Search

—We present results of petrographic, mineralogical, and geochemical study of all types of rocks of a multiphase pluton and consider the chemical evolution of igneous and metasomatic rocks of the Baga Gazriin Chuluu pluton, based on new precise analytical data. At the early stage of their formation, the pluton granites were already enriched in many trace elements (Li, Rb, Cs, Be, Nb, Ta, Th, and U), F, and HREE relative to the upper continental crust. They show strong negative Ba, Sr, La, and Eu anomalies, which is typical of rare-metal Li–F granites. The geochemical evolution of the Baga Gazriin Chuluu multiphase pluton at the postmagmatic stage was marked by the most intense enrichment of greisens and microclinites with lithophile and ore elements (Sn, W, and Zn) and the formation of ore mineralization. In the permeable rift zone where the Baga Gazriin Chuluu pluton is located, the fluid–magma interaction took place under the impact of a mantle plume. High-temperature mantle fluids caused melting of the crustal substratum, which determined the geochemical specifics of Li–F granite intrusions. Genesis of granitic magma enriched in Li, F, Rb, Sn, and Ta is possible at the low degrees of melting of the lower crustal substratum. The Baga Gazriin Chuluu pluton formed in the upper horizons of the Earth’s crust, where magma undergoes strong differentiation and the saturation of fluids with volatiles can lead to the postmagmatic formation of metasomatites of varying alkalinity (zwitters (greisens), microclinites, and albitites) producing rare-metal mineralization. By the example of the early Mesozoic magmatism area of Mongolia, it is shown that the formation of granites and associated rare-metal minerals is due to the interaction of mantle fluids with the crustal material and the subsequent evolution of granitic magmas.

This article presents first tephrochronological data on the volcanic activity in the valley of the Jom-Bolok River (East Sayan Mountains, Siberia), which is the largest manifestation of the Holocene eruptions in Central Asia. The data results from our study of the proglacial Kaskadnoe-1 Lake situated near the Jom-Bolok basalt field. The lake sediments include a series of tephra-rich layers. Radiocarbon dating of the lake sediments provided a robust age model which allowed us to build timing of eruptions formed the Jom-Bolok volcanic field. We recognize two large phases of volcanism separated by almost 5 thousand years dormant phase. The first phase is traced back to ca. 14.3 cal ka BP and lasted until 6.3 cal ka BP. Ten clusters of microtephra layers in the sediments of the first phase show 300-800 years recurrence of the volcanic events weakening upward. The event of 14.3-13.3 cal ka BP probably represents the strongest eruptions of the Jom-Bolok. The second phase started ca. 1.6 cal ka BP and highly likely continues in our days. Its strongest eruptions occurred between 1.6 and 0.8 cal ka BP with periodicity of 200 years. This tephrostratigraphy shows a multiplicity of the Jom-Bolok volcanic events and amplifies the earlier built scheme resulted from investigations of the stratified basalts, pyroclasts and lake damming events.

The first results of tephrochronological studies of Late Pleistocene–Holocene volcanic eruptions in the Zhom-Bolok River valley (Eastern Sayan) are reported. Based on the analysis of bottom sediments of the glacial lake Kaskadnoe-1, a detailed chronological scale of volcanic activity in this region has been worked out. It is established that the eruptions began at least 14 300 years ago and continued, gradually weakening, to 6280 years ago. In this period, ten groups of peaks of sediment enrichment in microtephra are registered for Kaskadnoe-1 Lake. The average interval between the stages of eruptions in this period was ~500–600 years. Between 14 300 and 13 400 years ago, the most large-scale volcanic events took place, eruptions were practically not interrupted, and the explosive activity was the most intense. The volcanic activity in the region in the period between 6280 and 2000 years ago is characterized by a relative rest. Later, eruptions resumed reaching peak values 1150 years ago and completing 800 years ago.

During the solid-phase synthesis of Be–Mg–Al silicates, i.e., beryllian indialite (BI), the crystal-chemical proximity is due to the transfer of elements of chemical and structural inheritance. BI crystallization from a gel is carried out via the structural adaptation of multicomponent metastable phases transforming into BI, which is implemented in two stages: diffusion and the formation of microcrystallites fed by a melt film. During the sintering of dry oxides, parageneses are transferred from a previous temperature stage to the subsequent one so that several phases of the former stage are retained and the fraction of stable phases increases because of the consumption of intermediate phases. Phase transformations whereby BI is crystalized in the subsolidus region are interpreted from the viewpoint of the evolutionary regularities of mineralogenesis.

The joint crystallization of the stable phase (a number of solid solutions of chromium-containing beryllian indialite) and metastable phases (crystalline modifications) of the compound with β-quartz structure (Si0.64Al0.28Mg0.21Be0.09)IVO2 ~ Mg1.89Be0.81Al2.52Si5.76O18 admixed with Cr2O3 phases of khmaralite Mg1.21Cr0.01Be0.46Al1.78Si3.38O20 and spinel {(Mg0.95Be0.045Si0.005)IV(Al1.31Cr0.67Mg0.02)VI}O4 is performed using melted chromium–beryllian indialite preliminarily obtained by solid-phase synthesis as a precursor. Simultaneously, the residual X-ray amorphous melt of composition Mg1.83Cr0.01Be1.04Al2.64Si5.57O18 is hardened. Thus, a reconstructive transition from the beryllian indialite melt to a phase with the β-quartz structure is implemented, and the chemical similarity of these compounds is demonstrated. The rate of change in the crystallization isotherm of 2°С/h and increased heat outflow through the highly heat-conducting walls of the Pt–Rh crucible (taper) contribute to this process.

The FeS2–Ag–Pt–As system was studied using hydrothermal thermogradient synthesis (with internal sampling) of pyrite crystals at a temperature of 500°C and pressure of 1 kbar in ammonium chloridebased solutions. The modes of occurrence of precious metals (PM) were determined using atomic absorption spectrometry (AAS) in its version of statistical selections of analytical data on single crystals (SSADSC), electron microprobe analysis (EMPA), scanning electron microscopy with energy-dispersive spectrometry (SEM-EDS), atomic force microscopy (AFM), and X-ray photoelectron spectroscopy (XPS). The concentration of Pt in its structural mode in pyrite is as high as 10–11 ppm and is practically not correlated with the As concentration. The dualistic distribution coefficient of Pt between pyrite and hydrothermal solution is 21 ± 7 for the structural mode and 210 ± 80 for the surface-related mode of this element. No inclusions of either any Pt-bearing minerals or Pt itself was detected. Platinum is an element highly compatible with hydrothermal pyrite and is different in this sense from gold, and pyrite is underestimated as a potential concentrator of platinumgroup elements (PGE). The distribution of Ag in pyrite is highly heterogeneous. The likely reason for this is that the Ag solid solution cannot be quenched, and hence, the Ag concentrations broadly vary and are very unsystematically distributed in natural pyrite crystals. Assuming this hypothesis, the limit for Ag accommodation in FeS2 can be estimated using SSADSC at 0.09 ± 0.06 wt % under the experimental parameters, and the distribution coefficient of the structural Ag mode is thereby evaluated at 1400 ± 700. When crystallizing together with FeS2 proustite (Ag3AsS3) near its melting point, forms mixtures with dervillite (Ag2AsS2), in which Ag deficit is counterbalanced by excess divalent As. The limit of As incorporation into pyrite under these conditions is ≤0.1 wt %. SEM-EDS and XPS data indicate that the surface phases are of three types. In the course of crystal growth, practically two-dimensional nonautonomous phases (NP) are aggregated into submicroscopic and micrometer-sized crystalline bodies (mesocrystals) that largely inherit their unusual minor-element composition from NP and are enriched in Ag, Pt, As, and other minor elements. NP and mesocrystals are enriched in Al, which was transferred into them from the Al-bearing Ti alloy of the reaction containers. Silver occur in the volume of the crystals and on their surface as monovalent silver sulfide. Arsenic was detected mostly in the form of di- and trivalent arsenic sulfides. Pentavalent arsenic oxide was identified only on the surface of the crystals and can be easily eliminated by ion milling.

Results of assessment of the stability and homogeneity of the Be−Mg−Al−silicate glass synthesized by the authors and possibilities of its use as a quality control material (QCM) in the X-ray electron probe microanalysis (EPMA) of Be-bearing silicate materials, i.e., crystals and quenched melts (glasses), and also silicates and oxides are presented. The homogeneity of the samples was studied at the macro- (10–100 μm) and microlevels (1–10 μm) and assessed according to the scheme of dispersion analysis. A possibility of using the Be–silicate glass as a certified reference material for the determination of the concentrations of Mg, Al, Si was estimated using international reference materials of glasses and QCM of minerals of the known composition. The metrological performance of the experimental data obtained suggest that the studied glass can be used as a QCM in the EPMA of Be-bearing silicate materials, silicates, and oxides. The use of the Be-silicate glass as a certified reference material of composition in EPMA ensures acquisition of satisfactory data on the composition of minerals including cordierite and beryllium cordierite, beryllium indialite, beryl and also of metastable phases of chrysoberyl and compounds with the structure of β-quartz and petalite.

The Central Asian orogenic belt (CAOB) is one of the largest orogens on Earth characterized by complex and long-term history resulted from the development of the Paleo-Asian Ocean. In Late Cambrian–Ordovician, closure of Paleo-Asian basins and accretion of microcontinents, island arcs and back-arc basins to the southern margin of the Siberian continent resulted in large-scale magmatism recorded in Southern Siberia and Mongolia. The timing of terminal stage of subduction and start of accretion-collision stage is a matter of discussion for different parts of this extensive collage of terrains. In this paper, we report new data on whole-rock and mineral composition and 40Ar/39Ar age of calc-alkaline dolerites, lamprophyres and gabbros of Bayangol river area (Northern Mongolia) from the Dzhida terrain of CAOB. The Bayangol calc-alkaline mafic rocks were previously considered as a member of boninite-basaltic sequence originated in a primitive island arc. We found that the Bayangol calc-alkaline mafic rocks compose late mafic dikes, cutting terrigenous sediments and serpentinite melange with boninite blocks, and, therefore, calc-alkaline rocks are not related to boninites. We obtained 40Ar/39Ar age of 498.9 ± 7.7 Ma for amphibole from Bayangol lamprophyre. The calc-alkaline rocks preserve clinopyroxene (diopside, Mg# of 76–90) and amphibole (mainly pargasite and magnesiohastingsite, Mg# of 38–92), which crystallized at 1170–1210°С, 2.2–9.5 kbar and 760–1030°С, 2.0–7 kbar, respectively. Dolerites, lamprophyres and gabbros show variations in major oxides (3.3–14.9% MgO, 10.1–19.4% Al2O3, 0.09–0.92% P2O5) and immobile trace element systematics (e.g., Th and LREE enrichment, HFSE depletion) corresponding to island arc calc-alkaline series. Isotope compositions of Nd (eNd(T) = −0.7 to +2.5) and Sr (87Sr/86Sr(T) = 0.7049–0.7078) suggest 1–5% addition of recycled terrigenous sediment melts to mantle wedge source of the Bayangol mafic rocks. Relations of trace element and SiO2 abundances and Nd Sr isotope ratios are resulted from fractional crystallization and do not support in situ crustal contamination of Bayangol magmas. Within the Dzhida terrain, the Bayangol mafic rocks have Nd isotope compositions transitional between those of Late Cambrian island arc intrusions with juvenile signatures and Early Ordovician collisional granitoids with crustal characteristics. Therefore, the Bayangol calc-alkaline mafic rocks mark transition from subduction to the accretion-collision regime during accretion of Dzhida island arc to the Siberian continent as a result of Paleo-Asian Ocean closure in SW Transbaikalia and Northern Mongolia.

The typomorphic features of pyrite of the Sukhoi Log deposit were studied by a set of volumetric and surface methods: electron probe microanalysis, scanning electron and probe microscopy, powder X-ray diffraction, X-ray photoelectron and Auger electron spectroscopy, atomic-absorption spectrometry in the SSADSC (method of statistical sample of analytical data for single crystals) version, and atomic-emission spectrometry. Pyrite from the Sukhoi Log deposit has the following distinctive features: permanent presence of sulfite ion, which often dominates over other surface sulfur anions; weakly determined size dependence of the content of uniformly distributed Au owing to the presence of an internal concentrator of gold—dispersed carbonaceous material—in pyrite from ore zones; cell sculptures of the crystal faces, which appeared owing to the nanofragmentation of the growth surface; micro- and nanoinclusions of carbonaceous phases within crystals, associated with defects in their structure; and thin films enriched in O and C on the surface of and within the crystals. It has been shown that gold–sulfide mineralization at the Sukhoi Log deposit formed in a single ore-generating hydrothermal system, in which gold, sulfur, and carbon belonged to a microparagenesis. Some features (composition of surface, characteristics of submicroscopic structure, and elemental composition) evidence that the conditions of crystallization of pyrite in inter-ore space were different from the conditions of its genesis in the ore zones, which suggests the presence of at least two genetic types of pyrite. Carbonaceous micro- and nanoparticles and O- and C-containing films can favor an increase in the adsorption of gold from cyanide solutions on pyrite. To reduce this effect during gold recovery, a technique for surface modification should be elaborated. The ways for solving the most complicated problems dealt with the source of noble metals (NM) and the ore specialization of the deposit have been outlined. For this purpose, a detailed analysis of the main ore minerals for trace-element speciation is required. In the case of the magmatic source of NM, correlation between the contents of Au and PGE structural forms should exist. On the other hand, there is no correlation between the structural forms of Au or Pt and elements whose contents in fluid are determined by the host rock rather than the magmatic source.

It is established that solid-phase crystallization of beryllian indialite (BI), obtained by heating a sol-gel precursor of the initial compound of nonstoichiometric composition Mg1.81Be1.09Al2.27Si5.85O18, occurs through the following successive reconstructive transformation: phase with a quartz structure → phase with a petalite structure → BI. It is found that metastable phases with quartz and petalite structures, filled with excess cations, are transformed into equilibrium BI, which has no cationic excess, during this process. It is revealed that the structural aspect of the relationship between the phases involved in the transformation is in the inheritance of a hollow Si6O18 ring of preceding phase by subsequent phases and in the edge aggregation of polyhedra. The chemical relationship is in the approximately identical amounts of all species-forming components in the phase compositions.

ICPMS analyses of encrusting sponges indicate that their predominant chemical elements are, along with Si, by P, Al, Fe, Ca, S, Mg, K, Na, Cu, Mn, Zn, Ti, Ba, and Br. The sponges are most signifi� cantly enriched in Al > Cu > Ti > REE > Mn > P relative to their aqueous habitat and in Cu > I > Cd > P > Br > As ≥ S relative to the relatively rudaceous bottom sediments. One of the sources of elements occurring the aqueous habitat and being of vital importance for the activity of the sponges was proved to be the rock substrates.

The Saramta peridotite massif is located within the Sharyzhalgai complex, SW margin of the Siberian craton. The Saramta massif was formed in the Archean and then juxtaposed with granulites of crystalline basement of the Siberian craton. The Saramta harzburgites are highly refractory in terms of lack of residual clinopyroxene, olivine Mg-number (up to 0.937), and spinel Cr-number (∼0.5), suggesting high degree of partial melting. Detailed study of their microstructures shows that they have extensively reacted with a SiO 2 -rich melt, leading to the crystallization of orthopyroxene, clinopyroxene, amphibole and spinel at the expense of olivine. The major element compositions of the least reacted harzburgites are similar to the residues of refractory peridotites produced by the fractional melting (initial melting pressures >3 GPa and melt fractions ∼40%). Moreover, non-residual clinopyroxenes are highly depleted in Yb, Zr and Ti, but highly enriched in LREE. A two-stage history is proposed for the Saramta peridotite: (1) primitive mantle underwent depletion in the garnet stability field followed by melting in the spinel stability field; (2) refractory harzburgites underwent refertilization by SiO 2 -rich melt in supra-subduction zone. Rare Saramta lherzolites probably formed from more refractory harzburgites as a result of such a melt–rock reaction. The Saramta peridotites are similar to low-T coarse-grained peridotites of subcratonic mantle. Processes of their formation, as reflected by textures and composition of minerals of the Saramta peridotites, are characteristic of the early stages of subcratonic mantle formation.

The system magnetite-Au-hydrothermal solution was employed to continue studying the distribution coefficients of trace elements in system with real crystals. The role of surface nonautonomous phase (NP) is elucidated. The distribution coefficient of an Au structural admixture between magnetite and hydrothermal solution at the experimental conditions [450°C, 1 kbar (100 MPa), and fluid sampling by a trap] is, according to the most representative data, 1.0 ± 0.3, and Au is thus not an incompatible element in magnetite, in contrast to pyrite and arsenopyrite [1], minerals for which this coefficient is much lower than one. The NP is enriched in Au with respect to the rest of the crystal by a factor of more than 4000, and this results in an one order of magnitude increase in the bulk distribution coefficient. Similar to pyrite, the reason for the dualistic nature of the distribution coefficient is the presence of an NP, which contains ∼2000 ± 500 ppm Au. The NP occupies the approximately 330-nm surface layer of the crystal, and the chemically bound Au [Au(III), according to XPS data] admixture is evenly distributed with depth within the layer, which is the reason for the strongly determinate dependences of the concentrations of the evenly distributed Au admixture on the size and specific surface area of the crystal. The occurrence of an NP is controlled by the chemistry of the system. The partial substitution of Fe for Mn and the synthesis of a phase close to jacobsite MnFe2O4 results in the disappearance of both the NP itself and the size dependence of the Au concentration. The XPS spectra of O 1s and Fe 2p are used to analyze two models: (i) a single goethite-like (O2−/OH−∼ 1) phase of variable composition and Fe in more than one valence state and (ii) a heterogeneous structure of alternating domains of wuestite- and goethite-like NP. The reason for the “excess” admixture in the former instance can be vacancies at Fe sites, whereas that in the latter one is the interaction of the admixture with nanometer-in-size nanometer in-size strained domains on the surface of the crystal.

The use of trace elements (TE) as geochemical indicators is complicated by the dualism of their distribution coefficients D due to the additional (i.e., above the concentrations of an isomorphic component) incorporation of elements at structural defects of various nature (including the surface of the crystal). A pressing problem in this situation is to determine the true D values that pertain to the structural component of an admixture D str and evaluate effects of other modes of TE occurrence. Only upon distinguishing D str in the bulk coefficient D bulk it is possible to evaluate the ore potential of fluid in terms of certain TE from the composition of a mineral containing the TE. Pyrite synthesized in solutions of variable pH at 450°C and 1 kbar (100 MPa) at fluid portions sampled in a trap is utilized to demonstrate the role of a surface nonautonomous phase (NP) in the incorporation of gold in this mineral. The distribution coefficient of gold between pyrite and hydrothermal solution is 0.14 for “pure” pyrite and 0.05 for As-bearing pyrite (containing 0.02–0.05 wt % As), and these coefficients for NP are 310 and 170, respectively. This increases the D bulk for evenly distributed (“invisible”) gold by factors of four and nine. In contrast to the results of earlier studies conducted at room temperature and pressure or parameters close to them, our data demonstrate that the accumulation of “invisible” Au in pyrite is controlled not only by reducing adsorption with the development of Au(0) particles and films but also by Au incorporation in NP developing in the surface layer of the crystal approximately 500 nm thick as chemically bound Au [most likely as Au(I)]. The possible reason for the high absorption capacity of NP is the defect (pyrrhotite-like) structure, which is not saturated with bonds of excess S and sulfoxi onions.

The dependence of the intensity (I i ) of X-ray fluorescence on the size (D) of finely ground particles was studied for saturated and unsaturated samples. It was found that even at the wet grinding (addition of ethanol) of powders with D < 10 μm, the aggregation and covering of larger grains (α) with smaller grains of different compositions occur, which changes the character of the dependence I i = f(D), particularly, if fluorescence is emitted by the grains α. The nature of the observed effects is proved by the results of granulometric analysis and by electron probe X-ray analysis. In the transition from saturated to unsaturated samples, the discussed effects are enhanced.

The face effect was observed in all individual compounds with beryl structure having grown at a rate of 1–8 unit cells a second, with spontaneous crystallization in the system beryl (or cordierite)–Mg,Ca/Cl,F–impurity, where impurities are Ti, V, Cr and/or Fe oxides. By “face effect” is meant a difference in chemical compositions and symmetry of growth pyramids of crystallographically nonequivalent crystal faces. In the studied compounds of beryl structural type the simple crystal forms, prism and pinacoid, are characterized by hexagonal (pseudohexagonal in cordierite) and rhombic symmetry of face patterns. Degrees and characters of distinction of growth pyramids of prism and pinacoid faces of beryls, beryllian indialites, cordierites and their varieties vary in a wide range depending on the accumulation of species-forming and impurity elements. Most likely, this effect appears when crystallographically nonequivalent faces entrap chemically different “building units” (clusters spontaneously formed in a feeding medium and composed of fragments of beryl-type structure). It is supposed that the final isomorphous substitution of an impurity for the species-forming element occurs immediately on the face and/or attached unit. The character, direction, and degree of isomorphism depend on the initial composition of the medium, type of the main compound to be formed, and the presence of synchronous impurity phases. The same ratio of Si and Al concentrations in building units of the prism-pinacoid couple of cordierite with and without Ti and V suggests the persistent action of the patterns of these faces on the medium in the range of sampled concentrations of impurities and rates of growth. The contrasting face effect with respect to Si and Al in an indialite-cordierite individual is due to a drastic dissymmetry of the latter. The Ti- or V-related effects of cordierite faces are negligible. The study of beryllian indialite and its varieties showed that in some cases changes in Si-Al-Mg composition of units feeding faces with different crystallographic patterns are quite the same, with the considerable evolution of the total composition of the crystal-forming medium. As to the face effects of beryllian indialites (chiefly, with respect to Ti, V, Cr, and Fe impurities), they are owing to a considerable difference in degree of isomorphism with the participation of these elements in the building units that feed pinacoid and prism faces. Beryls demonstrate drastically contrasting face effects with respect to Ti (up to 0.20 f.u.) and in the presence of impurities, to Al (up to 0.21 f.u.); distinct effects were related to Cr (0.06 f.u.), in all varieties to Mg (up to 0.10 f.u.), and quite rarely, to Fe (0.01 f.u.), which is seen from comparison of the schemes of isomorphism implemented in prism and pinacoid. The effects under consideration and their evolution in the process of development of individuals are also clearly recognized from crystallochemical formulas of the building units feeding faces with different crystallographical patterns. By the example of beryl, it has been shown that the face effect inversion with respect to a component during the crystal growth indicates the inversion of the degree of isomorphous substitution of this element in the process of evolution of the individual (because of a drastic change in phase formation in the system) rather than the absence of relationship between face pattern and effect. Thus, in the system beryl–Mg,Ca/F,Cl–TiO2 the inversion of Mg-related effect (resulting from inversion of degrees of substitution Al3+ → Mg2+) is linked to a drastic increase in rutile amount (TiO2) crystallized together with beryl at a certain stage of solution cooling.

The boundaries of the regions of solid solutions with beryl-type structures and different symmetry and degree of order obtained by solid-phase synthesis have been refined in the beryllium indialite-cordierite-beryl system. The change in the compositions of real crystals in the course of synthesis and the character of the dependence of the melting point on the composition are taken into account.

A set of procedures was developed for the electron probe X-ray microanalysis (EPMA) of fish and animal bone tissues, solid snow sediments, and coal ashes. Various sample-preparation procedures and two variants of the optimum conditions for analytical measurements were considered. In calculating analyte concentrations, matrix effects were taken into account using ZAF methods and the function of generated X-ray distribution with depth, namely, by the PAP method and by adding a joint correction for absorption and atomic number within a biexponential Φ(ρz) model. For EPMA of small fractions 1–5 μm in size, an approach was proposed that takes into account the test sample size using an analytical expression. In calculations of concentrations for micron-sized particles, this approach can lower the relative deviation from the stoichiometric composition from 36.8% to 22.2%. In developing the procedures, the potentialities of the proposed technique were assessed and its metrological characteristics were determined.

Crystallization of the beryllium indialite (Mg2BeAl2Si6O18) melt is investigated. Crystallization at high cooling rates and thermal gradients gives metastable compounds (petalite structural type, mullite, sapphririne); under slow cooling and in the absence of thermal gradients, stable beryllium indialite crystallizes. The main reason for that, suggested from structural motifs of the formed phases, is that most of the magnesium exists in the melt as MgO4 in the former case or as MgO6 in the latter case. The same (petalite) structural type crystallizes from liquates differing only in a ratio of type-forming elements. The composition of metastable phases is determined.

A technique is proposed for X-ray electron-probe microanalysis (EPMA) of individual particles of solid snow sediment. Two sample preparation procedures were used to analyze the sediments. The PAP-method and the original bi-exponential model were used for matrix correction and the calculation of elemental composition. A method of calculating composition has been developed for approximately spherical individual particles, comparable in size with an area of X-ray generation. An analytical expression is proposed for determining the composition of these particles, taking account of their size. Including the size factor reduces the error of composition determination from 0.5–45 to 0.2–22% relative percent, for particles in the size range 1–3 μm.