See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222079407 Estimations of Clarkes for Carbonaceous biolithes: World averages for trace element contents in black shales and coals Article in International Journal of Coal Geology · April 2009 DOI: 10.1016/j.coal.2009.01.002 CITATIONS READS 573 1,438 2 authors, including: Yakov. Elievich Yudovich Komi Scientific Center 46 PUBLICATIONS 1,381 CITATIONS SEE PROFILE Some of the authors of this publication are also working on these related projects: Geochemistry of Titanum View project Lithochemistry basis View project All content following this page was uploaded by Yakov. Elievich Yudovich on 29 June 2018. The user has requested enhancement of the downloaded file. This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy International Journal of Coal Geology 78 (2009) 135–148 Contents lists available at ScienceDirect International Journal of Coal Geology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j c o a l g e o Estimations of Clarkes for Carbonaceous biolithes: World averages for trace element contents in black shales and coals M.P. Ketris, Ya.E. Yudovich ⁎ Institute of Geology, Komi Scientific Center, Ural Branch of the Russian Academy of Sciences, 167023 Syktyvkar, Morozova st., 100, ap. 49, Russia a r t i c l e i n f o Article history: Received 27 October 2008 Received in revised form 1 January 2009 Accepted 6 January 2009 Available online 14 January 2009 Keywords: Metalliferous black shales Coal Geochemistry Trace elements World average contents (black shale and coal Clarke values) a b s t r a c t Black shale and coal Clarke values are the average trace element contents in the World black shales and coals. These calculations are made in Russian geochemistry but up to now are poorly known in the West. Modern tables of black shale and coal Clarkes are presented, based on comprehensive calculations using very large amount of information (thousands analyses of black shales, coals, and coal ashes for trace elements). In black shale geochemistry, three figures were calculated for each main lithologies: terrigenous (+tuff), chert, and carbonate. Two Clarke estimations are presented, named “lithological” (K1) and “lithostratigraphical” (K2). In coal geochemistry, seven figures were calculated for each trace element: average content in hard coals and their ashes; average content in brown coals and their ashes; average content in all coals and their ashes; and coal affinity index (or “coalphile index”) = average content in all ashes/Clarke values of sedimentary rocks. The black shale and coal Clarkes presented here provide an important scientific base for many geochemical comparisons and issues. © 2009 Elsevier B.V. All rights reserved. 1. Introduction 1.2. Some history In 1923, the famous Russian geochemist A.E.Fersman1 introduced the term “clark” (= Clarke, in English), in honor of the prominent American scientist, one of the founders of geochemistry, F.W.Clarke (who worked many years as Chief Chemist at U.S. Geol. Survey), who first calculated average composition of various rocks, and later, the Earth's crust. In black shale geochemistry, a well-known estimation of average contents of trace elements in black shales was made by Vine and Tourtelot (1970). Their calculations, using 20 U.S. Phanerozoic black shale units, have been widely cited (by many researchers) because there were no other estimations. But, it is desirable to use more comprehensive world-wide data for more accurate conclusions. Our calculations are based on tens of thousands of analyses employing hundreds of sample mean values. Statistical processing of the data supported new Clarke estimations which seem to be more plausible. In coal geochemistry, there are known some attempts to calculate coal Clarkes for several trace elements. The first calculations were made by the Norwegian-German geochemist, another founder of geochemistry3, Goldschmidt (1935), at start of the 1930s. Goldschmidt organized systematical analyses of some European coals for trace elements, using modern (for that time) analytical methods – emission spectrographic and X-ray fluorescence analyses. Goldschmidt's estimations of trace element average contents for “ordinary” (common) and “enriched” coal ashes were, for many consequent years, the “compass” for geochemists dealing with coal. Later, we find some averages in studies by Krauskopf (1955) and Bethell (1962). The first wide (including most trace elements and most 1.1. Meaning of the Russian term Clarke Fersman defined “Clarke” as the average content of given chemical element in the Earth's crust and also in the hydrosphere. This term very soon took root in Russian literature, but, up to now, is nearly unknown to Western researchers2. With time, however, a sense of the term was strongly extended: many other geochemical averages were named as “Clarkes”, for example, “Clarkes of granites”, “Clarkes of basalts”, “Clarkes of sedimentary rocks”, etc., including “coal Clarkes”, i.e. average trace element contents in the World coals. For example, coal Clarke of Ge is 3.0 ± 0.3 ppm (hard coals) and 2.0 ± 0.2 ppm (brown coals) (Yudovich and Ketris, 2004). ⁎ Corresponding author. Tel.: +7 821 31 19 24. E-mail addresses: yudovich@geo.komisc.ru, EYuYa@yandex.ru (Y.E. Yudovich). 1 Fersman was the first, who gave a lecture course “Geochemistry” (Moscow, 1912). Later, he created many fundamental proceedings on geochemistry, for example, “Pegmatites” (1931), “Geochemistry” (4 volumes, 1933–1939), etc. 2 That is why, in our articles in As or Hg in coal (Yudovich and Ketris, 2005a,b), the terms like “coal Clarke of As” need obligatory the special explanation for the Western reader. 0166-5162/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.coal.2009.01.002 3 Two other founders of geochemistry, new science of 20th century, were Russian scientists, V.I.Vernadski and his follower, A.E.Fersman. Author's personal copy 136 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Fig. 1. Structure of the scientific field in black shales geochemistry. The areas with neighboring overlaps are shaded (Yudovich and Ketris, 1994, p. 4; 1997, p. 5). World coals) calculation of coal Clarkes was performed in the USSR (Yudovich et al., 1972). Later, Valkovič's (1983a,b) estimations (based on U.S. figures, but including some others) were developed. Valkovič (1983a,b) used frequency histograms for statistical evaluations of averages. It is a very useful working procedure, and it was independently widely used in our new calculations (Yudovich et al., 1985). The estimations based on thousands of analyses, were made as four figures for each trace element: (1, 2) for hard coals and their ashes, and (3,4) for brown coals and their ashes. These new coal Clarke figures (with detailed explanation of stepped calculation procedure), were cited in Russian, East European, Indian, and Chinese studies for many years. Unfortunately, it was in a Russian-language book, and for this reason, our figures remained practically unknown to the most Western researchers. Swaine (1990) published his excellent outline on trace elements in coal. He tried to embrace all the World coals, but his outline had large omissions of the East (Soviet) Block coals. It is of note, Swaine's world estimation were, as a rule, rough interval values; for example, as 10–40 ppm (and not, say, as 16 ± 5 ppm, where ±5 ppm means statistical standard deviation). In last decades of 20th century, great changes appeared in coal geochemistry due to environmental problems caused by wider consumption of coal in electric power plants. This enlarged demand for coal created abundant new geochemical studies, with analyses of coal on toxic elements, such as Be, Hg, Cd, Pb, As, Sb, Se, Cr, Mn, U, and some others. In addition, coal mining sharply increased in developing countries, such as Spain, Turkey, Greece, China, India, Brazil, Peru, and Nigeria, among others. These coals have been analyzed using wide international cooperation, for example, with U.S. Geological Survey, using modern analytical equipment. Therefore, the amount of coal-trace-element figures has greatly increased. The U.S. Geological Survey created their “Coal Quality (COALQUAL) Database” (Bragg et al., 1998), with more than 13,000 analyses for most trace elements. Finally, in recent decades, due to mostly Russian researchers (V.V. Seredin, S.I. Arbuzov and some others), the old problem became again actual: coal as industrial resource of trace elements (Sc, Ge, REE, Nb, Ta, Re, Au, and platinum group elements, PGE) (Arbuzov et al., 2000, 2003; Zharov et al., 1996; Seredin and Shpirt, 1995; Seredin and Finkelman, 2008). All the latter work necessitates new coal Clarkes calculation. 2. Estimation of trace elements Clarkes for black shales The so called “black shales” (sedimentary rocks enriched in Corg) are of great interest to geochemists, lithologists, paleontologists, mineralogists, ore prospectors, and oil geologists, chemists, and technologists in the field of oil shale industry (Fig. 1). For a geochemist, black shales are anomalous rocks enriched in P, U, Mo, V, Re, Se, Zn, Hg, and some other trace elements. For a lithologist, they are unusual sedimentary rocks, mixtures of organic Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 and inorganic matter at varying ratios. The inorganic matter may be of almost any composition: clay, silt, sand, carbonate, chert, phosphate, tuff, and may be deposited in nearly any facies. For a stratigrapher and paleontologist, in general, black shales present special stages in the history of sedimentary shell. For an ore prospector, black shale (or oil shale) species are the most likely oil-productive source beds. Finally, so called oil shales are also black shales containing “shale oil”, which can be extracted by means of pyrolysis. World “oil shale” reserves exceed petroleum reserves. All the topics above mentioned account for abundant references on black shale (and oil shale) geochemistry and geology. Many scientific congresses, symposiums, and workshops took place on these problems. For example, a symposium “Geochemistry, Mineralogy and Lithology of Black Shales” was held in 1987 in Syktyvkar (Russia) under the direction of the second author. Of special note was the UNESCO International Geological Correlation Program Project 254, entitled “Metalliferous Black Shales”. As a result, two of our Russian monographs were published (Yudovich and Ketris, 1988, 1994), and also short English outline with huge bibliography (Yudovich and Ketris, 1997). In the 1994-monograph, Clarke estimations for more than 60 trace elements in black shales were first calculated. Such calculations are difficult because of lithological diversity of black shales, among other reasons. 137 2.1. An “ideal” information file Such a file should include answers to the following questions: - where were the samples taken from? what is the correct name of the rock? what is the correct stratigraphic position of the rock samples? what are the analytical errors and accuracy? how many rock samples were analyzed? Such informational “ideal” is hardly achievable in most cases; as a rule, some information is not available in literature. As an example, information about Corg content was often lacking, although it is crucial for geochemical conclusions. As there are no Corg data, rough Corg estimation may be made using chemical analysis data (Yudovich and Ketris, 1988, p. 29). 2.2. Calculation method All calculation procedures may be divided into several stages (Yudovich and Ketris, 1994): 1) formation of so called “primary statistical samples”; 2) calculation of common median and its standard deviation; 3) estimation of geochemical background and geochemical anomalies; Fig. 2. Common median calculation (± standard deviation, shaded column): the example is for uranium in black shales (Yudovich and Ketris, 1994, p. 44; 1997, p. 31). Each point represents one “primary statistical sample”. Q1 and Q2 are first and third quartilies of frequency distribution. Black shale lithologies: 1 – carbonate, 2 – clay, 3 – silt and sand, 4 – chert, 5 – tuff, 6 – phosophate, 7 – shale of unclear lithology. Author's personal copy 138 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 4) formation of so-called “second statistical samples”; 5) “lithological Clarke K1” estimation; 6) “lithostratigraphic Clarke K2” estimation. 2.2.1. “Primary statistical samples” and common median calculations As a base (the “elementary unit”) for calculations, “a primary sample” was used. The primary sample is a group of analyses of lithologies representing to certain stratigraphic unit of some region. For example: “Lower Carboniferous (=Mississippian) black cherts of the Pai-Khoi”. We dealt with two different situations. First: primary samples were already available from the literature as average values. Such samples were sometimes heterogenous; for instance, “carbonates + claystones”, although such rocks are known to widely differ in trace elements content. Further, some data recorded in Soviet references were average values for entire sedimentary sequences, the so called “carbonaceous formations” (Sozinov et al., 1988).4 The term itself is not clearly defined and the averages cannot geochemical applications. Second: primary statistical samples were not found in the literature, but analyses were available for statistical processing. Recalling the above mentioned information about geological age, lithologies etc., the primary sample averages may be obtained. Given considerable data dispersion, we applied mean median calculations, while at small dispersions simple arithmetic means were calculated. Then “common medians” were graphically estimated. Each point of such graph represents one primary sample mean. The points are plotted on horizontal lines representing a certain stratigraphic level each (Fig. 2). The following step is a common median standard deviation calculation: pffiffiffiffiffi δMe = ðQ3 − Q1 Þ=2 N; ð1Þ where: - N is number of points (=primary samples), - Q1 and Q3 are the first and the third quartiles of frequency distribution. The median was tabulated in the form of Me ± δMe. The standard of the primary frequency distribution, δx, was calculated as: δx = Q3 − Me ð2Þ The geochemical background value (GB) is a frequency distribution part limited by the first and the third quartiles: GB = Q3 − Q1 ð3Þ Hence, 50% of primary sample means constitute a geochemical background. Positive geochemical anomalies (A) are determined by adding standard deviations to the common median: A0 = Q3 = Me + 1δx ðqanomalyqÞ A1 = Q3 + 1δx = Me + 2δx ðqstrong anomalyqÞ A2 = Q3 + 2δx = Me + 3δxðqvery strong anomalyqÞ ð4Þ ð5Þ ð6Þ Common median and quartiles values allow estimation of “true” element distribution in black shales. The median is particularly con- venient because its value is little influenced by individual extreme values, when a data population is large enough. It is clear, however, that common median cannot be used for Clarke estimations because it is dependent on the available set of data. Some black shales have been studied in detail and described in numerous geochemical publications. Others, on the contrary, are poorly known, hence they are represented only as few points (=primary samples means) on the graphs. For this reason, the common median value is more affected by better studied black shales than by less studied ones. A good example is Goldschmidt's reports on average values for West Europian coals. They had been cited for a long time as Clarke values for coals worlwide. But, as was subsequently shown, the average values proved inflated since geochemistry of West European coals is sharply anomalous (Yudovich et al., 1985). 2.2.2. “Second statistical samples” and “lithostratigraphic subclarkes” The following calculation step is averaging primary sample means belonging to a given age, region, and lithology. Several primary samples thus grouped make up a “secondary sample”, giving an opportunity to calculate “lithological subclarkes”. Unfortunately, scantiness of lithological information, typical for many geochemical reports, brought us to employ it in a very rough form: lithologies were restricted to the following three: carbonates, cherts, and terrigenous rocks, even without classifying the latter into clastic and clayey ones. The rocks of volcanic-sedimentary origin were grouped within one of the three types, more, however, with terrigenous rocks. Phosphate rocks needed to be distinguished for some elements (P, Sr, U). “Lithological subclarkes” were calculated for the following stratigraphic levels: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Recent + Holocene Paleogene + Neogene Cretaceous Jurassic Triassic + Permian Carboniferous (Mississippian + Pennsylvanian) Devonian Silurian Ordovician Cambrian Vendian Riphean = Upper Precambrian Karelian (Aphebian) = Middle Precambrian Upper Archean and 15. Lower Archean = Lower Precambrian These 15 units constitute four larger stratons: Phanerozoan (including Vendian, or “Eocambrian” according to Salop, 1982), Upper, Middle, and Lower Precambrian. It is necessary to average the means of the litologies (=second statistical samples) in order to obtain “lithostratigraphic subclarkes”. For example, data set named “Carboniferous Black Shales” consists of: two carbonate samples (20 analyses), five chert samples (143 analyses), 19 terrigenous (+volcanic-sedimentary) samples (950 analyses). Obviously, calculation of arithmetic mean by lithologies would be incorrect since of Carboniferous subdivisions are different in thickness. Hence, straton means must be “weighted” for abundance of lithologies. For weight coefficients we used values recommended by A.I. Yeliseev (1997, oral communication)5: cherts : carbonates : terrigenous rocks = 0:2 : 0:3 : 0:5 4 It is necessary to note, that the term “formation” in Russian geological literature has a different meaning than in the West. “Formations” as understood in Russia are nearly equal to “assemblages”; they denote much thicker stratigraphic units than implied by American usage. What is meant by “formation” in the USA is close to “member” in Russia (“sveeta” or свита in Russian). ð7Þ 5 These values are based on an extended study of the Russian plate/the Urals boundary zone [Eliseev, 1978, 1982; Puchkov, 1979]. Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 If phosphate rocks are present, carbonates make up 0.29 and phosphates 0.01. For example, mean Zn contents for Carboniferous carbonates, cherts, and terrigenous rocks make up 440, 400, and 160 ppm respectively. As a result, the weighted mean for Carboniferous black shales (“lithostratigraphic subclarke”) may calculated as follows: K2 ðCarboniferousÞ = 0:3 × 440 + 0:2 × 400 + 0:5 × 160 = 280 ppm: ð8Þ 2.2.3. “Lithological Clarke K1” estimation Having lithologic averages for each straton (for example, “mean Zn content in Carboniferous cherts”), one may calculate “common lithological Clarke K1” as shown in Fig. 3, left. First, lithologic subclarkes are calculated using 15 (or fewer) stratons (i is a straton's number): K1 ðCaÞ½carbonate rocks = Median KiCa K1 ðSiÞ½cherts = Median KiSi ð9Þ ð10Þ K1 ðAlÞ½terrigenous + volcanicsedimentary rocks = Median KiAl ð11Þ Second, “lithological Clarke K1” is calculated as a weighted mean: K1 ¼ 0:2K1 ðSiÞ þ 0:3K1 ðCaÞ þ 0:5K1 ðAlÞ ð12Þ In this type of calculation, preference is given to black shale lithologies; in spite of uncertainty in rock names, the Clarke values obtained provide comparisons which arrive at unexpected conclusions. For example, for Zn we have: K1 ðCaÞ ¼ 140 43; K1 ðSiÞ ¼ 160 30; and K1 ðAlÞ ¼ 140 20 ppm: ð13Þ Similar subclarke values for Zn in chert and carbonate black shales indicate mostly hydrogenic Zn-origin. Moreover, enriched Zn content 139 in black cherts has untrivial implications suggesting its biogenic origin (Yudovich and Ketris, 1994, 1997). 2.2.4. “Lithostratigraphic Clarke K2” estimation First, stratone subclarkes are grouped as follows: K1(Ph) – for Phanerozoic including Vendian, mean of 11 stratones; K1(pЄ3) – for Upper Precambrian (=Riphean); K1(pЄ2) – for Middle Precambrian (=Karelian); K1(pЄ1) – for Lower Precambrian, mean of two stratons.> Second, “lithostratigraphic clarke K2” is calculated by averaging the values: K2 ¼ 1=4½K1 ðPhÞ þ K1 ðpЄ3 Þ þ K1 ðpЄ2 Þ þ K1 ðpЄ1 Þ ð14Þ The procedure is aimed at considering the relative duration of geochrones, consequently, to some extent, corresponding black shale masses for each straton. It is obvious that K2 is 75% composed of prePhanerozoan stratons (3:1), whereas, for K1 estimations the corresponding ratio is 4:11. This is why, K2-values strongly depend on old metamorphic schists subclarkes. If the schists are enriched in any element as compared with Phanerozoan, K2-values will be higher than K1 and vice versa. Besides, if qtrueq black shale distribution in the sedimentary shell is different from 3:1, K2-estimation would be also invalid. 2.3. Some words about “true” Clarkes It is hard to say which of the estimations, K1 or K2, is more relevant. For this reason both the estimations are tabulated (see Tables 1 and 2). The closer are K1 and K2 values, the more reliable is the “true” Сlarke value, which is a mathematical expectation of a random value. Besides, similar K1, K2 and common median values are an important Fig. 3. Calculation of the qlithologicalq and qlithostratigraphicq Clarkes K1 and K2 (Yudovich and Ketris, 1994, p. 47; 1997, p. 34). See the text for comments. Author's personal copy 140 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Table 1 Some statistics of trace elements distribution in black shales, ppm Me ± δMe Geochemical background A0 A1 K1 ± δK1 K2 KCa 1 ±δ KSi 1 ±δ KAl 1 ±δ 74 ± 4 31 ± 2 4.7 ± 0.4 2.0 ± 0.9 190 ± 10 500 ± 20 40–120 15–50 2–7 0.5–10 100–300 270–800 120–170 50–70 7–10 10–15 300–400 800–1000 170–200 70–100 10–12 15–25 400–500 1000–1400 68 ± 5 33 ± 3 4.8 ± 0.6 2.5 ± 0.7 290 ± 10 590 ± 40 76 37 4.2 2.2 230 630 39 ± 4 24 ± 10 4.1 ± 0.8 0.7 ± 0.8 480 ± 40 540 ± 90 44 ± 6 19 ± 2 3.0 ± 1.6 4.8 ± 2.1 140 ± 20 740 ± 70 93 ± 9 44 ± 2 5.9 ± 1.0 2.6 ± 1.1 200 ± 10 560 ± 60 Cation- and anion-forming lithophiles with constant valency Be 127 7810 2.0 ± 0.1 1–3 Sc 140 6880 12 ± 1 7–20 Y 142 8620 26 ± 1 15–40 Yb 110 5640 2.8 ± 0.1 2.0–3.5 La 128 4495 28 ± 1 20–40 Ce 75 2800 58 ± 3 35–80 Eu 40 1500 1.2 ± 0.04 1.0–1.4 Lu 32 950 0.4 ± 0.01 0.30–0.45 Nd 50 1300 33 ± 2 15–45 Sm 44 1270 5.4 ± 0.3 2.5–7.0 Tb 45 1160 0.75 ± 0.03 0.5–0.9 Pr 15 375 4.2 ± 0.5 2.5–6.0 Gd 25 400 4.7 ± 0.4 2.5–6.0 Dy 15 340 3.0 ± 0.4 1–4 Er 15 375 1.9 ± 0.1 1–2 Ho 5 330 0.52 ± 0.09 0.2–0.6 Tm 14 33 0.40 ± 0.04 0.3–0.55 Ga 249 10400 16 ± 1 9–25 Ge 60 4630 2.4 ± 0.2 1.2–3.0 3–4 20–25 40–50 3.5–4.0 40–50 80–110 1.4–1.6 0.45–0.50 45–55 7.0–8.5 0.9–1.1 6.0–8.0 6.0–7.5 4–5 2–3 0.6–0.7 0.55–0.65 25–35 3.0–4.0 4–5 25–30 50–65 4.0–5.0 50–60 110–130 1.6–1.8 0.50–0.60 55–65 8.5–10 1.1–1.3 8–10 7.5–9.0 5–6 3–4 0.7–0.8 0.65–0.80 35–45 4.0–5.0 2.1 ± 0.1 11 ± 0.5 23 ± 1 2.6 ± 0.1 28 ± 2 58 ± 6 1.0 ± 0.4 0.35 ± 0.02 26 ± 2 4.6 ± 0.3 0.60 ± 0.03 4.3 ± 0.3 3.8 ± 0.4 2.7 ± 0.2 1.6 ± 0.1 0.38 ± 0.02 0.26 ± 0.02 17 ± 1 2.4 ± 0.1 2.6 12 23 2.8 26 55 1.0 0.35 29 4.5 0.66 4.4 4.3 2.9 1.8 0.50 0.35 17 2.5 1.6 ± 0.4 5.9 ± 0.5 12 ± 3.4 1.7 ± 0.2 22 ± 2 50 ± 12 0.47 ± 0.09 0.29 ± 0.02 12 ± 4.7 2.5 ± 0.3 0.28 ± 0.06 4.7 ± 1.6 2.7 1.2 ± 0.3 1.0 ± 0.25 0.07 0.20 13 ± 1 1.2 ± 0.2 2.2 ± 0.2 11 ± 1.1 25 ± 2 2.9 ± 0.2 29 ± 1 53 ± 22 1.5 ± 0.1 0.33 ± 0.05 36 ± 3 6.5 ± 0.4 0.84 ± 0.05 4.3 ± 0.3 1.4 ± 1.7 1.8 ± 0.3 1.5 ± 0.2 – – 14 ± 1 3.0 ± 0.3 2.4 ± 0.2 14 ± 1 29 ± 1 2.9 ± 0.2 31 ± 2 61 ± 6 1.1 ± 0.1 0.40 ± 0.01 25 ± 3 4.5 ± 0.5 0.60 ± 0.02 4.1 ± 0.5 5.5 ± 0.5 3.9 ± 0.4 2.0 ± 0.1 0.57 ± 0.04 0.30 ± 0.03 20 ± 1 2.8 ± 0.2 Cation- and anion-forming lithophiles with variable valency Ti 466 17700 3000 ± 100 1500–4600 Zr 251 13300 120 ± 5 60–190 Hf 49 1290 4.2 ± 0.2 2.5–6.0 Th 99 6310 7.0 ± 0.4 4–11 Sn 170 8040 3.9 ± 0.3 2–10 V 653 25200 205 ± 15 100–400 Nb 39 1800 11 ± 1 7–15 Ta 40 680 0.7 ± 0.04 0.5–1.0 Mo 495 18480 20 ± 1.5 6–60 W 36 1380 2.9 ± 1.0 0–15 U 240 8400 8.5 ± 0.8 4–25 Re 48 670 0.9 ± 0.3 0.2–3.5 4600–6200 190–260 6.0–7.5 11–15 10–15 400–600 15–20 1.0–1.3 60–100 15–25 25–40 3.5–6.0 6200–7800 260–330 7.5–9.0 15–19 15–20 600–800 20–25 1.3–1.6 100–140 25–35 40–55 6–9 2700 ± 100 120 ± 5 3.5 ± 0.3 7.2 ± 0.4 5.6 ± 0.3 180 ± 10 10 ± 0.7 0.66 ± 0.06 20 ± 3 7.8 ± 1.4 13 ± 2 0.4 ± 0.3 2800 200 4.5 7.8 5.7 180 15 0.8 14 2.7 9.9 0.8 1200 ± 100 74 ± 6 2.2 ± 1.6 3.9 ± 0.6 5.0 ± 0.7 99 ± 23 2.9 ± 1.0 0.17 ± 0.14 16 ± 7 8.1 ± 1.2 10 ± 1.9 0.5 ± 0.5 2100 ± 200 120 ± 10 2.9 ± 0.3 5.0 ± 0.6 4.0 ± 0.4 250 ± 30 17 ± 2 0.55 ± 0.22 29 ± 3 21 ± 5 13 ± 2 1.0 ± 0.2 3900 ± 100 150 ± 10 3.8 ± 0.5 7.4 ± 0.6 6.6 ± 0.4 200 ± 10 12 ± 1 0.8 ± 0.1 18 ± 3 2.4 ± 1.9 14 ± 3 0.2 ± 0.1 Metals-thiophiles B 196 P 364 F 47 Cl 25 Br 23 I 12 Cu 580 Ag 220 Au 148 Zn 489 Cd 57 Hg 54 In 7 Pb 436 Bi 20 Elements N n Typical cation-forming lithophiles Rb 80 5790 Li 57 4520 Cs 33 3170 Tl 18 2710 Sr 304 16650 Ba 314 15100 Non metals-thiophiles As 130 Sb 82 Se 94 Te 7 Fe-group elements Cr 562 Mn 534 Co 517 Ni 638 9460 14,900 1630 450 1070 310 25,740 9000 9120 18,540 2260 1420 176 20,520 2740 56 ± 3 1400⁎ ± 160 660 ± 70 300 ± 100 13 ± 3 1.6 ± 0.2 70 ± 3 1.0 ± 0.1 7.0 ± 1.0 130 ± 10 5.0 ± 0.6 0.27 ± 0.03 0.7 21 ± 1 1.1 ± 0.3 30–120 930–220 400–1500 100–1000 2–30 – 35–150 0.4–2.4 3–20 60–300 2–12 0.2–0.6 – 10–40 0–4 120–165 2200–3000 1500–2000 1000–2000 30–50 – 150–230 2.5–4.0 20–35 – 300–470 12–19 0.6–0.8 – 40–60 165–220 3000–3800 2000–3000 2000–3000 50–70 – 230–310 4.0–5.5 35–50 470–640 19–26 0.8–1.1 – 60–85 6–10 71 ± 5 1800⁎ ± 100 830 ± 70 420 ± 90 9.8 ± 1.7 1.3 ± 0.2 87 ± 9 1.6 ± 0.2 7.6 ± 3.9 140 ± 20 6.9 ± 1.4 0.23 ± 0.03 2.4 26 ± 1 2.0 ± 0.5 65 1300⁎ 760 380 12 1.3 140 1.7 9.7 160 6.2 0.32 – 25 1.8 4.3 ± 8 710 ± 70 640 ± 70 340 ± 80 13 ± 3 0.8 ± 0.1 55 ± 26 1.7 ± 0.6 4.9 ± 2.1 140 ± 43 8.3 ± 3.3 0.29 ± 0.07 0.3 26 ± 2 2.5 ± 0.6 87 ± 17 1200 ± 100 720 ± 130 400 ± 190 12 ± 4.6 – 100 ± 16 1.0 ± 0.3 8.5 ± 1.3 160 ± 30 9.0 ± 3.3 0.18 ± 0.03 0.01 17 ± 2 3.4 ± 0.7 81 ± 6 1200 ± 100 780 ± 100 470 ± 160 7.0 ± 4.3 1.6 ± 0.2 100 ± 8 1.8 ± 0.2 8.8 ± 7.7 140 ± 20 5.3 ± 1.3 0.22 ± 0.03 4.6 ± 3.2 29 ± 2 1.1 ± 0.8 4190 1930 1650 110 30 ± 3 5.0 ± 0.5 8.7 ± 1.4 2.0 ± 0.3 10–80 2–11 3–30 1.3–3.0 80–130 11–17 30–50 3–4 130–180 17–23 50–70 4–5 30 ± 3 5.6 ± 0.8 7.8 ± 1.0 2.1 58 5.3 9.3 1.8 34 ± 6 6.8 ± 0.8 8.0 ± 0.8 4.2 30 ± 11 8.8 ± 3.9 12 ± 2 0.3 27 ± 3 3.6 ± 0.4 6.6 ± 1.7 1.7 21,900 19,600 21,000 23160 96 ± 3 400 ± 20 19 ± 1 70 ± 2 50–160 200–800 10–30 40–140 160–220 800–1200 30–40 140–210 220–280 1200–1600 40–50 210–280 81 ± 5 440 ± 30 14 ± 1 67 ± 4 93 1100 18 67 45 ± 6 500 ± 50 11 ± 1.1 41 ± 6 86 ± 12 250 ± 30 11 ± 1.5 63 ± 7 100 ± 7 470 ± 30 17 ± 2 84 ± 6 (N – number of primary samples, n – number of analyses). Calculations by Marina P. Ketris, 1990. ⁎ The asterisk means that calculation was made with including 0.01 phosphate contribution, as: chert:carbonate:terrigenic (+tuff):phosphate = 0.20:0.29:0.50:0.01. Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Table 2 (continued) Table 2 Subclarkes for the black shale lithologies, ppm Elements Carbonate shales, m n KCa 1 Me ± σMe Elements Chert shales, m n Typical cation-forming lithophile elements Li 9 1580 24 ± 10 8 930 Rb 11 1720 39 ± 4 11 1700 Cs 5 1550 4.1 ± 0.8 5 490 Tl 5 1440 0.7 ± 0.8 3 380 Sr 11 2210 480 ± 40 12 4870 Ba 12 2740 540 ± 90 12 4760 KSi 1 Terrigenous and volcanic-sedimentary shales, KAl 1 Me ± σMe m 19 ± 2 47 ± 6 3.0 ± 1.6 4.8 ± 2.1 140 ± 20 740 ± 70 13 14 10 6 15 15 2010 1660 1140 890 8560 7630 44 ± 2 93 ± 9 5.9 ± 1.0 2.6 ± 1.1 200 ± 10 560 ± 60 3900 3650 4860 2350 1740 340 780 660 990 370 630 340 330 340 32 2820 550 5880 1870 2.4 ± 0.2 14 ± 1 29 ± 1 31 ± 2 61 ± 6 4.1 ± 0.5 25 ± 3 4.5 ± 0.5 1.1 ± 0.06 5.5 ± 0.5 0.60 ± 0.02 3.9 ± 0.4 0.57 ± 0.04 2.0 ± 0. 1 0.30 ± 0.03 2.9 ± 0.2 0.40 ± 0.01 20 ± 1 2.8 ± 0.2 Cation- and anion-forming elements Be 10 1420 1.6 ± 0.4 Sc 10 1750 5.4 ± 0.3 Y 10 2230 12 ± 3.4 La 10 1080 22 ± 3 Ce 8 660 37 ± 12 Pr 2 5 4.7 ± 1.6 Nd 4 260 12 ± 5 Sm 4 260 2.4 ± 0.3 Eu 3 260 0.42 ± 03 Gd 1 4 2.7 Tb 3 250 0.27 ± 0.03 Dy 2 2 1.2 ± 0.3 Ho 1 1 0.07 Er 2 5 1.0 ± 0.25 Tm 1 1 0.2 Yb 8 1790 1.8 ± 0.2 Lu 3 260 0.28 ± 0.01 Ga 12 1880 13 ± 1 Ge 9 1820 1.2 ± 0.2 with stable valency 9 2490 2.2 ± 0.2 11 1350 12 ± 2 10 1530 25 ± 2 9 710 31 ± 3 8 340 41 ± 48 2 29 4.3 ± 0.3 4 57 40 ± 10 4 200 5.0 ± 1.2 3 41 1.3 ± 0.3 3 27 1.4 ± 1.7 4 66 0.72 ± 0.21 2 36 1.8 ± 0.3 Cation- and anion-forming elements Ti 13 3750 1200 ± 100 Zr 11 2690 74 ± 6 Hf 3 260 2.2 ± 2.8 Th 8 1440 3.9 ± 0.6 Sn 10 1930 5.0 ± 1.3 V 14 4790 99 ± 23 Nb 6 1170 2.9 ± 1.0 Ta 3 3 0.17 ± 0.12 Mo 13 3770 16 ± 7 W 4 690 8.1 ± 1.2 U 10 1230 10 ± 1.9 Re 4 440 0.5 ± 0.5 with variable 11 4100 10 4470 3 41 8 1890 10 1660 12 6550 7 1750 3 41 12 6220 2 70 11 3480 3 110 n Me ± σMe 2 29 1.5 ± 0.2 8 910 2.8 ± 0.3 10 7 2640 930 14 ± 1 3.0 ± 0.3 13 14 15 13 12 6 9 9 8 7 8 6 4 6 3 13 7 15 10 valency 2100 ± 200 140 ± 10 3.1 ± 2.4 4.2 ± 0.7 4.0 ± 0.4 250 ± 30 17 ± 2 0.55 ± 0.06 29 ± 3 21 ± 5 13 ± 2 1.0 ± 0.2 15 15 8 14 13 15 11 5 15 10 15 4 9810 6040 780 2470 4480 13870 1880 430 8490 530 3700 120 3900 ±100 150 ± 10 3.8 ± 0.5 7.4 ± 0.6 6.6 ± 0.4 200 ± 10 12 ± 1 0.83 ± 0.09 18 ± 3 2.4 ± 1.9 14 ± 3 0.2 ± 0.1 87 ± 17 1200 ± 100 720 ± 130 400 12 ± 5 14 15 13 9 7 5 4560 8200 600 290 460 270 81 ± 6 1200 ± 100 780 ± 100 470 ± 160 7.0 ± 4.3 1.6 ± 0.2 100 ± 16 1.2 ± 0.2 8.5 ± 1.3 160 ± 30 9.0 ± 3.3 0.18 ± 0.03 0.01 17 ± 2 3.4 ± 0.7 15 14 10 15 8 9 5 15 8 13730 4810 6050 11030 1470 460 220 10280 1450 100 ± 8 1.9 ± 0.3 8.8 ± 7.7 140 ± 20 5.3 ± 1.3 0.22 ± 0.03 4.6 ± 3.2 29 ± 2 1.1 ± 0.8 Typical anion-forming lithophile elements B 9 1970 43 ± 8 10 2930 P 13 2980 710 ± 70 12 3490 F 9 820 640 ± 70 6 210 Cl 6 100 340 ± 80 2 58 Br 3 600 13 ± 3 2 11 I 4 39 0.77 ± 0.04 Metals-sulfophiles Cu 14 5840 Ag 9 730 Au, ppb 6 650 Zn 14 4130 Cd 5 680 Hg 8 560 In 1 3 Pb 13 4580 Bi 5 1140 55 ± 26 11 1.7 ± 0.6 12 4.9 ± 2.1 5 140 ± 43 12 8.3 ± 3.3 4 0.29 ± 0.07 5 0.3 1 26 ± 2 12 2.5 ± 0.6 2 Nonmetals-sulfophiles As 9 840 Sb 7 610 Se 9 370 Te 1 25 34 ± 6 6.8 ± 0.8 8.0 ± 0.8 4.2 10 9 9 1 900 480 730 2 30 ± 11 8.8 ± 3.9 12 ± 2 0.03 14 12 11 2 2460 840 550 80 27 ± 3 3.6 ± 0.4 6.0 ± 1.7 1.7 Elements-siderophiles Cr 13 3820 Mn 13 4540 45 ± 6 500 ± 50 12 12 5090 4970 86 ± 12 250 ± 30 15 15 13040 10100 100 ± 7 470 ± 30 6170 3320 2420 3390 110 400 1 5670 150 141 (continued on next page) Carbonate shales, KCa 1 m n Elements-siderophiles Co 11 3540 Ni 13 4290 Chert shales, KSi 1 Terrigenous and volcanic-sedimentary shales, KAl 1 Me ± σMe m n Me ± σMe m n Me ± σMe 11 ± 1.1 41.6 12 12 5410 5460 11 ± 1.5 63 ± 7 15 15 12010 13400 17 ± 2 84 ± 6 Calculations by Marina P. Ketris, 1990. m - number of stratons, n - number of analyses, Me - straton median, σMe – standard deviation of straton median. controlling factor. At any rate, it seems likely that they must not exceed the double-standard median limits: ðK1 ; K2 Þ b Me F 2δMe No doubt, all the preceding errors affect the final “true” estimation. Among them are: - incorrectly defined original lithologies; - inaccurate weight coefficients, both lithologic (0.2:0.3:0.5) and stratigraphic (1:3); - number of lithological subclarkes different from each other; - number of Phanerozoic black shales different from pre-Phanerozoic ones. For example, assumed weight coefficients for carbonate (0.3) and terrigenous shales (0.5) may differ for Paleozoic platforms (more than 0.3) and Mesozoan geosynclines (more than 0.5), respectively. As follows from the above discussion, we are very well aware of all the inaccuracy of the Сlarke values we propose. It should be noted, however, that they are no less accurate than the Clarke values for the Earth's crust and the ocean widely referred to in literature. From all the variety of strongly diverging values, Vinogradov (1962) simply chose the figures he thought adequate. No doubt our Clarke values will be revised in the not so distant future. Nevertheless, the Clarkes proposed here are presently quite useful. 3. Estimation of trace elements Clarkes for coals As was mentioned above, such calculations were made by several researchers, but there are many difficulties in the calculation. 3.1. Coal-basis and ash-basis Clarkes Up to 1970–1980, coal was mostly not directly analyzed for trace elements but through an analysis of coal ash. Standard coal ashing was performed at 750 °C (in USA and many West countries), or at 850 °C (in former USSR and now in Russia)6. It is well known that some elements may be almost fully (Hg, I, Br), or partly (Ge, Mo, etc.) volatilized by high-temperature ashing. Trace element loss may be minimized by low-temperature ashing (~130–150 °C) by means of radio-frequency exposure in oxygen plasma (Gluskoter, 1965). This excellent method, is, however, time-consuming and for this reason is not widely acceptable. So, up to end of 1980s, the content of trace element in coal (coalbasis content) was obtained by recalculation from the content in ash (ash-basis content). Such recalculation may lead to underestimation of coal-basis figures, and, as a result, to underestimation of coal-basis Clarkes. During the last decades of 20th century, several direct methods of coal analysis were introduced in coal geochemistry, and first of all – INAA, instrumental neutron activation analysis. So, many directly 6 Nonstandard ashing (550 °C, with an air access) used in USSR and USA for special intent – following analyses of ash for trace elements. Author's personal copy 142 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 obtained coal-basis figures appeared in the literature. Now a new opportunity appears – to recalculate coal-basis figures to ash-basis ones. If the analysis of coal ash was earlier simply “a technical tool” (because a direct analysis of coal was too hard and unreliable procedure), have we any need for the “ash Clarkes” today? Yes, we have such need: for the calculation of important geochemical values, “coal affinity indexes” (or “coalphile coefficients), as we will discuss below. 3.2. “Coal-affinity” (coalphile) indexes Goldschmidt (1935) first calculated the “enrichment coefficients” of coal ash by comparison of element content in coal ash and Earth's crust Clarke value. For example: the Earth's crust-As Clarke value was assumed to be 5 ppm, and As content in As-rich ashes was determined as 500 ppm; so, enrichment coefficient was 500/5 = 100. Later Yudovich (1978) used coal ash Clarke values (instead of “enriched ashes”) and Clarkes of sedimentary rocks for such calculation. These figures (enrichment coefficients) were named as “typomorph coefficients”, and were widely cited in Russian and Bulgarian literature7. More recently, this poor term was substituted for coalphile coefficient (index), or coal affinity index (Yudovich and Ketris, 2002). What does a coal affinity index mean? – It shows, how efficiently coal acted as a geochemical barrier for trace elements, during all its geologic history. The more coal concentrated trace elements from environment compared with sedimentary rocks, the greater would be the coal affinity index. A researcher could compare coal affinity indexes for different elements in given coal field, given coal basin, or province; for the coals of different rank; and for the same coal field (basin, province) but for different elements. For example, As coal affinity index is 50 ppm/ 11 ppm= about 5 (Yudovich and Ketris, 2005a,d), and Hg coal affinity index is 0.75 ppm /0.05 ppm = 15 (Yudovich and Ketris, 2005b,c,d). So, Hg is threefold more coalphile element than As. 3.3. Calculation method All calculation procedures may be divided into several stages (Ketris and Yudovich, 2002): 1) analytical data collection; 2) preliminary data processing and formation of basic tables; 3) using basic tables, formation of sufficiently homogeneous statistical data samples (data assemblages), and calculation of sample-averages; 4) using sample-averages, plotting of frequency histograms and evaluation of the Clarke value, as median. 3.3.1. Analytical data collection This is very time-consuming and the hardest part of all the work. We collected coal-analyses data nearly 45 years. From the literature, it is necessary to extract the following data: (a) coal locality (country, basin, coal field (deposit), coal seam; (b) coal rank and geological age; (c) ash yield on dry matter (Ad, %); (d) how units are used (%, ppm, mg/g etc.), and basis – coal or ash; (e) number of coal specimens (analyses). Unfortunately, only in few instances we could obtain all the information! Often, a part of information needed was lacking. For example, if the coal locality was lacking, such analytical data have to be rejected. The stratigraphic data were needed in order to not include different-age coals within an united statistical sample. We need also to know: if the data collected belong to coal beds or coal inclusions? The latter may be extremely enriched in trace elements (Yudovich, 7 See numerous papers published by Greta Eskenazy, Jordan Kortensky, Stanislav Vassilev, and some others. Fig. 4. A sketch showing a procedure of stepped averaging (Tkachev and Yudovich, 1975, p. 124). The steps (levels): a – initial, the analyses of coal specimens; b – coal beds; c – coal fields; d – coal areas (basins, provinces), e – final totality of the areas (= global totality of statistical samples). 1 – anyone object of the given step (level); 2 – specific object of this step (level), for which are shown its lower (composing) components. 1972, 2003), and the use of such figures may sharply change the sample average. The other important question arises: was the given coal influenced secondary processes? Such data may also distort a homogeneity of the statistical sample. In literature prior to 1980, a simultaneous analysis presentation on an ash- and on a coal basis was rare; so, we needed to know an ash yield in order to obtain the lacking figure by recalculation. If an ash yield was lacking, we tried to evaluate it using all the available information sources (reference books, special monographs, etc.). In many Soviet papers (where in fact, many thousand spectrographic analyses were averaged), the number of analyses was lacking. In such instances, we needed to get only some “conditional” number Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 143 Fig. 5. An example of the “appropriate" plot: zirconium (Yudovich, Ketris, 2006a,b,c, p. 285). N – number of analyses, n – number of statistical samples, Me – the median of statistical samples. of analyses. This means that our “number of analyses” sometimes did not contain enough accurate information (tends to be underestimated). 3.3.2. Preliminary data processing and formation of basic tables Preliminary data processing may start even during data collection: the individual figures (belonging to each coal specimen) being averaged – if such average is lacking in original text. As elemental information units, we assumed a coal bed- or coal-field averages. For example, if there were several figures for one coal bed, we calculated the coal-bed average and put it in the basic table. If there were the figures for three coal beds, these three figures were put in basic table, etc. If the set of analyses had very sharp anomalies (for example, one to two order-of-magnitude more than coal geochemical background), such figures, as a rule, were not included in average calculation. As a result, for each element we created two basic tables, for brown and hard coal, as following: Hard (brown) coals Country Coal basin (province, region, area) Coal field, bed Geologic n Ad E, ppm age on coal basis E, ppm on ash basis Reference (source of data) In such a table, n means the number of analyses, Ad – ash yield, %, E – a chemical element. 3.3.3. Stepped averaging and final median calculation After basic tables were formed, we could obtain the assemblage of statistical samples (and their averages), which would be used for the final Clarke calculation. Each statistical sample embraces the analyses of a coal basin or large coal area (region). All the analyses included are to be averaged. If the collected data allowed, we constructed two to four statistical samples for big coal basins (provinces) – for example, for some industrial regions (within Donets basin, Ukraine), coal fields, or other geographical units. The number of such “intra-basinal” statistical samples nearly corresponded with coal geological resources of the given basin. The full concept involves stepped (successive) averaging, from minor to large: one coal bed ⇒ several coal beds ⇒ coal field (deposit) ⇒ coal area (several coal fields) ⇒ coal basin or province ⇒ totality of the coal basins (=assemblage of statistical samples), i.e. Clarke value totality, comprising several dozen random (statistical) samples representing many thousands of analyses (Fig. 4). For example, coal Clarke value of As was calculated using totality, consisting of 119 statistical samples (and about 21,000 analyses) – for hard coals, and 66 statistical samples (and about 21,000 analyses) – for brown coals (Yudovich and Ketris, 2005a). On the steps, the averages were calculated as arithmetic mean (if the sample was more homogeneous), or as median (if the sample was less homogeneous). Following our good experience in black shale geochemistry (Yudovich and Ketris, 1994), we use a sample totality median (Me) Author's personal copy 144 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 as a Clarke value estimation. As an estimation of the median accuracy (σMe) we use the value pffiffiffiffiffi σ Me = ðQ3 − Q1 Þ=2 n; ð16Þ where Q1 and Q3 – two distribution quartiles, corresponding 1/4 and 3/4 of the cumulative frequency. Although a Clarke value (a median) is estimated analytically, it is very instructive to plot a frequency histogram. The distribution of trace element contents in coal is, as a rule, log-normal; that is why, for plotting the logarithmic scale is often used. The plot character could suggest: how “good” is the totality studied. If the histogram is “right” (for example – near to log-normal), the totality could be attested as a “good” one (Fig. 5); in opposite case, we can think the totality is too small and not representative of natural coals (Fig. 6). In such a case, we need to get some new analyses for a better Clarke estimation. In fact, the elements with “right” histograms nearly “do not react” upon new, supplementary analyses – their Clarke value could only little vary (as a rule – not more than within accuracy range of median, i.e. ±1σMe). On the contrary, some rare, poorly studied elements (such as Au, Pd, Tl, In, Te, Cl, I, and some others) may, in future, change their Clarkes more appreciably. The other (well known) histogram characteristic is its σMe value compared with Me value. If these figures are comparable (very large variance of the statistical distribution), the Clarke value is, of course, not very reliable; if σMe bb Me, the opposite is true. For example, average content of Pd in brown coal ashes was estimated only over eight statistical samples (nearly 50 analyses) as 0.066 ± 0.027 ppm. Such a figure appears to be too high and dubious, accounted for Chinese Pd-rich coals contribution. There is no doubt, as the sample assemblage is extended, the Pd Clarke in coal will be changed. For the REE, one other criterion exists: the picture of the standardized (normalized) curve. As example, Tm has some peak on such curve, that distinguishes Tm from the neighboring REE. It may mean that the Tm coal Clarke is now overestimated, and may be amended in future, as more information would be available. 3.3.4. Calculation results The results of the calculations are shown in Table 3 and 4. The last column in Table 3 represents coal affinity indexes, calculated using our coal Clarkes and modern weighed Clarkes for sedimentary rocks, calculated by Grigoriev (2003) using the Ronov's sedimentary shell model (Ronov, 1980; Ronov et al., 1990). It is of note, the coal affinity indexes are calculated on the base of averaged ash Clarkes, i.e. (element Clarke in hard coal ash + element Clarke in brown coal ash)/2. For example, lithium: (49 ppm + 82 ppm)/2 ≈ 66 ppm; 66 ppm/33 ppm = 2 (coal affinity index). It should be noted that some Clarkes in Table 3 are based on small statistical samples or/and not very reliable analyses. Such (preliminary) Fig. 6. An example of the irregular plot: iodine. Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 145 Table 3 Coal Clarke values for trace elements Elements Coals Brown Typical cation-forming lithophile elements Li 10 ± 1.0 Rb 10 ± 0.9 Cs 0.98 ± 0.10 Tl 0.68 ± 0.07 Sr 120 ± 10 Ba 150 ± 20 Coal ashes Hard All Brown Hard All Clarke value of sedimentary rocks 14 ± 19 18 ± 1 1.1 ± 0.12 0.58 ± 0.04 100 ± 7 150 ± 10 12 14 1.0 0.63 110 150 49 ± 4 48 ± 5 5.2 ± 0.5 5.1 ± 0.5 740 ± 70 900 ± 70 82 ± 5 110 ± 10 8.0 ± 0.5 4.6 ± 0.4 730 ± 50 980 ± 60 66 79 6.6 4.9 740 940 33 94 7.7 0.89 270 410 Cation- and anion-forming elements with stable valency Be 1.2 ± 0.1 2.0 ± 0.1 Sc 4.1 ± 0.2 3.7 ± 0.2 Y 8.6 ± 0.4 8.2 ± 0.5 La 10 ± 0.5 11 ± 1 Ce 22 ± 1 23 ± 1 Pr 3.5 ± 0.3 3.4 ± 0.2 Nd 11 ± 1 12 ± 1 Sm 1.9 ± 0.1 2.2 ± 0.1 Eu 0.50 ± 0.02 0.43 ± 0.02 Gd 2.6 ± 0.2 2.7 ± 0.2 Tb 0.32 ± 0.03 0.31 ± 0.02 Dy 2.0 ± 0.1 2.1 ± 0.1 Ho 0.50 ± 0.05 0.57 ± 0.04 Er 0.85 ± 0.08 1.00 ± 0.07 Tm 0.31 ± 0.02 0.30 ± 0.02 Yb 1.0 ± 0.05 1.0 ± 0.06 Lu 0.19 ± 0.02 0.20 ± 0.01 Ga 5.5 ± 0.3 6.0 ± 0.2 Ge 2.0 ± 0.1 2.4 ± 0.2 1.6 3.9 8.4 11 23 3.5 12 2.0 0.47 2.7 0.32 2.1 0.54 0.93 0.31 1.0 0.20 5.8 2.2 6.7 ± 0.5 23 ± 1 44 ± 3 61 ± 3 120 ± 10 13 ± 2 58 ± 5 11 ± 1 2.3 ± 0.2 16 ± 1 2.0 ± 0.1 12 ± 1 3.1 ± 0.3 4.6 ± 0.2 1.8 ± 0.3 5.5 ± 0.2 1.10 ± 0.10 29 ± 1 11 ± 1 12 ± 1 24 ± 1 57 ± 2 76 ± 3 140 ± 10 26 ± 3 75 ± 4 14 ± 1 2.6 ± 0.1 16 ± 1 2.1 ± 0.1 15 ± 1 4.8 ± 0.2 6.4 ± 0.3 2.2 ± 0.1 6.9 ± 0.3 1.3 ± 0.1 36 ± 1 18 ± 1 9.4 23 51 69 130 20 67 13 2.5 16 2.1 14 4.0 5.5 2.0 6.2 1.2 33 15 1.9 9.6 29 32 52 6.8 24 5.5 0.94 4.0 0.69 3.6 0.92 1.7 0.78 2.0 0.44 12 1.4 Cation- and anion-forming elements with stable valency Ti 720 ± 40 890 ± 40 Zr 35 ± 2 36 ± 3 Hf 1.2 ± 0.1 1.2 ± 0.1 Th 3.3 ± 0.2 3.2 ± 0.1 Sn 0.79 ± 0.09 1.4 ± 0.1 V 22 ± 2 28 ± 1 Nb 3.3 ± 0.3 4.0 ± 0.4 Ta 0.26 ± 0.03 0.30 ± 0.02 Mo 2.2 ± 0.2 2.1 ± 0.1 W 1.2 ± 0.2 0.99 ± 0.11 U 2.9 ± 0.3 1.9 ± 0.1 800 36 1.2 3.3 1.1 25 3.7 0.28 2.2 1.1 2.4 4000 ± 200 190 ± 10 7.5 ± 0.4 19 ± 1 4.7 ± 0.4 140 ± 10 18 ± 1 1.4 ± 0.1 15 ± 1 6.0 ± 1.7 16 ± 2 5300 ± 200 230 ± 10 9.0 ± 0.3 23 ± 1 8.0 ± 0.4 170 ± 10 22 ± 1 2.0 ± 0.1 14 ± 1 7.8 ± 0.6 15 ± 1 4650 210 8.3 21 6.4 155 20 1.7 14 6.9 16 Typical anion-forming lithophile elements B 56 ± 3 P 200 ± 30 F 90 ± 7 Cl 120 ± 20 Br 4.4 ± 0.8 I 2.3 ± 0.4 52 230 88 180 5.2 1.9 410 ± 30 1200 ± 100 630 ± 50 770 ± 120 32 ± 5 13 ± 2 260 ± 20 1500 ± 100 580 ± 20 2100 ± 300 32 ± 9 12.2 ± 5.4 335 1350 605 1440 32 12.6 Metals-sulfophiles Cu Ag Au, ppb Zn Cd Hg In Pb Bi 47 ± 3 250 ± 10 82 ± 6 340 ± 40 6.0 ± 0.8 1.5 ± 0.3 3740 170 3.9 7.7 2.9 91 7.6 1.0 1.5 2.0 3.4 72 670 470 2700⁎ 44 1100 CAI 2.0 0.84 0.86 5.5 2.7 2.3 4.9 2.4 1.8 2.2 2.5 2.9 2.8 2.4 2.7 4.0 3.0 3.9 4.3 3.2 2.6 3.1 2.7 2.8 11 1.2 1.2 2.1 2.7 2.2 1.7 2.6 1.7 9.3 3.5 4.7 4.7 2.0 1.3 0.53 0.73 0.01 15 ± 1 0.090 ± 0.020 3.0 ± 0.6 18 ± 1 0.24 ± 0.04 0.10 ± 0.01 0.021 ± 0.002 6.6 ± 0.4 0.84 ± 0.09 16 ± 1 0.100 ± 0.016 4.4 ± 1.4 28 ± 2 0.20 ± 0.04 0.10 ± 0.01 0.040 ± 0.020 9.0 ± 0.7 1.1 ± 0.1 16 0.095 3.7 23 0.22 0.10 0.031 7.8 0.97 74 ± 4 0.59 ± 0.09 20 ± 5 110 ± 10 1.10 ± 0.17 0.62 ± 0.06 0.11 ± 0.01 38 ± 2 4.3 ± 0.8 110 ± 5 0.63 ± 0.10 24 ± 10 170 ± 10 1.20 ± 0.30 0.87 ± 0.07 0.21 ± 0.18 55 ± 6 7.5 ± 0.4 92 0.61 22 140 1.2 0.75 0.16 47 5.9 31 0.12 6.0 43 0.80 0.068 0.043 12 0.26 3.0 5.1 3.7 3.3 1.5 11 3.7 3.9 23 Nonmetals-sulfophiles As Sb Se 7.6 ± 1.3 0.84 ± 0.09 1.0 ± 0.15 9.0 ± 0.7 1.00 ± 0.09 1.6 ± 0.1 8.3 0.92 1.3 48 ± 7 5.0 ± 0.4 7.6 ± 0.6 46 ± 5 7.5 ± 0.6 10.0 ± 0.7 47 6.3 8.8 7.6 1.2 0.27 6.2 5.3 33 Elements-siderophiles Cr Mn Co Ni Pd 15 ± 1 100 ± 6 4.2 ± 0.3 9.0 ± 0.9 0.013 ± 0.006 17 ± 1 71 ± 5 6.0 ± 0.2 17 ± 1 0.001 ± 0.002 16 86 5.1 13 0.0074 82 ± 5 550 ± 30 26 ± 1 52 ± 5 0.066 ± 0.027 120 ± 5 430 ± 30 37 ± 2 100 ± 5 0.007 ± 0.011 100 490 32 76 0.037 58 830 14 37 1.7 0.59 2.3 2.1 (continued on next page) (continued on next page) Author's personal copy 146 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Table 3 (continued) Elements Elements-siderophiles Ir Pt Coals Coal ashes Brown Hard 0.002 ± 0.006 0.065 ± 0.018 0.001 ± 0.0003 0.005 ± 0.003 All 0.002 0.035 Brown Hard 0.013 ± 0.031 0.22 ± 0.04 0.007 ± 0.003 0.038 ± 0.018 Clarke value of sedimentary rocks All CAI 0.010 0.13 Calculations by Marina P. Ketris, 2005. CAI – coal affinity indexes (“coalphile indexes”). CAI = average element concentration in coal ash/Clarke value in sedimentary rocks. ⁎ Cl Clarke value in sedimentary rocks – according to (Ronov et al., 1990). ⁎⁎The brackets mean that calculated Clarke value (based on small statistical samples or not very reliable analyses) is of minor adequacy (reliability). Such values could be strong changed in future. 9 The ± figures mean ±1σMe Clarkes could be strong changed in future; corresponding figures are shown in the brackets. 4. Discussion in coal geochemistry The data in Table 3 allow us to make some comparisons and raise issues. 4.1. The values from 1985 and 2005: a comparison. Due to new analytical methods and the great information gain, the 2005 Clarke estimations for some elements were appreciably changed compared with 1985 estimations. Among these elements: Li, Rb, Ag, Ti, Nb, As, Cl – for all coals; Cu, Be, B, Sc, Th – for brown coals; Hg – for hard coals. Other Clarke estimates also changed but less, only within 20–40%: Ga, Y, Yb, Р – for all the coals; Ba, B, Cd, Co – for brown coals; Zn, Sc, Mo, Mn – for hard coals. Finally, many Clarke values have kept within the calculation accuracy range: Sr, Ge, Zr, Sn, V, F, Cr, Ni – for all the coals; Hg, Zn, Mo, Mn – for brown coals; Rb, Be, В, Cu, Th – for hard coals. 4.2. Comparison of brown and hard coal Clarkes Such a comparison allows analysis of coal rank influence on coal geochemistry. Coal “metamorphism” is the thermal epigenetic process, involving hot brines and fluids influencing coal beds. These processes not only changed coal organic matter, but may greatly change the trace element contents by means of their input or output (Yudovich, 1978; Yudovich and Ketris, 2002). However, this (obvious) issue is far from universal. It is known that, in general, hard coals are Paleozoic, and brown coals are Mesozoic and Cenozoic. This means that in some instances the geochemical difference “brown coals vs. hard coals” may be primary, accounted for the large initial difference of Paleozoic and Mesozoic-Cenozoic coal-forming flora. As seen from Table 3, brown coals are enriched in B, U, and Mn. For all three elements, one can suggest their output during thermal metamorphism of the coal organic matter. More elements are enriched in hard coals compared with brown ones. Weak enrichments include Co, Ge, V, Pb, Se, and strong ones include Rb, Be, Zn and Ni. Only for the evident sulfophile elements (such as Pb, Se, and Zn), such enrichments could be caused by hydrothermal input related to coal metamorphism. However, for litho- and siderophile elements (such as Be, Cr, Co, Ni) such explanation appears to be dubious. It is not excluded that primary coal-formed flora acted here as actual factor of difference (Yudovich, 1978). Such (non-trivial) conclusion could highlight some problems dealing with the biosphere evolution. 4.3. Coal affinity indexes By analyzing ranges of the coalphile indexes (the last column in Table 3), one can divide the elements in four groups: (a) non-coalphile elements (coal affinity indexes are b1): I, Cl, Mn, Br, Rb, Cs; (b) weakly or moderate coalphile elements (coal affinity indexes range from 1 to 2): Ti, Zr, F, Cd, V, Ta, Cr, Y, Li, P; (c) coalphile elements (coal affinity indexes range from 2 to 5): Ni, Hf, Sn, La, Co, Ba, Sc, Nb, Sr, Th, Ga, Cu, REE, Zn, W, Au, In, Pb, U, B, Be; (d) highly coalphile elements (coal affinity indexes N5): Ag, Sb, Tl, As, Mo, Ge, Hg, Bi, Se. The greater is the coal affinity index, the greater is the contribution of an authigenic fraction of given trace element (represented by organic or micro-mineral forms), and the less is one of a clastogenic fraction (represented by macro-mineral forms (for example, silicatic). Due to the new (weighed) Clarkes of sedimentary rocks (Grigoriev, 2003), some earlier estimations of coal affinity have drastic changed. So, after the calculation of real evaporites contribution in sedimentary shell (Ronov, 1980; Ronov et al., 1990), the weighted Clarkes of halogens Cl, Br, and I in sedimentary rocks increased. This results in a corresponding sharp decrease in their coal affinity indexes: halogens “have transformed” from highly coalphile elements (Yudovich et al., 1985; Yudovich and Ketris, 2006b) to non-coalphile ones8. Also, the coalphile indexes of Au, Cd, Y, V, U, Cr have unexpectedly decreased, and, in turn, indices of Tl, Zn, Hf, and In have unexpectedly increased. The extreme coalphile indices of Bi and Se appear to be very doubtful. Such strange figures could be accounted for errors in the Clarke values for sedimentary rocks (and not for coal Clarkes errors?). It is of note, these elements are often not analyzed; due to analytical difficulties; see, for example, for Se (Yudovich and Ketris, 2006a). This may be a factor. 5. Conclusion Black shale Clarke values are the average trace element contents in the World black shales. It is well known that black shales (carbonaceous, mainly marine sedimentary rocks) often are metalliferous and/ or bituminous: enriched in U, V, Mo, Re, many other trace elements, and “shale oil”. That is why the Clarke values for black shale are very important for each researcher dealing with oil- or ore-bearing black shales. The tables of black shale Clarkes calculated by the authors in 1994 are presented here; these calculations are based on huge information. Coal Clarke values are the average trace element contents in the World coals. The modern table of coal Clarkes is presented here, calculated by the authors based on a very large amount of information (thousands analyses of coal and coal ashes for trace elements). The coal Clarkes are the scientific tool for many geochemical comparisons and issues. First of all, coal Clarkes allow to compare given (studied) coal with World geochemical background, and to conclude: is the coal “normal” (near to the Clarke level), enriched or, in 8 However, J.C.Hower believes [personal communication, December 2008] that halogens are coalphile on lithotype basis; vitrains have high Cl and Br where these elements are high. But, Cl (or Br) are not universally high, so averaging takes in low-Cl coals. Author's personal copy M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Table 4 Number of analyses for coal Clarkes calculated Elements Li Rb Cs Tl Sr Ba Be Sc Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Ga Ge Ti Zr Hf Th Sn V Nb Ta Mo W U B P F Cl Br I Cu Ag Au Zn Cd Hg In Pb Bi As Sb Se Cr Mn Co Ni Pd Ir Pt Brown coals (lignites) Hard (bituminous) coals n N n N 44 47 40 28 78 76 80 73 72 57 44 16 30 35 35 18 34 26 25 19 18 61 32 86 84 88 74 38 52 69 97 43 33 80 45 63 65 52 37 39 27 8 90 63 31 84 40 48 7 78 26 66 43 39 95 82 95 93 8 4 6 4914 2898 1808 1588 16,335 16,108 48,001 40,358 59,505 44,689 4196 243 1203 1875 1913 453 1468 989 719 436 365 31,120 1525 41,801 32,186 26,528 49,087 10,932 3586 66,318 83,964 28,961 1626 71,606 8591 7561 5800 5127 3149 1410 1061 52 69,855 13,248 1180 78,908 2251 3623 193 67,744 2196 21,092 5220 2323 46,136 23231 69,660 70,560 52 7 44 84 79 70 49 113 127 118 112 106 105 80 35 58 71 73 42 64 49 41 37 35 94 65 127 127 126 112 65 97 103 139 80 65 131 69 110 103 105 77 83 65 18 139 96 43 137 75 94 15 136 53 124 92 81 141 134 142 143 17 9 18 11866 12742 11765 6105 27,052 18,433 114,256 118,155 117,643 22,232 12,200 4871 6828 11,440 11,578 5625 8640 5663 5298 5272 4792 95,812 10,290 121,820 128,060 28577 101,390 10,376 18,356 90,502 99,450 10,588 8224 104,901 8817 19,282 13,834 14,812 11,249 9981 6807 264 75,121 11,799 2037 103,924 15,079 34,775 648 93,983 6739 22,466 13,662 16478 64,442 29,521 125,023 135,952 431 146 201 n – number of statistical samples, N – number of analyses. turn, impoverished in given trace element. Such comparisons are known of great importance for toxic elements (such as Hg, As, Se, Be etc.), as well for valuable elements, having industrial potential (such as Ge, Ga, Sc, Re, REE, PGE, etc.). Other important values have the “coal affinity indexes” (coalphile indexes) of trace elements (average element content in coal ash/ element Clarke in sedimentary rocks). These indexes show how 147 efficiently coal acted as a geochemical barrier for trace elements during all its geologic history. A comparison of coalphile indexes with each other allow to display a contribution of coal authigenic matter (organics, sulfides etc.) in coal inorganics. On the other hand, the comparison of the coalphile indexes of the same element, but in different coals, allow to see some non-obvious peculiarities of the coal fields or coal basins. See, for example, numerous Bulgarian papers for reference, such as on the coal geochemistry of In, W (Eskenazy, 1980, 1982); REE, Zr and Hf (Eskenazy, 1987a,b); Ta, Au (Eskenazy, 1990, 1992); As and Sb (Eskenazy, 1995); Be (Eskenazy, 2006); B (Eskenazy et al., 1994), Sr (Eskenazy and Mincheva, 1989), REE, U and Th (Kortenski and Bakardjiev, 1993); on coal inclusions with unique contents of trace elements (Vassilev et al., 1995), etc. Acknowledgements James Hower kindly and selflessly edited the original Russian-toEnglish translation of this manuscript. Greta Eskenazy and Vladimir Seredin (the reviewers) made some valuable notes and corrections. We are very appreciative to Editor-in-Chief and our reviewers. References Arbuzov, S.I., S.I., Ershov, V.V., Potseluev, A.A., Rikhvanov, L.P., 2000. Trace Elements in the Kuznetsk Basin Coals. Kemerovo, 248 pp. (in Russian). Arbuzov, S.I., Ershov, V.V., Rikhvanov, L.P., Usova, T.Yu., Kyargin, V.V., Bulatov, A.A., Dubovic, N.E., 2003. Trace Element Potential of the Minusinsk Basin Coals. Novosibirsk, 347 pp. (in Russian). Bethell, F.V., 1962. The distribution and origin of minor elements in coal. Brit. Coal Util. Res. Ass. Monthly Bull. 26, 401–429. Bragg, L.J., Oman, J.K., Tewalt, S.J., Oman, S.L., Rega, N.H., Washington, P.M., Finkelman, R.B., 1998. Coal quality (COALQUAL) database: version 2.0. Open-File Rep. U.S. Geol. Surv, pp. 97–134 (unpaginated CD-ROM). Eliseev, A.I., 1978. Formations of the Northern-Eastern Margin of the European Platform in the Paleozoan. Nauka, Leningrad. [“Science” Publ. House], 204 pp. (in Russian). Eliseev, A.I., , 1982. Comparative Analysis of the Platform Margins in the Paleozoan. Syktyvkar, Komi Sci. Center, 56 pp. Eskenazy, G., 1980. On the geochemistry of indium in coalforming process. Geochim. Cosmochim. Acta 44, 1023–1027. Eskenazy, G., 1982. The geochemistry of tungsten in Bulgarian coals. Int. J. Coal Geol. 2, 99–111. Eskenazy, G., 1987a. Rare earth elements in a sampled coal from the Pirin deposit, Bulgaria. Int. J. Coal Geol. 7, 301–314. Eskenazy, G., 1987b. Zirconium and hafnium in Bulgarian coals. Fuel 66, 1652–1657. Eskenazy, G., 1990. Geochemistry of tantalum in Bulgarian coals. Int. J. Coal Geol. 15, 137–149. Eskenazy, G., 1992. On the geochemistry of gold in Bulgarian coals. Geol. Balcan. 22, 47–58. Eskenazy, G., 1995. Geochemistry of arsenic and antimony in Bulgarian coals. Chem. Geol. 119, 239–254. Eskenazy, G., 2006. Geochemistry of beryllium in Bulgarian coals. Int. J. Coal Geol. 66, 305–315. Eskenazy, G., Mincheva, E., 1989. On the geochemistry of strontium in Bulgarian coals. Chem. Geol. 74, 265–276. Eskenazy, G., Delibaltova, D., Mincheva, E., 1994. Geochemistry of boron in Bulgarian coals. Int. J. Coal Geol. 25, 93–110. Gluskoter, H.J., 1965. Electronic low-temperature ashing of bituminous coal. Fuel 44, 285–291. Goldschmidt, V.M., 1935. Rare elements in coal ashes. Ind. Eng. Chem. 27, 1100–1102. Grigoriev, N.A., 2003. Average content of chemical elements in rocks composing the upper Earth's crust. Geokhimiya [Geochemistry] 7, 785–792 (in Russian). Ketris, M.P., Yudovich, Ya.E., 2002. Calculation procedure of the coal Clarkes. Lithogenes i geochimiya osadochnykh formatsiy Timano-Ural'skogo regiona. Syktyvkar, pp. 111–117 (Tr. In-ta geol. Komi nauch. tsentra UrO Ross. Acad. nauk, vyp. 111). [Proc. Inst. Geol. Komi Sci. Center, issue 111] (in Russian). Kortenski, J., Bakardjiev, S., 1993. Rare earth and radioactive elements in some coals from the Sofia, Svoge and Pernic Basins, Bulgaria. Int. J. Coal Geol. 22, 237–246. Krauskopf, K.B., 1955. Sedimentary deposits of rare metals. Econ. Geol. 50, 411–463. Puchkov, V.N., 1979. Batial Sequences of the Passive Geosynclinal Margins. Nauka, Moscow. [“Science” Pub. House], 260 pp. (in Russian). Ronov, A.B., 1980. Sedimentary Shell of the Earth. Nauka, Moscow. [“Science” Pub. House], 80 pp. (in Russian). Ronov, A.B., Yaroshevsky, A.A., Migdisov, A.A., 1990. Chemical Composition of the Earth's Crust and Geochemical Balance of Main Elements. Nauka, Moscow. [“Science” Pub. House], 192 pp. (in Russian). Salop, L.I., 1982. Earth's Geologic Development in Precambrian. Nedra, Leningrad. [“Entrails” Publ. House], 343 pp. (in Russian). Author's personal copy 148 M.P. Ketris, Y.E. Yudovich / International Journal of Coal Geology 78 (2009) 135–148 Seredin, V.V., Finkelman, R.B., 2008. Metalliferous coals: A review of the main genetic and geochemical types. Int. J. Coal. Geol. 70, 253–289. Seredin, V.V., Shpirt, M.Y., 1995. Metalliferous coals: a new potential source of valuable trace elements as by-products. In: Pajares, J.A., Tascón, J.M.D. (Eds.), Eighth ICCS Proc. Coal Science, vol. II. Elsevier, Amsterdam, pp. 1649–1652. Sozinov, N.A., Chistyakova, N.N., Kazantsev, V.A., 1988. Metalliferous Black Shales of the Kursk Magnetic Anomaly. Nedra, Moscow. [“Entrails” Publ. House], 149 pp. (in Russian). Swaine, D.J., 1990. Trace Elements in Coal. Butterworths, London. 278 pp. Tkachev, Ya.A., Yudovich, Ya.E., 1975. Statistical Processing of Geochemical Data: Methods and Problems. Nauka, Leningrad. [“Science” Publ. House], 233 pp. (in Russian). Valkovič, V., 1983a. Trace Elements in Coal, vol. 1. Chem. Rubber Co. Press, Boca Raton, Fla.. 210 pp. Valkovič, V., 1983b. Trace Elements in Coal, vol. 2. Chem. Rubber Co. Press, Boca Raton, Fla. 281 pp. Vassilev, S.V., Eskenazy, G.M., Tarassov, M.P., Dimov, V.I., 1995. Mineralogy and geochemistry of a vitrain lens with unique trace element content from the Vulche Pole coal deposit, Bulgaria. Geol. Balcan. 25, 111–124. Vine, J.D., Tourtelot, E.B., 1970. Geochemistry of black shale deposits – a summary report. Econ. Geol. 65, 253–272. Vinogradov, A.P., 1962. Average content of chemical elements in main types of igneous rocks of the Earth's crust. Geokhimiya (Geochemistry) 7, 555–571 (in Russian). Yudovich, Ya.E., 1972. Geochemistry of Coal Inclusions in Sedimentary Rocks. Nauka, Leningrad. [“Science” Pub. House], 84 pp. (in Russian). Yudovich, Ya.E., 1978. Geochemistry of Fossil Coals. Nauka, Leningrad. [“Science” Pub. House], 262 pp. (in Russian). Yudovich, Ya.E., 2003. Coal inclusions in sedimentary rocks: a geochemical phenomenon. A review. Int. J. Coal. Geol. 56, 203–222. Yudovich, Ya.E., Ketris, M.P., 1988. Geochemistry of Black Shales. Nauka, Leningrad. [“Science” Pub. House], 272 pp. (in Russian). View publication stats Yudovich, Ya.E., Ketris, M.P., 1994. Trace elements in Black Shales. Nauka, Ekaterinburg. [“Science” Pub. House], 304 pp. (in Russian). Yudovich, Ya.E., Ketris, M.P., 1997. Geochemistry of Black Shales. 1. Outline. 2. Bibliography and Index. Prolog, Syktyvkar. 212 pp. (in English). Yudovich, Ya.E., Ketris, M.P., 2002. Inorganic Matter of Coals. UrO RAN, Ekaterinburg. [Ural Division of the Russian Acad. Sci.], 422 pp. (in Russian). Yudovich, Ya.E., Ketris, M.P., 2004. Germanium in Coal. Komi Sci. Center Ural. Div. Rus. Acad. Sci., Syktyvkar. 216 pp. (in Russian). Yudovich, Ya.E., Ketris, M.P., 2005a. Arsenic in coal: a review. Int. J. Coal. Geol. 61, 141–196. Yudovich, Ya.E., Ketris, M.P., 2005b. Mercury in coal: a review. Pt. 1. Int. J. Coal. Geol. 62, 107–134. Yudovich, Ya.E., Ketris, M.P., 2005c. Mercury in coal: a review. Pt. 2. Int. J. Coal. Geol. 62, 135–165. Yudovich, Ya.E., Ketris, M.P., 2005d. Toxic Trace Elements in Coal. UrO RAN, Ekaterinburg. [Ural Division of the Russian Acad. Sci.], 655 pp. (in Russian). Yudovich, Ya.E., Ketris, M.P., 2006a. Сhlorine in coal: a review. Int. J. Coal. Geol. 67, 127–144. Yudovich, Ya.E., Ketris, M.P., 2006b. Selenium in coal: a review. Int. J. Coal. Geol. 67, 112–126. Yudovich, Ya.E., Ketris, M.P., 2006c. Valuable Trace Elements in Coal. UrO RAN, Ekaterinburg. [Ural Division of the Russian Acad. Sci.], 538 pp. (in Russian). Yudovich, Ya.E., Korycheva, A.A., Obruchnikov, A.S., Stepanov, Yu.V., 1972. Average contents of trace elements in fossil coals. Geokhimiya [Geochemistry] 8, 1023–1031 (in Russian). Yudovich, Ya.E., Ketris, M.P., Merts, A.V., 1985. Trace Elements in Fossil Coals. Nauka, Leningrad. [“Science” Pub. House], 239 pp. (in Russian). Zharov, Yu.N., Meytov, E.S., Sharova, I.G. (compilers), 1996. Valuable and Toxic Elements in Russian Raw Coals: a Reference Book. Nedra [“Entrails” Publ. House], Moscow, 239 pp. (in Russian).