Research Institute http:// www.esri.com/. Golden Software, Inc. 2012. Surfer (surface mapping system), ver 11 – Software. Golden Software, Inc.http://www.goldensoftware.com/. Stat Soft, Inc. 2012. STATISTICA (data analysis software system), version 11—Software. Stat Soft, Inc.〈www.statsoft.com〉.