Our research vision

We want to understand how complex and in especially biological fluids flow and which microscopic mechanisms and properties are responsible for the macroscopic flow phenomena.

Red Blood Cells

A project that is related to the flow properties of blood is the study of the contribution of red blood cells in blood clot formation. In order to test this hypothesis, we built up an integrated microfluidic holographic optical tweezers setup to study cell adhesion and its effect on the blood flow. Recently we extended our know-how to AFM based single cell spectroscopy. This has been applied to the platelet induced adhesion of red blood cells but we could also quantify for the first time the adhesion energies between red blood cells in their physiological discocyte shape in the case of rouleaux formation. The results could be well explained with a depletion model even if for longer times some indications for bridging were observed, too.

The supply of oxygen and nutrients and the disposal of metabolic waste in the organs depend strongly on how blood, especially red blood cells, flow through the microvascular network. Macromolecular plasma proteins such as fibrinogen cause red blood cells to form large aggregates, called rouleaux, which are usually assumed to be disaggregated in the circulation due to the shear forces present in bulk flow. This leads to the assumption that rouleaux formation is only relevant in the venule network and in arterioles at low shear rates or stasis. Thanks to an excellent agreement between combined experimental and numerical approaches, we show that despite the large shear rates present in microcapillaries, the presence of either fibrinogen or the synthetic polymer dextran leads to an enhanced formation of robust clusters of red blood cells, even at haematocrits as low as 1%. Robust aggregates are shown to exist in microcapillaries even for fibrinogen concentrations within the healthy physiological range. These persistent aggregates should strongly affect cell distribution and blood perfusion in the microvasculature, with putative implications for blood disorders even within apparently asymptomatic subjects.


We studied the mobility of nanoparticles in mucus and similar hydrogels as model systems to elucidate the link between microscopic diffusion behavior and macroscopic penetration of such gels. Differences in particle adhesion to mucus components were strongly dependent on particle coating. Particles coated with 2 kDa PEG exhibited a decreased adhesion to mucus components, whereas chitosan strongly increased the adhesion. Despite such mucoinert properties of PEG, magnetic nanoparticles of both coatings did not penetrate through native respiratory mucus, resisting high magnetic forces (even for several hours). However, model hydrogels were, indeed, penetrated by both particles in dependency of article coating, obeying the theory of particle mobility in an external force field. Comparison of penetration data with cryogenic scanning EM images of mucus and the applied model systems suggested particularly high rigidity of the mucin scaffold and a broad pore size distribution in mucus as reasons for the observed particle immobilization. Active probing of the rigidity of mucus and model gels with optical tweezers was used in this context to confirm such properties of mucus on the microscale, thus presenting the missing link between micro- and macroscopical observations. Because of high heterogeneity in the size of the voids and pores in mucus, on small scales, particle mobility will depend on adhesive or inert properties. However, particle translocation over distances larger than a few micrometers is restricted by highly rigid structures within the mucus mesh.

Non equilibrium fluctuations

Another approach on the microscopic scale is the study of Brownian motions in shear flow by use of optical tweezers. Shear-induced cross-correlations of particle fluctuations perpendicular and along stream-lines were investigated experimentally and theoretically. Direct measurements of the Brownian motion of micron-sized beads, held by optical tweezers in a shear-flow cell, showed a strong time-asymmetry in the cross-correlation, which was caused by the non-normal amplification of fluctuations. Complementary measurements on the single particle probability distribution substantiated this behavior and both results were consistent with a Langevin model. In addition, a shear-induced anti-correlation between orthogonal random-displacements of two trapped and hydrodynamically interacting particles was detected, having one or two extrema in time, depending on the positions of the particles. Recently, this work has been extended to study the correlations of the Brownian movement in active suspensions (Chlamydomonas reinhardtii).

DNA separation

In my time at Agilent Technologies, I worked on the electropheretic analysis of biological macro molecules (DNA, RNA and proteins) on highly integrated Lab-on-a-chip microstructures in the Research and development Department. I was member of a team that developed an RNA assay that is capable of qualifying microdissected samples and in the development of an automated high throughput Lab-chip platform. Furthermore, I was responsible for the surveillance of the chip production at Caliper Technologies, Palo Alto.


Another main topic in my research group is the flow behaviour of droplets. We characterized the detachment process of a droplet of a diluted DNA solution and in such a way it is possible to measure the elongational viscosity. DNA is a polyelectrolyte and its flexibility can be tuned by the amount of salt added to the solvent. It was shown that elongational viscosity increases with the flexibility of the polymer, in contradiction to the FENE simulations which predict that elongational viscosity is independent of polymer flexibility at higher shear rates. First success has been achieved in gaining insight on the mechanisms that determine the observed elongational rates, and some phenomena of the final detachment process which could be reproduced in the numerical simulations of Prof. Dr. Jens Eggers, Bristol. The experimental values of elongational viscosity were used to explain some features of turbulent drag reduction of diluted DNA solutions. It was shown explicitly for the first time that elongational viscosity is the pertinent quantity for the characterization of the ability of a polymer to perform turbulent drag reduction . In order to gain insight on the molecular conformations of the macromolecules in capillary break up we developed a high speed birefringence system that allows measuring the optical retardation and the thickness of the capillary filament simultaneously. Our experiments show that rigid molecules like Xanthan are rapidly oriented in the elongational flow but cause nevertheless a prominent elongational viscosity while elastic molecules like DNA are continuously stretched. More recently we could use our droplet measurements to show that human blood plasma is a viscoelastic fluid. The viscoelasticity of blood plasma has been debated for decades and indeed in shear flow there is no detectable viscoelasticity but in elongational flow it can be detected. To some extend this was possible because we improved our experimental with a superresolution

Flow instabilities of simple and complex fluids

We investigate the pattern formation in dissipative systems and the flow dynamics of simple, complex and biological liquids. The hydrodynamic model systems for the study of pattern formation are the Faraday-Experiment, the Saffman-Taylor-Instability, the Taylor-Couette-Experiment and channel flow. The Faraday experiment refers to the appearance of ordered capillary waves on a free liquid surface that is oscillated vertically. Usually, these surface waves respond with half of the driving frequency (subharmonic). However, both in Newtonian- and in polymeric-liquids we showed experimentally that a harmonic (synchronous to the drive) response exists too. In polymeric liquids the harmonic response is observed if the external drive period is similar to the relaxation time of the polymers while in Newtonian liquids it is to observe if damping from the bottom becomes significant. Due to the interaction of subharmonic and harmonic modes, complex, but still highly symmetric patterns like superlattices appear. Superlattices are well known in solid state surface physics, but on a macroscopic scale they have been found only recently. A major achievement was the first experimental characterization of the full capillary wave spectrum of the surface waves.

We also work on elastic instabilities of physical gels. We characterized an oscillatory instability in a Hele Shaw experiment and we could explain the phenomenon quantitatively (characteristic time and length scales) by use of the according viscometric data. Furthermore, it was shown that PVA (Polyinylalkohol) with Borax as a cross linker is a suitable model system for the study of melt fracture. Melt fracture is a serious technical problem in many industrial applications of processing of synthetic materials. For the first time it was now possible to characterize the phenomenon as an elastic instability, also called viscoelastic turbulence. Theoretical calculations of the critical velocity and the critical wave number have been found to be in very good agreement with the experimental results. In especially we could also show that that maximum drag reduction asymptote in intertial turbulence might be simply a state of inertial-elastic turbulence.

Wet granular materials

We could prove that wet sand that is known to be a yield stress fluid flows better than dry sand by measuring the force needed to push an idealized form of sand through a circular tube. The "sand" consisted of an aqueous slurry of tiny glass beads with a diameter of 145 microns, which is roughly that of a grain of sand. We found that less energy was required to push sand through the tube if it was wet than if it was dry. Dry sand usually flows freely because air voids can form between the grains, which keeps them apart and reduces friction. But when the sand is packed in a tube, there is no room for voids to occur and therefore air cannot lubricate the flow. As a result, friction causes the dry sand to jam. If the sand in the tube is wet, the water initially acts like glue, causing individual grains to bind together, just as in a sandcastle. But if sufficient force is applied, the bonds between grains are broken and the water acts as a lubricant, causing the sand to flow more easily. Furthermore, we showed experimentally that the sliding friction on sand is greatly reduced by the addition of some—but not too much—water. The formation of capillary water bridges increases the shear modulus of the sand, which facilitates the sliding. Too much water, on the other hand, makes the capillary bridges coalesce, resulting in a decrease of the modulus; in this case, we observe that the friction coefficient increases again. Our results, therefore, showed that the friction coefficient is directly related to the shear modulus; this has important repercussions for the transport of granular materials. In addition, the polydispersity of the sand is shown to also have a large effect on the friction coefficient.

↑ top