FFT变换与STEM结果分析

2017/05/30

工具：ImageJ

Tool: ImageJ

待分析图样：STEM Pt合金花样

Pattern: STEM or Pt-alloy catalyst

在分析纳米颗粒的结构的时候，有时候会用到STEM。可以直观的看到晶面间距，标定晶面间距以后，就可以根据d值确定hkl值。以下图为例。

STEM is commonly used for measure the d value of the different facets to confirm the crystalline structure. The example is shown in beneath:
4-6before_20160724_1557_08_ADF1_ImagePanel1

1号区域内的间距很容易测定，因为晶面清晰，可以测定多个晶面的间距，除以数目，简单算出d值。那么，2、3区域，虽然肉眼可见花样，但是很难数清楚，并且一旦放大像素就会失真。

The d value in 1 area is easy to be observed. Measure distance between ten or more line ,and then divide by the number of lines, d value of each two facets is calculated out. However, in area 2 and 3, it’s difficult for the low dissolution and crowed lines.

所以采用如下方式处理：

1.选定区域，矩形选框或者自由线条都可以

2.imageJ的菜单栏-Process-FFT-FFT 变换

FFT变换的意思是图样在某个方向上有一个sin函数的亮度变化，一个周期，那么，就会有个亮点正好出现在中间。所以四周亮点到中心的距离代表了这个方向上的波数。选定区域以后，计算机会自动将区域转化为2^n PIX 的正方形，然后计算。imageJ确定两个亮点的间距（中点必经过中心点），得到一共是n个像素（PIX）,用正方形的边长a除以n就得到了一个周期是a/n,即一个明条纹和一个暗条纹的长度是a/n个像素，即d

再用比例尺转化就可以知道d是多少nm。

We can deal with these areas like this:

1.Choose the area by rectangle or free hand selection.

2.menu-Process-FFT

Calculate the distance of two central symmetric white dots, divided by two, we got the value n (PIX). We can get the wavelength of this dark and bright pattern in this orientation. In another word, d=a/n (a is the PIX number of the FFT square’s side length), here, the unit of d is “PIX”. we can turn it into “nm” by scale.

例如这张图2区域经过FFT变换：

Here is FFT square picture of the 2 area:
FFT1