Powder morphology development during spray drying

Abstract

Spray drying is widely applied in food industry to convert liquid formulations into powders to facilitate transport and extend their shelf-life. Spray dried powders have superior quality due to their excellent reconstitution behavior and the relative mild drying process which preserves product quality. One of the key factors in determining product quality is the morphology of the primary powder particle, which influences reconstitution behavior and flowability. However, the complex phenomenon of morphology development, especially as function of material composition and drying conditions, has not been topic of in-depth scientific study. Better understanding of morphology development is expected not only to contribute to improved powder quality, but also to improved efficiency of spray drying operations. Specifically, lack of control on particle formation and stickiness behavior increases risk of fouling in spray drying towers, which leads to extended downtime and loss of material. The main objective of the research in this thesis was thus to create mechanistic understanding of morphology development of drying droplets, which was divided in two main research questions: How are skin formation and subsequent morphology development affected by the drying conditions or product formulation? Can the rheological properties of components at high concentration explain the morphology development during drying? To answer these research questions multiple methods were employed. Single droplet drying was used to observe morphology development, thin film drying was used to study skin formation and its effect on drying kinetics, and rheology was used to measure properties of model components. Finally, pilot-scale experiments were done to translate findings of this research to larger-scale spray drying. In chapter 2 the influence of drying temperature and droplet composition on the formed particle morphology was studied for model components whey protein and maltodextrin DE12. It was observed that after a period of constant drying rate, the particle morphologies evolve in interaction with droplet composition and drying temperature. Droplets with a high concentration of protein yielded smooth and hollow particles, whereas particles with more maltodextrin were wrinkled and had multiple smaller vacuoles. Particles dried at lower temperatures were more likely to form hollow particles, whereas particles dried at higher temperature had the tendency to become wrinkled with fixed ratio of maltodextrin to whey protein. It was hypothesized that hollow particle formation at lower temperatures is related to phase separation of the individual components. This was confirmed by Confocal Raman Imaging, where it was observed that at lower temperatures (40oC) demixing occurred and at higher temperature (90oC) the system was well mixed. This work was extended with the drying of micellar casein and lactose, and eventually to the drying of whole milk in chapter 3. It was found that the influence of carbohydrates (lactose and maltodextrin DE12) and proteins (micellar casein or whey protein) on morphology development is very different, since upon concentration protein systems will jam and undergo a colloidal glass transition, whereas carbohydrate systems will gradually increase in viscosity as a consequence of the concentration. Whey protein gives relatively rigid shells due to jamming of the ‘hard sphere’ proteins, while casein micelles behave as ‘soft spheres’ that can deform after jamming, which gives flexibility to the shell during drying. The influence of the carbohydrates on the final morphology was found much larger than the influence of the proteins. Caseins influenced morphology only in mixtures with lactose at higher concentrations due to its high voluminosity. Similar observations were done for whole milk, where fat appeared to have no influence and the morphology was mainly determined by the presence of casein. In casein and maltodextrin mixtures it was observed that maltodextrin subtly influenced particle morphology via the shape and smoothness of wrinkles. In chapter 4 the work on maltodextrin and whey protein was continued, however this time the effect of initial droplet size and dry matter content was studied. Especially, the latter provided insight on the skin formation during drying, and therefore these results were combined with viscosity measurements. Shear rate sweeps showed jamming of the whey protein at concentrations of ~50% (w/w), whereas maltodextrin remained liquid-like up to concentrations of ~70% (w/w). Morphology development of the latter components during single droplet drying showed that it could be influenced by altering initial droplet size and dry matter content. If droplets had a high initial dry matter (50% (w/w) morphology development started immediately, and the formed morphology could be related to the rheological behavior of the mixture at that concentration. This indicated that measuring the rheological properties at high concentrations can provide insight in morphology development. In chapter 5 a different approach was chosen to understand the drying behavior, using thin films as a model system. First, the drying kinetics were measured using a specially developed thin film dryer. Subsequently, the viscoelastic properties of nearly dried films were studied. Drying kinetics of thin films were monitored in a custom-built drying equipment. Subsequently, the rheological properties of equilibrated thin films were assessed by oscillatory shear measurements at relevant high dry matter contents (66-82 w/w%). During drying, the samples high in whey protein formed brittle films and had lower evaporation rates, compared to films high in maltodextrin. From rheology analysis it was observed that for whey protein rich (>25%) systems, the samples were in structural arrest at the dry matter contents measured. Maltodextrin on the other hand showed typical viscoelastic polymer behavior, although as little as 1% addition of whey protein altered the viscoelastic properties drastically. Lastly, these properties were related to vacuole formation during single droplet drying: samples that undergo structural arrest at a lower dry matter content, i.e. high in whey protein, will form less and larger vacuoles compared to samples that undergo structural arrest only at high concentration, i.e. high in maltodextrin. Finally, the results from our single droplet drying research were compared with the morphology development in a pilot scale spray dryer in chapter 6. This showed the importance of looking at drying from a product perspective, and emphasized the industrial relevance of the project. A pilot scale single stage spray dryer was used with a maximum evaporation capacity of 80 kg/h. The effect of composition on morphology was studied by drying maltodextrin DE12, whey protein, and mixtures thereof with a ratio of 75:25 and 50:50 (WP:MD). The analysis was carried out with the Malvern Morphologi 4 analyzer. The morphologies observed in pilot scale drying were similar to the observations seen in single droplet drying, which showed the usefulness of this method, despite the different time scales of drying. This shows that component properties and interactions play a more important role than the drying conditions, and shows the relevancy of the single droplet drying methods for observing morphology development. Moreover, altering the drying conditions (specifically reducing the inlet air temperatures) could increase the number of wrinkled particles from 19% to 34% for a 75:25 (WP:MD) powder, which showed that morphology could be steered by the drying conditions. In summary, the combined approach of single droplet drying and rheology showed great value in understanding morphology development in spray drying, despite the longer drying times in single droplet versus spray drying. Using this approach it was shown that morphology depended greatly on the component properties, and could be steered by altering the drying conditions. This knowledge can be applied in industry to steer their spray drying operations better on powder morphology and thus resulting powder properties. This know-how can also be exploited to develop new strategies to mitigate risk of fouling, where for example earlier locking influences stickiness. Outlook for further research could be directed at developing single droplet drying methods with small droplet size, relating the morphology to powder properties using pilot-scale spray drying studies, and modifying protein properties to steer powder functionality.