火成岩结构的二维定量化分析方法

近些年,定量化火成岩结构研究表明,利用常规的岩矿鉴定设备,获取不同尺度的火成岩二维岩相学照片,通过肉眼识别矿物颗粒,并借助图像处理和结构分析软件,可以准确地量化火成岩的结构特征。本文结合近些年国内外同行的研究成果,对火成岩二维定量化结构分析方法中常用的多种观测方式优缺点进行了总结。粒度在毫米级以下的火成岩的定量化结构参数,可以用偏光显微镜下的透射光、反射光、阴极发光和电子探针背散射成像中的两种或两种以上观测方式进行分析,并具有较高的精度和准确度。粒度小于0.03 mm的各种镁铁质矿物可用反射光和背散射图进行分析,灰度近似的镁铁质矿物可以利用图像处理软件赋予不同的彩色,提高颗粒间的辨识度。常规偏光显微镜下不易区分的长英质矿物和多数副矿物可用偏光显微镜阴极发光进行分析。粒度在毫米级以上的造岩矿物可以用光片或野外测量的方式进行定量分析。为了方便相关领域学者使用火成岩二维定量化结构分析方法,本文详细列出了具体的分析步骤,并结合一个玄武岩样品中的橄榄石斑晶数据结果,重点分析以下4个方面的问题:(1)如何准确识别矿物颗粒边界;(2)矿物含量和形态的确定;(3)分析区域面积和颗粒数的确定;(4)不同晶体群的区分。分析结果表明,颗粒数100~500颗时,晶体粒度分布(CSD)的截距和斜率、矿物含量、定向程度和粒状矿物的三轴比在误差范围内没有显著区别,但颗粒最大长度和聚集程度会被低估。当颗粒数小于300颗时,晶体空间聚集程度的R值会被高估0.05~0.2,这一点在以往的研究中没有得到充分重视。当颗粒数大于500颗时,所有结构参数都趋于稳定,且精度和准确度都会显著提高。目前多数研究者提供的结构参数往往与观测和统计方式有关,缺乏对应的原始数据,不方便同行间的对比研究,建议学者今后发表相关成果时,提供详细的分析步骤和最原始的数据。分析步骤重点说明包括:(1)聚集矿物边界的识别和处理方式;(2)晶体三维形态的确定方法,样品间CSD参数的变化是否是由形态参数变化引起;(3)能够准确识别的矿物颗粒最小粒度;(4)利用颗粒数较多的样品选取较小的不同区域重复分析3到5个不同区域,评估样品的均一性,并据此估计样品的分析精度。原始数据方面包括:(1)提供同一个样品至少一个不同区域的分析结果,如果是多个作者的研究成果,建议提供至少两人独立分析的结果用来评估数据的精度和准确度;(2)文章正文或附件中应该提供每个样品不同粒度间隔的颗粒数,样品原始的高分辨率矿物轮廓描绘图或图片分析的相关原始参数。火成岩出现复杂晶体群时,定量化的结构参数往往体现的是多种晶体群的混合特征,并且与不同晶体群的比例有关。未来的研究需要结合多种观测方式和微区成分分析重点识别不同晶体群的结构参数,对粒度和成分近似的多种晶体群的识别,还需要开发更多有效的方法,这对准确认识火成岩结构多样性的成因和岩浆作用过程都有重要意义。

Two dimensional quantitative textural analysis method for igneous rock

In recent years, quantitative textural studies of igneous rock have shown that by using conventional rock and mineral identification equipment, one can obtain two-dimensional petrographic photos of different scales of igneous rock, so that mineral particles can be identified with naked eyes. Furthermore, by using image processing and textural analysis software the textural characteristics of igneous rocks can be accurately quantify. In this paper, we summarized the advantages and disadvantages of various observation methods commonly used in the two-dimensional quantitative textural analysis in igneous petrology. The quantitative textural parameters of igneous rocks with grain size below millimeter level can be analyzed by two or more observation methods using polarized microscope with transmission, reflective light or cathodoluminescence and back scatter-electron imaging techniques to obtain digital images of high precision and accuracy. All kinds of mafic minerals with crystal size less than 0.03 mm can be analyzed by reflective light and back scatter-electron image. One can assign different colors to look-alike gray-scale mafic mineral images by image processing software to improve mineral identification. Polarized microscope cathodoluminescence can be used for the analyses of felsic and most accessory minerals which are difficult to distinguish under conventional polarized microscope. The rock-forming minerals with mm or above grain size can be quantitatively analyzed by polished large sections or field measurement. In order to facilitate researchers in related fields to use this method, we listed the specific analysis steps in detail. Some detailed analyses are made by combining the quantitative textural data of olivine phenocryst in a basalt sample, focusing on how to accurately identify the boundaries of mineral particles, to determine mineral content and shape, analysis area and crystal numbers, and to distinguish different crystal population. The results show that the intercept and slope of crystal size distribution (CSD), mineral content, alignment factor and crystal shape of granular minerals are not significantly different for particle numbers between 100-500, but the maximum crystal length and degree of particle aggregation are underestimated. For particle numbers less than 300, the R value of crystal spatial distribution pattern will be overestimated by 0.05-0.2, which were largely ignored in previous studies. All textural parameters tend to be stable and precision and accuracy are significantly improved for particle numbers above 500. At the present, most of the textural parameters provided by researchers are usually statistical results and often related to the observation and statistical methods, and the intercept and slope of CSD lack the corresponding original data, which is not convenient for the comparative study among peers. We suggest that detailed analysis steps be provided with the original data when publishing results related to textural parameters in the future. The recommended analysis steps are as follows: (1) Identify boundaries of aggregated minerals or deal with it concretely; (2) Determine crystal three-dimensional shape and whether CSD parameter change among samples is caused by shape parameter change; (3) Obtain minimum particle size of minerals that can be accurately measured; and (4) Estimate sample homogeneity and precision of analysis on samples with large particle numbers so 3-5 repeated measurements can be performed. The original data should (1) contain analysis of at least two different areas of the sample. For multiple authors, at least two independent analyses by different authors is recommended for precision and accuracy evaluation. And (2) the number of crystals in different crystal size intervals of each sample, the original high-resolution mineral outline of the sample or the relevant original parameters of image analysis should be provided in the text or appendix of the article. When multiple crystal populations appear in igneous rocks, the quantitative textural parameters often reflect the mixed characteristics of multiple crystal populations and are related to the proportion of different crystal populations. Future research needs to focus on identifying textural parameters of different crystal populations by combining multiple observation methods and micro-area composition analysis. More effective methods should be developed to identify crystal populations with similar grain size and composition, which is of great significance for the understanding of the genesis of the textural diversity of igneous rocks and magmatic processes.