熔盐相图智能数据库研究

建立了包括三千多个已测定的二元、三元熔盐系相图的熔盐相图智能数据库。在建库过程中 ,运用原子参数 -模式识别方法总结已知相图几何特征与其组分元素的原子参数 (离子半径和电负性 )关系的半经验规律 ,并将其用于未知相图几何特征的预报。还将若干不符合上述规律的已测相图列为可疑对象 ,对其重新作实验测定。通过这一做法 ,纠正了前人报道的忽略了某些中间化合物的相图 ,证明这些相图存在CsCaBr3 ,Cs2 CaF4等中间化合物。另一方面 ,还通过实验否定了前人声称发现的KNO3 -KI等相图的中间化合物。这说明 :原子参数 -模式识别算法和实验验证相结合 ,是一种对有疑问的相图评估的有效方法。熔盐相图智能数据库还具有根据原子参数 -模式识别方法对部分未知相图的若干特征的预报功能。初步研究结果表明 :这类智能数据库的研究方法对水 -盐系等相图研究也有应用价值。

Intelligent Data Base of Phase Diagrams of Molten Salt Systems

An intelligent data base of the phase diagrams of molten salt systems has been built in our laboratory.It includes more than 3000 known phase diagrams of binary and ternary molten salt systems for retrieval.In the course of building of this data base,atomic parameter-pattern recognition method has been used to extract the semi-empirical rules about the relationships between the geometrical features of phase diagrams and the atomic parameters(ionic radius and electronegativity)of the constituent elements of salts.Based on these semi-empirical rules found,the geometrical features of unknown phase diagrams can be predicted.Besides,some published phase diagrams are found to be the "exceptions"(outliers)of these semiempirical rules can be considered to be suspectable and should be re-measured by experimental work.By this way,a series of published phase diagrams has been correted in our laboratory.Some phase diagrams have been found to include more intermediate compounds,such as CsCaBr 3 and Cs 2CaF 4;while some intermediate compounds in KNO 3-Kl,KNO 3-KBr,KNO 3-KCl systems claimed by some authors have been disproved in our work.So it is reasonable to consider that atomic parameter-pattern recognition method together with experimental re-determination is an effective way for the assessment of phase diagrams.