连江黄石、山仔濑石和伊利石质田黄的矿物组成及微量元素特征研究

Mineral Components and Trace Element Characteristics of Lianjianghuang Stone, Shanzilai Stone and Illite Tianhuang

  • 摘要: 连江黄石、山仔濑石因其外观与伊利石质田黄极为相似,所以一直受到了人们的关注。采用显微镜、X射线粉末衍射仪、红外光谱仪、显微激光拉曼光谱仪、扫描电子显微镜、能谱仪和激光剥蚀等离子质谱仪等进行测试,对其矿物组成和微量元素特征进行了系统性研究。结果表明,连江黄石、山仔濑石与伊利石质田黄的主要矿物都是伊利石,但三者的主要差别是次要矿物的不同。伊利石质田黄的次要矿物有不定形碳和少量呈点状、浸染状分布的赤铁矿,连江黄石的次要矿物有呈浸染状和结晶程度较好的赤铁矿,以及聚集状分布的独居石,山仔濑石的次要矿物成分种类复杂,有呈浸染状和结晶程度较好的赤铁矿、独居石、重晶石、锐钛矿、磷钡铝石、刚玉以及石英。伊利石质田黄及其相似品连江黄石和山仔濑石的微量元素Ga-Sn和Ga-Ge两组二维投点图能较好的区分三个品种,而Ga-Ge-Sn和Sc-Cs-Ge的三维投点图的区分程度更加有效;另外利用微量元素成分判别分析建模对三种寿山石的分类正确率可达到97.7%,说明模型的分类效果较好,较为可靠,但该模型还需要更多的样品数据来完善。

     

    Abstract: The excavated stones from Shoushan known as Lianjianghuang stone and Shanzilai stone have sparked substantial curiosity due to their aesthetic resemblances to the illite Tianhuang stone. The mineral components and trace element characteristics of these stones were studied using microscope, X-ray powder diffractometer, microscopic laser Raman spectrometer, Fourier transform infrared spectrometer, scanning electron microscope, energy dispersive spectrometer, and laser denudation plasma mass spectrometer (LA-ICP-MS). The results indicate that illite is the predominant mineral in the Lianjianghuang stone, Shanzilai stone, and illite Tianhuang. The minor minerals' peculiarities are what set them apart from one another. Hematite and monazite are the minor minerals in Lianjianghuang stone; hematite, monazite, barite, anatase, barium, corundum, and quartz are the minor minerals in Shanzilai stone; and amorphous carbon and hematite are the minor minerals in illite Tianhuang. Ga-Sn and Ga-Ge have better differentiation of two sets of two-dimensional projection plots, and Ga-Ge-Sn and Sc-Cs-Ge have more direct and effective three-dimensional projection plots. The discriminant analysis approach achieves 97.7% classification accuracy for the three types of Shoushan stones, and the separation degree of the group centroid is good, indicating that the model's classification ability is good and dependable. But the model still needs more sample data for further optimization.

     

/

返回文章
返回