Portland cement is a controlled chemical mixture, comprising calcium, silicon, aluminum, iron and low quantities of other components. The mixture is combined with gypsum in the final grinding process to modulate the setting time. Lime and silica constitute around 85% of the mass. Limestone, shells, chalk combined with shale, clay, slate or blast furnace slag, silica sand, and iron ore are common materials implemented in its production.

The robustness and setting time of cement is governed by the efficiency of the chemical reactions among its components. At the same time, however, this efficiency is strongly affected by the particle shapes of the constituent materials. Consequently, a prediction of the behavior of the final product necessitates a comprehensive understanding of the particle shape composition.

Conventional non-imaging size analysis equipment will record sizes based on the assumption that all particles are round or spherical. Such data might thus distort the manner in which those particles will flow in production, or operate in their eventual usage.  This assumption that all particles are round will also misinform the user of the surface area properties of the particles being used which can greatly impact the rate of the chemical reaction.

Cement sample comprises of these multiple categories of particle shapes. The challenge is to utilize the power of Dynamic Image Analysis to identify and quantify the particle sub-populations, according to size and shape.  Of course, having a thumbnail image of all particles serves as objective evidence of what the actual particles are.  No need for guessing and no need for a scientific expert to analyze statistics to identify what is in your sample.

Procedures and Results

The following Particle Insight data indicates a selection of the size and shape outcomes for the Cement sample. These represent overall results for the total sample.  Statistics for each different shape measurement is available allowing users to establish pass/fail criteria not only on size, but other shape parameters as well.

To show how to utilize Smoothness to distinguish particle subgroups, the total sample has been separated into a “smooth” and a “rough” particles by selecting a smoothness threshold.  Users can test their assumed thresholds by confirming with the thumbnail images for their selection.  Again, no experts required.  Observing the thumbnail particle images, the logic of this separation becomes obvious; the two subgroups are discernible. The Smoothness measure will then be able to efficiently distinguish these subgroups. In the following screenshots, 0.84 is utilized as the Smoothness dividing line. The Comprehensive Statistics form exhibits typically implemented means and measures of spread.

In addition to the statistics overview presented herein, the Particle Insight post-processing function enables the presentation of the total distribution of any, or all, of the measures utilized, for all particle subgroups. It also permits the observation of particle thumbnails and two measure correlation diagrams.

Taking the “very smooth” subgroup as the sole focus, a lower limit of 0.89 on Smoothness can be implemented in the post-run processing function. To specifically examine the “very rough” subgroup, a lower limit of 0.60 on Smoothness can be implemented. It is also possible to observe an interdependence between particle smoothness and generalized size.

Conclusions

Examining the cement sample more closely, three or more particle subgroups, characterized by size and shape, could be separated out using Dynamic Image Analysis.  Using data-filtering criteria they Particle Insight system is capable of classifying and quantifying the amount of each particle sub-group.

The “stick-like” objects would possess low circularity with medium smoothness. The particularly irregular shapes could be distinguished with the use of the smoothness measure. Additional modifications could be introduced by filtering the majority, or even all, of the measures utilized.