GC-60-2.indb 1. INTRODUCTION From the early 1960’s to the late 80’s, extensive explo- ration for uranium ore took place in the former Yugo- slavia, including Croatia (BRAUN, 1984). Prospecting activities by the Croatian Geological Survey identified several locations in the Slavonian Mts. (Croatia) with an increased level of radioactivity (BRAUN et al., 1983), particularly in the area of Mts. Psunj and Papuk. The Carbon age of the metasediments of Radlovac creek – known as the Radlovac series – has been determined from the fossil flora (BRKIĆ et al., 1974). The position of this series, which includes rocks with higher back- ground radioactivity, is shown in Fig. 1. The tectonic setting and structural fabric of these and associated rocks have been described by JAMIČIĆ (1976, 1979). The Radlovac series is a metaschist formation, and an increased level of radioactivity has been identified in Uranium Minerals in the Radlovac Series Metasediments at Mt. Papuk, Croatia Stjepan ŠĆAVNIČAR1, Vladimir BERMANEC1, Goran KNIEWALD2, Delko BARIŠIĆ2 and Višnja OREŠČANIN2 several, apparently unrelated locations. This is also true for deeper strata (samples from the prospecting bore- holes). A detailed synthesis of this research was given by JURKOVIĆ (2003). In 1984 the Institute of Mineralogy and Petrology of the Faculty of Sciences in Zagreb attempted X-ray dif- fraction analysis of the uranium-bearing minerals in the Radlovac metaschists. Due to the complex mineralogi- cal structure of the rocks, the separation methods avail- able at that time (magnetic separation, granulometry, high-density liquids) did not yield satisfactory results, and the minerals responsible for the increased radioac- tivity were not unambiguously identified, even though both the light and heavy mineral fractions as well as composite rock samples were analyzed. However, it was inferred that the activity was due to the presence of uranyl arsenates of calcium, copper, iron or bari- um (ŠĆAVNIČAR, 19843). Therefore, the aim of the present study is to determine the uranium-bearing min- erals of those samples, using contemporary and state- of-the-art XRD, EDXRF, SEM with EDX spectrometer and gamma spectrometric methods. 2. MATERIALS AND METHODS Some of the original samples, which were analysed in 1984, were still available to the authors. These sam- ples (which show the strongest background activities) were taken from low-grade metamorphosed sedimen- tary rocks of the Radlovac series in Mt. Papuk, which crops out near the Ninkovac creek. The host rock has a schistlike structure and contains muscovite, quartz, pyrophyllite, chlorite, albite, microcline, chloritoid, and, occasionally, calcite and dolomite. This mineral assem- blage indicates a sequence of low-grade metamorphism. Uranium, originating from the surrounding weathering zones, was deposited in the form of U-oxysalts in these rocks. The radioactivity of the analyzed samples was determined by gamma spectrometry. Prior to gamma- Geologia Croatica 60/2 165–171 6 Figs. 2 Tabs. ZAGREB 2007 Key words: Uranium mineralization, Meta-torbernite, Meta-uranospinite, Meta-zeunerite, Mt. Papuk, Rad- lovac series, Epigenetic–sedimentary mineralisation. 1 University of Zagreb, Faculty of Science, Institute of Mineralogy and Petrology, Horvatovac bb, HR-10000 Zagreb, Croatia; e-mail: vberman@public.carnet.hr 2 Rudjer Bošković Institute, Department of Marine and Environmen- tal Research, POB 180, HR-10002 Zagreb, Croatia Abstract Applying a combination of different methods – X-ray diffraction, scanning electron microscopy, gamma-spectroscopy and X-ray spec- troscopy, a suite of uranium minerals, meta-torbernite, meta-uranos- pinite and meta-zeunerite was identified in metasediments of the Rad- lovac series at the Mt. Papuk area, Croatia. The accessory minerals galena, zircon, rutile, chalcopyrite and cuprite, as well as an unidenti- fied Ni-bearing phase are also present. The mineral assemblage indicates a sequence of epigenetic and supergene processes affecting the Radlovac series. 3 ŠĆAVNIČAR, S. (1984): Rentgenografska analiza mineralnog sas- tava radioaktivnih uzoraka stijena iz Slavonije.– Unpubl. report, Archive of the Croatian Geological Survey. 166 Geologia Croatica 60/2 spectrometry measurements, samples were placed into counting vessels of known geometry, sealed and stored for at least 4 weeks to allow ingrowth of gaseous 222Rn (half-life of 3.8 days) and its short-lived decay prod- ucts to equilibrate with the long-lived 226Ra precursor in the sample. At the end of the ingrowth period, sam- ples were counted on an InSpector gamma spectrom- eter (‘Canberra’ portable HPGe detector coupled with a 8192 channel analyser). The system was calibrated using standard reference materials supplied by the International Atomic Energy Agency (stream sediments IAEA–306, IAEA–313 and IAEA–314). Counting time was 80,000 seconds and recorded spectra were analyzed using the Canberra GENIE 2K software. The activi- ties of 40K were calculated from the 1460.75 keV-peak. Activities of 226Ra were calculated from the 609.4 keV peak of its 214Bi progeny as well as 228Ra activities from 911.1 keV-peak of its 228Ac progeny. Activities of 235U were calculated from 186 keV peak (after subtraction of the overlapping 226Ra peak, which was calculated previously from 214Bi). Activities of 238U were calcu- lated from the 235U activities assuming the 235U/238U activity ratio of 0.0460. This procedure was detailed by BARIŠIĆ (1989). 40K activity of 309.3 Bq/kg corre- sponds to 1% of total potassium, 238U activity of 12.44 Bq/kg corresponds to 1 ppm of uranium, while the 232Th activity of 4.06 Bq/kg corresponds to 1 ppm of thorium (BARIŠIĆ, 1996). Samples with the highest activities were used for XRD investigations. The diffraction patterns of all sam- ples were collected on a Philips ‘Xpert diffractometer using CuKα radiation. Powder diffraction data were collected in the range of 4–63° 2Θ, and a step size of 0.02°. The obtained d-spacings and relative intensities of the patterns obtained were compared with reference values (JCPDS 01–077–0124, 01–086–1787, 00–018– 0309). The diffractograms were analyzed using the Philips proprietary computer code enabling a precise pattern analysis and the identification of very weak dif- fraction peaks. For elemental analysis, each sample was disaggre- gated in an agate mortar, pressed into pellets and meas- ured by energy dispersive X-ray fluorescence method (EDXRF; OREŠČANIN et al., 2006, 2007). Samples were irradiated by X-rays generated from a 109Cd annu- lar source. The incident angle was 50°. Detection of the characteristic X-ray radiation from the sample was done with a liquid-N colled Si(Li) detector (Canberra) with the following characteristics: detector size=30 mm2, Si thickness=3 mm. Be window=25 µm, FWHM for 5.9 keV 55Fe 165 eV, the emerging angle was 74°, and the distance was 1.5 cm. Spectra were collected by Genie – 2000 software (Canberra, Meriden, CT, USA). Collect- ing time was 10,000 s. Spectral data were analyzed by Canberra’s WinAxil software. Elemental concentrations in the thick targets were calculated using the ‘Funda- mental parameters’ method from the WinFund package, with IAEA SRM SL–1 and SOIL 7 as the standards. The concentration (wi) of each element (i) was calcu- lated according to the following formula: wi = Nij / (Kij Dij FPij Aij,m Hij,m I i0 t), where Nij is the net peak area of the characteristic X- ray line j; Kij – calibration constant calculated on the basis of the standard reference material; Dij – detector efficiency at Eij energy; FPij – number of fundamen- tal parameters; Aij,m – attenuation of the X-rays due to matrix (m); Hij,m – enhancement of the X-rays due to the matrix (m); I – intensity distribution of the charac- teristic lines and of the continuum in the excitation; i0 – factor taking into account the overall intensity of the source (the activity of the radio isotope); t – measure- ment time. Scanning electron microscopy (SEM) of the sam- ples was done using a Tescan SE microscope VEGA TS5136 equipped with Oxford detectors for secondary and back-scattered electrons. The operating voltage of the SEM was 20 kV, the current was 20 nA. The micro- scope is capable of elemental microanalysis through an EDS detector and the INCA 250 software package. Fig. 1 Location map of the Radlovac series in the Mt. Papuk area. 167 3. RESULTS AND DISCUSSION 3.1. X-ray diffraction X-ray diffraction analysis was run on available samples and the presence of minerals belonging to the meta- autunite group was verified. As it was impossible to obtain pure mineral phases, there were numerous over- laps of diffraction peaks of the uranium minerals and other mineral phases such as quartz, micas and feld- spars. However, the meta-autunite mineral group could be unambiguously identified based on their strongest lines. The unit cell could not be calculated from this data set, which – even if available – would not provide information on the mineral species due to the com- plex crystallochemical characteristics of the group. As in the earlier investigation (ŠĆAVNIČAR, 19843), the samples investigated here contained muscovite, quartz, pyrophyllite, chlorite, albite, microcline, chloritoide, and occasionally dolomite and calcite. These minerals account for all diffraction peaks identified in the XRD patterns, except for one line of variable intensity around 8.8 Ǻ. This peak belongs to the most intensive line in the powder XRD pattern of the meta-autunite mineral group. This line was also identified in most of the sam- ples analyzed previously, but its intensity varies from barely discernible to conspicuous. The powder pattern of the chloritoid minerals presents a major difficulty in the identification proc- ess of uranium minerals present in the samples. Some chloritoid group minerals have a strong diffraction line (of 10–20% intensity) at 8.9 Ǻ, and its resolution from the line at 8.8 Ǻ, belonging to the meta-autunite min- erals, is difficult, if at all possible. Chemical analysis would probably have provided the requisite information for mineral identification, but such an analysis was not feasible due to the substantial intergrowth of the various mineral phases present in the samples. 3.2. Scanning electron microscopy investigation For the SEM analysis of the Radlovac series rocks, the samples were only sputtered with carbon and scanned for surface roughness (i.e. the surface topography of the sample). The different mineral phases were identi- fied based on the atomic numbers of the elements in the mineral crystal structures. In this way, the heavy ele- ments could be identified very easily. In the samples collected for the previous investigation, SEM scanning revealed the presence of numerous aggregated frag- ments of uranium minerals. The fragments are hypidi- omorphic to idiomorphic with a platy appearance. The plates grow in a subparallel arrangement and a tetrago- nal symmetry is clearly visible in octagonal sections. The plates have nearly perfect cleavage parallel to the base (Fig. 2a and b). The crystals are 2–20 μm in length, their aggrega- tions up to 0.5 cm in size. Smaller crystals have rec- tangular (Fig. 3a), the larger ones octogonal sections, which can be diagonally elongated (Fig. 3b). Several, rounded fragments of zircon were identified in sample ZMAGO–1–4–3a. 3.3. Energy dispersive X-ray analysis The energy dispersive X-ray analysis (EDX) showed that the principal elements in the uranium mineral struc- tures are U, As and Cu in one group of samples, and U, P, Cu and Ca in the other group. Minor amounts of Fe, Ba and Si were identified in the uranium minerals. Sup- ported by XRD investigations, it can be concluded that Fig. 2 a) Polished sample RA–4 with visible uranium mineralization (white grains are uranium minerals, grey grains are baryte crys- tals); b) A more detailed view of Fig. 2a showing meta-zeunerite platelets growing next to baryte. a b Šćavničar et al.: Uranium Minerals in the Radlovac Series Metasediments... 168 Geologia Croatica 60/2 several uranium minerals are present in the samples – meta-torbernite [Cu(UO2)2(PO4)2]·8H2O, meta-urano- spinite [Ca(UO2)2(AsO4)2]·8H2O, and meta-zeunerite [Cu(UO2)2(AsO4)2]·8H2O. The EDX spectra show that the meta-torbernite is a single phase, while there seems to a solid solution series between meta-uranospinite and meta-zeunerite resulting in possible variations in the composition of discrete crystal fragments. The aggregates of these uranium minerals are part- ly interspersed with the other rock-forming minerals, resulting in a poikiloblastic texture (porphyroblasts which are riddled with finer grained inclusions of other minerals) of the uranium minerals. Small inclusions of baryte within the uranium minerals are common (Fig. 4a). The baryte fragments have typical morphologies with rhombic sections. They are 2–20 µm in size and formed by several subparallel plates. Individual baryte crystals associated with other minerals are also repre- sented (Fig. 5), but their incorporation into the uranium minerals is more common (Fig. 4a). Small amounts of silica have been identified in the baryte. In some samples (i.e. sample ZMAGO–c) a copper oxide phase (most probably cuprite) was observed. It occurs in the form of needle-like crystals up to 200 µm in length, and below 1 µm in thickness (Figs. 6a and b). The primary form of uranium mineralization has not been identified, but the presence of chalcopyrite was established. This chalcopyrite could have been the primary source of copper for the deposition of cuprite, but also provided copper for meta-torbernite and meta- zeunerite. A fragment of galena, ca. 15 µm in size was observed next to cuprite in one of the samples. A nick- el-containing mineral phase was also identified. Fig. 3 a) Small crystals of meta-torbernite (2–10 µm) showing rectangular sections; the larger crystals have octagonal sections (sample RA–6); b) Crystals of meta-zeunerite showing similar morphology as meta-uranospinite (sample RA–6). a b Fig. 4 a) Meta-uranospinite with an incorporated baryte inclusion (sample RA–6). The irregular grey spots are probably areas of dehydration with a partial loss of crystal water. Tetragonal meta-uranospinite crystals have octagonal sections. Different intensities of gray probably indicate Ca–Cu exchange; b) EDX spectrum of the area where meta-uranospinite dominates over meta-zeunerite (sample RA–6). a b 169Šćavničar et al.: Uranium Minerals in the Radlovac Series Metasediments... 3.4. Energy dispersive X-ray fluorescence analysis (EDXRF) Table 1 presents concentrations of the major, minor and trace elements measured in three samples from Radlo- vac creek. XRF analysis of sample ZMAGO–1 indi- cated a substantial concentration of uranium at 0.15%. The PIKO sample was characterized by extremely high concentrations of U, Ba, As, and Cu, more than hundred times (U, Cu, Ba) and more than two thousand times (As) higher when compared to normal values that could be obtained in sandstone sediments. Arsenic was also highly enriched in the ZBZ material while U and Ni were elevated approximately 20 and 10 times respec- tively compared to background values (ZMAGO sam- ple). Elevated concentrations of K, Ca, Fe, and Ti have also been established in this sample. The presence of all these elements can be related to the formation of ura- nium minerals, although Ba has only been observed in baryte and not in the uranium mineral phases. Cop- per forms meta-torbernite and meta-zeunerite, but also cuprite (and the primary chalcopyrite – Figs. 6a and b). The source of titanium is the small, unevenly distribut- ed needles of rutile up to 10 µm in length and less than 1 µm thick (sample ZMAGO–1–4–3a). It is interest- ing to note that the presence of discrete nickel mineral Fig. 5 a) Discrete baryte crystals 10x18 µm in size. The uranium minerals were formed mainly around baryte grains, but baryte crystals also occur separately (sample RA–6). No uranium mineral grains are visible on this image; b) A more detailed view of Fig. 5a. a b Element Sample PIKO ± ZBZ ± ZMAGO ± K (%) 1.10 0.07 1.37 0.10 1.88 0.06 Ca (%) 1.29 0.07 0.18 0.03 0.34 0.02 Ti (%) 0.060 0.010 0.200 0.02 0.268 0.008 Fe (%) 0.764 0.002 1.644 0.004 1.589 0.001 U (ppm) 1500 200 230 40 16 3 Pb (ppm) 46 1 49 2 12 0.5 Ba (ppm) 3900 300 390 30 630 40 As (ppm) 2200 200 2000 200 n.d. Rb (ppm) 6.7 0.7 68.1 1.0 123.8 0.5 Sr (ppm) 170.3 0.9 28.4 0.7 51.0 0.3 Y (ppm) 27.8 0.8 138 2.0 40.4 0.5 Zr (ppm) 131.9 0.7 525.0 2.0 432.2 1.0 Mn (ppm) 320 20 620 40 330 10 Co (ppm) 3.8 0.1 7.1 0.3 5.8 0.1 Ni (ppm) 15 5 140 10 19 3 Cu (ppm) 1550 60 13 2 15.7 0.9 Zn (ppm) 167 3 42 3 39 1 Table 1 Elemental concentrations deter- mined by EDXRF system in three samples from the Radlovac creek. 170 Geologia Croatica 60/2 phases was established, even though the concentration of Ni is rather low. Apart from its mineralogical significance, the identi- fication of the uranium mineralization in the Mt. Papuk area is also important for a better understanding of the geology and formation of the Radlovac series. Unfor- tunately, the intricately intergrown uranium mineral fragments cannot be separated without substantial con- tamination by other phases, so the characteristics and content of structural (crystal) water cannot be measured with sufficient precision. The amount of crystal water in the uranium minerals is variable and changes with dehydration and rehydration, depending on the physi- co-chemical parameters of the environment in which it was formed, and/or whether it is currently in thermody- namic equilibrium. Apart from the obvious significance of the partial pressure of water vapour, an important parameter of the reaction pathways leading to uranium mineral formation is the redox potential of the environ- ment (BERMANEC et al., 2005). The identification of the oxidation state of uranium and the amount of crystal water in the uranium minerals would be of great impor- tance for a better understanding of the conditions in which the Radlovac series was deposited and the proc- esses which have impacted on the chemical and minera- logical diagenesis of this rock series. 3.5. Gamma spectrometry The activities of 40K, 232Th, 226Ra and 238U (Bq/kg) as well as the 238U/226Ra activity ratio are presented in Table 2. Concentrations of naturally occurring 40K range over an order of magnitude, indicating signifi- cant differences in the total potassium content (2.78% of potassium in sample RA–1 against 0.22% of potas- sium in sample ZMAGO). Similar differences in ura- nium activities/concentrations were also found. Such differences indicate either compositional variation in the main minerals of the analyzed sandstones and/or intensive weathering processes. This is corroborated by the measured 238U/226Ra activity ratio, which varies between 0.156 (in sample RA–1) and 1.224 as found in sample ZMAGO. In their early research of the Radlovac series, BRAUN et al. (1983) maintained that uranium was not in equilibrium with its decay products. A high content of 226Ra (a relatively short lived member of 238U natural decay chain) in comparison with uranium is very prob- ably the result of recent (up to a few hundred thousand years), uranium leaching processes. However, 238U/226Ra activity of 1,224 units in sample ZMAGO indicates a zone where uranium was recently deposited with urani- um enrichment. It seems that the highest uranium con- tent of approximately 150 ppm, in sample RA6, is only ‘the remaining’ uranium after intensive leaching. On the basis of recent 226Ra activity in that sample, it could be supposed that about 50% of the uranium content was weathered in the last few hundred thousand years. The uneven distribution of uranium minerals and accessory minerals in the investigated samples indi- cates that the protolyte was of inhomogenous composi- Fig. 6 a) Bundle of needle-like crystals of copper oxide, probably cuprite (sample RA–6); b) Radiating crystal aggregates of copper oxide, probably cuprite (sample Zmago–c). a b Sample 40K 232Th 238U/226Ra 226Ra 238U RA1 860.9±13.5 53.3±2.1 0.156 892.2±4.9 139.0±12.3 RA6 276.0±12.7 15.7±2.0 0.487 3 913±11 1 906±30.0 ZMAGO 68.6±6.5 15.4±2.3 1.224 813.4±5.8 995.4±20.0 Table 2 Activities of 40K, 232Th, 226Ra, 238U (Bq/kg) and the 238U/226Ra activity ratio in the analyzed samples. 171Šćavničar et al.: Uranium Minerals in the Radlovac Series Metasediments... tion. This is further corroborated by the fact that sev- eral uranium mineral phases have formed, the phases not being fully separate but with variable composition within zoned mineral grains. All such phases appear to be in thermodynamic equilibrium. It was not possible to fully describe the zonations, since the thickness of such grains is very small. 4. CONCLUSIONS The combination of methods used in this study enabled unambiguous determination of the U-bearing minerals of the Radlovac series as meta-torbernite, meta-uranos- pinite and meta-zeunerite. These minerals have not been previously identified in Croatia. Further research will be devoted to identification of the nickel-bearing mineral phases and the determination of the oxidation state of the uranium ion in the uranium minerals – if appropriate samples become available, in which the separation of discrete uranium mineral phases will be possible. Acknowledgements This work was supported by the Ministry of Science, Education and Sports of the Republic of Croatia, under Grants No. 098–0982934–2713, 098–0982934–2715, 119–0000000–1158 and 098–0982934–2742. 5. REFERENCES BARIŠIĆ, D. (1989): Određivanje 235U and 238U gama-spek- trometrijskom metodom pri energijama oko 186 keV [The determination of 235U and 238U by gamma-spectrometry method at energies around 186 keV – in Croatian].– Pro- ceedings of 15th Symposium of the Yugoslav Society for Radiation Protection (JDZZ), Prishtina, 91–94. BARIŠIĆ, D. (1996): Dose rate conversion factors, soil thick- ness and their influence on natural background dose rate in air above carbonate terrains.– Journal of Environmental Radioactivity, 31/1, 51–70. BERMANEC, V., WEGNER, R., KNIEWALD, G., RAKVIN, B., PALINKAŠ, L., RAJIĆ, M., TOMAŠIĆ, N. & FURIĆ, K. (2005): The role of uranium(V) ion in the chemical composition of meta-autunite from pegmatites of Quin- tos de Baixo, Brazil.– Neues Jahrbuch für Mineralogie, Abhandlungen, 181, 27–38. BRAUN, K. (1984): Istraživanje urana u SR Hrvatskoj (1978–1983) [Prospecting for uranium in SR Croatia (1978–1983) – in Croatian].– Savjetovanje “Rezultati dosadašnjih i pravci daljnjih istraživanja nukelarnih siro- vina”, Zbornik radova, 62–98, Beograd. BRAUN, K., DRAVEC, J., SLOVIĆ, V., CRNOGAJ, S., VAL- KOVIĆ, V. & MAKJANIĆ, J. (1983): Pojava mineraliza- cije urana na Papuku i Krndiji [Uranium mineralisation at Mt. Papuk and Mt. Krndija – in Croatian].– Geološki vjesnik, 36, 111–115. BRKIĆ, M., JAMIČIĆ, D. & PANTIĆ, N. (1974): Karbonske naslage u Papuku (sjeveroistočna Hrvatska) [Carbonife- rous deposits at Mt. Papuk (NE Croatia) – in Croatian].– Geološki vjesnik, 27, 53–58. JAMIČIĆ, D. (1976): Structural fabric of the metamorphosed rocks of Mt. Krndija and the eastern part of Mt. Papuk.– Bulletin scientifique JAZU, Section A, 21, 2–3. JAMIČIĆ, D. (1979): Prilog poznavanju tektonskih odnosa Papuka i Krndije [Contribution to the knowledge of tec- tonic relationships of Mts. Papuk and Krndija – in Croati- an].– Zbornik radova, 4. godišnji znanstveni skup Sekcije za primjenu geologije, geofizike i geokemije Znanstvenog savjeta za naftu HAZU, 199–206. JURKOVIĆ, I. (2003): Metalogenija južne Tisije – Mosla- vačka gora, Psunj, Papuk, Krndija [Metalogenesis of southern Tisija – Moslavačka gora, Psunj, Papuk, Krndija – in Croatian].– Rudarsko-geološko-naftni zbornik, 15/1, 1–17. OREŠČANIN, V., MIKELIĆ, L., LOVRENČIĆ, I., BARIŠIĆ, D., MIKULIĆ, N. & LULIĆ, S. (2006): Environmental contamination assessment of the surroundings of the ex- ferrochromium smelter Dugi Rat, Croatia.– J. Environ. Sci. Health A, 41/11, 2547–2555. OREŠČANIN, V., MIKELIĆ, L., SOFILIĆ, T., RASTOV- ČAN-MIOČ, A., UŽAREVIĆ, K., MEDUNIĆ, G., ELEZ, L. & LULIĆ, S. (2007): Leaching properties of electric arc furnace dust prior/following alkaline extraction.– J. Environ. Sci. Health A, 42/3, 323–329. Manuscript received July 13, 2007. Revised manuscript accepted November 23, 2007. 172 Geologia Croatica 60/2