2020 | 73/1 | 29–48 | 14 Figs. | 12 Tabs. | www.geologia-croatica.hr Journal of the Croatian Geological Survey and the Croatian Geological Society 1. INTRODUCTION There are several bentonite deposits and occurrences in Croatia and neighbouring countries formed by alteration of volcanic ash, in different geological settings with ages ranging from the Late Jurassic to the Middle Miocene. Some of these deposits were exploited in the past, and with technological advancements and new applications emerging, some of them could be economically viable again. Bentonite is used for numerous purposes, most commonly in drilling fluids, metal castings, civil-engineering, as pet-waste adsorbents, desic­ cants, environmental sealants etc. Other uses of bentonite, includ­ ing the pharmaceutical and cosmetic industries, nanotechnology and various other uses, such as barrier materials for storing nu­ clear waste or removing contaminants from wastewater, are gain­ ing momentum. Such diversity and abundance of possible appli­ cations calls for a thorough understanding of bentonite, including its different types and properties. Properties of different bentonites vary widely because of the variations in their mineralogical composition. The smectite spe­ cies is the most important factor determining bentonite proper­ ties, such as adsorption capacity or swelling behaviour. As the main constituent and the one giving bentonite its specific char­ acteristics, smectite is the obvious object of interest in most of the research done on bentonites. Different types of smectites were distinguished by GRIM & KULBICKI (1961), SCHULTZ (1969), BRIGATTI & POPPI (1981), and BRIGATTI (1983). They proposed trivial names, de­ rived from the provenance of the most typical samples, such as Wyoming, Tatatilla, Chambers, and Otay (Cheto), which differ in Comprehensive characterisation of bentonites from Croatia and neighbouring countries Zvonka Gverić1, Darko Hanžel2, Štefica Kampić1, Andrej Pleša3 and Darko Tibljaš*,1 1 University of Zagreb, Faculty of Science, Department of Geology, Division of Mineralogy and Petrology, Horvatovac 95, HR-10 000, Zagreb, Croatia; (*corresponding author: dtibljas@geol.pmf.hr) 2 Jožef Stefan Institute, Jamova cesta 39, SI-1000 Ljubljana, Slovenia 3 INA d.d., Exploration and production, Avenija Većeslava Holjevca 10, HR-10 000, Zagreb, Croatia doi: 10.4154/gc.2020.02 Abstract Bentonites are an important industrial resource and are also interesting from the mineralogical point of view. The main component of bentonites is a mineral from the smectite group in which chemical and structural variations, influencing bentonite properties, provide a great deal of in- teresting research topics. The aim of this study was to better describe bentonites from 11 known deposits from Croatia and neighbouring countries: Bednja, Bunarić, Draga, Divoselo, Sjeničak, Paripovac, Lončarski vis and Poljanska Luka (Croatia), Zaloška Gorica (Slovenia), Šipovo (Bos- nia and Herzegovina) and Vranjska Banja (Serbia). Thirteen samples were analysed using seve- ral available techniques in order to obtain the data necessary for currently accepted bentonite classification. The mineralogical composition was analysed using XRD and FTIR, and crystallo- chemical properties were investigated by thermal analysis, CEC determination with ammonium index cations, chemical analyses (ICP-AES and ICP-MS) and Mössbauer spectroscopy. The results showed that the main mineral constituent of most local bentonite deposits is a Fe-poor smectite, with a predominantly medium layer charge mostly as a result of octahedral substitu- tions, with calcium or sodium cations occupying the interlayer. Nevertheless, the variations be- tween samples are prominent enough to provide a good overview of the range of crystallo-chemi- cal properties which exist in different smectites resulting in varying bentonite properties. layer charge (particularly in its value and whether it originates from substitutions in the tetrahedral or octahedral sheets), high- temperature transformations, and Fe content. Unfortunately, these trivial names do not reflect some of the important chemical and structural parameters. Furthermore, several montmorillo­ nites could not be assigned to the proposed types according to this classification system. Therefore EMMERICH et al. (2009) proposed a new comprehensive classification, which allowed modifications and extension of the old one. It is based on the chemical composition (tetrahedral/octahedral charge distribu­ tion), the layer charge and exchangeable cations, the di- and tri- octahedral character, Fe content and the structure of the octa­ hedral sheet (cis­, trans-vacancy). The new classification (which is also used here) is based on standardised laboratory techniques used in characterisation of smectites. The most extensive research on bentonites in Croatia was done by BRAUN (1991) and it has served as a foundation for many later works. Some of the other research was undertaken for industrial purposes, which goes along with the fact that 9 samples (from 7 locations) used in this study were taken from active or once active bentonite mines where bentonite was used in drilling fluids and castings. The data gathered from this industrial-based research has been compiled by MARKOVIĆ (2002). Additiona- lly, bentonite deposits in the north­eastern part of Croatia have been studied by TIBLJAŠ (1996), and those in south Croatia by ŠEGVIĆ et al. (2006) and BIŠEVAC et al. (2007). Samples of bentonite contain different smectite, resulting in their varying properties. Considering the large array of possibili­ ties for bentonite use, it is important to be able to choose the Article history: Manuscript received September 25, 2019 Revised manuscript accepted January 17, 2020 Available online February 29, 2020 Keywords: bentonite deposits, smectite crystallo-chemical properties G eo lo gi a C ro at ic a Geologia Croatica 73/130 sample with desirable properties appropriate for the specific ap­ plication. Consequently, an accurate identification and classifica­ tion of the different smectites is important and the aim of this study was to reinvestigate Croatian bentonites by several availa- ble techniques and methods, and classify them in accordance with the criteria from contemporary literature. For comparison, a few samples of bentonites of similar ages from neighbouring countries were also studied. 2. GEOLOGICAL SETTING The investigated bentonites differ in age and depositional envi­ ronments (Fig. 1, Table 1) whereby the oldest, Bunarić (from the Maovice-Štikovo area), located in the External Dinarides, was deposited in Late Jurassic (Malmian-Kimmeridgian) Lemeš beds. The sedimentary succession, consisting of light­coloured platy limestone and/or dolomitic limestone and dolomite with chert intercalations, formed on the Adriatic Carbonate Platform in an intraplatform trough, connected to the open Tethyan realm (VLAHOVIĆ et al., 2005). Two bentonite layers exist within the succession, separated by 4 to 6 m of limestone and chert. The lower layer is approx. 3.5 m thick while the thickness of the up­ per layer varies between 1.5 and 5.5 m (BRAUN, 1991). Bentonites from Zaloška Gorica and Bednja were deposited during the Oligocene and the Early Miocene (Eggenburgian), re­ spectively, in the Slovenia-Zagorje Basin (SZB). The SZB is also known as the Trans-Tethyan Trench Corridor which acted as a marine connection between the Paratethys and the Mediterranean (RÖGL, 1998). This region formed prior to the Southern Panno­ nian Basin (SPB) installation at the south-eastern margin of the Paratethys Basin (MANDIC et al., 2012). The Zaloška Gorica deposit is located in the northern part of the Celje Basin within a bentonite belt extending for 15 km. The bentonite occurs in three exploitable layers in the upper part of an andesitic tuff horizon which is up to 100 m thick and is intercalated with the Oligocene clays overlying Triassic rocks. These clays are conformably overlain by the Badenian lithothamnium sandstones and conglomerates and quartz sands as the youngest Miocene sedi- ments (RIHTERŠIČ, 1958; DRŽAJ & LUKACS, 1968). In the Bednja area (known for its bentonite deposits), the Šaša clay pit is the largest. The irregular plate-shaped bentonite body which occurs there is 10 to 35 m thick and 165 m wide, tec­ tonically embedded within coarse-grained marine arenites. The body is composed of a random mixture of different varieties of altered pyroclastic rocks: volcanic agglomerates, lapilli tuff, banded and pelitic tuff (BRAUN, 1991). Based on the fossil con­ tent and superposition, the sediments were deposited during the Late Eggenburgian-Early Ottnangian and belong to the Vrbno Member of the Macelj Formation (AVANIĆ, 2012). Another group of deposits was formed in the SPB, prior to Paratethys flooding, in lakes closely related to the Dinaride Lake System (DLS) (MANDIC et al., 2012 and references therein). In the area around Gornja Jelenska, bentonite clays, formed by the alteration of andesitic vitroclastic tuff, were discovered in several deposits. A sample from the Draga deposit was used in this study. According to BRAUN (1991), three bentonite layers (0.15-0.9 m, 0.2-1.2 m and 0-2.0 m thick) occur in continental, most probably lacustrine Ottnangian, strongly tectonized sediments, repre­ sented by gravel, sand, silt, sandy and tuffitic marl, tuffite and sandy clay which alternate both vertically and laterally. All three bentonite layers have a sharp contact with the underlying layer, while they gradually transit to tuffite in the overlying layer. Based on palynological analyses, KRIZMANIĆ (1995) con­ cluded that the bentonite was deposited during the Late Karpatian to the Early Badenian in a swamp with gently inclined shores, but with deeper water areas. In the area around Šipovo (Bosnia and Herzegovina), several bentonite deposits (VUJNOVIĆ, 1981) formed in the intra-mon­ taine basin of the DLS (KRSTIĆ et al., 2001). In the Grabež and Sarići deposits, three bentonite layers are underlain by Lower Miocene sandy clay. The bentonite layers (approximately 2.5, 2 and 5.5 m thick) are separated by 1 - 5 m of clastic rocks. To the south, in the Babići and Greda deposits, two bentonite layers were de­ scribed by MILADINOVIĆ (1976) and VUJNOVIĆ (1981). There are some controversies concerning their age. According to litera­ ture, they were deposited in the Middle Miocene (MILADINOVIĆ, 1976), Middle to Late Miocene (VUJNOVIĆ, 1981) or even the Late Miocene (TRUBELJA & BARIĆ, 1979), but the newest pa- laeontological data (KRSTIĆ et al., 2001 and references therein) Figure 1. Sampling locations (abbreviations defined in Table 1), A = Austria; BIH = Bosnia and Herzegovina; H = Hungary; HR = Croatia, I = Italy; MNE = Monte- negro; RKS = Kosovo; SLO = Slovenia; SRB = Serbia. Table 1. List of samples. Sample Sampling location Age BU Bunarić, Central Dalmatia, S Croatia Late Jurassic (Kimmeridgian) (BRAUN, 1991) ZG* Zaloška Gorica, E Slovenia Oligocene (RIHTERŠIČ, 1958) BD Bednja, Hrvatsko zagorje, NW Croatia Early Miocene (Eggenburgian- Ottnangian) (BRAUN, 1991) DR Draga, Moslavina, Central Croatia Early Miocene (Ottnangian) (BRAUN, 1991) ŠI1 Sokolac, Šipovo, Central Bosnia and Herzegovina Early Miocene (KRSTIĆ et al., 2001) ŠI2 Greda, Šipovo, Central Bosnia and Herzegovina Early Miocene (KRSTIĆ et al., 2001) SJ Sjeničak, Central Croatia Middle Miocene (Badenian) (MANDIC et al., 2012) PR Paripovac, Central Croatia Middle Miocene (Badenian) (MANDIC et al., 2012) DI1 Divoselo, Lika, SW Croatia unknown DI2 Divoselo, Lika, SW Croatia unknown LV Lončarski vis, Slavonija, E Croatia Early Miocene (Karpatian) (MARKOVIĆ et al., 2018) VB* Vranjska Banja, SE Serbia Early to Middle Miocene (KRSTIĆ et al., 2001) PL Poljanska Luka, Hrvatsko zagorje, NW Croatia Middle Miocene (Badenian) (BRAUN, 1991) *samples from the collection of the Division of Mineralogy and Petrology, Faculty of Science. G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 31 show that the lacustrine sediments were deposited in the middle Early Miocene, i.e. between 19 and 17 Ma ago. Other lacustrine deposits are observable at Sjeničak and Pari- povac (30 km apart), in the Karlovac and Glina sub-depressions where a 0.5 m thick altered volcanic ash occurs in the upper part of the sediments deposited in the Early Badenian (≈16.0 Ma, based on 40Ar/39Ar dating). In Sjeničak, the bentonite is represented by a biotite­bearing montmorillonite clay layer, underlain by more than 6 m of interbedded sands, pelites and limestone, and is posi­ tioned 2 m below the transgression horizon, marked by a marine organogenic limestone of Badenian age. In Paripovac, the ash is grey, medium-size sand, with white calcite veins and biotite flakes. It is intercalated in a sedimentary sequence consisting of 2 m of clays and silts below the ash and about 1 m of clays and lacustrine limestone above the ash deposit (MANDIC et al., 2012). Bentonite from Divoselo most probably also belongs to this group of deposits, based on the similarity of the tuff mineral as­ sociations and mineral morphology with those from Neogene ba­ sins in Dalmatia, although its age and depositional environment is unknown. Clays were discovered under a thin Quaternary cover on the palaeorelief formed on Malmian carbonates. In the boreholes, usually one to three bentonite layers occur in the first 10 m (0.3-0.5, 1.6-2.0 and approx. 2 m thick). These are separated by tuff layers, though sometimes due to erosion the first one or two layers are missing. In one of the deeper boreholes, a 13.2 m thick clay layer was observed, and in another, bentonite clay was observed at a depth of 24.3 m (MARKOVIĆ, 2002). In the SPB, bentonites were also formed in marine environ­ ments. In Lončarski vis, (an area close to Gradac village), there are several quartz-trachyandesite quarries. Apart from the main volcanic body, smaller flows and tuff layers are interstratified within Karpatian marine sediments (breccia-conglomerates, gravels, sands and marls) (PAMIĆ, 1997 and references therein). Some tuff layers are altered to bentonite. Radiometric dating us­ ing 40Ar/39Ar dated tuffs at 16.96 ± 0.03 Ma, corresponding to Karpatian age (MARKOVIĆ et al., 2018). The Poljanska Luka bentonite deposit, also formed in the SPB after the Paratethys flooding, and contains four beds of ben­ tonite clay as a result of the alteration of vitroclastic tuff. Two lower layers, between 0.6 and 1 m thick, extend over 1800 m and were previously exploited. They are bound on the upper and lower surfaces by 0.5 m thick layers of altered tuffite. Bentonites are interstratified in Badenian marls, calcilutites and biomicrites, which were deposited in a lagoon environment (BRAUN, 1991; MARKOVIĆ, 2002). The bentonite sample from Vranjska Banja (Serbia) was taken from the collection of the Division of Mineralogy and Pe­ trology at the Faculty of Science, University of Zagreb, so its ex­ act geographic and stratigraphic positions are unknown. It origi­ nates from the area in which pyroclastic material is widespread and is better known for zeolite than bentonite deposits (SIMIĆ, 2001; SIMIĆ et al. 2014). According to KRSTIĆ et al. (2001) a large outcrop from Vranje to Vranjska Banja with prevailing tuff layers and subordinate tuffaceous sandstones and breccias, and an andesite lava flow, formed in a fresh water lacustrine environ­ ment (Serbian Lake) during the transition from the Lower into the Middle Miocene. According to the older sources (VUKANOVIĆ et al., 1977) this outcrop belongs to the Upper Eocene sedimentary succession within the Pčinja Palaeogene Ba­ sin comprising up to 750 m of pyroclastic rocks. The pyroclastics were deposited on land and in shallow fresh water (lacustrine, braided river) environments which prograded into a marine one. 3. MATERIALS AND METHODS The bentonite samples used in this study, named after the loca­ tions where they were collected from are: Bednja, Poljanska Luka, Draga, Lončarski vis, Paripovac, Sjeničak, Bunarić, two samples from Divoselo (Croatia); two samples from Šipovo (Bosnia and Herzegovina); Vranjska Banja (Serbia) and Zaloška Gorica (Slovenia) (Fig. 1, Table 1). Preparation of the samples included either grinding in an aga te mortar to prepare whole rock (WR) samples or separation of the clay-size (< 2 µm) fraction by centrifugation. Approxima- tely 50 g of sample was dispersed in approximately 600 ml of dis­ tilled water, then the resulting slurry was centrifuged using a Teht­ nica Centric 322A machine, configuration time calculated using Centrifuge software (KRUMM, 1994). The suspended clay-size particles were taken from the top of the suspension. WR samples were used to determine the mineral composition (XRD), cation exchange capacity (CEC) and chemical composition, while the clay-size fraction was used for clay mineral determination (XRD), FTIR analysis, chemical analysis, determination of the structural iron characteristics and thermal properties. 3.1. X-ray diffraction (XRD) XRD analyses were performed on both WR and clay-size frac­ tion samples using a Philips PW3040/60 X’Pert PRO diffractome- ter equipped with Cu tube (40 kV and 40 mA), graphite mono­ chromator and proportional counter. The following slits were used: mask 10 mm, 1/2° divergence slit, 1° and 2° antiscatter slits on tube and detector side respectively, and a 0.2 mm receiving slit. The samples were scanned in continuous mode at a speed of 0.02 °2θ/s. Diffraction patterns were processed using an X’Pert HighScore computer program (PANanalytical, 2004) and com­ pared with the Powder Diffraction File database and with the data from literature (BROWN, 1961; BRINDLEY & BROWN, 1980; MOORE & REYNOLDS, 1997). The clay-size fraction (< 2 μm) was analysed as a randomly oriented powder and was also used to make oriented mounts on glass slides (after MOORE & REYN­ OLDS, 1997). Oriented samples were analysed as air-dried, after leaving the glass slides overnight in ethylene glycol vapour, and after heating for 1 h at 400 °C and 550 °C. Clay minerals were identified by observing changes on diffractograms after different treatments using the flowchart from STARKEY et al. (1984). To differentiate between smectites and vermiculites, mont­ morillonite and beidellite, and to estimate the layer charge, clay­ size fractions were treated with magnesium, lithium and potas­ sium chloride solutions (concentrations: 3M for lithium chloride and 1M for magnesium and potassium chloride solutions). Ap­ proximately 200 mg of the sample (<2 µm fraction) was mixed with 30 ml of solution (done separately for each chloride solution) in cuvettes and left overnight on a shaker at 250 rpm. The suspen­ sion was then centrifuged at 3500 rpm for 15 minutes, the super­ natant then decanted and the sample washed three times with etha- nol solution (80%). Materials saturated with a cation of interest were dispersed in a few drops of distilled water and analysed as oriented glass slide mounts using XRD. Magnesium saturated samples were solvated with glycerol following the procedure pro­ posed by ŚRODOŃ (1980), i.e. left overnight covered with glyce- rol-soaked filter paper. Lithium saturated samples were heated to 200 °C and treated with glycerol (again covered overnight with glycerol-soaked filter paper) following GREENE-KELLY (1955). The approximate layer charge was inferred using part of the method described by CHRISTIDIS & EBERL (2003). The swell­ ing behaviour of potassium saturated samples, which were fully expandable in the original form was observed, adopting the prin­ G eo lo gi a C ro at ic a Geologia Croatica 73/132 ciple that low-charge smectites would swell to >16.6 Å after 12+ hours in a desiccator with ethylene-glycol vapours, while the high- charged ones would show a basal peak between 13.5 and 15.5 Å. A randomly oriented clay-size powder was additionally ana­ lysed in step-scan mode (step 0.02 °2Θ, time 5 s) in the range from 58.5 to 64.5 °2Θ to measure 060 peak positions and determine the b cell dimension (by multiplying d060 by 6). Quartz was used as an external standard. 3.2. Fourier Transform Infrared (FTIR) spectroscopy Approximately 2 mg of the clay size fraction was mixed with ap­ proximately 200 mg of KBr powder, homogenized and pressed into pellets with a diameter of 13 mm. The spectra were recorded immediately after pressing, after which the pellets were heated at 150 °C overnight and cooled down in a desiccator before repeated recording. The spectra were obtained using a Bruker TENSOR 27 FTIR spectrometer equipped with a KBr beam splitter. Thirty-two scans in the MIR (4000 – 400 cm-1) region were recorded for each sample with a resolution of 2 cm-1. The spectra were interpreted using data from the literature (MADEJOVÁ & KOMADEL, 2001; RUSSELL & FRASER, 1994). 3.3. CEC determination NH4 + ions were used as index cations to determine the cation ex­ change capacity following the procedure proposed by MINATO (1997) for zeolites. Homogenized portions of the WR sample (1 g) were dried and suspended in 30 ml 1M ammonium acetate so­ lution. After mixing, the suspension was centrifuged at 3500 rpm for 17 minutes and the precipitate was washed three times with ethanol. NH4 + cations were then exchanged with potassium cati- ons using 30 ml 1M potassium chloride solution, centrifuged at 3500 rpm for 8 minutes and the supernatant was preserved. This was repeated three times for each sample. The released ammo­ nium cations were measured using a Hach ammonium probe. 3.4. Thermal analyses Six samples (BD, BU, DI1, DR, PL and ŠI1) of about 20 mg were analysed using a Shimadzu TG/DTA instrument. The heating rate was 20 K/min in the 30 – 1000 °C range. The analysis was performed in an air current (flow rate: 100 ml/min), samples were placed in platinum crucibles and an empty crucible was used as a reference material. The instrument was calibrated using in­ dium, tin and aluminium for temperature and DTA correction. The remaining 7 samples were analysed using a Mettler To­ ledo TG/SDTA 851e. The heating rate was 10 K/min in the 30 – 1000 °C range. Samples (around 35 mg of each) were analysed in aluminium oxide crucibles in an oxygen atmosphere (flow rate: 150 ml/min) while nitrogen was used as a protective gas (50 ml/ min). An empty crucible was used as a reference material. Dehydration and dehydroxylation temperatures were read by observing the peak positions on DTA/DTG curves, no peak de­ composition was made. 3.5. Chemical analysis Major and trace elements both in the WR and clay-size fraction of the samples were analysed using ICP-AES (Spectro Ciros Vi­ sion instrument) and ICP-MS (Elan 9000 instrument) in the Bu­ reau Veritas Mineral Laboratories (Vancouver, Canada). The sample weight used was 0.2 g for each analysis. For major and some trace element analysis samples were fused with lithium metaborate and tetraborate and dissolved in diluted nitric acid. Other trace elements (Au, Ag, As, Bi, Cd, Cu, Hg, Mo, Ni, Pb, Sb, Se, Tl and Zn) were analysed using ICP-MS after the samples were dissolved in aqua regia. Carbon and sulfur were analysed on Leco Induction furnace CS230 and LOI was determined after burning the sample at 1000 °C. The chemical composition of the <2 µm fraction was used to calculate the chemical formulae of the samples. Considering the impurities found in this fraction (both by XRD and FTIR methods), corrections for the crystalline impurities were made. Their quantities were assumed based on XRD peak intensities and for some samples a trial­and­error type deduction, necessary to obtain reasonable formulae (e.g. Si<4, sum of octahedral cati- ons = 2, charge balance of the layer and interlayer cations) was undertaken. Mostly, the impurities were present only in relatively small amounts. The corrected chemical composition was then used to calcu­ late the structural formula based on half a unit cell with O10OH2 and 22 negative charges and calculated after STEVENS (1946). Cations were appointed to different positions as follows: – all Si atoms were assigned to the tetrahedral sheet – the rest of the spaces in the tetrahedral sheet were given to Al or Al and Fe3+ where Mössbauer spectroscopy showed tetrahedrally coordinated Fe3+ – the remaining Al and Fe3+ together with Fe2+ were assigned to the octahedral sheet Table 2. Mineral compositions of the samples (WR and <2 µm fractions) determined by XRD. Sample smectite kaolinite illite plagioclase quartz opal-CT calcite zeolite volcanic glass BU + X + + + X + ZG + X + X + + + X BD + X + X + + + + X + DR + X + X + + X ŠI1 + X + X + X + ŠI2 + X + X + X + SJ + X + X + PR + X + X + + DI1 + X + X + DI2 + X + X + + + X LV + X + VB + X + + + X + PL + X + + + X + present in the WR sample X present in the < 2 µm fraction G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 33 – the unoccupied spaces in the octahedral sheet were filled by Mg – Na, Ca, K and leftover Mg were assigned to the interlayer. 3.6. Mössbauer spectroscopy Fe distribution among the tetrahedral and octahedral sheets and its valence state was analysed using Mössbauer spectroscopy. Mössbauer spectra were recorded at room temperature using a constant acceleration Wissel spectrometer, in transmission mode with Co-57 source in Rh matrix. Velocity calibration was per­ formed using a thin α-Fe foil. The isomer shifts are expressed relative to α-Fe at room temperature. The spectra were fitted with the analysis code RECOIL (LAGAREC & RANCOURT, 1998) using quadrupole doublets of Lorentzian lineshape. 4. RESULTS 4.1. X-ray diffraction (XRD) Randomly oriented powder mounts of WR samples revealed that, apart from clay minerals, the samples contain variable amounts of other minerals: quartz, opal-CT, calcite, plagioclase feldspars Figure 2. XRD patterns of BU <2 µm fraction, oriented samples, (A) air dried, (B) glycolated, (C) heated to 400 °C, (D) heated to 550 °C, CuKa radiation. Figure 3. XRD patterns of DI1 sample (<2 µm) after Greene-Kelly test, lithium saturated sample heated at 200 °C (grey); after glycerol treatment (black), CuKa radiation. G eo lo gi a C ro at ic a Geologia Croatica 73/134 Ta bl e 3. d -v al ue s [ Å ] o f 0 0l a nd 0 60 p ea ks m ea su re d on o rie nt ed sa m pl e (e xc ep t f or 0 60 ) o f d iff er en tly tr ea te d <2 μ m fr ac tio n, a nd C V va lu es fo r E G tr ea te d K- sa tu ra te d sa m pl es . BU ZG BD A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG 00 1 14 .5 0 16 .7 8 9. 80 9. 70 18 .0 7 14 .1 2 14 .5 6 17 .3 8 9. 69 (b ) 9. 41 18 .0 6 16 .3 6 10 .6 6 17 .1 5 9. 85 9. 60 18 .0 9 15 .7 0 00 2 8. 54 6 4. 81 2 4. 75 6 9. 06 8. 28 6 8. 35 7 8. 86 (b ) 4. 86 8 8. 72 4. 82 2 4. 71 73 9. 03 00 3 4. 94 7 5. 52 8 3. 20 12 5. 91 9 4. 62 9 5. 67 3 5. 97 5 5. 56 (b ) 3. 19 57 5. 60 3. 20 (b , a ) 00 4 4. 48 4 3. 39 74 4. 43 4. 32 00 5 2. 99 31 3. 34 06 1. 92 93 3. 55 28 3. 42 82 3. 46 1 1. 93 92 3. 37 1. 93 31 3. 39 51 00 6 2. 78 00 2. 99 00 2. 84 28 2. 82 9 (b ) 2. 82 2. 82 57 CV 2. 72 2. 96 4. 40 06 0* 1. 49 83 1. 49 52 1. 50 23 D R ŠI 1 ŠI 2 A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG 00 1 12 .4 9 17 .2 0 9. 84 9. 71 18 .1 0 16 .4 0 14 .3 2 17 .3 2 10 .1 1 17 .6 0 18 .4 7 14 .2 3 15 .0 9 17 .1 5 9. 67 9. 65 18 .1 6 15 .7 7 00 2 4. 93 4 8. 58 6 4. 81 4 4. 74 5 9. 08 (b ) 8. 93 (b ) 8. 44 3 4. 88 9 9. 13 9. 14 8. 51 8 4. 74 4 4. 77 4 9. 02 3 00 3 5. 62 0 3. 15 04 5. 95 (b ) 5. 59 (b ) 4. 90 3 5. 61 1 5. 84 7 5. 88 8 4, 67 (b ) 4. 98 5. 62 3 5. 94 7 00 4 3. 11 50 4. 27 8 3. 56 5 4. 23 3 4. 55 1 4. 43 9 3. 42 6 00 5 3. 36 29 1. 93 02 3. 41 7 2. 99 3 3. 36 44 3. 50 88 3. 58 8 2. 77 9 (b ) 3. 00 3. 37 02 3. 41 80 00 6 2. 83 11 2. 79 05 3. 02 95 3. 00 2 2. 81 62 2. 82 4 (b ) CV 3. 62 1. 54 5. 69 06 0* 1. 49 91 1. 49 83 1. 49 68 SJ PR D I1 A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG 00 1 14 .8 8 17 .2 8 9. 59 18 .8 14 .0 7 (1 7. 56 ) 16 .7 4 14 .6 1 17 .1 5 9. 91 9. 68 17 .8 4 (1 4. 24 ) 17 .2 6 13 .6 1 17 .0 3 9. 70 17 .9 2 18 .2 5 16 .6 8 00 2 7. 19 5 8. 53 3 4. 74 9 9. 56 7. 13 (b ) 8. 68 6 8. 53 3 4. 81 9 4. 74 4 8. 92 1 8. 65 9 7. 41 0 8. 47 0 4. 78 9. 11 9 8. 89 3 00 3 5. 00 8 5. 65 7 4. 73 5 5. 71 2 4. 94 4 5. 66 1 5. 92 9 5. 73 8 5. 62 0 5. 93 8 5. 62 3 00 4 3. 59 03 4. 24 7 4. 70 2 4. 35 0 4. 25 75 4. 74 9 4. 32 35 3. 62 1 4. 50 02 00 5 3. 03 00 3. 39 43 3. 42 51 3. 02 74 3. 39 61 3. 43 87 3. 15 18 3. 36 93 3. 55 56 3. 42 48 00 6 2. 83 24 2. 85 79 2. 83 18 2. 86 82 2. 81 14 3. 01 64 2. 84 90 CV 1. 39 0. 29 2. 44 06 0* 1. 49 52 1. 49 75 1. 49 36 G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 35 D I2 LV VB A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG 00 1 15 .2 2 17 .4 6 10 .0 0 18 .0 7 18 .5 7 17 .2 7 15 .0 5 17 .3 3 10 .1 5 9. 70 18 .5 9 (1 5. 16 ) 16 .2 0 12 .9 (b , a ) 17 .1 3 9. 77 17 .8 5 18 .0 6 17 .2 6 00 2 8. 62 1 4. 84 9 9. 16 (b ) 9. 05 5 8. 64 8 8. 58 8 4. 82 7 4. 75 8 9. 16 2 6. 17 6 8. 52 1 4. 82 5 9. 35 7 8. 97 5 8. 68 3 00 3 5. 04 7 5. 65 9 6. 01 3 5. 65 8 4. 96 6 5. 65 1 6. 00 1 5. 66 7 5. 65 7 5. 95 5 5. 72 4 00 4 4. 28 9 4. 81 7 3. 12 33 4. 24 54 4. 60 6 4. 47 6 4. 31 47 00 5 3. 39 32 3. 42 99 2. 98 86 3. 38 63 3. 40 79 2. 48 81 3. 39 07 3. 44 31 00 6 2. 81 94 2. 86 36 2. 82 51 2. 83 94 2. 82 98 2. 86 27 CV 0. 74 2. 47 0. 42 06 0* 1. 49 75 1. 49 73 1. 49 67 PL A D EG 55 0 °C Li -2 00 -G ly M g- G ly K- EG 00 1 14 .9 5 17 .0 2 9. 84 9. 68 17 .8 2 17 .0 4 00 2 8. 49 6 4. 76 2 4. 70 5 8. 94 3 8. 71 6 00 3 4. 99 0 5. 63 6 3. 18 (b , a ) 5. 92 6 5. 68 2 00 4 4. 22 2 2. 50 00 5 3. 02 25 3. 37 89 1. 93 61 3. 55 91 3. 42 23 00 6 2. 81 83 2. 96 88 2. 86 27 CV 0. 94 06 0* 1. 49 75 *M ea su re d on ra nd om ly o rie nt ed sa m pl es (v al ue s i n pa re nt he se s i nd ic at e w ea ke r p ea k) b – br oa d a- a sy m m et ric G eo lo gi a C ro at ic a Geologia Croatica 73/136 and zeolite from the heulandite-clinoptilolite series (Table 2). Opal-CT and quartz were also detected in the <2 µm fraction of some samples. Both randomly oriented powder samples and ori­ ented clay-size mounts on glass slides showed that the main mine- ral in all of the samples is dioctahedral smectite (d060 ≈ 1.50 Å) with small amounts of kaolinite present in some samples. Smec­ tite was determined by a typical behaviour after EG treatment (d001 ≈ 17 Å) and after heating at 400 and 550 °C (d001 ≈ 10 Å) (Fig. 2, Table 3), while the first order kaolinite basal peak re­ mained unchanged around 7 Å after EG treatment and heating to 400 °C, collapsing after heating to 550 °C. A couple of samples showed the presence of an amorphous substance (presumed to be volcanic glass), visible as a “hump” between and 20 and 30 °2Θ, on both randomly oriented and oriented mounts. The Greene-Kelly test revealed the presence of beidellite in samples DI1 (Fig. 3), DI2 and ŠI1. After prolonged treatment with glycerol, an additional shift to 18 Å was observed for the VB and SJ samples. Mg-test confirmed the main mineral to be montmo­ rillonite in most of the samples (Fig. 4A). Some samples (LV, PR and SJ) showed two peaks after glycerol treatment, one between 17 and 18 Å and a smaller one (more prominent only in the case of the SJ sample) around 14 Å (Fig. 4B, Table 3). Samples saturated with K and treated with EG showed either full expansion to about 17 Å, negligible expansion with the 001 Figure 4. XRD patterns of <2 µm fraction of samples exchanged with Mg: (A) PL sample, before (grey) and after (black) glycerol solvation, showing behaviour typi- cal for smectite and (B) LV sample showing the splitting of the 001 peak after the same treatment, CuKa radiation. G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 37 peak remaining around 14 Å, or exhibited in-between values (Ta­ ble 3, Fig 5). Observed d060 values were in the range of values for mont­ morillonite: 1.492 – 1.504 Å (MOORE & REYNOLDS, 1997) (Table 3). 4.2. Fourier Transform Infrared (FTIR) spectroscopy Results of the FTIR analysis of the clay-size fraction (Table 4) are in accordance with the XRD results. A well-defined 3700 cm-1 band, as well as the 795 and 753 cm-1 doublet (Fig. 6) confirm the presence of kaolinite in some samples. The spectra also allowed Figure 5. Comparison of XRD patterns of (A) low- (PR), (B) medium- (LV), and (C) high-charged (ŠI1) smectites after K saturation and EG treatment of <2 µm fraction, CuKa radiation. Figure 6. FTIR spectrum of DI1 (<2 µm) sample with wave numbers of bands showing presence of kaolinite. G eo lo gi a C ro at ic a Geologia Croatica 73/138 a more detailed definition of some structural properties, e.g. cations in the octahedral coordination. The type of octahedrally coordinated cations pre­ sent in the smectite structure is visible from AlAlOH (915 cm-1), AlMgOH (840 cm-1) and AlFeOH (885 cm–1) deformation bands (MADEJOVÁ & KOMADEL, 2001) (Fig. 7). All samples showed the AlAlOH band, while the presence of the AlMgOH and AlFeOH bands varied (Table 4). A broad band around 3420 cm-1 representing the OH stretching of water is visible on all spectra. After heating the KBr pellets, its intensity diminished, how­ ever its complete disappearance was not achieved. 4.3. CEC determination CEC of the samples varied between 44 and 94 cmol(+)/ kg with one sample from Divoselo (DI2) showing an anomalously low value of 19 cmol(+)/kg (Table 5). 4.4. Thermal analyses All of the samples show a single or double endothermic peak assigned to the dehydration process between 100 and 200 °C, as well as one or more endothermic peaks at higher temperatures, between 450 and 700 °C as­ signed to dehydroxylation (Table 6). Some samples show exothermic peaks around 450 °C indicating organic mat­ ter oxidation and/or above 900 °C due to recrystalliza­ tion (Fig. 8). 4.5. Chemical analysis The chemical composition of WR and of the <2 µm frac­ tion of the samples is shown in Tables 7 and 8. Data for the <2 µm fraction were used to calculate the chemical formula of smectite (Table 9), while the WR chemistry, specifically the concentration of immobile elements, was used to deduce the composition of parent (volcanic) ma­ terial. 4.6. Mössbauer spectroscopy Mössbauer spectra recorded at room temperature differ significantly between samples (Fig. 9, Table 10). Some of them (BU and ŠI1) could be fitted with a single dou­ blet characteristic for Fe3+ in the octahedral coordina­ tion, while others had to be fitted with more doublets. Spectra of ŠI2 and DI1 were fitted with two doublets for Fe3+ in the octahedral coordination, while for the rest of the samples (except BD) one additional doublet attribu- ted to Fe2+ in octahedral coordination had to be used. For the BD sample, the fit was made with four doublets, three previously mentioned and an additional one for Fe3+ in tetrahedral coordination. Assignment to different Fe species was made in accordance with BARON et al. (2017), PELAYO et al. (2018) and references therein. There is no clear evidence of the presence of a magneti­ cally ordered phase, which indicates that most of the Fe present in the fine fraction of the investigated smectites is incorporated into the smectite structure. 5. DISCUSSION 5.1. Mineral and parent material composition The mineral composition of both the WR and clay-size fractions show the samples vary between almost monomineralic to polymineralic containing a number of Ta bl e 4. A ss ig nm en ts o f F TI R ba nd s o bs er ve d in th e <2 µ m fr ac tio ns o f b en to ni te s ( af te r M A D EJ O VÁ & K O M A D EL , 2 00 1) . M nt – m on tm or ill on ite , K ln – k ao lin ite . W av en um be rs [ cm -1 ] A ss ig ne m en t BU ZG BD D R ŠI 1 ŠI 2 SJ PR D I1 D I2 LV VB PL 37 04 36 95 O H st re tc hi ng o f i nn er -s ur fa ce h yd ro xy l g ro up s ( Kl n) 36 25 36 28 36 27 36 29 36 29 36 25 36 28 36 32 36 26 36 28 36 30 36 34 36 28 O H st re tc hi ng o f s tr uc tu ra l h yd ro xy l g ro up s ( M nt ) 16 30 16 30 16 31 16 31 16 32 16 27 16 22 16 29 16 26 16 27 16 31 16 22 16 27 O H d ef or m at io n of w at er 14 00 14 02 14 03 14 02 w ea k 14 01 14 02 N H 4 d ef or m at io n of N H 4B r 13 84 13 16 10 89 br oa d br oa d Si -O st re tc hi ng o f c ris to ba lit e 10 30 10 39 10 27 10 35 10 36 10 37 10 37 10 48 10 38 10 47 10 39 10 44 10 39 Si – O st re tc hi ng 91 3 91 3 in fle ct io n 91 5 91 2 91 5 91 8 92 0 91 2 91 8 91 6 91 6 91 6 A lA lO H d ef or m at io n (M nt ) in fle ct io n 88 0 in fle ct io n 88 9 88 4 in fle ct io n 88 4 in fle ct io n 88 5 A lF eO H d ef or m at io n (M nt ) 84 3 84 2 in fle ct io n 84 3 83 8 84 5 in fle ct io n 84 7 85 2 84 5 in fle ct io n 84 9 A lM gO H d ef or m at io n (M nt ) w ea k 80 0 w ea k 80 0 80 2 80 8 Si -O st re tc hi ng o f q ua rt z an d si lic a 79 5 79 5 79 5 79 8 79 7 79 5 Si -O st re tc hi ng o f c ris to ba lit e 79 8 79 3 Si – O (K ln ) 78 2 75 3 75 3 Si – O p er pe nd ic ul ar (K ln ) 69 7 69 2 Si -O , p er pe nd ic ul ar 62 5 62 7 62 6 62 4 62 6 62 4 62 6 Co up le d A l-O a nd S i-O , o ut -o f-p la ne ; S i-O o f c ris to ba lit e 60 7 52 0 52 9 52 3 52 0 52 7 52 2 52 5 52 5 53 3 52 5 52 2 52 5 52 1 A l – O – S i d ef or m at io n (M nt ) 46 6 47 2 47 0 47 0 46 8 46 7 47 0 46 8 46 9 46 7 46 7 46 9 46 7 Si – O – S i d ef or m at io n (M nt ) 43 0 42 4 Si – O d ef or m at io n G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 39 additional minerals (Table 2). Smectite is the dominant mineral in the <2 μm fraction. Kaolinite is present in 8 samples, and some silica minerals (quartz and/or opal-CT) remained in the <2 μm fraction after separation. The most problematic for further ana­ lyses was the presence of amorphous impurities (most likely vol­ canic glass) in samples ZG, DI1 and DI2. The Greene-Kelly test showed the majority of samples con­ tain montmorillonite and only some (DI1, DI2 and ŠI1) could un­ ambiguously be classified as beidellite following their expansion after lithium saturation and glycerol treatment. For the other sam­ ples (SJ, PR and VB) that showed swelling after the Greene-Kelly test, other analyses were not conclusive for the presence of beidel­ lite. The Mg-test was even less conclusive and showed some sam­ ples (SJ, PR and LV) retained (almost completely or partially) the 14 Å peak position after glycerol treatment, even when other analyses showed that their layer charge was not high enough for vermiculite. This could be due to incomplete glycerol saturation, or the presence of interstratification of differently charged layers which was confirmed by the irrational order of the 00l series. The determined mineral composition of the samples is in overall good correlation with the existing literature data Table 5. Measured CEC values (WR samples). Sample CEC [cmol(+)/kg] BU 60 ZG 60 BD 85 DR 94 ŠI1 82 ŠI2 91 SJ 78 PR 53 DI1 64 DI2 19 LV 85 VB 44 PL 72 Figure 7. FTIR spectrum of BD (<2 µm) sample with bands characteristic of the different octahedral cations indicated. Table 6. Observed temperatures of thermal reactions measured on <2 µm frac- tions of bentonites. Sample Dehydration temperature [°C] Dehydroxylation temperature [°C] BU 105 + shoulder at 165 660 ZG 125 + shoulder at 180 510 + weaker broad peak at 620 BD 110 broad peak below 600 + shoulder at 680 DR 105 peak at 635 + weaker broad peak at 480 ŠI1 110 + shoulder 180 485 ŠI2 135 + shoulder at 180 670 SJ 125 + shoulder at 175 broad 670 + weak broad at 520 PR 120 + shoulder at 160 680 DI1 100 + shoulder at 150 470 and a weak broad peak at 670 DI2 110 + shoulder at 165 650 LV 120 + shoulder at 170 670 VB 120 + shoulder at 160 asymmetric peak at 670 with a shoulder at 610 PL 100 + shoulder at 155 660 G eo lo gi a C ro at ic a Geologia Croatica 73/140 (BRAUN, 1991; MARKOVIĆ, 2002; MILADINOVIĆ, 1976; VUJNOVIĆ, 1981), with the exception of beidellite found in sam­ ples from Divoselo. FTIR spectra provided more information on dioctahedral smectite in the samples and confirmed the presence of some min­ eral impurities (Table 4). Kaolinite was confirmed in samples ZG and DI1; however, it was not confirmed in other samples where XRD indicated its presence. The 3700, 3620 cm-1 doublet is char­ acteristic of the kaolin group, as well as OH deformation bands at 938 and 916 cm-1 (RUSSELL & FRASER, 1994). In the inves­ tigated samples, the 3620 cm-1 band was overshadowed by a broad stretching band of the structural hydroxyl group in smectites and 938 and 916 cm-1 bands are probably also covered by smectite bands. The ZG and DI1 samples also showed two small bands around 795 and 753 cm-1 of roughly equal intensity, characteris­ tic for kaolinites (RUSSELL & FRASER, 1994). Silica impurities were also confirmed: a most notable band around 1090 cm-1 rep­ resenting Si-O stretching of cristobalite in samples BU, VB and PL and 795 cm-1 in samples BU, ŠI2, DI2, LV, VB and PL. Quartz or silica, determined by a Si-O stretching band at 800 cm-1 was found in samples BD, DR, PR and PL. Some discrepancy between the content of impurities shown between XRD and FTIR results could be the result of peak over­ lapping in recorded XRD patterns and FTIR spectra and/or the presence of impurities in concentrations below the detection limit. This demonstrates the benefits of using both methods to complement each other. Additionally, the thermal analysis showed an exothermal peak in the 450 °C region indicating pos­ sible presence of organic matter in some samples (BU, BD and PL) which have the highest TOT/C. The mineral composition of the samples is the result of the parent material chemical composition. With that in mind, immo­ bile element (Ti, Zr, Nb and Y) content was used to determine parent material composition from the discrimination diagram af­ ter WINCHESTER & FLOYD (1977) (Fig. 10). This diagram shows that most samples originate from acidic to neutral volcan­ oclastic material, with sample BD standing out as andesitic in composition. This can explain the highest Fe content in the BD Figure 8. DTA curves of three representative samples, ŠI1, DI1 and PL (<2 µm fractions) differing in dehydroxylation temperature range. In addition to endo- thermic peaks due to dehydration and dehyroxilation PL sample shows exo- thermal peak caused by organic matter oxidation while peaks due to recrystal- lization are visible on ŠI1 and DI1 DTA curves. Ta bl e 7. C he m ic al c om po si tio n of W R sa m pl es . A na ly te U ni t M D L BU W R ZG W R BD W R D R W R ŠI 1 W R ŠI 2 W R SJ W R PR W R D I1 W R D I2 W R LV W R VB W R PL W R Si O 2 % 0. 01 63 .6 6 49 .8 0 53 .7 5 53 .4 4 45 .5 5 51 .3 1 51 .4 2 52 .9 9 45 .1 2 64 .6 1 52 .5 9 65 .1 8 63 .1 9 A l 2O 3 % 0. 01 11 .0 1 21 .1 3 14 .0 3 15 .2 8 18 .2 8 17 .3 1 19 .8 5 18 .3 6 25 .7 8 14 .2 7 16 .7 7 13 .4 7 12 .4 6 Fe 2O 3 % 0. 04 0. 73 2. 54 5. 93 3. 15 4. 35 1. 97 2. 78 4. 08 4. 4 2. 41 2. 36 2. 24 1. 7 M gO % 0. 01 2. 78 3. 05 4. 15 4. 48 2. 25 3. 58 2. 43 2. 48 0. 77 0. 70 3. 62 1. 95 2. 27 Ca O % 0. 01 4. 58 0. 97 1. 4 1. 52 5. 62 2. 44 2. 29 2. 87 1. 63 1. 40 1. 93 1. 70 2. 36 N a 2 O % 0. 01 0. 24 0. 01 1. 75 1. 01 0. 02 0. 04 0. 05 0. 61 0. 05 1. 74 0. 20 0. 67 0. 31 K 2 O % 0. 01 1. 28 0. 11 0. 88 0. 41 0. 26 0. 20 0. 39 0. 20 0. 12 3. 48 0. 53 0. 78 0. 35 Ti O 2 % 0. 01 0. 11 0. 27 0. 63 0. 11 0. 53 0. 25 0. 29 0. 26 0. 22 0. 13 0. 20 0. 21 0. 09 P 2 O 5 % 0. 01 0. 02 <0 .0 1 0. 07 0. 02 0. 01 0. 05 0. 09 0. 11 0. 04 0. 01 0. 07 0. 05 0. 02 M nO % 0. 01 <0 .0 1 <0 .0 1 0. 02 0. 01 0. 03 <0 .0 1 0. 01 0. 01 0. 04 0. 05 <0 .0 1 0. 04 <0 ,0 1 Cr 2O 3 % 0. 00 2 0. 00 4 <0 .0 02 0. 00 4 0. 00 2 0. 00 8 <0 .0 02 <0 .0 02 <0 .0 02 0. 00 3 <0 .0 02 <0 .0 02 0. 00 3 0. 00 8 Ba pp m 5 17 3 46 15 3 40 2 80 7 29 8 13 8 14 1 11 58 61 7 26 6 16 9 22 1 N i pp m 20 30 <2 0 21 <2 0 28 <2 0 <2 0 <2 0 <2 0 <2 0 <2 0 <2 0 35 Sr pp m 2 56 60 35 0 24 0 16 6 30 40 12 6 21 44 11 8 31 2 15 8 Zr pp m 5 88 20 8 19 1 10 5 24 2 16 1 13 4 16 7 41 7 23 9 28 0 90 13 8 Y pp m 3 20 18 38 23 18 10 7 10 41 36 26 13 18 N b pp m 5 16 13 6 11 17 17 11 9 33 19 27 12 .3 12 Sc pp m 1 3 6 14 5 12 5 4 4 20 11 4 4 7 LO I % -5 .1 15 .5 22 .0 17 .2 20 .4 22 .9 22 .7 20 .3 17 .9 21 .6 11 .0 21 .5 13 .6 17 .1 Su m % 0. 01 99 .9 5 99 .9 0 99 .9 2 99 .9 1 99 .9 3 99 .8 6 99 .8 9 99 .8 7 99 .9 5 99 .8 6 99 .8 3 99 .8 8 99 .9 6 TO T/ C % 0. 02 0. 71 0. 04 0. 31 0. 19 0. 98 <0 .0 2 0. 05 0. 18 0. 09 0. 02 0. 02 0. 11 0. 24 TO T/ S % 0. 02 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 0. 02 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 M D L - m et ho d de te ct io n lim it G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 41 Ta bl e 8. C he m ic al c om po si tio n of < 2 µm fr ac tio n. A na ly te U ni t M D L BU < 2 µm ZG < 2 µm BD < 2 µm D R <2 µ m ŠI 1 <2 µ m ŠI 2 <2 µ m SJ < 2 µm PR < 2 µm D I1 < 2 µm D I2 < 2 µm LV < 2 µm VB < 2 µm PL < 2 µm Si O 2 % 0. 01 60 .5 5 50 .1 7 51 .1 3 53 .2 6 46 .7 7 50 .0 7 50 .2 2 52 .8 8 45 .2 52 .5 1 51 .5 3 62 .2 5 63 .2 6 A l 2O 3 % 0. 01 13 .2 8 22 .1 8 13 .8 5 15 .9 1 20 .4 7 17 .2 8 19 .7 8 17 .6 2 24 .5 6 17 .4 9 16 .8 4 14 .7 8 12 .3 3 Fe 2O 3* % 0. 04 1. 41 2. 47 5. 61 3. 58 4. 99 2. 28 2. 37 3. 55 4. 68 3. 86 2. 27 2. 17 1. 58 Fe O * % - 0. 00 0. 11 0. 91 0. 47 0. 00 0. 00 0. 14 0. 28 0. 00 0. 32 0. 32 0. 12 0. 07 M gO % 0. 01 3. 86 2. 98 4. 23 4. 03 2. 55 3. 59 2. 38 2. 57 0. 89 1. 94 3. 84 2. 19 2. 35 Ca O % 0. 01 2. 29 0. 97 1. 33 1. 36 1. 66 2. 56 2. 49 2. 26 1. 71 2. 53 1. 96 2. 00 2. 39 N a 2 O % 0. 01 0. 10 0. 07 2. 74 1. 25 0. 01 0. 06 0. 08 0. 09 0. 09 0. 27 0. 12 1. 09 0. 17 K 2 O % 0. 01 1. 09 0. 11 0. 50 0. 80 0. 24 0. 17 0. 16 0. 07 0. 15 0. 62 0. 31 0. 22 0. 20 Ti O 2 % 0. 01 0. 12 0. 25 0. 64 0. 20 0. 42 0. 25 0. 18 0. 14 0. 18 0. 16 0. 18 0. 08 0. 08 P 2 O 5 % 0. 01 <0 .0 1 <0 .0 1 0. 06 0. 07 0. 02 <0 .0 1 <0 .0 1 <0 .0 1 0. 02 <0 .0 1 0. 02 <0 .0 1 <0 ,0 1 M nO % 0. 01 <0 .0 1 <0 .0 1 0. 02 0. 02 0. 02 <0 .0 1 <0 .0 1 <0 .0 1 0. 03 0. 05 <0 .0 1 0. 02 <0 ,0 1 Cr 2O 3 % 0. 00 2 0. 00 6 <0 ,0 02 0. 00 4 0. 01 1 0. 00 9 <0 ,0 02 <0 ,0 02 <0 ,0 02 <0 ,0 02 <0 ,0 02 <0 ,0 02 <0 ,0 02 <0 ,0 02 Ba pp m 5 72 50 26 4 32 8 15 4 62 64 67 10 92 17 6 10 3 24 16 0 N i pp m 20 62 <2 0 <2 0 31 32 <2 0 <2 0 <2 0 21 <2 0 <2 0 <2 0 <2 0 Sr pp m 2 28 59 25 1 18 3 15 5 28 35 78 24 22 11 8 26 9 15 2 Zr pp m 5 81 23 2 20 2 97 16 5 99 57 84 27 0 23 4 20 8 68 10 3 Y pp m 3 19 20 34 21 15 9 5 8 32 27 23 7 13 N b pp m 5 21 13 6 12 17 16 7 7 14 25 29 9 10 LO I* * % -5 .1 17 .2 20 .6 18 .6 18 .8 22 .7 23 .6 22 .1 20 .4 22 .3 20 .1 22 .4 15 .0 17 .5 Su m % 0. 01 99 .9 2 99 .8 9 99 .9 1 99 .9 2 99 .9 3 99 .9 0 99 .9 3 99 .9 1 99 .9 3 99 .8 7 99 .8 5 99 .9 0 99 .9 6 TO T/ C % 0. 02 0. 25 0. 14 0. 44 0. 24 0. 15 0. 05 0. 13 0. 19 0. 14 0. 25 0. 08 0. 35 0. 40 TO T/ S % 0. 02 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 0. 02 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 <0 .0 2 0. 06 <0 .0 2 * va lu es c al cu la te d ac co rd in g to M ös sb au er sp ec tr a ** va lu e co rr ec te d fo r F eO Ta bl e 9. C al cu la te d ch em ic al fo rm ul ae o f i nv es tig at ed b en to ni te sa m pl es . Sa m pl es BU ZG BD D R ŠI 1 ŠI 2 SJ PR D I1 D I2 LV VB PL Co rr ec tio ns -2 5% o pa l-C T -1 0% k ao lin ite -1 % c al ci te -1 % k ao lin ite -2 % q ua rt z -1 % q ua rt z -1 % k ao lin ite -1 0% k ao lin ite -2 0% o pa l-C T -2 5% o pa l-C T Te tr ah ed ra l Ca tio ns Si 3. 72 A l 0. 28 Si 3. 80 A l 0. 20 Si 3. 91 Fe 3+ 0. 03 A l 0. 06 Si 3. 91 A l 0. 09 Si 3. 62 A l 0. 38 Si 3. 89 A l 0. 11 Si 3. 82 A l 0. 18 Si 3. 94 A l 0. 06 Si 3. 54 A l 0. 46 Si 3. 93 A l 0. 07 Si 3. 94 A l 0. 06 Si 3. 88 A l 0. 12 Si 3. 95 A l 0. 05 Ch ar ge (% o f t ot al la ye r c ha rg e) -0 .2 8 (3 5% ) -0 .2 0 (4 3% ) -0 .0 9 (1 5% ) -0 .0 9 (1 7% ) -0 .3 8 (6 7% ) -0 .1 1 (2 2% ) -0 .1 8 (4 0% ) -0 .0 6 (1 6% ) -0 .4 6 (9 4% ) -0 .0 7 (2 3% ) -0 .0 6 (1 3% ) -0 .1 2 (2 8% ) -0 .0 5 (1 2% ) O ct ah ed ra l Ca tio ns A l 1. 36 F e3+ 0. 11 M g 0 .5 3 A l 1. 59 F e3+ 0. 15 Fe 2+ 0. 01 M g 0 .2 5 A l 1. 17 Fe 3+ 0. 30 Fe 2+ 0. 06 M g 0 .4 7 A l 1. 34 F e3+ 0. 21 Fe 2+ 0. 03 M g 0 .4 2 A l 1. 51 F e3+ 0. 30 M g 0 .1 9 A l 1. 47 Fe 3+ 0. 13 M g 0 .4 0 A l 1. 59 Fe 3+ 0. 14 Fe 2+ 0. 01 M g 0 .2 6 A l 1. 49 F e3+ 0. 20 Fe 2+ 0. 02 M g 0 .2 9 A l 1. 66 F e3+ 0. 31 M g 0 .0 3 A l 1. 47 Fe 3+ 0. 22 Fe 2+ 0. 02 M g 0 .2 2 A l 1. 46 Fe 3+ 0. 13 Fe 2+ 0. 02 M g 0 .3 9 A l 1. 48 Fe 3+ 0. 15 F e2+ 0. 01 M g 0 .3 0 A l 1. 45 F e3+ 0. 12 Fe 2+ 0. 01 M g 0 .3 6 Su m o f o ct ah ed ra l c at io ns 2. 00 2. 00 2. 00 2. 00 2. 00 2. 00 2. 00 2. 00 2. 00 1. 93 2. 00 1. 94 1. 94 Ch ar ge (% o f t ot al la ye r c ha rg e) -0 .5 3 (6 5% ) -0 .2 6 (5 7% ) -0 .5 3 (8 5% ) -0 .4 5 (8 3% ) -0 .1 9 (3 3% ) -0 .4 0 (7 8% ) -0 .2 7 (6 0% ) -0 .3 1 (8 4% ) 0. 03 (6 % ) -0 .2 4 (7 7% ) -0 .4 1 (8 7% ) -0 .3 1 (7 2% ) -0 .3 7 (8 8% ) To ta l l ay er c ha rg e (T +O ) -0 .8 1 -0 .4 6 -0 .6 2 -0 .5 4 -0 .5 7 -0 .5 1 -0 .4 5 -0 .3 7 -0 .4 9 -0 .3 1 -0 .4 7 -0 .4 3 -0 .4 2 In te rla ye r Ca tio ns Ca 0. 26 M g 0 .0 7 N a 0 .0 2 K 0. 15 Ca 0. 09 M g 0 .1 2 N a 0 .0 1 K 0 .0 1 Ca 0. 06 M g 0 .0 2 N a 0 .4 1 K 0. 05 Ca 0. 11 M g 0 .0 4 N a 0 .1 8 K 0. 08 Ca 0. 14 M g 0 .1 1 K 0 .0 2 Ca 0. 21 M g 0 .0 2 N a 0 .0 1 K 0. 02 Ca 0. 20 M g 0 .0 1 N a 0 .0 1 K 0 .0 2 Ca 0. 18 N a 0 .0 1 Ca 0. 16 M g 0 .0 9 N a 0 .0 2 K 0. 02 Ca 0. 20 N a 0 .0 4 K 0 .0 6 Ca 0. 16 M g 0 .0 5 N a 0 .0 2 K 0. 03 Ca 0. 20 N a 0 .1 9 K 0 .0 3 Ca 0. 26 N a 0 .0 3 K 0. 03 Ch ar ge +0 .8 3 +0 .4 4 +0 .6 2 +0 .5 6 +0 .5 2 +0 .4 9 +0 .4 5 +0 .3 7 +0 .5 4 +0 .5 0 +0 .4 7 +0 .6 2 +0 .3 2 G eo lo gi a C ro at ic a Geologia Croatica 73/142 sample. The initial hypothesis that the samples containing beid­ ellite, and thus less SiO2, originate from more basic material was not confirmed. Still, in the case of altered pyroclastic rocks, ele- mental analyses should be taken cautiously as different materials could have been deposited in the same basin. Moreover, the volcanic material could have been separated during transport, especially in the case of distant depositional basins, which is mostly the case for the studied samples. Figure 9. Mössbauer spectra of (a) BU and (b) BD samples (<2 µm) recorded at room temperature. a b G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 43 5.2. Crystallo-chemical properties of the investigated smectites and their classification Some trends could be observed regarding the chemical composi­ tion of the investigated smectites (Tables 7 and 8): the main ele­ ments did not show large variation between WR and the <2 µm fraction with the exception of a much higher calcium concentra­ tion in WR of BU and ŠI1 samples, due to calcite presence. Among microelements, zirconium, usually present in zircon which is expected in the silt fraction, is more abundant in the WR, as well as strontium which can replace calcium in calcite and pla­ gioclase. In contrast, the Ni and Fe concentrations showed nega­ tive correlation with particle size. Structural chemical formulae calculated from the chemical composition of samples’ <2 μm fractions could be compromised with crystalline and especially amorphous impurities present (BRIGATTI, 1983). Therefore, the resulting chemical formulae (Table 9) represent an approximate distribution of cations in the smectite structure. They can show some chemical and structural features of the smectite, such as layer charge, its distribution and the chemical composition of the octahedral sheet. Nevertheless, additional analyses were used to confirm (or disprove) these. Some calculated formulae show inconsistencies, e.g. charge im­ balance or divergence from the results of other analyses. For ex­ ample, the calculated layer charge for BU and, to lesser extent, for BD samples is too high for smectites, yet the Mg­test did not show the presence of vermiculite in those samples. Samples DI2 and VB have a significant discrepancy between their calculated layer charge and the interlayer cations compensating for it, which could be explained by the presence of amorphous impurities in the DI2 sample, but remains speculative in the case of the VB sample. According to the chemical composition, the dominant cation in the interlayer is Ca for most samples, except BD and DR where Na is more abundant and VB in which both cations are equally represented. In the ZG sample, the calculated structural formula would suggest Mg is the dominant interlayer cation; however we cannot be sure of the accuracy of the formula due to amorphous impurities detected in this sample, it is probably more likely Ca as well (DRŽAJ & LUKACS, 1968). There are a number of smectite classification schemes in the existing literature (GRIM & KULBICKI, 1961; SCHULTZ, 1969; BRIGATTI & POPPI, 1981) taking into account different criteria, such as layer charge, chemical composition and structure. EMMERICH et al. (2009) proposed a new classification scheme, in order to account for many discrepancies between known samples of bentonites not fitting into historical classifica­ tion criteria, based on the following groups of analytical data: – layer charge, – structure of the octahedral sheet, – iron content, and – charge location. These four criteria were used to further describe the smec­ tite in the investigated bentonites. Figure 10. Content of immobile elements in analyzed bentonites plotted on the diagram for determination of original composition of altered rocks (WIN- CHESTER & FLOYD, 1977). Table 10. Mössbauer parameters obtained from the fit of the Mössbauer spectra recorded at room temperature on the studied smectites [mm/s], assignment to dif- ferent Fe-species and their relative concentrations. Width is expressed as half width at half maximum (HWHM), isomer shift is specified relative to α-Fe. Fe3+ (VI) (A) Fe3+ (VI) (B) Fe2+ (VI) Fe3+ (IV) δ Δ width area δ Δ width area δ Δ width area δ Δ width area Sample [mm/s] [mm/s] [mm/s] [%] [mm/s] [mm/s] [mm/s] [%] [mm/s] [mm/s] [mm/s] [%] [mm/s] [mm/s] [mm/s] [%] BU 0.35 0.48 0.23 100.0 - - - - - - - - - - - - ZG 0.36 0.53 0.25 61.0 0.36 1.00 0.48 35.0 1.06 2.69 0.42 4.8 - - - - BD 0.37 0.45 0.20 33.3 0.40 1.10 0.37 44.1 1.10 2.73 0.20 15.3 0.19 0.57 0.18 7.4 DR 0.35 0.43 0.26 63.4 0.39 1.24 0.30 23.9 1.13 2.66 0.22 12.7 - - - - ŠI1 0.36 0.50 0.27 100.0 - - - - - - - - - - - - ŠI2 0.36 0.52 0.17 52.0 0.36 0.87 0.38 48.0 - - - - - - - - SJ 0.36 0.52 0.20 58.0 0.38 1.08 0.36 36.0 1.11 2.57 0.30 6.0 - - - - PR 0.34 0.68 0.33 88.2 0.67 1.11 0.13 3.8 0.97 2.81 0.37 8.0 - - - - DI1 0.35 0.51 0.22 72.0 0.39 0.96 0.39 28.0 - - - - - - - - DI2 0.35 0.54 0.22 68.3 0.45 1.19 0.33 23.2 0.99 2.68 0.37 8.5 - - - - LV 0.35 0.43 0.23 50.0 0.38 1.11 0.39 36.0 1.16 2.82 0.22 13.5 - - - - VB 0.33 0.46 0.28 49.0 0.41 0.94 0.51 45.0 1.12 2.82 0.18 5.9 - - - - PL 0.35 0.58 0.26 48.0 0.36 1.09 0.44 47.0 1.30 2.40 0.14 4.9 - - - - S – broad peak with shoulder G eo lo gi a C ro at ic a Geologia Croatica 73/144 According to EMMERICH et al. (2009) the interlayer cations are not considered for classification, nevertheless the descriptive name can be extended with respect to the original or exchanged interlayer cations, which, on one hand indicate conditions during smectite formation, and on the other, are responsible for its prop­ erties. 5.2.1. Layer charge The layer charge was determined using both calculated and ob­ served data, bearing in mind that chemical formulae calculations are affected by the presence of impurities in the samples. The layer charge from the calculated chemical formulae showed most samples are in the medium charge range, with values 0.31-0.54 per formula unit (FU) or half unit cell, with a few exceptions: ŠI1 (0.57), BD (0.62), and especially BU (0.81). Smectites are defined as 2:1 phyllosilicates with expandable layers and a layer charge between approximately 0.2 and 0.6/FU (GUGGENHEIM et al., 2006); therefore, BU would not satisfy that criteria. However, due to the already mentioned difficulties in obtaining an accurate chemical formula, and thus a calculated layer charge, the empha­ sis is put on the swelling behaviour of Mg-saturated samples treated with glycerol when it comes to differentiation between smectites and vermiculites. All of the investigated samples (in Mg-form), except SJ and partly PR and LV swelled after glycerol solvation, even if for some of them (BD and BU were not among them) a longer time was necessary for full swelling. The SJ sam­ ple, which is medium charged according to both its chemical for­ mula and swelling behaviour of K-saturated form, showed split­ ting of the 001 peak: only a small portion of the peak moved to 17.6 Å while the majority remained at 14.0 Å. Samples saturated with K and treated with EG showed either full expansion to about 17 Å, negligible expansion (001 peak re­ maining around 14 Å) or showed in-between values. Accordi- ngly, based on the layer charge inferred from such swelling be­ haviours samples can be divided into three groups: those with low layer charge expanding fully, those with high layer charge (non-expanding) and the samples showing in-between values and an irrational series of 00l peaks. Coefficient of variation (CV) of d00l values for K-saturated samples analysed after EG treatment were used to determine the stacking order of the clay mineral layers. Fully expandable samples, i.e. those with low-charge lay­ ers (PR, VB, DI2 and PL) exhibit a rational order of basal peaks and have a CV below or slightly above 0.75% (value defined by BAILEY, 1982) indicating periodic stacking of layers. Samples with expansion behaviour that suggests their layers are medium- to high-charged show an irrational order of basal peaks and higher CV values (Table 3). ČÍČEL & MACHAJDÍK (1981) in­ vestigated the three types of layers (expandable, partly expanda- ble and non-expandable) in smectites saturated with monovalent ions with low solvation energy and concluded they are most com­ monly randomly distributed in the minerals. Likewise, the over­ all conclusion drawn about the layer charge approximations (also applicable to the results of Greene-Kelly and Mg-test) is that the studied samples may not be homogeneous, i.e. they could consist of different layers and thus show intermediate properties. This would explain the inconclusive behaviour in different analyses and can be backed up by evidence of CV values which suggest an irrational order sequence in 9 out of 13 samples. It should be noted however that CV values should be calculated on at least 10 basal peaks (BAILEY, 1982) which were not all present in this case. Furthermore, CV calculations are strongly influenced by inaccurate reading (done by software profile fitting), of the 001 peak position, which is especially true for wider peaks in the case of inhomogeneous samples with a small number of smectite lay­ ers in stacks. There are a few samples (most notably ŠI1) where there is a discrepancy between the calculated and observed layer charge (Fig. 11), so for the purposes of classification the observed (in­ ferred) layer charges were used, on account of the fact that the chemical data could be compromised with mineral and amor­ phous impurities. When comparing the calculated tetrahedral and octahedral charge with the observed swelling behaviour, the samples largely follow the trend postulated by SATO et al. (1992) (Fig. 12). On their diagram, sample ŠI1, with a higher charge in the tetrahedral sheet, which notably deviated from the linear correlation of cal­ culated and inferred layer charge (Fig. 11) shows behaviour in accordance with the theory that swelling depends not only on the total layer charge but also on its location. CEC was measured on WR rather than clay-size fractions since all the investigated bentonites are pure enough to not have Figure 11. The relationship between calculated and inferred layer charge (after EG-swelling behaviour of K-exchanged samples i.e. observed d001) showing relatively good correlation, with some outliers (notably ŠI1). Figure 12. Octahedral and tetrahedral sheet charge and observed swelling be- haviour of K-saturated <2 μm sample fractions after EG solvation plotted on the diagram by SATO et al. (1992). White circles – samples fully expanding to about 17 Å; black dots – negligibly expanding samples with 001 peak remaining around 14 Å; black squares –samples showing in-between values. G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 45 to be treated before prospective applications. Several samples show CEC values (Table 5) lower than those expected for smec­ tites. MEIER & NÜESCH (1999) denoted the lower CEC limit to be 65 ± 2 cmol(+)/kg; however some samples are even lower than this limit. In this case the issue is most likely due to the fact that WR samples contain variable amounts of mineral and amor­ phous phases which do not contribute to the CEC. 5.2.2. Structure of the octahedral sheet When it comes to the octahedral sheet structure, chemical data indicate aluminium abundance. XRD analysis, done on randomly oriented clay-size fraction in the 59 – 65 °2Θ region, showed all the investigated clay minerals to be dioctahedral (Table 3). Additional cations occupying the octahedral sheet are ferric, ferrous and magnesium cations; the last two contributing to the octahedral charge. The OH vibrations visible in the FTIR spectra are affected by the octahedral cations to which the OH group is coordinated (PETIT, 2006). The octahedral cation occupancy can be deduced from characteristic bands: around 920 cm-1 for AlAlOH, 885 cm-1 for AlFeOH and 845 cm-1 for AlMgOH. There is a correlation between the octahedral charge and the prominence of the diffe- rent bands. Beidellites (DI1, DI2 and ŠI1) have a more prominent AlAlOH band and other less prominent bands, while samples with a higher number of octahedral substitutions show other more prominent bands. Samples with a higher amount of octahedral Fe (e.g. BD) can also be distinguished from the FTIR spectra. Thermal analyses performed on clay-size samples provided additional information about the structure of the octahedral sheet. The nature of octahedral cations and their bonding strength affect the endothermic peak position. The dehydroxylation temperature increases as Fe–OH < Al–OH < Mg–OH (KÖSTER, 1993). DRITS et al. (1995) showed that the temperature depends not only on the mineral species, but also on the position of vacant octahe­ dra in dioctahedral clay minerals. They set the limit of the dehy­ droxylation peak temperature discriminating between a trans­va­ cant and cis-vacant octahedral sheet at 600 °C. This means that ideal trans-vacant (tv) montmorillonites should lose their OH- groups below 600 and ideal cis-vacant (cv) montmorillonites above 600 °C. WOLTERS & EMMERICH (2007) expanded on this and postulated a model for estimating the tv and cv sheets in a given sample. KOMADEL & MADEJOVÁ (2010) stated that beidellitic samples would show a lower dehydroxylation tempera­ ture, i.e. more trans­vacant sites. Dehydroxylation temperatures varied significantly between the analysed samples with some samples showing multiple dehy­ droxylation peaks (Table 6). Generally speaking “ideal” montmo­ rillonites lose their structural water in the 700 °C region, while “ideal” beidellites have their main dehydroxylation peak below 550 °C (GREENE-KELLY, 1953). The fact that some samples can­ not be attributed to either group shows that some of the samples are “non-ideal”, i.e. structurally inhomogeneous. A small number of samples and their low Fe content did not allow determination of the influence of Fe on the dehydroxylation temperature. In the present study, the peak modelling of the mass-spec­ trometer curves of evolved water proposed by WOLTERS & EM­ MERICH (2007) was not used; however, by observing the peak dehydroxylation temperature it was possible to deduce the pre­ dominant nature of the octahedral sheets (Table 6). The samples in which the dehydroxylation peak was below 550 °C show an exothermic peak around 900 °C attributed to the recrystallization process. 5.2.3. Iron content BD is the only sample in which Fe made up a considerable pro­ portion of the octahedral cations (almost 20%) and is considered ferrian. Samples coming close to the limit of 15% percent of the octahedral positions occupied by Fe ions are DI1 and ŠI1; how­ ever, those would not be considered ferrian, noting as well that other samples taken from the same deposits (DI2 and ŠI2) do not follow the same trend. Information about Fe speciation, its oxidation state and co­ ordination, is necessary for the correct structural formula calcu­ lations. In this study, Mössbauer spectroscopy was used to pro­ vide the required data (Table 10). It showed that Fe2+ is present in the majority of samples (9 out of 13). Its content is small but in some samples not negligible (> 10 % of total Fe). There is no ev­ ident relationship between the geological environment (lacustrine or marine) and the total iron or Fe2+ content. Traditionally, in for­ mula calculations, only Al for Si substitutions have been ac­ counted for, mostly due to the smaller ion size of Al in compari­ son with Fe3+, but also due to the difficulties with IVFe3+ measurement, which persist even in Mössbauer spectroscopy analyses. GATES et al. (2002) concluded that tetrahedral Fe3+ is unlikely to be present in smectites with < 34 wt% Fe2O3 but KAUFHOLD et al. (2017) found tetrahedral Fe3+ even in sam­ ples containing 2.1 wt% Fe2O3. In sample BD with 6.6 wt% (the highest Fe content among analysed samples) fitting of Mössbauer spectra revealed that 7.4 % of the total iron is present in the tet­ rahedral coordination. Nevertheless, it must be said that for this sample a reasonable fit could also be obtained without the doublet corresponding to IVFe3+. Some of the spectra showed only one or one dominant dou­ blet (with a quadrupole splitting value in the range 0.43-0.68 mm/s) characteristic for Fe3+ in the octahedral coordination. For others, an additional doublet attributed to Fe3+ in the octahedral coordination (with a quadrupole splitting value ranging between 0.87 and 1.24 mm/s) had to be used. Some authors attribute (e.g. COEY, 1980) these two doublets to Fe3+ in octahedral cis­ and trans- positions, respectively. Several authors (BARON et al., 2017; PELAYO et al., 2018 and references therein) doubt this in­ terpretation. The studied samples speak in favour of the latter: both BU and ŠI1 samples have only the first doublet and accord­ ing to the thermal properties the first is cis-vacant while the sec­ ond is trans-vacant. Other authors (PELAYO et al., 2018 and ref­ erences therein) relate these doublets to less and more distorted octahedral environments. PELAYO et al. (2018) attribute distor­ tion to the substitution of Al with Fe but for the studied samples no correlation of total iron content and % area of the second dou­ blet was observed. Sometimes, a reliable structural formula calculation, includ­ ing proper iron content evaluation, is problematic due to impuri­ ties, especially amorphous ones. In such cases, relationships be­ tween chemical and physical properties are used to evaluate the content of an element of interest. An example of such physical properties, which can be easily measured, are unit-cell dimen­ sions. It is known that the length of the b­unit cell edge is corre­ lated with structural Fe content (RADOSLOVICH, 1962; BRIG­ ATTI, 1983; KÖSTER et al., 1999; HEUSER et al., 2013). For the analysed set of samples, most of which have a low iron content, no correlation of Fe content and b­unit cell dimen­ sion was observed (Fig. 13). There are several possible explana­ tions for this. Such a lack of correlation can be caused by inac­ curate structural formulae resulting from impurities present in the sample, e.g. the very low unit cell dimension for given Fe G eo lo gi a C ro at ic a Geologia Croatica 73/146 content for DI2 sample could be because of the presence of amor­ phous material. Another source of error could be the calculation of the b­unit cell dimension by multiplying d060, since the ob­ served reflection is not a single one (DESPRAIRIES, 1983). The lack of correlation could also be explained by the fact that the b-dimension does not depend only on iron content (RA­ DOSLOVICH, 1962), and that for low iron content other factors (octahedral Mg and tetrahedral Al content) can be more impor­ tant. This is in agreement with the results by BRIGATTI (1983) who stated that correlation of the b-unit cell edge with the struc­ tural Fe content is valid only for samples with an iron content > 0.50 atoms per half-cell and none of the analysed samples ful­ fil that requirement. HEUSER et al. (2013) also observed poorer correlation for samples with low Fe contents. In contrast, KÖSTER et al. (1999) do not mention the lack of correlation; however, their regression was determined on only five samples of nontronites and Fe­rich smectites. Observed b-unit cell dimensions were compared with those calculated using different regressions (Table 11, Fig. 14). There is a poor correlation between the observed and calculated values for the regression of KÖSTER et al. (1999), which takes only the in­ fluence of Fe into account. Better correlation was observed for values calculated using the regression of RADOSLOVICH (1962), which takes into account other substitutions; however, all calculated values are greater than those observed. 5.2.4. Charge location Overall (with the exceptions of the SJ, VB and DI2 samples), the Greene-Kelly test results are in accordance with the distribution of layer charge which can be observed from the calculated chemi- cal formulae. They are also predominantly in accordance with the observed dehydroxylation temperatures which are, as ex­ pected, lower for beidellitic samples (KOMADEL & MADE­ JOVÁ, 2010). Based on the distribution of the layer charge between octa­ hedral and tetrahedral sheets, smectites can be divided into mont­ morillonites (90–100% of charge originating in octahedral sheet – O), beidellitic montmorillonites (50–89% O), montmorillonitic beidellites (10–49% O) and beidellites (0–9% O) (EMMERICH et al., 2009). Most of the studied samples fall into the “beidellitic mont­ morillonite” category, even though their octahedral charge ranges from 57, which is close to “montmorillonitic beidellite” to 88% O which is very close to “montmorillonite”. Only DI1 and ŠI1 can be classified as beidellite and montmorillonitic beidellite, re­ spectively. Those two samples are also the only ones consistently showing the higher amount of tetrahedral charge, i.e. other ana- lyses (chemistry, DTA) are in accordance with their beidellitic nature. 5.2.5. Interlayer cations Chemical analysis and subsequent formula calculation (Table 9) show that Ca is the dominant interlayer cation in the majority of the investigated bentonites. Na is dominant only in two samples (BD and DR), while one sample (VB) shows an approximately equal ratio of Ca and Na in the interlayer. The sample ZG shows quite large quantities of Mg in the interlayer, though this could be a miscalculation due to amorphous impurities (volcanic glass) present in the sample. DTA results largely confirm the chemical data when it comes to the interlayer cations. The majority of samples (except BD and DR) show a two-step dehydration process. This is in accordance with the fact that in both the BD and DR samples Na is the domi- nant interlayer cation. According to GREENE-KELLY (1953) and SCHULTZ (1969), Na-smectites show a single endotherm peak around 100-150 °C, while Ca-smectites have two distinct endotherm peaks. According to GREENE-KELLY (1953) this is the result of the premature loss of water during dehydration caused by limited stability of the monolayer Na-smectite com­ plex. In the VB sample in which chemical analysis determined Na and Ca in the interlayer, two distinct peaks are visible, though the higher temperature one is very small. Samples with Na cations in the interlayer showed an addi­ tional endothermic peak at around 850 °C. The presence of an NH4Br deformation band in IR spectra of some samples indicates NH4 + ions in their interlayer (MADE­ JOVÁ & KOMADEL, 2001). 6. CONCLUSIONS Thirteen bentonite samples from eleven known deposits in Cro­ atia, Bosnia and Herzegovina, Serbia and Slovenia contain dio- ctahedral smectite as the main component. Some of them are al­ most monomineralic while others contain a number of additional phases (kaolinite, illite, quartz, opal-CT, plagioclase feldspars, calcite, zeolite from the heulandite-clinoptilolite series, amor­ phous impurities, and organic matter), some of those also being present in the <2 μm fraction. Figure 13. Correlation of b-unit cell dimension and number of Fe ions obtained by structural formula calculation. Figure 14. Correlation of observed and calculated b-unit cell dimension (black dots – regression of RADOSLOVICH, 1962; gray dots – regression of KÖSTER et al., 1999). G eologia C roatica Gverić et al.: Comprehensive characterisation of bentonites from Croatia and neighbouring countries 47 Smectites show prominent variations in their crystallo- chemical properties resulting in different bentonite properties. They were classified (Table 12) according to the newest classifi­ cation scheme proposed by EMMERICH et al. (2009) based on: – layer charge – both swelling properties of K-saturated sam­ ples and calculated structural formulae showed (in spite of some discrepancies) that samples ranged from those with low layer charge to those with high-layer charge but most are medium to low-charged. Results showed that samples are most probably not homogeneous, i.e. they consist of layers of different charges. – octahedral sheet structure – all investigated smectites are dioctahedral. IR spectra and chemical analyses showed variations in the Al, Fe and Mg content of the octahedral sheet. Dehydroxy- lation temperature, as a proxy for the location of the vacancy in the octahedral sheet (tv vs. cv) varied significantly between the samples with some samples showing multiple dehydroxylation peaks indicating that some of the samples are structurally inho­ mogeneous. – iron content – most of the samples are non-ferrian. The ex­ ceptions are BD in which Fe accounts for almost 20% of the oc­ tahedral cations, and DI1 and ŠI1 samples which are close to the limit defined as 15% percent of the octahedral positions occupied by Fe ions. In the Mössbauer spectrum of the BD sample there is an indication of Fe3+ in tetrahedral coordination. Mössbauer spec­ troscopy showed that Fe2+ is present in the majority of samples (9 out of 13). Its content is small but in some samples not negli­ gible (> 10 % of total Fe). Overall, Mössbauer spectra differ sig­ nificantly between samples. – charge location – calculated structural formulae, which are in accordance with the results of Greene-Kelly tests and FTIR spectroscopy, showed that most of the samples fall into the “bei­ dellitic montmorillonite” category, even though their octahedral charge ranges from 57%, which is close to “montmorillonitic bei­ dellite” to 88% O (PL sample) which is very close to “montmo­ rillonite”. Samples DI1 and ŠI1 contain “beidellite” and “mont­ morillonitic beidellite, respectively. – type of interlayer cation - the dominant cation in the inter­ layer is Ca for most samples, except BD and DR where Na is more abundant and VB in which both cations are present in approxi­ mately equal amounts. ACKNOWLEDGMENT The research has been financially supported by the 119-1191155- 1156 project of the Croatian Ministry of Science and Education, research program P1–0112 by Slovenian Research Agency, and annual support over several years from the University of Zagreb. The authors are grateful to Dobroslav SUPAN†, Stanislav MEĐIMOREC, Franjo PERCELA, Frane MARKOVIĆ and Vanja BIŠEVAC for their help in providing samples and to Dražan JOZIĆ (University of Split, Faculty of Chemistry and Techno- logy) and Dominik CINČIĆ (University of Zagreb, Faculty of Science, Department of Chemistry) for recording thermal curves. The authors also express their gratitude to Anita GRIZELJ and an anonymous reviewer for their valuable constructive comments and suggestions. REFERENCES AVANIĆ, R. (2012): Litostratigrafske jedinice donjeg miocena sjeverozapadne Hrvatske [Lower Miocene lithostratigraphic units from north-western Croatia – in Croatian with English Summary].– Unpl. PhD Thesis, Faculty of Science, University of Zagreb, 162 p. BAILEY, S.W. (1982): Nomenclature for regular interstratifications.– Am. Mineral., 67, 394–398. DOI: 10.1180/CLAYMIN.1982.017.2.09 BARON, F., PETIT, S., PENTRÁK, M., DECARREAU, A. & STUCKI, J.W. (2017): Revisiting the nontronite Mössbauer spectra.– Am. Mineral., 102, 1501–1515. BIŠEVAC, V., ŠEGVIĆ, B., TIBLJAŠ, D. & LUGOVIĆ, B. (2007): Mineralogy of al­ tered tephra layers in the Upper Jurassic Lemeš deposits near Maovice (Dalmatia, Croatia).– In: ROCHA, F., TERROSO, D. & QUINTELA, A. (eds.): Euroclay 2007, Abstract Book. University of Aveiro, Aveiro, 75–76. BRAUN, K. (1991): Mineraloško-petrografske karakteristike i geneza ležišta bentonit­ nih glina Maovica, Gornje Jelenske, Bednje i Poljanske Luke [Mineralogical-pet- rographical characteristics and genesis of the bentonite deposits of Maovice, Gornja Jelenska, Bednja and Poljanska Luka – in Croatian, with an English Ab­ stract].– Prirodoslovna istraživanja, 63, Acta Geol., 21/1, 1–34. BRIGATTI, M.F. (1983): Relationship between composition and structure in Fe-rich smectites.– Clay Miner., 18, 177–186. doi: 10.1180/claymin.1983.018.2.06 BRIGATTI, M.F. & POPPI, L. (1981): A mathematical model to distinguish the mem­ bers of the dioctahedral smectite series.– Clay Miner., 16, 81–89. doi: 10.1180/ claymin.1981.016.1.06 BRINDLEY, G.W. & BROWN, G. (1980): The X-ray identification and crystal structures of clay minerals.– Mineralogical Society, London, 495 p. doi: 10.1180/mono-5 BROWN, G. (1961): The X-ray identification and crystal structures of clay minerals.– Mineralogical Society, Clay Minerals Group, London, 544 p. CHRISTIDIS, G.E. & EBERL, D.D. (2003): Determination of layer charge characteristics of smectites.– Clay. Clay Miner., 51, 644–655. doi: 10.1346/CCMN.2003.0510607 ČÍČEL, B. & MACHAJDÍK, D. (1981): Potassium- and ammonium-treated montmoril­ lonites. I. Interstratified structures with ethylene glycol and water.– Clay. Clay Miner., 29, 40–46. doi: 10.1346/CCMN.1981.0290106 COEY, J.M.D. (1980): Clay minerals and their transformations studied with nuclear tech­ niques: The contribution of Mössbauer spectroscopy.– Atom. Energy Rev., 18, 73–124. DESPRAIRIES, A. (1983): Relation entre le parametre b des smectites et leur contenu en fer et magnesium. Application a l’etude des sediments.– Clay Miner., 18, 165–175. doi: 10.1180/claymin.1983.018.2.05 Table 12. Proposed descriptive names of the smectite mineral dominant in the investigated samples, following classification criteria proposed by EMMERICH et al. (2009). Sample Proposed descriptive name BU Ca-saturated high-charge cv beidellitic montmorillonite ZG Mg,Ca-saturated medium-charge tv/cv beidellitic montmorillonite BD Na-saturated medium-charge ferrian tv/cv beidellitic montmorillonite DR Na- saturated medium-charge cv/tv beidellitic montmorillonite ŠI1 Ca,Mg- saturated high-charge tv montmorillonitic beidellite ŠI2 Ca- saturated medium-charge cv beidellitic montmorillonite SJ Ca- saturated medium-charge cv/tv beidellitic montmorillonite PR Ca- saturated low-charge cv beidellitic montmorillonite DI1 Ca- saturated medium-charge tv/cv beidellite DI2 Ca- saturated low-charge cv beidellitic montmorillonite LV Ca- saturated medium-charge cv beidellitic montmorillonite VB Ca,Na- saturated low-charge cv beidellitic montmorillonite PL Ca- saturated low-charge cv beidellitic montmorillonite Table 11. Observed (6 x d060) and calculated (after relationships of RADOSLOV- ICH, 1962 and KÖSTER et al., 1999) b - unit cell dimensions. Sample b - observed b – calculated after RADOSLOVICH (1962) b – calculated after KÖSTER et al. (1999) BU 8.9898 9.0158 8.9377 ZG (S) 8.9712 8.9908 8.9433 BD 9.0138 9.0299 8.9693 DR 8.9946 9.0107 8.9523 ŠI1 8.9898 9.0051 8.9591 ŠI2 8.9808 8.9990 8.9400 SJ 8.9712 8.9900 8.9422 PR 8.9850 8.9952 8.9501 DI1 8.9616 8.9937 8.9602 DI2 8.9850 8.9908 8.9523 LV 8.9838 8.9981 8.9422 VB 8.9802 8.9926 8.9433 PL 8.9850 8.9929 8.9400 S – broad peak with shoulder G eo lo gi a C ro at ic a Geologia Croatica 73/148 DRITS, V.A., BESSON, G. & MULLER, F. (1995): An improved model for structural transformations of heat-treated aluminous dioctahedral 2:1 layer silicates.– Clay. Clay Miner., 43, 718–731. doi: 10.1346/CCMN.1995.0430608 DRŽAJ, B. & LUKACS, E. (1968): Nekatere geološke in tehnološke značilnosti ben­ tonitov iz nahajališč v okolici Celja [Some geological and technological charac- teristics of bentonites in deposits from Celje surroundings – in Slovenian].– Ge­ ologija, 11, 129–136. EMMERICH, K., WOLTERS, F., KAHR, G. & LAGALY, G. (2009): Clay profiling: The classification of montmorillonites.– Clay. Clay Miner., 57, 1, 104–114. doi: 10.1346/CCMN.2009.0570110 GATES, W.P., SLADE, P.G., MANCEAU, A. & LANSON, B. (2002): Site occupan­ cies by iron in nontronites.– Clay. Clay Miner., 50, 223–239. doi: 10.1346/ 000986002760832829 GREENE-KELLY, R. (1953): Irreversible dehydration in montmorillonite. Part II.– Clay Mineral Bull, 2, 52–56. GREENE-KELLY, R. (1955): Dehydration of the montmorillonite minerals.– Mineral. Mag., 30, 604–615. doi: 10.1180/minmag.1955.030.228.06 GRIM, R.E. & KULBICKI, G. (1961): Montmorillonite: High temperature reactions and classification.– Am. Mineral., 46, 1329–1369. GUGGENHEIM, S., ADAMS, J.M., BAIN, D.C., BERGAYA, F., BRIGATTI, M.F., DRITS., V.A., FORMOSO, M.L.L., GALÁN, E., KOGURE, T. & STANJEK, H. (2006): Summary of recommendations of nomenclature committees relevant to clay mineralogy: report of the Association Internationale pour l’Etude des Argiles (AIPEA) Nomenclature Committee for 2006.– Clay Miner., 41, 863–877. doi: 10.1180/0009855064140225 HEUSER, M., ANDRIEUX, P., PETIT, S. & STANJEK, H. (2013): Iron-bearing smec­ tites: a revised relationship between structural Fe, b cell edge lengths and refrac­ tive indices.– Clay Miner., 48, 97–103. doi: 10.1180/claymin.2013.048.4.06 KAUFHOLD, S., STUCKI, J.W., FINCK, N., STEININGER, R., ZIMINA, A., DOHR­ MANN, R., UFER, K., PENTRÁK, M. & PENTRÁKOVÁ, L. (2017): Tetrahedral charge and Fe content in dioctahedral smectites.– Clay Miner., 52, 51–65. doi: 10.1180/claymin.2017.052.1.03 KOMADEL, P. & MADEJOVÁ, J. (2010): Identification and characterization of inter­ stratified clay minerals with spectroscopic and classical methods.– In: FIORE, S, CUADROS, J. & HUERTAS, F.J. (eds.): Interstratified Clay Minerals: Origin, Characterization and Geochemical Significance. AIPEA Educational Series, Pub. No. 1., 89–113. KÖSTER, H.M. (1993): Beschreibung Einzelner Tonminerale.– In: JASMUND, K. & LAGALY, G. (eds.): Tonminerale und Tone: Struktur, Eigenschaften, Anwendun­ gen und Einsatz in Industrie und Umwelt. Steinkopff Verlag, Darmstadt, 33–58. doi: 10.1007/978-3-642-72488-6_2 doi: 10.1180/claymin.1999.034.4.06 KÖSTER, H.M., EHRLICHER, U., GILG, H.A., JORDAN, R., MURAD, E. & ON­ NICH, K. (1999): Mineralogical and chemical characteristics of five nontronites and Fe-rich smectites.– Clay Miner., 34, 579–599. KRIZMANIĆ, K. (1995): Palynology of the Miocene Bentonite from Gornja Jelenska (Mt. Moslavačka Gora, Croatia).– Geol. Croat., 48, 147–154. KRSTIĆ, N., DUMURDŽANOV, N., OLUJIĆ, J., VUJNOVIĆ, L. & JANKOVIĆ- GOLUBOVIĆ, J. (2001): Interbedded tuff and bentonite in the Neogene lacustrine sediments of the central part of the Balkan Peninsula. A review.– Acta Vulcano­ logica, 13, 91–99. KRUMM, S. (1994): Centrifuge, Erlangen.– http://www.ccp14.ac.uk/ccp/web-mirrors/ krumm/html/software/winsoft.html LAGAREC, K. & RANCOURT, D.G. (1998): RECOIL Mössbauer Spectral Analysis Software for Windows.– Department of Physics, University of Ottawa. MADEJOVÁ, J. & KOMADEL, P. (2001): Baseline studies of the clay minerals society source clays: Infrared methods.– Clay. Clay Miner., 49, 5, 410–432. doi: 10.1346/ CCMN.2001.0490508 MANDIC, O., de LEEUW, A., BULIC, J., KUIPER, K.F., KRIJGSMAN, W. & JURIŠIĆ-POLŠAK, Z. (2012): Paleogeographic evolution of the Southern Pan­ nonian Basin: 40Ar/39Ar age constraints on the Miocene continental series of North­ ern Croatia.– Int. J. Earth Sci. (Geol Rundsch), 101, 1033– 1046. doi: 10.1007/ s00531-011-0695-6 MARKOVIĆ, S. (2002): Hrvatske mineralne sirovine [Croatian raw materials – in Cro­ atian].– Institut za geološka istraživanja, Zagreb, 544 p. MARKOVIĆ, F., KOVAČIĆ, M., ĆORIĆ, S., TIBLJAŠ, D., PEZELJ, Đ., HAJEK-TA­ DESSE, V., HERNITZ-KUČENJAK, M. & BAKRAČ, K. (2018): Miocene tuffs from the North Croatian Basin.– In: GANIĆ, M., CVETKO, V., VULIĆ, P., ĐURIĆ, D. & ĐURIĆ, U. (eds.): Book of Abstracts 17th Serbian Geological Con­ gress, Vrnjačka Banja, 166–168. MEIER, L.P., & NÜESCH, R. (1999): The Lower Cation Exchange Capacity Limit of Montmorillonite.– J. Coll. Interf. Sci., 217, 77–85. doi: 10.1006/jcis.1999.6254 MILADINOVIĆ, D. (1976): Bentoniti [Bentonites – in Bosnian].– In: VARIĆAK, D. (ed.): Mineralne sirovine Bosne i Hercegovine, Knjiga II, Ležišta nemetala [Mine- ral raw materials of Bosnia and Hercegovina, Book II, Deposits of non-metallic raw materials]. Geoinžinjering Sarajevo, 402–406. MINATO, H. (1997): Standardization of methods for zeolite speciality determination and techniques for zeolite resources utilization.– In: KIROV, G., FILIZOVA, L. & PETROV, O. (eds.): Natural Zeolites - Sofia ‘95, PENSOFT, Sofia, 282–292. MOORE, D.M. & REYNOLDS Jr. R.C. (1997): X-Ray Diffraction and the Identifica­ tion and Analysis of Clay Minerals.– Oxford University Press, Oxford, 378 p. PAMIĆ, J. (1997): Vulkanske stijene Savsko-Dravskog međuriječja i Baranje (Hrvatska) [Volcanic rocks from the Sava-Drava interfluve and Baranja in the South Pannonian Basin (Croatia) – in Croatian with an English Summary].– Nafta, Zagreb, 192 p. PANanalytical (2004): X’Pert HighScore Plus, Version 2.1. – Almelo. PELAYO, M., MARCO, J.F., FERNÁNDEZ, A.M., VERGARA, L., MELÓN, A.M. & PÉREZ DEL VILLAR, L. (2018): Infrared and Mössbauer spectroscopy of Fe-rich smectites from Morrón de Mateo bentonite deposit (Spain).– Clay Miner., 53, 17–28. doi: 10.1180/clm.2018.1 PETIT, S. (2006): Fourier Transform Infrared Spectroscopy.– In: BERGAYA, F., THENG, B.K.G. & LAGALY, G. (eds.): Dev. Clay Sci., Vol. 1. Handbook of Clay Science, Elsevier, Amsterdam, 909–918. doi: 10.1016/B978-0-08-098259-5.00009-3 RADOSLOVICH, E. (1962): The cell dimensions and symmetry of layer-lattice sili­ cates. II. Regression relations.– Am. Mineral., 47, 617–636. RIHTERŠIČ, J. (1958): Bentoniti v celjski kotlini [Bentonites in Celje Basin – in Slove­ nian].– Geologija, 4–13, 193–196. RÖGL, F. (1998): Paleogeographic considerations for Mediterranean and Paratethys sea­ ways (Oligocene to Miocene).– Ann. Naturhist. Mus. Wien, 99A, 279–310. RUSSELL, J.D. & FRASER, A.R. (1994): Infrared methods.– In: WILSON, M.J. (ed.): Clay Mineralogy: Spectroscopic and Chemical Determinative Methods, Chapman & Hall, London, 11–67. doi: 10.1007/978-94-011-0727-3_2 SATO, T., WATANABE, T. & OTSUKA, R. (1992): Effects of layer charge, charge lo­ cation, and energy change on expansion properties of dioctahedral smectites.– Clay. Clay Miner., 40, 103–113. doi: 10.1346/CCMN.1992.0400111 SCHULTZ, L.G. (1969): Lithium and potassium absorption, dehydroxylation tempera­ ture and structural water content of aluminous smectites.– Clay. Clay Miner., 17, 115–149. doi: 10.1346/CCMN.1969.0170302 ŠEGVIĆ, B., TOŠEVSKI, A, ŠTEVANIĆ, D., BIŠEVAC, V. & LUGOVIĆ, B. (2006): Altered tephra layers in the Upper Jurassic Lemeš deposits near Maovice (Dalmatia, Croatia): Clay mineralogy and basic soil mechanic properties.– In: VLAHOVIĆ, I., TIBLJAŠ, D., DURN, G. & BIŠEVAC, V. (eds.): 3rd Mid-Euro­ pean Clay Conference (MECC 06), Abstracts Book. Faculty of Science & Faculty of Mining, Geology and Petroleum Engineering, Zagreb, 104–104. SIMIĆ V. (2001): Bentoniti i bentonitske gline Srbije – stanje i perspektiva [Bentonites and bentonitic clays of Serbia – Actual state and prospects – in Serbian with an English Abstract].– Zbornik radova III međunarodnog savetovanja o površinskoj ekspolataciji glina, Ruma, 186–191. SIMIĆ, V., ŽIVOTIĆ, D., ANDRIĆ, N., RADOSAVLJEVIĆ-MIHAJLOVIĆ, A. & KAŠIĆ, V. (2014): Zeolite deposits and occurrences in Serbia – an overview.– In: ZEOLITE 2014, 9th International Conference on the Occurrence, Properties and Utilization of Natural Zeolites, Belgrade, Serbia, 217–218. ŚRODOŃ, J. (1980): Precise identification of illite/smectite interstratifications by X-ray powder diffraction.– Clay. Clay Miner., 28, 401–411. STARKEY, H.C., BLACKMON, P.D. & HAUFF, P.L. (1984): The Routine Mineralogi- cal Analysis of Clay-Bearing Samples.– U.S. Geol. Surv. Bull., 1563, Washington, 32 p. STEVENS, R.E. (1946) A system for calculating analyses of micas and related minerals to end members.– Contributions to Geochemistry, 1942-45. Geol. Surv. Bull., 950, 101–119. TIBLJAŠ, D. (1996): Zeoliti i drugi produkti alteracijskih procesa u oligocenskim i do- njomiocenskim piroklastitima na širem području Macelja [Zeolites and the other products of the alteration processes in Oligocene and Early Miocene pyroclastic rocks from the Macelj area – in Croatian, with an English Summary].– Unpl. PhD Thesis, Faculty of Science, University of Zagreb, 167 p. TRUBELJA, F. & BARIĆ, LJ. (1979): Minerali Bosne i Hercegovine [Minerals of Bos- nia and Herzegovina – in Croatian, with an English Summary].– Zemaljski muzej BiH, Sarajevo, 236–245. VLAHOVIĆ, I., TIŠLJAR, J., VELIĆ, I. & MATIČEC, D. (2005): Evolution of the Adriatic Carbonate Platform: Palaeogeography,main events and depositional dy­ namics.– Palaeogeogr. Palaeocl., 220, 333–360. doi: 10.1016/j.palaeo.2005.01.011 VUJNOVIĆ, L. (1981): Osnovna geološka karta SFRJ 1:100000. Tumač za list Bugojno L33−143 [Basic Geological Map of SFRY 1:100000, Geology of the Bugojno sheet – in Bosnian, with an English Summary].– Savezni geološki zavod, Beograd, 58 p. VUKANOVIĆ, M., DIMITRIJEVIĆ, N., DIMITRIJEVIĆ, M.N., KARAJIČIĆ, LJ. & RAKIĆ, M.O. (1977): Osnovna geološka karta SFRJ 1:100000. Tumač za list Vranje L34−54 [Basic Geological Map of SFRY 1:100000, Geology of the Vranje sheet – in Serbian, with an English Summary].– Savezni geološki zavod, Beograd, 53 p. WINCHESTER, J.A. & FLOYD, P.A. (1977): Geochemical discrimination of different magma series and their differentiation products using immobile elements.– Chem. Geol., 20, 325–343. doi: 10.1016/0009-2541(77)90057-2 WOLTERS, F. & EMMERICH, K. (2007): Thermal reactions of smectites – Relation of dehydroxylation temperature to octahedral structure.– Thermochim. Acta, 462, 80–88. doi: 10.1016/j.tca.2007.06.002