Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Over the past couple of years the topic of micro-and nanoswimmers has gained increasing interest, especially the use of Janus particles. The implementation of Janus particles as swimmers is widely reported, which is reflected in its ever-growing list of applications. However, research on their utilization at liquid-liquid interfaces is scarce. In order to apply Janus micro- and nanoswimmers at liquid-liquid interfaces, for example as microscopic sweepers, it is imperative that more research is performed in order to get a better understanding of how it influences the swimmers’ behaviour. The goal of this thesis was to provide fundamental research concerning the characterization and behaviour of sideways propelling microrods at the interface between water and decane and to compare it to the behaviour near the bottom wall. Polystrene microfibers with a diameter of 3 to 6µm were produced by electrospinning and coated with a bimetallic Pt/Au coating. Using ultrasonification these fibers were cut into pieces resulting in bimetallic Janus microrods. Next a spreading solution of rods was made and added into a specially designed cell with a circular well, filled with 5 wt.% hydrogen peroxide solution. The rods sank to the bottom and due to electrostatic repulsive forces remained on a height of approximately 550 nm above the bottom wall. Their movement was observed using an inverted optical microscope and image analysis showed that they propelled with a velocity of 2.16 ± 1.11 µm/s. Using a specially designed cell consisting of a well with an edge in the middle with water in the bottom half and purified decane on top, the behaviour of similar rods was examined at the water-decane interface by adding the same spreading solution on top of the decane layer. The rods would again sink down in the decane, but they were adsorbed to the interface, limiting their movement to the 2D interface. Analysis of the images, provided by the inverted optical microscope, revealed that the velocities were 2 to 4 times higher than near the wall. Furthermore the wetting behaviour of the microrods was examined by performing sessile drop contact angle measurements on polystyrene and both coating materials. Polystyrene proved to be hydrophobic but both coatings showed similar contact angles for water in decane and were considered neutral at first, but hydrophilic after contact with hydrogen peroxide. The position of the rods at the interface and the thickness of the Au coating was also examined by AFM measurements of the liquid-liquid surface. These measurements indicated that the rods protruded 1.34 ± 0.33 µm into the water phase and that the Au layer had a thickness of 21.4 nm.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Population balance modelling and its applications is an area that is growing rapidly. In chemical engineering, population balance equations (PBEs) can be used to model various heterogeneous population systems such as crystallization, precipitation, deposition, granulation, flocculation, drying and mixing, cell growth and many more processes. The control of heterogeneous population systems is important, these systems need to have certain characteristics that are imposed by the industry or user. Model predictive control can be selected to handle this as it is a well-established method with a simple algorithm, on the condition the system dynamics can be captured accurately in a model. Population balance modelling will generate a multi-variate partial differential equation of the particle system in the number density function. There exist many methods to solve the PBE for the number density function but none take into account the randomness and the uncertainty of the model that is obtained. The use of Gaussian processes (GPs) to solve regression problems in a non-parametric and probabilistic way, is a machine-learning method that has gained interest over the years. These GPs can be propagated through time by the use of numerical GPs. Numerical GPs can thus be applied as a solution method for the PBE to approximate the number density function over time. The strength of this method is that the model will include the uncertainty of the solution which can be implemented in the probabilistic framework of stochastic model predictive control (SMPC). The purpose of this thesis is to illustrate the feasibility of the implementation of a numerical GP solution scheme into a stochastic model predictive control scheme. First of all, the algorithm was created that combined the solution scheme of a numerical Gaussian process for a heterogeneous population system with a stochastic model predictive control algorithm. The newly developed method was then put to the test by applying it to a simple case study. The case study considered a crystallization process where the growth rate was assumed to be constant and the dilution rate became the controlled input. The control objective of the case study was to let the particle size distribution (defined as a number density function) converge to a defined steady state in a purposed time. The system could reach this control objective by adjusting the dilution rate starting from a chosen initial dilution rate. The investigation of this case study provides a proof of concept that numerical Gaussian processes can be combined with stochastic model predictive control for the determination of important characteristics of a heterogeneous population system.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Catalytic oxidation of HMF is among the most promising and extensively studied processes for FDCA production. A possible catalyst for this oxidation is a homogeneous Co/Mn/Br catalyst. However, the separation and recycling of the catalyst has proven to be challenging. Recycling of the metal ions does not only rely on economic and strategic incentives. The environmental impact of cobalt and manganese production is the main driving force for the recuperation of the metal ions from the waste stream in order to reuse them. Removing cobalt and manganese from the process waste stream is also adding value in the effort towards a more sustainable process. In this work, the application of cobalt and manganese removal from the waste stream, obtained by the production of FDCA using a homogeneous Co/Mn/Br catalyst, by metal ion separation processes through selective separation processes such as solvent extraction (SE) and electrodialysis (ED) is studied experimentally. For solvent extraction, polymer inclusion membranes (PIMs) are observed whereby a selective extractant is integrated in the membrane. Membrane based separation methods are known for their economical and sustainable way to extract and separate compounds. The effects of various operating conditions such as pH, residence time, current density, etc. have been investigated in order to characterize the performance of the techniques and determine optimal conditions. Leveraging the obtained experimental data, a conceptual design of the process is made. The experiments have shown that electrodialysis is a viable technique to remove cobalt and manganese metal ions from the waste stream. The optimal voltage has been determined. Furthermore, the effects of the presence of bromine and the concentration of acetic acid in the receiving solution has been investigated. It was found that a PIM with a composition of 55 wt% PVC and 45 wt % D2EHPA is able to separate manganese from cobalt. Furthermore, the optimal pH for the process has been determined. As the experiments are time intensive, addition of an electric current has been investigated to decrease the duration of the PIM tests. Finally, the effect of the applied electric current density on the separation performance of the PIM has been studied. A final two-stage experiment was conducted. In a first stage, an electrodialysis experiment has been performed using the determined, optimal conditions for the removal of cobalt and manganese from the waste stream. The aim of this experiment was to achieve the specified goal to lower the cobalt concentration in the dilute compartment or waste stream to 20 ppm. In a second stage, a PIM at optimal conditions has been used in order to separate the removed metal ions. Finally, a path forward has been suggested with some recommendations on what and how further research can fine tune the process and make it more cost efficient.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

For the purification of wastewater, a water purification installation is needed. In this installation, activated sludge or granules are used for the removal of pollutants from the wastewater. To monitor the working of the activated sludge or granules, a quantitative fluorescent image analysis can be used. The first goal of this work is finding a suitable fluorescent staining dye capable of staining all the biomass in activated sludge. This to make a quantitative comparison possible when the total biomass is stained in combination with other dyes with specific targets. The second goal is to design a filament detection method that can be used with a three-dimensional (3D) quantitative fluorescent staining. This to quantify the amount of filaments in activated sludge samples. For the quantitative fluorescent image analysis, different activated sludge samples of both flocs and granules are used. For the granules samples, also a slicing step was executed. The staining of the total biovolume is preformed with the dyes DAPI, Hoechst 33342, SYTO 9 and Rodamine 6G. Also ThT has been used as a validation step, wich only stains amyloids. For the staining, the procedures of Hoechst 33342, SYTO 9 and Rodamine 6G were optimised. After the staining, a 3D image was made in the image acquisitions step using a confocal laser scanning microscope (CLSM). In this step the image quality was optimised by changing the laser intensity and detector voltage. The image data was extracted using COMSTAT a program written in MATLAB and adapted from an existing program. In this program, different methods are compared to detect the total biovolume where the Otsu, cross-entropy and a new thresholding method was designed in this work. When comparing the used methods, the Otsu and new threshlding methode gave an under and over estimation respectively, the cross-entropy thresholding method seemed to give the best result. The comparison of the different stainings shows that the result of the SYTO 9, DAPI and Hoechst 33342 staining gave about the same biovolumes, the Rodamine 6G detected more biovolume. Not one of the dyes detected a higher biovolume than ThT, resulting in an incomplete staining of the biovolume for every dye. Though, when a washing step is introduced in the staining process of the Rodamine 6G dye, probably more biovolume could be detected. The designed filament detection method showed a good detection of filaments throughout different activated sludge flocs samples. The added value of the filament detection in 3D and of a specific dye is that it enables the comparison between different dyes. The created filament detection method is based on a two-dimensional (2D) method which is adapted for 3D images.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

A camphene-based suspension with alumina as the solid phase was prepared by mixing for 3 h at 600 rpm, at 60◦C with an overhead stirrer. By performing an amplitude stress sweep, it was found that the storage modulus (G