The pharmaceutical industry has long struggled with monitoring the characteristics of a drying mixture, a critical step in the production of drugs and compounds. Currently, there are two noninvasive characterization approaches that are typically used: the sample is imaged and individual molecules are counted, or scientists exploit scattered delicate to estimate the particle size distribution (PSD). The former is time-consuming and leads to increased waste, making the latter a more attractive option.
In recent years, MIT engineers and scientists have developed a physics- and machine-learning-based approach using scattered delicate that has been shown to improve manufacturing processes for pharmaceutical tablets and powders, increasing efficiency and accuracy and resulting in fewer failed batches. A fresh open-access paper, “Non-invasive powder size distribution assessment from a single spot image”, available in the journal, extends this work and introduces an even faster approach.
“Understanding the behavior of scattered light is one of the most important topics in optics,” says Qihang Zhang PhD ’23, an assistant professor at Tsinghua University. “With the advances in analyzing scattered light, we have also invented a useful tool for the pharmaceutical industry. Locating the trouble spot and solving it by investigating the basic principle is the most exciting thing for the research team.”
The paper proposes a fresh method for estimating PSD, based on pupil engineering, which reduces the number of frames needed for analysis. “Our learning-based model can estimate the powder size distribution from a single snapshot of the speckle, which consequently reduces the reconstruction time from 15 seconds to just 0.25 seconds,” the researchers explain.
“Our main contribution in this work is that we accelerated the particle size detection method by 60 times, with the collective optimization of both the algorithm and the hardware,” says Zhang. “This fast probe is capable of detecting size evolution in fast dynamic systems, providing a platform for studying process models in the pharmaceutical industry, including drying, mixing and blending.”
The work, a successful collaboration between physicists and engineers, is the result of an MIT-Takeda program. The collaborators are affiliated with three MIT departments: mechanical engineering, chemical engineering, and electrical engineering and computer science. George Barbastathis, a professor of mechanical engineering at MIT, is the senior author of the paper.