Skip to content

Crash course on GIWAXS data analysis

Intro

GIWAXS data analysis has 5 steps in general:

GIWAXS data analysis

  • Calibration
  • Data reduction
  • 2D image generation
  • Area integration
  • Peak fitting

More advanced data analysis includes:

  • Peak search
  • GIWAXS pattern simulation
  • Phase identification
  • ...

Calibration

Calibration is a method to determine the sample-detector distance and the relative detector postion. A material (thin film, plate, or powders on a flat surface) with known lattice parameter is measured at high incident angles (<2 degrees). The collected 2D image from the calibration material is used to derive the detector position.

The quantitative calculation can be done using a python package 'pyFAI'. pyFAI is written in python2, so the installation is complicate. Here is a guide for pyFAI installation:

pyFAI installation

  • Do not install pyFAI on your PC. Use NSLS-II jupyterhub if you have the access. Most of the packages you will need for GIWAXS data analysis is installed there.
  • Use SMI_analysis package. This is a more advanced GIWAXS analysis package developed by beamline scientist at NSLS-II SMI beamline. pyFAI is included in SMI_analysis package.
  • Use pyWAXS package developed by Toney Group. You will need to be added to ToneyGroup on Github to get access. (Recommanded)
  • Directly install pyFAI on your PC. Good luck!

If you have successfully installed any of the package above, you can start to do the calibration calculation. In the terminal, type:

conda activate smi_analysis (or pyWAXS)
pyfai-calib2

The pyFAI calibration UI will be opened: image1

You will need to input:

Parameters

  • Detector
  • Caliration materal
  • X-ray energy
  • 2d image for calibration (.tif)

to do the calibration calculation. The .poni file will be generated for the use in the next step (data reduction).

Data reduction

The data reduction process convert the 2d detector image to the scattering instensity in the reciprocal space (qxy vs qz). This conversion considers the Ewald sphere correction with the experiment geometry. A missing wedge will appear after this step (if you use a grazing incident geometry).

A simple data reduction code is below:

import numpy as np
import os
import pyFAI, fabio
import matplotlib.pyplot as plt
import matplotlib as mpl
import pygix
# from pylab import *
import scipy.io as sio
import scipy.integrate as integrate
import image

# %matplotlib inline
print("Using pyFAI version", pyFAI.version)

def pygix_reduction(dirr,filename,poni_dirr,poni_filename,orientation,incident_angle):
    path_poni=os.path.join(poni_dirr,poni_filename)
    pg = pygix.Transform()
    pg.load(path_poni)
    pg.sample_orientation =  orientation
    pg.incident_angle = incident_angle
    pg.tilt_angle = 0
    path=os.path.join(dirr,filename)
    data = fabio.open(path).data
    ii_reciprocal, qxy, qz =pg.transform_reciprocal(data, method = "bbox", unit = "nm")
    return ii_reciprocal, qxy, qz

dirr="temp"
filename="DD_S9_12110220_0001.tif"
poni_dirr="temp"
poni_filename="Demi.poni"
orientation=3
incident_angle=3
ii_reciprocal, qxy, qz=pygix_reduction(dirr,filename,poni_dirr,poni_filename,orientation,incident_angle)
# show reduced image
ii1=np.array(ii_reciprocal)
plt.imshow(ii1,
          extent=(np.min(qxy)/10,np.max(qxy)/10,np.min(qz)/10,np.max(qz)/10),
          origin = "lower")  
# plt.xlabel('q$_{xy}$ (1/A)')
# plt.ylabel('q$_{z}$ (1/A)')
plt.title('DD_S9')

# set color bar
plt.set_cmap('jet')
lb = np.nanpercentile(data, 10)
ub = np.nanpercentile(data, 90)
plt.clim(lb,ub)

image1

Area integration