Abstract
In order to determine the number concentration of asbestos, it is initially necessary to develop a method to identify the type of asbestos. Thus a SEM/EDX was used to obtain both physical and chemical information from known asbestos samples as reference samples. Based on these information, we could make a source profile matrix consisted of a glass fiber and 3 other types of asbestos such as chrysotile, crocidolite, and tremolite. After collinearity test was performed for these sources, we could successfully develop an expert system by C-language to separate and to identify various unknown types of fiber particles. The expert system was perfectly self-verified with original reference data. Then the program was extensively applied to survey indoor and outdoor environment such as a residential area, an elementary school, and underground store, and an auto junkyard. As a result for surveying, a total of 442 individual fibrous particles were well classified into 4 types of particle classes above mentioned; 5.4% of chrysotile, 4.1% of crocidolite, 3.6% of glass fiber, and 86.9% of unknown fibers in terms of number concentration. However, tremolite was not detected in the study sites. All the samples were satisfied with the recommendation level of 0.01 f/cc.
