Introduction

Apatite is a group of phosphate minerals containing high concentration of OH−, F− and Cl− in their crystals. The naturally occurring apatite has as general chemical formula Ca10(PO4)6(OH,F,Cl)2. Apatite is the tenth most abundant mineral on Earth and the most abundant naturally occurring phosphate [1, 2].

The uniqueness of apatite lies in the ability of its structure to accommodate and adapt to substitution in both cation and anion sites [1, 3, 4]. Rakovan et al. [5] reported the structural characterization of U6+ in apatite as the first step in the investigation of U and Th incorporation in its structure [6, 7]. The study of this group of minerals as potential solid-state radioactive waste repository was also reported in the literature [8, 9] and for this reason, many studies based on the sorption/adsorption of U and other radionuclides in the apatite structure have been carried out [10–13]. Apatite, a mineral fundamental in controlling the rare-earth and trace element variation in rocks [14], is valuable in the fission-track determination of rates and dates in geologic processes [15–17].

Francolite [18] is a particular species of the apatite group corresponding to carbonate apatite. The mineral was named after its occurrence at Wheal Franco, Devon, England. The information about the presence and role of carbonate in the structure of this mineral is still rather limited [19, 20]. Perdikatsis [21] studying pure francolite from the Epirus region (Greece) by X-ray diffraction (XRD), concluded that there is a phosphate substitution by carbonate groups. His results also agree with previous findings [22–24]. Schuffert et al. [25] tried to explain the substitution of phosphates by carbonates in natural francolites modelling the results obtained from the investigation of 37 samples of synthetic francolite with CO3 2−-concentration between 0.0 and 9.0 wt%. Infrared absorption spectra of a natural carbonate fluorapatite were presented in the literature in 1990 however without further chemical and X-ray diffraction information and data [26]. On the other hand, Ivanova et al. [27] and Fleet and Liu [28–31] presented Rietveld and single-crystal X-ray diffraction data describing a possible substitution mechanism of phosphate by carbonate ions in synthetic fluor- and hydroxy-apatite. Petkova and Yaneva [32] evidenced by infrared spectroscopy (IR), X-ray diffraction (XRD) and thermogravimetry/diffrential thermal analysis (TG/DTA) the isomorphic substitution of phosphate by carbonate groups in Syrian carbonate apatite upon mechanochemical activation and Kostova et al. [33] the positional redistribution of carbonate ions in the structure of sedimentary apatite as effect of high energy milling. The role of the water presence in the structure of francolite is also rather controversial and not completely understood. According to literature, the association of water with apatites may be a key factor in the preference of syn-thetic low temperature apatites for B-type carbonate substitution for phosphate [34, 35]. On the other hand, for particular species of apatite, such as Ca10(PO4)6(OH)2 or Sr10(PO4)6F2, presenting nearly total B-type carbonate substitution, there is no correlation between the number of water molecules in the channels and the weight percentage of carbonate. This lack of correlation would be expected only when there is no competition between water and carbonate in the apatite channels [36].

The presence of carbonate apatites as mineral component of vertebrate bones, teeth and kidney stones is also still quite controversial and not completely clarified.

The objective of this study was to investigate the presence of uranium in organic-rich phosphatized limestones of Epirus region (NW Greece) and especially its association with apatites and francolite crystals as well as with organic matter. It should be noted that there is no generally accepted structural study of the U location in naturally occurring apatites and in particular in francolites. The results of the study are expected to contribute to a better understanding of the geochemistry and migration/immobilization of uranium into the geological environment of NW Greece and additionally to the elucidation of the redox phenomena taking place.

Experimental part

Sampling area and treatment of the samples

The sampling region was mainly selected on the basis of previous unpublished internal reports by the Greek Atomic Energy Commission (GAEC) and Institute of Geology and Mineral Exploration (IGME) concerning radiometric irregularities in the Mesozoic sedimentary rocks of NW Greece and particularly in the mountainous Epirus region [37] The region under investigation, which belongs to the Ionian geotectonic zone of NW Greece (Epirus), is shown in Fig. 1 [38, 39]. The sedimentary formation, where the samples were collected is an organic rich brecciated phosphatized limestone. This particular formation extends in 2 km approximately along the road of the region of Perivleptos, which is located 10 km N—NW of the city of Ioannina. The sampling (in total six samples of 1–2 kg weight each) was performed from the outcrop of the first appearance of the formation to its end. Specifically, the sampling locations were selected by surveying the suspected sedimentary formations by means of a portable radiation counter with NaI(Tl) detector (Canberra-MCB2). The specimens presenting enhanced radiation consisted of brecciated phosphatized limestones and were partially rich in organic matter.

Fig. 1
figure 1

Location of the studied region (modified after Robertson et al. [40])

Full size image

The samples were crushed using a jaw crusher, pulverized by milling and washed with water. Heavy mineral separation was carried out in four steps in order to obtain concentrates of apatite crystals. In the first step, heavy liquid separation was performed using 80 % sodium polytungstate (LST Fastfloat, density: 2.8 kg/m3). This step was followed by magnetic separation of the >2.8 kg/m3 fraction by means of a Frantz magnetic separator. The non-magnetic fraction was further processed by a solution of methylene iodide and acetone (density: 3.2 kg/m3) in order to separate the apatite crystals and finally by pure methylene iodide (density: 3.3 kg/m3) to receive the zircon (ZrSiO4) microcrystals.

Sample characterization

The chemical characterization of the separated crystals was performed on a JEOL JXA-8230 Superprobe Electron Probe Microanalyzer at the Department of Petrology and Metalogenesis of the Institute of Geosciences and Sciences of the Paulista State University (Brasilia). Energy Dispersive Spectra (EDS) were acquired with a silicon drift detector with energy resolution (FWHM) of 129 eV for Mn-Kα, at 3000 cps; Wavelength Dispersive Spectroscopy (WDS) qualitative scans were performed with LDE1, TAP, PETL and LIFH crystals and dwell time of 60 ms in position steps of 50 µm. The standardization was performed measuring the FKα, CaKα and PKα in F-apatite, SiKα in wollastonite, UMα in UO2, ClKα in sodalite and SKα in pyrite and using acceleration voltage of 15 kV, current of 20 nA, intergration time of 10 s for peak and 5 s for background. For WDS quantitative analyses acceleration voltage was applied at 15 nA in order to minimize destruction of the crystals under investigation. The beam diameter was 10 µm.

Raman spectroscopy measurements were performed at the University of Heidelberg using a Horiba iHR320 Raman Spectrometer, equipped with a 532 nm Laser (green) and a Horiba Superhead as optional confocal optic.

Fission track measurements

The apatite and francolite grains were embedded in epoxy resin, cut and polished in order to reveal their internal surface. Uranium free mica foils were attached to the samples, which were irradiated by neutrons, along with CN5 neutron dosimeter glasses, at the FRM II high flux research reactor of the Technical University of Munich at Garching. After the irradiation and decay of the induced radioactivity, the mica foils were etched for 20 min by 48 % HF in ambient temperature (20 °C). The fission tracks were then visualized by optical microscopy.

Synchrotron radiation measurements

Synchrotron radiation (SR) micro-X-ray fluorescence (μ-XRF) and X-ray absorption near edge structure (μ-XANES) spectra were obtained using powders and polished sections in the SUL-X beamline of the Laboratory for Environmental Studies of the ANKA facility (Karlsruhe Institute of Technology, Germany) [41]. The sample area of the thin polished section of the phosphatized limestone to be analyzed by μ-ΧRF was selected using an optical microscope. The intensities of the Ca, P, S, Y and U fluorescence emission lines excited by 17.5 keV X-rays were simultaneously determined by a 7-element Si(Li) solid state detector. On areas with elevated uranium concentration μ-XANES data were recorded at the UL 3-edge (about 17.170 keV). The powder samples measured as pellets were prepared by mixing the sample with a binding agent (cellulose). The uranium minerals uraninite (UO2) and shoepite [UO2(OH)2] as well as Durango fluorapatite were used as reference materials.

Results and discussion

The geological formation under investigation seems to be a matrix supported poorly sorted monomictic breccia produced by rock fragmentation due to faulting or other tectonic processes. The matrix of the breccia consists primarily of the cementing material, but it also contains sand and/or silt sized clasts cemented together among the coarser clasts. According to previous work [37], the cement binding the clasts consists of black organophosphate material (organic matter and apatite) and calcite.

Τhe separated inorganic phases were mainly calcite, apatites and francolites (carbonate apatites). The latter had optically different appearance than the classic apatite and, more specifically, appeared as opaque crystals without a specific shape showing different shades of brown. The francolites were identified by Raman spectroscopy combined with SEM/EDS microanalysis. The Raman spectra obtained (Fig. 2) were characterized by bands belonging to both carbonates and phosphates groups [42, 43].

Fig. 2
figure 2

Raman spectrum of francolite

Full size image

The brown color of the francolite crystals was probably due to a coating created by the interaction of their surface with the organic matter [44, 45] or originating from the high participation of carbonate part in its lattice. The microprobe analysis of the crystals showed that the mineral was a uranium-enriched sulfur- and sodium bearing carbonate fluorapatites. The presence of uranium in the francolite crystals was additionally evidenced by the enhanced number of fission tracks revealed in the attached uranium free mica foils after neutron irradiation and etching (Fig. 3). The number of fission tracks originating from the francolite crystals could not be quantified because of their enhanced uranium content. This finding confirming the uranium presence in sedimentary francolite crystals has not been, to our knowledge, previously reported in the literature.

Fig. 3
figure 3

Apatite (a) and francolite (c) crystals and the corresponding induced fission tracks (b) and (d) in the attached uranium free mica foils after neutron irradiation

Full size image

On the other hand, only few studies on the chemical composition of francolites have appeared so far in the literature [22, 46, 47]. The elemental mapping of the studied francolite performed by microprobe showed that Ca, P, S, and Na were not homogenously distributed in the crystals (Fig. 4).

Fig. 4
figure 4

Microprobe elemental maps of a francolite crystal

Full size image

SR-based measurements of the U-bearing organic rich phosphatized limestone were performed in order to obtain information about the uranium oxidation state in the samples. SR μ-XRF maps were used to examine the distribution of uranium in the samples in order to select the target areas for μ-XANES measurements. The SR μ‐XRF study (Fig. 5) revealed that U was accumulated, along with P, Y and most likely S, in certain areas of the dark organic‐rich part of the rock containing, according to previous XRD- and SEM/EDS measurements, enhanced concentrations of apatites and amorphous organic matter. The areas with very low U and high Ca concentrations mainly consisted of calcite and, in a lesser extent, of dolomite.

Fig. 5
figure 5

Optical image of the polished section of the U-bearing limestone (upper left) and qualitative SR μ-XRF elemental maps. The dark side corresponds to the organic-rich component, where (U and P) is mostly accumulated, while the bright one to Ca-carbonate

Full size image

The position of the UL 3‐edge (2p3/2 orbital) in preliminary μ-ΧANES spectra indicated the presence of U4+ associated to apatite‐group minerals and/or organic matter. Figure 6 presents the position of the uranium edge of the investigated phosphatized limestone in comparison with those obtained for uraninite (UO2) and shoepite [UO2(OH)2] used as reference materials. Under oxidizing conditions, typical of surface waters and groundwater systems, the aqueous speciation of U6+ determine the partitioning of uranium onto mineral surfaces or in their structure by the reduction of U6+–U4+. Oxidized U6+ is highly soluble as the uranyl ion UO2 2+, whereas the solubility of U4+ is controlled by insoluble oxides such as uraninite. Oxidation state analyses contribute to a better understanding of the geochemistry and migration/immobilization of uranium into the geological environment and additionally to the elucidation of the redox phenomena taking place.

Fig. 6
figure 6

Preliminary UL 3-edge μ-XANES spectra

Full size image

The ability of apatite to accommodate several foreign elements, including actinides, in its lattice is well-known and of great environmental importance. Monovalent (Na+, K+), divalent (Sr2+, Pb2+, Ba2+, Mn2+, Cd2+), trivalent (REE3+), as well as tetravalent (Th4+, U4+) and hexavalent (U6+) cations have been reported to substitute calcium in the apatite structure [3, 5–7, 48]. Rakovan et al. [5], observed, analyzing extended X-ray absorption fine structure (EXAFS) data, the Ca-substitution (Ca1 site) by U6+ in synthetic apatite grown at 1350 °C. In addition, Luo et al. [6] reported, also based on XRD and XANES data, that U4+ can also be incorporated in Ca-sites of U-doped synthetic fluorapatite. Structure refinements indicated that, in this case, U substitutes Ca almost exclusively into the Ca2 sites with site occupancy ratios UCa2/UCa1 ranging from 5.00 to 9.33. The occurrence of U4+ in calcite, formed at low temperatures, is considered to be less possible.

Combining the SR μ-XRF maps of the bulk rock with those of the francolites carried out by SEM/EDS—could be concluded that the uranium was concentrated in the organic rich part of the phosphatized limestones and more specifically in the francolite structure. The positive correlation between uranium and phosphor additionally strengthen this conclusion [49, 50]. Previous attempts to explain the presence of uranium in sedimentary apatite (e.g. [51]) were based on the similarity of the ionic radii of the elements and on the assumption that uranium may partly substitute calcium in the carbonate apatite lattice. At this point, it should also be emphasized that all the studies about uranium in apatite structure concerned fluorapatites and chloroapatites [5, 6] and there were no specific studies about uranium in carbonate apatite/francolite ([Ca5(PO4,CO3)3(F,OH)]) occurring, along with fluor- and hydroxy-apatite.

Conclusions

The organic rich limestones of NW Greece (Epirus region) exhibit radioactivity due to the enhanced uranium content. Organic matter, apatites, francolites (carbonate apatites) and calcite are the main phases of the rock. Francolites were identified by combining microprobe data with Raman spectroscopy. μ-XRF elemental mappings applied to polished thin section of the sample showed that the uranium exclusively existed in the organic rich part (apatites + francolites + organic matter). Moreover, uranium was positively correlated with the P, while S was also present. The density of homogenously distributed fission tracks highlighted the high U content. μ-XANES study revealed that the uranium existed in the U4+ oxidation state. Furthermore, the fission track results showed that francolites could be a perfect candidate for thermochronology dating due to their high U content.