Moritz Moß

Single particle light scattering has gained popularity due to its methodological simplicity, short measurement times, high precision, and suitability for high-throughput particle analysis. This work quantitatively validates two flow cytometry prototypes with regard to particle size and concentration determination based on qualitative studies conducted previously. First, the influence of the interrogation volume on the distribution of the measured scattering signal was examined. The broadening of the scattering signals was found to depend on the diameter of the sample stream, with smaller diameters leading to less broadening. To detect particle sizes ranging from 100 to 1000 nm, an intensity range of at least five orders of magnitude must be detectable. The new device uses two linear amplifiers to enable coverage of six orders of magnitude. Validation of particle size measurements was performed using certified reference standards, resulting in a bias of 8.4 nm and a relative standard deviation of 0.27%. A working range of 30 to 1600 nm, depending on the material, was established. The analysis of a multimodal size distribution was demonstrated. Finally, particle concentrations in the range of 105 to 108 particles per ml could be accurately determined. The amount of fine and coarse fractions in multimodal systems could also be reliably identified. Thus, not only did the validation show excellent results but also offers opportunities for future research on highly complex multimodal systems with widely distributed properties, which has been made possible due to these new technical developments.

Applications for machine learning (ML) and deep learning (DL) are constantly growing and have already been adopted in the field of particle measurement technology. Even though analytical (ultra-)centrifugation (AC/AUC) is a widely used technique for characterizing dispersed particle systems, ML and DL have not yet been applied in this area. Data evaluation and interpretation in AC/AUC can be challenging and often requires expert knowledge. DL models can help, but their development is limited by a lack of annotated training data. One solution is to generate and use synthetic data instead. In the first part of this study, a model was trained to synthesize data from experiments using a combination of Variational Autoencoder (VAE) and Generative Adversarial Networks (GANs). The results appear highly realistic. Novice users could distinguish real from synthetic samples with only 63% accuracy. Then, a classifier was trained on experimental AC data to categorize real-world examples based on their underlying separation kinetics, testing different DL architectures. After initial training, the models were further fine-tuned with synthetic AC data. ResNet34 models achieved the best performance with 94% accuracy, comparable to an AC expert (91%), while inexperienced users reached only 53%. In the second part of our study, a regression model was trained for the analysis of sedimentation coefficients. Therefore, various generative models were developed and evaluated for synthesizing AUC data based on numerically simulated sedimentation boundaries. The best results were achieved by combining VAE and GAN architectures with embedded physical constraints. However, the generative networks have so far led to additional smearing of the profiles, resulting in a broadening of the sedimentation coefficient distribution and indicating that further refinement is necessary.