Crystallite size distribution of clay minerals from selected Serbian clay deposits

The BWA (Bertaut-Warren-Averbach) technique for the measurement of the mean crystallite thickness and thickness distributions of phyllosilicates was applied to a set of kaolin and bentonite minerals. Six samples of kaolinitic clays, one sample of halloysite, and five bentonite samples from selected Serbian deposits were analyzed. These clays are of sedimentary, volcano-sedimentary (diagenetic), and hydrothermal origin. Two different types of shape of thickness distribution were found – lognormal, typical for bentonite and halloysite, and polymodal, typical for kaolinite. The mean crystallite thickness (TBWA) seams to be influenced by the genetic type of the clay sample.


Introduction
The size distributions of crystallites can be measured by powder X-ray diffraction (XRD) because the widths of the XRD peaks broaden as the crystallite size decreases, if the influence of associated components on the degree of disorder of clay minerals (as presented for kaolinite by GALAN et al., 1994) is eliminated by adequate sample preparation.The interpretation of distribution and the shapes of crystallite thicknesses, measured by the Bertaut-Warren-Averbach (BWA) method, can then be related to crystal-growth mechanisms according to the theoretical approach of EBERL et al. (1998a).
The BWA technique has been applied to the measurement of illite particle thickness (EBERL et al., 1998b), to measure the crystallite size distribution of kaolin minerals ([UCHA et al, 1999), to explore crystal growth mechanisms for illite and smectite (SRODON et al., 2000;MYSTKOWSKI & SRODON, 2000), to study the diagenetic evolution of the crystallite thickness distribution of illitic material (KOTARBA & SRODON, 2000), weathering processes which affected smectite and illite/smectite ([UCHA et al., 2001), and crystallite-size changes of pyrophyllite during grinding (UHLIK et al., 2000).EBERL et al. (1998a) studied the growth mechanism of minerals based on the shapes of the crystal size distribution.
The main goal of this study was to measure the thickness and thickness distribution of kaolinite and smectite crystallites by the BWA technique and to compare the results with those obtained for similar clays from Slovakia and some other world deposits, and to check if the mean crystallite size depends on the origin of the clay.
This study is a part of the Project "Genesis of Natural Microporous Mineral Resources and their Application in Industry and Environmental Protection" which is performed by the Department of Geology of Mineral Deposits, Faculty of Natural Sciences, Comenius University, Bratislava, Slovakia, and Department for Exploration of Mineral Deposits, Faculty of Mining and Geology, University of Belgrade, Serbia.
Prior to analyses, < 2 mm fractions were separated from the bulk samples by sedimentation.Separated fine fractions were used for X-ray diffraction (XRD) analysis of oriented specimens.The oriented specimens were prepared by sedimentation of the clay suspension (10 mg/cm 2 ) onto glass slides.All specimens were analysed by XRD ussing a Philips PW 1710 diffractometer equipped with Cu radiation with a graphite monochromator.The step size was 0.02° 2 with a counting time of 5s for the oriented specimens.
The resulting basal reflections of the clay minerals were used for the determination of the mean crystallite thickness (crystallite = X-ray scattering domain) and thickness distribution by means of the BWA techniques (DRITS et al., 1998) using the MudMaster program (EBERL et al., 1996).The XRD method of crystallite size determination is based on the observation that XRD peaks broaden regularly as a function of decreasing crystallite size.The first basal reflection of all samples was subjected to BWA analysis in the recommended two theta intervals between 6 and 13° for the kaolinite (Fig. 1) and 2.5 to 7.5, for the smectites.All kaolinite samples, except the halloysite sample NB, contain illite in the clay fraction.Therefore, the illite peaks were chopped by the program PkChopr (Fig. 2).A longer XRD exposition time (5 s) was used to obtasin smooth XRD patterns for the analysis.
Scanning electron images were taken from fresh rock chips coated with gold using a Jeol JXA 840 scanning electron microscope (SEM).

Geological features of the studied deposits
Samples of 6 kaolinites, 1 halloysite and 5 smectites from deposits in three different geological environments and origin were used for this study.Kaolinites and smectites from sedimentary (originating in a weathering crust and transported into sedimentary basins), volcanosedimentary (formed by diagenesis of volcanic ash in a subaqual and subarial environment) and of hydrothermal origin were studied.The geological setting of each sample is indicated in Table 1.The analysed samples represent a selected collection of kaolinitic clays and bentonites from the economically most important deposits in Serbia, one sample of halloysite and one sample of Miocene tonstein, recently discovered in eastern Serbia.

Results and discussion
Typical XRD patterns of each genetic type of clay are shown in Fig. 3, and the results of the BWA measurements of the kaolinite, halloysite and smectite samples in Table 2.
The T BWA value of the sedimentary kaolinites studied varies between 5.55 and 7.91 nm, with an average value of 6.48 nm.The curves of all five samples are polymodal (Fig. 4), indicating that the samples consist of two or more generations of crystals with different thickness.The average T BWA of the Serbian sedimentary kaolinites is slightly higher than the T BWA of Slovakian sedimentary kaolinites, but, at the same time, significantly smaller than the T BWA of selected world kaolinites (Table 3).
The BWA measurements confirmed the previously obtained geological, mineralogical and geochemical data that the weathering conditions during the Upper Oligocene-Lower Miocene did not lead to the origin of a well-developed kaolinitic weathering crust, neither in Serbia (MAKSIMOVI] & NIKOLI], 1978; SIMI], 2004), nor in Slovakia (KRAUS, 1989).
The halloysite sample of hydrothermal origin has a rather small mean crystallite thickness of 4.25 nm and a polymodal distribution pattern (Fig. 4).The T BWA values of the Serbian and Slovakian halloysites are very similar, indicating a similar stage of hydrothermal alteration of the primary rocks.The distribution shapes of   these two hydrothermal halloysites are different, as the Slovakian sample has an asymptotic shape and the Novo Brdo halloysite a polymodal one.The polymodal distribution of the Novo Brdo halloysite seems to be a combination of one lognormal and one asymptotic distribution.The asymptotic distribution is typical for samples with small T BWA and could be characteristic for early stages of formation (EBERL et al., 1998a).
The tonstein sample of volcano-sedimentary origin also has a small mean crystallite thickness of 4.32 nm with a polymodal distribution pattern, but similar to lognormal (Fig. 4).The tonstein from the Jasenovac mine is generally weakly crystallized according to the XRD (Fig. 5) and at least two generations of kaolinitic minerals can be observed on the SEM image (Fig. 6), confirming the polymodal distribution shape.
The smectites from the volcano-sedimentary type have higher mean crystallite thickness with an average value of 9.56 nm than the smectites from the sedimentary bentonites with an average value of 5.56 nm.The crystallite size distributions for the volcano-sedimentary samples are lognormal (Fig. 7).Their shapes are quite different from the sedimentary types.The volcano-sedimentary smectites have identical distribution shapes with a theoretical lognormal distribution (Fig. 8A).The distribution of smectites from the sedimentary bentonites is different from the theoretical lognormal shape (Fig. 8B).
The mean thickness of smectites from selected world volcano-sedimentary bentonites varies from 6 to 9 nm (MYSTKOWSKI & SRODON, 2000;MOL, 2001).The relatively wide range of T BWA values of volcano-sedimentary smectites does not support the idea of using it to distinguish the origin in general.However, the measurement of T BWA has sense for the differentiation of the origin of a bentonite in smaller regions, as was observed for the Serbian bentonites.A similar difference was found for in situ volcano-sedimentary and transposed bentonites from middle Slovakia (both types were characterized by [UCHA et al., 1996).Smectites originating from the in situ alteration of andesitic volcanoclastics have higher T BWA values (up to 7 nm) in comparison with smectites originating by the redeposition of alteration products (5.5 nm).
Table 3.Average mean crystallite thickness (T BWA ) of Serbian, Slovakian and selected world kaolinites and halloysites.Values for the Slovakian and selected world kaolinites are from [UCHA et al. (1999).

Conclusions
Sedimentary kaolinites from five deposits in Serbia have a low mean crystallite thickness indicating a poorly developed kaolinitic weathering crust from which these clays were redeposited, a situation similar to Slovak kaolin deposits.The role of crystal disintegration during transport may also influence the crystallite size.The shape of the crystal size distribution is polymodal for all samples, most probably as a result of the presence of different kaolinite generations.
Volcano-sedimentary (diagenetic) tonstein from the Jasenovac coal mine has a very low mean crystallite thickness, typical for a weakly crystallized material, and a polymodal distribution shape, due to at least two generations of kaolinite particles.
Hydrothermal Novo Brdo halloysite also has a very low mean crystallite thickness and a polymodal distribution.
Two diverse shapes of the theoretically lognormal distributions were observed for the smectites.They correspond to different genetic types of bentonites -sedimentary and volcano-sedimentary.The mean crystallite thickness is also different in the sedimentary and volcano-sedimentary bentonites, with an average T BWA of 5.56 and 9.56 nm, respectively.This means that "in situ" alteration of volcanic ash under subaqual conditions led to the formation of well-crystallised smectite with thicker crystallites.
Fig. 1.Changes in the distribution and mean thickness of halloysite (sample NB) after using the incorrect analysed area (10-13° 2 ) in comparison with the recommended area (6-13° 2 ).

Fig. 2 .
Fig. 2. Example of the modification of an XRD pattern before BWA-analysis the by PkChopr program.

Fig. 7 .
Fig. 7. Crystallite size distribution of the smectites obtained by the BWA technique.

Fig. 8 .
Fig. 8.Comparison of the measured (BWA) and the theoretical lognormal distributions of smectite particles.A) volcano-sedimentary type, B) sedimentary type.

Table 1 .
Basic geological features of the studied clays.

Table 2 .
List of clays used for BWA-analysis and the input and output data.