Feature Size measurement: Porosity, Grain, and Reinforcement

  1. Take an image from representative area

  2. Make sure brightness/contrast/ sharpness (or focus) is good (user’s choice).

  3. Make sure the magnification is appropriate

  4. Do take low magnification and high magnification images of the same region to see distribution/similarity at low and high magnifications. The details obtained by different magnifications also help to appreciate the detailed features.

  5. Do take ‘similar magnification’ pictures for comparison among different samples.

  6. Lock the Aspect Ratio (without failure)

  7. Crop the unwanted portion (such as bottom part of this image) if that contains the image information.

  8. Insert scale-bar separately ‘after’ you have estimated the grain/feature size. Otherwise the software also grabs the feature information (usually at bottom of SEM images) and might provide incorrect details of analysis. Inserting the scale-bar at later stage provides a better presentation of image.

  9. Ensure that you are able to see the feature you are interested in (such as particle-size or grain size, or the size of second-phase)

  10. Adjust the brightness, contrast, and sharpness of features to ensure that boundary of features do not overlap and will not interfere during the analysis.

  11. Put the grid WITHOUT any bias. Do not orient or try to align or select a specific feature to fall on the grid point. The choice must be random.

  12. Choose appropriate grid size so as the grid is bigger than that of the feature size. Also, a few points of the second phase should be falling on the grid (and not all the points on the grid are falling on the second phase).

  13. Perform the analysis on multiple number of images, and then combine the data (for that sample) to obtain statistically reliable data.

  14. Perform the analysis (as provided for ASTM grain-size, or Jeffries method or grid-count method) on other set of samples and then compare: (i) with data of other sample (with different features/sizes) as that may show influence of processing condition or with different content of reinforcement, etc. and (ii) with data obtained from the software on the same sample. The closeness to this data will help you compare the quality of the analysis that you have performed.