Sticky, Squishy & Stuck: A soft matter approach to membrane failure

Ties van de Laar

Research output: Thesisinternal PhD, WU


Separation of components is carried out at many different levels, ranging from sieving of particles, to gas separation. In all these fields membranes are known as greener, more sustainable options, and therefore it is not surprising that their use has seen a very steady increase over the years. Membrane filtration processes are widely used in large number of industrial applications, ranging from dairy processing, to pharmaceutical fractionation and waste-water treatment. For membrane processes such as micro- and ultrafiltration, in which pores act as gatekeepers, one of the major challenges is fouling and the associated clogging of membrane pores. Such clogging and fouling occurs due to the gradual accumulation of solid or semisolid matter in and on the membrane during its operational lifetime. This accumulated layer is commonly referred to as a filter cake and blocks flow through the membrane, which reduces the lifetime of a membrane and requires extensive cleaning. Delaying or preventing fouling would increase the lifetime of membranes and would contribute to a more sustainable process. However, solving the complex problem of fouling requires fundamental insight into what occurs within a membrane at the microscale, and that is generally not the scale that is considered when developing membrane systems.

In this thesis we combined experiments and theory to better understand the processes occurring near, at and within filtration membranes in order to further our understanding of fouling and clogging. We create links between problems encountered in membrane technology and those studied in soft matter science. We focus on three major themes; i) the structure and dynamics of dense and non-ideal particle packings, ii) confined flow of complex fluids and iii) the mechanisms of pore clogging. By studying these problems, we have not only gained insight in the microscopic processes and mechanisms that underlie membrane filtration but also further the fundamental understanding of these classical soft matter phenomena.

In Chapter 2 we provide a technical overview of the various synthesis routes used in this thesis to design and make colloidal model systems that are tailored to the specific requirements of our various experiments. In Chapter 3 we explore the effect of attractive forces on the colloidal glass transition. We show that attraction drastically changes the behavior of a colloidal glass, resulting in a solid-solid transition. We provide the first experimental proof of the predicted discontinuity for this solid-solid transition. The presence of small depletants which induce attractive interactions are common in real-world processing streams; studying the attractive glass transition brings us one step closer to linking colloidal glasses and the behavior of the fouling layer that forms on top of a membrane.

While hard spheres are extensively studied in soft matter science, these systems represent a very idealized situation. In the daily practice of large-scale processing, soft deformable particles are expected to determine the system behavior. In Chapter 4 we derive a microscopic theory to explain how softness changes the behavior of colloidal glasses, in particular their fragility and strength. Our theory provides a universal description of the glass transition for soft colloidal systems. We use our theory to explain the mechanisms of fragility, which has a large influence on the ability to process glasses. As such, it will have a large influence on the properties of e.g. a filter cake. In Chapter 5 we take this research one step further and investigate how soft particles deform in dense packings, and show that both shrinkage due to compression and deformation must be taken into account. This research opens up a way to understand the behavior of soft particles moving through constricted membrane pores, and builds the foundation to quantify differences encountered during filtration of rigid and soft particles.

In Chapter 6 we show how flow through very narrow channels can be used to achieve a pre-fractionation. Our results explore the limits of this fractionation technique (called shear-induced diffusion) to understand if, and how such fractionation occurs at very high concentrations up to volume fractions of 0.6 or higher for deformable particles. Ultimately, this could lead to energy efficient fractionation processes that can be carried out at higher concentrations. Such a process would be radically different from current process technologies which focus on diluted feed streams. In Chapter 7 we develop a microfluidic device to study and visualize two-phase flow in a random porous network. We show that by understanding the complex displacement cascade we can tailor the displacing fluid properties in such a way that we can greatly enhance the displacement efficiency. These results illustrate the power of direct visualization at the microscopic scale to gain quantitative understanding of the important parameters. In Chapter 8 we make the first steps toward a novel flow sensor that would enable the direct, quantitative visualization of flows in extreme confinement (i.e. nm scale). These sensors would offer unprecedented visualization capabilities of flows even at the nanometer scales. We calibrate these sensors at the single-molecule scale, which is the first demonstration of optical force sensing in single molecules.

Finally, we investigate pore clogging, which leads to the build-up of a filter cake and is the start of decreased membrane performance. A large variety of factors play a role in how, when and where a clog forms. To disentangle these effects, we developed microfluidic tools to study clogging at the single pore level. In Chapter 9 we show how small changes in the pore geometry have a massive impact on the clogging behavior. The angle under which a constriction is placed with respect to the flow direction can delay clogging by at least a factor 4. Again, we show the large effects that slight attractive forces can have, leading to a decrease in clogging time of more than an order of magnitude. We account for both these effects by developing a new model based on transition-state theory. The generated insights can be used in the design of smarter, micro-engineered membranes in which the effects of clogging are delayed, and ideally prevented. In Chapter 10 we take our microfluidic approach one step further and modify our design to emulate a cross-flow membrane system; the most prevalent membrane filtration configuration in industry. We show that the flow through the membrane completely dictates the clogging behavior in the pore and we extend our model to explain these effects. Moreover, we show that, while the cross-flow flux does not influence the rate of primary clogging directly, it does influence the build-up of a cake layer and its effect on neighboring pores. These results help us understand which factors play a pivotal role in membrane clogging, and form a basis from which we could design new, and improved membrane systems.

The work in this thesis highlights how a soft matter approach, utilising the tools from soft matter science can help shed light on the complex processes that occur during membrane filtration. Through direct visualization at the microscopic scale, combined with quantitative theory we gain a deep, microscopic understanding of a wide variety of processes. In the General Discussion we put our work in a broader context, provide routes to develop the work further and reflect on how our results can be used in the future design of innovative mem-brane filtration systems.

Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Wageningen University
  • Schroen, Karin, Promotor
  • Sprakel, Joris, Co-promotor
Award date25 May 2018
Place of PublicationWageningen
Print ISBNs9789463437578
Publication statusPublished - 25 May 2018


  • Cum laude


Dive into the research topics of 'Sticky, Squishy & Stuck: A soft matter approach to membrane failure'. Together they form a unique fingerprint.

Cite this