Sorbents based on Fe(OH)₃ and aluminum oxide are widely used in oceanology for the recovery of cosmogenic radionuclides ⁷Be, ³²Si, ³²P, and ³³P from the seawater. It is also possible to use them for the recovery of the natural radionuclides ²¹⁰Pb, ²³⁴Th. A comparative study of the sorbents based on Fe(OH)₃ and acrylic fiber obtained through various impregnation methods was carried out, and their comparison with granulated aluminum oxide. The possibility of extracting trace amounts of phosphorus and beryllium under laboratory and field conditions with these sorbents was studied. The sorption of ⁷Be, ²¹⁰Pb, and ²³⁴Th on the natural content by the two-column method was investigated. It is shown that fiber samples obtained by oxidation with sodium ferrate and the “classical” method have the highest sorption characteristics.

The paper investigates the physicochemical regularities (kinetics and isotherm) of phosphorus and beryllium recovery by sorbents based on polyacrylonitrile (PAN) fiber and Fe(OH)₃ obtained by various methods: PAN or pre-hydrolyzed PAN with precipitation of FeCl₃ with ammonia, using ready-made or electrochemically generated Na₂FeO₄, pre-hydrolyzed PAN treated with an alkaline solution of Na₂FeO₄, as well as their comparison with granular aluminum oxide. The Langmuir, Freudlich and Dubinin–Radushkevich models show high performance of materials for sorption of stable P and Be used as tracers for the release of ⁷Be, ³²P, and ³³P from seawater. The obtained kinetic data are processed using kinetic models of intraparticle diffusion and the pseudo-first-order, pseudo-second-order, and Elovich models. Optimal conditions for obtaining sorbents are established, namely, the effect of NaOH concentration at the stages of preparation on the properties of sorbents based on the PAN fiber and Fe(OH)₃ obtained by various methods.

The sorption efficiency of phosphorus from seawater by aluminum oxide and sorbents based on Fe(OH)₃ obtained by various methods (using prepared sodium ferrate or precipitation of Fe(OH)₃ with ammonia) was assessed. It was shown that phosphorus was recovered most efficiently at a seawater flow rate of one-to-four column volumes per minute with a sorbent based on hydrolyzed polyacrylonitrile fiber with a precipitation of Fe(OH)₃ with ammonia. Based on the results obtained, a method for phosphorus isotopes recovery with this sorbent was suggested. Using this method, the seasonal variability of phosphorus biodynamics in the Balaklava coastal area was estimated. For this purpose, the short-lived isotopes of cosmogenic origin ³²P and ³³P were used. Volumetric activity profiles of ³²P and ³³P in particulate and dissolved forms were obtained. Based on ³²P and ³³P volumetric activity, indicators of phosphorus biodynamics were calculated: the time, rate, and degree of phosphorus circulation to inorganic and particulate organic forms. In spring and summer, elevated values of phosphorus biodynamic parameters were determined. This is explained by the peculiarity of the economic and resort activities of Balaklava, which negatively affect the state of the marine ecosystem. The results obtained can be used to assess the dynamics of changes in the content of forms of dissolved and suspended phosphorus, and the biodynamic parameters when performing a comprehensive environmental assessment of the state of coastal waters.

The Russian-Chinese parallel corpus (RuZhCorp) was created in 2016 by sinologists and computational linguists. So far, it has accumulated 1 074 texts and over 4.6 million words that are aligned on a sentence level. To produce word alignment for the entire corpus, we used deep neural networks trained both on the whole RuZhCorp and on a manually aligned at a word level gold dataset. Using the principles presented in previous publications, we compiled the first word-to-word alignment guideline for the Russian-Chinese language pair, which makes the manual alignment process less ambiguous and more consistent. The joint fine-tuning of the LaBSE deep learning model on RuZhCorp and the gold dataset achieved the best AER of 18.9%.