Particle Size And Shape Analysis With Light Scattering

Particle material is widely applied to medicine, chemical, metallurgy, electronics, mechanics, food, architecture, and environment protection. Many characteristics of particles are associated with particles' size and shape.Compared with other particle sizing techniques, light scattering method has many advantages: high speed; good reproducibility; wide measurement range; and no-destructibility. As to this promising method, more attention is paid to the measurement precision increase, inverse algorithm improvement, and on-line application, nowadays.The theory of particle size and shape measurement is systematically studied in this paper. The model of particle measurement with light scattering is set up, and its computer simulation is realized. This virtual experiment is an effective tool for instrument development. Thus, researchers can perform experiments on multiple factors easily.The particle size inversion algorithm plays an important role in laser sizer, it embodies the sizer's level. Various independent algorithms are studied in this paper. Preconditioned method can enhance the Chahine iterative algorithm's stability.Generally speaking, to increase the inverse size classes' number will benefit the measurement precision. However, the high inverse size classes' number makes inversion results more instable. In order to overcome the instability, two new inverse algorithms (Smooth Chahine and Smooth No-negative Projection) are proposed. The simulation and measurement show that these two algorithms can improve the instrument's precision remarkably without change of the original physical configuration of instrument. The new algorithms can improve the precision of multi-peak distribution more than twice. Commonly, the laser sizers are all on the supposition that the systems can focalize ideally. In other words, when there are no particles in sample cell, the incident light would focalize to an abstract point on the detector. In fact, the incident light can only focalize to a small facula. Considering the model adaptability, the researchers use the optics collimation system to ensure the parallelism of incident light, which increases the cost of instrument's development greatly. All the while, there is no quantitative description about the measurement precision influence because of no-ideal focalization. The light scattering model with no-ideal focalization is developed in this paper. The simulation and measurement show that the expansion of focus decreases the laser sizer's resolution and make the measuring particle size distribution incorrectly broadened to the small size direction, which is distinct for large particles. The systematic error can be corrected through the integral light energy coefficient matrix. Thus, the intractable problem can be solved through software correction instead of sophisticated hardware design.The scattering pattern not only contains the information of particle size, but also of their shape. Particle shape analysis using light scattering is still in the initial stage. The scattering model for no-spherical particle is set up, and the numerical simulation study is carried out in this paper.At present, the laser sizer's inverse model is commonly based on spherical particle, so the measurement error for no-spherical particles cannot be ignored. Simulation indicates that no-spherical particles' measuring distribution is broadened evidently. The correlation coefficient of scattering spectrum is an effective way to describe the particle shape, which shape is more uniform. Four basic shapes are used to describe the particle shape distribution in this paper, and the spectral analysis can give this shape distribution.The novel CMOS (Complementary Metal Oxide Semiconductor) is used as the detector of the instrument. The CMOS pixel array can be arrange into ring-shaped units for size measurement or wedge-shaped ones for shape analysis by program. An experiment setup, which can analysis particle size and shape simultaneously is built up. The practice shows that CMOS take...