C.G.P.H. Schroën

Classification and evaluation of microfluidic devices for continuous suspension fractionation

Membrane processes are well-known for separating and fractionating suspensions in many industries, but suffer from particle accumulation on the membrane surface. Currently, there are new developments using microfluidic devices for cell/DNA sorting and fractionation. We anticipate these devices are also applicable to fractionation of polydisperse and concentrated suspensions (e.g. foods), and may potentially have fewer problems with particle accumulation compared to membranes. This review article presents an overview of relevant microfluidic devices. We focus on their performance with respect to concentrated suspensions, as one finds in food industry. We give quantitative estimates on their yield, selectivity, and the potential for large-scale application. From this evaluation follows that deterministic ratchets seem most promising.

Membrane modification to avoid wettability changes due to protein adsorption in an emulsion/membrane bioreactor

This study addresses problems encountered with an emulsion/membrane bioreactor. In this reactor, enzyme- (lipase) catalyzed hydrolysis in an emulsion was combined with two in-line separation steps. One is carried out with a hydrophilic membrane, to separate the water phase, the other with a hydrophobic membrane, to separate the oil phase. In the absence of enzyme, sunflower oil/water emulsions with an oil fraction between 0.3 and 0.7 could be separated with both membranes operating simultaneously. However, two problems arose with emulsions containing lipase. First, the flux through both the hydrophilic and the hydrophobic membranes decreased with exposure to the enzyme. Second, the hydrophobic membrane showed a loss of selectivity demonstrated by permeation of both the oil phase and the water phase through the hydrophobic membrane at low transmembrane pressure. These phenomena can be explained by protein (i.e. lipase) adsorption to the polymer surface within the pores of the membrane. It was proven that lipase was present at the hydrophilic membrane and that this, in part, explains the flux decrease of the hydrophilic membrane. To prevent the observed loss of selectivity with exposure to protein, the hydrophobic polypropylene membrane (Enka) was modified with block copolymers of propylene oxide (PO) and ethylene oxide (EO). These block copolymers act as a steric hindrance for proteins that come near the surface. The modification was successful: After 10 days of continuous operation the minimum transmembrane pressure at which water could permeate through an F 108-modified membrane was 0.5 bar, the same value as that observed in the beginning of the experiment. This indicates that loss of selectivity due to protein adsorption is prevented by the modification of the membrane.

Suspension flow in microfluidic devices — A review of experimental techniques focussing on concentration and velocity gradients

Microfluidic devices are an emerging technology for processing suspensions in e.g. medical applications, pharmaceutics and food. Compared to larger scales, particles will be more influenced by migration in microfluidic devices, and this may even be used to facilitate segregation and separation. In order to get most out of these completely new technologies, methods to experimentally measure (or compute) particle migration are needed to gain sufficient insights for rational design. However, the currently available methods only allow limited access to particle behaviour. In this review we compare experimental methods to investigate migration phenomena that can occur in microfluidic systems when operated with natural suspensions, having typical particle diameters of 0.1 to 10 μm. The methods are used to monitor concentration and velocity profiles of bidisperse and polydisperse suspensions, which are notoriously difficult to measure due to the small dimensions of channels and particles. Various methods have been proposed in literature: tomography, ultrasound, and optical analysis, and here we review and evaluate them on general dimensionless numbers related to process conditions and channel dimensions. Besides, eleven practical criteria chosen such that they can also be used for various applications, are used to evaluate the performance of the methods. We found that NMR and CSLM, although expensive, are the most promising techniques to investigate flowing suspensions in microfluidic devices, where one may be preferred over the other depending on the size, concentration and nature of the suspension, the dimensions of the channel, and the information that has to be obtained. The paper concludes with an outlook on future developments of measurement techniques.

Mixed motion in deterministic ratchets due to anisotropic permeability.

Nowadays microfluidic devices are becoming popular for cell/DNA sorting and fractionation. One class of these devices, namely deterministic ratchets, seems most promising for continuous fractionation applications of suspensions (Kulrattanarak et al., 2008 [1]). Next to the two main types of particle behavior, zigzag and displacement motion as noted by the inventors (Huang et al., 2004 [2]) and (Inglis et al., 2006 [3]), we have shown recently the existence of a intermediate particle behavior, which we named mixed motion. In this paper we formulate the hypothesis that the occurrence of mixed motion is correlated with anisotropy in the permeability of the obstacle array. This hypothesis we base on the comparison of experimental observations of mixed motion and the flow lane distribution as obtained from 2-D flow simulations.

Wettability of tri-block copolymer coated hydrophobic surfaces Predictions and measurements

Abstract Hydrophobic surfaces with adsorbed tri-block copolymers are wetted by oil in spite of the hydrophilic buoy groups of the block copolymer that are present near the surface. The effect of the buoy group length of the adsorbed molecules on the wettability of hydrophobic surfaces is studied by contact angle measurements and by computer modelling. The computer model predicts an increase in interfacial free energy with increasing buoy group length for equilibrium adsorption of block copolymer from water. Molecules with large buoy groups occupy more lateral space; therefore the “bare” surface gets more exposed and the anchor groups contribute less to the interfacial free energy which thus increases with the buoy group length. The calculations showed that the variation of the interaction parameter between solvent and buoy group hardly influences the interfacial free energy. In contrast the interaction parameter between solvent and surface influences the interfacial free energy to a large extent because the oil/surface interactions have a lower energetic value as compared to water/surface interactions and therefore the interfacial free energy is lower than in water. The interfacial free energy varies slightly with increasing buoy group length, depending on the value chosen for the solvent/surface interaction parameter. Advancing and receding contact angles of hexadecane, sunflower oil and hydrolysate (partly hydrolysed sunflower oil) were measured on hydrophobic surfaces. All oil/water contact angles were small, indicating a hydrophobic apolar surface character. It was found that, for oils with a “good” interaction with the surface (hexadecane and sunflower oil), the contact angle has a minimum value at a certain buoy group length. For hydrolysate (less-strong interaction with the surface) the contact angle decreases monotonically with increasing buoy group length. The results for hexadecane, sunflower oil and hydrolysate are in reasonable agreement with the model predictions. The effect of increasing buoy group length is weak; both decreasing and increasing angles are found, depending on the type of oil used.

Preparation of a protein-repelling, yet hydrophobic membrane by treatment with non-ionic block copolymers.

Protein (lipase) adsorption at hydrophobic microfiltration membranes leads to more hydrophilic membrane properties. This is undesirable in case of removal of oil from protein-containing emulsions by membrane separation. Therefore, protein adsorption has to be prevented. This was done by modification of hydrophobic surfaces/ membranes with a polypropylene oxide/polyethylene oxide block copolymer (Synperonic, F108). The block copolymer adsorbs onto a hydrophobic surface in a ’brush’ conformation which results in an effective steric hindrance for proteins that come near the surface. From adsorption and wettability experiments it could be concluded that at F108-modified surfaces no protein adsorption and, hence, no wettability change takes place. With an F108-modified hydrophobic membrane, separation of oil from a lipase containing sunflower oil in water emulsion is possible.

Integration of reaction and recovery by a continuous emulsion enzyme reactor with in-line removal of the oil and waterphase by membrane separation.

The emulsion/membrane bioreactor consists of an emulsion contained in a stirred tank and two membrane units. In the emulsion an enzyme catalyzed reaction is carried out. The oil phase and the water phase are separated by a hydrophobic and a hydrophilic membrane respectively.